Highly-cited biomarker studies tend to report larger effect estimates for associations than subsequent meta-analyses do, says a literature review published in the Journal of the American Medical Association. John Ioannidis and Orestis Panagiotou combed through the ISI Web of Science to find studies that had been cited more than 400 times and had been published in a highly cited journal, and they searched for meta-analyses of those studies. From this, they evaluated 35 highly cited associations.