Sage Journals: Discover world-class research

Abstract

Background:

This study investigated the correlation between admission partial pressure of arterial oxygen (PaO₂) levels and both in-hospital mortality and 1-year all-cause mortality among patients diagnosed with coronavirus disease (COVID-19) pneumonia.

Methods:

This retrospective cohort study included patients with COVID-19 pneumonia admitted to the First Hospital of Jinzhou Medical University. Restricted cubic spline regression and logistic regression analyses were employed to assess the relation between PaO₂ levels and the risk of in-hospital mortality and all-cause mortality within 1 year. Subgroup analyses were performed, stratified by age, sex, presence of cardiac disease, diabetes, hypertension, whether supplemental oxygen was provided during arterial blood gas analysis, and severity of pneumonia.

Results:

The study included 737 participants with in-hospital and 1-year all-cause mortality rates of 15.7% and 26.7%, respectively. Restricted cubic spline analysis revealed an L-shaped association between admission PaO₂ levels and in-hospital mortality (P nonlinear <0.001) and a U-shaped relation with 1-year all-cause mortality (P nonlinear <0.001), with a nadir risk of 82 mmHg. Threshold analyses indicated an odds ratio of 0.931 (95% confidence interval (CI): 0.91–0.952) for in-hospital mortality and 0.951 (95% CI: 0.933–0.969) for 1-year all-cause mortality when PaO₂ was <82 mmHg. Conversely, when PaO₂ was ≥82 mmHg, the odds ratio for in-hospital mortality was 1.022 (95% CI: 0.991–1.055), and for 1-year all-cause mortality was 1.029 (95% CI: 1.004–1.054).

Conclusions:

This study revealed a nonlinear relation between PaO₂ levels at admission and both in-hospital mortality and 1-year all-cause mortality in patients with COVID-19 pneumonia, with a notable inflection point observed at approximately 82 mmHg.

Keywords

In-hospital mortality 1-year all-cause mortality coronavirus disease pneumonia cohort study

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the COVID-19 pandemic, which has profoundly impacted global health, societal structures, and economies.¹ A significant proportion of patients with COVID-19 pneumonia who experience hypoxemia require hospitalization.² Among these hospitalized individuals, approximately 15–30% may develop COVID-19-associated acute respiratory distress syndrome (ARDS).³ Despite intensive care unit (ICU) support, patients with severe cases continue to exhibit high mortality rates.^4,5 Hypoxemia in these patients is marked by a complex interplay of ventilation-perfusion mismatch, endothelial dysfunction, immunothrombosis, and neovascularization. These pathophysiological changes result in increased lung weight, altered compliance, and more pronounced dead space, leading to respiratory compromise and potentially pulmonary fibrosis.⁶ Supplemental oxygen therapy is crucial for mitigating hypoxemia and ensuring adequate tissue oxygenation.However, research indicates that excessive oxygen can worsen existing lung injuries in patients with COVID-19.⁷ The principal mechanisms involved include hyperoxia-induced epithelial cell apoptosis, disturbances in surfactant metabolism, increased oxidative stress from reactive oxygen species, and a pro-inflammatory response due to elevated angiotensin II following angiotensin-converting enzyme 2 receptor downregulation. Furthermore, imbalances in the pulmonary and gastrointestinal microbiota may increase susceptibility to secondary bacterial infections.⁷ Despite the critical importance of oxygen therapy, optimal oxygenation targets remain elusive, with no universally endorsed PaO₂ goal for critically ill patients.^8,9 Moreover, conventional indices, such as SaO₂/FiO₂ and PaO₂/FiO₂ ratios, may inadequately reflect hypoxemia severity in COVID-19 pneumonia or ARDS.^10,11 In addition, SpO₂ may not reliably predict hypoxia in COVID-19-related ARDS.¹² In view of the aforementioned studies, blood PaO₂ level is considered the most accurate indicator for assessing oxygenation status.¹³ Therefore, this study aimed to elucidate the relation between PaO₂ levels at admission, in-hospital, and 1-year all-cause mortality in patients with COVID-19 pneumonia.

Methods

Study design and participants

This retrospective cohort study included 997 individuals diagnosed with COVID-19 pneumonia between 1 December 2022 and 1 March 2023, at the First Affiliated Hospital of Jinzhou Medical University in Liaoning, China. Diagnosis of COVID-19 pneumonia is made by a combination of positive Reverse transcription polymerase chain reaction or rapid antigen testing conducted on nasal or nasopharyngeal swab samples and chest computed tomography (CT) imaging.

Relevant data on the patients’ demographic characteristics, medical history, treatment modalities, and laboratory findings were extracted from their original medical records. Demographic variables, including age, sex, and smoking status, were collected. Vital signs including body temperature (T), heart rate (HR), and respiratory rate (RR) were recorded. Hospital-related parameters, such as the length of hospital stay and utilization of mechanical ventilation (MV), noninvasive ventilation (NIV), and high-flow oxygen therapy were documented. Additionally, comorbidities, such as hypertension (HTN), diabetes mellitus, coronary heart disease (CHD), cerebrovascular disease, cancer, and chronic obstructive pulmonary disease were noted. Biochemical parameters, including aspartate aminotransferase (AST), alanine aminotransferase, albumin (ALB), urea, and creatinine levels, were measured. Hematological parameters included white blood cell (WBC) count, neutrophil percentage, lymphocyte percentage, platelet count, and hemoglobin (HB) level. Blood gas analyses included partial pressure of oxygen (PaO₂), partial pressure of carbon dioxide (PaCO₂), oxygen saturation (SaO₂), lactate levels, fraction of inspired oxygen (FiO₂%), and pH. Blood gas analyses and other hematological parameters were collected within the first 24 hours of admission. All parameters were evaluated using standard laboratory tests.

