We thank van Leeuwen et al.1 for an interesting comment on our article on fasting capillary blood glucose as a screening test for gestational diabetes mellitus (GDM). In general, we agree with them that a multivariate analysis would be the optimal method for evaluating the relative importance of each risk indicator on the risk of GDM.
However, traditional risk factors (first degree relative with diabetes mellitus, obesity, prior large for gestational age infant, or prior GDM) treated as an entity being either present or absent, have in clinical practice long been used as a screening test for GDM, and in Sweden, traditional risk factors are recommended by the National Board of Health and Welfare for use in pregnancy as a screening test for GDM.
In our present study, the main objective was to assess the diagnostic accuracy of fasting capillary blood glucose. To compare it with traditional risk factors as a screening test, we applied both tests to the same population using an oral glucose tolerance test (75 g) as the diagnostic reference test.
In this population, we did not find that adding traditional risk factors to the model substantially changed the test performance. We cannot exclude that a multivariate analysis could somewhat alter the conclusions on the combined model, although we consider it unlikely that it would have a large impact.
We still consider our main conclusion on fasting capillary blood glucose as an acceptable screening test for GDM is valid.