A total of 351 determinations of sexual hormones were carried out in 189 HIV-infected men in stable clinical condition. Highly active antiretroviral therapy (HAART) was associated with increased levels of both testosterone and 17β-estradiol, but not with luteinizing hormone (LH) and follicle-stimulating hormone (FSH). Protease inhibitors were more associated with testosterone, and non-nucleoside reverse transcriptase inhibitors with 17β-estradiol. The values of both hormones, but not those of LH and FSH, increased with respect to pre-treatment levels in those patients who initiated HAART.
Sexual dysfunction and hypogonadism have long been recognized in HIV-infected men, mainly involving patients with weight loss, fewer than 200 × 106 CD4 cells/l and opportunistic infections [1-4]. However, most studies on this topic were published before the era of highly active antiretroviral therapy (HAART), and the possible influence of antiretroviral treatment in these aspects is unknown. Furthermore, such studies included patients in advanced stages and, in some of them, the number of patients was relatively low.
We measured the serum sexual hormones in a numerous and well-characterized group of HIV-infected patients in good clinical condition, in particular to evaluate the influence that antiretroviral therapy could have on these hormones.
The study group included 189 HIV-infected men who underwent a total of 351 determinations of sexual hormones of the hypophyseal-gonadal axis. The patients' mean age was 36.8 years. The number of determinations per patient ranged from one to four, each separated by a mean interval of 27.4 weeks. Other clinical and laboratory data were also recorded at the time of each analytical determination. The patients were in stable clinical condition and did not report any severe, acute illness at the time of sampling.
The Spearman's coefficient correlation, Mann-Whitney U test, Kruskal-Wallis test, Wilcoxon matched-pairs signed-ranks test, and Friedman's test were used as appropriate for statistical analysis. Stepwise multiple linear regression was used to identify the independent association of diverse parameters with testosterone. A P value less than 0.05 for a two-sided test was considered statistically significant.
The mean CD4 cell counts for all determinations were 451.1 × 106cells/l, and 64.1% of the patients had an undetectable viral load. The mean testosterone level was 28.1 nmol/l, 17β-estradiol 188.7 pmol/l, follicle-stimulating hormone (FSH) 5.3 U/l and luteinizing hormone (LH) 6.2 U/l. We found 6.8% of all determinations of testosterone to be above the upper and 4.0% below the lower normal limit. For 17β-estradiol these values were 45.1 and 2.9%, for FSH 3.2 and 0.6%, and for LH 13.3 and 5.6%, respectively.
Testosterone correlated positively with FSH (r = 0.25, P < 0.0001) and LH (r = 0.31, P < 0.0001), and 17β-estradiol with FSH (r = 0.16, P = 0.003) and LH (r = 0.21, P < 0.0001). Neither testosterone nor 17β-estradiol showed any significant correlation with age, time since the onset of HAART, or CD4 cell counts. Patients with low levels of testosterone also had lower levels of both FSH (P = 0.04) and LH (P = 0.02) than the remaining patients.
All but three patients were grouped into five therapeutic groups according to their antiretroviral regimens at the time of sampling: no treatment (56 determinations), nucleoside reverse transcriptase inhibitors (NRTI) only (12 determinations), NRTI plus protease inhibitors (PI) (138 determinations), nucleoside plus non-nucleoside reverse transcriptase inhibitors (NNRTI) (108 determinations), and NRTI plus NNRTI plus PI (34 determinations). There were significant differences in the values of testosterone (P = 0.0006) and 17β-estradiol (P < 0.0001), but not in those of FSH (P = 0.39) and LH (P = 0.29), when the five therapeutic groups were considered altogether. Fig. 1 depicts the values of testosterone and 17β-estradiol in each therapeutic group. When the three antiretroviral drug classes underwent a multiple regression analysis, PI (P = 0.0006) and NNRTI (P = 0.044) were predictive of testosterone levels, and NNRTI (P < 0.0001) and PI (P = 0.02) of 17β-estradiol levels. In the subgroup of NNRTI-treated patients, those receiving nevirapine had higher levels of testosterone and 17β-estradiol than those treated with efavirenz (mean 30.9 versus 24.6 nmol/l, P = 0.002, and 250.7 versus 188.7 pmol/l, P = 0.0001, respectively).
The sequential determinations of testosterone and 17β-estradiol in the same patients did not reveal any significant difference. However, in the 15 patients for whom hormonal determinations were available before and after the onset of HAART, there was an increase in the mean levels of testosterone (from 23.2 to 37.6 nmol/l, P = 0.02) and of 17β-estradiol (from 139.5 to 189.1 pmol/l, P = 0.1) after the initiation of treatment. FSH experienced a non-significant decrease (P = 0.07) and LH a minor increase (P = 0.27) in these patients.
A multiple regression analysis that included all variables analysed revealed that 17β-estradiol (P < 0.0001), treatment with PI (P = 0.0005), LH (P = 0.001) and younger age (P = 0.02) were positive predictors of testosterone levels, and testosterone (P < 0.0001), treatment with NNRTI (P = 0.0004), and older age (P = 0.02) were significant predictors of higher 17β-estradiol levels.
