Pediatric Sepsis: Subphenotypes to Enrich Clinical Trial Entry Criteria.

To the Editor:

We read with great interest the article by Carcillo et al (1) published in a recent issue of Pediatric Critical Care Medicine. Studies with biomarkers in adults with sepsis and acute respiratory distress syndrome (ARDS) have shown that identification of specific subphenotypes could lead to a better identification of patients that could be more responsive to interventions. In ARDS, a combination of biomarkers and clinical data improved the understanding of the patient profiles and may influence entry criteria of clinical trials. Recent trials support that the presence of ARDS subphenotypes may demand distinct treatment approaches, regarding, for example, fluid management or other specific therapies (2). In Statins for Acutely Injured Lungs for Sepsis study (3), two subphenotypes (hyper-inflammatory subphenotype or not) were tested for distinct treatment response to statin. Although the use of statin did not show treatment effect, hyper-inflammatory” subphenotype patients had higher mortality and validated subphenotypes previously described and their usefulness for patients’ stratification. More recently, Seymour et al (4) analyzed more than 20,000 adult patients derivating different sepsis subphenotypes, and the results suggested that these subtypes may help to comprehend the distinct responses to treatments and outcomes.

Carcillo et al (5) previously hypothesized that three inflammatory phenotypes tested in the adult population could be applied to pediatric patients. In a prospective cohort multicenter study, published in 2017, they compared children with severe sepsis with one of these three phenotypes: 1) immunoparalysis-associated multiple organ failure (MOF); 2) thrombocytopenia-associated MOF; and 3) sequential MOF with hepatobiliary dysfunction; to children with severe sepsis and none of these phenotypes, hypothesizing that these phenotypes were associated to higher inflammation and worst outcomes, but this hypothesis had to be still confirmed in other multicenter studies of pediatric MOF patients. In 2016, Manzoli et al (6) had already showed that early MOF and later immunoparalysis were associated with worst outcomes, higher incidence of nosocomial infections, and death in septic pediatric patients. This phenotyping could be used in randomized controlled trial for earlier-outcome assessment and randomization could focus on more severe patients with worse outcomes.

Carcillo et al (1) also described other interesting findings of this very well designed study, including evaluation of nonspecific biomarkers that have been used for a long time in the assessment of septic patients, such as C-reactive protein (CRP) and ferritin, incorporated into phenotypic evaluation. The authors found higher CRP and ferritin levels in children with MOF and any of the phenotypes (and worse outcomes) when compared with MOF patients without any phenotype. These results corroborate previous literature. Our research group published recently (7) a prospective cohort of septic children showing that patients with worse outcomes had higher CRP and Pediatric Logistic Organ Dysfunction scores than the ones with better outcomes. Therefore, CRP, an old and nonexpensive biomarker, available in most units, can be a good prognosis marker too, especially when dynamically evaluated and associated with other biomarkers. Also, another prospective pediatric sepsis cohort has shown that higher serum ferritin values were significantly associated with unfavorable outcomes (8). This leads us to think that stratification of patients may be feasible at the bedside and be remembered and encouraged, even in low- and medium-resource settings.

The phenotyping of patients through the use of biomarkers gives us, not only a path for future trials, but also the possibility of stratifying patients to predict outcomes and to implement earlier clinical management and individualized treatment, in an attempt to improve outcomes in pediatric sepsis.