Sage Journals: Discover world-class research

Abstract

Background

Colorectal cancer (CRC) is the second commonest cancer in England. Incidence rates for colorectal adenomas and advanced colorectal neoplasia are higher in men than in women of all age groups. The male-to-female ratio for CRC incidence rates differs for different parts of the large bowel.

Objective

To summarize the current evidence on colorectal screening and prevention, focussing on potential differences in benefits between sexes and colorectal sites.

Methods

(i), We reviewed the evidence from randomized controlled trials (RCTs) of the impact of different screening approaches on CRC incidence and mortality, overall, for each sex separately, and for subsites of the large bowel. (ii) We reviewed studies comparing detection parameters for faecal immunochemical tests for haemoglobin (FIT) with guaiac FOBt (gFOBt). (iii) The role of aspirin in CRC prevention in the general population was reviewed using evidence from RCTs, with specific emphasis on the differences observed between sexes and lesion site.

Results

(i) Our intention-to-treat random-effects meta-analysis showed that once-only flexible sigmoidoscopy (FS) screening performed on average-risk individuals aged 55 + decreased CRC incidence by 18% and mortality by 28%, but sex-specific results were lacking. (ii) Modern quantitative FIT were superior to qualitative gFOBt in average-risk population screening in their ability to discriminate between individuals with and without colorectal neoplasia. Some recent FIT studies suggest varying operating characteristics in men and women. (iii) Evidence of an effect of aspirin on the incidence of CRC (in particular, proximal disease) in both sexes aged 40 and over at average-risk of CRC is emerging.

Conclusions

We encourage researchers of CRC screening and prevention to publish their results by sex where possible. Pilot studies should be undertaken before implementation of quantitative FIT in a national screening programme to establish the appropriate threshold. Finally, individual risk assessment for CRC and non-CRC events, will be necessary to make an informed decision on whether a patient should receive aspirin chemoprevention.

Keywords

Colorectal cancer screening flexible sigmoidoscopy faecal occult blood test guaiac immunochemical aspirin sex-specific site-specific

Introduction

In England, colorectal cancer (CRC) is the second most common cancer; in men, it is the third most common cause of cancer death after prostate and lung cancer, and in women the third after breast and lung cancer.¹ Incidence rates for colorectal adenomas (CRA) and advanced colorectal neoplasia (advanced CRA and CRC) are higher in men than in women of all age groups.^2,3 The male-to-female ratio for CRC incidence rates differs for different parts of the large bowel,⁴ suggesting different CRC aetiologies between sexes and the potential need for sex-specific recommendations in CRC prevention.

In England, age-standardized incidence rates already showed a high male-to-female ratio (close to 1.8) for rectosigmoid and rectal cancers about the time population-based data collection commenced in the early 1970s, whereas the ratio for colon cancer was close to 1.0¹, (web appendix Table A.0). However, between 1971 and 2005, the year before the implementation of the NHS Bowel Cancer Screening Programme (NHSBCSP), CRC incidence rates had increased by 24% in men but only by 10% in women, while CRC mortality rates had decreased by 32% in men and by 46% in women.⁵ In 2009, the male-to-female ratio for rectosigmoid and rectal cancers was still close to 1.8, whereas the male-to-female ratio for colon cancer had increased from 1.0 to almost 1.4, indicating that the comparatively higher increase in CRC incidence observed in men versus women over the last three decades had been driven by an increase in colon cancer. These ratios are a current feature of most high (but not low) incidence populations, ie. westernized countries, where the male-to-female ratio for rectal cancer is consistently higher than the corresponding ratio for colon cancer (2.0 versus 1.3).⁶

CRC incidence has also been observed to rise more rapidly in men than in women when emigrating from low to high CRC incidence areas.⁴ These observations suggest that in addition to a potential effect of sex on incidence of rectal cancer, environmental factors which arose in developed countries over the past few decades have rendered men more susceptible to colon cancer. These hypotheses are supported by the fact that the upper and lower parts of the large bowel are derived from two different embryonic tissues, with their own vascular supplies and biological environments,⁷ leading to differential sensitivity to sex hormones and consequent variation in carcinogenesis pathways between men and women.^4,8

Endoscopic screening with flexible sigmoidoscopy (FS) has been shown to be more sensitive than both guaiac and faecal immunochemical tests for haemoglobin (gFOBt and FIT, respectively) for the detection of adenomatous polyps, the precursor of CRC lesions.⁹

Results from RCTs have shown that screening by FOBt can reduce mortality from large bowel cancer.^10,11 To date only one trial in the United States¹² has demonstrated any significant reduction in the incidence of CRC, and in that trial a high proportion of subjects in the intervention arms underwent colonoscopy. The uptake of population-based FOBt screening is low compared with, for example, that of mammographic screening, and also varies by age and sex,^13,14 so that ‘per protocol’ trial estimates of efficacy in people actually screened are particularly relevant.

The randomized trials have used a guaiac-based test kit (gFOBt) which has relatively poor sensitivity to preclinical CRC.¹⁵ There is therefore a need to identify a test with increased sensitivity to early stage disease without loss of specificity.¹⁶ A recent report by the Cancer Care Ontario’s Faecal Immunochemical Tests Guidelines Expert Panel¹⁷ concluded that immunochemical methods of faecal blood detection (FIT) were superior to guaiac methods in terms of test performance characteristics and detection of CRC and advanced CRA based on a review of findings from two RCTs,^18,19 and two community-based cohort studies where colonoscopy was performed in all participants.^20,21

Alongside screening, chemo-intervention offers a potential complementary approach to cancer prevention. Chronic inflammation plays an important role in the development and progression of colorectal cancer.^22,23 Aspirin is a non-steroidal anti-inflammatory drug (NSAID) able to irreversibly acetylate cyclooxygenases-1 and -2 (COX-1 and -2), with a 100-to-200 fold higher affinity for COX-1.²⁴

In this paper, we perform meta-analyses of the RCTs of once-only FS and biennial gFOBt and summarize the data currently available overall, and for subgroup analyses by sex and large bowel sub-location. We also consider the results of biennial gFOBt in the context of a once-only FS screen.

In addition, we review diagnostic cohort studies assessing the gFOBt and FIT tests in the same patients, and derive the relative sensitivity (RSN), the relative false-positive rate (RFP), and the odds of a true positive for the detection of both advanced neoplasia and early CRA located in different parts of the large bowel.

Finally, in an attempt to assess the potential role of aspirin in the primary prevention of bowel cancer, (ie. as part of a national programme) we gathered the evidence currently available for average-risk populations and reviewed the results obtained for each sex and by tumour site. Any preferential effect of aspirin prophylaxis according to tumour site will be of interest in the context of a complementary approach to an already implemented prevention scheme. The potential benefits of aspirin in primary CRC prevention are presented alongside potential harms to assess whether chemoprevention with aspirin is one possible intervention to reduce the burden associated with this disease.

Methods

Data sources and study selection

We conducted electronic searches of the PubMed/MEDLINE database, and of the Cochrane Collaboration’s Library up to 31 December 2012.

Once-only FS screening in the general population

The following search terms were used concomitantly and the search restricted to the title/abstract of the article: ‘sigmoidoscopy’, and ‘random*’ (or ‘trial*’). In addition, reference lists of included studies and recent related reviews were also scanned.^10,11,25

Selected studies

Abstracts of the studies identified by the above searches were reviewed by the first author and the studies with the following criteria were selected:

Studies published in English;

Randomized controlled trials, single or double-blind;

Trials performed in individuals at average risk of colorectal neoplasia (general population-based screening /community screening);

Trials comparing flexible sigmoidoscopy with no intervention;

Trials recruiting individuals aged 55 or over at randomization;

Trials reporting results on incidence of advanced neoplasia, and/or mortality from colorectal cancer;

Trials with at least 5 years of follow-up for both incidence and mortality.

Biennial gFOBt screening in the general population

The following search terms were used concomitantly and the search restricted to the title/abstract of the article: ‘fecal' (or ‘faecal’), ‘blood test', and ‘random*’ (or ‘trial*’). In addition, reference lists of included studies and recent related reviews were also scanned.^10,11,25

Selected studies

Abstracts of the studies identified by the above searches were reviewed by the first author and the studies with the following criteria were selected:

Studies published in English;

Randomized controlled trials, open, single or double-blind;

Trials performed in individuals at average risk of colorectal neoplasia (general population-based screening/community screening);

Trials comparing gFOBt with no intervention;

Trials recruiting individuals aged 45 or over at randomization;

Trials reporting results on incidence of advanced neoplasia, and/or mortality from colorectal cancer;

Trials with at least 5 years of follow-up for both incidence and mortality;

Trials offering biennial gFOBt screening.

