A model of BMD changes after alendronate discontinuation to guide postalendronate BMD monitoring

Setting: Data were derived from the Fracture Intervention Trial Long Term Extension (FLEX), the largest multicenter clinical trial of its type to date.

Participants: Four hundred four women who had received an average of 5.1 years of alendronate during the Fracture Intervention Trial and were subsequently observed for 5 treatment-free years (on placebo) during the FLEX trial were used to estimate the change in BMD over time.

Results: IfamanagementchangesuchasalendronatereinitiationwouldbeconsideredwhenBMDT-score dropsbelow- 2.5, themodelshowsthatwomenwith total hipBMDgreaterthan -1.9 T-scores at the time of alendronate discontinuation have less than a 20% probability that at follow-up, monitoringBMDwillbe below the threshold within 5 years. The model performed similarly, and results are provided over a range of management change thresholds from -1.75 to -3 T-scores.

Conclusions: Using the tool developed in this analysis, it is possible to estimate when BMD repeat measurement after alendronate discontinuation could potentially be useful. Measuring BMD within 5 years after alendronate discontinuation is unlikely to change management for women with total hip BMD 0.6 T-scores above a prespecified retreatment threshold within the range of -1.75 to -3 T-scores.

Context: Women stopping alendronate are commonly monitored with serial bone mineral density (BMD) measurements, yet no information exists on how frequently or for whom these measurements should be performed.

Objective: The objective of the study was to develop a tool to guide post-alendronate BMD monitoring.

Design: A predictive model was constructed to estimate the time until a given percentage of women's BMD T-scores drop below a given threshold that indicates a management change (such as retreatment) would be considered. This model was then used to estimate the time it would take for groups of women defined by their baseline BMDs to drop below the given threshold.