Natural Course of Euthyroidism and Clues for Early Diagnosis of Thyroid Dysfunction: Tehran Thyroid Study

Abstract

Objective:

Considering the limited data available on the natural course of euthyroidism, this study was designed to evaluate the progression in time from euthyroidism to subclinical or overt hypo- or hyperthyroidism.

Methods:

This study was conducted within the framework of the Tehran Thyroid Cohort Study, in which 5783 individuals aged 40.4 ± 0.2 years were followed for six years. The overall loss to follow-up rate was 8.3%. After applying exclusion criteria, data of 4204 euthyroid subjects remained for analysis of a six-year natural course analysis. Thyroid function tests, clinical characteristics, and metabolic characteristics were assessed at baseline and every three years.

Results:

The annual incidence rates [confidence intervals (CI)] of subclinical and overt hypothyroidism were 7.62 [CI 7.39–7.85) and 2.0 [CI 1.94–2.06] per 1000 persons, respectively. For thyroid hyperfunction, the annual incidence rates of subclinical and overt hyperthyroidism were 0.92 [0.90–0.95) and 0.68 [0.66–0.70) per 1000 persons, respectively. Euthyroid persistency was 93.24% during 6 years. Predictive factors for conversion to thyroid dysfunction were thyrotropin, free thyroxine and thyroid peroxidase antibody levels, sex, and smoking. Criteria for early diagnosis of hypothyroidism (i.e., sensitivity of 94% and specificity of 82%, p < 0.0001) were obtained based on baseline and three-year follow-ups of thyroid function tests and thyroid peroxidase antibody. Early diagnosis of hypothyroidism was significantly associated with impaired glucose tolerance (relative risk with 3.03 [CI 1.36–6.75]; p = 0.007), high cholesterol (relative risk 2.46 [CI 1.45–4.18]; p = 0.001), obesity (relative risk 2.92 [CI 1.64–5.2]; p < 0.001), and hypertension (relative risk 1.68 [CI 1.53–1.84]; p < 0.04).

Conclusion:

This study shows that after a six-year follow-up in an iodine sufficient area, 6.7% of euthyroid subjects were found to progress to thyroid dysfunction, in particular subclinical hypothyroidism.

Introduction

The natural course of thyroid function in euthyroid individuals and its variability over time continues to be debated. Studies from various countries demonstrate different incidence estimations for conversion to subclinical and overt hypothyroidism, and hyperthyroidism (1). Moreover, interpretation and generalization of such estimations are difficult due mainly to variable definitions of disease states, the heterogeneous populations studied, and the historically insensitive measures of thyroid function (1). Subclinical thyroid diseases, especially subclinical hypothyroidism, are common conditions. The prevalence of subclinical hypothyroidism is reported to range between 1% and 10% and tends to be higher in the elderly, mostly in women and in those residing in iodine sufficient areas (1). Prevalences up to 12.4% have been reported for subclinical hyperthyroidism (1,2). In the Tehran Thyroid Study (TTS), the incidence rates of thyroid function abnormalities in 1000 subjects per year were reported as follows: clinical hypothyroidism 0.28 in women and 0.21 in men; subclinical hypothyroidism 11.59 in women and 4.69 in men; clinical hyperthyroidism 1.4 in women and 0.21 in men; and subclinical hyperthyroidism 5.72 in women and 3.62 in men (3).

Although studies evaluating the time course of subclinical hypo- and hyperthyroidism to overt diseases are abundant (4,5), data regarding the natural course of euthyroidism in general populations are scarce. In addition, factors that play a main role in the natural history in the transition of euthyroidism to thyroid dysfunction need to be determined. The magnitude of the complications caused by abnormal thyroid function is not entirely known. The rate of progression from the euthyroid state to thyroid disease is also a key factor in potential cost implications for health care systems, including screening programs (1,2,6 –9). Additionally, early detection and treatment of subclinical thyroid dysfunction is potentially beneficial, since some studies report the occurrence of various adverse health outcomes such as osteoporosis, hyperlipidemia, hypercholesterolemia, hyperhomocysteinemia, and cardiovascular and neuropsychiatric disorders in cases with thyroid diseases, especially among the elderly (2,6 –9).

In this large cohort study, we aimed to determine the natural course of thyroid functions over a six-year period, including rates of progression to subclinical hypo- or hyperthyroidism and overt hypo- or hyperthyroidism or the persistence of euthyroidism.

Subjects and Methods

Study design

The present study was conducted within the framework of the TTS (10), a population-based cohort study performed on residents of district 13 of Tehran with the aim of evaluating the prevalence and natural course of thyroid diseases and their long-term consequences in terms of ischemic heart disease and cardiovascular and all-cause mortality in the urban, iodine-sufficient population of Tehran, the capital of Iran. Details of the study methods have been previously described (10). The TTS is being conducted within the framework of the Tehran Lipid and Glucose Study (11), in which 15,005 residents, aged ≥3 years participated. Data collection is ongoing, designed to continue for at least 20 years, with 3-year intervals. Of 10,368 subjects aged ≥20 years, 5786 subjects who had serum samples for measuring thyroid function tests at baseline and follow-up phases were selected to participate in the TTS.

The ethics committee of the Research Institute for Endocrine Sciences (RIES) of Shahid Beheshti University of Medical Sciences approved the protocol for this study. Written informed consent was obtained from all subjects.

Settings

In brief, 5783 persons (2377 men and 3406 women) aged ≥20 years were evaluated in the TTS, at baseline and again after three and six years respectively (10,12). The cohort profile is given in Figure 1. In each phase, detailed interviews were conducted by trained interviewers. Demographic data and information regarding smoking habits, history of radioiodine ablation and thyroid surgery, and detailed information on medication use (e.g., antithyroid, thyroxine, or any medication interfering with thyroid function) were recorded.

FIG. 1.

Cohort profile of the Tehran Thyroid Study. Reasons for excluded cases are given, and the total number of euthyroid individuals for natural course analysis is displayed for each phase. Overall loss to follow-up rate was 8.3% over six years of follow-up.

Participants

The eligibility criteria for studying the natural course of euthyroidism were defined as no history of thyroid surgery or radioiodine ablation, no history of past or current use of thyroid preparations or antithyroid medications, not being under treatment with glucocorticoids, and not being pregnant at the time of recruitment. Patients lacking any eligibility criteria were excluded from analysis at each phase. Finally, 5086 (86.6%) individuals remained for the study of whom data of 3635 and 4204 subjects remained for natural course analysis after 3 and 6 years, respectively; data for 569 individuals were available only after 6 years. Information about follow-up visits has previously been described (10).

Variables

For analysis of outcomes, individuals at baseline and years 3 and 6 were categorized into four categories according to the normal reference ranges of thyroid function tests in Iranian populations (13): (1) subclinical hypothyroidism was defined as serum thyrotropin (TSH) level >5.06 mU/L with normal free thyroxine (FT4) level; (2) overt hypothyroidism was defined as serum TSH level >5.06 mU/L and FT4 < 0.91 ng/dL; (3) subclinical hyperthyroidism was defined as serum TSH level <0.34 mU/L and normal FT4 level; and (4) overt hyperthyroidism was defined as a TSH concentration <0.34 mU/L with serum FT4 concentration >1.55 ng/dL (13).

Baseline values of demographics, anthropometrics, and metabolic profiles including fasting blood glucose, two-hours postprandial glucose, lipid profile, thyroid function tests, smoking habits, and blood pressure were investigated as predicting and/or confounding variables.

Data sources and measurements

Fasting blood samples were drawn from all participants between 7:00 and 9:00 am into vacutainer tubes at each reassessment. Free thyroxine and TSH were determined on −70°C stored serum samples by the electrochemiluminescence immunoassay method, using Roche Diagnostics kits & Roche/Hitachi Cobas e-411 analyzer (Roche Diagnostics GmbH). Lyophilized quality control material (Lyphochek Immunoassay plus Control, Bio-Rad Laboratories) was used to monitor assay accuracy; intra- and interassay coefficients of variation (CVs) were 1.3% and 3.7% for FT4 and 1.5% and 4.5% for TSH determinations respectively. Thyroid peroxidase antibody (TPOAb) were assayed by an immuno-enzymometric assay using related kits (Monobind) and the Sunrise ELISA reader (Tecan Co.); intra- and interassay CVs were 3.9% and 4.7%, respectively. Total cholesterol (TC) was assayed using the enzymatic colorimetric method with cholesterol esterase and cholesterol oxidase. High-density lipoprotein cholesterol was measured after precipitation of the apolipoprotein B (apo B)-containing lipoproteins with phosphotungistic acid. Triglycerides (TGs) were assayed using glycerol phosphate oxidase. Both intra- and interassay CVs for TC, high-density lipoprotein cholesterol, and TGs were below 1.9%, 3%, and 2.1%, respectively. Fasting plasma glucose (FPG) was measured by the enzymatic colorimetric method, using glucose oxidase; inter- and intra-assay CVs at baseline and follow-up phases were both below 2.3%. All biochemical tests were performed on the day of sampling, using commercial kits (Pars Azmon Inc.) by the Selectra 2 auto-analyser (Vital Scientific).

Statistical analysis

Quantitative data are described as mean ± standard deviation for normally distributed data, based on the Kolmogorov-Smirnov test, while nonnormal variables are presented as median (interquartile range, P25–P75). Categorical data, explained by percentages, were tested by chi-squared, Fisher's exact, or Monte Carlo tests. Incidence rates of clinical hyperthyroidism, clinical hypothyroidism, and subclinical hypo- and hyperthyroidism were estimated per 1000 person-years. Also, relative risk (RR) with 95% confidence intervals have been used to describe the incidence of progression to thyroid dysfunction as compared with euthyroid persistency after three and six years of follow-up. Repeated measures multivariate analysis of variance (ANOVA) was conducted to investigate the changes of the quantitative data collected over six years of follow-up. Within-group variance of nonparametric data was assessed by Wilcoxon and Friedman tests. Between-group differences were analyzed by one-way ANOVA, Student's t-test, Kruskal-Wallis, and Mann-Whitney U tests, where indicated. RRs of outcomes linked with TSH and FT4 levels were adjusted for regression dilution bias by Rosner regression dilution ratios of the TTS (12). Receiver operating characteristic (ROC) curve analysis was also used to find the most sensitive/specific TSH/TPOAb cutoff levels for defining the natural course of thyroid function, by assuming nonparametric distributions for which the Youden index method was used, along with the corresponding sensitivity and specificity. Diagnostic odds ratios with 95% confidence intervals were calculated (positive likelihood ratio divided by negative likelihood ratio) to explain the effectiveness of diagnostic criteria, because diagnostic odds ratios are not sensitive to the low prevalence of outcomes (14). We also employed multinomial logistic regression to determine the effects of various demographic, clinical, and biochemical factors on progression to thyroid dysfunction or euthyroid persistency. Binary logistic regression was also used to determine the associations between early diagnosis of euthyroidism (based on the proposed criteria) and some metabolic outcomes. To adjust for multiple testing, the Benjamini-Hochberg method (15) was used to control the false discovery rate. All statistical analysis was carried out using the Stata statistical package program version 13 (Stata Corp.).

