When a child arrives at a hospital with COVID-19, clinicians face a difficult question: will this infection stay mild, or is it heading toward something dangerous? In adult medicine, blood tests that measure inflammation have often helped answer that question, and pediatricians have long hoped that simple, widely available markers could do the same for children. A new retrospective cohort study from Dr. Sardjito General Hospital in Yogyakarta, Indonesia, published in BMC Pediatrics, has now put three of the most accessible candidates to the test: platelet count, mean platelet volume, and the mean platelet volume-to-platelet ratio, often abbreviated MPR. The verdict is sobering. All three markers showed minimal capacity to distinguish severe from non-severe COVID-19 in pediatric patients, suggesting that the search for a reliable routine blood-based severity predictor in children must continue.
The biological rationale behind these markers is worth unpacking. Platelets are tiny cell fragments best known for their role in blood clotting, but they are also active participants in the immune response. During viral infections, the body’s inflammatory signals can alter both the number of platelets in circulation and the characteristics of individual platelets. Mean platelet volume, or MPV, reflects the average size of a platelet; larger platelets are typically younger and more enzymatically active, and elevated MPV is often interpreted as a sign of heightened platelet turnover under inflammatory stress. The MPR combines both dimensions by dividing MPV by the platelet count, yielding a composite index that some researchers have proposed as a more sensitive gauge of inflammation than either component alone. These markers are already used as predictors in pediatric sepsis and other critical illnesses, which is precisely why the Indonesian team wanted to know whether they could earn a place in COVID-19 care.
The study’s design was straightforward but rigorous. The researchers conducted a retrospective cohort study involving pediatric COVID-19 patients admitted to Dr. Sardjito General Hospital between January 2020 and December 2022, a window that spanned the pandemic’s most turbulent years in Indonesia. Rather than sampling a subset, they applied a total sampling approach, enrolling every patient who met the defined inclusion and exclusion criteria. The analysis combined bivariate and multivariate techniques, using logistic and linear regression to test associations between the blood markers and disease severity. Associations were expressed as odds ratios with 95 percent confidence intervals, and statistical significance was set at a p value below 0.05. In total, 297 children with confirmed COVID-19 were included, giving the study a reasonable statistical foundation for evaluating the markers’ performance.
The headline result came from receiver operating characteristic analysis, the standard statistical tool for judging how well a biomarker discriminates between two clinical states. In ROC analysis, the area under the curve, or AUC, quantifies discriminative power: an AUC of 1.0 indicates perfect separation, while an AUC of 0.5 indicates performance no better than a coin flip. For all three biomarkers, the AUC values approached 0.50, signifying a restricted capacity to differentiate between severe and non-severe COVID-19 cases. In plain terms, knowing a child’s platelet count, MPV, or MPR added almost nothing to the ability to predict whether that child’s illness would become severe. The inadequate discriminative efficacy revealed by the ROC analysis applied across the board, to every marker assessed.
The bivariate analysis told a more nuanced story, however, and it is here that the study’s technical findings deserve careful attention. The researchers identified cut-off values for each marker in association with severe COVID-19. For platelet count, the cut-off was 246,500 cells per microliter, but the association did not reach statistical significance (p = 0.089; OR = 1.57; 95 percent CI 0.93 to 2.64). For MPV, the cut-off was 9.25, again falling short of significance (p = 0.069; OR = 1.72; 95 percent CI 0.96 to 3.09). The MPR cut-off of 8.95 was the only marker to achieve significance in the bivariate analysis (p = 0.021; OR = 2.09; 95 percent CI 1.11 to 3.94), meaning that children with MPR above this threshold had roughly twice the odds of severe disease. Yet even this statistically significant association translated into weak discriminative performance when judged by ROC standards, a discrepancy that highlights the difference between an odds ratio and true predictive utility.
