Isfahan Thyroid Cohort Study (ITCS)

1Isfahan Endocrine and Metabolism Research Center, Isfahan University of Medical Sciences, Isfahan, Iran 2Department of Epidemiology and Biostatistics, Isfahan University of Medical Sciences, Isfahan, Iran 3Isfahan Clinical Toxicology Research Center, Isfahan University of Medical Sciences, Isfahan, Iran 4Department of Radiology, Isfahan University of Medical Sciences, Iran 5Department of Engineering and Mathematics, Sheffield Hallam University, Sheffield, UK


Introduction
Thyroid dysfunctions are the most common endocrine disorders after diabetes. 1 Thyroid hormones regulate differentiation, growth, and metabolism throughout life. 2 These hormones play important roles in regulating carbohydrate and lipid metabolism, 3,4 neural development and occurrence of cognitive impairment disorders such as Alzheimer's disease (AD). 5 Thyroid hormones also affect cardiovascular function, 6 and thyroid disorders are proved to be associated with heart failure, coronary artery disease, and even cardiovascular mortality risk. 7,8 The prevalence and pattern of thyroid disorders vary substantially across countries. 8,9 A study of congenital hypothyroidism (CH) in Isfahan (launched in 2002) reported permanent and transient rates of 1:748 and 1:1114 births, respectively, 10 while the incidence of permanent CH in Greece 11 and Saudi Arabia 12 were reported to be 1:1800 and 1:1400 births, respectively.
The prevalence rate of thyroid disorders depends on ethnicity, geographical regions, autoimmunity, genetic, and environmental factors. One of these environmental factors is iodine intake status. Isfahan is considered as an iodine replete area following the national salt iodization in 1994. 8 There is a very limited number of cohort studies that have investigated thyroid disorders in Iran, 13 and the Isfahan Thyroid Cohort Study (ITCS) is one of the few populationbased epidemiological studies, focusing on the prevalence and incidence of thyroid disorders. The Isfahan Thyroid Cohort Study has been carried out in two phases. The first phase aimed to investigate the prevalence of thyroid disorders, including hypothyroidism, hyperthyroidism, goiter, nodule, and iodine status, and to determine whether iodine intake had any effect on the prevalence of thyroid disorders in the community, subsequent to 15 years of salt iodization. Furthermore, ultrasonographic evaluation of thyroid was performed in a small portion of the population to assess volume, size, and echogenicity of nodules. The second phase investigated the incidence of thyroid dysfunction in those participants who were euthyroid in the first phase in 2006. The role of thyroid peroxidase antibody (TPOAb) and thyroid-stimulating hormone (TSH) in predicting thyroid dysfunction was also assessed in the second phase. The Isfahan Thyroid Isfahan Thyroid Cohort Study (ITCS) Cohort Study was carried out in the Isfahan Endocrine and Metabolism Research Center (IEMRC).

