An immediate improvement in haemodynamic variables and cardiac performance is achieved in chronic heart failure following diuretic therapy, primarily due to reductions in plasma and extracellular fluid volumes. Humoral markers of these alterations are increased plasma renin, angiotensin and aldosterone levels; these increase maximally over the ñrst week of treatment but attenuate during sustained therapy. There are reciprocal alterations in plasma α-atrial natriuretic peptide levels. These findings suggest that the initial volume contraction is maintained, though somewhat attenuated, during chronic therapy. The neurohumoral consequences of diuretic therapy are of particular interest in heart failure, as they may contribute to diuretic resistance. Activation of the renin-angiotensin system favours the proximal tubular reabsorption of sodium and water, which may result in dilutional hyponatraemia. Diuretics have both direct vascular and non-vascular (volume-dependent) haemodynamic actions. Together these substantially reduce the left heart filling pressure (-29%) with a consequent fall in cardiac output ( 10%). Systemic vascular resistance initially increases but subsequently normalizes, allowing cardiac output to return towards control values. Haemodynamic tolerance to diuretics does not usually occur during sustained oral therapy; additionally, echocardiographic contractility indices and exercise capacity may increase. The vasodilator activity of the diuretics is due to prostaglandin release; the initial pressor action is due to activation of the renin-angiotensin system. Direct pulmonary vasodilatation with improved pulmonary compliance remains an interesting possibility. Over the longer term, substantial reductions in left heart filling pressure during exercise occur at unaltered cardiac output. The impact of diuretic therapy on the underlying myocardial disease process is unknown. There is little direct evidence of ventricular remodelling or of improved intrinsic cardiac performance (as distinct from that due to altered loading conditions). In the absence of such information, it would appear reasonable to combine diuretic therapy with a vasodilator or angiotensin converting enzyme inhibitor, even in patients symptomatically well controlled on diuretic alone.