Abstract: Background：Transcutaneous carbon dioxide partial pressure (TcPCO2) and transcutaneous oxygen partial pressure (TcPO2) monitoring make it possible to continuously monitor neonatal PCO2/PO2 and reduce the number of blood collections. However, clinicians are very concerned about its safety and accuracy. Objective：To reflect the consistency of TcPCO2/TcPO2 monitoring values and PaCO2/PaO2 monitoring values through a systematic review and meta-analysis. Design：Systematic review and meta-analysis. Methods：A systematic literature retrieval was performed in English databases of Ovid MEDLINE, Ovid Embase and Chinese databases of SinoMed, Wanfang and CNKI. Search strategy was developed using key words of neotate, TcPCO2 and TcPO2. After deduplication, the literature was first screened by reading titles and abstracts by two leading authors independently who then finished the full-text screening. A self-made excel form was used to extract the year of publication, gestational age at birth, gestational age at monitoring, birth weight, weight at monitoring, sample size and the number of monitored pairs, the probe temperature, monitoring site, duration and equipment, blood gas values difference by TcPCO2/TcPO2and PaCO2/PaO2 , correlation coefficient and adverse events. Meta-analysis was performed by Stata 17. Main outcome measures：Difference in blood gas results between TcPCO2/TcPO2and PaCO2/PaO2. Results：A total of 875 articles were retrieved in both Chinese and English. After deduplication, preliminary screening, and full-text screening, 40 articles were included in the analysis, including 14 in Chinese and 26 in English. The summarized differences between TcPCO2 and PaCO2 in probe temperature subgroups of 44℃ vs 43℃, 42℃, 40℃ vs 41℃, and 38℃ vs 39℃ were 2.37(95%CI: 2.17 to 2.67), 2.48(95%CI: 1.17 to 3.79), 4.47(95%CI: 3.03 to 5.92), and 5.96(95%CI: 5.27 to 6.64) respectively, and the Pearson coefficient ranged from 0.78 to 0.89 among the four subgroups. For the deviation from PaCO2, it was 5% in the 43℃ vs 44℃ subgroup, 118% to 14.8% in the 40℃ vs 41℃ subgroups, and 17% in the 38℃ vs 39℃ subgroups. The summarized difference between TcPO2 and PaO2 in probe temperature subgroups of 44℃ vs 43℃, 42℃, 40℃ vs 41℃, and 38℃ vs 39℃ were -7.28(95%CI: -12.92 to -1.63), -5.61(95%CI: -6.61 to -4.62), -28.85(95%CI: -35.12 to -22.58) and -31.42(95%CI: -35.97 to -26.88) respectively, and the Pearson coefficient was 0.35~0.91 among the four subgroups. Eight studies described mild skin redness and no adverse events of skin burns were reported. Conclusion：Without controlling the monitoring site, duration, time point, equipment, and neonatal diseases and conditions, TcPCO2 values are strongly correlated with the PaCO2 values, which is about 7mmHg higher than PaCO2. Under standardized operations, TcPCO2 can replace PaCO2. Within 38℃ to 44℃, the commonly used clinical probe temperature, the higher the probe temperature, the more accurately TcPCO2 can reflect PaCO2, the less skin redness adverse events will occur. Since TcPO2 is weakly related to PaO2, it is not recommended to use TcPO2 to reflect PaO2.

Key words: Newborn, Transcutaneous carbon dioxide partial pressure, Transcutaneous oxygen partial pressure, Oxygen partial pressure, Artery