High‐Intensity Versus Non‐High‐Intensity Statins in Patients Achieving Low‐Density Lipoprotein Cholesterol Goal After Percutaneous Coronary Intervention

The study population was selected from the Samsung Medical Center (Seoul, Korea) PCI registry. Between February 2003 and December 2014, 8148 patients underwent PCI with drug‐eluting stents. Patients were ineligible for the study if there were no data on blood cholesterol level or medication information at discharge. Exclusion criteria were: (1) <20 years of age; (2) on maintenance hemodialysis; (3) had a history of liver cirrhosis; (4) received treatment for human immunodeficiency virus infection; (5) had undergone any organ transplantation; (6) had cardiopulmonary resuscitation or mechanical circulatory support during PCI; or (7) taking statin and ezetimibe combination therapy. After exclusion of 4013 patients, whether treatment target for LDL‐C was achieved or not during follow‐up after index PCI was reviewed in 4135 patients. We calculated time‐averaged LDL‐C using follow‐up LDL‐C levels measured from 4 weeks to 3 years after index PCI and regarded target for LDL‐C being achieved when time‐averaged LDL‐C was <1.8 mmol/L (70 mg/dL) or reduced at least 50% from baseline during follow‐up after index PCI based on the ESC/EAS guideline.⁵ Time‐averaged LDL‐C was calculated by the formula:where Li=consecutively measured follow‐up LDL‐C level, Ti=time period of measurement between Li‐1 and Li, and T1=time period between index PCI and L1.

Finally, 1746 patients who achieved target for time‐averaged LDL‐C during follow‐up were included in this study. We classified patients into 2 groups according to the intensity of statin prescribed at hospital discharge from admission for the index PCI: high‐intensity statin group (n=372) and non‐high‐intensity statin group (n=1374). Definition of statin intensity was based on the guideline from ACC/AHA.³ Atorvastatin 40 or 80 mg and rosuvastatin 20 mg were defined as high‐intensity statins, and the other statins were classified as non‐high‐intensity statins. (Details about statins used in the non‐high‐intensity statin group are shown in Table S1.) Rosuvastatin 40 mg has not been approved in Korea. The exclusion and division of the study is shown in a flow diagram in Figure 1. The Institutional Review Board of Samsung Medical Center approved this study and waived the requirement for written informed consent. The data, analytical methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.

Baseline characteristics and clinical outcome data were prospectively collected in our PCI registry by trained research coordinators using a standardized case report form and protocol. Patients were routinely followed up at 1, 6, and 12 months after the index procedure and annually thereafter. Further information was collected by telephone contact or medical records, if necessary. Data on statin intensity were collected from the electronic prescribing system of Samsung Medical Center. To assess whether the intensity of statin was maintained during follow‐up, information on medications that had been prescribed during the 3 years after index PCI were collected. We also collected high‐sensitivity C‐reactive protein (hs‐CRP) level measured during the 3 years after index PCI. Time‐averaged hs‐CRP was calculated by the same formula used in calculating time‐averaged LDL‐C. Follow‐up was considered complete if mortality was confirmed from the National Population Registry of the Korea National Statistical Office using a unique personal identification number or if the patient was contacted at the planned follow‐up interval.

The primary outcome was the occurrence of major adverse cardiovascular events (MACEs) during follow‐up, defined as a composite of cardiac death, myocardial infarction (MI), or stroke. Secondary outcomes included all‐cause death, target lesion revascularization, target vessel revascularization, and individual components of the primary outcome. All deaths were considered to be cardiac death unless a definite noncardiac cause could be established.⁷ MI was defined as elevated cardiac enzymes (troponin or myocardial band fraction of creatinine kinase) greater than the upper‐normal limit with ischemic symptoms or electrocardiography findings indicative of ischemia or MI at readmission requiring subsequent hospitalization (defined as emergency admission with principal diagnosis of MI).⁸ Stroke was defined as a neurological deficit attributed to an acute focal injury of the central nervous system by a vascular cause, including cerebral infarction, intracerebral hemorrhage, and subarachnoid hemorrhage.¹⁰ Target lesion revascularization was defined as repeat PCI of the lesion within 5 mm of stent deployment or bypass graft surgery of the target vessel. Target vessel revascularization was repeat revascularization of the target vessel by PCI or bypass graft surgery.⁷ All end points in this study were censored at 5 years after index PCI.

Categorical variables are presented as numbers of events and percentages and were compared using the chi‐square test or Fisher's exact test (for sparse data). Continuous variables are presented as mean±SD and were compared using the t test or Wilcoxon rank‐sum test. We assessed survival curves using the Kaplan–Meier method, along with a log‐rank test. We estimated the hazard ratios (HRs) of high versus non‐high‐intensity statin group in both univariate and multiple Cox regression models. In multiple Cox regression, variables that appeared to be related in the univariate analysis with a P<0.2 were considered in a step‐wise method to select predictors of outcomes. HRs of high versus non‐high‐intensity statin group were reported with their 95% confidence intervals (CIs).

