Monitoring Oxygen Metabolism Using the Oxygen Reserve Index in Patients Undergoing Robotic Prostatectomy: A Prospective Observational Study

Introduction

The basic method for intraoperative monitoring of oxygen metabolism is pulse oximetry (SpO2) measurement. This technique, however, only detects changes within the partial pressure of oxygen in arterial blood (PaO2) range of 0 to 100 mmHg [1]. SpO2 cannot be used to diagnose hyperoxia or predict impending hypoxia. A parameter that makes such diagnosis possible is the oxygen reserve index (ORi). It is determined using a special sensor developed by Masimo Corporation (USA), which analyses changes in the absorption of incident light in arterial and venous blood. The ORi measures the oxygenation reserve status in the range of mild hyperoxia: PaO2 of 100 to 200 mmHg. These values correspond to variations in the ORi from 0.00 (indicating PaO2 ≤100 mmHg) to 1.00 (indicating PaO2 ≥200 mmHg) [2]. During oxygen supplementation, when PaO2 exceeds 100 mmHg, arterial oxygen saturation (SaO2) peaks at 100%, but venous oxygen saturation (SvO2) continues to increase until it reaches a plateau at PaO2 of approximately 200 mmHg. It is the changes in SvO2 in the PaO2 range of 100–200 mmHg that provide a basis for determining the ORi [3].

The ORi also demonstrates high sensitivity in detecting low PaO2 values, allowing prediction of impending hypoxia when SpO2 values are still within the normal range [4]. This property can be used to predict the risk of hypoxia during endotracheal intubation and in patients with a difficult airway [5,6].

Minimally invasive surgery is becoming a dominant part of interventional medicine. This also applies to urology procedures. Robot-assisted radical prostatectomies involve less tissue trauma, less postoperative pain, and a shorter hospital stay. These procedures, however, require positioning the patient in a steep Trendelenburg position and establishing a pneumoperitoneum. These unfavourable conditions can significantly disrupt the body’s physiology, especially the respiratory and circulatory functions [7].

Changes occurring in the respiratory system are caused by a cranial displacement of abdominal contents and the diaphragm. These include increased airway resistance and peak airway pressure, decreased functional residual capacity, decreased lung compliance, and an altered ventilation-to-perfusion ratio, which can lead to atelectasis [8].

The aim of our study was to analyze intraoperative variations in oxygen metabolism using ORi and PaO2 under conditions of ventilation compromised by the Trendelenburg tilt and pneumoperitoneum during a robot-assisted prostatectomy.

Material and Methods

ETHICS APPROVAL AND PATIENT CONSENT:

The study protocol was approved by the Bioethics Committee of the Medical University of Lublin (decision no. KE-0254/238/11) and complies with the tenets of the Declaration of Helsinki. Every patient received information about the study and gave written consent to participate.

DESIGN AND PATIENTS:

This was an observational prospective study. One hundred patients scheduled for an elective robotic prostatectomy participated in this study from 2 January 2023 to 22 March 2024. The inclusion criteria for the study were the patient’s informed consent, age 18 years or older, scheduled surgery using the Versius surgical robot (CMR Surgical, England), patients preoperatively classified in American Society of Anesthesiologist classification (ASA) as ASA I, II, or III. The exclusion criteria were as follows: baseline hemoglobin level below 10 g dl⁻¹, a history of brain injury, a history of neurosurgical procedure, a history of multiple diseases (ASA IV and V patients, particularly respiratory or circulatory diseases), a comorbid central nervous system pathology, or a confirmed cerebral circulation disorder. Each patient was anesthetised by the same anesthesia care team. Each surgery was performed by the same surgical team assisted by a Versius surgical robot.

ANESTHESIA:

Immediately after admission to the operating room, standard monitoring of vital signs (heart rate [HR], blood pressure [BP], SpO₂) was implemented using the Infinity Delta XL monitor (Dräger, Germany). The radial artery was then cannulated under local anesthesia using a 20-gauge cannula (BD, Belgium). The FloTrac System (Edwards Lifesciences, Irvine, USA) was connected to the cannula to monitor hemodynamic parameters – cardiac output (CO), cardiac index (CI), stroke volume variation (SVV), stroke volume (SV), and mean arterial pressure (MAP) – and to obtain direct BP measurement. Hemodynamic parameters were measured using the Edwards HemoSphere Advanced Monitor. The arterial cannula was also used to collect blood samples for blood gas analysis. After pre-oxygenation with 100% oxygen, anesthesia was induced with fentanyl at a dose of 3 μg kg⁻¹ and propofol at a dose of 1 to 1.5 mg kg⁻¹. Muscle relaxation was achieved by administration of rocuronium (0.6 mg kg⁻¹), and then, when train of four was 0%, endotracheal intubation was performed. After intubation, mechanical ventilation was started in the pressure control ventilation mode with fraction of inspired oxygen (FiO₂) 0.45, maintaining tidal volume at 500 to 580 ml, and a respiratory rate of 12 to 18 min⁻¹, depending on the end-tidal carbon dioxide values (etCO₂). Desflurane inhalation (minimum alveolar concentration 0.8 to 1.2) and a continuous infusion of remifentanil (adjustable dose 0.01 to 0.2 μg kg⁻¹ min⁻¹) were used to maintain anesthesia. Rocuronium was also administered intraoperatively in a continuous infusion (0.3 to 0.4 mg kg⁻¹ h⁻¹).

