Abstract
BACKGROUND: How indices specific to respiratory compromise contribute to prognostication in patients with ARDS is not well characterized in general clinical populations. The primary objective of this study was to identify variables specific to respiratory failure that might add prognostic value to indicators of systemic illness severity in an observational cohort of subjects with ARDS.
METHODS: Fifty subjects with ARDS were enrolled in a single-center, prospective, observational cohort. We tested the contribution of respiratory variables (oxygenation index, ventilatory ratio [VR], and the radiographic assessment of lung edema score) to logistic regression models of 28-d mortality adjusted for indicators of systemic illness severity (the Acute Physiology and Chronic Health Evaluation [APACHE] III score or severity of shock as measured by the number of vasopressors required at baseline) using likelihood ratio testing. We also compared a model utilizing APACHE III with one including baseline number of vasopressors by comparing the area under the receiver operating curve (AUROC).
RESULTS: VR significantly improved model performance by likelihood ratio testing when added to APACHE III (P = .036) or the number of vasopressors at baseline (P = .01). Number of vasopressors required at baseline had similar prognostic discrimination to the APACHE III. A model including the number of vasopressors and VR (AUROC 0.77 [95% CI 0.64–0.90]) was comparable to a model including APACHE III and VR (AUROC 0.81 [95% CI 0.68–0.93]; P for comparison = .58.).
CONCLUSIONS: In this observational cohort of subjects with ARDS, the VR significantly improved discrimination for mortality when combined with indicators of severe systemic illness. The number of vasopressors required at baseline and APACHE III had similar discrimination for mortality when combined with VR. VR is easily obtained at the bedside and offers promise for clinical prognostication.
Introduction
Estimating mortality risk in ARDS is a crucial component of clinical decision-making and providing high-quality bedside care. Mortality in ARDS can be driven predominantly by a patient’s underlying ARDS risk factor (such as sepsis), underlying comorbidities, or the severity of lung injury, in part measured by the degree of hypoxemia.1 Respiratory variables alone do not fully capture mortality risk in ARDS given the heterogeneity of clinical risk factors and frequent concomitant multi-organ failure. Additionally, since respiratory failure is a substantial contributor to mortality in patients with ARDS, general severity of illness scores do not fully explain ARDS mortality risk.1 Thus, metrics that are specific to both severity of critical illness and respiratory failure are necessary to best predict mortality in ARDS.
Various metrics have been developed to risk stratify patients with ARDS based both on their general severity of illness and on their degree of respiratory dysfunction. The Acute Physiology and Chronic Health Evaluation (APACHE) III score is highly associated with hospital mortality risk for critically ill adults.2 APACHE III is not readily available in most clinical settings, however, and it is burdensome to calculate. The ventilatory ratio (VR) is an index of impaired ventilation that can be easily obtained at the bedside, correlates with pulmonary dead space, and is independently associated with an increased odds of hospital mortality even after adjusting for PEEP, hypoxemia, and severity of illness.3 A physiologic pulmonary dead-space fraction < 0.3 and VR of one are considered normal, and higher values represent worse physiological derangements. A VR ≥ 2 in ARDS (high VR) has been shown to be associated with a higher risk of mortality.3
Many previous studies of mortality predictors in ARDS have utilized large clinical trial populations. Importantly, however, mortality from ARDS is higher in observational studies compared to randomized controlled trials (RCTs), partly because observational studies often include subjects who would be excluded from RCTs, such as those with severe comorbidities.4 The aim of this research was to study the following questions in an observational cohort of critically ill subjects with comorbidities and illness severity that are often not represented by clinical trials: (1) Which candidate variables specific to respiratory compromise will improve the performance of a logistic regression model describing 28-d mortality? and (2) Will the number of vasopressors required at baseline have comparable discriminatory power for mortality to APACHE III?
QUICK LOOK
Current Knowledge
Estimating mortality risk in ARDS is essential for shared decision-making and high-quality bedside care. How various indices specific to respiratory compromise contribute to prognostication in patients with ARDS is not well characterized in general clinical populations.
What This Paper Contributes to Our Knowledge
The ventilatory ratio (VR) significantly improved discrimination for mortality when combined with indicators of severe systemic illness. For severity of illness, the number of vasopressors required at baseline and the APACHE III had similar discrimination for mortality when combined with VR. The VR can be applied practically and easily at the bedside and offers promise for clinical prognostication in routine care.
