Leukemia https://doi.org/10.1038/s41375-019-0666-7 ARTICLE Myelodysplastic syndrome An MDS-specific frailty index based on cumulative deficits adds independent prognostic information to clinical prognostic scoring R. Starkman 1 ● S. Alibhai 2 ● R. A. Wells 1 ● M. Geddes 3 ● N. Zhu 4 ● M. M. Keating 5 ● B. Leber 6 ● L. Chodirker 1 ● M. Sabloff 7 ● G. Christou 7 ● H. A. Leitch 8 ● E. St-Hilaire 9 ● N. Finn 9 ● A. Shamy 10 ● K. Yee 11 ● J. Storring 12 ● T. Nevill 13 ● R. Delage 14 ● M. Elemary 15 ● V. Banerji 16 ● M. Lenis 17 ● A. Kirubananthaan 17 ● A. Mamedov 17 ● L. Zhang 1 ● K. Rockwood 18 ● R. Buckstein 1 Received: 31 July 2019 / Revised: 30 October 2019 / Accepted: 17 November 2019 © The Author(s), under exclusive licence to Springer Nature Limited 2019 Abstract The frailty index (FI) is based on the principle that the more deficits an individual has, the greater their risk of adverse outcomes. It is expressed as a ratio of the number of deficits present to the total number of deficits considered. We developed an MDS-specific FI using a prospective MDS registry and assessed its ability to add prognostic power to conventional prognostic scores in MDS. The 42 deficits included in this FI included measurements of physical performance, comorbidities, laboratory values, instrumental activities of daily living, quality of life and performance status. Of 644 patients, 440 were eligible for FI calculation. The median FI score was 0.25 (range 0.05–0.67), correlated with age and IPSS/ IPSS-R risk scores and discriminated overall survival. With a follow-up of 20 months, survival was 27 months (95% CI 24–30.4). By multivariate analysis, age >70, FI, transfusion dependence, and IPSS were significant covariates associated with OS. The incremental discrimination improvement of the frailty index was 37%. We derived a prognostic score with five risk groups and distinct survivals ranging from 7.4 months to not yet reached. If externally validated, the MDS-FI could be used as a tool to refine the risk stratification of current clinical prognostication models. Introduction Determining optimal treatment for patients with myelodys- plastic syndromes (MDS) is challenging due to the hetero- geneity of their underlying health, largely reflecting comorbidities associated with advanced age [1]. Although some older adults may benefit from intensive therapies, as a group they tend to experience considerable treatment-related morbidity, are more prone to relapse, and have inferior sur- vival [2]. A number of prognostic scoring systems have been developed to assist in clinical decision making in MDS patients, but they are based on disease-related characteristics. These systems include the 1997 International Prognostic Scoring System (IPSS) [3] and the 2012 revision of the IPSS (IPSS-R) that incorporates cytogenetics, bone marrow blast percentages, and number or degree of cytopenias to determine a risk score [4]. An international initiative is underway to incorporate genomic data into existing prognostic scores but is not yet available [5, 6]. Existing scoring systems are based on the biology of the disease and fail to take into account patient characteristics other than age that are integral to clinical decision-making in older adults. Life expectancy and treatment tolerance may vary substantially, even among patients of similar ages, due to the heterogeneity of individual characteristics which are not measured by these scoring systems [2]. This hetero- geneity in risk among people of the same age is referred to as frailty. Frailty may also be defined clinically, as an age- related, multidimensional syndrome that gives rise to increased vulnerability [7]. Clinical frailty is common in older patients with cancer; as the degree of frailty increases, so too does the risk of mortality, postoperative complica- tions, and chemotherapy intolerance [8]. We previously demonstrated that the 9-point Clinical Frailty Scale (CFS) and the Charlson comorbidity index (CCI) were indepen- dently associated with overall survival (OS) in 445 MDS * R. Buckstein rena.buckstein@sunnybrook.ca Extended author information available on the last page of the article Supplementary information The online version of this article (https:// doi.org/10.1038/s41375-019-0666-7) contains supplementary material, which is available to authorized users. 1234567890();,: 1234567890();,: