- Email: [email protected]
Comparison of In-Hospital Outcomes for Beta-Blocker Use Versus Non–Beta Blocker Use in Patients Presenting With Cocaine-Associated Chest Pain
Comparison of In-Hospital Outcomes for Beta-Blocker Use Versus NoneBeta Blocker Use in Patients Presenting With Cocaine-Associated Chest Pain Zaher Fanari, MDa, Kevin K. Kennedy, MDb, Michael J. Lim, MDc, Abhay A. Laddu, MDc, and Joshua M. Stolker, MDc,* Beta blockers are indicated for management of acute coronary syndromes, but they generally are withheld in patients with cocaine-associated chest pain because of concerns for adverse outcomes related to the unique physiological effects of cocaine. Because few clinical studies have evaluated this interaction, we identified patients with toxicology screen results positive for cocaine treated for chest pain at 2 academic hospitals. Clinical characteristics and in-hospital outcomes were compared between patients with and without b-blocker therapy. We then constructed propensity scores to evaluate the independent relation between b-blocker use and the composite primary end point of myocardial infarction, stroke, ventricular arrhythmia, or all-cause mortality after adjusting for clinical characteristics. Of 376 consecutive patients with cocaine-related chest pain, b blockers were used in 164 (44%). Compared with no b blockers, patients treated with b blockers were more likely to describe anginal chest pain, to have known cardiovascular risk factors, and to receive other antiatherosclerotic therapies. Despite these higher risk clinical characteristics, patients treated with b blockers experienced similar peak troponin levels, individual adverse events, and rates of the composite primary end point (15.9% vs 12.3%, p [ 0.32). The primary end point also was similar after propensity score analysis (odds ratio 1.37, 95% confidence interval 0.64 to 2.93, p [ 0.42), including specific comparisons of beta-1 selective (odds ratio 1.83, 95% confidence interval 0.79 to 4.24) and nonselective (odds ratio 0.90, 95% confidence interval 0.33 to 2.42) b blockers, when compared with patients not receiving b blockers. In conclusion, no differences in outcomes were observed between patients treated versus not treated with b-blocker therapy in the setting of cocaine-related chest pain. Ó 2014 Elsevier Inc. All rights reserved. (Am J Cardiol 2014;113:1802e1806)
Several clinical studies, including a retrospective analysis by Rangel et al,1 have demonstrated the safety of b-blocker (BB) use in patients with cocaine-associated chest pain, including similar or lower rates of in-hospital complications and mortality.1e3 However, current guidelines recommend using other medications instead of BB therapy for patients with acute coronary syndrome (ACS) and recent cocaine ingestion, largely based on mechanistic studies reporting unopposed a-adrenergic stimulation with concurrent BB administration.4e6 To better characterize contemporary practice patterns in patients with ACS and cocaine intoxication, we evaluated the rates and types of BB therapy used in 2 urban hospitals, and we studied the interaction between BB use and adverse in-hospital outcomes in this population.
a Division of Cardiology, Christiana Care Health System, Newark, Delaware; bSaint Luke’s Mid America Heart and Vascular Institute, Kansas City, Missouri; and cCenter for Comprehensive Cardiovascular Care, Saint Louis University, St. Louis, Missouri. Manuscript received November 29, 2013; revised manuscript received and accepted March 4, 2014. See page 1806 for disclosure information. *Corresponding author: Tel: (314) 577-8877; fax: (314) 577-8861. E-mail address: [email protected] (J.M. Stolker).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.03.010
Methods We studied patients presenting to the 2 major hospitals affiliated with an academic teaching university from January 2000 to December 2007 with urine toxicology screen results positive for cocaine (immunoassay urine drug test; Ortho-Clinical Diagnostics, Rochester, New York) and reported cocaine use within the previous 24 hours. Of these cocaine-positive patients, we then identified those subjects with International Classification of Diseases, ninth revision, codes for chest pain, angina, or chest discomfort. All charts were reviewed manually by physician members of the research team to exclude patients with chest pain diagnosed as pulmonary in etiology while in the emergency department (ED), such as pneumonia or pulmonary embolus. We collected all cardiac medications administered within the first 24 hours of presentation, electrocardiographic (ECG) findings, telemetry strips, and notes in the hospital chart. Records were analyzed for information about clinical presentations, previously prescribed medications, clinical course in the ED, medications prescribed during the hospitalization, timing of BB therapy, and important in-hospital outcomes including death, myocardial infarction (MI), stroke, or ventricular arrhythmia. www.ajconline.org
Coronary Artery Disease/Beta Blockers and Cocaine Chest Pain
Chest pain was defined as typical versus atypical according to standard definitions.7 Typical chest pain was reported as a pressure-like sensation (heavy or squeezing feeling), particularly if associated with exertion. Atypical chest pain was either sharp or defined in a specific area. Hypertension, dyslipidemia, and diabetes were considered to be present if previously established in the patient’s medical history or if taking an antihypertensive, lipidlowering, or antihyperglycemic medication, respectively. Coronary artery disease was present if the patient had a previous coronary revascularization procedure, documented MI, or known coronary stenosis 50% in severity. Vital signs, laboratory studies, and electrocardiograms were those obtained clinically on presentation to the ED. Stress test and echocardiographic data were collected from studies obtained clinically, and diagnoses were those determined clinically (e.g., left ventricular hypertrophy and wall motion abnormality on echocardiogram). Risk scores from the Thrombolysis In Myocardial Infarction (TIMI) studies were calculated after collecting the TIMI risk factors during chart review.8 Acute MI during the index hospitalization was defined by a troponin-T level greater than the upper limit of normal (>0.1 ng/dl at our hospital laboratory), or by significant ST-segment elevations in 2 contiguous leads by electrocardiography, associated with chest pain or anginal equivalent.5 Acute MI associated with BB use was defined as a