ISSN 0021-2571 (print) · 2384-8553 (online) | Coden: AISSAW 56 (No. 3) | 257-402 (2020) Annali DELL’ISTITUTO SUPERIORE DI SANITÀ

ANNALI Volume 56 No. 3 2020 DELL’ISTITUTO SUPERIORE DI SANIT À - Volume 56, No. 3, 2020

A science journal for public health

www.iss.it Ann Ist Super Sanità 2020 | Vol. 56, No. 3 Annali dell’Istituto Superiore di Sanità A science journal for public health

Publication Annali dell’Istituto Superiore di Sanità is published quarterly and in special issues. Freely available online at www.iss.it/annali1 - www.annali-iss.eu

Annali dell’Istituto Superiore di Sanità is indexed in - CAB - CHEMABS - EMBASE/Excerpta Medica - FSTA - MEDLINE - SCOPUS - WEB OF SCIENCE

The Journal Impact Factor is 1.370

Annali Editorial Office Scientific Communication Service Istituto Superiore di Sanità Viale Regina Elena 299, 00161 Rome, Tel.: +39 06 49902945 Fax: +39 06 49902253 E-mail: [email protected] www.iss.it/annali1 - www.annali-iss.eu Papers to be presented for publication should be submitted online to www.annali-iss.eu. Instructions to Authors are available online at www.annali-iss.eu

Publishing support Il Pensiero Scientifico Editore, Rome Via San Giovanni Valdarno 8, 00138 Rome, Italy www.pensiero.it

Subscription information & terms Il Pensiero Scientifico Editore Tel.: +39 06 86282324 Fax: 06 86282250 E-mail: [email protected] Year 2020 Italy individual subscription € 57,00 | Italy institutional subscription € 67,00. Other countries € 67,00 Each quarterly issue € 21,00 Responsibility for the contents and opinions expressed on this journal rests solely with the Author(s). ISSN 0021-2571 (print), 2384-8553 (online) Coden: AISSAW 56 (No. 3)

©Istituto Superiore di Sanità 2020. Some rights reserved.

Reg. Stampa - Tribunale di Roma, n. 482 del 29 ottobre 1985 (cartaceo); n. 121 del 16 maggio 2014 (online)

Printed in September 2020 by Ti Printing s.r.l. Via Case Rosse 23, 00131 Rome, Italy Ann Ist Super Sanità 2020 | Vol. 56, No. 3 Annali dell’Istituto Superiore di Sanità A science journal for public health

President of the Istituto Superiore di Sanità Silvio Brusaferro

Responsible Director Paola De Castro Editor-in-chief Enrico Alleva Istituto Superiore di Sanità Assistant Editor Federica Napolitani Cheyne Istituto Superiore di Sanità Scientific Committee Enrico Alleva, Luca Busani, Pietro Comba, Paola De Castro, Paola Fattibene, Alessandro Giuliani, Carlo Petrini, Giovanni Rezza, Emanuele Scafato, Marina Torre Istituto Superiore di Sanità Editorial Office Annarita Barbaro, Maria Cristina Barbaro, Alessandra Fuglieni, Federica Napolitani Cheyne, Laura Radiciotti Istituto Superiore di Sanità OJS Site Administrator Daniele Cordella Istituto Superiore di Sanità Editorial Advisory Board Chris Beyrer Johns Hopkins Bloomberg School of Public Health, Baltimore, USA Jacqueline N. Crawley Laboratory of Behavioral Neuroscience, NIMH, NIH, Bethesda, USA Juana Martín de las Mulas González-Albo Dpto de Anatomía y Anatomía Patológica Comparada, Facultad de Veterinaria, UCO, Córdoba, Spain Jean-François Delfraissy Agence Nationale de Recherches sur le sida et les hépatites virales, Paris, Tony Fletcher London School of Hygiene & Tropical Medicine, London, UK Sui Huang Institute for Systems Biology, Seattle, USA Stefania Maggi CNR Aging Branch Institute of Neuroscience, Padova, Italy Francesco M. Marincola Infectious Disease and Immunogenetics Section, NIH, Bethesda, USA Ana Marušic´ Dept of Research in Biomedicine and Health, University of Split School of Medicine, Croatia Patrick Smeesters Université Catholique de Louvain, Louvain-La-Neuve, Kurt Straif ISGlobal, Barcelona and Visiting Professor at Boston College David Vlahov School of Nursing, University of California, San Francisco, CA, USA Bernard Zalc (Boris) Centre de Recherche de l’Institut du Cerveau et de la Moelle epiniere, UPMC, Paris, France

Graphic design of the cover: Massimo Delle Femmine, Istituto Superiore di Sanità Ann Ist Super Sanità 2020 | Vol. 56, No. 3

Annali dell’Istituto Superiore di Sanità

Vol. 56, No. 3 2020

Contents

Editorials

257 The increasing need for a new Italian legislation to facilitate execution of observational studies assuring ethics and highest standards of scientific and methodological quality Carlo Petrini, Giovanni Fiori, Gualberto Gussoni, Sara Cazzaniga, Giovannni Corrao, Valeria Lovato, Dario Manfellotto, Francesca Mastormauro and Alessandro Mugelli

260 Migration and dementia in Europe: towards a culturally competent approach and provision of care Marco Canevelli and Nicola Vanacore

Commentary

263 Could we fight healthcare-associated infections and antimicrobial resistance with probiotic-based sanitation? Elisabetta Caselli and Ivana Purificato

Original articles and reviews

267 Demographic and socio-economic determinants of poor HIV-risk perception at first HIV diagnosis: analysis of the HIV Surveillance data, Italy 2010-2016 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti, Alessia Mammone, Enrico Girardi, Barbara Suligoi and the HIV Surveillance System Group

277 Microbiological ascertainment in patients with pneumonia: the experience of a teaching hospital in Rome Marianxhela Dajko, Andrea Poscia, Brunella Posteraro, Domenico Speziale, Massimo Volpe, Silvia Mancinelli, Walter Ricciardi and Chiara de Waure

285 Pharmacology and legal status of cannabidiol Pietro Brunetti, Alfredo Fabrizio Lo Faro, Filippo Pirani, Paolo Berretta, Roberta Pacifici, Simona Pichini and Francesco Paolo Busardò Ann Ist Super Sanità 2020 | Vol. 56, No. 3

292 Italian pool of asbestos workers cohorts: asbestos related mortality by industrial sector and cumulative exposure Corrado Magnani, Stefano Silvestri, Alessia Angelini, Alessandra Ranucci, Danila Azzolina, Tiziana Cena, Elisabetta Chellini, Enzo Merler, Venere Pavone, Lucia Miligi, Giuseppe Gorini, Vittoria Bressan, Paolo Girardi, Lisa Bauleo, Elisa Romeo, Ferdinando Luberto, Orietta Sala, Corrado Scarnato, Simona Menegozzo, Enrico Oddone, Sara Tunesi, Patrizia Perticaroli, Aldo Pettinari, Francesco Cuccaro, Stefano Mattioli, Antonio Baldassarre, Francesco Barone-Adesi, Marina Musti, Dario Mirabelli, Roberta Pirastu, Alessandro Marinaccio, Stefania Massari, Daniela Ferrante, for the Working Group Studio Multicentrico Italiano di Coorti di lavoratori dell’Amianto (SMICA)

303 Suicidal behaviour and non-suicidal self-injury in children and adolescents seen at an Italian paediatric emergency department Laura Castaldo, Giulia Serra, Simone Piga, Antonino Reale and Stefano Vicari

315 A qualitative study of family carers views on how end-of-life communication contributes to palliative-oriented care in nursing home Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

325 Tailored screening and dedicated funding for direct acting antiviral drugs: how to keep Italy on the road to hepatitis C virus elimination? Loreta A. Kondili, Sarah Blach, Homie Razavi and Antonio Craxi

330 Lack of protection for measles among Italian nurses. A potential for hospital outbreak Luca Coppeta, Stefano Rizza, Ottavia Balbi, Savino Baldi and Antonio Pietroiusti

336 Cancer stem cell targeted therapies Ugo Testa, Elvira Pelosi and Germana Castelli

351 Male circumcision: ritual, science and responsibility Francesco Ventura, Fiorella Caputo, Marta Licata, Alessandro Bonsignore and Rosagemma Ciliberti

359 Bacterial coinfections in COVID-19: an underestimated adversary Lanfranco Fattorini, Roberta Creti, Carla Palma, Annalisa Pantosti and the Unit of Antibiotic Resistance and Special Pathogens

365 Schiavonia Hospital response to COVID-19 outbreak: a first single-center experience Elena Marcon, Francesca Scotton, Elena Marcante, Alberto Rigo, Jacopo Monticelli, Maria Emanuela Buggio, Claudio Pilerci, Domenico Montemurro and Patrizia Benini

373 COVID-19 mortality among migrants living in Italy Marco Canevelli, Luigi Palmieri, Valeria Raparelli, Ornella Punzo, Chiara Donfrancesco, Cinzia Lo Noce, Nicola Vanacore, Silvio Brusaferro, Graziano Onder and the Italian National Institute of Health COVID-19 Mortality Group

378 Coronavirus and birth in Italy: results of a national population-based cohort study Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore, Serena Donati and the ItOSS COVID-19 Working Group

Brief note 390 Neurophysiological monitoring in neonatal abstinence syndrome from cocaine Raffaele Falsaperla, Simona Zaami, Maria Giovanna Aguglia, Catia Romano, Agnese Suppiej and Luigi Memo Ann Ist Super Sanità 2020 | Vol. 56, No. 3

397 Book Reviews, Notes and Comments Edited by Federica Napolitani Cheyne

399 Publications from International Organizations on Public Health Edited by Annarita Barbaro cine (SIMeF)togetherwith the IstitutoSuperioredi normative reference. the conductionofotherobservational studieswithouta observational studiesonmedicinal products[1],leaving particular, theItalianlegislationcurrentlyregulates only and for the design of innovative research projects. In for aneffective control of the studyprotocols quality tries), buthasbeenproducingalsopotentiallimitations competitive, ifcomparedwithotherEuropeancoun- pensive authorizationprocesses(whichmakeItalyless updated. Thishasbeencausingnotonlylongandex- definitions, andbyanationallegislationthatneedtobe stacles, mostlycausedbyambiguousterminologyand ducting observationalstudiesinItalyfacedseveralob- RWE). Unfortunately, inthelastdecadescientistscon- evidence intherealworldsetting(realevidence, vational research,thatgreatlycontributestogenerate als (RCT, experimental studies) and also from obser the literature,fromrandomizedcontrolledclinicaltri- search question,thatcomesfromsystematicreviewsof makers needthebestavailableevidenceonagivenre- professionals, patients,regulatoryagenciesandpolicy Valeria Lovato 9 8 7 6 5 4 3 2 Carlo Petrini and methodologicalquality and thehigheststandardsofscientific observational studiesassuringethics legislation tofacilitateexecutionof The increasingneedforanewItalian E DOI: 10.4415/ANN_20_03_01 Ann IstSuperSanità2020|Vol. 56,No.3:257-259 1 San GiovanniCalibitaFatebenefratelli,Rome,Italy Milano Bicocca,Milan,Italy E-mail: [email protected]. Address forcorrespondence : CarloPetrini,Unità diBioetica,IstitutoSuperioreSanità, Via GianodellaBella34, 00161Rome,Italy. Società ItalianadiFarmacologia–UniversitàdegliStudiFirenze,Florence,Italy Società ItalianadiMedicinaFarmaceutica–AstraZeneca,Basiglio,Italy FADOI SocietàScientificadiMedicinaInterna–DipartimentodelleDisciplineMediche,Ospedale Società ItalianadiMedicinaFarmaceutica–Roche,Monza,Italy Centro diRicercaInteruniversitarioHealthcareResearch&Pharmacoepidemiology, UniversitàdegliStudidi Società ItalianadiMedicinaFarmaceutica–JanssenCilag,ColognoMonzese,Italy FADOI SocietàScientificadiMedicinaInterna,Milan,Italy Società ItalianadiMedicinaFarmaceutica–MediNeos,Modena,Italy Unità diBioetica,IstitutoSuperioreSanità,Rome,Italy Recently, theItalianSocietyofPharmaceutical Medi- To makesolid andeffectivedecisions,healthcare ditorial 1 , GiovanniFiori 6 , DarioManfellotto 2 , GualbertoGussoni 7 , FrancescaMastromauro - vacy Authority. Agency (AIFA), theHealthMinistryandDataPri- tively discussedwithrepresentativesoftheItalianDrug tific societies*andtheunderlyingconceptswereposi- later formallyapprovedbyseveralotherItalianscien- 14 May2019,n.52[3].Therecommendationswere of law 11 January 2018, n. 33 [2] and legislative Decree legislation onobservationalstudies,intheframework tions forthedefinitionofnewforthcomingnational ed a working group to formulate specific recommenda- and theSocietàItalianadiFarmacologia(SIF)activat- sociazioni dei Dirigenti Ospedalieri Internisti (FADOI) & Pharmacoepidemiology),theFederazionedelleAs- tro diRicercaInteruniversitarioHealthcareResearch Sanità (ISS),theUniversityofMilanoBicocca(Cen- venti Terapeutici (SIFEIT);GruppoItalianoDataManager(GIDM). cietà ItalianaperStudi di Economia ed Etica sul Farmacoe sugli Inter Italiana diStatisticaMedicaedEpidemiologia Clinica(SISMEC);So- one ItalianadiEmatologiaeOncologia Pediatrica(AIEOP);Società tica (SIMeF); Associazione Farmaceutici Industria (AFI); Associazi- Italiana diFarmacologia(SIF);Società ItalianadiMedicinaFarmaceu- dalieri Internisti(FADOI); IstitutoSuperiorediSanità(ISS);Società Milano Bicocca;FederazionedelleAssociazionideiDirigentiOspe- care Research&PharmacoepidemiologyUniversitàdegliStudidi * List atFebruary2020:CentrodiRicercaInteruniversitario 3 , SaraCazzaniga 8 andAlessandroMugelli 4 , GiovannniCorrao

9

Health- 5 , - 257 Editorial 258 Carlo Petrini, Giovanni Fiori, Gualberto Gussoni et al.

The scope of the recommendations was limited to ob- cal reasons for this choice. Further, the subject must servational studies defined as “collection and analysis provide a specific written consent, based on a compre- for scientific purposes of epidemiological, administra- hensive information about the purposes, the nature and tive, clinical and biometric data related to single human the methods of carrying out additional examination/s, subjects”. According to these recommendations, the and potential inconveniences associated with the proce- new provision should be mandatory and should regu- dures. It is also fundamental that the costs of additional late all the types of study conducted with observational procedures should not be borne by the National Health methodology, within biomedical and health field, pro- Service (NHS) nor by the subjects, but fully covered by

ditorial moted by public or private organizations. Therefore, the sponsor/promoter. Moreover, the Ethics Comittee

E the new regulation should cover observational studies should evaluate the need for the sponsor/promoter to with or without drugs, with or without additional diag- stipulate a specific insurance if the additional procedure nostic procedures, with primary or secondary data uses, is evaluated as invasive and risky for the subject. and should also include studies based on databases and Recommendations also consider some additional ac- complex data sources (for example data collected di- tions: fundamental information on all type of observa- rectly from patients via digital tools). tional studies should be entered in a national registry; In this context, each study protocol should receive a mapping of new and already existing healthcare data- a single competent evaluation, with a multi-sites and bases and registries should be activated to allow inves- nation-wide validity: this modus operandi is coherent tigators to have access to high quality data, also thanks with that indicated by the European Regulation n. to nationwide, transparent rules; participation to obser- 536/2014 [4] for clinical trials. Local evaluations at sites vational studies shouldn’t be limited to physicians only level should be limited to verify the presence of all the but also permitted to other healthcare professionals, needed resources (human, material and organizational) once they are adequately trained on ethical, method- for correctly executing the study and to evaluate the ological, regulatory and technical aspects. Last but not promoter’s proposal for administrative agreement. This least, forthcoming Italian regulation should be aligned promoter’s should also be facilitated using standard with already existing EU regulations and guidelines – templates (e.g. privacy information model, administra- especially EU Good Pharmacovigilance Practice (GVP) tive contract) with a nation-wide validity and by a stan- and the General Data Protection Regulation (GDPR) dardized national fee for the evaluation of observational – also by means of preliminary consultations among the studies across the country. stakeholders to harmonize different approaches and Authors suggest also some practical steps to make definitions and to solve some existing issues. this new authorization process feasible. It could be use- According to the authors’ opinion, the above summa- ful to define a list of Ethics Committees certificated by rized recommendations can facilitate the execution of the Ministry of Health as “expert” Ethics Committees observational studies in Italy assuring both ethics and for the evaluation of observational studies. Sponsor/ the highest standards of scientific and methodological Promoter could therefore obtain a single evaluation quality. For this reason, the authors strongly encourage submitting the documentation to one of these certifi- to adopt these guidelines to define the new national leg- cated Ethics Committees which should have appropri- islation on observational studies. ate expertise to evaluate typical complexities related to observational studies, such as the documentation relat- Acknowledgements ed to the privacy as well as the methods for collecting Thanks to SIMeF colleagues Silvia Ferrara (Janssen) and storing biological samples or for access to database and Antonietta Caputo (Roche) for the contribution of biological samples. in the definition and writing of the recommendations. Special attention should be addressed to observa- Special thanks to Michela Masoero (Medineos) for the tional studies where, for methodological reasons, the precious contribution in coordinating the inter-societies study protocol requires additional diagnostic and evalu- working group. ation procedures, e.g. being these procedures known and used in normal clinical practice but not routinely Conflict of interest statement applied for the cases to be included in the study. In There are no potential conflicts of interest or any fi- these specific situations, authors recommend that the nancial or personal relationships with other people or study protocol contains a specific section dedicated to organizations that could inappropriately bias conduct illustrating the scientific rationale and the methodologi- and findings.

REFERENCES

1. Agenzia Italiana del Farmaco. Determinazione 20 marzo sanitarie e per la dirigenza sanitaria del Ministero della 2008. Linee guida per la classificazione e conduzione degli salute. Gazzetta Ufficiale della Repubblica Italiana– Se- studi osservazionali sui farmaci. Gazzetta Ufficiale della rie Generale n. 25, 31 gennaio 2018. Repubblica Italiana – Serie Generale n. 76, 31 marzo 2008. 3. Decreto legislativo 14 maggio 2019, n. 52. Attuazione 2. Parlamento Italiano. Legge 11 gennaio 2018 n. 3. Delega della delega per il riassetto e la riforma della normativa in al Governo in materia di sperimentazione clinica di medi- materia di sperimentazione clinica dei medicinali ad uso cinali nonché disposizioni per il riordino delle professioni umano, ai sensi dell’articolo 1, commi 1 e 2, della legge 11 259 Legislation on observational studies

gennaio 2018, n. 3. Gazzetta Ufficiale della Repubblica and of the Council of 16 April 2014 on clinical trials on Italiana – Serie Generale n. 136, 12 giugno 2019. medicinal products for human use, and repealing Direc- 4. European Parliament, Council of the European Union. tive 2001/20/EC. Official Journal of the European Union Regulation (EU) 536/2014 of the European Parliament 27 May 2014; L158:1-76. ditorial E 260 Editorial discrepancy oftheresultsobtained inpopulation-based among migrantindividualsin Europearisesfromthe pothesis thatdementiamaybe largelyunderdiagnosed lenges andlimiteduseofdedicated services[8].Thehy- likely evenhigherinmigrants duetodiagnosticchal- agnosed worldwide[7].Therateofunderdetectionis creasing awareness,dementiaremainslargelyunderdi- creased riskofdementiaascomparedtonatives. grants; andii)somemigrantgroupsmighthaveanin- dementia islikelylargelyunderdiagnosedamongmi- and migrantsinEurope,providedtwomainresults:i) paring theprevalence/incidenceofdementiainnatives view and meta-analysis of epidemiological studies com- extended to theEuropean context [6]. Asystematic re- dementia [5].Theseobservationshaverecentlybeen minority groupshaveahigherlikelihoodofdeveloping tus influences dementia risk and that some migrant and Several studieshaveinfactshownthatthemigrantsta- cally variegatedpopulationslivingintheUnitedStates. and dementia has beenmostly explored in theethni- societal, andhealthcareperspective[4]. ground mightassumespecialrelevanceunderaclinical, these disturbancesinindividualswithamigrationback- dementia andcognitivedisorders.Theoccurrenceof of chronic,disablingdiseasesandconditionsincluding grants areincreasinglyexposedtotheriskandburden migrants [1]. Similarly to their native counterparts, mi- in 2019,accountedfor12%ofoverallinternational ers isgraduallyextendingtooldermigrants[3]who, attention ofthescientificcommunityand policymak- the structureofmigrants’populations.Accordingly, the sition (i.e.,populationaging)isprogressivelychanging ority [2].Inparticular, theongoingdemographictran- population isgrowinglyregardedasapublichealthpri- health attributesofthisincreasinglylargesharethe timated worldwide[1].Theneedofcharacterizingthe 1 2 Marco Canevelli and provisionofcare towards aculturallycompetentapproach Migration anddementiainEurope: E DOI: 10.4415/ANN_20_03_02 Ann IstSuperSanità2020|Vol. 56,No.3:260-262 Rome, Italy. E-mail:[email protected]. Address forcorrespondence: MarcoCanevelli, Dipartimento NeuroscienzeUmane, SapienzaUniversitàdiRoma,Viale dell’Università 30,00185 di Sanità,Rome,Italy Centro NazionaleperlaPrevenzionedelleMalattieePromozionedellaSalute,IstitutoSuperiore Dipartimento diNeuroscienzeUmane,SapienzaUniversitàRoma,Rome,Italy Despite major advancements in diagnostics and in- To date,the complex relationshipbetweenmigration In 2019,272millioninternationalmigrantswerees- ditorial 1,2 andNicolaVanacore 2 of ourpopulations.Encouragingly, several cross-cultur adapt servicesandprocedures tothegrowingdiversity mographic transformations,it isinsteadimperativeto diagnosis. Inlightoftheongoing andfuturesociode- cognitive examination and contribute to dementia mis- in migrantscanthereforepotentiallyresultabiased tural backgroundoftheindividual[11,12].Theiruse influenced by the language, educational level, and cul- have beendevelopedinWestern contextsandcanbe tinely adoptedcognitivescreeningandassessmenttools sensitive instrumentsandprocedures.Mostofrou- bances inmigrantsrequirestheadoptionofculture confirm thattheclinicalapproachtocognitivedistur ish-born counterparts[10]. 0.93, 95%CI0.91-0.95)incomparisontotheirSwed- (HR 0.85,95%CI0.83-0.88)andfemalemigrants lower incidenceofdementiaamongbothmalemigrants adults aged45yearsandolderinSwedendocumented individuals (2.7%)[9].Acohortstudyincludingall Western migrants(1.5%)ascomparedtoDanish-born lence wassignificantlylowerinnon-Western (1.4%)and tia in Denmark in2012, the observeddementia preva- overall populationofolderpeoplelivingwithdemen- in Europe. In a register-based study conducted in the parent lowerdementiafrequencyamongmigrantsliving have providedadditionalevidence supporting suchap- not includedintheabove-mentionedmeta-analysis dementia treatment.Two recentregistry-basedstudies viduals whoalsohaveareducedlikelihoodofreceiving mentia emergesaslessprevalentinforeign-bornindi- studies basedonhealthcarerecordsandregistries,de- minority groupsrelativetonatives.Onthecontrary, in higher prevalenceratesareobservedinmigrantsand to theethnicityormigrantstatusofparticipants), cedures (e.g.,makingadiagnosisofdementiamasked by adoptingculture-sensitivecognitivetoolsand/orpro- ence ofdementiaisascertainedatthecommunitylevel and registry-basedstudies[6].Infact,whenthepres- On theonehand,theseapparentlyconflictingresults - - 261 Migration and dementia in Europe

al cognitive tools have been developed and validated in tributes and factors that have already been individually multicultural European samples and are starting to be associated with dementia risk. They include educational adopted in “real world” memory clinics [13, 14]. More- level, socioeconomic status, health-related behaviors, over, the barriers and inadequacies that may challenge cardiovascular diseases, stress exposure [19]. Moreover, the provision of care to migrants with dementia are be- there are several related variables and circumstances, ing explored with ad hoc surveys and qualitative studies such as country of origin, ethnicity, acculturation, social targeting both professionals, migrants with cognitive support, reason for migration, that are extremely het- disturbances and their families [15, 16]. These research erogeneous and impede to approach the phenomenon

efforts represent fundamental preliminary steps to ad- under a unitary perspective. In this regard, a greater ef- ditorial just existing dementia facilities to the evolving attri- fort is needed to collect data on these determinants and E butes and needs of all older Europeans. to fill the existing evidence gap concerning those mi- Conversely, the lower dementia prevalence estimates grant groups such as refugees, asylum-seekers, and un- among migrants found in registry-based studies are documented migrants, that are more marginalized and likely affected by the fact that these individuals have a constantly underrepresented in research. Dementia risk reduced access to healthcare resources (e.g., services, should be assessed at the individual level, by adopting treatments) and are therefore less frequently captured a life-course perspective that implies the interaction of by healthcare records and databases. Migrants and mi- multiple factors with protective or triggering role. The use nority groups have been shown to underuse and present of reductionistic and categorical frameworks may instead later to dementia services [17] and diverse barriers to result in unbalanced risk assessments and even contrib- medical help seeking for dementia have been identi- ute to prejudices and stigmatizing behaviors among and fied [8]. Consequently, the underdiagnosis of demen- towards migrants. Migration should be more properly tia, commonly documented in nationwide registries at intended as a “modifier of dementia risk” [20]. As such, the general population level [18], may be even higher it constitutes a further layer of complexity that should be among migrants. adequately captured to deliver person-centered, cultur- Based on the discussed body of evidence, migration ally competent prevention, support, and care. emerges as an important determinant of individual trajectories of cognitive functioning, similarly to what Acknowledgements already observed for many health outcomes [2]. How- Marco Canevelli is supported by a research grant of ever, understanding migration as a dementia risk factor the Italian Ministry of Health for the project “Demen- is currently not legitimate and potentially misleading. In tia in immigrants and ethnic minorities living in Italy: fact, the migratory background may influence the one’s clinical-epidemiological aspects and public health per- susceptibility to cognitive decline by affecting diverse at- spectives” (ImmiDem)(GR-2016-02364975).

REFERENCES

1. United Nations – Department of Economic and Social G, et al. Prevalence and determinants of undetected de- Affairs, Population Division. International Migration mentia in the community: a systematic literature review 2019: Report (ST/ESA/SER.A/438); 2019. Available and a meta-analysis. BMJ Open. 2017;7(2):e011146. doi: from: www.un.org/en/development/desa/population/mi- 10.1136/bmjopen-2016-011146 gration/publications/migrationreport/docs/International- 8. Mukadam N, Cooper C, Livingston G. A systematic re- Migration2019_Report.pdf view of ethnicity and pathways to care in dementia. Int 2. Abubakar I, Aldridge RW, Devakumar D, Orcutt M, J Geriatr Psychiatry. 2011;26(1):12-20. doi: 10.1002/ Burns R, Barreto ML, et al. The UCL-Lancet Com- gps.2484. mission on Migration and Health: the health of a world 9. Stevnsborg L, Jensen-Dahm C, Nielsen TR, Gasse C, on the move. Lancet. 2018;392(10164):2606-54. doi: Waldemar G. Inequalities in access to treatment and- 10.1016/S0140-6736(18)32114-7 care for patients with dementia and immigrant back- 3. WHO Regional Office for Europe. Health of older refu- ground: A Danish nationwide study. J Alzheimers Dis. gees and migrants. 2018. Available from: www.euro.who. 2016;54(2):505-14. doi: 10.3233/JAD-160124 int/en/publications/abstracts/health-of-older-refugees- 10. Wändell P, Carlsson AC, Li X, Gasevic D, Sundquist J, and-migrants-2018 Sundquist K. Dementia in immigrant groups: A cohort 4. Canevelli M, Lacorte E, Cova I, Zaccaria V, Valletta M, study of all adults 45 years of age and older in Sweden. Raganato R, et al. Estimating dementia cases amongst mi- Arch Gerontol Geriat. 2019;82:251-8. doi: 10.1016/j. grants living in Europe. Eur J Neurol. 2019;26(9):1191-9. archger.2019.03.003 doi: 10.1111/ene.13964 11. Milani SA, Marsiske M, Cottler LB, Chen X, Striley 5. Moon H, Badana ANS, Hwang SY, Sears JS, Haley WE. CW. Optimal cutoffs for the Montreal Cognitive Assess- Dementia prevalence in older adults: Variation by race/ ment vary by race and ethnicity. Alzheimers Dement. ethnicity and immigrant status. Am J Geriatr Psychiatry. 2018;10:773-81. doi: 10.1016/j.dadm.2018.09.003 2019;27(3):241-50. doi: 10.1016/j.jagp.2018.11.003 12. Ranson JM, Langa KM, Llewellyn DJ. Predictors of de- 6. Selten J-P, Termorshuizen F, van Sonsbeek M, Bogers J, mentia misclassification when using brief cognitive as- Schmand B. Migration and dementia: a meta-analysis of sessments. Neurol Clin Pract. 2019;9(2):109-17. doi: epidemiological studies in Europe. Psychol Med. 2020 10.1212/CPJ.0000000000000566 (in press). doi: 10.1017/S0033291720000586 13. Nielsen TR, Andersen BB, Gottrup H, Lützhøft JH, 7. Lang L, Clifford A, Wei L, Zhang D, Leung D, Augustine Høgh P, Waldemar G. Validation of the Rowland Univer- 262 Marco Canevelli and Nicola Vanacore

sal Dementia Assessment Scale for multicultural screen- 17. Cooper C, Tandy AR, Balamurali TBS, Livingston G. A ing in Danish memory clinics. Dement Geriatr Cogn Dis- systematic review and meta-analysis of ethnic differences ord. 2013;36(5-6):354-62. doi: 10.1159/000354375 in use of dementia treatment, care, and research. Am J 14. Nielsen TR, Segers K, Vanderaspoilden V, Beinhoff U, Geriatr Psychiatry. 2010;18(3):193-203. doi: 10.1097/ Minthon L, Pissiota A, et al. Validation of a European JGP.0b013e3181bf9caf Cross-Cultural Neuropsychological Test Battery (CNTB) 18. Rizzuto D, Feldman AL, Karlsson IK, Dahl Aslan AK, for evaluation of dementia. Int J Geriatr Psychiatry. Gatz M, Pedersen NL. Detection of dementia cases in 2019;34(1):144-52. doi: 10.1002/gps.5002 two Swedish Health Registers: A validation study. J Al- 15. Nielsen TR, Vogel A, Riepe MW, de Mendonça A, Ro- zheimers Dis. 2018;61(4):1301-10. doi: 10.3233/JAD-

ditorial driguez G, Nobili F, et al. Assessment of dementia in 170572

E ethnic minority patients in Europe: a European Al- 19. Yaffe K, Falvey C, Harris TB, Newman A, Satterfield S, zheimer’s Disease Consortium survey. Int Psychogeriatr. Koster A, et al. Effect of socioeconomic disparities on 2011;23(1):86-95. doi: 10.1017/S1041610210000955 incidence of dementia among biracial older adults: pro- 16. Canevelli M, Lacorte E, Cova I, Cascini S, Bargagli AM, spective study. BMJ. 2013;347:f7051. doi: 10.1136/bmj. Angelici L, et al. Dementia among migrants and ethnic f7051 minorities in Italy: rationale and study protocol of the 20. Albert SM. Immigration: A modifier of dementia risk in ImmiDem project. BMJ Open. 2020;10(1):e032765. doi: old age? Am J Geriatr Psychiatry. 2019;27(3):251-3. doi: 10.1136/bmjopen-2019-032765 10.1016/j.jagp.2018.12.014 di Sanità,Rome,Italy 1 HAI directlycauseover1.1 billions€ofadditional fecting upto4millionpatients eachyearintheEU, tality, fact,af- aswellHAI-associatedcosts[3]. In AMR isassociatedwithincreased morbidityandmor pressure exertedbythewide useofantimicrobials, cancer [2]. has beenhypothesizedthatAMRmightkillmorethan rope. Thisfigurecouldrisetenfoldby2050,whenit during hospitalization,arereportedeachyearinEu- drug-resistant bacterialinfectionsalone,acquired antibiotic” era[1],andmorethan33000deathsfrom this, thecurrentperiodhasbeendefinedas“post- growth ofAMRtoalarminglevels.Consistentlywith (particularly inthehealthcaresettings)hasledto antibiotics inhumans,animalsandtheenvironment era”, themassiveandsometimesinappropriateuseof cillin in1928,whichopenedtheso-called“antibiotic HAI causalpathogens.Sincethediscoveryofpeni- with thegrowingantimicrobialresistance(AMR)of a globalconcernforhumanhealth,tightlyassociated Borsari 46,44121Ferrara, Italy. E-mail:[email protected]. Address forcorrespondence: ElisabettaCaselli, Dipartimento diScienzeChimichee Farmaceutiche, UniversitàdegliStudi diFerrara,Via Luigi 3 2 Elisabetta Caselli with probiotic-basedsanitation? infections andantimicrobialresistance Could wefighthealthcare-associated C DOI: 10.4415/ANN_20_03_03 Ann IstSuperSanità2020|Vol. 56,No.3:263-266 CIAS, UniversitàdegliStudidiFerrara,Italy serves exploration. the reportedresultsandarguethattheirusemayprovideanovelapproachwhichde- stably reduceAMRsurfacepathogens,finallyreducingHAIincidence.Herewediscuss the fightagainstAMRmayinvolveprobiotic-basedsanitationapproaches,astheymight tion. Arecentlypublishedstudy, theSAN-ICAstudy, performedinItaly, suggeststhat pathogens, apparentlynotsufficientlycontrolledbyconventionalchemical-basedsanita- associated pathogens.Thehospitalenvironmentitselfisinfactareservoirofresistant environment, a majorcauseistherampantantimicrobial resistance (AMR)oftheHAI- billion €associatedcosts.Besidesthepersistentmicrobialcontaminationofhospital patients intheEU,causingover33000deathsasadirectconsequenceand1.1 Healthcare-associated infections(HAI)affecteveryyearabout4millionhospitalized Abstract Dipartimento diSicurezzaAlimentare,AlimentazioneeSanitàPubblicaVeterinaria, Centro RicercheInquinamentoFisicoChimicoMicrobiologicodiAmbientiConfinatiadAltaSterilità– Dipartimento diScienzeChimicheeFarmaceutiche,UniversitàdegliStudiFerrara,Italy In thehospitalenvironment,duetoselective Healthcare-associated infections(HAI)represent omment 1,2 andIvanaPurificato ary 3 - infection preventionmeasures (AMRstewardship)[6]. including increasingAMRawareness, surveillance,and have beentakentolimitAMR andHAIoccurrence, ment, andfoodproduction. to veterinarymedicine,agriculturallivestockmanage- worldwide, notonlyrelatedtohumanhealthbutalso critical focusforsustainablehealthcareintheEUand fections extremely challenging [5]. Targeting AMR is a the therapeuticalapproachagainstassociatedin- pathogenicity butratherfortheirAMR,whichrenders the humanhealthnotonlyandasmuchfortheir drug-resistant bacteriawhichrepresentthreatsfor in theso-called“dirtydozen”,aglobalprioritylistof difficult-to treat HAI[5]. terococcus faecalis/faecium,allcausingthemajorityof bacter baumannii,Pseudomonasaeruginosa,andEn- Staphylococcus aureus,Klebsiellapneumoniae,Acineto- gens astheESKAPEgroup,includingEscherichiacoli, WHO hasdefinedthetop-threateningHAIpatho- sanitary costs[4].Consistentlywithsuchconcern,the Since thehospitalenvironment representsahugeres- Consistently withthis,inthe recentyearsactions All of them are also included

Istituto Superiore • • • • Key words probiotics hospital sanitation antimicrobial resistance infections healthcare-associated 263 Commentary 264 Elisabetta Caselli and Ivana Purificato

ervoir of resistant pathogens, a focus on the environ- vironment, by using a “microbial-based” sanitation in mental hygiene was included as one of the key points substitution of the conventional chemical-based one. in the fight against AMR in the 2017 EU anti-AMR In particular, the sanitation method (named Probiotic plan [6]. In Italy, the importance of defining and main- Cleaning Hygiene System, PCHS) was based on the taining adequate levels of environmental microbial application of eco-sustainable cleansers additioned contamination in the hospital environment has been with probiotics belonging to the Bacillus genus, namely recently highlighted in the 2018 guideline document by B. subtilis, B. pumilus and B. megaterium species. The ANMDO, the Italian association of Medical Hospital system was applied in substitution of conventional Directors, to control the risk of contracting an HAI dur- disinfection in the Internal Medicine wards of six ing hospitalization [7]. Italian public hospitals, where surface bioburden and ommentary Controlling the hospital environment microbial con- HAI incidence was simultaneously analysed, compar- C tamination appears thus a key point, so far addressed ing the 6-month initial period during which chemical by conventional sanitation procedures based on the use sanitation was used with the 6-month period receiv- of chemical-based detergents and disinfectants. Never- ing PCHS. About 12 000 hospital inpatients were theless, several reports show the presence of persistent observed and over 30 000 environmental samples ana- contamination in > 50% of sanitized hospital surfaces lysed. Overall, the authors report a stable 83% reduc- [8], mostly due to recontamination phenomena, which tion of surface ESKAPE pathogens during the PCHS are continuous and cannot be prevented by conven- period compared to what detected during pre-PCHS tional sanitizers. phase. Of note, the sampling was performed seven In addition, some disinfectants have been reported hours after sanitation, to let recontamination occur. A to potentially contribute to the selection of resistant likely mechanism of replacement by competitive exclu- strains. For example, in Gram-negative species adapted sion was hypothesized, as the reduction of the patho- to benzalkonium chloride a new resistance was most genic component of the surface microbiota was accom- frequently found to ampicillin, cefotaxime, and sulfa- panied by a concomitant increase of probiotics (finally methoxazole [9]. With the use of chlorhexidine a new representing about 70% of total surface microbiota). resistance was often found to ceftazidime, sulfamethox- Interestingly, the study reports an up to 99% decrease azole, imipenem, cefotaxime and tetracycline, as well of the drug-resistance genes harboured by the residual as against colistin [9, 10]. Since colistin is considered contaminating population, compared to the chemical a “last-resort” drug for the treatment of multi drug-re- sanitation period, as detected by the resistome analy- sistant (MDR) Gram-negative bacteria, this effect ap- sis of the entire microbial population. In particular, the pears particularly undesirable, also in light of recently resistances against aminoglycosides, fluoroquinolones, reported data on the prevalence of the plasmid-driven macrolides, methicillin, vancomycin, β-lactamases, and colistin resistance in the hospital environment in Italy colistin were reported to be significantly decreased. [11]. Cross-resistance to antibiotics was also found with The decrease of AMR is further detailed in a compan- triclosan, octenidine, sodium hypochlorite, and didecy- ion paper by the same authors [16], where a significant ldimethylammonium chloride. reduction of antimicrobial consumption (-60%) and re- Alternatives to disinfectants include ultraviolet (UV) lated costs (-75%) is reported. light, hydrogen peroxide protocols, no-touch technolo- Most importantly in a patients-safety perspective, the gies, self-disinfecting surfaces, and use of metals like use of PCHS was reported to be associated with a 52% iron, copper or silver [12]. However, the effectiveness reduction of HAI incidence compared with the use of of such technologies is highly dependent on several pa- conventional sanitation [15], and no infection caused rameters including the concentration, time of exposure, by Bacillus in any of the hospitalized inpatients, con- and the amount of original contamination, and lastly firming previous data on the safety of use ofBacillus they have high costs of implementation and limited use probiotics [17, 18]. to specific surface types. The SAN-ICA survey has diverse potential limita- On the other hand, the use of probiotics as a po- tions, as indeed recognized by the authors themselves. tential tool in reestablishing a healthy balance in po- The first potential bias is related to the study design, a tentially pathogenic microbiota has become more and pre-post intervention performed in the same hospitals, more popular for human, veterinary and environmental and further studies should include stepped wedge trials application. The recent findings on the human, animal or cluster randomized trials. Another potential bias con- and environmental microbiota has profoundly stimulat- cerns the difficulty of performing double-blind surveys ed the research in this direction, suggesting that drastic and controlling all the other factors potentially affect- antimicrobial approaches may be doomed to failure, ing HAI onset. Furthermore, collected data refer only sometimes even further favouring the colonization of to Internal Medicine wards, and more studies should potentially pathogenic microbes [13]. At the same man- be performed to verify the generalizability of the results ner, at the environmental level, it is hypothesized that a in other type of wards and patients. Last, a potential “super-sanitation” might not represent the solution for for confounding is represented by the lack of measure- pathogens elimination, whereas rather a replacement ment of hand hygiene over the study period, although of pathogens with beneficial microorganisms might be the authors report that there was an agreement not to more effective [14]. introduce measures to improve infection control in the In a recently published study, the SAN-ICA study enrolled hospitals. [15], this principle was investigated in the hospital en- Nevertheless, even taking into account these poten- 265 Antimicrobial resistance and probiotic-based sanitation

tial confounders, the effect in term of displacement of adopted, it would be advisable to monitor its effects in microbiota, and the size and characteristics of the HAI the treated environment, not only to verify its decon- reduction seem significant and may deserve attention. taminating effectiveness but also, perhaps more impor- This also in a cost-saving and sustainability perspective, tantly, to prevent potential avoidable side-effects such due to the low environmental impact and cost of probi- as further AMR induction and spread. The use of the otic-sanitizers. new molecular methodologies, such as next generation Overall, Bacillus spores have a long history of safe sequencing (NGS) or other molecular analyses, could uses in humans, including food preparation, agricul- be of great help in characterizing the microbiota and its ture [19], animal farms [20], human therapy of the evolution precisely and timely. gut [21], and their administration was recently shown However, although with the highlighted limitations, to be associated with S. aureus eradication in human the data reported in these studies are intriguing and ommentary C gut [22]. suggest that microbial-based approaches may deserve However, there is still a lack of clear evidence on how further exploration. A fruitful debate may be opened on exactly probiotics produce their benefits. It has been the possibility to consider the approaches modulating suggested that they can act by competitive exclusion the environmental microbiota as something potentially (competition for nutrition and space), and/or secretion able to balance it in such a way to contribute positively of antimicrobial compounds, but a combined effort at to the control of AMR and HAI. global level should be needed for implementing probi- otic screening and regulation relative to their final use. Conflict of interest statement One important point, especially in the light of the re- The Authors have no potential conflict of interests cent COVID 19 pandemics, would be to ascertain the and received no financial support. potential antiviral activity of such ecological sanitation, an aspect still not elucidated. Received on 2 October 2019. It should be emphasised that, whatever the sanitation Accepted on 29 May 2020.

REFERENCES

1. Bragg RR, Meyburgh CM, Lee JY, Coetzee M. Potential S. Spread of mcr-1-Driven colistin resistance on hospital treatment options in a post-antibiotic era. Adv Exp Med surfaces, Italy. Emerg Infect Dis. 2018;24(9):1752-3. Biol. 2018;1052:51-61. 12. Boyce JM. Modern technologies for improving cleaning 2. O’Neill J. Antimicrobial resistance. Tackling a crisis for and disinfection of environmental surfaces in hospitals. the health and wealth of nations. London: Review on an- Antimicrob Resist Infect Control. 2016;5:10. timicrobial resistance; 2014. 13. Vangay P, Ward T, Gerber JS, Knights D. Antibiotics, 3. Sydnor ER, Perl TM. Hospital epidemiology and infec- pediatric dysbiosis, and disease. Cell Host Microbe. tion control in acute-care settings. Clin Microbiol Rev. 2015;17(5):553-64. 2011;24(1):141-73. 14. Al-Ghalith GA, Knights D. Bygiene: The new paradigm of 4. European Center for Disease Prevention and Control. bidirectional hygiene. Yale J Biol Med. 2015;88(4):359-65. Point prevalence survey of healthcare-associated infec- 15. Caselli E, Brusaferro S, Coccagna M, Arnoldo L, Ber- tions and antimicrobial use in European acute care hos- loco F, Antonioli P, Tarricone R, Pelissero G, Nola S, pitals. Solna: ECDC; 2013. La Fauci V, et al. Reducing healthcare-associated infec- 5. World Health Organization. Global priority list of anti- tions incidence by a probiotic-based sanitation system. A biotic-resistant bacteria to guide research, discovery, and multicentre, prospective, intervention study. PLoS One. development of new antibiotics. WHO; 2017. 2018;13(7):e0199616. 6. European Center for Disease Prevention and Control. A 16. Caselli E, Arnoldo L, Rognoni C, D’Accolti M, Soffritti I, European one health action plan against antimicrobial re- Lanzoni L, Bisi M, Volta A, Tarricone R, Brusaferro S, et sistance (AMR). European Commission; 2017. al. Impact of a probiotic-based hospital sanitation on an- 7. Associazione Nazionale dei Medici delle Direzioni Os- timicrobial resistance and HAI-associated antimicrobial pedaliere. Linea guida sulla valutazione del processo consumption and costs: a multicenter study. Infect Drug di sanificazione ambientale nelle strutture ospedaliere Resist. 2019;12:501-10. e territoriali per il controllo delle infezioni correlate 17. Caselli E, Antonioli P, Mazzacane S. Safety of probiotics all’assistenza (ICA). Bologna: ANMDO; 2018. used for hospital environmental sanitation. J Hosp Infect. 8. Carling PC, Parry MF, Bruno-Murtha LA, Dick B. Im- 2016;94(2):193-4. proving environmental hygiene in 27 intensive care units 18. Caselli E, D’Accolti M, Vandini A, Lanzoni L, Camerada to decrease multidrug-resistant bacterial transmission. MT, Coccagna M, Branchini A, Antonioli P, Balboni PG, Crit Care Med. 2010;38(4):1054-9. Di Luca D, et al. Impact of a probiotic-based cleaning 9. Kampf G. Biocidal agents used for disinfection can en- intervention on the microbiota ecosystem of the hospi- hance antibiotic resistance in Gram-negative species. An- tal surfaces. Focus on the resistome remodulation. PLoS tibiotics (Basel). 2018;7(4). One. 2016;11(2):e0148857. 10. Wand ME, Bock LJ, Bonney LC, Sutton JM. Mecha- 19. Leyva Salas M, Mounier J, Valence F, Coton M, Thierry nisms of increased resistance to chlorhexidine and cross- A, Coton E. Antifungal microbial agents for food bio- resistance to colistin following exposure of Klebsiella preservation. A Review. Microorganisms. 2017;5(3). pneumoniae clinical isolates to chlorhexidine. Antimicrob 20. Mingmongkolchai S, Panbangred W. Bacillus probiotics: Agents Chemother. 2017;61(1). an alternative to antibiotics for livestock production. J 11. Caselli E, D’Accolti M, Soffritti I, Piffanelli M, Mazzacane Appl Microbiol. 2018;124(6):1334-6. 266 Elisabetta Caselli and Ivana Purificato

21. Lopetuso LR, Scaldaferri F, Franceschi F, Gasbarrini A. 22. Piewngam P, Zheng Y, Nguyen TH, Dickey SW, Joo HS, Bacillus clausii and gut homeostasis: state of the art and Villaruz AE, Glose KA, Fisher EL, Hunt RL, Li B, et al. future perspectives. Expert Rev Gastroenterol Hepatol. Pathogen elimination by probiotic Bacillus via signalling 2016;10(8):943-8b. interference. Nature. 2018;562(7728):532-7. ommentary C Barbara Suligoi Italy. E-mail: [email protected]. Address forcorrespondence : MariaDorrucci,Dipartimento diMalattieInfettive,Istituto SuperiorediSanità,Viale Regina Elena299,00161Rome, percentage ofthosediagnosed withlowCD4orat substantially since2010[7]. Europe [6].Further, theseproportionsdidnotchange nosed atalatestageofinfection [5],aswellinrestof Italy, roughly40-60%ofHIV-positive peoplearediag- cause theywereunawareofHIV-positive status[4].In living withHIVwasstillundiagnosed,mostlikelybe- It hasalsobeenestimatedthatnearly13%ofpeople ing about4000newHIVdiagnoseseveryyear[2,3]. was establishedinJuly2008[1],andithasbeenreport- years. TheItalianHIV-Surveillance System(IHIVS) has notundergoneasubstantialdeclineoverrecent INTRODUCTION tional levels. results couldberelevanttoaddresstargetedHIVtestingpoliciesatbothlocalandna- deprivation, loweducation,olderage,malegender, heterosexualandIDUgroups.Our risk perception.PoorerHIV-risk perceptionwasassociatedwithboth,leastandhigh Conclusions. InItalythereisahighpercentageofHIV-infected peoplewithpoorHIV- > 40years(AOR=1.59,CI:1.50-1.69),males1.30,1.20-1.40). ers (IDU)(AOR=1.82,CI:1.59-2.08),loweducation=1.74.1.20-2.54),age 2.33, CI:1.39-3.90];forheterosexuals(AOR=1.96,1.83-2.11),InjectingDrugUs- and highlydeprived[AdjustedOddsRatio(AOR)=1.58,CI:1.14-2.18AOR with high perception. A low risk perception was estimated in both areas, least deprived Results. Thestudyincluded18055individuals:27%withlow, 40%moderateand33% three groupsofresponsecorrespondingtoorderedlevelsHIV-risk perception. on thebasisofreasonsforundergoingtesting.Ordinallogisticmodelswereappliedwith Methods. AnobservationalstudywasconductedandHIV-risk perceptionwasestimated HIV-risk perception inHIVsurveillancedata. cent years.Forthisreason,weanalysedrisk-factorsandsocio-economicindicatorsof Introduction. HIVinfectionsinItalyhasnotundergoneasubstantialdeclineoverre- Abstract * 2 Maria Dorrucci Italy 2010-2016 analysis oftheHIVSurveillancedata, perception atfirstHIVdiagnosis: determinants ofpoorHIV-risk Demographic andsocio-economic DOI: 10.4415/ANN_20_03_04 Ann IstSuperSanità2020|Vol. 56,No.3:267-276 1 The membersoftheHIVSurveillanceSystemGrouparelistedbeforeReferences Istituto NazionaleMalattieInfettive“L.Spallanzani”,Rome,Italy Dipartimento diMalattieInfettive,IstitutoSuperioreSanità,Rome,Italy The stabletrendofnewHIV diagnoses andthehigh The trendofnewdiagnosesHIVinfectionsinItaly 1 1 , Vincenza Regine and theHIVSurveillanceSystemGroup 1 , PatrizioPezzotti reported contrastingresults.A studyfromUKshowed testing andHIV-perception. However, thesestudies ducted [9-13]onpossibleassociation between HIV- sion to opt for testing. Several studies have been con- perception mainlyreliesonits possibleeffectondeci- delay inHIV-testing. these factorscoulddetermine, consequently, astrong with thestigmarelatedtoapositivediagnosis[8].All belonging to arisk group, as well as thefear associated problem thankstoantiviraltreatment,thebeliefofnot the notion,forinstance,thatHIVisnolongerahealth personal perceptionofalowspreadHIVinfection: This termcaninvolveanumberoffactors,includingthe AIDS stagemaybeduetopoor“HIV-risk perception”. The importance of estimating individual HIV-risk 1 , AlessiaMammone * 2 , EnricoGirardi • • • Key words risk-perception HIV epidemiology 2 , 267 Original articles and reviews 268 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti et al.

that only a minority of those who perceived themselves with men, MSM); nationality (Italian vs non-Italian); to be at high risk of HIV was tested in the previous year geographical area of residence (regions grouped on the [14]. In other studies, conversely, there is evidence that basis of geographic area and classified by the Italian Sta- greater HIV-risk perception is related to subsequently tistics Institute (ISTAT) as: Northern (i.e., Piemonte, HIV testing [15-18]. Thus, it is crucial to confirm and Valle d’Aosta, , Lombardia, Trentino-Alto Adige, explore more about possible risk factors and/or deter- , Friuli-Venezia Giulia, Emilia- Romagna), Cen- minants associated with HIV-risk perception to address tral (i.e., Toscana, , , ), Southern targeted HIV testing policies at both local and national (i.e., , , , Puglia, , reviews

levels. , Sicilia, Sardegna); year of the first HIV diag- To our knowledge only one study on risk factors as- nosis (grouped as 2010-2015 and 2016).

and sociated with HIV-risk perception [14] was published The study also analyzed socio-economic indicators

considering the general population, while no studies estimated each year by ISTAT [19] and used by Euro- were conducted on HIV-risk perception using HIV-sur- stat [20, 21]. The socio-economic indicators are derived veillance data. For these reasons, we performed a study from indicators database collected by ISTAT at Nation- in Italy with the aim of analyzing risk-factors and socio- al level which includes 316 indicators available at re- articles economic indicators of HIV-risk perception using the gional level [19]. The values of the indicators are based Italian HIV-Surveillance System (IHVS). on data from the 2011 Italian population census; we considered in our analysis two regional indicators: edu- METHODS cation and deprivation [19]. For the present study we riginal Study design and population considered the estimates in 2013, i.e. the central year of O Observational study by using Italian HIV Surveil- the study period (2010-2016). lance System (IHIVS) [1]. In particular, the Education indicator reports the per- In particular, we considered all new HIV diagnoses centage of adults per each Italian region, who have at- aged 18-85 reported to the Italian HIV-Surveillance Sys- tained at most a lower secondary education level (i.e., at tem (IHIVS) between 2010 and 2016; this surveillance least 8 years of studies) according to the International system has a national coverage and is continuously fed Standard Classification of Education (ISCED level 2) by approximately 180 infectious disease clinics located [22]; the deprivation indicator (named “severe material in all the 20 Italian regions. All diagnoses included in deprivation rate”) is an indicator that “directly considers the present analysis were notified within the end of the lack of some goods and services which are consid- 2016. We did not consider the years before 2010 giv- ered essential for a decent life”; more specifically, it was en that the national surveillance system reached 100% estimated on the basis of percentage of families who territorial coverage only in 2010 [3]. The surveillance say at least three of nine deprivations such as, unable system collects the following information: demographic to sustain unforeseen expenses, a proper meal (i.e., pro- characteristics, clinical data (CD4 cell count, viral load, tein) at least once every two days, adequate heating the clinical stages) and reason for HIV testing. Diagnoses home [20]. that did not report the reason for HIV testing or lacking In our study we assigned the regional estimates (cor- demographical were excluded (nearly 30%). responding to the 20 Italian regions) both Education indicator and the Deprivation to the individual prov- Definition of HIV-risk perception inces of residence: the provinces are smaller Italian ar- Reason for HIV testing was used as a proxy of HIV- eas within regions (107 provinces overall with a mean risk perception: the reason for undergoing HIV-testing of 5-6 provinces in each region). Thus, we divided the was an open question, and the answer was collected by study population in tertiles in order to obtain groups of the clinician at first HIV diagnosis. Only one reason was similar sizes. to be provided. Thus, reason for HIV-testing was classi- fied in three groups, from the lowest to the highest risk Statistical analysis perception as follows: As first analysis we performed a graph of the box-plot 1) HIV-related symptoms; 2) check-up for diseases according to HIV-risk perception level, and then we other than HIV; 3) having engaged in behaviour at risk applied a ordinal logistic model with the three groups for HIV (i.e., unprotected sex). of response corresponding to the three ordered levels In order to verify if the given order of the three groups of HIV-risk perception, from the lowest to the highest could be reasonable from the lowest to the highest, we risk perception as already mentioned in the definition studied the association between CD4 at HIV diagnosis of HIV-risk perception paragraph. This just mentioned and the above mentioned groups (details in statistical analysis was performed in order to verify the previously analysis). discussed classification orderings. Thus, we assumed that, on average, the lower was the CD4 count at new Demographic and socio-economic covariates diagnosis the lower the HIV-risk perception. Individual characteristics from newly diagnosed HIV- Then we applied ordinal logistic models [22], with the positive individuals deemed as possible covariates as- three groups of response corresponding to the three or- sociated to different levels of risk-perception were: age dered levels of HIV-risk perception, from the lowest to at first diagnosis (aged > 40vs ≤ 40); gender (males vs the highest risk perception (see definition of HIV-risk females); HIV- risk category (heterosexuals, injecting perception). drug users, IDU), not indicated vs men who have sex Firstly, we performed the univariate analyses using 269 Determinants of poor HIV-risk perception

ordinal logistic models, thus estimating the crude Odds Table 1 Ratio (OR) for each of independent variables as well General characteristics and socio-economic indicators at diag- as for each indicator (education and deprivation). Sec- nosis (n = 18 055 new HIV diagnoses, Italy 2010-2016) ondly, we applied multilevel ordinal logistic regression Age models [23] using individual characteristics at the first median 39 yrs level (age, gender, HIV-risk groups, nationality, area of residence, year of diagnosis), and the level of depriva- IQR; range (31-47); (18-86) yrs tion and education at the second level (107 provinces age < 40 reviews within regions), as we already mentioned in describing age ≥ 40 these covariates. The between-provinces of residence Gender variance (i.e., random intercept) was estimated for all and multilevel models. It was regarded as significant when males 13 986 (77%) the p-value was < 0.001, indicating that there was a very females 4069 (23%) significant amount of variability in the odds of HIV-risk HIV-exposure category perception between provinces within geographical re-

heterosexuals 8379 (46%) articles gions of residence [24]. All analyses were applied using SAS statistical soft- MSM 7221 (40%) ware version 9.4. IDU 838 (5%)

not indicated 1617 (9%) riginal RESULTS Characteristics of the study population Nationality O From 2010 through 2016, 26 434 new HIV diagnoses 13 370 (74%) among adults were reported to the IHIVS. Of these, non-Italians 4685 (26%) we excluded 6557 (25% of the total diagnoses) diag- Geographical area noses with reason for HIV testing not indicated, and 1822 (7%) individuals without demographic data. The 11 364 (63%) studied population, therefore, consisted on 18 055 new 2604 (14%) HIV diagnoses (i.e., nearly the 70% of the total diagno- 4087 (23%) ses from 2010 to 2016). The characteristics of the study population (n = 18 Year of diagnosis 055) are summarized in Table 1, the median age of the 2010 2518 (14%) study participants was 39 yrs with 56% aged less than 40 2011 2640 (14%) yrs. Further, more than 70% were males and of Italian 2012 2931 (16%) nationality. Heterosexuals were 46%, followed by MSM (40%). The majority of diagnoses (more than 60%) were 2013 2664 (15%) performed among those resident in Northern Italy; al- 2014 2473 (14%) most a constant number of new HIV diagnoses were 2015 2491 (14%) observed each year in the study period (i.e., about 14% 2016 2338 (13%) each year). The tertiles (roughly 33% of new diagnoses) of edu- Education, tertiles1 cation and deprivation at regional level are also shown least level: 33%-39% 3971 (22%) in Table 1. The majority was included into regions with moderate level: 40%-41% 8328 (46%) moderate level of education, i.e. they lived in regions with education level equal to 40%-41%; whilst the 37% high level: 42%-52% 5756 (32%) lived in least deprived regions, i.e. in regions with a de- Deprivation, tertiles2 privation level up to a maximum of 14%. least level: 8%-14% 6749 (37%) The general characteristics of the study population, shown in Table 1, were almost similar to that of the moderate level: 15%-16% 5589 (31%) target population, i.e., all HIV diagnoses from 2010 to high level: 17%-43% 5717 (32%) 2016 with the only exception of not indicated HIV ex- 1We assumed as least educated, individuals living in regions with an education posure group that was more frequent among individu- level ranging from 33% to 39% i.e., the “rate” of adults with a lower secondary als excluded (22%) compared to 8% of those included education level ranging from 33% to 39%, as moderate level from 40% to 41%, and as high level from 42% to 52%. 2We assumed as least deprived, individuals in the studied population. Further, the majority (about living in regions with a percentage of families with material deprivation from the 60%) of missing data comes from only one region 8% to 14%, as moderate level from15% to 16%, and high level from 17% to in Central Italy. 43%.

Association between HIV-risk perception and CD4 at first diagnosis defined HIV-risk perception within CD4 categories: A lower CD4 count was observed according to the the lowest CD4 (< 200 cells/mm3) was mostly asso- level of HIV-risk perception as shown by the box-plots ciated when reason for testing was 1 (“because HIV- of CD4 in Figure 1. Further, in Figure 2 is shown the symptoms”), reason for testing = 2 (“check-up for probability of each of the three groups wherewith we diseases other than HIV”) mostly was associated with 270 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti et al. reviews

and

articles

riginal O

Figure 1 Box-plots of CD4 count according to level of HIV-risk perception at new diagnosis ; HIV-risk perception: LOW = “because HIV-symp- toms”, MODERATE = “check-up for diseases other than HIV”, HIGH = “having engaged in behaviour at risk for HIV”; see methods, definition of HIV-risk perception.

1.0

0.8

0.6

Probability 0.4

0.2

0.0 CD4<200 CD4 200-350 CD4>350

HIV-risk perception LOW MODERATE HIGH

Figure 2 Probability estimated by ordinal logistic model of each HIV-risk perception level within CD4 category HIV-risk perception: LOW = “because HIV-symptoms”, MODERATE = “check-up for diseases other than HIV”, HIGH = “having engaged in behaviour at risk for HIV”; see methods, definition of HIV-risk perception.

CD4 from 201 to 350 cells/mm3, the reason for test- HIV-risk perception in the Italian Surveillance ing = 3 (“having engaged in behavior at risk for HIV”) (IHIVS) mostly was associated with CD4 > 350 cells/mm3, Table 2 shows the main characteristics of the study which confirms the previously discussed classification population by HIV-risk perception stratified by the in- orderings, i.e. assuming that those with the poorest dividual characteristics and regional socio-economic HIV-risk perception were those reporting mostly low- indicators: 4804 (27%), 7327 (40%), and 5924 (33%) er CD4 at first diagnosis (see methods, definition of were classified from the poorest to the highest HIV-risk HIV-risk perception). perception, respectively. 271 Determinants of poor HIV-risk perception

Table 2 General characteristics and socio-economic indicators at diagnosis according the level of HIV-risk perception (low, moderate, high; n = 18 055 new HIV diagnoses, Italy 2010-2016)

Low1 Moderate2 High3 n = 4804 (27%) n = 7327 (40%) n = 5924 (33%) n % n % n % Age reviews <40 2135 44% 4278 58% 3705 63% ≥40 2669 56% 3049 42% 2219 37% and Gender males 3828 80% 5345 73% 4813 81% females 976 20% 1982 27% 1111 19% articles

HIV exposure category heterosexuals 2483 52% 3706 51% 2190 37% IDU 225 5% 424 6% 189 3%

MSM 1621 34% 2423 33% 3177 54% riginal

not indicated 475 9% 774 11% 368 6% O Nationality Italians 3587 27% 5232 39% 4551 34% non-Italians 1217 26% 2095 45% 1373 29% Geographical area Northern Italy 2276 20% 5167 70% 3921 66% Central Italy 1297 50% 547 8% 760 13% Southern Italy 1231 30% 1613 22% 1243 21% Year of diagnosis 2010 632 13% 1041 14% 845 14% 2011 697 14% 1026 14% 917 15% 2012 785 16% 1159 16% 987 17% 2013 715 15% 1081 15% 868 15% 2014 690 14% 941 13% 842 14% 2015 657 14% 1089 15% 745 13% 2016 628 14% 990 13% 720 12%

Education, tertiles4 least level: 33%-39% 1923 40% 832 11% 1216 21% moderate level: 40%-41% 904 19% 4495 61% 2929 49% high level: 42%-52% 1977 41% 2000 28% 1799 30%

Deprivation, tertiles5 least level: 8%-14% 2449 51% 2070 28% 2230 38% moderate level: 15%-16% 245 5% 3285 45% 2059 35% high level: 17%-43% 2110 44% 1972 27% 1635 27%

1Low = “because HIV-symptoms”, 2Moderate = “check-up for diseases other than HIV”: 44% other pathologies, 33%, routine health checks, 11% pregnancy or other reproductive health checks, 8% pre-donation testing blood or organs, 4% hospital recovery for non-HIV pathologies; 3High = “having engaged in behaviour at risk for HIV: 86% unprotected sex, 2% newly HIV diagnosed partner, 12% not specified risk behaviour;4 we assumed as least educated, individuals living in regions with an education level ranging from 33% to 39% i.e., the “rate” of adults with a lower secondary education level ranging from 33% to 39%, as moderate level from 40% to 41%, and as high level from 42% to 52%; 5we assumed as least deprived, individuals living in regions with a percentage of families with material deprivation from 8% to 14%, as moderate level from15% to 16%, and high level from 17% to 43%.

Of those classified as with low HIV-risk perception were aged less than 40 yrs, MSM, living in the South- the majority were aged greater than 40 yrs, males, ern Italy and living in region with moderate level of heterosexuals, living in central Italy and living in most education. educated regions and least deprived regions. Of those In particular, of those categorized as with moderate classified as aware of HIV-risk perception, the majority perception, 44% were tested during ascertainments re- 272 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti et al.

lated to non-HIV symptoms, 33% during routine health in Central Italy (Centre vs South); regional education checks, 11% for checks during pregnancy or related to (least educated vs moderate educated areas); HIV-ex- medically assisted reproduction, 8% pre-donation test- posure group (heterosexuals and IDU vs MSM); older ing of blood, or of organs; 4% during hospital recovery age; gender (males vs females). Our results were not for pathologies other than HIV. Of those categorized in directly comparable with other studies, because of dif- the third group and for whom the reason for testing was ferent methods/definitions of HIV-risk perception or awareness of behaviours at risk of HIV: 86% because studied population. of unprotected sex, 2% because newly HIV diagnosed However our findings can be compared with the stud- reviews

partner, 12% with a risk behaviour non specified. ies relative to late presentation for HIV care, since late presentation can be considered as a consequence of

and Demographic characteristics and socio-economic low risk perception as described by a Swiss HIV Cohort

indicators as determinants of HIV-risk perception Study, in which late presentation to HIV care was driv- at first HIV diagnosis en by late HIV testing, due to lack of perception about In Table 3 both crude Odds Ratios (OR) and adjusted HIV [25]. Furthermore, in our study, this is corrobo- OR (AOR) of demographic and socio-economic char- rated by the fact that those defined with lower HIV-risk articles

acteristics possibly associated with poorer HIV-risk perception were also, on average, those with the lowest perception vs higher levels of HIV-risk perception are CD4 at diagnosis, as well as those more likely to be with shown. In the univariate analysis all variables resulted HIV-symptoms at diagnosis, thus those who were more associated with poorer HIV-risk perception (Table 3, likely to be late presenters as shown in the analyses. riginal crude OR), except for gender with males less likely with This study found that the effect of regional depriva- O low perception [crude OR of males vs females: crude tion on low HIV-risk perception was U-shaped: specifi- OR = 0.91 (95% CI: 0.85-0.97)], for geographical area cally, we showed that both those most deprived and (those who lived in the North were less likely to be with those least deprived had a significant higher risk of low perception respect those who lived in the South: poorer HIV risk perception (58% and 33% increments crude OR = 0.72, 95% CI: 0.67-0.77); calendar year of for those most and least deprived, respectively). In the diagnosis was not associated with poorer HIV risk per- case of the most deprived regions, our results were simi- ception. Again, when adjusting for all provinces within lar to the findings of the Swiss study where late presen- region of residence as random effect [Table 3, AOR, col- tation for HIV-care has been found to be more frequent umn (a)], results were similar except for gender [AOR in individuals living in neighbourhoods of lower socio- of males vs females = 0.97 (95% CI: 0.91-1.03)] and for economic status [25], whereas this is the first study that nationality [AOR of Italians vs non Italians = 1.03 (95% found an unexpected association of least deprivation CI: 0.97-2.00)]. with low HIV-risk perception. Meanwhile, in a French In the full model, regional deprivation (both least study the European Deprivation Index was not associ- and most deprived vs moderated deprived) and regional ated to late diagnosis [26]. education (least educated vs moderate educated), older In our opinion this finding (U-shaped effect of depri- age (> 40 vs ≥ 40 yrs), gender (males vs females), HIV- vation) could reflect the large differences in the organi- risk group and living in Central Italy resulted all asso- zation and provision of HIV care (with possible delay ciated with poorer HIV-risk perception (multivariate in HIV diagnosis) between the 20 Italian regions. Two analysis with all variables: Table 3, AOR of column e). realities coexist in Italy [27]: on the one hand the re- Of note, when adding regional deprivation in the model gions with less deprivation (i.e., Northern regions) with we observed an effect modification for geographical consequent better accessibility to diagnostic facilities, area of residence in the multivariate analysis (Table 3, therefore with greater probability to be diagnosed also AOR of column d), suggesting that regional deprivation during controls far from the suspicion of HIV (second may have a different effect on those living in the North category of HIV-risk perception definition), from the respect to the South of Italy. other, the regions with greater deprivation (especially in The between provinces within region of residence the South) and thus, with lower access and consequent variance (i.e., random intercept) resulted in all models greater probability of late diagnosis (first category of with a p-value < 0.001, indicating that there were sig- poor HIV-risk perception). This remark could partly nificant differences between provinces within region of explain, why when adding the deprivation in the mul- residence, also after accounting for sociodemographic tivariate model we observed a change in the effect of variables (Table 3, see the row with random effects pa- the geographic area. In fact, when we took account of rameters). different regional distribution of deprivation thus, lower access in the South and higher access in the North [27], DISCUSSION the likelihood of poorer HIV-risk perception was higher In Italy we found that there is a fairly high proportion in the Northern respect to the Southern area; while of people HIV diagnosed with poor risk perception at instead the differences between Central and Southern diagnosis, given the reason for undergoing HIV-testing Italy persisted. was HIV-symptoms (27%) or non-HIV health checks We observed that regional education affected lower (40%). HIV-risk perception for least educated areas vs middle Independent risk factors associated with poorer HIV- educated areas: our result was, partly in contrast with a risk perception were: regional deprivation (both most study performed by a probability sample survey of the and least deprived vs moderate deprived areas); living British population on HIV risk perception that found 273 Determinants of poor HIV-risk perception

Table 3 Crude Odds Ratio (OR) and adjusted OR of lower HIV-risk perception vs higher HIV-risk perception estimated by ordinal logistic models (n = 18 055 new HIV diagnoses; 107 Italian provinces within 20 regions)

Univariate models Adjusted (a) Adjusted (b) Adjusted (c) Adjusted (d) Adjusted (e) crude OR p-val OR p-val OR p-val OR p-val OR p-val OR p-val Age > 40 vs ≤ 40 yrs 1.66 < 0.001 1.71 < 0.001 1.60 < 0.001 1.59 < 0.001 1.60 < 0.001 1.59 < 0.001 (1.58-1.76) (1.61-1.81) (1.50-1.69) (1.50-1.69) (1.50-1.69) (1.50-1.69) reviews

Gender

Males vs Females 0.91 0.003 0.97 0.350 1.30 < 0.001 1.30 < 0.001 1.30 < 0.001 1.30 < 0.001 and (0.85-0.97) (0.91-1.03) (1.20-1.40) (1.20-1.40) (1.20-1.40) (1.20-1.40) HIV-exposure category Hetero vs MSM 1.88 < 0.001 1.83 < 0.001 1.96 < 0.001 1.96 < 0.001 1.96 < 0.001 1.96 < 0.001 articles

(1.77-2.00) (1.72-1.95) (1.82-2.10) (1.83-2.10) (1.83-2.11) (1.83-2.11) IDU vs MSM 1.91 < 0.001 1.83 < 0.001 1.82 < 0.001 1.82 < 0.001 1.82 < 0.001 1.82 < 0.001 (1.68-2.17) (1.60-2.10) (1.59-2.08) (1.59-2.08) (1.59-2.08) (1.59-2.08) Not indicated vs 2.02 < 0.001 2.56 < 0.001 2.54 < 0.001 2.55 < 0.001 2.53 < 0.001 2.54 < 0.001

MSM (1.83-2.23) (2.29-2.85) (2.27-2.84) (2.28-2.85) (2.26-2.82) (2.27-2.83) riginal Nationaliy O Non Italian vs 1.10 0.002 1.03 0.380 1.06 0.126 1.05 0.133 1.06 0.134 1.06 0.138 Italian (1.10-1.04) (0.97-2.00) (0.98-1.13) (0.98-1.13) (0.98-1.13) (0.98-1.13) Geographical area Center vs South 1.80 < 0.001 1.84 < 0.001 1.73 0.003 1.54 0.007 2.65 < 0.001 2.11 < 0.001 (1.63-1.98) (1.28-2.64) (1.20-2.51) (1.11-2.13) (1.84-3.83) (1.37-3.24) North vs South 0.72 < 0.001 0.79 0.074 0.75 0.034 0.94 0.907 1.76 0.002 1.54 0.132 (0.67-0.77) (0.61-1.02) (0.57-0.99) (0.63-1.40) (1.23-2.52) (0.88-2.69) Years 2016 vs 1.07 0.118 1.00 0.947 1.00 0.994 1.00 0.908 0.99 0.796 1.01 0.782 2010-2015 (0.98-1.16) (0.91-1.10) (0.90-1.10) (0.91-1.11) (0.90-1.09) (0.92-1.12) Education Low vs moderate 2.86 < 0.001 2.87 < 0.001 _ 2.60 < 0.001 _ 1.74 0.004 (2.66-3.08) (2.19-3.76) (1.93-3.49) (1.20-2.54) High vs moderate 1.88 < 0.001 1.83 < 0.001 _ 1.64 0.023 _ 1.44 0.104 (1.77-2.00) (1.43-2.33) (1.07-2.52) (0.93-2.22) Deprivation Low vs moderate 2.19 < 0.001 2.17 < 0.001 _ _ 1.98 < 0.001 1.58 0.006 (2.05-2.34) (1.59-2.97) (1.48-2.67) (1.14-2.18) High vs moderate 2.49 < 0.001 2.60 < 0.001 _ _ 3.86 < 0.001 2.33 0.001 (2.33-2.67) (1.93-3.50) (2.63-5.67) (1.39-3.90 ) Random effects parameters variance; SE; _ _ see note 2 0.369; 0.057; 0.247; 0.040; 0.242; 0.039; 0.221; 0.036; p-value < 0.001 < 0.001 < 0.001 < 0.001 (a) OR adjusted for provinces within region of residence as random intercept; (b) OR adjusted for provinces within region of residence as random intercept + individual characteristics; (c) OR adjusted for provinces within region of residence as random intercept + individual characteristics + education; (d) OR adjusted for provinces within region of residence as random intercept + individual characteristics + deprivation; (e) OR adjusted for provinces within region of residence as random intercept + individual characteristics+ education + deprivation note 2: Random effects parameters ( variance; SE; p-val ) for model with: age (0.441; 0.066; < 0.001); gender (0.434; 0.066; < 0.001); HIV-exposure cat (0.450; 0.067; < 0.001); nationality (0.435; 0.065; < 0.001); geographical area (0.350; 0.054; < 0.001); years (0.436; 0.066; < 0.001); education (0.266; 0.043; < 0.001); deprivation (0.309; 0.048; < 0.001).

that having academic qualifications was associated with between regional education and deprivation. lower HIV risk perception [14] at individual level. Of Our results on demographic risk factors could be note, also in our analysis before adjusting for depriva- overlapped to the same factors reported so far among tion we observed a similar result, i.e., higher educated late presenters in Italy [3] and in rest of Europe [4, areas were associated with lower HIV risk perception, 7]. Demographic factors were also consistent with the but when adjusting for regional deprivation this effect characteristics of all new HIV diagnoses in Italy [2, 3]: tended to disappear, suggesting possible interactions older age, in fact, was associated with HIV-risk percep- 274 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti et al.

tion, and it reflect the progressive increase of the me- regions (110 Italian provinces vs 20 Italian regions), and dian age at new HIV diagnosis observed in the IHIVS we obtained similar results when comparing findings [3]. Older age was associated with late presentation at from univariate models with those from models adjust- diagnosis in Italian studies [3], as well as, in rest of Eu- ed for provinces of residence entered as random effects. rope [6, 7], supporting that late presentation in older Another limit could rely on the definition of HIV-risk adults could be a consequence of the higher HIV-risk perception especially on ordering of the three groups. perception observed in older individuals. Age as risk For instance, why does an opt-out health care check for factor of low HIV-risk perception was also reported by a a pregnant woman would indicate higher risk perception reviews

study conducted in the British population on HIV risk than a doctor’s visit due to HIV related symptoms? The perception and HIV testing: authors showed that low only argument as to why this classification ordering was

and risk perception was associated with being older for all reasonable relies on the fact we observed a lower CD4

age classes between 25 and 74 yrs independently from count within each group with the decrease of the HIV- educational level [14]. risk perception just as it has been defined. Regarding gender, males showed lower HIV-risk per- There are a number of strengths in this study. The large ception than females. This result indirectly confirmed study sample included the majority of new HIV-diagno- articles

that new Italian HIV diagnoses over time are more fre- ses performed by HIV surveillance in Italy as shown in a quent among males than females in our surveillance [2, previous study [32] which increases the generalizability 3], as well as the finding that male gender is generally of our results. Further, this is the first study on HIV-risk associated with late presentation in Italian and Euro- perception using data from a HIV Surveillance System. riginal pean studies [3, 4]. However, this result was in contrast Another strength is the use of socio-economic indica- O with the finding HIV-risk perception was lower among tors that are internationally estimated parameters [20, females with respect to males in the general British 21], and this could ensure the reproducibility of these population [14] and in a Swiss cohort [25]. analyses in other European countries. We found that heterosexual transmission and IDU In conclusion, our findings on risk factors of HIV-risk compared to MSM were associated with lower HIV-risk perception are those reported in the majority of studies perception. This finding was similar to that shown by for HIV late presenters (i.e., the consequence of low Porter et al., on factors associated with lack of percep- HIV-risk perception): older adults, males, heterosexuals tion of HIV in UK [28] before 1996, and to an Ital- and IDU vs MSM, and those living in Central Italy had ian cohort study on risk factors relative to late present- more likely a low HIV-risk perception at first diagnosis. ers after 1996 [29]. Higher perception of HIV among Further, different regional deprivation and education, MSM respect to heterosexual males was shown also seemed to contribute to the lower HIV-risk perception in a more recent UK study [14]. However, the finding in Italy, indicating that both individual and regional ap- that MSM showed higher HIV-risk perception respect proaches are important in health care policies. to IDU and heterosexuals was in contrast with the in- creasing trend of new diagnoses among MSM reported Acknowledgements by Italian HIV surveillance, and the higher proportion The Authors wish to thank all the regional represen- of undiagnosed cases estimated among MSM respect tatives of the HIV Surveillance System for their useful the other risk groups in Italy [4]. These findings sug- help and constant availability. We would like to thank gest that although a large part of MSM are aware of the colleague Lucia Pugliese of Istituto Superiore di HIV at-risk practices, yet a proportion of them engage Sanità involved in the management of Italian HIV Sur- in high-risk behaviours that feed the reservoir of new veillance System. undiagnosed infections [30]. These behaviours can be associated with perceived partner knowledge and rea- Contributors sons reflecting perceived gay- and HIV-related stigma, All Authors contributed to the design of the study. thus delaying HIV testing [31]. Dorrucci M and Regine V analysed the data and draft- This study must acknowledge some limitations. Fore- ed the manuscript. Pezzotti P and Mammone A con- most, any generalization from the results should be tributed with statistical advices; Girardi E contributed made with caution; because of the cross-sectional na- to draft data interpretation; Suligoi B coordinates the ture of our data we are not able to draw conclusions Italian HIV Surveillance and discussed the results. All about causal effect; further we cannot exclude possible Authors commented on drafts of the manuscript and distortion resulting from “ecological fallacy”. In fact, the approved the final version. regional education and deprivation level was assigned to new HIV diagnoses according to their area of residence; Funding as a consequence, a patient was defined as “less deprived Italian HIV Surveillance was supported by the fol- and/or less educated” because he or she lives in a “less lowing grants: Ministry of Health – CCM, Article 4, deprived and/or less educated areas”. Thus, these results paragraph 7. Enrico Girardi and Alessia Mammone co- should be confirmed at individual-level, or at least in Authors were supported by “Ricerca corrente IRCCS smaller geographical areas. However, in order to control INMI L. Spallanzani”. for this limitation we assigned the regional estimates of education and deprivation to the provinces of residence Disclaimer within regions of each individual residence: in fact Ital- The funding body had no role in the design, execu- ian provinces are smaller geographic areas respect to the tion, analysis or reporting of the research. 275 Determinants of poor HIV-risk perception

Ethics approval oli, Daniele Giuseppe Chirico (Calabria); Guglielmo The study was undertaken in full accordance with the Borgia (Campania); Erika Massimiliani (Emilia Ro- Italian Ministry of Health and Italian National Institute magna); Tolinda Gallo, Cinzia Braida (Friuli Venezia of Health (Istituto Superiore di Sanità) regulations. Giulia); Vincenzo Puro, Paola Scognamiglio, Ales- sia Mammone (Lazio); Giancarlo Icardi, Piero Luigi Data sharing statement Lai (Liguria); Maria Gramegna, Liliana Coppola, No additional data available. Alessandra Piatti, Annamaria Rosa, Danilo Cereda (Lombardia); Fabio Filippetti (Marche); Alessandra reviews

Conflict of interest statement Prozzo (Piemonte); Chiara Pasqualini (Molise); Peter We declare that we have no conflicts of interest. Mian, Oswald Moling, Leonardo Pagani (Provincia

Autonoma di Bolzano); Paolo Lanzafame, Lucia Col- and

Received on 19 February 2020. lini, Danila Bassetti (Provincia Autonoma di Trento); Accepted on 11 June 2020. Maria Chironna (Puglia); Maria Antonietta Palmas (Sardegna); Gabriella Dardanoni (Sicilia); Fabio Vol- The members of the Italian HIV Surveillance System ler, Monia Puglia, Lucia Pecori (Toscana); Anna Tosti, articles includes at 31 December 2016: Rita Papili (Umbria); Mauro Ruffier, Marina Giulia Manuela Di Giacomo, Viviana Faggioni, Luigi Verardo, Elisa Francesca Echarlod, Saveria Amo- Scancella (Basilicata); Francesco Locuratolo, Ga- roso (Valle d’Aosta); Francesca Russo, Filippo da Re briella Cauzillo (Abruzzo); Anna Domenica Mignu- (Veneto). riginal O

REFERENCES

1. Italia. Decreto Ministeriale 31 marzo 2008. Istituzione 2):19152. doi: 10.7448/IAS.17.3.19152 del sistema di sorveglianza delle nuove diagnosi di infezi- 10. Dearborn JL, Lewis J, Mino GPPreventing mother-to- one da HIV. Gazzetta Ufficiale n. 175. 28 luglio 2008. child transmissionin Guayaquil, Ecuador: HIV knowledge 2. Camoni L, Boros S, Regine V, Santaquilani M, Ferri M, and risk perception. Glob Public Health. 2010;5(6):649- Pugliese L, e Suligoi B. Aggiornamento delle nuove diag- 62. doi: 10.1080/17441690903367141 nosi di infezione da HIV dei casi di AIDS in Italia al 31 11. Filipe EM, Batistella E, Pine A, Santos NJ, Paiva V, Se- dicembre 2014. Not Ist Super Sanità. 2015;28(9, Suppl. gurado A, Hearst N. Sexual orientation, use of drugs and 1):3-47. risk perception among HIV-positive men in Sao Paulo, 3. Raimondo M, Boros S, Regine V, et al. Aggiornamento Brazil. Int J STD AIDS. 2005;16:56-60. delle nuove diagnosi di infezione da HIV dei casi di 12. Koh KC, Yong LS. HIV Risk Perception, Sexual Be- AIDS in Italia al 31 dicembre 2015. Not Ist Super Sanità. havior, and HIV Prevalence among Men-Who-Have- 2016;29(9, Suppl. 1)3-51. Sex-with-Men at a Community-Based Voluntary 4. Mammone A, Pezzotti P, Regine V, Camoni L, Puro V, Counseling and Testing Center in Kuala Lumpur, Ma- Ippolito G, Suligoi B, Girardi E. How many people are laysia. Interdiscip Perspect Infect Dis. 2014;236240. doi: living with undiagnosed HIV infection? An estimate for 10.1155/2014/236240 Italy, based on surveillance data. AIDS. 2016;30(7):1131- 13. Evangeli M, Baker LL, Pady K, Jones B, Wroe AL. 6. doi:10.1097/QAD.0000000000001034 What leads some people to think they are HIV-posi- 5. Camoni L, Raimondo M, Regine V, Salfa MC, Suligoi B. tive before knowing their diagnosis? A systematic re- Late presenters among persons with a new HIV diagnosis view of psychological and behavioural correlates of in Italy, 2010–2011. BMC Public Health. 2013;13:281. HIV-risk perception AIDS Care. 2016;28:943-53. doi: doi:10.1186/1471-2458-13-281 10.1080/09540121.2016.1139040 6. Mocroft A, Lundgren JD, Sabin ML, et al. Risk factors 14. Clifton S, Nardone A, Field N, Mercer CH, Tanton and outcomes for late presentation for HIV-positive C, Macdowall W, Johnson AM, Sonnenberg P. HIV persons in Europe: results from the Collaboration of testing, risk perception, and behaviour in the Brit- Observational HIV Epidemiological Research Europe ish population. AIDS. 2016;30:943-52. doi: 10.1097/ Study (COHERE). PLoS Med. 2013;10(9):e1001510. QAD.0000000000001006 doi:10.1371/journal.pmed.1001510 15. Evangeli M, Pady K, Abigail LW. Which psychological fac- 7. The late presenters working group in COHERE in Eu- tors are related to HIV testing? A quantitative systematic roCoord. Late presentation for HIV care across Europe: review of global studies. AIDS Behav. 2016;20(4):880- update from the Collaboration of Observational HIV 918. doi: 10.1007/s10461-015-1246-0 Epidemiological Research Europe (COHERE) study, 16. MacKellar DA, Valleroy LA, Secura GM, et al. Men sur- 2010 to 2013. Euro Surveill. 2015;20. doi: http://dx.doi. vey study. Unrecognized HIV infection, risk behaviors, org/10.2807/1560-7917. ES.2015.20.47.30070 and perceptions of risk among young men who have sex 8. Tully S, Cojocaru M, Bauch CT. Sexual behavior, risk with men – Opportunities for advancing HIV prevention perception, and HIV transmission can respond to HIV in the third decade of HIV/AIDS. Jaids. 2005;38;603-14. antiviral drugs and vaccines through multiple pathways. doi: 0.1097/01.qai.0000141481.48348.7e Sci Rep. 2015;28:15411. 17. Massari V, Lapostolle A, Grupposo MC, Dray-Spira R, 9. Corneli AL, McKenna K, Headley J, et al. EM-PrEP Costagliola D, Chauvin P. Which adults in the Paris met- Study. A descriptive analysis of perceptions of HIV risk ropolitan area have never been tested for HIV? A 2010 and worry about acquiring HIV among FEM-PrEP par- multilevel, cross-sectional, population-based study. BMC ticipants who seroconverted in Bondo, Kenya, and Pre- Infect Dis. 2015;15:278. doi: 10.1186/s12879-015-1006- toria, South Africa. J Int AIDS Soc. 2014;17(3, Suppl. 9 276 Maria Dorrucci, Vincenza Regine, Patrizio Pezzotti et al.

