Risk factors for kidney stone recurrence and early prediction models: a systematic review and meta-analysis.

Background: Recurrence of kidney stone disease (KSD) remains a significant challenge in current clinical practice. There is a paucity of systematic evidence of independent risk factors and performance of early prediction models for KSD recurrence. Hence, this paper intended to systematically evaluate the independent risk factors for KSD recurrence and the feasibility of early prediction. Methods: Available original studies about risk factors and prediction models for postoperative KSD recurrence were systematically searched for in PubMed, Web of Science, Embase, and Cochrane Library databases until July 2024. The risk of bias was appraised via the Newcastle-Ottawa Scale. Results: Eighteen studies on risk factors for KSD recurrence and 12 studies on prediction models were included. Younger age [odds ratio (OR) =1.428, 95% confidence interval (CI): 1.262-1.616], diabetes mellitus (OR =1.428, 95% CI: 1.262-1.616), hypertension (OR =1.428, 95% CI: 1.262-1.616), and family history of KSD (OR =1.428, 95% CI: 1.262-1.616) were independent risk factors for KSD recurrence. Current prediction models were mainly constructed based on these common clinical characteristics. The area under the curve for recurrence prediction was 0.71 (95% CI: 0.64-0.78) in the training set and 0.69 (95% CI: 0.60-0.79) in the validation set. Conclusions: There are independent risk factors for postoperative KSD recurrence. However, the predictive performance of the prediction models constructed based on these factors faces challenges. More efficient predictive factors should be incorporated in constructing models with higher predictive accuracy for postoperative KSD recurrence.