Comparison of gFOBt and FITcharacteristics

The following search terms were used concomitantly and the search restricted to the title/abstract of the article: ‘fecal' (or faecal), ‘blood test', immunochemical' (or ‘immunologic*'),'guaiac' (or ‘chemical'). In addition, reference lists of included studies and recent related reviews were also scanned.^17,25–30

Selected studies

Abstracts of the studies identified by the above searches were reviewed by the first author and the studies with the following criteria were selected:

Studies published in English;

Diagnostic cohort studies;

Studies performed in individuals at average risk of colorectal neoplasia and which excluded individuals with a personal or familial history of CRC;

Studies comparing directly at least one gFOBt and one FIT;

Studies assessing both tests in the same group of individuals;

Studies reporting results in individuals aged 50 and over (although lower limit for age range in population studied may be lower than 50);

Studies using colonoscopy or sigmoidoscopy as a reference standard to obtain a pathology diagnosis for individuals with positive test results for either gFOBt or FIT, or for all participating individuals;

Studies reporting the number of true positive and true negatives based on reference standard for colorectal carcinomas/cancers (CRC), advanced and/or early colorectal adenomas (CRA).

Prophylactic aspirin in the general population

For benefit of aspirin on incidence of and mortality from colorectal neoplasia, the following search terms were used concomitantly and the search restricted to the title/abstract of the article: ‘aspirin’, ‘random*’(or ‘trial*), and ‘colo*’ (or ‘rect*' or ‘bowel' or ‘cancer’). The reference lists of the initially selected studies were scanned for any additional relevant references. All publications describing results related to the final list of selected studies were retrieved to identify reports of harms from aspirin treatment.

For additional information on serious vascular events and other malignancies, recent systematic reviews and meta-analyses of RCTs comparing incidence of and mortality from those events between aspirin and control, were identified (January 2000 – December 2012) and scanned for any relevant results.The following search terms were used concomitantly and the search restricted as stated in brackets: ‘aspirin’ [title], ‘random*’ (or ‘trial*') [title/abstract], ‘mortality’ (or ‘fatal’, ‘death’, ‘incidence’, ‘occurrence’) [title/abstract], and ‘bleed*' (or ‘haemorrhag*’ or ‘hemorrhag*’) [title/abstract], OR ‘heart’ (or ‘cardi*’, ‘vascular’, ‘coronary’, ‘stroke’) [title/abstract], OR ‘cancer’ (or ‘carcinoma’,’neoplasm’, ‘malignant’, ‘tumour’) [title/abstract].

Selected studies

Abstracts of the studies identified by the above searches were reviewed by the first author and the studies with the following criteria were selected:

Studies published in English;

Randomized controlled trials, single or double-blind;

Trials performed in average-risk individuals (general population-based screening/community screening);

Trials comparing aspirin with placebo or no treatment;

Trials reporting results on incidence of and/or mortality from CRC, cancer, and/or vascular events;

Trials with at least 5-year follow-up for incidence and mortality;

Trials offering an oral dose of 75 mg aspirin or more, daily or on alternate days;

No restriction was applied to age at randomization. Primary prevention trials for which patients were selected on the basis of high risk factors for vascular disease or other diseases (eg. diabetes) were not included in the selection.

Statistical methods

A standard collection form was used to extract relevant information for each study included in the meta-analyses. Summary-level effect sizes were pooled using the inverse variance method (DerSimonian-Laird random-effects model) and heterogeneity (χ2) between the studies was assessed.³¹

Once-only FS screening and biennial gFOBt screening in the general population

We calculated a pooled estimate of the effect of FS or gFOBt compared with no intervention on incidence and mortality from colorectal cancer and on all-cause mortality in invited individuals. The meta-analyses were performed using intention-to-treat (ITT) summary-level rate or hazard ratio data and corresponding 95% confidence intervals (CI). Individual data were obtained for the Nottingham trial from the authors (JH Scholefield and SM Moss). We also obtained ‘per protocol’ point estimates adjusted for potential selection bias using the method of Cuzick et al.,³² with 95% CI derived using the profile likelihood method of Stryhn and Christensen.³³ Compliers were defined as those who completed their test during the first round of screening (prevalence screen).

Comparison of gFOBt and FIT characteristics

The relative sensitivity (RSN, the ratio of two sensitivities) and the relative false-positive rate (RFP, the ratio of two false-positive rates) of CRC/CRA detection using FITcompared with gFOBt were derived according to the method described by Cheng and Macaluso.³⁴ This allowed inclusion of data from studies where the reference standard was applied solely to the population of individuals with positive test results.³⁵

Briefly, if we denote by m₁' the number of true positive individuals identified by FIT, and by n₁' the number of true positive individuals identified by gFOBt, then an estimate of the RSN of FIT compared with gFOBt is m₁' / n₁'. Similarly, if we denote by m₁'' the number of false positive individuals for FIT, and by n₁'' the number of false positive individuals for gFOBt, an estimate of the RFP of FIT compared with gFOBt is m₁'' / n₁''. The denominators for the absolute sensitivity and specificity of each test (ie. the total numbers of true positive and true negative individuals, which could only be ascertained if all subjects received the reference standard investigation) cancel out in the calculation of the relative figure.

Most authors classified patients as having “advanced colorectal neoplasia” if they had CRC, at least one adenomatous polyp with diameter ≥ 10 mm, or at least one adenomatous polyp with diameter < 10 mm with tubular/villous components or high-grade dysplasia. Some authors also treated the presence of ≥3 polyps as advanced neoplasia, irrespective of their size or grade of dysplasia.³⁶ Patients were generally classified as having “early adenoma” if they only presented with adenomas <10 mm. Hyperplastic polyps were generally not included as neoplasia with the exceptions of Cheng et al.³⁷ and Graser et al.³⁸

For advanced neoplasia (CRC + CRA ≥ 10 mm) and both advanced and early neoplasia (CRC + all CRA), the number of individuals with a false positive test result (ie. m₁'', n₁'') was calculated as the difference between the total number of individuals with a positive result for a given test (ie. m1, n1) minus the total number of CRC and CRA ≥ 10 mm (or all-size CRA, respectively) actually detected by the test (ie. m₁', n₁').

Where available, data were reported according to the location of lesions: proximal (caecum, ascending, and transverse colon including splenic flexure) versus distal (distal to splenic flexure) location. Where more than one lesion was reported for a patient, the location was set to be the region where the largest size CRA was found.

The odds and corresponding 95% CI of individuals with CRC, advanced colorectal neoplasia or adenoma having a positive FIT only (c) versus a positive gFOBt only (a) were also derived where data for the total number of individuals with a true FIT positive (c + b = i), a true gFOBt positive (a + b = g), and concomitant true gFOBt and FITpositives (b) were available.

Prophylactic aspirin in the general population

The effectiveness of aspirin in preventing incidence of, and mortality from CRC in the average-risk populations was assessed overall, and by sex and site of neoplasia. ITT risk ratios (RR) are reported in the text wherever possible (otherwise, ‘per protocol’ values are reported). In addition, summary-level crude RR and corresponding 95% CI were derived using the original publications (unless the required data were unavailable) for the comparison of aspirin versus control, for all endpoints considered.

Results

Once-only Flexible Sigmoidoscopy screening in the general population

The primary objective of this review was to assess whether FS is an effective screening tool to reduce CRC incidence and CRC mortality, overall as well as by tumour location and by sex. An overview of the characteristics of the RCTs included in this review is presented in Table 1. Although the intervention is once-only FS, the findings prompt a range of further investigations and colonoscopic follow-up policies, which vary from study to study (web appendix Table A1.1). For example, the PLCO trial referred all patients with any polyp to a physician, and the SCORE trial referred all patients with polyps > 5 mm, high-risk polyps, or inadequate bowel preparation, for colonoscopy, whereas in the UKFSST, only patients with polyps ≥ 10 mm or high-risk polyps were referred for colonoscopy.

Table 1.

Overview of Flexible Sigmoidoscopy (FS) RCTs characteristics.

Overall incidence & mortality

Although the first large RCT of FS versus no intervention performed in Norway did not show any reduction in CRC incidence at 7-year follow-up (NORCCAP1, Figure 1), more recent trials reported significant reductions in CRC incidence after 10-year follow-up or more, overall amounting to a significant 18% reduction in an ITT analysis (Risk ratio (RR) = 0.82, 95% CI 0.74–0.90; Figure 1 and web appendix Table A1.2). This reduction approached 30% for subjects who actually participated in screening (‘per protocol’ RR = 0.73, 95% CI 0.65–0.82; web appendix Figure A1.1b and Table A1.2). FS screening also resulted in a significant decrease in overall CRC mortality (ITT RR = 0.72, 95% CI 0.65–0.80; Figure 2 and web appendix Table A1.3). The corresponding result for the ‘per protocol’ analysis was a RR of 0.58 (95% CI 0.47–0.71).

Figure 1.

Random-effects meta-analysis of RCTs comparing once-only FS versus no screening in average-risk individuals with respect to CRC incidence (ITT analysis).

Figure 2.

Random-effects meta-analysis of RCTs comparing once-only FS versus no screening in average-risk individuals with respect to CRC mortality (ITT analysis).

