Microprocessor-based long term cardiorespirography. I. Heart rate changes and apneic attacks.

Abstract

Cardiorespirography is a well-known method of continuous monitoring in neonatal intensive care. Apneic attacks, bradycardia and tachycardia are registered. In our experience we connected a cardiorespirography recorder to a microprocessor system. The processor consisted of a hardware part including a program (software) and a printer which provided printouts of alarm events. As alarm situations, which cause an alarm printout, we defined: 1. apneic episodes (duration 10, 20 or 30 seconds) 2. tachycardiac (beat-to-beat rate greater than 180/minute) 3. V-shaped and U-shaped bradycardia (beat-to-beat heart rate less than 80/min) and combinations. The reliability of the system of recognizing and classifying alarm situations was tested by comparing the alarm printouts with the simultaneously recorded cardiorespirograms. Fifty eight 12 hour records of 41 patients were evaluated. Six hundred alarm situations were counted. The alarm printouts were found in concordance with the cardiograms in all tachycardia alarms. Nearly all bradycardia (V-shaped, U-shaped bradycardia, combination of bradycardia and apnea) were correctly classified. A preset apnea duration of 10 seconds resulted in many false positive alarm printouts. With 20 second apnea time only few false positive alarms were seen, but nine apneic attacks were not recognized. Altogether 81.5% of alarm printouts were correct, 16.8% were false alarms, or V-shaped bradycardia were really U-bradycardia. Only 2% of all alarms were not recognized by the microprocessor system. We suggest to combine the microprocessor with a special alarm recorder, which is able to store beat-to-beat heart rate and respiration wave before alarm situations.