Coronary angiographic findings in patients with cocaine-associated chest pain☆

Introduction 

The standard protocol for patients undergoing an ischemic evaluation typically includes coronary angiography in those with myocardial infarction (MI) and those in whom there is a high suspicion of coronary disease (1). In patients for whom there is a low suspicion of coronary disease after MI is excluded, stress testing is typically performed. In patients with cocaine-associated chest pain, the optimal diagnostic strategy has not been defined. In those without MI, the value of additional risk stratification is unclear, because only a small minority undergoes further diagnostic testing (2). Although cocaine-associated MI is often thought to result from coronary vasospasm, the incidence of underlying coronary disease in patients with MI has been relatively high (3). Thus, these patients may represent a select subgroup of patients after cocaine use. In this study, we report the results of coronary angiography performed in a cohort of patients who initially presented to the Emergency Department (ED) complaining of chest pain or who had symptoms consistent with myocardial ischemia associated with cocaine use.

Materials and methods 

This study was performed at a 600-bed inner city hospital, which has approximately 85,000 Emergency Department visits a year. This study includes consecutive patients who underwent evaluation in the ED from June 1994 to January 2001 for symptoms consistent with myocardial ischemia that were associated with recent cocaine use, and who subsequently had coronary angiography within 5 weeks of the ED evaluation. Patients either acknowledged recent cocaine use or had a positive urine drug screen for cocaine (n = 70 [78%]). Drug screens were routinely performed on admitted patients who either had a history of cocaine use (within 5 days of the ED visit) or who were <50 years old without known coronary disease.

The evaluation and triage strategy for patients with chest pain at our institution has been reported previously (4). In brief, all patients underwent an initial evaluation performed by housestaff and attending physicians that included a history, physical examination and electrocardiogram (EKG). During the initial portion of the study, attending physicians were board certified in Internal Medicine; after 1996 to 2001 the ED was also staffed by board certified Emergency Physicians. High-risk patients (those with ischemic EKG changes [ST elevation or ST depression ≥ 1 mm or ischemic T wave inversion] or with typical symptoms and a history of coronary disease) were admitted to the coronary care unit (CCU) for exclusion of MI; all others underwent further risk stratification using rest myocardial perfusion imaging. Patients considered at moderate risk for unstable angina and low risk for MI (non-ischemic EKG, and either prolonged typical symptoms in a patient without known coronary disease [prior MI or revascularization], or atypical symptoms in a patient with known coronary disease) were injected with Technetium-99m sestamibi or tetrofosmin in the ED and were observed in the CCU for a “fast-track rule in” protocol after imaging. Patients considered at low risk for unstable angina (non-ischemic EKG, and either short-lived typical symptoms or prolonged atypical symptoms) underwent rest perfusion imaging directly from the ED. Those with negative rest perfusion imaging were discharged from the ED and scheduled to return for follow-up stress perfusion imaging within the subsequent 72 h, whereas those with positive rest imaging were admitted to the CCU. Patient data were collected prospectively, using a standardized data collection form, by a trained data extractor with retrospective collection of data from patient records to supply missing information.

All admitted patients underwent serial sampling for Creatine Kinase (CK), CK-MB by mass assay and TnI, with further diagnostic evaluation left to the discretion of the CCU attending cardiologist. Total serum CK activity (Vitros, Johnson and Johnson, Raritan, NJ) was determined using N-acetylcysteine activation without pre-treatment. Two different assay systems for CK-MB and TnI were used. From June 1996 to May 1998, the Opus magnum system (Behring Diagnostics, Boston, MA) was used for determination of CK-MB and TnI. From May 1998 to January 2001, the Bayer assay (Bayer Corp., Tarrytown, NY) was used. For CK-MB, an upper reference limit of 8.0 ng/mL was used for both assays. A diagnostic value of ≥ 1.0 ng/mL for the Opus Magnum assay and ≥ 0.3 ng/mL for the Bayer assay was used. Both diagnostic values were derived using local receiver operator characteristic curve analysis compared to CK-MB as the gold standard (5). A CK relative index (RI) was calculated using the following formula: CK-MB × 100/total CK. Diagnosis of MI was confirmed by an elevation of CK-MB ≥ 8.0 ng/mL with a relative index ≥ 4.0 in association with a characteristic rise and fall in markers. After June 1996, an associated increase in TnI was also required for diagnosis of MI.

