BP indicates blood pressure.
eAppendix 1. Search Documentation Details
eAppendix 2. Background, Methods, and Results of Systematic Review of Combination Drug Therapy to Evaluate for Potential Interaction of Effects
eAppendix 3. PRISMA Flow Charts for Each Drug Class and Detailed Systematic Review Characteristics and Summary of Included Systematic Reviews and Meta-analyses
eAppendix 4. List of Excluded Studies and Reasons for Exclusion
Kunal N. Karmali, Donald M. Lloyd-Jones, Mark A. Berendsen, David C. Goff, Darshak M. Sanghavi, Nina C. Brown, Liliya Korenovska, Mark D. Huffman. Drugs for Primary Prevention of Atherosclerotic Cardiovascular DiseaseAn Overview of Systematic Reviews. JAMA Cardiol. 2016;1(3):341–349. doi:10.1001/jamacardio.2016.0218
How effective and safe are aspirin, statins, blood pressure (BP)–lowering therapy, and tobacco cessation drugs for prevention of primary atherosclerotic cardiovascular disease (ASCVD)?
In this overview of systematic reviews, we show that aspirin, statins, and BP-lowering therapy reduce ASCVD risk by 10% to 25% and tobacco cessation drugs increase the odds of continued smoking abstinence by 88% to 188%. Adverse events were increased with aspirin, were not increased with statins, and were poorly reported among BP-lowering and tobacco cessation drug trials.
High-quality evidence supports use of these drugs for ASCVD primary prevention and tobacco cessation.
The Million Hearts initiative emphasizes ABCS (aspirin for high-risk patients, blood pressure [BP] control, cholesterol level management, and smoking cessation). Evidence of the effects of drugs used to achieve ABCS has not been synthesized comprehensively in the prevention of primary atherosclerotic cardiovascular disease (ASCVD).
To compare the efficacy and safety of aspirin, BP-lowering therapy, statins, and tobacco cessation drugs for fatal and nonfatal ASCVD outcomes in primary ASCVD prevention.
Structured search of the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects (DARE), Health Technology Assessment Database (HTA), MEDLINE, EMBASE, and PROSPERO International Prospective Systematic Review Trial Register to identify systematic reviews published from January 1, 2005, to June 17, 2015, that reported the effect of aspirin, BP-lowering therapy, statin, or tobacco cessation drugs on ASCVD events in individuals without prevalent ASCVD. Additional studies were identified by searching the reference lists of included systematic reviews, meta-analyses, and health technology assessment reports. Reviews were selected according to predefined criteria and appraised for methodologic quality using the Assessment of Multiple Systematic Reviews (AMSTAR) tool (range, 0-11). Studies were independently reviewed for key participant and intervention characteristics. Outcomes that were meta-analyzed in each included review were extracted. Qualitative synthesis was performed, and data were analyzed from July 2 to August 13, 2015.
From a total of 1967 reports, 35 systematic reviews of randomized clinical trials were identified, including 15 reviews of aspirin, 4 reviews of BP-lowering therapy, 12 reviews of statins, and 4 reviews of tobacco cessation drugs. Methodologic quality varied, but 30 reviews had AMSTAR ratings of 5 or higher. Compared with placebo, aspirin (relative risk [RR], 0.90; 95% CI, 0.85-0.96) and statins (RR, 0.75; 95% CI, 0.70-0.81) reduced the risk for ASCVD. Compared with placebo, BP-lowering therapy reduced the risk for coronary heart disease (RR, 0.84; 95% CI, 0.79-0.90) and stroke (RR, 0.64; 95% CI, 0.56-0.73). Tobacco cessation drugs increased the odds of continued abstinence at 6 months (odds ratio range, 1.82 [95% CI, 1.60-2.06] to 2.88 [95% CI, 2.40-3.47]), but the direct effects on ASCVD were poorly reported. Aspirin increased the risk for major bleeding (RR, 1.54; 95% CI, 1.30-1.82), and statins did not increase overall risk for adverse effects (RR, 1.00; 95% CI, 0.97-1.03). Adverse effects of BP-lowering therapy and tobacco cessation drugs were poorly reported.
