OUP user menu

Patterns and prognostic implications of low high-density lipoprotein levels in patients with non-ST-segment elevation acute coronary syndromes

Matthew T. Roe, Fang-Shu Ou, Karen P. Alexander, Laura Kristin Newby, Joanne M. Foody, W. Brian Gibler, William E. Boden, Erik Magnus Ohman, Sidney C. Smith Jr, Eric D. Peterson
DOI: http://dx.doi.org/10.1093/eurheartj/ehn364 2480-2488 First published online: 20 August 2008

Abstract

Aims The patterns and prognostic significance of low high-density lipoprotein (HDL) cholesterol levels have not been well characterized. We sought to determine the prevalence and prognostic significance of low HDL cholesterol levels in patients with non-ST-segment elevation acute coronary syndromes (NSTE ACS).

Methods and results We evaluated HDL levels among NSTE ACS patients [ischaemic ECG (electrocardiogram) changes and/or positive cardiac markers] from the CRUSADE [Can Rapid Stratification of Unstable Angina Patients Suppress Adverse Outcomes with Early Implementation of the ACC(American College of Cardiology)/AHA(American Heart Association) Guidelines] initiative treated at 555 US hospitals from January 2001 through June 2006. Clinical and angiographic characteristics, treatments, and in-hospital outcomes were analysed by categories of HDL levels measured during hospitalization. Among 93 263 NSTE ACS patients with HDL measurements, 16 854 (18.1%) had very low HDL levels (10–29 mg/dL), 32 185 (34.5%) had low HDL levels (30–39 mg/dL), 35 875 (38.5%) had normal HDL levels (40–59 mg/dL), and 8349 (9.0%) had high HDL levels (60–100 mg/dL). Patients with very low HDL levels were younger, more often male, and more commonly obese and diabetic. Patients with very low HDL levels had the greatest risk of multi-vessel coronary disease on angiography and in-hospital mortality compared with patients with normal and high HDL levels.

Conclusion Almost one-fifth of patients with NSTE ACS have very low HDL levels – a finding that adds incrementally to a greater burden of atherosclerosis and a higher risk of mortality. Consequently, strategies for mitigating the adverse prognosis associated with very low HDL levels warrant further exploration in patients with ACS.

  • High-density lipoprotein cholesterol
  • Non-ST-segment elevation acute coronary syndromes
  • Prognosis

Introduction

Lipid-lowering therapies have been recognized as a critical component of the early treatment recommendations for patients with acute coronary syndromes (ACS), as randomized clinical trials have shown that rapid lowering of elevated low-density lipoprotein (LDL) levels with high-dose statin therapy contributes to improvements in clinical outcomes.1,2 Low high-density lipoprotein (HDL) levels are a key element of the definition of metabolic syndrome—which has been shown to be associated with an increased risk of cardiovascular events in the general population—but the prognostic impact of low HDL levels has primarily been studied in patients with stable coronary disease or in those with risk factors of coronary disease.35 Although the revised practice guidelines for unstable angina and non-ST-segment elevation MI [collectively known as non-ST-segment elevation ACS (NSTE ACS)] designate a Class IC recommendation for measurement of fasting lipid panel within 24 h of admission and provide multiple recommendations for therapies and strategies to lower LDL cholesterol and triglyceride levels, there are no recommendations to guide the treatment and risk stratification of ACS patients who are found to have low HDL levels.6

We therefore studied the CRUSADE [Can Rapid Stratification of Unstable Angina Patients Suppress Adverse Outcomes with Early Implementation of the ACC(American College of Cardiology)/AHA(American Heart Association) Guidelines] National Quality Improvement Initiative database to characterize the incidence, associated clinical and angiographic characteristics, and prognostic significance of low HDL levels among patients with NSTE ACS.

Methods

Patient inclusion criteria

The CRUSADE inclusion criteria required ischaemic symptoms lasting at least 10 min within the previous 24 h together with either ischaemic ST-segment changes (ST-segment depression of at least 0.5 mm or transient ST-segment elevation of 0.5–1.0 mm) and/or positive cardiac markers (either elevated creatine kinase-MB or troponin levels). Patients were not eligible if they were transferred to a participating hospital >24 h after presentation to the initial hospital.

