Elevated Testosterone Levels in HIV-Infected Men

Abstract

Little is known about HIV-infected patients with serum testosterone levels in high normal to elevated ranges. An HIV-infected patient with hepatitis C and unexplained high serum testosterone levels prompted a retrospective chart review into the association of hepatitis C and serum testosterone levels greater than 1000 ng/mL in our clinic. The charts of 1419 male HIV patients were reviewed. Out of 1419 patients, 159 (11%) met the criteria for data analysis. A total of 8 patients had serum testosterone levels greater than 1000 ng/mL. There was no significant correlation between hepatitis C antibody positivity or presence of hepatitis C viremia as measured by viral load, nor was there any significant correlation with CD4+ cell counts. We found a weak positive association between years since reported diagnosis of HIV and high testosterone levels.

Introduction

Age-related conditions of bone mineral density loss, fatigue, and muscle loss are commonly addressed problems in HIV care, as people are living longer with HIV. ^{1 –3} Since the prevalence of androgen deficiency increases with age and as HIV-infected patients are living longer, testosterone screening for HIV-infected men has clinical implications in maintaining health. ⁴ Testosterone screening should occur in all HIV-infected men with symptoms of hypogonadism as well as in those with comorbidities such as hepatitis B and C and injection drug use. ^5,6

Given the long-standing association between HIV-infected men and low testosterone levels in the medical literature, an unusual observation was seen in an HIV-infected man with hepatitis C coinfection, showing markedly elevated serum total testosterone levels, at the Temple University Comprehensive HIV Practice during routine screening for serum testosterone. Because others have reported a markedly increased risk for the development of hepatocellular carcinoma in male patients infected with hepatitis C having serum testosterone levels more than 569 ng/mL, we set out to explore the relationship between hepatitis C infection and HIV-infected men with elevated serum testosterone levels greater than 1000 ng/mL. ⁷

Case Report

A 45-year-old African American male with a history of HIV/AIDS, chronic hepatitis C, and Kaposi sarcoma (KS) had been diagnosed with AIDS 10 years earlier and was adherent with his current regimen consisting of raltegravir and emtricitabine/tenofovir (TDF). He consistently had nondetectable HIV RNA levels and his immune status was stable with CD 4 + cell counts typically greater than 300 cells/mm³.

He had no complaints of erectile dysfunction nor did he complain of fatigue. The patient denied any use of topical testosterone or any type of topical or injectable hormonal supplements, and this was confirmed with the patient’s pharmacy. In addition to antiretroviral therapy (ART), the patient was taking loratadine 10 mg as needed for seasonal allergies, omeprazole 20 mg daily, diphenhydramine 25 mg as needed for insomnia, and ammonium lactate lotion 12% for dry skin.

The patient reported adherence with ART, denied use of recreational drugs or alcohol, and smoked 10 cigarettes per day. Physical examination revealed a well-developed male weighing 68 kg, some excoriated skin around the mid lower back, and darkened and discolored, firm, edematous lower extremities but was otherwise normal.

All laboratory assessments were performed though Quest diagnostics (900 Business Center Dr, Horsham, Philadelphia). Laboratory assessment revealed a total testosterone of 2398 ng/dL (normal 241-827 ng/dL), free testosterone was 18 pg/mL (normal 6-73 pg/mL), prolactin was 5.4 ng/mL (normal 2.0-18.0 ng/mL), triiodothyronine was 249 ng/dL (76-181 ng/dL), total free thyroxine was 0.9 ng/dL (normal 0.9-1.6 ng/dL), thyroid-stimulating hormone was 0.95 mIU/L (normal <1.22 IU/L), luteinizing hormone was 11.1 (normal for age 1.5-34.6 mIU/mL), and follicle-stimulating hormone was 15.3 mIU/mL (normal 1.6-8.0 mIU/mL).

