In many countries of the world, the fecal immunochemical test for hemoglobin (FIT) is the main pillar of colorectal cancer (CRC) screening programs. The age range targeted by these programs, typically ≥ 50 years, corresponds to the age range in which cardiovascular morbidities are increasingly prevalent. Given that CRC and cardiovascular disease share several risk factors, those with cardiovascular disease are an important target group of CRC screening. This includes patients treated with oral anticoagulants (OACs), which are indicated for the prevention and treatment of thromboembolic disease and the prevention of ischemic stroke in patients with atrial fibrillation. It is known for both types of OAC (i. e. vitamin K antagonists [VKAs] and direct oral anticoagulants [DOACs]) that they increase the risk of gastrointestinal (GI) bleeding.

This raises the question of whether or not OACs have an impact on FIT screening. This impact could be negative and/or positive. “Negative” refers to the concern that anticoagulation therapy may increase the chances of a false-positive test result, thereby decreasing the positive predictive value (PPV) of FIT. “Positive” refers to the possibility that OACs might cause minor bleeding from colorectal lesions and therefore unmask them, rendering them detectable through occult blood testing.

An intriguing paper published in the current issue of Endoscopy provides new insights into this research question [1]. It reports on a cross-sectional study conducted in the context of the first round of the Danish CRC screening program (2014–2016). Available data included information on FIT participation, antithrombotic treatment before FIT testing, the FIT result, use of follow-up colonoscopy after a positive FIT, and the findings at follow-up colonoscopy, as well as age, sex, and the Charlson co-morbidity index. Overall, 884 036 people aged 50–74 years were invited for FIT screening, of whom 551 570 people (62.4 %) participated. Other studies on this topic often used the old guaiac test rather than FIT or had a substantially smaller sample size.

What did this study show with regard to, first, the false-positive test results and the PPV? Among people receiving any antithrombotic treatment, 11.8 % tested positive, as compared with 6 % among people not receiving antithrombotic treatment. Considering those patients who were using only one type of antithrombotic treatment (i. e. not on combination therapy), DOACs were associated with the largest increase in FIT positivity, while VKAs and other antithrombotic drugs showed a modest increase. The PPV, assessed based on those with a positive test result undergoing follow-up colonoscopy, was lower in users of antithrombotic drugs compared with those without treatment; it was particularly low among those treated with DOACs. Therefore, users of these drugs were more likely to have a false-positive test (i. e. specificity was lowered).

In addition, there were interesting side results: overall, attendance for FIT screening was markedly higher among people who were not receiving antithrombotic treatment (64.6 %) than among users of DOACs (48.2 %); it was also higher than among users of VKAs (61.4 %) and users of acetylsalicylic acid (57.5 %). Furthermore, the uptake of follow-up colonoscopy after a positive FIT was higher among people not receiving antithrombotic treatment (86.5 %) than among users of DOACs (79.3 %) and other antithrombotic drugs as single therapy (80.0 %–83.5 %). This clearly indicates that there is uncertainty among users of antithrombotic drugs or their physicians with respect to FIT screening and colonoscopy, particularly regarding FIT and DOAC use.

“Even though the findings clearly show that OAC users are more likely to have a false-positive test, treatment cessation would be risky regarding the underlying cardiovascular disease. Furthermore, the detection rate was higher among OAC users, so there could be a beneficial effect as well.”

The second part of the results refers to the question of a potential increase in the sensitivity of FIT through treatment with OACs. To explore this question, the authors compared the detection rate (CRC and high risk adenomas) between participants using antithrombotic drugs vs. nonusers of these drugs (i. e. this analysis included all participants and not only those who tested positive). After adjusting for age, sex, and the Charlson co-morbidity index, the detection rate was about 30 % higher among OAC users compared with nonusers.

This might indicate that OACs “unmask” colorectal lesions, but caution is needed in the interpretation of this finding. First, GI bleeding not only leads to more false-positive tests but also increases the likelihood that colorectal lesions are detected by chance, so OACs may not specifically induce the bleeding from colorectal lesions. Second, the comparison was not randomized. Observational studies try to emulate randomization by confounder adjustment, but this is subject to the availability of information on relevant confounders. In this study, the information was limited to age, sex, and a co-morbidity score. Several risk factors for colorectal neoplasia possibly associated with OAC use were missing. Residual confounding could therefore have biased the results. The expected direction of this bias would be to overestimate the difference in detection rates between OAC users and nonusers, given the common risk factors of CRC and cardiovascular diseases so, in truth, there may be no difference or a smaller difference.

Should current guidelines, which do not recommend the cessation of antithrombotic treatment before fecal sampling, be revisited in view of this study’s findings? Even though the findings clearly show that OAC users are more likely to have a false-positive test, treatment cessation would be risky regarding the underlying cardiovascular disease. Furthermore, the detection rate was higher among OAC users, so there could be a beneficial effect as well. Whether this is due to chance (a general increase in the likelihood of GI bleeding through the use of OACs) or whether this is a real improvement of test accuracy (increased likelihood that colorectal lesions bleed due to the use of OACs) remains to be clarified – ideally, based on a randomized controlled trial (RCT). For acetylsalicylic acid, the putative unmasking effect that was assumed based on observational data could not be reproduced in an RCT [2] [3].

Another important point to address is the obvious uncertainty among OAC users regarding participation in FIT screening and follow-up colonoscopy. A systematic risk–benefit assessment of FIT screening for these patients is needed, considering the increased risk of bleeding at colonoscopy/polypectomy among these patients [4], as well as the findings of the current study with regard to false positives and detection rates. This will be important to allow OAC users to make an informed decision about participation in FIT screening.

Article published online:
24 February 2023

© 2023. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany