Clinical Phenotypes of Hypoxia in Patients with COVID-19

Для цитирования: О.В.Военнов, А.В.Турентинов, К.В.Мокров, П.С.Зубеев, С.А.Абрамов. Клинические варианты гипоксии у пациентов с COVID-19. Общая реаниматология. 2021; 17 (2): 16–26. https://doi.org/10.15360/1813-97792021-2-16-26 [На русск. и англ.] For citation: Оleg V. Voennov, Аlexey V. Тurentinov, Кonstantin V. Моkrov, Pavel S. Zubееv, Sergey А. Аbramov. Clinical Phenotypes of Hypoxia in Critically Ill Patients with COVID-19. Obshchaya Reanimatologiya= General Reanimatology. 2021; 17 (2): 16–26. https://doi.org/10.15360/1813-9779-2021-2-16-26 [In Russ. and Engl.]


Introduction
The novel coronavirus infection COVID-19, which has swept the world, is a real challenge for the healthcare systems of all countries, causing an enormous burden on hospitals and intensive care units [1,2].
One of the syndromes seen in severe COVID-19 is acute respiratory failure (ARF), which most commonly results in hospitalization to medical and intensive care units [2,3].
Respiratory damage in SARS-CoV2 infection is characterized by alveolar and interstitial exudative inflammation with predominant macrophage and monocyte infiltration, as well as focal airway epithelial desquamation, pulmonary hemorrhage, and type 2 pneumocyte proliferation [4,5].
These histological changes in the lungs result in impaired gas exchange and hypoxemia, which associates with a poor prognosis [6].
Many patients with severe COVID-19 have respiratory disorders with possible development of acute respiratory distress syndrome (ARDS) [5].
At the same time, a significant part of patients with acute respiratory failure present with mild or moderate respiratory impairment with unusually severe decrease in transcutaneous oxygen saturation (SpO 2 ) [7,8].
Such discrepancy between the clinical presentations of ARF and SpO 2 reduction measured by pulse oximetry has been defined as «silent» or «happy» hypoxia [9][10][11].
As intensive care physicians know, in patients with pneumonia a significant desaturation occurs which presents with labored breathing and clinical signs such as tachypnea, forced breathing, and impaired consciousness [12,13].
In COVID-19, even in significant desaturation confirmed by transcutaneous pulse oximetry, these signs and symptoms do not always occur [14].
The study of ARF and ARDS in COVID-19 allowed to distinguish 2 types of respiratory impairment in this category of patients, depending on lung compliance [15].
The nature of this unusual condition for clinicians is under discussion [16,17].
Tissue desaturation can be caused by arterial hypoxemia, microcirculatory disorders, RBC and hemoglobin damage [18].
Another viewpoint suggests that this phenomenon is also associated with peripheral vascular microthrombosis [19].
Meanwhile, microcirculatory disorders, arterial hypoxemia and hypercapnia, as well as RBC and hemoglobin damage undoubtedly result in tissue hypoxia and acidosis development [20].
Understanding the nature of «silent hypoxia» and its differences from severe ARF is essential for choosing the proper treatment in patients with микроциркуляции, повреждением эритроцитов и гемоглобина [18].
Besides, there are doubts about the appropriateness of using tissue saturation values as an absolute indication for initiating the invasive mechanical ventilation [5,24].
This issue has influenced the evolution of approaches to tracheal intubation in patients with COVID-19 and ARF worldwide [25][26][27].
In this regard, the interest in studying the relationship between the severity of hypoxemia and hypoxia and clinical manifestations of ARF in patients with COVID-19 allows making a proper and timely decision on the choice of ARF correction strategy. Moreover, the relationship between changes in acid-base status and blood gases and the severity of hypoxia and clinical manifestations in patients with COVID-19 has not been adequately discussed.
The aim of the study was to examine the clinical phenotypes of hypoxia in patients with COVID-19 in relation to the severity of acute respiratory failure.

