Predicting stillbirths - still a distant reality.

Abstract

Although there has been an encouraging 14.5% reduction in the global stillbirth rate between 1995 and 2009, it is still very high at 18.9/1000 births worldwide and 3.9/1000 births in high-income countries (Cousens et al. Lancet 2011;377:1319–30). All-cause stillbirth ranks fifth among the global causes of deaths and interestingly ranks before diarrhoea, HIV/AIDS, tuberculosis, traffic accidents and cancer (Mathers et al. World Health Organization, 2008). Despite this, stillbirths are hardly included in global health targets and, until recently, data were not routinely collected or collated by UN agencies – a stark example of the ‘scandal of invisibility’. The tragedy of a stillbirth has profound medical, social and psychological consequences. The ability to predict stillbirths early in pregnancy and to offer additional surveillance and interventions to high-risk mother/fetus is understandably an important part of antenatal care. This well-conducted systematic review by Conde-Agudelo et al. is therefore extremely timely and important despite being unable to identify a clinically useful first- or second-trimester test for predicting stillbirth as a sole category. This conclusion is unsurprising because (1) the definition of stillbirths varies between geographical regions from 20 to 28 weeks of gestation, (2) the aetiology of stillbirths is multifactorial, (3) there is paucity of high quality and adequately powered studies on the prediction of stillbirth, (4) most studies attempt to extrapolate a variety of biochemical and biophysical tests and thresholds without adequate validation, and (5) there is a significant heterogeneity between studies. There are some methodological issues in the conduct and reporting of this review as well. First, there is still a lack of consensus on the appropriateness of the use of summary values in the meta-analyses of diagnostic tests, although the authors elected to combine likelihood ratios. Secondly, significant methodological flaws in as many as 79% of the included studies and the heterogeneity in approximately 40% of the meta-analyses meant that some conclusions had to be drawn from single studies, further limiting credibility. These limitations notwithstanding, the conclusions are important for clinicians, researchers and policy-makers. We are reminded that despite the advances in antenatal and intrapartum care, in 2014 there is still no clinically useful test to predict stillbirth, challenging researchers and clinicians. Although uterine artery pulsatility indices and maternal serum PAPP-A levels show some promise in predicting stillbirths related to placental dysfunction, their utility and accessibility in low-income countries where 76.2% of the stillbirths occur, is questionable. Finally, we ponder the concept of ‘life expectancy at birth’. At the beginning of the third-trimester of pregnancy, when most babies are viable, the risk of stillbirth is still about 2%, a risk matched only when people reach their 80s (Froen et al. Lancet 2011;377:1353–66). We believe that we need to amend our conceptual mindset by considering ‘life expectancy at conception’ rather than beguiling ourselves with increasing ‘life expectancy at birth’, which deliberately excludes stillbirths from such a widely used comparative marker. Mounting global socio-political pressure in maternal-child health may realize this goal provided that counting of life starts ‘at the right time and everyone is counted’. The authors have no conflicts of interest to disclose.