The Association Between Alveolar–Arterial Oxygen Tension Difference and the Severity of COVID-19 in Patients

Over the past 3 years, the global pandemic of Severe Acute Respiratory Syndrome Coronavirus 2, referred to as Coronavirus Disease 2019 (COVID-19), has been, and continues to be, a significant threat to human health [1‐3]. First observed in Wuhan, China, in December 2019, the disease has caused significant economic losses and had considerably negative impacts, especially in terms of the death toll [4]. According to recent reports, 500 million have been diagnosed with COVID-19 and over 5 million have died, with the numbers continuing to rise [5]. As such, there is an urgent need to find biomarkers of disease severity and prognosis, which can facilitate early-stage intervention and ultimately save lives.

In clinical practice, the alveolar to arterial oxygen tension difference (P_A-aO₂) is used to evaluate the gas exchange function of the lungs [6], to aid in the decision of therapy [7], to measure the effect of therapy [8], and to predict the outcome in different patient groups [9]. Previous studies had reported that P_A-aO₂ value combined with low-dose chest computed tomography (CT) scan could serve as a rapid tool to select mild COVID-19 in need for hospitalization [10], and other studies had also indicated that the P_A-aO₂ value may be used as a early marker to predict severe pneumonia [11‐13]. Certain COVID-19 patients display hypoxemia and dyspnea with unclear incidences of abnormal P_A-aO₂ value, and there is an apparent association with severe events of COVID-19.

Thus, in the present study, to better understand the potential effects of P_A-aO₂ value on COVID-19 patients, we present the clinical features of COVID-19 patients with or without abnormal P_A-aO₂ value, and analyze the association between abnormal P_A-aO₂ value and the results of COVID-19 patients.

In the present study, 242 patients with laboratory-confirmed COVID-19 were recruited. Among the patients, 155 (64.05%) had abnormal P_A-aO₂ value and 87 (35.95%) had normal P_A-aO₂ value on admission.

Among the 155 patients with abnormal P_A-aO₂ value, the median age was 47 years old (IQR 36–61 years old), which was significantly higher than that of the normal PA-aO2 group of 41 years old (IQR 29–58 years, p = 0.032). Further, the patients in the abnormal P_A-aO₂ group were more likely to have symptoms of fever, cough, and shortness of breath than those in the normal P_A-aO₂ group (80.00% vs. 65.2%, p = 0.013; 50.32% vs. 33.33%, p = 0.011; and 39.35% vs. 24.14%, p = 0.016) (Table 1). In addition, the clinical outcomes of COVID-19 patients with normal P_A-aO₂ value and abnormal P_A-aO₂ value were analyzed, and the proportion of severe events in the abnormal P_A-aO₂ group was found to be higher than in the normal P_A-aO₂ group (23.23% vs. 10.34%, p = 0.013) (Table 2).

Moreover, patients with abnormal P_A-aO₂ values were more likely to develop severe events compared with patients with normal P_A-aO₂ value (HR 2.622, 95% CI 1.197–5.744, p = 0.016). After the modification for age and common comorbidities (hypertension and cardiovascular disease), patients with abnormal P_A-aO₂ value still displayed significantly increased risks of developing severe events than patients with normal P_A-aO₂ value (HR 2.986, 95% CI 1.220–7.309, p = 0.017) (Table 3).

The abnormal P_A-aO₂ value group was significantly older than the normal P_A-aO₂ group (p = 0.032), which may suggest that age was one of the factors responsible for the above change except for the P_A-aO₂ value. We then compared the baseline characteristics and clinic parameters of the abnormal P_A-aO₂ value group and the normal P_A-aO₂ group with COVID-19 matched according to age, and only found that the virus shedding time and the length of hospital stay was longer (p = 0.042, p = 0.016) in the abnormal P_A-aO₂ value group (Table S1). Additionally, we also examined the association of P_A-aO₂ value and disease severity in a multivariate logistic regression model after matching with age, and it still showed that the P_A-aO₂ value increased the risk of serious events (Table S2). Collectively, these results further proved our findings.

Subsequently, the dynamic processes and the differences between the abnormal P_A-aO₂ group and the normal P_A-aO₂ group were investigated in terms of the related laboratory coefficients. There were different trends of inflammatory biomarkers in the abnormal P_A-aO₂ group. The white blood cell (WBC) count exhibited an increasing tendency from T0 to T4, the C-reactive protein (CRP) level displayed a declining trend from T1 to T4, while the erythrocyte sedimentation rate (ESR) peaked at T2 and T3. In terms of coagulation indicators, platelets (PLTs) increased from T0 to T4, but activated partial thromboplastin time (APTT) decreased from T0 to T4. Prothrombin time (PT) declined from T0 to T3, but slightly increased in T4, and Fibrinogen (Fib) had two peaks in T0 and T1 (Table 4).

