Elevated low-density lipoprotein cholesterol (LDL-C) is associated with increased risk of myocardial infarction and is a target for disease prevention. The association between initial LDL-C and statin treatment in patients with non–ST-segment elevation myocardial infarction (NSTEMI)/unstable angina pectoris (UAP) has not been well characterized. We explored detailed LDL-C levels and statin treatment in 22,938 patients with NSTEMI/UAP enrolled in the Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the American College of Cardiology/American Heart Association Guidelines Registry (2003 to 2006). Patients reporting home statin use or previous cardiovascular disease were excluded. We examined statin receipt at discharge across 4 categories of baseline LDL-C: very low (<70 mg/dl), low (70 to 99 mg/dl), high (100 to 129 mg/dl), and very high (≥130 mg/dl). The largest proportion of patients had LDL-C ≥130 mg/dl (32.6%), followed by LDL-C 100 to 129 mg/dl (32.1%), LDL-C 70 to 99 mg/dl (24.9%), and LDL-C <70 mg/dl (10.4%). Compared with high LDL-C categories, patients in the lowest LDL-C category had their first NSTEMI/UAP event at a significantly older age and had higher rates of other cardiovascular risk factors (including hypertension and diabetes) but were less likely to have a family history of coronary artery disease. Overall, 80.3% of eligible patients with NSTEMI/UAP received statins at discharge, ranging from 63.8% in those with very low LDL-C (<70 mg/dl) to 88.1% in those with very high LDL-C (>130 mg/dl). In conclusion, >1/3 of patients with NSTEMI/UAP had an LDL-C level <100. Those with low LDL-C were older, had more co-morbidities, and were less likely to be prescribed a statin at discharge than those with higher LDL-C.
Elevated serum low-density lipoprotein cholesterol (LDL-C) is an established risk factor for incident myocardial infarction (MI). Randomized clinical trials have demonstrated that cholesterol-lowering medications, such as statins, reduce mortality in patients with existing coronary artery disease (CAD). Recent guidelines encourage the use of high-intensity statins in all patients with CAD regardless of LDL-C level, yet few studies have characterized patterns of treatment with regard to LDL-C values in patients who were not taking statins at the time of CAD admission. Despite studies demonstrating the efficacy of statins in improving survival in older patients after ACS, previous work suggests that older patients are less likely than younger patients to receive statin therapy. Nevertheless, it is unknown whether LDL-C measured during admission is associated with differential discharge statin receipt in all patients or just in the elderly patients. We examined the association between LDL-C at the time of hospitalization and receipt of discharge statins in patients with non–ST-segment elevation myocardial infarction (NSTEMI)/unstable angina pectoris (UAP) without a history of CAD who were not taking statins at admission.
Methods
We analyzed data from the Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes with Early Implementation of the American College of Cardiology/American Heart Association Guidelines (CRUSADE) NSTEMI/UAP Registry. Details of the CRUSADE study design have been previousl y published. Briefly, CRUSADE was a national hospital-based initiative conducted from 2001 to 2006 that tracked provider adherence to treatment guidelines, with the ultimate goal of improving the quality of care for patients with NSTEMI/UAP. Trained hospital personnel abstracted medical record data on demographics, signs and symptoms on presentation, relevant medical history, inhospital laboratory values (including cardiac biomarkers), inhospital therapies and acute interventions, therapies prescribed at discharge, and inhospital clinical outcomes. All participating sites were required to obtain approval from local institutional review boards (or equivalent) before enrolling patients in CRUSADE. The Duke University Medical Center Institutional Review Board reviewed this study and granted a waiver of the requirement for informed consent and authorization.
Starting from patients with data on clinically relevant variables to the analysis collected from 2003 to 2006 (n = 118,252), we excluded patients who were transferred out to other facilities (n = 14,018); those with previous MI, stroke, peripheral artery disease, percutaneous coronary intervention, or coronary artery bypass grafting (n = 54,716); those on home statin use (n = 11,720); those who died during admission (n = 1,303); those with missing discharge status (n = 323); and those with unmeasured LDL-C or out-of-range values (<20 or >998 mg/dl; n = 13,234). After these exclusions, our final study population consisted of 22,938 patients.
