Both perinatal death and cerebral palsy are more likely in babies whose birthweight is lower than expected for gestational age but data for lower gestational ages are inadequate. Babies whose birthweight is high for gestational age have an increased risk of perinatal death and possibly of cerebral palsy. Babies who grow slowly in utero are more likely to be born early and therefore weight standards based on birthweight may be too low at earlier gestational ages. In a European collaborative study (

, see also commentary, ibid: 1089–90) an attempt has been made to circumvent this difficulty by using fetal growth standards based on ultrasound estimations of weight during pregnancy of healthy babies born at term. Data were gathered from 13 cerebral palsy registers in eight countries of which those from three registers in two countries were excluded from the published analysis. The birthweights and gestational ages of 4503 singleton children with cerebral palsy born between 1976 and 1990 were compared with published reference standards from the North of England and from Sweden. Rates of cerebral palsy were calculated from local population data.

Using weight for gestation standards based on birthweight different patterns were seen for babies born before or after 32 weeks gestation. After 32 weeks rates of cerebral palsy were lowest at a weight for gestation Z score of between 1 and 2 (equivalent to 75th to 90th percentile). There was a reverse-J curve with the highest rates of cerebral palsy at very low and very high Z scores. Before 32 weeks the lowest cerebral palsy rates were at weight for gestation Z scores of between –1 and –2 (about 3rd to 10th percentile). Using fetal growth rate standards this different pattern at earlier gestational ages was no longer seen. (This finding seems difficult to explain and they offer no explanation). Now, at all gestational ages there was the same reverse-J curve with the lowest risk of cerebral palsy at Z scores of between 1 and 2. The findings were similar for all types of cerebral palsy. For babies born between 32 and 42 weeks of gestation the risk of cerebral palsy was increased 4–6 fold at birthweights below the tenth percentile compared with birthweights between 25th and 75th percentile. For birthweights above the 97th percentile the increase was 1.6–3.1-fold.

The optimum birthweight for avoiding cerebral palsy is at 75th to 90th percentile for gestational age (using fetal growth standards at lower gestations). The risk is increased at lighter or heavier birthweights for gestational age. Whether abnormal fetal growth is a cause or a result of cerebral palsy is not known and there is uncertainty about whether the abnormal growth is proportionate (affecting all aspects of growth) or disproportionate (affecting mainly weight). Either way, the abnormal growth starts long before birth.