Body composition indices align with thyroid variation in euthyroid Saudi adults: A cross-sectional study

Abstract

Objective

Although associations between obesity and thyroid function have been widely reported, few studies have examined euthyroid (clinically normal) adults using both bioelectrical impedance-derived body composition measures and clinically accessible central adiposity markers in the same analysis, particularly Saudi adults in sex-stratified models. We examined the associations between thyroid function (evaluated using thyroid-stimulating hormone, free thyroxine, and free triiodothyronine) and body composition in Saudi adults, considering sex-specific patterns.

Methods

We conducted a cross-sectional analysis including 551 euthyroid adults (aged ≥18 years, 59% men). Anthropometric characteristics and body composition parameters were evaluated. Serum thyroid-stimulating hormone, free triiodothyronine, and free thyroxine levels were measured. Correlations and logistic regression analyses were used to assess associations. Participants were categorized as having high thyroid-stimulating hormone, low free thyroxine, or low free triiodothyronine levels for an assessment of their obesity risk.

Results

Low free thyroxine levels were associated with higher odds of obesity, as classified using body mass index (odds ratio = 1.85; 95% confidence interval, 1.09–3.10) and waist circumference (odds ratio = 2.68; 95% confidence interval, 1.21–6.20). High thyroid-stimulating hormone levels alone demonstrated a limited association to obesity risk. However, in women, high thyroid-stimulating hormone levels were strongly associated with central obesity. In men, low free thyroxine level was the more prominent predictor, associated with increased odds of elevated waist circumference. Free triiodothyronine levels were positively correlated with lean body mass and negatively with fat percentage, as indicated by higher free triiodothyronine levels in individuals with more muscle and less fat.

Conclusions

Greater adiposity was associated with higher thyroid-stimulating hormone and lower free thyroxine levels in our population. These relationships differed by sex; the association of elevated thyroid-stimulating hormone levels with adiposity was more pronounced in women, whereas a stronger inverse free thyroxine–obesity link was observed in men.

Introduction

Worldwide, obesity affects more than one billion people. ¹ In 2022, one in eight people lived with obesity, and 43% of adults were overweight. Adult obesity has more than doubled since 1990. ¹ Saudi Arabia carries a high national obesity burden. In 2024, the General Authority for Statistics released the results of the Health Determinants Statistics, according to which, the obesity rate among individuals aged ≥15 years was 23.1%, and 45.1% of individuals in this age group were considered overweight. ² A 4-years surveillance study conducted in 2024 has reported adult obesity rates of 21%–22% between 2020 and 2023. ³ Systematic reviews have concluded that national estimates generally range from approximately 20% to 39% across surveys and subgroups.^4,5

Thyroid disorders are also common in the region, although prevalence estimates vary. A recent Saudi systematic review has reported that subclinical hypothyroidism in the general adult population ranges from 10.3% among primary healthcare visitors to 15.9% in adults undergoing thyroid function tests. ⁶ A hospital-based cohort study has reported much higher figures, possibly reflecting referral bias. ⁷ The 2024–2025 evidence synthesis has also noted wide regional variations and higher prevalence among high-risk groups, such as diabetic and pregnant women. ⁶

Obesity is a core component of metabolic syndrome, a cluster of abnormalities that is associated with increased cardiovascular risk. ⁸ Recent studies on euthyroid adults have extended this field by linking thyroid evaluation parameters with body composition, metabolic syndrome features, and surrogate markers for subclinical cardiovascular dysfunction, including intima-media thickness.^9,10 These findings suggest that subtle thyroid-related variations have relevance beyond adiposity alone and intersect with early cardiometabolic risk.

Furthermore, obesity and thyroid physiology influence each other. Adipose-derived leptin stimulates hypothalamic thyrotropin-releasing hormone (TRH), which raises the levels of thyroid-stimulating hormone (TSH). Peripheral deiodinases regulate local triiodothyronine (T3) availability by activating or inactivating thyroxine (T4). ¹¹ These pathways may produce small serum changes in free triiodothyronine (FT3), free thyroxine (FT4), and TSH levels without causing overt thyroid disease. ¹²

Population studies have demonstrated mixed associations within reference ranges. Several cohorts have described positive links between adiposity and TSH and inverse associations between adiposity and FT4. ¹³ Others have reported higher FT3 levels or FT3-to-FT4 ratios in heavier individuals, consistent with adaptive conversion. Evidence for sex-specific patterns is also accumulating, with some datasets reporting stronger TSH/waist associations in women. ¹⁴

In Saudi adults, neck circumference (NC) is a practical screening tool for evaluating metabolic risk. Studies have suggested cutoff points of 37.5–39.25 cm for men and 32.5–34.75 cm for women to detect obesity and metabolic syndrome. These locally derived thresholds support the use of NC along with waist measures. ¹⁵ However, an important gap remains. Prior studies have reported mixed associations and have not consistently examined detailed body composition measures and simple central adiposity markers together in euthyroid Saudi adults, particularly in analyses stratified by sex. The purpose of this study was to examine the associations between serum TSH, FT3, and FT4 levels and detailed measures of body composition and general and central adiposity in euthyroid Saudi adults and assess whether these associations differ by sex. We focused on body mass index (BMI), waist indices, waist-to-height ratio (WHtR), and NC because these capture general and central adiposities in clinical settings. Identifying these links may refine risk assessment, guide interpretation of borderline thyroid test results in primary care and support targeted weight management strategies.

