COVID-19 phenotypes: leading or misleading?

To the Editor:

We read the editorial by Bos et al. [1] with a mixture of interest, irritation and serious concern.

Our interest derives from a simple fact: the debate on terms like “typical acute respiratory distress syndrome (ARDS)” or “atypical ARDS” is not just a question of semantics; these terms represent concepts linked to specific clinical, mechanical and radiological criteria, and are not merely based on the severity of gas exchange. It should not be a surprise to the authors that different radiological patterns and mechanical characteristics should suggest different ventilatory strategies, each with possible benefits and harm. The management of individual patients needs to take into consideration various factors, and not just the gas exchange that currently defines ARDS. This is precisely the point of bringing attention to the novel “L” and “H” phenotypes of coronavirus disease 2019 (COVID-19) that bracket the extremes of the clinical encounter [2]. Usually, there is overlap, depending in large part on disease duration. The “L” and “H” were not intended to be tightly prescriptive nor mutually exclusive “bins” into which each patient falls, as we clearly stated previously [3]. Rather, the object was to alert clinicians, in order to avert potential harm from assuming usual ARDS associations between hypoxaemia and mechanics at all stages. In so doing, we hoped to help prevent use of high positive end-expiratory pressure when there is no benefit and, equally important, to avoid maintaining low pressures when higher pressures can be beneficial.

The irritation derives from the fact that Bos et al. [1] seem to have deliberately decided to ignore the pathophysiological “evidence” readily available and ventured into a philosophical and semantic discourse against “premature phenotyping”, and in so doing committing the greater sin of “premature adjudication”. After reading sentences such as “By needlessly clouding the clinical picture, false phenotypes […] upon inspection of patient data, simply do not exist”, it is not clear to us (and without a doubt to most readers) what sort of clear and self-evident truth we (and other authors) have been trying to cloud. The fact that COVID-19 patients with similar oxygenation efficiency may have markedly different compliance (and risk of ventilator-induced lung injury) is apparent to any clinician who has ever looked after a number of these patients. The reasoning put forward by the editorialists seems purely argumentative and inflammatory, as it seems to imply that what we propose is based on non-existent data, i.e. a perception that we invented.

Our concern derives from noting that the observations of Bos et al. [1] are expressed with a tone that goes beyond healthy and reasonable scientific debate. We note also with concern the conclusions of the editorial: “By prematurely phenotyping patients with COVID-19, we expose ourselves and our patients to considerable and preventable risk” and we invite the authors to express with clarity the risks they are referring to and how their argument is furthering the cause of patients and clinicians. Time and emerging literature will undoubtedly demonstrate where “truth” lies.

Shareable PDF