Mo1682 INTERVAL POST-COLONOSCOPY COLORECTAL CANCER IN A REFERRAL CENTER COHORT: A GOOD REASON TO IMPROVE THE COLONOSCOPY QUALITY

Abstract

Colonoscopy has become the preferred modality for colorectal cancer (CRC) screening in many countries. However, it is well recognized that colonoscopy is not perfectly sensitive for the detection of neoplasia and that CRC can be diagnosed within a short interval following a colonoscopy in which no cancer was detected. These cancers are also known as interval post-colonoscopy colorectal cancers (PCCRCs). Several clinical and endoscopy-related risk factors have been associated with the risk of developing PCCRC. However, results are inconsistent across studies. To investigate the prevalence, characteristics and risks factors for PCCRC. We conducted a nested case-control study. Eligible patients were affiliated to a health insurance company that provides health-care to 41.000 adults and centralizes medical care in one tertiary hospital. Inclusion criteria were: Individuals between 55 to 90 years old, with average risk for CRC, 5 or more years of health plan enrolment duration and a CCR diagnosed between Jan 2010 and Aug 2017. The CRC was defined as PCCRC if it was detected within 6–60 months of a colonoscopy in which no cancer was detected. PCCRC cases were matched, using diagnosis (reference) date, to randomly selected free-CRC controls, on age and sex. The proportion of PCCRC was defined as the number of PCCRCs divided by all CRCs diagnosed during study period. Student’s t-test, U Mann-Whitney test were used to compare PCCRC group, CRC group and free-CRC controls. The statistical significance was P < 0.05. Of a total of 134 patients that presented CRC during de study period, 11 (8.2%) were PCCRC. The estimated frequency of PCRC among the total number of colonoscopies performed during the study period (n: 16.473) was one PCCRC every 1500 colonoscopies. Age, sex, Charlson comorbidity index, and advanced cancer stage were not associated with PCCRC. PCCRCs were more frequently detected in the proximal colon (73% vs 32%, P < 0.01). As compared with index colonoscopy of patients without a PCCRC diagnosis, PCCRC were least likely to be associated with a complete colonoscopy that reaches cecal intubation (79% vs 94%, P < 0.03). Only 15% of PCCRC patients had an index colonoscopy in which a polypectomy was performed vs 38% of patients without PCCRC (P: 0.05). PCCRC patients were least likely to present adenomas in index colonoscopy compared with those colonoscopy of patients without PCCRC (5% vs 27%, P:0.05). In our cohort, 8% of all patients with CRC had PCCRC.