Ninety-two patients (53% male and 47% female) with NDMM, fulfilling the International Myeloma Working Group [18] (IMWG2014 diagnostic criteria for symptomatic MM), who developed sepsis or septic shock were enrolled. The patients had a median age of 62.5 (IQR 53–74) years. Among the 92 NDMM patients, a significant proportion of individuals developed sepsis or septic shock. In terms of MM treatment modalities, a notable proportion of patients underwent ASCT, highlighting its importance as a therapeutic approach in eligible MM patients. For patients who did not undergo ASCT, triplet regimens consisting of proteasome inhibitors, immunomodulatory drugs, and corticosteroids were commonly administered, reflecting the current standard of care for non-ASCT patients (63%) [19]. Specifically, across the entire cohort 34 received ASCT, 58 did not. The median value of time interval over the entire cohort was -30 (IQR -302;-21). There is no difference in median values of time interval (days) from the treatment initiation and the sepsis event when stratified for the qSOFA (p = 0.954), SIRS (p = 0.843). ASCT impact was also deemed neither statistically nor clinically significant in our cohort (p = 0.068) (Supplementary Table 1). Additionally, a proportion of patients received anti-CD38 monoclonal antibody therapy, which has shown efficacy in MM treatment (35%) (Table 1). No patients treated with CAR-T were recruited and all patients had different combination regimens (Dara-based or not-Dara-based) in the combination regimen as by protocols approved by the regulatory agency in Europe [20] (Supplementary Table 2).

All the patients received antimicrobial prophylaxis with trimetoprim/sulphamethoxazole and acyclovir as by recommendations [20].

Applying the Sepsis criteria, sepsis was diagnosed in 74 patients (80%), while 19 patients had septic shock (21%), indicating the severity of their condition. For the sepsis group the qSOFA score is 3 in 82% of patients, and the SIRS score is 4 in 80% of patients, reflecting the overall disease burden and organ dysfunction (Table 2).

Among the MM patient population, a considerable number of individuals, 60% precisely, presented with renal failure at the time of sepsis diagnosis, underscoring the vulnerability of the renal system in these patients and the impact of sepsis on their renal function (Table 2). Anemia, defined by a hemoglobin level below 10 g/dL, was observed in 71% of the cases, reflecting the hematological disturbances accompanying sepsis in MM patients (Table 2).

The presence of bone disease, as defined by the CRAB criteria (hypercalcemia, renal insufficiency, anemia, and lytic bone lesions), was prevalent in 83% of the patients, with lytic bone lesions being the most common manifestation (Table 2). Additionally, immunoparesis, characterized by a decrease in one or more immunoglobulin classes, was observed in 70% of the patients [21].

Microbiological analysis yielded valuable insights into the underlying pathogens responsible for sepsis in these MM patients. Respiratory tract infections accounted for a significant portion, constituting 40% of the cases. Within this group, gram-negative bacteria such as Klebsiella pneumoniae and Pseudomonas aeruginosa were found to be predominant causative agents, as they were reported in 28% of the population. Gram-positive bacteria, including Staphylococcus aureus and Streptococcus pneumoniae, were responsible for 23% of the infections, indicating their substantial contribution to the overall septic burden. Additionally, fungal infections [22], predominantly caused by Candida species (C. albicans, C. glabrata, C. tropicalis, C. parapsilosis), were identified in 20% of the cases, highlighting the opportunistic nature of these pathogens in immunocompromised individuals [23]. The analysis revealed the distribution of different types of infections and the prevalence of specific pathogens (Table 2). Notably, Listeria monocytogenes infections were detected in 9.78% of the cases, primarily occurring in patients receiving CD38-directed antibody therapy (78% vs. 30%, p = 0.004), which warrants attention in the management of these patients (Table 3).

Regarding the antimicrobial therapy, patients were treated as in Supplementary Table 3, following our Institutional and international Guidelines [24, 25] (Supplementary Table 4). Antimicrobial de-escalation was deemed essential upon the reporting of culture test results. Collaboration with infectious disease consultants or the sepsis team was employed to streamline therapy based on the identified pathogen and its antimicrobial sensitivity profile, along with clinical response and laboratory trends. Caution was advised regarding carbapenem use, reserved for essential cases, and considering substitution of vancomycin, teicoplanin, and daptomycin with cefazolin or oxacillin if methicillin-sensitive staphylococcus were isolated. The duration of antibiotic therapy was also selected by clinical progress and biomarker trends. Specifically, the antimicrobial therapies, matching the given schedule and posology (Supplementary Table 5), lasting longer than 7–10 days were deemed unnecessary, except in specific scenarios like slow clinical improvement.

PFS was deemed significantly longer in patients with albumin levels ≥ 3.5 than in patients with albumin levels < 3.5 (Log-rank test = 52.10, p < 0.001) and in patients with KS ≥ 80 vs < 80 (Log-rank test = 10.86, p = 0.001) (Fig. 1A and 1B). Moreover, PFS showed significant improvement among patients in earlier disease stages of MM, particularly those identified as ISS–I (Log-rank test = 73.27, p < 0.001) and R-ISS–I (Log-rank test = 45.19, p < 0.001), compared to others affected by more progressed forms of the underlying disease (Fig. 2A and B).

Kaplan–Meier estimates of PFS based on Albumin levels (A) and Karnofsky Performance Status Scale (B)

Kaplan–Meier estimates of PFS based on disease stages according to the International Staging System (ISS) (A) and the Revised International Staging System (R-ISS) (B)

Cox univariate analyses of progression-free survival (PFS) showed statistically significant HR in patients with SIRS 4 vs 2 (HR = 0.56, p = 0.037), albumin levels < 3.5 vs \(\ge\) 3.5 (HR = 5.04, p < 0.001), Karnofsky Performance Status Scale (KS) < 80 vs \(\ge\) 80 (HR = 2.01, p = 0.002), age (HR = 1.02, p = 0.008) and late-stage vs early-stage disease according to International Staging System (ISS) (HR = 4.76 and HR = 12.52, both p < 0.001) and Revised-International Staging System (R-ISS) (R-ISS III vs R-ISS I, HR = 7.38, p < 0.001) (Table 4). The multivariate model confirmed the results for age (HR = 1.02, p = 0.005) and R-ISS III vs R-ISS I, HR = 7.08, p < 0.001).

Ancillary to the main outcomes, death events were recorded and summarized in Fig. 3A. While we do not have specific data on sepsis-related mortality, our study provides an overview of death events observed in our MM cohort during the study period. Statistically powered perspective studies will aim to include specific data on sepsis-related mortality. While acknowledging that the analysis of overall survival (OS) is beyond the scope of this manuscript, given the actual median survival of myeloma patients, the OS since the diagnosis of multiple myeloma to the death event was performed by applying the Kaplan–Meier estimator. We analyzed the course of the survival to the event weighted by explanatory variables (qSOFA = 1 and qSOFA = 3, anemia, renal failure, bone disease, immunoparesis, cytogenetics, frailty, response, karnofsky, R-ISS, albumin, ISS). OS was significantly different in patients with albumin levels ≥ 3.5 than in patients with albumin levels < 3.5 (Log-rank test = 6.92, p = 0.008) and among patients in earlier disease stages of MM, particularly those identified as ISS–I (Log-rank test = 16.94, p < 0.001), compared to others affected by more progressed forms of the underlying disease (Fig. 3B and C).

Distribution of death events (A) and Kaplan–Meier estimates of OS based on Albumin levels (B) and disease stages according to the International Staging System (ISS) (C)