18. ISTAT. Indicatori territoriali per le politiche di sviluppo. diagnosed patients in France. HIV Med. 2018;19(3):238- ISTAT; 2020. Available from: www.istat.it/it/condizioni- 42. doi: 10.1111/hiv.12545 economiche-famiglie.http://www.istat.it/it/archivio/16777 26. Fantini MP, Lenzi J, Franchino G, et al. Mortality ame- 19. European Commission. Eurostat Statistical Books; 2014. nable to health care services and health inequalities Living conditions in Europe. 2014. Available from http:// among Italian Regions. Epidemiol Prev. 2014;38:100-7. ec.europa.eu/eurostat/. 27. Porter K, Wall PG, Evans BG. Factors associated with 20. European Commission. Eurostat Statistical Books; 2015. lack of awareness of HIV infection before diagnosis of Quality of life, fact and views. 2015. Available from: AIDS. BMJ. 1993;307:20-3 http://ec.europa.eu/eurostat/. 28. Raffetti E, Postorino MC, Castelli F, et al. The risk of reviews

21. O’Connell AA. Logistic regression models for ordinal re- late or advanced presentation of HIV infected patients sponse variables. Thousand Oaks, CA: Sage Publications; is still high, associated factors evolve but impact on

and 2006. overall mortality is vanishing over calendar years: results 22. Goldstein H. Multilevel statistical models, 4nd ed. from the Italian MASTER Cohort. BMC Public Health. Chichester, UK: John Wiley and Sons; 2011. 2016;16:878. doi: 10.1186/s12889-016-3477-z 23. Moerbeek M. The consequences of ignoring a level of nest- 29. Gios L, Mirandola M, Toskin I, et al. Bio-behavioural ing in multilevel analysis. Multivariate Behavioral Research. HIV and STI surveillance among men who have sex with

articles 2004;39:129-49. doi:10.1207/s15327906mbr3901_5 men in Europe: the Sialon II protocols. BMC Public 24. Hachfeld A, Ledergerber B, Darling K, Weber R, Calmy Health. 2016;16:212. doi: 10.1186/s12889-016-2783-9 A, Battegay M, Sugimoto K, Di Benedetto C, Fux CA, 30. Marcus U, Gassowski M, Drewes J. HIV risk perception Tarr PE, Kouyos R, Furrer H, Wandeler G; Swiss HIV and testing behaviours among men having sex with men Cohort Study. Reasons for late presentation to HIV care (MSM) reporting potential transmission risks in the pre- riginal in Switzerland. J Int AIDS Soc. 2015;18:20317. doi: vious 12 months from a large online sample of MSM O 10.7448/IAS.18.1.20317 living in . BMC Public Health. 2016;16:1111. 25. Cuzin L, Yazdanpanah Y, Huleux T, Cotte L, Pugliese P, 31. Camoni L, Regine V, Stanecki K, Salfa MC, Raimondo Allavena C, Reynes J, Poizot-Martin I, Bani-Sadr F, Del- M, Suligoi B. Estimates of the Number of People Living pierre C; Dat’AIDS Study Group. No relationship be- with HIV in Italy. Biomed Res Int. 2014;2014:209619. tween late HIV diagnosis and social deprivation in newly doi: 10.1155/2014/209619 Via Tronto 10/A,60120 Torrette diAncona, Italy. E-mail: [email protected]. Address forcorrespondence : MarianxhelaDajko, CentrodiEpidemiologia,Biostatistica eInformaticaMedica,Università PolitecnicadelleMarche, INTRODUCTION impact ofunknownetiologicaldiagnosisonantibiotictreatment andresistance. than 50%.Albeitinlinewithotherevidence,thisresultshould calltheattentionon Conclusion. Theproportionofmicrobiologicalascertaimentinpneumonia remainsless rather thaninnon-ICU(71%vs12%). to a primary diagnosis (23% vs11%) and inhospitalized in intensive careunit-ICU- tilated patients(41%vs11%),incaseswithsecondarydiagnosisofpneumoniacompared surgical compared to medical patients (27% elective admission. The overall mortality was 17%. Mortality was significantly higher: in (43% tients had significantly lower proportion of ascertainment compared tosurgical patients reported in secondary diagnosis instead of primary diagnosis (52% vs 42%). Medical pa - significantly higherinmalesthanfemales(51%vs40%)andcasesofpneumonia Results. 2819 records were identified, where 46% had a microbiological ascertainment, surveillance database. same hospital:theelectronichospitaldischargeregisterandmicrobiologylaboratory Methods. Thestudy relied ontherecordlinkageoftwoadministrativedatabases one ofthemostimportantteachinghospitalsinRome. proportion ofmicrobiologicalascertainmentinpneumonia-relatedhospitalizations appropriate therapeuticalapproach.Thisstudyisaimedatprovidingevidenceonthe portant in the light of avoiding unnecessary antibiotic treatment and choosing the most point becauseoftherelevantepidemiologicalburden.Theetiologicaldiagnosisisim- Objectives. Pneumoniastillremainsaproblemfromtheclinicalandpublichealthview- Abstract 8 7 6 5 4 3 2 Marianxhela Dajko in Rome the experienceofateachinghospital in patientswithpneumonia: Microbiological ascertainment DOI: 10.4415/ANN_20_03_05 Ann IstSuperSanità2020|Vol. 56,No.3:277-284 1 Universitario AgostinoGemelliIRCCS,Rome,Italy Cattolica delSacroCuore,Rome,Italy Jesi (Ancona),Italy Italy Massimo Volpe tality inadults[1-2]withamajor epidemiologicalrele- Dipartimento diMedicinaSperimentale,UniversitàdegliStudiPerugia,Italy Dipartimento Scienzedellasalutedonna,delbambinoedisanitàpubblica,FondazionePoliclinico Fondazione IRCCSPoliclinicoSanMatteodiPavia,Italy Fondazione PoliclinicoUniversitarioAgostinoGemelliIRCCS,Rome,Italy Istituto diMicrobiologia,FondazionePoliclinicoUniversitarioAgostinoGemelliIRCCS,Università Istituto diSanitàPubblica,SezioneIgiene,UniversitàCattolicadelSacroCuore,Rome,Italy ASUR Marche–AV2 –UOCISPPrevenzioneeSorveglianzaMalattieInfettiveCronicoDegenerative, Centro diEpidemiologia,BiostatisticaeInformaticaMedica,UniversitàPolitecnicadelleMarche,Ancona, Pneumonia isaleadingcause ofmorbidityandmor vs 67%)whereastherewerenotdifferencesbetweenpatients withemergencyand 5 , SilviaMancinelli 1 , AndreaPoscia 6 , Walter Ricciardi 2 vs 15%), in ventilated compared to not ven- , BrunellaPosteraro - nia, arethefourthmostcommon causeofdeathwith respiratory tractinfections(LTRI), including pneumo- vance worldwide.Infact,theevidence reportsthatlower 7 andChiaradeWaure 3 , DomenicoSpeziale

• • • • Key words 8 diagnosis ascertainment microbiological etiology pneumonia 4 ,

277 Original articles and reviews 278 Marianxhela Dajko, Andrea Poscia, Brunella Posteraro et al.

1.9 million people over 14 years of age dying every year diagnostic information is required, clinical setting and [3]. Furthermore, they are responsible for more disabil- clinicians’decision making. ity-adjusted life years lost around the world than any This study aimed at providing evidence about the pro- other category of disease, including cancer and cardio- portion of microbiological ascertainment of pneumonia vascular diseases [4]. In Europe, the cost of caring for in a big teaching hospital in Italy and at elucidating char- patients with pneumonia is estimated to be around € acteristics associated with the request of microbiological 10.1 billion annually [5]. Pneumonia represent 9.4% of ascertainment and its impact on mortality. severe infections [6] with a reported incidence varying reviews

considerably among countries (1.7-11.6/1000 person- METHODS years) [7-9] and increasing with age (7.65-15.3/1000 A retrospective cohort study has been performed at

and person-years in adults aged ≥ 65 years) [1-13]. Albeit the “A. Gemelli” Teaching Hospital including all dis-

pneumonia can affect anyone, it occurs especially with charged patients from November 13th 2010 to March increasing frequency in individuals whose immune sys- 26th 2013 with a diagnosis of pneumonia. For the pur- tem is deficient or compromised [14]. Also, infants and pose of this study the term “Pneumonia” was referred very young children are highly vulnerable, as well as the to pneumonia from any cause. Pneumonia related hos- articles elderly. In fact, at both extremes of age, the increased pitalizations were assessed on the basis of the Interna- risk relates in part to impaired immunity and often de- tional Classification of Diseases, 9th revision – Clinical termines hospitalization [15]. Some studies have dem- Modification (ICD-9-CM) codes [29] using the elec- onstrated an increasing trend towards hospitalization, tronic hospital discharge register. All patients’ records riginal in particular in the elderly [16-18]. In the USA indi- reporting a first-listed ICD-9-CM code discharge diag- O viduals older than 65 years of age account for almost nosis of pneumonia (003.22, 011.6, 052.1, 055.1, 073.0, two thirds of hospitalizations and 90% of deaths from 115.05, 115.15, 115.95, 130.4, 480-487.0, 495.7, 495.9, pneumonia [16]. Because of population ageing and 506.0, 507, 517.1, 770.0) were considered. Similarly, a the growing number of patients with multiple vulner- first-listed discharge diagnosis of meningitis or septice- abilities, beside the increased proportion of cases re- mia (003.21, 013.0, 036.0, 036.1, 047, 049.0, 049.1, quiring hospitalization, there has been also a growth in 053.0, 054.72, 072.1, 090.42, 091.81, 094.2, 098.82, the proportion of patients experiencing poor outcomes 100.81, 112.83, 114.2, 115.01, 115.11, 115.91, 130.0, [19]. Antimicrobial resistance may contribute to poor 320, 321, 322, 003.1, 020.2, 022.3, 031.2, 036.2, 036.3, outcomes and is linked to inappropriate use of antibi- 038, 054.5, 785.50, 785.59, 790.7, 790.8) in addition otics. Thus, information on the potential pathogens is to a diagnosis of pneumonia in another diagnostic field important for a proper antimicrobial treatment, avoids was considered to identify the study population. Micro- inappropriate antibiotics prescriptions and eventually, biological tests performed in the same time period on the antimicrobial resistance [20-22]. Furthermore, the respiratory tract (e.g., sputum) and blood cultures, as identification of the pathogens is a key factor for a good well as urine tests for Streptococcus pneumonia and Le- prognosis, especially in cases with mixed etiology, who gionella pneumophila antigens were collected from the often develop severe pneumonia, and have longer hos- Institute of Microbiology of the “A. Gemelli” Teaching pitalization and poorer outcomes [23-25]. Hospital. Electronic hospital discharge records were Although clinical variables are independently associ- used to collect data on patients’ demographic character- ated with the detection of a pathogen group, there are istics and hospitalization information. As for patients’ no reliable clinical predictors to distinguish causative characteristics the following information was recorded: aetiologies [25]. Therefore, diagnostic testing, with the date of birth, age, sex, health status at discharge (dead ability to detect the causative pathogens have the po- or not). Comorbidity was also assessed using ICD-9- tential to guide to a more rational use of antibiotics, CM codes and Charlson’s index syntax integrated to firstly by distinguishing between viral and bacterial in- Stata [30]. The Charlson Comorbidity Index (CCI) is a fections, and then by identifying specific pathogens and method of categorizing comorbidities of patients based their antibiotic resistance pattern. Patients gain more on the ICD-9-CM codes. Each comorbidity category from a rapid and effective use of antibiotics and soci- has an associated weight (from 1 to 6), based on the ety gains from the reduction of the unselective use of adjusted risk of mortality or resource use, and the sum antibiotics, that has been considered as a major factor of all the weights results in a single comorbidity score driving the emergence and spread of resistance. for a patient. A score of zero indicates that no comor- In the past 20 years, there has been a decline in in- bidities were found. A score equal to one represents the terest and perceived need for microbiological analysis in presence of one of the following diseases: myocardial pneumonia, to the point that the vast majority of patients infarction, congestive heart failure, peripheral vascular have no microbial pathogen diagnosis [26]. disease, cerebrovascular disease, dementia, chronic ob- Moreover, it should be noted that in many patients structive pulmonary disease, rheumatoid disease, pep- the etiology remains unknown even after routine di- tic ulcer disease, mild liver disease and diabetes. In con- agnostic workup. The etiology of pneumonia has still trast, a score of 2 indicates the presence of more serious not been well characterized [27] and remains unknown comorbidities such as complicated diabetes, hemiplegia nearly in 50% of cases [28]. The lack of information may or paraplegia, renal disease, cancer, moderate / severe be due to a range of common clinical scenarios depend- liver disease, metastatic cancer and AIDS or the pres- ing on the severity of infection, the potential conse- ence of more than one comorbidity. quences of an incorrect diagnosis, timescales in which Patients with weakened immune system were identi- 279 Microbiological ascertainment in pneumonia

fied according to the following codes: V58.12, 283.10, Table 1 287.31, 283.0, 279.3, ,279.2, 279.06, 279.03, 203.80, Characteristics of the study sample V08, 079.53, 042. N (%) With respect to organizational features, the follow- ing variables were extracted: type of discharge, hospital Diagnosis of pneumonia ward, length of stay (LOS), modality of admission, diag- Primary 1597 (56.65) nosis, Diagnosis Related Group (DRG), type of DRG. Secondary 1222 (43.35) Hospital wards were distinguished in intensive care Age (2802)* 71.3 [52.4-82.1] reviews unit (ICU) and non-ICU. Ventilation was assessed ac- cording to the following procedure codes: 96.04, 96.7, Sex

96.71, 96.72, 93.9. For patients with multiple eligible Females 1225 (43.46) and admissions in the study period, each hospitalization was Males 1594 (56.54) considered as an independent observation except for re- admissions until 30 days that were excluded because Modality of admission possibly related to the first admission. Emergency 611 (21.67) articles Data from the electronic hospital discharge register Elective 2208 (78.33) and from the microbiology laboratory database were DRG Type merged by means of the unique identity number which is automatically assigned to each patient once admitted Medical 2454 (87.05) riginal to the “A. Gemelli” Teaching Hospital and used during Surgical 365 (12.95) O all hospitalizations or services provided by the hospital. Hospital ward Data from the microbiology laboratory database were considered only if the microbiological ascertainment Non – ICU 2590 (91.88) was requested during the hospital stay. The determin- ICU 229 (8.12) istic record linkage allowed calculating the total num- Length of stay (days)* 12 [7-19] ber of microbiological ascertainments for each patient Comorbidity during the hospitalization, and assessing their positive and negative results. With this respect, microbiological Myocardial infarction 45 (1.60) ascertainment was defined as “positive” if the patient Congestive heart failure 250 (8.87) had at least one positive finding. The proportion of as- Peripheral Vascular Disease 30 (1.06) certainment was yielded with 95% Confidence Interval. Differences between groups (patients with and without Cerebrovascular disease 276 (9.79) microbiological ascertainment request; patients died or Chronic Obstructive Pulmonary Disease 315 (11.17) alive at discharge) have been analyzed through the ap- Rheumatoid Disease 6 (0.21) plication of Chi-squared test with continuity correction PUD (Peptic Ulcer Disease) 6 (0.21) for categorical variables and non parametric tests for continuous variables. All the significant variables, and HP/PAPL (Hemiplegia or Paraplegia) 17 (0.6) the ones with 0.05 < p < 0.25 detected in the univariate RD (Renal Disease) 202 (7.17) analysis, were included in a multiple regression analy- Cancer 366 (12.98) sis. The analysis was performed with a logistic binomil- regression model having death as outcome and the level Metastatic Cancer 102 (3.62) of statistical significance was set at 0.05. Stata 12 soft- AIDS 80 (2.84) ware was used for the deterministic record linkage and Dementia 43 (1.53) the statistical analysis. Microbiological ascertainment RESULTS Present 1303 (46.22) A total of 2819 records were included in the pres- Absent 1516 (53.78) ent study; 1225 (43.46%) referred to females; patients’ *median [interquartile range]. median age was 71.3 years (interquantile range, 52.4 to 82.1). 1705 records (60.48%) were referred to people over 64 years old. 50.16% of patients had comorbidi- a positive result. Blood culture was the most common ties, like cerebrovascular disease, congestive heart fail- requested ascertainment (31.1%) with a positivity find- ure, acute myocardial infarction, renal diseases, chronic ing in 25.17% of cases (Table 2). Naso-faringeal swabs obstructive pulmonary disease, with a predominance of cancer (12.98%). 1597 records (56.65%) had pneu- were obtained from 54 (1.92%) patients, a sputum speci- monia reported in primary diagnosis and 365 (12.95%) men from 479 (16.99%), a bronchial washing from 390 had a surgical DRG type. Among all patients only 229 (13.83%), a bronchial aspirate from 65 (2.31%), a pleural (8.12%) required intensive care. The median LOS was fluid specimen from 53 (1.88%). Eventually, urine speci- 12 days (interquantile range, 7 to 19). The characteris- men was obtained from 629 (22.31%) and the request of tics of the study sample are presented in Table 1. Legionella and streptococcal antigen were performed in A microbiological ascertainment was requested in 22.31% and 9.01% respectively. Diagnostic results were 1303 (46.22%) patients but only 743 (57.02%) yielded available from 743 individuals. We identified 80 different 280 Marianxhela Dajko, Andrea Poscia, Brunella Posteraro et al.

Table 2 Characteristics of collected specimen Rates Positive specimen Analyzed specimen N % [95% CI] N % [95% CI] Blood culture 878 31.1 [29.44-32.89] 221 25.17 [22.33-28.18] Urine 629 22.3 [20.79-23.9] 38 6.04 [4.31-8.20] Sputum 479 17 [15.62-18.43] 399 83.3 [79.65-86.53] reviews

Bronchial washing 390 13.8 [12.58-15.16] 304 77.95 [73.5-81.97]

and Bronchial aspirate 65 2.3 [1.78-2.93] 54 83.08 [71.73-91.24]

Pleural fluid 53 1.9 [1.41-2.45] 12 22.64 [12.28-36.21] Naso-faringeal swabs 54 1.9 [1.44-2.49] 34 62.96 [48.74-75.71] Biopsy material 4 0.1 [0.04- 0.36] 2 50 [6.76-93.24] articles

types of etiologic agents causing pneumonia. The most and admission modality were considered. As shown in riginal common pathogen was S. pneumoniae identified in 299 Table 5 only five factors were eventually associated with O (40.24%) patients (Supplementary Table 1, available on- mortality: age, elective admission, ventilation, second- line). ary diagnosis of pneumonia and admission in ICU (all Table 3 reports the results of the univariable analy- associated with increased mortality). sis investigating the association between patients’ and organizational characteristics and the request of the Table 3 ascertainment. The latter was significantly lower in Microbiological ascertainment proportion and the association females as compared to males; in patients with a pri- with patients’ and organizational characteristics mary diagnosis of pneumonia in comparison to patients with a secondary diagnosis of pneumonia; in patients Microbiological ascertainment with grouped Charlson index equal to 0 or 1 as com- Yes No p-value pared to 2 and in patients with a medical DRG. The N (%) N (%) median age of patients who had a microbiological as- Sex certainment was significantly lower, 66.8 (interquantile Female 496 (40.49) 729 (59.51) < 0.001 range, 49.3 to 78.5) as compared to 74.7 (interquantile range, 57.3 to 84.3) for the ones who had not it and the Male 807 (50.63) 787 (49.37) proportion of ascertainment was significantly lower in Age 66.8 [49.3-78.5] 74.7 [57.3-84.3] < 0.001 the pediatric age compared to the other groups. More Age groups than half of the pediatric group and of elderly did not have a microbiological ascertainment. There were no < 15 years 87 (34.66) 164 (65.34) < 0.001 significant differences between patients admitted from 15-64 years 529 (62.53) 317 (37.47) the emergency room and those with an elective admis- > 64 years 687 (40.29) 1018 (59.71) sion. Among 543 patients who required ventilation, 355 CCI (65.38%) had a microbiological ascertainment and 275 (77.46%) a positive result. 66 (66.67%) patients with 0 634 (45.12) 771 (54.88) < 0.001 compromised immunity (AIDS included) had a micro- 1 236 (41.55) 332 (58.45) biological ascertainment but only in 44 (66.67%) the 2 433 (51.18) 413 (48.82) etiology was identified. The total number of deaths was 467 (16.57%) across Diagnosis of pneumonia the 2819 records. Higher mortality was observed in Primary 672 (42.08) 925 (57.92) < 0.001 people over 64 years of age, 22.2% (N = 379) as com- Secondary 631 (51.64) 591 (48.24) pared to 9% (N= 84) and 1.6% (N = 4) in people from DRG Type 15 to 64 years old and younger than 15 years respective- ly. Several variables showed an association with mortal- Surgical 244 (66.85) 121 (33.15) < 0.001 ity in the univariate analysis in Table 4. Mortality was Medical 1059 (43.15) 1395 (56.85) significantly higher in patients with secondary diagnosis Hospital ward of pneumonia, hospitalized in ICU wards, with an elec- tive admission and a surgical DRG. Mortality was also Non-ICU 1135 (43.82) 1455 (56.18) < 0.001 higher in people having a microbiological ascertain- ICU 168 (73.36) 61 (26.64) ment. After fitting the model for the multiple regression Admission Modality analysis of mortality: age, sex, CCI, the microbiological Emergency 299 (48.94) 312 (51.06) 0.128 ascertainment, type of admission, ventilation, hospital ward diagnosis of pneumonia (primary or secondary) Elective 1004 (45.47) 1204 (54.53) 281 Microbiological ascertainment in pneumonia

Table 4 Table 5 Mortality association with patients’ and organizational charac- Mortality multiple regression teristics Variables OR [95% CI] p-value Dead Age 1.06 [1.05- 11.07] < 0.001 Yes No p-value N (%) N (%) stay in ICU vs non-ICU 16.34 [10.67- 25.01] < 0.001 Sex Ventilation 3.94 [2.91- 5.33] < 0.001

Female 190 (15.51) 1035 (84.49) 0.186 reviews

Secondary (vs primary) 1.84 [1.43- 2.37] < 0.001 Male 277 (17.38) 1317 (82.62) diagnosis of pneumonia

Surgical vs medical DRG 0.72 [0.49- 1.04] 0.077 and Age group < 15 years 4 (1.59) 247 (98.41) < 0.001 Male vs female 1.16 [0.91- 1.49] 0.230 15-64 years 84 (9.03) 846 (90.97) Presence of 1.19 [0.92- 1.53] 0.178 microbiological > 64 years 379 (22.23) 1326 (77.77)

ascertainment articles

CCI Charlson index 1 (vs 0) 0.78 [0.58- 1.08] 0.137 0 203 (14.45) 1202 (85.55) 0.009 Charlson index 2 (vs 0) 1.17 [0.88- 1.55] 0.274 1 110 (19.37) 458 (80.63)

Elective vs emergency 1.41 [1.01- 1.96] 0.041 riginal 2 154 (18.2) 692 (81.8) admission O Microbiological ascertainment Not Present 216 (14.25) 1300 (85.75) < 0.001 Present 251 (34.24) 482 (65.76) found that all or almost most of patients underwent a Age* 79.4 [69.6-87.1] 69.2 [49.2-80.6] < 0.001 diagnostic test [31, 32], but other evidence suggested a Length of stay 13 (5-23) 12 (7-9) < 0.001 lower proportion of etiological diagnosis [34]. This het- erogeneity is actually expected because of discordant or Diagnosis of pneumonia not conclusive recommendations released by scientific Primary 180 (11.27) 1417 (88.73) < 0.001 societies. The guidelines on CAP yielded by the Infec- Secondary 287 (23.49) 935 (76.51) tious Disease Society of America/American Thoracic So- DRG Type ciety (IDSA/ATS) suggests a microbiology ascertainment in cases of severely ill patients (including those in ICU) Surgical 100 (27.4) 265 (72.6) < 0.001 or if a pathogen likely to change antibiotic management Medical 367 (14.96) 2087 (85.04) is suspected [1]. Similarly, as for children, the Pediatric Hospital ward Infectious Diseases Society and the Infectious Diseases Society of America recommends to collect blood cul- Non-ICU 305 (11.78) 2285 (88.22) < 0.001 tures in case of moderate or severe CAP even though ICU 162 (70.74) 67 (29.26) the definition of the severity is not well established [35]. Admission Modality On the contrary, the ascertainment is suggested in case Emergency 72 (11.78) 539 (88.22) < 0.001 of Hospital-Acquired Pneumonia (HAP) or Ventilator- Associated Pneumonia (VAP) [36]. In this light, the Elective 395 (17.89) 1813 (82.11) elaboration on our results is complex due to the lack of Ventilation differentiation between types of pneumonia. Neverthe- Yes 222 (40.88) 321 (59.12) < 0.001 less, looking at the percentages of people both admitted No 245 (10.76) 2031 (89.24) in ICU and with comorbidities – indeed most likely to have moderate to severe CAP – it can be assumed that *median [interquartile range]. the proportion of microbiological ascertainment shown in our study is acceptable considering current guidelines and the evidence that there is a worldwide tendency to DISCUSSION over-testing considering IDSA/ATS guidelines [31]. This work pointed out that a microbiological ascertain- Independently of the proportion of microbiologi- ment in patients with a diagnosis of pneumonia in a big cal ascertainment, an etiological diagnosis was overall teaching hospital in Rome was requested in less than achieved in 26.4% of cases and this result is aligned with 50%. Blood culture was the most common requested as- other evidence coming from adults but also elderly and certainment and S. pneumoniae was the most common children [31, 32, 37]. The positivity rate found in differ- identified etiologic agent. If data about blood culture as ent specimens is in agreement with international litera- the most common diagnostic test and S. pneumoniae as ture and suggests that blood culture has low sensitivity the most common pathogen may be confirmed by other in detecting pneumonia aetiology [31, 35]. multicentre and international evidence [31], the com- The request of the ascertainment was significantly ment on the proportion of ascertainment is not straight- higher, with a proportion over 60%, in adults, in pa- forward. Some international studies performed on adult tients staying in ICU and in patients with a surgical patients with community-acquired pneumonia (CAP) DRG. Furthermore, it was higher in cases requiring 282 Marianxhela Dajko, Andrea Poscia, Brunella Posteraro et al.

ventilation and with underlying immunocompromised ertheless, because of the source used to collect data, conditions. These data seem to reflect the current rec- our study did not encompass other factors that have ommendations on etiological diagnosis of pneumonia. been shown important in the assessment of the propor- With respect to mortality, our overall results seem to tion of ascertainment and pneumonia outcomes, such be higher than those released by other studies such as as the specimen quality, potential antimicrobial treat- the REACH [32], but it should be taken into consid- ment before the hospital admission, possible change of eration that our work encompassed all patients with a the therapy after consulting the microbiological ascer- diagnosis of pneumonia independently by the fact that tainment result [40]. Another pitfall could raise from reviews

it was community or hospital acquired. Higher mortality the potential inappropriate use of ICD-9-CM codes was observed in people over 64 years of age, in patients which could lead to misclassification. Nevertheless, in

and with a secondary diagnosis of pneumonia, and in those our opinion, this problem could mainly affect the iden-

hospitalized in ICU, with an elective admission and with tification of potential comorbidities instead of the se- a surgical DRG. These results may be related to the more lection of eligible study population. Another significant severe clinical course in these categories of patients. limitation is the lack of laboratory data regarding the Considering the high frequency of S. pneumoniae among viral specimen that could possibly increase the propor- articles the etiological agents, the pneumococcal vaccination of tion of ascertainment. this group of patients could be widely worthwhile. In Among the strengths of the study we should mention fact, elderly as well as people with comborbidities are that many records were analysed and that selection bias considered as eligible candidates for pneumococcal vac- can be ruled out because all records reporting the se- riginal cination and physicians may benefit from the hospital lected ICD-9-CM were considered. Furthermore, this O stay of these patients to raise their awareness on the study is one of the few attempts to describe the actual need for vaccination [33]. A finding arising from univari- diagnostic management of pneumonia in the Italian able analysis that is worth discussing was that mortality hospital setting and to investigate how the latter relates was higher in people having a microbiological ascer- with the outcome. tainment. The evidence on this aspect is counteracting with data showing a reduced mortality in patients who CONCLUSION underwent multiple guideline concordant microbiologi- Hospital information data can be used to give a cal testing for CAP [34] and evidence suggesting that timely and inexpensive picture of several diseases that microbiological diagnosis is associated to worse clinical frequently require hospitalisation and are not included course and higher in-hospital mortality rate [38]. This in special surveillance systems or more analytical regis- could be explained by the delay in the laboratory results tries. The population ageing will strongly increase the and eventually in setting a specific therapy. In fact, the burden of pneumonia in the near future. Pathogen‘s results of microbiological ascertainment may allow to detection in pneumonia is essential to inform clinical change antibiotic treatment, but initial antibiotic treat- management decisions and research priorities especial- ment modification has been shown associated with ly in the fields of vaccine and antibacterial and antiviral higher resource use and costs as well as higher mortality development [41, 42]. [39]. This should call the attention on the importance of deepening the knowledge of the aetiology of pneumonia Conflict of interest statement in order to better direct the empirical therapy [31] and There is no conflict of interest that could compromise to counter the problem of antibiotic resistance. the impartiality of the research reported. Our study has some limitations. The most important one is that we did not make any distinction among dif- Funding ferent types of pneumonia because the study was only This research did not receive any specific grants or relied on administrative data flows. This aspect could fellowships from any funding agency in the public, com- limit the interpretation and generalizability of results mercial or non-profit sectors. but, on the other hand, makes it possible to have a thor- ough overview of current practice with respect to all pa- Received on 21 October 2019. tients discharged with a diagnosis of pneumonia. Nev- Accepted on 31 March 2020.

REFERENCE

1. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, 3. Roth GA, Abate D, Abate KH, Abay SM, Abbafati C, Campbell GD, Dean NC, et al. Infectious Diseases So- Abbasi N, et al. Global, regional, and national age-sex- ciety of America/American Thoracic Society consensus specific mortality for 282 causes of death in 195 coun- guidelines on the management of community-acquired tries and territories, 1980-2017: a systematic analysis for pneumonia in adults. Clin Infect Dis. 2007;44(Suppl. the Global Burden of Disease Study 2017. The Lancet. 2):S27-72. 2018;392(10159):1736-88. 2. Lim WS, Baudouin SV, George RC, Hill AT, Jamieson C, 4. Mizgerd JP. Lung Infection. A public health priority. Le Jeune I, et al. BTS guidelines for the management of PLoS Med. 2006;3(2):e76. community acquired pneumonia in adults: update 2009. 5. Welte T, Torres A, Nathwani D. Clinical and economic Thorax. 2009;64(Suppl. 3):iii1-55. burden of community-acquired pneumonia among adults 283 Microbiological ascertainment in pneumonia

in Europe. Thorax. 2012;67(1):71-9. J. Antibiotic stewardship in community-acquired pneu- 6. Nicoletti G, Schito G, Fadda G, Boros S, Nicolosi D, monia. Expert Rev Anti Infect Ther. 2017;15(4):351-9. Marchese A, et al. Bacterial Isolates from Severe Infec- 23. Jennings LC, Anderson TP, Beynon KA, Chua A, Laing tions and Their Antibiotic Susceptibility Patterns in Italy: RTR, Werno AM, et al. Incidence and characteristics of a Nationwide Study in the Hospital Setting. J Chemoth- viral community-acquired pneumonia in adults. Thorax. er. 2006;18(6):589-602. 2008;63(1):42-8. 7. Rozenbaum MH, Mangen M-JJ, Huijts SM, van der 24. Johansson N, Kalin M, Hedlund J. Clinical impact of Werf TS, Postma MJ. Incidence, direct costs and dura- combined viral and bacterial infection in patients with tion of hospitalization of patients hospitalized with com- community-acquired pneumonia. Scand J Infect Dis. reviews

munity acquired pneumonia: A nationwide retrospective 2011;43(8):609-15. claims database analysis. Vaccine. 2015;33(28):3193-9. 25. Huijskens EGW, Koopmans M, Palmen FMH, van Erkel

8. Ewig S, Birkner N, Strauss R, Schaefer E, Pauletzki J, AJM, Mulder PGH, Rossen JWA. The value of signs and and Bischoff H, et al. New perspectives on community-ac- symptoms in differentiating between bacterial, viral and quired pneumonia in 388 406 patients. Results from a mixed aetiology in patients with community-acquired nationwide mandatory performance measurement pro- pneumonia. J Med Microbiol. 2014;63(3):441-52. gramme in healthcare quality. Thorax. 2009;64(12):1062-9. 26. Bartlett JG. Diagnostic Tests for agents of community-

9. Sun X, Douiri A, Gulliford M. Pneumonia incidence acquired pneumonia. Clin Infect Dis. 2011;52(Suppl. articles trends in UK primary care from 2002 to 2017: population- 4):S296-304. based cohort study. Epidemiol Infect. 2019;147:e263. 27. Wunderink RG, Waterer G. Advances in the causes and 10. Vila-Corcoles A, Ochoa-Gondar O, Rodriguez-Blanco management of community acquired pneumonia in T, Raga-Luria X, Gomez-Bertomeu F. Epidemiology of adults. BMJ. 2017;j2471. community-acquired pneumonia in older adults: A popu- 28. Spoorenberg SM, Bos WJW, Heijligenberg R, Voorn riginal lation-based study. Respir Med. 2009;103(2):309-16. PG, Grutters JC, Rijkers GT, et al. Microbial aetiology, O 11. Torres A, Peetermans WE, Viegi G, Blasi F. Risk factors outcomes, and costs of hospitalisation for community- for community-acquired pneumonia in adults in Europe: acquired pneumonia; an observational analysis. BMC a literature review. Thorax. 2013;68(11):1057-65. Infect Dis. 2014;14(1):335. 12. Rivero-Calle I, Pardo-Seco J, Aldaz P, Vargas DA, Mas- 29. ICD - ICD-9-CM - International Classification of Dis- carós E, Redondo E, Díaz-Maroto JL, Linares-Rufo M, eases. Ninth Revision. Clinical Modification. Cdc.gov. Fierro-Alacio MJ, Gil A, Molina J, Ocaña D, Martinón- 2017. Available from: www.cdc.gov/nchs/icd.htm. Torres F; on behalf of NEUMOEXPERTOS group. Inci- 30. Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider dence and risk factor prevalence of community-acquired P, et al. Updating and Validating the Charlson Comor- pneumonia in adults in primary care in Spain (NEUMO- bidity Index and Score for Risk Adjustment in Hospital ES-RISK project). BMC Infect Dis. 2016;16(1):645. Discharge Abstracts Using Data From 6 Countries. Am J 13. Ramirez JA, Wiemken TL, Peyrani P, Arnold FW, Kelley Epidemiol. 2011;173(6):676-82. R, Mattingly WA, et al. Adults hospitalized with pneumo- 31. Carugati M, Aliberti S, Reyes LF, Franco Sadud R, Irfan nia in the United States: Incidence, epidemiology, and M, Prat C, et al. Microbiological testing of adults hospi- mortality. Clin Infect Dis. 2017;65(11):1806-12. talised with community-acquired pneumonia: an interna- 14. Mizgerd JP. Acute lower respiratory tract infection. N tional study. ERJ Open Res. 2018;4(4):00096-2018. Engl J Med. 2008;358(7):716-27. 32. Blasi F, Garau J, Medina J, Ávila M, McBride K, Oster- 15. Jain S, Williams DJ, Arnold SR, Ampofo K, Bramley AM, mann H, et al. Current management of patients hospital- Reed C, et al. Community-acquired pneumonia requir- ized with community-acquired pneumonia across Europe: ing hospitalization among US children. N Engl J Med. outcomes from REACH. Respir Res. 2013;14(1):44. 2015;372(9):835-45. 33. Di Nardo F, Calabrò GE, Ianuale C, Poscia A, Azzolini 16. Fry AM. Trends in hospitalizations for pneumonia among E, Volpe M, et al. Capturing the chance for pneumo- persons aged 65 years or older in the United States, 1988- coccal vaccination in the hospital setting. Ann Ist Super 2002. JAMA. 2005;294(21):2712. Sanità. 2017;53:291-8. Available from: www.iss.it/binary/ 17. Van Gageldonk-Lafeber AB, Bogaerts MAH, Verheij RA, publ/cont/ANN_17_04_04.pdf. Van Der Sande MAB. Time trends in primary-care mor- 34. Uematsu H, Hashimoto H, Iwamoto T, Horiguchi H, Ya- bidity, hospitalization and mortality due to pneumonia. sunaga H. Impact of guideline-concordant microbiologi- Epidemiol Infect. 2009;137(10):1472-8. cal testing on outcomes of pneumonia. Int J Qual Health 18. Woodhead M, Blasi F, Ewig S, Garau J, Huchon G, Ieven Care. 2014;26(1):100-7. M, et al. Guidelines for the management of adult lower 35. Bradley JS, Byington CL, Shah SS, Alverson B, Carter respiratory tract infections - Full version. Clin Microbiol ER, Harrison C, et al. The management of community- Infect. 2011;17:E1-59. acquired pneumonia in infants and children older than 3 19. Wroe PC, Finkelstein JA, Ray GT, Linder JA, Johnson months of age: Clinical practice guidelines by the Pediat- KM, Rifas-Shiman S, et al. Aging population and fu- ric Infectious Diseases Society and the Infectious Diseases ture burden of pneumococcal pneumonia in the United Society of America. Clin Infect Dis. 2011;53(7):e25-76. States. J Infect Dis. 2012;205(10):1589-92. 36. Kalil AC, Metersky ML, Klompas M, Muscedere J, 20. Cilloniz C, Martin-Loeches I, Garcia-Vidal C, San Jose Sweeney DA, Palmer LB, et al. Management of adults A, Torres A. Microbial etiology of pneumonia: epidemi- with hospital-acquired and ventilator-associated pneumo- ology, diagnosis and resistance patterns. Int J Mol Sci. nia: 2016 Clinical Practice Guidelines by the Infectious 2016;17(12):2120. Diseases Society of America and the American Thoracic 21. Avdic E, Cushinotto LA, Hughes AH, Hansen AR, Efird Society. Clin Infect Dis. 2016;63(5):e61-111. LE, Bartlett JG, et al. Impact of an antimicrobial stew- 37. Saldías Peñafiel F, O’Brien Solar A, Gederlini Gollerino ardship intervention on shortening the duration of thera- A, Farías Gontupil G, Díaz Fuenzalida A. Neumonía py for community-acquired pneumonia. Clin Infect Dis. adquirida en la comunidad en el anciano inmunocom- 2012;54(11):1581-7. petente que requiere hospitalización. Cuadro clínico, 22. Viasus D, Vecino-Moreno M, De La Hoz JM, Carratalà factores pronósticos y tratamiento. Arch Bronconeumol. 284 Marianxhela Dajko, Andrea Poscia, Brunella Posteraro et al.

2003;39(8):333-4 L, Durán G, et al. Reduction in pediatric invasive pneu- 38. Serrano L, Zalacain R, Urrutia A, Garcia M, Gomez mococcal disease in the Basque Country and Navarre, A, Castro S, et al. Relationship between etiologic diag- Spain, after introduction of the heptavalent pneumococ- nosis and prognosis in community-acquired pneumo- cal conjugate vaccine. Eur J Clin Microbiol Infect Dis. nia. European Respiratory Journal. 2015;46:PA2591. 2007;26(5):303-10. Available from: http://erj.ersjournals.com/lookup/doi/ 41. Cantais A, Mory O, Pillet S, Verhoeven PO, Bonneau J, 10.1183/13993003.congress-2015.PA2591. doi: 10.1183/ Patural H, et al. Epidemiology and microbiological inves- 13993003.congress-2015.PA2591 tigations of community-acquired pneumonia in children 39. Ostermann H, Garau J, Medina J, Pascual E, McBride admitted at the emergency department of a university reviews

K, Blasi F. Resource use by patients hospitalized with hospital. J Clin Virol. 2014;60(4):402-7. community-acquired pneumonia in Europe: analysis of 42. Ruuskanen O, Lahti E, Jennings LC, Murdoch DR. Viral

and the REACH study. BMC Pulm Med. 2014;14(1):36. pneumonia. Lancet Lond Engl. 2011;377(9773):1264- 40. Aristegui J, Bernaola E, Pocheville I, García C, Arranz 75. articles

riginal O 1 Rome, Italy. E-mail:[email protected]. Address for correspondence:Roberta Pacifici, Centro Nazionale Dipendenzee Doping,Istituto Superioredi Sanità, Viale Regina Elena299, 00161 CBD-based foods,cosmetics and medicinals. regulatory lawsoverthemarketing authorizationof tention tothelackofharmonization ofinternational legal statusofCBDindifferent countries,drawingat- the benzenering[2]. distributed ontheetherandalkylgroups,aswell structures in whichthe unpaired electronsare mainly of thecationfreeradicalsshowseveralresonance 2013, CBDhaspotentialantioxidantactivitybecause two hydroxylgroups and accordingtoBorgesetal. CBD 1-en-2-yl group[2].Fromthechemicalstructureof 4-pentylphenyl group,andinposition4withaprop- by amethylgroup,inposition32,6-dihydroxy- cannabinol (Δ9-THC,itissubstitutedinposition1 medicinal CannabissativaL.,suchasΔ9-Tetrahydro- attention hasbeenpaidtothecompoundspresentin alkyl resorcinolwithamonoterpeneunit[6]alotof is ameroterpenoidobtainedfromthealkylationofan described byMechoulametal.in1963[1,5].CBD isolated forthefirsttimein1940,anditsstructurewas the mostprevalentinfibre-typehemp[4].Itwas pharmacologically activecomponentoftheplantand ent inCannabissativaL.[2-4].Itisthesecondmajor ol [1]isoneofthemore100phytocannabinoidspres- 3-methyl-2-cyclohexen-1-yl]-5-pentyl-1,3-benzene-di- INTRODUCTION marketing authorizationofCBD-basedproducts. of CBD,tohighlightthelackharmonizationinternationalregulatorylawsover atric patients. The aim of this review was to focus on pharmacology and on legal status severe formsofepilepsy, “Lennox-Gastautsyndrome”and“Dravetsyndrome”,inpedi- CBD oral solution with the trade name of Epidiolex® for the treatment of two rare and ers. Mostimportantly, on2018UnitedStatesFoodandDrugAdministrationapproved particularly medicinalproductscontainingCBD,whichareeasilyavailableforconsum- fields, havedeterminedanexponentialmarketinggrowthoffoods,cosmeticsandin nalization policies and the high applicability in therapeutic and technologic-industrial mimetic effectsproducedbytetrahydrocannabinol(THC).Thelatestcannabisdecrimi- L. Itisnotassociatedwithpsychotropicactivityandcapabletomitigatethepsychoto- Cannabidiol (CBD)isthesecondmostabundantcannabinoidpresentinCannabissativa Abstract 2 Pietro Brunetti of cannabidiol Pharmacology andlegalstatus DOI: 10.4415/ANN_20_03_06 Ann IstSuperSanità2020|Vol. 56,No.3:285-291 Roberta Pacifici Centro NazionaleDipendenzeeDoping,IstitutoSuperiorediSanità,Rome,Italy Dipartimento diScienzeBiomedicheeSanitàPubblica,UniversitàPolitecnicadelleMarche,Ancona,Italy In thisreview, we describedthepharmacologyand Cannabidiol (CBD)or2-[(6R)-6-Isopropenyl- (Table 1),itis easy torecognizethepresence of 1 , AlfredoFabrizioLoFaro 2 , SimonaPichini 2 andFrancescoPaoloBusardò 1 , FilippoPirani

The firstaimofthismanuscript wasfocusedon:1) only 29wereincludedforthe purposesofthisreview. RESULTS ANDDISCUSSION framework ofthecurrentreport. determine theirrelevanceandappropriatenessinthe were screenedindependentlybytwooftheauthorsto lish sources were taken into consideration. All sources whereas forinstitutionalwebsitesbothItalianandEng- Only scientificarticleswritteninEnglishwereincluded, searched aloneandinassociationtoeachoftheothers. latory. ThemainkeywordscannabidiolandCBDwere use, adverseeffects,reactions,toxicity, legalandregu- ogy, pharmacokinetics, pharmacodynamics, medical were: cannabidiol,CBD,phytochemistry, pharmacol- 2019). Thesearchtermsusedonlyorincombination identify themostrelevantliterature(uptoNovember Academies ofSciences,Engineering,andMedicineto cines Agency, ItalianPublicAdministration,National ropean Food Safety Authority (EFSA), Italian Medi- and DrugAddiction,EuropeanMedicinesAgency, Eu- tration (DEA),EuropeanMonitoringCentreforDrugs ministration (FDA),USDrugEnforcementAdminis- World Health Organization, US Food and Drug Ad- ternational agenciesorinstitutionalwebsitesincluding: plinary databasessuchasPubMed,Web ofScience,in- MATERIALS ANDMETHODS The initialsearchproducedabout 2800sources,but A literature search was performed on multidisci- 1 , PaoloBerretta 1

• • • • • Key words 2 , policies legal status law epilepsy (CBD) cannabidiol

285 Original articles and reviews 286 Pietro Brunetti, Alfredo Fabrizio Lo Faro, Filippo Pirani et al.

Table 1 Cannabidiol (CBD) and main metabolites in human

Molecular Structure Chemical names formula

reviews H O H CAS 13956-29-1 C21H30O2 Cannabidiol H and

H O

Main metabolites in human articles

Molecular Structure Chemical names CYP 450 isoform formula riginal O

C21H30O3 7-Hydroxy Cannabidiol CYP2C19 CYP3A4

C21H30O3 6a-Hydroxy- CYP2C19 CYP3A4 cannabidiol

C21H30O3 6 β-Hydroxy- CYP2C19 CYP3A4 cannabidiol

C27H38O8 Cannabidiol β-D- UGT1A7, UGT1A9, Glucuronide UGT2B7

C21H28O4 7-Nor-7- CYP2C19 CYP3A4 carboxycannabidiol

pharmacology and toxicology of CBD including phar- Pharmacokinetics macokinetics, pharmacodynamics, adverse effects and Pharmacokinetics processes of CBD are dynamic interactions, while the second aim was to describe 2) and may change over time, depending on the route of the legal status of CBD and to provide an overview administration and the frequency or magnitude of ex- about American and European legislation, with a spe- posure [7]. CBD is highly lipophilic and for that reason cial focus on the Italian scenario. can easily pass across biological barriers and it is rapidly 287 Cannabidiol (CBD) and legal status

absorbed. Inhalatory and oral administrations are the smokers could confirm or deny that possibility. Indeed, most common routes for drug formulation containing CBD pharmaceutical preparation are based on tablets CBD [7, 8]. Oral absorption seems to have more vari- and oily drops [17]. ability and less bioavailability compared to inhalatory one, probably due to an intensive first-pass metabolism Pharmacodynamics [9]. Distribution of CBD is time dependent and begins To date CBD mechanisms of action are not fully elu- upon absorption, into fatty tissues and highly perfused cidated. Thanks to its high lipophilicity CBD crosses organs such as: brain, heart, lung, and liver, quickly the blood-brain barrier (BBB) [4] modulating central reviews decreasing its plasma concentrations [7, 9]. Plasma nervous system (CNS) [1]. Most central CBD effects protein binding, primarily to lipoproteins, of CBD, is are mediated by the activation of endocannabinoid sys-

similar to the one of THC and is about 97%. For this tem through the modulation of cannabinoids (CBr) 1,2 and

reason, CBD intake can cause concentration increase receptors, which are Gi/o protein-coupled receptors [11, of co-administered drugs, displacing them from protein 18] present at a high density in the frontal cortex, basal linkage [10] and causing possible adverse effects. CBD ganglia, hippocampus, and cerebellum, and at a minor is extensively metabolized in liver and in intestine by density in the hypothalamus, nucleus accumbens, and articles different isoforms of cytochrome P450 (CYP) such as: amygdala [11, 19]. CB1r are located predominantly at CYP2C19 and CYP3A4 [1, 10, 11] (all main metabo- the presynaptic terminals of neurons while postsyn- lites are reported in Table 1), forming predominantly aptic localization has rarely been observed [11]. Their 7-hydroxy-cannabidiol (7-OH-CBD) and 7-carboxy- activation inhibits the presynaptic release of many neu- cannabidiol (7-COOHCBD), and 6-hydroxy-cannabi- rotransmitters such as: γ-aminobutyric acid (GABA), riginal diol as minor metabolite [1, 8]. The metabolic pathway glutamate, acetylcholine, serotonin, and noradrenaline O of CBD may involve other isoforms of CYP450 such [18]. CB2r are principally expressed in peripheral tissues as CYP1A2, -2B6, -2C8, -2C9, -2E1, -2J2, and -3A5/7 [18] and are associated with cells governing immune [1, 9]. The hydroxylation reactions occur furthermore function [19], although the receptorial density in brain at positions 1”-5” of the aliphatic pentyl- and position is very low and predominantly located at mesolimbic 10 on the propenyl- substituent [8, 11]. These me- DA (DopAminergic) system [18]. CBD possesses a tabolites may be further oxidized to form dihydroxyl- very low affinity for the CB1 and CB2 receptors [3, 20, ated metabolites and CBDoic acid derivatives [11]. 21] acting as negative allosteric modulator of the CB1r

At last UGT1A7, UGT1A9, and UGT2B7 isoforms [1] of [22] or inverse agonist of CB2r [23]. Moreover, it can 5’-diphosphoglucuronosyltransferase (UGT), which increase concentrations of anandamide through the in- catalyze glucuronidation of xenobiotics, create more hibition of fatty acid amide hydrolase (FAAH), its main easily excreted products [12]. Metabolites are primarily degradative enzyme [4, 22], and the blockade of its re- excreted with feces and in a minor amount (16% ca.) uptake, promoted by fatty acid-binding protein (FABP) [13] with urine, while a large proportion of CBD (33%) [4, 18, 22]. For that reason CBD can modulate, through is excreted unchanged in feces [8]. The slow release, presynaptic CB1rs, the release of certain neurotransmit- the redistribution phenomenon from deep lipid-storage ters in particular key brain zones. The dampen of neu- compartments and the significant enterohepatic circu- ronal excitability through the reduction of glutamate lation contribute to a long terminal half-life elimination release, indirectly protects against the development of of CBD, with the average amount post inhalation of cannabis use disorder (CUD ) [22] and attenuates psy- 31±4 hours and from 2 to 5 days after repeated daily chotomimetic and anxiogenic effects induced by high administrations in chronic cannabis users [7, 9]. In a doses of THC in humans. These observations suggested series of recent studies involving the smoking of “light that this cannabinoid could possess antipsychotic and cannabis” containing 0.16% THC and 5.8% CBD [14- anxiolytic properties [5, 18, 20, 22]. Experimentally, 16], the highest CBD concentrations in oral fluid (OF), CBD was successfully used in humans for reducing psy- serum and blood were observed on the samples col- chotic symptoms of schizophrenia, thanks to its partial lected 0.5 h after the start of smoking and the com- agonist activity on dopamine D2 receptors similarly to pound was measurable in those biological fluids up to atypical antipsychotics [20]. It also reduces anxiety and 4 hours after administration [14]. Following smoking stress symptoms [3, 20]. In this regard, the anxiolytic of four “light cannabis” cigarettes with a one h inter- effect exerted by CBD has been mainly related to its val between each cigarette, CBD concentrations in agonist activity towards serotonin type 1A (5HT1A) re- blood serum and OF overlapped with those obtained ceptors [3]. Botanical preparation containing THC acid after smoking a single cigarette, suggesting that CBD precursor (THCA-A) CBD acid precursor (CBDA) and is poorly absorbed after repeated smoking and that this cannabigerol acid (CBGA) were successfully used for is not the preferential route to administer it. As a jus- the treatment of neurodegenerative diseases such as tification to this result, it has to be said that this is the Huntington and possibly Alzheimer and Parkinson [3, first-time that a product containing negligible amounts 20]. These evidences suggest that acute administration of psychotropic THC and significant concentration of of CBD may reduce withdrawal symptoms of drugs de- CBD (58 mg per cigarette) was smoked in a controlled pendence and may also contribute to improve cogni- clinical trial, and that participants were cigarette smok- tive performances [3]. CBD is a promising therapeutic ers with some “light cannabis” experience. In fact, given agent approved for reducing seizures in many children that CBD does not preferentially volatilize compared with Dravet syndrome, a severe treatment-resistant to THC, a trial on more experienced “light cannabis” form of childhood epilepsy [24]. Analgesic myorelaxant 288 Pietro Brunetti, Alfredo Fabrizio Lo Faro, Filippo Pirani et al.

and antiepileptic actions of CBD are achieved through most recent evidence on cannabinoids effectiveness on the increasing of inhibitory tone in cortical and stria- sleep refer to CBD therapeutic potential for the treat- tal membranes obtained from the inhibition of GABA ment of insomnia [29]. This capability is due to its anx- reuptake and the positive allosteric modulation of GA- iolytic, antipsychotic and neuroprotective properties,

BAA receptor [25]. Thanks to its lack of psychoactivity, which prompted for its potential use in epilepsy, sub- CBD is one of the most interesting compounds poten- stance abuse and dependence, post-traumatic stress, tially useful in various models of pathologies such as depression, and finally sleep disorders [5]. However, inflammatory and autoimmune disorders like multiple this beneficial CBD feature should be considered when reviews

sclerosis, arthritis, and cancer [6]a lot of attention has the substance is used by healthy individuals, where this been paid to the compounds present in medicinal Can- anxiolytic and relaxing effects may be dangerous in nor-

and nabis sativa L., such as Δ 9 -Tetrahydrocannabinol (Δ mal daily activities such as driving or working.

9 -THC. The agonist activity on peroxisome prolif- erator-activated receptor gamma (PPARγ) [26], the Legal status in United States of America competitive inhibition of adenosine uptake [23], the As cannabis compound, CBD was reported in the antagonism over adenosine receptor A2A [27] and the schedule I of Controlled Substances Act (CSA) which articles

activation of Transient Receptor Potential Subfamily V prohibits “manufacture, distribution, or dispensation, member 1 and 2 (TRPV1,2) [21, 28], mediated by CBD as well as its possession with intent to manufacture, produce several changes. Namely, significant reduction distribute, or dispense” of all those substances, that in release of IL-2 (Interleukin), TNF-α (Tumor Necro- have a high potential for abuse or have no currently ac- riginal sis Factor), IFN-c (Interferon), IL-6, IL-12, IL-17, eo- cepted medical use in treatment in the United States O taxin-1 (CCL11) and COX-2 (Cyclooxygenase), and of America (US) [30]. On June 25th, 2018 FDA ap- iNOS (inducible Nitric Oxide Synthases) expression proved Epidiolex® (CBD) oral solution for the treat- [27], decreasing the inflammatory event and exerting ment of two rare and severe forms of epilepsy, Lennox- neuroprotective actions after hypoxic ischemic expo- Gastaut syndrome and Dravet syndrome, in pediatric sure [6] or decreasing the leukocyte infiltration in brain patients [31]. This is the first FDA-approved drug that in some autoimmune diseases, like multiple sclerosis contains a purified drug substance derived from mari- [27]. As regards cancer, CBD has exhibited antiprolif- juana, schedule I of CSA [32]. For that reason, FDA erative and proapoptotic activities through the antago- prepared and transmitted a report of CBD to justify the nism over G protein-coupled receptors (GPR) 55 [3], necessity of re-scheduling CBD, due to scientific evi- modulating the tumorigenesis in different types of can- dences of its therapeutic benefits. On September 27th, cer, including breast, lung, colon, brain, and others [6]. 2018 the Department of Justice and DEA announced that Epidiolex®, was placed in schedule V, as a “drug Adverse effects and interactions with lower potential for abuse than Schedule IV con- CBD is by definition: the major nonpsychoactive phy- taining limited quantities of certain narcotics”, the least tocannabinoid derived from cannabis [5] and, thanks to restrictive schedule of the CSA [33]. It is necessary to its good safety profile and the lack of euphoric effects, is underline that CSA did not distinguish between CBD the most interesting cannabinoid [4]. However, CBD is obtained from recreational, medical or industrial can- not a biologically inert compound and its complex phar- nabis (hemp). On December 20th, 2018, with the Agri- macology offers tremendous therapeutic potential but, culture Improvement Act, hemp was definitely removed also, the potential for adverse effects and drug-drug in- from schedule I and became legal, so any cannabinoid teractions. Pharmacodynamics and pharmacokinetics or derivatives, including CBD, obtained from hemp are interactions from cannabinoids may occur via simple legal, if and only if that hemp is produced under the fed- competitive inhibition, noncompetitive (allosteric) inhi- eral law, while all other cannabinoids, produced in any bition, or at the level of gene expression of the cannabi- other setting, remain illegal and scheduled in the first noid receptors, biotransformation enzymes, and trans- table of CSA [34], except for the limited circumstances port proteins [9]. The metabolic routes of cannabinoids e.g. a medically approved benefit. On June 16th 2019, primarily involve cytochrome P450 oxidases (CYP) en- the FDA released a document recognizing the signifi- zymes which are implicated in the primary metabolism cant public interest in cannabis-derived compounds, and biotransformation of the majority of therapeutic expressing, at the same time, concern for the exponen- agents and xenobiotics [12]. It is easy to understand tial increase of illegal products that violate the Federal that these compounds are susceptible to, or complicit Food, Drug and Cosmetic Act (FD&C Act) [35]. From as, inhibitors or inducers of these enzymes. The de- 2015 to 2019 FDA sent several warring letters to com- crease in the metabolic activity of individual CYPs can panies that commercialized products containing CBD increase the plasma levels of their substrates, and symp- because they were marketed as unapproved new drugs toms of toxicity could appear. In the opposite direction, adding CBD to food [36]. It is essential to underline increased CYP activity will decrease the efficacy of its that aside Epidiolex ®, no other CBD drug products substrates, and can lead to the failure of a therapy [10]. were FDA-approved and for that reason they cannot The above reported studies on “light cannabis” smoking be distributed or sold in interstate commerce [35]. showed that after smoking up to 232 mg CBD in four CBD products are excluded from the dietary supple- hours time caused sleepiness. It has to be said that few ment definition under section 201(ff) of the FD&C Act data exist in the international literature concerning the [35, 37]. Nonetheless, ingredients as hulled hemp seed, side effects of CBD intake in healthy individuals. The hemp seed proteins powder and hemp seed oil, that are 289 Cannabidiol (CBD) and legal status

derived from parts of the cannabis plant that do not are “safe for human health when used under normal or contain CBD, might be able to be marketed as dietary reasonably foreseeable conditions of use” in accordance supplements or be added to food [35]. In the first case, with articles 3 and 4 of the Chapter II of the Cosmetic this can happen only if dietary supplements comply re- Regulation (Regulation (EU) No 1223/2009) [44], as quirements related to Current Good Manufacturing to say: if they not contain any substances listed in the Practices and are correctly labeled to eliminate any pos- Annex II and are “obtained from cannabis, cannabis sible risk of illness or injury [37]. In the second case, resin, cannabis extracts and cannabis tinctures originat- only after approval by FDA or if CBD addition to food ing from the seeds and leaves that are not accompanied reviews is generally recognized as safe [35]. At last, cosmetics by the fruiting tops of the cannabis plant” [43]. There- containing CBD are not restricted by any regulation. fore, whereas US legal status of CBD is currently well

However, all ingredients must comply with all appli- defined throughout Federal States, EU one is not ho- and cable requirements and cannot be used if they cause mogeneous at moment, nor extensively regulated, with adulteration or misbranding of the product or could be some countries, such as Italy, with no legal measures injurious to users [37]. So it seems that the 2018 Farm at all. Bill, the primary agricultural and food policy tool of the articles federal government explicitly preserved FDA’s author- Legal status in Italy ity to regulate products other than medicines contain- Currently, in Italy, there are no laws that ban CBD ing cannabis or cannabis-derived compounds (CBD in and at moment CBD is not yet registered as a medici- particular) under the FD&C Act. On the other hand, nal. Due to this law vacancy, some hemp shops freely the Farm Bill has no effect on state-legal cannabis offer CBD products (oil, crystals, etc). riginal programs, in which cannabis and derived compounds O (CBD in particular) were legalized for medical and rec- CONCLUSIONS reational purposes, that are declaimed illegal under the Thanks to its safe therapeutic profile and its lack of federal law [38]. psychoactivity, CBD is one of the most interesting com- pounds, with a lot of reported pharmacological effects Legal status in European Union in different models of pathologies, from inflammatory In European Union (EU) there is no harmonized and neurodegenerative diseases, to epilepsy, autoim- law on CBD use. Indeed, currently the criminal or mune disorders like multiple sclerosis, arthritis, schizo- administrative response towards CBD use is under phrenia, cancer and many others. Even if pharmaceuti- the responsibility of each EU Member State. Actu- cal and therapeutic profile of CBD is deeply notorious, ally, medicinal products containing CBD, such as: Sa- its legal status, in different countries around the world, tivex® and Epidiolex® (since September 19th, 2019), isn’t clear and harmonized. The widespread growth of are authorized in many EU countries, and, in some of markets selling products containing CBD (medicinal them, under certain conditions, are reimbursed by the products, foods and cosmetics) is not only an “Ameri- national health insurance system [39, 40]. Any medici- can” reality, but also involves many Europeans Coun- nal product containing CBD placed for sale or distri- tries, and according to recent market research studies bution on the market in a Member State must require it will continue growing exponentially over the next few a Community marketing authorization, released by the years. Actual legislative acts of different governments European Agency for the Evaluation of EMA, in ac- to control drugs, foods and cosmetics containing CBD, cordance with the Title I article 3 of Regulation (EEC) which are easily available for consumers, are nowadays N. 2309/93 [41]. The marketing authorization shall be inappropriate to support the international and intrana- valid throughout the Community. On the contrary, the tional market without a systematic revision and could rejection of a Community marketing authorization shall cause an important public health threat. constitute a prohibition on the placing on the market of the medicinal product throughout the Community Funding in compliance with the Title II article 12 of Regulation The study was partly supported by Department for (EEC) N. 2309/93 [41]. In the past few years there has Anti-Drug Policies, Presidency of the Council of Min- been an increase in the availability of cannabis based isters of Italian Government. products (herb, hemp, oils) that contain CBD, referred as “light cannabis” [17]. All extracts of Cannabis sativa Acknowledgements L. and any product to which CBD, synthetically ob- The authors thank Michele Sciotti, Laura Martucci, tained or not, is added as ingredient, have to be con- Antonella Bacosi and Simonetta Di Carlo for technical sidered as novel foods and strictly controlled under the assistance. novel food Regulation ((EU) 2015/2283) [42]. As in case of medicinal products, novel foods require a com- Conflict of interest statement munity authorization for the suppling in EU Member None of the authors have any conflict of interest. States market which is released by the EFSA [42, 43]. Others Cannabis based products containing CBD may Received on 15 December 2019. be used in cosmetics placed on the EU market if they Accepted on 17 March 2020. 290 Pietro Brunetti, Alfredo Fabrizio Lo Faro, Filippo Pirani et al.

REFERENCES

1. Huestis MA, Solimini R, Pichini S, Pacifici R, Carlier following a single and repeated administration of “light J, Busardò FP. Cannabidiol adverse effects and toxicity. cannabis”. Clin Chem Lab Med. 2019 Apr 8. pii: /j/cclm. Curr Neuropharmacol. 2019;17(10):974-89. doi: 10.217 ahead-of-print/cclm-2019-0119/cclm-2019-0119.xml. 4/1570159X17666190603171901 doi: 10.1515/cclm-2019-0119 2. Tura M, Mandrioli M, Toschi TG. Preliminary study. 16. Pacifici R, Pichini S, Pellegrini M, Tittarelli R, Pantano Comparison of antioxidant activity of cannabidiol (CBD) F, Mannocchi G, Rotolo MC, Busardò FP. Determina-

reviews and α-tocopherol added to refined olive and sunflower tion of cannabinoids in oral fluid and urine of “light can- oils. Molecules. 2019;24(19):3485. doi:10.3390/mol- nabis” consumers: a pilot study. Clin Chem Lab Med. ecules24193485 19;57(2):238-43. doi: 10.1515/cclm-2018-0566 and

3. Mannucci C, Navarra M, Calapai F, Spagnolo EV, Bu- 17. European Monitoring Centre for Drugs and Drug Ad- sardò FP, De Cas R, Ippolito FM, Calapai G. Neuro- diction. Developments in the European cannabis mar- logical aspects of medical use of cannabidiol. CNS Neu- ket. EMCDDA Papers. 2019:1-19. Available from: rol Disord Drug Targets. 2017;16:541-53. doi: 10.2174/1 www.emcdda.europa.eu/system/files/publications/11391/ 871527316666170413114210 TDAU19001ENN.pdf. articles

4. Pisanti S, Malfitano AM, Ciaglia E, Lamberti A, Ranieri 18. Galaj E, Xi Z-X. Potential of cannabinoid receptor li- R, Cuomo G, et al. Cannabidiol. State of the art and new gands as treatment for substance use disorders. CNS challenges for therapeutic applications. Pharmacol Ther. Drugs. 2019;33(10):1001-30. doi: 10.1007/s40263-019- 2017;175:133-50. doi:10.1016/j.pharmthera.2017.02.041 00664-w

riginal 5. Crippa JA, Guimarães FS, Campos AC, Zuardi AW. 19. McPartland JM, Guy GW, Di Marzo V. Care and feeding

O Translational investigation of the therapeutic potential of of the endocannabinoid system. A systematic review of cannabidiol (CBD). Toward a new age. Front Immunol. potential clinical interventions that upregulate the endo- 2018;9:2009. doi: 10.3389/fimmu.2018.02009 cannabinoid system. PLoS One 2014;9(3):e89566. doi: 6. Pellati F, Borgonetti V, Brighenti V, Biagi M, Benvenuti S, 10.1371/journal.pone.0089566 Corsi L. Cannabis sativa L. and nonpsychoactive cannabi- 20. Mandolini GM, Lazzaretti M, Pigoni A, Oldani L, noids: their chemistry and role against oxidative stress, in- Delvecchio G, Brambilla P. Pharmacological proper- flammation, and cancer. Biomed Res Int. 2018:1691428. ties of cannabidiol in the treatment of psychiatric dis- doi: 10.1155/2018/1691428 orders. A critical overview. Epidemiol Psychiatr Sci. 7. Huestis MA. Human cannabinoid pharmacokinetics. 2018;27(4):327-35. doi: 10.1017/S2045796018000239 Chem Biodivers. 2007;4(8):1770-804. doi:10.1002/ 21. Kossakowski R, Schlicker E, Toczek M, Weresa J, Ma- cbdv.200790152 linowska B. Cannabidiol affects the bezold-jarisch reflex 8. Taylor L, Crockett J, Tayo B, Morrison G. A Phase 1, via TRPV1 and 5-HT3 receptors and has peripheral sym- Open-Label, Parallel-Group, Single-dose trial of the pathomimetic effects in spontaneously hypertensive and pharmacokinetics and safety of cannabidiol (CBD) in normotensive rats. Front Pharmacol. 2019;10:500. doi: subjects with mild to severe hepatic impairment. J Clin 10.3389/fphar.2019.00500 Pharmacol. 2019;59(8):1110-9. doi: 10.1002/jcph.1412 22. Hindocha C, Freeman TP, Schafer G, Gardner C, 9. Foster BC, Abramovici H, Harris CS. Cannabis and Bloomfield MAP, Bramon E, et al. Acute effects of can- cannabinoids: kinetics and interactions. Am J Med. nabinoids on addiction endophenotypes are moderated 2019;132(11):1266-70. doi:10.1016/j.amjmed.2019.05. by genes encoding the CB1 receptor and FAAH enzyme. 017 Addict Biol. 2019:e12762. doi: 10.1111/adb.12762 10. Socała K, Wyska E, Szafarz M, Nieoczym D, Wlaz´ P. Acute 23. Liou GI, Auchampach JA, Hillard CJ, Zhu G, Yousufzai effect of cannabidiol on the activity of various novel anti- B, Mian S, et al. Mediation of cannabidiol anti-inflamma- epileptic drugs in the maximal electroshock- and 6 Hz-in- tion in the retina by equilibrative nucleoside transporter duced seizures in mice. Pharmacodynamic and pharmaco- and A2A adenosine receptor. Investig Ophthalmol Vis kinetic studies. Neuropharmacology. 2019;158:107733. Sci. 2008;49(12):5526-31. doi: 10.1167/iovs.08-2196 doi: 10.1016/j.neuropharm.2019.107733 24. Bakas T, van Nieuwenhuijzen PS, Devenish SO, Mc- 11. Zendulka O, Dovrteˇlová G, Nosková K, Turjap M, Gregor IS, Arnold JC, Chebib M. The direct actions of Šulcová A, Hanuš L, et al. Cannabinoids and cytochrome cannabidiol and 2-arachidonoyl glycerol at GABAA re- P450 interactions. Curr Drug Metab. 2016;17(3):206- ceptors. Pharmacol Res. 2017;119:358-70. doi: 10.1016/j. 26. doi:10.2174/1389200217666151210142051 phrs.2017.02.022 12. Brown JD, Winterstein AG. Potential adverse drug 25. Osborne AL, Solowij N, Babic I, Lum JS, Newell KA, events and drug-drug interactions with medical and con- Huang XF, et al. Effect of cannabidiol on endocannabi- sumer cannabidiol (CBD) use. J Clin Med. 2019;8(7). noid, glutamatergic and GABAergic signalling markers doi:10.3390/jcm8070989 in male offspring of a maternal immune activation (poly 13. Grotenhermen F. Pharmacokinetics and pharma- I:C) model relevant to schizophrenia. Prog Neuro-Psy- codynamics of cannabinoids. Clin Pharmacokinet. chopharmacol Biol Psychiatry. 2019;95:109666. doi: 2003;42(4):327-60. doi:10.1111/j.1365-2044.1982. 10.1016/j.pnpbp.2019.109666 tb01148.x 26. Nadal X, del Río C, Casano S, Palomares B, Ferreiro- 14. Pichini S, Mannocchi G, Berretta P, Zaami S, Pirani F, Vera C, Navarrete C, et al. Tetrahydrocannabinolic acid Pacifici R, Busardò FP. Δ9-tetrahydrocannabinol and is a potent PPAR agonist with neuroprotective activity. cannabidiol time courses in the sera of “Light Cannabis” Br J Pharmacol. 2017;174(23):4263-76. doi: 10.1111/ smokers: discriminating light cannabis use from illegal bph.14019 and medical cannabis use. Ther Drug Monit. 2019 Aug 27. Burstein S. Cannabidiol (CBD) and its analogs. A review 1. doi: 10.1097/FTD.0000000000000683 of their effects on inflammation. Bioorganic Med Chem. 15. Pacifici R, Pichini S, Pellegrini M, Rotolo MC, Gior- 2015;23(7):1377-85. doi: 10.1016/j.bmc.2015.01.059 getti R, Tagliabracci A, Busardò FP, Huestis MA. THC 28. Pumroy RA, Samanta A, Liu Y, Hughes TET, Zhao S, and CBD concentrations in blood, oral fluid and urine Yudin Y, et al. Molecular mechanism of TRPV2 channel 291 Cannabidiol (CBD) and legal status

modulation by cannabidiol. BioRxiv. 2019;8:48792. doi: from: www.fda.gov/news-events/public-health-focus/warn- 10.7554/eLife.48792 ing-letters-and-test-results-cannabidiol-related-products. 29. Babson KA, Sottile J, Morabito D. Cannabis, cannabi- 37. Food and Drug Administration. Federal Food, Drug, and noids, and sleep: a review of the literature. Curr Psychia- Cosmetic Act, Pubblic Law 116-22, June 24, 2019 2019. try Rep. 2017;19(4):23. doi: 10.1007/s11920-017-0775-9 Available from: https://legcounsel.house.gov/Comps/Fed- 30. US Department of Justice. Title 21 Code of Federal eral Food, Drug, And Cosmetic Act.pdf. Regulations PART 1308 – Schedules of Controlled Sub- 38. Hudak J. The Farm Bill, hemp legalization and the sta- stances §1308.11 Schedule I. Available from: www.deadi- tus of CBD. An explainer. Brookings Inst. 2018. Avail- version.usdoj.gov/21cfr/cfr/1308/1308_11.htm. able fro: www.brookings.edu/blog/fixgov/2018/12/14/the- reviews

31. US Food and Drug Administration. FDA News Release. farm-bill-hemp-and-cbd-explainer/. FDA approves first drug comprised of an active ingre- 39. European Monitoring Centre for Drugs and Drug Ad-

dient derived from marijuana to treat rare, severe forms diction. Cannabis legislation in Europe 2017. Available and of epilepsy | FDA. Available from: www.fda.gov/news- from: www.emcdda.europa.eu/system/files/publications/ events/press-announcements/fda-approves-first-drug- 4135/TD0217210ENN.pdf. comprised-active-ingredient-derived-marijuana-treat- 40. European Monitoring Centre for Drugs and Drug Ad- rare-severe-forms. diction. Medical use of cannabis and cannabinoids: ques-

32. US Department of Justice. Title 21 United States Code tions and answers for policymaking n.d. Available from: articles (USC) Controlled Substances Act, Subchapter I Control www.emcdda.europa.eu/system/files/publications/10171/ and Enforcement, Part A Introductory Provisions, §802. 20185584_TD0618186ENN_PDF.pdf. Definitions n.d. Available from: www.deadiversion.usdoj. 41. The Council of the European Communities. Council Regu- gov/21cfr/21usc/802.htm. lation (EEC) No 2309/93 of 22 July 1993 n.d. Available from: 33. Food and Drug Administration. FDA-approved drug https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/ riginal Epidiolex placed in schedule V of Controlled Sub- ?uri=CELEX:31993R2309&from=EN. O stance Act. Available from: www.dea.gov/press-releas- 42. The European Parliament and the Council of the Europe- es/2018/09/27/fda-approved-drug-epidiolex-placed- an Union. Regulation (EU) 2015/2283 of 25 November schedule-v-controlled-substance-act. 2015 on novel foods n.d. Available from: https://eur-lex. 34. Congress of the United States of America. Agricul- europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:320 ture Improvement Act Pubblic Law 115-334-Dec. 20, 15R2283&from=en. 2018 n.d. Available from: www.congress.gov/115/plaws/ 43. Norwinski EJ. EU Regulation of CBD in Foods and Cos- publ334/PLAW-115publ334.pdf. metics | Publications and Presentations | Arnold&Porter 35. Food and Drug Administration. FDA Regulation of can- n.d. Available from: www.arnoldporter.com/en/perspec- nabis and cannabis-derived products, including canna- tives/publications/2019/04/eu-regulation-of-cannabidiol- bidiol (CBD) | FDA n.d. Available from: www.fda.gov/ in-foods. news-events/public-health-focus/fda-regulation-canna- 44. The European Parliament and the Council of the Euro- bis-and-cannabis-derived-products-including-cannabidi- pean Union. Regulation (EC) N. 1223/2009 of 30 No- ol-cbd. vember 2009 on cosmetic products 2009. Available from: 36. Food and Drug Administration. Warning letters and test re- https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?ur sults for cannabidiol-related products | FDA n.d. Available i=CELEX:32009R1223&from=EN. 292 Original articles and reviews Lucia Miligi Enrico Oddone Elisa Romeo Dario Mirabelli industrial sectorsshowedincreased mortalityfrompleuralmalignancies,andmostalso the plantandperiodlevels.Asbestos relatedmortalitywassignificantlyincreased.All (cause knownfor95%)and1.5% losttofollow-up.Asbestosexposurewasestimatedat Results. Thestudyincluded51801 subjects (5741women):55.9%alive,42.6%died sex, regionandcalendarperiod. trial sectorforthe1970-2010period,majorcauses, usingreferenceratesbyage, ing, glasswork,harbors,insulationandotherindustries).SMRs werecomputedbyindus- Methods. Pool of 43 Italian asbestos cohorts (asbestos cement, rolling stock, shipbuild- workers. until the1992ban.We presentapooledcohortstudyonlong-termmortality in exposed Objective. Italyhasbeenalargeuserofasbestosandcontaining materials Abstract 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Corrado Magnani sector andcumulativeexposure asbestos relatedmortalitybyindustrial Italian poolofasbestosworkerscohorts: DOI: 10.4415/ANN_20_03_07 Ann IstSuperSanità2020|Vol. 56,No.3:292-302 28100 Novara,Italy. E-mail:[email protected]. Address forcorrespondence : CorradoMagnani, DipartimentodiMedicinaTraslazionale, UniversitàdelPiemonteOrientale, Novara, Via Solaroli17, Danila Azzolina Daniela Ferrante di lavoratoridell’Amianto(SMICA)* Stefano Mattioli (SMICA) arelistedbeforetheReferences *The membersoftheWorking GroupStudioMulticentricoItaliano diCoortilavoratoridell’Amianto l’Assicurazione controgliInfortunisulLavoro(INAIL),Rome,Italy Università degliStudidiTorino, Turin, Italy Dipartimento diSanitàPubblica,MedicinaSperimentaleeForense,UniversitàdegliStudiPavia,Italy Istituto NazionaleTumori IRCCSFondazioneG.Pascale,Naples,Italy formerly ARPAE Emilia-Romagna,UfficioProvincialediReggioEmilia,Italy Servizio diEpidemiologia,AziendaUnitàSanitariaLocale,IRCCSReggioEmilia,Italy Dipartimento diEpidemiologiadelServizioSanitarioRegionale,ASLRM1,Rome,Italy Dipartimento diSanitàPubblica,AziendaSanitariaLocale,Bologna,Italy Registro MesoteliomidelVeneto, SistemaEpidemiologico Regionale,ASL6,Padua,Italy Istituto perloStudio,laPrevenzioneeReteOncologica(ISPRO),Florence,Italy Dipartimento diMedicinaTraslazionale, UniversitàdegliStudidelPiemonteOrientale,Novara,Italy Dipartimento diMedicina,Epidemiologia,IgienedelLavoroeAmbientale,IstitutoNazionaleper Dipartimento diBiologiaeBiotecnologie“CharlesDarwin”,LaSapienzaUniversitàRoma,Rome,Italy Centro InterdipartimentaleperloStudiodegliAmiantiedialtriParticolatiNocivi“GiovanniScansetti”, Unità diEpidemiologiadeiTumori, UniversitàdegliStudidiTorino, Turin, Italy Dipartimento InterdisciplinarediMedicine,UniversitàdegliStudiBari,Italy Unità diMedicinadelLavoro,AziendaOspedaliero-UniversitariaCareggi,Florence,Italy Dipartimento diScienzeMedicheeChirurgiche,UniversitàdegliStudiBologna,Italy Unità diEpidemiologiaeStatistica,SanitariaLocaleBarletta-Andria-Trani, Barletta,Italy Dipartimento Prevenzione,ASURMarche,Senigallia,Italy 2 5 , GiuseppeGorini , FerdinandoLuberto 9 15,16 , SaraTunesi 1 12 , Tiziana Cena 1 , AntonioBaldassarre , fortheWorking GroupStudioMulticentricoItalianodiCoorti 1,16 , RobertaPirastu , StefanoSilvestri 1 , PatriziaPerticaroli 2 1 , Vittoria Bressan , ElisabettaChellini 6 , OriettaSala 17 , AlessandroMarinaccio 1 , AlessiaAngelini 13 , FrancescoBarone-Adesi 7 3 , CorradoScarnato , PaoloGirardi 10 , AldoPettinari 2 , EnzoMerler 1 , AlessandraRanucci 18 , StefaniaMassari 3 , LisaBauleo 3 1 , Venere Pavone , MarinaMusti 10 , FrancescoCuccaro 4 , SimonaMenegozzo

• • • • • Key words mesothelioma shipyards rolling stock glassworks asbestos 1

, 5 , 18 , 14 4 , ,

11 , 8 ,

293 Italian asbestos workers cohorts

from peritoneal and lung cancer and asbestosis, with exposure related trend. Increased mortality was also observed for ovarian cancer and for bladder cancer. Discussion. The study confirmed the increased risk for cancer of the lung, ovary, pleura and peritoneum but not of the larynx and the digestive tract. A large increase in mortality from asbestosis was observed. reviews

INTRODUCTION [12]. The present study investigated systematically mor- Italy has been an important producer and user of as- tality by industrial sector and cumulative exposure.

bestos and asbestos containing products. The consump- and tion of asbestos was estimated nationwide in 132 358 MATERIAL AND METHODS tons in 1970, gradually increasing to the 1980 peak The pooled cohort study aimed at the inclusion of all of 180 528 tons, and declining afterwards [1]. The Italian cohorts of workers in asbestos using factories. largest use of asbestos was in asbestos-cement pro- The candidate list was formed after a literature review articles duction, followed by fire proofing and thermal insula- also including unpublished reports and the personal ex- tion in shipbuilding and railway carriages. EU Direc- perience of participants. Eligible cohorts were identi- tives on permissible occupational exposure levels and fied, principal investigators invited, and when accepted on the limitation of asbestos use in some applications they were invited to share the protocol and to submit and products had been gradually enforced (Council Di- it to the ethical committee of competence. Cohort riginal rective 83/477/EEC of 19 September 1983). The use data were then updated and pooled. Participation was O of asbestos definitely ceased following a law-enforced restricted to cohorts with at least one follow-up inves- ban on production, import, export, and trading (Law tigation completed in the past and, once updated, an 257/1992) that become fully effective in 1994. How- observation period longer than 40 years. The final pool ever, asbestos in place is only being gradually removed included 43 cohorts, including a cohort of women with [2] and over 70 000 exposed were estimated as for the domestic exposure and a cohort of Italian crocidolite period 2000-2003, in the construction and in the asbes- miners in Australia. Supplementary Table 1 (available tos removal sectors [3, 4]. online) lists the cohorts with information on location, The burden of asbestos related deaths include malig- use of asbestos, number of workers, and references to nant mesothelioma, cancers of the lung, the larynx and previous studies. of the ovary, and asbestosis. Limited evidence of car- The study was submitted to the University of East- cinogenicity also exist for pharynx, stomach and colon ern Ethical Review Board (Authorization and rectum neoplasms [5]. CE 112/13, July 12th, 2013) and to the corresponding Current production worldwide is reduced to about Boards of participating institutions. Only anonymized 1.1 million tons, and is limited to chrysotile [6]. Its use data were pooled, nominal data remaining at the local is concentrated in emerging economy countries, where study level. information on work conditions, number of exposed The initial size included 54 436 subjects but quality workers and frequency of asbestos related diseases are control led to the exclusion of 2453 records (4.5%), limited [7, 8]. from all industrial sectors. Causes were: conflicting The present study is part of a large project aiming dates, incomplete working periods, hiring or retirement at the evaluation of the occurrence of asbestos related age out of normal range (n = 737); first employment deaths in the main industrial sectors interested by the after the asbestos ban, (fixed on 1.1.1993 – midterm use of asbestos in Italy. The study has multiple aims: between law approval and enforcement) (n = 594). Two to evaluate the burden of asbestos related diseases, to cohorts, from the asbestos cement factories of Eter- assess the occurrence of cancers with limited evidence nit-Bagnoli and of Fibronit-Broni, were limited to the of association with asbestos, and to investigate the role workers hired after 1.1.1950 (n = 1122) to ensure com- of time-related factors in disease occurrence. The study pleteness of follow-up. design consisted in the updating and pooling a large Workers employed in different cohorts were identi- number of Italian cohort studies. It started as part of fied, for a total of 178 workers of which 4 with 3 employ- the “Asbestos Project” coordinated by the Italian Na- ment records each. Their work histories were merged tional Institute of Health (ISS) [9], as prompted by the in the pooled analyses, for a total of 51 801 individual conclusions of the 2nd Governmental Asbestos Confer- records, and were left separate in the analyses of indi- ence [10] and was later extended as a part of the Asbes- vidual cohorts. The few (n. 47) workers employed in tos Special Program of the Italian Workers’ Compensa- different sectors were excluded from the sector-specific tion Authority (INAIL). analyses. A first report described the pooled cohort, providing For each factory and time period, the study investiga- overall results on cause-specific mortality and mortality tors had provided all available information, in particular time trends using external references [11]. A second re- estimates of the proportion of workers directly and in- port described the methodology and results of the esti- directly exposed to asbestos, the percentage of working mation of asbestos exposure in the cohorts under study time in tasks with asbestos exposure and the minimum and presented results on exposure levels and mortality and maximum levels of asbestos concentration for direct by cumulative exposure in the asbestos-cement sector and indirect exposures. Two expert industrial hygienists 294 Corrado Magnani, Stefano Silvestri, Alessia Angelini et al.

