Data from animal studies and from studies of patients with acute necrotizing ulcerative gingivitis (ANUG) have provided suggestive evidence for an association between ascorbate deficiency and disease risk. Further, there is biological plausibility for such an association, due to the role of ascorbate in collagen synthesis and leukocyte function. A case-control study of plasma ascorbate and ANUG was performed on 60 patients with a history of ANUG infection and 60 age-race-sex-matched controls. No cases had had active lesions for at least two months prior to their vitamin assay to avoid any potential reduction of dietary intake of ascorbic acid due to the presence of painful mouth lesions. According to results obtained by use of a modification of the 2,4-dinitrophenylhydrazine method for determination of total plasma ascorbate, the mean and standard error of the mean of plasma ascorbate for all ANUG cases was 0.07 ± 0.006 mmol/L; the mean for all controls was 0.10 ± 0.006 mmol/L. Paired differences in plasma ascorbic acid concentrations between cases and controls were significantly different from zero (p < 0. 001). The unadjusted relative risk (RR) of ANUG as obtained by conditional logistic regression for subjects whose plasma ascorbic acid concentration was at or below the median value for controls, relative to subjects with higher values, was 7.3 (90% confidence interval, 3.0 - 17.4; one-sided p value < 0.001). Patients with a history of ANUG ingested a daily average of 1.2 ± 0.2 servings of dietary ascorbic acid, as compared with a daily average of 1.9 ± 0.2 servings for healthy controls. The significant RR of ANUG for decreased plasma ascorbate persisted after data were adjusted for dietary intake (RR = 6.6 ; 2.7 - 16.1; p = 0.016). Similarly, smoking status only partly explained ascorbate differences, with a RR of ANUG of 4.4 (0.9 - 20.2 ; p = 0.057) adjusted for smoking. Adjusting for social class differences between cases and controls, however, explained the observed differences (RR = 1.00). We conclude that the significant association between decreased plasma ascorbate and ANUG even after healing of initial lesions was as hypothesized, although a causal interpretation of the findings was obscured by social class differences between cases and controls. © 1988, SAGE Publications. All rights reserved.