|Home | About | Journals | Submit | Contact Us | Français|
To determine whether medication interventions enhance the sensitivity and specificity of guaiac-based fecal occult blood testing (FOBT) when screening for colorectal cancer (CRC).
We searched PubMed-MEDLINE, CINAHL, and the Cochrane databases using the MeSH headings occult blood, feces/analysis, and guaiac/analysis, linking them to variations of anticoagulants, heparin, warfarin, iron, aspirin, nonsteroidal anti-inflammatory drugs (NSAIDs), clopidogrel, cyclooxygenase-2 inhibitors, and ascorbic acid (vitamin C). Study selections were limited to English studies involving humans.
All resulting titles and abstracts were reviewed for studies that included manipulation of medications associated with guaiac-based FOBT. If the study’s relevance was unclear from the abstract, the full article was reviewed. The search resulted in 31 pertinent studies.
No studies addressed the effects of medication interventions on the sensitivity or specificity of FOBT screening. Randomized controlled trials, however, showed no increase in the rate of positive results among those taking NSAIDs. The literature is mixed regarding the effect of NSAIDs on the positive predictive value of a positive FOBT result, although no change in positive predictive value has been shown for warfarin. Iron will not affect FOBT results in vivo. Ascorbic acid might inhibit positive FOBT results both in vitro and in vivo, but it has not been studied in screening populations.
Studies evaluating the effects of medication intervention on FOBT screening for CRC are limited by their lower quality and because they do not address sensitivity and specificity. Available evidence, however, does not suggest a benefit from withholding NSAIDs, anticoagulant medications, or iron during the screening period. These recommendations should be abandoned in order to maximize adherence to screening. Positive FOBT results obtained among patients taking these medications deserve full evaluation for CRC. Until further studies clarify the effect of ascorbic acid on FOBT screening, withholding this medication before testing seems prudent.
Fecal occult blood testing (FOBT) remains an appropriate screening option for colorectal cancer (CRC).1,2 However, FOBT is limited by a lack of patient adherence. A consistently reported barrier is the lack of physician endorsement for screening.3–9 Other barriers include embarrassment or unpleasantness (the “yuk” factor), anxiety regarding results, lack of symptoms or health concerns, and practical issues such as inconvenience and cost.10 Dietary restrictions11–13 and medication restrictions13,14 have also been implicated as barriers to screening.
The most recently reported survey data, published in 2007, showed that only 30.2% of eligible Canadians were adherent to CRC screening guidelines.15 Clearly, addressing barriers to screening is a worthy task for physicians who endorse screening for CRC. A recent systematic review suggested that dietary restrictions were unnecessary when performing FOBT,16 a position endorsed by Cancer Care Ontario.17 The Ontario panel also advised against restricting medications other than ascorbic acid before FOBT screening. A systematic review of bleeding risks among patients taking anticoagulant or antiplatelet medications concluded that these medications did not diminish the positive predictive value (PPV) of the FOBT.18
Manufacturers of guaiac-based FOBTs available in Canada continue to recommend medication restrictions before testing. The patient instructions for Hemoccult II and Hemoccult II SENSA recommend avoiding nonsteroidal anti-inflammatory drugs (NSAIDs) for 7 days, but allow up to 1 adult acetylsalicylic acid (ASA) tablet daily. Both recommend avoiding ascorbic acid supplements and citrus juices.19,20 The insert for Tri-Slide suggests consideration of avoiding gastric irritants such as NSAIDs and anticoagulant medications after discussion with the physician. Avoidance of ascorbic acid is also recommended.21 ColoScreen and Hema-Screen recommend continuing usual medications with the exception of ascorbic acid.22,23 Despite a lack of recommendations from manufacturers, 16% of primary care physicians,24 32% of gastroenterologists,25 and 10% of internal medicine residents26 have reported discontinuing anticoagulant medications before FOBT. Forty-seven percent of gastroenterologists have reported withholding iron supplements.25
This systematic review explores the evidence for medication restrictions before CRC screening with guaiac-based FOBT.
PubMed-MEDLINE, the Cochrane databases, and CINAHL were searched using the key words and MeSH terms occult blood, feces/analysis, and guaiac/analysis. These terms were linked to variations of anticoagulants, heparin, warfarin, clopidogrel, iron, aspirin, NSAIDs, cyclooxygenase-2 inhibitors, and ascorbic acid (vitamin C). The search was limited to English studies involving humans. References from relevant articles were also explored for further resources.
Our PubMed search revealed 417 articles. Cochrane and CINAHL database searches revealed no additional studies. All titles and abstracts were reviewed for studies that included the manipulation of medications before or during guaiac-based FOBT. If the study’s relevance was unclear from the abstract, the full article was reviewed. The resulting 31 pertinent studies are summarized in Tables 1 to to33.27–57
There is a common assumption that FOBT screening for patients taking NSAIDs or anticoagulant medications is adversely affected by the increased predisposition to upper gastrointestinal (GI) bleeding and by increased false-positive rates among these patients. If so, this reduction in specificity would result in unnecessary follow-up evaluations. There might also be a danger if one assumes that a positive FOBT from a patient taking these medications has a reduced PPV. Attributing a positive result to a medication effect could mistakenly result in neglecting further necessary evaluations.
Unfortunately, there are no studies that evaluate the effects of NSAIDs and anticoagulant medications on the sensitivity and specificity of guaiac-based FOBT screening. Such studies would necessitate evaluating subjects with negative FOBT results for lower GI pathology. Most studies look at either the rate of positive FOBT results or the PPV of a positive result.
Only 2 randomized controlled trials compared rates of positive guaiac-based FOBT results between users of NSAIDs and control subjects, with no significant difference found between groups.27,28 Several randomized trials compared rates of positive FOBT results between ASA and ibuprofen,29 between various formulations of ibuprofen,30 and among ASA, warfarin, or both in combination.31 None showed differences between groups, but there were no control groups.
Cross-sectional studies reported low rates of positive Hemoccult results (0% to 5.5%) among patients taking anti-inflammatory agents.32,33 The PPV for CRC or large adenomas of a positive FOBT result, however, was 25% in a group of rheumatoid disease clinic patients using NSAIDs,32 which compares favourably with the PPV for positive Hemoccult screening of up to 17.9% in the general population reported in a large Minnesota trial.58 This substantial PPV makes it clear that a positive result should not be attributed solely to NSAID use.
Two cohort studies looking at the effects of NSAIDs on FOBT showed inconsistent results. A non-randomized prospective crossover study of medical patients taking either no ASA, 81 or 325 mg of ASA daily, or warfarin showed no difference in the rates of positive Hemoccult II results between the control period and different treatments.34 Quantified fecal blood in controls increased slightly when they switched to 325 mg of ASA daily (P = .02). This difference, however, disappeared when controls were compared with all patients taking ASA in the study (P = .14). Conversely, in a cohort study of subjects screened with Hemoccult II, NSAID users had a significantly higher rate of positive results at 27% compared with 4% for those not taking NSAIDs (P < .01).35 These results, however, might not be valid because they relied on self-reported adherence to the instructions to withhold NSAIDs during the screening period. Only 26 of 1797 subjects admitting to continuing NSAIDs during the testing. Neither study reported follow-up of positive FOBTs, so PPV was undetermined.
Several cross-sectional studies examined patients with positive FOBT results to determine if the PPV was affected by NSAIDs. Two showed a sizeable PPV of around 20% among NSAID users with no difference from controls.36,37 Two others, however, noted lower PPVs among NSAID users than controls.38,39 However, as in the previous studies, PPVs among the NSAID users were substantial enough to justify further evaluation.
One study evaluated the effect of ASA on FOBT of patients with known CRC.40 It found that 600 mg of ASA given daily for 3 days did not increase the rate of positive Hemoccult II results in this group. Chromium-51 red blood cell labeling confirmed no increase in quantified fecal blood loss when taking ASA.
