Inglés (pdf)
Articulo en XML
Referencias del artículo
Traducción automática
Enviar articulo por email
Citado por SciELO 
Similares en
SciELO 
uBio 
Permalink
INTRODUCTION
More than 30 years have passed since the Physician´s Health Study was published. This painstaking work demonstrated a 44% risk reduction of myocardial infarction (MI) with aspirin (RR: 0.56; 95 %CI, 0.45 to 0.70; p<0.0001) 1. This effect was more pronounced in the group of individuals older than 50 years, while the presence of ulcer and transfusion demand as secondary events were not significant compared with the placebo group (RR: 1.22; 95%CI,0.98 to 1.53; p = 0.08 for ulcer). The conclusion of this work was a recommendation for the use of aspirin in the primary prevention of a first MI in healthy individuals. However, the US Food and Drug Administration (FDA) did not approve the professional labeling of aspirin for the prevention of MI because another similar trial, The British Doctor´s Trial, did not show benefit from aspirin administration for cardiovascular prevention 2. In 2003, the use of aspirin was updated in the primary prevention of cardiovascular disease 3. In addition to the two mentioned trials, three more trials were also analyzed: The Thrombosis Prevention Trial 4, The Hypertension Optimal Treatment Study 5, and The Primary Prevention Project 6, including 55.580 randomized participants (11.466 women). The studies mentioned above revealed a 32% reduction in the risk of a first MI and a 15% reduction in the risk of all important vascular events following aspirin´s treatment, demonstrating strong evidence of the use of aspirin in the primary prevention of MI. At that time, the US Preventive Services Task Force (USPTF) 7 and the American Heart Association recommended aspirin for men and women whose 10-year risk of a first coronary event was 10% or greater 8 since the benefits of a reduction of cardiovascular (CV) events outweighed the risks in most of the patients presenting this sort of cardiovascular risk.
Additionally, there are gender-specific differences in platelet function and response to aspirin 9,10. Women under 65 years old without known CVD have a minor response to aspirin therapy in the primary prevention of coronary artery disease 11; the dose recommended was 81-100 mg every other day 12.
The net benefits for persons who have started taking aspirin continue accumulating over time without a bleeding event. The net benefits, however, generally become progressively smaller with advancing age because of an increased risk for bleeding, and modeling data suggest that it may be reasonable to consider stopping aspirin use at around age 75 13 for primary prevention. Despite this, there is a gap in understanding the benefits and risks of giving aspirin to patients at moderate risk of CVD. Aspirin has been a primary preventive drug for cardiovascular and cerebrovascular diseases for years. What has happened recently to change the concept and the prescription for the use of aspirin in the primary prevention of CVD? Today, the use of aspirin for primary prevention has been a subject of debate. Based on well-conducted studies, organizations such as the European Society of Cardiology, European Association for Cardiovascular Prevention & Rehabilitation, and the USPSTF, have delivered an almost uniform verdict that substantially changed the aspirin prescription for primary CV prevention 13,14.
The Number Needed to Treat (NNT), calculated as the reciprocal of the absolute risk reduction percentage, is a concise, clinically useful parameter that provides quantitative information on the efficacy of therapeutic interventions. Moreover, NNT allows clinicians to understand how much effort is needed to prevent a given event. The NNT and its opposite, the Number Needed to Harm (NNH), can be helpful in medical decision-making, then the use of NNT or NNH could be the likelihood of obtaining a benefit or harm 15. This systematic review aims to study the effect of aspirin in the primary prevention of CVD, under the scope of fundamental trials that have been an essential guide in the prescription or not of aspirin throughout decades till nowadays. Based on this premise, we believe that using the NNT and NNH could guide physicians in deciding whether aspirin could be prescribed to prevent primary CV events.
METHODOLOGY
Search Strategy
The present review exclusively focuses on aspirin as a preventive drug for primary CVD treatment. The search was divided into two periods to differentiate the times when the aspirin prescription was relevant (from 1988 to 1999, Table 1a) to the one where aspirin was questioned for primary prevention (from 2000 to 2018, Table 1b). The later period has been considered the modern era of cardiovascular primary prevention 16.
