Effects of Levothyroxine Treatment on Fertility and Pregnancy Outcomes in Subclinical Hypothyroidism: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Abstract

Background:

Subclinical hypothyroidism, defined by elevated thyrotropin (TSH) and normal free thyroxine levels, is associated with adverse pregnancy outcomes, including preterm birth, pre-eclampsia, and small for gestational age. Despite the uncertainty regarding the effectiveness of levothyroxine (LT4) treatment on pregnancy outcomes in subclinical hypothyroidism, LT4 is widely administered with a pre-treatment threshold TSH level of 2.5 mU/L. The aim of this study is to investigate the efficacy of periconceptional LT4 treatment for subclinical hypothyroidism, including TSH levels >2.5 mU/L, and identify the characteristics of subclinical hypothyroidism that can benefit from LT4 treatment.

Methods:

We conducted a systematic review and meta-analysis of randomized controlled trials from inception to February 2023. We analyzed the pooled effects of LT4 on subclinical hypothyroidism before and during pregnancy. The main outcomes before pregnancy were live birth, pregnancy, and miscarriage. The main outcomes during pregnancy were live birth, miscarriage, and preterm birth. We conducted subgroup analyses to compare the effects of LT4 on subclinical hypothyroidism with TSH levels of 2.5–4.0 and >4.0 mU/L.

Results:

Of the 888 studies identified, 27 full-text articles were screened for eligibility. Five studies on pre-conception treatment with 768 participants and eight studies on treatment during early pregnancy with 2622 participants were analyzed. One of the two studies on pre-conception treatment in subclinical hypothyroidism with TSH >4.0 mU/L had high risk of bias and the other was composed of 64 participants. Pre-conception LT4 treatment had no significant effect in improving rates of live births and pregnancies, or reducing miscarriages (risk ratio [RR], 95% confidence interval): 1.41 (0.84–2.36), 1.73 (0.88–3.39), and 0.46 (0.11–2.00), respectively. LT4 treatment during pregnancy was not significantly associated with higher rates of live births (RR 1.03, 0.98–1.09) nor decreased miscarriage rates (RR 1.01, 0.66–1.53). The effect of LT4 treatment on preterm birth during pregnancy was significantly different depending on the TSH values (p = 0.04); a positive effect was shown in the subclinical hypothyroidism subgroup with TSH >4.0 mU/L (RR 0.47, 0.20–1.10), while no significant effect was observed in the subgroup with TSH 2.5–4.0 mU/L (RR 1.35, 0.79–2.31).

Conclusions:

Pre-conceptional LT4 treatment for subclinical hypothyroidism does not improve fertility or decrease the incidence of miscarriages. However, further well-designed studies are needed for pre-conceptional treatment, especially in TSH >4.0 mU/L. LT4 treatment during pregnancy had a positive effect on preterm birth; nevertheless, this was only applicable to subclinical hypothyroidism with TSH >4.0 mU/L.

Introduction

Subclinical hypothyroidism, defined by elevated thyrotropin (TSH) concentrations with normal free thyroxine (fT4) concentrations, is prevalent among women of reproductive age. Subclinical hypothyroidism is associated with multiple adverse pregnancy outcomes, including pregnancy loss, preterm birth, and preeclampsia.^1
–3 Subclinical hypothyroidism negatively affects the outcomes of artificial reproductive technologies (ART) proportional to the increase in TSH levels; hence, the American Thyroid Association guidelines recommends levothyroxine (LT4) supplementation to achieve TSH levels <2.5 mU/L.⁴

In real world clinical settings, considering the potential risk of TSH level elevation, LT4 is also prescribed to euthyroid women with TSH levels of 2.5–4.0 mU/L before or during pregnancy;^5,6 however, the beneficial effects in periconceptional subclinical hypothyroidism remain controversial. For thyroid peroxidase antibody (TPOAb)-positive euthyroid women, the limited effectiveness of LT4 has been demonstrated in multiple randomized controlled trials (RCTs).^7
–9 Hence, a systematic review of the literature to clarify the patient groups potentially benefiting from LT4 treatment in pregnancy is needed.

Several systematic reviews and meta-analyses have investigated the effects of LT4 on pregnancy outcomes.^10

–13 However, it is difficult to interpret the efficacy of LT4 in women with subclinical hypothyroidism because of the following limitations: most analyses are composed of studies with heterogeneous levels of evidence, including RCTs and observational studies; the participants in the studies are not limited to those with subclinical hypothyroidism but include those with euthyroidism (TSH levels <2.5 mU/L) or overt hypothyroidism; and the LT4 intervention was not necessarily initiated from early pregnancy.

Therefore, to overcome these limitations, we analyzed only RCTs and strictly extracted the outcomes of subclinical hypothyroidism in the studies that initiated LT4 treatment before or in early pregnancy. We assessed the effect of LT4 before and during pregnancy on subclinical hypothyroidism, including cases with TSH levels >2.5 mU/L, which is the population in which there are questions about the potential treatment benefit.

Materials and Methods

This systematic review was performed based on the principles outlined in the Meta-Analysis of Observational Studies in Epidemiology guidelines¹⁴ and the Cochrane Handbook for Systematic Reviews for Intervention. The study protocol was registered in advance in PROSPERO (CRD42023399795).

The PICO question was as follows:

Population (P): women with subclinical hypothyroidism who were planning to become or being pregnant.

