THERAPY OF ENDOCRINE DISEASE: Impact of iodine supplementation in mild-to-moderate iodine deficiency: systematic review and meta-analysis

Search strategy

We searched Medline and the Cochrane library for relevant articles published in the English language between January 1966 and April 2013. We used various combinations of the following terms: iodine, deficiency, supplementation, mild, moderate, pregnancy, goitre, IQ, childhood, development, neurological, thyroid, TSH, hypothyroidism, thyroiditis, hyperthyroidism, side effects, cost, salt, fortification, maternal and foetus. Additional publications were sourced from references in individual articles. Relevant articles were selected after reading through all titles and abstracts and full texts were obtained if the information contained in the title or abstract was insufficient to exclude the study.

Inclusion criteria

We selected studies for review if they were randomised controlled trials (RCTs), quasi-randomised trials, or prospective cohort or case–control studies which investigated: i) the effects of maternal iodine supplementation in pregnancy on a) maternal thyroid function, b) foetal thyroid function and c) child neurodevelopment; and ii) the effects of childhood iodine supplementation on child cognitive performance. Studies were chosen if: i) participants received iodine supplements; ii) an appropriate control group was included which comprised participants who either received no supplements or received a significantly lower dose of supplements; iii) thyroid function, thyroid volume or cognitive performance were determined as outcomes; and iv) the study was limited to populations with mild-to-moderate iodine deficiency as determined from the median population urinary iodine.

Quality of studies

The quality of randomised trials was assessed using the criteria set by Jadad et al. (21). Studies were graded as high, moderate and low quality according to specific scores on the randomisation procedure, blinding of participants and investigators, and withdrawal rates from the study. We graded the quality of non-randomised trials using the Newcastle–Ottawa scale (22). This assessed the key domains of study selection, comparability and exposure (case–control studies) or outcome (cohort studies). Points were awarded for each study as follows: maximum of 4 for study selection, 2 for comparability and 4 for exposure or outcome. Based on total scores, studies were graded as high (9–10 points), moderate (7–8 points) or low (<7 points) quality.

Data extraction

Study selection and data extraction were independently conducted by two reviewers (P N Taylor and O E Okosieme) using preset selection criteria. Differences were resolved by consensus or by referral to the other authors (C M Dayan and J H Lazarus). Extracted information included sample size, geographical area, baseline iodine status according to the population median UIC and iodine supplementation (preparation, dose and route of administration). The outcomes examined were thyroid function, thyroid volume, and scales of childhood neurodevelopment and cognition.

Statistical analysis

A meta-analysis of studies of iodine supplementation in pregnancy was not considered feasible due to wide disparities in study selection criteria and lack of controlled trials on newborn cognitive outcomes. We undertook a meta-analysis of RCTs, which addressed iodine supplementation and its effect on cognitive function in school-age children. For each cognitive test we calculated the standardised mean difference (SMD) and s.e.m. as the difference in change from baseline between treated and control participants divided by the pooled s.d. Improved cognitive performance from baseline was reported as a positive effect and scores for tests in which lower scores indicated better performance were multiplied by −1. Cognitive tests were grouped according to domains and combined averages were calculated where more than one test assessed a single cognitive domain. A global cognitive score was derived from the average of all the individual domains. Unadjusted and adjusted effect sizes were derived using a random effects model (23) and inverse variance method and heterogeneity was assessed with the χ² test. The analysis was performed with the Review Manager (RevMan) Software, version 5.2 (The Nordic Cochrane Centre, The Cochrane Collaboration, 2012).

Study selection

The study selection flow process is shown in Fig. 1. Nine RCTs (24, 25, 26, 27, 28, 29, 30, 31, 32) and eight observational studies (33, 34, 35, 36, 37, 38, 39, 40) were included in the review. Seven RCTs were reported on the effects of maternal supplementation on maternal thyroid function (26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40). Of these, four RCTs also contained data on neonatal thyroid function (27, 29, 30, 32). One RCT (32) and four observational studies (33, 36, 39, 40) addressed the impact of gestational iodine supplementation on infant neuropsychological function, while two RCTs investigated the impact of iodine supplementation on cognitive performance in school-age children (24, 25).

