• Henry E Wang, Douglas F Kupas, Paul M Paris, and Donald M Yealy.
• Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15213, USA. wanghe@msx.upmc.edu
• Prehosp Emerg Care. 2004 Jan 1; 8 (1): 1-9.

ObjectiveTo identify a set of clinical factors most strongly associated with the use of drug-facilitated intubation (DFI) in the out-of-hospital setting.MethodThe authors used data from a prospective, multicentered endotracheal intubation (ETI) observational cohort trial, including patients from 45 emergency medical services in Pennsylvania. Providers reported clinical, physiologic, and anatomic factors associated with each ETI effort. The authors included only data from the 23 services using DFI. They identified all non-arrest (presence of a pulse) adult patients. They included both successful and failed ETIs. They defined DFI cases as patients who received a sedative or neuromuscular-blocking agent to facilitate ETI. The authors also classified patients who underwent nasotracheal intubation as DFI. They defined control subjects as patients undergoing conventional oral ETI. They performed multivariate logistic regression to identify the clinical, physiologic, and anatomic factors characteristic of DFI. They examined alternate forms of the final prediction model.ResultsThe authors analyzed data from 208 nonarrest patients, including 92 DFIs and 116 control subjects. Of 34 factors potentially related to DFI, 17 were excluded on univariate analysis (likelihood ratio p>0.25). Multivariate logistic regression revealed the following as positively associated with DFI: presence of clenched jaw/trismus (odds ratio [OR], 2.33; 95% confidence interval [CI], 1.10-4.95; p=0.026); increased verbal Glasgow Coma Scale score (OR, 1.71; 95% CI, 1.29-2.26; p<0.001); use of cervical spine precautions (OR, 2.30; 95% CI, 1.15-4.62; p=0.018). Anterior vocal cords (OR, 0.27; 95% CI, 0.10-0.71; p=0.004) and laryngospasm (OR, 0.14; 95% CI, 0.02-1.17; p=0.025) were negatively associated with DFI. The model showed good fit (Hosmer-Lemeshow p=0.75) and discrimination (area under the curve=0.76).ConclusionsThe authors identified a set of predictors strongly associated with DFI. These data offer insight into the current use of DFI and support the development of consensus-based guidelines for this procedure.

11 keywords...

hide…