Time series analysis Between Serum Calcium Levels and 60-Day Mortality in Critically ill Patients with Multiple Myeloma: A Retrospective Cohort Study

Background Hypercalcemia is a well-recognized adverse prognostic factor in patients with multiple myeloma (MM), particularly in critically ill subgroups where its incidence is substantially higher than in the general MM population. However, the prognostic value of dynamic serum calcium fluctuations during intensive care unit (ICU) hospitalization remains underexplored, limiting the ability to refine risk stratification and personalized management. Objective This study aims to identify heterogeneous serum calcium trajectories in critically ill MM patients and investigate their association with 60-day mortality. Methods A retrospective cohort study was conducted using data from 254 critically ill MM patients extracted from the MIMIC-IV database. Latent Class Growth Modeling (LCGM) was applied to classify serum calcium trajectories over the first 28 days of hospitalization, with measurements collected at predefined time points-. Multivariable Cox proportional hazards regression models (adjusted for demographic, clinical, and disease severity covariates) and Kaplan-Meier (K-M) curves were used to assess the association between trajectory patterns and 60-day mortality. Sensitivity analyses included trajectory reclassification and E-value assessment to validate result robustness. Results Four distinct serum calcium trajectories were identified: low-level (n = 90, 35.43%), medium-level (n = 138, 57.48%), U-shaped (n = 9, 3.54%), and high-level (n = 17, 6.68%). Compared with the high-level trajectory (reference group), the low-level (HR = 0.17; 95% CI 0.07–0.41; P < 0.001) and medium-level (HR = 0.26; 95% CI 0.12–0.56; P = 0.001) trajectories were independently associated with significantly lower 60-day mortality. The U-shaped trajectory showed no significant mortality difference from the high-level trajectory (HR = 0.56; 95% CI 0.14–2.25; P = 0.416). K-M analysis confirmed the prognostic gradient across trajectories, and sensitivity analysis (E-value = 2.49) validated the robustness of these associations to unmeasured confounding. Conclusion Persistently high or rebound (U-shaped) serum calcium trajectories are independent predictors of increased 60-day mortality in critically ill MM patients, reflecting underlying aggressive disease biology or treatment resistance. Dynamic serum calcium monitoring via trajectory analysis refines early risk stratification and provides actionable insights for guiding personalized ICU management, addressing an unmet need in current clinical practice.