Author + information
- Received August 21, 2007
- Revision received October 18, 2007
- Accepted October 22, 2007
- Published online February 19, 2008.
- Michael Miller, MD, FACC⁎,1,⁎ (, )
- Christopher P. Cannon, MD, FACC†,2,
- Sabina A. Murphy, MPH†,
- Jie Qin, MS†,
- Kausik K. Ray, MD, MRCP‡,3,
- Eugene Braunwald, MD, MACC†,4,
- PROVE IT-TIMI 22 Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Michael Miller, University of Maryland Hospital, Division of Cardiology, Room S3B06, 22 South Greene Street, Baltimore, Maryland 21201.
Objectives The purpose of this study was to assess the impact of on-treatment triglycerides (TG) on coronary heart disease (CHD) risk after an acute coronary syndrome (ACS).
Background The PROVE IT-TIMI (Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis In Myocardial Infarction) 22 trial demonstrated that low-density lipoprotein cholesterol (LDL-C) <70 mg/dl was associated with greater CHD event reduction than LDL-C <100 mg/dl after ACS. However, the impact of low on-treatment TG on CHD risk beyond LDL-C <70 mg/dl has not been explored.
Methods The PROVE IT-TIMI 22 trial evaluated 4,162 patients hospitalized for ACS and randomized to atorvastatin 80 mg or pravastatin 40 mg daily. The relationship between on-treatment levels of TG and LDL-C and the composite end point of death, myocardial infarction (MI), and recurrent ACS were assessed 30 days after initial presentation.
Results Low on-treatment TG (<150 mg/dl) was associated with reduced CHD risk compared with higher TG in univariate analysis (hazard ratio [HR] 0.73, 95% confidence interval [CI] 0.62 to 0.87; p < 0.001) and in adjusted analysis (HR 0.80, 95% CI 0.66 to 0.97; p = 0.025). For each 10-mg/dl decrement in on-treatment TG, the incidence of death, MI, and recurrent ACS was lower by 1.6% or 1.4% after adjustment for LDL-C or non–high-density lipoprotein cholesterol and other covariates (p < 0.001 and p = 0.01, respectively). Lower CHD risk was also observed with TG <150 mg/dl and LDL-C <70 mg/dl (HR 0.72, 95% CI 0.54 to 0.94; p = 0.017) or low on-treatment TG, LDL-C, and C-reactive protein (<2 mg/l) (HR 0.59, 95% CI 0.41 to 0.83; p = 0.002) compared with higher levels of each variable in adjusted analysis.
Conclusions On-treatment TG <150 mg/dl was independently associated with a lower risk of recurrent CHD events, lending support to the concept that achieving low TG may be an additional consideration beyond low LDL-C in patients after ACS. (The PROVE IT-TIMI 22 trial; NCT00382460)
Epidemiologic surveys have observed that elevated levels of total cholesterol and low-density lipoprotein cholesterol (LDL-C) are associated with increased risk of coronary heart disease (CHD) (1), and therapeutic strategies that lead to a statistically significant reduction in LDL-C lower CHD event rates (2). The magnitude of CHD risk reduction as a consequence of LDL-C lowering often ranges between 25% and 35% (3). One potential impediment limiting further reduction in CHD events despite low on-treatment LDL-C is residual elevation in serum triglyceride (TG) levels (4). Historically, elevated TG has predicted CHD events in univariate analysis, only to weaken after adjustment for other covariates, including plasma glucose and high-density lipoprotein cholesterol (HDL-C), to which it is strongly and inversely correlated (5). Yet, even after adjustment for HDL-C, detailed evaluation of population-based prospective studies has disclosed an independent effect of TG on CHD events (6). Coupled with the knowledge that combined hyperlipidemia (i.e., elevated LDL-C and TG) promotes CHD to a significantly greater extent than either high LDL-C or TG alone (7), the present analysis was undertaken to test the hypothesis that low on-treatment levels of TG when added to low LDL-C would be superior to low LDL-C alone in reducing subsequent CHD events after an acute coronary syndrome (ACS).
