Outcomes | Number of participants (studies), follow-up | Certainty of the evidence (GRADE) | Relative effect (95% CI) | Anticipated absolute effects over 10 years | |
Risk with placebo | Risk difference with aspirin* | ||||
Total mortality Follow-up: range 3.8 to 10 years[1-4] | 161,660 (13 RCTs) | ⊕⊕⊕ MODERATE due to imprecision¶ | RR 0.97 (0.93 to 1.02) | 60-year-old personΔ | |
83 per 1000Δ | 2 fewer per 1000 (6 fewer to 2 more) | ||||
Myocardial infarction (MI) Nonfatal events Follow-up: range 3.8 to 10 years[1,2,4] | 142,566 (12 RCTs) | ⊕⊕⊕⊕ HIGH | RR 0.83 (0.76 to 0.90) | Low cardiovascular risk population◊ | |
27 per 1000§ | 5 fewer per 1000 (6 fewer to 3 fewer) | ||||
Moderate cardiovascular risk population◊ | |||||
83 per 1000§ | 14 fewer per 1000 (20 fewer to 8 fewer) | ||||
High cardiovascular risk population◊ | |||||
136 per 1000§ | 23 fewer per 1000 (33 fewer to 14 fewer) | ||||
Stroke Includes nonfatal ischemic and hemorrhagic strokes Follow-up: range 3.8 to 10 years[1,2,4] | 127,433 (12 RCTs) | ⊕⊕⊕ MODERATE due to imprecision¶ | RR 0.95 (0.85 to 1.06) | Low cardiovascular risk population◊ | |
23 per 1000§ | 1 fewer per 1000 (3 fewer to 1 more) | ||||
Moderate cardiovascular risk population◊ | |||||
65 per 1000§ | 3 fewer per 1000 (10 fewer to 4 more) | ||||
High cardiovascular risk population◊ | |||||
108 per 1000 | 5 fewer per 1000 (16 fewer to 6 more) | ||||
Major extracranial bleed¥ Follow-up: range 3.8 to 10 years[1,2,4-6] | 155,911 (11 RCTs) | ⊕⊕⊕⊕ HIGH | RR 1.46 (1.32 to 1.62) | Low cardiovascular risk population‡ | |
8 per 1000§ | 4 more per 1000 (3 more to 5 more) | ||||
Moderate cardiovascular risk population‡ | |||||
24 per 1000§ | 11 more per 1000 (8 more to 15 more) | ||||
High cardiovascular risk population‡ | |||||
40 per 1000§ | 18 more per 1000 (13 more to 25 more) | ||||
Colorectal cancer (incidence) Follow-up: median 18.3 years[7] | 14,033 (4 RCTs) | ⊕⊕ LOW due to imprecision† and risk of bias** | HR 0.76 (0.60 to 0.96) | Low colorectal cancer risk population: Anticipated absolute effect over 20 years¶¶ | |
30 per 1000ΔΔ | 7 fewer per 1000 (12 fewer to 1 fewer) | ||||
Moderate colorectal cancer risk population | |||||
53 per 1000ΔΔ | 12 fewer per 1000 (21 fewer to 2 fewer) | ||||
High colorectal cancer risk population | |||||
100 per 1000ΔΔ | 23 fewer per 1000 (39 fewer to 4 fewer) |
CI: confidence interval; HR: hazard ratio; NCI: National Cancer Institute; RCT: randomized controlled trial; RR: risk ratio.
* The risk difference in the aspirin group (and its 95% CI) is based on the estimated risk in the comparison group and the relative effect of the intervention (and its 95% CI).
¶ The 95% CI for the absolute effect includes no benefit of aspirin. We did not rate down for risk of bias, but this was a borderline decision. Three of the trials did not blind patients, caregivers, or outcome adjudicator. Sensitivity analyses in meta-analysis by Raju et al[8] did not show evidence of risk of bias.
Δ Control group risk estimate for 10-year mortality applies to a 60-year-old person (male or female) and comes from population-based data from Statistics Norway. Mortality increases with age (eg, 50-year-old male; 40 deaths per 1000 in 10 years) and is lower in females than in males (eg, 2.5% in females aged 50 years versus 4% in males aged 50 years).
◊ Risk groups correspond to low (5%), medium (15%), and high risk (25%) according to the Framingham score (or other risk tool) to estimate 10-year risk.
§ Control group risk estimates in low, moderate, and high cardiovascular risk groups are based on the Framingham score. We have used data from an individual patient data meta-analysis to provide estimated risks for patient-important outcomes not covered by the Framingham risk score. We have also adjusted for 20% overestimation associated with Framingham risk score.
¥ Major extracranial bleeds are usually from the gastrointestinal tract and are most often defined in those requiring transfusion or resulting in death.
‡ In the individual patient data meta-analysis, risk for future major bleeding correlated with risk for future cardiovascular events. Therefore, we make the assumption that a patient at low, medium, or high risk of future cardiovascular events (determined by Framingham score) will be at low, medium, or high risk of future major bleeding events, respectively.
† The 95% CI for absolute effect borders no benefit of aspirin.
** Treatment with aspirin during the included studies ranged from 2.6 to 6.9 years. Colorectal cancer incidence was determined using cancer and death registries for a median of 18.3 years without knowledge of post-treatment period aspirin use.
¶¶ Control group risk estimates based on NCI Colorectal Cancer Risk Predictor Tool.
ΔΔ Moderate control group risk estimate derived from meta-analysis by Rothwell et al.[7]Adapted from: Vandvik PO, Lincoff AM, Gore JM, et al. Primary and Secondary Prevention of Cardiovascular Disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141:e637S.
Do you want to add Medilib to your home screen?