HEMOCCULT II, HEMOQUANT 26% SENSITIVITY FOR COLORECTAL CANCER
This article was originally published in The Gray Sheet
HEMOCCULT II, HEMOQUANT 26% SENSITIVITY FOR COLORECTAL CANCER detection is found for both tests in a prospective multicenter study by David Ahlquist, MD, Division of Gastroenterology, Mayo Clinic, et al. published in the March 10 Journal of the American Medical Association. The low test sensitivity rate observed in the study indicates, Alquist et al. conclude, that fecal blood "appears to be a poor marker for colorectal neoplasia." The authors add that, using fecal occult blood screening, "most cancers and the vast majority of polyps will be missed." The study was undertaken to "define the validity of fecal blood," measured by the Hemoccult II and HemoQuant fecal occult blood tests, as a screening marker for colorectal neoplasia. The study also sought to "compare the diagnostic yields of the two tests at matched specificity." The Hemoccult II test is manufactured by San Jose, California-based SmithKline Diagnostics; the HemoQuant is an in-house test performed by Mayo Medical Laboratories and SmithKline Beecham's clinical lab division. Two patient groups with "anticipated higher than average" prevalence of colorectal neoplasia were used in the study. The groups included patients who had already undergone resection of colorectal cancer, identified from a large database, and their first-degree relatives. Stool samples from the group of 1,217 postresection patients were collected and tested each year prior to routine annual endoscopic or radiographic structural surveillance evaluations. The evaluations were done for three years following surgery. A total of 12,312 relatives participated in the study by returning stool samples; only those with positive FOBT tests underwent colorectal evaluations. The samples were tested with both FOBTs in different Mayo Clinic laboratories within 48 hours of receipt, and within five to 21 days of collection (seven day mean). In the postresection group, 2,946 sample kits were returned overall; colorectal evaluations followed in 2,293 instances. Out of the 2,293 evaluations, neither cancer nor polyps were found in 1,861 cases. However, in 87 of these cases one of the three stools tested positive using the Hemoccult test (4.7%), yielding a specificity of 95.3%. Forty-six malignant colorectal neoplasms and 402 polyps were identified among the 2,293 evaluations, including some overlap. Hemoccult screens were positive in 12 of the 46 colorectal cancers found, for a sensitivity rate of 26%. The sensitivity of HemoQuant (which differs from Hemoccult II in that it is a quantitative rather than qualitative test), varied inversely with the specificity and cutoff level of hemoglobin per gram of stool. At a specificity matched with Hemoccult's (95.3% at 2.1 mg of hemoglobin per gram of stool), 12 of the 46 cancers were found, again yielding a 26% sensitivity rate. For the group of relatives, exact test sensitivities could not be ascertained since colorectal evaluations on screen-negative patients could not be obtained. However, colorectal evaluations done on subjects with positive stool screens with either Hemoccult II or HemoQuant revealed 42 colorectal cancers. Overall, screens were Hemoccult-positive in 473 subjects or 3.8% of the total; the same percentage had HemoQuant levels higher than 2.0 mg of hemoglobin per gram of stool. Estimated sensitivities for the relatives' group at one-to- three years of follow-up was 25% to 33% for Hemoccult II and 29% to 43% for HemoQuant, numbers "not significantly different." The estimates were made "by factoring into the estimation an approximation of missed cancers based on close follow-up by telephone inquiries." Ahlquist et al. conclude that the low test sensitivities observed in the study "were sobering" in light of the "assumption" that fetal blood tests "reliably detect asymptomatic colorectal neoplasia." A concern was that "the low sensitivities in this study more accurately represent fecal blood test performance for the screen detection of colorectal neoplasia than do the much higher detection rates reported from selected groups of symptomatic referred patients." SmithKline Diagnostics, manufacturer of the Hemoccult II test, issued a press release March 10 disagreeing with the study's findings. The firm says that "samples were not collected and stored according to SmithKline Diagnostics' recommendations" and thus abnormally low sensitivity for cancer was observed due to "degradation of hemoglobin" in the stool. Previous studies using the Hemoccult test "led to the detection of colorectal cancer in 50 to 75 percent of the cases," according to the firm. The Ahlquist study indicates however that "it is unlikely that the method of pretest fecal storage accounted for the low Hemoccult sensitivity in this study, as HemoQuant, a test less affected by storage artifact, yielded similarly low sensitivity for cancer."
Sign in to continue reading.
New to Medtech Insight?
Start a free trial today!
Register for our free email digests: