Noninvasive Indices of MASLD Are Associated With Hypogonadism in Male Patients With Type 2 Diabetes Mellitus

Abstract

Context

Metabolic dysfunction–associated steatotic liver disease (MASLD) is the most prevalent chronic liver disease, affecting one-fourth of the adult population worldwide. Recent data found an association between MASLD and hypogonadism, but this relation in patients with type 2 diabetes mellitus (T2DM) is still unclear.

Objective

To evaluate in men with T2DM the association between total testosterone (TT) and noninvasive indices of hepatic steatosis (Fatty Liver Index [FLI], Hepatic Steatosis Index [HSI], Dallas Steatosis Index [DSI]) and fibrosis (AST to Platelet Ratio Index [APRI], Fibrosis-4 Index [FIB-4]), and their predictive cutoff values in identifying hypogonadism.

Methods

Cross-sectional study on 189 men with T2DM, without history of liver diseases and alcoholism, recruited on an outpatient basis. Interventions were andrological evaluation, metabolic parameters, TT, and liver indices. The main outcome measures were comparison of steatosis and fibrosis indices with testosterone levels and presence of hypogonadism. Receiver operating characteristic curves were used to identify cutoff values of liver indices in predicting low testosterone (<12 nmol/L).

Results

FLI, HSI, and DSI were negatively related with TT and were higher in the low-testosterone group than in the normal-testosterone group (FLI: 74.1 [61.4-93.5] vs 56.5 [32.1-78.2], P < .001; HSI: 41.5 [39.2-45.9] vs 40.1 [36.6-43.2], P = .005; DSI: 0.45 [−0.08-+1.04] vs −0.07 [−1.02-+0.58], P < .001). FLI and DSI also correlated with clinical symptoms of hypogonadism. No differences between groups were observed for APRI and FIB-4. FLI ≥63 was the best parameter as predictive index of low TT (sensitivity 73%, specificity 64%).

Conclusion

We found an association between noninvasive indices of steatosis and hypogonadism in patients with T2DM. These indices could be used to direct the patients to andrological evaluation.

type 2 diabetes, liver disorders, hypogonadism, metabolic syndrome

Nonalcoholic fatty liver disease (NAFLD) is defined by the evidence of hepatic steatosis on imaging or histology in the absence of secondary causes (1). Currently, it is recognized as the most prevalent chronic liver disease, affecting approximately one-fourth of the adult general population worldwide (2, 3). The prevalence is even higher in patients affected by metabolic syndrome, reaching up to 75% in presence of type 2 diabetes mellitus (T2DM), due to a bidirectional pathophysiological link between these 2 entities (4-6). NAFLD is a spectrum of histological liver disease, ranging from simple steatosis to chronic nonalcoholic steatohepatitis (NASH), and over time it can progress to cirrhosis and organ failure. Moreover, there is an increased risk of developing hepatocellular carcinoma (7). The metabolic syndrome and its components features, especially T2DM, are recognized as independent risk factors for this spectrum disease progression. Furthermore, the association of NAFLD with metabolic syndrome increases the all-cause, liver-specific, and cardiovascular mortality risk (8).

Given the strict connection between these entities, in 2020 an international panel of experts suggested a name change from NAFLD to metabolic dysfunction–associated fatty liver disease (MAFLD) (9, 10). In 2023 the multisociety Delphi consensus statement replaced both NAFLD and MAFLD with the nomenclature metabolic dysfunction–associated steatotic liver disease (MASLD), providing a nonstigmatizing description of the condition. In this new definition, the presence of at least 1 cardiometabolic risk factor in addition to hepatic steatosis is required (11).

Liver biopsy is the gold standard both for diagnosis and staging of MASLD, but it has limited application in routine clinical practice due to its invasiveness, cost, intrinsic small risk of complication, poor patient acceptability, and sampling variability (12). For this reason, many noninvasive biomarkers have been developed and are currently approved for assessment of MASLD in the clinical practice. Among them, the Fatty Liver Index (FLI), Hepatic Steatosis Index (HSI), and Dallas Steatosis Index (DSI) are 3 validated noninvasive indices of hepatic steatosis used for large-scale screening studies (13). These scores are associated with insulin resistance but they cannot be used to estimate the severity of steatosis (14). Fibrosis-4 index (FIB-4) and AST (aspartate aminotransferase) to Platelet Ratio Index (APRI) are 2 noninvasive markers of hepatic fibrosis, which is correlated with liver-related outcomes and mortality (15).

The hepatocytes express sex steroid receptors, therefore sex steroids can directly impact hepatic metabolism and detoxification mechanisms protecting against or promoting hepatic disease (16).

