Thomas Bump

Diagnosis of Atrial Fibrillation Using Electrograms from Chronic Leads: Evaluation of Computer Algorithms

This study compares the performance of three detection algorithms for the recognition of atrial fibrillation in chronic pacing leads. Multiple serial recordings were obtained of wideband and filtered electrograms from chronic atrial and ventricular leads in dogs for a period up to 55 days following implantation. Each dog was recorded in sinus rhythm and induced atrial fibrillation. Four days were chosen for processing: The day of implantation and a day in the first, second or third, and fifth weeks. Three signal processing methods were assessed for performance in detection of atrial fibrillation: software recognition of rate with automatic threshold control, amplitude distribution, and frequency spectral analysis. A software trigger for rate determination was adjusted to thresholds of 10, 20, and 30% of maximum baseline‐to‐peak amplitude. At 10%, a rate boundary anywhere between 420 and 560 beats per minute (bpm) perfectly separated atrial fibrillation from sinus rhythm even though atrial electrograms were contaminated with large QRS deflections and double‐sensing was present. At 20% and 30%, a rate boundary around 300 bpm could be used, but sensitivity and specificity were reduced to 90%. In amplitude distribution analysis, a percent of time within a baseline window provided perfect separation of atrial fibrillation from sinus rhythm. In all cases, the signal was within this window Jess than 43% of the time in atrial fibrillation, and more than 43% in sinus rhythm. In spectral analysis, frequency bands were examined for power content. In the 6 to 30 Hz band atrial fibrillation contained the greater power. Choosing 58% of total power as a discriminant, sensitivity and specificity of atrial fibrillation detection were 100% and 95% respectively.

Hemodynamic recovery during simulated ventricular tachycardia: role of adrenergic receptor activation

Ventricular tachycardia (VT) produces a wide variety of hemodynamic outcomes. Variations in autonomic nervous system response were studied in an animal model of VT. In 18 dogs anesthetized with chloralose VT was simulated by ventricular pacing (rate 240 bpm). Dynamic changes in left ventricular (LV) function were assessed during sinus rhythm and after VT was initiated, under variable autonomic conditions: ganglionic blockade with hexamethonium (n = 5), alpha-adrenergic blockade with terazosin (n = 7; 0.3 mg/kg), and beta-adrenergic blockade with propranolol (n = 6; 2 mg/kg). Micromanometers were used to measure LV pressure, and endocardial piezo crystals assessed changes in cavity size. Sinus interval, an index of autonomic tone, was determined immediately after tachycardia was terminated. Under control conditions the onset of simulated VT was accompanied by severe hypotension, with a decline in LV systolic pressure from 113 +/- 5 to 67 +/- 4 mm Hg within 10 seconds (p less than 0.05). Subsequently, during persistent tachycardia peak LV pressure recovered to sinus values, and maximum +dP/dt exceeded sinus values by 20 seconds (2604 +/- 413 vs 2112 +/- 184 mm Hg/sec; 20 seconds for VT vs sinus rhythm). Diastolic pressures were unchanged, and sinus rate accelerated. Ganglionic blockade with hexamethonium resulted in persistent hypotension, blunted +dP/dt, no change in diastolic pressures, and failure of the sinus rate to accelerate after the tachycardia. After beta blockade there was sustained hypotension (LV systolic pressure 78 +/- 4 vs 120 +/- 5 mm Hg; 20 seconds for VT vs sinus rhythm), maximum +dP/dt was blunted, and minimum diastolic ventricular pressure rose. This was due to an upward shift in the diastolic pressure-dimension relationship associated with prolongation of the time constant of LV relaxation. The sinus interval did not change. In contrast, tachycardia during alpha blockade produced a sustained fall in peak LV pressure; however, maximum +dP/dt recovered (2194 +/- 328 vs 2154 +/- 153 mm Hg/sec; 20 seconds for VT vs sinus rhythm), minimum diastolic LV pressure remained low, and sinus rate accelerated after ventricular tachycardia. Hemodynamic recovery during ventricular tachycardia is mediated by the response of the autonomic nervous system and requires both alpha-adrenergic vasoconstriction and beta-adrenergic augmentation of contraction and relaxation.

