- Email: [email protected]
Clinical Outcomes of Patients With Diabetic Nephropathy Randomized to Clopidogrel Plus Aspirin Versus Aspirin Alone (A post hoc Analysis of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance [CHARISMA] Trial)
Clinical Outcomes of Patients With Diabetic Nephropathy Randomized to Clopidogrel Plus Aspirin Versus Aspirin Alone (A post hoc Analysis of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance [CHARISMA] Trial) Arijit Dasgupta, MDa,*, Steven R. Steinhubl, MDb,c, Deepak L. Bhatt, MD, MPHd, Peter B. Berger, MDb, Mingyuan Shao, MSce, Koon-Hou Mak, MDf, Keith A.A. Fox, MB, ChBg, Gilles Montalescot, MD, PhDh, Michael A. Weber, MDi, Steven M. Haffner, MDj, Alexios P. Dimas, MDk, P. Gabriel Steg, MDl, and Eric J. Topol, MDm, for the CHARISMA Investigators No prospective randomized trial has specifically examined the long-term outcomes of clopidogrel use in patients with chronic kidney disease. This study aimed to determine the risks and benefits of long-term clopidogrel administration in patients with diabetic nephropathy, the most common form of chronic kidney disease. We performed a post hoc analysis of the CHARISMA trial, which randomly assigned patients without active acute coronary syndrome, but with established atherosclerotic disease (symptomatic) or multiple risk factors for atherosclerotic disease (asymptomatic), to clopidogrel plus aspirin versus placebo plus aspirin. All CHARISMA patients (n ⴝ 15,603) were separated into the 3 groups: nondiabetic patients, diabetic patients without nephropathy, and diabetic patients with nephropathy. Within each group, outcomes of patients randomly assigned to clopidogrel were compared with those of patients randomly assigned to placebo. Outcomes in the prespecified CHARISMA subgroups of asymptomatic and symptomatic patients were also compared with respect to study drug assignment and nephropathy status. Patients with nephropathy who received clopidogrel had no difference in bleeding, but experienced significantly increased cardiovascular (CV) and overall mortality compared with those randomly assigned to placebo. There were no differences in bleeding, overall mortality, or CV mortality for nondiabetic or diabetic patients without nephropathy who received clopidogrel versus placebo. In the asymptomatic cohort, patients with nephropathy randomly assigned to clopidogrel had significantly increased overall and CV mortality compared with placebo, whereas asymptomatic patients without nephropathy randomly assigned to clopidogrel had no significant mortality difference compared with placebo. In conclusion, this post hoc analysis suggested that clopidogrel may be harmful in patients with diabetic nephropathy. Additional studies are needed to investigate this possible interaction. © 2009 Published by Elsevier Inc. (Am J Cardiol 2009;103:1359 –1363)
a
Gill Heart Institute, University of Kentucky, Lexington, Kentucky; Geisinger Clinic, Danville, Pennsylvania; cThe Medicines Company, Zurich, Switzerland; dVA Boston Healthcare System and Brigham and Women’s Hospital, Boston, Massachusetts; eDepartment of Cardiovascular Medicine, Cleveland Clinic, Cleveland, Ohio; fGleneagles Medical Centre, Singapore; gUniversity and Royal Infirmary of Edinburgh, Edinburgh, Scotland; hInstitut de Cardiologie-CHU Pitie-Salpetriere, Paris, France; i SUNY Downstate Medical Center, Brooklyn, New York; jUniversity of Texas Health Science Center at San Antonio, San Antonio, Texas; kHygeia Hospital, Athens, Greece; lHopital Bichat, Paris, France; and mScripps Translational Science Institute, Scripps Clinic, La Jolla, California. Manuscript received October 24, 2008; revised manuscript received and accepted January 23, 2009. The CHARISMA trial was supported by the Sanofi Aventis, Bridgewater, New Jersey/Bristol-Myers Squibb, New York, New York, partnership. *Corresponding author: Tel: 859-323-8040; fax: 859-323-6475. E-mail address: [email protected] (A. Dasgupta). b
0002-9149/09/$ – see front matter © 2009 Published by Elsevier Inc. doi:10.1016/j.amjcard.2009.01.342
Clopidogrel exerts its beneficial effect on platelets through irreversible binding of its active metabolite to the P2Y12 receptor on platelets.1 Although the pharmacokinetics of this active metabolite of clopidogrel is poorly described, it is known that approximately half the clopidogrel labeled with radioactive carbon is excreted in urine, and early studies found higher levels of the major, but still inactive, metabolite in the blood of patients with severe kidney disease. Also, because renal impairment can cause changes in platelet surface proteins,2 it is possible that the effect of clopidogrel in patients with chronic kidney disease (CKD) may be different from that in patients with normal kidney function. One post hoc analysis of a randomized trial comparing clopidogrel plus aspirin with aspirin alone reported diminished efficacy and a nonsignificant trend toward harm in patients with CKD randomly assigned to dual-antiplatelet therapy.3 However, these results were not www.AJConline.org
1360
The American Journal of Cardiology (www.AJConline.org)
Figure 1. Post hoc analysis design. ASA ⫽ aspirin.