Preclinical studies demonstrated that testosterone could influence hepatic insulin sensitivity and lipid metabolism and so it might be partly responsible for the pathophysiology of inflammatory and fibrotic progression of MASLD (17, 18).

Some studies have found an association between MASLD and low total testosterone (TT) levels both in a murine model (19) and in men (20-25). Other studies instead correlated low free testosterone with the presence and severity of NASH and fibrosis (26, 27). A significant correlation between liver fibrosis (estimated with APRI score) and TT and sex hormone–binding globulin (SHBG) levels was also found in a cohort of HIV-infected patients, which is a group at very high risk of hypogonadism due to many mechanisms (primary testicular disorder, secondary hypothalamic–pituitary dysfunction, liver-derived SHBG elevation with consequent reduction of free testosterone) (28).

Metabolic syndrome is frequently associated with functional hypogonadism, a potentially reversible impairment of the hypothalamic–pituitary–testis axis, characterized by a mild lowering of testosterone concentrations with low or inappropriately normal luteinizing hormone (LH) (29, 30).

In a previous study in men with T2DM, our group found that cardiometabolic indices (Visceral Adiposity Index, Triglyceride Glucose Index, and Lipid Accumulation Product) are strongly negatively associated with TT levels (31).

To the best of our knowledge, only 1 study has evaluated the correlation of noninvasive indices of hepatic steatosis and fibrosis with testosterone plasma levels in men with T2DM finding higher noninvasive indices of steatosis, but not of hepatic fibrosis, in individuals with low TT than those with high testosterone concentrations (22).

Hence, the aim of our cross-sectional study was to further explore the association between testosterone concentrations, prevalence of hypogonadism, and MASLD in patients with T2DM by using noninvasive indices of hepatic steatosis and fibrosis.

Materials and Methods

In this cross-sectional study, 265 participants were prospectively recruited on an outpatient basis at the Division of Metabolic Diseases of the University Hospital of Padova (Italy) and were evaluated at the Unit of Andrology and Reproductive Medicine of the same hospital from January 1, 2016, to December 31, 2019. The inclusion criteria were (1) male gender; (2) age >18 years; (3) T2DM diagnosis; (4) consent to participate in the study; (5) availability of all the biochemical parameters to calculate the liver indices performed at the same laboratory of the University Hospital of Padova. The exclusion criteria were (1) ethanol consumption >20 g/day; (2) history of liver cirrhosis or other liver disease (viral hepatitis, autoimmune hepatitis, primary sclerosing cholangitis, primary biliary cirrhosis, drug-induced liver disease, hemochromatosis, Wilson disease, alpha1-antitrypsin deficiency); (3) clinical signs of uncompensated glycemic control (polyuria, polydipsia, weight loss, hyperosmolar hyperglycemic syndrome); (4) overt diabetic chronic complications: retinopathy, nephropathy (excluding microalbuminuria), neuropathy; (5) known forms of organic hypogonadism (such as Klinefelter syndrome, Kallmann syndrome, cryptorchidism, testicular trauma or torsion, cancer and cancer treatments, orchitis, pituitary disorders); (6) hormonal treatments or drugs interfering with testosterone levels. Glycated hemoglobin (HbA1c) levels, diabetes duration, and the type of antihyperglycemic treatment did not represent selection criteria.

All subjects performed a complete andrological evaluation with medical history, physical examination (including waist circumference, weight and height measurement to calculate body mass index [BMI]), and testicular examination (to assess volumes through Prader's orchidometer). All patients filled out IIEF-5 (International Index of erectile function-5), IPSS (International Prostatic Symptoms Score), and AMSS (Aging Male Symptom Score) questionnaires. The IIEF-5 questionnaire is used to assess the erectile function and to identify the presence of erectile dysfunction for values ≤21 (32); the IPSS questionnaire evaluates the presence of low urinary tract symptoms (33), and the AMSS questionnaire to assess the absence (AMSS ≤26) or presence of signs compatible with mild (AMSS 27-36), moderate (AMSS 37-49), or severe (AMSS ≥50) testosterone deficiency (34).

Morning blood samples (between 8:00 Am and 10:00 Am) were collected after overnight fasting for the following biochemical assays: glycemia, HbA1c, lipid profile (total cholesterol, high-density lipoprotein [HDL] cholesterol, triglycerides), creatinine, AST, alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), platelets, TT, and LH (ECLIA kit, RRID:AB_2800498). Low-density lipoprotein cholesterol was calculated by the Friedewald equation.