• Download Figure
• Download figure as PowerPoint slide

Flow process of study selection.

Citation: European Journal of Endocrinology 170, 1; 10.1530/EJE-13-0651

Maternal iodine supplementation and foeto-maternal thyroid function

All seven RCTs which determined the effects of maternal iodine supplementation on maternal thyroid function were conducted in European countries, namely Italy (26, 28), Denmark (27, 31), Belgium (29), Germany (30) and Spain (32). All studies were in the area of mild-to-moderate iodine deficiency with median population urinary iodine excretion (UIE) ranging from 36 to 109 μg/l. Iodine supplementation doses varied from 50 to 300 μg daily of iodide. Three studies were graded as moderate quality while four were low quality. The trials included a total of 641 participants who received various doses of iodine supplementation in pregnancy (n=370) or served as controls who either received no treatment (n=174), placebo (n=24), a lower dose of supplements (n=35) or iodised salt alone (n=38) (Table 1). The median onset of intervention was earlier than 12 weeks in five of the studies (28, 29, 30, 31, 32) while follow-up continued into the postpartum in five studies (26, 27, 30, 31, 32) (Table 1). Iodine was administered in the form of potassium iodide (KI) preparations, iodine-containing vitamin preparations or iodised salt and one trial included an intervention arm which received KI in combination with levothyroxine (l-T₄) (29).

Five observational studies also assessed the impact of iodine supplementation on maternal thyroid function (33, 34, 35, 37, 38). Of these, three were of moderate quality while two were low quality. The interpretation and comparative analysis of the studies was compounded by their divergent study designs. In some studies estimation of iodine intake was partly dependent on patient recall of iodine intake (37, 38) or derived from semi-quantitative food questionnaires (34). Iodine supplements were taken in the form of tablets or iodised salt and two studies compared long-term pre-gestation iodised salt intake with post-conception supplementation (37, 38).

Maternal iodine supplementation and neurodevelopmental function in the child

One RCT (32) and four observational studies (33, 36, 39, 40) examined the impact of maternal iodine supplementation on child neurodevelopment, evaluated between 6 and 18 months of age. These studies were of low to moderate quality and differed in design. Santiago et al. (32) compared different doses of supplementary iodide with iodised salt consumption alone while Berbel et al. (36) compared early with delayed or no iodine supplementation at all in mildly hypothyroxinaemic women. Two of the observational studies estimated supplementary iodine intake from a semi-quantitative food frequency questionnaire which relied on participant recollection of dietary and supplementary iodine intake (34, 35).

The outcomes of these studies in terms of mental development scales (MDS), psychomotor development scales (PDS) and developmental quotients (DQs) are presented in Fig. 2. No differences were observed in MDS in any of the studies and the differences which attained significance were seen with PDS (Fig. 2a and b). Velasco et al. (33) observed that the children of treated women had better behavioural performance and psychomotor performance than those of untreated mothers (Fig. 2b). In a second study from Spain by Berbel et al. (36), the DQ was found to be higher in children of women who started taking supplementation from early pregnancy (4–6 gestational weeks) and in addition had a FT₄ above the 20th percentile compared with infants of mothers who commenced supplements in the second or third trimesters with free T₄ between 0 and 10th percentiles (Fig. 2c). Delayed neurobehavioural performance was observed in none of the infants of mothers who received supplements from early gestation compared with 25 and 37% of children of mothers who started supplements from second or third trimesters respectively (36).