supported by a larger post hoc analysis of a different trial that showed no excess mortality risk with clopidogrel use in patients with CKD, even when analyzed by tertiles of renal function.4 The aim of this analysis was to further investigate the relation between CKD and long-term clopidogrel administration through analysis of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial. Methods Objectives and methods of the CHARISMA trial have been previously described.5 In brief, the CHARISMA trial randomly assigned 15,603 patients with clinically evident cardiovascular (CV) disease (symptomatic patients) or multiple atherothrombotic risk factors for CV disease (asymptomatic patients) to receive either clopidogrel 75 mg/day or matching placebo in addition to daily aspirin (by protocol 75 to 162 mg). Patients had no evidence of active acute coronary syndrome at the time of initial enrollment. Patients were followed up for a median of 28 months to assess the primary efficacy end point of first occurrence of myocardial infarction (MI), stroke, or CV death. The secondary efficacy end point was CV death, MI, stroke, and hospitalization for acute coronary syndrome. Patients were also followed up for specific safety end points regarding bleeding, which was characterized using the Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) criteria.6 The CHARISMA case report form used for data collection included identification of patients with a history of diabetic nephropathy (n ⫽ 2,009; 12.9%). Although creatinine was not recorded, preventing calculation of estimated creatinine clearance, the presence of diabetic nephropathy was 1 of the entry criteria for enrollment in the study, and its presence or absence at baseline was documented in all patients. The protocol requested that all patients with diabetes and microalbuminuria with albumin ⱖ30 g/mL be considered to have diabetic nephropathy. In our post hoc analysis, outcomes with respect to treatment with clopidogrel or placebo were compared for patients with diabetic nephropathy (n ⫽ 2,009), diabetic patients without nephropathy (n ⫽ 4,546), and nondiabetic patients (n ⫽ 9,047; Figure 1). End points of interest included the prespecified primary and secondary efficacy end points, individual thrombotic outcomes, and bleeding outcomes. Statistical comparisons of baseline demographic and clinical characteristics of the study drug groups (clopidogrel
and placebo) were performed using Pearson’s chi-square test for categorical data and 2-sample t test for continuous data. The Kaplan-Meier method was used to estimate event rates of clinical outcomes. Differences in event rates were measured using log-rank test. Cox proportional-hazards models were created to evaluate the association of diabetic nephropathy and randomized therapy without adjusting for covariates and by adjusting for baseline characteristics and medical history, respectively. Model selection in multivariable Cox regressions was validated by bootstrap resampling. A 2-sided p value ⬍0.05 was considered statistically significant. All analyses were performed using SAS software, version 8.2 (SAS Institute, Cary, North Carolina). Results Almost all CV events occurred significantly more frequently in diabetic patients with nephropathy (n ⫽ 2,009) and diabetic patients without nephropathy (n ⫽ 4,546) compared with nondiabetic patients (n ⫽ 9,047; Table 1). Patients with diabetic nephropathy had a higher case fatality rate of MI compared with diabetic patients without nephropathy and nondiabetic patients (20% vs 14% and 11%, respectively), but this higher rate was not statistically significant (p ⫽ 0.240). Baseline characteristics of patients with diabetic nephropathy assigned to clopidogrel (n ⫽ 1,006) versus placebo (n ⫽ 1,003) are listed in Table 2. There were no significant differences in demographic characteristics between the 2 groups. Rates of primary, secondary, individual thrombotic, and safety end points for patients with diabetic nephropathy based on study drug assignment are listed in Table 3. Patients with nephropathy randomly assigned to clopidogrel experienced a significant increase in overall mortality compared with placebo patients, as well as significantly increased CV mortality. A multivariable interaction test confirmed the significant increase in both overall and CV mortality in patients with diabetic nephropathy treated with clopidogrel versus placebo, with adjusted hazard ratios (HRs) of 1.8 (95% confidence interval [CI] 1.2 to 2.7, p ⫽ 0.006) and 1.7 (95% CI 1.1 to 2.9, p ⫽ 0.028), respectively. Kaplan-Meier methods were used to estimate event curves for CV mortality and showed an early increase in events in patients with nephropathy randomly assigned to clopidogrel that appeared to continue to widen with increasing duration of therapy (Figure 2). A formal interaction analysis found a significant interaction of diabetic nephropathy and clopidogrel on risk of CV death with an HR of 1.856 (95% CI 1.106 to 3.116, p ⫽ 0.019). Figure 3 shows the interaction of study drug assignment on rates of the primary end point (CV mortality, MI, and stroke) and overall mortality in patients with diabetic nephropathy, diabetic patients without nephropathy, and nondiabetic patients. Patients with diabetic nephropathy randomly assigned to clopidogrel experienced significantly increased overall mortality. No similar hazard was found in diabetic patients without nephropathy assigned to clopidogrel or nondiabetic patients assigned to clopidogrel. The CHARISMA trial found that the asymptomatic cohort (n ⫽ 3,284) had a significant increase in overall mortality in patients assigned to clopidogrel compared with
Coronary Artery Disease/Clopidogrel in Nephropathy
1361
Table 1 Efficacy end points in the CHARISMA patient cohort (n ⫽ 15,603) Outcome
Nondiabetic (n ⫽ 9,047)
Diabetic Without Nephropathy (n ⫽ 4,546)
Diabetic Nephropathy (n ⫽ 2,009)
