Survival in Very Preterm Infants: an International Comparison of 10 National Neonatal Networks
Total Page:16
File Type:pdf, Size:1020Kb
Survival in Very Preterm Infants: Kjell Helenius, MD, a, b Gunnar Sjörs, MD, c Prakesh S. Shah, MD, Msc, d, e Neena Modi, MD, f Brian Reichman, MBChB, g Naho AnMorisaki, MD,International PhD, h Satoshi Kusuda, MD, i Kei Lui, MD, jComparison Brian A. Darlow, MD, k Dirk Bassler, MD, of MSc, l Stellan Håkansson, MD, c Mark Adams, MSc, l Maximo Vento, MD, PhD, m Franca Rusconi, MD, n Tetsuya Isayama, MD, e Shoo K. Lee, MBBS, 10PhD, d, e Liisa National Lehtonen, MD, a, b on behalf Neonatal of the International Network Networks for Evaluating Outcomes (iNeo) of Neonates OBJECTIVES: abstract To compare survival rates and age at death among very preterm infants in 10 METHODS: national and regional neonatal networks. ’ A cohort study of very preterm infants, born between 24 and 29 weeks gestation and weighing <1500 g, admitted to participating neonatal units between 2007 and 2013 in the International Network for Evaluating Outcomes of Neonates. Survival was compared by using standardized ratios (SRs) comparing survival in each network to the survival RESULTS: estimate of the whole population. Network populations differed with respect to rates of cesarean birth, exposure – to antenatal steroids and birth in nontertiary hospitals. Network SRs for survival were – highest in Japan (SR: 1.10; 99% confidence interval: 1.08 1.13) and lowest in Spain (SR: ’ – 0.88; 99% confidence interval: 0.85 0.90). The overall survival differed from 78% to 93% among networks, the difference being highest at 24 weeks gestation (range 35% 84%). – ’ Survival rates increased and differences between networks diminished with increasing gestational age (GA) (range 92% 98% at 29 weeks gestation); yet, relative differences in survival followed a similar pattern at all GAs. The median age at death varied from 4 days CONCLUSIONS: to 13 days across networks. The network ranking of survival rates for very preterm infants remained largely unchanged as GA increased; however, survival rates showed marked variations at lower GAs. The median age at death also varied among networks. These findings warrant further assessment of the representativeness of the study populations, organization of perinatal services, national guidelines, philosophy of care at extreme GAs, and resources used for decision-making. WHAT’S KNOWN ON THIS SUBJECT: Survival rates of aDepartment of Pediatrics and Adolescent Medicine, Turku University Hospital, Kiinamyllynkatu, Turku, Finland; very preterm infants vary among regions. International bDepartment of Clinical Medicine, University of Turku, Turku, Finland; cNational Quality Registry for Neonatal collaborations have been established to compare Care, Department of Pediatrics/Neonatal Services, University Hospital of Umeå, Umeå, Sweden; dDepartment of Pediatrics, University of Toronto, Toronto, Ontario, Canada; eMaternal-Infant Care Research Centre, outcomes, but population-based comparisons have and Department of Pediatrics, Mount Sinai Hospital, Toronto, Ontario, Canada; fUnited Kingdom Neonatal been difficult to accomplish. Collaborative, Neonatal Data Analysis Unit, and Section of Neonatal Medicine, Department of Medicine, Imperial College London, Chelsea and Westminster Hospital, London, United Kingdom; gGertner Institute for Epidemiology WHAT THIS STUDY ADDS: This study shows variations in and Health Policy Research, Sheba Medical Center, Tel Hashomer, Israel; hDepartment of Social Medicine, survival and age at death among very preterm infants National Center for Child Health and Development, Tokyo, Japan iMaternal and Perinatal Center, Tokyo Women’s in 10 national neonatal networks and provides insight Medical University, Tokyo, Japan; jRoyal Hospital for Women, and National Perinatal Epidemiology and Statistics into differences that affect comparisons. The findings k Unit, University of New South Wales, Randwick, Australia; Department of Pediatrics, University of Otago, warrant investigation of the organization of perinatal Christchurch, New Zealand; lDepartment of Neonatology, University Hospital Zurich, University of Zurich, Zurich, Switzerland; mHealth Research Institute La Fe, Avenida Fernando Abril Martorell, Valencia, Spain; and nTIN care in participating networks. Toscane Online, Unit of Epidemiology, Meyer Children’s University Hospital, Florence, Italy and Regional Health Agency of Tuscany, Florence, Italy To cite: Helenius K, Sjörs G, Shah PS, et al. Survival in Very Preterm Infants: An International Comparison of 10 National Neonatal Networks. Pediatrics. 