(AA and SS) collected and evaluated for each plant and A fibre-type-weighted-Cumulative Exposure Index year of activity the information available on the work (fibre-type-weighted-CEI) was computed for each process, the plant layout, the use of asbestos in total and worker summing the fibre-type-weighted AEI over the by asbestos type, and the measurements of asbestos fi- entire period of activity. bres, as well as the information provided by investiga- tors. Data sources included published and unpublished fibre-type-weighted-CEI = fibre-type-weighted-AEI ∑ py reports. Company reports, exposure surveys, judicial py examinations, and narrative reports from workers [13] reviews

were used. Company-specific data were checked against Sensitivity analyses were conducted using factors de- the evidence from other factories with similar activity rived by other authors [15-17]. The same weights were

and included in our pooled study and from literature data to used also in the analyses for other Asbestos Related

identify inconsistencies (i.e. differences not explained by Diseases (ARDs). plant-specific characteristics) and to fill gaps in the data The AEI dimension was a concentration (fibres/ml) (e.g. estimating asbestos concentration). and the dimension of the fibre-type-weighted-AEI The experts estimated for each plant and year the was the equivalent concentration of chrysotile asbes- articles

proportion of exposed workers, the percentage of time tos fibres (fibres of chrysotile/ml). CEI and fibre-type- in asbestos exposing tasks and the minimum and maxi- weighted-CEI had the dimension of concentration mum concentrations of asbestos airborne fibres (f/ml, times years (f/ml × year), the latter being the equivalent from data measured in optical microscopy), for direct concentration of chrysotile asbestos fibres times years. riginal and indirect exposure separately. Supplementary Table 2 (available online) presents the ex- O As the cohort information did not include tasks and posure indices by sector and period of activity. jobs of individual workers, only plant and period-spe- Follow-up, and ascertainment of the causes of death, cific estimates were generated. An Average Exposure were carried out by each research unit, using agreed Index (AEI) was computed for each plant and year and procedures already tested in previous studies [18]. The it was applied to all members of the corresponding co- Registrar’s Offices of the town of residence provided hort. It summarized in one value the range of concen- the information on vital status. The causes of death tration for each plant and period. The geometric mean were provided by the Local Health Authority Registries between minimum and maximum levels, adjusted for of Causes of Death for deaths after 1985 and by the the average proportion of time in tasks with asbestos Registrar Office of the municipality of death for earlier exposure, was first calculated separately for direct and years. The underlying cause of death was coded accord- indirect exposures. The AEI value for each plant and ing to the ICD, 8th, 9th, and 10th Revisions, according to year was obtained as the weighted average of direct the date of death. The date of follow-up depended on and indirect exposures, weights being the respective the available update of files but it was required to be at proportional size of the workforce, as in the following least 31/12/2010. Limited to the regions where files of formula: residents and of causes of death are kept at the regional level ( for both, Veneto and Emilia-Romagna AEIpy= (Edpy * wdpy+Eipy * wipy) for causes of death), similar procedures were applied to the regional files. Each research unit forwarded to the where E = exposure geometric mean, w = propor- study coordination the anonymous database for each tional size of the workforce and d = direct, i= indirect cohort, with sex, date of birth, vital status and date of exposure, p = plant, and y = year follow-up, cause of death for decedents and dates of start and finish of each period of employment. From the AEI a Cumulative Average Exposure In- Statistical analyses were based on person-years (p-y) dex (CEI) was computed for the occupational history and standardized mortality ratios (SMRs; i.e. the ratio of each worker summing the contribution of all periods of observed to expected deaths using indirect standard- of activity: ization) [19]. Subjects contributed person-years up to their most recent date of observation. Duration of ex- CEI = AEI posure was computed by summing up the duration of ∑ py py all employment periods in the cohort. TSFE (latency) was computed from the date of first employment. Cu- A fibre-type-weighted-AEI was computed tak- mulative exposure was computed as described before.

ing into account the proportion of chrysotile (CHpy), Numerical variables were analyzed in classes; tertiles or

amosite (Apy), and crocidolite (CRpy) used in each plant other percentiles were defined on the basis of the cu- and year. The weights were the MM potency factors mulative distribution by industrial sector. Duration of estimated by Hodgson and Darnton for chrysotile, exposure was calculated by summing up all the work amosite and crocidolite (respective 1:14:71) [14]. The periods since the date of first employment. TSFE was fibre-type-weighted-AEI corresponds to the chrysotile calculated from the date of first employment until the equivalent asbestos concentration in fibres per ml. It most recent date of observation. was computed as: Reference rates were age-, period-, sex-, region- and cause-specific. Regional mortality rates were used, ac-

fibre-type-weighted AEIpy = AEIpy * (1*CHpy + 14 * Apy cording to the region of location of each plant. The set

+ 71 * CRpy) of rates was prepared by the ISS, using mortality and 295 Italian asbestos workers cohorts

population figures provided by the National Institute of computed according to the Poisson distribution of ob- Statistics (ISTAT) for years from 1970 on [20]. Present served deaths, at the 95% confidence value (95% CI) analyses were therefore limited to p-y and events occur- [19]. Data were prepared using MS Access and SAS ring after January 1st 1970. 9.2. Analyses were carried out using OCMAP plus, We computed SMRs for the major causes of death. STATA 11 and SAS 9.2. SMRs were stratified by gender, industrial sector, and a priori defined classes of calendar time and cumula- RESULTS tive exposure. Present report is focused on mortality by Table 1 provides some descriptive information of the reviews industrial sector and cumulative exposure. We included pooled cohort under study. It included 51 801 persons the causes of death associated with asbestos following (89% men and 11% women). The industrial activities

IARC evaluation, namely: pleural and peritoneal ma- were: asbestos-cement (13 076 workers); rolling stock and lignant neoplasm or MM, cancers of lung, larynx and construction and maintenance (23 810 in total; 12 789 ovary and asbestosis, and those with limited evidence of in private plants and 11 021 in the Italian Railways, association [5]. Respiratory and cardiovascular diseas- of which 2626 in the Major Maintenance Workshops es were included, as relevant for the evaluation of the - OGR), shipyards (5120) and ship furnishing (1170), articles

Healthy Worker Effect (HWE) [19]. The list of causes glassworks (3727), dockyards (1939), insulation (205), was decided a priori. asphalt rolls (413) and ovens construction (217). The Throughout the paper the number of observed and cohort also included a cohort of Italian miners in Witte- expected cases are abbreviated as Obs and Exp. SMRs noom and a cohort of asbestos-cement workers’ wives, are computed in percent, as (Obs/Exp) × 100. For with domestic exposure, that were not included in the riginal consistency, SMRs from other studies were reported analyses by industrial sector. To avoid possible confu- O in the discussion using the percent scale. Statistical sion a limited number (47) of workers active in different significance was set at 5%. Confidence intervals were sectors were not included in any of the analyses by in-

Table 1 Italian pool of asbestos workers cohorts. Description of the cohort Men Women Total

n % n % n % p-ya Industrial Asbestos-cement 10 714 23.3 2362 41.1 13 076 25.2 388 915 activity Rolling stock constr. maint. 23 099 50.1 711 12.4 23 810 46.0 755 034 Shipyards 5099 11.1 21 0.4 5,120 9.9 172 583 Glassworks 2966 6.4 761 13.2 3727 7.2 105 446 Insulation 205 0.4 - - 205 0.4 6482 Ship furniture 1150 2.5 20 0.3 1170 2.3 36 957 Dockyards and harbours 1938 4.2 1 0.02 1939 3.7 62 102 Asphalt rolls production 341 0.7 72 1.2 413 0.8 14 429 Industrial ovens const. 202 0.4 15 0.3 217 0.4 7107 Crocidolite miners 299 0.6 1 0.02 300 0.6 9314 Domestic exposure - - 1777 30.9 1777 3.4 55 658 Wks in multiple sectors 47 0.1 - - 47 0.1 1626 Status at alive 25 977 56.4 3010 52.4 28 987 55.9 - follow-up deceased b, c 19 394 42.1 2651 46.2 22 045 42.6 -

emigrated c 172 0.4 31 0.5 203 0.4 - Lost to follow-up 517 1.1 49 0.9 566 1.1 - Year of first <= 1949 6649 14.4 1514 26.4 8163 15.7 169 669 exposure 1950-1959 6647 14.4 1517 26.4 8164 15.8 247 211 1960-1969 13 896 30.2 1295 22.6 15 191 29.3 538 718 1970-1979 13 033 28.3 839 14.6 13 872 26.8 488 420 1980-1989 5461 11.9 553 9.6 6014 11.6 163 752 1990-1992 374 0.8 23 0.4 397 0.8 7883 Total 46 060 100 5741 100 51 801 100.0 1 615 653 ap-y computed from 1970; b1092 causes of death unknown (960 men and 132 women, in both sexes 5% of decedents); cBefore 1970: 1172 deaths (1024 men and 148 women), 32 emigrated (25 men and 7 women), 230 lost to follow-up (211 men and 19 women). 296 Corrado Magnani, Stefano Silvestri, Alessia Angelini et al.

dustrial sector. The number of workers in the insulation, were of special interest, as they closely estimate the oc- asphalt rolls and ovens construction sectors is small, but currence of mesothelioma. An increase in the SMRs for it was decided to include these sectors given the rel- pleural neoplasms, with exposure response trends was evance of the exposure and the absence of larger Ital- observed for the asbestos cement, railway rolling stock, ian cohorts. Follow-up was known for 98.5% of workers shipyards, ship furniture, glasswork. Increased SMRs (98.5% for men and 98.6% for women), the remaining were also observed in the Insulation and Industrial Ov- being lost or untraced after abroad migration. Overall ens sectors, accompanied by a statistically significant 42.6% of cohort members were dead: 42.1% among increased SMR in the highest exposure tertile, and in reviews

men and 46.2% among women. The cause of death was Dockyards, where an overall statistically significant in- known for 95.0% of decedents, with the same percent- crease was present but not an increasing trend. How-

and age in both sexes. ever, among Dockyard workers, the contrast in expo-

Supplementary Table 2 (available online) presents the sure was rather small (Supplementary Table 3) (available distribution of the exposure indices in the plants includ- online). ed in the pooled cohort study, summarized by industrial A statistically significant increase in mortality for lung sector and period. For the asbestos cement sector, more cancer was observed in most industrial sectors, in par- articles

details can be found in the specific analyses presented ticular in asbestos cement. Railway rolling stock work- by Luberto et al. [12]: the highest exposure in this sec- ers showed an increasing trend (statistically significant) tor was observed before 1974, and declined sharply af- and a statistically significant increase in the SMR for ter 1980. The decennium 1970-1979 was a period of the highest tertile. Glassworks also showed a statisti- riginal transition, with high exposures at the beginning and cally significant trend for lung cancer and an increased O a reduction in the last quinquennium, as observed in SMR in the highest category. Shipyards showed a sta- particular with consideration of the range and of the tistically significant SMR in the highest category. The median values of the different industrial sectors. Insulation sector also showed a suggestion for an in- SMR analyses were limited to subjects contributing creasing trend in SMRs for lung cancer. person-years after 1/1/1970, excluding therefore 1172 No industrial sector showed an increase in mortal- decedents (1024 men and 148 women), 32 emigrated ity for laryngeal cancer, confirming the overall results (25 men and 7 women) and 230 lost to follow-up (211 presented in Table 2, that did not show an increase in men and 19 women) before 1970. mortality for this neoplasm in this pooled cohort. Table 2 presents mortality by cause of death in the Mortality from asbestosis was significantly increased pooled cohort, with observed and expected deaths, and in the sectors of asbestos cement, railway rolling stock SMRs with 95% CI, by gender. Total mortality was in- and ship furniture, while uncertain results were ob- creased in both genders, with 1183 excess deaths, cor- served for the shipyard, industrial ovens and insulation responding to a 6% increase in mortality. Causes that industrial sectors, and no cases were observed in the showed a statistically significant increase in mortality in remaining sectors. both sexes were: all cancers, respiratory tract cancers, These occupational cohorts also showed some in- lung cancers, pleural and peritoneal malignancies, blad- creases in mortality from diseases associated to other der cancers, respiratory diseases, and asbestosis. Wom- specific occupational risk factors. Shipyard workers en also showed an increase for ovarian cancers and men showed an excess (5 cases vs 2.95 expected; SMR: for malignant neoplasms of unspecified site. The num- 169, 95% CI 55-396) of deaths from pneumoconiosis ber of deaths from asbestosis was in great excess in both not due to asbestos exposure and a similar excess was genders; an excess was also shown for the “other pneu- also observed for the ship furniture (11 vs 1.68, SMR: moconiosis” category, with 89 deaths observed (vs 49.16 655, 95% CI 327-1172) and the glassworks (25 vs 8.64, expected) in men and 2 (vs 0.14) in women. Mortality SMR: 289, 95% CI 187-427) sectors. did not show a statistically significant increase for laryn- Workers in the ship furniture and in the insulation geal cancer, cancers of the digestive tract or pharyngeal sectors presented an increase of deaths from malignant cancer. A statistically significant reduction in mortality neoplasm of the nose and paranasal sinuses: 3 observed was observed for neurological, cardiovascular, digestive vs 0.20 expected deaths (SMR: 1528, 95% CI 315-4466) and genitourinary diseases in men. Deaths from un- and 3 observed vs 0.03 expected deaths (SMR:10179, specified causes represented 1.5% of total deaths. 95% CI 2099- 29 747), respectively. Class (tertiles) limits for CEI and for fibre-type- Dockyard workers also showed a statistically signifi- weighted-CEI by industrial sector are presented in the cant increase in mortality from bladder cancer (17 obs Supplementary Table 3 (available online). vs 9.09 exp; SMR:187, 95% CI 109-299), that is not Supplementary Table 4 (also available online) (nine a priori associated to asbestos exposure but it is likely panels numbered from 4.1 to 4.9) presents the sum- associated to some exposures collinear with it, as the mary report of mortality by cause by industrial sector excess increased over the tertiles of cumulative asbestos in men. Some of the sectors, in particular asbestos ce- exposure. ment, ship furniture, glassworks, dockyards, insulation Analyses for women are limited to the industrial sec- and industrial ovens, showed an increase in total mor- tors where female occupation was large enough to have tality. Workers in the railway rolling stock construction meaningful results, namely the asbestos cement and the and maintenance, on the contrary showed a statistically glasswork sectors, and are presented in the Supplemen- significant reduction in total mortality. tary Table 5 (available online) (two panels). Malignant neoplasms of the pleura and peritoneum In both sectors, women showed a statistically signifi- 297 Italian asbestos workers cohorts

Table 2 Italian pool of asbestos workers cohorts. Observed (OBS) and expected (EXP) deaths, by gender and cause of death Men Women Causes of death OBS EXP SMR 95% CI OBS EXP SMR 95% CI All causes 18370 17551.8 105** 103 106 2503 2138.0 117** 112 122 Malignant neoplasm (MN) 7361 6293.7 117** 114 120 818 612.7 133** 124 143 MN lip, oral cavity and pharynx 149 191.5 78** 66 91 9 6.6 137 62 259 reviews MN digestive organs (incl 2198 2194.5 100 96 104 262 226.9 116* 102 130 peritoneum) and MN stomach 523 575.2 91* 83 99 44 47.9 92 67 123 MN small intestine 14 10.8 130 71 218 1 1.2 84 2 468 MN colon 408 413.2 99 89 109 62 52.8 117 90 150

MN rectum 173 180.4 96 82 111 22 20.3 108 68 164 articles

MN of liver and intrahepatic bile 378 380.4 99 90 110 25 28.9 87 56 128 ducts MN peritoneum 136 28.5 477** 400 564 35 5.2 675** 470 939 riginal MN respiratory organs 3207 2155.3 149** 144 154 217 62.6 347** 302 396 O MN larynx 141 162.9 87 73 102 2 1.6 124 15 448 MN lung 2415 1918.6 126** 121 131 78 54.6 143** 113 178 MN pleura 611 46.0 1328** 1224 1437 134 4.7 2844** 2383 3369 MN uterus 34 35.7 95 66 133 MN ovary 43 31.1 138* 100 187 MN prostate 352 361.4 97 87 108 MN bladder 291 249.2 117* 104 131 19 9.5 199** 120 311 MN kidney 157 160.7 98 83 114 6 10.2 59 22 129 Leukemia and lymphoma 446 434.2 103 93 113 47 50.7 93 68 123 MN unspecified site 220 158.3 139** 121 159 19 18.1 105 63 164 Psychiatric diseases 143 161.0 89 75 105 51 34.6 147* 110 194 Neurological diseases 275 361.2 76** 67 86 45 63.3 71* 52 95 Cardiovascular diseases 5452 6209.0 88** 85 90 909 912.2 100 93 106 Respiratory diseases 1413 1113.4 127** 120 134 154 108.7 142** 120 166 Digestive diseases 932 1034.5 90** 84 96 118 104.3 113 94 136 Genitourinary diseases 184 219.0 84* 72 97 31 27.8 112 76 158 Asbestosis 366 1.22 30072** 27070 33317 51 0.13 38961** 29009 51227 Other Pneumoconioses 89 49.16 181** 145 224 2 0.1 143** 153 5160 Accidents and violence 851 1004 85** 79 91 76 78.6 97 76 121 Poorly specified causes 230 120.9 190** 166 216 75 32.93 228** 179 286 *p < 0.05; **p < 0.01; (-) no cases.

cant increase in overall mortality. In the asbestos ce- DISCUSSION ment sector mortality was significantly increased for Asbestos exposure has been an important source of peritoneal and pleural malignancies, lung cancer, and risk for Italian workers, because of the large use of as- asbestosis. Mortality was also increased for ovarian can- bestos and asbestos containing materials. The Italian cer but not reaching statistical significance. Registry of Malignant Mesothelioma (ReNaM) has In the glasswork sector, only pleural neoplasm showed documented the large number of cases of malignant a statistically significant increase, based on 3 deaths and mesothelioma that occur each year in Italy, for a total with a trend for cumulative exposure. Ovarian cancers of 27 356 reported cases in the period 1993 to 2015 were also more than expected. In this sector “all cancer” [21]. Besides the analysis of the occurrence of malig- mortality showed a reduction, in particular for gastroin- nant mesothelioma it is relevant also to investigate the testinal cancers, while an increase was observed for the occurrence of other diseases in the subjects who have deaths from diseases of the digestive tract. It is interesting been occupationally exposed to asbestos, as the assess- to note that no similar variations were observed for men. ment of the global burden of asbestos related diseases 298 Corrado Magnani, Stefano Silvestri, Alessia Angelini et al.

is important for epidemiological surveillance and com- did not over-estimate MM SIRs. Loomis et al [29] came pensation purposes. to a similar conclusion after a review conducted mainly A large project under this perspective was launched on US data. They also analysed individual data in their in 2013, after the recommendations issued by the Sec- record and came to the conclusion that under-ascer- ond Italian Governmental Conference on Asbestos and tainment of pleural mesothelioma also occurred after Asbestos related Diseases [10], including a large num- the adoption of ICD 10th classification. ber of cohorts in different industrial sectors. The project The information available in our cohort did not in- already produced results on the cause specific mortality clude individual data on jobs and work activities. As- reviews

in the total cohort [11], on mortality in the large and bestos exposure therefore could not be quantitatively homogeneous sector of the asbestos cement production assessed within each factory/cohort at the job or depart-

and [12] and on the variation of pleural cancer mortality by ment level. Instead it had to be carried out at the factory/

latency [22]. This report is a first systematic presenta- cohort level only. Exposure assessment was based on an tion of the results of mortality analyses by cumulative average index (AEI) representing the plant and period exposure and industrial sectors, including all industrial average exposure, obtained by combining two distinct sectors in the cohort. exposure estimates, for workers with direct and indirect articles

The pooled study included a large number of sub- asbestos exposure respectively. The individual cumula- jects (over 50 000) active in different industrial sectors, tive exposure index (CEI) of cohort members was then forming 43 different cohorts. This large number of dif- calculated by applying the plant- and period-specific ferent study groups, followed up and investigated by AEI to the duration and timing of employment of each riginal different research groups makes unlikely the occurrence worker. AEI and CEI were weighted for the proportion O of systematic errors affecting the entire project and is of the different fibre types used in each plant and pe- therefore an element of strength of the project. Other riod and their estimated carcinogenic potency factor for characteristics include the very long follow-up, more pleural MM [14]. More recent estimates of fibre type than 40 years, and the significant number of women, potency [16], were used in sensitivity analyses [12], and that made possible gender-specific analyses. no relevant difference was observed. In this analysis we The follow-up results are satisfactory: only 1.3% of applied the same estimates of exposure intensity to men subjects were lost or emigrated and had to be classified and women, as available historical data on airborne as- as “unknown” status. The cause of death was known bestos fibre concentrations in the factories included in for over 95% of decedents in both sexes. Causes of the pooled cohort are not gender-specific. death were classified according to standard classifica- Our findings for the asbestos cement sector are simi- tions. SMR analyses were based on regional mortality lar to results observed in comparable studies presented rates, for increasing comparability with reference rates. in the international literature [12, 30]. The asbestos Analyses were limited to period from 1970 onwards to cement sector has been analysed in a specific project, increase comparability with reference mortality rates, whose report can be accessed for more details [12]. In that were available only from then [20]. both sexes mortality was increased for “all causes” and The study was based on mortality data, that are the “all malignant neoplasm (MN)”, and for all asbestos re- only information homogeneously available for the pop- lated diseases, namely MN of peritoneum, pleura, lung ulation and period of interest for the study. There was and ovary, and asbestosis, with an exposure response no specific code for MM of peritoneum and pleura in trend. The large size of the cohort enabled analyses fo- the 8th and 9th ICD revisions, and ICD 10th revision is in cused on the effect at long latency, not presented here, use from 2003 in Italy. We classified in the categories of showing a flattening of the increasing trend after 40 pleural malignant neoplasm the following codes: ICD8: years of latency, for pleural but not for peritoneal malig- 1630, ICD9: 1630-1639, ICD10: C38.4, C45.0, C45.9. nancies [22]. Further analyses of asbestosis and of risk The codes according to which we classified the cause in women are under progress. of death as peritoneal malignant neoplasm were: ICD8 Glass making is more frequent in a few regions, in and ICD9: 1580-1589, ICD10: C48, C45.1. More in- particular Veneto, Tuscany, and also Emilia-Romagna formation on the codes used for each cause of death [31]. Our cohort included two large plants with indus- were provided by Ferrante et al. [11]. The use of mortal- trial production of flat glass and glass containers. The ity data might cause random misclassification between use of asbestos is associated to the handling and pro- MM and other cancers, in particular metastasis or lung duction of melting glass, in ovens insulation, in guide cancer. Kopylev et al. [23] explored in a meta-analysis holes to convey glass drops to molds, and for personal the sensitivity of death certificatesvs MM diagnosis protection devices. Other carcinogens are present in and observed an underestimation of MM incidence the glass making factories environment, including in from mortality data. Other studies not included in that particular silica, polynuclear aromatic hydrocarbons review came to the same conclusion: 74.5% of pleural (PAHs) and for some special productions also arsenic MM cases were identified from mortality records in or other metals [32] making the interpretation of re- Italy [24] and 87% in Southern England [25]. Similar sults more complex. The pooled cohort did not include results were observed by Conti et al., who compared art glass, that instead was the focus of most published mortality and incidence for peritoneal MM in Italy studies [33, 34]. Our cohort was very large, more than [26]. Some reports on mortality in cohorts included in any of the cohort studies included in the metanalysis our pooled study [18, 27, 28] conducted a record link- by Lehnert et al [35], and included a large number of age with the Italian ReNaM data observing that SMRs women, therefore had more power to estimate the risk 299 Italian asbestos workers cohorts

for rare cancers. We observed a statistically significant ral malignancies (with exposure response trend) and si- excess of pleural malignancies, with an SMR of 377 in nonasal cancers, suggesting that workers were exposed men and 700 in women, with an increasing trend with also to wood dust [5]. This cohort also showed a statisti- cumulative exposure in both sexes, statistically signifi- cally significant increase in mortality from pneumoconi- cant in men. Only Plato et al., had reported on the oc- osis (14 cases), only partially explained by asbestosis (3 currence of mesothelioma, observing a non statistically cases). The excess for pneumoconiosis (not asbestosis) significant SIR of 147 based on 3 cases in the Swedish suggests the presence of other risk factors and calls for surveillance of occupational diseases related to asbestos further investigations of this occupational activity. reviews

[36]. We observed an excess of deaths from pneumoco- The sectors of “dockyards”, “insulation”, “industrial niosis but no cases of asbestosis, suggesting that silica oven production”, “asphalt rolls” are represented by a

was an important exposure. single cohort each, none of which can be assumed as and

The industrial sector of rolling stock construction representative of its sector. and maintenance in Italy has been investigated with Dockyard workers showed a large excess of pleural the study of individual cohorts and with the report of cancer, with no clear trend with asbestos cumulative ex- mesothelioma cases in railway and rolling stock work- posure and a statistically significant excess of lung and articles ers [37-39]. Massive asbestos exposure in the Italian bladder cancer, suggesting the presence of exposure to rolling stock construction and maintenance started in other carcinogens. The lack of exposure response trend 1956 and lasted until 1986, corresponding to the deci- can depend on the small variation of exposure but it sion of the Italian National Railways first to introduce is also possible that our estimation of intensity of as- asbestos lagging for fire prevention and later to system- bestos exposure did not catch appropriately the charac- riginal atically remove it because of the observation of increas- teristics of exposure in that environment. The possible O ing number of mesothelioma cases. Asbestos insulation confounding effect of smoking cannot be ruled out but was sprayed, also by external firms; crocidolite was used it must be noticed that there is no increase in mortality for the new carriages until 1970, while chrysotile and for cardiovascular diseases. A large increase in meso- amosite were used afterward. The firms building new thelioma occurrence has been observed in the cohort cars also carried on maintenance activities on existing of the British Naval Dockyard [43]. An Italian pool of cars, and these activities included asbestos removal or five dockyards, not included in the present study, also replacement. The current pool includes the studies con- showed a statistically significant increase in lung cancer sidered in the comprehensive review by Merler et al., deaths but did not provide information on mesothe- [37] and the studies published afterwards, as well as lioma [44]. some unpublished cohorts. It also included a cohort of The cohort of insulators is very small and represents maintenance workers in the local train depots. The pool only a tiny fraction of workers in this sector. Neverthe- of this industrial sector showed a healthy worker effect, less results correspond to the expectation showing an with lower than expected total and cardiovascular mor- excess of all asbestos related diseases [45], even if with tality. Mortality from pleural and peritoneal cancer was broad confidence intervals due to the small cohort size. increased, with a sharp exposure response relation, in The production of industrial ovens is peculiar as, de- particular for pleural cancer. A few cases of asbestosis spite only a fraction of the workers had been considered were observed, all in the second and third tertiles of to be actually exposed due to handling asbestos panels cumulative exposure. These results are consistent with [46] a statistically significant excess of mesothelioma the observations from the international literature [40], was observed, with an exposure-response relation. although in the comparison it is important to take into The small cohort of asphalt rolls workers was recent- consideration the structural differences in the railway ly investigated by Zanardi et al., [47] and the present system of the different countries, and in particular the study provided similar information, with no increase extension of the use of steam vs electrical engines. of asbestos related diseases. A statistically significant We could include two shipyard workers cohorts and excess was observed for “lip, oral cavity and pharynx”, one cohort from a related plant involved in the interior increasing by tertiles of cumulative asbestos exposure. furnishing of vessels. Other shipyards are active in Italy The asbestos sampled at the factory was chrysotile only, and have been investigated, but could not participate in although Zanardi et al., [47] did not exclude the pos- the present call for this pooled study. In particular, the sible occurrence of amphibole contamination. Genoa shipyard was analysed and published recently, Despite the interest in providing through this pooled providing a detailed analysis of risk by job [41]. Over- study a broad description of the industrial sectors most- all, they observed a statistically significant increase for ly involved by the use of asbestos, their representation cancers of the larynx, lung, and pleura, as well as for in the pool is not homogeneous. The sector of asbes- asbestosis and for liver cirrhosis. The shipyards included tos cement is very well represented, with all the largest in the present study were mainly active in the construc- firms included and a large number of medium and small tion and repair, and therefore may not be comparable ones [48]. On the other side, the representation of the to shipbreaking shipyards [42]. In our study, main result other industrial sectors is more limited, with examples was related to the increased risk for lung cancer and rather than a representative description. Nevertheless, pleural malignancies, both with the observation of an our results do highlight a generalised, increased risk of exposure response trend. The cohort of ship furniture asbestos related diseases and call for an extension of workers was characterized by a different pattern, as it this pioneering effort to the largest possible number of showed a statistically significant increase for both pleu- cohorts of asbestos exposed workers. Some industrial 300 Corrado Magnani, Stefano Silvestri, Alessia Angelini et al.

sectors were not represented altogether, such as the Received on 6 December 2019. textile and the friction material sectors or the cohort Accepted on 31 March 2020. of Balangero chrysotile miners [49]. Other industrial sectors in which only a limited proportion of workers Members of the Working group Studio Multicentrico have been interested by asbestos exposure, most often Italiano di Coorti di lavoratori dell’Amianto (SMICA): maintenance workers, cannot be evaluated appropri- Corrado Magnani1,16, Stefano Silvestri1, Alessia An- ately through a cohort study design, unless the jobs or gelini1, Alessandra Ranucci1, Danila Azzolina1, Tiziana department assignments of all workers are known. Cena1, Elisabetta Chellini2, Enzo Merler3, Venere Pa-

reviews 4 2 2

vone , Lucia Miligi , Giuseppe Gorini , Vittoria Bres- Acknowledgements san3, Paolo Girardi3, Lisa Bauleo5, Elisa Romeo5, Fer- 6 7 4 and The present project was started in the framework of dinando Luberto , Orietta Sala , Corrado Scarnato ,

the “Asbestos Project” organized by the Italian Na- Simona Menegozzo8, Enrico Oddone9, Sara Tunesi1, tional Institute of Health (ISS) (Ricerca corrente 2012: Patrizia Perticaroli10, Aldo Pettinari10, Francesco Progetto amianto). A special thank to Loredana Mus- Cuccaro11, Stefano Mattioli12, Antonio Baldassarre13, meci, for the scientific coordination of the “Asbestos Francesco Barone-Adesi1, Marina Musti14, Dario Mi- articles

Project” and for her helpful assistance. The authors rabelli15,16, Roberta Pirastu17, Alessandro Marinaccio18, wish to thank the reviewers and the editor for their Stefania Massari18, Daniela Ferrante1, Maria Nicoletta helpful comments. Ballarin3, Carol Brentisci15, Barbara Cortini2, Stefania Curti12, Manuela Gangemi15, Francesco Gioffrè3, Pa- riginal Funding trizia Legittimo12, Lucia Mangone6, Francesco Mari- O The cohort study was conducted thanks to grants nelli12, Pasqualina Marinilli4, Chiara Panato3, Fran- from the “Asbestos Project” of the Italian National In- cesca Roncaglia6, Cinzia Storchi6, Antonella Stura15, stitute of Health (ISS), (Current research 2012: asbes- Massimo Vicentini6, Simona Verdi2, Anna Maria Nan- tos project. Operative Unit 2 “Amedeo Avogadro” Uni- navecchia11, Lucia Bisceglia19 versity of Eastern Piedmont, Novara, Research Line 1,2.) and from the INAIL (Piano Ricerca 2016-2018, Affiliation of the Working group members: “Programma Speciale Amianto”, Ricerca BRIC id 55 1Dipartimento di Medicina Traslazionale, Università and Ricerca BRIC id 59). degli Studi del Piemonte Orientale, Novara, Italy 2Istituto per lo Studio, la Prevenzione e la Rete Onco- Authors contribution logica (ISPRO), Florence, Italy CM (study coordinator), PI the study, overview of 3Registro Mesoteliomi del Veneto, Sistema Epidemi- the study and critical revision of the article; SS overview ologico Regionale, ASL 6, Padua, Italy and analysis of exposure information, estimation of cu- 4Dipartimento di Sanità Pubblica, Azienda Sanitaria mulative exposure and critical revision of the article; Locale, Bologna, Italy AA overview and analysis of exposure information, es- 5Dipartimento di Epidemiologia del Servizio Sanitar- timation of cumulative exposure and critical revision of io Regionale, ASL RM 1, Rome, Italy the article; AR, TC, ST data management, data analysis 6Servizio di Epidemiologia, Azienda Unità Sanitaria and critical revision of the article; DA data analysis and Locale, IRCCS di Reggio Emilia, Reggio Emilia, Italy critical revision of the article; EC, EM, VP, LM, GG, 7formerly ARPAE Emilia-Romagna, Ufficio Provin- VB, PG, LB, ER, FL, OS, CS, SM, EO, PP, AP, FC, ciale di Reggio Emilia, Reggio Emilia, Italy SM, AB, MM design of the study, conduct of the study 8Istituto Nazionale Tumori IRCCS Fondazione and critical revision of the article; FB-A planning and G.Pascale, Naples, Italy overview of data analysis and critical revision of the ar- 9Dipartimento di Sanità Pubblica, Medicina Speri- ticle; DM design of the study, evaluation of exposure in- mentale e Forense, Università degli Studi di Pavia, Pa- formation and critical revision of the article; RP design via, Italy of the study, overview of mortality data analyses and 10Dipartimento Prevenzione, ASUR Marche, Senigal- critical revision of the article; AM design of the study, lia, Italy incidence data collection coordination, and critical revi- 11Unità di Epidemiologia e Statistica, Unità Sanitaria sion of the article; SM design of the study and critical Locale di Barletta-Andria-Trani, Barletta, Italy revision of the article; DF design of the study, design 12Dipartimento di Scienze Mediche e Chirurgiche, and conduct of data analysis and drafting of the article; Università degli Studi di Bologna, Bologna, Italy MNB, CB, BC, SC, MG, FG, PL,, LM, FM, PM, CP, 13Unità di Medicina del Lavoro, Azienda Ospedaliero- FR, CS, AS, MV, SV, AMN, LB contributed to the con- Universitaria Careggi, Florence, Italy duct of the study and to data collection. 14Dipartimento Interdisciplinare di Medicine, Uni- versità degli Studi di Bari, Bari, Italy Conflict of interest statement 15Unità di Epidemiologia dei Tumori, Università degli The authors declare that they have no competing in- Studi di Torino, Turin, Italy terests. No authors declared financial conflict of interest. 16Centro Interdipartimentale per lo Studio degli Ami- The following authors reported that they served as ex- anti e di altri Particolati Nocivi “Giovanni Scansetti”, pert witness in court trials on asbestos related diseases: Università degli Studi di Torino, Turin, Italy AA, AB, FB-A, LB, CM, LM, SM, EM, DM, MM, 17Dipartimento di Biologia e Biotecnologie “Charles EO, VP, SS. Darwin”, La Sapienza Università di Roma, Rome, Italy 301 Italian asbestos workers cohorts

18Dipartimento di Medicina, Epidemiologia, Igiene Rome, Italy del Lavoro e Ambientale, Istituto Nazionale per 19Agenzia Regionale Strategica per la Salute e il So- l’Assicurazione contro gli Infortuni sul Lavoro (INAIL), ciale, ARESS Puglia, Bari, Italy

REFERENCES

1. Virta RL. Worldwide asbestos supply and consumption for mesothelioma in humans. Toxicol Appl Pharmacol. reviews trends from 1900 through 2003: US Geological Survey 2019;371:1-2. Circular 1298. 2006. Available from: https://pubs.usgs. 18. Magnani C, Ferrante D, Barone-Adesi F, Bertolotti M, and

gov/circ/2006/1298/c1298.pdf. Todesco A, Mirabelli D, et al. Cancer risk after cessation 2. Silvestri S. Managing asbestos in Italy: twenty years after of asbestos exposure: a cohort study of Italian asbestos the ban. New Solut. 2012;22:489-96. cement workers. Occup Environ Med. 2008;65:164-70. 3. Mirabelli D, Kauppinen T. Occupational exposures to 19. Breslow NE, Day NE. Statistical methods in cancer carcinogens in Italy: an update of CAREX database. Int J research. Volume II--The design and analysis of cohort articles

Occup Environ Health. 2005;11:53-63. studies. IARC Sci Publ. 1987:1-406. 4. Silvestri S, Benvenuti A, Cavone D, Cavariani F, Di Mar- 20. Pirastu R, Ranucci A, Consonni D, De Santis M, Bruno zio D, et al. L’esposizione ad amianto nel settore edile: C, Conti S, et al. Reference rates for cohort studies in considerazioni generali ed analisi dei dati ReNaM. In: Italy: an essential tool in occupational and residential co-

Marinaccio A, Binazzi A, Di Marzio D, Scarselli A, Ve- hort studies. Med Lav. 2016;107:473-7. riginal rardo M, Mirabelli D, . Registro Nazionale dei Me- 21. Marinaccio A, Binazzi A, Bonafede M, Branchi C, Bu-

et al O soteliomi. Terzo Rapporto. Roma: Istituto Superiore per gani M, Marisa Corfiati M,et al. Il Registro Nazionale la Prevenzione e la Sicurezza sul Lavoro; 2010. p. 123-37. dei Mesoteliomi. VI Rapporto. Milano: Inail; 2018. 5. International Agency for Research on Cancer. IARC 22. Barone-Adesi F, Ferrante D, Chellini E, Merler E, Pa- monographs on the evaluation of carcinogenic risks to vone V, Silvestri S, et al. Role of asbestos clearance in ex- humans. Arsenic, metals, fibres, and dusts. A review of plaining long-term risk of pleural and peritoneal cancer: human carcinogens. Volume 100 C. IARC; 2012. a pooled analysis of cohort studies. Occup Environ Med. 6. US Geological Survey. Mineral commodity summaries. 2019;76:611-6. US Geological Survey; 2019. p. 26-27. doi: https://doi. 23. Kopylev L, Sullivan PA, Vinikoor LC, Bateson TF. Monte org/10.3133/70202434 Carlo analysis of impact of underascertainment of meso- 7. Frank AL, Joshi TK. The global spread of asbestos. Ann thelioma cases on underestimation of risk. Open Epide- Glob Health. 2014;80:257-62. miol J. 2011;4:45-53. 8. Pasetto R, Terracini B, Marsili D, Comba P. Occupation- 24. Bruno C, Comba P, Maiozzi P, Vetrugno T. Accuracy of al burden of asbestos-related cancer in Argentina, Brazil, death certification of pleural mesothelioma in Italy. Eur J Colombia, and Mexico. Ann Glob Health. 2014;80:263- Epidemiol. 1996;12:421-3. 8. 25. Okello C, Treasure T, Nicholson AG, Peto J, Møller H. 9. ISS. Progetto Amianto. 2016. Available from: www. iss. Certified causes of death in patients with mesothelioma it/ amianto. in south East England. BMC Cancer. 2009;9:28. 10. Ministero della Salute, Direzione Generale della Co- 26. Conti S, Minelli G, Ascoli V, Marinaccio A, Bonafede M, municazione e delle Relazioni Istituzionali del Minis- Manno V, et al. Peritoneal mesothelioma in Italy: trends tero della Salute. Atti della II Conferenza governativa and geography of mortality and incidence. Am J Ind sull’amianto e le patologie asbestocorrelate. 2012. Avail- Med. 2015;58:1050-8. able from: www.salute.gov.it 27. Luberto F, Amendola P, Belli S, Bruno C, Candela S, Gri- 11. Ferrante D, Chellini E, Merler E, Pavone V, Silvestri S, gnoli M, et al. Mortality study of asbestos cement work- Miligi L, et al. Italian pool of asbestos workers cohorts: ers in Emilia-Romagna. Epidemiol Prev. 2004;28:239-46. mortality trends of asbestos-related neoplasms after 28. Menegozzo S, Comba P, Ferrante D, De Santis M, long time since first exposure. Occup Environ Med. Gorini G, Izzo F, et al. Mortality study in an asbestos 2017;74:887-98. cement factory in Naples, Italy. Ann Ist Super Sanità. 12. Luberto F, Ferrante D, Silvestri S, Angelini A, Cuccaro F, 2011;47:296-304. Nannavecchia AM, et al. Cumulative asbestos exposure 29. Loomis D, Richardson DB, Elliott L. Quantitative rela- and mortality from asbestos related diseases in a pooled tionships of exposure to chrysotile asbestos and mesothe- analysis of 21 asbestos cement cohorts in Italy. Environ lioma mortality. Am J Ind Med. 2019;62:471‐7. Health. 2019;18:71. 30. Goodman M, Morgan RW, Ray R, Malloy CD, Zhao 13. Stewart PA, Stewart WF. Occupational case-control stud- K. Cancer in asbestos-exposed occupational cohorts: a ies: II. Recommendations for exposure assessment. Am J meta-analysis. Cancer Causes Control. 1999;10:453-65. Ind Med. 1994;26:313-26. 31. Marinaccio A, Binazzi A, Branchi C, Corfiati M, Di Mar- 14. Hodgson JT, Darnton A. Mesothelioma risk from chryso- zio D, Scarselli A. Il Registro Nazionale dei Mesoteliomi. tile. Occup Environ Med. 2010;67:432. IV rapporto. Milano: Inail; 2012. p. 127. 15. Hodgson JT, Darnton A. The quantitative risks of meso- 32. Wingren G, Axelson O. Mortality in the Swedish thelioma and lung cancer in relation to asbestos expo- glassworks industry. Scand J Work Environ Health. sure. Ann Occup Hyg. 2000;44:565‐601. 1987;13:412-6. 16. Garabrant DH, Pastula ST. A comparison of asbestos fi- 33. Wingren G. Mortality and cancer incidence in a Swedish ber potency and elongate mineral particle (EMP) poten- art glassworks--an updated cohort study. Int Arch Occup cy for mesothelioma in humans. Toxicol Appl Pharmacol. Environ Health. 2004;77:599-603. 2018;361:127-36. 34. Pirastu R, Bartoli D, Battista G, De Santis M, Iaia T, Orsi 17. Finkelstein MM. A comparison of asbestos fiber po- D, et al. Cancer mortality of art glass workers in Tuscany, tency and elongate mineral particle (EMP) potency Italy. Scand J Work Environ Health. 1998;24:386-91. 302 Corrado Magnani, Stefano Silvestri, Alessia Angelini et al.

35. Lehnert M, Behrens T, Tulowietzki J, Guldner K, Brüning Environ Res. 2014;132:370-8 T, Taeger D. Cancer in glass workers: a systematic re- 43. Hilliard AK, Lovett JK, McGavin CR. The rise and fall view and meta-analysis. Int Arch Occup Environ Health. in incidence of malignant mesothelioma from a Brit- 2019. doi: 10.1007/s00420-019-01460-1. ish Naval Dockyard, 1979-1999. Occup Med (Lond). 36. Plato N, Martinsen JI, Sparén P, Hillerdal G, Weiderpass 2003;53:209-12. E. Occupation and mesothelioma in Sweden: updated 44. Bonassi S, Ceppi M, Puntoni R, Valerio F, Vercelli M, incidence in men and women in the 27 years after the Belli S, et al. Mortality studies of dockyard workers (long- asbestos ban. Epidemiol Health. 2016;38:e2016039. shoremen) in Italy.Am J Ind Med. 1985;7:219-27. 37. Merler E, Silvestri S, Roberti M, Menegozzo M. I casi 45. Ulvestad B, Kjaerheim K, Martinsen JI, Mowe G, Ander- reviews

di mesotelioma maligno per esposizione lavorativa ad sen A. Cancer incidence among members of the Norwe- amianto nella costruzione, manutenzione, riparazione gian trade union of insulation workers. J Occup Environ

and e bonifica dei mezzi ferroviari: considerazioni generali Med. 2004;46:84-9. ed analisi dei dati ReNaM. In: Marinaccio A, Cauzillo 46. Fedi A, Blagini B, Melosi A, Marzuoli E, Ancillotti M, G, Chellini E, DeZotti R, Gennaro V, Gorini G, et al. Gorini G, et al. Assessment of asbestos exposure, mor- Registro Nazionale dei Mesoteliomi. Secondo Rapporto. tality study, and health intervention in workers formerly Roma: Istituto Superiore per la Prevenzione e la Sicurez- exposed to asbestos in a small factory making drying ma-

articles za sul Lavoro; 2006. chines for textile finishing and the paper mill industry in 38. Comba P, Pasetto R. Health impact of asbestos exposure Pistoia, Italy. Med Lav. 2005;96:243-9. in the occupational sector of construction and repair of 47. Zanardi F, Salvarani R, Cooke RM, Pirastu R, Baccini M, railway rolling stock. Eur J Oncol. 2004;9:87-90. Christiani D, et al. Carcinoma of the pharynx and tonsils 39. Pavone VL, Scarnato C, Marinilli P, Costellati L, Tiranno in an occupational cohort of asphalt workers. Epidemiol- riginal A, Alberghini V, et al. Mortality in a cohort of railway ogy. 2013;24:100-3. O rolling stock construction and repair workers in Bologna. 48. Mirabelli D, Cavone D, Luberto F, Menegozzo M, Med Lav. 2012;103:112-22 Mensi C, Merler E, et al. Il comparto della produzione 40. Hosoda Y, Hiraga Y, Sasagawa S. Railways and asbestos di Cemento Amianto. In: Marinaccio A, Binazzi A, Di in Japan (1928-1987) – epidemiology of pleural plaques, Marzio D, Scarselli A, Verardo M, Mirabelli D, et al. Reg- malignancies and pneumoconioses. J Occup Health. istro Nazionale dei Mesoteliomi. Terzo rapporto. Roma: 2008;50:297-307 Istituto Superiore per la Prevenzione e la Sicurezza sul 41. Merlo DF, Bruzzone M, Bruzzi P, Garrone E, Puntoni R, Lavoro; 2010. p. 105-21. Maiorana L, Ceppi M. Mortality among workers exposed 49. Ferrante D, Mirabelli D, Silvestri S, Azzolina A, Giovan- to asbestos at the shipyard of Genoa, Italy: a 55 years nini A, Tribaudino P, Magnani C. Mortality and meso- follow-up. Environ Health. 2018;17:94. thelioma incidence among chrysotile asbestos miners 42. Wu WT, Lin YJ, Shiue HS, Li CY, Tsai PJ, Yang CY, Liou in Balangero, Italy: A cohort study. Am J INd Med. SH, Wu TN. Cancer incidence of Taiwanese shipbreaking First published 14 November 2019. doi: https://doi. workers who have been potentially exposed to asbestos. org/10.1002/ajim.23071 Piazza Sant’Onofrio 4,00146Rome,Italy. E-mail:[email protected]. Address forcorrespondence: LauraCastaldo,Unità diNeuropsichiatriaInfantile,Dipartimento diNeuroscience,Ospedale PediatricoBambinoGesù, INTRODUCTION with moreseveredepressivesymptoms. for SI or SB, who also presentedNSSI, were mainly substance abusers, and presented sultations forsuicidalbehaviours,mainlyinfemaleadolescents. Thesubjectsadmitted Conclusions. Between2011and2016,there was asignificantgrowingneedforCPcon- female, substanceabuse,andmoreseveredepressivesymptoms. for SBorSI,3factorsweresignificantlyandindependentlyassociatedwithNSSI:being admitted forSBorSIreportedacurrentlifelonghistoryofNSSI.Inpatients tients admittedforSBwhencomparedtotheSIgroup.Almosthalf(48%)ofpatients of patientshospitalizedforSBorSI.Substanceabusewasmorefrequentlyfoundinpa- (53%), withasignificantprevalenceinfemale.Mooddisorderwasidentified82.5% male ratio,andanaverageageof15.5±1.8years.SBconsistedmainlyindrugpoisoning 2016, with a statistically significant increase over the years (p < 0.001), a 2.2:1 female/ Results. TherateofCPcforSB,SIandNSSIswitchedfrom7.7%in2011to19% Child PsychiatryUnitafteraCPc;and,correlationsbetweenNSSIandSB/SI. CPc forSB,SIandNSSI;riskfactorsSBinpatientsadmittedtothehospital’s 2011 to2016,whounderwentchildpsychiatryconsultation(CPc).We analyzed:ratesof Methods. RetrospectiveanalysisoftheclinicalrecordspatientsadmittedtoEDfrom served atanItalianpaediatricemergencydepartment(ED). suicidal behaviour(SB),ideation(SI)andnon-suicidalself-injury(NSSI)–ob- growing concern. The objective of the study is to describe non-fatal self-harm – including Introduction. Increasing rates of non-fatal self-harm amongst adolescents represent Abstract 5 4 3 2 1 Laura Castaldo department seen atanItalianpaediatricemergency self-injury inchildrenandadolescents Suicidal behaviourandnon-suicidal DOI: 10.4415/ANN_20_03_08 Ann IstSuperSanità2020|Vol. 56,No.3:303-314 Rome, Italy if comparedtonon-fatalself-harm [3,4],whichison to bethetipoficebergand isrelativelyuncommon alent in males than in females [2]. Suicide is considered to 1.4/100000person-year, andis4.4timesmoreprev- adolescents andyoungadults aged<24yearsamounts worldwide [1].InItaly, theincidenceofsuicideamongst death amongstsubjectsagedbetween15and29years nization (WHO),suicideisthesecondleadingcauseof ing adolescence.AccordingtotheWorld HealthOrga- Dipartimento diPsichiatria,UniversitàCattolicadelSacroCuore,Rome,Italy Dipartimento diPediatriadell’Emergenza,OspedalePediatricoBambinoGesù,Rome,Italy Unità diEpidemiologiaClinica,OspedalePediatricoBambinoGesù,Rome,Italy Dipartimento diNeuroscienzeUmane,SapienzaUniversitàRoma,Rome,Italy Unità diNeuropsichiatriaInfantile,DipartimentoNeuroscience,OspedalePediatricoBambinoGesù, Suicide andself-harmarethemainhealthissuesdur 1 , GiuliaSerra 1,2 , SimonePiga - 3 , AntoninoReale methods, inrepresentativeadolescent samplesfrom11 -burning, -biting,-hitting,and skindamagebyother prevalence in clinical samples [5, 6]such as self-cutting, Empowering Young LivesinEurope(SEYLE,anda50% within theEuropeanUnionfunded project,Savingand sectional assessmentofadolescentswasperformed from 11Europeancountries.\\n\\nMETHODS:Cross- by othermethods,inrepresentativeadolescentsamples self-cutting, -burning,-biting,-hitting,andskindamage a 10%prevalenceincommunitysamples[3,5]suchas average reportedby27%ofadolescents,withalmost 4 andStefanoVicari • • • • Key words risk factor non-suicidal self-injury suicidal behaviour self-harm 1,5

303 Original articles and reviews 304 Laura Castaldo, Giulia Serra, Simone Piga et al.

European countries.\\n\\nMETHODS: Cross-sectional a prior history of NSSI represents one of the strongest assessment of adolescents was performed within the predictors of future suicide attempts. NSSI has been European Union funded project, Saving and Empower- prospectively identified as a predictive factor of a new ing Young Lives in Europe (SEYLE. In a recent study onset of suicidal ideation (SI) and a new suicide at- involving eleven European countries [5]such as self- tempt at 12-24 months. In particular, later onset and cutting, -burning, -biting, -hitting, and skin damage by repetitive NSSI are associated with a two-fold increased other methods, in representative adolescent samples risk for SI and a three-fold increased risk for suicide at- from 11 European countries.\\n\\nMETHODS: Cross- tempts at 12-24 months. On the contrary, NSSI with reviews

sectional assessment of adolescents was performed onset during early adolescence and a short duration of within the European Union funded project, Saving and self-harm behaviours seem to be linked to a smaller risk

and Empowering Young Lives in Europe (SEYLE, the esti- for SB and SI in the following years [19].

mated prevalence of self-harm amongst Italian adoles- Several theories have been developed to explain the cents amounted to 21%, out of which, 16% concerned transition from NSSI to SB. According to Joiner’s in- occasional self-harm and 5% repetitive self-harm. terpersonal theory of suicide [20], repetitive self-harm Self-harm tends to occur at 12-14 years of age and may develop individuals’ capability to commit suicide. articles is more common in female adolescents, with a girl/boy This capability develops through individuals’ habitu- ratio of 5/6:1 [7]95.8%. Presentation at a clinical service ation to fear and pain, which increases the likelihood only involves one out of eight adolescents who self-harm that suicidal thoughts are turned into action. Based on in the community, and is more common amongst those this theory, a recent review highlighted how adolescent riginal who attempt self-poisoning [8, 9]. The estimated mor- self-injurers who attempted suicide reported a higher O tality index for non-fatal self-harm amounts to 1/370 number of NSSI and more dangerous behaviours (e.g. adolescents who are admitted to hospital for self-harm, substance abuse, aggression) compared to adolescent and to 1/3900 adolescents who report self-harm in the self-injurers who did not show suicidal behaviours [19]. community [10]. In 2014, WHO published Preventing suicide. A global These findings outline how massive the issue repre- imperative [21], highlighting the need to focus on the sented by community-occurring self-harm is, and high- predictors of suicide, i.e. psychiatric illnesses, substance- light an urgent need for identifying the risk factors for abuse-related behaviours, traumatisms, and prior self- progression from largely prevalent self-harm behaviours injurious behaviours, which can help to identify subjects identified at the community level to suicide, and for de- at risk at whom specific therapeutic interventions should veloping appropriate protocols to monitor and treat the be aimed. In that respect, WHO highlighted the need for subjects presenting to clinical services after self-harm. epidemiological data collection strategies to estimate the A first step towards identification of the risk factors population at risk and better plan economic resources, for progression of self-harm to suicide is classifying self- prevention strategies and treatment settings. injurious behaviours on the basis of the presence/ab- Answering this need for epidemiological data collec- sence of the intent to die. Following the growing need tion, our study aims at describing non-fatal self-harm – in- for suicidal risk stratification amongst self-injurious cluding SB, SI and NSSI – in the juvenile population seen behaviours, the last edition of the Diagnostic and Sta- at the emergency department (ED) of the largest chil- tistical Manual of Mental Disorders (DSM-5) [11] has dren’s hospital in Italy, at analyzing the risk factors for sui- proposed a distinction between: a) suicidal behaviour cidal behaviours and evaluating the differences between disorder, which includes all self-initiated suicidal behav- SB/SI + NSSI vs the SB/SI without-NSSI in children and iours; and b) non-suicidal self-injury (NSSI), described adolescents admitted to the child psychiatry unit (CPU). as intentionally repeated (≥ 5 times/year) self-inflicted destruction of body tissue without suicidal intent. Al- METHODS though in the DSM-5 they are classified as two separate Study design and selection of subjects diagnostic categories, the distinction between NSSI This is an epidemiological study based on a chart re- and suicidal behaviours (i.e. suicide attempt) is far from view of patients admitted to the ED of Bambino Gesù clear. These behaviours present with a continuum of Children’s Hospital (Ospedale Pediatrico Bambino symptoms’ severity [12] and share the same distal and Gesù – OPBG) from January 1st, 2011 to December proximal risk factors [13]. Thus, the risk for progression 31st, 2016, who required a child psychiatry consultation from one to the other is one of the main clinical chal- (CPc) (Figure 1). lenges for developmental psychiatrists working in emer- The OPBG is a teaching and research institute in La- gency departments. Suicidal behaviour (SB) and NSSI tium as well as the largest paediatric hospital in Italy. have been differentiated in terms of prevalence (NSSI The ED provides free medical assistance to all children is more prevalent than SB in the community), inten- aged from 0 to 18 years, 24 hours a day, 7 days a week. tion (NSSI is carried out without the intent to die), fre- In 2011-2016, the number of visits to the ED amounted quency (NSSI occurs more frequently, particularly so in to approximately 74 000 per average year. The hospital clinical samples), and lethality (NSSI often involves low has a total of 607 beds, out of which 8 in the CPU. lethality injuries, e.g., cutting, burning while SB tends Most patients come from Rome and the surrounding to involve higher lethality injuries, e.g., self-poisoning, areas, however, the hospital is also a main reference hanging, wrist cutting) [14] including nonsuicidal self- centre for patients coming from the South and the Cen- injury (NSSI. Longitudinal research studies [15-18] tre of Italy, and, to some extent, from the North of Italy nonfatal self-poisoning or self-injury have shown how and from abroad. 305 Suicidal behaviour and non-suicidal self-injury in children and adolescent

Patients admitted to the ED of Bambino Gesù Children’s Hospital from January 1st, 2011 to December 31st, 2016 reviews CP consultations Other consultations and

Analysis of accesses to the ED SB, SI and NSSI sex, age, data and reason for admission, outcome of the Other ED evaluations (discharge or hospitalization), method articles

reasons and outcome of SB, potential and actual lethality of SB

Hospitalization riginal

Discharge O

Analysis of hospitalizations in the CPU Hospitalization sex, age, reason for admission, Hospitalization for SB and SI duration of admission (days), for NSSI environmental anamnesis/ environmental risk factors, psychopathological evaluation and pharmacological treatment

Figure 1 Study design and selection of subjects. CP = child psychiatric; SB = suicidal behaviour; SI = suicidal ideation; NSSI = non-suicidal self-injury; ED = emergency department; CPU = Child Psychiatry Unit.

Data were collected from GIPSE (the regional soft- ment used at the CPU of Bambino Gesù Children’s ware for the management of admissions to emergency Hospital. The data collected included sex, age, physio- departments) and from the hospital’s online database logical, clinical and psychiatric anamnesis, including in- where all patients are routinely registered. We selected, formation about previous and current pharmacological out of the patients requiring CPc, all the subjects who treatment, history of substance abuse, reason for hos- were admitted to the ED for SB, SI or NSSI. Infor- pitalization, duration of hospital stay (days) and treat- mation about the patients included sex, age, date and ment course. In addition, details about potential envi- reason for admission, and the outcome of the ED evalu- ronmental, family-related and clinical risk factors for ation including discharge or hospitalization in the CPU suicidal behaviours, including family history of mood (please see flow chart of the study inFigure 1). For the or other psychiatric disorders, family history of suicidal subjects admitted to the ED for SB, the method and behaviours, previous suicidal or non-suicidal self-injury the outcome of the attempt, including potential or ac- behaviours, history of child neglect/experience of loss, tual lethality, were also considered. family conflicts, history of bullying and of sexual and/ The subjects hospitalized in the CPU following CPc or physical abuse were also evaluated. Readmissions to for SB or SI from January 2011 to December 2016 were CPU were also taken into account. extensively evaluated via standardized diagnostic inter- views to assess psychiatric disorders; standardized rat- Structured psychopathological assessment ing scales for symptoms and risk factors for suicidal be- Inpatients and at least one of their parents or the haviour were also assessed. For subjects with repeated carer (when neither parent was available) were inter- hospitalizations for SB or SI during the 6-year period viewed via the semi-structured Schedule for Affective being evaluated, only data regarding their first hospital- Disorders and Schizophrenia for School Age Children, ization were analyzed. Present and Lifetime version (K-SADS-PL) [22] to as- sess current and past psychopathological features and Assessment of subjects psychiatric disorders in children and adolescents ac- Information about subjects was collected at the time cording to the Diagnostic and Statistical Manual of of their first evaluation via the semi-structured assess- Mental Disorders, Fourth Edition (DSM-IV), criteria. 306 Laura Castaldo, Giulia Serra, Simone Piga et al.

Data were collected via the K-SADS-PL, including Whitney test was used for continuous variables. We substance-abuse related problems and exposure to trau- compared risk factors (family history of mood disorders, matic experiences (i.e. sexual/physical abuse and child SB, substance abuse and/or other psychiatric disorders; neglect/experience of loss). previous SB, NSSI, substance abuse, experience of loss/ Symptoms of anxiety and depression were further child neglect, parental separation/divorce, conflict with evaluated on the Multidimensional Anxiety Scale for parents, bullying, and sexual and/or physical abuse) Children (MASC) [23] and the Children’s Depression amongst patients admitted to the hospital’s CPU in Inventory (CDI) [24] respectively. MASC is a 39-item terms of reasons for hospitalization (SB vs SI) and sex. reviews

self-rated scale for children and adolescents that evalu- We also stratified the sample of subjects hospitalized ates anxiety symptoms and includes four subscales: at the CPU for SB/SI by the presence/absence of NSSI.

and physical symptoms, social anxiety, harm avoidance and Multivariate logistic regression models were developed

separation anxiety. CDI is a 27-item self-rated scale to assess independent predictors of NSSI in the SB/SI + used to assess depressive symptoms in children and NSSI group vs the SB/SI without-NSSI group. adolescents aged 7-17 years. According to the Italian Variables with p < 0.20 in univariate analysis were validation criteria, a 19-point cut-off suggests the pres- included in the multivariate logistic regression models. articles

ence of clinically significant depressive symptoms [25]. Final models were computed with a stepwise backward Children and adolescents completed both the CDI and procedure (likelihood ratio test, p < 0.05). The Hosmer- the MASC with the support of trained psychologists. Lemenshow test was used to measure the model per- When needed, in accordance with the scales’ standard formance (goodness-of-fit test). Statistical analysis was riginal instructions, questions were read and, if needed, ex- performed using Stata, version 13 (StataCorp). O plained to children. The Italian version of the Child Behaviour Checklist RESULTS for ages 6 to 18 years (CBCL/6-18) was completed by Subjects admitted to the ED for SB, SI or NSSI inpatients’ parents/caregivers to rate behavioural and Overall, out of 2237 admissions to the ED that re- emotional problems in children and adolescents [26, quired CPc between 2011 and 2016, 13.1% (n = 294) 27]. The CBCL is an extensively used tool that provides were for SB, SI and NSSI, the female/male ratio was scores for three broadband behaviour rating scales: in- 2.2:1 (204 females and 90 males), and the average pa- ternalizing symptoms, externalizing symptoms and tients’ age was 15.5 ± 1.8 years. total behavioural problems. Sub-items of these three With regard to temporal trends, CPc percentage for broadband scales included the eight syndrome scales SB, SI and NSSI switched from 7.7% in 2011 to 19% (withdrawn/depressed, somatic complaints, anxious/de- in 2016, with a statistically significant increase over the pression, social problems, thought problems, attention years (from 12 cases in 2011 to 118 in 2016, p < 0.001; problems, rule-breaking behaviour, and aggressive be- see Figure 2). SB was the main reason for CPc at the haviour). Scores were transformed into a T-score, based ED during the period being evaluated, with a propor- on age and gender norms: a clinical cut-point value for tion of 46.3% (n = 136); the proportion of admissions to domain-specific syndrome scales and broadband scales the ED for SI and NSSI amounted to 36.4% (n = 107) was determined as the minimum raw score correspond- and 17.3% (n = 51) respectively. ing to T-score ≥ 65. The psychometric properties of the The most common method for SB admitted to the CBCL and its validity and reliability for use amongst ED was drug poisoning (53%), followed by attempted Italian population have been demonstrated [27, 28]. or completed jumping out of a window (25%), wrist The Italian version of the Columbia-Suicide Severity cutting/stabbing (8.8%), ingestion of toxic substances Rating Scale (C-SSRS) was used to rate SB with rel- (7.4%), and hanging/jumping in front of a train (5.9%). evant actual or potential lethality and SI. The psycho- Females were more frequently admitted for drug poi- metric properties and validity of the C-SSRS have been soning compared to males (60% vs 40%, c2 = 5.3; p = demonstrated [29]. In addition, information about cur- 0.02). rent SB, SI and NSSI was asked in the K-SADS-PL in- Out of all the patients admitted to the ED for SB, terview’s screening section for affective disorders. 67% presented with mild-to-severe physical injuries. In We defined SB as any self-initiated suicidal behav- particular, 31% presented with moderate physical in- iour, and NSSI as any intentionally repeated (≥ 5 times/ juries, 14.7% with moderately severe physical injuries, year) self-inflicted damage to the surface of the body and another 14% with minor physical injuries. Severe without suicidal intent [11]. SI was defined as any ac- physical injuries requiring intensive care management tive thought about wanting to end one’s life/commit sui- concerned 7.4% of patients. Patients identified as hav- cide, including SI with/without method, intent to act or ing no physical injuries were most commonly males specific plan (“Yes” at C-SSRS items 2 to 5). (44% vs 27%, c2 = 3.9, p = 0.04). In most cases, the level of potential lethality was high (95.3% of patients with Statistical analysis actual lethality = 0). Categorical variables were summarized by absolute In 85% (n = 249) of CP consultations for SB, SI and frequencies and percentages, continuous variables by NSSI, hospitalization at the CPU was recommended, mean and standard deviation. To determine statistical this was also true for patients admitted more than once differences between the groups, the χ2 test or Fisher for the same reason. SB and SI more frequently re- exact test was used for categorical variables and χ2 test sulted in hospitalization (90% and 93.5% respectively) for trends as appropriate, whereas the T-test or Mann- compared to NSSI (52.9%). The average length of stay 307 Suicidal behaviour and non-suicidal self-injury in children and adolescent reviews

and

articles

Figure 2 Number of CP consultations for SB, SI and NSSI observed at the ED of Bambino Gesù Children’s Hospital in Rome from January 1st, riginal 2011 to December 31st, 2016. Mean annual rate of CP consultations between 2011 and 2016 amounted to 373, with an increasing trend from 155 in 2011 to 614 in 2016. Over the same 6-year period, 294 adolescents aged between 11 and 17 years (average age: O 15 years) accessed the ED for SB, SI and NSSI, with a statistically significant increase over the years (from 12 cases in 2011, to 118 in 2016; p < 0.001). CP = child psychiatric; SB = suicidal behaviour; SI = suicidal ideation; NSSI = non-suicidal self-injury; ED = emergency department.

was 11.5 ± 9.2 days. Subjects admitted to the CPU for Risk factors for SB and SI NSSI only (in absence of SB or SI, n = 27) were ex- Figure 3 shows the most frequent risk factors for SB cluded from the analysis of hospitalized subjects as the and SI identified in our sample, with a distribution by number was too little to allow comparison with the SB sex and reason for hospitalization (SB vs SI). and SI groups. The most frequent risk factors were NSSI (47.9%) and conflicts between patients and their parents Subjects hospitalized in the CPU for SB or SI (43.3%), with a significantly higher prevalence in fe- One-hundred and seventy-one (n = 171) patients males versus males (c2 = 19.9, p < 0.0001 and c2 = 4.6, were hospitalized in the CPU at least once in 2011- p = 0.032 respectively), followed by parental separation 2016 following CPc at the ED for SB or SI, with a fe- or divorce (33.7%), family history of mood disorders male/male ratio of 2/1 (66.1% females). Seventy-nine (31.8%) or other psychiatric disorders (24.3%), previ- patients were admitted for SI (46.2%), and 92 for SB ous suicidal behaviours (27.5%), experience of loss/ (53.8%), with no female/male ratio difference in terms child neglect (24.6%), substance abuse (16.4%) and of hospitalization, and an average length of stay of 12.6 bullying (15.9%). Substance abuse was more frequently ± 10 days. identified in the subjects hospitalized for SB (c2 = 4.2, The most frequent diagnosis was mood disorder, in p = 0.04) compared to the SI group. A history of sexual 82.5% of patients, with a significant prevalence in fe- and/or physical abuse was present in 7.8% of the sub- males (c2 = 17.4, p < 0.001, not shown in the Table 1). jects, mostly females (c2 = 3.7, p = 0.046). Out of these patients, 59.6% were affected by additional disorders, such as anxiety disorder (21.3%), behavioural SB/SI+NSSI group vs SB/SI without-NSSI group, disorder (19.1%), eating disorders (18.4%), personality bivariate analysis disorders (7.8%), or autism spectrum disorders/intellec- Eighty-two patients (48%) hospitalized for SB or SI tual disability (5.7%). reported a current or lifelong history of NSSI. Amongst Most patients (n = 146; 84.4%) were discharged from inpatients admitted for SB or SI, those with additional hospital with the prescription of a drug therapy, atypical NSSI were: mostly female, more likely to be diagnosed antipsychotics mainly (66.7%), followed by mood sta- with a mood disorder, more frequently presenting with bilizers (28.1%) and antidepressants (14.4%). The per- a comorbid eating disorder, and more likely to be iden- centage of each psychiatric diagnosis did not differed tified as substance abusers compared to subjects with- significantly in subjects admitted to the CPU for SBvs out NSSI. NSSI patients also had higher scores on the SI, except for substance abuse that was more frequently CBCL scales (internalizing symptoms and total behav- identified in the subjects hospitalized for SB (c2 = 4.2, p ioural problems), higher depressive and suicidal score = 0.04) compared to the SI group. on the CDI, and were more frequently discharged from Twenty-one percent (n = 35) of patients were referred hospital with a prescription of second-generation anti- to a post-acute care centre; 25.7% of patients required psychotics (Table 1). at least one or more hospitalization for SB or SI in the Furthermore, comparing SB methods, suicide at- six-year period being evaluated. tempts by jumping out of a window were more frequent 308 Laura Castaldo, Giulia Serra, Simone Piga et al.

Table 1 Factors associated with NSSI in juvenile subjects admitted for SB or SI: bivariate analysis

Factor Group, % or mean ± SD Total NSSI w/o-NSSI p-value N = 71 N = 82 N = 89 Female 66.1 82.9 50.6 < 0.001* Age 15.6±1.6 15.4±1.8 15.8±1.4 0.123 reviews Hospital stay (days) 12.6 14.3 11.1 0.038* Family history of mood disorders 31.8 28.2 35.7 0.327 and

Family history of suicide 8.1 5.1 11.4 0.229 Family history of substance abuse 9.4 11.5 7.0 0.347 Family history of other psychiatric disorder 24.3 22.1 26.8 0.507 articles

Substance abuse 16.4 23.2 10.1 0.021* Previous SB 27.5 26.8 28.1 0.854 Experience of loss/child neglect 24.7 18.5 30.3 0.075

riginal Parental separation/divorce 33.7 37.8 29.9 0.276

O Conflict with parents 43.8 46.3 41.4 0.516 Bullying 15.9 17.1 14.8 0.682 Sexual and/or physical abuse 7.8 9.8 6.0 0.362 Mood disorder diagnosis 82.5 90.2 75.3 0.01* Eating disorder diagnosis 16.4 25.6 7.9 0.002 Psychosis diagnosis 5.9 4.9 6.7 0.749 Anxiety disorder diagnosis 22.8 17.1 28.1 0.086 Behavioural disorder diagnosis 19.9 19.5 20.2 0.907 Disability diagnosis ASD/ID diagnosis 7.0 6.1 7.9 0.651 MD prescription 28.9 25.6 30.3 0.492 AD prescription 14.0 12.2 15.7 0.506 BDZ prescription 20.5 25.6 15.7 0.110 SGA prescription 66.7 75.6 58.4 0.017* MASC total score 57.6±12.7 58.9±13.1 56.3±12.2 0.190 CBCL externalizing symptoms 62.7±10.7 63.3±9.8 62.2±11.6 0.542 CBCL internalizing symptoms 72.2±8.4 74.2±7.5 70.2±8.9 0.004* CBCL total symptoms 68.8±8.6 70.4±8.3 67.1±8.7 0.013* CDI total score 24.4±10.6 27.8±10.1 20.9±9.9 < 0.001* CDI item 9 score 1.3±0.7 1.5±0.6 1.1±0.7 0.001* SB = suicidal behaviour; SI = suicidal ideation; NSSI = non-suicidal self-injury; w/o = without; ASD = Autism Spectrum Disorders ID = intellectual disability; MS = mood stabilizers; AD = antidepressants; BDZ = benzodiazepines; CDI = Children’s Depression Inventory; CBCL = Child Behavioral Checklist; MASC = Multidimensional Anxiety Scale for Children; SGA = second generation antipsychotic. * Variables significantly associated with NSSI (p < 0.05).

in the group without NSSI, while cutting and drug poi- DISCUSSION soning were more frequent in the NSSI group (33.3% The present study investigates the frequency of non- vs 10.5%, c2 = 6.4, p = 0.012; 1.8% vs 23.7%, c2 = 11, p = fatal self-harm in children and adolescents observed at 0.001 respectively, not shown in the table). the ED of the largest Italian children’s hospital high- lighting a growing demand for CP consultations for sui- Multivariate logistic regression model cidal and non-suicidal behaviours amongst adolescents The 3 factors that were significantly and independent- in the six-year period being evaluated. It contributes to ly associated with NSSI through a multivariate logistic the data collection that is aimed at improving the moni- regression model were: a) being female [OR = 5.1 (95% toring, the management and the prevention of this se- CI = 1.9-13.6), p = 0.001]; b) substance abuse [OR = vere public health problem as recommended by WHO 6.0 (95% CI = 1.8-19.7), p = 0.003]; and c) more severe [21]. depressive symptoms [OR = 1.07 (95% CI = 1.02-1.1), The study showed a 10-fold increase in CP consul- p = 0.003; Table 2]. tations for SB, SI and NSSI at the ED involving ado- 309 Suicidal behaviour and non-suicidal self-injury in children and adolescent

Total Female Male SI SB 70

** 60

) %

( * I 50 S nd reviews a

B S

r 40 o f

and d

tt e i m

d 30 a

s t n

e * i t a 20 articles p

n i s r

o *

ac t 10 F sk i riginal R O 0 Family Family Family Family Previous SB NSSI Substance Experience Parental Conflict with Bullying Sexual history mood history history history other abuse of loss/Child separation/ parents and/or disorders suicide substance psychiatric neglect Divorce physical abuse disorders abuse

Figure 3 Risk factors for SB and SI amongst patients admitted to the hospital’s CPU after a CPc at the ED, with a distribution by sex and rea- son for admission (SB vs SI). NSSI and conflicts between patients and their parents were significantly more prevalent in females vs males. Substance abuse was more frequently found in patients admitted for SB when compared to the SI group. SB = suicidal behaviour; SI = suicidal ideation; NSSI = non-suicidal self-injury; ED = emergency department; CPU = Child Psychiatry Unit; CPc = child psychiatric consultation; *p < 0.01; **p < 0.001.

lescents aged between 11 and 17 years (average age: increased rate of suicide in Italian juvenile population. 15 years), with a statistically significant increase over Indeed, suicide rate has decreased in the last 20 years the years (from 12 cases in 2011 to 118 in 2016, p < from 2.6/100 000 person-year in 1994 to 1.4/100 000 0.001; see Figure 2). These findings, in line with results person-year in 2015 [2]. from published epidemiological research studies run in SB was the main reason for CPc at the ED, followed other European countries [5, 10]such as self-cutting, by SI. Additionally, our study showed how SI and SB -burning, -biting, -hitting, and skin damage by other shared the same risk factors, with no significant differ- methods, in representative adolescent samples from 11 ences. This confirms how suicidal phenomena are to be European countries.\\n\\nMETHODS: Cross-sectional regarded as a psychopathological continuum that goes assessment of adolescents was performed within the Eu- from SI to suicidal behaviours and completed suicide ropean Union funded project, Saving and Empowering [33] accurate estimation of suicidal risk remains one of Young Lives in Europe (SEYLE, suggest that self-harm the most difficult and most important tasks that clini- (including both suicidal behaviour and non-suicidal cians face – especially considering recently collected self-injury) is a relevant phenomenon amongst adoles- data showing increase in suicide prevalence in Poland. cents presenting to the ED. This may be due to differ- More thorough estimation of suicidal risk in patients ent reasons: increased sensibility and attention amongst with SI requires taking under consideration not only the general population and within clinical services [30]; suicidal risk factors but also factors that are more spe- an earlier onset of mood and other psychiatric disorders cific for progression of SI to suicidal behaviors (SB. [31]; an earlier use of substances and sexual practices In line with previous findings [10, 34, 35] 62.9% [32] 332 students in grades 7 and 9 provided complete were female and 37.1% male. Overdoses (90.5% of data on episodes of deliberate self-harm in the previ- DSH episodes, more females than males accessed the ous 12 months and pubertal stage. Pubertal stage was ED for suicidal and non-suicidal behaviours, as high- assessed with the Pubertal Development Scale. RE- lighted by the much higher female/male ratio (2.2:1). SULTS: The prevalence of deliberate self-harm was For instance, females, more often than males, engage 3.7% with a more than twofold higher rate in females. in intentional self-harm as a way to deal with relational Late puberty was associated with a more than fourfold issues and emotional distress, and seem to be more in- higher rate of self-harm (odds ratio 4.6, 95% confidence clined to turn to clinical services before and after delib- interval 1.5-14. However, it should be noticed that this erate self-harm [36, 37]. increase of the number of self-harm cases presenting to We found that drug poisoning was the most frequent clinical services apparently does not correlate with an method used by the adolescents admitted to the ED for 310 Laura Castaldo, Giulia Serra, Simone Piga et al.

Table 2 Multivariable logistic regression model of factors associated with NSSI in adolescent subjects admitted for suicidal behaviour or suicidal ideation Variable Univariate logistic regression model Multivariable logistic regression model OR (95% CI) p OR Adj (95% CI) p Sex M 1 [Ref] 1 [Ref] reviews F 4.8 (2.3-9.7) <0.001 5.1 (1.9-13.6) 0.001* Age 1.6 (1-1.4) 0.125 N.I. and Substance abuse No 1 [Ref] 1 [Ref] Yes 2.7 (1.4-6.3) 0.024 6.0 (1.8-19.7) 0.003*

articles Hospital stay (days) 1.0 (1.0-1.1) 0.050

Experience of loss/child 0.5 (0.3-1.1) 0.077 neglect Mood disorder diagnosis N.I. riginal No 1 [Ref] O Yes 3.0 (1.3-7.3) 0.013 Eating disorder diagnosis N.I. No 1 [Ref] Yes 4 (1.6-10.1) 0.003 Anxiety disorder diagnosis 0.5 (0.3-1.10) 0.089 N.I. MASC total score 1.0 (0.9-1.0) 0.190 N.I. CBCL Internalizing symptoms 1.06 (1.02-1.1) 0.006 N.I. CBCL Total symptoms 1.04 (1.01-1.1) 0.025 N.I. CDI Total score 1.07 (1.03-1.1) 0.000 1.07 (1.02-1.1) 0.002* CDI Item 9 score 2.4 (1.5-3.9) <0.001 N.I. SGA prescription No 1 [Ref] Yes 2.2 (1.1-4.3) 0.018 BDZ prescription N.I. No 1 [Ref] Yes 1.8 (0.9-3.9) 0.113 Variables with p < 0.20 in univariate analysis were included in multivariate models. Final models were computed via a stepwise backward procedure (likelihood ratio test, p < 0.05). OR = Odds Ratio crude; OR Adj = Odds Ratio Adjusted for all variable; N.I. not included in the final model with a stepwise backward procedure. *Variables significantly and independently associated with NSSI. a Not included in the logistic regression model because directly linked to selected outcome.

SB, especially by females. Several different studies show Most of the patients (85%) were considered to be that females tend to use less lethal methods such as self- enough severe to require acute psychiatric hospitaliza- poisoning [10, 38], whilst males tend to use more lethal tion due to either a high risk of suicide or serious clini- methods such as hanging or asphyxiation, which are re- cal diagnosis, with more than 90% of the ED visits for ported to be the most common method for completed SB and SI resulting in hospitalization in the CPU, as suicide [10, 39]. NICE guidelines on the management of self-harm be- Most of the attempters (67%) suffered from a mild- haviour in juvenile populations recommend [40]. From to-severe physical injury; hospitalization or intensive this perspective, hospitalization represents a first step care management was required in 22% of cases. Poten- to prevent suicide, as it restricts the subjects’ access to tial lethality was observed to be high in 95.3% of the lethal methods [41-43] and allows for a specialist evalu- patients who did not suffer from any physical injury ation of self-harming adolescents who may have never after SB (e.g., attempted jumping out of a window). accessed clinical care [44]. This suggests how, beyond the degree of the severity of Our study showed that more than a quarter of pa- injuries, all potentially lethal behaviours engaged with tients (26%) underwent one or more readmissions in the intention to die should be carefully evaluated by CPU for SI or SB during the period being evaluated. clinicians. According to published studies [45, 46], rehospitaliza- 311 Suicidal behaviour and non-suicidal self-injury in children and adolescent

tion of children and adolescents with psychiatric disor- ported in a systematic review [57] clinical and taxo- ders is common within one year from first admission, nomic attention in non-suicidal self-injury (NSSI about and self-injury is a strong predictor of readmission. one third of female adolescents with eating disorders Hospital readmission may be due to poor social welfare present with NSSI which, in some cases, occurs after services and politics, inadequate inpatient-outpatient the eating problem has been overcome; on the other treatment continuity and/or the severity of mental dis- hand, NSSI increases the risk for eating disorders to be- orders [46-48]. come chronic. Emotional distancing and harming one’s More than 80% of the adolescents hospitalized in the own body, with coexistence of severe impulsivity and reviews

CPU for SB or SI were diagnosed with a mood disorder emotional dysregulation, seem to play a major role in (bipolar or major depressive disorder), often associated the pathogenesis of these two conditions.

with comorbid anxiety, behaviour, and/or eating disor- In line with findings from a recent study [19], we and ders. This finding is consistent with previous psycho- found that substance abuse was more frequently report- logical autopsy studies that have shown a higher than ed by adolescents hospitalized for SB than for SI. Ad- 90-95% prevalence of psychiatric disorders amongst the ditionally, substance abuse was significantly and inde- people who committed suicide [49]. Additionally, af- pendently associated with NSSI. For instance, frequent articles fective disorders are the most frequent cause of suicide use of substances in adolescence has been associated and have also been shown to be prevalent in adoles- with NSSI [58, 59] S.D.=0.87, and this relationship cents and adults with SB [3, 50]. has been linked to the psychopathological mechanism However, the absence of long-term prospective data of craving NSSI shares with abuse-related pathologies. did not allow us to differentiate between a bipolar dis- NSSI can be interpreted as a coercive behaviour carried riginal order and a major depressive disorder. For instance, out with need and urgency. Indeed, self-injurious behav- O most patients were hospitalized following a first major iours artificially stimulate the opioid system and share depressive episode; it is well known that most early on- with addictive disorders pathologic copying strategies set depressions develop into a bipolar disorder at a later to manage stressors, the development of tolerance, and stage [51, 52]. Indeed, most patients were discharged psychological strain when a self-injurious attempt is not from hospital with the prescription atypical antipsy- achieved [60]. Some evidence suggests that substance chotics (66.7%). In particular, NSSI patients were more abuse is linked to SB and to impulsivity [59]. frequently discharged with the prescription of second- The main limitation of the present study lies in the generation antipsychotics (SGA) compared to patients absence of long-term prospective data and the cross- without NSSI. For instance, SGA seem to have short- sectional/retrospective nature of the data collected. term effectiveness in reducing suicidal risk by improv- However, patients’ data collection was accurate and the ing mixed manic-depressive symptoms that are typical quality of the data high thanks to the hospital online of depression in adolescence and reducing anxiety and system and the involvement of trained psychiatrists in emotional dysregulation associated with NSSI [53]. CP consultations at the ED. Family conflicts were observed in almost half of the Another potential limitation lies in the fact that the patients hospitalized for SB or SI, mainly in females. study only considered visits at the ED. These account This may suggest how the quality of the relationships for a minority of self-harm episodes in adolescents at within family environments and a likely gender-specific the community level [9, 10], hence these data cannot vulnerability to conflicts with parents play an important be generalized. However, since accesses to the ED are role in suicidal behaviours as shown in previous reports the main route by which children and adolescents are [5, 35]. seen at clinical services [44], our findings may be rel- Almost half (48%) of the patients hospitalized for evant to highlight those cases that otherwise would nev- SB or SI reported a current or lifelong history of NSSI er be observed, to whom aftercare interventions could (Figure 3). These patients presented with higher psy- be provided. On the other hand, the high number of chopathological symptoms, in particular, more severe readmissions we observed highlight poor social welfare depressive symptoms and more intense SI, compared services and inadequate inpatient-outpatient treatment to patients without NSSI (Table 2). In line with our re- continuity, suggesting the need to invest more resources sults, previous research studies [18, 19, 54-56]factors in this area. that contribute to the transition from NSSI to suicide ideation and suicide attempts are unclear. To address CONCLUSIONS this gap, we investigated whether demographic charac- The present study showed a statistically significant teristics, child maltreatment, and psychiatric factors are increase in accesses to the ED for SB, SI and NSSI, associated with the level suicidality among adolescents mainly in adolescent females, over the last few years. with a history of self-injury. Participants were three In addition, there was a high prevalence of psychiatric groups of adolescent inpatient self-injurers (n = 397, disorders, mainly depression, in this patient popula- 317 female have shown in adolescents with both NSSI tion. As shown in the study, a much lower number of and suicide attempts higher rates of psychiatric disor- ED accesses was required for NSSI compared to SB, ders, in particular major depression and PTSD, greater although almost half of the patients hospitalized for SB severity of symptoms (e.g., SI, depression, hopeless- or SI reported a current or lifelong history of NSSI. The ness), and greater impulsivity. Moreover, NSSI patients subjects hospitalized for SI or SB, who also presented were more frequently diagnosed with a mood disorder, NSSI, were mainly substance abusers, and presented in most cases with a comorbid eating disorder. As re- with more severe depressive symptoms. 312 Laura Castaldo, Giulia Serra, Simone Piga et al.

In order to improve current treatment and inter- Conflict of interest vention strategies, future studies should focus on dis- The Authors declare that they have no conflict of in- tinguishing between self-harming adolescents and terest. adolescents who are at risk for suicidal behaviours. In addition, actions should be taken to guarantee that Authors’ contribution proper screening is provided to patients presenting for GS and LC designed the study. LC managed the lit- mental disorders and/or substance use. erature searches and analyses. GS and SP undertook As healthcare professionals working in emergency the statistical analysis, and LC wrote the first draft of reviews

departments are often the first who provide primary the manuscript. SV supervised the work. All authors health care to these patients, they need to be able to contributed to and have approved the final manuscript.

and recognize the factors that can increase the risk for at-

tempted or completed suicide. According to WHO Role of the funding source [21], adequate prevention and the treatment of depres- None of the funding sources had any influence on the sion and of substance abuse, as well as the follow-up study design, preparation and submission. of patients who attempted suicide, can reduce suicide articles rates. Increasing awareness of the epidemiological Acknowledgements trends of suicide attempts and self-inflicted injuries is None declared. a first and important step towards the development of effective strategies for preventing repeated or fatal sui- Received on 30 November 2019. riginal cidal behaviours. Accepted on 9 April 2020. O

REFERENCES

1. World Health Organization. Global Health Estimates 11. American Psychiatric Association. Diagnostic and Statis- 2015. Deaths by cause, age, sex, by Country and by Re- tical Manual of Mental Disorders, 5th Edition (DSM-5). gion, 2000-2015. Glob Heal Estim Tech Pap. 2017;1-81. Diagnostic Stat Man Ment Disord. 4th Ed TR. 2013;280. Available from: www.who.int/healthinfo/global_burden_ 12. Nock MK. Self-injury. Annu Rev Clin Psychol. disease/GlobalCOD_method_2000_2015.pdf?ua=1. 2010;6(1):339-63. Available from: www.annualreviews. 2. Istituto Nazionale di Statistica. Giornata Mondiale per la org/doi/10.1146/annurev.clinpsy.121208.131258. Prevenzione del Suicidio. Le statistiche dell’Istat; 2017. 13. Andover MS, Morris BW, Wren A, Bruzzese ME. The co- 3. Hawton K, Saunders KEA, O’Connor RC. Self-harm and occurrence of non-suicidal self-injury and attempted sui- suicide in adolescents. Lancet. 2012;379;2373-82. cide among adolescents: distinguishing risk factors and 4. McMahon EM, Keeley H, Cannon M, Arensman E, psychosocial correlates. Child Adolesc Psychiatry Ment Perry IJ, Clarke M, et al. The iceberg of suicide and self- Health. 2012;6(1):11. Available from: http://capmh. harm in Irish adolescents. A population-based study. Soc biomedcentral.com/articles/10.1186/1753-2000-6-11. Psychiatry Psychiatr Epidemiol. 2014;49(12):1929-35. 14. Hamza CA, Stewart SL, Willoughby T. Examining the 5. Brunner R, Kaess M, Parzer P, Fischer G, Carli V, Hoven link between nonsuicidal self-injury and suicidal behavior. CW, et al. Life-time prevalence and psychosocial cor- A review of the literature and an integrated model. Clin relates of adolescent direct self-injurious behavior: A Psychol Rev. 2012;32(6):482-95. Available from: http:// comparative study of findings in 11 European countries. dx.doi.org/10.1016/j.cpr.2012.05.003. J Child Psychol Psychiatry Allied Discip. 2014;55(4):337- 15. Hawton K, Harriss L. How often does deliberate self- 48. harm occur relative to each suicide? A study of varia- 6. Klonsky ED, Muehlenkamp JJ. Self-injury. A research tions by gender and age. Suicide life-threaten behav. review for the practitioner. J Clin Psychol. 2007;63:1045- 2008;38(6):650-60. Available from: www.atypon-link. 56. com/GPI/doi/abs/10.1521/suli.2008.38.6.650. 7. Hawton K, Harriss L. Deliberate self-harm by under- 16. Sinclair JMA, Hawton K, Gray A. Six year follow-up of 15-year-olds. Characteristics, trends and outcome. J a clinical sample of self-harm patients. J Affect Disord. Child Psychol Psych Allied Discip. 2008;49(4):441-8. 2010;121(3):247-52. 8. Hawton K, Rodham K, Evans E, Weatherall R. Deliber- 17. Wilkinson P, Kelvin R, Roberts C, Dubicka B, Goodyer I. ate self harm in adolescents. Self report survey in schools Clinical and psychosocial predictors of suicide attempts in England. Br Med J. 2002;325(7374):1207-11. and nonsuicidal self-injury in the adolescent depression 9. Madge N, Hewitt A, Hawton K, Wilde EJ De, Corcoran antidepressants and psychotherapy trial (ADAPT). Am J P, Fekete S, et al. Deliberate self-harm within an inter- Psychiatry. 2011;168(5):495-501. national community sample of young people. Compara- 18. Asarnow JR, Porta G, Spirito A, Emslie G, Clarke G, tive findings from the Child & Adolescent Self-harm in Wagner KD, et al. Suicide attempts and nonsuicidal self- Europe (CASE) Study. J Child Psychol Psychiatry Allied injury in the treatment of resistant depression in ado- Discip. 2008;49(6):667-77. lescents. Findings from the TORDIA study. J Am Acad 10. Geulayov G, Casey D, McDonald KC, Foster P, Child Adolesc Psychiatry. 2011;50(8):772–81. Pritchard K, Wells C, et al. Incidence of suicide, hospital- 19. Stewart JG, Esposito EC, Glenn CR, Gilman SE, Prid- presenting non-fatal self-harm, and community-occurring gen B, Gold J, et al. Adolescent self-injurers: Comparing non-fatal self-harm in adolescents in England (the ice- non-ideators, suicide ideators, and suicide attempters. J berg model of self-harm): a retrospective study. The Lan- Psychiatr Res. 2017;84:105-12. cet Psychiatry. 2017;5(2):167-74. Available from: http:// 20. Joiner TE. Why people die by suicide. Cambridge, MA, dx.doi.org/10.1016/S2215-0366(17)30478-9. US: Harvard University Press; 2005. 313 Suicidal behaviour and non-suicidal self-injury in children and adolescent

21. World Health Organization. Preventing suicing. A ple: Characteristics and subsequent mortality in a 20-year global imperative. WHO; 2014. 92 p. Available from: cohort of patients presenting to hospital. J Clin Psychia- www.who.int/mental_health/suicide-prevention/ try. 2007;68(10):1574-83. world_report_2014/en/%0Ahttp://apps.who.int/iris/ 35. Hawton K, Bergen H, Waters K, Ness J, Cooper J, Steeg bitstream/10665/131056/1/9789241564779_eng. S, et al. Epidemiology and nature of self-harm in children pdf?ua=1&ua=1. and adolescents: Findings from the multicentre study 22. Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Mo- of self-harm in England. Eur Child Adolesc Psychiatry. reci P, et al. Schedule for affective disorders and schizo- 2012;21(7):369-77. phrenia for school-age children-present and lifetime ver- 36. Biddle L, Gunnell D, Sharp D, Donovan JL. Fac- reviews

sion (K-SADS-PL): Initial reliability and validity data. J tors influencing help seeking in mentally distressed Am Acad Child Adolesc Psychiatry. 1997;36(7):980-8. young adults. A cross-sectional survey. Br J Gen Pract.

