Patients with atrial fibrillation (AF) are routinely treated with oral anticoagulants (OACs) in order to prevent complications such as stroke. Classes of OACs include vitamin K antagonists and the newer direct-acting oral anticoagulants (DOACs) such as apixaban, rivaroxaban, dabigatran, and edoxaban. Depending on the agent, rates of upper and lower gastrointestinal bleeding (GIB) can vary between 1% and 2% per year.1
When patients with AF treated with OACs develop GIB, a multitude of causes should be considered, one of which includes an occult GI malignancy. A research group led by Rasmussen et al from Denmark recently conducted a retrospective evaluation of nationwide “real-world” data to determine the absolute risks of colorectal cancer (CRC) in patients with AF treated with OACs with and without lower GIB.2
A total of 125,418 patients were included in the study; 85% received vitamin K antagonists with 15% received DOACs (dabigatran 10%, rivaroxaban 3%, and apixaban 2%). Use of aspirin (43%) and non-steroidal anti-inflammatory drugs (16%) was also reported.
Through a maximum of 3 years of treatment, 2576 instances of lower GIB were identified, of which 140 were associated with a diagnosis with CRC within the first year of the lower GIB. The absolute 6-month risk of lower GIB after starting OACs was 0.29% (95% CI, 0.24-0.36) in patients aged ≤65 years compared with 0.95% (95% CI, 0.79-1.14) in patients aged >85 years. The highest risk was found to be during the first month of receiving OACs.
The absolute 1-year risk of CRC after lower GIB ranged from 3.7% (95% CI, 2.2-6.2) to 8.1% (95% CI, 6.1-10.6) in patients aged ≤ 65 and those aged 76 to 80 years, respectively. These risks were lower compared with those patients in the same age groups who did not experience lower GIB (0.16% [95% CI, 0.15-0.16] to 0.53% [95% CI, 0.55-0.87] in patients aged ≤65 and 76 to 80 years old, respectively).
Similarly, patients with AF treated with OACs who experienced lower GIB had increased risk ratios of CRC compared with those patients without lower GIB. These risk ratios varied from 12.3 (95% CI, 7.9-19.0) for patients aged >85 years to 24.2 (95% CI, 14.5-40.4) in patients ≤ 65 years.
The majority of cancer events occurred within the first 2 months after experiencing lower GIB in all age groups. Younger patients (aged ≤65) were more likely to undergo endoscopic evaluation for lower GIB (60.6%; 95% CI, 55.3-65.7) compared with patients aged >85 (37.3%; 95% CI, 32.5-42.4).
This study highlights several key points. The authors found that the absolute risk of lower GIB after starting OACs was higher in older patients, albeit relatively low in both groups over a 6-month span. This risk was highest during the first month, so this is an important detail to elicit from a patient during evaluation for GIB.
The absolute 1-year risk of CRC in patients with AF treated with OACs was not trivial at approximately 4% to 8%. As such, CRC should be strongly considered as a potential etiology of lower GIB in these patients, and endoscopic evaluation should be considered depending on the patient’s overall clinical status. As the majority of CRC-related events occurred within the first 2 months of OAC treatment, this may be the ideal time to undergo endoscopic evaluation.
This study had several limitations to consider. As the data were taken solely from a Danish registry, all of the results may not be applicable to patients seen in the United States. Not all patients underwent endoscopic evaluation; therefore, the results may represent an underestimation of the true CRC risk. As overwhelming majority of patients included were receiving vitamin K antagonists, future studies will need to report more DOAC-specific data as these agents have become more frequently used. The exact determination of how a lower GIB was made was not detailed, as this can sometimes be clinically challenging in real-world settings. It would be interesting if the authors had additional data on other GI cancers outside of CRC, as lesions within the GI tract could also potentially contribute to GIB.
When considering colonoscopic evaluation in any patient, especially those highlighted in this study, it is important to weigh the risks and benefits of the actual procedure, as many of these patients can be at higher than average risk for complications, especially those with significant comorbidities and those patients aged >80 years. A systematic review and meta-analysis conducted by Day et al found that in patients aged ≥80 years, the cumulative risk of GI adverse events (perforation, bleeding, and cardiovascular/pulmonary) was 34.9 (95% CI, 31.9-38.0) per 1000 colonoscopies.3 The risk was higher compared with patients aged ≥65 years at 26 per 1000 colonoscopies (95% CI, 25-27).
Risk associated with colonoscopy can vary based on multiple factors including age, indication, and intraprocedural techniques such as polypectomy. However, the risks highlighted in Day et al are considerably higher than the cumulative risks for younger, healthier patients reported in the literature.4
1. Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet. 2014;383(9921):955-962.
2. Rasmussen PV, Dalgaard F, Gislason GH, et al. Gastrointestinal bleeding and the risk of colorectal cancer in anticoagulated patients with atrial fibrillation [published online February 7, 2020]. Eur Heart J. doi:10.1093/eurheartj/ehz964
3. Day LW, Kwon A, Inadomi JM, Walter LC, Somsouk M. Adverse events in older patients undergoing colonoscopy: a systematic review and meta-analysis. Gastrointest Endosc. 2011;74(4):885-896.
4. Kothari ST, Huang RJ, Shaukat A, et al. ASGE review of adverse events in colonoscopy. Gastrointest Endosc. 2019;90(6):863-876.