Iodine deficiency in pregnant women in Sweden: a national cross-sectional study

Adequate dietary iodine is essential for the production of thyroid hormones. Moderate-to-severe iodine deficiency during pregnancy increases the risk of cognitive impairment in the offspring [1‐4]. Observational studies suggest that even mild iodine deficiency during pregnancy may negatively affect verbal intelligence quotient and educational level in their children [5‐7], although a causal relation could not be confirmed in a recent randomized controlled trial of iodine supplementation in pregnant women with mild iodine deficiency [8].

Pregnant women have higher iodine requirements than before pregnancy and are vulnerable to iodine deficiency due to increased thyroid hormone production, increased renal iodine clearance, and transplacental transfer of iodine to the fetus [9]. The recommended daily iodine intake for pregnant women is 250 μg, higher than the 150 μg/day recommended to non-pregnant women of reproductive age [10]. A median spot urinary iodine concentration (UIC) during pregnancy of 150–249 µg/L indicates adequate iodine nutrition in a population [10].

Iodine deficiency was historically endemic in Sweden with reported goiter prevalence up to 60% in certain areas [11, 12]. Swedish authorities introduced voluntary iodine fortification of table salt in 1936 [11, 13]. The current level of iodine in fortified salt is 50 mg/kg salt and has remained the same since 1966 [14, 15]. The first national monitoring of the present salt iodization program was conducted in school-aged children (6–12 years) in 2007 and the findings implied iodine sufficiency [16, 17]: no goiter was observed [16] and the median UIC was 125 µg/L [17], within the recommended range of 100–299 µg/L [10, 18]. At the same time, a local study suggested mild iodine deficiency in the third trimester of pregnancy [19]. Furthermore, more than 75% of consumed salt in Sweden is non-iodized [personal communication with the salt industry, 2017], milk iodine concentration has declined [20], consumption of iodine-rich fish and dairy products is decreasing [14], and national authorities recommend reduced salt intake [21].

We conducted a national cross-sectional study in Sweden with the objectives of assessing iodine status and thyroid function in pregnant women and evaluating the impact of prenatal iodine supplementation on iodine status and thyroid function.

Women were given a plastic cup and were asked to provide ~ 20 mL of fresh midstream urine, directly after their routine visit to MHC. Midwives were instructed not to expose the urine samples to any dipsticks, due to risk of iodine contamination. Samples were transported in a cool box from the place of collection to the central laboratory. Aliquots of all urine samples were frozen at − 20 °C until analysis. Blood drops (50 μL) were collected by a finger prick and directly collected onto filter paper cards (IDBS-226, Perkin Elmer, CT). The DBS cards were dried at room temperature, placed in sealed plastic bags, and stored at 4 °C until analysis or frozen at − 20 °C before analysis.

We enrolled 743 pregnant women from 25 MHC centers (18% of all MHCs asked). The recruitment of the 25 MHC centers (Table 1) followed the intended plan, with one exception: the region with the highest population nativity (H1 region) contributed two out of the intended six MHC centers. The remaining four MHC centers were recruited from the H2 region, i.e. the region with the second highest population nativity. Subject characteristics are presented in Table 2. Iodine-containing supplements (≥ 150 µg iodine/day) were used by 34.8% of the women.

The overall median UIC (IQR) was 101 µg/L (61, 182; n = 737) with bootstrapped 95% confidence interval (CI) (95, 108), without difference across trimesters (p = 0.381, Table 3). The UIC:U-creatinine ratio (IQR) was 0.11 (0.07, 0.19; n = 737), with no difference across trimesters (p = 0.153, Table 3). The median UIC (IQR) was 149 µg/L (88, 253; n = 253) in supplement users (using supplements containing ≥ 150 µg iodine/day) and 85 µg/L (51,134; n = 440) in non-supplement users (using no supplements or supplements containing < 150 µg iodine/day) (p < 0.001). The bootstrapped 95% CI of median UIC was (132, 164) in supplement users and (79, 92) in non-supplement users (Fig. 1a). Levothyroxine was used by 8.9% (n = 66) of the women, being equally distributed among supplement and non-supplement users (p = 0.206). The median UIC after excluding those on levothyroxine treatment did not differ compared to the median UIC in the complete study group (p = 0.942, data not shown).

