Colon Cancer Screening in the Elderly: When Do We Stop?

Abstract

Colon cancer is a major cause of cancer death in the US. Screening studies can prevent colon cancer and are recommended for average risk persons beginning at 50 years of age. Compliance with these recommendations has increased, particularly among those over 65 years old who have an increased prevalence of cancer and for whom screening is covered by Medicare. However, the efficacy of screening (and surveillance) in this population has not been well studied. The effect of co-morbidities and potentially increased risk of procedures performed on older adults are important concerns. This review addresses the benefits and harms of screening in the elderly. Physician emphasis on health status, life expectancy and patient preferences is critical in decision-making regarding colon cancer screening in this patient population. The recent update in the recommendations of the US Preventive Services Task Force (USPSTF) are reviewed.

Colorectal cancer is a leading cause of cancer death in the United States (1, 2). In 2005, over 140,000 new cases were diagnosed with over 47,000 colorectal cancer deaths. The annual incidence of colon cancer is decreasing slowly, likely in part due to our screening effort and possibly to related lifestyle changes. Numerous governmental, professional and lay organizations have organized campaigns stressing the need to initiate a screening program at age 50 years for men and women at average risk of colon cancer and at younger ages for higher risk individuals (3–5).

At this point, the US population is aging. There are more than 35 million people over the age of 65. The population of very elderly, those persons over 85 years, is approximately 5 million. Preventive care in this population has only been addressed by a limited number of guidelines. The small but growing number of geriatricians has been instrumental in highlighting the screening concerns in this population. Nonetheless, there is little research in the area of preventive health in the elderly (6).

The centrality of screening for colon cancer prevention is based on the fact that colon cancer develops over several years. There is a transition from normal colonic mucosa via adenomatous polyp formation, with consequent progression to colon cancer. Data from observational studies estimate this time span to be 10 to 15 years, although it may be much more rapid in some hereditary polyposis syndromes. Some individuals may go directly from normal appearing colonic mucosa to cancer. This type of transition had traditionally been associated with certain hereditary/familial syndromes. However, it can be seen in others with so-called “flat lesions”. However, the typical time progression allows for early detection and intervention through screening programs. Adenomas or premalignant lesions are completely curable via endoscopic removal. However, even those patients diagnosed with early asymptomatic disease are more likely to have localized or at least regional cancer which has an excellent five-year survival rate (>90%) (3).

Colon cancer screening programs and modalities are numerous. Recently this topic was extensively reviewed with attention to applicability in screening and surveillance in the US population (3, 7). Although screening and surveillance are associated with a decrease in mortality by the removal of adenomatous polyps and the early detection of invasive disease, the majority of the eligible population is not being appropriately screened. Options for screening range from stool tests to detect blood loss to structural studies that detect both premaligant and invasive lesions. The joint guidelines from the American Cancer Society, U.S. Multisociety Task Force on Colorectal Cancer and the American Radiological Society made the following recommendations regarding the screening of average risk men and women beginning at age 50: flexible sigmoidoscopy to 40 cm or splenic flexure every 5 years, colonoscopy every 10 years, double contrast barium enema every 5 years, and computer tomography colonography every 5 years. These studies are encouraged because colon cancer prevention is the goal of screening. Studies revealing lesions will require subsequent colonoscopy for sampling and/or tissue removal. Stool studies are primarily designed to detect colon cancer. These include guaiac-based fecal occult blood testing, the fecal immunochemical test, and stool DNA tests. Importantly, the report emphasizes that regular performance of screening studies with a clear understanding by practitioners and patients of the requirements for preparation and follow up is critical to maximize screening efficacy. The importance of study quality was stressed, with concomitant realization of the limitation and complications of procedures. The authors note that advanced age and increased incidence of diverticular disease increase the complication rate for colonoscopy. However, they do not directly address screening in the elderly.

The joint guidelines also review screening and surveillance for colorectal cancer in increased-risk and high-risk persons, with accompanying recommendations. Increased-risk persons include those individuals with a personal history of adenomatous polyps or colorectal cancer and patients with a family history of colon neoplasias (3). Patients with documented or suspected familial adenomatous polyposis, hereditary nonpolyposis colon cancer and chronic inflammatory colitis (pancolitis and left-sided colitis) are classified as high-risk. The recommendations for screening this population are specific, but beyond the scope this report. Importantly, no targeted recommendations are made for the elderly.

