European Journal of Pediatrics最新文献

We evaluated three multimodal LLMs, ChatGPT (GPT-5.2), Gemini 3, and Microsoft Copilot, in pediatric ECG interpretation, focusing on clinically significant abnormalities and emergency arrhythmias with likelihood ratios as primary outcome measures. This prospective comparative diagnostic accuracy study (STARD/STARD-AI) included 264 pediatric patients with 12-lead ECGs (November 2024-November 2025). De-identified images were submitted via standardized zero-shot prompt. Three blinded pediatric cardiologists established the reference diagnosis by majority-vote consensus. Cases were classified as Tier 1 (normal), Tier 2 (abnormal, non-urgent), or Tier 3 (urgent). Two binary endpoints were assessed: clinically significant abnormality (Tier 2 + 3 vs Tier 1) and emergency abnormality (Tier 3 vs Tier 1 + 2). Clinically significant abnormalities were present in 54.5% of patients. AUC values ranged from 0.550 to 0.623, reflecting modest discrimination. For the clinically significant endpoint, + LR values were 2.05 (ChatGPT), 1.26 (Gemini), and 1.21 (Copilot); - LR values were 0.68, 0.55, and 0.81, indicating limited rule-in and insufficient rule-out utility. For the emergency endpoint, Gemini achieved 100% sensitivity (95% CI = 85.1-100.0) with - LR 0.07 (95% CI = 0.00-1.12) in a small subgroup (n = 22); however, specificity of 30.2% and + LR of 1.40 indicate overcalling rather than diagnostic precision. No model achieved clinically meaningful rule-in utility for either endpoint.

Conclusions: Current multimodal LLMs showed limited diagnostic utility in pediatric ECG interpretation, with + LR values near 1.0 across both endpoints. Standalone deployment is not supported; these tools may at most serve as adjunctive screening aids under clinician oversight.

What is known: • Deep learning algorithms trained on large ECG datasets perform well in adult populations, but evidence in pediatric ECG interpretation is limited. • General-purpose LLMs show variable accuracy in medical examinations; reliability in subspecialty domains such as pediatric cardiology remains unproven.

What is new: • This is the[FCA1] first head-to-head comparative diagnostic accuracy study of multimodal LLMs in pediatric ECG evaluation, using likelihood ratios as primary outcome measures. • All three LLMs showed limited rule-in utility (+LR near 1.0); Gemini achieved potentially meaningful rule-out performance for emergency arrhythmias (-LR = 0.07), but with wide confidence intervals reflecting the small emergency subgroup (n = 22). • Gemini's 100% sensitivity in the emergency subgroup reflects overcalling (specificity 30.2%) consistent with a triage/screening behavior rather than diagnostic precision.

The aim of this study was to identify the predictors of relapse and severe disease in non-systemic juvenile idiopathic arthritis (JIA), a heterogeneous childhood disease. Patients with JIA were grouped based on relapse status, and those with ≥ 2 relapses requiring biologics were classified as severe disease phenotype. A total of 142 patients (63.4% female) were included in the study. Seventy-three patients (51.4%) experienced at least one relapse after achieving remission, who were significantly characterized by female gender, younger age at diagnosis, positive ANA test, and longer disease duration. Ankle, elbow, metacarpophalangeal (MCP), and temporomandibular joint (TMJ) involvement was more prevalent in patients who had experienced at least one relapse. Longer disease duration, higher number of joints involved at the time of diagnosis, and MCP involvement were found as independent risk factors for relapse. Twenty patients (14%) were grouped as having a severe disease phenotype, characterized by a younger age at the time of diagnosis, longer disease duration, and a higher number of joints involved throughout the disease course. Younger age at diagnosis, ankle involvement, and TMJ involvement were found to be independent risk factors for a severe disease phenotype.

Conclusion: In our study, longer disease duration, MCP joint involvement, and a higher number of joints involved at the time of diagnosis were found to be associated with relapse in non-systemic JIA patients. On the other hand, younger age at diagnosis, ankle and TMJ involvement were associated with severe disease phenotype.