The overall geometric mean DBS-Tg (95% CI), after exclusion of subjects on levothyroxine (n = 66) was 22.1 μg/L (20.8, 23.5; n = 675), with no difference across trimesters (p = 0.579, Table 3). The geometric mean DBS-Tg (95% CI) was 19.1 µg/L (17.2, 21.3, n = 229) in supplement users and 24.4 µg/L (22.6, 26.3, n = 405) in non-supplement users (p < 0.001) (Fig. 1b). The prevalence of elevated DBS-Tg was 19%: 13.1% among supplement users and 22.5% of among non-supplement users (p = 0.004). After exclusion of subjects with S-TPOab positivity (n = 45), the geometric mean DBS-Tg was still lower (p = 0.002) in supplement users than in non-supplement users (data not shown).The concentration of thyroid function parameters (DBS-TSH, DBS-tT4, and S-TPOabs) and the prevalence of subclinical and clinical thyroid disorders after exclusion of subjects treated with levothyroxine (n = 66) are presented in Table 3. The prevalence of thyroid dysfunction was low overall, except for isolated hypothyroxinemia, present in 24.8% of pregnant women (Table 3). When stratified for trimesters, the frequency of isolated hypothyroxenimia was higher and isolated hyperthyroxenimia lower in third trimester than in the first trimester. Comparison between first and second trimester gave just marginal significances of the same direction, whereas second and third trimesters presented no differences (Table 3). Clinical thyroid disease and subclinical hyperthyroidism were not observed. There was no difference between supplement users and non-supplement users for: prevalence of levothyroxine medication (p = 0.206, n = 699), DBS-TSH (p = 0.633, n = 633), DBS-tT4 (p = 0.328, n = 633), and S-TPOab positivity (p = 0.107, n = 318). The two groups did not differ for any of the three entities of thyroid morbidity observed: subclinical hypothyroidism (p = 1.0, n = 631), isolated hypothyroxinemia (p = 0.633, n = 631), and isolated hyperthyroxinemia (p = 0.449, n = 631). After exclusion of S-TPOab positive subjects (n = 45), no difference was observed between supplement users and non-supplement users for DBS-TSH and DBS-tT4 (data not shown).

This is the first national study assessing iodine status and thyroid function of pregnant women in Sweden. The median UIC of 101 μg/L in the study population is far below the recommended optimal range for pregnant women of 150–249 μg/L [10] and the results suggest mild iodine deficiency.

Our results are consistent with previous data. Local and regional studies in pregnant women in Sweden, conducted more than 10 years ago [19, 30], report low UIC and elevated Tg concentration, indicating that iodine intake in pregnant women has been insufficient over a prolonged period. However, the overall iodine intake in school-aged children is adequate. A national study in children aged 6–12 years conducted by our group in 2007 found a median UIC of 125 μg/L [17], which is within the recommended range of 100–299 μg/L [10, 18], and goiter was not observed [16].

Iodine deficiency during pregnancy is frequently observed despite iodine sufficiency in school-aged children, e.g. in Norway, Denmark, Switzerland, USA [14, 31‐33]. A review of studies conducted simultaneously in school-aged children and pregnant women in countries where salt is the main dietary source of iodine, showed that median UIC in pregnant women typically meets the recommended 150 μg/L in populations where the median UIC in school-aged children exceeds 180 μg/L [34]. However, in countries with voluntary iodization, the coverage by iodized salt may be incomplete. Although the data are limited (personal communication with the salt industry, 2017), most of the household salt in Sweden appears to be iodized, whereas processed food and food served in restaurants, an important salt source for working women in Sweden, seems to have lower coverage by iodized salt. Therefore, despite voluntary iodization at 50 mg/kg salt [14, 15], above the recommended 20–40 mg/kg [10], iodine intake in Sweden is inadequate in pregnant women, probably due to poor coverage by iodized salt. A recent study conducted in populations with mandatory legislation and high coverage by iodized salt [35] demonstrated that a well implemented salt iodization program ensures adequate dietary intake in all population groups. National recommendations in Sweden to promote iodized salt by the food industry are warranted as well as consideration of applying this on a mandatory basis.

One-third of the women in our study consumed iodine-containing prenatal dietary supplements containing ≥ 150 μg iodine/day and the median UIC in iodine supplement users suggested adequate iodine intake. The additional supplemental iodine may have a positive impact on the thyroid gland, as suggested by the lower DBS-Tg in this group compared to non-supplement users. However, the DBS-Tg concentration in supplement users was still higher than expected for an iodine-sufficient population (~ 10 μg/L) [25]. It is possible that women in Sweden have inadequate iodine intake before conception and iodine supplementation is not sufficient to normalize DBS-Tg during pregnancy. This is supported by the elevated DBS-Tg observed in the first trimester. It is also possible that other nutrient deficiencies beyond iodine (e.g. selenium, iron [36, 37]) interact with thyroid metabolism. Finally, thyroid size, reflected by DBS-Tg, may be larger in the Swedish population, as shown for school-aged children when compared with an age-matched reference international population [16].

