Development of a Nomogram Predicting the Infection Stones in Kidney for Better Clinical Management: A Retrospective Study

Abstract

Purpose:

To establish the first comprehensive nomogram for prediction of infection stones before treatment for better perioperative treatment and postoperative prevention of infection stones.

Methods:

A total number of 461 patients with kidney stones who underwent mini-percutaneous nephrolithotomy and flexible ureteroscopy between January 2019 and March 2021 were retrospectively analyzed. Univariable analysis and multivariable logistic regression analysis were conducted to identify the predictors for infection stones. Furthermore, the nomogram was established as a predicted model for infection stones.

Results:

Among 461 patients with infrared spectroscopy stone analysis, 100 (21.70%) had infection stones and 361 (78.31%) had noninfection stones. Multivariate logistic regression analysis indicated that female (odds ratio [OR] 2.816, 95% confidence interval [CI] 1.148–6.909, p = 0.024), recurrent kidney stones (OR 8.263, 95% CI 2.295–29.745, p = 0.001), stone burden (OR 6.872, 95% CI 2.973–15.885, p < 0.001), HU (OR 15.208, 95% CI 6.635–34.860, p < 0.001), positive preoperative bladder urine culture (PBUC; OR 4.899, 95% CI 1.911–12.560, p = 0.001), positive urine leukocyte esterase (ULE; OR 3.144, 95% CI 1.114–8.870, p = 0.030), urine pH (OR 2.692, 95% CI 1.573–4.608, p < 0.001), and positive urine turbidity (OR 3.295, 95% CI 1.207–8.998, p = 0.020) were predictors for infection stone.

Conclusions:

For patients with kidney stones, female, recurrent kidney stones, stone burden (>601 mm²), HU (750–1000), positive PBUC, positive ULE, urine pH, and positive urine turbidity were predictors for infection stone. We established the first comprehensive model for identifying infection stones in vivo, which is extremely useful for the management of infection stones.