https://orcid.org/0000-0002-1238-7850
Abstract
To monitor the impact of a cautious iodine fortification (IF) on the incidence of overt hypothyroidism in two subpopulations with different levels of preexisting iodine deficiency (ID).
A 20-year (1997 to 2016) prospective population-based study identified all new cases of diagnosed overt biochemical hypothyroidism in two open cohorts: a western cohort with moderate ID (n = 309,434; 1 January 1997) and an eastern cohort with mild ID (n = 224,535; 1 January 1997). A diagnostic algorithm was applied to all thyroid function tests performed within the study areas, and possible new cases were verified individually. Mandatory IF of salt was initiated in mid-2000 (13 ppm). The current study is a part of the DanThyr study.
At baseline, standardized incidence rates (SIRs) of hypothyroidism were 32.9 and 47.3/100.000/y in the cohorts with moderate and mild ID, respectively. The SIR of hypothyroidism increased significantly in both cohorts after implementing mandatory IF, with peak values of 150% in 2014 to 2016 for the moderate ID cohort and 130% in 2004 to 2005 for the mild ID cohort. Significant increases in SIR were seen among the young and middle-aged participants of both cohorts, whereas no changes were seen among the elderly participants (≥60 years). The follow-up period for the mildly iodine-deficient cohort was restricted up to and including 2008.
The cautious initiation of the IF program in Denmark caused a sustained increase in hypothyroidism incidence among subjects residing in areas of moderate and mild ID but only among the young and middle-aged participants.
Iron deficiency (ID) is one of the most common types of specific nutritional deficiencies globally. Endemic goiter is common in severe iron-deficient areas. Endemic cretinism with mental retardation may also result from severe ID (1). Moderate and mild ID is associated with a high prevalence of toxic and nontoxic multinodular goiter (1, 2).
According to World Health Organization recommendations, daily iodine intake for men, nonpregnant women, and children above 12 years of age should be ∼150 µg (1). Worldwide iodine fortification (IF) programs cover around two-thirds of the world’s population (1) and have eradicated or reduced the risk of ID disorders in many countries.
It is well known that increasing the daily iodine intake in a population with mild and moderate ID causes iodine-induced hyperthyroidism, which is a transient increase in the incidence of hyperthyroidism (3). Furthermore, some researchers have reported increased occurrence of hypothyroidism after IF (4–6). Still, little is known about the magnitude and possible reversibility of this increase in hypothyroidism. Also, the role of the prefortification level of ID in relation to a specific increase in iodine intake remains to be elucidated.
Prior to 1998, ID was common in Denmark. Mild ID dominated in the eastern part of Denmark, and moderate ID dominated in the western part (7). The low iodine intake was associated with both toxic and nontoxic multinodular goiter, especially among the elderly population, and insufficient thyroid hormone production in pregnant women was suggested by an increase in sensitive TSH in late pregnancy (8–10). Consequently, voluntary IF of salt was cautiously (8 ppm) initiated in July 1998 (11). The IF program aimed to increase the average daily iodine intake by 50 µg/d. However, this was ineffective, and in July 2000 the iodization of bread and household salt was made mandatory at the level of 13 ppm (11, 12). An iodine monitoring program, The Danish Investigation of Iodine Intake and Thyroid Disorders (DanThyr), was established to monitor the implementation of salt iodization in Denmark (7).
The influence of IF on the incidence rate of hypothyroidism in Denmark has been described for the early years of IF (4). In the current study, we performed a long-term follow-up on the incidence of hypothyroidism in two areas of Denmark with previous mild ID and moderate ID. The follow-up time for the area with moderate ID was 17 years past mandatory IF, thus ending in December 2016; staff limitations caused a shorter follow-up period of 9 years in the area with mild ID, which ended in October 2008. The current study is part of the DanThyr study.
