Quality of Life at a 10-Year Follow-Up of Children Born Preterm with Post-Hemorrhagic Ventricular Dilatation: A Cohort Study

Preterm births represent a large global burden, with around 15 million babies being born preterm per year, and the numbers likely increasing [1]. The leading cause of childhood mortality is complications attributed to prematurity [2], and one major complication is intraventricular haemorrhage (IVH) [3] and subsequent post-haemorrhagic ventricular dilatation (PHVD) [4].

IVH after preterm birth is a significant worldwide problem, and both IVH and PHVD are associated with significant increased mortality [5], and strongly influence decisions around limitation or withdrawal of care [6]. Reducing IVH and other perinatal brain injury is a key part of the UK National Maternity Ambition [7] and while a small number of interventions may reduce the risk [8, 9], no effective treatment is currently available [10, 11].

However, despite this substantial health burden, there is uncertainty about the long-term outcomes as the children grow, and little is known about the later health-related quality of life (HRQoL). The World Health Organisation (WHO) defines quality of life as the “individuals” perception of their position in life, in the context of the culture and value systems in which they live, and in relation to their goals, expectations, standards, and concerns [12]. A number of studies have reported the health-related quality of life (HRQoL) of preterm infants, reporting impacts in emotional, social functioning, speech, and dexterity; although, overall the subjective experience of their quality of life may be similar to their peers [13‒15]. However, there is little literature on quality of life in preterm children after severe IVH and PHVD despite it being a clinically important event, and a major factor assessed in neonatal care when considering withdrawal of care in these sick infants [6].

The drainage, irrigation, and fibrinolytic therapy (DRIFT) trial was a randomised controlled trial of a novel neurosurgical and medical treatment, which followed up infants with IVH or PHVD to 10 years of age [10]. While the study showed benefits of DRIFT in cognitive outcomes, overall HRQoL measures were similar across the two treatment arms [10].

The aim was to assess the quality of life in the cohort of infants with severe neonatal IVH and resultant PHVD at 10 years of age; data were obtained from the DRIFT randomised controlled trial.

The trial recruited preterm neonates with severe (grade 3 or 4) IVH, and corresponding PHVD, in 4 clinical centres, in the UK and Poland, and randomised them to either receive DRIFT or standard non-surgical treatment. DRIFT is a therapy involving an intraventricular injection of a fibrinolytic, followed by continuous lavage for up to 72 h [16]. The standard care arm of the trial consisted of control of ventricular expansion by lumbar punctures, or by using a ventricular access device. In total, 77 babies were recruited between 2003 and 2006 [16], and the 10-year follow-up assessed a total of 52 of these children between February 2015 and April 2016 [10, 17]. Further details on the initial trial and patient flow are presented in Appendix A.

At the 10-year follow-up, parents were asked to complete questionnaires for two, validated, generic measures of their child’s HRQoL.

• The Health Utilities Index (HUI-3) [18] measures eight attributes of quality of life: vision, hearing, speech, emotion, cognition, ambulation, dexterity, and pain [19]. They are discriminated by a score of 1 (no restrictions to their HRQoL) to 5 or 6 depending on the level measured (representing poor HRQoL).

• The EQ-5D-5L [20] measures five domains: self-care, usual activities, mobility, anxiety/depression, and pain/discomfort. Each provides an ordinal scale ranging from one to five, with one being no problems experienced to five being extreme problems experienced.

Attribute/domain responses for these measures are mapped to a summary score based on general population “valuation sets” provided by the EQ-5D-5L [21] and HUI-3 developers [19]. For both the EQ-5D-5L and HUI-3, the summary HRQoL score is anchored at a maximum score of 1 (best health) and 0 (often considered equivalent to death) [22]. It should be noted that scores below 0 are possible with both measurements and represent an extremely poor HRQoL.

Demographic and neonatal data were obtained through medical records and specific research data collection sheets and recorded in a dedicated research database. Further data included the following:

• Demographic data: sex (male or female), gestational age at birth (categorised as less than 28 weeks, 28–32, or 33 and above), birthweight (categorised as <1,000 g, 1,000–1,499 g, 1,500–2,499 g, ≥2,500 g), maternal age (categorised as <30 or 30 years and above)

• IVH characteristics: IVH graded as 3 or 4 [23].

• Treatments provided: DRIFT treatment arm (DRIFT/standard) and subsequent ventriculo-peritoneal (VP) shunt (yes/no) before discharge.

Initially, the median response scores for each of the HRQoL questions were derived, and the final summary score calculated using published scoring algorithms [19, 21], alongside an estimate of how correlated the summary scores were. Next, we derived and compared the summary scores for the HRQoL measures, split by the demographics and clinical characteristics of the population (as categorised above). Finally, we derived a linear regression model to assess the multivariable relationships between the covariates and the overall HRQoL (derived separately for each measure). For this model, gestational age at birth, birthweight, and maternal age were included as continuous terms.

Categorical data are presented with the number percentage (n [%]), and continuous data are presented as median (inter-quartile range [IQR]) or mean (standard deviation or 95% CI). Univariable comparisons between groups were made using the Mann-Whitney test (or an extension of it to 3 or more groups [24]). Analysis was performed on STATA version 17.

From the 77 children recruited to the DRIFT trial, 52 survivors were enrolled in the 10-year follow-up. One participant had incomplete responses on both the HUI-3 and the EQ-5D-5L, leaving 51 participants for the analysis.

The majority of infants were male (n = 36 [70.6%]) (Table 1). Half were extremely preterm (<28 weeks at birth) (n = 25 [49.0%]) and had a birthweight between 1,000 and 1,499 g (n = 22 [42.3%]). Half of the children had grade 3 IVH (26 [50.9%]) and half had grade 4 IVH (25 [49.0%]).

The median HRQoL score at the 10-year follow-up was 0.66 (IQR 0.35–0.87) for the ED-5D and 0.52 (0.22–0.87) for the HUI-3. The distribution of summary scores and individual domains are shown in Figures 1 and 2, and while many children had scores above 0.8, some had scores below 0. The two summary scores were highly correlated (p < 0.001), with an r² of 0.70 (Table 2).

There was little evidence that either HRQoL score was different by sex, maternal age, birthweight, or gestational age categories at birth (Table 3). There was some weak evidence that increasing gestational age was associated with increasing EQ-5D-5L (p_trend = 0.059) but not HUI-3 scores (p_trend = 0.111). There was stronger evidence that increasing birthweight was associated with higher EQ-5D-5L (p_trend = 0.013) and HUI-3 summary scores (p_trend = 0.022). Children with grade 4 IVH had lower EQ-5D-5L (0.45 [IQR 0.11–0.71] versus 0.82 [IQR 0.62–0.92], p = 0.0016 and HUI-3 scores [0.34 vs. 0.80], p = 0.0004) than those with grade 3. While there was only weak evidence that infants who received a VP shunt had a lower EQ-5D-5L scores (0.60 [0.13–0.72] versus 0.76 [0.45–0.90], p = 0.0864), there was stronger evidence for a reduction in their HUI-3 scores (0.34 [−0.04–0.61] versus 0.80 [0.43–0.93], p = 0.0004). Overall, children who received DRIFT treatment had similar HRQoL scores (EQ-5D, p = 0.7919; HUI-3, p = 0.8696).

In the multivariable model (Table 4), IVH grade 4 was associated with substantially lower EQ5D (−0.31 [−0.53 to −0.08], p = 0.008) and HUI-3 (−0.31 [−0.53 to −0.08], p = 0.009) scores at 10 years. Once adjusted for IVH grade, there was little evidence for any association with either HRQoL scores by sex, gestational age, birthweight, need for a VP shunt, or DRIFT treatment arm (all p values >0.1).

The aim of this work was to evaluate, and describe, the quality of life in children who experienced IVH and PHVD. At 10 years of age, following severe IVH in the neonatal period, HRQoL measures are estimated at around 0.5–0.6, depending on the score and summary statistic used; substantially lower than the likely population mean of 0.93 (SD 0.13) [25]. Overall children scored a median of one or two in most of the domains, indicating slight or no problems, but some had very low scores, with around 10% scoring less than 0. Individual domains are difficult to assess in isolation, but a high level of disability was reported in cognition on the HUI-3 tool. Overall, children with lower birthweights or gestational ages reported lower HRQoL measures, but this association may be predominantly due to a higher risk of a more severe IVH.

The relatively small number of children providing HRQoL scores at 10 year follow-up limits the statistical power to identify small but clinically important associations between patient characteristics and subsequent quality of life. However, this cohort represents the largest to date with systematic and long-term HRQoL follow-up, which allowed us to signpost the most influential associations for future work. While the relatively small sample size may limit the generalisability of this study, the cohort was recruited from neonates gathered across the country and may represent the wider UK population with severe IVH [16]. We used parent proxy responses, and parental perceptions of HRQoL may be influenced by their own expectations, and the burden they have faced when caregiving [15, 26, 27]. Indeed, while parents may place less importance regarding psychosocial aspects such as self-esteem, cognition, and emotion than children [13, 14, 26, 28], proxy reporting was needed to ensure consistent results across the cohort [27]. Finally, to adjust for confounders, linear regression models were used assuming normally distributed residuals.

One previous study assessing HRQoL in children with PHVD at 5 years suggested a similar reduced overall HRQoL, but with impact in motor, emotional, social, and school domains [29], while other studies have noted that lower HRQoL outcomes may depend on physical function [15, 28]. However, many of the parents in our study reported their children experienced no problems with pain or anxiety (EQ-5D-5L), or hearing, speech, ambulation dexterity, emotion or pain (HUI-3). Furthermore, the age of assessment may be important, with previous work suggesting improvements in HRQoL after preterm birth as the children grow [15]. In this work, mobility and corresponding activities appear to be the most commonly impacted. A higher grade of IVH was strongly associated with worse HRQoL at 10 years of age, and while DRIFT has been shown to improve cognitive outcomes in this group of infants, it is unlikely to improve motor outcomes, and so the results of this analysis are consistent with our published analysis of trial outcomes [10].

While no further studies on the efficacy of DRIFT have been published since this initial trial, the proof of principle that washing blood out of the brain reduces cognitive disability has led to alternative techniques (e.g., neuroendoscopic lavage) being more widely adopted and investigated. Of note, in our cohort, the presence of a VP shunt was not independently associated with worse HRQoL later in life despite some work suggesting to have a VP shunt is a poor prognostic outcome [30].

There remains a lot of uncertainty regarding the neonatal care of children with severe IVH, and withdrawal of care in neonatal units is commonly performed due to concerns over future HRQoL [6]. In this work, the summary scores appear substantially lower than the estimated normative means (0.93 [SD 0.13]) [25], as well as likely lower than an English population cohort of all children born extremely preterm [14]. In that work, the HUI-3 scores of 11-year-old children, born before 26 weeks of gestation in 2006, had a median score of 0.76 (IQR 0.38–0.97) compared to 0.52 (0.22–0.87) in this work.

However, interpreting what the scores may mean to the child needs to be performed in context of the clinical case and the views of the family. Comparison with other, perhaps more common disease processes suggest that the absolute metrics in this cohort (e.g., a HUI-3 median score of 0.52) appear higher to those reported for children with autism spectrum disorders (0.41) [25] or trisomy 21 (0.34) [25]. These referenced values are from sampled children using a similar methodology, and from a cohort of children with a similar demographic and age range to this work. However, data were limited to families of children with disability or severe illness who had applied for financial support, and therefore may represent a more disabled cohort than other children with their conditions. Interestingly, the results are also similar than those seen in survivors of adult stroke (measured 6 months after an ischaemic stroke; mean HUI-3 0.44 (SD 0.37) [31], another condition where novel active post-event treatment has reduced mortality [32] and increased intact survival [33].

However, it should be noted that while the median (and mean) scores reported for this cohort appears to suggest a similar HRQoL to many other childhood conditions, the range seen was often broader (e.g., the IQR for the HUI-3 was 0.22–0.87), with a substantial proportion with scores below 0, suggesting that outcome is more varied and difficult to predict in this group. Scores below zero indicate a HRQoL where physical, mental, or social functioning are likely severely compromised, alongside limitations and reduced ability to engage in daily activities.

Overall, around one-fifth of all children who survive PHVD have a very low HRQoL, while a quarter have scores above 0.87 (similar to population norms). However, impact was not uniform across domains. A higher (grade 4) IVH grade was the only clear factor tested that was predictive of poorer HRQoL. These parent-reported data on their children’s HRQoL may be useful for professionals and parents when discussing ongoing care plans, and future work could usefully explore more predictors of better HRQoL, in survivors of neonatal PHVD.

We thank the families and the children for their support and commitment to the DRIFT study and for giving their time so generously to advance the knowledge and management of PHVD. This study would not have been possible without their significant contribution.

The members of the DRIFT10 study group also include Charlotte Lea, Grace J Young, Helen Miller, Grazyna Kmita, Cathy Williams, Aída Moure Fernández, Michelle Morgan, Adam Smith-Collins, N Jade Thai, Steven Walker-Cox, and Kristian Aquilina. We are extremely grateful to the whole DRIFT study team.

Ethical approval was granted by the NHS Health Research Authority, NRES Committee South West - Central Bristol (14/SW/1078). Parents gave written informed consent. The children gave their assent for follow-up.

A.R., C.W., S.J., P.B., I.P., A.W., K.L., and D.E.O. reported no conflicts of interest.

The original DRIFT study was funded by the James and Grace Anderson Trust and by Cerebra. The school-age follow-up study was funded by the National Institute of Health Research, Health Technology Assessment Programme (HTA 12/35/61). This manuscript presents independent research funded by the National Institute for Health Research (NIHR). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health. The funding sources 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.

A.R. participated in the study concept and design, analysis and interpretation of data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content. C.W., S.J., and I.P. participated in the study concept and design, analysis and interpretation of data, critical revision of the manuscript for important intellectual content, and obtaining funding. P.B. participated in interpretation of data and critical revision of the manuscript for important intellectual content, A.W. designed, obtained funding, led the original DRIFT trial, and participated in this study concept and design analysis and interpretation of data, and critical revision of the manuscript for important intellectual content. K.L. and D.E.O. participated in the study concept and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, and obtaining funding.

Karen Luyt and David Edward Odd contributed equally to this work.Trial registration number: ISRCTN80286058.