select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
An abstract presented at the American Society for Reproductive Medicine investigates whether preterm birth, preeclampsia and miscarriage can be predicted from peripheral blood throughout the first trimester with high reliability.
Prof. Basky Thilaganathan, Director, Fetal Medicine Unit, St George’s University Hospital Foundation Trust, said:
“Although this is an interesting study, I think the findings are overstated. The authors study a relatively small number of cases and controls in their series of experiments, so the sample size is small. The data used is already published in previous studies, and as far as we can tell the research is scientifically sound.
“My major concern about the study is that the authors rely on the receiver-operator characteristic (ROC) analysis to state that their findings have high predictive ability. However, for a ROC analysis to be reliable, a prospective screening study is required where the proportion of cases and controls represents that seen in a general population. By doing this analysis in a selected cohort with a much higher prevalence of cases than would be seen in a routine population, the ROC analysis will inevitably produce a falsely high test accuracy. The sensitivity and specificity of a test is altered by the prevalence of the disease in the population, and their tested cohort has a much higher prevalence of pathological cases than in the normal population. So, we can’t be sure the results reported by the authors can be applied to the general population.
“In terms of real world implications, it is far too early for there to be any – and there are no implications for women in the clinic today.”
Dr Jane Stewart, Consultant in Reproductive Medicine and Gynaecology, Newcastle Hospitals NHS Foundation Trust, said:
“Given how common miscarriage is and how diverse the causes may be, it is unlikely that a test linked to later pregnancy complications would produce a screening test of practical clinical value to predict first trimester loss. This research reports on only 16 pregnancy losses over two studies (120 pregnancies in four studies) so it is at a very early stage even in terms of predicting the more significant obstetric complications.”
Prof. Andrew Shennan, Professor of Obstetrics, King’s College London, said:
“Pre-eclampsia, preterm birth and miscarriage can all be related to placental problems and is has long been known that certain biochemical and biophysical markers in early pregnancy are related to them. This is therefore not surprising but the new technique and its predictive accuracy looks impressive. However there were less than 20 cases of each disease, so the results need to be confirmed. There are many unrelated causes of preterm birth and miscarriage as well, so a much larger sample is needed to confirm which causes it predicts. It remains unlikely that one test will predict all causes.”
Prof. Daniel Brison, Honorary Professor of Clinical Embryology and Stem Cell Biology; Scientific Director of the Department of Reproductive Medicine, University of Manchester, said:
“Although this is an exciting looking study in a much needed area, by its nature it is a small and preliminary study. Studies presented at conferences are usually preliminary, not final. Although these might seem exciting and cutting edge, there is unfortunately a high risk of them being wrong. We’d need larger follow-up studies to be sure whether these results are valid.
“Previous studies following up on conference abstracts show that a majority are never published as full peer-reviewed papers, which means of course that they turned out to be wrong when repeated in more patients, or contained such flaws as to make them unreliable.”
Prof. Marie-Louise Newell, Professor of Global Health, University of Southampton, said:
“This is not directly my field, but as an epidemiologist I would like to point out that this study reports on associations (not predictors) between a particular biomarker indicative of some placenta function during the first trimester of pregnancy (when the placenta is established). The study cannot predict – the findings cannot be taken to infer causality and so it is not correct to say that this would be a test to detect miscarriage (or even risk of miscarriage) during the first trimester. Further, the combined total of the four earlier studies of the same group was only 160, which is a relatively small number to base a generalised conclusion on. I note the authors themselves are cautious and suggest possible use of this particular biomarker in combination with other, established, screening tests.”
* Abstract title: ‘Preterm birth, preeclampsia and miscarriage can be predicted from peripheral blood throughout the first trimester with high reliability’ by E. E. Winger et al.
There is no paper.
Declared interests
Prof. Basky Thilaganathan: “I have no conflicts to declare.”
Dr Jane Stewart: “No COI.”
Prof. Marie-Louise Newell: “I have no conflict of interest here.”
None others received.