Methods

Study design and participants

The present study was a cross-sectional analysis of participant-level data from the same study population described in a previous report. ¹⁶ In brief, participants were recruited from community health centers and university clinics through convenience sampling. Adults were invited through social media between December 2021 and December 2023. Interested individuals attended the Nutrition Clinic at King Saud University, received study information, and signed informed consent. Eligible participants were adults aged ≥18 years who provided informed consent, completed the baseline visit, and had complete anthropometric and biochemical data available, including measured BMI and serum TSH, FT3, and FT4 levels. Participants were excluded for the following primary reasons: (a) pregnancy or lactation, n = 24; (b) use of mineral or vitamin supplements, n = 118; (c) use of antihypertensive medications, n = 67; (d) use of corticosteroids, antidepressants, hormone therapy, or thyroid-affecting medications, n = 73; (e) history of cardiovascular disease or stroke, n = 29; (f) history of type 1 diabetes or dyslipidemia, n = 154; and (g) known thyroid disease, n = 58. Among the 1300 individuals recruited, 1074 completed the baseline visit and 551 met the eligibility criteria for the present analysis. The study was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2024, and was approved by the Institutional Review Board of King Saud University (KSU-IRB-21-314), Riyadh, Saudi Arabia. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. ¹⁷

Anthropometric and body composition measurements

Height, weight, and waist and hip circumferences were measured by trained staff using standardized protocols. BMI was calculated as weight (kg)/height (m²). WHtR was computed by dividing the waist circumference (cm) by height (cm). NC was measured just below the larynx with the participant looking forward, using a nonstretchable tape (to the nearest 0.1 cm). Body composition was assessed using a multi-frequency bioelectrical impedance analysis (BIA) device (InBody 770, InBody Co., Ltd., Seoul, South Korea, manufactured 2018). The analyzer estimates body composition from segmental impedance measurements obtained at multiple frequencies. Measurements were taken when the participant was in light clothing and bare feet. The device-generated outputs used in this study were fat mass (kg), body fat percentage, muscle mass (kg and %), and total lean body mass. For some analyses, we categorized obesity indices into the following “at-risk” thresholds: (a) BMI ≥30 kg/m² (“obesity”); (b) elevated waist circumference (≥102 cm for men and ≥88 cm for women); (c) elevated WHtR (≥0.50); (d) high waist-to-hip ratio (WHR; ≥0.90 in men or ≥0.85 in women); and (e) high NC (≥38 cm in men or ≥34 cm in women). These values were based on standard cutoff values for metabolic risk in the Saudi population.

Biochemical measurements

Venous blood samples were collected after an overnight fast of at least 12 h during the morning clinic visit, between 9 and 11 am, for thyroid function tests and metabolic profiling. Serum TSH, FT3, and FT4 levels were measured using electrochemiluminescence immunoassays (Elecsys TSH, Elecsys FT3 III, and Elecsys FT4 III) on a Cobas e analyzer, Roche Diagnostics. The intra-assay and inter-assay coefficients of variation were <5% for all thyroid parameters. Fasting glucose levels were measured using the ARCHITECT Glucose assay. Insulin was measured using a chemiluminescent microparticle immunoassay (ARCHITECT Insulin assay). Total cholesterol, triglyceride, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) levels were measured using the ARCHITECT Cholesterol, ARCHITECT Triglyceride, ARCHITECT Ultra HDL-C, and ARCHITECT Direct LDL-C assays on the ARCHITECT ci system. We also calculated homeostasis model indices for insulin resistance homeostatic model assessment of insulin resistance (HOMA-IR) and homeostatic model assessment of beta cell function (HOMA-β). We categorized participants into “low FT3” (lower than the reference range of <2.3 pg/mL), “low FT4” (<0.8 ng/dL), and “high TSH” (>4.5 mIU/L) groups for comparison with individuals in the normal range.

Thyroid panel and exposure definitions

The exposures were serum TSH, free T4, and free T3 levels measured using the fasting samples at the same laboratory visit. We examined thyroid hormone levels within reference ranges and created indicator variables for high TSH, low FT4, and low FT3 using laboratory cutoff points.

Statistical analyses

Data were analyzed using the Statistical Package for Social Sciences (SPSS) software (v.27). Normality of quantitative variables was assessed using the Shapiro–Wilk test and by visual inspection of histograms and normal Q-Q plots. Variables with non-normal distributions were presented as median and interquartile range values and analyzed using nonparametric tests, including the Mann–Whitney U test for group comparisons and Spearman’s rank correlation for associations. Continuous variables were summarized as mean ± SD (or median and interquartile range values for non-normally distributed measures) and categorical variables as frequencies (%). Participants were categorized based on thyroid hormone levels: TSH level was classified as normal (0.4–4.5 mIU/L) or high (≥4.5 mIU/L), FT3 as normal (2.3–4.2 mIU/L) or low (<2.3 mIU/L), and FT4 as normal (0.8–1.8 mIU/L) or low (<0.8 mIU/L). Between-group comparisons were performed using independent t-tests or Mann–Whitney U test for continuous variables and the chi-square test for categorical variables. We also performed age- and sex-adjusted comparisons using analysis of covariance.

In the primary analyses, we examined associations between the thyroid markers, TSH, FT3, and FT4, and continuous adiposity and body composition outcomes using Spearman’s rank correlation coefficient. Because correlation analysis is nondirectional, no single dependent variable was specified for this analysis. Secondary analyses used logistic regression to estimate the odds of binary obesity-related outcomes stratified by thyroid status. The dependent variables were obesity, defined as BMI ≥30 kg/m², elevated waist circumference, elevated WHtR, elevated WHR, and high NC. The main independent variables were thyroid status categories, including high TSH, low FT4, and low FT3. We built three models: (a) unadjusted; (b) adjusted for age and sex; and (c) adjusted for additional covariates (smoking status and physical activity level—factors that differed by thyroid status or could confound relationships with obesity). In the sex-stratified analyses, we ran models separately for male and female participants. The results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). A two-sided p < 0.05 was considered statistically significant.

