In a cross-sectional study, Kannus and colleagues [1] found that differences in bone mineral content between playing and nonplaying arms were significantly larger in women who played tennis and squash at a national level than in controls. They also found that such differences correlated negatively with the "biological age" at which these women began playing and concluded that "bones of the playing extremity clearly benefit from active tennis and squash training" and that the benefit is greater if training starts at or before menarche.

The authors did not comment on two unexpected findings. First, upper-arm and forearm circumference and muscle strength measurements were not smaller in controls (who were recruited from local schools and workplaces) than in national-level players. We ask whether exercise increases bone mineral content without increasing muscle bulk and strength. The authors indicated that none of the controls was involved in intense physical training or activities or work that would affect the dominant or nondominant arm only. If controls were involved in some activities that involved both arms, a comparison of the differences between the two arms of the players and those of controls may not be appropriate.

Second, if exercise causes the difference in bone mineral content between the two arms, a greater amount of exercise would be expected to result in a larger difference. Years of training, however, did not correlate significantly with differences in bone mineral content seen between the dominant and nondominant arms of the players. Does this finding indicate that the maximum benefit from this kind of training is achieved within 5 years (the minimum training period of the study group)?