Sex hormone–binding globulin (SHBG) was 200 nmol/L (normal male 7-65 nmol/L), AM cortisol was 14.5 µg/dL (normal 4.0-22 µg/dI), and the adrenal androgen dehydroepiandiosterone sulfate was 17 µg/dL (normal 25-510 µg/dL). The elevated total testosterone was attributed to a physiologic response to maintain a normal free testosterone level in the face of elevated SHBG from hepatitis C infection.

We present our single-center experience with a retrospective chart review investigating the relationship of elevated testosterone levels ≥1000 ng/mL to hepatitis C antibody positivity in HIV-infected men not taking testosterone supplementation.

Methods

This study was a single-center retrospective chart review of HIV-infected men seen in the Temple Comprehensive HIV Program. Patients included in the study were HIV-infected men and 18 years of age and older with at least 1 recorded serum testosterone level from January 1, 2006, to December 1, 2011. HIV-infected men in the study had to have a documented confirmed laboratory diagnosis of HIV infection by positive enzyme-linked immunosorbent assay HIV antibody test confirmed by Western blot, p24 antigen assay, or quantitative HIV-1 RNA assay. Patients excluded from the study included any males with history of intramuscular or topical testosterone use.

Study Design

Data were obtained from Temple University Hospital’s electronic medical record system (EPIC), and the study was approved by Temple’s Institutional Review Board. Two reviewers (V.P. and E.N.C.) obtained data from the medical charts using a standardized collection form and process. We recorded total testosterone levels, CD 4 + cell counts, presence of hepatitis B surface antigen positivity, presence of hepatitis C virus (HCV) antibody positivity, hepatitis C viral loads, history of KS, reported drug use (past or present), protease inhibitor (PI) use (past or present), oral ketoconazole use for more than 6 months, chronic narcotic use greater than the equivalent of oxycodone 150 mg monthly, alcohol use of more than 7 ounces/week, age, presence of diabetes mellitus, body mass index (BMI) greater than 30, duration of HIV infection, race, and mode of HIV transmission.

Spearman rank correlation was used to measure the relationship between testosterone levels and these clinical variables.

Results

For the period from February 1, 2012, to April 15, 2012, the charts of 1419 male HIV patients were reviewed using EPIC. Of the 1419 patients, 159 (11%) met the criteria for data analysis. A total of 8 patients had serum testosterone levels greater than 1000 ng/mL. We also found a weak positive association between years since reported diagnosis of HIV and high testosterone levels.

Demographic Data

Of the 159 patients, 87 (54.7%) were African Americans, 27 (16.9%) were Hispanic, 8 (5.0%) were caucasian, and 37 (23.2%) were unknown. A total of 54 (33.9%) patients were noted to have HCV antibody positivity. Of the 54 patients who were hepatitis C positive, 40 (74.1%) had at least 1 documented hepatitis C viral load. Of note, 79 (49.6%) of all the patients had a history of drug use and 77 (48.4%) had a history of taking a PI in the past or present. The mean duration of years since diagnosis of HIV infection among our patients in the study was reported to be 11.1 years. The mean age was 45.4 years.

There was no significant correlation between hepatitis C antibody positivity (Spearman correlation .03, P = .319, N = 691) or presence of hepatitis C viremia as measured by the viral load (Spearman correlation .16, P = .200, N = 61) in HIV-infected men and testosterone levels. There was also no significant correlation with CD 4 + cell counts (Spearman correlation .06, P = .07, N = 759). Potential confounding risk factors such as ketoconazole use (Spearman correlation −.144, P < .0001, N = 759), PI use (Spearman correlation .032, P = .381, N=756), and BMI >30 (Spearman Correlation 0.012, P = .747, N = 722) were not associated with total testosterone levels ≥1000 ng/mL.

