A large observational study from Sweden found that only a quarter of patients started on drugs for prophylaxis of osteoporosis were still taking them after 3 years, dropping to 14% after 4 years; however long-term use did reduce risk of fractures significantly.

There is much evidence that persistence (long-term adherence) is poor with drugs for prophylaxis of osteoporosis, but reported values vary widely. The authors of this paper report experience from Sweden, using data from a historical cohort derived from Swedish prescribing registers. They extracted data on all patients aged 50 and over newly prescribed (no eligible prescription in previous five months) any one of alendronate, risedronate, strontium ranelate, and raloxifene between June 2005 and December 2009: other bisphosphonates were not widely used in Sweden during the study period, and HRT was no longer routinely recommended for this use. Patients on systemic glucocortocoid therapy were excluded, as were those appearing to have osteoporosis secondary to other conditions. Persistence was investigated using survival analysis., and medication possession ratio (MPR) was used to measure compliance in persistent patients. Switching between medications was included. The data were linked with national patient and mortality registers to allow analysis of adherence and fracture incidence, for which the primary outcome measure was hospitalised osteoporotic fractures.

There were 56,586 eligible treatment-naïve patients in the cohort, with a total follow-up of 61,322 patient-years of observation time. The mean age was 71 and most (86%) were women. Overall, treatment adherence was poor, with 23% filling only the index prescription and median time on treatment (earliest time at which half of participants had stopped) was just over a year (378 days). Persistence with bisphosphonates was highest and was lowest with strontium ranelate. Across all drugs and the whole study period, 51%, 35%, 25%, and 14% were still on treatment after 1, 2, 3, and 4 years, respectively.

In total, there were 2,980 identified fractures in the study cohort: only 2% had more than one reported fracture. Persistence with treatment was found to be significantly associated with the 3-year risk of fractures. The 3-year fracture risk for increasing periods of persistence compared with <1 month of treatment were: 1 month to 1 year treatment, hazard ratio (HR) 0.86 (p=0.091); 1 to 2 years, HR 0.67 (p<0.001); and 2 to 3 years, HR 0.59 (p<0.001). Because of the methods used, this indicates a consistent level risk reduction, not an increase over time.

The authors conclude that in this Swedish population, persistence with drugs for prophylaxis of osteoporosis is poor with barely half of those started on treatment still taking it after one year. This drops to a quarter after three years. The data did not allow analysis of how many patients might have restarted treatment at a later date, but this may be worth investigation. There was a clear association between continuing treatment and reduced risk of fractures that is consistent with the results of clinical trials. The authors comment that in Sweden, the recommended duration of osteoporosis treatment is 3 to 5 years, however most patients will not reach even the lower of these and they recommend additional research to examine factors affecting persistence and what the potential cost to society of non-persistence might be.