Biomarkers can help indicate if people are likely to develop a disease and how they will respond to treatment, but the connection may be overstated, according to a recent analysis by John Ioannidis and Orestis Panagiotou. Their paper, published in the June 1 issue of the Journal of the American Medical Association (JAMA), examined 35 studies that had been cited at least 400 times and were also part of meta-analyses. For 29 papers, the meta-analyses showed a weaker link between the biomarker and the disease than the initial study. For 30 papers, a weaker association was evident in the largest study of the biomarker. Dr. Ioannidis and Dr. Panagiotou observed that studies with the most optimistic conclusions tend to be cited. Commenting on the JAMA paper, Patrick Bossuyt of the University of Amsterdam noted that the results may be biased because meta-analyses tend to examine conflicting studies.

Source: Nature