key: cord-0881044-i0wpew1s
authors: Kim, Jin Hyoung; Baek, Ae-Rin; Lee, Song-I; Kim, Won-Young; Na, Yong Sub; Lee, Bo Young; Seong, Gil Myeong; Baek, Moon Seong
title: ROX index and SpO(2)/FiO(2) ratio for predicting high-flow nasal cannula failure in hypoxemic COVID-19 patients: A multicenter retrospective study
date: 2022-05-12
journal: PLoS One
DOI: 10.1371/journal.pone.0268431
sha: a72324ca488f92d2e70767a1c0f248ce22b97973
doc_id: 881044
cord_uid: i0wpew1s

BACKGROUND: The ratio of oxygen saturation (ROX) index, defined as the ratio of oxygen saturation (SpO(2))/fraction of inspired oxygen (FiO(2)) to respiratory rate, can help identify patients with hypoxemic respiratory failure at high risk for intubation following high-flow nasal cannula (HFNC) initiation; however, whether it is effective for predicting intubation in coronavirus disease 2019 (COVID-19) patients receiving HFNC remains unknown. Moreover, the SpO(2)/FiO(2) ratio has been assessed as a prognostic marker for acute hypoxemic respiratory failure. This study aimed to determine the utility of the ROX index and the SpO(2)/FiO(2) ratio as predictors of failure in COVID-19 patients who received HFNC. METHODS: This multicenter study was conducted in seven university-affiliated hospitals in Korea. Data of consecutive hospitalized patients diagnosed with COVID-19 between February 10, 2020 and February 28, 2021 were retrospectively reviewed. We calculated the ROX index and the SpO(2)/FiO(2) ratio at 1 h, 4 h, and 12 h after HFNC initiation. The primary outcome was HFNC failure defined as the need for subsequent intubation despite HFNC application. The receiver operating characteristic curve analysis was used to evaluate discrimination of prediction models for HFNC failure. RESULTS: Of 1,565 hospitalized COVID-19 patients, 133 who received HFNC were analyzed. Among them, 63 patients (47.4%) were successfully weaned from HFNC, and 70 (52.6%) were intubated. Among patients with HFNC failure, 32 (45.7%) died. The SpO(2)/FiO(2) ratio at 1 h after HFNC initiation was an important predictor of HFNC failure (AUC 0.762 [0.679–0.846]). The AUCs of SpO(2)/FiO(2) ratio at 4 h and ROX indices at 1 h and 4 h were 0.733 (0.640–0.826), 0.697 (0.597–0.798), and 0.682 (0.583–0.781), respectively. Multivariable analysis showed that the patients aged ≥70 years are 3.4 times more likely to experience HFNC failure than those aged <70 years (HR 3.367 [1.358–8.349], p = 0.009). The SpO(2)/FiO(2) ratio (HR 0.983 [0.972–0.994], p = 0.003) at 1 h was significantly associated with HFNC failure. CONCLUSIONS: The SpO(2)/FiO(2) ratio following HFNC initiation was an acceptable predictor of HFNC failure. The SpO(2)/FiO(2) ratio may be a good prognostic marker for predicting intubation in COVID-9 patients receiving HFNC.

a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

High-flow nasal cannula (HFNC) has been widely used in patients with hypoxemic respiratory failure. HFNC could reduce the rate of endotracheal intubation in patients with acute respiratory failure compared with conventional oxygen therapy [1] [2] [3] . However, failure of HFNC may cause delayed intubation and increased mortality [4] . Therefore, predicting HFNC failure and determining the appropriate timing of endotracheal intubation are important strategies for HFNC treatment.

The ratio of oxygen saturation (ROX) index, defined as the ratio of oxygen saturation (SpO 2 )/fraction of inspired oxygen (FiO 2 ) to respiratory rate, has been proposed to detect HFNC failure [5] . Roca et al. [6] reported that an ROX index cut-off of 4.88 measured 12 h after HFNC initiation was associated with a lower risk of need for intubation, suggesting that the ROX index can help identify patients at high risk for intubation. However, whether the ROX index is effective for predicting intubation and whether the cut-off of the ROX index is appropriate in patients with COVID-19 remain unknown.

