Editorial 

January 21, 2024

Context and Implications of the New Pediatric Sepsis Criteria

Erin F. Carlton, Mallory A. Perry-Eaddy, Hallie C. Prescott

JAMA. Published online January 21, 2024. doi:10.1001/jama.2023.27979

Sepsis has been recognized as an important cause of morbidity and mortality for more than 2800 years.¹ Derived from the Greek word sepo (σηπω, translated as “I rot,”), sepsis appears in writings of Homer, Hippocrates, Aristotle, and Galen.^1,2 Conceptually, sepsis denotes a “line in the sand” within the broad spectrum of host-microbe interactions. It indicates that not only is the microbe unwanted (in contrast to the many microbes that live on and within us and contribute to health and digestion),³ but the host’s attempts to eradicate the microbe are resulting in collateral damage.

Despite existing in the medical lexicon for millennia, sepsis was not formally defined until the 1992 American College of Chest Physicians and Society of Critical Care Medicine Consensus Conference.⁴ Now referred to as Sepsis-1, the sepsis definition was viewed by the consensus conference as an overabundant inflammatory response to infection, a systemic inflammatory response syndrome (SIRS), requiring 2 or more abnormalities of temperature, heart rate, respiratory rate, or white blood cell count. An updated definition in the early 2000s (Sepsis-2)⁵ expanded on the potential signs of an overabundant inflammatory response but was otherwise similar to Sepsis-1.

Pediatric-specific sepsis criteria were subsequently developed by an expert panel during the International Pediatric Sepsis Consensus Conference (IPSCC) and published in 2005.⁶ Similar to adult definitions (Sepsis-1/2),^4,5pediatric sepsis was defined by 2 or more SIRS criteria in the setting of confirmed or suspected infection, with severe sepsis denoting sepsis complicated by organ failure, and septic shock indicating sepsis with severe cardiovascular dysfunction. Acknowledging differences in pediatric physiology, the IPSCC definitions diverged from adult criteria to require at least 1 SIRS criteria be abnormal temperature or white blood cell count and to use age-specific SIRS criteria.

The adult sepsis definitions underwent a third update in 2016 (Sepsis-3), which departed from the SIRS-based definition.⁷ Sepsis was reconceptualized as life-threatening acute organ dysfunction secondary to a dysregulated host response to infection. The prior notion of sepsis as an overabundant inflammatory response was deemed too narrow—a recognition informed, in part, by dozens of negative trials of anti-inflammatory agents to treat sepsis.⁸At the same time, SIRS criteria were removed from the definition because they may merely reflect a noninjurious host response to infection.

In contrast to prior definitions based on expert consensus, Sepsis-3 used a data-driven approach to select the criteria for organ dysfunction. After analyzing data from nearly 1 million patients treated in 177 hospitals, the panel recommended using an increase in the Sequential (sepsis-related) Organ Failure Assessment (SOFA) score of at least 2 points to define life-threatening acute organ dysfunction. This recommendation was based on both predictive validity and feasibility of implementing SOFA scoring.

The Process for Developing the Pediatric Sepsis Criteria

Although Sepsis-3 criteria were not proposed for use in children, there has been increasing recognition that—in children as well as in adults—sepsis is more appropriately viewed as infection-triggered acute organ dysfunction.⁹ Therefore, in 2019, the Society of Critical Care Medicine appointed an international, multiprofessional task force to update the pediatric sepsis criteria. The Phoenix sepsis criteria,⁹ presented in this issue of JAMA,¹⁰ were selected through a modified Delphi consensus process and informed by several task force–led studies (Figure).^9,11,12

Figure.  Process to Develop the Phoenix Pediatric Sepsis Criteria

This figure displays the systematic program of inquiry used to develop the Phoenix sepsis criteria. The task force completed an international survey,⁹ a systematic review and meta-analysis,¹¹ and a cohort study,¹² which together informed the consensus selection of the Phoenix Sepsis Score.¹⁰

First, the task force fielded an international survey to determine how clinicians diagnose sepsis in practice.⁹There were 2835 respondents from 6 global regions, including 14% from lower-income settings. None of the existing definitions (IPSCC, Sepsis-3, World Health Organization [WHO]) were perceived as useful by the majority of respondents across all 6 domains of use (recognition, early recognition, disease classification, prognostication, benchmarking, epidemiology, and trial enrollment). Furthermore, 71% of respondents believed the term sepsisshould be limited to children with infection-related organ dysfunction. Next, the task force completed a systematic review and meta-analysis to evaluate associations between patient and clinical features with (1) development of sepsis among children with infection and (2) mortality among children with sepsis.¹¹ Review of 16 studies (9629 children) evaluating sepsis and 71 studies (154 674 children) evaluating outcomes confirmed that organ dysfunction is strongly associated with both sepsis and mortality. Overall, these studies supported a transition from SIRS-based to organ dysfunction–based criteria for pediatric sepsis.