There is little literature exploring the effects of anticoagulant medications on FOBT results. A single cross-sectional study of 849 patients taking dicumarol revealed a rate of positive Hemoccult II results of 15%.41 Further evaluation of most of these patients showed a PPV of 19% for CRC and adenomas.
Studies comparing the rates of positive FOBT results for subjects taking anticoagulant medications with the rates for controls showed mixed results. One study showed no difference in the rate of positive FOBT results between warfarin users and controls, nor was there a difference in quantified fecal blood.34 Two others showed an increased rate of positive FOBT results among those taking warfarin or heparin compared with controls.42,43 The PPV for CRC or adenomas among users was 12.5% in the first study, but the PPV for controls was not reported. In the second study the PPV of 19.8% was not significantly different from that for controls.
A single randomized trial compared rates of positive FOBT results among survivors of myocardial infarction who were randomized to receive warfarin, ASA, or a combination.31 There were no significant differences in the rates of positive results between the group receiving warfarin and those receiving ASA or ASA plus warfarin. Unfortunately, a control group was not included in this study.
Four studies compared the PPV of a positive FOBT result between users and nonusers of anticoagulant medications. Three showed no statistically significant difference in the PPV between users and nonusers.39,43,44 In the fourth, the PPV for large polyps or tumours was higher for warfarin users at 16.1% compared with 11.4% for nonusers (P < .01), but this difference disappeared after adjusting for age and sex.45
Only a single study evaluated the effect of clopidogrel on guaiac-based FOBT results.39 This cross-sectional study of patients who presented for colonoscopy following a positive screening FOBT result showed a PPV for advanced neoplasia of 7.3% for those taking clopidogrel. Among the control patients the PPV was 30.5% (P < .01).
The literature consistently shows that ferrous forms of iron can produce positive results using standardized guaiac-based test cards such as Hemoccult in vitro.46–48,56,59 However, it is inconsistent regarding the effect of oral iron on in vivo testing using standardized guaiac-based cards. A single small trial reported an increased rate of positive Hemoccult results in 10 subjects given 300 mg of ferrous sulfate 3 times daily and 300 mg of ferrous gluconate 3 times daily.46 Similarly, a single cross-sectional study of patients with GI symptoms compared subjects who reported taking and not taking iron. The rate of positive Hemoccult II results was higher among those reporting iron use.49 However, self-reported iron use did not influence Hemoccult II results in a group of healthy volunteers in this study. No previous or subsequent studies have been able to confirm the ability of oral iron supplementation to cause false-positive results when using standardized guaiac-based cards.47,48,50–54
In the only randomized, prospective, double-blind study reported, 78 healthy volunteers were treated with placebo, 1500 mg of ferrous gluconate daily, or 975 mg of ferrous sulfate daily for 2 weeks each in a crossover trial.54 There was only 1 positive result among 326 samples tested with Hemoccult II and Hemoccult SENSA. The lone positive sample was collected during the placebo phase of the study.
As early as 1936, Barrett noted the ability of ascorbic acid to reverse the colour change of a positive guaiac test.60 The most consistent recommendation among FOBT manufacturers regarding medication restrictions is to withhold ascorbic acid during testing. The primary reference on which this is based is a case report in which an anemic patient ingesting 2 g of ascorbic acid daily was shown to have repeatedly negative Hemoccult but positive benzidine-based FOBT results.55 Results of Hemoccult testing became positive when ascorbic acid was withheld. Subsequent in vitro studies showed that complete inhibition of Hemoccult occurred with a fecal ascorbic acid concentration of 15.4 mg/dL, although the concentration of blood present in the stool was not reported. Other in vitro studies have also confirmed the ability of ascorbic acid to inhibit positive guaiac reactions.59,61,62 The 2 lone human studies involving a total of 8 subjects showed an inhibitory effect of ascorbic acid in 6 of the subjects taking 1000 to 1500 mg daily.56,57
No research has studied the effects of NSAIDs and anticoagulant medications on the sensitivity and specificity of guaiac-based screening FOBTs. However, randomized controlled trials showed no difference in the rate of positive results among subjects taking these medications. Cohort studies revealed mixed findings, although the 1 study showing an increase in positive results depended on self-reported patient compliance with instructions. Most cross-sectional PPV studies (5 of 7) showed no difference in PPV among patients with positive FOBT results taking NSAIDs or anticoagulant medications. Such studies, however, are difficult to interpret because it is unknown whether all patients with positive results agreed to colonoscopy or whether a cohort existed that declined further investigation, a confounder referred to as transfer bias.63 Neither study that quantified fecal blood showed a difference between users and nonusers of these medications.
