- Email: [email protected]
Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD
Journal Pre-proof Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD Monika Sarkar, MD, MAS, Katherine Yates, ScM, Ayako Suzuki, MD, Ph.D, Joel Lavine, MD, Ph.D, Ryan Gill, MD, Ph.D, Toni Ziegler, Ph.D, Norah Terrault, MD, MPH, Sandeep Dhindsa, MD, FACE PII: DOI: Reference:
S1542-3565(19)31393-X https://doi.org/10.1016/j.cgh.2019.11.053 YJCGH 56893
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 22 November 2019 Please cite this article as: Sarkar M, Yates K, Suzuki A, Lavine J, Gill R, Ziegler T, Terrault N, Dhindsa S, Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD, Clinical Gastroenterology and Hepatology (2020), doi: https:// doi.org/10.1016/j.cgh.2019.11.053. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 by the AGA Institute
1
Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD
Short title: Testosterone and NASH Monika Sarkar, MD, MAS1 Katherine Yates, ScM2 Ayako Suzuki, MD, Ph.D3 Joel Lavine, MD, Ph.D4 Ryan Gill, MD, Ph.D5, Toni Ziegler, Ph.D6 Norah Terrault, MD, MPH1,7 Sandeep Dhindsa, MD, FACE8
1
Division of Gastroenterology and Hepatology, University of California, San Francisco, California
2
Department of Biostats and Epidemiology, Johns Hopkins University,
3
Division of Gastroenterology and Hepatology, Duke University, Durham, North Carolina
4
Division of Pediatric Gastroenterology and Hepatology, Columbia University, New York, New York
5
University of California, San Francisco, Department of Pathology
6
National Primate Research Center, University of Wisconsin, Madison, Wisconsin
7
Division of Gastroenterology and Hepatology, University of Southern California, Los Angeles, CA
8
Division of Endocrinology and Metabolism, Saint Louis University, Saint Louis, Missouri
Grant Support: The NASH CRN is supported by the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK). MS is supported by a K23 from the NIDDK (DK111944). We also acknowledge the UCSF Liver Center (P30 DK026743) for content support.
2
Abbreviations: non-alcoholic fatty liver disease (NAFLD) non-alcoholic steatohepatitis (NASH) NIH/NIDDK NASH Clinical Research Network (NASH CRN) NAFLD Activity Score (NAS) sex hormone binding globulin (SHBG) homeostasis model assessment of insulin resistance (HOMA-IR) triglycerides (TG) high-density lipoprotein cholesterol (HDL) low-density lipoprotein cholesterol (LDL)
Correspondence: Monika Sarkar, MD, MAS 400 Parnassus Ave, San Francisco, CA 94143 (415) 353-1888 E-mail: [email protected]
Conflict of Interest: MS: Site PI for an industry sponsored NAFLD clinical trial (Zydus Pharmaceuticals). KY: nothing to disclose AS: nothing to disclose JL: Consultant for Merck, Novo Nordisk, and Pfizer, grant support from Genfit. RG: nothing to disclose TZ: nothing to disclose NT: Advisory board Intercept; institutionalization grant support Gilead.
3 SD: Bayer (consultant), site PI for a pharmaceutical multicenter trial (TRAVERSE) to evaluate cardiovascular events after testosterone therapy, sponsored by Abbvie. No authors have any personal conflicts of interest to disclose.
Author contributions: M.S. put forth the hypothesis, planned and analyzed the study, and performed initial manuscript draft. K.Y. and S.D. analyzed data and edited manuscript. T.Z. performed the measurements of testosterone in the samples. N.T. helped to plan the study provided content review and manuscript edits. J.L., A.S. and R.G. provided content review and manuscript edits.
Keywords: Testosterone, NAFLD, NASH, obese
4 INTRODUCTION: With rising prevalence of obesity and diabetes, non-alcoholic fatty liver disease (NAFLD) is now a leading cause of chronic liver disease. One-third of obese or diabetic men have subnormal free and bioavailable testosterone concentrations(1). Several studies have further shown low testosterone to be associated with imaging-confirmed NAFLD in men(2), although it is unknown whether low testosterone confers increased risk of more clinically relevant manifestations of NAFLD, including non-alcoholic steatohepatitis (NASH) and NASH fibrosis. We therefore aimed to evaluate the association of testosterone with histologic features of NAFLD among a representative cohort of men from the multicenter NASH Clinical Research Network (NASH CRN).
