Statistics from Altmetric.com
Editor—Gray et al 1 recently reported an apparently new observation that “maternal hypertension has a protective effect against cerebral palsy in very preterm infants.” Do their data support this conclusion, which is different from the conclusion reached by two other case control studies2 3 on this topic?
In the same issue, Tin et al 4remind us that those infants who are difficult to follow up have different (worse) outcomes than those infants who could be followed up without great difficulty. They provide a key message that “studies where it is not possible to see some children for assessment might usefully include a calculation of what the total prevalence would be if there was a fivefold difference in the proportion with the condition in question among the children who were not seen.”
Gray et al’s study comprised 107 in the study group, of whom there were 101 survivors. At 2 years of age, four in the study group were lost to follow up as were two in the control group. Using the recalculation suggested by Tinet al, there is no longer a significant difference between the estimated incidence of cerebral palsy in the two groups.
Spinillo2 et al reported cerebral palsy in 2/92 (2.1%) cases and 4/184 (2.1%) controls. However, their two year follow up was 92 of 97 cases. Ascertainment of the outcomes of the remainder of the group, and applyingTin et al’s “correction” would not have changed their conclusion of no difference in the incidence of cerebral palsy between infants born to hypertensive mothers and controls.
The case control study by Szymonowicz and Yu3 reported 27 cases and 26 controls and came to exactly the same conclusion as Grayet al. All had a two year follow up, but it is not stated in their paper whether complete follow up was achieved or whether the sample was selected retrospectively on the basis of the availability of two year follow up data.
A prospective study designed to have 80% power to show a protective effect of pre-eclampsia of 25% against cerebral palsy, assuming a 5% cerebral palsy rate in the control group, would require more that 4000 babies in each study arm. Even to detect a 50% reduction from a 10% rate of cerebral palsy in the controls would require about 500 babies in each arm. The combined number of babies of hypertensive mothers in the three case control studies is 216.
In view of Tin et al’s timely demonstration of the importance of the effects of incomplete follow up and the above considerations, it is prudent to await further prospective studies before accepting Gray et al’s conclusion that “maternal hypertension has a protective effect against cerebral palsy in very preterm infants.”
Dr Gray responds: Professor Colditz cited the report of Tinet al 4 and concluded that a recalculation of our data (on the basis of those infants lost follow up) would reveal that there would no longer be a statistical difference in the estimated incidence of cerebral palsy between study and control groups.
While we reported neurodevelopmental outcome of the infants at 2 years, in general, cerebral palsy of functional significance can be reliably diagnosed by the age of 12 months, One of our four study infants not seen at 2 years was assessed at 1 year, while both control infants not seen at 2 years had follow up data at 1 year of age. None of these three infants had cerebral palsy. Thus cerebral palsy occurred in 0/98 (0%) study infants compared with 5/104 (4.8%) controls.
Tin et al suggested that one might include in results “a calculation of what the total prevalence would be it there was a fivefold difference in the proportion with the condition in question among children who are not seen.” As the incidence of cerebral palsy in the study children was 0%, a fivefold increase would remain 0%. If, however, one calculated the incidence of cerebral palsy in the total population (5/202, 2.5%) and made a fivefold increase, the incidence of cerebral palsy would be 12.5% in those infants lost to follow up. Thus in both scenarios, of our three infants, less than one would be expected to have cerebral palsy and the significance between the groups would remain.
We acknowledge that our results are at variance with two previous cohort studies,1-5 1-6 but they agree with the recent case control study of cerebral palsy in preterm infants. Murphyet al 1-7 found a decreased risk (relative risk 0.4; 95% confidence intervals 0.2–0.9) of cerebral palsy when maternal hypertension/pre-eclampsia was present. Accordingly, while a large multicentred cohort study would be useful, we believe that the current evidence does suggest that maternal hypertension protects against cerebral palsy in preterm infants. We therefore agree with Collins and Paneth,1-8 who concluded in their annotation that preterm infants born to mothers with pre-eclampsia have a lower risk of cerebral palsy than other preterm infants.
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.