The objectives were to describe mortality and causes of death in children with intraventricular hemorrhage (IVH) and to study neurodevelopmental outcomes.
The study was a secondary analysis of the French national prospective and population-based cohort EPIPAGE-2. Children were recruited in 2011. A standardized assessment was conducted at age 5. Children born before 32 weeks’ gestation and admitted to a NICU were eligible. Exposure was IVH defined by the Papile classification. Main outcomes were mortality, causes of death, and neurodevelopmental outcomes at age 5.
Among the 3468 children included, 578 (16.7%) had grade 1 IVH, 424 (12.2%) grade 2 IVH, and 114 (3.3%) grade 3 IVH; 144 (4.1%) had intraparenchymal hemorrhage (IPH). Mortality was 29.7% (36 of 114) for children with grade 3 IVH and 74.4% (109 of 144) for those with IPH; 67.6% (21 of 31) and 88.7% (86 of 97) of deaths, respectively, were because of withholding and withdrawing of life-sustaining treatment. As compared with no IVH, low-grade IVH was not associated with measured neurodevelopmental disabilities at age 5. High-grade IVH was associated with moderate and severe neurodevelopmental disabilities, reduced full-scale IQ, and cerebral palsy.
Rates of neurodevelopmental disabilities at age 5 did not differ between children without IVH and those with low-grade IVH. For high-grade IVH, mortality rate was high, mostly because of withholding and withdrawal of life-sustaining treatment, and we found a strong association with overall neurodevelopmental disabilities in survivors.
The literature about neurodevelopmental consequences of low-grade intraventricular hemorrhage is contradictory. For high-grade intraventricular hemorrhage, few studies are available on the prevalence and circumstances of deaths, including those consecutive to withholding and withdrawal of life-sustaining treatment.
Caregivers can be reassured when they inform parents about neurodevelopmental consequences of low-grade intraventricular hemorrhage. For high-grade intraventricular hemorrhage, studies should always investigate in-hospital mortality rate and causes of death to interpret the neurodevelopmental outcomes.
Very preterm children are at high risk of neurodevelopmental disabilities,1 and intraventricular hemorrhage (IVH) is a major risk factor. This cerebral injury can be detected by systematic cranial ultrasonography (cUS) during the early neonatal period and is usually described according to the Papile classification.2 In the most at-risk population of children born before 25 weeks’ gestation, 24% and 45% experience high-grade IVH (grade 3 IVH and intraparenchymal hemorrhage [IPH]) and IVH of any grade, respectively.3 Even in children born less premature, IVH remains frequent. Among children born before 28 and between 28 and 31 weeks’ gestation, the incidence of high-grade IVH was 15% and 5%, respectively, and that of IVH of any grade was 34% and 17%, respectively.3 Knowing the neurodevelopmental prognosis of children with IVH-IPH is crucial to guide medical decisions and to inform parents.
However, the consequences of low-grade IVH (grades 1 and 2 IVH) are still controversial. Despite the absence of direct injury to the brain parenchyma, low-grade IVH could be responsible for abnormal brain development via damage to maturing glial and neuronal precursor cells in the germinal matrix.4 Some studies showed an association of low-grade IVH with neurodevelopmental disabilities5–7 and others did not.8–10 In addition, most studies focused on neurodevelopment at 2 years,6–9 and few of those on later neurodevelopment were multicenter and prospective cohorts.11 Concerning high-grade IVH, neurodevelopmental consequences may be cerebral palsy and cognitive impairment.6 Increased in-hospital mortality is also reported,12 but few studies are available on the prevalence and circumstances of death, including those consecutive to withholding and withdrawal of life-sustaining treatment (WWLST). However, because of possible selection by mortality, these are essential elements for better understanding the long-term outcomes of children experiencing high-grade IVH.
The objectives of our study were to describe mortality and causes of death in preterm children born before 32 weeks’ gestation with IVH-IPH and to study neurodevelopmental outcomes in surviving children with low- or high-grade IVH in a prospective population-based cohort assessed at age 5.
Methods
Study Design and Participants
This study was a secondary analysis of the French national population-based and prospective cohort EPIPAGE-2 (Etude éPIdémiologique sur les Petits Ages GEstationnels-2). The study protocol, neonatal mortality and morbidity, and age-5 results for the overall cohort have been published.1,13–15 Children born before 32 weeks’ gestation without cerebral congenital malformation or chromosomal abnormality and admitted in NICUs were eligible. Children with missing IVH status and cystic periventricular leukomalacia were excluded.
The analyses at age 5 were approved by the National Data Protection Authority (CNIL DR-2016-290) and by appropriate ethics committees.
Measurement of IVH
The IVH-IPH grade was assigned according to the Papile classification.2 In the EPIPAGE-2 study, cUS was performed according to the unit protocols, and data were collected prospectively. For most children, at least 1 cUS was performed in the first 8 days of life, then another was performed before day 21 of life, and at about 36 weeks’ postmenstrual age. We considered the highest IVH grade diagnosed among all cUS measurements performed during the neonatal stay. In case of bilateral hemorrhage, the grade of the most severely injured side was retained. Quality of data were subjected to indirect controls. We examined the proportion of children affected by each grade of IVH by center to detect any abnormalities. No outlier center was found. For a previous study,16 all diagnosed grade 3 IVH and IPH cases were centrally reviewed to confirm the consistency of the cUS results.
