To investigate the correlation and accuracy of transcutaneous carbon dioxide partial pressure (PTCCO2) with regard to arterial carbon dioxide partial pressure (PaCO2) in severe obese patients undergoing laparoscopic bariatric surgery. Twenty-one patients with BMI>35 kg/m(2) were enrolled in our study. Their PaCO2, end-tidal carbon dioxide partial pressure (PetCO2), as well as PTCCO2 values were measured at before pneumoperitoneum and 30 min, 60 min, 120 min after pneumoperitoneum respectively. Then the differences between each pair of values (PetCO2-PaCO2) and. (PTCCO2-PaCO2) were calculated. Bland-Altman method, correlation and regression analysis, as well as exact probability method and two way contingency table were employed for the data analysis. 21 adults (aged 19-54 yr, mean 29, SD 9 yr; weight 86-160 kg, mean 119.3, SD 22.1 kg; BMI 35.3-51.1 kg/m(2), mean 42.1,SD 5.4 kg/m(2)) were finally included in this study. One patient was eliminated due to the use of vaso-excitor material phenylephrine during anesthesia induction. Eighty-four sample sets were obtained. The average PaCO2-PTCCO2 difference was 0.9 ± 1.3 mmHg (mean ± SD). And the average PaCO2-PetCO2 difference was 10.3 ± 2.3 mmHg (mean ± SD). The linear regression equation of PaCO2-PetCO2 is PetCO2 = 11.58+0.57 × PaCO2 (r(2) = 0.64, P<0.01), whereas the one of PaCO2-PTCCO2 is PTCCO2 = 0.60 + 0.97 × PaCO2 (r(2) = 0.89). The LOA (limits of agreement) of 95% average PaCO2-PetCO2 difference is 10.3 ± 4.6 mmHg (mean ± 1.96 SD), while the LOA of 95% average PaCO2-PTCCO2 difference is 0.9 ± 2.6 mmHg (mean ± 1.96 SD). In conclusion, transcutaneous carbon dioxide monitoring provides a better estimate of PaCO2 than PetCO2 in severe obese patients undergoing laparoscopic bariatric surgery.