Predictors of early morbidity and mortality in newly diagnosed multiple myeloma: data from five randomized, controlled, phase III trials in 3700 patients

Early morbidity and mortality affect patient outcomes in multiple myeloma. Thus, we dissected the incidence and causes of morbidity/mortality during induction therapy (IT) for newly diagnosed multiple myeloma (NDMM), and developed/validated a predictive risk score. We evaluated 3700 transplant-eligible NDMM patients treated in 2005–2020 with novel agent-based triplet/quadruplet IT. Primary endpoints were severe infections, death, or a combination of both. Patients were divided in a training (n = 1333) and three validation cohorts (n = 2367). During IT, 11.8%, 1.8%, and 12.5% of patients in the training cohort experienced severe infections, death, or both, respectively. Four major, baseline risk factors for severe infection/death were identified: low platelet count (<150/nL), ISS III, higher WHO performance status (>1), and age (>60 years). A risk score (1 risk factor=1 point) stratified patients in low (39.5%; 0 points), intermediate (41.9%; 1 point), and high (18.6%; ≥2 points) risk. The risk for severe infection/death increased from 7.7% vs. 11.5% vs. 23.3% in the low- vs. intermediate- vs. high-risk groups (p < 0.001). The risk score was independently validated in three trials incorporating quadruplet IT with an anti-CD38 antibody. Our analyses established a robust and easy-to-use score to identify NDMM patients at risk of severe infection/death, covering the latest quadruplet induction therapies. Trial registrations: HOVON-65/GMMG-HD4: EudraCT No. 2004-000944-26. GMMG-MM5: EudraCT No. 2010-019173-16. GMMG-HD6: NCT02495922. EMN02/HOVON-95: NCT01208766. GMMG-HD7: NCT03617731.


INTRODUCTION
Clinical outcomes markedly improved over the past two decades in multiple myeloma (MM) [1,2].Yet, the risk of early morbidity and mortality can limit the therapy-related benefit of long-term disease control in a substantial number of patients.Various studies have demonstrated that the risk for morbidity and mortality from adverse events, mainly severe infections, during treatment initiation exceeds the risk from MM progression [3][4][5][6][7].
Thus, better characterization and improved prediction of individual risk for severe infections and death are essential to develop advanced preventive measures.To date, no large analyses have evaluated early morbidity and mortality in transplant-eligible patients with newly diagnosed MM (NDMM) in the era of modern myeloma treatment.
The present multi-cohort analysis included 3700 transplanteligible patients with NDMM from five multi-center, phase III trials.All patients received novel agent-based triplet or quadruplet induction therapies.The aims of this study were (i) to dissect the incidence, timing, and causes of morbidity and mortality during induction therapy, and (ii) to develop and validate a predictive risk score to identify patients at excessive risk of severe infections and death during the early treatment phase.
All patients included in this analysis had untreated NDMM and were considered eligible for induction therapy followed by high-dose melphalan (200 mg/m 2 ) and autologous stem cell transplantation.Patients received at least a bortezomib-containing triplet induction regimen: bortezomibdoxorubicine-dexamethasone (PAD; HD4, MM5, HO65) or bortezomibcyclophosphamide-dexamethasone (VCD; MM5, EMN02/HO95).In the HD6 and HD7 trials, induction therapy included lenalidomide-bortezomibdexamethasone (RVd) with or without the anti-SLAMF7 mAb elotuzumab (HD6) or anti-CD38 mAb isatuximab (HD7).Information on trials, induction therapy schedules, number of cycles, and recommended use of antibacterial prophylaxis is summarized in Supplementary Table S1.All analyses were performed on individual patient-level data.Trials were conducted according to the European Clinical Trial Directive and the Declaration of Helsinki, and were approved by the local ethics committees.All patients gave written informed consent for participation in the respective clinical trials.

Definitions, assessments, and objectives
Patients who received at least one dose of trial medication were included and analyzed as treated.Eight patients from the HD6 trial receiving > 4 induction therapy cycles were excluded from the analysis.The induction period was defined from the first until the last dose of induction treatment plus 30 days, or until the start of stem-cell mobilization.
The primary endpoints of the study were rates of severe infection, death from any cause, or a combined endpoint of severe infection/death from any cause, whichever occurred first, during the induction period.Severe infections were defined as any infection of grade ≥3 according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE).In case of multiple severe infections, the first one was counted.

Statistics and general methods
Fisher's exact test and Wilcoxon test were used to compare categorical and continuous variables between trials.Univariable and multivariable logistic regression models were used to assess the impact of risk factors on endpoints, depicted as odds ratio (OR) and 95% confidence interval (95% CI).In case of complete separation, logistic regression with Firth correction was applied.In all pooled analyses, the trial effect was included in the model.Likelihood-ratio test between model, with and without parameter-trial interaction term, was used to assess heterogeneity of effect between trials.For multivariable models, multiple imputations of missing values (100 bootstrap samples) for baseline variables were done by applying the multivariate imputations using the chained equations (mice) algorithm [21].P values from the univariable analysis were adjusted for multiple testing using Benjamini-Hochberg correction to control the false discovery rate.P values < 0.05 were considered statistically significant.Analyses were performed with the statistical software R 4.0 (R Foundation for Statistical Computing, Vienna, Austria; https://www.Rproject.org/).

