The ability to provide an accurate prognosis in an intensive care unit is of major importance. Numerous risk scores have been developed to predict hospital mortality based on demographic, physiologic, and clinical data. These scores were universally developed in general medical or surgical intensive care units. Patients admitted to a cardiac care unit differ in many aspects from those admitted to general medical intensive care units. Few patients require mechanical ventilation and prolonged intensive care. Performance of risk scores developed for patients with acute myocardial infarction (AMI) in this subgroup is unknown. We prospectively studied 51 consecutive patients who were admitted to a cardiac care unit from September 2006 to March 2008 for AMI and received mechanical ventilation for >24 hours. Acute Physiology and Chronic Health Evaluation II (APACHE II), Thrombolysis In Myocardial Infarction, and Global Registry of Acute Coronary Events risk scores were calculated for each patient. Mortality rates were extrapolated based on these 3 risk scores. Twenty-two of 51 patients (43%) died in hospital. Age, mean arterial pressure, urea, albumin, hemoglobin, need for vasopressors, and estimated glomerular filtration rate were predictive of mortality. APACHE II and Global Registry of Acute Coronary Events scores were higher in nonsurvivors but Thrombolysis In Myocardial Infarction risk score was not predictive of mortality. APACHE II score had the highest value for area under receiver operator characteristics curve for mortality prediction. In conclusion, patients with AMI requiring mechanical ventilation have a high mortality rate. This risk is predicted by co-morbidities better than by direct cardiac parameters. Consequently, conventional AMI risk scores do not perform well in this very sick population and the APACHE II score better predicts their short-term outcome.
The present study prospectively identified prognostic factors for 30-day mortality in patients with acute myocardial infarction (AMI) who require mechanical ventilation for >24 hours. A secondary aim was to compare in this unique population the performance of the Acute Physiology and Chronic Health Evaluation II (APACHE II), Thrombolysis In Myocardial Infarction (TIMI), and Global Registry of Acute Coronary Events (GRACE) risk scores.
Methods
The study protocol was approved by the institutional review board of Soroka University Hospital (Beer Sheva, Israel), a 1,200-bed tertiary university hospital. From September 2006 to March 2008 all adult patients admitted to the 16-bed cardiac care unit were prospectively screened. Patients were enrolled if they were admitted for an AMI and required mechanical ventilation within 72 hours of admission and for ≥24 hours. MI was diagnosed based on troponin increase associated with ischemic symptoms or appropriate electrocardiographic changes.
Demographic, clinical, and laboratory data and treatment and survival data were obtained. APACHE II, TIMI and GRACE scores were prospectively calculated for each patient on admission: for APACHE II score calculations in sedated patients we assumed the Glasgow Coma Score to be missing. We applied the Knaus formula to calculate the expected mortality rate in the study cohort. GRACE score was calculated to assess in-hospital mortality. The TIMI risk score was used to predict risks of mortality, recurrent MI, and need for urgent revascularization. The outcome measurement was 30-day or in-hospital mortality (the latest of the 2).
Normally distributed continuous variables are presented as mean ± SD and compared by Student’s t test. Non-normally distributed continuous variables are presented as median and interquartile range and compared by the Mann-Whitney test. Discrete variables were compared using chi-square test or Fisher’s exact test when appropriate. For paired continuous variables comparison paired t test was applied. Correlation between continuous variables was assessed by Pearson correlation test. Three clinical scores were compared for the accuracy of mortality prediction by constructing receiver operating characteristics curves and calculating the area under the curve. In addition, a forward logistic regression model with death as an outcome measurement and 3 score results as covariates was created with a stay criterion of 0.10.
All reported p values are 2-sided and a p value <0.05 was considered statistically significant.
Results
Of 3,123 patients admitted to the cardiac care unit during the study period, 1,761 were diagnosed with MI and 51 (1.6%) were mechanically ventilated within 72 hours of admission for >24 hours. These patients constituted the study population.
Forty-three patients (84%) were diagnosed with ST-elevation MI (STEMI) and the rest had non-STEMI. For most patients (63%) the index event was a first manifestation of ischemic heart disease. In-hospital mortality for the present cohort was 43% (95% confidence interval 31 to 57%).
Table 1 lists baseline clinical and demographic characteristics in survivors and nonsurvivors. The 2 most significant differences between the 2 groups were age and renal failure.
| Variable | Survivors | Nonsurvivors | p Value |
|---|---|---|---|
| (n = 29) | (n = 22) | ||
| Age (years), mean ± SD | 70 ± 12 | 71 ± 13 | 0.006 |
| Men | 25 (86%) | 6 (27%) | 0.23 |
| ST-elevation myocardial infarction | 23 (79%) | 17 (77%) | 1.00 |
| Previous chronic ischemic heart disease ⁎ | 9 (31%) | 10 (46%) | 0.29 |
| Previous coronary artery bypass grafting | 2 (7%) | 3 (14%) | 0.64 |
| Previous percutaneous coronary intervention | 5 (17%) | 5 (23%) | 0.63 |
| Renal failure | 4 (14%) | 6 (27%) | 0.23 |
| Atherosclerotic risk factors | |||
| Diabetes mellitus | 11 (38%) | 11 (50%) | 0.39 |
| Dyslipidemia † | 18 (62%) | 9 (41%) | 0.13 |
| Current smoker | 13 (45%) | 9 (41%) | 0.89 |
| Hypertension | 11 (38%) | 11 (50%) | 0.13 |
| Dementia | 0 | 4 (18%) | 0.03 |
⁎ Previous myocardial infarction, typical angina, or evidence of coronary disease on angiogram or at stress testing.
