Association between overt, subclinical hypothyroidism and metabolic dysfunction-associated steatotic liver disease using magnetic resonance imaging

Abstract

BACKGROUND

In addition to an elevated risk of cirrhosis and hepatocellular carcinoma, along with metabolic dysfunction-associated steatotic liver disease (MASLD) represents the primary contributor to chronic liver disease, affecting 30% of the global population. Hypothyroidism is a common disorder that may influence MASLD development. Limited data have addressed the association between the whole spectrum of hypothyroidism (overt, subclinical) and MASLD.

AIM

To determine the association between the whole spectrum of hypothyroidism (overt, subclinical) and MASLD and its severity determinants.

METHODS

This observational investigation encompassed 144 adult participants from Egypt, consisting of 48 subjects diagnosed with obvious hypothyroidism, 48 with subclinical hypothyroidism and 48 healthy control subjects, subsequent to the removal of those with a history of alcohol intake, diabetes mellitus, prediabetes, or any etiologies of chronic liver disease, such as chronic viral hepatitis B and C. All participants underwent evaluation for MASLD employing ultrasound imaging, with diagnosis thereafter corroborated through magnetic resonance imaging, and hepatic fat percentage (HF%) was determined to gauge MASLD severity at Kasralainy Hospitals, Cairo University, Egypt.

RESULTS

The 34 of the 48 overt hypothyroid, 27 of the 48 subclinical hypothyroid and 3 of the control group were diagnosed to have MASLD. Mild steatosis was 38.2% and 63% whereas moderate steatosis was 61.8% and 37% among the overt and subclinical groups respectively. The statistically significant predictor for the risk of MASLD development was the thyroid-stimulating hormone (TSH) level (22.9 ± 26.4, P value < 0.001). TSH correlated positively with the degree of steatosis independently of other factors according to HF% (r = 0.69, P value < 0.001). Multivariate analysis proved TSH as an autonomous determinant for MASLD development and severity. The receiver operating characteristic curve, with a sensitivity of 83% and a specificity of 89%, demonstrated TSH level of > 6.1 mIU/L or higher is associated with an increased risk of developing MASLD.

CONCLUSION

Overt and subclinical hypothyroidism are directly related to MASLD development. TSH exhibits an increased association with HF% and is an independent risk variable associated with the onset and severeness of MASLD.

Key Words: Metabolic dysfunction-associated steatotic liver disease; Hypothyroidism; Overt; Subclinical; Predictors; Fibrosis; Severity; Ultrasonography; Magnetic resonance imaging

Core Tip: Metabolic dysfunction-associated steatotic liver disease (MASLD) has emerged as the foremost etiology of cirrhosis and hepatocellular carcinoma. Our study assessed the overlooked association between this global liver disease and the spectrum of hypothyroidism. The three study groups (overt, subclinical hypothyroid, control) were assessed for MASLD by ultrasonography. The diagnosis was confirmed by magnetic resonance imaging, hepatic fat percentage was calculated. Ultrasound results were compared with magnetic resonance imaging findings to determine degree of agreement between both diagnostic methods. Multivariate analysis identified the independent risk variables associated with the onset and severeness of MASLD. The receiver operating characteristic curve concluded that there is a certain thyroid-stimulating hormone level above which the risk of MASLD development increases.

INTRODUCTION

Hepatic fat accumulation is the defining characteristic of metabolic-associated steatotic liver disease (MASLD), the previously called non-alcoholic fatty liver disease (NAFLD), a chronic hepatic condition that is not attributable to excessive alcohol consumption. The new nomenclature emphasyzes metabolic dysfunction as the central driver in patients with MASLD and reduces stigma of “fatty and alcohol” allowing the co-existence of other liver disease etiologies. The clinical spectrum of MASLD is broad, including uncomplicated fatty liver and steatohepatitis, which may advance to liver fibrosis, hepatocellular carcinoma, and cirrhosis. On top of that, MASLD is underappreciated and independent risk factor for atherosclerotic cardiovascular diseases[1,2]. Over recent decades, the prevalence of MASLD has markedly increased, rendering it a principal cause of hepatic pathology worldwide. Consequently, the factors influencing MASLD have constituted a primary focus within medical research. Some studies suggest that hypothyroidism might play a crucial role in the pathogenesis of MASLD[3].

De novo lipogenesis, cholesterol metabolism, beta-oxidation, and carbohydrate metabolism are all influenced by thyroid hormones (TH), which are essential for the maintenance of metabolic homeostasis at both systemic and hepatic levels[4]. The studies assessing the impact of thyroid disorders on MASLD development are inconsistent, with limited data assessing MASLD development among the subclinical hypothyroid patients[5,6].

Evaluating the overlooked association between this global liver disease and the spectrum of hypothyroidism raises concern regarding three main points: Whether the spectrum of thyroid dysfunction should be incorporated into the current criteria of metabolic risk abnormalities for MASLD development; the need for routine surveillance of MASLD among patients with thyroid dysfunction; and, finally, whether the threshold for treatment of subclinical hypothyroidism should be lower than currently recommended by the endocrinal societies aiming to prevent the undesirable consequences of MASLD among these patients.

MATERIALS AND METHODS

Data and participants

The Endocrinology Outpatient Clinic, which is affiliated with the Internal Medicine department at Cairo University Hospitals, recruited 144 Egyptian adults between November 2023 and September 2025 for the current observational study. The 2012 guidelines of the American Thyroid Association[7] were followed to classify the 144 participants into three groups based on the inclusion criteria: The study included forty-eight patients each with overt hypothyroidism, subclinical hypothyroidism, and euthyroid controls. Gender, age, body mass index (BMI) were assessed to determine the comparability of the three groups. Informed permission was secured from all participants, and the Research Ethics Committee of the Faculty of Medicine at Cairo University sanctioned the research in November 2023.

