Abstract
This study aimed to analyze the infection risk factors for transurethral resection of the prostate (TURP) and establish predictive models to help make personalized treatment plans. Our study was designed one-center and retrospectively enrolled 1169 benign prostatic hyperplasia (BPH) patients. Risk factors were explored for postoperative infection. A TURP-postoperative infection (TURP-PI) model with infection prediction values was created. The improved-TURP-PI (I-TURP-PI) model, including extra new factors (pathogens species), was also built to see whether it could optimize the prediction abilities. At last, we developed a nomogram for better clinical application. Operation time, preoperative indwelling urinary catheter (PIUC), and positive preoperative urine culture were independent risk factors (all P < 0.05). Interestingly, pathogens species in pre-surgery urine (P(Enterococcus faecium) = 0.014, P(Pseudomonas aeruginosa) = 0.086) were also independent risk factors. Patients with positive Enterococcus faecium (37.50%) were most likely to have postoperative infection. We built two models with AUC(TURP-PI) = 0.709 (95% CI 0.656-0.763) and AUC(I-TURP-PI) = 0.705 (95% CI 0.650-0.760). The nomogram could help improve the prediction ability. To our knowledge, our study is the first to use pathogen species in urine before surgery as risk factors for infection prediction after TURP. TURP-PI and I-TURP-PI models have essential roles in predicting patients' postoperative infections and in better postoperative antibiotic decision-making.