ÌÇÐÄvlog

ImportanceÌý Up-to-date estimates of the health outcomes of preterm children are needed for assessing perinatal care, informing parents, making decisions about care, and providing evidence for clinical guidelines.

ObjectivesÌý To determine survival and neonatal morbidity of infants born from 22 through 34 completed weeks’ gestation in France in 2011 and compare these outcomes with a comparable cohort in 1997.

Design, Setting, and ParticipantsÌý The EPIPAGE-2 study is a national, prospective, population-based cohort study conducted in all maternity and neonatal units in France in 2011. A total of 2205 births (stillbirths and live births) and terminations of pregnancy at 22 through 26 weeks’ gestation, 3257 at 27 through 31 weeks, and 1234 at 32 through 34 weeks were studied. Cohort data were collected from January 1 through December 31, 1997, and from March 28 through December 31, 2011. Analyses for 1997 were run for the entire year and then separately for April to December; the rates for survival and morbidities did not differ. Data are therefore presented for the whole year in 1997 and the 8-month and 6-month periods in 2011.

Main Outcomes and MeasuresÌý Survival to discharge and survival without any of the following adverse outcomes: grade III or IV intraventricular hemorrhage, cystic periventricular leukomalacia, severe bronchopulmonary dysplasia, retinopathy of prematurity (stage 3 or higher), or necrotizing enterocolitis (stages 2-3).

ResultsÌý A total of 0.7% of infants born before 24 weeks’ gestation survived to discharge: 31.2% of those born at 24 weeks, 59.1% at 25 weeks, and 75.3% at 26 weeks. Survival rates were 93.6% at 27 through 31 weeks and 98.9% at 32 through 34 weeks. Infants discharged home without severe neonatal morbidity represented 0% at 23 weeks, 11.6% at 24 weeks, 30.0% at 25 weeks, 47.5% at 26 weeks, 81.3% at 27 through 31 weeks, and 96.8% at 32 through 34 weeks. Compared with 1997, the proportion of infants surviving without severe morbidity in 2011 increased by 14.4% (P < .001) at 25 through 29 weeks and 6% (P < .001) at 30 through 31 weeks but did not change appreciably for those born at less than 25 weeks. The rates of antenatal corticosteroid use, induced preterm deliveries, cesarean deliveries, and surfactant use increased significantly in all gestational-age groups, except at 22 through 23 weeks.

Conclusions and RelevanceÌý The substantial improvement in survival in France for newborns born at 25 through 31 weeks’ gestation was accompanied by an important reduction in severe morbidity, but survival remained rare before 25 weeks. Although improvement in survival at extremely low gestational age may be possible, its effect on long-term outcomes requires further studies. The long-term results of the EPIPAGE-2 study will be informative in this regard.

Previous cohort studies^1-7 suggest that survival of infants born before 27 weeks’ gestation has improved during the past 2 decades. However, disability rates remain high at these gestational ages.^8-10 Moreover, countries differ substantially in their organization of care, available resources, national laws, and cultural preferences regarding provision of proactive care. Therefore, to provide the best available information for parents and medical staff to use in making treatment decisions, mortality and morbidity must be monitored, and studies should be conducted in countries with different attitudes toward active care at early gestational ages.

By far, most of the important cohort studies^11-13 in the field have focused exclusively on infants born before 27 weeks’ gestation. However, even though infants born between 27 and 31 weeks are at lower relative risk of adverse outcomes, they represent a much larger proportion of preterm births. Hence, in absolute numbers, they account for most children with deficits.

We present the results of the EPIPAGE-2 (Etude Epidémiologique sur les Petits Ages Gestationnels 2) study, a national cohort of infants born at a gestational age of 22 through 34 weeks in France in 2011. Our objectives were to study survival and survival without severe neonatal morbidities. We also looked at perinatal interventions and compared the outcomes with those of a similar cohort from 1997, the EPIPAGE-1 study.¹³ Our hypothesis was that survival and survival without severe morbidity have improved during the past 15 years in France, except for extremely preterm infants.

Recruitment and data collection occurred only after families had received information and agreed to participate in the study. This study was approved by the National Data Protection Authority (Commission Nationale de l’Informatique et des Libertés) and by the appropriate ethics committees (Consultative Committee on the Treatment of Information on Personal Health Data for Research Purposes and Committee for the Protection of People Participating in Biomedical Research). Participants provided oral informed consent.