23. March JS, Parker JDA, Sullivan K, Stallings P, Conners 2004;54(501):248-53. and CK. The Multidimensional Anxiety Scale for Children 37. Fortune S, Sinclair J, Hawton K. Help-seeking before and (MASC). Factor structure, reliability, and validity. J after episodes of self-harm. A descriptive study in school Am Acad Child Adolesc Psychiatry. 1997;36(4):554-65. pupils in England. BMC Public Health. 2008;8. Available from: http://dx.doi.org/10.1097/00004583- 38. Sheridan DC, Hendrickson RG, Lin AL, Fu R, Horowitz

199704000-00019. BZ. Adolescent suicidal ingestion: national trends over a articles 24. Kovacs M. Children’s Depression Inventory CDI Man- decade. J Adolesc Heal. 2017;60(2):191-5. ual. New York Multi-Health Syst. 1992;1-800. Available 39. Istituto Nazionale di Statistica (ISTAT). Bilancio Demo- from: www.pearsonclinical.co.uk/Psychology/generic/Ch grafico Nazionale. Report Anno 2015. 2016;1-16. Avail- ildrensDepressionInventory(CDI)/Resources/Technical. able from: www.istat.it. pdf. 40. National Collaborating Centre for Mental Health. The riginal 25. Frigerio A, Pesenti S, Molteni M, Snider J, Battaglia short-term physical and psychological management and O M. Depressive symptoms as measured by the CDI in a secondary prevention of self-harm in primary and second- population of northern italian children. Eur Psychiatry. ary care. Self-harm. 2004;16:1-200. Available from: www. 2001;16(1):33-7. ncbi.nlm.nih.gov/pubmedhealth/PMH0015739/pdf/ 26. Achenbach TM, Dumenci L. Advances in empirically PubMedHealth_PMH0015739.pdf. based assessment. Revised cross-informant syndromes 41. Yip PSF, Caine E, Yousuf S, Chang S Sen, Wu KCC, and new DSM-oriented scales for the CBCL, YSR, Chen YY. Means restriction for suicide prevention. Lan- and TRF: Comment on Lengua, Sadowksi, Friedrich, cet. 2012;379:2393-9. and Fisher (2001). Vol. 69. J Consult Clin Psychology. 42. Zalsman G, Hawton K, Wasserman D, van Heeringen K, 2001;699-702. Arensman E, Sarchiapone M, et al. Suicide prevention 27. Frigerio A, Rucci P, Goodman R, Ammaniti M, Carlet strategies revisited: 10-year systematic review. Lancet O, Cavolina P, et al. Prevalence and correlates of mental Psychiatry. 2016;3(7):646-59. disorders among adolescents in Italy. The PrISMA study. 43. World Health Organization. Mental Health Gap Action Eur Child Adolesc Psychiatry. 2009;18(4):217-26. Program: scaling up care for mental, neurological and 28. Frigerio A, Cattaneo C, Cataldo M, Schiatti A, Molte- substance use disorders. Geneva: WHO; 2008. Available ni M, Battaglia M. Behavioral and emotional problems from: www.who.int/mental_health/mhgap_final_english. among Italian children and adolescents aged 4 to 18 years pdf. as reported by parents and teachers. Eur J Psychol As- 44. Ystgaard M, Arensman E, Hawton K, Madge N, van sess. 2004;20(2):124-33. Available from: http://econtent. Heeringen K, Hewitt A, et al. Deliberate self-harm in hogrefe.com/doi/abs/10.1027/1015-5759.20.2.124. adolescents. Comparison between those who receive 29. Posner K, Brown GK, Stanley B, Brent DA, Yershova K help following self-harm and those who do not. J Adolesc. V, Oquendo MA, et al. The Columbia-suicide severity 2009;32(4):875-91. rating scale. Initial validity and internal consistency find- 45. Blader JC. Symptom, family, and service predictors of ings from three multisite studies with adolescents and children’s psychiatric rehospitalization within one year adults. Am J Psychiatry. 2011;168(12):1266-77. of discharge. J Am Acad Child Adolesc Psychiatry. 30. Merikangas KR, He J, Burstein M, Swendsen J, Avenevoli 2004;43(4):440-51. S, Case B, et al. Service utilization for lifetime mental dis- 46. van Alphen NR, Stewart JG, Esposito EC, Pridgen orders in U.S. adolescents: results of the National Comor- B, Gold J, Auerbach RP. Predictors of rehospitaliza- bidity Survey-Adolescent Supplement (NCS-A). J Am tion for depressed adolescents admitted to acute psy- Acad Child Adolesc Psychiatry. 2011;50(1):32-45. Avail- chiatric treatment. J Clin Psychiatry. 2017;78(5):592- able from: http://dx.doi.org/10.1016/j.jaac.2010.10.006 8. Available from: www-ncbi-nlm-nih-gov.ezproxy. 31. Avenevoli S, Swendsen J, He JP, Burstein M, Merikan- library.tufts.edu/pmc/articles/PMC5313382/pdf/ gas KR. Major depression in the national comorbidity nihms839285.pdf%0Ahttp://www.ncbi.nlm.nih.gov/ survey–adolescent supplement: prevalence, correlates, pubmed/27529444%0Ahttp://www.pubmedcentral.nih. and treatment. J Am Acad Child Adolesc Psychiatry. gov/articlerender.fcgi?artid=PMC5313382. 2015;54(1):37-44.e2. 47. James S, Charlemagne SJ, Gilman AB, Alemi Q, Smith 32. Patton GC, Hemphill SA, Beyers JM, Bond L, Toum- RL, Tharayil PR, et al. Post-discharge services and psychi- bourou JW, McMorris BJ, et al. Pubertal stage and delib- atric rehospitalization among children and youth. Adm erate self-harm in adolescents. J Am Acad Child Adolesc Policy Ment Health. 2010;37(5):433-45. Psychiatry. 2007;46(4):508-14. 48. Daniel SS, Goldston DB, Harris AE, Kelley AE, Palmes 33. Rodzin´ski P, Rutkowski K, Ostachowska A. Progression GK. Review of literature on aftercare services among chil- of suicidal ideation to suicidal behavior from a perspec- dren and adolescents. Psychiatr Serv. 2004;55(8):901-12. tive of selected suicidological models. Psychiatr Pol. Available from: http://ps.psychiatryonline.org/article. 2017;51(3):515-30. Available from: http://psychiatriapol- aspx?articleID=88917 ska.pl/515530_.html. 49. Cavanagh JTO, Carson AJ, Sharpe M, Lawrie SM. Psy- 34. Hawton K, Harriss L. Deliberate self-harm in young peo- chological autopsy studies of suicide: A systematic re- 314 Laura Castaldo, Giulia Serra, Simone Piga et al.

view. Psychol Med. 2003;33:395-405. 56. Dougherty DM, Mathias CW, Marsh-Richard DM, Pre- 50. Houston K, Hawton K, Shepperd R. Suicide in young vette KN, Dawes MA, Hatzis ES, et al. Impulsivity and people aged 15-24. A psychological autopsy study. J Af- clinical symptoms among adolescents with non-suicidal fect Disord. 2001;63(13):159-70. self-injury with or without attempted suicide. Psychiatry 51. Akiskal HS. Developmental pathways to bipolarity. Are Res. 2009;169(1):22-7. juvenile-onset depressions pre-bipolar? J Am Acad Child 57. Cucchi A, Ryan D, Konstantakopoulos G, Stroumpa Adolesc Psychiatry. 1995;34(6):754-63. S, Kaçar AS, Renshaw S, et al. Lifetime prevalence of 52. Kochman FJ, Hantouche EG, Ferrari P, Lancrenon S, non-suicidal self-injury in patients with eating disorders: Bayart D, Akiskal HS. Cyclothymic temperament as a A systematic review and meta-analysis. Psychol Med. reviews

prospective predictor of bipolarity and suicidality in chil- 2016;46:1345-58. dren and adolescents with major depressive disorder. J 58. Giletta M, Scholte RHJ, Engels RCME, Ciairano S,

and Affect Disord. 2005;85(1-2):181-9. Prinstein MJ. Adolescent non-suicidal self-injury: A 53. Tondo L, Baldessarini RJ. Suicidal behavior in mood dis- cross-national study of community samples from Italy, orders. Response to pharmacological treatment. Curr the Netherlands and the United States. Psychiatry Res. Psychiatry Rep. 2016;18(9):88. 2012;197(1-2):66-72. 54. Muehlenkamp JJ, Ertelt TW, Miller AL, Claes L. Bor- 59. Rodríguez-Cintas L, Daigre C, Grau-López L, Barral C,

articles derline personality symptoms differentiate non-suicidal Pérez-Pazos J, Voltes N, et al. Impulsivity and addiction and suicidal self-injury in ethnically diverse adolescent severity in cocaine and opioid dependent patients. Addict outpatients. J Child Psychol Psychiatry Allied Discip. Behav. 2016;58:1049. 2011;52(2):148-55. 60. Victor SE, Glenn CR, Klonsky ED. Is non-suicidal self- 55. Brausch AM, Gutierrez PM. Differences in non-suicidal injury an “addiction”? A comparison of craving in sub- riginal self-injury and suicide attempts in adolescents. J Youth stance use and non-suicidal self-injury. Psychiatry Res. O Adolesc. 2010;39(3):233-42. 2012;197(1-2):73-7. 1 10126 Turin, Italy. [email protected]. Address forcorrespondence: SilviaGonella,Azienda Ospedaliero-UniversitariaCittà dellaSaluteeScienzadiTorino, CorsoBramante88-90, palliative-oriented care[3,4]. play apivotalroleinthetransitionfromcurative-to lished withhealthcareprofessionals(HCPs)[2]and end-of-life carewhenasharedcomfortgoalisestab- nesses-related deaths[1]. of dying from acutedeaths to chronic progressive ill- more appropriate,duetotheshiftinepidemiology fort andqualityoflife–asinpalliativecarebecoming at theend-of-lifewithcareaimedimprovingcom- ing home(NH)requiresadjustmentsofthecareplan INTRODUCTION standing andshareddecision-making. family carers’ preferences for care and treatment at the end-of-life, and increases under Conclusion. Clearandin-depthcommunicationprovidesinsightintoresidents’ preferences. ing knowledgeofresidents’preferences;andd)improvingfamilycarers’ cision-making betweenhealthcareprofessionalsandresidents/familycarers;c)improv- relative’s healthconditions,prognosis,andtreatmentsavailable;b)fosteringsharedde- tive-oriented carewereidentified:a)promotingfamilycarersunderstandingabouttheir Results. Fourmechanismsbywhichend-of-lifecommunicationcontributedtopallia- with inductivecontentanalysiswasadopted. prior from13differentNHs.Atwo-stepsanalysisprocessfirstlywithdeductiveandthen with 32 bereavedfamily carers whose relative had died between45days to 9months Methods. A descriptive qualitativedesign was performed. Interviews were conducted perspective ofbereavedfamilycarers. end-of-life communicationmaycontributetopalliative-orientedcareinNHfromthe munication mayworkisstillunknown.Therefore,weexploredthemechanismsbywhich recognized topromotepalliative-orientedcareinnursinghome(NH),howthiscom- Background. Althoughfamily-centeredcommunicationaboutend-of-lifecarehasbeen Abstract 3 2 Silvia Gonella in nursinghome contributes topalliative-orientedcare on howend-of-lifecommunication A qualitativestudyoffamilycarersviews DOI: 10.4415/ANN_20_03_09 Ann IstSuperSanità2020|Vol. 56,No.3:315-324 tention totheresident’s physicalneedsandtothepsy - participate inthedecision-making process,andpayat- changing conditions, to allow each person to actively ers shoulddevelopovertimein relationtotheresident’s Dipartimento diScienzedellaSanitàPubblicaePediatriche,UniversiàdegliStudiTorino, Turin, Italy Azienda OspedalieroUniversitariaCittàdellaSaluteeScienzadiTorino, Turin, Italy Dipartimento diBiomedicinaePrevenzione,UniversitàdegliStudiRomaTor Vergata, Rome,Italy Clear communicationbetween HCPsandfamilycar Family carers of NH residents are more satisfied with An increasingproportionofresidentslivinginnurs- 1,2, , InesBasso 3 , MarcoClari 3 andPaolaDiGiulio - tion andaboutwhattoexpect duringthedyingprocess. report negativeperceptions about clarityofinforma- poor [10].Previousauthors[11] foundfamilycarersto of-life care plans is often missed [9] and its quality is the opportunity to establish communication about end- likely to have acomfort care goal [6].Unfortunately, satisfied withphysiciancommunicationweremore ented care[8]andresidentswhosefamilycarerswere family carersfacilitatedtheprovisionofpalliative-ori- ing the time to discuss and establish a partnership with and acknowledgmentoftheircarerole[7].HCPstak- ment offamilycarersincare,assessmenttheirneeds, with familycarersaboutcareandprognosis,involve- stresses theimportanceofcontinuouscommunication life careimprovesthequalityofdying[6],literature carers [5]. chological andsocialneedsofresidentstheirfamily Since family-centered communication about end-of- 3 - • • • • • • Key words qualitative research decision-making communication palliative care nursing homes family

315 Original articles and reviews 316 Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

Poor communication negatively influenced family car- Table 1 ers experience with end-of-life care and was associated Demographics of family carers of nursing home residents (n = with poor resident outcomes [12]. 32) Although end-of-life communication has been recog- Family carers N (%) nized to promote palliative-oriented care in NH [13], Female gender 20 (62.5) how communication may work is still unknown. Hence, this study aimed at exploring the mechanisms by which Age, years, mean [range] 58 [35-71] end-of-life communication contributes to palliative- Education

reviews Middle 5 (15.6) oriented care in NH from the perspective of bereaved family carers. Secondary/university 27 (84.4)

and Marital status

METHODS Married/cohabitant 22 (68.8) Widowed/single 6 (18.7) Study design Divorced/separated 4 (12.5) A descriptive qualitative study was carried out. Employment The COnsolidated criteria for REporting Qualitative

articles Full-time 14 (43.8) studies (COREQ) guidelines were followed to report Retired 11 (34.4) the methodological aspects [14] (Supplementary Table 1, Housewife 4 (12.5) available online). Part-time/freelance 3 (9.3) Relationship to the resident riginal Setting, participants and sample selection Adult child 19 (59.4) O Fifty-two NHs were purposively sampled for geo- Nice/nephew 7 (21.9) Daughter-in-law 3 (9.4) graphical area and different size in Piedmont – north- Other† 3 (9.4) west Italy – to ensure the greatest possible variation of Frequency of visiting data and 20 adhered to the study on a voluntary basis. Daily 12 (37.5) Family carers were eligible if a) they were willing to Three/four times a week 5 (15.6) participate, and their relative b) spent the last 30 days Two/three times a week 13 (40.6) of their life in NH and c) had died in NH from 45 days Less than once a month‡ 2 (6.2) to 9 months prior the study started. The NH director Main contact person with the help of the direct staff purposefully identified Interviewee and one or more family carers 20 (62.5) Interviewee only 11 (34.4) bereaved family carers to include information-rich cases Interviewee and social worker 1 (3.1) related to the phenomenon under study (i.e., end-of-life communication). Family carers were then contacted ac- †Sister-in-law (1), cousin (1), wife (1). ‡Every 45 days (1); every 3 months (1). cording to the NH preferences: i.e. a) by phone call with preliminary contact by the director to inform of the study and request permission for contact by the explore emerging topics [16]. Box 1 provides an over- research team, b) interview directly scheduled by the view of the final interview guide. director, or c) a personalized letter of condolence with All interviews were digitally audio-recorded. a brief presentation of the project and the researchers’ Additional data were collected about: phone number. a) residents’ demographics and clinical information In all, 32 family carers across 13/20 adhering NHs (clinical records, Table 2). participated in the study. b) NHs’ profile and their working processes. NH di- Most family carers were female (n = 20) and the rectors filled in a semi-structured questionnaire that mean age was 58 years [range 35-71]. The relation with explored: i) NH organizational features (e.g., public the resident was adult child (n = 19), nice/nephew (n = or private profile, bed size,Table 3); ii) frequency of 7), daughter-in-law (n = 3), others (n = 3) (sister-in-law, end-of-life communication between NH staff and cousin, and wife), and the majority visited their relative family carers during the last 6 months on a 5-points daily (n = 12) (Table 1). Likert scale (never to always); and iii) activation of Family carers’ interview was recorded after a median the palliative care service during the last 6 months of 106 days (interquartile range 68-175) after their rela- (i.e., number of and reasons for activations). tive’s death. Two family carers had their interview over Data were collected from December 2018 to May telephone and three interviews were in double (two in- 2019. terviews with child and daughter-in-law, and one inter- view with two nieces). Mean duration of interviews was Transcription and analysis 38 minutes (SD 22.6). S.G. and I.B. transcribed the interviews verbatim. M.C. checked a random sample of transcripts for accu- Data collection racy. Analysis of the interview transcripts started shortly Semi-structured in-depth interviews [15] with be- after each interview. reaved family carers of NH residents were conducted The analysis process consisted of two stages: by two researchers (S.G. and I.B.). An open-ended in- 1. data were analysed using a deductive content analysis terview technique with follow-up questions related to [17] (i.e., theory-driven approach where codes de- the participants’ answers was employed. The interview rived from data are organized in an a-priori frame- guide was refined after the first 10 interviews to better work) referring to a communication framework that 317 How communication contributes to palliative care

Box 1 Interview guide 1. Please, can you tell me how your relative died? Have her/his health conditions gradually or suddenly worsened? 2. Have you talked about the care provided to your relative with the nursing home staff or the physician during her/his last week of life? Can you tell me a positive episode and a negative episode about communication with the nursing home staff or the physician? 3. The week before your relative’s death, have you thought that she/he would be died within 7 days? Why? 4. Have you felt involved in planning for care during the last week of life of your relative? How the nursing home staff/physician promoted your involvement? reviews

5. Have you talked about preferences for care at the end-of-life with your relative when she/he was at home or during her/his nursing home stay?

6. Do you know whether your relative talked about her/his preferences for care at the end-of-life with the nursing home staff/physician? and Have any meetings about preferences for care and treatments at the end-of-life taken place between HCPs and the resident or HCPs, resident and family carers? 7. Have the nursing home staff or the physician asked you which type of care or treatment you would have desired at the end-of-life for your relative?

8. Did your relative suffer from distressing symptoms (e.g, pain, difficulty in breathing) during her/his last week of life? Are you satisfied with articles the management of symptoms? 9. Feel free to add whatever. HCPs, Healthcare professionals. riginal O Table 2 Table 3 Demographics of nursing home residents (n = 32) Nursing home characteristics (n = 13) Residents N (%) Characteristics N Female gender 23 (71.9) Profile Private 11 Age, years, mean [range] 88.8 [75-99] Public 2 Length of residence, months, median [IQR] 18 [5-36] Bed size, mean (SD) 78 (34) Treatments in the last week of life (n = 133) ≤ 60 4 Antibiotics 12 (9.0) 61-119 6 Curative-oriented treatments 71 (53.4) ≥ 120 3 Intravenous hydration 24 (18.0) Beds for functionally independent residents, 25 (16) Blood test 12 (9.0) mean (SD) Peripheral cannulation 11 (8.3) Call to the emergency services 6 (4.5) Beds for functionally dependent residents, 53 (29) Bladder catheter 6 (4.5) mean (SD) Access to emergency department 5 (3.8) Other† 7 (5.3) Number of death per year, mean (SD) 20 (9) Palliative-oriented treatments 50 (37.6) Beds for hospice care in the structure, mean (SD) 0 Oxygen therapy 15 (11.3) Assessment of pain/discomfort 13 (9.8) Alzheimer unit 2 Hypodermic hydration 11 (8.3) SD, Standard deviation. Oral therapy withdrawal 4 (3.0) Opioid analgesics 4 (3.0) Other‡ 3 (2.3) 2. data not fitting this categorisation frame were anal- Cause of worsening conditions (n = 46) Pulmonary infection 11 (23.9) ysed through inductive content analysis (i.e., new cat- Severe dehydration 10 (21.7) egories and themes emerging by data supplemented Worsening of dementia 7 (15.2) the original framework) [17]. Urinary infection 6 (13.0) Analysis was aided by the software ATLAS.ti 6.2. Worsening of neurodegenerative disease 4 (8.7) Other§ 8 (17.4) Themes are illustrated by participants’ quotations that are identified by an alphanumeric code to ensure Cause of death Cardiac arrest 17 (53.1) confidentialy (e.g., NH1, FC1). NH refers to the facil- Cachexia 6 (18.7) ity where the participant’s relative died, FC refers to Dementia 3 (9.4) the interviewee. The progressive numbers indicate the Sepsis 3 (9.4) order in which facilities and participants were recruited. Other organ failure 3 (9.4) Details of the analysis process are provided in Supple- †Glycemic monitoring (2), parenteral nutrition (2), enteral nutrition (1), mentary Table 2 (available online). cardiopulmonary resuscitation (1), hospitalization (1) ‡Oral aspiration (2), palliative sedation (1) §Fall (two with trauma and one without complications), ab ingestis (2), Trustworthiness worsening of respiratory disease (2), pressure ulcer infection (1) The process of continuous self-reflection known as re- IQR, interquartile range. flexivity which was aimed at improving transparency in the researcher’s subjective role [19], was sought through had previously described the potential mechanisms repeated discussions within the team about alternative by which end-of-life communication may contribute interpretations of the results. Repeated reading of the to palliative-oriented care in NH [18]; interview transcripts and repeated discussions with co- 318 Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

authors of emerging categories, themes and illustrative with the physician was fortuitous and often family car- extracts were undertaken to validate the findings. An ers had to look for information by phone or going to audit trail was kept throughout all the analysis process. the physician’s office to be updated. No interviewee was Further strategies to improve trustworthiness are de- asked about what they would have desired for their rela- tailed in Supplementary Table 1 and 2 (available online). tive’s end-of-life care by any HCPs: - “The communication was direct, open, clear and got Ethics to the point” (NH4, FC9); The study was approved by the Ethics Commit- - “I never talked to the physician, I saw him only the reviews

tee of the University of Torino (Italy, reference day of my mother-in-law’s death” (NH11, FC28). 457626/10.12.2018). All participants gave their writ-

and ten informed consent after receiving oral and written Family carers reported both barriers and facilitators of

information about the aim of the study. All participants end-of-life communication. Among the former, different could stop the interview at any time and for any rea- linguistic and cultural background and limited family de- sons, and a protocol to manage the onset of emotional sire to ask were frequently reported; among the latter, distress was available. family carers reported their recognition of changes in articles

The timeframe of family carers’ recruitment was con- their relative’s health and the characteristics of the set- sistent with previous research [20], to reduce the emo- ting where communication took place. They appreciated tional burden on participants while avoiding recall bias. being provided information in a private environment Data collections forms about residents’ demograph- (e.g., nursing office, NH director office) or in places riginal ics and clinical information were anonymous with resi- closed to the relative’s room (e.g., bedside, corridor): O dents identified with a progressive number, which was - “I didn’t ask, I didn’t go too far in asking” (NH10, paired with their family carer’s interview. Data collec- FC24); tion forms as well as the transcribed interviews were - “I noticed that my uncle had some difficulties in stored in the archive of the University of Torino and breathing, therefore I asked the nurse to measure his could be accessed only by the research team. vital signs … thus I recognized the problem and acti- vated the process that led to his physician’s visit [...]. RESULTS Then, talking to the physician, we started to personal- Family carers’ experience ize my uncle’s therapy” (NH5, FC12). The original communication framework [18] was up- dated and finally consisted of six themes and 21 catego- Theme 2. Gradual shift to palliative-oriented care ries that describe family carers’ perspective about the Most family carers reported that end-of-life care was mechanisms by which end-of-life communication may palliative-oriented rather than curative-oriented, with contribute to palliative-oriented care in NH (Figure 1, oral therapy usually stopped and their relative sustained Supplementary Table 3 available online). Specifically, (1) by hypodermic hydration. Many interviewees stated HCP-resident and HCP-family carers end-of-life com- that their relative received psychological support, spir- munication may promote or hinder a (2) gradual shift itual support, had basic needs fully fulfilled and died to palliative-oriented care by acting on (3) family car- without distressing symptoms: ers understanding about their relative’s health condi- - “During the last week she was hydrated subcutane- tions, prognosis, and treatments available, (4) shared ously” (NH1, FC3); decision-making between HCPs and residents/family - “I don’t think she was in pain because she never com- carers, (5) knowledge of residents’ preferences, and (6) plained, she went out just like a candle”. (NH3, FC7). knowledge of family carers’ preferences. A large proportion of family carers felt emotionally Theme 1. Healthcare professional-resident supported by all those working in the NH and were re- and healthcare professional-family carers end-of-life laxed when coming back home to rest since they were communication sure to be called if their relative conditions should have Several HCPs were involved in end-of-life communi- worsened. They appreciated to be called in time for cations, including nurses, head nurses, nurse aides, phy- sharing the last moments with their relative: sicians and NH directors, although few family carers - “Here I had a practical and moral support, especially where engaged in multiprofessional communications. moral” (NH2, FC5); Communication was both verbal and non-spoken, - “It has been very important for us to spend some mo- generally face-to-face even if communication over ments more with our father when he was still lucid, phone was largely employed when compelling issues above all because we lived far […]. They allowed us, needed to be discussed with family carers or when fam- they warned us in time, they let us ... they gave us ily carers lived far. the time, they allowed us to say him goodbye” (NH1, End-of-life communication was often described as FC3). clear, honest and constant, but often occurring late in the dying process (usually one week before death, range However, those family carers that perceived poor one month to few days). Moreover, while communica- information and involvement in end-of-life, they also tion with nurses and nurse aides was generally frequent, complained of having missed their relative’s death since many interviewees were dissatisfied that the physician they were not called in time: was rarely or never seen in the NH. Communication - “They called me 3 minutes before she died… my work 319 How communication contributes to palliative care

Family cares’ Family carers Shared decision-making Resident’s preferences known understanding between HCPs and preferences known Low intensity Information received residents/family carers Family carers’ of care Acknowledgement Resident involvement in knowledge and Priority for of death as an expected end-of-life care decisions assumption quality of life event of human being Family involvement in Staff’s knowledge Awareness/unawareness end-of-life care decisions reviews of impending death Family advocacy Family guidance

Level of trust and Relief from avoiding decisions Shared decisions with HCPs and family unit articles

HCP-resident and HCP-family carers end-of-life communication Delayed communication riginal

Supportive communication O Absent or poor communication Communication barriers and facilitators HCPs involved in communication Gradual shift to palliative-oriented care At the resident level At the family level

Figure 1 Mechanisms by which end-of-life communication contributes to palliative-oriented care in nursing home according to family car- ers perspective. HCP, Healthcare professional.

place is 3 minutes far from the NH, I would have the nurse told me ‘Do not worry, she will recover’, wanted to be there. Five minutes before would have I answered ‘Please, don’t joke about these things!’ ” been enough… I’m still suffering…” (NH12, FC31). (NH11, FC25); - “All of us, family carers, nurses and nurses’ aides, Pathways of successful end-of-life communication understood that my mother was going to pass away” Four pathways by which end-of-life communication (NH13, FC32); may contribute to palliative-oriented care emerged: - “No one here expected my mother’s death, if you’ll talk to the nurses, they’ll tell you that we are still all Theme 3. Family carers understanding shocked” (NH12, FC29). Most family carers reported being constantly updat- ed about their relative’s health conditions and therapy. Family carers’ awareness was characterized by the They were informed by several HCPs, including the expectation that their relative’s death would occur in NH director, the NH staff, and the physician: a short time due to worsening conditions. Some in- - “I’ve never had any problems to know what was hap- terviewees reported to find peace when they acknowl- pening to my mom ... whatever persons I talked to, in edged the upcoming death: the office, nurses or nurse aides, all of them informed - “My mom’s health conditions got worse in the last me about how she was going on” (NH2, FC5). three months before death, it was a slow process that lasted two years” (NH10, FC24); Most interviewees understood that their relative was - “I felt relieved after deciding to let her go and told coming to the end of his/her lifespan in advanced old my children ‘Tomorrow I’m shopping new clothes for age, but this understanding was not always associated grandma, thus she will be nicely dressed when meet- with the awareness of impending death. Three scenari- ing my [dead] dad” (NH6, FC15). os concerning awareness emerged: i) family carers were aware of the changes but perceived HCPs did not share Family carers could become aware very late (i.e., few it with them; ii) both family carers and HCPs were days before death) with differences among family mem- aware; and iii) neither were aware. bers: - “Her conditions were severely deteriorated … when - “On Monday, I would have never said that Sunday all 320 Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

would have been over, I was aware that some symp- pass away quietly, little by little, and avoid hospital- toms got worse but I thought that he could go on a ization […] we felt supported in the decision” (NH5, little bit more” (NH8, FC19); FC11); - “My sister took longer than me to accept that ... I - “I’m not a physician, I don’t have any medical com- know she kept on asking ‘what can be done” (NH1, petence, so I trusted what they advised me” (NH9, FC3). FC22).

Nurses were the HCPs that more frequently support- When decisions were shared with HCPs, family car- reviews

ed family carers in developing awareness of impending ers felt supported, whereas absence of sharing foster a death: feeling of loneliness and the burden of responsibility:

and - “The nurse explained me that it was not phlegm - “I would have shared with the physician the decision

but a rattle that came from inside: the kidneys were to hospitalize my father, because sometimes we use- blocked, she did not pee anymore and water was ris- lessly accessed the Emergency Department … I felt ing, thus provoking this rattle” (NH6, FC14). quite alone and responsible for decisions” (NH8, FC19). articles Theme 4. Shared decision-making between healthcare professionals and residents/family carers The timing of information was a pivotal element to Family carers reported that their relative was rarely perceive involvement in care decisions; receiving infor- involved in decisions concerning end-of-life treatments, mation before care decisions had been taken, positively riginal such as adjustments of the analgesic therapy or the de- impacted on perceived involvement, while post hoc in- O cision to access the emergency department. However, formation were associated with the perception of poor this usually was a family carers’ desire: involvement: - “My mom was cognitively competent, we couldn’t say - “They never did anything without they told it me be- ‘Mommy, you are going towards the end of your life’, fore” (NH13, FC32); she was 86 years and would have felt bad if hearing - “They told me about the change after it was already these things” (NH11, FC27). implemented” (NH10, FC24).

Particularly in the case of dementia, family carers Sharing decisions to withdraw or withhold aggressive took decisions for their relative, but reported to need treatments among family carers was also common, even guidance in deciding to withdraw or withhold aggres- if discussions about the desired intensity of care usually sive treatments, and needed a confirmation by HCPs took place too late: that the decision was right: - “I never asked my brother what he thought about it - “The physician told us ‘You should not forget that … that day [the day before death], I said ‘I would be the aim is not to heal him but accompany him, you of this idea’, and my brother also stated ‘you are right, should accept that he does not want to eat anymore” these treatments are too aggressive and without any (NH1, FC3). meaning” (NH6, FC15). - “When I said ‘I don’t want aggressive treatments’, the nurse looked at me and nodded. I felt supported” Theme 5. Resident’s preferences known (NH2, FC5). Many family carers reported that their relative was reluctant to talk about her/his desired end-of-life care Others believed that HCPs were in charge to take and treatments, thus, they had to make assumptions; decisions, tended to hand over the responsibility to de- instead, they felt released if the relative had left written cide and felt relieved when they did not have to decide. preferences: Trusting NH staff was recognized as a way to hand over - “I would have preferred something written, I did it, treatment decisions; family carers usually trusted both because it seems to me to relieve my loved ones from the NH staff and the physician about treatments their any kind of choice” (NH10, FC23). relative was in need of and did not feel the need to look for alternative source of information: According to the family carers’ perspective it was un- - “I think that in these structures, they [NH staff] must likely that their relative had shared her/his preferences tell me what needs to be done” (NH7, FC17); for end-of-life care with HCPs. Few interviewees were - “I trusted them [NH staff], what they said and what sure that HCPs were aware of their relative’s end-of-life they did and therefore I didn’t feel responsible for tak- wishes: ing decisions for my mother-in-law” (NH6, FC13). - “They [NH staff] were absolutely informed, they were well aware… we still talked about it even if, it can look Decisions concerning hydration therapy, drug ther- ugly, we even smiled each other” (NH10, FC23). apy, pain management and above all hospitalization were usually shared between family carers and HCPs, Theme 6. Family carers’ preferences known although few family carers felt active partners in their A substantial number of family carers stated that relative’s care due to their perceived poor medical com- they preferred not aggressive care for their relative and petence that made them feel informed rather than in- shared their opinion with the NH staff, particularly volved: nurses. The priority was to avoid suffering, to improve - “We decided with them [NH staff] to allow him to the quality of their relative’s remaining life: 321 How communication contributes to palliative care

- “I told them [nurses] ‘My mother could not continue honest communication is essential to prepare for death to be bombed with antibiotics ... Thus, why persisting [26]. Instead, poor or ambiguous communication could in therapies? Let her to go …” (NH2, FC5); make understanding more difficult and increase the risk - “The most important thing is that she does not suf- of short awareness time [30]. Although end-of-life com- fer. I shared my opinion with the NH staff” (NH11, munication plays the most prominent role in promoting FC25). family carers understanding, it is one of the most ne- glected aspect of end-of-life care with poor or fortuitous End-of-life communication and activation communication and no or rare meetings with the physi- reviews of the palliative care service cian being not uncommon [24, 28, 29]. This is unfortu- During the last 6 months, 9/13 NHs discussed with nate since the link between end-of-life communication

family carers the opportunity to activate the palliative and improved end-of-life care quality has been demon- and care service while the others did not offer this oppor- strated [13]. Our findings suggest that communication tunity (Supplementary Table 4 available online). The NH should be a dynamic process and start as early as pos- director reported that resident’s health conditions un- sibile in the disease trajectory, rather than being an iso- responsive to curative treatments, uncontrolled pain, lated or occasional event limited to the last days of life. articles and the desire to provide good end-of-life were the rea- Therefore, HCPs should provide family carers timely in- sons that led to the activation of palliative care service, formation about their relative’s clinical course and prog- which was the result of family carers’ choice, physician’s nosis to promote understanding and to offer emotional choice, a shared decision between family carers and support, since high-quality palliative care should be both physician or due to the resident’s denial of any curative resident-centered and family-focused [31]. riginal treatments. Several elements of a shared decision-making O Hospice referral was discussed less frequently than emerged: family carers judged particularly worthy to the activation of the palliative care service (Supplemen- feel involved in end-of-life care decisions and greater the tary Table 4 available online). amount of information received, higher the perception of involvement. Our family carers generally desired to DISCUSSION remain involved in care planning and decision-making This study found that end-of-life communication may [5], and were an important partner in establishing the contribute to palliative-oriented care in NH according care goal particularly in the case of dementia [32]. Thus, to the family carers perspective by: a) promoting fam- encouragement to be part of the caring team reinforced ily carers understanding, b) fostering shared decision- their role as contributors to residents’ care. Several fam- making between HCPs and resident/family carers, c) ily carers were involved in their relative’s care through improving knowledge of resident’s preferences, and d) a constant monitoring of the basic care provided and improving knowledge of family carers’ preferences. by asking for physical, psychological and spiritual sup- Family understanding was mainly sustained by the port for their relative. This suggests that family advo- amount of information received and the awareness for cacy may improve the quality of end-of-life care [29], impending death. Particularly, our findings suggest that although family carers needed HCPs guidance towards residents were more likely to receive palliative-oriented the best care options for their relative [28]. However, care at their end-of-life when family carers understood most interviewees felt informed rather than involved the prognosis and clinical course of the disease [21-23]. and received information about changes in their rela- Consistently with previous quantitative studies [24], tive’s treatment after they were already implemented. most family carers were prepared for their relative’s This confirmed previous literature that highlighted poor death. However, the acknowledgement that the old family involvement with little substantive communica- relative was going to the end-of-life was often not suffi- tion regarding end-of-life care planning [33], only one cient to promote understanding [25]. We found aware- out of five family carers involved in the plan meetings ness to have both a cognitive dimension (i.e., need of [34] and only half of the decisions to withdraw or with- clinical information) and an affective dimension (i.e., hold treatments routinely discussed [35]. emotional preparation) [26]: one could be very knowl- Supportive end-of-life communication promotes trust edgeable about medical aspects but not emotionally and partnership between family carers and HCPs [3], prepared. This multidimensional nature of awareness in that is valued as much if not more than care activities of addition to the unpredictable trajectory of the illness daily living [36]. High levels of trust make family carers [27] may explain why understanding was often delayed to feel emotionally supported [31] and are associated and why some family carers were unprepared for their with positive experiences of HCPs-family communica- relative’s death. No or poor understanding poses a seri- tion [37], while lack of trust could be one of the reasons ous problem to provide optimal end-of-life care, since family carers may not accept prognostic information it may lead to an unaware decision-making and hinder about their relative [38]. Therefore, improving com- optimal care planning. Indeed, failure to implement a munication may increase family carers’ trust in HCPs timely plan of care was identified as an obstacle to pal- and provide a good basis to activate palliative care, also liation and end-of-life care [8]. avoiding the burden of decisions by handing over the In accordance with previous literature [26, 28, 29], final decision to HCPs [39]. In fact, if the shared de- our inteviewees expressed the need to be constantly up- cision-making steps have taken place, HCPs will know dated about their relative’s health conditions and treat- what the resident’s wishes are, and the care approach ments to develop understanding. Clear, frequent, and will be proposed accordingly [40]. In contrast, delayed 322 Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

or poor communication both between family carers and an in-depth picture of end-of-life communication expe- HCPs or within the family unit hinders shared decision rience in NH [46]. making [8]. Therefore, educational initiatives aimed at improving communication skills should be regularly of- CONCLUSIONS fered to HCPs. Clear communication between HCPs and residents/ Our study confirmed that resident’s involvement in family carers is essential to provide quality care, par- decisions was extremely limited, thus further increasing ticularly at the end-of-life. In-depth and thorough com- family carers perceived burden because of taking deci- munication that promotes family carers understanding reviews

sions on behalf of their relative [41]. Instead, literature about their relative’s health condition, prognosis and suggests that when residents were engaged in advance treatments available, fosters shared decision-making

and care planning, their family carers reported to feel more between HCPs and residents/family carers, and im-

prepared for death [24]. proves knowledge about residents’ and family carers’ Although exploration of resident’s wishes has been preferences for end-of-life care, may contribute to a identified as key feature for good palliative care [42] timely transition towards palliative-oriented care. because of knowing the relative’s preferences reduces Planning of end-of-life care requires ongoing com- articles

aggressive care at the end-of-life [43], the majority of munication with information about prognosis and our interviewees reported that their relative did not problems that are likely to occur during the disease share her/his wishes for end-of-life care neither with trajectory to achieve shared decisions. Because end- HCPs nor with family. This is consistent with previous of-life is often a protracted and unpredictable process riginal research [41] and explains why several interviewees that occurs during the NH stay, communication about O stated to desire written relative’s preferences for end-of- prognosis and residents’ and family carers’ preferences life care to avoid assumptions and the burden of deci- for care and treatments at the end-of-life should start sion. Anyway, all cognitively competent NH residents as early as possible to promote establishing palliative- should be engaged in discussions about pros and cons oriented care. of end-of-life treatment options as early as possible to promote informed decisions and allow HCPs planning Author’s contribution statement end-of-life care accordingly. However, these dialogues SG, IB and PDG were responsible for the concept are not easy and HCPs are not always prepared to start and design of the research. SG and IB collected the such discussions. When residents’ wishes are not avail- data. SG, IB and MC conducted the data screening able, HCPs should explore family carers’ desired goal and analysis. SG and PDG wrote the manuscript while of end-of-life care for their relative and engage them in IB and MC revised it critically. All authors gave final decisions [6]. approval of the submitted manuscript. Our family carers usually did not desire treatments to sustain their relative’s life, while the quality of life be- Acknowledgements came a central element of the decision-making process The authors would thank the nursing homes for their [32]. They generally preferred to avoid burdensome interest in this research and for their help in recruiting hospitalizations and leave their relative to die quietly in participants. The authors express their deepest grati- NH, being reassured that the NH could provide care tude to the family carers that participated to the study. with equal benefit as compared to hospital [44]. Our study suffered from two main limitations. Firstly, Funding recruitment was affected by the emotionally challeng- The authors received no financial support for the re- ing topic and ethical considerations [45]. However, search, authorship, and/or publication of this article. although the non-random sample that may have intro- duced a selection bias, family carers were drawn from Conflict of interest statement several NHs across a large geographical area and data The authors declared no potential conflicts of interest saturation was reached. Secondly, interviews with other with respect to the research, authorship, and/or publi- family carers and friends could have provided greater cation of this article. understanding of the communication experience of the interviewees; however, three interviews were in double Received on 31 December 2020. and all participants were engaged carers, thus providing Accepted on 14 April 2020.

REFERENCES

1. Bone AE, Gomes B, Etkind SN, Verne J, Murtagh FE, tia: A prospective study. Palliat Med. 2015;29(6):538-46. Evans CJ, et al. What is the impact of population age- doi:10.1177/0269216315570409 ing on the future provision of end-of-life care? Popula- 3. Gonella S, Basso I, De Marinis MG, Campagna S, Di tion-based projections of place of death. Palliat Med. Giulio P. Good end-of-life care in nursing home accord- 2018;32(2):329-36. doi:10.1177/0269216317734435 ing to the family carers’ perspective: A systematic review 2. van Soest-Poortvliet MC, van der Steen JT, de Vet HC, of qualitative findings. Palliat Med. 2019;33(6):589-606. Hertogh CM, Deliens L, Onwuteaka-Philipsen BD. doi:10.1177/0269216319840275 Comfort goal of care and end-of-life outcomes in demen- 4. Rabow MW, Hauser JM, Adams J. Supporting family 323 How communication contributes to palliative care

caregivers at the end of life: “they don’t know what they ing rigour, reliability and validity in qualitative research. don’t know”. JAMA. 2004;291(4):483-91. doi:10.1001/ Int J Ther Rehabil. 2014;21(12):560-8. doi: 10.12968/ jama.291.4.483 ijtr.2014.21.12.560 5. Majerovitz SD, Mollott RJ, Rudder C. We’re on the 20. Vandervoort A, Houttekier D, Vander Stichele R, Van same side: Improving communication between nursing der Steen JT, Van den Block L. Quality of dying in nurs- home and family. Health Commun. 2009;24(1):12-20. ing home residents dying with dementia: does advanced doi:10.1080/10410230802606950 care planning matter? A nationwide postmortem study. 6. van Soest-Poortvliet MC, van der Steen JT, de Vet HC, PloS one. 2014;9(3):e91130. doi: 10.1371/journal. Hertogh CM, Onwuteaka-Philipsen BD, Deliens LH. pone.0091130 reviews

Factors related to establishing a comfort care goal in 21. Mitchell SL, Teno JM, Kiely DK, Shaffer ML, Jones nursing home patients with dementia: a cohort study RN, Prigerson HG, et al. The clinical course of ad-

among family and professional caregivers. J Palliat Med. vanced dementia. N Engl J Med. 2009;361(16):1529-38. and 2014;17(12):1317-27. doi:10.1089/jpm.2014.0205 doi:10.1056/NEJMoa0902234 7. Van der Steen JT, Radbruch L, Hertogh CM, de Boer ME, 22. Van der Steen JT, Onwuteaka-Philipsen BD, Knol DL, Hughes JC, Larkin P, et al. White paper defining optimal Ribbe MW, Deliens L. Caregivers’ understanding of de- palliative care in older people with dementia: a Delphi mentia predicts patients’ comfort at death: a prospec-

study and recommendations from the European Asso- tive observational study. BMC Med. 2013;11(1):105. articles ciation for Palliative Care. Palliat Med. 2014;28(3):197- doi:10.1186/1741-7015-11-105 209. doi:10.1177/0269216313493685 23. Mitchell SL, Palmer JA, Volandes AE, Hanson LC, 8. Travis SS, Bernard M, Dixon S, McAuley WJ, Loving Habtemariam D, Shaffer ML. Level of care prefer- G, McClanahan L. Obstacles to palliation and end- ences among nursing home residents with advanced of-life care in a long-term care facility. Gerontologist. dementia. J Pain Symptom Manage. 2017;54(3):340-5. riginal 2002;42(3):342-9. doi:10.1093/geront/42.3.342 doi:10.1016/j.jpainsymman.2017.04.020 O 9. Mitchell SL, Teno JM, Kiely DK, Shaffer ML, Jones 24. Schulz R, Boerner K, Klinger J, Rosen J. Preparedness for RN, Prigerson HG, et al. The clinical course of ad- death and adjustment to bereavement among caregivers vanced dementia. N Engl J Med. 2009;361(16):1529-38. of recently placed nursing home residents. J Palliat Med. doi:10.1056/NEJMoa0902234 2015;18(2):127-33. doi:10.1089/jpm.2014.0309 10. Toles M, Song M-K, Lin F-C, Hanson LC. Perceptions 25. Cable‐Williams B, Wilson D. Awareness of impending of family decision-makers of nursing home residents with death for residents of long‐term care facilities. Int J Older advanced dementia regarding the quality of commu- People Nurs. 2014;9(2):169-79. doi:10.1111/opn.12045 nication around end-of-life care. J Am Med Dir Assoc. 26. Hebert RS, Schulz R, Copeland VC, Arnold RM. Prepar- 2018;19(10):879-83. doi:10.1016/j.jamda.2018.05.015 ing family caregivers for death and bereavement. Insights 11. Biola H, Sloane PD, Williams CS, Daaleman TP, Wil- from caregivers of terminally ill patients. J Pain Symp- liams SW, Zimmerman S. Physician communication with tom Manage. 2009;37(1):3-12. doi:10.1016/j.jpainsym- family caregivers of long‐term care residents at the end man.2007.12.010 of life. J Am Ger Soc. 2007;55(6):846-56. doi:10.1111/ 27. Barclay S, Froggatt K, Crang C, Mathie E, Handley j.1532-5415.2007.01179.x M, Iliffe S, et al. Living in uncertain times: trajecto- 12. Anderson RA, Issel LM, McDaniel Jr RR. Nursing homes ries to death in residential care homes. Br J Gen Pract. as complex adaptive systems: relationship between man- 2014;64(626):e576-83. doi:10.3399/bjgp14X681397 agement practice and resident outcomes. Nurs Res. 28. Frey R, Foster S, Boyd M, Robinson J, Gott M. Family 2003;52(1):12. doi:10.1097/00006199-200301000- experiences of the transition to palliative care in aged res- 00003 idential care (ARC): a qualitative study. Int J Palliat Nurs. 13. Gonella S, Basso I, Dimonte V, Martin B, Berchialla 2017;23(5):238-47. doi:10.12968/ijpn.2017.23.5.238 P, Campagna S, et al. Association between end-of-life 29. Shield RR, Wetle T, Teno J, Miller SC, Welch LC. Vigi- conversations in nursing homes and end-of-life care lant at the end of life: family advocacy in the nursing outcomes: A systematic review and meta-analysis. J Am home. J Palliat Med. 2010;13(5):573-9. doi:10.1089/ Med Dir Assoc. 2019;20(3):249-61. doi:10.1016/j.jam- jpm.2009.0398 da.2018.10.001 30. Valdimarsdottir U, Helgason AR, Furst CJ, Adolfsson 14. Tong A, Sainsbury P, Craig J. Consolidated criteria for re- J, Steineck G. Awareness of husband’s impending death porting qualitative research (COREQ): a 32-item check- from cancer and long-term anxiety in widowhood: a na- list for interviews and focus groups. Int J Qual Health tionwide follow-up. Palliat Med. 2004;18(5):432-43. Care. 2007;19(6):349-57. doi:10.1093/intqhc/mzm042 doi:10.1191/0269216304pm891oa 15. Kvale S. Interviews: An introduction to qualitative re- 31. Lopez RP, Mazor KM, Mitchell SL, Givens JL. What is search interviewing. Thousand Oaks, CA: Sage Publica- family-centered care for nursing home residents with ad- tions; 1994. vanced dementia? Am J Alzheimers Dis Other Demen. 16. Castillo-Montoya M. Preparing for interview research: 2013;28(8):763-8. doi:10.1177/1533317513504613 the interview protocol refinement framework. Qual Re- 32. Caron CD, Griffith J, Arcand M. End-of-life decision port. 2016;21(5):811-31. making in dementia: The perspective of family caregivers. 17. Graneheim UH, Lundman B. Qualitative content analy- Dementia. 2005;4(1):113-36. sis in nursing research: concepts, procedures and mea- 33. Gessert CE, Forbes S, Bern-Klug M. Planning end-of- sures to achieve trustworthiness. Nurse Educ Today. life care for patients with dementia: roles of families 2004;24(2):105-12. doi:10.1016/j.nedt.2003.10.001 and health professionals. Omega 2001;42(4):273-91. 18. Gonella S, Campagna S, Basso I, De Marinis MG, doi:10.2190/2mt2-5gyu-gxvv-95ne Di Giulio P. Mechanisms by which end-of-life com- 34. Reinhardt JP, Downes D, Cimarolli V, Bomba P. End- munication influences palliative-oriented care in nurs- of-life conversations and hospice placement: Association ing homes: A scoping review. Patient Educ Conuns. with less aggressive care desired in the nursing home. J 2019;102(12):2134-44. doi:10.1016/j.pec.2019.06.018 Soc Work End Life Palliat Care. 2017;13(1):61-81. doi:1 19. Darawsheh W, Stanley M. Reflexivity in research: promot- 0.1080/15524256.2017.1282919 324 Silvia Gonella, Ines Basso, Marco Clari and Paola Di Giulio

35. Toscani F, van der Steen JT, Finetti S, Giunco F, Pette- tient Educ Couns. 2015;98(10):1172-9. doi:10.1016/j. nati F, Villani D, et al. Critical decisions for older people pec.2015.06.022 with advanced dementia: a prospective study in long- 41. Bollig G, Gjengedal E, Rosland JH. They know!-Do they? term institutions and district home care. J Am Med A qualitative study of residents and relatives views on ad- Dir Assoc. 2015;16(6):535.e13-.e20. doi:10.1016/j.jam- vance care planning, end-of-life care, and decision-mak- da.2015.02.012 ing in nursing homes. Palliat Med. 2016;30(5):456-70. 36. Hullett CR, McMillan JJ, Rogan RG. Caregivers’ predis- doi:10.1177/0269216315605753 positions and perceived organizational expectations for 42. Goodman C, Amador S, Elmore N, Machen I, Mathie the provision of social support to nursing home residents. E. Preferences and priorities for ongoing and end-of-life reviews

Health Commun. 2000;12(3):277-99. doi: 10.1207/ care: a qualitative study of older people with dementia res- S15327027HC1203_04 ident in care homes. Int J Nurs Stud. 2013;50(12):1639-

and 37. Boogaard JA, Werner P, Zisberg A, van der Steen JT. 47. doi:10.1016/j.ijnurstu.2013.06.008 Examining trust in health professionals among fam- 43. Bischoff KE, Sudore R, Miao Y, Boscardin WJ, Smith ily caregivers of nursing home residents with advanced AK. Advance care planning and the quality of end-of-life dementia. Geriatr Gerontol Int. 2017;17(12):2466-71. care in older adults. J Am Geriatr Soc. 2013;61(2):209- doi:10.1111/ggi.13107 14. doi:10.1111/jgs.12105

articles 38. Cherlin E, Fried T, Prigerson HG, Schulman-Green 44. Saliba D, Kington R, Buchanan J, Bell R, Wang M, D, Johnson-Hurzeler R, Bradley EH. Communication Lee M, et al. Appropriateness of the decision to trans- between physicians and family caregivers about care at fer nursing facility residents to the hospital. J Am Ger the end of life: when do discussions occur and what is Soc. 2000;48(2):154-63. doi:10.1111/j.1532-5415.2000. said? J Palliat Med. 2005;8(6):1176-85. doi:10.1089/ tb03906.x riginal jpm.2005.8.1176 45. Aoun S, Slatyer S, Deas K, Nekolaichuk C. Family Care- O 39. Gjerberg E, Lillemoen L, Førde R, Pedersen R. End-of- giver Participation in Palliative Care Research: Challeng- life care communications and shared decision-making in ing the Myth. J Pain Symptom Manage. 2017;53(5):851- Norwegian nursing homes-experiences and perspectives 61. doi:10.1016/j.jpainsymman.2016.12.327 of patients and relatives. BMC Geriatr. 2015;15(1):103. 46. Hennings J, Froggatt K, Payne S. Spouse caregivers of doi:10.1186/s12877-015-0096-y people with advanced dementia in nursing homes: A lon- 40. Stiggelbout AM, Pieterse AH, De Haes J. Shared de- gitudinal narrative study. Palliat Med. 2013;27(7):683- cision making: concepts, evidence, and practice. Pa- 91. doi:10.1177/0269216313479685 1 Rome, Italy. E-mail:[email protected]. Address forcorrespondence: LoretaA.Kondili, CentroNazionaleperlaSaluteGlobale, IstitutoSuperiorediSanità,Viale Regina Elena299,00161 to treat. herence, buttheidentification oftheavailablepatients of HCVtherapyisnolonger treatmentefficacyorad- patients withHCVrelateddisease [1].Thelimitation vigorated publichealthinitiatives aimedatidentifying goals fortheeliminationofHCVby2030hasrein- tion (WHO)’s GlobalHealthSectorStrategy(GHSS) viduals. Thus,achievingtheWorld HealthOrganiza- also guaranteeaccesstotreatmentallinfectedindi- cies thatseektoidentifyundiagnosedinfectionsbut achievable target,itiscrucialtodefinehealthpoli- an urgentresponse.AsalifewithoutHCVisnow but itisamajorpublichealthproblemthatrequires INTRODUCTION the WHOeliminationtargetsby2030. care andtreatment,isofparamountimportance,inordertokeepItalyontrackachieve sources bothforcasefinding,throughactivescreeningandactivitiesrapidlinkageto Conclusions. EstablishinganadhocfundforDAAseachItalianRegion,bindingre- years laterthantheyear2030. be anachievablegoalinItaly. Theothereliminationtargetscouldbeachievableover7 number oftreatedpatients,wereassumed,onlythe65%HCVmortalityreductionwould creasing numberofnewlydiagnosedindividualsandasconsequence,adeclineinthe elimination targetswouldbeachieved.Consideringthe2019Scenario,inwhichade- Results. Consideringahightreatmentrate,asassumedbythe2018Scenario,allHCV ing theannuallyHCVtreatmentrateinItaly. Italy. The model assessed two treatment scenarios: 2018 Scenario and 2019 Scenario, us- the impactofdifferentHCVtreatmentratesinachievingeliminationgoals Methods. Amodellinganalysiswasconducted,usingthe“ItalyPolaris”model,toforecast regarding HCVtreatmentimplementationinItalianRegions. dividuals inkeyhigh-riskgroupsandthegeneralpopulation,howeverthereisconcern Organization (WHO)eliminationgoalsrecognizingtheneedtoidentifyundiagnosedin- but achievablegoal.InItaly, thereispoliticalwill,whichaimstoachievetheWorld Health Background andaims.HepatitisCvirus(HCV)eliminationforItalyisanambitious, Abstract 3 2 Loreta A.Kondili to hepatitisCviruselimination? how tokeepItalyontheroad funding fordirectactingantiviraldrugs: Tailored screeninganddedicated DOI: 10.4415/ANN_20_03_10 Ann IstSuperSanità2020|Vol. 56,No.3:325-329 Italy Dipartimento BiomedicodiMedicinaInternaeSpecialistica,UniversitàdegliStudiPalermo, Center forDiseaseAnalysisFoundation,Lafayette,Colorado,USA Centro NazionaleperlaSaluteGlobale,IstitutoSuperiorediSanità,Rome,Italy Initially inItaly, HCVtreatmentwithdirect act- Viral hepatitisisnotjustaclinicallyrelevantdisease, 1 , SarahBlach 2 , HomieRazavi

transformed theItalianNational HealthService(NHS) general principles.TheConstitutional Reformof2001 their territory, whiletheStateislimitedto formulating organizing, andfinancinghealth careserviceswithin Regions withtheirownbroad discretioninplanning, throughout thecountry. Italyisdividedinto twenty ever, accesstotreatmenthasnotbeenhomogeneous DAAs withoutanykindofbudgetrestriction.How- for InnovativeDrugs,whichguaranteedthereleaseof tion, whichwaspossiblethankstoadedicatedFund population hasbeentreatedanderadicatedtheirinfec- Since 2015,alargeproportionoftheHCV-infected all diagnosedpatientswereeligiblefortreatment[2]. Universal treatment strategy was introduced, meaning severe, progressive liver disease. In 2017, however, a ing antivirals(DAAs)wasprioritizedtopatientswith 2 andAntonioCraxi

3 • • • Key words WHO eliminationgoal infection hepatitis C 325 Original articles and reviews 326 Loreta A. Kondili, Sarah Blach, Homie Razavi and Antonio Craxi

from a substantially centralized system into a highly re- of treated patients will also decline. To evaluate the gionalized one. As a consequence of this decentraliza- progress of each scenario, the WHO elimination tar- tion, although a National Hepatitis Plan exists, indi- gets were calculated for Italy (Targets: 90% reduction vidualized models of HCV care persist and there are no in incidence of chronic HCV infections between 2015 uniform strategies across regional networks. Only two and 2030, 65% reduction in HCV-related deaths due to ( and Veneto) of 20 Regions throughout Italy have chronic HCV infection between 2015 and 2030, 90% developed adequate organizational and operational pol- diagnosis coverage of the HCV-infected population and icies regarding HCV elimination [3]. 80% treatment coverage of the eligible HCV-infected reviews

Considering these health systems challenges and the population by 2030) [1]. economic burden of HCV chronic infection, is it pos-

and sible to achieve the HCV elimination targets and what RESULTS

barriers will be faced? The results of each scenario for each elimination tar- get are reported in Figure 1. METHODS All WHO elimination targets would be achieved A modelling analysis was conducted using the “Italy considering the 2018 Scenario. This means that con- articles Polaris” model, grounded in the natural history of HCV sidering a high treatment rate, as it was reported dur- progression to forecast the impact of HCV in the dis- ing the year 2018, Italy could eliminate HCV infection ease burden in the general population as previously de- by 2030. Considering the 2019 Scenario, in which it scribed [2-6]. Briefly, an Excel-based Markov disease was assumed the number of newly diagnosed individu- riginal burden model was populated with HCV prevalence als each year would decrease and as consequence the O data in Italian general population, stratified by age and number of those treated will also decline, only the 65% sex, to quantify the annual HCV-infected population HCV mortality reduction would be an achievable goal. by liver disease stage, sex, and age. The model simu- The other elimination targets could be achievable over lates the natural history of the disease and forecasts 7 years later than the year 2030. If screening/diagnosis disease burden annually, assessed under two scenarios, increases from the screening guidelines but treatment 2018 Scenario and 2019 Scenario, using the annually, does not, then the outcomes will not change substan- reported DAA treatment rate in Italy [7, 8]. The inputs tially from the 2019 Scenario outcomes. of each scenario are reported in Table 1 (A, B). 2018 For each elimination target, the overall results are Scenario assumes that the same number of people schematically reported in Figure 2. are treated each year beginning in 2018, and that at least 30 000 patients are diagnosed per year [7]. 2019 DISCUSSION Scenario assumes that without an extensive screening Italy, a country that in 2018 was on track for achiev- policy the number of newly diagnosed individuals each ing the elimination goals [7], is no longer estimated to year will decrease. Additionally, as the number of newly achieve these targets, except the goal for liver mortal- diagnosed decreases, it is estimated that the number ity reduction [8]. Previously, Italy was estimated to have

Table 1 Key inputs by treatment scenario (A) 2018 Treatment scenario Years 2018 2019 2020 2021 2022 2023 Treated 56 300 56 300 56 300 56 300 56 300 56 300 Newly diagnosed 30 400 30 400 30 400 30 400 30 400 30 400 Fibrosis stage ≥F0 ≥F0 ≥F0 ≥F0 ≥F0 ≥F0 New infections 7200 7200 6000 5400 4800 4300 Treated age 15-85+ 15-85+ 15-85+ 15-85+ 15-85+ 15-85+ SVR* 98% 98% 98% 98% 98% 98%

(B) 2019 Treatment scenario Years 2018 2019 2020 2021 2022 2023 Treated 56 300 35 800 32 100 31 400 31 200 31 200 Newly diagnosed 30 400 27 800 25 800 23 800 22 000 15 400 Fibrosis stage ≥F0 ≥F0 ≥F0 ≥F0 ≥F0 ≥F0 New infections 7200 6500 6100 5700 5300 4900 Treated age 15-85+ 15-85+ 15-85+ 15-85+ 15-85+ 15-85+ SVR* 98% 98% 98% 98% 98% 98% *SVR: sustained virological response. 327 How to keep Italy on the road to hepatitis C virus elimination? reviews

and

articles

riginal O

Figure 1 The number of years needed to achieve the WHO elimination goals in Italy, by scenario. A) Target: 90% diagnosed by 2030 2018 Scenario: by 2030, 90% of viremic patients will be diagnosed. 2019 Scenario: by 2030, less than 90% of viremic patients will be diagnosed. It will take more than seven years, given the scenario-specific diagnosis and treatment rates, to achieve the treatment target. B) Target: 80% treated in 2030 2018 Scenario: the model estimates 80% of patients will be treated by 2030. 2019 Scenario: the model estimates 65% of patients will be treated by 2030. It will take an additional seven years, given the scenario-specific diagnosis and treatment rates, to achieve the treatment target. C) Target: 65% reduction in HCV correlated deaths This target is achievable in Italy with both scenarios. 2018 Scenario: the target is achievable by 2022. 2019 Scenario: the target is achievable 5 years later, by 2027. D) Target: 90% reduction of new HCV infections 2018 Scenario: the 90% reduction of new infections will be achieved by 2030. 2019 Scenario: the target is achievable 7 years later, in 2037. the highest HCV prevalence in Europe, which was most in 2015 delaying further progression to death, the 65% likely attributed to a wave of nosocomial transmission reduction of liver related deaths target was forecasted to occurring in the 1950s, followed by intravenous drug use be an achievable reality for Italy by 2022 [3]. Though a in the 1980s [9-11]. These earlier epidemics would have constantly high treatment rate, as observed in the 2018 resulted in a large number of HCV-infected individuals Scenario, would permit to achieve the WHO targets, progressing to more severe liver disease, such as cirrho- reducing the number of patients treated each year, be- sis. However, due to the extensive roll-out of treatment ginning in 2019 (the 2019 Scenario), would mean the

Figure 2 Year of achieving elimination targets (extrapolated from 2019 data). Current WHO Target is 2030. The Center for Disease Analysis Foundation. Picture available from: https://cdafound.org/dashboard/polaris/dashboard.html. 328 Loreta A. Kondili, Sarah Blach, Homie Razavi and Antonio Craxi

opposite [7, 8]. Italy would not be able to maintain the has important medium to long-term health and eco- treatment rate necessary to achieve the set targets. Ac- nomic benefits for the NHS [16], were well received by cording to the 2019 Scenario, which does not consider Italian policymakers, resulting in important political ac- an extensive screening strategy of the population, the tion to achieve the elimination goals. An amendment to pool of treated patients is predicted to run out by the the Milleproroghe Decree concerning HCV screening has year 2025 [3], leaving a large burden of HCV-infected recently been approved [17]. With this provision, 71.5 individuals undiagnosed, nor cured. However, if screen- million Euros are allocated in 2020-2021, to introduce ing and diagnosis increases due to new screening guide- a screening program at no cost to the patient, with the reviews

lines but treatment does not, then the outcomes will not final goal of identifying undiagnosed individuals with change substantially from the 2019 Scenario outcomes. chronic HCV infection. These screening strategies will

and As reported by mathematical models, Italy will be able to focus on the free of charge screening of key populations

achieve the elimination goals if at least 40 000 patients (current intravenous drug users) in the public services are treated annually, which is achievable only through an for drug addiction (SerT), people detained in prison, active screening campaign and consistent access to treat- and general population screening, first, of those born ment for newly diagnosed patients [7]. To address this is- between the years 1969-1989. articles

sue, a recent modelling analysis determined if universal In light of these efforts, it seems as though program- and birth cohort screening (those born between 1948- matic structures are in place and would support Italy 1987, with the highest HCV prevalence and proportion to once again be on track for the elimination of HCV of undiagnosed patients), were cost-effective measures by 2030. However, identifying undiagnosed HCV-in- riginal to improving diagnosis [12]. The analysis found that fected individuals and patients not yet on treatment re- O both scenarios were cost-effective compared to the sta- quires close partnership between the scientific, health, tus quo (treatment of linked-to-care patients only). The and political worlds at central and regional levels. This incremental cost-effectiveness ratio (ICER) varied from would involve uniting the local medical structures, gen- 3,552-6,758/quality adjusted life year (QALY), far lower eral medicine doctors, and prescribing centers around than the Willingness to Pay threshold of 25 000 Euros/ one goal: the elimination of HCV infection by the year QALY, under which the strategy is considered cost-ef- 2030. However, the roll-out of an active screening cam- fective in Italy. paign could be considered as a starting point, as effi- The study determined that a tailored, graduated cient screening should be supplemented with rapid link- birth cohort screening strategy, which first identified age-to-care and treatment of newly diagnosed patients. young populations (1968-1987 birth cohorts) at risk of To ensure rapid linkage-to-care means appropriate di- transmitting HCV and then expanded to identify older agnostic and therapeutic paths that focus diagnoses on populations (1948-1967 birth cohorts) before disease patients not yet on treatment must be constructed. An progression, had the highest cost-effectiveness profile in “ad hoc fund for HCV elimination” should be part of Italy [12]. With specific regard to the 1968-1987 birth each regional elimination strategy. However, providing cohort, the phenomenon of transmission through inject- HCV treatment to diagnosed individuals will be a chal- ing drug use needs to be mentioned. In Italy, injecting lenge. The “Innovative Drugs Fund”, through which the drug use has been a less important mode of transmission State guarantees the necessary funds for DAAs, expired when compared to other European countries, although in April 2020. This means that Italy’s 20 individual Re- it was mainly responsible for the 1980-1990 wave of gions will be faced with the challenge of finding a way infections. Most individuals who have acquired HCV to pay for the drugs, yet to date, no regional plans for through injection drug use were born between 1968 eliminating HCV infection exist. Though the new HCV and 1987. Because these individuals are younger than screening policies address one bottleneck to the elimi- those who acquired infection through iatrogenic means, nation of HCV, the lack of a dedicated fund for DAAs a higher proportion of them are in the F0-F3 fibrosis will stress an already overburdened regional budget. stage and thus, more likely to have gone undiagnosed This could result in discordant actions with little chance [13]. Screening these younger cohorts first would likely of success. In most Regions, in the absence of regional detect individuals at higher risk of infectiousness, de- elimination plans, the treatment curve will drastically creasing the potential to transmit new infections com- decrease and Italy will move further and further away pared to screening older patients who are more likely achieving HCV elimination targets. In other words, if already diagnosed and cured for their symptomatic dis- Italy is to achieve the WHO’s HCV elimination target, ease. Considering also that more than 20% of treated screening alone will obviously not be sufficient. A com- patients in 2019 had cirrhosis or very advanced liver dis- mitment is needed by the state to support the WHO ease and similar estimates of advanced disease could be elimination goals for chronic HCV infection, which re- found in undiagnosed individuals, DAAs should still be mains a silent public health threat in Italy. It is of fun- considered as life-saving drugs [2, 12, 13]. Additionally, damental importance that a national plan for providing numerous studies have shown that HCV-related disease treatment for HCV infection be an integral part of re- inflicts a huge economic and clinical burden as a result gional elimination plans. of HCV related extra-hepatic comorbidities. Early erad- ication of HCV before it progresses to advanced liver CONCLUSIONS disease could reduce these burdens [14, 15]. Italy can once again be “on track” for achieving the The tailored screening strategy, coupled with the re- WHO targets for HCV elimination which can be main- sults of a previous study that reported HCV treatment tained if adequate screening strategies are adopted to 329 How to keep Italy on the road to hepatitis C virus elimination?

identify undiagnosed individuals (as a starting point), in the modelling work for this project; the entire Cen- and immediately link them to a cure for viral eradica- ter for Disease Analysis Foundation team for producing tion (an end point and of equal priority). previous estimations, which created the basis for this The establishment of an ad hoc fund for hepatitis study design, on a voluntary basis; Sarah R. Scott for treatment within the National Plan for the Prevention her contributions to the editing of this manuscript. and Treatment of hepatitis C is necessary and should be supplemented by regional elimination plans that could Conflict of interest statement make the WHO’s HCV elimination targets attainable The Authors declare that there is no conflict of inter- reviews in Italy by 2030. est regarding the publication of this paper.

Acknowledgements Received on 25 April 2020. and

The Authors thank Ivane Grankrelidze who assisted Accepted on 8 June 2020.

REFERENCES articles

1. World Health Organization Global health sector strategy titis C virus infection in the general population: a on viral hepatitis 2016-2021. Towards ending viral hepati- community-based survey in southern Italy. Hepatology. tis. Geneva: WHO; June 2016. 1997;26:1006-101. doi:10.1002/hep.510260431 riginal 2. Italia. Ridefinizione dei criteri di trattamento per la te- 11. Andreoni M, Giacometti A, Maida I, Meraviglia P, Ri- O rapia dell’Epatite C cronica. (Determina n. 500/2017). pamonti D, Sarmati L. HIV-HCV co-infection: epide- Gazzetta Ufficiale della Repubblica Italiana - Serie miology, pathogenesis and therapeutic implications. Eur Generale n. 75, 24 marzo 2017. Rev Med Pharmacol Sci. 2012;16:1473-8. 3. Kondili LA, Robbins S, Blach S, et al. Forecasting hepati- 12. Kondili LA, Gamkrelidze I Blach S, et al. Optimization of tis C liver disease burden on real-life data. Does the hid- hepatitis C virus screening strategies by birth cohort in It- den iceberg matter to reach the elimination goals? Liver aly. Liver Intern. 2020 (in press) doi.org/10.1111/liv.14408 Int. 2018;38:2190-8. doi: 10.1111/liv.13901 13. Kondili LA, Andreoni M, Alberti A, et al. Estimated prev- 4. Razavi H, Waked I, Sarrazin C, et al. The present and alence of undiagnosed hepatitis C virus infected individu- future disease burden of hepatitis C virus (HCV) in- als in Italy: a mathematic model to accurately measure fection with today’s treatment paradigm. J Viral Hepat. HCV prevalence with a route of transmission granularity. 2014;21(Suppl 1):34-59. doi: 10.1111/jvh.12248 Hepatology. 2019;70(S1) AASLD abstract (poster) N. 5. Blach S, Zeuzem S, Manns M, et al. Global prevalence 0520 Friday 8 November 2020: p. 328A. and genotype distribution of hepatitis C virus infection in 14. Negro F, Forton D, Craxì A, et al. Extrahepatic morbidi- 2015:a modelling study. Lancet Gastroenterol Hepatol. ty and mortality of chronic hepatitis C. Gastroenterology. 2016;2:161-76. doi: 10.1016/S2468-1253(16)30181-9 2015;149:1345-60. doi: 10.1053/j.gastro.2015.08.035 6. Andriulli A, Stroffolini T, Mariano A, et al. Declining 15. Adinolfi LE, Petta S, Fracanzani AL, et al. Impact of hep- prevalence and increasing awareness of hepatitis C virus atitis C virus clearance by direct-acting antiviral treatment infection in Italy: a population-based survey. Eur J Intern on the incidence of major cardiovascular events. A pro- Med. 2018;53: 79-84. doi:10.1016/j.ejim.2018.02.015 spective multicentre study. Atherosclerosis. 2020;296:40- 7. Razavi H, Sanchez- Gonzales Y, Yuen C, Cornberg M. 7. doi: 10.1016/j.atherosclerosis.2020.01.010 Global timing of hepatitis C virus elimination in high-in- 16. Marcellusi A, Viti R, Kondili LA, Rosato S, Vella S, come countries. Liver Int. 2020;40:522-9. doi: 10.1111/ Mennini FS. Economic consequences of investing in liv.14324 anti-HCV antiviral treatment from the Italian NHS per- 8. CDA Foundation’s Polaris Observatory. Italy Progress spective: a real-world-based analysis of PITER data. Phar- Toward Elimination 2019. 2020 (updated). Available macoeconomics. 2019;37:255-66. doi:10.1007/s40273- from: https://cdafound.org/polaris/. 018-0733-3 9. Deuffic-Burban S, Deltenre P, Buti M, et al. Predicted 17. Italia. Decreto legislativo 30 dicembre 2019, n. 162 con- effects of treatment for HCV infection vary among Euro- versione in legge 28 febbraio 2020, n. 8. Screening na- pean countries. Gastroenterology. 2012;143:974-85. doi: zionale gratuito per eliminazione del virus HCV a norma 10.1053/j.gastro.2012.05.054 dell’articolo 25 sexies. Gazzetta Ufficiale della Repub- 10. Guadagnino V, Stroffolini T, Rapicetta M, et al. Preva- blica Italiana - Serie Generale n. 51, 29 febbraio 2020. lence, risk factors, and genotype distribution of hepa- (Suppl ord n. 10/L). 330 Original articles and reviews be easilyexposedwhencaring formeaslesinfectedpa- for almost1hour[4],healthcare workers(HCWs)can est incidenceobservedinthe 0-4 agegroup[3]. in people aging between 15 and 39 years, with the high- one complication.Morethan60%ofthecasesoccurred cent werenotvaccinated,and31%experiencedatleast (median age30years)werereported:eighty-sixper cases [2]. Region eachshoweda15%reductioninthereported contrast, theSoutheastAsiaRegionandAmericas and theWestern PacificRegiona230%increase;by the EasternMediterraneanregiona50%increase, 900% increase,theEuropeanregiona120% the sameperiodof2018,Africanregionrecordeda cases to the WHO [2]. Comparing the data referring to 31 2019,182countriesreported364808newmeasles deaths eachyear[1].InthetimeperiodJanuary1-July ity andmortality, accountingformorethan100000 still animportantvaccine-preventablecauseofmorbid- vaccine coverageinthe1980s[1].Nevertheless,itis curred yearlybeforetheincreaseinglobalmeasles acute viral disease.Morethan 2 milliondeathsoc- INTRODUCTION centage ofnurses,evenamongthoseworkinginhighriskareas. Conclusions. Ourstudyrevealedanon-protectiveantimeaslesIgGlevelinhighper showed highervaluesthanfemales(253.3vs214.6;p<0.01). 68.7%; p<0.01).Themeanvalueoftheanti-measlesIgGwas217.2±91.1AU/ml.Males IgG antibodieslevel.Protectionratewashigheramongworkersaging40years(82.6%vs Results. Thisstudyincluded358nurses.77.7%(217)hadaprotectivemeasles-specific work areaandlevelsmeasles-specificIgGwereevaluated. gata hospitalwhounderwenttheoccupationalhealthscreeningprogram.Gender, age, Methods. We retrospectivelyevaluatedtheclinicalrecordsofnursesworkinginTor Ver immunization statusformeaslesinnursesofalargehospitalRome. Previous studiesfoundhighrateofoperatorsunprotectedformeasles.We evaluatedthe time periodandahighriskoftransmissionforairbornepathogens,includingmeasles. Background andaims.Nurse’s jobinvolvesstayingclosetothepatientforanextended Abstract ²Dipartimento MedicinadeiSistemi,UniversitàdegliStudidiRoma“Tor Vergata”, Rome,Italy ¹Dipartimento diBiomedicinaePrevenzione,UniversitàdegliStudiRoma“Tor Vergata”, Rome,Italy. Luca Coppeta¹,StefanoRizza²,OttaviaBalbi¹,SavinoBaldi¹andAntonioPietroiusti¹ for hospitaloutbreak among Italiannurses.Apotential Lack ofprotectionformeasles DOI: 10.4415/ANN_20_03_11 Ann IstSuperSanità2020|Vol. 56,No.3:330-335 00133 Rome,Italy. E-mail: [email protected]. Address forcorrespondence: Luca Coppeta, Dipartimento di BiomedicinaePrevenzione, UniversitàdegliStudi di Roma“Tor Vergata”, Via Cracovia 50, Since theviruspersistsviable inaerosolsuspension In Italy, duringtheyear2019,1627measlescases Measles isacontagiousandpotentiallydangerous cases ofoccupationalcontagion [10]. Thus, nurses have been often involved in thereported transmission forairbornepathogens, includingmeasles. including theperformanceof maneuversathighriskof for a protracted time for the duration of the work shift, another pediatricinpatient[9]. patient toanHCW, withsubsequent transmissionto avoidable caseofnosocomialtransmissionfromonein- vere complicationsforthem[1].Baxietal.reportedone measles transmittedbyHCWsmayoftenresultinse- munological deficitthengeneralpopulation,therefore are morelikelytobeaffectedbychronicdiseasesorim- patients hasbeenwidelyreported[7-8].Hospitalized tients, in fact transmission from affected operators to HCWs maybecomeasourceofexposurefortheirpa- (115) involvedHCWs[6].Ofnote,incaseofinfection, ing the2018epidemic,alargenumberofmeaslescases erature duringthelastdecade[5],andalsoinItalydur healthcare facilitieshavebeenwidelyreportedinthelit- [4]. Notsurprisingly, cases of measles contagion inthe onset, when the diseaseisdifficult tobe recognized tected HCWs: viral shedding is higher before the rush population. Measlescanspreadfrompatientstounpro- estimated tobe13foldgreatercomparedthegeneral tients. Infact,theriskofacquiringmeaslesforHCWsis Nurse jobinvolvesaclosecontactwiththepatient

- - • • • Key words vaccination strategy measles transmission ofmeasles occupational

- 331 Lack of protection for measles among Italian nurses

Vaccination remains the only effective measure to RESULTS prevent the contagion, however, since the particular All the participants were eligible and therefore were closeness that nurses have with patients, even if measles included in the study. Main characteristics of study is mainly spread by the airborne route, the lack of strict population are shown in Table 1. We evaluated 358 adherence to alcohol-base hand rub before and after all nurses (male n = 72, and female n = 286). The mean patient contacts [11], and the delay of appropriate iso- age of study population was 42.5 years (range = 28-58 lation measures may contribute to outbreaks in hospital years); 128 HCWs (male n = 36, and female n = 92) settings [12, 13]. were younger or equal than 40 years old, whereas 230 reviews

In Italy, during a large population outbreak in 2017, (male = 36 and female = 194) were older than 40 years. vaccination for measles, mumps, rubella and varicella Of note, 68.7% (95% CI = 63.7-73.2) had a work se-

(MMRV) became compulsory for children 0-2 years old, niority of more than 15 years. As shown in Table 1, most and but it was barely recommended for HCWs, according of study participants worked in Surgery and Emergency to the national vaccine prevention plan approved by the areas. Among the 358 nurses, 268 (74.9%, 95% CI = Italian government in the same year. Unfortunately, in It- 70.4-79.1) showed a protective measles IgG titre. Table aly a growing proportion of health professionals including 2 shows the clinical and occupational characteristics of articles nurses could be categorized as vaccine-hesitant; HCW’s the study population (n = 358) in relation to their se- knowledge and attitudes about vaccination is a key deter- rological status. A higher, although not significant (p minant of their own vaccine acceptance [14, 15]. = 0.136), prevalence of serologically immune subjects In this study we aimed to evaluate the immunological was found in male gender (83.3%, 95% CI = -76.2-84.6) status for measles among nurses working in a university with respect to female nurses (78.4, 95% CI = 72.4- riginal hospital in Rome one year after the enactment of the 84.8). We observed a significantly higher percentage O national vaccine prevention plan. of nurses aged older than 40 years having a protective measles IgG level (82.6%, 95% CI = 77.8-87.2) in com- METHODS parison to younger colleagues (68.7%, 95% CI = 64.1- This was a retrospective prevalence study which has 72.3) (p <0.001). Moreover, a significant higher preva- been approved by the Independent Ethics Commit- lence of immune subjects was detected in nurses having tee of the University Hospital Policlinico Tor Vergata a work seniority longer than 15 years in comparison to (PTV), Rome, Italy. Exclusion criteria included diagno- the other nurses (82.9%, 95% CI = 77.7-85.3 vs 66.1%, sis of diabetes, liver disease, renal insufficiency, thyroid 95% CI = 62.4-70.2 respectively; p <0.01 ). disorders, heart failure, coagulopathy, history of any Regarding working area, we found the higher rate form of cancer, positive blood tests for HIV, hepatitis B, or hepatitis C. We analyzed a group of nurses (n 358) who under- Table 1 went periodic health surveillance screening in the Oc- Clinical and working characteristic of the study population cupational Medicine service in the year 2018. Each par- ticipant performed a single annual health surveillance Variables Study population % (CI 95%) (n = 358) check throughout the study period. For each patient the following data were recorded: age, gender, job se- Mean age (± SD), years 42.5 (6.4) niority working area and measles specific IgG antibod- Gender ies title. According to the literature data, antibodies Male 72 20.1 (16.2-24.3) serum value higher than 16.5 AU/ml was considered protective [9]. A chemo luminescence immunoassay Female 286 79.9 (75.7-83.8) (the LIAISON® Measles IgG assay) was used to per- Age class form a semi-quantitative determination of specific IgG ≤40 years 128 35.8 (30.7-41.1) antibodies for measles in plasma. >40 years 230 64.3 (58.9-69.3) Subjects with partial clinical and serological data, or positive measles-specific IgM antibodies were excluded Working length from the study. ≤15 years 112 31.3 (26.8-36.3) Nurses were divided into two subgroups according to >15 years 246 68.7 (63.7-73.2) their age: younger or equal to 40 years (≤40 y) and older Working area (>40 y). We extracted information on job task, senior- ity (years) and working area from occupational records. Medicine 64 17.9 (14.0-22.1) We studied the prevalence of serologically protected Infective 18 5.0 (2.8-7.3) nurses and compared the mean values of IgG specific Surgery 130 36.3 (31.3-41.3) antibodies among different gender and the age group. Statistical analysis was performed by means of SPSS Radiology 20 5..6 (3.4-7.8) analytic software (release 25). Chi Squared test for di- Emergency 86 24.0 (19.6-28.2) chotomous variables and t-test for continuous values Ambulatory 40 11.2 (7.8-14.5) were used to evaluate statistical significance. Logistic Measles serological immunity regression model was used to perform multivariate analysis. Only P values <0.001 were considered as sig- No 90 25.1 (20.9-29.6) nificant in our study. Yes 268 74.9 (70.4-79.1) 332 Luca Coppeta, Stefano Rizza, Ottavia Balbi et al.

Table 2 Clinical and working characteristic of the study population (n = 358) divided upon protective serogical immunity for measles (posi- tive for IgG >16.5 AU/ml) Variables Number of % (95% CI) of individuals immune p individuals immune for measles for measles Gender Male (n = 72) 58 80.6 (84.6-76.2) 0.136 reviews Female (n = 286) 210 78.4 (72.4-84.8)

and Age class

≤40 years (n = 128) 88 68.7 (64.1-72.3) <0.01 >40 years (n = 230) 190 82.6 (77.8-87.2) Working length articles

≤15 years (n = 112) 74 66.1 (62.4-70.2) <0.01 >15 years (n = 246) 204 82.9 (77.7-85.3) Working area

riginal Medicine (n = 64) 58 90.6 (85.2-96.0) <0.01 O Infective (n = 18) 11 61.1 (52.6-68.8) Surgery (n = 130) 94 72.3 (67.4-77.6) Radiology (n = 20) 20 100 (100.0-100.0) Emergency (n = 86) 66 76.7 (72.8-80-9) Ambulatory (n = 40) 30 75.0 (70.2-80.2)

of protected subjects in radiology (20/20; 100%) and pendently associated with age (odds ratio = 1.045, 95% medicine (58/64 = 90.6%, CI 95% = 85.2-96.0) areas CI = 1.005-1.087, p = 0.027), even after adjustment for whereas infectious disease, surgery and ambulatory work seniority, gender and working areas (Table 3). areas showed the lower percentage of serologically im- The suboptimal protection in HCW category is also mune operators (11/18 = 61.1%, 95% CI = 52.6-68.8); reflected in the number of nosocomial cases of measles. 94/130 = 72.3%, CI 95% = 67.4-77.6; and 30/40 = 75%, In 2017-2018 time period, the Service of Occupational CI 95% = 70.2-80.2, respectively). Surprisingly, in the Medicine of University Policlinic of Rome Tor Vergata emergency department, 20/86 nurses (23%) tested un- reported 17 cases of nosocomial measles, seven out of protected at serological screening. which in nurses. To avoid any possible confounding factors, we built up a logistic regression analysis; we considered age, gender DISCUSSION and seniority as independent variables. Moreover, since We found a high level of nurses serologically unpro- working area was significantly associated with protec- tected for measles, especially in workers aging equal or tive antibody titres at univariate analysis, this variable less than 40 years. In fact, among younger subjects who was also included in the logistic regression model. Since represent a large part of our sample due to the relatively infectious disease had the lower proportion of protected low mean age of our hospital workforce, 31.3% resulted operators we kept it separated. We therefore regrouped serologically non immunized. the other working areas in two main categories accord- Age showed a significant association with serological ing to care environment: 1) Emergency and Medicine status, standing for a proportional relationship of pro- Department vs 2) Surgical and Diagnostic Department tective antibodies titre with age. This finding is in keep- (including radiology and ambulatory). ing with previous reports [16]. Protective serology may The regression model confirmed that protective an- be both the effect of immunization or natural infection tibodies levels for measles were significantly and inde- among the individuals included in the study population.

