Effects of Maternal Iodine Nutrition and Thyroid Status on Cognitive Development in Offspring: A Pilot Study

Abstract

Background and Objective:

Maternal iodine nutrition and thyroid status may influence neurocognitive development in offspring. This study investigated the effects on the intelligence quotient (IQ) of children born to mothers with different levels of iodine supplementation, with or without the administration of levothyroxine (LT4), prior to and during pregnancy.

Patients and Methods:

This pilot, prospective, observational study included four study groups, each comprising 15 mother–child pairs, identified on the basis of maternal histories of iodized salt consumption and LT4 treatment prior to and during pregnancy. The groups were labeled as follows: iodine (I), no iodine (no-I), iodine + LT4 (I + T4), and no iodine + LT4 (no-I + T4). IQ tests were administered to children at 6–12 years of age with the Wechsler Intelligence Scale for Children–3rd Edition (WISC-III), with full-scale IQ (FSIQ), verbal IQ (VIQ), and performance IQ (PIQ) being evaluated.

Results:

Children of I and I + T4 mothers had similar verbal, performance, and FSIQs, which were 14, 10, and 13 points higher, respectively, than children born to no-I and no-I + T4 mothers. A positive association was found between VIQ and maternal urinary iodine (β = 1.023 [confidence interval (CI) 1.003–1.043]; p = 0.028), but not with maternal free thyroxine concentrations at any stage of pregnancy. Overall, the prevalence of borderline or defective cognitive function was more than threefold higher in the children of mothers not using iodized salt than of those mothers using it (76.9% vs. 23.1%, odds ratio 7.667 [CI 2.365–24.856], χ² = 12.65; p = 0.0001).

Conclusions:

Neuro-intellectual outcomes in children appear to be more dependent on their mothers' nutritional iodine status than on maternal thyroid function. These results support the growing body of evidence that prenatal, mild-to-moderate iodine deficiency adversely affects cognitive development later in life, with a seemingly greater impact on verbal abilities.

Introduction

Iodine is required for the synthesis of thyroid hormones (TH), which are essential for several human neurodevelopmental events during fetal and early postnatal life. Prior to the onset of fetal thyroid function, the mother is the only source of TH for the developing brain. By 16–20 weeks post-conception, the fetal thyroid gland is sufficiently mature, and both the fetus and the mother contribute to the regulation of fetal TH-dependent processes within the brain (1). Therefore, maternal and fetal thyroid insufficiency, occurring independently or concomitantly, may be responsible for brain damage, with severity dependent on fetal age as well as the magnitude and length of exposure of the developing brain to the hormonal deficiency. The most serious consequence of intrauterine TH deprivation is endemic cretinism due to severe gestational iodine deficiency (ID), which impairs TH synthesis in the mother and fetus from early pregnancy onwards (2). In addition to cretinism, minor neuropsychiatric and intellectual deficits have been described in infants and schoolchildren born to mothers exposed to mild to moderate ID, which may result in mild maternal thyroid function impairment (3 –6).

Several studies involving mildly iodine-deficient and iodine-sufficient populations have shown a strong association between either maternal subclinical hypothyroidism (7 –9) or hypothyroxinemia (10 –16) and impaired neuro-intellectual outcomes in progeny. However, other studies have failed to confirm these associations (17 –20), and some anecdotal reports have shown normal neurodevelopment in children born to severely hypothyroid mothers whose iodine intake during gestation was more than adequate (21 –23). The only randomized clinical trial to date that has investigated the effects of levothyroxine (LT4) given at a median gestational age of 13 weeks to mothers with mild gestational thyroid insufficiency (i.e., subclinical hypothyroidism and isolated hypothyroxinemia) and exposed to mild-to-moderate ID during pregnancy did not show improved cognitive outcomes in their offspring (24). Moreover, in recent years, a growing body of clinical evidence has emerged suggesting that gestational iodine supplementation, while improving infant cognitive development, does not have a clear impact on maternal thyroid function (25 –28).

Taken together, these data suggest that maternal iodine status during gestation may play a more critical role than maternal thyroid function in determining neuro-intellectual outcomes in progeny. To assess this hypothesis, a pilot prospective study was conducted to observe the long-term effects of different maternal iodine supplementation regimens with and without LT4 prior to and during pregnancy on the cognitive function of school-age children living in an ID area where endemic cretinism (29), endemic cognitive deficiency (30), and attention deficit and hyperactivity disorders (ADHD) (14) had previously been described.

Methods

Study design and participants

This study was observational in design and included 60 mother–child pairs from a small rural ID area in northeastern Sicily. Mean urinary iodine excretion (UIE) and goiter prevalence in local schoolchildren at conception (2000–2007) and at the time of cognitive evaluation (2012–2013) of the subjects were consistent with moderate (UIE 62.2 ± 38.2 μg/L; goiter in schoolchildren 16.3%) (14) and mild (UIE 95.6 ± 25.8 μg/L; goiter in schoolchildren 7.6%; unpublished data) ID, respectively.

The objective of this study was to assess global cognitive outcomes in children born to mothers under different conditions of iodine nutrition during gestation, either treated with LT4 or receiving no LT4 prior to and during pregnancy.

Between January 2000 and January 2007, 351 pregnant women (gestational age <12 weeks) were referred to the authors' antenatal clinic for the prevention and control of gestational thyroid disorders. At the time of enrollment, all women were administered a structured questionnaire about their dietary habits, specifically focusing on the regular use of iodized salt. They were then sampled for free triiodothyronine (fT3), free thyroxine (fT4), and thyrotropin (TSH), and were scheduled to have their thyroid function serially reassessed at six-week intervals throughout pregnancy. Serum antithyroglobulin (TgAb) and antithyroperoxidase (TPOAb) antibodies were also measured at initial and final sampling. With few exceptions, the women provided urine samples at every appointment. Adjustments to daily LT4 doses were made when needed in the women receiving semi-suppressive LT4 therapy for uni- or multinodular goiter (31) in order to maintain preconception TSH levels (0.1–0.5 mIU/L). All data were recorded in a specific database.