METHODS: NASH CRN participants were selected at random to include the full spectrum of NAFLD, from simple steatosis to NASH with varying severity of fibrosis. The NAFLD Activity Score (NAS) was also calculated (composite score from 0-8 composed of steatosis (0–3), hepatocyte ballooning (0–2) and lobular inflammation scores (0–3)). Fibrosis ranged from stages 0-4. NASH diagnosis included possible or definite NASH. Biopsies were centrally evaluated by NASH CRN pathologists. Testosterone and SHBG were measured from fasting blood samples within 6 months of biopsy. Free and bioavailable (free + albumin bound) testosterone were calculated(3). Covariates included demographics, anthropomorphic and fasting serum lipid and insulin resistance measures. The association of low free testosterone with NAFLD histology was determined by logistic and ordinal logistic regression as appropriate. Additional methodology provided in Supplementary materials.
RESULTS: Biopsy-confirmed NASH was present in 72% of men, and 18% (n=28) had advanced fibrosis (Table 1). Low free testosterone was present in 26% of men with NAFLD, including 24% of men less than 40 years old. Men with low free testosterone were more likely to have NASH versus simple steatosis (88% vs 67%, p=0.01), as well as advanced fibrosis (27% vs 14%, p=0.07). NASH prevalence increased with lower quartiles of free testosterone (in 88% of men in lowest quartile versus 68% in highest quartile (p=0.03)) (Supplemental Figure 1). Low free testosterone was associated with NASH, independent of age, waist circumference, insulin resistance, and triglycerides (Adjusted OR 3.5; 95% CI 1.01-11.3, p=0.03) and with
5 higher stage of fibrosis (Adjusted OR 2.1, 95% CI 1.0-4.1, p=0.04) (Supplementary Table 1). Free testosterone concentrations in men with and without NASH were 5.3 (3.7-8.9) and 7.2 (5.3-9.8) ng/dl, respectively (p=0.03). Free and bioavailable testosterone concentrations were highly correlated (r=0.98, p<0.001). Low bioavailable testosterone also remained associated with presence of NASH (Adjusted OR 3.7, 95%CI 1.2-12.0, p=0.03) and fibrosis severity (Adjusted OR 2.0, 95%CI 1.0-3.8, p=0.05).
DISCUSSION: Leveraging data from the NASH CRN Study Cohort we identified low free testosterone levels in a quarter of men with NAFLD, and a third of those with NASH, including nearly a quarter of men under age 40 years. Moreover, low testosterone remained independently associated with presence of NASH (versus simple steatosis), as well as more severe NASH fibrosis. We also selected a conservative cut-off to define “low testosterone” that is below the threshold where hypogonadal symptoms typically develop(4). This is of clinical relevance as NASH is not an entity known to be associated with hypogonadism in men.
Several studies have demonstrated lower testosterone levels in men with NAFLD, and/or an independent association low testosterone with imaging-confirmed NAFLD in men (2, 5). Testosterone replacement in men also improves insulin resistance, lipids, and visceral adiposity, supporting a more direct role of testosterone on metabolic risk factors for NAFLD/NASH (1, 6, 7). Low free testosterone has also been associated with noninvasive markers of fibrosis(8), although the current study is the first to evaluate this association with histologically-confirmed NAFLD. Using “gold standard” LC/MS-MS methodology, we further demonstrated progressively lower testosterone concentrations by NAFLD severity.
The current study has some limitations, including lack of control group without NAFLD, as well as high proportion of men with NASH, likely due to higher suspicion for disease prompting biopsy. As a cross sectional study, causality of low testosterone and NASH could not be evaluated. While testosterone might be protective against NASH, lower testosterone may also reflect worse metabolic health in men. Due to the relatively small number with advanced fibrosis, we lacked power to analyze advanced fibrosis as a dichotomized outcome,
6 although were able to perform a robust analysis of fibrosis severity, including adjustment for comprehensive metabolic parameters.
In conclusion, low free testosterone in men was independently associated with presence and severity of NASH. Testosterone measurements should be considered in future prospective studies of the natural history of NAFLD in men.
7
FIGURE LEGEND Supplementary Figure 1. Prevalence of NASH by Quartiles of Free Testosterone. *P < .05 as compared to Q1. P overall = .06 Q1: 0.99 – 3.97 ng/dL Q2: 3.98 – 5.72 ng/dL Q3: 5.73 – 9.11 ng/dL Q: 9.12 – 26.38 ng/dL
8 REFERENCES:
1. 2.
3. 4. 5.
6.
7.
8.
