Table 1. Characteristics of Screening Tests for Colorectal Cancer*

Screening Strategy for CRC Effectiveness in Reducing Incidence and Mortality from CRC Ability to Detect Cancers Likelihood of Generating False-Positive Results Adverse Effects
Digital rectal examination Case-control study found no difference in mortality
OR: 0.96 (0.56, 1.7) (8)

[Level II—poor]

Pathology data suggest <10% of CRCs are within reach of examination finger Unknown No direct adverse effects known
Office FOBT
(1 card)
Unknown Only 58% of cancers are detected on the first of 3 cards, suggesting lower sensitivity than 3-card testing (10)

[Level III—fair]

Little difference compared with 3-card testing (11;12)

[Level 3—poor]

No direct adverse effects known
Home FOBT
(3 cards), unrehydrated
Biennial testing: 2 trials found mortality reductions of: 15% (1%, 26%) (13) and 18% (1%, 32%) (14)

[Level I—good]

One-time sensitivity 30%-40%.

Unrehydrated FOBT finds about 25% of cancers (9).

Single test specificity: 96%-98%

5%-10% of patients will require colonoscopy over 10 years of biennial testing (9)

[Level III—good]

No direct adverse effects known
Home FOBT
(3 cards),
rehydrated
Annual: 33% mortality reduction (13%, 50%); (2)
cancer incidence reduction 20% (10%, 30%) (16)

Biennial: 21% mortality reduction (3%, 38%); (15)
cancer incidence reduction 17% (6%, 27%) (20)

[Level I—good]

Single test accuracy for cancer=50% (95% CI 30%, 70%); for advanced neoplasms, 24% (19%, 29%). (17)

Over 13 years, rehydrated FOBT finds 50% of cancers (2)

[Level III—good]

Single test specificity: 90%

Over 10 to 13 years, 38% of patients tested annually and 28% tested biennially with rehydrated FOBT required colonoscopy (2)

[Level III—good]

Inconvenience, adverse effects resulting from followup tests required after positive results
Sigmoidoscopy Small RCT found decreased CRC mortality with screening: RR 0.50 (0.10, 2.72) (18)

Case- control studies suggest 59% mortality reduction within reach of scope (31%, 75%) (3)

[Level I—fair]

[Level II—good]

One-time screening detects 68%-78% of advanced neoplasia (17,20)

[Level III—good]

N/A <1 in 10,000 perforation rate for diagnostic exams; bleeding occurs in 2.5% after diagnostic studies,
5.5% after procedures with polypectomy (19)

[Level III—good]

Combined FOBT and sigmoidoscopy Nonrandomized trial found 43% mortality reduction by adding FOBT to rigid sigmoidoscopy:

RR 0.57 (0.56, 1.19) (4)

[Level I—fair-poor]

One-time screening detects 76% of advanced neoplasia (17)

Increased yield when sigmoidoscopy added to FOBT (21-23)

[Level III—good]

N/A Sum of adverse effects from each test alone
Double contrast barium enema Unknown One-time sensitivity for cancer or large polyps: 48% (24%-67%) (24)

[Level III—fair]

One-time specificity 85% (82%, 88%) (24)

[Level III—fair]

Perforations: 1 in 25,000 in study with screening and symptomatic patients (25)

[Level III—poor]

Colonoscopy Case-control study:

CRC mortality:
OR: 0.43 (0.30, 0.63)

CRC incidence decreased by 40-60% (26)

[Level II—fair]

Sensitivity for large adenomas: >90%; sensitivity for cancers probably higher (27)

[Level III—good]

N/A Diagnostic procedures:
perforations: 1/2000

Polypectomy:
perforations: 1/500-1,000
bleeding: 1/100-500
death: 1/20,000 (5)

[Level III—fair-good]

*CRC, colorectal cancer; FOBT, fecal occult blood test; RCT, randomized controlled trial.
OR, odds ratio; RR, relative risk. Numbers following in parentheses represent 95% CI.

Evidence grades for each item are in brackets. Level I: evidence from one or more controlled trials; level II: evidence from cohort or case-control studies; level III: evidence from diagnostic accuracy studies or case series. For each level, the investigators have assigned a quality score based on methods described in Harris et al. 2001. (7) N/A=not applicable (see text).


Return to Document