Development of an automatic implanted drug infusion system for the management of cardiac arrhythmias

Conventional management of cardiac arrythmias relies on oral drug therapy which minimizes recurrence of the arrhythmia, but risks unpleasant side effects and even long-term toxicity. The authors propose acute management instead, from an implanted drug pump which automatically senses the onset of arrhythmia, delivers a pharmacokinetically-based infusion to terminate the episode, and discontinues drug delivery until the next occurrence. A bedside system consisting of a personal computer and conventional intravenous pump has been developed and tested in five dogs and 24 patients during a catheter electrophysiologic study. After detection of the arrhythmia plasma levels of the antiarrhythmia drug rose immediately to the therapeutic range and were subsequently well-controlled for 30 to 60 minutes. In all five dogs and in seven of the eight patients in whom atrial fibrillation was induced during the study, conversion to normal rhythm occurred within fifteen minutes. >

The effect of drugs and lead maturation on atrial electrograms during sinus rhythm and atrial fibrillation

Antitachycardia devices need more accurate means to identify arrhythmias. Previous studies have found that sinus rhythm can be distinguished from a variety of tachyarrhythmias by algorithms that are based on time-domain and frequency-domain analysis of intracardiac electrograms. Amplitude distribution analysis (time-domain) and power density spectral analysis (frequency-domain) are two of the techniques that have seemed to hold promise. However, previous studies have not evaluated whether lead maturation or drugs such as lidocaine, propranolol, verapamil, or isoproterenol can interfere with the ability of these algorithms to distinguish among cardiac rhythms. In the present study, five dogs had permanent atrial pacing leads placed. On a series of days, recordings were made from the atrial leads during sinus rhythm and induced sustained atrial fibrillation, both before and after administration of cardioactive drugs. For up to 1 month after implantation, progressive lead maturation did not prevent differentiation of atrial fibrillation from sinus rhythm by either amplitude distribution analysis or power density spectral analysis. However, the difference between the power density spectra of sinus rhythm and atrial fibrillation became progressively less with time. Isoproterenol, lidocaine, verapamil, and propranolol had no consistent effects on amplitude distribution analysis of atrial electrograms during sinus rhythm or atrial fibrillation. However, there were marked effects of drugs on amplitude distribution characteristics in individual dogs. Propranolol and lidocaine produced consistent changes in power density spectra during sinus rhythm and atrial fibrillation, respectively; both drugs reduced the ability of power density spectral analysis to differentiate sinus rhythm from atrial fibrillation.(ABSTRACT TRUNCATED AT 250 WORDS)

PROBABILITY DENSITY FUNCTION REVISITED : IMPROVED DISCRIMINATION OF VF USING A CYCLE LENGTH CORRECTED PDF

The probability density function (PDF) describes the fraction of time an electrogram signal spends at the baseline, In normal rhythm the signal is at baseline during the period between electrogram complexes, while in fibrillation the signal exhibits continuous activity and spends little time at baseline. However, time spent at the baseline is dependent on the rate of the rhythm, which limits the ability of the PDF algorithm to discriminate ventricular fibrillation from fast nonfibrillatory rhythms. A cycle length corrected version of the PDF algorithm has been formulated, which only examines the electrical activity between detected beats. The algorithm was developed utilizing a training set of 77 endocardial recordings and tested utilizing a test set of 90 endocardial and 56 epicardial recordings. Ventricular fibrillation was detected with 100% sensitivity and 98% specificity.

Retroperitoneal placement of an ICD generator: a solution for a difficult problem.

