Timothy J. Vittorio, MS, MD, St. Francis Hospital – The Heart Center, Division of Cardiology/Center for Advanced Cardiac Therapeutics, Roslyn, NY 11576-1348 E-mail: email@example.com
Statins do not appear to have a significant benefit in heart failure (HF) as they do in coronary artery disease (CAD). Significant evidence exists that low serum cholesterol levels may be harmful in HF. This study sought to determine the optimal low-density lipoprotein (LDL) level in patients hospitalized with acute HF. Patients were included if they presented to the hospital with acute HF and had a lipid panel drawn during admission. The primary outcome was all-cause mortality, and secondary outcomes were rates of major cardiovascular (CV) events, left ventricular assist device (LVAD) implantation, and orthotopic heart transplantation (OHT). A total of 2428 patients were followed for a mean of 2.9±2.2 years. For the entire cohort, when compared with those with LDL levels >130 mg/dL, all-cause mortality was higher in those with LDL levels <71 mg/dL (hazard ratio, 1.68; 95% confidence interval, 1.31–2.167; P<.01). Results were similar when analyzing patients with LVEF ≤40%, HF of ischemic etiology only, and in statin users. The rates of CV events, LVAD implantation, or OHT in any comparison did not differ. Low LDL levels (<71 mg/dL), similar to low total cholesterol levels, were associated with a poorer prognosis and higher overall mortality in patients with HF, regardless of etiology and systolic function.
Given the high prevalence of chronic congestive heart failure (CHF)1 and its associated poor prognosis, evidence supporting the optimal management of patients with CHF is important in reducing its burden on the population. One aspect of management that is currently debated is the management of dyslipidemia in the CHF population and the role of statins (3-Hydroxy-3-methylglutaryl coenzyme A reductase inhibitors).
The benefits of statin use in coronary artery disease in reducing death, cardiovascular events, stroke, and revascularization is well studied.2–4 The extension of these benefits to patients with CHF was shown in multiple small prospective trials, post hoc analyses, and retrospective studies.5–15 In addition, statins have a theoretical benefit in CHF patients for several reasons. Their use has been linked to lower levels of inflammatory cytokines in patients with ischemic and nonischemic cardiomyopathy and increased endothelial nitric oxide synthase activity.16,17 Despite these benefits, however, two randomized controlled clinical trials—the Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto miocardico-Heart Failure (GISSI-HF) and the Controlled Rosuvastatin Multinational Trial (CORONA)—have challenged the routine use of statins in patients with systolic heart failure (HF).18,19
The purpose of this study was to investigate the optimal low-density lipoprotein (LDL) levels and cardiovascular outcomes in patients admitted to the hospital with acute HF.
All patients who were admitted to the hospital with acute HF from January 2003 to July 2009 were included in the study. Patient records were obtained from the information technology department of the hospital using codes from the International Classification of Diseases, Ninth Revision that were related to HF and restricted to those with acute HF. This included patients admitted with a new diagnosis of acute HF and those with acute on chronic HF. Both systolic and diastolic etiologies were included. Additional data, including demographics, admission laboratory values, and imaging results were obtained through the electronic medical record system of the hospital. Patients were included if they had a lipid panel drawn at admission and recent imaging within 3 months preceding admission, and either echocardiography, myocardial perfusion imaging, or left ventriculography to evaluate for left ventricular ejection fraction (LVEF). A patient was considered to have an ischemic etiology if this was the primary reason for the patient’s HF and obstructive coronary artery disease was diagnosed by selective coronary arteriography. Patients were entered into the study on the date of first known admission with acute HF to the hospital and outcomes were obtained starting from that time point. There were no comorbidities that excluded a patient from entering the study. The institutional review board approved the study design and data collection.
The primary outcome of the study was all-cause mortality. Secondary outcomes included cardiovascular events (myocardial infarction or unstable angina as diagnosed by the primary operator at the time of cardiac catheterization), rates of left ventricular assist device (LVAD) implantation, and rates of orthotopic heart transplantation (OHT). Only those who received cardiac catheterization could reliably be included as having a cardiovascular event. Mortality data were ascertained through both the hospital medical records and the US Social Security Death Index. Patients who underwent LVAD implantation or OHT were excluded from the all-cause mortality analysis, as these could influence mortality rates.
