Bisphosphonates in Multiple Myeloma. A Systematic Review and Meta Analysis.

Bisphosphonates are currently used as supportive therapy for multiple myeloma (MM). In 2001 we conducted a Cochrane systematic review (SR) showing that Bisphosphonates reduce vertebral fractures and pain but have no effect on other important outcomes. Here we report an update of that SR.

A comprehensive literature search of MEDLINE, EMBASE, LILACS, Cochrane database of randomized controlled trials (RCTs) and www.clinicaltrials.gov and meetings abstracts from American Society of Clinical Oncology, American Society of Hematology and European Hematology Association was undertaken to identify all phase III RCTs published until January 2009. We extracted data regarding overall survival (OS), progression free survival (PFS), vertebral and non vertebral fractures, skeletal related events (SREs), pain, hypercalcemia, grade III-IV treatment related harms. The time to event data and dichotomous data were pooled under the random effects model as hazard ratios (HR) and risk ratios (RR) respectively. Heterogeneity was assessed using the chi square test and I² statistic. Indirect comparison of various bisphosphonates was conducted according to the methods developed by Bucher and Glenny et al and were extended to calculate HR/RR.

Seventeen RCTs were included enrolling 3,010 patients. In comparison with placebo / no treatment, the pooled analysis demonstrated a beneficial effect of bisphosphonates on prevention of pathological vertebral fractures (7 RCTs, 1116 patients) [RR= 0.74 (95% CI: 0.62 to 0.89), P = 0.001], SREs (6 RCTs, 1334 patients) [RR= 0.81 (95% CI: 0.72 to 0.92), P = 0.001] and on amelioration of pain (8 RCTs, 1281 patients) [RR = 0.75 (95% CI: 0.60 to 0.95), P = 0.01]. We found no significant effect of bisphosphonates on OS, PFS, hypercalcemia or on the reduction of non-vertebral fractures. There was statistically significant heterogeneity for OS and pain endpoints. The heterogeneity for the outcome of pain could be explained by the variation in the pain scales used to measure pain. However, we also found that the beneficial effect of bisphosphonates on pain reduction was greater in patients who were asymptomatic at the start of treatment [RR= 0.28 (95% CI: 0.12 to 0.67)] compared to symptomatic patients [RR= 0.83 (95% CI: 0.69 to 1.00)] (Test of interaction: p = 0.005). The heterogeneity for OS was attributed to one RCT with unrealistic treatment effects (“an outlier effect”). Results of indirect meta analysis were consistent with the results from direct comparisons for the outcomes of vertebral fractures, SREs and pain. The indirect meta analyses did not find the superiority of any particular type of bisphosphonate over others. There were no significant adverse effects associated with the administration of bisphosphonates. In fact, only two RCTs reported osteonecrosis of jaw (ONJ). We also identified 7 observational trials evaluating 1068 patients for ONJ. These studies suggest that ONJ may be a common event (range: 0%- 51%). Since ONJ was only sporadically reported in RCTs the results from observational studies may be an overestimate due to their non-controlled design.

Adding bisphosphonates to the treatment of MM reduces pathological vertebral fractures, SREs and pain but - from the published evidence - not mortality. Assuming the baseline risk of 20%-50% for vertebral fracture without treatment, we estimate that between 8 - 20 MM patients should be treated to prevent vertebral fracture(s) in one patient. Similarly, assuming the baseline risk of 31%-76% for pain amelioration without treatment, we estimate that between 5 - 13 MM patients should be treated to reduce pain in one patient. Also, with the baseline risk of 35%-86% for SREs without treatment, we estimate that between 6 - 15 MM patients should be treated to prevent SRE(s) in one patient. No bisphosphonate appears to be superior to others.

Glasmacher:Celgene: Employment, Equity Ownership.