- Email: [email protected]
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes Thibault Mura, MD, PhD1,2, Jean-Charles Picaud, MD, PhD3, Beatrice Larroque, MD, PhD4,5, Florence Galtier, MD1,2, Stephane Marret, MD, PhD6, Jean-Christophe Roze, MD, PhD7, Patrick Truffert, MD, PhD8, Pierre Kuhn, MD, PhD9, Jeanne Fresson, MD, PhD10, Gerard Thiriez, MD, PhD11, Catherine Arnaud, MD, PhD12,13, Gregoire Mercier, MD, PhD1, Marie-Christine Picot, MD, PhD1,2, Pierre-Yves Ancel, MD, PhD4,14, and Bernard Ledesert, MD15, on behalf of the Etude Epidemiologique sur les Petits Ages Gestationnels (EPIPAGE) Study Group* Objective To evaluate the relationship between preterm premature rupture of membranes (PPROM) and cognitive impairment in 5-year-old children born very preterm.
Study design The Etude Epidemiologique sur les Petits Ages Gestationnels Study is a population-based cohort of children followed up from birth to age 5 years recruited in 9 French regions in 1997. We analyzed data from singletons born between 24 and 32 weeks gestation categorized into 4 groups according to etiology of prematurity: infants born after PPROM, after idiopathic preterm labor, in a vascular context (Vasc), and to women with other complications (Other). Cognitive development at age 5 years was assessed using the Mental Processing Composite score of the Kaufman-Assessment Battery for Children. Results Among the 1051 children followed up to age 5 years, the mean Mental Processing Composite score was 93.6 19.7, and 13.3% of the children (140 of 1051) had cognitive impairment. After adjustment for potential confounders, the risk of cognitive impairment among infants in the PPROM group was not significantly different than that in the idiopathic preterm labor group (OR, 1.09; 95% CI, 0.62-1.92) and the Other group (OR, 1.36; 95% CI, 0.75-2.47), but was lower than that in the Vasc group (OR, 1.86; 95% CI, 1.16-2.97). In the PPROM group, the risk of cognitive impairment was greater when the latency period (ie, time from rupture to delivery) was <3 days (OR, 2.32; 95% CI, 1.07-5.02). Conclusion Preterm infants born after PPROM are not at increased risk for cognitive impairment in childhood, but the time between PPROM and birth may influence that risk. (J Pediatr 2013;163:435-40).
I
n Europe, between 1.1% and 1.6% of liveborn infants are born very preterm (ie, before 33 weeks gestational age).1 The improved survival of very preterm infants has been associated with an increased rate of neuromotor and cognitive development abnormalities in these survivors.2,3 With a prevalence of >10% in this population, cognitive impairment is one of the most common sequelae associated with very preterm birth.4 These cognitive sequelae often result from white matter lesions that go unidentified in the perinatal period.5 These lesions may be attributable to hemodynamic causes, and also may be the result of infectious6 or inflammatory processes, as is observed in 50% of births that occur after preterm premature rupture of membranes (PPROM).7 Infants born after PPROM also may be at increased risk for intraventricular hemorrhage and periventricular leukomalacia.8-10 Although PPROM accounts for one-third of preterm births,11 studies on the long-term outcomes of these infants are rare and have reported conflicting results. Of note is a possible increased risk of cerebral palsy12-14; however, this link has yet to be systematically verified.15,16 Only 1 study reported an increased risk of cognitive impairment in very preterm infants born after PPROM17; in that study, this risk was higher in neonates with a latency period (ie, time from
EPIPAGE IPL K-ABC MPC PPROM SGA Vasc
miologique sur les Petits Ages Gestationnels Etude Epide Idiopathic preterm labor Kaufman Assessment Battery for Children Mental Processing Composite Preterm premature rupture of membranes Small for gestational age Vascular
From the 1Clinical Investigation Center and Information Medical Department, University Hospital of Montpellier; 2 et de la Recherche Institut National de la Sante dicale, Montpellier, France; 3Department of Me Neonatology, Croix Rousse Hospital, Claude Bernard University Lyon 1, Lyon, France; 4Pierre and Marie Curie University, Paris, France; 5Epidemiological Research Unit on Perinatal Health and Women’s and Children’s et de la Recherche Health, Institut National de la Sante dicale, Villejuif, France; 6Department of Neonatal Me Medicine, Rouen University Hospital and the Institut et de la Recherche Me dicale Avenir National de la Sante Research Group, Institute for Biomedical Research, University of Rouen, Rouen, France; 7Department of Neonatology, Children’s Hospital, Nantes, France; 8 Department of Neonatology, Jeanne de Flandre Hospital, Lille, France; 9Department of Pediatrics II, Hautepierre Hospital, Strasbourg, France; 10Information Medical Department, Regional Maternity University Hospital, Nancy, France; 11Department of Pediatrics, Besanc¸on University Hospital, Besanc¸on, France; 12 et de la Recherche Institut National de la Sante dicale; 13Clinical Epidemiology Unit, Toulouse Me University Hospital, Toulouse, France; 14Institut National et de la Recherche Me dicale, de la Sante Epidemiological Research Unit on Perinatal Health and Women’s and Children’s Health, Tenon Hospital, Paris, gional de Sante , France; and 15Observatoire Re Montpellier, France *A list of EPIPAGE Study Group members is available at www.jpeds.com (Appendix). Supported by INSERM (National Institute of Health and Medical Research), Merck Sharp, Dohme-Chibret, la dicale (The Medical ReFondation de la Recherche Me nerale de la search Foundation), and la Direction Ge du Ministe re des Affaires Sociales (French Ministry Sante of Health). The authors declare no conflicts of interest. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2013.01.039
435
THE JOURNAL OF PEDIATRICS
www.jpeds.com
rupture to delivery) exceeding 48 hours compared with those with a shorter latency period. In a cohort of very preterm infants in the Etude Epidemiologique sur les Petits Ages Gestationnels (EPIPAGE) population, we investigated whether those born after PPROM were at greater risk for cognitive impairment at age 5 years. We also investigated whether the latency period between membrane rupture and birth influenced later cognitive outcome.