Study population and protocol
The study population originated from the PROVE IT-TIMI (Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis In Myocardial Infarction) 22 trial, a study prospectively designed to compare the effect of intensive versus standard therapy to reduce LDL-C, as previously reported (8,9). Briefly, 4,162 men and women hospitalized for ACS with total cholesterol <240 mg/dl, or <200 mg/dl if receiving lipid-lowering therapy, were randomly assigned to receive intensive therapy (atorvastatin 80 mg daily) or standard therapy (pravastatin 40 mg daily) for a mean follow-up period of 2 years. In addition, gatifloxacin versus placebo was tested concomitantly in a factorial design. Total cholesterol, TG, and HDL-C were measured using an enzymatic colorimetric assay (Roche Modular system, Roche Diagnostics, Indianapolis, Indiana) after the recommended 12-h overnight fast (10). The LDL-C was estimated using the formula: total cholesterol − (TG/5 + HDL-C) (11), or directly measured if TG exceeded 400 mg/dl. Lipid and lipoprotein levels were obtained at baseline, 1, 4, 8, 16, and 24 months, and the final visit. The composite end point of death, myocardial infarction (MI), or recurrent ACS was used as previously outlined (12).
Kaplan-Meier event rates for the composite end point of interest were determined during follow-up (at 2 years) after censoring patients with events within 30 days of the initial ACS event. Hazard ratios (HRs) and associated 95% confidence intervals (CIs) were calculated using selected cut-points as referents. Cut-points were derived from the National Cholesterol Education Program Adult Treatment Panel III guidelines (13,14) and included: LDL-C <70 mg/dl, the optional target goal in ACS patients; TG <150 mg/dl, the normal designate also used in the metabolic syndrome classification; and HDL-C <40 mg/dl in men and <50 mg/dl women, as similarly classified (14). In addition, other selected cut-points for TG (100 and 200 mg/dl) and non–HDL-C (100 and 130 mg/dl) were evaluated (13–15). The impact of TG <150 mg/dl with achieved dual parameters of low LDL-C (<70 mg/dl) and C-reactive protein (CRP; <2 mg/l) (16) on recurrent CHD events was also examined. A Cox proportional hazards model included clinically important variables (e.g., age, gender, smoking, hypertension, obesity, diabetes), potential confounders or effect modifiers (e.g., low HDL-C, peripheral vascular disease, prior statin therapy, prior ACS), and intervention (atorvastatin 80 mg/day vs. pravastatin 40 mg/day) to estimate the effect of on-treatment LDL-C and TG in the adjusted analysis. The proportional hazards assumption was evaluated for each model. Additionally, risk estimates were used to determine the relative decrease in hazard associated with each on-treatment change of 10 mg/dl for lipids and lipoproteins. We also assessed the relative change in hazard for each 10% lowering of TG that occurred between baseline and the first month of therapy in statin-naïve patients. All analyses were conducted on data through 2 years of follow-up. Statistical analysis was performed using Stata/SE version 9.2 (StataCorp, College Station, Texas).
Study population and baseline characteristics
Of the 4,162 patients initially enrolled in the PROVE IT-TIMI 22 trial, the present analysis focuses on the incidence of death, MI, or rehospitalization for ACS beginning after an event-censored interval of 30 days (n = 3,718), with follow-up through 2 years. Baseline characteristics (Table 1) identified a predominantly male cohort (79%), with a relatively high percentage of smokers (36%), hypertension (49%), and obesity (39%). At increasing TG quintiles, there were corresponding increases in LDL-C and decreases in HDL-C. A scatterplot analysis demonstrated a positive correlation between TG and LDL-C (Spearman rho = 0.20; p < 0.0001) and inverse correlation between TG and HDL-C (Spearman rho = −0.24; p < 0.0001).
Continuous risk estimates for LDL-C, TG, and HDL-C
A Cox proportional hazards model was used to evaluate LDL-C, non–HDL-C, TG, HDL-C, and CRP as continuous variables (Table 2). The univariate relative decrease in hazard associated with each 10-mg/dl reduction in LDL-C and non–HDL-C was 4.0% and 4.8%, respectively (p < 0.01), whereas each 10 mg/dl decrement in TG was associated with a 1.8% reduction in risk (p < 0.001). However, only TG remained statistically significantly associated with death, MI, and recurrent ACS after covariate adjustment that included LDL-C (p < 0.001) or non–HDL-C (p = 0.010).
Effect of lowering TG levels on the composite end point
A Cox proportional model was also used to assess the effect of TG lowering on the composite end point in statin-naïve subjects (n = 2,587). For each 10% lowering of TG that was attained during the first month of therapy, there was a 2.7% (p = 0.003) reduced incidence of subsequent events in unadjusted analysis and a 2.3% (p = 0.035) lower incidence after adjustment for other covariates including high LDL-C (>70 mg/dl) and low HDL-C (<40 mg/dl in men and <50 mg/dl and women).
