INTRODUCTION
Intraventricular hemorrhage (IVH) is a major neurological complication in preterm and critically ill neonates. In infants with moderate-to-severe IVH, posthemorrhagic ventricular dilatation (PHVD) refers to progressive ventricular enlargement after hemorrhage, whereas posthemorrhagic hydrocephalus (PHH) generally indicates a more advanced or clinically symptomatic stage. PHVD is associated with secondary brain injury and adverse motor and cognitive neurodevelopmental outcomes even before progression to overt hydrocephalus. Therefore, the early recognition of progressive PHVD is important to minimize further neurological injury and optimize subsequent management [1-3].
Several international studies have described the epidemiology, intervention timing, and clinical outcomes of PHVD in preterm infants. Multicenter studies have shown that infants who develop PHVD after IVH have a substantially higher risk of death or neurodevelopmental impairment than those without PHVD. In addition, previous studies suggest that interventions initiated before progression to advanced ventricular dilatation or overt PHH may be associated with improved outcomes compared with a delayed approach. These findings underscore the importance of early identification of progressive PHVD, longitudinal monitoring of ventricular enlargement, and timely intervention in at-risk infants [1-3].
In Korea, national cohort studies have reported the incidence and risk factors of IVH and PHH in very-low-birthweight infants and, more recently, have described the current status of PHH in infants born at 22 to 28 weeks of gestation with severe IVH. However, data from single-center neonatal intensive care unit (NICU) cohorts remain limited, particularly regarding the progression from grade II–IV IVH to PHVD, the burden of ventricular dilatation, timing and type of intervention, and short- and long-term outcomes in Korean neonates. Therefore, we conducted a single-center cohort study of neonates with grade II–IV IVH to examine PHVD development, associated clinical characteristics, treatment patterns, and outcomes at discharge and at 18–24 months of corrected age [4,5].
MATERIALS AND METHODS
1. Study design and population
This retrospective single-center cohort study was conducted in the NICU of Kyungpook National University Chilgok Hospital, Korea. Neonates admitted between January 2014 and December 2025 who were diagnosed with grades II–IV IVH were eligible for inclusion. Infants with life-threatening congenital anomalies, including central nervous system (CNS) malformations and genetic disorders, were excluded. Infants who died within 7 days of IVH diagnosis were also excluded because progression to PHVD could not be reliably assessed in this group. The final cohort was divided into those who developed PHVD and those who did not (control group).
2. Institutional monitoring and treatment protocol
Cranial ultrasonography was routinely performed between postnatal days 3 and 7 in infants born at <32 weeks of gestation or weighing <1,500 g. In infants who did not meet these criteria, cranial ultrasonography was performed when neurological symptoms were present or intracranial hemorrhage or hypoxic-ischemic brain injury was suspected; otherwise, it was performed before discharge.
After IVH was identified, head circumference was measured daily, and follow-up cranial ultrasonography was performed at 1- to 2-week intervals, with additional scans when clinical signs of increased intracranial pressure were observed. Ventricular size was assessed using the ventricular index (VI), anterior horn width (AHW), and thalamo-occipital distance (TOD). Neurosurgical intervention was considered based on multidisciplinary clinical judgment in infants with severe PHVD, defined by marked ventricular enlargement (VI >97th percentile +4 mm) or clinical signs of increased intracranial pressure. Temporalized cerebrospinal fluid diversion was performed using external ventricular drainage (EVD), and permanent diversion was achieved through ventriculoperitoneal (VP) shunt placement when indicated. Routine serial cerebrospinal fluid (CSF) tapping is not part of the standard management protocol. Limited therapeutic CSF tapping was attempted in two infants while awaiting EVD placement, and both ultimately underwent EVD.
3. Definitions and data collection
IVH was graded using cranial ultrasonography according to the Papile classification at the time of the initial diagnosis [6]. Although this classification was originally developed mainly for preterm infants, it was used in this study as a standardized radiological classification of hemorrhage severity. Interpretation in late preterm and term infants should be made with caution. PHVD was defined as progressive ventricular enlargement on serial cranial ultrasonography after IVH, independent of the initial IVH grade. PHVD was diagnosed when the ventricular size exceeded the postmenstrual age-specific reference range, such as a VI above the 97th percentile with corresponding enlargement of the AHW and/or TOD, or when serial follow-up scans demonstrated clear progression of ventricular dilatation. Severe ventricular dilatation was defined as VI >97th percentile +4 mm, AHW >10 mm, and TOD >25 mm [1,3,7]. The brain injury variables included periventricular infarction, cystic encephalomalacia, and cerebellar hemorrhage.
