The impact of levothyroxine exposure on delivery outcome in hypothyroid pregnant women (PETAL study): A five-year retrospective cohort study

Abstract

Background

Hypothyroidism affects 3% of pregnant women, and to date, no studies have addressed the impact levothyroxine-treated hypothyroidism on delivery outcome.

Methods

This retrospective cohort study was conducted among 750 women with a singleton pregnancy who gave birth between 2015 and 2019. Delivery modes were compared between 250 hypothyroid women exposed to levothyroxine and 500 euthyroid control women. The aim of this study was to determine the impact of levothyroxine exposure on delivery outcome.

Results

Multiple logistic regression showed no significant association between exposure to levothyroxine and the overall rate of caesarean delivery (aOR 1.1; 95% CI 0.8 to 1.6). Mean TSH concentrations were significantly higher throughout the pregnancy in hypothyroid women despite levothyroxine treatment. Maternal and neonatal outcomes in both groups were not different.

Conclusion

Hypothyroidism treated with levothyroxine during pregnancy according to local guidelines is not a significant risk factor for caesarean delivery.

Keywords

Perinatal outcome hypothyroidism levothyroxine caesarean delivery uterine contractions pregnancy

Introduction

Hypothyroidism affects between 2% and 5% of pregnant women.¹ Thyroid hormone deficiency is known to cause adverse effects on pregnancy that impact the delivery outcome, such as pre-eclampsia, placenta abruption, and prematurity.^1,2 Some studies have also demonstrated a possible association with dystocia and fetal malpresentation leading to caesarean deliveries (CD), and this was explained by the possible thyroid hormonal influence over myometrial cell contraction and relaxation.^3,4 Turunen et al. supported this finding while observing an increased CD rate for hypothyroid pregnant women (21.2%) compared to euthyroid pregnant women (15.7%).⁵ They also observed a persistent significant risk association between CD and hypothyroidism treated with levothyroxine (OR 1.1, 95% CI 1.1–1.2) compared to euthyroidism. This finding supposes a slight attenuating effect of levothyroxine over the impact of hypothyroidism on uterine contractility and vaginal delivery prognosis, although the thyroid function nor the prescribed doses of levothyroxine were not considered in this study.⁵

CD have increased from 18.0% to 28.8% over the last 25 years in Canada, with similar increases in other countries.^6,7 Both emergency and elective CD are associated with increased maternal morbidity and mortality secondary to post-partum hemorrhage, infection, thromboembolic events^8–10 and substantial costs for the health-care system.³ Besides fetal distress, dystocia is the main indication for CD.³

There is a paucity of data about the effect of levothyroxine treatment on the delivery mode of women diagnosed with overt or subclinical hypothyroidism. Our team showed in previous studies a possible dose-dependent effect of levothyroxine on the uterine contractility pattern. For instance, we demonstrated decreased contraction frequency and increased amplitude and duration of contractions in the myometrium biopsies from hypothyroid women who were treated with levothyroxine throughout their pregnancy, and who had an elective CD.¹¹ We conducted a second study where the same altered contraction pattern was observed in hypothyroid female rats that were supposedly over-supplemented with levothyroxine, although their thyroid function was not verified.¹² Therefore, exposure to an inappropriate dosage of levothyroxine may possibly affect the uterine contraction efficiency in peripartum.

The objective of this retrospective cohort study was to evaluate the association between hypothyroidism treated with levothyroxine (according to local clinical guidelines) and delivery outcome, while comparing hypothyroid pregnant women to euthyroid pregnant women, matched according to age, parity and gestational age at delivery as per previous studies.

Methods

Design and participants

We conducted a retrospective cohort study with matched groups at the Centre Hospitalier Universitaire de Sherbrooke (CHUS) in Quebec, Canada. Our population included all women with a singleton pregnancy who delivered at the CHUS after 24 weeks of gestation. Women with multiple gestation were excluded. A database of all deliveries from 2015 to 2019 was created by our archives’ info-center, presenting extracted data from medical records including the CHUS record number, date of delivery, maternal age, gestational age at delivery and parity. The ICD code for unspecified endocrine disorder (E34.9) was also extracted for each woman in the database where applicable. Records presenting this code included diagnoses of overt and subclinical hypothyroidism. Diagnosis was made by a medical physician either before the current pregnancy or at the first trimester of pregnancy through the hypothyroidism universal screening program, in presence of TSH serum concentrations above 3.5 mUI/L and clinical presentation.

Starting from the most recent delivery of 2019, each record associated with the E34.9 ICD code was reviewed by the first author (DO) to verify a diagnosis of hypothyroidism and treatment with levothyroxine. Through a convenient sampling method, the first 250 women found to have confirmed diagnosis of clinical or subclinical hypothyroidism (according to the local clinical guidelines¹³) and a confirmed treatment with levothyroxine were selected and then assigned to the hypothyroid levothyroxine-exposed group.