Outcomes
Outcomes were in-hospital mortality, neurodevelopmental disabilities at age 5 in surviving children, and disability-free survival. Between age 5 and 6, children had a standard assessment by trained pediatricians and neuropsychologists. Parents also completed questionnaires. A questionnaire was mailed to parents if the center-based assessment was not possible. Motor evaluation was performed with the Surveillance of Cerebral Palsy in Europe description of cerebral palsy classified according to the Gross Motor Function Classification System17 and with the Movement Assessment Battery for Children, second edition.18 For behavioral difficulties, the Strengths and Difficulties Questionnaire19 was used. Cognitive impairment was assessed with the Wechsler Preschool and Primary Scale of Intelligence, fourth edition.20 Children also underwent visual assessment with the Sander-Zanlonghi scale21 and hearing assessment. We built a composite variable describing overall neurodevelopmental disabilities that included cerebral palsy, developmental coordination disorders, full-scale IQ, behavioral difficulties, vision, and hearing.1 Diagnoses of epilepsy and autism spectrum disorders were reported in the parental questionnaires. Developmental support was defined as at least 2 visits during the preceding 12 months to a psychologist, psychiatrist, orthoptist, speech therapist, psychomotor therapist, or physiotherapist; care in a rehabilitation center; or special school needs. For the neurodevelopmental tests, reference data were derived from contemporary term-born children from the ELFE (Étude Longitudinale Française depuis l’Enfance) cohort22 assessed with the EPIPAGE-2 follow-up protocol as described in a previous study.1
Statistical Analyses
Perinatal characteristics were reported according to IVH status. Mortality, neurodevelopmental outcomes at age 5 and disability-free survival are described by IVH status in bivariable analyses. Statistical significance was set at P < .05. Because children born at 22 to 26 and 27 to 31 weeks’ gestation were recruited by an 8-month and a 6-month period, respectively, percentages and means are weighted by recruitment time in descriptive tables (weighting factor 1.35 for children born at 27 to 31 weeks’ gestation).13 The association between IVH and outcomes at age 5 was assessed with logistic, multinomial, and linear regression analyses for binary, categorical, and continuous outcomes, respectively. The models were fitted with the generalized estimating equation to consider multiple pregnancies. In the multivariable models, IVH was grouped into 3 classes (no IVH; low-grade IVH, ie, IVH grades 1 and 2; high-grade IVH, ie, IVH grade 3 and IPH) according to international studies and pathophysiology.2 The reference group consisted of children without IVH. Adjustment variables were selected with a directed acyclic graph23 and were sex, gestational age, inborn or outborn status, premature rupture of membrane, complete antenatal corticosteroids course, small-for-gestational-age, hemodynamic disorders in the first 3 days after birth, and sociodemographic variables (maternal country of birth and maternal educational level). At age 5, complete medical and neuropsychological assessments were available for 1892 of 3129 (60.5%) of surviving children.
Multiple imputations by chain equation (R package MICE24 ) were used to decrease the risk of attrition bias related to lost to follow-up. All variables in the imputation model and percentages of missing data are detailed in Supplemental Table 4. A total of 60 datasets were imputed, with 40 iterations for each dataset. Neurodevelopmental outcomes at age 5 for surviving children and disability-free survival were reported after imputation. Nonimputed neurodevelopmental outcomes at age 5 are described in the Supplemental Information. Analyses were performed on each of the imputed datasets, and the results were combined according to Rubin’s rules.25 To assess the impact of multiple imputations, the perinatal characteristics of children added by multiple imputation were compared with those of children with complete age-5 assessment. Sensitivity analyses were performed with complete cases.
All statistical analyses were performed with R 4.1.2 (R Core team).
Results
The eligible population consisted of 3646 children admitted to a NICU (Fig 1); 94 (2.6%) were excluded for missing IVH status and 84 (2.3%) for cystic periventricular leukomalacia. Among the 3468 remaining children, 578 (16.7%) had grade 1 IVH, 424 (12.2%) grade 2 IVH, 114 (3.3%) grade 3 IVH, and 144 (4.1%) IPH. The mean gestational age in each group was 28.7 (SD 2.1, range: 24.0–31.9), 27.8 (SD 2.1, range: 23.7–31.9), 27.4 (SD 2.1, range: 24.0–31.9) and 26.6 (SD 2.0, range: 23.1–31.9) weeks’ gestation. For children with no IVH, mean gestational age was 29.4 (SD 2.0, range: 22.9–31.9) weeks’ gestation. Perinatal characteristics are described in Supplemental Table 5. Among the 39.5% (1237/3129) of children alive who did not have a complete 5-year assessment, partial information was available for 27.8% (344/1237), from the mailed questionnaire for 20.8% (258/1237) or from partial assessment for 7.0% (82/1237) (Fig 1). Comparison of children with complete assessment at age 5 with those added in the analyses by multiple imputation revealed that more patients without any IVH and with lower socioeconomic status were imputed (Supplemental Table 6).
Mortality Rate
Mortality in the absence of IVH was 4.0% (99 of 2208) and increased from 6.7% (43 of 578) for children with grade 1 IVH to 74.4% (109 of 144) for those with IPH (Table 1). Death occurred mostly during the neonatal stay and was secondary to WWLST in 67.6% (21 of 31) of children with grade 3 IVH and in 88.7% (86 of 97) of those with IPH. None of the children with grade 1 and 2 IVH died of central nervous system injury. Conversely, central nervous system injury was the main cause of death in 42.4% (15 of 35) of children with grade 3 IVH and 73.5% (80 of 109) with IPH. In children with high-grade IVH, in-hospital death secondary to WWLST was associated with low gestational age, absence of antenatal corticosteroid therapy, early hemodynamic disorders, and central nervous system injury identified as main cause of death (Supplemental Table 7).
