Sudden cardiac death (SCD) accounts for more than 50% of the deaths due to heart disease (1), with the vast majority from ventricular tachyarrhythmias. Statin-type hypolipidemic drugs have been shown to reduce mortality and cardiac death in patients with coronary artery disease (CAD) (2- 3). These drugs also reduce the incidence of first acute major coronary events in subjects without overt CAD who have average levels of total cholesterol and low-density lipoprotein cholesterol (4). Recent data suggest that lipid-lowering drugs may have anti-arrhythmic effects (5- 6).

The Multicenter Automatic Defibrillator Implantation Trial (MADIT)-II database includes information regarding ventricular arrhythmia (VA) occurrence obtained from ICD interrogation as well as detailed drug information during follow-up (7- 8). The purpose of this study was to investigate whether statin use is associated with reduced VAs in patients receiving an ICD in the MADIT-II study.

The design, inclusion and exclusion criteria, and primary results of the MADIT-II study have been previously described (7- 8). Briefly, 1,232 patients who were ≥21 years old, had a myocardial infarction one month or more before entry, and an ejection fraction ≤0.30 were randomized to ICD implant plus conventional medical therapy or conventional medical therapy alone in a 3:2 ratio.

Of 742 patients randomized to defibrillator therapy, 720 actually received an ICD. The study population consisted of 654 ICD-treated patients in whom detailed information of drug therapy for every day in the trial was available. Lipid-lowering therapy (drug and dose) was prescribed at the discretion of the treating physician. The statins used were atorvastatin, simvastatin, pravastatin, lovastatin, cerivastatin, and fluvastatin.

During the three-month follow-up visits, the dates of initiation and cessation of medications were recorded and ICD devices were interrogated. The average patient follow-up was 17 months. The distribution of the percent of statin use during follow-up (Figure 1) was highly concentrated at both ends of the range. Thus, the cut points of ≥90% and ≤10% statin use were arbitrarily selected, and three groups of patients were identified: 1) patients who used statins for ≥90% of the time (n = 386); 2) patients who used statins 11% to 89% of the time (n = 116); and 3) patients who used statins ≤10% of the time (n = 152). As there were six statins used in this trial at varying dosages, no attempt was made to compare one statin and/or one dose to another.

The prespecified primary end point in this substudy was the first occurrence of cardiac death or appropriate ICD therapy (anti-tachycardia pacing or shock) for ventricular tachycardia (VT)/ventricular fibrillation (VF). The secondary end point was first appropriate ICD therapy for VT/VF, with censoring of death when it occurred. Tertiary end points included sudden and non-sudden cardiac death.

A core lab reviewed all stored electrocardiograms and determined the frequency of appropriate ICD therapy for treatment of VT or VF on the basis of the interrogation findings. Details of the ICD devices, their programming, and their interrogation were recently reported (9). An end point committee reviewed all mortality events, as recently reported (10).

The effect of statins on the cumulative probabilities of the end points were determined by the Kaplan-Meier method, with significance testing by the log-rank statistic. Proportional hazards regression analysis was used to evaluate the effect of statin use (yes/no) on the probability of ICD therapy for the combined end point VT/VF or cardiac death and for the end point VT/VF (11). Subsequently, a time-dependent proportional hazards analysis was used to evaluate the effect of statin use on these end points in a more precise manner. Statin use was treated as a time-dependent covariate, an approach that takes into account each day a patient was on or off a statin during the study, with the onset or offset of statin use on the day of change. The baseline variables that had differences between the three statin use groups (Table 1), using a p value <0.10, and variables that were thought to be of clinical relevance, were evaluated in the proportional hazards stepwise selection model. Only covariates with a p value <0.05 in the proportional-hazards model were included in the final model. All tests of statistical significance were two-sided. SAS version 9.1.3 (SAS Institute, Cary, North Carolina) was used in the data analyses.

Of the 720 patients in the ICD arm, 67% were taking lipid-lowering drugs (statins in 96%) at baseline, and of the 490 patients in the conventional therapy arm, 65% were taking lipid-lowering drugs (statins in 96%) at baseline. The baseline characteristics of the patients grouped by statin use (≤10%, 11% to 89%, ≥90% of the follow-up time in the trial) are presented in (Table 1). The group receiving statins for ≥90% of the follow-up time were younger and had lower blood urea nitrogen, less diuretic use, more coronary angioplasty, higher ejection fraction and greater beta-blocker use compared with the group receiving statins ≤10% of the follow-up time (p < 0.05).

