Alimentary Pharmacology & Therapeutics
`
`Comparative evaluation of a new bedside faecal occult blood test
`in a prospective multicentre study
`N . H O E P F F NE R * 1 , Y . M . S H AS T RI * 1 , E . H AN IS C H  , W. R O¨ S C Hà , J . M O¨ S S NE R § , W . F . C A S PA R Y *
`& J . S T E IN *
`
`*Department of Medicine I, ZAFES,
`Centre of Internal Medicine, Johann
`Wolfgang Goethe University, Frank-
`furt/Main;  Asklepiosklinik, Dreieich-
`Langen; àDepartment of Medicine II,
`Hospital Nordwest, Frankfurt/Main;
`§Department of Medicine II, Univer-
`sity Hospital Leipzig, Leipzig,
`Germany
`
`Correspondence to:
`Dr J. Stein, Department of Medicine I,
`ZAFES, J. W. Goethe University
`Frankfurt, Theodor-Stern-Kai 7,
`D – 60590 Frankfurt/Main, Germany.
`E-mail: j.stein@em.uni-frankfurt.de
`
`1These authors contributed equally to
`this study.
`
`Publication data
`Submitted 23 August 2005
`First decision 12 September 2005
`Resubmitted 29 September 2005
`Accepted 29 September 2005
`
`SUMMARY
`
`Background
`Faecal occult blood testing is an established method of colorectal neopl-
`asia screening. Guaiac-based tests are limited by poor patient compli-
`ance, low sensitivity, specificity and positive predictive value. Newer
`immunochemical-based tests, accurate but tedious, require a well-estab-
`lished laboratory set up. There is need for simpler immunochemical tests
`that can be performed at the out-patient clinic.
`
`Aim
`To compare the performance characteristics of a new bedside immuno-
`logical test strip device with a sensitive Guaiac-based and established
`immunochemical test for detection of faecal occult blood in patients
`undergoing colonoscopy.
`
`Methods
`A total of 389 consecutive patients from four centres who were referred
`for colonoscopy also provided the stool samples for detection of occult
`blood without dietary restrictions. Stool
`tests performed were (i)
`Guaiac-based,
`(ii)
`immunochemical
`enzyme-linked immunosorbent
`assay and (iii) bedside immunochemical strip test.
`
`Results
`At the optimal threshold level, the sensitivity and specificity of the beside
`immunochemical strip test for detection of significant colorectal neoplasia
`(adenomas >1.0 cm and carcinomas) were 60% and 95%, respectively.
`
`Conclusions
`This bedside immunochemical strip test proved to be a simple, convenient,
`non-cumbersome and accurate tool with similar performance characteris-
`tics for detection of any bleeding lesion including colorectal neoplasia when
`compared with an established immunochemical faecal occult blood test.
`
`Aliment Pharmacol Ther 23, 145–154
`
`ª 2006 Blackwell Publishing Ltd,
`doi:10.1111/j.1365-2036.2006.02702.x
`
`145
`
`Geneoscopy Exhibit 1029, Page 1
`
`

`

`146 N . H O E P F F N E R et al.
`
`INTRODUCTION
`
`Colorectal cancer (CRC) accounted for about 1 million
`new cases in 2002; its worldwide prevalence is only
`second to that of breast cancer. It is one of the leading
`causes of cancer death in western societies with 25-
`fold variation in its occurrence across the globe.1 In
`Europe, there were estimated to be 376 400 new cases
`of CRC amounting to 203 700 deaths in 20042 while
`in the US there are expected to be 145 200 CRC regis-
`trations and 56 290 deaths this year.3 It is estimated
`that about half of these deaths could have been pre-
`vented if patients would have undergone CRC screen-
`ing.4 It has been observed that most cases of sporadic
`CRC arise from pre-existing adenomatous polyps.