Inclusion and exclusion criteria

The subsequent criteria for inclusion were as follows: (1) Group A: Included 48 individuals exhibiting laboratory-confirmed overt hypothyroidism [low free thyroxine (FT4) and thyroid-stimulating hormone (TSH) > 10 mIU/L]; (2) Group B: Consisted of 48 patients with laboratory evidence of subclinical hypothyroidism (normal FT4 and TSH ≥ 4.5 mIU/L); and (3) Group C: Consisted of 48 healthy, age-, sex-, and BMI-matched subjects (30-35 kg/m²) with laboratory evidence of euthyroidism (normal FT4 and TSH < 4.5 mIU/L).

The exclusion criteria for the three groups were as follows: Diabetes mellitus, prediabetes, chronic kidney disease, history of alcohol consumption, and any persistent liver disease, encompassing persistent hepatitis B and C viral infections, primary biliary cirrhosis, hemochromatosis, and autoimmune hepatitis, Wilson’s disease, or drug-induced hepatitis. Additionally, individuals with severe malnutrition, expectant women, and those who were taking steatogenic medications (amiodarone, tamoxifen, methotrexate, corticosteroids, or estrogen) were excluded. Participants in the three groups underwent medical history evaluation, clinical examination, and laboratory tests, including complete blood count, alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum urea, serum creatinine, glycated hemoglobin, viral markers (hepatitis B surface antigen, bepatitis C virus antibodies, human immunodeficiency virus antibodies), TSH, FT4, serum total cholesterol, triglycerides, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) followed by abdominal ultrasonography (3.5 mHZ probe Logiq E10 device) and abdominal magnetic resonance imaging (MRI) [philips 1.5 tesla magnetic resonance (MR) scanner].

MRI scanning parameters

All recruited patients underwent MRI scan on a high field (1.5Tesla) equipped with a phased-array body coil, sequences were obtained as follows: (1) Modification of Dixon method involving a multi-breath-hold double gradient echo T1-weighted sequence was done. Scan parameters: Relaxation time = 104 milliseconds (in-phase and out-of-phase), echo time = 4.8 milliseconds [in-phase (IP)] and 2.1 milliseconds [out-of-phase (OP)], matrix 128 × 256, flip angle = 75^°, slice thickness: 10 mm, field of view = 35 cm, matrix size = 256 × 179, and single excitation. Acquisition time = 18 seconds. This sequence allowed simultaneous acquisition of both IP and OP image representing chemical shift sequences during a multibreath-hold interval required to cover the entire liver; (2) Axial T2W TRUFISP and fat suppression sequence: Axial STIR FSE sequence fast spinecho space; (3) Axial diffusion (diffusion-weighted images) images to exclude hepatic focal lesions or masses or any other liver pathology; and (4) No contrast was given.

Images post processing and data analysis were performed on a workstation by an expert lecturer of Diagnostic and Interventional Radiology Department, Kasralainy Hospitals, Cairo University, Egypt as follows: (1) Qualitative assessment was managed through visual assessment of fat suppression sequence where fatty liver appears darker (hypo-intense) confirming the diagnosis, checking for increased liver brightness (higher signal intensity) on IP compared to normal liver tissue on standard T1-weighted sequences and corresponding a signal drop on OP (OP sequences due to fat-water cancellation relative to the reference spleen; and (2) Quantitative assessment was achieved by obtaining pixel signal intensities (SI) from IP and OP images from selected regions of interest (ROI’s) of average area 1-1.5 cm². Three ROI’s were placed on the liver parenchyma segments VIII and V of the right hepatic lobe and II of the left hepatic lobe, respectively, excluding blood vessels, motion artifacts or partial volume effects. An average SI value of the liver IP and liver OP were obtained. In the spleen, three additional ROIs were placed to obtain an average of splenic SI on both IP and OP images respectively. Hepatic fat percentage (HF%) was calculated from the average pixel SI data for grading of hepatic steatosis as mild, moderate and severe, furthermore Hepatic Fat fraction was calculated for confirmation, as shown in Figures 1 and 2, using these formulae HF% = 100 × [liver SI IP/spleen SI IP] - [liver SI OP/spleen SI OP]/2 × [liver SI IP/spleen SI IP]. Percentage of 6%-26.1% represented mild steatosis, 26.2%-36.8% represented moderate steatosis and > 36.8% represented severe steatosis. Hepatic fat fraction = 100 × [liver SI IP-liver SI OP]/2 × liver SI IP; where ≥ 5% represented mild steatosis, ≥ 33% represented moderate steatosis and ≥ 66% represented severe steatosis[8,9].

Figure 1 Magnetic resonance imaging of one of our patients exhibiting mild steatosis. A-C: Magnetic resonance imaging scan of a female patient, demonstrate axial in-phase; D-F: Out-of-phase sequences showing region of interest, calculated hepatic fat percentage (6.8%) and fepatic fat fraction (10.5%) matching with mild steatosis.

Figure 2 Magnetic resonance imaging of one of our patients exhibiting moderate steatosis. A-C: Magnetic resonance imaging scan of a female patient, demonstrate axial in-phase; D-F: Out-of-phase sequences showing applied region of interest (ROI’s), calculated hepatic fat percentage (32.2%) and hepatic fat fraction (33%), matching with moderate steatosis.

