Autonomic Neuroscience: Basic and Clinical 83 (2000) 90–99 www.elsevier.com / locate / autneu Three-dimensional return map: a new tool for quantification of heart rate variability a b a a c Ruy S. Moraes , Elton L. Ferlin , Carisi A. Polanczyk , Luis E.P. Rohde , Lerida Zaslavski , c,d a,d, * Jorge L. Gross , Jorge P. Ribeiro a ´ Cardiology Division, Hospital de Clınicas de Porto Alegre, Porto Alegre 90035-003, Brazil b ´ Biomedical Engineering Division, Hospital de Clınicas de Porto Alegre, Porto Alegre 90035-003, Brazil c ´ Endocrinology Division, Hospital de Clınicas de Porto Alegre, Porto Alegre 90035-003, Brazil d Department of Medicine, Faculty of Medicine, Federal University of Rio Grande do Sul, 90035-003 Porto Alegre, Brazil Received 25 August 1999; received in revised form 29 December 1999; accepted 14 February 2000 Abstract Background: Several methods are used to study heart rate variability, but they have limitations, which might be overcome by the use of a three-dimensional return map. Objectives: To evaluate the performance of three-dimensional return map-derived indices to detect (1) sympathetic and parasympathetic modulation to the sinus node and (2) autonomic dysfunction in diabetic patients. Methods: Six healthy subjects underwent partial and total pharmacological autonomic blockade in a protocol that incorporated vagal and sympathetic predominance. Twenty-two patients with type 2 diabetes mellitus and 12 normal controls participated in the subsequent validation experiment. Three-dimensional return maps were constructed by plotting RR intervals versus the difference between adjacent RR n intervals [(RR )2(RR )] versus the number of counts, and four derived indices (P , P , P , MN ) were created for quantification. n 11 n 1 2 3 Results: Both indices P and MN were significantly increased after sympathetic blockade with propranolol, while all indices except P 1 1 were modified after parasympathetic blockade ( P , 0.05). During the validation experiments, P and MN detected differences between 1 normal controls, and diabetic patients with and without autonomic neuropathy. The overall accuracy of most three-dimensional indices to detect autonomic dysfunction, estimated by the area under the ROC curve, was significantly better than traditional time domain indices. Three-dimensional return map-derived indices also showed adequate reproducibility on two different recording days (intra-class correlation coefficients of 0.69 to 0.82; P , 0.001). Conclusions: Three-dimensional return map-derived indices are reproducible, quantify parasympathetic as well as sympathetic modulation to the sinus node, and are capable of detecting autonomic dysfunction in diabetic patients.  2000 Elsevier Science B.V. All rights reserved. Keywords: Pharmacological autonomic blockade; Time domain indices of heart rate variability; Frequency domain indices of heart rate variability; ´ Poincare plots; Diabetic autonomic neuropathy 1. Introduction plotting each RR interval (RR ) versus its successive n interval (RR ), have been proposed as an alternative n 11 Several methods have been proposed to study beat-to- methodology to evaluate heart rate variability (HRV). This beat fluctuations in heart rate in the time and frequency method was able to detect different HRV patterns in heart domains. Most of these indices are able to detect para- failure patients (Woo et al., 1992) and abnormal beat-to- sympathetic modulation to the sinus node, but we have beat heart rate dynamics preceding episodes of ventricular recently demonstrated that they may not appropriately tachyarrhythmias (Huikuri et al., 1996). Although return detect sympathetic modulation (Polanczyk et al., 1998). maps might display heart rate variability in an elucidative ´ Recently, two-dimensional return maps or Poincare plots, way, long periods of analysis result in superimposition of points that may hide potentially important data in the core of the distribution. Moreover, most of the information derived from this methodology has been solely based on the general morphology of the distribution and few at- ´ *Corresponding author. Cardiology Division, Hospital de Clınicas de tempts have been made to create indices of quantification Porto Alegre, Rua Ramiro Barcelos 2350, 90035-003 Porto Alegre, RS, (Hnatkova et al., 1995; Copie et al., 1996b; Huikuri et al., Brazil. Tel.: 155-51-316-8344; fax: 155-51-316-8657. E-mail address: jpribeiro@cpovo.net (J.P. Ribeiro) 1996). 0165-1838 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0165-1838(00)00095-3