Table 3 Logistic regression analysis for positive measels (IgG >16.5 AU/ml) Variables OR 95% CI for OR p Age (y) 1.021 1.001-1.092 0.047 Gender (male) 1.668 0.834-3.396 0.102 Working Length (>15 y) 1.042 0.443-2.496 0.956 Emergency Medicine Department 1.524 0.890-2.608 0.079 333 Lack of protection for measles among Italian nurses

It should be taken into account that a large majority of According to the national guidelines, the administra- our study population was born before 1985, a time at tion of two does MMR is strongly recommended for which the vaccination rate for MMR in Italy was lower HCWs having both no written documentation of vac- than 5%, it is therefore plausible that natural immuniza- cination and non-protective IgG titer. Based on the re- tion was the phenomenon underlying our data (Figure sults of previous studies [24-37], workplace vaccination 1) [13-15]. strategy should be offered since it showed to be highly Previous studies reported a paradoxical higher risk of cost-effective and to result in adequate level of protec- infection among young adults due both to lower rate tion among HCWs. reviews of natural immunization and inadequate MMR cover- Based on the results of our study actual national poli- age [17-21]. According to the Italian infectious disease cy regarding vaccine offer seems to miss the objective to

network [11], during the year 2018 the age group 15-39 reach adequate levels of protection among nurses work- and reported the majority of cases of measles and 4.6% of ing in high risk setting. cases involved HCWs. Median age of affected workers Our survey has some possible limitations: the num- was 35 years and 47% of them reported at least one ber of male nurses included in the study is relatively complication. low and so the lack of statistical significance in gen- articles

Our study showed a clear gender difference in sero- der difference may be due to sample size. Moreover, logical status: male HCWs had both higher although we did not evaluate the records of the previous vac- non-significant rate of serum immunity and a signifi- cination so a percentage subjects showing a non- cantly higher mean IgG level. Gender difference in im- protective titre could effectively be immune accord- mune status, in our opinion, can be explained by the ing to current recommendations. However, in Italy a riginal paradoxical effect of higher vaccination rate among national registry of vaccination is not present and a O women due to MMR vaccinations programs for rubella negligible part of study population had a written vac- prevention. After the administration of two doses of cine documentation. MMR vaccine, measles antibodies decline over years and are detectable in serum of vaccinated subjects for CONCLUSIONS almost 15 years, while natural infection induces both Our study shows low rate of serological immunity higher and more persistent antibodies response. among nurses working in a teaching hospital in Italy, Regarding employment area, we found a surprisingly even in exposure prone settings. Younger employees high rate of unprotected nurses working in high-risk hos- showed lower coverage rate and therefore a relatively pital setting such as infectious disease unit and emer- high risk of contagion. Current government policy re- gency department: these subjects should be worried garding the vaccine offer seems to be inadequate to about their risk of measles infection since they faced the reach acceptable level of immunity among HCWs. outbreak occurred in the previous year, when over 80 Since nosocomial transmission of measles represents a cases of measles were admitted in the facility [22]. Lack serious risk, occupational health service should increase of awareness and vaccine hesitancy might have resulted prevention activities, including workplace vaccination in low vaccine acceptance among those operators [23]. of non-immune subjects.

Vaccination coverage rates among Italian infants aged 24 months 100 90 80 70 60 50 40 30

% Vaccination coverage % Vaccination 20 10 0 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Years

Figure 1 Historical coverage rate for measles, mumps and rubella (MMR) vaccination in Italy (period 1985-2003). 334 Luca Coppeta, Stefano Rizza, Ottavia Balbi et al.

Acknowledgements Compliance with ethical standards The Authors would like to thank all health care work- All procedures performed in studies were in accor- ers of Occupational Medicine of Policlinico Tor Vergata dance with the ethical standards of the Institutional for supporting this study. Research Committee (approved with authorization number 170) and with the 1964 Helsinki Declaration Funding and its later amendments or comparable ethical stan- The study was not funded. dards. reviews

Conflicts of interest statement Received on 3 October 2019. None declared. Accepted on 19 May 2020. and

REFERENCES

articles 1. Moss WJ. Measles. Lancet. 2017;390(10111):2490-502. 13. Bernard H, Fischer R, Mikolajczyk RT, Kretzschmar M,

doi: 10.1016/S0140-6736(17)31463-0 Wildner M. Nurses’ contacts and potential for infectious 2. World Health Organization. New measles surveillance disease transmission. Emerg Infect Dis. 2009;15(9):1438- data for 2019. Available from: www.who.int/immuniza- 44. doi: 10.3201/eid1509.081475 tion/newsroom/new-measles-data-august-2019/en/. 14. Willy ME, Koziol DE, Fleisher T, Koo S, Mcfarland riginal 3. Filia A, Bella A, Del Manso M, Baggieri M, Marchi A, H, Schmitt J, et al. Measles immunity in a popula- O Bucci P, Magurano F, Nicoletti L, Rota MC. Morbillo & tion of health-care workers. Infect Cont Hosp Ep. Rosolia News, N. 58 gennaio 2020 Available from: www. 1994;15(1):12-7. epicentro.iss.it/problemi/morbillo/bollettino.asp. 15. National Integrated Measles-Rubella Surveillance Sys- 4. Advisory Committee on Immunization Practices (ACIP) tem. Measles in Italy: period 1 January – 31 December – Centers for Disease Control and Prevention (CDC). 2018 (W30). Rome: Istituto Superiore di Sanità; 1 Janu- Immunization of Health-Care Personnel. Recommenda- ary 2019. 48. Available from: www.epicentro.iss.it/mor- tions of the Advisory Committee on Immunization Prac- billo/bollettino/RM_News_2018_48%20def.pdf. tices. MMVR Recomm Rep. 2011;60(RR-7):1-15. 16. Scepanovic P, Alanio C, Hammer C, et al. Human ge- 5. Centers for Disease Control and Prevention (CDC). Pre- netic variants and age are the strongest predictors of vention of measles, rubella, congenital rubella syndrome, humoral immune responses to common pathogens and and mumps, 2013. Summary Recommendations of the vaccines. Genome Med. 2018;10(1):59. doi: 10.1186/ Advisory Committee on Immunization Practices (ACIP). s13073-018-0568-8 MMWR. 2013;62(RR04);1-34. 17. Rooney JA, Milton DJ, Hackler RL, Harris JH, Reynolds 6. Filia A, Bella A, Manso M, Rota C. Morbillo. Aspetti D, Tanner M, et al. The largest outbreak of measles in the epidemiologici. Roma: EpiCentro, Istituto Superiore di United States during 1999: imported measles and pockets Sanità. Available from: www.epicentro.iss.it/problemi/ of susceptibility. J Infect Dis. 2004;189(Suppl. 1):78-80. morbillo/epidItalia.asp. 18. Filia A, Bella A, Del Manso M, Baggieri M, Magu- 7. Thierry S, Alsibai S, Parent du Chatelet I. An outbreak of rano F, Rota MC. Ongoing outbreak with well over measles in Reims, eastern France, January-March 2008 - 4000 measles cases in Italy from January to end Au- a preliminary report. Euro Surveill. 2008;13(13). gust 2017 − what is making elimination so difficult? 8. De Swart RL, Wertheim-van Dillen PM, van Binnendijk Euro Surveill. 2017;22(37). doi: 10.2807/1560-7917. RS, Muller CP, Frenkel J, Osterhaus AD. Measles in a ES.2017.22.37.30614 Dutch hospital introduced by an immuno-compromised 19. Davidkin I, Jokinen S, Broman M, Leinikki P, Peltola H. infant from Indonesia infected with a new virus genotype. Persistence of measles, mumps, and rubella antibodies in Lancet. 2000;355(9199):201-2. an MMR-vaccinated cohort: a 20-year follow-up. J Infect 9. Ruchi Baxi, Oliver T. Mytton, Muhammad Abid, Anne Dis. 2008;197:950-6. Maduma-Butshe, Shabnam Iyer, Anyanate Ephraim, 20. LeBaron CW, Beeler J, Sullivan BJ, et al. Persistence of Kevin E. Brown, Éamonn O’Moore, Outbreak report: measles antibodies after 2 doses of measles vaccine in a nosocomial transmission of measles through an unvacci- postelimination environment. Arch Pediatr Adolesc Med. nated healthcare worker – implications for public health, J 2007;161:294-301. Publ Health. 2014;36(3):375-81. Available from: .https:// 21. Italia. Ministero della Salute. Le coperture vaccinali doi.org/10.1093/pubmed/fdt096. dell’eta pediatrica. [Vaccination coverage in childhood]. 10. Escombe AR, Oeser CC, Gilman RH, Navincopa M, Ti- Rome: Ministero della salute. Available from: www.salute. cona E, Pan W, Martinez C, Chacaltana J, Rodriguez R, gov.it/imgs/C_17_tavole_20_allegati_iitemAllegati_3_ Moore DA, et al. Natural ventilation for the prevention fileAllegati_itemFile_3_file.pdf. of airborne contagion. PLoS Med. 2007;4(2):e68. 22. Rota MC, Massaria M, Gabutti G, Guido M, De Donno 11. Cohen B, Hyman S, Rosenberg L, Larson E. Frequency A, Ciofi ML. Measles serological survey in the Italian of patient contact with health care personnel and visitors: population: interpretation of results using mixture model. implications for infection prevention. Jt Comm J Qual Vaccine. 2008;26:4403-9. Patient Saf. 2012;38(12):560-5. doi:10.1016/s1553- 23. Freund R, Krivine A, Prévost V, Cantin D, Aslangul E, 7250(12)38073-2 Avril MF, Claessens YE, Rozenberg F, Casetta A, Baix- 12. Beggs CB, Shepherd SJ, Kerr KG. Potential for airborne ench MT, Dumaine V, Launay O, Loulergue. Measles im- transmission of infection in the waiting areas of healthcare munity and measles vaccine acceptance among healthcare premises: stochastic analysis using a Monte Carlo model. workers in Paris, France. J Hosp Infect. 2013;84(1):38- BMC Infect Dis. 2010;10:247. doi: 10.1186/1471-2334- 43. doi: 10.1016/j.jhin.2013.01.002 10-247 24. Steingart KR, Thomas AR, Dykewicz CA, Redd SC. 335 Lack of protection for measles among Italian nurses

Transmission of measles virus in healthcare settings dur- of measles, mumps, rubella and varicella antibodies in ing a communitywide outbreak. Infect Control Hosp Epi- the United States population, 2009-2010. Open Forum demiol. 1999;20:115-9. Infect Dis. 2015;2 (1):ofv006. doi: 10.1093/ofid/ofv006. 25. Bechini A, Boccalini S, Tiscione E, Pesavento G, Man- 32. Rosario G, Gareca M, Kincaid H, Knouse MC. Using nelli F, Peruzzi M, et al. Progress towards measles and locally derived seroprevalence data on measles, mumps, rubella elimination in Tuscany Italy: the role of popula- rubella, and varicella by birth cohort to determine risks tion seroepidemiological profile. Eur J Public Health. for vaccine-preventable diseases during international 2010;22(1):133-9. travel. Travel Med. 2015;22(6):396-402. doi: 10.1111/ 26. Socan M, Berginc N. High seroprevalence of varicella, jtm.12235. reviews

measles, mumps, rubella and pertussis antibodies in 33. Coppeta L, Balbi O, Baldi S, Pietroiusti A, Magrini A. first-grade medical students. Wien Klin Wochenschr. Pre-vaccination IgG screening for mumps is the most cost-

2008;120(13-14):422-6. doi: 10.1007/s00508-008-1005- effectiveness immunization strategy among health care and 6 workers. Hum Vaccin Immunother. 2019;15(5):1135-8. 27. Coppeta L, Pietroiusti A, Lieto P, Ferraro M, Grelli S, doi: 10.1080/21645515.2018.1564442 Stillo M, Magrini A. Measles immunity in an Italian 34. Asari S, Deguchi M, Tahara K, Taniike M, Toyokawa M, teaching hospital. Occup Med. 2019;69(2):143-5. doi: Nishi I, et al. Seroprevalence survey of measles, rubella,

10.1093/occmed/kqy132 varicella, and mumps antibodies in healthcare workers articles 28. Porretta A, Quattrone F, Aquino F, Pieve G, Bruni B, and evaluation of a vaccination program in a tertiary care Gemignani G, Vatteroni ML, Pistello M, Privitera GP, hospital in Japan. Am J Infect Control. 2003;31:157-62. Lopalco PL. A nosocomial measles outbreak in Italy, 35. Coppeta L, Morucci L, Pietroiusti A, Magrini February-April 2017 Eurosurveillance. 2017;22(33). A.Cost-effectiveness of workplace vaccination against 29. Botelho-Nevers, Cassir N, Minodier P, Laporte R, Gau- measles. Hum Vaccin Immunother. 2019;1-4. doi: riginal tret P, Badiaga S, Thiberville DJ, Ninove L, Charrel R, 10.1080/21645515.2019.1616505 O Brouqui P. Measles among healthcare workers: a poten- 36. Coppeta L, Biondi G, Perrone S, Pietroiusti A. Suscep- tial for nosocomial outbreaks. Euro Surveill. 2011;16(2). tibility to measles among healthcare workers: a cross- 30. Coppeta L, Pietroiusti A, Morucci L, Neri A, Fer- sectional serological study. Infect Dis. 2020;52(6):443-5. raro M, Magrini A. Workplace vaccination against doi: 10.1080/23744235.2020.1739746 measles in a teaching hospital of Rome. J Hosp Infect. 37. Baer G, Bonhoeffer J, Schaad UB, Heininger U. Sero- 2019;101(3):364-5. doi: 10.1016/j.jhin.2018.11.022 prevalence and immunization history of selected vac- 31. Lebo E, Kruszon Moran D, Marin M, Bellini WJ, Schmid cine preventable diseases in medical students. Vaccine. S, Bialek SR, Wallace GS, McLean HQ. Seroprevalence 2005;23(16):2016-20. 336 Original articles and reviews poses that mutant tumor cells with a proliferative ad- context oftheCSCmodeland stochasticmodel,pro- describes tumor evolution that may occur both in the erogeneous tumor cells [2]. The clonal evolution model ment, resultinginthegeneration ofphenotypicallyhet- epigenetic factors related to the tumor microenviron- cell canbeinfluencedinavariablewaybygeneticand of tumorheterogeneity;thepotentialeach renewal ordifferentiation,thusgeneratingacondition have theabilitytobetumorigenicandcapableofself- this theory, thestochasticmodelproposesthatallcells phenomenon of tumor dormancy [1]. At variance with py basedonchemicaldrugsorradiationsandthe large majorityoftumorsrelapseafteraninitialthera- through thestudyofcancerpatients,suchasthat cell theoryexplainsmanyclinicalobservationsmade progeny thatisnon-tumorigenic[1].Thecancerstem of tumors,throughtheircapacitytodifferentiateintoa are capableofself-renewal and long-termmaintenance populations; likenormalstemcells,cancercells cells withstemcellfeatures,distinctfromothersub- sues, aremaintainedthroughthedifferentiationoffew ago andsuggestedthatcancers,aswellnormaltis- as thehierarchicalmodel,wasproposedseveralyears profiles. Thecancerstemcell(CSC)theory, knownalso repopulation capacitiesandtypicalgeneexpression tiate andmaintaintumors,long-termself-renewal stem cell-likeproperties,includingthecapacitytoini- THEORY INTRODUCTION: THECANCERSTEMCELL veloping efficientstrategiesofcancerstemcelltargeting. eradication. Ongoingstudiesinsometumorsstronglysupporttheclinicalutilityofde- tional anticancertherapieswithdrugstargetingcancerstemcellscouldleadto anti-cancer therapiesandthereissupportingevidencethatthecombinationofconven- fined asafixedproperty, duetocancerplasticity. according tospecificimmunophenotypicfeatures,butcancerstemnesscannotbede- to tumorrecurrenceandmetastasis.Cancerstemcellscanbeidentifiedinmanytumors inter-tumoral heterogeneity and the main cellular mediators of drug resistance, leading growth, areabletoself-renewanddifferentiate,themaindriversofintra- cer stemcells.Cancercellshavethepropertyofinitiatingandmaintainingtumor last threedecadessupporttheexistenceofstemcellsalsointumors,so-calledcan- a populationoffunctionallymaturetissueelements.Recentstudiescarriedoutinthe stem cells, that have the unique property of self-renewal and differentiation to generate Abstract Dipartimento diOncologia,IstitutoSuperioreSanità,Rome,Italy Ugo Testa, ElviraPelosiandGermanaCastelli Cancer stemcelltargetedtherapies DOI: 10.4415/ANN_20_03_12 Ann IstSuperSanità2020|Vol. 56,No.3:336-350 [email protected]. Address forcorrespondence: UgoTesta, DipartimentodiOncologia,Istituto SuperiorediSanità,Viale ReginaElena 299,00161Rome,Italy. E-mail: Growing evidences indicate that tumor cells display For theseproperties,cancerstemcellsrepresentattractivetargetsfordevelopingnew In normaldividingtissues,cellhomeostasisismaintainedbyrarecellularelements,the markers (CD34 evidence thatthesecellsdisplay stemcellmembrane of CSCsinacutemyeloidleukemia(AMLs),providing This assaywasinitiallyappliedtothecharacterization and currentlyconsideredasanassayofstemness[3]. into lethallyirradiatedrecipientimmunodeficientmice to reconstitutehematopoiesisfollowingtransplantation based ontheevaluationofcapacityagivencell torical studiesonhematopoieticstemcells(HSCs)and immunodeficient mice.Thisassaywasderivedfromhis- based on xenotransplantation assays performed into LEUKEMIC STEMCELLS display propertiesofCSCs[2]. multiple cellular states within the cells that functionally Thus, accordingtothisnewview, thereiscoexistenceof capacity, chemoresistance andmetastaticpotential[2]. culiar malignantcharacteristics,includingtumorigenic as astateorprocessbywhichcancercellsgainsomepe- ered nonasadistinctsubpopulationoftumorcells,but and CSCproperties,theidentityofCSCs,consid- property ofthesecellstotranslatebetweennon-CSC the cancerstemcellplasticity, relatedtothepeculiar passes thesedivergentmodelsandtakesintoaccount an expansiveprocess[2].ThecurrentviewofCSCsby- vantage outgrowothercellsandprogressivelyundergo tance because it was propaedeutic for the study of CSCs tance becauseitwaspropaedeutic forthestudyofCSCs tion ofleukemicstem cells wasoffundamentalimpor normal HSCs[4]. The assaycurrentlyusedtoassessCSCactivitywas The developmentofthestudies onthecharacteriza- + CD38 - ) identicaltothoseobserved in • • • • • • Key words solid tumors leukemia stem cells targeted therapy cancer stemcells cancer - 337 Cancer stem cell targeted therapies

of other tumors. Particularly, these studies have contrib- expression (<2% of CD34+ cells), <10% of undifferenti- uted to define the main determinants and parameters ated leukemic blasts and are enriched in nucleophos- essential for leukemic stem cell detection. Thus, it was min1 (NPM1)-mutated AMLs. In CD34- AMLs, both shown the essential role of the xenotransplantation as- CD34+ and CD34- cells act as leukemic stem cells and say for leukemic stem cell identification, related to the display similar molecular properties, thus suggesting immunodeficient mouse strains (NOD, NOD/SCID or that CD34+ and CD34- LSCs in these AMLs represent NOD/SCID/IL2R-g mice, with the more immunode- the same cells displaying plasticity in CD34 expression ficient mice being more recipient for engraftment and [17]. In line with these findings, in NPM1-mutated

- + reviews growth of leukemic stem cells) [reviewed in 5] or to the AMLs both some CD34 and CD34 cells are able to type of leukemic samples assayed for leukemic stem cell have the function of LSCs [18, 19].

content [6] or to the procedure of the xenotransplanta- The study of LSCs in NPM1-mutated AMLs allowed and tion assay (thus, using a humanized bone marrow os- also to evaluate their prognostic impact (the presence of sicle xenotransplantation model 87% of AML samples an increased level of CD34+/CD38- cells together with engrafted to high levels of bone marrow chimerism [7]. the positivity for leukemic molecular-specific markers, Thus, these studies have highlighted that the frequency such as mutated-NPM1 in minimal residual disease articles of leukemic stem cells (LSCs) varied consistently among was a negative prognostic factor) [19]; furthermore, the various leukemic samples and xenotransplantation mod- study of LSC subpopulations in NPM1-mutated AMLs els, thus indicating that the evaluation of LSCs is highly allowed to perform a reconstruction of the clonal archi- dependent on the model used [8]. tecture and dynamics of these leukemias with identifi- Furthermore, the studies carried out on the charac- cation of a putative pre-leukemic subclone [20]. These riginal terization of LSCs have contributed to show that these studies unequivocally showed that LSCs responsible for O cells are heterogeneous and can be identified according leukemia development in mice bear NPM1-mutations, to a membrane phenotype. Particularly, according to in line with the observation that these LSC subclones the expression of CD34, an antigen expressed on nor- are responsible for leukemia relapse [21]. mal HSCs and hematopoietic progenitor cells (HPCs), The studies on the characterization of LSCs in pri- AMLs can be subdivided into CD34+ (about 75% of mary AML samples showed that their number is highly cases) and CD34- (about 25%). Basically, these studies heterogeneous, ranging from 1 in 10 to 1 in 106 bulk showed that LSCs are rare cellular elements accounting AML cells and is significantly higher in poor and inter- for about 1 per 1×106 leukemic blasts and characterized mediate-risk AML patient samples at diagnosis than in for their property to give rise to leukemic engraftment the good-risk AML group [22]. that could be propagated for multiple serial transplants, Thus, these studies have shown the consistent het- thus supporting that these cells have self-renewal prop- erogeneity in the phenotypes of engrafting AML stem erties and produce a non-LSC cell progeny [9]. cells; furthermore, it is evident that the AML LSCs LSCs have been explored in detail in CDF34+ AMLs, engrafting immunocompromised mice do not neces- showing that these cells were predominantly resident sarily represent the founder AML clones responsible in the CD38+/CD38- fraction, but in 50% of cases were for leukemia relapse. The heterogeneity of LSCs is fur- present also at the level of CD34+/CD38+ fraction and, ther supported by a recent study showing that CD34+/ more rarely, at the level of CD34- cells [9]; in the major- CD38- cells present in various CD34+ AMLs are het- ity of these patients it was observed the coexistence of erogeneous and correlates with genetic risk groups and CD34+/CD38- cells resembling stem/multipotent pro- outcomes [23]. Importantly, these authors showed that genitor cells and of CD34+/CD38+ fraction, resembling the AMLs bearing immature LSCs defined as CD34+/ granulocyte-macrophage progenitors [10]; in line with CD38-/ALDH (aldheyde dehydrogenase)high showed a these observations, the CD34+/CD38- fraction possess- shorter overall survival, event-free survival and com- es a higher LSC content and CD34+/CD38- cells have plete remission rates [23]. In contrast, AML cells in a gene expression profile different from CD34+CD38+ which immature LSCs (CD34+CD38-) displayed more cells [10]. These findings have promoted parallel stud- mature phenotypes were associated with better out- ies aiming to define the immunophenotypic features comes [23]. Therefore, the most immature phenotype of CD34+/CD38- LSCs, providing evidence that sev- of LSCs represents a clinically relevant biomarker of eral cell membrane markers are upregulated on these negative clinical outcome [23]. In line with these ob- cells, including CD123, CD47, TIM3, CD96, CLL-1, servations, AMLs displaying CD34+/CD38-/ALDH- GPR56 and CD93 [11, 12]. CD123 was extensively ex- high LSCs have increased expression of LSC and HSC plored in AMLs, showing that its expression increases signatures, while AMLs with CD34+/CD38-/ALDHlow at relapse and is particularly overexpressed in some cells are enriched for a progenitor signature [24]. These AML subtypes, such as AMLs bearing FLT3-ITD (in- observations support the initial studies by Cheung et ternal tandem duplication) mutations [13-15]. More al. showing that high ALDH activity in AML blasts recently, it was shown that CD99 is a cell surface pro- defines a subgroup of leukemias with adverse prog- tein frequently overexpressed on AML stem cells: high nosis and high NOD/SCID engrafting potential [25]. CD99 expression on AML blasts enriches for function- The enrichment in stemness-related transcriptional al leukemic stem cells; furthermore, anti-CD99 mAbs programs in AMLs bearing ALDHhigh LSCs suggests exhibit anti-leukemic activity in AML xenografts [16]. that these leukemias derive from the malignant trans- Other studies have characterized LSCs in CD34- formation of stem cells or immature progenitors and AMLs: these AMLs are characterized by a low CD34 provides an explanation for the therapy resistance and 338 Ugo Testa, Elvira Pelosi and Germana Castelli

poor prognosis of these leukemias. The clinical sig- the progressive shift from a model of cancer stem cells nificance of this immature LSC population in AML is based on the normal tissue developmental hierarchies to strongly supported by a recent clinical study based on a more nuanced model of stemness taking into account the treatment of older AML patients with the B cell both the genetic and nongenetic influences contribut- lymphoma 2 (BCL-2) inhibitor Venetoclax in combina- ing to the functional acquisition of stem-like properties tion with the hypomethylating agent Azacitidine: this by tumor cells and to generation of tumor heterogeneity treatment resulted in deep and durable remissions, due [8] and the role of microenvironment which is a critical to the targeting of LSCs through metabolic damage driver of heterogeneity, plasticity and evolution within reviews

of these cells caused by disruption of the tricarboxylic the CSC population [41]. acid (TCA) cycle [26]. Given the limitations of marker-based selection and

and the plasticity of CSC state, it is evident that develop-

CANCER STEM CELLS IN SOLID TUMORS ment of optimal assays to measure stem cell function in The studies carried out on AMLs provide a clear ex- specimens of human tumors is strictly required. In spite ample of the evolution of the cancer stem cell model. these important limitations, a large number of studies, The use of equivalent cell sorting and xenografting tech- basically based on measuring tumor stemness through articles

niques, has led to the identification of cancer stem cells surface markers, have supported the view that high ex- in various solid tumors, starting from breast cancer, pression of CSC biomarkers in the large majority of where it was shown that as few as 100 CD44+CD24-/ solid tumors is a property associated with poor overall low breast cancer cells initiate tumor growth when trans- survival and/or disease-free survival compared with low riginal planted into immunodeficient mice [27], to colon can- or no expression of CSC biomarkers [42, 43]. Further- O cer [28-30], brain tumors [31], prostate cancer [32], more, elevated stemness biomarkers levels are associ- pancreatic cancer [33], ovarian cancer [34], lung can- ated with decreased tumor differentiation, increased cer [35]. Basically, these studies have shown that CSCs metastasis and advanced tumor stage [42, 43]. isolated from various solid tumors are able to generate tumors in xenograft assays and generate in vivo in the CANCER STEM CELLS immunodeficient animals non-tumor initiating cells, AND TUMOR INITIATION thus suggesting a hierarchical organization of these The cancer stem cell model implies that human tu- tumors [27-35]. The capacity to initiate and propagate mors can derive from the malignant transformation of tumor development varies considerably between differ- tissue stem cells and that these disorders can be preced- ent cells within a solid tumor and this variation is due ed by a stage of premalignant tumorigenesis occurring to a hierarchical relationship between tumorigenic and at the level of stem cells and generating a pre-malignant non-tumorigenic cells [27-35]. These studies have also condition. shown that there are many markers that enrich for can- Studies carried out in many tumors strongly support cer-initiating cells in solid tumors, including CD133 in the view that these tumors are initiated through the various tumors, CD44 in breast cancer, LGR5 in colon malignant transformation of stem cells. Initial stud- cancer, ALDH1 in breast and gastrointestinal cancers, ies carried out in AML patients showed that normal SOX2 in glioblastoma, but the identification of markers HSCs display one or two of the mutations observed in selective and sensitive for CSCs remains largely elusive. leukemic blasts and can be considered as preleukemic The discordant findings in the phenotype and proper- HSCs [43]. These findings have suggested that AMLs ties of CSCs in solid tumors may be also related to the develop through the acquisition of serial mutations at experimental conditions and xenograft assays used to the level of self-renewing HSCs and at the genomic mu- assess the tumorigenic potential. In this context, par- tational level it is suggested that AML genomes clon- ticularly challenging was the debate originated by the ally evolve from founder mutations [44]. Other studies study and characterization of CSCs in human mela- have shown that preleukemic HSCs harbor some, but nomas. In melanoma, cells separated according to the not all, of the mutations observed in the bulk leuke- expression of the neural crest stem cell marker CD271/ mic blasts [45]. Particularly, mutations in landscaping p75/NGFR displayed differing capacities for tumor genes, involved in chromatin changes, such as DNA formation in patient-derived xenografts (PDX) assays, methylation, histone modification and chromatin loop- CD271+ cells being more tumorigenic than CD271- ing are early events during AML development, while cells [36, 37]. However, Quintana et al. have shown mutations in genes involved in proliferation control are that using highly efficient PDX assays [38], 16 cell sur- late events [45]. Furthermore, CD34+ cells at remission face markers, including CD271, were found unable to harbor preleukemic mutations [45]. This conclusion is identify melanoma cells with enriched tumorigenic po- supported by the observation that AML patients where tential [39]. Proposed reasons for the discordant results preleukemic mutations were detected in purified stem/ include differences in the PDX assays used and in other progenitor cells at diagnosis, nearly all of these preleu- experimental procedures [39]. However, a more recent kemic mutations were also detected in CD34+ cells and study by Boyle et al definitely supported the view that also in mature hematopoietic cells [45]. The presence of regardless of the assays used, melanoma does not follow these preleukemic mutations at the level of differentiat- a CSC model associated with CD271 expression [40]. ed cells strongly supports the view that these preleuke- The findings observed in melanoma, as well as the dif- mic HSCs retain a normal differentiative function and, ficulties in defining reproducible phenotypes for these together with normal HSCs surviving to induction che- cells in the large majority of solid tumors supported motherapy, contribute to bone marrow reconstitution 339 Cancer stem cell targeted therapies

during the remission phase [45]. These initial observa- vided evidence that mutations in IDH1, IDH1, TP53, tions were now confirmed through the study of many DNMT3A, TET2 and spliceosome genes increase the AML patients in remission: about 40-50% of these odds of developing AML; importantly, all subjects with AMLs in remission retained mutations with a variant TP53 mutations and IDH1 and IDH2 mutations devel- allelic frequency ≥1-2% and mutation persistence was oped AML [58]. It is important to underline that in this most frequent in DNMT3A, SRSF2, TET2 and ASXL1, study the median time of AML development from the genes whose mutations are associated with preleukemic initial detection of clonal hematopoiesis was 9.6 years stem cells [46, 47]. Thus, these studies provide clear [50]. Although individuals with clonal hematopoiesis reviews evidence about the persistence during chemotherapy- do not have alterations in the number of peripheral induced remission of preleukemic clones, carrying a blood elements under steady-state conditions, under

subset of AML-related gene mutations; the presence of stress conditions, such as anticancer chemotherapy, de- and these mutations during remission is associated with an lay the hematopoietic reconstitution and determine a increased risk of leukemia relapse [45-47]. clinical condition requiring more patient’s transfusional In line with these observation, a recent experimen- support [59, 60]. tal model based on double mutant NPM1/DNMT3A Similar observations were recently reported for other articles knock-in mice, reproduced a condition of AML pro- tumors. Thus, human glioblastoma mutations were ob- gression mimicking leukemia development in humans, served also in normal neural stem cells present at the where leukemia is preceded by a period of extended level of the subventricular zone, away from the primary myeloid progenitor proliferation and self-renewal [48]. tumor [61]. This study provides the first direct evidence This self-renewal can be reversed by a small molecule that neural stem cells in human subventricular zone are riginal acting as an inhibitor of Menin-MLL binding [48]. the cells of origin that develop the driver mutations of O These observations give support to the idea that sub- glioblastoma [61]. Recent studies have shown that so- jects at high-risk of developing AML might benefit matic mutations and clonal expansions are observed from a preventive targeted epigenetic therapy [48]. at the level of most of normal tissues. Thus, studies of The discovery of clonal hematopoiesis provided a immunostaining of normal sun-exposed human skin strong support to the stem cell mutation model of leu- showed the existence of small clones, composed by few kemia development. Clonal hematopoiesis is a com- hundred cells contained TP53 mutations [62]. More re- mon, age-associated condition characterized by the cent studies of deep sequencing carried out on skin tu- expansion of HSC clones carrying recurrent somatic mor biopsies showed that sun-exposed skin cells display mutations and occurring in individuals without diagno- many somatic mutations related to events of ultraviolet sis of hematological malignancies [49, 50]. mutagenesis; furthermore, positive selection in some Clonal hematopoiesis is associated not only with ag- cancer genes occurs, leading to clonal expansions [63]. ing, but also with some age-related conditions, such as Thus, by middle age, sun-exposed skin is composed by high risk of hematological malignancies, cancer mortal- thousand clones with one in every four cells carrying a ity, cardiovascular diseases and inflammatory condi- positively selected cancer mutated gene [63]. Other re- tions. Thus, clonal hematopoiesis varies according to cent studies carried out in normal esophageal epithelium age from <1% in individuals younger than 40 years to showed a mutational burden lower than that detected 15-20% in old individuals (80-90 years) [51]. Clonal bin ultraviolet-exposed skin, but positive selection was hematopoiesis is associated with about 1% annual risk stronger, driving the development of clones carrying mu- of developing a hematological malignant condition and tations in cancer genes (such as NOTCH1 and TP53) a 2-fold to 4-fold higher risk of developing coronary colonizing most of the esophagus by mid age [64, 65]. artery disease, stroke and cardiovascular death, inde- Other recent studies showed tissue-specific mutation ac- pendently of traditional risk factors for cardiovascular cumulation at the level of adult stem cells of colon, small diseases [51, 52]. It is particularly important to point intestine and liver [66]; furthermore, RNA sequence out that the impact of clonal hematopoiesis in the risk analysis revealed macroscopic somatic clonal expansion of developing a hematologic malignancy is greatly influ- occurring at the level of many normal tissues [67]. enced by the context in which it occurs: this risk is low These observations support the stem cell origin of hu- when clonal hematopoiesis occurs in healthy subjects, man cancers and indicate that mutations occurring at whereas is clearly higher when clonal hematopoiesis is the level of the stem cells of normal tissues represent found in the cancer population and its presence in this an in initial event in cancer development, inducing the last setting represents an independent, cancer-related formation of premalignant clones (Table 1). mortality risk [53-56]. Cancer stem cells also essential cellular drivers of tu- Very recent studies have explored more in detail the mor progression, chemoresistance and tumor relapse. clinical significance of gene mutations occurring in pre- Thus, studies of characterization of the properties of leukemic stem cells, suggesting a distinction within the glioma stem cells have strongly supported a model of clonal hematopoiesis of a benign condition not evolv- cancer stem cell-related tumor progression in which ing to a hematological malignancy and a preleukemic slow cycling cancer stem cells give rise to a more cycling condition [57, 58]. Thus, Abelson et al. showed that progenitor cell population with pronounced self-main- pre-AML cases were distinct from control clonal he- taining capacities, which in turn generates non-prolifer- matopoiesis cases because display more mutations per ative more differentiated cells [68, 69]. Chemotherapy sample, higher variant allele frequencies and enrichment facilitates the expression of pre-existing drug-resistant of mutations in specific genes [57]. Desaiet al. pro- glioblastoma stem cells [68, 69]. 340 Ugo Testa, Elvira Pelosi and Germana Castelli

Table 1 Mutations observed in the preneoplastic lesions observed at the level of various tissues; the most recurrent mutations observed in the corresponding developed tumors are also shown (last column) Tumor type Cells of origin Preneoplastic lesions Mutations in preneoplastic Mutations in developed clones tumors Acute myeloid leukemia Hematopoietic Clonal hematopoiesis DNMT3A, TET2, ASXL1 DNMT3A, NPM1, FLT3, stem cells IDH1, IDH2, Esophageal carcinoma Basal Barrett’s esophagus NOTCH1, TP53, NOTCH2, FAT1, TP53, NOTCH1, KMT2D, reviews

esophageal squamous dysplasia NOTCH3 NFE2L2, FAT1, EP300 stem cells and

Colorectal cancer Intestinal stem Aberrant crypt foci AXIN2, ERBB2, PIK3CA, ATM, cell FBXW7, ERBB3, CDK12 Endometrial cancer Endometrial Endometrial hyperplasia PIK3CA, ARHGAP35, FBXW7, PTEN, PIK3CA, ARID1A, stem cell PIK3R1, FOXA2, ZFHX3, ERBB2, PIK3R1, CTNNB1, CTFC, ERBB3 KRAS, RNF43, ARID5B, TP53 articles Lung cancer Brochioalveolar Atypical adenomatous TP53, NOTCH1, FAT1, CHEK2, epithelial stem hyperplasia PTEN, ARID1A, ARID2, IDH1 cells Skin cancer Basal stem cell TP53, NOTCHG1, FAT1, NOTCH2, Squamous: NOTCH1, TP53, riginal NOTCH3 FLNB, NOTCH2, CDKN2A

O Basaloid: TP53, PTCH1, TERT, CDKN2A Melanoma Mature Benign nevus BRAF BRAF, NRAS, NF1, TERT, melanocytes/ CDKN2A Melanocyte stem cell Glioblastomna Subventricular Unknown TERT, EGFR, PTEN, TP53, PDGFR, TERT, TP53, PTEN, EGFR, zone stem cells IDH1, NF1, PIK3CA PDGFR, IDH1 Hepatocarcinoma Dedifferentiated Cirrhotic liver PKD1, KMT2D, STARD9, APOB, TP53, CTNNB1, ALB, PCLO, adult dysplastic hepatic ARSM1, ALB, ARID1A FLG, CSMD3, XIRP2 hepatocytes nodules

The studies carried out in various models have also subtypes are originated from different primary AMLs supported a great plasticity of the cancer stem cell prop- [72]. In all instances, the relapsing clone was character- erties at the level of the cellular elements composing a ized by an increased number of LSCs [72]. tumor. Thus, in human, as well as in murine, colorectal Relapse is an event where CSCs play a key role not cancers, LGR5+ cells were identified as cancer stem cells only in hematopoietic tumors, but also in several solid [70]; depletion of LGR5+ cells from tumor organoids tumors. Several studies support a role for CSCs in re- established from human biopsies determines an initial lapse of many solid tumors. The peculiar properties of tumor regression, followed by tumor regrowth due to CSCs explain in large part their role in tumor recur- generation of new LGR5+ cells from differentiated tu- rence: cancer stem cells largely survive to chemo-radio- mor cells, thus supporting a consistent plasticity of the therapy and after these treatments enter into a condi- CSC properties [70]. Studies in experimental models tion of quiescence and dormancy for different period of of colorectal carcinogenesis support a distinct role for time; however, following changes in their genome due different CSCs during tumor evolution: thus, tumor to their intrinsic genetic instability and in the tumor growth of the bulk tumors is maintained by LGR5- microenvironment these cells are reactivated, become cells that continuously replenish LGR5+ CSCs (in fact, proliferative and are responsible for tumor recurrence LGR5+ cell ablation cannot induce tumor regression); [73]. This model of CSC-mediated cancer recurrence liver tumor metastases are driven by LGR5+ CSCs [71]. is supported by experimental studies and by observa- Cancer stem cells are involved in relapse events, as tions in cancer patients undergoing standard treat- evidenced by the studies on AML patients. The major- ments. Thus, concerning experimental models, particu- ity of AML patients relapse through a cellular mecha- larly interesting was the observation made by Li et al. in nism involving leukemic stem cells belonging to a leuke- lung cancer suspensive tumor model, showing that in mic clone already present at the time of initial diagnosis this tumor high levels of cancer stem cells undergoing [72]. Particularly, two types of AML have been identi- asymmetric self-replicative cell divisions in latent tumor fied: i) the first AML type contains a rare population are the key issue to reactivate this dormant tumor [74]. of LSCs, with a stem-early progenitor-like immunophe- Thus, a change in environmental conditions, represent- notype; ii) the second AML type relapses through the ed by a high level of serum insulin Growth Fcator-1, can main CD33+ leukemic blast cell population and displays induce the quiescent-to-proliferative, progressive tumor growth factor dependency [72]. The study of gene ex- transition through promotion of CSC symmetric divi- pression profile supports that these two AML relapsing sions [74]. 341 Cancer stem cell targeted therapies

Several observations made in cancer patients sup- of cancers related by stemness features [80]. The stem- ported a role for CSCs in disease relapse in various solid ness features revealed the aggregation of high stemness tumors. Thus, Merlos-Suarez et al. observed that a gene tumors across distinct tumor types [80]. signature specific for adult intestinal stem cells pre- By multiplatform analysis of transcriptome, methy- dicts disease relapse in colorectal cancer patients [75]. lome, and transcription factors performed in the large Roy et al showed that disease relapse in head and neck majority of tumors above described, two cancer stem- squamous cell carcinoma patients is associated with in- ness indices were discovered: one was relative of epi- creased p38 MAPK expression in CSCs and p38-inhib- genetic features (mDNAsi); one was relative of gene reviews ited tumor cells show significantly reduced expression expression (mRNAsi) [81]. These two indices allowed of CSC markers [76]. Sun et al. showed that quiescent to evaluate the stemness features of the various tumors

CD13+ CSCs are enriched after chemotherapy in he- associated with oncogenic dedifferentiation [81]. The and patocellular carcinomas and serve as a reservoir for dis- oncogenic dedifferentiation is associated with muta- ease recurrence [77]. tions in genes encoding oncogenes and epigenetic mod- ifiers, perturbations in mRNA/miRNA transcriptional CANCER STEMNESS network, deregulation of signaling pathways and ex- articles

As above discussed, considerable controversy remains pression of genes involved in the control of self-renewal as to how best to define CSCs and the level to which the of normal and cancer stem cells, such as MYC, OCT4, various cancers are organized at cellular level accord- SOX2 [81]. Analysis of tumor microenvironment ing to a hierarchical structure [8]. Growing evidences showed a correlation between cancer stemness and im- support the view that stem cell-associated molecular mune checkpoint expression and infiltrating immune riginal features, defined as “stemness”, are a key biological de- cells [81]. The dedifferentiated oncogenic phenotype O terminant in cancer [8]. The stemness phenotype may was more prominent among metastatic tumors [81]. reflect either the presence of CSCs in a given tumor or, Smith et al have developed gene signature for normal alternatively, the acquisition of stem-like properties by human stem cells and have used these signatures to bet- non-CSC tumor cells or both [8]. However, whatever ter elucidate the relationship between epithelial cancers is the mechanism operating in a tumor responsible for and stem cell transcriptional programs: this approach stemness acquisition, stemness was progressively con- showed that the adult stem cell signature selected can- sidered as a key phenomenon in tumor development for cers with poor overall survival and genetic alterations of its strong association with poor outcomes, as initially oncogenic drivers, such as small neuroendocrine lung shown for acute leukemias [78]. In fact, this initial study, cancer, prostate and bladder cancers [82]. At the level through the definition of a set of genes preferentially ex- of gene expression, DNA methyltransferase expression pressed in leukemic cell populations enriched in LSCs correlated with adult stem cell gene expression signa- identified a 17 gene signature; this 17-gene leukemic ture status [82]. signature allowed to attribute a stemness score to AML, Miranda et al. explored in 22 cancers the relation- highly predictive of initial therapy resistance [78]. ship between cancer stemness, intra-tumoral hetero- The Cancer Genome Atlas (TCGA) network per- geneity and immune response [83]. The results of this formed a detailed molecular analysis of various cancer study showed that the stemness phenotype confers types integrating various molecular analyses involving immunosuppressive properties on tumors, resulting in study of tumor DNA (exome sequencing; genome se- microenvironments scarcely reactive to immunological quencing; DNA methylation and copy number evalu- challenge that foster and maintain intra-tumoral het- ation) RNA (mRNA and microRNA sequencing) and erogeneity were observed [83]. Finally, studies on can- proteins/phosphoproteins. This analysis allowed to cer cells with high stemness features showed that these discover molecular signatures supporting a taxonomy cells have intrinsic immunosuppressive properties [83]. differing from the current organ- and tissue-histology based classification [79]. The identification of inte- CANCER STEM CELL TARGETED THERAPIES grated cancer subtypes sharing mutations, copy num- Given the important role played by CSCs in tumor ber alterations, signaling pathway similarities that in- initiation, progression, relapse and drug resistance, it is fluence the appurtenance of a tumor to a molecular quite obvious that they represent an attractive target in subtype, independently of the tissue of origin or tumor clinical studies (Table 2). stage [79]. This approach showed that at least 10% of Thus, Dalerba et al. investigating the properties of patients can be differently classified and, in some in- colon cancer stem cells have identified caudal-type stances treated, on the basis of molecular taxonomy homeobox transcription factor 2 (CDX2) as a bio- [79]. More recently, the PanCancer study englobed the marker that could be used to quantify the number of multiplatform molecular analysis of 11 286 tumors from undifferentiated colon cancer cells, displaying proper- 33 cancer types [80]. This study provided evidence of ties of cancer stem cells [84]. Particularly, these au- clustering primarily organized by histology, tissue type thors identified biomarkers not expressed in ALCAM/ or anatomic origin; integrative clustering emphasized CD166-positive tumors (with stem-like properties), but the dominant role of cell-of-origin patterns [80]. Simi- present in ALCAM/CD166-negative tumors and iden- larities among histologically or anatomically related tified the CDX2 protein [84]. Only 4% of colon cancers cancers allowed to perform pan-cancer analyses, involv- had lost CDX2 protein expression and these patients ing gastrointestinal, gynecological, squamous cancers; displayed reduced 5-year disease-free survival; stage interestingly, this analysis allowed to define also a group II colorectal cancer patients with CDX2 loss treated 342 Ugo Testa, Elvira Pelosi and Germana Castelli

Table 2 Agents targeting CSC-associated surface markers, altered signaling pathways or mutated molecules in ongoing clinical trials

Drug name Target Disease Clinical Study Current status phase Identification

Vismodegib Hedgehog R/R Medulloblatoma II NCT00939484 Completed pathway Basal cell carcinoma II NCT01700049 Completed Sarcoma II NCT01700049 Completed SCLC II NCT01700049 Completed Pancreatic cancer II NCT01088815 Completed

reviews Ovarian cancer II NCT00739661 Completed Colorectal cancer II NCT00636610 Completed

and Venetoclax BCL2 Venetoclax+Azacitidine AML not eligible for standard therapy III NCT02993523 Active, completed Venetoclax+Azacitidine After allogenic stem cell transplantation III NCT04161885 Ongoing Venetoclax+Decitabine AML high-risk I NCT03844815 Not yet recruiting Venertoclax+Azartidine Or Decitabine AML not eligicle for standard therapy III NCT03941964 Ongoing

articles Venertoclax+Chemotherapy Refractory/relapsing pediatric AML I NCT03194932 Completed Venetoclax+low-dose cytarabine AML not eligible for standard therapy III NCT03069352 Completed Venetoclax+Dinaciclib Refractory/relapsing AML I NCT034844815 Ongoing

Daratumumab (HuMax®-CD38) CD38 Daratumumab Refractory/relapsing multiple myeloma II NCT00574288 Completed

riginal Daratumumab Refractory/relapsing multiple myeloma II NCT02944565 Completed

O Daratumumab Refractory/relapsing multiple myeloma II NCT03871829 Ongoing Daratumumab+Lenalidomide+Dex Multiple myeloma transplant-inelegible III NCT02252172 Ongoing Daratumumab+Lenalidomide Multiple myeloma transplant-inelegible III NCT02195479 Ongoing Daratumumab+Borteziomib+Dex Multiple myeloma transplant-eligible III NCT02541383 Ongoing +Thalidomide

Magrolimab (mAb Hu5F9-G4) CD47 Solid tumors I NCT02216409 Completed TTI-621 Solid tumors I NCT02663518 Recruiting IBI 188 Advanced malignancies I NCT03763149 Recruiting CC-90002 Hematologic neoplasms I NCT02641002 Completed AO-176 Solid tumors I NCT03834948 Recruiting SRF231 Solid tumors I NCT03512340 Recruiting Bivatuzumab mertansine Metastatic breast cancer I NCT02254005 Completed

Tagraxofusp (SL-401) CD123 AML, BPDCN I NCT03113643 Recruiting BPDCN after SCT II NCT04317781 Recruiting CD123+ AML, BPDCN II NCT043342962 Not yet recruiting KHK283 AML I NCT02181699 Completed Talacotuzumab AML III NCT02472145 Completed SGN-CD123A AML I NCT02848248 Terminated IMGN632 AML II NCT03386513 Recruiting XmAb 14045 CD123/CD4 AML II NCT02152956 Recruiting Flotetuzumab (MGD006) CD123/CD3 Refractory/relapsing AML II NCT02152956 Recruiting JNJ-63709178 CD123/CD3 AML III NCT02472145 Completed

Mylotarg (gemtuzumab ozogamicin) CD33 CD33+ refractory/relapsing AML IV NCT03727750 Recruiting Vadastuximab talirine (SGN-CD33A) AML I NCT01902329 Completed IMGN779 AML I NCT02674763 Recruiting

Napabucasin STAT3 Napabucasin+FOLFIRI Metastatic colo-rectal cancer III NCT03522649 Recruiting Napabucasin+low-dose gemcitabinbe Metastatic pancreatic adenocarcinoma III NCT03721744 Recruiting Napabucasin+nab-paclitaxel+ Metastatic pancreatic adenocarcinoma III NCT02993731 Recruiting gemcitabinbe

Ivosidenib IDH1 Ivosidenib Refractory/relapsing AML, MDS I NCT02074839 Recruiting Ivosidenib Myeloid neoplasms I NCT03564821 Recruiting Ivosidenib+chemotherapy Refractory/relapsing AML, MDS I NCT04250051 Recruiting Ivosidenib or enasidenib+chemotherapy Refractory/relapsing AML, MDS III NCT03839771 Recruiting Ivosidenib or enasidenib+chemotherapy Newly diagnosed AML I NCT02632708 Recruiting Ivosidenib or enasidenib+azacytidine AML, elderly patients I/II NCT02677922 Active, not recruiting Ivosidenib+nivolumab Refractory/relapsing AML, MDS II NCT04056910 Recruiting Ivosidenib+venetoclax ± azacytidine Refractory/relapsing AML, MDS I/II NCT03471260 Recruiting Olutasidenib + azacitidine or cytarabine Refractory/relapsing AML, MDS I/II NCT02719574 Recruiting Olutasidenib + ASTX 727 Refractory/relapsing AML, MDS I/II NCT04013880 Recruiting

Enasidenib IDH2 Enasidenib Advanced AML I/II NCT01915498 Completed Enasidenib IDH2-mutant myeloid neoplasms I NCT03515512 Recruiting Enasidenib AML post stem cell transplantation I NCT03728335 Recruiting Enasidenib Refractory/relapsing, high-risk MDS II NCT03744390 Recruiting Enasidenib Pediatric AML II NCT04203316 Recruiting Enasidenib+chemotherapy Refractory/relapsing AML II NCT03881735 Recruiting Enasidenib+azacytidine Refractory/relapsing AML II NCT03683433 Recruiting Enasidenib+azacytidine Refractory/relapsing, high-risk MDS II NCT03383575 Recruiting Enasidenib+azacitidine or AraC AML ≥ 60 years II NCT02577406 Not yet recruiting Enasidenib + CPX-351 Relapsed AML II NCT03825796 Recruiting 343 Cancer stem cell targeted therapies

with adjuvant chemotherapy displayed 5-year improved patients [26]. Particularly, Venetoclax, increasing ROS overall survival compared to those with CDX2 loss not- (reactive oxygen species) production targets leukemia treated with chemotherapy [84]. These findings were stem cells in older AML patients [26]. A clinical study confirmed in a more recent study carried out in a large involving elderly AML patients with a poor prognosis set of colon cancer patients [85]. Particularly, this study and very limited response to standard therapy showed showed that both patients with microsatellite instabil- that about 67% of patients receiving combined therapy ity-positive (corresponding to patients with mutations with Venetoclax and Azacitidine had complete remis- in DNA-repair pathway genes involved in mismatch re- sions, some of the remissions being prolonged [93]. reviews pair) and microsatellite stability CDX2-negative type II This drug represents an important step towards LSC- colon cancers display a negative prognosis [85] Studies targeted therapy. A third drug targeting CSCs, recently

in animal models of colon carcinogenesis have shown approved for the treatment of an AML subtype is repre- and that CDX2 acts as a suppressor of intestinal tumori- sented by IDH inhibitors. These drugs were developed genesis, thus explaining why its loss is associated with with the assumption that targeting leukemic mutations poor-prognosis colorectal cancer [86]. might be an effective strategy to eradicate AML ma- The study of the role of Hedgehog signaling pathway lignant clone. IDH1 and IDH2 mutations are pre-leu- articles in basal cell carcinoma has led to the clinical develop- kemic mutations and the eradication of the leukemic ment of Vismodegib, a specific inhibitor of this path- stem cell clone bearing these mutations could represent way, that targets cancer stem cells. Alterations in the an effective therapy in AMLs characterized by IDH Hedgehog signaling pathway have been involved in the mutations. Targeting IDSH1 with Ivosidenib, a specific pathogenesis of basal-cell carcinoma; most basal cell inhibitor, elicited 30% of complete remissions among riginal carcinomas are treated surgically, but local-advanced IDH1 mutated/relapsed AML patients, with a median O or metastatic basal-cell carcinomas require a medical response of 8 months [94]. Targeting IDH2 with Ena- treatment. Thus, a phase I clinical study using Vismo- sidenib, a specific inhibitor, resulted in an overall re- degib in patients with basal cell carcinoma locally ad- sponse rate of 40.3% and a median response duration vanced or in metastatic stage showed a 43% response of 5.8 months [95]. rate, with 21% of complete response [87, 88]. A long- The study of the effects of Enasidenib on IDH2- term evaluation of this study on 100 patients showed mutant AML is important because provides a number 48.5% of responses in the metastatic group and 60.3% of important indications on a drug targeting leukemia- of responses in the locally-advanced group; median specific alteration present in LSCs. Thus, Enasidenib overall survival was 33.4 months for metastatic patients induced complete remissions with persistence of mu- and not estimable for those with locally-advanced tu- tant IDH2 and normalization of HSC and progenitor mors [89]. These studies have supported the approval compartments, with emergence of functional neutro- of this drug by FDA for treatment of basal cell cancers phils bearing the mutant IDH2 allele [96]. The mas- at advanced stage. Recent studies have explored the sive induction of neutrophil differentiation may induce mechanisms through which Vismogedib could exert its a clinically relevant syndrome, the differentiation syn- inhibitory effect on cancer stem cells of basal cell car- drome that needs careful medical monitoring and treat- cinomas. Despite the consistent efficacy of Vismodegib ment [97]. Using sequential patient samples, the clonal in the treatment of basal cell carcinoma, residual dis- structure of hematopoietic cell populations at different ease persists in some patients; thus, Biehs et al., using a stages of differentiation was determined, showing that model of basal cell carcinoma have shown that Vismo- Enasidenib promoted cell differentiation from termi- degib treatment did not result in complete eradication nal or ancestral mutant clones; relapse arose by clonal and, quiescent, residual tumor progenitor cells undergo evolution or selection of terminal or ancestral clones a switch to a stem-like transcriptional program, re- [98]. The resistance to Enasidenib may be related in sembling that of interfollicular epidermis and isthmus, some instances to the acquisition of new IDH2 muta- whereas untreated tumors resemble hair follicle bulge tions represented by trans or cis dimer-interface muta- [90]. This differentiation switch was related to the acti- tions [99]. The analysis of the molecular abnormalities vation of Wnt pathway: thus, the combined treatment of IDH2 mutant AMLs showed that both IDH2-R140 with both Vismodegib and a Wnt inhibitor reduced the and IDH2-R172 mutations are equally responsive to residual tumor burden and enhanced tumor differentia- Enasidenib [100]. Furthermore, response and survival tion [90]. These findings were supported by a parallel were comparable among patients who, at study entry, study that characterized the slow-cycling population re- were in relapse, or were refractory to intensive or non- sidual after Vismodegib treatment: these residual cells intensive therapies [100]. Finally, a very recent study correspond to LGR5+ cancer stem-like cells, exhibiting reported the first results of a trial carried out in older high Wnt activity [91]. patients with newly diagnosed IDH2-mutant AML and The Hedgehog signaling pathway is upregulated in showing durable responses among responsive (about pancreatic adenocarcinoma cancer stem cells; however, 31%) patients [101]. the administration of Hedegehog inhibitors (Vismo- A promising approach for treatment consists in the degib or GDC-0449) in combination with Gemcitabine targeting of membrane antigens selectively or preferen- to pancreatic cancer patients did not lead to a decrease tially expressed on tumor cells, including CSCs. Par- of CSCs and to tumor inhibition [92]. ticularly, targeted therapies involving CD38, CD47 and Another drug, Venetoclax, a BCL2 inhibitor, was CD123 are under progress in some hematological ma- recently approved for the treatment of elderly AML lignancies, with promising results. 344 Ugo Testa, Elvira Pelosi and Germana Castelli

CD38 is highly expressed on myeloma cells, includ- anti-CD47 blocking antibody is currently being stud- ing a population of myeloma-initiating cells, character- ied in four different phase I clinical studies. The first ized by high expression of CD47, positivity for CD138 results observed on 22 pre-treated chemoresistant lym- expression and negativity for CD19 and CD45 expres- phoma patients were recently reported: the anti-CD47 sion [102]. Myeloma stem cells are also characterized inhibitory mAb Hu5F9-G4, combined with rituximab, by high expression of CD24 [103]. Despite deepening showed significant anti-tumor activity, with 33% com- responses to frontline therapy, most of multiple my- plete responses in diffuse large B-cell lymphomas and eloma patients never become minimal residual disease- 43% complete responses in follicular lymphomas [111]. reviews

negative and relapse with a drug-resistant disease whose A recent study, presented at the last ASCO Meeting, development is mediated by drug-resistant cancer stem reported the results of a phase Ib clinical trial involving

and cells. Bortezomib, thalidomide and dexamethasone the administration of Hu5F9-G4 alone (10 pre-treated

plus autologous stem cell transplantation is standard AML or MDS patients) or in combination with aza- treatment for transplant-eligible patients with newly cytidine (22 untreated AML or MDS patients ineligible diagnosed multiple myeloma; lenalidomide plus dexa- for induction chemotherapy) to AML or MDS patients, methasone or the combination of bortezomib, melpha- showing that this treatment was well tolerated with ro- articles

lan and prednisone are standard treatments for patients bust anti-leukemia activity and induction of complete with multiple myeloma not eligible for autologous stem responses and minimal residual disease negativity [112]. cell transplantation. Three recent clinical studies have Other recent studies have supported the targeting of shown that the addition of Daratumumab, a mono- CD123, the interleukin-3 receptor a (IL-3Ra), as a new riginal clonal antibody anti-CD38, improved the therapeutic therapeutic tool to target leukemic stem cells. CD123 O efficacy of these three standard therapeutic regimens: is widely overexpressed in various hematological ma- a) among patients with newly diagnosed multiple my- lignancies, including AML, B-ALL, CML, blasticplas- eloma who were ineligible for autologous SCT, the risk mocytoid dendritic neoplasm (BPDCN) [113]. Impor- of disease progression or death was significantly lower tantly, CD123 is expressed both at the level of LSCs among those receiving Daratumumab plus lenalido- and more differentiated leukemic blasts and, conse- mide and dexamethasone than among those treated quently, is an attractive therapeutic target [14-16, 114]. with lenalidomide and dexamethasone [104]; b) among Various agents have been developed as drugs targeting patients with newly diagnosed multiple myeloma who CD123 on malignant leukemic cells and on the normal were ineligible for autologous SCT, Daratumumab counterpart. Tagraxofusp (SL-401, Stemline Therapeu- combined with bortezomib, melphalan and prednisone tics), a recombinant protein composed of a truncated elicited a lower risk of disease progression or death than diphteria toxin payload fused to IL-3, was introduced the same regimen without Daratumumab [105]; c) Da- in therapy and was approved for use in patients with ratumumab administration before or after the standard BPDCN, a rare clinical condition characterized by high regimen plus SCT improved depth of response and CD123 expression at the level of leukemic blast cells, progression-free survival compared to standard regimen including LSCs [114]. This compound is under inves- plus SCT [106]. tigation for the treatment of other hematological ma- CD47 and its inhibitory receptor SIRPa form an in- lignancies. Various monoclonal antibodies anti-CD123, nate immune checkpoint that can be targeted using including bispecific monoclonal antibodies, are under anti-CD47 mAb; this system is formed by CD47 ex- evaluation for the treatment of AML minimal residual pressed on tumor cells and the inhibitory receptor SIR- disease or of relapsing/refractory AML. Finally, recent Pa, which is selectively expressed on myeloid cells, par- studies are exploring the potential therapeutic impact ticularly macrophages. CD47 is a “don’t eat me” signal of T cell expressing CD123 chimeric antigen receptors because it inhibits the phagocytosis of nonmalignant (CART) as a new immunotherapy for the treatment cells, such as red blood cells; inhibition of CD47-SIRPa of relapsing/refractory AML and BPDCN. The most promotes the lysis of opsonized cancer cells, often over- consistent clinical experience was performed with the expressing CD47, by macrophages and granulocytes CART 123 reported by Mardiros et al. in 2013 [115] [107]. CD47 up-regulation is an important mechanism and developed as a clinical drug by the Mustang Bio providing protection to normal HSCs during inflam- Inc. and called MB-102. Using MB-102, 7 AML and mation-mediated mobilization; importantly, leukemic 2 BPDCN relapsing patients were treated, with some stem/progenitor cells co-opt this capacity constitutively patients achieving a complete response [116, 117]. No over-expressing CD47 and thus exhibiting the ability to major toxicities were observed in these 9 patients [116, evade macrophage killing [108]. Furthermore, chemo- 117]. In 2018, the Food and Drug Administration has resistant leukemic cells overexpress CD47 and CD123 granted Orphan and Drug Designation to MB-102 for [109]. the treatment of BPDCN. High CD47 expression was observed in non-Hodgkin Current studies are attempting to target cancer stem lymphomas (NHLs) and correlates with negative prog- cells present in different tumors through the targeting nosis [110]. Blocking anti-CD47 antibodies enabled of signaling pathways that are activated in these cells phagocytosis of NHL cells and synergize with anti- and are essential for their survival and/or proliferation CD20 monoclonal antibody rituximab [110]. In mouse (such as PI3K/AKT, STAT, WNT/b-catenin, NOTCH) lymphoma models, the combined administration of an- or transcription factors (such as YAP1) essential for ti-CD47 and rituximab led to elimination of lymphoma tumorigenesis and maintenance of cancer stemness and to a curative effect [110]. Hu 5F9, a humanized [118]. One example is given by the ongoing clinical 345 Cancer stem cell targeted therapies

studies based on the use of STAT3 inhibitors. STAT3 is a cient strategies to obtain at clinical level the inhibition potential target of anticancer therapy because this tran- of cancer stem cells of frequent solid tumors is not sur- scription factor promotes stem cell-like characteristics, prising and is mainly related to the aggressive nature of survival, proliferation, metastatic potential and immune these cells. Therefore, a better understanding of the bi- evasion of tumor cells [119]. STAT3 is hyperactivated ology of cancer stem cells is absolutely required for the in gastrointestinal tumors, where it represents a par- development of possibly efficient therapeutic strategies. ticularly attractive potential therapeutic target. Particu- An example is given by the study of cancer stem cells larly, studies using Napabucasin (BB1608 or BB608), a in pancreatic cancer. This tumor is a prototype of drug reviews small-molecule STAT3 inhibitor, have shown an inhibi- resistance, with only about 30% of patients responding tion of STAT3-induced gene transcription and of tumor to current multidrug chemotherapy regimens: however,

spherogenesis [120]. In a mouse model of colon cancer these responses are quickly followed by tumor resur- and tumorigenesis, Napabucasin inhibited spleen and liver gence and progression. Tumor relapse was related to the metastases and inhibited cell signaling pathways, such presence of chemoresistant tumor cells and, notably, of as those implying NANOG, SOX2, MYC, b-catenin, cells that can be identified by some markers, including and supporting cancer stemness [120]. Several clini- CD24+/CD44+/ESA+ [33], CD133 [126], c-Met [127], articles cal trials are investigating the safety and efficacy of Nestin [128], DCLK1 [129] and Musashi [130]. Imag- Napabucasin in various gastrointestinal malignancies. ing studies directly support the role of cells identified Napabucasin monotherapy was investigated in a phase as putative cancer stem cells in tumor progression and III clinical trial (CO.23 trial), comparing its efficacy to drug resistance [130]. A recent study based on a com- that of a placebo in refractory advanced colon cancer: prehensive molecular evaluation of the core dependen- riginal the study failed to demonstrate a significant difference cies of pancreatic cancer stem cells by integrating their O in the survival of the patients treated with Napabuca- transcriptomic, epigenetic and genomic landscape, al- sin compared to those treated with placebo [121]. In lowed to identify dependence of these cancer stem cells a biomarker-guided analysis, pSTAT3-positive colorec- on inflammatory and immune mediators [131]. How- tal cancer patients showed a significant gain in overall ever, many potential hurdles pose an obstacle in the survival compared to those treated with placebo [121]. development of pancreatic cancer stem cells-targeted The CanStem303C trial (NCT02753127) is an ongoing therapy: i) consistent phenotypic and functional hetero- randomized phase III clinical study evaluating Napabu- geneity of cancer stem cells; ii) cancer stem cells plastic- casin in combination with 5-fluorouracile, leucovorin, ity, giving the opportunity to non-stem cancer cells to irinotecan (FOLFIRI) in previously treated colorectal transdifferentiate into new cancer stem cells; iii) inten- cancer patients [122]. sive desmoplastic stroma around tumor cells, reducing In pancreatic cancer, an initial phase I/II study drug penetration [132]. showed promising results derived from the adminis- The therapeutic targeting of CSCs in solid tumors tration of Napabucasin in association with paclitaxel is a very complex problem due their intrinsic plasticity and gemcitabine [123]. These promising results pro- and to the absence of reliable and stable markers for moted the development of a randomized phase III trial, their identification. A remarkable example is given by CanStem111P, using the same drug combinations in the study of colon CSCs. Colon CSCs express a variety treatment-naïve pancreatic cancer patients with meta- of markers, including CD133 [28, 29], CD44, CD166 static pancreatic adenocarcinoma (NCT02993731). [30], ALDH [133] and LGR5 [75]. Among these vari- However, very recently, this study was stopped for clini- ous markers particularly interesting are the properties cal futility. Similarly, no clinical benefit related to Napa- of LGR5, a marker of Wnt/β-catenin-dependent adult bucasin administration in patients with hepatocellular stem cells of the colon and a regulator of the Wnt carcinoma or gastric cancer [124]. pathway [134]. Lineage-tracing experiments provided These studies are a paradigmatic example of the con- evidence about the existence of LGR5+CSCs in un- sistent difficulties encountered in translating into clinic, perturbed colorectal cancers: these cells isolated from with patient’s benefit, preclinical observations support- organoid-derived xenografts express a gene program ing the targeting of a signaling pathway involved in can- typical of intestinal stem cells and propagate colon cer stemness. This difficulty is seemingly related to the tumors to recipient mice with great efficiency; fur- existence of compensatory mechanisms rendering can- thermore, these cells displayed both self-renewal and cer stem cells not strictly dependent for their survival differentiation capacities [135]. However, studies car- from the inhibited pathway. Another difficulty encoun- ried out in mouse tumors engineered to recapitulate tered in this type of approach is related to the extent the clinical progression of human colorectal cancer of inhibition of the pathway needed to be achieved to showed that selective LGR5+ cell ablation markedly obtain a significant therapeutic effect and its compat- inhibits the growth of primary tumors, but does induce ibility with toxicity problems. Another element of addi- complete tumor regression and tumors restart to grow tional complexity is related to the unexpected induction upon treatment cessation, due to LGR5-cells that re- of secondary effects of some inhibitors: thus, clinically- plenish the LGR5+ pool [71]. These observations have used MEK1/2 inhibitors inadvertently increase Wnt ac- modified the traditional view of a hierarchical organi- tivity and induce stem cell plasticity of colorectal cancer zation of colon cancers which was replaced by a more stem cells, thus revealing an important limiting side ef- dynamic model implying the existence of different cell fect induced by RAS pathway inhibition [125]. types within the tumor that can act as sources of CSCs The difficulties observed in the development of effi- [136]. In line with this view, a recent study showed 346 Ugo Testa, Elvira Pelosi and Germana Castelli

that most colorectal cancer metastases are seeded by of ribosomal transcription and protein biosynthesis in LGR5- cells, which possess the intrinsic capacity to dif- colorectal cancers occurs in a limited subset of tumor ferentiate into LGR5+CSCs independently of a specif- cells, LGR5+and LGR5-, characterized by elevated ic microenvironment and restore epithelial hierarchies levels of the RNA polymerase I subunit A; genetic ab- in metastatic tumors [137]. Thus, it is not surprising lation of these calls cause an irreversible growth arrest that several clinicopathologic studies have reached the of colorectal cancer cells [140]. conclusion that LGR5 expression in colorectal cancer is not associated with a poor prognosis, as might be Conflict of interest statement reviews

expected for a CSC marker [138, 139]. These observa- The authors declare that there is no conflict of inter- tions suggest that the best approach to eradicate co- est regarding the publication of this paper.

and lon CSCs would be based on targeting their function

rather than their identity. In line with this idea, a very Received on 31 August 2019. recent study by Morral et al. showed that the majority Accepted on 15 May 2020. articles REFERENCES

1. Battle E, Clevers H. Cancer stem cell revisited. Nat Med. 16. Chung SS, Eng WS, Hu W, Khalaj M, Garrett-Bakelman

riginal 2017;23:1124-34. FE, Tavakkoli M, et al. CD99 is a therapeutic target on 2. Hung HF, Yang T, Kao SY. Cancer stem cell theory: are we disease stem cells in myeloid malignancies. Sci Transl O moving past the mist? J Chin Med Assoc. 2019;82:814-8. Med. 2017;9:374. 3. Doulatov S, Notta F, Laurenti E, Dick JER. Hematopoi- 17. Quek L, Otto GW, Garnett C,Lhermitte L, Karamitros esis: a human perspective. Cell Stem Cell. 2012;10:120- D, Stoilova B, et al. Genetically distinct leukemic stem 36. cells in human CD34- acute myeloid leukemia are ar- 4. Bonnet D, Dick JE. Human acute myeloid leukemia is rested at a hemopoietic precursor-like stage. J Exp Med. organized as a hierarchy that originates from a primitive 2016;213:1513-35. hematopoietic cell. Nat Med. 1997;3:730-7. 18. Martelli MP, Pettirossi V, Thiede C, Bonifacio E, Mezza- 5. Goyama S, Wunderlich M, Mulloy JC. Xenograft soma F, Cecchini D, et al. CD34+ cells from AML with models for normal and malignant stem cells. Blood. mutated NPM1 harbor cytoplasmic mutated nucleo- 2015;125:2630-40. phosmin and generate leukemia in immunocompromised 6. Wang K, Sanchez-Martin M, Wang X, Knapp KM, Koche mice. Blood. 2010;116:3907-22. M, Vu L, et al. Patient-derived xenotransplants can reca- 19. Taussig DC, Vargaftiog J, Mairaki-Moud F, Griessinger pitulate the genetic driver landscape of acute leukemia. E, Sharrock K, Luke T, et al. Leukemia-initiating cells Leukemia. 2017;31:151-8. from acute myeloid leukemia patients with mutated 7. Reinisch A, Thomas D, Corces MR, Zhang X, Gutzinger nucleophosmin reside in the (CD34-) fraction. Blood. D, Hong WJ, et al. A humanized bone marrow model 2010;115:1976-84. enables improved engraftment of healthy and leukemic 20. Zeijlemaker W, Grob T, Meijer R, Hanekamp D, Kelder human hematopoietic cells. Nat Med. 2016;22:812-21. A, Carbaat-Ham JC, et al. CD34+CD38- leukemic stem 8. Thomas D, Majeti R. Biology and relevance of human cell frequency to predict outcome in acute myeloid leuke- acute myeloid leukemia stem cells. Blood. 2017;129:1577- mia. Leukemia. 2019;33:1102-12. 85. 21. Potter N, Miraki-Moud F, Ermini L, Titley I, Vijayaragha- 9. Kreso A, Dick JE. Evolution of the cancer stem cell mod- van G, Papaemmanuil E, et al. Single cell analysis of el. Cell Stem Cell. 2014;14:275-91. clonal architecture in acute myeloid leukemia. Leukemia. 10. Goardon N, Marchi E, Atzberger A, Quek L, Schuh A, 2019;33:1113-23. Soneji S, et al. Coexistence of LMPP-like and GMP-like 22. Griessinger E, Anjos-Alfonso F, Vargaftig J, Taussig DC, leukemia stem cells in acute myeloid leukemia. Cancer Lassailly F, Prebet T, et al. Frequency and dynamics of Cell. 2011;19:138-52. leukemia-initiating cells during short-term ex vivo culture 11. Pelosi E, Castelli G, Testa U. Targeting LSCs through informs outcomes in acute myeloid leukemia patients. membrane antigens selectively or preferentially expressed Cancer Res. 2016;76:2082-6. on these cells. Blood Cells Mol Dis. 2015;55:336-46. 23. Jones CL, Stevens BM, D’Alessandro A, Reisz JA, Culp- 12. Haubner S, Perna F, Kohnke T, Schmidt C, Berman S, Hill R, Nemkov T, et al. Inhibition of amino acid me- Augsberger C, et al. Coexpression profile of leukemic tabolism selectively targets human leukemia stem cells. stem cell markers for combinatorial targeted therapy in Cancer Cell. 2018;34:724-40. MAL. Leukemia. 2019;33:64-74. 24. Blume R, Rempel E, Manta L, Saeed BR, Wang W, Raffel 13. Testa U, Riccioni R, Militi S, Coccia E, Stellacci E, S, et al. The molecular signature of AML with increased Samoggia P, et al. Elevated expression of IL-3Rain acute ALDH activity suggests a stem cell origin. Leuk Lympho- myelogenous leukemia is associated with enhanced blast ma. 2018;59:2201-10. proliferation, increased cellularity, and poor prognosis. 25. Cheung AM, Wan TS, Leung JC, Chan LY, Huang H, Blood. 2002;100:2980-8. Kwong YL, et al. Aldheyde dehydrogenase activity in 14. Ho TC, LaMere M, Stevens BM, Ashton JM, Myers JR, leukemic blasts defines a subgroup of acute myeloid leu- O’Dwyer KM, et al. Evolution of acute myelogenous leu- kemia with adverse prognosis and superior NOD/SCID kemia stem cell properties after treatment and progres- engrafting potential. Leukemia. 2007;21:1423-30. sion. Blood. 2016;128:1671-8. 26. Pollyea DA, Stevens BM, Jones CL, Wintery A, Pei S, 15. Al-Mawali A, Gillis D, Lewis I. Immunoprofiling of leu- Minhajaddin M, et al. Venetoclax with azacytidine dis- kemic stem c ells CD34+/CD38-/CD123+ delineate rupts energy metabolism and targets leukemia stem FLT3/ITD-positive clones. J Hematol Oncol. 2016;9:61. cells in patients with acute myeloid leukemia. Nat Med. 347 Cancer stem cell targeted therapies

2018;24:1859-66. man acute myeloid leukemia affect epigenetic regula- 27. Al-Hajj M, Wicha MS, Benito-Hernandez A, Mor- tors and persist in remission. Proc Natl Acad Sci USA. rison SJ, Clarke MF. Prospective identification of tu- 2014;111:2548-53. morigenic breast cancer cells. Proc Natl Acad Sci USA. 46. Rothenberg-Thurley M, Amler S, Goerlich D, Kohnke T, 2003;100:3983-8. Konstandin ND, Schneider S, et al. Persistence of leu- 28. O’Brien CA, Pollet A, Gallinger S, Dick LE. A human kemic clones during first remission and risk of relapse in colon cancer cell capable of initiating tumor growth in acute myeloid leukemia. Leukemia. 2018;32:1598-608. immunodeficient mice. Nature. 2007;445:106-10. 47. Morita K, Kantarjian HM, Wang F, Yan Y, Bueso-Ramos 29. Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, C, Sasaki K, et al. Clearance of somatic mutations at re- reviews

Todaro M, Peschle C, et al. Identification and expan- mission. J Clin Oncol. 2018;36:1788-97. sion of human colon-cancer-initiating cells. Nature 48. Uckelmann HJ, Kim SM, Wong EM, Hatton C, Giovin-

2007;445:111-5. azzo H, Gadrey JY, et al. Therapeutic targeting of pre- and 30. Dalerba P, Dylla SJ, Park IK, Liu R, Wang X, Cho RW, et leukemia cells in a mouse model of NPM1 mutant acute al. Phenotypic characterization of human colorectal can- myeloid leukemia. Science 2020;367:586-90. cer stem cells. Proc Natl Acad Sci USA. 2007;104:10158- 49. Shlush LI. Age-related clonal hematopoiesis. Blood. 63. 2018;131:496-504.

