Journal of the American College of Cardiology Vol. 46, No. 8, by the American College of Cardiology Foundation ISSN /05/$30.

Transcription

1 Journal of the American College of Cardiology Vol. 46, No. 8, by the American College of Cardiology Foundation ISSN /05/$30.00 Published by Elsevier Inc. doi: /j.jacc FOCUS ISSUE: PROVE IT-TIMI 22 Relationship Between Uncontrolled Risk Factors and C-Reactive Protein Levels in Patients Receiving Standard or Intensive Statin Therapy for Acute Coronary Syndromes in the PROVE IT-TIMI 22 Trial Kausik K. Ray, MRCP, MD,* Christopher P. Cannon, MD, FACC,* Richard Cairns, MSC, David A. Morrow, MD, MPH,* Nader Rifai, PHD, Ajay J. Kirtane, MD, Carolyn H. McCabe, BS,* Allan M. Skene, PHD, C. Michael Gibson, MS, MD, Paul M. Ridker, MD, MPH,* Eugene Braunwald, MD, MACC,* for the PROVE IT-TIMI 22 Investigators Boston, Massachusetts OBJECTIVES This study sought to evaluate what set of factors correlate with higher or lower C-reactive protein (CRP) levels in patients receiving standard and intensive statin therapy. BACKGROUND C-reactive protein levels in blood are becoming recognized as a potential means of monitoring cardiovascular risk. Although statin therapy is known to reduce CRP levels, many patients have a high CRP level despite statin therapy. METHODS This study was a cross-sectional study of 2,885 patients from the Pravastatin or Atorvastatin Evaluation and Infection Therapy Thrombolysis In Myocardial Infarction 22 (PROVE IT-TIMI 22) trial, which assessed the relationship between uncontrolled cardiovascular risk factors and CRP level at four months after enrollment. RESULTS In a multivariate model, several risk factors were weakly but independently associated with higher CRP levels: age, gender (with or without hormone replacement therapy), body mass index 25 kg/m 2, smoking, low-density lipoprotein 70 mg/dl, glucose 110 mg/dl, high-density lipoprotein 50 mg/dl, triglycerides 150 mg/dl, and the intensity of statin therapy. A direct relationship between the number of uncontrolled risk factors present and CRP levels (p ) was observed for both statin regimens. Despite the presence of each uncontrolled risk factor, prior randomization to intensive statin therapy was associated with a lower CRP level (p ). Across all strata, defined by the number of uncontrolled risk factors present, CRP levels were lower among those receiving intensive statin therapy. CONCLUSIONS The use of intensive statin therapy lead to a lower CRP level independent of the presence of single or multiple cardiovascular risk factors. Even among patients receiving intensive statin therapy, a lower CRP level was observed in patients with the fewest coronary risk factors present, suggesting that control of multiple risk factors may be a means to further achieve lower CRP levels. (J Am Coll Cardiol 2005;46: ) 2005 by the American College of Cardiology Foundation Inflammation, and C-reactive protein (CRP) in particular, is an independent predictor of cardiovascular risk for the development of both coronary artery disease (CAD) in the healthy population (1) and recurrent events in patients with known CAD (2). It has also recently been observed that achieving a lower CRP level with statin therapy is associated with a lower risk of recurrent events in patients with acute coronary syndromes (ACS) (3), even after adjusting for all Framingham risk factors and prior revascularization (4). Further, greater reductions in CRP with statin therapy have From the *Brigham and Women s Hospital/Harvard Medical School, Nottingham Clinical Research Group, Children s Hospital, Boston, and the Division of Cardiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts. The PROVE IT-TIMI 22 study was funded by Bristol-Myers Squibb and Sankyo. Dr. Ray is funded by a British Heart Foundation International Fellowship. Manuscript received July 21, 2005; revised manuscript received August 3, 2005, accepted August 8, been associated with less progression of atherosclerosis in patients with stable CAD (5). However, despite statin therapy, many patients with CAD have a high CRP level (6), and an important clinical question that has arisen is what else might be done to further lower CRP among patients receiving statin therapy. We sought to identify the factors correlated with higher CRP levels among patients receiving two different statin regimens to identify the future targets for intervention that might further lower CRP levels. First, we hypothesized that the presence of cardiovascular risk factors (e.g., high body mass index [BMI], low high-density lipoprotein [HDL] levels) would be associated with differences in CRP levels even among patients receiving intensive therapy, and second, that intensive statin therapy would be an independent predictor of a lower CRP irrespective of risk factor burden.

