LAUSR

Introduction Osteolytic bone lesion is a hallmark of multiple myeloma (MM). Bisphosphonates have been shown to be effective inhibitors of osteoclastic activity, leading to improvement of skeletal related morbidity in MM. Based on their chemical structure, bisphosphonates can be classified into two categories: amino bisphosphonates (ibandronate, pamidronate and zoledronate) and non-amino bisphosphonates (clodronate, etidronate). Although both categories are efficient for the management of myeloma related bone disease, their effect on mortality has been controversial. Some studies demonstrated beneficial effect associated with the administration of amino bisphosphonates, while others showed no such effect. We evaluated if amino bisphosphonates improve survival compared with non-amino bisphosphonates, when added to standard therapy for myeloma. Methods To address our objectives, we conducted a systematic review and network meta-analysis of all randomized controlled trials (RCT) estimating the effect of bisphosphonates in MM. We identified trials by searching electronic databases such as MEDLINE, CENTRAL and EMBASE. We compared the effect of amino bisphosphonates group to non-amino bisphosphonates group, versus the reference placebo group or no treatment. The primary outcome was overall mortality. We conducted a network meta-analysis under a random effects model. Pooled hazard ratios (HR) with 95% confidence intervals were estimated. To further evaluate the effect between the groups, a test for subgroup differences was applied. All analysis were conducted using Revman and R. Results Overall 14 studies were identified from 1982 to 2015, enrolling a total of 2706 patients: 5 studies tested bisphosphonates with non-amino group and 9 with amino group. The results from the network meta-analysis (Figure 1) showed a decrease in overall mortality for the amino bisphosphonates (HR 0.76, 95% CI 0.62 to 0.94) compared to the non-amino bisphosphonates (HR 1.06, 95% 0.87 to 1.3) and placebo. The statistical test for subgroup differences was significant (Q=4.69, p=0.03). Conclusion Based on the totality of randomized evidence, we showed for the first time that amino bisphosphonates compared to non-amino bisphosphonates versus placebo or no treatment reduce mortality in MM. The use of amino bisphosphonate should be further encouraged as anti-myeloma treatment and not solely as supportive treatment. No relevant conflicts of interest to declare.