For each selected hypothyroid woman exposed to levothyroxine, two control euthyroid women (without an E34.9 ICD code) were matched according to maternal age, parity and gestational age at delivery. These characteristics influence the delivery mode and were also used to match groups in previous studies assessing the delivery outcome.^14–16 The control group consisted of 500 euthyroid pregnant women not exposed to levothyroxine.

Each woman could only be included once. For hypothyroid women, only their most recent pregnancy was included in the levothyroxine-exposed group. If a euthyroid woman had more than one delivery in the study period, only one of her pregnancies was considered for matching with a levothyroxine-exposed hypothyroid woman. The demographics, gestational and obstetric characteristics of the selected women were extracted retrospectively and anonymously by the first author (DO) from their medical records.

The primary outcome was overall caesarean rate (including rates of elective CD and emergency CD in labor) in relation with exposure to levothyroxine. Many potential confounding factors were considered in assessing this outcome, including maternal, gestational, obstetric and neonatal characteristics that may influence the delivery mode. The secondary outcomes were (1) rates of assisted vaginal delivery (VAD) in relation with exposure to levothyroxine, (2) indications for CD and VAD, and (3) maternal and neonatal morbidity and mortality.

To determine the potential effect of levothyroxine exposure on delivery outcome, the operative delivery indications were categorized as “uterine muscle dysfunction-related” indications or unrelated indications. Based on our initial hypothesis, the former category included dystocia and fetal malpresentation, in addition to uterine atony and maternal exhaustion as direct and indirect contributing factors to dystocia.^3,10,17 Previous history of CD due to the potential impact of tissue scarring on uterine muscle function is also included in this category, as well as post-term pregnancy and failure of induced vaginal delivery, representing other signs of uterine muscle rigidity and decreased sensitivity to uterotonic agents.^11,12,18 In comparison, the unrelated indications included fetal distress, the most frequent indication of operative delivery unrelated to the uterine function.

We also collected all available TSH concentration values throughout all trimesters of pregnancy for each participant. Since universal screening for hypothyroidism in the first trimester was implemented at the CHUS before 2015, all women had at least one TSH concentration measurement in the first trimester of pregnancy. Hypothyroid women were all followed with serial TSH concentration measurements and treated with adjusted dosage of levothyroxine according to the local hypothyroidism in pregnancy protocol used uniformly by all health care professionals in our center, targeting a TSH value under 3.5 mIU/L.¹³ Euthyroid women had routine TSH measurement in the first trimester and only occasionally thereafter, according to clinical indications.

The project was approved by the CIUSSS de l’Estrie-CHUS research ethics committee with a delegated consent process (exempt from requiring individual consent (2017–1764)).

Statistical analysis

A previous local data review done by the Department of Obstetrics and Gynecology of Université de Sherbrooke showed an increased CD delivery rate in hypothyroid women treated with levothyroxine, estimating a possible effect size of 20% when compared to non-hypothyroid women not exposed to levothyroxine. Based on this information, the size of the levothyroxine-exposed group and the control group have been established respectively as 250 and 500 women, for a power of 80% and an alpha level of 5%.

Statistical analysis was performed using IBM SPSS Statistics 27.0. Descriptive analysis was presented as frequencies and proportions for categoric variables, and as mean and standard deviations for continuous numerical variables. Differences between groups were evaluated with chi-square test or Fisher's exact test for categoric variables, and with Student t test or Mann-Whitney U test for continuous numeric variables. The risk association between levothyroxine exposure during pregnancy and delivery mode was assessed by simple and multivariable logistic regression models including variables with statistically or clinically significant associations. No regression model selection procedure was used. Significant confounding variables with p<0.05 in univariable analysis were all included in the multivariable model. TSH values during the pregnancy were not considered in the model as a marker of adequate dosage of levothyroxine; exposure alone to levothyroxine was considered in hypothyroid women in this model. Confidence intervals for non-adjusted and adjusted odds ratios were calculated with an alpha level of 5%.

Results

Between January 2015 and December 2019, 13 923 deliveries occurred at the CHUS. Among these women, 66.8% had a spontaneous vaginal delivery (n = 9302), 24.6% had a CD (n = 3424), and 8.6% had an VAD (n = 1197). We excluded 289 women for multiple gestation and 63 women for delivery before 24 weeks of gestation. A total of 3081 women had more than one delivery during this period, which represented repetitive medical records. Of the remaining 10 490 records, 750 women were selected according to inclusion criteria.

