Phosphodiesterase-5 inhibitors as novel cardioprotective agents – have we reached threshold for large-scale clinical trials?

We read with great interest the compelling findings of Andersson et al demonstrating strong association between on-demand use of phosphodiesterase-5 (PDE5) inhibitors (PDE5is) and a reduction in cardiovascular mortality and heart failure hospitalisation in patients followed up following their first myocardial infarction (MI). We previously reported in Heart a similarly strong association between PDE5i use and reduced mortality in a cohort of patients with type 2 diabetes and attendant high cardiovascular risk.1 In our study, we showed that the association was stronger in patients with prior MI, supporting a cardioprotective action, although we were unable to draw conclusions about underlying mechanisms of protection. As the authors concede, and in similarity to our study, the observational nature of their work and the presence of unmeasured confounding factors (eg, marital status, physical activity, diposable income and educational attainment) make it impossible to establish concrete conclusions. Even so, the work by Andersson et al adds to the weight of association supporting a cardioprotective action of PDE5is, and their study design allows for several key additional observations not previously reported. First, the effect on mortality shows a dose-dependent protective effect. Second, the same effect on mortality or heart failure was not observed in patients taking an alternative treatment for erectile dysfunction (local alprostadil), effectively eliminating erectile dysfunction as a source of bias. Third, their finding of significantly reduced rate of heart failure (HF)hospitalisation, independent of any effect on new incident MI or revascularisation, adds to an evidence base supporting this drug class in heart failure. While the largest to-date randomised trial of PDE5is (RELAX) failed to show benefit in heart failure with preserved ejection fraction,2 the majority of HF cases post-MI are of heart failure with reduced ejection fraction (HFrEF). PDE5is have been trialled in several small-scale clinical studies of HFrEF, demonstrating improvements in exercise capacity and quality of life scales (reviewed in ref3),and supported by evidence from animal models showing improved contractile function, reduced adverse cardiac remodelling and improved survival.4 Thus, the findings by Andersson et al are, to our knowledge, the first to provide clinical evidence that PDE5is in heart failure may modify hard end points such as mortality rate and hospitalisation. Finally, further studies in animal models show that PDE5is reduce infarct size and protect against ischaemia-induced arrhythmias when given around the time of MI +/- reperfusion (see ref 4 for review). While this is unlikely the mechanism underlying the reduction in HF events in this study (PDE5i use commenced >30 days after index MI and no difference in new MI events), protective effects of PDE5i cardioprotection may be even more pronounced in patients already taking a PDE5i at the time of their MI. Larger-scale randomised clinical trials, adequately powered for hard endpoints, are surely now urgently needed to investigate PDE5is as promising novel cardioprotective agents in these settings.