Inclusion and exclusion criteria

The inclusion criteria for this study were patients diagnosed with COVID-19 pneumonia and aged 18 years or older. The exclusion criteria were as follows: (1) patients lost to follow up, (2) pregnant individuals, and (3) patients lacking PaO₂ data or incomplete covariate information. Severe COVID-19 was defined by the presence of any of the following conditions: dyspnea with an RR ≥30/min; SaO₂ ≤ 93% on room air; arterial PaO₂/fraction of inspired oxygen (FiO₂) ≤ 300 mmHg; or lung lesion progression >50% on CT within 24–48 hours.¹⁴ Hyperoxemia and hypoxemia were defined as PaO₂ > 100 and <55–60 mmHg, respectively.⁹

Sample size

To ascertain the requisite sample size for this study, we consulted current literature indicating a 1-year all-cause mortality rate of 33.6% among patients with COVID-19.¹⁵ Utilizing this mortality rate, it was determined that a minimum sample size of 512 patients is necessary to attain a confidence level of 95% and a precision of 10%.¹⁶

Outcomes

The primary outcome was all-cause mortality during hospitalization and within 1 year after discharge. Additionally, the 1-year survival status of patients after discharge was determined through telephone follow up. The secondary outcome measure was in-hospital mortality. In-hospital death was defined as death occurring during hospitalization or within 24 hours of discharge.

Ethical considerations

This study adhered to the principles outlined in the Declaration of Helsinki and was approved by the Ethics Review Board of the First Affiliated Hospital of Jinzhou Medical University (KYLL2024384). Due to the retrospective nature of this study and the removal of all patient identifying information, informed consent was waived. The reporting of this study is in accordance with the STROBE guidelines.¹⁷

Statistical analysis

Descriptive statistics were used to summarize baseline patient characteristics. The Kolmogorov–Smirnov test was used to assess the normal distribution of variables. Continuous variables were expressed as mean ± standard deviation or median (interquartile range). Categorical variables were presented as absolute values and percentages. The chi-square test or Fisher's exact test was used to analyze categorical variables. Analysis of variance was used for normally distributed continuous variables, whereas the Kruskal–Wallis test was applied for skewed continuous variables.

Logistic regression models were used to determine the odds ratios (ORs) and 95% confidence intervals (CIs) for the association between PaO₂ levels and 1-year and in-hospital mortality. The study population was stratified into quartiles based on the PaO₂ values. The regression analysis included adjustments for various factors in the three models. Covariates were selected based on variables whose ORs changed by at least 10% after inclusion in the model, as well as clinical significance. Model 1: Adjusted for sex and age. Model 2: Adjusted for Model 1 plus HTN, diabetes mellitus, CHD, cancer, and chronic obstructive pulmonary disease. Model 3: Adjusted for Model 2 plus AST, ALB, urea, creatinine, neutrophil percentage, lymphocyte percentage, FiO₂%, and lactate. Restricted cubic spline (RCS) regression was used to assess the association between PaO₂ levels and in-hospital and 1-year mortality, respectively. A two-piecewise logistic regression model with smoothing was used to analyze the association threshold between PaO₂ and in-hospital and 1-year mortality after adjusting for the variables in Model 3. Likelihood ratio tests and the bootstrap resampling method were used to determine the inflection points.

Interaction and stratified analyses according to age, sex, HTN, diabetes mellitus, CHD, oxygen supplementation, and presence of severe pneumonia were also conducted. Statistical significance was set at p < 0.05, as determined by two-tailed testing. Statistical analyses were performed using R version 4.2.2 and Free Statistics Software version 1.8.¹⁸

Results

Patient characteristics and outcomes

Initially, 997 patients diagnosed with COVID-19 were included in this study. However, 197 patients were excluded from the analyses because of missing PaO₂ and covariate data. Additionally, 63 patients lost to follow up were excluded, leaving a final enrollment of 737 patients. Figure 1 illustrates a flowchart of the study participant screening process.

Figure 1.

The study's flow diagram.

Table 1 summarizes the demographic characteristics, vital signs, comorbidities, treatments, laboratory findings, and outcomes of the 737 participants. The mean age of the patients included in our study was 73.5 ± 12.2 years, with 425 (57.7%) being men. In-hospital mortality was observed in 116 patients (15.7%), while 1-year mortality occurred in 197 patients (26.7%).

Table 1.

Baseline demographic characteristics of the study population stratiﬁed by PaO₂.