In addition, to adjust the 2 groups for any inherent imbalance in their demographic and clinical characteristics, we advocated the propensity score method and balanced the 2 groups so that they were comparable with no undue influences from confounding factors. Propensity scores, which are the probability of being treated with a high‐intensity statin for each patient, were estimated using predicted probabilities from a multiple logistic regression analysis. Typically a full nonparsimonious model was fit using all variables in Table 1 (except aspirin and P2Y₁₂ inhibitors) and baseline lipid profile in Table 2 to predict the probability to be treated with a high‐intensity statin. The variable ratio, parallel, and pair‐wise nearest neighbor matching method created a matched data set to avoid a significant data loss from unmatched patients. A standardized mean difference of less than 10% for each matching covariate suggested an appropriate balance between the groups, as did a variance ratio near 1.0 between the 2 groups. To confirm the balance, McNemar's or Bowker's test of symmetry were used to compare categorical variables, whereas continuous variables were compared using a paired t test. Using the matched data set, the risk of outcomes was assessed using a conditional Cox regression model to obtain another set of HRs. For all analyses, we used SAS (version 9.2; SAS Institute Inc., Cary, NC) and R software (version 3.3; “MatchIt” and “survival” packages; R Foundation for Statistical Computing, Vienna, Austria) for Windows for statistical analyses.

Information about statin intensity during follow‐up is shown in Table 3 and Table S2. In the high‐intensity statin group, adherence to high‐intensity statin was 98.5% at 2 years and 87.3% at 3 years after the index PCI. In the non‐high‐intensity statin group, adherence to non‐high‐intensity statin was 99.2% at 2 years and 94.8% at 3 years after the index PCI. At the 3 years of follow‐up, adherence in the high‐intensity statin group was significantly lower than in the non‐high‐intensity statin group (87.3% versus 94.8%; P<0.01).

To determine whether the treatment benefits of high‐intensity statin observed in the overall population were consistent, we calculated the unadjusted HR for the MACEs in various subgroups (Figure 3). Benefits of high‐intensity statin were consistent, and there was no significant interaction between statin intensity and primary outcome in any subgroups.

Statin therapy should be initiated as early as possible to all patients who undergo coronary revascularization for coronary artery disease, regardless of baseline serum cholesterol level.¹¹ However, 2 major guidelines from the ACC/AHA and ESC/EAS recommend different lipid‐lowering strategies for secondary prevention in patients who undergo PCI.³ Whereas the ESC/EAS guideline focuses on decreasing LDL‐C to specific treatment target, ACC/AHA recommends the treatment using evidence‐based intensity statin therapy without specific cholesterol target.¹⁴ Such a distinction between the 2 guidelines has led to confusion in the clinical setting.⁶ During the study period, a high‐intensity statin was prescribed in 14.2% (774 of 5452) of all patients with available information on statin type and dose in our PCI registry. Although the rate of prescription of a high‐intensity statin increased to 41.9% (137 of 327) after the ACC/AHA published the guideline on the treatment of blood cholesterol in November 2013, non‐high‐intensity statins have still been used frequently for secondary prevention in patients undergoing PCI. These observations imply a lack of consensus in the treatment of LDL‐C for secondary prevention. In a meta‐analysis by the Cholesterol Treatment Trialists’ Collabolation, each 1‐mmol/L (approximately 40 mg/dL) reduction in LDL‐C reduced the risk of major vascular events by around 22%, regardless of statin type and dose.⁴ Moreover, recent trials demonstrated that nonstatin lipid‐lowering drugs, such as ezetimibe and cholesteryl ester transfer protein inhibitor, when added to statin, can further decrease LDL‐C and improve outcomes.¹⁵ These findings support LDL‐C target‐based strategy. However, several landmark studies on statin therapy for secondary prevention compared high‐intensity statin versus moderate‐intensity statin, not target LDL‐C, and demonstrated that high‐intensity statin was superior to moderate‐intensity statin for reducing the incidence of MACEs.¹ So far, there are limited data on comparison between statin intensity‐based strategy versus LDL‐C target‐based strategy, especially in patients with similar LDL‐C levels. Therefore, we compared long‐term clinical outcomes between the high‐intensity statin group and the non‐high‐intensity statin group among patients who achieved LDL‐C target after PCI.

