Vol. 23 | Weekly Issue 45 | 08 November 2018

Europe’s journal on infectious disease epidemiology, prevention and control

Vol. 23 | Weekly issue 45 | 08 November 2018

Surveillance and outbreak report Tick-borne encephalitis in Europe, 2012 to 2016 2 Julien Beauté, Gianfranco Spiteri, Eva Warns-Petit and Hervé Zeller Research articles Echovirus type 6 transmission clusters and the role of environmental surveillance in early warning, the Netherlands, 2007 to 2016 11 Susana Monge, Kimberley Benschop, Loes Soetens, Roan Pijnacker, Susan Hahné, Jacco Wallinga and Erwin Duizer High rates of meticillin-resistant Staphylococcus aureus among asylum seekers and refugees admitted to Helsinki University Hospital, 2010 to 2017 21 Tuomas Aro and Anu Kantele High prevalence of carriage of mcr-1-positive enteric bacteria among healthy children from rural communities in the Chaco region, Bolivia, September to October 2016 35 Tommaso Giani, Samanta Sennati, Alberto Antonelli, Vincenzo Di Pilato, Tiziana di Maggio, Antonia Mantella, Claudia Niccolai, Michele Spinicci, Joaquín Monasterio, Paul Castellanos, Mirtha Martinez, Fausto Contreras, Dorian Balderrama Villaroel, Esther Damiani, Sdenka Maury, Rodolfo Rocabado, Lucia Pallecchi, Alessandro Bartoloni and Gian Maria Rossolini Congenital brain abnormalities during a Zika virus epidemic in Salvador, Brazil, April 2015 to July 2016 44 Mariana Kikuti, Cristiane W. Cardoso, Ana P.B. Prates, Igor A.D. Paploski, Uriel Kitron, Mitermayer G. Reis, Ganeshwaran H. Mochida and Guilherme S. Ribeiro

www.eurosurveillance.org Surveillance and outbreak report Tick-borne encephalitis in Europe, 2012 to 2016

Julien Beauté¹, Gianfranco Spiteri¹, Eva Warns-Petit2,3, Hervé Zeller² 1. European Centre for Disease Prevention and Control (ECDC), Stockholm, Sweden 2. European Centre for Disease Prevention and Control (ECDC), Stockholm, Sweden (affiliation when the work was performed) 3. Direction Départementale de la Cohésion Sociale et de la Protection des Populations d’Ille-et-Vilaine, Rennes, France (current affiliation) Correspondence: Julien Beauté ([email protected])

Citation style for this article: Beauté Julien, Spiteri Gianfranco, Warns-Petit Eva, Zeller Hervé. Tick-borne encephalitis in Europe, 2012 to 2016. Euro Surveill. 2018;23(45):pii=1800201. https:// doi.org/10.2807/1560-7917.ES.2018.23.45.1800201

Article submitted on 18 Apr 2018 / accepted on 16 Aug 2018 / published on 08 Nov 2018

Since 2012, tick-borne encephalitis (TBE) is a notifi- most cases are infected by TBEV-Eu but cases infected able in the European Union. The European Centre for with TBEV-FE were reported in Estonia and Latvia [1] Disease Prevention and Control annually collects data and with TBEV-Sib in Estonia [3] and Finland [4]. from 28 countries plus Iceland and Norway, based on the EU case definition. Between 2012 and 2016, The typical course of the disease is biphasic. After a 23 countries reported 12,500 TBE cases (Ireland and median incubation period of 8 days, the first stage Spain reported none), of which 11,623 (93.0%) were consists of a few days of non-specific symptoms such confirmed cases and 878 (7.0%) probable cases. Two as fever, fatigue and body pain. After a symptom-free countries (Czech Republic and Lithuania) accounted for week, approximately one-third of infected persons can 38.6% of all reported cases, although their combined develop neurological conditions [5], ranging from mild population represented only 2.7% of the population meningitis to severe encephalitis [1]; increasing age under surveillance. The annual notification rate fluc- is a known risk factor for severe TBE. Infection with tuated between 0.41 cases per 100,000 population in TBEV-FE is associated with more severe disease with 2015 and 0.65 in 2013 with no significant trend over case fatality as high as 20–40% compared with 1–2% the period. Lithuania, Latvia and Estonia had the high- with TBEV-Eu [6]. est notification rates with 15.6, 9.5 and 8.7 cases per 100,000 population, respectively. At the subnational There is no curative treatment for TBE but a vaccine is level, six regions had mean annual notification rates available. This vaccine is highly immunogenic [7] and above 15 cases per 100,000 population, of which five the impact of mass vaccination in Austria is suggestive were in the Baltic countries. Approximately 95% of of good effectiveness [8]. Vaccine schedules for the two cases were hospitalised and the overall case fatal- vaccines licensed in Europe based on TBEV-Eu strains ity ratio was 0.5%. Of the 11,663 cases reported with require three doses followed by boosters [1]. In a posi- information on importation status, 156 (1.3%) were tion paper on TBE vaccination published in 2011, the reported as imported. Less than 2% of cases had World Health Organization (WHO) recommended that received two or more doses of TBE vaccine. TBE vaccination should be offered to all age groups in highly endemic areas (i.e. areas with TBE incidence Background above 5 cases per 100,000 population) [9]. Tick-borne encephalitis (TBE) is an infectious disease of the central nervous system caused by a flavivirus and In Europe, most cases occur during June-September usually transmitted by the bite of infected Ixodes spp. [10]. Ixodes spp. are found in large parts of Europe but These ticks can be found from western Europe to Japan areas at risk for TBE are mainly located in central and [1]. Less frequently, humans can be infected by drink- eastern Europe and the Baltic and Nordic countries [11]. ing contaminated milk. Many vertebrate species can be Between 2000–2010, the annual number of TBE cases infected by the TBE virus but ticks are the main reservoir reported in the European Union and European Economic for the virus. There are three subtypes of the TBE virus: Area (EU/EEA) fluctuated between 2,000–3,500 cases the European subtype (TBEV-Eu) is mainly transmitted [11,12]. Spikes in cases of TBE have occurred in some by I. ricinus while both the Far-eastern (TBEV-FE) and years, e.g. 2006, but this was likely a result of changes Siberian (TBEV-Sib) subtypes are mainly transmitted in human behaviour based on suitable weather condi- by I. persulcatus. Recent findings from Finland suggest tions (e.g. increased outdoor recreational activities) that I. ricinus can also transmit TBEV-Sib [2]. In Europe, [13]. More recently, some countries, e.g. Belgium and

2 www.eurosurveillance.org Box for variables with geographical information (probable European Union case definition for tick-borne place of infection and place of residence) followed the encephalitis nomenclature of territorial units for statistics (NUTS) of the EU [19]. A confirmed case is defined as any person meeting the clinical criteria i.e. symptoms of inflammation of the central nervous system We used population denominator data provided by the Statistical Office of the EU (Eurostat) for calculat- AND ing rates (data extracted on 22 September 2017). We • Has laboratory-confirmation i.e. at least one of the compared continuous variables by the Mann–Whitney following five: U test and categorical variables using the chi-squared test. We estimated annual rates of change and their • Tick-borne encephalitis (TBE) specific IgM and IgG antibodies in blood. 95% confidence intervals (CI) using a log-linear regression of notification rates over the period 2012–2016. • TBE specific IgM antibodies in cerebrospinal fluid. We assessed goodness of fit of linear regressions • Sero-conversion or fourfold increase of TBE-specific using F statistics. We used Stata software release 14 antibodies in paired serum samples. (StataCorp. LP, United States) for all data management and statistical analyses. • Detection of TBE viral nucleic acid in a clinical specimen.

• Isolation of TBE virus from clinical specimen. Results

A probable case is defined as any person meeting the clinical criteria and the laboratory criteria for a probable Case classification and notification rate case i.e. detection of TBE-specific IgM-antibodies in a Over the 2012–2016 period, 23 countries reported unique serum sample 12,500 TBE cases (Ireland and Spain reported no OR cases), of which 11,622 (93.0%) were confirmed cases and 878 (7.0%) probable cases (Table 1). We excluded Any person meeting the clinical criteria with exposure to a 31 cases with unknown classification (11 cases for common source (unpasteurised dairy products). Austria, 15 cases for Lithuania, four cases for Poland and one case for Slovenia). Cyprus, Iceland, Malta, and Portugal had no TBE surveillance and Denmark did not report any data. the Netherlands, reported possible new endemic foci Most countries (18/23) reported over 90% of cases as having found antibodies to the TBE virus in roe deer confirmed. Slovakia (552/638; 86.5%), France (36/44; and cattle [14,15] and in 2016, the Netherlands reported 81.8%), Hungary (131/171; 76.6%), Latvia (683/953; their first locally-acquired human case [16]. The map- 71.7%), and Poland (712/1,040; 68.5%), classified ping of endemic foci is essential to make recommen- the lowest proportions of their cases as confirmed. dations for vaccination programme and travel advice The mean annual notification rate was 0.54 cases per [17]. In 2011, the first attempt to collect TBE surveil- 100,000 population. lance data at the EU/EEA level underlined the need for an agreed case definition and systematic data collec- Importation tion [11]. Therefore, in 2012, the European Commission Of the 11,664/12,500 cases reported with informa- included TBE in the list of notifiable diseases in the EU/ tion on importation status, 156 (1.3%) were reported EEA [17]. as imported (Table 1). Importation status was missing for cases reported by Bulgaria, Croatia, and Finland. Here, we describe TBE cases reported in the EU/EEA All cases reported in Belgium, Luxembourg, and the between 2012 and 2016. United Kingdom (UK) were imported. Information on the probable country of infection was available for 152 Methods of these imported cases (97.4%). Top destinations for Since 2012, the European Centre for Disease Prevention travel-associated TBE were Austria (32 cases, 21.1% and Control (ECDC) requires all 28 EU Member States, of all imported cases), Sweden (19 cases, 12.5%) and plus Iceland and Norway, to annually report their TBE Finland (18 cases, 11.8%). Four countries (the Czech data to the European Surveillance System (TESSy) Republic, Germany, Lithuania, and Sweden) reported database using the EU case definition (Box) [18]. More 102/156 (65.3%) of all imported cases. Imported cases detailed information on surveillance systems is avail- were slightly younger than locally-acquired cases able elsewhere [10]. We included all cases reported (median age for imported cases: 46 years; locally- during the years 2012–2016 meeting the EU case defi- acquired cases: 48 years; p = 0.03) and more likely to nition in the analysis. be male (imported cases: 71% males; locally-acquired cases: 59% males; p < 0.01). TBE Information received included age, sex, date of disease onset, probable place of infection, place of Geographical distribution residence, importation status, hospitalisation status, Two countries (Czech Republic and Lithuania) accounted vaccination status, and clinical outcome. Coded values for 4,825/12,500 (38.6%) of all reported cases (Table 1).

www.eurosurveillance.org 3 NA NA NA NA NA NA NA NA NA NA 95% CI 0.7 to 28.1 - 90.5 to 0.3 - 39.0 to 8.2 - 13.9 to 41.3 - 30.2 to 17.7 42.0 to 112.7 to 42.0 - 31.1 to 40.2 - 24.3 to 19.9 - 28.3 to 32.7 - 29.1 to 16.0 - 19.6 to 49.9 - 31.2 to 38.9 - 78.6 to 19.2 - 42.9 to - 6.1 - 22.4 to 30.7 - 33.0 to 20.6 Trend NA NA NA NA NA NA NA NA NA NA 2.2 4.2 3.8 4.6 (%) 77.3 13.7 15.2 - 6.3 14.4 - 2.2 - 6.2 - 6.6 - 45.1 - 15.4 - 29.7 - 24.5 Annual variation 0 0 rate 9.51 0.35 0.55 0.01 0.18 0.01 2.36 6.95 0.36 4.79 0.54 0.93 0.92 0.02 0.02 0.06 2.44 0.68 8.68 15.64 < 0.01 < 0.01 < 0.01 < 0.01 Notification Notification

0 0 0 0 0 cases 6 (1.7) 2 (1.2) 1 (0.2) 3 (6.3) 2 (0.3) 5 (0.8) 1 (100) n (%)b 8 (0.8) 5 (100) 3 (100) Total cases 7 (15.9) 21 (1.8) 18 (0.9) 16 (0.6) 47 (3.5) 9 (21.4) 2 (50.0) 156 (1.3) Imported Unknown Unknown Unknown 1 7 5 5 2 2 4 0 0 53 47 44 171 716 572 253 953 144 393 638 1,180 1,445 2,522 1,040 2,303 12,500 Number of cases 1.11 rate 5.35 3.19 0.19 0.75 6.16 0.14 0.01 1.09 0.23 0.03 0.58 0.02 4.02 2.42 0.00 0.00 0.42 0.00 0.00 0.00 0.00 0.00 0.08 21.91 10.36 Notification Notification 2016 1 4 0 0 6 0 0 0 0 0 19 12 61 81 95 22 83 48 173 565 633 283 238 204 348 2,876 Number of cases NA 7.10 1.55 rate 3.31 0.17 2.75 3.01 0.01 0.41 0.01 0.18 0.01 0.39 1.24 0.27 0.92 0.03 0.62 0.02 0.24 0.00 0.00 0.00 0.00 8.82 11.50 Notification Notification 2015 1 1 1 5 2 9 0 0 0 0 of 10 79 26 62 24 84 68 NA 116 141 219 149 336 349 268 1,950 cases Number NA rate 0.51 0.31 6.31 2.14 1.85 0.33 0.01 0.01 0.01 0.95 0.25 7.44 0.54 3.90 0.43 4.85 0.00 0.00 0.00 0.00 0.00 0.00 0.86 11.99 < 0.01 Notification Notification 2014 1 1 9 2 0 0 0 0 0 0 13 31 81 47 23 83 NA 116 195 178 410 353 149 100 264 2,056 Number of cases NA NA rate 1.18 2.19 1.03 0.12 3.01 0.01 0.01 0.53 0.59 0.52 5.94 0.70 0.65 0.00 0.00 0.00 0.00 0.00 8.64 11.36 14.91 16.39 < 0.01 Notification Notification 2013 1 3 2 6 0 0 0 0 0 53 38 44 NA NA 114 163 419 100 625 307 487 230 226 209 3,027 Number of cases NA NA rate of encephalitis, tick-borne percentage imported of cases, notification rate per 100,000 population and trend, in 25 countries, European and Union 1.05 7.98 0.14 0.01 5.45 1.89 0.72 0.45 3.03 0.02 0.64 0.49 0.24 0.00 0.00 0.00 0.44 11.20 13.43 16.45 a < 0.01 < 0.01 Notification Notification 2012 1 7 3 3 2 0 0 0 45 39 38 44 NA NA 195 178 187 102 573 164 287 229 494 2,591 Number of cases onfidence interval; EU/EEA: European Union and European Economic Area; NA: not available; UK: United Kingdom. Number of cases includes both confirmed and probable cases. Poland Romania Bulgaria Croatia Norway Slovakia Slovenia Belgium Estonia Czech Republic France Spain Finland Greece Austria Hungary The Netherlands Country Germany Ireland Italy Latvia Lithuania Luxembourg Sweden UK EU/EEA cases with known importation status. European Economic Area, 2012–2016 (n = 12,500)European Economic Area, 2012–2016 a b Table 1 Table Number reported of cases CL: c

4 www.eurosurveillance.org Figure 1 Rate of locally acquired tick-borne encephalitis per 100,000 population, by place of infection, European Union and European Economic Area countries, 2012–2016

Tick-borne encephalitis mean notiﬁcation rate

per 100,000 population, 2012–2016

0.00 0.01–0.9 1.00–4.99 5.00–14.99 ≥ 15.00

Data not available No surveillance Not included

Statistical level

Country

NUTS 2

NUTS 3

Luxembourg

Malta

ECDC. Map produced on: 19 July 2018

Of the 23 countries that reported cases, 16 had mean notification rates below one case per 100,000 popula- Trend and Seasonality tion. Over the 2012–2016 period Lithuania, Latvia and The overall annual notification rate fluctuated between Estonia had the highest notification rates with 15.6, a minimum of 0.41 cases per 100,000 population in 9.5 and 8.7 cases per 100,000 population, respec- 2015 and a maximum of 0.65 in 2013 with no signifi- tively (Table 1 and Figure 1). Among the 23 countries cant trend over the period (annual variation of - 6.6% that reported cases, 17 had locally-acquired cases. Of (95% CI: - 29.1 to 16; p = 0.4) (Table 1). We observed these, 12 provided geographical information at NUTS3 significant trends for three countries: the TBE notifica- level, two at NUTS2 (Austria and Poland), and three tion rate increased at an annual rate of 14.4% (95%CI: did not have information at subnational level (France, 0.7 to 28.1) in Finland and 77.3% (95%CI: 42.0 to 112.7) the Netherlands, and Norway) (Figure 1). At the subna- in France and decreased at an annual rate of 24.5% tional level, six regions had mean annual notification (95%CI: 6.1 to 42.9) in Hungary. rates above 15 cases per 100,000 population: Utena county, Lithuania (44.5), Lääne-Eesti, Estonia (27.7), Of the 11,397 cases reported with onset date, 10,632 Kurzeme, Latvia (23.8), Alytus county, Lithuania (22.1), (93.3%) had an onset month May–October and 135 Panevėžys county, Lithuania (19.1) and Carinthia, (1.2%) had an onset month December–March (off-sea- Slovenia (15.1) (Figure 1).Twenty-nine regions in seven son) (Figure 2). We observed a comparable seasonal- countries (Estonia, Germany, Latvia, Lithuania, Poland, ity in the 12 countries reporting at least 100 cases over Slovenia and Sweden) had notification rates above the period with onset month (Figure 3). There were five cases per 100,000 population. In Lithuania, Telšiai peaks in 2012 (Estonia and Sweden), 2013 (Germany, County was the region with the lowest mean annual Hungary, Slovakia, Czech Republic and Slovenia), 2014 notification rate (5.6). (Austria), 2015 (Austria, Estonia, Finland, Sweden), and

www.eurosurveillance.org 5 Figure 2 Number of reported tick-borne encephalitis cases by month of onset, and 12-month moving average, 19 European Union and European Economic Area countries, 2012–2016

200 Number of cases 52-week moving average

150

100 Number of cases

0 week 1 week 27 week 1 week 27 week 1 week 27 week 1 week 27 week 1 week 27 2012 2012 2013 2013 2014 2014 2015 2015 2016 2016

Year and week of onset

Includes data from: Austria, Czech Republic, Estonia, France, Germany, Greece, Hungary, Italy, Latvia, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Romania, Slovakia, Slovenia and Sweden.

2016 (Czech Republic, Germany, Lithuania, Poland and Vaccination Slovakia). Of the 5,205 cases with known vaccination status, 5,066 (97.3%) were not vaccinated, 60 had received Demographics one or two doses (1.2%), 60 (1.2%) three doses or more Of the 12,470 cases reported with information on age, and 19 (0.4%) an unknown number of doses (Table 2). 6,782 (54.4%) were in the 40–69 years old group (Table Of the 20 cases with fatal outcome and known vac- 2). TBE was more common in males with a male-to- cination status, 19 were not vaccinated and one had female rate ratio of 1.5:1. Notification rates increased received one dose of the vaccine. The proportion of with age in both sexes, peaking at 0.89 cases per cases that received two doses or more of vaccine was 100,000 population in males aged 60–69 years, and higher in the extreme age groups compared with the then decreased in older age groups (Figure 4). At date other groups (2.5% in both cases aged 20 years or of disease onset, females (median 51 years, interquar- younger and 70 years or older). No imported cases had tile ratio (IQR): 35–62) were older than males (median received more than one vaccine dose. 47 years, IQR: 31–61) (p < 0.01). Discussion Outcome The European TBE surveillance data suggest a sta- Of the 8,081 cases reported with hospitalisation stable trend over the years 2012–2016 with no reported tus, 7,672 (94.9%) were admitted to hospital (Table 2). changes in national surveillance systems; continuing Of the 9,889 cases reported with known outcome, 48 the long-term trend observed in Europe since the mid- (0.5%) died and 247 (2.5%) had neurological sequelae. 1990s [12]. The number of TBE cases reported in Europe, The case fatality ratio did not differ significantly by excluding Russia, increased over the years 1990–1994, sex (0.5% in males vs 0.4% in females, p = 0.30). The probably reflecting the start of surveillance in many case fatality ratio was higher in older age groups (3.1% countries [12]. Over the following 15 years (1995–2009), in cases aged 80 years or older, 2.0% in cases aged the trend was stable with an annual number of TBE 70–79 years and < 0.5% in cases aged below 70 years). cases fluctuating between 2,000 and 4,000 cases. Peaks occurred when a set of countries reported unu- sually high numbers of TBE cases, e.g. 2006 and 2009 [12]. In 2013, several European countries experienced

6 www.eurosurveillance.org Figure 3 Number of reported tick-borne encephalitis cases by week of onset and 52-week moving average, 12 European Union and European Economic Area countries, 2012–2016

Hungary Finland Austria Estonia 50

Number of cases 10

Latvia Poland Slovakia Sweden 100

25 Number of cases

Germany Slovenia Czech Republic Lithuania 200

150

100

50 Number of cases 0

Jan 2012 Jan 2013 Jan 2014 Jan 2015 Jan 2016 Jan 2012 Jan 2013 Jan 2014 Jan 2015 Jan 2016 Jan 2012 Jan 2013 Jan 2014 Jan 2015 Jan 2016 Jan 2012 Jan 2013 Jan 2014 Jan 2015 Jan 2016

Month

Number of cases 12-month moving average

a peak in TBE cases, which resulted in the highest the notification rate almost tripled in 2016 compared number of TBE cases (> 3,000) observed in Europe that with previous years in the Alsace region where most year. An analysis carried out in eight European coun- cases occurred [20]; some newly identified foci such tries suggested that human behaviour in response to as the Alpine region could also have contributed to good weather conditions, e.g. increased outdoor rec- the upsurge in cases. However, the reasons behind reational activities, was the main explanation for the this increase are yet to be determined. In Finland, the 2006 spike rather than tick abundance [13]. emergence of new foci reported during 1995–2013 could partly explain the increase [21]. A decrease in The overall stable trend observed in TBE surveillance TBE cases was observed in Hungary over the years data is mainly driven by a few countries reporting the 2012–2016, to our knowledge there is no explanation majority of cases, potentially masking important dis- as to why. Trends at country level, such as these, may parities both between and within countries. For exam- mask changes at local level as TBE endemicity is very ple, two countries (Czech Republic and Lithuania) focal and countries do not have a uniform risk across accounted for 38.6% of all reported TBE cases, all territories/regions/counties etc. In Lithuania, which although their combined population represented only had the highest average annual notification rate, there 2.7% of the population of the 25 countries included in was an eightfold difference between counties with this analysis. All countries with average annual noti- highest and counties lowest TBE incidence. An analy- fication rate above one TBE case per 100,000 popu- sis of epidemiological patterns of TBE in Lithuania lation had a stable trend over the period. We only suggested different trends across counties with more observed an increase in Finland and France. In France, pronounced increases in eastern and northern parts

www.eurosurveillance.org 7 Table 2 groups: (i) tick abundance, (ii) population at risk, (iii) Main characteristics of reported cases of tick-borne surveillance characteristics. Factors related to tick encephalitis, European Union and European Economic Area countries, 2012−2016 (n = 12,500) abundance are multiple (e.g. land, weather, reservoirs etc.) and can be very focal. The impact of climate change is debated with possibly different effects in different Notification rate Number of settings. A study carried out in Sweden suggested that Characteristics Percent per 100,000 cases persons milder winters were associated with increased TBE Total 12,500 100 0.54 incidence in the mid-1980s [24]. Yet, a general circu- Age group (years) lation model predicted that TBE transmission could be < 20 1,402 11.2 0.29 disrupted by climate change with a contraction of TBE areas to higher altitudes in central Europe and north- 20–29 1,257 10.1 0.44 ern latitudes in Scandinavia [25]. This would result in 30–39 1,575 12.6 0.50 a decreased incidence in the coming decades but such 40–49 2,200 17.6 0.65 change would probably not be captured over a 5–year 50–59 2,474 19.8 0.77 period. Changes in human behaviour (e.g. increase of 60–69 2,108 16.9 0.80 at-risk outdoor recreational activities) can put peo- 70–79 1,183 9.5 0.63 ple at greater risk of exposure to ticks and thus TBE. ≥ 80 271 2.2 0.23 However, with increased vaccine coverage such risk Unknown 30 NA NA could be improved. Finally, better clinical awareness, Sex testing and reporting would improve the ability of the Female 5,118 40.9 0.43 surveillance system to detect cases. Male 7,381 59.1 0.65 Unknown 1 NA NA The geographical granularity of our data (at best Importation status NUTS3) does not allow fine monitoring of TBE foci, Imported 156 1.3 NA which countries are best placed to perform. However, during the first effort to collect TBE data at the EU/EEA Locally-acquired 11,507 98.7 NA level most of the recommendations were followed [11]. Unknown 837 NA NA We implemented standard EU case definition for TBE Hospitalisation and initiated routine collection of surveillance data Yes 7,672 94.9 NA from EU/EEA countries, to at least NUTS-3 geographi- No 409 5.1 NA cal level for most of the countries. ECDC encourages all Unknown 4,419 NA NA countries to report their cases at subnational level. Outcome Alive 9,594 97.0 NA The reported TBE cases followed a pronounced sea- Dead 48 0.5 NA sonality with most cases occurring during the warmer Neurological months May–October, which is likely due to human 247 2.5 NA complications habits with people spending a greater amount of time Unknown 2,611 NA NA outdoors in areas e.g. forests where ticks populations Vaccination status are high [1]. Cases infected during colder months are Four doses 24 0.5 NA possible, however, especially in central Europe. Three doses 36 0.7 NA Two doses 27 0.5 Cases of TBE are more common in older age groups, One dose 33 0.6 NA with the highest number of cases occurring in those Vaccinated unknown aged 40-69 years. The highest notification rate, in 19 0.4 NA doses those aged 60-69 years, most likely reflects high expo- Not vaccinated 5,066 97.3 NA sure to tick populations at an age where individuals Unknown 7,295 NA NA have increased time for outdoor recreational activity, but also fall into the known higher severity seen in NA: not available. older age groups [26].

Almost 95% (7,672/8,081) of reported TBE cases were admitted to hospital, which is not unexpected given that the clinical criteria used in the case definition of the country [22]. Similarly, diverging trends across selects severe cases. Even though the overall case regions were reported in Austria [23]. Decreasing trend fatality was relatively low, it was far from negligible in were observed in north-east of Lower Austria whilst older age groups at ca 2–3% above 70 years of age. the alpine regions in the west of Austria became highly Previous reviews suggested that a third of patients endemic. could suffer long-lasting sequelae [1]. Our analysis found a much lower proportion but it is likely that our Independently of what happens in animal reservoirs, data could not capture long-term sequelae that would we can classify factors driving TBE incidence in three

8 www.eurosurveillance.org Figure 4 Conclusion Notification rates of tick-borne encephalitis per 100,000 The overall TBE notification rate remained stable dur- population, by sex and age group and male-to-female rate ratio by age group, European Union and European ing 2012–2016. Surveillance at EU/EEA level helped Economic Area countries, 2012–2016 provide reliable and comparative data allowing better mapping of the disease risk both at the national and

1.0 subnational level. Countries with regions where the Female disease is highly endemic should consider strength- Male ening information campaigns on preventive measures 0.8 against tick bites as well as introducing TBE vaccine recommendations if these are not already proposed. 0.6 ECDC encourages countries to report better quality and more complete data on TBE diagnoses, particularly 0.4 on the sub-national geographic distribution and on imported cases.

