Iron Deficiency Is a Risk Factor for Thyroid Dysfunction During Pregnancy: A Population-Based Study in Belgium

Abstract

Background:

Iron deficiency affects thyroid hormone synthesis by impairing the activity of the heme-dependent thyroid peroxidase. The prevalence of iron deficiency is elevated particularly in pregnant women. This study aimed to investigate the effects of iron status on thyroid function in a nationally representative sample of mildly iodine-deficient pregnant women.

Methods:

The study population comprised a sample of pregnant women in Belgium during the first and third trimesters of pregnancy (n = 1241). Women were selected according to a multistage proportional-to-size stratified and clustered sampling design. Urine and blood samples were collected, and a questionnaire was completed face to face with the study nurse. Concentrations of free thyroxine (fT4), total thyroxine (T4), free triiodothyronine, thyrotropin (TSH), thyroglobulin (Tg), thyroid peroxidase antibodies, Tg antibodies, hemoglobin, serum ferritin (SF), soluble transferrin receptor, urinary iodine concentrations (UICs) were measured and body iron stores (BIS) were calculated.

Results:

Median UICs were 117 and 132 μg/L in the first and third trimesters of pregnancy, respectively (p < 0.05). The frequency of SF <15 μg/L was 6.2% in the first trimester and 39.6% in the third trimester of pregnancy (p < 0.05). UIC was a significant predictor of serum Tg concentrations (p < 0.01) but not of thyroid hormone or TSH concentrations. The frequency of fT4<percentile 10th in the third trimester of pregnancy was 24% and 14% in pregnant women with negative BIS and positive BIS, respectively (p < 0.05). SF and BIS were significant predictors of fT4 and T4 in the first trimester of pregnancy (p < 0.05). Hemoglobin was a significant predictor of fT4 in both trimesters (p < 0.01) and for T4 in the third trimester (p = 0.015).

Conclusion:

Iron deficiency, but not mild iodine deficiency, is a determinant of serum fT4 and T4 in pregnant women. Correcting iron deficiency may help to maintain optimal thyroid function, in addition to preventing anemia during pregnancy.

Introduction

Hypothyroxinemia during pregnancy has been associated with impaired cognitive function in offspring (1,2). Severe iodine deficiency is one of the main nutritional determinants of hypothyroxinemia in pregnant women. Mild iodine deficiency occurs frequently in many countries including industrialized nations (3). During pregnancy, mild iodine deficiency induces a homeostatic response to compensate for low iodine intake. However, apart from an increase in maternal and neonatal thyroid volume associated with higher thyroglobulin (Tg) concentrations, thyroid hormone and thyrotropin (TSH) concentrations are generally not affected in pregnant women and newborns (4).

Recently iron deficiency has also been identified as a determinant of thyroid function (5,6). Iron status directly affects thyroid hormone synthesis by impairing the activity of the heme-dependent thyroid peroxidase (7). Iron deficiency decreases, and iron supplements increase the efficacy of iodine supplements in iodine-deficient children (8,9). Iron deficiency is extremely frequent particularly among pregnant women, peaking at 28% and 44% in industrialized countries and developing countries, respectively (10). The public health implications arising from the effect of iron deficiency on the thyroid could be even more relevant than mild iodine deficiency considering the very high prevalence of iron deficiency in both women of childbearing age and pregnant women. During pregnancy, the requirement of both iodine and iron increases, exposing pregnant women to a higher risk of combined deficiency. Daily iron requirements are estimated at 4 and 6 mg during the second and third trimesters of pregnancy, respectively (11,12).

This study hypothesized that iron status is a nutritional determinant of thyroid function in pregnant women. Therefore we aim to investigate the effects of iron status on thyroid function in the first and third trimesters of pregnancy among a nationally representative sample of mildly iodine-deficient pregnant women in Belgium.

Subjects and Methods

Sampling

The target population comprised all pregnant women in Belgium in the first and third trimesters of pregnancy. Women were selected according to a multistage proportional-to-size stratified and clustered sampling design, with the antenatal clinic as unit of cluster. Because mild iodine deficiency is considered likely to affect thyroid function during pregnancy and because previous population-based data suggested that the prevalence of iodine deficiency is higher in the south, compared with the north of the country, the clinics were stratified by region (13). A 30-cluster survey was performed in both the north and the south of Belgium with a total of at least 20 women per individual cluster.

In each region, all antenatal clinics were ranked by province and size based on the number of deliveries over the past year. Sixty clusters of four clinics were selected using systematic sampling (although accounting for the number of past deliveries) to have sufficient replacement clinics available in case some refused to participate. Of these 60 clusters, 30 clusters were randomly selected and within each cluster the first clinic was invited to participate. In each clinic, all gynecologists were invited to participate to level out a possible gynecologist-related effect.

A questionnaire about sociodemographic and socioeconomic characteristics, smoking and alcohol consumption, thyroid diseases, and use of iodine-containing supplements was completed face to face with the study nurse. All women under current treatment with thyroid hormones or antithyroid drugs were excluded.

Body mass index (BMI) was recorded by the gynecologist during the first consultation for both first- and third-trimester pregnant women.

The study was approved by the Medical Ethics Committee of the Erasme Hospital in Brussels, and informed written consent was obtained from the pregnant women.