Methods
Study Design and Oversight
This was a secondary analysis of a single-center prospective observational cohort study of subjects with ARDS who required endotracheal intubation and mechanical ventilation and were admitted to an ICU in the Moffitt-Long Hospital (Parnassus Campus) at the University of California, San Francisco Medical Center. This study was approved by the institutional review board (#17-21982).
Human Subjects: Patient Selection and Informed Consent
Daily screening of the electronic medical records from July 1, 2017–March 12, 2019, was used to identify mechanically ventilated patients who were endotracheally intubated and met the Berlin criteria for ARDS (including PaO2/FIO2 < 300 mm Hg, PEEP ≥ 5 cm H2O, and bilateral infiltrates on chest radiography not fully explained by cardiogenic pulmonary edema).5 Subjects were eligible for inclusion if ARDS had been present for ≤ 72 h at the time of enrollment. Patients were excluded if they were younger than 18, had an expected survival of < 96 h, had a previous diagnosis of pulmonary hypertension; or were receiving mechanical circulatory support (extracorporeal membrane oxygenation [ECMO], ventricular-assist device, or intra-aortic balloon pump), on pulmonary vasodilators, or had been subject to externally induced cardioplegia (such as in the setting of cardiopulmonary bypass) within 24 h. Informed consent was obtained for all participants. If the patient was unable to provide consent, consent was obtained from a surrogate.
Data Collection
Clinical data were extracted from the electronic medical record. Data collected at enrollment included subject demographics, primary and secondary risk factors for ARDS, comorbid conditions, and components of the APACHE III score. PaO2/FIO2, oxygenation index (OI, [FIO2 × mean airway pressure × 100]/PaO2), ventilator parameters (plateau airway pressure, mean airway pressure, PEEP, tidal volume [VT], respiratory system compliance, driving pressure), mode of ventilation, and pulmonary dead-space fraction (VD/VT) were also recorded at enrollment. VD/VT was calculated from the Enghoff modification of the Bohr equation utilizing simultaneous mixed-expired CO2 measured with volumetric capnography (NICO Cardiopulmonary Management System, Philips Respironics, Murrysville, Pennsylvania) and arterial blood gas measurements.6 Two sequential calculations of VD/VT were made using measurements with a 5-min interval, and these 2 values were then averaged to represent the patient’s baseline pulmonary dead-space fraction. VR was calculated as (minute ventilation [mL/min] × PaCO2 [mm Hg]/predicted body weight [kg] × 100 × 37.5).7 Ventilator-free days (VFD) were calculated as days alive and free of mechanical ventilation to 28 d. ICU-free days were calculated as days alive and outside of the ICU to 28 d. For both outcomes, subjects who died before 28 d were assigned zero.
Predictors and End Point
The primary end point was 28-d mortality. Prespecified candidate respiratory variables for the model were baseline PaO2/FIO2, OI, VR, and radiographic assessment of lung edema (RALE) score.8
Statistical Analyses
Continuous variables are expressed as mean ± SD or median (interquartile range [IQR]). Categorical variables are presented as count (percentage). For between-group comparisons, Fisher exact test was used to compare categorical variables; 2-sided unpaired t test was used for normally distributed continuous variables, and the Mann-Whitney U test was used for non-normally distributed continuous variables. Spearman correlation coefficients were used to describe all correlations and investigate collinearity between variables. Multivariable logistic regression utilizing manual stepwise backward selection and likelihood ratio testing for nested models was used for analyses of the association between candidate variables and 28-d mortality with appropriate model checking. Areas under receiver operating curve (AUROCs) were compared by the Delong test for non-nested models.9 A 2-sided P value of < .05 was considered significant. All analyses were performed using Stata SE 16.1 (StataCorp, College Station, Texas).
The APACHE III score was used in the models as an indicator of severity of illness because it is highly associated with hospital mortality risk for critically ill adults.2 Vasopressors were also tested as an indicator of a subject’s severity of illness and risk for multisystem organ failure and because the number of vasopressors a patient is receiving is much easier to calculate than APACHE III in a clinical setting.
