Risk of Progression of Gestational Subclinical Hypothyroidism and Hypothyroxinemia to Overt Hypothyroidism After Pregnancy is Associated with Underlying Thyroid Autoimmunity

Subclinical hypothyroidism (SH) is defined as elevated serum thyrotropin (TSH) with normal free thyroxine (FT4) levels. Isolated hypothyroxinemia (IHT) is defined as normal TSH with low FT4 levels. In iodine sufficient areas, SH affects about 3.5% of pregnancies. ¹ The estimated prevalence of maternal IHT in pregnancy varies widely from 1–31% depending on iodine status and FT4 cutoff used for diagnosis (<2.5th vs. <5th vs. <10th percentiles). ¹ Although maternal SH and IHT in pregnancy have been associated with adverse pregnancy and child neurocognitive outcomes, ² clinical trials have not shown a clear benefit of levothyroxine treatment. ² Thus, the treatment of subclinical thyroid hypofunction in pregnancy is debated. The American Thyroid Association (ATA) guideline recommends treatment of SH in pregnancy, especially in the presence of elevated thyroid peroxidase (TPO) antibody based on robust data from observational studies. ³ On the contrary, the American College of Obstetrics and Gynecologists (ACOG) guideline does not recommend treatment of SH in pregnancy based on clinical trials data. ⁴ The ATA guideline does not recommend treatment for IHT in pregnancy, while ACOG does not address IHT. ^3,4 For patients diagnosed with SH or IHT in pregnancy, there is scant data regarding potential progression to overt hypothyroidism (OH) postpartum and no clear recommendations regarding postpartum follow-up.

In this issue of Thyroid, Varner et al. presents a secondary pooled analysis of placebo-treated groups from two multicenter randomized controlled trials of subclinical thyroid hypofunction in pregnancy. ⁵ The original 2017 parent trials enrolled women with SH or IHT and with singleton pregnancy before 20-week gestation. Participants were randomly assigned to receive levothyroxine or placebo. ⁶ The current study evaluated progression of gestational SH (n = 307) and IHT (n = 229) to OH at 1 and 5 years postpartum in participants who received placebo during the trial. At each postpartum follow-up, history of diagnosis of or treatment for hypothyroidism was ascertained and serum thyroid function was measured. OH was defined as either positive history or serum values of TSH ≥4.0 mIU/L and FT4 < 0.86 ng/dL. OH was more commonly diagnosed in the SH group than the IHT group at both years 1 and 5 (13.4% vs. 3.1%; p < 0.001 and 15.6% vs 2.6%, p < 0.001, respectively). In SH group, participants with positive TPO antibody were more likely to develop OH at both years 1 and 5 compared with those with negative TPO antibodies [26.7% vs. 6.5%, OR 5.3 (95% CI: 2.6–10.7) and 30.5% vs. 7.5%, OR 5.4 (95% CI: 2.8–10.6), respectively]. In IHT group, participants with positive TPO antibodies were more likely to develop OH at year 5 compared with those with negative TPO antibodies [20% vs. 1.8%, OR 13.4 (95% CI: 2.1–84.1)]. In multivariable analysis of the combined groups, the presence of positive TPO antibodies and baseline SH were significantly associated with the development of OH at 1 or 5 years postpartum.

This is the largest study to date with multiple thyroid function measurements that evaluated long-term progression of untreated gestational subclinical thyroid hypofunction. The findings of this study are in line with previous studies. A smaller UK study of 65 patients with SH and 44 patients with IHT diagnosed at 28-week gestation reported TSH elevation >4.5mIU/L or levothyroxine treatment at 5 years postpartum in 24.6% of the SH group and 4.5% of the IHT group. ⁷ However, the majority of patients with postpartum hypothyroidism had normal FT4 levels, suggesting persistent SH rather than development of OH. In another study of 467 Indian women diagnosed with and treated for SH during pregnancy with treatment discontinuation at delivery, 13% developed OH at 2 years postpartum. ⁸ In both studies, a higher degree of TSH elevation and TPO antibody positivity were risk factors of developing future thyroid dysfunction, as seen in the current study.

Certain limitations should be considered in interpreting the results. There was no comparison group of euthyroid participants. A 2022 UK study reported development of SH in 8% and OH in 1% of euthyroid patients with positive TPO antibody and not treated with levothyroxine during the follow-up of 12 months, much lower than findings of the current study. ⁹ However, thyroid function assessment and the majority of progression occurred preconception in the aforementioned study. Another limitation is that the diagnostic criteria for OH at follow-up included history of diagnosis or treatment of hypothyroidism. The authors were unable to elucidate the circumstances of treatment initiation and those with SH may have been misclassified as developing OH. It is also notable that the majority of those with OH were not diagnosed clinically but only at the time of thyroid function measurements for research purpose. This may be partly due to TSH cutoff of ≥4.0mU/L used, leading to the detection of milder, less clinically significant disease. In addition, hypothyroidism was not a fixed diagnosis but one that changed throughout pregnancy for many of the participants, with possible misclassification of those with transient thyroid dysfunction. The authors reports that 53% in SH group had at least one normal TSH and 67% in IHT group had at least one normal FT4 levels in the second half of pregnancy.

At 5 years postpartum, Varner et al. found that about 23% of women with gestational SH and about 5% of women with gestational IHT developed OH. The risk was much higher than has been reported in euthyroid women with thyroid autoimmunity preconception, although further studies are needed to assess postpartum risk in those without thyroid dysfunction in pregnancy. The presence of underlying thyroid autoimmunity was the most important risk factor for progression. A recent meta-analysis by Osinga, et al. similarly reported TPO antibody positivity to be the single most significant risk factor of development of thyroid hypofunction in pregnancy. ¹⁰ These findings could be used to argue in favor of routine measurement of TPO antibody titer in patients with thyroid hypofunction detected in pregnancy, in order to risk-stratify for worsening thyroid function both during gestation as well as postpartum. These findings also may suggest value in routine long-term monitoring of thyroid function in patients with gestational thyroid hypofunction. Further research is needed to determine potential benefit of treatment of subclinical thyroid hypofunction in pregnancy on progression to OH postpartum.