1817MO - Long-term patient reported outcomes (PRO) and hematologic toxicity among patients (pts) who received granulocyte-colony stimulating factors (G-CSF) during chemotherapy (CT) for early breast cancer (EBC)

Background

Short-term side-effects of G-CSF were reported in EBC, including bone pain and transitory leukocytosis and thrombocytopenia. Few data exist on the long-term effects of G-CSF on quality of life and hematological toxicity.

Methods

EBC pts (stage I-III) treated with CT were identified from the prospective, longitudinal, multicenter CANTO cohort (NCT01993498). PRO data (EORTC QLQ-C30/BR23) and blood cell counts were obtained at diagnosis (baseline) and at year (Y)1, Y2, and Y4 after diagnosis. Our primary outcome was QLQ-C30 pain score. Exploratory outcomes included all QLQ domains, C30 summary score (Husson O, 2020), leucocyte and platelet count. The independent variable was G-CSF use. Multivariable mixed models examined associations of G-CSF use with outcomes.

Results

Among 2926 pts included from 2012-15, 49% (n=1427) used G-CSF and 51% (n=1499) did not. Baseline mean age was 53 years (SD=11), >80% pts received non dose-dense anthracyclines + taxanes regimens. Pts receiving G-CSF were older, had higher stage BC and worse baseline PRO (Table). Over time, pain score was not statistically different between groups (adjusted overall mean difference [adjMD] vs no-use [95% CI]: +1.32 [-0.28 to +2.92]). Consistently, remaining PRO were substantially similar between groups, including summary score (adjMD vs no-use [95% CI]: -1.04 [-2.01 to -0.08]). PRO had a differential longitudinal evolution between groups, with slightly worse scores at Y4 among pts receiving G-CSF, of trivial clinical significance (Table). No major differences were observed in leucocyte or platelet count (adjMD vs no-use [95% CI]: -280/mm³ [-393 to -167/mm³] and -133/mm³ [-4326 to +4060/mm³], respectively). Table: 1817MO

Mean differences* (95% CI) over time in QLQ-C30 scores among pts treated with G-CSF vs not

*all estimates are obtained from a mixed model including G-CSF use, time, G-CSF use by time interaction, socio-demographics, comorbidities, tumor and treatment-related variables; ^ˆdifferences <6 points are considered trivial (Cocks K, 2011)

Conclusions

This is the first large study assessing long-term and late associations of G-CSF use in EBC. G-CSF seemed overall well-tolerated, with no major clinical impact on PRO and hematologic toxicity.

Clinical trial identification

NCT01993498.

Editorial acknowledgement

Legal entity responsible for the study

UNICANCER.

Funding

This research was supported by the French Government under the “Investment for the Future” program managed by the National Research Agency (ANR), grant n° ANR-10-COHO-0004.

Disclosure

I. Vaz-Luis: Speaker Bureau/Expert testimony: Novartis; Amgen; AstraZeneca; Kephren. A. Di Meglio: Honoraria (self): Thermo Fisher. All other authors have declared no conflicts of interest.