Compared with individuals with more preserved renal function, those with moderate chronic kidney disease (CKD) (estimated glomerular filtration rate [eGFR] <60 ml/min/1.73 m²) are at increased risk of myocardial infarction, stroke, and vascular death (1- 3). However, 2 recent trials have found statin therapy to be ineffective among high-risk patients with severe renal failure undergoing maintenance hemodialysis (4- 5). Given these data, questions have been raised in the clinical community about the relative efficacy of statin therapy among those with moderate CKD. This is particularly relevant in primary prevention where available data are scarce; in the WOSCOPS (West of Scotland Coronary Prevention Study) primary prevention trial, pravastatin had no significant benefit on clinical cardiovascular events among those with CKD, despite evidence in the same trial of efficacy among those with eGFR ≥60 ml/min/1.73 m² and evidence of efficacy based on a meta-analysis of secondary prevention trials conducted among patients with and without CKD (6).

The recently completed JUPITER (Justification for the Use of statins in Prevention–an Intervention Trial Evaluating Rosuvastatin) trial of 17,802 men and women with low-density lipoprotein cholesterol (LDL-C) <130 mg/dl but elevated high-sensitivity C-reactive protein (hsCRP) (≥2 mg/l) demonstrated a 44% reduction in major vascular events and a 20% reduction in all-cause mortality for those allocated to rosuvastatin 20 mg as compared with placebo during a median follow-up of 1.9 years (maximum 5 years) (7). Because the JUPITER trial included 3,267 primary prevention patients with eGFR <60 ml/min/1.73 m² at trial entry, we had the opportunity to perform a secondary analysis of these data according to the presence or absence of moderate CKD.

The study population derived from the JUPITER trial, a randomized, double-blind, placebo-controlled trial designed to investigate whether rosuvastatin 20 mg daily compared with placebo decreases the rate of first-ever cardiovascular events among apparently healthy men over age 50 years and women over age 60 years with LDL-C <130 mg/dl at increased vascular risk due to hsCRP ≥2 mg/l (7- 8).

Full details of the trial protocol, procedures, and methods of confirming clinical end points and ascertaining adverse events have been previously presented. Trial exclusion criteria included treatment within 6 weeks of screening with any lipid lowering therapies, current use of hormone replacement therapy, evidence of hepatic dysfunction, creatinine >2.0 mg/dl, diabetes, uncontrolled hypertension, prior malignancy, uncontrolled hypothyroidism, or a recent history of alcohol, drug abuse, or other medical condition that might compromise safety. Because a core scientific hypothesis of the JUPITER trial related to underlying low-grade inflammation, individuals with conditions such as severe arthritis, lupus, or inflammatory bowel disease were excluded, as were individuals taking immunosuppressant agents. As previously reported, allocation to rosuvastatin was associated with a 44% reduction in the trial primary end point (nonfatal myocardial infarction, nonfatal stroke, hospital stay for unstable angina, arterial revascularization, or cardiovascular death) (7) as well as a 43% reduction in the pre-specified secondary end point of incident venous thromboembolism (8).

Study participants had baseline creatinine levels measured, which were used to calculate eGFR with the Modified Diet and Renal Disease method. Baseline clinical characteristics of the study population were compared between those with moderate CKD at baseline and those with baseline eGFR ≥60 ml/min/1.73 m²; in these analyses, significance was evaluated with t tests for continuous variables and the chi-square test for proportions. In stratified analyses according to eGFR at study entry, Cox proportional hazards models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for the comparison of event rates between those allocated to rosuvastatin or placebo. All analyses were performed on an intention-to-treat basis. Analyses were performed for the JUPITER pre-specified primary end point (nonfatal myocardial infarction, nonfatal stroke, hospital stay for unstable angina, arterial revascularization, or confirmed cardiovascular death) and for combined end points that additionally included incident venous thromboembolism and all-cause mortality. Per protocol, number needed to treat [NNT] values were calculated as the reciprocal of the absolute difference between risks of the outcome of interest based on Kaplan-Meier estimates. Estimated 95% CIs for the NNT were based on inversion of the CIs for risk differences with standard errors of risks estimated by Greenwood's formula. Consistent with prior JUPITER publications, 5-year NNT values were computed on the basis of 4-year absolute rates projected over an average 5-year period according to the methods of Altman and Anderson (9).

Of participants in the JUPITER trial, 3,267 (18%) had baseline eGFR <60 ml/min/1.73 m²), whereas 14,528 (82%) had higher levels. Seven participants did not have eGFR values available. Among those with reduced eGFR, 3,253 had stage 3 impairment (eGFR between 30 and 59 ml/min/1.73 m²) and 14 had stage 4 impairment (eGFR between 15 and 29 ml/min/1.73 m²).