Results

The prevalence of thyroid dysfunction at baseline (with no exclusion criteria) was found to be 118 (2%) and 127 (2.2%) for overt hypothyroidism and hyperthyroidism, respectively and for subclinical hypothyroidism and hyperthyroidism it was 316 (5.5%) and 213 (3.7%), respectively. A total of 5086 (86.6%) individuals aged 40.4 ± 0.2 years were euthyroid at baseline. Data of excluded individuals are shown in Figure 1. After 3 years, 3400 (93.5%) individuals remained euthyroid, of whom 3920 (93.2%) showed normal thyroid status after 6 years. For thyroid hypofunction, the annual incidence rates [with CIs] of subclinical and overt hypothyroidism were 7.62 [CI 7.39–7.85] and 2.0 [CI 1.94–2.06] per 1000 persons, respectively. A total of 81 subjects who were euthyroid at baseline became subclinically hypothyroid at after 3 years and all of them returned again to an euthyroid status again after 6 years. Regarding hyperfunctioning of the thyroid, the annual incidence rates of subclinical and overt hyperthyroidism were 0.92 [CI 0.90–0.95] and 0.68 [CI 0.66–0.70] per 1000 persons, respectively. Table 1 displays the annual incidence rates and risk for progression to thyroid dysfunction.

Table 1.

Risk of Conversion from Euthyroidism to Thyroid Dysfunctions After Three and Six Years of Follow-Up

Outcomes	Incidence, N (%)	Incidence rate [CI] ^a	RR [CI]	p-Value
Three-year follow-up
Euthyroid persistence	3400 (93.54%)	–	–
Progression to subclinical hypothyroidism	155 (4.26%)	12.49 [12.08–12.89]	1.41 [1.27–1.56]	<0.0001
Progression to subclinical hyperthyroidism	38 (1.05%)	3.11 [3.01–3.21]	0.035 [0.028–0.044]	<0.0001
Progression to hypothyroidism	21 (0.58%)	1.72 [1.67–1.78]	0.008 [0.006–0.011]	<0.0001
Progression to hyperthyroidism	21 (0.58%)	1.72 [1.67–1.78]	0.014 [0.010–0.019]	<0.0001
Six-year follow-up^b
Euthyroid persistence	3920 (93.24%)	–	–
Progression to subclinical hypothyroidism	194 (4.61%)	7.62 [7.39–7.85]	1.55 [1.41–1.69]	<0.0001
Progression to subclinical hyperthyroidism	23 (0.55%)	0.92 [0.90–0.95]	0.03 [0.022–0.041]	<0.0001
Progression to hypothyroidism	50 (1.19%)	2.00 [1.94–2.06]	0.016 [0.013–0.019]	<0.0001
Progression to hyperthyroidism	17 (0.40%)	0.68 [0.66–0.70]	0.017 [0.012–0.023]	<0.0001

Euthyroid persistent group considered as reference.

Per 1000 person-years.

The incidence rates of six-year follow-up show the final outcome of individuals, so they are cumulative of incidences at three-year follow-up in cases of persistent conditions, though for cases of normalization from thyroid dysfunctions to euthyroidism, the rates are diluted.

χ², chi-squared; CI, 95% confidence interval; RR, relative risk.

Baseline demographic, anthropometric, and laboratory characteristics for the six-year natural courses of euthyroid subjects are shown in Table 2, indicating that progression to subclinical (n = 136) and overt hypothyroidism (n = 41) was more frequently observed in females (p < 0.001). While TSH, FT4, and TPOAb levels were within the normal reference range at baseline, they differed significantly between groups of natural course (e.g., hypothyroid vs. persistently euthyroid individuals) (p < 0.001).

Table 2.

Baseline Demographic, Anthropometric and Laboratory Characteristics of Euthyroid Subjects

