Background: The present study was undertaken to assess the possible causes of hypokalemia, which occurs in 10-36% of continuous ambulatory peritoneal dialysis (CAPD) patients. Methods: Twenty CAPD patients enrolled in the study were assigned to two groups according to their serum potassium levels. Group Ⅰ comprised ten patients with hypokalemia (serum potassium level<3.5 meqll for three consecutive months); Group Ⅱ comprised ten patients with normokalemia. Patients with diabetes, or those receiving medications such as beta-blockers, angiotensin-converting-enzyme inhibitors, or diuretics were excluded. Patients in both groups were monitored for daily total potassium loss (TKL) including dialysate (DKL) and urinary potassium loss (UKL), and fasting serum insulin concentration. Results: The underlying biochemical values including sodium, calcium, and albumin were similar in both groups. DKL and TKL were significantly increased in Group Ⅱ (DKL, 30.10±3.31 versus 37.35±9.78 mEqld, p=0.048; TKL, 32.54±3.39 versus 44.21±8.29 mEqld in Groups Ⅰ and Ⅱ respectively) (p=0.001). The average fasting serum insulin level was higher in Group Ⅰ than in Group Ⅱ (35.06±4.99 versus 27.17±5.37 mUN, p=0.003). Conclusion: The findings suggested hypokalemia in CAPD patients did not result from excessive potassium loss in the dialysate, but was associated with fasting hyperinsulinemia. We hypothesize glucose loading during CAPD stimulated hyperinsulinemia resulted in hypokalemia by shifting potassium to the intracellular space.