2 1418 Ray et al. JACC Vol. 46, No. 8, 2005 CRP, Risk Factors, and Intensive Statin Therapy October 18, 2005: Abbreviations and Acronyms ACS acute coronary syndrome BMI body mass index BP blood pressure CAD coronary artery disease CI confidence interval CRP C-reactive protein HDL high-density lipoprotein hs-crp high-sensitivity CRP IQR interquartile range LDL low-density lipoprotein PROVE IT TIMI 22 Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis In Myocardial Infarction 22 TG triglyceride METHODS Patient population. The design of the Pravastatin or Atorvastatin Evaluation and Infection Therapy Thrombolysis In Myocardial Infarction 22 (PROVE IT- TIMI 22) protocol has been previously described (7). Briefly, 4,162 patients who had a total cholesterol level of 240 mg/dl and who had been hospitalized for an ACS within the previous 10 days were enrolled. Eligible patients were randomly assigned to pravastatin 40 mg (standard therapy) or atorvastatin 80 mg daily (intensive therapy), and gatifloxacin versus placebo, in a 2 2 factorial design. Because levels of metabolic markers and CRP during ACS may not reflect a true baseline, we used a cross-sectional study design and included only patients in whom CRP, lipid component, and glucose measurements were available at four months (n 2,885) for this analysis. The analyses were performed at four months to allow full stabilization of CRP levels, lipids, and glucose after ACS (3,7). Measurement of variables. Low-density lipoprotein (LDL), HDL, triglycerides (TG), glucose, and high sensitivity CRP (hs-crp) were measured in fasting samples in the core laboratories (7). Total cholesterol, HDL, and TG were measured using an enzymatic colorimetric method using the Roche Modular system (LabCorp, Raritan, New Jersey). The LDL level was obtained by calculation (total cholesterol [TG/5 HDL]). The CRP level was measured using the validated Denka Seiken assay for hs- CRP (N.R., Children s Hospital, Boston, Massachusetts). Blood pressure (BP), weight, and current medication use were recorded at each visit, and height and smoking status were measured at enrollment. Statistical analysis. The CRP levels were not normally distributed, and therefore are reported as median and interquartile range (IQR), and groups were compared using the Wilcoxon rank sum test; CRP was subsequently log transformed at the model fitting stage. Other continuous variables were compared using t tests, and categorical variables were compared using chi-squared tests. An F test was used to test for the presence of interactions between metabolic factors (grouped by quartiles) in a linear regression model for log CRP. Initial analyses combined the two statin arms to assess the distribution of CRP as a function of potential risk factors. The on-treatment uncontrolled risk factors at four months were defined as BMI 25 kg/m 2 (World Health Organization cut off for overweight), BP 130/85 mm Hg, glucose 110 mg/dl, TG 150 mg/dl (Adult Treatment Panel [ATP] III cut offs), and HDL 50 mg/dl. The presence of a linear correlation between log CRP and each risk factor was assessed using the Pearson product moment correlation coefficient, and the two statin treatments were compared using the Fisher z test. Relationships between log CRP and a range of metabolic factors were assessed for non-linearity and are graphically represented by fitting a robust local regression model (Loess plot) (8). The relationship between incremental risk factor burden (number of uncontrolled risk factors present) and CRP level was determined for each statin regimen, and differences in CRP level across multiple groups were assessed using a Kruskal-Wallis test. Finally, linear regression analysis was used to identify predictors of log CRP in a multivariate model. The statistical analyses were performed by the TIMI study group and the Nottingham Clinical Research group. RESULTS Characteristics at baseline and at four months. There were 2,885 patients on the study drug in whom metabolic markers and CRP measurements were available at four months (69% of the initial study cohort). The analysis cohort of 2,885 patients did not differ significantly from the 1,277 excluded patients in terms of age, gender, BMI, or history of diabetes, hypertension, or smoking. There were no significant differences in baseline characteristics between treatment strategies (Table 1). However, at four months, intensive statin therapy was associated with a lower LDL (p ), TG (p ), CRP (p ), and HDL (p 0.002) compared with standard therapy. Relationship between uncontrolled risk factors and CRP level. The relationship between month 4 CRP level (log transformed) and a number of variables of interest was assessed initially among all patients. The following correlation coefficients were significantly different from zero: age (r 0.056), BMI (r 0.19), on treatment systolic blood pressure (r 0.085), glucose (r 0.098), HDL (r 0.056), LDL (r 0.17), and TG (r 0.12), but all correlations were weak. A graphical representation of the complex and varied relationship between log CRP and each metabolic risk factor is shown in Figures 1A to 1F using a loess plot. Current smoking, diabetes, and the presence of uncontrolled metabolic risk factors at month 4 (specifically BP 130/85 mm Hg, fasting glucose 110 mg/dl, TG 150 mg/dl, and BMI 25 kg/m 2 ) were each associated with a significantly higher CRP level compared with the absence of

3 JACC Vol. 46, No. 8, 2005 October 18, 2005: Ray et al. CRP, Risk Factors, and Intensive Statin Therapy 1419 Table 1. Patient Characteristics at Baseline and Month 4 by Statin Therapy Characteristics Standard Therapy (n 1,432) Intensive Therapy (n 1,453) p Baseline Age, yrs (mean SD) Male (%) Diabetes (%) Smoking current (%) Smoking past (%) Hypertension (%) Extent of disease 2 vessels with 50% stenosis (%) BMI, kg/m 2 (mean SD) Month 4 Beta-blockers (%) ACE inhibitors (%) Aspirin (%) Clopidogrel (%) Insulin (%) Oral hypoglycemics (%) HRT (% of females) Systolic BP, mm Hg (mean SD) Diastolic BP, mm Hg (mean SD) LDL, mg/dl (mean SD) HDL, mg/dl (mean SD) Triglycerides, mg/dl (mean SD) Glucose, mg/dl (mean SD) CRP, mg/l (median and interquartile range) 2.11 ( ) 1.29 ( ) ACE angiotensin-converting enzyme; BMI body mass index; BP blood pressure; CRP C-reactive protein; HDL high-density lipoprotein; HRT hormone replacement therapy; LDL low-density lipoprotein. the relevant risk factor, when both statin arms were combined (Table 2, part A). Considering subgroups defined by the presence of individual uncontrolled risk factors, use of intensive statin therapy was associated with a significantly lower CRP compared with standard therapy (p in the presence of each uncontrolled metabolic or lifestyle risk factor) (Table 2, part B). Of the candidate variables studied in univariate analyses, female gender was associated with the highest CRP level, especially when combined with hormone replacement therapy (Table 2, part A). Irrespective of hormone replacement therapy use, women receiving intensive therapy had significantly lower CRP levels (Table 2, part B). Correlations of lipid levels and CRP level. The relationship between CRP level and the lipid parameters was also assessed for each statin regimen. Weak but significant correlations were observed between CRP (log transformed) and LDL among patients receiving pravastatin 40 mg or atorvastatin 80 mg (r 0.04 and r 0.15, respectively), although these correlation coefficients were significantly different (Fisher z test, p 0.002). Similarly weak but significant correlations were seen between CRP and TG (r 0.08 for pravastatin 40 mg, r 0.12 for atorvastatin 80 mg) and between CRP and HDL (r 0.07 for both statin regimens); however, no significant differences in these coefficients were observed between regimens (Fisher z test). With an apparent interaction between the statin regimen, LDL, and CRP, the relationships between CRP (log scale) and LDL were graphically reflected using a Loess plot for each statin regimen (Fig. 2). Multivariate analysis. In a multivariate model for log CRP that included age, gender, lifestyle factors, metabolic abnormalities, and treatments potentially associated with differences in CRP, several variables were seen to have a