132

ORIGINAL ARTICLE

Diagnostic accuracy of faecal occult blood tests used in screening for colorectal cancer: a systematic review J A Burch, K Soares-Weiser, D J B St John, S Duffy, S Smith, J Kleijnen and M Westwood .................................................................................................. J Med Screen 2007;14:132–137

See end of article for authors’ affiliations

............... Correspondence to: Jane Burch, Research Fellow, Centre for Reviews and Dissemination (CRD), University of York, York YO10 5DD, UK; [email protected] Accepted for publication 3 July 2007

...............

Objective To determine the accuracy of guaiac and immunochemical faecal occult blood tests (FOBTs) for the detection of colorectal cancer in an average-risk screening population. Methods Fifteen electronic databases, the internet, key journals and reference lists of included studies were searched. We included diagnostic accuracy studies that compared guaiac or immunochemical FOBTs with any reference standard, for the detection of colorectal cancer in an average-risk adult population, with sufficient data to construct a 2  2 table. Results Fifty-nine studies were included. Thirty-three evaluated guaiac FOBTs, 35 immunochemical FOBTs and one evaluated sequential FOBTs. Sensitivities for the detection of all neoplasms ranged from 6.2% (specificity 98.0%) to 83.3% (specificity 98.4%) for guaiac FOBTs, and 5.4% (specificity 98.5%) to 62.6% (specificity 94.3%) for immunochemical FOBTs. Specificity ranged from 65.0% (sensitivity 44.1%) to 99.0% (sensitivity 19.3%) for guaiac FOBTs, and 89.4% (sensitivity 30.3%) to 98.5% (sensitivity 5.4%) for immunochemical FOBTs. Diagnostic case–control studies generally reported higher sensitivities. Sensitivities were higher for the detection of CRC, and lower for adenomas, in both the diagnostic cohort and diagnostic case–control studies for both guaiac and immunochemical FOBTs. Conclusions Immudia HemSp appeared to be the most accurate immunochemical FOBT, however, there was no clear evidence to suggest whether guaiac or immunochemical FOBTs performed better, either from direct or indirect comparisons. Poor reporting of data limited the scope of this review, and the use the Standards for Reporting of Diagnostic Accuracy guidelines is recommended for reporting future diagnostic accuracy studies.

INTRODUCTION

T

he estimated lifetime risk of colorectal cancer (CRC) is 5–6% with almost 50% of CRC patients eventually dying of their disease.1,2 In England and Wales over 30,000 new cases are diagnosed each year.3 Mortality from CRC can be reduced by early detection of cancer and removal of adenomas.4 Randomized controlled trials (RCTs) have shown that annual or biennial screening in asymptomatic people over the age of 50 years using faecal occult blood tests (FOBTs) can reduce CRC mortality by 15–33%.5–10 This systematic review was commissioned to evaluate the comparative diagnostic performance of guaiac and immunochemical FOBTs, in the context of a decision to introduce screening for CRC, using FOBT, in the UK. RCTs can provide data on the comparative clinical outcomes of patients undergoing different screening regimens, while diagnostic accuracy studies provide information on the comparative ability of different screening tests to distinguish patients with disease from those without (diagnostic accuracy data). Therefore, diagnostic accuracy data can be considered a reasonable substitute for an RCT comparing tests, where it is combined with good evidence of a prognostic link between the parameter being tested and the disease state of interest and of the availability of an effective intervention. In this instance, a policy decision had already been made to implement FOBT screening in the UK, so that the starting assumption for the review was, by implication, that these conditions were satisfied.

Journal of Medical Screening

2007

Volume 14

Number 3

Guaiac FOBTs detect the haem moiety of haemoglobin; the pseudoperoxidase activity of haem releases oxygen from hydrogen peroxide, which reacts with guaiac to form a blue dye.11–13 Immunochemical FOBTs use monoclonal or polyclonal antibodies raised against the globin moiety of human haemoglobin, detecting intact human haemoglobin or its very early degradation products.11,14 Immunochemical tests can be grouped into several types, including enzyme immuno assay (EIA), reverse passive haemagglutination (RPHA), latex agglutination inhibition assay and latex agglutination assay. Guaiac FOBTs are more comparable. Pooling of data was precluded, for both guaiac and immunochemical FOBTs, due to clinical and statistical heterogeneity. Therefore, the narrative discusses the diagnostic performance of FOBTs according to FOBT type (guaiac or immunochemical). A UK screening programme has now been implemented, and this paper summarizes the evidence presented in the systematic review commissioned to inform the decision as to which FOBT, or combination of FOBTs, was to be used for the screening of an average-risk UK population. In July 2006, men and women in the UK aged 60–69 years were invited for screening with the guaiac FOBT, Hema-Screen (Alpha Laboratories); it is planned for screening to be repeated every two years. The first phase of the Australian National Bowel Cancer Screening Program started in August 2006, with men and women aged 55 and 65 years being offered screening with the automated version of the immunochemical FOBT Immudia HemSp, Magstream HemSp (Fujirebio Inc.).