Table 1a Features of randomized controlled studies on the aspirin primary prevention in cardiovascular diseases (1989-1999).
| Study | Number of individuals/follow-up (years) | Intervention design | Primary events | Secondary events | Outcomes in primary events | Outcomes in secondary events: |
|---|---|---|---|---|---|---|
| The British Doctor´s Trial (1988)2 | 5139 (3429 vs 1710)/6 | Aspirin 500mg/d/ vs no aspirin | NFMI: | Not specified | NFMI: 1.03 (0.71-1.49), p=0,96. Stroke: 1.377 (0.72-2.60), p=0.40 | Not specified |
| Physician´s health study (PHS)(1989)1 | 22071 (11037 vs 11034)/5 | Aspirin 325 mg/betacarotene every other day vs placebo | NFMI, nonfatal stroke, and death for CVD | Ulcer, hemorrhage of any cause, and transfusion | For non-fatal MI: 0.56 (CI:0.45-0.70), p<0,0001. For Stroke: 1.22 (0.93-1.60), p=0.15. For CV Mortality: 0.96 (0.60-1.54), p=0.87 | Hemorrhage of any cause: 2979 (27%) in aspirin group vs 2248 (20,3%) RR=1.32 (1.25-1.40). p<0,0001 |
| The Thrombosis Prevention Trial (TPT) (1998)4 | 5085 (2545 vs 2540)/6 | Aspirin 75 mg/d vs placebo | Ischemic Heart Disease: Coronary death, fatal and NFMI | Major, Intermediate and minor hemorrhage | For NFMI: 0.69 (0.53-0.89), p<0,0049. For fatal MI: 1.13 (0.78- 1.63), p= 0,57. For coronary death: 0.81 (0.66-0.99), p<0,048 | Hemorrhage of any cause: 1.24 (1.12-1.37), p<0.0001. Major bleeding: 1.52 (1.01-2.28), p<0.055. Intermediate: 2,00 (0.61-6.65),p=0.38 |
| The Hypertension Optimal Treatment (HOT) Study(1998)5 | 18790 (9399 vs 9391)/4 | Aspirin 75 mg/d vs placebo | Major CV events: fatal and NFMI, stroke, and other CV deaths | Major fatal, nonfatal hemorrhage, and minor hemorrhage | Major CV events, 0·85 (0·73-0·99), p<0.03. Fatal and NFMI: 0·64 (0·49-0·85) p<0,002. Stroke: 0·98 (0·78-1·24), p=0.88 CV death: 0·95 (0·75-1·20), p=0.65 | Any type of hemorrhage: 1.74 (1.44-2.09), p<0,0001. Major hemorrhage: 1.84 (1.38-2.46), p<0,0001 |
NFMI: Non-fatal Myocardial Infarction; CV: Cardiovascular; CVD: Cardiovascular disease.
Table 1b Features of randomized controlled studies on the aspirin primary prevention in cardiovascular diseases (2001-2018).