EPIPAGE-2 is a national, prospective, population-based study scheduled to follow up preterm children to the age of 12 years. Infants born at 22 through 34 completed weeks’ gestation in France were eligible for inclusion. Only one region, which accounts for 2% of all births in France, did not participate. The study began March 28, 2011. Recruitment took place at birth in all maternity units in the participating regions. The number of infants required according to our sample size calculations¹⁴ was provided by an 8-month recruitment period for births at 22 through 26 weeks, a 6-month period for 27 through 31 weeks, and a 5-week period for 32 through 34 weeks. During recruitment, members of the regional coordinating committees visited all maternity units to ensure the identification of all eligible children.

The births in our study population were defined to comprise live births, stillbirths, and terminations of pregnancy for maternal (severe maternal diseases) or fetal (severe growth restriction and oligohydramnios) reasons other than congenital anomalies. In all, 2381 births were eligible at 22 through 26 weeks, 3478 at 27 through 31 weeks, and 1376 at 32 through 34 weeks, with 176, 221, and 142 parental refusals, respectively. The study thus included 2205 births at 22 through 26 weeks, 3257 at 37 through 31 weeks, and 1234 at 32 through 34 weeks. Total births in the 25 French regions in 2011 (National Institute of Statistics and Economic Studies; ) were used to estimate preterm birth rates, taking into account months of inclusion and differences in recruitment periods according to gestational age at birth.

In each center, one obstetric and one pediatric study coordinator were responsible for data acquisition, validation, and quality control. Data were collected from medical records and obstetric and neonatal staff. Data on stillbirths and terminations of pregnancy were collected at the time of delivery. Data on live-born infants were collected prospectively during hospitalization until discharge or death. Gestational age was defined as the best obstetric estimate combining last menstrual period and ultrasonogram assessment. Extensive data were collected about pregnancy, delivery, and the neonatal period to investigate pregnancy complications, decisions about terminations of pregnancy, the child’s condition at birth, neonatal diseases, organization of care, treatment, and attitudes toward care. Only selected perinatal data were considered for this study: level of care of the institution, antenatal corticosteroid use, vaginal or cesarean delivery, indicated preterm delivery (defined as a birth after induction of labor or cesarean delivery before the onset of labor), and use of surfactants and postnatal corticosteroids. Questionnaires were completed online, with a secure interface that protected the confidentiality and privacy of data and personal information. The EPIPAGE coordination team used a centralized system to monitor and validate inclusion and data collection at the national level.

The primary outcome was infant survival, defined as the number of children discharged home alive. The secondary outcome was survival to discharge without severe neonatal morbidity. Severe neonatal morbidity was defined as any of the following outcomes: severe intraventricular hemorrhage (IVH), defined as IVH associated with ventricular dilatation (grade III IVH) and intraparenchymal hemorrhage (ie, large unilateral parenchymal hyperdensity or a large unilateral porencephalic cyst)¹⁵; cystic periventricular leukomalacia (cPVL) (ie, periventricular white matter echolucencies at ultrasonography); stages II and III necrotizing enterocolitis, according to the staging of Bell et al¹⁶; stage 3 or higher retinopathy of prematurity, according to the international classification¹⁷ and/or laser treatment; and/or severe bronchopulmonary dysplasia (BPD), defined as administration of oxygen for at least 28 days plus need for 30% or more oxygen and/or mechanical ventilatory support or continuous positive airway pressure at 36 weeks’ postmenstrual age.¹⁸ Among infants admitted to the neonatal intensive care unit (NICU), 98.0% of survivors born before 33 weeks, 88.0% of those born at 33 weeks, and 66.0% of those born at 34 weeks had at least one cranial ultrasonogram assessment. Survivors at 33 and 34 weeks without cranial ultrasonogram assessment were considered not to have severe cerebral lesions, and those with no funduscopic examination were considered not to have severe retinopathy of prematurity.