Incidence & mortality by site

Incidence of distal CRC was significantly reduced (UKFSST, SCORE & PLCO, ITT RR = 0.69, 95% CI 0.63–0.75; Figure 3 and web appendix Table A1.2), as was mortality from distal CRC (SCORE & PLCO, ITT RR = 0.58, 95% CI 0.40–0.83; web appendix Figure A1.5a and Table A1.3). Reduction in mortality from rectosigmoidal cancer was also reported in NORCCAP1 among screening attenders (HR = 0.24, 95% CI 0.08–0.76, not adjusted for self-selection bias⁴⁰). Incidence of proximal CRC was also found to be significantly reduced (ITT RR = 0.91, 95% CI 0.84–0.99; Figure 3 and web appendix Table A1.2), although the data were more variable between trials: the UK trial showed little impact on incidence of proximal CRC compared with the United States and Italy interventions. Mortality from CRC located at proximal sites was not significantly different between intervention and control groups (Italy and United States, ITT RR = 0.95, 95% CI 0.77–1.20; web appendix Figure A1.5a and Table A1.3), even after adjusting for participation (Italy, ‘per protocol’ RR = 0.78, 95% CI 0.45–1.35; web appendix Figure A1.5b and Table A1.3).

Figure 3.

Random-effects meta-analysis of RCTs comparing once-only FS versus no screening in average-risk individuals with respect to CRC incidence at proximal and distal sites (ITT analysis).

Incidence & mortality by sex

A secondary objective was to determine whether the protective effects of FS intervention differed between the male and female populations. CRC incidence data by sex showed significant reductions for both men (SCORE & PLCO ITT RR = 0.78, 95% CI 0.66–0.94), and women (ITT RR = 0.83, 95% CI 0.71–96; web appendix Figure A1.3a and Table A1.2). Interestingly, SCORE & PLCO each reported very similar effects of FS by CRC site, but opposite trends by sex. The SCORE trial conferred more advantage to women (ITT RR = 0.72, 95% CI 0.55–0.96 for women, and RR = 0.88, 95% CI 0.71–1.09 for men), while the PLCO trial conferred more benefit to men (ITT RR = 0.73, 95% CI 0.66–0.87 for men and RR = 0.86 95% CI 0.76–0.98 for women). Again, those two trials adopted a different polyp follow-up strategy. The decreased incidence with FS observed for men in the PLCO trial translated into a significant reduction in mortality from CRC (ITT RR = 0.66, 95% CI 0.53–0.81), while the mortality reduction observed for women was not significant (ITT RR = 0.87, 95% CI 0.68–1.12; web appendix Table A1.3). Data on CRC mortality by sex have so far only been published for the United States trial.

Potential harms

Alongside the benefits of FS screening, it is also necessary to consider the potential harms. Whitlock et al.⁴⁴ estimated the number of serious complications with FS from a meta-analysis of eight studies in average-risk populations to be eight times less common (0.34 per 1,000 procedures, 95% CI 0.06–1.9) than with colonoscopy (2.8, 95% CI 1.5–5.2). In another review,⁹ the most common adverse event associated with FS was non-fatal bleeding (0.8–30.6 per 1,000 procedures). To our knowledge, no difference in incidence of adverse events between male and female patients has been reported.

All-cause mortality

In the more recent RCTs, all-cause mortality was consistently slightly lower in the intervention group. ‘Per protocol’ estimates were available for the SCORE & UKFSST trials (RR = 0.95, 95% CI 0.90–1.00; web appendix Figure A1.7 and Table A1.4).

Biennial gFOBt screening in the general population

Meta-analyses of biennial gFOBt and endoscopic follow-up of positive test results have been published previously.⁴⁵ The characteristics of the RCTs included in this review are presented in Table 2 (additional details can be found in web appendix Table A2.1).

Table 2.

Overview of biennial guaiac faecal blood test (gFOBt) RCTs characteristics.

Incidence & mortality

Using data with the longest follow-up available, we confirmed the absence of any significant reduction in CRC incidence by repeated gFOBt screening overall, by site or sex, including where ‘per protocol’ analyses were performed (Figure 4 and web appendix Table A2.2), and the presence of a significant decrease in CRC mortality in an ITT analysis after a median follow-up of at least 14 years (ITT RR = 0.88, 95% CI 0.83–0.94; Figure 5a and web appendix Table A2.3). The smallest reduction was observed for the Nottingham trial analysed 19.5 years after the end of the screening period, although larger reductions in CRC mortality were observed in earlier analyses of this trial (web appendix Table A2.5). ‘Per protocol’ and subgroup data were available for the Nottingham and Funen trials, which both used unrehydrated gFOBt. The meta-analysis adjusted for screening attendance for the overall data revealed a slightly stronger reduction in mortality (‘per protocol’ RR = 0.82, 95% CI 0.72–0.93; web appendix Figure A2.4b and Table A2.3). CRC mortality was similarly decreased for both proximal (ITT RR = 0.81, 95% CI 0.67–0.98) and distal (ITT RR = 0.85, 95% CI 0.74–0.98; web appendix Figure 2.5a and Table A2.3) cancers. Screening using gFOBt was equally effective at reducing CRC mortality in males and females (ITT RR = 0.84, 95% CI 0.73–0.98 for men and RR = 0.83, 95% CI 0.71–0.97 for women; web appendix Figure A2.6a and Table A2.3).

Figure 4.

Random-effects meta-analysis of RCTs comparing biennial gFOBt versus no screening in average-risk individuals with respect to CRC incidence (ITT analysis).

Figure 5.

Random-effects meta-analysis of RCTs comparing biennial gFOBt versus no screening in average-risk individuals with respect to CRC mortality (ITT analysis).

Potential harms

Screening using gFOBt per se does not cause any serious adverse events, although subsequent diagnostic testing by colonoscopy may do so¹⁴; gFOBt true specificity for any dysplasia is fairly high (ie. 96–98% for Hemoccult-II as determined in patients who had all undergone colonoscopy^49,50), therefore limiting the number of unnecessary colonoscopies. In addition, the anxiety caused by the receipt of an abnormal gFOBt test was shown to decrease the day after colonoscopy, and remained low one month later in patients with false positive results.⁵¹

All-cause mortality

All-cause mortality did not differ between the intervention and control groups (web appendix Figure A2.7 and Table A2.4).

Comparison of gFOBt and FIT characteristics in the general population

We reviewed data from prospective studies which performed both guaiac (gFOBt) and faecal immunochemical tests for haemoglobin (FIT) tests in the same average-risk individuals. An overview of the characteristics of the prospective diagnostic studies included in this review is presented in Table 3 (additional details can be found in web appendix Table A3.1). Test specifications are quoted according to standards set by the World Endoscopy Organization CRC Screening Committee’s Expert Working Group on FIT.¹⁶

Table 3.

Overview of characteristics of prospective diagnostic cohort studies comparing gFOBt and FIT in the same groupof screened average-risk individuals (age ≥ 50 except where*).

Overall comparison of test characteristics

Because all studies except one²¹ only referred for colonoscopy individuals with at least one positive test, relative measures were used to compare both tests. gFOBt and FIT were compared for their ability to detect advanced colorectal neoplasia (CRC + CRA, size ≥ 10 mm) (Table 4). In early studies, qualitative FIT Hemeselect (Beckman Coulter Inc., CA, USA) displayed moderate superiority in terms of sensitivity over unrehydrated gFOBt Hemoccult-II (Beckman Coulter Inc., CA, USA) without clear improved specificity while it displayed lower sensitivity but superior specificity (RFP 0.4–0.5) compared with rehydrated Hemoccult SENSA (Beckman Coulter Inc., CA, USA). More recently, qualitative FIT ImmunoCare-C (CARE diagnostic, Austria, cut-off 50 ng Hb / mL buffer) was observed to have both superior sensitivity and specificity compared with Hemoccult-II, while another qualitative FIT, Prevent ID CC (Preventis GmbH, Germany, cut-off 10 ng Hb/mL buffer), was reported to have improved sensitivity but inconsistent specificity results (RFP 0.3–6.2). HemSp/MagStream HT (Fujirebio Inc. Japan) has reportedly been assessed at different thresholds; however, its measurement method does not allow for linear quantitative measurement and only results obtained using the manufacturer's instructions cut-off of 100 µg Hb/g faeces could therefore be considered for comparative purposes.¹⁶ At this threshold, HemSp/MagStreamHT was found to have superior sensitivity but poorer specificity than Hemoccult-II in two different settings,^21,58 see Table 4 for details). In recent years, quantitative FITs Hemoglobin-Haptoglobin ELISA (Immundiagnostik AG, Bensheim, Germany), FOB Gold (Sentinel Diagnostics CH. SpA, Italy), and OC-Sensor (Eiken Chemical Co., Ltd, Japan) have been developed, which show superiority over unrehydrated Hemoccult-II in terms of sensitivity. Hb-Hp ELISA also showed superior specificity in one study,⁵⁶ while relative specificity of OC-Sensor to Hemoccult-II appeared to vary greatly in the hands of separate investigators (ie. compare Park et al.²¹ and Faivre et al.⁶¹ at 30 µg Hb/g faeces, Table 4), and that of FOB Gold was poorer at two very different thresholds (ie. compare Graser et al.³⁸ and Faivre et al.⁶¹ at 30 µg Hb/g faeces , Table 4). OC-Sensor also showed higher sensitivity for the detection of combined advanced and early neoplasia, and for the detection of early adenomas compared with Hemoccult-II; specificity was still inconsistent between studies (Webappendix Tables A3.3–A3.4).