Assays

Maternal serum free thyroxine (fT4), free triiodothyronine (fT3), TSH, Tg, thyroid peroxidase antibodies (TPO-Ab), and thyroglobulin antibodies (Tg-Ab) were measured using a third-generation chemiluminescence immunoassay (Roche). Total thyroxine (T4) was measured by radioimmunoassay (Cisbio). Serum Tg assay was standardized against Certified Reference Material for human Tg (CRM-457). Reference values were 10.3–21.9 pmol/L for fT4, 2.8–7.1 pmol/L for fT3, 58–142 nmol/L for T4, and 0.1–4.0 mU/L for TSH as recommended (14). The reference range was 0–115 U/mL for TPO-Ab. TPO-Ab titers >34 U/mL and Tg-Ab titers >115 U/mL were considered positive.

Hypothyroxinemia and hypotriiodothyroninemia were defined as fT4 and fT3 concentrations below the 10th percentile.

Urinary iodine concentrations (UICs) were measured in duplicate by a modification of the Sandell–Kolthoff reaction using spectrophotometric detection (15). The sensitivity of the assay was 12 μg/L. The Erasme laboratory participated successfully in the Program to Ensure the Quality of Urinary Iodine Procedures, as defined by the U.S. Centers for Disease Control and Prevention (16). Urinary creatinine concentrations were determined by a colorimetric method based on the Jaffe reaction (17).

Hemoglobin and mean cell volume were measured using a Beckman Coulter ACT8 Hematology Analyzer (Beckman Coulter, Inc.). Serum ferritin (SF) and soluble transferrin receptor (sTfR) concentrations were measured by immunoassay (RAMCO).

Anemia was defined as Hb <110 g/L in the first and third trimesters (18). Severe anemia during pregnancy was defined as Hb <70 g/L. SF concentration <15 μg/L indicates iron depletion in all stages of pregnancy, and levels of <12 μg/L are associated with total depletion of iron stores (19,20). In this study, iron-deficiency anemia was defined as Hb <110 g/L and SF <15 μg/L. SF concentrations were not corrected for inflammation and, in particular, for C-reactive protein concentrations. Iron deficiency was further investigated through use of more sensitive indicators, specifically sTfR concentrations and body iron stores (BIS). Pregnant women with sTfR values more than 8.5 mg/L are said to suffer from iron deficiency (21), although there are little data on the use of sTfR values in pregnancy. BIS (mg/kg), and positive or negative values were calculated as [log(sTfR/SF) 2.8229]/0.1207 (22).

Statistical analyses

Statistical analyses were carried out using SPSS for Windows, version 26.0. (SPSS, Inc.). Normally distributed data were expressed as the mean (SD); non-normally distributed data (UIC, TSH, and Tg) were expressed as the median (25–75 percentile) and were log-transformed for further analysis. For continuous variables, differences between trimesters were tested using the Mann–Whitney test or a t-independent test. A multiple linear regression analysis was carried out with thyroid hormones, log serum TSH, and log serum Tg as dependent variables. Independent variables known or likely to be determinants of the dependent variables, gestational age, age, race/ethnicity, seasons, and log UIC/Cr, were used to build the model using the Enter method. Markers of iron status were then included separately in the model. Multiple logistic regression analysis was used to analyze the risk of hypothyroxinemia, hypotriiodothyroninemia, low T4, high TSH, and high Tg controlling for age, gestational age, and BMI.

All women undergoing treatment with thyroid hormones or antithyroid drugs, as identified from the questionnaire, were excluded from the study analysis.

A sample size calculation based on an estimated 70% prevalence of iodine deficiency (23), a 95% confidence interval, a desired precision of 5%, and a design effect of two resulted in a required sample size of 645 women in each region.

Results

Characteristics of the pregnant women

The characteristics of the 1241 pregnant women are shown in Table 1. Most pregnant women took multivitamin supplements during pregnancy. This number was even significantly higher among third-trimester pregnant women (p < 0.001). All the multivitamins prescribed for pregnant women in Belgium contained both iodine (150 μg) and iron (10 mg). In the first trimester of pregnancy, 20% of women smoked or consumed alcohol. In the third trimester, the frequency of pregnant women smoking was significantly lower (p < 0.001), but alcohol consumption was the same compared with the first trimester of pregnancy.

Table 1.

Characteristics of Pregnant Women by Trimester

	First trimester No. of subjects, 594		Third trimester No. of subjects, 647		All women No. of subjects, 1241
	Value	n	Value	n	Value	n
Age, years, mean ± SD	28.2 ± 5.1	594	28.8 ± 5.1	646	28.5 ± 5.1	1240
Gestational age, weeks, mean ± SD	9.9 ± 2.7	594	34.0 ± 3.6	647	22.5 ± 12.5	1241
BMI, kg/m²	24.4 ± 5.0	589	24.3 ± 5.1	635	24.4 ± 5.0	1224
Race/ethnicity
Caucasian, %	77.6	594	71.3	647	77.4	1241
Asian, %	1.7	594	2.8	647	2.3	1241
African (north), %	13.8	594	12.9	647	13.4	1241
African (black), %	4.3	594	4.3	647	4.3	1241
Hispanic, %	0.7	594	0.8	647	0.7	1241
Not known, %	1.8	594	1.8	647	1.8	1241
Smoking
Yes, %	21.0	594	13.7^a	647	17.2	1241
No, %	78.9	594	86.1^a	647	82.7	1241
Not known, %	0	594	0.2	647	0.01	1241
Alcohol
Yes, %	20.0	594	20.7	647	20.4	1241
No, %	78.9	594	78.8	647	78.6	1241
Not known, %	1.0	594	0.9	647	0.9	1241
Supplement use
Yes, %	72.4	594	79.4^a	647	76.1	1241
No, %	27.6	594	20.1^a	647	23.7	1241
Not known, %	0.0	594	0.5	647	0.2	1241

p < 0.001.

