MATERIALS AND METHODS
This study was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and MetaAnalysis (PRISMA) statement. The protocol for this review was registered with PROSPERO under registration number CRD42018110340. As a systematic review and metaanalysis are classified as nonhuman subjects research, no institutional review board approval was required. The PubMed, EMBASE, Web of Science, and CENTRAL databases were searched from inception to September 10, 2018, or inception to 2019 if exact date ranges were not available, to identify studies involving thyroid dysfunction in women with a history of RPL. The search query included the following terms: ‘‘thyroid antibodies’’ OR ‘‘thyroid antibodies in pregnancy’’ OR ‘‘thyroid antibody-positive’’ OR ‘‘thyroid antibody levels’’ OR ‘‘thyroid peroxidase antibodies’’ OR ‘‘thyroglobulin antibodies’’ OR ‘‘thyroid peroxidase antibody’’ OR ‘‘thyroglobulin antibody’’ OR ‘‘thyroid autoantibodies’’ AND ‘‘pregnancy’’ OR ‘‘hypothyroidism’’ OR ‘‘subclinical hypothyroidism’’ OR ‘‘thyroid autoimmunity’’ OR ‘‘Thyroid autoantibodies’’ OR ‘‘Thyroid peroxidase antibodies’’ OR ‘‘thyroglobulin antibodies’’ AND ‘‘recurrent pregnancy loss’’ OR ‘‘recurrent miscarriage’’ OR ‘‘habitual abortion’’ OR ‘‘recurrent spontaneous abortion’’ OR ‘‘recurrent abortion.’’ All studies were imported into RefWorks 2.0 (ProQuest, Ann Arbor, MI), and duplicate studies were removed electronically and manually.
Overt Hypothyroidism and RPL
Question: Is overt hypothyroidism associated with RPL?.
No studies examining the potential associations between overt hypothyroidism and RPL were identified in this systematic review.
Question: Does treatment improve pregnancy outcomes in women with RPL who have overt hypothyroidism?
No studies examining whether treating overt hypothyroidism in women with RPL improves pregnancy outcomes were identified in this systematic review.
Subclinical Hypothyroidism and RPL
Question: Is subclinical hypothyroidism associated with RPL?
Five eligible studies addressed whether subclinical hypothyroidism is associated with RPL. Descriptions of each study, including their definitions of RPL and subclinical hypothyroidism, and their results are shown in. A meta-analysis of the prevalence of subclinical hypothyroidism in women with RPL was conducted from these five studies, and the pooled prevalence was calculated to be 12.6% (95% confidence interval [CI], 0%–35.2%; I 2 99%).
Question: Does treatment with levothyroxine improve subsequent pregnancy outcomes in women with a history of RPL who have subclinical hypothyroidism?.
One study reported on subsequent pregnancy outcomes after treatment with levothyroxine in women with RPL who had subclinical hypothyroidism, as shown. The definition of RPL as well as the dose and timing of levothyroxine administration and the results are included in the table.
Thyroid Autoimmunity and RPL
Question: Is thyroid autoimmunity associated with RPL?
Twenty-six eligible studies addressed whether thyroid autoimmunity is associated with RPL (21, 26–50). includes descriptions of each study, including their definitions of RPL as well as thyroid autoimmunity, thyroid status, and results. Fourteen of the 26 studies showed a statistically significant association between thyroid autoimmunity and RPL, whereas 12 studies showed no association.
Question: Does thyroid autoimmunity predict future pregnancy outcomes in women with RPL?
Six studies have been published on the predictive value of thyroid autoimmunity on the prognosis of future pregnancies of women with RPL, as shown in Supplemental. Three studies reported no association between thyroid autoimmunity and future risk of pregnancy loss in women with RPL. However, three other studies reported that thyroid autoimmunity, mainly TPOAb positivity, is associated with worse outcomes in future pregnancies in women with RPL.
Question: What are the potential mechanisms by which thyroid autoimmunity might impact RPL?
Two studies have been published examining the potential mechanisms of action for how thyroid antibodies may cause miscarriage in women with RPL. Lazzarin et al. published a cross-sectional study reporting that 30 of 46 (65%) euthyroid, Italian women with RPL—defined as two or more consecutive first-trimester pregnancy losses—who tested positive for either TPO or thyroglobulin antibodies had abnormal responses to thyroid-releasing hormone (TRH), possibly indicating some level of subtle thyroid dysfunction despite otherwise normal thyroid function tests. Mariee et al. published a prospective cohort study that assessed uterine natural killer (NK) cells in 42 English women with a history of RPL, defined as three or more consecutive first-trimester pregnancy losses. Thyroid status was not presented. The mean proportion of uterine NK cells was 7.1% in women who tested positive for TPO antibodies, compared with 10.6% in those who tested negative.
Concurrent Subclinical Hypothyroidism and Thyroid Autoimmunity and RPL
Question: Does treatment with levothyroxine improve subsequent pregnancy outcomes in women with a history of RPL who have both subclinical hypothyroidism and thyroid autoimmunity as opposed to the treatment of thyroid autoimmunity alone?
One study has examined whether treatment of combined subclinical hypothyroidism and thyroid autoimmunity in women with RPL results in improved pregnancy outcomes compared with treating thyroid autoimmunity alone, as shown. The American Thyroid Association recommends treatment of combined subclinical hypothyroidism and thyroid autoimmunity in the general pregnant population based on an increased risk of adverse pregnancy outcomes. However, based on a single study in the RPL population, there is no added benefit to treating combined subclinical hypothyroidism and thyroid autoimmunity as compared to treating thyroid autoimmunity alone. Thus, given that there appears to be no benefit in treating thyroid autoimmunity in women with RPL, treatment of combined subclinical hypothyroidism and thyroid autoimmunity cannot be recommended based on the currently available evidence. However, further research is needed on this topic before definitive conclusions can be drawn.
Author: Allan C. Dong, M.D., Jessica Morgan, M.D., Monica Kane, M.P.H., Alex Stagnaro-Green, M.D., M.H.P.E., M.H.A., and Mary D. Stephenson, M.D., M.Sc.