`The metastatic CRC [stage IV tumour node metasta-
`sis (TNM)] has a dismal survival of <5% when com-
`pared with early CRC (stage I TNM).5 The prevalence
`of the disease, with the proven ability of screening to
`reduce mortality makes population-wide screening of
`CRC a lucrative health goal. An ideal screening test
`must be non-invasive, acceptable to patients, simple,
`have high sensitivity and specificity,
`inexpensive,
`cost-effective, easy to implement across a large popu-
`lation and should be effective in reducing morbidity
`and mortality from CRC in the screened population.6
`The most widely accepted non-invasive method for
`detecting CRC is faecal occult blood testing (FOBT). It
`is routinely used world over for CRC screening and
`surveillance and has been in use now for more than
`three decades. Screening in asymptomatic populations
`with FOBTs followed by colonoscopy if tested positive
`have reduced mortality rates by 15–33%.7–10 However,
`sensitivity of FOBTs ranges from 26% to 69% for
`detecting CRC.11–14
`There are two main types of commercially available
`FOBTs: ‘Guaiac-based tests’ in which Guaiac blue col-
`our change occurs because of the pseudoperoxidase
`activity of haeme in the stool; ‘immunochemical tests’
`which detect globin portion of human haemoglobin
`(Hb) in the stool using anti-Hb antibodies.15 A major
`disadvantage of Guaiac tests is the fact that they are
`not specific for human blood and can produce false-
`positive results when meat, fruit, or vegetables con-
`taining peroxidase have been ingested. Also bleeding
`in the upper gastrointestinal (GI) tract secondary to
`aspirin and non-steroidal anti-inflammatory drugs
`(NSAIDs) can produce false-positive Guaiac-based
`tests.16 This may lead to contraindicate colonoscopic
`examinations thus increasing the cost and risk of
`
`screening programmes. False-negative tests may occur
`with ingestion of high doses of vitamin C, which will
`prevent detection of cancer and will provide a false
`sense of security to the patient. Because of these short-
`comings Guaiac-based tests are being replaced by
`immunochemical faecal occult blood tests (IFOBTs),
`despite being more expensive.
`The IFOBTs, which are widely used as screening
`tests in Japan, have been shown to be superior to the
`Guaiac-based tests in sensitivities, with similar specif-
`icity, as reported both in known CRC17 and in asymp-
`tomatic populations.13, 18 This is because IFOBTs do
`not detect blood from the upper GI tract, do not react
`with plant peroxidases and animal Hb. Thus, there is
`no need of dietary or medical restriction. One of the
`studies has shown 66% increase in patient participa-
`tion with the elimination of dietary restrictions.19 Fur-
`thermore, sensitivity of IFOBTs is high for adenomas,17
`intramucosal cancer20 and Dukes A cancer.21 On the
`basis of these reports,
`it is expected that screening
`with IFBOTs may provide larger effects on the inci-
`dence reduction of advanced CRC when compared with
`the Guaiac-based tests. However, most IFOBTs require
`a suitable well-equipped laboratory, i.e. an enzyme-
`linked immunosorbent assay (ELISA) reader, limiting
`their widespread use because of logistics and higher
`costs.
`Colonoscopy is considered the gold standard for
`CRC screening. But
`it is expensive, requires skilled
`endoscopist and may lead to life-threatening compli-
`cations like perforation (one in 300 procedures).22
`However, complication rate varies in different studies
`and one of the recent studies in which more than
`22 000 patients underwent screening colonoscopies
`(only diagnostic) by experienced endoscopists without
`any major complication or mortality.23
`In a recent pilot study, we demonstrated a new
`immunological test strip device (bedside IFOBT) for the
`rapid, quantitative detection of FOB (Prevent ID CC,
`Preventis GmbH, Bensheim, Germany) and compared it
`with a widely used Guaiac-FOBT keeping ELISA-IFOBT
`as the gold standard. We showed that this bedside
`IFOBT is much more sensitive (76% vs. 30%) for detec-
`tion of FOB than Guaiac-based test whereas specificity
`of both tests were comparable.24 However, we did not
`perform colonoscopy in these patients.
`In the present prospective study, we have evaluated
`the performance characteristics of this new bedside
`test for detection of FOB in patients undergoing colon-
`oscopy. Experienced endoscopists carefully examined
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 2
`
`

`

`the entire large bowel, recorded and reported the pres-
`ence of any bleeding lesion that could cause positive
`FOBT. If any such lesion was found it was biopsied or
`snared if indicated. We also compared it with another
`established IFOBT (ELISA) and a sensitive Guaiac-
`based FOBT.
`Our study is peculiar and different from previous
`studies in that the patient population was not only of
`CRC screening patients but also constituted the
`patients referred to or admitted with colonic symptoms
`and many of them had known clinical diagnosis. Thus,
`we evaluated performance characteristics of FOBT not
`only in detecting neoplasia but also in detecting pres-
`ence of microscopic faecal blood whether
`it was
`because of neoplasia,
`inflammatory bowel disease
`(IBD) or dysentery.
`
`MATERIAL AND METHODS
`
`Ethics
`
`The study protocol was approved by the ethics com-
`mittee of Frankfurt University Hospital. Written
`informed consent was obtained from all participants
`before entering the study.
`
`Patients and healthy controls
`
`Patients from four different centres who were referred
`for colonoscopy were invited to participate in the
`study. There were 237 patients who either had known
`clinical diagnosis (e.g. IBD) or were symptomatic sug-
`gestive of colonic disease while as 150 healthy
`patients underwent CRC screening. Patients with active
`GI bleeding, menstruation and past history of total
`colectomy were not included in the study.
`Patients given appointments for colonoscopy were
`asked to provide one stool sample for detecting faecal
`blood. No dietary restrictions were advised to the
`patients. Stool sample was collected before starting the
`bowel preparation in a sterile box for Guaiac and
`immunological tests whereas for the bedside IFOBT we
`provided patients with a simple stool-collecting device
`(see below). Stool samples of these patients with or
`without pathological findings were tested. In a stand-
`ard questionnaire patient’s clinical history was recor-
`ded.
`After standard bowel wash colonoscopies were per-
`formed by the endoscopists, who were not aware of
`FOBT findings, till the caecum using the standard col-
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
`B E D S I D E I M M U N O L O G I C A L F O B T 1 4 7
`
`onoscopes. Patients who were poorly prepared or in
`whom caecum could not be reached were not included
`in the study. Histology was obtained from the routine
`biopsies, polypectomies and/or from surgery. Adeno-
`mas were classified as
`large (>10 mm) or
`small
`(<10 mm) as judged during colonoscopies using biopsy
`forceps as measure to assess polyp size. All data were
`collected prospectively and recorded on standardized
`questionnaire forms. If a neoplastic obstructive lesion
`was found during colonoscopies, it was biopsied and
`the patient was included in the study despite colono-
`scopy not being complete.
`
`Faecal occult blood tests
`
`Stool samples were applied to the three windows of
`a Guaiac-FOBT Haemoccult
`(Beckman Coulter,
`Inc.,
`Fullerton, CA, USA) and further processed without
`rehydration as indicated by the manufacturer. On an
`especially Guaiac resin impregnated filter paper stool
`sample was placed. Once it had dried alcohol-stabilized
`hydrogen peroxide was put on the paper. A positive
`test was indicated by colour change to blue after 60 s.
`A human Hb ELISA (Immundiagnostik AG, Bens-
`heim, Germany) was used as a highly sensitive
`IFOBT. This test is based on Hb antibodies-precoated
`microtitre plates. To 100 mg of faeces wash buffer
`was added, mixed and centrifuged. About 100 L of
`this solution was incubated on the microtitre plates
`precoated with anti-Hb antibodies. After washing
`anti-Hb a peroxidase-labelled antibody was added,
`and was incubated for 1 h. Then, the substrate was
`added and incubated for 20 min and finally the stop
`solution was added so the colour changed from blue
`to yellow. Absorption was determined with an ELISA
`reader at 450 nm against reference value. According
`to the manufacturer a cut-off level of £10 g Hb/mL
`of stool was used.
`An immunological FOB test strip device (Prevent ID
`CC) was used. It is a one-step immunochromatographic
`assay for rapid quantitative detection of human faecal
`blood. One of
`its major advantages, which should
`make it more acceptable to the patients, is the simpli-
`city and non-messy sample collection. Figure 1 dem-
`onstrates the convenience of this test. The stick of the
`sample collection device was sticked into the faeces
`(approximately 2 cm),
`it was then inserted into the
`collection device containing an extraction buffer
`solution and the assembly was shaken thoroughly.
`This procedure was repeated three times sampling the
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 3
`
`

`

`148 N . H O E P F F N E R et al.
`
`faeces from different sites. If the test was not per-
`formed within 1 day of sample collection it was stored
`at 2–8 C (as per the manufacturer’s instructions). Two
`drops of this extracted sample were loaded into the
`sample opening of the test strip device, liquid faecal
`extract flows from the sample opening of the test
`device through the conjugate-release pad.