Descriptive statistics regarding maternal, gestational, and neonatal characteristics are presented in Table 1. There was no difference between groups regarding the matching variables. Of the women exposed to levothyroxine, 64.4% had a diagnosis of hypothyroidism prior to pregnancy. Also, women exposed to levothyroxine showed a significantly higher body mass index (BMI) at the beginning of the pregnancy (27.38 vs. 26.00 kg/m², p = 0.02), although the proportion of obese women was similar (29.2% vs. 22.8%, p = 0.06). Women exposed to levothyroxine had higher rates of diabetes, hypertension and hypothyroidism prior to the pregnancy. In contrast, control women were significantly more affected by other medical conditions not related to thyroid function (23.6% vs. 15.6%, p = 0.01), asthma and other cardiopulmonary conditions being the most frequent in this group. Chorioamnionitis was significantly more frequent in the control group (4.0% vs. 1.2%, p = 0.04), and there were otherwise no significant differences between groups regarding gestational characteristics. There were also no significant differences between groups in terms of neonatal characteristics.

Table 1.

Maternal, gestational, and neonatal cohort characteristics.

	Levothyroxine-exposed group (n = 250)	Control group (n = 500)	p value
Maternal characteristics
Maternal age (years)	29.77 ± 4.62	29.77 ± 4.62	1.00
Parity	0.78 ± 0.95	0.78 ± 0.95	1.00
Primiparous	50.4%	50.4%	1.00
Pre-pregnancy conditions
BMI (kg/m²)	27.38 ± 7.23	26.00 ± 6.18	0.02
Obesity (> 30 kg/m²)	29.2%	22.8%	0.06
Diabetes (type I or type II)	2.0%	0.4%	0.04
Hypertension	3.6%	1.2%	0.03
Preexisting hypothyroidism	64.4%	0.8%	<0.001
Other medical conditions	15.6%	23.6%	0.01
Gestational characteristics
TPL	1.2%	0.8%	0.69
PPROM	0.4%	1.0%	0.67
Abruptio placentae	1.6%	1.6%	1.00
Gestational diabetes	14.0%	12.2%	0.49
Gestational HTN	6.8%	5.8%	0.59
Pre-eclampsia	2.0%	3.2%	0.35
IUGR	1.6%	3.2%	0.20
Placental anomaly	0.8%	1.0%	1.00
Oligohydramnios	0.4%	1.6%	0.29
Polyhydramnios	1.2%	1.0%	1.00
Chorioamnionitis	1.2%	4.0%	0.04
Neonatal characteristics
Gestational age (weeks)	38.85 ± 1.31	38.85 ± 1.31	1.00
Prematurity (<37 weeks)	3.2%	3.2%	1.00
Term (≥37, <40 weeks)	62.4%	62.4%	1.00
Post-term (≥40 weeks)	34.4%	34.4%	1.00
Birth weight (g)	3333.02 ± 455.78	3316.79 ± 492.00	0.66
Umbilical cord blood pH	7.24 ± 0.07	7.21 ± 0.23	0.17
APGAR score <7 at 5 min	4.0%	3.6%	0.79
NCU transfer	10.4%	9.6%	0.73
Stillbirth	0.0%	0.2%	1.00

Data is presented as mean ± standard deviation or proportion, as appropriate.

p values were calculated from Student t-test or Mann–Whitney U test for continuous variables, and from χ² test or Fisher's exact test for dichotomous variables.

BMI: body mass index; HTN: hypertension; TPL: threatened preterm labor; PPROM: premature preterm rupture of membranes; IUGR: intra-uterine growth retardation; PPH: post-partum hemorrhage; NCU: neonatal care unit (including intermediate and intensive neonatal care). Placental anomaly included: placenta previa (n = 2), placenta accreta (n = 1) grade 3 placenta (n = 2), hypertrophic placenta (n = 1).

Results in bold represent statistically significant results.

The prevalence of the operative delivery modes and their respective indications are depicted in Table 2. No difference was observed in the prevalence of overall CD in the levothyroxine-exposed group (26.8% vs. 24.8%, p = 0.553), compared to the control group. When analyzing the indications of overall CD rate, uterine atony was more frequent in the control group compared with the levothyroxine-exposed group (7.3% vs. 0.0%, p = 0.028). Otherwise, there was no difference between groups regarding any of the operative delivery indications, whether related or unrelated to possible uterine muscle dysfunction.

Table 2.

Indications for each operative delivery mode, including overall, elective and emergency caesarean delivery, and assisted vaginal delivery.