Rate per 100,000 population 0.2

Acknowledgements 0.0 <20 20–29 30–39 40–49 50–59 60–69 70–79 ≥80 We would like to thank all people involved in the surveil- Age (years) lance of TBE in the participating countries as well as the data managers at ECDC, without whom surveillance of TBE at EU/EEA level would not be possible. In addition, we would like to thank the following persons for their contribution to TBE surveillance in Europe: Heidemarie Holzmann and be diagnosed later in time from acute infection and Karin Stiasny (Austria); Tinne Lernout, Vanessa Suin and thus not reported. Marjan Van Esbroeck, (Belgium); Sanja Kurecic Filipovic, Iva Pem Novosel and Goranka Petrovic (Croatia); Bohumír Vaccination remains the most effective protective Kříž Marek Malý and Helena Šebestová (Czech Republic); Markku Kuusi, Jussi Sane and Pirjo Turtiainen (Finland); measure against TBE [27]. However, studies have Isabelle Leparc-Goffart (France); Doris Altmann and Wiebke reported vaccine failures, especially in older age Hellenbrand (Germany); Theano Georgakopoulou, Kassiani groups [28]. We found that 87/5,205 (1.7%) of cases Gkolfinopoulou, Anna Papa, and Danai Pervanidou (Greece); were supposedly vaccinated (at least two doses of vac- Jeff Connell and Sarah Jackson (Ireland); Giovanni Rezza, cine), mostly in extreme age groups. This would be in Caterina Rizzo and Giulietta Venturi (Italy); Antra Bormane, Tatjana Klemjacionoka and Irina Lucenko (Latvia); Saulius line with results from studies suggesting that age and Čaplinskas, Aidas Spiečius and Milda Žygutienė (Lithuania); number of vaccine doses were the most important fac- Eelco Franz and Agnetha Hofhuis (The Netherlands); Solveig tors determining the immunological response to vac- Jore and Heidi Lange (Norway); Cornelia Ceianu, Romana cination [29]. The extended period between doses Rebreanu and Anca Sirbu (Romania); Edita Staroňová and may mean that people are less likely to comply to the Elena Tichá (Slovakia); Marta Grgic-Vitek, and Maja Socan (Slovenia). We are also grateful to Silviu Lucian Ionescu for recommendations as shown in Germany where com- his support in GIS. pliance after the first dose was low [30]. Another reason for not receiving or completing TBE vaccination is cost. TBE vaccination is not reimbursed in most EU/EEA Conflict of interest countries and the willingness to pay for vaccination may not be sufficient to ensure uptake in residents or None declared. visitors frequenting areas considered high risk for tick populations and TBE [31]. A survey published in 2008, reported that Austria, Finland, Germany, Hungary, Authors’ contributions Latvia and Slovenia included TBE in their routine vac- EWP and HZ initiated the work. GS and JB ran the analysis. cination programme at least for some specific groups All authors contributed to the interpretation of the findings. or areas [32]. JB wrote the first draft of the manuscript. All authors revised the manuscript, providing substantial intellectual input. In this study, we only found a few TBE cases in international travellers. Cases that are resident in countries References with little or no risk of TBE are less likely to be vacci- 1. Lindquist L, Vapalahti O. Tick-borne encephalitis. Lancet. nated or diagnosed [33]. Increased awareness of TBE 2008;371(9627):1861-71. https://doi.org/10.1016/S0140- is required to improve vaccination coverage in travel- 6736(08)60800-4 PMID: 18514730 lers and promote the best practices to avoid tick bites. 2. Jääskeläinen A, Tonteri E, Pieninkeroinen I, Sironen T, Voutilainen L, Kuusi M, et al. Siberian subtype tick-borne Currently, the WHO recommends vaccination of travel- encephalitis virus in Ixodes ricinus in a newly emerged focus, lers who are at risk of TBE exposure during outdoors Finland. Ticks Tick Borne Dis. 2016;7(1):216-23. https://doi. org/10.1016/j.ttbdis.2015.10.013 PMID: 26548609 activities in rural endemic areas during the period of 3. Golovljova I, Vene S, Sjölander KB, Vasilenko V, Plyusnin A, transmission [34]. Lundkvist A. Characterization of tick-borne encephalitis virus from Estonia. J Med Virol. 2004;74(4):580-8. https://doi. org/10.1002/jmv.20224 PMID: 15484275

www.eurosurveillance.org 9 4. Jääskeläinen AE, Sironen T, Murueva GB, Subbotina N, 22. Radzišauskienė D, Žagminas K, Ašoklienė L, Jasionis A, Alekseev AN, Castrén J, et al. Tick-borne encephalitis virus Mameniškienė R, Ambrozaitis A, et al. Epidemiological in ticks in Finland, Russian Karelia and Buryatia. J Gen Virol. patterns of tick-borne encephalitis in Lithuania and clinical 2010;91(11):2706-12. https://doi.org/10.1099/vir.0.023663-0 features in adults in the light of the high incidence in recent PMID: 20660147 years: a retrospective study. Eur J Neurol. 2018;25(2):268-74. 5. Taba P, Schmutzhard E, Forsberg P, Lutsar I, Ljøstad U, https://doi.org/10.1111/ene.13486 PMID: 29053888 Mygland Å, et al. EAN consensus review on prevention, 23. Heinz FX, Stiasny K, Holzmann H, Kundi M, Sixl W, Wenk M, diagnosis and management of tick-borne encephalitis. Eur et al. Emergence of tick-borne encephalitis in new endemic J Neurol. 2017;24(10):1214-e61. https://doi.org/10.1111/ areas in Austria: 42 years of surveillance. Euro Surveill. ene.13356 PMID: 28762591 2015;20(13):9-16. https://doi.org/10.2807/1560-7917. 6. Mansfield KL, Johnson N, Phipps LP, Stephenson JR, Fooks ES2015.20.13.21077 PMID: 25860391 AR, Solomon T. Tick-borne encephalitis virus - a review of an 24. Lindgren E, Gustafson R. Tick-borne encephalitis in Sweden emerging zoonosis. J Gen Virol. 2009;90(8):1781-94. https:// and climate change. Lancet. 2001;358(9275):16-8. https://doi. doi.org/10.1099/vir.0.011437-0 PMID: 19420159 org/10.1016/S0140-6736(00)05250-8 PMID: 11454371 7. Demicheli V, Debalini MG, Rivetti A. Vaccines for preventing 25. Randolph SE, Rogers DJ. Fragile transmission cycles of tick- tick-borne encephalitis. Cochrane Database Syst Rev. 2009; borne encephalitis virus may be disrupted by predicted climate (1):CD000977. PMID: 19160184 change. Proc Biol Sci. 2000;267(1454):1741-4. https://doi. 8. Heinz FX, Stiasny K, Holzmann H, Grgic-Vitek M, Kriz B, Essl A, org/10.1098/rspb.2000.1204 PMID: 12233771 et al. Vaccination and tick-borne encephalitis, central Europe. 26. Lindquist L. Tick-borne encephalitis. Handb Clin Neurol. Emerg Infect Dis. 2013;19(1):69-76. https://doi.org/10.3201/ 2014;123:531-59. https://doi.org/10.1016/B978-0-444-53488- eid1901.120458 PMID: 23259984 0.00025-0 PMID: 25015503 9. Vaccines against tick-borne encephalitis: WHO position paper. 27. Heinz FX, Holzmann H, Essl A, Kundi M. Field effectiveness Wkly Epidemiol Rec. 2011;86(24):241-56. PMID: 21661276 of vaccination against tick-borne encephalitis. Vaccine. 10. European Centre for Disease Prevention and Control. Tick- 2007;25(43):7559-67. https://doi.org/10.1016/j. borne encephalitis. In: ECDC. Annual epidemiological report vaccine.2007.08.024 PMID: 17869389 for 2016. Stockholm: ECDC; 2018. Available from: https://ecdc. 28. Andersson CR, Vene S, Insulander M, Lindquist L, Lundkvist A, europa.eu/sites/portal/files/documents/AER_for_2016-TBE.pdf Günther G. Vaccine failures after active immunisation against 11. European Centre for Disease Prevention and Control (ECDC). tick-borne encephalitis. Vaccine. 2010;28(16):2827-31. https:// Epidemiological situation of tick-borne encephalitis in doi.org/10.1016/j.vaccine.2010.02.001 PMID: 20167301 the European Union and European Free Trade Association 29. Lindblom P, Wilhelmsson P, Fryland L, Matussek A, Haglund M, countries. Stockholm: ECDC; 2012.Available from: https:// Sjöwall J, et al. Factors determining immunological response ecdc.europa.eu/sites/portal/files/media/en/publications/ to vaccination against tick-borne encephalitis virus in older Publications/TBE-in-EU-EFTA.pdf individuals. PLoS One. 2014;9(6):e100860. https://doi. 12. Süss J. Tick-borne encephalitis 2010: epidemiology, risk org/10.1371/journal.pone.0100860 PMID: 24967619 areas, and virus strains in Europe and Asia-an overview. Ticks 30. Jacob L, Kostev K. Compliance with vaccination against tick- Tick Borne Dis. 2011;2(1):2-15. https://doi.org/10.1016/j. borne encephalitis virus in Germany. Clin Microbiol Infect. ttbdis.2010.10.007 PMID: 21771531 2017;23(7):460-3. https://doi.org/10.1016/j.cmi.2017.01.012 13. Randolph SE, Asokliene L, Avsic-Zupanc T, Bormane A, Burri PMID: 28143786 C, Gern L, et al. Variable spikes in tick-borne encephalitis 31. Slunge D. The Willingness to Pay for Vaccination against Tick- incidence in 2006 independent of variable tick abundance but Borne Encephalitis and Implications for Public Health Policy: related to weather. Parasit Vectors. 2008;1(1):44. https://doi. Evidence from Sweden. PLoS One. 2015;10(12):e0143875. org/10.1186/1756-3305-1-44 PMID: 19068106 https://doi.org/10.1371/journal.pone.0143875 PMID: 26641491 14. Jahfari S, de Vries A, Rijks JM, Van Gucht S, Vennema H, 32. Donoso Mantke O, Schädler R, Niedrig M. A survey on cases of Sprong H, et al. Tick-Borne Encephalitis Virus in Ticks and Roe tick-borne encephalitis in European countries. Euro Surveill. Deer, the Netherlands. Emerg Infect Dis. 2017;23(6):1028-30. 2008;13(17):18848. PMID: 18445446 https://doi.org/10.3201/eid2306.161247 PMID: 28518024 33. Haditsch M, Kunze U. Tick-borne encephalitis: a 15. Roelandt S, Suin V, Riocreux F, Lamoral S, Van der Heyden S, disease neglected by travel medicine. Travel Med Infect Van der Stede Y, et al. Autochthonous tick-borne encephalitis Dis. 2013;11(5):295-300. https://doi.org/10.1016/j. virus-seropositive cattle in Belgium: a risk-based targeted tmaid.2013.07.003 PMID: 23916617 serological survey. Vector Borne Zoonotic Dis. 2014;14(9):640- 34. Steffen R. Epidemiology of tick-borne encephalitis (TBE) 7. https://doi.org/10.1089/vbz.2014.1576 PMID: 25229702 in international travellers to Western/Central Europe and 16. de Graaf JA, Reimerink JH, Voorn GP, Bij de Vaate EA, de Vries conclusions on vaccination recommendations. J Travel Med. A, Rockx B, et al. First human case of tick-borne encephalitis 2016;23(4):taw018. PMID: 27087558 virus infection acquired in the Netherlands, July 2016. Euro Surveill. 2016;21(33):30318. https://doi.org/10.2807/1560- 7917.ES.2016.21.33.30318 PMID: 27562931 17. Amato-Gauci A, Zeller H. Tick-borne encephalitis joins the License and copyright diseases under surveillance in the European Union. Euro Surveill. 2012;17(42):20299. PMID: 23098821 This is an open-access article distributed under the terms of 18. European Commission. Commission Regulation (EC). the Creative Commons Attribution (CC BY 4.0) Licence. You Commission Implementing Decision 2012/506/EU of 8 may share and adapt the material, but must give appropriate August 2012 amending Decision 2002/253/EC laying down case definitions for reporting communicable diseases to credit to the source, provide a link to the licence, and indi- the Community network under Decision No 2119/98/EC of cate if changes were made. the European Parliament and of the Council. Luxembourg: Publication Office of the European Union. Available from: This article is copyright of the authors or their affiliated in- https://publications.europa.eu/en/publication-detail/-/ publication/10ed460f-0711-11e2-8e28-01aa75ed71a1/ stitutions, 2018. language-en 19. European Commission. Commission regulation (EC). No 1059/2003 establishing the common classification of territorial units for statistics (NUTS). Luxembourg: Publication Office of the European Union. Available from: https://eur-lex.europa.eu/ legal-content/EN/TXT/?qid=1535630753511&uri=CELEX:3200 3R1059 20. Velay A, Solis M, Kack-Kack W, Gantner P, Maquart M, Martinot M, et al. A new hot spot for tick-borne encephalitis (TBE): A marked increase of TBE cases in France in 2016. Ticks Tick Borne Dis. 2018;9(1):120-5. https://doi.org/10.1016/j. ttbdis.2017.09.015 PMID: 28988602 21. Tonteri E, Kurkela S, Timonen S, Manni T, Vuorinen T, Kuusi M, et al. Surveillance of endemic foci of tick-borne encephalitis in Finland 1995-2013: evidence of emergence of new foci. Euro Surveill. 2015;20(37):30020. https://doi.org/10.2807/1560- 7917.ES.2015.20.37.30020 PMID: 26535471

10 www.eurosurveillance.org Research article Echovirus type 6 transmission clusters and the role of environmental surveillance in early warning, the Netherlands, 2007 to 2016

Susana Monge1,2, Kimberley Benschop1, Loes Soetens1,3, Roan Pijnacker¹, Susan Hahné¹, Jacco Wallinga1,3, Erwin Duizer¹ 1. Centre for Infectious Disease Control (CIb), National Institute for Public Health and the Environment (RIVM), Bilthoven, the Netherlands 2. European Programme for Intervention Epidemiology Training (EPIET), European Centre for Disease Prevention and Control, (ECDC), Stockholm, Sweden 3. Department of Biomedical Data Sciences, Leiden University Medical Center, Leiden, The Netherlands Correspondence: Kimberley Benschop ([email protected])

Citation style for this article: Monge Susana, Benschop Kimberley, Soetens Loes, Pijnacker Roan, Hahné Susan, Wallinga Jacco, Duizer Erwin. Echovirus type 6 transmission clusters and the role of environmental surveillance in early warning, the Netherlands, 2007 to 2016. Euro Surveill. 2018;23(45):pii=1800288. https://doi.org/10.2807/1560-7917. ES.2018.23.45.1800288

Article submitted on 01 Jun 2018 / accepted on 11 Sep 2018 / published on 08 Nov 2018

Background: In the Netherlands, echovirus type 6 (E6) Background is identified through clinical and environmental enter- Enteroviruses (EVs) are viruses of ovirus surveillance (CEVS and EEVS). Aim: We aimed the Picornaviridae family mainly transmitted via the to identify E6 transmission clusters and to assess the faecal–oral route, but also through eye, nose, and role of EEVS in surveillance and early warning of E6. mouth secretions. Humans are the only reservoir of the Methods: We included all E6 strains from CEVS and four species classified A to D, within which more than EEVS from 2007 through 2016. CEVS samples were 100 types have been described [1]. For public health from patients with enterovirus illness. EEVS sam- purposes, EVs are mainly classified into polioviruses ples came from sewage water at pre-specified sam- and non-polio EVs (NPEVs). As the polio eradication pling points. E6 strains were defined by partial VP1 approaches, NPEVs are receiving increased attention sequence, month and 4-digit postcode. Phylogenetic as sources of morbidity and mortality [2,3], being the E6 clusters were detected using pairwise genetic dis- most common cause of viral central nervous system tances. We identified transmission clusters using a (CNS) infections [4-8]. They can present as sporadic combined pairwise distance in time, place and phy- cases or as clusters or even cause large outbreaks logeny dimensions. Results: E6 was identified in 157 involving severe cases with neurological or respiratory of 3,506 CEVS clinical episodes and 92 of 1,067 EEVS complications [1-3]. Echovirus type 6 (E6, EV-B species) samples. Increased E6 circulation was observed in detections have been increasing in recent years in the 2009 and from 2014 onwards. Eight phylogenetic Netherlands, where it caused 13% of all NPEV infections clusters were identified; five included both CEVS and in 2016 [9,10]. It is one of the most frequently detected EEVS strains. Among these, identification in EEVS types circulating in Europe [8,11-13] and other regions did not consistently precede CEVS. One phylogenetic [14-17], and has been implicated in meningitis outbreaks cluster was dominant until 2014, but genetic diversity [10,18-21]. increased thereafter. Of 14 identified transmission clusters, six included both EEVS and CEVS; in two of The World Health Organization [22] suggests three them, EEVS identification preceded CEVS identifica- reasons for EV surveillance unrelated to polio eradi- tion. Transmission clusters were consistent with phy- cation: to detect and respond to outbreaks, to estab- logenetic clusters, and with previous outbreak reports. lish disease burden data for public health planning, Conclusion: Algorithms using combined time–place– and to perform virological investigation and research. phylogeny data allowed identification of clusters not Serotyping and characterisation of circulating NPEVs, detected by any of these variables alone. EEVS iden- including rapid identification and close monitoring of tified strains circulating in the population, but EEVS emerging strains, is of public health relevance [23]. In samples did not systematically precede clinical case the Netherlands, poliovirus surveillance is performed surveillance, limiting EEVS usefulness for early warn- through clinical EV surveillance (CEVS), mainly in stool ing in a context where E6 is endemic. samples [9,24], complemented with environmental EV surveillance (EEVS) in sewage water [24]. Exclusion of poliovirus is based on virus isolation and sequencing of EV-positive samples. Results of both surveillance

www.eurosurveillance.org 11 Table Number of samples analysed within the clinical and environmental enterovirus surveillance systems, the Netherlands, 2007–2016 (n=4,795)

Number of samples analysed Samples positive for E6 n % (column) n % (row) Faecal 2,847 76.4 109 3.8 Respiratory 403 10.8 6 1.5 Cerebrospinal fluid 362 9.7 44 12.2 CEVS: sample Other 96 2.6 1 1.0 Unknown 20 0.5 0 0.0 Total 3,728 100 160 4.3 a 13 villages 451 42.3 38 8.4 11 schools 428 40.0 43 10.1 EEVS: sampling point 1 asylum seeker centre 188 17.4 11 5.9 Total 1,067 100 92 8.6 b

CEVS: clinical enterovirus surveillance system; EEVS: environmental enterovirus surveillance system; E6: echovirus type 6. a Multiple samples allowed per patient, therefore the total number of positive samples is higher that total number of positive cases (160 positive samples in 157 cases). b In one sample two distinct E6 types were identified, therefore total number of distinct E6 strains isolated in the environment is 93.

systems also provide valuable information on circulat- samples from patients with EV-related illness are ana- ing NPEVs. lysed at clinical laboratories. Laboratories are strongly advised to send positive samples to the National In the past, EEVS demonstrated to be more sensitive Institute for Public Health and the Environment (RIVM) than surveillance based on clinical cases alone [24,25]. or to a TypeNed laboratory for typing and exclusion of NPEVs isolates from EEVS are also phylogenetically poliovirus [31]. RIVM predominantly receives samples related to clinical isolates, suggesting its potential from the central part of the Netherlands. EEVS is per- for characterising the extent of NPEVs circulation and formed in sewage water from secondary schools or diversity [13,14,24,26,27] or recognising outbreaks residential areas in regions at higher risk of poliovi- [15,27]. More importantly, as demonstrated for polio- rus introduction, in the area known as the bible belt, virus [28], NPEV infections have been preceded by where the uptake of vaccination is low because of reli- silent circulation of the same strain in the environment gious beliefs [32]. In addition, ad hoc locations can be [14,26,29,30], suggesting a role of EEVS in early warn- included, for example, downstream from asylum seeker ing. Incorporating data on time and space in the phy- centres [24] or from the residence of known shed- logenetic analysis could help investigate the relation ders [33]. Samples are collected approximately every between environmental and clinical NPEVs strains, and 6 weeks, varying by season or following polio alerts assess if EEVS captures ongoing transmission chains [24,33,34]. Samples from CEVS typed at the RIVM rea- and could act as early warning. sonably overlap with the areas covered by the EEVS.

The public health importance of E6, together with its Laboratory methods detectability in sewage water [14,15,24], make it a good Laboratory procedures for the surveillance systems candidate to evaluate the full potential of environmen- have been previously described in detail [24]. Briefly, tal EV surveillance. Our objective was to identify trans- RNA was isolated from sewage water cultures with mission clusters of E6 in the Netherlands using time, positive cytopathic effect and 5’ UTR RT-PCR-positive place and phylogeny data of clinical and environmen- clinical samples by automated extraction using the tal strains, with the final goal of assessing the added LC Nucleic Acid isolation kit (MagnaPure96, Roche, value of environmental surveillance in general and, Almere, the Netherlands). RNA was eluted in 50 µL elu- specifically, for early warning of E6 outbreaks. tion buffer and amplified in the semi-nested RT-PCR described by Nix et al. [35]. The sensitivity of this Methods RT-PCR, defined as the equivalent of the lowest dose of cultured infectious virus detected by this RT-PCR, Design and setting was 0.126 50% tissue culture infective dose (TCID50) We included all E6-positive clinical cases and envi- for EV-B (prototype virus coxsackievirus B3). The 350– ronmental samples identified from 2007 through 400 nt fragments of the VP1 gene were purified using 2016 within CEVS and EEVS. Data from CEVS was ExoSAP-IT and sequenced at Baseclear (Leiden, the anonymised. Both surveillance systems have been Netherlands). The partial VP1 sequences were edited previously described [9,24,31]. Briefly, for the CEVS, using BioNumerics version 7.1 and used as input in the

12 www.eurosurveillance.org Figure 1 Echovirus type 6 detected in clinical (n = 157) and environmental surveillance (n = 93), the Netherlands, 2007–2016

A. E6 detections and total number of samples by source (N=250) B. Proportion of samples with detection of E6 by source (N=250)

80 160 140

120 Total number of samples 60 100 80 60 40 40 10 20 Percentage 8 0 20 6 4 2 Number of E6-positive samples

0 0 2011m1 2015m1 2017m1 2013m1 2014m1 2012m1 2016m1 2010m1 2011m1 2015m1 2017m1 2007m1 2013m1 2014m1 2012m1 2009m1 2016m1 2010m1 2008m1 2007m1 2009m1 2008m1

Month Month

E6 environmental E6 clinical Enviromental samples Clinical samples Total environmental Total clinical

E6: echovirus type 6.

Panel A: Total number of E6 detections and total number of samples analysed per month from the environmental and the clinical surveillance systems. Panel B: Percentage of samples positive for E6 among the total number of samples by month.

EV genotyping tool (http://www.rivm.nl/mpf/enterovi- of 15% to define subtype-specific phylogenetic clusters rus/typingtool/), which has an automated algorithm using Cluster Picker 1.2.4 [38]. to assign the species and (sub)type of the sequences entered [36]. We analysed E6 transmission clusters by applying the algorithm developed by Ypma et al. [39]. Briefly, it Data analysis uses pairwise distances in time, place and phylogeny Distribution of E6 strains across time, place and phy- to calculate the relative distance in each dimension logeny was described. Date of sample collection was between any two cases (i.e. number of cases between used or, if missing, date of reception at RIVM. Samples any two cases in an ordered distance sequence) and from the same patient within 10 days were consid- multiplies the three dimensions to obtain a combined ered as a single episode. E6 strains were described by distance measure. A hierarchical clustering tree is con- source (clinical or environmental) and month. Dot maps structed by sequentially connecting pairs of cases with represented geographical location of strains as a ran- the smallest distance, resulting in a clustering tree in dom point within the polygon of the four-digit postcode which length of the branches represents the combined (PC4) or, if missing, the municipality. For environmental distance. Significance of the clusters given their height samples, we used the location of the sewage sampling and size is calculated by bootstrapping methods [39]. point and for clinical samples, the location of patients’ In this study we considered clusters with p < 0.05. In residence or, if missing, of the clinical laboratory. addition, we chose a cut-off value of the tree height at 15% below which clusters are identified. This cut-off is Phylogenetic analysis was based on partial VP1 gene. set ad hoc and is guided by a plausibility assessment The Enterovirus-B echovirus type 6 D’Amori strain of the identified clusters, including the analysis of GenBank no. AY302558 was used for sequence align- intra-cluster correlation and inter-case distances (algo- ment using MEGA 7.0.21. Sequences were omitted if rithm available on github.com/lsoetens/ClusterViz/). they were shorter than 270 nt or if they overlapped only Finally, we described the distribution of the identified partially with the fragment from position 2,611 to 2,881. transmission clusters across time, place and phylog- Phylogenetic trees were built using the maximum-like- eny. Analyses were performed using R 3.4.0. lihood method with 1,000 bootstrap replicates and a threshold of 70%. Genetic diversity was assessed by Accession numbers pairwise distribution [37]. We used a genetic threshold VP1 partial sequences were deposited in GenBank; the accession numbers are listed in the Supplement.

www.eurosurveillance.org 13 Figure 2 Distribution of echovirus type 6 strains across Geographical distribution of echovirus type 6 in clinical time, place and phylogeny (n = 157) and environmental surveillance (n = 93), the Netherlands, 2007–2016 E6 showed long periods of low incidence with peaks of increased circulation in 2009 and from 2014 onwards, compatible with an epidemic pattern (Figure 1). The number of EV-positive samples received for typing within the CEVS was variable, with peaks in the number of samples received coinciding with peaks in E6 Groningen detections. The variability in the number of sewage water samples was lower. To remove this interference, we calculated the proportion of samples positive for E6 in both systems. After doing that, the increase in 2014 was more evident in sewage than in clinical cases (Figure 1). No periodic pattern or long-term trend was evident. Importantly, there was no evidence that Amsterdam increases in environmental detections of E6 preceded the occurrence of clinical cases.

PC4 or municipality was available for all environmental surveillance locations and for 124 cases; for the remaining 33 cases, the PC4 of the virology laboratory was used. E6 cases were widely distributed in the central Eindhoven part of the country (Figure 2), matching the area covered by the CEVS and the EEVS. No evident geographical clustering was present overall, nor when looking at individual years (data not shown).

Partial VP1 sequences encompassing at least 270 nt were available for 188 isolates: 58 (62%) environmental and 130 (83%) clinical isolates. Eight phylogenetic Red dots: clinical samples; blue dots: environmental samples. clusters were identified (Figure 3 and Supplement Boundaries represent municipalities, points represent random locations for cases within the 4-digit postcode polygon, and Figure S1). Two contained only environmental isolates: shaded grey areas represent 4-digit postcodes from which at least cluster 2 (n = 6) and cluster 8 (n = 2) and one contained one sample was received in the clinical enterovirus surveillance system. only clinical isolates: cluster 4 (n = 3). Five clusters included both: cluster 1 (31 environmental and 75 clinical isolates), cluster 3 (nine environmental and 30 clinical isolates), cluster 6 (three environmental and four clinical isolates), cluster 5 (three environmental and Results three clinical isolates) and cluster 7 (two environmental and one clinical isolate). There were 14 clinical isolates Description of samples from the clinical and and two environmental isolates that did not belong to environmental enterovirus surveillance any phylogenetic cluster. Between 2007 and 2016, RIVM received 1,067 sewage water samples from 25 locations within EEVS, and Cluster 1 was the dominant strain up to late 2014, with 3,728 samples from 3,506 clinical episodes in 3,452 simultaneous detections in EEVS and CEVS (Figure patients within CEVS (Table). E6 was detected in 92 4). In 2015 and 2016, other phylogenetic clusters (8.6%) samples from EEVS. There was one sample with appeared. Cluster 3 appeared first in an environmental two distinct E6 viruses, therefore a total of 93 differ- sample in October 2014, followed by a clinical case in ent E6 isolates were analysed. E6 was detected in 157 November 2014, and again in the EEVS in June 2015 fol- (4.5%) clinical episodes, hereafter referred to as cases. lowed by the CEVS from July 2015 onwards. Cluster 5 Of all episodes, 1,756 (50%) were male, 1,332 (38%) appeared first in a clinical sample in December 2015 female and 418 (12%) of unknown sex, with similar pro- and was thereafter detected in other clinical cases portions of E6 positivity (chi-squared test: p = 0.335). and in the EEVS. Also cluster 6 was first detected in Median age (based on 3,124 episodes) was 3.1 months a clinical case in November 2015 and then sporadi- overall (interquartile range (IQR): 1.0 months–2.5 years) cally during 2016 in other clinical cases and the EEVS. and 4.3 months in E6 cases (IQR: 0.9 months–18.3 Cluster 7 appeared simultaneously in the CEVS and years; Mann–Whitney U test: p = 0.339). EEVS in August 2016. All phylogenetic clusters were geographically dispersed throughout the Netherlands (Supplement Figure S2).

14 www.eurosurveillance.org Figure 3 Phylogenetic tree of echovirus type 6 sequences in the VP1 positions 2,611–2,881, depicting eight identified phylogenetic clusters and 14 transmission clusters, the Netherlands, 2007–2016 (n = 188)

Maximum-likelihood method with 1000 bootstrap replicates, analysed in MEGA 7.0.21.

Red dots: echovirus type 6 isolates from environmental samples (n = 58) and black dots from clinical samples (n = 130). Background colours represent phylogenetic clusters. Letters represent transmission clusters.