BMI, body mass index.

Iodine and iron status

The median UIC indicates that pregnant women were at risk of mild iodine deficiency particularly during the first trimester (Table 2). By contrast, iron deficiency was significantly more frequent during the third trimester of pregnancy (p < 0.001). The prevalence of iron deficiency varies considerably according to the biomarker use to assess iron deficiency. Based on SF concentrations below 15 μg/L and negative BIS, iron deficiency prevalence was more than 6- and 10-fold higher in the third-trimester compared with first-trimester pregnant women. The significantly higher frequency of iron deficiency in third-trimester pregnant women was associated with a significant decrease of mean Hb (p < 0.001).

Table 2.

Thyroid Function, Iodine and Iron Status of Pregnant Women by Trimester

	First trimester No. of subjects, 594		Third trimester No. of subjects, 647		All women No. of subjects, 1241
	Value	n	Value	n	Value	n
Thyroid function
TSH, mU/L	1.3 (0.8–1.9)	590	1.6 (1.0–2.2)^a	646	1.4 (0.9–2.0)	1236
fT4, pmol/L	14.3 ± 2.1	500	11.2 ± 1.6^a	646	12.6 ± 2.4	1236
fT4 < percentile 10, %, no.	0.5	3	15.6^b	101	8.4	104
fT3, pmol/L	4.5 ± 0.6	590	3.8 ± 0.5^a	645	4.2 ± 0.7	1235
fT3 < percentile 10, %, no.	0.5	3	16.9^b	109	9.1	112
T4, nmol/L	137 ± 29	598	144 ± 28^a	588	141 ± 28	1186
TPO-Ab positive, %, no.	4.9	29	3.3	21	4.0	50
Tg-Ab positive, %, no.	5.2	31	2.6^b	17	3.9	48
Tg, μg/L	20 (11 –32)	588	22 (12–38)^b	643	21 (11–35)	1231
UIC, μg/L	117 (70–189)	585	132 (74–237)^a	644	124 (73–212)	1229
UIC, μg/g Cr	104 (71–174)	585	140 (81–237)^a	644	120 (75–205)	1229
Hb, g/L	127 ± 10	590	118 ± 11^a	644	122 ± 11	1234
MCV, μm³	91.4 ± 5.5	588	92.3 ± 6.0^b	648	91.9 ± 5.8	1236
SF, μg/L	47 (29–75)	593	17 (11–28)^a	643	28 (15–50)	1236
sTfR, mg/L	3.2 (2.7–3.9)	588	4.3 (3.4–5.6)^a	646	3.7 (3.0–4.8)	1234
BIS, mg/kg body weight	8.1 (6.3–10.2)	587	3.6 (1.2–5.8)^a	643	5.9 (2.9–8.4)	1230
Prevalence of ID, %, no.
Based on SF <15 μg/L	6.2	37	39.6^a	255	23.7	292
Bases on SF <30 μg/L	25.4	593	78.2^a	643	52.9	1236
Based on sTfR >8.5 mg/L	0.5	3	6.9^a	45	3.8	48
Based on negative BIS	1.5	9	15.5^a	100	8.8	109

p < 0.001.

p < 0.05.

BIS, body iron stores; fT3, free triiodothyronine; fT4, free thyroxine; ID, iron deficiency; MCV, mean corpuscular volume; SF, serum ferritin; sTfR, soluble transferrin receptor; T4, thyroxine; Tg, thyroglobulin; Tg-Ab, thyroglobulin antibodies; TPO-Ab, thyroid peroxidase antibodies; TSH, thyrotropin; UIC, urinary iodine concentration.

Thyroid function in first- and third-trimester pregnant women

In third-trimester pregnant women, mean serum fT4 and fT3 concentrations were significantly lower and T4 higher (p < 0.001) compared with first-trimester pregnant women (Table 2). Mean serum TSH and Tg concentrations were significantly higher (p < 0.05), and hypothyroxinemia and hypotriiodothyroninemia were about 16-fold higher in the third trimester compared with the first trimester. The overall prevalence of thyroid autoimmunity was about 5% in the first trimester and lower in the third trimester, but the difference was significant only for Tg-Ab (p < 0.05).

Thyroid function and iodine status

UIC was not a predictor of serum fT4, T4, fT3, or TSH concentrations in first- and third-trimester pregnant women (Table 3). Maternal age and vitamin supplements were predictors of TSH in the first and third trimesters of pregnancy. In contrast to thyroid function, UIC was a strong predictor of serum Tg concentrations both during the first and third trimesters of pregnancy. Smoking and alcohol consumption were negatively associated with Tg concentrations, but for alcohol the association was significant only in the first trimester of pregnancy (p = 0.026).

Table 3.