documented increase in angina, troponin, or ECG changes after BB administration. Stroke and ventricular arrhythmias were identified clinically by treating physicians during the chest pain hospitalization. The primary end point for this study was the composite clinical outcome of MI, stroke, ventricular arrhythmia, or all-cause mortality. Characteristics of patients with and without in-hospital BB use within the first 24 hours were compared using chi-square for categorical, t test for continuous, and Wilcoxon rank sum and other appropriate nonparametric tests for variables without normal distribution. Given the significant differences in baseline characteristics between the groups of patients with and without BB use, we constructed propensity scores to evaluate the independent relation between BB use and the composite primary end point.9e11 The propensity score for an individual is defined as the conditional probability of being treated (in this case, getting BBs) given the individual’s covariates or characteristics, and it is designed to balance a large number of potential confounders equally across 2 observational cohorts of patients.10 Because quintile stratification removes 90% of the bias due to unbalanced factors,11 we stratified patients in our cohort by propensity score quintile and constructed a logistic regression model adjusted for propensity score to calculate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for experiencing the composite primary end point. With many clinicians choosing nonselective BBs in these patients, as a secondary analysis we reran the regression models to compare rates of the primary end point between patients receiving beta-1 selective BBs (atenolol and metoprolol) and those with no BB therapy, and again for nonselective BBs (carvedilol and labetalol) versus no BB therapy. Statistical significance was defined as p 0.05. All analyses were performed using SAS, version 9.2 (SAS Institute, Cary, North Carolina). The
1803
Table 1 Baseline characteristics, treatments, and disposition of patients with and without b-blocker (BB) therapy Variable
BB Therapy
p Value
Yes (n ¼ 164) No (n ¼ 212) Patient characteristics Age (yrs) Women White Black Chest pain typical for angina pectoris TIMI risk score 0 1 2 3 History of hyperlipidemia History of hypertension Diabetes mellitus Known coronary artery disease Lung disease Heart rate (beats/min) Respiratory rate (breaths per minute) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Hemoglobin (g/dl) Creatinine (mg/dl) Total cholesterol (mg/dl) Low-density lipoprotein cholesterol (mg/dl) High-density lipoprotein cholesterol (mg/dl) Triglycerides (mg/dl) Other medications ordered Aspirin Clopidogrel Systemic anticoagulation Nitrates Angiotensin-converting enzyme inhibitors Disposition Cardiac intensive care unit Medical intensive care unit Telemetry floor Home from ED
46 9 32% 10% 81% 17%
42 11 26% 13% 76% 8%
<0.001 0.18 0.36 0.23 0.012
44% 32% 18% 6% 23% 80% 25% 33%
69% 23% 5% 3% 11% 38% 14% 13%
<0.001 <0.001 0.005 <0.001
9% 91 22 21 5
8% 94 25 21 7
0.58 0.29 0.50
152 36
138 32
<0.001
89 26
80 22
<0.001
13 3 2.1 2.4 155 46 96 34
14 2 1.7 2.3 163 47 100 43
0.16 0.20 0.26 0.54
41 14
44 18
0.22
107 94
113 83
0.67
84% 23% 23% 50% 72%
57% 8% 9% 36% 24%
<0.001 <0.001 <0.001 0.007 <0.001
22% 19% 58% 1%
8% 11% 77% 4%
<0.001 0.038 <0.001 0.20
<0.001
Data are expressed as proportion (%) or mean SD.
study protocol was reviewed and approved by the university’s Institutional Review Board. Results A total of 376 patients with positive urine toxicology screen results for cocaine were admitted with chest pain, of whom 164 (44%) received a BB in the first 24 hours. The type of BB given was metoprolol in 74, carvedilol in 43, labetalol in 44, and atenolol in 3 patients. As listed in Table 1, patients who received BBs were older and had more cardiovascular risk factors, more established coronary
1804
The American Journal of Cardiology (www.ajconline.org)
Table 2 Cardiovascular findings between patients with and without b-blocker (BB) therapy Cardiac Characteristic
BB Use
ECG changes ST-segment depression 1 mm ST-segment elevation 1 mm T-wave inversion Cardiac biomarkers CKMB at presentation (ng/ml) Peak CKMB (ng/ml) Troponin T at presentation (ng/ml) Peak troponin T (ng/ml) Echocardiographic data Left ventricular hypertrophy Wall motion abnormality at rest Left ventricular ejection fraction Stress test Obtained during hospitalization Abnormal, when obtained
p Value
Yes (n ¼ 164)
No (n ¼ 212)
7 7 56
3 4 26
3.2 (1.8e5.5) 3.2 (1.7e7.1) 0.04 (0.01e0.10) 0.04 (0.01e0.12)
4.5 4.9 0.04 0.04
0.10 0.19 <0.001
(2.5e10.5) (2.7e12.2) (0.01e0.12) (0.01e0.21)
0.006 <0.001 0.12 0.14
36 32 0.41 0.17
35 28 0.49 0.17
0.84 0.57 0.006
22 25
11 21
0.005 0.71
Data are expressed as proportion (%), mean SD, or median (interquartile range) given the lack of normal distribution for cardiac biomarkers. CKMB ¼ creatine kinase-MB subfraction. Table 3 In-hospital outcomes before and after propensity score matching Outcome
Composite end point Death Stroke MI Arrhythmia
BB (%)
p Value
Yes (n ¼ 164)
No (n ¼ 212)
16 3 3 5 7
12 6 3 2 5
artery disease, higher presenting blood pressures, and higher TIMI risk scores. In addition, these patients were more likely to be treated with other medications for ischemic heart disease and were more likely to be admitted to an intensive care unit. Cardiovascular findings between patients with and without BB therapy are listed in Table 2. There were no significant differences in ST-segment depression or elevation on electrocardiogram nor admission or peak troponin, but both admission and peak MB subfraction of creatine kinase levels were higher in the patients not receiving BBs. Patients treated with BBs were noted to have worse left ventricular systolic function by echocardiography, but aside from higher rates of performing inpatient stress tests, all other echocardiographic and stress test findings were similar between patients with and without BB therapy. Despite higher baseline risk overall, patients with BB therapy experienced similar rates of the composite end point and its components, compared with patients without BB therapy (Table 3). When constructing the propensity scores, baseline clinical characteristics before treatment were incorporated: age, gender, race, character of chest pain, hypertension, dyslipidemia, diabetes, known human immunodeficiency virus infection, previous lung disease, previous coronary artery disease, TIMI risk score, blood pressures, and initial ECG
0.32 0.13 1.00 0.18 0.52
