Original Scientific Paper Characterizing differences in mortality at the low end of the fitness spectrum in individuals with cardiovascular disease Sandra Mandic a,b,c , Jonathan Myers a,b , Ricardo B. Oliveira a,b , Joshua Abella a and Victor F. Froelicher a,b a Veterans Affairs Palo Alto Health Care System, b Stanford School of Medicine, Palo Alto, California, USA and c School of Physical Education, University of Otago, Dunedin, New Zealand Received 18 January 2009 Accepted 14 July 2009 Background A graded but nonlinear relationship exists between fitness and mortality, with the greatest mortality differences occurring between the least-fit (first, Q1) and the next-least-fit (second, Q2) quintiles of fitness. The purpose of this study was to compare clinical characteristics, exercise test responses, and physical activity (PA) patterns in Q1 versus Q2 in patients with cardiovascular disease (CVD). Design Observational retrospective study. Methods A total of 5101 patients with a history of CVD underwent clinical treadmill testing and were followed up for 9.1 ± 5.5 years. Patients were classified into quintiles of exercise capacity measured in metabolic equivalents. Clinical characteristics, treadmill test results, and recreational PA patterns were compared between Q1 (n = 923) and Q2 (n = 929). Results Q1 had a nearly two-fold increase in age-adjusted relative risk of cardiovascular mortality compared with Q2 (hazard ratio: 3.79 vs. 2.04, P < 0.05; reference: fittest quintile). Q1 patients were older, had more extensive use of medications, and were more likely to have a history of typical angina (35 vs. 28%), myocardial infarction (30 vs. 24%), chronic heart failure (25 vs. 14%), claudication (15 vs. 9%) and stroke (9 vs. 6%) compared with Q2 (all comparisons: P < 0.05). Recent and lifetime recreational PA was not different between the two groups. Conclusion Greater severity of disease in the least-fit versus the next-least-fit quintile likely contributes to but cannot fully explain marked differences in mortality rates in CVD patients. To achieve potential survival benefits, our results suggest that unfit CVD patients should engage in exercise programs of sufficient volume and intensity to improve fitness. Eur J Cardiovasc Prev Rehabil 00:000–000  c 2010 The European Society of Cardiology European Journal of Cardiovascular Prevention & Rehabilitation 2010, 00:000–000 Keywords: cardiovascular disease, fitness, mortality, physical activity Introduction An inverse, graded, and dose-dependent relationship between physical fitness and mortality has been consis- tently reported in epidemiological studies, irrespective of the presence or absence of cardiovascular disease (CVD) [1–4]. This relationship is typically nonlinear, with the greatest decline in risk of mortality observed between the least-fit (first, Q1) and the next-least-fit (second, Q2) groups of individuals with and without CVD [1,3–6]. However, factors that may explain the steep mortality gradient in the lowest end of the fitness spectrum remain largely unknown. We recently observed that lower recent physical activity (PA) rather than differences in clinical characteristics, in part, explained the striking difference in mortality rates between Q1 and Q2 in apparently healthy individuals [7]. It remains unknown whether differences in clinical characteristics reflecting differences in severity of underlying disease, PA patterns, behavior, environmental, or other factors (e.g. genetics) explain the steep mortality gradient between Q1 and Q2 in patients with CVD. Correspondence to Sandra Mandic, PhD, Senior Lecturer, Exercise Physiology, School of Physical Education, University of Otago, PO Box 56 Dunedin, New Zealand Tel: +64 3 479 5415; fax: +64 3 479 8309; e-mail: sandra.mandic@otago.ac.nz Part of this work was presented as abstracts at the European Congress of Cardiology 2008 (Munich, Germany) and American Heart Association Annual Meeting 2008 (New Orleans, USA). 1741-8267  c 2010 The European Society of Cardiology DOI: 10.1097/HJR.0b013e32833163e2 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.