significant correlation with CRP level (Table 3). In this population in which everyone received a statin, prior randomization to atorvastatin 80 mg was associated with a 27% lower CRP compared with pravastatin 40 mg (95% confidence interval [CI], 20 to 34; p ) independently of LDL, TG, and HDL levels. Beyond intensity of statin therapy, a number of metabolic and lifestyle risk factors were also associated with higher CRP levels, in particular smoking (45% higher; 95% CI, 32 to 59) and BMI 25 (42% higher; 95% CI, 27 to 58), p for each. Significant but weak correlations between age and CRP level were also observed, with an average CRP level that was 1.2% higher per year (95% CI, 0.8 to 1.6, p ). Female patients not receiving hormone replacement therapy (in the four months after ACS) had CRP levels 47% higher than male patients (95% CI, 31 to 65), whereas those that did take hormone replacement therapy were much higher (137% compared with male patients; 95% CI, 97 to 285), both p Both glucose and lipid abnormalities, such as LDL 70 mg/dl, TG 150 mg/dl, and HDL 50 mg/dl, were each associated with higher average CRP levels (by 19%, 20%, 15%, and 23%, respectively). A clinical event between enrollment and month 4 was associated with a 35% higher CRP level (95% CI, 11 to 65; p 0.003), whereas high blood pressure and diabetes or treatments such as

4 1420 Ray et al. JACC Vol. 46, No. 8, 2005 CRP, Risk Factors, and Intensive Statin Therapy October 18, 2005: Figure 1. The correlation between median C-reactive protein (CRP) level and 95% confidence interval (CI) (log scale, y axis) across a range of metabolic risk factors (x axis) is shown for the whole cohort using a Loess plot: (A) versus glucose (mg/dl), (B) versus triglycerides (TG) (mg/dl), (C) versus high-density lipoprotein (HDL) (mg/dl), (D) versus body mass index (BMI), (E) versus systolic blood pressure (mm Hg), (F) versus diastolic blood pressure (mm Hg). ATP Adult Treatment Panel; WHO World Health Organization. aspirin, angiotensin-converting enzyme inhibitors, clopidogrel, and gatifloxacin were not significantly related to CRP level in this model. Interaction between metabolic risk factors and CRP level. Potential interactions between metabolic risk factors and CRP level were assessed in patients receiving atorvastatin 80 mg. Stratifying by LDL and glucose quartiles at four months suggested an interaction between increasing LDL and glucose (P I 0.034) on the CRP level (Fig. 3). Similarly, a statistically significant interaction was observed between high TG and low HDL on CRP level (P I 0.006). Statistically significant interactions for other combinations of risk factors (glucose, LDL, TG, and HDL) and CRP level were not observed. Table 2. The Relationship Between CRP at 4 Months and Individual Risk Factors (Both Arms Combined), and a Comparison of CRP By Statin Therapy in Subjects With Individual Risk Factors Risk Factor Present (A) All Patients Risk Factor Absent p Standard Therapy (B) Risk Factor Present Intensive Therapy Age 65 yrs 1.71 ( ) 1.65 ( ) ( ) 1.42 ( ) Current smoker 2.06 ( ) 1.45 ( ) ( ) 1.71 ( ) Diabetes 1.94 ( ) 1.60 ( ) ( ) 1.49 ( ) Females not on HRT 2.41 ( ) 1.49 ( ) ( ) 1.93 ( ) Females on HRT 3.70 ( ) 1.49 ( ) ( ) 2.73 ( ) Systolic 130 mm Hg and diastolic 85 mm Hg 1.94 ( ) 1.62 ( ) ( ) 1.34 ( ) Glucose 110 mg/dl 1.95 ( ) 1.58 ( ) ( ) 1.44 ( ) TG 150 mg/dl 2.08 ( ) 1.45 ( ) ( ) 1.67 ( ) HDL 50 mg/dl 1.70 ( ) 1.52 ( ) ( ) 1.33 ( ) BMI 25 kg/m ( ) 1.19 ( ) ( ) 1.40 ( ) CRP values shown as median (interquartile range). For the gender analyses, females either on or off HRT is compared vs. males (A). In part B the effect of statin regimen among women on or off HRT is compared. BMI body mass index; CRP C-reactive protein; HDL high-density lipoprotein; HRT hormone replacement therapy; TG triglycerides. p