Transmission clusters with a few cases from transmission cluster H (Figure Using the time–place–phylogeny clustering algorithm, 5). Phylogenetic cluster 2 as a whole was identified as we identified 14 transmission clusters (designated A to transmission cluster G and was the most robust cluster. N) comprising 122 (65%) isolates. Most transmission Cluster K comprised isolates from phylogenetic cluster clusters were dense in time, place and genetics, with 4 along with isolates that were not part of a cluster the exception of clusters J, K, N and M, which showed based on phylogeny alone. Possible spurious cluster- a significantly larger variation on the geographical ing was identified by analysing within-cluster genetic dimension than the non-clustered isolates (however distance (Supplement Table S1). In cluster J, very high geographical distances in the Netherlands are small) genetic distance (21.5%) was identified between just (Supplement Figure S3). In addition, clusters C, D, E one isolate (from phylogenetic cluster 4) and the rest and H showed high intra-cluster correlation, indicating of isolates (from phylogenetic cluster 1). After exclud- large internal cohesion in these clusters (Supplement ing this sequence, the maximum intra-cluster genetic Figure S4). For the other transmission clusters, the distance was 1.3%. In cluster C, maximum distance Spearman rank correlation coefficient between pair- was 8.3%, but decreased to 0% after the exclusion wise dimensions was either non-existent, very low, of one sequence. Cluster M, had pairwise distances non-significant or could not be estimated because of higher than 1.5% (corresponding to 4 nt difference) in zero variance in one or both dimensions. 78% of the pairs, corresponding to isolates from phylogenetic clusters 3, 5, 6 and 7. Cluster M also had the Seven transmission clusters (A, C, D, H, J, L, N) were highest distances in phylogeny and time, spanning 12 identified within phylogenetic cluster 1, and four (B, months, as well as large geographical distance (Figure E, F, I) within phylogenetic cluster 3. Cluster N was 5, Supplement Figures S4 and S6). the one responsible for the E6 peak of 2009, together

www.eurosurveillance.org 15 Figure 4 Monthly distribution of the eight phylogenetic clusters of echovirus type 6, the Netherlands, 2007–2016 (n = 188)

A. Environmental surveillance (n = 58)

9 8 7 6 5 4 3 2 1 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

B. Clinical surveillance (n = 130)

9 8 7 6 5 4 3 2 1 0

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Cluster 1 Cluster3 Cluster 5 Cluster 7 Non-clustered

Cluster 2 Cluster 4 Cluster 6 Cluster 8

Six of the 14 transmission clusters included both envi- that environmental detections systematically preceded ronmental and clinical isolates (E, H, J–M), another detection through surveillance based on clinical cases. six contained only clinical isolates (A–D, I, N) and To our knowledge, this is the first study to jointly ana- two included only environmental isolates (F and G), lyse environmental and clinical samples of EVs using although cluster G corresponded to the asylum seeker time, place and phylogeny to detect transmission centre, for which no clinical information was available. clusters. Combining the three dimensions allowed Only in two (E and L) of the six transmission clusters identification of clusters not detected by any of these including both clinical and environmental isolates the variables alone. Possible spurious clustering could environmental E6 detection preceded the detection in be detected, probably explained by the simultaneous CEVS. endemic circulation of different E6 strains in the same geographical region and time, as has been described Discussion by other authors [40]. It is not very likely that cluster In the Netherlands, E6 presented an epidemic pat- M corresponded to a true transmission cluster, given tern, with dispersed geographical distribution and the high genetic diversity of the isolates (more than increasing genetic diversity. Environmental and clini- 4 nt in 78% of the pairs), which fell within different cal samples belonged to joint phylogenetic clusters phylogenetic groups, as well as the high inter-patient and, to a lesser extent, to joint transmission clusters, distances in time and in space. However, most of the confirming the usefulness of environmental surveil- transmission clusters fell within single phylogenetic lance to characterise E6 strains causing disease in clusters and were highly plausible. Separate transmis- human populations. However, there was no indication sion clusters of phylogenetically similar strains is most probably explained by two separate introductions of

16 www.eurosurveillance.org Figure 5 Monthly distribution of the 14 transmission clusters (A to N) of echovirus type 6, the Netherlands, 2007–2016 (n = 188)

10 None C F I L A D G J M B E H K N

6 Number of cases 4

2008 2010 2012 2014 2016 Date onset disease

the same strain or by a circulating strain causing two low public health importance. The detection of novel distinct outbreaks. Our transmission clusters were also EVs in the environment can indicate emerging strains concordant with previous outbreak reports. Cluster that may cause outbreaks in the short or medium term. N explained the period of increased E6 circulation in In Finland, close genetic relatives of the echovirus 30 2009. Interestingly, the nine clinical isolates identified strains that caused an epidemic in 2009 had already as transmission cluster I corresponded to an outbreak been isolated in the environment a few years before that was detected between June and August 2016 and the outbreak [29]. A recent study in France found silent comprising 10 cases from the same province, of whom circulation of EV D68 in periods were no clinical cases nine had neurological presentation; the 10th case from were being reported [26] and in Greece, an E6 strain the outbreak was not in our study [10]. This supports found in sewage water in 2006 was correlated with the validity of our cluster detection method and high- cases from an aseptic meningitis outbreak 1 year later lights its usefulness for automated processing of com- [30]. This has raised the question of whether environ- plex data. Timely detection could potentially contribute mental surveillance could serve as early warning of to an earlier response to these clusters. Although there increases in EV infections in the population. is no effective antiviral therapy against EVs, case management, enhanced hygiene measures and vigilance of Environmental surveillance was able to capture emerg- close contacts can contribute to control and limit the ing strains causing disease in human populations, as spread of potentially severe EV infections. 83% of all environmental isolates were clustered phylogenetically with clinical isolates. However, 45% did At the frontier of polio elimination, also some other not belong to any transmission cluster and an addi- countries have implemented environmental EV sur- tional 19% belonged to transmission clusters that did veillance [8,13,29,30]. Previous studies have demon- not include cases captured by clinical surveillance. Our strated the potential of environmental specimens to results indicate that the usefulness of environmental contribute to the characterisation of circulating EVs in surveillance for early warning of E6 is very limited, at general and E6 in particular [14,41]. Environmental sur- least in contexts such as the Netherlands where E6 veillance overcomes the reference bias present in clini- circulation is established and active typing of clinical reports. While clinical surveillance only captures cal samples is performed. The usefulness of environ- severe cases for whom physicians seek microbiological mental surveillance could potentially be improved by confirmation, environmental data provides a broader increasing sampling frequency, although the feasibility picture of circulating EVs. On the other hand, this could and cost-effectiveness of such a programme is ques- result in identification of strains with low virulence and tionable. Also, usefulness as early warning could be

www.eurosurveillance.org 17 higher in countries where clinical surveillance is sub- Conclusion optimal or where E6 or other EV types are not endemic. Transmission clusters are reliably identified by For example, in Scotland, HPeV3 was not detected in jointly analysing time, place and genetic information. the sewage sludge/sediment collected in Edinburgh in Therefore, increasing efforts should be made to col- 2009 and only appeared in sewage one week before lect accurate time and place information within EV the first clinical case was diagnosed in early 2010 [13]. surveillance systems, along with genetic information. Automated algorithms can provide detection of such One of the transmission clusters detected the spread clusters and can potentially help control efforts to limit of E6 within an asylum seeker centre. The average stay the spread of highly pathogenic EVs. Environmental in this centre was between 3 and 7 days, while in the surveillance proved to be valuable to characterise EVs, EEVS, the cluster was detectable during 3 weeks, indi- in this case E6, causing disease in human populations, cating subsequent transmission, rather than a unique but showed limited benefit for early warning for this shedder. Except in the sewage from the asylum seeker endemic EV type. centre, this strain was not detected elsewhere, indicating that there was probably no transmission outside of this centre. Acknowledgements The technical assistance by Ron Altena, Edin Jusic, Dani Atto, A limitation of our study is that the CEVS is based on Jeroen Cremer, Bas van der Veer and Anne Marie van den voluntary referral of samples from clinical laborato- Brandt is well appreciated. ries [31]. Moreover, there are no standard criteria for seeking microbiological confirmation of suspected EV Funding: Dutch Ministry of Health, Welfare and Sports. cases, and this is done according to the expert opinion of the clinician in charge. This makes the system sensitive to variations in the frequency of EV testing, Conflict of interest although we tried to control for this by analysing E6 None declared. as a proportion of all EV. Finally, since there are other laboratories in the Netherlands performing EV typing, not all EV clinical identifications in the Netherlands Authors’ contributions were captured by this study and the exact population ED, KB and SM initiated the study. SM, KB and LS performed coverage is not known. Environmental surveillance in the analyses under the supervision of ED and JW. SM draft- the Netherlands is optimised for poliovirus exclusion ed the first version of the manuscript. SM, KB, LS, RP, SH, in specific areas and may not provide a good represen- JW and ED were involved in interpreting the results, made tation of the general population, especially since the substantial contributions to the discussion, and critically area covered (the bible belt) is considered to be rela- reviewed the manuscript. All the authors have read and approved the final version. tively socially isolated and circulating strains can differ from other regions [24,32]. References The VP1 fragment analysed in our study covered only 1. Centers for Disease Control and Prevention (CDC). Non-polio 270 nt. While EV types can be assigned with this frag- enterovrius. Atlanta: CDC. [Accessed: 20 Mar 2018]. Available ment, the fragment is too short to carry out a more in from: www.cdc.gov/non-polio-enterovirus/about/overview. depth phylogenetic analysis. As such, the study only html 2. European Centre for Disease Prevention and Control (ECDC). gives a first insight into possible clustering, which Enterovirus detections associated with severe neurological should be confirmed with complete VP1 sequencing. symptoms in children and adults in European countries, 8 August 2016. Rapid risk assessment. Stockholm: ECDC; 2016. We used a 15% threshold for genetic homologues, Available from: https://ecdc.europa.eu/en/publications-data/ which corresponds to the level of genetic variability rapid-risk-assessment-enterovirus-detections-associated- severe-neurological observed in other studies. In Greece, in patients with 3. European Centre for Disease Prevention and Control (ECDC). CNS infection in 2006 and 2007, mean genetic distance Outbreak of enterovirus A71 with severe neurological between groups was 22% [8] and in Poland, includ- symptoms among children in Catalonia, Spain, 16 June 2016. Rapid risk assessment. Stockholm: ECDC; 2016. Available ing also environmental isolates, nucleotide sequence from: https://ecdc.europa.eu/en/publications-data/rapid-risk- divergence was up to 19% [41]. As a limitation of the assessment-outbreak-enterovirus-a71-severe-neurological- symptoms-among time–place–phylogeny clustering algorithm, we used 4. Grist NR, Bell EJ, Assaad F. Enteroviruses in human disease. 15% of the maximum distance as a threshold for clus- Prog Med Virol. 1978;24:114-57. PMID: 360295 ter detection, and the number and composition of the 5. Khetsuriani N, Lamonte A, Oberste MS, Pallansch M. Neonatal enterovirus infections reported to the national enterovirus resulting clusters is sensitive to this choice. Using surveillance system in the United States, 1983-2003. Pediatr 20% of the maximum distance or higher resulted in Infect Dis J. 2006;25(10):889-93. https://doi.org/10.1097/01. inf.0000237798.07462.32 PMID: 17006282 transmission clusters that were less consistent with 6. de Ory F, Avellón A, Echevarría JE, Sánchez-Seco MP, the phylogenetic clusters previously defined, showed Trallero G, Cabrerizo M, et al. Viral infections of the central less plausibility and failed to identify the E6 cluster nervous system in Spain: a prospective study. J Med Virol. 2013;85(3):554-62. https://doi.org/10.1002/jmv.23470 PMID: detected in 2016 [10]. 23239485 7. Lee HY, Chen CJ, Huang YC, Li WC, Chiu CH, Huang CG, et al. Clinical features of echovirus 6 and 9 infections in children.

18 www.eurosurveillance.org J Clin Virol. 2010;49(3):175-9. https://doi.org/10.1016/j. 2018;101:11-7. https://doi.org/10.1016/j.jcv.2018.01.008 jcv.2010.07.010 PMID: 20729140 PMID: 29414181 8. Papa A, Skoura L, Dumaidi K, Spiliopoulou A, Antoniadis A, 24. Benschop KSM, van der Avoort HG, Jusic E, Vennema H, van Frantzidou F. Molecular epidemiology of Echovirus 6 in Greece. Binnendijk R, Duizer E. Polio and measles down the drain: Eur J Clin Microbiol Infect Dis. 2009;28(6):683-7. https://doi. Environmental enterovirus surveillance in the Netherlands, org/10.1007/s10096-008-0685-1 PMID: 19130106 2005 to 2015. Appl Environ Microbiol. 2017;83(13):e00558-17. 9. van der Sanden SM, Koopmans MP, van der Avoort HG. https://doi.org/10.1128/AEM.00558-17 PMID: 28432101 Detection of human enteroviruses and parechoviruses as part 25. Hovi T, Shulman LM, van der Avoort H, Deshpande J, Roivainen of the national enterovirus surveillance in the Netherlands, M, DE Gourville EM. Role of environmental poliovirus 1996-2011. Eur J Clin Microbiol Infect Dis. 2013;32(12):1525-31. surveillance in global polio eradication and beyond. https://doi.org/10.1007/s10096-013-1906-9 PMID: 23780695 Epidemiol Infect. 2012;140(1):1-13. https://doi.org/10.1017/ 10. Benschop KS, Geeraedts F, Beuvink B, Spit SA, Fanoy EB, S095026881000316X PMID: 21849095 Claas EC, et al. Increase in ECHOvirus 6 infections associated 26. Bisseux M, Colombet J, Mirand A, Roque-Afonso AM, with neurological symptoms in the Netherlands, June to Abravanel F, Izopet J, et al. Monitoring human enteric viruses August 2016. Euro Surveill. 2016;21(39):30351. https://doi. in wastewater and relevance to infections encountered in the org/10.2807/1560-7917.ES.2016.21.39.30351 PMID: 27719751 clinical setting: a one-year experiment in central France, 2014 11. Harvala H, Jasir A, Penttinen P, Pastore Celentano L, Greco D, to 2015. Euro Surveill. 2018;23(7):17-00237. . Available from: Broberg E. Surveillance and laboratory detection for non-polio http://dx.ddoi.org/10.2807/1560-7917.ES.2018.23.7.17-00237 enteroviruses in the European Union/European Economic https://doi.org/10.2807/1560-7917.ES.2018.23.7.17-00237 Area, 2016. Euro Surveill. 2017;22(45):16-00807. https:// PMID: 29471623 doi.org/10.2807/1560-7917.ES.2017.22.45.16-00807 PMID: 27. Wieczorek M, Ciąćka A, Witek A, Kuryk Ł, Żuk-Wasek A. 29162204 Environmental surveillance of non-polio enteroviruses in 12. Cabrerizo M, Trallero G, Echevarría JE, Moreno-Docón A, Poland, 2011. Food Environ Virol. 2015;7(3):224-31. https:// Pena MJ, Pérez-Ruiz M, et al. Molecular characterization of doi.org/10.1007/s12560-015-9195-3 PMID: 25862480 enteroviruses associated with neurological infections in Spain, 28. van der Avoort HG, Reimerink JH, Ras A, Mulders MN, van 2008. J Med Virol. 2013;85(11):1975-7. https://doi.org/10.1002/ Loon AM. Isolation of epidemic poliovirus from sewage jmv.23693 PMID: 23893817 during the 1992-3 type 3 outbreak in The Netherlands. 13. Harvala H, Calvert J, Van Nguyen D, Clasper L, Gadsby N, Epidemiol Infect. 1995;114(3):481-91. https://doi.org/10.1017/ Molyneaux P, et al. Comparison of diagnostic clinical samples S0950268800052195 PMID: 7781736 and environmental sampling for enterovirus and parechovirus 29. Savolainen-Kopra C, Paananen A, Blomqvist S, Klemola P, surveillance in Scotland, 2010 to 2012. Euro Surveill. Simonen ML, Lappalainen M, et al. A large Finnish echovirus 2014;19(15):20772. https://doi.org/10.2807/1560-7917. 30 outbreak was preceded by silent circulation of the same ES2014.19.15.20772 PMID: 24762664 genotype. Virus Genes. 2011;42(1):28-36. https://doi. 14. Wang H, Tao Z, Li Y, Lin X, Yoshida H, Song L, et al. org/10.1007/s11262-010-0536-x PMID: 20960045 Environmental surveillance of human enteroviruses in 30. Kyriakopoulou Z, Dedepsidis E, Pliaka V, Tsakogiannis D, Shandong Province, China, 2008 to 2012: serotypes, temporal Ruether IG, Krikelis A, et al. Complete nucleotide sequence fluctuation, and molecular epidemiology. Appl Environ analysis of the VP1 genomic region of Echoviruses 6 Microbiol. 2014;80(15):4683-91. https://doi.org/10.1128/ isolated from sewage in Greece revealed 98% similarity AEM.00851-14 PMID: 24837389 with Echoviruses 6 that were characterized from an 15. Tao Z, Song Y, Wang H, Zhang Y, Yoshida H, Ji S, et al. aseptic meningitis outbreak 1 year later. Clin Microbiol Intercity spread of echovirus 6 in Shandong Province, China: Infect. 2011;17(8):1170-3. https://doi.org/10.1111/j.1469- application of environmental surveillance in tracing circulating 0691.2010.03381.x PMID: 21489045 enteroviruses. Appl Environ Microbiol. 2012;78(19):6946-53. 31. Benschop KS, Rahamat-Langendoen JC, van der Avoort https://doi.org/10.1128/AEM.01861-12 PMID: 22843520 HG, Claas EC, Pas SD, Schuurman R, et al. VIRO-TypeNed, 16. Abedi GR, Watson JT, Pham H, Nix WA, Oberste MS, Gerber systematic molecular surveillance of enteroviruses in SI. Enterovirus and human parechovirus surveillance - the Netherlands between 2010 and 2014. Euro Surveill. United States, 2009-2013. MMWR Morb Mortal Wkly Rep. 2016;21(39):30352. https://doi.org/10.2807/1560-7917. 2015;64(34):940-3. https://doi.org/10.15585/mmwr. ES.2016.21.39.30352 PMID: 27719752 mm6434a3 PMID: 26334674 32. Ruijs WL, Hautvast JL, van der Velden K, de Vos S, Knippenberg 17. Papadakis G, Chibo D, Druce J, Catton M, Birch C. Detection H, Hulscher ME. Religious subgroups influencing vaccination and genotyping of enteroviruses in cerebrospinal fluid coverage in the Dutch Bible belt: an ecological study. BMC in patients in Victoria, Australia, 2007-2013. J Med Virol. Public Health. 2011;11(1):102. https://doi.org/10.1186/1471- 2014;86(9):1609-13. https://doi.org/10.1002/jmv.23885 2458-11-102 PMID: 21320348 PMID: 24474149 33. Duizer E, Rutjes S, de Roda Husman AM, Schijven J. Risk 18. Thoelen I, Lemey P, Van Der Donck I, Beuselinck K, Lindberg assessment, risk management and risk-based monitoring AM, Van Ranst M. Molecular typing and epidemiology following a reported accidental release of poliovirus in of enteroviruses identified from an outbreak of aseptic Belgium, September to November 2014. Euro Surveill. meningitis in Belgium during the summer of 2000. J Med Virol. 2016;21(11):30169. https://doi.org/10.2807/1560-7917. 2003;70(3):420-9. https://doi.org/10.1002/jmv.10412 PMID: ES.2016.21.11.30169 PMID: 27020766 12767006 34. Duizer E, Ruijs WL, van der Weijden CP, Timen A. Response 19. Luchs A, Russo DH, Cilli A, Costa FF, Morillo SG, Machado to a wild poliovirus type 2 (WPV2)-shedding event following BC, et al. Echovirus 6 associated to aseptic meningitis accidental exposure to WPV2, the Netherlands, April 2017. Euro outbreak, in São Joaquim da Barra, São Paulo, Brazil. Braz J Surveill. 2017;22(21):30542. https://doi.org/10.2807/1560- Microbiol. 2008;39(1):28-31. https://doi.org/10.1590/S1517- 7917.ES.2017.22.21.30542 PMID: 28597830 83822008000100007 PMID: 24031172 35. Nix WA, Oberste MS, Pallansch MA. Sensitive, seminested PCR 20. Siafakas N, Goudesidou M, Gaitana K, Gounaris A, Velegraki amplification of VP1 sequences for direct identification of all A, Pantelidi K, et al. Successful control of an echovirus 6 enterovirus serotypes from original clinical specimens. J Clin meningitis outbreak in a neonatal intensive care unit in central Microbiol. 2006;44(8):2698-704. https://doi.org/10.1128/ Greece. Am J Infect Control. 2013;41(11):1125-8. https://doi. JCM.00542-06 PMID: 16891480 org/10.1016/j.ajic.2013.02.014 PMID: 23706831 36. Kroneman A, Vennema H, Deforche K, v d Avoort H, Peñaranda 21. Kim HJ, Kang B, Hwang S, Hong J, Kim K, Cheon DS. Epidemics S, Oberste MS, et al. An automated genotyping tool for of viral meningitis caused by echovirus 6 and 30 in Korea in enteroviruses and noroviruses. J Clin Virol. 2011;51(2):121-5. 2008. Virol J. 2012;9(1):38. https://doi.org/10.1186/1743- https://doi.org/10.1016/j.jcv.2011.03.006 PMID: 21514213 422X-9-38 PMID: 22336050 37. Benschop KS, de Vries M, Minnaar RP, Stanway G, van 22. World Health Organization Regional Office for Europe (WHO der Hoek L, Wolthers KC, et al. Comprehensive full-length Europe). Enterovirus surveillance guidelines: Guidelines for sequence analyses of human parechoviruses: diversity and enterovirus surveillance in support of the Polio Eradication recombination. J Gen Virol. 2010;91(Pt 1):145-54. https://doi. Initiative. Copenhagen: WHO Europe; 2015. Available org/10.1099/vir.0.014670-0 PMID: 19759239 from: http://www.euro.who.int/en/publications/abstracts/ 38. Ragonnet-Cronin M, Hodcroft E, Hué S, Fearnhill E, Delpech enterovirus-surveillance-guidelines.-guidelines-for- V, Brown AJ, et al. Automated analysis of phylogenetic enterovirus-surveillance-in-support-of-the-polio-eradication- clusters. BMC Bioinformatics. 2013;14(1):317. https://doi. initiative org/10.1186/1471-2105-14-317 PMID: 24191891 23. Harvala H, Broberg E, Benschop K, Berginc N, Ladhani S, 39. Ypma RJ, Donker T, van Ballegooijen WM, Wallinga J. Susi P, et al. Recommendations for enterovirus diagnostics Finding evidence for local transmission of contagious and characterisation within and beyond Europe. J Clin Virol. disease in molecular epidemiological datasets. PLoS One. 2013;8(7):e69875. https://doi.org/10.1371/journal. pone.0069875 PMID: 23922835 www.eurosurveillance.org 19 40. Cabrerizo M, Trallero G, Simmonds P. Recombination and evolutionary dynamics of human echovirus 6. J Med Virol. 2014;86(5):857-64. https://doi.org/10.1002/jmv.23741 PMID: 24114692 41. Wieczorek M, Krzysztoszek A, Ciąćka A, Figas A. Molecular characterization of environmental and clinical echovirus 6 isolates from Poland, 2006-2014. J Med Virol. 2017;89(5):936- 40. https://doi.org/10.1002/jmv.24709 PMID: 27736044

License and copyright This is an open-access article distributed under the terms of the Creative Commons Attribution (CC BY 4.0) Licence. You may share and adapt the material, but must give appropriate credit to the source, provide a link to the licence, and indicate if changes were made.

This article is copyright of the authors or their affiliated institutions, 2018.

20 www.eurosurveillance.org Research article High rates of meticillin-resistant Staphylococcus aureus among asylum seekers and refugees admitted to Helsinki University Hospital, 2010 to 2017

Tuomas Aro1,2, Anu Kantele1,2,3 1. Department of Internal Medicine, Clinicum, Medical Faculty, University of Helsinki, Helsinki, Finland 2. Inflammation Center, University of Helsinki and Helsinki University Hospital, Helsinki, Finland 3. Karolinska Institutet, Stockholm, Sweden Correspondence: Anu Kantele ([email protected])

Citation style for this article: Aro Tuomas, Kantele Anu. High rates of meticillin-resistant Staphylococcus aureus among asylum seekers and refugees admitted to Helsinki University Hospital, 2010 to 2017. Euro Surveill. 2018;23(45):pii=1700797. https://doi.org/10.2807/1560-7917.ES.2018.23.45.1700797

Article submitted on 30 Nov 2017 / accepted on 28 Apr 2018 / published on 08 Nov 2018

Introduction: Antimicrobial resistance is increasing Introduction rapidly in countries with low hygiene levels and poorly Antimicrobial resistance (AMR), a major health prob- controlled antimicrobial use. The spread of resist- lem worldwide, surges most rapidly in regions with low ant bacteria poses a threat to healthcare worldwide. level of hygiene and poor control of antimicrobial use Refugees and migrants from high-prevalence coun- [1]. AMR spreads across the globe and its extent has tries may add to a rise in multidrug-resistant (MDR) been recognised by international bodies at the high- bacteria in low-prevalence countries. However, respec- est level: in 2016 AMR was addressed at a General tive data are scarce. Methods: We retrospectively col- Assembly session of the United Nations as the greatest lected microbiological and clinical data from asylum and most urgent global health risk [2]. seekers and refugees treated at Helsinki University Hospital between January 2010 and August 2017. A major concern about AMR is its spread to healthcare Results: Of 447 asylum seekers and refugees (Iraq: settings in low-prevalence countries with severe con- 46.5%; Afghanistan: 10.3%; Syria: 9.6%, Somalia: sequences: treatment failures, increase in the number 6.9%); 45.0% were colonised by MDR bacteria: 32.9% of serious infections, and dramatic cost implications had extended-spectrum beta-lactamase-producing [3]. To prevent the spread of AMR to hospitals, patients Enterobacteriaceae (ESBL-PE), 21.3% meticillin-resist- with particular risk of colonisation and infection with ant Staphylococcus aureus (MRSA), 0.7% carbap- multidrug-resistant (MDR) bacteria should be identi- enemase-producing Enterobacteriaceae (CPE), 0.4% fied and subjected to infection control measures at the multiresistant Pseudomonas aeruginosa (MRPA), 0.4% admission stage. Numerous studies have identified multiresistant Acinetobacter baumannii (MRAB); no international travel as a major risk factor for colonisa- vancomycin-resistant Enterococcus (VRE) were found. tion: approximately one third of visitors to high-prev- Two or more MDR bacteria strains were recorded for alence regions acquire MDR bacteria during ordinary 12.5% of patients. Multivariable analysis revealed tourist travel [4-18]. Refugees and migrants who have geographical region and prior surgery outside Nordic lived for years in high-risk regions are presumed to countries as risk factors of MRSA colonisation. Young have even higher levels of colonisation. age (< 6 years old), short time from arrival to first sample, and prior hospitalisation outside Nordic countries The European migrant crisis began in 2015 when over 1.2 were risk factors of ESBL-PE colonisation. Conclusion: million first-time asylum seekers (this group includes We found MDR bacterial colonisation to be common both refugees and migrants) applied for international among asylum seekers and refugees arriving from protection in European Union countries [19,20]. Finland current conflict zones. In particular we found a high received 32,476 applications for asylum in 2015, an prevalence of MRSA. Refugees and migrants should, almost 10-fold increase on previous years [21]. therefore, be included among risk populations requir- ing MDR screening and infection control measures at Like international travel, migration may contribute sub- hospitals. stantially to the spread of AMR [22-26]. Refugees and migrants mostly come from countries with considerably higher rates of MDR bacteria than Finland and, moreover, they may have journeyed through other

www.eurosurveillance.org 21 Figure 1 multiresistant Acinetobacter baumannii (MRAB) and Flowchart showing multidrug-resistant bacteria found in multiresistant Pseudomonas aeruginosa (MRPA). The samples from asylum seekers and refugees admitted to Helsinki University Hospital, Finland, January 2010 to same screening guidelines were applied to patients August 2017 (n = 447) who had been hospitalised outside the Nordic countries during the previous 12 months before admission

Patients tested for both to our hospital. MRSA and MRGN n = 6,423 Using the HUCH infectious diseases database, SAI, we compiled a list of patients who had been sampled for Origin of family name: Origin of family name: both MRSA and multiresistant Gram-negative (MRGN) Finnish/Swedish non-Finnish/non-Swedish n = 4,104 n = 2,319 bacteria at hospital admission. Among these, we selected those with a non-Finnish/non-Swedish name

Refugees/ Non-refugees/ and, after screening their patient charts, included only asylum seekers asylum seekers n = 447 n = 1,872 those who were asylum seekers or refugees (Figure 1). According to the Finnish Medical Research Act, review by an ethics committee is only required for research MRSA or MRGN MRSA and MRGN positive negative involving an intervention. The study protocol was n = 201 n = 246 approved by the research board of the Inflammation Center, Helsinki University Hospital, Finland.

ESBL-PE MRSA n = 147 n = 95 Collection of patient data For background information, we collected data from the patient records on sex, age, country of origin, date VRE CPE of arrival in Finland, prior hospitalisation and surgery n = 0 n = 3 (as recorded by the clinician), and determined Charlson Comorbidity Index (CCI) [27] for each subject. In addi-

MRPA MRAB tion, we collected data covering the results of bacte- n = 2 n = 2 rial cultures (blood, urine, stool), reason for admission, clinical diagnosis (ICD-10) [28] at discharge, MDR bacterial infections identified, and deaths. For further CPE: carbapenemase-producing Enterobacteriaceae; ESBL-PE: analysis, patients were grouped by geographical region extended-spectrum beta-lactamase-producing Enterobacteriaceae; MRAB: multiresistant Acinetobacter baumannii; MRGN: according to their country of origin. Here we applied a multiresistant Gram-negative bacteria; MRPA: multiresistant classification (Figure 2) modified from United Nations Pseudomonas aeruginosa; MRSA: meticillin-resistant Staphylococcus aureus; VRE: vancomycin-resistant Enterococcus. geoscheme (Europe, North Africa and Middle East, sub-Saharan Africa, Asia, other) [29].

Microbiological methods According to the hospital guidelines, swabs for screen- high-prevalence regions [1]. Accurate data on colonisa- ing patients for MDR bacteria are collected as follows: tion rates in this population are required to estimate MRSA samples are taken each with a separate swab transmission risk and to prepare infection control from the nostrils (one swab for both), pharynx and guidelines for hospitals. This retrospective study inves- rectum or perineum. MRGN bacteria samples are col- tigates the prevalence of various MDR bacteria among lected as rectal swabs. In addition, swabs are taken asylum seekers and refugees hospitalised in Finland, from wound infections when applicable. The screening and seeks risk factors that can be used to identify comprises two sets of samples, and where possible the those at highest risk of colonisation. samples are to be collected on consecutive days.

Methods While some minor modifications to the routine laboratory practices took place during the study period, at Selection of participants the time of the last sampling, MDR bacterial analyses Helsinki University Hospital (HUCH) provides second- were carried out as follows: MRSA was screened by ary and tertiary care for the 1.6 million inhabitants of overnight enrichment on eMRSA broth (Copan Italia, southern Finland. During the study period, from January Brescia, Italy) or selective in-house MRSA enrichment 2010 to August 2017, our hospital’s infection control broth [30] followed by culture on CHROMagar MRSA guidelines stated that all asylum seekers and refugees (CHROMagar, Paris, France), and confirmed with S. admitted to hospitals should be screened at entry for aureus-specific nuclease and mecA gene qPCR [30]. meticillin-resistant Staphylococcus aureus (MRSA), VRE were screened by enrichment Enterococcosel vancomycin-resistant Enterococcus (VRE), Broth (BBL, Cockeysville, MD, United States of America extended-spectrum beta-lactamase-producing (USA)) and followed by culture on CHROMagar VRE Enterobacteriaceae (ESBL-PE), carbapene- media. Positive findings were confirmed by in-house mase-producing Enterobacteriaceae (CPE), PCR as described by Suppola et al. [31].

22 www.eurosurveillance.org Figure 2 Rate of multidrug-resistant bacteria among asylum seekers and refugees admitted to Helsinki University Hospital, by country of origin or geographical region, Finland, January 2010 to August 2017 (n = 447)

Percentage of MDR carriers by country/region of origin

0–20 % Europe (incl. Russia) 21–40 %

41–60 % MDR 15.4 % ESBL–PE 11.5 % MRSA 7.7 % n = 26

North Africa & Middle East Asia MDR 56.0% ESBL-PE 41.4% MDR 38.6 % MRSA 28.0% ESBL–PE 31.6 % n = 268 MRSA 8.8 % n = 57

Afghanistan

MDR 34.8 % Sub-Saharan Africa ESBL–PE 26.1 % MRSA 10.9 % MDR 24.4% n = 46 ESBL-PE 15.1% MRSA 12.8% Syria Iraq n = 86

MDR 55.8 % MDR 57.2 % ESBL–PE 48.8 % ESBL–PE 40.4 % MRSA 20.9 % MRSA 31.3 % n = 43 n = 208

ESBL-PE: extended-spectrum beta-lactamase-producing Enterobacteriaceae; MDR: multidrug-resistant bacteria; MRSA: meticillin-resistant Staphylococcus aureus.

For nine patients, country of origin was not known. One patient was from Mexico (not shown in the map).

