Is heart failure a debatable end-point for biphosphonate treatment in older osteoporotic population?


Dear Sir,

In the recent issue of the Journal of Internal Medicine, we were interested to read an intriguing article by Grove et al. [1] on the occurrence of heart failure with bisphosphonate or raloxifene treatment in a nationwide retrospective cohort study. All users of bisphosphonates and raloxifene (= 102 342; mean age 70.5 ± 11.4 years) were included in the ‘exposed’ group and for each user three age- and gender-matched subjects (= 307 026; mean age 70.5 ± 11.4 years) from the general population comprised the control group. The authors reported that the absolute risk of heart failure was 4.4% in the exposed group and 3.7% in the control group (< 0.01). Although the adjusted hazard ratio (HR) of heart failure was significantly increased in users of bisphophonates (HR 1.41, 95% confidence interval 1.34–1.48), raloxifene was not associated with an increased risk of heart failure. Furthermore, users of alendronate, a nitrogen-containing bisphosphonate, showed a dose-dependent reduction in heart failure risk, in contrast to users of other bisphosphonates.

Randomized controlled trials have demonstrated the efficacy of bisphosphonates in reduction of fracture risk in osteoporotic populations [2, 3]. However, various safety issues that were not noted in clinical trials have now emerged with postmarketing surveillance and increasing clinical experience. Although the effects of bisphosphonate therapy on several cardiovascular end-points (mortality, myocardial infarction and arrhythmias) have been studied, the link between these drugs and heart failure has not been investigated specifically in previous studies. Grove et al. [1] have identified an important issue with regard to bisphosphonate use; however, the study results should be interpreted with caution in view of several limitations. First and most important is the lack of echocardiographic data such as left ventricular ejection fraction. Secondly, age-associated physiological changes and the presence of several risk factors predispose older patients to develop heart failure, even when left ventricular ejection fraction is normal or near normal. Heart failure with a preserved ejection fraction is particularly common in older hypertensive women, and hypertension plays a key role in its pathophysiology, as in the study by Grove and colleagues [1], which included a higher rate of hypertension amongst bisphosphonate users than amongst control subjects. Thirdly, no data were presented regarding the occurrence of arrhythmias in the study groups, such as atrial fibrillation, which predispose patients to the development of heart failure with both reduced and preserved left ventricular ejection fraction.

As mentioned in this study, the various bisphosphonates act differently from each other because of their additional chemical groups; for example, the nitrogen-containing alendronate inhibits the enzyme farnesyl pyrophosphate synthase (FPPS) [4]. It has been suggested that FPPS may play a role in hypertensive cardiac hypertrophy, and knockout of genes encoding this enzyme reduced cardiac hypertrophy in rats [5]. Thus, bisphosphonates that inhibit FPPS may indirectly reduce hypertensive cardiac hypertrophy, which may explain why alendronate reduced the occurrence of heart failure in this study. Furthermore, another possible explanation for the increased heart failure risk is longer follow-up times for treatment with raloxifene or bisphosphonates other than alendronate.

In conclusion, this study demonstrated that patients exposed to bisphosphonates have an increased risk of heart failure compared with the general population. However, inclusion of older women in the exposed group with a higher prevalence of hypertension and many other risk factors for heart failure should be kept in mind whilst interpreting the study results. In addition, the decreased risk of heart failure associated with alendronate in this study was probably due to inhibition of FPPS.

Conflict of interest statement

None of the authors has any conflicts of interest to declare.