Comorbidities and a deteriorated functional status have been demonstrated to affect progression free survival (PFS) and overall survival (OS) in various cancer pts, but have as yet not been formally evaluated in MM. Structured comorbidity analyses seem to be exceedingly needed in MM, as these pts are typical elderly and some are frail. The study aim was to develop and validate an easily assessable and cost effective MM-risk score and compare this to previously established comorbidity indices (CIs).
In initially 127 MM pts consecutively treated at our institution between 1997 and 2003, we determined age, Karnofsky-Performance Status, hypertension, diabetes, secondary malignancies, pain, liver-, heart-, lung-diseases and renal impairment (eGFR). Via uni- and multivariate Cox regression analyses, we developed a simple risk score, termed Freiburger comorbidity index (FCI), that consists of 3 risk factors only. We also compared previously established CIs, namely Kaplan Feinstein (KF), Charlson-Comorbidity (CCI), Satariano (SI) and Hematopoietic cell transplantation-specific comorbidity index (HCT-CI), assessing PFS and OS in ‘low- risk’ (scoring ≤median CI points) vs. ‘high-risk group’ pts (>median CI points). In another set of 466 consecutive MM pts treated in our department thereafter (2003-2009), we re-evaluated comorbidity, organ function and MM risk factors via uni- and multivariate analyses and determined the prognostic value of the FCI, KF and HCT-CI on PFS and OS.
The multivariate analysis on the initial 127 MM pts identified the Karnofsky-Performance Status <70%, moderate or severe lung impairment and eGFR <30ml/min/1.73m2 as most relevant prognostic factors, with hazard ratios (HR) for decreased OS of 2.2, 2.8 and 2.9, respectively. When incorporating these risk factors into the FCI, we identified a largely different median OS: with 0, 1 and 2–3 risk factors this was 118, 53 and 25 months, respectively (p<0.005). The initial comparison of all 4 CIs (KF, CCI, SI and HCT-CI) identified the KF and HCT-CI as most relevant: ’low-risk’ vs. ‘high-risk’-pts had a strikingly different median OS of 98 vs. 44 months (p=0.007), and 81 vs. 41 months (p=0.002), respectively. The confirmation analysis in 466 MM pts verified the prognostic relevance of the FCI, KF and HCT-CI, showing median OS differences between ‘low-risk’ vs. ‘high-risk’ pts of 113 vs. 39, 143 vs. 36 and 117 vs. 49 months, respectively (log rank-test p< 0.0001 each). The multivariate analysis of prognostic factors revealed that high-risk cytogenetics, elevated LDH, increased bone marrow infiltration, ß2-microglobulin (ß2-MG) and impaired albumin were most relevant. In our current test and validation analysis, we are re-evaluating all MM pts treated between 1997 and 2009, to confirm the value of specific organ function, prognostic risks and comorbidity scores in order to develop a weighted FCI for future analyses. Clinical characteristics of this combined pt set showed a median age of 62 years (27-90); 86% had stage II/III disease by Durie&Salmon and 18% stage B disease. The descriptive comorbidity evaluation revealed a Karnofsky-Performance Status <70% in 64% (median 70%; range: 20–100%), moderate or severe lung disease in 18% and eGFR <30 in 16%. Current results also indicate that weighted comorbidity scores- as realized with the in MM valuable KF and HCT-CI - allow to distinguish significantly different MM risk groups. The by our group developed FCI is currently a non-weighted, but simple, concise and cost effective comorbidity tool consisting of 3 factors only, namely Karnofsky-Performance Status, lung- and renal function, which is of importance to develop further to a weighted score as well as in combination with other indispensable risk factors.
The new FCI, KF and HCT-CI proved as useful tools for risk evaluation in MM. The development of a weighted FCI within a test- and validation-set is currently being pursued as well as the inclusion of other prognostic parameters therein (such as ISS). Our results highlight the strong debate on comorbidity scores and comprehensive risk assessment in MM which have not as yet moved into routine clinical practise, but appear appealing and useful to implement into future analyses and clinical trials.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.