OR indicates odds ratio; CI, confidence interval.
Tsai TT, Maddox TM, Roe MT, Dai D, Alexander KP, Ho PM, Messenger JC, Nallamothu BK, Peterson ED, Rumsfeld JS, National Cardiovascular Data Registry FT. Contraindicated Medication Use in Dialysis Patients Undergoing Percutaneous Coronary Intervention. JAMA. 2009;302(22):2458-2464. doi:10.1001/jama.2009.1800
Author Affiliations: Denver VA Medical Center, Denver, Colorado (Drs Tsai, Maddox, Ho, and Rumsfeld); University of Colorado Denver (Drs Tsai, Maddox, Ho, Messenger, and Rumsfeld); Colorado Cardiovascular Outcomes Research Group, Denver (Drs Tsai, Maddox, Ho, and Rumsfeld); Duke University Medical Center, Duke Clinical Research Institute, Durham, North Carolina (Drs Roe, Dai, Alexander, and Peterson); and Ann Arbor VA Medical Center, University of Michigan Hospital, Ann Arbor (Dr Nallamothu).
Context The US Food and Drug Administration guides clinicians through drug labeling of medications that are contraindicated or not recommended for use in specific patient groups. Little is known about the use of such medications and their effects on outcomes in clinical practice.
Objective To investigate the use of the contraindicated/not-recommended agents enoxaparin and eptifibatide among dialysis patients undergoing percutaneous coronary intervention (PCI) and their association with outcomes.
Design, Setting, and Participants Data from 829 US hospitals on 22 778 dialysis patients who underwent PCI between January 1, 2004, and August 31, 2008.
Main Outcome Measures In-hospital bleeding and death.
Results Five thousand eighty-four patients (22.3%) received a contraindicated antithrombotic; of these patients, 2375 (46.7%) received enoxaparin, 3261 (64.1%) received eptifibatide, and 552 (10.9%) received both. Compared with patients who did not receive a contraindicated antithrombotic, patients who did had higher rates of in-hospital bleeding (5.6% vs 2.9%; odds ratio [OR], 1.93; 95% confidence interval [CI],1.66-2.23) and death (6.5% vs 3.9%; OR, 1.68; 95% CI, 1.46-1.95). After multivariable adjustment, patients receiving contraindicated antithrombotics had significantly higher risks of in-hospital bleeding (OR, 1.66; 95% CI, 1.43-1.92) and death (OR, 1.24; 95% CI, 1.04-1.48). In 10 158 patients matched by propensity scores, receipt of contraindicated antithrombotics remained significantly associated with in-hospital bleeding (OR, 1.63; 95% CI, 1.35-1.98) but not in-hospital death (OR, 1.15; 95% CI, 0.97-1.36).
Conclusions In a sample of dialysis patients undergoing PCI, 22.3% received a contraindicated antithrombotic medication. In propensity-matched analysis, receipt of these medications was significantly associated with an increased risk of in-hospital major bleeding.
In the United States, medication errors are implicated in more than 100 000 deaths annually.1 Medication errors include adverse drug reactions related to inappropriately prescribed or administered drugs.2 To minimize inappropriate medication use, the US Food and Drug Administration (FDA) guides pharmaceutical manufacturers and clinicians through drug labeling of which medications are contraindicated or not recommended for use in specific patient groups. Few studies have evaluated the frequency of use and association between receipt of these medications and outcomes in clinical practice.
Patients receiving dialysis are an increasing population, projected to exceed 2 million patients worldwide by 2010.3 Since many medications are eliminated through the kidneys, dialysis patients are at elevated risk of medication-related adverse events.4- 7 Thus, many medications are contraindicated in dialysis patients.
A particularly challenging area of treatment of dialysis patients is percutaneous coronary intervention (PCI) because several antithrombotic medications are not recommended for use in dialysis patients. Eptifibatide, a platelet glycoprotein IIb/IIIa inhibitor, is renally cleared and its use contraindicated in dialysis patients because of increased bleeding risk. Similarly, enoxaparin, a low-molecular-weight heparin, is an anticoagulant that is renally cleared and its use not recommended in dialysis patients because of the risk of bleeding complications.8- 10 Non–renally cleared antithrombotic alternative medications are available, such as abciximab and unfractionated heparin.
The extent to which contraindicated antithrombotics are used in dialysis patients undergoing PCI and the association between use of these medications and clinical outcomes is unknown. Using data from the National Cardiovascular Data Registry (NCDR), we evaluated the frequency with which enoxaparin, eptifibatide, or both were administered to dialysis patients undergoing PCI, and we examined the association between the receipt of these medications and in-hospital bleeding and death.
The NCDR CathPCI registry, cosponsored by the American College of Cardiology (ACC) and the Society for Cardiovascular Angiography and Interventions, has been previously described.11,12 The registry catalogs data on patient and hospital characteristics, clinical presentation, treatments, and outcomes for PCI procedures from more than 800 sites across the United States. Data are entered into NCDR-certified software at participating institutions and exported in a standard format to the ACC. There is a comprehensive data quality program, including both data quality report specifications for data capture and transmission and an auditing program. The definitions variables were prospectively defined by a committee of the ACC and are available at http://www.ncdr.com. Race/ethnicity data were recorded by chart abstractors as part of the standard case report form. The options were Caucasian, black, Hispanic, Asian, Native American, and other. In our analysis, patients were categorized as white or nonwhite race (black, Hispanic, Asian, Native American, and other).
This study included all dialysis patients undergoing PCI who were treated with 1 or more antithrombotic agents. Between January 1, 2004, and August 31, 2008, 1 357 758 PCI patients from 829 hospitals were enrolled in the CathPCI registry, among whom 24 656 (1.8%) were receiving chronic dialysis. We then excluded those patients receiving warfarin (n = 1362; 5.5%) because of its unknown influence on antithrombotic choices and other infrequently used antithrombotics, tirofiban (n = 126; 0.51%), argatroban (n = 70; 0.28%), lepirudin (n = 23; 0.09%), thrombolytics (n = 189; 0.77%), dalteparin (n = 67; 0.27%), and nadroparin (n = 41; 0.17%), to decrease heterogeneity in the comparison groups. The final analytic cohort was 22 778 dialysis patients undergoing PCI.
The independent variables of interest were receipt of eptifibatide, enoxaparin, or both. Eptifibatide is contraindicated in dialysis patients because it is predominantly renally cleared and is associated with an increased risk of bleeding. The manufacturer's package insert states treatment with eptifibatide is contraindicated in patients with dependency on renal dialysis; dosing information for dialysis patients is excluded. Pharmacologic references such as Lexi-Comp also state that dialysis is a contraindication to its use.13 Enoxaparin similarly has predominantly renal clearance and an associated increased risk of bleeding among patients on dialysis.8- 10 The warnings and precautions section of the package insert states that its use in patients with renal impairment has been associated with hyperkalemia and increased bleeding; dosing information for dialysis patients is excluded. Pharmacologic references such as Lexi-Comp do not recommend its use, noting that enoxaparin has not been FDA approved for use in dialysis patients and that serious bleeding complications have been reported with use in patients who are dialysis dependent.13
The study outcomes were in-hospital major bleeding and all-cause in-hospital death. In-hospital major bleeding is defined in NCDR as bleeding requiring a transfusion, prolonging the hospital stay, or causing a decrease in hemoglobin of more than 3 g/dL at 1 or more of the following sites: percutaneous entry site, retroperitoneum, gastrointestinal, genital-urinary, or other/unknown.
Patients were classified into 2 groups: receipt of enoxaparin and/or eptifibatide or receipt of noncontraindicated antithrombotics (unfractionated heparin, bivalirudin, or abciximab). Differences between groups were compared using χ2 tests for categorical variables and the Wilcoxon rank sum or Kruskal-Wallis test for continuous variables.
To assess the association between contraindicated antithrombotics and in-hospital major bleeding and death, we used multivariable analysis with generalized estimating equations to adjust for potential confounders and within-hospital correlations. Variables from the NCDR PCI mortality model14 and NCDR PCI bleeding model15 were used for in-hospital mortality risk adjustment and in-hospital major bleeding risk adjustment, respectively. Patients transferred to another hospital were excluded from the death analysis; patients who died the same day as the procedure were excluded from the bleeding analysis.