Conclusions and Relevance
This overview demonstrates high-quality evidence to support aspirin, BP-lowering therapy, and statins for primary ASCVD prevention and tobacco cessation drugs for smoking cessation. Treatment effects of each drug can be used to enrich discussions between health care professionals and patients in primary ASCVD prevention.
In 2011, the US Department of Health and Human Services launched the Million Hearts initiative to prevent 1 million heart attacks and strokes during 5 years.1 Million Hearts includes a clinical emphasis on ABCS, that is, aspirin for high-risk patients, blood pressure (BP) control, cholesterol level management, and smoking cessation. Projections estimate that optimizing ABCS management could reduce the burden of atherosclerotic cardiovascular disease (ASCVD) substantially.2 To support the Million Hearts goal and to identify new models of care delivery and payment, the Center for Medicare & Medicaid Innovation launched the Million Hearts Cardiovascular Risk Reduction Model in 2016, a cluster-randomized payment model test to evaluate the effect of value-based payment to incentivize ASCVD risk assessment and reduction. This effort requires rigorous and transparent quantification of the treatment effects of aspirin, BP-lowering therapy, statins, and tobacco cessation drugs in primary ASCVD prevention.3 Despite the wealth of evidence-synthesis activities within individual drug classes, we are not aware of any reports that have summarized the efficacy and safety of all 4 drug classes highlighted in the Million Hearts initiative for primary ASCVD prevention.
An overview of systematic reviews is a novel approach to appraise and synthesize results from multiple systematic reviews into a single, useful document that can be used to guide health care professionals and policy makers.4- 7 To address this evidence gap, we performed an overview of systematic reviews to compare the efficacy and safety of aspirin, BP-lowering therapy, statins, and tobacco cessation drugs on fatal and nonfatal outcomes for primary ASCVD prevention.
This overview followed guidelines outlined by the Cochrane Collaboration to synthesize the effects of multiple interventions for an overarching clinical question using data from published systematic reviews.4 We established a protocol and published it in the PROSPERO International Prospective Register of Systematic Reviews.8 The clinical question guiding this overview is presented in the PICOTSS (patient, intervention, comparators, outcomes, timing, setting, and study design) format (Box). We also performed a supplemental systematic review to evaluate for potential interactions when combination drug therapy is used but found none (details about the search and results are outlined in eAppendices 1 and 2 in the Supplement).
Adults ≥18 years of age
People without prevalent ASCVD
Systematic reviews or trials that included people with Alzheimer disease, end-stage renal disease, macular degeneration, and aortic stenosis were excluded
Tobacco cessation drugs (ie, nicotine replacement therapy, varenicline tartrate, bupropion hydrochloride)
Fatal and nonfatal cardiovascular events, including myocardial infarction and stroke (ASCVD)
Adverse events as reported by authors of included studies
Fatal and nonfatal ischemic heart disease events, including myocardial infarction, angina, and coronary revascularization
Fatal and nonfatal cerebrovascular events, including stroke and transient ischemic attack
Total and nonfatal ASCVD events
Total and low-density lipoprotein cholesterol levels
Systolic and diastolic BP
Health-related quality of life
Studies of any duration
Systematic reviews of randomized and quasi-randomized clinical trials
Abbreviations: ASCVD, atherosclerotic cardiovascular disease; BP, blood pressure.
We included systematic reviews of randomized and quasi-randomized clinical trials comparing the pooled treatment effects of aspirin, BP-lowering therapy, statins, or tobacco cessation drugs against placebo or usual care in adults (≥18 years of age) without prevalent ASCVD. For systematic reviews that included a combination of individuals with and without prevalent ASCVD, we included reviews that reported effect estimates for participants defined as receiving primary prevention. When reports included primary and secondary prevention populations, we included only those reports with less than 10% of the population with prevalent ASCVD.9 Systematic reviews or trials in which drugs were used to treat or control chronic conditions (eg, Alzheimer disease, rheumatoid arthritis, renal disease, macular degeneration, and aortic stenosis) were excluded.
The primary outcomes for our review were (1) all-cause mortality; (2) fatal and nonfatal cardiovascular events, including myocardial infarction and stroke (ASCVD); and (3) adverse events as reported by the authors of included reviews. Secondary outcomes were (1) fatal and nonfatal ischemic heart disease events, including myocardial infarction, angina, and coronary revascularization; (2) fatal and nonfatal cerebrovascular events, including stroke and transient ischemic attack; (3) total nonfatal ASCVD events; (4) total and low-density lipoprotein cholesterol levels; (5) systolic and diastolic BP; (6) health-related quality of life using validated instruments; and (7) direct costs.