In 1997, the EPIPAGE-1 study, a comparable prospective, population-based cohort study, took place in 9 regions of France.^13,19 Eligible infants for this comparison are those born alive between 22 and 34 weeks’ gestation from the 1997 and 2011 cohorts in the same 9 regions. Outcome measures are survival to discharge and survival without neonatal morbidity. The latter was defined as above, except for BPD (oxygen supplementation at 36 weeks’ postmenstrual age) because information on its severity was not available in 1997. Cohort data were collected from January 1 through December 31, 1997, and from March 28 through December 31, 2011.¹⁴ Analyses for 1997 were run for the entire year and then separately for April to December; the rates for survival and morbidities did not differ. Data are therefore presented for the whole year in 1997 and the 8-month and 6-month periods in 2011.

Data reported on the birth certificate (ie, gestational age at birth, vital status at birth, and neonatal death) were available and usable without parental consent for children not included in the study. We compared the survival of the participating and nonparticipating (because of parent refusal) newborns. Then, for the infants whose parents agreed to participate, we analyzed survival to discharge, severe neonatal morbidity, survival without severe morbidity, and some obstetric and neonatal interventions, according to gestational age. To examine trends over time, we compared survival and neonatal morbidity in the EPIPAGE-1 study¹⁹ with those in the EPIPAGE-2 study according to gestational age. For each week of gestation, we report exact 95% binomial CIs of survival rates and their differences. All tests were 2-sided; P < .05 was considered statistically significant. All statistical analyses were performed with SAS statistical software, version 9.3 (SAS Institute Inc).

In 2011, preterm birth rates were 4.4 per 1000 total stillbirths and live births and 2.1 per 1000 live births before 27 weeks’ gestation, 8.4 and 7.5 at 27 through 31 weeks, and 17.8 and 17.3 at 32 through 34 weeks, respectively.

Participation rates were 92.6% among infants born at 22 through 26 weeks, 93.6% among those at 27 through 31 weeks, and 89.7% among those at 32 through 34 weeks. In each gestational age group, the proportion of live births was slightly higher among participants than nonparticipants. Survival among live births did not differ significantly between the 2 groups (eTable 1 in the Supplement).

The proportion of live-born infants increased with gestational age from 13.5% at 22 weeks to 98.5% at 34 weeks (Table 1). Only one infant born at 22 through 23 weeks (ie, 0.1% of all births and 0.7% of live births) survived to discharge (Table 1). Survival rates were 14.4% of all births and 31.2% of live births at 24 weeks, 41.8% and 59.1% at 25 weeks, 59.6% and 75.3% at 26 weeks, 73.7% and 86.3% at 27 through 28 weeks, 88.0% and 96.6% at 29 through 31 weeks, and 96.7% and 98.9% at 32 through 34 weeks, respectively (Table 1). Among infants who died, the proportion whose deaths followed a decision to limit intensive care varied from 80.9% at 22 through 24 weeks and 70.3% at 25 through 26 weeks to 57.0% at 27 through 31 weeks. The median age at death was the day of birth for infants born at 22 through 24 weeks, 5 days (interquartile range, 1-16 days) for those born at 25 through 26 weeks, and 7 days (interquartile range, 1-22 days) for those born at 27 through 31 weeks.

Among the live-born infants at 22 weeks, 36.2% were born in level III hospitals compared with 61.8% at 23 weeks, 77.4% at 24 weeks, 85.0% at 25 through 26 weeks, 84.8% at 27 through 31 weeks, and 50.1% at 32 through 34 weeks (Table 2). Among extremely preterm infants not born in a level III hospital, 54.3% were postnatally transferred to a NICU; this proportion varied from 4.3% at 22 through 23 weeks to 45.3% at 24 weeks and 90.7% at 25 through 26 weeks. Postnatal transfers to a NICU reached 91.8% at 27 through 29 weeks but decreased to 56.3% at 30 through 31 weeks and 17.0% at 32 through 34 weeks. The percentage of infants exposed to antenatal corticosteroids was very low at 22 (1.8%) and 23 (12.3%) weeks but increased to 56.7% at 24 weeks and 78.4% at 25 through 26 weeks (Table 2). Cesarean rates were 6.3% at 22 through 23 weeks and 13.5% at 24 weeks compared with 34.0% at 25 weeks and 59.9% at 26 weeks; this rate reached 69.8% at 27 through 31 weeks. Few of the infants born at 22 through 23 weeks were admitted to NICUs (6.1%); this percentage increased to 60.8% at 24 weeks, 91.9% at 25 weeks, 95.6% at 26 weeks, and 98.9% at 27 through 31 weeks. Among infants admitted to the NICU, 96.7% of those born at 24 through 26 weeks received surfactant and 24.0% received postnatal corticosteroids.

