From syndrome to subtypes: transforming ARDS care with real‑time biomarker profiling

Main image credit: 

Acute Respiratory Distress Syndrome (ARDS) remains one of the most challenging conditions treated in intensive care, affecting critically ill patients each year and carrying a substantial risk of death despite advances in supportive care. One of the major obstacles to developing effective therapies has been the recognition that ARDS is not a single disease, but rather a syndrome comprising patients with different underlying biological processes and responses to treatment. Increasingly, researchers are turning to precision medicine approaches that seek to identify meaningful patient subgroups and tailor interventions accordingly. The recently published PHIND study represents a significant milestone in this field, providing the first prospective evidence that biologically distinct inflammatory subphenotypes of ARDS can be identified rapidly at the bedside and linked to markedly different clinical outcomes. 

The PHIND Study is the first large, multicentre, observational study with the aim of prospectively identifying inflammatory subphenotypes in a real-world clinical setting. Conducted across 30 intensive care units in the UK and Ireland, the study recruited adults with ARDS or acute hypoxaemic respiratory failure (AHRF) who required advanced respiratory support, including invasive mechanical ventilation, non-invasive ventilation or high-flow nasal oxygen. 

ARDS is a severe form of respiratory failure characterised by widespread inflammation in the lungs. Previous retrospective analyses had suggested that ARDS is not a single disease entity, but rather consists of biologically distinct subphenotypes. In particular, researchers have consistently identified a hyperinflammatory subtype, associated with more severe illness and poorer outcomes, and a hypoinflammatory subtype, associated with lower mortality. However, until now, identifying these subgroups in real time had not been possible, limiting their usefulness in clinical practice and research.

The PHIND study addressed this challenge by using a near-patient diagnostic platform capable of measuring key inflammatory biomarkers within approximately one hour. Blood samples were collected from patients within 72 hours of the onset of ARDS or AHRF. A 4 mL blood sample was processed at the bedside using the Randox MultiSTAT immunoanalyser and InflamiSTRAT assay panel. The system measured concentrations of interleukin-6 (IL-6) and soluble tumour necrosis factor receptor-1 (TNFR1), two biomarkers strongly associated with inflammatory activity.

These biomarker measurements were combined with arterial bicarbonate values obtained from routine blood gas analysis. The three variables were then entered into a validated logistic regression model that generated the probability of a patient belonging to the hyperinflammatory or hypoinflammatory subtype. Patients with a probability of 0.5 or greater were classified as hyperinflammatory, while the remainder were classified as hypoinflammatory. Importantly, this process enabled researchers to assign a subphenotype shortly after enrolment, demonstrating the feasibility of real-time biological classification in a critical care setting. 

Between November 2019 and September 2023, 1,853 patients were screened and 525 were recruited. Following withdrawals and a small number of technical assay failures, 490 patients were successfully subphenotyped. Of these, 89 patients (18%) were classified as hyperinflammatory and 401 patients (82%) as hypoinflammatory. 

The clinical characteristics of the two groups mirrored findings from earlier retrospective studies. Hyperinflammatory patients exhibited higher levels of circulating inflammatory biomarkers, greater metabolic acidosis, more frequent sepsis, increased need for vasopressor support and renal replacement therapy, and higher illness severity scores. Despite these systemic differences, measures of lung injury severity were broadly similar between the two groups, highlighting the importance of biological rather than purely clinical classification. 

The study’s primary outcome was all-cause mortality at 60 days. Results demonstrated a substantial difference between the two subphenotypes. Mortality in the hyperinflammatory group was 51%, compared with 28% in the hypoinflammatory group. After adjusting for factors such as age, sex and study site, patients in the hyperinflammatory group remained almost three times more likely to die within 60 days. 

Secondary outcomes reinforced these findings. Hyperinflammatory patients experienced significantly higher 28-day mortality, fewer ventilator-free days and longer durations of mechanical ventilation among survivors. They also required longer stays in intensive care and spent more time in hospital overall. Among patients initially receiving high-flow nasal oxygen, those classified as hyperinflammatory were substantially more likely to deteriorate and require intubation than hypoinflammatory patients. 

The investigators also assessed the accuracy of the near-patient testing platform by comparing its classifications with those obtained using established laboratory-based assays, including ELISA and the ELLA platform. Agreement between methods was excellent, supporting the reliability of the bedside approach while offering a much faster turnaround time than conventional laboratory testing. 

Overall, the PHIND study provides the first prospective evidence that inflammatory subphenotypes of ARDS can be identified rapidly and reliably in routine clinical practice. By demonstrating clear differences in mortality and other clinically important outcomes between hyperinflammatory and hypoinflammatory patients, the study strengthens the case for precision medicine approaches in critical care. The ability to identify biologically distinct patient groups at the bedside could help future clinical trials target therapies to those most likely to benefit, potentially accelerating the development of effective treatments for ARDS.