We also developed a propensity score for the receipt of contraindicated antithrombotics using variables from the NCDR PCI mortality model.13 Based on the estimated logits, we selected matched pairs, conducted a McNemar test to determine whether in-hospital major bleeding or death was different between patient groups, and computed an odds ratio (OR) to test the association between contraindicated antithrombotics and in-hospital major bleeding or death.
In secondary analyses, multivariable regression analyses were repeated within the anticoagulant and antiplatelet medication classes. To evaluate the association of enoxaparin with outcomes, we restricted the cohort to patients who received the anticoagulants bivalirudin, enoxaparin, or unfractionated heparin, adjusting for receipt of antiplatelet agents. To evaluate the association of eptifibatide with outcomes, we restricted the cohort to patients who received eptifibatide or abciximab, adjusting for receipt of anticoagulants. Then, the association between receipt of contraindicated antithrombotics and outcomes was assessed in strata of patients who underwent PCI in the setting of an acute coronary syndrome (ACS) vs non-ACS PCI. A first-order interaction was assessed for ACS status and outcome. Analyses were performed using SAS software version 9.0 (SAS Institute, Cary, North Carolina).
This study was reviewed by the Duke University Health System institutional review board and was determined to meet the definition of research not involving human subjects.
Baseline characteristics of the study population are shown in Table 1. Patients receiving a contraindicated antithrombotic were more likely to have heart failure, lung disease, or active tobacco use; to present with ACS; and to be cared for at private or community centers with lower procedural volume. Patients who did not receive a contraindicated antithrombotic were more likely to have nonwhite race, diabetes, and history of prior PCI and to be treated at centers that were urban, had training programs, and had higher procedural volume.
Overall, 5084 patients (22.3%) received a contraindicated antithrombotic medication; 2375 (46.7%) received enoxaparin, 3261 (64.1%) received eptifibatide, and 552 (10.9%) received both. In the overall cohort, in-hospital major bleeding occurred in 3.6% (805/22 778) and in-hospital death occurred in 4.5% (963/22 778). Table 2 shows the frequency of major bleeding sites stratified by the receipt of a contraindicated antithrombotic. Gastrointestinal bleeding is the most prevalent cause of bleeding in patients who received a contraindicated antithrombotic, whereas percutaneous bleeding was the most prevalent cause in those who did not receive a contraindicated antithrombotic. Approximately one-third of patients had an alternative or unknown site of bleeding. Table 3 shows the causes of in-hospital death. As expected, most deaths were due to cardiac causes, and the data do not differentiate hemorrhage-related deaths.
In unadjusted analysis, patients who received contraindicated antithrombotics experienced higher rates of in-hospital major bleeding (5.6% vs 2.9%; OR, 1.93; 95% confidence interval [CI], 1.66-2.23) and death (6.5% vs. 3.9%; OR, 1.68; 95% CI, 1.46-1.95). With multivariable regression analysis, there remained a significantly increased risk of major bleeding (adjusted OR, 1.66; 95% CI, 1.43-1.92) and death (adjusted OR, 1.24; 95% CI, 1.04-1.48) associated with the use of contraindicated antithrombotics (Figure 1).
The propensity analysis yielded 5079 matched pairs; baseline characteristics of these patients were well balanced without any significant differences. In the propensity-matched cohort, receipt of contraindicated antithrombotics was significantly associated with increased in-hospital major bleeding (OR, 1.63; 95% CI, 1.35-1.98), but no significant association was found with in-hospital death (OR, 1.15; 95% CI, 0.97-1.36) (Figure 1).
In secondary analyses within anticoagulant and antiplatelet classes, patients receiving enoxaparin had significantly higher rates of in-hospital major bleeding (5.0% vs 3.9% vs 3.0%, P < .001) and death (6.0% vs 5.4% vs 3.0%, P < .001) compared with patients who received unfractionated heparin or bivalirudin, respectively. After risk adjustment, receipt of enoxaparin remained associated with in-hospital major bleeding (adjusted OR, 1.28; 95% CI, 1.04-1.59) and in-hospital death (adjusted OR, 1.35; 95% CI, 1.06-1.72) compared with unfractionated heparin. Compared with patients who received bivalirudin, patients receiving enoxaparin did not have a significantly higher risk of major bleeding (adjusted OR, 1.17; 95% CI, 0.93-1.47) but had a significantly higher risk of death (adjusted OR, 1.44; 95% CI, 1.11-1.88). The use of eptifibatide was not significantly associated with a higher risk of unadjusted in-hospital major bleeding (6.2% vs 5.4%; OR, 1.20; 95% CI, 0.93-1.54) or death (6.8% vs 7.1%; OR, 0.96; 95% CI, 0.76-1.20). After adjustment, the use of eptifibatide was not significantly associated with in-hospital major bleeding (adjusted OR, 1.23; 95% CI, 0.95-1.61) or in-hospital death (adjusted OR, 1.04; 95% CI, 0.78-1.37).