Patients’ countries of origin in the various geographical regions:

Europe: Albania, Georgia, Greece, Kosovo*, Russia, Serbia, Ukraine;

North Africa and the Middle East: Iraq, Iran, Jordan, Kuwait, Lebanon, Morocco, Syria, Turkey, Yemen;

Sub-Saharan Africa: Burundi, Cameroon, Congo, Ethiopia, Eritrea, Gambia, Ghana, Kenya, Nigeria, Rwanda, Somalia, South Africa, Sudan, Uganda, Zambia;

Asia: Afghanistan, Azerbaijan, Kazakhstan, China, Kyrgyzstan, Myanmar, Nepal, Pakistan, Sri Lanka.

*This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the International Court of Justice Opinion on the Kosovo Declaration of Independence.

www.eurosurveillance.org 23 Table 1 Background characteristics of asylum seekers and refugees admitted to Helsinki University Hospital, by geographical region of origina, Finland, January 2010–August 2017 (n = 447 patients)

North Africa and Sub-Saharan Other or Europe Asia Total Middle East Africa unknown

Patient attributes n = 26 n = 57 n = 447 n = 268 n = 86 n = 10 n % n % n % n % n % n % Sex Male 8 30.8 135 50.4 32 37.2 31 54.4 4 40.0 210 47.0 Female 18 69.2 133 49.6 54 62.8 26 45.6 6 60.0 237 53.0 Median age (years) 24 26 27 20 27 25 Age group (years) 0–5 3 11.5 36 13.4 7 8.1 6 10.5 2 20.0 54 12.1 6–15 3 11.5 32 11.9 3 3.5 15 26.3 0 0.0 53 11.9 16–25 9 34.6 64 23.9 31 36.0 14 24.6 3 30.0 121 27.1 26–35 8 30.8 98 36.6 30 34.9 12 21.1 4 40.0 152 34.0 > 35 3 11.5 38 14.2 15 17.4 10 17.5 1 10.0 67 15.0 Charlson Comorbidity Index 0 points 24 92.3 251 93.7 80 93.0 52 91.2 10 100.0 417 93.3 1 point 0 0.0 3 1.1 1 1.2 2 3.5 0 0.0 6 1.3 2–3 points 2 7.7 11 4.1 4 4.7 3 5.3 0 0.0 20 4.5 4–5 points 0 0.0 1 0.4 1 1.2 0 0.0 0 0.0 2 0.4 > 5 points 0 0.0 2 0.7 0 0.0 0 0.0 0 0.0 2 0.4 Mean time from arrival to Finland 78 127 100 119 363 119 to first sample (days) Prior treatment history Prior hospitalisation abroad 5 19.2 54 20.1 15 17.4 6 10.5 0 0.0 80 17.9 Prior invasive procedure abroad 2 7.7 34 12.7 9 10.5 4 7.0 0 0.0 49 11.0 aPatients’ countries of origin in the various geographical regions: Europe: Albania, Georgia, Greece, Kosovo*, Russia, Serbia, Ukraine; North Africa and Middle East: Iraq, Iran, Jordan, Kuwait, Lebanon, Morocco, Syria, Turkey, Yemen; Sub-Saharan Africa: Burundi, Cameroon, Congo, Ethiopia, Eritrea, Gambia, Ghana, Kenya, Nigeria, Rwanda, Somalia, South Africa, Sudan, Uganda, Zambia; Asia: Afghanistan, Azerbaijan, Kazakhstan, China, Kyrgyzstan, Myanmar, Nepal, Pakistan, Sri Lanka; Other or unknown: Mexico, unknown. *This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the International Court of Justice Opinion on the Kosovo Declaration of Independence.

ESBL-PE and CPE were analysed by plating directly on SPSS 24.0.0.0 software (IBM Corp., Armonk, NY, CHROMagar ESBL and CHROMagar KPC or CHROMagar USA). In univariate analyses for categorical variables, mSuperCARBA, respectively. ESBL-PE species iden- chi-squared test, Fisher’s exact test or binary logistic tification was confirmed by MALDI-TOF (Vitek-MS, regression analysis was applied. For continuous vari- bioMérieux, Marcy l’Étoile, France) and resistance by ables, we used binary logistic regression. Chi-squared standard EUCAST method [32]. CPE species were con- test and Fisher’s exact test were two-sided. Variables firmed with in-house carbapenemase gene PCR. for the multivariable model were selected using the p value limit of 0.2 in the univariate model. Time MDR-P. aeruginosa and MDR-A. baumannii were between arrival and first sample was not known for all screened from ESBL and KPC SuperCARBA plates. cases; missing values were taken into account by mul- Cultures were tested by MALDI-TOF for species tiple imputations, assuming that data were missing at identification. Isolates resistant to meropenem random. In the multivariable model, we used backward for Acinetobacter, and both meropenem and ceftazi- selection with Akaike information criteria (AIC) so as dime for Pseudomonas, were analysed by PCR for car- to choose the most informative explanatory variables bapenemase genes as previously described [33]. for the final model. From several highly correlated variables, only one was included. Statistics Data were entered on Microsoft Excel 2013 spread- sheets, and statistical analyses were conducted using

24 www.eurosurveillance.org Table 2 Background characteristics of asylum seekers and refugees admitted to Helsinki University Hospital, by country of origina, Finland, January 2010–August 2017 (n = 447 patients)

Iraq Afghanistan Syria Somalia Nigeria Other or unknown Total

Patient attributes n = 208 n = 46 n = 43 n = 31 n = 16 n = 103 n = 447 n % n % n % n % n % n % n % Sex Male 109 52.4 25 54.3 19 44.2 11 35.5 4 25.0 42 40.8 210 47.0 Female 99 47.6 21 45.7 24 55.8 20 64.5 12 75.0 61 59.2 237 53.0 Median age (years) 26 19 25 22 30 26 25 Age group (years) 0–5 27 13.0 4 8.7 5 11.6 1 3.2 1 6.3 16 15.5 54 12.1 6–15 20 9.6 12 26.1 9 20.9 2 6.5 0 0.0 10 9.7 53 11.9 16–25 54 26.0 13 28.3 9 20.9 17 54.8 4 25.0 24 23.3 121 27.1 26–35 76 36.5 10 21.7 15 34.9 8 25.8 10 62.5 33 32.0 152 34.0 > 35 31 14.9 7 15.2 5 11.6 3 9.7 1 6.3 20 19.4 67 15.0 Charlson Comorbidity Index 0 points 197 94.7 43 93.5 38 88.4 28 90.3 16 100.0 95 92.2 417 93.3 1 point 1 0.5 2 4.3 1 2.3 1 3.2 0 0.0 1 1.0 6 1.3 2–3 points 7 3.4 1 2.2 4 9.3 2 6.5 0 0.0 6 5.8 20 4.5 4–5 points 1 0.5 0 0.0 0 0.0 0 0.0 0 0.0 1 1.0 2 0.4 > 5 points 2 1.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 2 0.4 aThe data are only presented for the five most common countries (in total, the study population originated in 41 different countries).

Results (13.3%; 11/83), or presented as dermatological infection (16.9%; 14/83) or acute gastroenteritis (7.2%; 6/83). Subject characteristics The majority of patients (81.4%) had a non-infectious According to the HUCH infectious diseases database disease and nearly one third (29.8%; 133/447) of the SAI, 6,423 patients were screened for both MRSA and admissions were associated with pregnancy. MRGN bacteria at admission between 1 January 2010 and 23 August 2017. A total of 2,319 patients with a Findings of multidrug-resistant bacteria non-Finnish/non-Swedish name were selected and Almost half of the patients (45.0%; 201/447) were col- their patient records screened (Figure 1). The final study onised by MDR bacteria: 32.9% had ESBL-PE, 21.3% population included 447 patients with refugee or asy- MRSA, 0.7% CPE, 0.4% MRPA and 0.4% MRAB (Table lum seeker status stated in their patient records. The 3). No cases with VRE were recorded. Two or more increase in the number of refugees and asylum seekers MDR strains were found in 12.5%. The proportion of admitted during the study period was evident: the vast MDR bacterial carriers was found greatest among the majority (86.8%) were hospitalised after the beginning patients from Iraq and Syria (57.2% and 55.8% respec- of 2015. As shown by the demographic data presented tively), followed by Afghanistan (34.8%) and Somalia in Tables 1 and 2, their median age was 25 years, and (25.8%) (Table 4). Analysis by geographical region 53.0% were female. The patients had 41 different coun- (Table 3) showed the highest MDR bacterial colonisa- tries of origin, the majority in Iraq (46.5%), Afghanistan tion rates for patients from North Africa and the Middle (10.3%), Syria (9.6%), and Somalia (6.9%). Most of the East (56.0%). Patients from Asia and sub-Saharan subjects (93.3%) were healthy, with a CCI-score of 0. Africa had MDR bacterial carriage rates of 38.6% and Prior hospitalisation outside the Nordic countries was 24.4%, respectively, while the result for those from recorded for 17.9%, an invasive procedure for 11.0%, Europe was 15.4% (Figure 2). and treatment in an intensive care unit abroad for 4.0% of the patients. The median time from arrival in Finland Infections with multidrug-resistant bacteria to first MDR bacterial sampling was 59 days. The date Ten of the MDR bacterial carriers (5.0%; 10/201) had a of arrival in Finland was available for 283/447 patients clinical MDR bacterial infection verified by culture, with (63.3%). wound infections and urinary tract infections (UTI) as the most common manifestations. In six cases (60.0%; Cause of hospitalisation 6/10) the agent was identified as ESBL-PE, and in four An infectious disease was the primary diagnosis for (40.0%; 4/10) as MRSA; 4.1% (6/147) of the ESBL-PE 83/447 patients (18.6%), most commonly confined and 4.2% (4/95) of the MRSA carriers presented with to the respiratory (21.7%; 18/83) or urogenital tract a clinical infection. Among patients with a clinical

www.eurosurveillance.org 25 Table 3 The number of carriers of multidrug-resistant bacteria among asylum seekers and refugees admitted to Helsinki University Hospital, by geographical region of origina, Finland, January 2010-August 2017 (n = 447 patients)

North Africa and Sub-Saharan Europe Asia Other or unknown Total Middle-East Africa

Patient attributes n = 26 n = 57 n = 10 n = 447b n = 268 n = 86 n % n % n % n % n % n % Any MDR bacteria 4 15.4 150 56.0 21 24.4 22 38.6 4 40.0 201 45.0 MRSA 2 7.7 75 28.0 11 12.8 5 8.8 2 20.0 95 21.3 ESBL-PE 3 11.5 111 41.4 13 15.1 18 31.6 2 20.0 147 32.9 VRE 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 MRAB 0 0.0 1 0.4 1 1.2 0 0.0 0 0.0 2 0.4 MRPA 1 3.8 1 0.4 0 0.0 0 0.0 0 0.0 2 0.4 CPE 0 0.0 3 1.1 0 0.0 0 0.0 0 0.0 3 0.7 Multiple MDR bacteria

2 7.7 48 17.9 5 5.8 1 1.8 0 0.0 56 12.5 ≥ 2 classes or strains of MDR bacteria Clinical infection with MDR bacteria 0 0.0 8 3.0 1 1.2 1 1.8 0 0.0 10 2.2 Proportion of MDR bacteria carriers with clinical MDR bacterial infection 0 5.3 4.8 4.5 0 5.0 (%)

CPE: Carbapenemase-producing Enterobacteriaceae; ESBL-PE: extended-spectrum beta-lactamase-producing Enterobacteriaceae; MDR: multidrug-resistant; MRAB: multiresistant Acinetobacter baumannii; MRPA: multiresistant Pseudomonas aeruginosa; MRSA: meticillin- resistant Staphylococcus aureus; VRE: Vancomycin-resistant Enterococcus. aPatients’ countries of origin in the various geographical regions: Europe: Albania, Georgia, Greece, Kosovo*, Russia, Serbia, Ukraine; North Africa and Middle East: Iraq, Iran, Jordan, Kuwait, Lebanon, Morocco, Syria, Turkey, Yemen; Sub-Saharan Africa: Burundi, Cameroon, Congo, Ethiopia, Eritrea, Gambia, Ghana, Kenya, Nigeria, Rwanda, Somalia, South Africa, Sudan, Uganda, Zambia; Asia: Afghanistan, Azerbaijan, Kazakhstan, China, Kyrgyzstan, Myanmar, Nepal, Pakistan, Sri Lanka; Other or unknown: Mexico, unknown. bSome patients had two or more different ESBL-PE, CPE or MRSA strains. *This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the International Court of Justice Opinion on the Kosovo Declaration of Independence.

infection as primary diagnosis, 10.8% (9/83) had an t044 (6.4%), and t127 (6.4%). Spa-types seemed to MDR bacterial infection. differ by country of origin (Table 4). The median times from arrival in Finland to first sample among MRSA- UTIs were caused by ESBL E. coli (one patient) or carriers and non-carriers were 97 and 51 days, respec- ESBL Klebsiella pneumoniae (one patient). As for wound tively (Table 5); the difference was not statistically infections, MRSA was cultured in four cases (one with significant. osteomyelitis) and ESBL Enterobacter cloacae in one. One patient with a wound infection had ESBL Proteus Multiresistant Gram-negative bacteria findings mirabilis detected both in bacterial culture of wound CPE strains (K. pneumoniae, E. coli, Acinetobacter bau- tissue and blood. Furthermore, there was one patient mannii, and Proteus mirabilis) were found in the sam- with a finding of ESBL Shigella, and in one case ples of two patients from Iraq and one from Syria; both ESBL E. coli was cultured from perianal abscess. No Iraqis had multiple CPE strains. MRPA was recorded for deaths were directly attributed to infections with MDR two patients, one Russian, the other Iraqi. One refugee bacteria. from Cameroon and one from Iraq were screened positive for MRAB. Meticillin-resistant Staphylococcus aureus findings One third of the patients (32.9%; 147/447) were colo- The frequency of MRSA findings was 21.3% (95/447) nised by ESBL-PE, with the highest frequency seen (Table 3). The highest rates were seen among patients among patients of North African and Middle Eastern ori- coming from Iraq (31.3%; 65/208) and Syria (20.9%; gin (41.4%; 111/268). Nearly half of the Syrians (48.8%; 9/43), followed by those from Somalia (12.9; 4/31) and 21/43) were ESBL-PE carriers, while Iraqis and Afghans Afghanistan (10.9%; 5/46) (Table 4). The most com- had carriage rates of 40.4% (84/208) and 26.1% (12/46), mon Staphylococcal protein A (spa)-types of MRSA respectively (Table 4). Most ESBL-PE strains were E. isolates were t304 (24.4%), t386 (11.5%), t223 (10.3%), coli (90.2%), followed by K. pneumoniae (6.1%), Proteus

26 www.eurosurveillance.org Table 4 The number of carriers of multidrug-resistant bacteria among asylum seekers and refugees admitted to Helsinki University Hospital, as presented by five most common countries of origina, Finland, January 2010–August 2017 (n = 447 patients)

Other or Iraq Afghanistan Syria Somalia Nigeria Total unknown

Patient attributes n = 208 n = 46 n = 43 n = 31 n = 16 n = 447b n = 103 n % n % n % n % n % n % n % Any MDR bacteria (%) 119 57.2 16 34.8 24 55.8 8 25.8 2 12.5 32 31.1 201 45.0 MRSA 65 31.3 5 10.9 9 20.9 4 12.9 1 6.3 11 10.7 95 21.3 ESBL-PE 84 40.4 12 26.1 21 48.8 6 19.4 1 6.3 23 22.3 147 32.9 VRE 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 MRAB 1 0.5 0 0.0 0 0.0 0 0.0 0 0.0 1 1.0 2 0.4 MRPA 1 0.5 0 0.0 0 0.0 0 0.0 0 0.0 1 1.0 2 0.4 CPE 2 1.0 0 0.0 1 2.3 0 0.0 0 0.0 0 0.0 3 0.7 n % n % n % n % n % n % n % Multiple MDR bacteria

41 19.7 1 2.2 7 16.3 3 9.7 0 0.0 4 3.9 56 12.5 ≥ 2 classes or strains of MDR bacteria Clinical infection with MDR 5 2.4 1 2.2 3 7.0 1 3.2 0 0.0 0 0.0 10 2.2 bacteria Proportion of MDR bacteria carriers with clinical MDR 4.2 6.3 12.5 12.5 0 0 5.0 bacterial infection (%) spa- spa- spa- spa- spa- spa- spa- n n n n n n n type type type type type type type t304 17 t304 2 2 t021 1 t127 t304 19

t386 8 t786 2 2 t223 1 t852 t127 1 t386 9 The most common c t223 6 t044 1 MRSA spa-types 1 t363 1 t037 t1784 1 t1931 1 t223 8

t044 4 t223 1 1 t790 1 t267 t8154 1 t044 5

t690 3 t267 1 1 t6238 1 t346 t127 4 t991 3 t386 1

CPE: carbapenemase-producing Enterobacteriaceae; ESBL-PE: extended-spectrum beta-lactamase-producing Enterobacteriaceae; MDR: multidrug-resistant; MRAB: Multiresistant Acinetobacter baumannii; MRPA: multiresistant Pseudomonas aeruginosa; MRSA: meticillin- resistant Staphylococcus aureus; spa: staphylococcal protein A; VRE: vancomycin-resistant Enterococcus. aIn total, the study population originated in 41 different countries. bSome patients had two or more different ESBL-PE, CPE or MRSA strains. cThe spa-types of some MRSA samples were not available.

mirabilis (2.4%) and Enterobacter cloacae (0.6%). The (OR 1.7, 95% CI: 0.9–3.3) in univariate analysis. No sta- median time from arrival in Finland to first sample was tistical differences were seen in co-resistance between significantly shorter among ESBL-PE-carriers (38 days) the various geographical regions. than non-carriers (83 days) (Table 5). Risk factor analysis of meticillin- Extended-spectrum beta-lactamase-producing resistant Staphylococcus aureus colonisation Enterobacteriaceae with co-resistance to other In multivariable analysis, geographical region and prior antimicrobials invasive procedure outside the Nordic countries were Co-resistance to other antibiotics proved common found to be independent risk factors of MRSA colonisa- among the ESBL strains: two thirds (62.2%; 102/164) tion. No other risk factors were identified. (Table 6) of the ESBL-PE isolates showed decreased sensitivity to co-trimoxazole and more than one third were co- Risk factor analysis of extended-spectrum resistant to levofloxacin (40.2%; 66/164) or tobramycin beta-lactamase-producing Enterobacteriaceae (34.1%; 56/164). Resistance to levofloxacin was associ- colonisation ated with co-resistance to tobramycin (OR 6.0, 95% con- We identified by univariate analysis the follow- fidence intervals (CI): 3.0–12.2) but not co-trimoxazole ing factors as predisposing to ESBL-PE: young age,

www.eurosurveillance.org 27 Table 5 The results of univariate and multivariable risk factor analyses of extended-spectrum beta-lactamase-producing Enterobacteriaceae carriage among asylum seekers and refugees admitted to Helsinki University Hospital, Finland, January 2010–August 2017 (n = 447)

Patients ESBL-PE Non-carriers OR (95% Cl) p value in AOR (95% CI) p value in carriers in univariate univariate in multivariable multivariable Risk factor (n = 447) (n = 147) (n = 300) analysis analysis analysisa analysis n n % n % Sex Female 237 68 28.7 169 71.3 1.0 NA NA NA Male 210 79 37.6 131 62.4 1.5 (1.0–2.2) 0.045 NA NA Age group (years) < 0.001 0–5 54 30 55.6 24 44.4 1.0 NA 1.0 NA 6–15 53 16 30.2 37 69.8 0.3 (0.2–0.8)b 0.009 0.2 (0.1–0.5)b 0.001 16–25 121 30 24.8 91 75.2 0.3 (0.1–0.5)b < 0.001 0.3 (0.2–0.7)b 0.003 26–35 152 44 28.9 108 71.1 0.3 (0.2–0.6)b < 0.001 0.4 (0.2–0.8)b 0.010 > 35 67 27 40.3 40 59.7 0.5 (0.3–1.1)b 0.096 0.5 (0.2–1.2)b 0.123 Geographical regionc < 0.001 Europe 26 3 11.5 23 88.5 1.0 NA 1.0 NA North Africa and 268 111 41.4 157 58.6 5.4 (1.6–18.5)d 0.007 7.7 (2.1–28.1)d 0.002 Middle East Sub-Saharan Africa 86 13 15.1 73 84.9 1.4 (0.4–5.2)d 0.649 1.4 (0.4–5.9)d 0.617 Asia 57 18 31.6 39 68.4 3.5 (0.9–13.3)d 0.062 5.9 (1.4–24.1)d 0.014 Other or unknown 10 2 20.0 8 80.0 1.9 (0.3–13.6) d 0.516 7.0 (0.8–64.9)d 0.086 Prior hospitalisation abroad No or not specified 367 100 27.2 267 72.8 1.0 NA 1.0 NA Yes 80 47 58.8 33 41.3 3.8 (2.3–6.3) < 0.001 4.1 (2.3–7.4) < 0.001 Prior invasive procedure abroad No or not specified 398 118 29.6 280 70.4 1.0 NA NA NA Yes 49 29 59.2 20 40.8 3.4 (1.9–6.3) < 0.001 NA NA Prior ICU care abroad No or not specified 429 134 31.2 295 68.8 1.0 NA NA NA Yes 18 13 72.2 5 27.8 5.7 (2.0–16.4) 0.001 NA NA Time from arrival to first sample (days, 59 38 83 0.91 (0.85–0.96)e 0.002 0.91 (0.86–0.97) 0.002 median)e Charlson Comorbidity Index (points) 0.063 0 417 133 31.9 284 68.1 1.0 NA NA NA 1 6 1 16.7 5 83.3 0.4 (0.05–3.7)f 0.439 NA NA 2–3 20 13 65.0 7 35.0 4.0 (1.5–10.2)f 0.004 NA NA 4–5 2 0 0 2 100 NA NA NA NA > 5 2 0 0 2 100 NA NA NA NA

AOR: adjusted odds ratio; CI: confidence interval; ESBL-PE: extended-spectrum beta-lactamase-producing Enterobacteriaceae; ICU: intensive care unit; NA: not applicable; OR: odds ratio; SD: standard deviation. aWhen selecting the variables for the multivariable model with binary logistic regression, the Akaike information criteria (AIC) was used. Time between arrival and first sample was not known for all cases; missing values were taken into account by multiple imputations, assuming that data were missing at random. bCompared with the youngest age group. cPatients’ countries of origin in the various geographical regions: Europe: Albania, Georgia, Greece, Kosovo*, Russia, Serbia, Ukraine; North Africa and Middle East: Iraq, Iran, Jordan, Kuwait, Lebanon, Morocco, Syria, Turkey, Yemen; Sub-Saharan Africa: Burundi, Cameroon, Congo, Ethiopia, Eritrea, Gambia, Ghana, Kenya, Nigeria, Rwanda, Somalia, South Africa, Sudan, Uganda, Zambia; Asia: Afghanistan, Azerbaijan, Kazakhstan, China, Kyrgyzstan, Myanmar, Nepal, Pakistan, Sri Lanka. dCompared with Europe. en = 283; only in 283/447 cases arrival date to Finland was known. Analysed as a continuous variable, OR and adjusted OR given per 30 days. fCompared with those with 0 points. *This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the International Court of Justice Opinion on the Kosovo Declaration of Independence.

28 www.eurosurveillance.org geographical region, short time from arrival to first As we expected, the clinical ESBL-PE infections gener- sample, prior hospitalisation, prior invasive procedure ally proved to be UTIs; MDR bacterial wound infections abroad, and prior ICU care (Table 5). The final multi- were mostly attributed to MRSA. Data on the frequency variable analysis revealed geographical region of ori- of MDR bacterial carriers developing an infection with gin, age under 6 years, short time from arrival to first the MDR bacterial strain are scarce. Among returning sample and prior hospitalisation abroad as independ- travellers treated on an infectious and tropical dis- ent risk factors for colonisation. As for the geographi- eases ward in France from 2012 to 2013, Epelboin et cal region of origin, the ESBL-PE colonisation rates al. [40] found that one in five MRGN bacterial carri- were significantly higher among North African, Middle ers identified (21.7%, 5/23) also had a clinical MRGN Eastern and Asian patients than Europeans (Table 5). bacterial infection, albeit in a very small sample. The Of factors with p value under 0.2 in univariate analysis, difference between our rates and those of Epelboin male sex, prior invasive procedure abroad, prior ICU et al. (5.0% vs 21.7%, respectively) can probably be care, and general health (CCI score) were eliminated explained by two points related to the selection of from the final model on the basis of the AIC. population. Firstly, our data included a high proportion of healthy obstetric patients with no MDR bacterial Discussion infections. Secondly, in the investigation by Epelboin Data remain scarce on AMR carriage rates in refugees all patients had a clinical infection, while ours looked and migrants [24]. The present investigation reveals retrospectively at all patients, admitted for any reason, MDR bacterial colonisation and a high prevalence of who at admission to a tertiary hospital had been sam- MRSA of 21.3% among asylum seekers and refugees pled for both MRSA and MRGN bacteria, and whom we seeking healthcare at a tertiary hospital in Finland. could identify as asylum seekers and refugees from their patient records. Indeed, of our patients who had Although focusing on AMR, we also collected general infectious disease as primary diagnosis, 10.8% had data on refugees and migrants and causes of their an MDR bacterial infection. In our previous study look- hospitalisation. Although 79% of the asylum seeking at 1,122 Finnish patients hospitalised abroad for ers who have arrived in Finland since 2015 have been diverse reasons, we found that 11.4% (38/333) of those male [34], in our data no clear sex difference in hos- colonised with MDR bacteria had an infection with the pital admission was seen; the discrepancy can prob- same MDR bacteria [39]. ably be explained by our high number of obstetric hospitalisations. Overall, our patients were in good Colonisation by MRSA general health; the greatest number of visits (29.8%) A major finding emerging from our investigation was were reported for the specialty of obstetrics, while that MRSA carriage in asylum seekers and refugees is infectious diseases accounted for fewer than one in considerably more common than expected. This pro- five cases. Respiratory, skin, and urinary tract infec- portion significantly exceeds those of German stud- tions were the most common causes among infectious ies that have reported rates from 5.6% to 9.8% (8/143 diseases. Obviously, these are also common causes of and 32/325) [37,41], and rates of 15.7% recorded at hospitalisation among the host population. The pro- Swiss asylum seeker reception centres [42]. MRSA car- portion of minors corresponded to that of refugee and riage also proved more prevalent in the study popula- migrant minors arriving in Finland (24.9%) [35]. tion than among regular travellers hospitalised in the (sub)tropics: Khawaja et al. reported 6.6% (25/377) for Level of colonisation by any multidrug- 2010–2013 [39], Kaspar et al. 2% for 2012–2013 [43], resistant bacteria Nemeth et al. 1.2% for 2009–2011 [44], and Birgand Nearly half of the patients were colonised by MDR bac- et al. 3.8% for 2012–2013 [45]. This relevant and valu- teria, a finding consistent with recent research focusing able finding should be taken heed of when drawing up on refugees admitted to hospitals [36-38]. In Germany, infection control guidelines for hospital admission of Tenenbaum et al. reported an MDR bacterial carriage refugees and asylum seekers. rate of 33.8% (110/325) in a paediatric population in 2015 to 2016 [36], and Reinheimer et al. rates of 52.1% An analysis of spa-types revealed that the MRSA strains (61/117) in 2015 to 2016 and 66.4% (95/143) in 2015 differ from those most commonly seen in clinical sam- in two studies with both adult and paediatric patients ples in Finland (t008, t172, t067) [46]. Clear differences [37,38]. The various countries of birth may account for were found between those with different countries of the differences: in the previous studies, most of the origin, suggesting that the strains had not originated patients were Syrians and Afghans, whereas in ours, in Finland but that refugees had been MRSA carriers nearly half were Iraqis. The high overall rate of MDR on arrival. The median time from arrival in Finland to bacterial colonisation among refugees and asylum first sample was almost twice as long among MRSA seekers justifies infection control measures, accord- carriers than among non-carriers. A tendency was seen ing with rates of 55.2% demonstrated in another group for MRSA rates to grow as their stay in Finland length- routinely subjected to isolation at hospitals in low- ened, but the difference was not statistically signifi- prevalence countries, namely travellers with a recent cant (Table 6). history of hospitalisation in the (sub)tropics [39].

www.eurosurveillance.org 29 Table 6 The results of univariate and multivariable risk factor analyses of meticillin-resistant Staphylococcus aureus colonisation among asylum seekers and refugees admitted to Helsinki University Hospital, Finland, January 2010 to August 2017 (n = 447)

MRSA Patients Non-carriers carriers OR (95% Cl) p value in AOR (95% CI) p value in

Risk Factor (n = 447) (n = 352) in univariate univariate in multivariable multivariable (n = 95) analysis analysis analysisa analysis n n % n % Sex Female 237 47 19.8 190 80.2 1.0 NA NA NA Male 210 48 22.9 162 77.1 1.2 (0.8–1.9) 0.435 NA NA Age group (years) 0.806 0–5 54 13 24.1 41 75.9 1.0 NA NA NA 6–15 53 11 20.8 42 79.2 0.8 (0.3–2.1)b 0.681 NA NA 16–25 121 22 18.2 99 81.8 0.7 (0.3–2.1)b 0.369 NA NA 26–35 152 36 23.7 116 76.3 1.0 (0.5–2.0)b 0.954 NA NA > 35 67 13 19.4 54 80.6 0.8 (0.3–1.8)b 0.535 NA NA Geographical regionc 0.002 Europe 26 2 7.7 24 92.3 1.0 NA 1.0 NA North Africa and Middle 268 75 27.6 197 72.4 4.7 (1.1–20.2)d 0.040 4.5 (1.0–19.7)d 0.044 East Sub-Saharan Africa 86 11 13.4 71 86.6 1.8 (0.4–8.5)d 0.482 1.7 (0.4–8.4)d 0.500 Asia 57 5 8.8 52 91.2 1.2 (0.2–6.4)d 0.870 1.2 (0.2–6.4)d 0.864 Other or unknown 10 2 20.0 8 80.0 3.0 (0.4–24.9)d 0.309 3.2 (0.4–26.9)d 0.279 Prior hospitalisation abroad No or not specified 367 72 19.6 295 80.4 1.0 NA NA NA Yes 80 23 28.8 57 71.3 1.7 (1.0–2.9) 0.072 NA NA Prior invasive procedure abroad No or not specified 398 78 19.6 320 80.4 1.0 NA 1.0 NA Yes 49 17 34.7 32 65.3 2.2 (1.2–4.1) 0.017 2.0 (1.1–3.9) 0.033 Prior ICU care abroad No or not specified 429 89 20.7 340 79.3 1.0 NA NA NA Yes 18 6 33.3 12 66.7 1.9 (0.7–5.2) 0.208 NA NA Time from arrival to first 59 97 51 1.05 (0.99–1.10)e 0.079 NA NA sample (days, median)e Charlson comorbidity index (points) 0.626 0 417 93 22.3 324 77.7 1.0 NA NA NA 1 6 1 16.7 5 83.3 0.7 (0.1–6.0)f 0.743 NA NA 2–3 20 1 5.0 19 95.0 0.1 (0.02–1.4)f 0.100 NA NA 4–5 2 0 0 2 100 NA NA NA NA > 5 2 0 0 2 100 NA NA NA NA

AOR: adjusted odds ratio; CI: confidence interval; ICU: intensive care unit; MRSA: meticillin-resistant Staphylococcus aureus; NA: not applicable; OR: odds ratio; SD: standard deviation. aWhen selecting the variables for the multivariable model with binary logistic regression, the Akaike information criteria (AIC) was used. Time between arrival and first sample was not known for all cases; missing values were taken into account by multiple imputations, assuming that data were missing at random. bCompared with the youngest age group. cPatients’ countries of origin in the various geographical regions: Europe: Albania, Georgia, Greece, Kosovo*, Russia, Serbia, Ukraine; North Africa and Middle East: Iraq, Iran, Jordan, Kuwait, Lebanon, Morocco, Syria, Turkey, Yemen; Sub-Saharan Africa: Burundi, Cameroon, Congo, Ethiopia, Eritrea, Gambia, Ghana, Kenya, Nigeria, Rwanda, Somalia, South Africa, Sudan, Uganda, Zambia; Asia: Afghanistan, Azerbaijan, Kazakhstan, China, Kyrgyzstan, Myanmar, Nepal, Pakistan, Sri Lanka. dCompared with Europe. en = 283, only in 283/447 cases arrival date to Finland was known. Analysed as a continuous variable, OR and adjusted OR given per 30 days. fCompared with those with 0 points. *This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the International Court of Justice Opinion on the Kosovo Declaration of Independence.