What is known: • JIA is a chronic joint disease characterized by relapses and remissions, and approximately half of patients experience at least one relapse after remission.

What is new: • Multiple relapses and bDMARD requirement may define a severe disease course. Younger age, ankle, and TMJ involvement may predict severe disease course. • Relapses may occur with longer disease duration; not every relapse implies a severe disease course.

Pediatric early warning scores (PEWS) are widely used to detect clinical deterioration, yet their prognostic performance and clinical utility in low- and middle-income emergency departments (EDs) remain uncertain. We evaluated a modified Brighton PEWS for predicting 24-h ICU admission in a resource-limited pediatric ED and developed an internally validated time-to-event model. We conducted a 12-month prospective cohort of children aged 1 month-15 years triaged as emergency severity index level 2-3 in a tertiary pediatric ED. Modified Brighton PEWS was recorded at arrival. The primary outcome was ICU admission within 24 h. Time from ED arrival to ICU admission was analysed using Cox regression adjusted for age, sex and comorbidity, with bootstrap internal validation. We compared a PEWS-only model with an enhanced model incorporating routinely available ED variables. Performance was assessed by 24-h calibration, discrimination and clinical utility using decision-curve analyses (DCA). Among 572 children, 157 (27.5%) were admitted to ICU within 24 h. ICU admission occurred in 7.1%, 45.3%, and 53.9% of children with PEWS ≤ 3, 4-5, and ≥ 6, respectively (log-rank p < 0.001). The multivariable time-to-event model showed good discrimination (C-index 0.81) and acceptable 24-h calibration. Baseline PEWS predicted 24-h ICU admission with an AUC of 0.80; a threshold ≥ 4 achieved high sensitivity with reasonable specificity. DCA indicated net benefit of PEWS-based risk stratification across clinically relevant thresholds.

Conclusion: In a resource-limited pediatric ED, a modified Brighton PEWS at arrival provided robust prediction of 24-h ICU admission and can underpin risk-stratified escalation pathways.

What is known: • PEWS are widely used to identify children at risk of clinical deterioration. • Evidence on prognostic accuracy and clinical utility in LMIC pediatric emergency departments remains limited.

What is new: • In a resource-limited pediatric ED, a modified Brighton PEWS at arrival predicted 24-h ICU admission. • Decision-curve analysis supported PEWS-based risk stratification across clinically plausible thresholds.

Infective endocarditis (IE) in children is rare but potentially severe. Unlike adults, no pediatric-specific guidelines exist to guide echocardiography in children with bacteremia. We aimed to identify risk factors for IE in pediatric bacteremia and propose a selective, risk-based approach to echocardiographic evaluation. An 8-year prospective surveillance included all children (0-18 years) hospitalized with true bacteremia at Soroka University Medical Center between 2015 and 2022. A bacteremia episode was defined as the growth of a single organism during hospitalization. Episodes in which echocardiography was performed were included in the comparison between definite/possible IE, classified according to modified Duke criteria, and episodes in which IE was rejected. Among 2810 bacteremia episodes, echocardiography was performed in 573 (20%); 569 were analyzed. IE was classified as definite in 3/569 (0.5%) and possible in 228/569 (40.1%), while in 338 (59.4%) episodes, IE was rejected. Only 7/231 (3.0%) definite/possible IE cases had diagnostic echocardiographic findings. Independent predictors of IE included high-risk cardiac anomalies (prosthetic valve or material and cyanotic heart disease), prolonged bacteremia, Staphylococcus aureus bacteremia, and embolic or immunologic phenomena. Antimicrobial resistance, central venous catheters, and prior IE were not associated with an increased risk. Physical examination findings, including murmurs, did not differentiate IE from non-IE episodes.

Conclusion:  The rate of definite IE in pediatric bacteremia was extremely low, and echocardiographic yield was limited. Echocardiography should be reserved for children with high-risk cardiac conditions, embolic or immunologic phenomena, or additional risk factors as persistent bacteremia particularly Staphylococcus aureus.