		Outcomes of euthyroid status after six years of follow-up					Difference
Baseline characteristics	Total (4204)	Persistence (N = 3920)	Subclinical hypothyroidism (N = 194)	Subclinical hyperthyroidism (N = 23)	Hypothyroidism (N = 50)	Hyperthyroidism (N = 17)	p-Value
Sex
Male	2258 (44.6%)	1805 (46%)	53 (27.3%)	9 (39.1%)	8 (16%)	6 (35.3%)	<0.001
Female	2803 (55.4%)	2115 (54%)	141 (72.7%)	14 (60.9%)	42 (84%)	11 (64.7%)
Smoking^a
No	3647 (87.9%)	3393 (87.7%)	176 (91.7%)	18 (81.8%)	48 (98%)	12 (70.6%)	0.006
Yes	502 (12.1%)	476 (12.3%)	16 (8.3%)	4 (18.2%)	1 (2%)	5 (29.4%)
Age (years)	38.00 (29.00–51.00)	38 (28–51)	37.00 (30.00–48.00)	41.00 (30.00–52.00)	37.00 (32.00–56.00)	39.00 (32.00–46.00)	0.816
BMI (kg/m²)	26.12 (23.15–29.40)	26 (23.1–29.4)	26.23 (23.43–29.29)	25.20 (22.21–28.13)	27.72 (24.14–30.85)	27.18 (24.57–30.22)	0.393
Wrist circumference (cm)	16.60 (15.60–17.50)	16.7 (15.6–17.6)	16.30 (15.50–17.10)	16.60 (15.20–17.80)	16.70 (16.00–18.00)	16.00 (15.50–17.00)	0.003
Waist circumference (cm)	87.00 (77.00–95.50)	87 (77–95)	85.50 (77.00–95.00)	88.00 (74.00–97.00)	93.00 (79.00–101.00)	87.00 (78.00–95.00)	0.658
Hip circumference (cm)	100.00 (94.00–106.00)	99 (94–105)	102.00 (96.00–109.50)	100.00 (93.00–103.00)	103.00 (99.00–113.00)	103.00 (95.00–108.00)	0.004
Waist-to-hip ratio	0.87 (0.80–0.93)	0.87 (0.80–0.93)	0.83 (0.77–0.92)	0.88 (0.82–0.93)	0.86 (0.83–0.96)	0.82 (0.79–0.88)	0.022
TSH (mU/L)	1.60 (0.98–2.57)	1.5 (1.01–2.2)	3.59 (2.68–4.31)	0.51 (0.41–0.86)	3.32 (2.26–4.17)	1.74 (0.89–2.41)	<0.001
FT4 (ng/dL)	1.23 ± 0.24	1.23 ± 0.19	1.19 ± 0.16	1.21 ± 0.18	1.04 ± 0.15	1.28 ± 0.20	<0.001
FT4 (pmol/L)	15.82 ± 3.05	15.86 ± 2.39	15.27 ± 2.05	15.53 ± 2.27	13.43 ± 1.88	16.45 ± 2.62	<0.001
TPOAb (U/L)	5.59 (3.33–11.44)	5.1 (3.1–9.2)	11.64 (4.66–82.87)	5.22 (2.60–6.88)	14.38 (4.95–66.81)	6.48 (3.90–16.48)	<0.001

Quantitative data with normal distribution are shown as mean ± standard deviation, while nonnormal data are presented as median (interquartile range: P25–P75). Categorical data are reported as counts (percentages).

Smoking data had 1.3% random missing.

BMI, body mass index; FT4, free thyroxine; TPOAb, thyroperoxidase antibodies; TSH, thyrotropin.

Data regarding the most important predictors of natural course of euthyroid subjects are illustrated in Table 3. Based on the regression dilution ratios–adjusted RR of multivariate multinomial logistic regression, each mU/L increase in TSH was associated with a 9.32-fold greater risk of progression to subclinical hypothyroidism and a 5.40-fold increased risk of overt hypothyroidism, whereas other variables were controlled. Smoking was only a significant risk factor for overt hyperthyroidism.

Table 3.

Predictors of Natural Course of Euthyroid Individuals Based on Multinomial Logistic Regression

	Univariate model				Multivariate model
			95% CI for RR				95% CI for RR
Natural course of euthyroidism after six years follow-up	p-Value	RR	Lower	Upper	p-Value	RR	Lower	Upper
Subclinical hypothyroidism
TSH	<0.001	3.87	3.45	4.47	<0.001	3.70	3.19	4.28
Adjusted^*	<0.001	10.31	8.46	13.22	<0.001	9.54	7.39	12.27
FT4 (ng/dL)	0.001	0.23	0.10	0.54	0.863	0.92	0.36	2.38
Adjusted^*	0.001	0.076	0.018	0.339	0.863	0.86	0.17	4.58
FT4 (pmol/L)	0.001	0.89	0.83	0.95	0.863	0.99	0.92	1.07
Adjusted^*	0.001	0.82	0.72	0.91	0.863	0.98	0.86	1.13
TPOAb	<0.001	1.004	1.003	1.005	<0.001	1.003	1.001	1.004
Female	<0.001	2.27	1.65	3.13	0.065	1.44	0.98	2.13
Smoke	0.102	0.65	0.38	1.09	0.346	1.35	0.72	2.51
Hypothyroidism
TSH	<0.001	3.15	2.47	4.02	<0.001	2.70	2.07	3.52
Adjusted^*	<0.001	7.23	4.75	11.01	<0.001	5.54	3.51	8.76
FT4 (ng/dL)	<0.001	0.001	0.0001	0.0045	<0.001	0.003	0.0003	0.024
Adjusted^*	<0.001	0	0	0	<0.001	0	0	0.001
FT4 (pmol/L)	<0.001	0.56	0.48	0.66	<0.001	0.63	0.54	0.75
Adjusted^*	<0.001	0.36	0.28	0.48	<0.001	0.45	0.34	0.60
TPOAb	<0.001	1.005	1.003	1.006	<0.001	1.004	1.002	1.005
Female	<0.001	4.48	2.10	9.57	0.251	1.61	0.71	3.64
Smoke	0.059	0.15	0.02	1.08	0.382	0.40	0.05	3.14

Relative risks of outcomes linked with TSH and FT4 levels were adjusted regarding regression dilution bias by Rosner regression dilution ratios (RDR) of Tehran Thyroid Study [10] (exp(LN (RR) /RDR)).

