An epidemiologic perspective on biomarkers

The authors discuss biological markers from an epidemiologic perspective, emphasizing the importance of integrating biomarkers into large-scale observational and intervention studies. Whereas any biologic phenomenon can be considered a biomarker, an intermediate end-point is defined as being on the causal pathway between exposure and disease. An intermediate end-point is a valid surrogate for a disease in relation to a given exposure if, and only if, that exposure causes a similar change in the occurrence of both the intermediate end-point and the disease. Cancer studies using surrogate end-points may be shorter, smaller and cheaper than those using malignancy per se as an outcome. Three types of studies may be carried out to determine whether a given biomarker is an intermediate end-point and whether it can serve as a surrogate: (i) exposure-marker studies, (ii) marker-disease studies, and (iii) studies comprising all three elements, exposure, marker, and disease. The authors discuss statistical aspects of these three types of studies and provide examples from investigations of alcohol-hormone-breast cancer, diet-epithelial proliferation markers-large bowel adenomatous polyps, and reproductive risk factors-human papillomavirus infection-cervical cancer.