Efficacy and cost-effectiveness of alendronate for the prevention of fractures in postmenopausal women in Norway
Health technology assessment
|Updated
This HTA replaces Report 05-2010, ISBN 978-82-8121-331-9, which was withdrawn because of errors in the estimation of treatment effect.
Download
Key message
This HTA replaces Report 05-2010, ISBN 978-82-8121-331-9, which was withdrawn because of errors in the estimation of treatment effect.
Background: The Norwegian guidelines for prevention and treatment of osteoporosis and osteoporosis-related fractures recommend treatment with bisphosphonates for women with T-score less than -1.6 and previous fractures and also for women with T-score less than or equal to -2.5 without previous fracture. Only women with T-score equal to or less than -2.5 who have previous fractures will have their drug expenses reimbursed.
The guideline was last revised in 2005. Since then, the price of alendronate has been reduced by 80%. The University of Oslo has asked the Norwegian Knowledge Centre for the Health Services to evaluate how this price reduction affects the cost-effectiveness of alendronate.
Methods: We developed a model based economic evaluation with a lifetime perspective. The model follows a hypothetical cohort of women with respect to fractures of the hip, spine and wrist, late effects after fractures and mortality.
During the course of the model costs and health effects are accumulated as a result of the fractures. Half of the women receive treatment with a combination of alendronate, calcium and vitamin D. The other half only receives calcium and vitamin D. The estimated efficacy of alendronate in combination with calcium and vitamin D compared to calcium and vitamin D only was based on a systematic review of the literature.
Conclusions:
- Alendronate is likely to be a cost-effective alternative for women aged 65 and 75 years old with a T-score of equal to or less than -2.5 with no previous fracture and for women with a T-score of equal to or less than -2.0 who has suffered a previous fracture.
- The scarcity of efficacy data for women with a T-score above -2.5 without a previous fracture makes the inferences for these groups very uncertain.
Summary
Background
Norway has one of the highest incidences of osteoporosis-related fractures in the world. Norwegian guidelines for prevention and treatment of osteoporosis and osteoporosis-related fractures recommend treatment with bisphosphonates for women with a T-score less than -1.6 and previous fracture and also for women with T-score equal to or less than -2.5 without previous fracture. Only women with a T-score equal to or less than -2.5 who have suffered a previous fracture will have their drug expenses reimbursed.
The guidelines were last revised in 2005. Since then the price of alendronate, the most widely used bisphosphonates, has declined by 80% due to the introduction of generic competition. The Institute of Health Management and Health Economics at the University of Oslo has asked The Norwegian Knowledge Centre for the Health Services to evaluate the cost-effectiveness of alendronate for post-menopausal women after this price reduction.
Method
We developed a Markov model with three possible fracture events: fracture of the hip, vertebra and wrist. The model also contains four possible sequelae health states: mild, moderate and severe hip fracture sequela and vertebral fracture sequela. We performed analyses for women aged 55, 65 and 75 years old with T-scores of -1.5, -2.0 and -2.5 without previous fracture and T-score -2.0 with a previous fracture. Treatment with 70 mg alendronate per week in combination with calcium and vitamin D was compared to treatment with calcium and vitamin D only. Treatment was assumed to last for five years. The model followed the women from the age at treatment initiation until they all were one hundred years of age or dead.
Efficacy data were based on a review from the Cochrane Collaboration. We updated their literature search and conducted our own meta-analyses in order to obtain efficacy estimates for all of the groups requiring analysis.
In order to assess the robustness of our results, we performed one-way sensitivity analyses, probabilistic sensitivity analyses and we also calculated the expected value of perfect information on groups of parameters.
Results
In the probabilistic sensitivity analysis, the probability that alendronate is cost-effective for women with a T-score of less than -2.0 without previous fractures, varies from 0 % for the 55 years old, to 37 % for the 75 years old. For women with a T-score of less than -2.0 and a previous fracture and women with a T-score of -2.5 and no previous fracture, the likelihood that alendronate can be considered cost-effective varies from just below 10 % for the 55 years old, while it exceeds 90 % for the 65 and 75 years old.
We initially assumed that the Norwegian threshold value was NOK 500 000 per quality adjusted life year. Taking into account that this is not an official threshold, we also assessed how sensitive the conclusions were to this assumption. For women aged 75 years old with a T-score between -2.0 and -2.5, women aged 55 with a T-score less than -2.5 without a previous fracture and for women aged 55 with a T-score of -2.0 and a previous fracture, the conclusion may change if the willingness-to-pay is above NOK 500 000 per QALY.
The conclusions are uncertain for women aged 75 with a T-score between -2.0 and -2.5 without a previous fracture. According to the value of information analysis, the conclusion for this group is most affected by the uncertainty regarding the efficacy estimates. Further research on the efficacy of alendronate for women without previous fractures will reduce the decision uncertainty.
Discussion
The results of this analysis indicate that treatment with alendronate is likely to be cost-effective for women aged 65 and 75 years old with a T-score less than -2.5 without previous fracture and for women aged 65 and 75 years old with a T-score less than -2.0 with a previous fracture.
The conclusions are most uncertain for women aged 75 with a T-score of less than -2.0 and no previous fracture, women aged 55 with a T-score of -2.0 and a previous fracture and women aged 55 with a T-score of less than -2.5 and no previous fracture. Relatively small changes in the efficacy estimates of alendronate or the assumed willingness-to-pay per quality adjusted life year may change the conclusions for these groups.
All models are simplifications of reality and the study results are uncertain due to assumptions made and uncertainty in included parameters. We have only included fractures of the hip, vertebra and wrist as outcomes. In reality low bone mineral density increases the risk of all types of fractures. This simplification implies an underestimation of the cost-effectiveness of alendronate.
We did not include any side effects of alendronate in the model because there were no differences in the risk of side effects between treatment and control arms in the randomised controlled trials. Randomised controlled trials may however not be the appropriate study design to detect rare side effects, side effects that take long to develop or side effects that are more likely to occur in subpopulations.
Further research on the efficacy of alendronate on women at low risk of fracture may reduce the decision uncertainty for these women.
Conclusion
Alendronate is likely to be a cost-effective alternative for women aged 65 and 75 years old with a T-score of equal to or less than -2.5 with no previous fracture and for women with a T-score of equal to or less than -2.0 who has suffered a previous fracture.
The scarcity of efficacy data for women with a T-score above -2.5 without a previous fracture makes the inferences for these groups very uncertain.