We searched the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects (DARE), Health Technology Assessment Database (HTA), MEDLINE, EMBASE, and PROSPERO International Prospective Systematic Review Trial Register from January 1, 2005, to June 17, 2015. We limited retrieval to English language systematic reviews. One of us (M.A.B.) who was an experienced information specialist performed all searches. Detailed search strategies with explanations of databases and search filters are included in eAppendix 1 in the Supplement.
We identified additional eligible studies by searching the reference lists of included systematic reviews, meta-analyses, and health technology assessment reports. We contacted study authors when necessary to identify further information that we may have missed.
Two of us (K.N.K. and M.D.H.) independently performed all tasks for study selection, data extraction, evidence synthesis, and quality assessment. Any discrepancies here and throughout were resolved through consensus or recourse to a third investigator (D.M.L.-J.).
We screened titles and abstracts and then full texts to identify relevant systematic reviews for inclusion. For studies that fulfilled the inclusion criteria, we independently abstracted key participant and intervention characteristics and reported data on prespecified outcomes using standardized data extraction templates. We also extracted pooled effect estimates for outcomes that were meta-analyzed in each included review. We reported dichotomous data as risk ratios (RRs) or odds ratios (ORs) with 95% CIs. We reported continuous data as mean differences with 95% CI.
Data were analyzed from July 2 to August 13, 2015. We independently assessed the methodologic quality of each systematic review using the Assessment of Multiple Systematic Reviews (AMSTAR) tool.10 Because many systematic reviews included information from overlapping trials, we did not perform a separate meta-analysis of pooled effect estimates. Instead, we performed a qualitative synthesis for each drug intervention and reported the treatment effect from the most comprehensive and highest-quality systematic reviews as recommended by the Cochrane Collaboration and as has been performed in other overviews.4,6,11,12
We used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to rate the overall quality of the evidence.13 We created an adapted Summary of Findings table based on the methods described in the Cochrane Handbook for Systematic Reviews of Interventions to convey key information about the best treatment effect for each drug intervention and overall confidence in this estimate.4
We reported results that evaluated potential differences in drug effects between men and women and people with and without type 2 diabetes mellitus when these results were presented by authors of reviewed studies. We were unable to quantify differences by race/ethnicity owing to limitations of data availability and reporting.
We revised our search strategy by using a precision-maximizing filter after our initial search with a sensitivity- and specificity-balancing filter produced unfeasibly large results. We also restricted our retrieval to English-language trials because of time limitations for project completion. Last, we elected to include systematic reviews of tobacco cessation drugs that included some trials involving participants with known vascular disease. We made this decision because reviews addressed smoking cessation as the primary efficacy outcome and provided limited data on cardiovascular outcomes or mortality.
We performed 4 independent searches for systematic reviews evaluating the effects of aspirin, BP-lowering therapy, statins, and tobacco cessation drugs in the primary prevention of ASCVD. We present a PRISMA study flow in aggregate (Figure) and by drug class (eAppendix 3 in the Supplement).
Our search identified 1967 records, of which 145 were evaluated as full-text articles after title and abstract screening. In total, we selected 35 systematic reviews, available as 37 reports for inclusion. These consisted of 15 systematic reviews evaluating the effect of aspirin,14- 29 4 systematic reviews of BP-lowering therapy,30- 33 12 systematic reviews of statins,34- 46 and 4 systematic reviews of tobacco cessation drugs in a primary prevention setting.47- 50 Reasons for exclusion of each full text report and references to the excluded studies are presented by drug class in eAppendix 4 in the Supplement.
