New Zealand 1986 Very Low Birthweight Follow-up Study: the third decade

Following advances in neonatal intensive care during the 1980s, there have been major improvements in the survival of infants born at very low birthweight (VLBW; <1,500g) and/or very preterm (VP; <32 weeks), such that currently over 90% of New Zealand infants go home following birth. VLBW/VP birth has the potential to impact development and later function of major organs as the third trimester of pregnancy is a time of maximum foetal growth and development; yet these infants spend much of this period ex utero and experience a wide range of health issues.^1,2 Although most research has focussed on childhood and early adolescent outcomes, there are a number of older^3,4 and more recent^5–9 overviews of the outcome of VLBW/VP infants as young adults. These reports suggest that while the majority of VLBW/VP graduates are doing well by their third decade, they do face several important challenges that adults born preterm and the health professionals caring for them should be aware of. Population-based New Zealand data, as provided by the New Zealand Very Low Birthweight longitudinal study (NZVLBW), are also vital to inform national health priorities and to identify and address inequities. There have been some 30 publications from the NZVLBW study on the adult outcomes at 28 years, but almost all have been in the international literature. The aim of this report, therefore, is to summarise the main published findings in one place and discuss the implications for these vulnerable individuals and for health service provision.

The NZVLBW study

The NZVLBW cohort had its genesis in 1986 when all 413 VLBW infants who were admitted to a newborn unit that year were enrolled in a prospective audit of retinopathy of prematurity (ROP), with 338 infants (82%) surviving to discharge home.^10,11 As shown in Figure 1, this cohort was subsequently retraced and followed up at ages 7–8 years,^12–14 22–23 years¹⁵ and at a mean (standard deviation [SD]) 28.5 (1.1) years. At the 28-year follow-up, 250 VLBW young adults (77% of survivors) participated; 229 VLBW adults and all 100 controls (see below) came to Christchurch for 2 days of medical and neurological assessments between February 2013 and November 2016, with the remaining 21 VLBW participants answering a questionnaire only. A comparison group of same-age controls, born healthy at term, were recruited at the 22-year follow-up (n=69), either through peer nomination by a cohort member or random sampling from electoral rolls, aiming to ensure balance with respect to infant sex, ethnicity and region of birth. At 28 years, 39 of these controls were seen again together with 61 further recruits using the same methodology as before (n=100). With these numbers of participants, the study has 80% power at alpha=.05 to detect mean between-groups differences of .30 SD or greater on continuous outcomes, and odds ratio (OR) in the region of 2.0–3.5 for dichotomous outcomes (depending on the base rate). This suggests the study has adequate power to detect effect sizes in the small-to-moderate range. The study was approved by the Upper South B Regional Ethics Committee, superseded by the New Zealand Southern Health and Disability Ethics Committee (URB/12/05/015). All participants gave written informed consent.

Table 1 describes the socio-demographics and perinatal characteristics of the VLBW cohort participants and of the controls. There were no substantial differences between VLBW participants and non-participants, and between participants clinically assessed (n=229) and not assessed (n=21), except that the assessed group included fewer children with a prior history of moderate/severe neurosensory impairment (5.9% vs 18.9%, p<0.001).

Table 2 lists the assessment measures, with full details of procedures available in our published protocol¹⁶ and in the primary reference for each outcome (see below).

View Figure 1, Table 1–4.

Key reported outcomes at 28 years

General health^17,18

In the previous year, 70% of VLBW adults and 81% of term controls had seen their primary care physician.¹⁷ Nevertheless, both unresolved and unrecognised physical health problems warranting medical attention were common in both groups (44% VLBW, 38% controls). These included body mass index (BMI) >30 plus a raised fasting insulin level, increased blood pressure (systolic >139mmHg or diastolic >90mmHg), abnormal echocardiogram or abnormal respiratory function tests. In the VLBW cohort, 31% were current smokers compared with 21% of controls. Although there were few differences in clinical oral health between the groups based on an oral health interview and standardised dental examination, proportionally fewer VLBW adults visited a dentist for check-ups or regularly cleaned between their teeth.¹⁸

Growth and metabolic markers¹⁹

Compared with term-born controls, VLBW young adults were 5cm shorter (mean difference, 95% confidence interval [CI]: males −4.9 [−7.3–−2.6]; females −5.3 [−7.2–−3.3]) and lighter (mean difference: males −4kg; females −10kg). BMI was no different for males but slightly lower for females. Waist-to-hip ratios were no different. These results are similar to international research, with some evidence BMI values follow societal trends.¹⁹