What is known: • Pediatric infective endocarditis (IE) is rare, particularly in children without high-risk cardiac anomalies. • Unlike in adults, there are no specific guidelines on when to perform echocardiography in children with bacteremia.

What is new: • Definite IE was very uncommon among children with bacteremia, and only 3% of definite/possible IE episodes had diagnostic echocardiographic findings. • IE was independently associated with high-risk cardiac anomalies, persistent and S. aureus bacteremia, but not with central venous catheters, antimicrobial resistance or prior IE.

The purpose of this study was to evaluate whether a nurse-driven high-flow nasal cannula (HFNC) weaning protocol reduces HFNC duration and hospitalization in infants with severe bronchiolitis. This prospective, randomized controlled study was conducted in two tertiary pediatric intensive care units (PICUs) during a single bronchiolitis season. Infants aged 1-24 months requiring HFNC for severe bronchiolitis were randomized to standard care or a nurse-driven protocol group. The intervention included structured assessments using the Wang Bronchiolitis Severity Score (WBSS), and ROX index (SpO₂/FiO₂ divided by respiratory rate). Trained nurses initiated weaning based on predefined criteria. The primary outcome was total HFNC duration. Secondary outcomes included time to first weaning, PICU and hospital length of stay, need for non-invasive ventilation (NIV), intubation, readmission rates, nurse satisfaction, and protocol adherence. A total of 110 patients were included (55 per group). The nurse-driven group had significantly shorter median HFNC duration (34.00 h vs. 50.00 h, p = .001) and earlier weaning initiation (12.00 h vs. 20.00 h, p < .001). PICU stay (3.00 vs. 4.00 days, p = .005) and hospital stay (6.00 vs. 7.00 days, p = .001) were reduced. No significant differences were found in NIV use (p = .670), intubation (p = .450), or readmissions (p = .100). Nurse satisfaction was 96.0%, and protocol adherence was 92.0%.

Conclusion:  A nurse-driven HFNC protocol is associated with reduced therapy duration and hospitalization in infants with bronchiolitis without increasing adverse outcomes. Empowering nurses with standardized tools may support timely and effective respiratory care in the PICU.

Trial registration: ClinicalTrials.gov (NCT06621641), registered on November 22, 2025.

What is known: • High-flow nasal cannula (HFNC) therapy is widely used in infants with bronchiolitis, despite inconsistent evidence regarding its impact on length of stay and the need for escalation of care. • Lack of standardized weaning criteria may contribute to prolonged HFNC use and increased resource utilization.

What is new: • This randomized controlled study demonstrates that a nurse-driven HFNC weaning protocol significantly reduces HFNC duration and hospital length of stay in infants with severe bronchiolitis. • Structured, score-guided nurse-driven weaning can be safely implemented in the PICU without increasing the need for non-invasive ventilation or intubation.

The objective of this study is to compare conservative versus conventional oxygen saturation targets on clinical outcomes in critically ill children requiring respiratory support. This open label randomized controlled trial was conducted in a tertiary care pediatric intensive care unit between May 2023 and November 2024. Children aged 1 month to 15 years requiring invasive or noninvasive oxygen therapy were randomized to conservative (SpO₂ 88-92%) or conventional (SpO₂ 94-99%) oxygen saturation targets. Fraction of inspired oxygen was titrated to maintain assigned targets. The primary outcome was a composite of death and organ support at 30 days. Secondary outcomes included mortality at 7 and 30 days, duration of respiratory support and oxygen therapy, length of PICU and hospital stay, and oxidative stress assessed by serum malondialdehyde (MDA) at baseline and day 7. A total of 178 children were randomized. Baseline characteristics and illness severity were comparable between groups. The median composite outcome score was 8 (IQR 4-20.25) in the conservative group and 10 (IQR 5-20) in the conventional group (p = 0.15). Duration of respiratory support (4 vs 6 days; p = 0.003) and oxygen therapy (8 vs 100 h; p < 0.001) were significantly lower in the conservative group. Mortality at 7 and 30 days, organ support days, length of stay, and MDA levels were similar.Conclusion: Conservative oxygen saturation targets (88-92%) were safe and resulted in similar mortality and organ dysfunction compared with conventional targets, with reduced duration of respiratory support and oxygen therapy. Larger multicenter trials are needed to confirm these findings.Trial registration: The trial is registered in Clinical Trial Registry India (CTRI) (registration number CTRI/2023/11/060105, date of registration 21/11/2023.

Preterm infants on invasive breathing support are at increased risk for acquired infections called ventilator-associated pneumonias (VAP). Wide heterogeneity exists in diagnostic criteria and management of VAP. The objective of this study is to determine the demographic characteristics and prevalence of VAP according to three different existing criteria used in the neonatal intensive care unit (NICU). We conducted a retrospective study including infants born at < 33 weeks' gestational age (GA) from 3 different NICUs in Western Canada who have been diagnosed with VAP during 2018-2022. The incidence of VAP was categorized based on physicians' discretion, Canadian consensus, and National Healthcare Safety Network (NHSN) criteria. There were 185 neonates diagnosed to have VAP at physicians' discretion, with incidences of 4.6%, 8.9%, and 5.7% in the three NICUs. Among the VAP per physicians' diagnoses, 18.2-55.0% of infants fulfilled the VAP criteria of Canadian consensus, and 10.9-12.5% of them fulfilled the NHSN criteria. Conclusion: We found that 4.6-8.9% of infants born at < 33 weeks GA had VAP diagnosed during their NICU stay. About one-tenth of these cases fulfilled the NHSN criteria. Therefore, there is an urgency to develop neonatal-specific and validated diagnostic criteria for VAP. What is Known: • Preterm infants (33 weeks gestational age) are at increased risk for ventilator-associated pneumonia (VAP), a serious lung infection associated with mechanical ventilation. However, there is significant heterogeneity in the diagnostic criteria and management of VAP, which have not been validated in the neonatal population What is New: • 4.6-8.9% of infants are diagnosed with VAP during their NICU stay. However, only around one-tenth meet the NHSN diagnostic criteria. • These discrepancies highlight an urgent need to develop neonatal-specific, validated diagnostic criteria for VAP.

Food protein-induced enterocolitis syndrome (FPIES) is a non-immunoglobulin E-mediated food allergy that typically causes delayed gastrointestinal symptoms following the ingestion of a trigger food. FPIES is a rare condition with limited epidemiological data. This study aimed to investigate the demographic characteristics, triggers, clinical presentations, and outcomes of children with acute FPIES at the University Children's Hospital Zurich, a tertiary-level children's hospital in Switzerland, with a dual focus on confirmed and suspected FPIES cases. This retrospective study covered a 10-year period (2010-2020). Electronic medical records were screened for patients diagnosed with acute FPIES and those with symptoms suggestive of FPIES according to international diagnostic criteria. Those paediatric patients with an acute FPIES diagnosis or a history highly suggestive of FPIES were included. A total of 109 patients with acute FPIES were identified. Cow's milk (CM) was the most common food trigger (24%), followed by eggs (12%) and fish (10%). A single food trigger was identified in 56% of the patients. The median age of onset was 6 months (interquartile range (IQR), 4 months). Gender had no effect on observed FPIES cases. Data on FPIES tolerance and therefore resolution of FPIES symptoms and/or successful reintroduction of the triggering food were limited, with sufficient follow-up information available for only 37% of confirmed and 25% in suggestive cases: 27.5 months for CM (IQR, 15.75 months), 67.5 months for fish (IQR, 33.5 months), and 32 months for egg (IQR, 14.5 months).  Conclusion: The study contributes practical epidemiological and clinical insights into FPIES triggers, patterns, and patient characteristics, and provides further regional insights to navigate this complex condition. What is Known:  • FPIES is a rare, non-IgE-mediated food allergy, particularly affecting infants and young children. • FPIES is often under-recognized, resulting in delayed diagnosis and management. What is New: • By providing robust real-world data on acute FPIES, this study provides a decade-long overview of acute FPIES cases, detailing key food triggers, identifying gaps in clinical practice and delays in diagnosis, and contributing regional epidemiologic data that may not be apparent in short-term studies and underscore geographic variability in presentation and triggers.

The association between hypospadias, its severity, and anogenital distance remains inconsistent. This study aims to investigate the association between anogenital distance and hypospadias through a meta-analysis. We conducted a comprehensive search in international databases (PubMed, Web of Science, Cochrane Library, Embase, Medline) and Chinese databases (China National Knowledge Infrastructure, Wanfang Data, VIP database). We searched for studies on the association between anogenital distance and hypospadias. The search period extended until November 1, 2025. Thirteen studies involving 6905 male infants were included in the meta-analysis. We calculated the standardized mean difference and 95% confidence interval for anogenital distance using a random-effects model. The results indicated that children with hypospadias had a shorter anogenital distance (standardized mean difference, - 1.24; 95% confidence interval, - 1.74 to - 0.74). A negative correlation was observed between anogenital distance and hypospadias severity. Children with more severe hypospadias had a shorter anogenital distance. The standardized mean differences were - 1.05 (95% confidence interval, - 2.23 to 0.13) for distal hypospadias, - 1.66 (- 3.14 to - 0.18) for middle hypospadias, and - 3.00 (- 4.70 to - 1.31) for proximal hypospadias.

Conclusion:  Our study suggests that shortened anogenital distance is associated with hypospadias and may correlate with disease severity, consistent with the hypothesis of fetal androgen disruption. However, due to the observational nature of the included studies and significant heterogeneity, causal inference is limited. Anogenital distance may serve as a research endpoint, but prospective studies are needed to validate its etiological and clinical utility.

Trial registration: The study protocol has been registered in PROSPERO ( https://www.crd.york.ac.uk/prospero ) under the registration number CRD42023407183.

What is known: • In animal studies, anogenital distance is negatively correlated with hypospadias severity. • The association between anogenital distance and hypospadias has yielded inconsistent results in human studies.

What is new: • Anogenital distance is negatively correlated with the severity of hypospadias in human studies, which supports the role of impaired fetal androgen production/action in the pathogenesis of hypospadias. • Anogenital distance may serve as an indicator for reproductive health assessment endpoints.

Empirical data on post-injury disability in children is limited and deriving disability weights is urgently needed. The aim of this study is to utilise pooled data to determine adequate disability weights in children and adolescents. Five longitudinal prospective cohort studies of pediatric injury survivors aged 5-17 years were pooled (N = 1972) to create case-based pediatric weights for established nature-of-injury classifications using four types of EQ-5D utility scores to represent overall health-related quality of life. Healthy population norms from seven countries formed a sensitivity analysis. Disability weights (DW) were calculated at 1 month, 4 months, 6 months, and 12 months post injury to produce two disability weights per injury class: 12-month residual disability weights (12dw) and 12-month annualised weights (12aw). DW for asphyxiation/non-fatal submersion, spinal cord lesion, fracture of the femur, fracture of pelvis, and fracture of vertebral column, were more than twice that of the lowest DW for fracture of clavicle, scapula, or humerus, and fracture of radius or ulna. 12dw for moderate/severe traumatic brain injury (TBI) was 19% higher than minor TBI and 25% higher for 12aw. Conclusion: Different DW should be applied to DALY calculations for children and adolescents compared to adults. The calculation of these DWs is complex and warrants further investigation. This study confirms that injury is often a chronic disorder and burden of disease for children and adolescents and estimates should reflect this situation. What is Known: • Traumatic injuries pose substantial threats to children's health, education and social inclusion • Much-needed disability weights, used in calculating the years lived with disability (YLDs), has been limited by the lack of empirical data on postinjury disability in children What is New: • This study confirms injury is often a chronic disorder and burden of disease for children and adolescents • Injury-group disability weights for children and adolescents are provided for YLD calculations, and these weights differ to those used for adults.