FIT and Colonoscopy in Familial Colorectal Cancer Screening
FIT and Colonoscopy in Familial Colorectal Cancer Screening
Overall, 6472 first-degree relatives of 972 index cases were randomly assigned to either direct colonoscopy or FIT screening during 3 consecutive years. Of these, 4554 (70.4%) were not contacted or declined the invitation to participate in the study. Among the 1918 (29.6%) who were contacted, 352 were excluded (170 in the colonoscopy group and 182 in the FIT group), as they did not meet the inclusion criteria (Figure 1). Of all eligible first-degree relatives, the final study sample consisted of 782 first-degree relatives from 543 index cases in the colonoscopy group and 784 first-degree relatives from 429 index cases in the FIT group.
(Enlarge Image)
Figure 1.
Inclusion, exclusion, and screening test outcomes in first-degree relatives of patients with colorectal cancer. Overall, 2616 first-degree relatives refused the invitation to participate and 1938 were not contacted after being randomly assigned to either interval FIT or colonoscopy because they were not residents of the reference health area served by our hospital, did not attend the appointment made by the high-risk CRC clinic or had died. Criteria for exclusion were a personal history of inflammatory bowel disease, colorectal polyps or CRC, a family history of hereditary CRC, having previously undergone CRC screening with fecal occult blood testing, colonoscopy or sigmoidoscopy, having abdominal symptoms or refusal to participate.
No significant differences between the 2 groups were found with regard to age group, sex, kinship, age of the index case, rate of complete colonoscopies and compliance with the assigned strategy (Table 1). Crossover was significantly higher among participants assigned to the FIT group (6.5%) as compared with the colonoscopy group (1.9%) (P < .001) (Table 1).
According to screening actually performed, 798 (97.5%) of 818 first-degree relatives belonging to the colonoscopy group finally underwent colonoscopy and in the FIT group, 724 (96.8%) of 748 first-degree relatives underwent at least 1 FIT (Figure 1). Among first-degree relatives receiving FIT, 272 (37.5%) completed 3 tests, 254 (35.1%) completed 2 tests and 198 (27.3%) completed 1 test. The majority (72.6 %) of participants completed at least 2 FIT. Of these, 112 (15.5%) tested positive and all of these underwent colonoscopy. Table 2 shows the diagnostic yield in patients with a positive FIT after the performance of the first, the second, or the third test. Twenty-three of 29 (79.1%) advanced neoplasias were detected after the first FIT. Advanced neoplasia was found in 48 (6.0%) first-degree relatives undergoing screening colonoscopy and in 29 (4.0%) of those screened by FIT (OR = 1.53; 95% CI: 0.95–2.46; P = .08). In addition, there was no significant difference in overall advanced neoplasia detection rate between groups when stratified according to first-degree relative age or sex, index-case age, kinship, or the number of relatives with CRC (Table 3). Tumor staging showed 3 CRC stage I and 1 stage II in the FIT group and 4 CRC stage I and 2 stage III in the colonoscopy group.
In the intention-to-screen analysis (Table 4), advanced neoplasia was detected in 44 (5.6%) subjects in the colonoscopy group vs 33 subjects (4.2%) in the FIT group (OR = 1.41; 95% CI: 0.88–2.26; P = .14). Nonadvanced adenomas were found in 148 (18.9%) subjects in the colonoscopy group vs 52 (6.6%) in the FIT group (OR = 3.49; 95% CI: 2.49–4.91); P < .001). In the per-protocol analysis (Table 4), advanced neoplasia was detected in 43 (5.8%) subjects in the colonoscopy group vs 28 (3.9%) in the FIT group (OR = 1.56; 95% CI: 0.95–2.56; P = .08). Nonadvanced adenomas were detected in 148 (19.8%) subjects in the colonoscopy group vs 38 (5.4%) in the FIT group (OR = 4.71; 95% CI: 3.22–6.89; P < .001).
Supplementary Table 1 shows the results of the equivalence analysis, for which we estimated a 5%–20% higher participation rate in the FIT group than in the colonoscopy group. Overall, the 95% CI for the difference in probability of detecting advanced neoplasia between both strategies was within the pre-established ±3% error margin, both in the intention-to-screen analysis and in the per-protocol analysis.