Results

Descriptive characteristics

The final analyses sample included 551 euthyroid adults; among them, 323 (58.6%) were men and 228 (41.4%) were women. The mean values of evaluated parameters were as follows: age, 28.8 ±11.0 years; BMI, 26.0 ± 6.1 kg/m²; waist circumference, 82.8 ±17.6 cm; NC, 34.2 ±4.1 cm; fat mass, 25.0 ±12.4 kg; body fat percentage 33.8% ± 10.5%; muscle mass, 25.4 ± 6.8 kg; and lean body mass, 46.1 ±11.5 kg. Detailed anthropometric, body composition, and metabolic characteristics are presented in Table 1.

Table 1.

Descriptive statistics stratified by FT3 status.

	FT3
	Normal2.3–4.2 pg/mL	Low<2.3 pg/mL	p-value	Adjusteda
N	512	39
Age (years)	28.3 ± 10.6	34.9 ± 13.5	0.005
Women/Men	225/287	3/36	<0.001
Weight (kg)	71.7 ± 19.3	64.4 ± 14.4	0.005	0.011
Height (cm)	165.6 ± 9.1	159.3 ± 7.0	<0.0001	<0.0001
BMI (kg/m2)	26.0 ± 6.1	25.4 ± 5.5	0.500	0.510
Waist circumference (cm)	83.4 ± 17.4	75.4 ± 18.6	0.012	<0.0001
Waist-to-height ratio	0.5 ± 0.1	0.5 ± 0.1	0.328	0.254
Waist-to-hip ratio	0.8 ± 0.1	0.8 ± 0.1	0.008	<0.0001
NC (cm)	34.4 ± 4.1	31.9 ± 2.8	<0.0001	<0.0001
Fat (kg)	25.1 ± 12.5	23.8 ± 10.4	0.462	0.509
Fat (%)	33.7 ± 10.5	35.0 ± 10.3	0.471	0.401
Muscle mass (kg)	25.6 ± 6.8	22.1 ± 5.1	<0.0001	<0.0001
Muscle mass (%)	36.3 ± 6.3	34.8 ± 6.1	0.149	0.086
Lean body mass (Kg)	46.6 ± 11.6	39.4 ± 6.8	<0.0001	<0.0001
Water (%)	12.6 ± 2.3	11.6 ± 1.5	<0.0001	0.001
Cholesterol (mg/dL)	171.1 (146.6–194.4)	174.8 (154.3–215.4)	0.081	0.011
Triglycerides (mg/dL)	61.1 (49.2–87.7)	64.7 (47.4–101.4)	0.517	0.206
LDL-C (mg/dL)	108.7 (89.7–130.3)	110.2 (88.2–132.3)	0.902	0.424
HDL-C (mg/dL)	53.4 (46.8–62.3)	60.7 (47.8–68.1)	0.033	0.004
Glucose (mmol/l)	5.1 (4.7–5.5)	5.0 (4.8–5.5)	0.997	0.084
Insulin (µIU/mL)	9.9 (7.0–13.9)	9.6 (6.3–11.4)	0.144	0.115
HOMA-IR	2.2 (1.6–3.2)	2.1 (1.2–2.8)	0.145	0.591
HOMA-β	35.6 (24.6–50.3)	31.1 (20.4–38.8)	0.017	0.044

Data are presented as mean ± SD for normally distributed variables and median (Quartile 1–Quartile 3) values for non-normally distributed variables; p <0.05 is considered statistically significant. ^aIndicates p-values adjusted for age and sex.

BMI: body mass index; FT3: free triiodothyronine; HDL-C: high-density lipoprotein cholesterol; HOMA-IR: homeostatic model assessment of insulin resistance; HOMA-β: homeostatic model assessment of beta cell function; LBM: lean body mass; LDL-C: low-density lipoprotein cholesterol; NC: neck circumference.

Compared with participants with normal FT3 levels, those with low FT3 levels (<2.3 pg/mL) were older and were more likely to be male. After adjusting for age and sex, individuals with low FT3 levels exhibited significantly lower weight (p = 0.011), height (p < 0.001), waist circumference (p < 0.001), NC (p < 0.001), muscle mass (p < 0.001), lean body mass (p < 0.001), and mean arterial pressure (p = 0.006) as well as higher HDL-C (p = 0.004) and lower HOMA-β (p =0.044) levels. However, no significant differences were observed in the BMI or WHtR (Table 1).

Those with low FT4 levels (<0.8 ng/dL) had higher adiposity and lower lean body weight than those with normal FT4 levels. After adjusting for age and sex, individuals with low FT4 levels demonstrated significantly higher BMI (p = 0.003), fat mass (p < 0.001), fat percentage (p < 0.001), and triglyceride levels (p = 0.032). They also had lower muscle mass (p < 0.001), muscle percentage (p < 0.001), and lean body mass (p = 0.003). In addition, low FT4 levels were associated with increased WHtR (p = 0.020), WHR (p = 0.040), and NC (p = 0.023) (Supplementary Table 1). There was no significant difference in the BMI, waist circumference, fat mass, or lean mass between participants with high TSH (>4.5 mIU/L) and normal TSH (0.4–4.5 mIU/L) levels. However, individuals with high TSH levels exhibited significantly higher fasting glucose levels (p = 0.032); no significant differences were observed in insulin resistance (HOMA-IR level, p = 0.652) or beta cell function (HOMA-β level, p = 0.698) (Supplementary Table 2).