Variables	Total (n = 737)	Q1 (n = 181)(20.5–59.9)	Q2 (n = 317)(60–79.9)	Q3 (n = 183)(80–99.5)	Q4 (n = 56)(100–188)	p
Sex, n (%)						0.462
Male	425 (57.7)	107 (59.1)	190 (59.9)	97 (53)	31 (55.4)
Age (years)	73.5 ± 12.2	75.8 ± 10.8	73.7 ± 11.3	71.5 ± 13.7	71.7 ± 14.9	0.004
LOS (days)	12.0 (8.0, 16.0)	12.0 (7.0, 17.0)	12.0 (9.0, 15.0)	11.0 (7.0, 14.0)	12.0 (9.0, 17.2)	0.037
Smoking status n (%)						0.58
Never	598 (81.1)	147 (81.2)	249 (78.5)	152 (83.1)	50 (89.3)
Ever	98 (13.3)	26 (14.4)	46 (14.5)	22 (12)	4 (7.1)
Current	41 (5.6)	8 (4.4)	22 (6.9)	9 (4.9)	2 (3.6)
T (°C)	36.8 ± 0.6	36.9 ± 0.7	36.8 ± 0.6	36.7 ± 0.5	36.9 ± 0.7	0.008
HR (bpm)	86.4 ± 16.8	91.8 ± 18.6	85.8 ± 15.6	82.9 ± 15.8	84.0 ± 16.0	<0.001
RR (bpm)	21.1 ± 4.1	22.0 ± 4.9	21.0 ± 4.1	20.6 ± 3.5	20.4 ± 2.8	0.007
CHD, n (%)	222 (30.1)	55 (30.4)	97 (30.6)	57 (31.1)	13 (23.2)	0.705
DM, n (%)	201 (27.3)	47 (26)	92 (29)	45 (24.6)	17 (30.4)	0.665
HTN, n (%)	366 (49.7)	94 (51.9)	160 (50.5)	88 (48.1)	24 (42.9)	0.642
COPD, n (%)	59 (8.0)	10 (5.5)	27 (8.5)	17 (9.3)	5 (8.9)	0.527
CVD, n (%)	147 (19.9)	36 (19.9)	67 (21.1)	34 (18.6)	10 (17.9)	0.885
Cancer, n (%)	94 (12.8)	15 (8.3)	36 (11.4)	28 (15.3)	15 (26.8)	0.002
HFNC, n (%)	91 (12.3)	50 (27.6)	27 (8.5)	11 (6)	3 (5.4)	<0.001
NIV, n (%)	38 (5.2)	24 (13.3)	11 (3.5)	3 (1.6)	0 (0)	<0.001
MV, n (%)	28 (3.8)	14 (7.7)	9 (2.8)	2 (1.1)	3 (5.4)	0.008
Severe pneumonia, n (%)						<0.001
Yes	310 (42.1)	137 (75.7)	122 (38.5)	37 (20.2)	14 (25)
No	427 (57.9)	44 (24.3)	195 (61.5)	146 (79.8)	42 (75)
One year mortality, n (%)	197 (26.7)	83 (45.9)	73 (23)	23 (12.6)	18 (32.1)	<0.001
In hospital mortality, n (%)	116 (15.7)	69 (38.1)	30 (9.5)	12 (6.6)	5 (8.9)	<0.001
FiO₂%	27.3 ± 9.9	27.5 ± 12.9	27.5 ± 9.6	26.1 ± 6.8	30.3 ± 9.1	0.042
PaO₂/FiO₂	287.7 ± 99.9	200.5 ± 60.1	276.1 ± 72.4	352.6 ± 77.3	422.8 ± 116.0	<0.001
pH	7.4 ± 0.1	7.4 ± 0.1	7.4 ± 0.1	7.4 ± 0.0	7.4 ± 0.1	0.718
PaCO₂ (mmHg)	36.0 ± 8.0	34.7 ± 7.6	36.6 ± 8.8	35.8 ± 6.2	37.3 ± 9.0	0.05
PO₂ (mmHg)	73.1 ± 20.6	49.5 ± 8.4	70.0 ± 5.7	87.1 ± 5.3	120.3 ± 18.6	<0.001
SaO₂%	91.6 ± 8.4	81.6 ± 11.6	93.3 ± 2.5	96.5 ± 1.1	98.2 ± 0.9	<0.001
cLac (mmol/L)	1.8 (1.3, 2.5)	2.1 (1.6, 2.9)	1.8 (1.3, 2.3)	1.5 (1.2, 2.0)	1.5 (1.2, 2.2)	<0.001
WBC (10^9/L)	6.5 (4.7, 9.0)	8.2 (5.7, 10.9)	6.3 (4.6, 8.4)	6.1 (4.5, 8.3)	5.4 (4.0, 7.3)	<0.001
NEU%	76.0 ± 13.4	80.3 ± 15.3	75.7 ± 12.5	73.6 ± 11.9	71.7 ± 12.7	<0.001
Lymphocyte%	12.9 (7.4, 19.6)	8.3 (5.2,13.2)	13.3 (7.6, 19.8)	15.8 (10.2, 22.3)	17.4 (12.7, 22.9)	<0.001
HB (g/L)	124.9 ± 28.1	125.7 ± 35.2	124.4 ± 20.1	127.1 ± 33.0	117.4 ± 22.5	0.146
PLT (10^9/L)	202.2 ± 93.4	204.1 ± 105.2	209.8 ± 92.4	197.2 ± 87.9	169.8 ± 67.4	0.024
ALT (U/L)	23.0 (15.0, 37.0)	27.0 (18.0, 39.0)	24.0 (15.0, 38.0)	21.0 (14.5, 34.0)	17.5 (12.0, 28.0)	0.004
AST (U/L)	26.0 (18.0, 38.0)	29.0 (21.0, 43.0)	26.0 (19.0, 38.0)	24.0 (18.0, 34.0)	20.0 (16.0, 31.8)	<0.001
ALB (g/L)	33.6 ± 5.0	31.5 ± 4.3	33.5 ± 4.6	35.5 ± 5.2	34.6 ± 5.4	<0.001
Urea (umol/L)	6.1 (4.8, 8.7)	7.1 (5.2, 10.0)	6.1 (4.8, 8.4)	5.8 (4.4, 7.7)	6.0 (4.6, 9.2)	<0.001
Creatinine (umol/L)	72.4 (58.8, 95.9)	73.8 (58.9, 98.9)	71.5 (59.4, 97.3)	71.3 (58.9, 91.8)	70.7 (53.2, 92.4)	0.592

ALB: albumin; ALT: alanine aminotransferase; AST: aspartate aminotransferase; CHD: coronary heart disease; cLac: lactate; COPD: chronic obstructive pulmonary disease; CVD: cerebrovascular disease; HB: hemoglobin; HTN: hypertension; HR: heart rate; DM: diabetes mellitus; HFNC: high flow nasal cannula; LOS: length of stay; T: body temperature; RR: respiratory rate; MV: mechanical ventilation; NEU%: neutrophil percentage; NIV: noninvasive ventilation; PaCO₂: partial pressure of carbon dioxide; PaO₂: partial pressure of oxygen; PLT: platelet; Q: quartiles; WBC: white blood cell count.