Methods All parents of infants born at 24-32 weeks gestational age between January 1, 1997, and December 31, 1997, in 9 French regions (representing more than one-third of the country) were invited to participate in the EPIPAGE study.18 The children in the study cohort were followed up at age 2 month, at age 9 months, and then every year up to age 5 years. Out of the 2855 liveborn neonates included in the EPIPAGE cohort, 25 had no available information concerning the etiology of prematurity (Figure; available at www.jpeds.com) and were excluded from this analysis. We also excluded 885 neonates from multiple pregnancies (to avoid a confounding bias) and 9 with serious malformations. Owing to the number of preterm neonates born at 32 weeks gestational age, all regions were given the option of including only 1 of every 2 infants born at exactly 32 weeks in the follow-up; 2 regions chose this option (52 infants). In these regions, children whose mother was born on an odd-numbered day were selected. Finally, a total of 1884 children met our inclusion criteria and were eligible for analysis. This study was approved by the French data protection agency (Commission Nationale de l’Informatique et des Libertes). Maternal, obstetric, and neonatal data were collected at birth following a standardized protocol. All children were invited for a checkup at age 5 years. The parents were contacted in writing, and if no response was received, a system of followup contacts by phone or mail was established. The checkup involved medical, social, and psychological assessments, including a cognitive evaluation performed by a psychologist using the Kaufman Assessment Battery for Children (K-ABC).19 We divided the infants into 4 mutually exclusive groups according to the pregnancy complications that led to preterm delivery.11 Infants born after PPROM without other pregnancy complications (PPROM group) were compared with infants born after idiopathic preterm labor (IPL group), defined as spontaneous onset of labor before rupture of membranes; infants born in a vascular context (Vasc group) to mothers with hypertension, defined as systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg during pregnancy, and small for gestational age (SGA) infants, defined as a birth weight <10th percentile for gestational age of the liveborn infants in our sample (with these 2 groups aggregated because intrauterine growth restriction in preterm birth frequently has a vascular cause); and infants born to mothers with other complications (Other group) resulting in preterm delivery, including antepartum hemorrhage without SGA status or maternal hypertension. 436
Vol. 163, No. 2 The K-ABC has been validated in France for use in children aged 2.5-12.5 years.19 The K-ABC Mental Processing Composite (MPC) score, considered equivalent to IQ score, is a global measure of cognitive ability. This score was standardized to a mean of 100 (SD 15) using published French standards. In accordance with the World Health Organization’s definition, cognitive impairment was defined as an MPC score <70 (difference of >2 SD).20 The following data were included in the analysis: sociodemographic data (ie, sex, maternal nationality, maternal educational level, and family socioeconomic status), antepartal and peripartal data (ie, gestational age, expressed as completed weeks of amenorrhea, antenatal corticosteroids, and mode of delivery), and neonatal data (ie, grade 3-4 intraventricular hemorrhage, periventricular leukomalacia, maternal-fetal infection, ulcero-necrotizing enterocolitis, bronchopulmonary dysplasia, and postnatal corticosteroid use). The c2 test was used to analyze sociodemographic, antenatal, peripartal, and neonatal characteristics associated with the pregnancy complications leading to preterm delivery in the 4 study groups. ANOVA was used to compare mean KABC MPC scores. Logistic regression was used to evaluate the relationships between each of the pregnancy complication groups and cognitive impairment at age 5 years. We initially examined these relationships after adjustment for gestational age at birth (treated categorically using gestational age groups, as identified in Table I) and for the sociodemographic covariates deemed potentially important on univariate analysis (P < .20 for the association with pregnancy complication groups). Later, we made adjustments for antenatal, peripartal, and neonatal covariates meeting the same criteria (P < .20 for the associations with pregnancy complication groups). We considered these additional factors separately because they might be involved in the causal relationship between pregnancy complications and cognitive impairment, and thus could act as intermediate factors. Adjusting for their effect helped us explain the associations between pregnancy complication groups and cognitive impairment. We addressed the risk of selection bias potentially induced by missing values by carrying out a sensitivity analysis using data imputation on both outcome variable and covariates. We used the multiple imputation by Markov chain Monte Carlo method21 with the MI and MIanalyze procedures implemented in SAS version 9 (SAS Institute, Cary, North Carolina).22 The imputed dataset was generated by performing 50 imputation cycles; 9.6% of the data were imputed. To examine the influence of the latency period on cognitive outcome, we divided the population of neonates born after PPROM into 2 groups according to the interval between PPROM and birth. The threshold value for classifying neonates as either “short” or “long” latency corresponded to the median value of the latency period in our sample (3 days). The 2 latency groups were compared by following the same statistical procedure as described previously. Statistical analyses were performed at the conventional 2-tailed a level of 0.05 using SAS version 9.1 (SAS Institute). Mura et al
ORIGINAL ARTICLES
August 2013
Table I. Characteristics of children born before 33 weeks gestation Gestational age at birth, weeks, n (%) 24-26 27-29 29-32 Antenatal steroids, n (%) Vaginal delivery, n (%) Materno-fetal infection, n (%) Bronchopulmonary dysplasia, n (%) Postnatal steroids, n (%) Intraventricular hemorrhage, n (%) Periventricular leukomalacia, n (%) Necrotizing enterocolitis, n (%) Breastfeeding, n (%)
PPROM
IPL
Vasc
Other
28/318 (8) 94/318 (29) 196/318 (61) 236/311 (76) 189/314 (60) 116/304 (38) 65/317 (20) 52/315 (16) 14/314 (4) 39/304 (12) 10/314 (3) 72/299 (24)
22/200 (11) 66/200 (33) 112/200 (56) 117/195 (60) 172/198 (87) 21/159 (13) 38/159 (23) 24/159 (15) 5/158 (3) 9/150 (6) 3/158 (2) 35/152 (23)
15/372 (4) 99/372 (26) 258/372 (69) 286/364 (79) 11/368 (3) 35/368 (9) 88/366 (24) 80/371 (22) 5/367 (1) 31/340 (9) 15/369 (4) 89/349 (25)
7/161 (4) 52/161 (32) 102/161 (63) 121/158 (77) 23/159 (14) 35/200 (17) 46/200 (23) 32/199 (16) 10/197 (5) 22/191 (11) 9/199 (4) 40/188 (21)
Results Out of the 1884 neonates from the EPIPAGE study who met our inclusion criteria (Figure), 247 died before hospital discharge, in the maternity ward or neonatal intensive care unit (12% of the PPROM group, 16% of the IPL group, 11% of the Vasc group, and 14% of the Other group), and 586 were not assessed with the K-ABC at age 5 years (37%, 44%, 28%, and 37% of the 4 groups, respectively). Our study cohort of 1051 neonates included 318 in the PPROM group (30%), 372 in the Vasc group (36%), 200 in the IPL group (19%), and 161 in the Other group (15%). The reasons for classifying the 586 children as “nonresponders” (ie, not evaluated with the K-ABC) included parental refusal (n = 78), death between the time of discharge from the maternity hospital and age 5 years (n = 14), nonattendance at the 5-year follow-up visit (n = 427), and child’s refusal to take the test (n = 67). The nonresponders had a lower overall maternal educational level and lower family socioeconomic status, and a higher rate of periventricular leukomalacia compared with responders, but there was no difference between the 2 groups in terms of gestational age at birth (data not shown). The characteristics of mothers and infants according to the etiology of prematurity are summarized in Tables I and II. The 4 groups differed significantly in terms of infant sex,
P value .004 <.001 <.001 <.001 .709 .167 .054 .114 .530 .736
maternal education and marital status, gestational age at birth, use of antenatal corticosteroids, mode of delivery, and presence of maternal-fetal infection. The mean MPC score in the overall study cohort was 93.6 19.7, and 140 of the 1051 children (13.3%) had cognitive impairment (MPC score <70). There were no significant differences among the 4 study groups in terms of MPC scores or rate of cognitive impairment (Table III). After adjustment for potential confounding factors (ie, sociodemographic characteristics and gestational age), our analysis showed no evidence of an increased risk of cognitive impairment in children in the PPROM group compared with those in the IPL and Other groups, but there is a decreased risk compared with the Vasc group. This difference with the Vasc group persisted after adjustment for potential intermediary factors (ie, antenatal and postnatal corticosteroid use, intraventricular hemorrhage, and periventricular leukomalacia). A sensitivity analysis with multiple imputation for missing data performed for the 1051 responders and 536 nonresponders confirmed the robustness of these results, showing a lower risk in the PPROM group compared with the Vasc group (OR, 1.54; 95% CI, 1.00-2.37; P = .04) and an absence of increased risk compared with the IPL group (OR, 1.15; 95% CI, 0.72-1.83) and the Other group (OR, 1.26; 95% CI, 0.76-1.83) after adjustment for potential confounding factors.
Table II. Sociodemographic characteristics by pregnancy complications leading to prematurity Family socioeconomic status* Executive, n (%) Intermediate professional, n (%) Administrative/public service, self-employed, student, n (%) Shop assistant, service worker, n (%) Manual worker or unemployed, n (%) Maternal educational level Primary school or no school, n (%) Secondary school first part, n (%) Secondary school second part, n (%) University, n (%) Maternal nationality other than French, n (%) Female sex, n (%) Mother married or cohabitating at time of birth, n (%)
P value
PPROM
IPL
Vasc
Other
41/317 (13) 76/317 (24) 73/317 (23) 53/317 (17) 74/317 (23)
20/200 (10) 45/200 (22) 45/200 (22) 39/200 (19) 51/200 (25)
56/372 (15) 109/372 (29) 90/372 (24) 48/372 (13) 69/372 (19)
31/160 (19) 39/160 (24) 41/160 (26) 25/160 (16) 24/160 (15)
.063
18/314 (6) 159/314 (51) 57/314 (18) 80/314 (26) 29/312 (9) 148/368 (46) 255/280 (91)
7/196 (4) 92/196 (47) 49/196 (25) 48/196 (25) 16/198 (8) 75/200 (37) 152/179 (85)
26/370 (7) 138/370 (37) 77/370 (21) 129/370 (42) 33/371 (9) 207/372 (56) 321/339 (94)
6/158 (4) 71/158 (45) 33/158 (21) 48/158 (30) 10/160 (6) 75/161 (47) 136/148 (92)
.013
.701 <.001 .002
*Family socioeconomics status is defined as the higher of 2 parental levels.
