HDL‐C Levels and Revascularization Procedures in Coronary Heart Disease Patients Treated With Statins to Target LDL‐C Levels

Abstract

Background:

A low level of high‐density lipoprotein cholesterol (HDL‐C) is a strong predictor for cardiovascular disease morbidity and mortality at all low‐density lipoprotein cholesterol (LDL‐C) concentrations.

Hypothesis:

We evaluated this association in routine clinical practice among statin‐treated coronary heart disease patients who achieved LDL‐C target levels. This association also exists in routine clinical practice.

Methods:

A retrospective dynamic cohort included all male coronary heart disease patients of the Sharon‐Shomron district, Clalit Health Services, Israel, with LDL‐C levels <100 mg/dL and who were receiving statins (≥6 purchases/y) from January 1998 to June 2008. Data were collected on demographic variables; coexistence of hypertension, diabetes mellitus, and peripheral vascular diseases; details of revascularization procedures; and lipid levels. The outcome variable was revascularization procedure, by either percutaneous intervention or coronary artery bypass graft.

Results:

The study group of 909 male patients was stratified into quintiles, based on mean HDL‐C levels: Q1 (n = 179): ≤26.4 mg/dL; Q2 (n = 190): 26.4–≤30.0 mg/dL; Q3 (n = 191): >30.0–≤34.0 mg/dL; Q4 (n = 186): >34.0–≤41.0 mg/dL; Q5 (n = 163): >41.0 mg/dL. During the study period, 307 (33.8%) of the cohort required ≥1 revascularization procedure. Those in the highest quintile underwent significantly fewer procedures (40.8% for Q1 vs 16.6% for Q5, P<0.001). This significant effect of the highest HDL‐C quintile was not influenced by any variable.

Conclusions:

The protective effect of high HDL‐C levels, regardless of other risk factors, in preventing revascularization procedures was confirmed in the routine clinical practice among statin‐treated CHD patients who reached LDL‐C level <100 mg/dL. Possible additional benefits of using agents to raise HDL‐C levels should be investigated. © 2011 Wiley Periodicals, Inc.

The authors have no funding, financial relationships, or conflicts of interest to disclose.

Introduction

A large body of evidence demonstrates that statin treatment to reduce low‐density lipoprotein‐cholesterol (LDL‐C) levels to <100 mg/dL in high‐risk patients reduces the relative risk for major cardiovascular disease (CVD) events. Nevertheless, on average, only one‐third of the events are prevented.1, 2, 3 The evaluation of other treatment targets beyond LDL‐C suggests high‐density lipoprotein cholesterol (HDL‐C) is the most promising target.3, 4

Low HDL‐C levels are strong predictors for CVD morbidity and mortality at all LDL‐C concentrations.2, 5, 6, 7 This association was also reported among statin‐treated patients with optimal LDL‐C levels.8, 9, 10 However, these studies were based on either subanalysis of a statin clinical study8 or included specific populations, such as coronary heart disease (CHD) patients,8 patients who underwent percutaneous intervention (PCI),9 or males with various comorbidities.10

Furthermore, their outcomes also differed. On the other hand, compelling findings were reported recently from subanalysis of the Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) trial data. High‐density lipoprotein cholesterol concentrations were not predictive of residual vascular risk among patients in a primary prevention setting treated with potent statins and who attained very low LDL‐C concentrations.11

The aim of this study was to determine retrospectively the association between HDL‐C levels and CHD morbidity (ie, revascularization procedures) in statin‐treated CHD patients who reached LDL‐C target levels (<100 mg/dL) in the everyday clinic setting.

Methods

Study Population

The study population consisted of a dynamic cohort that included members of Clalit Health Services (CHS), the largest health maintenance organization in Israel. Clalit Health Services insures more than 50% of the population of the Sharon‐Shomron district in central Israel. The district's population is mostly urban.

All male CHD patients receiving statins (≥6 purchases/y) who had LDL‐C levels of <100 mg/dL from January 1, 1998 until June 30, 2008, were eligible. Excluded were patients who had <1 month follow‐up and those whose LDL‐C levels were ≥100 mg/dL at any time during the follow‐up period, defined as from the first measured lipid profile until June 30, 2008, until the first revascularization procedure, or death.

