Association Between Distal Hyperplastic Polyps and Proximal Neoplastic Lesions: A Prospective Study Among 5,613 Subjects

Abstract

We evaluated the association between distal neoplasia and synchronous proximal neoplasia among asymptomatic subjects in a colorectal cancer screening programme, as current evidence on this association is inconclusive. We recruited Chinese asymptomatic screening participants aged 50-70 years who underwent colonoscopy in Hong Kong from 2008-2014. 206 subjects with distal advanced neoplasia or cancer were excluded. The association between distal pathology (tubular adenoma [TA]; hyperplastic polyps [HPs]; no polyps) and proximal pathology (proximal neoplasia [PN]; proximal advanced neoplasia [PAN]) was assessed by multivariate regression models – overall and stratified by the Asia Pacific Colorectal Screening (APCS) scoring system (score 4-7: high-risk; 0-3: lower-risk). A total of 5,842 screening participants underwent colonoscopy. We excluded 206 subjects who were found to have distal advanced neoplasia or distal cancer, and 23 subjects who were intolerant to the laxative regimen or whose bowel preparation was poor. Their average age was 57.6 years (SD 4.87). The prevalence of PN in the “no distal polyp” group, “distal HP” group and “distal TA” group was 14.8%, 19.3% and 29.4%, respectively. The corresponding prevalence of PAN was 1.8%, 3.2% and 3.5%. Participants with distal HPs did not have significantly higher odds of PN (adjusted odds ratio [AOR]=1.24, 95% C.I. 0.97-1.59, P = .089) or PAN (AOR=1.77, 95% C.I. 1.00-3.13, P = .052), except in lower-risk subjects where the odds of PAN was marginally higher in the “distal HP” group than the “no distal polyp” group (AOR=1.97, 95% C.I. 1.01-3.85, P = .048). Overall, the “distal polyp” group had significantly lower odds of PN than the distal TA group (AOR=0.55, 95% C.I. 0.40-0.76, P < .001). The increased risk of PN and PAN among those with distal HPs was modest. Recommending colonoscopy for every patient with distal HPs detected by screening sigmoidoscopy is not supported by the present findings.