Data collection

Data were collected for CRUSADE anonymously during the initial hospitalization, and, because no patient identifiers were collected, patient informed consent was not required. The institutional review board of each institution approved participation in CRUSADE. Data collected included the use of acute medications (within 24 h of presentation), use and timing of invasive cardiac procedures, laboratory results (including measured lipid levels), discharge medications, and lifestyle modification interventions.

Analysis population

From a starting population of 180 842 NSTE ACS patients treated at 557 hospitals in USA from January 2001 through June 2006, we excluded 69 126 patients who did not have measured HDL values recorded during the index hospitalization, 11 273 patients who were transferred out due to inability to track clinical outcomes data after transfer given current US privacy laws, 6644 patients with recorded HDL levels that were performed either before or after hospitalization, 532 patients with HDL levels <10 mg/dL or >100 mg/dL, and four patients with missing data on death status. The final analysis population thus consisted of 93 263 patients from 555 hospitals.

Designation of high-density lipoprotein categories

Patients were analysed by categories of HDL levels defined as very low HDL levels (10–29 mg/dL), low HDL levels (30–39 mg/dL), normal HDL levels (40–59 mg/dL), and high HDL levels (60–100 mg/dL). These categories were chosen based upon classification of HDL levels by the Adult Treatment Panel III from the National Cholesterol Education Program.7

Statistical analysis

For descriptive analyses, continuous variables are presented as medians with interquartile percentiles, and categorical variables are expressed as percentages. Baseline characteristics and presentation features were compared between study population and excluded patients. Wilcoxon rank-sum test was used for continuous variables and Pearson χ2 test was used for categorical variables. Among study population, baseline characteristics, treatment profiles, and procedure use were compared across different HDL categories. To test for trend of a patient’s baseline characteristics and in-hospital care patterns with respect to different HDL categories, χ2 rank correlation statistics was used for continuous variables and χ2 rank-based group means score statistics was used for categorical variables to take into account the ordinal nature of HDL categories.

We explored the univariate relationship between factors associated with HDL levels as continuous measures including: men vs. women, older (age ≥65 years) vs. younger (age <65 years) patients, and patients with diabetes mellitus vs. patients without diabetes mellitus. We also determined demographic and clinical factors associated with very low HDL levels (10–29 mg/dL). Variables entered into the model were chosen based on clinical judgment prior to modelling. Variables included age, sex, race, body mass index, history of diabetes, hypertension, smoking status, family history of coronary artery disease, prior myocardial infarction (MI), prior heart failure, prior percutaneous coronary intervention, prior coronary artery bypass graft surgery, history of renal insufficiency, prior stroke, treatment with any home lipid-lowering agent, and recorded LDL and triglyceride values. In addition, because patients within a hospital are more likely to be treated in a similar way, generalized estimating equation models with exchangeable working correlation structure were used to adjust for correlations among clustered responses (e.g. within-hospital correlations).8 The method produces estimates similar to those obtained from ordinary logistic regression, but the estimated variances of the estimates are adjusted for the correlation of outcomes within each hospital.

To explore the continuous relationship between HDL values and unadjusted in-hospital mortality rates, an additive spline transformation of HDL values was used.911 Another spline transformation was used to adjust for age and sex because these variables were shown to be significantly associated with patterns of HDL values. Two multivariable logistic regression models with generalized estimating equations using exchangeable working correlation structure were used to explore the independent association between different HDL categories and the likelihood of severe coronary disease on angiography and in-hospital mortality. Severe coronary disease on angiography was defined as a stenosis ≥50% in the left main coronary artery and/or stenoses ≥50% in each of the three major coronary arteries for patients who underwent angiography. Variables in the model for severe coronary disease were the same as the variable list mentioned earlier. Patients who died before they had the opportunity to undergo angiography were not accounted for in this model. Variables in the model for in-hospital mortality included the previously mentioned variables plus systolic blood pressure and heart rate on presentation, positive cardiac markers, ischaemic ST-segment changes on the electrocardiogram, and signs of heart failure at presentation because these variables have been shown to be significant predictors of in-hospital mortality in the CRUSADE database.12 Odds ratios (OR) and 95% confidence intervals (CI) were determined for the different HDL categories compared with normal HDL values (40–59 mg/dL) to examine the variation of the strength of its influence on the outcomes evaluated.

