Abstract
Background
Expert guidelines for treatment of cardiac arrest recommend administration of epinephrine every three to five minutes. However, different dosing periods of epinephrine have not been systematically assessed.
Objective
We evaluated the association between epinephrine dosing frequency and survival to hospital discharge in adults with an in-hospital cardiac arrest (IHCA).
Methods
Using data from 2000–2009 in the Get With the Guidelines(GWTG)-Resuscitation IHCA registry (formerly the National Registry of Cardiopulmonary Resuscitation [NRCPR]), we examined the association between epinephrine dosing period and survival to hospital discharge. Epinephrine dosing period was defined as the time between the first epinephrine dose and the resuscitation endpoint, divided by the total number of epinephrine doses received subsequent to the first epinephrine dose. Generalized estimating equations were used to construct multivariable logistic regression models, adjusted for patient and arrest characteristics.
Results
Included were 20,909 eligible IHCA events from 505 GWTG-Resuscitation participating hospitals. Compared to an epinephrine dosing period of 4 to <5 minutes per dose, survival to hospital discharge was significantly higher in patients with an epinephrine dosing period of 6 to <10 minutes per dose: for 6 to <7 min/dose, adjusted odds ratio [OR], 1.41 (95% CI: 1.12, 1.78); for 7 to <8 min/dose, adjusted OR, 1.30 (95%CI: 1.02, 1.65); for 8 to <9 min/dose, adjusted OR, 1.79 (95%CI: 1.38, 2.32); for 9 to <10 min/dose, adjusted OR, 2.17 (95%CI: 1.62, 2.92). This pattern was consistent for both shockable and non-shockable cardiac arrest rhythms. Moreover, for the majority (87%) of cardiac arrests due to non-shockable rhythms, an epinephrine dosing period of 1 to <3 minutes/dose was associated with lower rates of survival.
Conclusion
In this large, observational, national registry of in-hospital cardiac arrest, we found that epinephrine dosing at a less frequent dosing period than recommended by consensus guidelines was associated with improved survival of in-hospital cardiac arrest. Our findings suggest that clinical trials may be needed to determine the role and dose frequency of epinephrine in the treatment of in-hospital cardiac arrest.
INTRODUCTION
More than 200,000 in-hospital cardiac arrests (IHCA) occur annually in the U.S. with a survival rate of less that 20%.(1–4) Despite being a common and high risk problem, several recommended therapies in resuscitation guidelines lack a supportive evidence base. For example, the recommendation to use epinephrine as a vasopressor agent in cardiopulmonary resuscitation (CPR), for all presenting cardiac rhythms, has remained largely unchanged in resuscitation guidelines despite unproven survival benefit.(5, 6)
The rationale for recommending epinephrine use in CPR is based on its ability to augment blood pressure and increase coronary perfusion through systemic vasoconstriction. However, epinephrine also stimulates cardiac adrenoreceptors and therefore may have detrimental effects on the heart during ischemia and upon reperfusion after return of spontaneous circulation (ROSC).(5, 7) The appropriate epinephrine dosing regimen to balance these opposing effects remains unclear. Administration of high dose epinephrine (5–10X standard dose) has not been shown to improve survival compared with the standard one milligram dose.(8, 9) However, little is known about the effect of the frequency of epinephrine dosing and survival after cardiac arrest. This gap in knowledge is important because current guidelines recommend the administration of epinephrine every three to five minutes, based primarily on expert opinion. Accordingly, we sought to examine the association between epinephrine dosing frequency and survival to hospital discharge in adults with an IHCA.
