OBJECTIVE: To assess the early adult health, education, and social functioning of a national cohort of very low birth weight (VLBW) infants born in 1986 in New Zealand compared with a control group of young adults born the same year. The impact of neurodevelopmental disability at age 7 to 8 years on young adult outcomes was examined.
METHODS: Participants underwent a comprehensive face-to-face interview that included standardized assessment tools and previously used custom written survey items.
RESULTS: We interviewed 230 VLBW young adults (71% of survivors) and 69 controls at age 22 to 23 years. VLBW young adults were 5.6 kg lighter and 4.2 cm shorter than controls, had lower rates of tertiary education/training (percentage difference [95% confidence interval]: −13.1 [−21.6 to −1.8]) and university degree completion (−15.6 [−28.0 to −4.8]), had more often been welfare dependent (23.5 [10.2 to 35.0]), had few or no friends (20.5 [7.2 to 32.2]), and more often had wheeze in the past year (20.1 [9.0 to 28.6]). However, in many areas there were no differences between the VLBW cohort and controls, and VLBW adults rated their overall quality of life and behavioral functioning similarly to their peers. Those with previous disability had poorer social, occupational, and physical functioning than other VLBW young adults.
CONCLUSIONS: Despite some evidence of health, educational, and social difficulties, former VLBW young adults obtained similar scores across many aspects of health and social functioning as their same age peers, with some differences largely confined to those with previous disability.
- very low birth weight
- young adults
- long-term outcome
- New Zealand/epidemiology
- health status
- educational status
- quality of life
- SGA —
- small for gestational age
- VLBW —
- very low birth weight
- VP —
- very preterm
What’s Known on This Subject:
Several studies have suggested that very low birth weight young adults have increased risks of physical and health problems, educational underachievement, and poorer social functioning than their peers, but there are limited population-based and longitudinal data.
What This Study Adds:
Former VLBW young adults in this national cohort scored as well as term controls on many measures of health and social functioning, including quality-of-life scores, with some differences largely confined to those with disability at age 7 to 8 years.
Infants born very low birth weight (VLBW; <1500 g) or very preterm (VP; <32 wk gestation) account for 1% to 2% of live births but between 50% and 75% of the workload of NICUs.1 Although survival rates for these infants now exceed 90%,2 what is most important for the individual, their families, and society is the quality of that survival in the longer term. Knowledge of longer term outcomes is also crucial to inform current neonatal care.
There is now clear evidence demonstrating increased rates of health problems among surviving VLBW infants compared with those born at term in their early years.3 During childhood, VLBW infants are more likely to experience poor growth, whereas neurosensory impairments, including cerebral palsy, are seen in up to 10%. Even more common are neurobehavioral impairments, with up to 40% of school-aged children born VP experiencing cognitive delay, learning problems, executive deficits, educational underachievement, and clinically significant emotional and behavioral adjustment problems, especially attention-deficit/hyperactivity disorder.4–10 Data are now emerging from a growing body of cohort and cross-sectional studies, particularly during adolescence but also early adulthood, showing that these impairments may persist and in turn impact the preterm survivor’s longer term functioning and life-course opportunities.11–17
In 1986, we enrolled all New Zealand VLBW infants admitted for neonatal intensive care in a prospective study of retinopathy of prematurity,18 and surviving children were followed up at ages 7 to 8 years.8,9 This cohort is unique in being population-based, characterized by high rates of antenatal steroid exposure, and with comprehensive information on child health and development during middle childhood. The aim of the current study was to retrace the cohort at 23 years and to assess their health, educational and occupational achievement, and social functioning relative to a comparison group of young adults born at term in the same year.