Mortality Rate and Cause of Death According to Intraventricular Hemorrhage Status
| Population Study, N = 3468 | ||||||
|---|---|---|---|---|---|---|
| No IVH, N = 2208 (63.7%) | Grade 1 IVH, N = 578 (16.7%) | Grade 2 IVH, N = 424 (12.2%) | Grade 3 IVH, N = 114 (3.3%) | IPH, N = 144 (4.1%) | Pa | |
| All children | ||||||
| Mortality rate, n/N (%b) | 99/2208 (4.0) | 43/578 (6.7) | 52/424 (11.1) | 36/114 (29.7) | 109/144 (74.4) | <.001 |
| In-hospital death | 92/99 (92.4) | 37/43 (85.1) | 47/52 (89.5) | 35/36 (97.5) | 109/109 (100.0) | .003 |
| After discharge death | 7/99 (7.6) | 6/43 (14.9) | 5/52 (10.5) | 1/36 (2.5) | 0/109 (0.0) | |
| Among children who died in a NICU (N = 320): | ||||||
| Death secondary to WWLST, n/N (%b) | 29/86 (32.7) | 19/33 (57.5) | 15/43 (33.3) | 21/31 (67.6) | 86/97 (88.7) | <.001 |
| Postnatal age at death in days, mean (SDb) | 16.4 (24.7) | 43.5 (64.8) | 25.6 (43.6) | 34.1 (66.7) | 10.7 (11.2) | <.001 |
| Cause-specificc mortality rate, n/N (%b) | ||||||
| Respiratory distress syndrome | 41/92 (42.5) | 19/37 (51.6) | 20/47 (42.2) | 9/35 (25.4) | 23/109 (20.9) | <.001 |
| Necrotizing enterocolitis | 10/92 (11.7) | 4/37 (11.9) | 5/47 (11.7) | 1/35 (2.5) | 0/109 (0.0) | |
| Late-onset sepsis | 17/91 (18.9) | 5/37 (13.5) | 13/47 (26.6) | 4/35 (11.9) | 2/109 (2.0) | |
| Central nervous system injury | 0/92 (0.0) | 0/37 (0.0) | 0/47 (0.0) | 15/35 (42.4) | 80/109 (73.5) | |
| Other | 18/92 (20.9) | 7/37 (17.4) | 5/47 (10.4) | 4/35 (11.0) | 2/109 (2.0) | |
| Unknown | 6/92 (6.0) | 2/37 (5.6) | 4/47 (9.1) | 2/35 (6.8) | 2/109 (1.6) | |
| Population Study, N = 3468 | ||||||
|---|---|---|---|---|---|---|
| No IVH, N = 2208 (63.7%) | Grade 1 IVH, N = 578 (16.7%) | Grade 2 IVH, N = 424 (12.2%) | Grade 3 IVH, N = 114 (3.3%) | IPH, N = 144 (4.1%) | Pa | |
| All children | ||||||
| Mortality rate, n/N (%b) | 99/2208 (4.0) | 43/578 (6.7) | 52/424 (11.1) | 36/114 (29.7) | 109/144 (74.4) | <.001 |
| In-hospital death | 92/99 (92.4) | 37/43 (85.1) | 47/52 (89.5) | 35/36 (97.5) | 109/109 (100.0) | .003 |
| After discharge death | 7/99 (7.6) | 6/43 (14.9) | 5/52 (10.5) | 1/36 (2.5) | 0/109 (0.0) | |
| Among children who died in a NICU (N = 320): | ||||||
| Death secondary to WWLST, n/N (%b) | 29/86 (32.7) | 19/33 (57.5) | 15/43 (33.3) | 21/31 (67.6) | 86/97 (88.7) | <.001 |
| Postnatal age at death in days, mean (SDb) | 16.4 (24.7) | 43.5 (64.8) | 25.6 (43.6) | 34.1 (66.7) | 10.7 (11.2) | <.001 |
| Cause-specificc mortality rate, n/N (%b) | ||||||
| Respiratory distress syndrome | 41/92 (42.5) | 19/37 (51.6) | 20/47 (42.2) | 9/35 (25.4) | 23/109 (20.9) | <.001 |
| Necrotizing enterocolitis | 10/92 (11.7) | 4/37 (11.9) | 5/47 (11.7) | 1/35 (2.5) | 0/109 (0.0) | |
| Late-onset sepsis | 17/91 (18.9) | 5/37 (13.5) | 13/47 (26.6) | 4/35 (11.9) | 2/109 (2.0) | |
| Central nervous system injury | 0/92 (0.0) | 0/37 (0.0) | 0/47 (0.0) | 15/35 (42.4) | 80/109 (73.5) | |
| Other | 18/92 (20.9) | 7/37 (17.4) | 5/47 (10.4) | 4/35 (11.0) | 2/109 (2.0) | |
| Unknown | 6/92 (6.0) | 2/37 (5.6) | 4/47 (9.1) | 2/35 (6.8) | 2/109 (1.6) | |
IPH, intraparenchymal hemorrhage; IVH, intraventricular hemorrhage; WWLST, withholding and withdrawal of life-sustaining treatments.
Pearson’s χ2 test for categorical variables and linear-model ANOVA for continuous variables.
All percentages and means were weighted over the recruitment time (weighting factor 1.35 for children born at 27 to 31-wk’ gestation).
Specific cause of death were defined according to Patel et al. N Eng J Med 2015.
Neurodevelopmental Disabilities Among Surviving Children
At age 5, severe overall neurodevelopmental disabilities and cerebral palsy rates increased with IVH severity. Mean full-scale IQ decreased with IVH severity. The imputed and nonimputed neurodevelopmental outcomes at age 5 are detailed in Table 2 and in Supplemental Table 8.
