Evaluation of two adaptive sodium nitroprusside control algorithms.

Computer control of sodium nitroprusside infusion may be safer and provide better control of arterial blood pressure than is achieved with manual control. In a series of test maneuvers in 20 mongrel dogs, the performance of two adaptive control algorithms (controllers) was compared and their safety tested. The controllers were set to infuse sodium nitroprusside to decrease mean arterial pressure and maintain it 20 to 30 mm Hg below the control pressure. Then, sequentially, the right atrium was paced to simulate a supraventricular tachydysrhythmia, the right ventricle was intermittently paced to simulate ventricular extrasystoles, large tidal volumes were given to simulate a respiratory-therapy maneuver, the catheter was clamped to simulate clotting, an air bubble was introduced, and the infused sodium nitroprusside concentrations were either doubled or halved. Next, 500 ml of blood was drawn. Then, in sequence, positive end-expiratory pressure was applied, the right atrium was paced, and large tidal volume breaths were given to cause the blood pressure to fluctuate. When the controllers were turned on, mean arterial pressure reached the set point and remained within 5 mm Hg of the target pressure after 8.6 +/- 0.9 minutes (mean +/- SEM). The controllers properly handled the differences in sodium nitroprusside sensitivity and the catastrophic challenges presented in the experiments. When the animals were not being disturbed, stability was maintained and the blood pressure was kept well within 5 mm Hg of the desired pressure. The controllers rejected all invalid pressure signals during testing.(ABSTRACT TRUNCATED AT 250 WORDS)