Suetonia C. Palmer, MB ChB, PhD; Lucia Di Micco, MD; Mona Razavian, MB BS; Jonathan C. Craig, MB ChB, DCh, MM, PhD; Vlado Perkovic, MB BS, PhD; Fabio Pellegrini, MSc; Massimiliano Copetti, MSc, PhD; Giusi Graziano, MSc; Gianni Tognoni, MD; Meg Jardine, MB BS, PhD; Angela Webster, MB BS, PhD; Antonio Nicolucci, MD; Sophia Zoungas, MD, PhD; Giovanni F.M. Strippoli, MD, PhD, MPH, MM
Financial Support: No specific external funding was received for this project. Dr. Palmer received support from an unrestricted Amgen Dompé Consorzio Mario Negri fellowship.
Potential Conflicts of Interest: Dr. Palmer: Grant: Amgen Dompé. Dr. Razavian: Grant: Amgen. Dr. Perkovic: Grants/grants pending (money to institution): Heart Foundation of Australia, Johnson & Johnson, Servier, Oxford University; Board membership: Baxter; Board membership (money to institution): Boehringer Ingelheim, Vitae, Abbott, Reata and Abbott; Payment for lectures including service on speakers bureaus (money to institution): Roche. Dr. Jardine: Grant (money to institution): Royal Australasian College of Physicians. Dr. Nicolucci: Board membership: Merck Sharp & Dohme; Grants/grants pending (money to institution): Merck Sharp & Dohme, Novo Nordisk, Sanofi Aventis, Eli Lilly, Johnson & Johnson. Dr. Zoungas: Board membership: Merck Sharp & Dohme, Novo Nordisk, Boehringer Ingelheim, Sanofi Aventis, Bristol-Myers Squibb/AstraZeneca; Payment for lectures including service on speakers bureaus (money to institution): Merck Sharp & Dohme, Novo Nordisk, Sanofi Aventis, Bristol-Myers Squibb/AstraZeneca, Novartis, Servier; Payment for development of educational presentations: MediMark Australia. Disclosures can also be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M11-2512.
Reproducible Research Statement:Study protocol, statistical code, and data set: Available from Dr. Strippoli (e-mail, mailto:firstname.lastname@example.org).
Requests for Single Reprints: Giovanni F.M. Strippoli, MD, PhD, MPH, MM, Department of Clinical Pharmacology and Epidemiology, Consorzio Mario Negri Sud, Via Nazionale 8/a, 66030, Santa Maria Imbaro, Italy; e-mail, mailto:email@example.com.
Current Author Addresses: Dr. Palmer: Department of Medicine, University of Otago Christchurch, 2 Riccarton Avenue, Christchurch 8140, New Zealand.
Dr. Di Micco: Division of Nephrology, Univerity of Naples Federico II, Via Tasso 91/B, 80127 Naples, Italy.
Drs. Razavian, Perkovic, Jardine, and Zoungas: The George Institute for Global Halth, Level 10, King George V Building, 83-117 Missenden Road, Camperdown NSW 2050, Australia.
Drs. Craig and Webster: University of Sydney School of Public Health, A27, University of Sydney, NSW 2006, Australia.
Mr. Pellegrini and Dr. Copetti: Scientific Institute Casa Sollievo della Sofferenza, Unit of Biostatistics, Polioambulatorio Giovanni Paolo II, Viale Padre Pio, 71013 San Giovanni Rotondo, Foggia, Italy.
Ms. Graziano and Drs. Tognoni, Nicolucci, and Strippoli: Consorzio Mario Negri Sud, Via Nazionale 8/a, 66030, Santa Maria Imbaro, CH, Italy.
Author Contributions: Conception and design: S.C. Palmer, J.C. Craig, V. Perkovic, M. Jardine, A. Webster, G.F.M. Strippoli.
Analysis and interpretation of the data: S.C. Palmer, L. Di Micco, J.C. Craig, V. Perkovic, F. Pellegrini, M. Copetti, G. Graziano, G. Tognoni, M. Jardine, A. Webster, A. Nicolucci, S. Zoungas, G.F.M. Strippoli.
Drafting of the article: S.C. Palmer, L. Di Micco, M. Copetti, G.F.M. Strippoli.
Critical revision of the article for important intellectual content: S.C. Palmer, L. Di Micco, J.C. Craig, V. Perkovic, F. Pellegrini, M. Copetti, G. Tognoni, M. Jardine, A. Webster, S. Zoungas, G.F.M. Strippoli.
Final approval of the article: S.C. Palmer, J.C. Craig, V. Perkovic, F. Pellegrini, M. Copetti, G. Tognoni, M. Jardine, A. Webster, A. Nicolucci, S. Zoungas, G.F.M. Strippoli.
Statistical expertise: F. Pellegrini, M. Copetti, G. Graziano.
Administrative, technical, or logistic support: V. Perkovic, G.F.M. Strippoli.
Collection and assembly of data: S.C. Palmer, L. Di Micco, M. Razavian, V. Perkovic, M. Jardine, G.F.M. Strippoli.
Palmer SC, Di Micco L, Razavian M, Craig JC, Perkovic V, Pellegrini F, et al. Effects of Antiplatelet Therapy on Mortality and Cardiovascular and Bleeding Outcomes in Persons With Chronic Kidney Disease: A Systematic Review and Meta-analysis. Ann Intern Med. 2012;156:445-459. doi: 10.7326/0003-4819-156-6-201203200-00007
Download citation file:
Published: Ann Intern Med. 2012;156(6):445-459.
Antiplatelet agents are used to prevent cardiovascular events; however, treatment effects may differ in persons with chronic kidney disease (CKD) because atherosclerotic disease is less prevalent, whereas bleeding hazards may be increased in this population.
To summarize the effects of antiplatelet treatment on cardiovascular events, mortality, and bleeding in persons with CKD.
Embase and Cochrane databases through November 2011 without language restriction.
Randomized trials that included adults with CKD and compared antiplatelet agents with standard care, placebo, or no treatment.