Predictors of Serum Free Thyroxine, Thyroxine, Free Triiodothyronine, Thyrotropin, and Thyroglobulin in Pregnant Women in a Multiple Linear Regression Model by Trimester Including Separately the Indicators of Iron Status

Multivariate regression	First trimester			Third trimester				First trimester			Third trimester
Multivariate regression	β	p	R²	β	p	R²		β	p	R²	β	p	R²
For serum fT4		<0.0001	0.082		0.005	0.036
Maternal age	0.000	0.763		−0.003	0.010		BIS	0.005	0.033	0.089	0.001	0.538	0.036
Gestational age	−0.011	<0.0001		0.000	0.727		sTfR	0.000	0.995	0.082	−0.020	0.537	0.036
BMI	−0.006	<0.0001		−0.001	0.215		SF	0.057	0.010	0.092	0.009	0.639	0.036
UIC/Cr	−0.006	0.819		0.015	0.383		Hb	0.019	0.008	0.089	0.020	<0.0001	0.062
Vitamins	0.014	0.395		−0.018	0.173
Smoking	0.021	0.227		−0.043	0.004
Alcohol	0.031	0.076		−0.030	0.463
Ethnicity^a	0.016	0.065		−0.039	0.343
For serum T4		<0.0001	0.229		0.003	0.037
Maternal age	−0.376	0.027		−0.426	0.016		BIS	0.639	0.020	0.239	−0.228	0.441	0.039
Gestational age	3.561	<0.0001		−0.419	0.083		sTfR	4.624	0.502	0.229	3.943	0.477	0.038
BMI	0.214	0.197		0.285	0.112		SF	8.556	0.002	0.245	−2.292	0.514	0.039
UIC/Cr	−0.858	0.776		−0.914	0.759		Hb	0.633	0.471	0.234	2.034	0.015	0.047
Vitamins	−1.124	0.589		−1.371	0.546
Smoking	2.534	0.238		−4.149	0.105
Alcohol	5.170	0.017		−2.969	0.174
Ethnicity^a	−0.326	0.768		−3.361	0.004
For serum fT3		<0.0001	0.152		<0.0001	0.050
Maternal age	−0.018	<0.0001		−0.006	0.042		BIS	−0.006	0.214	0.157	0.006	0.166	0.054
Gestational age	−0.034	<0.0001		−0.006	0.094		sTfR	0.329	0.010	0.162	−0.002	0.984	0.050
BMI	0.004	0.199		0.012	<0.0001		SF	−0.025	0.630	0.155	0.107	0.048	0.057
UIC/Cr	0.027	0.637		0.019	0.686		Hb	0.076	<0.0001	0.188	0.043	0.001	0.068
Vitamins	0.018	0.640		0.009	0.793
Smoking	−0.119	0.003		−0.075	0.057
Alcohol	0.018	0.666		−0.030	0.379
Ethnicity	0.034	0.102		0.018	0.324
For serum TSH		0.001	0.048		0.003	0.038
Maternal age	−0.007	0.026		−0.005	0.019		BIS	−0.001	0.897	0.048	0.006	0.099	0.044
Gestational age	−0.014	0.010		0.009	0.001		sTfR	−0.241	0.045	0.055	−0.080	0.210	0.041
BMI	0.008	0.009		0.000	0.825		SF	−0.047	0.337	0.050	0.063	0.119	0.043
UIC/Cr	−0.098	0.064		−0.051	0.137		Hb	0.023	0.139	0.050	−0.002	0.830	0.039
Vitamins	−0.117	0.001		−0.059	0.024
Smoking	−0.047	0.210		0.030	0.304
Alcohol	−0.018	0.628		0.003	0.896
Ethnicity	0.017	0.388		0.021	0.126
For serum Tg		<0.0001	0.083		<0.0001	0.088
Maternal age	−0.005	0.140		0.000	0.956		BIS	−0.008	0.183	0.087	0.005	0.363	0.090
Gestational age	0.008	0.192		0.016	0.001		sTfR	0.071	0.617	0.084	−0.090	0.403	0.089
BMI	0.001	0.806		−0.002	0.589		SF	−0.081	0.156	0.087	0.051	0.456	0.090
UIC/Cr	−0.168	0.007		−0.169	0.004		Hb	−0.008	0.676	0.085	0.009	0.587	0.090
Vitamins	0.039	0.358		0.077	0.080
Smoking	−0.189	<0.0001		−0.181	<0.0001
Alcohol	−0.099	0.026		−0.060	0.158
Ethnicity	0.019	0.414		0.062	0.006

Caucasian, Asian, African, North African, and Hispanic.

Thyroid function and iron status

The frequency of low fT4 (Fig. 1A), but not low T4 (data not shown), was significantly greater in pregnant women with negative BIS (p = 0.035) or Hb <110 g/L (p = 0.012) during the third trimester of pregnancy (the data for the first trimester are not shown because the number of negative BIS or Hb <110 g/L was small). Mean fT4 was also significantly lower in third-trimester pregnant women with negative BIS (p = 0.033) or Hb <110 g/L (p = 0.001). The frequency of low fT3 (Fig. 1B) was significantly greater (p = 0.040), and mean fT3 lower (p = 0.041) in pregnant women with Hb <110 g/L but not in pregnant women with negative BIS. Serum fT4, fT3, or T4 was not significantly different in pregnant women with SF <15 μg/L compared with pregnant women with SF ≥15 μg/L.

FIG. 1.