30 www.eurosurveillance.org Colonisation by multiresistant bacteria analyses have not been included in refugee/migrant The rising number of CPEs has aroused great concern studies [36-38,41,48,50]. in Europe [47]. CPE findings have been reported for 0–2.1% of refugee patients [37,41,48]. Likewise, our As independent risk factors of MRSA colonisation, we study found the CPE rates to be low. Of course, even recognised geographical region and prior invasive pro- at these levels, they exceed the background rate in cedure outside the Nordic countries. The latter accords Finland [49]; special attention is warranted, since they with previous studies showing prior healthcare con- may rise further, running parallel to increasing AMR tact to be a risk factor of MRSA colonisation at hospi- prevalence in the various countries of origin. tal admission [56,57]. In our research, the highest risk of MRSA colonisation was seen among patients from As for MRAB and MRPA colonisation, two cases of each North Africa and Middle East, which is in line with an were identified; none with VRE were found, a valuable investigation by Stenhem et al. analysing imported piece of information for professionals planning infec- MRSA cases in Sweden from 2000 to 2003 [58]. Some tion control measures in hospitals. of the other risk factors of MRSA colonisation such as prior infections or antibiotic treatments, occupation ESBL-PE were the most commonly recorded MDR bac- (e.g. healthcare worker) or contact with livestock could teria. Detected among one third of the refugees and not be covered, due to the retrospective nature of our migrants, the rates remained lower than in two studies study [57]. by Reinheimer et al. [37,38] in Germany, reporting 52.1% (61/117) and 58.7% (84/143) prevalence, yet accorded For ESBL-PE colonisation, the final multivariable analy- with rates identified in samples from unaccompanied ses revealed as independent risk factors geographical refugee minors in Frankfurt in 2015 (35.3%, 42/119) region, young age (< 6 years old), short time from arrival [50]. The colonisation rates among refugees resembled to first sample, and prior hospitalisation outside the those reported for regular travellers (20–70%) [51,52], Nordic countries. Geographical region has been identi- the figures probably reflecting country-related back- fied in virtually all traveller studies analysing ESBL-PE ground colonisation rates. Traveller studies show sig- risk factors [5,12-14,16,52]. The highest risk has been nificant differences depending on destination [51,52]; linked with Asia and the Middle East [5-7,9-18], a find- the highest numbers are seen among visitors to major ing that agrees with our results. Likewise, prior hos- risk regions such as the Indian subcontinent. pitalisation has been established as a risk factor for MDR bacterial carriage among travellers [59]. One It appears, however, that the initial ESBL-PE rates of previous result contradictory to ours showed a corre- asylum seekers and refugees on arrival in Finland may lation between young age and lower rates of ESBL-PE actually have been higher than recorded here: those colonisation [13]. Indeed, refugee children and travel- with longer time since arrival had a lower ESBL-PE car- ling minors cannot be regarded as comparable: tourist riage frequency than those sampled soon after immi- children’s exposure to food/drink contaminated with gration. Indeed, recent follow-up studies show that intestinal microbes is of shorter duration, and, moreo- travellers’ ESBL-PE carriage tends to be transient and ver, trying to avoid diarrhoea, their parents probably detectable only for a few months after return [15,53]. select less risky food for them. While the acquisition rates reported in traveller studies are based on samples collected soon after travel- Traveller studies have also identified a number of lers’ return, in the present study the median time from other risk factors, such as occurrence of diarrhoea arrival to sampling was approximately 8 weeks. Further [5,6,9,10,13-15] and use of antibiotics [6,13,15-17]. research is needed into carriage duration among refu- Unfortunately, these predisposing factors could not be gees and migrants or other people with a recent history included in our analyses since such data could not be of long-term exposure to MDR bacteria. consistently drawn from our patient records.

A substantial proportion of our ESBL-PE isolates Limitations proved co-resistant to levofloxacin, tobramycin or co- Due to the retrospective design of our study, the data trimoxazole, which accords with the results of stud- were limited to those available in patient records. Some ies exploring colonisation among travellers [4,9,11,12]. relevant factors such as antibiotic use could not be Co-resistance to levofloxacin correlated with resistance analysed. Information concerning the itineraries of the to tobramycin but not co-trimoxazole. This finding may refugees and asylum seekers was lacking, and dates of be related to a genetic linkage between the resistance arrival in Finland were recorded for only 63.3% of the mechanisms of the first two [54,55]. patients. The research was conducted in a tertiary hospital which is reflected in the selection of the patients. Risk factor analysis Females were over-represented in our study population To identify potential risk factors of MRSA and ESBL-PE (53.0%), probably because of the large proportion of colonisation, we conducted univariate and multivari- pregnancy-related hospital visits. After the beginning able analysis of the items derived from the patient of 2015 only 21.2% of all asylum seekers arriving in records. To our knowledge, until now, risk factor Finland were female [60].

www.eurosurveillance.org 31 5. Tängdén T, Cars O, Melhus A, Löwdin E. Foreign travel is Conclusions a major risk factor for colonization with Escherichia coli Our study shows considerable carriage rates for MDR producing CTX-M-type extended-spectrum beta-lactamases: a prospective study with Swedish volunteers. Antimicrob Agents bacteria among refugees and asylum seekers admitted Chemother. 2010;54(9):3564-8. https://doi.org/10.1128/ to a tertiary hospital in Finland. The data suggest that AAC.00220-10 PMID: 20547788 6. Kennedy K, Collignon P. Colonisation with Escherichia coli these patients should be considered a risk group that resistant to “critically important” antibiotics: a high risk requires both screening of MDR bacteria and infection for international travellers. Eur J Clin Microbiol Infect Dis. 2010;29(12):1501-6. https://doi.org/10.1007/s10096-010- control measures at entry to hospitals in low-preva- 1031-y PMID: 20835879 lence countries. In particular the refugee and migrant 7. Peirano G, Laupland KB, Gregson DB, Pitout JDD. Colonization population’s considerable MRSA colonisation rate war- of returning travelers with CTX-M-producing Escherichia coli. J Travel Med. 2011;18(5):299-303. https://doi.org/10.1111/ rants attention in healthcare settings. j.1708-8305.2011.00548.x PMID: 21896092 8. Weisenberg SA, Mediavilla JR, Chen L, Alexander EL, Rhee Note KY, Kreiswirth BN, et al. Extended spectrum beta-lactamase- producing Enterobacteriaceae in international travelers *This designation is without prejudice to positions and non-travelers in New York City. Woodrow CJ, editor. on status, and is in line with United Nations Security 2012;7(9):e45141. 9. Ostholm-Balkhed A, Tärnberg M, Nilsson M, Nilsson LE, Council Resolution 1244/99 and the International Hanberger H, Hällgren A. Travel-associated faecal colonization Court of Justice Opinion on the Kosovo Declaration of with ESBL-producing Enterobacteriaceae: incidence and risk Independence. factors. 2013 May 14;68(9):2144-53. 10. Lausch KR, Fuursted K, Larsen CS, Storgaard M. Colonisation with multi-resistant Enterobacteriaceae in hospitalised Danish patients with a history of recent travel: a cross-sectional Acknowledgements study. Travel Med Infect Dis. 2013;11(5):320-3. https://doi. org/10.1016/j.tmaid.2013.06.004 PMID: 23810306 Authors thank Jukka Ollgren (National Institute for Health 11. Paltansing S, Vlot JA, Kraakman MEM, Mesman R, Bruijning and Welfare, Helsinki, Finland) for expert advice in statis- ML, Bernards AT, et al. Extended-spectrum β-lactamase- producing enterobacteriaceae among travelers from the tical analyses, Sointu Mero for valuable comments on the Netherlands. Emerg Infect Dis. 2013;19(8):1206-13. https:// description of microbiological methods, and Topi Kairenius doi.org/10.3201/eid1908.130257 PMID: 23885972 for graphic design of Figure 2. This work was supported 12. Kuenzli E, Jaeger VK, Frei R, Neumayr A, DeCrom S, Haller S, et by the Finnish Governmental Subsidy for Health Science al. High colonization rates of extended-spectrum β-lactamase Research, the Sigrid Juselius Foundation, the Finnish Cultural (ESBL)-producing Escherichia coli in Swiss travellers to South Foundation, and the Scandinavian Society for Antimicrobial Asia- a prospective observational multicentre cohort study looking at epidemiology, microbiology and risk factors. BMC Chemotherapy Foundation. The funding sources had no in- Infect Dis. 2014;14(1):528. https://doi.org/10.1186/1471-2334- volvement in study design, collection, analysis and interpre- 14-528 PMID: 25270732 tation of data, devising manuscript, and decision to submit 13. Kantele A, Lääveri T, Mero S, Vilkman K, Pakkanen SH, the article for publication. Ollgren J, et al. Antimicrobials increase travelers’ risk of colonization by extended-spectrum betalactamase-producing Enterobacteriaceae. Clin Infect Dis. 2015;60(6):837-46. https://doi.org/10.1093/cid/ciu957 PMID: 25613287 Conflict of interest 14. Lübbert C, Straube L, Stein C, Makarewicz O, Schubert S, Mössner J, et al. Colonization with extended-spectrum TA declares no conflicts of interest; AK has received hon- beta-lactamase-producing and carbapenemase-producing Enterobacteriaceae in international travelers returning to orary for lectures (Pfizer, MSD, Valneva, Immuron) and an Germany. Int J Med Microbiol. 2015;305(1):148-56. https://doi. investigator-initiated grant (Pfizer), and has on two occa- org/10.1016/j.ijmm.2014.12.001 PMID: 25547265 sions consulted an advisory board (Valneva), none of these 15. Ruppé E, Armand-Lefèvre L, Estellat C, Consigny P-H, El Mniai relevant to the current manuscript. A, Boussadia Y, et al. High Rate of Acquisition but Short Duration of Carriage of Multidrug-Resistant Enterobacteriaceae After Travel to the Tropics. Clin Infect Dis. 2015;61(4):593-600. https://doi.org/10.1093/cid/civ333 PMID: 25904368 Authors’ contributions 16. Angelin M, Forsell J, Granlund M, Evengård B, Palmgren H, Johansson A. Risk factors for colonization with extended- Study concept and design AK; acquisition of data TA; analysis spectrum beta-lactamase producing Enterobacteriaceae and interpretation of results TA, AK; drafting of manuscript in healthcare students on clinical assignment abroad: A prospective study. Travel Med Infect Dis. 2015;13(3):223-9. TA, AK; statistical analysis TA; final approval of version pub- https://doi.org/10.1016/j.tmaid.2015.04.007 PMID: 25982453 lished TA, AK. 17. Reuland EA, Al Naiemi N, Kaiser AM, Heck M, Kluytmans JAJW, Savelkoul PHM, et al. Prevalence and risk factors for carriage of ESBL-producing Enterobacteriaceae in Amsterdam. J Antimicrob Chemother. 2016;71(4):1076-82. https://doi. References org/10.1093/jac/dkv441 PMID: 26755493 1. World Health Organization (WHO). Antimicrobial resistance: 18. Barreto Miranda I, Ignatius R, Pfüller R, Friedrich-Jänicke B, global report on surveillance. Geneva: WHO; 2014. Available Steiner F, Paland M, et al. High carriage rate of ESBL-producing from: http://www.who.int/drugresistance/documents/ Enterobacteriaceae at presentation and follow-up among surveillancereport/en/ travellers with gastrointestinal complaints returning from India and Southeast Asia. J Travel Med. 2016;23(2):tav024. https:// 2. World Health Organization (WHO). At UN, global leaders doi.org/10.1093/jtm/tav024 PMID: 26858272 commit to act on antimicrobial resistance. Geneva: WHO; 2016. Available from: http://www.who.int/mediacentre/news/ 19. Eurostat. Record number of over 1.2 million first time releases/2016/commitment-antimicrobial-resistance/en/ asylum seekers registered in 2015. Luxembourg: Eurostat; 2016. Available from: http://ec.europa.eu/eurostat/ 3. Schwaber MJ, Navon-Venezia S, Kaye KS, Ben-Ami R, documents/2995521/7203832/3-04032016-AP-EN.pdf Schwartz D, Carmeli Y. Clinical and economic impact of bacteremia with extended- spectrum-beta-lactamase- 20. The UN Refugee Agency (UNHCR). Global Trends: Forced producing Enterobacteriaceae. Antimicrob Agents Chemother. Displacement in 2015. Geneva: UNHCR; 2016. Available from: 2006;50(4):1257-62. https://doi.org/10.1128/AAC.50.4.1257- http://www.unhcr.org/statistics/unhcrstats/576408cd7/unhcr- 1262.2006 PMID: 16569837 global-trends-2015.html 4. Tham J, Odenholt I, Walder M, Brolund A, Ahl J, Melander E. 21. Finnish Immigration Service. The Finnish Immigration Service’s Extended-spectrum beta-lactamase-producing Escherichia statistical summary 2015. Helsinki: Finnish Immigration coli in patients with travellers’ diarrhoea. Scand J Infect Dis. Service; 2016. Available from: http://www.migri.fi/ 2010;42(4):275-80. download/65019_Tilastograafit_2015_EN_final.pdf

32 www.eurosurveillance.org 22. MacPherson DW, Gushulak BD, Baine WB, Bala S, Gubbins Clin Microbiol Infect. 2017;23(9):673.e1-8. https://doi. PO, Holtom P, et al. Population mobility, globalization, org/10.1016/j.cmi.2017.02.003 PMID: 28196696 and antimicrobial drug resistance. Emerg Infect Dis. 40. Epelboin L, Robert J, Tsyrina-Kouyoumdjian E, Laouira S, 2009;15(11):1727-32. PMID: 19891858 Meyssonnier V, Caumes EMDR-GNB Travel Working Group. High 23. van der Bij AK, Pitout JDD. The role of international travel in Rate of Multidrug-Resistant Gram-Negative Bacilli Carriage and the worldwide spread of multiresistant Enterobacteriaceae. Infection in Hospitalized Returning Travelers: A Cross-Sectional J Antimicrob Chemother. 2012;67(9):2090-100. https://doi. Cohort Study. J Travel Med. 2015;22(5):292-9. https://doi. org/10.1093/jac/dks214 PMID: 22678728 org/10.1111/jtm.12211 PMID: 25997830 24. de Smalen AW, Ghorab H, Abd El Ghany M, Hill-Cawthorne GA. 41. Heudorf U, Albert-Braun S, Hunfeld K-P, Birne F-U, Schulze J, Refugees and antimicrobial resistance: A systematic review. Strobel K, et al. Multidrug-resistant organisms in refugees: Travel Med Infect Dis. 2017;15:23-8. https://doi.org/10.1016/j. prevalences and impact on infection control in hospitals. GMS tmaid.2016.12.001 PMID: 27919742 Hyg Infect Control. 2016;11:Doc16. PMID: 27579250 25. Maltezou HC. Antibiotic resistance and the refugee crisis in 42. Piso RJ, Käch R, Pop R, Zillig D, Schibli U, Bassetti S, et Europe - Preemptive action is indicated. Travel Med Infect Dis. al. A Cross-Sectional Study of Colonization Rates with 2016;14(2):69-70. https://doi.org/10.1016/j.tmaid.2016.03.009 Methicillin-Resistant Staphylococcus aureus (MRSA) PMID: 27044610 and Extended-Spectrum Beta-Lactamase (ESBL) and 26. Nellums LB, Thompson H, Holmes A, Castro-Sánchez E, Carbapenemase-Producing Enterobacteriaceae in Four Swiss Otter JA, Norredam M, et al. Antimicrobial resistance among Refugee Centres. Huang Y-C, editor. PLOS One. 2017 Jan migrants in Europe: a systematic review and meta-analysis. 13;12(1):e0170251-12. Lancet Infect Dis. 2018;18(7):796-811. https://doi.org/10.1016/ 43. Kaspar T, Schweiger A, Droz S, Marschall J. Colonization with S1473-3099(18)30219-6 PMID: 29779917 resistant microorganisms in patients transferred from abroad: 27. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method who needs to be screened? Antimicrob Resist Infect Control. of classifying prognostic comorbidity in longitudinal studies: 2015;4(31):31. https://doi.org/10.1186/s13756-015-0071-6 development and validation. J Chronic Dis. 1987;40(5):373-83. PMID: 26213620 https://doi.org/10.1016/0021-9681(87)90171-8 PMID: 3558716 44. Nemeth J, Ledergerber B, Preiswerk B, Nobile A, Karrer S, Ruef 28. World Health Organization (WHO). International statistical C, et al. Multidrug-resistant bacteria in travellers hospitalized classification of diseases and related health problems. abroad: prevalence, characteristics, and influence on clinical 10th Revision. Volume 2. Instruction manual. 2010 ed. outcome. J Hosp Infect. 2012;82(4):254-9. https://doi. Geneva: WHO; 2011. Available from: http://www.who.int/ org/10.1016/j.jhin.2012.08.017 PMID: 23103249 classifications/icd/ICD10Volume2_en_2010.pdf 45. Birgand G, Armand-Lefevre L, Lepainteur M, Lolom I, Neulier 29. United Nations Statistics Division (UNSD). Standard country C, Reibel F, et al. Introduction of highly resistant bacteria into or area codes for statistical use (M49). New York: UNSD. a hospital via patients repatriated or recently hospitalized in [Accessed 6 Apr 2018]. Available from: https://unstats.un.org/ a foreign country. Clin Microbiol Infect. 2014;20(11):O887-90. unsd/methodology/m49/ https://doi.org/10.1111/1469-0691.12604 PMID: 25069719 30. Pasanen T, Korkeila M, Mero S, Tarkka E, Piiparinen H, 46. Jaakola S, Lyytikäinen O, Rimhanen-Finne R, Salmenlinna Vuopio-Varkila J, et al. A selective broth enrichment combined S, Savolainen-Kopra C, Liitsola K, et al. Infectious Diseases with real-time nuc-mecA-PCR in the exclusion of MRSA. in Finland 2016. Helsinki: National Institute for Health APMIS. 2010;118(1):74-80. https://doi.org/10.1111/j.1600- and Welfare; 2017. Available from: http://www.julkari.fi/ 0463.2009.02562.x PMID: 20041874 handle/10024/135619 31. Suppola JP, Kolho E, Salmenlinna S, Tarkka E, Vuopio-Varkila 47. Albiger B, Glasner C, Struelens MJ, Grundmann H, J, Vaara M. vanA and vanB incorporate into an endemic Monnet DLEuropean Survey of Carbapenemase-Producing ampicillin-resistant vancomycin-sensitive Enterococcus Enterobacteriaceae (EuSCAPE) working group. Carbapenemase- faecium strain: effect on interpretation of clonality. J Clin producing Enterobacteriaceae in Europe: assessment by Microbiol. 1999;37(12):3934-9. PMID: 10565910 national experts from 38 countries, May 2015. Euro Surveill. 2015;20(45):30062. https://doi.org/10.2807/1560-7917. 32. The European Committee on Antimicrobial Susceptibility ES.2015.20.45.30062 PMID: 26675038 Testing (EUCAST). Breakpoint tables for interpretation of MICs and zone diameters. Version 7.1. Växjö: EUCAST; 2017. 48. Ravensbergen SJ, Lokate M, Cornish D, Kloeze E, Ott A, Available from: http://www.eucast.org/ast_of_bacteria/ Friedrich AW, et al. High Prevalence of Infectious Diseases and previous_versions_of_documents/ Drug-Resistant Microorganisms in Asylum Seekers Admitted to Hospital; No Carbapenemase Producing Enterobacteriaceae 33. Pasanen T, Koskela S, Mero S, Tarkka E, Tissari P, Vaara M, until September 2015. Cleary PR, editor. PLOS One. 2016 May et al. Rapid Molecular Characterization of Acinetobacter 4;11(5):e0154791-9. baumannii Clones with rep-PCR and Evaluation of Carbapenemase Genes by New Multiplex PCR in Hospital 49. European Centre for Disease Prevention and Control (ECDC). District of Helsinki and Uusimaa. Mokrousov I, editor. 2014 Jan Surveillance of antimicrobial resistance in Europe 2016. 21;9(1):e85854-8. Available from: http://dx.plos.org/10.1371/ Annual Report of the European Antimicrobial Resistance journal.pone.0085854 Surveillance Network (EARS-Net). Stockholm: ECDC; 2017. Available from: https://ecdc.europa.eu/en/publications-data/ 34. Finnish Immigration Service. Statistics - International antimicrobial-resistance-surveillance-europe-2016 protection. Helsinki: Finnish Immigration Service. [Accessed 22 Nov 2017]. Available from: http://statistics.migri. 50. Heudorf U, Krackhardt B, Karathana M, Kleinkauf N, Zinn fi/#decisions/23330?start=540 C. Multidrug-resistant bacteria in unaccompanied refugee minors arriving in Frankfurt am Main, Germany, October to 35. Finnish Immigration Service. Asylum applications 1/2015- November 2015. Euro Surveill. 2016;21(2):30109. https://doi. 12/2015. Helsinki: Finnish Immigration Service; 2017. Available org/10.2807/1560-7917.ES.2016.21.2.30109 PMID: 26838714 from: http://statistics.migri.fi/#applications/23330/49?start= 540&end=551 51. Hassing RJ, Alsma J, Arcilla MS, van Genderen PJ, Stricker BH, Verbon A. International travel and acquisition of multidrug- 36. Tenenbaum T, Becker K-P, Lange B, Martin A, Schäfer P, resistant Enterobacteriaceae: a systematic review. Euro Weichert S, et al. Prevalence of Multidrug-Resistant Organisms Surveill. 2015;20(47). PMID: 26625301 in Hospitalized Pediatric Refugees in an University Children’s Hospital in Germany 2015-2016. Infect Control Hosp Epidemiol. 52. Woerther P-L, Andremont A, Kantele A. Travel-acquired 2016;37(11):1310-4. https://doi.org/10.1017/ice.2016.179 ESBL-producing Enterobacteriaceae: impact of colonization PMID: 27523304 at individual and community level. J Travel Med. 2017 Apr 1;24(suppl_1):S29-S34. 37. Reinheimer C, Kempf VAJ, Göttig S, Hogardt M, Wichelhaus TA, O’Rourke F, et al. Multidrug-resistant organisms detected in 53. Arcilla MS, van Hattem JM, Haverkate MR, Bootsma MCJ, refugee patients admitted to a University Hospital, Germany van Genderen PJJ, Goorhuis A, et al. Import and spread of June‒December 2015. Euro Surveill. 2016;21(2):30110-5. extended-spectrum β-lactamase-producing Enterobacteriaceae https://doi.org/10.2807/1560-7917.ES.2016.21.2.30110 PMID: by international travellers (COMBAT study): a prospective, 26794850 multicentre cohort study. Lancet Infect Dis. 2017;17(1):78- 85. https://doi.org/10.1016/S1473-3099(16)30319-X PMID: 38. Reinheimer C, Kempf VAJ, Jozsa K, Wichelhaus TA, Hogardt M, 27751772 O’Rourke F, et al. Prevalence of multidrug-resistant organisms in refugee patients, medical tourists and domestic patients 54. Cantón R, González-Alba JM, Galán JC. CTX-M Enzymes: Origin admitted to a German university hospital. BMC Infect Dis. and Diffusion. Front Microbiol. 2012;3:110. https://doi. 2017;17(1):17. https://doi.org/10.1186/s12879-016-2105-y org/10.3389/fmicb.2012.00110 PMID: 22485109 PMID: 28056820 55. Ranjan A, Shaik S, Hussain A, Nandanwar N, Semmler T, 39. Khawaja T, Kirveskari J, Johansson S, Väisänen J, Djupsjöbacka Jadhav S, et al. Genomic and Functional Portrait of a Highly A, Nevalainen A, et al. Patients hospitalized abroad as Virulent, CTX-M-15-Producing H30-Rx Subclone of Escherichia importers of multiresistant bacteria-a cross-sectional study. coli Sequence Type 131. Antimicrob Agents Chemother. 2015;59(10):6087-95. https://doi.org/10.1128/AAC.01447-15 PMID: 26195517 www.eurosurveillance.org 33 56. McKinnell JA, Miller LG, Eells SJ, Cui E, Huang SS. A systematic literature review and meta-analysis of factors associated with methicillin-resistant Staphylococcus aureus colonization at time of hospital or intensive care unit admission. Infect Control Hosp Epidemiol. 2013;34(10):1077-86. https://doi. org/10.1086/673157 PMID: 24018925 57. Zhou YP, Wilder-Smith A, Hsu LY. The role of international travel in the spread of methicillin-resistant Staphylococcus aureus. J Travel Med. 2014;21(4):272-81. https://doi.org/10.1111/ jtm.12133 PMID: 24894491 58. Stenhem M, Örtqvist A, Ringberg H, Larsson L, Olsson Liljequist B, Haeggman S, et al. Imported methicillin- resistant Staphylococcus aureus, Sweden. Emerg Infect Dis. 2010;16(2):189-96. https://doi.org/10.3201/eid1602.081655 PMID: 20113546 59. Siegel JD, Rhinehart E, Jackson M, Chiarello LHealthcare Infection Control Practices Advisory Committee. Management of multidrug-resistant organisms in health care settings, 2006. Am J Infect Control. 2007;35(10) Suppl 2;S165-93. https://doi. org/10.1016/j.ajic.2007.10.006 PMID: 18068814 60. Eurostat. European statistics. Luxembourg: Eurostat. [Accessed 4 Jan 2017]. Available from: http://ec.europa.eu/eurostat/data/ database

This article is copyright of the authors or their affiliated institutions, 2018.

