Retroperitoneal laparoscopic（RPL）surgery，utiliz⁃ ing retroperitoneal carbon dioxide（CO₂）insufflation，is currently established as a safe and reliable technique for specific urologic procedures. Although arterial blood gas（ABG）remains the golden standard for monitoring arterial blood carbon dioxide partial pressure（P_aCO₂）， it is invasive and lacks consistency. The frequent need for ABG analysis also contributes significantly to iatro⁃ genic anemia，particularly in critically ill patients and infants.

End ⁃tidal carbon dioxide（P_ETCO₂）is a commonly used noninvasive method for predicting P_aCO₂ in me⁃ chanical ventilated patients. However，the accuracy of P_ETCO₂ can be influenced by various factors such as sur⁃ gical position，as well as severe cardiovascular or pul⁃ monary diseases. Another noninvasive method for mon⁃ itoring CO₂ partial pressure is transcutaneous carbon dioxide partial pressure（P_TCCO₂）. This method has been widely accepted and is reported to provide better accuracy in predicting P_aCO₂ compared to P_ETCO₂ under many circumstances during laparoscopic surgery. In a study by XUE et al. ^[1]，in patients undergoing pro⁃ longed pneumoperitoneum laparoscopic surgery，88% and 17% of the samples showed a clinically acceptable difference（ ≤ 5mmHg）between P_TCCO₂ ⁃ P_aCO₂ and P_ETCO₂⁃P_aCO₂，respectively.

Retroperitoneoscopic surgery，a minimally inva⁃ sive surgical technique used for treating urinary system conditions，involves operating in the space behind the peritoneal cavity，which is enclosed by the posterior ab⁃ dominal wall. This area contains loose connective tis⁃ sue and adipose tissue and spans from the neck to the pelvis. During retroperitoneoscopic procedures，a surgi⁃ cal cavity is created by separating the peritoneum and posterior abdominal wall. However，this blunt dissec⁃ tion leads to significant surgical trauma，potentially re⁃ sults in higher CO₂ absorption compared to intraperito⁃ neal laparoscopic techniques^[2] . Despite this，there is inconsistency in CO₂ absorption between intraperitoneal and retroperitoneal pneumoperitoneum. KADAM et al. ^[3] found that CO₂ absorption does not depend on the route of surgery. They found no significant difference in CO₂ absorption between laparoscopic and retroperi⁃ toneal nephrectomy，with only subcutaneous emphyse⁃ ma notably increasing CO₂ absorption. Similarly，NG et al. ^[4] suggested that retroperitoneoscopy does not exhib⁃ it higher CO₂ absorption compared to transperitoneal laparoscopy for renal or adrenal surgeries. However，in STREICH et al.’s study^[5]，they discovered that the ret⁃ roperitoneal approach results in greater CO₂ absorption than intraperitoneal insufflation in urologic surgeries.

Given this inconsistency，the aim of this study was to investigate the accuracy of two distinct CO₂ par⁃ tial pressure monitoring techniques（P_TCCO₂ and P_ETCO₂） and their correlation with P_aCO₂ in patients undergoing retroperitoneoscopic surgery.

1 Materials and methods

1.1 Materials

This prospective observational study received ap⁃ proval from the Institutional Ethics Committee of the First Affiliated Hospital of Nanjing Medical University and was registered on www.ClinicalTrials.gov（NCT03 226041）. Initially，patients who were classified as the American Society of Anesthesiologists（ASA）Ⅰ-Ⅲ and scheduled for retroperitoneoscopic urologic surgery were screened. Those with severe cardiovascular or re⁃ spiratory diseases，such as coronary heart disease， chronic obstructive pulmonary disease（COPD），asth⁃ ma，a history of smoking or lung surgery（lobectomy or simple wedge resection），and individuals with morbid obese[body mass index（BMI）≥ 30 kg/m² ]were ex⁃ cluded. Subsequently，written consent was obtained from each participant before the surgery.

