Seventy frail elderly patients attending a day hospital were studied to investigate the incidence of postural hypotension in control patients on no diuretic treatment (n = 30), patients on loop diuretics (frusemide) for mild cardiac failure (n = 20) and patients on thiazide diuretics for mild cardiac failure (n = 20). The results showed that patients on thiazide diuretics had a higher incidence of postural hypotension (reduction in systolic blood pressure greater than 20 mmHg on standing after 2 minutes) than patients on loop diuretics (12 out of 20 vs 4 out of 20, p less than 0.05). Mean plasma potassium levels were lower in the thiazide group than in the frusemide group (p less than 0.05) and this correlated significantly with change in systolic blood pressure from supine to standing position (r = -0.56, p less than 0.01). These findings indicate that a loop diuretic (frusemide) is as safe if not safer than thiazides in older patients and the latter have an unjustified reputation of being safer first-line drugs for the treatment of cardiac failure.