Of survivors at 24 through 26 weeks, 12.9% had severe IVH, 2.4% had cPVL, 25.6% had severe BPD, 6.0% had retinopathy of prematurity stage 3 or higher, and 5.1% had stage 2 or 3 necrotizing enterocolitis (Table 3). In this group, 299 infants (59.2% of survivors and 34.1% of live births) were discharged home without severe neonatal morbidity. The percentage of such survivors ranged from 0% at 23 weeks to 65.3% (47.5% of live births) at 26 weeks (Figure 1). Of the 206 infants who survived with severe neonatal conditions, 23.8% had 2 or more conditions. Of survivors at 27 through 31 weeks, 2254 (87.6%; 81.3% of live births) were discharged home without severe neonatal morbidity; the percentage of survivors ranged from 71.9% (57.6% of live births) at 27 weeks to 93.5% (90.6% of live births) at 31 weeks (Figure 1). Among the 320 infants with severe neonatal morbidities, 8.4% had 2 or more morbidities. At 32 through 34 weeks, 1080 infants (97.9% of survivors and 96.8% of live births) were discharged home without severe neonatal morbidity. One infant had 2 severe conditions.

Among infants born alive at 22 through 23 weeks in the 9 regions studied in 1997, none survived in 1997 or 2011, and the chance of survival at 24 weeks did not change between the studies (Figure 2A and eTable 2 in the Supplement). Survival increased in these regions by 11.2% (95% CI, −0.5% to 22.9%) at 25 weeks, 18.1% (95% CI, 8.2% to 28.1%) at 26 weeks, 12.8% (95% CI, 4.8% to 20.8%) at 27 weeks, 12.3% (95% CI, 6.1% to 18.6%) at 28 weeks, 7.1% (95% CI, 2.7% to 11.5%) at 29 weeks, 4.7% (95% CI, 1.4% to 8.0%) at 30 weeks, and 2.1% (95% CI, −0.2% to 4.4%) at 31 weeks. Although median age at death did not change at 22 through 24 weeks, it increased significantly at 25 through 26 weeks. Between 1997 through 2011, the rates of antenatal corticosteroid use, indicated preterm deliveries, and surfactant use increased significantly in all gestational-age groups, except at 22 through 23 weeks (Figure 2C-E and eTable 3 in the Supplement).

Survival without neonatal morbidity did not change significantly at 24 weeks between 1997 (2.4%) and 2011 (7.4%) (Figure 2B). It increased by 16.2% (95% CI, 6.7% to 25.8%) at 25 weeks, 19.0% (95% CI, 9.1% to 28.8%) at 26 weeks, 16.3% (95% CI, 6.4% to 26.2%) at 27 weeks, 17.8% (95% CI, 9.2% to 26.5%) at 28 weeks, 16.6% (95% CI, 8.7% to 24.5%) at 29 weeks, 6.3% (95% CI, 0.7% to 11.8%) at 30 weeks, and 5.9% (95% CI, 1.8% to 10.1%) at 31 weeks. Among survivors at 24 through 26 weeks, the rates of necrotizing enterocolitis (P = .005), BPD (P = .004), cPVL, and severe retinopathy of prematurity decreased between 1997 and 2011, although not significantly for the cPVL (P = .07) and severe retinopathy of prematurity (P = .11) (eTable 4 in the Supplement). At 27 through 31 weeks, the prevalence of cPVL decreased by 3% (P < .001) and BPD by 4% (P < .001). Only cPVL decreased among infants born at 32 through 34 weeks (P = .03) (eTable 4 in the Supplement).

The results of the EPIPAGE-2 study, a national, prospective, population-based cohort study of births at 22 through 34 weeks’ gestation, indicate that survival and survival without severe neonatal morbidity improved significantly between 1997 and 2011 for infants born at 25 through 31 weeks. By contrast, neither survival nor survival without morbidity improved for infants born before 25 weeks.