1.2 Methods

1.2.1 Sample size

Based on our preliminary study，a sample size of 45 achieves 90% power to detect a mean of paired differences of 5.0 mmHg，with an estimated standard devi⁃ ation of differences of 9.9 mmHg，at a significance level α of 0.05 using a paired t⁃test. Given the possibility of loss to follow ⁃ up，we increased the sample size by 10%，resulting in a required sample size of 50.

1.2.2 determination of P_TCCO₂ and P_ETCO₂

After entering the operating room，a16⁃G intrave⁃ nous（Ⅳ）catheter was inserted into the median cubi⁃ tal vein for fluid and drug administration，while a20⁃G arterial catheter was cannulated in the non ⁃ operated radial artery for continuous blood pressure（BP）monitor⁃ ing and ABG sampling. The arterial catheter was flushed with 500 mL of heparinized saline using a pres⁃ sure bag. Standard monitoring including electrocardio⁃ gram（ECG），saturation of pulse oxygen（SpO₂），and arte⁃ rial BP was performed for all patients before anesthesia， with these values recorded as baseline values. Anesthe⁃ sia induction comprised propofol（1.5-2.5 mg/kg），fen⁃ tanyl（2-4 μg/kg），and rocuronium（0.6 mg/kg）. Fol⁃ lowing intubation，patients were ventilated with volume control ventilation（VCV）using60% oxygen（2 L/min）. The P_ETCO₂ values were maintained ideally between 35-45 mmHg by adjusting tidal volume，respiratory rate， and aspiration ratio（inspiratory∶expiratory，I∶E），with an upper limit of 50 mmHg allowed. P_ETCO₂ was mea⁃ sured by side stream spirometry（Mindray，BeneView T6，Shenzhen，China），while P_TCCO₂ was measured with the TCM ⁃ 4 monitor（Radiometer，Copenhagen， Denmark）. Before placement，calibration was per⁃ formed by a trained author（LIU Shijiang）according to the manufacturer’s recommendation. The electrode was then placed onto the patient’s chest wall of the non⁃operated side，which was cleaned with alcohol to facilitate the adhesion of the disk to the skin，with the electrode working temperature set at 44℃. P_aCO₂ was determined using a blood gas/electrolyte analyzer （GEM premier 3000，Instrumentation Laboratory Co. MA 01730 ⁃ 2443，USA）. Before ABG sampling，pa⁃ tients’hemodynamic was relatively stable for at least 5 min to ensure a stable P_aCO₂. P_TCCO₂ and P_ETCO₂ were recorded simultaneously with ABG sampling.

Anesthesia was maintained with propofol，sevoflu⁃ rane，and remifentanil to limit the BP and heart rate （HR）fluctuations within 20% of baseline values. Pa⁃tients requiring a vasopressor to maintain hemodynamic stability or experiencing a hemodynamic fluctuation ex⁃ ceeding20% were excluded. Data collected before the use of a vasopressor or hemodynamic instability could still be used for analysis. Close communication was maintained with surgeons during the surgery. Patients were excluded if the peritoneum had been ruptured， but the data collected before peritoneal rupture was re⁃ tained. Patient’s temperature was continuously moni⁃ tored and maintained above36℃，while the room tem⁃ perature was set at 23-25℃. The retroperitoneal CO₂ pressure was maintained at 12-15 mmHg during the surgery. P_aCO₂，P_ETCO₂，and P_TCCO₂ of each patient were measured at four time points：before CO₂ insuffla⁃ tion，30，60 and 90 min after CO₂ insufflation.

If P_ETCO₂ exceeded 50 mmHg during the surgery， adjustments could be made such as increasing respira⁃ tory rate，adjusting tidal volume，increasing the flow of fresh oxygen，reducing pneumoperitoneum pressure within the surgeon’s acceptable range，or pausing the operation or closing the pneumoperitoneum if neces⁃ sary to enhance CO₂ excretion and lower P_ETCO₂ levels.

1.3 Statistical analysis

Statistical analyses were performed using Graph⁃ Pad 8.0 software（GraphPad Prism，La Jolla，California， USA）. Quantitative data were presented as means ± standard deviation（$\stackrel{-}{x}\pm s$）or median with interquartile range[M（P₂₅，P₇₅）]depending on the type of distribu⁃ tion. A difference of ≤5 mmHg between P_aCO₂ and P_ETCO₂，or between P_aCO₂ and P_TCCO₂，was considered within the clinical acceptable range. Categorical vari⁃ ables，presented as frequencies（proportions）[n（%）]， were analyzed using the chi⁃square test or Fisher’s exact test as appropriate. Pearson correlation coefficient was employed to assess the correlation between P_ETCO₂ and P_aCO₂，as well as the correlation between P_TCCO₂ and P_aCO₂. Additionally，linear regression analysis was uti⁃ lized to model and quantify these relationships. Bland⁃ Altman analysis was used to compare the agreement be⁃ tween P_aCO₂ and P_ETCO₂，or between P_aCO₂ and P_TCCO₂. P <0.05 is considered statistically significant.

2 Results

Ninety ⁃ seven patients were initially assessed for eligibility，of whom 31 were excluded due to severe complications，a history of surgery，or morbid obesity. Within the remaining66 subjects，6 were reluctant to participate and 10 were excluded because phenylephrine was used during anesthesia induction and surgery. Fi⁃ nally，50 patients were included in this study. A detailed flowchart of participant enrollment was shown in Figure 1. In addition，as shown in Table 1，the50 patients（16 women）have a mean age of（42.14 ± 14.42）years and a BMI of（23.34 ± 2.99）kg/m² . Thirty ⁃ one underwent partial nephrectomy，while the rest underwent nephrec⁃ tomy，urethroplasty，and renal cyst excision. The mean duration of CO₂ pneumoperitoneum was 91.42（30.00-192.00）min.

The values of P_aCO₂，P_ETCO₂，and P_TCCO₂ were re⁃ corded at the following four time points：before，and 30，60，90 min after pneumoperitoneum. After exclud⁃ ing values with a P_ETCO₂ <35 mmHg or >50 mmHg，157 samples were finally analyzed. As shown in Table 2，the average level of P_aCO₂，P_ETCO₂，and P_TCCO₂ at eachtime point were presented. Either P_aCO₂，P_ETCO₂ or P_TCCO₂ increased and reached a plateau at 30-60 min， and slightly decreased at 90 min after pneumoperito⁃ neum.

BMI：body mass index；ASA：American Society of Anesthesiolo⁃ gists；IAP：intra⁃abdominal pressure.

The correlation analysis between P_aCO₂ and P_ETCO₂， as well as P_aCO₂ and P_TCCO₂，was performed at different time points or different P_ETCO₂ levels. As shown in Table 2 and 3，a statistically significant correlation with P_aCO₂ was observed for both P_ETCO₂ and P_TCCO₂. Moreover， the correlation coefficient with P_aCO₂ was consistently greater for P_TCCO₂ compared to P_ETCO₂，whether detect⁃ ed at each time point or with P_ETCO₂ maintained at 35-40，40-45，or 45-50 mmHg.

Furthermore，based on linear regression analysis， both P_ETCO₂ and P_TCCO₂ were closely correlated with P_aCO₂. The linear regression equations were as follows： P_ETCO₂=0.60×P_aCO₂+9.10（r ² =0.62，P <0.001，Figure 2A），P_TCCO₂=1.07 × P_aCO₂-7.30（r ² =0.83，P <0.001， Figure 2B）.

Additionally，the average P_aCO₂ ⁃P_ETCO₂ difference was（13.20 ± 4.43）mmHg，and the average P_aCO₂ ⁃ P_TCCO₂ difference was（4.35 ± 2.56）mmHg（P <0.05）.Among all157 samples，a difference ≤5 mmHg or ≤3 mmHg between P_aCO₂ and P_ETCO₂ was observed in 5（3.2%）and 1（0.6%）sample，respectively. However，a difference ≤5 mmHg or ≤3 mmHg between P_aCO₂ and P_TCCO₂ was recorded in 101（64.3%）and 57 （36.3%）samples，respectively（P <0.001）. According to Bland ⁃Altman analysis，the95% limit of agreement （LOA）of P_aCO₂ versus P_ETCO₂ was 4.53 to 21.88 mmHg （Figure 3A）and P_aCO₂ versus P_TCCO₂ was-3.18 to 10.48 mmHg（Figure 3B）.