Dhindsa S, Ghanim H, Batra M, and Dandona P. Hypogonadotropic Hypogonadism in Men With Diabesity. Diabetes Care. 2018;41(7):1516-25. Jaruvongvanich V, Sanguankeo A, Riangwiwat T, and Upala S. Testosterone, Sex Hormone-Binding Globulin and Nonalcoholic Fatty Liver Disease: A Systematic Review and Meta-Analysis. Ann Hepatol. 2017;16(3):382-94. Vermeulen A, Verdonck L, and Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab. 1999;84(10):3666-72. Wu FC, Tajar A, Beynon JM, Pye SR, Silman AJ, Finn JD, et al. Identification of late-onset hypogonadism in middle-aged and elderly men. N Engl J Med. 2010;363(2):123-35. Sarkar M, VanWagner LB, Terry JG, Carr JJ, Rinella M, Schreiner PJ, et al. Sex Hormone-Binding Globulin Levels in Young Men Are Associated With Nonalcoholic Fatty Liver Disease in Midlife. The American journal of gastroenterology. 2019;114(5):758-63. Dhindsa S, Ghanim H, Batra M, Kuhadiya ND, Abuaysheh S, Sandhu S, et al. Insulin Resistance and Inflammation in Hypogonadotropic Hypogonadism and Their Reduction After Testosterone Replacement in Men With Type 2 Diabetes. Diabetes care. 2016;39(1):82-91. Kapoor D, Goodwin E, Channer KS, and Jones TH. Testosterone replacement therapy improves insulin resistance, glycaemic control, visceral adiposity and hypercholesterolaemia in hypogonadal men with type 2 diabetes. European journal of endocrinology / European Federation of Endocrine Societies. 2006;154(6):899-906. Fujihara Y, Hamanoue N, Yano H, Tanabe M, Akehi Y, Nomiyama T, et al. High sex hormone-binding globulin concentration is a risk factor for high fibrosis-4 index in middle-aged Japanese men. Endocr J. 2019.
Table 1. Cohort Characteristics Among Men with Biopsy-Confirmed NAFLD by Free Testosterone (Free T) Level. Height, weight, waist and hip measurements were taken in duplicate while standing and wearing light clothing and averaged for analyses. Waist circumference was measured midway between the iliac crest and bottom of the rib cage. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (meters) squared. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following equation: (fasting glucose [mg/dL] fasting insulin [mU/L]) /405. Dyslipidemia was defined as triglycerides > 150ng/dL, high-density lipoprotein (HDL) < 40ng/dL, low-density lipoprotein (LDL) > 130ng/dL, or total cholesterol > 200ng/dL. Obesity was defined as BMI ≥ 30 kg/m2. Low free testosterone defined as <4ng/dl.
Age (years) Race, n (%) White Asian Black Other Hispanic, n (%) Body mass index (kg/cm2) Obesity, n (%) Waist circumference (cm) Waist-to-hip ratio Total cholesterol (mg/dL) HDL (mg/dL) LDL (mg/dL) Triglycerides (mg/dL) Dyslipidemia, n (%) Type 2 Diabetes, n (%) HOMA-IR Total testosterone (ng/dL) Free testosterone (ng/dL) Bioavailable testosterone (ng/dL) Sex hormone binding globulin (nmol/L) NASH, n (%) NAFLD Activity Score (NAS) Advanced fibrosis (stages 3-4), n (%)
All men (n=159) 47 (37-55)
Low Free T (n=41) 48 (38-59)
Normal Free T (n=118) 47 (37-54)
131 (82) 14 (9) 3 (3) 11 (6) 24 (15)
33 (80) 4 (10) 2 (5) 2 (5) 5 (12)
98 (83) 10 (8) 1 (1) 9 (8) 19 (16)
33 (30-35)
34 (32-37)
32 (30-35)
0.14
113 (71)
33 (81) 114 (101-121)
80 (68) 107 (100-115)
0.12
0.98 (0.94-1.02)
0.99 (0.94-1.02)
0.98 (0.95-1.02)
0.46
184 (162-221)
183 [156-219]
187 (162-222)
0.60
38 (33-43) 115 (90-140)
38 (34-42) 119 (89-138)
38 (33-44) 114 (91-143)
0.93 0.68
166 (106-238)
143 (97-239)
170 (107-235)
0.20
134 (84)
35 (85)
99 (84)
0.93
41 (26)
18 (44)
23 (19)
0.002
4.4 [2.9-6.5]
4.7 (3.2-7.6)
3.8 (2.7-5.3)
0.45
285 [195-391]
149 (117-212)
321 (253-422)
<0.001
5.7 [4.0-9.1]
3.1 (2.5-3.7)
7.7 (5.4-9.9)
<0.001
139 [96-216]
77 (57-89)
183 (132-241)
<0.001
25 [16-38]
29 (16-46)
24 (16-36)
0.28
115 (72)
36 (88)
79 (67)
0.01
4 [3-6]
5 (4-6)
4 (3-5)
0.001
28 (18)
11 (27)
17 (14)
0.07
108 (100-118]
p value 0.21
0.30 0.55
0.07
Supplementary Methods Sex hormone measures: Our primary predictor was low free testosterone. Low bioavailable testosterone was secondarily evaluated to ensure consistency of findings in light of potential low serum albumin levels in patients with liver disease. Testosterone is largely bound to sex hormone binding globulin (SHBG, 44%) and albumin (54%), thus total testosterone does not reflect the bio-availability of testosterone at the cellular level. As insulin resistant states and obesity are known to be associated with low SHBG concentrations, free testosterone or bioavailable (non-SHBG bound) testosterone measurements are essential in obese men to assess the gonadal status.
