AbstractBACKGROUND & AIMS:
Fecal immunochemical tests (FITs) are used in colorectal cancer (CRC) screening. We compared detection of CRCs and colorectal neoplasms by FITs using fresh samples (collected into buffer-filled tubes) vs frozen samples, and we assessed the effects of seasonal variations in ambient temperature on test performance.METHODS:
We performed a prospective study of 3466 individuals (50% male; mean age, 62 years) undergoing screening colonoscopies at 20 gastroenterology practices in southern Germany from November 2008 through September 2014. Frozen stool samples (collected and frozen by patients through February 2012, n = 1644) and fresh stool samples (collected by patients into buffer-filled tubes after February 2012, n = 1822) were obtained; hemoglobin (Hgb) concentrations were measured by using a commercial, quantitative FIT (cutoff value for positive result, 17 μg Hgb/g feces). Colonoscopy results were used as the gold standard, with results categorized as CRC, advanced adenoma, non-advanced adenoma, or no colorectal neoplasm. Differences in detection of colorectal neoplasms with fresh vs frozen samples were compared by using Wilcoxon rank sum test (continuous variables) and Fisher exact test (categorical variables). We also compared test performance when samples were collected during different seasons (based on outdoor temperature less than 8°, 8°–15°, or more than 15°).RESULTS:
Of the samples analyzed by FIT, 12.8% of frozen stool samples (95% confidence interval [CI], 11.3%–14.5%) and 8.7% of fresh stool samples (95% CI, 7.5%–10.1%) had positive results (P value for difference < .001). When adjusting the Hgb cutoff value to produce the same percentage of positive results for fresh and frozen samples (10% and 5%), FIT with frozen vs fresh samples detected colorectal neoplasms with similar levels of sensitivity and specificity. For example, at cutoff values that produced 5% positive results for each sample type, FIT detected advanced neoplasms with 27.8% sensitivity when frozen samples were used (95% CI, 21.4%–35.1%) and 25.6% sensitivity when fresh samples were used (95% CI, 19.8%–32.1%). Specificity values were 97.7% when frozen samples were used (95% CI, 96.8%–98.4%) and 97.6% when fresh samples were used (95% CI, 96.7%–98.3%). We did not observe any differences in detection of neoplasms during different seasons that were based on outdoor temperature.CONCLUSIONS:
In a prospective study of 3466 individuals who underwent screening colonoscopies and received FITs, we found that use of fresh vs frozen samples slightly affected positivity rates and the proportions of CRCs or adenomas detected at the recommended Hgb cutoff value. However, after we adjusted Hgb cutoff values to produce equal proportions of positive results for fresh vs frozen samples, the performance of the FIT was similar with each sample type. Season of sample collection (based on outdoor temperature) did not affect detection of CRC using either sample type in this study from Middle Europe.