Candidate variables

Participants across the three categories were matched for sex, age, and BMI. The variables studied were TSH, FT4, total cholesterol, triglycerides, LDL, HDL, AST, ALT, and MASLD assessment by ultrasound, with diagnosis confirmed by MRI and HF%. Furthermore, the ultrasound results were compared to those of the MRI to determine the degree of agreement between both diagnostic tools.

Outcome

HF% of 6%-26.1%, as calculated by the MR scans, were classified as mild steatosis, 26.2%-36.8% as moderate steatosis, and greater than 36.8% as severe steatosis[9]. Figures 1 and 2 illustrate mild and moderate steatosis on MRI in two of our patients. The subjects were then classified into non-MASLD and MASLD groups to ascertain the risk variables associated with the onset and severity of MASLD. Variables compared between the two groups included TSH, FT4, duration of hypothyroid disorder, ALT, AST, FIB-4 index, total cholesterol, triglycerides, HDL, LDL, and HF%. Finally, TSH was correlated with each of the variables.

Feature selection

A sequential logistic regression analysis was performed to assess the independent effect of all variables that influence the onset and severeness of MASLD. Factors with a threshold of significance below 0.100 were selected. The regression coefficient is a measure of the impact of each variable on the model, while also accounting for the influence of other factors.

Sample size

This observational study sought to evaluate the association between MASLD and overt, or subclinical hypothyroidism in comparison to healthy controls. According to Tahara et al[10] and Xu et al[11]: The proportion of NAFLD in subclinical hypothyroidism was 0.343, while in the control group it was 0.1077. Consequently, 48 people were necessary for evaluation in each group (resulting in a total sample of 144 participants) to dismiss the null hypothesis, concerning the proportion of the population affected by NAFLD among individuals with clinical and subclinical hypothyroidism and healthy controls is the same, with a statistical power of 0.8. The risk of a type I error associated with this null hypothesis test is 0.05. The sample size was established utilizing the PS software (version 3).

Statistical analyses

Data analysis was conducted using SPSS version 28. The quantitative data were expressed as means, standard deviations, medians, and ranges, where appropriate. The categorical data were displayed simultaneously as percentages and frequency counts. The normality of the data was evaluated utilizing the Kolmogorov-Smirnov test in conjunction with the Shapiro-Wilk test. The Student’s t-test was utilized to compare the two groups concerning numerically distributed variables following a normal distribution.

In contrast, the Mann-Whitney test was utilized to assess quantitative variables demonstrating a non-normal distribution. Whenever appropriate, either the analysis of variance or the non-parametric Kruskal-Wallis H test was utilized to compare quantitative data among more than two groups. Categorical data categories were analyzed utilizing the χ² test or Fisher's exact test, as appropriate. All statistical analyses employed a two-tailed approach, and statistical significance was established at a P value threshold of 0.05 or less.

Dalia Abdelfatah, a lecturer in the Biostatistics Department at Cairo University’s National Cancer Institute, Egypt, reviewed the statistical methods used in this investigation.

RESULTS

Summary of the sociodemographic and laboratory characteristics of the three groups

Our study was conducted on 144 participants fulfilling the inclusion criteria, including three groups: 48 overt hypothyroid patients, 48 subclinical hypothyroid patients, and 48 euthyroid controls. Table 1 demonstrates that the three groups exhibited similarity regarding sex, age, and BMI. The mean age ranged from 36 years to 39 years, while the mean BMI spanned from 30.7 kg/m² to 31.5 kg/m².

Table 1 Sociodemographic characteristics of the three study groups, n (%).

Variable	Control (n = 48)	Subclinical hypothyroidism (n = 48)	Overt hypothyroidism (n = 48)	P value
Gender
Female	38 (79.2)	40 (83.3)	41 (85.4)	0.792
Male	10 (20.8)	8 (16.7)	7 (14.6)
Age (years), mean ± SD	39 ± 12	39 ± 8	36 ± 9	0.209
BMI (kg/m2), mean ± SD	30.7 ± 2.3	30.7 ± 2.6	31.5 ± 2.4	0.193

P value < 0.05 is considered significant. BMI: Body mass index.

According to Table 2, the TSH levels of overt hypothyroid patients were substantially elevated in comparison to those of the subclinical and control groups. Furthermore, Individuals exhibiting subclinical and overt hypothyroidism displayed markedly elevated levels of low-density lipoprotein, triglycerides, and total cholesterol in comparison to the control cohort. Furthermore, free T4 levels were markedly decreased in individuals exhibiting both overt and subclinical hypothyroidism in comparison to healthy control subjects. Nevertheless, no statistically significant disparities were detected across the research groups regarding AST, ALT, and HDL values.

Table 2 Laboratory characteristics of study groups, mean ± SD.

Variable	Overt hypothyroidism (A)	Subclinical hypothyroidism (B)	Control (C)	P value	Pairwise comparison
TSH (mIU/L)	27.6 ± 29	7.1 ± 2.4	2.5 ± 1	< 0.001	A vs B: < 0.001
					A vs C: < 0.001
					B vs C: 0.553
FT4 (ng/dL)	1 ± 0.4	1.1 ± 0.2	1.2 ± 0.2	< 0.001	A vs C: < 0.001
					B vs C: 0.009
					A vs B: 0.250
ALT (U/L)	35 ± 64.5	18.7 ± 10.2	20.6 ± 9.5	0.077	A vs B: 0.115
					A vs C: 0.203
					B vs C: 1.000
AST (U/L)	29.8 ± 33.9	20.8 ±7.4	19.7 ± 6.5	0.033	A vs B: 0.100
					A vs C: 0.051
					B vs C: 1.000
Total cholesterol (mg/dL)	206.8 ± 37.9	178 ± 30.8	151.6 ± 35.5	< 0.001	A vs B: < 0.001
					A vs C: < 0.001
					B vs C: < 0.001
High density lipoprotein (mg/dL)	50.4 ± 12.5	48.9 ± 12.5	46.3 ± 18.7	0.381	A vs B: 1.000
					A vs C: 0.510
					B vs C: 1.000
Low density lipoprotein (mg/dL)	130.8 ± 34.2	110.4 ± 26.3	87.2 ± 36.6	< 0.001	A vs B: 0.008
					A vs C: < 0.001
					B vs C: 0.002
Triglycerides (mg/dL), median (range)	107 (31-356)	84 (31-322)	73.5 (37-344)	0.004	A vs B: 0.032
					A vs C: 0.005
					B vs C: 0.564