| Study | Number of individuals/follow-up (years) | Intervention design | Primary events | Secondary events | Outcomes in primary events | Outcomes in secondary events: |
|---|---|---|---|---|---|---|
| The Primary Prevention Project (PPP) (2001)6 | 4495 (2226 vs 2269)/3.6 | Aspirin 100 mg/d/vitamin E vs no aspirin | CV Death, NFMI, and non-fatal stroke. | Major fatal /non-fatal Hemorrhage | CV deaths: 0·56 (0·31-1.00), p=0.06. All MI: 0·69 (0·38-1·23), p=0,27. Stroke: 0·67 (0·36-1·27), p=0,29 | Any type of hemorrhage: 4.07 (1.67-9.96), p<0,0015 Major hemorrhage: 1.74 (1.32-2.30), p<0,0001 |
| The Primary Prevention of Cardiovascular Disease in Women (WHS) (2005)19 | 39876 (19934 vs 19942)/10 | Aspirin 100mg every other day vs placebo Patients; >45 year old | NFMI, non-fatal stroke, or death from cardiovascular causes | GI bleeding Requiring transfusion | Major CV events: (RR, 0.91; (0.80 to 1.03); p=0.13. Stroke: RR;0.83; (0.69-0.99); p=0.04). Fatal or NFMI: RR: 1.02; (0.84-1.25); p=0.83, or death from cardiovascular causes RR 0.95; (0.74-1.22); p=0.68) | Major bleeding: RR, 1.40; (1.07- 1.83); p=0.02 |
| The Japanese Primary Prevention Project (JPPP) (2014)20 | 14 464 (7220 vs 7244)/6.5 | Aspirin 100mg daily vs no aspirin | Composite primary outcome was death from CV causes (MI, stroke, and other CV causes), non-fatal stroke (ischemic or hemorrhagic, including undefined cerebrovascular events), and NFMI | Secondary Outcomes included in divide endpoints. | Composite outcome: HR:0.94 (0.77-1.15); p= 0.54. NFMI: HR:0.53 (0.31-0.91), p= 0.02. TIA:HR: 0.57 (0.32-0.99);p=0.04 | Extracranial (major) hemorrhage: HR:1.85 (1.22-2.81); p =0.004). |
| The ASPREE Trial (2018)21 | 19114 (9525 vs 9589)/4.7 | Aspirin 100mg daily vs placebo | Endpoints included major hemorrhage and cardiovascular disease (fatal coronary heart disease, NFMI, fatal or non-fatal stroke, or hospitalization for heart failure). | Major Hemorrhage as secondary endpoint | For CVD: HR: 0.95; (0.83-1.08), p=NS. Fatal, NFMI: HR:0.93 (0.76-1.15). Stroke: HR: 0.89 (0.71-1.11) | Major Hemorrhage: HR: 1.38; (1.18-1.62); p<0.001 |
| The ARRIVE Trial (2018)22 | 12546 (6270 vs 6276)/5 years | Aspirin vs Placebo Eligible patients were aged 55 years (men) or 60 years (women) and older and had an average cardiovascular risk | The primary efficacy endpoint was a composite outcome of cardiovascular death, MI, unstable angina, stroke, or TIA. | Safety Endpoints were hemorrhagic events and incidence of other adverse events | HR:0·96; (0·81-1·13), p=0·6038).; NFMI: HR 0·90, (0·67-1·20); p=0·4562 | |
| The ASCEND Trial (2018)23 | 15,480 (7740 vs 7740)/7.4 | Aspirin 100mg vs placebo in diabetic patients | The primary efficacy outcome (MI., stroke or TIA, or death from any vascular cause, excluding any confirmed intracranial hemorrhage) | The primary Safety outcome was the first major bleeding event | Serious vascular events: 0.88; (0.79-0.97), p=0.01. NFMI: 0.98 (0.80-1.19). Stroke: 0.88 (0.73-1.06) | Major bleeding events:1.29 (1.09-1.52); p=0.003 |
NFMI: Non-Fatal Myocardial Infarction; CV: Cardiovascular; TIA: Transient ischemic attack; HR: Hazard ratio; RR: Relative risk.
The bibliography search was conducted in PUBMED by MEDLINE and Google Scholar under the following MESH (Medical Subject Headings) terminology: aspirin in primary prevention, aspirin in myocardial infarction, aspirin in stroke, aspirin in cardiovascular death or mortality, aspirin in bleeding or hemorrhage; those terms were connected thru a Boolean “and” with randomized controlled trials (Fig. 1). Additionally, the term number needed to treat was used for the complementary bibliography. The primary endpoint to report was a composite of non-fatal myocardial infarction (NFMI), non-fatal ischemic stroke (NFIS), and cardiovascular mortality (CVM). The primary bleeding outcome was major bleeding, as stated by the studies. The exclusion criteria were: a) studies with less than 4000 participants, b) systematic reviews on aspirin treatment because there are good reviews about it 16-18, c) a combination of antiplatelet or anticoagulants treatments with aspirin (The Thrombosis Prevention Trial assessed warfarin and aspirin alone and in combination but data for participants who received warfarin were excluded from the analysis), d) duplicate publications and e) those works that do not contain the composite as the endpoint.