Frequency of low fT4 and mean serum fT4 in pregnant women with N-BIS, P-BIS, and with Hb <110* and Hb ≥110 g/L at the third trimester of pregnancy (A). Frequency of low fT3 and mean serum fT3 in pregnant women with Hb <110* and Hb ≥110 g/L at the third trimester of pregnancy (B). fT3, free triiodothyronine; fT4, free thyroxine; N-BIS, negative body iron stores; P-BIS, positive body iron stores.

Figure 2 shows the relationship between serum fT4 and Hb in first- and third-trimester pregnant women. Serum fT4 was significantly and positively associated with Hb both in the first (p < 0.01) and third trimesters (p < 0.001). The same relationship was found for fT3 and Hb (data not shown).

FIG. 2.

The relationship between serum fT4 and Hb in first-trimester (A) and third-trimester pregnant women (B).

Table 3 shows the results of the multiple regression with fT4, fT3, T4, TSH, and Tg as the dependent variables and including separately the indicators of iron status after controlling for age, gestational age, BMI, UIC, vitamins, smoking, alcohol, and ethnicity.

BIS, SF, and Hb were significant predictors of serum fT4 concentrations in first-trimester pregnant women (p < 0.05), but only Hb was a predictor of fT4 in the third trimester of pregnancy (p < 0.0001). BIS (p = 0.020) and SF (p = 0.002) were predictors of serum T4 concentrations in the first trimester, and Hb was a predictor of T4 in the third trimester of pregnancy (p = 0.015).

For serum fT3 concentrations, sTfR was a predictor in the first trimester (p = 0.010) and Hb in both trimesters (p < 0.01). Finally, none of the iron status biomarkers was associated with TSH or Tg concentrations.

Table 4 shows the results of the multiple logistic regression, including hypothyroxinemia, hypotriiodothyroninemia, low T4, high TSH, and high Tg as dependent variables after including separately the indicators of iron status after controlling for age, gestational age, and BMI. In the logistic regression with Tg as a dependent variable, we include in the model UIC as well (the data for the first trimester are not shown because the number of negative BIS, Hb <110 g/L, or SF <15 μg/L was small).

Table 4.

Predictors of Hypothyroxinemia, Hypotriiodothyroninemia, Low Thyroxine, High Thyrotropin, and High Thyroglobulin in Pregnant Women in a Multiple Logistic Regression by Trimester

Multivariate regression	First trimester		Third trimester
Multivariate regression	OR [CI]	p	OR [CI]	p
Hypothyroxinemia
Maternal age	1.32 [1.01–1.72]	0.045	1.01 [0.98–1.06]	0.408
Gestational age	3.16 [0.73–13.72]	0.125	0.99 [0.93–1.05]	0.723
BMI	0.94 [0.68–1.31]	0.717	1.03 [0.99–1.07]	0.110
Positive BIS	—	—	0.53 [0.31–0.90	0.019
Hb >110 g/L	—	—	0.51 [0.32–0.82]	0.005
SF >15 μg/L	—	—	0.88 [0.57–1.38]	0.585
Hypotriiodothyroninemia
Maternal age	1.04 [0.85–1.27]	0.733	1.03 [0.99–1.07]	0.188
Gestational age	7.43 [1.31–42.25]	0.024	0.99 [0.94–1.06]	0.045
BMI	0.93 [0.69–1.25]	0.617	0.90 [0.85–0.95]	<0.0001
Positive BIS	—	—	0.56 [0.32–0.97]	0.039
Hb >110 g/L	—	—	0.59 [0.37–0.96]	0.033
SF >15 μg/L	—	—	0.55 [0.36–0.85]	0.008
Low T4
Maternal age	0.98 [0.93–1.03]	0.347	1.04 [0.98–1.11]	0.170
Gestational age	0.65 [0.59–0.72]	<0.0001	1.13 [1.02–1.25]	0.017
BMI	0.94 [0.89–0.99]	0.037	1.02 [0.96–1.09]	0.628
Positive BIS	—	0.974	0.70 [0.31–1.60]	0.398
Hb >110 g/L	—	0.751	0.68 [0.33–1.42]	0.307
SF >15 μg/L	—	0.095	1.15 [0.58–2.29]	0.685
High TSH
Maternal age	0.99 [0.92–1.06]	0.709	0.98 [0.93–1.03]	0.372
Gestational age	0.90 [0.80–1.02]	0.107	1.16 [1.08–1.26]	<0.0001
BMI	1.04 [0.98–1.11]	0.193	0.97 [0.92–1.02]	0.272
Positive BIS	—	—	1.97 [0.90–4.32]	0.088
Hb >110 g/L	—	—	1.77 [0.89–3.48]	0.101
SF >15 μg/L	—	—	1.47 [0.87–2.49]	0.150
High Tg
Maternal age	0.98 [0.94–1.00]	0.051	0.99 [0.95–1.02]	0.374
Gestational age	0.99 [0.93–1.05]	0.758	1.07 [1.03–1.12]	0.002
BMI	1.00 [0.97–1.04]	0.735	0.97 [0.94–1.00]	0.118
UIC/Cr	0.42 [0.23–0.76]	0.004	0.30 [0.18–0.52]	<0.0001
Positive BIS	—	—	0.72 [0.46–1.15]	0.173
Hb >110 g/L	—	—	1.03 [0.69–1.53]	0.881
SF >15 μg/L	—	—	0.75 [0.53–1.06]	0.099

Low T4, T4 below the percentile 10. High TSH, TSH above the percentile 10. High Tg, Tg above the percentile 50. The data for the first trimester are not shown because the number of negative BIS or Hb <110 g/L was small.

