By Richard Kallet, MS, RRT, FCCM

Director of Quality Assurance, Respiratory Care Services, San Francisco General Hospital

Mr. Kallet reports no financial relationships relevant to this field of study.

SYNOPSIS: This retrospective observational study compared patients with acute respiratory distress syndrome classified as presenting with direct vs. indirect lung injury and found distinct differences in traditional predictors of hospital mortality between these subgroups.

SOURCE: Luo L, Shaver CM, Zhao Z, et al. Clinical predictors of hospital mortality differ between direct and indirect acute respiratory distress syndrome. Chest 2016 Sept 20. [Epub ahead of print].

Luo et al conducted a post-hoc analysis of 417 patients with acute respiratory distress syndrome (ARDS) categorized as presenting with either direct lung injury (pneumonia, aspiration) or indirect injury (sepsis, pancreatitis). Those with direct ARDS exhibited significantly higher lung injury scores, lower SAPS II and APACHE II scores, as well as fewer non-pulmonary organ failures compared to indirect ARDS. However, mortality was not different between the direct and indirect groups (28% vs. 31%, respectively; P = 0.49). Both intensive care and hospital length of stay were lower in those with direct ARDS. Ventilator-free days also was higher in direct vs. indirect ARDS, but this did not reach statistical significance (21 [2-24] vs. 15 [2-23] days, respectively; P = 0.054).

When the entire cohort was analyzed, factors such as age, lung injury score, and the number of organ failures were predictive of mortality, whereas the presence of diabetes was not. Both increasing age and lung injury score were only predictive of mortality in direct ARDS, and the presence of diabetes was protective in those with direct injury (odds ratio [OR], 0.47; 95% confidence interval [CI], 0.22-0.99; P = 0.04). In contrast, the only predictor of hospital mortality for indirect ARDS was the number of organ failures; the presence of diabetes was not protective (OR, 1.41; 95% CI, 0.64-3.09; P = 0.39).


Beginning in the late 1990s, studies suggesting that ARDS presents differently according to whether the primary site of injury was the alveolar epithelium (“direct,” “pulmonary”) vs. the vascular endothelium (“indirect,” “extrapulmonary”) focused on the effects of positive end-expiratory pressure (PEEP) and alveolar recruitment maneuvers to improve oxygenation and chest compliance. These studies suggested that direct pulmonary insults tend to result in more alveolar consolidation and less recruitable alveolar units. In contrast, indirect insults associated with trauma, sepsis, and pancreatitis primarily cause endothelial injury in the pulmonary and systemic vasculature, leading to global edema. The pulmonary significance of this is a pronounced reduction in chest wall compliance and increased congestive compressive atelectasis that is more amenable to PEEP and recruitment maneuvers.

A skeptical view of direct vs. indirect ARDS centers on two issues. First, alveolar tissue features an approximate width of 1 micron. It’s difficult to imagine that a substantial insult from caustic agents, such as gastric acid in aspiration or proteolytic enzymes in pancreatitis, would not seep across and injure other cell types in close proximity. Second, ARDS often presents as mixed etiologies (e.g., pneumonia leading to septic shock) rendering the distinction between direct and indirect injury essentially futile.

Some of the findings in the current study are important yet also unsurprising, such as the significant association of direct ARDS with age and mortality and indirect ARDS with organ failure and mortality. It’s apparent that the risk for pneumonia and aspiration and its association with mortality increases with age, particularly among the elderly, just as multiple organ failure, particularly in sepsis, also leads to higher mortality. In contrast, the positive effect of diabetes on mortality in direct (but not indirect) ARDS is a novel and interesting finding. Previous studies that investigated the interactions between diabetes and ARDS (without reference to direct vs. indirect injury) have produced contradictory results.

In their discussion, the authors bolstered their findings by citing several pre-clinical studies that found diabetes and hyperglycemia produce opposite effects on the magnitude of damage in direct (less damage) vs. indirect (greater damage) lung injury.

Other intriguing findings on direct vs. indirect ARDS come from recent studies by the ARDS Network.1 When considered together, these studies quell some of the skepticism regarding the importance of categorizing ARDS by injury vectors. Increased biomarker levels associated with alveolar epithelial injury (e.g., surfactant protein D, RAGE) were found in ARDS associated with pneumonia and aspiration. Likewise, indirect ARDS associated with sepsis demonstrated higher biomarker levels associated with alveolar endothelial injury (e.g., angiopoietin-2 and von Willebrand factor antigen). Direct ARDS compared to indirect ARDS was associated with lower proinflammatory mediator levels (IL-6 and IL-8), which corresponded with a strong trend toward lower mortality (29% vs. 35%, respectively; P = 0.06). These biomarkers previously have been associated with higher mortality in ARDS without reference to direct vs. indirect injury. However, investigators were unable to detect any interaction between ARDS subtypes and the prognostic value for any biomarker. Nonetheless, the two studies described above suggest the possibility that continued investigation of differences between direct and indirect ARDS, particularly in genetic propensities and biomarker expression, may lead to effective pharmacologic therapies targeting specific etiologies.


  1. Calfee CS, Janz DR, Bernard GR, et al. Distinct molecular phenotypes of direct vs. indirect ARDS in single-center and multicenter studies. Chest 2015;147:1539-1548.