Neurodevelopmental Outcomes by Intraventricular Hemorrhage Status, Imputed Data
| Neurodevelopmental Outcomes at Age 5 | Children Alive at Age 5, N = 3129, Imputed Data | |||||
|---|---|---|---|---|---|---|
| No IVH, N = 2109 (67.4%) | Grade 1 IVH, N = 535 (17.1%) | Grade 2 IVH, N = 372 (11.9%) | Grade 3 IVH, N = 78 (2.5%) | IPH, N = 35 (1.1%) | Pa | |
| Overall neurodevelopmental disabilitiesb, % (95% CI)c | ||||||
| None | 47.8 (47.5–48.0) | 44.2 (43.7–44.7) | 45.6 (45.0–46.1) | 36.6 (35.4–37.9) | 45.6 (43.6–45.6) | <.001 |
| Mild | 35.4 (35.2–35.7) | 35.7 (35.2–36.1) | 35.0 (34.5–35.6) | 25.0 (23.9–26.2) | 18.9 (17.3–20.5) | |
| Moderate | 10.2 (10.1–10.4) | 11.0 (10.7–11.3) | 10.6 (10.2–10.9) | 23.7 (20.6–24.8) | 15.3 (14.0–16.8) | |
| Severe | 6.6 (6.5–6.7) | 9.1 (8.9–9.4) | 8.8 (8.5–9.2) | 14.7 (13.8–15.6) | 20.2 (18.6–21.8) | |
| Developmental coordination disordersd (N= 1534) | ||||||
| Total MABC-2 score ≤ fifth centilee, % (95% CI)c | 8.2 (8.1–8.4) | 10.2 (9.9–10.5) | 13.8 (13.4–14.3) | 13.0 (11.9–14.1) | 7.9 (6.6–9.5) | <.001 |
| Behavioral difficulties: | ||||||
| Total SDQ score ≥ 90th centilee, % (95% CI)c | 10.7 (10.5–10.8) | 12.1 (11.7–12.3) | 11.3 (10.9–11.7) | 9.9 (9.1–10.7) | 10.1 (9.0–11.4) | .80 |
| Full-scale IQf | ||||||
| Mean (SD) | 94.7 (16) | 93.6 (16) | 93.1 (16) | 89.1 (18) | 88.4 (20) | .04 |
| Total IQ score < −2 SDe (< 79), % (95% CI)c | 13.8 (13.7–14.0) | 15.8 (15.5–16.2) | 17.2 (16.8–17.7) | 24.8 (23.7–25.9) | 31.1 (29.2–32.9) | <.001 |
| Moderate or severe visual disabilityg, % (95% CI)c | 1.2 (1.1–1.2) | 1.9 (1.8–2.0) | 1.9 (1.7–2.1) | 7.5 (6.9–8.3) | 5.5 (4.7–6.5) | .009 |
| Moderate or severe hearing disabilityh, % (95% CI)c | 0.9 (0.8–0.9) | 1.4 (1.3–1.5) | 0.8 (0.7–0.9) | 3.8 (3.3–4.3) | 0.0 (0.0–0.0) | .09 |
| Cerebral palsy | ||||||
| All cerebral palsy, % (95% CI)c | 6.2 (6.1–6.3) | 8.9 (8.6–9.1) | 6.2 (5.9–6.1) | 19.7 (18.7–20.8) | 29.7 (27.9–31.5) | <.001 |
| By categories, % (95% CI)c | ||||||
| No cerebral palsy | 93.8 (93.7–93.9) | 91.1 (90.9–91.4) | 93.8 (93.5–94.1) | 80.3 (79.2–81.3) | 70.3 (68.5–72.1) | <.001 |
| Cerebral palsy GMFCS-1 | 2.3 (2.2–2.4) | 4.1 (3.9–4.3) | 2.1 (1.9–2.2) | 1.9 (1.5–2.2) | 13.4 (12.1–14.8) | |
| Cerebral palsy GMFCS-2/5 | 3.9 (3.8–4.0) | 4.7 (4.5–4.9) | 4.1 (3.9–4.3) | 17.9 (16.9–18.9) | 16.3 (14.9–17.8) | |
| Epilepsy, % (95% CI)c | 1.6 (1.5–1.7) | 2.4 (2.2–2.5) | 2.6 (2.4–2.8) | 4.5 (4.0–5.1) | 6.4 (5.5–7.4) | <.001 |
| Autism spectrum disorders reported by parents, % (95% CI)c | 1.5 (1.4–1.5) | 2.9 (2.8–3.1) | 2.2 (2.1–2.4) | 4.5 (4.0–5.1) | 2.3 (1.7–2.9) | <.001 |
| Developmental supporti, % (95% CI)c | 40.5 (40.3–40.7) | 42.5 (42.1–43.0) | 45.2 (44.6–45.8) | 54.1 (52.8–55.4) | 52.9 (51.0–54.9) | .02 |
| Neurodevelopmental Outcomes at Age 5 | Children Alive at Age 5, N = 3129, Imputed Data | |||||
|---|---|---|---|---|---|---|
| No IVH, N = 2109 (67.4%) | Grade 1 IVH, N = 535 (17.1%) | Grade 2 IVH, N = 372 (11.9%) | Grade 3 IVH, N = 78 (2.5%) | IPH, N = 35 (1.1%) | Pa | |
| Overall neurodevelopmental disabilitiesb, % (95% CI)c | ||||||
| None | 47.8 (47.5–48.0) | 44.2 (43.7–44.7) | 45.6 (45.0–46.1) | 36.6 (35.4–37.9) | 45.6 (43.6–45.6) | <.001 |
| Mild | 35.4 (35.2–35.7) | 35.7 (35.2–36.1) | 35.0 (34.5–35.6) | 25.0 (23.9–26.2) | 18.9 (17.3–20.5) | |
| Moderate | 10.2 (10.1–10.4) | 11.0 (10.7–11.3) | 10.6 (10.2–10.9) | 23.7 (20.6–24.8) | 15.3 (14.0–16.8) | |
| Severe | 6.6 (6.5–6.7) | 9.1 (8.9–9.4) | 8.8 (8.5–9.2) | 14.7 (13.8–15.6) | 20.2 (18.6–21.8) | |
| Developmental coordination disordersd (N= 1534) | ||||||
| Total MABC-2 score ≤ fifth centilee, % (95% CI)c | 8.2 (8.1–8.4) | 10.2 (9.9–10.5) | 13.8 (13.4–14.3) | 13.0 (11.9–14.1) | 7.9 (6.6–9.5) | <.001 |
| Behavioral difficulties: | ||||||
| Total SDQ score ≥ 90th centilee, % (95% CI)c | 10.7 (10.5–10.8) | 12.1 (11.7–12.3) | 11.3 (10.9–11.7) | 9.9 (9.1–10.7) | 10.1 (9.0–11.4) | .80 |
| Full-scale IQf | ||||||
| Mean (SD) | 94.7 (16) | 93.6 (16) | 93.1 (16) | 89.1 (18) | 88.4 (20) | .04 |
| Total IQ score < −2 SDe (< 79), % (95% CI)c | 13.8 (13.7–14.0) | 15.8 (15.5–16.2) | 17.2 (16.8–17.7) | 24.8 (23.7–25.9) | 31.1 (29.2–32.9) | <.001 |
| Moderate or severe visual disabilityg, % (95% CI)c | 1.2 (1.1–1.2) | 1.9 (1.8–2.0) | 1.9 (1.7–2.1) | 7.5 (6.9–8.3) | 5.5 (4.7–6.5) | .009 |
| Moderate or severe hearing disabilityh, % (95% CI)c | 0.9 (0.8–0.9) | 1.4 (1.3–1.5) | 0.8 (0.7–0.9) | 3.8 (3.3–4.3) | 0.0 (0.0–0.0) | .09 |
| Cerebral palsy | ||||||
| All cerebral palsy, % (95% CI)c | 6.2 (6.1–6.3) | 8.9 (8.6–9.1) | 6.2 (5.9–6.1) | 19.7 (18.7–20.8) | 29.7 (27.9–31.5) | <.001 |
| By categories, % (95% CI)c | ||||||
| No cerebral palsy | 93.8 (93.7–93.9) | 91.1 (90.9–91.4) | 93.8 (93.5–94.1) | 80.3 (79.2–81.3) | 70.3 (68.5–72.1) | <.001 |
| Cerebral palsy GMFCS-1 | 2.3 (2.2–2.4) | 4.1 (3.9–4.3) | 2.1 (1.9–2.2) | 1.9 (1.5–2.2) | 13.4 (12.1–14.8) | |