• Download Figure
• Download figure as PowerPoint slide

Studies investigating the effect of maternal iodine supplementation on neurodevelopmental indices in the child. Mental development scales (MDS; a), psychomotor development scales (PDS; b) and developmental quotient (DQ; c) were determined in children of mothers who received iodine supplements or no supplements during pregnancy. Children were tested between 6 and 18 months. The MDS and PDS presented are adjusted for age of child at the time of testing and psychologist administering the test and were typified to a mean of 100 with s.d. of 15 points. The mean score for each index is denoted as white circles with black lines on either side of the circles representing the s.d. KI, potassium iodide; I-salt iodised salt, I-Sup Iodine supplements. For (c) ^aP<0.05 group 1 vs 2 and ^bP<0.001 group 1 vs 3.

Citation: European Journal of Endocrinology 170, 1; 10.1530/EJE-13-0651

A third study by Murcia et al. (40) showed that infants of mothers who self-reported a supplementary iodine intake of >150 μg/day scored lower on the psychomotor development index (PDI) than infants born to mothers with intake of <100 μg/day (Fig. 2b) from supplements only. This study also reported a higher risk of having a PDI <85 in children of women with iodine intake >150 μg/day (OR 1.8, CI 1.0, 3.3). A more recent study by the same group has demonstrated a significantly lower PDS in association with supplementary iodine intake >150 μg/day in three additional areas of Spain (Fig. 2b) (39). In the RCT by Santiago et al. (32), no differences in MDS or PDS were demonstrable between the infants of women who were randomised in the first trimester to receive either iodised salt only, or 200 or 300 μg daily of KI (Fig. 2a and b).

Iodine supplementation in childhood and cognitive function

A number of studies have examined the relationship between cognitive function in children and iodine nutrition status, but the majority of such studies were observational or conducted in severely iodine-deficient areas and were excluded from further analysis (Fig. 1). Excellent reviews of these studies have been published elsewhere (2, 4, 41, 42). We excluded one randomised trial in which iodised salt was introduced into the area during the course of the study thus compelling the authors to undertake a post hoc analysis based on final urine iodine concentrations (43). In the analysis they showed that children with improved UICs scored better on a combination of mental development tests than those with no improvement (43).

Only two RCTs which determined cognitive performance in children living in areas with mild-to-moderate iodine deficiency were suitable for meta-analysis. In a double blind trial, Zimmermann et al. (25) measured cognitive and motor performance in 310 Albanian children aged 10–12 years (median UIC 44 μg/l), who were randomly assigned to receive 400 mg of intramuscular iodine or placebo (25). Compared with placebo, the iodine-treated group showed improvements in four out of seven tests of cognition and motor performance, namely, rapid target marking, symbol search, rapid object naming and Raven's Coloured Progressive Matrices (P<0.0001) (25). Gordon et al. (24) in a marginally iodine-deficient New Zealand population randomly assigned 166 children aged 10–13 years (median UIC 63 μg/l) to receive a daily tablet of 150 μg of iodine or placebo for 28 weeks (24). Children who received iodine supplementation had improved iodine status and in addition scored higher on two out of four subtests for cognitive performance, namely, picture concepts (P=0.023) and matrix reasoning (P=0.040) (24).

For the meta-analysis cognitive tests were categorised into the following domains: i) perceptual reasoning; ii) processing speed; iii) working memory; and iv) global cognitive index. The global cognitive index was derived from the average of the scores in each of the domains. Unadjusted SMDs of the change in cognitive scores from baseline were computed from the raw scores reported by the authors, while adjusted SMDs were derived from the reported mean-adjusted treatment effects. s.e.m.-adjusted treatment effects were calculated using the recommended formula in the Cochrane handbook (44). The results of the analysis for individual domains are presented in Table 3, while Fig. 3 shows the forest plots for the global cognitive index. Beneficial effects of iodine supplementation were seen for both adjusted and unadjusted global indices with mild heterogeneity observed between the studies. For individual unadjusted domain scores, benefits were seen for processing speed but not for perceptual reasoning or working memory, while for the adjusted domains iodine was beneficial for perceptual reasoning although this was associated with significant heterogeneity.

Table 3

Results of meta-analysis^a on the impact of iodine supplementation on childhood cognitive performance.

