Derivation and evaluation of a termination of resuscitation clinical prediction rule for advanced life support providers

Abstract

The primary aim was to derive a new termination of resuscitation (TOR) clinical prediction rule for advanced life support paramedics (ALS) and to measure both its pronouncement rate and diagnostic test characteristics. Secondary aims included measuring the test characteristics of a previously derived and published basic life support termination of resuscitation (BLS TOR) clinical prediction rule [Morrison LJ, Visentin LM, Kiss A, et al. Validation of a rule for termination of resuscitation in out-of-hospital cardiac arrest. N Engl J Med 2006;355(5):478-87] on the same cohort of patients for comparison purposes. Secondary data analysis of adult cardiac arrests treated by ALS in rural and urban EMS systems participating in the OPALS study (data extracted from Phase III). A previous study for a basic life support termination of resuscitation (BLS TOR) clinical prediction rule proposed Termination of Resuscitation if the patient had no return of spontaneous circulation (ROSC) before transport; no shock administered; EMS personnel did not witness the arrest [Morrison LJ, Visentin LM, Kiss A, et al. Validation of a rule for termination of resuscitation in out-of-hospital cardiac arrest. N Engl J Med 2006;355(5):478-87]. Multivariable logistic regression was used to examine the relationship between these variables, additional Utstein variables, and the primary outcome of survival to hospital discharge. Diagnostic test characteristics were measured for both the ALS TOR and BLS TOR models on this derivation cohort. Four thousand six hundred and seventy-three cardiac arrest patients were included; 3098 (66%) were male, mean (S.D.) age 69 (15); 239 (5.1%; 95% CI 4.5-5.8) survived to hospital discharge; 3841 patients had no ROSC (82%) and of these only three survived (0.08%; 95% CI 0.02, 0.23). The final ALS TOR model associated with survival, included: ROSC (OR 260.9; 95% CI 96.3, 706.7), bystander witnessed (OR 2.0; 95% CI 1.3, 3.1), bystander CPR (OR 2.8; 95% CI 1.9, 4.1), EMS witnessed (OR 12.3; 95% CI 7.1, 21.3) and shock prior to transport (OR 6.4; 95% CI 4.1, 10.1). A new ALS TOR clinical prediction rule based on these variables was 100% sensitive (95% CI 99.9-100) for survival and had 100% negative predictive value (95% CI 99.9-100) for death. Under the ALS TOR clinical prediction rule, 30% of patients would be pronounced in the field. The BLS TOR clinical prediction rule, was 100% sensitive (95% CI 99.9, 100), had 100% negative predictive value (95% CI 99.9-100) and the field pronouncement rate was 48%. Cardiac arrest patients may be considered for prehospital ALS TOR when there is no ROSC prior to transport, no shock delivered, no bystander CPR and the arrest was not witnessed by bystanders or EMS. A single EMS termination clinical prediction rule for all levels of providers would be optimal for EMS systems to implement. Prospective evaluation of the ALS TOR clinical prediction rule in the hands of ALS providers will be required before implementation.