Clinical and perinatal data collected from electronic medical records included gestational age, birth weight, sex, Apgar scores at 1 and 5 minutes, delivery mode, maternal factors, and neonatal morbidities. Small for gestational age (SGA) was defined as birth weight below the sex-specific 10th percentile for gestational age, based on the Korean reference for birth weight by gestational age and sex [8]. Maternal factors included pregnancy-induced hypertension (PIH), preterm premature rupture of membranes lasting >18 hours, pathologic chorioamnionitis, gestational or overt diabetes mellitus, and complete antenatal steroid exposure. Neonatal morbidities included respiratory distress syndrome, moderate-to-severe bronchopulmonary dysplasia, hemodynamically significant patent ductus arteriosus requiring treatment, necrotizing enterocolitis stage IIb or higher, culture-proven sepsis, and retinopathy of prematurity requiring treatment [9-11].
Meningitis documented during hospitalization was classified as a concomitant CNS infection and analyzed separately from EVD- or shunt-related ventriculitis, which is considered an intervention-related complication.
4. PHVD subgroup classification and intervention-related variables
Infants with PHVD were categorized into the intervention and conservative groups according to whether neurosurgical intervention was performed during hospitalization. The intervention group included infants who underwent EVD, VP shunt placement, or both, whereas the conservative group included infants who were managed without neurosurgical intervention.
Ventricular dilatation characteristics included postnatal day and corrected age at PHVD diagnosis, interval from IVH to PHVD diagnosis, maximum VI, maximum AHW, maximum TOD, postnatal day of maximum ventricular dilatation, and corrected age at maximum ventricular dilatation. Excess VI above the 97th percentile (ΔVI) was calculated as the difference between the measured maximum VI and the age-specific 97th percentile reference value. Intervention-related variables included intervention type, timing of the first intervention, and intervention-related complications.
5. Outcomes
The short-term outcomes included discharge status, corrected age at discharge, head circumference at discharge, head circumference below the 10th percentile, and seizure or anticonvulsant use.
Long-term outcomes were assessed during follow-up visits conducted at 18 to 24 months of corrected age. Neurodevelopmental follow-up was performed by neonatologists and/or pediatric neurologists and was based on neurological examinations in conjunction with developmental assessments using the Korean Developmental Screening Test for infants and children and/or the Bayley Scales of Infant Development, second edition, or the Bayley Scales of Infant and Toddler Development, third edition, as available [12-14]. The outcomes evaluated included head circumference, head circumference below the 10th percentile, seizure or anticonvulsant use, developmental delay in each domain, cerebral palsy, visual impairment, and hearing impairment. Developmental delay in each domain was defined as either a Bayley composite score <70 or a recommendation for further evaluation on the corresponding Korean Developmental Screening Test domain; ‘follow-up needed’ was not considered developmental delay. Because the Bayley Scales provide composite scores, whereas the Korean Developmental Screening Test for infants and children uses categorical recommendations rather than directly comparable scores, individual test results were not presented in a unified format. Instead, a harmonized binary definition was applied across the instruments. Any developmental delay was defined as a delay in at least one of the following domains: cognitive, motor, or language.
Cerebral palsy was diagnosed by pediatric rehabilitation specialists, and functional severity was classified using the Gross Motor Function Classification System [15]. Visual impairment was defined as clinically significant ophthalmological abnormalities identified during follow-up, including strabismus, optic nerve abnormalities, abducens nerve palsy, and other vision-related disorders documented in the medical records. Hearing impairment was defined as abnormal hearing evaluation or documented hearing loss requiring follow-up or intervention.
6. Statistical analysis
Continuous variables were presented as mean±standard deviation or median (interquartile range [IQR]), as appropriate, and categorical variables were presented as numbers (percentages). Comparisons between the PHVD and control groups and between the intervention and conservative groups were performed using the independent t-test or Mann-Whitney U-test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables, as appropriate. A prespecified sensitivity analysis was performed after excluding infants with antenatal IVH or ventriculomegaly. Supplementary analyses included a subgroup analysis restricted to infants born at <35 weeks of gestation and multivariable logistic regression analyses for selected neurodevelopmental outcomes at 18 to 24 months of corrected age adjusted for gestational age and birth weight. Neurodevelopmental outcomes at 18 to 24 months of corrected age were also descriptively reviewed by treatment subgroup within the PHVD cohort. Because the conservative group was small, no formal statistical comparisons were performed between treatment subgroups. Statistical significance was defined as a two-sided P-value of <0.05. All the statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.).