Based on ROC analysis (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/thy), TSH cutoff values for prediction of subclinical and overt hypothyroidism after 6 years of follow-up were TSH >2.48 mU/L (sensitivity = 84%, specificity = 80%) and TSH >2.57 mU/L (sensitivity = 70%, specificity = 79%), respectively. The TPOAb cutoff values for prediction of subclinical and overt hypothyroidism were TPOAb >11.04 U/L (sensitivity = 53.6%, specificity = 79.3%) and TPOAb >12.12 U/L (sensitivity = 56%, Specificity = 80.7%), respectively. Based on the cutoffs, diagnostic criteria have been provided (Table 4) to predict progression to subclinical and overt hypothyroidism. We found that simultaneous examination of thyroid function tests followed by repeated examination after 3 years had a sensitivity of 94% and specificity of 82% for detecting hypothyroidism at the end of 6 years of follow-up. Moreover, early diagnosis of hypothyroidism was significantly associated with impaired glucose tolerance (2-hour plasma glucose ≥200 mg/dL; RR 3.033 [CI 1.36–6.75]; p = 0.01), high cholesterol (TC ≥240 mg/dL; RR 2.46 [CI 1.45–4.18]; p < 0.01), moderate risk TG elevation (TG = 200–400 mg/dL; RR 1.79 [CI 1.15–2.79]; p = 0.01), obesity (body mass index >30 kg/m²; RR 2.92 [CI 1.64–5].2; p < 0.01), and hypertension (RR 1.68 [CI 1.53–1.84]; p = 0.04).

Table 4.

Diagnostic Characteristics ^* of Thyroid Hormones' Cutoffs to Predict Overt and Subclinical Hypothyroidism at 6-Years Follow-Up

No.	Criteria for prediction	Youden index	Sensitivity	Specificity	DOR [CI]	RR [CI]	AdjRR ^a [CI]	p-Value
	Hypothyroidism
	Baseline
1	TSH ≥2.57 mU/ml	0.494	70 (65.2–74.8)	79.4 (75.1–83.7)	9.0 [4.9–16.5]	3.4 [1.9–6.2]	8.2 [2.9–23.1]	<0.0001
2	FT4 ≤ 1.07 ng/dL	0.552	74 (69.4–78.6)	81.2 (77.1–85.3)	12.3 [6.5–23.2]	3.9 [2.1–7.3]	11 [3.7–33]	<0.0001
3	TPOAb ≥12.12 U/L	0.367	56 (50.8–61.3)	80.7 (76.5–84.9)	5.3 [3.0–9.4]	2.9 [1.67–5]	No need to adjust	<0.0001
4	Simultaneous at baseline	0.486	96.6 (94.6–98.5)	52 (46.8–57.3)	–	–	–	–
	Three-year follow-up
5	TSH ≥3.46 mU/ml	0.662	78.3 (73.9–82.7)	87.9 (84.5–91.4)	26.2 [12.9–53.2]	6.4 [3.2–12.8]	24.61 [7.4–81.3]	<0.0001
6	FT4 ≤ 1.03 ng/dL	0.573	71.7 (67–76.5)	85.6 (81.8–89.3)	15.0 [7.9–28.8]	5 [2.6–9.4]	16.60 [5.5–50.6]	<0.0001
7	TPOAb ≥13.75 U/L	0.397	56.5 (51.3–61.7)	83.2 (79.3–87.4)	6.4 [3.6–11.6]	3.4 [1.9–6]	No need to adjust	<0.0001
8	Simultaneous^b at year 3	0.599	97.3 (95.6–99.1)	62.6 (57.5–67.7)	–	–	–	–
9	Sequential^c (Nos. 4 and 8)	0.760	94 (91.5–96.5)	82 (78–86.1)	71.4 [31.4–162.3]	41 [22.7–74.1]	–	<0.0001
	Subclinical hypothyroidism^d
	Baseline
1	TSH ≥2.48 mU/ml	0.638	84 (77.4–90.6)	79.8 (72.6–87.1)	20.7 [14.0–30.7]	4.14 [4.–4.3]	11.6 [11–12.2]	<0.0001
2	TPOAb ≥11.04 U/L	0.329	53.6 (44.6–62.6)	79.3 (72–86.6)	4.4 [3.3–5.9]	2.59 [2.5–2.7]	No need to adjust	<0.0001
3	Simultaneous at baseline	0.559	92.6 (87.9–97.3)	63.3 (54.6–72)	–	–	–	–
	Three-year follow-up
4	TSH ≥2.58 mU/ml	0.671	89.2 (83.6–94.8)	77.9 (70.5–85.4)	29.1 [17.5–48.5]	4 [3.9–4.1]	11 [10.5–11.6]	<0.0001
5	TPOAb ≥11.99 U/L	0.335	51.6 (42.6–60.6)	81.9 (74.9–88.8)	4.8 [3.5–6.7]	2.9 [2.8–2.9]	No need to adjust	<0.0001
6	Simultaneous^b at year 3	0.586	94.8 (90.7–98.8)	63.8 (55.1–72.5)	–	–	–	–
7	Sequential^c (Nos. 3 and 6)	0.744	87.7 (81.8–93.7)	86.7 (80.6–92.8)	46.6 [32.4–67.1]	9.9 [9–11]	–	<0.0001

Cutoff points for thyroid functions tests are extracted based in receiver operating characteristic analysis.

Adjusted (adj) RR, relative risks of outcomes linked with TSH and FT4 levels were adjusted regarding regression dilution bias by RDR of the Tehran Thyroid Study [10] (exp(LN (RR) /RDR)).