3 Discussion

Our results demonstrated that P_TCCO₂ estimated P_aCO₂ more accurately than P_ETCO₂ in patients undergo⁃ ing retroperitoneoscopic urologic surgery. Of note，the correlation between P_aCO₂ and P_TCCO₂ was consistently higher than that between P_aCO₂ and P_ETCO₂ at all time points，even when P_ETCO₂ was maintained within the ranges of 35-40 mmHg，40-45 mmHg，or 45-50 mmHg.

A difference of ≤ 5 mmHg between two measure⁃ ments is generally considered clinically acceptable，in⁃ dicating interchangeability. In the present study，the difference between P_aCO₂ and P_TCCO₂ was ≤5 mmHg in 101 out of 157 measurements（64.3%）whereas the dif⁃ ference between P_aCO₂ and P_ETCO₂ was ≤5 mmHg in only 5 out of 157 measurements（3.2%）. Compared to other studies defining acceptable bias as ≤3 mmHg^[6-8]， 36.3% of P_TCCO₂ values and only 0.6% of P_ETCO₂ val⁃ ues fell within this threshold in our study. Our findings suggested that P_TCCO₂ showed a greater accuracy than P_ETCO₂ in predicting P_aCO₂，with more values within 3 mmHg or 5 mmHg of P_aCO₂. Additionally，the mean P_aCO₂ ⁃ P_ETCO₂ difference was（13.20 ± 4.43）mmHg （95%CI：4.53-21.88 mmHg）. In contrast，the mean P_aCO₂ ⁃ P_TCCO₂ difference was（4.35 ± 2.56）mmHg （95% CI：-3.18 to 10.48 mmHg）. Taken together， these results indicate that P_TCCO₂ estimated P_aCO₂ more accurately than P_ETCO₂ in patients undergoing retroperitoneoscopic urologic surgery.

Retroperitoneoscopic surgery provides a minimally invasive approach to treating urinary system disease. The retroperitoneum refers to the space behind the peritoneal cavity，bounded by the posterior abdominal wall. This space is filled with adipose and loose con⁃ nective tissue，extending from the neck to the pelvis， and it is highly vascularized. During the operation，ex⁃ tensive tissue dissection is required to create the retro⁃ peritoneal space，potentially increasing CO₂ absorption compared to intraperitoneal laparoscopy. Consequently， hypercarbia may occur.

Clinically，hypercapnia principally impacts the cerebrovascular and cardiovascular system. Elevated P_aCO₂ causes cerebral vasodilation and increases intra⁃ cranial pressure despite autoregulatory mechanisms. Moreover，acute hypercapnia may increase the release of catecholamines due to β ⁃ adrenergic stimulation. This may be detrimental in procedures like retroperito⁃ neoscopic adrenalectomy，especially for pheochromocy⁃toma，as heightened catecholamine levels exacerbate hemodynamic instability. Given the intermittence of ABG analysis，a continuous，non⁃invasive method to ac⁃ curately predict P_aCO₂ during retroperitoneoscopic sur⁃ gery is necessary. P_TCCO₂ was found to be equivalent or even superior to P_ETCO₂ in predicting P_aCO₂ in differ⁃ ent populations^[9-11] . However，the correlation between P_TCCO₂ and P_aCO₂ remains unclear in retroperitoneo⁃ scopic surgery. In this study，P_aCO₂ and P_TCCO₂ showed a stronger correlation than P_aCO₂ and P_ETCO₂ across all subject groups（0.83 vs. 0.62）. Subgroup analysis re⁃ vealed a declining correlation between P_ETCO₂ and P_aCO₂ as P_ETCO₂ rose from 35-40 mmHg to 45-50 mmHg（0.41 to 0.18）. In contrast，P_aCO₂ ⁃ P_TCCO₂ correlation re⁃ mained high at 0.72 when P_ETCO₂ exceeded 45 mmHg. This indicates P_TCCO₂ monitoring may have greater ac⁃ curacy and sensitivity for detecting hypercapnia than P_ETCO₂. In other studies，the correlation between P_aCO₂ and P_ETCO₂ values or between P_aCO₂ and P_TCCO₂ values was higher than those in the present study，especially at baseline. This difference may be attributed to hemo⁃ dynamic fluctuations during anesthesia induction and positional changes，which increased the mismatch of the ventilation/perfusion（V/Q）ratio.

A patient position has a considerable influence on the accuracy of P_ETCO₂ monitoring. The lateral position often used in retroperitoneoscopic surgery can increase intrathoracic pressure and pulmonary pressures while decreasing venous return. Collectively，these effects re⁃ duce pulmonary blood flow，creating a mismatch be⁃ tween alveolar ventilation and perfusion. Thus，the dif⁃ ference between P_aCO₂ and P_ETCO₂ in lateral position was greater than those in other positions^[12-13] . Further⁃more，study had shown that P_TCCO₂ monitoring more ac⁃ curately predicted P_aCO₂ than P_ETCO₂ monitoring in trendelenburg position^[10] . Similarly，our study found more P_TCCO₂ than P_ETCO₂ values were within ≤3 mmHg or ≤5 mmHg of P_aCO₂. Whether the position affects P_TCCO₂ accuracy remains unclear，although tissue per⁃ fusion and electrode position can significantly affect ac⁃ curacy^[14-16] . Therefore，we chose the front chest wall in lateral position to ensure sufficient blood flow through the electrode.

In spite of the P_TCCO₂ can precisely estimate P_aCO₂， many technical factors can still affect the accuracy of P_TCCO₂ monitoring，including monitor factors（penetra⁃ tion of air bubbles，incorrect electrode positions，dam⁃ age of electrode membranes，and inaccurate calibra⁃ tion，etc.）and patient factors（skin blood perfusion， skin thickness，edema，dehydration，vascular active drug and anoxic acidosis. etc.）^[8，17-18] . Heating elec⁃ trodes can improve the reaction time，and increase lo⁃ cal blood flow through capillary arterialization but re⁃ duces measurement accuracy. NISHIYAMA et al. ^[19] thought the electrodes should be heated to at least 43℃ to guarantee more accurate estimates of P_aCO₂ and PaO₂. SϕRENSEN et al. ^[20]found that lower elec⁃ trode temperature increases the system error of mea⁃ sured values in premature and neonates. However， higher temperatures increase skin burn risk. Hence， we chosed 44℃ P_ETCO₂ values of 35-50 mmHg based on our data.

In conclusion，P_TCCO₂ demonstrated superior accu⁃ racy over P_ETCO₂ for estimating P_aCO₂ in retroperitoneo⁃ scopic urologic surgery. Although P_TCCO₂ may not re⁃ place the application of P_ETCO₂，it provides a promising continuous，non⁃invasive approach for monitoring P_aCO₂ and an early warning for hypercapnia.

Acknowledgements

We sincerely thank Professor YIN Changjun（1964-2015）for his excellent contributions to the Department of Urology in the First Affiliated Hospital of Nanjing Medical Chiversity. The authors acknowledge the indi⁃ viduals who contributed to this study and those who provided professional cares for the patients involved： WANG Zengjun，LI Pengchao，SHAO Pengfei and LI Jie（Department of Urology，the First Affiliated Hospi⁃tal of Nanjing Medical University）. We are greatly in⁃ debted to Professor ZHANG Kai from the Pancreas In⁃ stitute of Nanjing Medical University，the Pancreatic Center and Department of General Surgery at The First Affiliated Hospital of Nanjing Medical University for his guidance in statistical analysis.

参考文献