I like to have a prejudice overthrown from time to time: it helps to persuade me that my other prejudices are reasonable because I am a reasonable man who is open to the evidence. This is especially the case where the prejudice is one that I do not really care much about. I can give it up without much regret.
A paper in a recent edition of the New England Journal of Medicine overthrew one such minor prejudice, namely that the more thoroughly a person was investigated for an occult cancer after suffering an unexpected deep vein thrombosis (DVT) or pulmonary embolus (PE), the more likely that one would be found.
The association between spontaneous DVT and cancer has long been known. Its discoverer was Armand Trousseau, a great French physician of the middle of the nineteenth century, who noticed that people who suffered DVTs often had cancers such as that of the pancreas, in those days always diagnosed at post mortem. By a strange and tragic coincidence, Trousseau himself suffered a DVT and a few months later was dead – of pancreatic cancer. This is a story from medical history that, once heard, is never forgotten.
It has been found that about 10 per cent of people have diagnosable cancer within a year of having had a DVT or PE. So it seems to stand to reason that if people who suffer such events are investigated up hill and down dale immediately afterwards, some cancers will be caught earlier and treated, and therefore survival will be increased.
Some Canadian researchers tested this hypothesis. They randomly divided 854 patients with either DVT or PE into two groups: those who were tested for cancer by simple methods such as physical examination, blood tests and chest x-ray, and those who, in addition to all these, had CT scans of the abdomen and pelvis. They were then followed up for a year to see whether there was any difference in outcome.
As is so often the case with what stands to reason, it did not stand. Overall mortality was 1.4 percent and 1.2 percent respectively in the two groups, cancer mortality alone being 1.4 percent and 0.9 percent. The time taken to diagnose cancer in the two groups were 4.2 and 4 months respectively. There was no difference between the groups in rate of the recurrence of DVT or PE. Four occult cancers were missed (or later developed) in patients who were not scanned, five in those who were. In other words, there was no difference between the groups, at least none that was statistically significant.
The overall rate of detection of cancers was surprisingly low by comparison with the generally accepted figures, 3.9 percent instead of 10 percent. This could have been either because of better detection or because of reduced incidence in the population of those cancers that present as DVT or PE.
Therefore the addition of scanning – faith in which has reached almost religious proportions – added nothing to simpler and less costly investigations. Moreover, it not only added to cost but added a hazard. It has been estimated that, with the amount of radiation received by the patients who were subjected to scanning, one in 460 women and one in 498 men will develop a cancer that they would not otherwise have developed. In other words, a screening procedure to detect cancer not only failed to do so, but added (very slightly, it is true) to the risk of developing a cancer later in life. One patient in this experiment who had a DVT or PE and was allocated to the CT group will develop a cancer as a result.