The SpO 2 /FiO 2 (SF) ratio, which correlates with the partial pressure of oxygen PaO 2 /FiO 2 (PF) ratio [7] , can also be utilized as a prognostic marker for acute hypoxemic respiratory failure [8] . Notably, the SF ratio has been reported as a reliable factor for predicting failure of HFNC [9] or non-invasive ventilation [10] in clinical practice when arterial blood gas sampling is not readily available. Moreover, unlike traditional respiratory failure, COVID-19 often presents with silent hypoxia, a condition where the patient has no abnormal respiratory pattern despite severe hypoxia [11] ; therefore, respiratory rate may not be crucial in predicting HFNC failure in COVID- 19 . Hence, we hypothesized that the SF ratio had a comparable prediction ability with the ROX index in patients with severe COVID-19. We aimed to determine the predictive function of the ROX index and the SF ratio in COVID-19 patients after HFNC initiation. Furthermore, we evaluated the factors affecting mortality in these patients.

This multicenter study was conducted at seven university-affiliated hospitals in Korea with nationally designated isolation units. Patients with confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection were classified based on severity, and patients with moderate to severe infections were admitted to nationally designated treatment facilities. Consecutive hospitalized patients diagnosed with SARS-CoV-2 infection between February 10, 2020, and February 28, 2021, were retrospectively reviewed. We analyzed the electronic medical records of patients aged �18 years who received HFNC. As we enrolled patients who received HFNC in the advanced state of the disease, those who received the therapy following discontinuation of mechanical ventilation were not included in the study group. Patients were excluded if they did not receive HFNC, were intubated after 72 hours from HFNC weaning, or had a do-not-intubate order.

This study was approved by the Institutional Review Board of the Chung-Ang University Hospital (approval number 2103-009-19360) and the local institutional review boards of all other participating centers. Informed consent was waived owing to the retrospective nature of the study.

The following data were collected from electronic medical records of patients with SARS--CoV-2 infection: age, sex, smoking, symptom at admission, body mass index, Charlson comorbidity index, comorbidities, CURB-65 (Confusion, urea nitrogen, respiratory rate, blood pressure, age 65 � years), SOFA (sequential organ failure assessment) score, APACHE (acute physiology and chronic health evaluation) II score, vital signs at admission, duration of fever, laboratory findings, initial chest x-ray findings (normal or unilateral vs. bilateral or multifocal), treatment during hospitalization (remdesivir, antibiotics, antifungal agents, vasopressor, continuous renal replacement therapy, and corticosteroids), length of hospital stay, PF ratio at HFNC application, HFNC duration, and in-hospital mortality.

The decision to initiate HFNC or mechanical ventilation was made by the treating clinical team at each center. According to the Surviving Sepsis Campaign: guidelines on the management of critically ill adults with COVID-19 [12] , HFNC is encouraged to use in acute hypoxemic respiratory failure despite conventional oxygen therapy. The Korean Society of Critical Care Medicine published COVID-19 treatment guidelines indicating mechanical ventilation in the following circumstances [13] :

(1) Persistent or recurrent hypoxia: SpO 2 <92% / PaO 2 <60 mmHg (2) Respiratory rate: 30-35 breaths/min (3) Hemodynamic instability (4) Loss of consciousness/mind alteration (5) Treatment failure of HFNC or conventional oxygen devices (6) Cardiac arrest or respiratory arrest (7) Other: the decision of the physician.

Based on the guidelines, the treating clinical team determined the decision of intubation considering the composite assessment of signs of acute respiratory failure. We calculated the ROX index and SpO 2 /FiO 2 ratio at 1 h, 4 h, and 12 h after HFNC application. Primary outcome was failure of HFNC, which was defined as the need for subsequent intubation and use of mechanical ventilation despite HFNC application [14] .

(interquartile range), and Mann-Whitney U test was used for between-group comparisons. The receiver operating characteristic (ROC) curve analysis was used to evaluate discrimination of prediction models for HFNC failure. The sensitivity and specificity of the scores were determined, and the cutoff point corresponded to the maximum of the Youden's index. Univariate and multivariable logistic regression analyses were performed using the forward selection method to identify the factors associated with HFNC failure and in-hospital mortality.

Candidate variables for inclusion in the multivariable regression model were variables with p < 0.1 in the univariate analyses and clinical parameters. Calibrations of the models were evaluated using the Hosmer-Lemeshow goodness-of-fit test. P-values of < 0.05 were considered statistically significant. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 26.0 (IBM Corporation, Armonk, NY, USA).

A total of 1,565 patients with SARS-CoV-2 infection were admitted in seven hospitals, 488 of which received oxygen therapy (Fig 1) . HFNC was initiated in 157 patients, and subjects who were intubated after HFNC weaning (n = 3) and have a do-not-intubate order (n = 21) were excluded from the analysis.

The characteristics of the patients with SARS-CoV-2 infection who were treated with HFNC are presented in Table 1 . The HFNC failure group comprised older patients with higher Charlson comorbidity index, SOFA score, and APACHE II score than the HFNC success group. In addition, patients in the HFNC failure group had longer duration of fever, higher incidence of hypertension, higher use of vasopressor, and increased use of continuous renal replacement therapy. No significant differences in initial vital sign measurements, chest x-ray and laboratory results, PF ratio at HFNC application, and the use of corticosteroid were noted between the two groups.

The univariate analysis revealed that age, Charlson comorbidity index, CURB-65, SOFA score, APACHE II score, hypertension, 1 h SF ratio, 4 h SF ratio, 1 h ROX index, and 4 h ROX index were associated with HFNC failure (Table 2) . On the other hand, the multivariable analysis revealed that patients aged �70 years are 3.4 times more likely to experience HFNC failure than those aged <70 years (HR 3.367 [95% CI: 1.358-8.349], p = 0.009). Moreover, the SF ratio at 1 h (HR 0.983 [95% CI: 0.972-0.994], p = 0.003) was significantly associated with HFNC failure.

Changes in respiratory variables during HFNC are shown in Table 3 . The ROX index and the SF ratio after HFNC initiation were significantly higher in HFNC success group. The SF ratio at 1 h after HFNC initiation was an important predictor of HFNC (AUC 0.762; 95% confidence interval [CI]: 0.679-0.846) (Fig 2) . The AUCs of the SF ratio at 4 h and the ROX indices at 1 h and 4 h after HFNC initiation were 0.733 (95% CI: 0.640-0.826), 0.697 (95% CI: 0.597-0.798), and 0.682 (95% CI: 0.583-0.781), respectively.

An SF ratio at 1 h after HFNC initiation <166 had a sensitivity of 79.3% and a specificity of 37.3% for predicting HFNC failure (S1 Table) . An ROX index at 1 h after HFNC initiation <8.54 had a sensitivity of 59.6% and a specificity of 26.1% for predicting HFNC failure. The AUCs of the ROX index and the SF ratio at 12 h after HFNC initiation were 0.619 (CI: 0.500-0.738) and 0.686 (CI: 0.575-0.797), respectively.

Age, Charlson comorbidity index, CURB-65, SOFA score, and APACHE II score were factors associated with in-hospital mortality in patients treated with HFNC as revealed by the univariate analysis (Table 4 ). According to the multivariable analysis results, the odds of death of patients aged �70 years were 4.9 times higher than those aged <70 years (HR 4.867 [95% CI: 1.772-13.364], p = 0.002). Patients with SOFA score �3 were 4.4 times more likely to die than those with SOFA score <3 (HR 4.437 [95% CI: 1.770-11.120], p = 0.001).

In this multicenter cohort study in tertiary hospitals in Korea, we evaluated the factors associated with HFNC failure and mortality in COVID-19 patients requiring HFNC. We found an HFNC failure rate as high as 53% and an overall mortality of HFNC-treated patients of 24% at referral. After HFNC application, the ROX index and SF ratio between the HFNC success group and the HFNC failure group were significantly different. Age and SF ratio at 1 h after HFNC initiation were significantly associated with HFNC failure. Furthermore, the SF ratio at 1 h and 4 h after HFNC initiation had an acceptable predictive ability for HFNC failure. Mortality of COVID-19 patients was significantly associated with older age and higher SOFA score.

Since the advent of SARS-CoV-2, there have been reports on the usefulness of the ROX index to predict intubation risk in COVID-19 patients receiving HFNC [14] [15] [16] [17] [18] [19] were consistently lower in the HFNC failure group than in the HFNC success group. The AUCs of the ROX indices at 4 h or 6 h after HFNC initiation ranged from 0.70-0.798 [14, 17, 18] . In our study, the predictive function of the ROX index was not remarkable. Compared with those of previous studies, our cohort included relatively older patients with moderate hypoxemia. In addition, ROX indices after HFNC initiation were higher than those reported in previous studies. We attribute the low predictive power of the ROX index to the rapid worsening of respiratory failure in older patients. In cases of acute deterioration in older COVID-19 patients that is not adequately reflected by the ROX index, more accurate or easily applicable alternatives are needed. Furthermore, as the PaO 2 /FiO 2 ratio can be more accurate than the SpO 2 /FiO 2 ratio, a modified ROX index using the PaO 2 /FiO 2 ratio to respiratory rate may be a better prognostic factor that needs validation in the future [20] . Catoire et al. [21] suggested that the AUCs of the SF ratio for PF ratio of 300 and 400 mmHg were high (0.918 and 0.901), which could be used as a hypoxemia screening tool in the emergency department. Using multivariable logistic regression analysis, Patel et al. [22] found that the SF ratio is a significant predictor of intubation risk in COVID-19-related hypoxemic patients. In our study, the SF ratios after HFNC initiation had an acceptable predictive power (AUC: 0.762 and 0.733). Hu et al. [23] reported that both the ROX index at 6 h and the SF and PF ratios at 6 h are accurate predictors of HFNC failure (AUC: SF ratio of 0.786 and PF ratio of 0.749). Arterial blood gas analysis cannot be performed frequently because COVID-19 patients are isolated in a closed room and the number of healthcare personnel is limited. Moreover, respiratory rate may be difficult to objectively monitor because the accuracy of respiratory rate measurements by healthcare professionals is suboptimal [23] . On the other hand, the SF ratio can be calculated by objectively measuring pulse oximetry and FiO 2 ; therefore, we suggest that SF ratios can be a useful tool for predicting intubation in COVID-19 patients.

Whether delayed intubation is associated with higher mortality in COVID-19 patients remains controversial. Physicians managing COVID-19 patients may attempt to avoid intubation whenever possible because of the risk of aerosol dispersion, ventilator-associated pneumonia, or complications, such as unplanned extubation. A recent study revealed no association between time-to-intubation and mortality or further lung injury in critically ill COVID-19 patients [24] . In contrast, Hyman et al. [25] reported that the timing of intubation is significantly associated with mortality, with an adjusted hazard ratio for mortality of 1.03 for each day of delay in intubation. Self-inflicted lung injury associated with delayed intubation could also aggravate lung damage [26] , resulting in higher mortality. In addition, 73% of hypoxemia cases and 18% of cardiac arrest cases occurred during emergency intubation of COVID-19 patients [27] . These events are more likely to occur when intubation is performed during aggravated hypoxemia due to delayed intubation. The results of this study indicate that older age and higher SOFA score are important factors for mortality in COVID-19 patients receiving HFNC. Therefore, close observation is required to avoid delayed intubation in older patients with higher SOFA score. This study has some limitations. First, owing to the retrospective nature of this study, the ROX index may have been calculated inaccurately because of errors in measuring respiratory rate. In addition, respiratory rate can often be neglected owing to time constraints and lack of clinical resources [28] . Therefore, given the availability of limited resources due to the COVID-19 pandemic, our results suggest that the SF ratio can be a useful alternative. Second, the timing of intubation was not standardized at each center. Physicians' experience may have played a crucial role in the decision of intubation. However, the intubation criteria are not significantly different between COVID-19 and conventional respiratory failure and trained board-certified intensivists oversaw decision-making. Third, this study included relatively small number of patients who received HFNC. We performed a post-hoc power analysis with a sample size of 133 and type 1 error of 0.05, achieving a power of 86% for HFNC failure prediction. The post-hoc power analysis ensures sufficient power of the sample size. Nonetheless, prospective randomized controlled studies with a larger number of patients and standardized protocols are needed in the future.

