Introduction
The risk of postoperative pulmonary complications (PPCs) was evaluated using preoperative pulmonary function tests and risk scoring systems, such as the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) score (Supplementary Table 1) [1–8].
An oxygen reserve index (ORi) is a noninvasive measurement that ranges from 0 to 1, representing increases in the oxygen profile in areas where peripheral oxygen saturation (SpO2) is ≥ 100% (mild hyperoxia) [9–11]. ORi reflects venous oxyhemoglobin saturation that is sensitive to mild hyperoxia characterized by an arterial blood oxygen partial pressure (PaO2) ranging between 100 and 200 mmHg [12]. Although the ORi is primarily used to monitor oxygenation during surgery, its application is expanding to other areas, including sedation and one-lung ventilation (OLV) [13,14]. To date, however, few studies have focused on the use of the ORi as a tool for evaluating lung function.
We hypothesized that the degree of increase in the ORi following oxygen administration would reflect the oxygenation capacity of the lungs, thereby predicting the occurrence of PPCs [15,16]. Therefore, in this study, we aimed to investigate whether an increase in ORi during the preoxygenation phase before general anesthesia induction is associated with the occurrence of PPCs.
Materials and Methods
Written informed consent was obtained from all participants, and ethical approval was granted by the Institutional Review Board of Kyung Hee University Hospital (KHUH 2022-05-032) on June 7, 2022. This prospective observational study was conducted between September 2022 and March 2023 at the operating room of a single tertiary medical center. Data on patients’ demographic characteristics and the occurrence of PPCs were collected. The trial was conducted in accordance with the Declaration of Helsinki, 2013 and was registered with the Clinical Research Information Service (No. KCT0007700; registration date: September 16, 2022). The study protocol is available from the Clinical Research Information Service. This study complied with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist.
Participants
Adults aged 18–99 years, who were scheduled for elective surgery under general anesthesia and underwent preoperative pulmonary function tests, including the measurement of the diffusion capacity of the lungs for carbon monoxide (DLCO), were eligible for inclusion. Pregnant patients or those with severe obesity (body mass indices > 35 kg/m2) were excluded from the analysis owing to differences in their respiratory physiology and a high incidence of postoperative complications compared with other participants [17,18]. In addition, patients with tracheostomies were excluded because of differences in the respiratory drive and airway anatomy [19]. Finally, because the ORi is measured using peripheral blood flow, patients with blood circulation or peripheral blood flow disorders were also excluded.
Procedures
After entering the operating room, a Masimo Radical 7 pulse oximeter probe (Masimo Radical 7; Masimo Corp.) was attached to the fourth finger of the patient to measure the ORi. The initial heart rate, blood pressure (BP), SpO2, and ORi were recorded for all patients. After measuring initial ORi values, preoxygenation was performed for 3 min using tidal volume breathing with 100% O2 at a flow of 10 L/min, with the ORi recorded every minute. The three-minute preoxygenation ORi served as the primary variable of interest in the study. After preoxygenation, anesthesia induction, maintenance, and emergence were performed according to standard practice guidelines. During anesthesia induction, 1–2 mg/kg of propofol, 0.6–0.8 mg/kg of rocuronium, and 0.5–1 μg/kg of remifentanil were administered, followed by tracheal intubation. For maintenance of anesthesia, 3–12 mg/h/kg of propofol and 0.05–0.2 μg/min/kg of remifentanil were administered intravenously, targeting a bispectral index of 30–60. Postoperatively, 2 mg/kg sugammadex was administered, spontaneous breathing was confirmed, and tracheal extubation was performed. Intraoperative mechanical ventilation was set to the volume control mode with a fraction of inspired oxygen (FiO2) of 0.5, tidal volume of 6–8 ml/ideal body weight, respiratory rate of 10–16 breaths/min, and positive end-expiratory pressure of 5 mmHg for both lung ventilation. During OLV, the volume control mode was set and an FiO2 of 0.8, tidal volume of 5–6 ml/ideal body weight, respiratory rate of 18–20 breaths/min, and positive end-expiratory pressure of 5 mmHg were established. OLV was performed only in specific participants when required during thoracic surgery, and not in all participants. If arterial BP monitoring was required, arterial cannulation was performed based on the patient’s condition or type of surgery. Arterial blood gas analysis and ORi measurements were assessed every hour following arterial cannulation.
Outcomes
The primary outcome was the occurrence of composite PPCs on postoperative day 30. Composite PPCs were defined as the occurrence of at least one pulmonary complication, in accordance with the European Joint Task Force published guidelines for perioperative clinical outcomes [20]. Complications included respiratory infections, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm, and aspiration pneumonitis. Postoperative outcomes were evaluated via inpatient and outpatient medical chart reviews on postoperative day 30.