Simultaneous sensitivity = Sensitivity 1 + Sensitivity 2 − (Sensitivity 1 × Sensitivity 2); Simultaneous specificity = Specificity 1 × Specificity 2. Sensitivity 1 is related to baseline and sensitivity 2 is related to data at 3-years follow-up.

Sequential sensitivity = Sensitivity 1 × Sensitivity 2; Sequential specificity = Specificity 1 + Specificity 2 − (Specificity 1 × Specificity 2).

It is clear that the diagnostic criteria for subclinical hypothyroidism is applicable when free thyroxin levels are within normal range.

DOR, diagnostic odds ratio (positive likelihood ratio/negative likelihood ratio).

Early diagnosis of subclinical hypothyroidism was also significantly associated with impaired fasting glucose (fasting blood sugar = 100–125 mg/dL; RR 1.49 [CI 1.13–1.96]; p = 0.005), high cholesterol (TC ≥240 mg/dL; RR 1.60 [CI 1.15–2.23]; p = 0.005), and obesity (body mass index >30 kg/m²; RR 1.54 [CI 1.14–2.07]; p = 0.002).

Figure 2 illustrates the longitudinal variations of TSH, FT4, and TPOAb levels (over three TTS phases) among various outcomes of euthyroid individuals. For crude values, the significant time (p < 0.001) and time × groups (p < 0.0001) interaction indicates the different natural courses of thyroid functions and TPOAb among euthyroid individuals and thyroid disorders. After adjustment for sex, smoking status, and baseline waist to hip ratio (WHR) (baseline difference of groups), TSH, and FT4 showed significant time (p < 0.05) and time × groups (p < 0.0001) interaction. However, adjusted TPOAb levels did not differ significantly across the three phases of the TTS. Adjusted TPOAb levels showed a significant group (p < 0.001) and time × groups (p < 0.0001) interaction.

FIG. 2.

Longitudinal changes of thyrotropin (TSH) (a), free thyroxine (FT4) (c), and thyroperoxidase antibody (TPOAb) (e), indicating natural course of euthyroid individuals during six years. Data are expressed as mean [confidence interval]. Repeated measures analysis of variance revealed significant time effects and time × groups (natural courses of thyroid status) for the Tehran Thyroid Study cohorts. Repeated measures analysis of covariance (ANCOVA) shows the longitudinal changes of TSH (b), FT4 (d), and TPOAb (f) during the six-year follow-up, after adjustment for sex, smoking, and waist-to-hip ratio.

Discussion

The main findings of our study are as follows: (1) euthyroid persistency was as high as 93% during 6 years of follow-up; (2) conversion rates to subclinical hypothyroidism after 3 and 6 years were 4.3% and 4.6% respectively; (3) progression rates to overt hypothyroidism after 3 and 6 years were 0.6% and 1.2% respectively; (4) progression rates to subclinical and overt hyperthyroidism were approximately 0.6% and 0.4% respectively; and (5) criteria for early diagnosis of thyroid hypofunction were provided, which might be helpful in predicting the progression to subclinical and overt hypothyroidism.

The historical cohort of the Whickham Survey documented an incidence of thyroid disorders in a randomly selected sample of 2779 British adults to be 3.5/1000/year [CI 2.8–4.5] for spontaneous hypothyroidism of females and 0.6/1000/year [CI 0.3–1.2] for men (16). Moreover, the incidence of hyperthyroidism in women was reported to be 0.8 per 1000 person-years [CI 0.5–1.4] and was negligible in men (16). Both raised serum TSH and positive antithyroid antibodies were the paramount predictors of progression to hypothyroidism in the Whickham Survey and the study done by Guel et al. (16,17), results in line with our findings.

In addition, in the Colorado Thyroid Diseases Prevalence study, 88.3% of the total population was euthyroid, with men being more frequently euthyroid than women, revealing a prevalence similar to that of our study (86.6%), and consistent with findings in iodine-replete populations (1). The natural history of euthyroidism most probably depends on the presence of underlying autoimmunity, TSH levels (even in normal range), age, and sex; however, long-term prospective data on euthyroid individuals are rare, making it difficult to determine their definitive course (1,2,16). The results of the Amsterdam autoimmune thyroid disorders (AITD) cohort study showed that overt hypothyroidism and hyperthyroidism developed in 4.8% and 1.7% of female individuals, respectively during a 5-year follow-up of those prone to the disorders; however, the occurrence of hypothyroidism was reported only in females with family history of autoimmune thyroid diseases in this study (18). Another nested case-control study conducted within the framework of the Amsterdam AITD cohort study revealed that higher TSH and lower FT4 are predicting factors for conversion from euthyroidism to autoimmune hypothyroidism (19), results in agreement with those of the TTS.

The incidence of subclinical hypothyroidism in 3057 U.S. euthyroid individuals (≥65 years of age) has been reported to be 2.7% (84 of 3057) over a 2-year period (20). The prevalence of subclinical hypothyroidism has been reported to be 3–12% (16,21), and 1–6% for subclinical hyperthyroidism (1,2,22), findings in line with those of our study.

Studying the natural course of euthyroid individuals progressing to thyroid over-activity in large populations is crucial; to our knowledge there is no prospective data documenting this in large cohorts; the majority of studies evaluating the spontaneous changes in subclinical hyperthyroid patients report that progression to overt hyperthyroidism is rare (5,22,23). The retrospective cohort of the Thyroid Epidemiology, Audit, and Research Study showed an annual incidence of 30–40 per 100,000 population (with a median follow-up of 4.3 years) (23), an incidence different from that of our study (92 per 100,000 population during median follow-up of 6.34 years), and that of the Whickham Survey (101.9 per 100,000) (16).