Symptoms were considered consistent with typical angina if they were similar to prior angina or infarction symptoms or were described as pressure, tightness, burning, heaviness, squeezing, crushing, shortness of breath or indigestion. Electrocardiograms were interpreted by a cardiologist blinded to the clinical findings and patient outcomes. Maximal ST-segment elevation or depression was measured 0.08 ms from the J point. Q waves were considered indicative of prior MI if they were > 0.04 s in width and > 25% of the R wave in depth. An EKG was considered consistent with acute MI if there was ≥ 1 mm in two contiguous limb leads or ≥ 2 mm ST elevation in two contiguous precordial leads, consistent with ischemia if there was ≥ 1 mm ST depression (not associated with left ventricular hypertrophy) or T wave inversion that was symmetrically inverted ≥ 2 mm. Patients were considered to have prior MI if they reported a prior MI or had significant Q waves on the EKG. Patients were considered to have prior coronary disease if there was a history of MI or they had undergone prior revascularization.

Coronary angiography was performed using the Judkins technique with views of the coronary arteries obtained in multiple projections. Significant coronary artery disease was defined as ≥ 50% stenosis in the left main coronary artery, a major coronary artery or its branches, or bypass graft.

Results are presented as mean ± 1 standard deviation. Continuous and categorical results were compared using Student’s t test and chi-square analysis, respectively. A p value < 0.05 was considered significant.

Results 

From June 1994 to January 2001, 734 patients underwent evaluation for symptoms thought to be consistent with myocardial ischemia after cocaine use, of whom 90 underwent coronary angiography within 5 weeks of the ED evaluation. Demographic variables and cardiac risk factors are shown in Table 1. Thirty-two patients (36%) had evidence of prior coronary disease, which consisted of prior MI in 29 (32%), coronary artery bypass surgery in 3 (4%) and percutaneous coronary intervention (PCI) in 9 (10%). MI was diagnosed in 27 patients (30%), and an additional four patients had TnI elevations without meeting CK-MB criteria for MI, resulting in 31 patients (34%) who had MI or TnI elevations.

Table 1. Demographic Variables of All the Patients, and Separated by the Presenced or Absence of Significant Disease

		N (%)	Significant Disease (n = 45)	No Significant Disease (n = 45)
	Age (years)	42 ± 8	41 ± 7	42 ± 9
	Male	62 (69)	30 (67)	32 (71)
	Hypertension	49 (54)	25 (56)	24 (53)
	Diabetes	14 (54)	10 (22)	4 (8.9)
	Tobacco use	70 (78)	37 (82)	33 (73)
	Family history of CAD	35 (39)	17 (38)	18 (40)
	Elevated cholesterol	32 (36)	21 (47)	11 (24)*
	Number of Risk Factors	2.1 ± 1.1	2.4 ± 1.2	1.9 ± 1.3*
	History of CAD	32 (36)	23 (51)	9 (20)**
	Prior revascularization	11 (12)	9 (20)	2 (4.4)
	ST elevation on EKG	13 (14)	10 (22)	3 (6.7)*
	Prior MI, history or EKG	29 (32)	20 (44)	9 (20)**
	Prior MI by EKG alone	12 (26)	10 (22)	2 (4)**
	MI on admission	27 (30)	22 (82)	5 (18)**
	MI or (+) TnI	31 (34)	24 (77)	7 (23)**

* p < 0.05. ** p < 0.01; CAD = coronary artery disease; EKG = electrocardiogram; MI = myocardial infarction; TnI = troponin I.

Indications for angiography included elevated markers in 31 (34%), reversible defect on stress testing in 15 (17%), positive rest myocardial perfusion imaging in 12 (13%), ischemic EKG in 6 (7%), congestive heart failure or decreased ejection fraction in 8 (9%), known prior coronary disease in 8 (9%) and high-risk chest pain characteristics or recurrent chest pain in 10 (11%). Significant coronary disease was present in 45 patients (50%) (Table 1). Extent of coronary disease is shown in Figure 1. Single vessel disease was present in the majority of patients who had significant disease. The patients with significant coronary disease were significantly more likely to have MI or TnI elevations, an EKG consistent with acute infarction, a history of MI and an elevated cholesterol. Other risk factors were not significantly different, including age.

View Large Image Download to PowerPoint

Figure 1. Extent of coronary disease.

In the subgroup of patients without evidence of myonecrosis (n = 59), the only significant predictors of significant coronary disease were prior MI, prior revascularization, a history of coronary disease and an increased cholesterol (Table 2).

Table 2. Demographics and Risk Factors in Patients Without Myonecrosis Based on the Presence or Absence of Coronary Disease

		Significant Disease (n = 21)	No Significant Disease (n = 38)
	Age (years)	43 ± 7	43 ± 10
	Male	16 (76)	27 (71)
	Hypertension	12 (57)	21 (54)
	Diabetes	5 (24)	4 (11)
	Tobacco use	15 (71)	26 (68)
	Family history of CAD	8 (38)	15 (40)
	Elevated cholesterol	13 (62)	8 (21)**
	History of CAD	14 (67)	6 (16)**
	Prior revascularization	8 (38)	3 (5.3)**
	Prior MI	11 (52)	6 (16)**

** p < 0.01. History of CAD (coronary artery disease) indicates a history of myocardial infarction or previous revascularization. Abbreviations as in Table 1.

