Establishing A Predictive Model For The Development Of Infectious Calculus In Patients With Urinary Calculi And Its Analysis

ObjectivesTo analyze the potential preoperative risk factors that affect the development of infectious calculus following stone lithotripsy for urinary calculi,and to develop a nomogram for predicting the probability of postoperative infectious calculus according to the identified independent risk factors.MethodsWe retrospectively analyzed the clinical data from consecutive 477 cases of urinary calculi diagnosed with color Doppler ultrasonography,KUB/IVU or CT examination,with subsequent treated by stone lithotripsy(percutaneous nephrolithotomy,ureteroscopic lithotripsy,laparoscopy ureterolithotomy,open operation,etc)between September 2016 and November 2018 in our clinical department.The calculi specimens were investigated by Fourier transform infrared spectroscope(FT-IR),The patients were categorised into two groups according to whether the patients developed infectious calculus: Infectious calculus group and No infectious calculus group.Univariate and multivariate logistic regression analysis were done to determine the predictors of infectious calculus.R software was adopted to construct nomogram of the development of infectious calculus prediction model,the concordance index(C-index)to evaluate the accuracy of model,the Bootstrap method used in internal validation,and the calibration curve drawn.ResultsThe stone lithotripsy procedures of the 477 cases were performed successfully,and the patients were categorised into Infectious calculus group(n=309)and No infectious calculus group(n=168),the incidence rate of infectious calculus was identified in 64.78%.Univariate Logistic regression analysis revealed the urine p H(OR=1.995,95%CI 1.556-2.558,P < 0.001),uric acid(OR=0.996,95%CI 0.994-0.998,P<0.001),stones distribution(OR=1.573,95%CI 1.047-2.363,P=0.029),gender(OR=1.869,95%CI 1.259-2.774,P=0.002)and fever(OR=2.391,95%CI 1.232-4.639,P=0.010)were significantly differences among two groups;Multivariate Logistic regression analysis revealed the urine p H(OR=1.927,95%CI,1.484-2.502,P<0.001),uric acid(OR=0.997,95%CI,0.995-1.000,P=0.031),stones distribution(OR=1.745,95%CI,1.126-2.703,P=0.013)and fever(OR=2.37,95%CI :1.184-4.744,P=0.015)were the independent predictors of the development of infectious calculus.The nomogram based on these results was well fitted to predict a probability,and the concordance index(C-index)was 0.703(95%CI,0.655 to 0.751)in the nomogram model sample and the mean absolute error in the validation sample is 0.013,the value clinical application of which was well.ConclusionIn patients with urinary calculi,higher urine p H,lower uric acid,urethral or ureteral calculi and fever are identified as independent risk factors to affect the development of infectious calculus.A nomogram based on these preoperative clinical independent risk factors for infectious calculus has good discrimination and accuracy which could be helpful for screening the patients and achieving digitalized and individual predictions.