The chi-square test was used for comparison of categorical data and the Kruskal-Wallis test was used for comparison of continuous variables. Survival curves were generated using the Kaplan-Meier method with differences assessed using a log-rank test. A multivariable analysis using the Cox proportional hazard model was performed to ascertain the independent effects of demographic and clinical variables. These variables included age, sex, hypertension, diabetes mellitus, dyslipidemia, tobacco use, chronic kidney disease, cirrhosis, use of medications, and laboratory values (sodium, hemoglobin, renal function, liver transaminases, and albumin). Analyses were performed on the entire cohort, the subgroup of patients with LVEF ≤40%, and the subgroup of patients with ischemic heart disease. Furthermore, we did a subgroup analysis specifically of those taking statin therapy to further remove confounders of acute illness on cholesterol measurements and to look at the effects of statin use on outcomes. All analyses were performed using SPSS 17.0 software (IBM Corporation, Armonk, NY).
Baseline Clinical Characteristics
A total of 2428 patients were admitted with acute HF. The baseline clinical characteristics of these patients are summarized in Table I. Continuous variables are shown as median with interquartile ranges and categorical variables are shown as a percentage. The main cardiovascular characteristics of patients are listed in Table II.
Table I. Baseline Clinical Characteristics
Abbreviations: ACE, angiotensin-converting enzyme; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ARB, angiotensin receptor blocker; BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; HDL, high-density lipoprotein; LDL, low-density lipoprotein. Age and laboratory values shown with median and interquartile range.
LDL ≤70 (n=1048)
LDL 71–100 (n=801)
LDL 101–130 (n=381)
LDL >130 (n=198)
Chronic kidney disease
Other lipid-lowering agents
ACE inhibitor or ARB
Total cholesterol, mg/dL
Table II. Cardiovascular Characteristics
Abbreviation: LVEF, left ventricular ejection fraction.
LVEF ≤40% (% of Total)
LVEF >40% (% of Total)
Total (% of Total)
Patients were followed for a mean of 2.9±2.2 years. During this time there were 1199 (49.4%) deaths, 484 (19.9%) cardiovascular events, 59 (2.4%) LVAD implantations, and 78 (3.2%) OHTs. The Kaplan-Meier curve for all-cause mortality of the total cohort is shown in Figure 1. For patients with an LVEF ≤40%, the Kaplan-Meier curve is shown in Figure 2. Patients with known ischemic disease, regardless of LVEF, are shown in Figure 3. The Cox regression analyses with associated adjusted hazard ratios are summarized in Table III. These values are compared with the reference subgroup of patients with LDL >130 mg/dL. Rates of cardiovascular events, LVAD implantation, and OHT are summarized in Table IV.
Table III. Risk of Mortality With Various LDL Values
Patients with low LDL levels <71 mg/dL had significantly higher all-cause mortality when compared with patients with LDL levels >130 mg/dL (hazard ratio, 1.68; 95% confidence interval, 1.31–2.17; P<.01). Results were similar when comparing between depressed LVEF and when looking specifically at patients with ischemic heart disease. There was no statistical difference when comparing between groups for secondary outcomes of cardiovascular events, LVAD implantation, or OHT.
Subgroup Analysis of Statin Therapy
There were 975 patients on statin therapy. Of these, there were 565 (57.9%) patients with depressed LVEF ≤40% and 481 (49.3%) patients with known ischemic etiology for their HF. The results of this subgroup analysis are also summarized in Table III and Table IV.
Our study demonstrates a higher all-cause mortality rate in patients with LDL values <71 mg/dL when compared with those >130 mg/dL. This difference was also significant when restricting analysis to only those patients taking statins, a subgroup analysis meant in part to exclude those who may have a physiologic decrease in circulating LDL levels, as discussed below.
Statin use and LDL levels in patients with HF have been extensively investigated. The three most recent observational and registry studies examining statin therapy in HF patients were completed in 2006–2007, although the first studies of this nature date back to 2004.20 In 2006, Go and colleagues21 examined the Kaiser Permanente registry of CHF patients and showed that statins reduced all-cause mortality (14.5 per 100 person-years with statin therapy vs 25.3 per 100 person-years without statin therapy). Subsequently, Folkeringa and colleagues22 examined the European registry of new HF hospitalizations and again found that statin use reduced all-cause mortality independent of etiology of HF (odds ratio, 0.42). Most recently, Dickinson and colleagues23 performed a post hoc analysis from the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) of patients with New York Heart Association (NYHA) functional class II or III and LVEF <35%, which also showed that statin use reduced all-cause mortality in HF of any etiology (hazard ratio, 0.70).