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
437
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 163, No. 2
Table III. Relationship between pregnancy complications leading to prematurity and cognitive impairment (MPC score <70) Cognitive impairment (MPC score <70)
PPROM IPL Vasc Other
n
K-ABC MPC score, mean ± SD*
%
OR
95% CI
aORz
95% CI
aORx
95% CI
318 200 372 161
95.4 20.2 93.4 19.0 91.7 19.3 94.6 19.6
11.0 12.5 15.9 13.0
1.00 1.15 1.52 1.21
0.66-1.99 0.97-2.38 0.68-2.16
1.00 1.09 1.86 1.36
0.62-1.92 1.16-2.97 0.75-2.47
1.00 1.06 1.67 1.26
0.59-1.93 1.01-2.76 0.66-2.37
†
*P = .09. †Crude OR. zAdjusted for sex, gestational age at birth, mother’s educational level, and mother’s socioeconomic status. xAdjusted for sex, gestational age at birth, mother’s educational level, mother’s socioeconomic status, antenatal corticosteroid use, intraventricular hemorrhage, cystic periventricular leukomalacia, and postnatal corticosteroid use.
To evaluate the influence of latency period on cognitive development, we considered the 294 infants of the 318 in the PPROM group with data on the time interval between rupture of membranes and birth. This latency period was short (<3 days) in 132 infants and longer ($3 days) in 162 infants. There were no significant differences in sociodemographic characteristics between the short-latency and long-latency groups, but the former group had a lower rate of maternal antenatal corticosteroid use (63% vs 86%; P < .01), was less premature (mean gestational age at birth, 30.1 2.1 weeks vs 29.6 2.0 weeks; P = .01), had a higher rate of intraventricular hemorrhage (8% vs 1%; P < .01), and had a lower rate of bronchopulmonary dysplasia (14% vs 27%; P < .01). Other characteristics investigated did not differ significantly between the 2 groups. Cognitive development was significantly worse in the short-latency group, with a rate of cognitive impairment double that of the long-latency groups (15.2% vs 7.4%) (Table IV). Finally, an adjustment for potential intermediate factors (ie, antenatal corticosteroid use, intraventricular hemorrhage, and bronchopulmonary dysplasia) led to a largely unchanged estimated OR (OR, 2.09; 95% CI, 0.91-4.82), suggesting that the association is not completely explained by these factors.
Discussion Infants born after PPROM had no increased risk of cognitive impairment at age 5 years compared with those born after IPL or to mothers with other disorders. This risk even appeared to be lower than in infants born to hypertensive mothers or at SGA (Vasc group). In addition, the risk of cognitive impairment did not appear to be homogeneous among
infants in the PPROM group. Infants born after a short latency period (<3 days) were at greater risk compared with those born after a longer latency period (>3 days). One strength of the present study is its large prospective cohort, which comprised one-third of the preterm births in France during the study period, which limited the risk of recruitment bias or lack of statistical power. In addition, usable cognitive data were available from approximately two-thirds of eligible children at age 5 years, which was satisfactory in view of the geographic dispersion and mobility of the families. This rate is similar to rates reported in other longitudinal studies performed in this type of population.23,24 In our cohort, subjects who were lost to follow-up (nonresponders) more frequently presented with a certain number of factors associated with the risk of cognitive impairment, as reported previously.25 To ensure that the difference in the nonresponder rate between groups had not led to a selection bias in our main analysis, we performed a sensitivity analysis with multiple imputation for missing data. Our results were robust, indicating that selection bias did not confound interpretation of our data. Given that the improved survival of very preterm infants is associated with an increased prevalence of cognitive developmental abnormalities,2,3 it was important to differentiate these 2 outcomes (mortality and cognitive development) and to focus on the cognitive impairment. We found indeed the lowest rate of mortality before discharge to home and the highest rate of cognitive impairment in the Vasc group. Conversely, infant in the PPROM group exhibited lower rates of both cognitive impairment and mortality compared with those in the IPL and Other groups. Thus, the low rate of cognitive impairment in PPROM group should not be related to
Table IV. Relationship between cognitive impairment and latency period in 294 children born after PPROM Cognitive impairment (MPC score <70) Latency period $3 days <3 days
n 162 132
K-ABC MPC score, mean ± SD z
98.4 (18.4) 92.2 (19.6)
n (%) x
12 (7.4) 20 (15.2)
OR*
95% CI
aOR†
95% CI
1 2.23
1.04-4.76
1 2.32
1.07-5.02
*Crude OR. †Adjusted for gestational age. zSignificantly different (P < .01) from infants born after a short latency period. xSignificantly different (P = .03) from infants born after a short latency period.
438
Mura et al
August 2013 a survival effect (high mortality resulting in selection of children free of cognitive impairment). These results are consistent with those of Blumenfeld et al,26 who concluded that PPROM was not associated with an increased risk of neonatal mortality in any gestational age group. The method we used to form the different groups according to the etiology of prematurity also played an important part in our analyses. Our hierarchical method enhanced the comparability of the PPROM and IPL groups because it eliminated SGA infants and infants born after maternal hypertension from these 2 groups. Considering the results for our secondary endpoint, which show an association between cognitive impairment and latency period, it is possible that a classification differentiating shorter-latency and longer-latency periods after PPROM would have led to different conclusions. However, because no publication has yet reported a difference in risk according to latency period,16,27,28 an analysis that differentiated the 2 groups a priori according to latency period did not seem justified. Our results differ from those published by Spinillo et al in 1995.17 That study identified an increased risk of neurodevelopmental impairment in preterm infants born after PPROM compared with those born after spontaneous preterm labor in the absence of any pregnancy-related medical complications (the equivalent of our IPL group). However, there were several major methodological differences between that study and the present study. First, in the former study, neurodevelopmental outcome was measured at 2 years using the Bayley Scales of Infant and Toddler Development, which have a poor predictive value for the cognitive development of school-aged children.29 Second, the judgment criterion used in that study was a composite criterion that did not allow for the distinction between cognitive deficit and neurologic impairment. Some previous studies have identified PPROM as a likely risk factor for cerebral palsy,12,13 which may seem inconsistent with our findings. However, the level of evidence in those case-control studies was low, and this relationship has not been systematically verified.15 Finally, PPROM might be a risk factor for cerebral palsy, while not being linked to cognitive impairment. The increased risk of cognitive impairment in infants with a short latency period (<3 days) in our cohort is a new finding. This result helps qualify the results of our primary objective, given the similar rates of cognitive impairment in children with a short latency period and children in the Vasc group (15.2% vs 15.9%). Pasquier et al30 identified an increased risk of neonatal mortality in preterm infants born at <30 weeks gestational age after PPROM with a short latency period (<48 hours). Results at the limit of significance also have been reported for neonatal morbidity and mortality in infants born after PPROM with a short latency period (<48 hours) compared with those with very long latency period (>14 days).31 Other studies have reported no links among intraventricular hemorrhage, ventricular dilatation, and latency period,16,27 but the absence of significance in these studies may be explained by a lack of statistical power or the use of a different threshold to define short and long
ORIGINAL ARTICLES latency periods than that used in the present study. This result may suggest the existence of greater fetal and cerebral injury in infants born after PPROM with a short latency period. It is likely that this finding is not related to the specific relationship between PPROM and chorioamniotitis, given that several studies have reported an increased risk of histological chorioamniotitis with latency period.16,27 An alternative explanation could be that the infection/inflammation associated with PPROM is more severe during preterm labor not responding to tocolysis, leading to a short latency period and to more significant cognitive sequelae. It is also possible that placental infection contributes little to the relevant cerebral lesions, and that other factors merit investigation. Finally, as in all observational studies, the possibility remains that unmeasured confounders or uncontrolled selection biases may play a role in these associations. The present study suggests that overall, infants born after PPROM are not at greater risk for cognitive impairment compared with infants born preterm due to another cause. This conclusion is complemented by the finding of an increased risk of cognitive impairment in infants born after PPROM with a short latency period (<3 days) compared with those with a longer latency period. Considering the crucial importance of gestational age at birth in the occurrence of cognitive sequelae, and in view of our present results, a significant features of this study is that it does not provide a rationale for early induction of labor or cesarean delivery in women with preterm rupture of membranes. n Submitted for publication May 30, 2012; last revision received Dec 7, 2012; accepted Jan 16, 2013. Reprint requests: Thibault Mura, MD, PhD, Centre d’Investigation Clinique ^ pital Saint Eloi, CHRU Montpellier, 80 Ave Augustin Fliche, 34000 1001, Ho Montpellier, France. E-mail: [email protected]
References 1. Buitendijk S, Zeitlin J, Cuttini M, Langhoff-Roos J, Bottu J. Indicators of fetal and infant health outcomes. Eur J Obstet Gynecol Reprod Biol 2003; 111(Suppl 1):S66-77. 2. Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJ. Cognitive and behavioral outcomes of school-aged children who were born preterm: a meta-analysis. JAMA 2002;288:728-37. 3. Ancel PY, Livinec F, Larroque B, Marret S, Arnaud C, Pierrat V, et al. Cerebral palsy among very preterm children in relation to gestational age and neonatal ultrasound abnormalities: the EPIPAGE cohort study. Pediatrics 2006;117:828-35. 4. Larroque B, Ancel PY, Marret S, Marchand L, Andre M, Arnaud C, et al. Neurodevelopmental disabilities and special care of 5-year-old children born before 33 weeks of gestation (the EPIPAGE study): a longitudinal cohort study. Lancet 2008;371:813-20. 5. Holling EE, Leviton A. Characteristics of cranial ultrasound whitematter echolucencies that predict disability: a review. Dev Med Child Neurol 1999;41:136-9. 6. Wu YW, Colford JM Jr. Chorioamnionitis as a risk factor for cerebral palsy: a meta-analysis. JAMA 2000;284:1417-24. 7. Menon R, Fortunato SJ. Infection and the role of inflammation in preterm premature rupture of the membranes. Best Pract Res Clin Obstet Gynaecol 2007;21:467-78. 8. Ancel PY, Marret S, Larroque B, Arnaud C, Zupan-Simunek V, Voyer M, et al. Are maternal hypertension and small-for-gestational age risk
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
439
THE JOURNAL OF PEDIATRICS
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19. 20.
