JUPITER: major implications for vascular risk assessment

ABSTRACT

This Editorial comments on the recently published JUPITER trial (Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin), the further evidence it provides for supporting the role of statins in primary prevention and the major implications this may hold for vascular risk assessment and clinical practice guidelines.

Key words: :

• HsCRP
• Primary prevention
• Statins
• Vascular risk assessment

There is convincing evidence that statins reduce vascular morbidity and mortality in patients with established vascular disease or diabetes mellitus (DM)1,2. However, the evidence regarding the efficacy of statins in the primary prevention setting is more limited3–7. In subjects without established vascular disease, lovastatin and atorvastatin significantly reduced coronary heart disease (CHD) morbidity3,4. Pravastatin significantly reduced CHD morbidity in one study7 and both CHD morbidity and mortality in another5, but did not improve outcomes in the Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)6. The recently published Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) further supports the role of statins in primary prevention8.

JUPITER included 17 802 subjects (11 001 men older than 50 years of age and 6801 women older than 60 years of age) with low density lipoprotein cholesterol (LDL-C) levels < 130 mg/dl [3.4 mmol/l; median levels 108 mg/dl (2.8 mmol/l)], high sensitivity C-reactive protein (hsCRP) levels > 2 mg/l and without established vascular disease8. Subjects were randomly allocated to rosuvastatin 20 mg/day or placebo. The trial was stopped early after a median follow up of 1.9 years because of a significant reduction in the primary endpoint (myocardial infarction, stroke, arterial revascularization, hospitalization for unstable angina or cardiovascular mortality) by 44% [hazard ratio (HR) 0.56, 95% confidence interval (CI) 0.46–0.69; p < 0.00001]8. There was also a significant reduction in non-fatal and any myocardial infarction, non-fatal and any stroke, as well as in arterial revascularization (all, p ≤ 0.003). JUPITER is also the first statin trial to show a reduction in all-cause mortality (p = 0.02) in subjects without vascular disease or DM. Moreover, JUPITER showed a significant reduction in vascular events in women without established vascular disease. Previous primary prevention studies either did not enrol women5 or were not powered to show risk reduction in women3,4,7.

The findings of JUPITER might mandate a modification of the existing guidelines for lipid-lowering treatment in subjects without CHD or CHD equivalents. According to current guidelines, drug therapy is optional when LDL-C levels are between 100 and 130 mg/dl (2.6–3.4 mmol/l) and 10-year risk is 10–20%9. In JUPITER, rosuvastatin equally reduced the primary endpoint regardless of the baseline Framingham score (≤ or > 10%)8. These findings suggest that men older than 50 years and women older than 60 years with LDL-C levels < 130 mg/dl (3.4 mmol/l) but with hsCRP levels > 2 mg/l (i.e., the JUPITER population) could benefit from statin treatment regardless of the Framingham calculated score.

The JUPITER trial also has implications regarding the role of hsCRP testing in risk stratification. In prospective studies, hsCRP levels were independently associated with increased vascular morbidity and mortality10–12. Previous studies showed that considering hsCRP levels improves the predictive accuracy of the Framingham risk engine11,13. In JUPITER, rosuvastatin improved the outcome (HR 0.63, 95% CI 0.44–0.92; p = 0.01) of patients with hsCRP levels > 2 mg/l and no other risk factors except increased age8. Current guidelines state that measuring hsCRP levels in patients with estimated 10-year risk between 10 and 20% may help risk evaluation14. However, the benefits of treatment based on this strategy were uncertain until now14. The results of JUPITER suggest that hsCRP testing might be a useful tool in selecting subjects without established vascular disease who will benefit from statin treatment. However, some limitations of hsCRP testing should be mentioned. Several hsCRP assays are currently in use14. Moreover, hsCRP testing is not widely available and is costly. In JUPITER, 36.1% of the screened population did not have hsCRP levels > 2 mg/l8. Moreover, 52.2% of the screened population had LDL-C levels > 130 mg/dl (3.4 mmol/l) and the hsCRP levels in these patients were not reported8. In the Framingham study, mean hsCRP levels were 2.67 and 2.23 mg/l in men and women, respectively15. Therefore, a significant number of subjects will need to be tested in order to identify those with elevated hsCRP levels. In addition, due to the intraindividual variability of hsCRP measurements, it is recommended that hsCRP levels should be measured twice, at least 2 weeks apart14. These limitations will further increase the direct and indirect cost of hsCRP testing.

In JUPITER, LDL-C and hsCRP levels at 12 months were lower (50 and 37%, respectively) in the rosuvastatin group than in the placebo group8. These differences were sustained during the study8. It is not clear whether rosuvastatin reduced risk due to a fall in LDL-C levels, hsCRP levels or other actions. Previous statin trials in patients with established CHD suggested that a decrease in hsCRP levels is associated with a delay in the progression of coronary atherosclerosis16 and a reduced risk for vascular events17. This decrease in risk was independent of LDL-C lowering16,17. It is also not clear whether hsCRP plays a direct role in atherogenesis or is just a marker of the inflammatory process18–20. Polymorphisms in the CRP-encoding gene that result in increased hsCRP levels were associated with increased vascular risk in some studies21 but not in others22–24.

