Accuracy of Screening for Fecal Occult Blood on a Single Stool Sample Obtained by Digital Rectal Examination

Abstract & Commentary

Synopsis: A single fecal occult blood test from a digital rectal exam specimen does not provide adequate screening for colorectal neoplasia and cannot replace standard at-home 6-sample fecal occult blood testing.

Source: Collins JF, et al. Ann Intern Med. 2005;142:2:81-85.

Many physicians routinely screen for advanced colonic neoplasia by occult blood testing of stool obtained from a digital rectal exam. Several limited studies have suggested that fecal occult blood testing (FOBT) on specimens obtained at rectal exam could produce similar positive predictive values for colon pathology. Collins and colleagues have performed the first large-scale trial to elucidate the sensitivity and specificity of digital FOBT vs standard 6-sample FOBT by performing colonoscopy on all patients with both positive and negative FOBT results. Also, 3121 asymptomatic patients in 13 VA medical centers participated out of 17,732 persons screened for study inclusion. As expected, virtually all participants were men with an average age of 63 years (50-75 years of age). All had digital rectal exams including FOBT on stool specimens so obtained, and all performed at-home collection of 2 smears from each of 3 spontaneously passed stools. Although patients were told to restrict red meat, vitamin C, and aspirin, compliance was not monitored. Because some patients did not collect their at-home FOBT specimens and because stool was not always available in the rectal vault for testing of the digitally obtained specimen, 2,665 patients were evaluable. Most participants were Caucasian; 1,218 patients (45.7%) had no polypoid lesions of any kind. In 438 patients, hyperplastic or other non-neoplastic lesions were identified. Of 1,656 patients without adenomas, digital FOBT was positive in 41 (specificity 97.5%; CI, 96.8-98.3%). At least one window of the Hemoccult II™ tests was positive in 101 of these patients (specificity 93.9%; CI, 92.7-95.1%). In 725 patients, 1 or more tubular adenomas less than 10mm in size were identified. Digital FOBT was positive in 4% and the 6-sample test was positive in 6.3%.

284 patients were found to have advanced neoplasia (10.7% of the whole population included). Results of digital FOBT were positive in 14 of these patients vs 68 testing positive with the 6-sample FOBT. The sensitivity of the 6-sample test for advanced neoplasia was 23.9% vs only 4.9% for digital FOBT. Specificity for advanced neoplasia was 93.8% vs 97.1%. There was no evidence that adding the digital FOBT to the 6-sample home technique would identify significant additional pathology.

Collins et al admitted that a positive digital FOBT is significant and that it should mandate colonoscopic evaluation. Importantly, their major conclusion is that digital FOBT is not adequate as screening for advanced colon neoplasia.

Comment by Malcolm Robinson MD, FACP, FACG

As this paper points out, there are limitations to its findings. First, since virtually no women were included in this VA population, results are not necessarily generalizable to women. Second, this study used rehydration of the Hemoccult II slides, a technique that may increase sensitivity while decreasing specificity (no longer recommended for FOBT). It is also possible (even likely) that some neoplastic lesions were missed although these were highly experienced colonoscopists. In a somewhat cute editorial in the same issue of Annals of Internal Medicine, physicians are chided for their ignorance regarding appropriate colorectal cancer screening, and they are urged to put their Hemoccult slides in a locked drawer, labeled: For Emergency Use Only. However, I am not certain that the dismissal of digital FOBT is altogether reasonable. Many of our patients refuse to prepare at home slides for FOBT, and tests that are not performed will certainly do no good. Also, it is hard to be impressed with the highly unsatisfactory results of our gold standard test of 6-sample at-home FOBT. After all, even done properly, this test misses more than three quarters of advanced colon neoplasia. For my money, we need better tests. In a recent article, fecal DNA testing was discussed. Although not yet ready for widespread adoption, this seems to me to be a more promising direction than any approach that depends on occult blood detection. After all, many advanced neoplastic lesions don’t bleed at all. However, all of them must be shedding cells. We need to identify the signature (genetic or otherwise) of colon neoplasia as a means of directing colonoscopic intervention. For now, I don’t agree that we should abandon digital rectal exams including FOBT. An inadequate test is better than none at all. On the other hand, I agree with Collins et al that we should not pretend that a negative digital FOBT rules out significant colon neoplasia—as it certainly does not. We should redouble our efforts to achieve universal adherence to at-home FOBT in the pertinent patient populations, and we should try even harder to come up with a test that really works.

Dr. Robinson, Emeritus Clinical Professor of Medicine, University of Oklahoma College of Medicine Oklahoma City, OK, is Associate Editor of Internal Medicine Alert.