Article Text
PDF
Abstract
Context Enhanced platelet inhibition by clopidogrel decreases the risk of ischemic events but carries a risk for a concomitant increase in bleeding.
Objectives To compare the efficacy and safety of two clopidogrel loading regimens (300mg vs. 600mg) in patients undergoing percutaneous coronary intervention (PCI) at one month after start of therapy.
Data sources A systematic literature search of MEDLINE, EMBASE, CENTRAL, and Web of Science databases using predefined search terms for relevant articles in any language.
Study selection and data extraction Randomised controlled trials and non-randomised studies reporting adjusted effect estimates were included. Summary estimates of the risks ratios (RRs) with therapy were calculated using a random-effect model. Outcomes evaluated were combined major adverse cardiovascular events (MACE) and major bleedings.
Results Seven studies met the inclusion criteria and included 25,383 patients. A 600mg clopidogrel loading was associated with a 34% relative risk reduction of MACE (RR=0.66; 95% confidence intervals CI=0.52-0.84; p<0.001). Sub-analysis revealed a 47% risk reduction of MACE in randomised trials (RR=0.53; 95%CI=0.32-0.88; p=0.01) and a 31% relative risk reduction in non-randomised trials (RR=0.69; 95%CI=0.54-0.90; p=0.005) in patients receiving 600mg loading with clopidogrel. In patients suffering from acute coronary syndrome, 600mg clopidogrel loading was associated with a 24% relative risk reduction in MACE (RR=0.76; 95%CI=0.60-0.95; p=0.02). Importantly, the 600mg clopidogrel loading dose was not associated with an increased risk of major bleedings (RR=0.91; 95%CI=0.73-1.15; p=0.44).
Conclusions This meta-analysis demonstrates that intensified clopidogrel loading with 600mg reduces the rate of major cardiovascular events without increase in major bleeding compared to 300mg in patients undergoing PCI during one month follow-up.
- Clopidogrel
- loading dose
- MACE
- bleeding
- PCI
- platelets
- coronary intervention
- coronary artery disease (CAD)
- pharmacokinetics/phramacodynamics
- epidemiology
Statistics from Altmetric.com
- Clopidogrel
- loading dose
- MACE
- bleeding
- PCI
- platelets
- coronary intervention
- coronary artery disease (CAD)
- pharmacokinetics/phramacodynamics
- epidemiology
Introduction
Rapid and sufficient platelet inhibition is a therapeutic goal in the management of patients undergoing percutaneous coronary intervention (PCI), especially in those suffering from an acute coronary syndrome (ACS) with the need for acute PCI.1 Addition of the oral ADP receptor antagonist clopidogrel to aspirin has become a standard in those patients.2 As clopidogrel inhibits platelets in a dose-dependent manner, dosing of clopidogrel is an important variable influencing the patient's outcome.3–6 The standard clopidogrel loading and maintenance doses (300 mg/75 mg) were chosen on the basis of dose-finding studies in healthy volunteers,7–9 but without appropriate dose finding in the target population. Since initial reports showed insufficient platelet inhibition by clopidogrel in patients treated with 300/75 mg of clopidogrel,10 11 several studies have been performed to investigate whether higher doses of clopidogrel might inhibit platelet aggregation more extensively, and whether this has an impact on major adverse cardiac events or major bleedings.3–6 As several recent studies comparing these two clopidogrel loading regimens provide additional data, we performed an up-to-date systematic review of all currently available trials. As enhanced platelet inhibition carries a risk for a concomitant increase in bleeding, we were particularly interested in whether a 600 mg loading dose of clopidogrel was associated with a higher incidence of bleeding events. In this respect, we performed a meta-analysis, in which we compared the safety and efficacy of standard clopidogrel loading dose (300 mg) versus higher dose (600 mg) in patients undergoing PCI during a 1-month follow-up.
Methods
This meta-analysis was performed and reported according to established methods.12 We searched MEDLINE, EMBASE, CENTRAL and Web of Science using predefined search terms (‘clopidogrel’ and ‘dose’ or ‘dosing’ or ‘dosage’) until March 2009. We checked the references of retrieved studies for additional trials. We used no language restrictions. Both full-text articles and meeting abstracts were included. Preliminarily relevant citations were screened at the title and abstract level, and retrieved, if potentially pertinent, as full reports. Eligible reports were assessed for methodological quality. To be included, selected studies had to be randomised controlled trials or non-randomised studies reporting adjusted effect estimates. Inclusion criteria were: patients should use clopidogrel; patients should undergo percutaneous coronary intervention (PCI); the duration of follow-up should be 1 month, and the study should supply data on the clopidogrel loading dose, rate of major adverse coronary events (MACE) and/or rate of bleedings, and follow-up duration. Non-randomised studies reporting the patient's outcome after two different clopidogrel dosing regimens had to report adjusted effects to be included. According to the common epidemiological understanding, we assumed that the non-randomised studies are characterised by a better generalisability and the more complete description of adverse events compared with randomised trials. However, the non-randomised trials were small in size and did not contribute to the estimates of side effects, but it was an a priori decision to include the non-randomised trials, and we did not want to risk bias by excluding studies driven by the effect. Nonetheless, we explicitly investigated potential influences of the study design. Exclusion criteria were duplicate reports, the lack of control group and ongoing studies.
