Discriminating the subtype of hepatic glycogen storage disease by routine blood biochemistry
In brief
Routine blood tests classified pediatric liver glycogen disease subtypes with 86.5% accuracy
In a validation group of children with genetically confirmed disease, a decision tree using triglycerides, lactate, and creatine kinase correctly classified 86.5% of hepatic glycogen storage disease subtypes. Positive predictions were especially accurate for types I and III, but the model is exploratory and needs prospective testing at multiple centers before its role in care is clear.
- Journal
- Pediatric investigation (Q2)
- Published
- 18 November 2026
- Study design
- Prospective / inception cohort
- Evidence level
- Level 2, Moderate (CEBM 2b)
- Authors
- Xueyuan Zhang, Yuchuan Li, Weicheng Wu, Shiqi Lu, Yi Lu, Xinbao Xie, et al.
- PMID
- 43055007
- DOI
- 10.1002/ped4.70094
Why clinicians should know about it
- Picked for Biochemistry (medical) (top studies of the week, 22 November 2026): Uses routine biochemistry to distinguish inherited metabolic disease subtypes
Abstract
IMPORTANCE: Early discrimination of major hepatic glycogen storage disease (GSD) subtypes is crucial for effective GSD management. OBJECTIVE: To discriminate between major hepatic GSD subtypes-types I, III, and VI/IX-using routine blood biochemical parameters. METHODS: A total of 159 pediatric patients with genetically confirmed hepatic GSD between July 2014 and July 2025 were retrospectively collected and were divided into a training set and a validation set. A random forest algorithm was first applied to identify the most predictive biochemical variables. Subsequently, a classification and regression tree model was developed based on the variables. Model performance was evaluated using overall accuracy, sensitivity, specificity, positive predictive value (PPV), and Cohen's kappa. Bootstrap resampling was used to assess model stability. RESULTS: The blood levels of triglycerides (TG), lactate (LAC), and creatine kinase (CK) were first identified as the top predictors. The resulting decision tree model classified GSD subtypes: TG ≥3.0 mmol/L and LAC ≥4.4 mmol/L for GSD I; either TG ≥3.0 mmol/L, LAC < 4.4 mmol/L, and CK ≥135.0 U/L, or TG <3.0 mmol/L, and CK ≥198.0 U/L for GSD III; TG <3.0 mmol/L and CK <198.0 U/L for GSD VI/IX. In the validation cohort, the model achieved an overall accuracy of 86.5% (kappa = 0.744), with a PPV of 91.7% for GSD I and 100% for GSD III. INTERPRETATION: This study provides an exploratory decision tree based on clinical laboratory parameters that may help discriminate among hepatic GSD subtypes, pending prospective multicenter validation.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.