Association of Timing of Epinephrine Administration With Outcomes in Adults With Out-of-Hospital Cardiac Arrest

This cohort study evaluates survival and functional status outcomes associated with timing of epinephrine administration in adults with out-of-hospital cardiac arrest with or without initial shockable cardiac rhythms.


Time-dependent propensity score
We calculated propensity score as the time-varying probability of receiving epinephrine which was estimated from the hazard component at any given minutes after ALS arrival from Fine-Gray regression model with time-dependent covariates, time-independent covariates, competing risk events, and a censoring. [1][2][3][4] The time-dependent covariates were shock delivery (if a patient received shock), AAM (if a patient received AAM), and departure from the scene (if a patient was transported) after ALS arrival. The time-independent covariates were patient age, sex, race, location of arrest, etiology of arrest, witness status, layperson CPR, shock delivery before ALS arrival, and EMS response time. The time-dependent and time-independent covariates are presented in Table 1. We used spline functions (B-spline) for continuous variables (age and EMS response time). We chose these covariates a priori based on their association with survival from prior knowledge, biologic plausibility, and adequate ascertainment. 5-10 We included prehospital ROSC and TOR before epinephrine administration as competing risks in the model because (1) epinephrine administration never occurs after ROSC and TOR except cases with re-arrest after ROSC, (2) our interest was timing of epinephrine for initial arrest, and (3) ROSC and TOR are informative censorings. We modelled hospital arrival as a censoring because our main interest was the timing of the first prehospital epinephrine administration.

Linear assumption in a model treating the timing of epinephrine as a continuous variable
We assumed a linear relation between each outcome and the timing of epinephrine administration. To investigate the robustness of this linear assumption, we demonstrated the overlapping of RRs with 95% CIs in two models. We also explored a spline regression model assuming a nonlinear relationship between each outcome and the timing of epinephrine Support; Pediatric Life Support; Neonatal Life Support; Education, Implementation, and Teams;and First Aid Task Forces. Circulation. 2019;140(24):e826-e880. 11.
ALS indicates advanced life support; DNR do-not resuscitate; EMS emergency medical services; OHCA out-of-hospital cardiac arrest. eFigure 2. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) by the timing of the administration for patients with shockable OHCA (matching without replacement).
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 2B). eFigure 2A: Change of RR per minute = -4.6% (95% CI -7.7% to -1.4%, p-value for the interaction=0.005). eFigure 2B: Change of RR per minute = -4.8% (95% CI -8.4% to -1.0%, pvalue for the interaction=0.02). eFigure 2C: Change of RR per minute = 0.7% (95% CI -1.0% to 2.5%, p-value for the interaction=0.409). ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation. eFigure 3. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) by the timing of the administration for patients with nonshockable OHCA (matching without replacement).
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 3B). eFigure 3A: Change of RR per minute = -2.0% (95% CI -6.9% to 3.0%, p-value for the interaction=0.425). eFigure 3B: Change of RR per minute = -1.8% (95% CI -9.3% to 6.3%, pvalue for the interaction=0.300). eFigure 3C: Change of RR per minute =0.03% (95% CI -1.3% to 1.4%, p-value for the interaction=0.963). ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation. eFigure 4. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) by the timing of the administration for patients with shockable OHCA excluding those who had ROSC or TOR within 5 minutes after ALS EMS arrival on the scene.
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 4B). eFigure 4A: Change of RR per minute = -6.3% (95% CI -8.5% to -4.1%, p-value for the interaction<0.001). eFigure 4B: Change of RR per minute = -7.4% (95% CI -10.1% to -4.7%, pvalue for the interaction<0.001). eFigure 4C: Change of RR per minute = 0.4% (95% CI -0.9 to 1.7%, p-value for the interaction=0.52). ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation; TOR termination of resuscitation. eFigure 5. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) by the timing of the administration for patients with nonshockable OHCA excluding those who had ROSC or TOR within 5 minutes after ALS EMS arrival on the scene.
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 5B). eFigure 5A: Change of RR per minute = -3.2% (95% CI -6.3% to -0.03%, p-value for the interaction=0.048). eFigure 5B: Change of RR per minute = -5.0% (95% CI -9.8% to -0.07%, pvalue for the interaction=0.03). eFigure 5C: Change of RR per minute = 1.3% (95% CI 0.4% to 2.3%, p-value for the interaction=0.007). ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation; TOR termination of resuscitation. eFigure 6. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) by the timing of the administration for patients with bystander witnessed shockable OHCA.
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 6B). eFigure 6A: Change of RR per minute = -5.4% (95% CI -7.8% to -3.0%, p-value for the interaction<0.001). eFigure 6B: Change of RR per minute = -6.5% (95% CI -9.3% to -3.5%, pvalue for the interaction<0.001). eFigure 6C: Change of RR per minute = 0.4% (95% CI -1.1% to 2.0%%, p-value for the interaction=0.56). ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation. eFigure 7. Association of epinephrine administration with survival to hospital discharge (A), favorable functional outcome at hospital discharge (B), and prehospital ROSC (C) for patients with bystander witnessed nonshockable OHCA.
Point estimates of the association of epinephrine with outcomes (solid lines) were reported with 95% CIs (dot lines), treating timing of epinephrine administration after ALS EMS as a continuous variable. Plots indicate point estimates of the association of epinephrine with outcomes with 95% CIs, treating timing as a categorical variable. The plots were placed at median time for each categorized time. We rounded decimal points of the numbers of imputed patients with favorable functional status (eFigure 7B). eFigure 7A: Change of RR per minute = -2.8% (95% CI -7.2% to 1.8%, p-value for the interaction=0.23). eFigure 7B: Change of RR per minute = -5.8% (95% CI -11.7% to 0.4%, pvalue for the interaction=0.09). eFigure 7C: Change of RR per minute = 1.6% (95% CI 0.4% to 2.9%, p-value for the interaction=0.01) ALS indicates advanced life support; CI confidence interval; EMS emergency medical services; OHCA out-of-hospital cardiac arrest; ROSC return of spontaneous circulation.