Interpretation of results and implications for clinical practice
The observed patterns of racial differences in development of hypertensive disorders of pregnancy are consistent with those observed in the development of cardiovascular diseases in non-pregnant women.39 In a large UK registry of 1,068,318 patients between 1997 and 2010 from 225 general practices across England, black, compared to white patients, were more likely to present with ischaemic stroke and intracerebral haemorrhage while South Asian patients had significantly higher hazard ratios for angina and myocardial infarction.40 This study also showed that the median age of first cardiovascular disease diagnosis was substantially lower in women of black and South Asian racial origin than white women.
Prediction of PE and GH necessitates first, data obtained from large prospective observational studies with accurate recording of maternal demographic characteristics and medical history and the appropriate infrastructure for obtaining the necessary outcome measures, and second, multiple logistic regression analysis which defines the independent contribution of each risk factor. The data from the FMF study fulfil these criteria and there are several elements from the maternal history that contribute to PE and GH; in defining the specific contribution of one risk factor, such as black racial origin, it is essential that all other factors are taken into account. In the development of the FMF competing risks model for prediction of PE a wide range of maternal factors are taken into account to derive the prior risk which is then adjusted with the addition of biomarkers to obtain the posteriorrisk.41
This systematic review and meta-analysis has highlighted the weakness of such approach in defining the contribution of one specific risk factor such as racial origin. Although the combined number of patients arising from such studies can be very large the heterogeneity between individual studies and the lack or minimal adjustment for confounders produces results that cannot be used for accurate prediction of the outcome under investigation.