RESULTS
1. Study population
During the study period, 5,426 neonates were admitted to the NICU at Kyungpook National University Chilgok Hospital. Among them, 73 were diagnosed with grades II–IV IVH. One infant with a life-threatening congenital anomaly was excluded. An additional eight infants who died within 7 days after IVH diagnosis were excluded because progression to PHVD could not be reliably assessed. Consequently, 64 infants were included in the final study.
Among these, 22 developed PHVD, whereas 42 did not serve as controls. Of the 22 infants with PHVD, 16 underwent neurosurgical intervention and six received conservative management. Within the intervention group, five underwent EVD alone, six underwent EVD followed by VP shunt placement, and five underwent VP shunt placement without prior EVD (Figure 1).
2. Baseline characteristics and perinatal morbidities according to posthemorrhagic ventricular dilatation development
Table 1 presents the baseline characteristics and perinatal morbidities of the infants with PHVD and controls. Infants with PHVD had a higher gestational age at birth (34.6 weeks [IQR, 29.9 to 37.8] vs. 28.6 weeks [IQR, 27.3 to 32.7],P=0.006) and higher birth weight (2,270 g [IQR, 1,419 to 3,320] vs. 1,230 g [IQR, 939 to 1,805], P=0.006) than controls, and these differences remained significant in the sensitivity analysis. Among maternal factors, PIH was less frequent in the PHVD group (0.0% vs. 23.8%, P=0.012), and this difference remained significant in sensitivity analysis. Apgar scores, sex, cesarean delivery rate, SGA, and neonatal morbidities did not differ significantly between groups.
3. Intraventricular hemorrhage severity, associated brain injury, and concomitant central nervous system infection according to posthemorrhagic ventricular dilatation development
Table 2 presents IVH severity, associated brain injury, and concomitant CNS infection according to PHVD development. Infants with PHVD had a higher corrected age at IVH diagnosis than controls (36.1 weeks [IQR, 30.1 to 39.6] vs. 29.6 weeks [IQR, 27.8 to 35.2], P=0.011) and a more severe IVH grade distribution (P=0.001), which remained significant in sensitivity analysis. Among associated brain injuries, cystic encephalomalacia was more frequent in the PHVD group (59.1% vs. 23.8%, P=0.005) and remained significant in sensitivity analysis, whereas other associated brain injury variables, including periventricular infarction and cerebellar hemorrhage, did not differ significantly. Meningitis was rare in both groups.
4. Intraventricular hemorrhage and ventricular dilatation characteristics according to treatment group in infants with posthemorrhagic ventricular dilatation
Table 3 presents the severity of IVH, ventricular dilatation, and associated brain injury characteristics according to treatment group in infants with PHVD. Figure 2 depicts the distribution of ventricular measurements according to treatment group. The postnatal day of PHVD diagnosis tended to be later in the intervention group than in the conservative group (25.0 days [IQR, 9.0 to 31.0] vs. 8.0 days [IQR, 3.0 to 10.0], P=0.057), whereas corrected age at PHVD diagnosis did not differ significantly. The interval from IVH to PHVD diagnosis was also longer in the intervention group (10.0 days [IQR, 3.8 to 21.0] vs. 1.0 days [IQR, 0.0 to 6.0], P=0.073), although this difference was not statistically significant. IVH grade distribution, bilateral IVH, and antenatal IVH/ventriculomegaly did not differ significantly between the groups.
The intervention group showed greater ventricular enlargement than the conservative group, with higher maximum VI (19.9 mm [IQR, 16.1 to 25.0] vs. 14.7 mm [IQR, 14.1 to 16.0], P=0.036), greater ΔVI (6.9 mm [IQR, 3.2 to 12.5] vs. 2.3 mm [IQR, 1.8 to 2.4], P=0.010), and higher maximum AHW (16.4 mm [IQR, 12.2 to 20.3] vs. 9.8 mm [IQR, 9.6 to 10.7], P=0.005). Among indicators of severe ventricular dilatation, only AHW >10 mm was more frequent in the intervention group (93.8% vs. 33.3%, P=0.009). Cystic encephalomalacia was also markedly more frequent in the intervention group (81.2% vs. 0.0%, P=0.001), whereas other brain injury variables did not differ significantly.