High-flow nasal cannula oxygen therapy is superior to conventional oxygen therapy but not to noninvasive mechanical ventilation on intubation rate: a systematic review and meta-analysis

Can High-flow Nasal Cannula Reduce the Rate of Endotracheal Intubation in Adult Patients With Acute Respiratory Failure Compared With Conventional Oxygen Therapy and Noninvasive Positive Pressure Ventilation?: A Systematic Review and Meta-analysis

High flow nasal cannula compared with conventional oxygen therapy for acute hypoxemic respiratory failure: a systematic review and meta-analysis. Intensive care medicine

Failure of high-flow nasal cannula therapy may delay intubation and increase mortality

Predicting success of high-flow nasal cannula in pneumonia patients with hypoxemic respiratory failure: The utility of the ROX index

An Index Combining Respiratory Rate and Oxygenation to Predict Outcome of Nasal High-Flow Therapy. American journal of respiratory and critical care medicine

Comparison of the SpO2/ FIO2 ratio and the PaO2/FIO2 ratio in patients with acute lung injury or ARDS

Utility of SpO2/FiO2 ratio for acute hypoxemic respiratory failure with bilateral opacities in the ICU

SpO(2)/FiO(2) as a predictor of high flow nasal cannula outcomes in children with acute hypoxemic respiratory failure

Predicting non-invasive ventilation failure in children from the SpO 2 /FiO 2 (SF) ratio. Intensive care medicine

Misconceptions of pathophysiology of happy hypoxemia and implications for management of COVID-19

Surviving Sepsis Campaign: guidelines on the management of critically ill adults with Coronavirus Disease 2019 (COVID-19)

Protocol of hospitals dedicated to infectious disease 2020

High-Flow Nasal Cannula Therapy in COVID-19: Using the ROX Index to Predict Success

Validity of ROX index in prediction of risk of intubation in patients with COVID-19 pneumonia. Advances in respiratory medicine

High-flow nasal cannula for Acute Respiratory Distress Syndrome (ARDS) due to COVID-19. Multidisciplinary respiratory medicine

High-Flow Nasal Cannula for COVID-19 Patients: A Multicenter Retrospective Study in China. Frontiers in molecular biosciences

Application of high-flow nasal cannula in hypoxemic patients with COVID-19: a retrospective cohort study

The utility of high-flow nasal oxygen for severe COVID-19 pneumonia in a resource-constrained setting: A multi-centre prospective observational study

Success or Failure of High-Flow Nasal Oxygen Therapy: The ROX Index Is Good, but a Modified ROX Index May Be Better. American journal of respiratory and critical care medicine

Assessment of the SpO(2)/FiO(2) ratio as a tool for hypoxemia screening in the emergency department

Retrospective analysis of high flow nasal therapy in COVID-19-related moderate-to-severe hypoxaemic respiratory failure

The ROX index can be a useful tool for the triage evaluation of COVID-19 patients with dyspnoea

PubMed Central PMCID: PMC7448713 (NIH), and he received funding from Clew Medical. Drs. Blum and Auld received support for article research from the NIH. Dr. Martin disclosed he currently serves as president-elect for the Society of Critical Care Medicine

Timing of Intubation and In-Hospital Mortality in Patients With Coronavirus Disease

Patient self-inflicted lung injury: implications for acute hypoxemic respiratory failure and ARDS patients on non-invasive support. Minerva anestesiologica

Emergency tracheal intubation in 202 patients with COVID-19 in Wuhan, China: lessons learnt and international expert recommendations. British journal of anaesthesia

Accurate respiratory rates count: So should you! Australasian emergency nursing journal

The SpO 2 /FiO 2 ratio following HFNC initiation was an acceptable predictor of HFNC failure. The SpO 2 /FiO 2 ratio may be a good prognostic marker for intubation in COVID-19 patients receiving HFNC. The timing of intubation during HFNC treatment should be carefully tailored to patients, and close monitoring is needed, particularly in older COVID-19 patients.Supporting information S1