Epidemiology of fecal mutagenicity.

Fecal mutagenicity presents unique difficulties of measurement (63); nonetheless, population studies of fecal mutagenicity typify the common methodological problems of colorectal cancer epidemiology. For example, correlational studies have linked fecal mutagenicity both to colorectal cancer incidence and to dietary practices postulated to increase the risk of colorectal cancer. The threat of an ecologic fallacy, however, is just as strong for biochemical assays as for other epidemiologic data. Fecal mutagenicity may represent, along with high-fat or low-fiber intake, a mere correlate of some true risk factor still to be elucidated in high-risk populations. Dietary trials, on the other hand, may directly confirm the influence of diet on fecal mutagenicity (or another presumed "intermediate end point") but cannot directly address the relationship of mutagenicity to colorectal cancer risk. Case-control comparisons of fecal mutagenicity might initially seem to promise more compelling evidence regarding colorectal cancer risk, but they are actually unsuitable, since the possibility of disease affecting this "exposure" is so strong. Specifically, fecal mutagenicity might be affected by the clinical tests required to diagnose colorectal cancer; it might also be affected by the malignancy itself, or indirectly by the dietary modifications that can accompany gastrointestinal illness. The resultant biases threaten to confound a case-control comparison as much as recall bias can affect interview data. A proper case-control study of fecal mutagenicity and colorectal cancer must therefore consider the effect of diagnostic workup on each mutagenicity assay, before the test is used to compare case and control specimens. In addition, case subjects should be followed through hospitalization and recovery, in the hope that for some cured patients the measurements after surgery will exclude any effect of disease on exposure. A case-control study incorporating these extra efforts is now underway. It seems reasonable to state that most biochemical assays incorporated into case-control studies of colorectal cancer will require similar modifications. Moreover, even if an effect of disease or of health care on biochemical measurements is excluded, an additional concern regarding the role of fecal mutagenicity or other assays in case-control studies will remain. The use of biochemical tests in a case-control context is analogous to asking questions regarding current smoking or current diet in a case-control interview. The current measurements reflect past values only to the degree that the exposure variable remains stable during the subject's lifetime.(ABSTRACT TRUNCATED AT 400 WORDS)