To reduce variability in the determination of TT, only the blood samples analyzed at the Central Laboratory of the University Hospital of Padova (ECLIA kit, RRID:AB_2783736) were considered (n = 189). Plasma concentration of TT has been considered normal (≥12 nmol/L), low (8-12 nmol/L), and frankly low (<8 nmol/L), according to guidelines on functional hypogonadism (29). Levels of LH >9.4 UI/L have been considered as elevated (29).

The FLI (35), HSI (36), and DSI (37) were selected as noninvasive indices of MASLD. APRI (38) and FIB-4 (39) were selected as noninvasive indices of hepatic fibrosis. FLI, HSI, DSI, APRI, and FIB-4 were all calculated using automatic online calculators (https://www.mdcalc.com/calc/10001/fatty-liver-index; https://www.mdapp.co/hepatic-steatosis-index-hsi-calculator-357/; https://dsi.wustl.edu/; https://www.hepatitisc.uw.edu/page/clinical-calculators/apri; https://www.hepatitisc.uw.edu/page/clinical-calculators/fib-4).

The formulas for the noninvasive indices of hepatic steatosis and fibrosis are the following:

• FLI (35) = (e^{0.953×loge(triglycerides)+0.139×BMI+0.718×loge(GGT)+0.053×waist circumference −15.745})/(1 + e^{0.953×loge(triglycerides)+0.139×BMI+0.718×loge(GGT)+0.053 × waist circumference −15.745}) × 100

• HSI (36) = 8 × ALT/AST ratio + BMI + 2 (if T2DM) + 2 (if female)

• DSI (37) = −9.4 + 0.316 (if age ≥50 and female) + 2.4 (if known diabetes) + 0.02 * (equals 0 if diabetic; if not diabetic equals the glucose concentration in mg/dL) + 0.3 (if known hypertension) + 0.5 (if Hispanic/Asian/Other race/ethnicity) + Ln triglycerides in mg/dL + 0.4 if ALT 13.5-19.49 IU/L + 1.1 if ALT 19.5-40 IU/L + 1.5 if ALT >40 IU/L + 0.7 if not Black and BMI 25-27.49 kg/m² and +1.4 if not Black and BMI 27.5-34.9 kg/m² + 1.9 if not Black and BMI 35-37.49 kg/m² + 2.6 if not Black and BMI >37.5 kg/m² − 0.2 if Black and BMI 25-27.49 kg/m² +0.8 if Black and BMI 27.5-34.9 kg/m² + 0.8 if Black and BMI 35-37.49 kg/m² + 1.8 if Black and BMI >37.5 kg/m²

• APRI (38) = (AST/upper normal limit of AST) × 100/platelets

• FIB-4 (39) = (age × AST)/(platelets × ALT^1/2)

The study protocol follows the standard clinical approach and the principles outlined in the Declaration of Helsinki. The study was approved by the Ethics Committee of the University Hospital of Padova (protocol no. AOP2845). Informed consent to collect the data anonymously for scientific purposes was obtained from the study participants.

Statistical Analysis

Statistical Package for the Social Sciences software IBM SPSS Statistics, Version 25.0, Armonk (NY) was used for statistical analysis. Since the variables were not normally distributed (by Kolmogorov–Smirnov test), continuous variables were expressed as medians and interquartile ranges (IQRs), while categorical were expressed as percentage.

Within the population analyzed, subdivided into 2 groups by values of TT, comparisons between clinical and biochemical variables were performed by nonparametric tests, in particular the Mann–Whitney test for continuous variables and chi-square test for categorical variables.

In cases of more than 2 groups (population subdivided by quartiles of FLI, HIS, and DSI), Jonckheere's trend test (for continuous variables) and chi-square test for trend (for categorical variables) were used for between-group comparisons. Spearman's coefficient (r_s) was used for binary correlations.

Multiple linear regression analysis was performed to assess whether testosterone concentrations were independently associated with noninvasive indices of MASLD or fibrosis. Variables included in the equations of each noninvasive index were excluded from the list of independent variables because of the risk of multicollinearity.

Receiver operating characteristic curves were performed and analyzed to assess sensitivity, specificity, and optimal cutoff of the different variables in predicting TT <12 nmol/L.

P < .05 was considered to be statistically significant.

Results

Table 1 reports the clinical and biochemical data of overall patients (n = 189). Based on BMI, 46.4% of patients were overweight; 26.8%, 4.9%, and 2.3% were class 1, class 2, and class 3 obesity, respectively.

Table 1.

