LAUSR

We read with great interest the editorial by I.R. Reid published in Expert Opinion on Drug Safety [1]. The Author addressed cardiovascular safety of romosozumab. He thoroughly discussed data from the ARCH trial [2] that showed a significant imbalance in cardiovascular risk between the patients taking romosozumab and the controls taking alendronate: romosozumab treatment was associated with a hazard ratio of 1.87 (95% CI from 1.11 to 3.17, P = 0.002) for major adverse cardiovascular events. After considering other trials reporting on romosozumab safety, doubts on the biological plausibility of cardiovascular damage and high efficacy in fracture prevention the Author concluded that the safety uncertainties introduced by the ARCH study should not result in denying romosozumab treatment to individuals at very high risk of fracture. A crucial point discussed by the Author is the possible cardioprotective effect exerted by alendronate: the imbalance in major adverse cardiovascular events found in the ARCH trial may depend on favorable effects of alendronate. The results of the pivotal placebo-controlled FRAME trial that showed balanced cardiovascular outcomes between romosozumab and placebo [3] is in line with the possible cardioprotective role of alendronate (which has been suggested by several studies, although not definitively demonstrated) [4,5] and the absence of cardiovascular harm due to romosozumab. However, clinicians who take care of people at high risk of fracture, have a pragmatic issue to address: ‘Is romosozomab better than alendronate for my patient?’ (obviously, the no-treatment option is not considerable in clinical practice for patients at very high risk of fracture). Irrespectively of the reason(s) underlying the imbalance in major adverse cardiovascular events, (protection by alendronate or damage by romosozumab) a very large randomized controlled trial has shown the significant imbalance [2]. Risks and benefits of the 2 options (romosozumab or alendronate) should be assessed by estimating number needed to treat and number needed to harm before choosing the best treatment option. In the ARCH trial, the absolute risk of having a major cardiovascular event in the first year of treatment was 1.1% in the alendronate arm and 2.0% in the romosozumab group, whereas new clinical fractures occurred in 5.4% of the patients treated with alendronate and 3.9% of those treated with romosozumab. The estimated number needed to treat and number needed to harm (romosozumab versus alendronate) were 67 (67 patients treated to save one fracture) and 111 (111 patients treated to have one extra major cardiovascular event), respectively. Overall, one major cardiovascular event is expected to occur with romosozumab every 1.7 clinical fractures prevented. The number needed to treat decreases if the whole study period is considered: at 24 months the absolute risk of sustaining a new clinical fracture was 13% in the alendronate-to-alendronate group versus 9.7% in the romosozumab-to-alendronate group (number needed to treat = 30). An extra major cardiovascular event every 3.7 clinical fractures saved casts doubts on the risk-benefit ratio of choosing romosozumab instead of alendronate. Other calculations can be performed, including the reduction in morphometric vertebral fractures among the benefits due to romosozumab treatment, but the final risk-benefit ratio is only modestly ameliorated. Prof. Reid considered a possible alternative interpretation of the safety concerns raised by the ARCH trial: the imbalance in major cardiovascular events may be due to chance. The hypothesis is attractive and it is suggested by detailed analyses of the cardiovascular event data, but ‘some uncertainty remains’ as acknowledged by prof. Reid. Furthermore, the ‘chance-hypothesis’ is challenged by data from 2 published studies. Firstly, the Bridge randomized controlled trial [6,7] showed a numerical imbalance in the cardiovascular serious adverse events between romosozumab treatment (8 events; absolute risk = 4.9%) and placebo (2 events; absolute risk = 2.5%) in 245 men. Secondly, a large pharmacovigilance analysis of the US Food and Drug Administration Adverse Event Reporting System showed a significant association between romosozumab use and the occurrence of major cardiovascular events (OR 4.07, 95% CI: 2.39–6.93). Excluding from romosozumab treatment the patients at high cardiovascular risk is the reasonable way to reduce the estimated number needed to harm thus ameliorating the riskbenefit ratio. In Italy, the regulatory agency recommends a thorough calculation of individual cardiovascular risk in each candidate to romosozumab treatment. Furthermore, the