The 1986 cohort comprised all 413 VLBW infants who were live born and admitted for NICU care in New Zealand, 338 (82%) of whom survived to discharge.18 At birth, 58% had received antenatal steroids, 132 (32%) weighed <1000 g, 126 (31%) were <28 wk gestation, 103 (25%) were small for gestational age (SGA), and 95 (23%) were born to mothers who identified as Maori. Survival for those born <1000 g and at <28 wk was 64%. At follow-up at age 7 to 8 years, an additional 12 children had died, and we assessed 298 children (96% of survivors who resided in New Zealand; 91% of all survivors). Of these, 5% had severe disability, 5% had moderate disability, and 15% had mild disability, which was principally an IQ between 1 and 2 SDs below the mean.8 Compared with the Christchurch Health and Development Study cohort studied at the same age, the VLBW cohort had higher rates of behavioral problems (conduct disorders, attention problems, anxiety/withdrawal) and poorer school achievement.9,19
In the current study, the cohort was traced through their addresses at age 7 to 8 years, grandparents' addresses, general practitioners, National Health Index codes (a unique person identifier used within the New Zealand health system since 1993) via the electoral rolls and, in a limited number of cases, by local advertisements. Checks were made with Statistics New Zealand to see if any individuals had died since the last follow-up. A comparison group of individuals who were born at term in New Zealand and who were not admitted for NICU care was recruited, initially through peer nomination by cohort members (total of 40) and, because of time constraints, subsequently also via random sampling from electoral rolls (total of 29), ensuring balance with respect to the gender, ethnicity, and regional distribution of the sample. There were no substantive differences between those recruited by either method in terms of characteristics or outcome.
After written informed consent, all participants underwent a face-to-face interview conducted by 1 researcher, which lasted ∼2 hours. This interview examined individual functioning across a range of domains including the following: current living arrangements; education, income, and employment; physical and mental health; family, peer, and partner relationships; individual behavior and everyday executive difficulties; and quality of life, self-esteem, and overall life satisfaction. The interview incorporated a series of standardized assessment tools including the following: the Short Form 36 Health Survey, version 2,20 which allows quality-of-life scores to be derived; the Behavior Rating Inventory of Executive Function–Adult Version21; the Barkley adult self-report scales of attention-deficit/hyperactivity disorder and oppositional defiant disorder symptoms22; components of the Composite International Diagnostic Interview23 for assessment of mental disorders; the Inventory of Parent and Peer Attachments24; and the Rosenberg self-esteem scale.25 This information was supplemented by custom written survey items used previously in the Christchurch Health and Development Study26 to assess education, employment, income, health history, social and partner relationships, and life satisfaction. In addition, participants were assessed for height, weight, waist circumference, and body mass index (BMI). A more detailed description of the measures used in the present analysis is provided in Supplemental Table 7.
Between-group comparisons were tested for statistical significance by using the χ2 test of independence for comparison of percentages, Poisson regression for analysis of count outcomes, and the t test for independent samples or analysis of variance for differences in means of other continuous outcomes. Effect size estimates were summarized by the mean or percentage difference in outcome between groups and the associated 95% confidence interval. In supplementary analyses, the VLBW group was further stratified to test for variation in outcome by previous neurosensory disability at age 7 to 8 (none, mild, moderate/severe). Analyses were also conducted to compare outcomes for those in the VLBW cohort who were SGA at birth (birth weight <10th percentile) relative to those who were normally grown. Finally, to examine whether effect size differences were influenced by between-group differences in participant characteristics (age at assessment, gender, ethnicity, parental education, socioeconomic status), the observed outcome comparisons were adjusted by using multiple regression methods (logistic regression for dichotomous outcomes, Poisson regression for count outcomes, and multiple linear regression for other continuous outcomes).
The study had 80% power at α = 0.05 to detect a mean or percentage difference in outcome between groups with effect sizes (Cohen’s d) in the range of d = 0.3 to 0.5, depending on the nature of the outcome being assessed. These calculations suggest the study had adequate power to detect small to moderate effect sizes. Where the case group is further stratified (eg, by previous disability), power is reduced for subgroup analyses. In these cases, the design has adequate power to detect moderate to large effect sizes (d = 0.5–0.8).
The study was approved by the Multi-region Ethics Committee of the Ministry of Health (New Zealand).
Of the national VLBW cohort born in 1986, 324 were believed to be alive in 2008, with 2 known deaths since 7 to 8 years of age (of unknown causes), and 251 were traced and approached to take part in this study. This report concerns the 230 who consented to participate (92% of those approached; 71% of the surviving cohort), together with 69 controls.