The number of first-degree relatives who needed to be screened to detect one advanced neoplasia was 25 in the FIT group and 16 in the colonoscopy group. However, the numbers of subjects who needed to undergo colonoscopy to find any advanced neoplasia were 4 and 18, respectively.
Table 5 summarizes the follow-up data. A follow-up colonoscopy was performed in 475 of 485 first-degree relatives (97.9%) with at least 2 negative FIT results, resulting in the detection of 12 advanced adenomas. In addition, an interval colonoscopy was performed in 53 of 127 first-degree relatives (41.7%) with a single negative FIT result, resulting in the detection of 4 advanced adenomas. In total, 528 colonoscopies were performed in 612 first-degree relatives (86.3%) with a negative FIT result, resulting in the detection of 16 advanced adenomas. Therefore, 16 (39%) of 41 advanced adenomas were not detected by FIT. In 8 of these patients (19.5% of the 41), advanced adenomas were detected after 3 negative FIT results and in 4 cases (9.8%) after 1 and 2 negative FIT results, respectively. Eleven of 16 advanced adenomas not detected by FIT were located proximal to the splenic flexure.
In the colonoscopy group, 128 of 324 first-degree relatives (39.5%) without advanced neoplasia at baseline colonoscopy who had completed 60 months of follow-up underwent colonoscopy with 2 advanced adenomas detected. In addition, 15 first-degree relatives underwent an interval colonoscopy, with no lesions detected. CRC was not found in either group.
Post-polypectomy surveillance was performed in 20 of 25 (80%) first-degree relatives with advanced adenoma in the FIT group and in 38 of 42 (90.5%) in the colonoscopy group: 5 and 2 advanced adenomas were detected, respectively.
At study completion, 703 (97.0%) subjects in the FIT group and 756 (94.7%) in the colonoscopy group were successfully contacted after a median follow-up of 55 months (range, 3–94 months) from inclusion in either study group. No cases of interval cancer were recorded in either group. Twelve deaths in the FIT group and 6 in the colonoscopy group were recorded, all unrelated to CRC. Overall, 21 (2.9%) subjects in the FIT group and 42 (5.2%) in the colonoscopy group were lost to follow-up.
Results
Overall, 6472 first-degree relatives of 972 index cases were randomly assigned to either direct colonoscopy or FIT screening during 3 consecutive years. Of these, 4554 (70.4%) were not contacted or declined the invitation to participate in the study. Among the 1918 (29.6%) who were contacted, 352 were excluded (170 in the colonoscopy group and 182 in the FIT group), as they did not meet the inclusion criteria (Figure 1). Of all eligible first-degree relatives, the final study sample consisted of 782 first-degree relatives from 543 index cases in the colonoscopy group and 784 first-degree relatives from 429 index cases in the FIT group.
(Enlarge Image)
Figure 1.
Inclusion, exclusion, and screening test outcomes in first-degree relatives of patients with colorectal cancer. Overall, 2616 first-degree relatives refused the invitation to participate and 1938 were not contacted after being randomly assigned to either interval FIT or colonoscopy because they were not residents of the reference health area served by our hospital, did not attend the appointment made by the high-risk CRC clinic or had died. Criteria for exclusion were a personal history of inflammatory bowel disease, colorectal polyps or CRC, a family history of hereditary CRC, having previously undergone CRC screening with fecal occult blood testing, colonoscopy or sigmoidoscopy, having abdominal symptoms or refusal to participate.
No significant differences between the 2 groups were found with regard to age group, sex, kinship, age of the index case, rate of complete colonoscopies and compliance with the assigned strategy (Table 1). Crossover was significantly higher among participants assigned to the FIT group (6.5%) as compared with the colonoscopy group (1.9%) (P < .001) (Table 1).
