17 patients with chronic ventilatory failure (including 14 with chronic obstructive pulmonary disease) were studied to determine the causes of carbon dioxide retention and the chronic effect of medroxyprogesterone acetate on ventilatory drive and acid-base status. Carbon dioxide retention in patients with high mechanical loads occurred concomitantly with a higher than normal inspiratory effort (mouth occlusion pressure) and normal minute ventilation to carbon dioxide production ratio (Ve/Vco(2)); but with shortened inspiratory time (1.3+/-0.1 vs. 1.8+/-3 s), increased breathing frequency (17+/-1 vs. 14+/-1 breaths/min), low tidal volume (0.57+/-0.03 vs. 0.88+/-0.04 L), and high dead space to tidal volume ratio (0.63+/-0.02 vs. 0.39+/-0.07). Using a randomized application of treatment and placebo conditions, it was shown that 4 wk of medroxyprogesterone acetate caused significant reductions in Paco(2) (from 51+/-1 to 42+/-1 mm Hg) in 10 of 17 patients. This "correction" of Paco(2) in these patients was associated with increases in mouth occlusion pressure (14%), tidal volume (11%), and alveolar ventilation (15%) compared to placebo, although inspiratory time remained shortened. Arterial and lumbar cerebrospinal fluid pH was alkaline compared to placebo in patients who "corrected" Paco(2). No change was noted in lung mechanics or core temperature. Common prerequisites for correction of Paco(2) with medroxyprogesterone acetate treatment were the ability to significantly lower Paco(2) upon acute voluntary hyperventilation and to increase tidal volume rather than breathing frequency in response to the drug. We attribute chronic CO(2) retention in these patients to alterations in respiratory cycle timing and to a neuromuscular inspiratory effort which is adequate for the level of tissue CO(2) production, but inadequate in the presence of mechanical and ventilation-perfusion abnormalities to normalize arterial blood gases.