Microprocessor-based long term cardiorespirography. II. Status evaluation in term and premature newborns.

In 1965 URBACH et al. and RUDOLPH et al. [35, 39] described a loss of heart rate variability in severely ill neonates. In this study we investigated the correlation between instantaneous heart rate patterns and status diagnosis. We used a microprocessor-based cardiorespirography system. Seventy five newborn infants (51 prematures and 24 term neonates) were studied for about 12 hours each. Twenty nine patients had a second record after the first investigation. Parameters were: Type of frequency and oscillation, long time variability (LTV), short time variability (STV) and the newly introduced P-value (maximal difference between two successive R-peaks in five minutes). We found clear differences between the study groups. With increasing severity of illness mean values ("group mean values") of long time variability, short time variability and P-value decreased. Fixed heart rate became predominant. The most pronounced loss of heart rate variability was seen in infants with severe intracranial bleeding, thus offering a tentative diagnosis. For statistical analysis long time variability and the silent oscillation type have been proved as best parameters for this diagnosis. Severely decreased heart rate variations also have been seen in infants with acute renal failure--possibly because of brain edema--, after application of muscle relaxants, repeated doses of sedatives, and after prolonged anesthesia. Otherwise, the heart rate variability was probably dependent on age and gestational age in prematures and newborn infants without intracranial bleeding. It is possible to use microprocessor-based long time cardiorespirography as a simple screening method for the diagnosis of neonatal intracerebral bleeding. In future experiences transcutaneous measurements of oxygen tension should be included.