Only two prospective randomized controlled clinical trials have been performed to date. In 2007, the CORONA trial studied elderly patients with ischemic cardiomyopathy, LVEF <40% with NYHA class III or IV symptoms, or LVEF <35% with NYHA class II symptoms. The study showed that the composite primary outcome of cardiovascular death plus nonfatal myocardial infarction or stroke in the group treated with rosuvastatin 10 mg daily vs the placebo-treated group did not differ.18 In 2008, the GISSI-HF examined patients with NYHA II class through IV and LVEF <40%. Similarly, the study found that the composite primary outcome of death plus cardiovascular hospitalizations in the group treated with rosuvastatin 10 mg daily vs the placebo-treated group did not differ.19 Our study aligns with these two studies in that patients with reduced LVEF on statin therapy have significantly lower rates of long-term survival, regardless of the etiology of their HF.
The relationship between cholesterol and mortality in HF patients has also been investigated. Horwich and colleagues24 were the first to introduce serum total cholesterol as a novel prognostic indicator for patients with advanced HF, suggesting that lower serum total cholesterol predicted lower long-term rates of survival. This work was substantiated by Rauchhaus and colleagues,25 where lower serum total cholesterol was independently associated with a worse prognosis. Most recently, lower total cholesterol levels were shown to independently predict increased in-hospital mortality risk in those admitted with acute HF.26
Several mechanisms for the apparent benefit of higher total serum cholesterol levels have been proposed. The three most popular theories are the endotoxin lipoprotein hypothesis, the coenzyme Q10 (ubiquinone) hypothesis, and the selenoprotein hypothesis. The endotoxin lipoprotein hypothesis, proposed by Rauchhaus and colleagues,27 states that patients with HF experience mesenteric venous and bowel wall edema, which, in turn, causes bacterial translocation and endotoxin release into the bloodstream. Lipoproteins are then needed to bind endoxtoxin and decrease the systemic release of proinflammatory cytokines. The ubiquinone hypothesis suggests that statin use decreases available levels of ubiquinone, which is needed for the production of ATP in order to meet the metabolic demands of cells.28 Lastly, the selenoprotein hypothesis reasons that statins interfere with the enzymatic isopentenylation of selenocysteine tRNA and prevent its maturation to a functional tRNA molecule. This results in a decrease in available selenoproteins, which are a necessary precursor to selenoprotein enzymes that serve a similar antioxidant role to ubiquinone.29
It is possible, however, that decreased levels of cholesterol may simply be an effect of advanced HF. It has been thought that decreased levels of LDL are associated with cardiac cachexia, which is a state independently associated with increased mortality in HF.30 Alternatively, in vitro studies have suggested that circulating cytokines may act to decrease lipoprotein levels by decreasing hepatic lipoprotein production and increasing LDL receptor activity.31 These circulating cytokines may instead be the causative agent of the increased mortality in HF rather than a decrease in lipoprotein levels.
This study has several limitations. As mentioned above, there is an inherent limitation in a nonrandomized, observational, retrospective study, and, as such, cause and effect of the influence of LDL levels on outcomes cannot be conclusively proven from our study. Moreover, the retrospective design of the study does not allow the acquisition of pertinent patient details and nuances, which would provide a plausible explanation for the relatively low utilization of β-adrenoceptor blockers (<59%) and angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (<51%), as well as for the relatively high 3-year mortality rate (49.4%) observed in our study, as compared with other recent studies. In addition, it has to be noted that the group with LDL <71 mg/dL was a sicker population (eg, increased prevalence of chronic kidney disease and cirrhosis, lower serum hemoglobin, and albumin) and this may have adversely affected the 3-year mortality rate. On the other hand, we recognize that LDL levels can be influenced by medications and disease states, which may not have been constant during the follow-up period. A prospective trial would be required to measure LDL levels over the follow-up period and ensure that levels were maintained within a prespecified range. Despite the above limitations, and given the large size of the cohort and the significant results after adjusting for characteristics and potential confounders, we feel that our study provides useful information.
Our study calls to attention the important correlation between decreased levels of serum cholesterol and increased mortality in advanced HF. Moreover, extending that correlation it further indicates that low levels of LDL cholesterol are also associated with a poor prognosis in HF.
Disclosure: There are no relationships with industry to disclose.