p Value
Overall CV death/MI/stroke
569 (6.3%) 223 (2.5%)
Nonfatal MI
170 (1.9%)
Nonfatal stroke
221 (2.4%)
Overall death
379 (4.2%)
Overall MI
192 (2.1%)
160 (8.0%) 1.3 (1.1–1.6) 82 (4.1%) 1.7 (1.3–2.2) 51 (2.5%) 1.4 (1.0–1.9) 42 (2.1%) 0.9 (0.6–1.2) 118 (5.9%) 1.5 (1.2–1.8) 60 (3.0%) 1.4 (1.1–1.9) 12 (0.6%) 2.5 (1.2–5.1)
⬍0.001
CV death
378 (8.3%) 1.4 (1.2–1.5) 162 (3.6%) 1.5 (1.2–1.8) 121 (2.7%) 1.4 (1.1–1.8) 132 (2.9%) 1.2 (1.0–1.5) 248 (5.5%) 1.3 (1.1–1.6) 134 (2.9%) 1.4 (1.1–1.8) 19 (0.4%) 1.7 (0.9–3.2)
Fatal MI
22 (0.2%)
⬍0.001 0.004 0.110 ⬍0.001 0.002 0.021
Values expressed as number (percent) and HR (95% CI). Table 2 Baseline characteristics of patients with diabetic nephropathy (n ⫽ 2,009) randomly assigned to clopidogrel versus placebo Variable
Placebo Clopidogrel p Value (n ⫽ 1,003) (n ⫽ 1,006)
Age (yrs) Women White Hispanic Asian Black Other Smoking status Current Former Hypertension Hypercholesterolemia Congestive heart failure Previous MI Atrial fibrillation Previous stroke Previous transient ischemic attack Peripheral arterial disease Previous percutaneous coronary intervention Previous coronary artery bypass grafting Previous carotid endarterectomy Previous peripheral angioplasty
63.0 63.1 326 (32.5%) 345 (34.3%) 690 (68.8%) 692 (68.8%) 154 (15.4%) 141 (14.0%) 90 (9.0%) 93 (9.2%) 31 (3.1%) 53 (5.3%) 38 (3.8%) 27 (2.7%)
0.7 0.4 0.08
0.8 157 (15.7%) 447 (44.6%) 865 (86.2%) 804 (80.2%) 68 (6.8%) 197 (19.6%) 31 (3.1%) 87 (8.7%) 43 (4.3%) 162 (16.2%) 113 (11.3%)
161 (16.0%) 458 (45.5%) 892 (88.7%) 777 (77.2%) 76 (7.6%) 183 (18.2%) 33 (3.3%) 82 (8.2%) 38 (3.8%) 148 (14.7%) 102 (10.1%)
0.1 0.1 0.5 0.4 0.8 0.7 0.6 0.4 0.4
153 (15.3%) 132 (13.1%)
0.2
29 (2.9%) 66 (6.6%)
27 (2.7%) 54 (5.4%)
0.8 0.2
those assigned to placebo (5.4% vs 3.8%; p ⫽ 0.04), as well as an increase in CV mortality (3.9% vs 2.2%, respectively; p ⫽ 0.01). Because nearly half (44%) the patients enrolled in the asymptomatic cohort had a diagnosis of diabetic nephropathy, we also evaluated outcomes in the asymptomatic cohort based on the presence of diabetic nephropathy and study drug assignment. In asymptomatic patients with diabetic nephropathy randomly assigned to clopidogrel (n ⫽ 750) compared with asymptomatic patients with diabetic nephropathy randomly assigned to placebo (n ⫽ 721), rates of CV (4.1% vs 2.2%, respectively; p ⫽ 0.037) and overall
mortality (6.1% vs 3.5%, respectively; p ⫽ 0.018) were significantly higher in the clopidogrel cohort. There was no significant difference in CV or overall mortality in asymptomatic patients without diabetic nephropathy with regard to randomized drug assignment. However, the directionality of mortality trends was similar to those with diabetic nephropathy. CV mortality was 3.6% in clopidogrel versus 2.2% in placebo (p ⫽ 0.081) and overall mortality was 4.7% in clopidogrel versus 4.1% in placebo (p ⫽ 0.539). When the symptomatic cohort (n ⫽ 12,153) was analyzed with respect to nephropathy and study drug assignment, results suggested a benefit for clopidogrel use in patients without diabetic nephropathy, but no benefit in patients with diabetic nephropathy. Randomization to clopidogrel versus placebo in symptomatic patients without diabetic nephropathy (n ⫽ 11,624) was associated with a significant reduction in the combined primary end point of CV death, MI, and stroke (6.7% vs 7.7%; p ⫽ 0.048), although there was no significant difference in CV (2.6% vs 3.0%; p ⫽ 0.184) or overall mortality rates (4.3% vs 4.9%; p ⫽ 0.130). Symptomatic patients with diabetic nephropathy (n ⫽ 529) experienced no significant difference in results when randomly assigned to clopidogrel versus placebo (primary combined end point, 11.4% vs 12.0%, p ⫽ 0.837; CV death, 7.5% vs 5.1%, p ⫽ 0.245; and overall mortality, 10.2% vs 6.9%, p ⫽ 0.161). The frequency of bleeding in patients with diabetic nephropathy assigned to clopidogrel tended to be higher compared with the placebo-treated diabetic nephropathy group, but this increase was not statistically significant. GUSTO severe bleeding occurred in 2.6% of patients with nephropathy randomly assigned to clopidogrel versus 1.5% of patients with nephropathy randomly assigned to placebo (HR 1.8, p ⫽ 0.075; Table 3). In patients without diabetic nephropathy, there was also no significant difference in GUSTO severe bleeding rates between patients randomly assigned to clopidogrel (1.5%) versus placebo (1.3%; p ⫽ 0.282). When patients with diabetic nephropathy who had a severe or moderate bleeding event were excluded from the analysis of
1362
The American Journal of Cardiology (www.AJConline.org)
Table 3 Efficacy and safety end points in patients with diabetic nephropathy based on drug assignment Variable
Clopidogrel (n ⫽ 1,006)
Placebo (n ⫽ 1,003)
p Value
HR (95% CI)
Overall death CV death Overall CV death/MI/stroke Nonfatal MI Nonfatal stroke Overall CV death/MI/stroke/hospitalization Hospitalization GUSTO severe bleeding GUSTO moderate bleeding
73 (7.3%) 51 (5.1%) 85 (8.4%) 22 (2.2%) 20 (2.0%) 166 (16.5%) 97 (9.6%) 26 (2.6%) 28 (2.8%)
45 (4.5%) 31 (3.1%) 75 (7.5%) 29 (2.9%) 22 (2.2%) 161 (16.1%) 104 (10.4%) 15 (1.5%) 24 (2.4%)
0.008 0.023 0.405 0.347 0.766 0.784 0.634 0.075 0.543
1.6 (1.1–2.4) 1.7 (1.1–2.6) 1.1 (0.8–1.6) 0.8 (0.4–1.3) 0.9 (0.5–1.7) 1.0 (0.8–1.3) 0.9 (0.7–1.2) 1.8 (0.9–3.3) 1.2 (0.7–2.0)
Discussion
Figure 2. Kaplan-Meier estimates of cumulative incidence of CV mortality. *Kaplan-Meier curves truncated to 30 months. Clop ⫽ clopidogrel; DN ⫽ diabetic nephropathy; Plac ⫽ placebo.
Figure 3. Forest plot of the effect of drug assignment and diabetic nephropathy status on the primary end point and overall mortality.
mortality, adjusted HRs for overall and CV mortality in patients randomly assigned to clopidogrel compared with placebo were essentially unchanged (overall mortality HR 1.7, 95% CI 1.1 to 2.7, p ⫽ 0.014; CV mortality HR 1.8, 95% CI 1.1 to 3.1, p ⫽ 0.028, respectively).
