BACKGROUND: Most trials on the reliability of constant-load treadmill testing use one pair of treadmill settings (speed, grade) only. The question of whether the results can be applied to tests with different settings is left open. Also, claudication distances measured with differing settings are not comparable, rendering the comparison of the results from different trials difficult. This study evaluates the reliability of constant-load testing with various workloads and compares them with claudication distances achieved with walking at normal speed on level ground but also evaluates whether metabolic equivalent (MET) normalization can be used to translate the results of different treadmill tests into each other. METHODS: Fifteen patients with claudication underwent repeated treadmill testing with different treadmill settings, including speeds of 2.0, 3.2, and 4.0 km/h (1.25, 2.0, and 2.5 mph, respectively) and grades of 0% and 12%. The walking capacity was also tested on level ground with a speed chosen by the individual patient. Results of virtual treadmill tests with all possible combinations of the speeds and the grades used were predicted from real tests, with MET normalization. The relationship between real and predicted claudication distances was tested with regression modeling. RESULTS: Reliability coefficients (RCs) for the absolute claudication distance (ACD) were superior to RCs for the initial claudication distance. RCs for ACD ranged between 0.61 and 0.95, with increasing values found with increasing workloads. The best coefficients for the regression of measured on predicted claudication distances were achieved with a model on the basis of a power function (r = 0.963). The model was only appropriate for the prediction of group mean results from clinical trials but cannot be applied to single patient data. For proof of concept, the model was tested with six published studies in which the claudication distances of a patient sample were double treadmill tested with different workloads. The result of the second test was predicted from the first test, and estimated and measured claudication distances were compared. The mean difference (all trials) was 7.9%, whereas the maximum difference amounted to 16.5%. CONCLUSION: For an optimal treadmill test reliability, higher workloads should be used and ACD should be preferred over initial claudication distance. MET normalization provides the basis for the comparability of treadmill test results achieved with different test conditions.
BACKGROUND: Most trials on the reliability of constant-load treadmill testing use one pair of treadmill settings (speed, grade) only. The question of whether the results can be applied to tests with different settings is left open. Also, claudication distances measured with differing settings are not comparable, rendering the comparison of the results from different trials difficult. This study evaluates the reliability of constant-load testing with various workloads and compares them with claudication distances achieved with walking at normal speed on level ground but also evaluates whether metabolic equivalent (MET) normalization can be used to translate the results of different treadmill tests into each other. METHODS: Fifteen patients with claudication underwent repeated treadmill testing with different treadmill settings, including speeds of 2.0, 3.2, and 4.0 km/h (1.25, 2.0, and 2.5 mph, respectively) and grades of 0% and 12%. The walking capacity was also tested on level ground with a speed chosen by the individual patient. Results of virtual treadmill tests with all possible combinations of the speeds and the grades used were predicted from real tests, with MET normalization. The relationship between real and predicted claudication distances was tested with regression modeling. RESULTS: Reliability coefficients (RCs) for the absolute claudication distance (ACD) were superior to RCs for the initial claudication distance. RCs for ACD ranged between 0.61 and 0.95, with increasing values found with increasing workloads. The best coefficients for the regression of measured on predicted claudication distances were achieved with a model on the basis of a power function (r = 0.963). The model was only appropriate for the prediction of group mean results from clinical trials but cannot be applied to single patient data. For proof of concept, the model was tested with six published studies in which the claudication distances of a patient sample were double treadmill tested with different workloads. The result of the second test was predicted from the first test, and estimated and measured claudication distances were compared. The mean difference (all trials) was 7.9%, whereas the maximum difference amounted to 16.5%. CONCLUSION: For an optimal treadmill test reliability, higher workloads should be used and ACD should be preferred over initial claudication distance. MET normalization provides the basis for the comparability of treadmill test results achieved with different test conditions.
Authors: Lotte M Kruidenier; Saskia P A Nicolaï; Edith M Willigendael; Rob A de Bie; Martin H Prins; Joep A W Teijink Journal: BMC Cardiovasc Disord Date: 2009-03-02 Impact factor: 2.298
Authors: Marie Gernigon; Alexis Le Faucheur; Dominique Fradin; Bénédicte Noury-Desvaux; Cédric Landron; Guillaume Mahe; Pierre Abraham Journal: Medicine (Baltimore) Date: 2015-05 Impact factor: 1.889