The Risk of Lower Gastrointestinal Bleeding in Low-dose Aspirin Users

W.-C. Chen; K.-H. Lin; Y.-T. Huang; T.-J. Tsai; W.-C. Sun; S.-K. Chuah; D.-C. Wu; P.-I. Hsu


Aliment Pharmacol Ther. 2017;45(12):1542-1550. 

In This Article


According to our review of the relevant literature, our nationwide cohort is the largest used in a study specifically evaluating the risk of LGIB in low-dose ASA users. We observed that low-dose ASA use was associated with an increased incidence of LGIB. NSAIDs, steroids, SSRIs, PPIs and H2RAs were independent risk factors for LGIB. Our study provides evidence on the risk of low-dose ASA-related LGIB. Although the risk is low, the present findings may facilitate the management of low-dose ASA users.

The incidence of LGIB within 1 year in low-dose ASA users was 0.20% in our study; this result is consistent with a previous study on NSAIDs/ASA-related LGIB.[8] ASA causes gastrointestinal complications, including peptic ulcers and ulcer bleeding. Nevertheless, observational studies have revealed a heterogeneous association between low-dose ASA and LGIB.[11–13] ASA use has been reported to increase the risk of diverticular bleeding.[11,12] ASA-related LGIB could be attributed to the antiplatelet activity of ASA, which induces pre-existing lesions, such as diverticula, and causes direct injury of the gastrointestinal mucosa.[8] However, ASA use alone was not significantly associated with LGIB in a Japanese study.[13] Moreover, most studies have primarily investigated the risk of LGIB associated with NSAIDs, ASA and anticoagulants. Our results revealed that low-dose ASA users are at an increased risk of LGIB. In addition to the prevention of thromboembolic diseases, a recent systemic review with meta-analysis suggested that low-dose ASA is as effective as flexible sigmoidoscopy or guaiac faecal occult blood testing in colorectal cancer prevention.[14] With the increasing use of low-dose ASA, patients requiring this antiplatelet therapy should be informed about its adverse effects on the lower gastrointestinal tract. However, it is uncertain whether low-dose ASA users with LGIB should discontinue ASA. A study showed that continuation of ASA use is associated with an increased risk of recurrent bleeding, but a reduced risk of severe cardiovascular events and death in ASA users with a history of LGIB.[9] Further studies are warranted to evaluate the risk of recurrent bleeding in patients switching from low-dose ASA to clopidogrel treatment.

The risk factors for ASA-related GIB include a previous history of ulcers or GIB, corticosteroids use, anticoagulant therapy and use of NSAIDs other than ASA.[4] NSAID use was significantly associated with LGIB in our study. Moreover, NSAIDs induce direct injury of the upper gastrointestinal mucosa. Furthermore, oral and parenteral NSAIDs markedly increased the risk of lower gastrointestinal adverse events, including LGIB.[15] Concomitant use of NSAIDs and low-dose ASA was more likely associated with LGIB than the use of NSAIDs alone.[13] Our results imply that LGIB symptoms should be closely monitored in low-dose ASA users receiving NSAIDs.

Several studies have reported conflicting results on the association between steroid use and GIB. A meta-analysis revealed that the risk of GIB is not increased in patients using steroids. Nevertheless, a recent large meta-analysis revealed that steroid use is associated with an increased risk of GIB.[16] Furthermore, steroids increased the risk of UGIB in low-dose ASA users.[17] Steroids were associated with LGIB in our study. Although the mechanism underlying steroids-related LGIB is not well defined, steroids may impair the process for repairing preexisting lower gastrointestinal lesions, thus increasing the risk of LGIB.

The antiplatelet properties of SSRIs have been well demonstrated.[18] A study on patients with congestive heart failure revealed that SSRIs provide additional antiplatelet protection in patients receiving ASA.[19] However, several studies have showed that patients using SSRIs are at an increased risk of GIB.[20,21] Concomitant use of SSRIs and ASA has been reported to further increase the risk of UGIB.[22,23] We observed that the concomitant use of SSRIs in low-dose ASA users is associated with an increased risk of LGIB, which may be due to the enhanced antiplatelet properties that cause bleeding of the pre-existing lower gastrointestinal lesions, similar to that observed after using low-dose ASA alone.

Nitrovasodilators have been associated with a decreased risk of UGIB in patients taking NSAIDs.[24,25] However, a subsequent study did not identify nitrates as a protective factor against UGIB in low-dose ASA users.[26] Although nitrates were significantly associated with LGIB in univariate analysis in our study, the association was not significant in multi-variate analysis. Therefore, the effects of nitrates on the risk of LGIB in low-dose ASA users are yet to be clarified.

Anti-secretory drugs are typically associated with a decreased risk of UGIB in ASA users.[25] However, these drugs do not exert protective effects on the lower gastrointestinal tract. Gastroprotective agents, including PPIs and H2RAs, were unexpectedly found to be associated with LGIB in our cohort. In a Japanese study on patients receiving low-dose ASA, PPIs and H2RAs were risk factors for GIB in univariate analysis.[27] Another study comparing ASA users and controls reported an increased risk of UGIB in patients with a history of PPI use and a current or past history of H2RA use.[28] Increased LGIB risk was observed in ASA users receiving PPIs.[8,29,30] PPIs are considered as an indicator of poor comorbidity status.[31] Moreover, patients using PPIs and H2RAs were likely to receive NSAIDs, which might cause direct injury of the lower gastrointestinal tract. Furthermore, PPIs might affect the gut microbiome and increase the risk of Clostridium difficile infection or microscopic colitis,[32,33] that might increase the risk of diarrhoea and LGIB. PPIs and H2RAs have mostly been used in patients who received a diagnosis of peptic ulcers as proved by endoscopy or upper gastrointestinal series. These patients remained at a risk of LGIB even after PPI or H2RA use.

Our study has several limitations. First, the observation period was 1 year; therefore, the study results may not be applicable to long-term low-dose ASA users. Second, detailed patient information were not available because of the inherent limitations of the NHIRD. Patient compliance was not examined, and the effects of medications, including low-dose ASA, on LGIB should be interpreted cautiously. Third, the use of over-the-counter medications was not evaluated in this study; consequently, the actual risk of low-dose ASA-related LGIB might have been overestimated. Fourth, more patients in the low-dose ASA group experienced ischaemic stroke and used pro-bleeding drugs. Patients using antiplatelet drugs are more likely to have comorbidities and use pro-bleeding drugs.[34] However, low-dose ASA remained an independent risk factor for LGIB after adjustments for age, gender, comorbidities and pro-bleeding drugs. Finally, ICD-9-CM was used in the NHIRD during the study period. Studies using 10th Revision of International Classification of Diseases are warranted to evaluate the risk of LGIB in long-term low-dose ASA users.

In conclusion, the risk of LGIB was higher in low-dose ASA users than in ASA nonusers in this nationwide cohort. Low-dose ASA, NSAIDs, steroids, SSRIs, PPIs, and H2RAs were independent risk factors for LGIB. Patients requiring antiplatelet therapy should be informed about the risk of LGIB before low-dose ASA use, and a strategy should be developed to prevent and monitor LGIB in low-dose ASA users.