Doubtful validity of the Yo-Yo intermittent recovery test.

Abstract

Dear Editor-in-Chief: A recent article in Medicine & Science in Sports & Exercise ® (3) aimed at investigating the reproducibility of the “Yo-Yo intermittent recovery test” in 17 male subjects. In addition, the validity of the test was evaluated in 10 of 12 elite soccer players with regard to its sensitivity for seasonal changes and its relationship to video-borne analyses of running distances during soccer matches. It is concluded that the Yo-Yo intermittent recovery test yields a high reproducibility and represents a valid means for the assessment of soccer-specific fitness. However, these statements raise several issues of concern. The reproducibility of the Yo-Yo test was assessed in 17 male subjects (no information given about current sport activity) with an average endurance capacity. The test-retest coefficient of correlation was relatively high: r = 0.98. It remains questionable if these results can be easily transferred to elite soccer players. With regard to validity, the conclusion that “test performance was closely related to match performance in soccer” is not justified. The significant correlation between Yo-Yo test performance and the amount of high-intensity running during soccer matches (r = 0.71; Fig. 7) seems questionable and is mainly due to the surprisingly heterogeneous test results of the subject population. The Yo-Yo test result of the worst-performing player (1160 m) was located more than 20% below the second worst one and is rarely acceptable in an elite soccer player. Moreover, this player, as well as four other subjects, was considered more than once in the correlational analysis, which renders the observations dependent from each other and represents statistical flaw. We recalculated the correlation from the manuscript’s figure, and without the results of this player, no significant linear relationship remained. When additionally eliminating the second outlier, who performed about 20% better than his second-best counterpart and was taken for the calculations with no less than three test performances, the remaining data would have the shape of a heap of points without any tendency at all. The variability of the 10 players being left would be in accordance with interindividual differences that can be expected in elite soccer players. The authors confirm a high sensitivity of the Yo-Yo test for performance changes. However, an increase in performance of 25% over the duration of a few weeks (corresponding to the preparation period) is dubious in competitive athletes and exceeds what is commonly reported in the literature (1,2,4). Such a change is larger than endurance gains in untrained subjects after comparable training periods. Injury-related training breaks might be the reason for such unexpectedly large performance changes during the season (Fig. 5A). However, to evaluate the test sensitivity for realistic performance changes, it is mandatory to test players longitudinally under circumstances of continuous training. From our point of view, particularly the validity (i.e., the sensitivity for relevant changes in performance and the relationship between test results and high-intensity running) of the Yo-Yo intermittent recovery test has to be critically questioned. A reevaluation under more standardized conditions would be worthwhile. Wilfried Kindermann Tim Meyer