Data for populations, interventions, outcomes, and risk for bias were extracted. Quality of evidence for treatment effects on myocardial infarction, death, and bleeding was summarized by using Grading of Recommendations Assessment, Development, and Evaluation guidelines.
Nine trials (all post hoc subgroup analyses for CKD) involving 9969 persons who had acute coronary syndromes or were undergoing percutaneous coronary intervention and 31 trials involving 11 701 persons with stable or no cardiovascular disease were identified. Low-quality evidence has found that in persons with acute coronary syndromes, glycoprotein IIb/IIIa inhibitors or clopidogrel plus standard care compared with standard care alone had little or no effect on all-cause or cardiovascular mortality or on myocardial infarction but increased serious bleeding. Compared with placebo or no treatment in persons with stable or no cardiovascular disease, antiplatelet agents prevented myocardial infarction but had uncertain effects on mortality and increased minor bleeding according to generally low-quality evidence.
Data for antiplatelet agents in persons with CKD are frequently derived from post hoc analyses of trials of broader populations. Definitions for bleeding outcomes and trial duration were heterogeneous.
Benefits for antiplatelet therapy among persons with CKD are uncertain and are potentially outweighed by bleeding hazards.
Antiplatelet agents are given for cardiovascular prevention to patients with chronic kidney disease (CKD).
This review found that glycoprotein IIb/IIIa inhibitors or clopidogrel increased major bleeding but had little or no effect on myocardial infarction, death or coronary revascularization in patients with CKD who had acute coronary syndromes or were undergoing percutaneous coronary revascularization. Antiplatelet regimens in other patients with CKD who had or were at risk for cardiovascular disease increased minor bleeding, reduced myocardial infarction, and had uncertain effects on mortality.
Evidence was weak, derived primarily from subgroup analysis of trials.
Risks of antiplatelet agents may outweigh potential benefits in some patients with CKD.
Chronic kidney disease (CKD) is an important public health challenge. Approximately 10% to 15% of the adult population worldwide has CKD, and prevalence is increasing because of the epidemics of diabetes mellitus and obesity (1). Chronic kidney disease causes illness and premature death. Nearly 50% of persons aged 70 years or older (1) and between 33% and 50% of persons with acute myocardial ischemia have CKD (2). Even in early-stage CKD, the risk for premature cardiovascular disease is increased by 25% to 30% (3) and is more than 30- to 50-fold higher in persons with end-stage kidney disease (4).
Antiplatelet agents are widely used to prevent cardiovascular events by inhibiting intravascular thrombosis. Antiplatelet drugs reduce vascular deaths by 15% and serious cardiovascular events by 20% in persons at high risk for a vascular event (5). Extrapolating these benefits of antiplatelet therapy to persons with CKD is problematic because nonatherosclerotic conditions (cardiac failure, sudden cardiac death, and arrhythmia) are more common causes of cardiovascular events in persons with CKD than in the general population (5, 6). The bleeding risk of antiplatelet agents may be greater (7) among persons with CKD because of impaired hemostasis (8).
Treating complications of CKD imposes an important economic burden. Health costs of treating a person with CKD are nearly 3-fold higher than those for a person without CKD, and the cost of treating end-stage kidney disease is 10-fold higher (9, 10). Together, increasing use of health resources, continued poor outcomes, and emergence of performance measures directed to the care of persons with CKD (11) necessitate careful evaluation of all health care interventions in this growing population. The aim of our study was to summarize the benefits and harms of antiplatelet agents in persons with CKD, focusing on cardiovascular events, mortality, and bleeding.
We conducted a systematic review based on standard methods, including a published, peer-reviewed protocol (12) and reporting in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (13).
We searched Embase from 1980 to November 2011, the Cochrane Central Register of Controlled Trials through Issue 4 of 2011, and the Cochrane Renal Group's specialized register through November 2011 without language restriction by using a search strategy designed by a specialist information manager (Appendix Table 1). The Cochrane Renal Group's register was populated by weekly Ovid MEDLINE AutoAlerts, quarterly searches from the Cochrane Central Register of Controlled Trials, and hand-searching. We contacted investigators to request unpublished data for persons who had CKD at baseline. We received and included additional data from these investigators for 3 trials that were not identified by our initial search (14–16). We screened the reference lists of retrieved publications, including a meta-analysis (5), for potentially eligible trials.
We included trials that compared antiplatelet agents with placebo, standard care, or no treatment in adults with CKD (as defined by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative criteria ) or trials of broader populations for which data for participants with CKD could be disaggregated. Pediatric trials were excluded. We excluded 5 trials reporting follow-up shorter than 2 months (18–22) because we wanted to focus on longer-term outcomes. Sensitivity analyses, including only trials with follow-up longer than 1 year, were conducted.
Two or more independent authors screened the title and abstract of retrieved citations and reviewed the full text of potentially eligible citations to identify trials that fulfilled the inclusion criteria. For included trials, we extracted data on population characteristics; interventions; and outcomes, including fatal or nonfatal myocardial infarction or stroke, death (total and cause-specific), coronary artery revascularization, major or minor bleeding (Appendix Table 2) (6, 14–16, 23–58), end-stage kidney disease, all-cause hospitalization, and treatment withdrawal. Risk of bias was assessed according to standardized methods (59).
Relative risks (RRs) and 95% CIs were calculated from the numbers of events and participants at risk for events. When crude event rates were not provided, the reported risk ratio was extracted (27). Relative risks and CIs were then summarized by using an exact bivariate random-effects meta-analysis method following the approach proposed by Stijnen and colleagues (60). Sensitivity analyses to check for robustness of our findings were performed with the “inverse of the sample size in the opposite treatment arm” (61) and the arcsine methods (62). If exact bivariate random effects could not be used because crude event data were not available, we used a Bayesian meta-analysis following Greenland's “data equivalents” approach (63).