P value < 0.05 is considered significant. TSH: Thyroid-stimulating hormone; FT4: Free thyroxine; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase.

MASLD assessment among the three study groups

The three groups were assessed for the presence of MASLD using ultrasonography, with diagnosis confirmed by MRI.

As shown in Table 3, ultrasonography identified 35 patients with MASLD in the overt hypothyroid group, whereas MRI confirmed the diagnosis in only 34 patients. However, the results of both modalities matched in the 27 patients in the subclinical hypothyroid group and in the three subjects in the control group diagnosed with MASLD. The comparison of MASLD prevalence assessed by ultrasonography and MRI across the three clinical groups was statistically significant (P < 0.001 for both modalities), with high agreement between ultrasonography and MRI (kappa = 0.99, P < 0.001).

Table 3 Prevalence of metabolic dysfunction-associated steatotic liver disease among the three study groups, n (%).

Variable	Overt hypothyroidism (A) (n = 48)	Subclinical hypothyroidism (B) (n = 48)	Control (C) (n = 48)	P value	Paire-wise comparison
MASLD by ultrasound
Non-MASLD	13 (27.1)	21 (43.8)	45 (93.8)	< 0.001	A vs B: 0.135
MASLD	35 (72.9)	27 (56.3)	3 (6.3)		A vs C: < 0.001
					B vs C: < 0.001
MASLD by MRI
Non-MASLD	14 (29.2)	21 (43.8)	45 (93.8)	< 0.001	A vs B: 0.203
MASLD	34 (70.8)	27 (56.3)	3 (6.3)		A vs C: < 0.001
					B vs C: < 0.001

MASLD: Metabolic dysfunction-associated steatotic liver disease; MRI: Magnetic resonance imaging.

Moreover, pairwise comparisons indicated a statistically significant association between MASLD and overt hypothyroidism compared with the control group (A vs C: P < 0.001) and between MASLD and subclinical hypothyroidism compared with the control group (B vs C: P < 0.001).

Sociodemographic and laboratory characteristics of the MASLD vs the non-MASLD group

The participants’ data were then categorized into non-MASLD and MASLD groups to identify the risk factors for the development of MASLD and its severity.

The sociodemographic profiles of the groups (MASLD and non-MASLD) demonstrated no statistically significant variations in terms of gender, age, and BMI (P values = 0.961, 0.019, and 0.091, respectively). Nevertheless, a statistically significant distinction exists in the duration of thyroid dysfunction (overt and subclinical) between MASLD (mean 4.2 ± 1.8 years) and the non-MASLD (mean 2.1 ± 1.2 years) groups (P value < 0.001), as illustrated in Table 4. The laboratory features of MASLD participants, in comparison to non-MASLD subjects, demonstrated that MASLD patients exhibited substantially elevated levels of TSH total cholesterol, triglycerides, and low-density lipoprotein. Furthermore, free T4 concentrations were markedly decreased in MASLD patients relative to non-MASLD people. Notwithstanding the lack of substantial variations in HDL, ALT, and AST between the two cohorts, the mean ALT (30.3 ± 56.5 vs 20.3 ± 10.1) and AST (27.2 ± 29.6 vs 27.2 ± 29.6) levels were elevated in the MASLD group compared to the non-MASLD group, as seen in Table 5.

Table 4 Sociodemographic characteristics of metabolic dysfunction-associated steatotic liver disease and non-metabolic dysfunction-associated steatotic liver disease groups, n (%).

Variable	MASLD	Non-MASLD	P value
	n = 64/144 (%)1	n = 80/144 (%)1
Gender
Female	53 (44.5)	66 (55.5)	0.961
Male	11 (44)	14 (56)
Age (years), mean ± SD	36 ± 9	40 ± 10	0.019
BMI (kg/m2), mean ± SD	31.3 ± 2.2	30.6 ± 2.6	0.091
Duration of hypothyroidism (years), mean ± SD	4.2 ± 1.8	2.1 ± 1.2	< 0.001

¹Percentages were calculated within rows.

P value < 0.05 is considered significant. BMI: Body mass index; MASLD: Metabolic dysfunction-associated steatotic liver disease.

Table 5 Laboratory characteristics of metabolic dysfunction-associated steatotic liver disease and non-metabolic dysfunction-associated steatoic liver disease groups, mean ± SD.