Materials and Methods
A multicenter prospective study with participation of 60 patients (27 men and 33 women) with severe COVID-19 associated with symptomatic ARF hospitalized in infectious disease hospitals of the University Hospital of Privolzhsky Research Medical University of the Ministry of Health of Russia and municipal hospital No. 33 (Nizhny Novgorod) was conducted. The diagnosis of COVID-19 and the severity of lung lesions were established based on the criteria specified in the Interim Clinical Guidelines of the Ministry of Health of the Russian Federation [25]. The mean age of the patients was 70 (58; 77) years.
The study included patients with reduced SatO 2 (<93%) on spontaneous breathing which required, according to the Interim Clinical Guidelines for the Treatment of Patients with COVID-19, correction of respiratory impairment. The study did not include patients who had clinical manifestations of sepsis, shock, multiple organ failure syndrome, or coma at the time of examination.
All patients on admission to ICU were divided into 2 groups according to the nature of respiratory impairment. Group 1 included patients having no breathing difficulties and no clinical signs of forced breathing, with respiratory rate (RR) under 25/min and transcutaneous oxygen saturation <93%. Group 2 consisted of patients complaining of breathing difficulties and having clinical signs of forced breathing, with RR >25/min and transcutaneous oxygen saturation <93%. The severity of disease at the time of study enrollment was assessed using the NEWS scale. The severity of lung lesions (grading) was based on chest computed tomography results (CT1 -up to 25%, CT2 -up to 50%, CT3 -up to 75%, CT4 -more than 75% of both lungs area involved). Characteristics of patients in the groups are presented in Table 1.
Patients were treated according to the current Interim Clinical Guidelines for the Diagnosis and Treatment of Patients with COVID-19, with antiviral, antimicrobial, anti-inflammatory, anticoagulant drugs and a stepwise (escalation) approach for ARF management [25].
Statistical analysis was performed using Microsoft Office Excel and Statistica 6.0 software. The Shapiro-Wilk test was used to check the normality of distribution of variables. Taking into account sampling asymmetry, values of discrete and continuous variables were presented as median and percentiles, Me (Q1; Q3). Qualitative variables were presented as numbers of cases (n), the percentage and the standard deviation of the percentage (p±sp). Small-group criteria were used for comparative analysis. Statistical significance of group differences for quantitative variables was determined by Mann-Whitney U test. Comparative assessment of statistical significance of differences for percentages was performed using the c 2 criterion.