A P-value <0.05 was considered significant for all tests. All analyses were performed using SAS software (version 9.0, SAS Institute, Cary, NC, USA).

Results

Clinical characteristics of the study population vs. excluded patients

A total of 87 579 patients were excluded as previously detailed. The patients who were excluded from the analysis were older, less likely male, and had a greater percentage of co-morbidities, previous cardiac events, and previous revascularization procedures compared with the 93 263 patients in the study population (Table 1).

View this table:
Table 1

Baseline characteristics of excluded patients compared with the study population

Study population (N = 93 263)Excluded patients (n = 87 579)P-valuea
Demographics
 Age (years)b66 (55, 77)70 (58, 80)<0.0001
 Male sex61.5%58.2%<0.0001
 Body mass index (kg/m2)b28.1 (24.7, 32.3)27.4 (24.0, 31.6)<0.0001
 White race79.9%80.7%<0.0001
Past medical history
 Diabetes mellitus31.1%34.9%<0.0001
 Hypertension68.6%71.1%<0.0001
 Current/recent smoking30.2%24.1%<0.0001
 Hyperlipidaemia50.0%48.8%0.0003
 Prior myocardial infarction26.9%32.3%<0.0001
 Prior heart failure14.1%21.3%<0.0001
 Prior stroke9.1%11.8%<0.0001
 Prior percutaneous coronary intervention19.7%22.9%<0.0001
 Prior coronary artery bypass graft17.9%21.0%<0.0001
 Prior renal insufficiency11.1%15.9%<0.0001
Home medications
 Statin31.8%37.0%<0.0001
 Non-statin lipid-lowering agent5.0%5.0%0.84
 Any lipid-lowering agentc34.3%39.6%<0.0001
Presenting features
 Positive cardiac markers91.1%87.3%<0.0001
 Electrocardiographic findings
  ST-depression32.7%35.7%<0.0001
  Transient ST-elevation6.8%6.5%
  Both1.7%1.7%
 Heart rate (beats/min)b82 (70, 97)84 (70, 100)<0.0001
 Systolic blood pressure (mm Hg)b146 (126,166)143 (122,163)<0.0001
 Signs of heart failure19.9%26.3%<0.0001
  • aP-values for comparison between study population and excluded patients.

  • bPresented as median (25th, 75th percentile).

  • cAny lipid-lowering agent included patients who received either statin and/or non-statin lipid-lowering agent at home.

Clinical characteristics by high-density lipoprotein categories

Among the 93 263 patients analysed, 16 854 (18.1%) had very low HDL levels (10–29 mg/dL), 32 185 (34.5%) had low HDL levels (30–39 mg/dL), 35 875 (38.5%) had normal HDL levels (40–59 mg/dL), and 8349 (9.0%) had high HDL levels (60–100 mg/dL). Patients with very low HDL levels were younger, more often male and obese, and more commonly had diabetes mellitus, prior MI, and prior revascularization procedures compared with patients with higher HDL levels. Total cholesterol and LDL levels were lowest among patients with very low HDL levels, but the total cholesterol/HDL ratio and triglyceride levels were highest in this group (Table 2).

