Construction and validation of a prediction system for postherpetic neuralgia: based on meta-analysis

Abstract

Postherpetic neuralgia (PHN) is a chronic and intractable pain disorder. Existing PHN prediction systems have limited clinical applicability. A systematic search of the literature was conducted in PubMed, Web of Science, Embase and the Cochrane Library between 1950 and 2025, finally 44 studies were included (PROSPERO: CRD420251011300). Training (n = 417) and validation (n = 228) datasets were obtained from First Affiliated Hospital of Zhejiang Chinese Medical University and Third People's Hospital of Hangzhou. Based on the effect sizes from the meta-analysis, a linear regression model was fitted using the training set to develop the PHN Prediction Scoring Scale (PPSS), which was subsequently validated in the independent validation set. Seven key risk factors identified from the meta-analysis were selected for scoring,, including age (odds ratio(OR) (95% confidence interval): 1.66 (1.47, 1.88)), acute pain (OR: 2.61 (2.24, 3.05)), prodromal pain (OR: 1.79 (1.38, 2.31)), severe rash (OR: 2.64 (1.95, 3.56)), rash on the head(OR: 1.96 (1.8, 2.13)), diabetes (OR: 1.34 (1.24, 1.44)), immunosuppression (OR: 2.23 (1.95, 2.56)). Based on the above factors and weights, the PPSS were constructed. The results showed that on the validation set the area under curve of PPSS was 0.759 (0.693-0.824), accuracy was 0.684 (0.624-0.745), precision was 0.515 (0.417-0.612), recall was 0.693 (0.589-0.798), specificity was 0.680 (0.606-0.754). These findings suggest that the PPSS holds promise for generalization and application in clinical practice.