| Cerebral palsy GMFCS-2/5 | 3.9 (3.8–4.0) | 4.7 (4.5–4.9) | 4.1 (3.9–4.3) | 17.9 (16.9–18.9) | 16.3 (14.9–17.8) | |
| Epilepsy, % (95% CI)c | 1.6 (1.5–1.7) | 2.4 (2.2–2.5) | 2.6 (2.4–2.8) | 4.5 (4.0–5.1) | 6.4 (5.5–7.4) | <.001 |
| Autism spectrum disorders reported by parents, % (95% CI)c | 1.5 (1.4–1.5) | 2.9 (2.8–3.1) | 2.2 (2.1–2.4) | 4.5 (4.0–5.1) | 2.3 (1.7–2.9) | <.001 |
| Developmental supporti, % (95% CI)c | 40.5 (40.3–40.7) | 42.5 (42.1–43.0) | 45.2 (44.6–45.8) | 54.1 (52.8–55.4) | 52.9 (51.0–54.9) | .02 |
GMFCS, Gross Motor Function Classification System; MABC-2, Movement Assessment Battery for Children, second edition; SCQ, Social Communication Questionnaire; SDQ, Strengths and Difficulties Questionnaire.
Pearson’s χ2 test for categorical variables and linear-model ANOVA for continuous variables.
Includes cerebral palsy, vision, hearing, full-scale IQ, developmental coordination disorders, and behavioral difficulties.
All percentages and means were weighted over the recruitment time (weighting factor 1.35 for children born at 27 to 31-wk’ gestation).
Among children without cerebral palsy, severe or moderate sensory disabilities, and with full-scale IQ > −2 SDs of the distribution related to the reference group born at term.
Cut-off point of the distribution related to the reference group of term-born children from the ELFE (Étude Longitudinale Française depuis l’Enfance) cohort assessed with the EPIPAGE 2 follow-up protocol.
Full-scale IQ, measured by the Wechsler Preschool and Primary Scale of Intelligence, fourth edition.
Moderate or severe visual disability was defined as binocular visual acuity <3.2 of 10.
Moderate or severe hearing disability was defined as uni- or bilateral hearing loss (>70 dB) not corrected or partially corrected with a hearing aid.
Developmental support was defined as at least 2 consultations with a psychologist, psychiatrist, orthoptist, speech therapist, occupational therapist, or physiotherapist during the 12 last months; follow-up in any type of rehabilitation center; or schooling in special school.
In imputed multivariable models and compared with the reference group of children without IVH, low-grade IVH was not associated with any measured neurodevelopmental disability (Table 3). Subgroup analysis of children born before 28 weeks’ gestation and complete-case sensitivity analyses (Supplemental Table 9) were in line with this result.
Association Between Intraventricular Hemorrhage and Neurodevelopmental Outcomes, Multivariable and Imputed Models
| Neurodevelopmental Outcomes at Age 5 | Children Alive at Age 5, N = 3129 | ||
|---|---|---|---|
| Grades 1 and 2 IVH, N = 907 | Grade 3 IVH and IPH, N= 113 | Pb | |
| aOR or Mean Differencea (95% CI) | aOR or Mean Differencea (95% CI) | ||
| Overall neurodevelopmental disabilitiesc | |||
| None | 1 | 1 | .04 |
| Mild | 0.98 (0.80 to 1.21) | 0.79 (0.45 to 1.38) | |
| Moderate | 1.05 (0.76 to 1.45) | 2.35 (1.21 to 4.55) | |
| Severe | 1.25 (0.85 to 1.85) | 2.28 (1.01 to 5.15) | |
| Developmental coordination disordersd | |||
| Total MABC-2 score ≤ fifth centilee | 1.28 (0.91 to 1.82) | 0.97 (0.32 to 2.92) | .38 |
| Behavioral difficulties: | |||
| Total SDQ score ≥ 90th centilee | 1.07 (0.79 to 1.45) | 0.86 (0.38 to 1.96) | .84 |
| Full-scale IQf | |||
| Total IQ score (mean difference) | −0.5 (−2.0 to 1.0) | −4.5 (−8.6 to −0.3) | .10 |
| Total IQ score < −2 SD (<79)e | 1.13 (0.85 to 1.50) | 2.04 (1.07 to 3.88) | .08 |
| Cerebral palsy | |||
| All cerebral palsy | 1.09 (0.75 to 1.61) | 3.45 (1.92 to 6.22) | <.001 |
| By categories | |||
| No cerebral palsy | 1 | 1 | .001 |
| Cerebral palsy GMFCS-1 | 1.18 (0.66 to 2.11) | 2.19 (0.84 to 5.71) | |
| Cerebral palsy GMFCS-2/5 | 1.04 (0.64 to 1.68) | 4.23 (2.11 to 8.49) | |
| Developmental supportg | 1.01 (0.84 to 1.22) | 1.38 (0.85 to 2.23) | .42 |
| Neurodevelopmental Outcomes at Age 5 | Children Alive at Age 5, N = 3129 | ||
|---|---|---|---|
| Grades 1 and 2 IVH, N = 907 | Grade 3 IVH and IPH, N= 113 | Pb | |
| aOR or Mean Differencea (95% CI) | aOR or Mean Differencea (95% CI) | ||
| Overall neurodevelopmental disabilitiesc | |||
| None | 1 | 1 | .04 |
| Mild | 0.98 (0.80 to 1.21) | 0.79 (0.45 to 1.38) | |
| Moderate | 1.05 (0.76 to 1.45) | 2.35 (1.21 to 4.55) | |
| Severe | 1.25 (0.85 to 1.85) | 2.28 (1.01 to 5.15) | |
| Developmental coordination disordersd | |||
| Total MABC-2 score ≤ fifth centilee | 1.28 (0.91 to 1.82) | 0.97 (0.32 to 2.92) | .38 |
| Behavioral difficulties: | |||
| Total SDQ score ≥ 90th centilee | 1.07 (0.79 to 1.45) | 0.86 (0.38 to 1.96) | .84 |
| Full-scale IQf | |||
| Total IQ score (mean difference) | −0.5 (−2.0 to 1.0) | −4.5 (−8.6 to −0.3) | .10 |
| Total IQ score < −2 SD (<79)e | 1.13 (0.85 to 1.50) | 2.04 (1.07 to 3.88) | .08 |
| Cerebral palsy | |||
| All cerebral palsy | 1.09 (0.75 to 1.61) | 3.45 (1.92 to 6.22) | <.001 |