5 JACC Vol. 46, No. 8, 2005 October 18, 2005: Ray et al. CRP, Risk Factors, and Intensive Statin Therapy 1421 achieved the LDL target of 70 mg/dl (n 1,113) (Fig. 5A), or those who did not (LDL 70 mg/dl, n 1,772) (Fig. 5B), the relationship of increasing number of uncontrolled risk factors present and a higher CRP level was still evident. In addition, there continued to be an advantage for intensive statin therapy over standard therapy across all levels of overall risk factor burden. Additionally, patients defined by the presence or absence of the metabolic syndrome had lower CRP levels on intensive therapy (p for each) (Table 4). Figure 2. The correlation between median C-reactive protein (CRP) level and 95% confidence interval (CI) (log scale, y axis) is shown using a Loess plot against low-density lipoprotein (LDL) (mg/dl) for intensive and standard statin therapy. Modifiable risk factor control and achieved CRP level. Finally, the relationship between the number of persistent uncontrolled risk factors present and CRP level was assessed for each statin regimen. Presence or absence of each risk factor was defined using existing established targets (BMI 25 kg/m 2, TG 150 mg/dl, HDL 50 mg/dl, BP 130/85 mm Hg, glucose 110 mg/dl, LDL 70 mg/dl, current smokers). The relationship between the number of uncontrolled risk factors present and CRP level is graphically represented in Figure 4. Among patients allocated to standard therapy, CRP level was 3.8 mg/l (IQR, 1.9 to 7.8) when six to seven uncontrolled risk factors were present and 1.0 mg/l (IQR, 0.7 to 2.1) when none were present, p for difference across the groups. In patients allocated to intensive therapy, the corresponding CRP levels were lower and ranged from 2.4 mg/l (IQR, 1.7 to 5.7) to 0.8 mg/l (IQR, 0.4 to 1.2), p for difference across the groups. Even when restricting the analysis to patients who DISCUSSION Statin therapy lowers CRP levels (9), and a recent analysis has suggested that achieving a lower CRP level with statin therapy in patients with ACS is associated with a lower rate of recurrent cardiovascular events and less progression of atherosclerotic plaques (3,5). Despite the use of intensive statin therapy, CRP levels remain high in some patients (6). Therefore, consideration of how to lower the CRP levels further in this patient population beyond the use of statin therapy requires an understanding of which factors could potentially influence CRP levels. The present analysis shows that a more intensive statin regimen consistently achieved lower CRP levels at four months, as compared with patients randomized to standard-dose statin therapy independent of lipid levels and the presence of uncontrolled cardiovascular risk factors. However, even among patients receiving intensive statin therapy, CRP levels vary widely, ranging from 0.8 mg/l to 2.4 mg/l. In addition to the intensity of the statin regimen, we observed that the presence of a number of potentially modifiable cardiovascular risk factors such as BMI 25 kg/m 2 and smoking, and to a lesser extent, LDL 70 mg/dl, glucose 110 mg/dl, HDL 50 mg/dl, and TG 150 mg/dl, were independently associated with higher CRP levels (Table 3). In apparently healthy and statin naive populations, the presence of individual metabolic or coronary risk factors is Table 3. Multivariate Model of Independent Predictors of Log (CRP) All Patients Exp (est)* 95% CI p Intercept ( ) Age (baseline) ( ) Female on HRT (month 4) vs. male ( ) Female not on HRT (month 4) vs. male ( ) Current smoker (baseline) ( ) BMI 25 kg/m 2 (baseline) ( ) HDLC 50 mg/dl (month 4) ( ) LDL 70 mg/dl (month 4) ( ) Glucose 110 mg/dl (month 4) ( ) Clinical event pre-month ( ) TG 150 mg/dl (month 4) ( ) Atorvastatin 80 mg ( ) Baseline diabetes, systolic blood pressure 130 mm Hg and diastolic blood pressure 85 mm Hg (4 months), aspirin, clopidogrel, angiotensin-converting enzyme inhibitor, and gatifloxacin were also included in the model but did not achieve statistical significance. *The effect of each risk factor on CRP is quantified using the exponential of the estimate from the model for log(crp); e.g. atorvastatin 80 mg is associated with a CRP level of times that of patients on pravastatin 40 mg, that is 27.2% lower. BMI body mass index; CI confidence interval; CRP C-reactive protein; HDLC high-density lipoprotein cholesterol; HRT hormone replacement therapy; LDL low-density lipoprotein; TG triglycerides.

