Cytomegalovirus reactivation in patients with multiple myeloma receiving bispecific antibodies.

TO THE EDITOR:

Cytomegalovirus (CMV) reactivation (CMVr) is a recognized complication in immunocompromised hosts, including those with hematological malignancies receiving intensive immunosuppression or hematopoietic stem cell transplantation. 1 , 2 CMV may present as DNAemia, symptomatic infection, or end-organ disease, increasing morbidity, mortality, and health care costs. 3 In multiple myeloma (MM), CMVr appears as new phenomenon reflecting cumulative immune dysfunction related to host factors (prolonged survival, persistent lymphopenia, and hypogammaglobulinemia) and specific treatment regimens including corticosteroids, daratumumab, or chimeric antigen receptor T-cell (CAR-T) therapy. 4 , 5

Bispecific antibodies (BsAbs), which redirect T cells toward malignant plasma cells via simultaneous CD3 and tumor antigen engagement (B-cell maturation antigen [BCMA], FcRH5, and GPRC5D), have reshaped MM management due to high response rates. 6 , 7 However, complications such as cytokine release syndrome (CRS) and its immunosuppressive management may heighten infection susceptibility. 8 , 9 , 10 Data on CMVr during BsAb therapy remain scarce, and recommendations for monitoring or prophylaxis are lacking. 11 , 12 This study assessed the incidence, clinical features, risk factors, and outcomes of CMVr in CMV-seropositive patients with MM receiving BsAbs.

We retrospectively reviewed adult patients (aged ≥18 years) with MM who received BsAbs for >1 month between January 2021 and December 2024 at the Peter MacCallum Cancer Centre (Victoria, Australia). Ethics approval was granted by the local Human Research Ethics Committee (reference number QA/118023/PMCC). This study was conducted in accordance with the Declaration of Helsinki. Only CMV-seropositive (CMV immunoglobulin G positive) patients were included. Each BsAb treatment episode was analyzed separately. Clinical data, prior therapies, laboratory parameters, and antiviral management were extracted from medical records. CRS and immune effector cell–associated neurotoxicity syndrome were defined and graded per published criteria. 13 Persistent lymphopenia was defined as ≤1 lymphocyte per 10 9 /L within the 3 months preceding BsAb initiation; hypogammaglobulinemia as serum immunoglobulin G <400 mg/dL. CMVr was classified as CMV DNAemia or clinically significant CMV infection (cs-CMVi; symptomatic infection or end-organ disease) per consensus definitions. 3 CMV testing was clinician directed, and routine monitoring was not standard practice. Antiviral therapy initiation was clinician dependent, guided by viral load kinetics and clinical context rather than a single predefined threshold. Associations were explored using univariate and multivariable logistic regression.

Among 54 patients with MM treated with BsAbs, 39 were CMV seropositive with 45 BsAb treatment episodes. Baseline characteristics are summarized in Table 1 . Patients had received a median of 5 prior therapy lines, including autologous hematopoietic stem cell transplantation (91%), immunosuppressants in the last 90 days (91%), and CAR-T therapy (9%). Hypogammaglobulinemia and persistent lymphopenia occurred in 84% and 71% of episodes, respectively. Four patients (9%) had previous CMVr. BsAb targets were FcRH5 (42%), GPRC5D (38%), or BCMA (20%). Median BsAb therapy duration was 3 months (interquartile range [IQR], 1-12), with frequent discontinuation due to disease progression or toxicity. Any-grade CRS occurred in 58% of episodes; 38.5% received corticosteroids, and 23% received tocilizumab.

