Prompt percutaneous coronary intervention is associated with improved survival in patients presenting with cardiac arrest. Few studies, however, have focused on patients with cardiac arrest not selected for coronary angiography. The aim of the present study was to evaluate the clinical characteristics and outcomes of patients with cardiac arrest denied emergent angiography. Patients with cardiac arrest were identified within a registry that included all catheterization laboratory activations from 2008 to 2012. Logistic regression and proportional-hazards models were created to assess the clinical characteristics and mortality associated with denying emergent angiography. Among 664 patients referred for catheterization, 110 (17%) had cardiac arrest, and 26 of these patients did not undergo emergent angiography. Most subjects (69%) were turned down for angiography for clinical reasons and a minority for perceived futility (27%). After multivariate adjustment, pulseless electrical activity as the initial arrest rhythm (adjusted odds ratio [AOR] 13.27, 95% confidence interval [CI] 1.76 to 100.12), <1.0 mm of ST-segment elevation (AOR 10.26, 95% CI 1.68 to 62.73), female gender (AOR 4.45, 95% CI 1.04 to 19.08), and advancing age (AOR 1.10 per year, 95% CI 1.04 to 1.16) were associated with increased odds of withholding angiography. The mortality rate was markedly higher for patients who were denied emergent angiography (hazard ratio 3.64, 95% CI 2.05 to 6.49), even after adjustment for medical acuity (hazard ratio 2.29, 95% CI 1.19 to 4.41). In conclusion, older subjects, women, and patients without ST-segment elevation were more commonly denied emergent angiography after cardiac arrest. Patients denied emergent angiography had increased mortality that persisted after adjustment for illness severity.
Current guidelines indicate that emergent coronary angiography should be performed in survivors of cardiac arrest who have persistent electrocardiographic evidence of ischemia. These same guidelines suggest that subjects resuscitated from cardiac arrest without evidence of ongoing ischemia may be considered for emergent angiography. Previous research has demonstrated that standard electrocardiography has a poor negative predictive value for the presence of an angiographic coronary culprit lesion. Because of this, some have suggested that all resuscitated patients presenting with cardiac arrest should undergo emergent cardiac catheterization. With this in mind, emergency room physicians tend to activate cardiac catheterization teams soon after patients with cardiac arrest present. Public reporting of cardiac catheterization outcomes, however, may lead some interventional cardiologists to decline emergent angiography in cases that are perceived as futile, including patients with cardiac arrest. A thorough characterization of patients with cardiac arrest who are denied emergent angiography in the current era would thus be informative. To address this, we sought to identify the clinical characteristics and outcomes of patients with cardiac arrest denied emergent angiography.
Methods
All patients presenting to an urban trauma center (San Francisco General Hospital) or a tertiary care center (University of California, San Francisco) referred for emergent angiography for a potential ST-segment elevation myocardial infarction, including patients after cardiac arrest, were enrolled in the ACTIVATE-SF registry. As described previously, this registry includes all emergency physician–initiated cardiac catheterization laboratory activations from October 2008 through July 2012. The present analysis was focused on patients within this registry who presented with cardiac arrest and were successfully resuscitated. For the purposes of this study, cardiac arrest was defined as the absence of spontaneous circulation as documented by emergency medical services or emergency department physicians. The present project was reviewed and approved with a waiver of the requirement for consent by the institutional review board at the University of California, San Francisco.
Clinical information was collected from ambulance and emergency department records. Electrocardiograms from the initial presentation were deidentified and independently evaluated by 2 cardiologists blinded to clinical outcomes. Clinically significant ST-segment elevation was defined as >1.0 mm of ST-segment elevation in 2 contiguous electrocardiographic leads. Emergent coronary angiography was defined as a cardiac catheterization that was requested by the emergency department and deemed appropriate to perform by the treating interventional cardiologist. In some cases, the treating interventional cardiologist deemed the procedure inappropriate or futile given the clinical status of the patient. Laboratory values and echocardiographic data obtained later in the hospital course were retrieved from the electronic medical record. All study data were collected and managed using the Research Electronic Data Capture reporting system hosted at the University of California, San Francisco.