Variable	MASLD	Non-MASLD	P value
ALT (U/L)	30.3 ± 56.5	20.3 ± 10.1	0.123
AST (U/L)	27.2 ±29.6	20.5 ± 7.6	0.054
TSH (mIU/L)	22.9 ± 26.4	4.1 ± 3	< 0.001
FT4 (ng/dL)	1 ± 0.3	1.2 ± 0.2	< 0.001
Total cholesterol (mg/dL)	191.8 ± 40.2	168.4 ± 39.4	< 0.001
High density lipoprotein (mg/dL)	48.7 ± 11.6	48.4 ± 17.1	0.892
Low density lipoprotein (mg/dL)	118.3 ± 32.1	102.4 ± 39.3	0.009
Triglycerides (mg/dL, median (range)	106 (33-356)	76 (31-344)	< 0.001

P value < 0.05 is considered significant. ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; FT4: Free thyroxine; TSH: Thyroid-stimulating hormone; MASLD: Moderate metabolic dysfunction-associated steatotic liver disease.

Correlations of TSH with laboratory and imaging parameters

TSH exhibited a statistically significant positive association with total cholesterol, triglycerides, LDL, HF%, and AST within the MASLD group, in contrast to the non-MASLD one. (P value < 0.001, 0.004, < 0.001, < 0.001, and 0.006, respectively). The strongest correlation was observed with hepatic fat content, showing a significant positive correlation, as shown in Table 6.

Table 6 Correlation between thyroid stimulating hormone and other laboratory and imaging parameters.

Variable	TSH (mIU/L)		Degree of correlation
	r	P value
Age (years)	-0.12	0.163	Non-significant correlation
BMI (kg/m2)	0.08	0.348	Non-significant correlation
ALT (U/L)	0.12	0.164	Non-significant correlation
AST (U/L)	0.23	0.006	Significant little positive correlation
Total cholesterol (mg/dL)	0.38	< 0.001	Significant fair positive correlation
Triglycerides (mg/dL)	0.24	0.004	Significant little positive correlation
HDL (mg/dL)	0.13	0.131	Non-significant correlation
LDL (mg/dL)	0.33	< 0.001	Significant fair positive correlation
Fib4 score	0.26	0.168	Non-significant correlation
HF%	0.69	< 0.001	Significant good positive correlation

r is the correlation coefficient, which ranges from -1 to 1; P value < 0.05 is considered significant. BMI: Body mass index; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; HDL: High density lipoprotein; LDL: Low density lipoprotein; Fib4: Fibrosis-4 index; HF%: Hepatic fat percentage.

MRI classified the degree of steatosis in the MASLD group based on HF% as mild (6%-26.1%), moderate (26.2%-36.8%), or severe (> 36.8%). Twenty-three patients had mild steatosis, 31 had moderate steatosis, and none had severe steatosis. The degree of steatosis was higher in the overt hypothyroid group than in the subclinical group. Mild steatosis was observed in 38.2% and 63% of the overt and subclinical groups, respectively, whereas moderate steatosis was observed in 61.8% of the overt group and 10% of the subclinical group, with none exhibiting severe steatosis. This distribution showed a statistically significant difference in MRI-measured steatosis degree among the three groups (P value = 0.029), as illustrated in Table 7.

Table 7 The degree of steatosis among the three groups by ultrasonography and magnetic resonance imaging, n (%).

Variable	Overt hypothyroidism (A)	Subclinical hypothyroidism (B)	Control (C)	P value
Sonar
Mild MASLD	14 (40)	17 (63)	3 (100)	0.046
Moderate MASLD	21 (60)	10 (37)	0 (0)
Degree by MRI
Mild MASLD	13 (38.2)	17 (63)	3 (100)	0.029
Moderate MASLD	21 (61.8)	10 (37)	0 (0)

MASLD: Moderate metabolic dysfunction-associated steatotic liver disease; MRI: Magnetic resonance imaging.

Demographic and laboratory characteristics of the mild MASLD group vs the moderate MASLD one

Concerning the demographic and laboratory characteristics of the mild MASLD group compared to the intermediate MASLD one, cases with moderate MASLD had significantly higher TSH levels, a higher HF%, and substantially lower FT4 levels compared to mild cases (P value 0.005, < 0.001, and 0.039, respectively). Although there were no statistically significant differences in ALT or AST between the two groups, the mean ALT (19.2 ± 8.1 vs 42.1 ± 79.6) and AST (20.5 ± 5.3 vs 34.2 ± 41.4) values were found to be elevated in cases of MASLD group relative to the mild MASLD one as demonstrated in Table 8.

Table 8 Comparing demographic and laboratory characteristics among the mild and moderate metabolic dysfunction-associated steatotic liver disease groups, mean ± SD.

Variable	Mild MASLD	Moderate MASLD	P value
Age (years)	36 ± 9	36 ± 9	0.940
BMI (kg/m2)	31 ± 2	31.7 ± 2.3	0.238
Duration of hypothyroidism (years)	4.2 ± 1.6	4.3 ± 2	0.821
ALT (U/L)	19.2 ± 8.1	42.1 ± 79.6	0.106
AST (U/L)	20.5 ± 5.3	34.2 ± 41.4	0.065
TSH (mIU/L)	13.8 ± 16.1	32.6 ± 31.5	0.005
FT4 (ng/dL)	1 ± 0.2	0.9 ± 0.4	0.039
Total cholesterol (mg/dL)	186.2 ± 32.4	197.7 ± 47	0.253
High density lipoprotein (mg/dL)	48.8 ± 10.4	48.5 ± 13	0.919
Low density lipoprotein (mg/dL)	114.2 ± 25.8	122.6 ± 37.7	0.299
HF (%)	15.9 ± 7.8	30.8 ± 3.5	< 0.001
Triglycerides (mg/dL), median (range)	105 (39-322)	107 (33-356)	0.777
Fib4 score, median (range)	0.3 (0.1-0.8)	0.3 (0.1-0.7)	0.438

P value < 0.05 is considered significant. MASLD: Moderate metabolic dysfunction-associated steatotic liver disease; BMI: Body mass index; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; TSH: Thyroid-stimulating hormone; FT4: Free thyroxine; HF: Hepatic fat; Fib4: Fibrosis-4 index.

