was designed to evaluate 3 different methods of circulatory support: standard CCC, opep-chest manual compression (OCMC), and direct mechanical ventricular assistance (DMVA). DMVA is a method of open-chest circulatory support using a glass assistor cup that fits over the heart and alternately compresses and expands the ventricles to provide systole and diastole. The systolic duration, rate of compression, and force of assistor compression can be controlled by the main drive system. Ventricular fibrillation was induced in 15 dogs. They initially were given 10 rain of CCC using a Thumper TM. At the end of this period [hey were divided into 3 groups for continued resuscitation. Group I received 10 min of OCMC at 60/min followed by 10 rain at 90/min. DMVA was then applied for similar periods at rates of 60 and 90. In Group II, DMVA preceded OCMC. The experimental design was otherwise the same as in Group I. The data from this group were used to determine whether the order in which OCMC or DMVA was applied had any significant effects. Group IlI continued to receive CCC at 60/min for an additional 40 min. The total arrest time for all 3 groups was 50 min. Analysis of Groups I and II showed no significant difference in the order of application of OCMC and DMVA. CCC produced a cardiac index (CI) of 780 (19% of normal) with a mean arterial pressure (MAP) of 26 mm Hg (23% of normal). Compared to CCC, both forms of open-chest resuscitation produced higher values for all indices. OCMC at 60/ rain maintained a CI of 2,069 (52% of normal) with a MAP of 50 mm Hg (45% of normal). DMVA at the same rate produced a CI of 2,780 (70% of normal) with a MAP of 72 mm Hg (65% of nor-mal}. These represent significant increases when compared to OCMC at 60 (P < .005 for CI and < .0005 for MAP). Changing from standard CCC to DMVA at 90/rain produced the greatest hemodynamic improvements: diastolic pressure increased by 380%; MAP, by 250%; and CI, by 340%. With DMVA at 90, the systolic pressure, stroke index, and CI could not be statistically distinguished from case-controlled pre-arrest values. This study indicates that DMVA is capable of long-term circulatory support during ventricular fibrillation.