Mortality after the index hospitalization was ascertained using the Social Security Death Index. To emphasize discrimination, surviving subjects not identified in the death index were censored at the time of their last health care encounter. For the purposes of this analysis, the time to death or censoring was recorded in days from the date of the presenting arrest.
Summary statistics are reported as mean ± SD for continuous variables and as medians and interquartile ranges for continuous data not normally distributed. Normally distributed continuous variables were compared using Student’s t tests, and variables not normally distributed were compared using Kruskal-Wallis analysis of variance by ranks. All proportions were evaluated using chi-square or Fisher’s exact tests. To identify clinical characteristics associated with denial of emergent coronary angiography, directed acyclic graphs were created to identify covariates associated with the decision to withhold emergent angiography. Candidate covariates included age, gender, clinical symptoms (chest pain, dyspnea), therapeutic interventions (vasopressors, intubation, hypothermia), electrocardiographic findings (magnitude of ST-segment elevation), and cardiac arrest characteristics (arrest rhythm, arrest location). Candidate covariates were retained in the backward stepwise logistic regression model for the decision to withhold emergent coronary angiography using a conservative p value of <0.20. The final covariates that remained in the model were age, gender, electrocardiographic findings, arrest rhythm, and arrest location. Kaplan-Meier survival plots were stratified according to the performance of emergent angiography. Log-rank tests were used to assess the differences between the survival functions. To adjust for illness severity, a Cox proportional-hazards model was created using age, gender, intubation, hypothermia, and a lack of emergent angiography as covariates to determine the association between denial of emergent coronary angiography and mortality. This model was repeated in a sensitivity analysis excluding the cases that were deemed clinically futile. All statistical analyses were performed using Stata version 12 (StataCorp LP, College Station, Texas). A p value <0.05 was considered statistically significant.
Results
Six hundred sixty-four patients presented with clinical characteristics leading to emergent activation of the cardiac catheterization laboratory from October 2008 to July 2012, of whom 110 (17%) were successfully resuscitated from cardiac arrest. Twenty-six of the 110 patients (24%) with cardiac arrest referred for emergent coronary angiography by the emergency department were turned down by the catheterization team. Similarly, 100 of the 554 patients (18%) without cardiac arrest were also denied angiography (p = 0.16). Among the 84 subjects who went on to angiography after cardiac arrest, culprit lesions necessitating percutaneous coronary intervention were present in 60 (72%).
As listed in Table 1 , patients with cardiac arrest who did not undergo emergent coronary angiography were older (p <0.01) and were more frequently women (p <0.03). Pulseless electrical activity was the predominant initial arrest rhythm in patients who were turned down for angiography, compared with ventricular fibrillation for those who underwent the procedure (p <0.03). Symptoms traditionally associated with acute myocardial ischemia, such as chest pain and dyspnea, were significantly less common at presentation in patients denied emergent angiography (p <0.01). Finally, those who did not undergo angiography had lower recorded heart rates (p <0.01), respiratory rates (p <0.01), and systolic blood pressure (p <0.02) on presentation, leading to numerically greater use of mechanical ventilation (p = 0.06) and vasopressors (p = 0.24). As listed in Table 2 , the degree of maximum ST-segment elevation was similar in those who underwent emergent angiography and those that did not (p = 0.58). Electrocardiographic evidence of left ventricular hypertrophy, however, was more common in subjects who did not proceed to cardiac catheterization (p <0.01).