31. Singh SK, Clarker ID, Terasaki M, Bonn VE, Hawkins 50. Calvillo-Arguelles O, Jaiswal S, Shlush LI, Moslehi JJ, articles C, Squire J, et al. Identification of human brain tumor Schimmer A, et al. Connections between clonal hema- initiating cells. Nature. 2004;432:396-401. topoiesis, cardiovascular disease, and cancer: a review. 32. Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ. JAMA Cardiol. 2019;4:380-7. Prospective identification of tumorigenic prostate cancer 51. Jaiswal S, Fontanillas P, Flannick J, Manning A, Grau- stem cells. Cancer Res 2005;65:10946-51. man PV, Mar BG, et al. Age-related clonal hematopoi- riginal 33. Li C, Heidt DC, Dalerba P, Burant CF, Zhang L, Ad- esis associated with adverse outcomes. N Engl J Med. O say V, Wicha M, et al. Identification of pancreatic cancer 2014;37:2488-98. stem cells. Cancer Res. 2007;67:1030-7. 52. Jaiswal S, Natarajan P, Silver AJ, Gibson CJ, Bick AG, 34. Zhang S, Balch C, Chan MW, Lai HC, Matei D, Schilder Shvartz E, et al. Clonal hematopoiesis and risk ath- DM, et al. Identification and characterization of ovarian erosclerotic cardiovascular disease. N Engl J Med. cancer-initiating cells from primary human tumors. Can- 2017;377:111-21. cer Res. 2008;68:4311-20. 53. Coombs CC, Zehir A, Devlin SM, Kishtagari A, Syed A, 35. Eramo A, Lotti F, Sette G, Pilozzi E, Biffoni M, Di Vir- Jonsson P, et al. Therapy-related clonal hematopoiesis in gilio A, et al. Identification and expansion of the tumori- patients with non-hematologic cancers is common and genic lung cancer stem cell population. Cell Death Differ. associated with adverse clinical outcomes. Cell Stem 2008;15:504-14. Cell. 2017;21:374-82. 36. Boiko AD, Razorenova OV, van de Rijn M, Swetter SM, 54. Gibson CJ, Lindsley RC, Tchekmedyian V, Mar BG, Shi Johnson DL, Ly DP, et al. Human melanoma-initiating G, Jaiswal S, et al. Clonal hematopoiesis associated with cells express neural crest nerve growth factor receptor adverse outcomes after autologous stem-cell transplanta- CD271. Nature. 2010;466:133-7. tion for lymphoma. J Clin Oncol. 2017;35:1598-605. 37. Civenni G, Walter A, Kobert N, Mitic-Probst D, Zipser 55. Gillis NK, Ball M, Zhang Q, Ma Z, Zhao YL, Yoder SJ, M, Belloni B, et al. Human CD271-positive melanoma et al. Clonal hematopoiesis and therapy-related myeloid stem cells associated with metastasis establish tumor het- malignancies in elderly patients: a proof-of-concept, case- erogeneity and long-term growth. Cancer Res. 2011;71: control study. Lancet Oncol. 2017;18;112-21. 3098-109. 56. Takahahsi K, Wang F, Kantarjian HP, Doss D, Khanna K, 38. Quintana E, Shackleton M, Sabel MS, Fullen DR, John- Thompson E, et al. Preleukemic clonal haemopoiesis and son TM, Morrison SJ. Efficient tumour formation by risk of therapy-related myeloid neoplasms: a case control single human melanoma cells. Nature. 2008;456:593-8 study. Lancet Oncol. 2017;18:100-11. 39. Quintana E, Shackjkleton M, Foster HR, Fullen DR, 57. Abelson S, Collard G, Ng SWK, Weissbrod O, Men- Sabel MS, Johnson TM, et al. Phenotypic heterogeneity delon-Cohen N, Niemeyer E, et al. Prediction of acute among tumorigenic melanoma cells from patients that is myeloid leukemia risk in healthy individuals. Nature. reversible and not hierarchically organized. Cancer Cell. 2018;559:400-4. 2010;18:510-23. 58. Desai P, Mencia-Trionchant N, Savenkov O, Simon 40. Boyle SE, Fedele CG, Corbin V, Wybacz E, Szeto P, MS, Cheang G, Lee S, et al. Somatic mutations precede Levin J, et al. CD271 expression on patient melanoma acute myeloid leukemia years before diagnosis. Nat Med. cells is unstable and unlinked to tumorigenicity. Cancer 2018;24:1015-23. Res. 2016;76:3965-77. 59. Murphy T, Zou J, Daher-Reyes G, Arruda A, Gupta V, Mc- 41. Prayer BC, Xie Q, Bao S, Rich JN. Cancer stem cells: Namara CJ, et al. Impact of preleukemic mutations and the architects of the tumor ecosystem. Cell Stem Cell. their persistence on hematologic recovery after induction 2019;24:41-53. chemotherapy for AML. Blood Adv. 2019;3:2307-11. 42. Saygim C, Matei D, Majeti R, Reizes O, Lathia JD. Tar- 60. Arends CM, Galan-Sousa J, Hoyer K, Chan W, Jager M, geting cancer stemness in the clinic: from hype to hope. Yoshida K, et al. Hematopoietic lineage distribution and Cell Stem Cell. 2019;24:25-40. evolutionary dynamics of clonal hematopoiesis. Leuke- 43. Lathia JD, Liu H, Matei D. The clinical impact of cancer mia. 2018;32:1908-19. stem cells. The Oncologist. 2020:25:123-31. 61. Lee JH, Lee JE, Kahng JY, Kim SH, Park JS, Yoon SJ, 44. Jan M, Snyder TM, Corces-Zimmerman MR, Vyas P, et al. Human glioblastoma arises from subventricu- Weissman IL, Quake SR, et al. Clonal evolution of pre- lar zone cells with low-level driver mutations. Nature leukemic hematopoietic stem cells precedes human acute 2018;560:243-7. myeloid leukemia. Sci Transl. 2012;4:149ra118. 62. Jonson AS, Kunala S, Price GJ, Restifo RJ, Spinelli HM, 45. Corces-Zimmerman MR, Hong WJ, Weissman IL, Persing JA, et al. Frequent clones of p53-mutated kerati- Medeiros BC, Majeti R. Preleukemic mutations in hu- nocytes in normal human skin. Proc Natl Acad Sci USA. 348 Ugo Testa, Elvira Pelosi and Germana Castelli

1996;93:14025-9. 80. Hoadley KA, Yau C, Hinoue T, Wolf DM, Lazar AJ, Drill 63. Martincorena I, Roshan A, Gerstung M, Ellis P, VanLoo E, et al. Cell-of-origin patterns dominate the molecular P, McLaren S, et al. High burden and pervasive positive classification of 10,000 tumors from 33 types of cancer. selection of somatic mutations in normal human skin. Cell. 2018;173:291-304. Science. 2015;348:880-6. 81. Malta TM, Sokolov A, Gentles AJ, Burzykowski T, Pois- 64. Martincorena I, Fowler JC, Wabik A, Lawson A, Abascal son L, Weinstein JN, et al. Machine learning identifies F, Hall MW, et al. Somatic mutant clones colonize the stemness features associated with oncogenic dedifferen- human esophagus with age. Science. 2018;362:911-7. tiation. Cell. 2018;1763:338-54. 65. Yokoyama A, Kokiuchi N, Yoshizato T, Nannya Y, Su- 82. Smith BA, Balamis NG, Nanjundiath A, Shen KM, Tsai reviews

zuki H, Takeuchi Y, et al. Age-related remodeling of oe- BL, Zhang Q, et al. A human adult stem cell signature sophageal epithelia by mutated cancer drivers. Nature. marks aggressive variants across epithelial cancers. Cell

and 2019;565:312-8. Rep. 2018;24:3353-66. 66. Blokzijl F, de Ligt J, Jager M, Sasselli V, Roerink S, Sasaki 83. Miranda A, Hamilton PT, Zhang AW, Pattnaik S, Becht N, et al. Tissue-specific mutation accumulation in human E, Mezheryevski A, et al. Cancer Stemness, intratumoral adult stem cells during life. Nature. 2016;538:260-4. heterogeneity and immune response across cancers. Proc 67. Yizhak K, Aguet F, Kim J, Kubler K, Grimsby J, et al. Natl Acad Sci USA. 2019;116:9020-9.

articles RNA sequence analysis reveals macroscopic somat- 84. Dalerba P, Sahoo D, Paik S, Guo X, Yothers S, Song N, et ic clonal expansion across normal tissues. Science. al. CDX2 as a prognostic biomarker in stage II and stage 2019;364:eeW0726. III colon cancer. N Engl J Med. 2016;374:211-22. 68. Lan X, Jorg DJ, Cavalli FMG, Richards LM, Nguyen 85. Hansen TF, Kjaewr-Frifeldt S, Eriksen AC, Lindebjerg LV, Vannez RJ, et al. Fate mapping of human glioblas- J, Jansen LH, Sorensen FB, et al. Prognostic impact of riginal toma reveals an invariant stem cell hierarchy. Nature. CDX2 in stage II colon cancer: results from two nation- O 2017;549:227-32. wide cohorts. Br J Cancer. 2018;119:1367-73. 69. Park NI, Guilhamon OP, Desai K, McAdam RE, Langille 86. Balbinot C, Armant O, Elarouci N, Marisa L, Martin E, E, O’Connor M, et al. ASCL1 reorganizes chromatin to De Clara E, et al. The Cdx2 homeobox gene suppresses direct neuronal fate and suppress tumorigenicity of glio- intestinal tumorigenesis through non-cell-autonomous blastoma stem cells. Cell Stem Cell. 2017;21:411-21. mechanisms. J Exp Med. 2018;215:911-26. 70. Shimokawa M, Ohta Y, Nishikori S, Metano M, Takano 87. Sekulic A, Migden MR, Oro AE, Dirix L, Lewis KD, A, Fujii M, et al. Visualization and targeting of LGR5+ Hainswoth KD, et al. Efficacy and safety of vismo- human colon cancer stem cells. Nature. 2017;545:187- degib in advancer basal-cell carcinoma. N Engl J Med. 92. 2012;366:2171-9. 71. De Sousa-Melo F, Kurtova AV, Harnoss JM, Kjuvin N, 88. Sekulic A, Migden MR, Lewis K, Hainwoth JD, Solomon Hoeck JD, Hung J, et al. A distinct role for Lgr5+ stem JA, Yoo S, et al. Pivotal ERIVANCE basal cell carcinoma cells in primary and metastatic colon cancer. Nature. (BCC) study: 12-month update of efficacy and safety 2017;543:676-80. of vismodegib in advancer BCC. J Am Acad Dermatol. 72. Shlush LI, Mitchell A, Heisler L, Abelson S, Ng SWK, 2015;72:1021-6. Trotman-Grant A, et al. Tracing the origins of re- 89. Sekulic A, Migden MR, Basset-Seguin N, Garbe C, Ge- lapse in acute myeloid leukemia to stem cells. Nature. sieritch A, Lao CD, et al. Long-term safety and efficacy 2017;547:104-8. with vismodegib in patients with advanced basal cell car- 73. De Angelis ML, Francescangeli F, La Torre F, Zeuner A. cinoma: final update of the pivotal ERIVANCE BCC Stem cell plasticity and dormancy in the development of study. BMC Cancer. 2017;17:322. cancer therapy resistance. Front Oncol. 2019;9:626. 90. Bihs B, Dijkgraaf GJP, Piskol R, Alicke B, Boumahdi 74. Li L, Li JC, Yang H, Zhang X, Liu LL, Li Y, et al. Ex- S, Peale F, et al. A cell identity switch allows residual pansion of cancer stem cell pool initiates lung cancer re- BCC to survive Hedgehog pathway inhibition. Nature. currence before angiogenesis. Proc Natl Acad Sci USA. 2018;562:429-33. 2018;115:E8948-E8957. 91. Sanchez-Danes A, Larsimont JC, Liagre M, Munoz- 75. Merlos-Suarez A, Barriga FM, Jung P, Iglesias M, Cés- Conselo E, Laponge G, Brisebarre A, et al. A slow-cycling pedes MV, Rossell D, et al. The intestinal stem cell sig- Lgr5 tumor population mediated basal cell carcinoma re- nature identifies colorectal cancer stem cells and predicts lapse after therapy. Nature. 2018;562:434-8. disease relapse. Cell Stem Cell. 2011;8:511-24. 92. Kimn EJ, Sahai V, Abel EV, Griffith KA, Greenson JK, 76. Roy S, Roy S, Kar M, Padhi S, Saha A, Anuja K, et al. Takebe N, et al. Pilot clinical trial of hedgehog pathway Role of p38 MAPK in disease relapse and therapeutic inhibitor GDC-0449 (vismodegib) in combination with resistance by maintenance of cancer stem cells in head gemcitabine in patients with metastatic pancreatic ad- and neck squamous cell carcinoma. J Oral Pathol Med. enocarcinoma. Clin Cancer Res. 2014;209:5937-45. 2018;47:492-501. 93. DiNardo CD, Pratz K, Pullarkat V, Jonas BA, Arellano M, 77. Sun L, Zhang L, Chen J, Li C, Sun H, Wang J, et al. Becker PS, et al. Venetoclax combined with decitabine or Activation of tyrosine metabolism in CD13+ cancer stem azacitidine in treatment-naive, elderly patients with acute cells drives relapse in hepatocellular carcinoma. Cancer myeloid leukemia. Blood. 2019;133:7-17. Res Treat. 2020;52:604-21. 94. Di Nardo CD, Stein EM, de Botton S, Roboz GJ, Alt- 78. NG SW, Mitchell A, Kennedy JA, Chen WC, McLeod man JK, Mims AS, et al. Durable remissions with Ivo- J, Ibrhimova N, et al. a 17-gene stemness score for sidenib in IDH1-mutated relapsed or refractory AML. N rapid determination of risk in acute leukemia. Nature. Engl J Med. 2018;378:2386-98. 2017;540:433-7. 95. Stein EM, Di Nardo CD,Pollyea DA, Fathiu AT, Ro- 79. Hoadley KA, Yau C, Wolf DM, Cherniak D, Tambrero boz GJ, Altman JK, et al. Enasidenib in mutant IDH2 D, Ng S, et al. Cancer Genome Atlas Research Network relapsed or refractory acute myeloid leukemia. Blood. multiplatform analysis of 12 cancer types reveals molecu- 2017;130:722-31. lar classification within and across tissue of origin. Cell. 96. Amatangelo MD, Quek L, Shih A, Stein EM, Roshal M, 2014;158:929-44. David MD, et al. Enasidenib induces acute myeloid leu- 349 Cancer stem cell targeted therapies

kemia cell differentiation to promote clinical response. cytidine in AML and MDS patients: initial phase 1b re- Blood. 2017;130:732-41. sults. J Clin Oncol 2019;37(15 suppl):7009. 97. Fathi AT, Di Nardo CD, Kline I, Kenvin L, Gupta I, Attar 113. Testa U, Pelosi E, Frankel A. CD123 is a membrane bio- EC, et al. Differentiation syndrome associated with Ena- marker and a therapeutic target in hematologic malignan- sidenib, a selective inhibitor of mutant isocitrate dehy- cies. Biomarker Res. 2014;2:4. drogenase 2: analysis of a phase 1/2 study. JAMA Oncol. 114. Pemmaraju N, Lane AA, Sweet KL, Stein AS, Vasu S, 2018;4:1106-10. Blum W, et al. Tagraxofusp in blastic plasmacytoid den- 98. Quek L, David MD, Kennedy A, Metzner M, Amatange- dritic-cell neoplasm. N Engl J Med. 2019;380:1628-37. lo M, Shih A, et al. Clonal heterogeneity of acute myeloid 115. Mardiros A, Dos Santos C, McDonald T, Brown CE, reviews

leukemia treated with the IDH2 inhibitor Enasidenib. Wangh X, Budde LE, et al. T cells expressing CD123- Nat Med. 2018;24:1167-77. specific cytolytic effector functions and anti-tumor ef-

99. Intlekofer AM, Shih AH, Wang B, Nazir A, Rusterburg fects against human acute myeloid leukemia. Blood. and AS, Albanese SK, et al. Acquired resistance to IDH in- 2013;122:3138-48. hibition through trans or cis dimer-interface mutations. 116. Budde LE, Song JY, Kim Y, Blanchard S, Wagner J, Stein Nature. 2018;559:125-9. AS, et al. Remissions of acute myeloid leukemia and blas- 100. Stein EM, Di Nardo CD, Fathi AT, Pollyea DA, Stone tic plasmacytoid dendritic cell neoplasm following treat-

RM, Altman JK, et al. Molecular remission and response ment with CD123-specific CART cells: a first-in-human articles patterns in patients with mutant-IDH2 acute myeloid clinical trial. Blood. 2017;130:abst. 811. leukemia treated with enasidenib. Blood. 2019;133:676- 117. Budde LE, Schuster SJ, Del Real M. CD123CAR dis- 87. plays clinical activity in refractory/relapsed acute myeloid 101. Pollyea DA, Tallman MS, de Botton S, Kantarjian HM, leukemia (AML) and blastic plasmocytoid dendritic cell Collins R, Stein AS, et al. Enasidenib, an inhibitor of mu- neoplasm (BPDCN): safety and efficacy results from a riginal tant IDH2 proteins, induces durable remissions in older phase 1 study. Miami, FL, USA: AACR Tumor Immunol- O patients with newly diagnosed acute myeloid leukemia. ogy and Immunotherapy; 2018. Leukemia 2019;33:2575-84. 118. Shibata M, Hoque MO. Targeting cancer stem cells: a 102. Kim D, Park CY, Medeiros BC, Weissman IL. CD19- strategy for effective eradication of cancer. Cancers. CD45 low/- CD38high/CD138+ plasma cells enrich 2019;11:732. for human tumorigenic myeloma cells. Leukemia. 119. Huyhn J, Chaud A, Gough D, Ernst M. Therapeutically 2012;26:2530-7. exploiting STAT3 activity in cancer using tissue repair as 103. Gao M, Bai H, Jethava Y, Wu Y, Zhu Y, Yang Y. Identi- a road map. Nat Rev Cancer. 2019;19:82-96. fication and characterization of tumor-initiating cells in 120. Li Y, Rogoff HA, Keates S, Gao Y, Murikipudi S, Mikule multiple myeloma. J Natl Cancer Inst. 2020;112:507-15. K, et al. Suppression of cancer relapse and metastasis by 104. Facon T, Kumar S, Piesner T, Orlowski RZ, Moreau P, inhibiting cancer stemness. Proc Natl Acad Sci USA. Bahlis S, et al. Daratumumab plus lenalidomide and 2015;112:1839-44. dexamethasone for untreated myeloma. N Engl J Med. 121. Jonker DJ, Nott L, Yoshino T, Gill S, Shapiro J, Ohtsu A, 2019;380:2104-2115. et al. Napabucasin versus placebo in refractory advanced 105. Mateos MV, Dimopoulos MA, Cavo M, Suzuki K, colorectal cancer: a randomized phase 3 trial. Lancet Jakubowiak A, Knop S, et al. Daratumumab plus bortezo- Gastroenterol Hepatol. 2018;3:263-70. mib, melphalan, and prednisone for untreated myeloma. 122. Grothey A, Shah MA, Yoshino T, Cutsem EV, Taieb J, Xu N Engl J Med. 2018;378:518-28. R, et al. CanStem303C trial: a phase III study of Napa- 106. Moreau P, Atai M, Hulin C, Arnulf B, Belhadj K, Ben- bucasin (BBI-608) in combination with 5-fluorouracil boubker L, et al. Bortezomib, thalidomide, and dexa- (5-FU), leucovorin, irinotecan (FOLFIRI) in adult pa- methasone with or without daratumumab before and tients with previously treated metastatic colorectal cancer after autologous stem cell transplantation for newly diag- (mCRC). J Clin Oncol. 2017;35:TPS3619. nosed multiple myeloma (CASSIOPEIA): a randomized, 123. Bekaii-Saab TS, Starodub A, El-Rayes CF, Shahda S, open-label, phase 3 study. Lancet. 2019;394:29-38. O’Neil BH, Noonan AM, et al. Phase 1b/2 trial of cancer 107. Matlung HL. The CD47- SIRPa signaling axis is an stemness inhibitor napabucasin (NAPA) + nab-paclitaxel innate immune checkpoint in cancer. Immunol Rev. 8nPTX) and gemcitabine (Gem) in metastatic pancreatic 2017;276:145-64. adenocarcinoma (mPDAC). J Clin Oncol. 2018;36:4110. 108. Jaiswal S, Jamieson C, Pang WW, Park CY, Chao MP, 124. Sonbol MB, Ahn DH, Bekaii-Saab T. Therapeutic target- Majeti R, et al. CD47 is up-regulated on circulating he- ing strategies of cancer stem cells in gastrointestinal ma- matopoietic stem cells and leukemia cells to avoid phago- lignancies. Biomedicines. 2019;7:17. cytosis. Cell. 2009;138:271-85. 125. Zhan T, Ambrosi G, Wandmacher AM, Rauscher B, 109. Yan B, Chen Q, Shimada K, Tang M, Li H, Gurumurthy Betge J, Rindtorff N, et al. MEK inhibitors activate Wnt A, et al. Histone deacetylase inhibitor targets CD123/ signaling and induce stem cell plasticity in colorectal can- CD47-positive cells and reverse chemoresistance in acute cer. Nat Commun. 2019;10:2197. myeloid leukemia. Leukemia. 2019;33:931-44. 126. Hermann POC, Huber SL, Herrler T, Aicher A, Eliwart 110. Chao MP, Alizadeh AA, Tang C, Myklebust JH, Var- JW, Guba M, et al. Distinct populations of cancer stem ghese B, Gill S, et al. Anti-CD47 antibody synergizes cells determine tumor growth and metastatic activity in with rituximab to promote phagocytosis and eradicate human pancreatic cancer. Cell Stem Cell. 2007;1:313-23. non-Hodgkin lymphoma. Cell. 2010;142:699-713. 127. Li C, Wu JJ, Hynes M, Dosch J, Sarkar B, Welling TH, 111. Advani R, Flinn I, Popplewell L, Forero A, Bartlett NL, Pasca di Magliano M, et al. C-Met is a marker of pancre- Ghosh N, et al. CD47 blockade by Hu5F9-G4 and atic cancer stem cells and therapeutic target. Gastroen- rituximab in non-Hodgkin’s lymphoma. N Engl J Med. terology. 2011;141:2218-27. 2018;379:1711-21. 128. Kawamoto M, Ishiwata T, Cho K, Uchida E, Korc M, 112. Sallman D, Donnellan WB, Asch AS, Lee DJ, Malki MA, Naito Z, et al. Nsetin expression correlates with nerve Marcucci G, et al. The first-in-class anti-CD47 antibody and retroperitoneal tissue invasion in pancreatic cancer. Hu5F9-G4 is active and well tolerated alone or with aza- Hum Pathol. 2009;40:189-98. 350 Ugo Testa, Elvira Pelosi and Germana Castelli

129. Rasheed ZA, Yang J, Wang Q, Kowalski J, Freed I, Murter 135. Cortina C, Turon G, Stork D, Hernando-Momblona C, et al. Prognostic significance of tumorigenic cells with X, Sevillano M, Aguilera M, et al. A genome editing mesenchymal features in pancreatic adenocarcinoma. J approach to study cancer stem cells in human tumors. Natl Cancer Inst. 2010;102:340-51. EMBO Mol Med. 2017;9:869-79. 130. Fox RG, Lytle NK, Jaquish DV, Park FD, Ito T, Bajaj 136. De Sousa E Melo E, de Sauvage FJ. Cellular plasticity J, et al. Image-based detection and targeting of thera- in intestinal homeostasis and disease. Cell Stem Cell. py resistance in pancreatic adenocarcinoma. Nature. 2019;24:54-64. 2016;534:407-11. 137. Fumagalli A, Oost KC, Kester L, Morgner J, Bornes L, 131. Lytle NK, Ferguson LP, Rajbhandari N, Gilroy K, Fox Bruens L, et al. Plasticity of LGR5-negative cancer cells reviews

RG, Deshpande A, et al. A multiscale map of the stem drives metastasis in colorectal cancer. Cell Stem Cell. cell state in pancreatic adenocarcinoma. Cell. 2019;177: 2020;26:569-78.

and 572-86. 138. Ziskin JL, Dunlap D, Yaylaoglu M, Fodor IK, Forrest 132. Tsai KK, Chan TS, Skaked Y. Next viable routes to tar- WR, Patel R, et al. In situ validation of an intestinal stem geting pancreatic cancer stemness: learning from clinical cell signature in colorectal cancer. Gut. 2013;62:1012-23. seatbacks. J Clin Med. 2019;8:702. 139. Ihemelandu C, Naeem A, Paradiso E, Berry D, Chal- 133. Huang EH, Hynes MJ, Zhang T, Ginestier C, Dontu dekas K, Harris BT, et al. Clinicopathologic and prog-

articles G, Appelaman H, et al. Aldehyde dehydrogenase 1 is a nostic significance of LGR5, a cancer stem cell marker marker for normal and malignant human colonic stem in patients with colorectal cancer. Colorect Cancer. cells (SC) and tracks SC overpopulation during colon tu- 2019;8:CRC11. morigenesis. Cancer Res. 2009;69:3382-9. 140. Morral C, Stanisavljevic J, Hernando-Momblona X, 134. De Lau W, Peng WC, Gros P, Clevers H. The R-spondin/ Mereu E, Alvarez-Varela A, Cortina C, et al. Zonation of riginal LGR5/RNF43 module module: regulator of Wnt signal ribosomal DNA transcription defines a stem cell hierar- O strength. Genes Dev. 2014;28:305-16. chy in colorectal cancer. Cell Stem Cell. 2020;26:845-61. and RosagemmaCiliberti Genoa, Italy. E-mail:[email protected]. Address forcorrespondence: AlessandroBonsignore, DipartimentodiScienzedellaSalute, UniversitàdegliStudidiGenova, Via DeToni 12,16132 to fourpossiblecategories: and urethralabnormalities. concealed penis (concealed penis), micropenis, obesity, ture, penoscrotalfusion,sunkenpenis(buriedpenis), als withpeniledefects,suchascongenitalcurva- However, theprocedureiscontraindicatedin individu- order toalleviatethepainandphysiologicalstress[4]. elasticity andthelengththatitreaches. foreskin to beremovedis quantified onthebasis ofits skin. In the first phase of the procedure, the amount of prepuce sothattheglandisalwaysfreefromitshoodof be about37.7%,whileinItalyitis2.6%[3]. central Asia.Worldwide, theprevalenceisestimatedto prevalence isalmostuniversalintheMiddleEastand ed StatesandpartsofSouth-eastAsiaAfrica.Its it isalmostuniversal)forreligiousreasons,intheUnit- cially commonintheMuslimworldandIsrael(where men has been circumcised [2]. The procedure is espe- world [1]. Indeed, it is estimated that one out of three dure andoneofthemostwidespreadpracticesin INTRODUCTION measures inMC. mentation oftargetedcampaignsmayraiseawarenesstheimportancepropersafety should performMConminorsfreeofchargeor, atleast,forareducedfee.Theimple- there isapressingneedforthoroughdiscussionofwhethertheNationalHealthService Conclusion. public healthsetting. available inthevariousRegionsandtorisksengenderedbyexcludingMCfrom Italian healthcaresystemwithparticularregardbothtotheheterogeneityofservices admissibility ofMCwithinacomparativeperspective,weexaminethefeatures Results anddiscussion.Afterillustratingthehistoricalethicaloutlinesofmoral unqualified personsinunhygienicconditions. of malecircumcision(MC)proceduresperformedforculturalandreligiousreasonsby Introduction andobjectives.InItaly, fourminorshavediedinthelastyearasaresult Abstract 2 Francesco Ventura and responsibility Male circumcision:ritual,science DOI: 10.4415/ANN_20_03_13 Ann IstSuperSanità2020|Vol. 56,No.3:351-358 1 cum Dipartimento diBiotecnologieeScienzedellaVita, UniversitàdegliStudidell’Insubria,Varese, Italy Dipartimento diScienzedellaSalute,UniversitàdegliStudiGenova,Genoa,Italy 1) therapeuticcircumcision(resolution ofphimosis, Generally speaking,MCcan beclassifiedaccording Topical orlocalanaestheticsarenormallyutilisedin Male circumcision(fromtheLatincircumcidere: Male circumcision(MC)isbothanancientproce- = around and In ordertoadequatelysafeguardpublichealth,particularlythatofminors, caedere = 1 , FiorellaCaputo cut) consists of excising the 1 1 , MartaLicata cir - tion, reasonsunexplainedbytherequester). vent urinarytractinfectionsininfancy); tis, etc.); chronic irritationoftheglans,refractorybalanoposthi- ing veryseriousinjuries. authorised practiceofthemedical professionandcaus- The manwasarrestedforinvoluntary manslaughter, un- of Libyanoriginknowntocarry outsuchinterventions. gery. Therefore,thewomanhadturnedtoaUScitizen her paediatrician,whohadrefusedtoperformthesur twins tocircumcision,hadrequested information from 35-year-old Nigerianmother, before subjectingher the PoliclinicoGemelliinRomeandwassaved.The cedure, washospitalizedintheintensivecareunitat tions. Histwinbrother, whounderwentthesame pro- circumcision athomeanddiedofseriouscomplica- Rome, a2-year-old childofNigerianoriginunderwent most recent,inGenoa. in Rome,two the Province of Reggio Emilia and, the hygienic conditions have recentlybeenrecorded: one cumcision carried out by unqualified operators in un- In Italy, four deathsconnectedwithritualmalecir 4) circumcisionforother, personalreasons(imita- 3) ritualcircumcision(typicalofJudaismandIslam); 2) prophylacticcircumcision(e.g.innewbornstopre- In March2019,a5-month-old childofGhanaian In December2018,intheMonterotondoareaof 2 , AlessandroBonsignore • • • • • Key words ethics parental decision-making adverse events infants circumcision male

1

- - 351 Original articles and reviews 352 Francesco Ventura, Fiorella Caputo, Marta Licata et al

origin, who had been admitted to Santa Maria Nuova problems, including masturbation [7]. The cultural and Hospital in Reggio Emilia, died as a result of very se- social climate in which the procedure arose and spread rious complications following ritual circumcision. The in the second half of the 19th century, in English-speak- procedure had been performed by a practitioner of Af- ing countries, is described by the medical historian Ed- rican origin who used to carry out this type of interven- ward Wallerstein, who asserts that, “within the miasma tion in African communities in Emilia. of myth and ignorance, there emerged the theory that A few months later (November 2019), another masturbation caused many and various disorders. It death involved a 5-month-old child of Ghanaian ori- therefore seemed logical to some doctors to perform reviews

gin, who was resident in Scandiano (Reggio Emilia). genital surgery on both sexes, in order to prevent mas- The child was urgently hospitalized for cardiac arrest at turbation; the main technique implemented in males

and Sant’Orsola Hospital in Bologna, where he arrived in a was circumcision. This was especially true in English-

desperate condition following a circumcision procedure speaking countries, as it was in line with the mid-Vic- performed at home. The ongoing investigations focus torian attitude to sex, which was considered sinful and on a “holy man”, who is a compatriot of the parents and debilitating” [6]. As medical knowledge increased, how- is known within the African community to be willing to ever, the rationale behind preventing masturbation was articles carry out this kind of intervention. The man, who is a questioned and, subsequently, abandoned. resident of Modena, is currently under investigation for In 2007, on the basis of scientific evidence, the inter- involuntary manslaughter. national community recommended that “male circum- The most recent case concerns a newborn of Nige- cision now be recognized as an additional important in- riginal rian origin who died in Genoa (April 2019) following tervention to reduce the risk of heterosexually acquired O circumcision performed at home. Following the child’s HIV infection in men” [8, 9]. Moreover, in 2012, the death, his mother and grandmother, aged 25 and 50 American Academy of Pediatrics reported that MC was years, both of Nigerian origin, were arrested on charges able to significantly reduce the risk of contracting uri- of involuntary manslaughter. A 34-year-old Nigerian nary tract infections, some other sexually transmitted man, known within the community as a subject capable infections and carcinoma of the penis [4]. The report of carrying out circumcision, is accused of being the recommended that, at the beginning of pregnancy, par- material perpetrator of the intervention. He was arrest- ents should regularly be informed of the benefits and of ed by the judicial police while trying to flee the country, the low risk of MC, and that payment for MC by third having learned of the child’s death. parties was justified [4, 10]. These results were con- In all four cases, death was due to severe haemor- firmed by a recent systematic review of the literature, rhage after the procedure. In all cases, circumcision was which revealed that performing MC during early infan- carried out by foreign nationals who did not possess the cy was associated with a lower incidence of urinary tract necessary skills. infections, yielded benefits that were up to 200 times These events have prompted debate, not least within greater than the procedural risks involved, and was also the medical community, as to whether MC should be cost-saving [11]. The factors contributing to complica- made available by the National Health Service (NHS). tions clearly include the training and experience of the operator, the characteristics of the instruments used HISTORY AND ETHICAL CONSIDERATIONS and the sterility of the environment in which the proce- Therapeutic MC was regarded as a common medi- dure is carried out [12]. cal procedure up to the late Victorian period. In 1870, Nevertheless, the scientific validity of MC for pro- the orthopaedic surgeon Lewis Albert Sayre, one of the phylactic purposes, and also its cost/benefit ratio, con- founders of the American Medical Association, started tinue to be the subject of controversy [13]. practising it in order to treat serious motor difficulties In Italy, the CNB (National Bioethics Committee), [5]. while acknowledging that prophylactic MC is a criti- From the ethical standpoint, circumcision for thera- cal issue, does not consider this procedure per se to be peutic reasons does not raise issues that are any differ- unjustified or unacceptable from the ethical standpoint ent from those connected with any other therapeutic [14]. Thus, according to the CNB, in the absence of intervention. It is therefore subject to the ethical prin- compelling reasons to refrain from performing MC, ciples of both “non-maleficity” (which mandates re- the procedure may be deemed admissible, provided of spect for good medical practice) and autonomy, which course that it is carried out in conformity with the crite- is expressed through respect for both the free and re- ria of good medical practice and, in the individual case, sponsible self-determination of the individual and for supported by a specific scientific judgement. individual privacy. Moreover, when it is performed Ritual circumcision is certainly the procedure most for therapeutic reasons, circumcision, like any other hotly debated from the ethical point of view. healthcare procedure, must necessarily be carried out There is no consensus among anthropologists as to by a qualified operator. the origin of the practice of MC. Ritual male circumci- MC for prophylactic purposes is more debatable. sion is certainly a very widespread custom and is prac- This practice arose in the 19th century, when the ae- tised by many different peoples, from the ancient East- tiology of most diseases was still unknown [6]. The as- ern Mediterranean to Africa and pre-colonial Australia, sumption that a tight foreskin would inflame the nerves though not by populations of the Indo-Germanic lan- and cause systemic disorders prompted the adoption of guage group, nor by those of non-Semitic upper Asia. prophylactic MC for the prevention of a wide range of According to some historians, the procedure was, at 353 Male circumcision

least in some periods, carried out in order to humiliate recovery) and involves the use of ritual objects (a knife enemies [15, 16]. However, given the geographic exten- with a particular blade, a protective shield, a container sion of the practice, it is not possible to individuate a for the prepuce); these prescriptions constitute a pre- single satisfactory explanation. cise personal obligation on the parents, or whoever The historian, anatomist and Egyptologist Sir Grafton stands in for them, and their fulfilment is regarded as Elliot Smith claimed that MC had been practised for an act of devotion. more than 15 000 years [17]. Moreover, clear evidence From the medical standpoint, it is deemed preferable of the existence of this procedure dates back to more to perform circumcision in the neonatal period rather reviews than 4300 years ago. Indeed, in the great Egyptian ne- than at an older age. Indeed, in the neonatal period, cropolis of Saqqara, not far from Cairo, a decoration the procedure is not only operationally simpler, but also

on a wall of the tomb of Ankhamahor (an important displays a low rate of complications; this is due to the and

Egyptian functionary and high priest who lived during healing capacity of the newborn and to the fact that the VI Dynasty) depicts male circumcision, performed suturing is not generally necessary in such subjects [24]. presumably as a ritual before admission to the priest- Consequently, beyond religious motivations, perform- hood [18]. Although Egyptologists have not fully ascer- ing circumcision in the neonatal period is also safer and articles tained what the purpose of circumcision was in ancient less costly. Egypt, it is thought that the practice served to certify Despite its great diffusion, ritual circumcision con- the passage to adulthood among members of the higher tinues to be an extremely controversial practice in view social classes, and that it was performed during a pub- of the various rights and values involved. In particular, lic ceremony [17]. Further testimony to the practice of recognition of the cultural and/or religious rights of riginal MC among the ancient Egyptians can also be seen in minority communities raises complex issues for liberal O the so-called Ebers Papyrus, purchased in 1873-1874 democracies and their constitutions. Indeed, from the at Thebes by the German archaeologist Georg Moritz ethical point of view, legislation is faced with the dif- Ebers [19]. ficulty of reconciling the need to safeguard minorities In sum, hygiene, preparation for sexual life, a rite of with the protection of the rights of the individuals who passage and initiation to adulthood, tribal identity and live within these very minorities. A specific aspect of adherence to a religious belief are the reasons most fre- this dilemma is seen in the relationship between minors quently cited to explain the meaning of ritual circumci- and their families who belong to a cultural and religious sion. community that is “different” from that of the majority. References to circumcision can also be found both Thus, the perspective that emerges is one within in the Hebrew Bible (Genesis: 17, 9-14; Leviticus: 12, which the debate is not limited to merely legal aspects, 3) and in the Christian Bible (N.T.: Acts of the Apos- however important these may be, nor to the apodictic tles, 15), which considers the practice unnecessary. assertion of an ethical point of view; rather, questions By contrast, Coptic Christians perform circumcision of tolerance need to be examined in their political-pru- as a rite of passage [20]. In South Africa, circumci- dential dimension. sion is prescribed by some Christian groups, while it The safeguard of minors, who are obviously unable is frowned upon by others. The Ethiopian Orthodox provide valid consent, constitutes a particularly criti- Church requires MC, and its prevalence is almost uni- cal aspect [25, 26]. Indeed, the decision to circumcise versal among Orthodox men in Ethiopia. For Muslims, the minor is usually taken only by parents or guardians, circumcision is considered essential. Although it is not without the consent of the minor. explicitly mandated in the Koran, it is attributed to the Citing the principles of good medical practice, the Prophet Mohammed; for this reason, the practice has British Medical Association (BMA) asserts that cir- taken on the nature of Sunnah, or the tradition of the cumcision for purely prophylactic or ritual purposes Prophet. It is also recognised in the Hadith (sayings is not automatically justified by parental consent and and deeds of the Prophet). urges doctors to inform parents of the issues involved in The age at which ritual circumcision is performed, its an invasive medical operation. The Association speci- modalities and settings, the fate of the excised prepuce, fies that doctors must act within the confines of the law and the figure and function of the circumciser are all and of their own conscience and weigh the benefits and extremely variable [21, 22]. As pointed out by Abdul risks (also psychological) of circumcision in each spe- Wahid Anwer, et al., although MC is a religious rec- cific case. ommendation, cultural and social norms (such as the On the basis of the laws in force in the United States actual possibility to organise a sumptuous feast to cele- and of international declarations on human rights, cir- brate the circumcision of a son) are major determinants cumcision violates a minor’s absolute rights to the pro- of when circumcision is performed. Thus, belonging to tection of his physical integrity, autonomy and freedom certain ethnic groups is a risk factor for delayed circum- to choose his own religion. This means that the doctor cision [23]. Muslims celebrate MC either within the is legally obliged to protect children from unnecessary family or as a community event, and the procedure is medical interventions [13]. Thus, in the absence of usually performed some years after the birth of the child medical justification, parental consent to the circumci- (though always before puberty). In the Jewish commu- sion of a newborn would not be legally valid, in that par- nity, by contrast, CM is performed on the eighth day ents may authorise a non-therapeutic procedure only if after the child’s birth (unless adverse clinical conditions it is in the best interest of the child [27]. necessitate postponement of the procedure until after On 28 September 2013, the Swedish Civic Ombud- 354 Francesco Ventura, Fiorella Caputo, Marta Licata et al

sperson for Children and representatives of four leading infant circumcision, if performed after six months of medical associations in Sweden asserted that circumcis- age and by qualified health personnel [31]. ing a child without medical justification and without the The possibility to have one’s child circumcised, child’s consent is contrary to the child’s human rights should this be deemed necessary for the child’s own and to the fundamental principles of medical ethics good, must nevertheless be subordinated to the regu- [28]. lations of medical practice, the administration of effi- Indeed, according to the Convention of New York, cacious analgesic treatment, the provision of complete minors (aged at least 12 years, or even younger if ca- information for the parents, and proper consideration reviews

pable of discerning) have the right to be consulted and for the child’s wishes. Moreover, the law explicitly per- to express their opinion on all procedures that concern mits circumcision to be performed – during the first six

and them. Thus, in the absence of compelling reasons, it months after the birth of the child – by persons who

seems questionable to exclude minors from a choice are not physicians, provided that they possess compe- that may impact significantly on their physical and tence equivalent to that of a physician in performing mental health and which can easily be postponed until this specific procedure. This specification takes into ac- such time as they are able to express an informed opin- count the fact that circumcision for religious reasons is articles

ion regarding a totally personal choice. However, this often carried out by “ministers of the faith”, who occupy growing recognition of the will of the minor may clash a certain position within the religious community. As with the recognition of the (constitutional) right of par- a rule, these individuals possess the necessary skills to ents to religious freedom and to bring up their children perform the operation, as well as being well-versed in riginal in accordance with the precepts of the religious faith to the rites that accompany it. O which they belong. If we are to give adequate consideration to the vari- MALE CIRCUMCISION AND PUBLIC ous interests involved, the increased emphasis on the HEALTH IN ITALY: PROBLEMS rights of minors must not give rise to simplistic and AND OPPORTUNITIES dangerous commonplaces. Although circumcision Unlike the practice of female genital mutilation, impacts on the physical integrity of the minor, it may which is expressly forbidden by Italian law (Law n. 7 nevertheless contribute positively to the development of 09/01/06: “Provisions concerning the prevention and of his personality, constituting an overall benefit in the prohibition of practices of female genital mutilation”) physical, psychological, social and relationship spheres [32], MC is commonly admissible. [29]. Indeed, circumcision is an identifying sign of be- In Italy, the conviction that the Jewish practice of longing to a religious community; for parents, it con- male circumcision conforms to Italian law seems to stitutes a fundamental element of the structure of the find confirmation in some provisions of Law 101/1989 minor’s identity and is able to strengthen the rights of “Norms regulating relations between the State and the the child and his family as an organic unit. Moreover, Union of Italian Jewish Communities” [33]. In a not if properly performed, it does not result in impairment very recent case (9 November, 2007) discussed by the or alteration of the male’s sexual and reproductive Court of Padua, the Court did not contest the lawful- function. ness of ritual male circumcision in itself, in that “ritual This consideration has prompted the CNB to view circumcision may be viewed as being aimed at achiev- the practice within the operational framework of article ing a better state of health, a bodily form corresponding 19 of the Italian Constitution, which grants complete to the idea of physical perfection and psychic satisfac- freedom of cultural and ritual expression at both the tion of the individual person, not least with a view to individual and collective levels [14]. In addition to conforming with an ethnic or cultural identity”. In par- this, and taking into account the elementary nature of ticular, the judge specified that “Although it has been the procedure, some members of the CNB have even emphasised that male circumcision constitutes a viola- claimed that dedicated ministers should be allowed to tion of the psychophysical integrity of a subject who, perform ritual circumcision, as long as they are recog- owing to his young age, is generally unable to express nised as possessing adequate competence. In this re- his consent effectively (it is the parents who decide dur- gard, a recent German experience appears to be sig- ing the exercise of their right/duty to bring up their child nificant. in accordance with the principles of their culture), it In June 2012, the Court of Cologne in Germany seems difficult to contest the notion that this procedure ruled that non-therapeutic MC performed on minors – given that it is free from the negative physical, psycho- constituted an irreversible bodily lesion and violated logical and symbolic connotations that characterise fe- the individual’s right to physical integrity and self-de- male genital mutilation, and probably also owing to the termination, and that the procedure should be delayed influence of Judaism – has long been amply accepted by until the minor was old enough to decide in a free and Western custom and culture” [34]. The Court, however, informed manner. In addition, the Court stated that also ruled that MC was a practice that necessarily had doctors who carry out this surgical operation could be to be performed by medical personnel, in that “it results prosecuted [30]. in an impairment of the physical integrity which can- However, after a lively public debate concerning the not prescind from careful evaluation of the subject who cultural and religious traditions of infant male circumci- undergoes it, on account of the potential negative con- sion, Germany’s parliament introduced a modification sequences that it might have on health, and which must (§1613d) to the Civil Code that explicitly permits male be performed in accordance with good clinical practice 355 Male circumcision

and the subsequent assurance of care”. supported by doctors in Turin, who have stressed the Support for the implementation of MC in the hos- possible risks and complications of an invasive and mu- pital setting stems both from the desire to foster the tilating procedure that is imposed on a minor without integration of ethnic minorities and from the need to any medical justification. ensure that circumcision is carried out as safely as pos- In this sense, it must be stressed the legal position of sible in these communities. At present, however, par- healthcare professionals performing such a procedure: ents wishing to have their children circumcised for re- a delicate matter that should be taken into account es- ligious and/or cultural reasons are faced with different pecially in times when medico legal claims are always reviews responses from the various regional healthcare systems. more frequent, both from a criminal and civil point of The Tuscany Region, for example, included MC in view. An important aspect that could limit the availabil-

the list of essential services offered to all citizens (LEA) ity of medical staff providing this service on a nation- and back in 2002 (Dgr n. 561/2002); in Friuli Venezia Gi- wide basis. ulia, the procedure is available on payment of a charge Therefore, also ritual circumcision has to be preceded equal to the tariff applied to therapeutic circumcision by a careful evaluation of newborn and children’s con- (Dgr n. 600/2010); in Turin, in the Piedmont Region, ditions, balancing them with an identifiable psychic articles ritual MC is available for Muslim children in a day-sur- well-being linked to conforming to an ethnic or cultural gery setting at a reduced charge. identity, and respecting all the rules of good clinical In the Lazio Region, the Umberto I Polyclinic has practice. inaugurated an outpatient clinic where ritual circumci- In recent years, the issue of MC has gained greater sion can be performed on children aged 3 years or more attention in Italy, as a result of the increased number of riginal by authorised, specialised personnel. The service is pro- foreign families who carry out the procedure, usually for O vided on payment and is carried out under a regime of religious and/or cultural reasons. freelance activity; the fee agreed upon with the Rab- According to the data elaborated by the AMSI (As- binic Office of Rome and the Islamic Cultural Centre sociation of Doctors of Foreign Origin in Italy) in col- in Italy is € 200,00. laboration with the Medical Council of Rome (Section In the other regions, however, ritual MC is complete- for relations with the municipalities and foreign affairs; ly unavailable in public healthcare facilities. Section for rehabilitation), 11 000 ritual circumcisions The Union of Italian Jewish Communities, in con- are carried out annually on citizens of foreign origin liv- formity with the dispositions issued by the bodies of ing in Italy; of these, 5000 are performed in Italy and European Judaism, has set up a register of Mohalim 6000 in the various countries of origin. Of the 5000 pro- (circumcisers), whose standard educational curriculum cedures performed in Italy, 35% are carried out clandes- is certified by recognised international Jewish bodies tinely, at home or in other unprotected environments, (OU, UME, Initiation Society, central Rabbinate/Israe- and not by doctors [36, 37]. li Ministry of Health). The Mohalim must also pledge Unfortunately, precise statistical data on the rele- to observe the protocol of the procedure, which con- vance of the phenomenon, processed in Italy, are lack- tains a set of rules to safeguard the health of the new- ing. Infants of the Jewish community are, in fact, almost born. Clearly, it is not feasible for all the various reli- always circumcised in their community, while converted gions and ethnic groups which practice circumcision to adults carry out this procedure privately [38]. The same avail themselves of certified circumcisers; it is therefore problems affect the Islamic and Muslim communities. desirable that referral centres be instituted in hospitals. Based on broader reports, such as those provided by The heterogeneity within the NHS, together with Caritas, the General Secretary of the Islamic Cultural other factors (lack of adequate support by the com- Center in Italy hypothesized that, within the Muslim munity to which the individual belongs, scant financial community in Italy, MC involve a population of about resources of the family, lack of information), may well 40 000-45 000 newborn and children [39]. result in circumcision being carried out in unsafe condi- At the same time, Viviani et al., [40] underline that, tions. at least in some cases, ritual circumcisions have been The Italian Authority for Children and Adolescents wrongly labelled as “therapeutic” thus benefiting from recently addressed the issue of ritual circumcision in a the services of the national health system. note of recommendation to the Minister of Health; this In order to prevent the potential occurrence of severe invoked intervention to safeguard the health of newborn lesions and even death, increasing attention has been and children who risk suffering severe, and even lethal, focused on the question of whether the Italian NHS complications as a result of procedures performed out- should perform circumcision on demand. The CNB side healthcare facilities, and proposed the introduction claims that there are no reasons of an ethical or health- of a tariff scheme that would make ritual circumcision care nature that should prompt the State to impose on accessible to all income groups [35]. Such norms, how- the collectivity the cost of ritual MC practices that are ever, have aroused considerable protest. Specifically, directed not to the advantage of all members of society doctors in some Italian Regions have objected to ritual (regardless of their religious beliefs), but only of those circumcision, invoking their professional independence who belong to a specific religious confession [14]. and, in particular, citing the impossibility, from a deon- The above-mentioned regional experiments obviously tological and moral standpoint, of subjecting a person raise questions with regard to the difficulty of adequate- to any treatment whatsoever that does not have a medi- ly justifying the use of public resources to support the cal purpose. Moreover, this rejection has been further exercise of religious freedom and the right to health of 356 Francesco Ventura, Fiorella Caputo, Marta Licata et al

only those who profess some specific religious faiths. to carry out such procedures clandestinely and with dis- Nevertheless, the need to safeguard the right to health regard for good clinical and healthcare practice. It also (particularly of vulnerable subjects) could justify an ap- provides further food for thought concerning the need proach that permits non-therapeutic circumcision to be to legislate on this issue. offered by the NHS. At the same time, healthcare personnel (particularly The rise in the number of serious, and even fatal, obstetricians, gynaecologists and paediatricians) should incidents connected with MC procedures, however, implement campaigns to inform and educate women prompts careful consideration of the possibility of al- who wish to have their child circumcised. This approach reviews

lowing such operations on minors to be performed in may help to raise awareness of the risks of MC, of the NHS or NHS-affiliated facilities on payment of a fee importance of implementing proper safety measures

and that disadvantaged families can afford. Indeed, circum- and, not least, of the rights of children who are cur-

cision that is carried out clandestinely in unsuitable rently unable to decide for themselves, but who may environments, by unqualified persons, without proper one day decide not to adhere to the religion of their instruments and without the necessary hygiene, asep- parents. In the light of the case reported above, this sis and assistance, carries a high social cost in terms of type of education should also be targeted to healthcare articles

the harm that may be done to these particularly vulner- professionals operating in the field. able subjects. As circumcision is a surgical procedure, it must be carried out in a suitable healthcare setting and CONCLUSIONS by properly trained personnel. Indeed, like any surgical The medical, ethical, cultural and juridical issues sur- riginal operation, this procedure carries a risk of complications, rounding non-therapeutic circumcision are many and O which can be defined as early (such as haemorrhage, in- complex, and are deeply rooted in Western civilisation, fection, damage to the glans and urinary retention) and which is unequivocally founded on respect for the fun- late (such as phimosis, torsion of the penis and urethro- damental rights of the person and on the safeguard of cutaneous fistula) [41]. the psychophysical health of every member of society. The rate of complications varies enormously – from This respect clashes with the will to belong to commu- 0.06% to 55% – and depends, among other things, on nities in which circumcision is regarded as a primary the indications for the procedure, the operator who per- irrefutable symbol. Thus, the theme of circumcision is forms it and the place where it is performed. The lowest in some way paradigmatic of the complexity of today’s complication rates are seen when the procedure is car- pluralistic, multi-ethnic and multicultural society, in ried out in sterile conditions and by qualified surgeons. which various demands, traditions, rights and cultural, Indeed, in order to reduce the risk of complications, it religious and ideological references come into conflict is recommended that the operation be performed by or are, at least, difficult to reconcile. At the same time, qualified personnel who are trained in this type of inter- however, this scenario may constitute a “multicultural vention [42]. The procedures to be followed in order to laboratory” in which the various protagonists are called perform circumcision safely are described by the World upon not only to expound and defend their own ideas, Health Organisation in its “Manual for male circum- but also, and especially, to listen to the reasons of oth- cision under local anaesthesia” [43]. Serious adverse ers, thereby avoiding coercive methods and fostering di- events are rare and seem to be associated with a range alogue among religious leaders. Thus, on the one hand, of factors, such as age at the time of circumcision, op- ritual circumcision, being bereft of therapeutic indica- erator training and experience, the sterility of the envi- tions, impacts on the anatomical integrity of the new- ronment in which the procedure is carried out, and the born; on the other hand, however, it does not impair indications for the procedure itself. In particular, the the functionality of the organ. Moreover, as it is based qualifications and experience of the personnel are ma- on cultural and religious factors that are deeply rooted, jor factors in ensuring a good outcome in both the long to the extent that it is deemed essential in some com- and short term, while circumcisions practised for ritual munities, it may be claimed that it should be provided reasons – and consequently outside the clinical context in conditions of the greatest safety. – and for religious or traditional reasons are burdened Like any other surgical procedure, circumcision must by higher complication rates. be performed in such a way as to safeguard the child’s The possibility that MC could be performed by the health. It is therefore essential that it be carried out in NHS has also been entertained by both the Italian conditions of proper hygiene and sterility, in a suitable Society of Paediatrics [44] and the National Federa- healthcare setting and by experienced medical person- tion of Medical Council (FNOMCeO) [45]. With re- nel. Likewise, procedures of analgesia must be imple- gard to this latter Council, a study that we conducted mented in order to minimise the pain and suffering on a sample of 10 Provincial Associations of Surgeons caused by the operation. After the procedure, adequate (about 10% of the total) brought to light one case in post-operative monitoring should be carried out in or- which a regularly qualified doctor was reported to the der to prevent both short- and long-term complications. Council, and subsequently suspended for six months Obviously, these conditions of safety cannot be guar- (the maximum sanction) from practising medicine, for anteed in the domestic setting, in environments other having performed MC at the patient’s own home and in than proper healthcare facilities, or in the absence of the total absence of the necessary conditions of health qualified personnel. Indeed, as illustrated by the recent and hygiene. This case shows that, albeit very rarely, news stories, ritual MC that is performed in the ab- even qualified medical practitioners may be prepared sence of these requisites can have a tragic outcome, and 357 Male circumcision

deaths, mainly due to haemorrhage caused by technical of the fact that the sociocultural situation experienced shortcomings or to the lack of prophylaxis, are being by the minors of these communities differs from that of reported with increasing frequency. their parents and ancestors in their countries of origin. In the light of the above considerations, we feel that Furthermore, the intercultural perspective, as opposed the NHS should make every effort to render ritual cir- to the antagonism or clash of cultures, should consti- cumcision available free of charge. Indeed, the payment tute an ethical commitment in all sectors of social life, of a charge may clash significantly with the symbolic from education to health care. and ritual motivation of the act, thus hindering recourse Finally, we believe it is necessary to find a compro- reviews to its performance under a national health regime. mise between the Italian guidelines about circumcision Aware that the solution proposed creates further prob- (no surgical intervention before 4 years of age except

lems in terms of the allocation of resources, we never- in case of urinary retention, infection or inflammation) and theless feel that priority must be given to safeguarding and the religious dictates indicating that the circumci- the health of minors. sion has to be done during the first month of life, also in Moreover, we think that this proposal, which is order to prevent medico legal litigation. prompted by reasons of a prudential nature, cannot be Lastly, awareness-raising campaigns should target articles regarded as an exhaustive solution to a problem that is parents and the communities involved, in order to pub- complex and multiform (even in its operational modali- licise the potentially lethal risks associated with circum- ties) and which requires diversified strategies of action. cision outside the proper healthcare setting. The complexity of the problem is certainly exacerbated by the difficulty of identifying common ground with Authors’ contribution statement riginal those communities that are culturally less integrated All the Authors equally contributed to this work. O into our social fabric and which find it hard to accept certain modes of relating to the western world in which Declaration of funding they live. Indeed, these issues call into question the This research received no specific grant from any foundations of the relationship between the whole and funding agency in the public, commercial, or not-for- its parts, between the community and its members, be- profit sectors. tween the state and the individual. Thus, they manifest the difficulty of legislation both in reconciling the rights Conflict of interest statement of individuals with those of the collectivity, and in safe- The Authors declare that they have no conflict of in- guarding the rights of individual persons vis-a-vis the terest. community to which they belong. It is therefore essen- tial to promote constant dialogue, in order to promote Received on 7 September 2019. a relationship of trust and gradually to foster awareness Accepted on 19 May 2020.

REFERENCES

1. Cox G, Morris BJ. Why circumcision: from pre-history 9. Ledikwe JH, Nyanga RO, Hagon J, Grignon JS, Mpofu to the twenty-first century. In: Bolnick DA, Koyle MA, M, Semo BW. Scaling-up voluntary medical male cir- Yosha A (Ed.). Surgical guide to circumcision. London: cumcision – what have we learned? HIV AIDS (Auckl). Springer; 2012. 2014;6:139-46. doi: 10.2147/HIV.S65354 2. Arie S. Circumcision: Divided we fall. BMJ. 2010; 10. Brown JL. Medical-legal risks associated with circumci- 341:c4266. doi: 10.1136/bmj.c4266 sion of newborn males: need for revised consent. AAP 3. Morris BJ, Wamai RG, Henebeng EB, Tobian AA, Klaus- News. 2013;34.4:1. ner JD, Banerjee J, Hankins CA. Estimation of country- 11. Morris BJ, Kennedy SE, Wodak AD, Mindel A, Golovsky specific and global prevalence of male circumcision. D, Schrieber L, Lumbers ER, Handelsman DJ, Ziegler Popul Health Metr. 2016;14:4. doi: 10.1186/s12963-016- JB. Early infant male circumcision: Systematic review, 0073-5 risk-benefit analysis, and progress in policy. World J Clin 4. American Academy of Pediatrics Task Force on Circum- Pediatr. 2017;6(1):89-102. cision. Male circumcision. Pediatrics. 2012;130(3):e756- 12. World Health Organization, the Joint United Nations 85. doi: 10.1542/peds.2012-1990 Programme on HIV/AIDS (UNAIDS). Male circum- 5. Gollaher DL. From ritual to science: The medical cision. Global trends and determinants of prevalence, transformation of circumcision in America. J Soc Hist. safety and acceptability. 2007. Available from: http:// 1994;28(1):5-36. data.unaids.org/pub/report/2007/jc1360_male_circumci- 6. Wallerstein E. Circumcision: The uniquely American sion_en.pdf. medical enigma. Urol Clin North Am. 1985;12(1):123- 13. Svoboda JS, Adler PW, Van Howe RS. Circumcision is un- 32. ethical and unlawful. J Law Med Ethics. 2016;44(2):263- 7. Darby R. The masturbation taboo and the rise of routine 82. doi: 10.1177/1073110516654120 male circumcision: A review of the historiography. J Soc 14. Comitato Nazionale di Bioetica. La circoncisione: profili Hist. 2003;27:737-57. bioetici. 1998. Available from: http://bioetica.governo. 8. World Health Organization. WHO and UNAIDS an- it/it/documenti/pareri-e-risposte/la-circoncisione-profili- nounce recommendations from expert consultation on bioetici/. male circumcision for HIV prevention. Available from: 15. Gollaher DL. Circumcision: A history of the world’s most www.who.int/mediacentre/news/releases/2007/pr10/en/. controversial surgery. New York: Basic Books; 2000. 358 Francesco Ventura, Fiorella Caputo, Marta Licata et al

16. Alanis MC, Lucidi RS. Neonatal circumcision: a review blica Italiana – Serie Generale n. 14, 18 gennaio 2006. of the world’s oldest and most controversial operation. 33. Italia. Legge 8 marzo 1989, n. 101 “Norme per la rego- Obstet Gynecol Surv. 2004;59(5):379-95. lazione dei rapporti tra lo Stato e l’Unione delle Comu- 17. Dunsmuir WD, Gordon EM. The history of circumci- nità ebraiche italiane”. Gazzetta Ufficiale della Repub- sion. BJU Int. 1999;83(Suppl 1):1-12. blica Italiana – Serie Generale n. 69, 23 marzo 1989. 18. Aggleton P. “Just a snip? A social history of male circum- 34. Tribunale di Padova, 9 novembre 2007. Brevi note in cision”. Reprod Health Matters. 2007;15(29):15-21. tema di circoncisione rituale maschile, esercizio abusivo 19. Ebers G, Stern L. Papyros Ebers das hermetische Buch della professione e lesioni. Giur Merito. 2008:2590 e ss. über die Arzneimittel der alten Ägypter, 2. Leipzig: Glos- 35. Autorità Garante per l’infanzia. Segnalazione ai sensi reviews

sar und Text; 1875. dell’articolo 3, comma 1, legge n. 112 del 2011. Circonci- 20. Riggs T. Christianity: Coptic christianity, in worldmark sione rituale maschile. Available from: www.garanteinfan-

and encyclopedia of religious practices: religions and denomi- zia.org/sites/default/files/segnalazione-ministro-salute. nations. Farmington Hills: Thomson Gale; 2006. pdf. 21. Sahin F, Beyazova U, Aktürk A. Attitudes and practices 36. Associazione di Medici di Origine Straniera in Italia. No regarding circumcision in Turkey. Child Care Health Dev. al massacro dei bambini da circoncisione clandestina. 2003;29(4):275-80. AMSI; 2019. Available from: www.amsimed.org.

articles 22. Mahmood K, Nagra, ZM, Ahmad S, Malik MA, Hameed 37. Sottosegretariato Italiano alla Salute. Comunicato n. 40, S. Circumcision trends in Pakistan. J Uni Med Dent Coll. 5 aprile 2019. Available from: www.salute.gov.it/portale/ 2015;1:17-21. news/p3_2_4_1_1.jsp?lingua=italiano&menu=salastampa 23. Anwer AW, Samad L, Iftikhar S, Baig-Ansari N. Re- &p=comunicatistampa&id=5170. ported male circumcision practices in a muslim-major- 38. Angelucci A. Dietro la circoncisione: la specialità confes- riginal ity setting. Biomed Res Int. 2017;2017:4957348. doi: sionale. Il caso paradigmatico dell’Italia e di altri mod- O 10.1155/2017/4957348 elli. In: Angelucci A. Dietro la circoncisione. La sfida 24. Sabzehei MK, Mousavi-bahar SH, Bazmamoun H. Male della cittadinanza e lo spazio di libertà religiosa in Eu- neonatal circumcision. J Compr Ped. 2012;4(1):49-53. ropa. Torino: G. Giapichelli Editore; 2018. Available 25. Alfano L, Bonsignore A, Ciliberti R. Therapeutic from: www.giappichelli.it/media/catalog/product/sum- choices and care of minors: a recent story. Med Histor. mary/9788892114104.pdf. 2019;3(1):52-7. 39. Bombardieri F. Mappatura dell’associazionismo islamico. 26. Ciliberti R, Alfano L, Baldelli I, De Stefano F, Bonsignore In Angelucci A, Bombardieri M, Tacchini D (Eds). Islam A. Self-determination, healthcare treatment and minors e integrazione in Italia. Venezia: Marsilio; 2016. in Italian clinical practice: ethical, psychological, juridical 40. Viviani F, Bobbo S, Malaguti S, Paolini D. NOCIRC and medical-legal profiles. Acta Biomed. 2018;89(1):34- of Italy: Scientific activities 2006-2009. In: Denniston 40. doi: 10.23750/abm.v89i1.6368 GC, Hodges FM, Milos MF (Eds). Genital autonomy, 27. Svoboda JS, Van Howe RS, Dwyer JG. Informed consent springer science & business media. Springer; 2010. doi: for neonatal circumcision: an ethical and legal conun- 10.1007/978-90-481-9446-9_9 drum. J Contemp Health Law Policy. 2000;17(1):61-133. 41. Ben Chaim J, Livne PM, Binyamini J, Hardak B, Ben- 28. Dwyer JG (Ed). The Oxford handbook of children and Meir D, Mor Y. Complications of circumcision in Is- the law. New York: Oxford University Press; 2018. rael: a one year multicenter survey. Isr Med Assoc J. 29. Angelucci A. Liberta religiosa e circoncisione in Italia: 2005;7(6):368-70. una questione di specialità confessionale. Stato, Chiese 42. Edler G, Axelsson I, Barker GM, Lie S, Naumburg E. Se- e pluralismo confessionale. Rivista telematica. 2016;35. rious complications in male infant circumcisions in Scan- Available from: www.statoechiese.it/images/uploads/ar- dinavia indicate that this always be performed as a hospi- ticoli_pdf/angelucci.m_libert.pdf?pdf=liberta-religiosa-e- tal-based procedure. Acta Paediatr. 2016;105(7):842-50. circoncisione-in-italia-una-questione-di-specialita-con. doi: 10.1111/apa.13402 30. Fateh-Moghadam B. Criminalizing male circumcision? 43. World Health Organization. Manual for male circumci- Case Note: Landgericht Cologne, Judgment. German sion under local anaesthesia. Geneva: WHO; 2009. Law Journal. 2012;13(9):1131-45. 44. Società Italiana di Pediatria. Circoncisione, proce- 31. Swatek-Evenstein M. Limits of enlightenment and the dure eterogenee. SIP; 2019. Available from: www.sip. law. On the legality of ritual male circumcision in Europe it/2019/03/29/circoncisione-procedure-eterogenee/. today. Merkourios. 2013;9(77):42-50. 45. Federazione Nazionale degli Ordini dei Medici Chirurghi 32. Italia. Legge 9 gennaio 2006, n. 7 “Disposizioni concern- e degli Odontoiatri. La circoncisione sia inserita nei LEA. enti la prevenzione e il divieto delle pratiche di mutilazi- FNOMCeO; 2019. Available from: https://portale.fnom- one genitale femminile”. Gazzetta Ufficiale della Repub- ceo.it/anelli-fnomceo-la-circoncisione-sia-inserita-nei-lea/. and theUnitofAntibioticResistanceSpecialPathogens 1 Rome, Italy. E-mail:[email protected]. Address forcorrespondence: LanfrancoFattorini, DipartimentodiMalattieInfettive, IstitutoSuperiorediSanità,Viale Regina Elena299,00161 Staphylococcus aureus, in previousinfluenzapandemics,affectCOVID-19patients.Mycoplasmapneumoniae, Current literatureshowsthatsecondarybacterialinfections,althoughlessfrequentthan Abstract * Lanfranco Fattorini an underestimatedadversary Bacterial coinfectionsinCOVID-19: DOI: 10.4415/ANN_20_03_14 Ann IstSuperSanità2020|Vol. 56,No.3:359-364 and when they were admitted to the Intensive Care Units decreased survivalofCOVID-19 patients,particularly reports showed that secondary infections significantly ing spread of the disease worldwide. However, some navirus (COVID-19) are limited due to the still ongo- the severediseasecausedby theSARS-CoV-2 coro- WITH THE2020PANDEMIC SECONDARY INFECTIONSASSOCIATED virus pandemic. on bacterialcoinfectionsreportedinthe2020corona- monia severity[1].Here,weexploredcurrentliterature viral respiratory diseases, and lead toincrease in pneu- Secondary bacterialinfectionscanbecomplicationsof notic originSARS-CoVandMERS-CoV, respectively. (MERS) [3], both caused by the coronavirus of zoo- [2] andthe2012MiddleEastrespiratorysyndrome the 2002severeacuterespiratorysyndrome(SARS) in the2009swineinfluenzapandemic[1]andduring cently, secondary bacterial infections were also reported monia, particularlycausedbyS.pneumoniae.Morere- influenza virusalone,butbysubsequentbacterialpneu- with influenzapandemicof1918werenotcausedby of casesinameta-analysis[1].Mostdeathsassociated common causes,andratesrangingbetween1135% and with enza pandemicsarewelldescribedintheliterature, INTRODUCTION stewardship principles. patients, with more attention to evidence-based studies and respect for the antimicrobial revise frequentandempiricprescriptionofbroad-spectrumantibioticsinCOVID-19 infections by nosocomial antibiotic-resistant bacteria. This highlightsthe urgency to in patientsadmittedtheintensivecareunits,andthosediseasescanbeduetosuper- COVID-19 coinfectionsarealsoreported.However, bacterialcoinfectionratesincrease The membersoftheUnitAntibioticResistanceandSpecialPathogensarelistedbeforeReferences Dipartimento diMalattieInfettive,IstitutoSuperioreSanità,Rome,Italy Data regarding secondary respiratory infections in Secondary bacterialinfectionsassociatedwithinflu- Klebsiella Staphylococcus aureusbeingreportedasthemost Streptococcus pneumoniae, spp. are the main species isolated. Of note, Legionella pneumophila, 1 , RobertaCreti Haemophilus influenzae, 1 , CarlaPalma Streptococcus pneumoniae, the death-risk[6]. pose challengesinantibiotic therapy, and toincrease nosocomial, antibiotic-resistant pathogenisknownto resistant with bacteriaintheICU-death group,carbapenem- the generalwards.InCOVID-19 patientscoinfected 8.7%, respectively, ofpatientstransferredfromICUto spectively, of patients in the ICU, and in and othervirusesoccurredin55.6,44.444.4%,re- deaths associatedwithcoinfectionsbybacteria,fungi respectively, thanthosenotseverelyill.Furthermore, have coinfectionswithbacteria,fungiandotherviruses, illness were14.2,18.2and2.9timesmorelikelyto 221 COVID-19patientsinWuhan, thosewithsevere and hadapositivecultureofnewpathogen[4,5]. clinical symptomsorsignsofpneumoniabacteremia, ary infectionswerediagnosedwhenpatientsshowed next-generation sequencing(NGS)methods.Second- viruses andCOVID-19virus,usingreal-timePCRor and fungalexaminationsforcommonrespiratory aspirates) andbloodweretestedforroutinebacterial geal swabs,sputum,bronchoalveolarlavages,bronchial survivors [5].Respiratoryspecimens(nasalandpharyn- was reportedin50%ofnon-survivorsandonly1% care and0%inno-ICU care. A secondary infection 31% ofthemrequiringmechanicalventilationinICU observed in10%and15%,respectively, ofpatients,with performed inWuhan, China,secondaryinfectionswere involving 41and191COVID-19patients,respectively, (ICU). InthestudiesofHuangandZhouetal.[4,5] In anotherstudy, Zhangetal.[6]reportedthatin Mycobacterium tuberculosis- 1 , AnnalisaPantosti Acinetobacter baumanniiwasisolated.This 1* Haemophilus 1

• • • • • • Key words intensive careunits therapy antibiotics superinfections bacterial coinfections COVID-19 26.1, 13.0 and

359 Original articles and reviews 360 Lanfranco Fattorini, Roberta Creti, Carla Palma et al.

COMMON SPECIES INVOLVED IN COVID-19 MYCOBACTERIUM TUBERCULOSIS-COVID COINFECTIONS COINFECTIONS The species of the microorganisms identified in CO- During the SARS and MERS epidemics, few coin- VID-19 positive specimens are reported in Table 1 [7- fections involving Mycobacterium tuberculosis (Mtb) 19]. Overall, the 13 studies performed on a total of were reported [22, 23]. However, in the COVID-19 733 patients showed that viral coinfections, including pandemic the World Health Organization and other mainly influenza virus and rhinovirus/enterovirus, oc- institutions published several documents including curred in 17.2% of patients (126/733), while bacterial sustainability of tuberculosis (TB) services [24] and reviews

coinfections due to both Gram-positive and Gram- information on similarity, differences and interactions negative species and Mycoplasma pneumoniae occurred between these two dangerous respiratory pathogens, to

and in 11.7% (86/733) patients, and fungal coinfections in anticipate the impact of COVID-19 on TB patients and

1.8% (13/733) patients. The bacterial species more fre- TB control programmes [25, 26]. These publications quently isolated were, in ranking order, M. pneumoniae, were important to better tackle TB and COVID-19 S. aureus, Legionella pneumophila, Haemophilus spp., pandemics worldwide. Indeed, clinical data were re- Klebsiella spp., Pseudomonas aeruginosa, Chlamydia ported in studies regarding TB-COVID coinfections articles spp., S. pneumoniae, A. baumannii. Patients in the ICU [27, 28] and characterization of patients who died with were 522/733 [8, 10, 14-16, 18, 19], and 1.3% of them Mtb and COVID-19 [29]. (7/522) developed nosocomial super-infections with For instance, Stochino et al. [27] reported TB-CO- antibiotic-resistant S. aureus, Klebsiella pneumoniae, P. VID coinfections in 20/24 TB patients admitted to the riginal aeruginosa, or A. baumannii. Apparently, no antibiotic- phthisiology unit of the hospital of Sondalo (Northern O resistant strains were isolated outside the ICU. Italy). In the 3-4 weeks following COVID-19 diagno- Another study [20] reported bacteremia by clinical sis, the clinical course of TB and COVID-19 coinfec- pathogens in 21/643 blood cultures (3.3%) from CO- tion was generally benign, but follow-up was limited to VID-19 patients, with respiratory sources being con- a few weeks, not allowing assessment of longer-term firmed in two cases (a community acquiredK. pneu- outcomes. Following analysis of the dynamics of the moniae and a ventilator associated Enterobacter cloacae). infection spread in the hospital, it was apparent that All other bacteremias were attributed to non-respirato- the outbreak was due to insufficient control practices ry sources. No pneumococcal, legionella or influenza associated with a higher vulnerability of TB patients. infections were detected. No patient was admitted in the ICU. Furthermore, Overall, bacterial infections reported in COVID-19 Tadolini et al. [28] reported TB-COVID infections in patients were less frequent and different from those 49 patients with current or former TB occurring in 8 causing lower respiratory tract infections in influenza countries (Belgium, Brazil, France, Italy, Russia, Sin- pandemics [1], with S. pneumoniae being rarely isolat- gapore, Spain, Switzerland). Patients were treated ed. It must be noted that bacteria were mainly cultured with first-line TB drugs and, in case of MDR-TB, with from nasopharyngeal samples, while lower respiratory second-line drugs. Medications for COVID-19 in- samples were less available also due to safety concern cluded anti-viral drugs (lopinavir/ritonavir, darunavir/ for performing bronchoalveolar lavage. Bacterial diag- cobicistat) and antibiotics (azithromycin). Diagnosis, nosis was performed by routine methods (not detailed treatment and outcome details of the 49 COVID-19 in the papers), and by multiplex PCR kits for rapid de- patients showed various clinical profiles, thus larger tection of a wide range of respiratory pathogens, most- studies are necessary. Finally, after preliminary analy- ly viruses. Thus, coinfections by bacterial species not sis of 8 deaths occurring in the 69 TB-COVID coin- included in multiplex PCR kits [7-9], or not searched fections reported by Stochino et al. [27] and Tadolini during emergency, could have been under-estimated, et al. [28], Motta et al. [29] concluded that i) higher so as to undervalue their contribution in COVID-19 mortality was likely to occur in elderly patients with severity and mortality. Early and rapid diagnosis and comorbidities, ii) TB might not be a major determi- drug susceptibility testing of mixed bacterial infec- nant of mortality, iii) migrants had lower mortality, tions by culture-independent approaches such as, for probably because of their younger age and lower num- instance, NGS methods and Nanopore metagenomics ber of comorbidities. [21], could better guide/adjust antibiotic therapy so as to prevent fatal outcomes, particularly in case of Multi TREATMENT OF BACTERIAL INFECTIONS IN Drug Resistant (MDR) bacteria. THE COVID-19 PANDEMIC Rapid detection of bacterial infections may also limit The respiratory symptoms of patients with COV- development of virus super-spreaders, defined as pa- ID-19 pneumonia admitted to hospital with fever and tients infecting ≥ 10 persons each. For instance, in Sin- dry cough can mimic those of atypical bacterial pneu- gapore, during the SARS-CoV outbreak, two patients monia, making difficult to distinguish patients with hospitalized with bacterial infections were co-infected hospital acquired and ventilator associated pneumonia with SARS-CoV, and caused 76% of SARS-CoV infec- (VAP). A biomarker used to differentiate bacterial from tions in a healthcare facility [2]. Thus, to contain cur- viral infections is procalcitonin [4-6, 8, 10, 11, 14, 16- rent COVID-19 pandemic it is important to triage and 18], a peptide whose serum levels increase during bac- isolate patients with known bacterial infections in des- terial but not viral infections. ignated wards, and to apply efficient infection control To decrease chances of VAP in the ICU, most CO- measures, in order to limit virus super-spreading. VID-19 patients were empirically treated with antibiot- 361 Bacterial coinfections in COVID-19

ics. The principles of antibiotic stewardship should be Overall, since COVID-19 pandemic is still ongoing, considered, but in the case of severely ill patients, the and transfer of patients in the ICU continues, the use concern surrounding the pandemic forced clinicians to of antibiotics will steady raise and increase develop- start treatment with antibiotics. Indeed, Table 1 shows ment and transmission of MDR strains in the health- that 88.3% of COVID-infected patients (476/539) care systems. Thus, when the probability of a bacterial were treated with broad-spectrum antibiotics including infection is low, antibiotic treatment of COVID-19 pa- third-generation cephalosporins, quinolones, carbapen- tients should be re-evaluated, and stopped if not nec- ems. The choice of empiric regimens should take into essary. Antibiotics should be reserved for patients with reviews account possible side effects (e.g. QT prolongation, the most severe respiratory presentations [20, 31]. diarrhoea), local epidemiology of drug resistance, and

impact of drug resistance on the patient. In some coun- CONCLUSION and tries, bacteria are resistant to at least one antibiotic From current reports, incidence of bacterial coinfec- class, therefore empiric broad-spectrum therapy could tions in COVID-19 cases is lower than in previous in- have limited effect particularly in hospital-acquired in- fluenza pandemics. However, coinfection rates increase fections. In case of sepsis, inadequate antibiotic therapy in patients admitted to the ICU. Super-infections by articles may increase mortality [30]. antibiotic-resistant bacteria occur in 1.3% of patients in

Table 1 Review of recent literature describing the species involved in COVID-19 coinfections riginal O N of bacterial or fungal N of viral Clinical sample Diagnostic N of pts treated Pts Country Reference coinfections / N of coinfections* / N methods with antibiotics treated COVID-19 patients (pts) of COVID-19 pts (AB), antifungals in the tested (%) tested (%) (AF), antivirals ICU (%) (AV) / N of total pts treated (%). [AB resistances (AB-R)] Mycoplasma pneumoniae: REV: 12/67 (17.9) Nasopharyngeal Routine diagnosis, AB (doxycycline, NR United [7] 1/67 (1.5); OCV: 10/67 (14.9) swabs BioFire Film Array moxifloxacin): 8/9 Kingdom Staphylococcus aureus: IAV: 3/67 (4.5) Respiratory Panel 2 (88.9) 1/49 (2); RSV, ADV, HMV: plus (17 viruses, 4 AV (oseltamivir): 1/9 Haemophilus influenzae: 3/67 (4.5) bacteria) (BioMerieux) (11.1) 3/49 (6.1) Escherichia coli: 1/49 (2) M. pneumoniae: 1/42 (2.4); REV: 22/42 (52.4) Nasopharyngeal RT-PCR (Respiratory NR 14.2 USA [8] Chlamydia pneumoniae: OCV: 7/42 (16.7) samples panel) 2/42 (4.8) IAV: 1/42 (2.4) NR RSV: 4/42 (9.6) HMV: 2/42 (4.8) PIV: 3/42 (7.1) Haemophilus REV: 2/20 (10) Nasopharyngeal RT-PCR, NGS, ResPlex NR NR China [9] parainfluenzae: 4/20 (20.0); IAV, IBV: 2/20 (10) swabs II V2.0 kit Respiratory Klebsiella aerogenes: 1/20 RSV: 1/20 (5) panel (17 viruses, 3 NR (5); Candida albicans: bacteria) (Qiagen) 1/20 (5) Acinetobacter baumannii Not found Throat-swab Real time RT-PCR AB (cephalosporins, 23 China [10] + Klebsiella pneumoniae + specimens, quinolones, Aspergillus flavus: 1/99 (1); sputum, carbapenems, Candida spp: 4/99 (4) endotracheal tigecycline): 70/99 aspirates (70.7); AF: 15/99 (15.2); AV (oseltamivir, gangiclovir, lopinavir/ritonavir): 75/99 (75.8). [AB-R: 1 A. baumannii] M. pneumoniae: 4/20 (20) IAV, IBV: 3/20 (15) Pharyngeal swabs Nucleic acid tests NR NR China [11] RSV, CMV: 2/20 (10) from pediatric patients Legionella pneumophila: IAV, IBV: 34/68 (50) Acute phase IgM antibodies NR NR China [12] 6/68 (8.8); serum by indirect M. pneumoniae: 8/68 immunofluorescence (11.8) Enterobacter cloacae: 2/29 RSV, ADV: 2/28 (7.1) Sputum, serum Culture, IgG and IgM AB (mostly NR China [13] (6.9); antibodies in blood moxifloxacin): 66/67 C. albicans: 2/29 (6.9); (98.6); A. baumannii: 1/29 (3.4); AF: 8/67 (11.9); Chlamydia: 2/28 (7.1), AV (mostly by IgG umifenovir): 66/67 (98.6); Continues 362 Lanfranco Fattorini, Roberta Creti, Carla Palma et al.

Table 1 Continued

N of bacterial or fungal N of viral Clinical sample Diagnostic N of pts treated Pts Country Reference coinfections / N of coinfections* / N methods with antibiotics treated COVID-19 patients (pts) of COVID-19 pts (AB), antifungals in the tested (%) tested (%) (AF), antivirals ICU (%) (AV) / N of total pts treated (%). [AB resistances (AB-R)] reviews Gram-positive cocci and IAV, CMV: 2/11 Sputum, serum Culture, IgM AB (mostly 9.1 China [14] Gram-negative bacilli (18.2), by IgM antibodies in blood moxifloxacin): 4/11 (not identified): 1/11 (9.1) (36.4); and

AF (caspofungin): 1/11 (9.1); AV (mostly umifenovir): 8/11 (72.7).

articles A. baumannii + A. flavus: Not found Upper and lower Culture, multiplex PCR AB (meropenem, 40 France [15] 1/5 (20); respiratory tract tigecycline, colistin): samples, blood, 1/5 (20); urine, stools AF (isavuconazole): 1/5 (20); AV (remdesivir): riginal 3/5 (60). O Pseudomonas spp 1/21 IAV: 2/21 (9.6) Nasopharyngeal NR NR 81 USA [16] (4.8) PIV: 1/21 (4.8) samples L. pneumophila 1/7 (14.3) IAV: 2/7 (28.6) Sputum, Real time RT-PCR AB (cephalosporins, NR China [17] endotracheal macrolides, aspirates quinolones): 7/7 (100) AV (oseltamivir, gangiclovir, umifenovir); 7/7 (100) S. aureus 7/289 (2.4); IAB, IBV: 2/289 (0.7) Upper and lower Culture, respiratory AB (not specified): 36.3 USA [18] Streptococcus pneumoniae REV 2/289 (0.7) respiratory tract viral panel, PCR 271/289 (93.8) 4/289 (1.4); ADV, HMV: 2/289 samples, blood, AV (remdesivir): Pseudomonas aeruginosa (0.7) urine 28/289 (9.7) 1/289 (0.3); [AB-R: 3 methicillin Klebsiella spp 2/289 (0.7); resistant S. aureus] E. coli 1/289 (0.3); Bordetella holmesii 1/289 (0.3); Stenothrophomonas maltophilia 1/289 (0.3); Aspergillus niger 1/289 (0.3) Oral bacterial flora 15/289 (5.2) K. pneumoniae 2/55 (3.6) NR Respiratory tract NR AB (not specified): 100 China [19] P. aeruginosa 1/55 (1.8) samples, blood 49/52 (94.2) Serratia marcescens 1/55 AV (remdesivir, (1.8) lopinavir): 23/52 Aspergillus spp 2/55 (3.6) (44.2) C. albicans 1/55 (1.8) [AB-R: 2 K. pneumoniae, 1 P. aeruginosa] *Influenza A virus (IAV), influenza B virus (IBV), other coronavirus (OCV), rhinovirus/enterovirus (REV), respiratory sincytial virus (RSV), adenovirus (ADV), human metapneumovirus (HMV), parainfluenza 3 virus (PIV), cytomegalovirus (CMV). Not reported (NR).

ICU and 0% in no-ICU care. M. tuberculosis-COVID-19 Annalisa Pantosti, Fabrizio Barbanti, Romina Camilli, coinfections are also reported. Overall, despite frequent Alessandra Ciervo, Rosanna Dattilo, Maria Del Grosso, prescription of broad-spectrum antibiotics, antimicro- Giulia Errico, Daniela Fortini, Aurora Garcia Fernan- bial stewardship principles should be re-considered to dez, Federico Giannoni, Maria Giufré, Monica Imperi, avoid development and transmission of drug resistant Claudia Lucarelli, Fabiola Mancini, Monica Monaco, organisms in healthcare facilities. Francesca Mondello, Fernanda Pimentel de Araujo, Maria Luisa Ricci, Maria Scaturro, Patrizia Spigaglia, The members of the Unit of Antibiotic Resistance Antonella Torosantucci, Laura Villa. and Special Pathogens of the Department of Infectious Giulia Errico is a fellow of the European Program Diseases, Istituto Superiore di Sanità, Rome, Italy, are for Public Health Microbiology Training (EUPHEM), listed below: European Centre for Disease Prevention and Control, Lanfranco Fattorini, Roberta Creti, Carla Palma, Stockholm, Sweden. 363 Bacterial coinfections in COVID-19

Conflict of interest statement and findings of this study. There are no potential conflicts of interest or any fi- nancial or personal relationships with other people or Received on 28 June 2020. organizations that could inappropriately bias conduct Accepted on 10 July 2020.

REFERENCES reviews 1. Morris DE, Cleary DW, Clarke SC. Secondary bacterial CT features in pediatric patients with COVID-19 infec- infections associated with influenza pandemics. Front tion: Different points from adults. Pediatr Pulmonol. and

Microbiol. 2017;8:1041. doi: 10.3389/fmicb.2017.01041 2020;55(5):1169-74. doi: 10.1002/ppul.24718. 2. Wilder-Smith A, Green JA, Paton NI. Hospitalized pa- 12. Xing Q, Li G, Xing Y, Chen T, Li W, Ni W, Deng K, tients with bacterial infections: a potential focus of SARS Gao R, Chen C, Gao Y, Li Q, Yu G, Tong J, Li W, Hao transmission during an outbreak. Epidemiol Infect. G, Sun Y, Zhang A, Wu Q, Li Z, Pan S. Precautions 2004;132(3):407-8. doi: 10.1017/S0950268803001869 are needed for COVID-19 patients with coinfection of articles

3. Memish ZA, Perlman S, Van Kerkhove MD, Zum- common respiratory pathogens. medRxiv. 2020. doi: la A. Middle East respiratory syndrome. Lancet. 10.1101/2020.02.29.20027698 2020;395(10229):1063-77. doi: 10.1016/S0140- 13. Wang Z, Yang B, Li Q, Wen L, Zhang R. Clinical Fea- 6736(19)33221-0 tures of 69 Cases with Coronavirus Disease 2019 in Wu-

4. Huang C, Wang Y, Li X, Ren L, Zhao J, Hu Y, Zhang L, han, China. Clin Infect Dis. 2020. ciaa272. https://doi. riginal

Fan G, Xu J, Gu X, Cheng Z, Yu T, Xia J, Wei Y, Wu W, org/10.1093/cid/ciaa272. O Xie X, Yin W, Li H, Liu M, Xiao Y, Gao H, Guo L, Xie J, 14. Dong X, Cao YY, Lu XX, Zhang JJ, Du H, Yan YQ, Akdis Wang G, Jiang R, Gao Z, Jin Q, Wang J, Cao B. Clinical CA, Gao YD. Eleven faces of coronavirus disease. Al- features of patients infected with 2019 novel coronavi- lergy. 2020. doi: 10.1111/all.14289 rus in Wuhan, China. Lancet. 2020;395(10223):497-506. 15. Lescure FX, Bouadma L, Nguyen D, Parisey M, Wicky doi: 10.1016/S0140-6736(20)30183-5 PH, Behillil S, Gaymard A, Bouscambert-Duchamp M, 5. Zhou F, Yu T, Du R, Fan G, Liu Y, Liu Z, Xiang J, Wang Donati F, Le Hingrat Q, Enouf V, Houhou-Fidouh N, Y, Song B, Gu X, Guan L, Wei Y, Li H, Wu X, Xu J, Valette M, Mailles A, Lucet JC, Mentre F, Duval X, Des- Tu S, Zhang Y, Chen H, Cao B. Clinical course and camps D, Malvy D, Timsit JF, Lina B, van-der-Werf S, risk factors for mortality of adult inpatients with CO- Yazdanpanah Y. Clinical and virological data of the first VID-19 in Wuhan, China: a retrospective cohort study. cases of COVID-19 in Europe: a case series. Lancet Lancet. 2020;395(10229):1054-62. doi: 10.1016/S0140- Infect Dis. 2020;20(6):697-706. doi: 10.1016/S1473- 6736(20)30566-3 3099(20)30200-0 6. Zhang G, Hu C, Luo L, Fang F, Chen Y, Li J, Peng Z, 16. Arentz M, Yim E, Klaff L, Lokhandwala S, Riedo FX, Pan H. Clinical features and short-term outcomes of 221 Chong M, Lee M. Characteristics and Outcomes of patients with COVID-19 in Wuhan, China. J Clin Virol. 21 Critically Ill Patients With COVID-19 in Washing- 2020;127:104364. doi: 10.1016/j.jcv.2020.104364. ton State. JAMA. 2020;323(16):1612-4. doi: 10.1001/ 7. Easom N, Moss P, Barlow G, Samson A, Taynton T, jama.2020.4326 Adams K, Ivan M, Burns P, Gajee K, Eastick K, Lillie 17. Yu N, Li W, Kang Q, Xiong Z, Wang S, Lin X, Liu Y, PJ. Sixty-eight consecutive patients assessed for CO- Xiao J, Liu H, Deng D, Chen S, Zeng W, Feng L, Wu VID-19 infection: Experience from a UK Regional in- J. Clinical features and obstetric and neonatal outcomes fectious diseases Unit. Influenza Other Respir Viruses. of pregnant patients with COVID-19 in Wuhan, China: 2020;14(4):374-9. doi: 10.1111/irv.12739 a retrospective, single-centre, descriptive study. Lancet 8. Richardson S, Hirsch JS, Narasimhan M, Crawford Infect Dis. 2020;20(5):559-64. doi: 10.1016/S1473- JM, McGinn T, Davidson KW; and the Northwell CO- 3099(20)30176-6 VID-19 Research Consortium, Barnaby DP, Becker 18. Crotty MP, Akins RL, Nguyen AT, Slika R, Rahmanzadeh LB, Chelico JD, Cohen SL, Cookingham J, Coppa K, K, Wilson MH, Dominguez EA. Investigation of subse- Diefenbach MA, Dominello AJ, Duer-Hefele J, Falzon quent and co-infections associated with SARS-CoV-2 L, Gitlin J, Hajizadeh N, Harvin TG, Hirschwerk DA, (COVID-19) in hospitalized patients. medRxiv. 2020. Kim EJ, Kozel ZM, Marrast LM, Mogavero JN, Osorio doi: 10.1101/2020.05.29.20117176 GA, Qiu M, Zanos TP. Presenting characteristics, comor- 19. Yang X, Yu Y, Xu J, Shu H, Xia J, Liu H, Wu Y, Zhang bidities, and outcomes among 5700 patients hospital- L, Yu Z, Fang M, Yu T, Wang Y, Pan S, Zou X, Yuan ized with COVID-19 in the New York City area. JAMA. S, Shang Y. Clinical course and outcomes of critically ill 2020;323(20):2052-9. doi: 10.1001/jama.2020.6775 patients with SARS-CoV-2 pneumonia in Wuhan, China: 9. Ai J-W, Zhang H-C, Xu T, Wu J, Zhu M, Yu Y-Q, Zhang a single-centered, retrospective, observational study. Lan- H-Y, Shen Z, Li Y, Zhou X, Zang G-Q, Xu J, Chen W-J, cet Respir Med. 2020;8(5):475-81. doi: 10.1016/S2213- Li YJ, Xie D-S, Zhou M-Z, Sun JY, Chen JZ, Zhang 2600(20)30079-5 W-H. Optimizing diagnostic strategy for novel coronavi- 20. Hughes S, Troise O, Donaldson H, Mughal N, Moore rus pneumonia, a multi-center study in Eastern China. LS. Bacterial and fungal coinfection among hospitalised medRxiv. 2020. doi: 10.1101/2020.02.13.20022673 patients with COVID-19: A retrospective cohort study in 10. Chen N, Zhou M, Dong X, Qu J, Gong F, Han Y, Qiu a UK secondary care setting. Clinical Microbiology and Y, Wang J, Liu Y, Wei Y, Xia J, Yu T, Zhang X, Zhang L. Infection. doi: 10.1016/j.cmi.2020.06.025 Epidemiological and clinical characteristics of 99 cases of 21. Charalampous T, Kay GL, Richardson H, Aydin A, Bal- 2019 novel coronavirus pneumonia in Wuhan, China: a dan R, Jeanes C, Rae D, Grundy S, Turner DJ, Wain descriptive study. Lancet. 2020;395(10223):507-13. doi: J, Leggett RM, Livermore DM, O’Grady J. Nano- 10.1016/S0140-6736(20)30211-7. pore metagenomics enables rapid clinical diagnosis of 11. Xia W, Shao J, Guo Y, Peng X, Li Z, Hu D. Clinical and bacterial lower respiratory infection. Nat Biotechnol. 364 Lanfranco Fattorini, Roberta Creti, Carla Palma et al.