When the blood markers were tested together with clinical variables in regression analysis, the picture shifted in an interesting direction. Three factors emerged as independent predictors of COVID-19 severity in children: an MPV of 9.25 or higher (p = 0.031; OR = 1.953; 95 percent CI 1.062 to 3.593), the presence of comorbidity (p = 0.13; OR = 6.526; 95 percent CI 1.496 to 28.465), and moderate-to-severe malnutrition (p = 0.18; OR = 2.210; 95 percent CI 1.147 to 4.257). The comorbidity odds ratio of 6.5 stands out: children with pre-existing health conditions had more than six times the odds of severe COVID-19 compared with those without. Malnutrition more than doubled the odds. MPV retained its independent association, though with a more modest effect size. Platelet count and MPR, by contrast, dropped out of the final model entirely, unable to hold their ground once clinical variables were accounted for.
Why did these markers perform so poorly in children when they have shown more promise in other contexts? The answer likely lies in the distinctive immunology of pediatric COVID-19. Clinical manifestations in children are more heterogeneous than in adults, and several factors may influence disease severity, including the interplay between viral biology and the developing immune system. SARS-CoV-2 enters cells through the angiotensin-converting enzyme 2 receptor with assistance from the transmembrane serine protease 2, and the downstream inflammatory cascade that drives severe disease in adults, including acute respiratory distress syndrome, appears to unfold differently in children. Many children mount a more controlled immune response, and the platelet changes that reflect inflammatory stress in adults may simply be blunted or absent in pediatric patients. The heterogeneity of pediatric presentations, from asymptomatic infection to multisystem inflammation, makes a single blood marker an inherently blunt instrument.
The study’s conclusion is unambiguous: platelet count, MPV, and MPR had minimal capacity to predict COVID-19 severity in pediatric patients. This finding carries practical weight for clinicians in Indonesia and beyond, where these markers are attractive precisely because they are cheap, fast, and available on any standard hematology analyzer. The temptation to rely on them for triage decisions is real, particularly in resource-limited settings. The results suggest that such reliance would be misplaced. Instead, the regression analysis points toward clinical variables, comorbidity and nutritional status, as the more meaningful signals, reinforcing that a child’s overall health context matters more than any single laboratory value drawn from a complete blood count.
There are also important caveats to keep in view. As a retrospective study based on secondary data from pre-existing medical records, the research depended on the quality of documentation from the pandemic years, and the ethics committee appropriately waived the requirement for individual informed consent given the use of archival records. The study was conducted at a single center, and the markers were evaluated in a specific population during a defined period. The published version is an early release, subject to further edits before the final Version of Record appears. Still, the statistical approach was sound, the sample of 297 patients was substantial for a pediatric COVID-19 cohort, and the total sampling design minimized selection bias within the inclusion criteria.
The broader lesson for pediatric COVID-19 research is that the search for reliable severity biomarkers must look elsewhere, or at least in combination. The failure of these three markers does not mean that blood-based prediction is impossible in children; it means that platelet indices alone are insufficient. Future studies may need to explore composite scoring systems that integrate clinical variables, nutritional assessment, and multiple laboratory parameters, or to investigate markers of the specific inflammatory pathways that drive severe disease in young patients. For now, the message from Yogyakarta is clear and clinically useful: in children with COVID-19, a routine platelet panel cannot tell clinicians which child will deteriorate. Attention should turn instead to the factors that the data did identify, the chronic conditions and nutritional deficits that shape a child’s resilience against the virus, and that remain far more powerful predictors than any single number on a hematology report.
Subject of Research: Platelet markers show minimal predictive value for COVID-19 severity in children, Indonesian cohort study finds
Article Title: Predictive performance of platelet count, mean platelet volume, and mean platelet volume-to-platelet ratio in pediatric COVID-19: a cohort retrospective study in Indonesia
Article References: Beandda, A. S., Setyati, A., Citta, A. N., Susanto, N. A., & Arguni, E. (2026). Predictive performance of platelet count, mean platelet volume, and mean platelet volume-to-platelet ratio in pediatric COVID-19: a cohort retrospective study in Indonesia. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07608-6
Image Credits: AI Generated
DOI: 10.1186/s12887-026-07608-6
Keywords: COVID-19, pediatrics, platelet count, mean platelet volume, biomarkers, disease severity, Indonesia, BMC Pediatrics, Predictive, performance, platelet, count
News Source: Kristina Jarvis. (October 18, 2026). Platelet markers fall short as severity predictors in children with COVID-19, Indonesian study finds. Scienmag.
