Objective: To evaluate heart rate variability (HRV) by power spectral analy
sis of heart rate and its relationship to intracranial pressure (ICP), cere
bral perfusion pressure (CPP), and outcomes in children with acute traumati
c head injury.
Design: Prospective, case series.
Setting: Pediatric intensive care unit in a level II trauma center/children
's hospital.
Subjects: Fifteen critically ill children with documented acute traumatic b
rain injury and four control subjects.
Interventions: None.
Measurements and Main Results: The normalized total power from 0.04 to 0.15
Hz was used to quantify low-frequency HRV and from 0.15 to 0.40 Hz to quan
tity high-frequency HRV. The ratio of low- to high-frequency (LF/HF) power
was used as a measure of sympathetic modulation of heart rate. The power sp
ectral data from the 5-min samples were averaged over each hour of data col
lection, and an hourly LF/HF ratio was obtained based on a 60-min electroca
rdiogram collection (twelve 5-min segments). The daily mean LF/HF ratio was
calculated from the hourly LF/HF measurements. We found no linear correlat
ion between the LF/HF ratio and either ICP or CPP (p = NS). There was a sig
nificant decrease in the LF/HF ratio when the intracranial pressure was >30
mm Hg (p < .001) or the cerebral perfusion pressure was <40 mm Hg (p < .00
1). Children with a Glasgow Coma Scale score of 3-4 had a lower LF/HF ratio
compared with those who had a Glasgow Coma Scale score of 5-8 (p < .005).
Patients who progressed to brain death had a markedly lower LF/HF ratio (p
< .001), with a significant decrease after the first 4 hrs of hospitalizati
on. Patients with more favorable outcomes had significantly higher LF/HF ra
tios.
Conclusions: Our findings suggest that an ICP of >30 mm Hg or a CPP of <40
mm Hg may be associated with marked autonomic dysfunction and poor outcome.
We speculate that HRV power spectral analysis may be a useful adjunct in d
etermining the severity of neurologic insult and the prognosis for recovery
in children. The LF/HF ratio may be helpful not only in identifying those
patients who will progress to brain death but also in predicting which pati
ents will have favorable outcomes.