Measuring interval cancers in population-based screening using different assays of fecal occult blood testing: the District of Florence experience.

The fecal occult blood test (FOBT) has demonstrated its efficacy in reducing mortality from colorectal cancer (CRC). The guaiac-based FOBT has been criticized for its low sensitivity. In this study, two different assays for FOBT (guaiac or an immunochemical test based on reversed passive hemagglutination [RPHA]) were tested for comparison within a population-based screening program for colorectal cancer in the province of Florence (Italy). The proportional incidence method was used to calculate sensitivity for both FOBTs, according to rank of screening (first or repeat), age at entry (two groups of 50 to 59 and 60 to 70 years old) and lesion site (colon or rectum). When comparing FOBTs, the sensitivity multivariate Poisson regression was used to adjust for other variables. The sensitivity after the first 2 years was 50% (95% confidence interval [CI] 34% to 63%) for the guaiac test versus 82% (95% CI 67% to 92%) for RPHA. At multivariate analysis the risk of developing an interval cancer after a guaiac test is almost 3 times that after RPHA (rate ratio = 2.64; 95% CI 1.3 to 5.4). Our study confirms that RPHA is more sensitive than the guaiac test. The assumption that FOBT screening for CRC has to be based on a guaiac test should be reconsidered, and RPHA should be recommended as the standard FOBT for screening purposes.