2017;140(6):e20171264 Downloaded from www.aappublications.org/news by guest on September 29, 2021 PEDIATRICS Volume 140, number 6, December 2017:e20171264 ARTICLE Helenius et al 2017 ROUGH GALLEY PROOF Survival in Very Preterm Infants: An https://doi.org/10.1542/peds.2017-1264 December 2017 International Comparison of 10 National Neonatal Networks 6 140 Pediatrics METHODS Preterm birth remains the major Population and Study Design however, a number of neonatal units cause of infant mortality1 in high- have not consented for participation income countries. Studies published in iNeo. Covariates in the last decade have shown ’ This cohort study included infants variability in survival rates and – born at 24 to 29 weeks gestation, outcomes among health care settings 2 12 weighing <1500 g at birth, and as well as within countries. GA was determined by the best registered in participating network Results from the Euro-Peristat – – estimate based on either early databases from 2007 to 2013 project showed that neonatal ultrasound, the timing of the last (2008 2013 for UKNC, 2009 2013 survival among very preterm infants menstrual period, or physical for TuscanNN). The criterion of 1500 g improved by 29% between 2004 examination after birth, in that order. was used because some networks and 2010 in 18 European countries, Antenatal steroid use was defined as limited their data collection to including those with high baseline maternal receipt of antenatal steroids 13 infants with a birth weight <1500 g. z survival. Such comparisons of at any time before birth. Birth weight Stillborn infants and delivery room different health care organizations scores were calculated according to deaths were not included because – worldwide can deliver important birth weight references appropriate all networks do not routinely collect 15 17 benchmarking information and for each country. Mode of data on these infants. Because potentially provide impetus for delivery was recorded as vaginal or of variability in practices and reviewing clinical practices, norms, ’ cesarean. Multiple births included philosophy of care provision at and guidelines. ≥ ’ twins and higher-order multiples. <24 weeks GA, we only included Outcomes The International Network for neonates born at 24 weeks gestation. Infants admitted for the Evaluating Outcomes of Neonates ’ (iNeo) is a nonprofit collaboration first time to neonatal care after 36 The primary outcome was survival among 10 national and regional weeks corrected GA were excluded. to discharge from the hospital for neonatal networks: the Australian Data on infant characteristics infants admitted to neonatal care. and New Zealand Neonatal Network and outcomes were extracted, as 14 Survival to transfer to a step-down (ANZNN), the Canadian Neonatal previously described. Transferred hospital was used when follow-up Network (CNN), the Finnish Medical infants were included only once. was available only until the transfer. Birth Register (FinMBR), the Israel Population coverage of networks We analyzed survival in 1-week GA Neonatal Network (INN), Neonatal deviate from national birth statistics increments. As a secondary outcome, Research Network Japan (NRNJ), ’ for various reasons. All neonatal we compared postnatal age at death the Spanish Neonatal Network units in Finland, New Zealand, and among nonsurvivors. Neonatal (SEN1500), Sweden s National Tuscany participate in their national morbidities among iNeo survivors Quality Registry for Neonatal Care 18 and regional neonatal networks; have been recently published. (SNQ), the Swiss Neonatal Network Data Management however, in Australia, apart from (SwissNeoNet), Tuscany Neonatal all tertiary neonatal units, only a Network (TuscanNN), and the United few lower-level units participate in Kingdom Neonatal Collaborative the ANZNN. In Canada, only tertiary For most networks, defined data (UKNC). The overarching aim of the neonatal units participate in the elements were either collected iNeo collaboration is to compare CNN. In Israel, all neonatal units from patient records by designated outcomes of very preterm infants, participate, but infants with a birth abstractors according to network understand reasons for variations, weight >1500 g are not included. policies or entered directly and, if possible, identify areas for In Japan and Spain, not all neonatal into a central online database improvement in care practices based 14 units participate in the national by participating units. For the on findings. networks. In Sweden, all neonatal UKNC, data were obtained from å By using the iNeo data set, the units presently participate in the the National Neonatal Research objective of this study was to SNQ; however, the Sk ne region is Database managed by the Neonatal compare gestational age (GA)-specific not included in this study because Data Analysis Unit, which contains survival rates for very preterm it joined in 2011. In Switzerland, all a predefined extract from electronic infants in all neonatal networks tertiary and large secondary