Overall, determining euthyroidism or progression to thyroid dysfunction mostly depends on our definitions for TSH normal values. In the present study, the cutoff value for abnormal TSH was >5.06 mU/mL, which is the upper normal limit of TSH values in the TTS (13). Huber et al. considered a TSH level of <6 mU/L as a predictor of lower likelihood of progression to thyroid hypofunction (24). Walsh et al. showed that the risk of hypothyroidism is very low in males with a baseline TSH ≤2.5 mU/L and increases progressively across higher categories of TSH. The use of TSH limits of 2.5 and 4.0 mU/L, in the presence of thyroid antibodies provides a clinically useful estimate of the long-term risk of hypothyroidism and Walsh et al. concluded that female sex and baseline TSH were the strongest risk factors for the occurrence of hypothyroidism at follow-up (25). One prospective study reveals that only a limited number of the cohort of patients with subclinical hypothyroidism develop overt hypothyroidism over time and the majority remain in the subclinical state after 10 years (24). Hence, it is reasonable to assume that this stability in thyroid function tests is present among euthyroid subjects over long periods of time. Moreover, in the case of high serum TSH with normal FT4 levels, spontaneous normalization of TSH has been observed (26,27). A study by Rosário et al. showed that most patients with subclinical hypothyroidism and a TSH ≤10 mU/L did not progress to overt hypothyroidism and after 3 years of follow-up, spontaneous normalization of TSH levels was observed in 20% of the study population (26). However, some may say that subclinical hypothyroidism might be an independent risk factor for cardiovascular and metabolic events, requiring therapeutic intervention regardless of spontaneous normalization or persistence. The significantly higher risk of these adverse events among our patients with early hypothyroidism might support this hypothesis. Interventional studies need to assess whether therapeutic interventions are necessary for patients with early diagnosis of hypothyroidism and subclinical hypothyroidism.

There is evidence supporting our results demonstrating that TPOAb are a risk factor for progression to thyroid hypofunction (16,24,27). In the Whickham study, subjects with subclinical hypothyroidism and TPOAb negative status had at least a 2–3% annual risk of conversion to overt hypothyroidism, a rate that almost doubled (4–5%) if TPOAb were present (16). Some studies reveal that euthyroid individuals with positive TPOAb levels have at least a 2% annual rate of progression to subclinical or overt hypothyroidism, whereas antibody negative euthyroid subjects have lower rates compared with the studies mentioned above (16,17).

Our study does have some limitations. First, although a detailed history was obtained at the time of the baseline examination, drug history was self-reported. Second, the study has a follow-up period of six years, which is not that long. Moreover, this study is underpowered to investigate subclinical (n = 23) and overt (n = 17) hyperthyroidism, for which much larger numbers are needed. The study also has several strengths. First, free serum T4 was measured, so that thyroid status is more credible than if total T4 concentrations were available. Second, to our knowledge, this study is one of the few studies aiming to determine the natural course of thyroid function in a euthyroid population and third, the study has a prospective nature with follow ups every three years. We should note that, these data are of no direct clinical relevance and merely provide insight into the natural course of euthyroid thyroid function.

In conclusion, our data clearly demonstrate that over 93% of euthyroid subjects remained euthyroid over six years of follow-up and that the factors predicting progression to thyroid dysfunction are baseline TSH, FT4, TPOAb levels, sex, and smoking.

Footnotes

Acknowledgments

We would like to express our gratitude the laboratory personnel of the RIES for their assistance and to acknowledge Ms. Niloofar Shiva for critical editing of English grammar and syntax of the manuscript. This study was supported by financial grants from the Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran.

A.A. and F.Z. contributed to conception and design of trial, collection and assembly of data, interpretation of data, manuscript writing, and final approval of manuscript. M.B. and S.G. conducted the statistical analyses. M.B. and Z.G. wrote the primary draft of manuscript. L.M. and M.T. contributed to collection and assembly of data and manuscript writing. All authors contributed to and have approved the final manuscript.

Author Disclosure Statement

No competing financial interests exist.

References

Canaris

, Manowitz

, Mayor

, Ridgway

. 2000. The Colorado thyroid disease prevalence study. Arch Int Med, 160:526–534.

Hollowell

, Staehling

, Flanders

, Hannon

, Gunter

, Spencer

, Braverman

. 2002. Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab, 87:489–499.

Delshad

, Mehran

, Tohidi

, Assadi

, Azizi

. 2012. The incidence of thyroid function abnormalities and natural course of subclinical thyroid disorders, Tehran, I.R. Iran. J Endocrinol Invest, 35:516–521.

Park

, Oh

, Jeon

. 2013. Prospective observation of 5-year clinical course of subclinical hypothyroidism in korean population. J Korean Med Sci, 28:1622–1626.

Rosario

. 2010. Natural history of subclinical hyperthyroidism in elderly patients with TSH between 0.1 and 0.4 mIU/l: a prospective study. Clin Endocrinol, 72:685–688.

Biondi

, Cooper

. 2008. The clinical significance of subclinical thyroid dysfunction. Endocr Rev, 29:76–131.

Cai

, Ren

, Shi

. 2011. Blood pressure levels in patients with subclinical thyroid dysfunction: a meta-analysis of cross-sectional data. Hypertens Res, 34:1098–1105.

