Table 3.
Results of trials of Hemoccult screening for colorectal cancer
| Mandel et al (1993)8 Minnesota | Hardcastle et al (1996)9 Nottingham | Kronborg et al (1996)10 Funen | Kewenter et al (1994)11 Gothenburg | Winawer et al (1993)12 New York | |
|---|---|---|---|---|---|
| Years of trial follow up | 13 | Median 7.8 | 10 | Median 8.3 | 9 |
| Early CRC: | At mean 3 year follow up (107 349 subjects) | At mean 3.2 year follow up | At mean 2.25 year follow up (27 700 subjects) | At study entry | |
| No of cancers (rate/10 000) | Initial screen approx 87 (28) First year control approx 45 (29) | Screen 181 (34) Control 123 (23) | Screen 147 (47) Control 115 (37) | Screen 61 (44) Control 20 (14) | Regular attenders: Screen 11 (15) Control 1 (5) First time attenders: Screen 26 (45) Control 17 (25) |
| Relative incidence (95% CI) | 0.96 (0.67 to 1.37) | 1.47 (1.17 to 1.85) | 1.28 (1.00 to 1.63) | 3.03 (1.83 to 5.03) | 1.39 (0.79 to 2.44) |
| CRC at latest follow up: | |||||
| No of cancers (rate/10 000) | Annual screen 323 (207) Biennial screen 323 (207) Control 356 (231) | Screen 893 (117) Control 856 (112) | Screen 481 (155) Control 483 (156) | Screen 371 (108) Control 379 (111) | Regular attenders: Screen 72 (100) Control 22 (104) First timers: Screen 57 (98) Control 48 (72) |
| Cancers detected by Hemoccult screening | Annual screen 50% Biennial screen 39% | Screen 27% | Screen 25% | Screen 28% | All CRC 39% |
| Staged CRCs that were early stage (Dukes’s type A) | 91% staged Annual screen 33% Biennial screen 29% Control 25% | 97% staged Screen 21% Control 12% | 95% staged Screen 23% Control 12% | All staged Screen 26% Control 9% | All staged Screen 35% Control 27% |
| Sensitivity of Hemoccult for colorectal cancer | Rehydrated 92% Unhydrated 81% (Stated by authors) | Unhydrated 64% (Stated by authors) | Unhydrated 46% | Overall (most rehydrated) 81% | Not calculable from data |
| Time after screening for interval cancers to occur | 1 year | 2 years | 2 years | 2 years | — |
| Positive predictive value of Hemoccult for colorectal cancer | 2.2% (rehydrated) to 5.6% (unhydrated) | 1st screen 9.9%, 2nd screen 11.9% | 1st screen 17.7%, 2nd screen 8.4%, 3rd screen 16.3%, 4th screen 10.8%, 5th screen 10.2% | 1st screen (most unhydrated) 5.0%, 2nd screen (all rehydrated) 4.2% | Overall 10.7% |
| No of CRC deaths at latest follow up (rate/10 000) | Annual screen 82 (53) Biennial screen 117 (75) Control 121 (79) | Screen 360 (47) Control 420 (55) | Screen 205 (66) Control 249 (80) | Screen 121 (35) Control 138 (40) | Regular attenders: Screen 24 (33) Control 6 (28) First timers: Screen 12 (21) Control 22 (33) |
| Relative risk (95% CI) of CRC death with screening | Annual screen 0.67 (0.51 to 0.89) Biennial screen 0.95 (0.74 to 1.23) Total 0.81 (0.65 to 1.02) | 0.86 (0.74 to 0.99) | 0.82 (0.68 to 0.99) | 0.88 (0.69 to 1.12) | Regular attenders 1.18 (0.48 to 2.88) First timers 0.63 (0.31 to 1.27) Total 0.87 (0.53 to 1.43) |
CRC=Colorectal cancer.