This post hoc analysis of a large-scale blinded randomized trial evaluated whether the presence of diabetic nephropathy influenced the clinical benefit of dual-antiplatelet therapy with clopidogrel and aspirin versus aspirin alone for the prevention of CV events. Surprisingly, clopidogrel was associated with an increase in overall and CV mortality in such patients, although patients with diabetic nephropathy were at significantly higher risk of CV events and might have been predicted to derive greater benefit from more intense antiplatelet therapy, rather than harm. The interaction between randomized therapy and diabetic nephropathy was highly significant and, given the large percentage of patients with diabetic nephropathy in the asymptomatic cohort, may in part explain the surprising finding of increased mortality in asymptomatic patients randomly assigned to clopidogrel described in the CHARISMA trial. This increase in mortality was not caused by a significant increase in bleeding risks in patients with nephropathy receiving clopidogrel, thus suggesting an independent negative interaction. One limitation to the findings of this analysis was the use of GUSTO criteria to assess bleeding rates. The GUSTO definition of bleeding is relatively insensitive to all except very significant events. Therefore, although there was no statistically significant increase in bleeding rates in clopidogrel-treated patients with diabetic nephropathy compared with placebo-treated patients with diabetic nephropathy, there was a trend to more severe bleeding in the clopidogrel group, with a 1.1% absolute increase. Thus, it is possible that unrecorded bleeding events could have contributed to the observed excess in mortality in clopidogrel-treated patients. Another limitation was the lack of quantitative assessment of the degree of renal dysfunction in both the diabetic and nondiabetic populations. This was caused by the lack of collection of creatinine values at the time of enrollment into the CHARISMA trial. This precluded the ability to calculate estimated glomerular filtration rates in either population and limited the complete analysis of the negative interaction between nephropathy and clopidogrel use. Additionally, no quantitative data regarding amount of proteinuria in patients with nephropathy were collected prospectively in the original trial. Nevertheless, all patients enrolled in the trial had the presence or absence of diabetic nephropathy docu-
Coronary Artery Disease/Clopidogrel in Nephropathy
mented at the time of enrollment, and although the degree of nephropathy was not quantified, by either creatinine or degree of proteinuria the highly significant negative interaction between nephropathy and mortality suggests the existence of a relation that deserved further investigation. Similarly, hemoglobin A1c also was not recorded at the time of patient enrollment into CHARISMA, but the results of this analysis suggests that nephropathy status and not diabetes alone was related to increased mortality in patients assigned to clopidogrel. Therefore, it seems unlikely that this extra information would have added to the interaction analysis. Interestingly, although clopidogrel-treated patients with diabetic nephropathy had significantly increased CV mortality, the other components of the trial’s primary end point, MI and stroke, were not significantly increased in patients with nephropathy randomly assigned to clopidogrel. If this increase in CV mortality was secondary to increased risk of thrombotic events, it would be expected that a similar trend toward an increase in MI and stroke would have been identified. Because it was not, this increased the likelihood that the finding of increased CV mortality was a spurious finding. However, randomized blinded placebocontrolled trials of the oral glycoprotein IIb/IIIa antagonists also found a significant increase in mortality, but not other thrombotic CV end points,7 without a clear pathophysiologic explanation. Therefore, it was also possible that this finding was real. Although it was possible that the negative effects of clopidogrel and diabetic nephropathy found in this study were caused by chance, a similar relation between CKD and the diminishing benefit of dual-antiplatelet therapy was described in another study, which suggested a potentially important clinical relation.3 Although a post hoc analysis of the Clopidogrel in Unstable Angina to prevent Recurrent Events (CURE) found no significant relation between randomized therapy and renal function, when renal function was broken down by population tertiles of calculated glomerular filtration rate, they found the lowest benefit of randomization to clopidogrel therapy in patients in the lowest tertile of renal function.4 If the relation between renal dysfunction and clopidogrel was real, its pathophysiologic explanation was difficult to postulate. One possible explanation was that clopidogrel was a prodrug that required conversion to its active metabolite through cytochrome P-450 in sequential steps.8 Studies have shown that patients with renal failure had lower metabolic clearance of many drugs,9,10 especially those metabolized by the liver through the cytochrome P-450 pathway.11 Correlation existed between severity of renal failure and degree of decrease in metabolism of the drugs.12 This provided a mechanism by which there may be a reduced plasma level of the active metabolite of clopidogrel in patients with renal dysfunction, and these patients may thus derive less benefit than patients without renal dysfunction. Although this provided a framework for why patients with renal dysfunction might not derive benefit from clopidogrel, it did not suggest why these patients should experience harm with clopidogrel versus placebo.
1363
Another explanation may lie with possible pleiotropic effects of clopidogrel. Studies have reported differing results when examining effects of clopidogrel on inflammation. Some studies suggested a proinflammatory effect and others suggested an anti-inflammatory effect.13–15 The true impact of clopidogrel may ordinarily be balanced by attenuation of platelet activity through P2Y12 antagonism versus regulation of chemokines. In patients with renal dysfunction, the beneficial effects of clopidogrel may be reduced and outweighed by increased chemokine activation, with a net effect of increased adverse outcomes. 1. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329 –1339. 2. Gawaz M, Bogner C. Changes in platelet membrane glycoproteins and platelet-leukocyte interaction during hemodialysis. Clin Invest 1994; 72:424 – 429. 3. Best PJ, Steinhubl SR, Berger PB, Dasgupta A, Brennan DM, Szczech LA, Califf RM, Topol EJ. The efficacy and safety of long-term dual antiplatelet therapy in patients with mild or moderate chronic kidney disease: results from the Clopidogrel for the Reduction of Events During Observation Trial (CREDO) Trial. Am Heart J 2008;155:687– 693. 4. Keltai M, Tonelli M, Mann JF, Sitkei E, Lewis BS, Hawken S, Mehta SR, Yusuf S; CURE Trial Investigators. Renal function and outcomes in acute coronary syndrome: impact of clopidogrel. Eur J Cardiovasc Prev Rehabil 2007;14:312–318. 5. Bhatt DL, Fox KA, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706 – 1717. 6. The GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673– 682. 7. Cannon CP, McCabe CH, Wilcox RG, Langer A, Caspi A, Berink P, Lopez-Sendon J, Toman J, Charlesworth A, Anders RJ, et al. Oral glycoprotein IIb/IIIa inhibition with orbofiban in patients with unstable coronary syndromes (OPUS-TIMI 16) Trial. Circulation 2000;102: 149 –156. 8. Farid NA, Payne CD, Small DS, Winters KJ, Ernest CS II, Brandt JT, Darstein C, Jakubowski JA, Salazar DE. Cytochrome P450 3A inhibition by ketoconazole affects prasugrel and clopidogrel pharmacokinetics and pharmacodynamics differently. Clin Pharmacol Ther 2007; 81:735–741. 9. Touchette MA, Slaughter RL. The effect of renal failure on hepatic drug clearance. DICP 1991;25:1214 –1224. 10. Talbert RL. Drug dosing in renal insufficiency. J Clin Pharmacol 1994;34:99 –110. 11. Leblond F, Guevin C, Demers C, Pellerin I, Gascon-Barre M, Pichette V. Downregulation of hepatic cytochrome P450 in chronic renal failure. J Am Soc Nephrol 2001;12:326 –332. 12. Konishi K. Pharmacokinetics of cefmenoxime in patients with impaired renal function and in those undergoing hemodialysis. Antimicrob Agents Chemother 1986;30:901–905. 13. Waehre T, Damås JK, Pedersen TM, Gullestad L, Yndestad A, Andreassen AK, Frøland SS, Semb AG, Hansteen V, Gjertsen E, et al. Clopidogrel increases expression of chemokines in peripheral blood mononuclear cells in patients with coronary artery disease: results of a double-blind placebo-controlled study. J Thromb Haemost 2006;4: 2140 –2147. 14. Steinhubl SR, Badimon JJ, Bhatt DL, Herbert JM, Luscher TF. Clinical evidence for anti-inflammatory effects of antiplatelet therapy in patients with atherothrombotic disease. Vasc Med 2007;12:113–122. 15. Szalai C, Duba J, Prohászka Z, Kalina A, Szabó T, Nagy B, Horváth L, Császár A. Involvement of polymorphisms in the chemokine system in the susceptibility for coronary artery disease (CAD). Coincidence of elevated Lp(a) and MCP-1-2518 G/G genotype in CAD patients. Atherosclerosis 2001;158:233–239.