The cumulative probabilities of time to appropriate therapy for the combined end point VT/VF or cardiac death and for just VT/VF by statin usage are presented in (Figure 2)A and (Figure 2)B, respectively. When comparing the ≥90% statin versus the ≤10% statin use groups, there was a lower cumulative probability of VT/VF or cardiac death (p < 0.01) and of VT/VF (p = 0.06). Statin use was associated with a significant reduction in SCD (p < 0.01), but no meaningful reduction in non-sudden cardiac death (p = 0.21) in the ≥ 90% statin group compared with the ≤10% statin group ((Figure 3)A and Figure 3B).

Time-dependent statin use, using a time delay of zero days, was associated with a hazard ratio of 0.65 (95% confidence interval [CI] 0.49 to 0.87; p < 0.01) for the combined end point cardiac death (n = 42) or appropriate VT/VF therapy (n = 153), whichever comes first, and 0.72 (95% CI 0.52 to 0.99; p = 0.046) for the single end point appropriate VT/VF therapy, after adjusting for relevant covariates, which were beta-blocker use, blood urea nitrogen >25 mg/dl, and New York Heart Association functional class ≥2 (Table 2). These hazard ratios translate into risk reductions of 35% for cardiac death or VT/VF and 28% for VT/VF, respectively.

Time-dependent statin use, using a time delay of four weeks, was associated with adjusted hazard ratios of 0.77 (95% CI 0.58 to 1.04; p = 0.08) and 0.79 (95% CI 0.57 to 1.11; p = 0.17) for the two specified end points, respectively.

The principal finding of this analysis is that statin use was associated with fewer VT/VF episodes. There was a 28% reduction in the risk of a first VT/VF episode with statin use.

This study adds to the findings from several other studies. De Sutter et al. (5) examined 78 patients with coronary disease and life-threatening VA treated with ICD therapy. Patients who received lipid-lowering therapy had fewer VA episodes compared with those who did not (6 of 27 or 22%, vs. 29 of 51 or 57%, p < 0.05). Mitchell et al. (6) examined VA recurrence rates in patients who had CAD and near-fatal VA and received an ICD. There was a 40% reduction in the risk of VA recurrence in the group that received lipid-lowering therapy (n = 83; 79% statins) compared with the group that did not (n = 279) (reduction in hazard 0.40, 95% CI 0.15 to 0.58), after adjusting for baseline inequities. Chiu et al. (12) studied 281 patients who had CAD and underwent ICD implantation and observed a reduction in first ICD therapy for VA (adjusted hazard ratio 0.60, p = 0.01) among patients who used statin therapy (n = 154) compared with those who did not (n = 127).

Statins have also been reported to improve regulation of coronary arterial tone and nitric oxide-mediated endothelial function, inhibit cell proliferation, stabilize atherosclerotic plaques, have anti-inflammatory properties, and provide anti-oxidant effects. The mechanism by which statins reduce VAs may relate indirectly to one or more of these effects. For example, the anti-oxidant and anti–cell-proliferative effects of statins may play a role in plaque stabilization and thus contribute to an anti-arrhythmic effect by reducing ischemia-related ventricular tachyarrhythmias. Of interest, statin therapy has been associated with improved survival in heart failure patients independent of its effect on cholesterol levels (13).

Dietary polyunsaturated fatty acids have been reported to favorably alter the structure of the phospholipid part of the cardiac cell membrane and provide immediate anti-arrhythmic effects in lab experiments (14- 16). We speculate that statins may similarly alter the lipid portions of the membrane through which the transmembrane segments of ion channels penetrate, thereby affecting ion-channel conductances. Statins may also have some membrane-stabilizing properties. The exact mechanism by which statins may reduce arrhythmias is unknown.

Several lines of evidence from our study favor the hypothesis that statins have anti-arrhythmic properties. First, a dose-response effect was present. Second, the time-dependent analysis revealed a greater decrease in VA in patients taking statin therapy compared with those not taking a statin, using a time delay of zero days rather than four weeks. This suggests an immediate effect of these drugs rather than a delayed effect related to a slowing of the rate of progression of atherosclerosis. Third, the findings have biologic plausibility related to potential lipid-altering effects in the cardiac cell membrane.

This was an observational study and statin therapy was not randomized. The higher statin-use subjects may have been less sick than the lower statin-use subjects, and there may have been incomplete statistical adjustment for observed imbalances. In addition, there may be imbalance in other relevant variables that were not measured.

In the MADIT-II study, statins were associated with a reduced risk of cardiac death or VT/VF and with reduced VT/VF episodes. These findings suggest that statins have anti-arrhythmic properties.