Maximilian Burger, Hans-Martin Fritsche, Michael Rink, Armin Pycha, Margit Fisch, Thomas Seisen, Matthias May, Philipp Nuhn, Michael Gierth, Felix K.-H. Chun, Luis A. Kluth, Malte W. Vetterlein, Oliver W. Hakenberg, Atiqullah Aziz, Joachim Noldus, Philipp Gild, Chris Protzel, Nicolas von Landenberg, Florian Roghmann
To evaluate the effect of perioperative blood transfusion (PBT) on recurrence-free survival (RFS), overall survival (OS), cancer-specific mortality (CSM), and other-cause mortality (OCM) in patients undergoing radical cystectomy (RC), utilizing a contemporary European multicenter cohort.
Patients and Methods
The PROspective MulticEnTer RadIcal Cystectomy Series (PROMETRICS) includes data on 679 patients who underwent RC at 18 European tertiary care centers in 2011. The association between PBT and oncological survival outcomes was assessed using Kaplan-Meier, Cox regression, and competing-risks analyses. Imbalances in clinicopathological features between patients receiving PBT vs. not were mitigated using conventional multivariable adjusting as well as inverse probability of treatment weighting (IPTW).
Overall, 611 patients had complete information on PBT, and 315 (51.6%) received PBT. The two groups (PBT vs. no PBT) differed significantly with respect to most clinicopathological features including perioperative blood loss (median: 1000ml; IQR: 600-1500ml vs. median: 500ml; IQR: 400-800ml; P<0.001).
Independent predictors of receipt of PBT in multivariable logistic regression analysis were female gender (odds ratio (OR)=5.05; 95% confidence interval (CI)=[2.62-9.71]; P<0.001), body mass index (OR=0.91; 95% CI=[0.87-0.95]; P<0.001), type of urinary diversion (OR=0.38; 95% CI=[0.18-0.82]; P=0.013), blood loss (OR=1.32; 95% CI=[1.23-1.40]; P<0.001), neoadjuvant chemotherapy (OR=2.62; 95% CI=[1.37-5.00]; P=0.004), and ≥pT3 tumors (OR=1.59; 95% CI=[1.02-2.48]; P=0.041). In 531 patients with complete data on survival outcomes, unweighted and unadjusted survival analyses revealed worse OS, CSM, and OCM for patients receiving PBT (P<0.001, P=0.017, and P=0.001, respectively). After IPTW-adjustment, those differences no longer held true. PBT was not associated with RFS (HR=0.92; 95% CI=[0.53-1.58]; P=0.8), OS (HR=1.06; 95% CI=[0.55-2.05]; P=0.9), CSM (sub-HR=1.09; 95% CI=[0.62-1.92]; P=0.8) and OCM (sub-HR=1.00; 95% CI=[0.26-3.85]; P>0.9) in IPTW-adjusted Cox regression and competing-risks analyses. The same held true in conventional multivariable Cox and competing-risks analyses, where PBT could not be confirmed as a predictor of any given endpoint (all P values >0.05).
Our findings could not demonstrate an adverse effect of PBT on oncological outcomes after adjusting for baseline differences in patient characteristics.
This article is protected by copyright. All rights reserved.