Respiratory mechanics and mortality in coronavirus disease 2019 acute respiratory distress syndrome: A retrospective cohort study.

BACKGROUND: The association between commonly monitored respiratory parameters, including compliance and oxygenation and clinical outcomes in acute respiratory distress syndrome (ARDS) from coronavirus disease 2019 (COVID-19) remains unclear, limiting prognostication and the delivery of targeted treatments. Our project aim was to identify if any such associations exist between clinical outcomes and respiratory parameters.
METHODS: We performed a retrospective observational cohort study of confirmed COVID-19 positive patients admitted to a single dedicated intensive care unit at a university hospital from March 27 to April 26, 2020. We collected information on baseline clinical and demographic characteristics and initial respiratory parameters. Our primary outcome was in-hospital mortality.
RESULTS: A total of 22 patients met criteria for ARDS and were included in our study. Nine of the 22 (40.9%) patients with ARDS died during hospitalization. The initial static respiratory system compliance of survivors was 39 (interquartile range [IQR] 34, 55) and nonsurvivors was 27 (IQR 24, 33, P < 0.01). A lower respiratory system compliance was associated with an increased adjusted odd of in-hospital mortality (odds ratio 1.2, 95% confidence interval 1.01, 1.45 P = 0.04).
CONCLUSION: In our cohort of 22 patients mechanically ventilated with ARDS from COVID-19, having lower respiratory system compliance after intubation was associated with an increased risk of in-hospital mortality, consistent with ARDS from non-COVID etiologies. Copyright:

INTRODUCTION

Coronavirus disease 2019 (COVID-19) began to spread rapidly across the world with significant health and economic impacts. At the time of this publication, COVID-19 has infected more than 13 million people and caused over 584,000 deaths.[1] In severe cases, COVID-19 results in acute respiratory distress syndrome (ARDS).[2] ARDS was defined in accordance to the Berlin Definition as respiratory distress occurring within 1 week of a known clinical insult or new/worsening respiratory symptoms, new opacities on chest imaging, not fully explained by cardiac failure or fluid overload and graded based on severity of hypoxemia using the PaO2:FiO2 ratio (with at least 5 cmH2O of PEEP) ranging from mild 200–300, moderate 100–200, and severe <100.[3] Diagnosis of ARDS is clinically important as it is associated with mortality ranging from 27% to 45% with increasing severity based on the Berlin Definition.[3] Recent evidence shows that ARDS from COVID-19 may have phenotypic differences compared to ARDS from other etiologies.[4] Unfortunately, the initial respiratory parameters associated with clinical outcomes in this population have been poorly characterized, limiting prognostication, and the delivery of targeted treatments. We therefore evaluated patients admitted to a university hospital to determine the association between commonly recorded respiratory measurements and in-hospital mortality in COVID-19 ARDS. We hypothesized that lower respiratory system compliance would be associated with an increased risk of in-hospital mortality.

METHODS

Study design and setting

We conducted a retrospective observational cohort study of confirmed COVID-19 positive patients admitted to a single intensive care unit (ICU) designated for COVID-19 at an academic urban hospital with over 780 beds, from March 27 to April 26, 2020. The study was reviewed by our University Institutional Review Board (843,444, approved 6/22/2020) and determined to meet eligibility criteria for exemption and a waiver of HIPAA authorization (45 CFR 46.106, category 4)

Study participants

We included patients with laboratory confirmed COVID-19 infection (based on GeneXpert®; Cephied, Sunnyvale, CA, USA testing) who were admitted to our ICU and who met Berlin criteria for ARDS.[3] Patients with pre-existing restrictive lung pathology based on their medical history were excluded. All other patient populations including transplant recipients and pregnant patients were included in the cohort. Patients were followed until death or hospital discharge, which ever came first.

Study variables

For our primary exposure, we defined static respiratory system compliance as the lowest value of compliance in the 1st48 h following mechanical ventilation. Per our institutional guidelines, respiratory mechanics were performed immediately following intubation and subsequently at least twice daily. Patients were ventilated using a volume-control mode via SERVO-U® Ventilators (Maquet Medical Systems USA): Only patients requiring invasive mechanical ventilation were included in this analysis; patients were ventilated based on institutional lung protective ventilation protocol utilizing tidal volumes between 4 and 8cc/kg of ideal body weight while maintaining plateau airway pressures <30 cmH2O. Our primary outcome was in-hospital mortality. We collected additional variables through review of the electronic medical record (Epic System Corporation, Verona, Wisconsin, USA) that included patient demographics, clinical and laboratory data, and medications. Initial partial pressure of arterial oxygen (PaO2) to fraction of inspired oxygenation (FiO2), or P: F ratio, was defined similar to compliance.

