Maximal exercise test results in subacute stroke

Objectives: To establish the feasibility and reliability of graded maximal exercise testing, suitability of standard indexes of maximum oxygen consumption ((V)over dot O(2)max), and evidence of trial-to-trial practice effects in subacute stroke. Design: Descriptive, cross-sectional study. Setting: Rehabilitation hospital. Participants: Consecutive sample of 35 participants (mean age, 65.7 +/- 2.6y; mean days poststroke, 17.6 +/- 2.2d). Interventions: Not applicable. Main Outcome Measures: Graded maximal exercise test using semirecumbent cycle ergometry. A subset (n=20) performed repeated tests to determine test-retest reliability and presence of practice effects. Results: Thirty-one (89%) participants completed the exercise test without symptomatic responses (mean peak level of oxygen consumed [VO(2)peak], 10.7 mL.kg(-1)-min(-1); peak work rate, 39W). Twelve (34%), 15 (44%), and 3 (9%) participants achieved oxygen consumption per unit time ((V)over dot O-2) plateau, respiratory exchange ratio, and heart rate criteria, respectively. Although test-retest correlations were high (intraclass correlation coefficient range,.67-.87), higher VO(2)peak (1.0 mL.kg(-1).min(-1), P=.04) and work rate (7.3W, P=.01) were observed with repeated testing, with no associated increase in reaching standard criteria for (V)over dot O(2)max. Conclusions: This work has important implications for developing guidelines for measuring aerobic capacity early after stroke. Although maximal exercise testing using semirecumbent cycle ergometry is feasible, standard (V)over dot O(2)max criteria are not consistently appropriate. At least I practice trial is recommended before the actual evaluation is performed.