In analyses stratified by ACS status, among patients undergoing PCI in the setting of ACS, the use of any contraindicated antithrombotic (adjusted OR, 1.82; 95% CI, 1.52-2.17), enoxaparin vs unfractionated heparin (adjusted OR, 1.28; 95% CI, 1.01-1.62), or eptifibatide vs abciximab (adjusted OR, 1.39; 95% CI, 1.04-1.85) was associated with an increased risk of in-hospital major bleeding (Figure 2). Similarly, the use of any contraindicated antithrombotic (adjusted OR, 1.20; 95% CI, 1.00-1.46), enoxaparin vs unfractionated heparin (adjusted OR, 1.36; 95% CI, 1.05-1.74), or enoxaparin vs bivalirudin (adjusted OR, 1.51; 95% CI, 1.15-1.98) was associated with an increase in in-hospital death. However, there was no significant association with in-hospital major bleeding in the enoxaparin vs bivalirudin group (adjusted OR, 1.14; 95% CI, 0.88-1.47) or with in-hospital death in the eptifibatide vs abciximab group (adjusted OR, 1.13; 95% CI, 0.84-1.51). In the non-ACS PCI strata, all comparisons were nonsignificant (data not shown); however, the test for interaction was nonsignificant, and thus effect modification by ACS status was not confirmed.
Using data from the NCDR, we found that approximately 20% of dialysis patients undergoing PCI received enoxaparin, eptifibatide, or both, which are contraindicated or not recommended for use in patients receiving dialysis. In risk-adjusted analyses, the use of these medications was associated with increased in-hospital major bleeding and possibly in-hospital death. These findings were consistent in the overall cohort using multivariable regression and in a propensity-matched cohort analysis. The associations were particularly evident among patients undergoing PCI in the setting of ACS. This study therefore demonstrates that these medications are used in clinical practice despite FDA-directed labeling, and their use is associated with adverse patient outcomes.
The results of this study illustrate the problem of medication errors in the United States, as well as the need to make patient safety a priority on the health care agenda.16 Few prior studies have evaluated this aspect of medication errors, namely the use of contraindicated medications in specific patient subgroups and the association of such medication use with patient outcomes. Among patients receiving chronic dialysis, prior studies have identified inappropriate medication dosing and selection in ambulatory dialysis centers.7,17,18 These studies have been limited by small numbers of patients, and the association between receipt of medications and outcomes has been largely unknown. The current study is the first, to our knowledge, to evaluate the increasing population of dialysis patients undergoing PCI with regard to receipt of contraindicated medications and their outcomes.
There are several potential reasons for the use of enoxaparin and eptifibatide among dialysis patients undergoing PCI. One is the ease of administration of enoxaparin compared with unfractionated heparin. Enoxaparin is administered either as an intravenous bolus or a subcutaneous injection without a continuous intravenous drip, which simplifies emergency and inpatient management as well as transfer protocols. As a result, enoxaparin is a common drug choice in preprinted order sets for ACS management. Second, compared with abciximab, eptifibatide is often less expensive and more widely available in many hospitals and has generally become the predominant glycoprotein inhibitor for PCI.19 Third, despite FDA-guided warnings in the medication package inserts and Internet-based pharmacy resources noting that these agents are not recommended for use in dialysis patients,13 there is a paucity of direct evidence from clinical trial data regarding the use of these agents because dialysis patients are usually excluded from such trials. Moreover, there have been no previous large studies on the use and outcomes of these agents in routine clinical practice. The lack of trial evidence and prior studies may have contributed to the use of these agents despite labeling.