www.jmedscreen.com

Guaiac versus immunochemical FOBTs

METHODS The systematic review was performed in accordance with the Centre for Reviews and Dissemination guidelines,15 and guidelines on the meta-analysis of diagnostic tests.16–22 We search 15 databases for published and unpublished studies, with no language restrictions, up to November 2004. The internet, bibliographies of included studies and systematic reviews, key journals and conference proceedings were also searched. In March 2007, the search was repeated in MEDLINE (including pre-MEDLINE) and EMBASE. Details of the search strategies are reported elsewhere.23 Diagnostic cohort studies (diagnosis has not been determined prior to recruitment into the study, and all participants undergo both the index test and a reference standard) were eligible if they compared a guaiac and/or immunochemical FOBT with any reference standard, in an average-risk adult population, and had sufficient data to construct a 2  2 table. Diagnostic case–control studies (diagnosis has been determined prior to recruitment into the study; all participants undergo the index test; diagnostic status is the reference standard) were included whether or not the reference standard used to verify disease status was reported. Diagnostic cohorts were derived using data from the screened arm of effectiveness trials if they reported the use of a reference standard after both a positive and negative FOBT, and had a dropout rate less than 15%. Studies evaluating flushable FOBTs, stool DNA markers, or tests for albumin or calprotectin were excluded. The methodological quality of include studies was assessed using the QUADAS tool.16 All stages of the review process were conducted in duplicate; disagreements were resolved by consensus, or referral to a third reviewer. Sensitivity, specificity, positive and negative likelihood ratios, and diagnostic odds ratio, with 95% confidence intervals (CIs), were calculated for each study, and results presented in ROC space. Data were not pooled where Cochrane Q was o0.0524 and/or I2 was >75%.25,26 Where >10 studies were included in any pooled group, regression analyses were undertaken to investigate potential sources of observed heterogeneity. Full methods and results are reported elsewhere.23

133

claims and re-screening. Poor reporting was a problem for many studies, particularly in relation to the information available to those interpreting tests.

Diagnostic accuracy of FOBTs Sensitivities of guaiac FOBTs for the detection of all neoplasms (CRC and adenomas) ranged from 6.2% (Haemoccult; specificity 98.0%)28 to 83.3% (KryptoHaem; specificity 98.4%)35 when data were derived from diagnostic cohort studies. Diagnostic case–control studies generally reported higher sensitivities, from 47.4% (Kryptohaem; specificity 98.4%)36 to 73.4% (Shionogi B; specificity 60.3%).79 Sensitivities were higher for the detection of CRC (ranging from approximately 25–96% for Haemoccult, 62–79% for Haemoccult Sensa and 27% for Shionogi B), and lower for adenomas (ranging from approximately 4–19% for Haemoccult for the detection of all adenomas, and 4–33% for the detection of adenomas of 1 cm or larger). No one guaiac FOBT was consistently better than the others. Specificity was 80% or higher for all tests on all measured outcomes. Sensitivities of immunochemical FOBTs for the detection of all neoplasms, ranged from 5.4% (OC Light; specificity 98.5%)57 to 62.6% (Immudia HemSp; specificity 94.3%)63 when data were derived from diagnostic cohort studies. Diagnostic case–control studies generally reported higher sensitivities, from 25.6% (MonoHaem; specificity 98.3%)80 to 97.7% (Immudia HemSp; specificity 98.8%).74 Immudia HemSp appeared to be the most accurate immunochemical FOBT, with sensitivities ranging from 43.3% (specificity 79.2%)80 to 97.7% (specificity 98.8%)74 and specificity from 77.0% (sensitivity 54.7%)78 to 99.0% (sensitivity 80.2%82 and 87.5%73). A higher sensitivity for the detection of CRC was only apparent for OC Light (23.7–86.1%); sensitivities of immunochemical FOBTs ranged from approximately 2% (Flexsure) to 98% (MonoHaem). There was no obvious reduction in accuracy for the detection of adenomas, with sensitivities ranging from 4% (OC Light) to 63% (Immudia HemSp) for all adenomas, and 28% (Flexsure) to 67% (Immudia HemSp) for adenomas of 1 cm or larger. Specificity was 89% or higher for all named immunochemical FOBTs on all measured outcomes.

RESULTS Fifty-nine studies met the inclusion criteria, 23 evaluating guaiac FOBTs,27–49 25 immunochemical FOBTs50–74 and 10 both.75–84 One study evaluated a guaiac and immunochemical FOBT in sequence, but did not report results for the two tests separately.85 The search in March 2007 identified two further studies (see addendum).86,87

Study quality The universal use of an appropriate reference standard to exclude disease was a particular issue for studies in this review. Seventeen out of 23 guaiac FOBT studies reported using colonoscopy after a positive FOBT (diagnostic cohort studies)/in cases (diagnostic case–control studies)30,33,3739,42-48,75,76,79,81,82 and seven after a negative FOBT/in controls.30,33,47,48,75,79,81 Twenty-one out of 25 immunochemical FOBT studies reported using colonoscopy after a positive FOBT/in cases52–55,57,60,61,63–69,71,75,76,78,79,81,82 and 16 after a negative FOBT/in controls.52,55,57,60,61,63–66,68,69,71,75,78,79,81 Where reported, other reference standards used included barium enema, sigmoidoscopy, cancer registry, follow-up, health insurance www.jmedscreen.com

