Summary of Review Findings

Table 9 summarizes the evidence reviewed for this update. In adults at increased cardiovascular risk but without prior cardiovascular events, statin therapy was associated with reduced risk of clinical outcomes compared with placebo or no statin use, based on pooled evidence from 19 trials with 6 months to 6 years of followup. Although the trials evaluated diverse patient populations (e.g., patients with hyperlipidemia, diabetes, hypertension, early-onset cerebrovascular disease, or elevated CRP levels), findings were generally consistent across trials in favoring statin therapy versus placebo or no statin for various individual cardiovascular outcomes (NNT to prevent 1 event ranged from 123 [MI] to 263 [stroke]), cardiovascular mortality (NNT, 500), and composite cardiovascular outcomes (NNT, 72). Pooled results indicated a decreased risk of all-cause mortality (15 trials; RR, 0.86 [95% CI, 0.80 to 0.93]; I²=0%; ARD, -0.40%; NNT, 250 after 1 to 6 years), cardiovascular mortality (10 trials; RR, 0.82 [95% CI, 0.71 to 0.94]; I²=0%; ARD, -0.20%; NNT, 500 after 2 to 6 years), stroke (13 trials; RR, 0.71 [95% CI, 0.62 to 0.82]; I²=0%; ARD, -0.38%; NNT, 263 after 6 months to 6 years), MI (12 trials; RR, 0.64 [95% CI, 0.57 to 0.71]; I²=0%; ARD, -0.81%; NNT, 123 after 2 to 6 years), revascularization (7 trials; RR, 0.63 [95% CI, 0.56 to 0.72]; I²=0%; ARD, -0.66%; NNT, 152 after 2 to 6 years), and composite cardiovascular outcomes (13 trials; RR, 0.70 [95% CI, 0.63 to 0.78]; I²=36%; ARD, -1.39%; NNT, 72 after 1 to 6 years). Our analyses include the large HOPE-3 trial,¹⁰³ which was identified during an update search and reported results consistent with previously published trials. Findings were generally robust in sensitivity and stratified analyses based on trial quality, duration of followup, baseline TC or LDL-C levels, exclusion of trials that were stopped early, and exclusion of trials that enrolled a small proportion of patients with prior cardiovascular events. A challenge in interpreting the NNT is that estimates vary across studies depending on the baseline risk of the population and the duration of followup.

Table 9

Summary of Evidence.

Our findings regarding benefits of statin therapy were generally consistent with recent high-quality systematic reviews^105,137-139 that primarily focused on patients without prior cardiovascular events, though there was variability in inclusion criteria (e.g., inclusion of trials in which a small proportion of patients had prior cardiovascular events, trials of patients with specific conditions such as severe kidney disease, or trials of statins for prevention of noncardiovascular outcomes such as Alzheimer's disease), use of individual patient data,¹³⁷ and methods for analyzing outcomes (e.g., events that occurred during statin therapy or inclusion of events that occurred after treatment was discontinued). For all-cause mortality, our point estimate was very similar to the estimates reported in recent systematic reviews,^105,137,138 though in one of the reviews the difference was not statistically significant (RR, 0.91 [95% CI, 0.83 to 1.01]).¹³⁷ However, that review did not include the large, recently published HOPE-3 trial.¹⁰³

Effects of statins also appeared to be similar in patient subgroups defined according to demographic characteristics such as age, sex, and race/ethnicity and clinical characteristics such as presence of diabetes or renal dysfunction. For hypertension, three trials found no clear differences in estimates of effects of statins when patients were stratified according to presence or absence of hypertension.^74,83,103 However, the large ASCOT-LLA trial (n=10,305), which enrolled patients with treated or untreated hypertension and at least three other cardiovascular risk factors, found that statin therapy was associated with no clear effect on cardiovascular mortality (HR, 0.90 [95% CI, 0.66 to 1.23]), though results for other cardiovascular outcomes and all-cause mortality were generally consistent with other trials. The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT-LLT) (n=10,355), which focused on patients with stage 1 or 2 hypertension and at least one other cardiovascular risk factor, was excluded because about 15 percent of patients had prior CHD. It found no clear effects of statin therapy versus placebo on all-cause mortality, cardiovascular mortality, stroke, or fatal or nonfatal MI (RR estimates ranged from 0.91 to 0.99).¹⁴⁰ Challenges in interpreting the results of ALLHAT-LLT are use of an open-label design with high crossover (resulting in a modest reduction in LDL-C levels of about 24 mg/dL with statin therapy) and lower than projected sample size, resulting in decreased statistical power.¹⁴¹