A summary of the 15 systematic reviews of aspirin is provided in eAppendix 3 in the Supplement. Type 2 diabetes–specific treatment effects were provided in 6 reviews.18- 20,24,28,29 The latest search ran through 2013.17 The reviews included data from 6 to 9 primary prevention trials with 95 456 to 102 621 participants. Trials generally included adults aged 40 to 70 years, and the weighted mean (SD) age of participants in the most comprehensive report was 57 (4) years.25
Authors of included studies used a variety of criteria to judge trial quality, including risk for bias assessments, Jadad scores,51 Delphi process,52 and the US Preventive Services Task Force quality criteria. Trial quality was generally assessed as high, and risk of bias was categorized as low. Systematic review quality, as measured by AMSTAR rating, ranged from a score of 5 of 11 to 11 of 11. The most comprehensive and highest rated systematic review was an HTA by Sutcliffe et al.25
Treatment effects of aspirin on all-cause mortality, composite cardiovascular events, and individual component cardiovascular events were broadly comparable across all systematic reviews. The systematic review by Sutcliffe et al25 reported a 6% reduction in all-cause mortality (RR, 0.94; 95% CI, 0.88-1.00) and a 10% reduction in major cardiovascular events (RR, 0.90; 95% CI, 0.85-0.96). Treatment effects on other outcomes are provided in eAppendix 3 in the Supplement.
Systematic reviews that reported type 2 diabetes–specific treatment effects demonstrated similar relative benefits compared with the general primary prevention group; however, all of the upper 95% CIs included the possibility of no improvement. Sex-stratified analyses were reported in 4 systematic reviews,14,16,27,29 of which one was an individual participant data meta-analysis of 6 primary prevention aspirin trials.14 These analyses demonstrated similar reductions in the composite CVD outcome in men (RR, 0.88; 95% CI, 0.78-0.98) and women (RR, 0.88; 95% CI, 0.76-1.01). However, the effects were driven by a reduction in major coronary events in men (RR, 0.77; 95% CI, 0.67-0.89) and ischemic stroke in women (RR, 0.77; 95% CI, 0.59-0.99).
Reviews varied in their definition of clinically important bleeding, but all 15 reported increased risk for bleeding and other hemorrhagic complications with aspirin therapy. The analysis from the Antithrombotic Trialists’ Collaboration, which accounted for differences in person-years of follow-up and used a standardized definition for major bleeding in pooled trials (major gastrointestinal tract and extracranial bleeding that was fatal or required blood transfusion), reported a 54% increased risk for major bleeding with aspirin therapy (RR, 1.54; 95% CI, 1.30-1.82) and a 32% increased risk for hemorrhagic stroke (RR, 1.32; 95% CI, 1.00-1.75).14 Data on other bleeding risks are shown in eAppendix 3 in the Supplement. Data on health-related quality of life and direct costs were not reported in the identified systematic reviews.
We rated the quality of evidence for the effect of aspirin on all-cause mortality as moderate, which was downgraded because of imprecision of the treatment effect. We rated the quality of evidence for the effect of aspirin on reducing major cardiovascular events as high. We also rated the quality of evidence for the effect of aspirin on increasing major bleeding events as high.
We identified 4 systematic reviews of BP-lowering therapy in primary ASCVD prevention30- 33; 2 of these reviews30,33 reported the effects of BP reduction in individuals with mild or grade 1 hypertension (systolic BP, 140-149 mm Hg; diastolic BP, 90-99 mm Hg). Type 2 diabetes–specific treatment effects were not reported. The latest search ran through 2014 and included individual participant data from the Blood Pressure Lowering Treatment Trialists’ Collaboration.33 The most comprehensive systematic reviews included 25 trials with 163 131 participants31 and 27 trials with 108 297 participants.32 Only 8 trials overlapped between these 2 reviews owing to variations in search strategies, search time frames, inclusion criteria, and classification of primary prevention by authors of the included reviews. The mean age of participants ranged from 30 to 80 years, and the weighted mean age in the most comprehensive systematic review was 62 years.32 A summary of systematic review characteristics and detailed characteristics from the full data abstraction are provided in eAppendix 3 in the Supplement.
Authors of included reviews used a variety of criteria to judge trial quality, including risk of bias and GRADE assessments. These authors reported a range of quality assessments from very low to high quality for trials evaluating the effects of calcium channel blockers and diuretics. The AMSTAR ratings ranged from a score of 4 of 11 to 9 of 11.
