Health & Medical Endocrine disease

Testosterone Therapy in Men With T2D and Metabolic Syndrome

Testosterone Therapy in Men With T2D and Metabolic Syndrome

Discussion


In this systematic review and meta-analysis, using rigorous inclusion criteria, we have identified seven double-blind placebo-controlled RCTs that assess the effects of testosterone treatment on glucose metabolism in men with T2D and/or the metabolic syndrome and low to low-normal testosterone levels. Although the RCTs were of moderate to high quality, there was a large between trial heterogeneity. Differences in the choice of HOMA-IR-based methodology to estimate insulin resistance was identified as a major contributor as accounting for HOMA-IR methodology eliminated significant between trial heterogeneity. Collectively, the results indicate that testosterone treatment modestly improved insulin resistance in RCTs that estimated insulin resistance by using the HOMA1 equation. In contrast, in RCTs that used a computer-based model to determine insulin resistance (HOMA2), insulin resistance was not improved. While HOMA1 gives a good approximation of beta cell function relative to gold standard insulin clamp studies, the HOMA2 model attempts to account for variations in hepatic and peripheral glucose resistance, for increases in the insulin secretion curve for plasma glucose concentrations above 10 mm, and for the contribution of circulating proinsulin [https://www.dtu.ox.ac.uk/homacalculator/]. HOMA2 may thus provide a more precise physiological basis for estimation of insulin resistance. Therefore, testosterone therapy results in improved glucose-insulin profiles as assessed in the fasting state using HOMA1, which would be consistent with a beneficial effect to ameliorate insulin resistance. However, this may be modulated indirectly via factors which are captured by HOMA2, possibly explaining the disparity in trial results when the newer model is used.

Testosterone treatment had no significant effect on glycaemic control, assessed by HbA1c. Although we did not identify evidence of publication bias, we are aware of one unpublished RCT in 180 men with T2D with baseline HbA1c of 7·0–9·5% and total testosterone < 10·4 nm. In this RCT, testosterone treatment had no significant effect on HbA1c or on HOMA-IR, compared to placebo. Insufficient information precluded inclusion of this RCT into the meta-analysis, although it is expected that inclusion of this relatively large negative study would have further reduced the difference in glycaemic outcomes between testosterone and placebo groups. Similarly to the trial by Gianatti, this unpublished RCT recruited exclusively patients with established T2D. Reductions in insulin resistance with testosterone treatment were predominantly reported in RCTs that included men with the metabolic syndrome but without established T2D. This raises the possibility that testosterone treatment may be more effective in improving glycaemic outcomes in men with the metabolic syndrome compared to men with established T2D.

Our results differ from previous meta-analyses in this area, which have generally shown more favourable effects of testosterone treatment on glucose metabolism. These previous meta-analyses were smaller, including between 228 and 483 participants, showed larger between trial heterogeneity (I 38–82%), and were not restricted to placebo-controlled double-blind RCTs, but instead included nonblinded, open-label trials. In addition, the two more recent studies using more stringent HOMA2 modelling were not included.

Given that men in all the studies included had relatively well controlled T2D at baseline, the effects of testosterone treatment in men with poorly controlled T2D are unknown. Of note, RCTs in such populations would be more difficult to conduct, given the efficacy of standard antidiabetic medications. In addition, baseline testosterone levels in RCT participants were only modestly reduced, and whether testosterone treatment improves glucose metabolism in men with more marked reductions in testosterone levels remains unknown. Experimental studies of induced hypogonadism have not identified a serum testosterone threshold below which insulin resistance increases. Consistent with this, the inverse relationship of insulin resistance with testosterone levels in men with T2D does not have a clear breakpoint and remains present even in men with testosterone levels extending into the normal range. Marked reductions in testosterone are relatively uncommon in men with T2D and require careful assessment for underlying classical hypogonadism. Such men may well require testosterone treatment irrespective of glycaemic considerations.

