Background
Until the middle of the 1990s, Germany was considered a region with mild-to-moderate iodine deficiency. The improved iodine fortification program implemented in 1993 elevated the median urinary iodine excretion levels to the lower recommended level [
1‐
3] and reduced goiter prevalence in schoolchildren [
2].
The monitoring of iodine fortification programs is important in order to observe the benefits of iodine fortification in populations and to recognize unintended effects early. Ideally iodine deficiency disorder (IDD) prevention should result in a decrease of IDD without significant increase in the prevalence of hypothyroidism and autoimmune thyroid disorders [
4,
5]. An increase in the prevalence of hypothyroidism may already be induced by moderate increase in intake of iodine [
5,
6]. Thus, iodine fortification of salt should always be introduced cautiously.
The main consequences of long-term iodine deficiency in adults are a high prevalence of goiter, thyroid nodules, and hyperthyroidism. Data from the first cohort of SHIP (SHIP-0), which started a few years after the introduction of the efficient IDD prevention program in Germany, demonstrated a high prevalence of goiter, thyroid nodules, and hyperthyroidism in the general adult population of Northeast Germany [
1]. The question arises as to whether a further decade of IDD prevention program is sufficient to observe a decrease in the prevalence of IDD in adults.
Indirectly, the improved iodine supply in Northeast Germany is mirrored by findings from the five-year follow-up examinations of SHIP, in which the normalization rate of baseline goiter was higher than its incident rate [
7]. Similar tendencies were observed for thyroid nodules and hyperthyroidism. Also, the incidence rate of positive autoantibodies to thyroperoxidase (anti-TPO Abs) was lower than its normalization rate [
7].
With long-term improved iodine supply in Germany we now aim to investigate the change in prevalence of IDD over the last decade, based on two independent cross-sectional studies. Against this background, the rationale of our study was to investigate the change in the prevalence of thyroid disorders between SHIP-0 (1997–2001) and SHIP-TREND (2008–2012). Given a stable iodine supply, we expect a reduction in the prevalence of IDD such as goiter and hyperthyroidism and a nearly stable prevalence of autoimmune thyroid disorders during the past decade. Particularly younger age groups should have benefited from the improved iodine supply.
Discussion
In the present study we investigated the effects of long-term IDD prevention by comparing the prevalence of thyroid disorders between 2000 and 2010 (7 and 17 years after the initiation of the iodine fortification program, respectively) based on population-based data from the same study region of Northeast Germany.
We demonstrate a decrease in the prevalence of goiter in men and an increase in median serum TSH levels. In contrast, the prevalence of thyroid nodules increased, but this could most likely be explained by the better resolution of the thyroid ultrasound device in SHIP-TREND than in SHIP-0 [
16]. Our data, therefore, argue for a decline of IDD during the last decade, while the prevalence of thyroid diseases related to iodine repletion such as hypothyroidism or positive anti-TPO Abs was stable or even decreased during that time period.
In contrast to the generally expected trend of subclinical thyroid disorders and, of very important note, self-reported diagnosed thyroid disorders and treatments with thyroid medication increased. This is clearly against the aims of an IDD prevention program, but the question is whether this finding can be related to the improved iodine status of the Northeast German population. We assume that a few alternative factors may explain this phenomenon.