Treatment effects of BP-lowering therapy on all-cause mortality, composite cardiovascular outcomes, and individual component cardiovascular outcomes were broadly comparable across systematic reviews. The systematic review by Law et al,32 the most comprehensive review included in our study, reported an 11% reduction in all-cause mortality (RR, 0.89; 95% CI, 0.84-0.95), a 16% reduction in coronary heart disease events (RR, 0.84; 95% CI, 0.79-0.90), and a 36% reduction in stroke (RR, 0.64; 95% CI, 0.56-0.73). Treatment effects standardized to a 10–mm Hg reduction in systolic BP and a 5–mm Hg reduction in diastolic BP were also reported (Table).
The effects of BP-lowering treatments in persons with mild or grade 1 hypertension were reported in 2 systematic reviews.31,33 The review by Sundström et al33 was an update to the review by Diao et al30 and included individual participant data from 8 additional trials (10 comparisons) from the Blood Pressure Lowering Treatment Trialists’ Collaboration for a total of 13 trials and 15 266 participants with grade 1 hypertension. Sundström et al33 reported a 22% reduction in all-cause mortality (OR, 0.78; 95% CI, 0.67-0.92), a 14% reduction in total cardiovascular events (OR, 0.86; 95% CI, 0.74-1.01), a 25% reduction in cardiovascular deaths (OR, 0.75; 95% CI, 0.57-0.98), and a 28% reduction in stroke (OR, 0.72; 95% CI, 0.55-0.94). Overall, cardiovascular event rates were low in the trials, so effect estimates for total cardiovascular events and coronary events were imprecise.33
Type 2 diabetes–specific treatment effects were not reported for BP-lowering treatment. Sundström et al33 reported no interaction by sex for all-cause mortality, total CVD events, coronary heart disease events, stroke events, or heart failure events, but a borderline interaction was present for CVD mortality (OR, 0.58 [95% CI, 0.41-0.81] for men; 1.19 [95% CI, 0.74-1.91] for women; P = .02).
Only 1 systematic review30 reported an increased risk for treatment withdrawals owing to adverse events, and this was derived from 1 trial (RR 4.80, 95% CI 4.14-5.57). However, the risk of bias for this outcome was high. The updated search by Sundström et al33 noted that data on treatment withdrawal were limited but equally common in the active and control groups when available. Data on health-related quality of life and direct costs were not reported in any review.
We rated the quality of evidence for the effect of BP-lowering therapy on reducing all-cause mortality, coronary heart disease, and stroke as high. We rated the quality of evidence for the effect of BP-lowering therapy on treatment withdrawals as low, downgraded because of study limitations and inconsistency.
We identified 13 reports of 12 systematic reviews of trials that investigated the effects of statins in primary prevention.34- 46 Type 2 diabetes–specific treatment effects were reported in 3 reviews,36,39,41 and sex-specific treatment effects were reported in 3 systematic reviews.34,37,43 The latest search ran through 2012.45
The 2 most comprehensive reviews included 18 to 20 primary prevention trials involving 56 934 to 63 899 participants.42,45 The mean age of participants generally ranged from 50 to 75 years, and mean age in the most comprehensive review was 57 years.45 Two reports from the Cholesterol Treatment Trialists’ Collaboration37,38 were included in our overview. One of these reported treatment effects for 70 025 participants without prevalent vascular disease,37 and the other reported treatment effects for 53 152 participants with a 5-year predicted risk for a major vascular event of 10% or less (>90% without prevalent vascular disease).38 A summary of systematic review characteristics is provided in eAppendix 3 in the Supplement.
The authors of included reviews reported that trials generally had a low risk of bias, but many were funded by pharmaceutical companies.45 The AMSTAR ratings ranged from a score of 5 of 11 to 11 of 11. The most comprehensive and highest rated systematic review was by Taylor et al.45
Treatment effects of statins on all-cause mortality, total CVD events, myocardial infarction, and stroke were broadly comparable across all systematic reviews. The systematic review by Taylor et al45 reported a 14% reduction in all-cause mortality (OR, 0.86; 95% CI, 0.79-0.94), a 25% reduction in major cardiovascular events (RR, 0.75; 95% CI, 0.70-0.81), and reductions in fatal and nonfatal coronary heart disease and stroke events.45 The mean (SD) baseline low-density lipoprotein cholesterol level in the Cholesterol Treatment Trialists’ Collaboration trials was 143.1 (27.1) mg/dL (to convert to millimoles per liter, multiply by 0.0259), and the mean (SD) difference in low-density lipoprotein cholesterol levels between participants in the statin regimen and controls was 41.8 mg/dL. Treatment effects standardized per 38.7–mg/dL reduction in levels of low-density lipoprotein cholesterol were also reported by the Cholesterol Treatment Trialists’ Collaboration38 for all-cause mortality (RR, 0.91; 95% CI, 0.85-0.97) and major vascular events (RR, 0.75; 95% CI, 0.70-0.80).
