More of Less is More

Some folks at BU review evidence about a diagnostic test, called CT pulmonary angiograms. (I think you can only read the abstract, but I shall summarize well.)

A pulmonary embolism is a blockage in an artery in the lung. Sounds bad, right? And indeed it can be: symptoms can include respiratory distress, chest pain, and rapid heart beat and it can be fatal. The treatment is to take anticoagulant drugs, normally warfarin. This present a risk of bleeding but it beats dying. So we want to diagnose them all and make sure we don't miss any, right?

That's what doctors thought to and they were accordingly delighted when CT scanning came along. It is now possible to inject iodine into the blood stream and create very high quality images of the arterial network in the lungs, which means they can find even tiny embolisms. This has largely replaced an older method called ventilation-perfusion scanning, in which the patient inhales a radioactive gas and has a radioactive material injected, and a gamma ray camera can then image where air and blood are getting to within the lungs. Although that sounds scary, it actually subjects the person to a lower radiation dose than does CT scanning. However, it is less sensitive.

Doctors love the CT pulmonary angiography scan because they don't want to miss any of those "silent killers." But, what the BU researchers find is that since CTPA was introduced, the diagnosis of pulmonary embolism has increased by 80%, but deaths from pulmonary embolism have barely changed. It turns out that the procedure can detect tiny (isolated sub-segmental, referring to a small branch artery) embolisms, but apparently, these are harmless. In fact the body can reabsorb them, and it may well be that evolution has actually favored the lungs performing the function of filtering out tiny blood clots originating elsewhere. In one study, of 65 patients found to have these tiny embolisms who did not receive anticoagulants, none had any adverse effects.

As the authors summarize, anticoagulation is actually a leading cause of medication-related death. In one case series of patients given anticoagulants for isolated sub-segmental embolisms, 5.3% had major bleeding. Being told you have a condition that could kill you, and having to take a dangerous medication, is obviously unpleasant for people. And the cost? In 2006, hospital admission for a pulmonary embolism cost $44,000.

But the drug companies and medical device companies push these procedures hard. Hospitals need to have the latest scanners so they can appear cutting edge and high tech, but in order to justify paying for them, they have to use them. We're getting far too many of these scans, which increase the risk for cancer as well as costing money and finding all sorts of incidental abnormalities which aren't actually dangerous but have to be investigated, at great expense and angst. We should knock it off.