LDL-C classification was based on the earliest measured LDL-C value during hospitalization. Patients were classified into 1 of 4 LDL-C categories: very low (<70 mg/dl), low (70 to 99 mg/dl), high (100 to 129 mg/dl), and very high (>130 mg/dl). LDL-C was treated as a categorical variable for all statistical analyses. Distributions of baseline patient and event characteristics were compared across LDL-C categories using the chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables. We used multivariable logistic regression models with generalized estimating equations with exchangeable working correlation matrix to estimate the association between LDL-C categories and statin receipt at hospital discharge, accounting for clustering of patients within hospitals. The low LDL-C category (<70 mg/dl) was treated as the referent category in all regression models. We used a 4-phase modeling approach: (1) unadjusted, (2) adjustment for age and gender, (3) adjustment for variables from the previously validated CRUSADE inhospital mortality model, lipid values, and inhospital treatments (catheterization [within 48 hours, >48 hours, or none] and revascularization [percutaneous coronary intervention or coronary artery bypass grafting]), and (4) adjustment for hospital characteristics (teaching status, bed size, hospital capabilities, and region). In a sensitivity analysis, we stratified by age (<80 and ≥80 years) to assess associations in younger and older subgroups.
For all variables used in adjusted regression analyses, the percentage of missing values ranged from 0% to 3%. For categorical variables, missing values were imputed to the most frequent nonmissing value. For continuous variables, missing values were imputed to the median of nonmissing values. Continuous variables were fit with restricted cubic splines to account for nonlinearity with the outcome. All statistical analyses were performed using SAS software, version 9.2 (SAS Institute, Cary, North Carolina).
Results
Baseline characteristics of the study population are presented in Table 1 . The median age in this population was 60 years (interquartile range 51 to 73), 38.9% of patients were women and 79.3% were white. The largest proportion of patients had LDL-C >130 mg/dl, followed by LDL-C 100 to 129 mg/dl, 70 to 99 mg/dl, and <70 mg/dl. Median patient age (interquartile range) decreased with increasing LDL-C, from 67 years (53 to 79) in the very low LDL-C category to 57 years (49 to 68; p <0.0001) in the very high LDL category. Compared with those in higher LDL-C categories, patients with very low LDL-C were more likely to be female; more likely to have higher rates of hypertension, diabetes, congestive heart failure, and renal insufficiency; less likely to be white; and less likely to report current or recent smoking.
| Variable ∗ | LDL-C (mg/dL) | p-value † | |||
|---|---|---|---|---|---|
| <70 (n=2386) | 70-99 (n=5714) | 100-129 (n=7354) | ≥130 (n=7484) | ||
| Demographics | |||||
| Age (years), median (IQR) | 67.0 (53.0, 79.0) | 64.0 (53.0, 76.0) | 60.0 (51.0, 72.0) | 57.0 (49.0, 68.0) | <0.0001 |