a
Gill Heart Institute, University of Kentucky, Lexington, Kentucky; Geisinger Clinic, Danville, Pennsylvania; cThe Medicines Company, Zurich, Switzerland; dVA Boston Healthcare System and Brigham and Women’s Hospital, Boston, Massachusetts; eDepartment of Cardiovascular Medicine, Cleveland Clinic, Cleveland, Ohio; fGleneagles Medical Centre, Singapore; gUniversity and Royal Infirmary of Edinburgh, Edinburgh, Scotland; hInstitut de Cardiologie-CHU Pitie-Salpetriere, Paris, France; i SUNY Downstate Medical Center, Brooklyn, New York; jUniversity of Texas Health Science Center at San Antonio, San Antonio, Texas; kHygeia Hospital, Athens, Greece; lHopital Bichat, Paris, France; and mScripps Translational Science Institute, Scripps Clinic, La Jolla, California. Manuscript received October 24, 2008; revised manuscript received and accepted January 23, 2009. The CHARISMA trial was supported by the Sanofi Aventis, Bridgewater, New Jersey/Bristol-Myers Squibb, New York, New York, partnership. *Corresponding author: Tel: 859-323-8040; fax: 859-323-6475. E-mail address: [email protected] (A. Dasgupta). b
0002-9149/09/$ – see front matter © 2009 Published by Elsevier Inc. doi:10.1016/j.amjcard.2009.01.342
Clopidogrel exerts its beneficial effect on platelets through irreversible binding of its active metabolite to the P2Y12 receptor on platelets.1 Although the pharmacokinetics of this active metabolite of clopidogrel is poorly described, it is known that approximately half the clopidogrel labeled with radioactive carbon is excreted in urine, and early studies found higher levels of the major, but still inactive, metabolite in the blood of patients with severe kidney disease. Also, because renal impairment can cause changes in platelet surface proteins,2 it is possible that the effect of clopidogrel in patients with chronic kidney disease (CKD) may be different from that in patients with normal kidney function. One post hoc analysis of a randomized trial comparing clopidogrel plus aspirin with aspirin alone reported diminished efficacy and a nonsignificant trend toward harm in patients with CKD randomly assigned to dual-antiplatelet therapy.3 However, these results were not www.AJConline.org
1360
The American Journal of Cardiology (www.AJConline.org)
Figure 1. Post hoc analysis design. ASA ⫽ aspirin.
supported by a larger post hoc analysis of a different trial that showed no excess mortality risk with clopidogrel use in patients with CKD, even when analyzed by tertiles of renal function.4 The aim of this analysis was to further investigate the relation between CKD and long-term clopidogrel administration through analysis of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial. Methods Objectives and methods of the CHARISMA trial have been previously described.5 In brief, the CHARISMA trial randomly assigned 15,603 patients with clinically evident cardiovascular (CV) disease (symptomatic patients) or multiple atherothrombotic risk factors for CV disease (asymptomatic patients) to receive either clopidogrel 75 mg/day or matching placebo in addition to daily aspirin (by protocol 75 to 162 mg). Patients had no evidence of active acute coronary syndrome at the time of initial enrollment. Patients were followed up for a median of 28 months to assess the primary efficacy end point of first occurrence of myocardial infarction (MI), stroke, or CV death. The secondary efficacy end point was CV death, MI, stroke, and hospitalization for acute coronary syndrome. Patients were also followed up for specific safety end points regarding bleeding, which was characterized using the Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) criteria.6 The CHARISMA case report form used for data collection included identification of patients with a history of diabetic nephropathy (n ⫽ 2,009; 12.9%). Although creatinine was not recorded, preventing calculation of estimated creatinine clearance, the presence of diabetic nephropathy was 1 of the entry criteria for enrollment in the study, and its presence or absence at baseline was documented in all patients. The protocol requested that all patients with diabetes and microalbuminuria with albumin ⱖ30 g/mL be considered to have diabetic nephropathy. In our post hoc analysis, outcomes with respect to treatment with clopidogrel or placebo were compared for patients with diabetic nephropathy (n ⫽ 2,009), diabetic patients without nephropathy (n ⫽ 4,546), and nondiabetic patients (n ⫽ 9,047; Figure 1). End points of interest included the prespecified primary and secondary efficacy end points, individual thrombotic outcomes, and bleeding outcomes. Statistical comparisons of baseline demographic and clinical characteristics of the study drug groups (clopidogrel
and placebo) were performed using Pearson’s chi-square test for categorical data and 2-sample t test for continuous data. The Kaplan-Meier method was used to estimate event rates of clinical outcomes. Differences in event rates were measured using log-rank test. Cox proportional-hazards models were created to evaluate the association of diabetic nephropathy and randomized therapy without adjusting for covariates and by adjusting for baseline characteristics and medical history, respectively. Model selection in multivariable Cox regressions was validated by bootstrap resampling. A 2-sided p value ⬍0.05 was considered statistically significant. All analyses were performed using SAS software, version 8.2 (SAS Institute, Cary, North Carolina). Results Almost all CV events occurred significantly more frequently in diabetic patients with nephropathy (n ⫽ 2,009) and diabetic patients without nephropathy (n ⫽ 4,546) compared with nondiabetic patients (n ⫽ 9,047; Table 1). Patients with diabetic nephropathy had a higher case fatality rate of MI compared with diabetic patients without nephropathy and nondiabetic patients (20% vs 14% and 11%, respectively), but this higher rate was not statistically significant (p ⫽ 0.240). Baseline characteristics of patients with diabetic nephropathy assigned to clopidogrel (n ⫽ 1,006) versus placebo (n ⫽ 1,003) are listed in Table 2. There were no significant differences in demographic characteristics between the 2 groups. Rates of primary, secondary, individual thrombotic, and safety end points for patients with diabetic nephropathy based on study drug assignment are listed in Table 3. Patients with nephropathy randomly assigned to clopidogrel experienced a significant increase in overall mortality compared with placebo patients, as well as significantly increased CV mortality. A multivariable interaction test confirmed the significant increase in both overall and CV mortality in patients with diabetic nephropathy treated with clopidogrel versus placebo, with adjusted hazard ratios (HRs) of 1.8 (95% confidence interval [CI] 1.2 to 2.7, p ⫽ 0.006) and 1.7 (95% CI 1.1 to 2.9, p ⫽ 0.028), respectively. Kaplan-Meier methods were used to estimate event curves for CV mortality and showed an early increase in events in patients with nephropathy randomly assigned to clopidogrel that appeared to continue to widen with increasing duration of therapy (Figure 2). A formal interaction analysis found a significant interaction of diabetic nephropathy and clopidogrel on risk of CV death with an HR of 1.856 (95% CI 1.106 to 3.116, p ⫽ 0.019). Figure 3 shows the interaction of study drug assignment on rates of the primary end point (CV mortality, MI, and stroke) and overall mortality in patients with diabetic nephropathy, diabetic patients without nephropathy, and nondiabetic patients. Patients with diabetic nephropathy randomly assigned to clopidogrel experienced significantly increased overall mortality. No similar hazard was found in diabetic patients without nephropathy assigned to clopidogrel or nondiabetic patients assigned to clopidogrel. The CHARISMA trial found that the asymptomatic cohort (n ⫽ 3,284) had a significant increase in overall mortality in patients assigned to clopidogrel compared with
Coronary Artery Disease/Clopidogrel in Nephropathy
1361
Table 1 Efficacy end points in the CHARISMA patient cohort (n ⫽ 15,603) Outcome
Nondiabetic (n ⫽ 9,047)
Diabetic Without Nephropathy (n ⫽ 4,546)
Diabetic Nephropathy (n ⫽ 2,009)
p Value
Overall CV death/MI/stroke
569 (6.3%) 223 (2.5%)
Nonfatal MI
170 (1.9%)
Nonfatal stroke
221 (2.4%)
Overall death
379 (4.2%)
Overall MI
192 (2.1%)
160 (8.0%) 1.3 (1.1–1.6) 82 (4.1%) 1.7 (1.3–2.2) 51 (2.5%) 1.4 (1.0–1.9) 42 (2.1%) 0.9 (0.6–1.2) 118 (5.9%) 1.5 (1.2–1.8) 60 (3.0%) 1.4 (1.1–1.9) 12 (0.6%) 2.5 (1.2–5.1)
⬍0.001
CV death
378 (8.3%) 1.4 (1.2–1.5) 162 (3.6%) 1.5 (1.2–1.8) 121 (2.7%) 1.4 (1.1–1.8) 132 (2.9%) 1.2 (1.0–1.5) 248 (5.5%) 1.3 (1.1–1.6) 134 (2.9%) 1.4 (1.1–1.8) 19 (0.4%) 1.7 (0.9–3.2)
Fatal MI
22 (0.2%)
⬍0.001 0.004 0.110 ⬍0.001 0.002 0.021
Values expressed as number (percent) and HR (95% CI). Table 2 Baseline characteristics of patients with diabetic nephropathy (n ⫽ 2,009) randomly assigned to clopidogrel versus placebo Variable
Placebo Clopidogrel p Value (n ⫽ 1,003) (n ⫽ 1,006)
Age (yrs) Women White Hispanic Asian Black Other Smoking status Current Former Hypertension Hypercholesterolemia Congestive heart failure Previous MI Atrial fibrillation Previous stroke Previous transient ischemic attack Peripheral arterial disease Previous percutaneous coronary intervention Previous coronary artery bypass grafting Previous carotid endarterectomy Previous peripheral angioplasty
63.0 63.1 326 (32.5%) 345 (34.3%) 690 (68.8%) 692 (68.8%) 154 (15.4%) 141 (14.0%) 90 (9.0%) 93 (9.2%) 31 (3.1%) 53 (5.3%) 38 (3.8%) 27 (2.7%)
0.7 0.4 0.08
0.8 157 (15.7%) 447 (44.6%) 865 (86.2%) 804 (80.2%) 68 (6.8%) 197 (19.6%) 31 (3.1%) 87 (8.7%) 43 (4.3%) 162 (16.2%) 113 (11.3%)
161 (16.0%) 458 (45.5%) 892 (88.7%) 777 (77.2%) 76 (7.6%) 183 (18.2%) 33 (3.3%) 82 (8.2%) 38 (3.8%) 148 (14.7%) 102 (10.1%)
0.1 0.1 0.5 0.4 0.8 0.7 0.6 0.4 0.4
153 (15.3%) 132 (13.1%)
0.2
29 (2.9%) 66 (6.6%)
27 (2.7%) 54 (5.4%)
0.8 0.2
those assigned to placebo (5.4% vs 3.8%; p ⫽ 0.04), as well as an increase in CV mortality (3.9% vs 2.2%, respectively; p ⫽ 0.01). Because nearly half (44%) the patients enrolled in the asymptomatic cohort had a diagnosis of diabetic nephropathy, we also evaluated outcomes in the asymptomatic cohort based on the presence of diabetic nephropathy and study drug assignment. In asymptomatic patients with diabetic nephropathy randomly assigned to clopidogrel (n ⫽ 750) compared with asymptomatic patients with diabetic nephropathy randomly assigned to placebo (n ⫽ 721), rates of CV (4.1% vs 2.2%, respectively; p ⫽ 0.037) and overall
mortality (6.1% vs 3.5%, respectively; p ⫽ 0.018) were significantly higher in the clopidogrel cohort. There was no significant difference in CV or overall mortality in asymptomatic patients without diabetic nephropathy with regard to randomized drug assignment. However, the directionality of mortality trends was similar to those with diabetic nephropathy. CV mortality was 3.6% in clopidogrel versus 2.2% in placebo (p ⫽ 0.081) and overall mortality was 4.7% in clopidogrel versus 4.1% in placebo (p ⫽ 0.539). When the symptomatic cohort (n ⫽ 12,153) was analyzed with respect to nephropathy and study drug assignment, results suggested a benefit for clopidogrel use in patients without diabetic nephropathy, but no benefit in patients with diabetic nephropathy. Randomization to clopidogrel versus placebo in symptomatic patients without diabetic nephropathy (n ⫽ 11,624) was associated with a significant reduction in the combined primary end point of CV death, MI, and stroke (6.7% vs 7.7%; p ⫽ 0.048), although there was no significant difference in CV (2.6% vs 3.0%; p ⫽ 0.184) or overall mortality rates (4.3% vs 4.9%; p ⫽ 