Nutritional status outperforms platelet markers in predicting prolonged dengue hospitalization in children

Children hospitalized with dengue infection in Yogyakarta, Indonesia, stayed longer than four days in nearly one third of cases, and the strongest independent predictor of that prolonged stay was not the platelet count that clinicians have traditionally used as a marker of dengue severity, but the child's nutritional status. That is the central finding of a retrospective cohort study published in BMC Pediatrics, which followed pediatric dengue patients admitted to Dr. Sardjito General Hospital between 2019 and 2024. The researchers set out to test whether nutritional status, platelet count, platelet indices, and clinical warning signs could predict prolonged hospitalization, defined as a length of stay exceeding four days, and to compare how a platelet-only model, a nutrition-only model, and a full combined model performed against each other.

The study population comprised 104 children who met the inclusion criteria, of whom 102 survivors entered the prolonged length-of-stay analyses. Prolonged hospitalization occurred in 32 of those 102 patients, or 31.4 percent. The team analyzed nutritional status alongside a panel of platelet parameters, including mean platelet volume, platelet distribution width, plateletcrit, and platelet large cell ratio, together with clinical warning signs.

Bivariate analysis and multivariable logistic regression were used to identify independent associations, and the predictive performance of the three models was evaluated using receiver operating characteristic curve analysis, with 5-fold cross-validation for internal validation. Nutritional status was significantly associated with prolonged hospitalization, whereas platelet count, platelet indices, and warning signs were not. In the full multivariable model, both undernutrition and overweight or obesity remained independently associated with a prolonged stay.

Platelet large cell ratio showed only a weak, borderline inverse association that the authors explicitly characterize as exploratory and requiring confirmation. The platelet-only model showed poor discrimination, with an area under the curve of just 0.533, which fell to 0.494 after cross-validation. By contrast, the nutrition-only and full models demonstrated good apparent discrimination, with areas under the curve of 0.777 and 0.855 respectively.

After 5-fold cross-validation, these figures decreased to 0.729 and 0.754, indicating that the models retained acceptable but more modest discrimination when tested on data not used to fit them. The incremental gain in discrimination from adding platelet parameters and warning signs to nutritional status was small after internal validation, amounting to only 0.025 in cross-validated area under the curve. The nutrition-only model also retained higher cross-validated sensitivity than the full model, 84.4 percent versus 65.6 percent, and a higher negative predictive value, 90.7 percent versus 82.3 percent.

Length of stay is an important indicator in pediatric dengue care because it reflects disease severity, healthcare utilization, and hospital resource allocation. Dengue infection remains a major cause of pediatric hospitalization in tropical countries, including Indonesia, and the authors note that the predictive value of combining nutritional status, platelet indices, and warning signs for prolonged hospitalization had remained limited and heterogeneous. Their findings suggest that the traditional focus on platelet parameters may add little to the prediction of which children will occupy a hospital bed for an extended period, at least in the form studied here.

The full model showed better apparent discrimination than the platelet-only and nutrition-only models among the children studied, and all models retained acceptable discrimination on internal cross-validation. Nutritional status showed the strongest independent association with prolonged hospitalization, while the borderline inverse association of platelet large cell ratio should be considered exploratory. The authors emphasize that external validation and further adjustment for disease severity are needed before any clinical application, and the retrospective design means the findings describe associations within a single hospital cohort rather than causal effects.

The analysis was conducted at a single referral hospital in Yogyakarta over a six-year period, so the results may not generalize to other settings or healthcare systems, and the cross-validated performance figures are internal estimates rather than independent tests of the models in new populations. For researchers, the study reframes a familiar question: rather than asking whether falling platelet counts signal a complicated dengue course, it asks which children are most likely to remain hospitalized for more than four days, and points to nutritional status at both extremes, undernutrition and overweight or obesity, as the dominant independent factor in this Indonesian pediatric cohort.

Subject of Research: Factors associated with prolonged hospitalization in children with dengue infection: a retrospective cohort study from Indonesia

Article Title: Factors associated with prolonged hospitalization in children with dengue infection: a retrospective cohort study from Indonesia

Article References: Tahalele, M., Nurnaningsih, Sritanjung, O., Susanto, N. A., Haposan, J. H., & Arguni, E. (2026). Factors associated with prolonged hospitalization in children with dengue infection: a retrospective cohort study from Indonesia. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07860-w

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07860-w

Keywords: dengue, pediatrics, hospitalization, nutritional status, platelet count, Indonesia, retrospective cohort, length of stay, undernutrition, logistic regression, ROC analysis, Factors