Baseline faecal occult blood concentration as a predictor of incident colorectal neoplasia: longitudinal follow-up of a Taiwanese population-based colorectal cancer screening cohort

Summary

Background

Despite widespread use of the immunochemical faecal occult blood test (iFOBT), little is known about the subsequent risk of developing colorectal neoplasia for participants with negative iFOBT results. We investigated whether the concentration of faecal haemoglobin at the first screen is predictive of the subsequent incidence of colorectal neoplasia in those with a negative screening result.

Methods

Between 2001 and 2007, we did a prospective cohort study within the Keelung community-based iFOBT screening programme for residents aged 40–69 years, using a cutoff faecal haemoglobin concentration of 100 ng/mL to classify attendees as negative and positive groups for further clinical investigations. 44 324 participants with negative findings and 1668 with a positive result at the first screen (854 non-referrals who refused colonoscopy and 814 with a false-positive result as assessed by colonoscopy) were followed up to ascertain cases of colorectal neoplasia. We investigated the association between baseline faecal haemoglobin concentration and risk of incident colorectal neoplasia, after adjusting for possible confounders.

Findings

Median follow-up was 4·39 years (IQR 2·53–6·12) for all 45 992 participants, during which the incidence of colorectal neoplasia increased from 1·74 per 1000 person-years for those with baseline faecal haemoglobin concentration 1–19 ng/mL, to 7·08 per 1000 person-years for those with a baseline concentration of 80–99 ng/mL. The adjusted hazard ratios (HRs) increased from 1·43 (95% CI 1·08–1·88) for baseline faecal haemoglobin concentration of 20–39 ng/mL, to 3·41 (2·02–5·75) for a baseline concentration of 80–99 ng/mL (trend test p<0·0001), relative to 1–19 ng/mL. These results did not change when we included repeated iFOBT measurements. Non-referrals had the highest risk of incident colorectal neoplasia (adjusted HR 8·46 [6·08–11·76]).

Interpretation

Quantitative faecal haemoglobin concentration at first screening predicts subsequent risk of incident colorectal neoplasia. During follow-up, risk stratification based on faecal haemoglobin could help clinicians, with particular attention being paid to those with higher initial faecal haemoglobin concentrations, especially those just under the threshold taken to indicate presence of colorectal neoplasia.

Funding

None.

Introduction

Population-based screening for colorectal cancer by use of the immunochemical faecal occult blood test (iFOBT, also known as the faecal immunochemical test [FIT]) is widely done;1, 2, 3, 4, 5, 6, 7, 8, 9, 10 however, particularly in large population-based screening programmes, those with negative findings at the first screen (eg, faecal haemoglobin concentration <100 ng/mL with the OC-Sensor method) often consider themselves at low risk of developing colorectal neoplasia. These individuals are less likely to participate in subsequent screening rounds, as noted in several colorectal cancer screening programmes.11, 12, 13, 14, 15 Moreover, uptake of repeated screening is generally low in Asian countries, where population-based iFOBT screening has recently begun.3, 16 Participants who have a negative result at the first screen and do not attend further screening (non-attenders) may go on to develop symptomatic colorectal cancer, because the disease can arise at any time or because a tumour could have been missed. Previous studies reported detection of advanced colorectal neoplasia in around 15–30% of those with faecal haemoglobin concentrations between 50 ng/mL and 100 ng/mL at the first screen who were referred for colonoscopy.4, 6, 10

Therefore, it is worth investigating whether participants who have a negative result at the first screen can be stratified according to subsequent risk of developing colorectal adenoma and cancer. Faecal haemoglobin concentration is currently used to determine whether participants are referred for further clinical investigation. We propose use of faecal haemoglobin concentration to predict subsequent incidence of adenoma and colorectal cancer in individuals who have a negative result at the first screen of a population-based programme (including new cases and false negatives). We postulated that the higher the concentration of faecal haemoglobin, the higher the subsequent risk of developing adenoma and colorectal cancer. We also assessed whether the subsequent risk of colorectal neoplasia was increased for individuals with faecal haemoglobin concentrations of 100 ng/mL or higher at the first screen, but who did not attend colonoscopy when invited (non-referrals), or in whom colonoscopy did not find disease (false-positives), because in neither case is the natural progression of disease interrupted by treatment. Population-based screening for colorectal cancer, using iFOBT and faecal haemoglobin concentration higher than 100 ng/mL as the cutoff for diagnostic testing, was initiated as part of the Keelung Community-based Integrated Screening programme (KCIS), in Keelung, Taiwan, in 1999.¹⁷ Long-term follow-up is now available for this study, providing a unique opportunity to study the association between faecal haemoglobin concentration and subsequent risk of colorectal cancer and adenoma in individuals with a negative result at the first screen, and to quantify the subsequent risk in non-referrals and false-positives.