The aim of this study was to assess the validity of the multistage shuttle run test in predicting maximal oxygen uptake (VO2 max) in athletes trained in sports with different physical demands. Over a 14-day period, 10 male long-distance runners (continuous high-intensity exercise) and 10 male squash players (intermittent high-intensity exercise) performed, in random order, the multistage shuttle run test and a maximal treadmill protocol of increasing elevation which elicited VO2 max. Compared with direct measurement of VO2 max, the shuttle run test significantly underpredicted the VO2 max of the group of runners (n = 10; P < 0.01) and of the athletes as a whole (n = 20; P < 0.01). The correlation (r) between VO2 max determined by the shuttle run test and by the treadmill protocol was 0.61 (P < 0.05) for the squash players, 0.71 (P < 0.05) for the runners and 0.67 (P < 0.01) for the athletes as a whole. As the subjects represented athletes trained specifically in sports with different physical demands, this may explain why the correlations between the shuttle run test and direct measurement of VO2 max are less robust than those reported in previous studies. Furthermore, our results indicate that there are sport-specific differences when predicting VO2 max from the multistage shuttle run test.