BACKGROUND: Repeated procurement of a blood gas specimen in patients with sepsis is likely to increase the risk of infection and iatrogenic anemia. Non-invasive transcutaneous carbon dioxide pressure (PtcCO2) as a surrogate of invasive arterial carbon dioxide pressure (PaCO2)，which is increasingly suggested for its accuracy and sustainability in recent years. The objective of this study was to explore the accuracy and clinical application of PtcCO2 in patients with sepsis. METHODS: After approval of the local ethics committee PtcCO2 and PaCO2 were determined in n = 65 ICU patients with sepsis and n = 32 ICU non-sepsis patients. Bland–Altman method was used for statistical analysis, p＜0.05 was considered statistically significant. RESULTS: Statistical analysis revealed good correlation between PtcCO2 and PaCO2 in septic patients (r=0.591, P<0.0001) as well as in non-sepsis patients (r=0.642, P<0.0001). Bland–Altman analysis revealed a bias and precision of 3.4±8.7 mmHg for the septic patients and 2.2 ± 5.7 mmHg for the non-sepsis patients, both of them indicated good accuracy. During the first 6 hours of standardized treatment in septic shock patients, lactate clearance was linearly correlated with PtcCO2 difference (r=0.479, P=0.005<0.05), suggesting that early changes in PtcCO2 levels may be associated with tissue perfusion. CONCLUSIONS: PtcCO2 was a good alternative to PaCO2 in septic patients, and early changes in PtcCO2 may indicate changes in lactate clearance.