2019;37(7):783-92. doi: 10.1038/s41587-019-0156-5 L, De Souza-Galvão ML, Dominguez-Castellano A, 22. Alfaraj SH, Al-Tawfiq JA, Altuwaijri TA, Memish ZA. Dourmane S, Jachym MF, Froissart A, Giacomet V, Go- Middle East respiratory syndrome coronavirus and pul- letti D, Grard S, Gualano G, Izadifar A, Le Du D, Royo monary tuberculosis coinfection: Implications for in- MM, Mazza-Stalder J, Motta I, Ong CWM, Palmieri F, fection control. Intervirology. 2017;60(1-2):53-5. doi: Rivière F, Rodrigo T, Rossato Silva D, Sánchez-Montalvá 10.1159/000477908 A, Saporiti M, Scarpellini P, Schlemmer F, Spanevello 23. Wong CY, Wong KY, Law TS, Shum TT, Li YK, Pang WK. A, Sumarokova E, Tabernero E, Tambyah PA, Tiberi S, Tuberculosis in a SARS outbreak. J Chin Med Assoc. Torre A, Visca D, Murguiondo MZ, Sotgiu G, Migliori 2004;67(11):579-82. PMID: 15720073 GB. Active tuberculosis, sequelae and COVID-19 co- reviews

24. World Health Organization. COVID-19: Considerations infection: first cohort of 49 cases. Eur Respir J. 2020. doi: for tuberculosis (TB) care. WHO; 2020. www.who.int/ 10.1183/13993003.01398-2020.

and docs/default-source/documents/tuberculosis/infonote-tb- 29. Motta I, Centis R, D’Ambrosio L, García-García JM, covid-19.pdf. Goletti D, Gualano G, Lipani F, Palmieri F, Sánchez- 25. Wingfield T, Cuevas LE, MacPherson P, Millington KA, Montalvá A, Pontali E, Sotgiu G, Spanevello A, Sto- Squire SB. Tackling two pandemics: a plea on World Tu- chino C, Tabernero E, Tadolini M, van den Boom M, berculosis Day. Lancet Respir Med. 2020;8(6):536-8. Villa S, Visca D, Migliori GB. Tuberculosis, COVID-19

articles doi: 10.1016/S2213-2600(20)30151-X and migrants: preliminary analysis of deaths occurring in 26. Togun T, Kampmann B, Stoker NG, Lipman M. Antici- 69 patients from two cohorts. Pulmonology. 2020. doi: pating the impact of the COVID-19 pandemic on TB 10.1016/j.pulmoe.2020.05.002. patients and TB control programmes. Ann Clin Micro- 30. Rhee C, Kadri SS, Dekker JP, Danner RL, Chen H-C, biol Antimicrob. 2020;19(1):21. https://doi.org/10.1186/ Fram D, Zhang F, Wang R, Klompas M, CDC Prevention riginal s12941-020-00363-1. Epicenters Program. Prevalence of Antibiotic-Resistant O 27. Stochino C, Villa S, Zucchi P, Parravicini P, Gori Pathogens in Culture-Proven Sepsis and Outcomes Asso- A, Raviglione MC. Clinical characteristics of CO- ciated With Inadequate and Broad-Spectrum Empiric An- VID-19 and active tuberculosis co-infection in an tibiotic Use. JAMA Network Open. 2020;3(4):e202899. Italian reference hospital. Eur Respir J. 2020. doi: doi: 10.1001/jamanetworkopen.2020.2899 10.1183/13993003.01708-2020 31. Huttner BD, Catho G, Pano-Pardo JR, Pulcini C, 28. Tadolini M, Codecasa LR, García-García JM, Blanc FX, Schouten J. COVID-19: don’t neglect antimicrobial stew- Borisov S, Alffenaar JW, Andréjak C, Bachez P, Bart PA, ardship principles! Clin Microbiol Infect. 2020;26(7):808- Belilovski E, Cardoso-Landivar J, Centis R, D’Ambrosio 10. doi: 10.1016/j.cmi.2020.04.024 Maria EmanuelaBuggio veneto.it. Address forcorrespondence: ElenaMarcon,Ospedale diSchiavonia,Via Albere 30,35043Monselice,Padua,Italy. E-mail:elena.marcon@aulss6. unit (ULSS)6Euganeahealthcare authorityarea[4]. cases, andofthese2368(25%) wereinthelocalhealth on 31stMarch,therewere9625 confirmedCOVID-19 intensive careunit(ICU)beds [1-3].InVeneto Region, increased hospitalizationrateandgrowingdemandof NHS, intermsofremodelingservices,copingwiththe Such diseaserepresentedachallengetotheItalian the EuropeanUnionbySARS-CoV-2 pandemic. INTRODUCTION during theemergencysituation. but itallowedtoconsolidateanorganizationalmodelwhich couldanswertohealthneeds the pointofviewMedicalDirectionHospital challengeshadbeenmany Conclusions. SHchangeditsorganizationthreetimesinlessthantwomonths.From the surroundingareas. as thelargestCOVIDHospitalinVeneto, catchingexclusivelyCOVID-19patientsfrom tal contamination.Thelastperiodwascharacterizedbythere-organizationoffacility completely separatedareas,namedCOVID-19andCOVID-free,topreventintra-hospi- pital isolation.Inthesecondperiodhospitalreopenedanditwasdividedintotwo Results. Thefirstperiod,afterinitialcases’identification,wascharacterizedbythehos- munication, surveillanceonHCW). organizational modificationsareanalyzed(hospitalactivities,logisticalchanges,com- Material andmethods.Threeperiodscanbeidentified;ineachonethemostimportant the pandemic. scribe howSHwasabletoadjustitsservicescopingwiththeepidemiologicalstagesof providing optimalcaretothepatientsandstaffsafety. Theaimofthisarticleistode- building, managementofspaces,humanresourcesandsupplies,inordertocontinue infection, SHhadtorearrangetheclinicalservicesintermsofstructuralchanges of COVID-19 in Veneto Region. As a result of the underlying concomitant spread of Introduction. On21February2020,SchiavoniaHospital(SH)detectedthefirst2cases Abstract 5 4 3 2 Elena Marcon a firstsingle-centerexperience to COVID-19outbreak: Schiavonia Hospitalresponse DOI: 10.4415/ANN_20_03_15 Ann IstSuperSanità2020|Vol. 56,No.3:365-372 1 Italy Padua, Italy Italy Direzione Sanitaria,AziendaULSSn.6Euganea,Padua,Italy Direzione MedicadiPresidio,OspedalePioveSacco,AziendaULSSn.6Euganea,Padua,Italy Servizio ProfessioniSanitarie,OspedalediSchiavonia,AziendaULSSn.6Euganea,Monselice,Padua, Dipartimento diScienzeCardio-Toraco-Vascolari eSanitàPubblica,UniversitàdegliStudidiPadova, Direzione MedicadiPresidio,OspedaleSchiavonia,AziendaULSSn.6Euganea,Monselice,Padua, The ULSS6Euganeaisone ofthemostpopulous Italy hasbeenoneofthemostimpactedcountryin 1 , FrancescaScotton 3 , ClaudioPilerci 2 , ElenaMarcante 4 , DomenicoMontemurro approximately 180000inhabitants. trict. ThisHealthDistrictcovers 46municipalitiesfor mately 3300inhabitants,in the SouthernPaduaDis- Veneto occurred inVo’ Euganeo,atownofapproxi- Hospital operate. two SubacuteHospitalNodesandoneHubUniversity of fivedistricts.InthiscontextfourNetworkHospitals, 2019, itcounted931582inhabitants.Itiscomposed covering anareaofover2127sqkm.On1January consistingof101municipalities ULSS oftheRegion The firstclusteroftheSARS-CoV-2 epidemic in 2 , AlbertoRigo

1 1

, JacopoMonticelli andPatriziaBenini • • • • Key words safety hospital management Italy COVID-19 outbreak 1 , 5 365 Original articles and reviews 366 Elena Marcon, Francesca Scotton, Elena Marcante et al.

The Schiavonia Hospital is the major Hospital Presid- COVID-19 Hospital because it is newly built, organized ium of the Southern Padua District. It was built in 2014 in separate modules and with technologies and systems and it has been carried out in project financing. In the suitable for supporting high volumes of medical gas same area there are also two Subacute Hospital Nodes (oxygen). This involved important structural and orga- principally dedicated to post-acute care. The Clinical nizational changes also because before the COVID-19 Laboratory of Schiavonia Hospital is the microbiology emergency Schiavonia Hospital was not a hospital for reference center for the whole ULSS 6. treatment of infectious diseases. On 20 February 2020, the Hospital Infectious Dis- On 21 February 2020 three Task Forces of different reviews

eases specialist identified two patients, both residing decisional levels were set up: in Vo’ Euganeo and admitted (on 16 and 19 February, • a Regional Task Force (RTF), composed by the Gov-

and respectively) with fever, type 1 respiratory failure and ernment of Veneto Region, a regional scientific com-

interstitial pneumonia, preceded by mild diarrhea. The mittee and Civil Protection that coordinated the re- patients were tested for common pneumotropic patho- gional emergency; gens (common bacterial respiratory pathogens, influ- • an ULSS 6 Emergency Task Force (ETF), formed enza A/B viruses, parainfluenza 1-2-3-4 viruses, RSV, by ULSS 6 General Manager, Chief Medical Officer articles adenovirus, coronaviruses 229E-HKU1-OC43-NL63, and local health authorities that coordinated both the human metapneumovirus, rhinovirus/enterovirus, Bor- hospital response and the Prevention Department’s detella pertussis, Chlamydophila pneumoniae, Mycoplasma activities; pneumoniae, Legionella spp.), for less frequent patho- • an in-hospital Crisis Unit (HCU), composed by the riginal gens (acid-fast bacteria, HSV, VZV, CMV, Pneumocys- Director and all staff of the Hospital Medical Direc- O tis jiroveci, Leptospira spp., Coxiella spp.) and fungal tion, the Directors of Medical, Surgical and Diagnos- biomarkers (galactomannan, serum (1,3)-β-D-glucan). tic Departments, the Occupational Physician, the All tests resulted inconclusive. After a careful medical Director of Healthcare Professions and the Project history, it appeared that the two patients used to at- Financing Manager. tend the same local bar frequented by Chinese people. In the next paragraphs each period is described in de- Although patients had no history of recent travel to tails, analyzing the following items: endemic areas, it was considered appropriate to pro- • short description of beds disposition; ceed with performing nasopharyngeal swabs for SARS- • hospital activities; CoV-2. On the next day, Schiavonia Hospital confirmed • logistical changes; the first 2 cases of COVID-19 in Veneto. • communication; This article aims to describe the measures taken by • surveillance on HCW. Schiavonia Hospital to manage the current evolving challenges, including identification of risks, strategies RESULTS to prevent the transmission of COVID-19 in health- First period: Veneto outbreak and hospital’s isolation care setting, maintaining an efficient response for all (21 February - 7 March 2020) patients with urgent conditions and an adequate staff Immediately after the identification of the first cases support [5-8]. the RTF ordered the isolation of Vo’ Euganeo town. The ETF, according to the decision of the Italian MATERIAL AND METHODS Ministry of Health, ordered the closure of the hospital Since 21st February three different periods can be facility giving indications on the measures that had to identified. be taken to manage the emergency [11]: The first period (21 February - 7 March 2020) started • nobody was allowed to enter or leave the facility; with the identification of the first two cases at Schiavo- • as a precautionary measure, access to the wards was nia Hospital and of the outbreak in Vò Euganeo town. prohibited for non-healthcare workers; In this phase the hospital was completely closed to out- • nasopharyngeal swabs for SARS-CoV-2 were per- patient, except for saving-life therapies. formed to all people present in the hospital at the The second period (8 - 15 March 2020) was charac- time of closure, giving priority to patients’ relatives to terized by the hospital reopening. This could happen allow them to leave the facility in presence of a nega- because all the nasopharyngeal swabs for SARS-CoV-2 tive result. The HCU decided to test outpatients, fol- carried out in health-care workers (HCW) tested nega- lowed by HCW and then hospitalized patients; tive, the trend of the epidemic was not yet known and • swabs to inpatients were scheduled for the following external stakeholders and local mayors made pression day; for a return to routine activities. The facility was di- • Personal Protective Equipment (PPE) were provided vided into two treatment areas, named COVID-19 and to everybody; COVID-free, according to the regional programming, • a cordon sanitaire was established to secure the area, which identified Schiavonia Hospital as the only ULSS keep order and speed up swabs transport to the Vene- 6 facility that had to reserve an area to treat COVID-19 to reference laboratory, located in Padua; patients [9]. In this phase it was also established the • six special field tents were arranged outside the hospi- national lockdown following the surge of the pandemic. tal, to be used in case of need. In the third period (16 March 2020 - 3 May 2020) At the time of closure there were 47 patients in the Schiavonia Hospital was identified as a completely Emergency Room: 15 with triage yellow code, 9 with dedicated COVID-19 Hospital [10]. It was designated triage green code and 23 with triage white code. The 367 How hospital faced COVID-19 emergency situation

Director of the Emergency Room invited patients in the Hospital activities waiting area, not yet taken in charge, to move towards Although the hospital was closed, HCU and ETF as- other ULSS 6 hospital facilities in order to be evaluated sessed the clinical activities that had to remain guar- and taken in charge as soon as possible. anteed: clinical laboratory and intra-hospital surgical The HCU immediately established a quick form to emergencies for inpatients, dialysis and chemotherapy, screen all present people, in particular revealing presence/ as life-saving therapies, both for inpatients and outpa- absence of respiratory symptoms and body temperature. tients. Over 600 swabs were performed during the first Dialysis and chemotherapy services could remain reviews night. The next day the total amount rose to 1231 swabs open for outpatients because the areas dedicated to (Table 1). Laboratory capacity reached saturation very these activities had independent entrance gates.

early, therefore, in order to have the response that al- According to ETF priority lists, surgery and outpa- and lowed the subject to leave the hospital, the waiting time tient clinics had to be re-evaluated in order to define progressively increased to on average of 18 h. urgent cases that needed to be moved to other ULSS 6 No HCW tested positive, while two other patients hospitals or to outpatient settings. had positive swabs (also these patients came from Vo’ The ER was closed, so the Emergency Service of articles

Euganeo). Transports, coordinated at provincial level, diverted all The total number of people resident in Vo’ Euganeo requests to other hospitals of ULSS 6 (Table 2). at the moment of the closure of the hospital was 16 (in- cluding relatives, patients and HCW), of whom 6 tested Logistical changes positive: 4 patients and 2 relatives of one of them. The These events required a rapid reorganization of the riginal four patients were hospitalized in three different Units: hospital as a whole. O internal medicine, ICU and orthopedics; the one re- Access control was essential to reduce the risk of viral covered in ICU died on the night of February 21. His transmission to other patients or HCW. For this reason, death was the first one in Italy due to COVID-19; none specific active gates were created: of the positive patients had significant comorbidities. • gates for the entrance of HCW and other hospital staff: The three positive patients were transferred to Pad- entry times were scheduled to avoid overcrowding ua University Hospital, where the Infectious Diseases (6.30-8.00 for HCW, 8.00-8.30 for administrative Unit was ready for the isolation of confirmed cases. staff): moreover, relatives could give personal belong- The HCU therefore defined four “transit wards” ings to dedicated staff who provided to deliver them where COVID-19 patients had stayed: ER, internal to hospitalized patients; medicine, ICU and orthopedics (Figure 1). The ETF • gate for dialysis and oncology: independent access decided that patients admitted in non-transit wards where outpatients were identified and registered; who tested negative, could be discharged, instead those • gate for pharmaceutical assistance: independent access admitted in transit wards had to remain quarantined in through which medical supplies were distributed on hospital even if dischargeable. appointment to outpatients. At each gate: people had to fill in a form with their Surveillance on HCW personal details; body temperature was measured us- In the early phase of the outbreak, the situation was ing infrared thermometers to recognize any person with rapidly evolving so hospital guidelines were modified fever; surgical masks were distributed to protect against frequently. Due to excessive increasing of laboratory droplet transmission. response times it was established that: Relatives’ visits were suspended. In exceptional cases • all HCW entering shifts had to undergo nasopharyn- (end-of-life stage) only one person was allowed to enter, geal swabs for SARS-CoV-2; wearing proper PPE, after Medical Direction authori- • HCW could leave the hospital without having to wait zation. for swab result if asymptomatic at the end of the shift; As the HCU ordered the closure of the hospital can- • if the swab tested negative HCW could continue nor- teen and inner bar to avoid crowding, the catering was mal work shifts; reorganized through distribution of meals in disposable • if the swab tested positive or they were symptomatic trays both for HCW and inpatients, maintaining the they had to stay home, quarantined; possibility to order differentiated meals according to • all HCW were provided with appropriate PPE. patients’ nutritional needs.

Table 1 Total number of nasopharyngeal swabs for SARS-CoV-2 performed during hospital closure and tested positive Profile N swabs Positive test Healthcare workers 695 0 Administrative staff 46 0 Patients 344 4 Others 146 2 Total 1231 6 368 Elena Marcon, Francesca Scotton, Elena Marcante et al.

PERIOD 1

General Nephrology Geriatrics Medicine 4 25 ward beds 13 General Medicine 48 ward beds Neurology General Medicine Level 2 22 ward beds 25 ward beds

reviews C1 C2 C3

and Multi-specialist Surgery Week Surgery

26 ward beds 25 ward beds Orthopedics 43 ward beds General Surgery Day Surgery Level 1 25 ward beds 8 ward beds C1 C2 C3 articles

Intensive Care Unit Coronary Care Unit Surgical Unit 12 ward beds 10 ward beds riginal O PERIOD 2

Geriatrics 28 ward beds General Medicine 50 ward beds Neurology Level 2 22 ward beds C1 C2 C3

Multi-specialist Surgery 26 ward beds Orthopedics COVID 1 Sub-intensive area 43 ward beds General Surgery 52 ward beds Level 1 25 ward beds C1 C2 C3

Intensive Care Unit Coronary Care Unit Surgical Unit 14 ward beds 10 ward beds

PERIOD 3

COVID 4 COVID 3 COVID 2 50 ward beds 50 ward beds 50 ward beds Next activation Level 2

C1 C2 C3

COVID 5 COVID 1 50 ward beds Sub-intensive area Next activation 52 ward beds Level 1

C1 C2 C3

Intensive Care Unit Intensive Care Unit B Intensive Care Unit C 14 ward beds 10 ward beds 14 + 12 ward beds

Figure 1 Changes in the arrangement of beds in the different Periods of pandemic; in Period 1 transit wards are highlighted with thick edge. 369 How hospital faced COVID-19 emergency situation

Communication pital was completely sanitized following WHO proce- • external communication: Communication took place dures. through press releases and updated posts on social Hospital activities media (Facebook, Instagram) to keep the population The hospital was divided into two separated areas: informed about the rapidly changing situation; one dedicated to patients without infection and one to • internal communication: The ETF communicated with COVID-19 patients. COVID-19 area was activated in the hospital by registered letters. The HCU inter- a modular way based on the availability of HCW who faced with the Department Directors who commu- had been moved from other hospital sectors that had reviews

nicated the decisions taken to the respective Units. been closed. The COVID-free area was organized in order to guar-

Second period: hospital re-opening antee (Figure 1, Table 2): and

(8 - 15 March 2020) • 100 beds for internal medicine (including activities Before the re-opening, ordered by the ETF, the hos- of neurology, cardiology, nephrology and geriatrics); articles Table 2 Reallocation of hospital activities during COVID-19 emergency

Hospital Period 1 Period 2 Period 3

activities riginal Guaranteed Relocated Relocated Guaranteed Relocated Relocated Guaranteed Relocated Relocated to to other to to other to to other O subacute ULSS6 subacute ULSS6 subacute ULSS6 hospital hospitals hospital hospitals hospital hospitals nodes nodes nodes Clinical X X* X X* X X* Laboratory Dialysis X X X Oncology- day X X X hospital Radiotherapy X X X Radiology X*** X* X X*** X* Emergency X X X** Room Time X X X dependent pathologies (hemodynamic, stroke-trauma network, birth points) Specialist X X X X outpatient visit Ordinary hospitalization for acute patients: Medical Area X X X (Covid free) Surgical Area X X X (Covid free) Psychiatry X X X Day surgery X X X Outpatient X X X surgery Digestive X X X Endoscopy Transfusion X X X activities Pharmaceutical X X X assistance * for outpatients. ** downgraded to Firt Aid Point. *** for inpatients. 370 Elena Marcon, Francesca Scotton, Elena Marcante et al.

• 94 beds for surgical patients (orthopedics, general • COVID-19 Area: logistical aspects and training of HCW surgery, ENT, urology and gynecology) and 18 beds This area was completely separated from the others, of obstetrics; even with specially built-in walls. Both accesses for am- • 8 beds for pediatrics; bulance and for HCW were completely dedicated and • outpatient services (oncology, psychiatry, radiothera- no access of external people was allowed. py, radiology, outpatient surgery, digestive endoscopy, Inside each ward a clean path and a dirty path were clinical laboratory). established, with specific filter rooms where HCW Moreover, the ER, including gynecological and pedi- could wear and take off their PPE. reviews

atric ERs, were fully reactivated. All HCW were trained on correct PPE utilization The COVID-19 area was composed of: with a course of about 8 hours, organized by the Oc-

and • an Intensive Care Unit that increased from 12 to 14 cupational Physician. Moreover, Internal Medicine and

beds; Pneumology Directors organized seminars to focus on • a sub-intensive area of 52 beds of which 8 dedicated some typical aspects of the pathology and management to patients who needed non-invasive ventilation and of non-invasive-ventilation. 44 beds, comparable to an infectious disease unit, Communication articles

managed by a multi-professional team composed by • external communication: in addition to the first period internists, geriatricians, anesthesiologists and infec- people were encouraged to use telematic booking tious disease specialists. systems in order to avoid crowding and maintaining physical distancing measures. riginal Logistical changes • internal communication: as in period 1. O • Active gates Emergency room: it was separated, logistically and Third period: becoming a COVID-19 hospital physically, in two parallel and non-intersecting paths, (16 March 2020 - 3 May 2020) one for patients with suspect COVID-19 (called R1 - Hospital activities Risk 1) and one for patients considered not infected According to RTF, the ETF, ordered the suspension (R2 - Risk 2). of all activities for outpatients at Schiavonia Hospital The distribution of patients to R1 or R2 was due to with the exception of dialysis, psychiatric services and a score determined by a specific pre-triage managed by radiotherapy as these units were located in different a nurse who evaluated the presence of fever, respira- buildings and could guarantee an independent access tory symptoms and an epidemiological link to a positive to outpatients (Table 2). person. The RTF identified the total number of beds that had The pre-triage, for people who came to the emergen- be guaranteed in the newly-formed COVID Hospital. cy room autonomously, was performed in a field tent The ETF and HCU decided how to activate them (Fig- installed near the ER entrance. The subsequent hospital ure 1): admission continued to be performed separated for sus- • 50 beds of intensive care unit divided in 3 sectors: pected or confirmed positive patients and for patients ICU A of 14 beds (previously ICU of 12 beds), ICU considered non-infected in order to avoid as much as B of 10 beds (previously coronary care unit) and ICU possible the contact between the two groups and re- C of 26 beds (previously operating theatre and recov- duce the length of stay in an often overcrowded area. ery room); Specific gates for oncology and dialysis • 52 beds dedicated to patients who needed non-inva- Gates for the entrance of hospital staff: two different sive ventilation (COVID 1); entrances for HCW were established in order to avoid • 200 beds of Infectious Disease, distributed in 4 units bottlenecks. Unlike the previous organization, the sani- of 50 beds (COVID 2-5). tary control (temperature and presence of respiratory All hospitalized patients were COVID-19 positive. symptoms) was done at each unit entrance where PPE Also, patients who tested positive in other ULSS 6 ERs were also delivered. were admitted to Schiavonia Hospital which became Gate for outpatients or visitors: at the main entrance, the biggest Covid Hospital in Veneto Region. dedicated staff controlled the access of people in order It is important to underline that the activation of ICU to avoid overcrowding. Their body temperature was mea- and Infectious Disease units occurred in a modular way, sured and specific informative leaflets explaining the ap- on the basis of the availability of human resources and propriate hygienic measures to adopt were distributed. technological devices. HCU decided to implement wall signpost and floor For this reason, the ETF defined an internal HCW signage to facilitate following established routes. redistribution plan, both for physicians and nurses, • Management of waiting rooms and of access for visitors which involved all the hospitals of ULSS 6 in order to The following precautionary measures were adopted: meet Schiavonia Hospital’s increased need of human invite people to hand hygiene with hydro-alcoholic so- resources. lution; arrange seats to guarantee the safety distance On 7 April 2020 the total number of activated beds of at least 1 meter; prohibition of entry for care-givers, comparing to the ETF programmation was the follow- except those for of minor or disabled patients; limiting ing: 34 beds out of 50 of ICU, 52 of semi-intensive care the visiting time to the wards to two hours a day for and 100, out of 200 of infectious disease. only one person per patient. No entry was allowed in The HCU, according to ETF, proceeded to draw up the COVID-19 area. the plan for the reallocation of activities for outpatients 371 How hospital faced COVID-19 emergency situation

and the reorganization of the emergency and urgent ing constant surveillance. A field tent was set up near care network, with particular regard to time-dependent the ER entrance to test HCW considered high risk con- pathologies. tacts of a positive co-worker. In addition: On 7 April 2020, 22 out of 695 HCW (3.2%) had a • ER was downgraded to First Aid Point maintaining positive nasopharyngeal swab (18 nurses and 4 medical the separation into two different areas (R1 and R2).; doctors). • the Clinical Laboratory became the only one in the A total of 10 out of 22 HCW were operating in the ULSS 6 authorized by the National Government to sub-intensive area (COVID 1) and specifically in the reviews

process and store nasopharyngeal swabs for SARS- sector dedicated to non-invasive ventilation. The use of CoV-2; non-invasive mechanical ventilation can produce aero-

• twenty-five special bio-containment chambers were sol; therefore, the level of protection was increased to and

set up inside the hospital rooms in order to overcome prevent further transmission in this high risk ward. All the lack of negative pressure rooms. These chambers, HCW were provided with waterproof overalls, filtering normally used in the military field, assure high air face pieces (FFP3), face shields, gloves and boots. change volumes and prevent external air contamina- The Risk Manager identified 3 different hazard ar- articles

tion each time the door is opened, avoiding the en- eas: high, medium and low risk. According to this clas- vironmental transmission of potentially contagious sification and to the professional profile of each HCW, pathogens. HCU was able to provide weekly an adequate number Medical assistance in COVID 2-3 Units was provided of PPE’s kits. by physicians regardless of specialty, e.g. also surgeons riginal were assigned to the COVID-19 areas principally for Communication O back-office activities (e.g. to compile medical records, • External communication to contact patient’s relatives etc.). In comparison to previous periods the HCU decided According to their clinical conditions COVID-19 to approve two new projects: patients could be distinguished into different severity - tablets, configured for video-calling, were given to hos- classes, based on the respiratory impairment level and pitalized patients in COVID area to facilitate commu- chest X-ray findings, which allowed to identify the ap- nication with relatives; propriate care setting. - activation of a dedicated telephone number to keep During the observation period 16 March - 7 April relatives in touch with physicians. It was available ev- 2020, 329 hospitalizations (of which 39 in ICU), 5 ery day from 9am to 6pm. An operator collected re- transfers from the ICU to sub-intensive unit and 52 quests from patients’ family and transmitted them to deaths were registered; 142 patients were discharged the medical staff in order to answer as soon as pos- with an average of 6 discharges /day, no re-admission sible. was reported. • Internal communication In addition to period 1, a psychological support ser- Logistical changes vice for to hospitalized patients and HCW was acti- Compared with the second period no modifications vated. in active gates was done, except for the visitors’ gate HCW faced enormous pressure, the emergency situ- that was closed as the access to hospital was prohibited. ation caused psychological problems such as stress, The filter zones inside COVID-19 wards was en- anxiety, sleep deprivation and fear. A psychological hanced: dedicated staff was always present to carry out intervention plan was developed by a team of psychia- the sanitary control and deliver PPE in order to reduce trists and psychologists of the hospital’s Mental Health waste and improve efficiency. Pre-established PPE kits Unit for both patients and HCW: were made available according to the biological risk ex- - video call interview for hospitalized patients who posure of each HCW. needed psychological support for situations of stress The Medical Direction of the Hospital evaluated dai- related to their particular situation; ly the available stocks. - video call or in person visits for HCW who needed All hospital waste were considered infected so waste advices and support to deal with acute work-related separation was suspended and the service of waste col- discomfort. lection was enhanced. Also medical records were considered contaminated: CONCLUSIONS the indication was to insert them in a double plastic bag In less than two months Schiavonia Hospital changed and quarantine them in a designated area for at least its organization three times, according to regional health 14 days. programming. From the point of view of the Medical Corpses were managed according to the rules for Direction of the Hospital the challenges had been many corpses of patients with highly diffusive disease. and major: first of all the need to respond adequately to an health emergency and to a high number of patients, Surveillance on HCW and, not less important, the need to guarantee adequate The exposure and possible infection of HCW posed a logistical, technological, administrative and personnel massive challenge to the delivery of medical service. In support in order to allow all HCW to work safely. response to such situation, strategies to contain the in- An important aspect was the provision of specific hospital COVID-19 spreading were intensified ensur- medical technologies (eg. medical ventilators, devices 372 Elena Marcon, Francesca Scotton, Elena Marcante et al.

for non-invasive ventilation, High-Flow Oxygen, C- Urgent care was guaranteed in separated settings for PAP masks, etc.) and PPE. There is risk of a shortage COVID-19 positive and negative patients, moreover all of protective equipment during a pandemic, especially scheduled activities for COVID-19 negative patients if prolonged. Measures adopted to avoid this and an were carried out routinely. According to the regional re- inappropriate use of PPE included administrative con- quest, an Emergency Plan has been defined to provide trols and risk mapping related to HCW exposure to a further organizational model able to respond rapidly infection. Another improvement action to avoid inap- to a possible second wave of the epidemics. propriate use and waste was obtained by individual dis- reviews

tribution of pre-packed sets of PPE in filter areas before Authors’ contribution HCW started their shift. The study was designed by EM, FS, EM and DM

and The most critical aspect was human resources avail- with input from all coauthors. The manuscript was

ability. To overcome this a lot of not urgent services drafted by EM, FS, EM; PB, CP and AR provided criti- both in Schiavonia Hospital and in other facilities were cal revision for intellectual content and input in writing. closed to reallocate HCW to COVID-19 units, accord- JM and RV participated editing the manuscript. ing to ULSS 6 staff redistribution plan. Thanks to this articles we were able to open COVID-19 areas in a modular Acknowledgements way (about 20 beds for week) according also to the The Authors are grateful to Risk Manager Arseni A, epidemiological phase of the pandemic. To maximize Occupational Physician Fioretto M, Project Financing the utilization of medical resources many doctors were Manager Melinu G, Healthcare Professions Coordina- riginal asked to practice outside their area of specialty or ex- tors Spanò A, Sturaro W, Stanziale S, Alfonso M for O pertise and, differentiating tasks within the multi-pro- their assistance, Cesaro M for linguistic revision. fessional team, efficiency and safety were maintained as high as possible. Conflict of interest statement Since 4 May the hospital developed a process of grad- None. ual reopening, starting from outpatient services: clini- cal laboratory, ambulatory surgery unit and specialistic Ethics approval ambulatories. Not required. Since 8 June also surgical and medical wards were completely operative even though with fewer hospital Received on 9 May 2020. beds than in the pre-epidemic period. Accepted on 13 July 2020.

REFERENCES

1. European Centre for Disease Prevention and Control pital organization in response to a changing epidemic. (ECDC). Coronavirus disease 2019. Available from: Emerg Care J. 2020;16:8969. www.ecdc.europa.eu/en/covid-19-pandemic. 6. Wang C, et al. A novel coronavirus outbreak of global 2. World Health Organization (WHO). Coronavirus disease health concern. Lancet. 202015;395(10223):470-3. doi: (COVID-2019) situation reports. Situation report – 71. 10.1016/S0140-6736(20)30185-9 Available from: www.who.int/docs/default-source/coro- 7. Del Rio, Malani PN. Covid-19 - New insights on a rapid- naviruse/situation-reports/20200331-sitrep-71-covid-19. ly changing epidemic. JAMA. 2020 Feb 28. doi: 10.1001/ pdf?sfvrsn=4360e92b_8. jama.2020.3072 3. World Health Organization (WHO). Statement on the 8. Liu JW et al. Epidemiologic study and containment of second meeting of the International Health Regula- nosocomial outbreak of severe acute respiratory syn- tions Emergency Committee regarding the outbreak of drome in a medical center in Kaohsiung, Taiwan. Infect novel coronavirus (2019-nCoV). Available from: www. Control Hosp Epidemiol. 2006;27(5):466-72. who.int/news-room/detail/23-01-2020-statement-on-the- 9. Italia. Regione Veneto. Area Sanità e Sociale, nota regio- meeting-of-the-international-health-regulations-(2005)- nale, 6 Marzo 2020. emergency-committee-regarding-the-outbreak-of-novel- 10. Italia. Regione Veneto. Area Sanità e Sociale, nota regio- coronavirus-(2019-ncov). nale, 15 Marzo 2020. 4. Istituto Superiore di Sanità (ISS). Integrated surveillance 11. Italia. Regione Veneto. Comunicato n. 278. Coronavirus: of COVID-19 in Italy. Available from: www.epicentro.iss. ordinanza del Ministro Speranza d’intesa con Governa- it/coronavirus/sars-cov-2-sorveglianza-dati. tore del Veneto per comuni e aree colpite. Available from: 5. Coen D, et al. Changing emergency department and hos- www.regione.veneto.it/article-detail?articleId=4281001. Chiara Donfrancesco Sanità, Via Gianodella Bella34,00162Rome,Italy. E-mail:[email protected]. Address forcorrespondence: MarcoCanevelli,Centro NazionaleperlaPrevenzionedelle MalattieelaPromozionedellaSalute, IstitutoSuperioredi (n =2,687)betweenFebruary21 the medicalchartsofarepresentativesamplepatientsdeceasedinItalianhospitals acteristics of natives and migrants dying with COVID-19 were explored by considering with COVID-19bynative/migrantstatus.Themortalityratesanddetailedclinicalchar and migrantsinItalytoinvestigatetheclinicalcharacteristicsofindividualsdying We aimed to compare COVID-19-specific and all-cause mortality rates among natives Abstract 6 5 4 3 2 1 Marco Canevelli living inItaly COVID-19 mortalityamongmigrants DOI: 10.4415/ANN_20_03_16 Ann IstSuperSanità2020|Vol. 56,No.3:373-377 the References *The membersoftheItalianNationalInstituteHealthCOVID-19MortalityGrouparelistedbefore di Sanità,Rome,Italy di Sanità,Rome,Italy behaviors, socialnetworks,and accessibilitytohealth- [1]. Finally, differentages,lifestyles,health-related origin whentheyareinpoor healthorbeforedeath of migrantsgenerallyreturning totheircountriesof [1]. The“salmonbias”reflects thecommonbehavior likely todecidemigrateandbenefitfrommigration tion ofhealthyandrobustindividualsthataremore The “healthymigrant”hypothesisconsistsoftheselec- miological patterncanbeexplainedbyseveralfactors. countries [1,2].Thisapparentlyparadoxicalepide- ity relativetonativepopulationsinhigh-incomehost have amortalityadvantage,definedaslowermortal- their countryofbirth)haverepeatedlybeenshownto INTRODUCTION native counterparts. mortality advantageforCOVID-19andareexposedtotheriskofpooroutcomesastheir except fortheyoungerageatdeath.InternationalmigrantslivinginItalydonothavea The clinicalphenotypeofmigrantsdyingwithCOVID-19wassimilartothatnatives tive all-causemortalityratesestimatedinItaly2018(97.4%and2.6%,respectively). the COVID-19-relateddeaths(97.5%and2.5%,respectively)weresimilartorela- migration backgroundwereidentified.Theproportionsofnativesandmigrantsamong set providedbytheItalianNationalInstituteofStatistics.Overall,68individualswitha among nativesandmigrantslivinginItalywasderivedbythelastavailable(2018)data- was assignedbasedontheindividual’s countryofbirth.Theexpectedall-causemortality Graziano Onder President, IstitutoSuperiorediSanità,Rome,Italy Dipartimento diMalattieInfettive,IstitutoSuperioreSanità,Rome,Italy Dipartimento diMedicinaSperimentale,SapienzaUniversitàRoma,Rome,Italy Dipartimento MalattieCardiovascolari,Endocrino-MetabolicheeInvecchiamento,IstitutoSuperiore Dipartimento diNeuroscienzeUmane,SapienzaUniversitàRoma,Rome,Italy Centro NazionaleperlaPrevenzionedelleMalattieePromozionedellaSalute,IstitutoSuperiore International migrants(i.e.,peoplelivingoutside 3 1,2 andtheItalianNationalInstituteofHealthCOVID-19MortalityGroup* , LuigiPalmieri 3 , CinziaLoNoce st andApril29 3 , Valeria Raparelli 3 , NicolaVanacore th , 2020.Themigrantornativestatus with COVID-19inItalyascompared withnatives. ing theclinicalcharacteristics ofmigrantsdeceased the expected mortality rate for migrants and at assess- current rateofCOVID-19-related deathsdiffersfrom international migrantsareestimatedworldwide[5]. important publichealthimplicationssince272million national migrantpopulations.Addressingthisissuehas impact oftheongoingCOVID-19pandemicininter date, nostudieshavespecificallyexploredthemortality eases, doesnotapplytoinfectiousdiseases[1,4].To observed forthemajorityofnon-communicabledis- factors [3]. care resourceshavebeenproposedascontributing The presentstudyaimedat evaluating whetherthe Nevertheless, themortalityadvantageinmigrants, 4 , OrnellaPunzo 1 , SilvioBrusaferro 5 , - • • • • Key words global health SARS-CoV-2 migration COVID-19 6 ,

- 373 Original articles and reviews 374 Marco Canevelli, Luigi Palmieri, Valeria Raparelli et al.

METHODS and Oceania); ii) European countries with High Migra- At the outset of the COVID-19 outbreak, the Ital- tory Pressure (Central and Eastern Europe); iii) Non- ian National Institute of Health (Istituto Superiore di European countries with High Migratory Pressure (Af- Sanità, ISS) launched an integrated national surveil- rica, Asia, Central and South America). lance system to collect information on all individuals The Kolmogorov-Smirnov and Shapiro-Wilk tests with COVID-19 throughout the country [6]. All deaths were used to verify the normal distribution of con- occurring in patients with confirmed COVID-19 were tinuous variables. The characteristics of natives and tracked by the surveillance system. COVID-19-related migrants were compared by means of ANOVA analy- reviews

deaths were defined as those occurring in patients who ses for normally distributed variables, non-parametric tested positive for SARS-CoV-2 through RT-PCR, in- Mann–Whitney U-test for skewed variables, and Fish-

and dependently from pre-existing diseases that may have er’s exact test for categorical variables.

caused or contributed to death. This study was performed in line with the principles In the present analysis, mortality rates and the de- of the Declaration of Helsinki. On February 27th, 2020, tailed clinical characteristics of natives and migrants the Italian Presidency of the Council of Ministers au- dying with COVID-19 were explored by considering thorized the collection and scientific dissemination of articles the medical charts of a sample of patients deceased data concerning the COVID-19 epidemics by the ISS in Italian hospitals between February 21st and April and other public health institutions [8]. 29th, 2020, representative in terms of age, sex, and geographical distribution of overall COVID-19-related RESULTS riginal deaths occurred in the country [7]. Specifically, 2,687 Overall, 68 individuals (39.7% women) with a migra- O medical charts consecutively collected from the Italian tion background were identified in the study popula- regions and autonomous provinces were centrally ana- tion. The most common countries of birth of migrants lyzed. The migrant or native status was assigned based were Albania (n = 15) and France (n = 8) (Table 1). The on the individual’s country of birth [1], as indicated on proportions of natives and migrants among the COV- the chart and/or on the tax/fiscal code. People born ID-19-related deaths (97.5% and 2.5%, respectively) abroad have their four-character code according to were similar to the relative all-cause mortality rates es- the country of birth, all of them beginning with letter timated in Italy in 2018 (97.4% and 2.6%, respectively) Z (e.g., Albania is identified by the code Z100). The (Table 1). Accordingly, migrants’ mortality distribution following data were extracted from the charts: demo- in the three categories of countries of origin did not graphics; comorbidities; symptoms at onset; in-hospi- substantially differ from what was expected based on tal complications; treatments received; admission to the 2018 national data (Table 1). intensive care unit; time from onset to hospitalization, Migrants were younger at the time of death com- SARS-CoV-2 testing, and death; time from hospitaliza- pared to natives (71.1, SD 13.1 years vs 78.3, SD 10.8 tion to death. years; p < 0.001) (Table 2); no statistically significant The expected all-cause mortality among natives and mi- differences were found for age at death according to grants living in Italy was derived by the 2018 data provid- the profile of migrants’ countries of birth (p = 0.30). ed by the Italian National Institute of Statistics (ISTAT) A lower prevalence of ischemic heart disease and hy- (http://dati.istat.it/Index.aspx?DataSetCode=DCIS_ pertension was documented among migrants relative to MORTALITA1#). World countries were categorized natives (17.5% vs 28.7%, p = 0.03, and 55.6% vs 68.5, p according to the ISTAT classification in: i) Highly de- = 0.02, respectively) (Table 2). No other significant dif- veloped countries (European countries except those of ferences were observed between groups for comorbidi- Central-Eastern Europe, North America, Israel, Japan, ties, sex distribution, symptoms at onset, treatments

Table 1 Expected all-cause deaths and observed COVID-19 related deaths in natives and migrants living in Italy All-cause deaths* COVID-19 related deaths n % n % 95% CI Natives 616,729 97.4 2,619 97.5 96.8-98.0 Migrants 16,402 2.6 68 2.5 2.0-3.2 Highly developed countries 5,203 0.8 15 0.6 0.3-0.9 European HMP countries 2,362 0.4 25 0.9 0.6-1.4 Non-European HMP countries 8,837 1.4 28 1.0 0.7-1.5 Total 633,131 100 2,687 100 *2018 ISTAT data (http://dati.istat.it/Index.aspx?DataSetCode=DCIS_MORTALITA1#) Migrants are grouped according to the ISTAT classification of world countries: i) Highly developed countries (European countries except those of Central-Eastern Europe, North America, Israel, Japan, and Oceania); ii) European countries with High Migratory Pressure (Central and Eastern Europe); iii) Non-European countries with High Migratory Pressure (Africa, Asia, Central and South America). Migrants’ countries of birth: Albania (n = 15); France (n = 8); India (n = 4); Libya (n = 3); Romania (n = 3); Switzerland (n = 3); Tunisia (n = 3); Ukraine (n = 3); Ethiopia (n = 2); Morocco (n = 2); Philippines (n = 2); Algeria (n = 1); Argentina (n = 1); Austria (n = 1); Belgium (n = 1); Bulgaria (n = 1); Dominican Republic (n = 1); Ecuador (n = 1); Egypt (n = 1); Germany (n = 1); Ghana (n = 1); Honduras (n = 1); Macedonia (n = 1); Paraguay (n = 1); Peru (n = 1); Russia (n = 1); Serbia-Montenegro (n = 1); Somalia (n = 1); Sudan (n = 1); United States of America (n = 1); Uruguay (n = 1). 375 COVID-19 mortality in migrants

Table 2 Characteristics of individuals deceased with COVID-19 in Italy by migration status Natives Migrants p-value (n = 2,619) (n = 68) Demographics Age 78.3 ± 10.8 71.1 ± 13.1 < 0.001 Female sex 853 (32.6) 27 (39.7) 0.24 reviews

Comorbidities Ischemic heart disease 734 (28.7) 11 (17.5) 0.03 and Atrial fibrillation 570 (22.3) 14 (22.2) 0.57 Heart failure 417 (15.9) 10 (14.7) 0.55 Stroke 271 (10.6) 5 (7.9) 0.33

Hypertension 1,753 (68.5) 35 (55.6) 0.02 articles

Type 2 diabetes 790 (30.9) 24 (38.1) 0.14 Dementia 407 (15.9) 8 (12.7) 0.31 Chronic obstructive pulmonary disease 423 (16.5) 12 (19.0) 0.35 riginal

Active cancer (last 5 years) 408 (15.9) 11 (17.5) 0.43 O Chronic liver disease 101 (3.9) 3 (4.8) 0.46 Chronic renal failure 522 (20.4) 11 (17.5) 0.35 HIV 6 (0.2) 0 (0.0) 0.86 Autoimmune disease 96 (3.8) 5 (7.9) 0.09 Obesity 282 (11.0) 6 (9.5) 0.45 Number of comorbidities 0 98 (3.8) 2 (5.1) 0.79 1 381 (14.9) 4 (10.3) 2 546 (21.3) 9 (23.1) ≥ 3 1,533 (59.9) 24 (61.5) Symptoms Fever 1,928 (76.2) 52 (80.0) 0.56 Dyspnea 1,861 (73.5) 43 (66.2) 0.20 Cough 974 (38.5) 22 (33.8) 0.52 Diarrhea 144 (5.7) 2 (3.1) 0.58 Hemoptysis 15 (0.6) 0 (0.0) 1.00 Treatments Antibiotics 2,138 (85.0) 62 (91.2) 0.22 Antivirals, hydroxychloroquine, chloroquine 1,444 (57.4) 46 (67.6) 0.11 Steroids 924 (36.8) 31 (45.6) 0.16 Clinical course Admission to intensive care unit 487 (20.3) 19 (31.7) 0.22 Time from onset to SARS-CoV-2 testing (days) 5 (3-8) 4 (2-10) 0.99 Time from onset to hospitalization (days) 4 (2-7) 4 (1-9) 0.60 Time from onset to death (days) 10 (7-16) 11 (7-17.5) 0.57 Time from hospitalization to death (days) 5 (3-10) 6 (3-12) 0.36 Data are expressed as mean ± standard deviation, n (%), or median (IQR). Missing data for the overall sample (n = 2,687): comorbidities n = 64 (2.4%); symptoms n = 92 (3.4%); treatments n = 105 (3.9%); clinical course n = 231 (8.6%). received, admission to intensive care units, and times DISCUSSION to clinical milestones (all p > 0.05). Almost all individu- To the best of our knowledge, this study reported for als of both groups (97.0% of natives and 96.9% of mi- the first time data on international migrants deceased grants) developed acute respiratory distress syndrome during the COVID-19 pandemic in Italy. The distribu- as life-threatening complication. tion of COVID-19-related and all-cause mortality in 376 Marco Canevelli, Luigi Palmieri, Valeria Raparelli et al.

migrants and natives in Italy is not substantially differ- pursue ideals of equity and universality in public health ent suggesting that the mortality advantage in migrants actions, communication, and information sharing. does not apply to COVID-19. These data provide in- sights to inform the current debate on Italian media Authors’ contributions and social networks regarding the concerns that mi- Marco Canevelli designed the study, had full access to grants are spared or only marginally affected by SARS- all data in the study, and takes responsibility for the in- CoV-2 infection [9]. tegrity and accuracy of the data analysis. Luigi Palmieri The clinical phenotype of migrants, investigated in was responsible for data analysis. Chiara Donfrancesco reviews

a limited but representative sample of individuals de- and Cinzia Lo Noce were responsible for supervision of ceased in hospital settings, was similar to that of natives data collection and data management.

and except for the younger age at death, that was somehow Marco Canevelli, Luigi Palmieri, Valeria Raparelli,

expected given the different shape of the age pyramids Nicola Vanacore, and Graziano Onder drafted the man- of the two populations [10]. Indeed, only 7.1% of mi- uscript. All Authors (including all members of the Ital- grants living in Italy versus 24.7% of natives are older ian National Institute of Health COVID-19 Mortality than 65 years [11]. Accordingly, all-cause mortality Group) contributed to data collection and discussion. articles

rates over the age of 60 years are markedly lower among All Authors reviewed and approved the final version of foreign-born individuals residing in Italy relative to na- the manuscript. The corresponding Author attests that tives [12]. Nevertheless, the present findings might not all listed authors meet authorship criteria and that no be representative of COVID-19-related outcomes in the others meeting the criteria have been omitted. riginal overall, heterogeneous population of migrants. First, the O countries of origin of the sampled migrant individuals Funding do not perfectly match with those of the broader mi- Authors have no funding source to disclose for the grant population living in Italy, even if they mirror their present study. main attributes (e.g., income and migratory pressure) (Table 1) [5]. Moreover, our estimates most likely ap- Acknowledgements ply to migrants who are well-integrated and have access Marco Canevelli is supported by a research grant of to healthcare resources and services. This hypothesis the Italian Ministry of Health (GR-2016-02364975) for seems corroborated by the similar demographic and the project “Dementia in immigrants and ethnic minor- clinical characteristics of migrants and natives deceased ities living in Italy: clinical-epidemiological aspects and for COVID-19. We cannot exclude that more marginal- public health perspectives” (ImmiDem). ized migrants such as refugees, asylum seekers, and un- documented migrants (who also do not have an Italian Conflicts of interest statement tax code) behave differently in terms of clinical pheno- Authors have no competing interests to disclose for type and COVID-19-related outcomes. These particu- the present study. larly vulnerable individuals, usually underrepresented in mortality analyses [1], might have a higher risk of Received on 24 June 2020. contracting infectious diseases, including COVID-19, Accepted on 15 July 2020. and experiencing poorer outcomes due to poor living conditions, difficulties at adopting restrictive measures, The members of the Italian National Institute of scarcity of protective equipment, limited access to test- Health COVID-19 Mortality Group are: ing procedures and treatments [13]. In this regard, a Luigi Palmieri, Luigi Bertinato, Gianfranco Bram- limitation of the present study is the lack of informa- billa, Giovanni Calcagnini, Marco Canevelli, Federica tion on diverse determinants (e.g., reason for migration, Censi, Elisa Colaizzo, Chiara Donfrancesco, France- duration of the migration, acculturation, socioeconomic sco Facchiano, Marco Floridia, Marina Giuliano, Ti- status, ethnicity) that have already been shown to influ- ziana Grisetti, Yllka Kodra, Martin Langer, Ilaria Lega, ence health outcomes in the variegated migrant popula- Cinzia Lo Noce, Fiorella Malchiodi Albedi, Valerio tion and may also affect the course of COVID-19 [14]. Manno, Eugenio Mattei, Paola Meli, Giada Minelli, Future international, prospective, adequately powered Manuela Nebuloni, Lorenza Nisticò, Marino Nonis, and ad hoc studies are needed to disentangle the role of Graziano Onder, Lucia Palmisano, Nicola Petrosillo, these factors in understanding the ongoing pandemic. Flavia Pricci, Ornella Punzo, Valeria Raparelli, Paolo In conclusion, our study confirms that, with obvious Salerno, Manuela Tamburo De Bella, Domenica Taru- inter-individual variability, natives and migrants are all scio, Dorina Tiple, Brigid Unim, Luana Vaianella, Ni- on the same boat in dealing with the pandemic and rein- cola Vanacore, Monica Vichi, Emanuele Rocco Villani, forces the belief that, even in this contingency, we must Amerigo Zona.

REFERENCES

1. Aldridge RW, Nellums LB, Bartlett S, et al. Global patterns 2. Wallace M, Khlat M, Guillot M. Mortality advantage of mortality in international migrants: a systematic review among migrants according to duration of stay in France, and meta-analysis. Lancet. 2018;392(10164):2553-66. 2004-2014. BMC Public Health. 2019;19(1):327. doi: doi: 10.1016/S0140-6736(18)32781-8 10.1186/s12889-019-6652-1 377 COVID-19 mortality in migrants

3. Reus-Pons M, Vandenheede H, Janssen F, Kibele EUB. nesso all’insorgenza di patologie derivanti da agenti virali Differences in mortality between groups of older mi- trasmissibili. (Ordinanza n. 640). Gazzetta Ufficiale – Se- grants and older non-migrants in Belgium, 2001-09. Eur rie Generale n. 50 del 28 febbraio 2020. Available from: J Public Health. 2016;26(6):992-1000. doi: 10.1093/eur- www.gazzettaufficiale.it/eli/id/2020/02/28/20A01348/ pub/ckw076 SG. 4. Vanthomme K, Vandenheede H. Migrant mortality dif- 9. Meli E. Immigrati e vaccino anti-Tbc. Come stanno dav- ferences in the 2000s in Belgium: interaction with gen- vero le cose. Corriere della Sera. 2020. Available from: der and the role of socioeconomic position. Int J Equity www.corriere.it/salute/malattie-rare/20_marzo_25/coro- Health. 2019;18(1):96. doi: 10.1186/s12939-019-0983-5 navirus-immigrati-vaccino-antitubercolare-come-stanno- reviews

5. United Nations, Department of Economic and Social Af- cose-f13a84ba-6eb2-11ea-925b-a0c3cdbe1130.shtml. fairs, Population Division. International Migration 2019: 10. Migration and migrant population statistics - Statistics

Report (ST/ESA/SER.A/438). UN; 2019. Available from: Explained. Available from: https://ec.europa.eu/eurostat/ and www.un.org/en/development/desa/population/migration/ statistics-explained/index.php?title=Migration_and_mi- publications/migrationreport/docs/InternationalMigra- grant_population_statistics tion2019_Report.pdf. 11. Eurostat. Population (Demography, Migration and Pro- 6. Riccardo F, Ajelli M, Andrianou X, et al. Epide- jections). Available from: https://ec.europa.eu/eurostat/

miological characteristics of COVID-19 cases in It- web/population-demography-migration-projections/data/ articles aly and estimates of the reproductive numbers one database month into the epidemic. medRxiv. 2020 Apr 11. doi: 12. Ministero dell’Interno. 1° Rapporto sugli immigrati in 10.1101/2020.04.08.20056861 Italia. 2007. 7. Onder G, Rezza G, Brusaferro S. Case-fatality rate and 13. Kluge HHP, Jakab Z, Bartovic J, D’Anna V, Severoni S. characteristics of patients dying in relation to COVID-19 Refugee and migrant health in the COVID-19 response. riginal in Italy. JAMA. 2020;323(18):1775-6. doi: 10.1001/jama. Lancet. 2020;395(10232):1237-9. doi: 10.1016/S0140- O 2020.4683 6736(20)30791-1. 8. Italia. Presidenza del Consiglio dei Ministri, Diparti- 14. Pareek M, Bangash MN, Pareek N, et al. Ethnicity and mento della Protezione civile. Ordinanza 27 febbraio COVID-19: an urgent public health research priority. 2020. Ulteriori interventi urgenti di protezione civile in Lancet. 2020;395(10234):1421-2. doi: 10.1016/S0140- relazione all’emergenza relativa al rischio sanitario con- 6736(20)30922-3 378 Original articles and reviews nient forestablishinganational prospectivepopulation- ian ObstetricSurveillanceSystem (ItOSS)[7]isconve- should bebasedonrigorously collected data. TheItal- for obstetricprovidersand neonatologists, andthis UK [4]. 6]. Available data come primarily from China, and the of deliveriesamongwomenaffectedbyCOVID-19[3- eral population[1,2]. able datasuggestthatitisanalogoustoofthegen- women totheSARS-CoV-2 illness,thecurrentlyavail- with regardstoanincreasedsusceptibilityofpregnant comes [1].Althoughthereisnoconclusiveinformation plications aswellworsematernalandneonatalout- women are athigherrisk of developing respiratory com- SARS, andMERSoutbreaks,welearnedthatpregnant number ofclinicalcases.FromtheH1N1influenza, one oftheEuropeancountriesreportinghighest INTRODUCTION to moderateillness. are similartothosedescribedforthegeneralpopulation,mostwomendevelopingmild Conclusions. nor newbornsdied.SixpercentoftheinfantstestedpositiveforSARS-CoV-2 atbirth. least onedrugagainstSARS-CoV-2 infection.Caesareanratewas32.9%,nomothers Region. Overallonethird of thewomendevelopedapneumoniaand49.7%assumedat birth was2.1per1000maternitiesatanationalleveland6.9/1000intheLombardy Results. and April22,2020inanyItalianhospital. all women with confirmed SARS-CoV-2 infection who gave birth betweenFebruary25 Materials andmethods.National population‐based prospective cohort study involving SARS-CoV-2 infection. clinicians by describinghospitalcareoffered to womenwhogavebirthwithconfirmed Introduction. Abstract 2 1 Alice Maraschini cohort study results ofanationalpopulation-based Coronavirus andbirthinItaly: DOI: 10.4415/ANN_20_03_17 Ann IstSuperSanità2020|Vol. 56,No.3:378-389 Viale ReginaElena299, 00161Rome,Italy. E-mail:[email protected]. Address forcorrespondence : SerenaDonati,Centro NazionalediPrevenzionedelle Malattie ePromozionedellaSalute,Istituto SuperiorediSanità, *the membersoftheItOSSCOVID-19Working GrouparereportedbeforetheReferences Rome, Italy and theItOSSCOVID-19Working Group* Dipartimento diBiomedicinaePrevenzione,UniversitàdegliStudiTor Vergata, Rome,Italy Centro NazionalediPrevenzionedelleMalattieePromozionedellaSalute,IstitutoSuperioreSanità, There isacriticalneedtodevelopclinicalguidance To date,therehavebeenseveralpublishedcaseseries Following theChineseepidemic,Italyiscurrently The incidencerateofconfirmed SARS-CoV-2 infectioninwomenwhogave Clinical featuresandoutcomesofCOVID-19inwomenwhogavebirth The studywasimplementedtoprovideguidancedecision-makersand 1 , EdoardoCorsi 2 , MicheleAntonioSalvatore SARS-CoV-2 infectionestimate,COVID-19 pneumo- sponse toSARS-CoV-2 frommaternalperipheral blood. features ofCOVID-19pneumonia and/orantibodyre- swab and/orchestX-rayorcomputertomography(CT) for theSARS-CoV-2 virusthroughanasopharyngeal scriptase-polymerase chainreaction(RT-PCR) testing study protocolforthediagnosisincludes:reversetran- firmed duringpregnancy. Themethodssetforthinthe any ItalianhospitalwithaSARS-CoV-2 infectioncon- collects informationonallwomenwhogavebirthin MATERIALS ANDMETHODS addressed. SARS-CoV-2 virus from mothers to newborns is also ian hospitalsforchildbirth.Possibletransmissionofthe with confirmedSARS-CoV-2 infectionadmittedtoItal- and postpartuminItaly. based cohort study on COVID-19 in pregnancy, birth The study’s core outcomes include incidence rate of The ItOSSnationalpopulation‐basedcohortstudy This paperdescribeshospitalcareofpregnantwomen 1 , SerenaDonati

• • • • • Key words cohort studies COVID-19 pregnancy outcome SARS-CoV-2 pregnancy 1

379 COVID-19 and birth in Italy

nia, preterm birth, mode of delivery, invasive respira- place over the same study period in Italy, the incidence tory support, intensive care unit (ICU) admission, and rate of confirmed SARS-CoV-2 infection in women who maternal and neonatal severe morbidity and mortality. gave birth was 2.1 per 1000 (CI 95% 1.8-2.4) materni- The ItOSS network of trained reference clinicians ties at a national level, 3.9/1000 (CI 95% 3.2-4.6) in working in Italian public and private maternity units Northern, 1.0/1000 (CI 95% 0.6-1.6) in Central, and covering 91% of total births [7] has been extended 0.2/1000 (CI 95% 0.1-0.5) in Southern Italy. The rate in to reach nationwide coverage for the present study. was 6.9/1000 (CI 95% 5.5-8.5) maternities. Through this system, cases are rapidly notified and data During the 14 days prior to diagnosis, 41.1% of the reviews on maternal and neonatal management collected. In- women reported having had contact with a probable formed consent to the participation is acquired from case (32.9%) or entering health care facilities with con-

any woman at study enrolment. firmed SARS-CoV-2 cases (13.7%). and

A multidisciplinary expert group of clinicians has re- Table 1 describes the socio-demographic characteris- vised the data entry form and its online version has been tics of the cohort stratified by COVID-19 pneumonia pre-tested. The form is designed to collect information occurrence. Women’s median age is 32 years (q1-q3 = regarding the woman’s socio-demographic characteris- 29-36). Women without Italian citizenship are 18.5% of articles tics, medical and obstetrical history, pneumonia diagno- the entire cohort, respectively 27.7% and 14.4% of the sis and treatment, mode of delivery, and maternal and group with and without pneumonia (p-value = 0.049). neonatal outcomes. Previous comorbidities were significantly higher (p- In case of maternal death from SARS-CoV-2 infec- value = 0.023) amongst the pneumonia group (34.0%) tion, the ItOSS maternal mortality surveillance system compared to unaffected women (17.2%), obesity be- riginal will allow verification and provide further information. ing the most frequent condition. None of the women O Since this is an observational study, the cohort size smoked until the end of pregnancy, and foetal growth depends on the incidence of the disease; therefore, a restriction was diagnosed in 2.0% of the cases. formal power calculation has not been performed. As reported in Table 2, multiparas are 69.2% of the The data are collected and processed by personnel re- cohort and 67.1% of the mothers gave birth vaginally, sponsible for ensuring confidentiality and security. This 25.5% under epidural analgesia. Caesarean section analysis reports hospitalized cases from February 25 to (CS) rate was 32.9% overall, 48.9% among the pneu- April 22, 2020, for whom complete data have been re- monia group and 25.3% among the unaffected group ceived by April 11, 2020. (p-value = 0.004). COVID-19 indication to CS con- The incidence rates of women with confirmed SARS- cerned 7.5% of the entire cohort. General anaesthesia CoV-2 infection who gave birth with a 95% confidence was performed in 5 cases. interval (CI) were estimated at a national level, by geo- Overall, 19.2% of the cohort gave birth preterm, graphical area, and for the Lombardy Region. Denomi- 12.3% due to spontaneous onset of preterm delivery (n nator estimates are based on the national Birth Registry = 18), and 6.9% due to iatrogenic labour induction (n providing the most recent available data on deliveries = 3) and urgent/emergency CS (n = 7). Preterm birth (year 2018) assuming an annual reduction of 3% in <37 weeks concerned 31.9% of the women affected by births. The data analysis focuses on descriptive statis- pneumonia compared to 13.1% of the unaffected (p- tics stratified by the occurrence of pneumonia. Signifi- value = 0.007). Among preterm births, the majority cant differences between the two groups were assessed were late preterm (Table 2). through the Pearson’s Chi-squared test or the Fisher’s On hospital admission, 28.1% of the women were as- exact test for categorical variables and the Mann-Whit- ymptomatic. The onset of clinical symptoms occurred ney U test for continuous variables. Data analyses were in 9.5% of the cases on the day of delivery, and in 90.5% performed at the Italian National Health Institute us- before it, the median value being 8 days (range 1-52 ing the Statistical Package Stata/MP 14.2. days). Overall, fever (47.9%), cough (46.6%), and gen- The Ethics Committee of the Italian National Insti- eral weakness (35.6%) were the most common symp- tute of Health approved the project (Prot. 0010482 CE toms at presentation. Table 3 highlights the higher 01.00, Rome 24/03/2020). percentage of symptoms amongst women affected by This study has not received any funding. pneumonia, 31.9% reporting dyspnoea vs 5.1% of the unaffected (p-value <0.001). RESULTS One third of the women developed COVID-19 pneu- From February 25 to April 22, 2020, 146 women who monia; oxygen saturation <95% and abnormal results of gave birth in any Italian Obstetric Unit with confirmed blood gas test concerned respectively 27.7% and 35.6% SARS-CoV-2 infection during pregnancy were notified of the women with pneumonia compared to 4% of the to ItOSS. The diagnosis of SARS-CoV-2 infection was unaffected. confirmed for 142 patients by RT-PCR testing through Over 80% of the women who developed pneumonia nasopharyngeal swab and in 4 cases through chest X-ray. received at least one pharmacological treatment (Table Out of the total cases, 126 (86.3%) were notified by 3), hydroxychloroquine was the most frequently ad- 5 Regions and 2 Autonomous Provinces located in the ministered drug alone (4.3%) and in combination with North of the country. Among the cases reported in the other medical therapies (69.6%). The most frequently North, 84 (57.5%) were identified in the Lombardy adopted therapeutic scheme included hydroxychloro- Region. quine in association with antivirals and antibiotics con- Among an estimated 70 343 maternities that took cerning 13.8% of the cohort and 37.0% of the group 380 Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore et al.

Table 1 Women’s socio-demographic characteristics by occurrence of COVID-19 pneumonia Characteristics Total No COVID-19 COVID-19 p-value (N = 146) pneumonia pneumonia (N = 99) (N = 47) n % n % n %

Maternal agea

reviews <30 40 (28.0) 23 (24.0) 17 (36.2) 0.174

30-34 53 (37.1) 35 (36.5) 18 (38.3) and

≥35 50 (35.0) 38 (39.6) 12 (25.5) Citizenship Not Italian 27 (18.5) 14 (14.1) 13 (27.7) 0.049 Italian 119 (81.5) 85 (85.9) 34 (72.3) articles

Country of birth Italy and western Europe 112 (76.7) 83 (83.8) 29 (61.7) 0.010 East Europe 10 (6.8) 7 (7.1) 3 (6.4) riginal Africa 9 (6.2) 4 (4.0) 5 (10.6) O South/Central America 11 (7.5) 4 (4.0) 7 (14.9) Asia 4 (2.7) 1 (1.0) 3 (6.4) Educational level ≤8 years 18 (12.3) 10 (10.1) 8 (17.0) 0.469 >8 years 86 (58.9) 59 (59.6) 27 (57.4) Missing 42 (28.8) 30 (30.3) 12 (25.5) Previous comorbidities No 113 (77.4) 82 (82.8) 31 (66.0) 0.023 Yes 33 (22.6) 17 (17.2) 16 (34.0) Obesity 22 (15.1) 11 (11.1) 11 (23.4) 0.067 Autoimmune disease 4 (2.7) 2 (2.0) 2 (4.3) 0.387 Diabetes 6 (4.1) 4 (4.0) 2 (4.3) 0.649 Hypertension 5 (3.4) 1 (1.0) 4 (8.5) 0.042 Other 3 (2.1) 2 (2.0) 1 (2.1) 0.691 Smoking in pregnancy Never 118 (80.8) 78 (78.8) 40 (85.1) 0.719 Quit before or during pregnancy 14 (9.6) 11 (11.1) 3 (2.1) Missing 14 (9.6) 10 (10.1) 4 (8.5)

a 3 missing values in “No COVID-19 pneumonia” group.

with pneumonia. Among women with pneumonia, the conditions due to severe morbidity (2 acute respiratory subgroup at higher risk of worse outcomes – defined distress syndromes, 1 respiratory failure, 1 preeclampsia, by the presence of at least one previous comorbidity 2 postpartum haemorrhages, and one thrombosis). In- or C-reactive protein >10mg/100ml or dyspnoea – re- vasive ventilatory support concerned 11 women (7.5%), ceived more often the combination of the three drugs. orotracheal intubation 2 mothers (1.4%). ICU admis- The women belonging to this subgroup, compared to sion regarded 7 patients (4.8%), on average for 7 days. those at lower risk, registered the longest hospital stay None required extracorporeal membrane oxygenation (median value 13 days), received more often invasive and none died. Two percent of the women were trans- ventilatory support due to severe morbidity and were ferred from another hospital, and 18 were still hospital- admitted more frequently to ICU. ized at the end of the study period. The median hospital Antenatal corticosteroids for foetal lung maturation stay was respectively 10 and 4 days for the women with were administered respectively to 10.9% and 4.0% of and without pneumonia (p-value <0.001). the groups with and without pneumonia. Overall, 2 stillbirths were detected respectively at 30 Severe adverse maternal outcomes have been rare, af- and 35 weeks of pregnancy (Table 5). There were 143 fecting almost exclusively the group with pneumonia, as singletons and 3 sets of twins, 85.0% of the newborns described in Table 4. Seven women (4.8%) were in critical weighed ≥2500 gr, 3.4% <1500 gr. Median Apgar index 381 COVID-19 and birth in Italy

Table 2 Obstetric characteristics by occurrence of COVID-19 pneumonia Characteristics Total No COVID-19 COVID-19 p-value (N = 146) pneumonia pneumonia (N = 99) (N = 47) n % n % n % Parity

Nulliparae 45 (30.8) 32 (32.3) 13 (27.7) 0.569 reviews

Multiparae 101 (69.2) 67 (67.7) 34 (72.3) and

Multiple pregnancy No 143 (97.9) 97 (98.0) 46 (97.9) 0.691 Yes 3 (2.1) 2 (2.0) 1 (2.1) Mode of delivery articles Vaginal 98 (67.1) 74 (74.7) 24 (51.1) 0.001 Elective CS 12 (8.2) 9 (9.1) 3 (6.4) Urgent/emergency CS due to maternal/foetal 25 (17.1) 14 (14.1) 11 (23.4)

indication riginal

Urgent/emergency CS due to COVID-19 11 (7.5) 2 (2.0) 9 (19.1) O

PPROMa No 134 (95.0) 90 (94.7) 44 (95.7) 0.587 Yes 7 (5.0) 5 (5.3) 2 (4.3) Gestational age at birth ≤32 6 (4.1) 2 (2.0) 4 (8.5) 0.017 33-36 22 (15.1) 11 (11.1) 11 (23.4) ≥37 118 (80.8) 86 (86.9) 32 (68.1) Median (q1-q3) 38 (37-39) 38 (37-40) 38 (36-39) CS: Caesarean Section; PPROM: Preterm Premature Rupture of Membranes; q1: 25% percentile; q3: 75% percentile. a 5 missing values (4 in “No COVID-19 pneumonia” group and 1 in “COVID-19 pneumonia” group). was 9 at 1 minute and 10 at 5 minutes; at 5 minutes, 1 ity of the population. In the current phase, following infant had Apgar index <4 (Table 4). Admission to neo- the lockdown, the early identification of new cases and natal intensive care unit (NICU) concerned 23 new- their contact traceability and isolation will be decisive borns (15.6%), 18 of whom were preterm, including 6 for the containment of the virus circulation. <32 weeks of gestation. Four infants developed severe Compared to the reference population of women morbidity (1 interstitial pneumonia and 3 respiratory giving birth in Northern Italy [9, 10], the higher pro- distress syndrome). Two were breeches, one affected by portion of multiparas (69% vs 50%) among the women spina bifida, and one was macrosoma. None of them with SARS-CoV-2 infection confirms the hypothesis of tested SARS-CoV-2 positive at birth and, overall, no greater circulation of the virus in families with children neonatal death was recorded. who are often asymptomatic. Nine newborns (6.1%) tested positive for SARS- Clinical features of the detected pneumonia are simi- CoV-2, 5 were tested on the day of delivery, 1 the day lar to those described by previous studies [2, 6, 11, 12]. after, and 3 after 6-9 days from birth. Out of the 5 new- The majority of cases are mild/moderate, and similarly borns with positive swabs collected within 24 hours to the UKOSS study [4] previous comorbidities are sig- from birth, 4 were delivered vaginally and 1 by pre-la- nificantly associated with pneumonia (p-value = 0.023). bour CS. Three of the positive infants were admitted In Italy, black and minority ethnicities are present in to NICU, and none of them developed a severe illness. a lower proportion than in the UK where a significa- tive association with COVID-19 has been detected [4]. DISCUSSION Hospitalized women without Italian citizenship devel- The incidence rates of SARS-CoV-2 positive wom- op significantly more often pneumonia than the Italians en who gave birth in Italy (2.1/1000), in the North (p-value = 0.049), probably due to delayed access to (3.9/1000), the Centre (1/1000) and the South of the healthcare services. country (0.2/1000) as well as in the Lombardy Region Preterm birth, which is one of the feared outcomes (6.9/1000) reflect the same variation in circulation of of COVID-19, was overall 21.2% in the UK cohort [4], the virus among geographical areas as that detected in 17.4% in the WHO review [6], and 19.2% in Italy com- the general population [8]. This observation impacts on pared to the 7% figure among women who gave birth in the present different seroprevalence and susceptibil- northern Italy [8, 9]. 382 Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore et al.

Table 3 Diagnosis and medical therapy by occurrence of COVID-19 pneumonia

Characteristics Total No COVID-19 COVID-19 p-value (N = 146) pneumonia pneumonia (N = 99) (N = 47) n % n % n % Symptoms reviews

Fever 70 (47.9) 39 (39.4) 31 (66.0) 0.003 Cough 68 (46.6) 38 (38.4) 30 (63.8) 0.004 and Tiredness 52 (35.6) 25 (25.3) 27 (57.4) <0.001 Muscle/joint pain 28 (19.2) 14 (14.1) 14 (29.8) 0.015 Sore throat 27 (18.5) 16 (16.2) 11 (23.4) 0.261 articles

Rhinorrhea 23 (15.8) 17 (17.2) 6 (12.8) 0.528 Dyspnea 20 (13.7) 5 (5.1) 15 (31.9) <0.001 Headache 16 (11.0) 7 (7.1) 9 (19.1) 0.024

riginal Vomiting/Diarrhea 14 (9.6) 5 (5.1) 9 (19.1) 0.005 O Chest pain 5 (3.4) 3 (3.0) 2 (4.3) 0.650 Conjunctivitis 3 (2.1) 3 (3.0) 0 (0.0) 0.553 No symptoms 41 (28.1) 37 (37.4) 4 (8.5) <0.001 Imaging techniques No exams 62 (42.5) 62 (62.6) 0 (0.0) <0.001 Chest X-ray 51 (34.9) 26 (26.3) 25 (53.2) Chest CT 6 (4.1) 2 (2.0) 4 (8.5) Lung ultrasound 6 (4.1) 5 (5.1) 1 (2.1) Association of different techniques 21 (14.4) 4 (4.0) 17 (36.2) Vital signs and laboratory reports Body temperature >37.5 °C 34 (23.3) 11 (11.1) 23 (48.9) <0.001 Lymphopenia (<1500 mm3) 73 (50.0) 40 (40.4) 33 (70.2) 0.008 CRP values >10mg/100ml 34 (23.3) 17 (17.2) 17 (36.2) 0.077 Oxygen saturation <95% 17 (11.6) 4 (4.0) 13 (27.7) <0.001

Blood gas test a Not performed 96 (67.6) 80 (82.5) 16 (35.6) <0.001 Performed. normal results 26 (18.3) 13 (13.4) 13 (28.9) Performed. abnormal results 20 (14.1) 4 (4.1) 16 (35.6)

Drugs administeredb HCQ + Antivirals + Antibiotics 20 (13.8) 3 (3.0) 17 (37.0) <0.001 HCQ + Antibiotics 14 (9.7) 6 (6.1) 8 (17.4) Empirical antibiotics 14 (9.7) 12 (12.1) 2 (4.3) HCQ + Antivirals 8 (5.5) 1 (1.0) 7 (15.2) HCQ alone 8 (5.5) 6 (6.1) 2 (4.3) Antivirals + Antibiotics 5 (3.4) 1 (1.0) 4 (8.7) Targeted antibiotics 2 (1.4) 2 (2.0) 0 (0.0) Antivirals alone 1 (0.7) 1 (1.0) 0 (0.0) No pharmacological treatment 73 (50.3 67 (67.7) 6 (13.0) Antenatal corticosteroids 9 (6.2) 4 (4.0) 5 (10.9) 0.132 CRP: C-reactive protein; CT: Computed tomography; HCQ: hydroxychloroquine. a 4 missing values (2 in “No COVID-19 pneumonia” group and 2 in “COVID-19 pneumonia” group). b 1 missing value in “COVID-19 pneumonia” group. 383 COVID-19 and birth in Italy

Table 4 Maternal outcomes by occurrence of COVID-19 pneumonia Characteristics Total No COVID-19 COVID-19 p-value (N = 146) pneumonia pneumonia (N = 99) (N = 47) n % n % n % Severe morbidity 7 (4.8) 3 (3.0) 4 (8.5) 0.215

Non invasive respiratory support 28 (19.2) 6 (6.1) 22 (46.8) <0.001 reviews

Invasive respiratory support 11 (7.5) 2 (2.0) 9 (19.1) 0.001 and

Orotracheal intubation 2 (1.4) 1 (1.0) 1 (2.1) 0.544 ICU admission 7 (4.8) 2 (2.0) 5 (10.6) 0.035 Extracorporeal membrane oxygenation 0 (0.0) 0 (0.0) 0 (0.0) - Maternal death 0 (0.0) 0 (0.0) 0 (0.0) - articles ICU: Intensive Care Unit.

Table 5 Foetal and neonatal outcomes by occurrence of COVID-19 pneumonia riginal O Characteristics Total No COVID-19 COVID-19 p-value (N = 149) pneumonia pneumonia (N = 101) (N = 48) n % n % n % Stillbirth 2 (1.3) 1 (1.0) 1 (2.1) 0.542 Livebirth 147 (98.7) 100 (99.0) 47 (97.9) Neonatal birthweight (g) <1500 5 (3.4) 3 (3.0) 2 (4.3) 0.267 1500-2499 17 (11.6) 9 (9.0) 8 (17.0) ≥2500 125 (85.0) 88 (88.0) 37 (78.7) Apgar 1 min >7 128 (87.1) 91 (91.0) 37 (78.7) 0.039 Apgar 5 min >7 140 (95.2) 98 (98.0) 42 (89.4) 0.022 NICU admission 23 (15.6) 10 (10.0) 13 (27.7) 0.007 Neonatal morbidity 4 (2.7) 1 (1.0) 3 (6.4) 0.106 Neonatal death 0 (0.0) 0 (0.0) 0 (0.0) Neonatal positive SARS-CoV-2 test No 138 (93.9) 94 (94.0) 44 (93.6) 0.594 Positive test <24 hrs of age 5 (3.4) 4 (4.0) 1 (2.1) Positive test ≥24 hrs of age 4 (2.7) 2 (2.0) 2 (4.3) NICU: Neonatal Intensive Care Unit.

Mothers requiring critical care were 9% in the UK [4] The observation that almost 60% of the enrolled and 7.5% in Italy. NICU admissions concerned respec- women did not have risky contacts during the 14 days tively 26% in the UK [4], 6.2% of the infants described prior to symptom onset raises the challenge of the im- by the WHO review [6], and 4.8% of the Italian new- pact of asymptomatic infections and the opportunity borns (Table 4) but this variability could be related to dif- to consider screening policies for pregnant women at ferent local admission policies to NICU regarding quar- hospital admission, currently available in few Italian Re- antine or neonatal observation. The UK cohort reported gions and/or hospitals. 5 maternal and 2 neonatal deaths compared to zero The proportion of women with pneumonia receiving deaths reported in Italy and China. The interpretation pharmacological treatment against SARS-CoV-2 in- of these differences is not straightforward even though fection in this study is high, 73.9% receiving hydroxy- the lower prevalence of minority ethnicities could play a chloroquine alone or in association with antivirals and/ role as well as the different pattern of drug prescriptions. or empirical antibiotics (Table 3). Lopinavir amongst The data are preliminary and collection is still ongo- antivirals is the most frequently used, probably due to ing; nevertheless, they confirm a better course of the the experience of its use in HIV positive women. No disease compared to H1N1 flu and SARS and MERS differences on maternal and neonatal outcomes have epidemics [1, 2]. been detected according to the drugs administered. 384 Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore et al.