Patient characteristics and treatment completion
The training cohort included 1333 patients (192, 596, and 545 patients from the HD4, MM5, and HD6 trials, respectively).The median patient age was 58 (range 27-70) years with 534 (40.1%) patients aged > 60 years.The median duration of induction treatment was 89 (range 2-281) days.In total, 1261 (94.6%) patients completed regular induction treatment and 1103 (91.9%) received antibacterial prophylaxis.The baseline characteristics of patients in the training cohort are listed in Supplementary Table S2.
The majority of severe infections, deaths, or a combination of both occurred during the first two induction cycles: 119/158 (75.3%), 16/24 (66.7%), and 124/167 (74.3%;Fig. 1A-C).The median time from the start of induction therapy to the first severe infection, death, or a combination of both was short (severe infections: 36 [range 1-119] days; death: 66 [range    Identification of factors influencing the risk of severe infections, death, or the combined endpoint of severe infection/death during induction therapy We aimed to identify risk factors associated with all three endpoints.Owing to the relatively small number of events, analyses were conducted in the pooled training cohort, accounting for trial effects (Fig. 2).The effects of risk factors for each trial included in the training cohort (HD4, MM5, HD6) are shown in Supplementary Fig. S1.

Multivariable model on predictors for risk of severe infection/ death during induction therapy
Based on the findings from the logistic regression analyses, we aimed to confirm significant risk factors for the most clinically relevant, combined endpoint, severe infection/death, in a multivariable model.To account for parameters commonly used in clinical practice to assess infection, low baseline white blood cell counts and elevated C-reactive protein levels were included in the analysis as well.The multi-variable model identified four independent risk factors for the combined endpoint of severe  infection/death during induction therapy: low platelet count (OR = 2.05, 95% CI: 1.28-3.33,p = 0.003), ISS stage III (OR = 1.93, 95% CI: 1.23-2.96,p = 0.004), WHO performance status > 1 (OR = 1.83, 95% CI: 1.10-3.08,p = 0.021), and age > 60 years (OR = 1.73, 95% CI: 1.22-2.43,p = 0.002; Table 1).

DISCUSSION
Our results demonstrate that a combination of parameters readily accessible in the clinic-low platelet count (<150/nL), ISS stage III, WHO > 1, and age >60 years-consistently predicted risk of severe infection and severe infection/death in patients treated with concurrent, modern induction therapies, including quadruplets with an anti-CD38 mAb.To our knowledge, this is the largest, pooled analysis of individual patient-level data on early morbidity and mortality during novel agent-based induction therapy in transplant-eligible patients with NDMM, comprising 3700 patients treated between 2005 and 2020.Induction therapies in our dataset included a broad variety of widely used standard-of-care [22] triplet combinations such as RVd and VCD, as well as quadruplet combinations with the anti-CD38 mAb isatuximab plus RVd.Thus, the validated risk score can be considered a novel important tool to inform clinicians on the individual risk of early morbidity and mortality in patients with NDMM.Furthermore, it can serve as a benchmark in the design of future prospective clinical trials aiming to reduce early morbidity/mortality or tailor supportive care.
In line with prior analyses [5-7, 23, 24], our study showed that severe infections remain the major cause of early morbidity and mortality in NDMM, and by far exceed the risk of disease progression.This holds true, even though the incidence of severe infections and death has decreased over time in subsequent trials with the introduction of novel agents and the use of antibacterial prophylaxis in most patients (i.e., 91.9% in our training cohort).Thus, preventing early severe infections and death is paramount to achieving optimal outcomes in MM patients.
The composition of our risk score highlights that both diseasespecific factors and tumor burden (ISS stage III, low platelet count), as well as host factors (poor performance status, older age), contribute to early morbidity and mortality.A recent study including 1347 pooled NDMM patients (of whom 847 were transplant-eligible) from the Spanish study group found similar predictors for severe infections (serum albumin, MM immunoglobulin subtype, male sex, European Cooperative Oncology Group [ECOG] performance status) [23].The study evaluated mostly Our study has a few limitations.It includes rather young and fit transplant-eligible NDMM patients treated within clinical trials.However, such systematically evaluated and detailed safety and endpoint data are rarely available outside of clinical trials.
Further, we could not evaluate the impact of antibacterial prophylaxis, since the majority of patients received antibacterial prophylaxis during induction therapy.Despite the positive results on the use of antibacterial prophylaxis with levofloxacin in the TEAMM trial (ISRCTN51731976) [25], its broad use in clinical routine remains controversial.Foremost, it remains an open question whether the use of antibacterial prophylaxis is only beneficial in patients at high risk of severe infections.For example, the TEAMM trial did not provide evidence of a clear benefit with Fig. 3 A novel risk score to predict severe infections/death during novel agent-based induction therapy in transplant-eligible NDMM patients.A Bar plots on incidence of severe infections, death, or both during induction therapy within the low-, intermediate-and high-risk groups in the training cohort.B Incidence of severe infections during induction therapy within the low-, intermediate-and high-risk groups.C Incidence of death during induction therapy within the low-, intermediate-and high-risk groups.D Incidence of severe infections/death during induction therapy within the low-, intermediate-, and high-risk groups.NDMM newly diagnosed multiple myeloma.levofloxacin use vs. placebo in younger, transplant-eligible patients or patients receiving cotrimoxazole prophylaxis [25].Our risk score would allow stratification of patients in clinical trials investigating strategies to prevent infections, including antibacterial prophylaxis.Further, the optimal duration of antibacterial prophylaxis is not known and possible toxic effects (i.e.levofloxacin-induced tendinopathy and neuropathy) should be considered, especially in MM patients [26].Lastly, 21% of patients in the TEAMM trial withdrew consent and 44% of patients received a thalidomide-based treatment, which is not considered a standard-of-care [25].Based on our proposed risk score, preventive strategies may be investigated in prospective clinical trials in a contemporary treatment setting.Concepts for such trials may include initial dose reduction of chemotherapy and glucocorticoids, active monitoring for patients at risk (i.e. by digital wearables), differential use of antibacterial prophylaxis, or substitution of immunoglobulins.
Another limitation is that our study could not dissect the effect of glucocorticoid dose or intensity during induction therapy, as variable glucocorticoid doses and dose intensities were used within and among the trials analyzed.These cannot be disentangled from the overall trial effect, which is accounted for in our analysis.Yet, our risk score was robust, after validation across a variety of induction regimens and accompanying glucocorticoid therapies.However, as shown previously [27,28], treatment with low-dose glucocorticoids (i.e.dexamethasone once weekly) reduces early morbidity and mortality and is a standard-of-care.
In conclusion, our study highlights the importance of early severe infections and death in the era of novel agent-based therapy in patients with NDMM.Based on our risk score, patients at high risk of early, severe infections and death can be easily identified upfront, when evaluated for the latest quadruplet induction therapies including an anti-CD38 mAb.