OR (95% CI) After Propensity Score Matching 1.37 0.66 1.20 1.67 1.24
(0.64e2.93) (0.14e3.04) (0.27e5.36) (0.44e6.35) (0.38e3.98)
p Value
0.42 0.60 0.81 0.46 0.72
findings and cardiac biomarkers. C statistic for the risk score was 0.82, and whisker plot construction confirmed significant differences between the groups of patients receiving versus not receiving BBs, so that prediction of BB therapy using the propensity score was valid. After incorporating propensity score quintiles into the multivariable analysis to account for baseline differences between groups, findings were similar to the unadjusted comparisons, as the presence of BB therapy during hospitalization was not related to the composite clinical end point after adjusting for propensity score quintile (OR 1.34, 95% CI 0.64 to 2.93, p ¼ 0.42, compared with no BB therapy). Furthermore, no difference was noted for any of the individual clinical outcomes after propensity score analysis (see Table 3). Similar results were noted in the secondary analysis of BB subtypes, as there was no independent relation between beta-1 selective BBs (OR 1.83, 95% CI 0.79 to 4.24) and patients not receiving BBs, or between nonselective BBs (OR 0.90, 95% CI 0.33 to 2.42) and no BB therapy. Discussion In this retrospective analysis of patients evaluated in the ED with cocaine-related chest pain, we found that
Coronary Artery Disease/Beta Blockers and Cocaine Chest Pain
1805
Figure 1. Mechanisms of myocardial ischemia related to cocaine exposure. NO ¼ nitric oxide.
44% of patients received BB therapy within the first 24 hours of presentation, despite contemporary clinical guidelines recommending against this treatment.4e6 Compared with those not receiving BB therapy, patients treated with BBs had higher rates of cardiovascular risk factors and antiatherosclerotic therapies at baseline, despite relatively similar clinical presentations. Most notably, BB use was not related to adverse clinical outcomes either before or after multivariable adjustment, even when using propensity score analysis to account for differences in demographic and clinical characteristics between groups. These findings suggest that clinicians are continuing to use BBs in patients with active cocaine intoxication, particularly in those at higher risk for experiencing adverse cardiovascular events, and that BB use in these at-risk subjects may be safe. Extensive study of the relation between cocaine and myocardial ischemia has been performed, with proposed mechanisms including catecholamine accumulation,12,13 thrombosis,14e19 premature atherosclerosis,14 and coronary vasoconstriction.20 The net effect of these pathophysiological insults is the promotion of myocardial ischemia through multiple pathways (Figure 1), many of which could be modified with BB therapy. Clinical guidelines recommend withholding of BB therapy in patients with cocaine-associated ACS, primarily because of concerns about unopposed aadrenergic stimulation4e6 and the potential for coronary vasoconstriction, increased blood pressure, and decreased coronary blood flow with BB therapy.21e25 However, recent clinical studies have challenged this presumption, and some have even suggested lower incidence of ACS in patients with cocaine use who are treated with BBs.1e3 In the analysis by Rangel et al,1 in which BB use at hospital discharge and longer term mortality also were evaluated, cocaine users treated with BBs had no differences in ECG presentation, troponin levels, ventricular arrhythmias, or death. Another study also demonstrated no difference in adverse outcomes when comparing beta-1 selective and nonselective BBs.3 Our analysis thus expands on these data by creating propensity scores to account for the significant differences in baseline characteristics between patients with and without BB therapy for cocaine-associated chest pain. In addition, we compared
the outcomes of patients receiving beta-1 selective and nonselective BBs against those subjects receiving no BB therapy, in separate confirmatory analyses, again demonstrating no independent relation between either subgroup of BBs and adverse clinical events during the chest pain hospitalization. Whether the cardioprotective effect of BBs outweighs the risks associated with unopposed a-adrenergic stimulation in selected patients with high-risk features for ACS, or in specific subgroups of patients with cocaine intoxication, requires further study in a prospective trial. The findings from our study should be interpreted in the context of several limitations. First, the retrospective design is subject to selection bias and data quality issues related to chart documentation. We tried to mitigate some of the bias attributed to administrative data (i.e., International Classification of Diseases, ninth revision, codes alone) by way of physician review of each clinical chart. We also used propensity score analysis to help minimize unmeasured confounding, but other variables may have influenced our results outside of the clinical characteristics collected in our study. Of note, adverse events occurred equally in both BBand noneBB-treated patients, despite the significantly higher risk profile in the BB-treated subgroup. This suggests that BB therapy was at least neutral for its relation with the composite clinical outcome, and it could even potentially have had a protective effect in these higher risk patients. Another limitation is the inability to distinguish precise timing of cocaine ingestion, as urine toxicology screen results may remain abnormal for 48 to 72 hours after cocaine use.26,27 Despite this concern, our data likely reflect “realworld” management of cocaine chest pain in the ED, as cocaine exposure may contribute to myocardial injury for up to 4 days after ingestion.28 Another important consideration is the possibility that other diseases contributed to or interacted with the outcomes we measured, as patients who use cocaine are more likely to have other serious (and often undertreated) diseases. Interactions among these conditions, the drugs used to treat them, and BB therapy in patients with cocaine toxicity is beyond the scope of our study but should be evaluated in future analyses of patients with cocaine exposure. Despite these limitations, BB treatment in the setting of cocaine-induced chest pain may be safer than
1806
The American Journal of Cardiology (www.ajconline.org)
previously suggested. Prospective study of BBs versus other ACS medications for higher risk patients is warranted to help guide clinical management of patients with cocainerelated chest pain.
