HOW RISK-BLIND ARE CLINICAL DECISIONS ABOUT OSTEOPORORSIS THERAPY BASED ON BONE DENSITY?

     Purpose: To evaluate, in the context of a randomized trial of a decision aid regarding osteoporosis treatment, the extent to which guideline-based treatment recommendations based on bone mineral density T-score criteria would lead to medication starts across groups with different 10-year osteoporotic fracture risk.

    Methods: In an ongoing randomized trial of an osteoporosis treatment decision aid, over 9 months we identified 91 consecutive postmenopausal women receiving care in primary care clinics who had bone mineral density test T-score results of -1 or less and who were not on osteoporosis medication.  Using the recently released WHO osteoporosis fracture risk (FRAX) calculator which takes into account not only bone mineral density but also  other osteoporosis risk factors (age, BMI, previous fracture, parental hip fracture, current smoking, glucocorticoid use, rheumatoid arthritis, secondary osteoporosis, alcohol intake of 3 or > units per day), we estimated the 10-year risk of a major osteoporotic fracture (e.g., clinical spine, forearm, hip or shoulder fracture) for each woman and tabulated the range of risk in each of three categories: femoral neck T-score of -1.0 to -2.0  (WHO osteopenia;  NOF do not treat unless you have risk factors and  T-score   -1.5 to -2.0); -2.0 to -2.5 (WHO osteopenia; NOF treat group); and < -2.5 (WHO osteoporosis; NOF treat group).

   Results: The following table shows the distribution of risk among the three bone mineral density based categories.  

Patients in each of these T-score categories, n (%)
Risk Levels	-1.0 to -2.0	-2 to -2.5	< -2.5
<10%	6(11)	1(4)	0(0)
10-20%	24(44)	6(24)	0(0)
21-30%	17(31)	10(40)	2(17)
31-40%	3(6)	6(24)	7(58)
41-50%	4(7)	1(4)	1(8)
>50%	0(0)	1(4)	2(17)

   Conclusion: Bone mineral density based treatment decisions are in large part blind to fracture risk. The use of the FRAX calculator should support a risk-sensitive approach to osteoporosis treatment.  Effective risk-based decision aids may assist with this task in the future.