Clinically useful estimates of ˙VO2max from treadmill tests(GXT) may be made using protocol-specific equations. In many cases, GXT may proceed more effectively if the clinician is free to adjust speed and grade independent of a specific protocol. We sought to determine whether˙VO2max could be predicted from the estimated steady-state˙VO2 of the terminal exercise stage. Seventy clinically stable individuals performed GXT with direct measurement of ˙VO2. Exercise was incremented each minute to optimize clinical examination. Measured˙VO2max was compared to the estimated steady-state ˙VO2 of the terminal stage based on ACSM equations. Equations for walking or running were used based on the patient's observed method of ambulation. The measured ˙VO2max was always less than the ACSM estimate, with a regular relationship between measured and estimated ˙VO2max. No handrail support: ˙VO2max = 0.869·ACSM - 0.07; R2 = 0.955, SEE = 4.8 ml·min-1·kg-1 (N = 30). With handrail support: ˙VO2max = 0.694·ACSM + 3.33; R2 = 0.833, SEE = 4.4 ml·min-1·kg-1(N = 40). The equations were cross-validated with 20 patients. The correlation between predicted and observed values was r = 0.98 and 0.97 without and with handrail support, respectively. The mean absolute prediction error (3.1 and 4.1 ml·min-1·kg-1) were similar to protocol-specific equations. We conclude that ˙VO2max can be predicted independent of treadmill protocol with approximately the same error as protocol-specific equations.