Variables are reported as number or percentages as appropriate. Risk ratios (RR) were computed from individual studies and pooled according to the inverse variance random effect method with 95% CIs using RevMan version 5.0 (The Cochrane Collaboration). When non-randomised studies reported adjusted odds ratios (OR), OR were converted to risk ratios (RR) according to ‘the Cochrane Handbook for Systematic Reviews of Interventions’: RR=(OR/1–ACRx(1–OR)), where ACR denotes assumed control risk. Adjusted hazard ratios were accepted as risk ratios. We performed a sensitivity analysis to prove the robustness of effect estimates against within-study bias risk. Sensitivity analyses were conducted using random effect estimates and RRs and by (1) excluding sequentially one single study; (2) adding subgroups: randomised trials with low bias risk, randomised trials with high bias risk and non-randomised trials; and (3) calculating fixed-effect models to identify small study effects (when heterogeneity <50%). Six variables were included for assessment of bias risk: adequate sequence generation, allocation concealment, blinding to treatment arm, blinded treatment assessment, intention to treat analysis and incomplete outcome data assessment. Trials without blinded outcome assessment and/or blinding to treatment arm were assumed as studies with a high bias risk. In addition, numbers needed to treat were computed using the inverse of underlying risk as indicated by the control event rate (only data from randomised trials were used). We assessed the results for heterogeneity by examining the forest plots and then calculating a Q statistic, which we compared with a χ2 distribution, and the I2 index. A small study effects and/or reporting bias was appraised by graphic inspection of funnel plots and quantified by using the Egger regression test to calculate two-tailed p values. Unadjusted p values are reported throughout, with hypothesis testing set at the two-tailed significance level of 0.05.
Results
Study selection
Our search identified 881 reports, of which 11 studies investigating the impact of clopidogrel loading dose (300 mg vs 600 mg) reported clinical outcome during follow-up. Two non-randomised studies have not been included, as adjusted effect estimates were not reported.13 14 One randomised study15 and one non-randomised study reporting adjusted effect estimates16 have not been included, as the follow-up period was shorter or longer than 1 month. Only seven studies met our inclusion criteria: five randomised3–6 17 and two non-randomised18 19 (25 383 patients; figure 1). Details of the design of the included studies are listed in tables 1, 2.
Flow of articles through the review process.
Characteristics of included trials: study design, patient population, follow-up, loading dose
Characteristics of included studies: inclusion and exclusion criteria, lost to follow-up and period of recruitment
Impact of clopidogrel loading dose (600 vs 300 mg) on MACE
During a 1-month follow-up, 5.1% (570 of 11 148) of patients experienced a MACE in the 300 mg loading group compared with 3.8% (539 of 14 235) in the 600 mg loading group. A 600 mg clopidogrel loading reduced the risk of MACE by 22% when calculated with a fixed-effect model (risk reduction RR=0.78; CI 95% CI 0.69 to 0.88; p<0.001; I2=43%) or by 34% in the random-effect model (RR=0.66; 95% CI 0.52 to 0.84; p<0.001; I2=43%; figure 2).
Forest plot of major adverse cardiovascular events (MACE) for comparison between high clopidogrel loading dose (600 mg) and standard clopidogrel loading dose (300 mg) in patients undergoing PCI.
Sensitivity analysis: all randomised trials
A 600 mg clopidogrel loading reduced the risk of MACE by 47% compared with a 300 mg loading dose in randomised trials3–6 17 (RR=0.53; 95% CI 0.32 to 0.88; p=0.01; I2=57%; random-effect model).
The estimated number of patients who needed to be treated with a high clopidogrel loading dose, as compared with a low clopidogrel loading dose, to prevent one MACE at 1 month would be 113 (95% CI 67 to 354).