As shown in Figure 1, the initial population consisted broadly of two groups: the first included 201/351 (57.3%) women who reported using iodized salt prior to becoming pregnant; the second comprised 150/351 (42.7%) women who had never used iodized salt prior to conception. Women who had not used iodized salt regularly or had not used it for at least two years (32) (n = 64/201) were excluded from the former group; women who had started using iodized salt upon becoming pregnant (n = 24/150) were excluded from the latter group. The groups were further characterized according to whether mothers had uninterruptedly received LT4 treatment for uni- or multinodular goiters prior to and during pregnancy. The four groups identified were labeled iodine (I; n = 96 women), iodine + LT4 (I + T4; n = 41 women), no iodine (no-I; n = 88 women), and no iodine + LT4 (no-I + T4; n = 38 women). Since this was a pilot study, the first 15 mother–child pairs from each group who met the inclusion criteria were selected to make up the study groups. The size of 60 mother–child pairs (15 for each group) corresponds to 25% of the final study size of approximately 240 subjects, which was calculated by taking into account an expected difference of a mean IQ score of 13.5 points (26) between two main groups (I vs. no-I). Assuming α = 0.050, we estimated that 119 subjects per group would be needed in the final study in order to obtain a power level of 0.80.

FIG. 1.

Study population and design.

Inclusion criteria for mothers were: (i) age >18 years; (ii) uncomplicated singleton pregnancy; (iii) term delivery; (iv) no severe or chronic diseases (including thyroid autoimmune diseases); (v) no major postpartum complications or depression; (vi) TH determination throughout gestation (at least one testing at the 1st trimester and two at each subsequent trimester); (vii) full disclosure of diet and lifestyle information during and after pregnancy; and (viii) informed consent. Inclusion criteria for children were: (i) age 6–14 years; (ii) no major neonatal complications; (iii) no congenital hypothyroidism; (iv) no severe/chronic diseases; (v) no major cognitive deficits; (vi) regular school attendance; and (vii) approval for cognitive test administration.

Mother–child pairs included in the study (n = 60) did not differ from those not included (n = 291) in any of the variables reported in Table 1, except for the M/F ratio, which was significantly lower in the I and no-I groups compared with the whole sample (0.25 vs. 0.94, p = 0.031).

Table 1.

Clinical and Socioeconomic Characteristics of Parents and Their Children

	I	no-I	I + T4	no-I + T4	p-Values
Parent data
Maternal age at delivery, years
M ± SD	29.1 ± 4.7	27.7 ± 4.3	30.1 ± 4.3	28.1 ± 2.9	NS
Median (IQR)	31 (27–32)	28 (25–30.5)	30 (27.5–33)	28 (26.5–30.5)
Range	22–34	19–34	22–36	22–33
Parity
Nulliparous (%)	40	33.3	35.7	26.6	NS
≥1 pregnancy (%)	60	66.6	64.3	73.4
Obstetrical delivery
Vaginal delivery (%)	66.6	73.3	71.4	80	NS
Cesarean section(%)	33.3	26.7	28.6	20
Adjusted maternal urinary iodine, μg/L
Median	90.1	48.1	120.2	72	I vs. no-I 0.00046; I vs. I + T4 0.0005
(IQR)	(82.8–100.5)	(43–60)	(112–135.7)	(61.5–87.1)	I vs. no-I + T4 0.002; no-I vs. I + T4 0.00005
Range	70–104	32.9–76.1	106–152.6	54–99.8	no-I vs. no-I + T4 NS; I + T4 vs. no-I + T4 0.00005
LT4 dose, μg/day
Preconception (M ± SD)	n.a.	n.a.	102.7 ± 24.1	107.1 ± 24.8	NS
Gestational (M ± SD)	n.a.	n.a.	160.7 ± 37.3	150.1 ± 41.7
SES
M ± SD	18.5 ± 7.4	16.8 ± 5.6	19.8 ± 6.5	21.7 ± 10.6	NS
Cohabiting parents
Yes (%)	93.3	93.3	92.9	100
No (%)	6.7	6.7	7.1	0	NS
Child data
Sex
F (%)	80	80	35.7	60	I + T4 vs. I < 0.05; I + T4 vs. no-I <0.05
M (%)	20	20	64.3	40
Gestational age at birth, weeks
M ± SD	38.6 ± 1.3	38.6 ± 1.1	38.7 ± 1.4	38.8 ± 1.1	NS
Median	38	39	38	38
(IQR)	(37–39.5)	(38–39.5)	(37.5–39.5)	(38–39)
Range	37–41	37–41	37–41	37–40
Birth weight, g
M ± SD	3250 ± 280	3239 ± 267	3276 ± 265	2945 ± 298	NS
Median	3186	3131	3165	3316
(IQR)	(3000–3445)	(3007–3435)	(3065–3540)	(3120–3417)
Range	2850–3920	2750–3920	2570–3910	2975–3780
Age at cognitive testing, years
M ± SD	9.8 ± 2.1	9.5 ± 2.3	9.4 ± 2.4	9.7 ± 2.6	NS
Median (IQR)	10 (8.2–11.7)	9 (7.6–11.7)	9 (7.3–11.8)	10 (8.1–11.6)
Range	6.8–12.5	6.4–13	6.4–13	7.3–13
Urinary iodine, μg/L
Median	97	105	98	106	NS
(IQR)	(88–119)	(84–125)	(85.5–121)	(89–149.5)
Range	70–145	68–190	75–155	82–185

Data are expressed as M ± SD, median (IQR), and range for continuous variables, and as percentages for categorical variables. p-Values were calculated using the Kruskall–Wallis test (to compare all groups) and the Mann–Whitney U-test (to perform the two-by-two comparison between groups) for continuous variables and χ² for categorical variables.

I, iodine group; no-I, no iodine group; I + T4, iodine + levothyroxine group; no-I + T4, no iodine + levothyroxine group; NS, not significant; IQR, interquartile range; SES, socioeconomic status; n.a., not applicable.

Procedures

Cognitive outcome

Each child's intelligence quotient (IQ) was assessed using the Wechsler Intelligence Scale for Children—Third Edition (WISC-III) (33,34), which was administered blind by trained psychologists. Three IQ scores (the full-scale IQ [FSIQ], the Verbal IQ [VIQ], and the Performance IQ [PIQ]), all with a mean of 100 and standard deviation of 15, were calculated. Full-scale IQ scores of 70–85 and <70 were considered to be indicative of borderline or defective cognitive function, respectively.