PaO₂ and in-hospital mortality

Univariate analysis demonstrated associations between in-hospital mortality and various factors, including age, length of stay, T, HR, RR, CHD, HTN, FiO₂%, arterial PaO₂, PaO₂/FiO₂ ratio, arterial PaCO₂, arterial SaO₂%, lactate concentration, WBC count, percentage of neutrophils, percentage of lymphocytes, HB, platelet count, AST, ALB, urea, creatinine, use of high-flow nasal cannula (HFNC), NIV, invasive MV, and severe pneumonia (Supplemental Table S1).

When PaO₂ was analyzed in quartiles, a significant inverse association was observed between PaO₂ and in-hospital mortality after adjusting for potential confounders. Compared with observations in individuals with PaO₂ in the first quartile (20.5–59.9 mmHg), the adjusted OR values for PaO₂ in the second quartile (60–79.9 mmHg), third quartile (80–99.5 mmHg), and fourth quartile (100–188 mmHg) were 0.2 (95% CI: 0.12–0.35, p < 0.001), 0.19 (95% CI: 0.09–0.4, p < 0.001), and 0.22 (95% CI: 0.08–0.64, p = 0.005), respectively (Table 2). Consequently, the association between PaO₂ and in-hospital mortality demonstrated an L-shaped curve (nonlinear, p < 0.001) in RCSs (Figure 2(a)).

Figure 2.

(a) The nonlinear relationship between PaO₂ and in-hospital mortality in patients with COVID-19 pneumonia. Only 99.5% of the data is displayed. (b) Nonlinear relationship between PaO₂ and all-cause mortality at 1 year in COVID-19 pneumonia patients; adjusted for age, sex, hypertension, diabetes mellitus, coronary heart disease, cancer, chronic obstructive pulmonary disease, aspartate aminotransferase, albumin, urea, creatinine, neutrophil percentage, lymphocyte percentage, FiO₂%, and lactate.

Table 2.

Association of PaO₂ levels with in-hospital and 1-year all-cause mortality in different models.

Variable	n.total	n. event%	Unadjusted		Model1		Model2		Model3
Variable	n.total	n. event%	OR (95% CI)	P	OR(95% CI)	P	OR(95% CI)	P	OR(95% CI)	p
In-hospital mortality
PaO₂ (mmHg)	737	116 (15.7)	0.95 (0.94∼0.96)	<0.001	0.95 (0.94∼0.97)	<0.001	0.95 (0.94∼0.96)	<0.001	0.96 (0.95∼0.98)	<0.001
PaO₂ quartiles
Q1 (20.5–59.9)	181	69 (38.1)	1 (Ref)		1 (Ref)		1(Ref)		1 (Ref)
Q2 (60–79.9)	317	30 (9.5)	0.17 (0.1∼0.27)	<0.001	0.18 (0.11∼0.29)	<0.001	0.17 (0.1∼0.28)	<0.001	0.2 (0.12∼0.35)	<0.001
Q3 (80–99.5)	183	12 (6.6)	0.11 (0.06∼0.22)	<0.001	0.12 (0.06∼0.24)	<0.001	0.12 (0.06∼0.23)	<0.001	0.19 (0.09∼0.4)	<0.001
Q4 (100–188)	56	5 (8.9)	0.16 (0.06∼0.42)	<0.001	0.17 (0.06∼0.45)	<0.001	0.16 (0.06∼0.43)	<0.001	0.22 (0.08∼0.64)	0.005
P for trend				<0.001		<0.001		<0.001		<0.001
One-year mortality
PaO₂ (mmHg)	737	197 (26.7)	0.98 (0.97∼0.99)	<0.001	0.98 (0.97∼0.99)	<0.001	0.97 (0.96∼0.98)	<0.001	0.98 (0.97∼0.99)	0.002
PaO₂ quartiles
Q1 (20.5–59.9)	181	83 (45.9)	5.89 (3.48∼9.97)	<0.001	5.39 (3.15∼9.22)	<0.001	6.41 (3.67∼11.21)	<0.001	4.05 (2.2∼7.46)	<0.001
Q2 (60–79.9)	317	73 (23)	2.08 (1.25∼3.46)	0.005	1.99 (1.18∼3.35)	0.009	2.18 (1.28∼3.71)	0.004	1.77 (1.01∼3.12)	0.047
Q3 (80–99.5)	183	23 (12.6)	1 (Ref)		1 (Ref)		1 (Ref)		1 (Ref)
Q4 (100–188)	56	18 (32.1)	3.3 (1.62∼6.71)	0.001	3.39 (1.62∼7.09)	0.001	3.23 (1.51∼6.92)	0.003	2.86 (1.26∼6.49)	0.012
P for trend				<0.001		<0.001		<0.001		0.002

Model 1 adjusts for age, sex.

Model 2 adjusted for age, sex, hypertension, diabetes, coronary heart disease, cancer, and chronic obstructive pulmonary disease.

Model 3 adjusted for age, sex, hypertension, diabetes mellitus, coronary heart disease, cancer, chronic obstructive pulmonary disease, aspartate aminotransferase, albumin, urea, creatinine, neutrophil percentage, lymphocyte percentage, FiO₂%, and lactate.

CI: conﬁdence interval; OR: odds ratio; Q: quartiles; Ref: reference.

In the threshold analysis, the OR for in-hospital mortality was 0.931 (95% CI: 0.91–0.952, p < 0.001) in the participants with PaO₂ < 82 mmHg (Table 3). This implies that the risk of in-hospital mortality decreases by 6.9% with every 1 mmHg increase in PaO₂. No association between PaO₂ and in-hospital mortality was observed when PaO₂ was ≥82 mmHg (Table 3), indicating that the risk of in-hospital mortality ceases to decrease with increasing PaO₂.

Table 3.

Threshold effect analysis of the relationship between PaO₂ levels and both in-hospital mortality and 1-year all-cause mortality.

Threshold of PaO₂	OR	95% CI	p
In-hospital mortality
<82	0.931	0.91∼0.952	<0.001
≥82	1.022	0.991∼1.055	0.1704
Likelihood ratio test			<0.001
One-year mortality
<82	0.951	0.933∼0.969	<0.001
≥82	1.029	1.004∼1.054	0.0201
Likelihood ratio test			<0.001

Adjusted for age, sex, hypertension, diabetes mellitus, coronary heart disease, cancer, chronic obstructive pulmonary disease, aspartate aminotransferase, albumin, urea, creatinine, neutrophil percentage, lymphocyte percentage, FiO₂%, and lactate.

CI: conﬁdence interval; OR: odds ratio.

PaO₂ and 1-year all-cause mortality

Univariate analysis revealed associations between 1-year mortality and factors, including age, T, HR, RR, CHD, HTN, cancer, FiO₂%, PaO₂/FiO₂ ratio, arterial PaO₂, arterial SaO₂, lactate concentration, WBC count, percentage of neutrophils, percentage of lymphocytes, HB, platelet count, AST, ALB, urea, creatinine, use of HFNC, NIV, invasive MV, and severe pneumonia (Supplemental Table S2).

When PaO₂ was analyzed by quartiles, a significant nonlinear relation between PaO₂ and 1-year all-cause mortality was observed after adjusting for potential confounders. Compared with observations in individuals in the third quartile of PaO₂ (80–99.5 mmHg), the adjusted OR values for PaO₂ and 1-year all-cause mortality in the first quartile (20.5–59.9 mmHg), second quartile (60–79.9 mmHg), and fourth quartile (100–188 mmHg) were 4.05 (95% CI: 2.2–7.46, p < 0.001), 1.77 (95% CI: 1.01–3.12, p = 0.047), and 2.86 (95% CI: 1.26–6.49, p = 0.012), respectively (Table 2). Consequently, the association between PaO₂ and 1-year all-cause mortality demonstrated a U-shaped curve (nonlinear, p < 0.001) in RCS (Figure 2(b)).

In threshold analysis, a threshold value of approximately 82 mmHg was identified. Below this threshold, the OR for 1-year mortality was 0.951 (95% CI: 0.933–0.969, p < 0.001), while above this threshold, the OR for 1-year mortality was 1.029 (95% CI: 1.004–1.054, p = 0.0201) (Table 3). This represents a 4.9% reduction in the 1-year risk of all-cause mortality for every 1 mmHg increase in PaO₂ when PaO₂ was below 82 mmHg and a 2.9% increase in the 1-year risk of all-cause mortality for every 1 mmHg increase in PaO₂ when PaO₂ was 82 mmHg or higher.

Subgroup and interaction analyses

Subgroup analyses were conducted to assess the potential impact of the relation between PaO₂ and in-hospital and 1-year all-cause mortality. Stratification by age, diabetes mellitus, CHD, HTN, oxygen supplementation, and severe pneumonia revealed no significant interactions in any subgroup (Figure 3). However, the p-value for the sex interaction was <0.05, warranting further exploration.

Figure 3.

(a) The relationship between PaO₂ and in-hospital mortality according to basic features. (b) The relationship between PaO₂ and all-cause mortality at 1 year according to basic features. Except for the stratification component itself, each stratification factor was adjusted for all other variables (age, sex, hypertension, diabetes mellitus, coronary heart disease, cancer, chronic obstructive pulmonary disease, aspartate aminotransferase, albumin, urea, creatinine, neutrophil percentage, lymphocyte percentage, FiO₂%, and lactate).

Discussion

This retrospective cohort study revealed a robust correlation between PaO₂ levels at admission and both in-hospital and 1-year all-cause mortality in patients diagnosed with COVID-19. Moreover, even after adjusting for covariates, the independent nonlinear relations persisted. During the threshold analysis, it was observed that when PaO₂ levels fell below 82 mmHg, the OR for in-hospital mortality was 0.931 (95% CI: 0.91–0.952, p < 0.001), and for 1-year all-cause mortality, it was 0.951 (95% CI: 0.933–0.969, p < 0.001). This indicates that with each 1 mmHg rise in PaO₂, there was a corresponding decrease in the risk of in-hospital death by 6.9% and a decrease in the 1-year risk of death by 4.9%.Conversely, when PaO₂ levels were at or above 82 mmHg, participants exhibited an OR for in-hospital mortality of 1.022 (95% CI: 0.991–1.055, p = 0.1704), and for 1-year all-cause mortality, it was 1.029 (95% CI: 1.004–1.054, p = 0.0201). These findings suggest that there is no discernible association between each 1 unit increase in PaO₂ and in-hospital mortality. However, there was a notable 2.9% increase in the 1-year risk of all-cause mortality.

