Study design: The study is conducted on the basis of comparative-repeated measures. Objectives: The objective of this study is to assess the reliability of non-invasive cardiac output (CO) measurements in individuals with tetraplegia (TP) at rest and during exercise using Innocor, and to test the hypothesis that CO measurements are less reliable in TP than in able-bodied (AB) individuals. Setting: Ambulatory volunteers, Switzerland. Methods: Nine male motor-complete TP (C5-C7) and nine pair-matched AB performed repeated CO measurements at rest and during submaximal arm-crank and wheelchair exercises in four different test sessions. Within-and between-day reliabilities were compared between TP and AB. Results: Mean differences between measurements at rest (TP vs AB, within-day: 0.1±0.5 vs 0.2 ± 0.6 l min À1 , between-day: À0.7 ± 0.6 vs À0.1 ± 0.8 l min À1 ), during arm-crank (TP vs AB, withinday: 0.1±0.9 vs 0.5±0.7 l min À1 , between-day: À0.3±1.1 vs 0.0±1.1 l min À1 ) and wheelchair exercises (TP vs AB, within-day: 0.3 ± 1.2 vs -0.1 ± 0.8 l min À1 , between-day: 0.1 ± 1.1 vs 0.5±0.9 l min À1 ) were not significantly different between TP and AB (all P40.05). Coefficients of variation in TP (within-day, rest: 6.8%, arm-crank: 9.6% and wheelchair: 10.8%; between-day, rest: 11.9%, arm-crank: 11.2% and wheelchair: 10.3%) and in AB (within-day, rest: 7.7%, arm crank: 6.8% and wheelchair: 6.0%; between-day, rest: 9.2%, arm crank: 8.5% and wheelchair: 8.0%) indicated acceptable reliability. Conclusion: In contrast to our hypothesis, we found non-invasive CO measurements using Innocor to be as reliable in TP as they are in AB. Consequently, Innocor can be recommended for repeated assessments of CO in TP within routine diagnostics or for evaluation of training progress.