440
www.jpeds.com
factors for severe intraventricular hemorrhage and cystic periventricular leukomalacia? Results of the EPIPAGE cohort study. Am J Obstet Gynecol 2005;193:178-84. Bauer M, Fast C, Haas J, Resch B, Lang U, Pertl B. Cystic periventricular leukomalacia in preterm infants: an analysis of obstetric risk factors. Early Hum Dev 2009;85:163-9. Spinillo A, Capuzzo E, Stronati M, Ometto A, De Santolo A, Acciano S. Obstetric risk factors for periventricular leukomalacia among preterm infants. Br J Obstet Gynaecol 1998;105:865-71. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75-84. Drougia A, Giapros V, Krallis N, Theocharis P, Nikaki A, Tzoufi M, et al. Incidence and risk factors for cerebral palsy in infants with perinatal problems: a 15-year review. Early Hum Dev 2007;83:541-7. Murphy DJ, Sellers S, MacKenzie IZ, Yudkin PL, Johnson AM. Casecontrol study of antenatal and intrapartum risk factors for cerebral palsy in very preterm singleton babies. Lancet 1995;346:1449-54. Beaino G, Khoshnood B, Kaminski M, Pierrat V, Marret S, Matis J, et al. Predictors of cerebral palsy in very preterm infants: the EPIPAGE prospective population-based cohort study. Dev Med Child Neurol 2012; 52:e119-25. Kurkinen-Raty M, Koivisto M, Jouppila P. Perinatal and neonatal outcome and late pulmonary sequelae in infants born after preterm premature rupture of membranes. Obstet Gynecol 1998;92:408-15. Park SH, Kim HJ, Yang JH, Choi JS, Lim JE, Oh MJ, et al. Neonatal brain damage following prolonged latency after preterm premature rupture of membranes. J Korean Med Sci 2006;21:485-9. Spinillo A, Capuzzo E, Stronati M, Ometto A, Orcesi S, Fazzi E. Effect of preterm premature rupture of membranes on neurodevelopmental outcome: follow-up at two years of age. Br J Obstet Gynaecol 1995;102:882-7. Larroque B, Breart G, Kaminski M, Dehan M, Andre M, Burguet A, et al. Survival of very preterm infants: EPIPAGE, a population-based cohort study. Arch Dis Child Fetal Neonatal Ed 2004;89:F139-44. Kaufman AS, Kaufman NL. Kaufman Assessment Battery for Children. Circle Pines (MN): American Guidance Service; 1983. World Health Organization. Multiaxial classification of child and adolescent psychiatric disorders: the ICD-10 classification of mental and behavioral disorders in children and adolescents. Geneva, Switzerland: World Health Organization; 1996.
Vol. 163, No. 2 21. Schafer JL. Analysis of incomplete multivariate data. New York: Chapman and Hall; 1997. 22. Yuan YC. Multiple imputation for missing data: concepts and new development (version 9.0). Rockville, MD: SAS Institute. Available from: http://www2.sas.com/proceedings/sugi25/25/st/25p267.pdf. Accessed March 4, 2013. 23. Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005;352:9-19. 24. Wolke D, Meyer R. Cognitive status, language attainment, and prereading skills of 6-year-old very preterm children and their peers: the Bavarian Longitudinal Study. Dev Med Child Neurol 1999;41: 94-109. 25. Hille ET, Elbertse L, Gravenhorst JB, Brand R, Verloove-Vanhorick SP. Nonresponse bias in a follow-up study of 19-year-old adolescents born as preterm infants. Pediatrics 2005;116:e662-6. 26. Blumenfeld YJ, Lee HC, Gould JB, Langen ES, Jafari A, El-Sayed YY. The effect of preterm premature rupture of membranes on neonatal mortality rates. Obstet Gynecol 2010;16:1381-6. 27. McElrath TF, Allred EN, Leviton A. Prolonged latency after preterm premature rupture of membranes: an evaluation of histologic condition and intracranial ultrasonic abnormality in the neonate born at <28 weeks of gestation. Am J Obstet Gynecol 2003;189:794-8. 28. Manuck TA, Maclean CC, Silver RM, Varner MW. Preterm premature rupture of membranes: does the duration of latency influence perinatal outcomes? Am J Obstet Gynecol 2009;201:414.e1-e6. 29. Hack M, Taylor HG, Drotar D, Schluchter M, Cartar L, WilsonCostello D, et al. Poor predictive validity of the Bayley Scales of Infant Development for cognitive function of extremely low birth weight children at school age. Pediatrics 2005;116:333-41. 30. Pasquier JC, Bujold E, Rabilloud M, Picaud JC, Ecochard R, Claris O, et al. Effect of latency period after premature rupture of membranes on 2 years infant mortality (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2007;135:21-7. 31. Pasquier JC, Picaud JC, Rabilloud M, Claris O, Ecochard R, Moret S, et al. Neonatal outcomes after elective delivery management of preterm premature rupture of the membranes before 34 weeks’ gestation (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2009;143:18-23.
Mura et al
ORIGINAL ARTICLES
August 2013
Appendix Members of the EPIPAGE study group include: Institut National de la Sante et de la Recherche Medicale U149: B. Larroque (national coordinator), P. Y. Ancel, B. Blondel, G. Breart, M. Dehan, M. Garel, M. Kaminski, F. Maillard, C. du Mazaubrun, P. Missy, F. Sehili, K. Supernant, L. Marchand Alsace: M. Durand, J. Matis, J. Messer, A. Treisser (Hopital de Hautepierre, Strasbourg) Franche-Comte: A. Burguet, L. Abraham-Lerat, A. Menget, P. Roth, J.-P. Schaal, G. Thiriez (CHU St Jacques, Besancon) Haute-Normandie: C. Leveque, S. Marret, L. Marpeau (Hopital Charles Nicolle, Rouen) Languedoc-Roussillon: P. Boulot, G. Cambonie (Hopital Arnaud de Villeneuve, Montpellier), A.-M. Donadio, B. Ledesert (ORS Montpellier)
Lorraine: M. Andre, J. Fresson, J. M. Hascoet (Maternite Regionale, Nancy) Midi-Pyrenees: C. Arnaud, S. Bourdet-Loubere, H. Grandjean (Institut National de la Sante et de la Recherche Medicale U558, Toulouse), M. Rolland (Hopital des Enfants, Toulouse) Nord-Pas-de-Calais: C. Leignel, P. Lequien, V. Pierrat, F. Puech, D. Subtil, P. Truffert (Hopital Jeanne de Flandre, Lille) Pays-de-Loire: G. Boog, V. Rouger-Bureau, J.-C. Roze (Hopital Mere-Enfant, Nantes) Paris-Petite-Couronne: P. Y. Ancel, G. Breart, M. Kaminski, C. du Mazaubrun (Institut National de la Sante et de la Recherche Medicale U149, Paris), M. Dehan, V. ZupanSimunek (Hopital Antoine Beclere, Clamart), M. Vodovar, M. Voyer (Institut de Puericulture, Paris).
