The necessity, the methodology and clinical benefit of physical training were evaluated in 85 patients with valvular heart surgery. In 21 patients who had undergone the exercise stress test before surgery, corrective valvular surgery increased maximal oxygen uptake (max.VO2) from 15.4 before to 18.9 ml/min/kg, while 11 (52%) of them failed to increase exercise capacity over 5 METs. Of the total 85 patients, 41 (58%) again showed reduced exercise capacity in the initial postoperative stress test, suggesting that a return to work might be difficult without rehabilitation in these patients. We then compared two training programs: program A of short duration and frequent exercise and program B of longer duration and less frequent exercise, both using a bicycle exercise set at an intensity of 70% of the max.VO2. Both programs similarly increased max.VO2, while patients preferred program A, suggesting that exercise of longer duration could not be tolerated because of deconditioning. Program A was then prescribed to 62 patients, and it increased max.VO2 from 18.2 to 20.7 ml/min/kg after 4 weeks training without any complication. In 9 patients who served as controls undergoing no physical training, no spontaneous improvement in exercise capacity was observed. Of the 76 patients who received either program A or B, 28 patients failed to increase the max.VO2 by 10% or more. These patients presented atrial fibrillation, a cardiothoracic ratio greater than or equal to 60% or exercise-induced ST depression more frequently, suggesting that residual cardiac dysfunction might inhibit the training effects.(ABSTRACT TRUNCATED AT 250 WORDS)