The Database

All medical information from the primary care clinics is recorded in the CHS computerized database and can be accessed at the level of the individual patient. Each family physician is responsible for updating the computerized medical records during all visits and following any admission to a regional hospital. The database includes demographic data, all diagnoses, laboratory values, medications, and medical procedures. This database was the source of data for our study.

Study Outcome and Predictor Variables

The study outcome variable was a revascularization procedure, either PCI or coronary artery bypass graft. Revascularization procedures are performed at the regional hospitals and are automatically documented in the CHS database. This makes them the most reliable cardiovascular endpoints in our database.

Predictor variables included age, history of hypertension, diabetes mellitus (DM), cerebrovascular accident, and peripheral vascular disease (PVD), as well as revascularization procedure before entry, and mean values of total cholesterol, triglycerides, HDL‐C, and LDL‐C.

Laboratory Methods

Biochemical analyses of all blood samples were performed on fresh samples in a core laboratory facility. The laboratory is authorized to perform tests according to the international quality standard ISO‐2009. All measured biochemical markers were identified with the use of a BM/Hitachi 917 automated analyzer (Boehringer Mannheim GmbH, Mannheim, Germany). The study was approved by the local institutional ethics committee.

Statistical Analysis

The study subjects were stratified into quintiles based on their mean HDL‐C levels during the study period. The baseline demographic and clinical characteristics of these 5 patient groups were compared.

For continuous variables, analysis of variance was used to determine associations with both independent and dependent variables. For nonparametric variables, the χ ² test was used.

Logistic regression analysis was used to determine the best model predicting a revascularization procedure. The HDL‐C level quintiles were forced into the model and all other predictor variables were entered into the model and then removed one at a time if not significantly adding to the model. They were then added back one at a time and kept in the model if they added to the model. The adequacy of the model was determined by area under the curve (AUC) calculation. Subgroup analysis of the protective effect of those in HDL‐C Q5 level was performed, adjusting for all variables that add significantly to the model unless defining the subgroup. All statistical analyses were performed using SPSS statistical software for Windows, version 17.0 (SPSS Inc., Chicago, IL).

Results

The cohort included 909 male patients, mean age 68 ± 10 years, who were followed from 1 month to 7.5 years (mean follow‐up, 5.3 ± 2.3 y). Hypertension, DM, PVD, cerebrovascular events, and a history of revascularization procedures were prevalent (59.5%, 46.5%, 17.8%, 14.5%, and 29.7%, respectively). There were 8.8 ± 6.7 HDL‐C and 6.8 ± 5.5 LDL‐C level measurements that were averaged. During the follow‐up, 307 (33.8%) members of the cohort required ≥1 revascularization procedure.

Subjects in the highest HDL‐C quintile were older, had shorter follow‐up duration, had higher LDL‐C and lower triglyceride levels, and fewer had DM, hypertension, or a previous revascularization procedure. Furthermore, those in the highest HDL‐C level quintile underwent significantly fewer revascularization procedures with either PCI or coronary artery bypass graft during the study period (40.8% for Q1 vs 16.6% for Q5, P<0.001) (Table 1).

Table 1.

Demographics, Clinical Parameters, and Study Outcomes According to Quintile of HDL Cholesterol