Discussion

We found no significant association between HIV coinfection with hepatitis C and having a total testosterone level ≥1000 ng/mL. Our study results did demonstrate a weak positive association between years since reported diagnosis of HIV and high testosterone levels (Spearman correlation .237, P < .0001, N = 762) which may reflect only a retention in care effect with subsequent use of ART.

In HIV-infected patients, both serum levels of SHBG (the primary binding protein of circulating testosterone) and its affinity for testosterone are increased, independent of CD4 + cell counts. ^8,9 In addition, in patients with chronic hepatitis and liver cirrhosis, the percentage of bound testosterone can become significantly increased. ¹⁰ The level of liver fibrosis in patients with chronic HCV correlates with an increase in SHBG and a decrease in free testosterone levels, further driving the hypothalamic–pituitary–gonadal axis toward homeostasis, resulting in elevated total testosterone levels. ¹¹ We postulate that this mechanism explains the serum testosterone levels measured in our patient at greater than 1000 ng/mL.

Confounding the picture can be the use of common concomitant medications such as PIs, chronic narcotics, and ketoconazole, all of which can reduce serum testosterone levels. As expected, none of these medications were linked to serum testosterone levels greater than 1000 ng/mL in our study.

A study in the highly active antiretroviral therapy (HAART) era has reported that 6.8% of a cohort of HIV-infected men had serum testosterone in the elevated range, which agrees with our finding of 5%. ¹² There has been 1 report of a link between HIV-infected men with KS and high testosterone levels in the pre-HAART era, which found that serum testosterone levels in HIV-infected men with KS and CD4 + cell counts greater than 200 cells/mm³ like our patient had a 60% increase compared to that of HIV-negative controls; however, few explanations were postulated. ¹³ Finally, others found that serum testosterone levels greater than 569 ng/mL were associated with a 4-fold increase in the risk of hepatocellular carcinoma in 1 cohort, which, unfortunately, was not tested for the presence of HIV infection. Independently, the risk of heptatocellular carcinoma held a 37-time relative risk in patients with HCV over no infection with hepatitis C. ⁷

The major limitation to this study was using single random total testosterone levels for the analysis. Two morning measurements of serum total testosterone are usually sufficient to diagnose testosterone deficiency. Furthermore, the test for total testosterone can be unreliable in both the low and the high ranges of detection. ¹⁴ Total testosterone levels can be unreliable if binding proteins are altered; however, free testosterone levels are not routinely ordered for screening of HIV-infected men for hypogonadism. In addition, commercially available free testosterone levels can give misleading results. If abnormalities of SHBG are suspected, free testosterone, bioavailable testosterone levels (free plus albumin—bound), or calculated free testosterone levels should be considered. Total testosterone is commercially available usually as automated immunoassay platforms, and considerable variability exists among the various methods. The gold standard for free testosterone measurements remains centrifugal ultrafiltration and equilibrium dialysis, which are not available outside the research laboratories. ¹⁵ Calculated free testosterone levels using total testosterone and SHBG are most commonly used in clinical practice; the estimation of free testosterone concentration using SHBG and total measured testosterone remains the best approach for the estimation of free testosterone levels. ^16,17 Calculated free testosterone levels correlate well with those obtained by equilibrium dialysis. ¹⁸ Web-based calculators are available to the clinicians to calculate free testosterone concentrations using available values for free testosterone, SHBG, and albumin. ¹⁹

In conclusion, our findings did not show any association between total testosterone levels ≥1000 ng/mL and hepatitis C infection in HIV-infected men. However, we did find that 5% of men screened for total testosterone in our HIV program had serum testosterone levels greater than 1000 ng/mL. Given the progress that we have made over the past 3 decades in HIV treatment, further studies are warranted to explore the relationship of sex hormones in HIV-infected men in a larger sample size using morning testing with total testosterone levels and calculated free testosterone levels. Although its clinical significance is as yet unknown, associations between elevated total testosterone levels and risk of hepatocellular carcinoma in HIV-infected patients with hepatitis C coinfection should also be further investigated.