Cardiopulmonary and Noninvasive Hemodynamic Responses to Exercise Predict Outcomes in Heart Failure JONATHAN MYERS, PhD, MYOWONG, BA, CHANDANA ADHIKARLA, MD, MADHAVI BOGA, MD, SRIDEVI CHALLA, MD, JOSHUA ABELLA, MD, AND EUAN A. ASHLEY, MRCP, DPhil Palo Alto, California ABSTRACT Background: An impaired cardiac output response to exercise is a hallmark of chronic heart failure (HF). We determined the extent to which noninvasive estimates of cardiac hemodynamics during exercise in combination with cardiopulmonary exercise test (CPX) responses improved the estimation of risk for ad- verse events in patients with HF. Methods and Results: CPX and impedance cardiography were performed in 639 consecutive patients (mean age 48 6 14 years), evaluated for HF. Clinical, hemodynamic, and CPX variables were acquired at baseline and subjects were followed for a mean of 460 6 332 days. Patients were followed for the com- posite outcome of cardiac-related death, hospitalization for worsening HF, cardiac transplantation, and left ventricular assist device implantation. Cox proportional hazards analyses including clinical, noninvasive hemodynamic, and CPX variables were performed to determine their association with the composite end- point. There were 113 events. Among CPX variables, peak oxygen uptake (VO 2 ) and the minute ventila- tion (VE)/carbon dioxide production (VCO 2 ) slope were significant predictors of risk for adverse events (age-adjusted hazard ratio [HR] 1.08, 95% confidence interval [CI] 1.05e1.11 for both; P ! .001). Among hemodynamic variables, peak cardiac index was the strongest predictor of risk (HR 1.08, 95% CI 1.0e1.16; P 5 .01). In a multivariate analysis including CPX and noninvasively determined hemody- namic variables, the most powerful predictive model included the combination of peak VO 2 , peak cardiac index, and the VE/VCO 2 slope, with each contributing significantly and independently to predicting risk; an abnormal response for all 3 yielded an HR of 5.1 (P ! .001). Conclusions: These findings suggest that noninvasive indices of cardiac hemodynamics complement established CPX measures in quantifying risk in patients with HF. (J Cardiac Fail 2013;19:101e107) Key Words: Exercise testing, oxygen uptake, cardiac output, mortality. Chronic heart failure (HF) is characterized by the inabil- ity of cardiac output to meet the metabolic demands of the tissues, which is particularly evident during physical activ- ity. 1 The cardiopulmonary exercise test (CPX) response to exercise has therefore become well established as a tool to help predict outcomes in patients with HF. This is due in part to the fact that oxygen uptake (VO 2 ) derived from CPX is considered to be a surrogate for cardiac output, given the Fick equation (VO 2 5 cardiac output  arteriovenous oxygen difference). In recent years, useful variables derived from CPX have been extended beyond peak VO 2 to include indices of ventilatory inefficiency, including the minute ventilation (VE)/carbon dioxide production (VCO 2 ) slope, the oxygen uptake efficiency slope (OUES), and os- cillatory breathing patterns. 2 These indices of ventilatory in- efficiency have been demonstrated to provide both superior and complementary information to peak VO 2 in stratifying risk in HF. 2e4 The prognostic utility of these markers is at- tributable to the fact that they are also related in part to the degree of impairment in the cardiac output response to exercise. 2e4 Ideally, cardiac output during exercise would be deter- mined directly to gauge the severity of HF; however, direct measurement of cardiac output is a difficult undertaking. Determining cardiac output directly requires invasive mea- surements, is time consuming, is expensive, carries an in- herent risk, and is difficult to measure accurately during From the Veterans Administration Palo Alto Health Care System, Division of Cardiovascular Medicine, Stanford University, Palo Alto, Cal- ifornia. Manuscript received September 13, 2012; revised manuscript received November 19, 2012; revised manuscript accepted November 27, 2012. Reprint requests: Jonathan Myers, PhD, VA Palo Alto Health Care Sys- tem, Cardiology 111C, 3801 Miranda Ave, Palo Alto, CA 94304. Tel: 650- 493-5000 x64661; Fax: 650-852-3473. E-mail: drj993@aol.com See page 106 for disclosure information. 1071-9164/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.cardfail.2012.11.010 101 Journal of Cardiac Failure Vol. 19 No. 2 2013