Intervention (I): LT4 treatment before pregnancy and/or started in early pregnancy (at least within 20 weeks of pregnancy).

Comparator (C): placebo or no treatment.

Outcome (O): the main outcomes of preconception LT4 treatment were live births, pregnancies, and miscarriages. The main outcomes of LT4 treatment during early pregnancy were live births, miscarriages, and preterm births. Secondary outcomes included hypertensive disorders of pregnancy, preeclampsia, gestational diabetes mellitus, caesarean section, placental abruption, neonatal admission, birth weight, small-for-gestational age, and neurodevelopmental outcomes.

Study design (inclusion criterion): RCTs were included. Case reports, case series, and observational studies were excluded.

Definition of subclinical hypothyroidism

Subclinical hypothyroidism was defined as a normal fT4 level with an elevated TSH level. In the present study, the lower limit of the reference range of TSH was defined as 2.5 mU/L both before and during pregnancy. Considering the physiological fluctuations of TSH in early pregnancy,¹⁵ a normal fT4 and a TSH level of 2.5–4.0 mU/L are usually categorized as subclinical hypothyroidism during pregnancy, but as euthyroid before pregnancy. However, we defined pre-conceptional TSH levels >2.5 mU/L as subclinical hypothyroidism to investigate the efficacy of LT4 in line with actual clinical practice.^4,16

Search strategy

We searched the PubMed, Cochrane Central Register of Controlled Trials, Embase, ClinicalTrials.gov, and WHO International Clinical Trials Registry Platform from inception to February 21, 2023. We updated the search on January 19, 2024. The search words included terms for “subclinical hypothyroidism,” “pregnancy,” or “levothyroxine.” The details of the search strategy for each database are presented in Supplementary Table S1.

Study selection

Two reviewers (A.S. and H.S.) independently screened the titles and abstracts generated from the search results. Disagreements were resolved through a discussion between the two reviewers or a consultation with senior reviewers, if necessary. We contacted the authors for additional information, the extracted outcomes limited to subclinical hypothyroidism.

Data extraction

One reviewer (A.S.) extracted the data, and the other two reviewers (S.K. and M.I.) checked the extracted data independently. The extracted data included the following: first author's name, year of publication, country of study, study population, number of participants (the randomized number and the number of subclinical hypothyroidism patients included), the cut-off TSH level for subclinical hypothyroidism, baseline characteristics (TPOAb positivity, ART, and TSH levels at randomization), details of intervention methods (timing of LT4 treatment initiation, LT4 initial dose and subsequent dose adjustment, and TSH levels after LT4 treatment), and outcomes. Live births, one of the main outcomes, were estimated using the number of participants, miscarriages, and stillbirths in cases where the exact number of live births was not shown in the original study.

Risk of bias

Two reviewers (A.S. and H.S.) independently assessed the risk of bias across the seven domains using the Cochrane Collaboration risk-of-bias tool.¹⁷ The risk of bias in each domain was classified as low, high, or unclear. Disagreements were resolved through a discussion between the two reviewers or a consultation with senior reviewers (E.O. and N.A.), if necessary.

Statistical analysis

All meta-analyses were performed using the RevMan 5.4 (Cochrane Collaboration, Oxford, UK). The pooled effect size was described as risk ratio (RR) with 95% confidence intervals (CIs) for dichotomous outcomes and mean difference and 95% CIs for continuous outcomes. Heterogeneity was assessed using the chi squared (I² ) test.

We revised the protocol to use the random effect model regardless of the heterogeneity. The level of statistical significance was set as p < 0.05. Outcomes with p-values <0.1 were considered positive or negative effects. We performed subgroup analyses by dividing the included studies into those where subclinical hypothyroidism was defined by TSH levels of 2.5–4.0 and >4.0 mU/L, respectively. If necessary, studies that did not have a TSH cut-off value of 4.0 mU/L were allocated depending on the upper limit of the reference range. Sensitivity analyses were performed to exclude studies with a high risk of bias.

Results

Study selection and characteristics

A flowchart of the search and selection process for eligible studies is presented in Supplementary Figure S1. The search identified 888 records after excluding 266 duplicates. After screening titles and abstracts, 27 records were identified as potentially relevant. We assessed the full-text articles for these 27 studies and contacted the authors for additional information if necessary. In the studies with women who were euthyroid, only the studies presenting the outcomes for TSH levels >2.5 mU/L separated from the outcomes for TSH levels <2.5 mU/L were included.

In studies with several types of thyroid abnormalities, we contacted the authors to provide outcomes restricted to subclinical hypothyroidism, excluding overt hypothyroidism, isolated hypothyroxinemia, and euthyroidism. We excluded studies that did not provide specific results for subclinical hypothyroidism. The list of the excluded studies through the full-text review process is presented in Supplementary Table S2. Finally, 768 participants in five reports with pre-conception LT4 treatment^7
–9,18,19 and 2622 participants in eight reports with LT4 treatment during early pregnancy^{20

–27} were included. The main characteristics of the included reports are shown in Table 1 (preconception) and Table 2 (during pregnancy).

Table 1.