Clinical and biochemical data of patients and comparison between men with low vs normal TT

	Overall (n = 189)	TT < 12 nmol/L (n = 74)	TT≥12 nmol/L (n = 115)	P
Age (years)	58.0 (51.0-65.0)	58.0 (51.0-64.5)	58.0 (51.0-65.0)	.540
Weight (kg)	86.7 (77.5-96.5)	91.5 (80.0-99.5)	85.0 (76.3-93.5)	.005
BMI (kg/m2)	28.5 (25.8-31.2)	29.6 (26.9-33.7)	28.0 (25.0-30.8)	.003
Waist circumference (cm)	101.0 (94.0-110.0)	105.0 (99.5-118.0)	99.5 (93.0-107.8)	<.001
Glycemia (mg/dL)	143.0 (123.0-180.0)	155.0 (129.3-183.8)	137.0 (120.0-165.5)	.004
HbA1c (%)	7.1 (6.6-8.1)	7.5 (6.6-8.6)	7.0 (6.5-7.8)	.004
Total cholesterol (mg/dL)	174.0 (158.0-196.0)	175.5 (158.5-206.8)	173.0 (157.5-193.5)	.243
HDL cholesterol (mg/dL)	44.0 (38.0-54.0)	41.0 (35.0-46.5)	46.0 (40.0-55.5)	<.001
Non-HDL cholesterol (mg/dL)	115.6 (91.8-134.0)	121.5 (95.2-147.0)	112.0 (89.9-129.0)	.055
Triglycerides (mg/dL)	118.0 (86.0-177.0)	169.5 (117.3-223.8)	101.0 (82.0-134.5)	<.001
AST (U/L)	22.0 (17.0-28.0)	23.0 (19.0-26.8)	22.0 (17.0-28.0)	.513
ALT (U/L)	26.0 (20.0-36.0)	26.0 (21.0-38.8)	26.0 (18.3-34.0)	.293
GGT (U/L)	27.0 (18.0-42.0)	30.0 (18.3-43.8)	25.0 (18.0-37.5)	.192
PLT (109/L)	212.0 (181.0-243.0)	204.0 (175.5-239.5)	218.0 (185.0-243.0)	.078
LH (IU/L)	5.1 (3.9-6.8)	5.4 (4.0-8.0)	5.0 (3.8-6.5)	.183
TT (nmol/L)	13.0 (10.3-16.5)	9.5 (7.44-11.0)	15.1 (13.2-18.6)	<.001
Right testicular volume (mL)	14.0 (12.0-18.0)	14.0 (12.0-18.0)	14.0 (12.0-18.0)	.569
Left testicular volume (mL)	14.0 (12.0-16.8)	14.0 (12.0-16.1)	14.0 (12.0-17.5)	1.000
Erectile dysfunction (%)	68.1%	77.6%	74.6%	.644
IIEF-5	15.0 (7.0-21.0)	12.0 (7.0-20.0)	17.0 (8.3-21.8)	.155
IPSS	6.0 (2.0-11.0)	6.5 (3.3-12.8)	5.0 (2.0-10.3)	.260
AMSS	31.5 (24.0-40.0)	34.5 (29.0-41.5)	29.0 (22.0-38.0)	.001
FLI	63.8 (40.0-84.7)	74.1 (61.4-93.5)	56.5 (32.1-78.2)	<.001
HSI	40.5 (37.6-44.4)	41.5 (39.2-45.9)	40.1 (36.6-43.2)	.005
DSI	0.15 (−0.66-0.78)	0.45 (−0.08-1.04)	−0.07 (−1.02-0.58)	<.001
FIB-4	1.13 (0.90-1.50)	1.14 (0.91-1.62)	1.12 (0.85-1.46)	.360
APRI	0.27 (0.20-0.34)	0.28 (0.23-0.36)	0.25 (0.19-0.31)	.078

Data are expressed as median (interquartile range) for continuous variables and as absolute number (%) for categorical variables. Statistically significant P values are in bold.

Abbreviations: ALT, alanine aminotransferase; AMSS, Aging Male Symptom Score; AST, aspartate aminotransferase; APRI, AST to Platelet Ratio Index; BMI, body mass index; DSI, Dallas Steatosis Index; FIB-4, fibrosis-4 index; FLI, Fatty Liver Index; GGT, gamma-glutamyl transferase; HbA1c, glycated hemoglobin; HSI, Hepatic Steatosis Index; IIEF-5, International Index of erectile function-5; IPSS, International Prostatic Symptoms Score; LH, luteinizing hormone; PLT, platelets; TT, total testosterone.

Seventy-four patients (74/189, 38.9%) had TT values <12 nmol/L. Of these, 83.8% had low–normal LH levels (≤9.4 UI/I, hypogonadotropic hypogonadism), and 16.2% had elevated LH levels (hypergonadotropic hypogonadism).