Correlation analyses

Correlation analyses revealed significant associations between thyroid function markers (FT3, FT4, TSH) and obesity indices, with FT3 showing the strongest correlations. Higher FT3 levels were positively associated with height (r = 0.37, p < 0.001), muscle mass (r = 0.28, p < 0.001), and lean body mass (r = 0.29, p < 0.001) and inversely related to fat percentage (r = −0.25, p < 0.001) and cholesterol (r = −0.22, p < 0.001). FT4 levels showed weaker but significant negative correlations with BMI (r = −0.12, p = 0.005), fat mass (r = −0.14, p = 0.001), and lipid levels (cholesterol: r = −0.14, p = 0.001); however, FT4 levels were positively associated with muscle-related measures (muscle mass %: r = 0.22, p < 0.001). TSH levels showed minimal associations, with only a weak positive correlation with fat mass (r = 0.09, p = 0.040) and no significant relationships with most obesity indices (Table 2).

Table 2.

Spearman’s rank correlation coefficients between thyroid function markers and anthropometric and metabolic variables.

Parameters	FT3		TSH		FT4
	r	p-value	r	p-value	r	p-value
Age (years)	−0.31	<0.0001	−0.06	0.151	−0.19	<0.0001
Weight (kg)	0.10	0.018	0.07	0.093	−0.02	0.671
Height (cm)	0.37	<0.0001	0.00	0.994	0.20	<0.0001
BMI (kg/m2)	−0.07	0.109	0.08	0.056	−0.12	0.005
Waist circumference (cm)	0.13	0.002	0.04	0.388	−0.01	0.808
Waist-to-height ratio	0.01	0.758	0.04	0.342	−0.09	0.028
Waist-to-hip ratio	0.22	<0.0001	0.03	0.502	0.07	0.117
NC (cm)	0.28	<0.0001	0.02	0.623	0.09	0.029
Fat (kg)	−0.09	0.030	0.09	0.040	−0.14	0.001
Fat (%)	−0.25	<0.0001	0.07	0.096	−0.22	<0.0001
Muscle mass (kg)	0.28	<0.0001	0.02	0.597	0.13	0.003
Muscle mass (%)	0.28	<0.0001	−0.07	0.115	0.22	<0.0001
Lean body mass (kg)	0.29	<0.0001	0.05	0.267	0.14	0.001
Log cholesterol (mg/dL)	−0.22	<0.0001	0.05	0.237	−0.14	0.001
Log triglycerides (mg/dL)	−0.07	0.094	0.04	0.377	−0.14	0.001
Log LDL-C (mg/dL)	−0.11	0.011	0.06	0.141	−0.11	0.013
Log HDL-C (mg/dL)	−0.19	<0.0001	−0.03	0.448	−0.07	0.087
Log glucose (mmol/l)	−0.03	0.456	0.07	0.081	−0.06	0.141
Log insulin (µIU/mL)	0.11	0.008	−0.03	0.550	−0.09	0.028
Log HOMA-IR	0.11	0.009	0.02	0.579	−0.10	0.016
Log HOMA-β	0.17	<0.0001	0.03	0.496	−0.10	0.018

Values represent Spearman’s coefficients (r); p-value <0.05 is considered statistically significant.

BMI: body mass index; FT3: free triiodothyronine; FT4: free thyroxine; HDL-C: high-density lipoprotein cholesterol; HOMA-IR: homeostatic model assessment of insulin resistance; HOMA-β: homeostatic model assessment of beta cell function; LBM: lean body mass; LDL-C: low-density lipoprotein cholesterol; NC: neck circumference; TSH: thyroid-stimulating hormone; WC: waist circumference; WHR: waist-to-hip ratio; WHtR: waist-to-height ratio.

Multivariable obesity risk in the overall sample

Logistic regression analysis revealed differential associations between thyroid dysfunction and obesity risk indices. High TSH levels (>4.5 mIU/L) showed limited associations with obesity risk indices and were significantly associated only with elevated WHR in the adjusted models (OR = 2.91; 95% CI, 1.06–8.36; p = 0.042). The association with low FT3 level was not significant after adjustment for covariates. A low FT4 level (<0.8 ng/dL) showed the strongest and most consistent associations with obesity risk, related to significantly increased odds of obesity (OR = 1.85; 95% CI, 1.09–3.10; p = 0.020), high waist circumference (OR = 2.68; 95% CI, 1.21–6.20; p = 0.017), WHR (OR = 2.69; 95% CI, 1.09–7.18; p = 0.037), and WHtR (OR = 2.08; 95% CI, 1.23–3.53; p =0.006) in the fully adjusted models (Table 3).

Table 3.

Odds of obesity risk indices by high TSH, low FT3, and low FT4 levels.

