SpO2/FiO2 ratio: an estimation of PaO2/FiO2 ratio in invasive mechanically ventilated patients in high altitude areas Authors: Hongmei Yang, Yifeng Chen, Wenlu Wang, Chaoyong Zhou, Xiaolin Zhu, Tianming Cun, Chengyang He, Yaowu Chen Published: 2026 Dear Editor, Acute respiratory distress syndrome (ARDS) is a critical illness frequently encountered in intensive care units (ICUs), characterized by high morbidity and mortality rates.1 Its core pathophysiological change is oxygenation impairment due to diffuse alveolar damage, necessitating reliable indicators for clinical assessment of oxygenation function.2 The arterial oxygenation index (arterial partial pressure of oxygen (PaO2)/inspired oxygen fraction (FiO2) ratio), which reflects the relationship between PaO2 and FiO2, has been incorporated into the Berlin ARDS diagnostic criteria since 2011. It serves as a cornerstone for classifying disease severity, guiding treatment strategies, and evaluating clinical outcomes in patients with acute respiratory failure.3 Trends in the PaO2/FiO2 ratio not only support clinical decision-making but also play a pivotal role in recent clinical research and trials. Blood gas analysis requires measuring PaO2 and calculating the PaO2/FiO2 ratio.4 However, this procedure has limitations, particularly for ICU patients requiring repeated dynamic monitoring. It is cumbersome, increases risks of infection and bleeding, and imposes significant nursing burdens and healthcare costs. Moreover, in resource-limited settings such as high-altitude regions, frequent arterial blood sampling is practically challenging due to limited medical equipment, technical personnel, and patient conditions. Consequently, exploring non-invasive, reliable, and reproducible alternatives to the PaO2/FiO2 ratio has become a research priority. Previous studies have suggested that the ratio of pulse oxygen saturation (SpO2) to FiO2 (SpO2/ FiO2 ratio) can be used to assess and categorize the severity of respiratory failure without the need for arterial blood sampling. Early research demonstrated a strong linear correlation between the SpO2/FiO2 and PaO2/FiO2 ratios when SpO2 is below 97%. Current methods for estimating PaO2 from SpO2 include simulations of the hemoglobin dissociation curve and linear correlations between SpO2/FiO2 and PaO2/FiO2 ratios, primarily compared using Rice and Ellis equations.567 However, most existing studies are based on data from sea-level or low-altitude populations, and their applicability in special geographic environments, particularly high-altitude regions, lacks systematic evaluation. In high-altitude settings, reduced atmospheric pressure, lower oxygen partial pressure, and chronic hypoxiainduced compensatory mechanisms significantly alter SpO2 interpretation and oxygen transport dynamics, potentially leading to biases in PaO2/ FiO2 ratio estimates derived from conventional algorithms. Furthermore, most prior research focused on spontaneously breathing or noninvasively ventilated patients, while the correlation and agreement between SpO2/FiO2 and PaO2/ FiO2 ratios in critically ill hypoxemic patients requiring invasive mechanical ventilation remain underexplored. Our preliminary studies in high-altitude regions showed significant correlations (r = 0.75–0.90) between estimated PaO2/FiO2 ratios (using these equations) and measured values, but this was limited to non-invasively ventilated patients.89 Thus, systematically investigating the correlation and agreement between SpO2/FiO2 ratios, estimated PaO2/FiO2 ratios, and directly measured PaO2/FiO2 ratios in critically ill hypoxemic patients undergoing invasive mechanical ventilation at high altitudes holds critical clinical value. Additionally, evaluating the performance disparities among estimation methods and the accuracy of SpO2/ FiO2 ratios in identifying and diagnosing severe hypoxemia in this specific population will provide essential insights for oxygenation monitoring and clinical decision-making in high-altitude critical care settings. Patients receiving invasive mechanical ventilation support in the ICU of Lijiang Municipal People’s Hospital were selected. A combined retrospective-prospective cohort study was conducted at a single center to maximize data collection while maintaining methodological consistency. The study design leveraged the stable clinical environment of our ICU, where protocols, equipment, and personnel remained consistent throughout the study period. Data from hospitalized patients between June 2022 and May 2023 were first collected retrospectively, followed by prospective data collection from June 2023 to April 2024. Based on Rice equation, SpO2/FiO2 = 64 + 0.84 × PaO2/FiO2, the measured SpO2 was substituted into the formula to calculate SpO2/FiO2 ratio, yielding an estimated PaO2/FiO2 ratio. Using Ellis equation, SpO2 was substituted into the formula to calculate PaO2, thereby deriving an estimated PaO2/FiO2 ratio.56 A total of 487 patients were enrolled during the study period. Patients were stratified into two groups based on positive end-expiratory pressure levels: positive end-expiratory pressure < 8 cmH2O (1 cmH2O = 0.0981 kPa) (n = 257) and positive end-expiratory pressure ≥ 8 cmH2O (n = 230). Baseline characteristics are summarized in Additional Table 1. The majority of patients in both groups were female (65.0% vs. 67.8%), with median age of 62 (50, 73) years and a median Sequential Organ Failure Assessment score of 6 (5, 8). No statistically significant differences were observed in age, weight, or SpO2 between groups (all P > 0.05). However, significant differences were noted in PaO2, FiO2, measured PaO2/FiO2 ratio, SpO2/FiO2 ratio, Rice equation-estimated PaO2/FiO2 ratio, Ellis equation-estimated PaO2/FiO2 ratio, Sequential Organ Failure Assessment score, duration of mechanical ventilation, and hospital length of stay (all P < 0.005). Additional Table 1Demographic and clinical dataTotal (n=487)PEEP < 8 cmH2O (n=257)PEEP ≥ 8 cmH2O (n=230)PAge (yr)62 (50−73)62 (50−73)62 (50−71)0.125Sex (female)323 (66.3)167 (65)156 (67.8)0.507Height (cm)168 (162−170)168 (164−170)168 (162−170)< 0.001Weight (kg)62 (55−70)60 (50−70)63 (58−70)0.198FiO2 (%)60 (50−60)60 (41−60)60 (60−65)< 0.000SpO2 (%)97 (93−99)97 (93−100)97 (94−99)0.154PaO2 (mmHg)100 (77−148)100 (76−140)102 (78−155)< 0.001Directly measured PaO2/FiO2 (mmHg)186 (131−260)205 (145−264)170 (125−257)< 0.001Rice equation-estimated PaO2/FiO2 (mmHg)120 (106−152)122 (112−176)118 (90−122)< 0.001Ellis equation-estimated PaO2/FiO2 (mmHg)126 (112−174)138 (125−181)126 (104−164)< 0.001SpO2/FiO2 ratio165 (153−192)167 (158−212)163 (140−167)< 0.001Mechanical ventilation duration (d)5 (3−8)5 (3−8)5 (3−9)< 0.001Days of hospital stay6 (4−11)6 (4−10)7 (3−11)< 0.001SOFA score6 (5−8)6 (5−7)6 (5−8)< 0.001The PEEP threshold of 8 cmH2O (1 cmH2O = 0.0981 kPa) was selected based on: (1) clinical practice patterns where this level represents a transition point toward more intensive respiratory support; (2) physiological considerations regarding significant alveolar recruitment effects at higher PEEP levels; (3) alignment with published validation studies of oxygenation indices that used similar thresholds to stratify lung injury severity; and (4) consistency with institutional protocols for managing patients with varying degrees of respiratory failure. Continuous variables following a normal distribution are presented as mean ± standard deviation and were analyzed with independent-samples t-test; skewed continuous variables are expressed as median (interquartile range) and were analyzed with the Mann−Whitney U test. FiO2: Fraction of inspired oxygen; PaO2: partial pressure of arterial oxygen; PEEP: positive end-expiratory pressure; SOFA: sequential organ failure assessment; SpO2: speripheral capillary pulse oximetry oxygen saturation. Demographic and clinical data The PEEP threshold of 8 cmH2O (1 cmH2O = 0.0981 kPa) was selected based on: (1) clinical practice patterns where this level represents a transition point toward more intensive respiratory support; (2) physiological considerations regarding significant alveolar recruitment effects at higher PEEP levels; (3) alignment with published validation studies of oxygenation indices that used similar thresholds to stratify lung injury severity; and (4) consistency with institutional protocols for managing patients with varying degrees of respiratory failure. Continuous variables following a normal distribution are presented as mean ± standard deviation and were analyzed with independent-samples t-test; skewed continuous variables are expressed as median (interquartile range) and were analyzed with the Mann−Whitney U test. FiO2: Fraction of inspired oxygen; PaO2: partial pressure of arterial oxygen; PEEP: positive end-expiratory pressure; SOFA: sequential organ failure assessment; SpO2: speripheral capillary pulse oximetry oxygen saturation. The correlation between directly measured PaO2/ FiO2 ratios, Rice/Ellis equation-estimated PaO2/ FiO2 ratios, and SpO2/FiO2 ratios was assessed using intraclass correlation coefficient analysis. A strong correlation was observed (intraclass correlation coefficient = 0.76, P < 0.005, 95% confidence interval 0.73–0.79). In patients with severe hypoxemia (PaO2/FiO2 ratio ≤ 150), different estimation methods and the SpO2/FiO2 ratio demonstrated comparable diagnostic accuracy, with areas under the receiver operating characteristic curve (AUC) exceeding 0.70 (Figure 1 and Additional Table 2). The predictions of diagnostic performance using different estimation methods