Fig. 1 Flow chart of literature search strategy according to the MESH terminology. Note: some data might be lost in the early stages of the search.
Finally, four studies from 1988-1998 1,2,4,5 and six studies from 2001-2018 6,19-23 were included in the analysis (Tables 1a and 1b). This article has been assessed according to the PRISMA 2020 statement and checklist 24. The bias risk in each study was assumed according to a systematic review published previously 16, following the Cochrane risk of bias assessment, and the Jadad scale was used to evaluate the quality of the randomized controlled studies (< 3: high risk of bias, ≥ 3: low risk of bias) 25. Ethical approval was not required for conducting this study.
Statistical approach
NNT and NNH
NNT or NNH would be the number of patients to be treated to obtain one benefit or one harm in a predefined period 26. The number needed to treat is simply the reciprocal of the absolute risk difference obtained from the percentage of events in the control group minus the percentage of events in the experimental group (also named as absolute difference). Depending on the treatment, when the difference in the two groups proportions is significant, the NNT is small, and vice versa. NNT is a concise, clinically helpful presentation of the effect of an intervention. The NNT and the NNH are calculated as the inverse of the absolute reduction (ARR) or absolute increment of the risk (ARI). A 95% CI for NNT can be constructed by simply inverting and exchanging the limits of a 95% CI from the ARR. When the result shows an ARR or ARI with 95% CI extended from negative to positive values, it means that the zero is included, and the NNT is infinity thus, we need to separate two intervals using via infinity (∞) and indicate that the treatment may be helpful or harmful 27.
As Altman has mentioned 27, the terms NNT and NNH may not be appropriate to denote benefit and harm. He proposes the abbreviations of NNTB (benefit) and NNTH (harm). However, we maintain the conventional abbreviations of NNH and NNT to refer to benefit or harm, respectively. We also constructed an arbitrary classification of the NNT or NNH effect as follow: Net benefit, Uncertain benefit (treatment could be harmful), Uncertain harm (treatment could be helpful), and net harm.
Because the included studies have different follow-up periods, we have to make assumptions about it and make a “time adjustment” because if we want to be able to compare these NNTs, it is necessary to adjust all of them to refer to the same tracking time. So it is necessary to uniform the time of these studies to obtain the same relative interpretation of the results regarding benefit and harm. For this purpose, we have to use the following formula:28,29
Rearranging the terms, we have;
In the present study, we have adjusted five years as a hypothetical time for all the included studies to calculate NNT and NNH. We used a computer program for ARR, NNT, and NNH to obtain the data and their 95% CI (https://www.graphpad.com/quickcalcs/NNT1/). Additionally, as an effect measure, a Relative Risk and its 95% CI were estimated using the Comprehensive Meta-Analysis program (Biostat, Englewood, NJ) alongside the relative weight and random effect. As statistical parameters, the consistence for heterogeneity (I2) was determined as low (<25%), moderate (25% to 75%), and high (>75%) by testing the chi-square calculation of each meta-analysis (Cochrane Q) according to the Higgins formula 30; and statistical significance was fixed as p<0.05.
The forest plot for meta-analysis of each primary effective endpoint and safety was constructed under NNT and NNH parameters, with all included studies performing a logarithmic scale with infinity value in the middle of the scale according to previous reference 30, the NNT 95%CI (benefit) values are shown to the left and NNH 95% CI (harm) values on the right with the overall estimate.
RESULTS
Table 2a shows the total results of the endpoints in the trials made in the last year of the 20th century (four trials with a total of 51,085 participants), with a net benefit for NFMI (NNT= 156) confirmed by a significant relative risk reduction of 31% [RR,95%CI: 0.69(0.61-0.79); p<0,0001; I2= 10.5%].