There were no differences in the fasting lipid profile of VLBW and control young adults. Most overseas studies have similar lipid measurement findings but some report higher low-density lipoprotein cholesterol (LDL-C) levels in preterm adults.²⁰ There were also no between-group differences in fasting glucose, fasting insulin, glycated haemoglobin (HbA_1c) or HOMA-IR levels (a measure combining both glucose and insulin levels to provide data on insulin resistance).¹⁹ Previous reports on glucose homeostasis are conflicting with some showing increased fasting insulin and HOMA-IR levels in VLBW²¹ and extremely low birthweight (ELBW; <1,000g)²² young adults.

The metabolic syndrome, associated with an increased risk of cardiovascular disease, is defined as central obesity plus any two of elevated triglycerides, reduced high-density lipoprotein cholesterol (HDL-C), elevated fasting glucose or elevated blood pressure.²³ We found a modest but non-significant increase in the rate of metabolic syndrome between VLBW adults (17%) and controls (12%). On logistic regression, male sex, gestational age <28 weeks, Māori/Pacific ethnicity and a BMI >90th percentile at 7–8 years were all significant predictors for having metabolic syndrome at age 28 years with an OR close to 3.¹⁹ A Finnish study reported an increased risk of metabolic syndrome (OR=3.7) in young adults born <34 weeks gestation.²⁴

Cardiovascular health^19,25,26

Systolic blood pressure (BP) was significantly higher in VLBW adults than in controls (mean difference, 95% CI: 4.6mmHg [1.8–7.5]), but there were no differences in diastolic BP. Systolic BP was significantly higher in males but not females.¹⁹ Higher systolic BP has been consistently reported in VLBW/VP adults with a mean difference of 4.2mmHg in one meta-analysis, but diastolic BP results vary.^20,27

Premature birth disrupts cardiac development with reduced cardiomyocyte endowment.²⁸ Among the NZVLBW cohort there were significant between-group differences in some aspects of cardiovascular structure and function and in epigenetic markers of cardiovascular development. In VLBW adults compared with controls, the left ventricular (LV) mass, volume and right ventricular (RV) dimensions were reduced.^25,26 In addition, minor decreases in ventricular function were noted in VLBW adults (decreased LV stroke volume, LV cardiac output, RV strain).^25,26 Several studies have now shown VLBW/VP adults to have smaller cardiac dimensions compared with term-born controls and reduced biventricular function (at rest and with exercise), but reports differ as to whether relative ventricular mass is decreased or increased.^25,29

We investigated endothelial function using the EndoPAT system reporting a reactive hyperaemic index (RHI) score. VLBW adults had significantly higher RHI scores, indicating stiffer microvasculature.²⁵ Two other groups have also used EndoPAT: one in ELBW adults with similar results to ours,³⁰ and one in extremely premature (EP; <28 weeks) adolescents when no differences were found.³¹ Ultrasound of the brachial or carotid artery has been undertaken by other groups with one report of increased carotid intimal medial thickness in VLBW adults compared with controls.³² In our study we also found increased LV and arterial elastance (measures of stiffness) on echocardiography, a change that is normally associated with ageing.²⁵

Between-group epigenetic differences were evaluated in a sub-group of the cohort, comparing DNA methylation at birth (using archived newborn blood spots) and at 28 years.³³ VLBW infants had altered methylation that is predicted to lead to perturbations of underlying gene signalling pathways involved in cardiovascular development and hypertrophy.³³ The altered methylation profiles at birth showed associations with cardiovascular and respiratory health in adulthood and may identify infants at higher risk of future adverse health outcomes.