Compared with patients with normal TT (≥12 nmol/L), men with TT <12 nmol/L had worse metabolic parameters (BMI, waist circumference, lipid profile, glycemic control with a mean difference in HbA1c of 0.5%) and worse AMSS results. Age, LH, testicular volume, and liver enzymes were not different.

All 3 noninvasive indices of MASLD (FLI, his, and DSI) were higher in subjects with low testosterone than in subjects with normal testosterone (P < .001, P = .005, P < .001). In contrast, none of the noninvasive indices of hepatic fibrosis (FIB-4 and APRI) was different between groups.

We then divided the subjects on the basis of the quartiles of FLI, HSI, and DSI. Increasing the quartiles of FLI, his, and DSI, the level of TT decreases (P = .001) and the percentage of men with TT <12 nmol/L and <8 nmol/L significantly increases (Fig. 1 and Table 2). HbA1c was significantly different in the quartile groups, with a difference of approximately 1% between first and last quartile for all the noninvasive indices of steatosis (Table 2). Furthermore, AMSS was also different in the quartile groups of FLI and DSI (Table 2).

Figure 1.

Prevalence of TT <12 nmol/L according to FLI, HSI, and DSI quartiles (their cutoffs are expressed in Table 2). TT, total testosterone; FLI, Fatty Liver Index; HSI, Hepatic Steatosis Index; DSI, Dallas Steatosis Index.

Table 2.

Comparison of the study population according to FLI, HSI, and DSI quartiles

	FLI < 39.25	39.25 ≤ FLI < 64	64 ≤ FLI < 85	FLI≥85	P trend
Age (years)	61.0 (53.5-68)	57.0 (52.0-65.0)	56.5 (48-0-61.8)	59.0 (51.0-66.0)	.178
HbA1c (%)	6.9 (6.5-7.5)	6.9 (6.5-7.8)	7.1 (6.6-8.0)	8.1 (7.0-8.9)	<.001
LH (IU/L)	5.0 (4.1-6.8)	5.1 (3.7-6.1)	5.1 (4.2-7.3)	5.0 (3.5-6.8)	.701
TT (nmol/L)	13.4 (12.3-17.3)	14.6 (11.7-18.0)	11.8 (9.9-16.5)	10.6 (7.7-13.0)	<.001
TT <12 nmol/L (%)	16.2%	28.6%	52.4%	60%	<.001
TT <8 nmol/L (%)	2.3%	7.1%	9.5%	26.7%	.002
IIEF-5	18.0 (7.0-22.0)	18.0 (9.0-22.0)	11.0 (7.0-18.0)	14.0 (7.0-18.8)	.047
IPSS	5.0 (2.0-9.0)	6.0 (2.0-11.8)	5.0 (2.0-10.5)	8.0 (4.0-15.0)	.07
AMSS	28.5 (22.0-35.0)	30.0 (24.0-41.3)	33.0 (24.0-41.0)	34.0 (28.0-41.0)	.008

	HSI < 37.46	37.46 ≤ HSI < 40.55	40.55 ≤ HSI < 44.46	HSI≥44.46	P trend
Age (years)	60.5 (53.3-68.0)	60.5 (52.0-65.0)	58.0 (49.5-65.0)	56.0 (48.3-59.8)	.003
HbA1c (%)	6.8 (6.5-7.3)	7.0 (6.4-8.0)	7.3 (6.8-8.2)	7.6 (6.8-8.5)	<.001
LH (IU/L)	4.9 (3.7-6.4)	5.6 (4.3-7.4)	5.3 (4.0-6.8)	4.8 (3.5-6.1)	.762
TT (nmol/L)	15.1 (12.4-18.1)	13.3 (11.2-19.3)	12.0 (10.5-14.2)	12.0 (7.5-14.4)	<.001
TT <12 nmol/L (%)	21.7%	36.2%	48.9%	50%	.017
TT <8 nmol/L (%)	4.3%	6.4%	6.4%	28.2%	<.001
IIEF-5	17.5 (7.8-21.3)	11.0 (7.0-22.0)	12.0 (6.0-21.0)	16.5 (9.0-20.0)	.830
IPSS	7.0 (2.0-12.0)	5.0 (2.0-12.5)	6.0 (2.0-8.0)	7.0 (2.0-11.5)	.932
AMSS	31.0 (26.8-39.3)	28.0 (22.0-40.0)	33.0 (24.5-41.0)	32.5 (26.0-37.8)	.715