METHODS
Study Design
We used data submitted to the Get With the Guidelines (GWTG)-Resuscitation registry (formerly the National Registry of Cardiopulmonary Resuscitation [NRCPR]), a prospective, multicenter registry of patients with IHCA. The study design and data collection of GWTG-Resuscitation have been previously described.(1) Briefly, an IHCA was defined as any patient with unresponsiveness, apnea, and absence of a central pulse. Standardized Utstein definitions are employed within GWTG-Resuscitation to ensure uniform reporting for cardiac arrest variables.(10) Data accuracy within GWTG-Resuscitation is ensured by testing and certification of data entry personnel, use of case-study methods for newly enrolled hospitals before submission of data, data entry software with numerous built in data checks for missing or outlying values, and a reabstraction process that has demonstrated a mean error rate of 2.4% of all data.(1, 11)
Patient Population
Our study included IHCA events from January 1, 2000 through November 23, 2009 among adult inpatients in a general hospital ward bed or an intensive care unit bed at the time of cardiac arrest. For an admission containing multiple cardiac arrest events, only the index (first) cardiac arrest was included. We focused our analysis on patients in the general hospital ward and intensive care units and therefore excluded 28,331 arrests occurring in other hospital areas (Figure 1). Because our study sought to evaluate the association between epinephrine dosing frequency and survival, we excluded 22,212 arrests without at least two epinephrine doses, as dosing period could not be calculated (see below). To avoid confounding with use of vasopressin and other vasopressors, we excluded 31,117 arrests in which a vasopressor infusion was in place at the time of cardiac arrest, and 18,703 arrests during which a vasopressor other than epinephrine was administered. We excluded 9,006 arrests for which our independent variable of interest, epinephrine dosing period, could not be determined, owing to missing or inconsistent times for epinephrine administration or unknown epinephrine dose count (earlier versions of the GWTG-Resuscitation data fields did not include dose count). Finally, we excluded 6,831 arrests with outlier epinephrine dosing periods (i.e., <1 minute or >=10 minutes per dose). Our final study cohort was comprised of 20,909 IHCA for whom epinephrine dosing period could be determined.
Figure 1.
Study Cohort. Of the initial 141,324 in-hospital cardiac arrest events in the Get With the Guidelines(GWTG)-Resuscitation, 20,909 eligible patients were included in the final study population.
Epinephrine Dosing Period
Epinephrine dosing period was defined as the amount of time between the first epinephrine dose and the resuscitation endpoint (either death or return of spontaneous circulation [ROSC] for at least 20 minutes), divided by the total number of epinephrine doses received subsequent to the first epinephrine dose. Epinephrine dosing period was classified in pre-specified categories: 1 to <3 minutes/dose, 3 to <4 minutes/dose, 4 to <5 minutes/dose, 5 to <6 minutes/dose, 6 to <7 minutes/dose, 7 to <8 minutes/dose, 8 to <9 minutes/dose, and 9 to <10 minutes/dose. Epinephrine dosing period is meant to estimate the epinephrine dosing pattern intended by the prescribing physician leading the resuscitation efforts. That physician is apt to verbally prescribe an amount of time, in minutes, between epinephrine doses (e.g., 3 minutes, 4 minutes or 5 minutes per guideline recommendation). The time period prior to the first dose of epinephrine was not considered in the epinephrine dosing period because this initial phase of resuscitation is inherently less organized and likely to have included time when the lead physician was not present to prescribe medications. However, for purpose of secondary analysis, an alternate independent variable, “total epinephrine dosing period”, was used and was defined more simply as the total cardiac arrest duration divided by total number of epinephrine doses administered.
Statistical Analysis
Baseline characteristics among categories of epinephrine dosing period are described. To evaluate the association between epinephrine dosing period and survival to hospital discharge, we constructed multivariable logistic regression models. Generalized estimating equations (GEE) with a logit link, an exchangeable correlation matrix, and a robust variance estimator, were used to examine the association between epinephrine dosing period and survival to hospital discharge, adjusted for patient and event characteristics and for clustering of cardiac arrest cases within hospitals.(12, 13) For our independent variable, we chose an epinephrine dosing period of 4 to <5 minutes/dose as our reference category, as this is the recommended dosing frequency of current resuscitation guidelines.
Covariates included in the adjusted models were chosen a priori based on their association with survival from prior studies of IHCA and biologic plausibility. These included: age, sex, race, patients’ preadmission status of home/self care, admission illness category, pre-arrest comorbidities (congestive heart failure [CHF] this admission, prior CHF, myocardial infarction [MI] this admission, prior MI, acute cerebrovascular accident [CVA], baseline central nervous system deficits, respiratory insufficiency, pneumonia, renal insufficiency, hepatic insufficiency, diabetes, septicemia, metastatic or hematologic cancer, major trauma), coexisting requirement for mechanical ventilation during the cardiac arrest, hospital location of the cardiac arrest (intensive care, ward, ward telemetry), whether the arrest was witnessed, initial pulseless rhythm (asystole, pulseless electrical activity [PEA], ventricular fibrillation [VF], pulseless ventricular tachycardia [VT], whether the arrest occurred on a weekend or at night, cardiac arrest duration, and time to first epinephrine dose.
Covariates missing more than 0.5% of values were assessed for differential distribution of missing data among categories of the primary predictor epinephrine dosing period. Hypothermia treatment was left out of the regression model as most hospitals did not submit information on its use. For continuous variables, assumptions of linearity were tested with partial residual plots.(14) Because the associations between cardiac arrest duration and time to first epinephrine dose with survival to discharge were non-linear, cubic splines with five knots were used to transform these variables for model inclusion.(15)
As a secondary analysis, we repeated the above analyses, stratified by whether the first identifiable rhythm was shockable (VF or pulseless VT) or non-shockable (asystole or PEA). For the analyses with the subgroup of shockable rhythms, time to defibrillation was added to the model as a continuous variable. We also examined the association between epinephrine dosing period and survival using the alternate independent variable, total epinephrine dosing period. Sensitivity analyses were conducted on the primary and secondary analyses. First, the analyses were repeated sequentially after culling cardiac arrests of greater durations (>60, >50, >40, >30, >20 minutes), since patients with transient ROSC are potentially overrepresented among cardiac arrests of greater duration. Second, the analyses were repeated sequentially after culling cardiac arrests with greater duration of time-to-first-epinephrine-dose (>30, >20, >10 minutes). All analyses were performed using Stata/IC 11.1 (StataCorp, College Station, TX). All analyses were specified a priori, with the exception of a post-hoc exploratory analysis where we removed the cardiac arrest duration covariate from our model in order to better understand the difference of the unadjusted and adjusted outcomes.
RESULTS
Of 20,909 IHCA, 1100 (5.3%) patients received epinephrine every 1 to <3 minutes, 2381 (11.4%) every 3 to <4 minutes, 3712 (17.8%) every 4 to <5 minutes, 4184 (20.0%) every 5 to <6 minutes, 3373 (16.1%) every 6 to <7 minutes, 2676 (12.8%) every 7 to <8 minutes, 1961 (9.4%) every 8 to <9 minutes, and 1522 (7.3%) every 9 to <10 minutes (Figure 2). Baseline characteristics of IHCA events among categories of epinephrine dosing period are presented in Table 1. There were no significant differences in the pattern of missing data among categories. In the overall study population, the mean age was 68 ± 16 years, 61% of patients were male, 68% were white and 23% were of black race. The most common admission illness category was medical noncardiac (53%) followed by medical cardiac (27%). Thirty percent of the IHCA events occurred in general hospital beds unmonitored by telemetry, 34% in general hospital beds monitored by telemetry, and 36% in intensive care beds. The most common presenting cardiac rhythm was asystole (46%), followed by PEA (39%), while shockable rhythms comprised less than 13% of events. The mean cardiac arrest duration was 19 ± 10 minutes.
Figure 2.
Frequency distribution of in-hospital cardiac arrest events eligible for analysis, by category of epinephrine dosing period.
Table 1.
Baseline characteristics of 20,909 in-hospital cardiac arrest events by categories of epinephrine dosing period.
| Characteristic | Epirephrine Dosing Inteasily (minutes/dose)* | |||||||
|---|---|---|---|---|---|---|---|---|
