Infections Associated with Bendamustine – a Retrospective Multicenter Study

Background and aims:

Bendamustine is a chemotherapeutic drug with structural similarities to both alkylating agents and purineanalogues. We aimed to determine the incidence and characteristics of infectious-related adverse events in the "real life" setting and to identify predictors for infection.

Methods:

A retrospective cohort study of patients with lymphoproliferative malignancies treated with any bendamustinecontaining regimen between 2010 and 2013.

Demographic, clinical, and laboratory data including types of infection, therapy and outcome were collected from patients’ files until the latest follow up available and for at least 6 months after bendamustineadministration. Primary outcome was any infection. Risk factors of patients who developed infection and those who did not were compared. The Kaplan Meier Model, using log rank analysis, was used to compare time-to-infection or death between study groups. The Cox Proportional Hazards model was used to analyze multivariate effects of risk.

Results:

183 patients from 3 centers in Israel (Beilinson, Sheba and HaEmek) were included. Hematological disorders included: non Hodgkin lymphoma in 72% of patients (30% follicular lymphoma, 11% mantle cell lymphoma, 7% marginal zone lymphoma, 9% diffuse large B cell lymphoma, 4% lymphoplasmacytic lymphoma, and 11% other) , chronic lymphocytic leukemia in 20% and multiple myeloma in 8% of patients. The most common regimen was bendamustine-rituximab (73% of pts). 21% received bendamustine alone. Bendamustine was administered either at the dosage of 70 or 90 mg/sqm iv on days 1,2 (mean dose - 77.8 +/-13.04), and rituximab was administered at a dose of 375 mg/sqm iv on day 1. Therapy was administered every 4 weeks up to 6 courses. The mean number of cycles administered was 4.54+/-1.78. Mean age at first cycle was 68 +/-10.8. 60% of patients were males, 86% had advanced disease, 60.6% had bone marrow involvement,60% received bendamustine as third line treatment and above, 65% received growth factors, 42% received prior rituximabmaintenance. Antibiotic prophylaxis was administered to only 2 patients.

86 patients (47%) developed at least one infection. The mean number of infections per patient was 2.26+/-1.89. 38 patients with infection died during follow up (20% of all patients). 76 patients (41.5%) developed a bacterial infection, 8 patients (4.4.%) developed a fungal infection and 20 patients (11%) developed a viral infection (Table). There was no difference in the timing from the first cycle of bendamustine to infection development with a median of 154 days to development of bacterial infection, 154 days for development of viral infection and 348 days for development of fungal infections (p=0.37). 38% of patients with infection developed grade 3-4 leukopenia, 45% developed grade 3-4 neutropenia, and 65% developed grade 3-4 lymphopenia.

Factors significantly associated with time to infection or death on univariate analysis were: advanced treatment line (3^rd line treatment and above vs. 1^st -2^ndline treatment), no growth factor administration, and no prior rituximab maintenance. On multivariable analysis, the only factor that remained an independent predictor for infection or death was advanced treatment line with HR of 1.18 (1.053-1.323), p = 0.004.

Conclusions:

The infection and sepsis rate associated with bendamustineadministration in the "real life" setting is 47% and 11%, respectively, consistent with previous data from randomized controlled trials. Most infections are bacterial. The major predictor for the development of infection is treatment line. This issue should be taken into consideration in heavily pretreated patients.