By Charles S. Abrams, MD
2008-05-01
Dr. Abrams indicated no relevant conflicts of interest.
Anderson DR, Kahn SR, Rodger MA, et al. Computed
tomographic pulmonary angiography vs. ventilation-perfusion lung
scanning in patients with suspected pulmonary embolism: a randomized
controlled trial. JAMA. 2007;298:2743-53.
For several decades, ventilation-perfusion (V/Q) scans have been
used to detect pulmonary embolisms in patients. It is clear that a
normal V/Q scan excludes the diagnosis of a pulmonary embolism, and a
high probability V/Q scan essentially confirms the diagnosis. The
challenge is what to do with patients who have a non-diagnostic (i.e.,
low to intermediate probability) V/Q scan.
Computed tomographic pulmonary angiography (CTPA) is an alternative
test used to detect pulmonary embolisms that has several advantages
over the V/Q scan. It provides a clear (yes or no) result, and it is
also capable of detecting non-thrombotic causes for a patient's
symptoms. The downsides of CTPA are that this technique exposes the
patient to more radiation than V/Q scans and that it may cause
dye-induced nephrotoxicity. In spite of these tradeoffs, CTPA has
largely replaced V/Q testing for pulmonary embolic disease. Although
CTPA has become widely accepted as a standard test for this disease,
there is controversy over whether small-vessel disease can be imaged as
well by CTPA as by V/Q scans.
In this single-blinded, noninferiority, multicenter clinical trial,
701 patients were randomized to undergo CTPA and 716 were randomized to
V/Q scanning. The results showed that 19.2 percent of the CTPA group
and 14.2 percent of the V/Q group were diagnosed with a pulmonary
embolism and treated with anticoagulation. Of the patients who were not
initially diagnosed with a pulmonary embolism, 0.4 percent of the
patients in the CTPA group and 1.0 percent of the patients in the V/Q
scan were subsequently diagnosed to have thromboembolic disease.
This study by Anderson and colleagues convincingly
demonstrates that CTPA is at least as effective as V/Q scans at
diagnosing pulmonary embolisms. In fact, the incidence of detected
pulmonary embolisms was 30 percent higher in the group randomized to be
analyzed by CTPA ― a statistically significant difference. Some
patients who had negative V/Q scans were subsequently diagnosed as
having a pulmonary embolism once they had a CTPA; this suggests that
CTPA might be an even more sensitive diagnostic test than V/Q scans.
These findings raise the interesting question about
whether our technology is getting too good. Is CTPA capable of
diagnosing emboli that are too small to be clinically significant? In
the Anderson study, approximately 7 percent of patients diagnosed as
having a thromboembolism by CTPA had emboli only in small subsegmental
pulmonary vessels. Do we really know that these small clots require
therapy? It should be noted that eight patients in the Anderson trial
who had no obvious pulmonary embolism diagnosed by either CTPA or V/Q
scans ultimately developed a pulmonary embolism (including one patient
who died of a clot). This implies that these eight symptomatic patients
probably had very small pulmonary emboli that were missed by their
initial imaging study. We need to keep in mind that the true value of
treating a pulmonary embolism with anticoagulation is to prevent the
next one, rather than the thromboembolism that has already occurred.
Using this rationale and the currently available data, we should be
treating all pulmonary emboli regardless of their size, and perhaps
even seeking out technologies that diagnose smaller and smaller clots.
back to top
