BACKGROUND: The arterial partial pressure of CO(2) (P(aCO(2))) can be grossly estimated by the end-tidal partial pressure of CO(2) (P(ETCO(2))). This principle is used in SmartCare (Dräger, Lübeck, Germany), which is an automated closed-loop system that uses P(ETCO(2)) to estimate alveolar ventilation during mechanical ventilation. OBJECTIVE: To assess whether the maximum P(ETCO(2)) value (instead of the averaged P(ETCO(2)) value) over 2-min or 5-min periods improves P(aCO(2)) estimation, and determine the consequences for the SmartCare system. METHODS: We continuously monitored breath-by-breath P(ETCO(2)) during ventilation with SmartCare in 36 patients mechanically ventilated for various disorders, including 14 patients with COPD. Data were collected simultaneously from SmartCare recordings, every 2 min or 5 min, and through a dedicated software that recorded ventilation data every 10 s. We compared the maximum and averaged P(ETCO(2)) values over 2-min and 5-min periods to the P(aCO(2)) measured from 80 arterial blood samples clinically indicated in 26 patients. We also compared SmartCare's classifications of patient ventilatory status based on averaged P(ETCO(2)) values to what the classifications would have been with the maximum P(ETCO(2)) values. RESULTS: Mean P(aCO(2)) was 44 ± 11 mm Hg. P(aCO(2)) was higher than averaged P(ETCO(2)) by 10 ± 6 mm Hg, and this difference was reduced to 6 ± 6 mm Hg with maximum P(ETCO(2)). The results were similar whether patients had COPD or not. Very few aberrant values (< 0.01%) needed to be discarded. Among the 3,137 classifications made by the SmartCare system, 1.6% were changed by using the maximum P(ETCO(2)) value instead of the averaged P(ETCO(2)) value. CONCLUSIONS: Use of maximum P(ETCO(2)) reduces the difference between P(aCO(2)) and P(ETCO(2)) and improves SmartCare's classification of patient ventilatory status.