A, Treatment changes following monitoring DXA for 196 scans. B, Treatment changes following significant decrease in bone mineral density (BMD) for 36 scans. A change was considered to be due to DXA if (1) drug treatment was started or changed in the setting of significantly decreased BMD, (2) treatment was stopped in the setting of stable or significantly increased BMD, or (3) treatment was stopped in a patient with osteopenia whose fracture risk score (determined by the Fracture Risk Assessment Tool [FRAX]; World Health Organization) would not merit treatment. A change was considered to be not due to DXA if (1) drug treatment was changed because of adverse effects or patient preference, (2) treatment was changed in the setting of stable or significantly increased BMD, or (3) treatment was stopped in the setting of significantly decreased BMD. A treatment change had to occur within 6 months of DXA or by the next clinical encounter.
Combs BP, Rappaport M, Caverly TJ, Matlock DD. “Due” for a Scan: Examining the Utility of Monitoring Densitometry. JAMA Intern Med. 2013;173(21):2007-2009. doi:10.1001/jamainternmed.2013.8998
Opinions differ on the utility of monitoring dual-energy x-ray absorptiometry (DXA) to assess responses to treatment for low bone mineral density (BMD).1- 3 Some argue that routine monitoring DXA may be unnecessary because approximately 98% of postmenopausal women treated with alendronate sodium experience an increase in BMD, and variation in subsequent BMD measurements by DXA may obscure the treatment effects.4
This study aimed to understand the utility of monitoring DXA scans by assessing (1) clinician rationale for ordering monitoring DXA and (2) the treatment changes that follow among average-risk women who are receiving treatment for low BMD. We hypothesized that monitoring DXA would rarely lead to treatment changes.
We identified 1782 patients at the University of Colorado Hospital, Aurora, who had undergone more than 1 DXA scan between January 1, 2003, and December 31, 2011. After excluding men (n = 120), those receiving medications or with conditions known to cause secondary osteoporosis (n = 580), and women not receiving treatment (n = 533), 549 women receiving treatment for low BMD remained. Of these, we reviewed the medical records from a random sample of 92 patients. Clinician rationale for ordering monitoring DXA and the treatment changes that followed were assessed.
Monitoring DXA was defined as any DXA performed on a patient being treated for low BMD, excluding the first DXA scan. All other scans were considered to be screening studies.
To explore clinician rationale for ordering monitoring DXA, we reviewed all the documentation within 6 months prior to the DXA scan or at the last clinical encounter. Quotations from the ordering clinician were recorded and categorized by topic. Changes in treatment following DXA were recorded as being due to monitoring DXA or due to factors other than monitoring DXA (Figure).
Of 1647 DXA scans in 549 patients, the mean (SD) number of scans per patient was 3.0 (1.1) (range, 2-10 scans), with a mean interval of 2.4 years between scans. The mean age of our population was 68.4 years, 70 patients (76%) were white, and 91 (99%) were treated with bisphosphonates. For the 92 patients under review, a total of 196 monitoring DXA scans were performed. The mean 10-year probability of hip or major osteoporotic fracture as determined by the Fracture Risk Assessment Tool (FRAX; World Health Organization) was 3.5% and 13.3%, respectively.
The primary rationale for ordering scans was that they were “due” (177 of 196 scans [90%]). Other rationale for monitoring DXA and representative quotations from the ordering clinician are reported in the Table. Most scans (165 [84%]) resulted in no treatment changes (Figure, A). Among the 36 scans showing a significant decrease in BMD, 26 (72%) resulted in no treatment changes (Figure, B).
Our data indicate that clinicians frequently order monitoring DXA scans out of a perception that they are due and rarely make changes in treatment based on the results. Even when DXA showed a significant decrease in BMD, treatment changes were uncommon. We are aware of no other studies that have assessed clinician rationale for ordering monitoring DXA or treatment changes that follow interpretation of the results.
We suspect that the frequency of monitoring DXA reflects adherence to professional guidelines, many of which recommend routine monitoring every 1 or 2 years while the patient is receiving treatment.2,3 Nonetheless, clinicians may feel uncomfortable escalating treatment on the basis of BMD changes because decreases in BMD during treatment do not reliably predict future fracture risk.5 Notably, most patients who lose BMD during the first year of treatment regain much of that in the following year even if the treatment is not changed.6 The findings from this single-center study, however, may not be generalizable to other institutions.
How often monitoring DXA should be used is uncertain, although finite health resources and our obligation to avoid unnecessary interventions require us to reconsider routine use of monitoring DXA among average-risk women who are receiving treatment for low BMD.
Corresponding Author: Brandon P. Combs, MD, Division of General Internal Medicine, University of Colorado Denver School of Medicine, 8111 E Lowry Blvd, Ste 120, Denver, CO 80230 (firstname.lastname@example.org).
Published Online: July 22, 2013. doi:10.1001/jamainternmed.2013.8998.
Author Contributions: Dr Combs had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Combs, Caverly, Matlock.
Acquisition of data: Combs, Rappaport.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Combs, Rappaport.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Matlock.
Administrative, technical, and material support: Combs.
Study supervision: Combs.
Conflict of Interest Disclosures: None reported.
Previous Presentations: This study was presented as an abstract at the 36th Annual Meeting of the Society of General Internal Medicine; April 26, 2013; Denver, Colorado.