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
440.e1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 163, No. 2
Figure. Flow chart of the study cohort. *In 2 regions, follow-up was proposed for only 1 of every 2 infants born at exactly 32 weeks gestational age. **Including 31 of 253 infants (12%) with a latency period >3 days, 31 of 286 infants (11%) with a latency period <3 days (P = .60) and 7 of 38 infants with no information available on latency period.*** Among the 494 survivors with no available K-ABC MPC score, 8 did not complete the K-ABC because of a sensory or motor disability or a language disorder. The distribution of these children was not significantly associated with the etiology of prematurity (P = .63).
440.e2
Mura et al
Study design The Etude Epidemiologique sur les Petits Ages Gestationnels Study is a population-based cohort of children followed up from birth to age 5 years recruited in 9 French regions in 1997. We analyzed data from singletons born between 24 and 32 weeks gestation categorized into 4 groups according to etiology of prematurity: infants born after PPROM, after idiopathic preterm labor, in a vascular context (Vasc), and to women with other complications (Other). Cognitive development at age 5 years was assessed using the Mental Processing Composite score of the Kaufman-Assessment Battery for Children. Results Among the 1051 children followed up to age 5 years, the mean Mental Processing Composite score was 93.6 19.7, and 13.3% of the children (140 of 1051) had cognitive impairment. After adjustment for potential confounders, the risk of cognitive impairment among infants in the PPROM group was not significantly different than that in the idiopathic preterm labor group (OR, 1.09; 95% CI, 0.62-1.92) and the Other group (OR, 1.36; 95% CI, 0.75-2.47), but was lower than that in the Vasc group (OR, 1.86; 95% CI, 1.16-2.97). In the PPROM group, the risk of cognitive impairment was greater when the latency period (ie, time from rupture to delivery) was <3 days (OR, 2.32; 95% CI, 1.07-5.02). Conclusion Preterm infants born after PPROM are not at increased risk for cognitive impairment in childhood, but the time between PPROM and birth may influence that risk. (J Pediatr 2013;163:435-40).
I
n Europe, between 1.1% and 1.6% of liveborn infants are born very preterm (ie, before 33 weeks gestational age).1 The improved survival of very preterm infants has been associated with an increased rate of neuromotor and cognitive development abnormalities in these survivors.2,3 With a prevalence of >10% in this population, cognitive impairment is one of the most common sequelae associated with very preterm birth.4 These cognitive sequelae often result from white matter lesions that go unidentified in the perinatal period.5 These lesions may be attributable to hemodynamic causes, and also may be the result of infectious6 or inflammatory processes, as is observed in 50% of births that occur after preterm premature rupture of membranes (PPROM).7 Infants born after PPROM also may be at increased risk for intraventricular hemorrhage and periventricular leukomalacia.8-10 Although PPROM accounts for one-third of preterm births,11 studies on the long-term outcomes of these infants are rare and have reported conflicting results. Of note is a possible increased risk of cerebral palsy12-14; however, this link has yet to be systematically verified.15,16 Only 1 study reported an increased risk of cognitive impairment in very preterm infants born after PPROM17; in that study, this risk was higher in neonates with a latency period (ie, time from
EPIPAGE IPL K-ABC MPC PPROM SGA Vasc
miologique sur les Petits Ages Gestationnels Etude Epide Idiopathic preterm labor Kaufman Assessment Battery for Children Mental Processing Composite Preterm premature rupture of membranes Small for gestational age Vascular
From the 1Clinical Investigation Center and Information Medical Department, University Hospital of Montpellier; 2 et de la Recherche Institut National de la Sante dicale, Montpellier, France; 3Department of Me Neonatology, Croix Rousse Hospital, Claude Bernard University Lyon 1, Lyon, France; 4Pierre and Marie Curie University, Paris, France; 5Epidemiological Research Unit on Perinatal Health and Women’s and Children’s et de la Recherche Health, Institut National de la Sante dicale, Villejuif, France; 6Department of Neonatal Me Medicine, Rouen University Hospital and the Institut et de la Recherche Me dicale Avenir National de la Sante Research Group, Institute for Biomedical Research, University of Rouen, Rouen, France; 7Department of Neonatology, Children’s Hospital, Nantes, France; 8 Department of Neonatology, Jeanne de Flandre Hospital, Lille, France; 9Department of Pediatrics II, Hautepierre Hospital, Strasbourg, France; 10Information Medical Department, Regional Maternity University Hospital, Nancy, France; 11Department of Pediatrics, Besanc¸on University Hospital, Besanc¸on, France; 12 et de la Recherche Institut National de la Sante dicale; 13Clinical Epidemiology Unit, Toulouse Me University Hospital, Toulouse, France; 14Institut National et de la Recherche Me dicale, de la Sante Epidemiological Research Unit on Perinatal Health and Women’s and Children’s Health, Tenon Hospital, Paris, gional de Sante , France; and 15Observatoire Re Montpellier, France *A list of EPIPAGE Study Group members is available at www.jpeds.com (Appendix). Supported by INSERM (National Institute of Health and Medical Research), Merck Sharp, Dohme-Chibret, la dicale (The Medical ReFondation de la Recherche Me nerale de la search Foundation), and la Direction Ge du Ministe re des Affaires Sociales (French Ministry Sante of Health). The authors declare no conflicts of interest. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2013.01.039
435
THE JOURNAL OF PEDIATRICS
www.jpeds.com
rupture to delivery) exceeding 48 hours compared with those with a shorter latency period. In a cohort of very preterm infants in the Etude Epidemiologique sur les Petits Ages Gestationnels (EPIPAGE) population, we investigated whether those born after PPROM were at greater risk for cognitive impairment at age 5 years. We also investigated whether the latency period between membrane rupture and birth influenced later cognitive outcome.
Methods All parents of infants born at 24-32 weeks gestational age between January 1, 1997, and December 31, 1997, in 9 French regions (representing more than one-third of the country) were invited to participate in the EPIPAGE study.18 The children in the study cohort were followed up at age 2 month, at age 9 months, and then every year up to age 5 years. Out of the 2855 liveborn neonates included in the EPIPAGE cohort, 25 had no available information concerning the etiology of prematurity (Figure; available at www.jpeds.com) and were excluded from this analysis. We also excluded 885 neonates from multiple pregnancies (to avoid a confounding bias) and 9 with serious malformations. Owing to the number of preterm neonates born at 32 weeks gestational age, all regions were given the option of including only 1 of every 2 infants born at exactly 32 weeks in the follow-up; 2 regions chose this option (52 infants). In these regions, children whose mother was born on an odd-numbered day were selected. Finally, a total of 1884 children met our inclusion criteria and were eligible for analysis. This study was approved by the French data protection agency (Commission Nationale de l’Informatique et des Libertes). Maternal, obstetric, and neonatal data were collected at birth following a standardized protocol. All children were invited for a checkup at age 5 years. The parents were contacted in writing, and if no response was received, a system of followup contacts by phone or mail was established. The checkup involved medical, social, and psychological assessments, including a cognitive evaluation performed by a psychologist using the Kaufman Assessment Battery for Children (K-ABC).19 We divided the infants into 4 mutually exclusive groups according to the pregnancy complications that led to preterm delivery.11 Infants born after PPROM without other pregnancy complications (PPROM group) were compared with infants born after idiopathic preterm labor (IPL group), defined as spontaneous onset of labor before rupture of membranes; infants born in a vascular context (Vasc group) to mothers with hypertension, defined as systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg during pregnancy, and small for gestational age (SGA) infants, defined as a birth weight <10th percentile for gestational age of the liveborn infants in our sample (with these 2 groups aggregated because intrauterine growth restriction in preterm birth frequently has a vascular cause); and infants born to mothers with other complications (Other group) resulting in preterm delivery, including antepartum hemorrhage without SGA status or maternal hypertension. 436
