You'll want to discuss this with your doctor. Why the recommendation to treat that small aneurysm and cut your life short by two years?
Leaving Tiny, Unruptured Intracranial Aneurysms Untreated
Why Is It So Hard?
JAMA Neurol. 2018;75(1):13-14. doi:10.1001/jamaneurol.2017.2559
-
Original InvestigationManagement of Tiny Unruptured Intracranial AneurysmsAjay Malhotra, MD, MMM; Xiao Wu, BS; Howard P. Forman, MD, MBA; Charles C. Matouk, MD; Dheeraj Gandhi, MD; Pina Sanelli, MD, MPH
One of the frustrations
that drove me back to focusing more on delivery of care and less on
research was the difficulty in changing practice with the results of
research. Nearly 20 years ago, my research group started to study the
question of which intracranial aneurysms should be treated. The results
were startling: the vast majority of small, unruptured aneurysms should
be left untreated, even if the published evidence was off target by a
large margin.1
However, these results had little effect. In spite of 2 decades of
largely confirmatory evidence for very small aneurysms (arbitrarily set
at ≤3 mm in diameter) showing that coil embolization is not as safe as
some believe and that rupture and growth rates are extremely low, many
continue to recommend treatment for most of these aneurysms.
The updated analysis produced by Malhotra and colleagues2 in this issue of JAMA Neurology
paints the picture very clearly. The best approach for patients with
aneurysms measuring 3 mm or less in diameter is to ignore the aneurysms:
there is no need for follow-up imaging and certainly no need to try to
treat them. Treating very small aneurysms is projected to reduce a
person’s healthy lifespan by nearly 2 years. That’s right—you try to
treat that little aneurysm and you are likely to knock a couple of years
off someone’s life. Could the estimates used in the model change the
recommendation? The data would have to be unbelievably wrong to change
the recommendation. For example, the annual rupture rate of an untreated
aneurysm would need to be more than 1.7% to change the recommendation,
and no study has shown a rate close to this, with a best estimate being
0.23%. Furthermore, one can always argue that the underlying studies
could be better, but what happened to that oath we all took, “First do
no harm…”? With the best estimate suggesting we shorten a patient’s
lifespan by 2 years when we treat a very small aneurysm, we should stop
treating now rather than waiting for better data to change course.
Perhaps even more remarkable in this analysis is the
conclusion that we do more harm than good from monitoring patients with
magnetic resonance angiography regardless of whether the duration
between scans is 1, 2, or 5 years, and this is without any consideration
of cost. It would seem counterintuitive that more monitoring could be
harmful when the monitoring itself is completely safe. However, we have
seen from prostate-specific antigen testing for prostate cancer that the
result of a screening test may worsen health outcomes.3 Monitoring can push us to intervene and these interventions can have negative effects. It is an old story.
I suspect that these results, as with all those before
them, will be ignored by some practitioners. I would like to examine
several arguments I have heard through the years to justify treating
such small aneurysms.
One frequent argument for proceeding with treatment even
when the evidence recommends avoiding it is that the potential for the
aneurysm to rupture produces anxiety that will affect a patient’s
quality of life, making treatment the right decision. However, when it
is clear that the risk of rupturing the aneurysm or causing a stroke is
greater by treating the aneurysm than leaving it alone, this concern
should be reduced. We all know how suggestible patients can be, which
makes it even more critical that we carefully educate them, perhaps even
scripting the way we introduce the choices. Proper education and
counseling are much safer and more appropriate interventions for this
anxiety.
Another argument is that the procedural risks used in
these models are inflated and do not reflect those at one’s own
institution. In fact, the estimates in the models are based on the
published literature. Several studies have shown that there is a
tendency to underreport adverse outcomes in case series such as these,4
in part related to discomfort in accurately reporting adverse outcomes,
so published estimates are probably underestimates. Furthermore,
studies have shown that practitioners routinely underestimate their
complication rates. The solution here is for physicians to distrust
their own intuitions about local institutional complication rates and
outcomes. The literature is almost certainly more accurate and should
not be ignored. Your results are probably poorer than those published by
your colleagues.
