Objective: To investigate changes in back pain in postmenopausal women with severe osteoporosis who received teriparatide for 24 months or switched at 12 months to raloxifene or no active treatment.
Study design and methods: This prospective, controlled, randomised, open-label, 2-year study enrolled 868 postmenopausal women with osteoporosis and a recent fragility fracture. After 12 months of teriparatide (20 microg/day), 507 patients were randomised to further teriparatide (n = 305), raloxifene 60 mg/day (n = 100), or no active treatment (n = 102) for another 12 months (substudy 1); in substudy 2, 199 patients continued teriparatide. All received calcium and vitamin D supplementation. Back pain was self-assessed by patients using a visual analogue scale (0-100 mm). Changes in back pain were analysed using a mixed model for repeated measures.
Results: During year 1, back pain decreased from a mean (SD) of 48.9 mm (24.0) at baseline by 11.5 mm (p < 0.001) in the total study population. In substudy 1, mean change in back pain from month 12 (randomisation) to 24 months was -2.2, -4.4 and +0.7 mm in the teriparatide (p = 0.076), raloxifene (p = 0.041), and no active treatment groups (p = 0.751). There were no between-group differences from randomization to 18 or 24 months. In a sensitivity analysis excluding patients with low baseline back pain (VAS < 30 mm), mean change from randomisation to endpoint was significant for teriparatide (-3.9 mm, p = 0.006) and raloxifene (-6.3 mm, p = 0.018) groups. Subgroup analyses of 503 patients who received teriparatide for up to 2 years showed that patients with a recent vertebral fracture had a greater decrease in back pain than those without (p < 0.05). Those with and without mild back pain (>or=30 mm), and those with and without severe back pain (>or=60 mm) at baseline all had a statistically significant reduction in back pain after 24 months (p < 0.05).
Conclusions: Teriparatide treatment is associated with significant reductions in back pain regardless of the presence of recent vertebral fracture. These results need to be considered with caution due to the open-label design of the study.