Anemia has been associated with adverse outcomes in patients with acute coronary syndromes (ACS). However, the underlying pathophysiologic mechanisms have not been well elucidated. We sought to determine the independent relation between the hemoglobin level and recurrent ischemia in patients with non–ST-segment elevation ACS using continuous electrocardiographic monitoring. In the Integrilin and Enoxaparin Randomized Assessment of Acute Coronary Syndrome Treatment (INTERACT) trial, 746 patients presenting with non–ST-segment elevation ACS underwent continuous ST-segment monitoring for 48 hours. The data were analyzed independently at a core laboratory. We stratified the study population according to their hemoglobin level on presentation. The primary outcome of the study was recurrent ischemia, defined as ST-segment shifts on continuous electrocardiographic monitoring. Of the 705 patients with analyzable data, 64 had a baseline hemoglobin level <120 g/L, 259 had a level of 120 to 139 g/L, 315 had a level of 140 to 159 g/L, and 67 had a level >160 g/L. The corresponding rates of recurrent ischemia were 39.1%, 22.0%, 15.6%, and 11.9% (p for trend <0.001). A lower hemoglobin level was associated with advanced age, co-morbidities, and a higher GRACE risk score. In multivariable analysis adjusting for these confounders, lower hemoglobin levels retained a significant independent association with recurrent ischemia (p for trend = 0.004). In conclusion, a lower hemoglobin level at presentation was independently associated with recurrent ischemia detected by continuous electrocardiographic monitoring in the setting of non–ST-segment elevation ACS. This suggests that anemia might predispose patients to recurrent ischemia, which could be an important underlying mediator of worse outcomes in patients with lower hemoglobin levels.
Although recurrent ischemia after acute myocardial infarction (MI) has been associated with a poor prognosis, the clinical symptoms—such as angina pectoris or angina equivalents—often underestimate the entire ischemic burden. It is now well-recognized that most episodes of myocardial ischemia are asymptomatic. Thus, patient-reported symptoms and standard electrocardiography alone might fail to identify much of the ischemic burden. Continuous electrocardiographic (ECG) monitoring is a validated tool for the detection of recurrent ischemia in the setting of acute coronary syndromes (ACS), thereby providing valuable pathophysiologic insights into the potential adverse effects of anemia in ACS. Using data from the Integrilin and Enoxaparin Randomized Assessment of Acute Coronary Syndrome Treatment (INTERACT) trial, we sought to determine whether an independent relation exists between the hemoglobin level and recurrent ischemia detected by continuous ECG monitoring in patients with non–ST-segment elevation ACS.
Methods
The study population included patients enrolled in the INTERACT trial. The study design has been previously published. The INTERACT trial was a prospective, randomized, controlled study that enrolled 746 patients, involving 50 centers across Canada, from September 2000 to December 2001. Its main objective was to compare the efficacy and safety of enoxaparin, relative to unfractionated heparin, in patients with high-risk non–ST-segment elevation ACS receiving aspirin and the glycoprotein IIb/IIIa inhibitor eptifibatide.
Patients were eligible to enter the study if they had presented with ischemic chest pain at rest within 24 hours before enrollment, with evidence of ST-segment deviation on the electrocardiogram (either ST-segment depression of >0.1 mV or transient ST-segment elevation of >0.1 mV in 2 contiguous leads), and/or elevated serum biomarkers (troponin I or T >3 times the upper reference level or creatinine kinase-MB greater than normal). Patients with noninterpretable ST-segment findings on the baseline electrocardiogram (e.g., left bundle branch block) were excluded, as were patients with significant anemia at presentation (hemoglobin level <110 g/L in men and <90 g/L in women). Patients were closely monitored with daily blood counts as a part of the study protocol. All patients provided written informed consent. The study was approved by the institutional review board of each participating hospital.
In the INTERACT study, the patients were randomized to receive enoxaparin (1 mg/kg subcutaneously twice daily) or unfractionated heparin (70 U/kg bolus; 15 U/kg/hour infusion, titrated to an activated partial thromboplastin time of 1.5 to 2 times control) for 48 hours. All patients also received aspirin and eptifibatide (180 μg/kg intravenous bolus followed by a 2-μg/kg/min infusion). The use of other medications, coronary angiography, and revascularization was left to the discretion of the treating physicians.