0.130). Symptomatic patients with diabetic nephropathy (n ⫽ 529) experienced no significant difference in results when randomly assigned to clopidogrel versus placebo (primary combined end point, 11.4% vs 12.0%, p ⫽ 0.837; CV death, 7.5% vs 5.1%, p ⫽ 0.245; and overall mortality, 10.2% vs 6.9%, p ⫽ 0.161). The frequency of bleeding in patients with diabetic nephropathy assigned to clopidogrel tended to be higher compared with the placebo-treated diabetic nephropathy group, but this increase was not statistically significant. GUSTO severe bleeding occurred in 2.6% of patients with nephropathy randomly assigned to clopidogrel versus 1.5% of patients with nephropathy randomly assigned to placebo (HR 1.8, p ⫽ 0.075; Table 3). In patients without diabetic nephropathy, there was also no significant difference in GUSTO severe bleeding rates between patients randomly assigned to clopidogrel (1.5%) versus placebo (1.3%; p ⫽ 0.282). When patients with diabetic nephropathy who had a severe or moderate bleeding event were excluded from the analysis of
1362
The American Journal of Cardiology (www.AJConline.org)
Table 3 Efficacy and safety end points in patients with diabetic nephropathy based on drug assignment Variable
Clopidogrel (n ⫽ 1,006)
Placebo (n ⫽ 1,003)
p Value
HR (95% CI)
Overall death CV death Overall CV death/MI/stroke Nonfatal MI Nonfatal stroke Overall CV death/MI/stroke/hospitalization Hospitalization GUSTO severe bleeding GUSTO moderate bleeding
73 (7.3%) 51 (5.1%) 85 (8.4%) 22 (2.2%) 20 (2.0%) 166 (16.5%) 97 (9.6%) 26 (2.6%) 28 (2.8%)
45 (4.5%) 31 (3.1%) 75 (7.5%) 29 (2.9%) 22 (2.2%) 161 (16.1%) 104 (10.4%) 15 (1.5%) 24 (2.4%)
0.008 0.023 0.405 0.347 0.766 0.784 0.634 0.075 0.543
1.6 (1.1–2.4) 1.7 (1.1–2.6) 1.1 (0.8–1.6) 0.8 (0.4–1.3) 0.9 (0.5–1.7) 1.0 (0.8–1.3) 0.9 (0.7–1.2) 1.8 (0.9–3.3) 1.2 (0.7–2.0)
Discussion
Figure 2. Kaplan-Meier estimates of cumulative incidence of CV mortality. *Kaplan-Meier curves truncated to 30 months. Clop ⫽ clopidogrel; DN ⫽ diabetic nephropathy; Plac ⫽ placebo.
Figure 3. Forest plot of the effect of drug assignment and diabetic nephropathy status on the primary end point and overall mortality.
mortality, adjusted HRs for overall and CV mortality in patients randomly assigned to clopidogrel compared with placebo were essentially unchanged (overall mortality HR 1.7, 95% CI 1.1 to 2.7, p ⫽ 0.014; CV mortality HR 1.8, 95% CI 1.1 to 3.1, p ⫽ 0.028, respectively).
This post hoc analysis of a large-scale blinded randomized trial evaluated whether the presence of diabetic nephropathy influenced the clinical benefit of dual-antiplatelet therapy with clopidogrel and aspirin versus aspirin alone for the prevention of CV events. Surprisingly, clopidogrel was associated with an increase in overall and CV mortality in such patients, although patients with diabetic nephropathy were at significantly higher risk of CV events and might have been predicted to derive greater benefit from more intense antiplatelet therapy, rather than harm. The interaction between randomized therapy and diabetic nephropathy was highly significant and, given the large percentage of patients with diabetic nephropathy in the asymptomatic cohort, may in part explain the surprising finding of increased mortality in asymptomatic patients randomly assigned to clopidogrel described in the CHARISMA trial. This increase in mortality was not caused by a significant increase in bleeding risks in patients with nephropathy receiving clopidogrel, thus suggesting an independent negative interaction. One limitation to the findings of this analysis was the use of GUSTO criteria to assess bleeding rates. The GUSTO definition of bleeding is relatively insensitive to all except very significant events. Therefore, although there was no statistically significant increase in bleeding rates in clopidogrel-treated patients with diabetic nephropathy compared with placebo-treated patients with diabetic nephropathy, there was a trend to more severe bleeding in the clopidogrel group, with a 1.1% absolute increase. Thus, it is possible that unrecorded bleeding events could have contributed to the observed excess in mortality in clopidogrel-treated patients. Another limitation was the lack of quantitative assessment of the degree of renal dysfunction in both the diabetic and nondiabetic populations. This was caused by the lack of collection of creatinine values at the time of enrollment into the CHARISMA trial. This precluded the ability to calculate estimated glomerular filtration rates in either population and limited the complete analysis of the negative interaction between nephropathy and clopidogrel use. Additionally, no quantitative data regarding amount of proteinuria in patients with nephropathy were collected prospectively in the original trial. Nevertheless, all patients enrolled in the trial had the presence or absence of diabetic nephropathy docu-
Coronary Artery Disease/Clopidogrel in Nephropathy
mented at the time of enrollment, and although the degree of nephropathy was not quantified, by either creatinine or degree of proteinuria the highly significant negative interaction between nephropathy and mortality suggests the existence of a relation that deserved further investigation. Similarly, hemoglobin A1c also was not recorded at the time of patient enrollment into CHARISMA, but the results of this analysis suggests that nephropathy status and not diabetes alone was related to increased mortality in patients assigned to clopidogrel. Therefore, it seems unlikely that this extra information would have added to the interaction analysis. Interestingly, although clopidogrel-treated patients with diabetic nephropathy had significantly increased CV mortality, the other components of the trial’s primary end point, MI and stroke, were not significantly increased in patients with nephropathy randomly assigned