Pituitary antibodies in women with Hashimoto's thyroiditis: prevalence in diagnostic and prediagnostic sera.

Abstract

Pituitary antibodies have been reported with greater frequency in patients with Hashimoto's thyroiditis than in healthy controls, although there is significant variability in the strength of the association and the methodologies used. We designed a nested case-control study to characterize the prevalence of pituitary antibodies at the time of the clinical diagnosis of Hashimoto's thyroiditis, as well as at 2, 5, and 7 years before diagnosis. Active component female service member cases (n=87) and matched female controls (n=107) were selected using the Defense Medical Surveillance System database (DMSSD) between January 1998 and December 2007. Pituitary antibodies were measured by immunofluorescence using human pituitary glands collected at autopsy as the substrate. At diagnosis, pituitary antibodies were present in 9% of cases with Hashimoto's (8 of 87) and 3% of controls (3 of 107). When the data were analyzed using a conditional logistic regression model, which takes into account the matching on age and work status, pituitary antibodies increased the odds of having Hashimoto's thyroiditis by sevenfold (95% confidence interval from 1.3 to 40.1, p=0.028), after adjusting for components of the DMSSD-category-termed race and for thyroperoxidase antibodies. Before diagnosis, pituitary antibodies were positive in 3 of the 11 subjects (2 cases and 1 control) at the -2-year time point, and negative in all 11 subjects at the -5- and -7-year time points. In summary, using a nested case-control design, we confirm that pituitary antibodies are more common in Hashimoto's thyroiditis and suggest that they appear late during its natural history.