Exposure	Outcome	Unadjusted		Model 1		Model 2
		OR (95% CI)	p-value	OR (95% CI)	p-value	OR (95% CI)	p-value
High TSH	NC risk (≥38 cm in men; ≥34 cm in women)	1.04 (0.53, 1.99)	0.903	2.45 (0.69, 8.66)	0.17	3.26 (0.79, 13.53)	0.106
	BMI	1.17 (0.54, 2.38)	0.666	1.34 (0.61, 2.77)	0.437	1.30 (0.59, 2.72)	0.496
	WC risk (≥102 cm in men, ≥88 cm in women)	1.12 (0.54, 2.22)	0.749	1.96 (0.79, 4.79)	0.141	2.12 (0.83, 5.38)	0.113
	WHR risk (≥0.90 in men, ≥0.85 in women)	1.26 (0.62, 2.44)	0.509	2.89 (1.06, 8.22)	0.041	2.91 (1.06, 8.36)	0.042
	WHtR risk (≥0.50)	1.18 (0.62, 2.26)	0.615	1.55 (0.75, 3.18)	0.232	1.51 (0.72, 3.16)	0.268
Low FT3	NC risk (≥38 cm in men; ≥34 cm in women)	0.12 (0.03, 0.33)	<0.001	0.33 (0.05, 1.64)	0.209	0.24 (0.04, 1.24)	0.104
	BMI	0.70 (0.28, 1.57)	0.419	0.58 (0.22, 1.35)	0.234	0.54 (0.20, 1.29)	0.192
	WC risk (≥102 cm in men, ≥88 cm in women)	0.20 (0.05, 0.57)	0.009	0.36 (0.07, 1.28)	0.149	0.29 (0.06, 1.10)	0.096
	WHR risk (≥0.90 in men, ≥0.85 in women)	0.32 (0.11, 0.76)	0.019	1.21 (0.31, 4.26)	0.77	1.26 (0.32, 4.40)	0.722
	WHtR risk (≥0.50)	0.80 (0.41, 1.54)	0.51	0.67 (0.30, 1.45)	0.314	0.61 (0.26, 1.35)	0.227
Low FT4	NC risk (≥38 cm in men; ≥34 cm in women)	0.34 (0.20, 0.55)	<0.001	1.25 (0.46, 3.49)	0.657	1.50 (0.53, 4.40)	0.453
	BMI	1.77 (1.10, 2.83)	0.017	1.80 (1.08, 2.98)	0.023	1.85 (1.09, 3.10)	0.02
	WC risk (≥102 cm in men, ≥88 cm in women)	0.75 (0.45, 1.21)	0.25	2.31 (1.08, 5.09)	0.032	2.68 (1.21, 6.20)	0.017
	WHR risk (≥0.90 in men, ≥0.85 in women)	0.57 (0.34, 0.93)	0.027	2.51 (1.04, 6.56)	0.048	2.69 (1.09, 7.18)	0.037
	WHtR risk (≥0.50)	1.57 (1.03, 2.40)	0.038	2.07 (1.24, 3.48)	0.005	2.08 (1.23, 3.53)	0.006

Data are presented as odds ratios (95% CI); Model 1 was adjusted for age only when sex stratified and for age and sex in pooled models. Model 2 was additionally adjusted for smoking, physical activity, and supplementation.

BMI: body mass index; FT3: free triiodothyronine; FT4: free thyroxine; HDL-C: high-density lipoprotein cholesterol; HOMA-IR: homeostatic model assessment of insulin resistance; HOMA-β: homeostatic model assessment of beta cell function; LBM: lean body mass; LDL-C: low-density lipoprotein cholesterol; NC: neck circumference; TSH: thyroid-stimulating hormone; WC: waist circumference; WHR: waist-to-hip ratio; WHtR: waist-to-height ratio; OR: odds ratio; CI: confidence interval.

The analysis of male participants revealed distinct patterns of association between thyroid dysfunction and obesity risk indices. A high TSH level was not significantly associated with any obesity indices in the unadjusted or adjusted models. The small sample size limited analyses for low FT3 levels; however, available estimates showed no significant associations. Notably, low FT4 levels demonstrated strong positive associations with central adiposity markers in men, significantly increasing the odds of elevated waist circumference (OR = 6.24; 95% CI, 1.57–41.83; p = 0.022) and showing borderline significance for association with high WHtR (OR = 4.74; 95% CI, 1.01–32.39; p = 0.055) in the fully adjusted models. Although low FT4 levels showed a trend toward increased obesity risk (OR = 2.46; 95% CI, 0.90–6.81; p = 0.079), this association did not reach statistical significance after adjustment (Supplementary Table 3).

The analysis of female participants showed significant sex-specific associations between thyroid dysfunction and obesity risk. High TSH levels demonstrated particularly strong associations with central adiposity in women, with significantly higher odds of elevated waist circumference (OR = 4.83; 95% CI, 1.13–17.97; p = 0.023) and high WHR (OR = 5.45; 95% CI, 1.94–27.23; p = 0.042) in the fully adjusted models. Low FT4 levels were significantly associated with elevated WHtR (OR = 1.84; 95% CI, 1.04–3.23; p = 0.034) after adjustment for covariates; however, in the unadjusted model, the association with obesity (OR = 1.84, p = 0.039) was nonsignificant when accounting for confounders. Notably, low FT3 levels showed no significant associations with any obesity indices in women (all p > 0.05; Supplementary Table 4).

Discussion

Summary of main findings

In this cross-sectional study involving Saudi adults, we observed significant associations between body composition and thyroid function within the clinically normal range. Higher adiposity was linked to higher TSH levels and lower FT4 levels. These relationships showed notable sex-specific patterns; TSH levels were more strongly associated with central obesity measures (e.g. waist circumference) in women, whereas in men, lower FT4 levels were associated with larger waist size. Notably, the TSH and FT4 levels of all participants were in the euthyroid reference range; however, variation within that range correlated with body fat distribution. Collectively, our findings suggest that higher body fat is associated with higher TSH levels and lower peripheral thyroid hormone activity within the normal range, with sex-based differences. The added value of this study is not the general observation that adiposity and thyroid function are interrelated, which is already established, but the characterization of these associations within a population of euthyroid Saudi adults using both BIA-derived body composition measures and clinically accessible central adiposity indices, with sex-stratified analyses.