Effects of statin treatment in participants with diabetes were reported in 3 systematic reviews36,39,41 and demonstrated similar proportional benefits compared with the general primary prevention group. Sex-specific treatment effects were also reported in 3 systematic reviews,34,37,43 of which one was an individual participant data meta-analysis.37 After adjusting for baseline differences in prognostic characteristics and 5-year vascular risk, no sex-specific heterogeneity was seen.37 Additional statin treatment effects are provided in eAppendix 3 in the Supplement.
Data on adverse effects of statin treatment were included in 7 systematic reviews and included risks for cancer, elevation of creatine kinase levels, rhabdomyolysis, elevation of liver enzyme levels, hemorrhagic stroke, type 2 diabetes, and serious adverse effects.34,35,37,38,42,45,46 Taylor et al45 reported no evidence of increased risk for overall adverse effects (defined as cancers, myalgia and rhabdomyolysis, type 2 diabetes, hemorrhagic stroke, and other adverse effects leading to treatment discontinuation) among individuals treated with statins compared with control or placebo (RR, 1.00; 95% CI, 0.97-1.03). However, the risk for the individual outcome of type 2 diabetes for those treated with statins increased by 18% (RR, 1.18; 95% CI, 1.01-1.39).
Data on health-related quality of life were not reported. Data on cost-effectiveness were reported in Taylor et al45 from 3 statin trials, all demonstrating cost-effectiveness of statin therapy in primary prevention. In the West of Scotland Coronary Prevention Study (WOSCOPS), statin treatment led to 2460 years of life at £8121 (or $12 788) per life-year gained. In the Justification for the Use of Statins in Primary Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) trial, statin therapy had a cost-effectiveness of £25 796 (or $40 315) per quality-adjusted life-year. Last, in analyses from the Collaborative Atorvastatin Diabetes Study (CARDS), Taylor et al45 reported an incremental cost-effectiveness ratio of £2320 (or $3626) per quality-adjusted life-year at 10 years.
We rated the quality of evidence for the effect of statins on reducing the risk for all-cause mortality, major cardiovascular events, coronary heart disease events, and stroke as high. We rated the quality of evidence for the safety of statins as moderate, downgraded because of the indirectness of evidence.
We included 4 systematic reviews of tobacco cessation drugs.47- 50 None of the reviews differentiated between trials that included participants with and without prevalent vascular disease. Moreover, the reviews reported limited information about the effects of tobacco cessation drugs on cardiovascular outcomes or risk factors. The primary efficacy outcome for most reviews was continuous smoking cessation at 6 months. Type 2 diabetes–specific or sex-specific treatment effects were not reported. The latest search ran through 2012.48 The mean age among participants in the most comprehensive systematic review was 57 years.48
The most comprehensive systematic review was an overview of several Cochrane systematic reviews by Cahill et al48 that addressed drug therapy for smoking cessation. This overview synthesized information from 267 trials of 101 804 participants. Among the trials, 150 assessed nicotine replacement therapy, 72 trials assessed antidepressants (primarily bupropion hydrochloride), and 24 trials assessed nicotine receptor partial agonists (primarily varenicline tartrate). A summary of systematic review characteristics is provided in eAppendix 3 in the Supplement.
Authors reported that trials of nicotine replacement therapy and bupropion did not clearly describe methods of randomization or allocation concealment and were therefore at high risk of bias. Trials of varenicline were more contemporary and generally at lower risk of bias, although they were industry funded.48 The AMSTAR ratings ranged from 7 of 11 to 11 of 11, and the highest rated systematic review was by Cahill et al48 (eAppendix 3 in the Supplement).