	DSI < −0.71	−0.71 ≤ DSI < 0.15	0.15 ≤ DSI < 0.79	DSI≥0.79	P trend
Age (years)	61.0 (52.5-68.0)	56.0 (50.5-64.5)	58.0 (52.3-64.5)	56.0 (48.5-61.5)	.039
HbA1c (%)	6.9 (6.5-7.3)	7.4 (6.5-8.3)	7.0 (6.6-8.0)	7.8 (6.6-8.7)	.004
LH (IU/L)	5.0 (3.8-6.8)	5.3 (3.9-7.1)	5.3 (4.0-7.1)	5.0 (3.9-6.7)	.824
TT (nmol/L)	14.4 (13.1-18.5)	12.0 (10.1-15.4)	14.0 (11.3-17.7)	11.8 (8.7-13.0)	<.001
TT <12 nmol/L (%)	12.8%	46.8%	41.3%	55.3%	<.001
TT <8 nmol/L (%)	2.1%	10.6%	10.9%	21.3%	.033
IIEF-5	18.0 (8.0-22.0)	14.0 (7.0-22.0)	11.0 (8.0-20.0)	15.5 (7.0-19.8)	.163
IPSS	5.0 (2.0-9.0)	6.0 (2.0-9.0)	5.0 (2.0-10.3)	8.0 (3.5-13.0)	.085
AMSS	28.0 (22.0-36.0)	31.5 (25.8-38.3)	32.0 (23.0-43.5)	34.0 (28.3-40.8)	.038

Data are expressed as median (interquartile range) for continuous variables and as absolute number (%) for categorical variables. Statistically significant P values are in bold.

Abbreviations: AMSS, Aging Male Symptom Score; HbA1c, glycated hemoglobin; IIEF-5, International Index of erectile function-5; IPSS, International Prostatic Symptoms Score; LH, luteinizing hormone; TT, total testosterone.

Spearman's correlation showed that TT was positively correlated with HDL cholesterol and inversely correlated with other metabolic parameters (BMI, waist circumference, glycemia, HbA1c, non-HDL cholesterol, and triglycerides) (Table 3). Moreover, a significant negative correlation with AMSS, FLI (r_s = −.360, P < .001), HSI (r_s = −0.278, P < .001), and DSI (r_s = −0.332, P < .001) was found. No correlation with FIB-4 and APRI was found. Interestingly, LH did not correlate with either the steatosis index or the fibrosis index (Table 3).

Table 3.

Correlations of total testosterone with clinical and biochemical variables

	rs	P value
Age	0.023	.755
Weight	−0.212	.004
BMI	−0.244	<.001
Waist circumference	−0.291	<.001
Glycemia	−0.167	.021
HbA1c	−0.188	.01
Total cholesterol	−0.122	.094
HDL cholesterol	0.223	.002
Non-HDL cholesterol	−0.149	.041
Triglycerides	−0.427	<.001
AST	0.003	.963
ALT	−0.073	.323
GGT	−0.051	.506
PLT	0.073	.324
LH	−0.023	.755
Right testicular volume	0.037	.623
Left testicular volume	0.074	.316
IIEF-5	0.065	.385
IPSS	−0.103	.170
AMSS	−0.167	.027
FLI	−0.360	<.001
HSI	−0.278	<.001
DSI	−0.332	<.001
FIB-4	0.002	.983
APRI	−0.062	.405

Statistically significant P values are in bold.

ALT, alanine aminotransferase; AMSS, Aging Male Symptom Score; APRI, AST to Platelet Ratio Index; AST, aspartate aminotransferase; BMI, body mass index; DSI, Dallas Steatosis Index; FIB-4, fibrosis-4 index; FLI, Fatty Liver Index; GGT, gamma-glutamyl transferase; HbA1c, glycated hemoglobin; HSI, Hepatic Steatosis Index; IIEF-5, International Index of erectile function-5; IPSS, International Prostatic Symptoms Score; LH, luteinizing hormone; PLT, platelets; TT, total testosterone.

In multiple linear regression analysis, after excluding the parameters that composed the liver indices, TT concentrations (dependent variable) remained inversely associated with FLI, HSI, and DSI (explanatory variables) independently from potential confounders (Table 4). Weight and waist circumference were also not included in this regression model, because of the risk of overadjustment due to collinearity (weight is used for the calculation of BMI; waist circumference and BMI are both measures of adiposity). Similarly, BMI and waist circumference were not considered in the multiple regression analysis because they are included in the formula for the calculation of FLI, HSI, and DSI.

Table 4.

