Table 1 Summary of studies included in the systematic review and meta-analysis

Bernardi et al.; 2013 [32]

University of Chicago

Cohort study

N = 39

United States

July 2004–December 2011

TSH > 2.5mIU/L with normal fT4 level (0.9–1.7 ng/dL) measured before conception

Initiated if SCH was identified pre-conception

Initial dose and adjustment unknown.

Live-birth rates

Among women with a history of > 2 pregnancy losses and SCH, women who received treatment for SCH did not have an increased live birth rate compared to women who did not receive levothyroxine treatment (per-pregnancy live birth rate: SCH treated: 48% (22/46) vs. SCH untreated: 52% (12/23).

Maraka et al. 2016 [33]

Mayo Clinic

Retrospective

Cohort

N = 366

Rochester, Minnesota

January 2011–December 2013

TSH 2.5–10 mIU/L in first trimester & 3–10 mIU/L in other trimesters normal fT4 (> 0.8 ng/dL)

Started at median GA 9.1 weeks

Pregnancy loss defined as miscarriage and stillbirth, preterm delivery (< 37 weeks), premature rupture of membranes, placental abruption, gestational diabetes, gestational hypertension, pre-eclampsia, eclampsia, intrauterine growth restriction, birth weight, Apgar score at 5 min, admission to the neonatal intensive care unit, neonatal death (during immediate postpartum period until discharge of the mother) and duration of hospital stay.

Treated pregnancies had a lower birth weights and no Apgar score less than 7. Other pregnancy outcomes were statistically similar between groups.

Pregnancy loss OR: 2.44; 95% CI: 0.80–8.87

Preterm delivery OR: 3.06; 95% CI: 0.96–12.28

Gestational diabetes OR: 3.31; 95% CI: 0.91–16.57

Gestational hypertension OR: 0.64; 95% CI: 0.23–1.93

Pre-eclampsia OR: 3.37; 95% CI: 0.66–26.84

Premature rupture of membranes OR: 0.71; 95% CI: 0.29–1.79

Intrauterine growth restriction OR: 1.45; 95% CI: 0.23–28.1

Placenta previa and placental abruption: not enough events to do multivariate analyses

NICU admission OR: 1.94; 95% CI: 0.38–15.36

Birth weight < 2500 g OR: 16.4; 95% CI: 2.7–326.9

Neonatal death and congenital malformations: not enough events to do multivariate analyses

Al-Anbari, 2017

Prospective study N=149

High Institute of Infertility diagnosis and assisted reproductive technologies/Al-Nahrain University, Iraq

TSH > 2.5mIU/L prior to conception

Initiated if SCH was identified pre-conception

Pregnancy rate, miscarriage rate, multiple pregnancy rate and live birth rate

Significantly increased pregnancy rate (19/75 among levothyroxine treated versus 8/74 among women not given treatment).

No multiple pregnancies in both groups.

No difference in miscarriage rate (2/75 among levothyroxine treated versus vs 2/74 among women not given treatment).

Maraka et al. 2017 [34]

OptumLabs Data Warehouse

Retrospective

Cohort

N = 5405

United States

January 2010–December 2014

TSH 2.5–10 mIU/L from 1 month prior to 3 months after first prenatal visit fT4 0.8 ng/dL or total thyroxine 7.5 mcg/dL

Started at median GA 28.7 weeks before birth

Unknown initial dose.

Median dose: 50 mcg daily (range 25–300mcg daily)

Pregnancy loss defined as miscarriage and still-birth, preterm delivery, preterm labor, premature rupture of membranes, placental abruption, gestational diabetes, gestational hypertension, pre-eclampsia, poor fetal growth, tachycardia

Treatment of SCH was associated with decreased risk of pregnancy loss but was associated with increased risk of other adverse pregnancy related outcomes.

Pregnancy loss OR: 0.62; 95% CI: 0.48–0.82

Preterm delivery OR: 1.6; 95% CI: 1.14–2.24

Preterm labor OR: 1.14; 95% CI: 0.89–1.46

Premature rupture of membranes OR: 0.97; 95% CI: 0.66–1.42

Placental abruption OR: 1.60; 95% CI: 0.65–3.93

Gestational diabetes OR: 1.37; 95% CI: 1.05–1.79

Gestational hypertension OR: 1.27; 95% CI: 0.88–1.82

Pre-eclampsia OR: 1.61; 95% CI: 1.10–2.37

Poor fetal growth OR: 1.12; 95% CI: 0.84–1.5

Tachycardia OR: 1.77; 95% CI: 1–3.11

Nazarpour et al. 2017 [35]

Shahid Beheshti Medical University

RCT

N = 1294

Tehran, Iran

September 2013–February 2016

TSH 2.5–10 mIU/L fT4 1–4.5 TPOAb positive

Started 4 to 8 d after prenatal visit, the mean GA at initial visit was 10.8 weeks.