34 www.eurosurveillance.org Research article High prevalence of carriage of mcr-1-positive enteric bacteria among healthy children from rural communities in the Chaco region, Bolivia, September to October 2016

Tommaso Giani1,2, Samanta Sennati¹, Alberto Antonelli², Vincenzo Di Pilato², Tiziana di Maggio¹, Antonia Mantella², Claudia Niccolai², Michele Spinicci², Joaquín Monasterio³, Paul Castellanos⁴, Mirtha Martinez⁵, Fausto Contreras⁵, Dorian Balderrama Villaroel⁵, Esther Damiani⁶, Sdenka Maury⁷, Rodolfo Rocabado⁸, Lucia Pallecchi¹, Alessandro Bartoloni2,9, Gian Maria Rossolini2,10 1. Department of Medical Biotechnologies, University of Siena, Siena, Italy 2. Department of Experimental and Clinical Medicine, University of Florence, Florence, Italy 3. Servicio Departamental de Salud (SEDES) de Santa Cruz, Santa Cruz, Bolivia 4. Servicio Departamental de Salud (SEDES) de Tarija, Tarija, Bolivia 5. Servicio Nacional de Sanidad Agropecuaria e Inocuidad Alimentaria (SENASAG), Ministerio de Desarrollo Rural y Tierras, Santa Cruz, Bolivia 6. Instituto Nacional de Laboratorios de Salud (INLASA), Ministerio de Salud, La Paz, Bolivia 7. Unidad Epidemiología, Ministerio de Salud, La Paz, Bolivia 8. Servicios Generales de Salud, Ministerio de Salud, La Paz, Bolivia 9. Infectious and Tropical Diseases Unit, Careggi University Hospital, Florence, Italy 10. Clinical Microbiology and Virology Unit, Careggi University Hospital, Florence, Italy Correspondence: Gian Maria Rossolini ([email protected])

Citation style for this article: Giani Tommaso, Sennati Samanta, Antonelli Alberto, Di Pilato Vincenzo, di Maggio Tiziana, Mantella Antonia, Niccolai Claudia, Spinicci Michele, Monasterio Joaquín, Castellanos Paul, Martinez Mirtha, Contreras Fausto, Balderrama Villaroel Dorian , Damiani Esther, Maury Sdenka, Rocabado Rodolfo, Pallecchi Lucia, Bartoloni Alessandro, Rossolini Gian Maria. High prevalence of carriage of mcr-1-positive enteric bacteria among healthy children from rural communities in the Chaco region, Bolivia, September to October 2016. Euro Surveill. 2018;23(45):pii=1800115. https://doi.org/10.2807/1560-7917.ES.2018.23.45.1800115

Article submitted on 12 Mar 2018 / accepted on 10 Jul 2018 / published on 08 Nov 2018

Background: The mcr-1 gene is a transferable resist- lineages and mcr-genetic supports. Conclusion: This ance determinant against colistin, a last-resort anti- high prevalence of mcr-1-like carriage, in absence of microbial for infections caused by multi-resistant professional exposure, is unexpected. Its extent at Gram-negatives. Aim: To study carriage of antibiotic- the national level should be investigated with priority. resistant bacteria in healthy school children as part Possible causes should be studied; they may include of a helminth control and antimicrobial resistance unrestricted use of colistin in veterinary medicine and survey in the Bolivian Chaco region. Methods: From animal breeding, and importation of mcr-1-positive September to October 2016 we collected faecal sam- bacteria via food and animals. ples from healthy children in eight rural villages. Samples were screened for mcr-1- and mcr-2 genes. Background Antimicrobial susceptibility testing was performed, The mcr-1 gene is a transferable colistin resistance and a subset of 18 isolates representative of individuals determinant that was first described among entero- from different villages was analysed by whole genome bacterial strains isolated from animals and humans sequencing (WGS). Results: We included 337 children in China. The gene encodes a phosphoethanolamine (mean age: 9.2 years, range: 7–11; 53% females). The transferase that modifies the colistin target by addition proportion of mcr-1 carriers was high (38.3%) and of phosphoethanolamine to the 1’ or 4’ phosphate group present in all villages; only four children had previ- of lipid A, which reduces its affinity to colistin [1,2]. ous antibiotic exposure. One or more mcr-1-positive Discovery of mcr-1 was considered highly alarming, isolates were recovered from 129 positive samples, given the role that colistin has recently regained as yielding a total of 173 isolates (171 Escherichia coli, a last-resort antibiotic for treatment of infections 1 Citrobacter europaeus, 1 Enterobacter hormaechei). caused by multi-resistant Gram-negative patho- No mcr-2 was detected. Co-resistance to other antimi- gens such as carbapenem-resistant Enterobacterales crobials varied in mcr-positive E. coli. All 171 isolates and Acinetobacter baumannii [1,3]. were susceptible to carbapenems and tigecycline; 41 (24.0%) were extended-spectrum β-lactamase Subsequent to its discovery, several studies producers and most of them (37/41) carried blaCTX- have revealed a global distribution of mcr-1, with

M-type genes. WGS revealed heterogeneity of clonal an overall higher prevalence among Escherichia www.eurosurveillance.org 35 Figure 1 Geographical locations of the surveyed communities and the proportion of mcr-1-positive samples, Chaco, Boliva, September–October 2016 (n = 8 communities)

BOLIVIA

La Paz

Santa Cruz de la Sierra

PALMARITO 13/46 (28.3%; 95% CI: 15.3–41.3)

Sucre IVAMIRAPINTA Gutierrez 17/67 (25.4%; 95% CI: 15–35.8) TETA PIAU / KURUPAITY 18/42 (42.9%; 95% CI: 27.9–57.9) Lagunillas Charagua SAN ANTONIO DEL PARAPETÍ 9/47 (19.1%; 95% CI: 7.9–30.3)

Tarĳa Villa Montes CHIMEO 16/39 (41.0%; 95% CI: 25.6–56.4) Yacuiba TARAIRÍ 33/41 (80.5%; 95% CI: 68.4–92.6) CAPIRENDITA 4/16 (25.0%; 95% CI: 3.8–46.2)

PALMAR CHICO 19/39 (48.7%; 95% CI: 33– 64.4)

CI: confidence interval.

The surveyed communities are located in the Bolivian Chaco, in rural areas of five municipalities (indicated with stars). For each community, the proportion of mcr-1-positive samples versus the total number of collected samples is reported, along with 95% CI. Major Bolivian cities are indicated by circles.

coli and Salmonella enterica, and occasional occurrence been detected [6]. More recently, additional trans- in other enterobacterial species. Most mcr-1-positive ferable mcr genes (mcr-2, mcr-3, mcr-4, mcr-5, mcr- strains were of animal origin, and farm animals were 6, mcr-7 and mcr-8) have been reported, for which the identified as the principal reservoir of mcr-1 genes [1,4]. global epidemiology remains to be clarified [7-13]. Investigation of archival strains dated the presence of mcr-1 back to at least the 1980s [5]. As with other In South America, mcr-1 genes have been reported from resistance genes, minor allelic variants of mcr-1 have several countries in isolates from humans, animals

36 www.eurosurveillance.org Table 1 Features of the study population, stratified by mcr-1-positive and -negative children, Chaco, Bolivia, September–October 2016

Total mcr-1-negative mcr-1-positive Characteristics % % % P value (n/N) (n/N) (n/N) Sex Male 158/337 47 100/208 48 58/129 45 0.58 Female 179/337 53 108/208 52 71/129 55 Age (years) Mean (95% CI) 9.3 (9.1–9.4) NA 9.3 (9.1–9.5) NA 9.2 (9.0–9.5) NA 0.81 Median (IQR) 9 (8–10) NA 9 (8–10) NA 9 (8–10) NA Prior antibiotic usea 4/337 1 3/208 1 1/129 1 0.58

CI: confidence interval; IQR: interquartile range; NA: not applicable. aIn the last 15 days.

and food [14-27]. Recently, the Pan American Health economy is mostly based on subsistence farming and Organisation (PAHO) section of the World Health local animal husbandry. Organization (WHO) recommended to implement and strengthen surveillance and epidemiological investiga- In each community, children were selected among tion of plasmid-mediated transferable colistin resist- those attending primary school, starting from the ance in its Member States [14]. In Bolivia, mcr-1 has third year and possibly including the upper years, to thus far been reported in a Citrobacter braakii that achieve a number of ca 50 individuals per site when- was isolated from a ready-to-eat food sample [21], as ever possible. This sample size corresponded to that well as in a few clinical isolates of E. coli referred from recommended by WHO for cluster sampling in helminth various departments to the National Institute of Health control programmes in healthy school children [33]. Laboratories (INLASA) (data not shown). Previous use of antibiotics during the last 15 days During the last two decades we carried out several sur- was investigated by a questionnaire administered to veillance studies in the Bolivian Chaco region, docu- parents. menting a high prevalence of resistance to old and more recent antibiotics in commensal and pathogenic Laboratory analyses bacteria from humans [21,28-32]. Screening for mcr-1- and mcr-2-positive strains in faecal In 2016, a new surveillance study was carried out in samples a population of healthy school children from several One faecal sample for each child was collected during rural communities in this region to investigate the a two-month period from September to October 2016; prevalence of intestinal parasites and the carriage of the samples were transferred to the Laboratories of antibiotic-resistant bacteria. Here we report about an Camiri or Villa Montes Hospitals within 6 hours and unexpected and high rate of faecal carriage of mcr- were plated onto MacConkey agar. After incubation at 1-positive Enterobacterales in this population. 35 °C for 24 hours, the bacterial growth (representative of the total enterobacterial microbiota) was collected Methods with a sterile swab in an Amies transport medium and was shipped to Italy. Each sample was then sub- Study population and setting cultured on MacConkey agar again, and the bacterial The study population consisted of healthy school chil- growth was resuspended in Brain Heart Infusion broth dren living in eight rural communities of the Chaco plus 20% (v/v) glycerol and stored at -70 °C pending region, in south-eastern Bolivia (between longitude further analyses. 63°66 and 63°18 east and latitude 19°49 and 21°88 south, Figure 1). In these communities, the population To screen for the presence of mcr-1- and mcr-2-positive lives in houses mostly constructed of mud and sticks, strains, the preserved suspensions of total enterobac- with packed earth floors and straw or corrugated metal terial microbiota were thawed and 10 μl were inocu- roofs. There is no wired electricity and no sewage sys- lated onto McConkey supplemented with colistin (2 tem. The main water sources are small ponds, in which mg/L, MCC medium). After incubation at 35 °C for 24 animals also bathe and drink, and outdoor taps. The hours, a loopful of the bacterial growth (taken either from confluent growth or from isolated colonies of

www.eurosurveillance.org 37 Table 2 Susceptibility of mcr 1-positive Escherichia coli isolates to various antimicrobials, Chaco, Bolivia, September–October 2016 (n = 171)

AMC PTZ CAZ CTX FEP MEM ERT GEN CIP TIG COL Total n % n % n % n % n % n % n % n % n % n % n % n % 72 42.1 171 100 130 76.0 128 74.9 130 76.0 171 100 171 100 138 80.7 62 36.3 171 100 168 1.7 171 100

AMC: amoxicillin/clavulanate (clavulanate at fixed concentration of 4 mg/L); CAZ: ceftazidime; CIP: ciprofloxacin; COL: colistin; CTX: cefotaxime; ERT: ertapenem; FEP: cefepime; GEN: gentamicin; MEM: meropenem; PTZ: piperacillin/tazobactam (tazobactam at fixed concentration of 4 mg/L); TIG: tigecycline. Numbers and percentages of susceptible isolates are given.

different morphologies) was resuspended in 300μl of randomly selected individuals if co-colonisations were normal saline, and half of the bacterial suspension was not detected and the two non-E. coliisolates bore mcr- used to prepare a crude DNA extract by heating at 99 °C 1. For the latter, species identification was carried for 15 minutes. The crude extracts were then screened out by the analysis of housekeeping genes [39,40]. for the presence of mcr-1 and mcr-2 genes by real-time Bacterial genomic DNA of these 18 selected mcr-pos- (RT) PCR, as described previously [34]. In the case of a itive isolates, extracted using the phenol-chloroform positive result, the remaining bacterial suspension was method [41], was subjected to WGS with a MiSeq used to inoculate the MCC medium to obtain isolated platform (Illumina, Inc., San Diego, California, United colonies, and all isolated colonies of different morphol- States (US)) using a 2x300 paired-end approach. ogy were then tested for the presence of mcr genes by Raw reads were assembled using SPAdes 3.5 [42]. An RT-PCR. The mcr-positive isolates were identified using average of 120 contigs per strain was obtained, with MALDI-TOF mass spectrometry (Vitek MS, bioMérieux, an average N50 of 163 Kb. Draft genomes have been Marcy-l’Etoile, France). deposited in the National Center for Biotechnology Information (NCBI) WGS database under the BioProject When a sample yielded two or more mcr-1-positive PRJNA427943 (accession numbers: PQTO00000000; isolates of the same species, clonal relatedness of PQTN00000000; PQTM00000000; PQTL00000000; the isolates was investigated by random amplifica- PQTK00000000; PQTJ00000000; PQTI00000000; tion of polymorphic DNA (RAPD) profiling, as described PQTH00000000; PQTG00000000; PQTF00000000; previously [35]. The three mcr-positive isolates that PQTE00000000; PQTD00000000; PQTC00000000; were colistin susceptible were subjected to mcr gene PQTB00000000; PQTA00000000; PQSZ00000000; amplification and sequencing using previously PQSY00000000; PQSX00000000). Resistance genes described primers and conditions [34]. and plasmid content were investigated using the ResFinder and PlasmidFinder tools available at the Antimicrobial susceptibility testing Center for Genomic Epidemiology at https://cge.cbs. Antimicrobial susceptibility testing was carried out dtu.dk/services/ResFinder/. Clonal relatedness was using reference broth microdilution [36]. Minimum investigated by in silico determination of the multilo- inhibitory concentration (MIC) results were interpreted cus sequence typing (MLST) profile obtained by the according to the European Committee on Antimicrobial MLST 1.8 software (available at https://cge.cbs.dtu. Susceptibility Testing (EUCAST) clinical breakpoints dk/services/MLST/) using the assembled WGS as input [36]. data.

Analysis of extended-spectrum β-lactamases Statistical analysis All isolates showing a ceftazidime and/or cefotaxime Statistical analysis of the data was performed with MIC > 1 mg/L were screened for extended-spectrum STATA 11.0 (StataCorp, College Statio, Texas, US). β-lactamases (ESBL) production by a combination Frequencies and percentages with 95% confidence disk test using ceftazidime and cefotaxime as sub- intervals (CI) for categorical variables, medians and strates and clavulanic acid as an inhibitor [37]. ESBL- interquartile ranges (IQR) for continuous variables positive isolates by phenotypic testing were subjected were calculated. Mann–Whitney test was used to com- to RT-PCR for the detection of blaCTX-M ESBL genes, as pare median age. Chi-squared test was used to inves- described previously [38]. tigate the association of mcr-1 carriage with sex and prior antibiotic use. Results were considered signifi- Whole genome sequencing cant when the p value was ≤ 0.05. A subset of 18 mcr-1-positive isolates were subjected to whole genome sequencing (WGS) analysis. This Ethical statement subset comprised two E. coli isolates per community: Written informed consent was always obtained from randomly selected individuals co-colonised from parents or legal guardians. The investigation by two different mcr-1-positive E. coli or from two was planned and carried out within a collaboration

38 www.eurosurveillance.org Table 3 Features of mcr-1-positive isolates subjected to whole genome sequencing analysis, Chaco, Bolivia, September–October 2016 (n = 18)

Additional Isolate Subject mcr variant and mcr contig Community Species resistance Acquired resistance genesb STc code code genetic contextd size (bp) trait(s)a AMC; GEN; blaTEM-1B; aac (3)-IV; aph (4)- 12A 1 Escherichia coli 48 mcr-1-pap(IncI2) 61,600 CIP Ia; fosA3; floR; qnrB19; tet(A)

Palmarito blaTEM-1B; aac (3)-IV; aph(3’)-Ia; aph (4)-Ia; strA; 12B 1 E. coli GEN; CIP 744 mcr-1-pap(IncI2) 60,992 strB; catA1; floR; oqxA; oqxB; sul2; tet(A) 155A 2 E. coli FSe ND 10 mcr-1-pap(IncI2) 60,547

Ivamirapinta blaTEM-1A; aadA1; aadA2; strA; 155B 2 E. coli AMC; CIP strB; cmlA1; floR; qnrB19; sul2; 206 mcr-1-unkf 2,943 sul3; tet(A); dfrA8 blaTEM-1B; strA; strB; floR; 86A 3 E. coli AMC; CIP 2,705 mcr-1-unkf 2,942 qnrB19; sul2; tet(A); dfrA1 blaTEM-1B; aadA1; aadA2; strA; mcr-1.5-pap- Tetapiau/Kurupaity 86B 3 E. coli AMC; CIP strB; cmlA1; floR; sul1; sul2; 2,936 13,7897 ISApl1(IncHI1) sul3; tet(A); tet(B); dfrA1; dfrA12 Citrobacter 67A 4 AMC qnrB19; qnrB28 NA mcr-1-pap(IncI2) 60,321 europaeus AMC; CIP; blaCTX-M-55; aadA1; aadA2; GEN; CAZ; 173A 5 E. coli aadA5; rmtB; fosA3; cmlA1; floR; 1,286 mcr-1-unkf 6,134 CTX; FEP; qnrB19; sul3; tet(A); dfrA17 San Antonio del (ESBL) ParapetÍ AMC; CIP; blaCTX-M-55; blaTEM-1B; GEN; CAZ; 173B 5 E. coli aadA1; aadA2; rmtB; cmlA1; 1,286 mcr-1-unkf 2,863 CTX; FEP; floR; qnrB19; sul3; tet(A) (ESBL) aadA1; aadA2; strA; strB; 224A 6 E. coli AMC cmlA1; floR; qnrB19; sul2; sul3; 2,705 mcr-1-pap(IncI2) 59,561 tet(A); tet(B); dfrA14 TarairÍ blaTEM-1B; aadA1; aadA2; strA; ∆ISApl1- 224B 6 E. coli AMC; CIP strB; cmlA1; floR; QnrB19; sul2; 7,570 mcr-1-pap- 52,737 sul3; tet(A); tet(B); dfrA14 ISApl1(IncHI1) blaTEM-1B; aadA5; strA; strB; 306A 7 E. coli AMC 69 mcr-1-pap(IncI2) 63,921 sul1; sul2; dfrA17 blaTEM-1B; blaOXA-1; aadA1; Palmar Chico 306B 7 E. coli AMC; CIP 10 mcr-1-pap(IncI2) 64,425 sul1; tet(X) Enterobacter 301B 8 FSe ND - mcr-1-pap(IncI2) 63,943 hormaechei blaTEM-1B; aadA1; floR; sul3; 286A 9 E. coli AMC 117 mcr-1-pap(IncI2) 59,748 Capirendita tet(A); tet(C); dfrA1 295B 10 E. coli FSe ND 711 mcr-1-pap(IncI2) 56,317 blaTEM-1B; aac (3)-IV; aadA1; AMC; GEN; aadA2; aph(3’)-Ia; cmlA1; floR; 274A 11 E. coli 7,571 mcr-1-unkf 2,943 CIP qnrB19; sul2; sul3; tet(A); tet(M); dfrA12 Chimeo blaCTX-M-55; blaTEM-1B; AMC; CAZ; blaOXA-10; aac(6‘)Ib-cr; aacA4; 274B 11 E. coli CTX; FEP aadA1; strA; strB; fosA3; cmlA1; 3,056 mcr-1-pap(IncI2) 60,652 (ESBL) floR; qnrB19; qnrVC4; sul2; tet(A); dfrA14

AMC: amoxicillin/clavulanate (clavulanate at fixed concentration of 4 mg/L); CAZ: ceftazidime; CIP: ciprofloxacin; COL: colistin; CTX: cefotaxime; ESBL: extended-spectrum β-lactamase; FEP: cefepime; FS: fully susceptible; GEN: gentamicin; NA: not applicable; ND: none detected; ST: sequence type; unk: unknown. aAll isolates were resistant to colistin; additional resistance traits referred to the panel of tested drugs reported in Table 2. bAcquired resistance genes as determined by analysis with the ResFinder software. cSequence-types were assigned using the Warwick scheme (http://enterobase.warwick.ac.uk/species/index/ecoli). dIf the gene was linked with a known plasmid backbone, the plasmid replicon type is reported in brackets. eThe isolate was susceptible to all tested agents except colistin. fIn these cases it was not possible to reveal the nature of flanking regions due to the presence of repeated sequences flanking the gene. For each isolate, the epidemiological data, additional resistance profile, acquired resistance genes content, sequence type and mcr genetic context are reported.

www.eurosurveillance.org 39 agreement between the Ministry of Health of the In silico MLST analysis of the 16 E. coli isolates revealed Plurinational State of Bolivia and the University of a considerable diversity, with only a few isolates from Florence, Italy, and with the support of the Guaraní different villages belonging to the same sequence type political organisation (Asamblea del Pueblo Guaraní). (ST). All but one of the couples isolated from the same Ethical approval for the study was obtained from the individual belonged to different STs (Table 3). above-mentioned institutions (see Acknowledgements section). Analysis of the acquired resistance genes showed a remarkable diversity and a variety of patterns (Table Results 3). The number of known acquired resistance genes Faecal specimens were obtained from 337 healthy varied from 0 to 16 (median: 9). Overall, the resistance school children in eight rural communities of the gene content was consistent with the susceptibility Bolivian Chaco region (Figure 1). Children (179 females; profile. The three ESBL-positive E. coli isolates carried

53%;) were aged 7 to 11 years (mean: 9.2 years). the blaCTX-M-55 variant previously reported in Bolivia [30]. Previous antibiotic exposure was only reported for four Analysis of the mcr-1 carrying contigs revealed that children. in 13 isolates the mcr-1 gene was linked to backbone regions typical of IncI2 or IncHI1 plasmids, suggesting mcr-1 carriage a plasmid location, with some plasmid diversity. In the All 337 samples of enterobacterial microbiota yielded remaining five isolates, it was not possible to deter- some growth (from scanty to vigorous) on the MCC mine the nature of flanking regions due to the presence medium, and 129 (38.3%) yielded a positive result of repeated sequences flanking the gene (Table 3). for mcr-1. Positive samples were detected in children from each village, although at variable rates (range: Discussion 19.1–80.5%; Figure 1). No mcr-2 genes were detected. Our study revealed a very high prevalence of carriage One or more mcr-1-positive isolates were recovered from of mcr-1-positive strains among healthy children living each of the 129 samples, yielding a total of 173 posi- in rural communities of the Bolivian Chaco. Carriage tive isolates, including 171 E. coli, one Citrobacter spp. of mcr-1-positive strains in healthy humans has been and one Enterobacter spp.. Multiple mcr-1-positive iso- investigated in a limited number of studies, mostly from lates from the same sample consisted of either two or Asian countries [41-53]. The prevalence rates detected three E. coli isolates of different colonial morphology in such studies have usually been low (< 5%), except in a and RAPD profile (in 32 and 5 samples, respectively), or group of chicken farmers from Vietnam, where a 34.7% in an E. coli plus an Enterobacter spp. (in one sample). carriage rate of mcr-1-positive E. coli was detected and No differences were found in the demographic charac- attributed to professional exposure to mcr-1-positive teristics, sex or age, of children carrying mcr-1-positive animals [45]. Therefore, to our best knowledge, we pre- Enterobacterales or children without mcr-1 carriage sent the highest rate of mcr-1 carriage thus far reported (Table 1), nor were there any differences in the living in healthy humans. conditions of the communities with different proportions of carriers (data not shown). In our study, professional exposure could be excluded as a reason for the high prevalence of mcr-1 carriage, Antimicrobial susceptibility of mcr-1-positive as well as human use of colistin. Overall, only four isolates children had prior exposure to antibiotics and the use Colistin susceptibility testing showed that the major- of colistin in Bolivia is occasional and limited to infec- ity (n = 170; 98.3%) of the mcr-1-positive isolates were tions by some multi-drug resistant pathogens in large resistant to colistin (MIC range: 4–> 8 mg/L), while urban hospitals (data not shown). However, colistin is only three E. coli (from different villages) were colis- available with no restrictions for veterinary use and tin-susceptible (all with an MIC of 2 mg/L) (Table 2). in animal breeding [54], and we hypothesise that this Sequencing of mcr amplicons from the latter isolates could have played a major role in the selection of colis- showed identity with mcr-1, suggesting that the colistin-resistant strains in the animal population and the tin susceptible phenotype was not due to mutations environment. Moreover, the introduction of mcr-posi- inactivating the gene. Variable resistance rates to tive strains via imported food and/or food-producing other antimicrobial agents were observed, including animals from countries where their prevalence was fluoroquinolones, expanded-spectrum cephalospor- found to be high (e.g. Brazil) [15,22] could also repre- ins, β-lactamase plus inhibitor combinations and gen- sent a source of such strains. Poor sanitation and close tamicin. All isolates were susceptible to carbapenems contact with animals, which characterise the studied and tigecycline (Table 2). setting, may lead to a high level of environmental contamination and facilitate cross-transmission of colistin- Diversity of the mcr-1-positive isolates resistant strains and colistin resistance genes between WGS analysis of the subset of 18 mcr-1-positive iso- different environments, resulting in a high prevalence lates confirmed the identification of the two non-E. in humans who are not directly exposed to the drug. coli isolates as Citrobacter europaeus and Enterobacter hormaechei, respectively (Table 3), two species in In our case, only a minority of the mcr-positive iso- which mcr-1 was not previously reported. lates showed resistance to other antimicrobials, and

40 www.eurosurveillance.org aspects; the Guaraní political organization (Asamblea del no carbapenem resistance was detected, leaving a Pueblo Guaraní) supported the field work and conducted the number of therapeutic options in case of infection. interviews. However, the potential risk of spread of the mcr-1 gene to extensively resistant isolates through transferable plasmids mechanisms should not be underestimated. Conflict of interest

Genomic analysis of a subset of the mcr-1-positive E. None declared. coli isolates, representative of different communities and of different isolates from the same child, revealed a remarkable heterogeneity in terms of clonal lineages Authors’ contributions and genetic supports. Therefore, the observed TG and SS analysed the data and drafted the manuscript; AA, epidemiological scenario could not be ascribed to the VDP, CN did the molecular analysis and genome sequencing; TM, AM and LP produced phenotypic data and handled the expansion of a single mcr-1-positive clone, nor even samples; MS, MM, FC, JM, PC, DBV, ED, SM and RR collect- to the spread of a single plasmid. The diversity of the ed the samples and participated in the coordination of the genetic background of the mcr-1 genes underlined the survey; AB and GMR coordinated the survey and edited the ability of this gene to transfer itself among different manuscript. clones (and even different species) and plasmids. Interestingly, we detected for the second time in South America the mcr-1.5 variant, previously described in References an E. colistrain from Argentina [23]. 1. Poirel L, Jayol A, Nordmann P. Polymyxins: antibacterial activity, susceptibility testing, and resistance mechanisms encoded by plasmids or chromosomes. Clin Microbiol Rev. Our study has some limitations. First, the presence 2017;30(2):557-96. https://doi.org/10.1128/CMR.00064-16 of animal or environmental reservoirs of mcr-positive PMID: 28275006 2. Liu YY, Wang Y, Walsh TR, Yi LX, Zhang R, Spencer J, et isolates and the direct transmission between humans al. Emergence of plasmid-mediated colistin resistance and animals/environment could not be demonstrated, mechanism MCR-1 in animals and human beings in China: a microbiological and molecular biological study. Lancet since we did not collect any samples from animals or Infect Dis. 2016;16(2):161-8. https://doi.org/10.1016/S1473- the environment. Second, apart from mcr-2, we did 3099(15)00424-7 PMID: 26603172 not search for other recently described mcr-variants 3. Giamarellou H. Epidemiology of infections caused by polymyxin-resistant pathogens. Int J Antimicrob that could be responsible for resistance observed Agents. 2016;48(6):614-21. https://doi.org/10.1016/j. in other isolates. Third, the study was designed as a ijantimicag.2016.09.025 PMID: 27865627 4. Schwarz S, Johnson AP. Transferable resistance to colistin: a cross-sectional survey, in which one sample from each new but old threat. J Antimicrob Chemother. 2016;71(8):2066- individual was collected. It would be interesting to 70. https://doi.org/10.1093/jac/dkw274 PMID: 27342545 investigate the prevalence of mcr-1 carriage in adults 5. Shen Z, Wang Y, Shen Y, Shen J, Wu C. Early emergence of mcr-1 in Escherichia coli from food-producing animals. Lancet and the duration of carriage over time to understand Infect Dis. 2016;16(3):293. https://doi.org/10.1016/S1473- if and how much humans could represent a major 3099(16)00061-X PMID: 26973308 reservoir in this setting. It would also be interesting to 6. Rebelo AR, Bortolaia V, Kjeldgaard JS, Pedersen SK, Leekitcharoenphon P, Hansen IM, et al. Multiplex PCR further characterise, in more detail, the plasmid sup- for detection of plasmid-mediated colistin resistance ports of the mcr-1 and other resistant determinants. determinants, mcr-1, mcr-2, mcr-3, mcr-4 and mcr-5 for surveillance purposes. Euro Surveill. 2018;23(6):17-00672. Investigations on these aspects are underway. https://doi.org/10.2807/1560-7917.ES.2018.23.6.17-00672 PMID: 29439754 7. Xavier BB, Lammens C, Ruhal R, Kumar-Singh S, Butaye P, In conclusion, our findings prompt the need to rap- Goossens H, et al. Identification of a novel plasmid-mediated idly monitor the extent of human and animal carriage colistin-resistance gene, mcr-2, in Escherichia coli, Belgium, June 2016. Euro Surveill. 2016;21(27):30280. https://doi. rates and environmental contamination by mcr genes org/10.2807/1560-7917.ES.2016.21.27.30280 PMID: 27416987 with a one-health approach, and to introduce policies 8. Yin W, Li H, Shen Y, Liu Z, Wang S, Shen Z, et al. Novel banning the non-therapeutic use of colistin. This was plasmid-mediated colistin resistance gene mcr-3 in Escherichia coli. MBio. 2017;8(4):e01166-17. https://doi.org/10.1128/ also recently highlighted by the PAHO/WHO, which mBio.01166-17 PMID: 28811346 encouraged the implementation of animal-human sur- 9. Carattoli A, Villa L, Feudi C, Curcio L, Orsini S, Luppi A, et veillance, as well as actions to prevent and control al. Novel plasmid-mediated colistin resistance mcr-4 gene in Salmonella and Escherichia coli, Italy 2013, Spain and the spread of mcr-positive microorganisms, such as Belgium, 2015 to 2016. Euro Surveill. 2017;22(31):30589. the monitoring of colistin use in human food produc- https://doi.org/10.2807/1560-7917.ES.2017.22.31.30589 PMID: 28797329 tion [14]. In Europe, knowledge of mcrcarriage among 10. Borowiak M, Fischer J, Hammerl JA, Hendriksen RS, Szabo I, healthy individuals is still limited [47,53]. While avail- Malorny B. Identification of a novel transposon-associated phosphoethanolamine transferase gene, mcr-5, conferring able data suggest a very low occurrence, it will be inter- colistin resistance in d-tartrate fermenting Salmonella enterica esting to study human and animal carriage rates and subsp. enterica serovar Paratyphi B. J Antimicrob Chemother. 2017;72(12):3317-24. https://doi.org/10.1093/jac/dkx327 environmental contamination in different countries and PMID: 28962028 settings. 11. AbuOun M, Stubberfield EJ, Duggett NA, Kirchner M, Dormer L, Nunez-Garcia J, et al. mcr-1 and mcr-2 variant genes identified in Moraxella species isolated from pigs in Great Britain from 2014 to 2015. J Antimicrob Chemother. 2017;72(10):2745-9. Acknowledgements https://doi.org/10.1093/jac/dkx286 PMID: 29091227 12. Yang YQ, Li YX, Lei CW, Zhang AY, Wang HN. Novel plasmid- The Bolivian Ministry of Health and the Regional Health mediated colistin resistance gene mcr-7.1 in Klebsiella Departments approved the study design, including its ethical