View this table:
Table 2

Baseline characteristics by high-density lipoprotein category

High-density lipoprotein (mg/dL)P-valuea
Overall (N = 93 263)10–29 (n = 16 854)30–39 (n = 32 185)40–59 (n = 35 875)60–100 (n = 8349)
Demographics
 Age (years)b66 (55, 77)62 (52, 74)64 (54, 75)68 (57, 78)72 (60, 81)<0.001
 Male sex61.5%76.6%70.0%53.1%34.3%<0.001
 Body mass index (kg/m2)b28 (25, 32)29 (26, 33)29 (25, 33)27 (24, 32)26 (22, 30)<0.001
 White race79.9%83.9%82.0%78.0%72.3%<0.001
Past medical history
 Diabetes mellitus31.1%36.8%32.6%28.8%24.0%<0.001
 Hypertension68.6%67.6%67.5%69.5%70.2%<0.001
 Current/recent smoking30.2%37.7%32.7%26.1%23.0%<0.001
 Hyperlipidaemia50.0%51.8%51.5%49.3%43.0%<0.001
 Prior myocardial infarction26.9%30.4%27.8%25.4%22.5%<0.001
 Prior heart failure14.1%15.1%13.1%14.1%15.9%0.12
 Prior stroke9.1%9.0%8.8%9.4%9.5%0.01
 Prior percutaneous coronary intervention19.7%23.5%21.0%18.1%13.8%<0.001
 Prior coronary artery bypass graft17.9%22.0%19.4%15.9%12.0%<0.001
 Prior renal insufficiency11.1%12.6%11.0%10.5%11.4%<0.001
Home medications
 Statin31.8%31.8%32.8%32.1%27.0%<0.001
 Non-statin lipid-lowering agent5.0%6.9%5.3%4.2%3.5%<0.0001
 Any lipid-lowering agentc34.3%35.4%35.4%34.2%28.9%<0.001
Presenting features
 Positive cardiac markers91.1%90.7%91.0%91.4%91.1%0.02
 Electrocardiographic findings
  ST-depression32.7%31.8%32.9%32.9%32.9%0.27
  Transient ST-elevation6.8%6.7%6.9%6.8%6.7%
  Both1.7%1.6%1.8%1.7%1.4%
 Heart rate (beats/min)b82 (70, 97)82 (70, 97)80 (69, 95)82 (70, 98)86 (72, 102)<0.001
 Systolic blood pressure (mm Hg)b146 (126, 166)142 (123, 162)145 (126, 166)147 (127, 168)148 (127, 169)<0.001
 Signs of heart failure19.9%19.5%18.2%20.6%24.8%<0.001
Recorded lipid values
 Total cholesterol (mg/dL)b170 (142, 203)155 (126, 188)167 (140, 199)175 (148, 207)187 (161, 217)<0.001
 High-density lipoprotein cholesterol (mg/dL)b39 (32, 48)26 (23, 28)35 (32, 37)46 (43, 51)67 (63, 75)<0.001
 Total cholesterol/high-density lipoprotein cholesterol ratiob4.3 (3.4, 5.6)6.2 (5.1, 7.6)4.9 (4.0, 5.8)3.7 (3.1, 4.5)2.7 (2.3, 3.2)<0.001
 Low-density lipoprotein cholesterol (mg/dL)b100 (76, 127)92 (68, 118)101 (78, 128)103 (79, 131)98 (74, 126)<0.001
 Triglycerides (mg/dL)b126 (87, 190)163 (111, 253)142 (100, 208)111 (78, 161)86 (62, 122)<0.001
  • aP-values for trend across high-density lipoprotein categories.

  • bPresented as median (25th, 75th percentile).

  • cAny lipid-lowering agent included patients who received either statin and/or non-statin lipid-lowering agent at home.

The continuous distribution of HDL levels demonstrated a shift towards lower HDL levels among men compared with women, among patients age <65 years compared with those age ≥65 years, and among patients with diabetes mellitus vs. those without diabetes mellitus (Figures 1A–C). The factors most strongly associated with very low HDL levels included male sex, lower measured LDL values, higher measured triglyceride values, current smoking, lack of use of home lipid-lowering agents, younger age, higher body mass index, and diabetes mellitus (Table 3).

Figure 1

Comparison of the continuous distribution of high-density lipoprotein (HDL) (mg/dL) levels among selected categories of patients: (A) men vs. women; (B) age <65 years vs. age ≥65 years; (C) diabetes mellitus vs. no diabetes mellitus.

View this table:
Table 3

Factors associated with very low high-density lipoprotein (10–29 mg/dL) levelsa

Variableχ2Adjusted odds ratio95% Confidence intervalsP-value
Male sex13412.122.04–2.21<0.001
Low-density lipoprotein cholesterol value (per 10 mg/dL increase)9090.910.91–0.92<0.001
Triglyceride value (per 10 mg/dL increase)4121.041.03–1.04<0.001
Current/recent smoking1931.301.25–1.35<0.001
Use of home lipid-lowering agents1540.770.74–0.80<0.001
Age (per 10 year increase)1130.920.91–0.93<0.001
Body mass index (per 5 unit increase)841.071.05–1.08<0.001
Diabetes mellitus781.201.15–1.24<0.001
White race521.231.16–1.30<0.001
Prior coronary artery bypass graft surgery471.191.13–1.25<0.001
Prior heart failure431.181.13–1.24<0.001
Prior renal insufficiency221.141.08–1.21<0.001
Prior percutaneous coronary intervention131.081.03–1.12<0.001
Hypertension120.940.90–0.97<0.001
Prior myocardial infarction61.051.01–1.090.016
Prior stroke51.071.01–1.130.028
  • aVariables included in the model that were not significantly associated with very low high-density lipoprotein levels were family history of coronary artery disease.