Our results are consistent with a growing body of evidence linking obesity to subtle changes in thyroid function. Population studies worldwide have documented that individuals with higher BMI tend to have higher normal-range TSH levels and slightly lower circulating thyroid hormones.^13,18 For example, a study by Al-Musa et al. has demonstrated that the mean TSH level increased across BMI categories (normal, overweight, obese) among euthyroid adults. ¹⁹ In that study, obese participants had significantly higher average TSH levels than lean individuals, despite all values being within the normal range. ¹⁹ Similarly, a meta-analysis involving >100,000 adults has demonstrated a positive correlation between TSH levels and BMI and an inverse correlation between FT4 levels and BMI; each 1-mIU/L increase in the TSH level was associated with an average increase of 0.21 kg/m² in the BMI. Meanwhile, each unit increase in the FT4 level corresponded to a −0.14-kg/m² lower BMI. ¹³ These effect sizes are modest but consistent; within the euthyroid range, heavier individuals tend to have higher TSH and lower FT4 levels than leaner individuals. This consistency is biologically plausible and may also reflect methodological overlap. Similar to the above-mentioned meta-analysis, our study focused on euthyroid adults and evaluated thyroid variations within the reference range rather than overt thyroid disease. In addition, our use of anthropometric and body composition measures in a sample of relatively young adults may have increased the sensitivity to detect subtle adiposity-related thyroid variation that might be obscured in studies with broader clinical heterogeneity. However, the magnitude of association in our data was smaller than in some clinic-based cohorts, which may reflect our exclusion of participants with known thyroid disease, major comorbidities, and medications affecting thyroid hormone levels.

Our findings are also consistent with studies from the Gulf region and other populations. A clinical study in Riyadh, Saudi Arabia, has reported that morbidly obese euthyroid patients had slightly elevated TSH and lower FT3 levels compared with age-matched lean controls. ¹⁸ In the combined group, BMI positively correlated with TSH levels and negatively correlated with FT3 levels. Most regional reports showing higher TSH levels in obese individuals were conducted in clinical or obesity-enriched samples, which may have increased their likelihood of detecting mild thyroid variation associated with greater adiposity. ¹⁹ However, there have been inconsistencies in the reported associations with T3 levels. Some studies (often in moderately obese individuals) have reported normal or even higher T3 levels in obese participants, ²⁰ whereas others (particularly in studies on morbid obesity) have reported lower FT3 levels, as observed in a Saudi bariatric-clinic cohort. ¹⁸ These discrepancies likely reflect both contextual and methodological differences across studies. Studies reporting lower FT3 levels in obese participants were conducted in morbidly obese or bariatric-clinic populations,^18,19 in whom long-standing obesity, greater metabolic impairment, and chronic inflammation may be more likely to suppress peripheral thyroid hormone signaling. In contrast, studies reporting normal or higher T3 levels in obese individuals often included broader community samples or participants with less severe obesity, in whom adaptive T4-to-T3 conversion may have been preserved. ²⁰ Differences in iodine exposure, exclusion of comorbid illness, and whether adiposity was assessed using BMI alone or in combination with central and body composition measures may also contribute to variations across studies. Our study was community-based, restricted to euthyroid adults, and incorporated both general and central adiposity indices, which may explain why FT4 levels showed stronger and more consistent associations than FT3 levels.

Notably, our investigation of sex differences extends the previous literature. An extensive cohort study involving >5000 Italian obese patients has demonstrated that TSH levels increase with higher BMI, predominantly in women, whereas in men, FT4 levels tend to decrease as adiposity increases. ²¹ In that study, female participants showed a significant trend toward higher TSH levels across incremental BMI classes, whereas in male participants, the opposite trend was observed for FT4 levels (i.e. FT4 levels declined with increasing obesity). ²¹ Our results are consistent with these reported sex differences. The similarity is notable despite important differences in the study design. The Italian cohort was substantially larger and included patients with obesity across BMI classes, whereas our study included euthyroid Saudi adults from community and clinic recruitment and examined not only BMI but also waist and body composition measures. The convergence of findings across these different settings suggests that the sex-specific pattern is robust; however, differences in age structure, obesity severity, fat distribution, and hormonal milieu could still influence the strength of association in individual studies. Our stronger waist measures-related findings in women may reflect the use of central adiposity markers to understand the sex-specific metabolic burden better than BMI alone.

Physiological mechanisms linking adiposity to thyroid function

Several biological mechanisms could be responsible for the observed associations between body composition and thyroid function. Adipose tissue is an active endocrine organ; an important adipokine, leptin, may play a central role in thyroid regulation. Leptin levels increase with greater fat mass and serve as a nutritional signal to the hypothalamus. Experimental evidence has shown that leptin stimulates hypothalamic TRH neurons, increasing pituitary TSH secretion. ²² During energy deficits (such during as fasting and cachexia), a declining leptin level contributes to a drop in TRH/TSH levels to conserve energy. Conversely, in obesity, chronically elevated leptin levels may drive a compensatory increase in TSH release. ¹¹ Our findings are consistent with the hypothesis that higher TSH in heavier individuals may be related to the stimulatory effect of leptin on the hypothalamus. Other data are consistent with this model; for instance, Mele et al. have reported that leptin is a significant positive predictor of TSH levels in nonsmoking obese individuals. ²¹ Therefore, leptin provides a physiological link between increased fat mass and upregulation of the hypothalamic–pituitary–thyroid (HPT) axis.