Characteristics of the Included Reports: Levothyroxine Before Pregnancy

Study	Country	Participants SCH or EU	Number		ART (%)	TPOAb positivity (%)	RPL (%)	Initial LT4 (mcg/day)	Target TSH values (mU/L)	TSH in LT4 arm (mU/L) ^b
Study	Country	Participants SCH or EU	Randomized	Included ^a (LT4:CTRL)	ART (%)	TPOAb positivity (%)	RPL (%)	Initial LT4 (mcg/day)	Target TSH values (mU/L)	Basal	At pregnancy
Dhillon-Smith et al.⁸	United Kingdom	EU	952	145:143	12.5	100	19.5	50	No adjustment	2.10 (1.51–2.74)	No data
Kim et al.¹⁹	Korea	SCH, TSH ≥4.5 mU/L	64	32:32	100	79.7	No data	50	<2.5 in the first trimester	6.6 (1.7)	2.9 (1.0)
Abdel Rahman et al.¹⁸	Egypt	SCH, TSH ≥4.0 mU/L	70	35:35	100	No data	No data	50	Within reference range	4.7 (0.5)	1.1 (0.3)
van Dijk et al.⁹	Netherland, Belgium, Denmark	EU^c	187	39:30	12.3	100	100	0.5–1.0/kg	No adjustment	2.10 (1.40–3.11)	No data
Wang et al.⁷	China	EU	600	180:97^c	100	100	0	25 or 50	0.1–2.5	2.94 (2.04–3.74)	No data

The analyzed number in the meta-analysis, excluding participants with TSH <2.5 mU/L and normal fT4.

Shown as median (IQR) or mean (SD).

Provided information from authors.

fT4 were not evaluated.

ART, assisted reproductive technology; CTRL, control; EU, euthyroidism; IQR, interquartile range; LT4, levothyroxine; RPL, recurrent pregnancy loss; SCH, subclinical hypothyroidism; SD, standard deviation; TPOAb, anti-thyroid peroxidase antibody; TSH, thyrotropin.

Table 2.

Characteristics of the Included Reports: Levothyroxine During Pregnancy

Study	Country	TSH cutoff value for SCH (mU/L)	Number		TPOAb positivity (%)	Initial LT4 (mcg/day)	Target TSH value (mU/L)	Basal TSH in LT4 arm (mU/L) ^b	Gestational age of LT4 initiation (weeks) ^c
Study	Country	TSH cutoff value for SCH (mU/L)	Randomized	Included ^a (LT4:CTRL)	TPOAb positivity (%)	Initial LT4 (mcg/day)	Target TSH value (mU/L)	Basal TSH in LT4 arm (mU/L) ^b	Gestational age of LT4 initiation (weeks) ^c
Casey et al.²⁰	United States	3 or 4	677	339:338	Not shown	100	0.1–2.5	4.5 (4.4–4.7)	The randomization was at 16.6 (3.0)
Alcazar Lazaro et al.²¹	Spain	2.5–4.9 mU/L	363	181:182	16.5	0.5 per kg	<3.0	2.53 (0.99)	First visit was at 12.55 (2.0), <13
Lazarus et al.²²	United Kingdom, Italy	Cohort specific 97.5th percentile value	16,346	209:219	Not measured	150	0.1–1.0	Not shown	Median 13.4
Leng et al.²³	China	2.5	1736	243:251 RPL; 131:136 Non RPL; 112:115	0	50	No adjustment	3.74 (1.28) 3.32 (0.78)	4–8 days after first antenatal visit; <12
Nazarpour et al.²⁴	Iran	2.5	131	56:58^d	100	Determined by BW and TSH value	No adjustment	3.8 (2.8–4.8)	4–8 days after first antenatal visit, median 10–12, <20
Nazarpour et al.²⁵	Iran	2.5	366	183:183	0	Determined by BW and TSH value	No adjustment	4.5 (4.4–4.7)	4–8 days after first antenatal visit, 11.4 (4.1), <20
Nazarpour et al.²⁶	Iran	4.0	527	62:56	Not shown	Determined by BW and TSH value	No adjustment	Not shown	4–8 days after first visit
Zhao et al.²⁷	China	2.5	62	31:31	32.3	25	<3.0	Not shown	2nd day after diagnosis, 8–10

The analyzed number in the meta-analysis, excluding participants with isolated hypothyroxinemia, overt hypothyroidism, and TSH <2.5 mU/L with normal fT4.

Shown as median (IQR) or mean (SD). median (95% CI) in Casey et al.²⁰

Shown as median (IQR) or mean (SD), range.

Provided information from authors.

BW, body weight; CI, confidence interval; SCH, subclinical hypothyroidism; TPOAb, anti-thyroid peroxidase antibody.

Quality of the included studies

The risk-of-bias assessment is shown in Figure 1. Nine reports had a high risk of blinding participants and personnel owing to the lack of a placebo.^{7,19,21

–27} One report was considered to have a high risk of other biases as it had 100% pregnancy outcomes in the intervention arm without any biochemical miscarriages.¹⁸ In six reports, we performed a meta-analysis extracting the participants who were eligible for our study: a TSH level over 2.5 mU/L with normal fT4 levels.^{7
–9,22–24}

FIG. 1.

Risk-of-bias assessment and trial quality assessment. (a) Pre-conception LT4 treatment. (b) LT4 treatment during pregnancy. LT4, levothyroxine.

We excluded euthyroidism with a TSH level <2.5 mU/L, isolated hypothyroxinemia, and overt hypothyroidism.^{7
–9,22–24} All these six reports originally had low risk of random sequence generation before the extraction of eligible participants for the present study. However, randomization may not be maintained through the extraction process in our analysis.

