Antithyroid Microsomal Autoantibodies and HLA-DR5 Are Associated With Postpartum Thyroid Dysfunction: Evidence Supporting an Autoimmune Pathogenesis*

Abstract

Antithyroid microsomal antibodies (AMA) and human leukocyte antigen (HLA) haplotypes were assessed as markers for the occurrence of postpartum thyroid dysfunction (all forms) and postpartum painless thyroiditis with transient hyperthyroidism, respectively. AMA titers and thyroid function tests were measured sequentially in 261 mothers from delivery to up to 1 yr postpartum. To test for the association of HLA haplotypes in the subgroup of women with postpartum painless thyroiditis with hyperthyroidism, typing for HLA-A, -B, -C, - DR, and -DQ antigens was carried out in a selected group of 38 mothers with this syndrome. Their results were compared to those in 60 women with hypothyroid goitrous autoimmune thyroiditis and 98 regional controls. In the AMA study, 40 (15%) of the 261 mothers had positive AMA titers (titer, ≤l:80) at delivery, 54 (21%) 2–4 months postpartum, 69 (26%) 5-7 months postpartum, and 60 (23%) 1 yr postpartum. Among 55 mothers who developed postpartum thyroid dysfunction, 25 (47%) had positive AMA at delivery [relative risk (RR), 10.0; P < 0.001; sensitivity, 45%; specificity, 95%]. Two to 4 months postpartum, 46 mothers had thyroid dysfunction, of whom 35 (76%) had positive AMA, while only 19 of the 215 euthyroid mothers (9%) had positive tests for AMA. Positive AMA tests at these times resulted in a RR for postpartum thyroid dysfunction of 32.8 (P < 0.0001; sensitivity, 76%; specificity, 91%). Five to 7 months postpartum, 55 mothers had thyroid dysfunction, of whom 47 (86%) had positive AMA, while only 22 of 206 euthyroid mothers (11%) were positive (RR, 59.0; P < 0.0001; sensitivity, 86%; specificity, 90%). In the HLA studies, DR5 occurred in 11 of 38 (29%) women with postpartum painless thyroiditis with hyperthyroidism vs. 12 of 98 (12%) controls, resulting in a RR of 2.9 (P < 0.05). HLA-DR5 was present in 22 of 60 (37%) patients with goitrous autoimmune thyroiditis (RR, 4.2; P < 0.01). HLA-DR4 and HLA-DQw3 were slightly but not significantly increased in women with postpartum painless thyroiditis with hyperthyroidism and goitrous autoimmune thyroiditis compared to those in normal controls. DQwl was reciprocally decreased (P < 0.025) in the women with goitrous autoimmune thyroiditis, but not in the women with postpartum painless thyroiditis with hyperthyroidism compared to that in normal controls. We conclude that 1) there is a significant association between the occurrence of postpartum thyroid dysfunction and the presence of serum AMA, which reaches maximal sensitivity 5-7 months post-partum; 2) postpartum painless thyroiditis with hyperthyroidism and goitrous autoimmune thyroiditis are significantly associated with HLA-DR5 in our region; however, no significant associations could be established between goiter size, positive AMA, or family history of thyroid disease and HLA haplotypes; and 3) the association of postpartum painless thyroiditis with these markers supports the theory that this syndrome is mediated by autoimmune mechanisms in the majority of affected women and probably represents a variant of goitrous autoimmune thyroiditis which is activated in the postpartum period.