Disclosures The authors have no conflicts of interest to disclose. 1. Rangel C, Shu RG, Lazar LD, Vittinghoff E, Hsue PY, Marcus GM. Beta-blockers for chest pain associated with recent cocaine use. Arch Intern Med 2010;170:874e879. 2. Dattilo PB, Hailpern SM, Fearon K, Sohal D, Nordin C. Beta-blockers are associated with reduced risk of myocardial infarction after cocaine use. Ann Emerg Med 2008;51:117e125. 3. Ibrahim M, Maselli DJ, Hasan R, Hamilton A. Safety of beta-blockers in the acute management of cocaine-associated chest pain. Am J Emerg Med 2012;31:613e616. 4. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2005;112:IV1e203. 5. Antman EM, Anbe DT, Armstrong PW, Bates ER, Green LA, Hand M, Hochman JS, Krumholz HM, Kushner FG, Lamas GA, Mullany CJ, Ornato JP, Pearle DL, Sloan MA, Smith SC Jr, Alpert JS, Anderson JL, Faxon DP, Fuster V, Gibbons RJ, Gregoratos G, Halperin JL, Hiratzka LF, Hunt SA, Jacobs AK. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1999 Guidelines for the Management of Patients with Acute Myocardial Infarction). Circulation 2004;110:e82e292. 6. McCord J, Jneid H, Hollander JE, de Lemos JA, Cercek B, Hsue P, Gibler WB, Ohman EM, Drew B, Philippides G, Newby LK. Management of cocaine-associated chest pain and myocardial infarction: a scientific statement from the American Heart Association Acute Cardiac Care Committee of the Council on Clinical Cardiology. Circulation 2008;117:1897e1907. 7. Thomas HL. “Chapter 12. Chest discomfort” (Chapter). In: Fauci AS BE, Kasper DL, Hauser SL, Longo DL, Jameson JL, Loscalzo J, ed. Harrison’s Principles of Internal Medicine, 18th ed., 2011. 8. Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T, Papuchis G, Mautner B, Corbalan R, Radley D, Braunwald E. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;284: 835e842. 9. Austin PC. The performance of different propensity-score methods for estimating relative risks. J Clin Epidemiol 2008;61:537e545.
10. Austin PC. The relative ability of different propensity score methods to balance measured covariates between treated and untreated subjects in observational studies. Med Decis Making 2009;29:661e677. 11. Rosenbaum PR. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 12. Muscholl E. Effect of cocaine and related drugs on the uptake of noradrenaline by heart and spleen. Br J Pharmacol Chemother 1961;16:352e359. 13. Whitby LG, Hertting G, Axelrod J. Effect of cocaine on the disposition of noradrenaline labelled with tritium. Nature 1960;187:604e605. 14. Foltin RW, Ward AS, Haney M, Hart CL, Collins ED. The effects of escalating doses of smoked cocaine in humans. Drug Alcohol Depend 2003;70:149e157. 15. Kugelmass AD, Oda A, Monahan K, Cabral C, Ware JA. Activation of human platelets by cocaine. Circulation 1993;88:876e883. 16. Moliterno DJ, Lange RA, Gerard RD, Willard JE, Lackner C, Hillis LD. Influence of intranasal cocaine on plasma constituents associated with endogenous thrombosis and thrombolysis. Am J Med 1994;96: 492e496. 17. Rezkalla SH, Mazza JJ, Kloner RA, Tillema V, Chang SH. Effects of cocaine on human platelets in healthy subjects. Am J Cardiol 1993;72: 243e246. 18. Rinder HM, Ault KA, Jatlow PI, Kosten TR, Smith BR. Platelet alphagranule release in cocaine users. Circulation 1994;90:1162e1167. 19. Stenberg RG, Winniford MD, Hillis LD, Dowling GP, Buja LM. Simultaneous acute thrombosis of two major coronary arteries following intravenous cocaine use. Arch Pathol Lab Med 1989;113: 521e524. 20. Tazelaar HD, Karch SB, Stephens BG, Billingham ME. Cocaine and the heart. Hum Pathol 1987;18:195e199. 21. Ramoska E, Sacchetti AD. Propranolol-induced hypertension in treatment of cocaine intoxication. Ann Emerg Med 1985;14:1112e1113. 22. Rappolt RT Sr, Gay G, Inaba DS, Rappolt N, Rappolt RT Jr. Use of Inderal (propranolol-Ayerst) in I-a (early stimulative) and I-b (advanced stimulative) classification of cocaine and other sympathomimetic reactions. Clin Toxicol 1978;13:325e332. 23. Rappolt RT, Gay G, Inaba DS, Rappolt NR. Propranolol in cocaine toxicity. Lancet 1976;2:640e641. 24. Smith M, Garner D, Niemann JT. Pharmacologic interventions after an LD50 cocaine insult in a chronically instrumented rat model: are betablockers contraindicated? Ann Emerg Med 1991;20:768e771. 25. Vargas R, Gillis RA, Ramwell PW. Propranolol promotes cocaineinduced spasm of porcine coronary artery. J Pharmacol Exp Ther 1991;257:644e646. 26. Das G. Cardiovascular effects of cocaine abuse. Int J Clin Pharmacol Ther Toxicol 1993;31:521e528. 27. Om A, Ellahham S, DiSciascio G. Management of cocaine-induced cardiovascular complications. Am Heart J 1993;125:469e475. 28. Hollander JE, Hoffman RS. Cocaine-induced myocardial infarction: an analysis and review of the literature. J Emerg Med 1992;10:169e177.
