If you are reading this blog, you must already be aware that in the
United States, reimbursement for measurement of bone mineral density using DXA
has been cut substantially over the last few years for all users except in
hospital radiology suites. This has already led to a marked decrease in the
number of private offices still offering DXA as an important diagnostic and
As I have noted several times in the past, osteoporosis is still
under-recognized and under-treated, so this decreasing availability and use of
DXA is contrary to the important nationwide discussions about the needs to
improve preventive care.
If that was not enough threat to those at risk of sustaining an
osteoporosis-related fracture later in life (that includes every one of us
unless we unfortunately die prematurely from something else), recent articles
in the British Medical Journal1,2 have accepted that
baseline DXA is important but have gone on to suggest that follow-up DXA to
monitor progression or regression of disease is unnecessary. The current issue
of the Journal of Bone and Mineral Research includes two
point/counterpoint articles for and against serial measurement of
What messages are being put forward in these four articles, and which
are putting forward a sustainable opinion?
The articles in the BMJ, one a retrospective analysis of
two articles published in 1996 and 1998 (what took you so long?) and the second
an editorial, reasoned that the change in BMD was not the most reliable
predictor of reduction in fracture occurrence and therefore serial BMD
measurement was not helpful and should be discontinued.
The point/counterpoint articles in the JBMR provided very
reasoned arguments for and against serial DXA. To get the full gist of their
conflicting opinions, you will have to read the articles in full. In this blog,
I want to emphasize just a couple of points.
To me, the take-home message was not whether serial DXA is a good or not
good thing but that we have to be mindful whenever we translate data from
randomized, double blind, placebo-controlled clinical trials into our
clinical practice. Participants in these clinical trials are quite unlike the
majority of patients you see day to day in your office, and this seems to hold
true no matter what disease state or therapy is the subject of the trial.
In order to conduct and complete a randomized, controlled trial in the
most cost-efficient, timely manner, the investigators are carefully selected and
the trial participants even more so. Study subjects are carefully screened for
potentially confounding health issues and for concomitant medication use.
Investigators usually have to demonstrate that they can recruit study subjects
in a timely manner, retain patients in the study and optimize adherence to
study medication. Monitors not related to the drug company sponsoring the
trial, or to the investigators conducting the trial, oversee every item
documented about patient compliance to therapy, every reported adverse event,
and make sure that every study outcome parameter has been obtained correctly.
In the field of osteoporosis, it has been reported from one such
potential study center that fewer than 20% of patients referred to the practice
for ongoing care were eligible for any one of the four randomized, controlled
trials for which they were trying to enroll subjects.5
Randomized, controlled trials are critical for development of new
therapies â€” try to imagine where we would be without them â€” but they
are often not representative of the patient in your office right now. Case in
point â€” the 66-year-old woman I saw recently who had sustained three
fragility fractures during her three years of oral bisphosphonate therapy. I
have good reason to believe that she was adherent to her medication as she
claimed to be: She was certainly compulsive enough to hand me two double-spaced
typed pages, the first one and a half pages listing her prescribed medications
with the list of her concomitant diseases taking up the last half page.
Assuming she did take her bisphosphonate on schedule on an empty stomach with a
full glass of water, nothing else to eat or drink for 30 to 60 minutes all the
while sitting or standing, who knows how any of the other medications may have
affected the effectiveness of therapy or which of her diseases was overwhelming
the effectiveness of therapy?
She was/is a candidate for osteoporosis therapy, but she is also,
without question, a candidate for serial DXA â€” as often as once a year!
Not every patient will be like this, but at the same time, we cannot
practice â€œcookie cutterâ€� medicine. In some patients, particularly
those who have not yet fractured, a baseline DXA with follow-up two or three
years later may be appropriate. Interim measurement of a marker of bone
resorption would also be appropriate during the first six months or so of
therapy. Patients need early reinforcement that therapy is working as it
should. In others the clinical situation will dictate more frequent BMD and
I am confident that the authors of the articles arguing against serial
DXA are aware of this, and many of them do monitor their patients on therapy. I
am less confident that those insurance carriers who have to pay for the serial
DXA will see it that way, or if they do, you will have to go through hoops to
convince them one patient at a time. Time is of course something we all have
aplenty of in day-to-day practice.
For more information:
1. Bell KJ. BMJ. 2009;338:b2266. PMID: 19549996.
2. Compston J. BMJ. 2009;338:b1276. PMID: 19549994.
3. Watts NB. J Bone Miner Res. 2009;24:1643-1646. PMID: 19712042.
4. Adler RA. J Bone Miner Res. 2009;24:1647-1648. PMID: 19702482.
5. Dowd R. Osteoporos Int. 2000;11:533-536. PMID: 10982170.