Results and Discussion
Group 1 patients, when compared with Group 2 patients, were younger, had smaller lung lesion area on CT scan and lower scores on the NEWS severity scale.
The capillary refill time was less than 3 sec in all Group 1 patients, and their blood lactate level was not elevated. Oxygen therapy with flow rate of 5-15 l/min in prone position helped correct ARF. Meanwhile, RR decreased down to 16-22/min, on the average, to 18 (17; 20) per minute, HR decreased down to 80-90/min, on the average, to 85 (83;88) per minute. SpO 2 increased to the mean of 93 (92; 95)%. During the treatment, 4 patients in Group 1 required short-term (one day long) noninvasive CPAP or high flow oxygen therapy (Table 3). There were no fatal outcomes in Group 1. The hospital stay in this group was 12-16 days.
Group 2 patients complained of breathing difficulties and shortness of breath. Anxiety with forced breathing was noted in 15 patients. RR ranged from 25 to 46 per minute, HR from 99 to 138 per minute, SpO 2 from 65 to 85%. SpO 2 was less than 80% in 24 patients, while being within the 81-85% range in 6 other patients. Notably, unlike Group 1 patients, the six mentioned patients had increased RR up to 30 or more per minute. Venous blood pH was in the range of 7.13 2 7.27, pCO 2 was 55 to 97 mm Hg, BE was from -9 to 5 mmol/l pO 2 was 14-39 mm Hg, SО 2 was 40-60%. Arterial blood pO 2 was 41-69 mm Hg, SО 2 was 50-84%, and pCO 2 was >45 mm Hg ( Table 2).
Как следует из представленных результатов, у пациентов 1-й группы, в сравнении с were 10 fatal outcomes (33%) in this group. The hospital stay ranged from 8 to 37 days. All patients who died were admitted late (more than 10 days after the disease onset) and had high body temperature during the entire prehospital period.
As follows from the results, Group 1 patients, compared with Group 2 patients, were younger, had less severe disease and lung involvement, no hypercapnia and severe tachypnea and tachycardia. Moderate desaturation ranging from 80 to 93% was not associated with signs of tissue hypoxia such as acidosis, increased venous blood lactate, critical drop in venous blood oxygen saturation and pO 2 , prolonged capillary refill time, and impaired consciousness. It was probably due to adequate oxygen delivery to the tissues (pvO 2 over 40 mm Hg, SvO 2 over 60%), i. e. arterial hypoxemia was not accompanied by tissue hypoxia and even compensated by increased oxygen transport and utilization. Besides, the compensatory increase of BE prevented the development of acidosis. Thus, moderate arterial hypoxemia without hypercapnia, acidosis, or tissue hypoxia is typical for «silent hypoxia».
Arterial hypoxemia in the majority of patients in Group 1 (n=26), as well as oxygen desaturation were corrected by prone positioning and oxygen therapy with the flow rate of 5-15/min. This suggests that ARF in Group 1 patients was associated with the development of pulmonary atelectasis and «dead space» in posterior lung areas due to interstitial infiltration (edema) of lung tissue, and simple change of body position allowed to improve ventilation-perfusion ratio, arterial blood oxygenation, and condition of patients [10].
Thus, development of ARF in Group 2 patients is probably associated with progressive alveolar infiltration, atelectasis of lung tissue, increased dead space and pulmonary shunting [10,28].
The correction of intrapulmonary blood shunting requires both increased pressure in the respiratory system to open the alveoli and prevent their closure, and improved pulmonary blood flow [28,29]. Invasive ventilation fully affects only respiratory component of shunt formation [28,30]. Perhaps, this can explain the failure of mechanical ventilation in most patients in Group 2 (10 out of 14 patients died).
Group 1 patients were found to have no impaired tissue perfusion based on capillary refill time measurement. This technique, though, has low reliability in assessing microcirculatory disorders, therefore, abnormal oxygen saturation can be due to arterial hypoxemia. In 13 patients of Group 2, the capillary refill time was prolonged which suggests that reduced oxygen saturation could be caused both by severe arterial hypoxemia and microcirculatory and tissue perfusion disorders associated with hypercapnia, acidosis, and microvascular thrombosis [29,31].
At the same time, oxygen saturation in 6 patients from Group 2 was the same as in Group 1 patients. However, a difference in RR was observed. In Group 1, the RR was below 25 per minute, while in Group 2, it exceeded 30 per minute. Such difference can probably be explained by hypercapnia and acidosis in Group 2 patients. Thus, we can assume that SpO 2 is highly informative in clinical assessment of hypoxia severity, but RR is more important when assessing the severity of ARF in spontaneously breathing patients.
Thus, some patients with severe COVID-19 develop so-called «silent hypoxia», which presents with tissue desaturation but without severe respiratory signs and symptoms such as tachypnea, forced breathing, impaired consciousness. «Silent hypoxia» is essentially different from severe ARF. Our data suggest that patients with «silent hypoxia» contrary to those with severe ARF are characterized by younger age, less severe lung damage and disease severity, as well as moderate arterial hypoxemia, which is not accompanied by hypercapnia, acidosis, and tissue hypoxia. «Silent hypoxia» or, more properly, hypoxemia without tissue hypoxia, рапия, СРАР и НИВЛ в 14 случаях также оказалось недостаточно для коррекции ОДН, и возникла необходимость в переводе на ИВЛ.

Conclusion
In patients with COVID-19, two clinical phenotypes of hypoxia can be distinguished. The first pattern is characterized by decreased SpO 2 (80-93%) and lack of tachypnoea (defined as RR >25 per minute) which indicates moderate arterial hypoxemia without tissue hypoxia and acidosis («silent hypoxia»). It is typical for younger patients and is associated with less lung lesions and lower disease severity than in patients with severe ARF. Hypoxemia can be corrected by prone positioning and oxygen therapy and does not require initiating mechanical ventilation. The second pattern of hypoxia is characterized by significant arterial hypoxemia and hypercapnia with tissue hypoxia and acidosis. Noninvasive or invasive mechanical ventilation is required for its correction.