First, an inadequate prescription of thyroxine due to therapeutic misjudgment may not be ruled out. Data from the Framingham study demonstrated a high prescription rate of thyroxine in elderly individuals for non-thyroid indications [
17]. In Europe there are also examples for misuse of thyroxine. Findings of the DanThyr study from Denmark showed that levothyroxine might have been prescribed to the wrongly diagnosed thyroid diseases such as hypothyroidism and hyperthyroidism and their subclinical forms [
18]. Second, the diagnosis of thyroid disorders depends on the definition of TSH reference intervals. For example, diagnosis of subclinical hypothyroidism critically depends on the upper TSH reference limit [
19]. As demonstrated in our study there is a substantial increase in median serum TSH levels between SHIP-0 and SHIP-TREND. In consequence, the formerly low upper reference limit established with data from SHIP-0 [
10], still applied in clinical practice, does not adequately represent the current situation and might have led to an increased rate of thyroxine prescription. Therefore, we recently established a new reference interval based on data from SHIP-TREND, which clearly indicates that the upper reference limit has to be raised substantially in our study region [
11]. Third, inappropriate treatments may be responsible for the increase in prevalence of diagnosed thyroid disorders and thyroid medication use. Therapeutic decisions may commonly be based on TSH reference values only, which according to recent guidelines is considered a common error [
20], as it neglects the symptoms and complaints of the patient [
21]. Finally, increased awareness about thyroid disorders in recent years, leading to the frequent use of thyroid laboratory tests, may have contributed to the higher prevalence of diagnosed thyroid disorders and consequently, use of thyroid medications. Additionally, the increased prescription of thyroxine for the thyroid nodules in SHIP-TREND may not be considerable, and further the data are not comparable for the different devices used in the two cohorts. However, owing to the increased prevalence of thyroid nodules in SHIP-TREND, this assumption may not be ruled out entirely.
Iodine intake in a population is the major determinant of thyroid disorders. Even a small change in the level of iodine intake in a population can change the frequency of thyroid-related disorders [
22]. The median iodine excretion levels throughout Germany in recent years are reportedly declining [
23]. The German health interview and examination survey for adults (DEGS 2008–2011) revealed a drastic decrease (61 μg/L) in iodine excretion levels, comparing the results (117 μg/L) to the previously conducted German health interview and examination survey for children and adolescents (KiGGS 2003–2006) [
3]. Though in comparison to DEGS, in our study the median iodine excretion levels decreased slightly between SHIP-0 and SHIP-TREND, but importantly without detrimental effects. Median iodine excretion levels were lower in females (particularly in younger women) than in males in both studies. This difference may be explained by the higher intake of fluids by women compared to men [
24]. A greater than normal intake of water will increase the urine volume, leading to lower urinary creatinine and lower iodine concentrations, but usually will not reduce the amount of creatinine excreted daily [
25]. Thus, the iodine/creatinine ratio adjusts for differences in fluid intake and may provide a better estimate of the true iodine intake in young females [
24,
26]. This argument was supported by the DEGS conclusions [
23]. On the other hand, there was a marginal but significant increase in median thyroid volume in women, while goiter prevalence remained nearly unchanged. In contrast, thyroid volume and goiter prevalence decreased in men. This sex-specific difference may be explained by a suboptimal iodine status in women compared to men.
When improving iodine supply there is always a risk to overcompensate, which might result in a high prevalence of autoimmune thyroid disorders. A sudden and strong increase in iodine intake may provoke the formation of thyroid antibodies [
4,
5,
27]. In our study, there is no indication of a substantial increase in autoimmune thyroid disorders parallel to the iodine fortification program, which is in line with the longitudinal findings of the SHIP cohort [
7]. We observed a decrease in the prevalence of elevated and positive anti-TPO Abs between SHIP-0 and SHIP-TREND. Previous trend studies have focused on the influence of sudden or periodic increase in iodine supply among the population and of thyroid autoimmunity [
28,
29]. In contrast, we studied the effect of consistently optimal iodine supply long-term on the same outcome. In a Chinese study conducted four years after initiation of iodine fortification, [
5] no difference in the five-year incidence rates of thyroid autoimmunity was detected. In contrast, a Danish registry study, studying all new cases of thyroid autoimmunity, four to five years after mandatory iodine fortification, demonstrated an increase in incidence (32.0 %) compared to baseline (20.0 %). However, this increase was in response to increase of iodine intake from previous moderate and mild to mandatory iodine intake in an overall cohort [
28]. Similarly, a marked increase in incidence of elevated anti-TPO Abs was observed in a study from Slovenia 10 years after the increase in iodine fortification, leading to an improvement of the iodine supply from mildly deficient to sufficient [
29].
In our study, prevalence of increased and positive anti-TPO Abs was higher in females than in males, which is in concordance with other studies [
5,
30]. In the Danish study, [
28] prevalence of positive anti-TPO Abs showed an age-dependent increase with time, which was not observed in our study, as it was highest in middle-aged (30–60 years) women. However, the age-dependent prevalence of autoimmune thyroid disorders is strongly related to the iodine repletion history or sudden increase in iodine intake in a region [
31]. Hence, the decrease in prevalence of elevated and positive anti-TPO Abs argues for an optimal but slightly declining iodine supply in our study region over the past two decades.
