Some tests of little use for cancer in colon

Oregon takes part in a national study that finds stool samples taken at doctors' offices lead to faulty, possibly fatal results

(reprinted with permission from the Oregonian, front page, January 18, 2005)

by Andy Dworkin

A common screening test for colon cancer is practically meaningless and should not be used, according to new studies by researchers in Oregon and elsewhere.

Tests for blood in stool samples that doctors gather during routine rectal exams are inaccurate, may give patients false hope of being cancer-free and might result in preventable deaths. Researchers said doctors should switch to better ways of checking for the nation's second-deadliest cancer.

"Colon cancer screening does save lives. But it has to be done properly," said Dr. David Lieberman, one of the researchers and chief of gastroenterology at Portland's Veterans Affairs Medical Center. "What we showed in our study is this is an inadequate test and just about worthless."

Lieberman's group checked the accuracy of various colon-cancer screening tools by giving 2,665 volunteers a digital rectal exam followed by a fecal occult blood test. That uses a chemical-impregnated card to find blood hidden in stool samples. The samples can be gathered at home by patients or during digital exams, which doctors often give during physicals or routine gynecologic checkups.

Each participant also got a colonoscopy, which found cancer or advanced precancerous growths in 284 of the volunteers.

Of those 284 people, only 14 had blood show up in office stool sample tests, a 4.9 percent rate. Tests that checked multiple samples patients gathered at home, on three successive days, found blood in 68 of the cases, a 23.9 percent rate. That result means the take-home tests could be useful, but not the one-sample test, Lieberman said.

The in-office test fell short by many measures: Looking just at cancers, not precancerous polyps, the in-office test found 9.5 percent of cases seen on colonoscopy versus 42.9 percent for take-home tests. And a negative result on the in-office test basically doesn't change a patient's odds of actually having a worrisome growth; negative results on the take-home test lower those odds by a small, but significant, amount.

The study involved mostly male patients at 13 Veterans Affairs hospitals, including Portland's. Results appear in today's Annals of Internal Medicine along with a survey that showed nearly one-third of doctors use the useless in-office stool screening test.

"The two articles paint a disappointing picture of over-reliance on an inaccurate test," the journal's editor, Dr. Harold Sox, wrote in an editorial. He said the discredited test's popularity may explain why U.S. colon cancer deaths are falling at a slower rate than scientists expected.

Among cancers, only lung cancer kills more U.S. residents than colon tumors, the American Cancer Society says. In 2004, the group estimates that 146,940 U.S. residents were diagnosed with colorectal cancer and 56,730 died of the disease. That includes 1,790 new cases and 690 deaths in Oregon, and 2,720 new cases and 1,050 deaths in Washington.

Colon cancer death rates have dropped over the years, probably because of factors including better health habits and screening tests. Those tests can find not only cancers but also precancerous polyps, which doctors can remove during a colonoscopy. So screening can prevent cancer, as well as detect existing cases early for better treatment.

National health groups say that, starting at age 50, people should choose any of several methods of colon-cancer screening: Fecal occult blood tests on samples patients gather at home on three successive days. Using several samples seems more accurate than one in-office sample. Long-term studies suggest these tests, done yearly, can cut death rates by up to one-third. Sigmoidoscopy, which uses a small tube and scope to examine the lower one-third of the colon. Experts recommend repeating this test every five years. An X-ray exam, including a barium enema, every five years. Colonoscopy, which uses a scope to study the whole large intestine, repeated every 10 years.

Lieberman said those tests are not perfect but can find many suspicious growths when repeated at the proper intervals. When colon cancer is diagnosed early, more than 90 percent of treated patients survive for at least five years. But statistics show less than one-half of U.S. adults get the recommended screening.

Although many tests check for suspicious growths, only sigmoidoscopy and colonoscopy can take samples to check for cancer. Doctors say a full colonoscopy should follow any suspicious findings on another screening test.

But the survey found bad follow-up habits are common: Almost 30 percent of doctors followed a positive in-office fecal blood test with a repeat blood test. Repeating a bad test is more than useless: It could be dangerous if doctors see no blood on a second test and decide against a colonoscopy.

The survey, led by federal researchers, didn't ask doctors why they used the in-office tests. Lieberman said doctors may screen a sample they gather in a rectal exam because they worry patients will not collect and return home stool samples or complete other screening tests.

While colonoscopy finds many growths, it is more expensive, risky and invasive than other screening tests. Ideally, Lieberman said, doctors could use a simpler way to identify high-risk patients and refer only them for colonoscopy.

He added that certain people at higher risk, such as those with a family history of colon cancer, should be tested before age 50.

© 2005, Oregonian Publishing Co. All rights reserved. Used with Permission of The Oregonian.