The correlation between the ORi and conventional pulmonary function test parameters was also evaluated. Preoperative pulmonary function tests included forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC), and DLCO. Based on these results, obstructive airway disease (predicted FEV1/FVC < 70%), restrictive airway disease (predicted FVC < 80%), and decreased diffusing capacity (predicted DLCO < 80%) were defined.
To offset the heterogeneity in the types of surgery, a subgroup analysis was conducted among patients who underwent thoracic surgery.
Statistical analysis
A pilot study comprising 28 patients was conducted using the same procedure to obtain the required sample size. ORi was measured after 3 min of preoxygenation in all 28 patients, among whom 20 patients did not develop PPCs (No-PPC group), while eight developed PPCs (PPC group). The mean ± standard deviation of the three-minute ORi was 0.69 ± 0.25 and 0.54 ± 0.28 in the No-PPC and PPC groups, respectively. Using G-power analysis (University of Düsseldorf), the required number of participants was calculated to be 138 with an α error of 0.05, a power of 0.95, and an effect size of 0.566 (t tests; means: difference between two independent means [two groups]). Considering a 10% dropout rate, the target number of participants was calculated to be 154.
This study was conducted under the hypothesis that a lower ORi during the preoxygenation phase would affect the occurrence of PPCs (H0: preoxygenation ORi in the PPC group = preoxygenation ORi in the non-PPC group; H1: preoxygenation ORi in the PPC group < preoxygenation ORi in the non-PPC group). To compare demographic differences between the non-PPC and PPC groups, a Chi-squared test was performed on categorical variables, and a normality test was performed on continuous variables, followed by the Wilcoxon rank-sum tests as a non-parametric analysis. A logistic regression analysis was used to identify factors affecting the occurrence of PPCs. Univariate logistic regression analysis was performed on variables that exhibited significant differences in patient characteristics and were previously described as clinically relevant. Multivariate logistic regression was performed to confirm whether the preoxygenation ORi remained effective after adjusting for covariates. All variables with P < 0.2 in the univariate logistic regression analysis and previously described clinically relevant factors were included in the multivariate logistic regression analysis. Covariates were adjusted for sex, American Society of Anesthesiologists physical status class, surgical department, and ARISCAT class. Multicollinearity among the variables was evaluated via a generalized variance inflation factor (GVIF), with a GVIF < 2 indicating no multicollinearity. A receiver operating characteristic (ROC) curve analysis was conducted to determine the cutoff value of the ORi for its association with the occurrence of PPCs using the maximum Youden index (sensitivity [%] + specificity [%] – 100).
A post hoc power analysis (z tests and logistic regression) was performed based on the data obtained in the current study. A subgroup analysis was conducted for patients who underwent thoracic surgery. All statistical analyses were performed using SPSS software (version 22.0; IBM Corp.). Differences were considered statistically significant at P values < 0.05.
Results
Study participants and patient characteristics
We screened 211 patients, 57 of whom were excluded from and 154 of whom were included in the study (Fig. 1). Among the enrolled patients, 78 did not develop PPCs (No-PPC group), while 76 developed PPCs (PPC group) (Table 1 and Supplementary Fig. 1). No significant differences were found between the two groups in terms of demographic data or preoperative pulmonary function test results (Table 1 and Supplementary Table 2). Although one-minute preoxygenation ORi values did not differ significantly between the two groups, two- and three-minute preoxygenation ORi values were significantly higher in the No-PPC group than in the PPC group (Table 1 and Fig. 2). The ARISCAT scores were significantly higher in the PPC group (Table 1). The postoperative hospital length of stay was significantly longer in the PPC group than in the No-PPC group (Table 1).
Outcomes
Univariate logistic regression analysis revealed that the two- and three-minute preoxygenation ORi, thoracic surgery, and ARISCAT class (intermediate or high) were all significantly associated with the occurrence of PPCs (Table 2), while multivariate logistic regression analysis showed that three-minute preoxygenation ORi, thoracic surgery, and ARISCAT class (intermediate or high) were independently associated with the occurrence of PPCs (Table 2 and Supplementary Fig. 2). According to the estimated GVIFs (< 2.0), multicollinearity had a minimal impact on the results (Table 2). Considering its clinical utility, multivariable logistic regression analysis was performed using the same variables to evaluate the effect of a 0.1 increase in the ORi, with the odds ratio (OR) calculated as 0.69 (95% CI [0.56–0.83], P < 0.001) (Supplementary Table 3).
After a ROC curve analysis of PPCs occurrence, the area under the curve (AUC) of the two- and three-minute preoxygenation ORi and ARISCAT scores were 0.62, 0.64, and 0.71, respectively (Fig. 3A and Supplementary Table 4). In this study, the occurrence of PPCs was not associated with conventional pulmonary function test parameters, such as FEV1, FVC, and DLCO (Fig. 3B and Supplementary Table 4).
The intraoperative correlation coefficient between the PaO2 and ORi was r = 0.70 (P < 0.001) in 442 pairs of 130 patients (Supplementary Fig. 3). No significant correlation was observed between the three-minute preoxygenation ORi and conventional pulmonary function test parameters (Supplementary Fig. 4), nor between preoperative hemoglobin levels and the three-minute preoxygenation ORi (Supplementary Fig. 5).
Subgroup analysis
A subgroup analysis was performed on patients who underwent thoracic surgery. The one-minute preoxygenation ORi did not significantly differ between the two groups (No-PPC group: 0.52 [0.00–0.67]; PPC group: 0.40 [0.13–0.51], P = 0.260), whereas the two-minute preoxygenation ORi (No-PPC group: 0.68 [0.62–0.86]; PPC group: 0.61 [0.50–0.73], P = 0.049, Fig. 4A) and three-minute preoxygenation ORi (No-PPC group: 0.86 [0.68–1.00]; PPC group: 0.63 [0.53–0.76], P = 0.006, Fig. 4B) were significantly higher in the No-PPC group. Univariate logistic regression analysis based on the occurrence of PPCs revealed that the three-minute preoxygenation ORi, body mass index, and ARISCAT class (intermediate) were significantly associated with the occurrence of PPCs (Table 3). Multivariate logistic regression analysis showed that the three-minute preoxygenation ORi was independently associated with the occurrence of PPCs (Table 3 and Supplementary Fig. 6).
ROC curve analysis, the AUCs of the two- and three-minute preoxygenation ORi, and ARISCAT scores were 0.65, 0.72, and 0.56, respectively (Fig. 4C and Supplementary Table 5). During OLV, the correlation coefficient between PaO2 and ORi was r = 0.87 (P < 0.001) for 93 pairs of 68 patients (Fig. 4D).
Post hoc power analysis
Based on the data obtained in this study, a post hoc power analysis was performed to determine the effect of the three-minute preoxygenation ORi on the occurrence of PPCs. Based on 78 patients in the non-PPC group and 76 patients in the PPC group, with an effect size of 0.48 for the three-minute preoxygenation ORi and an α error of 0.05, the power of the study was calculated to be 0.90.
Discussion
In this study, we found that a low ORi after 3 min of preoxygenation before general anesthesia was significantly associated with the occurrence of PPCs. The increase in the ORi following preoxygenation can be assumed to reflect the gas exchange capacity of the lungs, and a failure to show a substantial increase in ORi may indicate a higher risk of PPCs. Moreover, a low ORi during the preoxygenation phase was found to be more strongly associated with PPCs than traditional lung function tests. This consistent result was also observed in the subgroup analysis of patients who underwent thoracic surgery. These findings suggest that the ORi may serve as a useful tool for predicting PPCs, even among patients who underwent thoracic surgery.
Analysis of a subgroup of patients who underwent thoracic surgery showed that the ARISCAT score was not associated with the occurrence of PPCs. Previous studies have highlighted the limitation of the ARISCAT score in patients undergoing lung resection [8,21,22]. In the ARISCAT scoring system, surgical incision sites for intrathoracic and preoperative SpO2 ≤ 90% are assigned the highest score of 24 points (Supplementary Table 1) [8]. Consequently, patients who underwent thoracic surgery had a minimum ARISCAT score of 24 points because of the surgical incision site. In the ROC curve analysis, the cutoff value was 26 points for all patients and 40 points for patients undergoing thoracic surgery. As such, when applied to patients undergoing thoracic surgery, the discriminatory power of the ARISCAT score approach between patients was weak, resulting in a low AUC for the occurrence of PPCs. In patients undergoing thoracic surgery, a low ORi during the preoxygenation phase, rather than ARISCAT scores, may be associated with the occurrence of PPCs. In addition, the ARISCAT score requires an evaluation of the duration of surgery, and an accurate score can only be calculated postoperatively. However, the ORi can be measured preoperatively, and preoxygenation ORi is independent of the surgical type or duration. Measurement of the preoxygenation ORi may therefore aid in postoperative patient management, such as ventilator use or intensive care unit admission.
Although conventional pulmonary function tests were not associated with the occurrence of PPCs, a low ORi during the preoxygenation phase was significantly associated with the occurrence of PPCs. The ORi, which is comparable to applying the method of SpO2, represents the degree of oxygenation, which is typically monitored through the patient’s finger. However, SpO2 reflects only the range where PaO2 is approximately < 100 mmHg [23], whereas ORi quantifies mild hyperoxia within the PaO2 range of 100–200 mmHg [24]. The ORi that increases when oxygen is provided to the patient may be affected by all three categories of lung function assessment as follows [25]: respiratory mechanics, lung parenchyma, and cardiopulmonary interactions. Therefore, the ORi values taken after oxygen administration may represent overall lung function.