Medications and procedures

Acute (<24 h) medications were used to a similar degree among patients with very low and low HDL levels, but were used less commonly among patients with normal and high HDL levels (Table 4). Similar patterns of use of cardiac catheterization and revascularization procedures were found.

View this table:
Table 4

Acute medications and invasive procedures by high-density lipoprotein category

High-density lipoprotein (mg/dL)P-valuea
Overall (N = 93 263)10–29 (n = 16 854)30–39 (n = 32 185)40–59 (n = 35 875)60–100 (n = 8349)
Medications < 24 hb
 Aspirin95.5%95.4%95.7%95.6%95.0%0.22
 Heparinc88.5%88.3%88.9%88.5%87.1%0.02
 Beta-blockers87.2%88.6%87.3%86.8%85.0%<0.001
 Angiotensin-converting enzyme inhibitors/angiotensin receptor blockers52.4%52.6%52.3%52.5%51.6%0.57
 Lipid-lowering therapiesd56.7%57.9%57.8%56.3%52.0%<0.001
 Glycoprotein IIb/IIIa inhibitors47.3%49.5%49.6%46.2%38.6%<0.001
 Clopidogrel51.8%54.1%53.6%50.7%45.3%<0.001
Invasive procedures
 Catheterization84.6%85.8%86.5%83.8%78.4%<0.001
  Catheterization < 48 h65.1%66.4%67.6%64.1%56.9%<0.001
 Percutaneous coronary intervention49.9%52.0%52.9%48.4%39.6%<0.001
 Coronary artery bypass graft12.9%14.0%13.6%12.6%9.3%<0.001
  • aP-values for trend across high-density lipoprotein categories.

  • bMedication use reported only for patients without listed contraindications to each medication class.

  • cIncludes unfractionated and low molecular weight heparin.

  • dLipid-lowering agent included patients received either statin and/or non-statin lipid-lowering agent at home.

Coronary disease severity by high-density lipoprotein levels

Patients with very low and low HDL levels were more likely to have multi-vessel disease demonstrated on angiography, whereas patients with normal or high HDL levels were more likely to have insignificant coronary disease (Table 5). After adjusting for other clinical factors, including recorded LDL and triglyceride levels, very low HDL levels (adjusted OR 1.16; 95% CI 1.10–1.22) and low HDL levels (adjusted OR 1.10; 95% CI 1.05–1.14) were independently associated with an increased likelihood of severe coronary disease (left main disease and/or three-vessel disease) relative to normal HDL levels. In contrast, high HDL levels were associated with a lower likelihood for severe coronary disease (adjusted OR 0.86; 95% CI 0.80–0.91).

View this table:
Table 5

Extent of coronary disease by high-density lipoprotein category

High-density lipoprotein (mg/dL)P-valuea
Overall (N = 93 263)10–29 (n = 16 854)30–39 (n = 32 185)40–59 (n = 35 875)60–100 (n = 8349)
Extent of coronary artery diseaseb
 Insignificant disease11.2%9.0%9.5%12.0%19.6%<0.001
 One-vessel disease32.4%31.6%31.8%33.0%33.4%
 Two-vessel disease27.4%28.5%28.2%26.8%23.5%
 Three-vessel disease29.1%30.9%30.5%28.2%23.5%
Left main diseasec8.5%9.0%8.7%8.3%7.7%<0.001
  • aP-values for trend across high-density lipoprotein categories.

  • bSignificant coronary disease defined as a lesion ≥50% in a native coronary artery. One-, two-, or three-vessel disease refers to significant stenoses in the major native coronary arteries (left anterior descending, left circumflex, and right coronary arteries) or a branch vessel.