Several studies have explored the association between hyperoxia and mortality outcomes in critically ill patients, particularly those with COVID-19-related ARDS. A study found no significant difference in outcomes between hyperoxemia and normoxemia groups among invasively ventilated COVID-19 ARDS patients,¹⁹ consistent with previous research indicating that hyperoxia alone is not independently linked to in-hospital or 6-month mortality²⁰ and that the overuse of oxygen does not correlate with mortality.²¹ Moreover, investigations have revealed a U-shaped relation between PaO₂ levels and mortality and morbidity in patients admitted to ICUs. The lowest mortality rates were observed at PaO₂ levels ranging from 100 to 150 mmHg.⁸ However, it is important to note that such supraphysiological PaO₂ values are generally not recommended due to potential adverse effects. Recent observational studies have further elucidated this U-shaped relation. For instance, Boyle AJ identified that patients with a time-weighted PaO₂ range of 93.8 to 105 mmHg exhibited the lowest risk of death among those with ARDS.²² In addition, de Jonge et al.²³ demonstrated a similar U-shaped relation between PaO₂ levels and mortality in patients in the ICU receiving mechanical ventilation, supporting the notion that optimal oxygenation levels are crucial for patient outcomes.

In the context of invasively ventilated patients with COVID-19, Tsonas et al.¹⁹ reported that 28-day mortality exhibited a plateau relation with PaO₂ levels on day 1 and a U-shaped relation on day 2, with a minimum PaO₂ threshold of 75 mmHg. These findings underscore the complexity of oxygenation management in critically ill patients and highlight the need for individualized approaches to optimize outcomes while avoiding hypoxia and hyperoxia. This investigation delineated a nonlinear association between PaO₂ at admission and both in-hospital and 1-year all-cause mortality in patients with COVID-19 pneumonia, with the lowest mortality risk observed at a PaO₂ of approximately 82 mmHg. It has been underscored that while maintaining PaO₂ within the normal range, careful titration is essential to avert hypoxemia and excessive hyperoxemia.²⁴

Oxygen plays a dual role in immune response modulation and tissue homeostasis. It enhances the respiratory burst of neutrophils and exerts antibacterial effects by inhibiting bacterial proliferation in tissues.²⁵ However, the administration of hyperbaric oxygen can lead to oxidative stress,²⁶ posing a particular threat to pulmonary tissues. For instance, one study demonstrated that after administering 28% oxygen to healthy volunteers for only 1 hour, exhaled breath condensate measurements revealed elevated levels of both interleukin-6 and 8-isoprostane.²⁷

Numerous investigations have highlighted the prevalence of excessive oxygen administration and resulting hyperoxemia among mechanically ventilated patients with SARS-CoV-2 pneumonia. Exposure to hyperoxemia has been correlated with an increased risk of ICU mortality and ventilator-associated pneumonia.²⁸ Hyperoxia has multiple effects on various organs. It can directly inflict tissue damage by triggering excessive production of reactive oxygen species, surpassing physiological antioxidant defenses.²⁹ This phenomenon results in increased apoptosis and an augmented release of endogenous damage-associated molecular pattern molecules, which, in turn, stimulate inflammatory responses, particularly in the pulmonary milieu.³⁰ Moreover, hyperoxia-induced vasoconstriction may result from diminished NO levels.³¹

In the pulmonary context, hyperoxia causes oxidative stress and inflammation,³² potentially compromising the surface-active substance system, thereby precipitating alveolar collapse and diminishing lung compliance.³³ Additionally, excessive oxygen administration may impede mucosal ciliary clearance and attenuate the antimicrobial capacity of immune cells, predisposing individuals to ventilator-associated pneumonia.³⁴ Given these detrimental effects, it is prudent to avoid excessive oxygenation, with recommendations suggesting the maintenance of PaO₂ levels below 100 to 120 mmHg.³⁵

COVID-19 pneumonia manifests with three characteristic features: dysregulation of hemostasis leading to heightened clotting propensity, inflammation, and endothelial cell injury, all of which significantly contribute to the development of respiratory failure, particularly in severe cases. Early-stage COVID-19 pneumonia primarily affects the periphery.^36,37 Owing to the inability of the body to retain oxygen efficiently, even brief episodes of hypoxemia can rapidly induce irreversible tissue damage, precipitating organ dysfunction (including hypoxic brain injury, stroke, myocardial infarction, and acute kidney injury) or mortality.³⁸ Hypoxemia in COVID-19 arises from a heterogeneous mismatch between pulmonary ventilation and perfusion, which is primarily driven by immunothrombosis, endothelial dysfunction, and neovascularization. Historically, clinicians have favored aggressive oxygen supplementation to mitigate these risks. However, the necessity of oxygen therapy varies among individuals, with patients having hypoxemia clearly requiring supplemental oxygen therapy. Nonetheless, oxygen requirements should be routinely reassessed, and indiscriminate induction of hyperoxemia should be avoided unless supported by a compelling clinical justification to mitigate the potential risks associated with hyperoxia.^39,40 Studies have indicated that PaO₂ < 60 and PaO₂ > 100 are associated with unfavorable prognoses.⁴¹

This study had several strengths. Primarily, it offers a comprehensive examination of the relation between arterial PaO₂ and 1-year all-cause and in-hospital mortality among patients afflicted with COVID-19 pneumonia. This rigorous analysis provides invaluable insights into the development of oxygen therapy strategies tailored to patients with COVID-19 pneumonia. Second, the PaO₂ acquisition method employed in this study was both straightforward and cost-effective. Third, the use of multivariate and subgroup analyses bolsters the robustness and consistency of the findings.