Linear regression analysis to identify independent associates of testosterone concentrations, considering the 3 indices FLI, HSI, and DSI

Covariate	Unstandardized B	Standardized B	95% CI for standardized B	P value
FLI	−0.065	−0.313	−.099 to −.031	<.001
Age	−0.058	−0.101	−.148 to .031	.200
HbA1c	−0.415	−0.100	−1.083 to .253	.221
Total cholesterol	−0.003	−0.025	−.036 to .029	.832
Non-HDL cholesterol	−0.012	−0.095	−.041 to .018	.429
AMSS	0.007	0.014	−.074 to .089	.858
HSI	−0.242	−0.268	−.376 to −.108	<.001
Age	−0.075	−0.130	−.160 to .010	.084
HbA1c	−0.231	−0.055	−.858 to .396	.469
Total cholesterol	0.008	0.059	−.024 to .040	.615
Triglycerides	−0.015	−0.289	−.024 to–.007	.001
Non-HDL cholesterol	−0.013	−0.104	−.040 to .015	.355
AMSS	−0.017	−0.034	−.090 to .056	.641
DSI	−1.546	−0.338	−2.223 to −.859	<.001
HbA1c	−0.378	−0.093	−.980 to .224	.217
Total cholesterol	0.002	0.012	−.030 to .033	.914
Non-HDL cholesterol	−0.012	−0.094	−.039 to .016	.406
AMSS	−0.019	−0.039	−.092 to .053	.598

Statistically significant P values are in bold.

Abbreviations: AMSS, Aging Male Symptom Score; DSI, Dallas Steatosis Index; FLI, Fatty Liver Index; HbA1c, glycated hemoglobin; HSI, Hepatic Steatosis Index; HDL, high-density lipoprotein.

Finally, receiver operating characteristic curves were generated to determine the parameters that are predictive of TT <12 nmol/L (Fig. 2). FLI had a better predictive performance of TT <12 nmol/L than HSI and DSI (area under the curve = 0.718 vs area under the curve = 0.624 and area under the curve = 0.677). FLI ≥63, HSI ≥40, and DSI ≥0.05 were found to be the cutoff levels with the best sensitivity (73%, 70% and 70%, respectively) and specificity (64%, 52% and 52%, respectively).

Figure 2.

Receiver operating characteristic curves for TT <12 nmol/L. Comparison between FLI, HSI, and DSI.

Discussion

In this cross-sectional study, we found that noninvasive indices of steatosis (FLI, HSI, and DSI) were negatively and independently associated with TT plasma concentrations and prevalence of hypogonadism in men with T2DM. The prevalence of low TT increased accordingly with the quartiles of FLI, HSI, and DSI, reaching up to 50% to 60% in men with the highest values. FLI and DSI were also correlated with AMSS score, highlighting their association not only with biochemical hypogonadism (TT levels), but also with clinical symptoms of hypogonadism.

Linear regression analysis excluded any potential confounding factor related to the state of diabetes control. In addition, we identified cutoff values for these indices that are associated with low TT with good sensitivity and specificity. In contrast, no statistically significant associations were found between noninvasive indices of fibrosis (APRI and FIB-4) and TT levels.

Previous studies, using several techniques for the diagnosis of MASLD, have shown an inverse association between MASLD and testosterone levels in men. Using abdominal ultrasound, Zhang et al described that association in men with T2DM (40). Interestingly, they did not find the same correlation in women with T2DM. Polyzos et al reached the same conclusion using 2 noninvasive indices of steatosis (HSI and triglyceride to HDL C ratio) in a general male population (22). Moreover, they found no association between TT and noninvasive indices of hepatic fibrosis (APRI, FIB-4, and NAFLD fibrosis score). A cross-sectional analysis conducted in the United States showed that TT levels were inversely related to the risk of MASLD, independently of age, obesity, and lifestyle (20, 23). In line with these findings, a meta-analysis of 10 cross-sectional studies enrolling 2995 subjects confirmed the correlation, pointing out that patients with the more severe form of MASLD had lower TT levels (41). Consistent with these studies, we concluded that in men with T2DM, noninvasive indices of MASLD are significantly related to low serum TT.

Guidelines do not agree on whether patients with T2DM should be investigated for hypogonadism, and the Endocrine Society, as well as other scientific societies, does not consider T2DM as a condition at risk for hypogonadism, unless it is associated with obesity and does not recommend routine testing for testosterone deficiency (29, 30, 42). Our study not only found a high prevalence of low TT in men with T2DM, but suggests that noninvasive indices of MASLD could be used to identify the patients at higher risk for hypogonadism. In particular, a FLI ≥63 was the best predictive index of TT <12 nmol/L, with a good sensitivity (73%) and satisfying specificity (64%). On the other hand, patients with known hypogonadism and T2DM should be evaluated for MASLD, given the high risk of being affected by that condition and given the increased risk of all-cause, liver-specific, and cardiovascular mortality (8).

