In our previous studies of the determinants of carotid plaque area (CPA), we used a linear multiple regression model, which permitted us to control for the presence of known risk factors in order to reveal the contribution of putative new risk factors. We recognized that this approach could identify patients whose observed CPA was excessive when considering traditional risk factors. Subjects whose observed CPA markedly exceeded the expected CPA were easily identified because of their deviation from the regression line that was derived using all members of the study sample. We classified such subjects as having 'unexplained atherosclerosis' relative to the overall study sample when traditional risk factors were included as independent variables. We then examined the plasma homocyst(e)ine concentration in members of the subgroup with 'unexplained atherosclerosis'. We found a significantly higher mean plasma concentration of homocyst(e)ine in the subgroup with 'unexplained atherosclerosis', compared to rest of the study sample (20.4 +/- 4.3 vs. 13.2 +/- 3.2 mumol/l, P < 0.005). We also found that significantly more subjects with 'unexplained atherosclerosis' had plasma homocyst(e)ine concentrations in excess of 14 mumol/l compared to the rest of the study sample (52 vs. 33%, P < 0.002). We thus propose that systematic identification of subjects with 'unexplained atherosclerosis' relative to the rest of a well-characterized study sample might be a useful approach to identify subjects in whom there are newer, non-traditional determinants of predisposition to atherosclerosis.