Modelling and control of neuromuscular blockade.

摘要

Standard administration of neuromuscular blockade drugs is in the form of large doses when required. This administration results in repeated overdosing, and creates problems for the surgery and for the patient, including loss of fine control and inability to intervene intra-operatively, and increased recovery time, curarization and reduced function post-operatively.;The goals of this thesis were to produce a system overcoming the mentioned limitations, and to prove its efficacy in a prospective, blinded, randomized, controlled, clinical trial. Many novelties were enacted, including: (1) Monitoring techniques for improved modeling and sensing; (2) Inter-conversion of neuromuscular stimuli protocols; (3) Description of NMB through relaxation and pseudo-occupancy; (4) Linearization of the neuromuscular junction; (5) Modelling techniques for nonlinear systems in noisy environments; (6) Model adaptation schemes; and (7) Simulation of intrapatient variance. The result was an adaptive control computer program, the Neuromuscular Blockade Advisory System (NMBAS). The NMBAS suggests rocuronium dosage and administration time based on a patient model and the history of the patient's response, to avoid the problems associated with conventional NMB drug administration.;The NMBAS was compared to standard care in a prospective, randomized, controlled, clinical trial (N = 60). The NMBAS group experienced fewer peri-operative adverse events, and had reduced danger of inadequate reversal. Drug use and the incidence of postoperative adverse events were not different.;Other items investigated included: using body mass index (BMI) to reduce overparalysis; stimulation current according to patient weight and BMI; and intrapatient variance due to anesthetic, blood loss and tissue loss.;By application of modelling and control to NMB, the problems of standard administration could be remedied by controlling and adjusting the dosing to the patient's needs. Prior work has had some success but with limitations: controllers were conservative; setpoints tended towards irreversibility; computer control was instituted after induction; monitoring with single twitch required delays at the start to tune the sensor; the systems were not tested in blinded, randomized, controlled, clinical trials.