We tested for heterogeneity of treatment effects between studies with the Cochran Q and I2 test statistics (64). Potential sources of heterogeneity in intervention effects were explored by prespecified subgroup analysis (type of antiplatelet regimen or stage of CKD [predialysis, dialysis, or kidney transplantation]) by reporting results of analyses when 4 or more studies were available for each subgroup. To assess potential bias from small study effects, we constructed funnel plots displaying the log RR on the horizontal axis and the SE of the log RR on the vertical axis. To evaluate the presence and extent of publication bias, we used the Egger regression test (65).
We conducted sensitivity analyses, excluding shorter trials (<12 months) and those published only in internal drug company reports. We summarized the quality of evidence according to the Grading of Recommendations Assessment, Development, and Evaluation guidelines (66). We conducted analyses by using SAS, version 9.1 (SAS Institute, Cary, North Carolina) (67); WinBugs, version 1.4.3 (Imperial College and Medical Research Council, Cambridge, United Kingdom); and Comprehensive Meta-analysis, version 2 (Biostat, Englewood, New Jersey). Details of the SAS macro routines based on the linear and nonlinear mixed models (namely Proc MIXED and Proc NLMIXED) are available from the authors on request.
This project received no specific funding. Dr. Palmer received support from an unrestricted Amgen Dompé Consorzio Mario Negri fellowship. The authors had full responsibility for data collection, data interpretation, and writing of the report. Drs. Palmer and Strippoli had full access to all of the data and had the final responsibility to submit the manuscript for publication.
Searching identified 1460 publications. Of these publications, 196 were reviewed in full text (Figure 1) and 40 eligible trials or patient subgroups of randomized trials (21 670 participants) were included (Appendix Table 3) (6, 14–16, 23–58). Thirty-six trials (20 942 participants) provided extractable data for inclusion in meta-analyses (6, 14–16, 23–31, 34–37, 39–45, 47–58). We included unpublished subgroup data for persons with CKD that were provided by the investigators of 12 trials (11 732 participants) (6, 14–16, 23–25, 28, 30, 37, 50, 58). Appendix Table 4 provides reasons for missing data in the meta-analyses.
Figure 1. Summary of evidence search and selection.
ACS = acute coronary syndrome; CENTRAL = Cochrane Central Register of Controlled Trials; CKD = chronic kidney disease; PCI = percutaneous coronary intervention.
* Four trials could not be included in the meta-analyses because data were not provided in an extractable format (32, 33, 38, 46).
Information for 4 trials (14, 43–45), including 2 internal company reports (44, 45), were available only in a previously published meta-analysis of antiplatelet agents (5). For 2 studies, the most complete information was provided in conference proceedings (32, 46), although neither of these trials provided extractable data for meta-analyses. Appendix Table 5 describes the sources of additional unpublished data.
Nine trials (9969 participants) provided information on antiplatelet treatment among persons with CKD who presented with an acute coronary syndrome or were scheduled to undergo coronary artery intervention and were considered at high risk for subsequent vessel closure (14–16, 23–28). All data for these trials were published (26, 27) or unpublished (14–16, 23–25, 28) post hoc analyses for the subgroup of participants with CKD from larger trials. Trials provided data for glycoprotein IIb/IIIa inhibitors (abciximab, eptifibatide, or tirofiban) (7 trials, 5471 participants) (14–16, 23–26) or clopidogrel (2 trials, 4498 participants) (27, 28), and all involved coadministration of aspirin with (14, 15, 23–26) or without (16, 27, 28) heparin as nonrandomized interventions. Median follow-up was 12 months.
The remaining 31 trials provided data for antiplatelet treatment among 11 701 persons with CKD who had stable or no cardiovascular disease. Twelve trials assessed the effects of antiplatelet agents on mortality, progression of kidney disease, or safety in 6970 patients with glomerulonephritis (4 trials, 119 participants) (29, 31, 33, 34) or diabetic nephropathy (6 trials, 2990 participants) (30, 32, 35–38) or who had an impaired glomerular filtration rate regardless of cause (2 trials, 3861 participants) (6, 58). These trials generally involved administration of aspirin (6, 30, 36, 38), dipyridamole (36), aspirin and dipyridamole (29, 31, 32, 36), or a thienopyridine (clopidogrel or ticlopidine) (34, 37). One trial involved administration of aspirin as a co-intervention (37). Median follow-up was 12 months.
Seventeen trials provided data for antiplatelet treatment in 4471 persons receiving or who would soon require dialysis (39–54, 58). These trials involved administration of a range of antiplatelet agents (aspirin, dipyridamole, clopidogrel, sulfinpyrazone, ticlopidine, or picotamide), and 3 involved administration of additional antiplatelet agents or oral anticoagulation as nonrandomized co-interventions (39, 49, 52). Trials were generally of shorter duration (median, 6 months). Four trials administered antiplatelet treatment to 260 kidney transplant recipients (55–58).
Nine trials of acute coronary syndromes or percutaneous coronary intervention generally had low risk of bias, with a high proportion reporting adequate allocation concealment (78%), intention-to-treat analysis (89%), blinding of outcome assessors (100%), and freedom from selective outcome reporting (89%). However, all were post hoc analyses of trials of broader populations. In more than 75% of the remaining 31 trials, methods for random sequence generation, allocation concealment, blinding of outcome assessors, completeness to follow-up, or the risk for selective reporting or other biases were unclear or inadequate (Appendix Figure).
Appendix Figure. Risk of bias in included studies.
Numbers of trials in each category are shown.
Figure 2 shows the overall results of all meta-analyses. The Table summarizes the quality of the available evidence (4, 66–73). Meta-analysis by using exact bivariate random effects is reported because sensitivity analyses suggested that the effect estimates were robust regardless of the statistical model used.
Figure 2. Summary of treatment effects for antiplatelet agents in persons with chronic kidney disease.
Summary estimates are provided by using random-effects meta-analysis. ACS = acute coronary syndrome; PCI = percutaneous coronary intervention.