| Variable | Emergent Angiography | p Value | |
|---|---|---|---|
| Denied (n = 26) | Performed (n = 84) | ||
| Age (yrs) | 80 ± 9 | 63 ± 16 | 0.01 |
| Men | 13 (50%) | 62 (74%) | 0.03 |
| Race | 0.61 | ||
| Asian | 7 (27%) | 23 (28%) | |
| Black | 2 (8%) | 6 (7%) | |
| Hispanic | 4 (15%) | 9 (11%) | |
| Non-Hispanic White | 12 (46%) | 44 (53%) | |
| Cardiac arrest type | 0.03 | ||
| Pulseless electrical activity | 11 (42%) | 13 (15%) | |
| Ventricular tachycardia | 1 (4%) | 2 (2%) | |
| Ventricular fibrillation | 9 (35%) | 51 (61%) | |
| Other | 5 (19%) | 18 (21%) | |
| Cardiac arrest location | 0.05 | ||
| Emergency department | 2 (8%) | 22 (26%) | |
| Out of hospital | 24 (92%) | 62 (74%) | |
| Coronary artery disease | 10 (38%) | 18 (21%) | 0.08 |
| Previous percutaneous coronary intervention | 5 (22%) | 8 (11%) | 0.16 |
| Previous coronary bypass | 3 (12%) | 6 (7%) | 0.57 |
| Diabetes mellitus | 5 (21%) | 13 (17%) | 0.68 |
| Hypertension | 12 (50%) | 32 (43%) | 0.53 |
| Hyperlipidemia | 5 (22%) | 11 (15%) | 0.45 |
| Heart failure | 5 (21%) | 3 (4%) | 0.02 |
| Substance Abuse | |||
| Drugs | 0 | 9 (13%) | 0.10 |
| Tobacco | 0 | 8 (13%) | 0.10 |
| Vital signs | |||
| Heart rate (beats/min) | 60 (0–100) | 94 (54–110) | 0.01 |
| Respiratory rate (breaths/min) | 12 (0–17) | 19 (16–22) | 0.00 |
| Systolic blood pressure (mm Hg) | 77 (0–127) | 108 (60–152) | 0.02 |
| Creatinine (mg/dl) | 1.7 ± 0.7 | 1.5 ± 1.2 | 0.17 |
| Therapeutic interventions | |||
| Intubation | 24 (92%) | 61 (73%) | 0.06 |
| Hypothermia | 6 (23%) | 26 (31%) | 0.47 |
| Vasopressors | 12 (46%) | 27 (32%) | 0.24 |
| Variable | Emergent Angiography | p Value | |
|---|---|---|---|
| Denied (n = 26) | Performed (n = 84) | ||
| Recovered rhythm | 0.14 | ||
| Sinus rhythm | 14 (58%) | 52 (63%) | |
| Atrial fibrillation | 1 (4%) | 13 (16%) | |
| Atrial flutter | 0 (0%) | 1 (1%) | |
| Junctional rhythm | 1 (4%) | 5 (6%) | |
| Third-degree heart block | 2 (8%) | 4 (5%) | |
| ST-segment elevation magnitude (mm) | 0.58 | ||
| <1 | 10 (40%) | 24 (29%) | |
| 1–1.9 | 4 (16%) | 10 (12%) | |
| 2–2.9 | 3 (12%) | 10 (12%) | |
| >3 | 8 (32%) | 39 (47%) | |
| Bundle branch block | 0.42 | ||
| None | 16 (67%) | 59 (71%) | |
| Left | 1 (4%) | 9 (11%) | |
| Right | 7 (29%) | 15 (18%) | |
| Left ventricular hypertrophy ∗ | 8 (33%) | 6 (7%) | 0.01 |
The treating interventional cardiologist deemed emergent coronary angiography clinically unnecessary in 18 of 26 patients (69%). In an additional 7 patients (27%), emergent angiography was perceived as likely futile during continued resuscitation efforts in the emergency department. One patient (4%) was turned down for angiography for reasons that were not clearly documented. None of the subjects had documented contraindications to cardiac catheterization, nor did any patients or their surrogates refuse the procedure when it was offered. A multivariate model was constructed to identify clinical characteristics associated with denial of emergent coronary angiography. As listed in Table 3 , increasing age (adjusted odds ratio [AOR] 1.10, 95% confidence interval [CI] 1.04 to 1.16), female gender (AOR 4.45, 95% CI 1.04 to 19.08), pulseless electrical activity as the presenting rhythm (AOR 13.27, 95% CI 1.76 to 100.12), and <1.0 mm of ST-segment elevation on postresuscitation electrocardiography (AOR 10.26, 95% CI 1.68 to 62.73) were associated with increased odds of denying cardiac catheterization.