Vol. 163, No. 2 The K-ABC has been validated in France for use in children aged 2.5-12.5 years.19 The K-ABC Mental Processing Composite (MPC) score, considered equivalent to IQ score, is a global measure of cognitive ability. This score was standardized to a mean of 100 (SD 15) using published French standards. In accordance with the World Health Organization’s definition, cognitive impairment was defined as an MPC score <70 (difference of >2 SD).20 The following data were included in the analysis: sociodemographic data (ie, sex, maternal nationality, maternal educational level, and family socioeconomic status), antepartal and peripartal data (ie, gestational age, expressed as completed weeks of amenorrhea, antenatal corticosteroids, and mode of delivery), and neonatal data (ie, grade 3-4 intraventricular hemorrhage, periventricular leukomalacia, maternal-fetal infection, ulcero-necrotizing enterocolitis, bronchopulmonary dysplasia, and postnatal corticosteroid use). The c2 test was used to analyze sociodemographic, antenatal, peripartal, and neonatal characteristics associated with the pregnancy complications leading to preterm delivery in the 4 study groups. ANOVA was used to compare mean KABC MPC scores. Logistic regression was used to evaluate the relationships between each of the pregnancy complication groups and cognitive impairment at age 5 years. We initially examined these relationships after adjustment for gestational age at birth (treated categorically using gestational age groups, as identified in Table I) and for the sociodemographic covariates deemed potentially important on univariate analysis (P < .20 for the association with pregnancy complication groups). Later, we made adjustments for antenatal, peripartal, and neonatal covariates meeting the same criteria (P < .20 for the associations with pregnancy complication groups). We considered these additional factors separately because they might be involved in the causal relationship between pregnancy complications and cognitive impairment, and thus could act as intermediate factors. Adjusting for their effect helped us explain the associations between pregnancy complication groups and cognitive impairment. We addressed the risk of selection bias potentially induced by missing values by carrying out a sensitivity analysis using data imputation on both outcome variable and covariates. We used the multiple imputation by Markov chain Monte Carlo method21 with the MI and MIanalyze procedures implemented in SAS version 9 (SAS Institute, Cary, North Carolina).22 The imputed dataset was generated by performing 50 imputation cycles; 9.6% of the data were imputed. To examine the influence of the latency period on cognitive outcome, we divided the population of neonates born after PPROM into 2 groups according to the interval between PPROM and birth. The threshold value for classifying neonates as either “short” or “long” latency corresponded to the median value of the latency period in our sample (3 days). The 2 latency groups were compared by following the same statistical procedure as described previously. Statistical analyses were performed at the conventional 2-tailed a level of 0.05 using SAS version 9.1 (SAS Institute). Mura et al
ORIGINAL ARTICLES
August 2013
Table I. Characteristics of children born before 33 weeks gestation Gestational age at birth, weeks, n (%) 24-26 27-29 29-32 Antenatal steroids, n (%) Vaginal delivery, n (%) Materno-fetal infection, n (%) Bronchopulmonary dysplasia, n (%) Postnatal steroids, n (%) Intraventricular hemorrhage, n (%) Periventricular leukomalacia, n (%) Necrotizing enterocolitis, n (%) Breastfeeding, n (%)
PPROM
IPL
Vasc
Other
28/318 (8) 94/318 (29) 196/318 (61) 236/311 (76) 189/314 (60) 116/304 (38) 65/317 (20) 52/315 (16) 14/314 (4) 39/304 (12) 10/314 (3) 72/299 (24)
22/200 (11) 66/200 (33) 112/200 (56) 117/195 (60) 172/198 (87) 21/159 (13) 38/159 (23) 24/159 (15) 5/158 (3) 9/150 (6) 3/158 (2) 35/152 (23)
15/372 (4) 99/372 (26) 258/372 (69) 286/364 (79) 11/368 (3) 35/368 (9) 88/366 (24) 80/371 (22) 5/367 (1) 31/340 (9) 15/369 (4) 89/349 (25)
7/161 (4) 52/161 (32) 102/161 (63) 121/158 (77) 23/159 (14) 35/200 (17) 46/200 (23) 32/199 (16) 10/197 (5) 22/191 (11) 9/199 (4) 40/188 (21)
Results Out of the 1884 neonates from the EPIPAGE study who met our inclusion criteria (Figure), 247 died before hospital discharge, in the maternity ward or neonatal intensive care unit (12% of the PPROM group, 16% of the IPL group, 11% of the Vasc group, and 14% of the Other group), and 586 were not assessed with the K-ABC at age 5 years (37%, 44%, 28%, and 37% of the 4 groups, respectively). Our study cohort of 1051 neonates included 318 in the PPROM group (30%), 372 in the Vasc group (36%), 200 in the IPL group (19%), and 161 in the Other group (15%). The reasons for classifying the 586 children as “nonresponders” (ie, not evaluated with the K-ABC) included parental refusal (n = 78), death between the time of discharge from the maternity hospital and age 5 years (n = 14), nonattendance at the 5-year follow-up visit (n = 427), and child’s refusal to take the test (n = 67). The nonresponders had a lower overall maternal educational level and lower family socioeconomic status, and a higher rate of periventricular leukomalacia compared with responders, but there was no difference between the 2 groups in terms of gestational age at birth (data not shown). The characteristics of mothers and infants according to the etiology of prematurity are summarized in Tables I and II. The 4 groups differed significantly in terms of infant sex,
P value .004 <.001 <.001 <.001 .709 .167 .054 .114 .530 .736
maternal education and marital status, gestational age at birth, use of antenatal corticosteroids, mode of delivery, and presence of maternal-fetal infection. The mean MPC score in the overall study cohort was 93.6 19.7, and 140 of the 1051 children (13.3%) had cognitive impairment (MPC score <70). There were no significant differences among the 4 study groups in terms of MPC scores or rate of cognitive impairment (Table III). After adjustment for potential confounding factors (ie, sociodemographic characteristics and gestational age), our analysis showed no evidence of an increased risk of cognitive impairment in children in the PPROM group compared with those in the IPL and Other groups, but there is a decreased risk compared with the Vasc group. This difference with the Vasc group persisted after adjustment for potential intermediary factors (ie, antenatal and postnatal corticosteroid use, intraventricular hemorrhage, and periventricular leukomalacia). A sensitivity analysis with multiple imputation for missing data performed for the 1051 responders and 536 nonresponders confirmed the robustness of these results, showing a lower risk in the PPROM group compared with the Vasc group (OR, 1.54; 95% CI, 1.00-2.37; P = .04) and an absence of increased risk compared with the IPL group (OR, 1.15; 95% CI, 0.72-1.83) and the Other group (OR, 1.26; 95% CI, 0.76-1.83) after adjustment for potential confounding factors.
Table II. Sociodemographic characteristics by pregnancy complications leading to prematurity Family socioeconomic status* Executive, n (%) Intermediate professional, n (%) Administrative/public service, self-employed, student, n (%) Shop assistant, service worker, n (%) Manual worker or unemployed, n (%) Maternal educational level Primary school or no school, n (%) Secondary school first part, n (%) Secondary school second part, n (%) University, n (%) Maternal nationality other than French, n (%) Female sex, n (%) Mother married or cohabitating at time of birth, n (%)
P value
PPROM
IPL
Vasc
Other
41/317 (13) 76/317 (24) 73/317 (23) 53/317 (17) 74/317 (23)
20/200 (10) 45/200 (22) 45/200 (22) 39/200 (19) 51/200 (25)
56/372 (15) 109/372 (29) 90/372 (24) 48/372 (13) 69/372 (19)
31/160 (19) 39/160 (24) 41/160 (26) 25/160 (16) 24/160 (15)
.063
18/314 (6) 159/314 (51) 57/314 (18) 80/314 (26) 29/312 (9) 148/368 (46) 255/280 (91)
7/196 (4) 92/196 (47) 49/196 (25) 48/196 (25) 16/198 (8) 75/200 (37) 152/179 (85)
26/370 (7) 138/370 (37) 77/370 (21) 129/370 (42) 33/371 (9) 207/372 (56) 321/339 (94)
6/158 (4) 71/158 (45) 33/158 (21) 48/158 (30) 10/160 (6) 75/161 (47) 136/148 (92)
.013
.701 <.001 .002
*Family socioeconomics status is defined as the higher of 2 parental levels.
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
437
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 163, No. 2
Table III. Relationship between pregnancy complications leading to prematurity and cognitive impairment (MPC score <70) Cognitive impairment (MPC score <70)
PPROM IPL Vasc Other
n
K-ABC MPC score, mean ± SD*
%
OR
95% CI
aORz
95% CI
aORx
95% CI
318 200 372 161
95.4 20.2 93.4 19.0 91.7 19.3 94.6 19.6
11.0 12.5 15.9 13.0
1.00 1.15 1.52 1.21
0.66-1.99 0.97-2.38 0.68-2.16
1.00 1.09 1.86 1.36
0.62-1.92 1.16-2.97 0.75-2.47
1.00 1.06 1.67 1.26
0.59-1.93 1.01-2.76 0.66-2.37
†
*P = .09. †Crude OR. zAdjusted for sex, gestational age at birth, mother’s educational level, and mother’s socioeconomic status. xAdjusted for sex, gestational age at birth, mother’s educational level, mother’s socioeconomic status, antenatal corticosteroid use, intraventricular hemorrhage, cystic periventricular leukomalacia, and postnatal corticosteroid use.
