The purpose of this study was to assess the validity of the 1 min stage version of the 20 m shuttle run multistage test to predict VO2max in adults (53 males and 24 females, 19 to 47 years old). The maximal shuttle run speed (S-MAS) was thus compared to the VO2max attained during a multistage treadmill test (TE-VO2max), the retroextrapolated VO2max at time zero of the O2 recovery curve of the shuttle run (SR-VO2max) and the VO2max predicted from a previously developed regression equation with children (Mercier et al., 1983) and putting 18 year olds in the regression equation for all adults which yielded (SP-VO2max = -27.4 + 6.0 S-MAS). The latter regression equation was very similar to the ones obtained with the two other criteria and was thus retained to ensure a smooth transition between children and adults in predicting VO2max. Correlations and standard errors of the estimate between S-MAS and TE-VO2max (r = 0.90 and Syx = 4.4) or SR-VO2max (r = 0.87 and Syx = 4.7) were quite good. TE-, SR- and SP-VO2max were also similar (mean +/- SD = 49.4 +/- 10.1, 48.8 +/- 9.3 and 47.1 +/- 8.3 mL.kg-1.min-1, respectively, p greater than 0.05). It was concluded that the 20 m shuttle run is a valid test to predict VO2max in adults.