The prevalence of antithyroid peroxidase antibodies (ATPO) and/or of thyroid dysfunction was studied in 422 patients with chronic viral hepatitis C, B, and D. Baseline results were compared with those during and 6 months after interferon alfa (IFN-alpha) therapy. The overall prevalence of ATPO among untreated patients was 14.1%, with no significant differences between chronic hepatitis C, B, or D, as well as between males and females. However, high ATPO titers (> or = 18 IU/mL) clustered significantly among females (8.7% vs. 3.4%; P = .022), especially those with chronic hepatitis C (11.2% vs. 3.6%; P = .036). Before treatment, 3.7% of the patients had thyroid dysfunction, mostly hypothyroidism (3.5%), the latter increasing to 14.3% among patients with ATPO titers > or = 18 IU/mL. IFN-alpha treatment significantly increased overall thyroid dysfunction (9.7%; P = .001) and hypothyroidism (7.8%; P = .01), particularly among patients with high baseline ATPO (38.5%; P = .0002). Six months after stopping IFN-alpha treatment, the prevalence of thyroid dysfunction was 8.0%, still significantly higher than at baseline. By multivariate analysis, the only predictor positively associated with pre- or on-treatment hypothyroidism was the baseline titer of the ATPO antibodies (relative risk [RR], 3.0 and 3.8 per each log titer increase, respectively). In conclusion, patients with chronic viral hepatitis on IFN-alpha treatment exhibit an almost threefold increase of baseline thyroid dysfunction, persisting long after the end of therapy. High ATPO titers, clustering among females, particularly those with hepatitis C, represent the only predictor of pre- and on-treatment hypothyroidism by multivariate analysis. Patients with chronic viral hepatitis, especially females, should be tested for ATPO and thyroid function and monitored during and posttreatment for free thyroxin (FT4) and thyroid-stimulating hormone (TSH) levels.