[CI], 95% confidence interval.

The risk hypothyroxinemia was significantly lower in pregnant women with positive BIS than in pregnant women with negative BIS (p = 0.019). Pregnant women with Hb >110 g/L had also a lower risk of hypothyroxinemia than women with Hb <110 g/L (p = 0.005). The risk of hypotriiodothyroninemia was also lower in pregnant women with positive BIS, Hb >110 g/L, and also for pregnant women with SF >15 μg/L. The risk of low T4, high TSH, and high Tg was not significantly modified by iron status.

Discussion

In this population-based study, we demonstrate that iron deficiency, but not mild iodine deficiency, is a significant determinant of fT4 and T4 in pregnant women. These findings have important implications since iron deficiency is extremely frequent and because hypothyroxinemia during pregnancy can adversely affect the neurodevelopment of offspring (1,2). Our results are in line with previous studies reporting a higher risk of hypothyroxinemia, not just in iron-deficient pregnant women but also in the general population (6,24 –26). To our knowledge, this is the first population-based study showing such an association in pregnant women in a mildly iodine-deficient country. In a previous study carried out in Switzerland among second- and third-trimester pregnant women, iron status was found to be a determinant of serum T4 and TSH (6). In our population, fT4 and T4 were associated with BIS and SF in first-trimester pregnant women unlike serum TSH. These differences may be explained by a higher prevalence of iron deficiency in pregnant women from Switzerland than in Belgium. The frequency of negative BIS was 9% and 28%, respectively, in Belgium and Switzerland.

There are few other regional or hospital-based studies on iron status and thyroid function in pregnancy. Studies from two iodine-sufficient regions in China reported a significant association between iron status and fT4, but not with TSH in first-trimester pregnant women and in women of childbearing age (24,25). The prevalence of iron deficiency (ID) in China was low, only 3% of pregnant women were iron deficient. Our findings not only corroborate the results from China but also suggest that iron deficiency is a determinant of total T4 in the first trimester of pregnancy. In our study, Hb was associated with fT4 and fT3 during both the first and third trimesters. Our results indicate that anemia or iron deficiency is associated with hypothyroxinemia. In fact, during pregnancy, iron deficiency is the main nutritional determinant of Hb. Therefore, anemia can be considered in pregnant women a surrogate of severe iron deficiency. The effect of anemia on fT4 is thus secondary to the decreased heme-dependent thyroid peroxidase activity by the insufficient iron dietary supply and not due to a direct effect of low Hb on fT4.

In Belgium, a hospital-based study showed a small but significant negative correlation of SF with TSH and a positive correlation with fT4 during early pregnancy (27). The median SF in these pregnant women was 10 μg/L much lower than in our population-based study. Taken together, these data suggest that iron deficiency may affect both serum fT4 and also T4 concentrations in pregnant women.

The effect of iron status on thyroid function is not just limited to pregnant women but also affects the general population, including women of childbearing age (25). In a recent national survey of adults in a Spanish population, a significant association was found between ID and hypothyroxinemia and hypotriiodothyroninemia, but not with serum TSH (26).

In Belgium, pregnant women, particularly during the first trimester of pregnancy, are at risk of mild iodine deficiency. During the third trimester, iodine status improves because most pregnant women start taking iodine-containing multivitamins (150 μg iodine) when pregnancy is confirmed (3,28). Unlike mild iodine deficiency, the frequency of iron deficiency is relatively low in first-trimester pregnant women but increases significantly with gestational age, despite iron supplementation (14–28 mg). Two-thirds of pregnant women in Belgium take iodine-iron-containing multivitamins by the end of pregnancy although in contrast to iodine status, iron status was worse in the third-trimester compared with first-trimester pregnant women.

Interestingly, despite the fact that iron deficiency was more frequent in the third trimester of pregnancy, iron deficiency was not a determinant of fT4 concentrations in third-trimester pregnant women. This finding may be explained by the presence of other determinants of fT4 during the third trimester, which exert greater effect on fT4 concentrations than ID. Among those, there is the well-known decrease of fT4 concentrations in the third trimester, even in healthy pregnant women, attributed to poor analytic performances of fT4 immunoassays (29). In addition, physiologic hemodilution during the third trimester may also override the effect of iron deficiency on fT4 (29).

The effect of iron deficiency on fT4 and T4 in first-trimester pregnant women is of concern, given the critical role of thyroid hormones in the neurodevelopment of the fetus during the first trimester of pregnancy. At a population level, the lowering effect of iron deficiency on fT4 and T4 may affect an important number of pregnant women. In our population, 6% of first-trimester pregnant women had an SF <15 μg/L indicating iron deficiency, and 25% had an SF <30 μg/L, the cutoff for iron deficiency in the general population indicating suboptimal iron stores. At a European level, the prevalence of women of childbearing age with suboptimal BIS (SF <30 μg/L) affects 40–55% of women, while the prevalence of iron deficiency is estimated at 24–85% (30). Consequently, the prevalence of women with prepregnancy inadequate BIS is extremely high even in industrialized countries exposing pregnant women to a higher risk of thyroid dysfunction.

