Can smoking history, peripheral inflammation, and nutritional status discriminate Parkinson's disease? Development and validation of a clinically accessible nomogram.

Research overview

The text below is the source article’s original abstract under a verified reuse license. cc by

BackgroundParkinson's disease (PD) is a major public health challenge in China, with rising prevalence and limited early diagnostic tools. Systemic inflammation and nutritional status are increasingly recognized as key modulators of PD pathogenesis, but integrated predictive models remain lacking.ObjectiveTo develop and internally validate a simple, clinically accessible nomogram for individualized probability estimation of currently having PD using routinely available demographic, lifestyle, and laboratory parameters.MethodsThis retrospective study enrolled 2,221 participants (1,135 PD patients and 1,086 healthy controls) from a single center between December 2023 and December 2025. The cohort was randomly split into a training set (n = 1,555) and an internal validation set (n = 666). Least absolute shrinkage and selection operator (LASSO) regression was used for predictor selection from 25 candidate variables. Multivariate logistic regression was performed to identify independent associated factors and construct a nomogram. Model performance was assessed by discrimination (area under the receiver operating characteristic curve, AUC), calibration (calibration curve with Brier score), and clinical utility (decision curve analysis).ResultsLASSO regression identified three core associated factors: smoking history, Systemic Immune-Inflammation Index (SII100), and Prognostic Nutritional Index (PNI). Multivariate logistic regression confirmed all three as independent associated factors: smoking history (OR = 0.66, 95% CI: 0.52-0.85, p = 0.001), SII100 (OR = 1.30, 95% CI: 1.25-1.36, p p ConclusionThe nomogram developed in this study may help identify individuals with higher probability of PD diagnosis in a cross-sectional setting using only three routine variables (smoking history, SII100, and PNI). The model maintained moderate discrimination (AUC ~ 0.80) and acceptable but imperfect calibration (Brier scores: 0.185 in training, 0.188 in validation), though calibration in the training set showed some deviation from the ideal line. Decision curve analysis supported its positive net benefit in real-world clinical decision-making. As a promising tool for cross-sectional diagnostic classification, it requires external validation to confirm its generalizability in large-scale populations.

Related topics