Physiology-Based Approach to PEEP Titration in COVID-19 ARDS

The heterogeneous and dynamic clinical behaviors of patients with ARDS in response to PEEP continue to challenge all known methods to optimize this variable, raising important conceptual and research questions regarding how to accurately estimate the “best PEEP.” The immense number of ARDS patients treated during the COVID-19 pandemic put the spotlight on the previously documented, yet frequently unrecognized, physiologic diversity of the condition we label as ARDS.¹ The oxygenation-based definition and severity assignment of ARDS encourage the use of oxygenating efficiency as a convenient target for PEEP titration. But is this a wise practice when lung protection is also at issue? The clinical and physiologic implications of such heterogeneity are nicely illustrated by the data presented by Scaramuzzo et al in this issue of the Journal.² The authors evaluated the effect of 3 different PEEP levels on ventilation/perfusion (V̇/Q̇) mismatch, shunt, and alveolar dead space among subjects with ARDS due to COVID-19 with 2 different phenotypes of respiratory mechanics. Subjects were classified as relatively low compliance/high elastance (phenotype H) or as relatively high compliance/low elastance (phenotype L).^2,3 Bedside measurements of gas exchange and alveolar ventilation were entered into a mathematical model that estimated pulmonary shunt and states of low or high V̇/Q̇ matching in response to mean PEEP values ranging between ∼5–15 cm H₂O.

In the overall population, the salient findings included: (1) shunt, V̇/Q̇ mismatch, and alveolar dead space were not significantly influenced by PEEP changes; and (2) lung mechanics were impaired by raising PEEP, as evidenced by decreases of respiratory system compliance (C_RS) without significant P_aO2/F_IO2 variation. In fact, raising PEEP did not improve C_RS in any study subject. Notably, for the 2 defined ARDS phenotypes, PEEP produced opposing effects on shunt, with a decrease in the phenotype L and an increase in phenotype H. No other differences in V̇/Q̇ mismatch were observed in response to PEEP changes. These findings indicate a low potential for functional recruitment of lung tissue in the overall sample of subjects.⁴ Significant differences may exist between the method used here for recruitment estimation (improved respiratory mechanics)⁵ and more precise assessments of recruitment as quantified by imaging (eg, gasless tissue regaining aeration on analysis by computed tomography).⁶ A clinical manifestation of this ambiguity is illustrated by the decrease in shunt observed in the phenotype L subgroup, in which it is unlikely that lung units already opened experienced significant improvements in V̇/Q̇ mismatch or P_aO2/F_IO2. Considering the known effect of raising PEEP on mean airway pressure and the corresponding increases in arterial P_aO2 expected from that intervention,⁷ the absence of V̇/Q̇ mismatch and P_aO2/F_IO2 variations with augmented PEEP raises concern regarding the accuracy of the automatic lung parameter estimator (ALPE) and the use of noninvasive estimations of cardiac output instead of direct measurements. Another intriguing possibility to explain the discordance between expected and observed gas exchange responses to PEEP may relate to the abnormalities of vascular dysregulation and extensive shunting through the bronchial circulation that distinctively characterize COVID-19 ARDS.^8-9

Since PEEP had opposite effects on shunt, with a decrease in the phenotype L (higher compliance) and an increase in phenotype H, the authors concluded that C_RS could influence the effect of PEEP on V̇/Q̇ mismatch. We point out, however, that although recruitment of functional lung units usually is the key factor explaining improvement of P_aO2/F_IO2 other well-recognized mechanisms such as PEEP-related declines of cardiac output may strongly contribute.¹⁰ Moreover, mean lung compliance in ARDS populations, though usually low, ranges widely. Some patients who have relatively mild compliance reductions may experience large contributions to improved hypoxemia from PEEP via mechanisms other than lung recruitment.^4,10 When managing ARDS, COVID-19 associated or not, recognizing the existence of such clinical diversity and physiologic heterogeneity should help the clinician to consider all major factors contributing to addressable cardiopulmonary dysfunction and thereby formulate a personalized approach to mechanical ventilation that protects both the lung and the heart.