The patients were monitored with 7-lead, 3-channel, continuous, ST-segment monitoring (Applied Cardiac Systems, Laguna Hills, California) for the first 48 hours after randomization. The ST-segment analysis was initially performed by an automated algorithm and then reviewed twice by a cardiologist blinded to clinical data, treatment assignment, or outcome. We defined a significant ST-segment shift (measured at 80 ms after the J point) as horizontal or downsloping ST-segment depression of ≥0.1 mV below baseline or ≥0.1 mV ST elevation above baseline, lasting for >1 minute, and separate from other episodes of ST deviation by ≥1 minute. For patients with ST-segment depression on the baseline electrocardiogram, additional ST-segment shifts were required to qualify as an ischemic episode. All continuous ST-segment monitoring data were analyzed centrally at the Canadian Heart Research Centre. These results were not available to the treating physicians and therefore did not influence the management decisions. The present study focused on early continuous ST-segment monitoring performed in the first 48 hours. Of the 746 patients enrolled in the INTERACT study, 41 (5.5%) did not have analyzable ST-segment data because of technical problems or poor quality or incomplete data. Thus, the data from 705 patients were analyzed. The primary outcome was recurrent ischemia, defined as a significant ST-segment shift, during the first 48 hours.
Continuous variables are presented as the median with the interquartile range and categorical data as percentages. We stratified the study population a priori according to their presenting hemoglobin level: <120, 120 to 139, 140 to 159, and >160 g/L. Trends across the hemoglobin groups were examined using the Kendall tau-b test and chi-square test for trend. We also examined the nadir (lowest level) hemoglobin level in the first 48 hours in relation to the ST-segment shifts.
To assess the independent relation between the hemoglobin level at presentation and recurrent ischemia, we constructed a multivariable logistic regression model to adjust for the Global Registry of Acute Coronary Events (GRACE) risk score, gender, diabetes, hypertension, history of myocardial infarction and revascularization, and treatment assignment (enoxaparin vs unfractionated heparin). These co-morbid conditions have been previously recognized to be independently associated with recurrent ischemia. The GRACE risk score is a validated predictor of in-hospital mortality, and has 8 components: age, heart rate, systolic blood pressure, cardiac arrest, Killip class, ST-segment deviation, positive biomarker findings, and serum creatinine level. All variables were collected at presentation. Model discrimination and calibration were assessed using the c-statistic and Hosmer-Lemeshow goodness-of-fit test, respectively. We used the Statistical Package for Social Sciences, version 15.0 (SPSS, Chicago, Illinois), for data analysis and defined statistical significance as 2-sided p <0.05.
Results
Table 1 summarizes the baseline demographic and clinical characteristics of the 746 study patients stratified according to their hemoglobin level at presentation. The study population consisted of 513 men (68.8%) and 233 women (31.2%). Overall, patients with lower hemoglobin levels were more likely to be older, women, and non-smokers, to present with a worse Killip class, and to have multiple co-morbidities, including a history of diabetes, hypertension, myocardial infarction, and revascularization.
| Characteristic | Baseline Hemoglobin (g/L) | ||||
|---|---|---|---|---|---|
| <120 (n = 67) | 120–139 (n = 278) | 140–159 (n = 331) | ≥160 (n = 70) | p Value for Trend | |
| Age (years ⁎ ) | 72 (63, 79) | 67 (56, 75) | 61 (53, 71) | 57 (50, 65) | <0.001 |
| Women | 56.7% | 45.7% | 19.3% | 5.7% | <0.001 |
| Previous myocardial infarction | 41.8% | 30.6% | 26.0% | 20.0% | 0.002 |
| Prior percutaneous coronary intervention | 19.4% | 10.8% | 8.5% | 8.6% | 0.02 |