DATA AVAILABILITY
After the publication of this article, data collected for this analysis and related documents will be made available to others upon reasonably justified request, which needs to be written and addressed to the attention of the corresponding author, Elias K. Mai, at the following e-mail address: elias.mai@med.uni-heidelberg.de.The EMN, HOVON and GMMG, via the corresponding author Elias K. Mai, are responsible for evaluating and eventually accept or refuse every request to disclose data and their related documents, in compliance with the ethical approval conditions, in compliance with applicable laws and regulations, and in conformance with the agreements in place with the involved subjects, the participating institutions, and all the other parties directly or indirectly involved in the participation, conduct, development, management, and evaluation of this analysis. Fig.1A).

Fig. 1
Fig. 1 Incidence of severe infections and deaths or a combination of both during induction therapy in the training cohort.A Frequency of the first severe infection per induction cycle in the overall training cohort and every single trial (HD4, MM5, HD6).The colors indicate the primary infection sites.B Incidence of death per induction cycle in the overall training cohort.The colors indicate the leading causes of death.C Incidence of severe infections and death per induction cycle in the overall training cohort and each single trial (HD4, MM5, and HD6).CNS central nervous system, FUO fever of unknown origin, GI gastrointestinal, GU genitourinary, MM multiple myeloma, TEE thromboembolic event.

Fig. 2
Fig.2Logistic regression analyses on factors influencing the risk of severe infections, death or the combined endpoint of severe infections/deaths during induction therapy in the training cohort.Forest plots on factors influencing A risk of severe infections, B risk of death, and C risk of severe infections and/or death.All logistic regression analyses accounted for trial effects.P values from the univariable analysis were adjusted for multiple testing.BMI body mass index, ISS International Staging System, LDH lactate dehydrogenase, ULN upper limit of normal, WHO World Health Organization.

Fig. 4
Fig.4Validation of the risk score to predict severe infections and/or death during induction therapy in transplant-eligible NDMM patients in the HO65, EMN02/HO95, and HD7 trials.Validation of the risk score in the A HO65 trial, B EMN02/HO95 trial, and C HD7 trial.Bar plots show the incidence of severe infections, death, or both during induction therapy according to low-, intermediate-, and high-risk groups.NDMM newly diagnosed multiple myeloma.

Table 1 .
Multivariable model on factors influencing the combined endpoint of severe infection/death during induction therapy in the training cohort.ISS International Staging System, LDH lactate dehydrogenase, GMMG German-speaking Myeloma Multicenter Group, ULN upper limit of normal, WHO World Health Organization.