`Principally it is a sandwich immunoassay in which
`gold-labelled anti-Hb antibodies bind to Hb (if present)
`in the sample. These complexes migrate over the test
`membrane until it passes the test zone, which consists
`of
`immobilized gold-labelled antibody–Hb complex
`and
`generates
`a
`violet
`band within
`10 min
`(Figure 2a,b). To prevent false-negative an internal
`(control) line consisting of anti-IgG antibody has been
`added which indicates that the test has run properly. It
`has a very high sensitivity and can detect human Hb
`up to 10 g/mL.
`The bedside FOBT was easy to use, and results did
`not depend on operator experience. All the three FOBTs
`were performed and reported by an experienced techni-
`cian who was blinded to the patient’s clinical profile.
`
`Statistical analysis
`
`For the statistical comparison of the non-parametrical-
`ly distributed data, the Kruskall–Wallis ANOVA test was
`used (STATISTICA Version 5.0, StatSofts Inc., Tulsa, OK,
`USA). Sensitivity, specificity, positive predictive (PPV)
`and negative predictive (NPV) values of all the three
`FOBTs were determined.25 Specificity was calculated
`for CR neoplasia as a whole group and not separately
`for adenoma and CRC. We used chi-square and Fish-
`er’s exact tests to compare differences in proportions
`amongst the tests. The P-values of <0.05 were consid-
`ered as statistically significant (for sensitivity and spe-
`cificity). Methods
`for proportions were used to
`calculate 95% confidence interval (CI).
`
`Figure 1. Sample collection for
`the bedside immunochemical
`faecal occult blood test (IFOBT;
`immunological test strip device;
`Prevent ID CC).
`
`RESULTS
`
`A total of 387 consecutive patients underwent 407
`tests, i.e. colonoscopies and FOBTs simultaneously. In
`14 patients tests were repeated twice while in three
`patients they were repeated thrice. All
`the patients
`with repeat test were known cases of IBD. Their age
`ranged from 5 to 96 years (median age: 51 years).
`There was slight preponderance of females; 186 males
`and 201 females. The study duration was from Janu-
`ary 2002 to December 2004. The major indications
`for colonoscopy were CRC screening,
`IBD, clinical
`suspicion of CR neoplasia or IBS, unexplained anae-
`mia or diarrhoea, positive FOBT, etc. Comparative
`results of
`the test characteristics of asymptomatic
`screened patients
`and symptomatic patients
`are
`shown in Table 1. About 54 patients had evidence of
`CRC on colonoscopy (Table 2), the sensitivity of these
`tests to detect
`faecal Hb was 37.0%, 74.0% and
`77.7% for the Guaiac-based test, the bedside IFOBT
`and the ELISA (cut-off <10 ng/mL)
`respectively
`(Table 3). Of 18 adenomas diagnosed, six were large
`(>10 mm) and 12 were small, i.e. <10 mm. The sensi-
`tivity of faecal Hb for any neoplasia (CR carcinoma
`plus large adenomas) was 29.1%, 59.7%, and 63.8%
`respectively (Table 3). Of 164 patients with normal
`colonic mucosa, 16 (9.7%) were positive with the
`Guaiac test, nine (5.5%) with the bedside IFOBT,
`whereas six (3.7%) were found positive with ELISA-
`IFOBT, i.e. false-positive.
`About 141 (34.6%) patients revealed a lesion (bleed-
`ing) on colonoscopy which is fairly high when com-
`pared with screening population. This is because, ours
`was a hospital-based study, and mostly included
`symptomatic patients unlike the screening studies that
`usually includes asymptomatic healthy individuals.
`Active bleeding colonic lesions were found in 67
`(43.2%) of total 155 IBD patients examined, 29 with
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 4
`
`

`

`B E D S I D E I M M U N O L O G I C A L F O B T 1 4 9
`
`(a)
`
`Absorbent
`pad
`
`Control
`band
`
`Positive
`band
`
`Liquid faecal
` extract
`
`Nitrocellulose
`membrane
`
`Conjugate
`release pad
`
`(b)
`
`Positive
`
`Negative
`
`Figure 2. (a) Principle of the
`bedside immunochemical faecal
`occult blood test (IFOBT; immu-
`nological test strip device). After
`binding of faecal haemoglobin
`(blue dots) by antihaemoglobin
`antibodies, immuncomplexes are
`detected by a second nitrocellu-
`lose membrane-coupled anti-
`body, resulting in the
`appearance of a positive band.
