Writing in the February 22, 2011 online edition of the Journal of Urology, researchers from Kaiser Permanente managed care plans in California examined the association between oral NSAID use and ED in a large, ethnically diverse cohort of men enrolled in the California Men’s Health Study [Gleason et al. 2011]. Subjects for this prospective cohort study included males aged 45 to 69 years of age who had enrolled in the managed care plans from January 2002 to December 2003.
NSAID exposure of interest, as determined by pharmacy data and self-reported use, was defined as receiving more than a 3-month supply of 1 or more prescription NSAIDs, any prescription for 3 or more doses per day, or a self-report of using NSAIDs at least 5 days per week; all types of prescription and over-the-counter NSAIDs were included in data gathering, and acetaminophen was excluded. Erectile dysfunction was assessed by a validated questionnaire.
Of the 79,130 men surveyed and providing full data, roughly 47% were NSAID users and 29% of all subjects reported moderate or severe ED. Overall, 35% of frequent NSAID users versus 24% of nonusers had ED of some degree. NSAID use and erectile dysfunction appeared to be strongly correlated with age; regular NSAID use increased from about 35% in men <50 years old to 55% in men >60 years of age, and ED correspondingly increased from 16% to 46%, respectively.
Overall, compared with men who did not take NSAIDs regularly, the frequent use of NSAIDs increased the odds of self-reported ED to some extent by about 70% (Odds Ratio = 1.72; 95% CI: 1.67-1.77 [see notes below about odds ratios]). When data were adjusted to take into account other factors that might influence ED — age, race/ethnicity, smoking status, diabetes mellitus, hypertension, hyperlipidemia, peripheral vascular disease, coronary artery disease, and body mass index — NSAID use increased the odds of suffering ED only by roughly 20% (OR=1.22; CI: 1.18-1.27, which was significantly less than the odds without taking those other factors into account). These associations persisted across age groups, and when using only 3 days/week rather than 5 days/week NSAID exposure as a criterion, or when also including acetaminophen use in the mix.
The authors conclude that their observational study suggests that regular NSAID use is associated with a greater chance of developing ED beyond what would be expected due to age and other comorbidities. Older men exposed to NSAIDs appear to be somewhat more susceptible to these drugs’ effects on ED. While the use of NSAIDs should not necessarily be avoided for this reason, this potential side effect of ED should be considered in the risk/benefit analysis when recommending the regular use of NSAIDs in men.
COMMENTARY: Further research would be appropriate to better understand and confirm this association of NSAIDs and ED. Meanwhile, it could be one more potential side effect that should be added to the long list of NSAID safety concerns elaborated in our ongoing series of UPDATES on this topic [here].
A strength of this large study by Gleason and colleagues is that it was adequately powered and prospectively designed in part to answer the specific research questions examined; thus, the approach avoided many of the pitfalls we have previously noted can be associated with retrospective data-mining studies [UPDATE here]. Still, data were not gathered (or, at least not included in their report) for resolving certain questions, such as: Which came first, NSAID use or ED? Beyond the comorbidities that were adjusted for in the study, what were the pain conditions for which NSAIDs were being taken? Might these conditions have affected development of ED? Would discontinuation of NSAIDs help to resolve ED in these patients? Therefore, based on the reported findings of this study, while NSAIDs might be a contributing factor it cannot be presumed unequivocally that NSAID use is causative of ED.
Technical Notes: The following additional observations may aid in interpreting the somewhat complex, possibly misleading, presentation of data in this study…
- The article abstract contains data that can be easily misunderstood and some of the information seems inconsistent with data presented in the body of the report. For our discussion above, and following, we referred to data described in the body of the report.
- The authors’ use of Odds Ratios to summarize their data is questionable. While presentation of data in terms of odds is favored by gamblers and some statisticians, many authorities on evidence-based medicine disparage this approach as being unhelpful and potentially misleading for clinical decision-making purposes. Knowing the odds and odds ratios may be essential for successful betting on horse races but an understanding of Risks and Risk Ratios is more helpful for medical practice.
- Odds and risks are not the same thing, although they often are confused with each other. Risk is more straightforward and represents the probability of an outcome — eg. an event or effect, like ED — occurring in a group of subjects, often expressed as a percentage of subjects in the group experiencing the outcome. Whereas, the odds of an outcome is the probability of the outcome occurring divided by the probability of the outcome not occurring in the group. For healthcare providers, odds are conceptually much more difficult to understand and translate into clinically meaningful terms.
- From data in the report by Gleason et al. we calculated the overall unadjusted Risk Ratio as being 1.46, which is a medium-sized effect and suggests that the frequent use of NSAIDs incurs a 46% greater risk of developing ED than might occur in men not using NSAIDs. This is in contrast to the 70% greater odds of ED denoted by the odds ratio in the report; so, clearly, risks and odds are not direcly comparable.
- The absolute risk increase of ED conferred by NSAIDs is only 11%, which further implies that for every 9 men frequently using NSAIDs 1 additional case of ED might occur, beyond what would normally occur without NSAID use (NNT= 1/.11 = 9.1). This might be of some clinical importance.
- However, the authors do not include data in their report to calculate adjusted risks or risk ratios, taking into account comorbidity factors that also might influence ED, but these risks and their effect sizes would most certainly be much smaller and possibly of only minor clinical significance.
- It seems of some concern that the authors do not indicate precisely how they made modifications — eg, mathematical weightings, etc. — to adjust their data for the potentially confounding factors of comorbidity (eg, diabetes, hypertension, etc.).
- This is not to say that the authors intended to obfuscate the issues by the creative use of statistics; however, in many studies data presentations seem guided more by the sophisticated analytical capabilities of computer programs than by clinical common sense. One critic has stated: “It is difficult to see any justification for the use of odds in what purports to be scientific study. It is just another example of the misleading effects of statistical computing packages when they are used without understanding or with disinterest” [John Brignell, 2006, at http://www.numberwatch.co.uk/].