Guaiac or immunochemical? Two diagnostic cohort studies reported direct comparisons between guaiac and immunochemical FOBTs.75,76 The guaiac FOBT Haemoccult Sensa was more sensitive (43.1% versus 33.1%) and less specific (90.7% versus 97.5%) than the immunochemical FOBT Flexsure for the detection of all neoplasms, but less sensitive (63.3% versus 79.1%) and less specific (90.1% versus 96.9%) for the detection of CRC in one study.75 The other study reported Haemoccult Sensa as having the highest sensitivity for the detection of CRC (78.6%; specificity 86.7%), followed by Immudia HemSp (immunochemical) (68.2%; specificity 94.4%), and unrehydrated Haemoccult (guaiac) had the lowest sensitivity (37.5%; specificity 97.7%).76 Eight diagnostic case–control studies reported direct comparisons. Four studies evaluated accuracy for the detection of all neoplasms; three indicated that an immunochemical FOBT had the highest sensitivity,78,80,82 and all four the highest specificity.79 For the detection of CRC, two studies reported that an immunochemical FOBT had the highest sensitivity and specificity,81,84 and two that a guaiac FOBT had higher sensitivity and comparable specificity.77,83 Journal of Medical Screening

2007

Volume 14

Number 3

134

Burch et al. 1

0.9 0.8 Sensitivity

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.1

0.2

0.3

0.4 0.5 0.6 0.7 False positive rate

0.8

0.9

1

Figure 1 Results for the detection of all neoplasms as reported in the cohort studies plotted in ROC space for guaiac Haemoccult (), Haemoccult Sensa (m), KryptoHaem (’), and immunochemical Flexsure (J), Immudia HemSp (n), OC Light (&) and the SPA Test (})

1 0.9 0.8 Sensitivity

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

0.1

0.2

0.3

0.4 0.5 0.6 0.7 False positive rate

0.8

0.9

1

Figure 2 Results from all studies for the detection of all neoplasms plotted in ROC space. Black symbols are results from cohort studies, and white symbols from case–control studies. Circles are results for guaiac FOBTs and triangles for immunochemical

Therefore, immunochemical FOBTs were not consistently better than guaiac FOBTs. With so few direct comparisons, attempts were made to draw indirect comparisons by visual inspection of the results of the diagnostic cohort studies plotted in ROC space (Figure 1). There was no clear indication that either guaiac or immunochemical FOBTs performed better. We restricted indirect comparisons to diagnostic cohort studies due to the potential overestimation of accuracy in diagnostic case–control studies.88,89 The impact of study design can be seen in Figure 2, where results from both diagnostic cohort and diagnostic case–control studies are plotted in ROC space. It can be seen that the sensitivity of both guaiac and immunochemical FOBTs was generally higher in diagnostic case–control studies.

DISCUSSION In diagnostic cohort studies, KryptoHaem had the highest sensitivity of the guaiac tests for detecting all neoplasms, and non-rehydrated Haemoccult and Haemoccult Sensa for detecting CRC. However, studies from which these data were derived had limitations, including a lack of information regarding the patient spectrum,35,49 or use of an inappropriate reference standard (barium enema, sigmoidoscopy, or questionnaire, general practitioner contact and Journal of Medical Screening

2007

Volume 14

Number 3

cancer registry).35,40,49 Immudia HemSp seemed the most sensitive immunochemical FOBT for the detection of all neoplasms in the diagnostic cohort studies.63 This test also performed more consistently than other immunochemical FOBTs for the detection of CRC, and the diagnostic cohort study evaluating Immudia HemSp was of better quality than those evaluating other immunochemical FOBTs. The number of colonoscopies required due to differing positivity rates will be an issue for any national screening programme. If 60% of the people aged between 60 and 69 years in the UK (3,360,000 people) complied with FOBT screening,90 the positivity rates reported in the diagnostic cohort studies in this review suggest that between 36,960 and 538,000 colonoscopies after guaiac tests, and 60,480 and 302,400 after immunochemical tests, would be required if used as a single test. These values reflect the uncertainty of the evidence available; no firm conclusions can be drawn from these data regarding the number of colonoscopies that would be required after implementation of screening and how this might vary with the type of FOBT used. We conducted extensive literature searches, assessed study quality, included studies in any language, and attempted to maximize available data by deriving diagnostic cohorts from the intervention arms of screening studies. Few studies reported direct comparisons between FOBTs; where reported, results were inconsistent and often conflicting, though there was some indication of a better overall performance for immunochemical FOBTs. Therefore the review had to rely upon less reliable indirect comparisons, which also showed no clear evidence to suggest that either guaiac or immunochemical FOBTs performed better. Further limitations of the current review included the more frequent use of studies of diagnostic case–control design to evaluate immunochemical FOBTs, and the more frequent publication of these results in Japanese or Chinese requiring translation, increasing the potential for data extraction errors. Diagnostic case–control studies are prone to bias and over-estimation of diagnostic accuracy, particularly when cases are people with more severe disease and controls are healthy individuals.88,89 When the control group is derived from patients with other diagnoses, accuracy may be under- or over-estimated, depending upon the diagnoses of the control group.91,92 This bias, associated with the study design, may obscure clinically significant differences between the different test methodologies. The effect of this bias was evident for Haemoccult, OC Light and Immudia HemSp, where diagnostic case–control studies reported consistently higher sensitivities than the diagnostic cohort studies. We regarded colonoscopy as the gold standard for diagnosis; other reference standards (sigmoidoscopy, referral to a cancer registry, follow-up, re-screening) were considered inappropriate for the target condition. Sigmoidoscopy only examines the distal bowel; FOBTs tend to be better at detecting lesions in this region, therefore accuracy may be overestimated, particularly for immunochemical FOBTs which detect intact globin that may be degraded during transit from a lesion in the proximal bowel. This bias may be offset to some degree by the greater volume of bleeding from cancers of the proximal colon.93,94 Referral to the cancer registry and repeat screening may result in interval (false negative) results being missed, either due to loss to followup or being wrongly classified as true negative. The use of reference standards that may not identify the target condition in those with a negative FOBT could result in an underestimation of the number of false-negatives and hence an over-estimation of sensitivity. www.jmedscreen.com