Effect of pre-conception LT4 treatment

Data on live births were available in all of the five included studies (Fig. 2a).^7
–9,18,19 LT4 treatment did not improve the live birth rate compared with the controls (RR, 95% CI): 1.41, 0.84–2.36; p = 0.20). A high level of statistical heterogeneity was found on pooling the data from these five studies (p < 0.001, I² = 82%).

FIG. 2.

Forest plots presenting the effects of pre-conception LT4 on (a) live birth, (b) pregnancy, and (c) miscarriage in women with subclinical hypothyroidism. CI, confidence interval; TSH, thyrotropin.

Pregnancy rate data were available for three of the included studies (Fig. 2b).^7,18,19 All these three studies were conducted during ART treatment, and the majority of the participants were TPOAb-positive women. LT4 treatment did not improve the pregnancy rate compared with the controls (RR 1.73, 0.88–3.39; p = 0.11). Significantly high heterogeneity existed among the three studies (p = 0.003; I² = 83%).

Miscarriage data were available for three of the included studies (Fig. 2c).^7,18,19 All these three studies were conducted during ART treatment, and the majority of the participants were TPOAb-positive women. LT4 treatment did not decrease miscarriage rates compared with the controls (RR 0.46, 0.11–2.00; p = 0.30). Moderately high heterogeneity existed among the three studies (p = 0.12, I² = 52%).

Although tests for subgroup differences showed no significance in all of the three primary outcomes (live birth, I² = 73.9%, p = 0.05; pregnancy, I² = 50.9%, p = 0.15; and miscarriage, I² = 65.5%, p = 0.09, respectively), the potential positive effects for each outcome were more evident in the studies of subclinical hypothyroidism with TSH levels >4.0 mU/L than with TSH levels of 2.5–4.0 mU/L.

The subgroup with TSH levels of 2.5–4.0 mU/L in the analyses on pregnancy and miscarriage is composed of a single study,⁷ as we could not obtain data specifically focused on a TSH above 2.5 mU/L in the other two studies.^8,9

In the sensitivity analysis to exclude the study with high risk in other biases,¹⁸ the effect on live birth was significantly different between the subgroups with TSH levels of 2.5–4.0 and >4.0 mU/L (I² = 80.5%, p = 0.02). In the subgroup with TSH levels of >4.0 mU/L, consisting of a single study,¹⁹ LT4 treatment significantly improved the live birth rate compared with the controls (RR 2.13, 1.07–4.21; p = 0.03). On the other hand, in the subgroup with TSH levels of 2.5–4.0 mU/L, LT4 treatment did not improve the live birth rate compared with the controls (RR 0.93, 0.76–1.15; p = 0.50) with low risk of heterogeneity (I² = 0%, p = 0.54). LT4 treatment did not improve the pregnancy rate and decrease miscarriage rates compared with the controls (RR 1.20, 0.88–1.64; p = 0.24, and RR 0.50, 0.02–14.59; p = 0.69, respectively). Tests for subgroup differences showed no significance in pregnancy and miscarriage (I² = 0%, p = 0.49; and I² = 72.1%, p = 0.06). These results of the sensitivity analyses are shown in Supplementary Figure S2.

Effect of LT4 treatment during pregnancy

Data on live births were available for six of the included studies (Fig. 3a). The pooled estimates suggested that LT4 treatment had no significant effect on live birth rate compared with the controls (RR 1.03, 0.98–1.09; p = 0.23). Significantly high heterogeneity existed among the six studies (p < 0.001; I² = 87%). Among four of the six studies, we performed subgroup analysis to compare the effects in subclinical hypothyroidism with TSH levels of 2.5–4.0 and >4.0 mU/L (Fig. 3b). The two subgroups showed no significant difference (I² = 0%, p = 0.58).

FIG. 3.

Forest plots presenting the overall effects of LT4 during early pregnancy and subgroup analysis to compare the effects between TSH levels over 4.0 mU/L and TSH levels between 2.5 and 4.0 mU/L on (a, b) live birth, (c, d) miscarriage, and (e, f) preterm birth in women with subclinical hypothyroidism. (a, c, e) Results of the overall analysis. (b, d, f) Results of the subgroup analysis among studies with eligible data. The participants in Nazarpour et al.^24,25 were divided into two subgroups according to the information provided by the corresponding author.

Miscarriage data were available for five of the included studies (Fig. 3c). We performed subgroup analysis to divide the five studies into two subgroups (I² = 78.3%, p = 0.03): women with and without recurrent pregnancy loss. Miscarriage rates among women with recurrent pregnancy loss were significantly decreased by LT4 treatment (RR 0.54, 0.37–0.79; p = 0.002), while the recurrent pregnancy loss subgroup was composed of a single study with 267 participants. However, miscarriage rates were not decreased among women without recurrent pregnancy loss (RR 1.01, 0.66–1.53; p = 0.97). Among three of the four studies without recurrent pregnancy loss, we performed subgroup analysis to compare the effects of LT4 treatment in subclinical hypothyroidism with TSH levels of 2.5–4.0 and >4.0 mU/L (Fig. 3d). There was no significant difference between the two subgroups (I² = 0%, p = 0.38).

