Underestimation of Vertebral Fractures After Denosumab Discontinuation.

Abstract

To the Editor: We read with interest the article on vertebral fractures after stopping denosumab by Cummings and colleagues.1 Denosumab discontinuation is followed by a rebound effect with an increase of CTX above baseline values for 2 years and a decrease of BMD to baseline value after 1 year.2 Spontaneous clinical vertebral fractures, most often multiple, occur in the 8 to 16 months (median, 11.2 months) following the last denosumab injection.3 The risk of clinical vertebral fractures in the 12 months following the last denosumab injection was evaluated at 7.3%.4 This risk is certainly higher with a follow-up of 16 months and the inclusion of nonclinical fractures. Cummings and colleagues1 calculate an exposure-adjusted participant incidence of vertebral fractures (radiological and clinical) of 7.1 per 100 participant-years, from which only 39.3% were clinical fractures. We think that the difference between the previously published data and the actual study is because the duration of the follow-up was not sufficient to capture all the vertebral fractures. Moreover, we found that this study has methodological caveats. In the Fracture Reduction Evaluation of Denosumab in Osteoporosis (FREEDOM) trial and its extension, the yearly incidence of new radiological vertebral fractures was 2.22% to 3.15% in the placebo group and 0.70% to 1.86% in the denosumab group.5 In the present post hoc analysis, the rate of new vertebral fractures per 100 participant-years increased to 8.5% (5.5% to 11.5%) after discontinuation of placebo. This rebound effect in the placebo group is surprising. Moreover, the authors mention an annual risk of vertebral fractures of 7% (5.2% to 8.7%) during the on-treatment period with placebo, 2.5 to 3.0 times higher than what was published in the FREEDOM trial. During the on-treatment period with denosumab, the rate of new radiological vertebral fractures was 1.2% (0.9% to 1.6%), similar to what was published in the FREEDOM trial and its extension. After discontinuing denosumab, the rate of new vertebral fractures per 100 participant-years increased to 7.1% in the combined group. This combined group was composed of 327 women of the FREEDOM trial and 678 women of the FREEDOM EXT trial. The median time of follow-up after denosumab discontinuation (7 months after the last dose) was 0.5 years (0.2 to 1.4 years) and 0.2 years (0.1 to 0.7 years), respectively. As expected, the median (Q1, Q3) follow-up time was longer in participants with vertebral fractures than in those without vertebral fractures: 1.3 years (0.5 to 1.4 years) versus 0.5 years (0.2 to 1.4 years) in FREEDOM, and 1.4 years (0.4 to 3.4 years) versus 0.2 years (0.1 to 0.6 years) in FREEDOM EXT. We conclude that the duration of follow-up was not sufficient to capture all the vertebral fractures occurring after denosumab discontinuation. A follow-up of at least 16 months after the last denosumab injection is required, or 9 months after the 7 months from the last dose according to the definition used by the authors. Thus, 25% to 33% of the vertebral fractures were not captured for the 327 women who discontinued FREEDOM, and 66% to 75% of the vertebral fractures were not captured for the 678 women who discontinued FREEDOM EXT. We can thus estimate with a follow-up of 16 months after the last injection of denosumab, that the risk of vertebral fracture is close to 15%. All authors state that they have no conflicts of interest.