This study was a cohort study meaning that the investigators evaluated a large population of men and compared those that received testosterone replacement therapy prescriptions with those who did not and compared their risk of non-fatal heart attacks.
My intent is to discuss this topic and evaluate what this study tells us, and what is does NOT tell us, about the risks of testosterone replacement therapy. As with all studies, the results may not be as they appear in the author’s conclusions.
How the study was conducted
The investigators used the Truven Health Marketscan to collect data. This database includes diagnoses, procedures, and prescriptions. They evaluated the available data from 2006 to 2010. There was available data on enrollment history, date of birth, gender, ICD-9 diagnosis codes, treatments, and outpatient prescriptions. There was no data available on reasons for testosterone prescription, race, laboratory findings, occupational, environmental, or lifestyle factors.
The investigators then placed the patient data into cohorts provided they had at least 22 months of continuous data on the individuals. The idea was to have 12 months of data before a prescription (pre-prescription) and 90-180 days of data after a prescription (post-prescription). They compared testosterone replacement prescriptions as the treatment or test group and a PDE5 prescription (i.e. viagra) as the control group for comparison.
The testosterone prescription group included any new testosterone replacement therapy prescription that did not contain estrogen. There were 55,593 individuals in this group. They did not differentiate one form of testosterone replacement from another. There were 167,279 new PDE5 prescriptions in this timeframe. There was no data on how much medication was actually consumed in either group.
The post-prescription group was followed until one of the following occurred: they had a diagnosis of non-fatal MI, they refilled their first prescription, or 90 days after their first prescription whichever occurred first. Men with a history of MI prior to their first testosterone (or PDE5) prescription were excluded.
The authors used a pre/post-prescription ratio, what they called a rate ratio (RR), to analyze the rate of MI before and after the use of testosterone. They further broke the information down by age.
For all men: They found an increased RR of 1.36 for all men who were prescribed testosterone. Before testosterone replacement, the rate of MI among all ages was 3.48 MIs per 1000 men. After testosterone replacement, the rate for this same group was 4.75 MIs per 1000 men.
For men <65 years of age: The RR was 1.17 which was slightly increased. Before testosterone replacement, the MI rate was 3.22 per 1000 men. After testosterone replacement, the rate was 3.76 per 1000 men.
For men >65 years of age: The RR was much higher at 2.19. Before testosterone replacement, the MI rate was 5.27 per 1000 men but increased to 11.52 per 1000 men after starting testosterone replacement. It should be noted that there were only 20 MI’s in the 7,054 men > 65 years of age after starting testosterone replacement revealing relatively low numbers in an aging population.
This study does show an increased risk of non-fatal myocardial infarction (MI) after starting testosterone replacement therapy in men 65 yrs of age or older in the first 90 days after initiating therapy.
However, there are several issues in this study.
First, they only followed these men for 90 days. This study does not evaluate a longer-term look at the risks or benefits of testosterone therapy. It is conceivable that longer term use of testosterone lowers risk but would not be revealed in this study. Interestingly, when patients didn’t refill their testosterone prescription they followed them through to 180 days and noted a decreased risk (back to baseline) of MI in these patients. They attributed this return to baseline risk to the lack of testosterone replacement. However, they did not evaluate those that did continue to take testosterone for the same period of time. It is possible that this effect was simply due to stabilization in their physiology independent of testosterone replacement.
Next, there is no information on prescribed dose or testosterone levels. It is impossible to accurately determine the effect of testosterone replacement independent of dose. As I’ve stated before, testosterone replacement therapy (TRT) is an art and there are multiple factors that must be addressed in order to optimize each patient’s physiology. Several of these factors could increase the risk of MI if not optimally controlled. Increased estrogen levels could play a significant role but this data was unavailable in this study.
This type of data is epidemiologic. We are making observations only. The problem is that many people will read this study, take it at face value, and try to apply this across all patients who may benefit from testosterone replacement therapy (TRT).
There may be truth in this study but there are multiple potential confounding issues. I agree that further studies are needed.
Other studies show different results
There are multiple studies showing that testosterone therapy is beneficial for cardiovascular disease. Read the post on low testosterone for more information on this topic.