Objective To determine the survival and neurological outcome at 2 years of age of extremely low birthweight (ELBW, birth weight 500–999 g) infants born in the state of Victoria compared with term controls, and contrasted with ELBW cohorts from previous eras.
Design and setting A population-based cohort study of consecutive ELBW infants born during 2005 in the state of Victoria, and also in 1979–1980, 1985–1987, 1991–1992 and 1997.
Participants All 257 live births free of lethal malformations weighing 500–999 g in 2005, 220 randomly selected term, normal birthweight (birth weight >2499 g) controls, and equivalent cohorts born in earlier eras.
Main outcome measures Survival rates and quality-adjusted survival rates at 2 years of age, contrasted between cohorts.
Results Of 257 ELBW live births in 2005, 66.9% survived to 2 years of age, significantly lower than the survival rate of 75.2% for 1997 (odds ratio (OR) 0.67, 95% CI 0.45 to 0.99, p=0.046), but not after adjustment for confounders of birth weight, gestational age and gender (adjusted OR 0.73, 95% CI 0.46 to 1.16, p=0.18). This was a reversal of the steady increase in survival rates up to 1997. Rates of blindness, severe developmental delay and severe disability were significantly lower in 2005 than in ELBW survivors from previous eras. Consequently the difference in the quality-adjusted survival rates between 2005 and 1997 was only −3.8% (95% CI −11.4% to 3.7%, p=0.32).
Conclusions Regional survival rates for ELBW infants have plateaued since the late 1990s, but the neurosensory outcome in survivors has improved in 2005.
Statistics from Altmetric.com
Survival rates for extremely low birthweight (ELBW, birth weight 500–999 g) live births increased rapidly in the state of Victoria from the late 1970s to the late 1990s as improvements in perinatal and neonatal intensive care were applied to progressively more infants.1 However, the rates of long-term impairments and disabilities in survivors remained relatively constant and were much higher compared with term, normal birthweight (NBW, birth weight >2499 g) controls.1 Survival rates cannot increase indefinitely and there will always be some live born infants who die, even at term. However, it is important to re-evaluate neonatal intensive care for ELBW infants at intervals to ensure that standards and outcomes are at least maintained, even if they cannot be improved further. Also, some strategies that have improved long-term outcomes, such as caffeine therapy for apnoea of prematurity,2 have appeared in the 2000s. The effects of these therapies on whole populations of at-risk infants should be determined.
What is already known on this topic
▶ Survival rates for ELBW infants have risen with advances in perinatal care in recent decades, but cannot increase indefinitely.
▶ Neurological impairments and disabilities remain too high in ELBW infants compared with term, normal birthweight controls.
What this study adds
▶ Survival rates for ELBW infants have plateaued in the 2000s.
▶ Neurological outcome in ELBW survivors has improved in the 2000s.
The aims of this study were to evaluate the outcomes of neonatal intensive care for all ELBW live births in the state of Victoria in 2005 compared with results from both NBW, term controls from the same period, and earlier ELBW cohorts born in the state. The reason the cohort was selected by birth weight rather than gestational age was to allow comparisons between the ELBW cohort born in 2005 with previously reported results for ELBW infants born in the state in 1979–1980, 1985–1987, 1991–1992 and 1997.1 Gestational age was more certain only for cohorts born from the 1990s in the state of Victoria. Because there had been no important advances in perinatal care between the late 1990s and the mid 2000s that might have improved survival, it was hypothesised that survival rates would not have increased since the late 1990s, but that neurosensory outcomes in survivors would have improved with the widespread use of caffeine.
Over the whole period from 1979 until now there have been three level III perinatal centres (high-risk obstetric referral centres with a neonatal intensive care unit (NICU)) in the state of Victoria, and another NICU located within a standalone children's hospital. All four NICUs are located in Melbourne, the capital city of Victoria. Approximately 25% of Australia's population (now 22 million) live in Victoria, and approximately three-quarters of the population of Victoria live in Melbourne. Since the 1970s there has been a policy in the state favouring the transfer of mothers to a level III perinatal centre, wherever possible, when their babies are likely to require intensive care after birth.