Table 4.

Comparison of gFOBt and FIT for their ability to detect advanced colorectal neoplasia (CRC + CRA ≥ 10 mm):ratio of sensitivities (RSN) and ratio of false-positive rates (RFP).

Where data were available, the proportional odds of true positives were derived and random-effects meta-analysis results suggested that, irrespective of the type of guaiac or immunochemical method used, FIT sensitivity was overall superior to gFOBt for the detection of CRC alone, advanced colorectal neoplasia, or for the detection of any neoplasia (odds for CRC and all-size CRAs = 4.8, 95% CI 3.0–7.5; Figure 6 and web appendix Table A3.5).

Figure 6.

Comparison of gFOBt and FIT for their ability to detect colorectal neoplasia: meta-analysis of odds of positive FIT-only versus gFOBt positive-only. NS: Not Specified, NA: Not Applicable.

Influence of tumour site

The results obtained for HemSp/MagStream HT (100 µg Hb/g faeces, Guittet et al.⁵⁹) and OC-Sensor (cut-offs lower than 20 µg Hb/g faeces²¹) suggest little difference between CRC locations for advanced neoplasia: superior sensitivity and specificity were observed both for cancer of the proximal and distal colon (web appendix Tables A3.6 and A3.7). However, Guittet et al.⁵⁹ reported increased sensitivity of HemSp/MagStream HT (100 µg Hb/g faeces) for the detection of distal versus proximal early CRA (size < 10 mm, RSN = 5.1 versus 3.9).

Influence of sex

The influence of sex on the comparative performance of the tests to detect advanced colorectal neoplasia has not been investigated with enough power in any of the diagnostic cohort studies reviewed. A recent study nested within the German screening colonoscopy programme reported a significantly greater sensitivity of Hemoccult-II gFOBt in men than in women, while the opposite was observed for specificity.⁶² ImmoCARE-C (50 ng Hb/mL buffer), one of the FIT shown to have better overall sensitivity and specificity compared with Hemoccult-II⁴⁹ also showed the same trend. In contrast, a recent Spanish study found the odds ratio of having a false positive result to be at least 2-fold higher (OR = 2.82, 95% CI 2.16 – 3.68) in women compared with men when using either of the FOBt approaches, Hemascreen (Immunostics, Spain), and OC-Sensor (20 µg Hb/g faeces).⁶³

Prophylactic aspirin in the general population

An overview of the parallel group RCTs included in this review of the effect of aspirin in average-risk populations is presented in Table 5.

Table 5.

Characteristics overview of RCTs comparing aspirin versus control treatment in average-risk individuals with respect to incidence of, and mortality from, serious vascular and malignant events. .

Overall incidence & mortality from colorectal neoplasia: Influence of sex and tumour site

A significant decrease in CRC incidence was observed when the available long-term follow-up data (10 + years, range 12–23) were pooled in a random-effects meta-analysis (crude RR = 0.86, 95% CI 0.71–1.00; Figure 7 and web appendix Table A4.1). Both the men-only UK “British Doctors' Aspirin Trial” (BDAT) and the women-only United States “Women's Health Study” (WHS) reported significant reduction, but only after 10 or more years of follow-up (BDAT 10 years or more HR = 0.64, 95% CI 0.42–0.97⁶⁴; WHS post 10-year trial HR = 0.57, 95% CI 0.40–0.82⁶⁹), while the men-only United States “Physicians’ Health Study” (PHS) did not after a mean follow-up of 12 years. Colorectal endoscopic screening was ruled out as an underlying cause for the difference between treatment groups in the WHS trial.⁶⁹

Figure 7.

Random-effects meta-analysis of RCTs comparing aspirin versus control treatment in average-risk individuals with respect to incidence of CRC events.

Results on the effect of aspirin on CRC incidence by site were not directly available for male BDAT and PHS trials. However, a meta-analysis by Rothwell et al.,⁶⁵ which pooled data from the BDAT trial with three other trials that had recruited male patients at higher risk of vascular disease, revealed that allocation to aspirin for 5 years or longer significantly reduced the risk of colon cancer (ITT adjusted HR = 0.75, 95% CI 0.58–0.97) due to a dramatic reduction in the risk of proximal colon cancer (ITT adjusted HR = 0.35, 95% CI 0.20–0.63), but no reduction in distal colon cancer risk (ITT adjusted HR = 1.14, 95% CI 0.69–1.86). The risk of rectal cancer was also significantly reduced (ITT adjusted HR = 0.58, 95% CI 0.36–0.92). Those results were directly related to findings on CRC mortality by site in the same meta-analysis, ie. significantly reduced mortality from proximal (ITT adjusted HR = 0.24, 95% CI 0.11–0.52), but not from distal (ITT adjusted HR = 1.24, 95% CI 0.58–2.65) colon cancer, and mortality from rectal cancer reduced by over 50% (ITT adjusted HR = 0.47, 95% CI 0.26–0.87). In the female WHS trial, no significant decrease in CRC incidence was observed at any site at the end of the 10-year aspirin treatment (ITT adjusted HR = 0.92, 95% CI 0.71–1.21, in colon; ITT adjusted HR = 0.86, 95% CI 0.60–1.25, in proximal colon; ITT adjusted HR = 0.94, 95% CI 0.63–1.40, in distal colon; ITT adjusted HR = 1.20, 95% CI 0.70–2.04, in rectum).⁶⁸ However, at 17-year follow-up, the observed significant reduction in CRC incidence was primarily due to a reduction in proximal colon cancer (HR = 0.72, 95% CI 0.54–0.96).⁶⁹

Allocation to aspirin resulted in a non-significant reduction in CRC deaths at a median 23-year follow-up in the male-only BDAT trial (ITT non-adjusted OR = 0.73, 95% CI 0.49–1.10⁶⁵), while the female-only WHS trial found no evidence that would suggest aspirin to be effective in reducing CRC mortality after 10 or 17-year follow-up (n = 39,876; HRs not reported;^68,69 web appendix Table A4.2).

Incidence of & mortality from other malignancies

There was no evidence of an effect of aspirin on incidence of any invasive neoplasm in either sex (web appendix Figure A4.1 and Table A4.1), although both BDAT and WHS reported a non-significant effect of aspirin on mortality from invasive neoplasm (web appendix Figure A4.1 and Table A4.2, pooled RR = 0.92, 95% CI 0.80–1.10).

Incidence of & mortality from vascular events

We assessed the effect of aspirin treatment in the average-risk population RCTs in terms of major vascular events (Figure 8 and web appendix Tables A4.1–A4.2).

Figure 8.

Random-effects meta-analysis of RCTs comparing aspirin versus control treatment in average-risk individuals with respect to incidence of, and mortality from serious vascular and malignant events.

Our random-effects meta-analysis of crude risk ratios suggests an overall near significant increase in incidence of (30%) and mortality from intracranial bleeds (70%, Figure 8a). Aspirin had no impact on ischaemic strokes incidence and mortality in men, but was effective in reducing ischaemic stroke incidence in the WHS trial (Figure 8b). Overall, there was no significant benefit of aspirin on myocardial infarction (MI) incidence and mortality, although the male-only PHS trial reported a significant decrease in incidence (48%) and a non-significant decrease in mortality (64%, Figure 8c). Incidence of major extracranial bleeds (mainly GI haemorrhages) was overall significantly increased (40%) - although one of the male-only trial (BDAT) did not report any difference – while mortality from extracranial bleeds was decreased by 35%: the excess bleeds tended to be non-fatal bleeds (Figure 8d). At the end of the treatment period, the overall relative risk of fatal plus non-fatal major vascular events (strokes, ischaemic heart disease, and major extra-cranial bleeds) was 1.02 (95% CI 0.84–1.24) for the BDAT trial, 0.83 (95% CI 0.72–0.97) for the PHS trial, and 0.99 (95% CI 0.88–1.12) for the WHS trial (web appendix Table A4.1).

Discussion

The results of our intention-to-treat random-effects meta-analysis confirm that once-only FS screening performed in average-risk individuals aged 50+ at randomization significantly decreased CRC incidence by 18% and mortality by 28% although the study with shorter follow-up (NORCCAP1) contributed to significant heterogeneity in the ITT incidence data.

The effect of once-only FS on incidence of CRC occurring in the distal bowel was stronger than that on incidence in the proximal (31% versus 9% reduction), and was accompanied by a significant decrease in mortality from distal CRC but not from proximal CRC (42% versus 5%), although the evidence for this mainly came from one study (PLCO). The ITT effect was the same for both sexes, although women who underwent the once-only FS intervention were reported to have a greater reduction in CRC incidence compared with men who undertook the intervention in one trial (SCORE, 21% versus 45%), while compliance was reported to be higher in men than in women in two of the other trials analyzed (ie. UKFSST⁷⁰and PCLO⁷¹). In contrast, a significant reduction in CRC mortality (34%, ITT analysis) was only observed for the male population in the PLCO trial which reported on this outcome.