Data on preterm births were available for six of the included studies (Fig. 3e). The pooled estimates suggested that LT4 treatment showed a positive effect in decreasing the incidence of preterm birth, compared with the case for the controls, without statistical significance (RR 0.67, 0.43–1.04; p = 0.08). Moderately high heterogeneity existed among the six studies (p = 0.07, I² = 51%).

We performed a subgroup analysis with four of the six analyzed studies to compare the effects of LT4 treatment in subclinical hypothyroidism with TSH levels of 2.5–4.0 and >4.0 mU/L (Fig. 3f). There was significant difference between the two subgroups (I² = 76.3%, p = 0.04). Although LT4 treatment showed positive effect to decrease the incidence of preterm birth in the TSH >4.0 mU/L subgroup (RR 0.47, 0.20–1.10; p = 0.06), no effect was observed on preterm birth in the TSH 2.5–4.0 mU/L subgroup (RR 1.35, 0.79–2.31; p = 0.28).

Other pregnancy outcomes were analyzed, including: hypertensive disorders of pregnancy, preeclampsia, gestational diabetes mellitus, caesarean section, placental abruption, neonatal admission, small for gestational age, neonatal birth weight, and neurodevelopmental outcomes in childhood (intelligence quotient). No significant effect of LT4 treatment was observed for any of these outcomes (Supplementary Fig. S3).

Nazarpour et al.,²⁶ in a follow-up study of Nazarpour et al.,^24,25 investigated neurodevelopmental outcomes at the age of 3 years using the Ages and Stages Questionnaires to show no significant effect of LT4 in four of the five domains (communication, gross motor, fine motor, problem-solving, and social-personal domains), excluding gross motor.

Discussion

To the best of our knowledge, our study is the first to exclusively focus on RCTs and to perform subgroup analyses focused on pregnant patients with subclinical hypothyroidism, including the subgroups of TSH levels of 2.5–4.0 and >4.0 mU/L.

Previous meta-analyses included studies of variable quality levels of evidence and variable eligibility of participants; euthyroidism with TPOAb positivity, subclinical hypothyroidism, and overt hypothyroidism.^{10,11,13,28

–31} Hence, there have been discrepancies in the results, and the subgroups that can derive the most benefit from LT4 treatment remain unclear, that is, the TSH cutoff values, TPOAb positivity, the number of previous pregnancies lost, and the regimen of LT4 treatment.

However, there has been a marked increase in periconceptional LT4 prescriptions and the threshold for initiating LT4 treatment has declined over the past decade.^5,6,32

Although statistical significance was not observed in any subgroup analyses involving pre-conceptional LT4, the most benefit was observed in the subgroup where subclinical hypothyroidism was defined by TSH levels of >4.0 mU/L (live birth: RR 4.04, 0.94–17.25; p = 0.06, pregnancy: RR 2.20, 0.94–5.18; p = 0.07, miscarriage: RR 0.28, 0.07–1.06; p = 0.06).

The three high-quality RCTs among TPOAb positive women who have experienced infertility or pregnancy loss showed consistent result that pre-conception LT4 did not improve live births in subclinical hypothyroidism with TSH levels between 2.5 to 4.0 mU/L (p = 0.50, I ² = 0%).^7
–9 However, further studies are needed for TSH levels of >4.0 mU/L, considering the small sample size and the fact that one of the two studies, Rahman 2010, was subject to a high risk in other bias.¹⁸ Based on the current evidence, we believe that the target population for pre-conception LT4 treatment should be narrowed down to those with subclinical hypothyroidism with TSH levels of >4.0 mU/L.

In the present analysis of LT4 administration during pregnancy, no effects on live births or miscarriages were observed. While the critical period for miscarriage is the early phase within the first trimester of pregnancy,³³ LT4 treatment was initiated at almost the end of the first trimester in many of the included studies. Thus, the present analysis of LT4 treatment during pregnancy might be inappropriate for assessing the effect on miscarriages and live births.

Alcazar Lazaro et al.²¹ and Leng et al.²³ did not show significantly decreased incidence rates of miscarriages with LT4 treatment, although LT4 was initiated within the first trimester. The limited impact of LT4 on miscarriage might be because the estimated risk of subclinical hypothyroidism may not be as important as the other risk factors for early pregnancy loss, including maternal age, diabetes mellitus, and obesity.

Leng et al. showed a significant effect of LT4 in preventing miscarriage in subclinical hypothyroidism patients with recurrent pregnancy loss.²³ Two observational studies investigated the effect of LT4 on recurrent pregnancy loss in subclinical hypothyroidism patients and found no significant effects.^34,35 No other RCTs have primarily investigated the effects of LT4 in women with recurrent pregnancy loss. Van Dijk et al.⁹ and the subgroup analysis by Dhillon-Smith et al.⁸ showed no significant effect of LT4 in preventing miscarriage in TPOAb-positive euthyroid participants with recurrent pregnancy loss. Although not statistically significant, van Dijk et al. showed a tendency of reduced pregnancy loss rates in the LT4 group compared with that in the placebo group (23% and 33%, respectively).⁹ Further studies are needed to clarify whether recurrent pregnancy loss is treatable with LT4 in subclinical hypothyroidism.