Sensitivity analysis: randomised trials with a low bias risk (trials reporting blinded outcome assessment)
Four randomised trials have been assumed as studies with a low bias risk.3–5 17 In trials with a low bias risk, a 600 mg clopidogrel loading dose was associated with a 42% RR reduction in MACE compared with a 300 mg loading dose (RR=0.58; 95% CI 0.33 to 1.00; p=0.05; I2=52%; random-effect model; figure 3).
Forest plot of major adverse cardiovascular events (MACE) for comparison between high clopidogrel loading dose (600 mg) and standard clopidogrel loading dose (300 mg) in three categories of included studies: randomised trials with low risk of bias, randomised trials with high risk of bias and non-randomised trials.
Sensitivity analysis: randomised trials with a high bias risk (trials not reporting blinded outcome assessment)
One randomised trial has been assumed as a study with a high bias risk.6 In this trial, a 600 mg clopidogrel loading dose was associated with a 61% RR reduction in MACE (RR=0.39; 95% CI 0.17 to 0.89; p=0.03; figure 3).
Sensitivity analysis: non-randomised trials
In two non-randomised studies reporting adjusted effect estimates of MACE,18 19 a higher clopidogrel loading dose was associated with a 31% RR reduction in MACE (RR=0.69; 95% CI 0.54 to 0.90; p=0.005; I2=0%; figure 3).
Test for interaction
An interaction test yielded χ2=4.05, df=2 and p=0.13, indicating no difference between the effects of the three study type categories (randomised trials with low or high bias risk and non-randomised trials). All three study subtypes showed the same direction of the effect, and the summary effect remained unchanged independent of within-study bias components, indicating that our summary effect is robust, and the presenting summary effect is justified.
Sensitivity analysis: patients presenting with ACS at admission
Four randomised trials3 6 17 19 included 20 904 patients undergoing PCI due to ACS (table 1). Overall, 5% (494 of 10 018) of patients experienced a MACE in the 300 mg loading group compared with 4% (435 of 10 886) in the 600 mg loading group. A higher clopidogrel loading dose was therefore associated with a 24% RR reduction in MACE in patients suffering from ACS (RR=0.76; 95% CI 0.60 to 0.95; p=0.02; I2=32%; random-effect model, figure 4). Using a fixed-effects model, the RR reduction after 600 mg clopidogrel loading was 19% (RR=0.81; 95% CI 0.72 to 0.92; p<0.001).
Forest plot of major adverse cardiovascular events (MACE) for comparison between high clopidogrel loading dose (600 mg) and standard clopidogrel loading dose (300 mg) in patients presenting with acute coronary syndrome (ACS) at admission.
When calculated with the random-effect model, the risk reduction in MACE in patients presenting with ACS was lower than the risk reduction in the overall study population (RR: 24% vs 34%; figures 2 and 4). In the presence of heterogeneity (43%), a random-effects meta-analysis weights the studies more equally than a fixed-effect analysis. This difference therefore might be due to the presence of small-study effects, in which the intervention effect was more beneficial, as the discrepancy in the results diminished when the fixed-effects model was used (RR: 19% vs 22%). A second possible explanation might be that patients presenting with myocardial infarction are treated with a number of drugs, including anticoagulants and high-dose aspirin, which might contribute to the lower overall effect of clopidogrel. Third, the result might be due to chance.
Impact of clopidogrel loading dose (600 vs 300 mg) on the composite of myocardial infarction and death
Five randomised studies reported the frequencies of myocardial infarction and death.3–6 17 Overall, 4.7% (425 of 9036) of patients experienced a myocardial infarction or death in the 300 mg loading group compared with 3.8% (346 of 8931) in the 600 mg loading group, resulting in a 48% relative risk reduction in the 600 mg group (RR=0.52; 95% CI 0.29 to 0.99; p=0.05; I2=52%; random-effect model).
Impact of clopidogrel loading dose (600 vs 300 mg) on major bleedings
All randomised trials3–6 17 reported rates of major bleedings, and one non-randomised study reported an adjusted effect estimate of major bleeding.19 Overall, 1.7% (177 of 10 189) of patients experienced a major bleeding in the 300 mg loading group compared with 1.9% (206 of 11 089) in the 600 mg loading group. Interestingly, a 600 mg clopidogrel loading dose was not associated with an increased risk of major bleedings (RR=0.91; 95% CI 0.73 to 1.15; p=0.44; I2=0%; random-effect model; figure 5). To minimise the effect of small studies, we computed the effect estimate with the fixed-effects model: the result remained unchanged (RR=0.91; 95% CI 0.73 to 1.15; p=0.44). Although the use of a 600 mg loading dose was associated with a slightly higher rate of major bleeds (1.9% vs 1.7%), the 600 mg loading dose corresponded to a 9% RR reduction in major bleedings, which was likely due to chance. The reason for this apparent discrepancy is that we weighted studies by the amount of information they contribute by the inverse variances of their effect estimates. This gives studies with more precise results (narrower confidence intervals) more weight. The pooled estimate of major bleedings was therefore driven by the results of the HORIZONS-AMI trial, representing 70.3% of weight in this analysis. After exclusion of the HORIZONS-AMI trial from the analysis, the result remained unchanged (RR=1.03; 95% CI 0.67 to 1.57).
