Review of the PROVE IT-TIMI 22 Trial

Intensive versus moderate lipid lowering with statins after acute coronary syndromes

NEJM 2004;350:1495-504.

Background The benefit of cholesterol lowering with statins compared to placebo had been demonstrated in multiple trials across a range of cholesterol starting levels in patients with and without a history of coronary artery disease.

The average reduction in cholesterol from baseline levels in these trials was in the range of 25% to 35%. Guidelines at the time this trial was undertaken recommended an LDL treatment target of <100 mg/dl.

The Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis in Myocardial Infarction 22 (PROVE IT-TIMI 22) trial sought to test the hypothesis that more intensive LDL lowering with atorvastatin 80 mg daily would reduce a composite endpoint of cardiovascular events compared to less intensive LDL lowering with pravastatin 40 mg daily in patients hospitalized with ACS.

Patients Patients had to be ≥18 years of age and hospitalized for either AMI or high-risk unstable angina in the preceding 10 days. They had to be in stable condition and were to be enrolled after a PCI, if one was planned. Finally, they had to have a total cholesterol level ≤240 mg/dl. For patients already on long term lipid lowering therapy they had to have a total cholesterol level of ≤200 mg/dl. Exclusion criteria including coexisting conditions that shortened expected survival to less than 2 years, were receiving any statin at a dose of 80 mg daily at the time of index event, had undergone PCI within the previous 6 months or CABG within the previous months or were scheduled to undergo CABG in response to the index event, had serious hepatic illness or had a creatinine level >2 mg/dl.

Baseline characteristics The average age of patients was 58 years and 78% were men; 90% were white. The index event was unstable angina in 30%, NSTEMI in 36% and STEMI in 34%. Seventy percent of patients underwent PCI for the treatment of their index event. Less than 20% of patients had a prior MI, 18% had diabetes and 50% had hypertension. The median total cholesterol was 180 mg/dl, LDL was 106 mg/dl, HDL was 39 mg/dl and triglycerides were 150 mg/dl.

Procedures Eligible patients were randomly assigned to receive either atorvastatin 80 mg or pravastatin 40 mg daily in a double-blind fashion. Patients were seen for follow-up visits and received dietary counseling at 30 days, at 4 months and every 4 months thereafter until their final study visit in August or September 2003. Blood samples were obtained at randomization, at 30 days, at 4, 8, 12, and 16 months and at the final study visit for measurement of lipids and other components that were part of the safety assessment. Measurements were made at core laboratories. The dose of either study drug could be cut in half if liver enzymes or creatinine kinase was elevated (levels not specified) or in the presence of myalgias (criteria not defined).

Endpoints The primary endpoint was the time from randomization to the first occurrence of a component of the primary composite endpoint, which included: all-cause death, nonfatal MI, documented unstable angina requiring hospitalization, revascularization with PCI or CABG if either were performed 30 days after index event, and stroke. Secondary endpoints were occurrence of each component of the primary composite endpoint as well as a composite of death from coronary heart disease, nonfatal MI or revascularization occurring at least 30 days after index event and another composite consisting of death from coronary heart disease and nonfatal MI.

The study was designed to test whether pravastatin 40 was non-inferior to atorvastatin 80 with respect to the event rate for the primary composite endpoint at 2 years. To claim non-inferiority for pravastatin, the one-sided 95% confidence interval for the relative risk of the 2-year event rate in the pravastatin group could not exceed 1.17, corresponding to a hazard ratio of 1.198. Assuming a two-year event rate of 22% in the atorvastatin group and that the two treatments had equivalent efficacy, it was determined that 2000 patients per group would give the study a statistical power of 87 percent and that this power would be preserved if follow-up continued until 925 end-point events had occurred.

Basically, if the upper bound of the 95% CI showed that pravastatin 40 was not worse than atorvastatin 80 by more than 17%, the investigators could claim it was noninferior.

Results 4,162 patients were included in the final analysis; 2,063 in the pravastatin group and 2,099 in the atorvastatin group. Average follow-up time was 2 years. Compared to atorvastatin, pravastatin did not meet the prespecified criteria for noninferiority set by the investigators. The rates of the primary composite endpoint in the pravastatin and atorvastatin groups were 26.3% and 22.4%, respectively and the upper bound of the 95% CI was 25% (it had to be 17% or less to claim noninferiority). The investigators performed further analyses which demonstrated the superiority of atorvastatin compared to pravastatin; however, the presentation of these results in the main manuscript are not in a traditional format due to the prespecified hypothesis that was tested.

The only individual component of the primary composite endpoint that was significantly reduced at 2 years by atorvastatin was revascularization. The rates of death in the atorvastatin and pravastatin groups were 2.2% and 3.2%, respectively and the difference was not statistically significant. Post-hoc power for this endpoint was only 51% at an alpha of 0.05.

The authors report that subgroup data showed consistent effects; however, inspection of the forest plot reveals the possibility of treatment effect heterogeneity with no benefit for atorvastatin 80 compared to pravastatin 40 for patients ≥65 years of age, those taking a statin at baseline, those with baseline LDL <125 mg/dl, and those whose index event was a STEMI.   

The rates of treatment discontinuation in the pravastatin and atorvastatin groups were 33% and 30%, respectively at 2 years. Medication was discontinued due to elevated LFT’s in 1.1% of patients on pravastatin compared to 3.3% of patients on atorvastatin (p<0.001). It was discontinued for myalgias or elevated creatine kinase in 2.7% of pravastatin group compared to 3.3% of atorvastatin group (p=NS). There were no documented cases of rhabdomyolysis in either group.

At the time of randomization, a median of seven days after the onset of the index event, the median LDL cholesterol levels were 106 mg/dl before treatment in each group. The LDL cholesterol levels achieved during follow-up were 95 mg/dl in the pravastatin group and 62 mg/dl in the atorvastatin group (P<0.001).

Conclusions In patients with ACS, pravastatin 40 mg did not meet non-inferiority compared to atorvastatin 80 mg. The difference in a composite endpoint that included 5 individual components was statistically lower in the atorvastatin group with a number needed to treat of 25 patients.

The achieved difference in LDL lowering was significantly greater in the atorvastatin group. Notably, the only individual component of the primary endpoint that was significantly reduced was revascularization.

Inspection of various subgroups suggests the possibility of treatment effect heterogeneity for several important subgroups including patients ≥65 years of age.

The trial helped establish lower LDL targets for secondary prevention but also highlights the diminishing gains of such therapy.