When comparing the differences between the abnormal P_A-aO₂ group and the normal P_A-aO₂ group in terms of relevant laboratory parameters, the following findings were made: the WBC count and PLTs were significantly higher in T3 and T4 in the abnormal P_A-aO₂ group; CRP and Fib were significantly higher in T0 and T1 in the abnormal P_A-aO₂ group; ESR was significantly higher in T0–T2 in the abnormal P_A-aO₂ group; D dimer (D–D) was significantly higher in T2 and T3 in the abnormal P_A-aO₂ group; lymphocytes (Lys) were significantly lower in T1 in the abnormal P_A-aO₂ group; regarding two direct indicators of liver function, alanine aminotranspherase (ALT) was significantly higher in T0 in the abnormal P_A-aO₂ group, and aspartate aminotransferase (AST) exhibited no significant differences during the study period; regarding indirect indicators, albumin (ALB) was significantly lower in T1 and T2 in the abnormal P_A-aO₂ group, and the total bilirubin (TBil) also exhibited no significant differences during the study period; regarding indicators of kidney function, serum creatinine (Cr) was significantly lower in T2 in the abnormal P_A-aO₂ group and urea nitrogen (BUN) exhibited no significant differences during the study period (Table 4).

In the present study, 242 laboratory-confirmed COVID-19-infected patients were recruited, and the clinical data were evaluated. Among those recruited, 155 (64.05%) patients had abnormal P_A-aO₂ value on admission and the abnormal P_A-aO₂ was related to subsequent severe events. Meanwhile, the data revealed a link between abnormal P_A-aO₂ value and several significant markers of inflammation/coagulopathy/fibrinolysis, especially for the early-stage inflammation parameters.

First, the demographic features, clinical symptoms, and results of the abnormal P_A-aO₂ group and the normal P_A-aO₂ group were compared. The median age of the abnormal P_A-aO₂ group was higher than the normal P_A-aO₂ group. At the same time, more severe events and more clinical symptoms, such as fever, fatigue, and shortness of breath, were found in the abnormal P_A-aO₂ group. As reported in previous studies, the P_A-aO₂ value is used to evaluate the gas exchange function of the lungs [6], and an abnormal P_A-aO₂ value indicates deficient pulmonary oxygenation [14]. Similarly, a number of studies have indicated that severe events generally occur in older patients [15, 16]. Moreover, COVID-19 always affects the lung tissue, and can impair the oxygen exchange to the blood of patients [17]. Previous studies have also reported that P_A-aO₂ value may serve as an early marker to predict severe pneumonia, which was consistent with our findings [10‐12]. Such findings can explain the present results in which abnormal P_A-aO₂ was related to severe events. Therefore, an assumption could be made that, for patients not displaying obvious hypoxemia in the early stage of the disease, those with abnormal P_A-aO₂ may already be suffering with compensatory hyperventilation and may deteriorate further. Hence, COVID-19 patients with abnormal P_A-aO₂ value should be adequately monitored, even when there are no signs of hypoxemia. Additionally, according to recent studies, COVID-19 death is generally caused by a severe case of the disease [18]. However, in the present study, there were only 2 deaths. Therefore, no relationship analysis on abnormal P_A-aO₂ and mortality could be performed, but the rate of developing severe events and the P_A-aO₂ value were significantly related according to the present research. Additionally, an observational prospective study has reported that the P_A-aO₂ value could be used to predict survival though with limited samples [13]. Hence, an assumption can be made that abnormal P_A-aO₂ value on admission may also be a potential predictive element for death, but such an assumption needs to be verified with a larger sample of studies.

Further findings have reported that the inflammation/coagulopathy/fibrinolysis parameters in the abnormal P_A-aO₂ group, such as C-reactive protein, erythrocyte sedimentation rate, fibrinogen, WBC count, and D–D were higher, while PLTs, alanine aminotranspherase, and albumin were lower compared with the normal P_A-aO₂ group. Of the parameters, the increase in WBC count and the decrease in PLTs could be regarded as hematologic biomarkers in COVID-19 patients [19, 20]. The increases in C-reactive protein and erythrocyte sedimentation rate could be regarded as inflammatory biomarkers prior to indications of critical findings with CT in COVID-19 patients [20, 21]. In respect of coagulation biomarkers, there was an increase in D–D as one of two biomarkers (the other biomarker being prothrombin time) with severe systemic disease in COVID-19 patients [20, 22, 23]. These findings indicate that abnormal P_A-aO2 values might contribute to a stronger inflammation/coagulopathy/fibrinolysis response in COVID-19 patients. However, the underlying mechanisms that account for this phenomenon remain unknown. Further studies will be performed by our group to determine which pathway dominates. Collectively, the present results show that the P_A-aO₂ value was associated with hematologic parameters, biochemical parameters, inflammatory parameters, coagulation parameters, and severe events, thereby demonstrating that the P_A-aO₂ value may be a novel potential biomarker for severe COVID-19 in patients.

Notably, the present research has several limitations. Due to the study being retrospective, there is less evidence than in prospective and interventional studies. Additionally, only a brief description of the predictive value of abnormal P_A-aO₂ on severe events in patients with COVID-19 has been provided, while the association between abnormal P_A-aO₂ value and death could not be explored due to the limited sample size.

An abnormal P_A-aO₂ value has been found to be common in COVID-19 patients, is highly related to severe event development, and could be a potential biomarker for the prognosis of COVID-19 patients.