| 1–3 min/dose | 3–1 min/dose | 4–5 min/dose | 5–6 min/dose | 6–7 min/dose | 7–8 min/dose | 8–9 min/dose | 9–10 min/dose | |
| (n=1100) | (n=2381) | (n=3712) | (n=4184) | (n=3373) | (n=2676) | (n=1961) | (11=1522) | |
| Age, mean (SD), y | 67.0 (16.1) | 67.6 (15.8) | 67.6 (15.9) | 67.4 (15.8) | 67.7 (16.0) | 67.5 (15.7) | 68.9 (14.8) | 68.2 (15.1) |
| Male sex, no.(%) | 674 (61.3) | 1443 (60.6) | 2277 (61.3) | 2544 (60.8) | 2071 (61.4) | 1642 (61.4) | 1165 (59.4) | 882 (58.0) |
| Race/ethnicity, no.(%) | ||||||||
| White | 760 (69.1) | 1597 (67.1) | 2479 (66.8) | 2850 (68.1) | 2362 (70.0) | 1799 (67.2) | 1374 (70.1) | 1015 (66.7) |
| Black | 223 (20.3) | 571 (24.0) | 850 (22.9) | 967 (23.1) | 740 (21.9) | 612 (22.9) | 411 (21.0) | 367 (24.1) |
| Other | 117 (10.6) | 213 (9.0) | 383 (10.3) | 367 (8.8) | 171 (8.0) | 265 (9.9) | 176 (9.0) | 141) (9.2) |
| Preadmission = home/self care, no.(%) | 796 (72.4) | 1676 (70.4) | 2660 (71.7) | 3048 (72.9) | 2452 (72.7) | 1995 (74.6) | 1470 (75.0) | 1115 (73.3) |
| Preadmission Cerebral Performance Category, no.(%)** | ||||||||
| 1 (normal) | 458 (42.8) | 979 (42.4) | 1596 (44.4) | 1837 (45.3) | 1482 (45.2) | 1186 (45.5) | 875 (45.8) | 646 (44.1) |
| 2 (disability, independent ADLs) | 300 (28.0) | 648 (28.1) | D73 (27.0) | 1107 (27.3) | 883(26.9) | 761 (29.2) | 523 (27.4) | 422 (28.8) |
| 3 (disability, not independent ADLs) | 127 (11.9) | 312 (13.5) | 476 (13.2) | 510 (12.6) | 423 (12.9) | 299 (11.5) | 230 (12.0) | 188 (12.8) |
| 4 (coma) | 47 (4.4) | 65 (2.8) | 87 (2.4) | 127 (3.1) | 75 (2.3) | 66 (2.5) | 50 (2.6) | 42 (2.9) |
| Unknown | 138 (12.9) | 304 (13.2) | 467 (13.0) | 477 (11.8) | 417 (12.7) | 295 (11.3) | 252 (122) | 167 (11.4) |
| Admission Illness category, no.(%) | ||||||||
| Medical cardiac | 317 (28.8) | 605 (25.4) | 999 (26.9) | 1068 (25.5) | 928 (27.5) | 745 (27.8) | 557 (28.4) | 443 (29.1) |
| Surgical cardiac | 52 (4.7) | 109 (4.6) | 138 (3.7) | 166 (4.0) | 134 (4.0) | 110 (4.1) | 84 (4.1) | 57 (3.8) |
| Medical noncardiac | 558 (50.8) | 1265 (53.1) | 2008 (54.1) | 2282 (54.6) | 1803 (53.5) | 1432 (53.5) | 1017 (51.9) | 796 (52.4) |
| Surgical noncardiac | 136 (12.4) | 347 (14.6) | 474 (12.8) | 577 [13.8) | 445 (13.2) | 332 (12.4) | 265 (13.5) | 196 (12.9) |
| Trauma | 36 (3.3) | 54 (2.3) | 87 (2.4) | 87 (2.1) | 61 (1.8) | 56 (2.1) | 36 (1.8) | 27 (1.8) |
| Pre-arrest comorbidity, no.(%)** | ||||||||
| CHF this admission | 182 (17.5) | 338 (15.4) | 624 (18.0) | 651 (16.7) | 580 (18.2) | 463 (18.3) | 325 (17.7) | 252 (17.6) |
| Prier CHF (before admission) | 216 (20.8) | 476 (21.6) | 731 (21.1) | 779 (19.9) | 708 (22.2) | 519 (20.5) | 423 (23.1) | 320 (22.4) |
| MI this admission | 126 (12.1) | 232 (106) | 403 (11.6) | 455 (11.7) | 403 (12.7) | 328 (13.0) | 229 (12.5) | 184 (12.9) |
| Prior MI before admission | 177 (17.0) | 317 (14.4) | 532 (15.4) | 592 (15.2) | 537 (16.9) | 390(154) | 306 (16.7) | 245 (17.1) |
| Acute CVA | 60 (5.8) | 92 (4.2) | 139 (4.0) | 182 (4.7) | 138 (4.3) | 123 (4.9) | 86 (4.7) | 72 (5.0) |
| Baseline CNS deficits | 146 (14.0) | 306 (13.9) | 473 (13.6) | 548 (14.0) | 420 (13.2) | 336 (13.3) | 248 (13.5) | 201 (14.1) |
| Respiratory insufficiency | 446 (42.9) | 854 (38.8) | 1308 (37.8) | 1448 (37.1) | 1158 (36.4) | 974 (38.5) | 687 (37.5) | 532 (37.2) |
| Pneumonia | 165 (15.9) | 362 (16.51 | 522 (15.1) | 589 (15.1) | 475 114.9) | 408 (16.1) | 258 (14.1) | 219 (15.3) |
| Renal insufficiency | 330 (31.7) | 687 (31.2) | 1126 (32.5) | 1212 (31.0) | 1077 (33 S) | 872 (34.5) | 613 (33.4) | 487 (34.0) |