| By categories | |||
| No cerebral palsy | 1 | 1 | .001 |
| Cerebral palsy GMFCS-1 | 1.18 (0.66 to 2.11) | 2.19 (0.84 to 5.71) | |
| Cerebral palsy GMFCS-2/5 | 1.04 (0.64 to 1.68) | 4.23 (2.11 to 8.49) | |
| Developmental supportg | 1.01 (0.84 to 1.22) | 1.38 (0.85 to 2.23) | .42 |
aOR, adjusted odds ratio; GMFCS, Gross Motor Function Classification System; MABC-2, Movement Assessment Battery for Children, second edition; SDQ, Strengths and Difficulties Questionnaire.
Generalized estimating equation for multinomial or binomial outcome model to account for nonindependence of outcomes related to multiple births or linear models for continuous outcomes. Reference category for the exposure is children without any IVH. The models are imputed and multivariable models included the following variables: IVH (3 categories) + gestational age (4 categories) + sex (binomial) + inborn/outborn status (binomial) + small-for-gestational-age (binomial) + ≥ 1 complete antenatal corticosteroids cure (binomial) + premature rupture of membranes (binomial) + hemodynamic disorders (binomial) + maternal educational level (4 categories) + maternal country of birth (5 categories).
Wald test.
Including cerebral palsy, vision, hearing, full-scale IQ, developmental coordination disorders, and behavioral difficulties.
Among children without cerebral palsy, severe or moderate sensory disabilities, and with full-scale IQ >−2 SD of the distribution related to the reference group born at term.
Cut-off point of the distribution related to the reference group of term-born children from the ELFE (Étude Longitudinale Française depuis l’Enfance) cohort assessed with the EPIPAGE 2 follow-up protocol.
Full-scale intelligence quotients, measured by the Wechsler Preschool and Primary Scale of Intelligence, fourth edition.
Developmental support was defined as at least 2 consultations with a psychologist, psychiatrist, orthoptist, speech therapist, occupational therapist, or physiotherapist during the 12 last months; follow-up in any type of rehabilitation center; or schooling in special school.
In imputed multivariable models, high-grade IVH was associated with moderate (odds ratio [OR], 2.35; 95% CI, 1.21 to 4.55) and severe neurodevelopmental disabilities (OR, 2.28; 95% confidence interval [CI], 1.01 to 5.15]), reduced full-scale IQ (mean difference, −4.5; 95% CI, −8.6 to −0.3) and cerebral palsy overall (OR, 3.45; 95% CI, 1.92 to 6.22) (Table 3). Complete-case sensitivity analyses confirmed these results except for full-scale IQ (Supplemental Table 9).
Disability-free Survival
Disability-free survival decreased with IVH severity (no IVH: 45.8% [95% CI, 45.6% to 46.1%]; grade 1 IVH: 41.2% [95% CI, 40.7% to 41.6%]; grade 2 IVH: 40.5% [95% CI, 40.0% to 41.0%]; grade 3 IVH: 25.7% [95% CI, 24.7% to 26.7%]; IPH: 11.6% [95% CI, 11.0% to 12.2%], imputed data). In imputed and multivariable models, disability-free survival did not differ between children with low-grade IVH and those without IVH (OR, 0.97; 95% CI, 0.80 to 1.17) but was significantly lower for children with high-grade IVH (OR, 0.35; 95% CI, 0.23 to 0.52).
Discussion
This study provides crucial and robust data on survival and neurodevelopmental outcomes at age 5 of preterm children with IVH-IPH. Children with low-grade IVH had higher mortality rates than those without IVH, but deaths were caused by other organ failures than to central nervous system injuries. Furthermore, low-grade IVH was not associated with an increase in any measured neurodevelopmental disability at age 5. Among children with high-grade IVH, mortality rate was high (50%) and WWLST was frequent. At age 5, high-grade IVH was strongly associated with overall neurodevelopmental disabilities, reduced full-scale IQ, and cerebral palsy and inversely with disability-free survival.
The most important finding of this study was the absence of any measured neurodevelopmental disability at age 5 associated with low-grade IVH. In our analyses, neither overall neurodevelopmental disabilities nor development coordination disorders, behavioral difficulties, full-scale IQ, and cerebral palsy were associated with low-grade IVH after controlling for confounding factors. These findings confirmed the results from assessments at ages 2, 3, and 8 of a smaller number of children finding no neurodevelopmental consequence of low-grade IVH.8–10 A recent multicentric cohort study5 found low-grade IVH associated with cerebral palsy, but models were bivariable logistic regressions, and this association may be caused by confounding bias. Finally, as with all very preterm children, neurologic development should be monitored in children with low-grade IVH. However, in line with the results of this study, parents should be reassured concerning isolated low-grade IVH consequences.