Free and bioavailable testosterone concentrations were calculated from the concentrations of total testosterone, SHBG and albumin by the formulae of Sodergard and Vermuelen et al, which have been validated across a wide range of total testosterone, SHBG, and albumin levels. Total testosterone concentrations were measured by liquid chromatography tandem mass spectrometry (LC-MS/MS). The sensitivity of the assay (limit of quantification) was 7.625 ng/ml ng/dl for testosterone. The inter-assay coefficient of variation (CV) was determined from 9 quality controls with a CV of 9.4 at the low range, 8.05 at the medium range, and 9.35 at the high range of the standard curve. SHBG concentrations were measured by Quantikine enzyme immunoassay (EIA, R&D Systems, DSHBGOB). The sensitivity of the assay was 0.006 nmol/L. The quality control provided an intra-assay CV of 1.57% and an interassay CV of 1.9%. Normal ranges for subnormal free and bioavailable testosterone concentrations are not well defined. Laboratories are encouraged to establish their own ranges. In a large epidemiological study designed to evaluate the association of testosterone concentrations with symptoms of hypogonadism, calculated free testosterone <4.6 ng/dl was predictive of sexual symptoms. We did not record sexual symptoms in our study. In the absence of symptoms and in order to avoid borderline low testosterone levels, we conservatively chose 4 ng/dl to define subnormal free testosterone, and 95 ng/dl for low bioavailable testosterone.
Statistical Analysis: Group comparisons were performed by two-sided t tests, Mann-Whitney rank sum tests, and χ2 tests as appropriate. Pearson correlation and multiple linear regression analyses between variables were performed as indicated. Data that were not normally distributed (Kolmogorov-Smirnov test) were log-transformed to perform the parametric statistical tests. Multivariate models included covariates selected a priori based on known associations
with the study outcome and predictor. Logistic regression was used to assess presence of NASH (versus simple steatosis) as a dichotomous outcome while ordinal logistic regression was used to assess the association with severity of fibrosis, from stages 0-4. P- values < 0.05 were considered statistically significant.
Supplementary Table. Low Free Testosterone (Free T) is Associated with Presence of NASH (vs Simple Steatosis) and Fibrosis Severity Among Men with NAFLD (n=159) Fibrosis Severity* Presence of NASH Unadjusted OR Adjusted OR** Unadjusted OR Adjusted OR** (95% CI), p value (95% CI), p value (95% CI), p value (95% CI), p value 4.20 3.53 2.14 2.05 Low Free T (1.39-12.64) (1.09-11.33) (1.11-4.10) (1.03-4.05) (<4ng/dL) 0.01 0.03 0.02 0.04 1.02 1.02 1.05 1.05 Age (0.99-1.05) (0.99-1.05) (1.02-1.08) (1.02-1.08) (per year) 0.11 0.29 <0.001 <0.001 1.06 1.04 1.03 1.02 Waist circum. (1.02-1.09) (1.00-1.08) (1.01-1.06) (0.99-1.04) (per 1 cm) 0.001 0.02 0.01 0.21 1.14 1.06 1.03 1.02 HOMA-IR (1.00-1.29) (0.95-1.19) (0.99-1.07) (0.98-1.06) (per unit) 0.04 0.27 0.17 0.43 Triglycerides 1.07 1.06 1.02 1.01 (per 20 (0.99-1.15) (0.98-1.15) (0.99-1.04) (0.99-1.03) mg/dL) 0.09 0.13 0.08 0.28 *Ordinal logistic regression; **Adjusted for age, waist circumference, HOMA-IR, and triglycerides Abbreviations: Odds ratio (OR), Confidence interval (CI), Homeostasis model assessment of insulin resistance (HOMA-IR)
S1542-3565(19)31393-X https://doi.org/10.1016/j.cgh.2019.11.053 YJCGH 56893
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 22 November 2019 Please cite this article as: Sarkar M, Yates K, Suzuki A, Lavine J, Gill R, Ziegler T, Terrault N, Dhindsa S, Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD, Clinical Gastroenterology and Hepatology (2020), doi: https:// doi.org/10.1016/j.cgh.2019.11.053. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 by the AGA Institute