HDL‐C Levels (mg/dL), (No. of Subjects)	Q1, ≤26.4 (179)	Q2, >26.4–≤30.0, (190)	Q3, >30.0–≤34.0, (191)	Q4, >34.0–≤41.0, (186)	Q5, >41.0 (163)	P Value
Mean HDL‐C level (mg/dL)	22.4 ± 0.3	28.6 ± 0.3	32.5 ± 0.3	37.9 ± 0.3	48.1 ± 0.3	<0.001
Age (y)	59 ± 1	61 ± 1	62 ± 1	65 ± 1	67 ± 1	<0.001
Mean follow‐up (y)	5.8 ± 0.2	5.9 ± 0.2	5.3 ± 0.2	4.9 ± 0.2	4.6 ± 0.2	<0.001
Serum lipid levels (mg/dL)
Total cholesterol	144 ± 2	149 ± 2	149 ± 2	151 ± 2	157 ± 2	<0.001
LDL‐C	67 ± 1	71 ± 1	75 ± 1	76 ± 1	78 ± 1	<0.001
Triglycerides	245 ± 11	216 ± 11	180 ± 11	154 ± 11	124 ± 12	<0.001
Comorbidities, n (%)
Previous revascularization procedure	62 (34.6)	58 (30.5)	70 (36.6)	60 (32.3)	43 (26.4)	0.008
Previous CVA	73 (40.8)	84 (44.2)	68 (35.6)	55 (29.6)	27 (16.6)	0.63
PVD	35 (19.6)	35 (18.4)	26 (13.6)	30 (16.1)	36 (22.1)	0.58
DM	105 (58.7)	101 (53.2)	98 (51.3)	62 (33.3)	57 (35.0)	<0.001
Hypertension	105 (58.7)	135 (71.1)	102 (53.4)	112 (60.2)	87 (53.4)	0.006
Study outcomes, n (%)
All‐cause mortality	12 (6.7)	12 (6.3)	18 (9.4)	13 (7.0)	12 (7.4)	0.41
Revascularization procedures	73 (40.8)	84 (44.2)	68 (35.6)	55 (29.6)	27 (16.6)	<0.001

Abbrevations: CVA, cerebrovascular accident; DM, diabetes mellitus; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; PVD, peripheral vascular disease.

The variables that significantly added to the model predicting revascularization procedures were hypertension and length of follow‐up. Protective variables included a past revascularization procedure, older age, and being in the highest HDL‐C level quintile. There was a trend for a dose‐response relationship with those in Q4 having a borderline significant protective effect. The model was adequate with an AUC of 70% (Table 2). There were no significant interactions by any variable with the consistent protective effect of Q5 (Table 3).

Table 2.

Variables That Significantly Add to the Model Predicting a Revascularization Procedure According to HDL‐C Level

Predictor Variable	Units	OR	95% CI
Hypertension	Yes/no	2.18	1.60–2.97
Follow‐up period	Years	1.14	1.06–1.22
Age	Years	0.98	0.97–1.00
Previous revascularization procedure	Yes/no	0.61	0.44–0.86
HDL‐C Q4	Yes/no	0.70	0.49–1.01
HDL‐C Q5	Yes/no	0.38	0.24–0.58

Abbreviations: CI, confidence interval; HDL‐C, high‐density lipoprotein cholesterol; OR, odds ratio; Q, quintile.

Table 3.

Subgroup Analysis of the Protective Effect of Those in HDL‐C Q5^a

Variable	Q5 (OR)	95% CI
DM
Yes	0.33	0.16–0.69
No	0.39	0.22–0.68
Hypertension
Yes	0.43	0.25–0.73
No	0.33	0.16–0.69
Previous CVA
Yes	0.09	0.02–0.44
No	0.43	0.27–0.69
PVD
Yes	0.20	0.07–0.60
No	0.42	0.26–0.68
Age ≥70 years
Yes	0.23	0.10–0.52
No	0.46	0.28–0.78
Follow‐up period ≥6.5 years
Yes	0.53	0.29–1.00
No	0.30	0.16–0.57
Triglyceride level ≥150 mg/dL
Yes	0.46	0.22–0.96
No	0.33	0.19–0.58
LDL‐C level >70 mg/dL
Yes	0.43	0.26–0.71
No	0.31	0.13–0.75

Abbrevations: CI, confidence interval; CVA, cerebrovascular accident; DM, diabetes mellitus; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; OR, odds ratio; PVD, peripheral vascular disease; Q, quintile.

^aAdjusted for all variables that add significantly to the model unless defining the subgroup.

Discussion

We found an inverse association between HDL‐C levels >41.0 mg/dL and the need for revascularization procedures in male CHD patients treated with statins and who had achieved target LDL‐C levels <100 mg/dL.

High‐density lipoprotein is the main particle of the reverse cholesterol transport, and low levels are a strong independent predictor for CVD morbidity and mortality at all LDL‐C concentrations.2, 5, 6, 7 Nevertheless, it is not clear whether this association always exists in statin‐treated patients who achieve low LDL‐C levels. Previous studies revealed similar findings, but differed in design, study groups, and outcome parameters.