On-treatment effects using various cut-points of LDL-C, TG, and non–HDL-C
The subsequent series of analyses focused on events based on National Cholesterol Education Program cut-points for LDL-C and TG. At 30 days, 34.6% of patients had TG ≥150 mg/dl. Between 30 days and the 2-year follow-up, significantly fewer events occurred among patients with LDL-C <70 mg/dl (13.0%) than with LDL-C ≥70 mg/dl (16.2%) (HR 0.81, 95% CI 0.68 to 0.96; p = 0.015) (Fig. 1A). Similarly, fewer events occurred with TG <150 mg/dl (13.2%) than with TG ≥150 mg/dl (17.6%) (HR 0.73, 95% CI 0.62 to 0.87; p < 0.001) in univariate analysis (Fig. 1B) and after adjustment for age, gender, high LDL-C, low HDL-C, smoking, hypertension, obesity, diabetes, prior statin therapy, prior ACS, peripheral vascular disease, and treatment effect (HR 0.80, 95% CI 0.66 to 0.97; p = 0.025). A Cox proportional model further examined the relationship between achieved LDL-C and TG at 30 days and risk of recurrent events. Compared with LDL-C ≥70 mg/dl and TG ≥150 mg/dl (referent), lower CHD risk was observed with low on-treatment TG (<150 mg/dl) and LDL-C (<70 mg/dl) (HR 0.72, 95% CI 0.54 to 0.94; p = 0.017), with a graded response among patients with LDL-C ≥70 mg/dl and TG <150 mg/dl (HR 0.85, 95% CI 0.67 to 1.08; p = 0.180) in adjusted analysis (Fig. 2). Low HDL-C was not associated with increased risk of CHD events after covariate adjustment (HR 1.01, 95% CI 0.84 to 1.22; p = 0.911). Similarly, evaluation of LDL-C <70 mg/dl with varying cut-points of TG (100 and 200 mg/dl) and non–HDL-C (100 and 130 mg/dl), identified a statistically significantly lower risk with TG <200 mg/dl (HR 0.60, 95% CI 0.45 to 0.81; p = 0.001) and a nonsignificant trend with TG <100 mg/dl (HR 0.82, 95% CI 0.61 to 1.10; p = 0.189), non–HDL-C <130 mg/dl (HR 0.79, 95% CI 0.61 to 1.02; p = 0.067), and non–HDL-C <100 mg/dl (HR 0.83, 95% CI 0.66 to 1.05; p = 0.123) compared with higher levels of LDL-C, TG, or non–HDL-C in adjusted analysis (Table 3).
Attainment of LDL-C <70 mg/dl and TG <150 mg/dl with atorvastatin 80 mg/day and pravastatin 40 mg/day
The proportion of patients with achieved LDL-C <70 mg/dl or TG <150 mg/dl, the dual parameter (LDL-C <70 mg/dl and TG <150 mg/dl), or the triple parameter (LDL-C <70 mg/dl, CRP <2 mg/l, and TG <150 mg/dl) at 30 days stratified by treatment arm is shown in Table 4. Overall, a higher percentage of atorvastatin 80 mg/day patients attained any of these parameters than pravastatin 40 mg/day patients (p < 0.0001 for each comparator). However, even with the more intensive lipid-lowering regimen, only 56.1% of atorvastatin-treated patients attained LDL-C <70 mg/dl and TG <150 mg/dl, and only 35% achieved all 3 parameters.
On-treatment effects of LDL-C <70 mg/dl, CRP <2 mg/l, and TG <150 mg/dl
The association between TG and recurrent CHD events was further assessed in the presence of achieved LDL-C <70 mg/dl and CRP <2 mg/l, previously noted to be associated with improved event-free survival (16). The risk of death, MI, or recurrent ACS after achieving 1, 2, or all 3 parameters (i.e., LDL-C <70 mg/dl, CRP <2 mg/l, and/or TG <150 mg/dl) is shown in Table 5. Between 30-day and 2-year follow-up, the risk of the composite end point was 28% lower in the presence of any 1 parameter (p = 0.017), 32% lower with any 2 parameters (p = 0.007), and 41% lower when all 3 parameters were achieved compared with their absence (p = 0.002) in an adjusted analysis.
In this analysis of the PROVE IT-TIMI 22 trial, the most noteworthy finding was the reduced risk of CHD with low on-treatment TG (<150 mg/dl) that was independent of the level of LDL-C. For each 10-mg/dl decline in on-treatment TG, we observed a 1.6% lower risk of the composite end point (p < 0.001) after adjustment for LDL-C and other covariates. Moreover, the combination of low LDL-C (<70 mg/dl) and low TG (<150 mg/dl) was associated with the lowest event rates compared with higher LDL-C, higher TG, or both. Our observations on the relationship between lower event rates with reduced on-treatment TG are consistent with 2 recent studies. They include a large Chinese prospective study that found TG to be predictive of CHD mortality, even in the setting of low total cholesterol (17), and the Prospective Cardiovascular Münster study stratified by HDL-C, which identified a higher CHD risk with high TG (>150 mg/dl) in subjects at all levels of LDL-C (18). Taken together, aiming for low on-treatment levels of LDL-C and TG may be particularly effective after ACS where residual CHD risk remains elevated despite recent diagnostic and therapeutic advancements (19).