The strengths of the EPIPAGE-2 study include the population-based cohort design and prospective enrollment of infants born prematurely in France in 2011. Standardized definitions of outcomes and systematic and prospective collection of all information available (eg, all cranial ultrasonograms) from a national sample of more than 8000 preterm births (22-34 weeks’ gestation) allowed us to look at the effects associated with a wide range of gestational ages on survival and on major neonatal morbidities in our population. The accuracy of the gestational age estimates was improved by the very high rate (>98%) of women with early ultrasonogram assessments.

One limitation is that 7% of eligible infants were not included because of parental refusal. However, the survival status of all patients, including those who refused to participate, was available. Furthermore, the percentage of survival in these 2 groups did not differ significantly. Therefore, the effect of this selection was very slight.

These 2 EPIPAGE studies^13,19 made it possible to determine the changes in mortality and morbidity between 1997 through 2011. We studied neonatal conditions known to be prognostic for long-term outcomes. Although the studies had a common design, more extensive data were collected in 2011 than in 1997. Hence, we may have underestimated changes between the 2 periods for survival without morbidity and morbidity rates in general. However, because we restricted our comparisons to severe neonatal conditions, defined similarly in each study, we assume that the influence of this difference was slight.

One important result of our study is that less than 1% of infants born at 22 through 23 weeks survived. In this population, 80.9% of deaths occurred after a decision to limit intensive care, mostly within the first day of life. The general policy in France is not to intervene before 24 weeks’ gestation; infants born earlier receive palliative but not intensive care.²⁰ We compared French results with those of large contemporary international studies^1,2,4,6,7,21 conducted in the middle to late 2000s (eTable 5 in the Supplement). The more active perinatal management at the limit of viability in other countries has resulted in higher survival rates than those in our population at extremely preterm gestational ages.^1,2,4,6,7 At 24 weeks, survival remained low in France, reflecting the lack of consensus and heterogeneity of perinatal management for these infants. In this group, as among those born at 22 through 23 weeks, deaths occurred within a day of birth after a decision to limit intensive care. This timing contrasts with the timing of death in those countries that report high rates of perinatal interventions and survival.^1,2,6,7 Active perinatal interventions and survival became more frequent in France at 25 weeks, but survival rates remained higher in the United States, Japan, and Sweden up to a gestational age of 27 weeks.^5,6

International comparisons emphasize that the potential for survival among extremely preterm infants is 10% to 50% higher than our results. They also suggest that active management of extremely preterm infants can improve survival for those born at higher gestational ages. In France, the extension of withholding care to less premature infants, because of fears about immediate and long-term adverse outcomes, might also explain our results at 25 through 27 weeks. However, results of comparisons such as those noted above should be interpreted with caution because differences in gestational age measurement and in the distinction between stillbirths and live births cannot be excluded.²² One way to clarify the role of these issues would be to design multinational cohort studies with standardized methods. In addition, meta-analysis of outcomes using patient-level data might allow better assessment of country-level differences in outcomes.

There is a widespread consensus that the aim of neonatal care should be to resuscitate infants with a reasonable likelihood of an acceptable quality of life, but identification of strategies for better outcomes remains difficult. Uncertainty about long-term outcomes at the limit of viability influences treatment decisions at extremely low gestational ages in France. The results of previous studies^6,23-25 of trends in short-term morbidity and longer-term outcomes of infants born at gestational ages close to this limit make it difficult to predict the effect of a more proactive management of these infants on their survival without morbidity. Hence, consideration of this potential effect must examine the possible and problematic nature of the trade-off between improved survival and increased risk of severe long-term adverse health outcomes for infants born before 25 weeks.

Results of our trend study during a 15-year period (1997-2011) reveal that survival without morbidity increased by 14.1% for infants born at 25 through 29 weeks. This finding indicates that 1 of every 7 infants had a more favorable outcome in 2011 compared with 1997. Hence, the total number of children surviving without short-term and perhaps also long-term severe adverse outcomes has increased over time.