There are several important considerations when interpreting the results of this study. First, patients and hospitals participating in the NCDR may not be representative of all US practice. However, the CathPCI registry represents more than 800 hospitals across the United States and thus captures a significant portion of PCI procedures nationally. Furthermore, hospitals enrolled in registry programs like CathPCI may be more committed to quality assurance, which could mean that the use of contraindicated antithrombotics in dialysis patients found in this study is underestimated. Second, the CathPCI registry does not capture the dosage or timing (before or during the PCI laboratory visit) of medication administration and cannot evaluate for dosing errors or distinguish between patients who received more than 1 class of antithrombotic medication at the same time or in succession (ie, on admission and then in the catheterization laboratory). However, eptifibatide and enoxaparin, independent of their dose and timing of administration, are considered contraindicated in dialysis patients.
Third, as in all observational studies, unmeasured confounders that could influence the receipt of contraindicated antithrombotics must be considered. However, a wide array of variables was available for risk adjustment, and multivariable regression in the overall cohort and the propensity-matched analysis yielded similar results. Fourth, hemorrhagic strokes were not differentiated from ischemic strokes on the data collection form and thus were not included in the major bleeding end point. The rate of bleeding events may therefore be underestimated. Fifth, we were unable to assess long-term outcomes following discharge. However, the CathPCI registry, with its large numbers of patients, is well suited to address safety issues in patient subsets poorly represented in smaller registries. Finally, this study does not attempt to define the optimal antithrombotic therapy in dialysis patients undergoing PCI. However, our study addresses a clinically relevant safety issue in a growing subset of patients undergoing PCI.
In a large national cohort study, we found that approximately 20% of dialysis patients undergoing PCI received eptifibatide or enoxaparin. The use of these medications was associated with increased risk of in-hospital major bleeding. These results appear to validate the FDA-directed labeling of eptifibatide and enoxaparin as contraindicated or not recommended for use in dialysis patients. Because non–renally cleared antithrombotic alternatives are available, this study supports avoiding the use of enoxaparin and eptifibatide in dialysis patients undergoing PCI. Educational efforts targeting clinicians who prescribe these medications and quality improvement interventions, such as amending clinical pathway order sets to include consideration of renal function, are urgently needed.
Corresponding Author: Thomas T. Tsai, MD, MSc, Denver VA Medical Center, Cardiology Section 111B, 1055 Clermont St, Denver, CO 80220 (email@example.com).
Author Contributions: Dr Tsai had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Tsai, Ho, Rumsfeld.
Acquisition of data: Roe.
Analysis and interpretation of data: Tsai, Maddox, Roe, Dai, Alexander, Messenger, Nallamothu, Peterson.
Drafting of the manuscript: Tsai, Maddox, Dai.
Critical revision of the manuscript for important intellectual content: Tsai, Roe, Alexander, Ho, Messenger, Nallamothu, Peterson, Rumsfeld.
Statistical analysis: Dai.
Administrative, technical, or material support: Tsai, Roe, Ho, Messenger, Rumsfeld.
Study supervision: Maddox, Roe, Ho, Messenger, Rumsfeld.
Financial Disclosures: Dr Roe reported receiving significant research support from Bristol-Meyers Squibb, Sanofi-Aventis, Schering Plough, and Eli Lilly and serving as a consultant to Schering Plough and the speakers’ bureaus for Schering Plough, Bristol-Meyers Squibb, and Sanofi-Aventis. Dr Peterson reported receiving significant research support from the partnership between Bristol-Myers Squibb and Sanofi and from Schering Plough. No other disclosures were reported.
Funding/Support: This analysis was funded by the NCDR CathPCI Registry, which oversees the collection and management of data in the registry. Dr Ho is supported by a VA Health Services Research and Development Service (HSR&D) Career Development Award.
Role of the Sponsor: The funding agency had no role in the design and conduct of the study; in the analysis and interpretation of the data; or in the preparation of the manuscript. The Research and Publications committee for the NCDR CathPCI Registry reviewed and approved the final version of the manuscript.
Disclaimer: Dr Peterson, a contributing editor for JAMA, was not involved in the editorial review of or the decision to publish this article.
Previous Presentation: Presented in part at the 2009 Scientific Sessions of the American Heart Association; November 14-18, 2009; Orlando, Florida.