Guaiac versus immunochemical FOBTs

Guaiac tests are generally best at detecting large, more distal lesions. They depend upon peroxidase or pseudoperoxidase activity in faeces not specific to human haemoglobin. A number of variables may influence guaiac test results that do not influence the results of immunochemical tests. These include ingestion of animal haemoglobin/ myoglobin in red meat, fruits and vegetables high in peroxidase activity, and aspirin or other medication that may produce gastrointestinal bleeding, each of which may cause false-positive results,95 and high doses of vitamin C96 and faecal dehydration97 which may cause false-negative results. Insufficient data were available to support investigation into the impact of study design, type of test, patient spectrum, reference standards used, the threshold used to define a positive result and test-specific details (e.g. rehydration and dietary restrictions for guaiac FOBTs) upon diagnostic accuracy.

Future research Well-designed diagnostic cohort studies directly comparing guaiac and immunochemical FOBTs are required to determine the true comparative accuracy of these different testing strategies. The possible need for retesting, and impacts upon the infrastructure requirements and cost of a screening programme, particularly in relation to the number of colonoscopies that would be required in the UK, should be examined. Studies should ideally recruit a representative screening population, use colonoscopy to confirm diagnosis regardless of the FOBT result, measure the detection of CRC and adenomas and report the results separately and combined, and allow outcome assessors access to clinical information that would be available in practice, but blind them to other information. Other areas requiring research include: the impact of dietary restrictions for guaiac FOBTs on compliance; the acceptability of different FOBTs to screenees; the impact of false-positive and false-negative results on psychiatric morbidity, future self-referral with onset of symptoms, and attendance at rescreening sessions; and the impact on accuracy of variability in the skills and experience of those conducting and interpreting tests. The availability of tests that allow automated analysis must be considered, and their impact on accuracy by removing the human factors associated with manual development of tests determined. The standard of reporting of primary studies was a problem throughout the review and the authors recommend the use of the Standards for Reporting of Diagnostic Accuracy (STARD) checklist to improve the quality of reporting of future diagnostic studies.98–100

Implications for policy Sensitivities derived from diagnostic cohort studies were generally low, being o50% in 89% of studies for the detection of all neoplasms, and 50% of studies for the detection of CRC. Therefore the ‘miss rate’ is likely to be high regardless of the FOBT used, particularly when small adenomas are included in the definition of the target condition. The overall test accuracy, compliance, handling of the specimens, cost-effectiveness and availability in the UK must be considered when selecting an FOBT for use as a screening tool. Compliance may be better with immunochemical tests due to an easier collection method, less contact with the stool specimen and no requirement for dietary restrictions; there is insufficient evidence to confirm this. In contrast, immunochemical FOBTs may be more www.jmedscreen.com

135

costly. There have also been concerns raised in the UK relating to the postal service handling immunochemical FOBTs. However, when the Australian screening programme was introduced, approval was granted for the transfer of immunochemical tests via the postal service.

Addendum The search in March 2007 identified two further studies that met the inclusion criteria.86,87 One was a diagnostic cohort study in which 3090 patients underwent colonoscopy and an unspecified immunochemical FOBT. For the detection of all neoplasms, sensitivity was 53% and specificity 99.6%; for the detection of CRC, sensitivity was 52.6% and specificity 87.2%.86 The other was a diagnostic case–control study in which 57 people with a diagnosis of cancer and 44 volunteer controls underwent a guaiac FOBT; sensitivity was 51% and specificity 98%.87 These results are comparable with those of other studies of similar design included in the review, and do not alter the conclusions.

ACKNOWLEDGEMENTS We thank Ruth Lewis for her help during the development of the protocol and the early stages of this review. We thank the advisory panel: Mr SP Halloran, Consultant Clinical Biochemist, Royal Surrey County Hospital; Dr Alphons GH Kessels, Departments of Clinical Epidemiology and Medical Technology Assessment, Maastricht University Hospital; Mr B Saunders, Consultant Physician & Senior Lecturer, Wolfson Unit for Endoscopy, St Mark’s Hospital, Harrow; Professor R J C Steele, Department of Surgery and Molecular Oncology, Ninewells Hospital and Medical School, Dundee and Professor David Weller, Centre for Public Health and Primary Care Research, University of Edinburgh, for their comments on the protocol and draft report. We also thank the authors we contacted for further information and who responded to us, in particular, Professor JA Allison (USA) and Dr LL Lim (China), for providing us with data prior to publication. This work was undertaken by The Centre for Reviews and Dissemination, which received funding from the UK NHS Cancer Screening Programme. The views expressed in the publication are those of the authors and not necessarily those of the UK NHS Cancer Screening Programme.

