The use of symptoms generated by head up tilt (HUT) is not a useful tool in identifying chronic fatigue syndrome (CFS). We investigated whether heart rate variability (HRV) assessed early during HUT might be useful. A sample of 46 female subjects (24 with CFS and 22 sedentary, age-matched healthy controls; CON) who had exhibited no difference in time to syncope during tilt was examined for HRV responses to 10 min of 70 degrees HUT after 5 min of baseline in the supine position. HRV data were analyzed by the method of coarse graining spectral analysis. Variables compared between groups included mean and standard deviation (SD(RRI)) of RR intervals (RRI), amplitudes of low- (A(LF); 0.04-0.15 Hz) and high-frequency (A(HF); >0.15 Hz) harmonic as well as aperiodic, fractal (A(FR); 1/f(beta)) spectral components, the spectral exponent beta, and the difference in these values between baseline and HUT for each subject. In the supine baseline, only mean RRI was significantly (P < 0.01) lower in CFS than in CON. During HUT, however, mean RRI (P < 0.01), SD(RRI) (P < 0.01), A(HF) (P < 0.05), and A(FR) (P < 0.01) were significantly lower in CFS than in CON. When the difference in values between baseline and HUT for each subject was examined, only the difference for A(FR) (deltaA(FR)) was significantly (P < 0.01) lower in CFS than in CON, suggesting that A(FR)is a disease-specific response of HRV to HUT. When a cut-off level was set to deltaA(FR) = -2.7 msec, the sensitivity and the specificity in differentiating CFS from controls were 90% and 72%, respectively. The data suggest that a decrease in aperiodic fractal component of HRV in response to HUT can be used to differentiate patients with CFS from CON.