For effects in subgroups defined by sex, our findings are in accordance with a pooled analysis on the effects of statins in women enrolled in JUPITER,⁷⁴ AFCAPS/TexCAPS,⁵⁴ and MEGA,⁸³ which reported pooled estimates for all-cause mortality (RR, 0.78 [95% CI, 0.53 to 1.15]) and cardiovascular events (RR, 0.63 [95% CI, 0.49 to 0.82]) that were similar to our pooled estimates.⁸¹ Results from a good-quality systematic review on the effect of statins in women that included trials^140,142 in which more than 10 percent of the population had prior CVD events also reported similar estimates for all-cause mortality (3 studies; RR, 0.90 [95% CI, 0.60 to 1.35]; I²=11%) and CHD events (6 studies; RR, 0.78 [95% CI, 0.64 to 0.96]; I²=7%).¹⁴³

Benefits did not appear to be restricted to patients with severely elevated lipid levels, as similar effects were observed in subgroups stratified according to baseline TC or LDL-C level^{54,60,83,96, 103,107} and trials that excluded patients with severe dyslipidemia but included those who had other cardiovascular risk factors.^{60,63,65,69,74} Similarly, trials that stratified patients according to a baseline global cardiovascular risk score reported similar risk estimates in those classified as higher and lower assessed risk.^54,74,103 Given similar RR estimates, however, the absolute benefits of statin therapy will be greater in patients at higher baseline risk. This has implications for determining the cardiovascular risk threshold used to select patients for statin treatment (e.g., 10-year risk >7.5% vs. >10%). In JUPITER, which enrolled patients with an LDL-C level of less than 130 mg/dL and a CRP level of 2.0 mg/L or greater, a post-hoc analysis found that the incidence of cardiovascular events in patients with at least one additional cardiovascular risk factor was nearly twice as high as in those without additional risk factors (15.5 vs. 7.7 events per 1,000 patient-years),^107,144 resulting in a NNT to prevent 1 cardiovascular event about twice as high in the subgroup without additional risk factors, based on a similar estimate of relative effect.⁷⁴

We found no evidence that statin treatment in adults without prior cardiovascular events is associated with increased risk of withdrawal due to adverse events, serious adverse events, cancer, or elevated liver enzymes versus placebo or no statin therapy. Our findings are generally consistent with recent systematic reviews, some of which also included trials of statins for secondary prevention.^{34,35,105,145} Similar to other meta-analyses of trials of primary and secondary prevention,^31,146 we found no increased risk of muscle-related harms with statin use, although some observational studies of patients taking statins for various indications found an increased risk of myopathy compared with nonuse.¹⁴⁷ While none of the trials found that statins were associated with increased risk of myalgia versus placebo, one recent trial of healthy, statin-naïve subjects reported an increased risk of myalgia using predefined criteria (including resolution after discontinuation of study drug and recurrence on rechallenge) with high-intensity statin therapy (atorvastatin 80 mg/day) versus placebo for 6 months that was just below the threshold for statistical significance (9.4% vs. 4.6%; RR, 2.03 [95% CI, 0.97 to 4.26]),¹⁴⁸ and a trial of patients who discontinued statins due to muscle symptoms found that rechallenge with a statin was associated with increased risk of muscle symptoms versus placebo.¹⁴⁹ The HOPE-3 trial found that statins were associated with increased risk of cataract surgery, an unanticipated finding. None of the other primary prevention trials evaluated risk of cataracts or cataract surgery. A systematic review that included secondary prevention trials and observational studies reported findings discordant with HOPE-3, as statins were associated with a decreased risk of cataract incidence (OR, 0.81 [95% CI, 0.71 to 0.93]) and surgery (OR, 0.66 [95% CI, 0.61 to 0.71]).¹⁵⁰