Ideally, a best PEEP simultaneously (1) achieves appropriate gas exchange, (2) prevents tidally phasic airway collapse, (3) avoids alveolar overdistention, (4) does not compromise hemodynamics, and (5) accomplishes these while minimizing its contribution to the inflation energy load per breathing cycle (lowering the risk for ergotrauma¹¹). In this context, what does personalization of the best PEEP actually mean at the beside? In reality, any PEEP selection implies a compromise among these primary objectives, balancing the time-dependent lung mechanics of ARDS with genuine gas exchange needs and the prevention of complications associated with PEEP titration. Given the inherent unreliability of an oxygenation-based PEEP strategy, we agree with Scaramuzzo et al² that a better physiologic approach is required. Available data¹² call for a systematic reassessment of pulmonary mechanics as the clinical condition of the patient improves or deteriorates. The components of this approach might be as follow. (1) Assessing lung recruitability. Current practice favors setting PEEP that limits airway driving pressure and plateau pressure,¹² utilizing indicators such as dual before and after PEEP increment pressure-volume curves¹³ or response to a PEEP change with one of an array of tools to assess mechanics. These include the ratio of estimated recruited volume to the total volume increment,¹⁴ computed tomography analysis of aeration,⁴ bedside lung ultrasound,¹⁵ and electrical impedance tomography.¹⁶ (2) Lung mechanics and PEEP titration target. In some clinical scenarios, stable recruitment has been assumed when tidal compliance of the respiratory system increases in response to a PEEP increment.⁵ Setting PEEP to optimize C_RS can overcome some (but not all) pitfalls associated with the use of oxygenation targets. Shortcomings of an oxygenation-based approach are especially evident among patients with ARDS having relatively preserved C_RS, those in the organizing or fibrotic phase of inflammation, and those—such as COVID-19 ARDS—with inherently atypical pulmonary vascular pathology.^3,8,9 In these subpopulations, targeting hypoxemia may lead to the use of excessively high values of PEEP with negative consequences for hemodynamics. When focusing on mechanics, the clinician’s goal should be achieving the highest possible compliance with the lowest possible PEEP rather than attempting to achieve a specific C_RS target.¹² C_RS will vary from patient to patient, and within the same patient the compliance of aeratable lung units is influenced by stage of disease, ventilator settings, chest wall stiffness, and body positioning.^17-18 (3) Assessing the effect of PEEP on hemodynamics. Although the effects of positive-pressure ventilation on right-ventricular preload (reduction in venous return) and afterload (increased in pulmonary vascular resistance at higher lung volumes) are generally understood,¹⁹ the hemodynamic consequences of PEEP are still frequently neglected at the bedside. When compromise is recognized, it is routinely addressed with additional intravenous fluids or vasoactive medications instead of an optimized ventilatory approach. Since right-ventricular dysfunction develops over time under the influence of deleterious ventilatory strategies, the early stages of such myocardial compromise can present in a subclinical fashion,²⁰ raising important research questions regarding how to identify at the beside an optimal PEEP that includes right-ventricular function, for example, with the conjoint use of echocardiographic tools such as speckle-tracking echocardiography.²¹ (4) Adapting lung-protective mechanical ventilation strategies to clinical progression. Apart from the natural progression of the lung’s inflammatory process toward resolution or fibrosis, any interventions such as diuresis, airway clearance, body repositioning, or treatment of the underlying cause of ARDS can change the clinical and physiologic profile of the patient and consequently the patient’s response to PEEP. Therefore, for any given PEEP value, periodic reassessment of its need and adequacy constitutes an important quality measure for any mechanical ventilation strategy.¹²

The small cohort of subjects studied by Scaramuzzo et al² may not be representative of the wide clinical spectrum that characterizes ARDS. In fact, subjects were enrolled within a median of 5 d of intubation, and no data are offered regarding the potential time spent on noninvasive ventilatory support prior to intubation. Moreover, the accuracy and potential applicability of ALPE as a feasible bedside tool require confirmation. Nevertheless, their helpful study underscores the necessity for future ARDS management strategies that optimize PEEP based on integrating the diverse elements of changing pathophysiology that track hemodynamics and energetic cost, as well as the gas exchange and mechanics indicators already in use.