Although heart rate variability (HRV) decreasing has
been usually described in obstructive sleep apnea
syndrome (OSAS), some studies have recently questioned
the validity of spectral HRV analysis in presence of
respiratory and arrhythmic disorders. Fractal analysis of
HRV is an emerging nonlinear technique overcoming
these limitations and allowing short term HRV
assessment during hypo/apnea phases. The aim of this
study is to analyse the Fractal features in sleep apnea in
order to find as these characteristics could change during
abnormal respiration patterns in OSAS. We studied 30
polysomnographic recordings of severe OSAS (AHl30)
pts. (age 55±9) and 10 PR of normal subjects (age 46±4).
Hypo/apnea phases and related beat-to-beat time series
have been detected and classified by automated
algorithms and manually verified by expert technicians.
Fractal analysis was performed by the Higuchi algorithm
(FD). Results showed that while FD does not significantly
differ between Normals (l.6J±0.09) and normal breath
epochs in OSAS, it significantly (p<0.005) tends to a less
fractal structure from normal breath (l.60±0.15) to
hypopneas (l.52±0.13), obstructive (l.50±0.12) and
mixed apneas (l.48±0.1l) epochs, with a significant
Dunn's multiple comparisons post test only between
normal breath vs. obstructive and mixed apneas.