`The formation of a control band
`indicates the correct function of
`the test and unimpeded flow of
`faecal extracts through the
`device. (b) Photograph of
`Prevent ID CC test strips from a
`faecal specimen negative or
`positive for faecal occult blood.
`
`Table 1. Test characteristics in screening and symptomatic patients
`
`Diagnosis
`
`N
`
`Positive
`Guaiac-based test, N (%)
`
`Positive bedside
`IFOBT, N (%)
`
`Positive
`ELISA-IFOBT, N (%)
`
`Screening patients
`2 (33.3)
`6
`Colorectal neoplasia*
`17 (11.4)
`150
`Normal
`19 (12.2)
`156
`Total
`Symptomatic patients/patients with known disease
`Colorectal neoplasia*
`66
`19 (28.8)
`IBD
`155
`39 (25.2)
`Others
`30
`3 (10)
`Total
`251
`61 (24.3)
`
`4 (66.7)
`8 (5.3)
`12 (7.7)
`
`39 (59)
`68 (43.9)
`3 (10)
`110 (43.8)
`
`4 (66.7)
`5 (3.3)
`9 (5.8)
`
`42 (63.6)
`68 (43.9)
`3 (10)
`113 (45.0)
`
`* CRC + large adenomas.
`CRC, colorectal cancer; IFOBT, immunochemical faecal occult blood test; IBD, inflammatory bowel disease; ELISA, enzyme-
`linked immunosorbent assay.
`
`ulcerative colitis (UC) and 38 with Crohn’s disease
`(CD). About 88 patients had inactive IBD (i.e. no evi-
`dence of bleeding lesion during colonoscopy), 22 UC
`and 66 CD. Two patients with pseudomembranous
`colitis had all the three FOBTs-positive in them. Sensi-
`tivity and specificity of rapid bedside test was found
`to be statistically significant in detecting any bleeding
`lesion (except adenomas) when compared with Guai-
`ac-based tests.
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
`DISCUSSION
`
`The CRC and adenomas bleed intermittently26 and po-
`sitivity of FOBTs depend on degree of blood loss. Gen-
`erally 2 mL of blood is necessary in stool to produce
`positive Guaiac-based tests.5 When compared with
`CRC, adenomas bleed less frequently and also the
`bleeding is small in amount,27 i.e. why sensitivity to
`detect them is lower. Also in the present study, the
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 5
`
`

`

`150 N . H O E P F F N E R et al.
`
`Diagnosis
`
`Total
`number
`
`Positive
`Guaiac-based
`test, N (%)
`
`Positive
`bedside IFOBT,
`N (%)
`
`Positive
`ELISA-IFOBT,
`N (%)
`
`Table 2. Endoscopic diagnoses
`in 407 consecutive patients
`
`Normal*
`Colorectal cancer
`Adenoma
`Large (>10 mm)
`Small (<10 mm)
`IBD
`UC
`Active
`Inactive
`CD
`Active
`Inactive
`Diverticulosis
`Pseudomembranous colitis
`Hyperplastic polyps
`
`164
`54
`18
`6
`12
`155
`51
`29
`22
`104
`38
`66
`10
`2
`4
`
`16 (9.7)
`20 (37.0)
`1 (5.6)
`1 (16.6)
`0 (0)
`39 (25.2)
`
`14 (48.3)
`2 (9.1)
`
`16 (42.1)
`7 (10.6)
`2 (20)
`2 (100)
`0 (0)
`
`9 (5.5)
`40 (74.0)
`3 (16.7)
`2 (33.3)
`0 (0)
`68 (43.9)
`
`27 (93.1
`1 (4.5)
`
`36 (94.7)
`4 (6.0)
`0 (0)
`2 (100)
`0 (0)
`
`6 (3.7)
`42 (77.7)
`4 (22.2)
`3 (50.0)
`0 (0)
`68 (43.9)
`
`28 (96.6)
`1 (4.5)
`
`36 (94.7)
`3 (4.5)
`0 (0)
`2 (100)
`0 (0)
`
`* Includes 13 patients with irritable bowel syndrome and 15 patients with unexplained
`diarrhoea.
