Among men participating in the MACS CVD substudy for whom an AM blood sample was available for T assay, the overall prevalence of biochemical hypogonadism was 8.7% (9.3% among HIV-infected men and 7.2% among HIV-uninfected men). Prevalence of hypogonadism, when men on TRT (n = 71) were included in the group of hypogonadal men, was significantly higher in HIV-infected compared with HIV-uninfected men (24.0% v. 7.8%). In the men not on TRT, the diagnosis of hypogonadism would have been missed among one-third of HIV infected men if TT only had been measured, while no HIV-uninfected men would have gone undiagnosed. Our results underscore the importance of using free T for hypogonadism diagnosis among HIV-infected men.
Hypogonadism prevalence among HIV-infected men has been reported as high as 70% [1–3]. In our cohort, when including men on TRT, the prevalence of hypogonadism in the HIV-infected men was 24.0%. Prevalence varies widely depending on the hypogonadism definition used and the demographic and clinical characteristics of the cohort, which likely accounts for the differences observed between our findings and other groups. Of note, hypogonadism among HIV-infected men has remained common despite successful ART . Advancing age, higher body mass index (BMI)  and HCV co-infection  are associated with hypogonadism among HIV-infected men, and contribute to the persistence of hypogonadism despite successful ART.
We performed exploratory analyses among the HIV-infected hypogonadal men, comparing those who had a normal TT but low FT with men who had a low TT. Our objective was to evaluate whether we could provide guidance to clinicians as to which HIV-infected men would require FT testing; we found that men co-infected with HCV were more likely to have normal TT but low FT levels. Hence, HIV/HCV-coinfected men may represent a population at particular risk for misdiagnosis if TT alone is used to ascertain the presence of hypogonadism.
Consistent with findings from prior studies [2, 7, 8] we found that SHBG levels were higher among HIV-infected men. The exact pathophysiologic mechanisms accounting for this finding are unclear. Based on analyses of this same cohort, we have previously reported that SHBG increases with increasing age, black race, and HCV infection and decreases with increasing BMI . Elevations in TT levels that occur as a result of elevations in SHBG decrease the diagnostic utility of the TT assay among HIV-infected men.
There are several limitations to this study. A diagnostic evaluation of hypogonadism is usually initiated because a patient reports symptoms such as low libido, fatigue, and low energy. Hypogonadism is a clinical diagnosis, and therefore the significance of biochemical hypogonadism diagnosed by laboratory assays is unclear. Commercially available testosterone assays vary in quality, and our assays, performed at a preeminent lab, may not have generated results that are representative of those readily available to clinicians using commercial assays. It is unclear how well FT calculated by Vermeulen equation reflects equilibrium dialysis, which is the gold standard. In our analysis, men with samples drawn after 12 noon were excluded, with an unknown effect on the outcome. Men receiving T therapy, most of whom were HIV-infected, were excluded. We presume that these men were receiving T therapy for diagnosed hypogonadism. By excluding these men, we may have introduced a selection bias because they may have been diagnosed using a low TT assay alone, possibly resulting in enrichment of numbers of included men with normal T and low FT. Even in this scenario, however, the number of HIV-infected men whose hypogonadism would have gone undetected by using TT measurement alone would still have been 11% (rather than the 32.3% we found) if all of the men who were receiving TRT had both low T and low free T levels. This represents a rate of hypogonadism still significantly higher than observed among the HIV-uninfected men. Finally, luteinizing hormone/follicle stimulating hormone (LH/FSH) levels were not performed; therefore, we were unable to determine if hypogonadism was primary or secondary.
The use of TT levels assayed from morning specimens to diagnose hypogonadism among HIV-infected men may result in about 30% missed cases. Morning free T levels are more sensitive diagnostically and should be measured in all HIV-infected men in whom hypogonadism is suspected.