A woman experienced multiple episodes of pocket erosion after placement of an implantable cardioverter defibrillator culminating in a pocket infection. The device was placed in a retroperitoneal location, and the patient has not developed further episodes of pocket erosion or infection. Interrogation of the device in this position and its subsequent replacement have been straightforward.

Computer analysis of right ventricular pressure for improved discrimination of ventricular tachyarrhythmias

A computer algorithm suitable for an implantable device is devised that incorporates analysis of both right ventricular electrogram and pressure to assess the hemodynamic stability of a ventricular tachycardia. The algorithm was tested on recordings from 5 patients during pretachycardia control (N=5), stable (N=9) and unstable tachycardias (N=6) induced during electrophysiological studies and after tachycardia termination (N=10). Computer measurements of beat-by-beat and 5-beat averages of cycle length and hemodynamic parameters (pulse pressure, maximum positive dP/dt, and maximum negative dP/dt) at 5, 10, 15, and 30-second intervals after tachycardia induction and termination agreed to within 5% of manual measurements. Right ventricular pulse pressure, when normalized to pretachycardia control and expressed as a ratio, could perfectly distinguish hemodynamically stable from unstable ventricular tachycardia, both within individual patients and in the group, in as early as five seconds after tachycardia induction.<<ETX>>

Adaptive and sampled-data control of blood levels of antiarrhythmic drugs

Open-loop infusion techniques can be improved by a computer-controlled drug-infusion pump using a pharmacokinetic model for estimation of the blood level. Unfortunately, in individual patients the pharmacokinetic behavior of most drugs varies considerably from population averages. The authors tested the use of occasional blood samples in direct (closed-loop) and indirect (adaptive) feedback schemes for delivery of the antiarrhythmic drugs procainamide and disopyramide using computer simulation. A third scheme tested was a combination of these two (closed-loop adaptive). Pharmacokinetic parameters of 12 patients were used in computer simulations of feedback control. Using an open-loop infusion, the average mean-square error of the plasma level was 22.8% of the target concentration. With feedback control, this error was reduced to 5.1% or better. It is concluded that systems in which feedback of actual plasma levels can occur within three minutes may greatly improve control of drug infusion.<<ETX>>

Detection And Distinction Of Atrial Arrhythmias From Bipolar Electrograms

Several implantable devices are now available which attempt control of atrial tachyarrhythmias through pacing. If detection of tachycardia is inaccurate in these devices, attempts at electrical termination may be made inappropriately or withheld altogether. We have studied two computer algorithms for the analysis of atrial electrograms. They are based on rate and rcgularity of rate of the atrial electrogram, which are measures used in current antitachycardia devices. We have applied these algorithms to atrial elcctrograms from patients during sinus rhythm, supraventricular tachycardia, atrial flutter and atrial fibrillation. A rate limit was able to distinguish successfully each of the tachyarrhythmias from sinus rhythm but not from each other. The regularity algorithm failed to discriminate between sinus rhythm and the three tachyarrhythmias. Accurate diagnosis of atrial tachyanhythmias may require the application of more than one algorithm on both atrial and ventricular electrograms.

Implantable microprocessor-based devices for the management of arrhythmia

Microprocessors of small size and low power are being used in implantable devices for control of cardiac rhythm. The authors have developed two applications of this technology: (1) signal processing algorithms for accurate arrhythmia identification in an antitachycardia pacemaker or defibrillator and (2) control algorithms for pharmacokinetically based infusion of antiarrhythmic drugs from an implanted pump. For tachycardia identification they have used rate and regularity, amplitude-distribution histograms, Fourier spectra, and autocorrelation analysis of the atrial and ventricular electrograms. For automatic antiarrhythmic drug infusion, they have used procainamide and disopyramide in patients and animals and have studied the efficacy of these drugs for acute management of paroxysmal atrial fibrillation. In one dog they implanted a complete system which recognizes atrial fibrillation and automatically dispenses the drug from a subcutaneous reservoir.<<ETX>>