If such medications increase the risk of upper GI bleeding, why would this not be reflected in a reduction in specificity of guaiac-based FOBTs? This inconsistency likely occurs because blood originating more proximally looses its pseudoperoxidase activity as it transits the gut.64 It is this pseudoperoxidase activity that allows hemoglobin to oxidize guaiac, inducing the blue colour change. During gut transit hemoglobin is broken down to porphyrins, thereby losing this ability. While the specificity of guaiac-based FOBTs has not been evaluated for patients taking NSAIDs or anticoagulant medications, a study using immunochemical FOBTs demonstrated that the use of NSAIDs or anticoagulant medications among high-risk or minimally symptomatic individuals had no adverse effect on specificity when explicitly studied.65
Most important, studies evaluating PPV in patients taking NSAIDs and anticoagulant medications consistently show that the PPV is sufficiently high to justify further evaluation for CRC when a positive guaiac-based FOBT result is encountered.
The effect of iron supplementation on guaiac-based FOBTs has been laid to rest in the literature. While iron can cause a positive guaiac response in vitro, this is not the case in vivo.
While it is clear that as a strong reducing substance, ascorbic acid is able to inhibit a positive guaiac reaction, it is challenging to apply these study results to clinical situations because several unknowns remain. It was shown that in elderly subjects taking 200 mg of ascorbic acid daily, the average stool concentration of ascorbic acid was 2.7 mg/dL.66 However, the degree of absorption and, subsequently, the stool concentration of ascorbic acid when supplementing with higher doses is unknown. Likewise, the amount of fecal occult blood typically present in patients with early, asymptomatic CRC and the concentration of fecal ascorbic acid required to inhibit a positive guaiac response for that amount remain undetermined. Fortunately, daily dosing of ascorbic acid is rarely required, so abstaining during FOBT to prevent false-negative results seems reasonable.
The research evaluating the effects of medications on guaiac-based FOBT screening is varied and of generally poor quality. The preponderance of the literature, however, fails to reveal significant differences in the rates of positive test results or PPV among patients taking NSAIDs, anticoagulant medications, or iron supplements. When positive screening results are encountered, the PPV is sufficiently high to justify further workup in these patients. Ascorbic acid can theoretically inhibit positive guaiac reactions, although this has not been shown in screening populations. However, until further research clarifies its effect on sensitivity, withholding high-dose ascorbic acid supplementation during FOBT screening seems a prudent recommendation.
Fecal occult blood testing (FOBT) remains an effective screening option for colorectal cancer. However, it is frequently limited by lack of patient adherence. Among the barriers is patients’ willingness to follow recommended medication interventions. This systematic review explores the evidence regarding medication restrictions before screening with guaiac-based FOBT. Most of the literature fails to reveal significant differences in the rates of positive test results or positive predictive value among patients taking nonsteroidal anti-inflammatory drugs, anticoagulant medications, or iron supplements. Ascorbic acid can theoretically inhibit positive guaiac reactions, so until further research clarifies its effect on sensitivity, withholding high-dose ascorbic acid supplementation during FOBT screening seems to be a reasonable recommendation.
This article has been peer reviewed.
Cet article a fait l’objet d’une révision par des pairs.
This article is eligible for Mainpro-M1 credits. To earn credits, go to www.cfp.ca and click on the Mainpro link.
Dr Konrad conducted the literature review and prepared and revised the manuscript. Dr Katz contributed to the concept and design of the review, revised drafts of the manuscript, and read and approved the final manuscript.