`IBD, inflammatory bowel disease; UC, ulcerative colitis; CD, Crohn’s disease; IFOBT, im-
`munochemical faecal occult blood test; ELISA, enzyme-linked immunosorbent assay.
`
`Parameters
`
`Guaiac-based test
`
`Bedside IFOBT
`
`ELISA-IFOBT
`
`Sensitivity (%)
`Adenoma
`Cancer
`Cancer + adenoma
`Ulcerative colitis
`Crohn’s disease
`Specificity (%)
`Cancer + adenoma
`Ulcerative colitis
`Crohn’s disease
`PPV (%)
`Adenoma
`Cancer
`Cancer + adenoma
`Ulcerative colitis
`Crohn’s disease
`NPV (%)
`Adenoma
`Cancer
`Cancer + adenoma
`Ulcerative colitis
`Crohn’s disease
`
`5.56 (0.14–27.29)
`37.0 (24.29–51.26)
`29.1 (19.05–41.07)
`48.2 (29.45–67.47)
`42.1 (26.31–59.18)
`
`18.9 (3.58–41.42)
`74.0* (60.35–85.04)
`59.7* (47.50–71.12)
`93.1 (77.23–99.15)
`94.7* (82.25–99.36)
`
`22.2* (6.41–47.64)
`77.7* (64.40–87.96)
`63.8* (51.71–74.88)
`96.5 (82.24–99.91)
`94.7* (82.25–99.36)
`
`90.2 (84.64–94.32)
`90.9 (70.84–98.88)
`89.3 (79.36–95.63)
`
`94.5 (89.84–97.46)
`95.4* (77.16–99.88)
`93.9* (85.20–98.32)
`
`96.3* (92.21–98.65)
`95.4 (77.16–99.88)
`95.4* (87.29–99.05)
`
`5.9 (0.15–28.69)
`55.6 (38.10–72.06)
`56.7 (39.49–72.90)
`87.5 (61.65–98.45)
`69.6 (47.08–86.79)
`
`25.0 (5.49–57.19)
`81.6 (67.98–91.24)
`82.6 (69.67–91.77)
`96.4 (81.65–99.91)
`90.0 (76.34–97.21)
`
`40.0 (12.16–73.76)
`87.5 (74.75–95.27)
`88.4 (76.56–95.65)
`96.6 (82.24–99.91)
`92.3 (79.13–98.38)
`
`89.7 (84.02–93.88)
`81.3 (74.89–86.70)
`74.4 (67.72–80.28)
`57.1 (39.35–73.68)
`72.8 (61.81–82.13)
`
`91.2 (85.86–94.98)
`91.7 (86.49–95.40)
`84.2 (78.16–89.18)
`91.3 (71.96–98.93)
`96.9 (89.16–99.62)
`
`91.9 (86.72–95.48)
`92.9 (87.99–96.30)
`85.9 (79.99–90.56)
`95.4 (77.16–99.88)
`96.7 (89.32–99.63)
`
`* P < 0.05 compared with Guaiac-based test.
`Values in parentheses indicate 95% CI.
`IFOBT, immunochemical faecal occult blood test; PPV, positive predictive value; NPV,
`negative predictive value; ELISA, enzyme-linked immunosorbent assay; CI, confidence
`interval.
`
`Table 3. Sensitivities, specifici-
`ties, PPV and NPV of three
`FOBTs in the different diag-
`nostic groups
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 6
`
`

`

`sensitivity and specificity of all the three FOBTs per-
`formed were lowest
`for detecting adenomas when
`compared with all the other lesions.
`The results of IFOBT performed on 1 day of stool sam-
`ple has been shown to be almost as specific as Guaiac-
`based testing performed on 3 days of stool sample.21We
`have also used 1-day stool sample for conducting FOB-
`Ts. Faecal Hb loses its antigenicity at room temperature
`mostly because of its degradation by faecal flora,28 so
`we stored stool samples at 4 C before testing.