The overview by Cahill et al48 reported treatment effects of tobacco cessation drugs on 6 months of continuous smoking abstinence, which was primarily verified by biochemical means. The overview reported that nicotine replacement therapy, bupropion, and varenicline all increased the odds of smoking cessation at 6 months compared with placebo (OR for nicotine replacement therapy, 1.84 [95% CI, 1.71-1.99]; OR for bupropion, 1.82 [95% CI, 1.60-2.06]; and OR for varenicline, 2.88 [95% CI, 2.40-3.47]). Cahill et al48 also reported the effects of a limited number of tobacco cessation drugs on CVD events. However, individual studies were underpowered for this end point, events were poorly reported, and many trials included participants with prevalent CVD. Treatment estimates for bupropion and varenicline for smoking cessation outcomes are listed in eAppendix 3 in the Supplement.
Cahill et al48 also reported the effects of tobacco cessation drugs on serious adverse events for a limited number of therapies. For nicotine replacement therapy, limited information on serious adverse effects was reported in trials.48 Data for bupropion and varenicline were reported (RR, 1.29 [95% CI, 0.99-1.69] and 1.06 [95% CI, 0.72-1.55], respectively). Data on health-related quality of life and direct costs were not reported.
We rated the quality of evidence for the effect of varenicline on smoking cessation as high and the quality of evidence for the effect of bupropion and nicotine replacement therapy on smoking cessation as moderate, downgraded because of study limitations. We rated the quality of evidence for the safety of tobacco cessation drugs as moderate, downgraded because of indirectness of evidence. We rated the quality of evidence for the effect of tobacco cessation drugs on all-cause mortality and cardiovascular outcomes as low, downgraded because of study limitations, inconsistency, and imprecision.
We performed an overview of systematic reviews that synthesized evidence of the efficacy and safety of drugs that can be used in primary ASCVD prevention to achieve targets set by the Million Hearts initiative and that will be used in the Million Hearts Cardiovascular Risk Reduction model. High-quality evidence suggests that aspirin, BP-lowering therapy, and statins reduce the risk for ASCVD events from 10% to 25% among individuals without prevalent ASCVD. High-quality evidence also suggests that BP-lowering therapy and statins reduce the risk for all-cause mortality by 11% and 14%, respectively. No heterogeneity in treatment effect was found among subgroups of individuals with type 2 diabetes or by sex. Moderate- to high-quality evidence suggests that tobacco cessation drugs increase the odds of continued abstinence by 88% to 188%, but the direct effects on cardiovascular events are uncertain. A summary of findings is presented in the Table.
Adverse effects of drug therapy were not reported or were poorly reported in many systematic reviews, and many authors of the included reviews noted underreporting of adverse events in the individual trials. Our search demonstrated high-quality evidence that aspirin increases the risk for major bleeding by 54% and moderate-quality evidence that statins do not increase the overall risk for adverse events, but the risk for type 2 diabetes was increased among individuals taking statins. Adverse effects of BP-lowering therapy and tobacco cessation drugs were poorly reported.
Our overview has several strengths. First, we focused this evidence synthesis on systematic reviews and meta-analyses of randomized clinical trials, because randomized clinical trials represent the highest-quality evidence to determine the effects of health care interventions. Second, we used a comprehensive, transparent search strategy to identify studies and followed a prespecified protocol to guide our evidence synthesis, noting any deviations from protocol. Third, we performed all title screening, data extraction, and quality assessments in duplicate to minimize potential bias in generation of this overview. Fourth, we used a validated instrument (the AMSTAR tool) to assess the methodologic quality of included systematic reviews and factored this quality assessment to guide our conclusions regarding the effects of pharmacologic interventions. This systematic process, with study quality assessment using standardized tools, could be used as a potential model for more rapid development of trustworthy guidelines.
