Objectives: This study tested the hypothesis that edema during hypothermic cardioplegia is caused by the hypotonicity of the perfusate at cold temperatures. Methods: The volume of isolated human and rabbit atrial myocytes was measured by video microscopy under nonischemic conditions. Each cell served as its own control. Results: After equilibration in 37°C physiologic buffer (Tyrode's solution), exposure to 9°C St. Thomas' Hospital solution for 20 minutes caused human atrial cells to swell by 20% and rabbit atrial cells to swell by 10%. Cell volume fully recovered on rewarming in 37°C physiologic solution. Cell swelling was due to the composition of St. Thomas' Hospital solution rather than hypothermia alone. Exposure to 9°C physiologic solution did not significantly affect cell volume. Swelling of myocytes was largely prevented by replacing most of the Cl- in St. Thomas' Hospital solution with an impermeant anion so that the product of the concentrations of K+ and Cl- were the same as in the physiologic solution. Conclusions: This study suggests that cell swelling during hypothermic cardioplegia is caused in part by the composition of the cardioplegia solution. The volume of cardiac myocytes appears to follow a Donnan equilibrium in the cold, and the perfusate KCl product determines water movement. Thus, the tonicity of hyperkalemic cardioplegic solutions can be adjusted to a physiologic value by replacing most Cl- by an impermeant anion. Following this simple principle, a reformulation of cardioplegic solutions may be able to minimize iatrogenic myocardial edema.