Abstract
The diagnosis of hypogonadism in people living with HIV (PLWH) remains challenging by the lack of a standardised diagnostic algorithm. Since sexual hormone-binding globulin levels are commonly increased in PLWH, guidelines recommend assessing free testosterone (FT) along with total testosterone levels. We compared different online equations available to estimate FT levels and found a good correlation amongst all algorithms. Estimating FT levels increased diagnostic accuracy of hypogonadism and therefore should be encouraged in clinical practice in PLWH with clinical symptoms of hypogonadism, even when total testosterone levels are normal.
Keywords
Introduction
Hypogonadism remains a common condition of endocrine disruption found in men living with HIV1,2 A relatively smaller proportion of female patients also suffer from androgen abnormalities, although this is likely underdiagnosed.3–5 The real prevalence of TD in people living with HIV (PLWH) is still a matter of debate, 6 in part due to the difficulties in its diagnosis, related to non-specific clinical symptoms and the lack of a standardised diagnostic algorithm.3,7
Laboratory diagnosis of hypogonadism relies on the measurement of morning serum total testosterone (TT) (7–11 a.m.) on at least two occasions along with luteinizing hormone (LH) and sexual hormone-binding globulin (SHBG) measurements.3,7 Since SHBG levels are commonly increased in PLWH,8,9 current guidelines recommend assessing free testosterone (FT) in individuals with suspected clinical hypogonadism to improve diagnostic accuracy. 10
The gold standard for FT measurement is by equilibrium dialysis, however this may be costly and not routinely available, while commercially available FT assays might be unreliable. 11 Alternatively, the measurement of serum SHBG along with TT and albumin allows for the calculation of FT levels using previously derived online equations. 12 However, there is still limited data comparing different methods of calculating FT levels and how the use of different diagnostic methods can change the reported prevalence of hypogonadism in PLWH.
Methods
We conducted a retrospective data collection from 817 PLWH aged between 40-80 years old, attending Chelsea and Westminster Hospital NHS Foundation Trust to receive HIV care between 2009–2020, aimed at analysing the correlation between different equations available to estimate FT levels and to estimate the prevalence of biochemical hypogonadism in our cohort. Our internal care protocol recommends the routine titration of TT, SHBG and albumin levels in PLWH over the age of 50. Patients were excluded from the analysis if data on TT, SHBG or albumin were missing or if they were on testosterone replacement therapy (TRT). This study was approved by the Trust as a service evaluation and quality improvement audit.
Calculated free testosterone (cFT) was determined using the free online calculator available at the ISSAM website (http://www.issam.ch/freetesto.htm), using the levels of TT and SHBG (measured by radioimmunoassay) and serum albumin applied to the equation developed by Vermeulen et al. (Eq(V) 13 Four other empirical equations (Eqs) for estimating cFT: Eq(S): Sodergard et al. 14 ; Eq (NW): Nanjee and Wheeler 15 ; Eq (LH):Ly and Handelsman 13 ; and Eq (EK): Emadi-Konjin et al. 16 were compared to the ISSAM online calculator. Linear regression models were used for statistical analysis.
Biochemical hypogonadism was defined according to British Society for Sexual Medicine (BSSM) guidelines on male adult TD as: TT<8 nmol L−1 and/or cFT<0.180 nmol L−1 (confirmed biochemical hypogonadism) and TT<12 nmol L−1 and/or cFT<0.225 nmol L−1 (possible hypogonadism). 7 While the diagnostic range for hypogonadism in female patients is less clear, we used a threshold of TT<0.8668 nmol L−1 and/or cFT<0.0052 nmol L−1 (below age 50) and of TT<0.6934 nmol L−1 and/or cFT<0.0035 nmol L−1 (above age 50) as indicative of androgen deficiency in female patients.17,18
Results
Summary of participant characteristics.
Data are presented as median (interquartile range) unless stated otherwise.
Abbreviations: ARV, antiretroviral therapy; HIV, Human Immunodeficiency; RNA, Ribonucleic Acid Virus; SD, standard deviation; SHBG, sexual hormone-binding globulin.
aCalculated with the online algorithm available at http://www.issam.ch/freetesto.htm
Overall, 13.7% (46/335) of male individuals presented with TT<12 nmol L−1, of which 2.9% (10/335) had TT<8 nmol L−1. SHBG levels were increased in 30.4% (102/335) of tested individuals and significantly associated with age (p = 0.0435). Using the ISSAM calculator, 14.9% (50/335) had cFT <0.225 nmol L−1, with 6.3% (21/335) exhibiting cFT<0.180 nmol L−1.- Overall, 16/65 (26.2%) male patients identified to have biochemical hypogonadism presented with low or borderline cFT (<0.225 nmol L−1), despite having normal TT and thus would have been missed if FT levels were not calculated along with TT. There was strong positive linear correlation between all tested equations for calculating cFT (Figure 1). The empirical formula by Handelsman Eq (LH), had the lowest R2 value compared to other formulae, but a strong positive linear relationship was still observed. Linear correlation between cFT (ISSAM) and cFT using four different empirical equations (male patients, n = 335).