To evaluate the influence of latency period on cognitive development, we considered the 294 infants of the 318 in the PPROM group with data on the time interval between rupture of membranes and birth. This latency period was short (<3 days) in 132 infants and longer ($3 days) in 162 infants. There were no significant differences in sociodemographic characteristics between the short-latency and long-latency groups, but the former group had a lower rate of maternal antenatal corticosteroid use (63% vs 86%; P < .01), was less premature (mean gestational age at birth, 30.1 2.1 weeks vs 29.6 2.0 weeks; P = .01), had a higher rate of intraventricular hemorrhage (8% vs 1%; P < .01), and had a lower rate of bronchopulmonary dysplasia (14% vs 27%; P < .01). Other characteristics investigated did not differ significantly between the 2 groups. Cognitive development was significantly worse in the short-latency group, with a rate of cognitive impairment double that of the long-latency groups (15.2% vs 7.4%) (Table IV). Finally, an adjustment for potential intermediate factors (ie, antenatal corticosteroid use, intraventricular hemorrhage, and bronchopulmonary dysplasia) led to a largely unchanged estimated OR (OR, 2.09; 95% CI, 0.91-4.82), suggesting that the association is not completely explained by these factors.
Discussion Infants born after PPROM had no increased risk of cognitive impairment at age 5 years compared with those born after IPL or to mothers with other disorders. This risk even appeared to be lower than in infants born to hypertensive mothers or at SGA (Vasc group). In addition, the risk of cognitive impairment did not appear to be homogeneous among
infants in the PPROM group. Infants born after a short latency period (<3 days) were at greater risk compared with those born after a longer latency period (>3 days). One strength of the present study is its large prospective cohort, which comprised one-third of the preterm births in France during the study period, which limited the risk of recruitment bias or lack of statistical power. In addition, usable cognitive data were available from approximately two-thirds of eligible children at age 5 years, which was satisfactory in view of the geographic dispersion and mobility of the families. This rate is similar to rates reported in other longitudinal studies performed in this type of population.23,24 In our cohort, subjects who were lost to follow-up (nonresponders) more frequently presented with a certain number of factors associated with the risk of cognitive impairment, as reported previously.25 To ensure that the difference in the nonresponder rate between groups had not led to a selection bias in our main analysis, we performed a sensitivity analysis with multiple imputation for missing data. Our results were robust, indicating that selection bias did not confound interpretation of our data. Given that the improved survival of very preterm infants is associated with an increased prevalence of cognitive developmental abnormalities,2,3 it was important to differentiate these 2 outcomes (mortality and cognitive development) and to focus on the cognitive impairment. We found indeed the lowest rate of mortality before discharge to home and the highest rate of cognitive impairment in the Vasc group. Conversely, infant in the PPROM group exhibited lower rates of both cognitive impairment and mortality compared with those in the IPL and Other groups. Thus, the low rate of cognitive impairment in PPROM group should not be related to
Table IV. Relationship between cognitive impairment and latency period in 294 children born after PPROM Cognitive impairment (MPC score <70) Latency period $3 days <3 days
n 162 132
K-ABC MPC score, mean ± SD z
98.4 (18.4) 92.2 (19.6)
n (%) x
12 (7.4) 20 (15.2)
OR*
95% CI
aOR†
95% CI
1 2.23
1.04-4.76
1 2.32
1.07-5.02
*Crude OR. †Adjusted for gestational age. zSignificantly different (P < .01) from infants born after a short latency period. xSignificantly different (P = .03) from infants born after a short latency period.
438
Mura et al
August 2013 a survival effect (high mortality resulting in selection of children free of cognitive impairment). These results are consistent with those of Blumenfeld et al,26 who concluded that PPROM was not associated with an increased risk of neonatal mortality in any gestational age group. The method we used to form the different groups according to the etiology of prematurity also played an important part in our analyses. Our hierarchical method enhanced the comparability of the PPROM and IPL groups because it eliminated SGA infants and infants born after maternal hypertension from these 2 groups. Considering the results for our secondary endpoint, which show an association between cognitive impairment and latency period, it is possible that a classification differentiating shorter-latency and longer-latency periods after PPROM would have led to different conclusions. However, because no publication has yet reported a difference in risk according to latency period,16,27,28 an analysis that differentiated the 2 groups a priori according to latency period did not seem justified. Our results differ from those published by Spinillo et al in 1995.17 That study identified an increased risk of neurodevelopmental impairment in preterm infants born after PPROM compared with those born after spontaneous preterm labor in the absence of any pregnancy-related medical complications (the equivalent of our IPL group). However, there were several major methodological differences between that study and the present study. First, in the former study, neurodevelopmental outcome was measured at 2 years using the Bayley Scales of Infant and Toddler Development, which have a poor predictive value for the cognitive development of school-aged children.29 Second, the judgment criterion used in that study was a composite criterion that did not allow for the distinction between cognitive deficit and neurologic impairment. Some previous studies have identified PPROM as a likely risk factor for cerebral palsy,12,13 which may seem inconsistent with our findings. However, the level of evidence in those case-control studies was low, and this relationship has not been systematically verified.15 Finally, PPROM might be a risk factor for cerebral palsy, while not being linked to cognitive impairment. The increased risk of cognitive impairment in infants with a short latency period (<3 days) in our cohort is a new finding. This result helps qualify the results of our primary objective, given the similar rates of cognitive impairment in children with a short latency period and children in the Vasc group (15.2% vs 15.9%). Pasquier et al30 identified an increased risk of neonatal mortality in preterm infants born at <30 weeks gestational age after PPROM with a short latency period (<48 hours). Results at the limit of significance also have been reported for neonatal morbidity and mortality in infants born after PPROM with a short latency period (<48 hours) compared with those with very long latency period (>14 days).31 Other studies have reported no links among intraventricular hemorrhage, ventricular dilatation, and latency period,16,27 but the absence of significance in these studies may be explained by a lack of statistical power or the use of a different threshold to define short and long
ORIGINAL ARTICLES latency periods than that used in the present study. This result may suggest the existence of greater fetal and cerebral injury in infants born after PPROM with a short latency period. It is likely that this finding is not related to the specific relationship between PPROM and chorioamniotitis, given that several studies have reported an increased risk of histological chorioamniotitis with latency period.16,27 An alternative explanation could be that the infection/inflammation associated with PPROM is more severe during preterm labor not responding to tocolysis, leading to a short latency period and to more significant cognitive sequelae. It is also possible that placental infection contributes little to the relevant cerebral lesions, and that other factors merit investigation. Finally, as in all observational studies, the possibility remains that unmeasured confounders or uncontrolled selection biases may play a role in these associations. The present study suggests that overall, infants born after PPROM are not at greater risk for cognitive impairment compared with infants born preterm due to another cause. This conclusion is complemented by the finding of an increased risk of cognitive impairment in infants born after PPROM with a short latency period (<3 days) compared with those with a longer latency period. Considering the crucial importance of gestational age at birth in the occurrence of cognitive sequelae, and in view of our present results, a significant features of this study is that it does not provide a rationale for early induction of labor or cesarean delivery in women with preterm rupture of membranes. n Submitted for publication May 30, 2012; last revision received Dec 7, 2012; accepted Jan 16, 2013. Reprint requests: Thibault Mura, MD, PhD, Centre d’Investigation Clinique ^ pital Saint Eloi, CHRU Montpellier, 80 Ave Augustin Fliche, 34000 1001, Ho Montpellier, France. E-mail: [email protected]
References 1. Buitendijk S, Zeitlin J, Cuttini M, Langhoff-Roos J, Bottu J. Indicators of fetal and infant health outcomes. Eur J Obstet Gynecol Reprod Biol 2003; 111(Suppl 1):S66-77. 2. Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJ. Cognitive and behavioral outcomes of school-aged children who were born preterm: a meta-analysis. JAMA 2002;288:728-37. 3. Ancel PY, Livinec F, Larroque B, Marret S, Arnaud C, Pierrat V, et al. Cerebral palsy among very preterm children in relation to gestational age and neonatal ultrasound abnormalities: the EPIPAGE cohort study. Pediatrics 2006;117:828-35. 4. Larroque B, Ancel PY, Marret S, Marchand L, Andre M, Arnaud C, et al. Neurodevelopmental disabilities and special care of 5-year-old children born before 33 weeks of gestation (the EPIPAGE study): a longitudinal cohort study. Lancet 2008;371:813-20. 5. Holling EE, Leviton A. Characteristics of cranial ultrasound whitematter echolucencies that predict disability: a review. Dev Med Child Neurol 1999;41:136-9. 6. Wu YW, Colford JM Jr. Chorioamnionitis as a risk factor for cerebral palsy: a meta-analysis. JAMA 2000;284:1417-24. 7. Menon R, Fortunato SJ. Infection and the role of inflammation in preterm premature rupture of the membranes. Best Pract Res Clin Obstet Gynaecol 2007;21:467-78. 8. Ancel PY, Marret S, Larroque B, Arnaud C, Zupan-Simunek V, Voyer M, et al. Are maternal hypertension and small-for-gestational age risk
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
439
THE JOURNAL OF PEDIATRICS
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19. 20.