	Levothyroxine- exposed group (n = 250)	Control group (n = 500)	p value
	n (%)	n (%)
Overall caesarean delivery	67 (26.8)	124 (24.8)	0.553
Possible uterine muscle dysfunction	56 (83.6)	103 (83.1)	0.927
Dystocia	20 (29.9)	40 (32.3)	0.732
Uterine atony	0 (0.0)	9 (7.3)	0.028
Abnormal fetal presentation	16 (23.9)	18 (14.5)	0.106
Maternal exhaustion	0 (0.0)	0 (0.0)	1.000
Post-term pregnancy	1 (1.5)	11 (8.9)	0.059
Failed vaginal delivery	1 (1.5)	8 (6.5)	0.164
Previous CD	23 (34.3)	36 (29.0)	0.450
Fetal distress	15 (22.4)	40 (32.3)	0.151
Elective caesarean delivery	33 (13.2)	51 (10.2)	0.219
Possible uterine muscle dysfunction	31 (93.9)	43 (84.3)	0.302
Abnormal fetal presentation	10 (30.3)	12 (23.5)	0.490
Previous CD	22 (66.7)	31 (60.8)	0.585
Emergent caesarean delivery in labor	34 (13.6)	73 (14.6)	0.712
Possible uterine muscle dysfunction	25 (73.5)	60 (82.2)	0.302
Dystocia	20 (58.8)	40 (54.8)	0.696
Uterine atony	0 (0.0)	8 (11.0)	0.053
Abnormal fetal presentation	6 (17.6)	6 (8.2)	0.191
Post-term pregnancy	1 (2.9)	11 (15.1)	0.098
Failed vaginal delivery	1 (2.9)	7 (9.6)	0.431
Previous CD	1 (2.9)	5 (6.8)	0.662
Fetal distress	15 (44.1)	40 (54.8)	0.304
Assisted vaginal delivery	22 (8.8)	50 (10.0)	0.599
Possible uterine muscle dysfunction	10 (45.5)	17 (34.0)	0.355
Dystocia	2 (9.1)	4 (8.0)	1.000
Uterine atony	1 (4.5)	3 (6.0)	1.000
Abnormal fetal presentation	1 (4.5)	1 (2.0)	0.521
Maternal exhaustion	1 (4.5)	5 (10.0)	0.660
Post-term pregnancy	5 (22.7)	7 (14.0)	0.493
Fetal distress	15 (68.2)	43 (86.0)	0.107

p values were calculated from χ² test or Fisher's exact test for dichotomous variables.

CD: caesarean delivery; failed vaginal delivery: failure to attempt induction/operative vaginal delivery; previous CD: previous history of caesarean delivery.

Results in bold represent statistically significant results.

The association between exposure to levothyroxine and delivery outcome was evaluated with a univariable logistic regression model, presented in Table 3. Variables with a statistically significant association with overall CD rate in the univariable analysis were then analyzed in the multivariable model (Table 4), along with other clinically-significant variables that are known to influence the delivery outcome, including advanced maternal age^3,8,12 and placental abnormalities (i.e. placenta previa, placenta accreta, etc.).^8,11 Variables that were significant for elective or emergency CD were also included in the multivariable model, as the overall CD outcome represents the combination of both CD categories. After adjustments, the odds ratios related to levothyroxine exposure showed no association with the overall CD rate, elective and emergency CD, or VAD.

Table 3.

Univariable logistic regression analysis to establish the risk association between exposure to levothyroxine during pregnancy and the delivery mode.