Besides LDL-C and hsCRP lowering, other actions of rosuvastatin might have contributed to the improved outcome in JUPITER. Triglyceride (TG) levels decreased from 118 to 99 mg/dl (1.3–1.1 mmol/l) at 12 months in the rosuvastatin group (17% lower than TG levels in the placebo group; p < 0.001) and this effect persisted at 48 months8. Elevated TG levels appear to be associated with increased vascular risk25. In some statin trials in patients with CHD, the reduction in TG levels correlated with a lower event rate26–28. However, this association was not observed in other secondary prevention studies or in subjects without vascular disease29–33. Rosuvastatin did not modify HDL-C levels significantly; this might be due to the ‘satisfactory’ baseline HDL-C levels [49 mg/dl (1.3 mmol/l)]. It would therefore be of interest to report if there was a rosuvastatin-associated increase in HDL-C levels in the participants of JUPITER who had low baseline values of this protective lipoprotein. In other statin trials, the HDL-C levels were relevant to outcome34. Interestingly, there was also a very small but significant difference in estimated glomerular filtration rate (eGFR) at 12 months in the rosuvastatin group (66.8 vs. 66.6 ml/min/1.73 m² in the placebo group; p = 0.02)8. Previous statin trials in high risk patients with or without vascular disease reported similar findings35–41. This preservation of renal function might play a role in vascular risk reduction during statin treatment38,39 and might have contributed to the reduced risk of events in JUPITER. Statins also reduced vascular risk in patients with chronic kidney disease 4, 42–45. Patients with serum creatinine levels > 2.0 mg/dl (176.8 μmol/l) were excluded from JUPITER but it would be interesting to evaluate the effects of rosuvastatin on renal function and vascular events in a subgroup of patients with the highest baseline creatinine levels (or lowest eGFR). This subgroup might show greater changes in renal function39.

There was a significantly increased risk for developing type 2 DM in the rosuvastatin group (270 vs. 216 reports; p = 0.01)8. This difference was observed within 24 months8. Median glycated haemoglobin (HbA_1c) levels were also marginally but significantly higher at 24 months in patients allocated to rosuvastatin (5.9 vs. 5.8%; p = 0.001)8. However, fasting blood glucose levels [98 mg/dl (5.4 mmol/l) in both groups] at 24 months and the risk of newly diagnosed glycosuria at 12 months were similar in the two groups8. Type 2 DM in JUPITER was physician-reported and was not adjudicated by the endpoint committee8. In the Controlled Rosuvastatin Multinational Trial in Heart Failure (CORONA; 5011 patients with heart failure treated with rosuvastatin 10 mg/day or placebo for 32.8 months), there was no significant increase in newly diagnosed type 2 DM (100 vs. 88 cases, respectively; p = 0.40)46. Other long-term studies of rosuvastatin did not specifically report the effects of rosuvastatin on glucose levels or the incidence of newly diagnosed type 2 DM47–49. However, this information must be available because glucose was measured. Similar trends were observed in other statin trials as discussed by the authors of the JUPITER study8. In other rosuvastatin trials of shorter duration and smaller size, there was no change in glucose levels or in insulin resistance (expressed as homeostasis model assessment index)50,51.

In JUPITER, rates of myopathy, elevated alanine aminotransferase levels > 3 times the upper limit of reference range and other adverse events were similar in the rosuvastatin and placebo groups8. Rosuvastatin 20 mg/day appears to lower LDL-C levels to a similar degree as atorvastatin 80 mg/day and more than simvastatin 80 mg/day52. However, both atorvastatin and simvastatin at these doses are associated with increased risk of side-effects and treatment discontinuation53–56.

In conclusion, the findings of JUPITER have important implications for clinical practice. This trial suggests that hsCRP testing is a useful tool in identifying subjects without vascular disease who should receive statins despite a low (e.g., Framingham 10-year risk ≤ 10%) or intermediate calculated risk. Existing guidelines might need to be revised. For example, the threshold for LDL-C levels for considering statin treatment in primary prevention may need to be evaluated in the light of plasma hsCRP levels. JUPITER also has major cost implications if the benefits observed can not be reproduced by ‘generic’ statins. The problem is how do you answer this question? Furthermore, can we really estimate true cost in a trial that was prematurely discontinued after only an average follow-up of 1.9 years?

Acknowledgement

Declaration of interest: This editorial was written independently; no company or institution supported it financially. Some of the authors have attended conferences, given lectures and participated in advisory boards or trials sponsored by various pharmaceutical companies (including AstraZeneca, sponsors of the JUPITER trial). No professional writer was involved in the preparation of this editorial.

Konstantinos Tziomalos is supported by a grant from the Hellenic Atherosclerosis Society.