Study participants with moderate CKD were older, more likely to be female, more likely to have a family history of premature atherothrombosis, and less likely to smoke (Table 1). Median baseline levels of LDL-C, high-density lipoprotein cholesterol, triglycerides, apolipoprotein A, apolipoprotein B, and hsCRP were somewhat higher among those with moderate CKD, whereas blood pressure, glucose, and hemoglobin A1c were similar. Within each eGFR category, there was no imbalance between study characteristics among those allocated to rosuvastatin or placebo.

Irrespective of statin or placebo allocation, at trial conclusion 111 participants in the moderate CKD group suffered a primary trial end point (incidence rate 1.51 per 100 person-years) as compared with 282 participants in the group with more preserved eGFR (incidence rate 0.95 per 100 person-years). Thus, in crude analysis, those with lower eGFR at study entry were at significantly higher risk of developing a study primary end point during follow-up (HR: 1.54, 95% CI: 1.23 to 1.92, p = 0.0002). Those with lower eGFR were at increased risk of developing myocardial infarction, stroke, or cardiovascular death (HR: 1.44, 95% CI: 1.08 to 1.92, p = 0.02) as well as arterial revascularization (HR: 1.53, 95% CI: 1.13 to 2.08, p = 0.008). All-cause mortality was similar in the 2 groups (incidence rates 1.19 vs. 1.11 per 100 person-years, respectively; HR: 1.06, 95% CI: 0.85 to 1.33, p = 0.61) as were rates of venous thromboembolism (0.31 and 0.24 per 100 person-years, respectively; HR: 1.26, 95% CI: 0.79 to 2.02, p = 0.35).

Adjustment for baseline differences between groups had little influence on these rates; after adjustment for age, sex, smoking, and drug allocation, the HR for the primary trial end point comparing those with CKD with those with preserved renal function remained 1.54 (95% CI: 1.22 to 1.96, p = 0.0004).

Among those with moderate CKD at study entry, 71 suffered a primary trial end point among those allocated to placebo (incidence rate 1.95 per 100 person-years) as compared with 40 among those allocated to rosuvastatin (incidence rate 1.08 per 100 person-years, HR: 0.55, 95% CI: 0.38 to 0.82, p = 0.002) (Table 2).

An almost identical effect of rosuvastatin was observed among those with more preserved renal function; in this subgroup, 180 suffered a primary trial end point among those allocated to placebo (incidence rate 1.21 per 100 person-years) as compared with 102 among those allocated to rosuvastatin (incidence rate 0.69 per 100 person-years, HR: 0.57, 95% CI: 0.45 to 0.72, p < 0.001).

(Figure 1)A presents cumulative incidence curves for the primary end point among those with and without moderate CKD, according to rosuvastatin assignment. (Figure 1)B and C show similar data for the primary end point plus all-cause mortality and venous thromboembolism.

As also shown in (Table 2), rosuvastatin had similar efficacy for most individual clinical outcomes within the JUPITER trial when comparing participants with and without moderate CKD. A possible exception was total mortality, which was reduced 44% by rosuvastatin among those with moderate CKD (HR: 0.56, 95% CI: 0.37 to 0.85, p = 0.005) as compared with 12% among those with more preserved eGFR (HR: 0.88, 95% CI: 0.72 to 1.09, p = 0.25) (p for interaction = 0.048). Effects of rosuvastatin were consistent across all pre-specified subgroups evaluated (Figure 2).

Among those with moderate CKD, the NNT at 5 years was 14 (95% CI: 9 to 30) for the primary end point and 9 (95% CI: 6 to 17) for the end point that also included total deaths and venous thromboembolism. Among those with more preserved renal function, the corresponding values were 35 (95% CI: 22 to 82) and 25 (95% CI: 16 to 58).

As previously reported (7), there is no evidence in the JUPITER trial that rosuvastatin led to any impairment of renal function as measured by eGFR at the 12-month visit; if anything, in the trial as a whole, median eGFR at 12 months was marginally improved among those randomly allocated to rosuvastatin as compared with placebo (66.8 vs. 66.6 ml/min/1.73 m², p = 0.02). In analyses stratified by baseline eGFR, similar results were observed. Among those with eGFR <60 ml/min/1.73 m² at baseline, the median eGFR levels at 12 months were 53.0 and 52.8 ml/min/1.73 m² in the rosuvastatin and placebo groups, respectively (p = 0.44). Among those with eGFR ≥60 ml/min/1.73 m² at baseline, the median eGFR levels at 12 months were 70.5 and 70.0 ml/min/1.73 m² in the rosuvastatin and placebo groups, respectively (p = 0.007).

