Decreased nonvertebral fracture incidence with zoledronate use for >3 years: post hoc analysis of a Randomized...

The anti-fracture efficacy of most osteoporosis drugs is assessed in trials of up to 3 yr duration. However, treatment of osteoporosis is required long-term, so it is important to know whether effectiveness of treatment changes over time. To date, this information has only been available from open extensions of the treatment groups from clinical trials. Such data are subject to selective loss of frailer patients from the cohort, and no placebo comparator group is available to permit reliable measurement of anti-fracture efficacy. The 6-yr Auckland zoledronate RCT administered zoledronate or placebo every 18 mo to 2000 osteopenic women aged >65 yr. It is longer than most osteoporosis trials, and only 3.6% of participants withdrew or were lost to follow-up. Therefore, anti-fracture efficacy can be validly compared between the first and second halves of this study. There were 178 non-vertebral fractures in the placebo group and 108 in the zoledronate group (excluding fractures of the hands, feet, and face). Fracture rates per 1000 women-years in the placebo group were 28 (95% CI 23, 35) and 32 (26, 40) in years 1-3 and years 4-6, respectively. In the zoledronate group, fracture rates were 23 (18, 30) and 13 (9, 18) per 1000 women-years in the first and second halves of the study. The rate ratios for non-vertebral fractures were 0.83 (0.60, 1.14) in years 1-3, and 0.40 (0.27, 0.58) in years 4-6 (between-period comparison, p < .05). Rate ratios for total fragility fractures, which include vertebral fractures also, showed a similar pattern: years 1-3, 0.76 (0.56, 1.02) and years 4-6, 0.39 (0.29, 0.53). These findings suggest that the efficacy for non-vertebral fracture prevention of anti-resorptive drugs has been underestimated because of the short-term evaluation of their effects. Further, it indicates that longer term treatment with these drugs may yield greater benefit to patients.