Biochemical evaluation

Serum concentrations of TSH, fT4, fT3, TgAb, and TPOAb were measured using the electrochemiluminescence immunoassay (ECLIA) system assay (Cobas; Roche Diagnostics, Mannheim, Germany). The detection limits of the assays were as follows: 0.005 mIU/L for TSH, 0.30 pmol/L for fT4, 0.39 pg/mL for fT3, 5 IU/mL for TPOAb, and 10 IU/mL for TgAb. The manufacturer's reference ranges were 0.27–4.20 mIU/L for TSH, 12–22 pmol/L for fT4, and 2–4.4 pg/mL for fT3. TPOAb and TgAb were considered positive for values >34 IU/mL and >115 IU/mL, respectively. Precision profiles showed inter- and intra-assay coefficients of variation <5% over the entire measurement range. UIE, expressed as μg/L, was measured using the modified Sandell–Kolthoff reaction (35) in random urine samples collected from the mothers throughout gestation and from the children at the time of cognitive evaluation. Since several urine samples were collected from each mother (M ± SD = 5.9 ± 0.9; range 4–7), adjusted mean UICs accounting for day-to-day variations in each subject were calculated (36) in order to estimate the overall median UIE for each study group. The study was approved by the local ethics committee.

Statistical analysis

Numerical data are expressed as means ± standard deviations (M ± SD), medians, and interquartile ranges (IQR), and categorical variables as numbers and percentages.

Although most variables under examination were normally distributed (Kolmogorov–Smirnov test), because of our sample size, the nonparametric approach was used to ensure valid asymptotic results. Accordingly, the Mann–Whitney (comparison of independent samples), Wilcoxon (comparison of dependent samples), and Friedman (comparison over the observation period of several dependent groups) tests were used. The chi-square test was used to assess the association between categorical variables. Finally, logistic regression models were used to assess the dependence of suboptimum cognitive outcomes (IQ scores <85, i.e., −1 SD) on various explanatory variables and confounders. Covariates were either continuous/discrete measures (maternal age; gestational age at birth; birth weight; child age at time of cognitive evaluation; maternal serum fT3, fT4, and TSH at the various points during gestation; maternal UIE; and family socioeconomic status [SES] evaluated by means of the Hollingshead Index) (37,38), or categorical variables (maternal parity [uniparous vs. pluriparous]; sex; breastfeeding [some vs. no]; major maternal stressful life-events [yes vs. no]; and maternal/paternal education [low, medium, high]). Statistical analyses were performed using SPSS Statistics for Windows v17.0 (SPSS, Inc., Chicago, IL). A p-value of <0.05 was considered to be statistically significant.

Results

Although all children successfully completed the tests administered, one result was excluded from analysis because the mother of the child concerned had helped him complete the test. Therefore, the results refer to 59 mother–child pairs (14 for the I + T4 group and 15 for each of the other groups).

No differences in epidemiological, socioeconomic, and clinical parameters were observed between the four study groups, except for child sex distributions and maternal UIE (Table 1). The latter, calculated on a total of 88 (I group), 83 (no-I group), 89 (I + T4 group), and 90 (no-I + T4 group) samples, respectively, was significantly higher in women who regularly used iodized salt (I and I + T4 groups) than it was in women who had never used it (no-I and no-I + T4 groups).

Changes in maternal thyroid function parameters over the course of pregnancy are reported in Table 2. Mothers in the I group displayed serum fT4 concentrations that were always significantly higher than that of those in the no-I group. The cognitive performances of children born to mothers in these two groups also differed, with significantly higher FSIQ and VIQ scores in the I than in the no-I offspring (FSIQ: I 93.1 ± 11.8 vs. no-I 81.7 ± 13.5; VIQ: I 90.1 ± 14.2 vs. no-I 80.3 ± 12.3; p = 0.028 and 0.034, respectively; Fig. 2A). Similar differences were found between I + T4 and no-I + T4 offspring (FSIQ: I + T4 96.1 ± 18.2 vs. no-I + T4 81.3 ± 13.1; VIQ: I + T4 97.2 ± 18.6 vs. no-I + T4 79.6 ± 13.9; p = 0.038 and 0.017, respectively), despite their mothers having had similar serum fT4 concentrations during pregnancy (Fig. 2B). Table 3 shows full comparisons between groups. Interestingly, there were no statistically significant differences in any of the three quotients between children born to supplemented mothers (i.e., I group vs. I + T4 group) and those born to mothers not using iodized salt (i.e., no-I group vs. no-I + T4 group). Accordingly, children were grouped in relation to their mothers' nutritional iodine status and irrespective of LT4 treatment to form the I ± T4 group (including 29 children: 15 from the I group and 14 from the I + T4 group) and the no-I ± T4 group (including 30 children: 15 from the no-I group and 15 from the no-I + T4 group). Mothers in the I ± T4 group displayed higher serum fT4 concentrations than mothers in the no-I ± T4 group at early gestation and at 13–18 weeks (p = 0.006 and 0.0008, respectively), but not at subsequent follow-up.

FIG. 2.

Mean maternal free thyroxine concentrations (line graphs) and median (interquartile range, minimum, maximum) cognitive outcomes (box plots) in (A) iodine (I) and no iodine (no-I) and (B) iodine +levothyroxine (I + T4) and no-I + T4 groups. *p < 0.05; **p ≤ 0.003; ***p ≤ 0.001.

Table 2.

