Geographical Disparities in Multiple Myeloma Survival in Three French Regions

Introduction: Within the past decade, new therapeutic agents for Multiple Myeloma (MM) have improved survival of patients included in clinical trials. The aim of this study was to measure and compare characteristics and survival of patients with MM in different French regions in the general population.

Methods: All incident cases with MM (ICD-O-3 codes, 9732/3 and) between January 1, 2008 and December 31, 2009 in the French population-based Hematological Malignancies registries (Calvados, Côte d'Or and Gironde) were included. The end point of follow-up was January 1, 2015. Specific clinical information was recorded from clinical charts. Incidence rates were calculated using estimations of population by age, sex, year and geographical area and rates were standardized using the world standard population (WSP). We focused our analysis only on MM cases considered as symptomatic, i.e. treated patients and/or those with at least one CRAB criteria (Calcium elevation, Renal insufficiency, Anemia, lytic bone lesions). In symptomatic MM, probabilities of Net Survival were estimated with the non-parametric Pohar-Perme estimator and age-standardized according to the International Cancer Survival Standards. To model effects of covariate on the Excess Mortality Rate (EMR), the flexible model of Nelson et al. was used using an adapted strategy.

Results: A total of 346 patients were diagnosed during 2008-2009 in the three registries. The WSP incidence rates were 2.8, 3.1 and 3.8 per 100 000 habitants/year in each of the three regions (not significantly different). We observed geographical variation about sex ratio (from 1 to 1.9) and according to age-specific incidence rates. Between the three regions, the median age at diagnosis ranged from 69 to 75 years old and only 4% to 10% of MM cases were considered as smoldering myeloma. In symptomatic MM (N = 321 cases), the proportion of treated patients was comparable between the three regions (87%). Significant differences as regards to geographical areas, were found for the ISS score with frequencies of score III ranging from 20% to 35%. Geographical differences were found in the age standardized net survival with 5-years probabilities of survival of 59%, 45% and 38% according to each area. In a multivariate analysis, the EMR was significantly affected by age (p<0.001), geographical area (p = 0.01), platelet count (p < 0.001), albumin (p = 0.014) and treatment (p < 0.001) with adjustment on covariables sex and ISS score.

Conclusion: This population-based study showed an area-related difference in MM survival in France mainly due to discrepancies in patients' characteristics at diagnosis. As the region effect influences survival independently of patients' characteristics, access to diagnosis in different regions may also be critical.