................

Authors’ affiliations Jane Burch, Research Fellow, Centre for Reviews and Dissemination (CRD), University of York, York YO10 5DD, UK Karla Soares-Weiser, Director, Enhance Reviews, Kfar-Saba, Israel James St John, Associate Professor, The Cancer Council Victoria, Melbourne, Australia Steven Duffy, Information Officer, CRD, University of York, York, UK Stephen Smith, Consultant Clinical Biochemist, University Hospitals of Coventry & Warwickshire, Coventry, UK Jos Kleijnen, Director, Kleijnen Systematic Reviews Ltd, York, UK Marie Westwood, Senior Research Fellow, CRD, University of York, York, UK

REFERENCES 1 2

Strul H, Arber N. Fecal occult blood test for colorectal cancer screening. Ann Oncol 2002;13:51–6 Mathers CD, Boschi-Pinto C, Lopez AD, Murray CJL. Cancer Incidence, Mortality and Survival by Site for 14 Regions of the World. Geneva: World Health Organization, 2001

Journal of Medical Screening

2007

Volume 14

Number 3

136 3 4 5 6 7 8 9 10 11

12 13 14 15

16 17 18 19 20 21 22 23

24 25 26 27

28 29 30

31

32 33

Burch et al. National Statistics. Cancer Statistics Registrations: Registrations of Cancer Diagnosed in 2003, England Series MB1 no. 34. London: The Stationery Office, 2005 Bond JH. Fecal occult blood test screening for colorectal cancer. Gastrointest Endosc Clin N Am 2002;12:11–21 Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365–71 Hardcastle JD, Chamberlain JO, Robinson MH, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472–7 Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occultblood test. Lancet 1996;348:1467–71 Mandel JS, Church TR, Ederer F, Bond JH. Colorectal cancer mortality: effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 1999;91:434–7 Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA 2003;289:1288–96 Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale: update based on new evidence. Gastroenterology 2003;124:544–60 Young GP, St John DJ, Winawer SJ, Rozen P. Choice of fecal occult blood tests for colorectal cancer screening: recommendations based on performance characteristics in population studies: a WHO (World Health Organization) and OMED (World Organization for Digestive Endoscopy) report. Am J Gastroenterol 2002;97:2499–507 Ransohoff DF, Lang CA. Screening for colorectal cancer with the fecal occult blood test: a background paper American College of Physicians. Ann Intern Med 1997;126:811–22 Walsh JM, Terdiman JP. Colorectal cancer screening: clinical applications. JAMA 2003;289:1297–302 Pearson SM. MDA evaluation of 11 faecal occult blood test kits. In: Martin SM, ed. Proceedings of Pathology. Birmingham, London: IBMS, 2000;114 NHS Centre for Reviews and Dissemination. Undertaking Systematic Reviews of Research on Effectiveness: CRD’s Guidance for Those Carrying Out or Comissioning Reviews CDR Report 4. 2nd edn York: NHS Centre for Reviews and Dissemination, 2001 Whiting P, Rutjes AW, Dinnes J, Reitsma J, Bossuyt PM, Kleijnen J. Development and validation of methods for assessing the quality of diagnostic accuracy studies. Health Technol Assess 2004;8:1–234 Honest H, Bachmann LM, Khan K. Electronic searching of the literature for systematic reviews of screening and diagnostic tests for preterm birth. Eur J Obstet Gynecol Reprod Biol 2003;107:19–23 Honest H, Khan KS. Reporting of measures of accuracy in systematic reviews of diagnostic literature. BMC Health Serv Res 2002;2:4 Smidt N, Rutjes AW, van der Windt DA, et al. Quality of reporting of diagnostic accuracy studies. Radiology 2005;235:347–53 Whiting P, Rutjes AW, Reitsma JB, Glas AS, Bossuyt PM, Kleijnen J. Sources of variation and bias in studies of diagnostic accuracy: a systematic review. Ann Intern Med 2004;140:189–202 Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med 2002;21:1525–37 Deeks JJ. Systematic reviews in health care: systematic reviews of evaluations of diagnostic and screening tests. BMJ 2001;323:157–62 Soares-Weiser K, Burch J, Duffy S, Smith S, Kleijnen J, Westwood M. Diagnostic accuracy and cost-effectiveness of faecal occult blood tests (FOBT) used in screening for colorectal cancer: a systematic review 2007. CRD Report 36. http://www.york.ac.uk/inst/crd/pdf/report36.pdf Fleiss JL. The statistical basis of meta-analysis. Stat Methods Med Res 1993;2:121–45 Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60 Allison JE, Feldman R, Tekawa IS. Hemoccult screening in detecting colorectal neoplasm: sensitivity, specificity, and predictive value: longterm follow-up in a large group practice setting. Ann Intern Med 1990;112:328–33 Bang KM, Tillett S, Hoar SK, Blair A, McDougall V. Sensitivity of fecal hemoccult testing and flexible sigmoidoscopy for colorectal cancer screening. J Occup Med 1986;28:709–13 Bennett DH, Hardcastle JD, Moshakis V, et al. A comparison of screening for colorectal cancer by flexible sigmoidoscopy and faecal occult blood testing (FOBT) with FOBT alone. Gastroenterology 1996;110:A491 Bhattacharya I, Sack EM. Screening colonoscopy is as effective as colonoscopy for a positive fecal occult blood test (FOBT) for the detection of colorectal neoplasia in average risk asymptomatic patients 60–75 years old. Gastroenterology 1997;112(4 Suppl):A538 Brevinge H, Lindholm E, Buntzen S, Kewenter J. Screening for colorectal neoplasia with faecal occult blood testing compared with flexible sigmoidoscopy directly in a 55–56 years’ old population. Int J Colorectal Dis 1997;12:291–5 Castiglione G, Grazzini G, Poli A, Bonardi R, Ciatto S. Hemoccult sensitivity estimate in a screening program for colorectal cancer in the Province of Florence. Tumori 1991;77:243–5 Collins JF, Lieberman DA, Durbin TE, Weiss DG. and the Veterans Affairs Cooperative Study #380 Group. Accuracy of screening for fecal occult blood on a single stool sample obtained by digital rectal examination:

Journal of Medical Screening

2007

Volume 14

Number 3

34 35 36 37 38

39

40 41 42 43

44

45 46 47

48 49

50 51 52 53

54

55 56 57

58 59 60

a comparison with recommended sampling practice. Ann Intern Med 2005;142:81–5 Foley DP, Dunne P, Dervan PJ, Callaghan TWO, Crowe J, Lennon JR. Leftsided colonoscopy and hemoccult screening for colorectal neoplasia. Eur J Gastroenterol Hepatol 1992;4:925–36 Klug W, Knoch HG. [What does the Krypto-Haem test accomplish in occult hemorrhage from the large intestine?] [in Japanese]. Z Arztl Fortbild (Jena) 1983;77:516–19 Lampe J, Jacobasch KH, Uhlig K. Krypto-haem SSWR, ein testbrief zum nachweis von okkultem blut im stuhl. Dtsch Gesundh Wes 1982;37:175–7 Launoy G, Smith TC, Duffy SW, Bouvier V. Colorectal cancer massscreening: estimation of faecal occult blood test sensitivity, taking into account cancer mean sojourn time. Int J Cancer 1997;73:220–4 Lieberman DA, Weiss DG. For the Veterans Affairs Cooperative Study Group 380. One-time screening for colorectal cancer with combined fecal occult-blood testing and examination of the distal colon. N Engl J Med 2001;345:555–60 Mandel JS, Bond JH, Bradley M, et al. Sensitivity, specificity, and positive predictivity of the Hemoccult test in screening for colorectal cancers: the University of Minnesota’s Colon Cancer Control Study. Gastroenterology 1989;97:597–600 Michalek AM, Cummings KM, Gamble D. The use of a cancer registry in a mass screening program for colorectal cancer. Cancer Detect Prev 1988;11:353–7 Murakami R, Otani T, Nakanishi K, et al. Diagnostic validity of fecal occult blood tests for detecting gastroenterological cancers. Jpn J Cancer Res 1992;83:141–5 Niv Y, Lev-El M, Fraser G, Abuksis G, Tamir A. Protective effect of faecal occult blood test screening for colorectal cancer: worse prognosis for screening refusers. Gut 2002;50:33–7 Parikh A, Ramamoorthy R, Kim KH, Holland BK, Houghton J. Fecal occult blood testing in a noncompliant inner city minority population: increased compliance and adherence to screening procedures without loss of test sensitivity using stool obtained at the time of in-office rectal examination. Am J Gastroenterol 2001;96:1908–13 Rasmussen M, Kronborg O, Fenger C, Jorgensen OD. Possible advantages and drawbacks of adding flexible sigmoidoscopy to hemoccult-II in screening for colorectal cancer: a randomized study. Scand J Gastroenterol 1999;34:73–8 Rennert G, Rennert HS, Miron E, Peterburg Y. Population colorectal cancer screening with fecal occult blood test. Cancer Epidemiol Biomarkers Prev 2001;10:1165–8 Ribet A, Frexinos J, Escourrou J, Delpu J. Occult blood tests and colorectal tumours. Lancet 1980;1:417 St John DJ, Young GP, McHutchison JG, Deacon MC, Alexeyeff MA. Comparison of the specificity and sensitivity of Hemoccult and HemoQuant in screening for colorectal neoplasia. Ann Intern Med 1992;117:376–82 Sung JJ, Chan FK, Leung WK, et al. Screening for colorectal cancer in Chinese: comparison of fecal occult blood test, flexible sigmoidoscopy, and colonoscopy. Gastroenterology 2003;124:608–14 Winawer SJ, Schottenfeld D, Miller D, et al. Detection of early colon cancer and colonic polyps. In: Nieburgs HE, ed. Prevention and Detection of Cancer: Proceedings of the Third International Symposium on Detection and Prevention of Cancer New York 1976, Part II, vol. 2. New York: Marcel Dekker, 1980;2103–10 Chen HS, Sheen-Chen SM. Influence of age and gender on surveillance for colorectal tumors in low-risk asymptomatic population. Anticancer Res 2002;22:399–403 Chen K, Jiao DA, Zheng S. et al. Diagnostic value of fecal occult blood testing for screening colorectal cancer. China Natl J New Gastroenterol 1997;3:166–8 Cheng TI, Wong JM, Hong CF, et al. Colorectal cancer screening in asymptomaic adults: comparison of colonoscopy, sigmoidoscopy and fecal occult blood tests. J Formos Med Assoc 2002;101:685–90 Gondal G, Grotmol T, Hofstad B, Bretthauer M, Eide TJ, Hoff G. The Norwegian Colorectal Cancer Prevention (NORCCAP) screening study: baseline findings and implementations for clinical work-up in age groups 50–64 years. Scand J Gastroenterol 2003;38:635–42 Itoh M, Takahashi K, Nishida H, Sakagami K, Okubo T. Estimation of the optimal cut off point in a new immunological faecal occult blood test in a corporate colorectal cancer screening programme. J Med Screen 1996;3:66–71 Kawai K. Clinical study on MonoHaem, an immunological fecal occult blood test reagent. Nippon Daicho Komonbyo Gakkai Zasshi 1987;40:308–13 Kim YH, Kim J, Son HJ, et al. Immunochemical fecal occult blood test for detection of rectosigmoid cancer in asymptomatic adults. Gastroenterology 1998;114:A625 Liu HH, Huang TW, Chen HL, Wang TH, Lin JT. Clinicopathologic significance of immunohistochemical fecal occult blood test in subjects receiving bidirectional endoscopy. Hepatogastroenterology 2003;50: 1390–2 Morikawa T, Kato J, Yamaji Y, et al. Pitfall of immunological fecal occult blood test: a comparison with total colonoscopy. Gastrointest Endosc 2004;59:AB286 Nakama H, Fattah A, Zhang B, Uehara Y, Wang C. A comparative study of immunochemical fecal tests for detection of colorectal adenomatous polyps. Hepatogastroenterology 2000;47:386–9 Nakama H, Abdul Fattah AS, Zhang B, Kamijo N, Fujimori K, Miyata K. Detection rate of immunochemical fecal occult blood test for colorectal