LT4 treatment was shown to reduce the incidence of preterm births during pregnancy. Notably, this effect was observed only in the subgroup with TSH levels >4.0 mU/L, but not in the subgroup with TSH levels of 2.5–4.0 mU/L. The subgroup with TSH levels of 2.5–4.0 mU/L mainly comprised TPOAb-negative women. In an independent-patient-data meta-analysis by the Consortium on Thyroid and Pregnancy, the association between subclinical hypothyroidism and preterm birth was no longer apparent after additional adjustment for TPOAb positivity. Further studies investigating the association between TPOAb positivity and TSH levels are needed to identify the optimal TSH target for LT4 treatment to prevent preterm birth.

In the present study, no significant negative effects of LT4 on adverse pregnancy outcomes were observed. In the retrospective cohort study, Maraka et al. reported a higher risk of preterm birth, preeclampsia, and gestational diabetes mellitus in the LT4 group than in the control group.³⁶ The follow-up study of the Controlled Antenatal Thyroid Screening Study²⁰ for children around the age of 9 of the mothers with suboptimal gestational thyroid function showed increased risk of attention deficit hyperactivity disorder symptoms and behavioral difficulties in the overtreated group compared with the untreated group.³⁷

Recent independent patient data meta-analyses have shown that high-to-normal fT4 levels within the reference range, which are compatible with LT4 treatment, are associated with an increased risk of lower neonatal birthweight.³⁸ Moreover, the low-to-normal TSH levels within the reference range were shown to be associated with an increased risk of preeclampsia.³ These findings indicate that LT4 overtreatment, even within the normal range, may lead to potential harm.

Yang et al. showed a U-shaped association between pre-conception TSH levels and adverse pregnancy outcomes, including preterm birth and small for gestational age, in a large-scale population-based cohort study, indicating that the optimal target ranges of TSH and fT4 might be quite narrow.³⁹ Two of the studies on pre-conceptional LT4^8,9 and four of the studies on LT4 during pregnancy^23

–27 had no dose adjustment after treatment initiation, raising concerns about overtreatment. The effects of LT4 might have been underestimated due to the potential overtreatment. Reconsidering the possibility of overtreatment and verifying the optimal target for LT4 treatment is necessary.

Since the early 2010s, several guidelines have recommended LT4 treatment before⁴⁰ or during pregnancy^4,16,40 at a TSH threshold of 2.5 mU/L, despite a lack of sufficient evidence. However, over the past half decade, multiple RCTs have been conducted to determine the appropriate treatment targets. The results of the present study suggest there is a need to reconsider the TSH threshold for LT4 treatment both before and during pregnancy.

The main strength of our study is the strict definition of subclinical hypothyroidism and the subgroup analysis based on TSH values. We exclusively analyzed RCTs and extracted the results of eligible participants, excluding those with isolated hypothyroxinemia, overt hypothyroidism, or euthyroidism with TSH levels <2.5 mU/L. A limitation of our study is the lack of important data, including those for TPOAb positivity and TSH values after LT4 treatment.

In conclusion, in this systematic review and meta-analysis, pre-conception LT4 treatment had no effects on live births, pregnancies, or miscarriages in subclinical hypothyroidism with TSH levels of 2.5 to 4.0 mU/L. Evidence investigating the effects of LT4 treatment for subclinical hypothyroidism with TSH levels >4.0 mU/L is still insufficient. In secondary analyses, we found that LT4 treatment in early pregnancy has the potential to reduce preterm birth rates, but this benefit was observed only among individuals with TSH levels >4.0 mU/L. Further well-designed clinical trials are needed to determine the beneficial effects of LT4 treatment, stratifying for both the TSH and TPOAb levels.

Footnotes

Acknowledgments

The authors thank Prof. Fahimeh Ramezani Tehrani for sending additional data for this study on Nazarpour 2017 and 2018. The data enabled us to perform a detailed subgroup analysis. This study was conducted in collaboration with members of the committee of the Japan Thyroid Association (JTA) to investigate the diagnosis and treatment of subclinical thyroid dysfunction during pregnancy. The authors also thank Dr. Akira Hisinuma, chair of the JTA, and Dr. Masanobu Yamada, chair of the committee. They thank Cochrane Japan (https://japan.cochrane.org/) for executing the search. They also thank Editage for their writing support on the manuscript.

Authors' Contributions

A.S.: Study selection and assessment (lead), formal analysis (lead), writing–original draft (lead), and review and editing (equal). N.A.: Conceptualization (lead), writing–review and editing (equal). H.S.: Study selection and assessment (supporting), methodology (equal), formal analysis (supporting), and review and editing (equal). E.O.: Methodology (lead), review and editing (equal). M.I., A.Y., S.K., M.K., K.H., Y.H., A.Y., H.N., S.K., M.K.-K., A.I., and T.S.: Conceptualization (supporting), review and editing (equal).

Author Disclosure Statement

N.A., H.N., and A.Y. received speaking fees from ASKA Pharmaceutical Co., Ltd. The other authors have nothing to declare.

