Abstract
Background
Infection is one of the leading causes of mortality in multiple myeloma (MM) due to disease-related immune dysfunction and treatment-induced cytopenias (Blimark et.al., 2025). The relationship between longitudinal immune parameters and infection-related death remains incompletely understood but is particularly relevant in the current era of immune-modulating agents. This study aimed to explore trends in absolute neutrophil count (ANC), absolute lymphocyte count (ALC), and immunoglobulin G (IgG) levels, and their association with infection-related mortality in a real-world cohort of MM patients.
Methods
We retrospectively reviewed MM patients diagnosed between 2021-2023 at Northwell Health. After excluding those without confirmed MM, no outpatient follow-up, or who died before induction, 173 patients were included. Data collected included demographics; diagnostic data; lab results at baseline, 3 months, and 6 months from diagnosis; body mass index (BMI); treatment history; infection rates; and use of IVIG and antimicrobial prophylaxis. Neutropenia was defined as ANC < 1.0 ×109/L, lymphopenia as ALC < 0.5 ×109/L, and hypogammaglobulinemia as IgG < 400 mg/dL.
Results
Among the 173 patients, 54% were female, with a median age at diagnosis of 68.9 years (IQR 60.8–76.9) and a median BMI of 28 (IQR 24.1–32). Racial distribution included 39.3% White, 35.3% Black/African American, 6.4% Asian, 0.6% American Indian/Alaskan Native, 14.5% Other, and 4.0% not reported. High-risk cytogenetics were present in 41 patients. R-ISS staging showed 32 patients (18%) in stage I, 70 (40%) in stage II, and 37 (21%) in stage III; data were missing for the remainder. At diagnosis, 22 patients (13%) had hypogammaglobulinemia, 10 (6%) had neutropenia, and 38 (22%) had lymphopenia.
Antiviral prophylaxis was given to 149 patients (86%), and antibiotic prophylaxis to 12 (7%). IVIG prophylaxis was documented in 2 patients (1.1%). Within 6 months of diagnosis, 12 patients (7%) developed bloodstream infections, 10 (6%) with gram-positive bacteremia and 5 (3%) with gram-negative bacteremia. Viral infections occurred in 21 patients (12%), including COVID-19 (n=8), hepatitis B (n=1), herpes zoster (n=2), and parainfluenza (n=1). At a median follow-up of 559 days, 51 patients (29%) had died, with infection listed as the cause of death in 25 cases.
Patients who died of infection had lower median ANC at diagnosis (3.57 vs. 5.18 ×10⁹/L), 3 months (2.83 vs. 3.93 ×10⁹/L), and 6 months (2.93 vs. 3.24 ×10⁹/L), though differences were not statistically significant (p=0.08, 0.06, and 0.70, respectively).
Among deceased patients, ALC did not differ significantly between infection-related and other causes of death. Median ALC was similar at diagnosis (1.25 vs. 1.62 ×10⁹/L, p=0.813), 3 months (1.10 vs. 1.07 ×10⁹/L, p=0.543), and 6 months (1.14 vs. 1.04 ×10⁹/L, p=0.702), suggesting ALC may not predict infection-related mortality in this cohort.
Median IgG levels were 1299 mg/dL at diagnosis, 764 mg/dL at 3 months and 623 mg/dL at 6 months (n=173). Among deceased patients, those who died of infection had lower baseline IgG (729 vs. 1408 mg/dL), though not statistically significant (p=0.09). IgG levels at 3 and 6 months also showed no significant differences by cause of death (p=0.46 and 0.47, respectively).
Discussion
Infection was the leading cause of death in this cohort—nearly half of all cases—most often due to COVID-19. This highlights the ongoing need for precautions in this vulnerable population. Patients who died of infection showed consistently lower median ANC over time, though not statistically significant, likely due to limited sample size. ALC remained similar across groups, suggesting limited utility as a predictive marker. A trend toward lower baseline IgG was seen in infection-related deaths, but not at later time points. While no single immune parameter predicted infection-related mortality, early declines in ANC and IgG may help identify higher-risk patients for closer monitoring or intervention. Further validation with larger cohort is required to establish the association of ANC and IgG at diagnosis with infection related mortality.
