Meta-analysis: Testosterone treatment not linked to short-, medium-term CV events in men

Erin D. Michos, MD, MHS, FACC, FAHA, FASE, FASPC

In an editorial I co-wrote that was published alongside this study (Michos ED, et al. Lancet Healthy Longev. 2022;doi:10.1016/S2666-7568(22)00115-5), I stated that I would still be cautious about testosterone therapy until the larger dedicated CV outcome trial TRAVERSE is published.

Meta-analyses are only as good as the underlying studies that are being pooled together. Unfortunately, these studies were still short-term (ranging from 3 months to 3 years), with mean follow-up of only 9 months, and that is really too short of a time period to assess CV safety or efficacy. The pooled studies were not powered for a primary CV endpoint. Longer-term data area really needed. Some, but not all, prior studies have suggested a worrisome signal of CV harm for testosterone therapy in older men. A 2013 observational study from the Veterans Affairs system suggested a 29% increased risk for CVD in men receiving testosterone therapy compared with controls (Vigen R, et al. JAMA. 2013;doi:10.1001/jama.2013.280386). The TOM trial was small (209 men) but was stopped early due to a fourfold increase in CV events among men treated with testosterone compared with controls (Basaria S, et al. N Engl J Med. 2010;doi:10.1056/NEJMoa1000485). The Testosterone Trials CT angiography study, also small (170 men), showed increased progression of noncalcified plaque and total coronary plaque in men randomly assigned  to testosterone compared with those assigned placebo (Budoff MJ, et al. JAMA. 2017;doi:10.1001/jama.2016.21043). Together these trials led to the FDA warning that testosterone products should not be used simply for low testosterone associated with aging and that there is possible increased risk for MI and stroke with use. There were other studies that didn’t show this excess risk, but none of them were sufficiently powered for CV outcomes.

In my own research, postmenopausal women with higher endogenous testosterone levels had increased risk for CV events (Zhao D, et al. J Am Coll Cardiol. 2018;doi:10.1016/j.jacc.2018.01.083), and premenopausal women with polycystic ovary syndrome (PCOS) with hyperandrogenism also have elevated risk. Women, on average, have delayed onset of CVD about 10 years after men. This raises the question about whether higher testosterone levels are associated with CV risk, explaining the sex differences we see in CVD risk.  

This current meta-analysis about testosterone replacement therapy from pooled trials is reassuring about the safety for short-term use, but we really don’t have data for long-term safety yet until the TRAVERSE 5-year CV outcomes trial is published. I would still be cautious about using testosterone therapy in men at high CV risk until we have better long-term safety data. We shouldn’t be using testosterone therapy simply to treat a number of low T (generally defined as < 300 ng/dL), but only if accompanied with symptoms of hypogonadism (low libido, erectile dysfunction, loss of muscle mass, decreased muscle mass and strength). Ultimately it comes down to shared decision making, a conversation about how symptomatic a man is and his underlying CV risk profile. I would want to make sure his lipids and BP are well controlled before starting, and if there is uncertainty about his CV risk, a coronary artery calcium score could help refine risk assessment.

In sum, I don’t think the current analysis changes the current recommendation. I would certainly continue therapy in men who are already taking this and tolerating it, but tell them that we still don’t know long-term effects. But at this time, I would not newly start men at high CV risk on therapy until we had more data, or at least provide a warning of caution.