Phoenix sepsis criteria fail to detect 95% of hospitalised children

sepsis, 25 July 25/ in E-News, Research

A large-scale prospective study across Australia and New Zealand has revealed that the internationally recognised Phoenix sepsis criteria identify fewer than 5% of children hospitalised with suspected sepsis, raising significant concerns about diagnostic accuracy and patient safety in paediatric emergency medicine.

The research, conducted by the Murdoch Children’s Research Institute (MCRI) and published in The Lancet Regional Health – Western Pacific on 22 July 2025, examined 6,232 children with suspected sepsis across 11 hospitals, finding that only 306 met the Phoenix sepsis criteria – a detection rate that has prompted calls for more comprehensive diagnostic approaches.

Critical diagnostic gaps emerge in multi-centre analysis

The study, part of the Paediatric Research in Emergency Departments International Collaborative (PREDICT) network, screened 822,072 children over three years between April 2021 and December 2023. Of the 6,232 children identified with suspected sepsis, representing 0.8% of all emergency department presentations, merely 306 children (<0.1%) fulfilled the Phoenix sepsis criteria.

“The Phoenix criteria didn’t pick up 95 per cent of children hospitalised with sepsis,” said Associate Professor Elliot Long, lead researcher from MCRI. “This high rate of missed cases is troubling for clinicians, researchers, policymakers and families because it does not tell the whole story.”

The Phoenix sepsis score, developed by the Society of Critical Care Medicine, utilises a four-organ dysfunction model combining cardiovascular, neurological, respiratory, and coagulation system derangements. A score of ≥2 within the first 24 hours of hospitalisation indicates sepsis under these criteria.

Mortality outcomes highlight clinical significance

Perhaps most concerning, the study revealed that more than half of the children who died within 90 days did not meet the Phoenix criteria. Of the 87 patients who died, only 42 fulfilled the Phoenix sepsis criteria, representing a significant diagnostic blind spot for this life-threatening condition.

Children meeting Phoenix criteria demonstrated substantially worse outcomes, with 80.1% requiring intensive care unit admission compared to 17.3% in the overall cohort. These patients also required significantly more intensive interventions, including vasoactive infusion (47.1% vs 2.9%) and mechanical ventilation (47.7% vs 4.0%).

“Children who fulfilled Phoenix criteria had very high rates of vasoactive infusions and mechanical ventilation compared to the overall cohort,” the authors noted in their analysis. The mortality rate among Phoenix-positive children was 13.7% within 90 days, compared to 1.4% in the overall suspected sepsis population.

Healthcare burden underestimated by current criteria

The research highlights a fundamental disconnect between clinical reality and diagnostic criteria. While the Phoenix score successfully identified the most critically ill children – those requiring intensive organ support – it failed to capture the broader spectrum of sepsis-related illness requiring hospital care.

“By failing to identify these children much earlier, before being admitted to intensive care, causes delays to treatment and impacts their recovery,” Associate Professor Long explained. The study found that intensive care and hospital length of stay were significantly longer for Phoenix-positive patients (median 79.8 hours vs 48.4 hours in ICU, and 189.8 hours vs 69.7 hours in hospital).

Global implications for paediatric sepsis management

The findings carry international significance, as sepsis affects approximately 25 million children globally each year, resulting in 3 million deaths. The World Health Organization has identified paediatric sepsis as a global health priority, making accurate diagnostic tools essential for effective treatment and resource allocation.

Professor Franz Babl, senior author from MCRI, emphasised the clinical urgency: “We haven’t substantially changed how we manage childhood sepsis in the past 20 years. There are enormous knowledge gaps in how to best treat sepsis. Some treatments can even cause additional damage to patients, rather than helping them.”

Research drives innovation in sepsis treatment

The study’s implications extend beyond diagnostic concerns. Associate Professor Long has been awarded a AUS$5 million Medical Research Future Fund grant to test potential sepsis treatments through an innovative adaptive platform trial across Australia and New Zealand. This approach aims to remove traditional research obstacles by testing multiple treatment options simultaneously.

The research methodology involved children aged 0-18 years admitted to hospital for parenteral antibiotics with either a provisional sepsis diagnosis or treatment involving fluid boluses for impaired perfusion. This pragmatic approach captured real-world clinical decision-making rather than relying solely on diagnostic codes.

Improving sepsis care pathways

The authors conclude that while the Phoenix criteria provide valuable standardisation for reporting sepsis epidemiology, “the healthcare and societal burden of sepsis is much greater than might be inferred when the Phoenix criteria alone are used for case ascertainment.”

These baseline epidemiological data will prove critical for improving sepsis care pathways and designing future interventional studies in community-acquired sepsis. The research underscores the urgent need for more sensitive diagnostic tools that can identify the full spectrum of sepsis-related illness in children, with the aim of saving lives through earlier intervention and treatment.

Reference

Long, E., Borland, M. L., George, S., et. al. (2025). Epidemiology of community acquired sepsis in children in Australia and New Zealand: a multicentre prospective cohort study. The Lancet Regional Health – Western Pacific, 101608. https://doi.org/10.1016/j.lanwpc.2025.101608