Most of the data in the literature concerning high-grade IVH outcomes focused on cerebral palsy and little on neurocognitive outcomes.7 This study confirmed the strong association of high-grade IVH with cerebral palsy and found an association between high-grade IVH and reduced full-scale IQ as is described in a few articles.8,26 The lack of association of high-grade IVH with developmental coordination and behavioral disorders should be interpreted with caution because the number of children assessed for these outcomes was small, with possible lack of power to show a difference. Finally, because of high mortality rate, reporting neurodevelopmental disabilities and neonatal mortality or considering a combined outcome (disability-free survival) is essential when studying neurodevelopmental outcomes of children with high-grade IVH.
This study also evaluated the rate of WWLST in cases of high-grade IVH in France in 2011. Overall, 29.7% and 74.4% of children with grade 3 IVH and IPH died, and death was secondary to WWLST in 67.6% of children with grade 3 IVH and 87.4% of those with IPH. A study by Sheehan et al27 in the United States, with recruitment between 1993 and 2013, showed similar high WWLST rates in children with high-grade IVH. In this work, 32% of children with grade 3 IVH and 66% of those with IPH died, and death was secondary to WWLST in 80% of children with grade 3 IVH and 93% of those with IPH. However, there is great variation in WWLST practices between countries. Hellman et al showed that among very preterm children with severe neurologic injury born in Canada between 2008 and 2010, only 31% died,28 and this study found discussion regarding WWLST for 28% of very preterm children with severe neurologic injury. In our study, characteristics of children with high-grade IVH and WWLST decisions suggest that high clinical instability and poor prognosis may lead to more WWLST. The high rate of WWLST in France has probably decreased for unilateral and less severe IPH since 2011 because of studies showing that the severity of impairment is associated with bilateral injuries and the number of territories involved in the IPH.27 New approaches to end-of-life decisions in very preterm infants, with better involvement of parents,29 have also changed practices. Unfortunately, we lacked the data to document this. To guide the difficult decision to continue, withhold, or withdraw life-sustaining treatment in case of high-grade IVH, future studies should compare neurodevelopmental outcomes of surviving children with high-grade IVH across countries with different in-hospital mortality rates. Future research is needed to determine to what extent a decreased in-hospital mortality rate among children with high-grade IVH would increase neurodevelopmental disability rate.
Our study has some strengths. The first is related to the design of the EIPAGE-2 cohort. This is a national, prospective, population-based cohort with data collection starting from the antenatal period, so many known potential confounding factors, both antenatal and postnatal, could be considered. The sample size allowed for adjustment on all potential confounders, even though IVH of higher grades was rare among survivors at age 5. The second strength was related to the standardized 5-year assessment. The tests and information collected covered all aspects of the child’s development, including sensory, motor, behavioral and cognitive domains using standardized tools that allowed for comparison with other studies. By use of a parental questionnaire, data on developmental interventions were also available. Finally, we selected the confounders by using a directed acyclic graph after a thorough review of the literature. Use of a directed acyclic graph limits the risk of over-fitting or residual confounding bias.23 Because sociodemographic factors are strongly associated with neurodevelopmental outcomes,1 we decided to consider them in 2 variables: mother’s education level and country of birth.
However, the study has some limitations. Missing data rate at 5 years was 39.5% (1237 of 3129), which suggested possible attrition bias. In different preterm cohorts, the proportion of children lost to follow-up at 2 years is usually between 10% and 50%,30 so our proportion of lost to follow-up was within the expected range, especially since it was a 5-year follow-up. We considered missing data by multiple imputations. Because of the systematic socioeconomic differences between participating and nonparticipating children in cohorts of very preterm children, multiple imputations are necessary.30 We included in our imputation model all perinatal characteristics, including socioeconomic variables, 2-year follow-up data, and the available 5-year outcomes. Finally, we conducted sensitivity analyses with the unimputed data, and results were consistent with the main analyses. The second limitation of the study was the measurement of exposure with a risk of classification bias caused by errors in assessing IVH. According to the center, cUS was performed by a radiologist or a trained neonatologist, and there was no blinded review of the images when the data were collected. Nevertheless, the Papile classification is easy to apply and is widely recognized in all participating centers. Consistently, we did not find any outlier NICU in terms of IVH rate, and there was no abnormality in the perinatal characteristics associated with IVH in the whole cohort. Furthermore, all high-grade IVH diagnoses were centrally reviewed for a previous study to confirm the consistency of cUS.16 Finally, regrouping IVH into 3 classes (no IVH, low-grade IVH, high-grade IVH) indirectly decreased the risk of classification bias between grades 1 and 2 IVH and between grade 3 IVH and IPH. The third limitation was the lack of specificity in IVH diagnosis. Data on whether the IVH was unilateral or bilateral was not collected, nor was the location or size of the brain injury for IPH. These data are rarely reported in the literature, but they should be analyzed in future cohort studies. Otherwise, for bilateral IPH, the risk is to underestimate the association with neurodevelopmental disabilities.
Conclusions
No children with low-grade IVH died of central nervous system injury. Among survivors, low-grade IVH was not associated with any measured neurodevelopmental disability at age 5. In children with high-grade IVH, risk of death was high, mainly secondary to WWLST. High-grade IVH was strongly associated with cerebral palsy, reduced full-scale IQ, and overall neurodevelopmental disabilities. Further studies should compare neurodevelopmental outcomes of surviving children with high-grade IVH across countries with different in-hospital mortality rates to appreciate the potential impact of this injury in less selected populations.