Forest plot of major bleedings for comparison between high clopidogrel loading dose (600 mg) and standard clopidogrel loading dose (300 mg) in patients undergoing PCI.
Despite differences in the definition of major bleeds, at least three points might indicate that the statistical methods used were robust: (1) the absence of heterogeneity (0%); (2) four of seven studies reported major bleeds according to the TIMI criteria and two studies had very similar criteria (fatal, requiring intervention or transfusion; table 1); and (3) the random and fixed-effects models revealed the same results.
The estimated number of patients needed to be treated with 600 mg of clopidogrel for one additional patient to be harmed in the form of major bleeding at 1 month is 18 720 (95% CI –infinity to −500 and 474 to +infinity; calculated only with data from randomised trials).
Discussion
The current meta-analysis including 25 383 patients has important clinical implications. High clopidogrel loading (600 mg) resulted in a 34% relative risk reduction in combined major cardiovascular events without any increase in major bleedings compared with 300 mg loading.
The approved loading dose of clopidogrel is 300 mg.20 Increasing evidence, however, indicated that higher clopidogrel doses improve the antiplatelet effect and patients’ outcome, which is due to the higher absolute level of platelet inhibition and more rapid onset of action.4 21 22 The recent STEMI guidelines of the European Society of Cardiology therefore recommend a 600 mg loading dose in patients undergoing PCI.23 Our meta-analysis supports and strengthens these recommendations of the European Society of Cardiology23 and American Heart Association24 on the use of intensified clopidogrel loading in patients undergoing PCI. There are no recommendations regarding loading with 600 mg of clopidogrel in elective patients, which leaves the choice of loading dose in those patients at the discretion of the interventionalist. Unfortunately, as none of included studies selectively investigated whether patients undergoing elective PCI might benefit from a higher clopidogrel loading dose, our meta-analysis does not provide any additional information in regard to this patient population. The ISAR-SWEET trial provides interesting data regarding clopidogrel loading regimen in patients undergoing elective PCI.25 The ISAR-SWEET trial demonstrated that abciximab did not reduce the rates of death and myocardial infarction in patients with diabetes after pretreatment with 600 mg of clopidogrel.25 Interestingly, the incidence of the primary end point in the group receiving 600 mg of clopidogrel without abciximab was lower (8.6%)25 than that previously reported with 300 mg of clopidogrel (10.6–14.8%).26 These findings suggests that pretreatment with 600 mg of clopidogrel at least 2 h before the procedure may obviate the need for abciximab during elective percutaneous coronary interventions in patients at low to intermediate risk.25
This meta-analysis shows that there was no difference in major bleedings between both dose regimens (figure 5). In contrast to our findings, concerning safety issues under a high clopidogrel loading dose, aspirin has been shown to enhance bleeding risk with increasing doses without any additional effects in efficacy.27 As haemorrhagic complications are associated with increased morbidity and mortality,28 the results of this meta-analysis showing no increase in bleeding risk with 600 mg clopidogrel loading have important clinical implications.
Despite the large relative effect of high-dose clopidogrel, the number needed to treat (NNT) of 113 is mostly determined by the relatively low baseline risk. However, compared with well-established interventions such as angiotensin-converting enzyme inhibitors for prevention of cardiovascular death or myocardial infarction in patients with heart failure (NNT=118)29 or antiplatelet drugs (mostly aspirin) to avoid death in patients after myocardial infarction (NNT=153),30 31 the absolute effect of clopidogrel can be regarded as reasonably beneficial.