| Associated Factor | OR | 95% CI | p Value | AOR | 95% CI | p Value |
|---|---|---|---|---|---|---|
| Age (per yr) ∗ | 1.09 | 1.05–1.13 | 0.001 | 1.10 | 1.04–1.16 | 0.001 |
| Female gender | 2.81 | 1.13–7.00 | 0.026 | 4.45 | 1.04–19.08 | 0.044 |
| Field cardiac arrest | 4.26 | 0.93–19.51 | 0.062 | 4.80 | 0.64–35.94 | 0.127 |
| Cardiac arrest rhythm | ||||||
| Unknown initial rhythm | 1 (reference) | 1 (reference) | ||||
| Pulseless electrical activity | 3.05 | 0.85–10.90 | 0.087 | 13.27 | 1.76–100.12 | 0.012 |
| Ventricular tachycardia | 1.80 | 0.13–24.16 | 0.657 | 7.79 | 0.20–301.23 | 0.271 |
| Ventricular fibrillation | 0.64 | 0.19–2.15 | 0.465 | 1.96 | 0.35–10.86 | 0.440 |
| Postresuscitation ST-segment elevation (mm) | ||||||
| >3 | 1 (reference) | 1 (reference) | ||||
| 2–2.9 | 1.46 | 0.33–6.54 | 0.619 | 1.29 | 0.13–13.13 | 0.830 |
| 1–1.9 | 1.95 | 0.49–7.80 | 0.345 | 8.47 | 0.98–73.35 | 0.053 |
| 0–0.9 | 2.03 | 0.70–5.86 | 0.190 | 10.26 | 1.68–62.73 | 0.012 |
∗ Unadjusted or adjusted odds of denying emergent angiography for each increasing year of age relative to the mean age for the entire cohort.
The length of stay and mortality for patients with cardiac arrest are listed in Table 4 . The median follow-up period was 8 days (interquartile range 2 to 361), reflecting the high in-hospital mortality in this population. Among subjects who survived to hospital discharge, the median follow-up period was 184 days (interquartile range 7 to 786). Thirty-five percent of subjects denied emergent coronary angiography died on the first hospital day, compared with 2% of those who underwent the procedure (p <0.01). Similarly, the in-hospital (p <0.01) and 30-day mortality (p <0.01) among subjects denied angiography were increased compared with those who underwent the procedure. As shown in Figure 1 , patients who did not undergo angiography had an increased unadjusted incident mortality rate compared with those selected for angiography (hazard ratio [HR] 3.65, 95% CI 2.05 to 6.49), a finding that remained significant after multivariate adjustment for the severity of illness (HR 2.29, 95% CI 1.19 to 4.41). Among patients selected for angiography, percutaneous coronary intervention did not significantly reduce the HR for mortality compared with patients who did not undergo intervention (HR 0.79, 95% CI 0.41 to 1.55). As shown in Figure 2 , we performed a sensitivity analysis that excluded all subjects who were denied emergent catheterization because of perceived futility. In this smaller cohort (n = 103), the unadjusted HR (2.69, 95% CI 1.39 to 5.18) confirmed the findings, and the adjusted hazard ratio (1.51, 95% CI 0.73 to 3.11) showed the appropriate directional trend toward a higher incident mortality rate for patients denied emergent angiography compared with those who underwent the procedure.