6 1422 Ray et al. JACC Vol. 46, No. 8, 2005 CRP, Risk Factors, and Intensive Statin Therapy October 18, 2005: Figure 3. The interaction between quartiles of low-density lipoprotein (LDL) and glucose on C-reactive protein (CRP) levels (intensive statin therapy arm). associated with a higher CRP level (10 12). In addition, higher CRP levels have been observed, with an increase in the number of components of the metabolic syndrome in healthy statin-naive populations (13). In the present analysis of ACS patients in which all subjects received statin therapy, the highest CRP levels were similarly observed among patients who had the greatest number of uncontrolled metabolic and/or lifestyle risk factors present. On average, a 0.3 to 0.5 mg/l higher CRP level was observed with each risk factor that was present (Fig. 4). Conversely, the lower the number of risk factors present (i.e., BMI 25 kg/m 2,HDL 50 mg/dl, TG 150 mg/dl, glucose 110 mg/dl, BP 130/85 mm Hg, LDL 70 mg/dl, and abstinence from smoking), the lower the CRP level. Irrespective of the presence of a given risk factor (Table 2, part B) or cumulative risk factor burden (Fig. 4), randomization to intensive versus standard statin therapy was associated with an approximately 1 mg/l lower CRP level. Thus, the Figure 4. The relationship between the number of uncontrolled risk factors present and the median C-reactive protein (CRP) level for standard and intensive therapy. Risk factors were defined as body mass index (BMI) 25 kg/m 2, triglycerides 150 mg/dl, high-density lipoprotein (HDL) 50 mg/dl, glucose 110 mg/dl, blood pressure (BP) 130/85 mm Hg, low-density lipoprotein (LDL) 70 mg/dl, current smoker. p across the range of risk factors for each statin regimen. use of intensive statin therapy is associated with a significantly lower CRP level, independent of a patient s cardiovascular risk factor profile. Our observation of significant associations between the presence of several uncontrolled risk factors and higher CRP levels, even among intensively treated patients, identifies these factors as important correlates of CRP in this population. We also observed statistically significant interactions between combinations of risk factors (specifically glucose and LDL and between TG and HDL) and a higher CRP level, suggesting that the relationship between the presence of multiple risk factors and a higher CRP level may extend beyond simply additive risk factor burden alone. Adverse biological interactions between risk factors such as glucose and LDL have been observed in vitro (14). The present data suggest that CRP levels could be a useful method of monitoring such interactions in vivo, beyond any additive effect. In trials of stable patients with risk factors, individual interventions such as use of thiazolidinediones (15,16), fibrates (17,18), ezetimibe (19), endocannabinoid receptor blockers (20,21), increased exercise (22,23), weight loss (24 28), a Mediterranean diet (29), intensive glycemic control (30), and smoking cessation (31) have reported significant lowering of CRP levels. Therefore, in ACS patients, beyond the use of intensive statin therapy, interventions that control for these clinical risk factors may be an important means of achieving a further reduction in CRP level. The observed correlations between CRP level and the presence of cardiovascular risk factors or use of different cardioprotective regimens suggest that the CRP level could be considered a useful global barometer of an individual s cardiovascular risk profile, and potentially of the effectiveness of interventions that modify known cardiovascular risk factors. Thus, as has been proposed, CRP levels may be a useful single test to help titrate or monitor the effectiveness of lifestyle change or risk factor intervention in future clinical trials of patients with CAD (32). This could be particularly useful in a multi-factorial approach to managing ACS patients for the

7 JACC Vol. 46, No. 8, 2005 October 18, 2005: Ray et al. CRP, Risk Factors, and Intensive Statin Therapy 1423 Figure 5. (A) The relationship between the number of uncontrolled risk factors present and median C-reactive protein (CRP) level for standard and intensive therapy in patients with low-density lipoprotein (LDL) 70 mg/dl. Across the range of risk factors, p for atorvastatin 80 mg and p for pravastatin 40 mg. (B) The relationship between the number of uncontrolled risk factors present and median CRP for standard and intensive therapy in patients with LDL 70 mg/dl. Risk factors at target were defined as body mass index (BMI) 25 kg/m 2, triglycerides 150 mg/dl, high-density lipoprotein (HDL) 50 mg/dl, glucose 110 mg/dl, blood pressure (BP) 130/85 mm Hg, current smoker. p across the range of risk factors for each statin regimen. monitoring of lifestyle changes such as the amount of exercise, dietary modification, or degree of weight loss, for which clear targets are perhaps less well established. Beyond seeing the change in CRP level, it is important