The ELBW infants comprised all consecutive live births of birth weight 500–999 g born in the state of Victoria in 2005, excluding some live births that were late terminations of pregnancy because of lethal anomalies. Data on the total number of live births and their survival rates were collected prospectively from multiple sources, including all individual hospitals in the state through the Victorian Perinatal Data Collection Unit, the four NICUs and the Newborn Emergency Transport Service, the sole retrieval service for the state.3 Controls comprised randomly selected NBW (birth weight >2499 g) and term (37–42 weeks) live births taking place in each of the three maternity units affiliated with the three level III perinatal centres, matched with the ELBW or extremely preterm (EPT, gestational age <28 weeks) survivors according to the mother's health insurance status (private or public, as a proxy for social class), the main language spoken in her country of birth (English or other), and the child's gender. The outcome for the EPT cohort born in 2005 has been reported elsewhere.4
The survival rate was determined at 2 years of age, corrected for prematurity, when survivors were assessed by paediatricians and psychologists blinded to perinatal details. Impairments included cerebral palsy (CP), blindness (diagnosed by paediatric ophthalmologists during the first 2 years of life), deafness (requiring hearing aids or more advanced) and developmental delay. Development was assessed with the Bayley Scales of Infant and Toddler Development (Bayley-III), and Cognitive Scale and Language Composite Scale scores were obtained relative to the mean and SD for the controls on the respective scales. The scores for ELBW infants were compared with the controls, rather than the test norms, because the mean scores for the controls for the Cognitive and Language Composite Scales were significantly higher than would be expected from the test norms.5 The rates of developmental delay in the ELBW infants compared with same-age controls would have been substantially underestimated if we had simply relied on the normative mean of 100 and SD of 15.5 Mild developmental delay was defined as a score on either scale from −2 SD to less than −1 SD, moderate developmental delay was a score on either scale from −3 SD to less than −2 SD, and severe developmental delay was a score on either scale of less than −3 SD compared with the mean and SD for the controls. Children unable to complete psychological testing because of presumed severe developmental delay were assigned a score of −4 SD. The criteria for a diagnosis of CP included abnormal tone and loss of motor function,6 and its severity was assessed by the Gross Motor Function Classification System (GMFCS).7 Severe disability was defined as severe CP (unlikely ever to walk; GMFCS level 4 or 5), blindness, or severe developmental delay; moderate disability was defined as moderate CP (not walking at 2 years of age, but expected to walk eventually; GMFCS level 2 or 3), deafness, or moderate developmental delay; mild disability was defined as mild CP (walking at 2 years of age; GMFCS level 1), or mild developmental delay.1 Neurosensory utilities for survivors were assigned according to the severity of the disability: 0.4 for severe, 0.6 for moderate, 0.8 for mild and 1 for no disability.1 Utilities were multiplied for children with multiple disabilities, and therefore the lowest utility possible for survivors was 0.064. Infants who died had a utility of 0. Infants who survived but were not assessed were assigned a utility of 1. Utilities were summed and divided by the number of live births to calculate quality-adjusted survival rates.
Data were analysed by SPSS for Windows.8 Proportions were compared by χ2 analysis, and OR and 95% CI calculated, and by logistic regression analysis to adjust for confounding variables, calculating ORs and 95% CIs from the regression coefficients. Differences between groups for skewed interval data and for ordered categories were compared by Mann–Whitney U test or by χ2 analysis for linear trend. Quality-adjusted survival rates were compared by Mann–Whitney U test and Student t test. Because conclusions were the same for both methods, only the t test results, including mean differences and 95% CIs, are reported. Results for the 2005 ELBW cohort were compared with the control group, as well as with earlier ELBW cohorts born in the state of Victoria, particularly those born in 1997; methods and results from the 1979–1980, 1985–1987, 1991–1992 and 1997 cohorts have all been reported previously.1 In this earlier report lethal anomalies had been included in the previous cohorts—they were excluded in this study for comparison of survival rates.
The Research and Ethics Committees at the Royal Women's Hospital, the Mercy Hospital for Women and Monash Medical Centre, Melbourne approved these follow-up studies. Written informed consent was obtained from parents of term controls. Follow-up was considered routine clinical care for the ELBW survivors.
There were 267 ELBW live births in the state of Victoria in 2005, 10 of which were terminations of pregnancy for lethal anomalies. Almost 19% of these births took place outside a level III unit, and just over a half were female (table 1). Demographic data for the controls are shown in table 1.