| Hepatic insufficiency | 79 (7.6) | 166 (7.6) | 246 (7.1) | 283 (7.2) | 240 (7.5) | 197 (7.8) | 125 (6.8) | 92 (6.4) |
| Diabetes | 324 (31.2) | 672 (30.6) | 1083 (31.3) | 1237 (31.7) | 1006 (31.6) | 841 (33.3) | 613 (33.4) | 467 (32.6) |
| Septicemia | 139 (13.4) | 338 (15.4) | 480 (13.9) | 537 (13.7) | 458 (14.4) | 374 (14.8) | 241 (13.3) | 194 (13.6) |
| Cancer (metastatic or hematologic) | 150 (14.4) | 339 (15.4) | 536 (15.5) | 616 (15.8) | 501 (15.7) | 365 (14.4) | 276 (15.1) | 209 (14.6) |
| Major Trauma | 46 (4.4) | 80 (3.6) | 115 (3.3) | 122 15.1) | 81 (2.5) | 78 (3.1) | 56 (3.1) | 58 (2.7) |
| Interventions already in place. No.(%) | ||||||||
| Mechanical ventilation | 212 (19.3) | 422 (17.7) | 594 (16.0) | 647 (15.5) | 496 (14.7) | 408 (15.3) | 244 (12.4) | 247 (16.2) |
| IHCA location, no.(%) | ||||||||
| Intensive care | 476 (42.5) | 878 (36.9) | 1324 (35.7) | 1481 (35.4) | 1218 (36.1) | 969 (36.2) | 699 (35.7) | 543 (35.7) |
| Inpatient, monitored by telemetry | 264 (24.0) | 585 (24.6) | 902 (24.3) | 1015 (24.3) | 855 (25.4) | 699 (26.1) | 509 (26.0) | 387 (25.4) |
| Inpatient, unmonitored by telemetry | 369 (33.6) | 918 (38.6) | 1486 (40.0) | 1688 (40.3) | 1300 (38.5) | 1008 (37.7) | 753 (38.4) | 592 (38.9) |
| Arrest characteristics, no.(%) | ||||||||
| Unwitnessed | 321 (29.2) | 789 (33.1) | 1262 (34.0) | 1448 (34.6) | 1114 (33.0) | 879 (32.9) | 628 (32.0) | 495 (32.5) |
| First pulseless rhythm | ||||||||
| VF | 96 (8.7) | 205 (8.6) | 284 (7.7) | 340 (8.1) | 269 (8.0) | 222 (8.3) | 188 (9.6) | 140 (9.2) |
| VT | 56(5.1) | 84 (3.5) | 136 (3.7) | 174 (4.2) | 133 (3.90) | 130 (4.9) | 86 (4.4) | 62 (4.1) |
| PEA | 405 (36.8) | 850 (35.7) | 1425 (38.4) | 1650 (39.4) | 1282 (38.0) | 1089 (40.7) | 793 (40.4) | 651 (42.8) |
| Asystole | 509 (46.3) | 1164 (48.9) | 1764(47.5) | 1903 (45.5) | 1589 (47.1) | 1150 (43.0) | 827 (42.2) | 620 (40.7) |
| Unknown | 34 (3.1) | 78 (3.3) | 103(2.8) | 117 (2.8) | 100(3.0) | 85 (3.2) | 67 (3.4) | 49 (3.2) |
| Night time, 11pm to 6am. no(%) | 367(33.4) | 940 (39.5) | 1534(41.3) | 1690 (40.4) | 1317 (391) | 1034 (38.6) | 811 (41.4) | 617 (40.5) |
| Weekend, Fri 11p – Mon 7a, no.(%)** | 336(33.7) | 757 (35.3) | 1221 (36.5) | 1444 (38.0) | 1165 (38.2) | 913 (37.8) | 678 (39.0) | 515 (37.6) |
| Treatment characteristics | ||||||||
| Time to 1st Cpi dose, mcan(SD) | 3.5(4.7) | 3.1 (3.3) | 3.1 (3.4) | 3.2 (3.5) | 3.1 (3.3) | 3.1 (3.6) | 3.0 (3.5) | 3.0 (3.5) |
| Arrest duration, mean(SD), min | 11.1(8.1) | 15.2 (8.2) | 17.2 (8.5) | 18.6 (9.6) | 20.4 (10.4) | 21.4 (11.4) | 22.5 (12.2) | 23.0 (12.4) |
| Arrest duration after 1st Epi, mean(SD) | 7.6 (6.6) | 12.0 (7.9 | 14.1(8.1) | 15.4 (9.3) | 17.3 (10.0) | 18.3 (10.8) | 19.5 (11.9) | 19.9 (12.2) |
| Hypothermia after ROC, no.(%)** | 10(1.7) | 16 (1.8) | 25(2.1) | 31 (2.3) | 31 (2.6) | 17 (1.8) | 10(1.3) | 14 (2.2) |
Epinephrine dosing period categories are inclusive of lower bound, not-inclusive of upper bound
Missingness less than 0.5% except for the following: pre-arrest comorbidity (6.3%), weekend (Fri 11p – Mon 7a, 9.8%), admission cerebral performance score (14.9%), and hypothermia (treatment) after ROC (63.8%)