1
Low Testosterone Is Associated With Nonalcoholic Steatohepatitis (NASH) and Severity of NASH Fibrosis in Men With NAFLD
Short title: Testosterone and NASH Monika Sarkar, MD, MAS1 Katherine Yates, ScM2 Ayako Suzuki, MD, Ph.D3 Joel Lavine, MD, Ph.D4 Ryan Gill, MD, Ph.D5, Toni Ziegler, Ph.D6 Norah Terrault, MD, MPH1,7 Sandeep Dhindsa, MD, FACE8
1
Division of Gastroenterology and Hepatology, University of California, San Francisco, California
2
Department of Biostats and Epidemiology, Johns Hopkins University,
3
Division of Gastroenterology and Hepatology, Duke University, Durham, North Carolina
4
Division of Pediatric Gastroenterology and Hepatology, Columbia University, New York, New York
5
University of California, San Francisco, Department of Pathology
6
National Primate Research Center, University of Wisconsin, Madison, Wisconsin
7
Division of Gastroenterology and Hepatology, University of Southern California, Los Angeles, CA
8
Division of Endocrinology and Metabolism, Saint Louis University, Saint Louis, Missouri
Grant Support: The NASH CRN is supported by the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK). MS is supported by a K23 from the NIDDK (DK111944). We also acknowledge the UCSF Liver Center (P30 DK026743) for content support.
2
Abbreviations: non-alcoholic fatty liver disease (NAFLD) non-alcoholic steatohepatitis (NASH) NIH/NIDDK NASH Clinical Research Network (NASH CRN) NAFLD Activity Score (NAS) sex hormone binding globulin (SHBG) homeostasis model assessment of insulin resistance (HOMA-IR) triglycerides (TG) high-density lipoprotein cholesterol (HDL) low-density lipoprotein cholesterol (LDL)
Correspondence: Monika Sarkar, MD, MAS 400 Parnassus Ave, San Francisco, CA 94143 (415) 353-1888 E-mail: [email protected]
Conflict of Interest: MS: Site PI for an industry sponsored NAFLD clinical trial (Zydus Pharmaceuticals). KY: nothing to disclose AS: nothing to disclose JL: Consultant for Merck, Novo Nordisk, and Pfizer, grant support from Genfit. RG: nothing to disclose TZ: nothing to disclose NT: Advisory board Intercept; institutionalization grant support Gilead.
3 SD: Bayer (consultant), site PI for a pharmaceutical multicenter trial (TRAVERSE) to evaluate cardiovascular events after testosterone therapy, sponsored by Abbvie. No authors have any personal conflicts of interest to disclose.
Author contributions: M.S. put forth the hypothesis, planned and analyzed the study, and performed initial manuscript draft. K.Y. and S.D. analyzed data and edited manuscript. T.Z. performed the measurements of testosterone in the samples. N.T. helped to plan the study provided content review and manuscript edits. J.L., A.S. and R.G. provided content review and manuscript edits.
Keywords: Testosterone, NAFLD, NASH, obese
4 INTRODUCTION: With rising prevalence of obesity and diabetes, non-alcoholic fatty liver disease (NAFLD) is now a leading cause of chronic liver disease. One-third of obese or diabetic men have subnormal free and bioavailable testosterone concentrations(1). Several studies have further shown low testosterone to be associated with imaging-confirmed NAFLD in men(2), although it is unknown whether low testosterone confers increased risk of more clinically relevant manifestations of NAFLD, including non-alcoholic steatohepatitis (NASH) and NASH fibrosis. We therefore aimed to evaluate the association of testosterone with histologic features of NAFLD among a representative cohort of men from the multicenter NASH Clinical Research Network (NASH CRN).
