ObjectiveIntravenous administration of vasopressin during cardiopulmonary resuscitation (CPR) may be more effective than optimal doses of epinephrine. The main purpose of this study was to determine whether intraosseous vasopressin achieves serum drug levels comparable with intravenous doses during CPR and, additionally, to evaluate the effects of intraosseous vasopressin during CPR.DesignProspective, randomized laboratory investigation using an established porcine model with instrumentation for measurement of hemodynamic variables, blood gases, and return of spontaneous circulation.SettingUniversity hospital laboratory.SubjectsTwelve domestic pigs.InterventionsAfter 4 mins of untreated ventricular fibrillation and 3 mins of CPR, 12 pigs were randomized to be treated with intravenous administration of vasopressin (0.8 unit/kg vasopressin; n = 6) or intraosseous vasopressin (0.8 unit/kg vasopressin; n = 6). Defibrillation was performed 5 mins after drug administration to attempt the return of spontaneous circulation.Measurements and Main ResultsAt both 90 secs and 5 mins after drug administration, intravenous and intraosseous administration of vasopressin resulted in comparable mean (+/- SEM) coronary perfusion pressure (43 +/- 4 vs. 44 +/- 3 and 30 +/- 2 vs. 37 +/- 2 mm Hg, respectively) and vasopressin plasma concentrations (13,706 +/- 1,857 vs. 16,166 +/- 3,114 pg/mL and 10,372 +/- 883 vs. 8246 +/- 2211 pg/mL, respectively). All animals in both groups were successfully resuscitated; pigs that received intraosseous vasopressin had a significantly higher (p < .05) mean arterial (92 +/- 6 vs. 129 +/- 12 mm Hg) and coronary perfusion pressure (84 +/- 11 vs. 119 +/- 11 mm Hg) at 5 mins of return of spontaneous circulation.ConclusionsIntraosseous vasopressin resulted in comparable vasopressin plasma levels, hemodynamic variables, and return of spontaneous circulation rates as did intravenous vasopressin. Intraosseous vasopressin may be an alternative for vasopressor administration during CPR, when intravenous access is delayed or not available. (Crit Care Med 1999; 27:1565-1569)