Abstract
Background
The predictive value of the respiratory rate‑oxygenation (ROX) index for a high-flow nasal cannula (HFNC) in patients with COVID-19 with acute hypoxemic respiratory failure (AHRF) may differ from patients without COVID-19 with AHRF, but these patients have not yet been compared. We compared the diagnostic accuracy of the ROX index for HFNC failure in patients with AHRF with and without COVID-19 during acute emergency department (ED) visits.
Methods
We performed a retrospective analysis of patients with AHRF treated with an HFNC in an ED between October 2020 and April 2022. The ROX index was calculated at 1, 2, 4, 6, 12, and 24 h after HFNC placement. The primary outcome was the failure of the HFNC, which was defined as the need for subsequent intubation or death within 72 h. A receiver operating characteristic (ROC) curve was used to evaluate discriminative power of the ROX index for HFNC failure.
Results
Among 448 patients with AHRF treated with an HFNC in an ED, 78 (17.4%) patients were confirmed to have COVID-19. There was no significant difference in the HFNC failure rates between the non-COVID-19 and COVID-19 groups (29.5% vs. 33.3%, p = 0.498). The median ROX index was higher in the non-COVID-19 group than in the COVID-19 group at all time points. The prognostic power of the ROX index for HFNC failure as evaluated by the area under the ROC curve was generally higher in the COVID-19 group (0.73–0.83) than the non-COVID-19 group (0.62–0.75). The timing of the highest prognostic value of the ROX index for HFNC failure was at 4 h for the non-COVID-19 group, whereas in the COVID-19 group, its performance remained consistent from 1 h to 6 h. The optimal cutoff values were 6.48 and 5.79 for the non-COVID-19 and COVID-19 groups, respectively.
Conclusions
The ROX index had an acceptable discriminative power for predicting HFNC failure in patients with AHRF with and without COVID-19 in the ED. However, the higher ROX index thresholds than those in previous publications involving intensive care unit (ICU) patients suggest the need for careful monitoring and establishment of a new threshold for patients admitted outside the ICU.
1
Introduction
A high-flow nasal cannula (HFNC) is routinely used as a noninvasive procedure in the management of acute hypoxemic respiratory failure (AHRF) in patients with COVID-19. During the COVID-19 pandemic, HFNCs have been extensively used outside the intensive care unit (ICU) such as in the emergency department (ED) [ , ]. Failure of an HFNC may increase the risk of delayed intubation and mortality in patients with AHRF [ , ]. Known factors associated with HFNC failure include heart rate, respiratory rate, organ dysfunction scores, and the respiratory rate‑oxygenation (ROX) index, which is defined as the ratio of oxygen saturation (SpO 2 )/fraction of inspired oxygen (FiO 2 ) to respiratory rate [ , ].
The ROX index has been increasingly applied to predict the outcome of HFNC placement in patients with pneumonia with AHRF. The first study using the ROX index to predict HFNC failure in ICU patients with pneumonia found that an ROX index greater than or equal to 4.88 measured 12 h after HFNC placement was associated with a decreased risk of the need for intubation [ , ]. Several studies have evaluated the predictive accuracy of the ROX index for HFNC failure during the COVID-19 pandemic; however, the findings are inconsistent due to differences in the clinical setting, cutoff values used, and population [ , ]. Moreover, unlike traditional respiratory failure, COVID-19 often presents with silent hypoxia, a condition in which the patient has no abnormal respiratory patterns despite severe hypoxia [ ].
Therefore, we hypothesized that the predictive value of the ROX index for HFNC failure may be different in patients with AHRF with and without COVID-19, especially in those presenting for acute ED visits. In addition, the threshold value of the ROX index for predicting HFNC failure needs to be confirmed. We aimed to compare the predictive value of the ROX index for HFNC failure between patients with AHRF with and without COVID-19.
2
Materials and methods
2.1
Study design and population
This single-center, retrospective, observational study was conducted in an ED with an annual census of approximately 120,000 visits at Asan Medical Center, a university-affiliated teaching hospital in South Korea. Among adult patients aged 18 years or older who visited the ED from October 2020 to April 2022, the medical records of those with AHRF who underwent HFNC placement (AIRVO2, Fisher & Paykel Healthcare) in the ED were reviewed. Patients with a prescription code for HFNC were identified. Indications for initiating HFNC were based on the following criteria: clinical signs of respiratory failure (such as the use of accessory muscles, respiratory rate >20 /min, paradoxical abdominal movement, desaturation despite nasal cannula or venturi therapy), partial pressure of arterial oxygen (PaO 2 ) < 60 mmHg or a PaO2/FiO 2 ratio < 300 with supplemental oxygen [ , ]. We excluded patients who (1) were transferred from or to other centers during HFNC placement, (2) had a “do-not-intubate” order, (3) were electively intubated (e.g., surgery or bronchoscopy), (4) were intubated or switched to other devices within one hour, (5) underwent HFNC placement after extubation, (6) had incomplete data, or (7) withdrew a “do-not-intubate” order after 72 h. Our institutional review board approved the review of patient data and waived the requirement for informed consent (IRB no. 2022–0589).