* Number of events and number of participants from the CURE (Clopidogrel in Unstable Angina to Prevent Recurrent Events) trial (27) were not available and therefore not included in the summary totals for cardiovascular death, all-cause mortality, or major or minor bleeding.
Low-quality evidence found that antiplatelet treatment plus standard care had little or no effect on myocardial infarction (7 trials, 5261 participants; RR, 0.89 [CI, 0.76 to 1.05]), with no evidence of significant heterogeneity (Table and Figure 3). Including only trials of glycoprotein IIb/IIIa inhibitors provided similar treatment effects (6 trials, 4850 participants; RR, 0.87 [CI, 0.74 to 1.03]) (14–16, 23–25). In the 4 trials (1786 participants) reporting myocardial infarction at 1 year, the RR was 0.79 (CI, 0.37 to 1.72) with additional antiplatelet therapy (14, 15, 24, 28). One trial provided data on only 6 fatal or nonfatal strokes in 411 persons, finding that clopidogrel in addition to standard therapy had uncertain effects on stroke (RR, 0.51 [CI, 0.09 to 2.77]) (Figure 3).
Figure 3. Effect of antiplatelet agents on cardiovascular, mortality, and bleeding outcomes in persons with CKD and acute coronary syndromes or undergoing percutaneous coronary intervention.
Summary estimates are provided by using random-effects meta-analysis. Only trials reporting ≥1 event are shown. CKD = chronic kidney disease; CREDO = Clopidogrel for the Reduction of Events During Observation; CURE = Clopidogrel in Unstable Angina to Prevent Recurrent Events; EPIC = Evaluation of 7E3 for the Prevention of Ischemic Complications; EPILOG = Evaluation in PTCA to Improve Long-Term Outcome with Abciximab GP IIb/IIIa Blockade; EPISTENT = Evaluation of Platelet IIb/IIIa Inhibitor for Stenting; IMPACT-II = Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis-II; NA = not applicable; PRISM-PLUS = Platelet Receptor Inhibition in Ischemic Syndrome Management in Patients Limited by Unstable Signs and Symptoms; PURSUIT = Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy; RAPPORT = ReoPro and Primary PTCA Organization and Randomized Trial.
Low- or very low-quality evidence found that antiplatelet therapy in addition to standard treatment had little or no effect on all-cause mortality (8 trials, 9347 participants; RR, 0.89 [CI, 0.75 to 1.05]) and uncertain effects on cardiovascular mortality (2 trials, 4498 participants; RR, 0.96 [CI, 0.79 to 1.16]) without heterogeneity in the analyses (Table and Figure 3). Including only trials that evaluated glycoprotein IIb/IIIa inhibitors provided a similar effect estimate for mortality (6 trials, 4849 participants; RR, 0.86 [CI, 0.69 to 1.07]) (14–16, 23–25). No data were available for the effects of glycoprotein IIb/IIIa inhibitors on cardiovascular mortality. Limiting analyses to trials reporting outcomes at 1 year did not meaningfully alter the risk for death (5 trials, 5873 participants; RR, 0.89 [CI, 0.74 to 1.07]) (14, 15, 24, 27, 28).
All 9 trials (9863 participants) in persons with CKD and acute coronary syndromes or requiring percutaneous coronary intervention provided information on major and minor bleeding events (Figure 3). Definitions of bleeding outcomes varied. Major bleeding was defined as an intracranial hemorrhage, a decrease in hemoglobin level of 50 g/L or more, or a decrease in hematocrit level of 15% or more in 7 trials (14–16, 23–25, 28); a decrease in hemoglobin level of more than 40 g/L, bleeding necessitating transfusion of 2 units or more of blood, bleeding necessitating corrective surgery, or intracranial or retroperitoneal hemorrhage in 1 trial (26); or substantially disabling bleeding, intraocular bleeding leading to loss of vision, or bleeding necessitating transfusion of 2 units or more of blood in 1 trial (27). Minor bleeding included blood loss, a decrease in hemoglobin level of 30 g/L or more, or no observed blood loss with a decrease in hemoglobin level of 40 g/L or more (6 trials) (14, 16, 23, 25); interruption of study medication (1 trial) (27); or bleeding that was incompletely defined (2 trials) (26, 28).
According to low-quality evidence, antiplatelet therapy in addition to standard care increased major (RR, 1.40 [CI, 1.07 to 1.86]) and minor (RR, 1.47 [CI, 1.25 to 1.72]) bleeding, although significant heterogeneity was present in the analyses (Table and Figure 3). Excluding the 2 trials that randomly assigned participants to clopidogrel (27, 28) resulted in similar effect estimates for major (7 trials, 5365 participants; RR, 1.36 [CI, 0.78 to 2.38]) and minor (RR, 1.38 [CI, 1.18 to 1.61]) bleeding with persistent heterogeneity in both analyses. Antiplatelet therapy increased major bleeding in analyses limited to trials reporting outcomes at 1 year or more (5 trials, 5868 participants; RR, 1.47 [CI, 1.00 to 2.10]) (14, 15, 24, 27, 28). In the 5 trials that reported data for hemorrhagic stroke (14–16, 24, 25), treatment hazards of antiplatelet therapy were uncertain (4035 participants; RR, 1.08 [CI, 0.47 to 2.49]) (Figure 2).
Antiplatelet treatment had little or no effect on coronary artery revascularization (7 trials, 5265 participants; RR, 0.93 [CI, 0.84 to 1.04]) (Figure 3). No data were available for treatment effects on all-cause hospitalization, end-stage kidney disease, or withdrawal from treatment.
Moderate-quality evidence showed that antiplatelet therapy reduced myocardial infarction (10 trials, 9133 participants; RR, 0.66 [CI, 0.51 to 0.87]) but had uncertain effects on stroke (10 trials, 9133 participants; RR, 0.66 [CI, 0.16 to 2.78]) (Table and Figure 4) (6, 30, 37, 43, 44, 49, 50, 52, 53, 58). There was no significant heterogeneity in these analyses. Data were insufficient to perform prespecified subgroup analyses based on the type of antiplatelet regimen or stage of kidney disease. No data were provided by trials enrolling only kidney transplant recipients.