Several clinical studies, including a retrospective analysis by Rangel et al,1 have demonstrated the safety of b-blocker (BB) use in patients with cocaine-associated chest pain, including similar or lower rates of in-hospital complications and mortality.1e3 However, current guidelines recommend using other medications instead of BB therapy for patients with acute coronary syndrome (ACS) and recent cocaine ingestion, largely based on mechanistic studies reporting unopposed a-adrenergic stimulation with concurrent BB administration.4e6 To better characterize contemporary practice patterns in patients with ACS and cocaine intoxication, we evaluated the rates and types of BB therapy used in 2 urban hospitals, and we studied the interaction between BB use and adverse in-hospital outcomes in this population.
a Division of Cardiology, Christiana Care Health System, Newark, Delaware; bSaint Luke’s Mid America Heart and Vascular Institute, Kansas City, Missouri; and cCenter for Comprehensive Cardiovascular Care, Saint Louis University, St. Louis, Missouri. Manuscript received November 29, 2013; revised manuscript received and accepted March 4, 2014. See page 1806 for disclosure information. *Corresponding author: Tel: (314) 577-8877; fax: (314) 577-8861. E-mail address: [email protected] (J.M. Stolker).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.03.010
Methods We studied patients presenting to the 2 major hospitals affiliated with an academic teaching university from January 2000 to December 2007 with urine toxicology screen results positive for cocaine (immunoassay urine drug test; Ortho-Clinical Diagnostics, Rochester, New York) and reported cocaine use within the previous 24 hours. Of these cocaine-positive patients, we then identified those subjects with International Classification of Diseases, ninth revision, codes for chest pain, angina, or chest discomfort. All charts were reviewed manually by physician members of the research team to exclude patients with chest pain diagnosed as pulmonary in etiology while in the emergency department (ED), such as pneumonia or pulmonary embolus. We collected all cardiac medications administered within the first 24 hours of presentation, electrocardiographic (ECG) findings, telemetry strips, and notes in the hospital chart. Records were analyzed for information about clinical presentations, previously prescribed medications, clinical course in the ED, medications prescribed during the hospitalization, timing of BB therapy, and important in-hospital outcomes including death, myocardial infarction (MI), stroke, or ventricular arrhythmia. www.ajconline.org
Coronary Artery Disease/Beta Blockers and Cocaine Chest Pain
Chest pain was defined as typical versus atypical according to standard definitions.7 Typical chest pain was reported as a pressure-like sensation (heavy or squeezing feeling), particularly if associated with exertion. Atypical chest pain was either sharp or defined in a specific area. Hypertension, dyslipidemia, and diabetes were considered to be present if previously established in the patient’s medical history or if taking an antihypertensive, lipidlowering, or antihyperglycemic medication, respectively. Coronary artery disease was present if the patient had a previous coronary revascularization procedure, documented MI, or known coronary stenosis 50% in severity. Vital signs, laboratory studies, and electrocardiograms were those obtained clinically on presentation to the ED. Stress test and echocardiographic data were collected from studies obtained clinically, and diagnoses were those determined clinically (e.g., left ventricular hypertrophy and wall motion abnormality on echocardiogram). Risk scores from the Thrombolysis In Myocardial Infarction (TIMI) studies were calculated after collecting the TIMI risk factors during chart review.8 Acute MI during the index hospitalization was defined by a troponin-T level greater than the upper limit of normal (>0.1 ng/dl at our hospital laboratory), or by significant ST-segment elevations in 2 contiguous leads by electrocardiography, associated with chest pain or anginal equivalent.5 Acute MI associated with BB use was defined as a documented increase in angina, troponin, or ECG changes after BB administration. Stroke and ventricular arrhythmias were identified clinically by treating physicians during the chest pain hospitalization. The primary end point for this study was the composite clinical outcome of MI, stroke, ventricular arrhythmia, or all-cause mortality. Characteristics of patients with and without in-hospital BB use within the first 24 hours were compared using chi-square for categorical, t test for continuous, and Wilcoxon rank sum and other appropriate nonparametric tests for variables without normal distribution. Given the significant differences in baseline characteristics between the groups of patients with and without BB use, we constructed propensity scores to evaluate the independent relation between BB use and the composite primary end point.9e11 The propensity score for an individual is defined as the conditional probability of being treated (in this case, getting BBs) given the individual’s covariates or characteristics, and it is designed to balance a large number of potential confounders equally across 2 observational cohorts of patients.10 Because quintile stratification removes 90% of the bias due to unbalanced factors,11 we stratified patients in our cohort by propensity score quintile and constructed a logistic regression model adjusted for propensity score to calculate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for experiencing the composite primary end point. With many clinicians choosing nonselective BBs in these patients, as a secondary analysis we reran the regression models to compare rates of the primary end point between patients receiving beta-1 selective BBs (atenolol and metoprolol) and those with no BB therapy, and again for nonselective BBs (carvedilol and labetalol) versus no BB therapy. Statistical significance was defined as p 0.05. All analyses were performed using SAS, version 9.2 (SAS Institute, Cary, North Carolina). The