Return of spontaneous circulation occurred in 7558 (36.2%) patients and survival to hospital discharge occurred in 1470 (7.0%) patients. The proportion of patients surviving to hospital discharge by category of epinephrine dosing period is presented in Figure 3. Compared to a reference epinephrine dosing period of 4 to <5 minutes/dose, an epinephrine dosing period of 6 to <10 minutes/dose was associated with higher rates of survival to discharge: for 6 to <7 min/dose, adjusted odds ratio [OR], 1.41 (95% CI: 1.12, 1.78); for 7 to <8 min/dose, adjusted OR, 1.30 (95%CI: 1.02, 1.65); for 8 to <9 min/dose, adjusted OR, 1.79 (95%CI: 1.38, 2.32); for 9 to <10 min/dose, adjusted OR, 2.17 (95%CI: 1.62, 2.92); test for linear trend p<0.001. For cardiac arrests due to shockable rhythms, a similar pattern was observed between epinephrine dosing period and survival (Appendix I, p<0.001). For cardiac arrests due to non-shockable rhythms, the overall pattern of association was similar (p<0.001) and an epinephrine dosing frequency of 1–3 minutes/dose was associated with lower survival (adjusted OR, 0.63 [95% CI: 0.47, 0.85], Appendix II).
Figure 3.
Survival to discharge by category of epinephrine dosing period (the amount of time between the first epinephrine dose and the resuscitation endpoint, divided by the total number of epinephrine doses received subsequent to the first epinephrine dose).
The results of the secondary analysis that assessed the alternate predictor, total epinephrine dosing period, are presented in Figure 4. A pattern of increasing survival with increasing total epinephrine dosing period was similarly observed when the total epinephrine dosing period was used (p<0.001).
Figure 4.
Survival to discharge by category of total epinephrine dosing period (total cardiac arrest duration divided by total number of epinephrine doses administered).
In the sensitivity analysis performed to understand the potential confounding of longer cardiac arrest duration, the sequential culling of cardiac arrests of greater duration did not alter the pattern of association between epinephrine dosing period and survival to hospital discharge. The pattern of association between epinephrine dosing period and survival to hospital discharge for cardiac arrests of duration less than 20 minutes yielded similar results (Figure 5). Furthermore, when we removed the cardiac arrest duration covariate from our primary regression model, the associations between epinephrine dosing period and survival to hospital discharge reverted to be in general alignment with the overall unadjusted model.
Figure 5.
Survival to discharge by category of epinephrine dosing period for in-hospital cardiac arrests of duration less than 20 minutes.
DISCUSSION
In this large, national registry of IHCA, we found that epinephrine dosing at higher dosing period (i.e., less frequent dosing) than recommended by consensus guidelines was associated with improved survival. This pattern was consistent for both shockable and non-shockable cardiac arrest rhythms. Moreover, for the majority (87%) of cardiac arrests due to non-shockable rhythms, an epinephrine dosing period of 1 to <3 minutes/dose was associated with lower rates of survival.
The primary predictor, epinephrine dosing period, was designed as a marker of the epinephrine dosing pattern intended by the prescribing physician leading the resuscitation effort. Our findings remained consistent when total epinephrine dosing period over the full duration of cardiac arrest was used as an alternate predictor. It remains unclear if the association of improved survival with increased epinephrine dosing period was caused by the potential harmful effects of epinephrine or if another unmeasured factor, such as interrupted CPR, was at play.
