The Clinical Effectiveness of Denosumab (Prolia®) for the Treatment of Osteoporosis in Postmenopausal Women, Compared to Bisphosphonates, Selective Estrogen Receptor Modulators (SERM), and Placebo: A Systematic Review and Network Meta-Analysis

To assess the effectiveness and safety of denosumab (Prolia®) compared to bisphosphonates (alendronate, ibandronate, risedronate, zoledronate), selective estrogen receptor modulators (SERMs; bazedoxifene, raloxifene) or placebo, for the treatment of osteoporosis in postmenopausal women (PMW). Systematic searches were run in PubMed, Embase & Cochrane Library on 27-April-2022. Randomized controlled trials (RCTs) that included osteoporotic PMW allocated to denosumab, SERMs, bisphosphonates, or placebo were eligible for inclusion. RCTs were appraised using Cochrane Risk of Bias 2.0. Bayesian network and/or pairwise meta-analyses were conducted on predetermined outcomes (i.e. vertebral/nonvertebral fractures, bone mineral density [BMD], mortality, adverse events [AEs], serious AEs (SAEs), withdrawals due to AEs, AEs caused by denosumab discontinuation). A total of 12 RCTs (k = 22 publications; n = 25,879 participants) were included in the analyses. Denosumab, reported a statistically significant increase in lumbar spine (LS) and total hip (TH) BMD, compared to placebo. Similarly, denosumab also resulted in a statistically significant increase in TH BMD compared to the raloxifene and bazedoxifene. However, relative to denosumab, alendronate, ibandronate and risedronate resulted in significant improvements in both femoral neck (FN) and LS BMD. With regards to vertebral fractures and all safety outcomes, there were no statistically significant differences between denosumab and any of the comparator. Relative to placebo, denosumab was associated with significant benefits in both LS and TH BMD. Additionally, denosumab (compared to placebo) was not associated with reductions in vertebral and nonvertebral fractures. Finally, denosumab was not associated with improvement in safety outcomes, compared to placebo. These findings should be interpreted with caution as some analyses suffered from statistical imprecision.