Singh

, Duggal

, Molnar

, Maldonado

, Barsano

, Arora

. 2008. Impact of subclinical thyroid disorders on coronary heart disease, cardiovascular and all-cause mortality: a meta-analysis. Int J Cardiol, 125:41–48.

Rodondi

, den Elzen

, Bauer

, Cappola

, Razvi

, Walsh

, Asvold

, Iervasi

, Imaizumi

, Collet

, Bremner

, Maisonneuve

, Sgarbi

, Khaw

, Vanderpump

, Newman

, Cornuz

, Franklyn

, Westendorp

, Vittinghoff

, Gussekloo

, Thyroid Studies

. 2010. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA, 304:1365–1374.

10.

Azizi

, Amouzegar

, Delshad

, Tohidi

, Mehran

, Mehrabi

. 2013. Natural course of thyroid disease profile in a population in nutrition transition: Tehran Thyroid Study. Arch Iran Med, 16:418–423.

11.

Azizi

, Ghanbarian

, Momenan

, Hadaegh

, Mirmiran

, Hedayati

, Mehrabi

, Zahedi-Asl

. 2009. Prevention of non-communicable disease in a population in nutrition transition: Tehran Lipid and Glucose Study phase II. Trials, 10:5.

12.

Amouzegar

, Beigy

, Gharibzadeh

, Azizi

. 2014. Underestimation of thyroid dysfunction risk due to regression dilution bias in a long-term follow-up: Tehran Thyroid Study (TTS). Horm Metab Res, 46:440–444.

13.

Amouzegar

, Delshad

, Mehran

, Tohidi

, Khafaji

, Azizi

. 2013. Reference limit of thyrotropin (TSH) and free thyroxine (FT4) in thyroperoxidase positive and negative subjects: a population based study. J Endocrinol Invest, 36:950–954.

14.

Glas

, Lijmer

, Prins

, Bonsel

, Bossuyt

. 2003. The diagnostic odds ratio: a single indicator of test performance. Journal of clinical epidemiology, 56:1129–1135.

15.

Benjamini

, Hochberg

. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol, 57:289–300.

16.

Vanderpump

, Tunbridge

, French

, Appleton

, Bates

, Clark

, Grimley Evans

, Hasan

, Rodgers

, Tunbridge

, et al. 1995. The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickham Survey. Clin Endocrinol, 43:55–68.

17.

Geul

, van Sluisveld

, Grobbee

, Docter

, de Bruyn

, Hooykaas

, van der Merwe

, van Hemert

, Krenning

, Hennemann

, et al. 1993. The importance of thyroid microsomal antibodies in the development of elevated serum TSH in middle-aged women: associations with serum lipids. Clin Endocrinol, 39:275–280.

18.

Strieder

, Tijssen

, Wenzel

, Endert

, Wiersinga

. 2008. Prediction of progression to overt hypothyroidism or hyperthyroidism in female relatives of patients with autoimmune thyroid disease using the Thyroid Events Amsterdam (THEA) score. Arch Int Med, 168:1657–1663.

19.

Effraimidis

, Strieder

, Tijssen

, Wiersinga

. 2011. Natural history of the transition from euthyroidism to overt autoimmune hypo- or hyperthyroidism: a prospective study. Eur J Endocrinol, 164:107–113.

20.

Somwaru

, Rariy

, Arnold

, Cappola

. 2012. The natural history of subclinical hypothyroidism in the elderly: the cardiovascular health study. J Clin Endocrinol Metab, 97:1962–1969.

21.

Aghini-Lombardi

, Antonangeli

, Martino

, Vitti

, Maccherini

, Leoli

, Rago

, Grasso

, Valeriano

, Balestrieri

, Pinchera

. 1999. The spectrum of thyroid disorders in an iodine-deficient community: the Pescopagano survey. J Clin Endocrinol Metab, 84:561–566.

22.

Parle

, Franklyn

, Cross

, Jones

, Sheppard

. 1991. Prevalence and follow-up of abnormal thyrotrophin (TSH) concentrations in the elderly in the United Kingdom. Clin Endocrinol, 34:77–83.

23.

Vadiveloo

, Donnan

, Cochrane

, Leese

. 2011. The Thyroid Epidemiology, Audit, and Research Study (TEARS): the natural history of endogenous subclinical hyperthyroidism. J Clin Endocrinol Metab, 96:E1–8.

24.

Huber

, Staub

, Meier

, Mitrache

, Guglielmetti

, Huber

, Braverman

. 2002. Prospective study of the spontaneous course of subclinical hypothyroidism: prognostic value of thyrotropin, thyroid reserve, and thyroid antibodies. J Clin Endocrinol Metab, 87:3221–3226.

25.

Walsh

, Bremner

, Feddema

, Leedman

, Brown

, O'Leary

. 2010. Thyrotropin and thyroid antibodies as predictors of hypothyroidism: a 13-year, longitudinal study of a community-based cohort using current immunoassay techniques. J Clin Endocrinol Metab, 95:1095–1104.

26.

Rosario

, Bessa

, Valadao

, Purisch

. 2009. Natural history of mild subclinical hypothyroidism: prognostic value of ultrasound. Thyroid, 19:9–12.

27.

Diez

, Iglesias

. 2004. Spontaneous subclinical hypothyroidism in patients older than 55 years: an analysis of natural course and risk factors for the development of overt thyroid failure. J Clin Endocrinol Metab, 89:4890–4897.