Lower mortality following coronary arterial revascularization in patients taking statins

Abstract

Objectives. Numerous studies have suggested that statins have beneficial non-lipid-lowering effects, including reduction of systemic inflammatory response following surgery. We wanted to evaluate the effect of preoperative statin treatment on complications and operative mortality after coronary arterial revascularization. Design. We performed a retrospective study of 720 consecutive patients who underwent on-pump coronary artery bypass grafting (CABG) (n = 513) or off-pump (OPCAB) (n = 207) in Iceland from 2002–2006. Patients taking statins preoperatively (n = 529) were compared with those not taking statins (n = 191). Predictors of complications and operative mortality were evaluated by univariate and multivariate analysis. Results. Cardiovascular risk profiles were similar. However, hypertension was more common in the statin group but EuroSCORE was slightly lower. Operative mortality was significantly lower in patients taking statins (1.7% vs. 5.8%, p < 0.001). There were no significant differences in the incidence of major complications. Multivariate analysis showed that preoperative statin treatment was an independent predictor of lower operative mortality (OR = 0.33, p = 0.043), even after adjusting for EuroSCORE, acute operations, advanced age, or other medications. Conclusions. In this non- randomized study, patients taking statins had lower operative mortality than the controls after adjusting for multiple confounders. The reason for this might be linked to pleiotropic effects of statins.

Key words::

• statins
• coronary artery bypass grafting
• mortality

Introduction

Reducing the levels of low-density lipoprotein (LDL) is one of the cornerstones of lowering the mortality and morbidity of patients with ischemic heart disease (IHD) (1,2). Lipid-lowering statins were approved for clinical use in 1987, and they have become an inherent part of the primary and secondary prevention of cardiovascular disease (2). Furthermore, statins have been shown to have numerous pleiotropic effects such as reducing the systemic inflammatory response (3,4) and reducing endothelial dysfunction (5), in addition to preventing plaque rupture, thrombus formation and coronary artery occlusion (6).

Coronary artery revascularization is an effective treatment for IHD, and remains the therapy of choice for advanced coronary artery disease. In spite of improvements in surgical technique and pre- and postoperative care, the morbidity and mortality in patients following CABG or OPCAB is substantial. With an aging population, the number of high-risk procedures will probably increase (7). To improve outcomes in cardiac surgery, it is important to optimize strategies to limit postoperative complications. There is accumulating evidence for the benefits of preoperative statin therapy in patients undergoing myocardial revascularization, and current guidelines recommend statins for these patients (1). In a meta-analysis involving more than 30,000 patients, Liakopoulos et al. found that statins significantly reduced early mortality in patients undergoing CABG (8). Other studies have shown reduced operative and medium-term mortality following cardiac surgery in patients taking statins, regardless of their baseline lipid status (9). In a recent randomized trial, Mannacio et al. showed that normolipemic patients taking statins had less operative myocardial damage (10). However, the randomized ARMYDA-3 trial did not find any effects of statins on postoperative mortality (11). As each study arm only included 100 patients, the study was probably underpowered to detect differences in rare events such as operative mortality (12). A recent meta-analysis from the Cochrane Central Register, that included all randomized studies (a total of 11 studies, n = 984) on statin treatment in relation to surgical coronary arterial revascularization, showed no impact on short-term mortality or post-operative stroke. However, preoperative statin therapy reduced the incidence of post-operative atrial fibrillation and length of stay was shorter, both regarding ICU and total hospital stay. Furthermore, a trend toward lower incidences of perioperative myocardial infarction and renal failure was seen in the statin group, but the difference did not reach statistical significance (13).

Testing of the effects of statins on operative mortality following coronary artery revascularization in a prospective manner is difficult, as reflected by the low number of studies from the Cochrane group. Almost all patients are now being treated with statins preoperatively and discontinuing the treatment for research purposes would be unethical. However, by making use of our detailed centralized database of CABG/OPCAB outcomes from an era where almost a quarter of the patients were not taking statins preoperatively, we have been able to test our hypothesis that statins may affect operative mortality.

Material and methods

This was a retrospective non-randomized nationwide study that included all patients who had undergone coronary artery revascularization at Landspitali University Hospital, the single institution performing open heart surgery in Iceland, between January 1, 2002 and December 31, 2006.

Patients were identified through two different computerized diagnosis and operation registries at our institution, with a 100% cross-match between registries. During these five years, a total of 720 patients had CABG/OBCAB operations, representing approximately 80% of all patients who underwent cardiac surgery in Iceland during the same period. Patients who underwent concomitant heart surgery (e.g. aortic or mitral valve surgery) were omitted.