to understand the exact mechanisms by which different drugs or interventions modify the CRP level. In our analysis, high-dose statin therapy was associated with a lower CRP, even among patients with an LDL that has achieved target level ( 70 mg/dl) (Fig. 5A) and that could not be explained by adjusting for known clinical correlates. This gap in the CRP level between standard and intensive statin therapy, which remained even after adjusting for risk factor burden, may have been related to greater pleiotropic effects related to statin dose, e.g., on the Rho/Rho kinase pathways (33), effects on lipid rafts (34), greater endothelial nitric oxide synthase generation (35), or cytokine switching in T lymphocytes (immune modulation) (36), which were not directly measured but which could contribute to a reduction in inflammation. Further mechanistic studies are warranted to pinpoint the exact mechanisms by which different interventions reduce inflammation. Although the relationship of CRP to LDL is weak, there is a small component of CRP that was associated with achieved LDL. Thus, some component of the reduction in CRP by intensive statin therapy seems to be attributable to the very low LDL levels achieved. With respect to this relationship, some differences were observed (Figs. 2 and 5) between standard and intensive therapy that deserve further exploration. Further study of these issues using different doses of the same statin and different statins at equivalent LDL-lowering doses could help elucidate the relationship between LDL and CRP. Study limitations. This analysis is post hoc, and therefore there are important limitations in interpretation. This analysis was cross-sectional in design and did not assess the effect of specific interventions other than statin regimen. Because levels of metabolic markers during ACS do not reflect true basal levels, we cannot comment on the possible correlation or lack of correlation between changes in these markers and changes in CRP in this analysis. In combined analyses of men and women, any HDL cut point would have been arbitrary. The mean HDL at four months was 43 mg/dl, and so we deliberately used a higher HDL cut point of 50 mg/dl, which approximated the inflection point on the Loess plot and is also the ATP III cut point in women (Fig. 1). Similar results were obtained using an HDL level of 40 mg/dl. Despite these limitations, this analysis is the most robust observation yet that shows that the lower CRP level Table 4. Relationship Between Statin Regimen and CRP Levels in Patients With and Without the Metabolic Syndrome Standard Therapy Intensive Therapy p Absence of metabolic syndrome (0 2 risk factors) 1.76 (0.9, 4.07) 1.13 (0.55, 2.73) (n 941) (n 1,060) Presence of metabolic syndrome (3 5 risk factors) 2.59 (1.22, 5.29) 1.59 (0.77, 3.33) (n 808) (n 698) C-reactive protein (CRP) levels (median and interquartile range) are shown in mg/l. Note the presence of the metabolic syndrome was defined (on treatment) at 4 months, as the presence of 3 or more of the following risk factors (body mass index 25 kg/m 2, triglycerides 150 mg/dl, high-density lipoprotein 50 mg/dl, glucose 110 mg/dl, blood pressure 130/85 mm Hg).

8 1424 Ray et al. JACC Vol. 46, No. 8, 2005 CRP, Risk Factors, and Intensive Statin Therapy October 18, 2005: associated with intensive statin therapy holds up well across many subgroups and analyses. CONCLUSIONS Among patients receiving a statin as secondary prevention after ACS, randomization to high-dose statin therapy was independently associated with a significantly lower CRP level irrespective of the number or type of risk factors present. Beyond intensity of statin therapy, control of risk factors may be a further means of achieving a lower CRP level in ACS patients. In view of the observed correlations between CRP and modifiable cardiovascular risk factors or cardioprotective regimens, the CRP level could be a useful global barometer of the effectiveness of lifestyle and therapeutic interventions that modify cardiovascular risk. Reprint requests and correspondence: Dr. Christopher P. Cannon, The TIMI Study Group, 350 Longwood Avenue, 1st Floor, Boston, Massachusetts cpcannon@partners.org. REFERENCES 1. Ridker PM, Cushman M, Stampfer MJ, et al. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336: Haverkate F, Thompson SG, Pyke SD, et al. Production of C-reactive protein and risk of coronary events in stable and unstable angina. European Concerted Action on Thrombosis and Disabilities Angina Pectoris Study Group. Lancet 1997;349: Ridker PM, Cannon CP, Morrow D, et al. C-reactive protein levels and outcomes after statin therapy. N Engl J Med 2005;352: Ridker PM, Cannon CP, Braunwald E. Response to C-reactive protein levels and outcomes after statin therapy (letter). N Engl J Med 2005;352: Nissen SE, Tuzcu EM, Schoenhagen P, et al. Statin therapy, LDL cholesterol, C-reactive protein, and coronary artery disease. N Engl J Med 2005;352: Ridker PM, Morrow DA, Rose LM, et al. Relative efficacy of atorvastatin 80 mg and pravastatin 40 mg in achieving the dual goals of low-density lipoprotein cholesterol 70 mg/dl and C-reactive protein 2 mg/l: an analysis of the PROVE-IT TIMI-22 trial. J Am Coll Cardiol 2005;45: Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med 2004;350: S-PLUS 6 for Windows Guide to Statistics V. Seattle, WA: Insightful Corporation, Albert MA, Danielson E, Rifai N, et al. Effect of statin therapy on C-reactive protein levels: the pravastatin inflammation/crp evaluation (PRINCE): a randomized trial and cohort study. JAMA 2001;286: Rohde LE, Hennekens CH, Ridker PM. Survey of C-reactive protein and cardiovascular risk factors in apparently healthy men. Am J Cardiol 1999;84: Bo M, Raspo S, Morra F, et al. Body fat and C-reactive protein levels in healthy non-obese men. Nutr Metab Cardiovasc Dis 2004;14: Aronson D, Bartha P, Zinder O, et al. Association between fasting glucose and C-reactive protein in middle-aged subjects. Diabetes Med 2004;21: Ridker PM, Buring JE, Cook NR, et al. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events: an 8-year follow-up of 14,719 initially healthy American women. Circulation 2003;107: Zhang J, Ren S, Sun D, et al. Influence of glycation on LDL-induced generation of fibrinolytic regulators in vascular endothelial cells. Arterioscler Thromb Vasc Biol 1998;18: Haffner SM, Greenberg AS, Weston WM, et al. Effect of rosiglitazone treatment on nontraditional markers of cardiovascular disease in patients with type 2 diabetes mellitus. Circulation 2002;106: Wang TD, Chen WJ, Lin JW, et al. Effects of rosiglitazone on endothelial function, C-reactive protein, and components of the metabolic syndrome in nondiabetic patients with the metabolic syndrome. Am J Cardiol 2004;93: Rizos E, Kostoula A, Elisaf M, et al. Effect of ciprofibrate on C-reactive protein and fibrinogen levels. Angiology 2002;53: Wang TD, Chen WJ, Lin JW, et al. Efficacy of fenofibrate and simvastatin on endothelial function and inflammatory markers in patients with combined hyperlipidemia: relations with baseline lipid profiles. Atherosclerosis 2003;170: Ballantyne CM, Houri J, Notarbartolo A, et al. Effect of ezetimibe coadministered with atorvastatin in 628 patients with primary hypercholesterolemia: a prospective, randomized, double-blind trial. Circulation 2003;107: Van Gaal LF. Rio-Europe. Conference of European Society of Cardiology, Munich, Germany, September Despres J-P. Rio Lipids Trial. Conference, American Heart Association, New Orleans, Louisiana, November Church TS, Barlow CE, Earnest CP, et al. Associations between cardiorespiratory fitness and C-reactive protein in men. Arterioscler Thromb Vasc Biol 2002;22: Okita K, Nishijima H, Murakami T, et al. Can exercise training with weight loss lower serum C-reactive protein levels? Arterioscler Thromb Vasc Biol 2004;24: Ryan AS, Nicklas BJ. Reductions in plasma cytokine levels with weight loss improve insulin sensitivity in overweight and obese postmenopausal women. Diabetes Care 2004;27: Wegge JK, Roberts CK, Ngo TH, et al. Effect of diet and exercise intervention on inflammatory and adhesion molecules in postmenopausal women on hormone replacement therapy and at risk for coronary artery disease. Metabolism 2004;53: Tchernof A, Nolan A, Sites CK, et al. Weight loss reduces C-reactive protein levels in obese postmenopausal women. Circulation 2002;105: Milani RV, Lavie CJ, Mehra MR. Reduction in C-reactive protein through cardiac rehabilitation and exercise training. J Am Coll Cardiol 2004;43: Dansinger ML, Gleason JA, Griffith JL, et al. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial. JAMA 2005; 293: Esposito K, Marfella R, Ciotola M, et al. Effect of a Mediterraneanstyle diet on endothelial dysfunction and markers of vascular inflammation in the metabolic syndrome: a randomized trial. JAMA 2004; 292: Schaumberg DA, Glynn RJ, Jenkins AJ, et al. Effect of intensive glycemic control on levels of markers of inflammation in type 1 diabetes mellitus in the diabetes control and complications trial. Circulation 2005;111: Frohlich M, Sund M, Lowel H, et al. Independent association of various smoking characteristics with markers of systemic inflammation in men: results from a representative sample of the general population (MONICA Augsburg Survey 1994/1995). Eur Heart J 2003;24: Bhatt DL, Topol EJ. Need to test the arterial inflammation hypothesis. Circulation 2002;106: Laufs U, Liao JK. Targeting Rho in cardiovascular disease. Circ Res 2000;87: Hillyard DZ, Jardine AG, McDonald KJ, et al. Fluvastatin inhibits raft dependent Fcgamma receptor signaling in human monocytes. Atherosclerosis 2004;172: Laufs U, La Fata V, Plutzky J, et al. Upregulation of endothelial nitric oxide synthase by HMG CoA reductase inhibitors. Circulation 1998;97: Shimada K, Miyauchi K, Daida H. Early intervention with atorvastatin modulates TH1/TH2 imbalance in patients with acute coronary syndrome: from bedside to bench. Circulation 2004;109:e213 4; author reply e213 4.