Figure 4. Effect of antiplatelet agents on cardiovascular, mortality, and bleeding outcomes in persons with CKD at risk for or with stable cardiovascular disease.
Summary estimates are provided by using random-effects meta-analysis. Only trials reporting ≥1 event are shown. CHARISMA = Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance; CKD = chronic kidney disease; ETDRS = Early Treatment Diabetic Nephropathy Study; GI = gastrointestinal; HOT = Hypertension Optimal Treatment; STOP = Shunt Thrombotic Occlusion Prevention by Picotamide; UK-HARP-I = First United Kingdom Heart and Renal Protection.
Limiting meta-analyses to trials in which follow-up duration was 12 months or more did not meaningfully alter the risk estimates for these outcomes (6 trials, 7721 participants; RR for myocardial infarction, 0.69 [CI, 0.52 to 0.92], and RR for stroke, 0.90 [CI, 0.18 to 4.42]) (6, 30, 37, 49, 53, 58). Excluding the trial with data reported only in an internal report (44) did not alter the risk estimate for myocardial infarction (9 trials, 8848 participants; RR, 0.67 [CI, 0.51 to 0.88]).
In low-quality evidence, antiplatelet therapy had uncertain effects on all-cause (21 trials, 10 632 participants; RR, 0.87 [CI, 0.61 to 1.24]) (6, 29–31, 34, 36, 37, 39, 40, 42–45, 48–50, 52–54, 57, 58) and cardiovascular (16 trials, 8706 participants; RR, 0.91 [CI, 0.60 to 1.36]) (6, 29–31, 34, 36, 37, 39, 40, 42–45, 49, 57, 58) mortality, with no significant heterogeneity in analyses (Table and Figure 4). Risk for death from any cause differed in prespecified subgroups of trials according to the drug used. In 5 trials, aspirin (4340 participants; RR, 0.80 [CI, 0.61 to 1.06]) had uncertain effects on risk for death (6, 30, 36, 48, 58), whereas thienopyridines may increase mortality (7 trials, 3452 participants; RR, 1.47 [CI, 1.02 to 2.12]) (P < 0.01 for subgroup interaction) (34, 37, 42–44, 52, 54).
Subgroup analysis to explore the effects of CKD stage was not possible. Estimates for all-cause (13 trials, 8942 participants; RR, 0.89 [CI, 0.59 to 1.33]) (6, 29–31, 34, 36, 37, 45, 48, 49, 53, 57, 58) and cardiovascular (11 trials, 8186 participants; RR, 0.92 [CI, 0.56 to 1.51]) (6, 29–31, 34, 36, 37, 45, 49, 57, 58) mortality did not differ when analyses were restricted to trials in which outcomes were reported during follow-up of 12 months or longer. Excluding 2 trials with data available only in internal reports (44, 45) did not alter the effects of antiplatelet treatment on all-cause (19 trials, 9444 participants; RR, 0.93 [CI, 0.62 to 1.39]) or cardiovascular (14 trials, 7518 participants; RR, 0.94 [CI, 0.55 to 1.60]) mortality.
Eighteen trials (10 230 participants) reported major bleeding events (6, 36, 37, 39–45, 48, 49, 51–54, 57, 58), and 8 trials (7202 participants) reported minor bleeding events (6, 29, 36, 37, 40, 42, 52, 58). No trials in this clinical setting reported data specifically for hemorrhagic stroke. Definitions of major and minor bleeding were generally not well-defined and were not centrally adjudicated with blinding to treatment allocation, with the exception of 2 trials (37, 58). Major bleeding included intracerebral or substantial hemodynamic compromise (37); gastrointestinal bleeding (39); hemarthrosis, nasal bleeding, shunt hemorrhage, and bleeding at injection site (42); confirmed retroperitoneal, intra-articular, intraocular, or intracranial bleeding or causing the hemoglobin level to decrease by 20 g/L or more and requiring hospital admission or transfusion (53); or hospital admission or death (58). Minor bleeding included occult gastrointestinal or rectal polyp bleeding and ecchymoses (29); not needing transfusion or causing hemodynamic compromise (37); or epistaxis, ecchymoses, or bruising (58).
According to low-quality evidence, antiplatelet therapy significantly increased minor bleeding (RR, 1.70 [CI, 1.44 to 2.02]) but had uncertain effects on major bleeding (RR, 1.29 [CI, 0.69 to 2.42]) (Table and Figure 4). There was no heterogeneity in these analyses. Treatment with antiplatelet agents in trials of 1 year or longer was associated with increased major bleeding (10 trials, 8696 participants; RR, 1.55 [CI, 1.07 to 2.26]) (6, 36, 37, 45, 48, 49, 51, 53, 57, 58). Excluding the 2 trials for which data were available only in internal reports (44, 45) also resulted in significantly increased major bleeding in the remaining trials (16 trials, 9042 participants; RR, 1.50 [CI, 1.06 to 2.12]). Prespecified subgroup analyses to assess treatment effects in subgroups of trials based on stage of CKD or antiplatelet regimen were not possible.
Antiplatelet therapy had uncertain effects on end-stage kidney disease, hospitalization, and withdrawal from treatment (Figure 2). No data were available for coronary artery revascularization in persons with CKD who have stable or no cardiovascular disease.
No asymmetry was observed in the funnel plots for the outcomes of myocardial infarction, all-cause mortality, or major bleeding in trials involving persons at risk for or with stable cardiovascular disease or in those involving acute cardiovascular disease (Egger regression test, P > 0.10 for all). This finding suggests that unpublished studies did not cause bias of effect estimates.