Of the female patients screened, only 11 had adequate data for cFT calculation. cFT values estimated by the four tested equations showed strong positive linear relationship to the cFT values (ISSAM) as all R2 values were above 0.97 (average: 0.9942). Similarly, Eq (LH) was least reliable for calculating cFT in female patients or when TT and cFT were below 0.6 and 0.0107 nmol L−1, respectively.
Discussion
This study confirms the validity of online equations to estimate cFT levels in PLWH. Furthermore, it highlights the need for considering further investigations in patients with normal TT and clinical symptoms suggestive of TD to explore if they may benefit from testosterone replacement therapy, as there might be the effect of a raised SHBG in PLWH affecting serum total testosterone measurements.9,8 We found that SHBG levels were increased in 30.4% of tested individuals and significantly associated with age (p = 0.0435) as previously reported. 19 Other possible causes for increased SHBG concentrations include hypothyroidism, chronic liver disease and hepatitis C. On the other hand, circulating testosterone levels decrease with age and in several comorbidities such as diabetes mellitus, dyslipidemia, metabolic syndrome and obesity.19,20
Prior to the advent of effective antiretroviral therapy (ART), hypogonadism was more frequently found among men with AIDS, in the current ART there is an increasing prevalence of forms of hypogonadism that are secondary to chronic medical conditions and partially reversible. 21 In these types of hypogonadism – often classified as functional hypogonadism – pituitary gland LH and FSH levels are inappropriately normal or low.22,23 The clinical relevance of such abnormalities in androgen levels depends on the presence of signs and symptoms of testosterone deficiency including fatigue, weight loss, loss of libido or erectile dysfunction or depressive symptoms. 23 However, some of these symptoms may be subtle and overlapping with those arising from co-existing co-morbidities and therefore a low threshold for suspicion of testosterone deficiency should be kept in PLWH.
In this study we found an overall prevalence of biochemical hypogonadism of 17.9% in males. In previous studies, the reported prevalence of male hypogonadism in PLWH remains variable, with studies quoting values ranging from <10% to >50%3,24 A meta-analysis by Santi et al. reported an overall prevalence of hypogonadism in HIV-infected men of 26% that increased up to 40% when FT was considered. 6 Interestingly, Pezzaioli et al. reported a 20.2% prevalence of overt biochemical hypogonadism in a single centre analysis including 94 HIV+ men of 169 screened, in line with the 18.8% prevalence recorded in this study. 25
This study is limited by the absence of relevant clinical information required to confirm the diagnosis and the absence of serum gonadotropins to distinguish between primary and secondary hypogonadism. However the scope of this work was to assess the agreement between different methods to estimate serum testosterone levels and how this could impact on overall prevalence of biochemical hypogonadism. Other limitations include the underrepresentation of non-white race and female patients limiting the generalizability of the results. Despite the known pitfalls in measuring testosterone levels in women with radioimmunoassays and in the absence of direct assays to measure free testosterone, the recommendation is also to measure SHBG and calculate free testosterone using online algorithms. However, our results in the female population need to be interpreted with caution and further validated with a larger sample. Other limitations include the absence of at least a second TT measurement and information on the time of measurement of serum TT was not always available and therefore the effects of diurnal variations in testosterone concentrations could not be eliminated. 26
In summary, our results indicate that online algorithms give a reliable prediction of cFT levels if TT, SHBG and albumin are available. Its use should be encouraged in clinical practice in patients with clinical features suggestive of hypogonadism even if TT levels are normal. In our cohort, 4.8% (16/335) of the male participants had normal TT levels and thus would have been missed if FT was not calculated. Since SHBG levels are often elevated in PLWH, 9 the estimation of FT levels may increase diagnostic accuracy of hypogonadism as previously shown.12,25 This is particularly relevant in ageing men, in whom clinical diagnosis of hypogonadism is even more challenging as symptoms are often nonspecific and mimicked by other prevalent disorders.21,22
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Author Contributions
SW and OH collected and analysed the data and wrote the manuscript; BP analysed the data and wrote the final version of the manuscript; AM and MB designed the study and contributed to revise the final version of the manuscript; All the authors read and approved the final version of the manuscript.