Table 2a Summarized results of the endpoints in the included studies, period 1988-1998 (n= 4).
| ENDPOINTS | Aspirin events (%) n: 26410 | No Aspirin events (%) n:24675 | RR(95%CI), p value; I2 | ARR% (95%CI) | ARI% (95%CI) | NNT (95%CI) | NNH (95%CI)* | Observations |
|---|---|---|---|---|---|---|---|---|
| NFMI | 385 (1.46) | 518 (2.10) | 0.69 (0.61-0.79) <0.0001; 10.5% | 0.64 (0.41-0.87) | 156 (115 - 244) | Net benefit of aspirin treatment | ||
| NFIS | 338 (1.28) | 300 (1.22) | 1.02 (0.83-1.25) 0.87; 11.5% | 0.06 (-0.13 to 0.26) | 1667 (770 to ∞ to 390) | Uncertain harm (aspirin could be helpful) | ||
| CV Mortality | 463 (1.75) | 383 (1.55) | 1.12 (0.99-1.30) 0.08; 0% | 0.20 (-0.02 to 0.42) | 500 (4986 to ∞ to 237) | Uncertain harm (aspirin could be helpful) | ||
| Composite Outcome | 1186 (4.49) | 1201 (4.87) | 0.92 (0.85-0.99) 0.046; 0% | 0.38 (0.01-0.74) | 266 (135 -10158) | Net benefit of aspirin treatment | ||
| Major Bleeding | 205 (0.78) | 109 (0.44) | 1.79 (1.42-2.26) <0.0001; 0% | 0.33 (0.20-0.47) | 299 (213 - 500) | Net harm of aspirin treatment |
The composite outcome shows an NNT of 266 and an 8% relative risk reduction [RR,95%CI: 0.92(0.85-0.99); p=0.046; I2= 0%]. The major bleeding revealed a 79% increase in the relative risk [RR,95%CI; 1.79(1.42-2.26); p<0.0001; I2= 0%, NNH=299;].
Table 2b points out the same elements of the previous table with 105,975 participants and six trials made in two decades of the 21st century. The result of the studies carried out showed a benefit with the treatment for NFIS (NNT= 553) with a 12% in relative risk reduction [RR,95%CI: 0.88 (0.80-0.98); p<0.01; I2= 0%].
Table 2b Summarized results of the endpoints in the included studies, period 2001-2018 (n= 6).
| ENDPOINTS | Aspirin events (%) n: 52915 | No Aspirin events (%) n: 53060 | RR (95%CI) p value; I2 | ARR% (95%CI) r | ARI% (95%CI) | NNT (95%CI) | NNH (95%CI) | Observations |
|---|---|---|---|---|---|---|---|---|
| NFMI | 669 (1.26) | 718 (1.35) | 0.93 (0.82-1.05) 0.21; 9.6% | 0.09 (-0.05 to 0.23) | 1112 (435 to ∞ to2086) | Uncertain benefit (aspirin could be harmful) | ||
| NFIS | 721 (1.36) | 819 (1.54) | 0.88 (0.80-0.98) <0.01; 0% | 0.18 (0.04 -0.33) | 553 (303- 2500) | Net benefit of aspirin treatment | ||
| CV Mortality | 508 (0.96) | 551 (1.04) | 0.92 (0.82-1.04) 0.20; 0% | 0.08 (-0.04 to 0.20) | 1250 (505 to ∞ to 2419) | Uncertain benefit (aspirin could be harmful) | ||
| Composite Outcome | 1898 (3.59) | 2088 (3.94) | 0.91 (0.86- 0.97) 0.003; 0% | 0.35 (0.12 - 0.58) | 288 (174- 839) | Net benefit of aspirin treatment | ||
| Major Bleeding | 964 (1.82) | 662 (1.25) | 1.57 (1.30-1.91) <0.0001; 0% | 0.57 (0.43-0.72) | 175 (139- 233) | Net harm of aspirin treatment |
The composite outcome presented an NNT of 288 with a significant relative risk reduction of 9% [RR,95%CI: 0.91 (0.86-0.97); p<0.003; I2=0%], and for major bleeding, there was a 57% increase of the relative risk [RR,95%CI: 1.57 (1.30-1.91); p<0.0001; I2= 0%, NNH= 175;].