While there is evidence that testosterone treatment can improve glucose metabolism in preclinical studies by a variety of cellular mechanisms, the degree of which they are operative in men is unknown. Testosterone treatment modestly increases muscle mass and decreases fat mass, changes expected to be metabolically favourable. Of the meta-analysed RCTs, only one study reported changes in body composition assessed by rigorous methodology. Despite the expected decrease in fat mass and increase in muscle mass, in that study insulin resistance was not improved. Similarly, in the unpublished RCT, insulin resistance was not improved despite a significant increase in lean body mass. In a recent study of 57 obese men, testosterone treatment did not improve insulin sensitivity assessed by euglycaemic clamps, despite significant decreases in fat mass and increases in total fat mass, and in a chemical castration study of healthy young men, no changes in insulin resistance were observed despite significant increases in fat mass. One explanation for this apparent paradox is the observation that in men with T2D testosterone treatment preferentially reduced the amount of subcutaneous, but not of the metabolically more active visceral fat. Similarly, testosterone treatment had no significant effect on visceral adipose tissue in most but not all RCTs conducted in overweight or obese men not specifically selected for the presence of T2D and/or the metabolic syndrome. Additional studies are needed to clarify whether testosterone acts selectively on subcutaneous compared to visceral fat, or whether specific circumstances exist which lead to one or other reservoir being affected.

Given that the metabolic syndrome and T2D are slowly progressive conditions, it remains possible that longer duration of testosterone treatment, beyond the maximum duration of 12 months in current RCTs (Table 1) may have more marked effect on glucose metabolism. While uncontrolled registry studies have demonstrated progressive improvements in glucose metabolism in men treated with testosterone up to 6 years, these observations have yet to be confirmed in controlled trials.

With respect to the effects of testosterone treatment on other cardiovascular risk factors, effects in the different RCTs were relatively modest and consistent with findings from RCTs in of testosterone therapy in men from the general population. Effects on lipids were modest and may be neutral from a cardiovascular perspective as both decreases in pro-atherogenic lipid fractions (total cholesterol, LDL cholesterol, lipoprotein a) and in the theoretically cardioprotective HDL cholesterol were reported. Triglyceride levels did not change, consistent with the absence of documented changes in visceral fat. While none of the RCTs found an effect on blood pressure, modest decreases in CRP were seen, although whether this is independent of changes in body composition is not known.

Testosterone treatment did not have a significant effect on constitutional symptoms suggestive of androgen deficiency, as assessed by the relatively nonspecific Aging Male Symptoms score. Heterogeneity in the instruments used to assess sexual function precluded a meta-analysis of sexual function. Some but not all RCTs reported improvements in sexual parameters although these were variable between the studies, including erectile function, sexual desire and overall sexual function. In one RCT, sexual and constitutional symptoms were worse in men with depression and microvascular complications, but did not correlate with testosterone levels, suggesting that these nonspecific symptoms are confounded by comorbidities. This provides a potential explanation for the relatively modest and inconsistent symptomatic benefit of testosterone treatment in this population, which is distinct from the marked symptomatic benefits in men with classical, pathologically based hypogonadism who have unequivocally low testosterone levels and objective evidence of androgen deficiency.

In these relatively short term RCTs as expected, serious adverse events were few. Consistent with men in the general population, an increase in haematocrit was the most commonly reported adverse event, occurring in four of the seven RCTs. Cardiovascular events were numerically lower in men receiving testosterone a difference largely driven by the TIMES2 study. Although the number of events was very low, this may provide some assurance given the uncertainties surrounding the cardiovascular safety of testosterone treatment.

Strengths of the present meta-analysis include, in contrast to previous meta-analyses, the strict selection of double-blind placebo-controlled RCTs including two recent RCTs using more rigorous HOMA2 methodology to estimate insulin resistance, which allowed identification of sources of heterogeneity and the quantification of summary estimates across all published RCTs. We used the PRISMA statement to conduct our research, and carefully evaluated study quality using the full 25-item CONSORT checklist. Limitations include the limited number and relative modest size of included studies, lack of access to individual patient level data and the fact that included studies not always fully complied with CONSORT reporting criteria, which may weaken the precision of some of our estimates. Moreover, while none of the RCTs used 'gold standard' measurements of insulin resistance, HOMA-IR measurements have shown to correlate well with clamp-based technologies.

In conclusion, the results of this meta-analysis do not support the routine use of testosterone treatment to improve glucose metabolism or constitutional symptoms in men with relatively well controlled T2D and/or the metabolic syndrome and modest reductions in testosterone levels. This reinforces recommendations that, at the current state of evidence, lifestyle measures and use of standard therapy to optimize glycaemic control and comorbidities should remain the first line approach.



You might also like on "Health & Medical"

Leave a reply