Statistical methods

Our analysis was based on a convenience sample of COVID-19 patients admitted to our hospital. We therefore did not perform a sample size calculation and results should be interpreted as exploratory. We reported baseline demographic and clinical characteristics of our sample as number with proportions for the categorical variables and medians with interquartile ranges (IQR) for the continuous variables. Baseline variables were compared between survivors and non-survivors utilizing the Chi-square test for the categorical variables and Mann–Whitney test for the continuous variables. We then performed logistic regression to define the association between respiratory system compliance and in-hospital mortality. Variables with a P < 0.20 on univariate regression were included into the multivariable model. Patient age was retained a priori. Finally, we divided respiratory system compliance and P: F ratio into quartiles and depicted the proportion of in-hospital mortality within each separate quartile using an alluvial plot. We used P < 0.05 to indicate statistical significance. Stata 16.0 (College Station, TX, USA: StataCorp) was utilized for all calculations.

RESULTS

Our initial sample included 24 patients with COVID-19 infection who received mechanical ventilation [Figure 1]. We excluded 1 patient who did not meet Berlin criteria for ARDS and 1 patient with restrictive lung disease. The characteristics of the 22 included patients in our cohort are shown in Table 1. Patients had a median age of 59 years and a median body mass index of 31.4. Nine (40.9%) patients met criteria for severe ARDS. The median initial tidal volume of ideal body weight was 6 mL/kg (IQR 6, 6.7), and the median respiratory rate was 28 (IQR 24, 30). Eleven (50%) of our patients were treated with systemic steroids and 15 (68.2%) of them received therapeutic anticoagulation. Twenty-one of 22 patients in our cohort received 5 days of hydroxychloroquine treatment (1 patient did not due to prolonged QT interval), 6 patients received remdesivir (5 survivors and 1 nonsurvivor), and 2 patients were enrolled in a blinded remdesivir versus placebo trial. No other investigational COVID-19 treatment modalities were administered. Acute Physiology and Chronic Health Evaluation-II scores upon ICU admission were higher in nonsurvivors – median score 28 (IQR 24, 31) compared to survivors – median score 21 (IQR 16, 25); P = 0.013. Other baseline characteristics did not differ between survivors and nonsurvivors.

Figure 1

Consort diagram of inclusion and exclusion criteria

Table 1

Patient characteristics of entire cohort, survivors, and nonsurvivors

	Total cohort (n=22)	Survivors (n=13)	Nonsurvivors (n=9)	Significance
Patient characteristics
Male, n (%)	7 (32)	4 (31)	3 (33)	1.00
Female, n (%)	15 (68)	9 (69)	6 (67)
Age (IQR)	59 (44-67)	58 (42-63)	61 (54-73)	0.14
BMI (IQR)	31.4 (25.6-37.6)	31.5 (25.6-37.6)	30.5 (28.9-33.8)	0.69
Race, n (%)
White	7 (32)	4 (31)	3 (33)	0.61
African American	10 (45)	5 (38)	5 (56)
Asian	2 (9)	2 (15)	0
Hispanic	3 (14)	2 (15)	1 (11)
ARDS characteristics
Berlin category, n (%)				0.31
Mild	8 (36)	4 (31)	4 (44)
Moderate	5 (23)	2 (15)	3 (33)
Severe	9 (41)	7 (54)	2 (22)
Initial P:F (IQR)	147 (90233)	95 (90-233)	193 (178-231)	0.30
Initial Crs (IQR)	34 (28-48)	39 (34-55)	27 (24-33)	<0.01
Initial TV (cc/kg of ideal body weight) (IQR)	6 (6-6.7)	6 (6-6.7)	6 (6-6)	0.38
Respiratory rate (IQR)	28 (24-30)	28 (20-38)	30 (27-34)	0.10
ICU LOS (IQR)	18.5 (12-30)	22 (16-30)	13 (10-19)	0.16
Ventilator days (IQR)	15.5 (10-22)	19 (14-22)	13 (8-17)	0.26
Prone positioning, n (%)	7 (32)	6 (46)	1 (11)	0.09
Adjuvant therapy, n (%)
Steroids	11 (50)	8 (62)	3 (33)	0.39
No steroids	11 (50)	5 (38)	6 (67)
Therapeutic anticoagulation	15 (68)	10 (77)	5 (56)	0.38
No therapeutic anticoagulation	7 (32)	3 (23)	4 (44)
Hydroxychloroquine	21 (95)	12 (92)	9 (100)	1.00
Remdesivir	6 (27)	5 (38)	1 (11)	0.16
Vasopressors	13 (59)	7 (54)	6 (67)	0.67
Other characteristics
APACHE II score (IQR)	24 (19-29)	21 (16-25)	28 (2431)	0.013
Diabetes, n (%)	8 (36)	7 (54)	1 (11)	0.07
Smoking history, n (%)	1 (5)	1 (8)	0	1.00
Superimposed infection, n (%)	11 (50)	7 (54)	4 (44)	1.00

IQR: Interquartile range, ICU: Intensive care unit, LOS: Length of stay, ARDS: Acute respiratory distress syndrome, BMI: Body mass index, APACHE: Acute Physiology and Chronic Health Evaluation

In-hospital follow-up was complete for our cohort. Nine of the 22 (40.9%) patients with ARDS died during hospitalization. Of our total sample, 9 (40.9%) patients were able to be extubated. Eight of these patients then survived to discharge. Furthermore, 7 (31.8%) patients received a tracheostomy. Five of these patients were then liberated from the ventilator and discharged from the hospital. Of the 13 patients who survived their hospitalization, 5 were discharged to home, 6 were discharged to acute rehab, and 2 were discharged to long-term acute care hospitals.