Chronic inflammation and insulin resistance, common in obesity, may also mediate thyroid hormone changes. Proinflammatory cytokines (interleukin 6 (IL-6) and tumor necrosis factor-alpha (TNFα)) suppress aspects of thyroid signaling and could contribute to the “low T3 syndrome” in severe obesity, analogous to nonthyroidal illness syndrome, although usually in a milder form. In our study, we did not observe a clear correlation between insulin resistance (HOMA-IR) and thyroid levels, ¹⁸ implying that hyperinsulinemia might not be the primary driver. Overall, the HPT axis may reset in obesity; the pituitary requires a higher leptin (and fat mass) signal to maintain energy homeostasis, leading to mild elevation in the TSH level. Meanwhile, peripheral tissues adjust T4/T3 conversion in complex ways. Whether these changes increase the basal metabolic rate or are simply a result of obesity remains debatable. The slightly elevated TSH level in obesity may be an adaptive response to promote thyroid hormone production and metabolism (the “central adaptation hypothesis”). However, it is equally likely that weight gain leads to mild functional hypothyroidism (a higher TSH set-point). Elements of both causation and adaptation may be involved, forming a bidirectional feedback loop between thyroid function and adiposity.

Sex-specific associations and interpretation

In our population, the relationship between thyroid indices and adiposity differed by sex, which has also been observed in previous studies. ²¹ In female participants, TSH levels demonstrated a robust positive association with central obesity (e.g. higher TSH levels correlated with larger waist circumference and higher body fat percentage). In contrast, among men, FT4 levels were more strongly inversely related to waist circumference; men with larger waistlines tended to have lower FT4 levels, and the correlation between TSH levels and BMI was weaker in men. These sex differences likely reflect hormonal, behavioral, and body composition factors. Methodological differences may also matter because studies that rely mainly on BMI can miss sex-specific variation in fat distribution, whereas waist circumference, WHR, and NC may better capture the central adiposity patterns that were most strongly linked to variations in thyroid hormone levels in our female subgroup. Furthermore, women typically have a higher body fat percentage and circulating leptin levels at a given BMI than men. A specific increase in fat mass may result in higher TSH levels in women due to higher leptin secretion. Men, conversely, accumulate more visceral (intraabdominal) fat, which is metabolically active and associated with insulin resistance. Visceral fat can influence deiodinase activity in the liver or muscle and may contribute to lower FT4 levels (and tendency toward higher T3 levels) in men with central obesity. Taken together, women’s thyroid function may be more sensitive to total fat mass, whereas men’s thyroid parameters may reflect visceral fat burden. Furthermore, estrogen increases thyroid binding globulin (TBG) levels, raising the total T4/T3 levels. In our study, however, we focused on FT4/FT3 levels; therefore, TBG levels were less relevant. Estrogen and progesterone may also modulate immune function, affecting autoimmune thyroiditis prevalence, which is higher in women, and possibly HPT axis sensitivity. Men with obesity often have a relative testosterone deficiency, which can influence metabolism. Although speculative, such hormonal milieu differences could modulate how thyroid function interacts with weight. For instance, postmenopausal women who have lower estrogen levels may show different patterns than premenopausal women; unfortunately, our sample was not large enough to stratify women by menopausal status to explore this association.

Overall, our sex-specific findings highlight that the thyroid–adiposity relationship varies among individuals. Women’s TSH levels are closely associated with adiposity (especially centrally), whereas men exhibit a clearer decline in FT4 levels with increasing obesity, a pattern also reported by Mele et al. ²¹ This indicates that clinicians should be mindful of patient sex when interpreting mild thyroid changes in the context of weight. It also leads to the question of whether women are more prone to obesity-related TSH elevations (and possibly related symptoms) and whether men are more prone to obesity-related low-normal FT4 levels and their metabolic consequences? Further studies are needed to address these research questions.

Clinical implications

The observed associations between body composition and thyroid function may have practical implications for evaluating and managing thyroid health in obese patients. First, caution should be exercised while interpreting borderline thyroid function results in obese patients. For instance, mildly elevated TSH levels in an asymptomatic obese patient may not necessarily signify primary thyroid disease requiring treatment; they may indicate a physiological elevation related to adiposity. ¹⁹ Our findings are consistent with this interpretation since several individuals with the highest TSH levels in our sample were women with significant central obesity without overt thyroid illness. Clinical guidelines currently advise against over diagnosing subclinical hypothyroidism; our data support the idea that in obese patients with high-normal or slightly high TSH levels, weight should be considered a contributing factor. In such cases, weight loss and lifestyle modification should be pursued, followed by reassessment of thyroid function before the prescription of lifelong levothyroxine therapy. Weight reduction has been shown to reverse some thyroid hormone-related changes: for instance, weight loss after bariatric surgery significantly lowers TSH levels in obese euthyroid patients. ²³ Interventional studies have reported that TSH can decline after substantial weight loss; however, our cross-sectional data do not address reversibility. ²⁴ Treating obesity (with diet, exercise, and bariatric intervention if indicated) can serve as a dual-benefit strategy, improving metabolic health and normalizing thyroid function tests without medication.

Another implication of our findings is that adjusted reference ranges or tailored interpretation of thyroid tests in populations with high obesity prevalence may be needed. The upper limit of normal TSH level may be higher in obese populations than in lean populations. ²¹ For example, in an Italian obesity cohort, the 97.5^th percentile of TSH was ∼5.0 mIU/L, whereas ∼4.0 mIU/L is often cited for general populations. ²¹ Although not yet adopted in practice, this raises the question of whether laboratories and clinicians should interpret a higher TSH in an obese patient as “normal-for-weight.” Our study provides data supporting a similar trend in the Saudi region. We do not recommend using weight-specific cutoffs; however, shifting clinical reference ranges may prevent unnecessary treatment in patients who are borderline euthyroid by standard criteria but otherwise healthy apart from excess weight.