www.eurosurveillance.org 41 pneumoniae. J Antimicrob Chemother. 2018;73(7):1791-5. antimicrobial resistance among healthy children in Bolivia. https://doi.org/10.1093/jac/dky111 PMID: 29912417 Trop Med Int Health. 1998;3(2):116-23. https://doi. 13. Wang X, Wang Y, Zhou Y, Li J, Yin W, Wang S, et al. Emergence org/10.1046/j.1365-3156.1998.00201.x PMID: 9537273 of a novel mobile colistin resistance gene, mcr-8, in NDM- 29. Bartoloni A, Pallecchi L, Riccobono E, Mantella A, Magnelli producing Klebsiella pneumoniae. Emerg Microbes Infect. D, Di Maggio T, et al. Relentless increase of resistance to 2018;7(1):122. https://doi.org/10.1038/s41426-018-0124-z fluoroquinolones and expanded-spectrum cephalosporins PMID: 29970891 in Escherichia coli: 20 years of surveillance in resource- 14. Pan American Health Organization (PAHO)/World Health limited settings from Latin America. Clin Microbiol Infect. Organization (WHO). Epidemiological Alert: Enterobacteriaceae 2013;19(4):356-61. https://doi.org/10.1111/j.1469- with plasmid-mediated transferable colistin resistance, public 0691.2012.03807.x PMID: 22414066 health implications in the Americas, 10 June 2016. Washington, 30. Bartoloni A, Sennati S, Di Maggio T, Mantella A, Riccobono D.C./Geneva: PAHO/WHO; 2016. Available from: http://www. E, Strohmeyer M, et al. Antimicrobial susceptibility and paho.org/hq/index.php?option=com_docman&task=doc_view emerging resistance determinants (blaCTX-M, rmtB, fosA3) in &Itemid=270&gid=35007&lang=en clinical isolates from urinary tract infections in the Bolivian 15. Fernandes MR, Moura Q, Sartori L, Silva KC, Cunha MP, Chaco. Int J Infect Dis. 2016;43:1-6. https://doi.org/10.1016/j. Esposito F, et al. Silent dissemination of colistin-resistant ijid.2015.12.008 PMID: 26686940 Escherichia coli in South America could contribute to the global 31. Bartoloni A, Riccobono E, Magnelli D, Villagran AL, Di Maggio spread of the mcr-1 gene. Euro Surveill. 2016;21(17):30214. T, Mantella A, et al. Methicillin-resistant Staphylococcus https://doi.org/10.2807/1560-7917.ES.2016.21.17.30214 PMID: aureus in hospitalized patients from the Bolivian Chaco. 27168587 Int J Infect Dis. 2015;30:156-60. https://doi.org/10.1016/j. 16. Rapoport M, Faccone D, Pasteran F, Ceriana P, Albornoz E, ijid.2014.12.006 PMID: 25486009 Petroni A, et al. MCR Group. First description of mcr-1-mediated 32. Pallecchi L, Malossi M, Mantella A, Gotuzzo E, Trigoso colistin resistance in human infections caused by Escherichia C, Bartoloni A, et al. Detection of CTX-M-type beta- coli in Latin America. Antimicrob Agents Chemother. lactamase genes in fecal Escherichia coli isolates from 2016;60(7):4412-3. https://doi.org/10.1128/AAC.00573-16 healthy children in Bolivia and Peru. Antimicrob Agents PMID: 27090181 Chemother. 2004;48(12):4556-61. https://doi.org/10.1128/ 17. Delgado-Blas JF, Ovejero CM, Abadia-Patiño L, Gonzalez-Zorn AAC.48.12.4556-4561.2004 PMID: 15561825 B. Coexistence of mcr-1 and blaNDM-1 in Escherichia coli from 33. World Health Organization (WHO)/Department of control Venezuela. Antimicrob Agents Chemother. 2016;60(10):6356-8. of neglected tropical diseases. Helminth control in school- https://doi.org/10.1128/AAC.01319-16 PMID: 27431212 aged children, a guide for managers of control programs. 18. Dalmolin TV, Martins AF, Zavascki AP, de Lima-Morales D, Geneva: WHO; 2012. Available from:http://www.who.int/ Barth AL. Acquisition of the mcr-1 gene by a high-risk clone of neglected_diseases/resources/9789241548267/en/ KPC-2-producing Klebsiella pneumoniae ST437/CC258, Brazil. 34. Coppi M, Cannatelli A, Antonelli A, Baccani I, Di Pilato V, Diagn Microbiol Infect Dis. 2018;90(2):132-3. https://doi. Sennati S, et al. A simple phenotypic method for screening org/10.1016/j.diagmicrobio.2017.09.016 PMID: 29169659 of MCR-1-mediated colistin resistance. Clin Microbiol Infect. 19. Saavedra SY, Diaz L, Wiesner M, Correa A, Arévalo SA, 2018;24(2):S1198-743X(17)30457-3. Reyes J, et al. Genomic and Molecular characterization of 35. Pacheco AB, Guth BE, Soares KC, Nishimura L, de Almeida DF, clinical isolates of Enterobacteriaceae harboring mcr-1 in Ferreira LC. Random amplification of polymorphic DNA reveals Colombia, 2002 to 2016. Antimicrob Agents Chemother. serotype-specific clonal clusters among enterotoxigenic 2017;61(12):e008441-17. https://doi.org/10.1128/AAC.00841-17 Escherichia coli strains isolated from humans. J Clin Microbiol. PMID: 28893788 1997;35(6):1521-5. PMID: 9163473 20. Ortega-Paredes D, Barba P, Zurita J. Colistin-resistant 36. European Committee on Antimicrobial Susceptibility Testing Escherichia coli clinical isolate harbouring the mcr-1 gene in (EUCAST). Breakpoint tables for interpretation of MICs and Ecuador. Epidemiol Infect. 2016;144(14):2967-70. https://doi. zone diameters. Version 8.0. Växjö: EUCAST; 1 Jan 2018. org/10.1017/S0950268816001369 PMID: 27586373 Available from:http://www.eucast.org/fileadmin/src/media/ 21. Sennati S, Di Pilato V, Riccobono E, Di Maggio T, Villagran AL, PDFs/EUCAST_files/Breakpoint_tables/v_8.0_Breakpoint_ Pallecchi L, et al. Citrobacter braakii carrying plasmid-borne Tables.pdf mcr-1 colistin resistance gene from ready-to-eat food from a 37. European Committee on Antimicrobial Susceptibility Testing market in the Chaco region of Bolivia. J Antimicrob Chemother. (EUCAST). EUCAST guideline for the detection of resistance 2017;72(7):2127-9. https://doi.org/10.1093/jac/dkx078 PMID: mechanisms and specific resistances of clinical and/or 28333311 epidemiological importance. Version 2.0. Växjö: EUCAST; July 22. Monte DF, Mem A, Fernandes MR, Cerdeira L, Esposito F, 2016. Available from:http://www.eucast.org/fileadmin/src/ Galvão JA, et al. Chicken meat as a reservoir of colistin- media/PDFs/EUCAST_files/Resistance_mechanisms/EUCAST_ resistant Escherichia coli strains carrying mcr-1 genes in South detection_of_resistance_mechanisms_170711.pdf America. Antimicrob Agents Chemother. 2017;61(5):e02718-16. 38. Giani T, Antonelli A, Caltagirone M, Mauri C, Nicchi J, https://doi.org/10.1128/AAC.02718-16 PMID: 28193665 Arena F, et al. Evolving beta-lactamase epidemiology in 23. Tijet N, Faccone D, Rapoport M, Seah C, Pasterán F, Ceriana Enterobacteriaceae from Italian nationwide surveillance, P, et al. Molecular characteristics of mcr-1-carrying plasmids October 2013: KPC-carbapenemase spreading among and new mcr-1 variant recovered from polyclonal clinical outpatients. Euro Surveill. 2017;22(31):30583. https://doi. Escherichia coli from Argentina and Canada. PLoS One. org/10.2807/1560-7917.ES.2017.22.31.30583 PMID: 28797330 2017;12(7):e0180347. https://doi.org/10.1371/journal. 39. Ribeiro TG, Novais Â, Branquinho R, Machado E, Peixe L. pone.0180347 PMID: 28678874 Phylogeny and comparative genomics unveil independent 24. Fernandes MR, Sellera FP, Esposito F, Sabino CP, Cerdeira L, diversification trajectories of qnrB and genetic platforms Lincopan N. Colistin-resistant mcr-1-positive Escherichia coli on within particular Citrobacter species. Antimicrob Agents public beaches, an infectious threat emerging in recreational Chemother. 2015;59(10):5951-8. https://doi.org/10.1128/ waters. Antimicrob Agents Chemother. 2017;61(7):e00234-17. AAC.00027-15 PMID: 26169406 https://doi.org/10.1128/AAC.00234-17 PMID: 28416556 40. Chavda KD, Chen L, Fouts DE, Sutton G, Brinkac L, Jenkins SG, 25. Aires CAM, da Conceição-Neto OC, Tavares E Oliveira TR, Dias et al. Comprehensive genome analysis of carbapenemase- CF, Montezzi LF, Picão RC, et al. Emergence of the plasmid- producing Enterobacter spp.: new insights into phylogeny, mediated mcr-1 gene in clinical KPC-2-producing Klebsiella population structure, and resistance mechanisms. MBio. pneumoniae Sequence Type 392 in Brazil. Antimicrob Agents 2016;7(6):e02093-16. https://doi.org/10.1128/mBio.02093-16 Chemother. 2017;61(7):e00317-17. https://doi.org/10.1128/ PMID: 27965456 AAC.00317-17 PMID: 28438940 41. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a 26. Sellera FP, Fernandes MR, Sartori L, Carvalho MPN, Esposito F, laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Nascimento CL, et al. Escherichia coli carrying IncX4 plasmid- Press; 1989. mediated mcr-1 and blaCTX-M genes in infected migratory 42. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Magellanic penguins (Spheniscus magellanicus). J Antimicrob Kulikov AS, et al. SPAdes: a new genome assembly algorithm Chemother. 2017;72(4):1255-6. PMID: 28031274 and its applications to single-cell sequencing. J Comput Biol. 27. Rossi F, Girardello R, Morais C, Cury AP, Martins LF, da 2012;19(5):455-77. https://doi.org/10.1089/cmb.2012.0021 Silva AM, et al. Plasmid-mediated mcr-1 in carbapenem- PMID: 22506599 susceptible Escherichia coli ST156 causing a blood infection: 43. Bi Z, Berglund B, Sun Q, Nilsson M, Chen B, Tärnberg M, et al. an unnoticeable spread of colistin resistance in Brazil? Clinics Prevalence of the mcr-1 colistin resistance gene in extended- (Sao Paulo). 2017;72(10):642-4. https://doi.org/10.6061/ spectrum β-lactamase-producing Escherichia coli from human clinics/2017(10)09 PMID: 29160428 faecal samples collected in 2012 in rural villages in Shandong 28. Bartoloni A, Cutts F, Leoni S, Austin CC, Mantella A, Province, China. Int J Antimicrob Agents. 2017;49(4):493- Guglielmetti P, et al. Patterns of antimicrobial use and 7. https://doi.org/10.1016/j.ijantimicag.2016.12.018 PMID: 28263896

42 www.eurosurveillance.org 44. Purohit MR, Chandran S, Shah H, Diwan V, Tamhankar AJ, Stålsby Lundborg C. Antibiotic resistance in an Indian rural community: a ‘One-Health’ observational study on commensal coliform from humans, animals, and water. Int J Environ Res Public Health. 2017;14(4):386-13. https://doi.org/10.3390/ ijerph14040386 PMID: 28383517 45. Trung NV, Matamoros S, Carrique-Mas JJ, Nghia NH, Nhung NT, Chieu TTB, et al. Zoonotic transmission of mcr-1 colistin resistance gene from small-scale poultry farms, Vietnam. Emerg Infect Dis. 2017;23(3):529-32. https://doi.org/10.3201/ eid2303.161553 PMID: 28221105 46. Zhong LL, Zhang YF, Doi Y, Huang X, Zhang XF, Zeng KJ, et al. Co-production of MCR-1 and NDM-1 by colistin-resistant Escherichia coli isolated from a healthy individual. Antimicrob Agents Chemother. 2016;61(1):6. PMID: 27821458 47. Zurfluh K, Stephan R, Widmer A, Poirel L, Nordmann P, Nüesch HJ, et al. Screening for fecal carriage of MCR-producing Enterobacteriaceae in healthy humans and primary care patients. Antimicrob Resist Infect Control. 2017;6(28):28. https://doi.org/10.1186/s13756-017-0186-z PMID: 28316780 48. Lu X, Hu Y, Luo M, Zhou H, Wang X, Du Y, et al. MCR-1.6, a New MCR variant carried by an IncP plasmid in a colistin- resistant Salmonella enterica Serovar Typhimurium isolate from a healthy individual. Antimicrob Agents Chemother. 2017;61(5):e02632-16-13. 49. Wang Y, Tian GB, Zhang R, Shen Y, Tyrrell JM, Huang X, et al. Prevalence, risk factors, outcomes, and molecular epidemiology of mcr-1-positive Enterobacteriaceae in patients and healthy adults from China: an epidemiological and clinical study. Lancet Infect Dis. 2017;17(4):390-9. https://doi. org/10.1016/S1473-3099(16)30527-8 PMID: 28139431 50. Zhang X-F, Doi Y, Huang X, Li H-Y, Zhong L-L, Zeng K-J, et al. Possible transmission of mcr-1-harboring Escherichia coli between companion animals and human. Emerg Infect Dis. 2016;22(9):1679-81. https://doi.org/10.3201/eid2209.160464 PMID: 27191649 51. Zhang R, Huang Y, Chan EW, Zhou H, Chen S, Zhang R, et al. Dissemination of the mcr-1 colistin resistance gene. Lancet Infect Dis. 2016;16(3):291-2. https://doi.org/10.1016/S1473- 3099(16)00062-1 PMID: 26973306 52. Chen K, Chan EW, Xie M, Ye L, Dong N, Chen S. Widespread distribution of mcr-1-bearing bacteria in the ecosystem, 2015 to 2016. Euro Surveill. 2017;22(39):17-00206. 53. Gröndahl-Yli-Hannuksela K, Lönnqvist E, Kallonen T, Lindholm L, Jalava J, Rantakokko-Jalava K, et al. The first human report of mobile colistin resistance gene, mcr-1, in Finland. APMIS. 2018;126(5):413-7. https://doi.org/10.1111/apm.12834 PMID: 29696722 54. Servicio Nacional de Sanidad Agropecuaria e Inocuidad Alimentaria (SENASAG). [National Service of Agricultural Health and Food Safety]. Registro de Productos de Uso Veterinario e Insumos Pecuarios. [Registration of Products for Veterinary Use and Livestock Supplies]. Bolivia: SENASAG; 2018. Spanish. Available from: http://190.129.48.189/egp/ productosVeterinarios.html

This article is copyright of the authors or their affiliated institutions, 2018.

www.eurosurveillance.org 43 Research article Congenital brain abnormalities during a Zika virus epidemic in Salvador, Brazil, April 2015 to July 2016

Mariana Kikuti1,2, Cristiane W. Cardoso3, Ana P.B. Prates3, Igor A.D. Paploski1,2, Uriel Kitron1,4, Mitermayer G. Reis1,2, Ganeshwaran H. Mochida5,6,7, Guilherme S. Ribeiro1,2 1. Fundação Oswaldo Cruz, Salvador, Brazil 2. Universidade Federal da Bahia, Salvador, Brazil 3. Secretaria Municipal de Saúde de Salvador, Salvador, Brazil 4. Emory University, Atlanta, United States 5. Boston Children’s Hospital, Boston, United States 6. Massachusetts General Hospital, Boston, United States 7. Harvard Medical School, Boston, United States Correspondence: Guilherme S. Ribeiro ([email protected])

Citation style for this article: Kikuti Mariana, Cardoso Cristiane W., Prates Ana P.B., Paploski Igor A.D., Kitron Uriel, Reis Mitermayer G., Mochida Ganeshwaran H., Ribeiro Guilherme S.. Congenital brain abnormalities during a Zika virus epidemic in Salvador, Brazil, April 2015 to July 2016. Euro Surveill. 2018;23(45):pii=1700757. https://doi. org/10.2807/1560-7917.ES.2018.23.45.1700757

Article submitted on 01 Nov 2017 / accepted on 27 Jun 2018 / published on 08 Nov 2018

Background: North-eastern Brazil was the region most Introduction affected by the outbreak of congenital Zika syndrome Early in 2015, large outbreaks of acute exanthematous that followed the 2015 Zika virus (ZIKV) epidemics, illness were detected in several states in north-eastern with thousands of suspected microcephaly cases Brazil. By April 2015, Zika virus (ZIKV) was identified reported to the health authorities, mostly between as the aetiology of the illness [1,2]. A few months after late 2015 and early 2016. Aim: To describe clinical the epidemic peak in May 2015, an increase in new- and epidemiological aspects of the outbreak of con- borns with microcephaly was noted in north-eastern genital brain abnormalities (CBAs) and to evaluate Brazil [3] and promptly gathered global attention due the accuracy of different head circumference screen- to a possible link between gestational ZIKV infection ing criteria in predicting CBAs. Method: Between April and microcephaly. Since then, evidence for a causal 2015 and July 2016, the Centers for Information and association between in utero exposure to ZIKV and Epidemiologic Surveillance of Salvador, Brazil investi- microcephaly and other neurological complications has gated the reported cases suspected of microcephaly emerged [4-6]. The constellation of clinical manifesta- and, based on intracranial imaging studies, confirmed tions of congenital ZIKV infection may be referred to as or excluded a diagnosis of CBA. Sensitivity, specificity ‘congenital Zika syndrome’ [7]. and positive and negative predictive values of different head circumference screening criteria in predicting In light of the surge of microcephaly cases, the CBAs were calculated. Results: Of the 365 investigated Brazilian Ministry of Health (BMoH) declared a national cases, 166 (45.5%) had confirmed CBAs. The most public health emergency in November 2015 and initi- common findings were intracranial calcifications and ated a surveillance programme for identification of ventriculomegaly in 143 (86.1%) and 111 (66.9%) of suspected microcephaly cases [8]. All health facilities the 166 CBA cases, respectively. Prevalence of CBAs were required to report such cases [8-10] and encour- peaked in December 2015 (2.24 cases/100 live births). aged to report spontaneous abortions and stillbirths in Cases of CBAs were significantly more likely to have women with a history of a rash during pregnancy, in been born preterm and to mothers who had clinical a national reporting system. Although head circumfer- manifestations of arboviral infection during pregnancy. ence was immediately adopted as the primary criterion None of the head circumference screening criteria per- for screening cases suspected of congenital abnormali- formed optimally in predicting CBAs. Conclusion: This ties by Zika, not all children with neurological impair- study highlights the magnitude of neurological con- ment due to ZIKV present with microcephaly at birth sequences of the ZIKV epidemic and the limitations of [11]. Therefore, it is important to understand how well head circumference in accurately identifying children the criteria used to detect microcephaly can predict the with CBA. Gestational symptoms compatible with ZIKV congenital brain alterations of ZIKV. infection should be combined with imaging studies for efficient detection of suspect CBAs during ZIKV Here, we describe the characteristics of the cases with epidemics. congenital brain abnormalities (CBAs) confirmed by intracranial imaging studies among the reported cases

44 www.eurosurveillance.org Figure Epidemic curve of the reported cases suspected of microcephaly per week of birth by status of congenital brain abnormalitya, Salvador, Brazil, April 2015–July 2016 (n = 631b)

Conﬁrmeda (n = 165) Excluded (n = 198) Not investigated (n = 268)

Number of cases 20

0 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32

Epidemiological week 2015–2016

a Confirmation based on intracranial imaging studies. b The 631 suspected cases depicted are those with available data on date of birth.

of suspected microcephaly in Salvador, Brazil. We also From 12 December 2015 to 12 March 2016, the sus- identified clinical manifestations during pregnancy pected case definition was updated as newborns with that were associated with CBAs, and evaluated the head circumference ≤ 32 cm for term and the previous accuracy of different screening criteria based on head criteria for preterm newborns were maintained [9]. circumference for predicting CBAs. Lastly, on 13 March 2016, the suspected microcephaly case definition was changed to newborns with Methods head circumference < 3rd percentile of the World Health Organization (WHO) Child Growth Standards for term Suspected microcephaly case definition for and < 3rd percentile of the INTERGROWTH-21st standards mandatory reporting for preterm [10]. Mandatory reporting was performed From 17 November to 11 December 2015, the BMoH based solely on head circumference parameters of the defined suspected microcephaly cases as newborns newborns, regardless of a Zika diagnosis of the moth- with head circumference measures ≤ 33 cm for term ers. However, reporting of spontaneous abortions, (≥ 37 weeks) or < 3rd percentile of the Fenton Preterm stillbirths or pregnancies with any detected alterations Growth Chart for preterm (< 37 weeks) [8] and required in the fetal central nervous system in women with a mandatory reporting of newborns fulfilling this case self-reported history of rash during pregnancy was also definition. This criterion was also applied to cases encouraged regardless of head circumference, but not who were retrospectively identified during this period. mandatory [9,10].

www.eurosurveillance.org 45 Table 1 Clinical characteristics of suspected microcephaly cases, Salvador, Brazil, April 2015–July 2016 (n = 365)

Congenital brain abnormalities diagnosis Clinical characteristics Confirmed (n = 166) Excluded (n = 199) n % n % Male sex 76 45.8 65 32.7 Gestational age at birtha < 37 weeks 54 33.3 13 7.0 37–42 weeks 108 66.7 172 93.0 Head circumference Head circumference (cm), median and IQRb,c 30.0 28.0–31.0 32.0 31.0–32.0 Head circumference (cm), min–maxb,c 21.5–42.0 28.0–36.0 Intracranial imaging performed Ultrasound 136 81.9 192 96.5 Computed tomography 53 31.9 9 4.5 Magnetic resonance imaging 17 10.2 4 2.0 Image findings consistent with congenital malformations Intracranial calcifications 143 86.1 0 0.0 Ventriculomegaly 111 66.9 0 0.0 Agenesis of the corpus callosum 20 12.1 0 0.0 Dysgenesis of the corpus callosum 19 11.5 0 0.0 Lissencephaly 17 10.2 0 0.0 Cerebellar abnormalitiesd 9 5.4 0 0.0 Anencephaly 3 1.8 0 0.0 Other findings Oligohydramnios 17 10.2 7 3.5 Intrauterine growth restriction 14 8.4 2 1.0 Subependymal cysts 16 9.6 7 3.5 Arthrogryposis 11 6.6 0 0.0 Auditory abnormalitiese 20 19.2 10 11.5 Ophthalmological abnormalitiesf 13 18.8 3 4.5 Deathg 6 3.6 0 0.0

IQR: interquartile range. Suspected microcephaly cases may have undergone more than one imaging testing and presented with more than one finding. a Data available for 162 confirmed cases and 185 excluded cases. The lowest gestational age for the confirmed and excluded cases were 20 and 30 weeks, respectively. b Data on head circumference were not available for two confirmed cases. Data on head circumference of three confirmed cases and eight excluded cases were not considered in the analysis due to a late measurement without a precise date (> 28 days post birth). c Some of the cases from both groups of confirmed and excluded congenital brain abnormalities diagnosis had a head circumference size greater than the screening limits used to detect microcephaly. This is because reporting cases of spontaneous abortions, stillbirths or pregnancies with any detected alterations in the fetal central nervous system in women with self-reported history of rash during pregnancy was also encouraged, though not mandatory. d Cerebellar vermis agenesis or cerebellar hypoplasia. e Data on auditory abnormalities were available for 104 confirmed cases (86 by auditory screening method, three by brainstem evoked response audiometry test (BERA) and 15 by both) and for 87 excluded cases (79 by auditory screening method, three by BERA test and five by both). f Data on ophthalmological abnormalities were available for 69 confirmed cases and for 67 excluded cases. Among the confirmed cases, two had pigmentary abnormalities in macula, one abnormal red reflex, one bilateral chorioretinal atrophy and optic disc hypoplasia, one cataracts, one macular atrophy, one corneal excoriation, one chorioretinitis, one optic disc excavation, one coloboma, and three without information on type of abnormality. Among excluded cases, one had retinal haemorrhage, one retinal excavation and one cataracts. g Death occurred on the day of birth for two cases, within 2 days for one case, within 2 months for two cases and within 4 months for one case.

Investigation of reported cases suspected of Municipal Secretary of Health in charge of the investi- microcephaly gation of the reported suspected cases. Investigations Salvador, the fourth largest city in Brazil, was one of the were performed by reviewing medical records for north-eastern cities most affected by the microcephaly intracranial imaging studies. In addition, mothers of epidemic [3]. The Salvador Centers for Information and reported suspect microcephaly cases were interviewed Epidemiologic Surveillance (CIES) is the branch of the about clinical manifestations during pregnancy using a

46 www.eurosurveillance.org Table 2 Maternal clinical characteristics of suspected microcephaly cases, Salvador, Brazil, April 2015–July 2016 (n = 365)

Congenital brain abnormalities diagnosis Confirmed Excluded Characteristics OR 95% CI p value (n = 166) (n = 199) n/Na % n/Na % Mother’s age in years, median and IQRb 26 21–32 25 21– 31 NA NA 0.38 Type of gestation Multifetal 5/163 3.1 4/195 2.0 1.51 0.32–7.74 0.54 Single 158/163 96.9 191/195 98.0 1 NA Symptoms during pregnancy Feverc 66/155 42.3 42/187 22.5 2.56 1.56–4.21 < 0.001 Exanthema 118/161 73.3 68/193 35.2 5.04 3.12–8.19 < 0.001 Trimester of exanthemad First trimester 75/108 69.4 25/57 43.9 2.91 1.42–5.97 0.001 Second or third trimester 33/108 30.6 32/57 56.1 1 NA Other arboviral-infection-like symptoms reported during pregnancy Pruritus 72/133 54.1 36/143 25.2 3.51 2.05–6.04 < 0.001 Arthralgia 60/135 44.4 31/143 21.7 2.89 1.66–5.06 < 0.001 Myalgia 51/135 37.8 25/142 17.6 2.84 1.58–5.17 < 0.001 Headache 48/133 36.1 25/142 17.6 2.64 1.46–4.83 0.001 Retro-orbital pain 24/133 18.1 6/142 4.2 4.99 1.89–15.38 < 0.001 Conjunctival hyperaemia 21/133 15.8 9/142 6.3 2.77 1.16–7.14 0.01

NA: not applicable; CI: confidence interval (95%); IQR: interquartile range; OR: odds ratio. a The denominators in this column may vary as they are based on the number of cases with known information on the characteristic in question. b Data on mothers’ ages were not available for nine confirmed and four excluded cases. c Fever was defined as axillary temperatures of ≥ 37.8 °C. d Data on trimester of exanthema are presented for the mothers who had exanthema (n = 118 for the mothers of confirmed congenital brain abnormality cases, and n = 68 for those of the excluded cases). For 10 confirmed and 11 excluded cases the trimester of the exanthema was unknown.

standardised questionnaire [10]. After concluding the investigation, CIES updated the national reporting sys- Statistical analysis tem with the obtained information. An epidemiological curve of the temporal distribution of the suspected microcephaly cases stratified accord- Image-confirmed congenital brain abnormality ing to the confirmation status was constructed by epi- case definition demiological week of the date of birth. Records with In the present study, we analysed data on suspected incomplete information on date of birth were excluded microcephaly cases investigated by CIES up to 13 from the epidemiological curve. Prevalence of CBA per September 2016. According to the availability of data month was calculated dividing the number of imaging on prenatal or postnatal intracranial imaging studies, confirmed cases by the monthly number of live births the reported cases were classified as either investi- from mothers residing in Salvador. To estimate the gated or not investigated. Investigated cases whose average annual prevalence of CBA, we divided the CBA intracranial ultrasound, computed tomography, or prevalence calculated for the complete study period by magnetic resonance imaging results reported intrac- the number of months in the study period and multi- ranial calcifications, ventriculomegaly, dysgenesis or plied the result by 12. Live birth data were obtained at agenesis of the corpus callosum, lissencephaly, cer- the National Birth Registration System (SINASC) [12]. ebellar abnormalities, or anencephaly were classified as confirmed CBA cases. Reports of hydrocephalus or Clinical characteristics of confirmed and excluded colpocephaly were consolidated as ventriculomegaly. cases of CBA were presented as frequencies and medi- Suspected microcephaly cases that underwent imaging ans. Two-tailed Fisher exact test and odds ratios with studies and did not exhibit any of the previous findings 95% confidence intervals (CI) were used to test for dif- were excluded from consideration as CBA cases. ferences in the frequencies of clinical characteristics of suspected cases and gestational characteristics of cases’ mothers between confirmed and excluded CBA cases. The Wilcoxon rank test was used to test

www.eurosurveillance.org 47 Table 3 Imaging findings of confirmed congenital brain abnormalities cases, according to the timing of maternal exanthema, Salvador, Brazil, April 2015–July 2016 (n = 151)

Pregnancy trimester of exanthema First Second Third Without exanthema Congenital brain abnormalities (n = 75) (n = 24) (n = 9) (n = 43) n % n % n % n % Intracranial calcifications 68 90.7 23 95.8 7 77.8 32 74.4 Ventriculomegaly 48 64.0 18 75 8 88.9 28 65.1 Agenesis of the corpus callosum 11 14.7 3 12.5 2 22.2 4 9.3 Dysgenesis of the corpus callosum 8 10.7 5 20.8 0 0.0 5 11.6 Lissencephaly 6 8.0 4 16.7 2 22.2 4 9.3 Cerebellar abnormalitiesa 6 8.0 1 4.2 0 0.0 1 2.3 Arthrogryposis 6 8.0 2 8.3 0 0.0 2 4.7 Oligohydramnios 12 16.0 1 4.2 0 0.0 3 7.0 Intrauterine growth restriction 9 12.0 3 12.5 0 0.0 1 2.3 Subependymal cysts 7 9.3 1 4.2 1 11.1 6 14.0 a Cerebellar vermis agenesis or cerebellar hypoplasia.