Gestational Changes in Maternal Thyroid Function Parameters

	I	no-I	I + T4	no-I + T4	p^a
fT3, pg/mL, M ± SD
≤12 weeks	3.8 ± 0.5	3.8 ± 0.4	4.2 ± 0.9	3.6 ± 0.5	I + T4 vs. no-I + T4 0.049
13–18 weeks	3.9 ± 0.6	3.9 ± 0.7	4.4 ± 1.0	3.8 ± 0.5	NS
19–24 weeks	3.7 ± 0.3	3.6 ± 0.5	3.9 ± 0.7	3.9 ± 0.6	I vs. no-I 0.026
25–30 weeks	3.8 ± 0.5	3.6 ± 0.8	3.9 ± 0.6	3.6 ± 0.5	NS
Week 31–term	3.6 ± 0.4	3.6 ± 0.7	3.8 ± 0.7	3.5 ± 0.4	NS
p-Values^b	0.021	NS	0.044	NS
fT4, pmol/L, M ± SD
≤12 weeks	17.9 ± 1.8	14.7 ± 1.8	19.1 ± 5.4	17.7 ± 2.9	I vs. no-I 0.005; no-I vs. I + T4 0.001; no-I vs. no-I + T4 0.002
13–18 weeks	16.4 ± 1.7	12.5 ± 2.1	17.4 ± 3.8	15.8 ± 3.7	I vs. no-I 0.001; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.009
19–24 weeks	14.7 ± 1.2	12.6 ± 2.3	16.4 ± 1.9	16.6 ± 2.9	I vs. no-I 0.003; I vs. I + T4 0.008; I vs. no-I + T4 0.049; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.0001
25–30 weeks	14.1 ± 1.3	12.3 ± 2.2	15.8 ± 1.9	16.5 ± 2.7	I vs. no-I 0.014; I vs. I + T4 0.034; I vs. no-I + T4 0.02; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.0001
Week 31–term	13.8 ± 1.2	12.2 ± 2.6	16.6 ± 1.5	16.4 ± 3.5	I vs. no-I 0.024; I vs. I + T4 0.0001; I vs. no-I + T4 0.021; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.001
p-Values^b	0.0001	0.004	0.001	0.0001
TSH, mIU/L, M ± SD
≤12 weeks	0.61 ± 0.35	0.82 ± 0.48	0.53 ± 0.48	0.72 ± 0.37	NS
13–18 weeks	0.77 ± 0.38	1.19 ± 0.46	0.55 ± 0.49	0.83 ± 0.61	I vs. no-I 0.017; no-I vs. I + T4 0.003; no-I vs. no-I + T4 0.036
19–24 weeks	0.98 ± 0.51	1.22 ± 0.72	0.49 ± 0.42	0.58 ± 0.32	I vs. I + T4 0.014; I vs. no-I + T4 0.033; no-I vs. I + T4 0.002; no-I vs. no-I + T4 0.003
25–30 weeks	0.99 ± 0.58	1.3 ± 0.67	0.44 ± 0.36	0.37 ± 0.27	I vs. I + T4 0.013; I vs. no-I + T4 0.003; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.001
Week 31–term	0.96 ± 0.41	1.45 ± 0.67	0.35 ± 0.32	0.33 ± 0.28	I vs. no-I 0.028; I vs. I + T4 0.001; I vs. no-I + T4 0.001; no-I vs. I + T4 0.0001; no-I vs. no-I + T4 0.0001
p-Values^b	0.001	0.001	NS	0.0001

Data are expressed as M ± SD. p-Values were calculated using ^athe Mann–Whitney U-test (comparisons between groups) and ^bthe Friedman test (comparisons within groups in different times).

Serum fT3 concentrations were significantly higher in I + T4 than in no-I + T4 and in I than in no-I women at very early and mid-gestation, respectively. Serum fT4 concentrations were significantly higher in I + T4 and no-I + T4 groups than in the no-I (at any point in time during gestation) and I groups (from week 19 to week 24 onwards). Also, serum fT4 concentrations were significantly higher in the I than in the no-I group at any point in time. Opposite changes, but similar statistical differences between the study groups, were observed in serum TSH concentrations over the study period.

fT3, free triiodothyronine; fT4, free thyroxine; TSH, thyrotropin.

Table 3.

FSIQ, VIQ, and PIQ According to Study Group

	I (n = 15)	no-I (n = 15)	I + T4 (n = 14)	no-I + T4 (n = 15)	p-Value
FSIQ
M ± SD	93.1 ± 11.8	81.7 ± 13.5	96.1 ± 18.2	81.3 ± 13.1	I vs. I + T4; no-I vs. no-I + T4 NS
CI	86.5–99.1	74.4–88.9	85.6–106.6	74.1–88.5	I vs. no-I 0.01
Median	95	82	98.5	80	I vs. no-I + T4 0.007
IQR	87–100.5	71–94.5	82.5–109.5	78–85.5	I + T4 vs. no-I 0.011
Range	67–110	59–101	66–121	55–112	I + T4 vs. no-I + T4 0.0089
VIQ
M ± SD	90.1 ± 14.2	80.3 ± 12.3	97.2 ± 18.6	79.6 ± 13.9	I vs. I + T4; no-I vs. no-I + T4 NS
CI	82.2–97.9	73.5–87.1	86.4–107.9	71.9–87.3	I vs. no-I 0.027
Median	88	80	97.5	81	I vs. no-I + T4 0.025
IQR	82.5–100.5	73–85	82.3–108	76.5–84	I + T4 vs. no-I 0.003
Range	65–115	59–109	69–129	45–112	I + T4 vs. no-I + T4 0.003
PIQ
M ± SD	98.2 ± 12.2	87.3 ± 15	97.4 ± 18.6	87.5 ± 13.2	I vs. I + T4; no-I vs. no-I + T4 NS
CI	91.4–104.9	79–95.6	87.6–108.1	80.2–94.8	I vs. no-I 0.019
Median	100	90	98	86	I vs. no-I + T4 0.014
IQR	92.5–103	77.5–100	87–111	81–94	I + T4 vs. no-I 0.031
Range	66–123	65–111	56–123	63–114	I + T4 vs. no-I + T4 0.037

Data are expressed as M ± SD; confidence intervals (CI), medians, and interquartile ranges (IQR) are also reported. p-Values were calculated using the Mann–Whitney U-test.

FSIQ, Full Scale Intelligence Quotient; VIQ, Verbal Intelligence Quotient; PIQ, Performance Intelligence Quotient; UIC, urinary iodine concentration.

FSIQ was on average 13 points higher in I ± T4 than in no-I ± T4 children (94.5 ± 15.1 vs. 81.5 ± 13.1; p = 0.00037), with the difference due to significantly better results on both verbal and performance subscales. More specifically, VIQ was almost 14 points (93.5 ± 16.6 vs. 79.6 ± 12.9; p = 0.00043) and PIQ 10 points (97.4 ± 15.9 vs. 87.4 ± 13.9; p = 0.0067) higher in I ± T4 than in no-I ± T4 children. It is noteworthy that while I ± T4 children performed similarly on verbal and performance subscales (p = 0.11), mean VIQ scores were significantly lower than PIQ scores for children from the no-I ± T4 group (p = 0.006).

Overall, nine children (3/9 born to I ± T4 women and 6/9 born to no-I ± T4 women) with FSIQs <2 SD (range 55–69 points) were diagnosed with mild mental retardation. Of the remaining 50 children, three from the I ± T4 group and 14 from the no-I ± T4 group (χ² = 10.2, p = 0.0014) exhibited borderline intellectual functioning (FSIQ: 70–84 points). Therefore, the overall prevalence of borderline or defective cognitive function was more than threefold higher in the children of mothers not using iodized salt compared with those mothers using it (76.9% vs. 23.1%, OR 7.667 [CI 2.365–24.856], χ² = 12.65; p = 0.0001).