After the induction of anesthesia, the patient was immobilised in a Vacuform 2.0 mattress (BuW Schmidt GmbH, Germany) to prevent them from sliding off the operating table when in a steep Trendelenburg position.

Pneumoperitoneum was induced by injecting carbon dioxide into the peritoneal cavity until a pressure of 15 mmHg was achieved. Patients were then placed in a steep Trendelenburg position at a 27° incline. During surgery, intra-abdominal pressure was maintained at 15 mmHg in all patients. At the end of the procedure, patients were returned to the horizontal position, and the insufflation gas was passively removed from the peritoneal cavity.

MONITORING:

After arriving at the operating room, all patients had SedLine EEG Sensor electrodes (Masimo Corporation, USA) attached to their forehead to monitor the depth of anesthesia, and an ORi RD rainbow SET-2 Adt sensor (Masimo Corporation, USA) attached to the index finger to monitor the ORi. These parameters were observed using an O3 Regional Oximetry Root device (Masimo Corporation, USA).

ORi was monitored at the following time points: before induction of anesthesia (BEG), after pre-oxygenation with 100% oxygen (PREOX), after induction of anesthesia (INDUCT), after placing the patient in the Trendelenburg position and establishing pneumoperitoneum (PP+T), then every 20 minutes (AFT 10, AFT 30, AFT 50, and AFT 70), after returning the patient to the initial position (ZERO), after extubation (EXTUB), and before transferring the patient to the post-anesthesia care unit (PACU).

Arterial blood samples for blood gases analysis were determined: G1 – before the induction of anesthesia (time point BEG), G2 – after the patient was placed in the Trendelenburg position and pneumoperitoneum was established (time point PP+T) and G3 – after returning the patient to the initial position (time point ZERO). Arterial blood gas analysis was performed using the GEM Premier 5000 device. During each anesthesia, a member of the research team was present in the operating room, whose only task was to record the monitored parameters.

ASSESSMENT OF OUTCOME MEASUREMENTS:

The main monitored parameter was ORi. The oxygen reserve index was analyzed at each time point of the study (BEG, PREOX, INDUCT, PP+T, AFT10, AFT30, AFT50, AFT70, ZERO, EXTUB, PACU). Along with the ORi measurements, we recorded MAP, SV, CO, CI, SpO₂, and Pi.

The existence of a relationship between the body’s oxygen management parameters (ORi, PaO₂, and SpO₂) and the patients’ BMI at different time points was also assessed. The relationship between ORi and hemodynamic parameters was monitored. To provide a more complete picture of changes in the body’s oxygen management, ORi was monitored in parallel with SpO_2, and ORi was compared with PaO₂ at selected study points (G1 – BEG, G2 – PP+T, G3 – ZERO).

STATISTICAL ANALYSIS:

