Reader Comments

Post a new comment on this article


Posted by agelessmd on 17 Feb 2014 at 02:55 GMT

In this paper, the authors compared the myocardial infarction and mortality rates of men prescribed testosterone for low T versus men prescribed PDE inhibitors. The glaring errors in this study include that the authors did not measure or know the testosterone levels of the men in the PDE-I cohort, nor did they measure or know the baseline or post-treatment T level in the testosterone treated group. The study extracted the data from a review of insurance submitted prescriptions and post-prescriptions insurance ICD diagnosis codes. What the authors compared in this paper was a group of men with presumably low testosterone (who may not have received adequate treatment for low T) against an unrelated cohort of men with unknown but presumed average testosterone. In essence, this PLOS article did not control for the variable they were supposedly studying - testosterone! Clearly, the two groups are not comparable and the study is of no value. Assuming the men given T prescriptions had low T, the authors may have inadvertently lent support to the more established findings across two decades of studies linking men with low testosterone levels to significantly higher levels of myocardial infarctions and mortality from all causes, since the PLOS study’s low testosterone group had higher levels of MIs and mortality for the first 3 months of treatment, but not after that time frame. However, it is hard to draw any conclusion from this paper as they did not have scientifically valid comparable cohort groups or reliable patient data.

No competing interests declared.


hooverr replied to agelessmd on 20 Feb 2014 at 22:44 GMT

The commenter is correct that we did not have information on pre or post-treatment levels of testosterone, as well as a number of other potentially interesting variables, and this will surely be of interest for subsequent investigations. However, what we were attempting to do was to document the effects of testosterone therapy as it is currently being used in the U.S. male population. We believe we have accomplished this, and the conclusion is that as it is currently being used, it appears to be related to a substantial increase in the risk of myocardial infarction.
The commenter is incorrect about our comparison group. The group of men using PDE-I was used only to assess the risk of MI that might be associated with behavioral changes associated with drugs used for related indications. The primary comparison group was the same group of men who were in the testosterone exposed group (in other words, they served as their own controls). Specifically, in the group of men given testosterone, we compared their rate of myocardial infarction in the year prior to their first prescription to their rate in the 90 day period following their first prescription. This resulted in the 2 to 3-fold increased risks reported. We then evaluated the risk of myocardial infarction during the period from 91 to 180 days after the first prescription in the subgroup of men who did not refill their prescription, a time period during which they likely took little or no testosterone. The risk for this group in this time period returned to the rate level observed in the year prior to their first prescription. Using study subjects as their own controls is a powerful study design, and the results appear to be a fairly strong indictment of testosterone therapy.

No competing interests declared.