Results
Of 322 patients screened for inclusion, 208 met criteria for ARDS within 72 h of screening. Of these 158 were excluded, and 50 subjects were enrolled (Fig. 1). Pneumonia was the most common primary ARDS etiology (46%) (Table 1). The baseline APACHE III score was 107 ± 30, and 82% of the subjects had vasopressor-dependent shock. Many subjects had major comorbid conditions such as cirrhosis, chronic lung disease, end-stage renal disease, and heart failure. Most subjects had at least one major comorbid condition. Most subjects had a PaO2/FIO2 < 200 mm Hg. Both VD/VT (0.59 ± 0.13) and VR (1.9 [1.6–2.3]) were elevated in this cohort. Eighty-four percent of subjects were managed with volume-controlled ventilation compared to 16% using pressure-regulated modes. Mortality at day 28 was 56% (Table 2). High mortality accounted for a median of zero VFD and ICU-free days. Median duration of mechanical ventilation among survivors was 9 (IQR 4–20) d. Median ICU length of stay among survivors was 11 (IQR 7–21) d, and median hospital length of stay was 27 (IQR 13–28) d.
Flow chart.
Baseline Characteristics
Clinical Outcomes
The APACHE III score and the nonpulmonary Sequential Organ Failure Assessment score were significantly higher among nonsurvivors as compared to survivors (Table 3). Nonsurvivors required a significantly greater number of vasopressors at baseline. VR was significantly higher among nonsurvivors (P = .002) (Fig. 2). Mortality among subjects with high VR (VR ≥ 2) was 78% versus 37% in the low VR group (P = .005).
Box plot of ventilatory ratio dichotomized by A: alive (n = 22) versus B: dead (n = 28) at day 28. Horizontal lines represent median; boxes represent upper and lower quartiles, and whiskers represent 1.5 times the interquartile range. Ventilatory ratio was significantly higher among nonsurvivors, P = .002 by Wilcoxon rank sum.
Baseline Characteristics by Mortality at Day 28
There was a strong negative correlation between PaO2/FIO2 and OI (ρ = −0.91, P < .001). OI was included as a prognostic indicator in candidate models instead of PaO2/FIO2 given the existing body of evidence that OI is an independent risk factor for mortality in adults with ARDS10-12 and given that PaO2/FIO2 is not consistently associated with mortality.12-14 Using backward selection and likelihood ratio testing, OI and the RALE score were eliminated from the model. VR significantly improved the model of 28-d mortality as compared to a model including APACHE III alone (P = .036 by likelihood ratio test). Adjusted for APACHE III, each 0.5-unit change in VR was associated with odds ratio for 28-d mortality of 1.78 (95% CI 0.78–3.23). The model including APACHE III and VR had an AUROC of 0.81 (95% CI 0.68–0.93). Univariable logistic regression for each candidate variable, the full candidate logistic regression model, and the final model for 28-d mortality using physiological variables are depicted in Table 4. VR significantly improved a model of 28-d mortality by likelihood ratio testing when added to vasopressor number at baseline (P = .01). The model including vasopressor number and VR had an AUROC of 0.77 (95% CI 0.64–0.90). This was comparable (P value for comparison = .58) to the model of APACHE III with VR, which had an AUROC of 0.81 (95% CI 0.68–0.93).
Candidate Logistic Regression Models of 28-Day Mortality Using Physiologic Variables
Discussion
In this observational cohort of 50 subjects that represents ARDS populations outside of clinical trials, the VR and the number of vasopressors required at baseline were strongly associated with mortality. Both clinical variables are easily obtained and could enhance bedside assessment of patient prognosis. VR has previously been demonstrated to be associated with mortality in large cohorts of patients,3,15 but this is the first study to our knowledge to test the association between VR and mortality in a small observational cohort. These findings support the generalizability of VR as a prognostic variable in general populations of patients with ARDS.