| Prior coronary artery bypass graft surgery | 10.4% | 11.5% | 9.4% | 0.0% | 0.03 |
| Prior stroke | 13.4% | 7.6% | 6.6% | 1.4% | 0.01 |
| Diabetes mellitus | 37.3% | 23.7% | 20.2% | 15.7% | 0.002 |
| History of dyslipidemia | 46.3% | 46.8% | 42.3% | 47.1% | 0.58 |
| Hypertension | 56.7% | 53.2% | 44.4% | 44.3% | 0.02 |
| Current smoker | 20.9% | 18.3% | 36.0% | 45.7% | <0.001 |
| Weight (kg ⁎ ) | 72 (60, 84) | 75 (64, 85) | 82 (73, 95) | 86 (74, 95) | <0.001 |
| Clinical presentation | |||||
| Heart rate (beats/min ⁎ ) | 76 (61, 84) | 72 (62, 84) | 70 (62, 80) | 69 (64, 78) | 0.16 |
| Systolic blood pressure (mm Hg ⁎ ) | 133 (116, 150) | 133 (122, 150) | 134 (120, 150) | 130 (114, 152) | 0.49 |
| Diastolic blood pressure (mm Hg ⁎ ) | 70 (63, 80) | 76 (68, 86) | 79 (68, 88) | 78 (67, 90) | 0.001 |
| Killip class I | 69.7% | 87.4% | 92.4% | 95.7% | <0.001 |
| Killip class II | 25.8% | 11.9% | 7.3% | 2.9% | |
| Killip class III or IV | 4.5% | 0.7% | 0.3% | 1.4% | |
| ST-segment deviation | 34.8% | 20.7% | 20.1% | 29.0% | 0.43 |
| Abnormal cardiac biomarker | 74.2% | 82.3% | 85.5% | 95.7% | 0.001 |
| Creatinine (umol/L ⁎ ) | 87 (73, 110) | 85 (72, 103) | 87 (76, 101) | 91 (78, 101) | 0.26 |
| GRACE risk score ⁎ | 135 (111, 162) | 118 (99, 141) | 112 (94, 131) | 108 (96, 130) | <0.001 |
The medications administered within the first 48 hours, in-hospital cardiac procedures, and 30-day outcomes are listed in Table 2 . No significant difference was found among the groups, except that patients with lower hemoglobin levels had a greater rate of calcium channel blocker use and a lower rate of percutaneous coronary intervention during their hospital stay. Most cardiac catheterizations and other cardiac interventions were performed >48 hours after randomization (i.e., after completion of continuous ECG monitoring). Patients with lower hemoglobin levels had worse outcomes at 30 days, including myocardial reinfarction and death.
| Variable | Hemoglobin (g/L) | ||||
|---|---|---|---|---|---|
| <120 (n = 67) | 120–139 (n = 278) | 140–159 (n = 331) | ≥160 (n = 70) | p Value for Trend | |
| Cardiac medications within first 48 h | |||||
| Aspirin | 95.5% | 96.8% | 95.8% | 91.4% | 0.19 |
| Clopidogrel | 14.9% | 17.6% | 15.4% | 11.4% | 0.38 |
| Angiotensin-converting enzymes inhibitors or angiotensin receptor blockers | 64.2% | 56.5% | 55.0% | 52.9% | 0.19 |
| β Blocker | 85.1% | 79.9% | 85.8% | 84.3% | 0.31 |
| Calcium channel blocker | 37.3% | 24.5% | 15.4% | 11.4% | <0.001 |
| Lipid-lowering agent | 44.8% | 52.5% | 50.8% | 50.0% | 0.80 |
| Intravenous nitrate | 35.8% | 34.9% | 30.5% | 40.0% | 0.81 |
| Oral nitrate | 62.7% | 54.7% | 49.8% | 52.9% | 0.10 |
| Unfractionated heparin | 59.7% | 51.4% | 42.9% | 58.6% | 0.19 |
| Low-molecular weight heparin | 40.3% | 48.6% | 57.1% | 41.4% | 0.19 |
| In-hospital cardiac procedures | |||||
| Cardiac catheterization | 46.3% | 66.5% | 65.0% | 60.0% | 0.25 |
| Time interval to cardiac catheterization (days) | 5 (3, 7) | 4 (3, 7) | 4 (3, 7) | 5.5 (3, 8) | 0.95 |
| Percutaneous coronary intervention | 10.4% | 26.6% | 32.6% | 35.7% | <0.001 |
| Time interval to percutaneous coronary intervention (days) | 4 (3, 12) | 5 (3, 10) | 5 (3, 9) | 4 (3, 10) | 0.60 |
| Coronary artery bypass graft surgery | 11.9% | 11.5% | 12.4% | 17.1% | 0.35 |
| 30-Day outcomes | |||||
| Death | 6.0% | 5.4% | 1.5% | 0.0% | 0.002 |
| Death/myocardial (re-)infarction | 7.5% | 10.8% | 4.2% | 4.3% | 0.014 |
Of the 746 patients enrolled, 705 had complete and analyzable continuous ECG data, and these patients formed the study cohort. Figures 1 and 2 show the rates of ST-segment shift detected through continuous ECG monitoring at 48 hours. The results are shown according to the hemoglobin level at presentation ( Figure 1 ) and according to the nadir level (lowest level) within the first 48 hours ( Figure 2 ). Statistically significant inverse relations were found between rates of ST-segment shift and the hemoglobin levels (both p for trend <0.001).