Study Participants
Isfahan is centrally located in Iran. It is the third-largest city of the country and is 1590 meters above sea level. The Isfahan Thyroid Cohort Study was launched in 2006 when the estimated population of Isfahan was around 2 million. Participants were included in the study if they were (a) residents of Isfahan, (b) Iranian citizens, (c) older than 20 years, and (d) willing to provide informed consent. A sample size of 2511 was considered based on a prevalence rate of 3% for hypothyroidism, 5% type one error rate, with a precision of 0.7%, after accounting for a maximum nonresponse rate of 10%. 8 The participants were selected using multistage cluster sampling. Forty blocks were randomly selected from the city map via the post office. The mailing addresses of houses in these 40 blocks were provided accordingly. Then, 960 houses were chosen randomly (24 houses in each block) to be approached.
Twenty trained staff went door-to-door to inform and invite all adult individuals to take part in the cohort study. The leader of the group was a retired employee of the Isfahan Health Center with several years of experience in epidemiological fieldwork. They hand-picked 20 experienced health center staff and trained them for the door-to-door mission. During the face-to-face invitation, the study team explained to the potential participants which type of tests and examinations would be carried out. The objectives of the study and its contribution to healthcare decision-making were also explained.
Interestingly, one of the main reasons that many people agreed to participate in the study was to take part in improving medical care conditions in the city. In total, 2600 general residents of Isfahan aged more than 20 years old were invited, and 2523 (97% participation rate) accepted the invitation. Informed verbal consent was obtained from all individuals before their recruitment in the study (in the third phase, informed written consent is obtained) (Figure 1).
An invitation was issued for general practitioners (GPs) to participate in the project. Seven volunteers were selected and trained. The principal investigator (PI), an endocrinologist, trained GPs to examine the thyroid, search for possible nodules or goiter, and fill the checklist. An appointment was scheduled for each participant to attend a temporary cohort unit. The initial stage of examinations was established in a temporary cohort unit, with enough rooms and space to receive a large number of participants. This temporary building was located in a convenient part of the city with easy access to major transport links. The appointment session included physical examination, blood sample collection, and filling a questionnaire checklist. Ultrasound imaging was performed on two randomly selected subgroups and urine samples were collected from them. The results of the tests were mailed to participants. Subjects with abnormal results (on physical examinations and/or laboratory tests) were recalled and examined by an endocrinologist (PI) in the IEMRC for further assessment. If the PI confirmed abnormal blood test and/or physical examination results, a treatment plan was suggested, and the patient was referred to clinics or specialists for future follow-up visits.
In the second phase of the study, 723 subjects were invited to take part, of whom 711 participated in the study for follow-up (98% response rate) ( Figure 1). Participants attended the IEMRC according to scheduled appointments. The visit consisted of a physical examination and blood sample collection. In this phase of the study, a trained, experienced GP and two endocrinologists performed the physical examination, and filled the checklists.

Follow-up
Participants were evaluated and followed up by the study team for thyroid disorders, such as hypothyroidism, hypothyroidism, goiter, and nodules.
• First phase: The study started in 2006.
• Second phase: This stage of the cohort study was initiated in 2011. • Third phase: The third phase is focused on autoimmune thyroid diseases. This phase is planned for 2020-2021; recruiting started in January, but was halted temporarily due to the COVID-19 crisis. All participants of the second phase (711 participants) are considered for the third phase. We estimated more than 500 participants for this phase. The following data are being collected for this phase: demographic and general data, anthropometric data, fertility and pregnancy history, medical history, blood test, and physical examination results (Table 1).

Loss to Follow-up
Due to significant changes in phone numbers, following upgrades made to the infrastructure of the city (Isfahan), and lack of funding to take other means of contacting participants, only 723 participants could be contacted in 2011, of whom 711 agreed to be engaged with the second phase of the study, nine had passed away, and three rejected our invitation because of personal reasons. The reasons for the deaths were road accidents (n = 2), cerebrovascular accident (n = 2), colorectal cancer (n = 1), lung cancer (n = 1), and heart attack (n = 3). Although the study experienced a significant attrition rate in the second phase, it still served its purpose of investigating thyroid disorder incidences and the role of TSH and TPOAb values in predicting thyroid disorders. In order to make sure that there is no selection bias due to the loss to follow-up in the 2 nd phase, we compared the characteristics of individuals between the followed and not-followed participants at the baseline.

Phase One
The first stage included an examination and face-to-face interview of participants by trained interviewers (GPs). The GPs examined the participants' thyroid gland for consistency and possible goiter [according to the WHO and clinical criteria 14,15 and nodules. Demographic data including gender, age, and education; medical history including previous laboratory results, thyroid disorders, and medication history; and parity and menopause status of all participants were collected.
General practitioners examined all participants for possible thyroid goiter and nodules.
Blood samples were collected from all subjects. These samples were analyzed and stored at the IEMRC preceded by a centrifuge and freezing (at -20°C) in a laboratory next to the temporary cohort unit. Serum TSH was assessed by immunoradiometric assay (Kavoshyar Co., Tehran, Iran). Serum TPOAb and TgAb were measured by Rapid ELISA (Genesis Diagnostic Products Corp.). Serum T4, T3, and T3RU were measured by radioimmunoassay (Kavoshyar Co., Tehran, Iran). TSH was measured for all participants, while urinary iodine concentration (UIC) and TPOAb were evaluated for about one-fourth and one-sixth of randomly selected subjects, respectively. Around one-tenth of randomly selected participants were assessed by ultrasonography. Ultrasound imaging was carried out to assess the volume, size, and echogenicity of nodules. Height and weight were also measured for this subgroup. In the second stage of the study, the participants with a history of thyroid diseases or abnormal physical examination and TSH analysis were recalled. Blood samples were collected from these patients again to measure TSH, T4, T3, and T3RU.