Materials and Methods
Population cohort
Two open cohorts were selected, one with moderate ID prior to IF [Aalborg city and the surrounding municipalities, n = 309,434 by 1 January 1997; median urinary iodine concentration (UIC): 45 µg/L in subjects not using iodine containing supplements, 53 µg/L if all subjects were included] (13) and one with only mild ID (located in the Danish capital Copenhagen, n = 224,535 by 1 January 1997; median UIC: 61 µg/L in subjects not taking iodine containing supplements, 68 µg/L if all subjects were included) (13). Detailed information on the composition of these two cohorts was provided yearly by Statistics Denmark (14). As part of a national structural reform, the boundaries of the Danish municipalities were restructured in January 2007. Therefore, the cohort size became smaller in the moderate ID area (n = 261,569 by 1 January 2007) but remained unaltered in the mild ID cohort.
The iodine status in the study areas changed during the study period. As part of the DanThyr study, two cross-sectional studies and a follow-up study were performed in the two cohort areas in parallel with the current study. The first cross-sectional study took place at baseline (1997 to 1998) before IF, the second study took place from 2004 to 2005, and a follow-up study to the initial cross-sectional study was conducted from 2008 to 2010. Median UICs in the Western and Eastern cohort were determined at each study [first: 45 vs 61 µg/L; second: 86 vs 99 µg/L (15); follow-up: 73 vs 76 µg/L (16)]. Thus, a small decrease in iodine intake level was observed during late mandatory IF. This decrease was likely caused by a diminished iodine intake from dairy products (17).
Identification of new cases of hypothyroidism
The register database used in the current study and the evaluation of the database have been described in detail previously (13). A brief summary follows.
In the western cohort with moderate ID, a single laboratory at Aalborg University Hospital handled all thyroid function tests. Three laboratories covered the eastern cohort with mild ID (the laboratories at Frederiksberg and Bispebjerg Hospitals and the General Practitioners Laboratory in Copenhagen).
All general practitioners, hospital departments, and private practice specialists in Denmark have unique referral identification numbers used for laboratory services. All Danish citizens have a unique identification number in the Centralized Person Register. The Centralized Person Register numbers and the unique referral identification numbers made it possible to identify potential new cases of hypothyroidism within the selected area by monitoring results of thyroid function tests as follows. All laboratory tests for serum TSH and estimates of T4 sampled within the cohort areas were recorded in a specially designed register database (13) (Fig. 1). Potential new cases of overt hypothyroidism were identified by an elevated serum TSH (>5.0 mU/L) combined with a low serum T4 estimate. Specific reference ranges were used to evaluate low T4 according to the standards of each laboratory as described previously (13). Cases previously recorded and verified in the database as either hyper- or hypothyroid were marked as “known.” Each potential new case was then scrutinized by contacting the requesting physician or through available hospital records. Hence, we identified and verified all new cases of diagnosed hypothyroidism among individuals tested within the cohort areas.
Flowchart showing how new cases of overt biochemical hypothyroidism were identified in both cohorts. Subjects on the list of possible new cases were individually verified through contact with the physician who requested the initial blood test and through available hospital records. TBG, thyroxine-binding globulin; TT4, total T4.
Statistical methods
We adjusted for changes in sex and age composition of the cohort and calculated the standardized incidence rate (SIR) using the method of direct standardization (18). The Danish population on 1 January 2005 was used as the standard population (14). Significance to baseline SIR was calculated using the 95% confidence intervals (CIs) of the SIR ratio (SIRR) (18). The SIR is given in cases per 100,000 person-years. For statistical analysis, we used SPSS Statistics for Windows, Version 24.0 (IBM Corp., Armonk, NY).
The current study was approved by the National Committee on Health Research Ethics in Denmark and by The Danish Data Protection Agency.
