[A predictive model based on risk factors for early mortality in patients with newly diagnosed multiple myeloma].

Objective: To investigate risk factors for early mortality (EM) in patients with newly diagnosed multiple myeloma (NDMM) and to build an EM-predictive model. Methods: In a cohort of 275 patients with NDMM, risk factors for EM at 6, 12, and 24 months after diagnosis (EM6, EM12, and EM24, respectively) were determined to establish a model to predict EM. Results: The rates of EM6, EM12, and EM24 were 5.5% , 12.7% , and 30.2% , respectively. The most common cause for EM was disease progression/relapse, accounting for 60.0% , 77.1% , and 84.3% of EM6, EM12, and EM24, respectively. EM6 was associated with corrected serum calcium >2.75 mmol/L and platelet count <100×10(9)/L, whereas risk factors for EM12 included age >75 years, ISS Ⅲ, R-ISS Ⅲ, corrected serum calcium >2.75 mmol/L, serum creatinine >177 µmol/L, platelet count <100×10(9)/L, and bone marrow plasma cell ratio ≥ 60% . In addition to the risk factors for EM12, EM24 was also associated with male sex and 1q21 gain. By multivariate analysis, age >75 years, platelet count <100×10(9)/L, and 1q21 gain were independent risk factors for EM24 but there were no independent risk factors significantly associated with EM6 and EM12. Using a scoring system including these three risk factors, a Cox model for EM24 was generated to distinguish patients with low (score<3) and high (score ≥ 3) risk. The sensitivity and specificity of the model were 20.7% and 99.2% , respectively. Further, an internal validation performed in a cohort of 183 patients with NDMM revealed that the probability of EM24 in high-risk patients was 26 times higher than that in low-risk patients. Moreover, this model was also able to predict overall survival. The median overall survival of patients with scores of 0, 1, 2, 3, 4, and 5 were 59, 41, 22, 17.5, and 16 months, respectively. Conclusion: In the study cohort, the EM6, EM12, and EM24 rates were 5.5% , 12.7% , and 30.2% , respectively, and disease progression or relapse were main causes of EM. An EM24-predictive model built on three independent risk factors for EM24 (age>75 years, platelet count<100×10(9)/L, and 1q21 gain) might predict EM risk and overall survival.