Phase Two
In the second phase of the study, all assessments were conducted on all participants. One GP interviewed participants and filled a detailed checklist. Then, physical examination was carried out. The following data were collected ( Age-standardized means were considered and compared using general linear models. Binary logistic regression analysis was performed in different models to determine the association between TPOAb as predictor of thyroid dysfunction, as well as TSH levels at year 2006 (first phase) with incidence of hypothyroidism. In all analyses, relative risk (RR) and 95% confidence interval (95% CI) were reported in crude and adjusted models; adjustment was made for age, gender, smoking, BMI, positive family history, parity, abortion, age of menopause, having history of goiter or nodules for both dependent variables and being positive TPOAb, TgAb and T4 for hypothyroidism. The ability of TPOAb, TSH, T4, T3, and T3RU to predict thyroid dysfunctions was examined by the receiver operating characteristic (ROC) curve and their respective areas under the curve, in which sensitivity is plotted as a function of 1-specificity. An area under the curve (AUC) above 0.7 was considered to be useful, while an AUC between 0.8 and 0.9 indicated good diagnostic accuracy. All statistical analyses were performed using SPSS 16 for Windows (SPSS Inc., Chicago, IL, USA)

Phase Two
The second phase of the study included 338 male and 372 female subjects with a mean ± SE age of 47.28 ± 0.45 years, ranging from 25 to 82 years. The incidence of thyroid dysfunction was reported as follows: overt hypothyroidism (2.7, 95% CI: 1.6-3.7), subclinical hypothyroidism (20.6, 95% CI: 18-23), overt hyperthyroidism (1.9, 95% CI: 1-2.7) and subclinical hyperthyroidism (2.7, 95% CI: 1.6-3.7) per 1000 (person-year). 1 The results of the second phase showed significantly  higher TSH levels in females, subjects with a positive family history of thyroid disease, and those with positive TPOAb compared to other participants. The median concentrations of TSH and TPOAb in 2011 (P = 0.001) were higher than these values in 2006, especially in women. The follow-up study after 6 years revealed 16.4% new cases of positive TPOAb, while 16.1% of previously positive TPOAb cases showed negative values. 21 Aging showed no association with the risk of thyroid dysfunction, TSH levels, or an increase in the TPOAb level. However, a higher incidence of thyroid dysfunction was observed in people older than 60 years, except for subclinical hyperthyroidism. Interestingly, subjects older than 45 years with positive TPOAb had a higher risk of thyroid dysfunction. 1 The TSH levels at the baseline showed a positive association with the risk of overt hypothyroidism (RR = 3.5: 95% CI; 1.27-10.0) and subclinical hypothyroidism (RR = 8.3: 95% CI; 4.81-14.60) in the second phase. 21 One of the other objectives of ITCS was to investigate the roles of TSH and TPOAb concentrations in predicting thyroid dysfunction. ROC curve analysis was used to determine the best cutoff values of TSH and TPOAb in 2006. A cutoff value of TSH at 2.05 mIU/L (AUC [95% CI]; 0.68 [0.44-0.92]; P = 0.05) was obtained for differentiating the patients with overt hypothyroidism from euthyroidism. However, this cutoff value was unattainable when only participants with negative TPO and TgAbs were included. 21 A TPOAb cutoff value of 38 IU/mL was determined for both overt hypothyroidism and hyperthyroidism 1 ( Table 3).
A comparison was made between characteristics of the followed (participated in phase 2) and not-followed individuals (those who were in the first phase but did not participate in the second phase) and the results are presented in Table 4. No significant difference was observed between the two groups except for age which was statistically but not clinically significant.

