Mechanical ventilation causes significant morbidity and mortality in patients with severe asthma. Hypoventilation may reduce this morbidity and mortality, but indicators to guide the degree of hypoventilation are unclear. We used a measure of pulmonary hyperinflation to assess the degree of airflow obstruction and to guide the extent and duration of hypoventilation. Ten patients who required mechanical ventilation for scute severe asthma were studied. All were sedated, paralyzed, and given an initial minute ventilation (VE) of 200 ml/kg/min. End-inspiratory lung volume (VEI) above FRC was measured from the total exhaled gas volume during 40 to 60 a of apnea. VEI was used to regulate VE to a safe level (VE(safe)), irrespective of Pa(CO2), by reducing the rate when VEI was > 20 ml/kg and increasing it when VEI was < 20 ml/kg. Each patient was weaned when VE(safe) resulted in Pa(CO2) less-than-or-equal-to 40 mm Hg (the weaning point). FRC was measured by computer analysis of anterior and lateral chest radiographs taken at the end of apnea. Using the weaning point criterion, 2 patients (Pa(CO2) < 40 mm Hg) were weaned shortly after arrival. The remaining eight (initial Pa(CO2), 63 +/- 17 mm Hg) continued hypoventilation until the weaning point was reached (30 +/- 29 h). The weaning point was reached by the VE required for Pa(CO2) 40 mm Hg decreasing concurrent with the VE(safe) increasing. All but 1 patient were successfully weaned within 24 h of the weaning point. initial FRC was elevated (4.8 +/- 1.9 L; predicted, 3.2 +/- 0.3 L; p = 0.06) and decreased in parallel with VEI when airflow obstruction had improved at the weaning point. FRC showed a high level of correlation with VEI (r = 0.85, p < 0.001). In conclusion, VEI was a valuable indicator of the severity of airflow obstruction and the level of FRC. VEI provided a reliable guide to the extent of hypoventilation and the duration of mechanical ventilation.