TSH exhibited a statistically significant positive correlation regarding the extent of hepatic steatosis in individuals diagnosed with MASLD (P value < 0.001), as presented in Table 9.

Table 9 Thyroid stimulating hormone level correlation to steatosis grading by ultrasound and magnetic resonance imaging.

Variable	TSH (mIU/L)		Paire-wise comparison
	mean ± SD	P value
Ultrasound
Non-MASLD (A)	4.1 ± 3	< 0.001	A vs B: 0.023
Mild MASLD (B)	13.4 ± 16		A vs C: < 0.001
Moderate MASLD (C)	32.6 ± 31.5		B vs C: < 0.001
Degree by MRI
Non-MASLD	4.1 ± 3	< 0.001	A vs B: 0.017
Mild MASLD	13.8 ± 16.1		A vs C: < 0.001
Moderate MASLD	32.6 ± 31.5		B vs C: < 0.001

P value < 0.05 is considered significant. TSH: Thyroid stimulating hormone; MASLD: Moderate metabolic dysfunction-associated steatotic liver disease; MRI: Magnetic resonance imaging.

Multivariate analysis

To evaluate the independent impacts of all factors affecting MASLD risk and severity, Variables exhibiting a significance level less than 0.100 were selected for inclusion in a stepwise logistic regression analysis. The regression coefficient indicates the effect of each variable while considering the influence of other factors in the model. The model indicated that TSH levels are the primary predictors of MASLD development. For every unit rise in TSH level, the incidence of MASLD escalates by 18% [95% confidence interval (CI) range for the odds ratio (OR): 1.04-1.33] as seen in Table 10.

Table 10 The variables that were significant in the stepwise logistic regression.

Variable	B	SE	OR	95%CI for OR	P value
TSH (mIU/L)	0.2	0.1	1.18	1.04-1.33	0.010

TSH: Thyroid-stimulating hormone; B: Regression coefficient; SE: Standard error; OR: Odds ratio; CI: Confidence interval.

Regarding MASLD severity, the model showed that TSH level was the most important predictor of MASLD severity. With each unit rise in TSH, the chance of MASLD severity increased by 4% (95%CI for OR: 1.01-1.06), as shown in Table 11.

Table 11 The variable that was significant in the stepwise logistic regression.

Variable	B	SE	OR	95%CI for OR	P value
TSH (mIU/L)	0.03	0.01	1.04	1.01-1.06	0.011

TSH: Thyroid-stimulating hormone; B: Regression coefficient; SE: Standard error; OR: Odds ratio; CI: Confidence interval.

Moreover, the receiver operating characteristic (ROC) curve, with a sensitivity of 83% and a specificity of 89%, showed that TSH > 6.1 mIU/L is correlated with an elevated likelihood of developing MASLD, as illustrated in Table 12 and Figure 3.

Figure 3 Receiver operating characteristic curve for thyroid stimulating hormone and risk of metabolic dysfunction-associated steatotic liver disease development.

Table 12 Receiver-operating characteristic curve, with sensitivity 83% and specifity 89%, showing thyroid-stimulating hormone > 6.1 (mIU/L) is associated with a higher risk of metabolic dysfunction-associated steatotic liver disease development.

Variable	Cut-off point	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	AUC	95%CI for AUC	P value
TSH (mIU/L)	> 6.1	83	89	86	87	0.88	0.82-0.93	< 0.001

PPV: Positive predictive value; NPV: Negative predictive value; AUC: Area under the curve; CI: Confidence interval TSH: Thyroid-stimulating hormone.

DISCUSSION

MASLD is one of the most prevalent metabolic diseases worldwide, with prevalence among adults estimated to be approximately 30% (around 1.66 billion individuals), and with global prevalence rising by 1.0% annually over the past three decades. MASLD is associated with an increased risk of liver-related complications as well as multiple extrahepatic manifestations. Liver-related complications include hepatic steatosis and metabolic dysfunction-associated steatohepatitis, with 40% of affected individuals at risk of progression to liver fibrosis and subsequent cirrhosis and HCC. Furthermore, extrahepatic manifestations include an increased risk of extrahepatic cancers, such as colon, esophageal, gastric, and pancreatic cancers, in addition to a fivefold higher risk of chronic kidney disease progression and a 3.3-fold higher risk of chronic kidney disease incidence. Moreover, cardiovascular disease is the leading cause of mortality in patients with MASLD due to a 1.5-fold increased risk of cardiovascular events. Consequently, the rising global burden of MASLD is alarming, underscoring the urgent need to prevent and treat metabolic risk factors[12].

On the other hand, hypothyroidism is a state of thyroid hormone deficiency that may be overt or subclinical, as defined by biochemical parameters. The prevalence of overt and subclinical hypothyroidism is estimated to be 0.4% and 9% respectively[13]. TH play a pivotal role in metabolic pathways, affecting almost every nucleated cell via nuclear thyroid hormone receptors; thyroid hormone receptor (THR)α is abundant in hepatic stellate cells, whereas THRB predominates in hepatocytes. TH regulate cholesterol and carbohydrate metabolism through direct effects on gene expression, as well as through cross-talk with other nuclear receptors, including peroxisome proliferator-activated receptor, liver X receptor, and bile acid signaling pathways. TH also modulate hepatic insulin sensitivity, which is particularly important for suppressing hepatic gluconeogenesis, highlighting the intertwined thyroid-liver relationship[14]. Our study aimed to assess the previously overlooked association between the full spectrum of thyroid dysfunction (overt and subclinical) and MASLD.

