Automated heparin-delivery system to control activated partial thromboplastin time: evaluation in normal volunteers.

BACKGROUND: Unfractionated heparin is used widely; however, control of the level of anticoagulation remains its greatest problem, with fewer than 35% of patients having activated partial thromboplastin times (aPTTs) within a range of 55 to 85 seconds in recent trials. METHODS AND RESULTS: We developed and tested a prototype of an automated heparin control system (AutoHep) in which a computer-based titration algorithm adjusted the heparin infusion to reach a target aPTT. In 1 study, 12 healthy male subjects received an intravenous infusion of heparin with the rate determined by AutoHep and were randomized to receive an initial bolus or no bolus of heparin preceding the infusion. A second study evaluated the automated blood sampling system in 12 subjects. Of the 344 end-point aPTT measurements, 78% were within +/-10 seconds of the target (prespecified primary end point), and 89% were within a +/-15-second range. The time to achieve a target aPTT was 93 minutes without and 150 minutes with an initial heparin bolus. The total percentage of time within the target range +/-15 seconds was 46 of 48 hours (96%). The automatic blood sampling system successfully obtained 96% of all scheduled samples. CONCLUSIONS: These results suggest that the AutoHep system has the potential to significantly improve aPTT control of intravenous heparin compared with current clinical practice.

Modeling and closed-loop pharmacologic control of the ventricular rate during induced atrial fibrillation in anesthetized dogs.

An automated drug delivery system that provides closed-loop feedback control of the ventricular rate during atrial fibrillation is described. The control system was designed using a mathematical model of the effect of esmolol infusion in the ventricular rate. The model was developed in system identification experiments with anesthetized dogs in which atrial fibrillation was induced and maintained by rapid atrial pacing. A control system of variable structure, which incorporates a transient controller and a regulator, was designed to perform satisfactorily over a wide range of subject responses to drug infusion. The transient controller brings the ventricular rate to the setpoint with little overshoot. When the ventricular rate is near the setpoint, the drug infusion rate is calculated by the regulator. The drug infusion rate is constrained to ensure smooth transitions in the hemodynamic state of the patient and for safety. Feasibility of the system was demonstrated in computer simulations and animal experiments.

Automated administration of lidocaine for the treatment of ventricular arrhythmias.

An automated lidocaine infusion system has been developed for the treatment of ventricular arrhythmias. A drug serum concentration which is in the therapeutic range is initially achieved and maintained by an open-loop infusion regimen. The lidocaine infusion rate is then adjusted using closed-loop control to regulate the rate of premature ventricular contractions (PVC rate) at a setpoint. The major system components are a portable computer, a computer-controlled drug infusion pump, and an arrhythmia monitor which is used to measure the PVC rate. During closed-loop control, the drug infusion rate is constrained to keep the predicted serum concentration below an upper limit (ceiling concentration) so that toxic serum concentrations are avoided. The infusion rate is also constrained to keep the predicted serum concentration above a minimum value (floor concentration) to ensure the maintenance of a therapeutic serum level during closed-loop control. The system has been evaluated in initial clinical trials.

A comparison of the noise sensitivity of nine QRS detection algorithms.

The noise sensitivities for nine different QRS detection algorithms were measured for a normal, single-channel lead II, synthesized ECG corrupted with five different types of synthesized noise. The noise types were electromyographic interference, 60 Hz powerline interference, baseline drift due to respiration, abrupt baseline shift, and a composite noise constructed from all of the other noise types. The percentage of QRS complexes detected, the number of false positives, and the detection delay were measured. None of the algorithms were able to detect all QRS complexes without any false positives for all of the noise types at the highest noise level. Algorithms based on amplitude and slope had the highest performance for EMG-corrupted ECG. An algorithm using a digital filter had the best performance for the composite noise corrupted data.

A microcomputer based controller for neuromuscular block during surgery.

A microcomputer based control system has been developed to automatically induce and maintain neuromuscular block during surgery. The system repeatedly adjusts the infusion rate of a muscle relaxant, succinylcholine, based on the evoked, rectified, and integrated electromyogram (EMG). The system was tested in 12 patients for a 30-min infusion period with a setpoint of 80% depression of the evoked, rectified, and integrated thenar EMG. The mean time to reach the setpoint for 10 of the patients was 5.5 (+/- 1.87 SD) min and the mean time for 95% recovery after infusion was stopped was 5.4 (+/- 0.83 SD) min. Average overshoot was 9.9% of the baseline (+/- 3.1% SD), and the average time within +/- 10% of the setpoint was 22.1 min (+/- 7.26 SD). The total dose of succinylcholine for these 10 patients ranged from 1.21 to 3.77 mg/kg with a mean of 1.92 mg/kg. The other two patients were relatively insensitive to the drug and the controller was unable to bring the response to the setpoint due to a ceiling placed on infusion rate by the control algorithm.