According to screening actually performed, 798 (97.5%) of 818 first-degree relatives belonging to the colonoscopy group finally underwent colonoscopy and in the FIT group, 724 (96.8%) of 748 first-degree relatives underwent at least 1 FIT (Figure 1). Among first-degree relatives receiving FIT, 272 (37.5%) completed 3 tests, 254 (35.1%) completed 2 tests and 198 (27.3%) completed 1 test. The majority (72.6 %) of participants completed at least 2 FIT. Of these, 112 (15.5%) tested positive and all of these underwent colonoscopy. Table 2 shows the diagnostic yield in patients with a positive FIT after the performance of the first, the second, or the third test. Twenty-three of 29 (79.1%) advanced neoplasias were detected after the first FIT. Advanced neoplasia was found in 48 (6.0%) first-degree relatives undergoing screening colonoscopy and in 29 (4.0%) of those screened by FIT (OR = 1.53; 95% CI: 0.95–2.46; P = .08). In addition, there was no significant difference in overall advanced neoplasia detection rate between groups when stratified according to first-degree relative age or sex, index-case age, kinship, or the number of relatives with CRC (Table 3). Tumor staging showed 3 CRC stage I and 1 stage II in the FIT group and 4 CRC stage I and 2 stage III in the colonoscopy group.
Main Study Outcomes
In the intention-to-screen analysis (Table 4), advanced neoplasia was detected in 44 (5.6%) subjects in the colonoscopy group vs 33 subjects (4.2%) in the FIT group (OR = 1.41; 95% CI: 0.88–2.26; P = .14). Nonadvanced adenomas were found in 148 (18.9%) subjects in the colonoscopy group vs 52 (6.6%) in the FIT group (OR = 3.49; 95% CI: 2.49–4.91); P < .001). In the per-protocol analysis (Table 4), advanced neoplasia was detected in 43 (5.8%) subjects in the colonoscopy group vs 28 (3.9%) in the FIT group (OR = 1.56; 95% CI: 0.95–2.56; P = .08). Nonadvanced adenomas were detected in 148 (19.8%) subjects in the colonoscopy group vs 38 (5.4%) in the FIT group (OR = 4.71; 95% CI: 3.22–6.89; P < .001).
Supplementary Table 1 shows the results of the equivalence analysis, for which we estimated a 5%–20% higher participation rate in the FIT group than in the colonoscopy group. Overall, the 95% CI for the difference in probability of detecting advanced neoplasia between both strategies was within the pre-established ±3% error margin, both in the intention-to-screen analysis and in the per-protocol analysis.
The number of first-degree relatives who needed to be screened to detect one advanced neoplasia was 25 in the FIT group and 16 in the colonoscopy group. However, the numbers of subjects who needed to undergo colonoscopy to find any advanced neoplasia were 4 and 18, respectively.
Follow-up
Table 5 summarizes the follow-up data. A follow-up colonoscopy was performed in 475 of 485 first-degree relatives (97.9%) with at least 2 negative FIT results, resulting in the detection of 12 advanced adenomas. In addition, an interval colonoscopy was performed in 53 of 127 first-degree relatives (41.7%) with a single negative FIT result, resulting in the detection of 4 advanced adenomas. In total, 528 colonoscopies were performed in 612 first-degree relatives (86.3%) with a negative FIT result, resulting in the detection of 16 advanced adenomas. Therefore, 16 (39%) of 41 advanced adenomas were not detected by FIT. In 8 of these patients (19.5% of the 41), advanced adenomas were detected after 3 negative FIT results and in 4 cases (9.8%) after 1 and 2 negative FIT results, respectively. Eleven of 16 advanced adenomas not detected by FIT were located proximal to the splenic flexure.
In the colonoscopy group, 128 of 324 first-degree relatives (39.5%) without advanced neoplasia at baseline colonoscopy who had completed 60 months of follow-up underwent colonoscopy with 2 advanced adenomas detected. In addition, 15 first-degree relatives underwent an interval colonoscopy, with no lesions detected. CRC was not found in either group.
Post-polypectomy surveillance was performed in 20 of 25 (80%) first-degree relatives with advanced adenoma in the FIT group and in 38 of 42 (90.5%) in the colonoscopy group: 5 and 2 advanced adenomas were detected, respectively.
At study completion, 703 (97.0%) subjects in the FIT group and 756 (94.7%) in the colonoscopy group were successfully contacted after a median follow-up of 55 months (range, 3–94 months) from inclusion in either study group. No cases of interval cancer were recorded in either group. Twelve deaths in the FIT group and 6 in the colonoscopy group were recorded, all unrelated to CRC. Overall, 21 (2.9%) subjects in the FIT group and 42 (5.2%) in the colonoscopy group were lost to follow-up.