Barter et al concluded, in a post hoc analysis of the Treating to New Targets (TNT) trial, that among CHD subjects with LDL‐C <70 mg/dL, those with the highest quintile of HDL‐C levels (>55 mg/dL) were at lower risk for major CVD events than those with the lowest quintile (<38 mg/dL).8 DeGoma et al evaluated only male patients with LDL‐C levels <60 mg/dL and various comorbidities (only 40%–60% of them used statins). They reported that the rate of myocardial injury or hospitalization from CHD was inversely related to HDL‐C levels, regardless of statin use or recent myocardial injury. A U‐shaped relationship was observed between HDL‐C and all‐cause mortality.10 Kini et al reported a strong inverse association between HDL‐C levels and all causes of mortality in patients with LDL‐C <70 mg/dL who underwent PCI. However, only 84.5% of their patients used statins.9 Our study group was homogeneous with highly compliant statin‐treated male CHD subjects whose LDL‐C levels were <100 mg/dL. Contrary to these findings, subanalysis of the JUPITER trial data revealed that HDL‐C concentrations were not predictive of residual vascular risk among primary prevention patients treated with potent statins who attained very low LDL‐C levels.11

The study results reveal that HDL‐C >41 mg/dL (the fifth HDL‐C level quintile) is significantly and independently associated with lower rates of revascularization procedures. This finding also correlates with the National Cholesterol Education Program (NCEP) III definition of low HDL‐C level in males as <40 mg/dL1 and the previous finding of HDL‐C levels as an independent predictor for CVD morbidity and mortality.2, 5, 6, 7

The study results also explored the relatively low HDL‐C levels in the Israeli population. Similar results in CHD patients were reported by the Bezafibrate Infraction Prevention (BIP) study group. Levels of HDL‐C ranged from 34 ± 10 mg/dL and 45 ± 12 mg/dL in men and women, respectively.12 In comparison, the quintiles of HDL‐C levels in the post hoc analysis of the TNT study were as follows: Q1, <38 mg/dL; Q2, 38–<43mg/dL; Q3, 43–<48 mg/dL; Q4, <48–<55 mg/dL; and Q5, >55 mg/dL.8 Genetic and environmental parameters might contribute to these differences.

Study Limitations

The main strength of our study is that it reflects 10 years of routine clinical practice in statin‐treated CHD patients who reached the LDL‐C target level. Other strengths were that the lipid values used were the means of numerous tests over time, which added to the predictive value of the tests; the large number of endpoints in the cohort that were consistent in the various subgroups; and those in Q4 showing a trend for a dose response decreased risk.

The main limitation was the retrospective design, although the dynamic cohort had increased strength for finding associations. Other limitations include the highly selected study group, as well as the lack of additional clinical outcomes, such as fatal and nonfatal coronary events, stroke, and hospitalizations. Nevertheless, our findings might be even more powerful in patients who are not fully compliant with statins. The statistical analysis included both univariate and multivariate analysis. The statistical methods did not add any limitations to the study.

The clinical implication of our study results is that the next step should be to evaluate whether a therapeutic intervention to increase HDL‐C levels in high‐risk patients who achieve the target LDL‐C levels would be translated into clinical benefits that are greater than those achieved with statins alone. Nicholls et al found that statin therapy is associated with regression of coronary atherosclerosis when LDL‐C is substantially reduced and HDL‐C is increased by more than 7.5%.13 Previously, the intravenously administered synthetically produced recombined apolipoprotein A‐1 Milano was found to have a marked effect in reducing the atheroma burden.14 Taylor et al reported that in statin‐treated, high‐risk patients who had LDL‐C levels <100 mg/dL and HDL‐C levels <50–55 mg/dL, extended‐release niacin induced significant regression decrease in carotid intima media thickness.15 Cholesterylester transfer protein inhibitors are considered the most promising medications to increase HDL‐C levels, especially after safety issues were clarified with one these agents, anacetrapib.16, 17 Nevertheless, it seems that HDL‐C function is also important and is independently related to both the presence and the extent of atherosclerosis.18

Conclusion

Levels of HDL‐C >41 mg/dL independently predict for a low rate of revascularization procedures in statin‐treated CHD patients who reach the LDL‐C target level (<100 mg/dL). Further studies are needed to clarify whether agents that raise HDL‐C levels would have additional beneficial effects.