Abstract
The format of the traditional guaiac faecal occult blood test (gFOBT), particularly the collection technique, might cause difficulties for some. A multistage evaluation of alternative tests was performed. Firstly, four tests with different faecal collection approaches were assessed: a focus group recommended further investigation of a wipe gFOBT. Secondly, 100 faecal samples were analysed using two wipe tests and the routine gFOBT: no differences were found. Thirdly, a wipe gFOBT was introduced. Over 21 months, 400 requests were made and 311 wipe kit sets were submitted for analysis: 153 (49.2%) were negative, 21 (6.8%) positive (all 3 kits positive), 96 (30.9%) weak positive (1 or 2 positive) and 41 (13.2%) un-testable. Forty-three participants were referred for colonoscopy. Outcome data were provided on 39 participants: nine declined colonoscopy, two were judged unsuitable, two did not attend, two were already in follow-up, 13 had normal colonoscopy and two normal barium enema, two had diverticular disease, two had a metaplastic polyp, four had a low-risk adenoma and one had a high-risk adenoma. No participant had cancer. Detection of significant neoplasia was small. The use of the wipe gFOBT was ceased: it cannot be recommended as a screening test for bowel cancer.
INTRODUCTION
Bowel (colorectal cancer) screening using guaiac-based faecal occult blood tests (gFOBTs) has been shown to reduce disease-specific mortality in randomized controlled trials. 1 As a result, three demonstration pilot screening rounds were carried out in Fife, Grampian, and Tayside in Scotland, 2 while all the necessary prerequisites for the Scottish Bowel Screening Programme (SBoSP) were being organized for the roll out that began in June 2007 and is now complete. 3 In the SBoSP, all individuals aged 50–74 years are invited to participate every two years.
The format of the traditional gFOBT used in the pilot screening rounds and as the initial investigation in the SBoSP requires the collection of two samples from each of three faecal specimens onto a single card. The collection may be seen as a somewhat unpleasant procedure that involves close handling of faeces. Moreover, the collection technique might lead to difficulties for some, particularly the visually impaired and those with poor manual dexterity. Any screening programme should facilitate participation by all who are eligible. Therefore, it was decided to investigate faecal testing kits with different collection methods to establish whether an alternative test could be adopted that was more suitable for participants with disabilities, to assess the analytical performance of the potentially suitable test, and to establish whether the alternative test objectively selected (a wipe gFOBT) was appropriate for ongoing use in the SBoSP.
METHODS
As a first stage, four gFOBT with different collection devices were selected for initial assessment. These were:
The gFOBT kit used in the pilots, which required two samples from each of three faeces collected by wooden stick and applied to the six windows of the card (hema-screen, Immunostics Inc, Ocean, NJ, USA); A second gFOBT that used a wipe of the anus after passing of faeces technique which transfers faeces to guaiac paper integral in the device: three separate wipes are required (HemaWipe, MedTek, Morrisville, NC, USA); A faecal immunochemical test (FIT) kit that used a small specimen collection tube with a lid with an integral stick with serrations: the stick is stabbed into the faeces and then replaced in the tube containing buffer and qualitative analysis done on an immunochemical test cassette (Instant-View, Alfa Scientific Designs Inc, Poway, CA, USA); and A second FIT that used a brush collection technique to sample faeces in the toilet by wiping the brush across the faeces or collecting some of the toilet water near the faeces and ‘painting’ the material from two specimens onto a single card which then had a immunochemical test strip inserted for qualitative analysis (InSure, Enterix Inc, Sydney, Australia).
After consideration of advantages and disadvantages by the authors, the Bowel Screening Services Manager discussed the four options with a focus group comprising 58 people (53 women and 5 men, aged from 58 to 92 years) suffering from varying degrees of arthritis who attended a bowel cancer awareness session.
Secondly, to assess analytical performance, 100 randomly selected samples submitted for routine gFOBT analysis in the Department of Biochemical Medicine, Ninewells Hospital and Medical School, Dundee, were analysed using HemaWipe (3 wipes), an almost identical wipe test, EZ Wipe (Immunostics Inc, Ocean, NJ, USA) (3 wipes) and hema-screen gFOBT (all 6 windows).
Test results were collected from the Scottish Bowel Screening Centre Laboratory and the colonoscopy and pathology data requested from, and provided by, the NHS Board of those participants referred with positive results.
RESULTS
On the basis of professional opinions and those of the focus group, EZ Wipe was adopted for use in the SBoSP. New concise instructions for use by participants were prepared, similar to those used in the SBoSP. Standard operating procedures for staff to respond to requests for the alternative test were generated.
Analysis of 100 faecal samples with Hema-Wipe against hema-screen gFOBT, and EZ Wipe against hema-screen gFOBT, showed no differences (kappa = 1.00).