for difference in maternal age and head circumference late reporting (> 28 days post birth) without a precise between confirmed and excluded CBA cases. In order to date on which head circumference was measured were investigate whether the imaging-detected CBAs varied excluded from the accuracy analysis. Data were ana- according to the presumptive timing of infection during lysed with Stata 14 [17]. pregnancy, the frequencies of each imaging abnormality in confirmed cases were compared according to the Ethics statement presence and timing of exanthema during pregnancy. A This investigation was performed using de-identi- two-tailed significance level of 0.05 was set. fied secondary data obtained by routine activities of the Epidemiological Surveillance Office/Municipal Imaging-confirmed and excluded CBA cases were used Secretariat of Health from Salvador, Bahia, Brazil. The to assess the accuracy of different head circumference Salvador Secretariat of Health and the Oswaldo Cruz microcephaly screening criteria for prediction of CBAs. Foundation Ethics Committee approved the study and The microcephaly screening criteria evaluated were granted a waiver of signed informed consent. those adopted by the BMoH from (i) November to 11 December 2015 [8]; (ii) from 12 December 2015 to 12 Results March 2016 [9]; and (iii) from 13 March 2016 to the pre- By 13 September 2016, Salvador CIES had received sent [10]; as well as (iv) the screening criteria recom- 650 reports of suspected microcephaly cases, who mended by the Pan American Health Organisation (< -2 were born between April 2015 and July 2016. Of these, standard deviation (SD) of the Fenton Preterm Growth review of medical records to retrieve results of intrac- Chart according to sex and gestational age for preterm ranial imaging studies was completed for 365 cases. newborns and < 3rd percentile of the WHO Child Growth Among those, 166 (45.5%) had imaging findings con- Standards according to sex for term newborns) [13]; (v) sistent with a diagnosis of CBA, while 199 (54.5%) did the Fenton Preterm Growth Chart (< -2 SD according to not. The epidemiological curve of the temporal distri- sex and gestational age) [14]; (vi) the INTERGROWTH- bution of reported cases (built for 631 cases with avail- 21st standards (< -2 SD according to sex and gestational able data on date of birth) peaked between week 47 age) [15]; and (vii) the WHO Child Growth Standards of 2015 (22–28 November) and week 4 of 2016 (24–30 for term newborns (< -2 SD according to sex for term January) for both the suspected microcephaly reported newborns) [16]. Some of the criteria require detailed cases and the imaging-confirmed CBA cases (Figure 1). information on gestational age (weeks and days) but Prevalence of imaging-confirmed CBA was 1.22 cases gestational age was recorded in full weeks in our data- per 100 live births in November 2015, 2.24 cases per set. We used the number of weeks plus 0 days in such 100 live births in December 2015, and 1.55 cases per cases. Accuracy of the criteria in predicting imaging- 100 live births in January 2016. Prevalence of imaging- confirmed CBAs was assessed by calculating sensitiv- confirmed CBA for the whole study period (April 2015 ity, specificity, positive and negative predictive values, to July 2016) was 0.34 per 100 live births and the one- and their respective 95% CIs. Records with no informa- year adjusted annual prevalence of image-confirmed tion on head circumference, gestational age, or with CBA was estimated as 0.26 per 100 live births. The

48 www.eurosurveillance.org last CBA case confirmed by imaging studies during the When the confirmed cases were classified according study period was born in epidemiological week 15 of to the timing of maternal rash during pregnancy (first, 2016 (10–16 April). second, third trimester or no rash), there were no statistically significant differences in the frequency of Male sex was more frequent among the imaging-con- CBAs, nor in the frequency of arthrogryposis, oligohy- firmed CBA cases (45.8%) than among excluded cases dramnios, and intrauterine growth restriction, between (32.7%) (odds ratio (OR): 1.74; 95% CI: 1.11–2.72; the four groups. An exception was observed for cal- p = 0.01) (Table 1). Confirmed cases were more likely to cifications, that were more frequently present when have been born preterm (< 37 weeks) (OR: 6.62; 95% CI: exanthema occurred during first trimester than when 3.35–13.79; p < 0.001) and had a lower head circumfer- compared with mothers who did not present exan- ence median (30 cm; interquartile range (IQR): 28–31) thema during pregnancy (p = 0.02) (Table 3). compared with excluded cases (32 cm; IQR: 31–32) (p < 0.001). Confirmed cases also had a broader head Among the different head circumference criteria used circumference range (range: 21.5 cm – 42 cm vs 28 cm to screen for microcephaly, the first criterion adopted – 36 cm; p < 0.001). The most frequent CBAs observed by the BMoH from November 2015 to December 2015 among confirmed cases were intracranial calcifica- was the one with the highest sensitivity (83.6%) and tions (86.1%) and ventriculomegaly (66.9%) (Table 1). lowest specificity (7.3%) in predicting the presence of Agenesis of the corpus callosum (12.1%), dysgenesis CBAs (Table 4). On the other hand, the INTERGROWTH- of the corpus callosum (11.5%), lissencephaly (10.2%), 21st standards had the lowest sensitivity (63.4%) and cerebellar abnormalities (5.4%), and anencephaly highest specificity (72.4%). Positive predictive value (1.8%) were identified in a minority of the confirmed was the highest for the INTERGROWTH-21st standards cases. Of note, intracranial calcifications were associ- (63.9%) and the lowest for the criterion adopted by ated with other brain lesions, being present in 83.8% BMoH from December 2015 to March 2016 (41.1%). (93/111) of those with ventriculomegaly, but in 19.7% Negative predictive value was the highest for the WHO (50/254) of those without it; in 94.1% (16/17) and Child Growth Standards (77.8%) and the lowest for the 36.5% (127/348) of those with and without lissenceph- BMoH criterion used between November and December aly, in 94.7% (18/19) and 36.1% (125/346) of those with 2015 (36.1%). and without dysgenesis of the corpus callosum, and in 85.0% (17/20) and 36.5% (126/345) of those with and Discussion without agenesis of the corpus callosum (p < 0.001 for In this study, we described a high prevalence of con- all the comparisons). In terms of other findings, arthro- firmed CBAs in Salvador, as high as 2.2% of the live gryposis was found in 6.6% of the confirmed cases, births in December 2015. The prevalence of image- but in none of the excluded cases. Oligohydramnios, confirmed CBA estimated for the study period adjusted intrauterine growth restriction, subependymal cysts, for one year was 52 times higher than the estimated and auditory and ophthalmological abnormalities were baseline prevalence of microcephaly in the north-east also statistically more frequent among the confirmed region (average of 5 cases per 100,000 live births per than the excluded cases (Table 1). Intracranial calci- year, between 2000 and 2014) [18]. Unfortunately, we fications were also more frequently observed among did not have information on serological or virological those with arthrogryposis than among those without ZIKV testing, which would allow ascertaining the aetiol- it (72.7% (8/11) vs 38.1% (135/354)) and among those ogy of such an outbreak. However, the peak of births of with oligohydramnios than among those without it babies with microcephaly occurred 30–33 weeks after (66.7% (16/24) vs 37.2% (127/341)) (p < 0.001 for both the peak of ZIKV epidemic in Salvador [3], and this is comparisons). consistent with the growing body of evidence suggesting that the first trimester of pregnancy is the period Maternal age at birth and type of gestation (single when ZIKV infections pose the highest risk of adverse vs multifetal) were not associated with imaging con- fetal outcome [3,19,20]. Taken together, it is reasonable firmation of CBAs (Table 2). Frequency of exanthema to assume that most of the imaging-confirmed cases in during pregnancy among mothers of children in the this study were due to congenital ZIKV infection. confirmed group was 73.3%. Among those, 69.4% had the rash during the first trimester and 30.6% during As we only considered cases with specific neuroimag- the second or third trimester (22.2% during the sec- ing findings as confirmed cases, we certainly under- ond and 8.3% during the third). Mothers of children estimated cases of congenital ZIKV infection. Several in the confirmed CBA group were more likely to have suspected cases had not been investigated by the time had exanthema during pregnancy (OR: 5.04; 95% CI: we analysed the data and the imaging modality most 3.12–8.19) than mothers of children whose diagnosis commonly used was prenatal or postnatal intracranial of CBA was excluded, especially in the first trimester ultrasound, which is not an optimal modality to detect (OR: 2.91; 95% CI: 1.42–5.97) when compared with the abnormalities of the corpus callosum and cerebral cor- second and third trimester. All other symptoms com- tex. In addition, suspected microcephaly cases were monly observed during arboviral infections were more reported based on birth head circumference, which frequent during pregnancy on the mothers of children could be well within normal limits in some cases of in the confirmed group (Table 2). congenital ZIKV infection [11]. Although reporting of

www.eurosurveillance.org 49 spontaneous abortions, stillbirths and fetuses pre- ventriculomegaly. Although these findings are not spe- senting alterations in the central nervous system was cific for congenital Zika syndrome, they have been fre- also encouraged, allowing us to confirm a few cases quently observed among laboratory-confirmed cases of with normal or large head circumference at birth, we congenital Zika syndrome [25-28]. Anencephaly has not could not evaluate whether there was an increase in been previously reported among laboratory-confirmed abortions and stillbirths in Salvador during the study cases of congenital ZIKV infection, and further studies period. On the other hand, some cases counted as con- are warranted to determine if it is part of the spectrum firmed could be due to other causes such as congenital of the congenital Zika syndrome. Auditory and ocular cytomegalovirus infection or genetic disorders, but the manifestations were present in approximately 20% of number of these cases is expected to be small, con- the confirmed cases. However, they were also found in sidering the baseline rate of microcephaly before the lower frequencies among the suspected microcephaly epidemic. In addition, in north-east Brazil (a region in cases with normal intracranial imaging studies. Since which Salvador was one of the epicenters for the ZIKV these manifestations have been linked to congenital outbreak), only 1.3% of confirmed cases of infection- ZIKV infection [29,30], it is important to further investi- related microcephaly during the 2015–16 period had gate whether ZIKV infection can cause auditory or ocu- laboratory evidence of syphilis, toxoplasmosis, cyto- lar lesions in the absence of structural malformations megalovirus, or herpes simplex [21]. in the brain and to monitor for long-term consequences in ZIKV-exposed babies born with no alterations in A similar increase in microcephaly cases was reported brain imaging studies. In addition, as preterm babies in other locations where ZIKV epidemics have occurred, may have more auditory and visual complications than such as Colombia, where the prevalence of micro- term babies, we investigated whether these sensory cephaly also increased around 6 months after the disorders were associated with premature delivery peak of ZIKV transmission in July 2016. However, the and found a higher frequency of auditory abnormali- microcephaly prevalence reported in Colombia peaked ties among preterm than term babies (26.5% vs 11.4%; at 17.7 cases per 10,000 live births, much lower than p = 0.01), but this difference was not observed for oph- observed in Salvador [22]. Potential reasons for this thalmological abnormalities. Further studies are nec- difference may include, variable intensity levels of ZIKV essary to determine the specific contribution of both transmission, differences in circulating ZIKV strains prematurity and ZIKV-related neurological injury on the and different case definitions and surveillance criteria. occurrence of sensory disorders in these children. Further, co-circulation of other arboviruses (dengue and chikungunya, for example), differences in mos- Frequency of exanthema during pregnancy among quito control measures, and prior exposure to yellow mothers of children in the confirmed group was 73.3%. fever vaccination could be contributing factors [22,23]. It has been previously estimated that only 20% of the Additionally, Brazil was the first country in the Americas ZIKV infections are symptomatic [31], but other studies to experience a large outbreak of ZIKV and to detect have shown similarly high frequencies of exanthema an increase in microcephaly cases, and this allowed among mothers who gave birth to children with congen- other countries as Colombia to issue recommendations ital Zika syndrome [7,32,33]. Recall bias and the sur- for delaying pregnancies, which might have resulted in veillance system associating a rash during pregnancy decreased risk of congenital abnormalities associated as a marker for microcephaly risk may have accounted with ZIKV infection during pregnancy [22]. for the high proportion of symptomatic women in this series. However, both confirmed and excluded CBA Female newborns were overrepresented among the cases originated from the same reported dataset and reported cases who had a CBA diagnosis excluded. only 35.2% of the mothers of children in the excluded This finding may be due to the application of the group reported a rash, representing a significant dif- same head circumference screening criteria for report- ference. Therefore, it is likely that symptomatic ZIKV ing boys and girls suspected of microcephaly until 12 infection during pregnancy truly poses a higher risk of March 2016 (period during which 85% of the suspect CBAs. Similar findings of an association of rash dur- cases had been reported), since head circumference ing pregnancy with increased CBA risk was previously of girls tend to be smaller than boys at the same ges- reported in Brazil [7,32,33], although an absence of tational age [24]. We also found that the frequency of such association was noted in the United States (US) preterm births among the confirmed CBA cases was [34]. significantly greater than in excluded cases. Although this finding suggests that congenital Zika syndrome Among mothers of children in the confirmed group who could be associated with preterm birth, we could not reported a history of rash during pregnancy, 69.4% determine from the available data whether the early had the rash during the first trimester, but, in addition, births were natural in their occurrence or due to a med- 22.2% had rash during the second trimester and 8.3% ical decision in the presence of fetal anomalies and during the third trimester. Data linking ZIKV infection in distress. the second and third trimester to congenital malformations are still scarce [27,32], and our findings reinforce The most frequent imaging findings among the con- that second and third trimester infections may also firmed cases were intracranial calcifications and lead to congenital Zika syndrome with CBAs.

50 www.eurosurveillance.org Head circumference-based criteria were primarily used of children with and without microcephaly or congeni- during the epidemic of congenital ZIKV infection. These tal abnormalities, who were exposed to ZIKV in utero can be easily applied in any clinical setting and do not are needed to fully understand the full spectrum of require any special equipment. Although microcephaly congenital Zika syndrome. in general is a risk factor for developmental delay, according to the US National Collaborative Perinatal Project, which followed a cohort of newborns, only Acknowledgements 11% of children with microcephaly (head circumfer- We thank health professionals in Salvador, Brazil, especial- ence ≤ -2SD) at birth had IQ ≤ 70 at 7 years of age [35], ly those working in surveillance activities, those providing suggesting that microcephaly has a relatively low spec- healthcare for children born with congenital malformations ificity in predicting poor neurodevelopmental outcome and those providing support for affected children’s families. We also thank the children reported as suspected cases of in the general population. Further, cases with congeni- Zika congenital syndrome and their parents. tal ZIKV infection without microcephaly at birth have been reported [36]. On the other hand, the severity of The Brazilian National Council of Technological and Scientific anomalies in neuroimaging during the neonatal period Development, the Brazilian Coordination for the Improvement has been shown to have a good prognostic value for of Higher Education, the Bahia Foundation for Research Support, the National Institute of Neurological Disease and predicting poor developmental outcomes in sympto- Stroke, the Manton Center for Orphan Disease Research, matic congenital cytomegalovirus infection [37,38], Boston Children’s Hospital Faculty Career Development which shares many clinical and radiological similarities Award and the David Rockefeller Center for Latin American to congenital ZIKV infection. Therefore, it is important Studies at Harvard University provided funding for the study, to evaluate the accuracy of different microcephaly cri- but they played no role in data analyses. teria in predicting CBAs.

In this regard, none of the criteria performed particu- Conflict of interest larly well. Overall, the INTERGROWTH-21st standards None declared. had a better performance, with sensitivity, specificity and positive and negative predictive values over 60%. A significantly low specificity of the criteria used by the Authors’ contributions BMoH until 11 March 2016 was also noted. There is only MK, CWC, APBP, UK, MGR, and GSR conceived the study; one small prior study with 31 cases that used imaging findings consistent with congenital infection as the CWC and APBP provided access to the data; reference criteria for sensitivity estimation of different head circumference criteria [39]. Compared with this MK and IADP analysed the data; prior study, our study revealed lower sensitivity and UK and GSR oversaw data analysis; much lower specificity. Combined with relatively low positive and negative predictive values, these data MK, GHM, and GSR wrote first draft of the manuscript; clearly demonstrate the limitations of head circumference in accurately identifying children with CBA during MK, CWC, APBP, IADP, UK, MGR, GHM, and GSR participated in the discussion of the study and critical reviewed and ap- ZIKV epidemics. proved the final manuscript.

Novel screening methods for congenital Zika syndrome during ZIKV epidemics that incorporate additional References parameters to head circumference are urgently needed 1. Campos GS, Bandeira AC, Sardi SI. Zika Virus Outbreak, Bahia, in order to detect the maximum number of affected chil- Brazil. Emerg Infect Dis. 2015;21(10):1885-6. https://doi. dren and fetuses for further specific health assistance, org/10.3201/eid2110.150847 PMID: 26401719 2. Zanluca C, Melo VC, Mosimann ALP, Santos GI, Santos CN, Luz without compromising specificity [40,41]. Prenatal and K. First report of autochthonous transmission of Zika virus in postnatal intracranial ultrasound triage performed by Brazil. Mem Inst Oswaldo Cruz. 2015;110(4):569-72. https:// doi.org/10.1590/0074-02760150192 PMID: 26061233 experienced ultrasonographers may be an efficient 3. Paploski IAD, Prates APPB, Cardoso CW, Kikuti M, Silva approach during outbreaks, especially for women who MMO, Waller LA, et al. Time Lags between Exanthematous experienced symptoms consistent with ZIKV infection Illness Attributed to Zika Virus, Guillain-Barré Syndrome, and Microcephaly, Salvador, Brazil. Emerg Infect Dis. during pregnancy. However, the importance of clini- 2016;22(8):1438-44. https://doi.org/10.3201/eid2208.160496 cal examination and follow-up of newborns, as well as PMID: 27144515 4. Rasmussen SA, Jamieson DJ, Honein MA, Petersen LR. Zika development of better serological and molecular tests, Virus and Birth Defects--Reviewing the Evidence for Causality. cannot be understated. N Engl J Med. 2016;374(20):1981-7. https://doi.org/10.1056/ NEJMsr1604338 PMID: 27074377 5. de Araújo TVB, Rodrigues LC, de Alencar Ximenes RA, de Our study highlights the magnitude of neurological Barros Miranda-Filho D, Montarroyos UR, de Melo APL, et al. consequences of the ZIKV epidemic in Salvador, Brazil, Association between Zika virus infection and microcephaly in Brazil, January to May, 2016: preliminary report of a case- further delineates congenital Zika syndrome, and control study. Lancet Infect Dis. 2016;16(12):1356-63. https:// identifies limitations of screening for congenital Zika doi.org/10.1016/S1473-3099(16)30318-8 PMID: 27641777 syndrome based on head circumference that was per- 6. Brasil P, Pereira JP Jr, Moreira ME, Ribeiro Nogueira RM, Damasceno L, Wakimoto M, et al. Zika Virus formed during the recent epidemics. Follow-up studies Infection in Pregnant Women in Rio de Janeiro. N Engl J www.eurosurveillance.org 51 Med. 2016;375(24):2321-34. https://doi.org/10.1056/ 22. Cuevas EL, Tong VT, Rozo N, Valencia D, Pacheco O, Gilboa NEJMoa1602412 PMID: 26943629 SM, et al. Preliminary Report of Microcephaly Potentially 7. Miranda-Filho DB, Martelli CMT, Ximenes RA, Araújo Associated with Zika Virus Infection During Pregnancy - TVB, Rocha MAW, Ramos RCF, et al. Initial Description Colombia, January-November 2016. MMWR Morb Mortal Wkly of the Presumed Congenital Zika Syndrome. Am J Public Rep. 2016;65(49):1409-13. https://doi.org/10.15585/mmwr. Health. 2016;106(4):598-600. https://doi.org/10.2105/ mm6549e1 PMID: 27977645 AJPH.2016.303115 PMID: 26959258 23. Butler D. Brazil asks whether Zika acts alone to cause 8. Secretaria de Vigilância em Saúde. Procedimentos birth defects. Nature. 2016;535(7613):475-6. https://doi. preliminares a serem adotados para a vigilância dos casos de org/10.1038/nature.2016.20309 PMID: 27466104 microcefalia no Brasil. [Preliminary procedures to be adopted 24. Barbier A, Boivin A, Yoon W, Vallerand D, Platt RW, Audibert F, for surveillance of microcephaly cases in Brazil]. Brasília – DF : et al. New reference curves for head circumference at birth, by Minisrério da Saúde ; 2015. Portuguese. gestational age. Pediatrics. 2013;131(4):e1158-67. https://doi. 9. Secretaria de Vigilância em Saúde. Protocolo de vigilância e org/10.1542/peds.2011-3846 PMID: 23509164 resposta à ocorrência de microcefalia relacionada à infecção 25. de Fatima Vasco Aragao M, van der Linden V, Brainer-Lima pelo vírus zika [Protocol for surveillance and response to AM, Coeli RR, Rocha MA, Sobral da Silva P, et al. Clinical occurence of microcephaly related to Zika virus infection]. features and neuroimaging (CT and MRI) findings in presumed Brasília - DF: Ministério da Saúde; 2015. Portuguese. Available Zika virus related congenital infection and microcephaly: from: http://portalsaude.saude.gov.br/images/pdf/2015/ retrospective case series study. BMJ. 2016;353:i1901. https:// dezembro/09/Microcefalia---Protocolo-de-vigil--ncia-e- doi.org/10.1136/bmj.i1901 PMID: 27075009 resposta---vers--o-1----09dez2015-8h.pdf 26. Hazin AN, Poretti A, Di Cavalcanti Souza Cruz D, Tenorio M, van 10. Secretaria de Vigilância em Saúde. Protocolo de vigilância e der Linden A, Pena LJ, et al. Computed Tomographic Findings resposta à ocorrência de microcefalia relacionada à infecção in Microcephaly Associated with Zika Virus. N Engl J Med. pelo vírus zika [Protocol for surveillance and response to 2016;374(22):2193-5. https://doi.org/10.1056/NEJMc1603617 occurence of microcephaly related to Zika virus infection]. PMID: 27050112 Brasília - DF: Ministério da Saúde; 2016. Portuguese. Available 27. Brasil P, Pereira JP Jr, Moreira ME, Ribeiro Nogueira from: http://combateaedes.saude.gov.br/images/sala-de- RM, Damasceno L, Wakimoto M, et al. Zika Virus situacao/Microcefalia-Protocolo-de-vigilancia-e-resposta- Infection in Pregnant Women in Rio de Janeiro. N Engl J 10mar2016-18h.pdf Med. 2016;375(24):2321-34. https://doi.org/10.1056/ 11. van der Linden V, Pessoa A, Dobyns W, Barkovich AJ, Júnior HV, NEJMoa1602412 PMID: 26943629 Filho ELR, et al. Description of 13 Infants Born During October 28. Melo AS, Aguiar RS, Amorim MM, Arruda MB, Melo FO, Ribeiro 2015-January 2016 With Congenital Zika Virus Infection Without ST, et al. Congenital zika virus infection: Beyond neonatal Microcephaly at Birth - Brazil. MMWR Morb Mortal Wkly microcephaly. JAMA Neurol. 2016;73(12):1407-16. https://doi. Rep. 2016;65(47):1343-8. https://doi.org/10.15585/mmwr. org/10.1001/jamaneurol.2016.3720 PMID: 27695855 mm6547e2 PMID: 27906905 29. Leal MC, Muniz LF, Ferreira TSA, Santos CM, Almeida LC, Van 12. DATASUS. Sistema Nacional de Nascidos Vivos [Live Birth Der Linden V, et al. Hearing Loss in Infants with Microcephaly Information System]. Brasília - DF: Ministério da Saúde; 2016. and Evidence of Congenital Zika Virus Infection - Brazil, Portuguese. [Accessed 11 Jun 2016]. Available from: http:// November 2015-May 2016. MMWR Morb Mortal Wkly Rep. tabnet.datasus.gov.br/cgi/deftohtm.exe?sinasc/cnv/nvuf.def 2016;65(34):917-9. https://doi.org/10.15585/mmwr.mm6534e3 13. Pan American Health Organization (PAHO). Lineamientos PMID: 27585248 preliminares de vigilancia de microcefalia en recién nacidos en 30. de Paula Freitas B, de Oliveira Dias JR, Prazeres J, Sacramento entornos con riesgo de circulación de virus Zika. [Preliminary GA, Ko AI, Maia M, et al. Ocular Findings in Infants With guidelines for surveillance of microcephaly in newborns in Microcephaly Associated With Presumed Zika Virus Congenital environments with risk of Zika virus circulation]. Washington- Infection in Salvador, Brazil. JAMA Ophthalmol. 2016;1-19. DC: PAHO; 2016.Portuguese. PMID: 26865554 14. Fenton TR, Kim JH. A systematic review and meta-analysis 31. Duffy MR, Chen T-H, Hancock WT, Powers AM, Kool JL, Lanciotti to revise the Fenton growth chart for preterm infants. BMC RS, et al. Zika virus outbreak on Yap Island, Federated States Pediatr. 2013;13(1):59. https://doi.org/10.1186/1471-2431-13-59 of Micronesia. N Engl J Med. 2009;360(24):2536-43. https:// PMID: 23601190 doi.org/10.1056/NEJMoa0805715 PMID: 19516034 15. Villar J, Cheikh Ismail L, Victora CG, Ohuma EO, Bertino E, 32. França GVA, Schuler-Faccini L, Oliveira WK, Henriques CMP, Altman DG, et al. International standards for newborn weight, Carmo EH, Pedi VD, et al. Congenital Zika virus syndrome in length, and head circumference by gestational age and sex: Brazil: a case series of the first 1501 livebirths with complete the Newborn Cross-Sectional Study of the INTERGROWTH- investigation. Lancet. 2016;388(10047):891-7. https://doi. 21st Project. Lancet. 2014;384(9946):857-68. https://doi. org/10.1016/S0140-6736(16)30902-3 PMID: 27372398 org/10.1016/S0140-6736(14)60932-6 PMID: 25209487 33. Ventura CV, Maia M, Travassos SB, Martins TT, Patriota 16. World Health Organization (WHO). The WHO child growth F, Nunes ME, et al. Risk Factors Associated With the standards. Geneva: WHO; 2016. [Accessed 31 Mar 2016]. Ophthalmoscopic Findings Identified in Infants With Available from: http://www.who.int/childgrowth/en/ Presumed Zika Virus Congenital Infection. JAMA 17. StataCorp. Stata Statistical Software: Release 14. College Ophthalmol. 2016;134(8):912-8. https://doi.org/10.1001/ Station, TX: StataCorp LP; 2015. jamaophthalmol.2016.1784 PMID: 27228275 18. Marinho F, de Araújo VEM, Porto DL, Ferreira HL, Coelho 34. Honein MA, Dawson AL, Petersen EE, Jones AM, Lee EH, Yazdy MRS, Lecca RCR, et al. Microcefalia no Brasil: prevalência e MM, et al. Birth Defects Among Fetuses and Infants of US caracterização dos casos a partir do Sistema de Informações Women With Evidence of Possible Zika Virus Infection During sobre Nascidos Vivos (Sinasc), 2000-2015. [Microcephaly in Pregnancy. JAMA. 2017;317(1):59-68. https://doi.org/10.1001/ Brazil: prevalence and characterization of cases from the Live jama.2016.19006 PMID: 27960197 Birth Information System (Sinasc), 2000-2015]. Epidemiol 35. Dolk H. The predictive value of microcephaly during the e Serviços Saúde. 2016 Oct;25(4):701-12. Portuguese. first year of life for mental retardation at seven years. Available from: http://www.iec.pa.gov.br/template_doi_ess. Dev Med Child Neurol. 1991;33(11):974-83. https://doi. php?doi=10.5123/S1679-49742016000400701&scielo org/10.1111/j.1469-8749.1991.tb14813.x PMID: 1743426 =S2237-96222016000400701 36. van der Linden V, Pessoa A, Dobyns W, Barkovich AJ, Júnior HV, 19. Kleber de Oliveira W, Cortez-Escalante J, De Oliveira WT, do Filho ELR, et al. Description of 13 Infants Born During October Carmo GMI, Henriques CMP, Coelho GE, et al. Increase in 2015-January 2016 With Congenital Zika Virus Infection Without Reported Prevalence of Microcephaly in Infants Born to Women Microcephaly at Birth - Brazil. MMWR Morb Mortal Wkly Living in Areas with Confirmed Zika Virus Transmission During Rep. 2016;65(47):1343-8. https://doi.org/10.15585/mmwr. the First Trimester of Pregnancy - Brazil, 2015. MMWR Morb mm6547e2 PMID: 27906905 Mortal Wkly Rep. 2016;65(9):242-7. https://doi.org/10.15585/ mmwr.mm6509e2 PMID: 26963593 37. Alarcon A, Martinez-Biarge M, Cabañas F, Quero J, García-Alix A. A Prognostic Neonatal Neuroimaging Scale for Symptomatic 20. Cauchemez S, Besnard M, Bompard P, Dub T, Guillemette-Artur Congenital Cytomegalovirus Infection. Neonatology. P, Eyrolle-Guignot D, et al. Association between Zika virus and 2016;110(4):277-85. https://doi.org/10.1159/000446690 microcephaly in French Polynesia, 2013-15: a retrospective PMID: 27344149 study. Lancet. 2016;387(10033):2125-32. https://doi. org/10.1016/S0140-6736(16)00651-6 PMID: 26993883 38. Kwak M, Yum M-S, Yeh H-R, Kim H-J, Ko T-S. Brain Magnetic Resonance Imaging Findings of Congenital Cytomegalovirus 21. de Oliveira WK, de França GVA, Carmo EH, Duncan Infection as a Prognostic Factor for Neurological Outcome. BB, de Souza Kuchenbecker R, Schmidt MI. Infection- Pediatr Neurol. 2018;83:14-8. https://doi.org/10.1016/j. related microcephaly after the 2015 and 2016 Zika virus pediatrneurol.2018.03.008 PMID: 29681488 outbreaks in Brazil: a surveillance-based analysis. Lancet. 2017;390(10097):861-70. https://doi.org/10.1016/S0140- 39. Victora CG, Schuler-Faccini L, Matijasevich A, Ribeiro E, Pessoa 6736(17)31368-5 PMID: 28647172 A, Barros FC. Microcephaly in Brazil: how to interpret reported

52 www.eurosurveillance.org numbers? Lancet. 2016;387(10019):621-4. https://doi. org/10.1016/S0140-6736(16)00273-7 PMID: 26864961 40. Sanz Cortes M, Rivera AM, Yepez M, Guimaraes CV, Diaz Yunes I, Zarutskie A, et al. Clinical assessment and brain findings in a cohort of mothers, fetuses and infants infected with ZIKA virus. Am J Obstet Gynecol. 2018;218(4):440.e1-36. https:// doi.org/10.1016/j.ajog.2018.01.012 PMID: 29353032 41. Levine D, Jani JC, Castro-Aragon I, Cannie M. How Does Imaging of Congenital Zika Compare with Imaging of Other TORCH Infections? Radiology. 2017;285(3):744-61. https://doi. org/10.1148/radiol.2017171238 PMID: 29155634

This article is copyright of the authors or their affiliated institutions, 2018.