Absolute risk reduction in osteoporosis: assessing treatment efficacy by number needed to treat

Postmenopausal osteoporosis is a chronic condition due to decreased bone mass, leading to reduced bone strength and increased fracture risk. Currently available pharmacological treatments include antiresorptive agents (bisphosphonates and raloxifene) and bone-forming agents (strontium ranelate and two different parathyroid peptides). Comparison via reduction in relative risk of fracture may produce artificially high reductions in fracture risk for some agents. Responder analysis based on absolute risk reduction (ARR, the arithmetic difference between events rates with and without treatment over a fixed time) and a related parameter, number needed to treat (NNT, the number of patients needed to treat over a fixed time to prevent one event) may provide more reliable parameters. We reviewed placebo-controlled, randomized, double-blind, pivotal phase 3 trials employed as part of the regulatory process, in order to calculate ARRs and NNTs for vertebral and hip fracture over 3 years for antiosteoporotic agents currently available in Europe. The NNT values to prevent one vertebral fracture over 3 years range from 9 for the strontium ranelate to 21 for ibandronate. NNT values for hip fracture over 3 years range from 48 for strontium ranelate to 91 for three of the bisphosphonates. Our analysis indicates that the bone-forming agent strontium ranelate may have the lowest NNT for the prevention of both vertebral and hip fracture. Responder analysis may enable translation of clinical trial results into guidance for routine clinical practice by indicating the amount of effort needed to prevent the same event in comparable populations with different treatment options.