The total of the results in the combined and separate primary endpoints of all studies are shown in Table 2c. From a total of 157,060 participants, there was a net benefit of 20% and 10% on the relative risk reduction for NFMI [RR= 0.80 (0.69 to 0.93); p=0.003; I2= 0%; NNT= 259] and Composite outcome [RR= 0.90 (0.85 to 0.99); p<0.001; I2= 0%; NNT= 292] respectively. Major bleeding presents a 60% increase in the relative risk [RR: 1.60 (1.38 to 1.85); p<0.0001; I2= 3%; NNH= 208]. The rest of the endpoints showed an uncertain result because the aspirin treatment could be harmful or helpful compared with the control.
Table 2c Summarized results of the endpoints in all included studies (n=10).
| ENDPOINTS | Aspirin events (%) n: 79325 n: | No aspirin events% n:77735 | RR (95%CI) p value; I2 | ARR% (95%CI) | ARI% (95%CI) | NNT (95%CI) | NNH (95%CI) | Observations |
|---|---|---|---|---|---|---|---|---|
| NFMI | 954 (1.20) | 1236 (1.59) | 0.80 (0.69-0.93) 0.003; 0% | 0.39 (0.27-0.50) | 259 (199-369) | Net benefit of aspirin treatment | ||
| NFIS | 1059 (1.34) | 1119 (1.44) | 0.93 (0.83-1.01) 0.09; 0% | 0.10 (-0.01 to 0.22) | 958 (454 to ∞ to 8910) | Uncertain benefit (aspirin could be harmful) | ||
| CV. Mortality | 971 (1.22) | 934 (1.20) | 0.96 (0.88-1.05) 0.39; 0% | 0.02 (-0.09 to 0.13) | 4433 (1167 to ∞ to 765) | Uncertain harm (aspirin could be helpful) | ||
| Composite Outcome | 3084 (3.89) | 3289 (4.23) | 0.90 (0.85-0.99) <0.001; 0% | 0.34 (0.15-0.54) | 292 (186-676) | Net benefit of aspirin treatment | ||
| Major bleeding | 1169 (1.47) | 771(1.00) | 1.60 (1.38-1.85) <0.0001;3% | 0.48 (0.37-0.59) | 208 (269-170) | Net harm of aspirin treatment |
NFMI: Non-fatal Myocardial Infarction, NFIS: Non-fatal Ischemic Stroke, CV: Cardiovascular, RR: Relative Risk, ARR: Absolute Risk Reduction, ARI: Absolute Risk Increase, NNT: Number Needed to Treat, NNH: Number Needed to Harm, CI: Confidence Interval. Note: In the present study, each individual study was adjusted to 5 years a hypothetical time to calculate NNT and NNH: NNT(hypothetic) = NNT(observed) X [Time(hypothetic)/Time (observed)].
When the result shows an ARR with 95%CI positive values means a net benefit (NNT); when the result shows an ARI with 95%CI means net harm (NNH). When the result shows an ARR or ARI with 95%CI extended from negative to positive values means that the zero is included thus we need to separate two intervals using via infinity (∞) and indicates that the treatment may be helpful or harmful 27.
Forest plots of the meta-analysis of aspirin treatment in the total included studies are shown in terms of NNT (benefit) or NNH (harm) for each primary endpoint.
The overall estimate points out that concerning aspirin treatment revealed a net benefit for NFMI (Fig. 2a), an uncertain benefit for NFIS (Fig. 2b), and uncertain harm for CV mortality (Fig. 2c). As we expected, net harm was associated with major bleeding (Fig. 2d)
Fig. 2 Forest plot of meta-analysis of the included studies calculating from ARR95%CI as effect measure the NNT or NNH with their respective 95%CI. Non-fatal myocardial infarction (A), Non-fatal ischemic stroke (B), Cardiovascular mortality (C), and Major bleeding (D) as endpoints. *Denotes Absolute Risk Increase (ARI). When the ARR95%CI includes zero we need to separate two intervals using via infinity (∞) in the middle of the scale 27.