| Men | 55.0 | 61.0 | 61.9 | 61.5 | <0.0001 |
| White | 74.3 | 79.9 | 80.4 | 79.4 | <0.0001 |
| Body mass index, median (kg/m 2 ) (IQR) | 26.6 (23.3, 31.2) | 27.4 (24.1, 31.8) | 28.1 (25.0, 32.2) | 28.7 (25.6, 32.8) | <0.0001 |
| Medical history | |||||
| History of hypertension | 63.9 | 58.2 | 54.2 | 51.5 | <0.0001 |
| Diabetes mellitus | 25.9 | 20.2 | 16.2 | 14.7 | <0.0001 |
| History of dyslipidemia | 22.3 | 21.3 | 27.7 | 38.1 | <0.0001 |
| Current/recent smoker | 31.7 | 31.9 | 35.5 | 40.7 | <0.0001 |
| Family history of coronary artery disease | 28.1 | 33.6 | 37.7 | 41.7 | <0.0001 |
| Prior congestive heart failure | 11.9 | 6.7 | 3.9 | 3.4 | <0.0001 |
| Signs and symptoms | |||||
| Signs of congestive heart failure | 22.5 | 16.2 | 11.3 | 10.9 | <0.0001 |
| Heart rate, median (IQR) | 87.0 (73.0, 104.0) | 82.0 (70.0, 98.0) | 80.0 (69.0, 94.0) | 81.0 (70.0, 95.0) | <0.0001 |
| Systolic blood pressure, median (mmHg) (IQR) | 141.0 (120.0, 162.0) | 143.0 (124.0, 164.0) | 148.0 (128.0, 168.0) | 150.0 (131.0, 170.0) | <0.0001 |
| ST depression by ECG | 27.1 | 28.3 | 27.3 | 28.6 | 0.148 |
| Transient ST elevation | 5.9 | 7.3 | 7.2 | 6.6 | |
| Neither | 65.6 | 62.8 | 64.0 | 63.1 | |
| Both | 1.4 | 1.6 | 1.6 | 1.7 | |
| Laboratory values | |||||
| Median HDL-C (mg/dL) (IQR) | 41.0 (31.0, 54.0) | 39.0 (32.0, 49.8) | 39.0 (32.0, 48.9) | 40.0 (33.0, 48.0) | <0.0001 |
| Median LDL-C (mg/dL) (IQR) | 58.0 (48.0, 64.0) | 87.0 (79.0, 94.0) | 114.0 (107.0, 122.0) | 150.0 (139.0, 169.0) | <0.0001 |
| Median triglycerides (mg/dL) (IQR) | 101.0 (68.0, 165.0) | 110.0 (76.0, 167.0) | 123.0 (86.0, 180.0) | 139.0 (99.0, 199.0) | <0.0001 |
| Median CrCl, (mL/min) (IQR) | 56.8 (38.0, 80.6) | 63.2 (44.7, 84.5) | 68.1 (49.3, 87.2) | 71.1 (52.7, 89.4) | <0.0001 |
| Hematocrit (%) (IQR) | 40.1 (35.8, 43.9) | 41.4 (37.9, 44.7) | 42.4 (39.2, 45.2) | 43.1 (40.0, 45.9) | <0.0001 |
∗ All data presented as percentages except age, BMI, systolic blood pressure, heart rate, HDL-C, LDL-C, triglycerides, hematocrit and creatinine, presented as median (interquartile range).
† p-values from chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables.
Patients in lower LDL-C categories received less invasive therapies during hospitalization than those in higher LDL-C categories, with lower rates of both diagnostic catheterization and revascularization in very low and low LDL-C groups ( Table 2 ). In general, patients with very low LDL-C were slightly less likely to receive inhospital aspirin, clopidogrel, glycoprotein IIb/IIIa inhibitors, heparin, and β blockers than those in higher LDL-C groups. Similar patterns were observed for discharge medications; patients with very low LDL-C were less likely to receive aspirin, clopidogrel, and β blockers. Discharge statin use increased with increasing LDL-C, with 88.1% of patients receiving statins with LDL-C >130 mg/dl compared with only 63.8% of those with LDL-C <70 mg/dl. Adverse inhospital clinical events, including cardiogenic shock and heart failure, were more common in patients with lower LDL-C compared with higher LDL-C.