The patients were divided into two groups for comparison: patients who had taken statins (n = 529, 78%) and those who had not (n = 191, 22%) at the preoperative evaluation for surgery. Clinical information was obtained from patient charts and surgical reports, was registered in a standardized data sheet, and reviewed by four of the authors (H.S., S.H., S.J.O., and T.G.). The information collected covered more than 100 variables, including patient gender, age, cardiovascular risk profile, history of smoking and arrhythmias, other preoperative medication (including anti-arrhythmic drugs, beta-blockers, and acetylsalicylic acid). Data on prior myocardial infarction and preoperative left ventricle ejection fraction (LVEF) were registered.

The number of vessels affected and presence of left main stem coronary disease was registered, along with whether surgery was performed electively or in an acute setting (less than 24 hours from decision to operation). Skin-to-skin time, aortic cross-clamp time, and time on cardiopulmonary bypass was compared. Height and weight were used to calculate body mass index (BMI). Heart failure symptoms before surgery were evaluated according to the New York Heart Association (NYHA) system. The European System for Cardiac Operative Risk Evaluation score (EuroSCORE) was calculated for each patient.

Postoperative complications were registered and grouped as minor (atrial fibrillation, surgical wound infection, pneumonia, drainage of pleural fluid) or major (stroke, mediastinitis, myocardial infarction (CK-MB elevation above 75 IU/L and new ST-changes or left-bundle branch block), renal failure (postoperative S-creatinine > 200 μmol/mL) requiring ultrafiltration or dialysis, reoperation for bleeding, multiple organ failure, or acute respiratory distress syndrome (ARDS)). Each complication was compared independently between the study groups.

Chest tube bleeding 24 hours postoperatively and transfusions were registered. Operative mortality was defined as death within 30 days of surgery. Patients were assigned a date and a cause of death or were identified as still living on December 1, 2010 using data from the Icelandic National Population Registry. Mean follow-up time was 61 months (range 0–97 months) and no patients were lost to follow-up.

Microsoft Excel (Microsoft Corp., Redmond, WA) was used for descriptive statistics but all other analyses were done using the R statistics package, version 2.5.1 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were compared using Welch's t-test or the Mann-Whitney U-test, based on whether the data were normally distributed. Categorical variables were compared using the Chi-square test or Fisher's exact test with 2 × 2 tables. Overall survival rate and cardiac-specific survival were estimated using the Kaplan-Meier method. Logistic regression was used to identify independent risk factors for 30-day mortality. The final model was created by manually modifying an automatically created model, which was done by performing a stepwise backwards reduction of a model including all variables with the univariate p-value of the difference between the two groups being less than 0.1. Calculated area under the receiver operating characteristic (ROC) curve was used to assess the classification properties of the finalized model. A p-value of less than 0.05 was considered statistically significant.

The study was approved by the Icelandic National Bioethics Committee and the Icelandic Data Protection Commission.

Results

A comparison of cardiovascular risk profile, operative factors, and EuroSCORE for patients who were taking statins (n = 529) and controls who were not (n = 191) is given in Table I.

Table I. Comparison of patient demographics in the statin group and control group for patients who underwent coronary arterial bypass grafting at Landspitali University Hospital, 2002–2006. Number of patients is given with percentage in parentheses, except for age, BMI, and EuroSCORE where mean (SD) is given. For skin-to-skin time, bypass time and X-clamp time, the mean value and range is shown.

	Statin group (n = 529)	Control group (n = 191)	p-value
Age, years (std.)	66.2 (9.2)	67.0 (9.2)	0.270
BMI, kg/m^2 (std.)*	28.3 (4.3)	27.0 (3.7)	< 0.001
Hypertension	343 (65)	100 (52)	0.003
Male	442 (83)	120 (78)	0.109
Diabetes	82 (16)	27 (14)	0.739
Family history of IHD**	276 (52)	68 (44)	0.091
Current smoker	133 (25)	41 (22)	0.346
Ejection fraction, %	54 (9.9)	51 (11.7)	< 0.001
EuroSCORE (st)	4.6	5.7	< 0.001
Acetylsalicylic acid	424 (80)	89 (47)	< 0.001
Beta-blocker	482 (76)	81 (42)	< 0.001
Ca-blocker	123 (23)	26 (14)	0.007
Acute operation, %	12 (2)	13 (7)	0.007
OPCAB, %	131 (25)	53 (28)	0.475
Skin-to-skin time, min (range)	204 (85–555)	207 (120–420)	0.381
Bypass time, min (range)	85 (28–366)	88 (32–246)	0.521
X-clamp time, min (range)	43 (13–191)	42 (19–106)	0.830

*BMI = body mass index, **IHD = ischemic heart disease.