Acknowledgments
We are grateful for the participation of all families of preterm children in the EPIPAGE-2 cohort study and for the cooperation of all maternity and neonatal units in France. We also thank the EPIPAGE-2 Study group – 5Y FUp and Laura Smales for editorial assistance. Members of EPIPAGE-2 Study group – 5Y FUp: Alsace: D Astruc, P Kuhn, B Langer, J Matis, C Ramousset; Aquitaine: X Haernandorena, P Chabanier, L Joly-Pedespan, M Rebola, MJ Costedoat, A Leguen, C Martin; Auvergne: B Lecomte, D Lemery, F Vendittelli, E Rochette; Basse-Normandie: G Beucher, M Dreyfus, B Guillois, V. Datin-Dorrière, Y Toure, D Rots; Bourgogne: A Burguet, S Couvreur, JB Gouyon, P Sagot, N Colas, A Franzin; Bretagne: J Sizun, A Beuchée, P Pladys, F Rouget, RP Dupuy, D Soupre, F Charlot, S Roudaut; Centre: A Favreau, E Saliba, L Reboul, E Aoustin; Champagne-Ardenne: N Bednarek, P Morville, V Verrière; Franche-Comté: G Thiriez, C Balamou, C Ratajczak; Haute-Normandie: L Marpeau, S Marret, C Barbier, N Mestre; Ile-de-France: G Kayem, X Durrmeyer, M Granier, M Ayoubi, O Baud, B Carbonne, L Foix L’Hélias, F Goffinet, PH Jarreau, D Mitanchez, P Boileau, C Duffaut, E Lorthe, L Cornu, R Moras, D Salomon, S Medjahed, K Ahmed; Languedoc-Roussillon: P Boulot, G Cambonie, H Daudé, A Badessi, N Tsaoussis, M Poujol; Limousin: A Bédu, F Mons, C Bahans; Lorraine: MH Binet, J Fresson, JM Hascoët, A Milton, O Morel, R Vieux, L Hilpert; Midi-Pyrénées: C Alberge, C Arnaud, C Vayssière, M Baron; Nord-Pas-de-Calais: ML Charkaluk, V Pierrat, D Subtil, P Truffert, S Akowanou, D Roche, M Thibaut; PACA et Corse: C D’Ercole, C Gire, U Simeoni, A Bongain, M Deschamps, M Zahed; Pays de Loire: B Branger, JC Rozé, N Winer, V Rouger, C Dupont, H Martin; Picardie: J Gondry, G Krim, B Baby, I Popov; Rhône-Alpes: M Debeir, O Claris, JC Picaud, S Rubio-Gurung, C Cans, A Ego, T Debillon, H Patural, A Rannaud; Guadeloupe: E Janky, A Poulichet, JM Rosenthal, E Coliné, C Cabrera; Guyane: A Favre, N Joly, Stouvenel A; Martinique: S Châlons, J Pignol, PL Laurence, V Lochelongue; La Réunion: PY Robillard, S Samperiz, D Ramful. Inserm UMR 1153: PY Ancel, H Asadullah, V Benhammou, B Blondel, A Brinis, ML Charkaluk, A Coquelin, V Delormel, M Durox, M Fériaud, L Foix-L’Hélias, F Goffinet, M Kaminski, G Kayem, K Khemache, B Khoshnood, C Lebeaux, E Lorthe, L Marchand-Martin, A Morgan, L Onestas, V Pierrat, J Rousseau, MJ Saurel-Cubizolles, D Tran, D Sylla, L Vasante-Annamale, J Zeitlin.
Dr Treluyer conducted the analyses and drafted the initial manuscript; Dr Torchin conceptualized and designed the study, supervised the analyses and the study, and reviewed and revised the manuscript; Dr Ancel conceptualized and designed the study, supervised the analyses and the study, reviewed and revised the manuscript, and obtained funding; Drs Benhammou, Chevallier, Gire, Pierrat, Baud, Jarreau, and Marret, and Ms Marchand-Martin critically reviewed the manuscript for important intellectual content; and all authors approved the final version as submitted and agree to be accountable for all aspects of the work.
FUNDING: The EPIPAGE 2 project was funded with support from the French Institute of Public Health Research/Institute of Public Health and its partners: the French Health Ministry, the National Institutes of Health and Medical Research (INSERM), the National Institute of Cancer, and the National Solidarity Fund for Autonomy (CNSA); the National Research Agency through the French EQUIPEX program of investments in the future (reference ANR-11-EQPX-0038 and ANR-19-COHO-001); the PREMUP Foundation; Fondation de France (Reference 11779); Fondation pour la Recherche Médicale (SPF20160936356); Program Hospitalier de Recherche Clinique Epinutri (DGOS13-040); Ministère de l’Enseignement Supérieur, De La Recherche et de L’Innovation (G13129KK); and Apicil Foundation (R20065KK). Ludovic Tréluyer benefited during this work from an annual research grant from Astra-Zeneca awarded by a jury of the French Society of Neonatology. The other authors received no additional funding. The funding organizations 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 decision to submit the manuscript for publication.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest relevant to this article to disclose.
Comments
Differences in Low-Grade IVH and Neurodevelopmental Outcome in European Cohorts: What Is the Take Home Message?
In the current study, Tréluyer et al. [2] performed a secondary analysis of the French national population-based and prospective cohort EPIPAGE-2 (Etude éPIdémiologique sur les Petits Ages GEstationnels-2). They found that as compared with no IVH, low-grade IVH was not associated with measured neurodevelopmental disabilities at age 5.
What is the take-home message? In neonatal practice, regional and institutional data vary. When counseling the families about the neurodevelopmental outcome for infants with grade 1 and 2 IVH, the practitioner should consider the regional data.
References:
1. Périsset A, Natalucci G, Adams M, Karen T, Bassler D, Hagmann C. Impact of low-grade intraventricular hemorrhage on neurodevelopmental outcome in very preterm infants at two years of age. Early Hum Dev. 2023;177-178:105721. doi:10.1016/j.earlhumdev.2023.105721
2. Tréluyer L, Chevallier M, Jarreau PH, et al. Intraventricular Hemorrhage in Very Preterm Children: Mortality and Neurodevelopment at Age 5. Pediatrics. 2023;151(4):e2022059138. doi:10.1542/peds.2022-059138