An appropriate strategy against confounding by indication was a major emphasis in our study design process.(16) We anticipated that longer cardiac arrest durations would be associated with differing epinephrine dosing periods, but it came as some surprise that this association was a negative and graded one (Table 1, p<0.001). In other words, on average, epinephrine was dosed less frequently when cardiac arrests were of greater duration, despite the lower survival potential of prolonged cardiac arrest. Less surprising, the primary outcome, survival to hospital discharge, was negatively and non-linearly associated with cardiac arrest duration. In anticipation of such complex observations, cubic splines were pre-specified as the flexible form in which to model two important time covariates, overall duration of cardiac arrest and time to first epinephrine dose.(15)
In the context of current treatment guidelines for IHCA, our results are consistent with those to be expected were the recommended epinephrine dosing period of every 3 minutes, 4 minutes or 5 minutes to be harmful compared with less frequent epinephrine dosing.
Our work should be interpreted in the context of the following limitations. First, confounding by indication may be present if treating providers used epinephrine more sparingly for patients who were immeasurably less sick (e.g., patients with intermittent return of pulse without meeting ROSC definition). If this were the case, however, we would expect that the sensitivity analyses performed after the sequential culling of cardiac arrests of greater duration would have affected our inference, and they did not. Second, time data were self-reported by hospitals and inaccuracies may exist. However, any inaccuracies in the time variables would be expected to be nondifferential and to bias our observations toward the null hypothesis. Third, we were unable to account for epinephrine dose amount in addition to dose frequency. However, the highest dose of epinephrine administered was recorded and <4% were greater than one milligram. Fourth, we were unable to adjust for post resuscitation hypothermia treatment even though two trials have suggested its clinical utility in out-of-hospital cardiac arrest.(17, 18) Fifth, our results may not be generalizable to other populations (e.g., children) or hospital locations. Finally, although data from 505 different facilities were included in our work, observations made among hospitals that participate in the NRCPR may not generalize to other hospitals.
In summary, we found that less frequent epinephrine dosing than recommended by consensus guidelines was associated with improved survival of in-hospital cardiac arrest. Our findings suggest that clinical trials may be needed to determine the role and dose frequency of epinephrine in the treatment of in-hospital cardiac arrest.
Supplementary Material
Acknowledgments
We thank the AHA for its support of GWTG-R; innumerable staff and data abstractors from GWTG-R hospitals for their time and effort; The Biostatistics Analysis Center staff at the University of Pennsylvania for GWTG-R database management; and hard-working hospital clinicians and staff who take initiatives to improve teamwork in the care of IHCA every day in hospitals around the world.
Funding/Support: Data acquisition was funded by Get With The Guidelines – Resuscitation which is supported by the American Heart Association and membership fees paid by participating hospitals. The remainder of the study was funded by a Fellowship training grant from Department of Health and Human Services, Public Health Services, Grant Type 6, Activity T32, ID Serial No. HP10002-21-00, Principal Investigator: Paula Lozano, MD, MPH.
Footnotes
Author Contributions: S.A.W. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: S.A.W., E.H., S.M.B., P.S.C., G.N.
Acquisition of data: S.A.W., P.S.C., G.N.
Statistical analysis and interpretation of data: S.A.W., E.H., S.M.B., P.S.C., C.L.B., G.N.
Drafting of the manuscript: S.A.W., E.H., S.M.B., P.S.C., G.N.
Critical revision of the manuscript for important intellectual content: S.A.W., E.H., S.M.B., P.S.C., C.L.B., A.L.F., G.N.
Author Conflicts of Interest Disclosures: S.A.W: volunteer GWTG-R investigator, includes travel and meal reimbursement for occasional GWTG-R work meetings; E.H.:XXXXX; S.M.B.: XXXXXXX; P.S.C.: XXXXX; C.L.B.: XXXX; A.L.F.: XXXXXX; G.N.: XXXXXX.
References
- 1.Peberdy MA, Kaye W, Ornato JP, Larkin GL, Nadkarni V, Mancini ME, et al. Cardiopulmonary resuscitation of adults in the hospital: A report of 14720 cardiac arrests from the national registry of cardiopulmonary resuscitation. Resuscitation. 2003 Sep;58(3):297–308. doi: 10.1016/s0300-9572(03)00215-6. [DOI] [PubMed] [Google Scholar]
- 2.http://www.nationmaster.com/graph/hea_hos_bed-health-hospital-beds.