METHODS: NASH CRN participants were selected at random to include the full spectrum of NAFLD, from simple steatosis to NASH with varying severity of fibrosis. The NAFLD Activity Score (NAS) was also calculated (composite score from 0-8 composed of steatosis (0–3), hepatocyte ballooning (0–2) and lobular inflammation scores (0–3)). Fibrosis ranged from stages 0-4. NASH diagnosis included possible or definite NASH. Biopsies were centrally evaluated by NASH CRN pathologists. Testosterone and SHBG were measured from fasting blood samples within 6 months of biopsy. Free and bioavailable (free + albumin bound) testosterone were calculated(3). Covariates included demographics, anthropomorphic and fasting serum lipid and insulin resistance measures. The association of low free testosterone with NAFLD histology was determined by logistic and ordinal logistic regression as appropriate. Additional methodology provided in Supplementary materials.
RESULTS: Biopsy-confirmed NASH was present in 72% of men, and 18% (n=28) had advanced fibrosis (Table 1). Low free testosterone was present in 26% of men with NAFLD, including 24% of men less than 40 years old. Men with low free testosterone were more likely to have NASH versus simple steatosis (88% vs 67%, p=0.01), as well as advanced fibrosis (27% vs 14%, p=0.07). NASH prevalence increased with lower quartiles of free testosterone (in 88% of men in lowest quartile versus 68% in highest quartile (p=0.03)) (Supplemental Figure 1). Low free testosterone was associated with NASH, independent of age, waist circumference, insulin resistance, and triglycerides (Adjusted OR 3.5; 95% CI 1.01-11.3, p=0.03) and with
5 higher stage of fibrosis (Adjusted OR 2.1, 95% CI 1.0-4.1, p=0.04) (Supplementary Table 1). Free testosterone concentrations in men with and without NASH were 5.3 (3.7-8.9) and 7.2 (5.3-9.8) ng/dl, respectively (p=0.03). Free and bioavailable testosterone concentrations were highly correlated (r=0.98, p<0.001). Low bioavailable testosterone also remained associated with presence of NASH (Adjusted OR 3.7, 95%CI 1.2-12.0, p=0.03) and fibrosis severity (Adjusted OR 2.0, 95%CI 1.0-3.8, p=0.05).
DISCUSSION: Leveraging data from the NASH CRN Study Cohort we identified low free testosterone levels in a quarter of men with NAFLD, and a third of those with NASH, including nearly a quarter of men under age 40 years. Moreover, low testosterone remained independently associated with presence of NASH (versus simple steatosis), as well as more severe NASH fibrosis. We also selected a conservative cut-off to define “low testosterone” that is below the threshold where hypogonadal symptoms typically develop(4). This is of clinical relevance as NASH is not an entity known to be associated with hypogonadism in men.
Several studies have demonstrated lower testosterone levels in men with NAFLD, and/or an independent association low testosterone with imaging-confirmed NAFLD in men (2, 5). Testosterone replacement in men also improves insulin resistance, lipids, and visceral adiposity, supporting a more direct role of testosterone on metabolic risk factors for NAFLD/NASH (1, 6, 7). Low free testosterone has also been associated with noninvasive markers of fibrosis(8), although the current study is the first to evaluate this association with histologically-confirmed NAFLD. Using “gold standard” LC/MS-MS methodology, we further demonstrated progressively lower testosterone concentrations by NAFLD severity.
The current study has some limitations, including lack of control group without NAFLD, as well as high proportion of men with NASH, likely due to higher suspicion for disease prompting biopsy. As a cross sectional study, causality of low testosterone and NASH could not be evaluated. While testosterone might be protective against NASH, lower testosterone may also reflect worse metabolic health in men. Due to the relatively small number with advanced fibrosis, we lacked power to analyze advanced fibrosis as a dichotomized outcome,
6 although were able to perform a robust analysis of fibrosis severity, including adjustment for comprehensive metabolic parameters.
In conclusion, low free testosterone in men was independently associated with presence and severity of NASH. Testosterone measurements should be considered in future prospective studies of the natural history of NAFLD in men.
7
FIGURE LEGEND Supplementary Figure 1. Prevalence of NASH by Quartiles of Free Testosterone. *P < .05 as compared to Q1. P overall = .06 Q1: 0.99 – 3.97 ng/dL Q2: 3.98 – 5.72 ng/dL Q3: 5.73 – 9.11 ng/dL Q: 9.12 – 26.38 ng/dL
8 REFERENCES:
1. 2.
3. 4. 5.
6.
7.
8.
