Abstract
Background: Both MM specific and patient specific parameters place patients at risk for death. Patient frailty has been delineated primarily by age and ECOG performance score (PS) and secondarily with creatinine, eGFR, and the International Staging System (ISS), which reflect both renal function and disease burden. In an attempt to improve upon these commonly used measures, a frailty score for elderly patients that combines functional status [Activity of Daily Living (ADL) and Instrumental Activity of Daily living scores], comorbidities [Charlson comorbidity index (CCI)] and age has been developed by the IMWG. Incorporation of these tools into a busy clinical practice is challenging due to time constraints. Therefore, the identification of frail patients with an easily applicable, rapid and objective tool remains a challenge and an unmet need. A reliable, non-invasively assessable, biochemical marker of frailty would be an ideal tool to identify patients at higher risk of death. It was our goal to determine the role of NT-proBNP, a well-established cardiovascular risk biomarker, as a marker of frailty in patients with MM and in predicting overall survival (OS).
Methods: Patients were elegible for this retrospective study if they were seen at the Mayo Clinic, Rochester, MN within 30 days of their myeloma diagnosis during the interval between 1/1/2007 and 12/31/2011. As part of the Mayo Clinic registration, all patients complete forms containing their past medical history, symptoms, and ADLs. Data from that first visit was abstracted and used to calculate a CCI. We excluded all patients with a concomitant diagnosis of light-chain AL amyloidosis. NT-proBNP concentration was measured in frozen sera collected within 30 days of diagnosis. NT-proBNP assay was run on the E170 Modular analyzer (Roche Diagnostics, Penzberg, Germany). We evaluated the prognostic role of NT-proBNP and these indices on OS using the Kaplan-Meier method.
Results: Among the 351 consecutive patients satisfying entry criteria, median age was 65 years (range 22-95), 33% were ≥70 years, and 56% were male. Twenty-eight percent were ISS stage III, 13% had a creatinine ≥2 mg/dL, 19% had PS ≥2, and 11% had ADL score ≥2, and 30% CCI ≥2. The median value of NT-proBNP was 109 ng/L (interquartile range: 30-375 ng/L). NT-proBNP concentrations differed in the three ISS stages (median: 122 ng/L stage I, 190 ng/L stage II and 1822 ng/L stage III, P<0.0001). Median OS was 5.6 years. The best cutoff of NT-proBNP predicting OS at one year was 301 ng/L (sensitivity: 57.88%, specificity: 74.29%; AUC=0.668) and distinguished two groups with different OS (median not reached vs. 35 months, P<0.0001). Variables predictive for OS are shown in the Table and included PS, age, CCI, ISS, NT-proBNP. On multivariate analysis, ISS and NT-proBNP were independent of each other as long as age, PS and/or CCI were excluded from the model. NT-proBNP withstood these clinical variables better than did ISS (data not shown). We built two different models (table). In model 1, NT-proBNP >301 ng/L was an independent predictor of survival while age ≥70 and PS ≥2 and CCI were not significant. However, if we considered ISS stage instead of NT-proBNP in the same model, ISS stage was not significant.
Conclusions: In this unselected cohort of patients with newly diagnosed multiple myeloma, NT-proBNP was a useful predictor of survival independent of age and PS. The threshold of 301 ng/L is remarkably similar to the well-established cutoff for heart failure of 300 ng/L. NT-proBNP outperformed the CCI and the ISS as risk factors. NT-proBNP is a widely available biomarker that could be added to the panel of laboratory tests of newly diagnosed MM patients and serve as a simple means of determining patient frailty in a busy clinical practice.
. | Univariate . | Multivariate . | Multivariate . | |||
---|---|---|---|---|---|---|
. | RR (95% CI) . | P . | Model 1 . | P . | Model 2 . | P . |
PS ≥2 | 3.4 (2.4-4.8) | <0.0001 | 2.5 (1.7-3.6) | <0.0001 | 2.8 (1.9-4.0) | <0.0001 |
Age ≥70 | 2.7 (1.9-3.6) | <0.0001 | 2.1 (1.5-2.9) | <0.0001 | 1.9 (1.3-2.7) | 0.0003 |
CCI >=2 | 1.9 (1.4-2.6) | <0.0001 | NS | NS | 1.5 (1.0-2.0) | 0.03 |
ISS stage | 1.1 (1.1-1.8) | 0.0004 | NI | NI | NS | NS |
NT-proBNP >301 ng/L | 2.3 (1.7-3.2) | <0.0001 | 1.6 (1.1-2.2) | 0.007 | NI | NI |
. | Univariate . | Multivariate . | Multivariate . | |||
---|---|---|---|---|---|---|
. | RR (95% CI) . | P . | Model 1 . | P . | Model 2 . | P . |
PS ≥2 | 3.4 (2.4-4.8) | <0.0001 | 2.5 (1.7-3.6) | <0.0001 | 2.8 (1.9-4.0) | <0.0001 |
Age ≥70 | 2.7 (1.9-3.6) | <0.0001 | 2.1 (1.5-2.9) | <0.0001 | 1.9 (1.3-2.7) | 0.0003 |
CCI >=2 | 1.9 (1.4-2.6) | <0.0001 | NS | NS | 1.5 (1.0-2.0) | 0.03 |
ISS stage | 1.1 (1.1-1.8) | 0.0004 | NI | NI | NS | NS |
NT-proBNP >301 ng/L | 2.3 (1.7-3.2) | <0.0001 | 1.6 (1.1-2.2) | 0.007 | NI | NI |
Merlini:Janssen-Cilag: Honoraria; Prothena: Honoraria; Millennium-Takeda: Honoraria; Pfizer: Honoraria. Kumar:Janssen: Research Funding; Celgene: Research Funding; Sanofi: Research Funding; Millenium/Takeda: Research Funding; AbbVie: Research Funding; Onyx: Research Funding; Celgene, Millenium, Sanofi, Skyline, BMS, Onyx, Noxxon,: Other: Consultant, no compensation,; Skyline, Noxxon: Honoraria.
Author notes
Asterisk with author names denotes non-ASH members.