Cerebral palsy and intrapartum careBMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6963.1229 (Published 05 November 1994) Cite this as: BMJ 1994;309:1229
Wrong denominator used
- E Blair
EDITOR, - Geraldine Gaffney and colleagues calculated a significantly increased odds ratio for cerebral palsy and for perinatal death according to two definitions of suboptimal intrapartum care.1 Care was defined as suboptimal if there was failure to respond in specified ways to one of a large number of suboptimal conditions or if instrumental delivery met one of two descriptions. Therefore care was only at risk of being classified as suboptimal in the presence of one (or more) of those suboptimal conditions. In order to conclude that suboptimal care was or was not associated with an outcome, the proportion of cases and controls receiving suboptimal care who were at risk of suboptimal care must be compared.
Gaffney and colleagues did not say the number of cases and controls experiencing the conditions that put them at risk of being classified as receiving suboptimal care. However, they gave sufficient information - for example, the frequency of low Apgar scores at five minutes, ominous cardiotocograms, and meconium staining of amniotic fluid - for readers to infer that (as anticipated) cases were more likely than controls to be classified as having fetal distress or poor neonatal condition. These were the two conditions that must be present for a subject to be at risk of receiving the types of suboptimal intrapartum care associated with high odds ratios of cerebral palsy or perinatal death. The odds ratios for cerebral palsy and perinatal death shown in table V are therefore too high.
A similar error was made by Richmond et al2 but the numbers of cases and controls at risk of being classified as receiving suboptimal care were provided, allowing an alternative analysis to be performed. Richmond et al found a significant association between cerebral palsy and suboptimal intrapartum care, the mean estimate of attributable proportion being 9.2% of cases at term. Analysis using only those at risk as the denominator suggested that the association was not significant and that the attributable proportion was only 2.1% of cases at term. The matched experimental design and lack of relevant information preclude a similar analysis of the study by Gaffney and colleagues.
- A Johnson,
- G Gaffney
EDITOR, - Eve Blair's central point is the issue that we addressed in detail in our discussion - that is, mothers of children with cerebral palsy have a higher frequency of both adverse antenatal factors and signs of intrapartum distress than the mothers of children who do not have cerebral palsy. (These numbers and proportions are shown in table III of our paper.) As we stated clearly in the text, these are potential confounding variables. In other words, they could be associated both with the risk of suboptimal care (more complex clinical decisions would be needed in the cases than in the controls) and with the outcome of cerebral palsy (the adverse antenatal factors and the factors underlying the intrapartum fetal distress may have a causal role in cerebral palsy). Within the constraint of the matched design, we adjusted for known confounding variables when this was possible; these are reported. Although our study was considerably larger than previous studies, its size and design (matched case-control) limited the analysis because too few matched pairs or triads could be included when some variables were adjusted for. We accept that had adjustment for the presence of intrapartum distress been possible, the odds ratio might have been reduced and confidence intervals widened. As we pointed out in previous correspondence,1 there is clearly a need for larger collaborative studies to clarify this point.
In the meantime, we consider that the carefully worded, cautious discussion and conclusions of our paper hold. Suboptimal intrapartum care seems to be a rare cause of cerebral palsy; the origins of cerebral palsy in singleton babies born at term lies mainly in the antenatal period.