In contrast with systematic reviews of primary and secondary prevention trials that reported a slightly increased risk of diabetes with statin therapy (OR, 1.09 [95% CI, 1.02 to 1.17]^112,151 and RR, 1.13 [95% CI, 1.03 to 1.23]¹⁵²), we found no increased risk of diabetes in six trials of patients without prior cardiovascular events (RR, 1.05 [95% CI, 0.91 to 1.20]; I²=52%). Another systematic review that limited analysis to primary prevention trials also found no increased risk of diabetes with statin use (4 trials; RR, 1.05 [95% CI, 0.84 to 1.32]).¹³⁸ However, results of individual primary prevention trials were inconsistent, with one large trial (JUPITER) showing increased risk of diabetes (3.0% vs. 2.4%; RR, 1.25 [95% CI, 1.05 to 1.49]).⁷⁴ A difference between JUPITER and the other trials in our analysis is that it was the only trial to evaluate high-potency statin therapy. Other analyses that included trials of statins for secondary prevention have suggested an association between intensity of statin dose and risk of incident diabetes.^{138,151, 153,154} In JUPITER, the risk of diabetes was increased in patients with risk factors for diabetes at baseline but not in persons without diabetes risk factors. Based on JUPITER, among patients with diabetes risk factors, 134 cardiovascular events were prevented for every 54 incident cases of diabetes, while among persons without diabetes risk factors, 86 cardiovascular events were prevented and no incident cases of diabetes were diagnosed.¹⁰⁹ One mechanism by which statins may increase risk of diabetes is through a modest increase in body weight.^155,156

Evidence on the association between statin use in adults without prior cardiovascular events and renal or cognitive harms was sparse but found no clear increase in risk. Our findings are consistent with a recent systematic review of RCTs and observational studies on the effect of statins on cognition that found no effect on incidence of Alzheimer's disease or dementia and no differences in performance on tests of procedural memory, attention, motor speed, global cognitive performance, executive function, declarative memory, processing speed, or visuoperception.³⁴ Unlike our review, this systematic review included trials of patients receiving statins for any reason, including for prevention of cognitive decline or dementia and for secondary prevention following a cardiovascular event. A recent cohort study in which most patients receiving statin therapy had a history of CVD found that statins and nonstatin lipid-lowering drugs were associated with similar risk of acute memory loss in the first 30 days following exposure, suggesting that either all lipid-lowering drugs cause acute memory loss or that the observed association is due to detection bias rather than a causal association.¹⁵⁷

Recent guidelines from the ACC/AHA³⁰ differ from prior ATP III guidelines¹⁶ in recommending fixed-dose statin therapy, with the intensity of therapy determined by cardiovascular risk factors, rather than titration of statin therapy to achieve target LDL-C levels. We identified no study that directly compared treatment with statins titrated to attain target cholesterol levels versus other fixed-dose or other treatment strategies. Although indirect comparisons based on trials of statins versus placebo or no statin that permitted dose titration compared with those that used fixed-dose therapy showed no clear differences in risk of all-cause or cardiovascular mortality, MI, or stroke, only three^52,54,83 of 18 trials permitted limited dose titration (no trial involved titration from low- to high-intensity statin therapy, and one trial only titrated within the low-intensity category), precluding strong conclusions.

Little direct evidence was available to determine effects of statin therapy intensity on clinical outcomes or adverse events. Two trials that directly compared different statin intensities were underpowered to evaluated clinical outcomes.^66,92 Indirect comparisons based on trials of statins versus placebo or no statin stratified according to the intensity of therapy were also limited, as most trials evaluated moderate-intensity therapy. For all-cause mortality, risk estimates were similar in trials of low-intensity (RR, 0.72 [95% CI, 0.52 to 1.00]; I²=0%), moderate-intensity (RR, 0.88 [95% CI, 0.80 to 0.97]; I²= 0%), and high-intensity (RR, 0.80 [95% CI, 0.67 to 0.97]; I²=0%) statins. For other clinical outcomes, there were too few trials of low- and high-intensity statins to conduct meaningful comparisons. A meta-analysis of individual patient data found an association between the degree of LDL-C reduction and reduced risk of clinical outcomes, potentially providing indirect evidence regarding the relative effectiveness of higher- versus lower-intensity statin therapy.⁴⁸ Although this analysis included trials of patients with prior cardiovascular events, estimates were similar in patients with an estimated 5-year risk of less than 5 percent or of 5 to 10 percent, a subgroup unlikely to include persons with prior cardiovascular events. A good-quality systematic review also found no clear effects of statin intensity on benefits or harms outcomes but categorized different statins as low (fluvastatin, lovastatin, pravastatin, simvastatin) or high (atorvastatin and rosuvastatin) potency without consideration of statin dose or estimated lipid-lowering effect.¹³⁸