Funding Information

This study was supported by a grant from the Children and Families Agency (Grant No.: 23DA0201). The funder had no role in the study. N.A. received the funding. The other authors do not receive any funding to support this study.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Table S1

Supplementary Table S2

References

Maraka

, Ospina

, O'Keeffe

, et al. Subclinical hypothyroidism in pregnancy: A systematic review and meta-analysis. Thyroid, 2016; 26(4):580–590; doi: 10.1089/thy.2015.0418

Korevaar

TIM

, Derakhshan

, Taylor

, et al. Association of Thyroid Function Test Abnormalities and Thyroid autoimmunity with preterm birth: A systematic review and meta-analysis. JAMA, 2019; 322(7):632–641; doi: 10.1001/jama.2019.10931

Toloza

FJK

, Derakhshan

, Mannisto

, et al. Association between maternal thyroid function and risk of gestational hypertension and pre-eclampsia: A systematic review and individual-participant data meta-analysis. Lancet Diabetes Endocrinol, 2022; 10(4):243–252; doi: 10.1016/S2213-8587(22)00007-9

Alexander

, Pearce

, Brent

, et al. 2017 Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease During Pregnancy and the Postpartum. Thyroid, 2017; 27(3):315–389; doi: 10.1089/thy.2016.0457

Turunen

, Vaarasmaki

, Leinonen

, et al. The increased trend of medical treatment for thyroid diseases during pregnancy: A 13-year national study. Eur Thyroid J, 2021; 10(3):230–236; doi: 10.1159/000515125

, Filion

, Reynier

, et al. Use of levothyroxine among pregnant women with subclinical hypothyroidism in the United Kingdom: A population-based assessment. Pharmacol Res Perspect, 2021; 9(5):e00848; doi: 10.1002/prp2.848

Wang

, Gao

, Chi

, et al. Effect of levothyroxine on miscarriage among women with normal thyroid function and thyroid autoimmunity undergoing in vitro fertilization and embryo transfer: A randomized clinical trial. JAMA, 2017; 318(22):2190–2198; doi: 10.1001/jama.2017.18249

Dhillon-Smith

, Middleton

, Sunner

, et al. Levothyroxine in women with thyroid peroxidase antibodies before conception. N Engl J Med, 2019; 380(14):1316–1325; doi: 10.1056/NEJMoa1812537

van Dijk

, Vissenberg

, Fliers

, et al. Levothyroxine in euthyroid thyroid peroxidase antibody positive women with recurrent pregnancy loss (T4LIFE trial): A multicentre, randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Diabetes Endocrinol, 2022; 10(5):322–329; doi: 10.1016/s2213-8587(22)00045-6

10.

Ding

, Liu

, Maraka

, et al. Pregnancy and neonatal outcomes with levothyroxine treatment in women with subclinical hypothyroidism based on new diagnostic criteria: A systematic review and meta-analysis. Front Endocrinol (Lausanne), 2021; 12:797423; doi: 10.3389/fendo.2021.797423

11.

Rao

, Zeng

, Zhou

, et al. Effect of levothyroxine supplementation on pregnancy loss and preterm birth in women with subclinical hypothyroidism and thyroid autoimmunity: A systematic review and meta-analysis. Hum Reprod Update, 2019; 25(3):344–361; doi: 10.1093/humupd/dmz003

12.

Rao

, Zeng

, Zhao

, et al. Effect of levothyroxine supplementation on pregnancy outcomes in women with subclinical hypothyroidism and thyroid autoimmuneity undergoing in vitro fertilization/intracytoplasmic sperm injection: An updated meta-analysis of randomized controlled trials. Reprod Biol Endocrinol, 2018; 16(1):92; doi: 10.1186/s12958-018-0410-6

13.

Akhtar

, Agrawal

, Brown

, et al. Thyroxine replacement for subfertile women with euthyroid autoimmune thyroid disease or subclinical hypothyroidism. Cochrane Database Syst Rev, 2019; 6(6):CD011009; doi: 10.1002/14651858.CD011009.pub2

14.

Stroup

, Berlin

, Morton

, et al. Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA, 2000; 283(15):2008–2012; doi: 10.1001/jama.283.15.2008

15.

Korevaar

TIM

, Medici

, Visser

, et al. Thyroid disease in pregnancy: New insights in diagnosis and clinical management. Nat Rev Endocrinol, 2017; 13(10):610–622; doi: 10.1038/nrendo.2017.93

16.

De Groot

, Abalovich

, Alexander

, et al. Management of thyroid dysfunction during pregnancy and postpartum: An Endocrine Society clinical practice guideline. J Clin Endocrinol Metab, 2012; 97(8):2543–2565; doi: 10.1210/jc.2011-2803

17.

Higgins

, Savović

, Page

, et al. Revised Cochrane risk-of-bias tool for randomized trials (RoB 2). Cochrane Collaboration, 2019. Available from: https://methods.cochrane.org/bias/resources/rob-2-revised-cochrane-risk-bias-tool-randomized-trials [Last accessed: February 5, 2024 ].

18.

Abdel Rahman

, Aly Abbassy

, Abbassy

. Improved in vitro fertilization outcomes after treatment of subclinical hypothyroidism in infertile women. Endocr Pract, 2010; 16(5):792–797; doi: 10.4158/ep09365.Or

19.

Kim

, Ahn

, Kang

, et al. Effect of levothyroxine treatment on in vitro fertilization and pregnancy outcome in infertile women with subclinical hypothyroidism undergoing in vitro fertilization/intracytoplasmic sperm injection. Fertil Steril, 2011; 95(5):1650–1654; doi: 10.1016/j.fertnstert.2010.12.004

20.

Casey

, Thom

, Peaceman

, et al. Treatment of subclinical hypothyroidism or hypothyroxinemia in pregnancy. N Engl J Med, 2017; 376(9):815–825; doi: 10.1056/NEJMoa1606205

21.