	Overall CD OR (CI)	Elective CD OR (CI)	Emergency CD OR (CI)	VAD OR (CI)
Variables
Levothyroxine exposure	1.1 (0.8–1.6)	1.3 (0.8–2.1)	0.9 (0.6–1.4)	0.9 (0.5–1.5)
Prematurity	1.6 (1.1–2.3)	7.9 (3.4–18.4)	0.7 (0.4–1.0)	0.7 (0.4–1.2)
Post-term delivery	1.7 (0.7–4.2)	8.4 (2.2–32.2)	0.5 (0.1–1.8)	0.7 (0.2–2.4)
AMA	1.2 (0.8–1.8)	1.5 (0.8–2.6)	1.0 (0.5–1.8)	1.3 (0.7–2.3)
Multiparity	0.6 (0.4–0.8)	3.2 (1.9–5.4)	0.1 (0.1–0.2)	0.2 (0.1–0.4)
Preexisting diabetes	4.0 (0.9–17.9)	1.3 (0.2–11.1)	4.6 (1.0–20.9)	---
Preexisting HTN	1.5 (0.5–4.4)	2.0 (0.6–7.3)	0.9 (0.2–4.1)	0.7 (0.1–5.2)
Obesity	1.8 (1.2–2.6)	1.7 (1.0–2.8)	1.6 (1.0–2.4)	0.7 (0.4–1.3)
TPT	0.5 (0.1–4.1)	---	1.0 (0.1–8.4)	---
PPROM	0.6 (0.1–5.0)	---	1.2 (0.1–10.4)	---
Abruptio placentae	0.3 (0.0–2.0)	0.7 (0.1–5.6)	0.0 (0.0-.)	0.9 (0.1–6.7)
Gestational diabetes	1.0 (0.6–1.6)	0.9 (0.5–1.8)	1.0 (0.6–1.9)	0.7 (0.3–1.6)
Gestational HTN	1.6 (0.9–3.0)	1.0 (0.4–2.5)	2.0 (0.9–4.0)	1.5 (0.6–3.5)
Pre-eclampsia	1.5 (0.6–3.7)	1.3 (0.4–4.6)	1.4 (0.5–4.3)	1.0 (0.2–4.3)
Oligohydramnios	3.6 (0.9–14.1)	1.2 (0.1–10.7)	4.6 (1.1–20.2)	1.1 (0.1–9.6)
Polyhydramnios	1.8 (0.4–7.5)	---	3.7 (0.9–15.6)	---
Abnormal placenta	1.1 (0.2–6.1)	0.8 (0.1–7.1)	1.3 (0.1–12.6)	4.4 (0.7–27.9)
Chorioamnionitis	3.2 (1.3–7.7)	1.00 (0.1–7.9)	3.1 (1.3–7.6)	1.6 (0.5–4.5)

CD: caesarean delivery; aOR: adjusted odds ratio; CI: confidence intervals; AMA: advanced maternal age (≥ 35 years old); HTN: hypertension; TPT: threatened preterm labor; PPROM: preterm premature rupture of membranes.

Abnormal placenta included: placenta previa (n = 2), placenta accreta (n = 1), grade 3 placenta (n = 2), hypertrophic placenta (n = 1).

Results in bold represent statistically significant results.

Table 4.

Multivariable logistic regression model to establish the risk association between exposure to levothyroxine during pregnancy and the delivery mode.

	Overall CD aOR (CI)	Elective CD aOR (CI)	Emergency CD aOR (CI)	VAD aOR (CI)
Significant variables
Levothyroxine exposure	1.1 (0.8–1.6)	1.3 (0.8–2.2)	1.0 (0.6–1.5)	0.9 (0.5–1.6)
Prematurity	1.7 (1.1–2.4)	8.1 (3.5–19.1)	0.7 (0.4–1.1)	0.7 (0.4–1.3)
Post-term delivery	1.6 (0.6–4.1)	12.3 (3.0–50.8)	0.5 (0.1–1.8)	0.6 (0.1–2.6)
AMA	1.2 (0.8–2.0)	1.1 (0.6–2.0)	1.5 (0.8–2.8)	2.1 (1.0–4.2)
Multiparity	0.6 (0.4–0.9)	3.4 (2.0–5.9)	0.1 (0.1–0.2)	0.2 (0.1–0.4)
Preexisting diabetes	3.7 (0.8–16.9)	0.8 (0.1–8.4)	6.4 (1.1–37.3)	---
Obesity	1.7 (1.2–2.5)	1.6 (1.0–2.7)	1.6 (1.0–2.7)	0.7 (0.3–1.2)
Oligohydramnios	3.6 (0.9–14.1)	1.2 (0.1–10.7)	4.6 (1.1–20.2)	1.1 (0.1–9.6)
Abnormal placenta	1.1 (0.2–6.1)	0.8 (0.1–7.1)	1.3 (0.1–12.6)	4.4 (0.7–27.9)
Chorioamnionitis	3.2 (1.3–7.7)	1.00 (0.1–7.9)	3.1 (1.3–7.6)	1.6 (0.5–4.5)

CD: caesarean delivery; aOR: adjusted odds ratio; CI: confidence intervals; AMA: advanced maternal age (≥ 35 years old). Abnormal placenta included: placenta previa (n = 2), placenta accreta (n = 1), grade 3 placenta (n = 2), hypertrophic placenta (n = 1). .

Table 5 presents the TSH concentration results of women in both groups at each trimester of pregnancy. There was a significant difference of TSH values between groups for each trimester of pregnancy, where women exposed to levothyroxine had higher mean concentrations pre-pregnancy and in all trimesters of pregnancy. In the first trimester of pregnancy, the 75th percentile value of TSH concentration was above 3.5 mIU/L in the levothyroxine-exposed group, meaning that the thyroid function of at least 25% of these hypothyroid women was not optimal. Respectively in the second and third trimesters of pregnancy, 70.0% and 45.5% of the exposed women presenting a TSH above the 75th percentile were also considered to have a suboptimal thyroid function. All the women from the control group had a TSH concentration value in the normal range throughout their pregnancy.