Within the JUPITER trial, the largest relative risk reductions were observed among those who not only reduced LDL-C more than 50% but also reduced hsCRP levels more than 50%; as reported previously, a 79% relative risk reduction in vascular events was observed in the JUPITER trial among those who achieved low levels of both LDL-C and hsCRP (10). We thus also sought to evaluate the reduction in both LDL-C and hsCRP according to baseline renal function. Among those with moderate CKD, rosuvastatin reduced LDL-C by 52% and hsCRP by 37%, whereas among those with more preserved renal function, LDL-C was reduced by 46% and hsCRP by 37% (all p values <0.001) (Table 3).

Adverse event rates associated with rosuvastatin were similar in the JUPITER trial among those with and without moderate CKD (Table 4).

In this secondary analysis of participants in the JUPITER trial, rosuvastatin was effective at reducing cardiovascular event rates and all-cause mortality among the 3,267 study participants who had evidence of moderate CKD at study entry. As anticipated, absolute rates of future vascular disease were higher among those with moderate CKD. Thus, absolute risk reductions associated with rosuvastatin were higher (and NNT values were lower) among JUPITER participants with eGFR <60 ml/min/1.73 m² when compared with those with higher eGFR levels. As such, these data support guidelines from the American Heart Association and the National Kidney Foundation to provide more aggressive cardiovascular prevention efforts among those with mild to moderately reduced renal function (2- 3,11).

The increased absolute vascular risk in the JUPITER trial among those with moderate CKD is consistent with prior observational cohorts and trials (1- 3,6). However, our data regarding the efficacy of statin therapy is consistent with some but not all prior data. For example, although the Pravasatatin Pooling Project found that pravastatin significantly reduced vascular event rates and mortality in a combined analysis of CKD patients participating in the WOSCOPS, CARE (Cholesterol and Recurrent Events), and LIPID (Long-term Intervention with Pravastatin in Ischemic Disease) trials, the primary prevention data in this overview were not statistically significant (6). Thus, our finding of benefit for rosuvastatin among a primary cardiovascular prevention cohort who have concomitant moderate CKD importantly extends prior data. Our observation of benefit with rosuvastatin 20 mg in a population with moderate CKD is also consistent with data from the secondary prevention Treat-to-New-Targets trial where atorvastatin 80 mg was superior to atorvastatin 10 mg among those with and without CKD (12). By contrast, our data conflict with those from both the German Diabetes and Dialysis Study (4) and the AURORA (A Study to Evaluate the Use of Rosuvastatin in Subjects on Regular Haemodialysis: an Assessment of Survival and Cardiovascular Events) trial (5), where those receiving maintenance hemodialysis failed to have a clinical benefit from statin therapy.

Patients receiving hemodialysis differ in many ways from those with moderate CKD, and thus our data demonstrate that null data from these settings should not be generalized to that of individuals without severe renal failure. For example, individuals receiving hemodialysis tend to have more calcification of arterial plaque than those with normal renal function. Nonetheless, our data raise a potential explanation for the failure of statins in the setting of hemodialysis; in both the German Diabetes and Dialysis Study and the AURORA study, LDL-C reductions were large, but reductions in hsCRP were small, likely the result of chronic and repeated pro-inflammatory stimulation associated with the dialysis membrane (4- 5). By contrast, within the JUPITER trial, those with and without moderate CKD had a similar magnitude of LDL-C and hsCRP reduction. Because both LDL-C reduction and hsCRP reduction have been independent predictors of clinical benefit in several statin trials (13- 16) as well as in the JUPITER trial (10), the possibility that the anti-inflammatory effects of statin therapy were blunted in the hemodialysis trials merits consideration. This hypothesis requires testing in future trials.

A limitation of our analysis is that creatinine >2 mg/dl was an exclusion criterion. Despite this exclusion, 3,276 individuals were randomized into the trial with eGFR <60 ml/min/1.73 m², reinforcing the observation that many individuals with creatinine levels often considered in the normal range nonetheless have impaired renal function. Our findings were similar when the Cockcroft-Gault equation was alternatively used to define eGFR. Because an entry glucose level of 126 mg/dl was an exclusion criterion for the JUPITER trial, our study is limited to those without overt diabetes. Finally, all JUPITER participants had elevated hsCRP, an inflammatory biomarker associated with increased risk of cardiovascular events and diabetes. Therefore it is unknown whether rosuvastatin would be effective in primary prevention among those with CKD and lower hsCRP levels.