DISCUSSION
The light of the results of this systematic review, we argued that aspirin has an ambiguous place in the prescription for primary CV prevention. NFMI from studies of the 20th century and NFIS in this century showed a net benefit with the aspirin treatment. In some instances, the treatment could be harmful, given negative ARR 95% CI values. The composite outcome results also reported benefits in the two groups of studies. Globally there was a net benefit of aspirin treatment for NFMI and the composite outcome but also a significant bleeding risk. This study demonstrates that the absolute risk reduction for cardiovascular events and absolute risk increase for major bleeding associated with aspirin use were of similar magnitude. In terms of NNT and NNH, this represents notable data on the aspirin prescription despite some evidence that indicates the efficacy of aspirin could be uncertain in the NNT/NNH ratio since all of the studies and the combined results have shown a relevant bleeding risk. These findings have similarities with a recent meta-analysis 31. A systematic review 32 revealed an ARR of 0.41% in the composite cardiovascular outcome with an NNT of 241 and an ARI of 0.47% for major bleeding risk, representing an NNH of 210. This confirmed a possible adverse effect of aspirin due to bleeding risk. Another systematic review 33 showed a reduction of 10% of MCE (major cardiovascular events) (0.90, 95% CI 0.85-0.96, p<0.001) with an ARR of 0.39% (95%CI: 0.18-0.61), which correspond to NNT of 253 (95%CI: 163-568) to prevent one single MCE, and the NNH (major bleeding) was 306. These results are similar to our work, indicating potential harm in aspirin use due to increased bleeding risk. Abdelaziz et al. 34 showed in an illustration the NNT for MI (357), ischemic stroke (500), TIA (370), and MACE (263) with the NNH for major bleeding (222), hemorrhagic stroke (1000), and GI bleeding (385), based on pooled data from 15 randomized controlled trials. Therefore, it was deduced that bleeding risk is present when calculating the NNTs over NNH with major bleeding. These findings suggest that the decision to use aspirin for primary prevention should be tailored to the individual patient based on the estimated CV risk. Another approach to evaluate aspirin treatment in the primary prevention of CV events is to compare benefits (prevention of MI and stroke) and harms (major GI bleeds and hemorrhagic stroke) using relative weights in the assessment of systematic reviews of the NNT and NNH as the number of person-years with treatment need to prevent one adverse event 35. This approach demonstrated a net benefit for aspirin; nevertheless, in the sensitivity analysis, aspirin was harmful due to greater relative weight for GI bleeds.
We tried to give a better scope of aspirin as a primary preventive treatment for CV events, but unfortunately, the controversy persists. To treat or not to treat with aspirin is the question, and this situation is under debate. To better understand this complex scenario, we follow the timing of the declarations of the USPSTF about the statements on aspirin prevention in CVD 7,36-38.
In this contemporary period, aspirin passed from being a good prescription for primary prevention of CVD at any age to a restriction for its use only in 40-59 years old individuals, with a strong recommendation against its use in older people. The last decision came from another six studies, whose results regarding the therapy with aspirin in primary prevention for CVD were unfavorable for its recommendation (Table 1b). Additionally, in an analysis of 17 RCT (164.862 participants) 39, aspirin did not show any significant reduction in all-cause mortality compared with placebo (RR:0.97;95%CI:0.93-1.01; p=0.13). However, when ≥ 65 years old patients were excluded, it significantly reduced all-cause mortality (RR 0.94; 95% CI 0.90-0.99; p = 0.01) in the aspirin group. These results concord with the age arguments recently expressed by the USPSTF. The European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice colleagues and other systematic reviews share the same opinion since its guidelines do not recommend aspirin for the primary prevention of CVD at all 14,40. However, those agreements can lead to misinterpretation or confusion for patients in whom aspirin therapy may be essential: such as those on primary or secondary prevention with an established CV risk of more than 10% 41.
The only explanation for this change is attributed to the bleeding risk that increases with age. However, there is a debate about this new scenario, and doubts must be cleared. For example, should they stop taking aspirin for people who have been using aspirin for years and do not have evidence of bleeding? If yes, what could we probably expect? So we could expect an increase in CV events in this particular group unless there is an indication of another alternative surrogate for aspirin prescription.
