Pancreatic Stone Protein in the Early Detection of Sepsis-Induced ARDS: A Biomarker Panel Approach

Pancreatic Stone Protein (PSP) is a valuable early biomarker for sepsis. Given its kinetics and accessibility, PSP may aid in identifying patients at risk of progressing to ARDS, particularly within the hyperinflammatory phenotype.

Introduction

Acute respiratory distress syndrome (ARDS) is defined by hypoxaemic respiratory failure, bilateral pulmonary infiltrates, and non-cardiogenic pulmonary oedema. It affects millions worldwide and is associated with high mortality and long-term morbidity. Despite advances in supportive care—such as lung-protective ventilation and conservative fluid strategies—targeted therapies remain lacking. The Berlin ARDS definition (Ranieri et al. 2012), based on clinical and radiologic criteria, does not capture early or preclinical disease stages or biological heterogeneity. A recent comprehensive update (Matthay et al. 2024) aims to enhance early ARDS recognition and management across varied healthcare settings. It broadens diagnostic criteria to include non-intubated patients on high-flow nasal oxygen (HFNO) or noninvasive ventilation (NIV) and introduces SpO₂/FiO₂-based criteria for resource-limited environments—enabling diagnosis where arterial blood gases or advanced imaging are unavailable.

Biomarkers, as defined by the FDA-NIH Biomarker Working Group, are measurable indicators of biological states or conditions. Recognition of the biological heterogeneity in ARDS has highlighted the potential role of biomarkers in improving early diagnosis, risk stratification, and therapeutic targeting—especially in sepsis-induced cases. Ideal biomarkers should be biologically relevant, validated, easily measurable, and capable of informing clinical decisions. Such tools, particularly in sepsis-related ARDS, have the potential to optimise treatment and potentially improve survival. However, further prospective validation—especially using point-of-care platforms—is needed to determine whether biomarker-defined phenotypes should guide ARDS management (Matthay et al. 2024).

ARDS Biomarkers and Sepsis-Related Subphenotypes

Biomarkers offer critical insight into the pathophysiological processes of ARDS. They support disease stratification by underlying mechanisms, as illustrated in Table 1, and summarised here (Ge et al. 2023):

• Inflammatory markers such as IL-6, IL-8, and TNF-α rise early, particularly in hyperinflammatory subphenotypes, and correlate with severity and mortality. They are also being explored as therapeutic targets.
   
• Endothelial injury markers like Angiopoietin-2 (Ang-2), von Willebrand factor (vWF), and ICAM-1 reflect vascular leakage and endothelial activation, with elevated levels linked to higher ARDS risk and worse outcomes, especially in sepsis and post-surgical patients.
   
• Alveolar epithelial injury markers including sRAGE, SP-D, and CC16 indicate damage to type I and II alveolar cells. Their concentrations align with disease severity, ventilation needs, and mortality.
   
• Coagulation/fibrinolysis markers such as PAI-1, thrombomodulin, and D-dimer signal coagulation imbalance, with elevations associated with poor outcomes.
   
• Extracellular matrix turnover markers, notably MMP-9 and TIMP-1, relate to fibrosis and inflammation. Their imbalance contributes to lung remodelling and ARDS progression.
   
• Oxidative stress markers like malondialdehyde (MDA) reflect lipid peroxidation and alveolar damage, and may be modifiable through targeted interventions (e.g., ferrostatin-1, obacunone).

Multi-component biomarkers (MCBs)—combinations of two or more markers—usually offer superior diagnostic and prognostic accuracy over single biomarkers (Whitney et al. 2020). These may derive from multiplex platforms or integrative omics approaches (e.g., proteomics, transcriptomics, imaging). As ARDS is a biologically heterogeneous condition, MCBs—often supported by machine learning—are increasingly used to define subphenotypes, guide treatment, and identify novel therapeutic targets.

Among the most relevant are the hypoinflammatory and hyperinflammatory ARDS subphenotypes (Table 2). Identified through latent class analyses, they differ in immune response, clinical presentation, and outcomes. Sepsis more often triggers the hyperinflammatory phenotype, marked by elevated cytokines, endothelial injury, metabolic acidosis, and increased mortality. These patients are more responsive to high PEEP and corticosteroids. In contrast, hypoinflammatory ARDS features a blunted immune response and better prognosis. Maddali et al. (2022) validated these phenotypes using routine clinical variables from electronic health records (EHRs). Machine learning models accurately classified patients (AUC 0.88–0.92), with hyperinflammatory ARDS consistently linked to higher 90-day mortality and improved outcomes under specific interventions.

Pancreatic Stone Protein - A Promising Biomarker

Lopes et al. (2022) highlighted Pancreatic Stone Protein (PSP) as a valuable early biomarker for sepsis. PSP, a 16 kDa C-type lectin secreted by the pancreas and intestine, acts as a damage-associated molecular pattern (DAMP) during systemic inflammation. It has outperformed C-Reactive protein (CRP) and procalcitonin (PCT) in sensitivity and specificity for early sepsis detection across varied settings. PSP levels rise earlier than clinical signs or other biomarkers and correlate with SOFA scores and ICU mortality. Importantly, PSP can be measured via rapid point-of-care tests. Its performance remains robust even in complex cases, such as patients with inhalation injuries, where traditional markers are less discriminating Klein et al. (2021). Also Klein et al. (2021b) showed that PSP reliably differentiated between septic and non-septic burn patients—sometimes up to 72 hours before clinical deterioration—enabling earlier intervention and improved outcomes. Given its kinetics and accessibility, PSP may aid in identifying patients at risk of progressing to ARDS, particularly within the hyperinflammatory phenotype. However, further studies are needed to confirm its utility across broader populations and evaluate its impact on clinical decision-making in sepsis-associated ARDS.