Of the 257 live births free of lethal anomalies in 2005, 66.9% survived to 2 years of age, significantly lower than the survival rate of 75.2% (170/227) for ELBW live births free of lethal malformations in the cohort born in 1997 (table 2). In the 1997 and 2005 cohorts combined, being female (OR 2.15, 95% CI 1.33 to 3.46, p=0.002), more mature (per week increase; OR 1.56, 95% CI 1.34 to 1.80, p<0.001) and heavier at birth (per 250 g increase; OR 3.12, 95% CI 1.79 to 5.44, p<0.001) were all independently associated with a higher survival rate. Adjusting for these variables made the difference in survival rates between 2005 and 1997 smaller and non-significant (adjusted OR 0.73, 95% CI 0.46 to 1.16, p=0.18). For the lightest ELBW infants (birth weight 500–749 g), the survival rate was significantly lower in 2005 compared with 1997 (table 2). Adjusting for gender, gestational age and birth weight (as a continuous variable) reduced the difference in survival rates between eras and it became non-significant (OR 0.56, 95% CI 0.31 to 1.004, p=0.052). Part of the explanation for the higher mortality in the 500–749 g subgroup was that more infants were not offered intensive care and died in 2005 (not offered intensive care: 2005, 25.2% (33/131); 1997, 16.2% (17/105); OR 1.74, 95% CI 0.91 to 3.34). Despite this, the mean gestational age for survivors in this birthweight subgroup was lower, not higher, in 2005 than in 1997 (mean completed weeks (SD): 2005, 25.1 (1.5); 1997, 25.4 (1.7); mean difference −0.3, 95% CI −0.9 to 0.3). However, for the heavier ELBW infants (birth weight 750–999 g), the survival rate was higher in 2005 compared with 1997, but the difference was not statistically significant (table 2). Adjusting for gender, gestational age and birth weight had little effect on the relationship (OR 1.29, 95% CI 0.57 to 2.88, p=0.54). The changing survival rates over time since the 1970s, both overall and in the 250 g birthweight subgroups, are shown in table 2. Up to 1997 there were statistically significant stepwise increases in survival rates between earlier eras, as reported elsewhere.1
The follow-up rates to 2 years of age for the 2005 cohorts were 96% (165/172) for the ELBW infants and 92% (202/220) for the controls. One control was deaf at the 2-year assessment, but none had CP or were blind. Overall, significantly fewer controls had CP, developmental delay, or neurodevelopmental disability compared with the ELBW infants, but there were no substantial differences between the groups for blindness or deafness (table 3).
There were no significant trends over time in rates of CP, deafness, no developmental delay and severe developmental delay in the ELBW survivors from all eras (table 4). There was, however, a significant decrease in the rate of blindness over all eras (table 4). Compared with the previous cohort born in 1997, there were significantly fewer ELBW survivors with severe developmental delay in 2005 (OR 0.24, 95% CI 0.09 to 0.60, p<0.01). There were no significant trends over time in overall rates of no neurosensory disability and severe neurosensory disability for ELBW survivors from all eras (table 4). However, compared with the cohort born in 1997 there were significantly fewer ELBW survivors with severe disability in 2005 (OR 0.24, 95% CI 0.09 to 0.60, p<0.01).
The quality-adjusted survival rates rose progressively between successive eras until 1997, but then fell slightly in 2005 (table 5). The trend in the quality-adjusted survival rate over time was similar in the 500–749 g birthweight subgroup, but the fall between 1997 and 2005, although larger, was still not statistically significant. In the 750–999 g birthweight subgroup the quality-adjusted survival rate increased between successive eras, but the increase between 1997 and 2005 was not statistically significant. Statistical comparisons between earlier eras have been reported previously.1
The major findings of this study are that survival rates to 2 years of age plateaued for ELBW infants in the late 1990s, and indeed fell significantly for infants of 500–749 g birth weight in 2005. However, the survivors in 2005 had lower rates of severe developmental delay and severe disability compared with the late 1990s, resulting in improved quality of survival. Therefore there was no significant difference in the quality-adjusted survival rates in 2005 compared with 1997. The one significant trend over a longer period of time, from the late 1970s to 2005, in ELBW survivors has been the fall in the rate of blindness. As in all of our previous cohorts,1 the rates of some impairments and disabilities in ELBW survivors from the 2005 cohort remained much higher than in NBW controls.
One reason for the relative improvement in developmental outcome in survivors in 2005 might be that caffeine was widely used in 2005, whereas it was not used at all prior to the 2000s. Caffeine not only improves short-term outcomes, lowering rates of patent ductus arteriosus and bronchopulmonary dysplasia,9 but it also reduces the rate of CP and improves cognitive function in early childhood.2 Also, as the survival rate was lower in the lighter birthweight subgroup in 2005 compared with 1997, it is possible that some of the infants of birth weight 500–749 g who died might otherwise have been at higher risk of adverse neurological outcomes.
Although most studies now report outcomes by gestational age, rather than by birth weight, the reports of outcomes by birth weight in this study are important to be able to compare current data with historical ELBW cohorts, and to evaluate neonatal intensive care over longer periods of time. Outcome studies by gestational age are a relatively recent phenomenon following the improved obstetric dating of gestational age over the past few decades. However, vigilance must be maintained to avoid repeating the mistakes of the past; the longer the collective memory, the less likely it is that previous mistakes will be forgotten. In fact, the survival rates and overall neurodevelopmental outcomes for infants of 23–27 weeks' gestational age are almost identical collectively with those for infants of birth weight 500–999 g born in Victoria in the same era, where we have been able to report them from the 1991–1992,10 199711 and 20054 cohorts, which may not be the case in all other geographical cohorts of such infants.