The overall prevalence of isolated hypothyroxinemia was 25% and did not differ between supplement users and non-users, but was higher in third trimester (32%) compared with first trimester (18%). The high prevalence of isolated hypothyroxinemia in our population contrasts with a local Swedish study [38], where only 2.8% of pregnant women presented isolated hypothyroxinemia, which was though measured in serum samples at median pregnancy week 18. The clinical importance of isolated hypothyroxinemia is debated; some studies indicate an association with impaired neurocognitive development of offspring [39], whereas others do not [40, 41]. The use of tT4 concentration, as a proxy for free T4, may induce bias when estimating the prevalence of hypothyroxinemia. Increased estradiol production during pregnancy leads to elevation of thyroxine-binding globulin, resulting in higher tT4 despite stable or lower free T4 during pregnancy. To compensate for this, an elevation of the upper limit of the reference range throughout pregnancy is recommended [26] and was applied in our study, but the precision of the lower limit of tT4 is uncertain. Furthermore, no reference range has specifically been defined for DBS-T4 and DBS-TSH during pregnancy, possibly contributing to misclassification. Uncertainty also prevails in terms of the free T4 reference range during pregnancy, free T4 being associated with tT4. In one study [42], free T4 decreased from first to second trimester despite iodine supplementation and TSH remained stable, suggesting the decrease in free T4 during pregnancy may be a physiological observation. The same trend during pregnancy was observed in our study population regarding isolated hypothyroxinemia; the opposite trend, though in lower prevalence, was noticed in isolated hyperthyroxenemia, which was more frequent in first trimester (8%) than in third trimester (3%), highlighting the need of proper reference intervals deriving from a thyroid healthy, iodine sufficient population of pregnant women.

The observed mild iodine deficiency in this study raises the question of national recommendation in Sweden for iodine supplementation of 150 µg/day during pregnancy, as recommended in several other countries [26, 43] and as supported by World Health Organization when the coverage of iodized salt is incomplete [10]. Several observational studies suggest an association between mild iodine deficiency during fetal life and poor educational level, motor skills, or verbal abilities in children 3–12 years of age [5, 6, 44, 45]. However, this was not confirmed in randomized controlled trials of iodine supplementation of pregnant women with mild iodine deficiency [8, 42], possibly due to lower than expected prevalence of iodine deficiency in the population studied in one trial [8] or due to low sample size in the other trial [42]. At the same time, other observational studies have indicated the opposite, i.e. iodine supplementation during fetal life was associated with higher risk of behavioral problems or poorer mental and psychomotor development in children [45‐47]. It has been proposed that a sudden increase of iodine intake blocks fetal thyroid function, leading to hypothyroidism and goiter, especially in the case of preconceptional iodine deficiency [48]. Women with preconceptional iodine deficiency may handle the restricted available iodine more effectively during pregnancy [49]. Taken together, the data supporting iodine supplementation during pregnancy in mild ID are inconclusive. A randomized clinical trial of prenatal iodine supplementation is currently ongoing in Sweden and the results will provide valuable data on the consequences of mild iodine deficiency on fetal development [50].

We recognize the low response rate of the MHC centers (18% of all asked) as a weakness of the study, but the reason given for declining participation was the high workload of midwives, which is unlikely to be associated with the outcome parameters. Although only two of the intended six MHC centers were included from the H1 region, the remaining four MHC centers were recruited from the region (H2) with the closest population density. The applied stratified cluster sampling methodology still ensured a representative sample of the Swedish pregnant population. Data on iron and selen status or other endocrine disruptors were not collected; we plan, however, to investigate the possible interaction between iodine, iron and selen within the framework of another study (ClinicalTrials.gov identifier: NCT02378246). We did not study the locality either, as sample size for each area was small and we followed the WHO recommendations for monitoring of iodization program by clinic-based cross-sectional surveys (10). Besides, a comparison between the different areas in the country was beyond the scope of this study, as we aimed to set the ground for recommendations to pregnant women nationally.

In conclusion, adequate iodine intake in school-aged children does not guarantee adequate intake in risk populations, even in a country with a high level of iodine fortification of salt. Pregnant women in Sweden have inadequate iodine nutrition and only 35% consume a daily supplement with iodine ≥ 150 μg. Women not taking iodine-containing supplements or supplements containing < 150 μg/day are exposed to mild iodine deficiency. The low iodine intake increases the risk of elevated thyroid activity to maintain euthyroidism. The long-standing national policy of universal salt iodization on a voluntary basis is not operating as intended and actions to improve its coverage must be urgently undertaken. We suggest that Swedish authorities encourage the food industry to use iodized salt instead of non-iodized salt and we urge regular monitoring of iodine status of the general Swedish population and of groups at particular risk of inadequate iodine intake, as a part of the national nutrition policy, supported by the National Food Agency. A randomized controlled trial in a population with mild iodine deficiency is highly warranted [50] to provide authorities with guidance on whether iodine supplementation during pregnancy should be recommended or not. If the coverage of iodized salt is not improved and mild iodine deficiency infers mental consequences, mandatory iodization of table salt may be considered.