Authors' response: “Lung ultrasound to predict gas-exchange response to prone positioning in COVID-19 patients: A prospective study in pilot and confirmation cohorts”

To the editor

With great interest we read the letter to the editor by Cheema and colleagues concerning our recent submission: “Lung ultrasound to predict gas-exchange response to prone positioning in COVID-19 patients: A prospective study in pilot and confirmation cohorts” [1]. This letter is greatly appreciated as it offers an opportunity to clarify some of our study's key points and issues.

In this study we performed a quantitative lung ultrasound assessment prior to prone positioning of critically ill COVID-19 patients. We found, in separate derivation and validation cohorts, that the obtained lung ultrasound score correlates to gas-exchange response during prone positioning. This is an interesting finding and unveils a potentially important underlying principle.

The extent (and localization) of pulmonary involvement predicts the magnitude of response to prone positioning. Patients with more healthy pulmonary tissue, globally and in anterior fields, have a better gas-exchange response to prone positioning. Conversely, prone positioning fails to optimize gas-exchange in patients with more extreme pulmonary affliction. In addition to mechanistic insight, it also offers therapeutic considerations: can we tailor management depending on disease configuration [2]?

As addressed by the authors of this letter to the editor, there are several important limitations to be discussed.

First, current findings originate from a single centre prospective cohort study. Climbing the evidence pyramid requires more robust designs and external validation. As mentioned by the authors, future well-powered multicenter trials with longer follow-up are needed. Initiatives implementing such strategies are currently underway [3].

Second, the lung ultrasound score is an imperfect surrogate, but currently still the most widely validated protocol for quantitative assessment of pulmonary involvement. Although the i-LUS is an interesting alternative, it is mechanistically uninformative and may be quantitatively inadequate. The i-LUS incorporates qualitatively distinct pathologies for only 10 % of its score and places them on the same continuum without data-driven regard for their interrelation. External validation against a gold standard, for both COVID-SOFA and i-LUS, is required to estimate their relative relevance on suitable outcomes [4,5].

Third, we agree that the presence of (cardiovascular) comorbidities may contextualize and contribute to the clinical cost-benefit assessment of prone positioning. However, we speculate that these may be of limited relevance to the underlying physiological mechanism. To test this hypothesis we performed univariate logistic regression analyses for hypertension (p = 0.877) and diabetes mellitus (p = 0.897) with gas-exchange response at 24 h, but found no significant relation.

Last, and most important, is whether to define prone-response by its impact on gas-exchange. Albert and colleagues have cast doubts on its clinical relevance by showing a lack of association between gas-exchange improvement during prone positioning and survival [6]. Some argue that its effect on survival may be mediated through mitigation of ventilator induced lung injury [7]. Nonetheless, until conclusive evidence is obtained gas-exchange parameters remain important targets, offer mechanistic justification, and are clinically familiar.

To conclude, lung ultrasound-detected disease configuration can be used to tailor management, but true utility depends on its ability to modify patient outcomes.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest

None of the authors have any financial or personal relationship with other people or organizations that could inappropriately influence this work. Competing interests do not exist for any of the authors.