Prevalence of Goiter and Thyroid Nodules before and after Implementation of the Universal Salt Iodization Program in Mainland China from 1985 to 2014: A Systematic Review and Meta-Analysis

Objectives We comprehensively estimated the prevalence of goiter and thyroid nodules (TNs) before and after the implementation of the Universal Salt Iodization (USI) program in mainland China and provided information for creating effective health policies. Methods PubMed, Google Scholar, CNKI, Chinese Wanfang and Chongqing VIP databases were searched for relevant studies from Jan 1985 to Feb 2014. Data from eligible citations were extracted by two independent reviewers. All analyses were performed with Stata 11.0 and SPSS 17.0. Results Eligible articles (N = 31; 4 in English and 27 in Chinese) included 52 studies (15 about goiter rates made before 1996 and 14 afterwards, and 23 about TNs). Our meta-analysis suggests a pooled prevalence for goiter before and after 1996 and for TNs of 22.8% (95% CI: 15.3%, 30.3%), 12.6% (95% CI: 9.4%, 15.8%) and 22.7% (95% CI: 18.3%, 27.0%), respectively. Egger's test of three independent categories revealed no evidence of publication bias (p = 0.101, 0.148 and 0.113, respectively). Conclusions The prevalence of goiter was reduced by almost half after 1996 in mainland China, so the USI program was considered beneficial. However, subgroup analysis suggests that both insufficient and excess iodine may be associated with goiter. The prevalence of goiter and TNs increased significantly after 2002, suggesting a risk of excessive iodine intake. Thus, salt iodization standardizations should be set according to local conditions.


Introduction
Iodine is an essential trace element required for the normal thyroid hormone activity, specifically that of thyroxine and triiodothyronine. Both insufficient and excessive iodine intake can cause thyroid-hormone disorders [1][2][3][4] and the presence of goiter and thyroid nodules (TNs) represents thyroid diseases. The term goiter describes the enlargement of the thyroid gland and the goiter prevalence of this condition is considered an important and sensitive long-term indicator of iodine intake [1,5]. The prevalence of goiter in school-age children is related to the severity of iodine deficiency. For example, a prevalence of 0.0-4.9% suggests no iodine deficiency; 5.0-19.9% indicates mild deficiency; 20.0-29.9% reveals moderate deficiency; and severe deficiency is observed at $30% [1]. Excessive iodine, however, can also lead to goiter [1].
China was once an iodine-deficient country; prior to the 1970s, 370 million people lived in iodine-deficient areas [15]. Iodine deficiencies manifested various ways, including goiter, cretinism, endemic mental retardation, and decreased fertility rates [1], all of which are generally classified as iodine deficiency disorders (IDDs) [1,16]. To reduce IDDs, in 1979, a program of local iodine fortification was introduced into these iodine-deficient areas; and in 1996, China launched the Universal Salt Iodization (USI) program [17]. Although the USI program reduced the prevalence of goiter, the median urinary iodine concentration (UIC) in school-age children simultaneously rose sharply-reaching 330 mg/L in 1997 and 306 mg/L in 1999. For this reason, in 2002, national standards for iodized salt were revised to reduce the iodine concentration at the production level [5]. In 2012, global data identified China as a region with more than adequate iodine intake [18]. Meanwhile, since the implementation of the USI program, a growing number of Chinese clinical endocrinologists have reported an increasing incidence of thyroid diseases, especially in recent years [19][20][21][22][23][24][25].
China has diverse environments with varied populations and socio-economic conditions, complicating unified epidemiological investigations for prevalence of goiter and TNs, most of which were limited to specific geographic areas or populations. Thus, data may not accurately represent these two diseases especially in light of the implementation of the USI program which may skew data for the epidemiology of goiter and TNs. Thus, to formulate appropriate local public health policies and  criteria, we must document the epidemiology and distribution of both diseases and use the few previous studies to provide comprehensive analyses of detailed explorations and secular trends, geographic properties, and iodine status.
Here, we describe a systematic review and meta-analysis of the prevalence of goiter and TNs before and after implementation of the USI program in mainland China from 1985 to 2014.