The data collected in the spreadsheet was analyzed statistically using MedCalc software (version 15.8 PL) and Statistica (version 13 PL). Since data on the correlation between ORi and PaO₂ at PP+T measurement point is limited in the literature, we calculated sample size post hoc based on our own data. Most medical studies consider a P value below 0.05 to reject the null hypothesis, thus a type I error (alpha) of 0.05 value was used. In the case of type II error (beta), we set a cut-off on 0.01 to achieve nearly 100% of statistical power. Considering Rho for correlation between ORi and PaO₂ at PP+T measurement point (primary endpoint) equal 0.521, the minimal study group was estimated as 58. Categorical variables were presented using absolute numbers and percentages. The normality of the distribution of continuous variables was tested using the D’Agostino-Pearson test. Because continuous variables turned out not to be normally distributed, the relationships between the variables were assessed using nonparametric tests. For the same reason, the median was used as a measure of data clustering, and the interquartile range and the minimum-maximum range were used to quantify dispersion. The Spearman rank correlation test was used to assess correlations between selected continuous variables including BMI and ORi, PaO₂, SpO₂ at different measurement points; ORi and oxygen metabolism parameters at different measurement points; Trendelenburg position time with ORi and PaO₂ at different measurement points. Statistically significant results of the correlation analysis were presented using scatter plots. Additionally, to assess independent correlations between BMI and selected oxygen metabolism parameters, a multivariable analysis (multiple regression) was performed including potential confounding factors. The co-variables for the multivariable analysis were selected based on clinical analysis of potential confounding factors, including ASA, age, presence of lung and heart disease, smoking, surgery and Trendelenburg position duration. After the backward elimination method, ASA and smoking were included as covariates in the final multivariable analysis of the correlations between BMI and the analyzed oxygen metabolism parameters. Similarly, to assess independent correlations between ORi and selected oxygen metabolism parameters in consecutive measurement points a multivariable analysis (multiple regression) was performed including potential confounding factors. The co-variables for the multivariable analysis were selected based on clinical analysis of potential confounding factors, including ASA, BMI, age, presence of lung and heart disease, smoking, surgery and Trendelenburg position duration. After the backward elimination method, the following covariates were included at specific ORi measurement points: BEG – ASA, Age, surgery, and Trendelenburg position duration; PP+T – BMI, surgery and Trendelenburg position duration, ZERO, EXTUB, INDUCT – BMI, Age, ASA; PACU - surgery and Trendelenburg position duration; PREOX – BMI; AFT10, AFT30, AFT50, AFT70 – BMI, surgery and Trendelenburg position duration. The distribution of continuous variables was compared between 2 independent groups using the Mann-Whitney U test. This test was used to compare ORi values between patients with and without lung disease. The distribution of continuous variables between individual time points was compared using the Friedman ANOVA (with post hoc analysis). This test was used to differentiate oxygen metabolism parameters, including ORi, PaO₂, and SpO₂ between subsequent measuring points. Statistically significant results of comparisons of continuous data between groups and between consecutive measurement points were visualised using box-and-whisker plots. The diagnostic usefulness of the ORi in detecting hyperoxia was assessed based on ROC curve analysis. For all statistical tests, results with an alpha error (P) less than 0.05 were considered statistically significant.

Results

SAMPLE DESCRIPTION:

Out of the total of 100 patients who underwent a robotic radical prostatectomy between January 2023 and March 2024, 96 were included in the study. Two patients were excluded due to serious comorbidities: 1 patient had dilated cardiomyopathy with heart ejection fraction 32% and the other had subarachnoid hemorrhage and had undergone carotid artery aneurysm surgery. Another patient was excluded from the study due to equipment failure, which required a change from a robotic to a laparoscopic approach. The fourth patient was disqualified because, due to cataract and glaucoma, he had to undergo the surgery at a less steep Trendelenburg position angle of 16 degrees.

The median age was 66 years. The median body mass index (BMI) was 27.7. In the study group, 67.7% of the patients were assessed as ASA II. Heart disease and lung disease were diagnosed in 65.6% and 14.6% of the patients, respectively. Regarding use of stimulants, 85.4% of the patients declared themselves as non-smokers. The median ORi value before the induction of anesthesia was 0. Detailed demographic and clinical data for the study group are shown in Table 1.

CIRCULATORY SYSTEM AND OXYGEN METABOLISM PARAMETERS:

The median SpO2 was significantly lower at BEG compared with the other measurement points. The median ORi was significantly lower at BEG than at the remaining measurement points: PREOX, INDUCT, PP+T, AFT10, AFT30, AFT50, AFT70, ZERO, and EXTUB (Figure 1).

By contrast, the median perfusion index (Pi) value was significantly higher at BEG compared with the PREOX, EXTUB, and PACU measurement points and significantly lower at BEG than at PP+T, AFT10, AFT30, AFT50, and AFT70. Details of the parameter comparisons for the successive measurement points are shown in Table 2.

In the study group, the median SVV was significantly higher at INDUCT compared with the PP+T, AFT10, AFT30, and ZERO measurement points. Additionally, etCO2 had a significantly lower median at INDUCT than at AFT30, AFT50, AFT70, and ZERO. Detailed parameter comparisons for each successive measurement point are shown in Table 3.

UNIVARIABLE ANALYSIS:

The univariable analysis showed that BMI was significantly, negatively correlated with selected oxygen metabolism variables. Significant correlations were observed for PaO₂ at the BEG, PP+T, and ZERO measurement points. Additionally significant negative correlations were observed between BMI and ORi at PREOX, PP+T and subsequent intraoperative measurement points (AFT10, AFT30, AFT50, AFT70). For SpO₂, significant negative correlations were found at the BEG, INDUCT, PP+T, AFT10, AFT30, AFT50, ZERO, and PACU measurement points.