Dhindsa S, Ghanim H, Batra M, and Dandona P. Hypogonadotropic Hypogonadism in Men With Diabesity. Diabetes Care. 2018;41(7):1516-25. Jaruvongvanich V, Sanguankeo A, Riangwiwat T, and Upala S. Testosterone, Sex Hormone-Binding Globulin and Nonalcoholic Fatty Liver Disease: A Systematic Review and Meta-Analysis. Ann Hepatol. 2017;16(3):382-94. Vermeulen A, Verdonck L, and Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab. 1999;84(10):3666-72. Wu FC, Tajar A, Beynon JM, Pye SR, Silman AJ, Finn JD, et al. Identification of late-onset hypogonadism in middle-aged and elderly men. N Engl J Med. 2010;363(2):123-35. Sarkar M, VanWagner LB, Terry JG, Carr JJ, Rinella M, Schreiner PJ, et al. Sex Hormone-Binding Globulin Levels in Young Men Are Associated With Nonalcoholic Fatty Liver Disease in Midlife. The American journal of gastroenterology. 2019;114(5):758-63. Dhindsa S, Ghanim H, Batra M, Kuhadiya ND, Abuaysheh S, Sandhu S, et al. Insulin Resistance and Inflammation in Hypogonadotropic Hypogonadism and Their Reduction After Testosterone Replacement in Men With Type 2 Diabetes. Diabetes care. 2016;39(1):82-91. Kapoor D, Goodwin E, Channer KS, and Jones TH. Testosterone replacement therapy improves insulin resistance, glycaemic control, visceral adiposity and hypercholesterolaemia in hypogonadal men with type 2 diabetes. European journal of endocrinology / European Federation of Endocrine Societies. 2006;154(6):899-906. Fujihara Y, Hamanoue N, Yano H, Tanabe M, Akehi Y, Nomiyama T, et al. High sex hormone-binding globulin concentration is a risk factor for high fibrosis-4 index in middle-aged Japanese men. Endocr J. 2019.
Table 1. Cohort Characteristics Among Men with Biopsy-Confirmed NAFLD by Free Testosterone (Free T) Level. Height, weight, waist and hip measurements were taken in duplicate while standing and wearing light clothing and averaged for analyses. Waist circumference was measured midway between the iliac crest and bottom of the rib cage. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (meters) squared. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following equation: (fasting glucose [mg/dL] fasting insulin [mU/L]) /405. Dyslipidemia was defined as triglycerides > 150ng/dL, high-density lipoprotein (HDL) < 40ng/dL, low-density lipoprotein (LDL) > 130ng/dL, or total cholesterol > 200ng/dL. Obesity was defined as BMI ≥ 30 kg/m2. Low free testosterone defined as <4ng/dl.
Age (years) Race, n (%) White Asian Black Other Hispanic, n (%) Body mass index (kg/cm2) Obesity, n (%) Waist circumference (cm) Waist-to-hip ratio Total cholesterol (mg/dL) HDL (mg/dL) LDL (mg/dL) Triglycerides (mg/dL) Dyslipidemia, n (%) Type 2 Diabetes, n (%) HOMA-IR Total testosterone (ng/dL) Free testosterone (ng/dL) Bioavailable testosterone (ng/dL) Sex hormone binding globulin (nmol/L) NASH, n (%) NAFLD Activity Score (NAS) Advanced fibrosis (stages 3-4), n (%)
All men (n=159) 47 (37-55)
Low Free T (n=41) 48 (38-59)
Normal Free T (n=118) 47 (37-54)
131 (82) 14 (9) 3 (3) 11 (6) 24 (15)
33 (80) 4 (10) 2 (5) 2 (5) 5 (12)
98 (83) 10 (8) 1 (1) 9 (8) 19 (16)
33 (30-35)
34 (32-37)
32 (30-35)
0.14
113 (71)
33 (81) 114 (101-121)
80 (68) 107 (100-115)
0.12
0.98 (0.94-1.02)
0.99 (0.94-1.02)
0.98 (0.95-1.02)
0.46
184 (162-221)
183 [156-219]
187 (162-222)
0.60
38 (33-43) 115 (90-140)
38 (34-42) 119 (89-138)
38 (33-44) 114 (91-143)
0.93 0.68
166 (106-238)
143 (97-239)
170 (107-235)
0.20
134 (84)
35 (85)
99 (84)
0.93
41 (26)
18 (44)
23 (19)
0.002
4.4 [2.9-6.5]
4.7 (3.2-7.6)
3.8 (2.7-5.3)
0.45
285 [195-391]
149 (117-212)
321 (253-422)
<0.001
5.7 [4.0-9.1]
3.1 (2.5-3.7)
7.7 (5.4-9.9)
<0.001
139 [96-216]
77 (57-89)
183 (132-241)
<0.001
25 [16-38]
29 (16-46)
24 (16-36)
0.28
115 (72)
36 (88)
79 (67)
0.01
4 [3-6]
5 (4-6)
4 (3-5)
0.001
28 (18)
11 (27)
17 (14)
0.07
108 (100-118]
p value 0.21
0.30 0.55
0.07
Supplementary Methods Sex hormone measures: Our primary predictor was low free testosterone. Low bioavailable testosterone was secondarily evaluated to ensure consistency of findings in light of potential low serum albumin levels in patients with liver disease. Testosterone is largely bound to sex hormone binding globulin (SHBG, 44%) and albumin (54%), thus total testosterone does not reflect the bio-availability of testosterone at the cellular level. As insulin resistant states and obesity are known to be associated with low SHBG concentrations, free testosterone or bioavailable (non-SHBG bound) testosterone measurements are essential in obese men to assess the gonadal status.