| Variable | LDL-C (mg/dL) | p-value ∗ | |||
|---|---|---|---|---|---|
| <70 (n=2386) | 70-99 (n=5714) | 100-129 (n=7354) | ≥130 (n=7484) | ||
| Acute medications | |||||
| Aspirin | 96.7 | 97.5 | 97.5 | 97.9 | 0.0063 |
| Clopidogrel | 53.1 | 57.8 | 60.4 | 59.8 | <0.0001 |
| Glycoprotein IIb/IIIa inhibitor | 48.7 | 55.4 | 59.5 | 60.7 | <0.0001 |
| Heparin (LMWH or UFH) | 89.8 | 90.5 | 91.2 | 91.1 | 0.1569 |
| Beta-blocker | 89.8 | 91.4 | 91.6 | 91.8 | 0.0309 |
| Statins | 45.8 | 50.4 | 53.3 | 58.5 | <0.0001 |
| In-hospital procedures | |||||
| Cardiac catheterization | 75.9 | 84.1 | 89.0 | 91.9 | <0.0001 |
| Percutaneous coronary intervention | 42.2 | 52.6 | 57.7 | 61.1 | <0.0001 |
| Coronary Bypass | 8.7 | 10.9 | 12.2 | 15.5 | <0.0001 |
| Discharge medications | |||||
| Aspirin | 94.0 | 95.8 | 96.0 | 96.8 | <0.0001 |
| Clopidogrel | 67.3 | 72.6 | 76.0 | 76.7 | <0.0001 |
| Beta blocker | 90.3 | 91.7 | 92.3 | 93.0 | <0.0001 |
| ACEI or ARB | 65.6 | 63.1 | 62.7 | 62.0 | 0.0175 |
| Statins | 63.8 | 72.5 | 83.4 | 88.1 | <0.0001 |
| In-hospital clinical events | |||||
| Cardiogenic shock | 1.7 | 1.2 | 0.9 | 1.0 | 0.0186 |
| Congestive heart failure | 7.3 | 5.1 | 3.5 | 3.4 | <0.0001 |
In unadjusted regression analyses (model 1), increasing LDL-C was associated with significantly increased odds of statin receipt at discharge ( Table 3 ). After adjustment for age and gender (model 2), these associations were slightly attenuated but remained significant. Estimates were further attenuated on adjustment for inhospital mortality risk variables, lipid values, and inhospital therapies but remained significant. After adjustment for hospital characteristics, increasing LDL-C was positively associated with discharge statin receipt.
| Model ∗ | LDL-C Category | Odds Ratio (95% CI) | p-value | Global p-value |
|---|---|---|---|---|
| Model 1: Unadjusted | LDL (70 – <100 mg/dL) | 1.52 (1.37,1.67) | <0.0001 | <0.0001 |
| LDL (100 – <130 mg/dL) | 2.75 (2.51,3.02) | <0.0001 | . | |
| LDL (≥130 mg/dL) | 3.81 (3.41,4.25) | <0.0001 | . | |
| Model 2: Adjusted for age and sex | LDL (70 – <100 mg/dL) | 1.44 (1.31,1.59) | <0.0001 | <0.0001 |
| LDL (100 – <130 mg/dL) | 2.54 (2.31,2.79) | <0.0001 | . | |
| LDL (≥130 mg/dL) | 3.45 (3.08,3.87) | <0.0001 | . | |
| Model 3: Adjusted for in-hospital mortality model variables, lipid values, and in-hospital treatments † | LDL (70 – <100 mg/dL) | 1.08 (0.94,1.24) | 0.2978 | <0.0001 |
| LDL (100 – <130 mg/dL) | 1.50 (1.26,1.78) | <0.0001 | . | |
| LDL (≥130 mg/dL) | 1.67 (1.33,2.10) | <0.0001 | . | |
| Model 4: Model 3 + hospital characteristics ‡ | LDL (70 – <100 mg/dL) | 1.08 (0.93,1.24) | 0.3133 | <0.0001 |
| LDL (100 – <130 mg/dL) | 1.51 (1.27,1.81) | <0.0001 | . | |
| LDL (≥130 mg/dL) | 1.69 (1.33,2.14) | <0.0001 | . |
∗ Since guidelines for statin use changed during the study period, study year was included in all the models.
† In-hospital mortality model covariates: age, SBP on admission, signs of heart failure, prior CHF, family history of CAD, BMI, heart rate on admission, diabetes, sex, race, hypertension, current/recent smoker, dyslipidemia, ECG findings, renal insufficiency, insurance status, and type of doctor that primarily cared for the patient; lipid values: triglycerides, total cholesterol, and HDL; in-hospital therapies: catheterization (within 48 hours, >48 hours, none) and revascularization (PCI or CABG).
‡ Hospital characteristics: teaching status, bed size, hospital capabilities, and region.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