Patients in the statin group had a higher BMI and were more often diagnosed with hypertension. Age, gender distribution, diabetes, and smoking were similar. Average left ventricular ejection fraction (LVEF) was slightly higher in the statin group (54.1% vs. 50.9%, p < 0.001) and the standard EuroSCORE was lower (4.6 vs. 5.7, p < 0.001). Patients taking statins were more often treated with beta-blockers, Ca-blockers, and acetylsalicylic acid.

Operation-related factors were similar between the two groups. In 92% of cases, the left internal mammary artery (LIMA) was grafted to the left anterior descending artery (LAD). The average number of grafts was 3.4 (range 1–6). On-pump CABG was performed on approximately 75% of patients in both groups, and aortic cross-clamp time was comparable. Chest tube bleeding and transfusions were similar.

Operative mortality of patients taking statins was significantly lower than in patients not taking statins (1.7% vs. 5.8%, p < 0.001). In Table II, postoperative complications are listed both separately and collectively as minor or major. Thoracocentesis was more common in the control group (17.9% vs. 11.1%, p = 0.032), but when other independent minor complications (pneumonia, atrial fibrillation, pleural effusion and urinary tract infection, superficial wound infection) were studied, no significant differences were noted between groups. The same was true of major complications such as stroke, mediastinitis, and reoperation for bleeding. Non-significant trends of lower incidence of ARDS and multiple organ failure (MOF) were seen for patients in the statin group (p = 0.250). Average length of hospital stay and days in intensive care were comparable for the two patient groups.

Table II. Comparison of minor and major complications in patients treated with statins and controls who were not, all of whom underwent coronary arterial bypass grafting at Landspitali University Hospital 2002–2006. The number of patients is given with percentage in parentheses, except for days in hospital and days in ICU, where median (range) is given.

	Statin group (n = 529)	Control group (n = 191)	p-value
Minor complications	282 (53)	88 (58)	0.328
Atrial fibrillation	205 (39)	89 (47)	0.071
Surgical wound infection	53 (10)	12 (6)	0.162
Pneumonia	32 (6)	13 (7)	0.845
Thoracocentesis	60 (11)	34 (18)	0.032
Urinary tract infection	17 (3)	10 (5)	0.299
Major complications	132 (25)	48 (21)	0.305
Stroke	11 (2)	5 (3)	0.884
Mediastinitis	3 (1)	3 (2)	0.399
Myocardial infarction	67 (13)	28 (15)	0.566
Renal dialysis/ultrafiltration	10 (2)	2 (1)	0.652
Reoperation for bleeding	16 (3)	9 (5)	0.389
Multiple organ failure/ARDS*	14 (3)	9 (5)	0.250
Days in hospital, median (range)	10 (1–75)	11 (1–96)	0.494
Days in ICU**, median (range)	1.0 (1–25)	1.0 (1–35)	0.252
Operative mortality ( ≤ 30 days)	9 (2)	9 (6)	< 0.001

*ARDS = acute respiratory distress syndrome, **ICU = intensive care unit.

A multivariate analysis on the effect of statin use on minor and major complications and mortality is shown in Table III. Beta-blockers and acetylsalicylic acid at admission were used as surrogate markers for optimal preoperative medical treatment in the model. After correcting for potential variables affecting operative mortality (age, ejection fraction, EuroSCORE, acute operation and surrogate markers for medical treatment (preoperative use of beta-blockers and acetylsalicylic acid), the preoperative use of statins significantly reduced the odds of operative mortality (OR = 0.33, 95% CI: 0.11–0.88, p = 0.046). Reducing the model by removing those components included in EuroSCORE (age, acute operation, pre-operative EF) did neither change the magnitude nor the statistical significance of the effects of pre-operative statin usage on 30-day mortality (data not shown). An interaction factor analysis between statins by age showed a non-significant interaction (p = 0.12). When patients operated acutely were excluded from the multivariate model limiting the analysis to the 695 patients undergoing non-acute operation, the contribution of statins to 30-day mortality was non-significant. Importantly, the area under the ROC curve in our model was 0.92, indicating that the model could differentiate efficiently between patients with respect to operative mortality.

Table III. The finalized multiple logistic regression analysis of factors affecting 30-day operative mortality. Odds ratio (OR) with 95% confidence interval (CI) and p-value is shown.

Risk factor for death within 30 days of surgery	OR (95% CI)	p-value
Advanced age	1.10 (1.02–1.18)	0.016
Ejection fraction (EF)	0.96 (0.92–1.00)	0.090
EuroSCORE (st)	1.28 (1.10–150)	0.002
Acute operation	5.91 (1.47–23.72)	0.012
Preoperative statin usage	0.33 (0.11–0.98)	0.043
Preoperative acetylsalicylic acid usage	1.16 (0.37–3.63)	0.805
Preoperative beta-blocker usage	0.88 (0.30–2.62)	0.818

Both over-all (p = 0.011) and cardiac-specific (p < 0.001) long-time survival was significantly better in the statin group. One- and five-year overall survival in the statin group was 96% and 90% respectively, compared with 91% and 81% in the control group. One- and five-year cardiac-specific survival in the statin group was 97% and 94% respectively, compared with 91% and 88% in the control group.