www.jmedscreen.com

Guaiac versus immunochemical FOBTs

61 62 63

64 65

66

67 68

69 70 71 72

73 74 75 76 77

78

adenomatous polyps with severe dysplasia. J Gastroenterol 1997;32: 492–5 Nakama H, Kamijo N, Fujimori K, Fattah AS, Zhang B. Diagnostic accuracy of immunochemical faecal occult blood test for gastric cancer. J Med Screen 1996;3:113–14 Nakama H, Zhang B, Kamijo N. Sensitivity of immunochemical fecal occult blood test for colorectal flat adenomas. Hepatogastroenterology 2004;51:1333–6 Nakama H, Zhang B, Fattah AA, Kamijo N, Zhang X. Characteristics of colorectal cancer that produce positive immunochemical occult blood test results on stool obtained by digital rectal examination. Can J Gastroenterol 2001;15:227–30 Nakama H, Zhang B, Fattah AS, Zhang X. Colorectal cancer in iron deficiency anemia with a positive result on immunochemical fecal occult blood. Int J Colorectal Dis 2000;15:271–4 Nakama H, Fattah AS, Zhang B, Kamijo N. Digital rectal examination sampling of stool is less predictive of significant colorectal pathology than stool passed spontaneously. Eur J Gastroenterol Hepatol 2000;12:1235–8 Nakama H, Zhang B, Abdul Fattah AS, Kamijo N, Fukazawa K. Relationships between a sign of rectal bleeding and the results of an immunochemical occult blood test, and colorectal cancer. Eur J Cancer Prev 2000;9:325–8 Nakama H, Kamijo N, Abdul Fattah AS, Zhang B. Validity of immunological faecal occult blood screening for colorectal cancer: a follow up study. J Med Screen 1996;3:63–5 Okamoto M, Kawabe T, Shiratori Y, et al. Risk for colorectal cancer in patients with positive fecal occult blood test: analysis on 5648 asymptomatic patients performed total colonoscopy. Gastroenterology 1997;112:A632 Tada M, Hosokawa N, Shimizu S, et al. Evaluation of a new fecal occult blood test. J Kyoto Prefectural Univ Med 1986;95:1331–7 Tada M. New immunological fecal occult blood test applying latex inhibition method]. Med Sci Pharmaceut 1988;20:1271–7 Zhang B, Nakama H, Fattah AS, Kamijo N. Lower specificity of occultblood test on stool collected by digital rectal examination. Hepatogastroenterology 2002;49:165–7 Zhou DY, Feng FC, Zhang YL, et al. Comparison of Shams’ test for rectal mucus to an immunological test for fecal occult blood in large intestinal carcinoma screening: analysis of a check-up of 6480 asymptomatic subjects. Chin Med J (Engl) 1993;106:739–42 Zhou PH, Zhang WZ. Experimental study on reverse passive hemagglutination for detection of human fecal occult blood. Natl Med J China 1987;67:671–2 Zhu WX, Lin JJ. Reverse passive hemagglutination for detection of fecal occult blood: a comparison with Japanese Immudia-Hem SP Kit. Chin Med J (Engl) 1988;101:519–22 Allison JE, Levin T, Sakoda L, et al. The new fecal occult blood tests have poor application sensitivity for significant polyps in average risk subjects. Gastroenterology 2002;122:A592–3 Allison JE, Tekawa IS, Ransom LJ, Adrain AL. A comparison of fecal occultblood tests for colorectal-cancer screening. N Engl J Med 1996;334:155–9 Kikkawa N, Takeuchi H, Kawahara T, Okamoto M, Funahashi S, Yamamoto A. [Comparison of chemical and immunological methods in fecal occult blood testing] [in Japanese]. Gan No Rinsho 1987;33:1049–51 Matsuse R, Uchida K, Miyachi N, et al. [Immunochemical detection of fecal occult blood by latex-agglutination reaction for hemoglobin and transferrin]. Latest Med Sci 1989;44:2636–41