It is evident that the current and historical iodine supply to a region determines the distribution of TSH values in the population [
1]. If the region is iodine-deficient the TSH reference interval tends to be lower than in regions with higher iodine supply [
10,
32‐
35]. However, when a region is moving toward sufficiency from deficiency, initially the TSH distribution may shift toward the left, as was demonstrated in the SHIP-0 population [
10]. With the persistence of improved iodine supply for several years, the TSH value distribution for the general population shifts toward the right [
11]. Even the TSH reference range established during the transition phase from deficiency to sufficiency may not reflect the current situation [
11]. In our study this shift cannot be explained by different populations, sex, and current smoking, as these distribution were quite similar between SHIP-0 and SHIP-TREND [
11]. The association between serum TSH levels and BMI has been previously reported [
36,
37] and although in SHIP-TREND the median BMI was slightly higher than in SHIP-0, the detected differences in BMI might be too small to explain this magnitudinal change in TSH values [
11]. We assume that the change in TSH is mainly related to the improved iodine supply. A similar association has been previously reported for the Danish population [
4]. In contrast to the studies from the US, we could not observe an increase in TSH levels with age [
38,
39]. It may be explained by the history of iodine deficiency in our region and more than 60 years of sufficiency in the US population [
40].
Iodine repletion in populations may result in a higher prevalence of hypothyroidism [
41]. A Chinese study, also conducted in a historically iodine-deficient region after iodine fortification, showed, in agreement with our study, no increase in incidence of overt hypothyroidism. However, there was a slight increase in subclinical hypothyroidism in that study [
5]. In contrast, based on hospital registry data, a Danish study demonstrated an increase in incidence of hypothyroidism from 38/100,000
.yr to 48.7/100,000
.yr after the increase of iodine intake from moderate to mild [
6]. Differences in results between our study and the Danish study can be mainly explained by two factors. First, in the Chinese study and in SHIP, the iodine level was almost consistent throughout the study period [
5,
42], while in the Danish study a periodic increase in iodine intake was introduced [
6]. Second, in our study and in the Chinese study the TSH reference range was adapted to the changing iodine status of the study population, which was not done in the Danish study [
5,
43]. Furthermore, the Chinese study concluded that a shorter follow-up time may be responsible for non-recording of overt hypothyroidism because of its long latency period. In contrast to the Danish and the Chinese studies, which had follow-up periods of eight and five years, our follow-up was conducted 10 years after baseline [
5]. Thus, the different follow-up periods are no explanation for the different findings across these studies. In our study there is equilibrium in prevalence of hypo- and hyperthyroidism defined by increased and decreased serum TSH levels in SHIP-TREND using regional reference limits, which might reflect adequate iodine supply over the last 19 years in our study region.
Strengths of our study are the population-based design of both studies, which were conducted in the same study region one decade apart. In both SHIP studies, examinations were executed in the same laboratory with a common approach to high standardization of measurements. A limitation is that in SHIP-0 and SHIP-TREND, different thyroid ultrasound devices were used, which particularly led to a higher detection rate of thyroid nodules in SHIP-TREND. Another limitation of our study is that we have not measured thyroglobulin (Tg) antibodies in the SHIP studies, which hampers the interpretation of our findings, since an increase in iodine supply may lead to an increase in thyroglobulin (Tg) antibody concentrations [
44]. However, the slight reduction in prevalence of positive anti-TPO Abs may indicate that the prevalence of autoimmune thyroid disorders is stable in our study region. A further limitation is that we have no data on fT3 and fT4 in SHIP-TREND, making it impossible to distinct overt and subclinical forms of thyroid dysfunction. Additionally, results may be affected by selection bias as the participation rate decreased to 50.1 % in SHIP-TREND as compared to 68.8 % in the baseline cohort, though inverse probability weights were calculated for SHIP-TREND study participation and used multiplicatively to the base-weights to reduce selection bias.