1 mcg/kg daily

Dose adjustment not described.

Preterm delivery, neonatal admission, placental abruption, still birth, GA mean birth weight, neonate height, birth head circumference, neonatal TSH

There were no significant differences in preterm delivery or neonatal admission between treated and untreated women with SCH (TSH < 4mIU/L) treated versus untreated but there were significant differences in preterm delivery and neonatal admission between treated and untreated women with TSH > 4mIU/L. There was also a significant decrease in neonatal admission among women treated for SCH versus women who did not receive treatment [number (%): 2 (3.6) vs.12 (20.7)] There were no significant differences in placental abruption, still birth, and gestational age between treated versus untreated women with SCH.

There were no significant differences in mean birth weight, head circumference and neonate TSH levels between study groups.

Nazarpour et al. 2018 [36]

Shahid Beheshti Medical University

RCT

N = 354

Tehran, Iran

September 2013–February 2016

TSH 2.5–10 mIU/L fT4 1–4.5 TPOAb

negative

Started 4–8 days after first prenatal visit, which was at 11.2–12.2 weeks of gestation

Dosed at 1 mcg/kg daily.

Dose adjustment not described

Preterm delivery, placental abruption, stillbirth, neonatal admission, birth weight, mean gestational age neonate height, birth head circumference neonatal TSH

Significant difference in pre-term delivery when TSH > 4 mIU/L and treated with levothyroxine versus no treatment, RR: 0.38; 95% CI: 0.15–0.98. There was no significant difference in the risk of other adverse pregnancy outcomes among women with SCH treated versus untreated.

Number (%) of outcomes treated versus untreated:

Preterm delivery: 18 (9.8) vs. 21 (11.5)

Neonatal admission: 8 (4.5) vs. 9 (4.9)

Placental abruption: 3 (1.6) vs. 0

Stillbirth: 0 vs. 0

Gestational age: 38.03 (1.4) vs. 37.9 (1.5)

Mean (standard deviation)

Birth weight: 3190.82 g (455.13) vs. 3203.1 g (497.1)

Neonate height: 50.1 cm (2.3) vs. 50.2 cm (2.7)

Birth head circumference: 34.6 cm (1.4) vs. 34.7 cm (1.6) Neonatal TSH (mIU/L): 1.1 (0.5–1.9) vs. 1.1 (0.5–2.1)

Casey et al. 2017 [37]

RCT

N = 677 women with SCH, 526 with hypothyroxinemia

United States

October 2006–October 2009

TSH > 3mIU/L or fT4 < 0.86 ng/dl with TSH between 0.08 and 3.99mIU/L

Started at GA 10–24 weeks; average GA 16.6 +/−  3 week standard deviation

Initial dose 100 mcg daily.

Monthly adjustment to maintain TSH 0.1–2.5 mIU/L

Max dose: 200 mcg daily

Treatment group normalized TSH by median of GA 21 weeks

Multiple pregnancy and neonatal outcomes. For maternal outcomes: preterm birth (<  34 and < 37 weeks), placental abruption, gestational hypertension, preeclampsia, gestational diabetes. For fetal and neonatal outcomes: stillbirth/miscarriage, neonatal death, Apgar score at 1 and 5 min, birth weight (< 10 percentile), head circumference, respiratory distress syndrome, necrotizing enteritis, bronchopulmonary dysplasia, respiratory therapy > 1 day, number of days in nursery. Annual cognitive testing over first 5 years, IQ at 5 yo. or general conceptual ability at 3 yo

No differences in adverse pregnancy and neonatal outcomes between levothyroxine treated versus placebo group. No differences in IQ score at the age of 5 years or death at the age of < 3 years between treated and untreated SCH (median IQ score for children in levothyroxine treated group: 97 (95% CI: 94–99) versus 94 (95% CI: 91–95) in placebo.

Lazarus et al. 2012 [38]

RCT

N = 794

Wales & Cardiff, UK Ospendale Sant’Anna, Turin, Italy

Period of time not mentioned but followed children at age 3 years

TSH > 97.5th percentile and f T4 < 2.5th percentile

Started at GA 12–13 weeks

Initial dose of 150 mcg daily.

Adjustment 6 weeks after beginning treatment and at 30 weeks GA to target TSH 0.1–1 mIU/L

IQ at 3 years of age

No significant differences in IQ scores at age 3 years between children born of mothers with SCH treated versus placebo during pregnancy: IQ score treated: 99.2 +/−  13.3 vs. untreated: 100 +/−  13.3.

Kim et al. 2011 [39]

RCT

N = 64

Seoul, South Korea

March 2006–September 2009

TSH > 4.5mIU/L with normal fT4

Started prior to IVF/ICSI treatment (before pregnancy).

