Major GI bleeding in older persons using aspirin: incidence and risk factors in the ASPREE randomised controlled trial

Abstract

Objective: There is a lack of robust data on significant gastrointestinal bleeding in older people using aspirin. We calculated the incidence, risk factors and absolute risk using data from a large randomised, controlled trial.

Design: Data were extracted from an aspirin versus placebo primary prevention trial conducted throughout 2010-2017 ('ASPirin in Reducing Events in the Elderly (ASPREE)', n=19 114) in community-dwelling persons aged ≥70 years. Clinical characteristics were collected at baseline and annually. The endpoint was major GI bleeding that resulted in transfusion, hospitalisation, surgery or death, adjudicated independently by two physicians blinded to trial arm.

Results: Over a median follow-up of 4.7 years (88 389 person years), there were 137 upper GI bleeds (89 in aspirin arm and 48 in placebo arm, HR 1.87, 95% CI 1.32 to 2.66, p<0.01) and 127 lower GI bleeds (73 in aspirin and 54 in placebo arm, HR 1.36, 95% CI 0.96 to 1.94, p=0.08) reflecting a 60% increase in bleeding overall. There were two fatal bleeds in the placebo arm. Multivariable analyses indicated age, smoking, hypertension, chronic kidney disease and obesity increased bleeding risk. The absolute 5-year risk of bleeding was 0.25% (95% CI 0.16% to 0.37%) for a 70 year old not on aspirin and up to 5.03% (2.56% to 8.73%) for an 80 year old taking aspirin with additional risk factors.

Conclusion: Aspirin increases overall GI bleeding risk by 60%; however, the 5-year absolute risk of serious bleeding is modest in younger, well individuals. These data may assist patients and their clinicians to make informed decisions about prophylactic use of aspirin.

Trial registration number: ASPREE. NCT01038583.

Keywords: aspirin; clinical trials; elderly; epidemiology; gastrointestinal bleeding.

Figure 1

Flow diagram of study participants with bleeding events.

Figure 2

Cumulative incidence of GI bleeding events according to trial arm.

Figure 3

Cumulative incidence of upper and lower GI bleeding events according to trial arm.

Figure 4

Incidence of GI bleeding according to subgroup. Diabetes=self-report of diabetes mellitus or fasting glucose ≥126 mg/dL (≥7 mmol/L) or on treatment for diabetes. Hypertension=systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or on treatment for high blood pressure. Chronic kidney disease (CKD)=estimated glomerular filtration rate < 60 mL/min/1.73 m² or albumin-to-creatinine ratio ≥3 mg/mmol. NSAID, non-steroidal anti-inflammatory drugs; PPI, proton pump inhibitor; SSRI, selective serotonin reuptake inhibitor.

Figure 5

Multivariable model of risk factors for GI bleeding (HR and 95% CI). Diabetes=self-report of diabetes mellitus or fasting glucose ≥126 mg/dL (≥7 mmol/L) or on treatment for diabetes. Hypertension=systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or on treatment for high blood pressure. Chronic kidney disease (CKD)=estimated glomerular filtration rate < 60 mL/min/1.73 m² or albumin-to-creatinine ratio ≥3 mg/mmol. NSAID, non-steroidal anti-inflammatory drugs; PPI, proton pump inhibitor; SSRI, selective serotonin reuptake inhibitor.

Figure 6

Absolute bleeding risk over 5 years for GI bleeding according to age and risk factors (shading represents 95% CI around the estimates). Red line=no risk factors, on placebo, green line=no risk factors but on aspirin, blue line=all risk factors and on placebo, purple line=aspirin and all risk factors; modelled risk factors were those significant in multivariable analyses.