Conflicting results have been published on prognostic significance of central venous to arterial PCO(2) difference (∆PCO(2)) after cardiac surgery. We compared the prognostic value of ∆PCO(2) on intensive care unit (ICU) admission to an original algorithm combining ∆PCO(2), ERO(2) and lactate to identify different risk profiles. Additionally, we described the evolution of ∆PCO(2) and its correlations with ERO(2) and lactate during the first postoperative day (POD1). In this monocentre, prospective, and pilot study, 25 patients undergoing conventional cardiac surgery were included. Central venous and arterial blood gases were collected on ICU admission and at 6, 12 and 24 h postoperatively. High ∆PCO(2) (≥ 6 mmHg) on ICU admission was found to be very frequent (64% of patients). Correlations between ∆PCO(2) and ERO(2) or lactate for POD1 values and variations were weak or non-existent. On ICU admission, a high ∆PCO(2) did not predict a prolonged ICU length of stay (LOS). Conversely, a significant increase in both ICU and hospital LOS was observed in high-risk patients identified by the algorithm: 3.5 (3.0-6.3) days versus 7.0 (6.0-8.0) days (p = 0.01) and 12.0 (8.0-15.0) versus 8.0 (8.0-9.0) days (p \textless 0.01), respectively. An algorithm incorporating ICU admission values of ∆PCO(2), ERO(2) and lactate defined a high-risk profile that predicted prolonged ICU and hospital stays better than ∆PCO(2) alone.