This could be related to the fact that men do develop more rectal cancers than women⁴, and women are known to develop more proximal tumours with microsatellite instability with age,⁴ which may have led to proportionally fewer lesions being identified by FS in women. The PLCO trial recruited patients up to the age of 74 in contrast to the other trials (upper limit of 64).

Significant 12% (ITT analysis) and 18% (‘per protocol’ analysis) decreases in CRC mortality were observed for biennial gFOBt screening after 14 to 20 years of follow-up. gFOBt screening was equally effective at reducing CRC mortality for both proximal and distal cancers (19% and 15% decrease, respectively), and in both male and female subjects (16% and 17% decrease, respectively). ‘Per protocol’ estimates were not available for each sex separately, but although female subjects have a lower CRC risk compared with men in the UK, they have been reported to be more likely to return their gFOBt kit.^14,72 Biennial gFOBt screening did not, however, demonstrate any effectiveness in decreasing CRC incidence in our meta-analysis of average-risk individuals aged 45+ at randomization. In contrast, FIT has been shown to reduce CRC incidence by over 20% in a large Italian cohort study after 11-year follow-up.⁷³

Endoscopic screening with once-only FS at age 50+ has been found to be more sensitive than both guaiac rehydrated⁷⁴ or unrehydrated^19,75–77) Hemoccult-II, and qualitative faecal immunochemical tests for haemoglobin (OC-Sensor,^19,77 Immudia-HemSp^78,79) for the detection of prevalent advanced CRA.⁹ The results from the long-term RCTs on the effect of the two screening approaches on incidence of CRC confirmed this observation.

Compliance with screening is an important factor in effectiveness of a screening programme. In the RCTs, compliance at first screen for gFOBt ranged from 57 to 67% (see Table 2) and for FS from 58 to 83.5% (see Table 1). For the latter, compliance in the general population is likely to be lower than among those expressing an interest to participate, while for repeated gFOBt screening compliance may diminish over time, although the UK pilot did not present evidence of substantial fall in compliance over three rounds.⁸⁰ Differences in compliance between sexes, by age and in different ethnic groups, are important and should be considered in relation to differences in risk. ‘Per protocol’ analyses of the RCTs are therefore valuable for comparisons of different modalities and subgroups. When ‘per protocol’ estimates were compared, once-only FS was still superior to repeated gFOBt screening with respect to CRC mortality reduction (42% versus 18% decrease).

Long-term studies of the impact of adding once-only FS to repeated FOBt on CRC incidence and mortality are not yet available. One trial reported significantly fewer CRC cases detected in the ‘gFOBt, plus FS group’ compared with the group allocated to ‘gFOBt only’ in the 24–62 months following the prevalent screen.⁷⁶ In a review of cost-effectiveness studies, no one screening strategy emerged as the most effective in terms of life-years saved, or as the preferred strategy for a given willingness to pay; discrepancies between modelling studies were mainly due to differences in the simulated duration of the adenoma-carcinoma sequence.⁸¹ Using a state transition modelling approach and newly available CRC incidence data in the absence of screening, Whyte et al.⁸² estimated that combining once-only FS at age 55 (based on UKFSST data) with biennial FIT age 60--74 (OC-Sensor with a threshold of 20 µg Hb/g faeces) would increase clinical (and economic) benefits over the use of a single screening modality in England. There was some evidence that once-only FS screening might benefit men more than women in terms of CRC mortality, probably due to the fact that men develop proportionally more distal tumours than women and develop advanced CRAs at an earlier age.³ The optimal age for the greatest quality-adjusted life year gain using once-only FS was estimated to be between 52 and 58.⁸²

Our comparison of gFOBt and FIT operating characteristics in community screening suggests that modern quantitative FIT OC-Sensor (Eiken Chemical Co., Ltd, Japan) outperforms gFOBt in terms of sensitivity while controlling for false-positives by allowing cut-off of haemoglobin concentrations to be set to a level acceptable to patients, but also according to local colonoscopy resources and health system resources, as well as according to patient characteristics, such as sex. This enhanced sensitivity was observed for CRC and advanced, as well as for early adenomas, although sensitivity was slightly worse for early adenomas, indicating a likely increase in the use of follow-up endoscopic resources. Results obtained so far with OC-Sensor do not suggest any difference in the performance of OC-Sensor compared with Hemoccult-II with respect to different parts of the colon.²¹ Some evidence is emerging suggesting that the tests operating characteristics may differ between sexes,^49,62,63 although further work is needed.

Comparative results of OC-Sensor versus Hemoccult-II sensitivity and specificity were consistent between diagnostic cohort studies and parallel group trials; sensitivity was found to be superior to both unrehydrated Hemoccult-II (RSN > 2.0 at cut-offs of 10 and 20 µg Hb/g faeces,^18,19 respectively) and Hemoccult SENSA (RSN > 1.5 at a cut-off of 14 µg Hb/g faeces,⁸³ rehydration status not reported), while specificity was inferior at those thresholds (RFP > 1.0).

Compliance at first screening round has generally been reported to be higher for FIT compared with gFOBt as assessed using meta-analysis of RCTs by Vart et al.⁸⁴ and by others^85,86, while Birkenfeld et al.⁸⁷ found overall comparable compliance for the two methods with taking the kits favouring gFOBT and kit return favouring FIT. Finally, Chubak et al.⁸⁸ found uptake of 1-stool OC-Sensor to be significantly higher than 2-stool InSureFIT and 3-stool Hemoccult-SENSA. A Cochrane review to compare diagnostic accuracy of gFOBT and FIT screening for the detection of advanced colorectal neoplasia in average-risk individuals is currently being undertaken.³⁰ A strategy of biennial FIT (OC-Sensor), age 60–74 was found to be cost saving and to result in quality-adjusted life year gains compared with the current screening programme of biennial gFOBT age 60–74.⁸² However, the results were highly sensitive to FIT threshold model parameters with higher thresholds of 30 and 40 µg Hb/g faeces associated with lower quality adjusted life year (QALY) gains but little difference in costs.

The efficacy of aspirin to prevent both non-fatal and fatal thrombotic events in patients at high-risk of vascular disease is well established.⁸⁹ Recently, several meta-analyses of RCTs have reported long-term treatment (>5 years) with aspirin to reduce CRC incidence and mortality after a latent period of about 10 years when using data from 51 RCTs initially set up to assess the effect of aspirin on vascular endpoints. Most RCTs were performed in populations at higher risk of vascular disease, with prior experience of a vascular event (secondary prevention trial), or in populations suffering from diabetes.^65,90,91 The effect was particularly evident for cancers of the proximal colon for which a significant 65% reduction was reported.⁶⁵

Our review of RCTs performed in average-risk individuals suggests a protective effect of aspirin on the incidence of CRC in both male and female individuals after at least 5 year of treatment and 10 years of follow-up, although relative estimates were quite variable. The large study of female individuals reported a significant 14% reduction in proximal colon⁶⁹, which was much smaller than the estimate reported for populations at high-risk of vascular disease.⁶⁵

Only one of the two large RCTs performed in males reported data on CRC mortality. This study showed a non-significant 27% reduction with a more likely favourable outcome for proximal CRC, while the only large female RCT reported no significant effect (exact ratios not reported). The 27% reduction is much less than the 42% reduction reported in the meta-analysis of males at higher risk of vascular disease.⁹¹ In a recent large cohort study of healthy individuals, CRC mortality was only modestly reduced in patients taking aspirin (adjusted HR after ≥ 5 years of use = 0.92, 95% CI 0.83 to 1.02⁹²).

The effect of aspirin in average-risk individuals appeared to be quite specific to CRC as the RCTs did not find an overall benefit of aspirin on incidence of or mortality from all cancers after a follow-up of at least 10 years, although other primary tumours of the gastro-intestinal tract have been reported to be sensitive to aspirin prophylaxis.^69,90 Those findings are in contrast to the reduced incidence of all cancers reported⁹¹ for both men and women in a meta-analysis of 12 RCTs for primary prevention of vascular events in individuals of various health status (OR = 0.76, 95% CI 0.66 – 0.88).

Overall, the effects of aspirin treatment on risk of major vascular events (strokes, ischaemic heart disease, and major extra-cranial bleeds) suggested (i) limited benefit of aspirin on primary prevention of vascular events in average-risk individuals, but also (ii) limited harm due to major haemorrhages.⁹³ These observations are in agreement with the results by Rothwell et al.⁹¹ which showed no significant difference between low-dose daily aspirin and control treatment on risk of vascular death, the trials' primary endpoint, in individuals of various health status (ITT unadjusted OR = 0.93, 95% CI 0.70–1.23). These observations were also in agreement with the results from a review of recent meta-analyses of primary prevention trials⁹⁴ showing no convincing evidence that aspirin reduces cardiovascular mortality although it had a modest significant reduction on total mortality.