Search Strategy
Goiter and TNs are independent diseases and iodine intake in China changed markedly after 1996 after the implementation of USI. Therefore, we conducted three meta-analyses-one on the prevalence of goiter before 1996 and one after 1996 and one analysis of the prevalence of TNs. We searched all English-  language reports of population-based studies on the prevalence of goiter and TNs using PubMed and Google Scholar, and searched all Chinese reports manually and on-line using the CNKI (Chinese National Knowledge Infrastructure), Chinese Wanfang and Chongqing VIP databases, from Jan 1985 to Feb 2014. Search key words were ''goiter,'' ''thyroid nodule(s),'' ''thyroid disorder(s),'' ''epidemiology'' and ''prevalence(s)''. We also scanned relevant reference lists and reviews to find additional studies. Attempts were made to contact authors of the identified papers for necessary details not given in the original texts. The PRISMA guideline for systematic reviews and meta-analyses was followed closely [26], see Checklist S1.

Inclusion and Exclusion Criteria
The following inclusion criteria were adopted for paper selection: (1) Data were acquired through population-based studies instead of hospital-based studies; (2) study participants were recruited from a random community-based sample, rather than from volunteers or those receiving routine-health examinations, and the subjects had to be living in mainland China; (3) the studies contained sufficient information to conduct pooled analysis of the prevalence; (4) if the same study data were reported in both English and Chinese, the English publication was included.
Studies were excluded if: (1) they were reviews or case reports; (2) the participants suffered from any related diseases or took medicines known to affect thyroid structure or function; (3) the study focused on participants with one underlying condition (such as pregnant women and smokers) or a certain occupation; (4) they were duplicate publications.

Data Extraction
The literature was searched independently by two reviewers. Any discrepancies between the extracted data of the two reviewers were reconciled through discussion. The literature-search process is shown in Figure 1. For all included studies, the first author's name, publication date, study year, age, location, sample size, events data and prevalence were recorded.

Iodine Status
Median UIC is recognized as the most practical biochemical marker for iodine nutrition [27]. Using the criteria from WHO/ UNICEF/ICCIDD [1], the iodine status was identified based on the median UIC, where the median UIC #99 mg/L is insufficient, 100-199 mg/L is adequate, 200-299 mg/L is more than adequate and a median UIC $300 mg/L is considered to be excessive.

Statistical Analysis
The pooled prevalence and 95% confidence intervals (CIs) were used to estimate the prevalence of goiter and TNs in mainland China. All meta-analyses were calculated for heterogeneity using the Chi-square based Q test and the I 2 test (25, 50 and 75% were considered low, moderate and high levels of heterogeneity, respectively) [28]. For a moderate or high level of heterogeneity we adopted a random-effects meta-analysis rather than using a fixed-effects model. Addressing heterogeneity and performing a secondary analysis required subgroup analysis. Publication bias was estimated through Egger's test. A p-value less than 0.05 indicated statistical significance. Meta-analyses were carried out in Stata Version 11.0 (Stata Corp LP, TX). Differences in prevalence among different groups were analyzed using the Chi-square test with SPSS Version 17.0 (SPSS Software, Chicago, IL).  Table 1 provides a descriptive summary of these studies.

Characteristics of Papers
All studies were based on general population samples. Before 1996, 436,182 people met the inclusion criteria for goiter; the criteria for diagnosis were in line with those of WHO, which was based on palpations. Another 14,359 people met the criteria for goiter after 1996. Goiter diagnosis after 1996 was based on palpation or ultrasound. After a diagnose with ultrasound, 59,098 people met the inclusion criteria for TNs.

Analysis of Heterogeneity and Publication Bias
We noted a significant overall heterogeneity within the studies (p,0.001, I 2 = 97.4-100.0%), which decreased through subgroup analyses. Egger's test for the three independent research categories revealed no evidence of publication bias (p = 0.101, 0.148 and 0.113, respectively).