Table 5.

TSH serum concentration value distribution for each trimester of pregnancy.

TSH serum concentration (mIU/L)	Levothyroxine- exposed group	Control group	p value
Pre-pregnancy
n = available data	188	349
Mean ± SD	3.94 ± 11.13	1.73 ± 0.84	<0.001
Median (range)	2.20 (115.99)	1.67 (5.78)
75th percentile	3.39	2.16
1st trimester
n = available data	229	470
Mean ± SD	4.02 ± 9.71	1.48 ± 0.79	<0.001
Median (range)	3.14 (145.00)	1.37 (3.50)
75th percentile	4.30	1.96
2nd trimester
n = available data	200	70
Mean ± SD	2.56 ± 2.91	1.40 ± 0.82	<0.001
Median (range)	2.10 (37.80)	1.19 (3.40)
75th percentile	3.03	1.98
3rd trimester
n = available data	168	30
Mean ± SD	2.76 ± 9.81	1.26 ± 0.76	0.002
Median (range)	1.68 (127.90)	1.16 (3.00)
75th percentile	2.48	1.80

p values were calculated from Mann–Whitney U test for continuous variables.

TSH: thyroid stimulating hormone; SD: standard deviation; T1, T2, and T3: first, second, and third trimester of pregnancy.

Discussion

In our study, we found no significant evidence regarding the previously-hypothesized risk association between levothyroxine exposure during pregnancy and overall CD rates. There was no significant association between exposure to levothyroxine and VAD. Our results also show no difference between groups concerning the specific indications of operative delivery, related or not to possible uterine muscle dysfunction, except for uterine atony in the overall CD rate. We observed that women exposed to levothyroxine were significantly less affected with uterine atony compared to control women, which could be explained as a possible less-exhausting uterine contraction activity when exposed to levothyroxine due to an absolute decreased number of contractions. As noticed in previous fundamental research, levothyroxine was shown indeed to decrease uterine contraction frequency.¹² Hence, contrasting with our initial hypothesis, levothyroxine treatment may possibly decrease the occurrence of some uterine dysfunction related to underlying hypothyroidism.^3,5,12 However, whether exposure to levothyroxine in pregnancy is a proper protective factor for uterine atony during labor needs further study.

Women exposed to levothyroxine were significantly more affected by pre-existing diabetes and hypertension, and tended to have a higher BMI at the beginning of the pregnancy, which could be related to their underlying hypothyroid condition that is known to be associated with metabolic and cardiovascular diseases, in addition to weight gain.^19,20 Control women were significantly more affected by other medical conditions unrelated to their thyroid function, which is likely explained by random occurrence. Otherwise, groups were comparable. Therefore, there was no difference in terms of maternal and fetal morbidity and mortality related to levothyroxine exposure.

The potential effects of the underlying hypothyroid condition should be considered while analyzing our results, as they might not represent the effect of levothyroxine alone. Based on the results our team previously demonstrated in fundamental research, we could presume a possible additional effect of levothyroxine due to overtreatment in hypothyroid pregnant women of our population.^11,12 Some clinical studies also documented an increased risk of CD and labor induction with hyperthyroidism, where women are exposed to high thyroxine serum concentrations due to a pathological overstimulation of thyroid hormone production.^21–23 Therefore, a prolonged uterine quiescent phase in late pregnancy could be hypothesized as a cause of fetal breech presentation, in addition to an inefficient uterine contraction pattern induced by a high thyroxine concentration state, resulting in an increase of overall CD rate.^{4,12,21,24–26} However, our study cannot support this hypothesis as the prescribed dosage, the women's adherence to levothyroxine therapy and their thyroid function could not be assessed for all patients due to missing data. More studies are then needed to explore the effect of inappropriately treated hypothyroidism on delivery outcome.

Our results have also shown significantly higher TSH serum concentrations at each trimester of pregnancy in hypothyroid women compared to control women, despite their levothyroxine treatment. The TSH concentration value distribution was also analyzed, to compare the thyroid function status in both groups. According to these results, at least 25% of women from the exposed group were treated with an insufficient dosage of levothyroxine in the first trimester of pregnancy, as their 75^th percentile TSH concentration value was above the targeted threshold for hypothyroidism to be controlled. This finding suggests potential underdosing of levothyroxine in at least 25% of hypothyroid women who were treated as recommended by the local guidelines, which could have influenced the outcomes observed in our study that differ with our initial hypothesis. However, whether this difference has a clinical impact is uncertain, and the missing data regarding the thyroid function of most of the euthyroid women in our study limits our interpretation.

Our study presented several limitations, the foremost being our sample size. Other studies investigating the impact of thyroid hormone variations and thyroid diseases on perinatal outcome were conducted with larger samples, often population-wide.^3,5,21,27 For instance, Turunen et al. observed, in their population-based cohort including 16 364 hypothyroid pregnant women, a risk association between hypothyroidism treated with levothyroxine and CD with a significant odds ratio of 1.13 (95% CI 1.06 to 1.21).⁵ Several studies have observed a significantly increased prevalence of breech presentation and dystocia as indications for CD in women diagnosed with hypothyroidism (without specifying their treatment status).^3,4,21 These tendencies are not significant in our study. Hence, a lack of power due to an overestimated effect size in the sample size calculation might explain our results. However, the 250 levothyroxine-exposed hypothyroid women were selected among 3182 eligible women, between the delivery dates of the first and the last selected women for this group, representing a prevalence of 7.9% of hypothyroidism during this period, which is slightly greater than the estimated hypothyroidism prevalence in the population.¹ Moreover, obtaining such non-significant results in a smaller sample causes us to question the clinical significance of the effect of levothyroxine and of hypothyroidism on a perinatal outcome that is known to be only significant in very large samples. More studies are needed to answer this question.

Another limitation is our control group. The ideal control group would have been hypothyroid women not treated with levothyroxine. We believe that euthyroid women represented the most appropriate control group in terms of ethical and clinical issues, as the absence of levothyroxine treatment in hypothyroid pregnant women can lead to significant obstetric, neurologic and intellectual sequelae for the future newborn, particularly in the case of clinical hypothyroidism.^1,2 Also, our study presents results that seem to contradict previous fundamental research suggesting both hypothyroidism and exposure to high-dosages of levothyroxine as risk factors for operative delivery, due to altered uterine muscle function.^11,12 The inclusion of adequately and inadequately-treated hypothyroid women in the levothyroxine-exposed group might have influenced our results, and this lack of information limits the analysis of our study.

Finally, selection bias and information bias while collecting data from medical records were additional limitations of our study, as some information might not have been documented and as participants were not randomly selected.

Conclusion

Rates of CD did not differ between our cohort of hypothyroid women treated with levothyroxine and a selected and matched population of euthyroid women. These results are discordant with existing fundamental and clinical findings on the subject. Additional studies must be carried out with larger samples to determine the clinical significance of the previously-described levothyroxine effect on obstetric outcome, while paying particular attention to hypothyroid women who are not treated appropriately with levothyroxine during their pregnancy.

Footnotes

Acknowledgments

We would like to thank Audrey Hamel-Thibault for logistic assistance, Samuel Lemaire-Paquette for statistical support and Ju-Hong Lee for medical record access.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

Ethical approval for this study was obtained from the CIUSSS de l’Estrie-CHUS research ethics committee (2017–1764).

Informed consent

This study was exempt from informed consent by its retrospective nature.

Guarantor

Dr Jean-Charles Pasquier.

Contributorship

Dr Jean-Charles Pasquier and Dr Diana Oprea researched literature, and Dr Jean-Charles Pasquier conceived the study. Dr Diana Oprea was also involved in protocol development, gaining ethical approval, patient selection, and data analysis. Dr Nadine Sauvé contributed to protocol adjustments, data analysis, and interpretation of results. Dr Diana Oprea wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved its final version.

ORCID iDs

Diana Oprea

Nadine Sauvé

References

Teng

Shan

Patil-Sisodia

, et al. Hypothyroidism in pregnancy. Lancet Diabetes Endocrinol 2013 Nov; 1: 228–237.

Nazarpour

Ramezani Tehrani

Simbar

, et al. Thyroid dysfunction and pregnancy outcomes. Iran J Reprod Med 2015 Jul; 13: 387–396.

Monen

Pop

Hasaart

, et al. Increased maternal TSH and decreased maternal FT4 are associated with a higher operative delivery rate in low-risk pregnancies: a prospective cohort study. BMC Pregnancy Childbirth 2015 Dec; 15: 267.

Vissenberg

Vrijkotte

TGM

van der Post

JAM

, et al. Abnormal thyroid function parameters in the second trimester of pregnancy are associated with breech presentation at term: a nested cohort study. Eur J Obstet Gynecol Reprod Biol 2016 Apr; 199: 169–174.

Turunen

Vääräsmäki

Männistö

, et al. Pregnancy and perinatal outcome Among hypothyroid mothers: a population-based cohort study. Thyroid 2019 Jan; 29: 135–141.

