The case for that shift rests on a fundamental biological reality: sepsis is not one disease. It is a syndrome of dysregulated host responses to infection, and in children this heterogeneity is particularly striking. Different pathogens activate fundamentally different immune cascades, from novel respiratory viruses such as influenza and respiratory syncytial virus to diverse Gram-negative and Gram-positive bacteria and fungi. Layered on top of this pathogen diversity are individual genetic polymorphisms that make each child’s initial response drastically different. Using peripheral blood RNA transcriptomic sequencing, international research teams have identified distinct sepsis response signatures, including the SRS1 and SRS2 endotypes, revealing a highly polarized biological spectrum hiding behind a single clinical label.

The divergence does not stop at molecular signatures. Some patients present early with a hyperinflammatory endotype, the so-called cytokine storm, driving widespread systemic endothelial activation and a surge in vascular permeability. Others, especially neonates or children with underlying immunodeficiencies, rapidly slide into the opposite state: a hypoinflammatory or immunoparalyzed endotype characterized by leukocyte reprogramming and global mitochondrial bioenergetic failure. Big data approaches and unsupervised machine learning algorithms such as latent class analysis have confirmed that pediatric sepsis comprises distinctly different severe sub-phenotypes, including shock-dominant forms, multiorgan dysfunction syndrome, and hyperferritinemic or macrophage activation-like syndromes. Two children with identical sepsis diagnoses may therefore harbor completely antithetical underlying host responses and mechanisms of injury.

Meanwhile, the way acute kidney injury itself is diagnosed is under intense scrutiny. For decades, staging has relied almost entirely on the Kidney Disease: Improving Global Outcomes criteria, based on elevated serum creatinine and decreased urine output. But these are crude functional filtration markers that lack microstructural pathological granularity. In pediatric critical care, baseline creatinine varies widely by age, developmental stage, sex, and muscle mass, and it rises with a significant physiological lag after renal injury. Early fluid resuscitation compounds the problem: pathological fluid accumulation and fluid overload dilute serum creatinine, frequently masking or delaying a timely diagnosis. Static urine output assessments, the editorial notes, fail to integrate mass balance and the actual fluid dynamics of the patient. Notably, the newly published KDIGO 2026 draft guidelines explicitly advocate incorporating novel structural and stress biomarkers alongside traditional functional metrics to redefine acute kidney injury and recognize subclinical structural damage before function declines.

The kidney itself is far from a uniform organ. Anatomically it is divided into cortex, outer medulla, and inner medulla, each with radically different vulnerabilities. The cortex receives abundant blood flow and drives filtration, whereas the medulla receives less than ten percent of total renal blood flow, and its specialized vasa recta countercurrent multiplier system leaves it naturally on the brink of hypoxia. In sepsis, systemic hemodynamic redistribution often produces severe, occult hypoxia in the renal medulla that standard monitoring never sees. The organ contains more than thirty distinct cell types, including glomerular endothelial cells, podocytes, and proximal and distal tubular epithelial cells, each reacting uniquely to circulating inflammatory cytokines, endotoxins, and damage-associated molecular patterns. The proximal tubule, particularly the S3 segment, is highly metabolically active and dependent on mitochondrial oxidative phosphorylation, making it the epicenter of structural damage.

Single-cell technologies have pushed this anatomical insight to a molecular level. Single-cell RNA sequencing shows that metabolically active proximal tubular cells and distinct nephron segments, such as the loop of Henle, follow divergent gene reprogramming and injury pathways during sepsis. Critically, two pediatric patients meeting identical KDIGO stage 3 criteria may harbor diametrically opposed internal microenvironments. One may experience a transient, functional alteration driven by relative volume deficit, rapidly reversible with titrated fluid resuscitation. The other may suffer persistent, damage-associated injury marked by acute tubular necrosis, severe microvascular endothelial barrier disruption, and widespread apoptosis. Expanding on initial adult cohorts, recent pediatric critical care research has validated distinct biological sub-phenotypes of acute kidney injury in critically ill children, including low-inflammatory/mild injury and hyperinflammatory/endothelial injury groups, whose clinical trajectories, renal recovery probabilities, and long-term mortality risks are fundamentally disparate.

