Initial aspirin studies did not show a statistically significant increase in bleeding, likely due to too few events and inadequate powering. Subsequent meta-analyses from multiple evaluations have consistently shown bleeding to be a risk.3,7 The risk for bleeding with aspirin has also been examined in multiple cohort studies, which has helped elucidate the risk in greater detail.
Epidemiologic data show that among patients who do not use nonsteroidal anti-inflammatory drugs (NSAIDs), the rate of upper gastrointestinal (GI) complications is 1 case per 1000 patient-years.20 Multiple studies have consistently shown that aspirin use increases the rate of significant upper GI bleeding over baseline risk (odds ratio [OR] = 1.54-1.58).3,21,22 Interestingly, these increases seem not to be influenced by other factors, such as comorbidities that increase the risk for ASCVD. Analysis of cancer prevention studies showed similar epidemiologic trends, with aspirin use exceeding a baseline bleeding risk of 0.7 cases of upper GI complications per 1000 patient-years (OR = 1.31-1.73).23
Other risk factors. Evaluation of risk factors for bleeding primarily comes from 2 studies.3,7 Most data concern the impact of individual factors on significant GI bleeding, with fewer data available for evaluating risk for intracerebral hemorrhage (ICH). Initial analysis of individual prospective studies showed little or no correlation between risk for bleeding and such factors as gender, age, or history of hypertension or ASCVD.21 Subsequent analysis of meta-data and large cohorts did show statistically significant impact on rates of bleeding across several factors (TABLE 13,7).
Of note is a large heterogeneous cohort study conducted in Spain. Data showed significant increases in baseline risk for GI bleeding in older men with a history of GI bleeding and NSAID use. The absolute risk for GI bleed in this group was potentially as high as 150 cases per 1000 patient-years, well above the risk level assumed for the average patient.24 A seemingly small OR of 1.5 could dramatically increase the absolute risk for bleeding in such patients, and it suggests that a generalized risk for bleeding probably shouldn’t be applied to all patients. Individuals may be better served by a baseline risk calculation reflecting multiple factors.
Due to the comparatively uncommon nature of ICH, fewer data are available to support definitive conclusions about its increased risk with aspirin use. Aspirin use appeared to increase the risk for ICH with ratios between 1.27 and 1.32 in meta-analyses (measured as an OR or as an RR),3,7,21 with an IRR of 1.54 in a cohort study.22 The only statistically significant factors suspected to increase the risk of ICH at baseline were smoking (RR = 2.18) and mean BP > 20 mm Hg over normal (OR = 2.18). Age, gender, and diabetes all showed a nonsignificant trend toward risk increase.7
Continue to: Risk based on dose and formulation