The initial respiratory system compliance of survivors was 39 (IQR 34, 55) and 27 (IQR 24, 33) in nonsurvivors (P < 0.01). The initial P: F ratio of survivors was 95 (IQR 90, 233) and 193 (IQR 178, 231) in nonsurvivors (P = 0.30). Lower respiratory system compliance was associated with a 1.19 increased unadjusted odds of in-hospital mortality (95% confidence interval [CI] 1.01–1.40, P = 0.04).

From multivariable regression adjusting for age and receipt of steroids, the adjusted odds ratio (OR) for the association between lower respiratory compliance and in-hospital mortality was 1.2 (95% CI 1.01, 1.45 P = 0.04). In contrast, initial P: F ratio was not associated with mortality (OR 1.01, 95% CI 1.00–1.02, P = 0.23) from univariate logistic regression.

Figure 2 depicts an alluvial plot on the association between quartiles of respiratory system compliance, P: F ratio, and mortality. This plot demonstrates that patients with the highest respiratory system compliance (quartile 4) all survived to hospital discharge, while none of the patients with the lowest respiratory system compliance (quartile 1) survived to hospital discharge. This relationship was not observed for initial P: F ratio.

Figure 2

An alluvial plot breaking down initial Crs and initial P: F ratio into quartiles and showing how they relate to mortality. In this graph, quartile 1 represents patients with the lowest (worst) Crs and P: F ratio, while quartile 4 represents patients with the highest (best) Crs and P: F ratio

DISCUSSION

In our cohort of 22 patients mechanically ventilated with COVID-19 ARDS, we found that lower respiratory system compliance was associated with an increased risk of in-hospital mortality but found no association between mortality and initial P: F ratio.

From our early clinical experience in taking care of critically ill COVID-19 patients, there was an identifiable discrepancy between the severity of hypoxemia and relatively preserved work of breathing, prompting us to examine the association between oxygen parameters, respiratory mechanics and mortality. We found that respiratory system compliance was associated with in-hospital mortality. Our results are consistent with a recent secondary analysis of the EFRAIM study examining immunocompromised patients with ARDS from non-COVID etiologies.[5] In this study, lower respiratory system compliance, higher plateau pressures and driving pressure were associated with mortality in contrast to ARDS severity, as defined by Berlin Criteria, where an association was not seen.[5] The results of our study are an indication that a focus toward mechanics and away from oxygenation should be examined across the growing body of literature and data being published on COVID-19, as respiratory mechanics seem to yield important prognostic information independent of hypoxemia and Berlin criteria. Relatedly, P: F ratios are likely more variable in COVID-19 due to the high levels of reversible shunt. Recent publications on both the COVID-19 ARDS population[67] and the classic pre-COVID-19 ARDS population[8] have shown an association between lower respiratory system compliance and increased mortality[78] while also noting a disassociation between static respiratory system compliance and oxygenation.[68] While this supports our observations that lower respiratory compliance is associated with increased mortality and not related to oxygenation, the fact that initial P: F ratios were higher in non-survivors is likely attributable to the small size of our study cohort.

The characteristics of our cohort are similar to populations described in other regions of the United States, particularly in terms of body mass index and ARDS severity.[9] However, prior studies have been limited by the duration of follow-up where >35% of patients still remained in the ICU.[910] Our study followed patients to hospital discharge, revealing that over 1 in 3 patients admitted to an ICU with COVID-19 ARDS die in the hospital. This is critical information for hospital preparedness.

There are potential limitations to our study. There is the potential for confounding from unmeasured variables. In addition, given our small sample size, we were constrained in the number of variables that could be adjusted for in our regression analysis. Finally, as a quaternary referral center, our results may not be generalizable to other institutions and hospitals.

In summary, we found that lower respiratory system compliance was associated with increased mortality in adult patients with COVID-19 ARDS, but initial P: F ratio was not. Our results yield important implications for treatment management and prognostication, guiding family discussions, and furthering our understanding of this complex disease.

Research quality and ethics statement

The authors of this manuscript declare that this scientific work complies with reporting quality, formatting, and reproducibility guidelines set forth by the EQUATOR Network. The authors also attest that this clinical investigation was determined to require Institutional Review Board/Ethics Committee review, and the corresponding protocol/approval number is 843444 (approved 6/22/2020).

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.