Alcazar Lazaro

, Lopez Del Val

, Garcia Lacalle

, et al. Slightly elevated thyrotropin levels in pregnancy in our clinical practice. Endocrinol Diabetes Nutr (Engl Ed), 2019; 66(10):620–627; doi: 10.1016/j.endinu.2019.03.020

22.

Lazarus

, Bestwick

, Channon

, et al. Antenatal thyroid screening and childhood cognitive function. N Engl J Med, 2012; 366(6):493–501; doi: 10.1056/NEJMoa1106104

23.

Leng

, Li

, Zhang

. Levothyroxine treatment for subclinical hypothyroidism improves the rate of live births in pregnant women with recurrent pregnancy loss: A randomized clinical trial. Gynecol Endocrinol, 2022; 38(6):488–494; doi: 10.1080/09513590.2022.2063831

24.

Nazarpour

, Ramezani Tehrani

, Simbar

, et al. Effects of levothyroxine treatment on pregnancy outcomes in pregnant women with autoimmune thyroid disease. Eur J Endocrinol, 2017; 176(2):253–265; doi: 10.1530/eje-16-0548

25.

Nazarpour

, Ramezani Tehrani

, Simbar

, et al. Effects of levothyroxine on pregnant women with subclinical hypothyroidism, negative for thyroid peroxidase antibodies. J Clin Endocrinol Metab, 2018; 103(3):926–935; doi: 10.1210/jc.2017-01850

26.

Nazarpour

, Ramezani Tehrani

, Sajedi

, et al. Lack of beneficiary effect of levothyroxine therapy of pregnant women with subclinical hypothyroidism in terms of neurodevelopment of their offspring. Arch Gynecol Obstet, 2023; 309(3):975–985; doi: 10.1007/s00404-023-06954-5

27.

Zhao

, Jiang

, Tian

, et al. Initiation timing effect of levothyroxine treatment on subclinical hypothyroidism in pregnancy. Gynecol Endocrinol, 2018; 34(10):845–848; doi: 10.1080/09513590.2018.1451836

28.

Reid

, Middleton

, Cossich

, et al. Interventions for clinical and subclinical hypothyroidism pre-pregnancy and during pregnancy. Cochrane Database Syst Rev, 2013; 5:CD007752; doi: 10.1002/14651858.CD007752.pub3

29.

, Shen

, Qin

. Effects of levothyroxine on pregnancy outcomes in women with thyroid dysfunction: A meta-analysis of randomized controlled trials. Altern Ther Health Med, 2017; 23(2):49–58.

30.

Wang

, Zhang

, Tan

, et al. Effect of levothyroxine on pregnancy outcomes in women with thyroid autoimmunity: A systematic review with meta-analysis of randomized controlled trials. Fertil Steril, 2020; 114(6):1306–1314; doi: 10.1016/j.fertnstert.2020.06.034

31.

Lau

, Benham

, Lemieux

, et al. Impact of levothyroxine in women with positive thyroid antibodies on pregnancy outcomes: A systematic review and meta-analysis of randomised controlled trials. BMJ Open, 2021; 11(2):e043751; doi: 10.1136/bmjopen-2020-043751

32.

Chang

, Huang

, Shen

, et al. Prevalence, trends, and characteristics of polypharmacy among US pregnant women aged 15 to 44 years: NHANES 1999 to 2016. Medicine (Baltimore), 2023; 102(22):e33828; doi: 10.1097/MD.0000000000033828

33.

Neill

. Management of Early Pregnancy Loss. JAMA, 2023; 329(16):1399–1400; doi: 10.1001/jama.2023.0933

34.

Bernardi

, Cohen

, Stephenson

. Impact of subclinical hypothyroidism in women with recurrent early pregnancy loss. Fertil Steril, 2013; 100(5):1326–1331; doi: 10.1016/j.fertnstert.2013.07.1975

35.

Yoshihara

, Sugiura-Ogasawara

, Goto

, et al. Levothyroxine and subclinical hypothyroidism in patients with recurrent pregnancy loss. Am J Reprod Immunol, 2021; 85(3):e13341; doi: 10.1111/aji.13341

36.

Maraka

, Mwangi

, McCoy

, et al. Thyroid hormone treatment among pregnant women with subclinical hypothyroidism: US national assessment. BMJ, 2017; 356:i6865; doi: 10.1136/bmj.i6865

37.

Hales

, Taylor

, Channon

, et al. Controlled Antenatal Thyroid Screening II: Effect of treating maternal suboptimal thyroid function on child behavior. J Clin Endocrinol Metab, 2020; 105(3):dgz098; doi: 10.1210/clinem/dgz098

38.

Derakhshan

, Peeters

, Taylor

, et al. Association of maternal thyroid function with birthweight: A systematic review and individual-participant data meta-analysis. Lancet Diabetes Endocrinol, 2020; 8(6):501–510; doi: 10.1016/s2213-8587(20)30061-9

39.

Yang

, Guo

, Fu

, et al. Preconception thyrotropin levels and risk of adverse pregnancy outcomes in Chinese Women Aged 20 to 49 Years. JAMA Netw Open, 2021; 4(4):e215723; doi: 10.1001/jamanetworkopen.2021.5723

40.

Garber

, Cobin

, Gharib

, et al. Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Endocr Pract, 2012; 18(6):988–1028; doi: 10.4158/EP12280.GL