The risk of bleeding associated with aspirin use has however been shown to be particularly high in individuals younger than 50 years in a prospective study with median follow-up of 5.7 years,⁹⁵ suggesting that the balance of benefits and harms may vary with life stages. In addition, a meta-analysis by the Antithrombotic Trialists’ (ATT) Collaboration⁹⁶ which used individual data from the BDAT, PHS, and WHS trials, among others, showed that aspirin increased the risk of GI bleeding in non-smokers compared with current smokers, and in patients with relatively lower 5-year predicted risk of coronary heart disease (<5%). Interestingly, the benefit of aspirin on the risk of fatal adenocarcinoma was also reported to be greatest in smokers.⁹⁷

Finally, other adverse effects of aspirin such as on renal function and the control of blood pressure^98,99 and, to a lesser extent, on gut lining integrity and dyspepsia¹⁰⁰ as well as the discomfort caused by minor bleeding events also need to be taken into consideration.

Sequestration of angiogenesis regulatory proteins has recently been shown to be elevated in platelets of CRC patients (all stages) compared with healthy individuals,¹⁰¹ and thrombocytosis (high platelet count) has been found to be associated with poor prognosis in colon cancer.¹⁰² However, whether thrombocytosis is causal or consequent to cancer development is to be determined before we can understand whether healthy people with elevated platelet levels are at higher risk of developing cancer, and/or metastasis.

The risk of vascular events was found to be significantly higher in non-responders to aspirin compared with responders in a meta-analysis of observational studies performed in patients at higher risk of vascular events.¹⁰³ High platelet reactivity to arachidonic acid, an indirect COX-1-dependent pathway, has been associated with cardiac adverse events in a prospective cohort study of patients undergoing percutaneous coronary intervention at 12-month follow-up¹⁰⁴, but not in another where COX-1 independent platelet function correlated with subsequent major adverse cardiovascular events at 2-year follow-up.¹⁰⁵ Whether similar patterns will be observed with respect to incidence of CRC in average-risk individuals remains to be elucidated.

The net clinical benefit over the risks presented by prolonged aspirin intake will need to be assessed on an individual basis taking into consideration each patient's risk factors.

Conclusion

Evidence for a different CRC aetiology between men and women is mounting. RCTs of biennial gFOBt screening starting at age 60 did not show a differential benefit in terms of CRC incidence or mortality between sexes, while a recent cross-sectional analysis suggests the introduction of gFOBt screening in the UK is having a greater differential impact on male than on female mortality,¹⁰⁶ in spite of lower compliance in men¹³ (see web appendix Table A2.10). In contrast, preliminary results from once-only FS screening trials performed at age 55 or over suggest greater benefit for men in term of mortality, although the evidence is still sparse. Since March 2013, the NHSBCSP has started rolling out a once-only FS programme in six areas in England where patients are invited for ‘bowel scope screening' around the time of their 55th birthday. In time, this will provide more insight into male-female potential differences in screening benefit. The optimal age for screening initiation has been proposed to be around 5 years higher for women than for men within average-risk populations, based on 10-year CRC incidence and mortality¹⁰⁷, but this assumes that screening is equally effective in both sexes, which we know not to be the case, men and women having different incidence of proximal and distal lesions and screening modalities preferentially detecting lesions located in right or left bowel. Finally, long-term effects of aspirin treatment on CRC incidence has now been shown in both men and women at average-risk of CRC or vascular events who started treatment aged 40 or over with particular benefit for proximal CRC.

Age at menopause and consequent sex hormone changes in women may be an important factor when it comes to choosing the optimal window for implementing the different approaches available for primary CRC prevention. Age- and sex-specific vascular risk factors will also have to be taken into account when considering aspirin chemoprevention as part of a national CRC prevention programme. Further investigation into the effectiveness of the various screening modalities by sex and site of lesions is warranted.

Footnotes

SWD and NJM contributed to this work as part of the programme of the Policy Research Unit in Cancer Awareness, Screening and Early Diagnosis. The Policy Research Unit in Cancer Awareness, Screening, and Early Diagnosis receives funding for a research programme from the Department of Health Policy Research Programme. It is a collaboration between researchers from seven institutions (Queen Mary University of London, UCL, King’s College London, London School of Hygiene and Tropical Medicine, Hull York Medical School, Durham University and Peninsula Medical School).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Cancer Research UK, Bowel Cancer Statistics [http://www.cancerresearchuk.org/cancer-info/cancerstats/types/bowel/].

Nguyen

Bent

Chen

. Gender as a risk factor for advanced neoplasia and colorectal cancer: a systematic review and meta-analysis. Clin Gastroenterol Hepatol 2009; 7(6): 676–681.

Ferlitsch

MRKPS

. Sex-specific prevalence of adenomas, advanced adenomas, and colorectal cancer in individuals undergoing screening colonoscopy. JAMA: The Journal of the American Medical Association 2011; 306(12): 1352–1358.

Koo

Leong

RWL

. Sex differences in epidemiological, clinical and pathological characteristics of colorectal cancer. Journal of Gastroenterology and Hepatology 2010; 25(1): 33–42.

Jones

Morris

Thomas

. Evaluation of bowel cancer registration data in England, 1996-2004. Br J Cancer 2009; 101(8): 1269–1273.

Curado

Edwards

Shin

. Cancer Incidence in Five Continents. IARC Scientific Publications 2007; IX(160): 466–481.

Distler

Holt

. Are right- and left-sided colon neoplasms distinct tumors? Digestive diseases 1997; 15(4--5): 302–311.

Leslie

Carey

Pratt

. The colorectal adenoma–carcinoma sequence. British Journal of Surgery 2002; 89(7): 845–860.

Littlejohn

Hilton

Macfarlane

. Systematic review and meta-analysis of the evidence for flexible sigmoidoscopy as a screening method for the prevention of colorectal cancer. British Journal of Surgery 2012; 99(11): 1488–1500.

10.

Hewitson

Glasziou

Watson

. Cochrane systematic review of colorectal cancer screening using the fecal occult blood test (Hemoccult): An update. American Journal of Gastroenterology 2008; 103(6): 1541–1549.

11.

Hewitson

Glasziou

Irwig

. Screening for colorectal cancer using the faecal occult blood test, Hemoccult. Cochrane database of systematic reviews (Online) 2007; Jan 24(1): CD001216–CD001216.

12.

Mandel

Church

Bond

. The Effect of Fecal Occult-Blood Screening on the Incidence of Colorectal Cancer. New England Journal of Medicine 2000; 343(22): 1603–1607.

13.

Steele

Kostourou

McClements

. Effect of gender, age and deprivation on key performance indicators in a FOBT-based colorectal screening programme. J Med Screen 2010; 17(2): 68–74.

14.

Logan

Patnick

Nickerson

. Outcomes of the Bowel Cancer Screening Programme (BCSP) in England after the first 1 million tests. Gut 2012; 61(10): 1439–1446.

15.

Lansdorp-Vogelaar

van Ballegooijen

Boer

. A novel hypothesis on the sensitivity of the fecal occult blood test. Cancer 2009; 115(11): 2410–2419.

16.

Halloran

Launoy

Zappa

. International Agency for Research on C. European guidelines for quality assurance in colorectal cancer screening and diagnosis. First Edition–Faecal occult blood testing. Endoscopy 2012; 44(Suppl 3): SE65–87.

17.

Rabeneck

Rumble

Thompson

. Fecal immunochemical tests compared with guaiac fecal occult blood tests for population-based colorectal cancer screening. Can J Gastroenterol 2012; 26(3): 131–147.

18.

van Rossum

van Rijn

Laheij

. Random comparison of guaiac and immunochemical fecal occult blood tests for colorectal cancer in a screening population. Gastroenterology 2008; 135(1): 82–90.

19.

Hol

van Leerdam

van Ballegooijen

. Screening for colorectal cancer: randomised trial comparing guaiac-based and immunochemical faecal occult blood testing and flexible sigmoidoscopy. Gut 2010a; 59(01): 62–68.

20.

Allison

Sakoda

Levin

. Screening for colorectal neoplasms with new fecal occult blood tests: update on performance characteristics. Journal of the National Cancer Institute 2007; 99(19): 1462–1470.

21.

Park

Ryu

Kim

Y-H

. Comparison of Guaiac-based and quantitative immunochemical fecal occult blood testing in a population at average risk undergoing colorectal cancer screening. Am J Gastroenterol 2010; 105(9): 2017–2025.

22.

Mantovani

Allavena

Sica

. Cancer-related inflammation. Nature 2008; 454(7203): 436–444.

23.

Greenhough

Smartt

HJM

Moore

. The COX-2/PGE2 pathway: key roles in the hallmarks of cancer and adaptation to the tumour microenvironment. Carcinogenesis 2009; 30(3): 377–386.

24.

Hyde

CAC

Missailidis

. Inhibition of arachidonic acid metabolism and its implication on cell proliferation and tumour-angiogenesis. International Immunopharmacology 2009; 9(6): 701–715.

25.

Bretthauer

. Colorectal cancer screening. J Intern Med 2011; 270(2): 87–98.

26.

Rabeneck

Zwaal

Goodman

. Cancer Care Ontario guaiac fecal occult blood test (FOBT) laboratory standards: evidentiary base and recommendations. Clinical Biochemistry 2008; 41(16--17): 1289–1305.

27.