www.jmedscreen.com

137 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98

99 100

Miyoshi H, Ohshiba S, Asada S, Hirata I, Uchida K. Immunological determination of fecal hemoglobin and transferrin levels: a comparison with other fecal occult blood tests. Am J Gastroenterol 1992;87:67–73 Miyoshi H, Shimamoto C, Hirata I, et al. [The clinical study of the new immunological detection test (stick EIA) for faecal occult blood: first report]. Prog Med 1988;8:1510–2 Nakama H, Kamijo N. Accuracy of immunological fecal occult blood testing for colorectal cancer screening. Prev Med 1994;23:309–13 St John DJ, Young GP, Alexeyeff MA, et al. Evaluation of new occult blood tests for detection of colorectal neoplasia. Gastroenterology 1993;104:1661–8 Takeshita T, Horiguchi J, Kinoshita T, et al. Immunological fecal occult blood testing with anti-human hemoglobin antibody]. Nippon Daicho Komonbyo Gakkai Zasshi 1985;38:369–75 Takeshita T, Sugimoto N, Ashizawa S. [Mass screening for colo-rectal cancer by fecal occult blood test]. Nippon Daicho Komonbyo Gakkai Zasshi 1982;35:561–6 Li S, Zhang C, Xu E. [Evaluation of combined test of sequential fecal occult blood and albumin in the screening of colorectal neoplasms]. Zhonghua Zhong Liu Za Zhi 1995;17:381–3 Nakazato M, Yamano HO, Matsushita HO, et al. Immunologic fecal occult blood test for colorectal cancer screening. JMAJ 2006;49:203–7 Kutzner N, Hoffmann I, Linke C, et al. Non-invasive detection of colorectal tumours by the combined application of molecular diagnosis and the faecal occult blood test. Cancer Lett 2005;229:33–41 Lijmer JG, Mol BW, Heisterkamp S, et al. Empirical evidence of designrelated bias in studies of diagnostic tests. JAMA 1999;282:1061–6 Westwood ME, Whiting PF, Kleijnen J. How does study quality affect the results of a diagnostic meta-analysis? BMC Med Res Methodol 2005;5:20 Atkin WS. Impending or pending? The national bowel cancer screening programme. BMJ 2006;332; doi: 10.1136/bmj.38797.494757.47 Rutjes AW, Reitsma JB, Vandenbroucke JP, Glas AS, Bossuyt PM. Case–control and two-gate designs in diagnostic accuracy studies. Clin Chem 2005;51:1335–41 Rutjes AW, Reitsma JB, Nisio MD, Smidt N, van Rijn JC, Bossuyt PM. Evidence of bias and variation in diagnostic accuracy studies. Can Med Assoc J 2006;174:469–76 Macrae FA, St John DJB. Relationship between patterns of bleeding and Hemoccult sensitivity in patients with colorectal cancers or adenomas. Gastroenterology 1982;82:891–8 Turunen MJ, Liewendahl K, Partanen P, Adlercreutz H. Immunological detection of faecal occult blood in colorectal cancer. Br J Cancer 1984;49:141–8 Macrae FA, St John DJ, Caligiore P, Taylor LS, Legge JW. Optimal dietary conditions for hemoccult testing. Gastroenterology 1982;82:899–903 Jaffe RM, Kasten B, Young DS, MacLowry JD. False negative stool occult blood tests caused by ingestion of ascorbic acid (vitamin C). Ann Intern Med 1975;83:824–6 Austoker J. Cancer prevention in primary care: screening for colorectal cancer. BMJ 1994;309:382–6 Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Standards for Reporting of Diagnostic Accuracy. Clin Chem 2003; 49:1–6 Bossuyt PM, Reitsma JB, Bruns DE, et al. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Ann Intern Med 2003;138:W1–2 Bossuyt PM, Reitsma JB. Standards for reporting of diagnostic accuracy. The STARD initiative. Lancet 2003;361:71

Journal of Medical Screening

2007

Volume 14

Number 3