Our observational study was conducted on 144 Egyptian adults recruited according to the inclusion and exclusion criteria to assess 48 overt hypothyroid patients, 48 patients diagnosed with subclinical hypothyroidism, along with an age-, sex-, and BMI-matched group of 48 control subjects. The variables studied were TSH, FT4, total cholesterol, triglycerides, LDL, HDL, AST, ALT, and MASLD assessment by ultrasound, with diagnosis confirmed by MRI and HF% calculated. Furthermore, the ultrasound results were compared to those of the MRI to determine the degree of agreement between both diagnostic tools.

Initially, total cholesterol, LDL, and triglyceride levels were markedly elevated in both overt and subclinical hypothyroid individuals compared to the euthyroid group, with substantially greater values seen in overt hypothyroidism relative to the subclinical group (P value < 0.001), as it is well established in previous literature. Dyslipidemia linked to hypothyroidism increases the risk of MASLD development, as diminished T3 levels result in decreased activity of sterol regulatory element-binding protein-2 (SREBP-2), reduced lipoprotein lipase, and decreased plasma cholesteryl ester transfer protein activity. Lowered effect on SREBP-2 causes a reduction of the cell surface LDL receptors, resulting in high levels of LDL and apolipoprotein B lipoproteins. Low lipoprotein lipase levels lead to decreased cholesterol and plasma triglyceride clearance, decreased cholesteryl ester transfer protein activity, and enhanced cholesterol shift from HDL to LDL[15]. The aforementioned alterations culminate in hepatic lipid buildup, resulting in heightened hepatic insulin resistance due to diminished insulin receptor substrate-mediated phosphatidylinositol 3-kinase/protein kinase B signaling, which enhances hepatic de novo lipogenesis and promotes the development of hepatocyte lipid droplets (specialized organelles for neutral lipid storage), giving rise to macrovesicular steatosis, cellular lipotoxicity, organelle dysfunction, pathogenic gut microbiota overabundance, and subsequent chronic hepatic inflammation and progressive liver injury[16].

In the subsequent analysis, thirty-four patients from the overt hypothyroid group, twenty-seven from the subclinical hypothyroid group, and three subjects from the euthyroid group were diagnosed with MASLD via MRI, demonstrating a statistically significant association between overt hypothyroidism and the onset of MASLD in comparison to the euthyroid group (P value < 0.001), as well as a statistically significant correlation between subclinical hypothyroidism and MASLD relative to the euthyroid group (P value < 0.001). The prevalence of advanced steatosis was significantly higher within the overt hypothyroid group relative to the subclinical one concerning the severity of MASLD. Mild steatosis was detected in 38.2% and 63% of the overt and subclinical groups, respectively. Moderate steatosis was detected in 61.8% of the overt group and 10% of the subclinical group, with none exhibiting severe steatosis according to the HF% calculated by the MR scans, with TSH significantly positively correlated with HF%.

Analysis of laboratory data indicated that individuals with MASLD had statistically significant elevations in TSH, total cholesterol, LDL, and triglycerides compared to those without MASLD, thereby identifying potential risk factors for MASLD development. The aforementioned variables were then entered into a stepwise logistic regression to assess independent effects; multivariate analysis revealed that TSH was the most significant predictors of MASLD risk. For each unit increase in TSH level, the probability of MASLD increased by 18% (95%CI for OR: 1.04-1.33), whereas TSH level emerged as the primary predictor of MASLD severity. For each unit increase in TSH, the risk of MASLD severity increased by 4% (95%CI for OR: 1.01-1.06). The results align with the Rotterdam research[17], a thorough population-based prospective cohort study, which showed that increased TSH levels are associated with an augmented risk of clinically severe fibrosis in the previously called NAFLD, as assessed by elastography. Tahara et al[10], Fan et al[18], and Kim et al[19] similarly posited that the spectrum of thyroid dysfunction is an independent predictor of the onset of MASLD. We used the ROC curve to determine the cut-off value for TSH associated with an enhanced risk of MASLD development, revealing that a TSH level > 6.1 mIU/L is associated with an increased risk of MASLD (sensitivity 83% and specificity 89%).

Mechanistically, hypothyroidism can induce hepatic steatosis independently of the associated dyslipidemia through various pathways. First, the loss of intracellular signal transducers and transcription factors (2/3) inhibition leads to phosphorylation events within the transforming growth factor beta signaling cascade, thereby enhancing fibrosis-related gene expression. Furthermore, TSH facilitates hepatic steatosis by activating SREBP signaling pathways, inhibiting bile acid synthesis, and diminishing cholesterol production through AMP-activated protein kinase-mediated phosphorylation of 3-hydroxy-3-methylglutaryl-coenzyme A reductase. TSH functions as an autonomous regulator of hepatic lipid and cholesterol equilibrium. Moreover, hepatic autophagy is impaired by diminished TH, as decreased T3 levels inhibit AMP-activated protein kinase activation and reduce autophagic flux, resulting in compromised lipophagy. Thyroid dysfunction may contribute to the development of steatohepatitis through oxidative stress, mitochondrial impairment, and the generation of reactive oxygen species induced by TSH, leading to lipid peroxidation, depletion of polyunsaturated fatty acids, and disruption of SREBP-1c regulation and peroxisome proliferator-activated receptor-α; thus, cytoprotective mechanisms are compromised in hypothyroidism. The alteration of tumor necrosis factor-alpha, adiponectin, leptin, and nuclear factor kappa-light-chain-enhancer of activated B cells signaling, exacerbated by TSH, establishes a deleterious feedback loop linking inflammation with insulin resistance. A crucial mechanism involves the activation of hepatic stellate cells via reduced THRα stimulation, which normally facilitates antifibrotic signaling, leading to collagen I deposition and increased extracellular matrix (ECM) stiffness via integrin-mediated signaling. Recent data indicate that the hepatic ECM transcends its role as a mere structural support, functioning as an active modulator of intercellular communication and thereby influencing lipid metabolism, immune homeostasis, and fibrogenesis. In hypothyroid states, enhanced ECM production, amplified by chronic inflammation, drives the progression from steatosis to fibrosis, cirrhosis, and ultimately hepatic failure. The reciprocal interaction between the thyroid and liver is illustrated by sphingolipids, specifically ceramide-derived sphingosine-1-phosphate, whose accumulation within the ECM microenvironment is crucial for immune activation and fibrotic transformation. The inhibition of sphingosine-1-phosphate is thought to induce thyroid inflammation, highlighting the bidirectional relationship between thyroid dysfunction and hepatic inflammation mediated through ECM pathways. Finally, the immunopathological etiology of some thyroid illnesses, such as Hashimoto’s thyroiditis, may lead to chronic systemic inflammation and disrupted adipokine signaling, which in turn fosters hepatic inflammation and ECM accumulation in the context of thyroid hormone dysregulation[20].