Patients with evidence of myonecrosis (n = 31) had a significantly higher prevalence of significant disease than those without myonecrosis (n = 59) (77% vs. 35%, respectively) (p < 0.001). Of the 27 patients who had CK-MB MI, 22 (82%) had significant disease. In contrast, significant disease was found in only 9 of the 42 patients (24%) who did not have either CK-MB MI, TnI elevation or a history of MI, and in 7 of the 39 patients (18%) who did not have either MI, TnI elevations or a history of coronary disease (both p < 0.001)

Discussion 

Our data show that underlying coronary disease is common in patients with cocaine-associated myocardial infarction. In contrast, in patients without evidence of myocardial necrosis, significant disease was less common and was infrequent in those without a history of MI or coronary disease.

Various etiologies for cocaine-associated MI have been proposed. In vivo studies have shown that acute cocaine use causes significant coronary artery spasm, especially in combination with tobacco use 6, 7. Cocaine also causes platelet aggregation and coronary thrombosis (8). In some patients without coronary disease, these properties, when combined with cocaine’s vasospastic actions, may result in acute vessel occlusion and MI. Cocaine is also associated with accelerated atherosclerosis (9). The sympathomimetic effects of cocaine, which can cause an increase in heart rate and blood pressure, combined with underlying coronary disease may cause oxygen demand to exceed supply, resulting in MI.

A number of factors suggest that underlying atherosclerosis, rather than vasospasm, is the most likely etiology of MI in these patients. We found a high incidence of significant coronary disease in the patients with evidence of myocardial necrosis. The distribution of risk factors in our study is similar to that of patients with premature coronary disease (10).

We found that most patients with evidence of myocardial necrosis had significant disease. Consistent with the young age of the patients, most had single vessel disease. This is similar to other studies reporting angiographic results in young patients with MI not related to cocaine use. For example, Cross et al. found that single vessel disease was present in 42% of 186 patients under the age of 60 who had MI, with 11% not having any significant disease (11). Similarly, in a study of 129 patients under the age of 40 presenting with MI who underwent angiography, Negus et al. found no significant disease in 16% of patients and single vessel disease in 53% (12). Similar results were reported by others 13, 14, 15.

Early descriptions of the incidence of coronary disease in patients with cocaine-associated MI were limited to case reports or case series. As summarized by Hollander and Hoffman, 34 of the 54 patients (55%) with cocaine-associated MI who underwent coronary angiography had either significant disease or thrombotic occlusion (16). In a multicenter study of 70 patients who had cocaine-associated MI, 42 of the 63 patients (67%) who underwent coronary angiography had significant disease. In addition, 18 of 20 patients who underwent provocative testing had reversible defects. In a previous report from our institution, Om et al. found that 10 of 12 patients with cocaine-associated MI had significant coronary disease (17). The results of these studies, as well as the current one, indicate that underlying coronary disease is common in patients with cocaine-associated MI. Therefore, further evaluation to determine the presence and extent of disease in these patients is warranted.

Less is known about the prevalence of coronary artery disease in patients who have MI excluded. Despite the large number of patients presenting to EDs, few patients with cocaine-associated chest pain undergo further risk stratification after MI exclusion 2, 16. However, data from stress, angiographic and long-term studies indicate that the incidence of underlying coronary disease is likely to be low. In a study from our institution, only 2% of patients had evidence of myocardial ischemia using acute myocardial perfusion imaging (18). Provocative testing, using treadmill or pharmacological methods with or without imaging, has demonstrated inducible ischemia in only a minority of patients 19, 20. In a long-term follow-up study, Hollander et al. found only 1 of 250 patients evaluated for possible myocardial infarction after cocaine use, most of whom had no additional evaluation, had a cardiac event in the following year (21).

Consistent with these results, we found that a minority of patients without MI had significant disease. The only significant predictors of significant disease in these patients were the variables associated with known coronary disease, such as prior MI and prior revascularization, and an elevated cholesterol. The increased incidence of cholesterol elevations may result from selective lipid sampling in patients in whom the angiogram was abnormal. Only 24% of patients without positive markers or a history of MI, and 18% of patients without positive markers or a history of coronary disease, had significant disease.

Conclusions 

The majority of patients who have cocaine-associated MI have significant coronary disease. In contrast, patients without myonecrosis have a low incidence of disease, and therefore coronary angiography should not be the routine first test to evaluate for ischemia in these patients.