A 42-Item Frailty Index Improves Survival Prediction in Myelodysplastic Syndromes

In a recent Leukemia paper, researchers reported on a new 42-variable frailty index (FI) for patients with myelodysplastic syndromes (MDS), which appears to improve on current prognostic systems and their ability to predict survival outcomes.

Existing scoring systems for MDS risk stratification such as the revised International Prognostic Scoring System (IPSS-R) are based on the biology of the disease and fail to take into account patient characteristics that affect outcomes and treatment tolerability, the authors explained. “If a clinician applies the risk score that we developed, which is made up of both disease and patient characteristics, one gets a more refined prognosis,” said Rena Buckstein, MD, director of the National MDS Registry and staff hematologist at Sunnybrook Odette Cancer Center in Ontario. “I plan to use this in my clinic to better prognosticate and tell patients what to expect of their disease and their survival.”

To construct the MDS-specific FI, Dr. Buckstein and researchers obtained data from the ongoing prospective Canadian MDS registry to identify baseline laboratory- and patient-related factors and evaluate their prognostic power. The registry included 644 patients with MDS, chronic myelomonocytic leukemia (CMML), and low blast count (<30% blasts) acute myeloid leukemia.

At time of enrollment, investigators calculated patients’ risk scores using the IPSS and IPSS-R and measured performance status, frailty, quality of life, disability, and physical performance using the following tests and assessment tools:

  • physical performance tests: grip strength test, 4-minute walk test, and timed 10-times chair stands test
  • comorbidities: Charlson Comorbidity Index
  • frailty: 9-point Clinical Frailty Scale and Lawton-Brody instrumental activity of daily living scale
  • quality of life: EQ-5D-3L and European Organisation for Research and Treatment of Cancer Quality-of-life QLQ-C30
  • fatigue: Global Fatigue Scale

A total of 58 variables met criteria for inclusion in the FI, meaning that each variable was associated with health status, generally increased in prevalence with age, included a range of systems, and was consistent from one iteration to the next when used serially. Of this group of 58, 42 were selected for the final FI.

The final FI included body mass index, congestive heart failure, dementia, diabetes, ability to shop or use a phone, ability to handle finances, hemoglobin levels, bilirubin levels, and grip strength, among others. Notably, all EORTC QLQ-C30 scale scores were excluded from the FI “because they require extensive calculation that is not feasible in a clinical setting,” the authors wrote.

A total of 440 patients from the registry were eligible for inclusion in the FI calculation. FI scores differed significantly by patients’ IPSS and IPSS-R scores, as well as age, but provided additional discriminatory power beyond IPSS and IPSS-R scores. Patients who were transfusion-dependent had a significantly higher mean FI, compared with those who were transfusion-independent.

The multivariable analysis revealed four co-variates that were significantly associated with overall survival:

  • age >70 (hazard ratio [HR]=1.49, 95% CI 1.12-1.98; p<0.01)
  • MDS-FI (p<0.0001 for all FI categories)
  • transfusion dependence (HR=1.34; 95% CI 1.01-1.79; p=0.04)
  • IPSS (p<0.0001 for all IPSS risk groups)

A limitation of the study includes the lack of adjustment for the timing of quality of life and physical performance testing, particularly in relation to transfusions and transfusion intensity, the authors noted.

To further evaluate the prognostic impact of the MDS-specific FI, Dr. Buckstein added that it should be compared with other measures of frailty currently used in clinical practice. However, these measures may not be helpful for risk stratification, making it difficult to compare these tools with the FI developed in this study. “Most other scores don’t include biologic function, like renal function and hepatic function, or other abnormalities,” she said. “Our FI is a more intuitive measure of physiologic vulnerability.” In addition to the 42-item MDS-FI, Dr. Buckstein hopes some other form of FI that stratifies risk will soon be tested in clinical trials.

For example, she would like to see this MDS-FI, or a modified version with fewer elements, deployed in future clinical trials for the purposes of risk stratification. “When we look at patients’ responses to a new treatment in the context of a randomized trial, we can then look to see if that benefit is seen in all FI categories,” she explained.

According to Dr. Buckstein, this study raises two new questions that require further explanation. “First, we need to see if the addition of an FI can refine or provide added value to the upcoming prognosis scoring system that incorporates somatic mutations, the IPSS-Rm, which we’re expecting to be unveiled later this year,” she said. “Second, we need to know what the most important critical elements of this FI are. We could consider these elements for inclusion in a future model. For example, perhaps this MDS-FI will work really well with only 15 elements instead of the 42 elements.”

Based on her findings and clinical experience, Dr. Buckstein encourages physicians and researchers to avoid staying “siloed” in the prognostication of patients with MDS, which often happens when only disease-related features are examined. “This is a disease where the median age of onset is in the 70s,” she said. “To just look at prognosis in isolation, based on blasts, cytogenetics, and blood count, without considering the patient in which the disease has taken hold, is naïve.”

The authors report no relevant conflicts of interest.

Reference

Starkman R, Alibhai S, Wells RA, et al. An MDS-specific frailty index based on cumulative deficits adds independent prognostic information to clinical prognostic scoring. Leukemia. 2019 December 6. [Epub ahead of print]