Development and validation of a risk prediction model for cerebral hyperperfusion syndrome after revascularization in adult moyamoya disease.

To identify risk factors for cerebral hyperperfusion syndrome (CHS) following revascularization surgery in adult moyamoya disease (MMD) patients and to develop and validate a corresponding risk prediction model. A systematic literature search was conducted in PubMed, Web of Science, the Cochrane Library, and CNKI for studies on CHS risk factors post-revascularization in adult MMD. Meta-analysis was performed using Review Manager 5.4. Significant risk factors from the meta-analysis were used to construct a logistic regression model. For clinical validation, 120 eligible patients from a tertiary neurosurgical center were enrolled (March–October 2025). The model's discrimination and calibration were evaluated. Fifteen studies involving 1,968 patients were included. The pooled CHS incidence was 23% (95% CI: 0.17–0.30), with significant heterogeneity across studies (I2 = 94%). Meta-analysis identified four independent risk factors: left-side surgery (OR = 4.08, 95% CI: 2.69–6.19, P < 0.001), advanced Suzuki stage (OR = 4.06, 95% CI: 2.02–8.17, P < 0.001), concomitant hypertension (OR = 5.38, 95% CI: 3.44–8.41, P < 0.001), and hemorrhagic onset type (OR = 3.20, 95% CI: 1.47–6.97, P = 0.003). The prediction model was: logit(P) = -1.046 + 1.406 × (surgical side) + 1.401 × (Suzuki stage) + 1.682 × (hypertension) + 1.163 × (onset type). In the validation cohort (CHS incidence 25.83%), the model showed an area under the curve (AUC) of 0.808 (95% CI: 0.724, 0.885). The optimal cutoff probability was 0.70, determined by maximizing Youden's index (0.529), at which the model demonstrated a sensitivity of 0.742 and a specificity of 0.787. The Hosmer–Lemeshow test indicated good calibration (χ2 = 6.037, P = 0.643). An evidence-based risk prediction model for CHS after revascularization in adult MMD was developed and validated, demonstrating good predictive performance. This model may assist in early risk stratification and personalized perioperative management.