Limitations

Our review had some limitations. Statistical heterogeneity was present in several pooled analyses. Therefore, we used the DerSimonian and Laird random-effects model to pool studies. The DerSimonian and Laird random-effects model may result in CIs that are too narrow when heterogeneity is present, particularly when the number of studies is small.⁵⁰ Therefore, we repeated analyses in which statistical heterogeneity was present using the profile likelihood method, which resulted in similar findings. To address statistical heterogeneity, we also performed sensitivity and subgroup analyses based on study quality, duration of followup, intensity of statin therapy, baseline lipid levels, and exclusion of trials that enrolled some patients with prior cardiovascular events. Although statistical heterogeneity remained present in some analyses, results were generally robust in sensitivity and stratified analyses.

We did not have access to individual patient data. Therefore, our findings are based on analyses of study-level data and our ability to analyze effects in subgroups was restricted to published reports. An individual patient data meta-analysis found that the effect of statins for primary prevention on all-cause mortality did not reach statistical significance (RR, 0.91 [95% CI, 0.83 to 1.01, though the estimate favored statins.¹³⁷ The HOPE-3 trial was not included in the analysis; however, because it had access to individual patient data, the meta-analysis was able to include some trials that we excluded because more than 10 percent of the population had prior cardiovascular events.^140,158 For trials that we included in which less than 10 percent of patients had prior cardiovascular events, the meta-analysis was also able to separately analyze patients without prior cardiovascular events; our analyses were based on results for the whole population. However, excluding the latter trials from our analyses did not affect our findings.

We also used indirect comparisons when direct evidence was unavailable or limited to evaluate effects of titrated versus fixed-dose statin therapy, intensity of statin therapy, and subgroup effects. Although findings based on indirect comparisons were generally consistent with available direct evidence, results based on indirect comparisons should be interpreted with caution.¹⁵⁹

We excluded non–English-language articles, which could result in language bias. However, some research suggests that English-language restriction has little effect on the conclusions of systematic reviews of topics other than complementary medicine, and we did not identify any large non-English trials of statins versus placebo referenced in other systematic reviews.^160,161 We only formally assessed for publication bias using statistical and graphical methods to assess for small sample effects when there were at least 10 studies, as research indicates that such methods can be misleading with smaller numbers of studies.⁵¹ We found no evidence of small sample effects but cannot exclude the possibility of publication bias in analyses based on smaller numbers of trials. Only two trials received no industry funding.^52,92 Although research has found an association between receipt of industry funding and biased estimates,^162-164 analyses of statin trials have found no association between funding source and degree of LDL-C reduction.¹⁶⁵

Emerging Issues/Next Steps

Determining the optimal methods for assessing cardiovascular risk has recently received increased scrutiny. Although the ACC/AHA guideline recommends the use of the newly developed Pooled Cohort Equation to predict risk,³⁰ some validation studies have found that it overpredicts cardiovascular risk, with more overestimation appearing to occur in more contemporary cohorts not used to develop the instrument.^{36,128,130,166} There is also ongoing interest in use of newer methods to supplement traditional risk factors for predicting cardiovascular risk, such as measurement of coronary artery calcium score, measurement of carotid intima-media thickness, CRP levels, and alternative lipid measures.^1,113