Basu

Smartt

. Comparison of diagnostic accuracy between immunochemical and guaiac based faecal occult blood tests for colorectal cancer detection. HSAC Report 2009; 2(5):

28.

van Dam

Kuipers

van Leerdam

. Performance improvements of stool-based screening tests. Best Practice & Research Clinical Gastroenterology 2010; 24(4): 479–492.

29.

Zhu

Nie

. Comparison of immunochemical and guaiac-based fecal occult blood test in screening and surveillance for advanced colorectal neoplasms: A meta-analysis. Journal of Digestive Diseases 2010; 11(3): 148–160.

30.

van Roon A, van Dam L, Zauber A, et al. Guaiac-based faecal occult blood tests versus faecal immunochemical tests for colorectal cancer screening in average-risk individuals (Protocol). The Cochrane Collaboration 2011, http://summaries.cochrane.org/CD009276/.

31.

Egger M, Davey Smith G, Altman D. Systematic reviews in health care: meta-analysis in context. BMJ Books 2001, 2nd ed.

32.

Cuzick

Edwards

Segnan

. Adjusting for non-compliance and contamination in randomized clinical trials. Statistics in Medicine 1997; 16(9): 1017–1029.

33.

Stryhn

Christensen

. Confidence intervals by the profile likelihood method, with applications in veterinary epidemiology [Abstract]. International Symposia on Veterinary Epidemiology and Economics (ISVEE), Vina del Mar, Chile 2003.

34.

Cheng

Macaluso

. Comparison of the accuracy of two tests with a confirmatory procedure limited to positive results. Epidemiology 1997; 8(1): 104–106.

35.

Sasieni

. Estimating prevalence when the true disease status is incompletely ascertained. Statistics in Medicine 2001; 20(6): 935–949.

36.

Smith

Young

Cole

. Comparison of a brush-sampling fecal immunochemical test for hemoglobin with a sensitive guaiac-based fecal occult blood test in detection of colorectal neoplasia. Cancer 2006; 107(9): 2152–2159.

37.

Cheng

Wong

Hong

. Colorectal cancer screening in asymptomatic adults: Comparison of colonoscopy, sigmoidoscopy and fecal occult blood tests. J Formos Med Assoc 2002; 101(10): 685–690.

38.

Graser

Stieber

Nagel

. Comparison of CT colonography, colonoscopy, sigmoidoscopy and faecal occult blood tests for the detection of advanced adenoma in an average risk population. Gut 2009; 58(2): 241–248.

39.

Thiis-Evensen

Hoff

Sauar

. Population-based surveillance by colonoscopy: effect on the incidence of colorectal cancer: telemark polyp study I. Scandinavian Journal of Gastroenterology 1999; 34(4): 414–420.

40.

Hoff

Grotmol

Skovlund

. Risk of colorectal cancer seven years after flexible sigmoidoscopy screening: randomised controlled trial. BMJ 2009; 338: b1846–b1846.

41.

Atkin

Edwards

Kralj-Hans

. Once-only flexible sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial. The Lancet 2010; 375(9726): 1624–1633.

42.

Segnan

Armaroli

Bonelli

. Once-only sigmoidoscopy in colorectal cancer screening: follow-up findings of the Italian Randomized Controlled Trial–SCORE. J Natl Cancer Inst 2011; 103(17): 1310–1322.

43.

Schoen

Pinsky

Weissfeld

. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy. The New England journal of medicine 2012; 366(25): 2345–2357.

44.

Whitlock

Lin

Liles

. Screening for colorectal cancer: a targeted, updated systematic review for the U.S. Preventive Services Task Force. Ann Intern Med 2008; 149(9): 638–658.

45.

Hewitson

. Review: faecal occult blood test screening reduces risk of colorectal cancer mortality. Evidence Based Medicine 2009; 14(1): DOI: 10.1002/14651858.

46.

Scholefield

Moss

Mangham

. Nottingham trial of faecal occult blood testing for colorectal cancer: a 20-year follow-up. Gut 2012; 61(7): 1036–1040.

47.

Kronborg

Jorgensen

Fenger

. Randomized study of biennial screening with a faecal occult blood test: Results after nine screening rounds. Scandinavian Journal of Gastroenterology 2004; 39(9): 846–851.

48.

Lindholm

Brevinge

Haglind

. Survival benefit in a randomized clinical trial of faecal occult blood screening for colorectal cancer. British Journal of Surgery 2008; 95(8): 1029–1036.

49.

Hundt

Haug

Brenner

. Comparative evaluation of immunochemical fecal occult blood tests for colorectal adenoma detection. Ann Intern Med 2009; 150(3): 162–169.

50.

Wong

Fedorak

Prosser

. The sensitivity and specificity of guaiac and immunochemical fecal occult blood tests for the detection of advanced colonic adenomas and cancer. Int J Colorectal Dis 2012; 27(12): 1657–1664.

51.

Parker

Robinson

MHE

Scholefield

. Psychiatric morbidity and screening for colorectal cancer. Journal of Medical Screening 2002; 9(1): 7–10.

52.

Petrelli

Michalek

Freedman

. Immunochemical versus guaiac occult blood stool tests: results of a community-based screening program. Surgical Oncology 1994; 3(1): 27–36.

53.

Robinson

MHE

Marks

Farrands

. Population screening for colorectal cancer: Comparison between guaiac and immunological faecal occult blood tests. British Journal of Surgery 1994; 81(3): 448–451.

54.

Allison

Tekawa

Ransom

. A Comparison of Fecal Occult-Blood Tests for Colorectal-Cancer Screening. New England Journal of Medicine 1996; 334(3): 155–160.

55.

Castiglione

Zappa

Grazzini

. Immunochemical vs guaiac faecal occult blood tests in a population-based screening programme for colorectal cancer. Br J Cancer 1996; 74(1): 141–144.

56.

Hoepffner

Shastri

Hanisch

. Comparative evaluation of a new bedside faecal occult blood test in a prospective multicentre study. Aliment Pharmacol Ther 2006; 23(1): 145–154.

57.

Dancourt

Lejeune

Lepage

. Immunochemical faecal occult blood tests are superior to guaiac-based tests for the detection of colorectal neoplasms. European Journal of Cancer 2008; 44(15): 2254–2258.

58.

Guittet

Bouvier

Mariotte

. Comparison of a guaiac based and an immunochemical faecal occult blood test in screening for colorectal cancer in a general average risk population. Gut 2007; 56(2): 210–214.

59.

Guittet

Bouvier

Mariotte

. Comparison of a guaiac and an immunochemical faecal occult blood test for the detection of colonic lesions according to lesion type and location. Br J Cancer 2009; 100(8): 1230–1235.

60.

Parra-Blanco

Gimeno-García

Quintero

. Diagnostic accuracy of immunochemical versus guaiac faecal occult blood tests for colorectal cancer screening. Journal of Gastroenterology 2010; 45(7): 703–712.

61.

Faivre

Dancourt

Denis

. Comparison between a guaiac and three immunochemical faecal occult blood tests in screening for colorectal cancer. European Journal of Cancer 2012; 48(16): 2969–2976.

62.

Brenner

Haug

Hundt

. Sex Differences in Performance of Fecal Occult Blood Testing. Am J Gastroenterol 2010a; 105(11): 2457–2464.

63.

Garcia

Mila

Binefa

. False-positive results from colorectal cancer screening in Catalonia (Spain), 2000--2010. J Med Screen 2012; 19(2): 77–82.

64.

Flossmann

Rothwell

. Effect of aspirin on long-term risk of colorectal cancer: consistent evidence from randomised and observational studies. The Lancet 2007; 369(9573): 1603–1613.

65.

Rothwell

Wilson

Elwin

C-E

. Long-term effect of aspirin on colorectal cancer incidence and mortality: 20-year follow-up of five randomised trials. The Lancet 2010; 376(9754): 1741–1750.

66.

Gann

Manson

Glynn

. Low-Dose Aspirin and Incidence of Colorectal Tumors in a Randomized Trial. Journal of the National Cancer Institute 1993; 85(15): 1220–1224.

67.

Sturmer

Glynn

Lee

I-M

. Aspirin Use and Colorectal Cancer: Post-Trial Follow-up Data from the Physicians' Health Study. Annals of Internal Medicine 1998; 128(9): 713–720.

68.

Cook

Lee

I-M

Gaziano

. Low-Dose Aspirin in the Primary Prevention of Cancer. JAMA: The Journal of the American Medical Association 2005; 294(1): 47–55.

69.

Cook NR: Aspirin and cancer: Evidence from randomized trials. 8th NCRI Cancer Conference 4–7 November 2012 2012, Parallel sessions Abstract 89.

70.

Wardle

Miles

Atkin

. Gender differences in utilization of colorectal cancer screening. Journal of Medical Screening 2005; 12(1): 20–27.

71.

Weissfeld

Schoen

Pinsky

. Flexible Sigmoidoscopy in the PLCO Cancer Screening Trial: Results From the Baseline Screening Examination of a Randomized Trial. Journal of the National Cancer Institute 2005; 97(13): 989–997.

72.