Several previous studies have reported a PPC incidence ranging from 5% to 40% [26–28]. In the present study, the incidence of PPCs was relatively high, at 49% (76/154). This may be attributed to the inclusion of patients who underwent thoracic (68/154 [44%]) or upper abdominal surgeries (16/154 [10%]) (Table 1 and Supplementary Table 2), both of which were associated with a higher incidence of PPCs. Additionally, to examine correlations between the preoxygenation ORi and conventional pulmonary function test parameters, the present study was conducted in patients who underwent conventional preoperative pulmonary function tests. These tests were performed at the discretion of the surgical department or pulmonologist depending on the patient’s age, type of surgery, and medical history. Because pulmonary function tests are not routinely performed for all surgical patients, the study cohort may have been more vulnerable to PPCs. The median ARISCAT score of all participants was 27 points, corresponding to the intermediate-risk category of the ARISCAT class. The high-risk category accounted for 20% of the cohort that also explains the high incidence of PPCs observed in this study.
Preoxygenation is typically performed for either 3 min of tidal volume breathing, eight deep breaths, or until end tidal O2 reaches a plateau [29,30]. In this study, an increase in ORi was observed during 3 min of tidal volume breathing in the preoxygenation phase. Most of the participants’ ORi values plateaued between 2 and 3 min. Moreover, a stronger association was detected between the three-minute preoxygenation ORi, measured after reaching a plateau, and the occurrence of PPCs than the two-minute preoxygenation ORi. Confirming the ORi plateau during the preoxygenation phase is therefore crucial for assessing maximum oxygenation status.
The correlation between the ORi and PaO2 has been investigated in previous research; for example, Applegate et al. [31] reported a correlation coefficient of 0.732 for PaO2 < 240 mmHg and noted that the correlation was not significant for PaO2 > 240 mmHg. Similarly, Fadel et al. [32] reported a correlation coefficient of 0.8193 for a PaO2 of 100–200 mmHg. Koishi et al. [33] reported a correlation coefficient of 0.8191 between the ORi and PaO2 during OLV. The ORi was designed to be particularly sensitive within a PaO2 range of 100–200 mmHg. During both lung ventilations, PaO2 frequently exceeded 200 mmHg. The results of the present study are thus consistent with previous findings because the correlation coefficient between the ORi and PaO2 during OLV (characterized by relatively mild hyperoxia) in the subgroup analysis was higher than that in the overall participant cohort. In addition, the correlation coefficients observed in the current study were comparable to those reported in previous studies. Because PaO2 levels during OLV were closer to the range in which ORi was the most sensitive, the correlation coefficient was higher, suggesting that the ORi is likely to be particularly useful in clinical practice during OLV.
This study had several limitations. First, tests that evaluate cardiopulmonary interactions, such as the six-minute walking test, were not performed due to time and cost constraints. Thus, further studies comparing the six-minute walking test and ORi in association with the occurrence of PPCs should be conducted. Second, this study did not include a low-risk patient group and the types of surgeries were heterogenous; bias could have been minimized if the study had been conducted using a single surgery type or patient group. However, we elected to include all patients who underwent preoperative pulmonary function testing to explore associations between ORi and the occurrence of PPCs, as well as its potential generalizability to various patient groups. We attempted to overcome this limitation through a subgroup analysis of participants who underwent thoracic surgery. Further research involving a wider range of patients is therefore needed. Third, we could not compare the ORi with PaO2 using arterial blood gas analysis in all participants. However, previous studies have shown a correlation between ORi and PaO2. Considering the risks and benefits of the procedure, additional arterial blood gas analysis was not performed because of its invasive nature. When arterial blood gas analysis was clinically indicated, PaO2 was measured and its correlation with ORi was analyzed. Fourth, the severity of PPCs was not evaluated [34]. Because research on the association between the ORi and occurrence of PPCs remains insufficient, assessment of the severity of PPCs would be premature. By evaluating the occurrence of composite PPCs, we observed a higher incidence; but additional research is needed that focuses on what influence ORi has on the severity of these complications. Lastly, the normal ORi range is unknown; thus, ORi could be more readily applied in practice if standardized ORi ranges (similar to that of SpO2) were to be established for different clinical scenarios. Further research is needed to identify the normal preoxygenation ORi range in healthy individuals.
In conclusion, a low ORi measured after 3 min of preoxygenation for general anesthesia was significantly associated with an increased risk of PPCs in surgical patients, including those undergoing thoracic surgery. This study demonstrated the potential of the ORi measured after oxygen administration as a valuable tool for assessing lung function that complements traditional lung function tests and scoring systems. To confirm the results of this study, larger multicenter studies across various surgical procedures are warranted.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
