Experimental evidence from animal models has provided a framework for our current understanding of autoimmune disease pathogenesis and supports the importance of genetic predisposition, molecular mimicry, and immune dysregulation. However, only recently has evidence emerged to support the role of immune dysregulation in human organ-specific autoimmune disease. In the current study of the "late" manifestation of autoimmune thyroid disease (AITD) in a cohort of human immunodeficiency virus (HIV)-positive patients following highly active antiretroviral therapy (HAART), we discuss how immune dysregulation and factors associated with the immunopathology of HIV infection fit the current understanding of autoimmunity and provide a plausible basis for our clinical observations. De novo diagnoses of thyroid disease were identified between 1996 and 2002 in 7 HIV treatment centers (5/7 centers completed the study). Patients were diagnosed as clinical case entities and not discovered through thyroid function test screening. Paired plasma specimens were used to demonstrate sequential rise in thyroid antibodies. Seventeen patients were diagnosed with AITD (median age, 38 yr; 65% were of black African or black Caribbean ethnicity; and 82% were female). The median duration of immune reconstitution was 17 months. Graves disease (GD) was diagnosed in 15 of 17 patients. One patient developed hashithyrotoxicosis with atypically raised C-reactive protein, and another developed hypothyroidism. One GD patient had associated secondary hypoadrenalism. The estimated combined prevalence of GD for 4 treatment centers for female patients was 7/234 and for males was 2/1289. The denominator numbers were matched controls, from 4 centers able to provide data, who commenced HAART during the same time (January 1996 to July 2002) and who did not develop clinical AITD. The mean baseline pre-HAART CD4 count was 67 cells/mL, and the mean increase from nadir to AITD presentation was 355 cells/mL. AITD patients were more likely than controls (95% confidence interval, chi-square test) to be severely compromised at baseline (as defined by a CD4 count < 200 cells/mL or the presence of an acquired immunodeficiency syndrome [AIDS]-defining diagnosis), and to experience greater CD4 increments following HAART. AITD may be a late manifestation of immune reconstitution in HIV-positive patients taking HAART, and immune dysregulation may be an important factor.