to clopidogrel. If this increase in CV mortality was secondary to increased risk of thrombotic events, it would be expected that a similar trend toward an increase in MI and stroke would have been identified. Because it was not, this increased the likelihood that the finding of increased CV mortality was a spurious finding. However, randomized blinded placebocontrolled trials of the oral glycoprotein IIb/IIIa antagonists also found a significant increase in mortality, but not other thrombotic CV end points,7 without a clear pathophysiologic explanation. Therefore, it was also possible that this finding was real. Although it was possible that the negative effects of clopidogrel and diabetic nephropathy found in this study were caused by chance, a similar relation between CKD and the diminishing benefit of dual-antiplatelet therapy was described in another study, which suggested a potentially important clinical relation.3 Although a post hoc analysis of the Clopidogrel in Unstable Angina to prevent Recurrent Events (CURE) found no significant relation between randomized therapy and renal function, when renal function was broken down by population tertiles of calculated glomerular filtration rate, they found the lowest benefit of randomization to clopidogrel therapy in patients in the lowest tertile of renal function.4 If the relation between renal dysfunction and clopidogrel was real, its pathophysiologic explanation was difficult to postulate. One possible explanation was that clopidogrel was a prodrug that required conversion to its active metabolite through cytochrome P-450 in sequential steps.8 Studies have shown that patients with renal failure had lower metabolic clearance of many drugs,9,10 especially those metabolized by the liver through the cytochrome P-450 pathway.11 Correlation existed between severity of renal failure and degree of decrease in metabolism of the drugs.12 This provided a mechanism by which there may be a reduced plasma level of the active metabolite of clopidogrel in patients with renal dysfunction, and these patients may thus derive less benefit than patients without renal dysfunction. Although this provided a framework for why patients with renal dysfunction might not derive benefit from clopidogrel, it did not suggest why these patients should experience harm with clopidogrel versus placebo.
1363
Another explanation may lie with possible pleiotropic effects of clopidogrel. Studies have reported differing results when examining effects of clopidogrel on inflammation. Some studies suggested a proinflammatory effect and others suggested an anti-inflammatory effect.13–15 The true impact of clopidogrel may ordinarily be balanced by attenuation of platelet activity through P2Y12 antagonism versus regulation of chemokines. In patients with renal dysfunction, the beneficial effects of clopidogrel may be reduced and outweighed by increased chemokine activation, with a net effect of increased adverse outcomes. 1. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329 –1339. 2. Gawaz M, Bogner C. Changes in platelet membrane glycoproteins and platelet-leukocyte interaction during hemodialysis. Clin Invest 1994; 72:424 – 429. 3. Best PJ, Steinhubl SR, Berger PB, Dasgupta A, Brennan DM, Szczech LA, Califf RM, Topol EJ. The efficacy and safety of long-term dual antiplatelet therapy in patients with mild or moderate chronic kidney disease: results from the Clopidogrel for the Reduction of Events During Observation Trial (CREDO) Trial. Am Heart J 2008;155:687– 693. 4. Keltai M, Tonelli M, Mann JF, Sitkei E, Lewis BS, Hawken S, Mehta SR, Yusuf S; CURE Trial Investigators. Renal function and outcomes in acute coronary syndrome: impact of clopidogrel. Eur J Cardiovasc Prev Rehabil 2007;14:312–318. 5. Bhatt DL, Fox KA, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706 – 1717. 6. The GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673– 682. 7. Cannon CP, McCabe CH, Wilcox RG, Langer A, Caspi A, Berink P, Lopez-Sendon J, Toman J, Charlesworth A, Anders RJ, et al. Oral glycoprotein IIb/IIIa inhibition with orbofiban in patients with unstable coronary syndromes (OPUS-TIMI 16) Trial. Circulation 2000;102: 149 –156. 8. Farid NA, Payne CD, Small DS, Winters KJ, Ernest CS II, Brandt JT, Darstein C, Jakubowski JA, Salazar DE. Cytochrome P450 3A inhibition by ketoconazole affects prasugrel and clopidogrel pharmacokinetics and pharmacodynamics differently. Clin Pharmacol Ther 2007; 81:735–741. 9. Touchette MA, Slaughter RL. The effect of renal failure on hepatic drug clearance. DICP 1991;25:1214 –1224. 10. Talbert RL. Drug dosing in renal insufficiency. J Clin Pharmacol 1994;34:99 –110. 11. Leblond F, Guevin C, Demers C, Pellerin I, Gascon-Barre M, Pichette V. Downregulation of hepatic cytochrome P450 in chronic renal failure. J Am Soc Nephrol 2001;12:326 –332. 12. Konishi K. Pharmacokinetics of cefmenoxime in patients with impaired renal function and in those undergoing hemodialysis. Antimicrob Agents Chemother 1986;30:901–905. 13. Waehre T, Damås JK, Pedersen TM, Gullestad L, Yndestad A, Andreassen AK, Frøland SS, Semb AG, Hansteen V, Gjertsen E, et al. Clopidogrel increases expression of chemokines in peripheral blood mononuclear cells in patients with coronary artery disease: results of a double-blind placebo-controlled study. J Thromb Haemost 2006;4: 2140 –2147. 14. Steinhubl SR, Badimon JJ, Bhatt DL, Herbert JM, Luscher TF. Clinical evidence for anti-inflammatory effects of antiplatelet therapy in patients with atherothrombotic disease. Vasc Med 2007;12:113–122. 15. Szalai C, Duba J, Prohászka Z, Kalina A, Szabó T, Nagy B, Horváth L, Császár A. Involvement of polymorphisms in the chemokine system in the susceptibility for coronary artery disease (CAD). Coincidence of elevated Lp(a) and MCP-1-2518 G/G genotype in CAD patients. Atherosclerosis 2001;158:233–239.