Table 1 describes the characteristics of the sample. VLBW participants were interviewed at a mean age that was 3 months younger than the controls (P < .001). There were no significant differences between the proportion of the 230 survivors assessed and the 94 survivors not assessed who weighed <1000 g at birth; were born at <28 wk gestation; were born SGA; were male, Maori, or Pacific Islander; or who had moderate/severe disability at 7 to 8 years of age (Supplemental Table 8). There was no effect of receipt of antenatal steroids related to any of the outcomes reported below.
Education, employment and income, or welfare dependence are shown in Table 2. The VLBW cohort had lower overall educational attainment, with half as many gaining a university degree. The trend was most marked for those with previous disability at age 7 to 8 years (34.4%, no qualifications; 19%, high school completion; 56.9%, tertiary enrollment; 3.4%, university degree). VLBW graduates were more likely to have ever been welfare dependent. There was a clear trend (P < .01) on all measures toward lower income, hours worked, and greater welfare dependence with increasing previous disability at age 7 to 8 years. (Further information on outcomes for the VLBW cohort stratified by variation in previous neurosensory disability at age 7–8 years is provided in the Supplemental Information and Supplemental Tables 9–13.)
Family, peer, and sexual/partner relationships are shown in Table 3. The VLBW cohort reported being closer to their parents/grandparents but had fewer friends. There were significant trends for fewer friends and poorer quality peer relationships as assessed by the Inventory of Parent and Peer Attachments scale with increasing disability at age 7 to 8 years (P < .05). In terms of current living arrangements, the VLBW cohort was more likely to be still living with their parents and less likely to be rooming with their peers (P < .05; data not shown). In addition, those with moderate/severe disability were more likely to be living with parents (56% vs 36% of those with no/mild disability vs 25% of controls) or in residential care (22% [5 individuals] vs none in the other groups) (P < .001). There was an overall trend toward lesser engagement by the VLBW cohort in sexual/partner relationships, and rates of engagement were all substantially lower for those with moderate/severe disability at age 7 to 8 years (P < .05).
Substance abuse, criminality, and mental health are shown in Table 4. The VLBW cohort had similar rates of tobacco, alcohol, and cannabis use compared with controls but lesser involvement with other illicit drugs. There was a consistent trend for those with moderate/severe disability at age 7 to 8 years to be much less likely to have any involvement with drugs, both legal and illicit (P < .05). There were no differences in the rates of Diagnostic and Statistical Manual of Mental Disorders, 4th edition, disorders or suicidal behaviors between the cohort and controls, and there was no apparent variability in previous disability.
Physical health and growth variables are shown in Table 5. As well as an overall higher incidence of hospital attendance in the past year and lifetime chronic illness and wheeze among the VLBW cohort, there were also higher lifetime incidences of cardiac (P = .02), vision (P = .04), and hearing (P = .03) problems (data not shown). Lifetime hospital admissions and severe or chronic illness was greatest for those with moderate/severe disability at age 7 to 8 years. The cohort was, on average, 5.6 kg lighter and 4.2 cm shorter than controls. The average weight/height differentials between the cohort and controls did not vary with gender. There was also no significant between-group variability in mean BMI or waist circumference by gender.
Executive functioning, behavior, quality of life, and overall functioning are shown in Table 6. The 2 groups had similar scores on all measures. There was a trend for poorer behavioral regulation composite scores (P = .03) on the Behavior Rating Inventory of Executive Function–Adult Version with increasing disability at age 7 to 8 years.
Compared with others in the VLBW cohort, those born SGA were, on average, 8.5 (95% CI: 3.2–12.8) kg lighter and 6.3 (95% CI: 3.6–9.0) cm shorter (P < .001). They were also more likely to have a lifetime diagnosis of asthma (46.2% vs 30.3%; P = .02) and wheeze in the past 12 months (36.9% vs 29.7%), but this difference was not significant (P = .29).