Vart

Snowball

. The impact of media coverage of the Flexible Sigmoidoscopy Trial on English colorectal screening uptake. Journal of Medical Screening 2012; 19(2): 83–88.

73.

Ventura L, Mantellini P, Grazzini G: The effect of introduction of immunochemical faecal occult blood testing on colorectal cancer incidence. International Cancer Screening Network Biennial Meeting, Sydney Australia October 23--25 2012 2012, Poster 51.

74.

Brevinge

Lindholm

Buntzen

. Screening for colorectal neoplasia with faecal occult blood testing compared with flexible sigmoidoscopy directly in a 55 – 56 years' old population. Int J Colorect Dis 1997; 12(5): 291–295.

75.

Berry

Clarke

Hardcastle

. Randomized trial of the addition of flexible sigmoidoscopy to faecal occult blood testing for colorectal neoplasia population screening. British Journal of Surgery 1997; 84(9): 1274–1276.

76.

Rasmussen

Kronborg

Fenger

. Possible advantages and drawbacks of adding flexible sigmoidoscopy to hemoccult-II in screening for colorectal cancer. A randomized study. Scand J Gastroenterol 1999; 34(1): 73–78.

77.

Hol

de Jonge

van Leerdam

. Screening for colorectal cancer: Comparison of perceived test burden of guaiac-based faecal occult blood test, faecal immunochemical test and flexible sigmoidoscopy. European Journal of Cancer 2010b; 46(11): 2059–2066.

78.

Segnan

Senore

Andreoni

. Randomized Trial of Different Screening Strategies for Colorectal Cancer: Patient Response and Detection Rates. Journal of the National Cancer Institute 2005; 97(5): 347–357.

79.

Segnan

Senore

Andreoni

. Comparing Attendance and Detection Rate of Colonoscopy With Sigmoidoscopy and FIT for Colorectal Cancer Screening. Gastroenterology 2007; 132(7): 2304–2312.

80.

Moss

Campbell

Melia

. Performance measures in three rounds of the English bowel cancer screening pilot. Gut 2012; 61(1): 101–107.

81.

Lansdorp-Vogelaar

Knudsen

Brenner

. Cost-effectiveness of colorectal cancer screening - an overview. Best Pract Res Clin Gastroenterol 2010; 24(4): 439–449.

82.

Whyte

Chilcott

Halloran

. Reappraisal of the options for colorectal cancer screening in England. Colorectal disease: the official journal of the Association of Coloproctology of Great Britain and Ireland 2012; 14(9): e547–561.

83.

Levi

Birkenfeld

Vilkin

. A higher detection rate for colorectal cancer and advanced adenomatous polyp for screening with immunochemical fecal occult blood test than guaiac fecal occult blood test, despite lower compliance rate. A prospective, controlled, feasibility study. International Journal of Cancer 2011; 128(10): 2415–2424.

84.

Vart

Banzi

Minozzi

. Comparing participation rates between immunochemical and guaiac faecal occult blood tests: A systematic review and meta-analysis. Preventive Medicine 2012; 55(2): 87–92.

85.

Deutekom

van Rossum

LGM

van Rijn

. Comparison of guaiac and immunological fecal occult blood tests in colorectal cancer screening: The patient perspective. Scandinavian Journal of Gastroenterology 2010; 45(11): 1345–1349.

86.

Schottinger J, Kanter M, Palmer-Toy D. Reducing the risk of overpapulation: Achieving more by doing less. International Cancer Screening Network Biennial Meeting, Sydney Australia October 23--25 2012 2012, Poster 47.

87.

Birkenfeld

Belfer

Chared

. Factors affecting compliance in faecal occult blood testing: a cluster randomized study of the faecal immunochemical test versus the guaiac faecal occult test. Journal of Medical Screening 2011; 18(3): 135–141.

88.

Chubak J, Bogart A, Laing S, Green B: Comparing uptake of fecal occult blood tests for colorectal cancer screening: Results from a randomized controlled trial. International Cancer Screening Network Biennial Meeting, Sydney Australia October 23-25 2012 2012, Poster 28.

89.

Lordkipanidze

. Advances in monitoring of aspirin therapy. Platelets 2012; 23(7): 526–536.

90.

Rothwell

Fowkes

FGR

Belch

JFF

. Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials. The Lancet 2011; 377(9759): 31–41.

91.

Rothwell

Price

Fowkes

FGR

. Short-term effects of daily aspirin on cancer incidence, mortality, and non-vascular death: analysis of the time course of risks and benefits in 51 randomised controlled trials. The Lancet 2012a; 379(9826): 1602–1612.

92.

Jacobs

Newton

Gapstur

. Daily Aspirin Use and Cancer Mortality in a Large US Cohort. Journal of the National Cancer Institute 2012; 104(16): 1208–1217.

93.

Siller-Matula

. Hemorrhagic complications associated with aspirin: An underestimated hazard in clinical practice? JAMA: The Journal of the American Medical Association 2012; 307(21): 2318–2320.

94.

Raju

Eikelboom

. The aspirin controversy in primary prevention. Curr Opin Cardiol 2012; 27(5): 499–507.

95.

De Berardis

GLGDEA

. ASsociation of aspirin use with major bleeding in patients with and without diabetes. JAMA: The Journal of the American Medical Association 2012; 307(21): 2286–2294.

96.

Antithrombotic Trialists' (ATT) Collaboration. Aspirin in the primary and secondary prevention of vascular disease: collaborative meta-analysis of individual participant data from randomised trials. The Lancet 2009; 373(9678): 1849–1860.

97.

Rothwell

Wilson

Price

. Effect of daily aspirin on risk of cancer metastasis: a study of incident cancers during randomised controlled trials. The Lancet 2012b; 379(9826): 1591–1601.

98.

Sharma

Jawad

. Adverse Effects of COX-2 Inhibitors. TheScientificWorldJOURNAL 2005; 5: 629–645.

99.

Kawada

Moriyama

Kitamura

. Towards developing new strategies to reduce the adverse side-effects of nonsteroidal anti-inflammatory drugs. Clin Exp Nephrol 2012; 16(1): 25–29.

100.

Lanas

McCarthy

Voelker

. Short-Term Acetylsalicylic Acid (Aspirin) Use for Pain, Fever, or Colds – Gastrointestinal Adverse Effects: A Meta-Analysis of Randomized Clinical Trials. Drugs in R&D 2011b; 11(3): 277–288.

101.

Peterson

Zurakowski

Italiano

Jr . VEGF, PF4 and PDGF are elevated in platelets of colorectal cancer patients. Angiogenesis 2012; 15(2): 265–273.

102.

Gay

Felding-Habermann

. Contribution of platelets to tumour metastasis. Nat Rev Cancer 2011; 11(2): 123–134.

103.

Crescente

Di Castelnuovo

Iacoviello

. Response variability to aspirin as assessed by the platelet function analyzer (PFA)-100. A systematic review. Thrombosis and Haemostasis 2008; 99(1): 14–26.

104.

Siller-Matula

Delle-Karth

Christ

. Dual non-responsiveness to antiplatelet treatment is a stronger predictor of cardiac adverse events than isolated non-responsiveness to clopidogrel or aspirin. International Journal of Cardiology 2013; 167(2): 430–435.

105.

Frelinger

Linden

. Association of Cyclooxygenase-1-Dependent and -Independent Platelet Function Assays With Adverse Clinical Outcomes in Aspirin-Treated Patients Presenting for Cardiac Catheterization. Circulation 2009; 120(25): 2586–2596.

106.

McClements

Madurasinghe

Thomson

. Impact of the UK colorectal cancer screening pilot studies on incidence, stage distribution and mortality trends. Cancer Epidemiology 2012; 36(4): e232–e242.

107.

Brenner

Hoffmeister

Arndt

. Gender differences in colorectal cancer: implications for age at initiation of screening. Br J Cancer 2007; 96(5): 828–831.

Screening and Primary prevention of Colorectal Cancer: a Review of sex-specific and site-specific differences

Abstract

Background

Objective

Methods

Results

Conclusions

Keywords

Introduction

Methods

Data sources and study selection

Once-only FS screening in the general population

Selected studies

Biennial gFOBt screening in the general population

Selected studies

Comparison of gFOBt and FITcharacteristics

Selected studies

Prophylactic aspirin in the general population

Selected studies

Statistical methods

Once-only FS screening and biennial gFOBt screening in the general population

Comparison of gFOBt and FIT characteristics

Prophylactic aspirin in the general population

Results

Once-only Flexible Sigmoidoscopy screening in the general population

Overall incidence & mortality

Incidence & mortality by site

Incidence & mortality by sex

Potential harms

All-cause mortality

Biennial gFOBt screening in the general population

Incidence & mortality

Potential harms

All-cause mortality

Comparison of gFOBt and FIT characteristics in the general population

Overall comparison of test characteristics

Influence of tumour site

Influence of sex

Prophylactic aspirin in the general population

Overall incidence & mortality from colorectal neoplasia: Influence of sex and tumour site

Incidence of & mortality from other malignancies

Incidence of & mortality from vascular events

Discussion

Conclusion

Footnotes

Funding

References