However, this study has several limitations. First, its retrospective design and confinement to a single center may have inherently skewed the results. Second, the absence of precise timing and underlying causes of death imposes constraints on data interpretation. Third, the recorded PaO₂ values represent only the initial measurements upon admission, precluding the assessment of dynamic fluctuations in PaO₂ throughout the hospitalization period. Finally, as a correlational study, it lacks the capacity to definitively infer causality.

Conclusions

In conclusion, this study highlights a curvilinear relationship between PaO₂ levels at hospital admission and both short-term and long-term mortality rates in patients with COVID-19 pneumonia, with a turning point around 82 mmHg. Physicians may consider adjusting treatment plans in a timely manner based on arterial oxygen partial pressure, aiming to make rational use of oxygen resources and potentially reducing the risk of mortality associated with hyperoxemia and hypoxemia.

Supplemental Material

sj-docx-1-sci-10.1177_00368504241310737 - Supplemental material for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study

Supplemental material, sj-docx-1-sci-10.1177_00368504241310737 for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study by Ruoqing Zhou and Dianzhu Pan in Science Progress

Supplemental Material

sj-doc-2-sci-10.1177_00368504241310737 - Supplemental material for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study

Supplemental material, sj-doc-2-sci-10.1177_00368504241310737 for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study by Ruoqing Zhou and Dianzhu Pan in Science Progress

Footnotes

List of abbreviations

Acknowledgements

We thank the Free Statistics team for providing technical assistance and valuable tools for data analysis and visualization.

Authors’ contributions

DP designed the study and performed the statistical analyses. RZ drafted the manuscript and collected data. All the authors have read and approved the final manuscript.

Availability of data and materials

The datasets generated and/or analyzed in the current study are available from the corresponding author upon reasonable request.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Ethics approval and consent to participate

The study was approved by the Ethics Review Committee of the First Affiliated Hospital of Jinzhou Medical University (approval number: KYLL2024384).

ORCID iD

Ruoqing Zhou

Supplemental material

Supplemental material for this article is available online.

References

WHO Coronavirus (COVID-19) Dashboard. Available online: https://covid19.who.int

Siddiqi

Mehra

. COVID-19 illness in native and immunosuppressed states: a clinical-therapeutic staging proposal. J Heart Lung Transplant 2020; 39: 405–407.

Attaway

Scheraga

Bhimraj

, et al. Severe COVID-19 pneumonia: pathogenesis and clinical management. Br Med J 2021; 372: n436.

Arentz

Yim

Klaff

, et al. Characteristics and outcomes of 21 critically ill patients with COVID-19 in Washington state. JAMA 2020; 323: 1612–1614.

Yang

, et al. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. Lancet Respir Med 2020; 8: 475–481.

Camporota

Cronin

Busana

, et al. Pathophysiology of coronavirus-19 disease acute lung injury. Curr Opin Crit Care 2022; 28: 9–16.

Dushianthan

Bracegirdle

Cusack

, et al. Alveolar hyperoxia and exacerbation of lung injury in critically ill SARS-CoV-2 pneumonia. Med Sci (Basel) 2023; 11(4): 70.

Barbateskovic

Schjorring

Russo

, et al. Higher versus lower fraction of inspired oxygen or targets of arterial oxygenation for adults admitted to the intensive care unit. Cochrane Database Syst Rev 2019; 2019(11): CD012631.

Demiselle

Calzia

Hartmann

, et al. Target arterial PO(2) according to the underlying pathology: a mini-review of the available data in mechanically ventilated patients. Ann Intensive Care 2021; 11: 88.

10.

Aboab

Louis

Jonson

, et al. Relation between PaO₂/FiO₂ ratio and FiO₂: a mathematical description. Intensive Care Med 2006; 32: 1494–1497.

11.

Gowda

Klocke

. Variability of indices of hypoxemia in adult respiratory distress syndrome. Crit Care Med 1997; 25: 41–45.

12.

Nguyen

Helias

Raia

, et al. Impact of COVID-19 on the association between pulse oximetry and arterial oxygenation in patients with acute respiratory distress syndrome. Sci Rep 2022; 12: 1462.

13.

Tobin

Jubran

Laghi

. P (ao(2)) /F (io(2)) ratio: the mismeasure of oxygenation in COVID-19. Eur Respir J 2021; 57(13): 2100274.

14.

Shi

Huang

, et al. Effect of human umbilical cord-derived mesenchymal stem cells on lung damage in severe COVID-19 patients: a randomized, double-blind, placebo-controlled phase 2 trial. Signal Transduct Target Ther 2021; 6: 58.

15.

Novelli

Raimondi

Carioli

, et al. One-year mortality in COVID-19 is associated with patients’ comorbidities rather than pneumonia severity. Respir Med Res 2023; 83: 100976.

16.

Hsieh

Bloch

Larsen

. A simple method of sample size calculation for linear and logistic regression. Stat Med 1998; 17: 1623–1634.

17.

von Elm

Altman

Egger

, et al. The strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577.

18.

Zhou

Pan

. Association between admission heart rate and in-hospital mortality in patients with acute exacerbation of chronic obstructive pulmonary disease and respiratory failure: a retrospective cohort study. BMC Pulm Med 2024; 24: 111.

19.