When systemic sepsis heterogeneity intersects with organ-level microenvironmental heterogeneity, the kidney becomes both a passive target of the immune storm and an active amplifier of inflammation and microcirculatory breakdown. Modern pathophysiological research shows that hyperinflammatory sepsis triggers widespread shedding and degradation of the vascular endothelial glycocalyx, the protective sugar-rich coating of blood vessels. Stripped of this layer, the microvascular endothelium exposes adhesion molecules, prompting neutrophils and platelets to aggregate, adhere, and roll along renal microvascular walls, producing severe spatial mismatching of capillary blood flow and localized microthrombi. This explains the dissociation between macrocirculation and microcirculatory tissue perfusion: even when systemic blood pressure is normalized and total renal blood flow preserved or increased, local renal microvessels remain trapped in severe, occult tissue hypoxia. To survive, tubular epithelial cells initiate gene reprogramming and metabolic downregulation, entering a protective, cell-cycle-arrested quiescent state.

This structure-function mismatch, the authors argue, is precisely why past large-scale trials in sepsis-associated kidney injury overwhelmingly failed: interventions were applied uniformly across broad clinical syndromes, ignoring the distinct biological sub-phenotypes of individual patients. To break the deadlock, recent international consensus reports, including the 23rd and 28th Acute Disease Quality Initiative workgroup statements, advocate transforming static functional diagnoses into a dynamic theragnosis framework built on predictive enrichment. The diagnostic core is a multimodal biomarker matrix. Structural injury markers such as neutrophil gelatinase-associated lipocalin (NGAL) and kidney injury molecule-1 (KIM-1) enable detection of subclinical stage 1S injury before creatinine rises, allowing clear stratification of functional versus damage-associated injury. Cell-cycle arrest markers, specifically the product of urinary tissue inhibitor of metalloproteinase-2 and insulin-like growth factor-binding protein 7 (TIMP-2 × IGFBP-7), and tubular stress indicators like urinary Dickkopf-3 provide early warning of children progressing toward irreversible tubular damage. Simultaneously profiling systemic biomarkers, including interleukin-6, chemokines, soluble thrombomodulin, angiopoietin-2, and olfactomedin-4, allows clinicians to map whether a patient carries a hyperinflammatory/endothelial-disruption endotype or a hypoinflammatory/energy-suppressed one.

Treatment, in turn, must be tailored to phenotype, because directed interventions show profound heterogeneity of effect across pediatric sub-phenotypes. The editorial, grounded in the landmark AKI1 (low-inflammatory/mild injury) and AKI2 (high-inflammatory/severe endothelial injury) sub-phenotype framework updated with pediatric evidence, maps dichotomous, phenotype-driven intervention strategies. Looking ahead, the authors call for a tripartite synergy: artificial intelligence-driven clinical decision support systems that integrate bedside point-of-care biomarker testing with electronic health record algorithms, such as an automated renal angina index paired with a secondary urinary NGAL tiering pathway; refined clinical trial endpoints that abandon renal replacement therapy initiation as an indiscriminate endpoint in favor of biologically aligned benchmarks such as RRT-free survival days, AKI duration, and fluid-balance-related organ damage; and adaptive platform and umbrella trial designs that introduce transcriptomic endotypes and segment-specific urinary biomarkers into multicenter research. By abandoning the one-size-fits-all approach, the editorial concludes, clinicians can build a precise diagnostic framework bridging systemic host responses and local renal microenvironments, ultimately rewriting the long-term renal outcomes and survival trajectories of every critically ill child with sepsis.

Subject of Research: Precision medicine approaches for pediatric sepsis-associated acute kidney injury based on systemic and organ-level heterogeneity

Article Title: Precision medicine in pediatric sepsis-associated acute kidney injury: when systemic heterogeneity meets organ-level heterogeneity

Article References: Wu, M.-Y., Zhao, Y., Chen, X.-H., & Ge, X.-H. (2026). Precision medicine in pediatric sepsis-associated acute kidney injury: when systemic heterogeneity meets organ-level heterogeneity. World Journal of Pediatrics. https://doi.org/10.1007/s12519-026-01096-5

Image Credits: AI Generated

DOI: 10.1007/s12519-026-01096-5

Keywords: pediatric sepsis, acute kidney injury, precision medicine, biomarkers, sub-phenotypes, serum creatinine, KDIGO guidelines, single-cell RNA sequencing, endothelial glycocalyx, machine learning, renal angina index, intensive care