In addition to iron deficiency, iodine deficiency, if severe enough, also increases the risk of hypothyroxinemia during pregnancy (31,32). The potentially adverse consequences of combined iron and iodine deficiency on fetal neurodevelopment must be considered, given the high frequency of both micronutrient deficiencies worldwide. In our population of pregnant women, iodine status was a determinant of Tg concentrations, but not of thyroid hormone or TSH concentrations. This situation is frequently observed in mildly iodine-deficient regions such as in Belgium. Iodine supplements taken during pregnancy by mildly iodine-deficient pregnant women significantly decreased Tg concentrations in mothers and in newborns, but did not modify thyroid hormone concentrations (4). Consequently, it is unlikely that mild iodine deficiency will increase the risk of hypothyroxinemia in our population of iron-deficient pregnant women. However, we have considered this eventuality (data not shown). We found that the frequency of hypothyroxinemia was no different in pregnant women with combined iodine and iron deficiency (grouped in quartiles according to UIC and SF) when compared with women with iron deficiency alone. The situation can be different in countries where iron and iodine deficiencies are more severe. The combined effect of both deficiencies could increase the risk of severe hypothyroxinemia during pregnancy and thus has an adverse effect on the development of fetus.

Iron deficiency during perinatal and early life may impair the optimal neurodevelopment of children (33,34). In addition, experimental rodent models of iron deficiency have shown that maternal hypothyroxinemia decreases neonatal triiodothyronine (T3) concentrations in the brain (35). The lowering of neonatal brain T3 concentrations was associated with a reduction of thyroid-responsive genes in the neonatal cerebral cortex and hippocampus (36). In addition, iron-deficient pups showed retarded sensorimotor skills. Interestingly, the occurrence of hypothyroxinemia precedes brain iron depletion, suggesting that brain and Hb synthesis is protected to some extent from iron deficiency at the expense of iron-containing enzymes such as TPO.

One strength of the present study is that it is a population-based study. In addition, and unlike previous studies, we also measured total T4 and fT3. The study design and the finding that iron deficiency was a determinant of fT4 and also of T4 reinforce the potential effects that iron status has on thyroid function.

The main limitation of the study is that pregnant women taking iodine–iron supplements were not excluded. The frequency of pregnant women taking supplements was high in the first trimester and even higher in the third trimester. The inclusion of pregnant women taking iron supplements may have underestimated the effect of iron deficiency on thyroid hormone concentrations. However, the fact that iron deficiency increases significantly in the third trimester of pregnancy suggests that the effect of iron supplementation has little impact on pregnant women's iron status and is unlikely to affect thyroid function.

In conclusion, this is the first population-based study suggesting that iron deficiency, unlike mild iodine deficiency, is a determinant of fT4 and T4 in pregnant women. Our results support the correction of iron deficiency before pregnancy or during early gestation, not just to avoid anemia but also to maintain optimal thyroid function and normal fetal development. Given the high prevalence of pregnant women and women of childbearing age with iron deficiency, our findings require urgent further investigation. The public health implications of iron deficiency on thyroid function may be even more relevant for pregnant women living in regions where the risk of severe iron and iodine deficiency is greater than in industrialized countries.

Footnotes

Acknowledgments

We thank the hospitals and gynecologists who agreed to participate and all the participating pregnant women.

Authors' Contributions

R.M.-R. and S.V. designed and coordinated the research. R.M.-R., B.C., and S.V. wrote the article. C.D., F.W., and C.F.P. contributed to data interpretation of the results and critically revised the draft versions of the article.

Disclaimer

The article has been submitted solely to this journal and is not published, in press, or submitted elsewhere.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by funding from the Belgian Federal Public Service of Health, Food Chain Safety and Environment.

References

Finken

MJJ

, van Eijsden

, Loomans

, et al. 2013. Maternal hypothyroxinemia in early pregnancy predicts reduced performance in reaction time tests in 5- to 6-year-old offspring. J Clin Endocrinol Metab, 98:1417–1426.

Henrichs

, Bongers-Schokking

, Schenk

, et al. 2010. Maternal thyroid function during early pregnancy and cognitive functioning in early childhood: the generation R study. J Clin Endocrinol Metab, 95:4227–4234.

Vandevijvere

, Amsalkhir

, Mourri

, et al. 2013. Iodine deficiency among Belgian pregnant women not fully corrected by iodine-containing multivitamins: a national cross-sectional survey. Br J Nutr, 109:2276–2284.

Glinoer

, De Nayer

, Delange

, et al. 1995. A randomized trial for the treatment of mild iodine deficiency during pregnancy: maternal and neonatal effects. J Clin Endocrinol Metab, 80:258–269.

Beard

, Borel

, Derr

. 1990. Impaired thermoregulation and thyroid function in iron-deficiency anemia. Am J Clin Nutr, 52:813–819.

Zimmermann

, Burgi

, Hurrell

. 2007. Iron deficiency predicts poor maternal thyroid status during pregnancy. J Clin Endocrinol Metab, 92:3436–3440.

Hess

, Zimmermann

, Arnold

, et al. 2002. Iron deficiency anemia reduces thyroid peroxidase activity in rats. J Nutr, 132:1951–1955.