5. Clinical course and timing of neurosurgical intervention in infants with posthemorrhagic ventricular dilatation
Figure 3 illustrates the clinical course and timing of the neurosurgical intervention in infants with PHVD. Supplementary Table 1 provides the individual clinical, imaging, treatment, and outcome characteristics of the infants who underwent the intervention. In the intervention group, IVH was diagnosed at a median postnatal age of 8.0 days, PHVD at 25.0 days, and the first neurosurgical intervention at 35.5 days. The median interval from IVH to PHVD diagnosis was 9.5 days, and the median interval from PHVD diagnosis to the first intervention was 6.0 days. EVD/shunt-related ventriculitis occurred in six of the 16 infants who underwent neurosurgical intervention, and VP shunt revision was required in two infants.
6. Discharge and neurodevelopmental outcomes according to posthemorrhagic ventricular dilatation development
Table 4 presents discharge and follow-up neurodevelopmental outcomes according to PHVD development. At discharge, infants with PHVD had a higher corrected age (46.9 weeks [IQR, 44.3 to 47.8] vs. 40.3 weeks [IQR, 38.5 to 42.4], P<0.001) and larger head circumference (37.0 cm [IQR, 35.0 to 38.0] vs. 33.5 cm [IQR, 32.1 to 35.6], P<0.001) than controls. Head circumference below the 10th percentile did not differ significantly between groups, whereas seizure or anticonvulsant use at discharge was more frequent in the PHVD group (19.0% vs. 2.4%, P=0.039).
At 18 to 24 months of corrected age, head circumference below the 10th percentile (73.3% vs. 34.8%, P=0.020), seizure or anticonvulsant use (31.2% vs. 4.3%, P=0.033), any developmental delay (75.0% vs. 25.0%, P=0.002), and cognitive (62.5% vs. 16.7%, P=0.003), motor (62.5% vs. 16.7%, P=0.003), and language delays (68.8% vs. 8.3%, P<0.001) were all more frequent in the PHVD group than in controls. Cerebral palsy, visual impairment, and hearing impairment were more frequent in the PHVD group; however, these differences were not statistically significant. The overall direction of the main findings was preserved in subgroup analysis restricted to infants born at <35 weeks of gestation and in supplementary analyses adjusted for gestational age and birth weight (Supplementary Tables 2, 3).
Among infants with PHVD, follow-up data were available for four conservatively managed and 12 surgically treated infants. As the conservative group was small, formal statistical comparisons were not performed. Developmental delay was observed in three of four conservatively managed infants (75.0%) and nine of 12 surgically treated infants (75.0%), whereas it was observed in six of 24 controls (25.0%). Cerebral palsy was observed only in the intervention group.
DISCUSSION
In this single-center cohort of neonates with grade II–IV IVH, PHVD developed in approximately one-third, which is within the range reported in previous studies [2,3,16-20]. Compared to controls, infants who developed PHVD had a more severe IVH grade distribution, a higher frequency of cystic encephalomalacia, and worse neurodevelopmental outcomes. In the PHVD cohort, the infants who underwent neurosurgical intervention showed greater ventricular dilatation and a markedly higher frequency of cystic encephalomalacia than those who were managed conservatively.
These findings align with previous studies showing that PHVD is an important sequela of IVH and is associated with a higher risk of adverse neurological outcomes. International studies have also suggested that greater ventricular enlargement and delayed intervention may be associated with less favorable neurodevelopmental outcomes [2,3,16-20]. In our cohort, these adverse effects were reflected in head growth and neurodevelopmental outcomes. Although infants with PHVD had a higher gestational age and birth weight than controls, their long-term neurodevelopmental outcomes were worse, suggesting that the observed differences are unlikely to be explained solely by baseline maturity. Nevertheless, these group differences should be interpreted with caution because the PHVD group included a greater proportion of relatively mature infants at birth, including late preterm and term infants. Infants with PHVD had a larger head circumference at discharge, likely reflecting older corrected age and hydrocephalus-associated skull expansion rather than preserved brain growth. However, at 18 to 24 months of corrected age, head circumference below the 10th percentile was more frequent in the PHVD group, despite similar absolute head circumferences between the groups. Previous studies have suggested that suboptimal head growth and relative microcephaly after neonatal brain injury are associated with adverse neurodevelopmental outcomes. Our findings support this interpretation [21]. Overall, the poorer neurodevelopmental outcomes observed in the PHVD group suggest a broad burden of impairment, rather than an isolated deficit in a single domain. This pattern supports the interpretation that PHVD reflects more severe brain injury and impaired brain growth potential than benign radiological findings.
