Successful withdrawal of digoxin in small highly selected groups of patients and in larger groups, in which many patients had no cardiac lesion, have led to speculation as to the value of maintenance digoxin following an episode of heart failure. Recent experience in general medical outpatients receiving maintenance digoxin, and also in many cases potent diuretics, suggests that such doubts are largely misplaced.Differences in bioavailability between formulations have hindered the prediction of dose requirements. Standardisation of dissolution rate has reduced this contribution to the measured variation in individual absorption, but substantial variation remains which is probably due to individual differences in gastrointestinal function.The best criterion by which to judge the appropriateness of a prescription would be the clinical response of the patient. Nevertheless, guidelines are required for initial treatment and even for long term treatment of those patients in whom there is no convenient indicator of therapeutic response. In both situations, renal function, measured by creatinine clearance, has been the best guide to the dose schedule required to achieve the desired mean steady state serum digoxin concentration. In adult patients with normal or only slightly impaired renal function, and normal atrio-ventricular conduction, the most practical procedure may be to start all patients on the same dose schedule.Symptoms of toxicity can occur at mean serum concentrations within the published therapeutic range. In these circumstances, reduction of the dose by small steps may abolish toxicity without loss of therapeutic response. Failure to reduce the dose encourages non-compliance. Further reduction during follow-up may be necessary because of deterioration of renal function or progression of cardiac disease.