Relationship between cognitive outcomes and possible confounders

In order to assess the possible dependency of IQs on explanatory variables and confounders, univariate logistic regression models were estimated (Table 4). These analyses showed a significant positive association of adjusted maternal UIEs with both FSIQ (exp β = 1.021 [CI 1.001–1.041]; p = 0.04) and VIQ (exp β = 1.022 [CI 1.002–1.042]; p = 0.03). After adjustment for maternal age, adjusted maternal UIE was the only factor independently associated with VIQ (exp β = 1.023 [CI 1.003–1.043]; p = 0.028).

Table 4.

Relationship Between Cognitive Outcomes and Possible Confounders (Univariate Logistic Regression Models)

	VIQ			PIQ			FSIQ
	exp β	CI	p	exp β	CI	p	exp β	CI	p
Maternal age	1.176	1.015–1.362	0.030	1.067	0.924–1.232	0.375	1.086	0.949–1.241	0.230
Maternal education	2.012	0.786–5.154	0.145	2.439	0.782–7.611	0.125	1.684	0.654–4.337	0.280
Major maternal stressful life events	1.192	0.071–20.011	0.903	0.947	0.081–11.145	0.966	0.781	0.047–13.117	0.864
Maternal parity	0.568	0.196–1.649	0.298	0.259	0.065–1.038	0.057	0.442	0.146–1.334	0.148
Paternal education	1.280	0.539–3.043	0.576	2.669	0.873–8.161	0.085	1.561	0.638–3.817	0.329
Family SES	1.004	0.940–1.073	0.905	0.994	0.925–1.069	0.880	0.980	0.916–1.048	0.548
Gestational age at birth	1.394	0.119–16.296	0.791	0.685	0.067–7.058	0.751	0.620	0.053–7.240	0.703
Sex	0.359	0.119–1.078	0.68	0.982	0.302–3.189	0.976	0.500	0.165–1.514	0.220
Breastfeeding	4.167	0.802–21.644	0.090	3.083	0.760–12.506	0.115	3.684	0.847–16.016	0.082
Child age at the time of cognitive evaluation	0.780	0.600–1.013	0.062	0.808	0.261–2.498	0.711	0.891	0.694–1.144	0.366
fT3
≤12 weeks	1.486	0.637–3.469	0.360	1.129	0.446–2.856	0.797	1.265	0.542–2.953	0.587
13–18 weeks	0.997	0.504–1.973	0.992	1.248	0.573–2.217	0.577	0.969	0.489–1.922	0.969
19–24 weeks	1.173	0.599–2.299	0.641	2.319	0.842–6.388	0.104	1.116	0.727–1.715	0.616
25–30 weeks	0.586	0.248–1.383	0.222	0.829	0.335–2.53	0.686	0.568	0.240–1.345	0.199
Week 31–term	0.651	0.263–1.651	0.355	1.553	0.574–4.201	0.385	0.707	0.289–1.733	0.449
fT4
≤12 weeks	1.192	0.980–1.450	0.079	1.053	0.881–1.259	0.569	1.103	0.928–1.311	0.268
13–18 weeks	1.217	1.015–1.459	0.068	1.010	0.855–1.194	0.902	1.165	0.980–1.385	0.084
19–24 weeks	1.234	0.991–1.537	0.060	1.125	0.897–1.411	0.308	1.155	0.937–1.425	0.177
25–30 weeks	1.137	0.928–1.393	0.214	1.003	0.810–1.243	0.975	1.031	0.847–1.254	0.763
Week 31–term	1.191	0.982–1.445	0.076	1.010	0.834–1.224	0.918	1.070	0.895–1.280	0.459
TSH
≤12 weeks	0.774	0.231–2.593	0.678	0.480	0.127–1.808	0.278	0.838	0.251–2.799	0.773
13–18 weeks	1.001	0.381–2.633	0.998	1.061	0.365–3.086	0.913	1.201	0.452–3.189	0.713
19–24 weeks	1.012	0.421–2.433	0.979	1.704	0.574–5.064	0.337	1.929	0.724–5.139	0.189
25–30 weeks	0.880	0.382–2.025	0.764	0.979	0.395–2.422	0.963	1.662	0.678–4.071	0.267
Week 31–term	0.908	0.411–2.009	0.812	0.856	0.366–2.004	0.721	1.405	0.614–3.217	0.421
Maternal UIC	1.017	1.001–1.033	0.038	1.012	0.995–1.028	0.163	1.013	0.998–1.029	0.098

Significant values are shown in bold.

Based on the high, although not statistically significant, exp β value associated with breastfeeding, the dependency of IQs on this factor in the two subgroups of mothers using and not using iodized salt was also calculated. However, no association was identified (exp β = 5.667 [CI 0.818–39.267]; p = 0.079).

Discussion

To examine associations of the iodine status of pregnant women or their thyroid function during pregnancy with the mental development of their children, this pilot, prospective, observational study was designed to compare school-age children born to women who had been using iodized salt long before becoming pregnant (supplemented mothers) with children born to women who had never used it (unsupplemented mothers). A further stratification based on maternal LT4 treatment that was started before conception and uninterruptedly continued during pregnancy, either in association or not in association with iodized salt consumption, eventually allowed the effect of maternal nutritional iodine status to be assessed against maternal serum fT4 concentrations on cognitive outcome in the progeny.

