Buccal mucosa (BM) cells have been used in human biomonitoring studies for detecting DNA adducts and chromosomal damage in an epithelial cell population. In the present study, we have investigated if human BM cells are suitable for use in the single-cell gel electrophoresis (SCGE)/Comet assay as an approach for estimating the exposure of epithelial cells to DNA-damaging agents. Our results indicate that only a few cells from BM cell samples yield comets that can be analyzed by current methods, and that the yield of cells with comets is independent of the percentage of viable BM cells in the sample. Data generated after enzymatic enrichment of viable cells and immunomagnetic separation of epithelial cells suggest that most of the BM cells that do form comets are probably leukocytes. Moreover, by reevaluating specific cells after running the Comet assay, we found that viable epithelial BM cells give rise to atypical comets that are not included in the analysis. Comparing DNA migration patterns between small groups of smokers and nonsmokers indicated that long-term smoking had no effect on the subpopulation of cells that yield typical comets. Our results indicate that the SCGE assay, as it is commonly performed, may not be useful for genotoxicity monitoring in human epithelial BM cells.