High-Quality Screening Colonoscopy Priority for GI Docs

SAN FRANCISCO—Current evidence suggests that colonoscopy screening can reduce mortality from colorectal cancer—if these procedures are performed in a high-quality manner, David A. Lieberman, MD, said at the 2006 Gastrointestinal Cancers Symposium.

"The rates of colonoscopy in the United States have been skyrocketing over the last 5 years, to the extent that some have suggested that we have inadequate resources to be able to perform all the colonoscopy that needs to be done," observed Dr. Lieberman, head of the Division of Gastroenterology, Oregon Health & Science University, Portland. However, he said, evidence that this strategy reduces colorectal cancer mortality is indirect at this time and mainly derived from clinical trials.

Indirect Support

The indirect support of a lifesaving benefit of colonoscopy screening comes from several sources, Dr. Lieberman said. Four trials have assessed fecal occult blood testing (FOBT) as primary screening for colorectal cancer, finding mortality was lower among screened participants overall (reduction, 15% to 33%) and more so among those who were adherent (reduction, 33% to 39%). One trial also found a reduced incidence of cancer (Mandel et al: N Engl J Med 343:1603-1607, 2000). "The identification of early-stage cancers was beneficial, but that was done with colonoscopy. The incidence reduction, which was attributed by the authors to detection and removal of adenomas, was also due to colonoscopy," he pointed out.

Sigmoidoscopy studies also lend indirect support to a lifesaving benefit of colonoscopy screening, he said. Results of case-control studies suggest that mortality from cancers in the part of the colon examined is reduced by 60%. Moreover, VA Cooperative Study 380 found that among men, the rate of detection of advanced adenomas was 70% if the colon was examined to its sigmoid-descending part, but 80% if the colon was examined to the splenic flexure (Lieberman et al: N Engl J Med 343:162-168, 2000). "The implication from these studies is that an endoscopic evaluation of the colon with detection and removal of polyps could reduce mortality," he said. "And if you examine more of the colon, theoretically, that could result in higher . . . mortality reduction."

According to Dr. Lieberman, more indirect support comes from the National Polyp Study, in which patients with adenomas underwent colonoscopy with polypectomy and were followed for almost 6 years (Winawer et al: N Engl J Med 329:1977-1981, 1993). During follow-up, the actual number of interval cancers was only about one-tenth of the expected number. In addition, a case-control study found that patients with colorectal cancer were only about half as likely as their healthy counterparts to have undergone colonoscopy (Muller et al: Ann Intern Med 123:904-910, 1995).

Compromising Factors

Despite this evidence suggesting benefit, a variety of factors may make colonoscopy screening less effective than thought, according to Dr. Lieberman. "We have known for a long time that colonoscopy is not perfect," he commented. For example, he said, chemoprevention studies in patients with adenomas who underwent colonoscopy with complete polypectomy have found a substantial rate of interval cancers 3 years later (1.7 to 2.4 per 1,000 person-years). These interval cancers may arise in several ways, he said.

First, 5% to 7% of colonoscopies performed in the United States are not complete to the cecum, suggesting some lesions are missed because of incomplete exams (Bowles et al: Gut 53:277-283, 2004). Second, he noted, in 2% to 12% of patients with adenomas measuring 1 cm or greater in diameter detected by CT colonography, the adenomas are not detected by colonoscopy, suggesting that other lesions are missed because they are overlooked.

Third, a trial of polyp prevention found that about a third of cancers occurring in the 3 years after a colonoscopy occurred at a site where a polyp had been previously removed, suggesting incomplete removal (Pabby et al: Gastrointest Endosc 61:385-391, 2005). Rates of missed lesions and of incomplete removal in routine clinical practice are unknown, Dr. Lieberman noted.

A fourth possible explanation for interval cancers is that polyps grow faster than thought. "We know that about 15% or so of sporadic cancers have microsatellite instability," he observed, and these appear to grow more rapidly than others.

Factors other than technical and biological ones may also compromise the effectiveness of colonoscopy screening, according to Dr. Lieberman. "Remember that any screening test is ultimately going to lead to colonoscopy if it is positive," he said. "That leads to surveillance, and if surveillance is not appropriate, that also can affect the effectiveness of a colonoscopy program."

Surveillance may be done too frequently, exposing patients to excess risk, he noted, and evidence suggests that in the United States, patients with hyperplastic polyps or adenomas measuring less than 1 cm are often advised to have colonoscopy more frequently than guidelines recommend (Mysliwiec et al: Ann Intern Med 141:264-271, 2004). Alternatively, surveillance may be done too infrequently, failing to catch high-risk and fast-growing lesions.

Yet another factor that influences the overall outcome of a colonoscopy screening program is the risk of complications, Dr. Lieberman said. In the VA Cooperative Study, 0.3% of patients had a major complication attributable to the procedure, with a higher rate observed in patients having therapeutic procedures than in those having diagnostic ones (Nelson et al: Gastrointest Endosc 55:307-314, 2002). In addition, about 16% of patients had minor complications, such as vasovagal events and transient oxygen desaturation. "Although these seem like minor complications, in an older patient with heart disease or lung disease, they might be somewhat more significant," Dr. Lieberman observed.

Cumulative Risk

The cumulative risk of complications is also unknown. "So if patients are getting multiple exams, does their risk accrue?" he asked. In one study, the estimated cumulative risk of serious complications over the next 30 years for a patient aged 50 entering a colonoscopy screening program was 1% with 3 exams spaced 10 years apart, 2% with 6 exams spaced 5 years apart, and 3% with 10 exams spaced 3 years apart (Ransohoff et al: Gastroenterology 128:1685-1695, 2005). The corresponding cumulative risks of death were 0.1%, 0.2%, and 0.3%. "Now this is simply a guess or an estimate of what could happen, but I think it gives us pause about the potential risks associated with colonoscopy," he said.

A final consideration in assessing the benefit of screening colonoscopy is its actual effect on mortality, Dr. Lieberman asserted. He noted that, theoretically, screening could identify a colorectal cancer, thereby succeeding in preventing death from that cancer, but still not prolong life.

Similarly, screening could identify adenomas that are removed, thereby succeeding in reducing the incidence of colon cancer, but still not prolong life. "I think these are questions that remain to be answered about whether or not colonoscopy is going to make a difference," he stated. "We would like to think that it will, but I think that at this point in time, there is still some uncertainty."

Effectiveness Models

Although there has not been an ideal study in clinical practice addressing the effectiveness of colonoscopy screening, models can provide some information regarding its impact on outcomes, costs, and resource use, Dr. Lieberman said. He discussed a recent analysis that estimated the cost-effectiveness of five different screening approaches supported by the US Preventive Services Task Force—FOBT, sigmoidoscopy, FOBT and sigmoidoscopy combined, double-contrast barium enema, and colonoscopy (Institute of Medicine, 2005). "The good news about this is that they are all within a range of what we could consider cost-effective in terms of the cost per added year of life," he said, with values ranging from $5,000 to $25,000. The cost per life-year gained was lowest with FOBT, while the rate of cancer prevention was highest with colonoscopy, he noted.

"I conclude that colonoscopy screening can reduce colorectal cancer mortality, but it depends on a high-quality examination to the cecum, a low rate of missed lesions, a low rate of incompletely removed lesions, and a low rate of adverse events," Dr. Lieberman said. "This means making sure we perform these exams with the highest quality. I think that has to be one of the highest priorities for the GI community today, to focus on how we can improve the quality of performing colonoscopy."