To our knowledge, this meta-analysis provides the first comprehensive summary of the benefits and risks of antiplatelet treatment in persons with CKD. In general, evidence is of low or very-low quality, with considerable variation in trial duration; heterogeneity in the definitions and assessment of bleeding outcomes; reliance on subgroup data from major trials, particularly for antiplatelet treatment in acute cardiovascular disease; and substantial methodological limitations in data for adults with CKD and stable cardiovascular disease. Antiplatelet treatment (generally glycoprotein IIb/IIIa inhibitors) given in addition to standard care in persons with acute coronary syndromes or those undergoing percutaneous coronary revascularization who also have CKD has little or no effect on myocardial infarction, death, or coronary revascularization but increases major and minor bleeding. Evidence for an association between additional antiplatelet therapy and stroke and cardiovascular death in persons with CKD and who have acute coronary syndromes or who are undergoing percutaneous coronary intervention is scant.
Antiplatelet regimens in persons with CKD who have or are at risk for cardiovascular disease reduce fatal or nonfatal myocardial infarction by approximately 33% but have uncertain effects on stroke or all-cause and cardiovascular mortality. Summary estimates for the effects of antiplatelet agents on major bleeding or hemorrhagic stroke are uncertain. Thienopyridines may increase mortality in persons with CKD and stable cardiovascular disease, although data are derived from subgroup analyses in few trials and are unreliable.
Overall, most trials of persons with CKD and stable or no cardiovascular disease have methodological or reporting limitations that reduce the reliability of the evidence. When absolute treatment effects are estimated, persons with acute coronary syndromes or who have undergone high-risk coronary intervention experience no reduction in myocardial infarction or subsequent need for revascularization with additional antiplatelet therapy. However, up to 2% of these persons may have serious bleeding. Twelve months of oral antiplatelet therapy may prevent myocardial infarction in 1% to 3% of persons at risk for myocardial infarction, but information about bleeding hazards and especially intracranial hemorrhage is of low or very low quality. Given the quality of the available evidence, specific trials evaluating antiplatelet agents in persons with CKD and coexisting acute or stable cardiovascular disease are required.
Outcome data for antiplatelet agents are scant in several important clinical settings. We currently have no data on persons receiving dialysis or kidney transplant recipients who have acute coronary syndromes or require coronary artery revascularization, and evidence for persons with earlier stages of CKD is entirely derived from post hoc analyses within larger trials. In addition, more and better evidence is required for long-term antiplatelet treatment in persons receiving dialysis or who have undergone kidney transplantation.
Evidence to support secondary prevention with low-cost antiplatelet drugs (such as aspirin) in persons with a recent occlusive myocardial event and coexisting CKD is not available, and extrapolating data from the general population may not be appropriate because the biology of arterial disease and causes of death in persons with CKD may confer a different risk–benefit tradeoff for therapy. As shown in the general population (74), widespread administration of antiplatelet agents may have uncertain value in persons with CKD because of the balance between reduced occlusive events (myocardial infarction) and the uncertain risk for major bleeding, including intracranial hemorrhage.
Although the values and preferences of patients with CKD are not well-understood (75), it seems unlikely that many patients would accept the risk for major bleeding to reduce the risk for myocardial infarction without proven reductions in death or the need for coronary revascularization. The benefits and hazards of antiplatelet therapy to prevent cardiovascular events may be particularly important for patients receiving hemodialysis and dialysis-related anticoagulation who have impaired hemostasis (73). However, these patients may also have greater absolute reductions in occlusive coronary events because of higher baseline risk (71).
Considering the totality of current evidence, using antiplatelet and related agents to prevent cardiovascular events in people with CKD may be prudent only in clinical trials that can further define the role of these drugs. Recent prespecified subgroup data from the PLATO (Platelet Inhibition and Patient Outcomes) trial indicates that ticagrelor, an oral purinergic receptor inhibitor cleared by extrarenal mechanisms, reduces mortality and major cardiovascular events better than clopidogrel among persons with CKD and acute coronary syndromes (76). This finding suggests that newer antiplatelet agents may potentially act as adjunctive therapy in persons with impaired kidney function. However, large placebo-controlled trials to assess the relative benefits and toxicity of these newer antiplatelet agents, in addition to standard care specifically in persons with CKD and acute cardiovascular disease, are needed.
An earlier collaborative meta-analysis (5) provided data for the effects of antiplatelet agents in patients receiving hemodialysis, but to our knowledge, outcome data for persons with earlier stages of CKD have not been previously summarized. In that previous meta-analysis (5), antiplatelet therapy reduced major cardiovascular events (nonfatal myocardial infarction, nonfatal stroke, and vascular death) by 41% in persons undergoing hemodialysis; however, CIs approached no effect, and the effects of treatment on individual components of the outcome were not reported. In that previous review (5), a meta-analysis that included only 46 events summarized bleeding risks with antiplatelet therapy in persons receiving hemodialysis, and estimates were—appropriately—considered unreliable.
We found a lack of a clear reduction in vascular deaths with antiplatelet treatment among persons with CKD; this finding is in contrast to findings observed in other populations at high risk for vascular events, including persons with a history of documented myocardial infarction and stroke (5). Our finding that antiplatelet agents reduce cardiovascular events in persons with CKD to a lesser extent than in other populations and have no certain effect on mortality in persons with CKD echoes similar treatment effects for statin therapy in persons with CKD (77). The competing mechanisms for cardiovascular disease in this population potentially explain this finding. Progressive kidney dysfunction is characterized by vascular stiffening and calcification, cardiomyopathy, hyperkalemia, and sudden cardiac death, in addition to occlusive vascular disease (78).
We found that the proportional increased risk for serious bleeding from antiplatelet agents were 20% to 40%, which is somewhat smaller than that seen in patients with documented chronic or acute cardiovascular disease (60%) (5). However, the substantially higher baseline risk for bleeding in persons with CKD (approximately 2.5% per year  compared with <1% in other at-risk populations ) means that absolute bleeding risks with antiplatelet therapy might be at least doubled in persons with CKD.