Finally, analyses were extended by using regression methods to include control for possible between-group differences in personal and family background characteristics (age at assessment, gender, ethnicity, parental education, family socioeconomic status). Control for these factors did not materially affect any of the observed differences in outcome between VLBW and controls.
New Zealand VLBW young adults show clear differences in health, educational achievement, welfare dependency, and social engagement compared with controls. VLBW young adults were, on average, 5.6 kg lighter and 4.2 cm shorter than controls. The VLBW young adults less often had a university degree and were more likely to have either no qualifications or high school qualifications only. They were less likely to be living independently. Approximately one-quarter of the VLBW cohort and controls were daily smokers, and one-third of both groups used alcohol weekly and had used cannabis in the past year, but fewer VLBW graduates had ever used other illicit drugs or been arrested. In addition, they had a higher lifetime number of severe or chronic illnesses and were almost 3 times more likely to have been admitted to a hospital in the past year. However, in many areas, there were no differences between the VLBW young adults and controls, and VLBW graduates rated their overall quality of life and behavioral functioning in a similar way to their term peers.
Several reports have shown that former VLBW/VP young adults do have lower median IQ scores than their peers, even after excluding those with neurosensory impairments and controlling for confounding factors; and fewer graduate from high school or enter tertiary education.11,12,27 We did not formally measure IQ in this study, but differences in educational outcomes between the VLBW young adults and controls were more marked in those with previous disability at 7 to 8 years. Similarly, there were clear trends for those with previous disability, and particularly those with moderate or severe disability at age 7 to 8, to have poorer social, occupational, and physical functioning, and to exhibit poorer behavioral regulation than other VLBW young adults.
There are a number of studies that have reported respiratory outcomes in former VLBW/VP late-adolescents,12,28,29 although few in young adults.30–33 In general, the evidence suggests that VLBW/VP survivors, particularly if they had chronic neonatal lung disease, have reduced forced expiratory volume in 1 second, lower forced vital capacity, and decreased diffusing capacity. In our cohort at 7 to 8 years of age, we found that 32% had a diagnosis of asthma at the time of follow-up, with a significant trend (P < .05) of greater risk of asthma with decreasing gestation.34 At 23 years the VLBW young adults were almost 3 times more likely than the controls to have experienced wheeze in the past 12 months (32% vs 12%; P = .02), with a greater proportion having a lifetime diagnosis of asthma, although this difference was now not significant.
An important finding for the VLBW young adults themselves was that relative to their term peers, there were no significant differences on the Short Form 36 physical or mental component scores or on measures of perceived overall functioning, self-esteem, or life satisfaction. Similar findings have been reported previously for self-reported quality-of-life scores, even when scores obtained from parents or other observers have been lower than those for controls.16,35–39
Our report is the first, to our knowledge, from New Zealand to document outcomes for former VLBW infants in young adulthood. Those of Maori descent, the indigenous population in New Zealand, were overrepresented in the cohort with 23% of mothers recording Maori ethnicity compared with 13% of all births in 1986. At follow-up, 26.5% of the cohort identified themselves as Maori and 5.7% as of Pacific Island descent. Overall, consistent with national data, Maori participants had poorer educational qualifications, higher welfare dependence, higher suicidal ideation/attempt, arrest/conviction, cannabis use, daily smoking, higher rates of pregnancy/parenthood, and higher BMI scores. However the number of Maori participants in the control group was too small to draw useful conclusions regarding the size of VLBW/control differences for Maori and non-Maori participants.
Strengths of our study include that it is a national cohort, more than half of whom received antenatal steroids, and that an assessment had been made in middle childhood. Data from other countries are now revealing a range of functional and health-related challenges faced by former VLBW infants in late adolescence and early adulthood.11,12 However, most cohorts were born in the 1970s or early 1980s when few infants received antenatal steroids and so may be less representative of modern neonatal intensive care than our study. Whereas antenatal steroids reduce mortality by ∼40%,40 it is interesting that in this study receipt of antenatal steroids had no effect on any long-term outcomes. In addition, in New Zealand few infants of ≤23 wk gestation were offered NICU care in 1986, which largely remains the case with contemporary care. Many studies in other countries have been hospital based and so may be subject to a degree of bias. Of the few population-based studies, most are regional or rely on constrained national registry data.41–44 The study that is most similar to ours is the Netherlands Project on Preterm and Small for Gestational Age Infants study in VLBW/VP infants born in 1983.45,46 As in our study, Hille et al,46 reporting from the POPS study cohort at age 19 years, found that VLBW survivors had greater difficulty establishing social contacts but did not have elevated risks of psychopathology compared with their term-born peers.
