Association Among Systemic Immune-Inflammation Index, Vitamin D Status and Clinical Subtypes of ADHD: A Retrospective Investigation

Abstract

Background: Attention-deficit/hyperactivity disorder (ADHD) has been linked to immune dysregulation and nutritional imbalance. The systemic immune-inflammation index (SII) is a composite inflammatory marker, while serum 25-hydroxyvitamin D (25(OH)D) reflects vitamin D status. Their joint association with paediatric ADHD remains incompletely characterised. Methods: In this retrospective case-control study, 78 treatment-naïve children with ADHD and 40 non-ADHD outpatient comparison children selected using frequency matching by age and sex were included. The most recent laboratory results available in the outpatient record were extracted. Given that some 25(OH)D values were derived from reports obtained outside our hospital, all available 25(OH)D records were manually re-checked against source documents and standardised to ng/mL prior to analysis. Group comparisons, whole-sample Spearman correlation analyses, logistic regression models including an interaction term between standardised SII and standardised 25(OH)D and exploratory receiver operating characteristic (ROC) analyses were performed. Sensitivity analysis and Firth penalised logistic regression were additionally conducted. Results: Compared with the comparison group, children with ADHD had significantly higher SII, C-reactive protein (CRP), neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) levels, together with lower serum 25(OH)D. The ADHD group was more likely to fall into lower clinically defined 25(OH)D categories, whereas the comparison group was more frequently in the ≥30 ng/mL category; this pattern remained directionally consistent after stratification by season of blood draw. In the whole study sample, SII was positively correlated with attention, impulsivity and oppositional scores, whereas serum 25(OH)D was inversely correlated with these symptom dimensions; SII was also inversely correlated with 25(OH)D. In the parsimonious model, the interaction term between standardised SII and standardised 25(OH)D was statistically significant (odds ratio (OR) = 3.06, 95% confidence interval (CI) 1.59–6.80, p = 0.002), and it remained significant in the broader sensitivity model and Firth penalised regression. In exploratory within-sample analysis, ROC curves yielded area under curves (AUCs) of 0.808 for SII, 0.781 for serum 25(OH)D and 0.836 for the parsimonious multivariable model.  Conclusions: Paediatric ADHD was associated with high inflammatory burden and low serum 25(OH)D status in this retrospective outpatient sample. The statistically significant interaction between SII and 25(OH)D suggests a stable pattern of effect modification on the logistic model scale. These findings support a potentially relevant immuno-nutritional pattern in ADHD, but prospective validation is needed before causal interpretation or clinical implementation can be considered.