It was found that children born to mothers with similar nutritional iodine status showed comparable intellectual abilities. Specifically, no differences in IQs were observed between I and I + T4 children or between no-I and no-I + T4 children, which was largely independent of mothers' gestational fT4 concentrations. Conversely, the children of mothers on LT4 treatment (I + T4 and no-I + T4) performed differently in IQ tests, despite their mothers having had consistently similar fT4 concentrations throughout gestation. Taken as a whole, these findings suggest that neuro-intellectual outcomes in children may be more dependent on their mothers' nutritional iodine status than on maternal T4 levels throughout gestation. Indeed, logistic regression models designed to assess the dependence of suboptimum cognitive outcomes (IQ <85 points) on various explanatory variables failed to show a significant association with maternal thyroid parameters at any stage in pregnancy. Conversely, maternal nutritional iodine status proved to be positively associated with cognitive outcomes. Overall, children born to supplemented mothers had FSIQ, VIQ, and PIQ scores that were 13, 14, and 10 points higher, respectively, than those recorded for children born to unsupplemented mothers. In addition, individual data analysis showed that >40% of the children studied had defective cognitive outcomes, and the vast majority were born to mothers not using iodized salt. Accordingly, the risk of impaired cognitive function for children exposed to higher degrees of maternal ID was found to be almost eight times that of children whose mothers had more adequate iodine status. Notably, a high frequency (approximately 20%) of defective IQs was observed, even among children of supplemented mothers. Although other potential causes of poor neuro-intellectual development may not be excluded, the present findings suggest that iodine supplementation by means of iodized salt alone may be inadequate to guarantee normal neurodevelopment. The region where the study was conducted is a small and iodine-deficient mountain area (approximately 600 m above sea level), which is quite far from the coast. The diet of the population in this area is mainly based on a Mediterranean-style food system. Staple foods include starchy foods, vegetables, legumes, milk and dairy products, olive oil, and fruits, which are almost exclusively locally produced. Seafood is only occasionally consumed, and the protein supply generally comes from meat obtained from animals raised in local small-scale livestock farms. Due to the long-standing ID in this area, this food system proves to be inadequate for providing physiological iodine requirements, as demonstrated by the description of several iodine-deficiency disorders affecting people living in this region (14,29,30). In these circumstances, the contribution of iodized salt to iodine nutrition becomes quite relevant, although it is unlikely to be sufficient for meeting the increased iodine requirements during pregnancy and lactation.

Since cognitive development is affected by several factors, it cannot definitely be ruled out that confounders other than those controlled for in the study are to some extent involved. However, in the absence of any association between cognitive outcomes and maternal thyroxinemia, the less favorable cognitive outcomes observed in the children of unsupplemented mothers might be the result of the developing brain being exposed to insufficient fetal TH levels as a consequence of low iodine availability. This hypothesis would provide a unifying explanation both for the absence of differing IQs in children born to mothers not using iodized salt and for the much better cognitive outcomes of children born to mothers with more adequate iodine intake during pregnancy. In this view, the relative contribution of the fetal thyroid to neurodevelopment would not be secondary to that of the mother, which is a hypothesis that is in agreement with the evidence that children affected by congenital hypothyroidism, even when treated early, may still exhibit subtle and specific neurocognitive deficits (39,40). The relevance of fetal thyroid contribution to neurodevelopment is further supported by the evidence of poorer cognitive outcomes reported in preterm infants with hypothyroxinemia compared with preterm euthyroid infants (41).

To some extent, the present results are in agreement with those of Lazarus et al. in the Controlled Antenatal Thyroid Screening Study (CATS), who reported that use of LT4 therapy for mild maternal hypothyroidism made no difference to cognitive outcomes (24). Possible explanations for the negative results of this study were recently provided by Taylor et al. in a retrospective analysis of almost 500 mother–child pairs from the CATS population. The authors of that study postulated that exposure to perchlorate during intrauterine life might have played a role in reducing fetal TH production by disrupting fetal thyroidal iodine uptake. In fact, maternal perchlorate levels in the highest 10% of the population were associated with increased odds of offspring having an IQ of <80, with a greater negative impact observed on VIQ than PIQ (42). Interestingly, similar results for VIQ have also been reported in a recent study conducted in the UK aimed at investigating the association between maternal iodine status and child IQ at age 8–9 years (5). In their study, Bath et al. clearly demonstrated that, far from being confined to infants of mothers with very low iodine status, the children of mild-to-moderate ID mothers also had an enhanced risk of suboptimum cognitive scores. A decline in VIQ with decreasing maternal iodine status was also observed in the present small series of children. Significant differences were found between VIQ and PIQ only in the children of unsupplemented mothers, which is a finding further confirmed by evidence of an association between low maternal urinary iodine and defective VIQ.

A relatively low VIQ score may suggest underlying central auditory processing disorders (cAPDs), which are characterized by difficulties in the perceptual processing of auditory information in the central nervous system of individuals with normal peripheral hearing (43,44). Persisting abnormalities in auditory processing and selective attention has been reported possibly to influence cognitive development in children with congenital hypothyroidism who are diagnosed and treated early (45). Similarly, a cAPD that influences working memory has recently been proposed to be responsible for reduced outcomes in areas of literacy in children of mothers with a gestational UIE below the current established cutoff point for sufficiency (150 μg/L) (6). Interestingly, causes of cAPDs may include misplaced cells in the auditory cortical areas and/or delay in myelin maturation (46,47). Both of these processes are known to be TH regulated, the first occurring at early stages of neurodevelopment under the control of both maternal and fetal TH, and the second beginning in mid/late gestation and continuing during postnatal development (48).

cAPDs cover a wide range of behavioral disorders, including specific learning disabilities, autism, and ADHD (44,49). The latter, reported by the authors in a population of children from the same area as those included in this study, was found to be strongly associated with ID-related early gestation maternal hypothyroxinemia (14).

This pilot study has several limitations, the first two being the relatively small size of the population studied and the fact that the design of the study did not include parental IQ testing. However, these limitations are to some extent attenuated by the fact that the children included in the study all grew up in the same narrow rural area and were from homogeneous socioeconomic and educational backgrounds. Indeed, no differences in the SES, evaluated by means of the Hollingshead Index, were observed between the different subgroups, and no association was found at logistic regression between children IQs and maternal and paternal education. In addition, stringent inclusion/exclusion criteria for both mothers and children were adopted, thus limiting the bias of known confounders potentially affecting overall results. Third, UIEs were not corrected for urinary volume. Nonetheless, several urine samples were obtained for each pregnancy, and adjusted mean UICs were calculated to account for day-to-day variations, which reduces the risk of misclassification of individual nutritional iodine status. Finally, the observational design of the current study and an inadequate power for the sample size do not allow us to draw firm conclusions, and therefore the results would theoretically need confirmation from larger, randomized controlled trials to determine the precise effects on child cognition of maternal iodine supplementation and LT4 during pregnancy. Whether such intervention studies in conditions of even mild ID would be ethical is currently disputed (50 –52). In our view, similar trials might be difficult to implement because they should include a study group of pregnant women (and their fetuses) deliberately exposed to the potential risks of gestational ID.