Discussion
We examined 52 epidemiological research studies covering 14 provinces, municipalities and autonomous regions in China. We systematically analyzed the prevalence of goiter and TNs prior to and after 1996, when China implemented the USI program. To our knowledge this is one of the first studies of this kind. Research indicates that the prevalence of goiter in children 8-10 years-of-age is an indicator of local iodine consumption [1]. Our analysis, however, focused on the general population, to provide epidemiological information concerning the disease itself. Iodine nutrition is certainly an important factor for goiter, and the association between iodine intake and goiter prevalence has been investigated extensively.
Indian researchers reported that the prevalence of goiter was 65.9% in an iodine-deficient area [59], which decreased to 27.7% two decades after the USI program [60]. In Denmark, the goiter prevalence was 13.4% in an area with mild iodine deficiency and 19.7% in an area with moderate iodine deficiency, indicating that even a modest increase in iodine intake might significantly reduce IDDs [61]. In China, the USI program was launched in 1996 to prevent IDDs, establishing an important time boundary for our study. We observed a trend similar to the study in India. The overall goiter prevalence was 22.8% prior to 1996 and this was halved (declined to 12.6%). Thus, the USI program might be beneficial in China.
Also, iodine excess may increase thyroid volume [1]. In our subgroup analyses, both before and after 1996, the prevalence of goiter significantly changed with iodine status, with data suggesting hormesis-emphasizing that overdoses of iodine, as well as deficiencies may be associated with high prevalence of goiter. Meanwhile, previous studies suggest a goitrogenic effect of excess iodine in school-age children and adults [62][63][64][65][66][67]. In areas with high iodine content in the drinking water, schoolage children had elevated median UIC and endemic goiter, suggesting that, in addition to adjusting the iodine content of salt, improving water quality is also necessary [62,63]. In a prospective study, 10 normal men accepted daily oral intake of excess iodine for 1 month. At the end of this treatment, their mean thyroid volume increased, which eventually returned to baseline within 4 weeks after iodine withdrawal [64]. LeMar reported that excess iodine from tetraglycine hydroperiodide tablets caused a reversible thyroid stimulating hormone (TSH)dependent thyroid enlargement [65]. For Peace Corps volunteers, after the removal of excess iodine from water, the mean serum iodine declined sharply, and the goiter rate fell from 44 to 30% [66]. In animal models, excessive iodine intake has been shown to lead to thyroid enlargement [67]. Thus, iodine excess should be scrutinized. In agreement with this concern, we observed that after 1996, the pooled goiter prevalence increased from 10.7% before 2002 to 25.1% afterwards, indicating an impending problem.
We also observed that altitude correlated with goiter: the plateau environment was associated with higher goiter frequency than residence in the plains and hills, a finding similar to trends in previous surveys that may be explained by deficiencies in natural iodine at high altitudes and limited educational and economic support for the people who reside there [68][69][70].
There is a longstanding controversy regarding the relationship between the prevalence of TNs and iodine nutrition because of the varied epidemiological situation across the globe. Among a Swedish population with adequate iodine, the rate of TNs was only 2.6% [71]. German data from a previously iodine deficient area revealed a 20.2% prevalence of TNs, with a smaller rate for men [72]. Mexico, which formerly was mildly iodine deficient, now consumes more than normal iodine intake and here the locally-identified TNs frequency was 19.6% [73]. Overall, the prevalence of TNs in mainland China was similar to that of other countries and regions of historic iodine deficiency and present states of iodine sufficiency. In addition, our data indicate that the prevalence of TNs after 2002 was higher than the rate prior to 2002-11.0% of those investigated from 1999-2001 were diagnosed with TNs, and this increased to 24.4% between 2002 and 2014.
Unlike goiter, which has a strong etiological relationship with iodine nutrition, the prevalence of TNs also depends on sex, age and head-and-neck radiation exposure history [74,75]. According to the present information, TNs are more common in cities than in the country, which may be explained by lifestyle choices [76,77]. Furthermore, females were at greater risk for TNs and this difference was not changed when females were further divided into rural, mixed and urban groups.
In conclusion, we report that the USI program in China successfully reduced the prevalence of goiter after 1996. However, the program might cause excesses iodine intake that may be associated with a high prevalence of goiter. The prevalence of TNs also increased over time. Also, 5% of TNs may be malignant [14], so an increased prevalence of TNs might predict more thyroid carcinoma patients. In view of this information, in 2012, China adjusted the iodized salt concentration to 20-30 mg/kg (previously 35 mg/kg) and provincial governments and health administrative departments were allowed to formulate local standards within 630% of recommended concentrations based on data for their area [78,79].
Our study has several limitations. First, variations in the quality of the selected papers exist. Uncertain data will confound investigations into potential influences on heterogeneity. Secondly, primary TNs articles did not use a completely unified diagnostic criterion, but they did diagnose with ultrasound. Third, data from qualified articles failed to cover most of mainland China provinces. Still, we included all available information about goiter and TNs from the past three decades and this report is the first to document the epidemiological status of both diseases before and after the USI program in mainland China. In addition, our work underscores the need for additional population-based studies in areas excluded from this analysis.
Our data show that the implementation of USI program is beneficial but may pose iodine risks; therefore, salt iodization standards should be set according to local conditions. Also, some epidemiology studies remain to be undertaken, and these are essential for comprehensive original data on the epidemiology and distribution of thyroid diseases. Our future work will include a national baseline study on thyroid diseases which will be implemented from 2014 to 2016 and these data will provide broad and accurate information for other researchers.