† Low-density lipoprotein >100 mg/dl, high-density lipoprotein <35 mg/dl, and/or triglycerides >200 mg/dl.
Table 2 lists physiologic and laboratory parameters captured on admission and at 24 hours. Mean arterial blood pressure was significantly higher in survivors over the first 24 hours. Degree of myocardial damage as reflected by levels of creatine kinase on admission and at 24 hours was similar between groups. Estimated creatinine clearance was lower in patients who eventually died both on admission and at 24 hours. Although creatinine clearance increased among survivors by 16 ± 37 ml/min over the first 24 hours (p = 0.04), a significant decrease of 12 ± 12 ml/min (p <0.001) was observed in nonsurvivors. Albumin and hemoglobin levels on admission and at 24 hours were significantly lower in nonsurvivors.
| Variable | Survivors | Nonsurvivors | p Value |
|---|---|---|---|
| (n = 29) | (n = 22) | ||
| Mean arterial blood pressure (mm Hg) | |||
| On admission | 83 ± 17 | 71 ± 18 | 0.03 |
| At 24 hours | 67 ± 16 | 58 ± 15 | 0.06 |
| Heart rate (beats/min) | |||
| At admission | 91 ± 21 | 88 ± 29 | 0.70 |
| At 24 hours | 85 ± 14 | 77 ± 15 | 0.07 |
| Oxygen saturation (%) | |||
| At admission | 89 ± 19 | 89 ± 20 | 1.00 |
| At 24 hours | 99 ± 1 | 95 ± 9 | 0.02 |
| Moderate to severe left ventricular dysfunction | 12 (41%) | 13 (59%) | 0.21 |
| Creatinine (mg/dl) | |||
| At admission | 1 ± 0.41 | 2 ± 1 | 0.003 |
| At 24 hours | 1 ± 1 | 3 ± 2 | <0.001 |
| Estimated glomerular filtration rate (ml/min) | |||
| At admission | 56 ± 22 | 34 ± 18 | 0.001 |
| At 24 hours | 72 ± 42 | 23 ± 19 | <0.001 |
| Albumin (g/dl) | |||
| At admission | 4 ± 0.60 | 3 ± 0.42 | 0.04 |
| At 24 hours | 3 ± 0.57 | 2 ± 0.42 | <0.001 |
| Creatine kinase (U/L) | |||
| At admission | 1,439 ± 1,681 | 1,902 ± 2,497 | 0.44 |
| At 24 hours | 3,460 ± 3,777 | 2,779 ± 2,595 | 0.49 |
| Hemoglobin (g/dl) | |||
| At admission | 14 ± 2 | 13 ± 2 | 0.01 |
| At 24 hours | 11 ± 2 | 10 ± 2 | 0.29 |
| White blood cell count (10 3 /mm 3 ) | |||
| At admission | 16 ± 6 | 15 ± 6 | 0.30 |
| At 24 hours | 15 ± 4 | 17 ± 7 | 0.33 |
| Urea (mg/dl) | |||
| At admission | 45 ± 23 | 55 ± 37 | 0.02 |
| At 24 hours | 66 ± 37 | 129 ± 66 | <0.001 |
| pH | |||
| At admission | 7.32 ± 0.14 | 7.29 ± 0.14 | 0.44 |
| At 24 hours | 7.41 ± 0.10 | 7.32 ± 0.17 | 0.15 |
Table 3 presents in-hospital management. Most patients (69%) were ventilated on admission and underwent cardiopulmonary resuscitation (55%). Median lengths of ventilation were 4 days (interquartile range 2 to 7) in survivors and 4 days (interquartile range 3 to 7) in nonsurvivors. Overall, 34 of 51 patients (67%) underwent angioplasty within 24 hours of admission: 28 patients (70%) in the STEMI group and 6 patients (55%) in the non-STEMI group. Twenty-one patients (41%) required an intra-aortic balloon pump. Use of an intra-aortic balloon pump was not associated with survival: mortality in patients who did or did not require an intra-aortic balloon pump was 43%. Thirty patients (59%) required vasopressors, of whom 16 (53%) did not survive 30 days compared to 29% in patients who did not require vasopressors (p = 0.07).
| Variable | Survivors | Nonsurvivors | p Value |
|---|---|---|---|
| (n = 29) | (n = 22) | ||
| Ventilated on admission | 19 (66%) | 16 (73%) | 0.58 |
| Cardiopulmonary resuscitation before mechanical ventilation | 18 (62%) | 10 (46%) | 0.24 |
| Electrical cardioversion | 10 (35%) | 6 (27%) | 0.58 |
| Primary percutaneous coronary intervention | 20 (69%) | 14 (64%) | 0.69 |
| Multivessel coronary disease | 16 (55%) | 14 (64%) | 0.48 |
| Intra-aortic balloon counterpulsation | 12 (41%) | 9 (41%) | 0.97 |
| Vasopressor use | 14 (48%) | 16 (73%) | 0.08 |
| Length of stay in cardiac care unit (days), median (interquartile range) | 5 (7–9) | 4 (3–8) | 0.11 |
| Length of stay in hospital (days), median (interquartile range) | 17 (10–32) | 12 (4–23) | 0.07 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