Although differences in cardiovascular structure and function were relatively minor at 28 years, registry studies of whole populations show decreasing gestation increases the risk of future heart failure, ischaemic heart disease and stroke.^8,34 There is also some evidence that even well VLBW adults who are non-smokers may have occult pulmonary vascular disease.³⁵

Respiratory function^36,37

Increased respiratory obstruction has been consistently shown in studies of VLBW/VP adults.³⁸ For the NZVLBW cohort, an obstructive respiratory pattern (forced expiratory volume in 1 second [FEV₁]/forced vital capacity [FVC] ratio [FEV₁/FVC]below lower limit of normal, the 5th centile of the predicted value) was demonstrated in 35% of VLBW adults and 14% of controls, being mild in the majority. On spirometry, expiratory flow variables were all significantly lower in VLBW adults compared with controls; those with a history of bronchopulmonary dysplasia were most affected.³⁶ Other tests showed that compared with controls, VLBW adults had evidence of increased gas trapping, reduced gas exchange efficiency and higher ventilatory inhomogeneity.³⁶

Compared with controls, VLBW adults self-reported exercising less frequently and less vigorously. On cardiopulmonary exercise testing (CPET), although mean values were within the normal range, VLBW adults showed significantly reduced exercise capacity compared with control participants, characterised by an approximately 9% reduction in VO₂peak (the maximum oxygen consumption an individual can use in 1 minute per kilogram of body weight) with similar decreases in peak work rate, oxygen pulse (an indicator of cardiac output) and anaerobic threshold (reflecting cardiovascular fitness).³⁷ Our analysis suggested that both impaired lung function and altered cardiac structure and function contributed.³⁷ A meta-analysis of CPET studies in children and adults to age 21 reported a similar reduction in VO₂peak, but there are few studies in older VLBW/VP adults.³⁹

Visual outcomes⁴⁰

ROP is a potentially blinding condition in preterm infants resulting from abnormal development of retinal vessels in association with higher blood oxygen than experienced in utero. A unique feature of the NZVLBW cohort is the prospective documentation of ROP in the neonatal period and comprehensive assessment of vision in young adulthood. VLBW adults who had ROP had reduced visual acuity compared with VLBW adults without ROP and controls.⁴⁰ Moderate visual impairment (any of visual acuity >0.3 LogMAR [poorer than 20/40 Snellen], myopia >2 D, hypermetropia >2 D or astigmatism >2 D in the better eye) occurred in 33% VLBW participants with ROP, 20% VLBW participants without ROP and 14% controls. Fewer VLBW adults (79% vs 96%) drove a car, although factors other than vision contributed to this, including poorer visuospatial functioning.^40,41 These data demonstrate that VP birth has consequences for vision even in the absence of ROP. The NZVLBW study established the New Zealand protocol for routine retinal examination in VLBW/VP infants and, with treatment becoming available in 1987, severe visual impairment from ROP is now a rare occurrence.

Renal function¹⁹

Preterm birth disrupts kidney development with decreased nephron endowment.⁴² We found no differences in serum creatinine, the estimated glomerular filtration rate or early-morning urinary albumin creatinine ratio.¹⁹ There are few other data on renal function in VLBW/VP adults but a report from the Dutch POPS study showed VLBW/VP young adults had reduced renal growth.⁴³

Biological ageing⁴⁴

Although ageing is universal, the rate at which we age may differ with adverse life experiences impacting an individual’s “biological age” trajectory.⁴⁵ Biological age metrics have also been shown to better predict mortality than chronological age.⁴⁶ We assessed whether VLBW adults in the New Zealand cohort had a more advanced physiological age than controls by summing the z-scores for 10 physiological markers that alter with age, spanning metabolic, cardiac, respiratory and renal function.⁴⁴ The mean difference (95% CI) in the total z-scores was 1.73 (0.82–2.64, p<0.001). This represents a mean shift of 0.47 SD in the distribution of test scores for VLBW adults relative to controls; a moderate effect, with VLBW adults on average having a more advanced physiological age than their term-born peers.⁴⁴

Cranial magnetic resonance imaging (MRI) scans⁴⁷

Cranial MRI scans at 28 years were undertaken on a subsample of 150 VLBW adults and 50 term-born controls. All VLBW adults born <28 weeks (n=53; excluding three with contra-indications and one unreadable scan) underwent an MRI alongside a random sample of the remaining VLBW cohort (total= 150). VLBW adults had significantly reduced mean global grey matter volume, and non-significantly reduced mean white matter volume.⁴⁷ Global brain volume was found to predict both perceptual and total IQ (see below).⁴⁸