Discussion

The main finding of this study was that the operative mortality of patients treated with statins was significantly lower than in patients who were not treated with statins (1.7% vs. 5.8%). Furthermore, after correcting for potential confounders in the two groups and surrogate markers for optimal medical treatment, such as EuroSCORE, ejection fraction, and acetylsalicylic acid usage, pre-operative statin treatment was significantly correlated with lower operative mortality. The use of an advanced multivariate statistical model is important since patients in the statin group seem to have been of a lower risk and were more frequently treated with other cardioprotective agents than statins (e.g. acetylsalicylic acid and beta-blockers). This was also reflected in better long-term survival of patients taking statins.

There is little debate about the benefits of statins in reducing the risk of cardiovascular diseases, and they have become a cornerstone in the medical therapy of atherosclerosis and secondary prevention (14). Statins are a well-tolerated and remarkably safe group of drugs (15), and the current guidelines of the European Society of Cardiology, the American College of Cardiology, and the American Heart Association recommend that patients undergoing CABG surgery should receive intensified statin treatment (with target LDL levels under 100 mg/dL) unless otherwise contraindicated (1).

In the present study, approximately 75% of operations were performed as CABG using cardiopulmonary bypass, which has been shown to induce a strong systemic inflammatory response (16). However, we did not find significant differences in operative mortality or complications between CABG or OPCAB patients.

A number of studies have shown beneficial effects of statins by mechanisms that appear unrelated and independent of reduction in cholesterol level. Statin use has been shown to improve endothelial function during the systemic inflammatory response by reducing endothelial adhesion of neutrophils (17,18), increasing neutrophil apoptosis (19), and reducing the release of inflammatory cytokines (20).

Currently, it remains unclear which statin provides the greatest therapeutic benefit regarding pleiotropic effects (12). It is known that all currently available statins reduce the biosynthesis of LDL-cholesterol by inhibiting 3-hydroxy 3-methylglutaryl-coenzyme A reductase, but to different extents (21). Recent data have indicated that the beneficial pleiotropic actions of statins appear to be class-dependent, as they are mainly attributable to the reduced downstream products of mevalonate (22), which effects the systemic inflammatory response (3,4), prevents endothelial dysfunction (23), and aids in preventing plaque rupture and coronary artery occlusion (6). This has raised questions about the optimal timing of statin therapy regarding the above-mentioned effects. One study showed that the improvement in endothelial and vascular function after statin treatment decreased as soon as statin therapy was withdrawn (24).

Data from animal models suggest that the positive effects of statins on the size of myocardial infarct were lost within three days if their use was discontinued, but they were maintained if statin was used on a continuous basis. The same study showed that large doses of statins were more favorable than a standard dose (25). The start of statin therapy postoperatively could also be of importance. Collard et al. investigated the effects of discontinuation and early initiation (with statin administered through the nasogastric tube) and found beneficial effects such as a reduction of all cause-mortality and of cardiac mortality within 30 days (25). The development of a statin for intravenous administration may therefore be desirable for use in the early postoperative period (12).

Despite their beneficial effects in patients undergoing surgical coronary revascularization, recent studies have shown that statins are under-used in this group of patients. In the study by Kulik et al. involving 9,284 CABG patients who were operated between 1995 and 2004, only 36% were prescribed statins within 3 months of surgery (15). In the present study, 73% of patients who underwent coronary arterial revascularization between 2002 and December 2006 received perioperative statin therapy.

The strength of this study is that our cohort consisted of patients from a whole population, all of whom were operated in a single center. The results were therefore less likely to be affected by tertiary referral. However the retrospective, non-randomized nature of the study is an obvious weakness. Despite vigorous statistical correction for confounding variables, unmeasured and thus uncontrolled variables related to pre-operative state or treatment of the patients cannot be excluded in a non-randomized observational study. Finally, preoperative LDL-levels were not available.

In summary, this retrospective study has shown that patients treated with statins had lower operative mortality than control patients who were not treated with statins. The mechanistic explanation is not clear, but in addition to beneficial effects of LDL-lowering pleiotropic effects of statins, such as anti-inflammatory effects and/or improvement in endothelial function, may also play a role.

Acknowledgements

We thank Gunnhildur Johannsdottir for secretarial help.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

This study was supported by grants from the Landspitali University Research Fund, the University of Iceland Research Fund, and the Helga Gudmundsdottir and Sigurlidi Kristjansson Memorial Fund.