It is well known that visceral obesity and metabolic syndrome are risk factors for hypogonadism (43). On the other hand, evidence is emerging that sex steroid impairment could have a role in the development of insulin resistance, metabolic syndrome, and MASLD (21).

In a murine model, Livingstone et al demonstrated that 5 alpha-reductase type 1 knockout mice fed with a high-fat diet developed greater weight gain, hyperinsulinemia, and hepatic steatosis (19). Another study demonstrated that testosterone deprivation in male mice due to castration induces hyperglycemia and caused the development of symptoms seen in the metabolic syndrome (44).

Human liver expresses androgen receptor in a sex-dependent (more in men than in woman) and age-dependent (almost undetectable before puberty, it increases in postpubertal life and gradually declines during aging) manner. It has been suggested that testosterone, through binding to the androgen receptor, regulates lipid homeostasis by inhibiting the uptake of triglycerides and by facilitating lipid mobilization from visceral fat, and therefore suppressing the development of hepatic steatosis (45, 46).

A recent clinical study showed an association between testosterone treatment and a significant reduction in liver fat in men with T2DM and low testosterone concentrations, hypothesizing that testosterone treatment might be advisable in patients with hypogonadal MAFLD (47, 48). In contrast, Lee et al did not find any improvement of MASLD in hypogonadal man after 12 months of testosterone replacement therapy vs placebo (49). However, more data are needed before drawing any definitive conclusion on the impact of sex steroids on human liver and on the relationship between MASLD and hypogonadism.

Even the relationship between testosterone and hepatic fibrosis is still unclear, partly because it is a less common condition than MASLD, and due to various methods used in the studies to evaluate hepatic fibrosis. In the literature there is conflicting evidence (22, 40). Miyauchi and al found an inverse association between free testosterone and FIB-4 in men with T2DM (50). However, serum free testosterone was directly measured using an enzyme-linked immunosorbent assay kit, a method considered inaccurate. Sarkar et al showed that low calculated free testosterone was independently associated with the presence and severity of biopsy-confirmed NASH and fibrosis (26). On the other hand, Dayton et al did not find any significant difference in testosterone levels (both TT and free testosterone levels, the last 1 measured by liquid chromatography tandem mass spectroscopy) with liver histology characteristic of fibrosis (analyzed with ultrasonography-guided liver biopsy) in patients with T2DM (51). In HIV-infected male patients without T2DM liver fibrosis, the APRI score positively correlated with both TT and SHBG. SHBG levels were associated with APRI independently of age, BMI, and HIV duration (28). Interestingly, patients with higher SHBG levels also had an increase in LH production as a compensation mechanism.

In the present study, we found no association between noninvasive indices of liver fibrosis (APRI and FIB-4) and TT serum levels. However, more data are needed on this topic before drawing any conclusion since the results of our study could also be underpowered to assess any difference in hepatic fibrosis. Furthermore, our study has a limitation in that calculated free testosterone was not available because SHBG and albumin were not routinely assessed. TT was not measured using the gold standard liquid chromatography tandem mass spectrometry but with immunoassay, which has already been demonstrated to be a reliable technique with good correlation with liquid chromatography tandem mass spectrometry (52, 53).

Our study has some other limitations. We could not clarify a causal relationship between TT and MASLD due to the cross-sectional nature of the study, and we used surrogate markers of MASLD while the gold standard remains the histological evidence.

In conclusion, we found that serum TT concentrations are independently associated with indices of MASLD in men with T2DM. In contrast, no statistically significant association was found with noninvasive indices of hepatic fibrosis. Further longitudinal studies are needed to establish causal associations between androgens and MASLD in men with T2DM, as well as the effect of T treatment in these patients.

Disclosures

The authors have no relevant financial or nonfinancial interest to disclose.

Data Availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Abbreviations

 
• AMSS

  Aging Male Symptom Score

 
• AST

  aspartate aminotransferase

 
• BMI

  body mass index

 
• DSI

  Dallas Steatosis Index

 
• FLI

  Fatty Liver Index

 
• GGT

  gamma-glutamyl transferase

 
• HDL

  high-density lipoprotein

 
• HSI

  Hepatic Steatosis Index

 
• IIEF-5

  International Index of erectile function-5

 
• IPSS

  International Prostatic Symptoms Score

 
• LH

  luteinizing hormone

 
• MAFLD

  metabolic dysfunction–associated fatty liver disease

 
• MASLD

  metabolic dysfunction–associated steatotic liver disease

 
• NAFLD

  nonalcoholic fatty liver disease

 
• NASH

  nonalcoholic steatohepatitis

 
• SHGB

  sex hormone–binding globulin

 
• T2DM

  type 2 diabetes mellitus

 
• TT

  total testosterone