In conclusion, there is evidence that brain development is vulnerable to TH deficiency. In addition to affecting maternal thyroid function, a less than adequate maternal iodine intake during pregnancy potentially leads to insufficient fetal TH output due to reduced iodine storage in the fetal gland. The present findings provide evidence that mild-to-moderate ID during fetal development, even with the lack of maternal thyroid insufficiency, may result in cognitive impairment later in life, with a seemingly greater impact on verbal abilities. Close monitoring of maternal iodine intake before and during pregnancy is therefore strongly recommended to prevent impaired neurodevelopment in offspring. Although the results of this preliminary study might be partly biased by the relatively small sample size, they may be informative for the design and implementation of future and adequately powered studies.

Footnotes

Author Disclosure Statement

No competing financial interests exist. No funding was secured for this study.

References

Williams

. 2008. Neurodevelopmental and neurophysiological actions of thyroid hormone. J Neuroend, 20:784–794.

Delange

. 2000. The disorders induced by iodine deficiency. Thyroid, 4:107–128.

Glinoer

, Delange

. 2000. The potential repercussions of maternal, fetal, and neonatal hypothyroxinemia on the progeny. Thyroid, 10:871–887.

Morreale De Escobar

, Obregon

, Escobar Del Rey

. 2000. Is neuropsychological development related to maternal hypothyroidism or to maternal hypothyroxinemia?. J Clin Endocrinol Metab, 85:3975–3987.

Bath

, Steer

, Golding

, Emmett

, Rayman

. 2013. Effect of inadequate iodine status in UK pregnant women on cognitive outcomes in their children: results from the Avon Longitudinal Study of Parents and Children (ALSPAC). Lancet, 27:331–337.

Hynes

, Otahal

, Hay

, Burgess

. 2013. Mild iodine deficiency during pregnancy is associated with reduced educational outcomes in the offspring: 9-year follow-up of the gestational iodine cohort. J Clin Endocrinol Metab, 98:1954–1962.

Haddow

, Palomaki

, Allan

, Williams

, Knight

, Gagnon

, O'Heir

, Mitchell

, Hermos

, Waisbren

, Faix

, Klein

. 1999. Maternal thyroid deficiency during pregnancy and subsequent neuropsychological development of the child. N Engl J Med, 341:549–555.

, Shan

, Teng

, Yu

, Li

, Fan

, Teng

, Guo

, Wang

, Li

, Chen

, Wang

, Chawinga

, Zhang

, Yang

, Zhao

, Hua

. 2010. Abnormalities of maternal thyroid function during pregnancy affect neuropsychological development of their children at 25–30 months. Clin Endocrinol (Oxf), 72:825–829.

Ghassabian

, Bongers-Schokking

, Henrichs

, Jaddoe

, Visser

, de Muinck Keizer-Schrama

, Hooijkaas

, Steegers

, Hofman

, Verhulst

, van der Ende

, de Rijke

, Tiemeier

. 2011. Maternal thyroid function during pregnancy and behavioral problems in the offspring: the Generation R study. Pediatr Res, 69:454–459.

10.

Pop

, Kuijpens

, van Baar

, Verkerk

, van Son

, de Vijlder

, Vulsma

, Wiersinga

, Drexhage

, Vader

. 1999. Low maternal free thyroxine concentrations during early pregnancy are associated with impaired psychomotor development in infancy. Clin Endocrinol (Oxf), 50:149–155.

11.

Pop

, Brouwers

, Vader

, Vulsma

, van Baar

, de Vijlder

. 2003. Maternal hypothyroxinaemia during early pregnancy and subsequent child development: a 3-year follow-up study. Clin Endocrinol (Oxf), 59:282–288.

12.

Kooistra

, Crawford

, van Baar

, Brouwers

, Pop

. 2006. Neonatal effects of maternal hypothyroxinemia during early pregnancy. Pediatrics, 117:161–167.

13.

Finken

, van Eijsden

, Loomans

, Vrij kotte

, Rotteveel

. 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.

14.

Vermiglio

, Lo Presti

, Moleti

, Sidoti

, Tortorella

, Scaffidi

, Castagna

, Mattina

, Violi

, Crisa

, Artemisia

, Trimarchi

. 2004. Attention deficit and hyperactivity disorders in the offspring of mothers exposed to mild-moderate iodine deficiency: a possible novel iodine deficiency disorder in developed countries. J Clin Endocrinol Metab, 89:6054–6060.

15.

Roman

, Ghassabian

, Bongers-Schokking

, Jaddoe

, Hofman

, de Rijke

, Verhulst

, Tiemeier

. 2013. Association of gestational maternal hypothyroxinemia and increased autism risk. Ann Neurol, 74:733–742.

16.

Henrichs

, Bongers-Schokking

, Schenk

, Ghassabian

, Schmidt

, Visser

, Hooijkaas

, de Muinck Keizer-Schrama

, Hofman

, Jaddoe

, Visser

, Steegers

, Verhulst

, de Rijke

, Tiemeier

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

17.

Behrooz

, Tohidi

, Mehrabi

, Behrooz

, Tehranidoost

, Azizi

. 2011. Subclinical hypothyroidism in pregnancy: intellectual development of offspring. Thyroid, 21:1143–1147.

18.

Oken

, Braverman

, Platek

, Mitchell

, Lee

, Pearce

. 2009. Neonatal thyroxine, maternal thyroid function, and child cognition. J Clin Endocrinol Metab, 94:497–503.

19.

Chevrier

, Harley

, Kogut

, Holland

, Johnson

, Eskenazi

. 2011. Maternal thyroid function during the second half of pregnancy and child neurodevelopment at 6, 12, 24, and 60 months of age. J Thyroid Res, 2011:426–427.

20.

Craig

, Allan

, Kloza

, Pulkkinen

, Waisbren

, Spratt

, Palomaki

, Neveux

, Haddow

. 2012. Mid-gestational maternal free thyroxine concentration and offspring neurocognitive development at age two years. J Clin Endocrinol Metab, 97:E22–E28.

21.

Liu

, Momotani

, Noh

, Ishikawa

, Takebe

, Ito

. 1994. Maternal hypothyroidism during early pregnancy and intellectual development of the progeny. Arch Intern Med, 154:785–787.

22.

Momotani

, Iwama

, Momotani

. 2012. Neurodevelopment in children born to hypothyroid mothers restored to normal thyroxine (T₄) concentration by late pregnancy in Japan: no apparent influence of maternal T₄ deficiency. J Clin Endocrinol Metab, 9:1104–1108.

23.

Downing

, Halpern

, Carswell

, Brown

. 2012. Severe maternal hypothyroidism corrected prior to the third trimester is associated with normal cognitive outcome in the offspring. Thyroid, 22:625–630.