Tsonas

van Meenen

Botta

, et al. Hyperoxemia in invasively ventilated COVID-19 patients-insights from the provent-COVID study. Pulmonology 2022; 30(3): 272–281.

20.

Raj

Bendel

Reinikainen

, et al. Hyperoxemia and long-term outcome after traumatic brain injury. Crit Care 2013; 17: R177.

21.

Gomes

Reboredo

Costa

, et al. Impacts of a fraction of inspired oxygen adjustment protocol in COVID-19 patients under mechanical ventilation: a prospective cohort study. Med Intensiva (Engl Ed) 2023; 47: 212–220.

22.

Boyle

Holmes

Hackett

, et al. Hyperoxaemia and hypoxaemia are associated with harm in patients with ARDS. BMC Pulm Med 2021; 21: 285.

23.

de Jonge

Peelen

Keijzers

, et al. Association between administered oxygen, arterial partial oxygen pressure and mortality in mechanically ventilated intensive care unit patients. Crit Care 2008; 12: R156.

24.

Lassen

Risgaard

Baekgaard

, et al. Determining a safe upper limit of oxygen supplementation for adult patients: a systematic review. BMJ Open 2021; 11: e45057.

25.

Soh

Pietrobon

Freiberger

, et al. Hyperbaric oxygen therapy in necrotising soft tissue infections: a study of patients in the United States nationwide inpatient sample. Intensive Care Med 2012; 38: 1143–1151.

26.

Moradkhan

Sinoway

. Revisiting the role of oxygen therapy in cardiac patients. J Am Coll Cardiol 2010; 56: 1013–1016.

27.

Carpagnano

Kharitonov

Foschino-Barbaro

, et al. Supplementary oxygen in healthy subjects and those with COPD increases oxidative stress and airway inflammation. Thorax 2004; 59: 1016–1019.

28.

Damiani

Casarotta

Carsetti

, et al. Too much tolerance for hyperoxemia in mechanically ventilated patients with SARS-CoV-2 pneumonia? Report from an Italian intensive care unit. Front Med (Lausanne) 2022; 9: 957773.

29.

Brueckl

Kaestle

Kerem

, et al. Hyperoxia-induced reactive oxygen species formation in pulmonary capillary endothelial cells in situ. Am J Respir Cell Mol Biol 2006; 34: 453–463.

30.

Zaher

Miller

Morrow

, et al. Hyperoxia-induced signal transduction pathways in pulmonary epithelial cells. Free Radic Biol Med 2007; 42: 897–908.

31.

Sjoberg

Singer

. The medical use of oxygen: a time for critical reappraisal. J Intern Med 2013; 274: 505–528.

32.

Damiani

Donati

Girardis

. Oxygen in the critically ill: friend or foe? Curr Opin Anaesthesiol 2018; 31: 129–135.

33.

Bailey

Martin

Zhao

, et al. High oxygen concentrations predispose mouse lungs to the deleterious effects of high stretch ventilation. J Appl Physiol (1985) 2003; 94: 975–982.

34.

Six

Jaffal

Ledoux

, et al. Hyperoxemia as a risk factor for ventilator-associated pneumonia. Crit Care 2016; 20: 195.

35.

Vincent

De Backer

. Circulatory shock. N Engl J Med 2013; 369: 1726–1734.

36.

Shi

Han

Jiang

, et al. Radiological findings from 81 patients with COVID-19 pneumonia in Wuhan, China: a descriptive study. Lancet Infect Dis 2020; 20: 425–434.

37.

Fox

Akmatbekov

Harbert

, et al. Pulmonary and cardiac pathology in African American patients with COVID-19: an autopsy series from New Orleans. Lancet Respir Med 2020; 8: 681–686.

38.

Leach

Treacher

. The pulmonary physician in critical care * 2: oxygen delivery and consumption in the critically ill. Thorax 2002; 57: 170–177.

39.

Moon

. Oxygen in acute illness: more or less? Crit Care Med 2015; 43: 1547–1548.

40.

Vincent

Taccone

. Harmful effects of hyperoxia in postcardiac arrest, sepsis, traumatic brain injury, or stroke: the importance of individualized oxygen therapy in critically ill patients. Can Respir J 2017; 2017: 2834956.

41.

Sartini

Massobrio

Cutuli

, et al. Role of SatO₂, PaO₂/FiO₂ ratio and PaO₂ to predict adverse outcome in COVID-19: a retrospective, cohort study. Int J Environ Res Public Health 2021; 18(21): 11534.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.11 MB

0.02 MB

Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study

Abstract

Background:

Methods:

Results:

Conclusions:

Keywords

Introduction

Methods

Study design and participants

Inclusion and exclusion criteria

Sample size

Outcomes

Ethical considerations

Statistical analysis

Results

Patient characteristics and outcomes

PaO2 and in-hospital mortality

PaO2 and 1-year all-cause mortality

Subgroup and interaction analyses

Discussion

Conclusions

Supplemental Material

sj-docx-1-sci-10.1177_00368504241310737 - Supplemental material for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study

Supplemental Material

sj-doc-2-sci-10.1177_00368504241310737 - Supplemental material for Nonlinear association between admission oxygen partial pressure and in-hospital and one-year all-cause mortality in patients with coronavirus disease pneumonia: A retrospective cohort study

Footnotes

List of abbreviations

Acknowledgements

Authors’ contributions

Availability of data and materials

Declaration of conflicting interests

Funding

Ethics approval and consent to participate

ORCID iD

Supplemental material

References

Supplementary Material

PaO₂ and in-hospital mortality

PaO₂ and 1-year all-cause mortality