OBJECTIVE The long-term cardiovascular effects of GH administration in adults are of major clinical importance, given the increasing use of such treatment. We have evaluated long-term cardiovascular effects of recombinant human GH (rhGH) substitution in GH deficient men. METHODS S.c. rhGH 0.5 U/kg/week or placebo was administered in a 6-month double-blind, cross-over study, followed (after a year without substitution) by a 42-month period of open GH substitution. METHODS We evaluated 7 GH-deficient men serially and compared the results with 21 men matched in terms of age and height. METHODS Investigations included exercise tests and Doppler-echocardiography to determine exercise capacity and cardiovascular performance. RESULTS Heart rate and systolic blood pressure at rest increased with GH substitution to the level of the controls, as did diastolic blood pressure after an initial reduction. Age-adjusted exercise capacity increased during the study and we found no evidence of ischaemic heart disease on exercise ECG. Stroke volume increased with GH substitution, thereby normalizing the initially reduced cardiac index. There was no significant change in left atrial or ventricular internal dimensions, systolic function as measured by fractional shortening, or diastolic function as measured by isovolumic relaxation time and left ventricular filling (A/E ratio). However, a lower atrial emptying index than that seen among controls might indicate some diastolic disturbance and there was a definite increase in left ventricular wall thickness compared with controls (to 25.1 +/- 1.5 vs 19.7 +/- 0.4 mm, P < 0.001). CONCLUSIONS We found that GH substitution in GH-deficient adults had a beneficial effect on physical performance and cardiac output. The concomitant increase in left ventricular mass index might be an effect of an excessive substitution dose.