Introduction
Respiratory system compliance (CRS) following intubation is associated with ICU outcomes and begins to decrease after a patient is placed on mechanical ventilation.1 Mechanically ventilated patients with COVID-19 have shown a clinical presentation of air-flow obstruction and increased resistance.2 A retrospective study of subjects who succumbed to COVID-19 infection reported that 85% of the deaths were directly attributed to COVID-19, and 62% required mechanical ventilation.3 A decrease in CRS and/or PaO2/FIO24,5 has been shown to affect mortality. However, some research suggests that PaO2/FIO2 is not associated with ICU mortality.6 Indeed, the effect of COVID-19 on the respiratory system varies for subjects of different sociodemographic backgrounds.7–10 Moderate to severe COVID-19 infection affects the respiratory system and increases the possibility of pneumonitis and ARDS.11,12 Preexisting health conditions13 or developing ARDS14 is negatively associated with the likelihood of survival in subjects with COVID-19. This retrospective analysis aimed to determine whether respiratory mechanics, oxygenation impairment, social demographics, and comorbid conditions among mechanically ventilated subjects with COVID-19 infection are associated with all-cause mortality.
Methods
This study reviewed the electronic medical record at a tertiary and large correctional care facility, the University of Texas Medical Branch at Galveston (UTMB), Galveston, Texas. Inclusion criteria were age ≥ 18 y, a positive COVID-19 diagnosis, and being invasively mechanically ventilated for at least 5 consecutive days between March 1, 2020–December 31, 2021. Of the 583 subjects reviewed, 168 did not meet the inclusion criteria and were excluded from the study. The study received institutional review board (IRB) exemption from the UTMB IRB committee (IRB number 21–0216) due to the study's retrospective nature.
The primary outcome was mortality, defined as death before hospital discharge. The independent variables of interest were compliance, airway resistance (Raw), and oxygenation impairment severity. Static compliance (CRS) was calculated as (tidal volume/[plateau pressure – PEEP]). Dynamic compliance (Cdyn) was calculated as (tidal volume/[peak inspiratory pressure – PEEP]). The difference between CRS and Cdyn is attributed to Raw. If subjects had missing data for plateau pressures, the nearest average days value within 24 h was used.
The severity of oxygenation impairment was categorized using the Berlin definition15 associated with ARDS as none, mild, moderate, and severe, using PaO2/FIO2 (> 300 not impaired, 200–300 mildly impaired, 100–200 moderately impaired, and < 100 was considered severely impaired). Oxygenation impairment severity was used rather than ARDS severity because the Berlin definition of ARDS could not be satisfied (ie, lack of radiologic imaging, etc).
We collected age at the time of hospital admission, sex (female and male), race/ethnicity (non-Hispanic white, Hispanic or Latino, non-Hispanic Black, and others), and multimorbidity (2 or more comorbid conditions). The comorbidities examined were diabetes, hypertension, asthma, history of stroke, liver disease, coronary artery disease, congestive heart failure, COPD, obesity, end-stage renal disease, chronic kidney disease, and history of cancer.
Data were statistically examined for normality of distribution with the Shapiro-Wilk test. Continuous variables were presented as mean and SD and categorical variables as frequencies and percentages. T tests (normally distributed variables) or Wilcoxon tests (non-normally distributed variables) for continuous variables and chi-square tests for categorical variables were performed to assess differences by mortality status.
Generalized estimating equation (GEE) regression models were performed to estimate the odds ratio and 95% CI of factors associated with mortality between March 2020–December 2021. GEE accounts for the effects of the repeated measures on our analysis and explains correlation in clustered data since each subject can have multiple observations and missing values. The model estimation excludes missing value observations. Models were fit for mortality with yes or no as the outcome. P < .05 level of significance was used. All tests were performed in SAS version 9.4 software (SAS Institute, Cary, North Carolina).
Results
The study included 415 subjects who met the study's criteria. Descriptive characteristics of the sample by mortality status are provided in Table 1. The average age was 59 y ± 14, with the majority of the population being male, Hispanic/Latino, not having multimorbidities, and severe oxygenation impairment. Cdyn was significantly higher in the surviving subjects on day 5 of mechanical ventilation. Cdyn, on day 1 of mechanical ventilation, CRS, and Raw did not significantly differ between groups.
