Identifying frailty in clinically fit patients diagnosed with hematological malignancies using a simple clinico-biological screening tool: The HEMA-4 study.

This study aims to develop and validate a simple score to estimate survival in the older population suffering from malignant hemopathies. We prospectively recruited 285 patients, aged ≥65 years, admitted to receive chemotherapy. At inclusion, a geriatric assessment was performed. Cox proportional hazards models were performed to assess correlations between vulnerabilities and one-year survival. We developed a frailty score, HEMA-4, based on the most powerful prognostic factors. It was externally confirmed with an independent cohort. In the development cohort, 206 patients were evaluable. Mean age was 76 years (range 65-90). The HEMA-4 score was created based on four independent predictive factors for survival: cognitive impairment (MMSE<27), comorbidities (≥2 on Charlson comorbidity index), CRP (≥10 mg/L) and low albumin level (<35 g/L). The population was stratified into three groups: good prognosis (score = 0-1, n = 141), intermediate prognosis (score = 2, n = 37) and poor prognosis (score = 3-4, n = 28). Their respective one-year survival was 74%, 51% (HR = 2.30; 95% CI =1.31-4.05; p < 0.01) and 36% (HR = 3.95; 95% CI =2.23-6.98; p < 0.01). In the validation cohort (n = 25), the one-year survival was 78% in the good prognosis group (n = 9) and 50% in the intermediate prognosis group (n = 6). The poor prognosis group had a median survival of four months in the development cohort and six months in the validation cohort (n = 10). The HEMA-4 score is a simple score that combines cognitive impairment, comorbidities, inflammation and low albumin level. Our data suggest that it predicts survival among older patients suffering from malignant hemopathies referred to receive chemotherapy regardless of their chronological age.

Informations de copyright

Copyright © 2021 Elsevier Ltd. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of competing interest None.