Results
Overall incidence rates of diagnosed hypothyroidism
The baseline SIR of hypothyroidism in the western cohort with moderate ID was 32.9 per 100,000 person-years, compared with 47.3 per 100,000 person-years in the eastern cohort with mild ID (SIRR, 0.69; 95% CI, 0.54 to 0.90). No significant increase in incidence rate of hypothyroidism was seen in either cohort during the initial voluntary IF (mid-1998 to mid-2000). The SIR of hypothyroidism increased significantly in both cohorts after mandatory IF (Fig. 2). In the western cohort with previous moderate ID, the incidence rate of hypothyroidism increased during the extended study period and peaked in the last 3 years of the study (2014 to 2016) (SIRR to baseline, 1.50; 95% CI, 1.25 to 1.81). In the eastern cohort, the increase in incidence rate came later and was significant from 2004 to 2005 (SIRR to baseline, 1.30; 95% CI, 1.06 to 1.60) and until 2007 only. The SIR of hypothyroidism reached a maximum in 2008, although this was not statistically significantly different from the baseline level.
The standardized incidence rates of hypothyroidism per 100,000 person-y (pyar) in the cohorts with mild (dotted line) and moderate (solid line) ID. The rates are sex and age standardized to the Danish population of the year 2005. Error bars indicate the 95% CIs for the incidence rates. Voluntary IF was initiated in July 1998 with 8 ppm iodine in table salt and salt used by the food industry. Mandatory IF was initiated in July 2000 with 13 ppm iodine in all table salt and salt used for the production of bread. Stars indicate statistical significance to baseline (closed star, moderate ID; open star, mild ID). The study period for the mildly iodine-deficient cohort was concluded by the end of September 2008.
For each time period, the incidence rate of hypothyroidism was higher in the cohort with mild ID compared with moderate ID (Fig. 2).
Sex-specific incidence rates of hypothyroidism
In female subjects, the SIR of hypothyroidism prior to IF was 53.8 and 72.2/100.000/y in the area of moderate and mild ID, respectively. SIR increased in female subjects in the mild ID cohort, reached statistical significance by the years 2002 to 2003, and seems to have reached a plateau. Thus, SIRR to baseline was 1.29 (95% CI, 1.02 to 1.62) in 2004 to 2005 and was 1.27 (95% CI, 1.01 to 1.61) in 2006 to 2007 (Fig. 3). The increase in SIR among female subjects in the moderate ID cohort was not statistically significantly until the last years of the study period (2014 to 2016; SIRR, 1.31; 95% CI, 1.06 to 1.63) (Fig. 3).
Sex-specific incidence rates with 95% CI of hypothyroidism per 100,000 person-y (pyar) among men (M) and women (F) for both cohorts. Stars indicate statistical significance from baseline (closed star, moderate ID; open star, mild ID). Incidence rates were standardized for age.
In male subjects, the SIR of hypothyroidism at baseline was 11.4 vs 21.8/100.000/y (moderate vs mild ID). The incidence rate of hypothyroidism increased after initiation of mandatory IF in the cohort with moderate ID (Fig. 3). The increase was statistically significant from the years 2001 to 2002 onward. Peak value was reached in 2012 to 2013 (SIRR, 2.54; 95% CI, 1.66 to 3.90) (Fig. 3). The incidence rate seen in men in the mild ID cohort did not change significantly.
Age-specific incidence rates of hypothyroidism
The incidence rates of hypothyroidism were highly correlated with age in both cohorts throughout the study period.
In the western cohort with previously moderate ID, the baseline SIRs of the young (20 to 39 years), middle-aged (40 to 59 years), and elderly (≥60 years) subjects were 14.5, 30.8, and 92.2/100.000/y, respectively (Fig. 4a). The incidence rate of hypothyroidism in the young subjects rose continuously until 2012 to 2013, when it leveled off (SIRR, 2.80; 95% CI, 1.67 to 4.70) compared with baseline. The middle-aged individuals from this cohort also experienced an early increase in SIR of hypothyroidism, reaching a peak value by 2014 to 2016 (SIRR, 2.16; 95% CI, 1.55 to 3.01) compared with baseline (Fig. 4a). The changes in the elderly subjects were nonsignificant throughout the study period.
