On Happy Hypoxia and on Sadly Ignored “Acute Vascular Distress Syndrome” in Patients with COVID-19

To the Editor:

We read with great interest the article by Tobin and colleagues (1) on the issue of silent hypoxemia, which is also known as happy hypoxia, and found it to be a nice review of physiologic mechanisms of dyspnea. The authors refer to the definitions and mechanisms of dyspnea in relation to blood gases, pulmonary insults, age, and disease. They also discuss the definitions and effects of hypoxia, the inaccuracies of pulse saturation, and the properties of the oxygen dissociation curve as well as the mechanisms of hypoxemia in patients with coronavirus disease (COVID-19). We agree that all the physiologic concepts recalled by Tobin and colleagues might, in isolation or together, contribute to a blunted ventilatory response to low levels of Pa_O2 and to its corollary subjective feeling of normality or the absence of dyspnea. Among these various factors, we do not believe that the poor correlation between oxygen saturation and arterial partial pressure at low levels of saturation can explain happy hypoxia because, as shown in the vignettes of their paper, patients have not only low oxygen saturation as measured by pulse oxymetry (Sp_O2) values but also very low levels of Pa_O2 (which, according to Tobin and colleagues’ Figure 1, should have led to ventilation levels well above 20 L/min), yet they consistently denied any difficulty with breathing. Similarly, although age and diabetes have a known blunting effect on the ventilatory response to hypoxia, many patients with happy hypoxia are in their 50s or 60s, wherein age effects are not expected to be great, and are not diabetic. Similarly, we would add that if dyspnea is subjective, $\dot{\mathrm{V}}$e levels of more than 20 L/min require obvious use of accessory muscles and visible increases in respiratory frequency that patients with happy hypoxia do not show.

We would like to advance that the main reason for the phenomenon of happy hypoxia is the presence of hypocapnia. We have shown several years ago that hypocapnia has such a powerful braking effect on the respiratory center that it can completely abolish any response to repeated exposure to very low Sp_O2 levels in normal subjects (2). We see no reasons why happy hypoxia should be limited, as Tobin and colleagues claim, to patients without hypocapnia. By the way, hypocapnia and its consequent alkalosis would tend to shift the oxygen dissociation curve to the left, counteracting the rightward shift due to fever.

As to the reasons for hypocapnic hypoxia without dyspnea, there is one that Tobin and colleagues do not mention and that we believe offers the best explanation, as follows: the presence of a right-to-left intrapulmonary shunt (3). Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is known to induce vascular proliferation in the lungs demonstrated both in anatomic and radiologic studies (4, 5). We have demonstrated a late right-to-left intrapulmonary shunt by contrast enhanced echocardiography in one patient with COVID-19 without radiologic lung lesions (unpublished observation). This right-to-left shunt will induce hypoxia, leading to a normal increase in ventilation. However, in face of a shunt, hyperventilation will not increase Pa_O2 but will certainly decrease Pa_CO2, with CO₂ being more diffusible than O₂. Thus, hypocapnia would develop, abolishing any further increase in ventilation and explaining the absence of enhanced respiratory efforts and, therefore, of dyspnea. This, we contend, is the initial insult of SARS-CoV-2 that has prompted us to coin the acronym “AVDS” for acute vascular distress syndrome (6). When lung lesions become prominent, showing either ground-glass opacities or consolidations, hypoxia could worsen but hypocapnia would lessen, with the consequent normalization of Pa_CO2 and the appearance of feelings of difficult breathing.

In conclusion, we believe it is now time to consider the intrapulmonary shunt as the key factor in patients with COVID-19 that accounts for both the presence of hypoxia and the absence of dyspnea in many of them.