Others have reported long-term outcomes by birth weight from cohorts born in the 2000s. Wilson-Costello et al12 reported the results of a large, single-centre, uncontrolled study of the outcomes of ELBW infants born in 2000–2005 compared with cohorts from the 1980s and 1990s. Of their cohort of 223, the survival rate of 71% at 20 months was similar to the current study; the rate had also plateaued since the 1990s. Also in agreement with the current study, the proportion of children at 20 months of age who were free of neurodevelopmental impairment increased over time. Tommiska et al13 compared data from the Finnish national register for ELBW infants born in 1996–1997 with those born in 1999–2000. Survival, including stillbirths, and the rates of CP and blindness did not change significantly over time; no control data were available.
In contrast to the results in the current study, a population-based Swiss study has shown an increase in survival for EPT infants at the limit of viability (22–25 completed weeks' gestation) between 2000–2001 and 2003–2004. Survival rose from 31% to 40%, with most of the increase due at 25 weeks' gestation; there were no controls and no long-term outcome data.14 Survival also rose from 36% in 1994–1999 to 47% in 2000–2005 in children born at 22–25 completed weeks' gestation in another region in England.15
The major strength of the current study is the recruitment of five geographically defined cohorts over three decades, using matched NBW, term controls as the basis for comparison. Another strength is the high follow-up rates for all cohorts, avoiding the problem of underestimating disability rates.16 Weaknesses include the imprecision with which cognitive tests in early childhood predict later outcome,17 but this is common in all studies reporting outcomes at this early age, including most randomised trials of perinatal interventions. It must be remembered that development is not a smooth process, and many children display early developmental delay, rather than cognitive impairment, and they will catch up later in childhood. From a clinical perspective, developmental outcomes at this age can certainly be used to identify children who may benefit from early intervention.
In conclusion, the increase in survival of ELBW infants over time has plateaued in the 2000s in the state of Victoria and the quality-adjusted survival rate has also been stable since the late 1990s. ELBW infants continue to have greater rates of adverse neurodevelopmental outcomes, particularly CP and developmental delay, compared with term controls, although there was less severe developmental delay and less severe disability in the latest cohort. Ongoing cohort studies of EPT and ELBW infants in Victoria are planned in order to continue monitoring outcomes. In the meantime, the search must continue for additional treatments to improve long-term outcomes, both before and after birth, such as magnesium sulphate given to women likely to deliver very preterm18 19 and docosahexaenoic acid supplementation added to the feedings of very preterm and tiny infants20. The effects of these treatments need to be re-evaluated when they are introduced into routine clinical practice.
Victorian Infant Collaborative Study Group Convenor: Lex W Doyle, MD, FRACP—Royal Women's Hospital, Murdoch Childrens Research Institute and University of Melbourne; Collaborators: Peter J Anderson, PhD—Murdoch Childrens Research Institute; Catherine Callanan, RN—Royal Women's Hospital; Elizabeth Carse, FRACP—Monash Medical Centre; Margaret P Charlton, M Ed Psych—Monash Medical Centre; Mary-Ann Davey, PhD—Victorian Perinatal Data Collection Unit; Noni Davis, FRACP—Royal Women's Hospital; Julianne Duff, FRACP—Royal Women's Hospital; Rod Hunt, PhD, FRACP—Royal Children's Hospital; Cinzia de Luca, PhD—Royal Women's Hospital; Marie Hayes, RN—Monash Medical Centre; Esther Hutchinson, DPsychClinNeuro—Royal Women's Hospital; Elaine Kelly, MA—Royal Women's Hospital and Mercy Hospital for Women; Marion McDonald, RN—Royal Women's Hospital; Gillian Opie, FRACP—Mercy Hospital for Women; Gehan Roberts, PhD, FRACP—Royal Women's Hospital and Royal Children's Hospital; Michael Stewart, FRACP—Royal Women's Hospital and Royal Children's Hospital; Linh Ung, BSc(Hons) —Royal Women's Hospital; Andrew Watkins, FRACP—Mercy Hospital for Women; Amanda Williamson, MA—Mercy Hospital for Women; Heather Woods, RN—Mercy Hospital for Women.
Funding Supported in part by the National Health and Medical Research Council, Australia—Project Grants 108702 and 454413.
Competing interests None.
Ethics approval This study was conducted with the approval of the Royal Women's Hospital, The Mercy Hospital for Women, and Monash Medical Centre, Melbourne.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.