1803
Table 1 Baseline characteristics, treatments, and disposition of patients with and without b-blocker (BB) therapy Variable
BB Therapy
p Value
Yes (n ¼ 164) No (n ¼ 212) Patient characteristics Age (yrs) Women White Black Chest pain typical for angina pectoris TIMI risk score 0 1 2 3 History of hyperlipidemia History of hypertension Diabetes mellitus Known coronary artery disease Lung disease Heart rate (beats/min) Respiratory rate (breaths per minute) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Hemoglobin (g/dl) Creatinine (mg/dl) Total cholesterol (mg/dl) Low-density lipoprotein cholesterol (mg/dl) High-density lipoprotein cholesterol (mg/dl) Triglycerides (mg/dl) Other medications ordered Aspirin Clopidogrel Systemic anticoagulation Nitrates Angiotensin-converting enzyme inhibitors Disposition Cardiac intensive care unit Medical intensive care unit Telemetry floor Home from ED
46 9 32% 10% 81% 17%
42 11 26% 13% 76% 8%
<0.001 0.18 0.36 0.23 0.012
44% 32% 18% 6% 23% 80% 25% 33%
69% 23% 5% 3% 11% 38% 14% 13%
<0.001 <0.001 0.005 <0.001
9% 91 22 21 5
8% 94 25 21 7
0.58 0.29 0.50
152 36
138 32
<0.001
89 26
80 22
<0.001
13 3 2.1 2.4 155 46 96 34
14 2 1.7 2.3 163 47 100 43
0.16 0.20 0.26 0.54
41 14
44 18
0.22
107 94
113 83
0.67
84% 23% 23% 50% 72%
57% 8% 9% 36% 24%
<0.001 <0.001 <0.001 0.007 <0.001
22% 19% 58% 1%
8% 11% 77% 4%
<0.001 0.038 <0.001 0.20
<0.001
Data are expressed as proportion (%) or mean SD.
study protocol was reviewed and approved by the university’s Institutional Review Board. Results A total of 376 patients with positive urine toxicology screen results for cocaine were admitted with chest pain, of whom 164 (44%) received a BB in the first 24 hours. The type of BB given was metoprolol in 74, carvedilol in 43, labetalol in 44, and atenolol in 3 patients. As listed in Table 1, patients who received BBs were older and had more cardiovascular risk factors, more established coronary
1804
The American Journal of Cardiology (www.ajconline.org)
Table 2 Cardiovascular findings between patients with and without b-blocker (BB) therapy Cardiac Characteristic
BB Use
ECG changes ST-segment depression 1 mm ST-segment elevation 1 mm T-wave inversion Cardiac biomarkers CKMB at presentation (ng/ml) Peak CKMB (ng/ml) Troponin T at presentation (ng/ml) Peak troponin T (ng/ml) Echocardiographic data Left ventricular hypertrophy Wall motion abnormality at rest Left ventricular ejection fraction Stress test Obtained during hospitalization Abnormal, when obtained
p Value
Yes (n ¼ 164)
No (n ¼ 212)
7 7 56
3 4 26
3.2 (1.8e5.5) 3.2 (1.7e7.1) 0.04 (0.01e0.10) 0.04 (0.01e0.12)
4.5 4.9 0.04 0.04
0.10 0.19 <0.001
(2.5e10.5) (2.7e12.2) (0.01e0.12) (0.01e0.21)
0.006 <0.001 0.12 0.14
36 32 0.41 0.17
35 28 0.49 0.17
0.84 0.57 0.006
22 25
11 21
0.005 0.71
Data are expressed as proportion (%), mean SD, or median (interquartile range) given the lack of normal distribution for cardiac biomarkers. CKMB ¼ creatine kinase-MB subfraction. Table 3 In-hospital outcomes before and after propensity score matching Outcome
Composite end point Death Stroke MI Arrhythmia
BB (%)
p Value
Yes (n ¼ 164)
No (n ¼ 212)
16 3 3 5 7
12 6 3 2 5
artery disease, higher presenting blood pressures, and higher TIMI risk scores. In addition, these patients were more likely to be treated with other medications for ischemic heart disease and were more likely to be admitted to an intensive care unit. Cardiovascular findings between patients with and without BB therapy are listed in Table 2. There were no significant differences in ST-segment depression or elevation on electrocardiogram nor admission or peak troponin, but both admission and peak MB subfraction of creatine kinase levels were higher in the patients not receiving BBs. Patients treated with BBs were noted to have worse left ventricular systolic function by echocardiography, but aside from higher rates of performing inpatient stress tests, all other echocardiographic and stress test findings were similar between patients with and without BB therapy. Despite higher baseline risk overall, patients with BB therapy experienced similar rates of the composite end point and its components, compared with patients without BB therapy (Table 3). When constructing the propensity scores, baseline clinical characteristics before treatment were incorporated: age, gender, race, character of chest pain, hypertension, dyslipidemia, diabetes, known human immunodeficiency virus infection, previous lung disease, previous coronary artery disease, TIMI risk score, blood pressures, and initial ECG
0.32 0.13 1.00 0.18 0.52
OR (95% CI) After Propensity Score Matching 1.37 0.66 1.20 1.67 1.24
(0.64e2.93) (0.14e3.04) (0.27e5.36) (0.44e6.35) (0.38e3.98)
p Value
0.42 0.60 0.81 0.46 0.72
findings and cardiac biomarkers. C statistic for the risk score was 0.82, and whisker plot construction confirmed significant differences between the groups of patients receiving versus not receiving BBs, so that prediction of BB therapy using the propensity score was valid. After incorporating propensity score quintiles into the multivariable analysis to account for baseline differences between groups, findings were similar to the unadjusted comparisons, as the presence of BB therapy during hospitalization was not related to the composite clinical end point after adjusting for propensity score quintile (OR 1.34, 95% CI 0.64 to 2.93, p ¼ 0.42, compared with no BB therapy). Furthermore, no difference was noted for any of the individual clinical outcomes after propensity score analysis (see Table 3). Similar results were noted in the secondary analysis of BB subtypes, as there was no independent relation between beta-1 selective BBs (OR 1.83, 95% CI 0.79 to 4.24) and patients not receiving BBs, or between nonselective BBs (OR 0.90, 95% CI 0.33 to 2.42) and no BB therapy. Discussion In this retrospective analysis of patients evaluated in the ED with cocaine-related chest pain, we found that
Coronary Artery Disease/Beta Blockers and Cocaine Chest Pain
1805
Figure 1. Mechanisms of myocardial ischemia related to cocaine exposure. NO ¼ nitric oxide.