Conversely, among the UKOSS cohort hydroxychloro- scribing hydroxychloroquine and antivirals observed in quine was not administered at all and 2% of the women the UK and Italy promotes a reflection on the deter- received antivirals [4]. At the end of May, the Italian mining factors that guide clinicians in deciding whether Medicines Agency (AIFA) suspended the authoriza- and to what extent they should confidently prescribe tion for the prescription by the National Health Service drugs for which conclusive evidence is still unavailable. of hydroxychloroquine and lopinavir/ritonavir for CO- The study’s strengths are the national population- VID-19, except for use in clinical studies. It is necessary based prospective design and the opportunity to anal- to study safety and effectiveness of medications used in yse data from the beginning of the epidemic. Another reviews

pregnant women to guide decision-making about treat- asset is the wealth of information contained in the data ment options for COVID-19 disease and associated collection form. Limitations include the analysis of pre-

and complications. liminary data while the pandemic is still underway and

Antenatal corticosteroids for foetal lung maturation the constraints linked to the impossibility of general- have been prescribed as recommended [2] to 4 women izing the results without taking into account the differ- <34 weeks and to 5 women at 34-35 weeks of gestation. ent prevalence of the condition by geographical area. The detected rate of CS in the Italian cohort is 32.9%, The lack of information regarding women infected in articles

higher compared to the rate of the Northern Regions the early stages of pregnancy is also a limit of the study, (26%) but considerably lower compared to 59% of the but the ItOSS will follow-up affected women currently UK cohort [4], 73.5% reported by the WHO review in the first trimester of pregnancy. [6], and 85% of the Chinese series [3-6]. As reported in riginal the UK, the majority of CS indications were not due to CONCLUSIONS O SARS-CoV-2 infection, which is not in itself necessarily The clinical presentation of SARS-CoV-2 infection in an indication for delivery, nor for CS. The proportion of women who gave birth appears to be similar to the gen- CS due to COVID-19 concerned in fact a minority of eral population. The lesson learned by reviewing hos- the total surgeries, 7.5% of Italian and 16% of UK co- pital care offered to affected Italian women who gave hort. Although international agencies are unequivocal in birth, confirmed the current priority of physical distanc- claiming that the disease is not an indication for CS and ing measures, the urgent need for stronger evidence on that the protection of birth physiology is a priority [2, the safety and effectiveness of medical therapy and a 13, 14], clinical practice seem not to follow the current continued commitment to face the challenge of re- recommendations. Italy, which has historically recorded specting and protecting childbirth physiology. These higher CS rates compared to the UK, on this occasion findings confirm the primary importance of compari- didn’t show the same significant increase in CS as in sons among population-based cohorts to support health other countries despite the early onset of the epidemic. professionals and decision-makers with evidence-based An issue still highly debated concerns the possibility recommendations. of mother to foetus transmission of SARS-CoV-2 virus. The WHO review shows that 6.6% of the newborns Acknowledgements tested as suspected to have COVID-19, the UK cohort We thank Silvia Andreozzi and Mauro Bucciarelli for reports 5% of positive infants, and the ItOSS cohort 6%. their valuable technical support and assistance to the Although evidence is sparse, vertical transmission can- operation of the web-based data collection system. We not be excluded [15, 16] but it appears to be rare and, thank Clarissa Bostford for language editing. for the most part, babies, who must be carefully moni- Our heartfelt thanks go to all the clinicians working tored, have a good prognosis. in the national network of maternity units (Appendix) Many case reports and small case series have already for the assistance offered to women and for collecting been published on COVID-19 in pregnancy; however, the data, we thank all women who agreed to participate there is a lack of population-based data that could allow in the study. incidence rates to be estimated and unbiased character- istics and outcomes to be described and compared [4]. Disclosure statement The scientific community, as well as international jour- The Authors and the working group members report nal peer review procedures, should better support the no conflict of interest. development and the dissemination of studies adopting population-based approaches to properly inform clini- Source of financial support cians and decision-makers. This study has not received any financial support. Pregnant women are often not included in clinical trials on drugs [17]. However, the good news is that Individual contribution to the manuscript recently the European Medicines Agency and Health Alice Maraschini: conceptualization, methodology, Canada under the aegis of the International Coalition of software, formal analysis, writing-review and editing; Medicines Regulatory Authorities have agreed on three Edoardo Corsi: methodology, investigation, data cu- priority areas for cooperation on observational research ration, writing-review and editing, project administra- during the outbreak of COVID-19 [18]. One area is tion; Michele Antonio Salvatore: methodology, formal devoted to research in pregnancy in order to examine analysis, writing-review and editing; Serena Donati: the impact of both Coronavirus disease and the use of conceptualization, methodology, formal analysis, writ- drugs on pregnant women infected with SARS-CoV-2 ing original draft, supervision; Ilaria Lega: investigation, and their unborn babies. The different practices in pre- writing-review and editing; Paola D’Aloja: investigation, 385 COVID-19 and birth in Italy

writing-review and editing; Letizia Sampaolo: literature 18ASST Papa Giovanni XXII, Bergamo, Italy review, writing-review and editing; Paola Casucci: in- 19Regione Emilia-Romagna, Bologna, Italy vestigation, data curation; Irene Cetin: investigation, 20 ASST Spedali Civili di Brescia, Italy data curation; Gabriella Dardanoni: investigation, 21IRCSS Giannina Gaslini, Genoa, Rome data curation; Franco Doganiero: investigation; Enrico 22ASST Fatebenefratelli Sacco, Milan, Italy Ferrazzi: investigation, data curation; Massimo Piergi- 23Azienda Ospedaliera S. Carlo, Potenza, Italy useppe Franchi: investigation, data curation; Livio Leo: 24Ospedale Centrale di Bolzano, Bozen, Italy investigation, data curation; Marco Liberati: investiga- 25Ospedale Santa Chiara, Trento, Italy

26 reviews tion, data curation; Mariavittoria Locci: investigation, Mater Dei Hospital, Bari, Italy data curation; Claudio Martini: investigation, data cu- 27Fondazione MBBM/Ospedale San Gerardo, Monza,

ration; Federico Mecacci: investigation, data curation; Italy and

Alessandra Meloni: investigation, data curation; Anna Domenica Mignuoli: investigation, data curation; Lu- Received on 9 June 2020. isa Mondo: investigation, data curation; Luisa Patanè: Accepted on 16 July 2020. investigation, data curation; Enrica Perrone: investiga- articles tion, data curation, writing review and editing; Federico Prefumo: investigation, data curation; Luca Ramenghi: APPENDIX. Reference clinicians and maternity units investigation, data curation; Valeria Savasi: investiga- participating in the project network tion, data curation; Sergio Schettini: investigation, data curation; Martin Steinkasserer: investigation, data cu- Piemonte riginal ration; Saverio Tateo: investigation, data curation; Vito Elena Amoruso Ospedale Sant’Andrea Vercelli; Alberto O Trojano: investigation, data curation; Patrizia Vergani: Arnulfo, Enrico Finale Stabilimento Ospedaliero Castel- investigation, data curation. li Verbania; Maria Bertolino, Andrea Guala Ospedale San Biagio Domodossola; Marisa Biasio, Ruggero Croc- The ItOSS COVID-19 Working Group co, Pietro Gaglioti, Luca Marozio, Clara Monzeglio Serena Donati1, Alice Maraschini1, Ilaria Lega1, Sant’Anna - AOU Città della Salute e della Scienza di Paola D’Aloja1, Letizia Sampaolo1, Michele Antonio Torino; Mario Canesi, Sara Cantoira Ospedale Maria Salvatore1, Edoardo Corsi2, Salvatore Alberico3, Paola Vittoria Torino; Paola Capelli Istituto SS. Trinità Borgo- Casucci4, Irene Cetin5, Gabriella Dardanoni6, Franco manero; Ilaria Careri, Giovanni Lipari Ospedale Martini Doganiero7, Massimo Piergiuseppe Franchi8, Enrico Torino; Luigi Carratta Ospedale S. Spirito Casale Monfer- Ferrazzi9, Livio Leo10, Marco Liberati11, Mariavittoria rato; Ilaria Costaggini Ospedale degli Infermi Rivoli; Tania Locci12, Claudio Martini13, Federico Mecacci14, Ales- Cunzolo Presidio Osp. Cardinal G. MASSAIA Asti; Enza sandra Meloni15, Anna Domenica Mignuoli16, Luisa De Fabiani Azienda Ospedaliera Ordine Mauriziano To- Mondo17, Luisa Patanè18, Enrica Perrone19, Federico rino; Alberto De Pedrini Ospedale Maggiore della Carità Prefumo20, Luca Ramenghi21, Valeria Savasi22, Sergio Novara; Fiorenza Droghini, Paola Rota Ospedale San- Schettini23, Martin Steinkasserer24, Saverio Tateo25, Vito ta Croce Moncalieri; Germano Giordano Ospedale SS. Trojano26, Patrizia Vergani27 Pietro e Paolo Borgosesia; Daniela Kozel Ospedale Civile SS. Antonio e Biagio Alessandria; Vincenzo Lio Ospedale 1Centro Nazionale di Prevenzione delle Malattie e Civico Chivasso; Francesca Maraucci Ospedale degli in- Promozione della Salute, Istituto Superiore di Sanità, fermi Biella; Simona Mazzola Ospedale Maggiore Chieri; Rome, Italy Maria Milano Ospedale Civile Mondovì; Giovanna Oggè 2Dipartimento di Biomedicina e Prevenzione, Univer- Ospedale maggiore SS. Annunziata Savigliano; Simona sità di Roma Tor Vergata, Rome, Italy Pelissetto Ospedale Civile di Ivrea; Pasqualina Russo 3IRCCS Burlo Garofolo, Trieste, Italy Presidio ospedaliero riunito Ciriè; Manuela Scatà Ospe- 4Regione Umbria, Perugia, Italy dale Civico di San Lazzaro Alba; Federico Tuo Ospedale 5Università degli Studi di Milano, Milan, Italy San Giacomo Novi Ligure/Tortona; Concetta Vardè Ospe- 6Regione Sicilia, Palermo, Italy dale Agnelli Pinerolo; Elena Vasario Azienda Ospedaliera 7Ospedale Civile Antonio Cardarelli, Campobasso, S. Croce e Carle Cuneo Italy 8Università degli Studi di Verona, Verona, Italy Valle D’Aosta 9Fondazione IRCCS Cà Granda, Ospedale Maggiore Livio Leo Umberto Parini Aosta Policlinico Milan, Italy 10Ospedale Beauregard Valle D’Aosta, Aosta, Italy Lombardia 11Università degli Studi Gabriele d’Annunzio, Chieti- Debora Balestrieri Ospedale di Cittiglio; Federica Pescara, Italy Baltaro Ospedale Niguarda di Milano; Pietro Barbaci- 12Università degli Studi di Napoli Federico II, Naples, ni, Elisabetta Venegoni Ospedale di Magenta; Miche- Italy le Barbato Ospedale di Melegnano; Lorena Barbetti 13Regione Marche, Ancona, Italy Ospedale di Esine; Paolo Beretta Ospedale di Como; 14Università degli Studi di Firenze, Florence, Italy Bruno Bersellini Ospedale di Sondrio; Stefano Bianchi 15Regione Autonoma della Sardegna, Cagliari, Italy Ospedale San Giuseppe di Milano; Giuseppina Bizzoni, 16Regione Calabria, Reggio Calabria, Italy Giancarlo Garuti Ospedale di Lodi; Antonia Botrugno 17ASL TO3, Turin, Italy Ospedale di Casalmaggiore; Donatella Bresciani Ospe- 386 Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore et al.

dale di Desenzano; Alessandro Bulfoni Pio X Humani- Spa; Tiziano Maggino Ospedale all’Angelo di Mestre; tas di Milano; Carlo Bulgheroni Ospedale di Gallara- Mario Marando Ospedale di Portogruaro e Ospedale di te; Orlando Caruso, Elena Pinton Ospedale di Chiari; San Donà di Piave; Giovanni Martini Ospedale di Valda- Massimo Ciammella Ospedale di Seriate; Elena Cresta- gno; Carlo Maurizio Ospedale di Mirano; Yoram J. Meir ni, Giulia Pellizzari Ospedale di Pieve di Coriano; Anto- Ospedale di Bassano del Grappa; Marcello Rigano Ospe- nella Cromi Ospedale di Varese; Rosa Di Lauro, Carla dale di Camposampiero; Cesare Romagnolo Ospedale di Foppoli Ospedale di Sondalo; Patrizia D’Oria, Ospedale San Bonifacio; Roberto Rulli Ospedale di Cittadella; Ma- di Alzano; Santina Ermito Ospedale di Piario; Massimo ria Grazia Salmeri Ospedale di Montebelluna; Marcello reviews

Ferdico Ospedale di Vimercate; Maria Fogliani, Guido Scollo Ospedale di Santorso; Francesco Sinatra Ospedale Stevanazzi Ospedale di Legnano-Cuggiono; Roberto Fo- di Conegliano; Gianluca Straface Casa di cura Abano;

and gliani Ospedale di Sesto San Giovanni; Ambrogio Frige- Fabio Gianpaolo Tandurella Ospedale di Pieve di Cado-

rio Ospedale di Rho; Eleonora Fumagalli Ospedale Ma- re e Ospedale San Martino – Belluno; Marco Torrazzina cedonio Melloni ASST FBF-Sacco di Milano; Roberto Ospedale di Bussolengo - Ospedale di Villafranca; Paolo Garbelli Brescia Istituto Clinico S. Anna; Daniela Gatti Lucio Tumaini Ospedale di Arzignano; Giuliano Zanni Ospedale di Manerbio; Giampaolo Grisolia, Serena Va- Ospedale di Vicenza articles ralta Ospedale di Mantova; Paolo Guarnerio Ospedale San Carlo di Milano; Enrico Iurlaro IRCCS Cà Gran- Friuli-Venezia Giulia da Ospedale Maggiore Policlinico-Mangiagalli Milano; Emanuele Ancona Pordenone S. Giorgio; Fabiana Stefano Landi Ospedale di Gravedona; Mario Leonardi Cecchini AO Udine; Michela De Agostini Civile Pal- riginal Ospedale di Iseo; Stefania Livio Ospedale Buzzi ASST manova; Gianpaolo Maso Burlo IRCSS Trieste; Edlira O FBF-Sacco di Milano; Anna Locatelli Ospedale di Ca- Muharremi Pordenone S.M. degli Angeli; Alessandra Ni- rate; Giuseppe Losa Ospedale di Melzo; Massimo Lo- coletti S. Daniele - Tolmezzo; Roberta Pinzano AO S.M. votti Como Valduce; Anna Minelli Ospedale di Gavardo; degli Angeli S. Vito; Lucia Zanazzo, Monfalcone Luisa Muggiasca Ospedale di Garbagnate; Giuseppe Nucera Ospedale di Busto Arsizio; Guido Orfanotti Liguria Ospedale di Desio; Alessandra Ornati Ospedale di Vi- Angelo Cagnacci Ospedale S. Martino di Genova; Luca gevano; Luisa Patanè ASST Papa Giovanni XXIII Ber- Ramenghi IRCCS Giannina Gaslini di Genova gamo; Antonio Pellegrino Ospedale di Lecco; Francesca Perotti, Arsenio Spinillo Fondazione IRCCS Policlinico Emilia-Romagna San Matteo di Pavia; Ezio Pozzi Ospedale di Broni Stra- Lorenzo Aguzzoli, Emanuele Soncini Ospedale S.M. della- Ospedale di Voghera, Federico Prefumo Spedali Nuova Reggio Emilia; Patrizio Antonazzo, Lucrezia Pi- Civili di Brescia; Anna Catalano Brescia Fondazione gnatti Ospedale Bufalini Cesena; Angela Bandini, Isabel- Poliambulanza; Marina Ravizza Ospedale San Paolo la Strada Ospedale G.B. Morgagni - L. Pierantoni For- di Milano; Aldo Riccardi Ospedale di Cremona; Ales- lì; Chiara Belosi Ospedale degli Infermi Faenza; Renza sia Chiesa Ospedale di Ponte San Pietro; Tazio Sacconi Bonini, Maria Cristina Ottoboni Ospedale Guglielmo Ospedale di Asola; Valeria Savasi Ospedale Sacco di Mi- Da Saliceto Piacenza; Fabrizio Corazza, Paola Pennac- lano; Ubaldo Seghezzi Ospedale di Saronno; Vincenzo chioni Ospedale Ss. Annunziata Cento; Fabio Facchi- Siliprandi Ospedale di Crema; Paolo Valsecchi Ospedale netti, Giliana Ternelli Azienda Ospedaliero-Universitaria San Raffaele; Laura Vassena Ospedale di Merate; Patri- Modena; Alessandro Ferrari, Cristina Pizzi, Ospedale zia Vergani Fondazione MBBM Ospedale San Gerardo S.M. Bianca Mirandola; Tiziana Frusca, Stefania Fieni Monza; Antonella Villa Ospedale di Treviglio; Matteo Azienda Ospedaliero-Universitaria Parma; Maria Cri- Zanfrà Ospedale di Tradate; Alberto Zanini Ospedale di stina Galassi, Federica Richieri, Nuovo Ospedale Ci- Erba vile Di Sassuolo S.P.A.; Francesco Giambelli, Carlotta Matteucci Ospedale S.M. Delle Croci Ravenna; Pantaleo Veneto Greco, Danila Morano Azienda Ospedaliero-Universita- Giuseppe Angeloni Ospedale di Piove di Sacco; An- ria Ferrara; Marinella Lenzi, Ilaria Cataneo, Ospedale tonio Azzena Ospedale di Vittorio Veneto; Roberto Bac- Maggiore Bologna; Gialuigi Pilu, Marisa Bisulli, Azienda cichet, Cristina Napolitano Ospedale di Oderzo; Valen- Ospedaliero-Universitaria Bologna; Maria Cristina Selle- tino Bergamini Ospedale Borgo Trento; Luca Bergamini ri Ospedale di Bentivoglio; Federico Spelzini, Elena De Ospedale di Chioggia; Enrico Busato, Monica Zannol Ambrosi, Ospedale degli Infermi Rimini; Paolo Venturini, Ospedale di Treviso; Pietro Catapano, Marco Gentile Francesca Tassinati Ospedale B. Ramazzini Carpi; Ste- Ospedale Mater Salutis - Legnago; Marcello Ceccaroni fano Zucchini, Barbara Paccaloni, Ospedale S.M. della Ospedale Sacro Cuore don Calabria Negrar; Gianluca Scaletta Imola Cerri Ospedale SS. Giovanni e Paolo - Venezia; Andrea Cocco Ospedale di Asiago; Enrico Di Mambro Ospedale Toscana di Adria - Ospedale di Rovigo; Carlo Dorizzi Ospedale di Andrea Antonelli, Carlotta Boni Ospedale Civile Ce- Schiavonia; Laura Favretti Ospedale S. Maria del Prato - cina; Maria Paola Belluomini, S. Francesco Barga - PO Feltre; Massimo Franchi Azienda Ospedaliera di Verona; Valle del Serchio e Generale Provinciale Lucca - PO San Franco Garbin Ospedale di Dolo; Maria Teresa Gervasi, Luca, Rosalia Bonura, S. Maria della Gruccia - Ospedale Daniela Truscia Azienda Ospedaliera di Padova; Dome- del Valdarno, Stefano Braccini, SS. Cosimo e Damiano nico Lagamba Ospedale di Castelfranco Veneto; Antoni- Pescia - Osp della Valdinievole, Giacomo Bruscoli e Pa- no Lo Re Ospedale P. Pederzoli - Casa di cura Privata squale Mario Florio, Nuovo Ospedale San Jacopo di Pi- 387 COVID-19 and birth in Italy

stoia, Giovanna Casilla, SS. Giacomo e Cristoforo Massa RR Foggia’ Gin. Uni.; Alessandro DalfieroCerignola; - PO Zona Apuana, Anna Franca Cavaliere, Ospedale Gerardo D’Ambrogio Galatina ‘Santa Caterina Novel- Santo Stefano Prato, Marco Cencini, Ospedali Riuniti la’; Nicola Del Gaudio Castellaneta; Paolo Demarzo della Val di Chiana, Venere Coppola e Laura Miglia- San Severo Teresa Masselli Mascia; Giovanni Di Vagno vacca, Ospedale Misericordia Grosseto, Barbara De Santi, Bari - San Paolo; Giuseppe Laurelli Casa Sollievo Dalla PO Felice Lotti Pontedera, Paola Del Carlo, Ospedale S. Sofferenza - S. Giovanni Rotondo; Roberto Lupo Galli- Giovanni Di Dio Torregalli, Carlo Dettori, Nuovo Ospe- poli; Nicola Macario Altamura; Antonio Malvasi, Bari dale di Borgo S. Lorenzo, Mariarosaria Di Tommaso e - Casa Di Cura Santa Maria; Guido Maurizio, Elisa- reviews

Serena Simeone, Careggi - CTO Firenze - AOU, Giu- betta Monteduro, Acquaviva ‘Miulli’; Andrea Morciano seppe Eremita, Civile Elbano Portoferraio, Sara Failli, Cardinale G. Panico Di Tricase; Lucio Nichilo Umberto I

Ospedale Area Aretina Nord Arezzo, Paolo Gacci, S.M. Corato; Anna Maria Nimis Francavilla Fontana; Anto- and

Annunziata Bagno a Ripoli - Ospedale Fiorentino Sud Est, nio Perrone Lecce Vito Fazzi; Elena Rosa Potì Brindisi Alessandra Meucci, Le Scotte Siena - Azienda ospedalie- ‘Perrino’; Sabino Santamato Monopoli Putignano; Emilio ra universitaria, Filippo Ninni, Riuniti Livorno, Barbara Stola Taranto; Antonio Tau Scorrano; Mario Vicino Bari Quirici, Ospedale Unico Versilia, Alessia Sacchi, Ospe- - Di Venere; Martino Vinci Martina Franca articles dale dell’Alta Val d’Elsa Poggibonsi, Cristina Salvestroni, Ospedale S. Giuseppe Empoli, Sara Zullino, Ospedali Pi- Basilicata sani Pisa - Azienda universitaria Giampiero Adornato Policoro; Francesco Bernasconi Melfi; Alfonso Chiacchio Lagonegro; Sergio Schettini Umbria Azienda Ospedaliera Regionale San Carlo - Potenza; Giu- riginal Gian Carlo Di Renzo, Giorgio Epicoco Azienda seppe Trojano Matera O Ospedaliera Santa Maria della Misericordia di Perugia; Leonardo Borrello Azienda Ospedaliera Santa Maria di Calabria Terni Michele Morelli, Ospedale Annunziata - AO Cosenza; Stefano Palomba Ospedali Riuniti di Reggio Calabria; Marche Morena Rocca Azienda Ospedaliera ‘Pugliese Ciaccio’ di Andrea Ciavattini, Sara D’Eusanio AOU - Ospedali Catanzaro Riuniti di Ancona; Filippo Di Prospero Ospedale di Civi- tanova Marche; Francesco D’Antonio Ospedale Santissi- Sicilia ma Annunziata Chieti Luigi Alio, Giuseppina Orlando ARNAS Civico di Cristina Benfratelli - Palermo; Salvatore Bevilacqua, Fa- Lazio brizio Quartararo Casa di cura Candela SPA - Palermo; Francesco Antonino Battaglia, Immacolata Marcucci Rocco Billone Civico Partinico e Dei Bianchi - Corleone; PO Santa Maria Goretti Latina; Marco Bonito Ospeda- Giuseppe Bonanno, Maria Paternò Arezzo - Ragusa, le San Pietro Fatebenefratelli Roma; Daniele Di Mascio Antonio Bucolo, Umberto I - Siracusa, Claudio Cam- Umberto I - Policlinico di Roma; Sergio Ferrazzani Poli- pione, Casa di cura prof. Falcidia - Catania, Giuseppe clinico Universitario Fondazione Agostino Gemelli - Roma; Canzone, S. Cimino-Termini Imerese, Angelo Caradon- Maria Grazia Frigo Fatebenefratelli San Giovanni Cali- na, V. Emanuele II - Castelvetrano, Sebastiano Caudullo bita - Isola Tiberina; Giorgio Nicolanti Ospedale Belcolle e Cosimo Raffone, AO Papardo - Messina, Giovanni Viterbo; Giancarlo Paradisi, Maria Rita Pecci Ospedale Cavallo, PO Maggiore - Modica, Antonio Cianci, Mi- Fabrizio Spaziani Frosinone chele Fichera V. Emanuele Rodolico - Catania; Salvato- re Corsello, Sergio Di Salvo, Immacolata Schimmenti Abruzzo Casa di cura Villa Serena - Palermo; Rosario D’Anna, Fabio Benucci Sant’Omero; Paola Caputo Sulmona; AOU G. Martino - Taormina, Maria Rosa D’anna, Buc- Sandra Di Fabio, Leonardo Di Stefano L’Aquila; An- cheri La Ferla - Palermo, Maria Di Costa, Basilotta- tonio Di Francesco Lanciano; Diego Gazzolo, Marco Catania, Giuseppe Ettore, ARNAS Garibaldi Nesima Liberati Chieti; Anna Marcozzi Teramo; Francesco Ma- - Catania, Giovanni Falzone, Umberto I - Nicosia, Mi- trullo Vasto; Maurizio Rosati, Gabriella Scorpio Pescara; chele Gulizzi e Francesco La Mantia, G. F. Ingrassia Giuseppe Ruggeri Avezzano - Palermo, Laura Giambanco, S. Antonio Abate - Erice e B. Nagar - Pantelleria, Salvatore Incandela, S. Giovanni Molise di Dio - Agrigento e Giovanni Paolo II - Sciacca; Lilli Franco Doganiero Ospedale di Campobasso Maria Klein S. Vincenzo - Enna e Barone Romeo - Patti; Michele La Greca, Venera Mille, M. SS. Addolorata - Campania Biancavilla; Luigi Li Calsi S. Giacomo d’Altopasso - Li- Luigi Cobellis, Annunziata Mastrogiacomo Ospedale cata; Emilio Lo Meo, Paolo Scrollo Cannizzaro - Ca- di Caserta; Maria Vittoria Locci AOU Federico II Napoli tania; Lucia Lo Presti, Santo Recupero S. Marco (ex V. Emanuele S. Bambino) - Catania e Ospedale Generale Puglia - Lentini; Vincenzo Miceli, S. Raffaele Giglio - Cefalù, Luca Loiudice, Bari - Presidio Mater Dei; Antonio Maria Pia Militello, S. Marta e S. Venera - Acireale, Alfio Belpiede Barletta ‘Mons. Dimiccoli’; Mariano Cantatore Mirenna, Istituto clinico Vidimura (ex Casa di cura Gret- L. Bonomo Andria; Ettore Cicinelli Bari - Policlinico Gi- ter e Lucina) - Catania, Umberto Musarra, G. Fogliani necologia; Aldo D’Aloia, Maurizio De Luca AOU ‘OO - Milazzo e Lipari, Pietro Musso, Abele Ajello - Mazara 388 Alice Maraschini, Edoardo Corsi, Michele Antonio Salvatore et al.

del Vallo, Michele Palmieri, V. Emanuele - Gela, Maria de - Lanusei; Gianfranco Puggioni S. Martino - Oristano; Giovanna Pellegrino, Basso Ragusa Mario - Militello Val Loredana Pagliara Nostra Signora di Bonaria - San Ga- di Catania, Francesco Pellegriti, Castiglione Prestianni- vino; Giulietta Ibba CTO - Iglesias; Caterina Tronci, SS Bronte; Antonio Perino Villa Sofia - Cervello - Palermo; Trinità - Cagliari; Giampiero Capobianco, AOU Sassari; Concetta Remigia Pettinato, Angelo Tarascio Gravina Alessandra Meloni Duilio Casula Monserrato AOU - Ca- - Caltagirone; Vincenzo Scattarreggia Barone Lombar- gliari; Francesca Palla S. Michele AO Brotzu - Cagliari do - Canicattì; Antonio Schifano R. Guzzardi - Vittoria; Calogero Selvaggio S. Elia - Caltanissetta; Luigi Triolo Provincia autonoma di Bolzano reviews

Casa di cure Triolo Zancla SPA - Palermo; Renato Vene- Martin Steinkasser, Micaela Veneziano Ospedale zia P. Giaccone - Palermo Centrale di Bolzano and

Sardegna Provincia autonoma di Trento Speranza Piredda Civile Alghero; Giangavino Peppi Pietro Dal Rì, Fabrizio Taddei UO Rovereto; Roberto Giovanni Paolo II - Olbia; Giovanna Pittorra S. Fracesco Luzietti UO Cles; Saverio Tateo UO Trento; Fabrizia - Nuoro; Gianfranco Depau Nostra Signora della Merce- Tenaglia UO Cavalese articles

REFERENCES riginal

O 1. Schwartz DA, Graham AL. Potential maternal and in- dataset/Rapporto-Cedap/v74r-mqr5. fant outcomes from (Wuhan) Coronavirus 2019-nCoV 9. Perrone E, Formisano D, Gargano G et al. [Birth in infecting pregnant women: lessons from SARS, MERS, Emilia-Romagna. 2018] Bologna: Emilia-Romagna Re- and other human Coronavirus infections. Viruses. gion, 2019 [cited 2020 May 28]. Italian. Available from: 2020;12(2):194. Published 2020 Feb 10. doi: 10.3390/ www.saperidoc.it/flex/cm/pages/ServeBLOB.php/L/IT/ v12020194 IDPagina/349. 2. Royal College of Obstetricians and Gynaecologists, Roy- 10. The COVID-19 Task force of the Department of Infec- al College of Midwives, Royal College of Paediatrics and tious Diseases and the IT Service Istituto Superiore di Child Health, Public Health England and Health Pro- Sanità – Epidemics COVID-19, National update 28th tection Scotland. Coronavirus (COVID-19) Infection in April 2020 [cited 2020 May 28]. Italian. Available from: Pregnancy. Information for healthcare professionals. Ver- www.epicentro.iss.it/coronavirus/bollettino/Bollettino- sion 9: Published Wednesday 13 May 2020 [cited 2020 sorveglianza-integrata-COVID-19_28-aprile-2020.pdf. May 28]. Available from: www.rcog.org.uk/globalassets/ 11. Yang Z, Wang M, Zhu Z, Liu Y. Coronavirus dis- documents/guidelines/2020-05-13-coronavirus-covid- ease 2019 (COVID-19) and pregnancy: a system- 19-infection-in-pregnancy.pdf atic review [published online ahead of print, 2020 Apr 3. Della Gatta AN, Rizzo P, Pilu G, Simonazzi G. COV- 30]. J Matern Fetal Neonatal Med. 2020;1-4. doi: ID-19 during pregnancy: a systematic review of reported 10.1080/14767058.2020.1759541 cases [published online ahead of print, 2020 Apr 18]. Am 12. Arabi S, Vaseghi G, Heidari Z, et al. Clinical characteristics J Obstet Gynecol. 2020;S0002-9378(20)30438-5. doi: of COVID-19 infection in pregnant women: a systematic https://doi.org/10.1016/j.ajog.2020.04.013 review and meta-analysis. medRxiv 2020.04.05.20053983 4. Knight M, Bunch K, Vousden N, Morris E, Simpson N, [preprint]. 2020 [cited 2020 May 28]. Available from: Gale C, O’Brien P, Quigley M, Blocklehurst P, Kurinc- https://doi.org/10.1101/2020.04.05.20053983. zuk JJ. Characteristics and outcomes of pregnant women 13. Poon LC, Yang H, Dumont S, et al. ISUOG Interim admitted to hospital with confirmed SARS-CoV-2 infec- Guidance on coronavirus disease 2019 (COVID-19) dur- tion in the UK: national population based cohort study. ing pregnancy and puerperium: information for health- BMJ. 2020;369:m2107. doi: https://doi.org/10.1136/bmj. care professionals - an update [published online ahead m2107 of print, 2020 May 1]. Ultrasound Obstet Gynecol. 5. Yan J, Guo J, Fan C, et al. Coronavirus disease 2019 2020;10.1002/uog.22061. doi: 10.1002/uog.22061 (COVID-19) in pregnant women: A report based on 116 14. World Health Organization. Clinical management of se- cases [published online ahead of print, 2020 Apr 23]. Am vere acute respiratory infection when novel coronavirus J Obstet Gynecol. 2020;S0002-9378(20)30462-2. doi: (nCoV) infection is suspected – Interim guidance – 13 10.1016/j.ajog.2020.04.014 March 2020 [cited 2020 May 28]. Available from: www. 6. World Health Organization Collaborating Centre for who.int/docs/default-source/coronaviruse/clinical-man- Women’s Health University of Birmingham. COVID-19 agement-of-novel-cov.pdf. in pregnancy (PregCOV-19 LRS) [cited 2020 May 28]. 15. Dong L, Tian J, He S, et al. Possible vertical transmission Available from: www.birmingham.ac.uk/research/who- of SARS-CoV-2 from and infected mother to her newborn collaborating-centre/pregcov/index.aspx. [published online ahead of print, 2020 Mar 26]. JAMA. 7. EpiCentro [Internet]. 2020 Feb 20 [cited 2020 May 28]. 2020;323(18):1846‐8. doi: 10.1001/jama.2020.4621 ItOSS – Italian Obstetric Surveillance System. Italian. 16. Patanè L, Morotti D, Giunta MR, et al. Vertical transmis- Available from: www.epicentro.iss.it/itoss/. sion of COVID-19: SARS-CoV-2 on the fetal side of the 8. Campi R, Cartabia M, Miglio D, Bonati M. Certificato placenta in pregnancies with COVID-19 positive moth- di Assistenza al Parto (CedAP) Regione Lombardia. ers and neonates at birth. Am J Obstet Gynecol MFM. Analisi dell’evento nascita anno 2017 – Laboratorio per 2020;100145. doi: 10.1016/j.ajogmf.2020.100145 la Salute Materno Infantile, Dipartimento di Salute Pub- 17. Costantine MM, Landon MB, Saade GR. Protection by blica, IRCCS – Istituto di Ricerche Farmacologiche Ma- exclusion: another missed opportunity to include preg- rio Negri Milano: Lombardia Region 2019 [cited 2020 nant women in research during the coronavirus disease May 28]. Italian. Available from: www.dati.lombardia.it/ 2019 (COVID-19) pandemic [published online ahead of 389 COVID-19 and birth in Italy

print, 2020 Apr 24]. Obstet Gynecol. 2020. doi: 10.1097/ to cooperate on observational research in the context AOG.0000000000003924 of COVID-19. Available from: www.ema.europa.eu/en/ 18. European Medicines Agency [Internet]. EMA/275870/ news/global-regulators-commit-cooperate-observational- 2020 [cited 2020 May 28]. Global regulators commit research-context-covid-19. reviews

and

articles

riginal O 390 Brief note Agnese Suppiej 1 caine and methamphetamine in the previous year [1]. caine andmethamphetamine inthepreviousyear[1]. used illicitdrugsduringpregnancy andinparticularco- timated that194thousandpregnant womenhadbeen tional SurveyonDrugUseand Health(NSDUH)es- among differentcountries.In 2017,theAmericanNa- a worldwide rising phenomenon with a great variability INTRODUCTION ability duringsubsequentdaysafterbirth,subclinicalseizures orbraininjuries. signs and symptoms, which cannot be evaluated by FNASTsuch asstartles and itsvari- borns inuteroexposedtopsychoactivedrugsandtheneurophysiological definitionof that allowstheevaluationofbehaviouralstateandclinical neurologicalsignsinnew- Discussion. Video-EEG recordingcouldbeconsideredanimportant andobjectivetool psychoactive drugs,excludingepilepticphenomena. of behaviouralstateandclinicalneurologicalsignsinnewbornsuteroexposedto of NAS due toprenatal exposure to cocaine as acomplementarytoolforthe evaluation Results. Anexample for thepotentialuse of suchmethodologyisdiscussedinacase to psychoactivedrugspresentingNAS. if theycanbecorrelatedtopossibleinjuriesincriticalneonatesasthoseexposedutero ous vEEGmonitoringcouldbeusedtodescribesomeclinicalmanifestationsandassess with the neurological development. Due to its feature to be a non-invasive tool continu- as physiologicalmotorparoxysms,thatdisappearfrombirthtoinfantageincorrelation normal newborns.vEEGallowstoproperlystudyandidentifyclinicalmanifestations information aboutbraincorticalfunctioningandevaluationofbackgroundactivityin Methods. Continuousv-Electroencephalography(vEEG)providesparticularlyuseful always efficaciousorpredictive. opioids NAS, in case of other psychotropic drugs and especially cocaine, the tool is not regimen, whenelevatedscoredarereported.WhereasFNASTissuccessfullyappliedfor Tool (FNAST) documents the infant withdrawal, andinitiate the appropriatetreatment pregnancy. Sinceforty decades,thestandardizedFinneganNeonatalAbstinenceScoring sudden discontinuationofexposuretopsychotropicdrugsabusedbythemotherduring Introduction. Neonatalabstinencesyndrome(NAS)inanewbornisresultofthe Abstract 4 3 2 Raffaele Falsaperla from cocaine in neonatalabstinencesyndrome Neurophysiological monitoring DOI: 10.4415/ANN_20_03_18 Ann IstSuperSanità2020|Vol. 56,No.3:390-396 Università diRoma, Viale Regina Elena366,00161Italy. E-mail: [email protected]. Address for correspondence: Simona Zaami, Dipartimento di Scienze Anatomiche, Istologiche, Medico-Legali e dell’Apparato Locomotore, Sapienza Università degliStudidiCatania,Italy Università diRoma,Rome,Italy Catania, Italy Dipartimento diScienzeMediche,SezionePediatria,UniversitàdegliStudiFerrara,Italy Sezione diNeuropsichiatriadell’InfanziaeAdolescenza,DipartimentoMedicinaClinicasperimentale, Dipartimento diScienzeAnatomiche,Istologiche,Medico-Legaliedell’ApparatoLocomotore,Sapienza UOC NeonatologiaeTerapia Intensiva,POSanMarcoAziendaO-UPoliclinicoVittorio Emanuele, Prenatal exposure to drugs abuse during pregnancy is Prenatal exposuretodrugsabuseduringpregnancyis 4 andLuigiMemo 1 , SimonaZaami 1 2 , MariaGiovannaAguglia (3.6%), andItalian(0.4%)pregnant women[3,4]. has beenreportedincohortsof Spanish(3.9%),Danish abuse andprescriptionpsychoactive drugsconsumption cisely, asignificantlynonnegligiblerate ofillicitdrug other thanopioidsmaybeequally high[2].Morepre- year andthenumberofpregnantwomenusingdrugs as many30000pregnantwomenusingopioidseach Similarly, inEurope,itisestimatedthattheremaybe 1 , CatiaRomano

• • • Key words

video-EEG syndrome abstinence neonatal cocaine intrauterine exposureto 3 , 391 EEC in cocaine NAS

Since gestational psychotropic drug consumption prove an eventual diagnosis of cocaine NAS. In this re- can be a serious hazard to the fetus, a careful moni- gard, it has to be said that, differently from what happens toring of drug use during pregnancy is crucial to assess with opioids, cocaine NAS presents a late onset and can- and eventually prevent prenatal exposure and provide not be immediately recognized and treated [5]. high quality obstetrical health care. Drug consumption In FNAST an important clinical sign to evaluate is the screening is essential to achieve identification of preg- motor paroxysmal that vary to tremor from seizures. In nant abusers and the most used screening method is the this concern, electroencephalography (EEG) was firstly note use of questionnaires. However, the use of standardized used in 1988 to examine neurologic and electroencepha- questionnaires presents some drawbacks, especially in lographic abnormalities in newborns prenatally exposed pregnancy period. Pregnant women may underestimate to cocaine [15]. Indeed, the EEG monitoring in neonates rief their consumption and/or are unwilling to disclose their is actually the unique diagnostic tool to well character- B habits during pregnancy due to fear of legal repercus- ize the motor phenomenon differentiating the onset of sions, guilt, memory biases. All these reasons, demon- startles from seizures and behavioural states anomalies. strated in several studies often render questionnaires Recently, Italian media highlighted several cases of unreliable to predict fetal drug exposure [5]. neonates presenting cocaine NAS in different cities The use of biomarkers in maternal and fetal matrices [16, 17], evidencing an epidemics of fetal exposure to (e.g. drugs and/or metabolites detection in maternal cocaine in the national territory, already evidenced in blood, urine, neonatal hair, meconium and breast milk) several other western countries [1, 18-20]. and clinical assessment by instrumental tools are funda- In this concern, local health professionals involved in mental to objectively assess gestational drug abuse and these upcoming cases may require updated information consequent prenatal exposure to them. Thus, biomark- and tools for differential diagnosis. ers measurement and instrumental tools use should al- Continuous video-electroencephalography (vEEG) ways complement questionnaires, as it has been shown provides particularly useful information about brain that self-report may underestimate prenatal consump- cortical functioning and evaluation of background ac- tion of substances of abuse [6-8]. tivity in normal newborns, allowing to properly study It has to be said that the valuable assessment of in and identify clinical manifestations as physiological mo- utero exposure to drugs of abuse can prevent its hazard- tor paroxysms, that disappear from birth to infant age ous consequences such as premature rupture of mem- in correlation with the neurological development [21]. branes, placental abruption, preterm birth, low birth- Continuous vEEG is also a useful tool in the intensive weight, small-for-gestational age, and admission to the care unit to define neurologic status and brain func- neonatal intensive care unit. Prenatal exposure to drugs tioning in critically ill neonates at high risk for adverse abuse may also result in neurodevelopmental problems neurologic injuries as hypoxic ischaemic encephalopa- as impairments of the growth and development of the thy, cerebral infections, to diagnose electroencephalo- brain and/or central nervous system and finally neona- graphic seizures and monitoring the responsiveness to tal abstinences syndromes with related neurological and anticonvulsivant treatment. Due to its feature to be a physiological consequences [9]. non- invasive tool, continuous vEEG monitoring could Neonatal abstinence syndrome (NAS) in a newborn be used to describe some clinical manifestations and as- is a result of the sudden discontinuation of fetal expo- sess if they can be correlated to possible injuries in criti- sure to substances that have been used or abused by the cal neonates as those exposed in utero to psychoactive mother during pregnancy [10]. drugs, such as cocaine, studying modified behavioural NAS has been described as a complex disorder that state and clinical neurological signs that onset in case primarily involves central nervous system, autonomic of neonatal cocaine abstinence syndrome in newborns, nervous systems and gastrointestinal system [11]. Clini- with the aim to exclude onset of epileptic phenom- cal signs typically develop within the first few days after ena and seizures in these babies. Neurological signs birth, although the timing of their onset, as well as their as tremors and numerous startles, constant irritability severity, can vary as a function of the drug abused by and anomalies in behavioural states and sleep and wake the pregnant woman [10]. In particular, NAS have been cycle characterize the onset of neonatal abstinence syn- most frequently reported in case of maternal use of opi- drome during subsequent days after birth. oids and cocaine [11, 12]. Fetal exposure to cocaine can be the cause, among The standardized Finnegan Neonatal Abstinence other signs, the onset of neurological signs as tremors, Scoring Tool (FNAST), ideated in 1975 by Loretta startles and irritability after birth. Tremors are motor Finnegan and subsequently modified, is used identify symptoms with muscular but not electroclinical corre- the withdrawal symptoms, document the infant’s with- late in vEEG monitoring. They could be very frequent drawal, and initiate the appropriate treatment regimen, during days after birth and have many extra cerebral when elevated scored are reported [13, 14]. Although causes. this scoring system has been extensively used for all the Startles are physiological motor manifestations that psychoactive drugs, Finnegan test was validated only for can be defined as “brainstem reflex originated in the NAS from opiates and up to now there are no validation bulbopontine reticular formation that gradually under- studies for other drugs of abuse such as cocaine. For this goes the inhibitory control of the corticospinal motor reason, the investigation of cocaine and metabolites in pathway”; we can consider as a basic alerting reaction in maternal and neonatal biological matrices and specific response to stress stimuli [22]. In NAS, tremors, irrita- instrumental and clinical tests on the newborn can im- bility and startles represent the excessive alerting reac- 392 Raffaele Falsaperla, Simona Zaami, Maria Giovanna Aguglia et al.

tion due to the effect of exposure to psycoactive drugs. by polygraphy/vEEG from the onset of neonatal absti- However, these neurological manifestations could hide nence syndrome to the time of hospital discharge and electroclinical seizure as brain injuries and need ad- sought to correlate the obtained observations with the equate treatments. Clinically, it is not possible to dif- clinical signs evaluated with FNAST. ferentiate electroclinical manifestations from motor signs. Video-EEG allows neurophysiological definition METHODS excluding an abnormal seizure pattern and to quantify Continuous vEEG monitoring allows to record elec- note the anomalies in behavioural states and sleep and wake trical signals of cortical cells, providing particularly use- cycle related to alerting reaction that characterize the ful information about electrical brain function. Cortical rief onset of neonatal abstinence syndrome. The rapid and electrical activity is defined background activity and it B continuous transition from quiet sleep to active sleep is related to EEG patterns typical for gestational age and wake can be define and objectively quantify with and behavioural state of newborns. Through vEEG nor- neurophysiological study. mal or abnormal electrical patterns changes that char- We here describe, just as an example, the case of a acterize the background activity in healthy or critical neonate with prenatal exposure to cocaine monitored newborns can be identified. Indeed, vEEG monitoring

Table 1 Comparative report of video-electroencephalography findings, Finnegan Neonatal Abstinence Scoring Tool and clinical manifes- tations Timing Finnegan Finnegan signs and symptoms EEG correlate to FCS EEG no correlate to AS/QS Score FCS At birth 0 - no symptoms or signs At 1 hour 5 - spontaneous mild tremors - high pitched cry At 4 hours 0 At 8 hours 0 At 12 hours 3 - spontaneous mild tremors At 16 hours 0 At 20 hours 2 - difficult feeding At 24 hours 6 - spontaneous mild tremors, - high pitched cry - excoriation At 48 hours 15 - sweeting - EDG - startles (AS/QS) normal AS/QS - spontaneous mild tremors - tremors without cortical - hiccups - polipnoic breath (FR > 60 b/min) correlate - grimacing - sleep < 2 hours after feeding - PMG (polipnoic breath) - tongue movements - high pitched cry - chewing - hyperactive Moro reflex - no other correlate - frequent arousal - excessive sucking - sneezing, regurgitation - regurgitation - suction artefacts At 52 hours 5 - mild tremors disturbed - excoriation At 72 hours 14 - sweeting - EDG - startles (QS/AS) normal AS/QS - mild tremors disturbed - motor artefacts without - oral movements - excessive sucking cortical activity - sleep < 1 hours after feeding - suction artefacts - excessive suction - increased muscle tone high pitched cry - excoriation - frequent yawning At 76 hours 10 - sleep after feeding < 1 h - increased muscle tone - excoriation - regurgitation At 80 hours 4 - excoriation - regurgitation - sleep after 3 h feeding Two weeks 1 - sleep after 3 h feeding - no motor artefacts - reduced normal AS/QS after birth - startles (QS/AS) - motor sucking artefact 393 EEC in cocaine NAS with associated polysomnographic data (breath, muscle Polygraphic pattern showed regular respiratory pattern. activity, heart rate, eyes movements) allows to define Regular respiration typical of QS is characterized by a behavioural state, distinguishing sleep from awake and variability between the slowest and fastest breaths is different phases of sleep (quiet or active) in relation to less than 20 breaths/min. AS phase was about 62%, QS electrical patterns changes. phase resulted about 34%, undetermined sleep phase Through vEEG electrical function showing normal was about 4%. And the first week after birth, sleep pat- or abnormal graphoelements (or named neonatal EEG tern resulted unchanged (AS 59% vs QS 32%). note features), electrical seizures with or without clinical cor- At the onset of NAS, at 48 hours of life, we recorded relate, rhythmic or periodic patterns can be defined. in an half hour of EEG study at the onset of NAS (at 48 The newborn, born with surgical delivery at 37 weeks hours of life) in active sleep phase and mainly in quiet rief of PMA, with regular adaptation (Apgar index 9/10), sleep 12 episodes of motor startles, some of them as- B weight 2870 g (50° centile), cranial circumference 33.5 sociated to grimacing, not correlated with cortical ictal cm (50° centile), length 48 cm (50° centile) has been activity, but with motor artifacts (Figure 1c and 1d ). admitted to Neonatal Intensive Care Unit (NICU) and Specifically, startles have been defined as physiologi- placed in continuous cardiac and respiratory monitor- cal clinical motor manifestations in quiet sleep phase ing. FNAST has been acquired because of the referred in healthy infant as “brainstem reflex that originates in maternal cocaine abuse during pregnancy. From the the bulbopontine reticular formation and that gradually first hour of life, the baby begun to show tremors and undergoes the inhibitory control of the corticospinal marked irritability, jitteriness and difficult consolability, motor pathway”; and hence they could be considered a so that wrapping, close feeding and reduction of senso- basic alerting reaction in response to stress stimuli [22]. rial stimulation was performed. We observed several episodes of startles (Figure 1a and Newborn was evaluated through clinical examination 1c) during first days after birth, respect following weeks and FNAST from birth every 4 hours to the disappear- from birth, with the continuous and sudden transition ance of symptoms and signs. Cardiac and respiratory from quiet to active sleep (figure active and quiet sleep) monitoring with continuous pulse oximetry was also that could define a sleep waking cycle disorder. These started. At 48 hours of life video-EEG with polygraphy, motor manifestations do not correlate to electroclinical EMG, ECG study lasting 90 min was initiated and we abnormal pattern. have revaluated at 72 hours and after two weeks, ac- Tremors and irritability have not a pathologic electric cording American Clinical Neurophysiology Society. correlate on cortical side; they are shown only as mo- At 48 hours after birth FNAST confirmed the onset tor artefact caused by movements and associated with of abstinence syndrome because of tremors, sweating, the onset of a regular rapid rhythm on electromyogra- excessive sucking. Toxicological urine test on the new- phy. The lack of cortical parossistic manifestations can born at 48 hours from birth confirmed the presence of exclude seizures and then the need of anticonvulsant high levels of cocaine metabolite as resulting of chronic treatment.Moreover the frenetic suction, as described maternal intake during gestation. Unfortunately, this re- in Finnegan Score, is valuable in polygraphy/vEEG sult cannot be confirmed in maternal biological matrices through the related electroclinical artefact. As demon- (mother hair or urine), since baby mother did not give strated in Figure 1e suction reflex is shown as a motor consent to any toxicological tests and once interviewed, artefact associated to the electrical manifestation due denied the use of cocaine or any other drug of abuse. to repetitive tongue movements (Figure 1e), without an Video-EEG with polygraphy, electrocardiogram electric correlate of seizures. (ECG), electromyography (EMG) and associated car- One week after, toxicological urine test and FNAST diac and respiratory monitoring was started and ac- resulted negative. The newborn showed only difficulties quired to define a complete sleep cycle, to study be- in feeding, normal sucking reflex. Motor phenomena, havioural states and define motor paroxysmal excluding previously defined as startles, disappeared. The vEEG onset of subclinical seizures. recording in sleep and wakefulness was repeated and A comparative report of video-EEG findings, FNAST showed a background characterized by continuous low and clinical manifestations is shown in Table 1. The voltage activity with interbursts of < 2sec and voltage of EEG tracing shows regular pattern characterized in 25 mcV in active sleep and alternating activity in quiet active sleep (AS) by continuous low-medium voltage sleep phase. The background appeared to be compat- (25-50 mcV pp) mixed frequency activity, theta waves ible with the correct chronological age of the newborn and slow delta activity in frontal and central regions, and motor artefacts and electrodermogram described anterior dysrhythmia, frontal sharp transients and ro- in previous EEG are decreased, sucking reflex was cor- landic activity in central regions (Figure 1a). The cor- related with motor suction artefact (Figure 1e). responding polygraphic pattern was characterized by irregular respiration and variable respiratory rate > 20 DISCUSSION breaths/min, typical of AS, wakefulness and transition NAS refers to the sequence of symptoms arising with sleep. In quite sleep (QS), the background EEG activ- the interruption of exposure to the substances of abuse ity was characterized by alternating activity, with slow taken by the mother during pregnancy. Initially valu- delta waves (50-50 mcV pp) alternate with periods of ated for abuse of opioids, the term has been extended lower voltage (25-50 mcV pp), lasting about 4-5 sec in to infants exposed to other drugs maternal abuse in frontal and central regions, theta activity in temporal prenatal age [11]. Finnegan scoring was introduced as regions, delta brushes in occipital regions (Figure 1b). “clinical” scoring systems to define neurological symp- 394 Raffaele Falsaperla, Simona Zaami, Maria Giovanna Aguglia et al. note

rief B

Figure 1 Electroencephalography tracing in a cocaine neonatal abstinence syndrome. Nikon Kohden Neurofax EEG 1200 was used. EEG re- Figure 1a cording has been started with the application on scalp of 9 cere- Startle in active sleep (motor artefact in electroencephalogra- bral electrodes, according to International 10-20 System, modified phy). for neonates. Extracerebral channels have been placed for analysis of polygraphic data. Electromyogram (EMG) has been recorded with electrode in right deltoid, respiration has been recorded through respiratory channel placed with sensor in the epigastric area. Electrocardiogram (ECG) has been recorded with sensor placed in left thorax side. Active Sleep: behavioral state characterized in healthy term neo- nate because the closure of eyes, intermittent periods of rapid eye movements, irregular breath, small and large body movements. On vEEG by a continuous low to medium voltage (25-50 mcV peak- to-peak) mixed frequency activity with a predominance of theta and delta and overriding beta activity. This activity is indistinguish- able from that of normal wakefulness.

Figure 1b Quiet Sleep: clinically characterized by eye closure, absence of rapid eye movements, occasional sucking activity or generalized myoclonic “startles.” EEG background, defined alternant, is char- acterized because of higher voltage bursts (50-150 mV pp), pre- dominantly of delta activity and lasting roughly 4 to 10 seconds, alternate with briefer, lower voltage (25-50 mV pp) interburst pe- riods composed mostly of mixed theta and delta activity. Figure 1d Startle: Rapid sequence of movements: grimacing and blinking, flexion of neck, trunk, hips and knees, arms adduction and fists clenching, in response to unexpected stimuli [22].

Figure 1e Figure 1c Sucking reflex: automatic stereotypic oral and tongue move- Startle in quiet sleep. startle has probably the meaning of a basic ments caused by brainstem-mediated reflex; Tremor: an involun- alerting reaction and appears during the quiet sleep phase; the tary, rhythmic oscillatory movement of equal amplitude around a EEG-trace shows motor artefact not associated to electric patho- fixed axis. Jitteriness: recurrent tremors. logic activity. 395 EEC in cocaine NAS

toms and signs and represents a rather subjective and zures after their initial month of life and a smaller sub- variable evaluation method that depends on the infant’s set continued even after 6 months of life, suggesting general conditions and usually correlates to course of long-term neurodevelopmental effects of early cocaine NAS [13]. exposure [26]. Several studies suggested that the development of the Even earlier, in 1988, Doberczak et al., focused on central nervous system could be affected by exposure to EEG patterns and evaluated that 90% of infants that cocaine in prenatal period and correlate with long-term prenatally exposed to cocaine displayed deficits in note neurodevelopmental effects in this cohort of infants neurophysiological behaviour, such as increased CNS [15]. Already in 1985, Chasnoff et al., noted increased irritability, and over 50% had abnormal EEGs. These tremulousness, more startle responses and deficient effects were transient for some of them, so that only rief interactive behavioural state organization in cocaine 20% of infants displayed abnormal EEGs by the second B exposed infants evaluated with Brazelton Neonatal Be- week, while only one of neonates identified in this study havioral Assessment [12]. displayed abnormal EEG patterns by 3 to 12 months of The study of the background and neurophysiological age. In particular EEG have shown multifocal anoma- patterns actually allows to define normal cerebral func- lies, lateralized anomalies and an asymmetric back- tion in healthy infants. Numerous studies demonstrated ground [15]. As a confirmation, in 2002 Scheret al., re- the usefulness of EEG monitoring in cocaine exposed ported that prenatal cocaine exposure affected reflexes, newborns for the characterization of neurophysiological motor maturity and autonomic stability in the newborn, anomalies and abnormal behavioural states. Indeed, in underlining the importance of electroencephalographic these infants, the reduced bioavailability of drug for in- (EEG) sleep patterns that can be used to assess cere- creased placenta catabolism near term or exposure ces- bral maturation and neurophysiologic organization of sation after delivery could determine altered behavioral the developing CNS [27]. state in postnatal EEG [23]. EEG sleep patterns could The transient neurological symptoms we have ob- be used to assess cerebral maturation and neurophysi- served which decreased with the passing of the time ologic organization of the developing CNS this cohort and disappeared at the second week of life could be of neonates. Clinical seizures in first days from birth, related to the manifestation of marked irritability and unrelated to structural brain damage have been report- jitteriness due to altered catecholaminergic pathway ed as an important complication in newborns exposed associated to the toxic effect of cocaine in first days to cocaine. [23]. after birth, as also Mirochnick et al., have previously Moreover, neonates with suspected NAS due to in described [28]. utero exposure to drugs show seizure-like clinical ac- In our case we also observed during active sleep tivity. The aim and the usefulness of FNAST is to de- phase other minor motor manifestations not described scribe the onset of SAN and define the severity of these in FNAST as oral and tongue repetitive movements, overexpressed motor manifestations (myoclonus dur- grimacing, that correlate to motor artifacts in absence ing sleep, tremors), physiological reflexes, seizure-like of cortical ictal activity. These motor phenomena are manifestation as jerking or rhythmic movement of ex- described as “the expression of brainstem release phe- tremities, apnoea, anomalies related to the involvement nomena, correlated to immaturity of the inhibitory con- of autonomic and central nervous system with a clinical trol” [18]. score. Sucking reflex that is valuated in FNAST, correlated However, excluding major neurological anomalies in EEG recording with muscular activity of masseter in due to hypoxic ischemic encephalopathy, metabolic temporal leads and slowing waves in frontal derivations disorders that cause seizures, the electrical abnormal due to repetitive tongue movements as described. These activity related to these manifestations can be excluded motor automatic stereotypic and repetitive movements or define only with vEEG monitoring [24]. could hide misunderstood subtle seizures, underling the Video-EEG monitoring can be a valid tool for iden- usefulness of vEEG analysis. Interpretation of neonatal tification of seizures-like episodes that can appear in a EEG is still challenging and background activity is fre- baby exposed to drugs or to define activity as motor quently intermixed with physiological artefacts, such as phenomena evaluated in FNAST. Video-EEG also al- ocular, muscle and movement artifacts, as in NAS [27, lows to define motor manifestation that FNAST does 29]. not analyze, always with the aim to exclude onset of In conclusion, whereas FNAST is essential to neonatal subclinical seizures or brain injuries in ex- promptly define the clinical presentation of newborn posed cocaine newborns. exposure to drugs and the onset of NAS, vEEG record- In fact, prenatal cocaine exposure is also associated ing could be considered an important and objective tool with increased risk of neonatal seizures and anomalies that allows the evaluation of behavioural state, clinical in brain wave activity in newborns [25]. and neurological signs in newborns in utero exposed to In agreement with this, early in 1990 Kramer et al., psychoactive drugs, excluding epileptic phenomena and observed that all the infants prenatally exposed to co- defining if exposure has determined EEG background caine experienced seizures within 36 hr of delivery with disturbances, to establish also neurological outcome in 50% of these infants showing repetitive seizures during these cohorts of neonates. their neonatal hospitalization stay, but did not have re- currences during a follow-up period of 4 to 12 months. Disclosure The other 50% continued to experience neonatal sei- None declared 396 Raffaele Falsaperla, Simona Zaami, Maria Giovanna Aguglia et al.

Conflict of interest statement Received on 17 February 2020. None declared Accepted on 29 May 2020.

REFERENCES note

1. Carlier J, La Maida N, Di Trana AG, Huestis MA, Pi- intrauterine cocaine exposure. J Pediatr. 1988;13(2):354- chini S, Busardò FP. Testing unconventional matrices 8. to monitor for prenatal exposure to heroin, cocaine, 16. Mento M. Positività alla cocaina, succede in alcuni os- rief amphetamines, synthetic cathinones, and synthetic opi- pedali italiani e le vittime sono dei neonati. Available B oids. Ther Drug Monit. 2020;42(2):205-21 doi: 10.1097/ from: www.newnotizie.it/2019/11/03/positivita-alla-co- FTD.0000000000000719 caina-neonati-coinvolti/. 2. Gyarmathy VA, Giraudon I, Hedrich D, Montanari L, 17. Bocci M. I medici: “Emergenza neonati in crisi di Guarita B, Wiessing L. Drug use and pregnancy – chal- astinenza”. Available from: www.repubblica.it/crona- lenges for public health. Eur Surveill. 2009;14(9):33-6. ca/2019/11/03/news/i_medici_emergenza_neonati_in_ 3. Cortes L, Almeida L, Sabra S, Muniesa M, Busardo FP, crisi_di_astinenza_-240102649/. Garcia-Algar O, Gomez-Roig MD. Maternal hair and 18. Joya X, Gomez-Culebras M, Callejón A, Friguls B, Puig neonatal meconium to assess gestational consumption C, Ortigosa S, Morini L, Garcia-Algar O, Vall O. Co- and prenatal exposure to drugs of abuse and psychoac- caine use during pregnancy assessed by hair analysis in tive drugs. Curr Pharm Biotechnol. 2018;19(2):136-43. a Canary Islands cohort. BMC Pregnancy Childbirth. doi: 10.2174/1389201019666180405163612 2012;12:2. doi:10.1186/1471-2393-12-2 4. Rausgaard NL, Ibsen IO, Jørgensen JS, Lamont RF, Ravn 19. Joya X, Fríguls B, Simó M, Civit E, de la Torre R, P. Prevalence of substance abuse in pregnancy among Palomeque A, Vall O, Pichini S, Garcia-Algar O. Acute Danish women. Acta Obstet Gynecol Scand. 2015; heroin intoxication in a baby chronically exposed to 94(2):215-219. doi: 10.1111/aogs.12528 cocaine and heroin: a case report. J Med Case Rep. 5. Chiandetti A, Hernandez G, Mercadal-Hally M, Al- 2011;5:288. doi: 10.1186/1752-1947-5-288 varez A, Andreu-Fernandez V, Navarro-Tapia E, Bas- 20. García-Algar O, Felipe A, Puig C, Monleón T, Pacifici R, tons-Compta A, Garcia-Algar O. Prevalence of pre- Pichini S. Fetal exposure to cocaine and infections during natal exposure to substances of abuse: questionnaire the first 18 months of life in infants from a Mediterranean versus biomarkers. Reprod Health. 2017;14(1):137. doi: city. Acta Paediatr. 2007;96(5):762-3. 10.1186/s12978-017-0385-3 21. The American Clinical Neurophysiology Society. Stan- 6. Joya X, Marchei E, Salat-Batlle J, García-Algar O, Cal- dardized EEG terminology and categorization for the varesi V, Pacifici R, Pichini S. Drugs of abuse in mater- description of continuous EEG monitoring in neonates: nal hair and paired neonatal meconium: an objective as- Report of the American Clinical Neurophysiology Soci- sessment of foetal exposure to gestational consumption. ety Critical Care Monitoring Committee. J Clin Neuro- Drug Test Anal. 2016;8(8):864-8. physiol. 2013;30:161-73. 7. Joya X, Pacifici R, Salat-Batlle J, García-Algar O, Pichini 22. Facini C, Spagnoli C, Pisani F. Epileptic and non-epi- S. Maternal and neonatal hair and breast milk in the as- leptic paroxysmal motor phenomena in newborns. J sessment of perinatal exposure to drugs of abuse. Bio- Matern Fetal Neonatal Med. 2016;29(22):3652-9. doi: analysis. 2015;7(10):1273-97. doi: 10.1002/dta.1921 10.3109/14767058.2016.1140735 8. Pichini S, Garcia-Algar O, Alvarez A, Gottardi M, Mar- 23. Legido A, Clancy R, Spitzer A, Finnegan L. Electroen- chei E, Svaizer F, Pellegrini M, Rotolo MC, Pacifici R. cephalographic and behavioral state studies in infants of Assessment of unsuspected exposure to drugs of abuse cocaine-addicted mothers. AJDC. 1992;146:748-52. in children from a Mediterranean city by hair testing. Int 24. Palla MR, Khan G, Haghighat ZM, Bada H. EEG find- J Environ Res Public Health. 2014;11(2):2288-98. doi: ings in infants with neonatal abstinence syndrome pre- 10.3390/ijerph110202288 senting with clinical seizures. Front. Pediatr. 2019;7:111. 9. Dodge NC, Jacobson JL, Jacobson SW Effects of Fetal doi: 10.3389/fped.2019.00111. Substance Exposure on Offspring Substance UsePedi- 25. Martin M, Graham D, McCarthy D, Bhide P, Stanwood atr Clin North Am. 2019;66(6):1149-61. doi: 10.1016/j. G. Cocaine-induced neurodevelopmental deficits and pcl.2019.08.010 underlying mechanisms. Birth Defects Res C Embryo To- 10. Kocherlakota P. Neonatal abstinence syndrome. Pediat- day. 2016;108(2):147-73. doi: 10.1002/bdrc.21132 rics 2014;134(2):e547-61. doi: 10.1542/peds.2013-3524 26. Kramer L, Locke G, Ogunyemi A, Nelson L. Neonatal 11. Stover MW, Davis JM. Opioids in pregnancy and neonatal cocaine-related seizures. J Child Neurol. 1990;5:60-4. abstinence syndrome. Semin Perinatol. 2015;39(7):561- 27. Scher M, Richardson G, Day N. Effects of prenatal co- 5. doi: 10.1053/j.semperi.2015.08.013 caine/crack and other drug exposure on electroencepha- 12. Chasnoff IJ, Burns W, Schnoll SH, Burns K A. Cocaine lographic sleep studies at birth and one year. Pediatrics. Use in Pregnancy. New Engl J Med. 1985;313(11):666- 2000;105(1 Pt 1):39-48. doi: 10.1542/peds.105.1.39 9. 28. Mirochnick M, Meyer J, Cole J, Herren T, Zucker- 13. Finnegan LP, Kron RE, Connaughton JF, Emich JP. As- man B. Circulating catecholamine concentrations in sessment and treatment of abstinence in the infant of the cocaine-exposed neonates: A pilot study. Pediatrics. drug-dependent mother. Int J Clin Pharmacol Biopharm. 1991;88(3):481-5. 1975;12(1-2):19-32. 29. Falsaperla R, Marino SD, Aguglia MG, Cupitò G, Pisani 14. Jansson LM, Velez M, Harrow C. The opioid exposed F, Mailo J, Suppiej A. Developmental EEG hallmark or newborn: Assessment and pharmacologic management. J biological artifact? Glossokinetic artifact mimicking an- Opioid Manag. 2009;5(1):47-55. terior slow dysrhythmia in two full term newborns. Neu- 15. Doberczak TM, Shanzer S, Senie RT, Kandall SR. Neo- rophysiol Clin. 2019;49(5):377-80. doi: 10.1016/j.neu- natal neurologic and electroencephalographic effects of cli.2019.07.015 Federica NapolitaniCheyne Edited by B Ann IstSuperSanità2020 DOI: 10.4415/ANN_20_03_19 cancer outcome. Tobacco products (chapter2.1),in- tion betweenlife styleandenvironmentalfactors with effective cancerpreventionstrategies. variation incancerincidence ratesiskeyfordesigning income countries.Theunderstanding ofthecauses ing in most higher-income countries, but not in lower- medium HDI.Moreover, cancermortality is declin- will havethegreatestimpactsoncountrieswithlowand (HDI), thepredictedincreasesincancerburden in countrieswithveryhighhumandevelopmentindex though thecancerincidenceburdeniscurrentlyhighest of cancerbetweencountriesovertimeisobserved.Al- sue. Agreatvariationofincidenceratesforseveraltypes that cancerisamongthemostrelevantpublichealth- 2016, almost30%wereduetocancer, thusconfirming deaths from noncommunicable diseases worldwide in pact onhumanhealth.Ofthe15.2millionpremature of thecontentthesesections. equalities andcancer. BelowIprovideabriefsummary gies. Moreover, onesection(section4)isfocusedonin- be conveyedintoimplementationofpreventionstrate- tion on statistics, causes and mechanisms of cancer can structured toguidethereaderonhowallinforma- diseases andtopromotepublichealth. is theonlywaytodecreaseburdenofalltypes vention istackledbutweareallawarethatprevention tion. BecauseofthespecificgoalIARC,cancerpre- ISS, ItalianNationalInstituteofHealth),i.e.preven- mission ofourInstitute(IstitutoSuperiorediSanità, focused onatopicthatishighlyrelevantalsoforthe investigators worldwide.Inparticular, thisnewWCRis lection ofpeerreviewedpaperspreparedbyrecognized made byIARCoranyotherauthoritybutisacol- its characteristicssinceit is notbasedon assessments cently published. This five-year publication is unique in Agency forResearchonCancer(IARC)hasbeenre- cer research for cancer prevention” by the International Section 2givesadetaileddescription ofthecorrela- Section 1providesanupdatedpictureofcancerim- This volume, which isorganized in sixsections,is The 2019World CancerReport(WCR)entitled“Can- ook R eviews

CC BY-NC-ND 3.0IGO. Licence: publications.iarc.fr/586. Available from: http:// Cancer; 2020. forAgency Research on Lyon, France: International BW,Stewart editors. Wild CP, Weiderpass E, prevention Cancer research for cancer REPORT WORLD CANCER |

Vol. 56,No.3:397-398 , N , otes

a n d C world. being increasinglyappliedby agenciesthroughoutthe in establishingbiological plausibility. Thisapproach is data fromstudiesinhumans arelackingandcanhelp provide independentevidence ofcarcinogenicitywhen (chapter 3.11).Dataonthe keycharacteristicscan recently applied in IARC Monographs evaluations the key characteristics of human carcinogens have been studies. To stresstherelevanceofmechanisticevidence, ising, buttheyhaveyettobeevaluatedinprospective prevention strategies (chapter 3.10)appearalso prom- the fight against cancer. Microbiota-targeted cancer system hasthepotentialtobecomenewfrontierin microenvironment (chapter3.9).Targeting the immune Immune cellsareessentialcomponentsofthetumour can beausefultoolforcancerpreventionstrategy. well as the analysis of epigenetic changes (chapter 3.8) Metabolomics ofhumanbiospecimens(chapter3.7)as tant contributorstoseveralcancersites(chapter3.6). productive andhormonalfactorsappeartobeimpor ment ofasignificantsubgrouphumancancers.Re- immunotherapy, are presentedaspromiseinthetreat- DNA repairdefects,eitherbysyntheticlethalityor ment. Precision therapies targeted to cancer-specific tion factors(chapter3.5)canmediatecancerdevelop- constitutive activationofpro-inflammatorytranscrip- pathways controlling genetic stability(chapter 3.4) or tational signature)(chapter3.3).Faultyregulationof environmental exposures(e.g.skincancerandUVmu- mutations incancertypesthatcorrelatewithdistinct eration sequencinghasallowedtoidentifyhighdensity etrance gene variants(chapters 3.1 and3.2).New gen- agents andweakgeneticsusceptibilityduetolow-pen- endogenous agentsortheinteractionbetweenthese tions. Theremainingcancersareduetoenvironmental/ mated tobeduehighlypenetrantinheritedmuta- tion andtherapy. About5-10%ofallcancersareesti- of themechanismscarcinogenesisforcancerpreven- logical andregulatorymeans. being amenabletointerventionsbybehavioral,techno- a preventablecontributortotheglobalcancerburden good newsisthatexposuretomostofthesefactors drugs (2.11)areallpotentialcancerriskfactors.The occupational carcinogens(2.10)andhormone-based polluted waterfromcontaminatedsoils(chapter2.9), environmental pollutantsincludingairpollutionand ity (chapter2.7),dietarycontaminants (chapter 2.8), overweight/obesity (chapter 2.6),lackofphysical activ- dietary patterns particularly those associated with tion (chapter2.4)andionizingradiation2.5), tion (chapter2.3),physicalagentssuchasUVradia- papilloma virus(HPV)(chapter2.2),alcoholconsump- fectious agentssuchasHelicobacterpyloriandhuman Section 4deals withinequalitiesthataffect cancer Section 3showshowimportantistheunderstanding omme n ts - 397 Book Reviews, Notes and Comments 398 Book Reviews, Notes and Comments have not been identified, its prevention must focus on For example,sinceexogenouscausesofprostatecancer cording tothetumourtypeorsubtypebeingconsidered cancer causationandpreventionmustbequalifiedac- The generaltakehomemessageisthatknowledgeof and leukaemia(chapter5.1)arespecificallyaddressed. (chapter 5.18),Non-Hodgkinlymphoma5.19) nervous systemtumors(chapter5.17),thyroidcancer cancer (chapter5.16),brainandotherprimarycentral (chapter 5.14),bladdercancer5.15),kidney 5.12), prostatecancer(chapter 5.13), testicularcancer trial cancer(chapter5.11),ovarian (chapter 5.9), cervical cancer (chapter 5.10), endome- (chapter 5.7),skincancer5.8),breast ter 5.5), liver cancer (chapter 5.6), pancreatic cancer stomach cancer(chapter5.4),colorectal(chap- cer (chapter5.2),oesophagealcancer5.3), types. Lungcancer(chapter5.1),headandneckcan- by thelegacyofcolonizationandracism(chapter4.7). social, political,andeconomicenvironmentsaswell enous peoples are significantly affected by the broader cancer burdenand,moregenerally, thehealthofIndig- standing, persistentcauseofinequity(chapter4.6).The ties incancerpreventionservicestheUSAisalong- cer screeningwillreducehealthinequalities.Dispari- that achievingrelativelyhighparticipationratesincan- pean Unioncountries(chapter4.5).Researchshows present betweenscreeningpracticesfollowedinEuro- and controlservices(chapter4.4).Basicdifferencesare to significantdisparitiesinaccesscancerprevention rise tovastsocioeconomicchangesand,consequently, The rapid economic development of India has given ization, andtheageingofpopulation(chapter4.3). nities alongwithrapideconomic development, urban- and dietarypatternsbetweenurbanruralcommu- could bepartiallyexplainedbydifferencesinlifestyles incidence andspectrumofcancertypesinChinathat oped countries.Thereisanurban-ruraldifferenceinthe tality rates 3-fold higher than those for women in devel- incidence rates2-foldhigherandcervicalcancermor where theseservicesarerare,presentcervicalcancer in Africa(chapter4.2).Women indevelopingcountries, tion, andtreatmentonitsincidencemortalityrate show theimpactofservicesforprevention,earlydetec- preventable diseasesuchascervicalcanceristakento within certaincountriesisdescribed.Theexampleofa ter 4.1).Theimpactofthesefactorsoncancerburden financial incomeorspecific livingenvironments(chap- terplay ofmanyfactors,includingeducationalstatus, prevention. Cancerinequalitiesaredrivenbythein- Section 5 deals with prevention of particular tumor - WCR, namelychapter2.9 obesity, and physicalinactivity(chapter6.9). municable diseasesshouldfocusontobacco,alcohol, pollutants). Preventionstrategiescommontononcom- of tobaccouseandeliminationpersistentorganic hazard-based regulation(examplesincludereduction suspected carcinogens(chapter6.8)particularlyby been effectiveinreducingexposuretoknownand stage disease(chapter6.7).Governmentalactionhas worse outcomeinpatientswithbothearly-andlate- cells ishighlypredictiveofmetastaticoutgrowthand early-stage disease.Thepresenceofcirculatingtumour tant opportunityforcancerreductionbydiagnosisof with screeningprogrammes(chapter6.6)areanimpor control. Emerginggenomictools(chapter6.5)together detection ofcancerisacriticalcomponent creased throughchemoprevention(chapter6.4).Early breast cancer, deathsfromsporadiccancermaybede- 6.3). As shown by the use of anti-estrogenic agents in tive forliverandcervicalcancer, respectively(chapter Vaccination againsthepatitisBvirusandHPViseffec- of canceraredirectlyassociatedwithvariousinfections. some low-incomecountries,uptoonethirdofallcases key factorsinthedesignofprogrammesandpolicies.In to nutrition,exercise,andweightgain(chapter6.2)are Interventions tochangebehavior(chapter6.1)related more thandoublethechanceofsuccessfullyquitting. tobacco cessationispresented.Cessationsupportcan communities andcountries.TheWHOperspectivefor multiple differentcancertypescanbedecreasedinall prevention strategies.Theburdenofdeathfromthe particular cancersites. dures can be meaningfully explored only with respect to detection ofprecancerouslesionsandscreeningproce- nisms ofcarcinogenesis. in thefieldofenvironmentalepidemiologyandmecha- worldwide of the research conducted in our Institute searchers ofourInstitute.Thistestifiestherecognition sources ofdamageandhereditarysyndromes” pair andgeneticinstability. Endogenousandexogenous lenge istocharacterizespecificrisks” “Contamination ofair, water, soil,andfood.Thechal- Finally, itis ofnotethattwochaptersthe2019 Section 6presentsthebasisfor, andoutcomesfrom, 2 1 Eugenia Dogliotti,MargheritaBignami“DNAre- Pietro Comba, Ivano Iavarone, Manolis Kogevinas Istituto SuperiorediSanità,Rome,Italy 1 and3.4 [email protected] 2 , seeasauthorsre- Eugenia Dogliotti - DOI: 10.4415/ANN_20_03_20 is basedonthree basicconcepts:first,babies havethe Free, Stay AIDSFreeframework.This framework countries approach the UN 2020 targetsofthe Start gains, newdevelopments and remainingchallenges as AIDS 2020;50p.Thisprogress reportreviewsrecent Geneva: JointUnited Nations Programme on HIV/ Start FreeStayAIDS Free-2020report. of cereals. and FoodSituation,especiallyregardingthecoverage another major GIEWS publication, Crop Prospects region. FoodOutlookmaintainsaclosesynergywith exports fromtheLatinAmericanandCaribbean the outlookforbanana,avocadoandothertropicalfruit global impactofthespreadAfricanSwineFeverand sues. Thecurrenteditionalsohasspecialreportsonthe well as meat and includes feature articles on topical is- ties, includingcereals,fish,sugar, oilcropsandmilkas short-term forecasts for a wide array of food commodi- markets. Eachreportprovidesmarketassessmentsand cusing ondevelopmentsaffectingglobalfoodandfeed 131932-1. of theUnitedNations2019;104p.ISBN978-92-5- Markets. Rome:FoodandAgricultureOrganization Food Outlook-BiannualReportonGlobal assimilated. that canbecontinuallyupdatedasmoreknowledgeis landscape, butalsowillhelpbuildresilientfoodsystems pating futurechallengesinashiftingglobalfoodsafety scanning andforesight,whichwillnotonlyaidinantici- benefits offorward-lookingapproachessuchashorizon methylmercury. There isalsoadedicatedsectiononthe cides, mycotoxins,andheavymetalswithemphasison pathogens andparasites,harmfulalgalblooms,pesti- hazards considered in the publication are foodborne are associated with climate change. The food safety some currentandanticipatedfoodsafetyissuesthat pose of this report is toidentify and attempt to quantify tions 2020; 176 p. ISBN 978-92-5-132293-2. The pur Food andAgricultureOrganizationoftheUnitedNa- safety. FoodsafetyandqualityseriesNo.8.Rome: Climate change: Unpacking the burden on food Annarita Barbaro Edited by on P Ann IstSuperSanità2020 NATIONS (FAO) ORGANIZATION OFTHEUNITED FOOD ANDAGRICULTURE AIDS (UNAIDS) UNITED NATIONS PROGRAMMEONHIV/ ublications P ublic Food Outlookisabiannualpublicationfo- H

ealt |

Vol. 56,No.3:399-401 from h

I nternational - learned fromtheCOVID-19pandemic. the progressthathasbeenlostwhileusinglessons HIV epidemicwillrequireclearlyfocusingonregaining plans foraddressingchildrenandadolescentswithinthe pandemic willpotentiallyincreasethesegaps.Future met andservicedisruptionscausedbytheCOVID-19 despite muchprogress,mostofthetargetswillnotbe The reportrecognizesthat,withlessthanayeartogo, a few months after the end date of December 2020. its finalyear, butfinalprogresswillnotbeknownuntil they canremainAIDSfree.In2020,theinitiativeisin ner, with access to optimal antiretroviral therapy so that identified, treated,andcaredforitinatimelyman- adolescents whodoacquireHIVhavetherighttobe the righttostayfreeofvirus;third,childrenand vention, children, adolescents, andyoungwomenhave vertical transmissionofHIV; second,throughHIVpre- right toentertheworldfreeofHIVbyeliminating change; second, improvingtheresilienceof livelihoods; first, enabling people to adapt to the impacts of climate ed by examples from across the world, in three areas: address thesechallenges.It outlines actions,support- opment. It also serves as a guide for concrete actions to and climate change inthe context ofsustainable devel- Report addressesthecritical linkages betweenwater 2020editionofthe World Water4. The Development SCO Publishing 2020; 219 p.ISBN: 978-92-3-100371- port 2020:waterandclimatechange. The United Nations world water development re- specific activitiesmentionedin thechecklists. nexes which providedetailedguidanceand resources for tors before and after schools reopen and sec- Evaluation, Education and Child Protection and Health,Nutrition,WASH, Monitoring key actions for guide isdividedintotwosections: inclusive, all-hazardsapproachtoschool reopening. The tool thatoutlinesthekeystepsneededfor a coordinated, aimstobeauser-friendly, practitioner-focused guide cerns exacerbatedbyCOVID-19schoolclosures. This protectioncon- address thelearninginequalitiesand spacesand transmission ofCOVID-19within learning tries and school communities must minimize the riskof - sites, minis pare to reopen schools and other learning UNESCO Publishing2020;24p. As governments pre- Safe back to school: a practitioner’s guide. USA: ORGANIZATION (UNESCO) SCIENTIFIC ANDCULTURAL UNITED NATIONS EDUCATIONAL,

O rganizations Checklists with lists of Technical an- Paris: UNE-

399 Publications from International Organizations 400 Publications from International Organizations vices forallinasustainablemanner. used, whileprovidingwatersupplyandsanitationser improving howwaterresourcesaremanagementand improved resilience–thatcanbeaddressedthrough climate change–intermsofadaptation,mitigationand the challenges,opportunitiesandpotentialresponsesto the hydrologicalcycle.Rather, itismeanttofocuson nical examinationoftheimpactsclimatechangeon resources. Thereportisnotmeanttobeapurelytech- es inthewayweuseandreuseEarth’s limitedwater drivers ofclimatechangewillrequiresubstantialchang- report concludes that reducing both the impacts and and, third,reducingthedriversofclimatechange.The the contractionofjobsinmiddle-skilloccupations. ates, whoselabormarketperspectivesarechallengedby dle-education vocational,educationandtraininggradu- examines thechanginglabormarketoutcomesformid- of jobsinmiddle-skilloccupations.Finally, Chapter5 fresh lookatjobpolarization,andthehollowingout dismissals, andenforcementissues.Chapter4takesa ment ofregulationsforcollectiveredundancies,unfair EPL indicators,whichnowincludeanimprovedassess- countries bydevelopinganewversionoftheOECD’s ployment protectionlegislation(EPL)acrossOECD curity. Chapter 3 provides a comparative review of em- ficult balancebetweenworkincentivesandincomese- in part-timeandlessstablejobsdiscussesthedif- uneven accesstounemploymentbenefitsforworkers the resultingeconomiccrisis.Chapter2investigates market consequencesoftheCOVID-19outbreakand Chapter 1 provides an initial assessment of the labor cuses on worker security and the COVID-19 crisis. 2020 editionoftheOECDEmploymentOutlookfo- Publishing 2020; 339p. ISBN 978-92-64-99828-5. The (print) ISBN978-92-2-032405-9(webPDF).Thepur Labour Office,2020;46p.ISBN978-92-2-032404-2 beyond. Apracticalguide.Geneva:International Teleworking duringtheCOVID-19pandemicand curity andtheCOVID-19 Crisis. Paris: OECD EmploymentOutlook2020:Worker Se- policymakers have been handling teleworking during the policymakers have beenhandlingteleworking duringthe number ofcase examplesregardinghowemployers and tions, themainissuesforconsiderations, andpresentsa The guide presents 8 focus areas, offers relevant defini- update theirownteleworking policiesandpractices. prises andpublicsectororganizations candevelopor flexible frameworkthrough which bothprivateenter makers inupdatingexistingpolicies, and toprovidea plicable toabroadrangeofactors,supportpolicy- recommendations foreffectiveteleworkingthatareap- pose ofthisguideistoprovidepracticalandactionable ORGANIZATION (ILO) INTERNATIONAL LABOUR OPERATION ANDDEVELOPMENT(OECD) ORGANISATION FORECONOMICCO- OECD - - - al andorganizationalperformance. ance, occupationalhealthandwell-being,individu- telework anditseffectsonworkingtime,work-lifebal- and companycasestudiesregardingtheincidenceof forms andeffects.Italsoanalyseslarge-scalesurveys regarding thedevelopment of telework,itsvarious India, Japan, the United States, and ten EUcountries It reviewsnationalexperiencesfromArgentina,Brazil, explaining the evolution of telework over four decades. premises. Thisbookoffersanewconceptualframework ing ICTs toperformworkfromoutsideofanemployer’s and alsoanevolutionoftelework,broadlydefinedasus- ical developmentshaveenabledadramaticexpansion Office 2019;352p. by JonC.Messenger. Geneva:InternationalLabour Perspective. Telework inthe21stCentury. AnEvolutionary relevant conventionsandrecommendations. alongside internationallabourstandardslaiddownin were designedtorespondthepublichealthcrisis, and referencestonationalgovernmentpolicies,which guide alsocontainsalistofavailabletoolsandresources of teleworkingarrangementsbeyondthepandemic.The from therecentmonthsthatarerelevantforfuture COVID-19 pandemicaddressingthelessonslearned against children, andincludeschildmaltreatment, bul- sonal violence that accounts for most acts of violence children. Indetail,thisreport focusesontheinterper evidence-based strategies for ending violence against against childrenthroughthelens ofthesevenINSPIRE Development Goals(SDG) targets onendingviolence in implementing activities to achieve the Sustainable report explorestheprogressthatcountrieshavemade tronic version)9789240004207(printversion).This ganization Geneva:World HealthOr against children 2020. Global statusreportonpreventingviolence tobacco andalcohol. World CancerReportaswellWHOglobalreportson able development.ThisreportcomplementstheIARC interventions canyieldvalueformoneyandsustain- vices. Pragmaticdecision-makingandevidence-based high-impact investmentsinprograms,policiesandser sage thatcancercanbecontrolledthroughstrategic, trophe, uncertainty, andresignationtoapositivemes- cancer fromonebasedonfearofdeath,financialcatas- returns. Thereportseekstochangethedialogueon an investment,withsubstantialhumanandeconomic presents newevidenceofthevaluecancercontrolas tools, andcurrentprioritiesincancercontrol.Italso (print version).Thisreportintroducestheprinciples, 9789240001299 (electronicversion),9789240001305 neva: World Health Organization 2020; 149 p. ISBN Ge- vesting wiselyandprovidingcareforall. WHO reportoncancer:setting priorities, in- WORLD HEALTH ORGANIZATION (WHO) 2020; 332p.ISBN9789240004191(elec- The ILOFutureofWork series.Edited ISBN:9781789903744. Technolog- - - - epidemic” by 2030, backed up by concrete milestones epidemic” by2030,backedupconcretemilestones Member Stateshavecommittedto“endingtheglobalTB global, regional,andcountry levels. LeadersofallUN in ordertoprogresstheresponseepidemic,at hensive andup-to-dateassessmentoftheTBepidemic since 1997withthepurposeofprovidingacompre- a global tuberculosis (TB) report every year published 156571-4. TheWorld Health Organization(WHO)has Health Organization2019;284p.ISBN978-92-4- Global tuberculosisreport2019.Geneva:World not coveredbythereport. three forms of violence – these forms of violence are and terrorism)–similarrootcausesunderlieall and thelikelihoodofcollectiveviolence(includingwar self-directed violence (including suicide and self-harm) personal violencecanincreasetheriskforsubsequent partner violence.Althoughchildhoodexposuretointer lying andothertypesofyouthviolence,intimate - 32% andchildren(aged<15years)for11%. TB casesin2018.Bycomparison,womenaccountedfor 57%ofall in men(aged≥15years),whoaccounted for both sexesinallagegroups,butthehighestburdenis TBaffectspeopleof global averagebeingaround130. new casesper100000populationyear, withthe among countries,fromfewerthanfivetomore500 recent years.Theburdenofdiseasevariesenormously TB in 2018, a numberthathas been relatively stablein 10.0 million(range,9.0–11.1million)peoplefellillwith Globally, anestimated estimated numberofTBcases. count formorethan99%oftheworld’s populationand were reportedby202countriesandterritoriesthatac- maintained byothermultilateralagencies.In2019,data WHO inannualroundsofdatacollection,anddatabases made. Thereportisbasedprimarilyondatagatheredby and increasedfinancingallindicatethatprogressisbeing and deaths,improvedaccesstoTBpreventioncare and targets. Globalindicatorsforreductionsin TB cases 401 Publications from International Organizations Ann Ist Super Sanità 2020 | Vol. 56, No. 3

Instructions to Authors

Annali dell’Istituto Superiore di Sanità is a peer reviewed quarterly science journal which publishes research articles in biomedicine, translational research and in many other disciplines of the health sciences. The journal includes the following material: original articles, reviews, commentaries, editorials, brief and technical notes, book reviews. The publication of Monographic Sections on Annali ISS has been discontinued. In case you wish to present a limited number of coordinated contributions on specific themes concerning priorities in public health, please contact the Editorial office.Annali follows the Recommendations for the Conduct, Reporting, Editing, and Publications of Scholarly Work in Medical Journals, issued by the International Committee of Medical Journal Editors (ICMJE) www.icmje.org.

MANUSCRIPT SUBMISSION Results, Discussion, Conclusions). In the Methods sec- Manuscripts should be submitted online to www.annali-iss.eu. tion a specific paragraph on the adopted statistical analy- The submission should include: sis should necessarily be included. • cover letter where the authors declare that the manuscript Each article should be accompanied by: has not been published or submitted for publication elsewhere; • an abstract of about 150 words; the abstract should be • manuscript; structured when required (such as in original articles); • tables and figures; • key words up to a maximum number of five (MeSH • author’s contribution statement (individual contribution headings, whenever possible. Refer to: www.nlm.nih.gov/ to the manuscript); mesh/meshhome.html). • conflict of interest statement (a conflict of interest exists Tables and figures should be kept to a minimum and be when authors or their institutions have financial or personal presented only if necessary. relationship with other people or organizations that could Authors should deal responsibly and effectively with se- inappropriately bias conduct and findings of the study); curity issues that might be raised by their papers (see: • permission to reproduce figures, if appropriate. Receipt of author’s paper will be acknowledged by an e- Statement on Scientific Publication and SecurityScience mail containing an identification number which should be 2003;299:1149). used in future correspondence. This journal has adopted the SAGER reporting Guide- lines for Sex and Gender Equity in Research. REVIEW PROCEDURE These guidelines apply to original research articles and re- Each paper submitted to Annali is subjected to the fol- view papers. Authors should use the terms sex and gender lowing procedures: carefully in order to avoid confusing both terms. Where • it is reviewed by the Editor-in-Chief for general suitability; subjects can also be differentiated by gender (shaped by • if it is judged suitable, qualified referees are selected and social and cultural circumstances), the research should be a peer review process takes place; conducted similarly at this additional level of distinction. • based on the recommendations of the referees and re- Where the subjects of research comprise organisms capa- plies of the authors, the Editor-in-Chief decides whether ble of differentiation by sex, the research should be de- the article should be accepted, modified or rejected; signed and conducted in a way that can reveal sex-related • once the paper has been accepted, authors will receive differences in the results, even if these were not initially proofs from the editorial office, which should be correct- expected. ed and returned (usually within three working days). Please consult the guidelines (https://researchintegrity- journal.biomedcentral.com/articles/10.1186/s41073-016- MANUSCRIPT PRESENTATION 0007-6). Please ensure that your manuscript follows these guide- Authors are also encouraged to use fair, accurate and lines. respectful language, but preferences can change and Manuscripts should be written in good English, as con- vary across groups and individuals and can also evolve cisely as possible to allow a clear understanding of the text. overtime. The following guidelines may help in use of a The title should be followed by the complete name of the authors, their affiliations – in the original language – town correct terminology in the area of HIV: https://www.cdc. and country. The name of the Working Group should ap- gov/stophivtogether/library/stop-hiv-stigma/fact-sheets/ pear at the end of the by-line; its composition should be cdc-lsht-stigma-factsheet-language-guide.pdf reported before the References, names and affiliations of https://www.hptn.org/resources/HIVLanguageGuide each member are required. The name and address, tele- https://unesdoc.unesco.org/ark:/48223/pf0000144725 phone and e-mail of the corresponding author should also The name of the bioresource (and identifier, if available) be indicated. On the same page a running head of no more which provided samples/data useful for the conduct of than 40 characters (including spaces) should be included. the study should be reported in extense, either in the Original articles should normally be organized into differ- Material and methods section or in the Acknowledge- ent sections (i.e.: Introduction, Materials and methods, ments. LENGTH OF THE TEXT with the Medline abbreviation of the US National Li- To provide a text that meets the requirements of our pub- brary of Medicine (www.nlm.nih.gov/bsd/aim.html). On- lication: line journal articles can be cited using, in addition to the • the letter to the Editor should be about 450 words; it complete citation, the DOI number. does not need an abstract; • the editorial should be no longer than 1000 words; edito- Articles in journal rials are submitted on invitation. Please contact the edito- Bozzuto G, Ruggieri P, Molinari A. Molecular aspects of rial office in advance if you wish to submit an editorial; tumor cell migration and invasion. Ann Ist Super Sanità. • the commentary, 2000 words; the commentary is an 2010;46(1):66-80. doi: 10.4415/ANN_10_01_09 opinion piece or reflection on recent papers previously published on Annali ISS or elsewhere; an abstract is re- Books and chapters in a book quired; please contact in advance the editorial office; Godlee F, Jefferson T. Peer review in health sciences. • the brief note, 3000 words, including about 15 referenc- London: BMJ Books; 1999. es, one table and one figure; Van Weely S, Leufkens HGM. Background paper: orphan • the article, 6000 words, including about 40 references, diseases. In: Kaplan W, Laing R (Eds). Priority medicines three tables and two figures; for Europe and the world – a public health approach to • the review should be no longer than 10 000 words, in- innovation. Geneva: World Health Organization; 2004. cluding no more than 100 references up to a maximum of four tables and three figures. Proceedings Fadda A, Giacomozzi C, Macellari V. Comparative FORMATTING GUIDELINES measurements to validate a new telemetric pressure in- Text soles system. In: 2. International Symposium on meas- • Use Times New Roman font, 10 point, single spaced; urement, analysis and modelling of human functions. 1. • do not use the automated features of your application Mediterranean Conference on measurement. Workshop (endnotes, headers, footers, especially for references); on evaluation check of traceability. Proceedings. Geno- • avoid using bold characters to emphasise words or sen- va: June 14-16, 2004. p. 425-7. tences within the text; • indicate clearly titles of chapters and subchapters avoid- Technical reports ing numbering. Della Seta M, Di Benedetto C, Leone L, Pizzarelli S, Siegmund U. ETHICSWEB technical guides. Manual for Tables and figures the creation of standards and guidelines for sharing infor- They should be understandable also without reference to mation about knowledge organization systems on ethics the text and should be numbered in Arabic numerals in a and science. Roma: Istituto Superiore di Sanità; 2011. consecutive and independent way according to their cita- (Rapporti ISTISAN, 11/32). tion within the paper. Tables should be presented on a separate sheet and pre- Legislation ceeded by a title. Each column within the table should Italia. Decreto legislativo 29 ottobre, n. 419. Riordinamen- have a heading. Abbreviations should be reported in full to del sistema degli enti pubblici nazionali, a norma degli in the legend. articoli 11 e 14 della legge 15 marzo 1997, n. 59. Gazzetta Figures should be loaded as separate files. The following Ufficiale – Serie Generale n. 268, 15 ottobre 1999. file formats are acceptable: JPEG, TIFF or EPS. Vecto- US Social Security Administration. Evidentiary require- rial images (graphs, flow charts, schemes, and other non ments for making findings about medical equivalence. bitmap material) should be in Excel, Adobe Illustrator, Final rules. Fed Reg. 2006 Mar 1;71(40):10419-33. Microsoft Power Point so as to allow the editorial format- ting of the material. The authors should check that each reference cited in the Figures are redrawn into the Annali style by our in-house text appears in the reference list and viceversa. Referenc- illustrators. es should not include works submitted for publication but Photographs must have a minimum resolution of 300 dpi. not yet accepted or unpublished results, etc. These can be Captions should be presented on a separate sheet and mentioned in the text in parentheses. contain a sufficient explanation of their object. They should be concise but comprehensive. CONVENTIONS All Latin or foreign words should be in italics. The au- REFERENCES thors should use internationally accepted abbreviations. All references in the text must be numbered in square All abbreviations should be spelled out in full the first brackets, i.e. [1, 2, 3-6], and mentioned at the end of the time they occur in the text, followed by the shortened article in the order in which they are quoted. They should term in parentheses; afterwards use the abbreviation conform to the “Recommendations for the Conduct, Re- only. porting, Editing, and Publications of Scholarly Work in Avoid abbreviations in the title of the manuscript. Medical Journals” (www.icmje.org), according to the fol- For writing symbols, quantities and units of measure- lowing examples. ments refer to the International Systems of Units (SI) and Titles of periodicals should be abbreviated in accordance the ISO standards. LICENSES AND AGREEMENTS Extended quotations and illustrative material taken from Please, contact the Editorial Office for any other information. other publications must be accompanied by the original Annali Editorial Office - Scientific Communication Service permission granted by the Authors and by the publisher Istituto Superiore di Sanità or in accordance with the copyright policy of the original Viale Regina Elena, 299 00161 Rome, Italy publication and the Creative Commons. Tel.: +39 06 49902945 Fax: +39 06 49902253 Responsibility for the contents and opinions expressed on E-mail: [email protected] this journal rests solely with the Author(s). www.iss.it/annali1 - www.annali-iss.eu