We performed serial determinations of levels of potassium in 198 patients with suspected or proven coronary arterial disease who underwent a dobutamine stress test, so as to investigate if the depression in the ST segment induced by the test may be due to hypokalemia. The test consisted of an intravenous infusion of dobutamine, starting with a dose of 5-mu-g / kg / min for 5 minutes and continuing with 10, 15, 20 and up to 40-mu-g / kg / min every 5 minutes (mean peak dose = 20-mu-g / kg / min). Serial 12-lead electrocardiograms were taken to detect changes in the ST segment. The double product changed with dobutamine from 8844 +/- 6000 to 15201 +/- 3030. The peak dose of dobutamine induced a small but significant decrease in levels of serum potassium, with a further decrease 10 minutes later. In the 198 patients, the plasma potassium changes from 4.22 +/- 4.8 to 3.86 +/- 0.35 mmol / 1 (P < 0.00001). The maximum decrease in potassium (0.56 +/- 0.49) occurred in the patients who received the highest dose of dobutamine (30 to 40-mu-g / kg / min). Only 17 patients reached levels lower than 3.5 mmol / 1, and 4 of them achieved levels lower than 3.1 mmol / 1. No correlation was found between depression of the ST segment equal to or greater than 1 mm on the electrocardiogram and the level of potassium after the test. No correlation was found between ventricular arrhythmias and levels of potassium. High doses of dobutamine, therefore, produce a small but significant decrease in potassium. As the depression of the ST segment observed on the electrocardiogram is not related to levels of potassium, it must be due to dobutamine-induced ischemia, which supports the value of the test.