44% of patients received BB therapy within the first 24 hours of presentation, despite contemporary clinical guidelines recommending against this treatment.4e6 Compared with those not receiving BB therapy, patients treated with BBs had higher rates of cardiovascular risk factors and antiatherosclerotic therapies at baseline, despite relatively similar clinical presentations. Most notably, BB use was not related to adverse clinical outcomes either before or after multivariable adjustment, even when using propensity score analysis to account for differences in demographic and clinical characteristics between groups. These findings suggest that clinicians are continuing to use BBs in patients with active cocaine intoxication, particularly in those at higher risk for experiencing adverse cardiovascular events, and that BB use in these at-risk subjects may be safe. Extensive study of the relation between cocaine and myocardial ischemia has been performed, with proposed mechanisms including catecholamine accumulation,12,13 thrombosis,14e19 premature atherosclerosis,14 and coronary vasoconstriction.20 The net effect of these pathophysiological insults is the promotion of myocardial ischemia through multiple pathways (Figure 1), many of which could be modified with BB therapy. Clinical guidelines recommend withholding of BB therapy in patients with cocaine-associated ACS, primarily because of concerns about unopposed aadrenergic stimulation4e6 and the potential for coronary vasoconstriction, increased blood pressure, and decreased coronary blood flow with BB therapy.21e25 However, recent clinical studies have challenged this presumption, and some have even suggested lower incidence of ACS in patients with cocaine use who are treated with BBs.1e3 In the analysis by Rangel et al,1 in which BB use at hospital discharge and longer term mortality also were evaluated, cocaine users treated with BBs had no differences in ECG presentation, troponin levels, ventricular arrhythmias, or death. Another study also demonstrated no difference in adverse outcomes when comparing beta-1 selective and nonselective BBs.3 Our analysis thus expands on these data by creating propensity scores to account for the significant differences in baseline characteristics between patients with and without BB therapy for cocaine-associated chest pain. In addition, we compared
the outcomes of patients receiving beta-1 selective and nonselective BBs against those subjects receiving no BB therapy, in separate confirmatory analyses, again demonstrating no independent relation between either subgroup of BBs and adverse clinical events during the chest pain hospitalization. Whether the cardioprotective effect of BBs outweighs the risks associated with unopposed a-adrenergic stimulation in selected patients with high-risk features for ACS, or in specific subgroups of patients with cocaine intoxication, requires further study in a prospective trial. The findings from our study should be interpreted in the context of several limitations. First, the retrospective design is subject to selection bias and data quality issues related to chart documentation. We tried to mitigate some of the bias attributed to administrative data (i.e., International Classification of Diseases, ninth revision, codes alone) by way of physician review of each clinical chart. We also used propensity score analysis to help minimize unmeasured confounding, but other variables may have influenced our results outside of the clinical characteristics collected in our study. Of note, adverse events occurred equally in both BBand noneBB-treated patients, despite the significantly higher risk profile in the BB-treated subgroup. This suggests that BB therapy was at least neutral for its relation with the composite clinical outcome, and it could even potentially have had a protective effect in these higher risk patients. Another limitation is the inability to distinguish precise timing of cocaine ingestion, as urine toxicology screen results may remain abnormal for 48 to 72 hours after cocaine use.26,27 Despite this concern, our data likely reflect “realworld” management of cocaine chest pain in the ED, as cocaine exposure may contribute to myocardial injury for up to 4 days after ingestion.28 Another important consideration is the possibility that other diseases contributed to or interacted with the outcomes we measured, as patients who use cocaine are more likely to have other serious (and often undertreated) diseases. Interactions among these conditions, the drugs used to treat them, and BB therapy in patients with cocaine toxicity is beyond the scope of our study but should be evaluated in future analyses of patients with cocaine exposure. Despite these limitations, BB treatment in the setting of cocaine-induced chest pain may be safer than
1806
The American Journal of Cardiology (www.ajconline.org)
previously suggested. Prospective study of BBs versus other ACS medications for higher risk patients is warranted to help guide clinical management of patients with cocainerelated chest pain.