It is also relevant to consider which specific bleeding complications are incurred by treatment when deciding whether the clinical benefits outweigh the potential risks of treatment in persons with CKD. Reversible hemorrhage from the gastrointestinal tract, skin, or dialysis access or surgical sites may be more acceptable treatment hazards than disabling bleeding into an eye, a joint, or the brain or bleeding requiring major surgery. However, insufficient data were available in the included trials to provide comprehensive information on specific types of bleeding caused by antiplatelet agents in this population and, consequently, on the relevant risk–benefit tradeoff needed to inform clinical decision making.
It is also important to remember that persons in the “real world” may have much higher risks for bleeding than trial participants. Therefore, the absolute numbers of persons with CKD affected by serious bleeding complications from antiplatelet treatment suggested in the Table may shift the balance toward excess harm.
Our meta-analysis quantifies the benefits and harms of antiplatelet agents in a large number of persons. However, it has limitations, largely because it relies on trial-level rather than individual-patient data. First, we could not assess whether the stage of kidney disease modified the effects of antiplatelet therapy. Second, definitions and assessments of bleeding were widely heterogeneous; therefore, estimates of hazards for specific bleeding events were less reliable. Third, overall trial duration varied greatly; in general, the longer-term effects (3 to 5 y) of antiplatelet treatment are uncertain.
Fourth, the amount of data available overall, and particularly that in persons with advanced CKD, was limited; as a result, insufficient power may explain some of the negative findings and highlights the need for trials of antiplatelet agents specifically targeting this population. In addition, reliance on data from post hoc analyses in larger trials may reduce the reliability of the summary findings (79). This is particularly true for evidence in persons with CKD and acute cardiovascular disease, for whom available data provide hypothesis-generating, rather than confirmatory, evidence for antiplatelet treatment effects in this population—especially for adverse events. Finally, we were unable to determine the relative benefits of antiplatelet agents in primary prevention of cardiovascular disease (treating persons without clinically evident cardiovascular disease) compared with secondary prevention (treating persons with established cardiovascular disease) because too few trials provided data for participants in the primary prevention setting.
In conclusion, evidence for antiplatelet agents in persons with CKD and various cardiovascular diseases is of low quality. Glycoprotein IIb/IIIa inhibitors or clopidogrel given in addition to standard care have little or no effect on death, myocardial infarction, or coronary revascularization and may increase major bleeding in persons with CKD and acute coronary syndromes or those having high-risk coronary artery intervention. Antiplatelet agents reduce myocardial infarction in persons with CKD but have uncertain effects on stroke and mortality and may increase bleeding. Bleeding hazards and lack of clear efficacy in reducing cardiovascular morbidity and mortality need to be acknowledged when patients with CKD are being counseled about acute or long-term antiplatelet therapy.
Annals staff will be at these upcoming meetings:
American College of Cardiology, 24–27 March 2012, Chicago
Digestive Disease Week, 20–22 May 2012, San Diego
American Thoracic Society, 20–22 May 2012, San Francisco
American Society of Clinical Oncology, 1–5 June 2012, Chicago
American Diabetes Association, 8–12 June 2012, Philadelphia
Stop by the ACP/Annals booth and register to be a peer reviewer or discuss your thoughts for submission or topic coverage with Annals staff.
The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.
William G.Kussmaul iii, Interventional Cardiology
Hahnemann University Hospital
March 23, 2012
Caution regarding CKD patients who have coronary stents
Palmer and co-authors have usefully collected information on the use of anti-platelet agents in patients who have both coronary and kidney disease. They conclude that the benefits in both primary and secondary prevention are uncertain, and there may be a greater risk of bleeding than in non-CKD patients. They imply that these agents should not be used in patient with CKD until better data are available.
Unfortunately they may have done a dis-service to those many CKD patients who may have coronary stents, and to their physicians. Almost all patients receiving stents are required to take aspirin and a thienopyridine (most often clopidogrel) for 30 days up to 12 months, depending on the type of stent. The goal is to prevent acute coronary stent thrombosis, a very dangerous event.
While it is certainly true that trials of anti-platelet therapy specifically in CKD patients are lacking, I am hoping that the publication of this systematic review will not result in the premature discontinuation of the drugs in stent patients. After all, the authors' first conclusion is that the evidence is of poor quality. This should be seen as true both for benefits and potential harms.
Evidence-based medicine is to be promoted and practiced. In the absence of good evidence, one can either withhold a treatment until it is proven efficacious, or give it unless it is proven harmful. In the case of coronary stenting, I would urge the latter approach while agreeing with the authors that better studies in patients with CKD and CAD are needed.
R. BrooksRobey, Nephrologist & Associate Chief of Staff for Research, Daniel J. O'Rourke
VA Medical Center & Geisel School of Medicine at Dartmouth
April 13, 2012
Antiplatelet therapy in chronic kidney disease
The systematic review and meta-analysis by Palmer et al. (1) found limited objective evidence to support antiplatelet therapy in patients with chronic kidney disease (CKD). We concur with this assessment, but feel that the authors introduced subjective bias into their analysis and interpretation of the data that potentially underestimates benefits while overemphasizing risks.
The authors previously reported reduced cardiovascular events and mortality in hypertensive CKD patients receiving low dose aspirin relative to their counterparts with normal kidney function (2, 3). Apparent benefits were observed in all outcome categories and were greatest in patients with Stage 3b or greater CKD (estimated glomerular filtration rate (eGFR) <45 mL/min/1.73 m2) where treatment was associated with up to 40-fold lower event rates (2). Benefits were markedly diminished when Stage 3a CKD patients were included in the analysis (eGFR <60 mL/min/1.73 m2). These data, included in their subsequent meta-analysis, account for over 40% of all patients in some outcome categories (1). Unfortunately, the study design ignored their previously defined major benefit threshold in favor of unstratified data analysis. We suspect that this decision was based on limited availability of properly stratified data and considerations of statistical and clinical heterogeneity. However, by combining datasets with established outcome differences, the authors knowingly introduced subjective bias into their analysis.
