Objective To determine the causes of excess minute ventilation in patients initiated on mechanical ventilation. Design Prospective study of recently intubated, mechanically ventilated patients. Setting The medical ICU in a county hospital. Patients Fifty-two mechanically ventilated medical ICU patients were studied within 36 hrs of intubation. Patients were all supported with volume-cycled ventilation in the assist-control mode. Interventions Timed expired gas collection and an arterial blood gas. Measurements and Main Results Measurements of minute ventilation and CO2 production (Vco2) were made from a timed expired gas collection. Paco2 was sampled during the gas collection and deadspace was determined. Minute ventilation, Vco2, deadspace, and Paco2 values in the patients were compared with predicted normal values, and excess minute ventilation due specifically to each component was calculated. Patients were separated clinically into groups: adult respiratory distress syndrome (ARDS), sepsis, obstructive lung disease, pneumonia, and drug overdose. Comparisons were then made between groups. Excess minute ventilation for the entire study population was secondary to increased deadspace (39%), low Paco2 (36%), increased Vco2 (15%), and the interactive effect of deadspace and Vco2, (10%). Vco2 contributed little to excess minute ventilation early in respiratory failure, even in the ARDS and sepsis groups. Deadspace contributed significantly to excess minute ventilation in all groups, especially in the ARDS group, where it accounted for 53% of the excess ventilation. Low Paco2 set-point was the predominant cause of excess minute ventilation in the sepsis group, where it contributed to 57% of their total excess minute ventilation. Conclusions Although all groups initiated on mechanical ventilation had an excess ventilatory requirement, the contribution of individual components varied considerably between clinical groups. (Crit Care Med 1991; 19:916)