  • cLeft main disease refers to a significant stenosis ≥50% in the left main coronary artery and is shown separately compared with two- or three-vessel coronary disease.

In-hospital mortality

Continuous high-density lipoprotein distribution

The continuous distribution of unadjusted in-hospital mortality rates by HDL values demonstrated that the highest mortality rates were seen among patients with the lowest HDL values, and the lowest unadjusted in-hospital mortality rates were seen among patients with HDL values between 30 and 40 mg/dL (Figure 2A). However, after adjusting for age and sex, the continuous distribution of HDL values demonstrated a linear decline in mortality from HDL values of 10 to 40 mg/dL and relatively flat mortality rates from HDL values of 40 to 100 mg/dL (Figure 2B).

Figure 2

Continuous relationship between in-hospital mortality rates and high-density lipoprotein (HDL) (mg/dL) values. Broken lines indicate 95% confidence intervals: (A) overall population unadjusted mortality rates; (B) overall population mortality rates adjusted for age and sex.

High-density lipoprotein categories

The incidence of unadjusted in-hospital mortality among HDL categories was 2.9% among patients with very low HDL levels, 2.4% among those with low HDL levels, 2.7% among those with normal HDL levels, and 3.4% among those with high HDL levels. Compared with patients with normal HDL levels, patients with very low HDL levels had a higher adjusted risk of mortality (adjusted OR 1.18; 95% CI 1.04–1.35), whereas patients with low HDL levels (adjusted OR 1.01; 95% CI 0.91–1.11) and those with high HDL levels (adjusted OR 1.07; 95% CI 0.93–1.24) had a similar adjusted risk of mortality.

Discussion

We have demonstrated that almost one-fifth of patients with NSTE ACS had very low HDL levels (10–29 mg/dL) measured during the hospitalization and that these patients were younger, more often male, and commonly demonstrated features of the metabolic syndrome. Patients with very low HDL levels had a higher risk of severe coronary disease on angiography and also had a higher risk of in-hospital mortality compared with those with normal HDL levels. Consequently, very low HDL levels appear to be an adverse prognostic indicator for patients with NSTE ACS and may represent a unique target for new therapeutic interventions.

Lower HDL levels have been shown to be associated with a higher risk of cardiovascular events and a greater burden of atherosclerosis, even among patients with lower LDL levels, including those who are treated with statins.4,5 However, because lower HDL levels are a key component of metabolic syndrome, the adverse prognosis and increased severity of coronary disease associated with lower HDL levels demonstrated in this analysis may be a result of the interactions between the components of the metabolic syndrome (abdominal obesity, hypertriglyceridaemia, lower HDL levels, hypertension, and hyperglycaemia) rather than solely due to the impact of lower HDL levels on atherosclerosis progression.3,1316 Notwithstanding these potential interactions, we have shown that lower HDL levels are highly prevalent among a large cohort of patients with ACS, tend to be associated with a greater likelihood of severe coronary disease on angiography and a higher risk of adverse outcomes, and appear to correlate with younger age upon presentation with ACS. In contrast, higher HDL levels were associated with older age at presentation with NSTE ACS – a finding that indicates that higher HDL levels may forestall symptomatic coronary disease progression until older age. Nonetheless, lower HDL levels appear to be a unique, common, and potentially modifiable risk factor for patients with ACS.

The recently revised ACC/AHA practice guidelines for NSTE ACS do not recommend specific strategies to treat patients found to have low HDL levels, and the revised Adult Treatment Panel III guidelines published in 2004 focus mostly upon strategies designed to lower LDL levels among high-risk patients (including those with recent ACS).6,17 Because a recent analysis demonstrated that low HDL levels are associated with a higher risk of cardiovascular events among patients with stable coronary disease treated with statins who had achieved the target LDL level <70 mg/dL, use of intensive statin therapy for ACS patients (a Class I guidelines recommendation) may not be sufficient to improve the adverse prognosis associated with low HDL levels.1,2,4,6 Although the cholesterol ester transfer protein inhibitor torcetrapib was shown to raise HDL levels by as much as 65% and to lower LDL levels by ∼20% (beyond the effects of concomitant statin therapy) in a number of recent studies, the therapy showed no impact on the progression of carotid or coronary artery atherosclerosis, and it was associated with a higher incidence of hypertension and death leading to the termination of a large phase III clinical trial.1821 As a result, alternative athero-protective treatment strategies designed to modulate HDL levels and/or HDL function will require further study (especially in the primary prevention population) to reduce the risk of future coronary disease, as well as in the ACS population to reduce the risk of subsequent adverse outcomes.2224 In the meantime, identification of patients with ACS who have low HDL levels should prompt the initiation of intensive dietary and lifestyle modification interventions that may mitigate some of the risks associated with this adverse prognostic indicator.