Discussion
This is the profile of the Isfahan Thyroid Cohort Study conducted since 2006. In the first phase, we captured a picture of thyroid disorders prevalence in Isfahan, Iran and in the second phase, we focused on the incidence of thyroid dysfunction and its risk factors, especially autoimmunity. The prevalence rates of subclinical and overt hypothyroidism in Isfahan (ITCS) and Tehran (TTS) were 5.8% vs. 5.5% and 2.8% vs. 2%, respectively. 8,13 The prevalence rates (subclinical and overt) for hyperthyroidism in Isfahan and Tehran were 1% vs. 3.7% and 0.8% vs. 2.2%, respectively. Over 6 years of follow-up, both in Isfahan and Tehran, the annual incidence rates of subclinical and overt hypothyroidism were 20.6 vs. 7.62 and 2.7 vs. 2 per 1000 person-years, respectively. 1,13 The incidence rates for subclinical and overt hyperthyroidism were 2.7 vs. 3.7 and 1.9 vs. 2.2 per 1000 person-years, respectively. In Isfahan, both the prevalence and incidence of hypothyroidism are higher than Tehran and the prevalence and incidence of hyperthyroidism in Tehran are higher than Isfahan. During the period of the study, both cities were iodine sufficient for many years through fortification of salt with iodine. Iodine supplement  increases the development of thyroid autoimmunity. In the Isfahan Thyroid Cohort Study, hypothyroidism and hyperthyroidism were associated with positive TPOAb and a cutoff value of 38 IU/mL was obtained for TPOAb to predict both hypothyroidism and hyperthyroidism. 1 In the Tehran thyroid study, the optimal cutoff points of TPOAb were 18.38 and 14.77 IU/mL to predict clinical and subclinical hypothyroidism, respectively. 13 The prevalence of thyroid nodules was 2.5% and 13.2% by physical examination and ultrasonography, respectively. 17,18 There is no report on nodules in the Tehran thyroid study. It appears that autoimmunity plays a major role in both thyroid dysfunctions. The prevalence of hyperthyroidism (overt and subclinical) was 1.8% in our study, which was consistent with previous reports 22,23 in iodine-sufficient areas. It is worth noting that the ITCS was conducted 15 years after the national salt iodization program and the area was considered as an iodine-replete-environment. 9 The prevalence of thyroid incidentalomas in our population was similar to previous studies in Seoul (13.2%, vs 13.4%), 24 the three cities of Gdańsk, Sopot, and Gdynia in Poland (8%), 25 and in Taipei (18.5%). 26 In 2011, the data obtained from the first phase, combined with the data collected from the second phase were used to estimate the incidence of thyroid disorders in our population. The annual incidence of overt and subclinical hypothyroidism was 2.7 and 20.6 per 1000, respectively. The incidence of overt and subclinical hyperthyroidism was 1.9 and 2.7 per 1000 person-years, respectively. This showed that the incidences of subclinical hypo-and hyperthyroidism were higher than overt hypoand hyperthyroidism in both genders after 6 years of follow-up. Moreover, we observed a positive association between almost all types of thyroid dysfunctions and positive TPOAb in various models. The median of TPO and Tg autoantibody levels were raised in various thyroid dysfunctions 1 . Accordingly, we are conducting the third phase of the ITCS in order to study the associations between TPOAb level and environmental factors such as infection. This stage of the cohort study started in January 2020, and has been paused since March due to the COVID-19 outbreak. The third phase will be resumed as soon as the crisis is managed.

Strengths and limitations Strengths
• This cohort study is one of the few population-based studies of thyroid in Iran. The Tehran Thyroid Study 13 was conducted in the East of Tehran, whereas the Isfahan Thyroid Cohort Study included all areas of the city. • A large number of participants were recruited with a 97% participation rate in the first phase. The study population was statistically appropriate to represent the general population. • First-hand and original data was assessed in this study instead of previously recorded medical history; the main data was collected by physical examinations and laboratory tests. • Physical examinations and TSH measurements were performed for all participants.

Limitations
• Due to the shortage of funds, ultrasound imaging was performed and TPOAb concentration was measured for only a small subgroup of the subjects in the first stage. • Another limitation that this study faced was that some information, including smoking, parity, gravity, and family history of thyroid disorders, were not recorded in the first phase. • Significant attrition occurred in the second phase of the study; however, it had no negative effects on the findings. In order to assess possible selection bias in the phase 2 due to the loss to follow-up, we compared the characteristics of the followed and not-followed individuals at the baseline. All characteristics show no statistically significant difference, but the age shows a difference of two years, which is also insignificant clinically.