Preoperative blood typing and screening for benign prostatic hyperplasia surgery

A retrospective analysis and real-world validation of a machine learning model for predicting intraoperative and immediate postoperative blood transfusion

Abstract

Introduction: Routine preoperative blood type and screen (T&S) testing is commonly performed prior to benign prostatic hyperplasia (BPH) surgery, despite the low transfusion risk associated with endoscopic procedures such as transurethral resection of the prostate (TURP) and holmium laser enucleation of the prostate (HoLEP). In contrast, open simple prostatectomy (OP) remains a high-bleeding-risk procedure. Nevertheless, current guidelines do not provide clear recommendations regarding selective T&S use. This study aimed to identify objective predictors of perioperative transfusion, determine evidence-based cutoffs, and prospectively validate a risk-based testing protocol.

Methods: We analyzed 1232 consecutive BPH surgeries (2017–2022) performed at a tertiary center: 173 OP, 740 TURP, and 319 HoLEP. Demographic, comorbidity, and perioperative data were retrieved and cross-referenced with transfusion records, categorized as intraoperative or within 24 hours postoperatively. First, receiver operating characteristic (ROC) analysis was used to define optimal transfusion risk thresholds for each variable. Second, all predictors were entered into a logistic regression model, achieving an area under the curve (AUC) of 0.81, followed by a machine learning gradient boosting (XGBoost) algorithm to address the low transfusion event rate. Third, based on six final risk factors, a multidisciplinary team (urology, anesthesiology, hematology) developed a selective T&S protocol, which was prospectively implemented in 309 patients undergoing BPH surgery in 2023.

Results: The final XGBoost model identified age ≥80 years, hemoglobin ≤11 g/dL, platelets ≤100 × 10⁹/L, international normalized ratio (INR) ≥1.3, prostate volume ≥100 mL, and Charlson Comorbidity Index >4 as high-risk criteria. In the retrospective cohort, transfusions occurred intraoperatively in 4.9% (OP 24%, TURP 1.9%, HoLEP 0.9%) and postoperatively in 8% (OP), 1.1% (TURP), and 1.5% (HoLEP). Due to the high transfusion rate observed in OP, preoperative T&S testing was maintained for open procedures, while omission was applied only to low-risk endoscopic surgeries. Among the 309 validation patients, intraoperative transfusions occurred in 0% of TURPs and 2.3% (n=5) of HoLEPs. Of these, three were correctly screened, one was a protocol deviation, and one had an undiagnosed coagulopathy. Overall, 35% of patients met none of the risk criteria and safely underwent surgery without T&S.

Conclusions: Our findings support shifting away from universal preoperative T&S testing toward a protocol emphasizing individualized risk. Based on our findings, routine preoperative T&S testing can be safely omitted in low-risk patients undergoing TURP or HoLEP, while it should be maintained for open procedures and for patients with predefined high-risk features. Future studies may further validate these criteria and the model.