Long-Term Outcome of Interferon-α-Induced Thyroid Autoimmunity and Prognostic Influence of Thyroid Autoantibody Pattern at the End of Treatment

Thyroid autoimmunity and dysfunction have been widely reported as side effects of interferon-α (IFN-α) treatment, but the literature lacks data regarding the long-term course of these complications, clinical observation being limited to 6–12 months off therapy. Our study is the first that has aimed to evaluate the natural history of IFN-related thyroid autoimmunity during a 6.2-yr follow-up after the IFN-α withdrawal as well as to investigate the potential role of the autoantibody pattern at the end of treatment to predict the long-term outcome. Our study group included 114 patients (79 males, 35 females), mean age 48 yr (range 23–67 yr) with no preexisting thyroid disease, undergoing a 12-month treatment with recombinant IFN-α for C virus-related chronic active hepatitis. Thyroid autoimmunity (serum TgAb and TPOAb) and function (serum FT4, FT3, TSH) were retrospectively evaluated at the end of IFN therapy, 6 months after IFN withdrawal and after a median period of 6.2 yr (range 5.5–8.4 yr). At the end of treatment, 78 patients were negative for thyroid autoantibodies (Abs−) and all but one of them remained so for the following evaluations. The remaining 36 patients had thyroid autoantibodies (Abs+) at the end of treatment, and they subsequently showed a heterogeneous behavior: 16 patients remained Abs+ for the whole length of the study (persistent thyroiditis); 10 patients became Abs− 6 months off therapy but were again Abs+ 6.2 yr later (remitting/relapsing thyroiditis); 10 patients reverted to autoantibody negativity at different observation times (transient thyroiditis). The absence of thyroid autoantibodies at the end of treatment was a protective factor for the successive development of thyroiditis (odds ratio: 0.02, confidence interval (CI) 95%: 0–0.1). On the contrary, the positivity for TgAb and/or TPOAb at high titers at the end of IFN treatment was significantly related to the highest risk of having chronic thyroiditis (odds ratio: 17.3, CI 95%: 3.2–91.7 for TgAb levels > 50 degree percentile; odds ratio: 7.3, CI 95%: 1.5–35.2 for TPOAb levels > 50 degree percentile). None of the patients showed overt thyroid dysfunction throughout the study, whereas a subclinical hypothyroidism was found in 12 patients. In all 12 cases, the functional abnormality was accompanied by the presence of thyroid autoantibodies. Eight of these 12 patients belonged to the group with persistent thyroiditis (P < 0.05). The absence of thyroid autoantibodies at the end of t