Assessment of Frailty Measures and Their Effect on Survival in a Population of Patients Registered At Marche Region Multiple Myeloma (MM) Registry

Abstract 2873

International staging system (ISS) and cytogenetics are the main prognostic factors of Multiple Myeloma (MM) but they reflect biologic characteristics of disease without taking into account individual host features. On the contrary, clinical characteristics of single patient could be substantial as to various points of view. For instance, in elderly MM patients, novel therapies reduction or interruption due to toxicity represent the major cause of unsatisfactory outcome. Therefore, it was empirically suggested different schedule of drugs in these “frail” patients but, how the “frailty” should be assessed in every single patient, is still unsettled. Advanced age, poor performance status (PS) and comorbidities are usually applied to recognize the “frailty” but it is not well known which of them are really prominent and whether these parameters, adjusted for conventional prognostic factors, still affect final outcome.

We analyzed a population of symptomatic MM diagnosed from 2007 to 2010 included in the Marche Region MM Registry, to assess the frequency of “frailty” features, such as age, PS, comorbidities, cytopenias, renal insufficiency (RI) and lytic bone lesions, and their role on the overall survival (OS) when adjusted for prognostic factors. Comorbidities were scored according to Charlson Comorbidity Index (CCI) that split patients in 4 categories according to number and type of comorbidity. Patients were treated with transplant or standard therapy according to their eligibility. Overall, 88% of patients were treated with new drug-based therapies and 12% with MP.

Median age of the 266 patients analyzed was 73 years (range 38–90). Twenty-four percent of patients had IgA MM, fifty patients (23%) had ISS stage=3 and 29/166 (17.5%) had unfavourable cytogenetics. Regarding “frailty” measures, 38% of patients had > 75 years, 39% had PS=2–4, 34% had 1 or more comorbidities. The most frequent comorbidities were hypertension (35%), heart diseases (22%), diabetes (15%), neurological diseases (16%), COBP (8%), secondary malignancies (8%) and chronic renal failure (6%). CCI ≥1 was detected in 51%. Increasing comorbitities number and CCI were associated with increased age although 37% of patients aged less than 65 years had CCI ≥2. Moreover, 35% had at least 2 cytopenias, 76% had bone disease and 14% had RI. Fifty patients (19%) died during follow-up. OS at 3 years was 74%. Univariate analysis performed on the total population determined age > 65 years (p=0.065), PS=2–4 (p<0.001), Hb < 10 g/dl (p=0.035), ISS=3 (p=0.017), unfavourable cytogenetics (p=0.074), cytopenias (p=0.024), RI (p=0.036) and CCI=1–3 (p=0.005) as factors that significantly decrease OS. Multivariate stepwise analysis selected ISS=3 (HR=1.6; p=0.033), PS=2–4 (HR=2.5; p=0.007) and CCI=1–3 (HR=2.1; p=0.028) as factors affecting worse OS. To obtain a clinical applicable prognostic model, we assigned 1 point to each unfavourable finding such as PS=2–4 or CCI=1–3. A “frailty score” (FS) was thus developed as low (0 point), intermediate (1 point) and high (2 points). Patients in the low category had 91% OS at 3 years vs 83% in the intermediate one (p=0.205) vs 36% in the high category (p< 0.0001). According to these results we performed a multivariate analysis adding FS (0–1 vs 2) to the above frailty covariates and to prognostic factors. This analysis was carried out splitting patients in two subgroups according to the age ≤65 years or older; high FS was recognized in 10% and 30% of subgroups, respectively. In younger patients, multivariate analysis selected ISS=3 (median OS 41.5 months vs NR; HR=5.2; p=0.006) and FS=2 (median OS 46.5 vs NR; HR=5.5; p=0.024;) as factors diminishing OS whereas only FS=2 (HR=3.5; p=0.002) affects worse survival in older ones. Patients with FS=2 had a median survival of 27 months vs NR in patients with FS=0-1.

This study demonstrated that, behind biologic features of MM, host characteristics are of utmost importance regarding final outcome prediction. Two easy, reliable and not time-consuming measures normally used in clinical practice such as PS and CCI, scored as we found, seem to be very useful tool to identify frail patients independently by age or other single frailty parameter. These findings maybe used to try to explain why outcome of some patients poorly improved despite new-drugs introduction and should be taking into account planning future clinical trials.