A randomized, double-blind, cross-over study comparing 50 mg hydrochlorothiazide plus 5 mg amiloride (HCTZ/A) with 50 mg hydrochlorothiazide plus 26 mmol potassium chloride (HCTZ/K) was conducted in 18 patients with mild essential hypertension (diastolic pressure 90-105 mmHg). The sequence of treatment was: placebo for 2 weeks, one active drug for 3 weeks, placebo for 2 weeks, the other active drug for 3 weeks. The two agents were significantly and equally efficacious in lowering the systolic and diastolic blood pressure. Baseline vs. treatment mean serum potassium levels were 3.82 vs. 3.78 mmol/l for HCTZ/A and 3.82 vs. 3.70 mmol/l for HCTZ/K. The decrease in serum potassium level from baseline was significant for both agents but not significantly different when the two treatment forms were compared. Both treatment forms elevated fasting serum cholesterol and glucose. Serum triglycerides and uric acid rose significantly with HCTZ/K. Amiloride may affect the tubular handling of uric acid causing increased uric acid excretion, thus counteracting thiazide-induced hyperuricemia. During 3 weeks' extension of the main study, 5 patients received HCTZ/A in double the original dose (100 mg/10 mg) and 6 patients received HCTZ/K in double the original dose (100 mg/52 mmol). No further blood pressure reduction was observed on treatment with these doses. The mean serum potassium levels did not decrease further on doubling the HCTZ/A dose, while a significant fall was observed for HCTZ/K (3.60 vs. 3.42 mmol/l) (p less than 0.05, single tailed t-test). Both drug combinations were well tolerated and side-effects were not significantly different from those during placebo administration. This study demonstrates that 50 mg hydrochlorothiazide plus 26 mmol potassium chloride are as effective as 50 mg hydrochlorothiazide plus 5 mg amiloride, both in reducing blood pressure and preventing hypokalaemia in the treatment of essential hypertension. A small extension study indicates that amiloride might be more effective than potassium chloride in preventing hypokalaemia when high doses (100 mg/day) of hydrochlorothiazide are administered.