Cancer screening in the elderly (those greater than 75 years of age) has been controversial and is not specifically addressed by most guidelines (6, 8). Elderly patients have not been included in randomized trials of the efficacy of screening studies. Recommendations for screening guidelines have been made based upon detection rates, but do not address the impact of co-morbidities, functional status, and life expectancy. These issues are more important considerations in the elderly. Walter (6) and other geriatricians have stressed the need to individualize screening (and surveillance) in the care of the elderly. They propose that these factors be weighed in the development of guidelines with upper age limits for screening. The basic factors that should be considered are life expectancy, risk of death from cancer and screening outcomes. A clear estimate of the benefits of the screening program weighed against the harm of the procedures is critical in decision-making. Often, it has been difficult to obtain an estimate of how these factors change with age. I will review these factors as they apply to colorectal cancer screening.

The central focus of screening is the increased risk of colon cancer in the elderly. The lifetime risk of colorectal cancer is ∼5% with the incidence in the population over 75 years at about 40 to 50 per 100,000 persons. This is compared to an incidence of 15 to 20 per 100,000 in persons 60–65 years. Another important factor and potential benefit of screening for colon cancer (especially with colonoscopy) is the increase incidence of right-sided colon neoplasias (9, 10). In the US and other countries, the incidence of right-sided advanced adenomas and cancers increases with age, increasing from 15% of lesions in patients at 50 years of age to over 35% in patients over age 75 (1). Right-sided lesions present symptomatically at advanced stages so early detection is important.

General estimates of the benefit of colon cancer screening are based on a low side-effect profile for the screening procedures themselves. Effective colorectal screening is ultimately dependent upon an acceptable complication rate associated with colonoscopy, which is required to remove adenomas and early lesions. As indicated above, most randomized trials have not included the elderly. Thus, estimates of the harm in this population have not been widely studied. Recent reviews (11, 12) have pointed to some of the concerns that should be entertained. First, preparation with electrolyte solutions and polyethylene glycol may be of increased concern given the increased prevalence of renal and cardiac dysfunction in the elderly. Although studies have not documented an increased prevalence of fluid and electrolyte imbalances, these complications are generally not reported but are of concern. The potential for falls given the frequent toileting associated with purgation is also a possibility. Overall, preparation is one of the most bothersome aspects of colonoscopy (13). In addition, sedation in the elderly is associated with increased sensitivity and sedation risk due to a number of physiologic changes with age (11, 14, 15). The elderly are subject to greater central nervous system depression and resultant respiratory compromise.

Benzodiazepines (midazolam being most commonly used) are associated with transient apnea in the elderly, who need to be closely monitored. Aspiration is another important concern given the age-related delays in reflexive glottis responses. The combination of increased volume of distribution of these agents and altered clearance may be associated with prolonged clearance of sedative agents. Thus, recovery from sedation may be delayed, affecting discharge time and post-procedure mobility.

Perforation is one of the major complications of colonoscopy, and in some studies, risk appears to be increased in the elderly (11, 12, 16). They have an increased likelihood of significant fixation and narrowing of the sigmoid colon as a result of diverticular disease and/or prior pelvic surgery. These procedural limitations are frequently encountered in elderly women. In one study, the risk of perforation was significantly greater with age, with increases of 1% per year of age (16). Co-morbidities were associated with a more than 50% increase in the rate of perforation (OR 1.52, P < 0.007). Bleeding is another complication associated with therapeutic colonoscopy. Although risk of hemorrhage itself is not increased in the elderly, tolerance of hemodynamic insults is decreased (8, 17).

The psychological stress of screening for colon cancer should be strongly considered. This may be particularly important in persons who have cognitive compromise. Parker et al. (18) noted high anxiety levels in patients who have positive fecal blood testing, with anxiety levels returning to normal one month after a negative test. However, given the high percentage of false positives associated with fecal occult blood testing, a number of patients may live through the “temporary experience of the diagnosis of cancer” while waiting for confirmatory studies and these fears may be difficult to dispel in some patients (6).

Life expectancy and the risk of dying from colon cancer are the most important components in the development of guidelines for screening in the elderly, which must be balanced with the benefits and risks of the screening procedures themselves.