There were several limitations to this analysis. First, HDL values were not recorded in a large proportion of the CRUSADE population during the analysis time period (38%), and another 10% of patients were excluded for other reasons. Because patients in this analysis were treated more commonly with evidence-based medications and more often underwent invasive cardiac procedures compared with the overall CRUSADE population, the exclusion of almost half of the original population may have biased the results and led to an incomplete characterization of the patterns and impact of HDL levels.25 Secondly, only in-hospital outcomes were collected, so the long-term impact of the findings could not be investigated. Thirdly, we could not fully characterize the extent of the metabolic syndrome in this analysis because waist circumference and fasting blood glucose levels were not collected. Fourthly, we designated HDL categories based upon clinically accepted classifications of HDL levels, but this type of analysis may have not fully characterized the impact of low HDL levels. Fifthly, home lipid-lowering therapies were used in only one-third of the analysis population, so the results of this analysis may have been influenced by the disproportionate use of lipid-lowering therapies that could not be fully accounted for in the multivariable analyses. Sixthly, recorded HDL levels during the hospitalization may differ from chronic HDL levels before presentation, so the impact of low HDL levels on the progression of coronary disease may not be accurately represented by in-hospital HDL levels. Seventhly, we demonstrated many significant differences across HDL categories but the statistical significance was likely due to the large number of patients analysed rather than to true clinical differences. Finally, we did not account for patients who died before they had an opportunity to undergo angiography in the modelling procedure that assessed the frequency of severe coronary artery disease among the HDL categories.

Funding

CRUSADE is funded by the Schering-Plough Corporation. Bristol-Myers Squibb/Sanofi-Aventis Pharmaceuticals Partnership provides additional funding support. Millennium Pharmaceuticals, Inc., also funded this work.

Acknowledgement

The authors would like to thank David Z. Bynum and Amanda McMillan for their editorial assistance.

Conflict of interest: M.T.R.: Speakers’ bureau for Millennium Pharmaceuticals, Inc., Bristol-Myers Squibb/Sanofi-Aventis Pharmaceuticals Partnership, Schering Corporation; research grants from Millennium Pharmaceuticals, Inc., Bristol-Myers Squibb/Sanofi- Aventis Pharmaceuticals Partnership, and Schering Corporation. L.K.N.: research grant support from Millennium Pharmaceuticals, Inc., Schering-Plough, Bristol-Myers Squibb/Sanofi-Aventis Pharmaceuticals Partnership; Speakers' bureau for Bristol-Myers Squibb/Sanofi-Aventis Pharmaceuticals Partnership. W.B.G.: research grants from Millennium Pharmaceuticals, Inc.; Schering Corporation; Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership. W.E.B.: Speakers’ bureau for Pfizer, Inc., Merck & Co. Inc., Sanofi, Glaxo SmithKline, and Kos Pharmaceuticals; research grants from Pfizer, Inc., Merck & Co. Inc., Aventis, Sanofi, and Kos Pharmaceuticals; consultant for Kos Pharmaceuticals. E.M.O.: research grants from Millennium Pharmaceuticals, Inc.; Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership; Schering Corporation; Berlex. S.C.S. Jr: ownership interest (stock): Johnson & Johnson, Medtronic; speaking honoraria or consulting for Sanofi, Bayer, Pfizer and Astra Zeneca. E.D.P.: research grants from Bristol-Myers Squibb, Bristol-Myers Squibb/Sanofi-Aventis Pharmaceuticals Partnership, Bristol-Myers Squibb/Merck, and Schering-Plough Corporation. F.-S.O., K.P.A., and J.M.F.: no relationship(s) to disclose.

References

References

The above article uses a new reference style being piloted by the EHJ that shall soon be used for all articles.

View Abstract