How Good are Colonoscopies?
Posted by medconsumers on January 1, 2009
Colonoscopy’s Benefit Found to be Overestimated
Regular visitors know that this Web site has never championed cancer screening tests. They are inevitably oversold to the public—benefits hyped, risks downplayed. Now colonoscopy has been exposed as a less than sure-fire lifesaving procedure. When colonoscopy screening first took off about nine years ago, people were often told by their doctors, “No one would die of colorectal cancer if every adult over age 50 had a colonoscopy.”
This inflated promise, based on no direct evidence, was downsized in time to: 90% of all colorectal cancers will be prevented. Now a new study, published this month in Annals of Internal Medicine, indicates that the true statistic may be nearer to 60%. It found that colonoscopy often misses polyps in the right, or ascending, colon where many colorectal cancers develop.
Why did it take so long to find this out and to be honest with the public? It’s not as if colonoscopy had ever been proven 90% effective or that disappointing information about colorectal cancer screening hasn’t been around for years. A 2000 San Francisco meeting of colorectal cancer researchers from all over the world is a case in point. Screening continually came up at this three-day conference as the most important way to reduce deaths. At the end of the conference, Dr. Wendy Atkin, a researcher from the UK, went up to the microphone to warn that there are only three randomized, controlled colon cancer screening trials worldwide, and all three showed screening is not lifesaving. True, a 30%reduction in colorectal cancer deaths was shown in the screened people, but this was offset by an equivalent increase in heart-related deaths.
As one of a handful of consumer advocates invited to attend this conference, I found her statement to be astounding and expected it to generate a heated discussion. Instead the waters folded over, and talk went on to other topics. I asked Dr. Atkin, as she was rushing to a taxi, whether researchers knew why people screened for colorectal cancer showed an increase in heart-related fatalities. She said they had no idea.
The first (and possibly only) doctor to share this disturbing bit of information with the public is H. Gilbert Welch, MD, professor at Dartmouth Medical School and author of the excellent 2004 book entitled, “Should I be Tested for Cancer? Maybe not and here’s why” (University of California Press). The three screening trials* involved the fecal occult blood test (FOBT). The participants received a colonoscopy only after the FOBT indicated the possibility of cancer. (Colonoscopy was not a screening procedure in the pre-1990s era when these studies were planned.) As for the question of why screening would cause an increase in cardiac deaths, Dr. Welch could only make a few guesses because no researchers have come up with, or perhaps ever looked for, a plausible explanation.
Dr. Welch’s book is not a rant against cancer screening tests. He takes a hard look at the supporting research and advocates realistic expectations. “Some believe that anyone who dies of cancer and wasn’t screened would have been saved had they had a test. But that’s not true. Sometimes cancers appear in the interval between screening tests. These interval cancers are growing rapidly and are more deadly than cancers detected by screening. Other times cancers are found by screening but people still die from the disease. Clearly, screening helps only in certain cases.” Dr. Welch also takes on the prevailing assumption—among doctors and consumers alike—that screening is a risk-free endeavor.
The new study that found colonoscopy misses many polyps did not randomly assign participants to have a colonoscopy or not. We’ll have to wait for that type of trial to see whether colonoscopy actually saves lives or whether colonoscopy screening merely trades one cause of death for another. Instead the new study drew information from the insurance claims for people living in the Canadian province of Ontario. It matched each of the 10,292 people who died of colon cancer with five people of the same age, sex and socioeconomic status who did not have colon cancer or any type of bowel disease.
The research team led by Nancy N. Baxter, MD, University of Toronto, checked how many had had colonoscopies and whether doctors had removed polyps. During the 1996 to 2001 study period, the colorectal cancer death rate was 7% among those who had ever had a colonoscopy and 9.8% among those who never had one. The decline in colon cancer deaths was “strongly associated” with the left side of the colon. And most of the missed cancers were located in the right (ascending) colon. Dr. Baxter and colleagues speculate that some of the colonoscopies might have been inadequately performed. About 30% were done by family doctors.
Doctor’s Skill Unknowable
This study comes on the heels of another one published last spring showing that many polyps are missed during colonoscopies that are performed too quickly. Unfortunately, people have no way of determining the skill of the doctor who will perform the colonoscopy. The American Cancer Society and the Centers for Disease Control and Prevention are reportedly working on developing quality standards for physicians to measure their performance and make improvements. The discussion about how to know whether you have a skilled colonoscopist who does many of these procedures is ongoing.
The renewed attention to colon cancer screening accuracy reminds me of an encounter with another researcher, this time at a 2004 conference. I recognized his name as a leading colorectal cancer researcher who has co-authored many important studies and is a strong proponent of screening. I asked him privately what we should make of that increase in heart-related deaths among people screened for colon cancer, half expecting him to deny any such finding. Instead, his response was surprisingly candid. “I guess we all have to die of something.”
Maryann Napoli, Center for Medical Consumers©
*Dr. Welch described the only one of the three trials to be conducted in the U.S.: the Minnesota Cancer Control Study—published in 1993 in the New England Journal of Medicine .