Age-specific incidence rates of hypothyroidism per 100,000 person-y (pyar) with 95% CI in the age groups 20 to 39, 40 to 59, and ≥60 y for the areas of moderate ID (a) and mild ID (b). Stars indicate statistical significance from baseline (closed star, moderate ID; open star, mild ID). Incidence rates were standardized for age and sex. IR, incidence rate.
In the eastern cohort with previous mild ID, the baseline SIRs of the young, middle-aged, and elderly subjects were 16.4, 42.2, and 148.3/100.000/y, respectively (Fig. 4b). After mandatory IF, SIR rose in young and the middle-aged subjects. It peaked in the younger age group by 2008 (SIRR, 2.62; 95% CI, 1.36 to 5.04) compared with baseline. The elevation in SIR among the middle-aged subjects was present from 2004 to 2007, with a peak value in 2006 to 2007 (SIRR compared with baseline, 2.19; 95% CI, 1.49 to 3.23). The SIR decreased by the end of the study period for the middle-aged subjects. The fluctuations seen in the elderly subjects during the study period showed no systematic change in incidence after IF.
Discussion
Principal observations
Implementation of IF in Denmark was followed by a marked increase in the overall incidence rate of diagnosed hypothyroidism. Peak incidence rates were reached by 2012 to 2016 in moderate ID and by 2004 to 2008 in mild ID. Significant increases in the incidence rates of hypothyroidism were seen in women from both cohorts after introduction of mandatory IF, whereas only men from the moderate ID cohort responded to IF with an increased incidence rate. The increased incidence rate of hypothyroidism was highly age dependent and was observed only in the young and middle-aged subjects from both cohorts. The incidence rate of hypothyroidism decreased by 2008 to a level not statistically significantly different from baseline level in middle-aged women in the mild ID cohort.
Comparison with other studies
A high occurrence of hypothyroidism has been widely reported in populations with a high stable iodine intake (19–22). Szabolcs et al. (19) investigated the prevalence of overt and subclinical hypothyroidism in three areas in Slovakia and Hungary with different urinary iodine levels (median UIC, 72, 100, and 513 µg/g creatinine, respectively) and found prevalence rates of overt hypothyroidism of 0.8%, 1.5%, and 7.6%, respectively (prevalence of subclinical hypothyroidism: 4.2%, 10.45%, and 23.9%).
A large cross-sectional study dealing with the correlation between iodine intake and the prevalence of thyroid disorders was published in 2016 by Shan et al. (20). Iodine excretion and thyroid disease frequency was surveyed in 15,008 subjects from 10 different Chinese cities after a 16-year period of varying mandatory salt iodization with an initial excessive iodine intake that was adjusted several times over the years to approach adequate iodine intake. Six cities were categorized as having adequate iodine intake (median UIC between 100 and 200 µg/L), and four cities had more than adequate iodine intake (median UIC between 200 and 300 µg/L). A higher prevalence of both overt and subclinical hypothyroidism was reported in the cities with more than adequate iodine intake compared with those with just adequate intake (prevalence rates: overt hypothyroidism, 1.3% vs 1.0%; subclinical hypothyroidism, 22.6% vs 12.7%).
Jeon et al. (21) investigated 6564 male and female participants of all age groups. The overall median UIC was 299.3 µg/L (interquartile range, 158.8 to 699.8 µg/L). The subjects were divided into six groups based on their UIC, ranging from moderate/severe ID to extremely excessive iodine excretion. In the iodine-excessive groups, the sensitive TSH levels were right-shifted when compared with those in the adequate iodine intake group, whereas the TSH levels were left-shifted in the iodine-deficient groups. This supports the correlation between the frequency of thyroid failure and iodine intake over a wide range (21).
The results of the current study are in accordance with the positive association found between the frequency of thyroid failure and the magnitude of iodine intake among different populations (19–22). Prior to IF, we found a 44% higher incidence rate of hypothyroidism in the eastern cohort with mild ID compared with the western cohort with moderate ID. It seems that very small differences in iodine intake influence the occurrence of hypothyroidism even within the ID range (23).
An increase in the occurrence of hypothyroidism has been reported from other areas undergoing iodine fortification (22, 24, 25). The Pescopagano survey of voluntary IF of salt in Italy described a markedly higher prevalence of clinical and subclinical hypothyroidism in the study population after 15 years (2.8% in 1995 vs 5.0% in 2010) (25). This was mainly due to an increased frequency of subclinical hypothyroidism in children (<15 years). In addition, they reported an increased prevalence of thyroid autoantibodies by the year 2010 compared with 1995 (19.5% vs 12.6%) (25).
An increase in median TSH (1.30 to 1.51 mU/L) was reported from the first cross-sectional study within the cohort areas at baseline (1997 to 1998) to the second cross-sectional study after implementation of mandatory IF (2004 to 2005) (26). This increase was in part caused by fewer cases of suppressed TSH but was also due to an increase in autoimmune thyroid failure, as suggested by the increased presence of thyroid autoantibodies (TPO-Ab and/or Tg-Ab) (27). An increase in the prevalence of thyroid autoantibodies in the population was evident from the first to the second cross-sectional study (TPO-Ab, 14.3% vs 23.8%; Tg-Ab, 13.7% vs 19.9%) (27). This increase was most pronounced in young adults who also had the highest increase in the incidence of hypothyroidism in the current study.
Monitoring the use of thyroid medication on a nationwide scale in Denmark before and after IF showed a marked increase in the use of thyroid hormone replacement therapy in areas with previously moderate and mild ID (28). This is in keeping with our findings of increased occurrence of thyroid failure in both cohort areas.
Prior to IF in Denmark, 50% of patients diagnosed with spontaneous autoimmune thyroid failure were 68 years or older in the two cohort areas of the current study (29). We found an increase in the incidence of hypothyroidism among only the young and middle-aged subjects. Evidently the mechanism responsible for increasing the incidence rate of hypothyroidism after introduction of IF does not significantly affect elderly patients.
The slight decrease in incidence rate of hypothyroidism observed in our cohort with previous moderate ID during the years 2008 to 2011 could be explained by a likely decrease in median UIC found during this period (from 86 µg/L in 2004 to 73 µg/L in 2008) (17). Likewise, the iodine intake in the western cohort may have risen during the following years, thus explaining the observed increase in incidence rate toward the end of the study (2012 to 2016). The incidence rate within the western cohort in the final years (2014 to 2016) is remarkably similar to the incidence rate observed in the eastern cohort at baseline value prior to IF. These similar incidence rates of hypothyroidism could result from equal median UIC between these two areas at the time periods specified. Newer studies of iodine excretion in the Western cohort are needed to confirm this.
Possible mechanisms involved
From the cross-sectional studies performed in parallel with our current cohort study as part of the DanThyr study complex, we have solid evidence that the prevalence of TPO antibody positivity increased significantly 4 to 5 years after the implementation of mandatory iodine fortification (27). The increase in prevalence of TPO-Ab positivity was most pronounced among the younger age groups and least among the elderly patients (27). Results from the same cross-sectional study clearly indicate a distinct association between elevated TSH and TPO-Ab level (27, 30). This is in accordance with the findings of the current study because the younger age group experienced the greatest increase in incidence rate of hypothyroidism, which is as expected given the substantial increase in the TPO-Ab prevalence of this age group. These results suggest that thyroid autoimmunity was a main cause of the increase in hypothyroidism incidence found in the current study (27). This is in keeping with other studies suggesting autoimmunity as a possible explanation for the increased occurrence of hypothyroidism after initiation of IF (5, 31–33). In addition, auto-regulatory processes within the thyroid gland, possibly involving the production of iodine containing derivates of arachidonic acid in response to increased iodine intake as a means to protect against an iodine overload, may be involved in inducing hypothyroidism (34).