`One of the studies has demonstrated an increase in
`sensitivity by 7.4% but worsening of specificity by 7.3%
`with rehydration of stool sample with Guaiac-based
`tests.7 We also did not rehydrate the stool sample before
`performing the Guaiac-based test. There is a broad
`range of variation in the sensitivity of Guaiac-based
`tests. According to Ahlquist and Shuber in a review of
`eight prospective Guaiac-based tests against colonosco-
`py, in screening population, showed that sensitivity for
`diagnosing CRC was 26% while for large adenomas it
`was 12%.29 We observed the sensitivity of Guaiac-based
`test for detecting CRC to be 37% and for large adenomas
`16.6%. Although the specificity for Guaiac-based tests
`lies in the range of 88–98% actually it is misleading as
`false-positives outnumber the true positives. Only one
`in 10 subjects with a positive Guaiac-based test harbour
`a CRC.15 Thus, despite high specificity PPV is low.
`The clinical value of technically comparable test
`strip devices for the rapid qualitative detection of
`other human proteins, e.g. cardiac troponin T or pros-
`tate-specific antigen, has already been proven.30, 31
`Here, we have utilized the same principle for detection
`of FOB using an immunological test strip device.
`
`B E D S I D E I M M U N O L O G I C A L F O B T 1 5 1
`
`We found the performance characteristics of the
`bedside IFOBT in detecting CRCs to be comparable
`with ELISA-IFOBT. The sensitivity of this new bedside
`IFOBT for diagnosing CR neoplasia was 59.7% vs.
`63.8% (P 0.29) for ELISA-IFOBT whereas the specifici-
`ty of these two tests were 94.5 vs. 96.3 (P 0.38)
`respectively. Thus,
`the sensitivity and specificity of
`bedside IFOBT was lower than that of ELISA-IFOBT,
`but not statistically significant.
`The sensitivity of both IFOBTs in the present study
`was in the same range or even better than that repor-
`ted by other authors, and far higher than the values
`reported for Guaiac tests without rehydration.32, 33 A
`large screening study revealed sensitivity of IFOBT for
`detecting advanced neoplasia and cancer to be 27%
`and 66% respectively.23
`Upon comparison of rate of positive CR neoplasia in
`earlier large studies we noticed that the positivity rate
`of FOBT varied significantly (Table 4). These differ-
`ences could be because of varied study populations
`and also inherent biases of these studies. As some of
`them were conducted in asymptomatic screening pop-
`ulations whereas others were performed in patients
`who were diseased. Also it is well evident that there is
`a wide range of variation in the rate of positivity of
`FOBT in detecting CR neoplasia, e.g.
`the range of
`HemOccult positivity is 47–84% while that of HemeSe-
`lect is 49–91%. Immunochemical tests were shown to
`be more sensitive than Guaiac-based tests.
`In a study using a simple brush sampling technique
`for collection of stool
`for immunochemical
`‘Insure
`FOBT’ demonstrated improved compliance by 66%.
`About
`82% patients
`preferred
`brush over
`the
`
`Table 4. Comparison of posit-
`ive FOBTs in patients with CR
`neoplasia (CRC and adenoma
`>10 mm; modified from
`Ref.32)
`
`Rate of positive lesions* (%)
`
`Reference
`
`Lesions
`
`HO
`
`SENSA
`
`HS
`
`ELISA-IFOBT
`
`Bedside IFOBT
`
`Ahlquist et al.12
`St John et al.44
`Gopalswamy et al.45
`Allison et al.13
`Greenberg et al.34
`Our study
`
`102
`152
`9
`142
`55
`60
`
`19
`75
`60
`38
`38
`35
`
`–
`84
`78
`47
`47
`–
`
`–
`91
`89
`49
`49
`–
`
`–
`–
`–
`–
`–
`76
`
`–
`–
`–
`–
`–
`71
`
`* Positive test in patients with lesion documented by endoscopy.
`HO, HemOccult II; SENSA, HemOccult II SENSA; HS, HemSelect; IFOBT, immunochemi-
`cal faecal occult blood test; CRC, colorectal cancer; ELISA, enzyme-linked immunosorb-
`ent assay.
`
`ª 2006 Blackwell Publishing Ltd, Aliment Pharmacol Ther 23, 145–154
`
` 13652036, 2006, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2006.02702.x, Wiley Online Library on [10/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1029, Page 7
`
`

`

`152 N . H O E P F F N E R et al.
`
`traditional sampling method using the spatula for per-
`forming FOBT.19 Similarly in this bedside IFOBT also
`the sample preparation is very simple with minimal
`stool handling, so we hope that it will improve patient
`acceptability for FOBTs.