and SF in patients with severe hypoxemia (SF ≤ 150 mmHg). The performance of each model was evaluated by the area under the curve. ROC: Receiver operating characteristic; SF: the ratio of pulse oximetry (SpO2) to inspiratory oxygen partial pressure (FiO2). Additional Table 2Association between Rice equation-estimated PaO2/FiO2 ratio, Ellis equation-estimated PaO2/FiO2 ratio, SpO2/FiO2 ratio, and severe hypoxemia (PaO2/FiO2 ≤ 150 mmHg)AUC95% CICut-off valueSensitivitySpecificity P Rice equation-estimated PaO2/FiO2 ratio0.810.77-0.85119.2450.8320.693< 0.001Ellis equation-estimated PaO2/FiO2 ratio0.710.66-0.76117.1550.5480.852< 0.001SpO2/FiO2 ratio0.730.68-0.78156.5350.5420.88< 0.001The ROC curves demonstrate the predictive performance of three non-invasive methods, Rice equation, Ellis equation, and the SpO2/FiO2 ratio, for identifying severe hypoxemia, defined as a directly measured PaO2/FiO2 ratio ≤ 150 mmHg. The analysis included 487 mechanically ventilated patients. The area under the curve for Rice equation-estimated PaO2/FiO2 ratio (0.81) was higher than that of both Ellis equation (0.71) and the SpO2/FiO2 ratio (0.73), suggesting superior discriminatory ability. AUC: Area under the ROC curve; CI: confidence interval; ROC: receiver operating characteristic. Association between Rice equation-estimated PaO2/FiO2 ratio, Ellis equation-estimated PaO2/FiO2 ratio, SpO2/FiO2 ratio, and severe hypoxemia (PaO2/FiO2 ≤ 150 mmHg) The ROC curves demonstrate the predictive performance of three non-invasive methods, Rice equation, Ellis equation, and the SpO2/FiO2 ratio, for identifying severe hypoxemia, defined as a directly measured PaO2/FiO2 ratio ≤ 150 mmHg. The analysis included 487 mechanically ventilated patients. The area under the curve for Rice equation-estimated PaO2/FiO2 ratio (0.81) was higher than that of both Ellis equation (0.71) and the SpO2/FiO2 ratio (0.73), suggesting superior discriminatory ability. AUC: Area under the ROC curve; CI: confidence interval; ROC: receiver operating characteristic. To evaluate the practical utility of the SpO2/FiO2 ratio, we compared the ranges of PaO2/FiO2 ratios calculated using two methods (Rice and Ellis equations) with the range of directly measured arterial blood gas PaO2/FiO2 ratios, as summarized in Additional Table 3. Additional Table 3Comparison of PaO2/FiO2 ratios calculated via different methods with directly measured valuesSpO2/FiO2 ratioDirectly measured PaO2/FiO2 ratio (mmHg)Rice equation-estimated PaO2/FiO2 ratio (mmHg)Ellis equation-estimated PaO2/FiO2 ratio (mmHg)> 300203−449288−439194−496250−299205−270221−273182−274200−249163−238162−218111−330150−199136−213102−160104−264100−14975−13543−10183−121Stratification was based on the severity of SpO2/FiO2 ratio, which was measured directly (line 1: normal/mild; lines 2−5: moderate to severe). Data are presented as interquartile range (25th to 75th percentile). Comparison of PaO2/FiO2 ratios calculated via different methods with directly measured values Stratification was based on the severity of SpO2/FiO2 ratio, which was measured directly (line 1: normal/mild; lines 2−5: moderate to severe). Data are presented as interquartile range (25th to 75th percentile). This study found correlations between PaO2/ FiO2 ratios estimated via two methods, SpO2/FiO2 ratios, and directly measured arterial blood gas PaO2/FiO2 ratios. The non-invasive nature of these indices holds significant clinical value for assessing oxygenation status in invasively ventilated patients, particularly in resource-constrained high-altitude ICUs where arterial blood sampling is challenging. SpO2/FiO2 ratio emerges as a reliable alternative for dynamic hypoxemia monitoring in such settings. Critically, both estimation methods demonstrated excellent diagnostic performance for severe hypoxemia (PaO2/FiO2 ratio ≤ 150 mmHg), offering a practical tool for high-altitude critical care. Prior studies have reported high concordance between non-invasive oxygenation indices and arterial blood gas-derived PaO2/FiO2 ratio. A previous study reported correlation coefficients ranging from 0.73 to 0.88 using their estimation formula, aligning with our findings (intraclass correlation coefficient > 0.76).9 However, the accuracy of non-invasive indices may be influenced by factors such as positive end-expiratory pressure levels, hemoglobin concentration, total bilirubin, determinants of hemoglobin-oxygen affinity (e.g., pH, temperature, carbon dioxide, 2,3-diphosphoglycerate, fetal hemoglobin), and extreme arterial saturation values—variables that often fluctuate in critically ill patients.4910 Against this backdrop, our study reinforces the applicability of SpO2/FiO2 ratio in high-altitude environments, where low ambient pressure and exaggerated alveolar–arterial oxygen gradients could theoretically compromise accuracy. Even under these conditions, SpO2/FiO2 ratio reliably reflects oxygenation status in invasively ventilated patients. Theoretically, hemoglobin dissociation curve-based estimation methods might offer superior precision in hypoxemia diagnosis. Three studies observed enhanced diagnostic performance with Ellis equation for moderate-to-severe hypoxemia.6711 However, our data show comparable diagnostic capabilities between estimation methods, with minimal variation when including patients with SpO2 > 97%. This discrepancy may stem from differences in study populations and environmental pressures. Our focus was on severe hypoxemia (not hyperoxia), where SpO2 values are less constrained by saturation limits.10 Furthermore, as a continuous, dynamic parameter, SpO2/FiO2 ratio provides unique advantages for real-time oxygenation trend monitoring, offering a simpler and more sensitive reference for ventilator parameter adjustments in ARDS management. Invasive mechanical ventilation operates within a closed circuit but is not pressure-limited; thus, patients ventilated at high altitudes often exhibit lower PaO2 and SpO2 values compared to sea-level counterparts. A previous study reported SpO2/ FiO2 ratios of 214–235 corresponding to arterial PaO2/FiO2 ratios of 200 in low-altitude residents.8 However, no studies have defined an universal SpO2/FiO2 ratio threshold corresponding to a given PaO2/FiO2 ratio in mechanically ventilated patients at altitudes exceeding 1500 meters. This gap complicates oxygenation assessment and introduces significant error margins in estimation methods, undermining the accuracy of hypoxemia diagnosis and treatment in high-altitude settings. To address this, we propose an equivalence table linking SpO2/FiO2 ratio ranges with estimated PaO2/FiO2 values (via both equations) and directly measured PaO2/FiO2 ratios. It may aid in diagnosing and following patients with varying degrees of hypoxemia. Notably, Orti et al.6 observed comparable diagnostic performance between SaO2/FiO2 ratios and estimated PaO2/FiO2 ratios for severe hypoxemia in high-altitude ventilated patients across diverse pathologies, supporting the feasibility of noninvasive oxygenation indices in high-altitude care. Our equivalence table further refines SpO2/FiO2PaO2/FiO2 relationships at high altitudes, providing specific quantitative criteria for clinical hypoxemia assessment and foundational data for developing altitude-adjusted ARDS diagnostic criteria. Certainly, this study has certain limitations. Firstly, single arterial blood gas sampling may introduce random errors, affecting the accuracy of hypoxemia diagnosis. Future studies could enhance data quality by increasing sampling frequency or integrating continuous monitoring technologies. Secondly, retrospective data collection may lead to information bias, which could be mitigated through prospective cohort studies. Lastly, while the sample includes high-altitude populations, it does not cover all altitude gradients. Subsequent research should expand the geographic scope to refine the assessment of oxygenation indices across different elevations. Conclusion: This study demonstrates significant correlations and strong predictive performance between SpO2/FiO2 ratios and PaO2/FiO2 ratios in invasively ventilated patients at high altitudes. SpO2/FiO2 ratio serves as a valid surrogate for PaO2/FiO2 ratio, offering a simple and reliable method for oxygenation monitoring in high-altitude clinical settings. This finding may enhance the management and outcomes of mechanically ventilated patients in high-altitude regions. The diagnostic efficacy of SpO2/FiO2 ratios and estimated PaO2/FiO2 ratios is comparable in severe hypoxemia, though further prospective studies are warranted to explore SpO2/FiO2 ratio applicability across diverse pathologies and validate its stability in larger, multicenter cohorts. This study complied with medical ethics standards and was reviewed and approved by the Medical Ethics Committee of Lijiang Municipal People’s Hospital (approval No. 2022048; dated October 26, 2022). Data collection was performed after obtaining informed consent from the patients or their next of kin. This study was supported by the National Clinical Key Specialty Project of Chinese Emergency Medicine (No. Z155080000004) and Science and Technology Bureau of Lijiang City (No. 2022LJSHFZ016). Additional files Additional Table 1: Demographic and clinical data. Additional Table 2: Association between Rice equation-estimated PaO2/FiO2 ratio, Ellis equation-estimated PaO2/FiO2 ratio, SpO2/FiO2 ratio, and severe hypoxemia (PaO2/FiO2 ≤ 150 mmHg). Additional Table 3: Comparison of PaO2/FiO2 ratios calculated via different methods with directly measured values.