Refining Our Risk Assessment of Colorectal Cancer in Men
Abstract & Commentary
By Allan J. Wilke, MD, Residency Program Director, Associate Professor of Family Medicine, University of Alabama at Birmingham School of Medicine—Huntsville Regional Medical Campus, Huntsville. Dr. Wilke reports no financial relationship to this field of study.
Synopsis: A prediction tool that factors age, history of smoking, BMI, and alcohol use can determine a man's risk for colorectal cancer.
Source: Driver JA, et al. Development of a risk score for colorectal cancer in men. Am J Med. 2007;120:257-263.
Colorectal cancer (CRC) is the second leading cause of cancer death (behind lung/bronchus) and the fourth most common cancer in men and women (behind prostate, breast, and lung/bronchus). The American Cancer Society estimated that there will be 106,680 new cases and 55,170 deaths in 2006. Yet, if detected and treated in an early stage, CRC is curable. Catch-22? Driver and colleagues believe the system breakdown occurs at the point of deciding to screen. Using data from the Physician's Health Study (PHS), they set out to devise a risk score that would identify individuals at high risk with the implicit assumption that these individuals would be targeted for screening. The PHS enrolled 22,071 American male physicians (40-82 years old at entry, 92% white) in a randomized, placebo-controlled study of aspirin and β-carotene in prevention of cardiovascular disease and cancer. The initial data were collected in 1982 and results of this study were published in 1989 and 1996. After excluding participants with a history of cancer at the initiation of the study and those for whom data on smoking, height, weight, exercise, alcohol use, and diabetes were missing, there were 21,581 left to study.
The authors reviewed the literature for "well accepted and biologically plausible risk factors." They chose age, body mass index (BMI), smoking status, alcohol use, intake of vegetables, intake of multivitamins, vitamin C, and vitamin E, intake of cold cereal, lack of vigorous physical exercise, and history of diabetes. Cases of colon or rectal cancer were reported by the physicians or their families and cancer deaths were confirmed by record review. In 20 years of follow-up, 485 of 21,581 physicians were diagnosed with CRC. As anticipated by their review of risk factors, advanced age, history of smoking, alcohol use, obesity, history of diabetes, and sedentary lifestyle were all significantly associated with the development of CRC in univariate analysis, but vegetable intake, eating cold cereal, and vitamin use were not. In multivariate analysis, diabetes and lack of vigorous exercise were no longer significant, but the other four factors were. The adjusted odds ratios (AOR) with 95% confidence intervals (95% CI) and prediction score points for the risk factors were as follows:
| Risk Factor |
AOR (95% Cl) |
Points |
| Age 50-59 |
2.25 (1.72-2.95) |
2 |
| Age 60-69 |
4.40 (3.36-5.77) |
4 |
| Age ≥ 70 |
6.25 (4.55-8.60) |
6 |
| History of smoking |
1.42 (1.17-1.72) |
1 |
| BMI 25.0-29.9 |
1.26 (1.05-1.52) |
1 |
| BMI ≥ 30 |
1.62 (1.09-2.42) |
2 |
| Alcohol use ≥ once/week |
1.36 (1.08-1.71) | 1 |
Individuals could score between 0 and 10 points. Because the number of physicians who scored 9 or 10 was so small (1%, thank goodness!), those two groups were made one. The predicted number of cases, the observed number of cases, and the predicted OR for each point score was as follows:
| Points |
Predicted # |
Observed # |
OR (95% Cl) |
| 0 |
5 |
6 |
1.00 |
| 1 |
23 |
28 |
1.48 (0.61-3.59) |
| 2 |
41 |
37 |
1.52 (0.64-3.61) |
| 3 |
52 |
47 |
2.25 (0.96-5.29) |
| 4 |
75 |
75 |
3.70 (1.61-8.52) |
| 5 |
81 |
85 |
5.35 (2.33-12.29) |
| 6 |
82 |
86 |
7.46 (3.25-17.12) |
| 7 |
66 |
58 |
7.76 (3.33-18.05) |
| 8 |
37 |
39 |
11.29 (6.19-37-81) |
| 9-10 |
22 |
24 |
15.29 (6.19-37.81) |
By combining the point groups into low (0-3), intermediate (4-6), and high (7-10) risk groups, the authors devised a simplified risk stratification and calculated the observed 20-year risk of developing CRC.
| Risk Group | Predicted # | Observed # | OR (95% Cl) | 20-year risk |
| Low | 121 | 118 | 1.00 | 1% |
| Intermediate | 238 | 246 | 3.07 (2.46-3.83) | 3% |
| High | 126 | 121 | 5.75 (4.44-7.44) | 6% |
Commentary
There are several groups making recommendations for screening for CRC, ranging from the United States Task Force on Preventive Services to the American Society for Gastrointestinal Endoscopy to the American College of Radiology. The Centers for Disease Control and Prevention (CDC) estimated that in 2004, 57% of adults ≥ 50 years had a fecal occult blood test within the last year and/or a colonoscopy within the last 10 years. This is up from 53% in 2001, but the rate lags considerably behind screening for breast (75%) and cervical cancer (85%). Since screening involves retrieving stool or the time and expense of colonoscopy or imaging, it is not too difficult to understand the population's reluctance.
This study's strengths lie in its use of a very large database, its prospective design, and its very long follow-up. However, is this tool, derived from data that is now a quarter of a century old, valid for someone who isn't American, male, or a physician? It may, but first it needs to be validated in different populations, especially ones with women. Then we will need to look at how we will put it into practice. You'll note it is not until a point total of 4 is reached that the OR 95% CI does not include the value 1.00 and becomes significant. In my limited understanding of statistics, combining the point groups makes the best sense, and certainly from a practical view, having just three risk groups will make patient education easier. You'll also note that "low risk" is not "no risk." As the data show, age is the most powerful risk factor.Physcians have incorporated, if not implemented, the maxim of "first colonoscopy at age 50." However, a man who is less than 50, but who smokes, is obese, and drinks alcohol frequently is at the same risk as a vigorous 69-year-old non-smoking teetotaler. The follow-up study should look at whether a patient's knowledge of his risk for CRC motivates behavior change.
A prediction tool that factors age, history of smoking, BMI, and alcohol use can determine a man's risk for colorectal cancer.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content