Hess

, Zimmermann

, Adou

, et al. 2002. Treatment of iron deficiency in goitrous children improves the efficacy of iodized salt in Côte d'Ivoire. Am J Clin Nutr, 75:743–748.

Zimmermann

, Adou

, Torresani

, et al. 2000. Persistence of goiter despite oral iodine supplementation in goitrous children with iron deficiency anemia in Côte d'Ivoire. Am J Clin Nutr, 71:88–93.

10.

Vandevijvere

, Amsalkhir

, Van Oyen

, et al. 2013. Iron status and its determinants in a nationally representative sample of pregnant women. J Acad Nutr Diet, 113:659–666.

11.

Bothwell

. 2000. Iron requirements in pregnancy and strategies to meet them. Am. J Clin Nutr, 72:257S–264S.

12.

Milman

. 2006. Iron and pregnancy—a delicate balance. Ann Hematol, 85:559–565.

13.

Vandevijvere

, Dramaix

, Moreno-Reyes

. 2011. Does a small difference in iodine status among children in two regions of Belgium translate into a different prevalence of thyroid nodular diseases in adults?. Eur J Nutr, 51:477–482.

14.

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, 27:315–389.

15.

Pino

, Fang

, Braverman

. 1996. Ammonium persulfate: a safe alternative oxidizing reagent for measuring urinary iodine. Clin Chem, 42:239–243.

16.

Caldwell

, Makhmudov

, Jones

, et al. 2005. EQUIP: a worldwide program to ensure the quality of urinary iodine procedures. Accreditation Qual Assur, 10:356–361.

17.

Hervey

. 1953. Determination of creatinine by the Jaffe reaction. Nature, 171:1125.

18.

Centers for Disease Control (CDC). 1989. CDC criteria for anemia in children and childbearing-aged women. Morb Mortal Wkly Rep, 38:400–404.

19.

Milman

. 2006. Iron and pregnancy—a delicate balance. Ann Hematol, 85:559–565.

20.

Milman

, Pedersen

, Visfeldt

. 1983. Serum ferritin in healthy Danes: relation to marrow haemosiderin iron stores. Dan Med Bull, 30:115–120.

21.

Akesson

, Bjellerup

, Berglund

, et al. 1998. Serum transferrin receptor: a specific marker of iron deficiency in pregnancy. Am J Clin Nutr, 68:1241–1246.

22.

Cook

, Flowers

, Skikne

. 2003. The quantitative assessment of body iron. Blood, 101:3359–3364.

23.

Delange

, Van Onderbergen

, Shabana

, et al. 2000. Silent iodine prophylaxis in Western Europe only partly corrects iodine deficiency; the case of Belgium. Eur J Endocrinol, 143:189–196.

24.

Teng

, Shan

, Li

, Yu

, et al. 2018. Iron deficiency may predict greater risk for hypothyroxinemia: a retrospective cohort study of pregnant women in China. Thyroid, 28:968–975.

25.

, Shan

, Li

, et al. 2015. Iron deficiency, an independent risk factor for isolated hypothyroxinemia in pregnant and nonpregnant women of childbearing age in China. J Clin Endocrinol Metab, 100:1594–1601.

26.

Maldonado-Araque

, Valdés

, Lago-Sampedro

, et al. 2018. Iron deficiency is associated with Hypothyroxinemia and Hypotriiodothyroninemia in the Spanish general adult population: di@bet.es study. Sci Rep, 8:6571.

27.

Veltri

, Decaillet

, Kleynen

, et al. 2016. Prevalence of thyroid autoimmunity and dysfunction in women with iron deficiency during early pregnancy: is it altered?. Eur J Endocrinol, 175:191–199.

28.

Moreno-Reyes

, Glinoer

, Van Oyen

, et al. 2013. High prevalence of thyroid disorders in pregnant women in a mildly iodine-deficient country: a population-based study. J Clin Endocrinol Metab, 98:3694–3701.

29.

Lee

, Spencer

, Mestman

, et al. 2009. Free T4 immunoassays are flawed during pregnancy. Am J Obstet Gynecol, 200:260–261.

30.

Milman

, Taylor

, Merkel

, Brannon

. 2017. Iron status in pregnant women and women of reproductive age in Europe. Am J Clin Nutr, 106:1655S–1662S.

31.

Glinoer

. 1997. The regulation of thyroid function in pregnancy: pathways of endocrine adaptation from physiology to pathology. Endocr Rev, 18:404–433.

32.

Vermiglio

, Lo Presti

, Scaffidi Argentina

, et al. 1995. Maternal hypothyroxinaemia during the first half of gestation in an iodine deficient area with endemic cretinism and related disorders. Clin Endocrinol (Oxf.), 42:409–415.

33.

Lozoff

, Jimenez

, Hagen

, et al. 2000. Poorer behavioral and developmental outcome more than 10 years after treatment for iron deficiency in infancy. Pediatrics, 105:E51.

34.

Lozoff

. 2007. Iron deficiency and child development. Food Nutr Bull, 28:S560–S571.

35.

, Wang

, Shan

, et al. 2016. Perinatal iron deficiency-induced hypothyroxinemia impairs early brain development regardless of normal iron levels in the neonatal brain. Thyroid, 26:891–900.

36.

Barks

, Fretham

SJB

, Georgieff

, et al. 2018. Early-life neuronal-specific iron deficiency alters the adult mouse hippocampal transcriptome. J Nutr, 148:1521–1528.