Furthermore, our results suggest that thyroid parameters could be markers for metabolic risk. In men, a low-normal FT4 level along with central obesity might identify those at higher metabolic syndrome or cardiovascular risk. A recent study has shown that a low-normal FT4 level is independently associated with the “metabolically unhealthy” obese phenotype (e.g. hypertension and dyslipidemia). ²⁵ Clinicians might therefore screen for other metabolic complications in obese patients with unusually low FT4 or high TSH levels. However, thyroid function alone should not be overinterpreted. Obese patients often have symptoms such as fatigue, low energy, and weight gain that mimic hypothyroid symptoms; therefore, differentiating whether these are due to true thyroid hormone deficiency versus weight-related effects is critical. Our findings reinforce that many obese patients with these symptoms have normal thyroid hormone levels (albeit on the higher end of normal TSH levels). Thus, addressing lifestyle factors may alleviate symptoms without thyroid pharmacotherapy in these cases.

Strengths and limitations

This study has several strengths. First, we used detailed body composition measurements beyond BMI, including waist circumference and body fat percentage, allowing a nuanced analysis of the relationship of central vs. general adiposity with thyroid function. Second, use of population-appropriate cutoffs for obesity (BMI and waist thresholds validated for Middle Eastern populations) improved the relevance of our findings in the local context. Another strength is that we excluded individuals with overt thyroid disease or significant comorbid conditions. By focusing on euthyroid participants without clinical or biochemical hyper/hypothyroidism, we could detect subtle physiological relationships without the confounding effect of thyroid medications or severe illness. Furthermore, our study contributes data from an under-studied population, Middle Eastern adults, who are underrepresented in global thyroid obesity research. This adds ethnic and environmental diversity to the literature.

We also acknowledge that there are important limitations. First, the cross-sectional design limits causal inference; we cannot determine whether higher TSH level contributed to weight gain, whether obesity caused the TSH level to rise, or whether there is a bidirectional relationship. Longitudinal studies that follow weight and thyroid changes over time are needed to clarify directionality. ¹⁹ Second, selection bias is possible. Our sample was drawn from adults attending health centers in an urban Saudi setting; thus, they may not represent the general population precisely. For example, health-conscious individuals or those with minor symptoms may have been more likely to undergo thyroid function and body composition evaluations, potentially skewing the sample. Our cohort also had a relatively wide age range but a moderate size, lending limited statistical power to certain subgroup analyses (e.g. stratifying by age or menopausal status).

Another limitation is that we did not have complete data regarding all potential confounders. Additionally, we could not systematically assess dietary iodine intake or urinary iodine. Given Saudi Arabia’s iodine sufficiency programs, overt iodine deficiency is likely uncommon. However, we cannot rule out that minor variations in iodine status could have affected individual thyroid function set-points. Thyroid autoimmunity is another factor to be considered. Although we excluded known thyroid disease, we did not measure thyroid peroxidase or thyroglobulin antibodies in all participants. Thus, a few “euthyroid” individuals may have had early, mild Hashimoto’s thyroiditis, which could elevate TSH levels; however, such cases would have been rare and would have introduced a bias toward stronger TSH–weight correlations if present. Finally, our study’s observational nature precludes us from confirming underlying mechanisms; we can only speculate on the involvement of leptin and deiodinase without performing direct measurements.

Despite these limitations, the consistency of our results with previous reports and the biological plausibility of our findings demonstrate the reliability of our conclusions. However, caution is advised when extrapolating these results beyond the context of this study. For example, our results should not be generalized to children, patients with thyroid disorders, or populations with very different lifestyle factors without further investigation.

Future research directions

Our findings suggest several avenues for future research. For instance, longitudinal studies should be conducted on euthyroid individuals to track how weight changes affect thyroid function over time (and vice versa). This type of study could help answer causality questions, such as whether an increase in TSH levels predicts subsequent weight gain or whether people who gain weight later exhibit elevations in TSH levels. Early evidence from other cohorts suggests a bidirectional relationship (higher baseline TSH levels are associated with future weight gain, and weight gain elevates TSH levels), in a self-reinforcing cycle. Prospective studies in Saudi or Gulf populations should be performed, especially if they incorporate serial measurements of weight, body composition, and thyroid hormone levels.

From a clinical research perspective, the impact of adjusting TSH reference ranges for BMI on diagnostic accuracy should be investigated. An algorithm or nomogram incorporating BMI or fat percentage when interpreting thyroid test values could be developed, and studies could test if this approach better predicts who will develop true hypothyroidism and who will respond to treatment. Additionally, further research is needed to clarify the health consequences of changes in thyroid hormone levels in obesity. For example, studies should evaluate whether obese patients with the highest TSH levels have worse outcomes in terms of metabolic or cardiovascular health than those with lower TSH levels. Long-term outcome studies (tracking the development of diabetes, dyslipidemia, or cardiac events in relation to normal baseline thyroid function) could answer these questions.

In this cross-sectional study, thyroid markers within the euthyroid range were associated with several body composition and adiposity measures, and these associations differed by sex. These findings should be interpreted as associations rather than evidence of directionality or causality. Further research should build on these findings, ultimately aiming to refine thyroid function interpretation in obese patients and understand whether modulating the thyroid-body weight axis can improve health outcomes. Continued investigation will contribute to more personalized and precise care in endocrinology, integrating metabolic status with thyroid health.

Supplemental Material

sj-pdf-1-imr-10.1177_03000605261444581 - Supplemental material for Body composition indices align with thyroid variation in euthyroid Saudi adults: A cross-sectional study

Supplemental material, sj-pdf-1-imr-10.1177_03000605261444581 for Body composition indices align with thyroid variation in euthyroid Saudi adults: A cross-sectional study by Deemah S Alrajhi, Sara AI-Musharaf, Tagreed A Mazi, Madhawi Aldhwayan, Mahmoud Abolmeaty and Ghadeer S Aljuraiban in Journal of International Medical Research