The PROVE IT-TIMI 22 trial (9) and the Heart Protection Study (20) served as the impetus for the National Cholesterol Education Program’s optional recommended LDL-C target of <70 mg/dl (14). Whereas lowering LDL-C to <70 mg/dl has been recommended, the impact of low on-treatment TG beyond achieved LDL-C <70 mg/dl has been less well defined. For example, although TG levels <150 mg/dl are defined as normal, the National Cholesterol Education Program does not recommend TG lowering as a primary target of therapy. Rather, non–HDL-C (e.g., total cholesterol − HDL cholesterol) has become a secondary target when TG levels exceed 200 mg/dl (13).
However, several lines of evidence support TG as a biomarker of CHD risk owing to the role of TG-rich lipoproteins in atherothrombosis. Following the hydrolysis of exogenously derived chylomicrons or endogenously secreted very-low-density lipoprotein, cholesterol-enriched remnant by-products enter the subendothelial space. In hypertriglyceridemic states, remnants accumulate, resulting in a proinflammatory and oxidative milieu that may enhance adhesion molecule expression, foam cell formation, and smooth muscle cell toxicity (21). Indirectly, high levels of TG may also be associated with hypertriglyceridemic HDL particles, which are thought to be less efficient in reverse cholesterol transport (22,23), as well as an increased proportion of small, dense LDL particles which may be more susceptible to oxidative modification (24,25). The fact that plasma TG correlates with atherogenic remnants (26), coupled with the excess risk of CHD with combined elevation of LDL-C and TG (27), supports the notion that low on-treatment LDL-C and TG may improve CHD risk beyond low LDL-C alone. The 1.6% reduction in risk associated with each 10-mg/dl decrement in on-treatment TG that is independent of the reduction in risk associated with decreases in LDL-C or non–HDL-C is noteworthy because it includes adjustment of other closely aligned TG covariates (e.g., diabetes, hypertension, obesity, and HDL-C). Moreover, the 2.3% lower incidence of recurrent CHD events associated with each 10% lower TG concentration (after adjustment for high LDL-C and low HDL-C) raises the possibility that additional reduction in CHD risk may be attained through efforts aimed at lowering both LDL-C and TG compared with lowering LDL-C alone. Accordingly, if a combined strategy of low LDL-C and low TG proves to be more effective in reducing CHD events than intensive LDL-C lowering alone, then additional strategies might be considered after ACS, including replacement of saturated and trans fats with mono- and polyunsaturated fats (28), especially derivatives such as omega-3 fatty acids that are cardioprotective and at high doses possess TG-lowering properties (29,30), or the addition of niacin- or fibrate-based therapy, which is currently under investigation for CHD event rate reduction beyond LDL-C lowering (31,32).
In the present analysis, lower HDL-C levels were not associated with an increased risk of death, MI, or recurrent ACS when used as either a categoric or continuous variable. Although placebo-assigned patients with low HDL-C have traditionally had the highest event rates (33), this association may have been partly minimized in the PROVE IT-TIMI 22 trial, because all patients received statin therapy that may have attenuated the excess risk associated with low HDL-C (34). The basis for this effect may reflect, in part, statin-mediated reduction in atherogenic lipoproteins and remnants (35) as well as the potential decrease in cellular adhesion molecule expression (36) that may be up-regulated in patients with low HDL-C (37). Previously, the combination of LDL-C <70 mg/dl and CRP <2 mg/l was shown to be associated with statistically significant reductions in recurrent MI or vascular death compared with higher levels of LDL-C and CRP (38). Moreover, in the PROVE IT-TIMI 22 trial, high CRP was also found to be associated with each of the factors comprising the metabolic syndrome (39). In the present study, there was a 41% lower risk of CHD events between 30 days after ACS and the 2-year follow-up with attainment of LDL-C <70 mg/dl, CRP <2 mg/l, and TG <150 mg/dl compared with higher levels in all 3 parameters after adjustment for other covariates. Mechanistically, if elevated TG represents in part a prothrombotic state (40,41), then low on-treatment levels of LDL-C, CRP, and TG may be a consideration in ACS patients, owing to the intimate linkage between lipids, inflammation, and thrombosis (42). This may justify consideration of other therapies (29–31,43–46) if future clinical trials demonstrate clinical benefit beyond LDL-C lowering.