Cognitive domains	SMD b	95% CI	P value	I2 c
Unadjusted analysisd
Perceptual reasoning	0.68	−0.10, 1.46	0.09	94% (P<0.0001)
Processing speed index	0.31	0.13, 0.49	0.0007	0% (P=0.95)
Working memory	0.05	−0.13, 0.23	0.55	0% (P=0.41)
Global cognitive index	0.41	0.13, 0.70	0.005	56% (P=0.13)
Adjusted analysise
Perceptual reasoning	0.55	0.05, 1.04	0.03	90% (P=0.001)
Processing speed index	0.21	−0.02, 0.44	0.08	59% (P=0.12)
Working memory	0.08	−0.07, 0.24	0.30	0% (P=0.91)
Global cognitive index	0.27	0.10, 0.44	0.002	26% (P=0.24)

^aStudies included in the meta-analysis (random effects analysis with inverse variance method): i) Gordon et al. (24) (iodine group=84, control group=82, baseline median urinary iodine concentration (UIC) 63 μg/l (24)). ii) Zimmermann et al. (25) (iodine group=159, control=151, baseline median UIC 44 μg/l (25)). P value for significance in SMD between iodine supplemented and control participants.

^bSMD, standardised mean difference between iodine supplemented and control participants.

^cTest of heterogeneity; corresponding P value in brackets.

^dUnadjusted SMD of the mean difference from baseline scores.

^eCognitive scores adjusted for baseline scores, sex, method of recruitment, cohort, ethnicity and household income (24) and adjusted for sex, school and baseline scores (25).

• Download Figure
• Download figure as PowerPoint slide

Forest plots showing effect of iodine supplementation on cognitive function (global cognitive index) in school-age children in mild-to-moderate iodine deficiency: (a) unadjusted SMD of the change from baseline. (b) Adjusted SMD of final cognitive scores adjusted for baseline scores, sex, method of recruitment, cohort, ethnicity and household income (24) and adjusted for sex, school and baseline scores (25). SMD, standardised mean difference.

Citation: European Journal of Endocrinology 170, 1; 10.1530/EJE-13-0651

Main findings of the review

Correction of mild-to-moderate iodine deficiency prevented increases in maternal and newborn thyroid volume and serum thyroglobulin. The effect of iodine on maternal thyroid function was less consistent, but three out of five RCTs which included an untreated control group showed that iodisation prevented the rise in serum TSH seen in untreated women while the other two RCTs showed no effect of iodine on TSH. Iodine supplementation at dose ranges of 200–300 μg daily was equally effective as iodised salt in optimising gestational thyroid indices. In a limited number of studies, long-term iodisation before pregnancy had more favourable effects on thyroid indices than recent iodisation and early institution of iodine supplements was more effective than initiation of treatment in late pregnancy. None of the controlled iodine intervention trials reported an excess frequency of overt thyroid dysfunction. In school-age children, iodine supplementation was associated with modest benefits on aspects of cognitive performance, including perceptual reasoning and global cognitive index. However, the impact of maternal iodine supplementation on newborn neurodevelopment remains uncertain due to a lack of appropriate controlled intervention trials.

Comparison to previous reviews

Early studies in severely iodine-deficient areas showed that iodine supplementation reduced the occurrence of endemic cretinism (2, 6). A meta-analysis by Bleichrodt & Born (45), comprising 21 observational or quasi-randomised studies, reported a 13.5 deficit in IQ in individuals with moderate to severe iodine deficiency. A more recent Cochrane review comprising 26 prospective trials has observed a trend towards goiter reduction in iodine-supplemented children but found mixed results for cognitive and psychomotor functions (41). In a second Cochrane review of six prospective controlled trials in the general population, favourable effects of iodine supplementation were observed on goiter reduction rates (42). While these reviews have addressed iodine deficiency across a spectrum of severity, only a limited number of reviews focused on iodine supplementation in marginal iodine deficiency states (4, 46). A previous review by Zimmermann (4) addressed the adverse effects of mild-to-moderate iodine deficiency in pregnancy and childhood and highlighted the benefits of maternal iodine supplementation on maternal thyroid indices and the need for further data on infant neurodevelopmental outcomes. The findings from the present systematic review corroborate the conclusions from these previous reviews and in addition provide an updated analysis including recent key trials (24, 25, 32).

