We thank Dr. Watts for his thoughtful comments on our article. There are several points we would like to expand upon. First and foremost, the use of the FRAX calculator is meant to be one means of identifying high-, medium-, and low-risk patients receiving glucocorticoids. Its utility in this patient population at the time the recommendations were being developed was not determined. In a recent position statement for the International Society for Clinical Densitometry and International Osteoporosis Foundation, Leib and colleagues suggest that the dose of glucocorticoids is a factor not fully accounted for by the FRAX calculation (Leib ES, Saag KG, Adachi JD, Geusens PP, Binkley N, McCloskey EV, et al. Official positions for FRAX Clinical regarding glucocorticoids: the impact of the use of glucocorticoids on the estimate by FRAX of the 10 Year Risk of Fracture From Joint Official Positions Development Conference of the International Society for Clinical Densitometry and International Osteoporosis Foundation on FRAX. J Clin Densitom 2011;14:212–9). It is therefore not meant to supplant clinical judgment.
The purpose of categorizing patients into high, medium, and low risk was to incorporate clinical risk factors for fracture other than bone density into the consideration for preventive and treatment strategies without creating overly complex recommendations. Our expert advisory panel discussed the use of the FRAX risk of hip fracture, but found that dividing high, medium, and low risk over 3% (the National Osteoporosis Foundation threshold for treatment based on absolute risk of hip fracture) was difficult, and thus, the focus alternatively was placed on the risk of a major osteoporotic fracture. Furthermore, with the version of FRAX that was used in 2008 when this decision for defining high-, medium-, and low-risk patients in part by a FRAX score was made, there were seemingly more examples in the sample patients given to the expert advisory panel where the hip fracture risk did not fit with the panels' clinical impression of risk. For example, as shown in Table 1, using FRAX, a 55-year-old white man had a higher risk of hip fracture than a 55-year-old white woman when the T score was −2.0 and other factors were equal.
|Major osteoporotic fracture, %||Hip fracture, %|
This discordant finding in FRAX further supported the panels' attention to the risk of major osteoporotic fracture over the risk of hip fracture in defining high risk. We would also add that, although hip fractures are the most concerning complication of osteoporosis, glucocorticoids likely have a preferential effect on trabecular bone, leading initially to a selectively greater risk of spine fractures. Of note, in the example that Dr. Watts gives, a patient with a T score of −2.6 would be high risk by hip, but not by major osteoporotic fracture risk. However, as noted in our article, high-risk patients were suggested to be those with a FRAX score of 20% or greater or a T score of −2.5 or less or a history of a fragility fracture. Therefore, this patient would be considered at high risk based on these other considerations. Based on Dr. Watts's comment that glucocorticoids increase hip fracture risk more than major osteoporotic risk, we would be less concerned about misclassification from high to medium risk since both would be treated based on the guidelines. We would be more concerned that some patients with a 10-year hip fracture risk of >3% could be misclassified as low risk, <10%, for major osteoporotic fracture and thus not receive treatment. We believe that this would occur infrequently.
In summary, we strongly encourage clinicians to consider other factors in assessing risk as shown in Table 1 of our original article, and as Dr. Watts appropriately points out, the FRAX hip fracture risk should be added to this list.