RESUMO
PURPOSE: To estimate the risk of lower gastrointestinal bleeding (LGIB) caused by malignant lesion in patients presenting with per-rectal bleeding (PRB), by using visual aid as an objective measurement of PRB colour. METHODS: This was a prospective observational study on patients presented with PRB to Family Medicine Specialty Clinic, who undergo flexible sigmoidoscopy (FS) or colonoscopy (CLN) from December 2012 to September 2013. Patients aged 40 years old or above, haemodynamically stable, with normal haemoglobin level were included. Patients with a history of previous colonic surgery, refused to have FS or CLN, with ophthalmologic diseases such as colour blindness were excluded. Parameters including subjective description of PRB colour, number of chosen red colour by patients, source and distance of bleeding from anal verge were recorded for analysis. Receiver operating characteristic (ROC) curve was used to identify the optimal cutoff level of colour for diagnosing colonic lesion. Diagnostic accuracy was assessed by area under the ROC curve (AUC). Accountability of this model was assessed by logistic regression. RESULTS: The dark PRB colour was associated with diagnosis of tumour (p < 0.001) and advanced neoplastic polyp (p < 0.001). The light PRB colour was associated with the diagnosis of piles (p < 0.001). The performance of our model to predict tumour or advanced neoplastic polyps by colour (AUC, 0.798) had a better discriminative power than that to predict colonic lesion alone (AUC, 0.610) by ROC curve analysis. CONCLUSION: Objective measurement of PRB colour accurately estimated the risk of LGIB caused by malignant lesion in patients presenting with PRB.