Skiffington

Metcalfe

Tang

, et al. Potential impact of guidelines for the prevention of cesarean deliveries in a contemporary Canadian population. J Obstet Gynaecol Can 2020 Jan; 42(6). DOI: 10.1016/j.jogc.2019.10.010.

Liu

Rusen

Joseph

, et al. Recent trends in caesarean delivery rates and indications for caesarean delivery in Canada. J Obstet Gynaecol Can 2004 Aug; 26: 735–742.

Schuit

Kwee

Westerhuis

, et al. A clinical prediction model to assess the risk of operative delivery: prediction of operative delivery. BJOG Int J Obstet Gynaecol 2012 Jul; 119: 915–923.

Kolettis

Craigo

. Thromboprophylaxis in pregnancy. Obstet Gynecol Clin North Am 2018 Jun; 45: 389–402.

10.

Archie

Roman Ashley

. Normal & abnormal labor & delivery. In: CURRENT diagnosis & treatment, obstetrics & gynecology. 11th. New York: The McGraw Hill Companies, 2013, pp.305–320.

11.

Corriveau

Pasquier

J-C

Blouin

, et al. Chronic levothyroxine and acute T3 treatments enhance the amplitude and time course of uterine contractions in human. Am J Physiol-Endocrinol Metab 2013 Mar 1; 304: E478–E485.

12.

Corriveau

Blouin

Raiche

, et al. Levothyroxine treatment generates an abnormal uterine contractility patterns in an in vitro animal model. J Clin Transl Endocrinol 2015 Dec; 2: 144–149.

13.

Pesant

M-H

. Évaluation de la TSH maternelle en préconception chez la femme enceinte. Sherbrooke: CIUSSS de l’Estrie-CHUS, 2019, Report No.: GPR-OB-002.

14.

Zeino

Carbillon

Pharisien

, et al. Delivery outcomes of term pregnancy complicated by idiopathic polyhydramnios. J Gynecol Obstet Hum Reprod 2017 Apr; 46: 349–354.

15.

Bodner-Adler

Bodner

Pateisky

, et al. Influence of labor induction on obstetric outcomes in patients with prolonged pregnancy: a comparison between elective labor induction and spontaneous onset of labor beyond term. Wien Klin Wochenschr 2005 Apr; 117: 287–292.

16.

Michlin

Oettinger

Odeh

, et al. Maternal obesity and pregnancy outcome. Isr Med Assoc J IMAJ 2000 Jan; 2: 10–13.

17.

Kooistra

Kuppens

SMI

Hasaart

THM

, et al. High thyrotrophin levels at end term increase the risk of breech presentation: TSH and breach delivery. Clin Endocrinol (Oxf) 2010 Nov; 73: 661–665.

18.

Tanos

Toney

. Uterine scar rupture - prediction, prevention, diagnosis, and management. Best Pract Res Clin Obstet Gynaecol 2019 Aug; 59: 115–131.

19.

Cooper

Ladenson

. The thyroid gland. In: Greenspan's basic and clinical endocrinology. 9th. San Francisco: The McGraw Hill Companies, 2011, pp.163–226.

20.

Thvilum

Brandt

Almind

, et al. Type and extent of somatic morbidity before and after the diagnosis of hypothyroidism. A nationwide register study. Soares P, editor. PLoS ONE 2013 Sep 16; 8: e75789.

21.

Männistö

Mendola

Grewal

, et al. Thyroid diseases and adverse pregnancy outcomes in a contemporary US cohort. J Clin Endocrinol Metab 2013 Jul 1; 98: 2725–2733.

22.

Aggarawal

Suri

Singla

, et al. Pregnancy outcome in hyperthyroidism: a case control study. Gynecol Obstet Invest 2014; 77: 94–99.

23.

Pillar

Levy

Holcberg

, et al. Pregnancy and perinatal outcome in women with hyperthyroidism. Int J Gynecol Obstet 2010 Jan; 108: 61–64.

24.

Kuppens

SMI

Kooistra

Wijnen

, et al. Maternal thyroid function during gestation is related to breech presentation at term: high maternal TSH related to breech position. Clin Endocrinol (Oxf) 2009 Oct 15; 72: 820–824.

25.

Miazga

Reed

Tunde-Byass

, et al. Decreasing cesarean delivery rates using a trial of labour after cesarean (TOLAC) bundle. J Obstet Gynaecol Can 2020 Sep; 42: 1111–1115.

26.

Sheiner

Levy

Feinstein

, et al. Risk factors and outcome of failure to progress during the first stage of labor: a population-based study. Acta Obstet Gynecol Scand 2002 Mar; 81: 222–226.

27.

Milanesi

Brent

. Management of hypothyroidism in pregnancy. Curr Opin Endocrinol Diabetes Obes. 2011 Oct; 18: 304–309.