Denmark has seen an increased focus on thyroid dysfunction over the course of our study period. Increased diagnostic activity could result and lead to more patients with hypothyroidism being identified as the study progressed. Indeed, substantial reductions in the prevalence of undiagnosed overt hypothyroidism were observed in both study areas in the cross-sectional cohorts after the initiation of IF (26). During the current study, we found a substantial increase in the incidence of TSH measurements performed within both areas. However, we cannot easily separate diagnostic activity from increased monitoring of patients with known thyroid disease.
The impact of IF on the distribution of the nosological subtypes of hypothyroidism remains to be elucidated and may contribute to explaining the mechanisms responsible for at least part of the increase in hypothyroid cases seen in this study.
Implications for IF programs
A small increase in the daily iodine intake of 50 µg or less caused increases of 30% and 50% in the incidence rate of hypothyroidism in subjects with mild and moderate ID, respectively. It seems that IF even at this low level induces increased hypothyroidism in moderately and mildly iodine-deficient populations. Thus, a cautious approach to IF in an iodine-deficient population is recommended. Specific attention may be given to the young and middle-aged subjects when implementing IF because the overall increase in incidence of hypothyroidism seemed to exclusively affect these age groups, whereas the incidence rate of hypothyroidism among elderly subjects seemed to be unaffected in both cohorts.
Strengths and limitations
Our inclusion of all thyroid function tests among general practitioners, specialists with private practice, and hospital departments reduced the referral biases frequently associated with studying the development in incidence rates of thyroid disease before and after IF (35). The present study thus ensures the inclusion of subjects with all degrees of overt biochemical hypothyroidism.
Our methods do not provide information about the duration or severity of the biochemically hypothyroid cases identified in the two cohorts. Some of the increase in incidence rate of diagnosed hypothyroidism may relate to mild or transitory cases of hypothyroidism. This is an important topic for future studies.
The relatively low frequency of hypothyroidism in areas with ID necessitated the inclusion of a substantial number of subjects in the cohorts to gain the statistical power for analysis. As such, not all subjects within the cohorts underwent thyroid function measurements each year given the resources required for such a task. The incidence rates described in the current study specifically refer to the incidence of diagnosed hypothyroidism and do not include potential cases of hypothyroidism where no thyroid function tests have been performed.
Treatment of subclinical hypothyroidism may have occurred more frequently as the study progressed. This would lead to an underestimation of the true level of hypothyroidism caused by IF because initiation of thyroid hormone substitution in subjects with subclinical hypothyroidism may cause their thyroid hormone levels to remain within the reference range.
Conclusion
Mandatory iodine fortification increased the incidence rate of hypothyroidism in both mildly and moderately iodine-deficient parts of Denmark in less than 4 years. This elevated level was present 17 years after mandatory IF. Further studies are needed to clarify the subtypes, severity, and duration of hypothyroidism after the introduction of IF.
Abbreviations:
- ID
iodine deficiency
- IF
iodine fortification
- SIR
standardized incidence rate
- SIRR
standardized incidence rate ratio
- UIC
urinary iodine concentration
Acknowledgments
Disclosure Summary: The authors have nothing to disclose.
References
World Health Organization. Assessment of iodine deficiency disorders and monitoring their elimination: a guide for programme managers. Third edition. Available at: www.unicef.org/ukraine/2_Guide_for_IDD_managers_eng.pdf. Accessed 6 October 2017.
Jensen HG, Strube M. Bekendtgørelse Om Tilsætning Af Jod Til Husholdningssalt Og Salt I Brød Og Almindeligt Bagværk M.v. Glostrup, Denmark: Foedevarestyrelsen, 1998.
Boyle P, Parkin DM. Statistical methods for registries. https://www.iarc.fr/en/publications/pdfs-online/epi/sp95/sp95-chap11.pdf. Accessed 20 November 2017.
