ORIGINAL ARTICLE Assessment of cardiovascular reactivity by fractal and recurrence quantification analysis of heart rate and pulse transit time JE Naschitz 1 , R Itzhak 2 , N Shaviv 1 , I Khorshidi 1 , S Sundick 1 , H Isseroff 1 , M Fields 1 , RM Priselac 1 , D Yeshurun 1 and E Sabo 1 1 Department of Internal Medicine A, Bnai-Zion Medical Center and Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, Haifa, Israel; 2 Rheumatology, Bnai-Zion Medical Center and Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, Haifa, Israel Methods used for the assessment of cardiovascular reactivity are flawed by nonlinear dynamics of the cardiovascular responses to stimuli. In an attempt to address this issue, we utilized a short postural chal- lenge, recorded beat-to-beat heart rate (HR) and pulse transit time (PTT), assessed the data by fractal and recurrence quantification analysis, and processed the obtained variables by multivariate statistics. A 10-min supine phase of the head-up tilt test was followed by recording 600 cardiac cycles on tilt, that is, 5–10 min. Three groups of patients were studied, each including 20 subjects matched for age and genderF healthy subjects, patients with essential hypertension (HT), and patients with chronic fatigue syndrome (CFS). The latter group was studied on account of the well-known dysautonomia of CFS patients, which served as contrast against the cardiovascular reactivity of the healthy population. A total of 52 variables of the HR and PTT were determined in each subject. The multivariate model identified the best predictors for the assessment of reactivity of healthy subjects vs CFS. Based on these predictors, the ‘Fractal & Recurrence Analysis-based Score’ (FRAS) was calculated: FRAS ¼ 76.2+0.04 * HR- supine-DET 12.9 * HR-tilt-R/L 0.31 * HR-tilt-s.d. 19.27 * PTT-tilt-R/L 9.42 * PTT-tilt-WAVE. The median values and IQR of FRAS in the groups were: healthy ¼1.85 (IQR 1.89), hypertensives ¼ +0.52 (IQR 5.78), and CFS ¼24.2 (5.34) (HT vs healthy subjects: P ¼ 0.0036; HT vs CFS: Po0.0001). Since the FRAS differed significantly between the three groups, it appears likely that the FRAS may recognize phenotypes of cardiovascular reactivity. Journal of Human Hypertension (2003) 17, 111–118. doi:10.1038/sj.jhh.1001517 Keywords: arterial hypertension; cardiovascular reactivity; tilt test; fractal analysis; recurrence quantitative analysis Introduction Cardiovascular reactivity is defined as the change in blood pressure (BP), heart rate (HR), or other haemodynamic parameters in response to physical or mental stimuli. 1 Various methods are employed for the study of cardiovascular reactivity. At 24 h ambulatory monitoring measures short- and long- term variabilities in BP and HR. In the laboratory, the cardiovascular reactivity is studied under psy- chical, cold pressor, or postural challenge, lower- body negative pressure, physical exercise, and combined mental and physical challenges. 2–4 The aim of these efforts is to measure BP and HR changes under various conditions, and compare them to baseline values. For this purpose, attention has to be paid not only to average values of BP and HR but also to fluctuations of the measurements around the average levels. The standard deviation supplies data on the signal dispersion around the mean, but does not provide information on the patterns that char- acterize the variability of the signal over a period of time. This has led to the development of other methods for the evaluation of cardiovascular varia- bility, among which spectral analysis has been widely used. 4 Spectral analysis allows the overall variance of the signal to be split into its various frequency components. The claimed specificity of low-frequency powers as markers of sympathetic tone and of high-frequency powers as markers of vagal tone is highly debated. 5 In fact, autonomic cardiovascular modulation is characterized by a high degree of nonlinearity between external stimuli and cardiovascular response. 4 Other methods have been proposed in an attempt to address the issue of nonlinear dynamics of the cardiovascular responses. Received 24 July 2002; revised 25 October 2002; accepted 31 October 2002 Correspondence: Dr JE Naschitz, Department of Internal Medicine A, Bnai-Zion Medical Center, Haifa 31048, PO Box 4940, Haifa 31048, Israel. E-mail: Naschitz@tx.technion.ac.il Journal of Human Hypertension (2003) 17, 111–118 & 2003 Nature Publishing Group All rights reserved 0950-9240/03 $25.00 www.nature.com/jhh