Our study evaluated the performance of a selective screening strategy for hyperglycaemia in pregnancy (HIP) based on the presence of risk factors (RFs; body mass index≥25kg/m², age≥35years, family history of diabetes, personal history of HIP or macrosomic infant) to diagnose HIP and to predict HIP-related events.

Women with no known diabetes who had undergone complete universal screening (early, before 22weeks of gestation and, if normal, in the second part of pregnancy) at our department (2012–2016) were selected, resulting in four groups of women according to the presence of HIP and/or RFs, with a predefined composite endpoint (preeclampsia or large-for-gestational-age infant or shoulder dystocia).

Included were 4518 women: 23.5% had HIP and 71.1% had at least one RF. The distribution among our four groups was: HIP−/RF− (n=1144); HIP−/RF+ (n=2313); HIP+/RF− (n=163); and HIP+/RF+ (n=898). HIP was more frequent when RFs were present rather than absent (33.1% vs 15.4%, respectively; P<0.001). Incidence of the composite endpoint differed significantly (P<0.0001) across groups [HIP−/RF− 6.3%; HIP−/RF+ 13.2%; HIP+/RF− 8.6%; and HIP+/RF+ 17.1% (HIP effect: P<0.05; RF effect: P<0.001; interaction HIP * RF: P=0.94)] and significantly increased with the number of RFs (no RF: 6.3%, 1 RF: 10.8%, 2 RFs: 14.7%, 3 RFs: 28.0%, 4–5 RFs: 25.0%; P<0.0001).

RFs are predictive of HIP, although 15.4% of women with HIP have no RFs. Also, irrespective of HIP status, RFs are predictive of HIP-related events, suggesting that overweight/obesity, the only modifiable RFs, could be targets of interventions to improve pregnancy prognosis.