Few other population-based studies from around the world provide up-to-date estimates of short-term prognosis of extremely, very, and moderately preterm infants and of changes during the past decade. International comparisons help to estimate the potential for survival and to identify appropriate interventions; they thus reveal areas for improvement in each country. In particular, they reveal that improvement in survival at extremely low gestational age is possible in France and in countries with similar practices. This finding should encourage health care professionals to reassess their attitudes toward care at extremely low gestational ages. This reassessment should include a complete analysis of neonatal morbidity and long-term sequelae, which have not yet been sufficiently evaluated, although they remain important factors in decision making. EPIPAGE-2 should provide further information on them as the children it studies age. Finally, specificities in the organization of care, health policies, laws, and available resources of each country must also be part of this discussion.

Accepted for Publication: November 18, 2014.

Corresponding Author: Pierre-Yves Ancel, PhD, Obstetrical, Perinatal, and Pediatric Epidemiology Team, Epidemiology and Biostatistics Sorbonne Paris Cité Research Center (U1153), INSERM, 53 avenue de l’Observatoire, 75014 Paris, France (pierre-yves.ancel@inserm.fr).

Published Online: January 26, 2015. doi:10.1001/jamapediatrics.2014.3351.

The EPIPAGE-2 Writing Group includes Pierre Kuhn, PhD; Bruno Langer, MD; Jacqueline Matis, MD; Xavier Hernandorena, MD; Pierre Chabanier; Laurence Joly-Pedespan, MD; Bénédicte Lecomte, MD; Françoise Vendittelli, PhD; Michel Dreyfus, MD; Bernard Guillois, MD; Antoine Burguet, PhD; Pierre Sagot, MD; Jacques Sizun, MD; Alain Beuchée, MD; Florence Rouget, MD; Amélie Favreau, MD; Elie Saliba, PhD; Nathalie Bednarek, PhD; Patrice Morville, MD; Gérard Thiriez, PhD; Loïc Marpeau, MD; Stéphane Marret, PhD; Gilles Kayem, PhD; Xavier Durrmeyer, MD; Michèle Granier, MD; Olivier Baud, PhD; Pierre-Henri Jarreau, PhD; Delphine Mitanchez, PhD; Pascal Boileau, PhD; Pierre Boulot, MD; Gilles Cambonie, PhD; Hubert Daudé, MD; Antoine Bédu, PhD; Fabienne Mons, PhD; Jeanne Fresson, PhD; Rachel Vieux, PhD; Corine Alberge, MD; Catherine Arnaud, PhD; Christophe Vayssière, MD; Patrick Truffert, PhD; Véronique Pierrat, PhD; Damien Subtil, PhD; Claude D’Ercole, MD; Catherine Gire, MD; Umberto Simeoni, MD; André Bongain, PhD; Loïc Sentilhes, PhD; Jean-Christophe Rozé, PhD; Jean Gondry, MD; André Leke, PhD; Michel Deiber, MD; Olivier Claris, PhD; Jean-Charles Picaud, PhD; Anne Ego, PhD; Thierry Debillon, PhD; Anne Poulichet, MD; Eliane Coliné, MD; Anne Favre, MD; Olivier Fléchelles, MSc; Sylvain Samperiz, MD; Duksha Ramful, MD; Bernard Branger; Valérie Benhammou, PhD; Laurence Foix-L’Hélias, PhD; Laetitia Marchand-Martin, MSc; Monique Kaminski, MSc.