440
www.jpeds.com
factors for severe intraventricular hemorrhage and cystic periventricular leukomalacia? Results of the EPIPAGE cohort study. Am J Obstet Gynecol 2005;193:178-84. Bauer M, Fast C, Haas J, Resch B, Lang U, Pertl B. Cystic periventricular leukomalacia in preterm infants: an analysis of obstetric risk factors. Early Hum Dev 2009;85:163-9. Spinillo A, Capuzzo E, Stronati M, Ometto A, De Santolo A, Acciano S. Obstetric risk factors for periventricular leukomalacia among preterm infants. Br J Obstet Gynaecol 1998;105:865-71. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75-84. Drougia A, Giapros V, Krallis N, Theocharis P, Nikaki A, Tzoufi M, et al. Incidence and risk factors for cerebral palsy in infants with perinatal problems: a 15-year review. Early Hum Dev 2007;83:541-7. Murphy DJ, Sellers S, MacKenzie IZ, Yudkin PL, Johnson AM. Casecontrol study of antenatal and intrapartum risk factors for cerebral palsy in very preterm singleton babies. Lancet 1995;346:1449-54. Beaino G, Khoshnood B, Kaminski M, Pierrat V, Marret S, Matis J, et al. Predictors of cerebral palsy in very preterm infants: the EPIPAGE prospective population-based cohort study. Dev Med Child Neurol 2012; 52:e119-25. Kurkinen-Raty M, Koivisto M, Jouppila P. Perinatal and neonatal outcome and late pulmonary sequelae in infants born after preterm premature rupture of membranes. Obstet Gynecol 1998;92:408-15. Park SH, Kim HJ, Yang JH, Choi JS, Lim JE, Oh MJ, et al. Neonatal brain damage following prolonged latency after preterm premature rupture of membranes. J Korean Med Sci 2006;21:485-9. Spinillo A, Capuzzo E, Stronati M, Ometto A, Orcesi S, Fazzi E. Effect of preterm premature rupture of membranes on neurodevelopmental outcome: follow-up at two years of age. Br J Obstet Gynaecol 1995;102:882-7. Larroque B, Breart G, Kaminski M, Dehan M, Andre M, Burguet A, et al. Survival of very preterm infants: EPIPAGE, a population-based cohort study. Arch Dis Child Fetal Neonatal Ed 2004;89:F139-44. Kaufman AS, Kaufman NL. Kaufman Assessment Battery for Children. Circle Pines (MN): American Guidance Service; 1983. World Health Organization. Multiaxial classification of child and adolescent psychiatric disorders: the ICD-10 classification of mental and behavioral disorders in children and adolescents. Geneva, Switzerland: World Health Organization; 1996.
Vol. 163, No. 2 21. Schafer JL. Analysis of incomplete multivariate data. New York: Chapman and Hall; 1997. 22. Yuan YC. Multiple imputation for missing data: concepts and new development (version 9.0). Rockville, MD: SAS Institute. Available from: http://www2.sas.com/proceedings/sugi25/25/st/25p267.pdf. Accessed March 4, 2013. 23. Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005;352:9-19. 24. Wolke D, Meyer R. Cognitive status, language attainment, and prereading skills of 6-year-old very preterm children and their peers: the Bavarian Longitudinal Study. Dev Med Child Neurol 1999;41: 94-109. 25. Hille ET, Elbertse L, Gravenhorst JB, Brand R, Verloove-Vanhorick SP. Nonresponse bias in a follow-up study of 19-year-old adolescents born as preterm infants. Pediatrics 2005;116:e662-6. 26. Blumenfeld YJ, Lee HC, Gould JB, Langen ES, Jafari A, El-Sayed YY. The effect of preterm premature rupture of membranes on neonatal mortality rates. Obstet Gynecol 2010;16:1381-6. 27. McElrath TF, Allred EN, Leviton A. Prolonged latency after preterm premature rupture of membranes: an evaluation of histologic condition and intracranial ultrasonic abnormality in the neonate born at <28 weeks of gestation. Am J Obstet Gynecol 2003;189:794-8. 28. Manuck TA, Maclean CC, Silver RM, Varner MW. Preterm premature rupture of membranes: does the duration of latency influence perinatal outcomes? Am J Obstet Gynecol 2009;201:414.e1-e6. 29. Hack M, Taylor HG, Drotar D, Schluchter M, Cartar L, WilsonCostello D, et al. Poor predictive validity of the Bayley Scales of Infant Development for cognitive function of extremely low birth weight children at school age. Pediatrics 2005;116:333-41. 30. Pasquier JC, Bujold E, Rabilloud M, Picaud JC, Ecochard R, Claris O, et al. Effect of latency period after premature rupture of membranes on 2 years infant mortality (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2007;135:21-7. 31. Pasquier JC, Picaud JC, Rabilloud M, Claris O, Ecochard R, Moret S, et al. Neonatal outcomes after elective delivery management of preterm premature rupture of the membranes before 34 weeks’ gestation (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2009;143:18-23.
Mura et al
ORIGINAL ARTICLES
August 2013
Appendix Members of the EPIPAGE study group include: Institut National de la Sante et de la Recherche Medicale U149: B. Larroque (national coordinator), P. Y. Ancel, B. Blondel, G. Breart, M. Dehan, M. Garel, M. Kaminski, F. Maillard, C. du Mazaubrun, P. Missy, F. Sehili, K. Supernant, L. Marchand Alsace: M. Durand, J. Matis, J. Messer, A. Treisser (Hopital de Hautepierre, Strasbourg) Franche-Comte: A. Burguet, L. Abraham-Lerat, A. Menget, P. Roth, J.-P. Schaal, G. Thiriez (CHU St Jacques, Besancon) Haute-Normandie: C. Leveque, S. Marret, L. Marpeau (Hopital Charles Nicolle, Rouen) Languedoc-Roussillon: P. Boulot, G. Cambonie (Hopital Arnaud de Villeneuve, Montpellier), A.-M. Donadio, B. Ledesert (ORS Montpellier)
Lorraine: M. Andre, J. Fresson, J. M. Hascoet (Maternite Regionale, Nancy) Midi-Pyrenees: C. Arnaud, S. Bourdet-Loubere, H. Grandjean (Institut National de la Sante et de la Recherche Medicale U558, Toulouse), M. Rolland (Hopital des Enfants, Toulouse) Nord-Pas-de-Calais: C. Leignel, P. Lequien, V. Pierrat, F. Puech, D. Subtil, P. Truffert (Hopital Jeanne de Flandre, Lille) Pays-de-Loire: G. Boog, V. Rouger-Bureau, J.-C. Roze (Hopital Mere-Enfant, Nantes) Paris-Petite-Couronne: P. Y. Ancel, G. Breart, M. Kaminski, C. du Mazaubrun (Institut National de la Sante et de la Recherche Medicale U149, Paris), M. Dehan, V. ZupanSimunek (Hopital Antoine Beclere, Clamart), M. Vodovar, M. Voyer (Institut de Puericulture, Paris).
Cognitive Impairment at Age 5 Years in Very Preterm Infants Born Following Premature Rupture of Membranes
440.e1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 163, No. 2
Figure. Flow chart of the study cohort. *In 2 regions, follow-up was proposed for only 1 of every 2 infants born at exactly 32 weeks gestational age. **Including 31 of 253 infants (12%) with a latency period >3 days, 31 of 286 infants (11%) with a latency period <3 days (P = .60) and 7 of 38 infants with no information available on latency period.*** Among the 494 survivors with no available K-ABC MPC score, 8 did not complete the K-ABC because of a sensory or motor disability or a language disorder. The distribution of these children was not significantly associated with the etiology of prematurity (P = .63).
440.e2
Mura et al