`Specificity is an important determinant of the cost
`of FOB screening. Even small changes in test specifici-
`ty result in large effects on the number of false-posi-
`tives in a screening study, which are responsible for
`unnecessary colonoscopic examinations. In our study
`the specificity defined by false-positive results if a
`normal colonic mucosa and no other cause of GI
`bleeding were shown to be between 89.3% and 90.9%
`for the Guaiac test, 93.9% and 95.4% for bedside
`IFOBT and 95.4% and 96.3% for the ELISA-IFOBT
`which is comparable with earlier reports in the litera-
`ture.
`One peculiar feature of our study is that subjects
`with variety of symptoms and with suspected or
`known clinical or endoscopic findings were included
`and underwent these tests. As the real focus of our
`work was to evaluate the relative performance charac-
`teristics of each of these three tests in diagnosing pos-
`itive FOB irrespective of clinical diagnosis (whether
`because of CR neoplasia or active IBD) unlike previous
`studies which had excluded patients with known clin-
`ical diagnosis-like IBD, etc.34, 35 Non-malignant bleed-
`ing lesions were diagnosed in 69 of total 141 patients
`showing positive bleeding lesions on colonoscopy.
`There are mainly two large reports in the literature
`mentioning diagnosis of IBD in patients undergoing
`FOBTs. Actually in these studies asymptomatic IBD
`were diagnosed upon performing colonoscopy in
`patients detected positive with FOBTs during CRC
`screening in large sample population.36, 37
`On comparing the costs of these tests, Guaiac-based
`FOBT costs around 1 €, ELISA-FOBT cost 15–20 €
`whereas for bedside IFOBT one has to pay 3–5 € for
`each test performed.
`Newer promising non-invasive methods for CRC
`screening include detection of stool DNA for muta-
`tions-like p53, K-ras, APC, BAT-26, etc. However, ini-
`tial enthusiasm with these tests could not be realized in
`large trials comparing it with Guaiac-based tests.38–41
`The major drawbacks of it are lower sensitivities, exor-
`bitant costs (up to 800$), compliance issue as patient
`needs to send an entire bowel movement with a mini-
`mum weight of 30 g to be sent to the laboratory, which
`may not be acceptable to many patients and last but
`not the least anxiety associated with false-positive tests.
`
`The inherent limitations of our study include, at first
`that a highly selected population undergoing colonos-
`copy for several ‘high-risk’ indications were included,
`and so care should be taken when applying our results
`to an average-risk asymptomatic population. Never-
`theless, we feel that useful information was obtained
`about
`the feasibility and relative specificity of this
`new bedside IFOBT in our population. And secondly
`the use of endoscopy as the standard for the presence
`of bleeding lesion. Though colonoscopy is widely con-
`sidered as the gold standard for detection of colonic
`bleeding lesions, miss rates of up to 6% in detection
`of adenomatous polyps are documented.42, 43 This
`might have caused underestimation of results of FOBT
`in our study.
`The biggest advantages of this new bedside IFOBT
`are its simplicity, rapidity (needs just about 10 min),
`convenience, no need of cumbersome ELISA readers
`and the fact that it can be performed at the patient
`bedside or in doctor’s office by any health care per-
`sonnel. Thus, performing FOBT may become as simple
`as doing serum blood glucose determination using
`glucometer. This advance should make screening for
`CRC an attainable goal and should increase patient
`compliance.
`To conclude, FOBT is the only method of CRC
`screening whose efficacy has been established in rand-
`omized-controlled trials. We have shown that the new
`bedside IFOBT,
`is more specific and appears to be
`more accurate than the Guaiac-based HemOccult test
`for the detection of significant colonic bleeding lesions
`including CR neoplasia. Although this constitutes a
`pilot study only,
`it provides some guidance as to
`which type of FOBT may be appropriate for large-scale
`screening programmes. Future larger-scale screening
`studies are required to evaluate it over a large asymp-
`tomatic sample size.
`
`CONFLICT OF INTEREST
`
`The authors have no conflicts of interest with the
`device manufacturers.
`
`ACKNOWLEDGEMENTS
`
`This work was supported by the Else Kro¨ ner-Fresen-
`ius-Foundation, Bad Homburg (Germany). Authors
`thank Mr Marc Neumann for helping with statistics.
`Also acknowledge the help rendered by Ms Maron
`Hagas and Ms Nada Povse for data management.
`
`ª 2