Implications for Research and Practice

Monitoring biomarker trends over time can also indicate response to treatment or signal resolution of lung injury. The concept that ARDS is a syndrome with different biological subtypes detectable ideally with rapid subphenotype identification (Calfee et al. 2014; Famous et al. 2017; Spadaro et al. 2019), without the need of high-cost high complexity multiplexed molecular assays, may eventually drive new approaches. The study by Villar et al. (2021) indicates that a biomarker panel (RAGE, CXCL16, and Ang-2, combined with the clinical marker (PaO₂/FiO₂,) was good in predicting ARDS (AUC=0.88). ICU death was related to the presence of ARDS, the need for invasive mechanical ventilation, and pulmonary/extrapulmonary origin of sepsis. Such small biomarker panels integrated with clinical features demonstrate strong predictive value.

Currently, treatment strategies for sepsis-induced ARDS, primarily focus on supportive care such as oxygen delivery similar as for the other forms of ARDS (Matthay et al. 2019). To develop more effective targeted therapies, a deeper understanding of the molecular mechanisms linking sepsis to ARDS is essential. A key contributor to lung injury is vascular hyperpermeability during the exudative phase, driven by immune responses and damage to epithelial and endothelial cells. Inflammatory mediators such as IL-17 and transcription factors like NF-κB regulate the cytokine storm that characterises ARDS onset. Genetic polymorphisms in IL-17 (rs763780, rs2275913, rs8193036) and NF-κB1 (rs3774934) have been associated with ARDS risk and prognosis  (Hu et al. 2020), particularly in septic patients. Moreover, damage-associated molecular patterns (DAMPs), such as PSP, released by apoptotic cells can be detected in plasma and may serve as early diagnostic markers of lung injury. These findings highlight the potential of using circulating biomarkers to predict and prevent ARDS progression in septic patients.

A recent study by Sun et al. (2024) explored serum Myeloid-related proteins 8 and 14 (Mrp 8/14) as potential predictors of sepsis-induced ARDS. Mrp 8/14 expression rises in LPS-induced lung injury and is closely linked to sepsis-related lung injury in animal models, as well as COVID-19 pneumonia and ARDS. In the retrospective study with 168 septic patients, Mrp 8/14 levels were significantly higher in those who developed ARDS. These levels correlated with clinical severity scores (PCT, IL-6, APACHE II, SOFA) and ICU resource use. Logistic regression identified Mrp 8/14 as an independent predictor of ARDS, outperforming PCT and traditional scoring systems. While the findings highlight Mrp 8/14’s potential for early risk assessment, challenges remain—particularly the timing and dynamic nature of such biomarker expression.

Patient heterogeneity underscores the need for refined risk models. Yao et al. (2023) developed a predictive model integrating transcriptional biomarkers with clinical parameters to identify septic patients at risk of ARDS. The model included six factors (BPI, MMP8, MME, shock, tumour, direct lung injury) and demonstrated strong predictive accuracy (AUROC = 0.86, and 0.97 in patients without direct lung injury). These findings support the potential of transcriptomic data to improve early risk stratification and guide targeted interventions.

Standardising assays and thresholds across studies is essential for broader adoption. In early diagnosis, markers such as sRAGE—indicating alveolar epithelial injury—can rise before hypoxaemia or radiographic changes. In a meta-analysis of 746 patients across eight ARDS cohorts, Jabaudon et al. (2018, 2021) found that baseline plasma sRAGE was independently associated with 90-day mortality. This underscores the prognostic role of epithelial damage in ARDS outcomes. Similarly elevated serial surfactant protein-D (SP-D) levels, but not angiopoietin-2, showed a trend in non-survivors (Yang et al. 2022). Lin et al. (2023), in a systematic review and meta-analysis of 32 studies involving 2,654 ARDS/ALI patients, confirmed SP-D as a promising diagnostic and prognostic biomarker. These results support its utility in early ARDS detection.

Future Directions

The future of ARDS management will depend on validated multi-component biomarker panels reflecting different injury pathways. Rapid bedside testing, integration with clinical decision support tools, and artificial intelligence will help realise real-time risk stratification and personalised treatment. Incorporating biomarkers into trial design will improve patient selection for therapeutic targeting. This holds great promise in transforming ARDS care, from the early diagnosis to better phenotyping and to monitor targeted intervention. However, challenges remain in standardisation within larger populations and prospective outcome-driven clinical validation. Advancing biomarker-guided precision medicine in sepsis-related ARDS will require cross-disciplinary collaboration and well-designed studies.

Disclaimer

Point-of-view articles are the sole opinion of the author(s) and are part of the ICU Management & Practice Corporate Engagement or Educational Community Programme.