The task force next completed a cohort study to identify and validate organ dysfunction–based pediatric sepsis criteria.¹² To support this task, they assembled a granular database from 6 US and 4 international sites, located in Bangladesh, China, Colombia, and Kenya. Nearly 175 000 children in the derivation cohort and 50 000 children in the validation cohort with confirmed or suspected infection were included. Data from all care settings in the 24 hours after presentation were used, including emergency department, inpatient, and intensive care units.

The overarching goal of the cohort study was to determine which organ dysfunction criteria best predict mortality among children with proven or suspected infection. First, the best criteria for predicting in-hospital mortality were identified for each of 8 individual organ systems (cardiovascular, coagulation, endocrine, hepatic, immunologic, neurological, kidney, respiratory), drawn from 5 existing scoring systems: IPSCC, Pediatric Logistic Organ Dysfunction-2 (PELOD-2), Pediatric Organ Dysfunction Information Update Mandate (PODIUM), pediatric SOFA, and Proulx. Next, these best-performing criteria for each individual organ system were considered for inclusion in an overall model predicting in-hospital mortality. A 4-system model (including cardiovascular, coagulation, neurological, and respiratory systems) achieved similar discrimination as the full 8-system model. The task force selected the 4-system model due to its simplicity and translated this model into an integer score (Phoenix Sepsis Score) to facilitate implementation into practice.

The Definitions

The Phoenix sepsis criteria defines sepsis as life-threatening organ dysfunction of the respiratory, cardiovascular, coagulation, and/or neurological systems, demonstrated by a Phoenix Sepsis Score of at least 2, in the setting of confirmed or suspected infection (Table).¹⁰ Septic shock was defined as sepsis with at least 1 point in the cardiovascular category (blood lactate ≥5 mmol/L [≥45.05 mg/dL], hypotension for age, or vasoactive use). These thresholds were selected based on group consensus, requiring more than 80% agreement among more than 80% of the task force.

Table.  Comparison of Phoenix Pediatric Sepsis Criteria With International Pediatric Sepsis Consensus Conference Criteria

In the validation study, children meeting the Phoenix criteria for sepsis in higher- and lower-resourced settings had 7% and 29% in-hospital mortality, respectively, or 8-fold higher than all children with infection. Children meeting Phoenix criteria for septic shock had 11% and 34% mortality, respectively. Overall, the Phoenix criteria out performed all other criteria, including IPSCC, for predicting in-hospital mortality.

The Meaning

The Phoenix sepsis criteria were created to improve clinical care, research, and benchmarking. They were informed by a systematic program of inquiry that intentionally incorporated diverse stakeholders and perspectives, from panel composition to inclusions in the survey, meta-analysis, and cohort studies. The broad applicability of sepsis definitions is important, particularly given the disproportionate burden of sepsis and sepsis-related mortality experienced in lower resourced settings¹³—a point highlighted by the 3-fold difference in mortality between US and international sites in the validation study. Continued efforts to decrease this large gap are paramount.

With the Phoenix criteria, pediatric sepsis is redefined as life-threatening organ dysfunction in the setting of infection. The organ dysfunction criteria selected are both data driven and pragmatic. As the field incorporates these criteria, there are several points to consider. First, the SIRS criteria are not included in the Phoenix definition but have value in clinical care particularly for assessing the presence of infection. The Phoenix criteria were developed and validated among children with proven or suspected infection. However, recognizing and confirming infection remains a challenge, with up to a third of patients diagnosed with sepsis having a noninfectious illness in hindsight.¹⁴ The SIRS criteria remain useful for assessing the presence of infection.⁷ Second, while Sepsis-3 requires at least 2 new SOFA points, Phoenix scoring does not specify that organ dysfunction must be new. Chronic organ dysfunction may indicate increased risk of mortality but be less responsive to acute treatments for sepsis. Enrollment into clinical trials should thus be mindful of acute vs chronic organ dysfunction. Third, kidney dysfunction (which is included in the IPSCC definition of severe sepsis and associated with mortality in prior studies of pediatric sepsis^15,16) was not included in the Phoenix criteria. Research studies should consider using the full 8-system model to further assess the implications of restricting to 4-organ systems.

Fourth, the cohort study considered only data from the first 24 hours of presentation and had high rates of missingness at some sites. Additional validation, particularly for hospital-onset sepsis, is warranted. Fifth, like Sepsis-3, the Phoenix criteria identify a sicker subset of patients than prior SIRS-based criteria. Some may worry this higher threshold could delay management of patients not meeting sepsis criteria. Just as patients with chest pain and a troponin leak warrant monitoring and treatment (but are not prioritized for immediate heart catheterization)—patients with infection need monitoring and treatment. Improvements in care should thus be judged not only by improved outcomes among patients with sepsis but also by decreased progression to sepsis among patients with infection.

The Phoenix sepsis criteria identify children with life-threatening organ dysfunction in the setting of infection. This sepsis definition is supported by a robust body of research, inclusive of diverse geographic and resource settings. These new definitions and shared conceptual understanding of pediatric sepsis will support improvements in the management, research, and outcomes of children with sepsis worldwide.