Results

During 1991–2 there were 429 consecutive live births with birthweights 500–999 g, of whom 242 (56.4%) survived their primary hospital stay. One child subsequently died before 2 years of age following genitourinary surgery at 1 year of age. Survival to 2 years of age was therefore 56.2%, a 48% relative increase compared with those in 1985–7 (37.9%), which in turn was a 49% relative increase compared with those in 1979–80 (25.4%), all differences being significant (table 1). Survival was significantly higher in heavier infants in each era (table 1). The larger increase in survival between 1979–80 and 1985–7 was in infants of birthweight 750–999 g; the small increase in survival in infants of birthweight 500–749 g was not significant. Survival increased significantly in both birthweight subgroups in 1991–2 compared with 1985–7, but the larger increase was in infants of birthweight 500–749 g (table 1).

Of the 265 normal birthweight controls for the 1991–2 cohort, two died before 2 years of age, three emigrated, the families of four refused the two year assessment, and 14 were lost to follow up. Of the 242 normal birthweight children assessed at 2 years of age, four had alternative psychological tests and four had severe developmental delay and scored < 50 on the MDI. The MDI for the remaining 234 normal birthweight controls was 115 (SD 20). Of the 242 normal birthweight controls assessed at 2 years of age, none had cerebral palsy, blindness or deafness; disability in the controls was therefore assessed solely on the basis of psychological tests.

Of the 241 ELBW survivors in 1991–2, 237 (98.3%) were assessed at 2 years of age, corrected for prematurity. Of the remaining four children, two emigrated, the families of one refused the two year assessment, and one was lost to follow up. Twenty two (9.3%) of the 237 survivors assessed had some form of cerebral palsy, predominantly quadriplegia (table 2). The disability imposed by the cerebral palsy was severe in nine children, moderate in four, and mild in nine. Five (2.1%) children were blind, three from retinopathy of prematurity and two from cortical blindness, both of whom had severe quadriplegia. Two (0.8%) children required hearing aids for sensorineural deafness, one of whom also had a DQ of −4 SD. The rates of cerebral palsy and deafness were not significantly lower than in children in the 1979–806 and 1985–75 cohorts assessed at 2 years of age, but the rate of blindness was significantly lower in 1991–2 compared with 1979–80 (table 2). Fourteen (5.9%) children, all of whom were unable to complete the MDI because of blindness, severe cerebral palsy, or severe developmental delay, had DQ scores <−3 SD, 15 (6.3%) children had DQ scores from −3 SD to <−2 SD, 33 (13.9%) had DQ scores from −2 SD to <−1 SD, and 174 (73.4%) had DQ scores >− 1 SD. One child in a non-English speaking family could not complete the MDI because of extreme shyness but was considered by the testing psychologist to be probably normal, although no standardised DQ score was assigned. As the child was free of other impairments, she was considered to have no disability; she has been fully assessed subsequently and is free of disability at 5 years of age. For ELBW children born in 1991–2 able to complete psychological tests, the mean MDI was 99 (SD 20, n = 220), significantly higher than for the children able to be tested in the 1979–80 cohort (mean MDI 91; SD 17, n = 74)(t = 3.1, P < 0.01), but not significantly different from the 1985–7 cohort (mean MDI 99; SD 18, n = 183). For the 1991–2 cohort, overall, sensorineural disability was severe in 16 (6.8%); nine (3.8%) children had multiple severe disabilities. Compared with the 1979–80 ELBW cohort, the overall rate of disability was significantly lower in the 1985–7 and 1991–2 ELBW cohorts; the rates of disability were not significantly different in the 1991–2 and 1985–7 ELBW cohorts (table 3). For the 1979–80 cohort, the overall sensorineural disability rate was significantly worse in survivors of birthweight 500–749 g compared with those of birthweight 750–999 g, but there was no significant difference between the birthweight subgroups for the 1985–7 and 1991–2 cohorts (table 3).

For the 1991–2 ELBW cohort compared with the normal birthweight controls, the overall rate of disability was significantly higher in ELBW survivors (Mann-Whitney U test, z = 6.4, P <0.0001) (table 4).

Discussion

The survival rates for ELBW infants in Victoria have risen progressively since the late 1970s, through the 1980s, and now into the 1990s. Survival rates rose more for larger infants, with birthweights 750–999 g, into the 1980s, but more for the smaller infants, with birthweights 500–749 g, into the 1990s. Improvements in perinatal care, such as antenatal steroids and exogenous surfactant (first used in Victoria in 1991) have been partly responsible, but recognition of rising survival rates and an increasing willingness and ability to treat tiny babies have also probably contributed.

As well as a higher survival, sensorineural outcome improved for survivors born in 1985–7 compared with those born in 1979–80.5 However, despite an improvement in survival, there is no indication of a further reduction in the rates of adverse sensorineural outcome in survivors from the 1980s into the 1990s. Moreover, the rates of sensorineural impairments and disabilities in ELBW infants born in the 1990s remain higher than for normal birthweight infants.

The different categorisation of disability for ELBW infants in 1991–2 based on psychological test scores for the normal birthweight controls highlights the importance of control groups. However, the number of ELBW and normal birthweight children born in 1991–2 surviving with severe disability was not affected by the reclassification. It would have been desirable to recruit control groups for the 1979–80 and 1985–7 ELBW cohorts to permit a more appropriate classification of disability for these groups.

The sensorineural outcome for the tiniest survivors of birthweight 500–749 g was significantly worse than for the larger infants of birthweight 750–999 g in only one of our three cohorts. In other regional studies Hack et al 3 reported higher rates of adverse sensorineural outcome for children of birthweight < 750 g compared with those of birthweight 750–1499 g, and Frenchet al 10 described higher sensorineural disability rates in children of birthweight 500–799 g compared with those of 800–999 g. However, it is important to realise that most of the survivors of 500–749 g birthweight in our study were not severely disabled, and the absolute numbers of children with disabilities of all severities were fewer for those of birthweight 500–749 g compared with those of birthweight 750–999 g in each era.

A comparison of the absolute number of ELBW and NBW children expected to survive per year might help to reduce anxiety that the improving survival rates for the tiniest babies will result in large numbers of severely disabled children in the community. Using the definitions of disability in our study, if the rate of severe disability of 1.7% in our normal birthweight control group truly reflects the rate for all such children born in Victoria, as the number of normal birthweight children born in Victoria 1991–2 was 123 10011 12 we would expect about 1000 normal birthweight children to have severe disability every year, compared with only eight children with severe disability per year in the 1991–2 ELBW cohort.

Although ELBW survivors are few, the number with adverse sensorineural outcome should be reduced if possible. Several treatments offer promise, such as antenatal magnesium sulphate13 and postnatal indomethacin,14 and are currently being evaluated in large multicentre randomised controlled trials, with the major endpoint being to reduce adverse sensorineural disabilities. Trials such as these should be strongly promoted and supported.

In conclusion, survivals of ELBW infants in this regional cohort have improved further in the early 1990s. However, rates of sensorineural disability in ELBW survivors have not improved since the mid-1980s and remain too high.