Clinical significance of the findings

Pending further studies therefore it is unclear whether the available data on maternal thyroid indices from controlled intervention trials can serve as a surrogate for future child intellectual development. Iodine is required for the syntheses of thyroid hormones which in turn are crucial for normal neuronal myelination, migration and glial differentiation (47). Observational studies show that children born to women with untreated hypothyroidism have a 7–10 points deficit in IQ compared with those of euthyroid mothers (48). Although a recent controlled antenatal trial, the CATS study, failed to show any benefits of treating maternal hypothyroidism on child IQ at 3 years of age, the women in this study had milder degrees of hypothyroidism than in earlier studies and in addition l-T₄ treatment in the CATS study was initiated relatively late in gestation (49). A further consideration is that maternal hypothyroidism carries an increased risk of miscarriages and pre-term deliveries, complications largely preventable by l-T₄ therapy (50, 51, 52). Thus, taken together it is plausible that optimisation of maternal thyroid status through iodine supplementation in marginally iodine-deficient areas will have benefits on newborn cognitive function amongst other outcomes. To explore this, further RCTs or cluster randomised control trials of iodine supplementation in pre-pregnancy or early in the first trimester are needed with analysis of subsequent offspring IQ.

The benefits of iodine supplementation observed in school-aged children will be relevant to health policy makers. Reduced school performance adversely affects a region's productivity and economic potential. In Germany, analysis from 1981 to 2001 indicated that endemic iodine-deficiency goiter was responsible for costs of approximately one billion Euros per year (53). A comprehensive cost–benefit analysis from the USA showed that the projected benefit of screening and treatment for congenital iodine deficiency including savings in long-term care and productivity losses were thrice the costs incurred from treatment (54). An analysis of universal salt iodisation in the developing world suggests approximate gains of 35 billion USA dollars with a cost:benefit ratio of 1:70 (55). The advantages of correcting mild iodine deficiency will no doubt be less dramatic than for severe deficiency, but substantial returns in productivity and reduced health care costs will still be made.

A major concern with iodine supplementation has been the risk of iodine-induced thyroid dysfunction. We found no evidence of an excess of thyroid dysfunction in the controlled iodine intervention trials in pregnancy. Furthermore, the incidence of PPTD observed in these trials was not higher than published rates in the general population (50). Epidemiological studies, however, show that sharp increases in iodine intake in severely iodine-deficient populations may precipitate hyperthyroidism especially in elderly individuals with longstanding thyroid autonomy (56). Less striking manifestations are reported in marginally iodine-deficient areas or where iodine prophylaxis has been gradually introduced (56, 57, 58). In Denmark and Switzerland transient increases in the incidence of hyperthyroidism were recorded in the aftermath of iodisation but with reversal to baseline rates occurring within years of iodisation (59, 60). Surveys from Denmark (58), Greece (61), Sri Lanka (62), China (63) and parts of Africa (64) have all documented increases in the occurrence of thyroid dysfunction or autoimmunity in the wake of iodisation. In addition, the prevalence of both TPO-Ab and Tg-Ab (albeit low titre) was higher 4–5 years after cautious iodine fortification of salt was introduced in Denmark, particularly in young women (65). These population-level increases in the adult incidence of autoimmune thyroiditis thus seem an inevitable by-product of iodisation but should not deter future efforts at iodisation as the potential adverse effects of iodine deficiency on child development far outweighs the risk of correctable hypothyroidism in adults.