Cognitive outcomes^48–50

IQ was assessed at 28 years using the Wechsler Abbreviated Scale of Intelligence (2nd edition).⁵¹ There was a substantial gap in IQ with VLBW adults scoring on average 9.4 points lower than term-born controls, after adjustment for perinatal and socio-economic factors.⁴⁸ In the VLBW group, IQ scores at ages 7–8 and 28 years were highly correlated (r=0.78), suggesting stability of IQ over time, as also reported by others.⁵² On regression analysis, parental education strongly predicted both verbal and total IQ and birthweight strongly predicted perceptual and total IQ, with a modest beneficial effect of longer duration of breastfeeding across all IQ domains.⁴⁸ A recent individual patient data meta-analysis of eight longitudinal population-based studies of VLBW/VP adults, including the NZVLBW study, similarly reported an unadjusted IQ difference of approximately 12 points between VLBW/VP adults and controls.⁵³

Executive functions (EF) consist of higher-level cognitive regulatory skills that help an individual plan, solve problems and adapt flexibly to new situations. We examined participants’ self-perceived EF abilities in their everyday lives using the Behaviour Rating Inventory of Executive Function (BRIEF-A) at ages 22 and 28 years. VLBW adults reported poorer EF than controls at both ages. However, between-group differences reduced with increasing age, suggesting that executive difficulties may be having less impact on their daily lives.⁴⁹ Formal testing of spatial working memory at age 28 showed that VLBW adults were less accurate, slower and less efficient than controls.⁵⁰ This is generally consistent with our IQ results and EF findings from other studies suggesting that cognitive and EF challenges persist into adulthood.⁵⁴ Importantly, we also found that VLBW individuals with poorer spatial working memory had lower levels of educational achievement and occupational/socio-economic success, even after accounting for the effects of their parents’ socio-economic backgrounds.^49,50

Mental health⁵⁵

At age 28, VLBW adults reported more mental health problems than controls, although between-group differences were small.⁵⁵ Rates of agoraphobia, social phobia and suicidal ideation were all slightly elevated in the VLBW cohort compared with controls, but not depression. There were no differences in substance use or violent/property offending. Compared with same-sex controls, female VLBW adults had higher relative risks of anxiety disorders than VLBW males, while VLBW males had higher relative risks of depression. VLBW adults born <28 weeks had higher rates of most mental health disorders than more mature VLBW adults and controls, and might be at increased risk of depressive disorder/symptoms. We did not directly question participants about symptoms of attention-deficit/hyperactivity disorder (ADHD) at age 28, but data from our 22–23 year assessment¹⁵ were included in an individual patient data meta-analysis involving eight adult preterm outcome studies. Results showed similar rates of self-reported ADHD symptoms in preterm adults and controls.⁵⁶

Psychosocial outcomes¹⁵

Relative to controls, at age 22, VLBW adults had lower rates of tertiary education/training, were less likely to be in paid employment, more often living with their parents and had a closer relationship with them.¹⁵ They also reported fewer friends and less engagement in sexual partner relationships. Importantly, however, there were no differences in self-reported life satisfaction or self-esteem.¹⁵ In previously unpublished data, we found that by 28 years the gaps in tertiary education (23% vs 33% attained a university degree), employment (72% vs 79% paid employment) and welfare dependence (20% vs 12%) were reduced, and living arrangements were similar for both groups. Differences in peer, sexual and partner relationships remained but there were no differences in the proportion who had become a parent (37% vs 32%). Both groups reported similar self-esteem but life satisfaction was now slightly reduced for VLBW adults overall, largely due to poorer quality of life (QoL) scores related to physical health issues. Health-related QoL outcomes have been reported to be no different between VLBW/VP young adults and controls in their late teens and early 20s in several studies.⁵⁷ However, with increasing age other studies have noted some decline in self-reported QoL, primarily resulting from physical issues.^58,59

Table 3 summarises our main study findings.

Discussion and implications

The first 1,000 days (conception to age 2) of a child’s life represents a sensitive period of development during which a child’s experiences and exposures play a major role in shaping their future health and development.⁶⁰ Despite their difficult start in life, the majority of VLBW adults were generally living healthy, productive lives similar to their term-born peers. Biomedical measures for VLBW adults were mostly within the normal range, although group mean values favoured controls (Table 3). Our summation score of biomedical measurements that are known to change with age suggests that VLBW adults may have a more advanced “physiological age” towards the end of their third decade than term-born controls and may be more vulnerable to early organ decline and future health issues. In an overview of national Scandinavian registry data, Crump has shown that in their fourth and fifth decades preterm-born adults have increased risks of chronic disorders involving various organ systems, even though the absolute risk for most individuals is low.⁸

It is established that peak function for many physiological systems, for instance respiratory function, is attained by the third decade of life, followed by a decline.² Peak physiological function is influenced by genetic variation and modified by key exposures during critical windows of development, as well as by a range of social and environmental factors. Equally the rate of decline from peak function will be influenced by multiple factors including socio-economic status and lifestyle choices, such as eating habits and exercise. The current data show that peak function in VLBW/VP adults is reduced compared with term-born peers for many systems, although still within the normal range for most. What is less clear is whether the pace-of-ageing from that peak will match that of controls or be accelerated and whether appropriate preventative healthcare can modify the rate of function decline. Ongoing longitudinal research with VP-born cohorts is essential to answer these questions.

