Skip navigation

DRE Screening for Prostate Cancer

Search
Results
Moreover
Cost
Comment

Screening for prostate cancer is a topic that raises blood pressure more than the intellectual level of debate. Some people want it because of the old argument that finding and treating cancer early must be good. Others bring us back to earth by reminding us that the issues come down to many more men dying with prostate cancer than from prostate cancer. The issue is pussycats and tigers - tigers being the aggressive cancers that we do perhaps need to catch early. Bandolier has visited this before (issues 26 , 37 and 43 ). But some newly published studies help clear the mind a bit.

If we are to screen, what population of men do we screen and what test do we use? A new meta-analysis [1] from Maastricht tells us how well digital rectal examination (DRE) does in unselected older men in primary care.

Search


MEDLINE and a specialist primary care database were searched electronically, and selected primary care journals hand searched without language restriction. Included studies had to compare DRE with biopsy or surgery as a reference diagnosis. Disease was usually excluded on the basis of a negative biopsy, and/or negative DRE combined with a negative PSA test or transrectal ultrasound. The study population had to be unselected with respect to prostate signs or symptoms.

Results


There were 14 studies with results on 21,821 men. The size of studies ranged between 315 and 6630 men. The lower age ranged between 45 and 60 years, and the upper age ranged between 70 and 90 years. The overall prevalence of biopsy or surgery confirmed prostate cancer was 3.7% (range 1.2% to 7.3%).

The paper provides information on a number of statistical outcomes. Thus in individual studies the likelihood ratio for a positive test ranged between 2.1 and 114. The negative predictive value was always above 95%.

Because Bandolier has problems with these statistical outcomes, the results of the studies are presented in a different way in the Figure. Studies were normalised to 1000 men, and the number calculated for:



Figure: Results of individual trials and weighted average (diamond) for DRE positive with cancer (True positive), DRE positive without cancer (False positive) and DRE negative with cancer (False negative)

So for every man detected by DRE and biopsy as truly having prostate cancer, four would be screened positive without actually having cancer, and one man with cancer would be missed.

Primary care group


These numbers can be applied to an average primary care group of 100,000 people. If screening with DRE were applied to all men over 45 years (about 17,000 of them), that would mean 359 cancers detected, 274 cancers missed, and 1454 men with positive DRE who would have to be biopsied to exclude them actually having cancer.

Moreover


Actually, it could be worse than this. Firstly we know that the usual sextant biopsy probably used in most of these studies has been shown itself to miss 35% of prostate cancers ( Bandolier 43 ). We also know that the PSA test, used to exclude cancer in a number of the studies, is normal in about 40% of men with cancer ( Bandolier 26 ). It is likely that the number of missed cancers is therefore higher than that suggested by the paper, perhaps missing as many cancers again.

If that were the case, screening for prostate cancer with DRE in primary care would mean that for every case of cancer found, two would be missed and four men unnecessarily sent for further investigation.

Cost


The implication of screening for prostate cancer on health service budgets is huge. An exercise to establish the actual and projected costs of screening with PSA in Canada [2], suggested that screening cost about £50 per man screened when the costs of treatment and the harms of treatment were added in.

Comment


Few of the studies discussed the method of digital rectal examination used or the sort of result from the digital rectal examination that constituted a positive test. The review considered a non-enlarged, smooth, symmetrical prostate with normal consistency to be normal, and test results were recalculated using this criterion where it was possible to do so. The review, concentrating on men without prostate problems, cannot speak about the usefulness of DRE for diagnosing prostate cancer in men with prostate symptoms.

Bandolier found the statistical outcomes unhelpful, which is why we calculated the results on the basis of actual numbers of men classified correctly or incorrectly, on the basis of 1000 men and a primary care group. This seems to be much easier to understand than sensitivity and specificity, or other ways of expressing results of diagnostic tests. As shown in Bandolier 61 , most doctors seem to agree with us that we need new ways of expressing and using test results.

The arguments for screening for prostate cancer are moot, and digital rectal examination alone would be unlikely to be used. When more sophisticated methods are evaluated, like PSA or newer laboratory tests, it will be interesting to see how many men are correctly and incorrectly diagnosed.

Reference:

  1. A Hoogendam, F Buntinx, HCW de Vet. The diagnostic value of digital rectal examination in primary care screening for prostate cancer: a meta-analysis. Family Practice 1999 16: 621-626.
  2. MD Krahn, A Coombs, IG Levy. Current and projected annual direct costs of screening asymptomatic men for prostate cancer using prostate-specific antigen. Canadian Medical Association Journal 1999 160: 49-57.
previous or next story in this issue