A comment arises regarding the different conduct over time on aspirin prescription, during the last decades of the 20th century and then the change of opinion in the two decades of the 21st century. The first studies on aspirin and prevention of cardiovascular risk focused mainly on the significant relative reduction of the risk of myocardial infarction (44% PHS, 31% TPT, and 36% HOT, only the BDT showed a non-significant 3%). Perhaps this finding eluded the attention to the side effects of bleeding caused by the administration of aspirin, which was barely mentioned in those papers, but was not given its due importance, despite the relevant relative increase in the risk of major bleeding (71% PHS, 52% TPT, and 84% HOT). On the contrary, the trials that were performed in the two decades of the 21st century addressed with great interest the risk of bleeding with aspirin prescription in the prevention of CVD. For this reason, these papers highlighted the relevance of major hemorrhage as a contraindication to aspirin intake. They concluded that serious but no fatal bleeding 42 is frequent with aspirin administration compared with the benefit achieved in CV prevention. As a representative example, the myocardial infarction in terms of RRR (Relative Risk Reduction for efficacy) vs. major bleeding as RRI (Relative Risk Increase for safety) is shown as follows: PPP (31% vs. 74%), WHS (2% vs. 40%), JPPP (47% vs. 85%), ASPREE (7% vs. 38%), ASCEND (2% vs. 29%) and the ARRIVE (10% vs 110%), all of them favoring the bleeding risk. The different points of view above about aspirin treatment changed the opinion of doctors and their patients regarding the routine use of aspirin as a primary preventive drug in CV events.
Another point of view of this conflictive situation is that the prescription of aspirin is unnecessary in some instances because it is an over-a-counter (OTC) drug that can be freely purchased, and patients have been buying the drug without considering the potential bleeding risk 41.
Although there are relevant evidence and guidelines as instruments for shared decision-making to help clinicians in the use of aspirin in primary prevention 43-46 and the search for a benefit versus risk prediction tool, we propose the NNT and NNH as valuable measures in the balancing benefit-risk with aspirin in the therapy of the effective primary prevention of CV events. Knowing or estimating NNT and NNH for an individual patient’s risk could be a guide for the overall or net value of a prophylactic intervention 47. This combination provides a good measure of the effort in avoiding an unfavorable outcome, weighed against possible associated risks. The calculation of these measures is straightforward, and also its interpretation so that physicians could make an individual clinical decision based on the results of the interventions for CV primary prevention guided by the calculation of how many patients can be treated to avoid one adverse event, counterbalancing with the collateral side effects of the aspirin prescription. However, although the calculation of the NNT is simple, we need to consider the treatment time to ensure its correct interpretation 48. An essential limitation of NNT and NNH is that these metrics are limited to dichotomous (rather than continuous) outcomes 16.
Limitations
The present study has several limitations: first, the dose of aspirin was different in the studies, oscillating between 75mg and 500mg with six trials using 100mg daily. Second, there were some difficulties in classifying endpoints (MI and Stroke are sometimes not defined as non-fatal, and several studies did not specify major bleeding as a safety endpoint). On the other hand, one crucial point that is missing is the evolution of the definition of non-fatal MI: in the “modern era”, the use of troponin captures minor MIs which were previously missed by ECG and/or CPK only, and on the other hand, we did not include total mortality as an outcome, and this could probably be a significant cause of bias. Third, in the studies made in the 20th century, particularly bleeding events were poorly reported despite the apparent evidence. Fourth, the studies were not analyzed, separating participants from high and low cardiovascular risk. Fifth, The NNT and NNH were calculated in hypothetical results that were expressed as five years of tracking time as a standard unit for all studies and their calculations, thus the results of this review need to be interpreted with prudence. Sixth, the forest plots for meta-analysis for NNT and NNH were constructed with a scale including infinity (∞) and quoting to separate in two confidence intervals when there were negative values.
CONCLUSION
Recent studies have demonstrated that aspirin should not be recommended in the primary prevention of CVD, although it has a place in the secondary prevention of CVD 17,45,49. The use of aspirin in primary cardiovascular disease was associated with a lower risk of cardiovascular events and an increased risk of major bleeding. NNT as a measure of effect and NNH to determine harm could be helpful in clinical share decision-making. However, as the “two faces of the coin” (not determined by a coin flip), it is essential to establish consistently that the benefit-risk for patients should be individualized and not be a clinical practice guide for everyone.
