Treatment Duration May Explain Inconsistent Risks Seen With Testosterone Therapy

Adding to the debate over whether testosterone replacement therapy raises the risks of mortality and cardiovascular events, a new observational study conducted in Canada suggests that length of exposure may explain what have thus far been disparate findings.

While long-term exposure to testosterone is associated with lower odds of death, cardiovascular events, and prostate cancer, short-duration treatment is tied to a higher likelihood of death and cardiovascular events, researchers reported online May 7, 2016, in the Lancet Diabetes & Endocrinology.

The current study alone cannot spur any change in practice, senior author Robert K. Nam, MD (Sunnybrook Health Sciences Centre, Toronto, Canada), cautioned to TCTMD. Its results “shouldn’t alter what physicians are doing right now,” he said. “What we can say is that the study addresses the controversy [and] answers the controversy quite elegantly, in that there is a camp that thinks that testosterone is not safe and there’s another camp that thinks it’s safe.”

The US Food and Drug Administration (FDA), which warned in 2015 that testosterone therapy might raise the risk of heart attack and stroke, falls into the first camp, he said. “Now this study [has shown] that initially it isn’t safe, because it increases cardiovascular [event] and mortality rates, but if you take it long enough, then it drops. And that makes biologic sense.

“It takes time for the testosterone to take effect,” Nam continued, adding that the FDA had based its conclusions on studies involving approximately 3 months of testosterone, “which is not enough. And in fact, patients who had cardiovascular events in the short term, we don’t think they were dying because they were taking testosterone, we think they were dying because they were still suffering from testosterone deficiency.” The average time to heart attack in the current study was 30 months, he pointed out.

“I want to make it clear that we’re not saying the FDA is wrong,” Nam stressed, saying that the study “adds to the information that the FDA put out. So it’s complementary.”

Any Versus None, Short Versus Long

For the intention-to-treat observational cohort study, Nam, lead author Christopher J.D. Wallis, MD (Sunnybrook Health Sciences Centre), and colleagues looked at 10,311 men ages 66 years and older who received new prescriptions for testosterone replacement therapy in Ontario between January 2007 and June 2012. They then matched them with 28,029 untreated controls based on age, region of residence, comorbidity, diabetes status, and year of treatment. Individuals with recent cardiovascular events or newly diagnosed prostate cancer were excluded.

Over a median follow-up of approximately 5 years, patients who received any testosterone replacement therapy had lower risks of mortality (HR 0.88; 95% CI 0.84-0.93) and prostate cancer (HR 0.86; 95% CI 0.75-0.99) and a nonsignificant trend toward a greater risk of cardiovascular events (HR 1.10; 95% CI 1.00-1.20) compared with controls who did not receive testosterone.

The investigators found that patients with the shortest exposures had significantly greater risks of dying or experiencing a cardiovascular event compared with controls, whereas those with the longest exposures had lower likelihoods of death, cardiovascular events, and prostate cancer.

Median duration of testosterone was 2 months, 9 months, and 35 months in the short, intermediate, and long exposure groups, respectively.

“Using a time-varying exposure, we found that patients during their first 120 days of therapy had an increased risk of early death, a risk that subsequently dissipated with time or ongoing treatment,” Wallis et al report.

Overall, the study “identified a beneficial association of long-term testosterone replacement therapy, but also showed that short-term durations of therapy might be associated with harm,” they say.

However, the decreased risk seen with long-term therapy “should be considered hypothesis-generating rather than definitive,” the researchers stress. “We do not believe that these data provide sufficient evidence for a new indication for use of testosterone in prevention of cardiovascular disease.” Instead, the findings could inform clinicians who are counselling older men who are considering whether to start testosterone replacement therapy, they suggest. “Because of potential biases in observational studies, these findings require validation in a randomized clinical trial.”

However, Nam predicted that such a trial is unlikely to happen because it would be prohibitively large. “Testosterone is available generically, so industry is not going to be too interested in funding this type of study,” he said.

Intent Matters

Writing in an accompanying editorial, Michael S. Lauer, MD (National Institutes of Health, Bethesda, MD), agrees that based on our current knowledge, it is unclear whether “testosterone therapy in older men is beneficial or safe,” and that a large-scale intention-to-treat randomized trial would be appropriate. “But in the meantime,” he says, “Wallis and colleagues’ thoughtful observational analyses not only provide important insights into a vexing clinical problem, but also remind us of an often underappreciated component—intention to treat—of rigorous clinical science.”

By looking only at patients with new prescriptions, the current study tries to get around the possibility that patients are doing better long-term only because they were healthy enough to survive any potential short-term risks carried by testosterone, he explains. “The intention-to-treat design suggests that the effects of testosterone on health are unlikely to be simple and monotonic.”

Acknowledging that these thoughts “might seem esoteric,” Lauer points out that the Nurses’ Health Study serves as a prime example of why intents are important.

Initially, that study found hormone-replacement therapy to benefit women, but the subsequent Women’s Health Initiative trial found myriad health risks. “Maybe important confounders, such as socioeconomic status, had been overlooked? As it turned out, another important omission was failure to incorporate an intention-to-treat design,” he notes, pointing out that when that approach was applied to the Nurses’ Health data, researchers “identified varying effects of treatment over time and similar effects to those seen in the Women’s Health Initiative randomized trial.”

Lauer concludes: “Intents, as well as proper analyses of them, matter.”

Related Stories:

• Post Angiography Testosterone Therapy Increases Cardiovascular Events in High-Risk Men
• HRT Increases Venous Thromboembolism and Stroke Risk, But Not CHD
• Menopause Summit Refutes Role of HRT in Cardiovascular Disease