24.

Lazarus

, Bestwick

, Channon

, Paradice

, Maina

, Rees

, Chiusano

, John

, Guaraldo

, George

, Perona

, Dall'Amico

, Parkes

, Joomun

, Wald

. 2012. Antenatal thyroid screening and childhood cognitive function. N Engl J Med, 366:493–501.

25.

Skeaff

. 2011. Iodine deficiency in pregnancy: the effect on neurodevelopment in the child. Nutrients, 3:265–273.

26.

Zimmermann

. 2009. Iodine deficiency. End Rev, 30:376–408.

27.

Berbel

, Mestre

, Santamaria

, Palazon

, Franco

, Graells

, Gonzalez-Torga

, de Escobar

. 2009. Delayed neurobehavioral development in children born to pregnant women with mild hypothyroxinemia during the first month of gestation: the importance of early iodine supplementation. Thyroid, 19:511–519.

28.

Taylor

, Okosieme

, Dayan

, Lazarus

. 2013. Therapy of endocrine disease: impact of iodine supplementation in mild-to-moderate iodine deficiency: systematic review and meta-analysis. Eur J Endocrinol, 170:R1–R15.

29.

Squatrito

, Delange

, Trimarchi

, Lisi

, Vigneri

. 1981. Endemic cretinism in Sicily. J Endocrinol Invest, 4:295–302.

30.

Vermiglio

, Sidoti

, Finocchiaro

, Battiato

, Lo Presti

, Benvenga

, Trimarchi

. 1990. Defective neuromotor and cognitive ability in iodine-deficient schoolchildren of an endemic goiter region in Sicily. J Clin Endocrinol Metab, 70:379–384.

31.

Gharib

, Papini

, Paschke

, Duick

, Valcavi

, Hegedüs

, Vitti

. 2010. AACE/AME/ETA Task Force on Thyroid Nodules. Endocr Pract, 16:1–43.

32.

Moleti

, Lo Presti

, Campolo

, Mattina

, Galletti

, Mandolfino

, Violi

, Giorgianni

, De Domenico

, Trimarchi

, Vermiglio

. 2008. Iodine prophylaxis using iodized salt and risk of maternal thyroid failure in conditions of mild iodine deficiency. J Clin Endocrinol Metab, 93:2616–2621.

33.

Wechsler D 1991 WISC-III: Wechsler Intelligence Scale for Children. The Psychological Corporation, New York.

34.

Orsini

, Picone

. 2006 WISC-III. Contributo Alla Taratura Italiana. Giunti O.S. Organizzazioni Speciali, Florence, Italy.

35.

Dunn

, Crutchfield

, Gutekunst

, Dunn

. 1993. Two simple methods for measuring iodine in urine. Thyroid, 3:119–123.

36.

Subcommittee on Interpretation and Uses of Dietary Reference Intakes, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes 2003 Appendix E—Adjustment of observed intake data to estimate the distribution of usual intakes in a group. In Dietary Reference Intakes: Applications in Dietary Planning. National Academies Press, Washington DC, pp 196–208.

37.

Hollingshead A 1975 Four Factor Index of Social Status. Department of Sociology, Yale University, New Haven, CT.

38.

Coscarelli

, Balboni

, Cubelli

. 2008. Il livello socio-culturale nella ricerca psicologica. Problemi concettuali e metodologici. Psicol Soc (Bologna), 3:387–408.

39.

Léger

. 2015. Congenital hypothyroidism: a clinical update of long-term outcome in young adults. Eur J Endocrinol, 172:R67–R77.

40.

Rovet

. 2014. The role of thyroid hormones for brain development and cognitive function. Endocr Dev, 26:26–43.

41.

Delahunty

, Falconer

, Hume

, Jackson

, Midgley

, Mirfield

, Ogston

, Perra

, Simpson

, Watson

, Willatts

, Williams

; Scottish Preterm Thyroid Group. 2010. Levels of neonatal thyroid hormone in preterm infants and neurodevelopmental outcome at 5 1/2 years: millennium cohort study. J Clin Endocrinol Metab, 95:4898–4908.

42.

Taylor

, Okosieme

, Murphy

, Hales

, Chiusano

, Maina

, Joomun

, Bestwick

, Smyth

, Paradice

, Channon

, Braverman

, Dayan

, Lazarus

, Pearce

. 2014. Maternal perchlorate levels in women with borderline thyroid function during pregnancy and the cognitive development of their offspring: data from the controlled antenatal thyroid study. J Clin Endocrinol Metab, 99:4291–4298.

43.

American Speech-Language-Hearing Association 2005 (Central) auditory processing disorders [Technical Report]. Available at: www.asha.org/policy (accessed October 2014 ).

44.

Bamiou

, Musiek

, Luxon

. 2001. Aetiology and clinical presentations of auditory processing disorders—a review. Arch Dis Child, 85:361–365.

45.

Oerbeck

, Reinvang

, Sundet

, Heyerdahl

. 2007. Young adults with severe congenital hypothyroidism: cognitive event related potentials (ERPs) and the significance of an early start of thyroxine treatment. Scand J Psychol, 48:61–67.

46.

Knipper

, Zinn

, Maier

, Praetorius

, Rohbock

, Köpschall

, Zimmermann

. 2000. Thyroid hormone deficiency before the onset of hearing causes irreversible damage to peripheral and central auditory systems. J Neurophysiol, 83:3101–3112.

47.

Boscariol

, Garcia

, Guimarães

, Montenegro

, Hage

, Cendes

, Guerreiro

. 2010. Auditory processing disorder in perisylvian syndrome. Brain Dev, 32:299–304.

48.

Patel

, Landers

, Li

, Mortimer

, Richard

. 2011. Thyroid hormones and fetal neurological development. J Endocrinol, 209:1–8.

49.

O'Connor

. 2012. Auditory processing in autism spectrum disorder: a review. Neurosci Biobehav Rev, 36:836–854.

50.

Zhou

, Anderson

, Gibson

, Makrides

. 2013. Effect of iodine supplementation in pregnancy on child development and other clinical outcomes: a systematic review of randomized controlled trials. Am J Clin Nutr, 98:1241–1254.

51.

Stagnaro-Green

, Sullivan

, Pearce

. 2012. Iodine supplementation during pregnancy and lactation. JAMA, 308:2463–2464.

52.

Bath

, Jolly

, Rayman

. 2013. Iodine supplements during and after pregnancy. JAMA, 309:1345.