The results of our meta-analysis addressing clopidogrel dosing are in line with those already published.32 Our meta-analysis, however, differs from the earlier analysis in several ways. First, we were able to include several recently published studies targeted at 600 mg clopidogrel loading that enrolled a large number of patients undergoing PCI. For example, the CURRENT-OASIS 7 trial comprised 68% of patients included in the analysis of randomised trials (17 232/25 383). CURRENT-OASIS 7 is the largest clinical trial to evaluate different dosing regimens of clopidogrel plus aspirin in patients with an ACS managed with an early invasive strategy with PCI. Our analysis comparing clopidogrel dosing has therefore a significantly higher statistical power compared with the previous meta-analysis because it is based on a 16-fold higher number of patients. Second, our study provides additional information on the effect of higher clopidogrel dosing in a subgroup of patients with ACS. Third, we analysed clinical outcome exactly at 1 month after PCI, whereas previous study mixed trials were with a 1-month follow-up and those with in-hospital follow-up. Fourth, we included only non-randomised trials reporting adjusted effect estimates, which reduces the risk of bias. Fifth, our meta-analysis focussed on the composite of MACE, whereas the previously published meta-analysis was on the composite of myocardial infarction and death. The composite of MACE is a more comprehensive patient relevant outcome than myocardial infarction or death. Therefore, we are convinced that our meta-analysis adds useful and clinically relevant information. When comparing the previously published meta-analysis and our paper, the direction and magnitude of the risk reduction in the composite endpoint of death and myocardial infarction were similar (48% vs 46%).
Interestingly, in some patients, even 600 mg of clopidogrel does not sufficiently inhibit platelet aggregation, which mirrors the interindividual variability in the response to clopidogrel.33–36 It has been shown that despite repeated intake of 600 mg of clopidogrel given at 24 h intervals up to 2400 mg, 15% of patients remained clopidogrel non-responders.22 37 According to the recent data, a low response to clopidogrel could be overcome by the use of prasugrel,38 as prasugrel is characterised by more rapid, more consistent and greater platelet inhibition compared with clopidogrel.39–42 The higher efficacy of prasugrel over clopidogrel has been shown in the TRITON-TIMI 38 trial, comparing prasugrel (60 mg/10 mg) with low-dose clopidogrel (300 mg/75 mg) in 13 608 patients with ACS undergoing PCI.43 Use of prasugrel was associated with a 19% RR reduction in the primary end point (death from cardiovascular causes, non-fatal myocardial infarction or non-fatal stroke) compared with clopidogrel (prasugrel: 9.9% vs clopidogrel: 12.1%).43 The ongoing TRIGGER PCI trial will answer the question of whether switching to prasugrel in patients with an insufficient platelet inhibition after loading with 600 mg of clopidogrel will reduce the rate of adverse cardiovascular outcomes in patients undergoing PCI.
Future objectives for the scientific community working on guidelines regarding the use of clopidogrel would be to (1) formulate a recommendation on the clopidogrel dosing in patients undergoing elective, acute and primary PCI, and (2) perform studies investigating the net benefit of higher clopidogrel dose regimens compared with novel platelet inhibitors.
Study limitations
This meta-analysis contains two non-randomised studies, which are more likely to contain selection, observer or publication bias and confounding, respectively. To minimise the influence of bias, we included only non-randomised trials reporting adjusted effects. Because of a number of missing studies, we could not analyse whether 600 mg clopidogrel loading improves outcome in patients undergoing elective PCI. Another important limitation is the different definition of the composite of MACE used in individual studies: all studies included death and myocardial infarction into the composite end point, and stroke and revascularisation procedures were included in 60% of studies. To minimise the heterogeneity of the MACE definition, we (1) included only end points reflecting effectiveness and (2) performed a subanalysis including only death and myocardial infarction as an end point. Similarly, the definition of major bleeds differed between included trials. Despite differences in the definition of major bleedings, at least four issues might indicate that the statistical methods used were robust: (1) the absence of heterogeneity (0%); (2) four of seven studies reported major bleeds according to the TIMI criteria, and two studies had very similar criteria (fatal, requiring intervention or transfusion; table 1); (3) the random and fixed-effects models yielded the same results; (4) effect estimates calculated with the Peto method, which can be used for calculation of effects when events are rare, revealed that the magnitude and direction of the effect remained unchanged. Moreover, although we tried to identify and include all relevant trials in our calculations, it is possible that some unpublished studies might have been missed. Considering clopidogrel dosing regimens, a double maintenance dose was used for 1 week (150 mg once daily)17 or 1 month (75 mg twice daily)4 after PCI in two included studies, which might have additionally accounted for a better outcome in higher-dose regimen groups. As the funnel plots were not ideally shaped, it is possible that some trials were selectively not published or that there are systematic differences between smaller and larger studies (‘small study effects’). To minimise the influence of small studies, we computed both fixed- and random-effect estimates.
Despite these limitations, the present meta-analysis adds important insights into the current knowledge about the dosing of clopidogrel loading.
Conclusion
In summary, our meta-analysis demonstrates that intensified clopidogrel loading with 600 mg of clopidogrel reduces the rate of major cardiovascular events without any increase in bleedings compared with 300 mg.
References
Footnotes
See Paper, p 106
Competing interests None.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.