In the present study, high‐intensity statin therapy was more effective in preventing MACE than non‐high‐intensity statin therapy in patients achieving treatment target for LDL‐C after PCI. There are several plausible explanations for our results. First, level of follow‐up time‐averaged LDL‐C was lower in the high‐intensity statin group than in the non‐high‐intensity statin group. As mentioned above, greater reduction in LDL‐C resulted in greater reduction in risk of MACE in meta‐analyses and randomized controlled trials. Second, percent LDL‐C reduction was greater in the high‐intensity statin group than in the non‐high‐intensity statin group. In a pooled individual patient‐level analysis of 3 large statin trials, Bangalore et al have reported that, among patients with attained LDL‐C ≤70 mg/dL, those with percent LDL‐C reduction of <50% had a significantly higher risk of cardiovascular event when compared with the group with percent LDL‐C reduction of ≥50%.¹⁷ Greater percent LDL‐C reduction in the high‐intensity statin group, compared with the non‐high‐intensity statin group, might have partly explain difference in the risk of MACEs. Third, however, observed benefit in the high‐intensity statin group was greater than expected benefit from the difference in time‐averaged LDL‐C between the 2 groups, when compared with previous data from meta‐analyses.⁴ This finding suggests that observed benefit in the high‐intensity statin group could not be explained by only the lipid‐lowering effect of statins and might be, at least partly, attributable to a non‐lipid‐mediated mechanism or pleotropic effects of statins. Statins have been reported to have diverse protective effects on the cardiovascular system: improvement of endothelial dysfunction¹⁸; modulation of inflammatory response and thrombogenesis¹⁹; and stabilization of plaque.²⁰ In particular, several trials demonstrated that pleotropic effects of statins are more potent in high‐intensity statins than in non‐high‐intensity statin.²¹ In the present study, we investigated follow‐up hs‐CRP to compare anti‐inflammatory effects between the 2 groups. Follow‐up hs‐CRP was significantly lower and reduction from baseline hs‐CRP was significantly greater in the high‐intensity statin group than in the non‐high‐intensity statin group, suggesting benefit from non‐lipid‐lowering effects of high‐intensity statins. Additionally, to find an explanation for the extreme differences between the 2 groups, we compared the clinical outcomes between the high‐intensity statin group (n=618) and the non‐high‐intensity statin group (n=3517) in all reviewed patients (n=4135), including the patients who failed to achieve treatment target for LDL‐C. Unlike the results of the patients who achieved LDL‐C goal, incidence of the primary outcome was not significantly different between the 2 groups (7.3% versus 10.0%; adjusted HR, 0.80; 95% CI, 0.55–1.16; P=0.24). Although it is difficult for us to explain the exact causes, the relatively small sample size, especially for the high‐intensity statin group, and nonrandomized nature of our study may have resulted in the extreme differences between the 2 groups.

Our study had several limitations. First, the study was a nonrandomized, observational study. Statin intensity was determined at the discretion of the attending physician and might have been influenced by several factors such as underlying demographics, clinical presentation at admission, baseline lipid values, and physician's preference. Although we performed propensity‐score–matched analysis and adjustments to overcome the potential bias that can influence the study outcome, unmeasured factors might have affected study outcomes. Second, 3622 patients were excluded because of lack of information on statin prescription and LDL‐C during follow‐up, although 8148 patients were screened at first. For this reason, a selection bias could have influenced the study results. Third, patients’ compliance to statin was not accurately evaluated. To evaluate the compliance indirectly, we used information of medication prescribed during follow‐up after PCI. As shown in Table 3, statin intensity at discharge after PCI was well maintained during follow‐up. Moreover, because we included only patients who achieved treatment target for LDL‐C, we believe most patients might have a good compliance to statins. The main cause of sudden drop of adherence to high‐intensity statin at the 3 years of follow‐up might be that some physicians reduced the intensity of statin from high to moderate because LDL‐C level had been well maintained under the treatment target during follow‐up. Fourth, hs‐CRP data were only available for 69.6% of the total population (1216 of 1746). Fifth, the median follow‐up duration of the high‐intensity statin group was shorter than that of the non‐high‐intensity statin group (3.1 [1.4–4.9] versus 4.5 [2.5–5.0] years). To assess whether difference in follow‐up duration influenced the study outcome, we truncated follow‐up period to 3 years and re‐evaluated the clinical outcome. Incidence of primary outcome was significantly lower in the high‐intensity statin group than that of the non‐high‐intensity statin group (2.4% versus 6.1%; adjusted HR, 0.41; 95% CI, 0.20–0.86; P=0.02). An adjusted HR of the primary outcome obtained from 3‐year follow‐up data was similar to that using 5‐year follow‐up data. Based on this similarity, we could estimate that the difference in follow‐up duration between the 2 groups did not significantly affect the clinical outcome. Last, MI and stroke might be relatively under‐reported given that this was not a randomized, controlled trial with rigorous follow‐up. However, ratios of MI and stroke to cardiac death or all‐cause death were comparable to, or higher than, those of the randomized, controlled trials²⁴ conducted in Korea. Moreover, it was highly unlikely that MI and stroke were selectively under‐reported in the high‐intensity statin group than in the non‐high‐intensity statin group.