Forty-Year-Old Bell’s Criteria Still Dominate How Trials Define Necrotizing Enterocolitis

Necrotizing enterocolitis, or NEC, remains one of the most feared diagnoses in any neonatal intensive care unit. The condition, in which portions of a premature infant’s bowel become inflamed and begin to die, can progress with terrifying speed from subtle feeding intolerance to full-thickness intestinal necrosis, perforation, sepsis and death. Despite decades of research, its underlying biology is only partially understood, and its reported incidence varies widely between neonatal networks and countries. Now a new systematic review has highlighted a deceptively simple problem that may be quietly undermining the entire field: the way researchers define NEC in clinical trials has changed remarkably little in nearly half a century.

The review, published in the Journal of Perinatology by a large international team led from Trinity College Dublin, set out to answer a focused question: how do randomised controlled trials, the most rigorous experiments in medicine, actually diagnose and stage NEC? The question matters because a trial is only as good as its outcome measures. If two trials use different definitions of the same disease, their results cannot be cleanly compared or combined in meta-analyses, and regulators and clinicians are left guessing about whether a treatment that appears to work in one setting will work in another.

NEC is a significant cause of morbidity and mortality for preterm neonates, and its stakes have only risen as survival at earlier gestational ages improves. Whole-population surveillance in England has documented the scale of severe disease across neonatal networks, and reviews of contemporary outcomes continue to report substantial mortality among infants who require surgery. The disease is understood to involve a destructive interplay between an immature intestinal barrier, an unstable and often dysbiotic gut microbiome, inflammation mediated in part by innate immune receptors such as toll-like receptor 4, and haemodynamic fragility of the preterm mesentery. Risk factors described in the literature include enteral feeding practices, the protective association of early human milk, maternal smoking, and even in-utero exposures such as indomethacin tocolysis.

Against this complicated biological backdrop sits a diagnostic framework born in 1978. In that year, Bell and colleagues published a staging system for neonatal necrotizing enterocolitis in the Annals of Surgery, designed to guide therapeutic decisions based on clinical staging. The scheme stratified suspected disease from stage one, or suspected NEC with nonspecific systemic signs, through stage two, in which radiographic findings such as pneumatosis intestinalis, gas trapped within the bowel wall, confirm the diagnosis, to stage three, advanced disease with perforation or profound systemic collapse. Walsh and Kliegman refined the criteria in 1986, producing the modified Bell’s staging that generations of neonatologists have since memorised.

Herein lies the conceptual tension that the Dublin-led review interrogates. As the authors point out, Bell’s criteria were never intended as a case definition of NEC. They were a staging tool, created for an era when the sickest infants were often older and more mature than the micro-preemies cared for today. With the survival of neonates at earlier gestations, the clinical phenotype of intestinal injury has shifted, and researchers have repeatedly questioned whether a single framework can capture what is now a heterogeneous spectrum of disease. Some have argued that spontaneous intestinal perforation, a condition with different pathology and outcomes, has been inappropriately lumped together with classic NEC in older trials, muddying the interpretation of surgical studies comparing laparotomy with peritoneal drainage.

To map how trials actually define the disease, the team performed a systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, the PRISMA guidelines that standardise how such evidence syntheses are conducted and reported. They searched PubMed to identify randomised controlled trials published in the last twenty years that used NEC in the study title and included NEC either as a primary outcome or as an inclusion criterion. This dual requirement ensured that every trial analysed genuinely placed NEC at the centre of its scientific question rather than treating it as a passing safety mention.

The screening funnel distilled a large body of literature into a focused evidence base. The initial search identified fifty-six randomised controlled trials, of which thirty-six proceeded to full-text analysis. The headline finding was striking in its consistency: thirty-three of the thirty-six trials used Bell’s criteria or the modified Bell’s criteria to define NEC, while only three trials deployed unique, author-created definitions. In other words, when researchers design the most rigorous experiments on NEC prevention and treatment, the overwhelming majority still anchor their case ascertainment to a staging framework conceived before the advent of modern neonatal intensive care as it exists today.

What does that anchoring mean in practice? Bell’s staging relies on a combination of nonspecific systemic signs, abdominal findings, radiographic evidence such as pneumatosis intestinalis or portal venous gas, and, at the severe end, surgical or autopsy confirmation. Its strengths are real: it is universally recognised, cheap to apply, and requires no specialised laboratory infrastructure. But its weaknesses are equally well documented. Interobserver agreement on stage one disease is notoriously poor, the criteria were never gestational-age adjusted, and they predate the biomarker and imaging revolution now reshaping neonatal diagnostics. Recent work has evaluated neutrophil CD64 as a surveillance marker, explored data-driven diagnostic algorithms integrating clinical and laboratory features, and compared abdominal ultrasonography with plain radiography for detecting disease and predicting severity.

The review’s authors situate their findings within a broader reform movement. A gestational age-specific case definition developed by the UK Neonatal Collaborative has been proposed to capture disease more accurately across the preterm spectrum, and the Vermont Oxford Network maintains its own surveillance definitions, as do the CDC’s NHSN surveillance criteria. Critical evaluations of current definitions have concluded that the field’s diagnostic heterogeneity impedes both research and drug development, with regulatory scientists arguing that a consensus case definition is a prerequisite for any licensed therapy for NEC. The authors of the new review conclude that while the consistent use of Bell’s and modified Bell’s criteria in trials is itself informative, international consensus on further modification of the definition would greatly contribute to both research and clinical practice, allowing greater consistency in staging and therefore optimal management.

The trials catalogued in the review span the full range of neonatal intervention research: prophylactic and therapeutic probiotics including Bifidobacterium breve and Lactobacillus strains, bovine lactoferrin, synbiotics, oral glutamine, enteral L-arginine, docosahexaenoic acid supplementation, bovine colostrum and oropharyngeal colostrum administration, donor human milk fortification, early versus delayed minimal enteral feeding, maternal dietary manipulation, early caffeine treatment and erythropoietin. Each of these trials judged success or failure largely through the lens of a 1978 staging system. If the field can converge on a modern, gestational-age-aware, biologically informed definition, one that perhaps integrates imaging advances, biomarkers and patient-centred research priorities championed by families through organisations such as the NEC Society, the resulting consistency could sharpen future trials, accelerate regulatory approval of preventives and treatments, and ultimately help clinicians identify, stage and treat this devastating disease more reliably.

Subject of Research: How necrotizing enterocolitis is defined and staged in randomised controlled trials involving preterm neonates

Article Title: Definitions of neonatal Necrotizing Enterocolitis (NEC) in randomised controlled trials: a systematic review

Article References: Ballantine, R. S., Campbell, E., Croitoru, O., Jackson, E., Lyne, E. J., McGoldrick, C., Murphy, S. M., Oganezova, K., Shukla, T., Yogesan, S. S., Trayer, J., Stewart, P., Carroll, S., Branagan, A., Roche, E., Tabassum, S., Abrahamsson, T., Embleton, N., Berrington, J., … Molloy, E. J. (2026). Definitions of neonatal Necrotizing Enterocolitis (NEC) in randomised controlled trials: a systematic review. Journal of Perinatology. https://doi.org/10.1038/s41372-026-02905-5

Image Credits: AI Generated

DOI: 10.1038/s41372-026-02905-5

Keywords: necrotizing enterocolitis, Bell's criteria, preterm infants, randomised controlled trials, systematic review, neonatology, clinical staging, PRISMA, gut microbiome, neonatal mortality, case definition, Journal of Perinatology