The New Zealand 1986 VLBW Follow-up Study is one of only two national longitudinal studies of VLBW/VP infants born in the 1980s—the other being from the Netherlands, which lacked a comparison group.⁶¹ However, there are several regional population-based cohorts and others based around regional referral hospital(s). Two cohorts of infants <26 weeks gestation born in the United Kingdom in 1995 and 2006 have been followed in the EPICure studies.⁶² These studies and a number of other trial-based cohorts contribute to the Adults Born Preterm International Collaboration (APIC),⁶³ which facilitates pooling of data and individual patient data meta-analyses resulting in increased statistical power. Saroj Saigal, who is a pioneer of longitudinal studies documenting outcomes following very premature birth, has also given space for ELBW young adults to voice their own perspectives on their life’s journey, challenges and accomplishments.⁶⁴ 

Strengths of our study include that it involves a prospectively enrolled, national population-based cohort with good retention to the third decade of life. We have used a comprehensive battery of tests and assessment undertaken at one centre. We have established international collaboration so that our data contribute to larger, often individual patient data, analyses.

Limitations include that there are some missing data, and our sample size means we are only powered to detect small-to-moderate between-group differences. The VLBW cohort was included in the original 1986 audit on the basis of birthweight (usual in the mid-1980s), so results are possibly affected by the increased proportion of small-for-gestational age (SGA; birthweight less than 10th centile) infants at higher gestations. However, results have also been reported for the sub-group born at <28 weeks gestation, although the smaller sample size for these analyses means reduced power to detect differences. Where differences or trends in outcome were noted for the total VLBW group compared with controls, these were often greater for those <28 weeks gestation or <1,000g birthweight. Since the VLBW cohort was born there have been many advances in neonatal intensive care. However, more than half of the cohort had been exposed to antenatal corticosteroids and routine care included assisted ventilation and parenteral nutrition, although surfactant therapy was not available and breastmilk feeding was lower than in contemporary cohorts.

In Table 4 we have listed a range of strategies we believe are important to maximise the health and wellbeing of VLBW/VP graduates throughout the lifespan. Early identification of developmental delay and provision of intervention services to support the cognitive needs of these children is critical. Educators within the school system need to be aware of the cognitive and learning needs of children born VLBW/VP so they can be proactively identified and addressed with timely remedial support to help optimise these children’s educational and occupational lifecourse opportunities.

VP birth is a lifetime condition and extra surveillance and monitoring is warranted, particularly for respiratory, cardiovascular, kidney and metabolic health. Optimising future health and welfare will be aided by healthy nutritional choices, encouraging regular exercise from childhood onwards, avoidance of smoking/vaping/mould exposure/air pollution and uptake of immunisation including for influenza and respiratory viruses. We recommend birthweight and gestation history is maintained in the medical records, or if absent questioned about at all visits with a health provider. We also recommend regular health checks to detect problems at an early stage.

VP birth has been associated with increased prevalence of mental health disorders in childhood and adolescence in a number of studies and a common preterm phenotype characterised by increased shyness and social withdrawal, increased susceptibility to attention and anxiety problems but low levels of risk-taking and oppositional behaviour.^65,66 However, the extent to which these traits persist into adulthood has been less certain.^55,67,68 Our mental health and psychosocial findings suggest at least some continuity in several of the challenges experienced earlier in development. Recognition of these issues might proactively help VLBW individuals self-monitor and develop strategies to accommodate these challenges to minimise impacts on everyday life.

Conclusion

The majority of VLBW young adults born in 1986 are living healthy, productive lives similar to their term-born peers. Most measures of physical health and welfare have means in the normal range, although these are less favourable for VLBW compared with term adults. It is possible that VLBW adults might experience chronic diseases earlier than their term peers, emphasising the importance of continued follow-up of this cohort.