This issue of Chest (see page 139) contains an excellent contribution by Stein and coauthors to the area of the diagnosis of pulmonary embolism. Using data obtained from the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED), the authors refute a recent paper published in another journal.(1) Using questionable data and questionable epidemiologic methods, that article claimed that a normal P(A-a)[O.sub.2] tension difference could be used to rule out pulmonary embolism. In the present study in Chest, it is demonstrated conclusively that about 20% of patients with angiographically documented pulmonary embolism have a normal P(A-a)[O.sub.2] tension difference. Using a normal difference to exclude pulmonary embolism would lead to a large number of false-negatives. This in turn would result in failure to treat despite the existence of effective treatment. In turn, failure to treat would lead to preventable deaths plus short- and long-term complications, including chronic pulmonary hypertension.
Some interesting points arise in this connection. One issue is illustrated by a recent lawsuit in which a 55-year old patient undergoing ventilation-perfusion scanning suffered cardiopulmonary arrest after a (normal) perfusion scan and before a ventilation scan could be performed. The defense position was that the arrest resulted from a pulmonary embolus. Four experts agreed. In support, they testified that without a ventilation scan one could not eliminate a pulmonary embolus as a cause of the acute cardiopulmonary arrest!
It should be stated emphatically that a negative perfusion scan rules out pulmonary embolism with very high probability.(2)(3)(4) If the ventilation scan and perfusion scan are normal, pulmonary embolism is not diagnosable and is almost certainly not present. If the ventilation scan shows regional defects without perfusion abnormalities, there is no pulmonary embolism. If the ventilation scan is equivocal but the perfusion scan is normal, there is no pulmonary embolism.
Why, then, are ventilation scans performed in patients with normal perfusion scans? Presumably, they are performed for two reasons: (1) ignorance; and (2) unenlightened self-interest. In many imaging departments, there is an additional charge for ventilation scanning; the profits accure to the imager and the hospital.
In the interest of patients, ventilation scanning in pulmonary embolism suspects should be (as is now usually the case) postponed until after the perfusion scan. If the perfusion scan is normal, the ventilation scan is not needed.
If such a policy is unenforceable by education, perhaps contemplated health-care plans will decline payment for ventilation scans when the perfusion scan is normal. Perhaps these plans may consider the role of ventilation scanning for pulmonary embolism generally.
Another issue involves the question whether P(A-a)[O.sub.2] tension differences are in some way valuable so that they might be advisable, even if sometimes misleading, in some patients suspected of having pulmonary embolism. After all, physicians are sufficiently wise so as not to be misled.
There are two problems with this approach:
1. Some physicians are not wise enough.
2. Some laboratories collect an additional fee for calculating the P(A-a)[O.sub.2] and A-a difference. When this calculation is done to exclude the diagnosis of pulmonary embolism, we have, in cost/benefit terms, mathematical nonsense; the cost is measurable, the benefit is zero (or worse, negative, when the true diagnosis is "ruled out"), and the resulting "fraction" is either infinity or nonexistent.
Finally, we pose a highly embarrassing question. How did the original article(1) supporting the use of P(A-a)[O.sub.2] tension differences to exclude a diagnosis of pulmonary embolism ever get by the process of peer review and become published? The vagaries of peer review are too extensive to discuss in this limited editorial.
In any case, Dr. Stein and Master Henry are to be congratulated for a contribution wherein they ingeniously used the excellent database from the PIOPED study to evaluate the misleading nature of P(A-a)[O.sub.2] tension differences to exclude a diagnosis of pulmonary embolism.
One must ask, however, why in the diagnosis of pulmonary embolism have so many physicians made so many errors?
Recall the a-A C[O.sub.2] tension difference used to diagnose pulmonary embolism(5) introduced by the senior author of this editorial (EDR) and widely used for over a decade. After he discovered that the "Robin" test lacked adequate sensitivity and specificity, he found no easy way to undo the harm. The test had developed its own constituency who refused to accept the fact that the test was not accurate. The use of the test persisted even after the originator of the test debunked its value.
Consider the widespread uncritical acceptance of V scanning including modification by various probability systems. The practice continues despite at least two adequately performed clinical trials(2)(3) indicating the inaccuracy of such systems. Who remembers [131.sup.I] or [125.sup.I]-labeled firbrinogen to detect venous thrombosis,(6) and by so doing, diagnose pulmonary embolism?
And now we have the P(A-a)[O.sub.2] tension difference. We must learn that bad, even noninvasive tests, can kill or harm patients. When will we ever learn?
REFERENCES
(1)McFarlane MJ, Imperiale TF. Use of the alveolar-arterial oxygen gradient in the diagnosis of pulmonary embolism. Am J Med 1994; 96:57-62
(2)The PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism: results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). JAMA 1990; 263:2753-59
(3)Hull RD, Hirsh J, Carter CJ, et al. Pulmonary angiography, ventilation lung scanning and venography for clinically suspected pulmonary embolism with abnormal perfusion lung scan. Ann Intern Med 1983; 98:891-99
(4)Robin ED. Overdiagnosis and overtreatment of pulmonary embolism; the Emperor may have no clothes. Ann Intern Med 1977; 87:775-81
(5)Robin ED, Julian DG, Travis DM, et al. A physiologic approach to the diagnosis of acute pulmonary embolism. N Engl J Med 1959; 260:586-89
(6)Hobbs JT, Davis JWL. Detection of venous thrombosis with [131.sup.iodine-labeled] fibrinogen in the rabbit. Lancet 1960; 2:134-35
COPYRIGHT 1995 American College of Chest Physicians
COPYRIGHT 2004 Gale Group