In their previous work, the authors confirmed associated bleeding risks in CKD (2) similar to those observed in the general population (4), but they concluded that the risks were "outweighed by the substantial benefits" of therapy (2). They reached a different conclusion in their subsequent review based on assumptions and speculation not fully supported by the data (1). Increased bleeding risks are presumably fully reflected in corresponding overall outcomes.
It is disconcerting that review of over three decades of investigation identified so few studies to inform the use of antiplatelet agents in CKD (1). Undue reliance on post hoc subgroup analysis of studies not designed nor powered to address this issue is also concerning (1). Nonetheless, the absence of unambiguous evidence of benefit does not constitute unambiguous evidence of an absence of benefit. Caution is warranted, but the totality of the evidence still seems to favor low dose aspirin therapy in CKD (1-3, 5). Acquired uremic platelet dysfunction is largely observed in end-stage kidney disease patients unrepresented in these analyses, so additional studies stratifying data by CKD stage, baseline hematocrit, and the presence or absence of pretreatment platelet functional abnormalities are needed to advance this line of inquiry.
1. Palmer SC, Di Micco L, Razavian M, Craig JC, Perkovic V, Pellegrini F, et al. Effects of Antiplatelet Therapy on Mortality and Cardiovascular and Bleeding Outcomes in Persons With Chronic Kidney Disease: A Systematic Review and Meta-analysis. Annals of Internal Medicine. 2012;156(6):445-59.
2. Jardine MJ, Ninomiya T, Perkovic V, Cass A, Turnbull F, Gallagher MP, et al. Aspirin is beneficial in hypertensive patients with chronic kidney disease: a post-hoc subgroup analysis of a randomized controlled trial. Journal of the American College of Cardiology. 2010;56(12):956-65.
3. Choi MJ, Fried LF. Chronic kidney disease and progression. Nephrology Self Assessment Program, American Society of Nephrology. 2011;10(5):491-3.
4. McQuaid KR, Laine L. Systematic review and meta-analysis of adverse events of low-dose aspirin and clopidogrel in randomized controlled trials. The American Journal of Medicine. 2006;119(8):624-38.
5. Daugirdas JT. Handbook of Chronic Kidney Disease Management. Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins Health; 2011.
Giovanni F.M.Strippoli, , Suetonia C. Palmer and Jonathan C. Craig
Consorzio Mario Negri Sud, Italy
May 6, 2012
Author reply to Kussmaul and Robey
Dr Kussmaul raises an important point - does randomised trial (RCT) evidence tell us whether patients who have chronic kidney disease (CKD) should receive aspirin and clopidogrel for 12 months after coronary artery stenting? In the general population, guidelines recommend antiplatelet therapy with aspirin and a P2Y12 inhibitor (clopidogrel, prasugrel or ticagrelor) for 12 months after percutaneous stent insertion, unless the risk of morbidity from bleeding outweighs the anticipated benefits, when earlier discontinuation from the P2Y12 inhibitor is reasonable(1). Applying the evidence from the general population to the setting of CKD may be inappropriate because of the increased risks of bleeding. Our systematic review has highlighted that data to address the question of dual antiplatelet therapy after stenting in the setting of CKD are scant. The benefits and harms of aspirin are unknown and the role of clopidogrel after coronary intervention is explored by two RCTs, that both observed uncertain benefits and harms for 12 months of clopidogrel treatment(2, 3). We need to acknowledge this current uncertainty and promote more research.
Drs Robey and O'Rourke query why our review concluding there is little benefit for antiplatelet agents in CKD is so different to that of a post-hoc analysis from a trial of long-term aspirin treatment (HOT study) which contributed data to our review(4). In that RCT, aspirin reduced risks of myocardial infarction by 69% (CI, 15-81), as well as cardiovascular (64% [CI, 10-86]) and all-cause mortality (49% [CI, 6-73]). It should be noted that the number of events in HOT were very small and therefore imprecise and this was a post-hoc subgroup analysis that was not pre-specified at randomization(5). When data from HOT are combined with all other RCTs, the totality of the evidence suggests that benefits of antiplatelet treatment are more modest or absent in people with stable cardiovascular disease and CKD and are consistent across trials of different antiplatelet agents or stages of CKD. We suggest CKD patients should be counseled about treatment uncertainties when considering antiplatelet therapy. We reject the notion of subjective bias in the analysis. The review was conducted using a peer-reviewed protocol specifying all aspects of conduct before data collection and analysis. We certainly agree that "it is disconcerting that review of over three decades of investigation identifies so few studies to inform the use of antiplatelet agents in CKD". Now we know this, it is time to act.
1. Levine GN, Bates ER, Blankenship JC, Bailey SR, Bittl JA, Cercek B, et al. 2011 ACCF/AHA/SCAI Guideline for Percutaneous Coronary Intervention: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines and the Society for Cardiovascular Angiography and Interventions. Circulation. 2011;124(23):2574-609.
2. Best PJ, Steinhubl SR, Berger PB, Dasgupta A, Brennan DM, Szczech LA, et al. The efficacy and safety of short- and 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 (CREDO) trial. Am Heart J. 2008;155(4):687-93.
3. Keltai M, Tonelli M, Mann JF, Sitkei E, Lewis BS, Hawken S, et al. Renal function and outcomes in acute coronary syndrome: impact of clopidogrel. Eur J Cardiovasc Prev Rehabil. 2007;14(2):312-8.
4. Jardine MJ, Ninomiya T, Perkovic V, Cass A, Turnbull F, Gallagher MP, et al. Aspirin is beneficial in hypertensive patients with chronic kidney disease: a post-hoc subgroup analysis of a randomized controlled trial. J Am Coll Cardiol. 2010;56(12):956-65.
5. Sun X, Briel M, Busse JW, You JJ, Akl EA, Mejza F, et al. Credibility of claims of subgroup effects in randomised controlled trials: systematic review. BMJ. 2012;344:e1553.
Cardiology, Emergency Medicine, Nephrology, High Value Care, Chronic Kidney Disease.
Results provided by:
Copyright © 2016 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use
This PDF is available to Subscribers Only