To the cessation of the offering of the alternative test (December 2008), 400 requests were made for wipe gFOBT. Participants were notified, as was their general practitioner (GP), and referred for colonoscopy if all three initial wipe gFOBT were positive. If all were negative, participants were sent the appropriate letter as used in the SBoSP. If at least one of the three wipe gFOBT was positive, another test pack was sent and, if any of the second kit set was positive, participants were notified, as was their GP, and referred for colonoscopy: if all were negative, participants were sent the appropriate letter. Those who sent in an un-testable kit set were sent another pack. Overall, 532 test packs of three wipes were issued: this comprises 400 initial and 132 follow-up and replacement packs: 311 wipe kit sets (58.5% of those issued) were submitted for analysis: 153 (49.2%) were negative, 21 (6.8%) positive (all 3 kits positive), 96 (30.9%) weak positive (1 or 2 tests positive) and 41 (13.2%) un-testable.
Forty-three screening algorithm positive participants were referred for colonoscopy. Outcome data were available on 39 participants and these are shown in Table 1. Classification of polyps was done using the guidelines of the British Society of Gastroenterology. 4
Outcomes for participants with positive wipe gFOBT results
DISCUSSION
Initial professional evaluation of the four candidate tests led to the decision that a FIT should not be used as an initial investigation because (a) the analytical detection limits are generally about 10 times lower than that of gFOBT, (b) positivity rates reported are about 10% in contrast to the 2% required for the SBoSP and (c) costs are significantly higher. 5 Thus, it was decided that a gFOBT was required as an alternative test. Interestingly, those questioned at the focus group meeting stated that, if a participant could manage their own personal hygiene, but had failing sight or manual dexterity difficulties, the wipe gFOBT would be the most suitable alternative.
Since the results obtained on analysis of 100 randomly selected faecal specimens submitted for routine gFOBT showed no differences, the analytical detection limits (sensitivity) of both wipe gFOBT studied were inferred to be as per hema-screen gFOBT, and thus are 0.6 mg haemoglobin/g faeces, as previously documented. 6 In consequence, it was hoped that the positivity rates, and thereby the colonoscopy demand, might be the same for the alternative wipe gFOBT as the traditional gFOBT used. With the wipe gFOBT, participants can be referred for colonoscopy either on first finding that all three wipes are positive (strong positive) or, following a similar algorithm to that used in the pilots 2 , on finding that at least one of the three in each of two successfully completed wipe tests are positive (weak positive). There were 49.2% negative, 6.8% strong positive, 30.9% weak positive and 13.2% un-testable results, respectively. However, for the same time period as this study, for the routine initial gFOBT in the SBoSP, the same percentages were 90.7, 0.4, 7.6 and 1.3: thus, the wipe gFOBT had many more positive, weak positive and un-testable results compared with the usual gFOBT, leading not only to a greater demand for repeat wipe gFOBT packs but, more importantly, to a much greater colonoscopy demand. The pilots and SBoSP were set up with screening algorithms designed so that the positivity rate was 2.0%, a rate that could be coped with by the scarce colonoscopy resources available in Scotland. This would be far exceeded if a wipe gFOBT was used in substantial numbers.
A large number (39%) of participants with positive wipe gFOBT results declined colonoscopy, were judged unsuitable for colonoscopy, did not attend or were already in follow-up: this may reflect the fact that those who asked for the alternative test did have some disability that impacted on the decision not to proceed with the second-line definite investigation. There are a number of plausible reasons for the high rate of un-testable wipe gFOBT: individuals with visual impairment or poor dexterity might have had real difficulties in performing the collection adequately: many of the wipe gFOBT returned were torn, had insufficient faeces, were received after the 14 days expiry time allowed or fewer than three kits were returned. There are also possible reasons why the positivity rate was high: it may be that those individuals who used the wipe gFOBT had taken dietary constituents or medicines, such as warfarin, aspirin or NSAID, that have been suggested to give false-positive results. 7 This postulate is borne out by the high number (39%) of normal colonoscopies and barium enema.
There appear to be no clinical trials using wipe gFOBT in bowel screening. The only investigation is on the utility of wipe gFOBT for simulated upper gastrointestinal tract bleeding: moreover, it was stated that the wipe approach did not produce false-positive results in those with haemorrhoidal disease. 8 Our data show that the wipe gFOBT gives a very low yield of significant disease and that the performance characteristics are different from those in the SBoSP, 9 so that, along with the other disadvantages discussed above, cannot be recommended as a suitable alternative test for use in bowel screening. This is not to say that the search for an alternative test for those with disabilities should cease. We firmly believe that any screening programme should be accessible to all who make the informed choice to participate. It may be that either imaginative alterations to the traditional gFOBT format or, more likely, the use of the newer compact collection devices that are available as collection devices for the single faecal sample required for quantitative FIT might assist in this laudable goal.