There are several important limitations associated with the present study. First, it was not designed a priori to address whether combined low on-treatment LDL-C and TG was superior to low on-treatment LDL-C alone. Second, greater intraindividual variability has been reported for TG than for LDL-C (47). Nevertheless, although standard statistical methods were used to account for known variability, a recent evaluation of repeat TG measurements over a 4- or 12-year period (n = 2,312) identified long-term stability of TG that was similar to blood pressure and other lipid measurements (6). Third, the results of the present study are not generalizable to nonstatin-based therapies. Finally, although the study assessed outcomes of patients who were observed to have TG <150 mg/dl at different levels of LDL-C, it remains to be confirmed in randomized trials whether, and to what extent, targeting a lower on-treatment TG (15) may provide additional clinical benefit.
Among patients receiving statin therapy after ACS, on-treatment TG <150 mg/dl was associated with a lower risk of recurrent CHD events independently of the level of LDL-C. These data lend support to the concept that achieving both a low LDL-C and a low TG may be important therapeutic parameters in patients following ACS.
Online Cardiosource Slide Set
↵1 Dr. Miller was funded by a Veterans Affairs Merit Award and a National Institutes of Health grant (HL-61369). Dr. Miller has received research grant support and honoraria for lectures from Abbott, AstraZeneca, Merck–Schering-Plough, Pfizer, Reliant, and Sanofi-Aventis.
↵2 Dr. Cannon has received research grant support from AstraZeneca, Bristol-Myers Squibb, GlaxoSmithKline, Merck, Sanofi-Aventis, and Schering-Plough.
↵3 Dr. Ray has received research grant support from Bristol-Myers Squibb and Pfizer and honoraria for lectures and consulting fees from Pfizer.
↵4 Dr. Braunwald has received research support from Bristol-Myers Squibb, Merck, and AstraZeneca and honoraria for lectures from Bristol-Myers Squibb, Sanofi, and Pfizer.
The PROVE IT-TIMI trial was funded by Bristol-Myers Squibb and Sankyo. See accompanying online Cardiosource Slide Set.
- Abbreviations and Acronyms
- acute coronary syndrome
- coronary heart disease
- C-reactive protein
- high-density lipoprotein cholesterol
- low-density lipoprotein cholesterol
- myocardial infarction
- Received August 21, 2007.
- Revision received October 18, 2007.
- Accepted October 22, 2007.
- American College of Cardiology Foundation
- Sarwar N.,
- Danesh J.,
- Eiriksdottir G.,
- et al.
- Manninen V.,
- Tenkanen L.,
- Koskinen P.,
- et al.
- Ray K.K.,
- Bach R.G.,
- Cannon C.P.,
- et al.
- Friedewald W.T.,
- Levy R.I.,
- Fredrickson D.S.
- Ray K.K.,
- Cannon C.P.,
- McCabe C.H.,
- et al.
- ↵(2002) Third Report of the Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (ATP III Final Report) (National Institutes of Health, Bethesda, MD) NIH Publication No. 02-5215.
- Grundy S.M.,
- Cleeman J.I.,
- Merz C.N.,
- et al.
- Miller M.,
- Seidler A.,
- Moalemi A.,
- Pearson T.A.
- Giugliano R.P.,
- Braunwald E.
- Greene D.J.,
- Skeggs J.W.,
- Morton R.E.
- Skeggs J.W.,
- Morton R.E.
- Austin M.A.,
- King M.C.,
- Vranizan K.M.,
- Krauss R.M.
- Imke C.,
- Rodriguez B.L.,
- Grove J.S.,
- et al.
- Brown B.G.,
- Boden W.
- Ballantyne C.M.,
- Herd J.A.,
- Ferlic L.L.,
- et al.
- Rezaie-Majd A.,
- Prager G.W.,
- Bucek R.A.,
- et al.
- Calabresi L.,
- Gomaraschi M.,
- Villa B.,
- Omoboni L.,
- et al.
- Ridker P.M.,
- Morrow D.A.,
- Rose L.M.,
- Rifai N.,
- Cannon C.P.,
- Braunwald E.
- Ray K.K.,
- Cannon C.P.,
- Cairns R.,
- et al.
- Gelfand E.V.,
- Cannon C.P.
- Gaziano J.M.,
- Hennekens C.H.,
- O’Donnell C.J.,
- Breslow J.L.,
- Buring J.E.