Within the PHVD cohort, infants who underwent intervention had larger maximal ventricular measurements, including higher VI, ΔVI, and AHW values, than those managed conservatively, and a markedly higher frequency of cystic encephalomalacia. These findings suggest that the intervention group included infants with a more severe disease course rather than a distinct treatment category. Accordingly, the need for neurosurgical intervention may serve primarily as a marker of severe PHVD and associated brain injury. This should be considered when interpreting outcome differences between the intervention and conservative groups [16,18].
The timing data in our cohort provided clinically relevant insights into the natural course of PHVD after IVH. In infants who ultimately required neurosurgical intervention, PHVD was not identified immediately after the initial hemorrhagic event, but evolved over days to weeks during serial follow-up. Notably, the intervention group not only had a later postnatal day of PHVD diagnosis, but also a longer interval from IVH to PHVD diagnosis (10.0 days vs. 1.0 days). In contrast, ventricular dilatation in the conservative group may have been identified relatively early after IVH but remained limited without further progression. Although this interpretation should be made with caution because of the small sample size, these findings suggest that the progression pattern of ventricular enlargement after IVH, rather than timing of PHVD diagnosis alone, may be more clinically informative in identifying infants who later require neurosurgical intervention. These findings also support the importance of serial cranial ultrasonographic follow-up after IVH to allow timely detection of progressive ventricular dilatation.
In our cohort, neurosurgical intervention was generally performed when PHVD reached a severe stage. Compared to the thresholds used in previous studies, our local practice appears closer to a relatively high-threshold intervention strategy [1,3,17,22,23]. In this setting, infants who underwent intervention had a greater burden of ventricular dilatation, cystic encephalomalacia, and adverse neurodevelopmental outcomes. Recent randomized and observational studies have suggested that earlier intervention at lower thresholds of ventricular dilatation may be advantageous [17,19,20]. In particular, the ELVIS trial showed that although the primary composite outcome did not differ in the unadjusted analysis, the lower threshold group had less brain injury on magnetic resonance imaging and a lower risk of adverse neurodevelopmental outcomes in the adjusted analyses [17]. Adverse neurodevelopmental outcomes were not limited to infants who underwent neurosurgical intervention but were also observed in conservatively managed infants in our cohort. However, because the number of conservatively managed infants was small, a formal comparison was not possible. Taken together, these observations suggest that PHVD itself may carry a clinically important neurodevelopmental risk and that refinement of local protocols toward earlier intervention in selected high-risk infants may be worth considering. However, the available evidence remains insufficient to define an optimal intervention threshold, and further prospective studies are needed [17,19,20].
Our study also provides complementary single-center data in a Korean setting, particularly on serial ventricular measurements, timing of intervention, and neurodevelopmental follow-up in infants with PHVD. Unlike previous Korean studies that focused mainly on very preterm or very-low-birthweight infants, our cohort included a substantial proportion of infants born at ≥32 weeks of gestation, including term infants, and these infants were more common in the PHVD group than in the controls [4,5]. Accordingly, the higher gestational age at birth, higher birth weight, and lower prevalence of maternal PIH observed in the PHVD group likely reflect the greater representation of relatively mature infants at birth, including late preterm and term infants, rather than the intrinsic features of PHVD itself. This difference in cohort composition should be considered when making direct comparisons with previous Korean cohorts, as previous studies focused predominantly on very preterm infants [4,5].
This study had several limitations. First, it was a retrospective single-center study with a relatively small sample size, particularly for subgroup analyses within the PHVD cohort, which may limit generalizability; only parsimonious supplementary adjusted analyses were performed. Second, cranial ultrasonography screening was more intensive in infants born at <32 weeks of gestation or weighing <1,500 g than in more mature infants, raising the possibility that mild or asymptomatic IVH in late preterm and term infants may have been detected later or under-ascertained. Third, follow-up neurodevelopmental data at 18 to 24 months of corrected age were not available for all infants. Fourth, this cohort included a substantial proportion of late preterm and term infants, in whom IVH may arise from etiologies distinct from those in very preterm infants, which may limit outcome comparability and complicate direct comparisons with preterm-focused studies. Finally, infants who died within 7 days of IVH diagnosis were excluded, which may have introduced a survival selection bias, particularly among the most immature and critically ill infants with severe IVH.
Nonetheless, this study has important strengths, including a detailed review of serial cranial ultrasonography, integration of ventricular measurements and intervention timing, and availability of both discharge and follow-up neurodevelopmental outcomes. These findings provide clinically relevant insights into PHVD progression, treatment, and outcomes in a Korean NICU cohort, although further prospective multicenter studies are needed to define the optimal timing and threshold of intervention, and to identify infants most likely to benefit from earlier treatment.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