Initial dose 50 mcg daily.

Adjustment: First trimester titration to maintain TSH < 2.5 mIU/L to max dose 100mcg daily (12/17) and 125mcg daily (1/17)

Embryo implantation rate, total amount and days of rhFSH administered, number of retrieved, mature, fertilized oocytes, and good quality embryos, clinical pregnancy rate per cycle, and miscarriage rate, preterm birth (< 34 weeks) and live birth (delivery of fetus> 20 weeks with signs of life)

Significant increase in embyro quality and implantation when women with SCH were treated versus untreated. Significant decrease in miscarriage and increase in live birth rate per cycle when women with SCH were treated versus untreated.

Miscarriage rate in treated versus untreated: 0/17 vs. 4/12

Live birth rate per cycle in treated versus untreated: 17/32 vs. 8/32

Ju et al. 2016 [40]

Beijing Friendship Hospital of Capital Medical University

Prospective cohort

N = 457

Beijing, China

October 2010–September 2013

TSH > 97.5th percentile and fT4 2.5th–97.5th percentile

Started at approx. GA 10 weeks

Initial dose 100 mcg daily.

Adjusted by 100 mcg to maintain TSH 2.5–97.5 percentile.

Premature rupture of membranes, fetal macrosomia, gestational diabetes, hypertensive disorders in pregnancy, postpartum hemorrhage, preterm labor, oligohydramnios, fetal distress, and low birth weight

Overall risk of pregnancy complications in control group significantly higher than in treated group (OR: 1.219; 95% CI 1.139–1.304). For single outcomes, there was a statistically increased risk of gestational diabetes among untreated versus treated women with SCH (OR: 1.938; 95% CI: 1.267–2.964) and fetal macrosomia (OR:3.081; 95% CI: 1.783–5.326).

Wang et al. 2012 [41]

Liaoning Provincial Key Laboratory of Endocrine Diseases

Prospective cohort

N = 196

Shenyang, China

2007–2009

TSH > 2.5 mIU/L with a fT4 between 12 pmol/L and 23.34 pmol/L in the first 12 weeks of pregnancy

Started at approximately 6 weeks.

Initial dose at 50 μg, 75 μg, or 100 μg daily

Adjustment every 4 weeks by serum TSH (mIU/L): 2.5–5: 50 μg daily 5–8: 70 μg daily > 8: 100 μg daily

Spontaneous abortion, anemia, hypertension, premature delivery, low birth weight, post-partum hemorrhage, Apgar score ≤ 7 at 5 min

Levothyroxine treatment in women with SCH decreased the incidence of spontaneous abortions compared to no treatment but this was not statistically significant.

Outcomes in treated versus untreated women with SCH:

Spontaneous abortions: 2 (7.14%) vs. 26 (15.47%)

Anemia: 3 (10.71%) vs. 35 (20.83%)

Hypertension: 1 (3.57%) vs. 2 (1.19%)

Premature delivery: 0 vs. 9 (5.36%)

Low birth weight: 0 vs. 4 (2.38%)

Hemorrhage: 0 vs. 2 (1.19%)

Apgar score < 7 at 5 min: 0 vs. 4 (2.38%)

Zhao et al. 2018 [42]

Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine

RCT

N = 93

Shanghai, China

January 2014–October 2016

T1:TSH > 2.5mIU/L T2: TSH > 3mIU/L

Started either in T1 at 8-10 weeks or T2 at 13–16 weeks.

Initial dose 25mcg daily.

Max dose: 100 mcg daily

Gestational hypertension, anemia, gestational diabetes, pre-eclampsia, premature labor, pregnancy loss, post-partum hemorrhage, low birth weight (< 2500 g)

Pregnancy complications: no significant different between treated versus untreated groups for individual outcomes. However, for combined outcomes: treatment given during T1 had significantly less complications than women who were not treated (Number of total adverse pregnancy outcomes among women treated at T1: 1/31 vs. 10/31 (treated at T2) vs. 12/31 (no treatment).

Zhang et al. 2017 [12]

Retrospective cohort

N = 9

ZhongDa, China

January 2014–May 2014

TSH 0.27–4.2 mIU/L and fT4 0.93–1.70 ng/dL (in T2)

Started in second trimester.

Dose at 50 mcg daily

Premature delivery (< 37 weeks), Apgar score, birth weight

No significant differences in the rate of premature delivery (0/1 in treated vs. 3/8 untreated), Apgar (10 +/−  0 in treated vs. 9.8+/−  0.61 untreated), and birth weight (3.67+/− 0.6 kg in treated vs. 3.48+/−  0.54 kg in untreated) in women with treated and untreated SCH.