Free and bioavailable testosterone concentrations were calculated from the concentrations of total testosterone, SHBG and albumin by the formulae of Sodergard and Vermuelen et al, which have been validated across a wide range of total testosterone, SHBG, and albumin levels. Total testosterone concentrations were measured by liquid chromatography tandem mass spectrometry (LC-MS/MS). The sensitivity of the assay (limit of quantification) was 7.625 ng/ml ng/dl for testosterone. The inter-assay coefficient of variation (CV) was determined from 9 quality controls with a CV of 9.4 at the low range, 8.05 at the medium range, and 9.35 at the high range of the standard curve. SHBG concentrations were measured by Quantikine enzyme immunoassay (EIA, R&D Systems, DSHBGOB). The sensitivity of the assay was 0.006 nmol/L. The quality control provided an intra-assay CV of 1.57% and an interassay CV of 1.9%. Normal ranges for subnormal free and bioavailable testosterone concentrations are not well defined. Laboratories are encouraged to establish their own ranges. In a large epidemiological study designed to evaluate the association of testosterone concentrations with symptoms of hypogonadism, calculated free testosterone <4.6 ng/dl was predictive of sexual symptoms. We did not record sexual symptoms in our study. In the absence of symptoms and in order to avoid borderline low testosterone levels, we conservatively chose 4 ng/dl to define subnormal free testosterone, and 95 ng/dl for low bioavailable testosterone.
Statistical Analysis: Group comparisons were performed by two-sided t tests, Mann-Whitney rank sum tests, and χ2 tests as appropriate. Pearson correlation and multiple linear regression analyses between variables were performed as indicated. Data that were not normally distributed (Kolmogorov-Smirnov test) were log-transformed to perform the parametric statistical tests. Multivariate models included covariates selected a priori based on known associations
with the study outcome and predictor. Logistic regression was used to assess presence of NASH (versus simple steatosis) as a dichotomous outcome while ordinal logistic regression was used to assess the association with severity of fibrosis, from stages 0-4. P- values < 0.05 were considered statistically significant.
Supplementary Table. Low Free Testosterone (Free T) is Associated with Presence of NASH (vs Simple Steatosis) and Fibrosis Severity Among Men with NAFLD (n=159) Fibrosis Severity* Presence of NASH Unadjusted OR Adjusted OR** Unadjusted OR Adjusted OR** (95% CI), p value (95% CI), p value (95% CI), p value (95% CI), p value 4.20 3.53 2.14 2.05 Low Free T (1.39-12.64) (1.09-11.33) (1.11-4.10) (1.03-4.05) (<4ng/dL) 0.01 0.03 0.02 0.04 1.02 1.02 1.05 1.05 Age (0.99-1.05) (0.99-1.05) (1.02-1.08) (1.02-1.08) (per year) 0.11 0.29 <0.001 <0.001 1.06 1.04 1.03 1.02 Waist circum. (1.02-1.09) (1.00-1.08) (1.01-1.06) (0.99-1.04) (per 1 cm) 0.001 0.02 0.01 0.21 1.14 1.06 1.03 1.02 HOMA-IR (1.00-1.29) (0.95-1.19) (0.99-1.07) (0.98-1.06) (per unit) 0.04 0.27 0.17 0.43 Triglycerides 1.07 1.06 1.02 1.01 (per 20 (0.99-1.15) (0.98-1.15) (0.99-1.04) (0.99-1.03) mg/dL) 0.09 0.13 0.08 0.28 *Ordinal logistic regression; **Adjusted for age, waist circumference, HOMA-IR, and triglycerides Abbreviations: Odds ratio (OR), Confidence interval (CI), Homeostasis model assessment of insulin resistance (HOMA-IR)