A limitation of our study is that we assessed both the VLBW cohort and controls by face-to-face interview only. Our follow-up included 71% of those known to be alive at 23 years. Young adulthood is a time of high mobility, and many members of our cohort are known to be overseas, such trips still being a “right of passage” for many young New Zealanders. There were, however, no differences in the basic demographic characteristics of those who did and did not participate in our study. Our follow-up rate is comparable to that reported in most international studies. For example, in the highly informative Helsinki Study32,47–49 there were 474 consecutive VLBW admissions to the regional NICU from January 1978 to December 1985; 335 were discharged alive, 95% survivors were traced, 255 residing in the Helsinki area were invited to participate, and 166 did so (50% overall). The Netherlands POPS study at 19 years achieved 62% follow-up,45 and studies of pulmonary function in late adolescence in VP infants from Melbourne, Australia, included 74% of survivors.29
Despite the relatively positive findings, it will be important to objectively assess health and neuropsychological functioning in the cohort, and we are currently undertaking a battery of medical and neuropsychological investigations at age 26 to 27 years. Knowledge of the range of health issues faced by VLBW young adults will inform future neonatal care in general and in New Zealand and will guide multidisciplinary intervention and provision of medical care and social support in childhood and adolescence.
We have shown that, whereas New Zealand VLBW young adults do experience lower levels of educational achievement and greater social isolation, there are no differences on many measures of health and social functioning, and the young people themselves report similar overall life functioning and quality of life to their same age peers.
We thank Mrs Nina Mogridge and Mrs Nicki McNeill for assistance in tracing the cohort. We also thank the young adults who participated in the study.
- Accepted September 19, 2013.
- Address correspondence to Brian A. Darlow, MD, FRACP, Department of Pediatrics, University of Otago Christchurch, PO Box 4345, Christchurch 8140, New Zealand. E-mail:
Dr Darlow conceptualized and designed the study and drafted the initial manuscript; Mr Horwood contributed to the study concept and design, carried out the initial analyses, and reviewed and revised the manuscript; Ms Pere-Bracken contributed to the study design, was responsible for all the data collection, and critically reviewed the manuscript; Dr Woodward contributed to the study design and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: All phases of this study were supported by a project grant from the New Zealand Child Health Research Foundation (Cure Kids).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Darlow BA,
- Cust AE,
- Donoghue DA
- Marlow N
- Aarnoudse-Moens CS,
- Weisglas-Kuperus N,
- van Goudoever JB,
- Oosterlaan J
- Doyle LW,
- Victorian Infant Collaborative Study Group
- Horwood LJ,
- Mogridge N,
- Darlow BA
- Doyle LW,
- Anderson PJ
- Darlow BA
- Horwood LJ,
- Darlow BA,
- Mogridge N
- ↵Ware J Jr, Kosinski M, Gandek B. SF-36 Health Survey: Manual & Interpretation Guide. Lincoln, RI: QualityMetric, Inc; 2005
- Roth RM,
- Isquith PK,
- Gioia GA
- Barkley RA,
- Murphy KR
- World Health Organization
- Rosenberg M
- Fergusson DM,
- Horwood LJ
- Doyle LW,
- Faber B,
- Callanan C,
- Freezer N,
- Ford GW,
- Davis NM
- Lindström K,
- Winbladh B,
- Haglund B,
- Hjern A
- Hille ETM,
- Weisglas-Kuperus N,
- van Goudoever JB,
- et al.,
- Dutch Collaborative POPS 19 Study Group
- Strang-Karlsson S,
- Räikkönen K,
- Pesonen AK,
- et al
- Copyright © 2013 by the American Academy of Pediatrics