2.2
Data collection
We collected demographic and clinical data, such as age, sex, and comorbid diseases such as hypertension, diabetes mellitus, hematologic malignancy, cancer, cardiovascular disease, cerebrovascular disease, heart failure, chronic kidney disease, chronic liver disease, by reviewing the electronic medical records. Vital signs and relevant laboratory data during the ED stay were collected. HFNC therapy was initiated with a temperature between 31 °C and 37 °C, a flow of 40–60 L/min, and the FiO 2 was adjusted to maintain an SpO 2 over 92%. The time from ED admission to HFNC therapy, time from HFNC placement to intubation, and length of HFNC therapy were also reviewed. The ROX index was calculated with the SpO 2 , FiO 2 , and respiratory rate at 1, 2, 4, 6, 12, and 24 h after the initiation of HFNC treatment. Patients were divided into a non-COVID-19 group and a COVID-19 group. COVID-19 was diagnosed as a positive result on a polymerase chain reaction test of nasopharyngeal exudate samples for SARS-CoV-2. The primary outcome was the development of HFNC failure, which was defined as a need for endotracheal intubation or death within 72 h of HFNC placement. The clinical characteristics and ROX index to predict HFNC failure in patients with AHRF with and without COVID-19 were compared.
2.3
Statistical analysis
Continuous variables are expressed as the mean with the standard deviation or as the median with the interquartile range depending on whether they were normally distributed, as assessed by the Kolmogorov-Smirnov test. Categorical data are presented as absolute numbers and percent frequencies. Differences between continuous variables were analyzed using Student’s t -test or the Mann-Whitney U test as appropriate. The Chi-square test or Fisher’s exact test was used to compare categorical variables. A receiver operating characteristic (ROC) curve with the area under the curve (AUC) was used to determine the accuracy of the ROX index in discriminating the failure of HFNC placement over time in the COVID-19 and non-COVID-19 groups. The optimal cutoff value of the ROX index for each time to discriminate the failure of HFNC placement was determined using Youden’s index. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), and negative likelihood ratio (NLR) were calculated to compare the predictive performance of the ROX index for HFNC failure of the traditional cutoff value of 4.88 and new cutoff values. The comparison of the two independent ROC curves was performed using DeLong’s test [ ]. A two-tailed P < 0.05 was considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 21.0 (IBM Corp., Armonk, NY, USA) and MedCalc® Statistical Software version 20.116 (MedCalc Software Ltd., Ostend, Belgium).
3
Results
A total of 850 patients with AHRF who required the use of an HFNC in the ED were identified during the study period, of whom 448 met the inclusion criteria ( Fig. 1 ). Several patients were excluded from the study for various reasons, such as do-not-intubate orders ( n = 216), transfer out ( n = 104), intubation within 1 h ( n = 35), transfer in ( n = 13), switch to another device within 1 h ( n = 12), inadequate information ( n = 11), elective intubation for a surgical procedure ( n = 5), HFNC placement after extubation ( n = 4), and withdrawal of do-not-intubate order after 72 h ( n = 2). Seventy-eight patients (17.4%) were confirmed to have COVID-19. Of the 370 patients without COVID-19, 261 (70.5%) were successfully weaned from HFNC therapy without intubation, whereas 109 (29.5%) were intubated or expired within 72 h of HFNC placement. Among 78 patients with COVID-19, 52 (66.7%) experienced HFNC success, and 26 (33.3%) experienced failure. No statistically significant differences between the two groups were found in terms of HFNC failure ( P = 0.498) ( Table 1 ).
