Benefit of Statin Therapy with Percutaneous Coronary Intervention
Abstract & Commentary
This is an observational report of a large number of patients undergoing percutaneous coronary intervention (PCI) at the Cleveland Clinic between 1993-1999, assessing whether subjects on a statin at the time of PCI had an improved clinical course over the next 6 months. A total of 5052 patients were included in the final analysis; 26.5% or 1337 were on a statin prior to their PCI. Information regarding subsequent statin administration in hospital or doing follow-up was not available and was not analyzed. Elaborate statistical techniques were used, including a propensity analysis to attempt to ascertain which clinical features are more likely to be associated with statin use. The first 1000 patients per year for 7 years who underwent PCI at the Cleveland Clinic were entered into a registry. Clinical events were analyzed during hospitalization at 30 days and 6 months. Patients were divided into 2 groups according to whether they were on a statin before the PCI. The statin-treated patients were substantially higher risk at baseline and had a greater burden of vascular disease; they were more likely to have diabetes, high cholesterol, hypertension, triple vessel disease, prior MI, and prior bypass surgery. ACE inhibitor and B-blocker use was greater in these individuals who underwent more complex PCI and had a greater likelihood of receiving stents or 2B3A inhibitors. Procedures success, ejection fraction, or unstable angina rates were comparable.
Results: There was a lower all-cause mortality at 30 days and 6 months in the individuals who were on a statin at the time of PCI. Early mortality at 30 days was 0.8% vs. 1.5%; hazard ratio, 0.53; P = 0.048. This mortality benefit was sustained at 6 months, 2.4% vs. 3.6%; hazard ratio, 0.67; P = 0.046. There were no differences in nonfatal MI or repeat revascularization during follow-up. The propensity analysis showed that within each decile of the population in the propensity score, there were no differences between statin- and nonstatin-treated groups. Using the COX proportional hazards model for all potential variables, the adjusted analysis still found that statin therapy was "an independent predictor for the survival at 6 months (hazard ratio, 0.65; P = 0.045)." Other predictors of mortality included advanced age, renal insufficiency, low body weight, low ejection fraction, and a history of PVD. B-blocker use and successful revascularization were associated with survival benefit. Chan and colleagues conclude that on the basis of the 7-year registry data, statin therapy initiated prior to PCI "seems to be associated with an early and sustained mortality reduction period." This benefit appeared by 30 days. They suggest the data are concordant with a prior report from the Cleveland Clinic relating to acute coronary syndrome (ACS) patients discharged on a statin (vs those who were not), which also demonstrated a lower 6-month mortality. Comparable findings were noted in a Swedish postinfarction observational report. Chan et al discuss 3 recent restenosis trials with statins, which in general have been negative; they stress that a fluvastatin trial (FLARE) found a lower rate of death or MI in patients pretreated with fluvastatin for 2-4 weeks prior to PCI. They suggest the possibility that nonlipid-lowering effects of statins may have contributed to the lower mortality in their own observational database, as well as in FLARE, while not necessarily decreasing the current rates or restenosis. They further suggest that anti-inflammatory effects of the HMG COA reductase inhibitors may be important. They subsequently point out the potential limitations of the trial, including its nonrandomized nature, the higher risk status of the PCI patients, and the differences in subsequent treatment between the statin and nonstatin subjects. The status of statin treatment subsequent to the PCI is not known: Chan et al believe that those individuals who were given a statin after PCI would dilute the positive results of the study. They conclude that the data suggest "the need to evaluate prospectively the impact of pretreatment of statins within a randomized trial design, in particular targeting patients with elevated inflammatory markers before coronary intervention" (Chan AW, et al. Circulation. 2002;105:691-696).
Comment by Jonathan Abrams, MD
The data by Chan et al are provocative and hypothesis generating, and concordant with several other recently published data sets indicating that patients who leave the hospital after ACS or MI on a statin have improved 6- and 12-month outcomes. Chan et al believe the results of MIRACL as being concordant with these observations. However, some individuals, including myself, were disappointed with that study, which showed no hard end point differences between those individuals discharged on a high dose of atorva-statin following ACS vs. those randomized to placebo. Re-hospitalization for myocardial ischemia was the sole benefit noted in that study; mortality or MI was not reduced. Nevertheless, Chan et al cite MIRACL as being consistent with the possibility that statins may induce a reduction in mortality independent from their lipid-lowering effects. They use the mortality data from the FLARE trial to support this. This is somewhat of a leap of faith, representing an interesting although somewhat questionable hypothesis relating to statin pleitropism. The mortality difference at 30 days and 6 months in this study is of interest and could be related to statin use, especially since the statin-treated cohort appeared to be at a significantly higher risk. They received more pharmacologic therapy and revascularizations. This and other unknown factors could contribute to the significant differences in survival. Furthermore, patients enrolled in this trial were not randomized, as only the first 1000 patients of each of 7 consecutive years were entered into the registry. This approach is somewhat arbitrary and is a design problem. The 2 groups are quite dissimilar with respect to clinical parameters. Data are not provided to as to how the percentage of overall statin use changed between 1993-1999. One would assume that in the last several years of the observational period, statin use went up considerably. The duration of statin use and the "drop in" rate after PCI are unknown. These data do support the current belief of the prevention community to ensure that a statin is provided to all patients hospitalized with acute MI or ACS. Furthermore, the American Heart Association’s "Get With the Guidelines" program emphasizes the administration of a statin to all such patients. The golden window of opportunity enhancing compliance is frequently cited. Nevertheless, the data supporting this recommendation are somewhat, but not fully, persuasive, yet it appears that the policy of a statin for all hospitalized with chest pain, with or without a PCI, has become the law of the land. A large RCT is needed. Finally, one must keep in mind the possible parallel to the hormone replacement story, where much observational data appeared to confirm a major benefit of estrogens in preventing CAD and/or decreasing its manifestations. The data supporting the benefit for a statin in the first year following ACS or after MI remain limited to observational studies, except in the 16-week MIRACL study. The event curves of 4S, CARE, and LIPID, indicate a prolonged period of time before the placebo and statin cohorts begin to separate; while these studies are in more stable individuals, we should remain somewhat cautious before completely accepting the recommendations coming out of the observational statin trials.
Dr Abrams is Professor of Medicine, Division of Cardiology, University of New Mexico, Alburquerque, NM.