- 3.Nichol G, Thomas E, Callaway CW, et al. Regional variation in out-of-hospital cardiac arrest incidence and outcome. JAMA. 2008 Sep 24;300(12):1423–31. doi: 10.1001/jama.300.12.1423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ehlenbach WJ, Barnato AE, Curtis JR, Kreuter W, Koepsell TD, Deyo RA, et al. Epidemiologic study of in-hospital cardiopulmonary resuscitation in the elderly. N Engl J Med. 2009 Jul 2;361(1):22–31. doi: 10.1056/NEJMoa0810245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Weil MH, Sun S, Tang W. Adrenergic agonists. In: Paradis N, Halperin H, Kern K, Wenzel V, Chamberlain D, editors. Cardiac Arrest: The Science and Practice of Resuscitation Medicine. 2nd. Cambridge University Press; 2007. pp. 639pp. 640–646. [Google Scholar]
- 6.Part 8: Adult advanced cardiovascular life support: 2010 american heart association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care – neumar et al. 122 (183): S729 – circulation [Internet].; cited 11/11/2010]. Available from: http://circ.ahajournals.org.offcampus.lib.washington.edu/cgi/content/full/122/18_suppl_3/S729.
- 7.Cummins RO, Hazinski MF. The next chapter in the high-dose epinephrine story: Unfavorable neurologic outcomes? Ann Intern Med. 1998 Sep 15;129(6):501–2. doi: 10.7326/0003-4819-129-6-199809150-00014. [DOI] [PubMed] [Google Scholar]
- 8.Ian J, Judith F. Adrenaline and vasopressin for cardiac arrest. In: Ian Jacobs, Judith Finn., editors. Adrenaline and vasopressin for cardiac arrest. Cochrane Database of Systematic Reviews: Protocols. Issue 3. Vol. 2001. John Wiley & Sons, Ltd; Chichester, UK: 2001. [DOI] [Google Scholar]
- 9.Vandycke C, Martens P. High dose versus standard dose epinephrine in cardiac arrest – a meta-analysis. Resuscitation. 2000 Aug 1;45(3):161–6. doi: 10.1016/s0300-9572(00)00188-x. [DOI] [PubMed] [Google Scholar]
- 10.Jacobs I, Nadkarni V, Bahr J, Berg RA, Billi JE, Bossaert L, et al. Cardiac arrest and cardiopulmonary resuscitation outcome reports: Update and simplification of the utstein templates for resuscitation registries: A statement for healthcare professionals from a task force of the international liaison committee on resuscitation (american heart association, european resuscitation council, australian resuscitation council, new zealand resuscitation council, heart and stroke foundation of canada, InterAmerican heart foundation, resuscitation councils of southern africa) Circulation. 2004 Nov 23;110(21):3385–97. doi: 10.1161/01.CIR.0000147236.85306.15. [DOI] [PubMed] [Google Scholar]
- 11.Chan PS, Krumholz HM, Nichol G, Nallamothu BK. Delayed time to defibrillation after in-hospital cardiac arrest. N Engl J Med. 2008 Jan 3;358(1):9–17. doi: 10.1056/NEJMoa0706467. [DOI] [PubMed] [Google Scholar]
- 12.Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: An orientation. Am J Epidemiol. 2003 Feb 15;157(4):364–75. doi: 10.1093/aje/kwf215. [DOI] [PubMed] [Google Scholar]
- 13.Hubbard AE, Ahern J, Fleischer NL, Van der Laan M, Lippman SA, Jewell N, et al. To GEE or not to GEE: Comparing population average and mixed models for estimating the associations between neighborhood risk factors and health. Epidemiology. 2010 Jul;21(4):467–74. doi: 10.1097/EDE.0b013e3181caeb90. [DOI] [PubMed] [Google Scholar]
- 14.Stata web books. regression with stata. chapter 2 – regression diagnostics. [Internet]. Available from: http://www.ats.ucla.edu/stat/stata/webbooks/reg/chapter2/statareg2.htm (accessed November 9, 2010)
- 15.Royston PSW. Multivariable modeling with cubic regression splines: A principled approach. The Stata Journal. 2007;7(1):45. [Google Scholar]
- 16.Psaty BM, Siscovick DS. Minimizing bias due to confounding by indication in comparative effectiveness research: The importance of restriction [commentary] JAMA: The Journal of the American Medical Association. 2010 Aug 24;304(8):897–8. doi: 10.1001/jama.2010.1205. Tue. [DOI] [PubMed] [Google Scholar]
- 17.Bernard SA, Gray TW, Buist MD, Jones BM, Silvester W, Gutteridge G, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med. 2002 Feb 21;346(8):557–63. doi: 10.1056/NEJMoa003289. [DOI] [PubMed] [Google Scholar]
- 18.Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med. 2002 Feb 21;346(8):549–56. doi: 10.1056/NEJMoa012689. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.