Practice Variation in the Care of Subclinical Hypothyroidism During Pregnancy

A National Survey of Physicians in the United States

Freddy J.K. Toloza; Naykky M. Singh Ospina; Rene Rodriguez-Gutierrez; Derek T. O'Keeffe; Juan P. Brito; Victor M. Montori; Spyridoula Maraka


J Endo Soc. 2019;3(10):1892-1906. 

In This Article


This study assessed clinicians' knowledge, perceptions, and clinical practices regarding SCH diagnosis, treatment, and effect on pregnancy in relation to the most recently published ATA guidelines.[1] Across a demographically diverse sample of clinicians and members of the largest endocrinology society in the United States, we found an awareness of the ATA guidelines and evidence of their effects on practice, with low adherence to the recommended TSH cutoff for the diagnosis of SCH during pregnancy and the indications for TPO-Ab status assessment as part of the diagnostic evaluation. When LT4 treatment was chosen by respondents, there was a small or very small expected reduction in maternofetal complications. We also found that only 50% of the clinicians who responded take patient preferences into consideration when determining treatment, and 19% take therapy adverse effects into consideration.

Interventional studies[13,27,28] have been unable to document the same benefits of LT4 treatment of SCH in pregnancy as seen in observational studies.[29–32] In this uncertain context, the ATA issued an updated version of its guidelines encompassing major changes in clinical practices for the management of SCH during pregnancy.[1,33] The new recommendations have been partially accepted;[33–35] as demonstrated here, practice is only partially concordant with ATA recommendations. This could be due to low-grade recommendations according to the guideline-grading hierarchy, which could be perceived by health care providers as lacking certainty in the evidence. However, we did not find a correlation between guideline adherence and the guideline-reported strength of recommendations or the quality of the evidence supporting the recommendations, except for those recommendations regarding treatment. Moreover, the extent to which adherence to these guidelines improves maternofetal outcomes remains uncertain.

One of the most controversial issues in the field of thyroid dysfunction and pregnancy is the appropriate screening approach in early pregnancy. The ATA guideline recommends neither for nor against universal screening for abnormal TSH concentrations in early pregnancy on the basis of uncertain benefits.[36] Contrary to this recommendation, according to our survey universal screening is the preferred method for pregnant women, a result that is consonant with findings from surveys of other medical societies.[18,19,21,37] The preference for universal screening may be driven by the inability of clinicians to identify at-risk women[38] and the well-known benefit of LT4 treatment for overt thyroid dysfunction during pregnancy. Furthermore, universal screening for thyroid dysfunction in pregnancy has been supported by some authors,[35,39] who argue that universal screening[40] for overt thyroid disease is justifiable and lack of clarity on the effect of optimum management of SCH is not an adequate rationale for inaction after detection.

Of note, most respondents are still using a TSH level >2.5 mIU/L for the diagnosis of SCH in the first trimester of pregnancy, as recommended in the older guidelines[10,41] and consistent with previous studies.[18] Although a TSH cutoff of >4.0 mIU/L was offered for cases in which a population-based cutoff is unavailable, only 5.4% of respondents selected this option. This may be because using a TSH cutoff of >4.0 mIU/L for diagnosis has been criticized, given that only 26% of the studies cited under this recommendation found an upper limit of normal for TSH ≥4.0 mU/L.[1,35,42–45] However, the benefit of detecting and treating women with a TSH level of 2.5 to 4.0 mIU/L remains uncertain while exposing patients to anxiety and treatment burden, factors that physicians rarely consider in treatment decisions as reported in the current study. In addition, although a TSH level >2.5 mIU/L was the most used cutoff for SCH diagnosis among respondents, only 18% of the respondents would treat a pregnant woman with a TSH level of 3.2 mIU/L and without thyroid autoimmunity. It is possible that for the respondents a TSH level >2.5 mIU/L is used mainly to define SCH and create awareness, leading to closer follow-up during pregnancy, but does not lead on its own to LT4 treatment initiation.

Although the decision to treat or not varied depending on the specific clinical scenario, when LT4 treatment was chosen, respondents expected a small or very small reduction in maternofetal complications across all clinical scenarios regardless of patients' characteristics. This trend may be explained by the absence or small size of effects on LT4-treated patients shown in interventional studies.[13,27,28,46,47]

Delivery of maternal T4 to the fetus through placental transference is crucial for optimal fetal brain development.[48–50] The use of LT4 + T3 or desiccated thyroid extracts produces a low T4/T3 ratio; as a result, the placental transfer of LT4 to the fetal brain may be insufficient.[49,51,52] Therefore, LT4 is the preferred drug during pregnancy. Despite the strong recommendation to not use other thyroid preparations, a small portion of the respondents still choose LT4 + T3 or desiccated thyroid extracts as pharmacological therapy during pregnancy.

Our findings are limited in their applicability by the relatively low response rate of society members. However, previous studies have shown similar response rates by clinically active members of the Endocrine Society.[24–26,53] This low response rate may have an effect on the generalizability of the present results, driven mainly by the potential selection bias of the respondents. Moreover, underrepresentation of non–endocrine clinicians (Obstetrics & Gynecology, internal medicine, and family medicine specialists) may have affected the results. In addition, we did not include history of miscarriages as a variable in the case scenarios, and this could have changed/influenced the respondents' answers. A history of miscarriage, as demonstrated by our results, is a strong factor in the decision to initiate LT4 therapy during pregnancy. Finally, it is important to note that it takes ~2 to 3 years to fully implement a new guideline in clinical practice. We performed this survey 18 months after the ATA guidelines were released, which may have contributed to the low adherence and may be a source of bias in the current study. Further studies in the field might intend to evaluate the practices of the other participating specialties in the care of pregnant women with SCH (obstetrics & gynecology, internal medicine, family medicine).

In summary, this national study assessed clinicians' knowledge and reported practices regarding the diagnosis, treatment, and effect of SCH on pregnancy and their concordance with the latest ATA guidelines. Despite recently updated guidelines, there is still wide variation in clinical practice regarding the care of pregnant women with SCH. Improvement may require multicentric collaboration to produce highly reliable, practical randomized trials of the comparative effectiveness and the impact on maternofetal and offspring outcomes of the most uncertain and commonly used treatment practices in the care of SCH during pregnancy.