Strong Bones: Are DXA Scans Overused?

Regardless of political opinions, gender, race, ethnicity, or medical specialty, the one thing every physician agrees upon is the need to control medical spending and reduce unnecessary testing.

As physicians, we need to become smarter and wiser. And many calls have been made for physicians to deploy the costly diagnostics in medicine a little smarter. The Choosing Wisely campaign is based on this very principle.

Over the last several years, it seems like most medical societies and associations have started encouraging conservative choices in medical spending. Recently, I read that one of the top hospitals in America, Johns Hopkins Hospital, has taken steps to make smarter decisions about use of medical tests, resulting in an estimated savings of $1.25 million per year.

One medical test in particular that appears to be little discussed in the area of conservation is the measurement of bone mineral density by the dual-energy x-ray absorptiometry (DXA) scan. Although Choosing Wisely and the USPSTF have said not to screen patients younger than 65, I believe there still needs to be a little more guidance to these recommendations.

The guidelines read "The USPSTF recommends screening for osteoporosis in women aged 65 years and older and in younger women whose fracture risk is equal to or greater than that of a 65-year old white women who has no additional risk factors."

Bad grammar notwithstanding, based on scientific evidence, the guidelines should read: the DXA scan should not be ordered for patients under 65, and not more frequently than every 10 years.

The CHOOSING WISELY recommendations also agree with the USPSTF recommendations, however, they support the use of recommendations as guidelines, not clinical rules.

"The therapeutic thresholds proposed in this Guide are for clinical guidance only and are not rules," they wrote.

Research

I am surprised that no medical societies have yet to come out with their own, similar recommendations. And, as we recently discussed on the Questioning Medicine podcast, the accuracy of the DXA scan is poor.

There are several factors that can alter the result of a DXA scan, which include, but are not limited to, the x-ray tech operating the machine, the way the patient lays on the table, and the clothes the patient is wearing during the scan. If the patient shifts at all, it effects the calibration of the machine.

Depending on the study you read, the variability between DXA scans can be as high as 5% to 6%. Although 5% to 6% might sound like a small amount, changes in bone growth are measured down to the hundredth and even thousandth decimal point. A 5% to 6% change takes years to overcome.

In a study by Seo et al, in the Journal of Bone Metabolism, researchers tested the efficacy of the DXA scan, and determined that the scan did not accurately reflect losses in bone strength either.

Seo et al, tested bone specimens that had been immersed in an HCL solution for anywhere from 10-100 minutes to degrade bone mineral density with different levels of demineralization. The specimens were then measured with a DXA scan and quantitative computed tomography.

But after all of the tests, and calculations, Seo et al, found that the actual bone mineral density of the specimens had changed far more than either the DXA scans or the quantitative computed tomography were able to show.

"Actual decrease of bone strength was much greater than that of [bone mineral density] by both DXA and [quantitative computer tomography]," Seo et al, wrote.

Research published in JAMA Internal Medicine, by Hillier et al., looked at 4,124 women, average age 72 ± 4 years, from 1998 to 1990 and then again 8 years later. Over an average of 5 years after the second bone mineral density measure, 877 women reported a nontraumatic nonspine fracture. Among those incidents, 275 were hip fractures. An additional 340 women reported spine fractures.

Hillier et al, made adjustments for age and weight change, and found that initial and repeat bone mineral density measures were similarly associated with fracture risk for nonspine (hazard ratio (HR) 1.6), spine (odds ratio 1.8-1.9), and hip (HR 2.0-2.2) fractures. Results reached P<0.001 significance for all models.

Areas under the receiver operating characteristic curves (AUC) did not reveal any significant differences that discriminated nonspine (AUC 0.65), spine (AUS 0.67-0.68), or hip (AUC 0.73-0.74) fractures between models with initial bone mineral density, repeat bone mineral density, or initial bone mineral density plus change in bone mineral density.

Even when Hillier et al, stratified by initial bone mineral density t scores of normal, osteopenic, or osteoporotic, high bone loss, or hormone therapy, the results remained unaltered.

"In healthy, older, postmenopausal women, repeating a measurement of [bone mineral density] up to 8 years later provides little additional value besides the initial bone mineral density measurement for predicting incident fractures," the authors wrote.

Redundancy

So based off of the research by Hillier et al, we see that getting a scan every 8 years didn't change patient outcomes.

Yet, Medicare still pays for DXA scans every 2 years, and the American Association of Clinical Endocrinologists recommends screening yearly until the patient is stable.

But there's redundancy with this approach. Much like with statin therapy, once the patient is on the treatment, why bother with repeated testing?

It's like using an umbrella to shield against the rain while you're swimming in a pool. When the patient is already being treated, retesting will not change the course of therapy.

Because the error in the DXA scan is so great and the decrease in bone growth is such a small number, even in the best case scenario it takes at least 10 years to get outside the test's margin of error.

The error in the test does not change when a patient is on bisphosphonate therapy. And it still takes well over 7 years, at best, to get outside the error of the test, and be able to certifiably say the drug is effective.

DXA scan costs also vary by region ranging from $150 to $300. Testing more frequently than every 10 years doesn't provide any additional clinical information as long as you have a result from one DXA scan. At any point in time, physicians can use the old numbers with the new patient characteristics to calculate the patient's risk stratification.

If all physicians were to stop the excessive use of unnecessary DXA scans, then by my very conservative estimate, Medicare alone would save over $50 million per year. Rescreening with a DXA scans is pointless, and if I had it my way, the practice would be banned.

Thank you for reading. Your thoughts are welcome. We are always available on Twitter @MedQuestioning and @AndrewBuelt. You can also email us at questioningmedicine@gmail.com.

For more Questioning Medicine:

Prostate Cancer Screening

Physician Roles in Lethal Injection

The Pain Management Fiasco

DXA Scan Overuse

Sham Lipid Management

EHRs and Attention Deficit

Andrew Buelt, DO, and Joe Weatherly, DO, are family medicine residents in St. Petersburg, Fla. Together, they co-produce the podcast Questioning Medicine, where they deconstruct issues confronting today's clinicians. In this guest blog, Buelt gives his take on the overuse of DXA scans.