499P - Incidence and Relevance of Proteinuria in Bevacizumab (BV)-Treated Patients (PTS): Pooled Analysis from Randomized Controlled Trials (RCTS)

Abstract

ABSTRACT Background Proteinuria (PU) is a recognized adverse event with BV treatment (tx); however it is not known if BV-related PU is associated with clinical outcomes. This analysis was undertaken to define the relationship between PU with BV tx and sequelae, including changes in kidney function. Methods A pooled safety database, comprising 22 phase 2/3 RCTs of BV across tumor types, was used to characterize PU events. Analysis pools were created based on data availability in individual studies. Raw and time-adjusted PU rates and data on the association between PU and arterial thromboembolic events (TEs), venous TEs, and infection were derived from Pool 1 (17 RCTs; n = 14,548). Data from Pool 2 (8 RCTs; n = 9,158) were used to estimate the association between lab-reported PU and changes in kidney function based on serum creatinine (sCr) levels. Severity of kidney function change was categorized using the RIFLE classification for acute kidney injury (AKI). Potential predictors of PU were also assessed. Results In Pool 1, the incidence rate of any-grade (gr) PU was 8.2% (733/8917) and 4.6% (257/5631) in BV and control pts, respectively; gr ≥3 PU occurred in 1.4% and 0.2%, respectively. Pts developing PU had a numerically increased rate of any-gr infection irrespective of tx (driven mostly by urinary tract infection); however, no association was found between PU and TEs (arterial or venous). In Pool 2, PU gr and tx appeared to slightly shift the rate of post-baseline AKI (Table). Evaluated risk factors (age, sex, history of hypertension or diabetes) were not found to be significantly predictive for gr ≥3 PU. Conclusions The use of laboratory events and pooled data confirm a modest increase in PU with BV tx. The development of PU in BV-treated pts was associated with a modest increase in the risk of infection, but not TEs, and a trend toward a decrease in kidney function. Whether these associations are causal cannot be determined by this analysis. Tx group Worst post-BL gr of PU a Pts without PU at BL, n Worst post-BL kidney function (RIFLE classification), n (%) Normal Risk b Injury c Failure d BV (n = 5098) 0 2960 2277 (76.9) 585 (19.8) 66 (2.2) 25 (0.8) 1 1381 1021 (73.9) 307 (22.2) 31 (2.2) 20 (1.4) 2 562 391 (69.6) 137 (24.4) 18 (3.2) 14 (2.5) 3 59 34 (57.6) 19 (32.2) 4 (6.8) 2 (3.4) No BV (n = 3186) 0 2144 1736 (81.0) 349 (16.3) 39 (1.8) 18 (0.8) 1 769 618 (80.4) 123 (16.0) 19 (2.5) 9 (1.2) 2 166 124 (74.7) 32 (19.3) 7 (4.2) 3 (1.8) 3 2 1 (50.0) 1 (50.0) — — Note: missing values for each row are not shown. BL, baseline. a CTCAE grade. No patients had gr 4 PU. b Increased sCR ×1.5 or estimated creatinine clearance (eCrCl) decrease >25%. c Increased sCr ×2 or eCrCl decrease >50%. d Increased sCr ×3, eCrCl decrease >75% or sCr ≥4 mg/dL. Disclosure B. McCall: Dr. McCall is an employee of and holds stock options in Genentech, Inc. N.F. Li: Dr. Li is an employee of and holds stock options in Genentech, Inc. L. Chu: Ms. Chu is an employee of and holds stock options in Genentech, Inc. P. Werner: Dr. Werner is an employee of F. Hoffmann-La Roche Ltd. A. Das: Dr. Das is an employee of and holds stock options in Genentech, Inc. R.J. Glassock: Dr. Glassock is a consultant to Genentech. All other authors have declared no conflicts of interest.