Search

Background: In 2015, the biosimilar filgrastim EP2006 became the first biosimilar approved by the US Food and Drug Administration for commercial use in the United States, marketed as Zarxio® (Sandoz). This phase III randomised, double-blind registration study in patients with breast cancer receiving (neo)adjuvant myelosuppressive chemotherapy (TAC; docetaxelþdoxorubicinþcyclophosphamide) compares reference filgrastim, Neupogen® (Amgen), with two groups receiving alternating treatment with reference and biosimilar every other cycle. Patients and methods: A total of 218 patients receiving 5 mg/kg/day filgrastim over six chemotherapy cycles were randomised 1: 1: 1: 1 into four arms. Two arms received only one product, biosimilar or reference (unswitched), and two arms (switched) received alternating treatments every other cycle (biosimilar then reference or vice versa over six cycles). Since the switch occurred from cycle 2 onwards, this analysis compared pooled switched groups to the unswitched reference group for efficacy during cycles 2-6. Safety was also assessed. Non-inferiority in febrile neutropenia (FN) rates between groups for cycles 2-6 was shown if 95% were within a pre-defined margin of -15%. Results: A total of 109 patients switched treatment, and 52 patients received reference in all cycles. Baseline characteristics were similar between groups. The incidence of FN was 0% (reference) versus 3.4% (n=3, switched) across cycles 2-6, with a difference of -3.4% (95% confidence interval: -9.65% to 4.96%), showing non-inferiority. Infections occurred in 9.3% (switched) versus 9.9% (reference). Hospitalisation due to FN was low (one patient in cycle 6; switched). Adverse events related to filgrastim were reported in 42.1% (switched) versus 39.2% (reference) (all cycles). Musculoskeletal/connective tissue disorders related to filgrastim occurred in 35.5% (switched) versus 39.2% (reference) (all cycles), including bone pain (30.8% versus 33.3%). No neutralising antibodies were detected. Conclusions: There were no clinically meaningful results regarding efficacy, safety or immunogenicity when switching from reference to biosimilar filgrastim/EP2006, or vice versa. [ABSTRACT FROM AUTHOR]

• Genotype by Environment by Management (GxExM) interactions hampers crop improvement. • Crop simulation revealed large/variable yield gaps (∼70 %). • Filling 80 % of yield gap would increase production by 8 M tons. • Modelling output and biophysical traits were clustered (HPU). • Grouping by HPU reduced GxExM interactions and will help refine breeding targets. The high degree of Genotype by Environment by Management (GxExM) interactions is a serious challenge for production and crop improvement efforts. This challenge is especially true for a crop like groundnut that is often grown as a rainfed crop in diverse environments and management, leading to considerable production fluctuations among regions and seasons. Developing a means to characterize the drivers of variable yield and to identify region specific breeding objectives were the main motivations for this research, using groundnut production in India, as a case study for rainfed crops. Historically, five groundnut production areas have been considered by Indian crop improvement programs. Our objectives were to assess the relevance of this zonation system and possibly to re-define production areas with a higher degree of similarities into homogeneous production units (HPUs). Towards this, we used yield gap analysis and the geo-biophysical characters of the production region to understand and deal with GxExM interactions. Weather and soil data, crop parameters, and management information data were collected and groundnut production was simulated at the district scale over 30 consecutive years. Consequently, the geographic distribution of the potential yields and the yield gaps were first estimated to understand the main production limitations in a given region. Large and variable yield gaps (with a mean of ∼70 %) were observed and results revealed a readily exploitable production gap (∼ 8 M tons), which might be bridged by following recommended agronomic practices. Water deficit limited the yield potential by an average of 40 %, although with large variability among districts. However, large and variable yield gaps remained. To resolve the unexplained variation, principal component and cluster analysis of agronomic model output together with geo-biophysical indicators for each district were carried out. This resulted in seven HPUs, having well-defined production-limiting constraints. Grouping by HPU greatly reduced variance in actual and simulated yields, as compared to grouping across all groundnut production zones in India. The HPU based approach delimited precise geographic regions within which HPU-specific GxM products could be designed by crop improvement programs to boost productivity. [ABSTRACT FROM AUTHOR]