Some believe that the fact that many ruptured aneurysms
are small means that the only way to prevent rupture is to treat small
aneurysms. Although this idea is logical on the surface, it presumes
that a static, small aneurysm existed for some number of months or years
before rupturing, but the literature does not support that supposition.
When we find and follow up small unruptured aneurysms, they almost
never grow and almost never rupture. It is possible that the small
aneurysms that rupture emerged days or even hours before rupture and
that the ones we find incidentally are disproportionately stable. This
possibility certainly fits the data.
A final argument I have heard to justify treatment of
very small aneurysms is that a particular aneurysm is at high risk for
rupture because of its configuration or location, the patient’s family
history, history of rupture of a different aneurysm, or some other
consideration. Undoubtedly, there are higher-risk tiny aneurysms, and
the literature suggests a few such risk factors. However, the overall
rates of rupture for tiny aneurysms are extremely low even when these
subgroups are included, so these factors are of questionable importance
in a given case.
Underlying these arguments are biases working behind the
scenes, perhaps even subconsciously. Although we would love to deny it,
we physicians are human and humans are by nature subject to an array of
biases.
An example of such a bias is optimism. Even the
proceduralists with the best skills have cases with complications. Going
to work every day requires that they minimize the emotional effect of
these complications on themselves. It would be too great a burden to
acutely feel a complete sense of responsibility for every avoidable
death or disabling complication, particularly in the highly risky and
complex area of cerebrovascular disease. Proceduralists must be built to
bounce back and move on; however, taken too far, this attitude can
create a bias toward underestimating the rate and outcome of these same
complications. This bias toward optimism about one’s own outcomes is not
universal, of course, and we have all seen both “cowboys” and
“turtles,” with the latter representing those we refer to when caution
is preferred.
Career advancement is a bias to which we are all
subject. For some, the longer the list of treated patients, the more
impressive the published case series, and the greater the bragging
rights and consequent referral base. Financial incentives also play into
treatment decisions more than any of us would like to admit. Whether a
physician benefits directly from treating more cases, as in private
practice or with similar incentive plans, or benefits indirectly by
achieving relative-value units that justify a high salary or bonus, the
incentive to do more in our health care system is a very strong one.
Physicians, department chairs, and hospitals are all complicit. One way
to reduce this bias is to make sure the physicians making the decision
about when treatment is necessary are not financially incentivized to do
so.
Finally, the bias to do things as they have always been
done is powerful in medicine. My generation of stroke physicians was
taught that unruptured aneurysms were time bombs that should be
identified and defused as soon as possible. It is extremely difficult to
unteach that belief, with years of experience and all our trusted
colleagues and mentors pushing in the opposite direction. The best
solution here is probably to allow for greater coordination of care in
multidisciplinary teams, where members keep each other up to date and
ensure alignment with the latest and most reliable evidence. Greater
devotion to tracking and achieving better outcomes could also help
change the calculus to constant evolution toward the best treatment
approach.
I hope this updated analysis will change the standard
for how we approach tiny unruptured aneurysms. More important, I hope it
is also a call to arms for creating a more responsive and responsible
system of care that reduces the outcome of our biases and more closely
adheres to the interests of the patient. Such a system would reward good
outcomes rather than more care, and would include direct, ongoing
feedback of outcomes to encourage integration of the best data elsewhere
and generate new data in the delivery of care. Given my own frustration
with the effect and pace of research, I am moving on to developing that
system. Wish me luck. I will need it.
Back to top
Article Information
Corresponding Author: S.
Claiborne Johnston, MD, PhD, Dell Medical School, University of Texas at
Austin, 1501 Red River, Stop Z0100, Austin, TX 78712 (clay.johnston@austin.utexas.edu).
Published Online: November 20, 2017. doi:10.1001/jamaneurol.2017.2559
Conflict of Interest Disclosures: None reported.
No comments:
Post a Comment