Development of a Frailty Score for Older Patients with Myelodysplastic Syndromes and Acute Myeloid Leukemia.

Treatment options in older patients (pts) with MDS/AML range from best supportive care (BSC) to intensive chemotherapy/hematopoietic stem cell transplantation (IC/HCT), with low-dose chemotherapy or novel non-intensive agents (e.g. hypomethylating agents; HA) as alternatives. Due to frequent age-related physical and/or mental impairments, intensive treatment is not always feasible. As the basis for treatment decision-making is not well defined, the generation of comprehensive assessments of age-specific functional and quality of life (QOL)-aspects in addition to disease-specific risk factor definition therefore is urgently needed. Geriatric Assessment (GA) is expected to offer rational support in this process.

Since January 2004, we have prospectively evaluated the prognostic impact of GA on overall survival (OS) in 195 consecutive pts ≥60 years (yrs) with AML (n=132) or MDS (n=63) in three participating centers, receiving either BSC or HA+BSC or IC/HCT. Of the pts receiving non-intensive treatment, 50% had MDS. GA included eight instruments evaluating QOL, activities of daily living, depression, mental functioning, mobility, comorbidities and performance status (PS). In addition, disease- and patient-specific laboratory parameters were obtained.

Median age of pts was 71 yrs (range: 60-87 yrs). The primary treatment allocation was BSC in 47 pts (median age: 75 yrs); HA+BSC in 66 pts (74 yrs); IC/HCT in 75 pts (68 yrs). 62% of IC/HCT pts received a matched related/unrelated stem cell transplantation. Application of age-specific tests at the different study centers was readily feasible. The initial multidimensional GA was associated with treatment allocation, age, hematological and functional parameters and treatment outcome. Multivariate analyses revealed impairments in activities of daily living (ADL: Barthel Test, HR: 2.22) and fatigue (measured by EORTC QLQ-C30; HR: 1.68) as significant prognostic parameters for overall survival. Both risk factors were combined to construct a simple risk score for survival. Conducting a Cox regression model with established risk factors, a high risk frailty score in the entire pt population was associated with an elevated HR of 4.17 (p<0.0001), while adverse cytogenetics (AML), blasts >20% and comorbidities >1 proved to be independently associated with HRs of 2.491 (p=0.0001), 2.756 (p=0.0005) and 1.495 (p=0.1281). When this score was applied to pts receiving sole BSC or HA+BSC, highly significant differences in OS could be demonstrated, with p=0.0035 and p<0.0001, respectively.

Our data demonstrate that GA is a useful and objective tool in the in-depth evaluation process prior to treatment allocation in elderly patients with MDS/AML. A simple prognostic score based solely on ADL and fatigue to predict outcome of patients treated non-intensively has been established. Validation in independent cohorts appears warranted.

No relevant conflicts of interest to declare.