| Total ( n = 448) | Non-COVID-19 ( n = 370) | COVID-19 ( n = 78) | P value | |
|---|---|---|---|---|
| Age, years | 68 (61–77) | 68 (61–77) | 67 (62–76) | 0.803 |
| Sex, male, n (%) | 293 (65.4) | 247 (66.8) | 46 (59.0) | 0.189 |
| Comorbidities, n (%) | ||||
| Hypertension | 225 (50.2) | 180 (48.6) | 45 (57.7) | 0.147 |
| Diabetes mellitus | 158 (35.3) | 128 (34.6) | 30 (38.5) | 0.516 |
| Hematologic malignancy | 30 (6.7) | 28 (7.6) | 2 (2.6) | 0.108 |
| Cancer | 119 (26.6) | 106 (28.6) | 13 (16.7) | 0.029 |
| Cardiovascular disease | 66 (14.7) | 55 (14.9) | 11 (14.1) | 0.863 |
| Cerebrovascular disease | 34 (7.6) | 29 (7.8) | 5 (6.4) | 0.665 |
| Chronic heart failure | 30 (6.7) | 29 (7.6) | 2 (2.6) | 0.108 |
| Chronic renal failure | 69 (15.4) | 54 (14.6) | 15 (19.2) | 0.303 |
| Chronic liver disease | 22 (4.9) | 19 (5.1) | 3 (3.8) | 0.632 |
| Laboratory findings | ||||
| White blood cell count, × 10 3 /uL | 10.7 (7.1–15.2) | 11.3 (7.7–16.0) | 8.2 (5.6–11.2) | < 0.001 |
| Hemoglobin, g/dL | 11.6 (10.0–13.4) | 11.4 (9.9–13.2) | 12.6 (10.8–14.0) | 0.004 |
| Platelet count, × 10 3 /uL | 224 (151–308) | 233 (156–313) | 178 (139–256) | 0.004 |
| Creatinine, mg/dL | 0.95 (0.70–1.66) | 0.94 (0.68–1.57) | 1.05 (0.81–1.74) | 0.068 |
| PaO 2 /FiO 2 ratio | 121 (78–176) | 125 (84–187) | 103 (65–140) | 0.001 |
| HFNC failure | 135 (30.1%) | 109 (29.5%) | 26 (33.3%) | 0.498 |
| Time from HFNC placement to intubation | 461 (188–1192) | 452 (193–1109) | 505 (162–1868) | 0.598 |
Table 1 shows the baseline demographics and clinical characteristics of the patients. The median age was 68 years, and the male-to-female ratio was 2:1. There was no difference in baseline characteristics between the non-COVID-19 and COVID-19 groups, except for the white blood cell count, hemoglobin level, and platelet count. The PaO 2 /FiO 2 ratio was lower in the COVID-19 group than in the non-COVID-19 group (median, 125 vs. 103; P = 0.001). The median time from HFNC placement to intubation was 505 min (162–1868 min) in the COVID-19 group and 452 min (193–1109 min) in the non-COVID-19 group.
Table 2 shows the ROX index for each time in the two groups. Significant differences were observed in the ROX index at all time points between the groups. The ROX index of the COVID-19 group was lower than that of the non-COVID-19 group. The difference in the ROX index was the largest at 24 h (median, 8.82 vs. 6.11; P < 0.001), followed by 12 h (median, 8.39 vs. 6.46; P < 0.001). The ROX index measured at 4 h showed the smallest difference between the two groups, with an absolute difference 0.70 (median, 7.20 vs. 6.50; P = 0.003). Both the COVID-19 and non-COVID-19 groups showed an increase in the ROX index over time, but the increase was more pronounced in the non-COVID-19 group (ROX 1h 6.81, ROX 24h 8.82), whereas the increase was less significant in the COVID-19 group (ROX 1h 5.41, ROX 24h 6.11). The accuracy of the ROX index in discriminating the failure of HFNC therapy was the best at 4 h after HFNC placement in the non-COVID-19 group (AUC = 0.75, 95% confidence interval [CI] = 0.69–0.80, P < 0.001) and at 6 h in the COVID-19 group (AUC = 0.83, 95% CI = 0.70–0.92, P < 0.001) ( Fig. 2 and Table 3 ). In the COVID-19 group, there was little difference in the AUC values between 1 h and 6 h (AUC 1h = 0.80 [0.68–0.89], AUC 2h = 0.82 [0.70–0.91], AUC 4h = 0.82 [0.70–0.91], AUC 6h = 0.83 [0.70–0.92]). The AUC values were higher in the COVID-19 group; however, there were no statistically significant differences except for the ROX index at 1 h (0.62 vs. 0.80; P = 0.008).
Related posts:
CT-based diagnosis in patients presenting with throat pain: A single institutional review
Droperidol undermining gastroparesis symptoms (DRUGS) in the emergency department
The long and the short of pediatric emergency department antibiotic prescribing: A retrospective observational study
Efficacy and safety of corticosteroid therapy in patients with cardiac arrest: A meta-analysis of randomized controlled trials
Further considerations regarding acute limb ischemia
Reply to: Atrial fibrillation and heart failure and the risk of false dichotomies
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