Affiliations of The EPIPAGE-2 Writing Group: University Hospital, Strasbourg, France (Kuhn, Langer, Matis); La Côte Basque Hospital, Bayonne, France (Hernandorena); University Hospital, Bordeaux, France (Chabanier, Joly-Pedespan); University Hospital Estaing, Clermont-Ferrand, France (Lecomte, Vendittelli); Department of Gynecology and Obstetrics, University Hospital, Caen, France (Dreyfus); Department of Neonatal Pediatrics and Intensive Care, University Hospital, Caen, France (Guillois); Department of Neonatal Pediatrics, University Hospital, Dijon, France (Burguet); Department of Gynecology and Obstetrics, University Hospital, Dijon, France (Sagot); University Hospital, Brest, France (Sizun); Department of Pediatrics, University Hospital, Inserm-Irset U 1085, Rennes, France (Beuchée, Rouget); Department of Neonatal Pediatrics and Intensive Care, University Hospital, Tours, France (Favreau); INSERM U 930, François Rabelais University, Tours, France (Saliba); Department of Neonatal Pediatrics, University Hospital, Reims, France (Bednarek, Morville); Department of Neonatal Pediatrics, University Hospital, Besançon, France (Thiriez); Department of Gynecology and Obstetrics, University Hospital, Rouen, France (Marpeau); Department of Neonatal Pediatrics and Intensive Care, Rouen University Hospital-Laboratory of microvascular endothelium and neonatal brain lesions, Rouen, France (Marret); Department of Obstetrics and Gynecology, Louis Mourier Hospital, University Hospitals Paris Nord Val de Seine (HUPNVS)), Assistance Publique-Paris Hospitals (APHP), Paris Diderot University, Paris, France (Kayem); Department of Neonatal Pediatrics and Intensive Care, CHI, CRC, Créteil, France (Durrmeyer); Department of Neonatal Pediatrics, Sud Francilien Hospital, Evry, France (Granier); Neonatal intensive care unit, Robert Debré Hospital, INSERM, UMR 676, Paris, France (Baud); Department of Neonatal Pediatrics and Intensive Care, Cochin Hotel Dieu Hospital, Paris, France (Jarreau); Department of Neonatal Pediatrics, Trousseau Hospital, Paris, France (Mitanchez); Department of Neonatal Pediatrics, Poissy Saint Germain University Hospital, Poissy, France (Boileau); Department of Obstetrics and Gynecology, Arnaud de Villeneuve Hospital, Montpellier, France (Boulot); Department of Neonatal Pediatrics and Intensive Care, Arnaud de Villeneuve Hospital, Montpellier, France (Cambonie); CAMSP, University Hospital, Montpellier, France (Daudé); Department of Neonatal Pediatrics, Mère-Enfant Hospital, Limoges, France (Bédu, Mons); Department of Medical Information, Adolphe Pinard Maternity Unit, Nancy, France (Fresson); Department of Neonatal Pediatrics and Intensive Care, Adolphe Pinard Maternity Unit, Nancy, France (Vieux); UMR 1027 INSERM, Paul-Sabatier Toulouse III University, Toulouse, France (Alberge, Arnaud); Department of Obstetrics and Gynecology, Toulouse, France (Vayssière); Department of Neonatal Pediatrics, Jeanne de Flandres Hospital, Lille, France (Truffert, Pierrat); Department of Gynecology and Obstetrics, Jeanne de Flandre Hospital, Lille, France (Subtil); Department of Gynecology and Obstetrics, Nord Hospital, Marseille, France (D’Ercole); Department of Neonatal Pediatrics and Intensive Care, Nord Hospital, Marseille, France (Gire); Department of Neonatal Pediatrics and Intensive Care, La Conception Hospital, Marseille, France (Simeoni); Department of Gynecology and Obstetrics, Archet Hospital, Nice, France (Bongain); Department of Obstetrics and Gynecology, Angers University Hospital, Angers, France (Sentilhes); Department of Neonatal Medicine, Angers University Hospital and INSERM CIC 004, Nantes, France (Rozé); Department of Obstetrics and gynecology, Amiens, France (Gondry); Department of Pediatrics, Amiens, France (Leke); Department of Pediatrics, Chambéry, France (Deiber); Department of Neonatal Pediatrics and Intensive Care, University Hospital, Lyon, France (Claris); Department of Neonatal Pediatrics and Intensive Care, La Croix Rousse Hospital, Lyon, France (Picaud); INSERM CIC003, University Hospital, Grenoble, France (Ego); Department of Neonatal Pediatrics, University Hospital, Grenoble, France (Debillon); University Hospital, Pointe à Pitre, Guadeloupe, France (Poulichet, Coliné); Department of Neonatal Pediatrics and Intensive Care, Cayenne Hospital, Cayenne, Guyane, France (Favre); University Hospital, Fort de France, Martinique (Fléchelles); Department of Neonatal Pediatrics and Intensive Care, University Hospital Felix Guyon, Saint-Denis, La Réunion, France (Samperiz, Ramful); Fédération des Réseaux de Santé en Périnatalité [FFRSP], Nantes, France (Branger); Inserm UMR1153, Perinatal and Pediatric Epidemiology Team, Paris, France (Benhammou, Foix-L’Hélias, Marchand-Martin, Kaminski).

Author Contributions: Dr Ancel had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: All authors.

Obtained funding: Ancel.