Our overview also has important limitations to acknowledge. First, we did not retrieve data from primary trials and therefore were limited to the information and judgments of the authors who wrote the systematic reviews. Selection criteria, search strategies, and definitions of primary prevention often varied between reviews, and authors of the included reviews often used different criteria to define primary prevention, which led to different numbers of trials for systematic reviews of the same drug. Second, our conclusions were limited by the available data. Although data were generally well reported for drug efficacy, limited information was reported on safety, particularly for BP-lowering therapy and tobacco cessation drugs. Third, we had limited ability to comment on the potential of differential treatment effects by race/ethnicity. Nevertheless, several reviews noted consistent proportional treatment effects regardless of baseline characteristics, suggesting that treatment effect does not demonstrate heterogeneity. Fourth, our treatment effect estimates were limited by the short-term horizon of the clinical trials. Thus, we may have underestimated the potential added benefits and risks of sustained treatment. Fifth, our overview does not include evidence from recent trials like the Japanese Primary Prevention Project (JPPP)53 or cost-effectiveness analyses of long-term statin use from the WOSCOPS.54 However, inclusion of these studies would not have changed our overall conclusions. For example, the point estimate for treatment effect from low-dose aspirin in the JPPP trial was similar to our reported effect, although with wider confidence intervals (hazard ratio, 0.94; 95% CI, 0.77-1.15), and primary prevention statin therapy was also shown to be cost-effective at 15 years of follow-up. Finally, our overview does not provide information on the added effects of lifestyle interventions such as diet, exercise, and weight loss in combination with drug therapy. However, essentially all included clinical trials of ASCVD prevention included background recommendations of therapeutic lifestyle change in combination with study drugs.
This overview of systematic reviews demonstrates high-quality evidence to support aspirin, BP-lowering therapy, and statins for primary ASCVD prevention and tobacco cessation drugs for smoking cessation. The overview provides reliable, evidence-based pooled estimates for these interventions on the lowered risk for primary ASCVD events and best-available evidence for the effect of tobacco cessation drugs on continuous abstinence at 6 months. These treatment effects can be used to enrich discussions between health care professionals and patients.
Correction: This article was corrected on February 1, 2017, to fix a conversion factor in the Table and Results section.
Corresponding Author: Mark D. Huffman, MD, MPH, Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, 680 N Lake Shore Dr, Ste 1400, Chicago, IL 60660 (firstname.lastname@example.org).
Accepted for Publication: February 10, 2016.
Published Online: April 27, 2016. doi:10.1001/jamacardio.2016.0218
Author Contributions: Drs Karmali and Huffman had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Karmali, Lloyd-Jones, Berendsen, Goff, Sanghavi, Brown, Huffman.
Acquisition, analysis, or interpretation of data: Karmali, Berendsen, Korenovska, Huffman.
Drafting of the manuscript: Karmali, Lloyd-Jones, Berendsen, Huffman.
Critical revision of the manuscript for important intellectual content: Karmali, Lloyd-Jones, Goff, Sanghavi, Brown, Korenovska, Huffman.
Statistical analysis: Karmali, Korenovska, Huffman.
Obtained funding: Goff, Lloyd-Jones.
Administrative, technical, or material support: Karmali, Berendsen, Brown, Korenovska, Huffman.
Study supervision: Sanghavi, Brown, Huffman.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Huffman reports receiving grants from the Center for Medicare & Medicaid Innovation (CMMI) via subcontract from The MITRE Corporation during the conduct of the study and grants from World Heart Federation outside the submitted work and serving as an associate editor for JAMA Cardiology and the coordinating editor of the Cochrane Heart Group US Satellite. Dr Lloyd-Jones reports receiving grants from the CMMI via subcontract from The MITRE Corporation during the conduct of the study. No other disclosures were reported.
Funding/Support: This study was supported by the CMMI through work conducted under contract No. HHSM-500-2012-00008I.
Role of Funder/Sponsor: This study was performed to provide a synthesis of available evidence to support development of a longitudinal cardiovascular risk calculator that will be used in the Million Hearts Cardiovascular Risk Reduction Model. The CMMI provided an initial description of key questions. However, the final protocol was developed independently by the authors. The authors were solely responsible for data collection, management, analysis, and interpretation. Members of CMMI and The MITRE Corporation reviewed a draft version of a full, more detailed report, and the final version used by the group explicitly addressed comments and suggestions provided. Final interpretation of the data, including judgments of evidence quality, was solely the responsibility of the authors. The sponsor had no role in the review or approval of the manuscript. The sponsor was involved in the decision to submit the manuscript for publication in JAMA Cardiology to provide transparency regarding the evidence base for the new cardiovascular risk calculator.
Disclaimer: Dr Huffman is an associate editor of JAMA Cardiology but was not involved in the editorial review or the decision to accept the manuscript for publication.
Additional Contributions: Janet S. Wright, MD, Million Hearts initiative, Centers for Disease Control and Prevention, provided input in preparing this manuscript. She received no compensation for this role.