Parental age and birth order were studied in 251 patients with cerebral palsy. No parental age or birth order effects were observed in spastic quadriplegia or diplegia, but a paternal age effect was detected in those with athetoid/ dystonic cerebral palsy and congenital hemiplegia. These observations indicate that some cases of athetoid/dystonic or hemiplegic cerebral palsy might arise by fresh dominant genetic mutation. (J7 Med Genet 1993;30:44-6)
MethodsAll cases were ascertained from the records of the Cheyne Centre for Spastic Children, London. All had been referred from general practitioners or paediatricians in south-east England and were seen and examined personally by one of the authors (JF) between 1955 and 1986. CP was defined as above, low birth weight (LBW) as < 2500 g, and prematurity as birth before 37 weeks' gestation.Cases with recognisable syndromes or malformations, genetic or metabolic disorders, kernicterus, other identifiable causes for the motor disorder (including postnatal factors), similarly affected relatives, ataxic CP, or asymmetrical quadriplegia were excluded.The records of 532 cases were examined. Of these, 172 had spastic diplegia (65% preterm, 69% LBW), 102 had congenital hemiplegia (23% preterm, 20% LBW), 104 had symmetrical spastic quadriplegia (26% preterm, 35% LBW), and 154 had athetoid/dystonic CP (14% preterm, 19% LBW). Sibship size, birth order, parental ages at the time of birth, and date of birth were recorded. Of the 532 cases, 399 were born in the UK to parents whose ages had been recorded; 251 of these were born between 1961 and 1983 which is the period for which the Registrar General has published annual mean parental ages for the UK population.'0 Methods of birth order and parental age analysis were those described by Emery.11 Birth order was analysed using the method of Haldane and Smith. 12 For each case, the observed birth order is recorded and the expected birth order and its variance (assuming no birth order effect) calculated from the sibship size and published tables." In each group of cases, the standard error of the difference between the sums of the observed and expected birth orders is the square root of the sum of the variances. Parental ages at the time of birth of each case were compared with mean UK population maternal and paternal ages for the corresponding year using paired t tests. For each group of cases, mean differences between observed and expected values of paternal age (dp), maternal age (dn)j and birth order (db) were calculated.As dp, dm) and db are mutually dependent variables, a method based on regression analysis was used to compare any statistically significant value of one with that predicted by the observed values of the other two."I 13 The difference between observed and predicted values is d,-b12 d2, where di is the mean increase in one of the three variables (dp, dm3 or db), d2 is the mean increase in one of the other two, and b12 is a regression coefficient given by s1r12/s2 44