Descriptive Characteristics of the Sample by Mortality Status
The results from the multivariable analysis are presented in Table 2. For every 1 mL/cm H2O increase in Cdyn, the odds of mortality decrease by 3% after controlling for age, sex, race/ethnicity, multimorbidity, oxygenation impairment, Raw, and CRS. Subjects with mild, moderate, or severe oxygenation impairment had greater odds of mortality (odds ratio 2.80 [95% CI 1.30–2.46], odds ratio 3.74 [95% CI 1.69–8.3], and odds ratio 10.3 [95% CI 4.65–23], respectively) than subjects with no oxygenation impairment after controlling for all study variables. For every 1 y increase in age, the odds of mortality increased by 1.05 after controlling for all study variables. Hispanic/Latinos had 1.8 times higher odds of mortality, and non-Hispanic Blacks had 0.53 lower odds of mortality compared to their non-Hispanic white counterparts after controlling for all study variables. Males had 1.5 times higher odds of mortality than females after controlling for all study variables. Subjects with multimorbidity had 0.68 lower mortality odds than subjects without multimorbidity. No other factors included in the analysis were significantly associated with mortality.
Multivariable Analyses for Mortality Outcome
Discussion
Our analysis aimed to determine whether respiratory mechanics, oxygenation impairment, sociodemographics, and multimorbidity among mechanically ventilated subjects with COVID-19 were associated with all-cause mortality. In agreement with other studies, we found an association between CRS and mortality.5 Oxygenation impairment likewise increased mortality odds. Although Koppurapu and colleagues2 found a clinical presentation of air-flow obstruction and increased resistance in mechanically ventilated subjects with COVID-19, our study did not find Raw significantly associated with mortality.
This study's mortality rate of 60% is greater than the 35% reported by an international, multi-center study of ICU subjects with COVID-19.1 Our findings are similar to previous studies that found that increasing age was significantly associated with mortality in the mechanically ventilated COVID-19 population.8,10,13,14 Our findings are likewise similar to studies that found a significant association between race/ethnicity and mortality,7–10 specifically Hispanics/Latinos. Although the literature shows conflicting results,9 we found that non-Hispanic Blacks had significantly lower odds of mortality than non-Hispanic whites. Non-Hispanic Blacks were significantly younger than non-Hispanic whites in our study, a fact which may have contributed to this finding. Although some studies disagree,7 our study found sex significantly associated with mortality.10,13 Surprisingly, in our analysis, multimorbidity was significantly associated with lower odds of mortality compared to not having multimorbidity. This finding may be attributed to the fact that subjects in our study with multimorbidity received more attentive care than those with less comorbid conditions or because the subjects with multimorbid conditions had significantly higher PaO2/FIO2 than those without multimorbidity.
This study has some limitations. First, the study was conducted in a single academic tertiary health center and included an unknown number of incarcerated subjects, limiting the generalizability to the United States population. Second, selection bias may arise because the incarcerated subjects may have more access to care than if they were not incarcerated. Third, using the nearest average of daily values for missing data points may affect estimates of CRS and Raw estimates, as plateau pressures were not gathered consistently throughout the studied population. Several conditions, such as high spontaneous breathing rates, the influx of severely ill subjects, overwhelming resources, and pressure control ventilation, potentially kept our respiratory therapists from consistently gathering plateau pressures. However, analysis conducted without using daily averages for missing plateau pressures produced similar findings. Lastly, the novelty of COVID-19 produced frequent changes in clinical practice and new publications frequently, so our study cannot be seen as exhaustive.
Further studies should examine the pulmonary function and quality of life longitudinally for mechanically ventilated patients with COVID-19 discharged alive. Findings from this study contribute to identifying factors that could prevent or reduce mortality rates in mechanically ventilated patients with COVID-19. Sociodemographics and health inequities concerning mortality in the COVID-19 population warrant future research. This research will likely require a meta-analysis from studies like ours, with a single significant minority (Hispanic/Latino) population, and other studies that include differing minority populations, such as non-Hispanic Blacks, that may be found in different parts of the United States.
In this study, we found that an increase in Cdyn was associated with lower odds of mortality. Older age, being male, Hispanic/Latino, and having oxygenation impairment were associated with higher mortality odds. Age, sex, oxygenation impairment, and Cdyn can be considered in managing mechanically ventilated subjects with COVID-19. Future research is needed to determine if mechanical ventilation protocols aimed at improving compliance in the first 5 days might decrease mortality rates in those infected with COVID-19.
Footnotes
- Correspondence: Khamron Micheals RRT MHA, School of Public and Population Health, University of Texas Medical Branch. E-mail: khmichea{at}utmb.edu
The authors have disclosed no conflicts of interest.
Mr Micheals receives financial support through CoBGRTE Predoctoral Scholarship and the Edgar and Grace Gnitzinger Endowment.
- Copyright © 2023 by Daedalus Enterprises