Understanding these mechanistic pathways provides opportunities for therapeutic interventions. For instance, resmetirom, oral THR-β agonist, the first Food and Drug Administration approved drug for the treatment of MASLD. Resmetirom proved to improve steatohepatitis, liver fibrosis by at least one stage and even lowering LDL. It works by mitigating chemically induced liver fibrosis by reducing collagen deposition, regulating mitochondrial bioenergetics, lipid metabolism, cholesterol homeostasis, and fatty acid oxidation and normalizing autophagic flux achieving[21,22].

The increasing prevalence of MASLD-associated HCC has emerged as a considerable problem, necessitating rigorous MASLD monitoring for at-risk individuals. MRI is regarded as the most precise non-invasive technique for identifying and measuring hepatic fat accumulation; however, its limited availability and high cost hinder its routine use[23]. Our work demonstrates strong concordance between MRI and ultrasonography, consistent with the results reported by Fishbein et al[24] and Kromrey et al[25], both of whom found that hepatic MRI and ultrasonography are proficient at detecting fat accumulation associated with fatty liver disease. In the aforementioned studies, ultrasonography identified hepatic steatosis in 37.8% of 112 individuals, whereas MRI detected steatosis in 40% of the same cohort. Consequently, we assert that ultrasonography, due to its cost-effectiveness, broad accessibility, and lack of radiation exposure, may be a feasible option for MASLD screening in hypothyroid individuals.

Despite the FIB-4 score being a straightforward, non-invasive tool, our findings showed no statistically significant association between FIB-4 and MASLD severity. Nonetheless, this may be attributed to the fact that the FIB-4 score was originally designed for individuals with human immunodeficiency virus/hepatitis C virus coinfection and includes platelet count and age in its evaluation[26]. Larger-scale studies with increased sample sizes are necessary to assess the sensitivity and specificity of the FIB-4 score in hypothyroid individuals with MASLD.

Liver enzymes have been proposed as useful in MASLD screening; our study showed that mean ALT and AST levels were higher in the MASLD group than in the non-MASLD group. They were higher in the moderate MASLD group than in the mild MASLD group, consistent with the study by Hossain et al[27], which reported that aminotransferase levels are independent predictors of mild and severe fibrosis. Elevated liver enzymes in MASLD reflect hepatocellular injury occurring in patients with thyroid dysfunction.

To strengthen the clinical implications, addressing the association between thyroid dysfunction and MASLD may promote screening, and treating MASLD patients for thyroid dysfunction may alleviate a substantial portion of the MASLD-related financial and health care burden. Conversely, screening patients with thyroid dysfunction for MASLD may facilitate early detection and appropriate prevention of hepatic steatosis progression. Furthermore, if additional research strengthens the association between subclinical hypothyroidism and MASLD at specific TSH levels, the cut-offs for treating subclinical hypothyroidism may be lowered compared with those currently recommended by endocrinology societies, to prevent the adverse consequences of MASLD in these patients. Finally, the spectrum of thyroid dysfunction may be incorporated into the current criteria for metabolic risk factors in MASLD development, enabling better recognition of diverse disease determinants.

Regarding the strengths of our study, to our knowledge, it is among the few studies to confirm the diagnosis of MASLD using an MR scanner with axial, coronal, and sagittal STIR FSE sequences, along with hepatic fat content calculation, and to define a specific TSH level above which the risk of MASLD development increases using a ROC curve, in addition to assessing the relationship between both overt and subclinical hypothyroidism and MASLD. Conversely, the limitations include the absence of a causal relationship and the lack of long-term observation of MASLD progression in well-controlled patients receiving medical treatment compared with uncontrolled patients.

CONCLUSION

Our research delineates a substantial association between the spectrum of hypothyroidism, encompassing both overt and subclinical forms, and the prevalence of MASLD. The TSH level constitutes an independent risk factor influencing the initiation and severity of MASLD. Elevated TSH levels correlate with an augmented chance of developing MASLD and a heightened probability of advancing to severe steatotic stages (increasing hepatic fat content). ROC curve research indicated that TSH > 6.1 mIU/L correlates with an elevated risk of MAFLD development.

ACKNOWLEDGEMENTS

We wish to extend our sincere appreciation to all personnel within the Department of Internal Medicine at Kasr Alainy, Cairo University.