CORRELATION BETWEEN THE ORI AND SELECTED VARIABLES: MULTIVARIABLE ANALYSIS:

In multivariable analysis, ORi remained independently associated with selected oxygen metabolism variables. Statistically significant positive correlations between ORi and PaO₂ persisted at BEG (Rho=0.412). Statistically significant positive correlations between ORi and PaO₂ as well as with SpO₂ persisted at PP+T (Rho=0.521 and Rho=528, respectively). Additionally, ORi was significantly negatively correlated with Pi at ZERO (Rho=−0.370), positively correlated with SpO₂ at EXTUB (Rho=0.486) and positively correlated with SpO₂ at AFT10 (Rho=0.580). Moreover, the statistically significant correlation between ORi and SpO₂ at PREOX was positive (Rho=0.371) and between ORi and Pi at PREOX it was negative (Rho=−0.385). There was statistically significant negative correlation between ORi and SVV at AFT30 (Rho=−0.334) and statistically significant positive correlations between ORi and SpO₂ at AFT30 (Rho=0.380), AFT50 (Rho=0.573), and AFT70 (Rho=0.434). Significant correlations were noted in the PACU: negative between ORi and Pi (Rho=−0.283) and positive between ORi and SpO₂ (Rho=0.532).

ORI AND HYPEROXIA: Figure 3 shows relationship between ORi and PaO2 measurements at 3 different time points. The value of the ORi parameter was characterized by a statistically non-significant 17.65% sensitivity and 100% specificity in detecting hyperoxia at the BEG measurement point (cut-off point: >0; AUC=0.588, 95% CI: 0.474–0.696; P=0.2956). At the PP+T time point, the value of the ORi parameter was characterized by a statistically significant 81.93% sensitivity and 100% specificity in detecting hyperoxia (cut-off point: >0; AUC=0.910, 95% CI: 0.831–0.960; P<0.0001).The value of the ORi parameter was characterized by a statistically significant 91.95% sensitivity and 100% specificity in detecting hyperoxia at ZERO time point (cut-off point: >0; AUC=0.960, 95% CI: 0.895–0.990; P<0.0001).

Discussion

LIMITATIONS:

Several limitations of this study should be considered. First, it was performed at a single tertiary referral center and involved a relatively limited number of participants, which may restrict the external validity and generalizability of the findings. The sample size (n=96) was determined pragmatically rather than based on a formal power calculation; consequently, the study may have been insufficiently powered to detect smaller but potentially relevant differences. All patients were anesthetized by the same anesthesia team, which ensured procedural consistency and minimized inter-operator variability. However, this may limit the reproducibility of the results in other clinical settings where anesthetic management may differ.

Second, arterial blood gas analyses results were obtained at only 3 intraoperative time points, which were considered crucial. However, this may have constrained the assessment of continuous or transient fluctuations in oxygen metabolism during the procedure.

Third, our study did not include a control group, which may distort the interpretation of the obtained ORi results. In the future, a control group of patients should be included to enable monitoring of the oxygen reserve index without the influence of harmful factors such as pneumoperitoneum or Trendelenburg position.

Although ORi demonstrated high sensitivity and specificity at selected intraoperative time points, this study focused exclusively on intraoperative correlations between ORi and physiological parameters. It was not designed to assess associations between ORi values and postoperative clinical outcomes, including pulmonary or cardiovascular complications. Therefore, the clinical significance of intraoperative ORi changes in relation to broader perioperative outcomes remains uncertain, and future studies should aim to link ORi monitoring with patient-centered pulmonary and cardiovascular endpoints.

Finally, as the investigation included only male patients undergoing robot-assisted prostatectomy, extrapolation of the results to other surgical populations should be undertaken with caution.

Future multicenter investigations involving larger cohorts and incorporating clinically relevant outcome measures are needed to validate these observations and to further elucidate the clinical utility of ORi monitoring in perioperative practice.

Conclusions

The oxygen reserve index is a parameter that demonstrates a positive correlation with SpO2 values. Furthermore, it strongly correlates with PaO2 values obtained from blood gas analysis. It is a sensitive and highly specific tool for detecting hyperoxia. Crucially, ORi measurements are non-invasive and are continuously monitored, with results obtained in real time, which is a clear advantage over the method of obtaining PaO2 values. It is important to remember that blood gas analysis is the most accurate method for determining blood oxygen tension. Nevertheless, ORi is a very good complementary tool, and in situations where a gas analyzer is unavailable or where there is concern about excessive blood loss, it can be a good indicator for better oxygen management during anesthesia. Further research is needed on the use of ORi in detecting impending hypoxia in conditions of steep Trendelenburg position and pneumoperitoneum.