In our series there was a relatively low prevalence of hypogonadism, as determined by low testosterone levels, in comparison with other studies that reported figures of 30-40% of HIV-infected men, mainly in those with fewer than 200 CD4 cells/μl, weight loss and opportunistic infections [3-10]. This apparent discrepancy may be explained by the good status performance of our patients compared with those of other studies. In addition, the influence of antiretroviral drugs on the testosterone levels observed in our study would have contributed appreciably to the relatively low prevalence of hypogonadism.
On the other hand, we found a positive correlation between testicular and hypophyseal sexual hormones, and patients with low levels of testosterone also had lower levels of gonadotrophins. These findings suggest that the mechanism of hypogonadism, as measured by low levels of testosterone, is not testicular. In fact, almost half of our hypogonadal patients had subnormal LH levels, and only 15% of patients with low testosterone had appropriately increased LH values. In this regard, previous reports have also implicated a hypophyseal deficit as the cause or as a contributing factor in the genesis of hypogonadism in HIV-infected patients [2,10-12].
Antiretroviral treatment seemed to influence the serum levels of both testosterone and 17β-estradiol. In fact, patients who initiated HAART experienced an increase in both hormones compared with their own pre-HAART values. Although both PI and NNRTI were associated with increased levels of both hormones, PI were more related to testosterone, and NNRTI to 17β-estradiol. To our knowledge, no study has analysed this aspect, and these associations have not been reported to date. Some degree of inhibition of testosterone metabolism by PI has been described [13], and given the numerous interactions of these drugs with the cytochrome P450 system, the inhibition of the catabolic pathways of these hormones in the liver could probably account, at least partly, for their increase in these patients.
Julio Collazos
Eduardo Martinez
José Mayo
Sofia Ibarra
Acknowledgements
The authors are indebted to Dr Margarita Esteban and Dr Carmen Mar for their collaboration with the hormonal determinations.
References
1. Newshan G, Taylor B, Gold R. Sexual functioning in ambulatory men with HIV/AIDS. Int J STD AIDS 1998, 9: 672-676.
2. Schambelan M, Grunfeld C. Endocrine abnormalities associated with HIV infection and AIDS. In:Textbook of AIDS medicine. Broder S, Merigan TC, Bolognesi D (editors). Baltimore: Williams and Wilkins; 1994. pp. 629-636.
3. Sellmeyer DE, Grunfeld C. Endocrine and metabolic disturbances in human immunodeficiency virus infection and the acquired immune deficiency syndrome. Endocr Rev 1996, 17: 518-552.
4. Dobs AS, Dempsey MA, Landeson PW, Polk BF. Endocrine disorders in men infected with human immunodeficiency virus. Am J Med 1988, 84: 611-615.
5. Kopicko JJ, Momodu I, Adedokum A, Hoffman M, Clark RA, Kissinger P. Characteristics of HIV-infected men with low serum testosterone levels. Int J STD AIDS 1999, 10: 817-820.
6. Ferrando SJ, Rabkin JG, Poretsky L. Dehydroepiandrosterone sulfate (DHEAS) and testosterone: relation to HIV illness stage and progression over one year. J Acquired Immune Defic Syndr 1999, 22: 146-154.
7. Wagner G, Rabkin JG, Rabkin R. Illness stage, concurrent medications and other correlates of low testosterone in men with HIV illness. J Acquired Immune Defic Syndr Hum Retrovirol 1995, 8: 204-207.
8. Coodley GO, Loveless MO, Nelson HD, Coodley MK. Endocrine function in HIV wasting syndrome. J Acquired Immune Defic Syndr 1994, 7: 46-51.
9. Bhasin S, Storer TW, Javanbakht M. et al. Testosterone replacement and resistance exercise in HIV-infected men with weight loss and low testosterone levels. JAMA 2000, 283: 763-770.
10. Raffi F, Brisseau JM, Planchon B, Remi JP, Barrier JH, Grolleau JY. Endocrine function in 98 HIV-infected patients: a prospective study. AIDS 1991, 5: 729-733.
11. Croxon TC, Chapman WE, Miller LK, Levitt CD, Senie R, Zumoff B. Changes in the hypothalamic pituitary-gonadal axis in human immunodeficiency virus-infected hypogonadal men. J Clin Endocrinol Metab 1989, 68: 317-321.
12. Salehian B, Jacobson D, Grafe M, McCutchan A, Swerdloff R. Pituitary-testicular axis during HIV infection: a prospective study. In:18th Annual Meeting of the American Society of Andrology. Tampa, Florida, 15-19 April 1993 [Abstract 9].
13. Inaba T, Fischer NE, Riddick DS, Stewart DJ, Hidaka T. HIV protease inhibitors saquinavir, indinavir and ritonavir: inhibition of CYP3A4-mediated metabolism of testosterone and benzoxazinorifamycin, KRM-1648, in human liver microsomes. Toxicol Lett 1997, 93: 215-219.
© 2002 Lippincott Williams & Wilkins, Inc.