The mechanism of iodine fortification in mild-to-moderately deficient areas will merit careful consideration. Salt iodisation has proven cost effective in many countries and more than 70% of households worldwide now have access to iodised salt (14, 66, 67). However, salt may not suffice as the sole vehicle of iodisation in some countries. Retail outlet surveys conducted in the UK for example showed that most commercial salt brands lacked adequate iodine and iodised salt was unlikely to contribute substantially to overall iodine nutrition (10, 20). In addition, recent successful public health campaigns aimed at preventing cardiovascular disease through reduced salt consumption may have instigated further reductions in population iodine intake (68). However, a recent WHO forum has indicated that strategies to reduce salt intake and increase iodine fortification should not necessarily be contradictory and such strategies could support each other (69). Alternative strategies such as iodisation of bread as used in Australia and New Zealand (70, 71, 72) have been largely unexplored in the UK. However, supplementary iodine intake may still be indicated in addition to food fortification in vulnerable sub-populations such as pregnant women and children (71, 72). Routine prenatal iodine supplementation is recommended by professional bodies in North America (73) and Europe (74) but it is unlikely that this is systematically adhered to in many European countries where gestational iodine status remains inadequate (75, 76, 77). The method of iodisation notwithstanding, strategies for systematic surveillance should constitute a necessary component of future iodisation programmes.

Overall, the evidence base suggests that iodine supplementation in countries with mild-to-moderate iodine deficiency is beneficial, although data from RCTs remain lacking for maternal iodine supplementation and offspring cognitive development. However, recent studies (15, 16) have indicated that maternal iodine deficiency during pregnancy increased the odds of children having low IQ scores. We therefore agree that whilst awaiting results from current trials of iodine supplementation in pregnancy, pregnant and breastfeeding women should be offered iodine supplementation (78). Results from these ongoing trials and monitored interventions of iodine supplementation in children should ultimately provide a sufficient evidence base to identify if iodine fortification in countries with mild-to-moderate iodine deficiency would have sufficient benefit to justify implementation.

Strengths and limitations of the review

Ours is the first systematic review to focus exclusively on the benefits of correcting mild-to-moderate iodine deficiency. We have applied stringent selection criteria thereby confining our review to RCTs or observational studies with comparable control groups. The significance of our findings is however limited by the small number of high quality controlled trials especially in relation to maternal iodisation and newborn cognitive function. A key constraint in iodine intervention trials has been in justifying a control group of untreated pregnant women given the established knowledge on the devastating effects of iodine deficiency on foetal well-being. Perhaps reflecting disparities in study designs, the available data from uncontrolled observational studies has been conflicting. For instance, some authors suggest that self-reported supplement intake in excess of 150 μg daily is associated with impaired foetal neurodevelopment (39, 40). This is a concern given that the adaptive mechanisms to counteract the thyroid inhibitory actions of an acute iodide load, or Wolff–Chaikoff effect, do not fully develop in the foetus until late gestation (79). Furthermore, iodine content of food and water (80) is highly variable and some individuals in marginally iodine-deficient countries will inevitably be exposed to higher iodine intake than the WHO-recommended daily upper limit of 500 μg (3). Further studies will thus be necessary to define optimal thresholds of iodisation in marginally iodine-deficient populations.

Conclusion

Correction of mild-to-moderate iodine deficiency improves cognitive performance in school-age children, but there is insufficient data on the developmental outcomes in early life. Large scale controlled trials are now needed to clarify whether gestational supplementation will benefit infant neurodevelopment in countries with marginal iodine deficiency. One such trial is currently in progress in India and Thailand and aims to recruit 800 women randomised to receive 200 μg of KI or placebo in early pregnancy (81). Outcomes will include infant neuropsychiatric function assessed at 12–18 months amongst other indices of thyroid and developmental function. Another randomised control trial is ongoing in Australia, the Pregnancy Iodine and Neurodevelopment in Kids (PINK) (82), which is recruiting women before the 20th week of pregnancy comparing 150 μg iodine with placebo; however, studies from European countries remain desirable. The findings of these studies and this report should substantially add to the evidence base for an iodisation policy in countries with mild-to-moderate iodine deficiency.