Disclosures The authors have no conflicts of interest to disclose. 1. Rangel C, Shu RG, Lazar LD, Vittinghoff E, Hsue PY, Marcus GM. Beta-blockers for chest pain associated with recent cocaine use. Arch Intern Med 2010;170:874e879. 2. Dattilo PB, Hailpern SM, Fearon K, Sohal D, Nordin C. Beta-blockers are associated with reduced risk of myocardial infarction after cocaine use. Ann Emerg Med 2008;51:117e125. 3. Ibrahim M, Maselli DJ, Hasan R, Hamilton A. Safety of beta-blockers in the acute management of cocaine-associated chest pain. Am J Emerg Med 2012;31:613e616. 4. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2005;112:IV1e203. 5. Antman EM, Anbe DT, Armstrong PW, Bates ER, Green LA, Hand M, Hochman JS, Krumholz HM, Kushner FG, Lamas GA, Mullany CJ, Ornato JP, Pearle DL, Sloan MA, Smith SC Jr, Alpert JS, Anderson JL, Faxon DP, Fuster V, Gibbons RJ, Gregoratos G, Halperin JL, Hiratzka LF, Hunt SA, Jacobs AK. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1999 Guidelines for the Management of Patients with Acute Myocardial Infarction). Circulation 2004;110:e82e292. 6. McCord J, Jneid H, Hollander JE, de Lemos JA, Cercek B, Hsue P, Gibler WB, Ohman EM, Drew B, Philippides G, Newby LK. Management of cocaine-associated chest pain and myocardial infarction: a scientific statement from the American Heart Association Acute Cardiac Care Committee of the Council on Clinical Cardiology. Circulation 2008;117:1897e1907. 7. Thomas HL. “Chapter 12. Chest discomfort” (Chapter). In: Fauci AS BE, Kasper DL, Hauser SL, Longo DL, Jameson JL, Loscalzo J, ed. Harrison’s Principles of Internal Medicine, 18th ed., 2011. 8. Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T, Papuchis G, Mautner B, Corbalan R, Radley D, Braunwald E. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;284: 835e842. 9. Austin PC. The performance of different propensity-score methods for estimating relative risks. J Clin Epidemiol 2008;61:537e545.
10. Austin PC. The relative ability of different propensity score methods to balance measured covariates between treated and untreated subjects in observational studies. Med Decis Making 2009;29:661e677. 11. Rosenbaum PR. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41e55. 12. Muscholl E. Effect of cocaine and related drugs on the uptake of noradrenaline by heart and spleen. Br J Pharmacol Chemother 1961;16:352e359. 13. Whitby LG, Hertting G, Axelrod J. Effect of cocaine on the disposition of noradrenaline labelled with tritium. Nature 1960;187:604e605. 14. Foltin RW, Ward AS, Haney M, Hart CL, Collins ED. The effects of escalating doses of smoked cocaine in humans. Drug Alcohol Depend 2003;70:149e157. 15. Kugelmass AD, Oda A, Monahan K, Cabral C, Ware JA. Activation of human platelets by cocaine. Circulation 1993;88:876e883. 16. Moliterno DJ, Lange RA, Gerard RD, Willard JE, Lackner C, Hillis LD. Influence of intranasal cocaine on plasma constituents associated with endogenous thrombosis and thrombolysis. Am J Med 1994;96: 492e496. 17. Rezkalla SH, Mazza JJ, Kloner RA, Tillema V, Chang SH. Effects of cocaine on human platelets in healthy subjects. Am J Cardiol 1993;72: 243e246. 18. Rinder HM, Ault KA, Jatlow PI, Kosten TR, Smith BR. Platelet alphagranule release in cocaine users. Circulation 1994;90:1162e1167. 19. Stenberg RG, Winniford MD, Hillis LD, Dowling GP, Buja LM. Simultaneous acute thrombosis of two major coronary arteries following intravenous cocaine use. Arch Pathol Lab Med 1989;113: 521e524. 20. Tazelaar HD, Karch SB, Stephens BG, Billingham ME. Cocaine and the heart. Hum Pathol 1987;18:195e199. 21. Ramoska E, Sacchetti AD. Propranolol-induced hypertension in treatment of cocaine intoxication. Ann Emerg Med 1985;14:1112e1113. 22. Rappolt RT Sr, Gay G, Inaba DS, Rappolt N, Rappolt RT Jr. Use of Inderal (propranolol-Ayerst) in I-a (early stimulative) and I-b (advanced stimulative) classification of cocaine and other sympathomimetic reactions. Clin Toxicol 1978;13:325e332. 23. Rappolt RT, Gay G, Inaba DS, Rappolt NR. Propranolol in cocaine toxicity. Lancet 1976;2:640e641. 24. Smith M, Garner D, Niemann JT. Pharmacologic interventions after an LD50 cocaine insult in a chronically instrumented rat model: are betablockers contraindicated? Ann Emerg Med 1991;20:768e771. 25. Vargas R, Gillis RA, Ramwell PW. Propranolol promotes cocaineinduced spasm of porcine coronary artery. J Pharmacol Exp Ther 1991;257:644e646. 26. Das G. Cardiovascular effects of cocaine abuse. Int J Clin Pharmacol Ther Toxicol 1993;31:521e528. 27. Om A, Ellahham S, DiSciascio G. Management of cocaine-induced cardiovascular complications. Am Heart J 1993;125:469e475. 28. Hollander JE, Hoffman RS. Cocaine-induced myocardial infarction: an analysis and review of the literature. J Emerg Med 1992;10:169e177.