Hypoxia in COVID-19: Sign of Severity or Cause for Poor Outcomes

In December 2019, reports of an outbreak of an unusual viral pneumonia were released from Wuhan, China, which was later found to be due to infection with novel severe acute respiratory syndrome coronavirus 2. In the following months, this viral infection, known as coronavirus disease 2019 (COVID-19), became a pandemic that now is engaging an ever-growing number of individuals across the globe, with a high rate of morbidity and mortality. Coronavirus disease 2019 not only has affected health care systems at the global levels but also has led to a substantial downturn in the world economy, both of which may increase the adverse effects of this infectious disease, particularly in resource-limited areas. Although the COVID-19 pandemic has led to many misfortunate life and livelihood losses, it has also resulted in solidarity at the global level in patient care and instigated innovative approaches to prevent, manage, or treat this disease.¹

There are several characteristics of COVID-19 which have caused considerable concern. The patterns of transmission of this virus are progressively recognized. It appears that transmission of this virus is mostly via droplets unless infected individuals undergo aerosol-generating procedures that result in the airborne mode of transmission. Indeed, places that implemented social distancing, hand washing, and face masks as a priority have had success in controlling the spread of this virus. The other concern is that asymptomatic individuals could also shed the virus and thus contribute to its rapid spread in communities.² Therefore, widespread testing and contact tracking of infected individuals could also result in a slower spread of this disease. Large areas of uncertainty exist regarding COVID-19, and these include the extent of immunity after recovery from COVID-19, inherent and environmental risk factors of more severe illnesses, and a global or local consensus on preventive, management, or therapeutic options for COVID-19.¹

One of the substantial challenges related to COVID-19 is the high incidence of multiorgan involvement as compared with other viral infections (ie, lungs, heart, kidney, gastrointestinal tract, coagulation system,³ etc). However, it seems that the respiratory system is one of the most commonly engaged organs. Coronavirus disease 2019–associated pneumonia could lead to acute respiratory distress syndrome (ARDS), and the characteristics of COVID-19–associated ARDS may be similar or differ from those seen in ARDS due to other causes. Specifically, COVID-19–associated ARDS engages older individuals and those with comorbid conditions (eg, hypertension and diabetes mellitus)⁴; it is associated with significant dyspnea⁴; it presents with different phenotypes (ie, L vs H phenotypes that differ by lung elastance, ventilation to perfusion ratio, right-to-left shunt, and lung recruitability⁵); it imposes hypoxia that could be due to high shunt physiology (ie, hypoventilated areas of the lung are hyperemic,⁶ particularly in the H phenotype); patients so afflicted require a prolonged period to resolve their ARDS⁷; and it has high mortality rates (51%).⁴ These patients frequently require a higher level of care in hospitals or intensive care units, and de-escalation to a lower level of care or discharge may require several days. These factors have led to tremendous pressure on the health care systems, particularly in COVID-19 hot spots. The ability to triage patients who may need care level escalation could not only assist with appropriate bed assignment and prevention of health care overflow but could also potentially improve patients’ outcomes by earlier initiation of preventive and management measures.

In this issue of Mayo Clinic Proceedings, Xie et al⁸ report the results of a retrospective cohort study of 140 patients with confirmed or presumed COVID-19 who presented with relevant signs and symptoms, with positive COVID-19 real-time reverse transcription–polymerase chain reaction test results present in most patients. These patients received medical attention in hospitals or intensive care units in medical centers in Beijing, China, over 1 month that the study was conducted. The authors reported low peripheral capillary oxygen saturation (Spo ₂; with the cutoff of 90%) after receiving oxygen support along with the presence of dyspnea to be a strong predictor of mortality. In addition, they suggested leukocytosis with a left shift along with C-reactive protein levels as a possible predictor of mortality in patients with COVID-19.

This study is commendable for being able to identify clinical and laboratory markers of outcomes in the COVID-19 pandemic. Hypoxia and dyspnea are both signs of lung involvement by severe acute respiratory syndrome coronavirus 2. These results mirror the pathophysiological processes of viral pneumonia, which, in turn, could result in worse outcomes. These markers are readily available at the bedside and could enhance the feasibility of appropriate and quick triage of patients with COVID-19 to a higher level of care; resources are thereby conserved and preventive and management measures more expeditiously initiated. In addition, these markers could be used to potentially enroll appropriate patients in much-needed clinical trials to find the proper treatment of this deadly disease.

Although this article is timely and adds significant value to the current and growing literature on the topic, there remain some unresolved questions that should be addressed in future investigations. The assessment of oxygen saturation in the arterial blood (partial pressure of oxygen, arterial [Pao ₂]) when it is estimated by pulse oximetry should be carefully interpreted. Estimated oxygen saturation by CO-oximeters (Spo ₂) could be different from measured arterial oxygen saturation by about ±4%.⁹ Therefore, validating the results of Xie et al by using measured arterial oxygen saturation could be the next step. Furthermore, to be able to correctly assess the lung capacity for gas exchange, knowing the fraction of inspired oxygen (FIo ₂) is a necessity. Achieving this information may be challenging in some clinical scenarios. For example, the estimates of FIo ₂ when nasal cannulae or face masks are used may be variable (eg, when 2 L is delivered by a nasal cannula is used, FIo ₂ could vary between 24% and 35%, depending on the tidal volume patients demand).^{9 , 10} Therefore, in the next models for the prediction of mortality in patients with COVID-19, using the ratio of Spo ₂ or Pao ₂ and FIo ₂ may be necessary. It is also essential to assess the correlation between lung gas exchange capacity and mortality in light of different phenotypes of ARDS (ie, L vs H phenotypes). When ventilation to perfusion ratio mismatch drives hypoxia, the delivery of higher FIo ₂ leads to higher Pao ₂ and Spo ₂ (L phenotype). In contrast, when shunt drives hypoxia, the effect of FIo ₂ on the improvement in oxygen saturation would be less evident.^{5 , 6}

Xie et al also reported dyspnea as a significant predictor of mortality in patients with COVID-19. Dyspnea is defined as a subjective sense of breathlessness, and it is often mistaken for tachypnea, hyperpnea, or hyperventilation. Therefore, it is important in future studies to assess this symptom more objectively. For example, describing its acuity, its presence at rest or exertion or in different positions, and its precipitating or alleviating factors may facilitate triage of these patients into a more appropriate level of care.

As our knowledge and understanding of COVID-19 and its pathophysiology progressively increase, the article by Xie et al represents remarkable progress in the field. In particular, this study links the outcomes of COVID-19–associated pneumonia with straightforward clinical signs and symptoms, a linkage with a clear and plausible pathophysiological basis.