Life expectancy decreases with advancing age, but not uniformly. Rather, life expectancy is highly dependent on co-morbidities (19–21). The impact of co-morbidity can be quantified by several processes, from estimates of health status and frailty to more quantitative approaches (22). The Charlson (-Deyo) co-morbidity index is one mechanism that quantifies impacts as follows: a score of 0 denotes the best health, with scores of ≥4 indicative of severe co-morbidity. A value of 1 is assigned to conditions such as congestive heart failure, cerebrovascular disease, dementia or mild liver disease, whereas diabetes with complications, hemiplegia and moderate to severe renal disease are assigned a value of 2. Metastatic solid tumors and AIDS have a value of 6 (23). Patients with higher Charlson scores were observed to have decreased five year survival (1, 19, 24). Using this index and a general assessment of health status, patients can be categorized as being in good, average or poor health. Life expectancy can thus be reported from the sickest to the healthiest quartiles (lower 25%, middle 50%, and upper 25%). As noted in Figure 1 (adapted from Walter et al (1)) depicting the life expectancy of women in the U.S., women of age 75 years have a life expectancy of 6.8 years in the sickest patients with high co-morbidity to 17 years in the healthiest quartile. In the 90 year old group, the range is from 2 years to 7 years. Similar data are presented for men (1). Therefore, the approach to patients of the same age might be significantly different given the patient's health status, which complicates the development of screening recommendations for the elderly.

Fig. 1.

Life Expectancy of Women in US.

The risk of dying from colon cancer is another important factor in the decision-making process for screening elderly patients. Given advancing age and decreased life expectancy with a concomitant increase in the risk of colon cancer, the impact of screening is not easily predictable. Ko and Sonnenberg (8) determined that the risk of dying from colon cancer remained stable in the healthiest group of elderly persons. Figure 2 is adapted from their data. Patients in poor health (lowest 25^th percentile) were progressively less likely to die from colon cancer, and therefore would not benefit from screening. The authors evaluated the number needed to screen (NNS) with the various screening regimes. Small numbers are favored in this evaluation. The healthiest patients (upper 25^th percentiles) in the elderly age groups (70–85 years) had NNS not dissimilar from that of the 50–54 years group. Patients with impaired health and presumed co-morbidities have high NNS, reflecting death from other causes and a low likelihood of benefit from the early diagnosis of colon cancer (20).

Fig. 2.

Risk of Dying of Colon Cancer Vs Age and Health Status.

Have these factors been used by practitioners in their screening for colon cancer in the elderly? Walter et al (23) reported the rates of screening in Veterans' Administration patients. They found that in 2001–2002, 46% of elderly patients were screened for colon cancer. However, only 47% of patients with no co-morbidities (and thus a much greater five-year survival rate) were screened. Indeed this screening rate did not differ significantly from the 41% of elderly patients with severe co-morbidities and life expectancies much less than five years. They found that the best predictor of screening was the patient's number of clinic visits. Thus, patients with severe co-morbidity and more than four visits were just as likely to be screened as healthy patients with no co-morbidities. There was a decrease in the rate of screening with age. However, patients who would benefit from screening, healthy elderly patients with no co-morbidities were not screened at a rate commensurate with their health status and colon cancer risk. An earlier report by Fisher al. (21) had similar findings and urged that resources be “directed toward screening the patients with the most life-years at stake”.

The US Preventive Health Task Force (2) recently published recommendations for screening in the elderly as an update to the 2002 guidelines (25–27). This update reconfirms the efficacy of screening techniques, including fecal occult blood testing, flexible sigmoidoscopy and colonoscopy, in the prevention and early detection of colorectal cancer. Colorectal cancer screening starting at age 50 for average risk persons is a cost-effective program. The task force did not recommend fecal DNA tests or Computed Tomography Colonography (CTC). The latter was not supported due to the radiation exposure, decreased availability, need for preparation and cost. In addressing screening in the elderly (27), they recommended that in patients over 75 years who have been previously screened on a regular basis, no further screening should be offered. In patients 75–85 years, who have not been screened previously, the decision regarding first time screening should be based on the patient's health status and competing risks. The task force did not address differences in life expectancy in the patients 75–85 years who have been screened previously. In addition, the panel recommended not screening patients over the age of 85 years. Again there was no specific comment regarding health status.

Finally, regarding screening paradigms in the elderly, the patient's preferences and perceptions are infrequently considered (28, 29). Walter and Covinsky (6) stress the importance of an assessment of the patient's preferences and values regarding invasive studies, discomfort versus concerns or uncertainty about cancer status. In depth discussion and accurate conveyance of the risks of cancer, benefits of studies and harms associated with studies may take considerable time and skill on the part of the practitioner (28). This process may require specialized training but should be accessible to the elderly, their caregivers and providers.

In summary, colon cancer screening is effective in the prevention of colon cancer death. The risk of harm is increased in this population. In the elderly, the benefits of screening must be assessed in the context of diminished life expectancy especially in patients with severe co-morbidities. Given the heterogeneity of health status and life expectancy in the elderly, a specific age at which to discontinue screening measures is difficult. Screening should be offered to healthy elderly with specific discussion with the elderly regarding their values and preferences.