Navigating the Extravascular Lung Waters of Pulmonary Edema in COVID-19 ARDS

Acute respiratory distress syndrome (ARDS) is marked by increased vascular and epithelial permeability, leading to alveolar flooding with protein-rich edema.(1) This hallmark pathophysiology drives the profound hypoxemia seen in ARDS. Several interventions in ARDS management, including conservative fluid balance strategies, titration of positive end-expiratory pressure (PEEP), or anti-inflammatory therapies, aim to modulate permeability-driven pulmonary edema. However, clinicians often rely on indirect metrics—such as oxygenation indices, lung compliance, or qualitative radiographic assessments—to infer edema severity and response to therapies. The development of reliable tools for tracking the evolution of lung edema in ARDS remains an unmet need.

In this issue of AJP Lung, Schippers et al. present a comprehensive analysis of pulmonary edema in COVID-19 ARDS, elucidating the interplay between extravascular lung water, vascular permeability, pulmonary physiology, chest imaging, and inflammatory plasma proteomics.(2)

The authors employed the pulse contour cardiac output (PiCCO) transpulmonary thermodilution technique to measure extravascular lung water index (EVLWi) and pulmonary vascular permeability index (PVPi) in 65 COVID-19 ARDS patients (Figure). This method involves injecting cold saline through a central venous catheter and detecting temperature changes with a thermistor-tipped arterial catheter(3). The resulting data yield several hemodynamic parameters, including EVLWi and PVPi. Normal EVLWi values are estimated to be <7ml/kg, with values exceeding 10ml/kg associated with pulmonary edema. An autopsy study reported that an estimated EVLWi of >14.6 ml/kg had a 99% positive predictive value for pulmonary edema.(4) PVPi, a dimensionless index, reflects the ratio of extravascular lung water to pulmonary blood volume and helps distinguish permeability from hydrostatic edema. PVPi values ≥3.0 have a sensitivity of 85% and specificity of 100% for diagnosing ARDS.(5)

Figure 1. Transpulmonary thermodilution (TPTD) measurements reveal distinct associations for extravascular lung water index (EVLWi) and pulmonary vascular permeability index (PVPi) in patients with COVID-19 acute respiratory distress syndrome (ARDS).

Left panel: TPTD is a bedside technique used to assess pulmonary edema. A known volume of cold saline is injected via a central venous catheter, and the resulting change in blood temperature is measured by a thermistor-tipped arterial catheter. From this, two key indices are calculated: EVLWi, reflecting fluid accumulation outside the pulmonary capillaries, and PVPi, which quantifies capillary leak. Right panel: In the study by Shippers et al.(2) patients with COVID-19 ARDS, elevated EVLWi is associated with worse lung compliance, decreased oxygenation, and increased chest CT ground glass opacities. PVPi demonstrated stronger correlations with inflammatory plasma proteins compared to EVLWi. Dashed panel: The distribution of EVLWi (blue) and PVPi values (red) in COVID-19 ARDS patients is shown, with the median and interquartile range (IQR) indicated. Proposed cut-offs to differentiate between normal lung function, hydrostatic edema, and increased permeability edema are overlaid (graph adapted from (3)). Figure created with BioRender.

Schippers et al. study revealed a broad range of EVLWi and PVPi values, with partially overlapping clinical associations. Higher EVLWi correlated with worse lung function, lower PaO₂/FiO₂ ratios, and reduced compliance. Patients in the highest EVLWi tertile (>17.1ml/kg) exhibited stiff lungs and significant gas exchange impairment, suggesting alveolar flooding and shunting. Elevated central venous pressure in these patients indicated that hydrostatic forces might contribute to edema formation. EVLWi also correlated with ground-glass opacities (GGO) on CT scans, underscoring the value of imaging in characterizing pulmonary edema. Interestingly, the radiographic assessment of lung edema (RALE) score derived from chest X-rays showed no such correlation, highlighting limitations in traditional imaging interpretation. In contrast, PVPi showed stronger associations with plasma proteins, particularly cytokines, chemokines, and extracellular matrix turnover markers (Figure).

These observations have important clinical implications. While transpulmonary thermodilution provides detailed insights, its invasive nature limits its use to selected patients. CT imaging, with its ability to discern GGOs and other densities, appears better suited for capturing EVLWi.(6) However, logistical challenges and the risks transporting critically ill patients to radiology suites highlight the need for optimizing bedside tools. Alternative approaches, such as lung ultrasound (7) or improving chest X-ray interpretation through use of AI algorithms, show promise. Preliminary data show that deep learning models can accurately detect and quantify pulmonary edema using chest radiography.(8) The findings by Shippers et al. also highlight the complex interplay of systemic inflammation, alveolar-capillary membrane damage, and hydrostatic factors in pulmonary edema. While PVPi reflects inflammation-driven permeability changes, EVLWi captures broader physiologic dysfunction, aligning more closely with radiographic abnormalities.

The plasma proteomic analysis offers valuable insights into pulmonary edema in COVID-19 ARDS. Acute-phase inflammatory pathways were strongly correlated to both pulmonary edema and vascular permeability, with PVPi showing greater protein enrichment and fold change compared to EVLWi, suggesting that the plasma proteome more accurately reflects vascular permeability. However, the directionality of this relationship remains unclear, as systemic inflammation may exacerbate vascular permeability, or increased permeability could facilitate the spillover of pro-inflammatory mediators from injured lungs. Surfactant protein D was elevated in the highest EVLWi tertile, potentially due to pneumocyte damage impairing fluid reabsorption. Additionally, hepatocyte growth factor activator inhibitor-1 (HAI-1) strongly correlated with PVPi, indicating its potential role in vascular permeability. HAI-1 modulates hepatocyte growth factor (HGF) activity, which has been associated with poor COVID-19 outcomes.(9) The interplay between HAI-1, HGF, and vascular permeability warrants further investigation in non-COVID ARDS to elucidate its broader implications and therapeutic potential.

Despite the detailed phenotyping in this study, several limitations should be acknowledged. Some of the findings may not be generalizable to non-COVID ARDS patients. The small sample size and cross-sectional design restrict the ability to draw causal infrerences, particularly regarding the directionality of the relationship between systemic inflammation and PVPi. The exclusive reliance on plasma samples without bronchoalveolar lavage fluid analysis or other type of lower respiratory tract sampling constrains the understanding of localized lung inflammation and the ability to directly profile the alveolar edema fluid.(10) Finally, this study did not evaluate the effects of specific therapies on EVLWi or PVPi.

Future research should prioritize the development of accessible radiologic tools for early detection and monitoring of pulmonary edema in ARDS, alongside rapid laboratory assays for biological phenotyping to enable personalized treatment approaches. Transpulmonary thermodilution analysis could play a role in specific patient populations with severe edema, particularly when central venous and arterial access is already established. In conclusion, Schippers et al. offer a comprehensive analysis of pulmonary edema in COVID-19 ARDS, integrating advanced physiological measurements, imaging, and proteomic data. Their findings illuminate the complex pathophysiology of ARDS and demonstrate the potential of multidimensional phenotyping to guide targeted interventions. This integrative approach paves the way for more nuanced, patient-specific strategies in managing the heterogeneous syndrome of ARDS.