Other clinical practices around use of statins may also be changing due to the release of the 2013 ACC/AHA guideline.³⁰ Recommendations in the ACC/AHA guideline differ substantially from the ATP III guideline in recommending fixed-intensity statin therapy without specific target LDL-C levels. Adoption of these recommendations could substantially impact practices related to lipid levels and other monitoring in patients taking statin therapy. The ACC/AHA also recommends a lower threshold for initiation of treatment with a statin in patients without prior cardiovascular events, which analyses indicate would substantially increase the number of patients eligible for therapy.^36,38-40

Although pitavastatin was recently approved by the U.S. Food and Drug Administration, no trial of statin therapy in patients without prior cardiovascular events evaluated this drug. Drugs in the proprotein convertase subtilisin/kexin type 9 (PCSK9) class have also been recently approved by the U.S. Food and Drug Administration for use with diet and maximally tolerated statin therapy in persons with familial hyperlipidemia or clinical atherosclerotic CVD who require additional LDL-C reduction. PCSK9 drugs reduce LDL-C levels by about 60 percent compared with standard therapy, including maximally tolerated statins, although evidence on effects on clinical outcomes is limited at this time.^167,168 More research is needed to understand the benefits and harms of this class of drugs in persons without prior cardiovascular events, including persons who cannot tolerate statin therapy.

Relevance for Priority Populations

Statin therapy appears to be similarly effective in younger and older adults, based on RR estimates. Because risk of cardiovascular events increases with age, however, statin therapy in older adults is associated with greater absolute benefits. For example, in the JUPITER trial, the NNT to prevent 1 cardiovascular event was 62 in persons age 70 years or older and 94 in those younger than age 70 years.⁷⁴ The trials of statin therapy included in this review reported no increased risk of muscle-related, liver-related, renal, oncologic, or cognitive adverse events versus placebo, but only one trial evaluated potential interactions between age and adverse events (and found no statistically significant interaction).⁷⁴ However, older persons may be at increased risk of adverse events due to use of concomitant medications or comorbid conditions, warranting additional research to fully understand the balance of benefits and harms in this population. Evidence regarding benefits and harms of statin therapy in persons older than age 80 years is very limited, as most trials were restricted to younger patients, and in trials that did enroll patients older than age 80 years, results were not reported for this subgroup.¹⁶⁹ We identified one trial of fluvastatin versus placebo in which half of the study population (n=1,229) was age 75 years or older. However, it was not designed to assess clinical outcomes and did not meet inclusion criteria.¹⁷⁰

Evidence on effects of statin therapy in racial/ethnic minorities was very limited. The only trials to report effects of statin therapy versus placebo found no significant interaction between effects of statins and race/ethnicity.^74,103 In trials that reported race/ethnicity, whites were the predominant group in all but one trial.¹⁰³

Future Research

Several research gaps limit the full understanding of benefits and harms of statin therapy. Trials that directly compare titrated statin therapy to target lipid levels versus fixed-dose therapy would help to inform optimal dosing strategies. Trials that directly compare higher- versus lower- intensity statin therapy and are powered to assess clinical outcomes are also needed. Additional research would be helpful for more definitively determining whether statin therapy is associated with increased risk of diabetes, cognitive harms, or cataracts. More research is also needed to clarify benefits and harms of statins in subgroups, including persons older than age 80 years. Evidence to determine whether effectiveness of statin therapy varies in racial/ethnic minorities remains sparse.

Additional research is needed to validate the predictive accuracy of the ACC/AHA Pooled Cohort Equation to predict cardiovascular risk, in order to help guide optimal methods for risk assessment and to understand how overprediction of cardiovascular risk would impact the number of patients treated and estimates of benefit. Studies that compare strategies based on global risk assessment scores versus presence of defined cardiovascular risk factors could help to further clarify optimal methods to select patients for statin therapy. Research is also needed to better understand how frequently cardiovascular risk assessment (including lipid testing) should be performed, ideally by directly comparing how different assessment intervals impact use of statin therapy as well as subsequent clinical outcomes.

Conclusions

In adults at increased cardiovascular risk but without prior cardiovascular events, statin therapy is associated with reduced risk of all-cause and cardiovascular mortality and cardiovascular events. Benefits appear to be present across diverse demographic and clinical subgroups, with greater absolute benefits in patients at higher baseline risk, and do not appear to be restricted to patients with marked hyperlipidemia.