Several measures of the severity of respiratory failure have been studied for their prognostic validity in ARDS. We tested PaO2/FIO2, OI, VR, and the RALE score because of their frequent use in clinical and research settings and their established associations with ARDS outcomes.3,8,10,12,16 In previous work, a novel composite score that includes PaO2/FIO2, the RALE score, and VR had high discrimination for the need for ECMO or death from severe pulmonary dysfunction.17 However, this score was derived from a single cohort without external validation and had poor discrimination for overall hospital mortality, so its generalizability remains unclear. In our study, the VR improved logistic regression models of 28-d mortality when combined with APACHE III, the most commonly used severity of illness score in ARDS, or with the number of vasopressors required at baseline. Additionally, a model that replaced the multicomponent APACHE III with the number of vasopressors required at baseline performed similarly to the model with APACHE III. Given its complexity and limited availability outside the research setting, the APACHE III score is rarely used in clinical practice, and these results indicate that the number of baseline vasopressors in patients with shock may represent similar prognostic information. These findings are especially important given that this cohort consisted of subjects who were critically ill with comorbidities and poor prognoses that would likely disqualify them from RCTs.
ARDS is a complex and heterogeneous syndrome that often involves both respiratory failure and multisystem organ dysfunction, leading to severe critical illness. Although the APACHE III score is a reliable measure of severity of illness, it is not designed to specifically measure the severity of lung injury or respiratory failure. Thus, we tested whether it would be valuable to supplement a measure of severity of illness with variables specific to respiratory compromise when modeling mortality in ARDS. Several indices of respiratory failure have been shown to have stronger independent associations with ARDS outcomes than the PaO2/FIO2.14,18,19 Pulmonary VD/VT is an independent predictor of mortality in ARDS19-21; however, estimation of dead space requires specialized equipment to measure the PaO2 (PCO2) in mixed expired air. In this study, VR improved a model of 28-d mortality including APACHE III, which suggests that VR captures a domain of ARDS severity and lung injury that is not reflected by APACHE III alone. These findings are also consistent with the existing literature demonstrating that VR is independently associated with mortality even after adjusting for oxygenation, PEEP, and severity of illness with APACHE II.3 Previous work has demonstrated the importance of identifying patients at risk of ARDS-attributable mortality,1 and our findings suggest that VR can play a valuable role in identifying these high-risk patients. When VR was dichotomized into high VR (≥ 2) and low VR (< 2), mortality was significantly greater among subjects with high VR.
Although the sample size of this study may limit its precision, we deliberately tested our hypotheses in a modest-sized cohort. This was done to elucidate which of the physiologic factors that was previously demonstrated to be associated with mortality in large clinical trials is strong enough prognosticator that it would also be valuable for routine clinical care. This study had very few exclusion criteria. As a result, and in contrast to many RCTs, our cohort included severely ill subjects with multiple comorbidities. Chronic respiratory failure, chronic liver disease, previous bone marrow transplantation, and prior lung transplantation are all medical problems that were represented in our cohort but have been exclusion criteria in many of the landmark ARDS studies.22-24 As was observed in the LUNG-SAFE study,25 subjects in observational cohorts are often more systemically ill and have higher mortality rates. We believe that observational studies such as this one have a valuable role in the generalizability of ARDS research, as emphasized in a prior publication in this journal.6
Conclusions
The results of this study underscore that VR was a valuable tool for mortality risk assessment and captured a domain of ARDS severity that is not reflected by general severity of illness indicators. Although VR improved a model of 28-d mortality when added to a severity of illness score, it also performed well on its own at discriminating between those who lived and those who died. This severely ill cohort is representative of ARDS in clinical practice more so than the carefully selected populations that generally meet all criteria for enrollment in clinical trials. Thus, VR, a respiratory variable that is convenient to calculate using information already collected in critically ill patients, may be valuable for determining mortality risk in clinical trials as well as risk assessment and shared decision making in the clinical setting.
Acknowledgments
We thank the subjects and families for their consent and participation in this study, without whom this research would not have been possible.
Footnotes
- Correspondence: Katherine D Wick MD, 505 Parnassus Avenue, UCSF, Moffitt Hospital, M-917, San Francisco, CA 94134–0624. E-mail: Katherine.wick{at}ucsf.edu
See the Related Editorial on Page 1208
Dr Wick has received grant support from NIH, No. 5T32GM008440-24. The remaining authors have disclosed no conflicts of interest.
This study was funded by grants from the National Institutes of Health: 5TL1TR001871-05 (Ms Siegel), R35 HL140026 (Dr Calfee), and R01 HL134828 (Dr Matthay). This research was also supported by a grant from Bayer Pharmaceuticals.
- Copyright © 2022 by Daedalus Enterprises