Study supervision: All authors.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the French Institute of Public Health Research/Institute of Public Health and its partners the French Health Ministry, the National Institute of Health and Medical Research, the National Institute of Cancer, and the National Solidarity Fund for Autonomy; grant ANR-11-EQPX-0038 from the National Research Agency through the French Equipex Program of Investments in the Future; and the PremUp Foundation.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.

Additional Contributions: We are grateful for the participation of all families of preterm infants in the EPIPAGE-2 cohort study and for the cooperation of all maternity and neonatal units in France.

The EPIPAGE-2 Study Group Collaborators include Alsace: D. Astruc, P. Kuhn, B. Langer, J. Matis (Strasbourg), C. Ramousset; Aquitaine: X. Hernandorena (Bayonne), P. Chabanier, L. Joly-Pedespan (Bordeaux), M. J. Costedoat, A. Leguen; Auvergne: B. Lecomte, D. Lemery, F. Vendittelli (Clermont-Ferrand); Basse-Normandie: G. Beucher, M. Dreyfus, B. Guillois (Caen), Y. Toure; Bourgogne: A. Burguet, S. Couvreur, J. B. Gouyon, P. Sagot (Dijon), N. Colas; Bretagne: J. Sizun (Brest), A. Beuchée, P. Pladys, F. Rouget (Rennes), R. P. Dupuy (St-Brieuc), D. Soupre (Vannes), F. Charlot, S. Roudaut; Centre: A. Favreau, E. Saliba (Tours), S. Leclercq; Champagne-Ardenne: N. Bednarek, P. Morville (Reims), M. Palot; Franche-Comté: G. Thiriez (Besançon), C. Balamou; Haute-Normandie: L. Marpeau, S. Marret (Rouen), C. Barbier RM; Ile-de-France: G. Kayem (Colombes), X. Durrmeyer (Créteil), M. Granier (Evry), M. Ayoubi, A. Baud, B. Carbonne, L. Foix L’Hélias, F. Goffinet, P. H. Jarreau, D. Mitanchez (Paris), P. Boileau (Poissy), C. Duffaut, E. Lorthe; Languedoc-Roussillon: P. Boulot, G. Cambonie, H. Daudé (Montpellier), A. Badessi, N. Tsaoussis; Limousin: A. Bédu, F. Mons (Limoges), C. Bahans; Lorraine: M. H. Binet, J. Fresson, J. M. Hascoët, A. Milton, O. Morel, R. Vieux (Nancy), L. Hilpert; Midi-Pyrénées: C. Alberge, C. Arnaud, C. Vayssière (Toulouse), M. Baron; Nord-Pas-de-Calais: M. L. Charkaluk, V. Pierrat, D. Subtil, P. Truffert (Lille), C. Delaeter; PACA et Corse: C. D’Ercole, C. Gire, U. Simeoni (Marseille), A. Bongain (Nice), M. Deschamps, C. Grangier; Pays de Loire: B. Branger (FFRSP), J. C. Rozé, N. Winer (Nantes), V. Rouger, C. Dupont; Picardie: J. Gondry, G. Krim (Amiens), B. Baby; Rhône-Alpes: M. Debeir (Chambéry), O. Claris, J. C. Picaud, S. Rubio-Gurung (Lyon), C. Cans, A. Ego, T. Debillon (Grenoble), H. Patural (Saint-Etienne), A. Rannaud; Guadeloupe: E. Janky, A. Poulichet, J. M. Rosenthal (Point à Pitre), E. Coliné; Guyane: A. Favre (Cayenne), N. Joly; Martinique: S. Châlons (Fort de France), V. Lochelongue; La Réunion: P. Y. Robillard (Saint-Pierre), S. Samperiz, D. Ramful (Saint-Denis); Inserm UMR S953: P. Y. Ancel, V. Benhammou, B. Blondel, M. Bonet, A. Brinis, M. L. Charkaluk, M. Durox, L. Foix-L’Hélias, F. Goffinet, M. Kaminski, G. Kayem, B. Khoshnood, C. Lebeaux, L. Marchand-Martin, V. Pierrat, M. J. Saurel-Cubizolles, D. Tran, L. Vasante-Annamale, J. Zeitlin.

Correction: This article was corrected on March 9, 2015, to fix errors in Group Information and Figure 2.