Table 1. Baseline characteristics of CMV-seropositive patients with MM with 45 episodes receiving BsAbs Characteristic All episodes n = 45, n (%) No CMV reactivation n = 34, n (%) CMV reactivation n = 11, n (%) Age, median (IQR), y 68 (57-72) 62 (57-72) 70 (64-73) Male sex 32 (71) 25 (73.5) 7 (64) Charlson Comorbidity Index, median (IQR) 2 (2-4) 3 (2-4) 2 (2-4) MM features  M component isotype  IgG 34 (76) 25 (73.5) 9 (82)  IgA 5 (11) 5 (14.5) 0  Light chain 6 (13) 4 (12) 2 (18)  Light chain isotype  Kappa 26 (58) 19 (56) 7 (64)  Lambda 19 (42) 15 (44) 4 (36)  International Staging System median, (IQR) 2 (2-3) 2 (2-2) 3 (1-3) Prior therapy lines, median (IQR) 5 (3-5) 4 (3-5) 5 (4-6)  >3 31 (69) 22 (65) 9 (82)  >5 11 (24) 7 (21) 4 (36) Prior hematopoietic stem cell transplant  Autologous 41 (91) 31 (91) 10 (91)  >1 18 (40) 15 (44) 3 (27)  Allogeneic 3 (7) 2 (6) 1 (9) Prior CAR-T ∗ 4 (9) 3 (9) 1 (9) Prior immunosuppressive therapy (last 90 days) 39/43 (91) 28/32 (87.5) 11/11 (100)  Steroids 37/43 (86) 27/32 (84) 10/11 (91)  Daratumumab 10/43 (23) 7/32 (22) 3/11 (27) BsAb treatment  Target  BCMA 9 (20) 6 (18) 3 (27)  FcRH5 19 (42) 14 (41) 5 (46)  GPRC5D 17 (38) 14 (41) 3 (27)  Duration of therapy, median (IQR), mo 3 (1-12) 3.5 (1-12) 3 (2-7)  Related adverse events  CRS  Any grade 26 (58) 18 (53) 8 (73)  Treated with steroids 10/26 (38.5) 7/18 (39) 3/8 (37.5)  Treated with tocilizumab 6/26 (23) 4/18 (22) 2/8 (25)  Grade >1 8 (18) 5 (15) 3 (27)  ICANS (any grade) 1 (2) 0 1 (9) Neutropenia <1 neutrophil x 10 9 /L 6/44 (14) 3/33 (9) 3/11 (27) Persistent lymphopenia 32 (71) 21 (62) 10 (91) Hypogammaglobulinemia (IgG <400 mg/dL) † 38 (84) 28 (82) 10 (91) Cotrimoxazole prophylaxis 44 (98) 33 (97) 11 (100) Valacyclovir prophylaxis 44 (98) 33 (97) 11 (100) Immunoglobulin replacement † 32 (71) 23 (68) 9 (82) Prior CMV reactivation 4 (9) 1 (3) 3 (27) ICANS, immune effector cell–associated neurotoxicity syndrome; Ig, immunoglobulin. ∗ Median time from CAR-T to BsAb therapy was 594 days (IQR, 513-670). † Hypogammaglobulinemia was assessed using the most recent measurement available before BsAb initiation and included patients receiving immunoglobulin replacement if it had been previously documented.

Baseline characteristics of CMV-seropositive patients with MM with 45 episodes receiving BsAbs

ICANS, immune effector cell–associated neurotoxicity syndrome; Ig, immunoglobulin.

Median time from CAR-T to BsAb therapy was 594 days (IQR, 513-670).

Hypogammaglobulinemia was assessed using the most recent measurement available before BsAb initiation and included patients receiving immunoglobulin replacement if it had been previously documented.

Eleven CMVr events were observed (24% of episodes), with an incidence rate of 3.13 cases per 100 person-months. Median time from BsAb initiation to CMVr was 48 days (IQR, 23-78). Seven episodes (7/45 [16%]) in 6 patients (6/39 [15%]) were CMV DNAemia, whereas 4 episodes (4/45 [9%]) in 4 patients (4/39 [10%]) were cs-CMVi (colitis, n = 1; esophagitis, n = 1; hepatitis, n = 1). Reactivation rates by target were BCMA 33%, FcRH5 26%, and GPRC5D 18%. The main characteristics of these episodes are shown in Table 2 . Blood CMV polymerase chain reaction (PCR) was positive in all patients. Fever prompted testing in 64% of episodes, sometimes accompanied by gastrointestinal symptoms (27%), cytopenias (18%), or liver test derangement (9%). Four cases (36%) were detected in nontargeted testing, being asymptomatic. Median peak viral load was substantially higher in cs-CMVi (9610 IU/mL; IQR, 1215-61 270 IU/mL) than DNAemia (789 IU/mL; IQR, 230-2239 IU/mL).

Table 2. Clinical and microbiological features, antiviral treatment, and outcomes of CMV reactivation episodes Characteristic Total n = 11, n (%) CMV DNAemia, n = 7, n (%) Clinically significant CMV infection, n = 4, n (%) Positive blood CMV PCR 11 (100) 7 (100) 4 (100)  Reason for PCR performance  Sporadic determination (asymptomatic) 4 (36) 3 (43) 1 (25)  Fever 7 (64) 4 (57) ∗ 3 (75)  + Gastrointestinal symptoms 3 (27) 1 (14) 2 (50)  + Cytopenias 2 (18) 1 (14) 1 (25)  + Liver function test abnormalities 1 (9) 0 1 (25)  CMV PCR load peak, median (IQR), IU/mL 1326 (290-2512) 789 (230-2239) 9 610 (1 215-61 270)  Time to first negative PCR, median (IQR), d 28 (15-36) 33 (15-39) 23 (22-36) Time from BsAb start to CMVr, median (IQR), d 48 (34-78) 48 (7-72) 61.5 (39.5-163.5) Bacterial coinfection 3 (27) 1 (14) 2 (50) CMVi CMV disease  End organ involved 1 (25)  Colitis 3 (75)  Esophagitis 1/3 (33)  Hepatitis 1/3 (33)  Classification 1/3 (33)  Possible 2/3 (67)  Proven 1/3 (33) Antiviral therapy  Valganciclovir 2 (50)  Foscarnet 2 (50) Related adverse events (nephrotoxicity) 1 (25) Treatment duration, median (IQR), d 21 (12-45.5) Concomitant hospitalization 2 (50)  Due to CMVr 6 (54.5) 4 (57) 1/2 (50) Infection resolution at day 30 1/6 (17) 0 3 (75) All-cause mortality at day 30 2 (18) 1 (14) 1 (25)  CMVr related 0 0 0 Recurrence of CMVr within 3 months (excluding death) 1/9 (11) 0 1/3 (33) MM relapse within 3 months (excluding death) 5/9 (56) 4/6 (67) 1/3 (33) Secondary antiviral prophylaxis (excluding death) 3/3 (100)  Duration, median (IQR), d 49 (34-63) ∗ Fever was attributed to alternative causes other than CMVr.

Clinical and microbiological features, antiviral treatment, and outcomes of CMV reactivation episodes

Fever was attributed to alternative causes other than CMVr.

All patients with cs-CMVi received antiviral treatment: valganciclovir (50%) or foscarnet (50%), with one switching to oral therapy. One foscarnet-treated patient developed nephrotoxicity. Median treatment duration was 21 days (IQR, 12-45). More than half of CMVr events (54.5%) occurred during hospitalization, but only 1 admission was CMV related. All-cause 30-day mortality after reactivation was 18% (95% confidence interval, 2.3-51.8). CMV recurrence within 3 months, excluding those who died, occurred in 11% of episodes. Among patients with cs-CMVi, 3 (75%) achieved clinical and microbiological resolution at day 30, and 1 patient (25%) died within 30 days from causes unrelated to CMV. All survivors received secondary prophylaxis with valganciclovir for a median of 49 days (IQR, 34-63). Univariate analysis suggested association between CMVr and persistent lymphopenia (91% vs 62%; P = .10) and prior CMVr (27% vs 3%; P = .04). Multivariate logistic regression did not identify statistically significant predictors ( supplemental Table 1 ).

In this retrospective cohort of heavily pretreated CMV-seropositive patients with MM receiving BsAbs, CMVr was a frequent early complication, occurring in nearly one-quarter of treatment episodes despite no routine screening. Although most events were self-limited DNAemia, cs-CMVi developed in 9% of episodes and was characterized by substantially higher viral loads. Persistent lymphopenia and prior CMVr emerged as potential risk signals, consistent with impaired CMV-specific cellular immunity. Our incidence, early timing (median, 48 days), and clinical spectrum align with growing evidence that T-cell–redirecting therapies increase CMVr risk compared with standard myeloma regimens. The reported data from clinical trials of BsAbs and early studies were variable, ranging from 3% to 22%, including both CMV DNAemia and cs-CMVi. 14 Systematic reviews and real-world cohorts reported higher infection rates with BCMA-targeting BsAbs and combination strategies, often noting viral reactivations in the first 1 to 3 months of therapy. 8 , 9 A similar pattern has been described in MM BCMA-directed CAR-T series, with notable rates of CMV DNAemia (17%-56%) and cs-CMVi (3%-15%), with corticosteroid exposure and cumulative immunosuppression identified as risk factors. 5 , 14

Several pathophysiological mechanisms plausibly explain heightened CMVr risk during BsAb therapy. BsAbs cause potent T-cell activation and, in many patients, subsequent T-cell exhaustion or depletion; loss of CMV-specific cellular immunity is central to loss of control of latent CMV. 6 Moreover, management of CRS with corticosteroids or other immunomodulators transiently impairs cellular immunity. 5 , 13 Age-related immunosenescence and cumulative prior therapies further limit immune reconstitution. 4 , 7 Clinicians should anticipate CMVr in this population. Early median onset and higher viral loads in patients with cs-CMVi might suggest a role for targeted surveillance with baseline CMV serology and periodic quantitative PCR during the high-risk period (eg, first 8-12 weeks), especially for patients with persistent lymphopenia or prior CMV events. Nevertheless, our data do not demonstrate improved outcomes with surveillance. Contemporary expert recommendations acknowledge the risk of viral reactivation in patients on T-cell–engaging therapies but lack consensus on universal prophylaxis or standardized monitoring. 10 , 11 , 15

Strengths of our work include detailed clinical characterization of CMV episodes, use of quantitative PCR data, and inclusion of consecutive BsAb-treated patients with MM. There is a clear knowledge gap and lack of studies specifically addressing CMVr in this population. However, several limitations temper interpretation. First, the retrospective design and nonstandardized testing protocol may have led to underascertainment or biased ascertainment of CMV DNAemia. Second, the small sample size limits statistical power. Third, the short median duration of BsAb therapy in our cohort limits the interpretation of late reactivation risk. Finally, we did not perform CD4 + T-cell counting or CMV-specific T-cell assays.

In conclusion, CMVr appears as a relatively common and potentially serious complication in CMV-seropositive patients with MM receiving BsAbs. This study could serve as an initial basis for recommending routine baseline CMV serology, a low threshold for PCR testing in symptomatic patients, and consideration of targeted PCR surveillance during the early treatment period for high-risk individuals (persistent lymphopenia and/or prior CMV). Prospective multicenter studies are needed to define optimal monitoring and prophylactic strategies.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Acknowledgments

Acknowledgments: The authors thank the Centro de Investigación Biomédica en Red de Enfermedades Infecciosas (CB21/13/00009) and University of Barcelona for providing the PhD student E.S.-E. with financial support. B.W.T. is supported by the Australian Government Medical Research Future Fund Investigator (Emerging leader Level 2) Fellowship (number 1195894).

Contribution: E.S.-E. and B.W.T. designed the research; V.Z. retrieved data; E.S.-E. collected and analyzed data; and all authors wrote the manuscript.

Footnotes

Data, including original anonymized data and response to specific questions, are available from the corresponding author, Enric Sastre-Escolà ( esastre@bellvitgehospital.cat ), on reasonable request.

The full-text version of this article contains a data supplement.

Supplementary Material

Supplemental Table

References

Associated Data

Supplementary Materials

Supplemental Table