AUTHORS
Sean Bowden, MD, Jaewon Lee, MD2, Maria-Jose Hernandez-Woodbine3, Pedro Luiz Lage Bodour Danielian, MD, MPH4, Ishan Iyengar Bhatia, MD5, Victoria Rossetti6, Faheem Nagpurwala7, Carolyn Alonso, MD8, Poorva Bindal, MD9
AFFILIATIONS
1 University of Massachusetts, Worcester, Massachusetts, United States
2 University of Massachusetts, Worcester, Massachusetts, United States
3 Norwalk Hospital, Norwalk, Connecticut, United States
4 University of Connecticut, Farmington, Connecticut, United States
5 Mount Sinai, New York, New York, United States
6 University of Massachusetts, Worcester, Massachusetts, United States
7 University of Massachusetts, Worcester, Massachusetts, United States
8 Beth Israel Deaconess Medical Center, Boston, Massachusetts, United States
9 University of Massachusetts, Worcester, Massachusetts, United States
ABSTRACT
Cytomegalovirus (CMV) DNAemia is increasingly recognized after chimeric antigen receptor (CAR) T-cell therapy, but its frequency, clinical severity, and prevention practices remain incompletely characterized. CAR T-cell recipients may be particularly vulnerable because of prior cytotoxic therapy, lymphodepletion, prolonged cytopenias, corticosteroid exposure, and delayed immune reconstitution. Published studies have used heterogeneous surveillance strategies and variable definitions of CMV DNAemia, clinically significant infection, and tissue-invasive disease, limiting interpretation of the available evidence. We conducted a systematic review and meta-analysis to estimate the pooled proportions of CMV DNAemia, CMV infection, and tissue-invasive CMV disease following CAR T-cell therapy and to describe reported CMV-directed prophylaxis practices.
Records identified through a prespecified search strategy were screened in Covidence by two independent reviewers, with disagreements resolved by additional reviewers. Eligibility criteria included studies reporting CMV occurrence after CAR-T therapy. Outcomes were categorized as CMV DNAemia, CMV infection, or tissue-invasive CMV disease according to consensus guidelines, defined as follows: CMV DNAemia was defined as the detection of CMV DNA within the bloodstream, whether replicating virus or not; CMV infection was defined as the detection of CMV in blood with attributable symptoms; and tissue-invasive CMV was defined as evidence of CMV infection causing end-organ damage. Random-effects meta-analyses of study-level event proportions were performed on the logit scale using restricted maximum likelihood estimation and modified Knapp-Hartung confidence intervals. Heterogeneity was assessed using I², and risk of bias was evaluated using the Newcastle-Ottawa Scale. Because follow-up duration and surveillance intervals varied, pooled estimates were interpreted as proportions observed over study-specific follow-up rather than fixed-time cumulative incidences.
Of 37,286 records identified, 15,861 underwent title and abstract screening, 158 full-text articles were assessed, and 38 studies were included. Among these, 31 studies comprising 5,894 patients reported 452 episodes of CMV DNAemia, yielding a pooled proportion of 7% (95% CI, 4%–12%; I²=93.2%). Ten studies comprising 2,333 patients reported 52 CMV infections, with a pooled proportion of 2% (95% CI, 1%–8%; I²=89.1%). Twelve studies comprising 897 patients reported 20 cases of tissue-invasive CMV disease, corresponding to a pooled proportion of 2% (95% CI, 1%–3%; I²=0%). CMV DNAemia was therefore more frequently reported than clinically defined infection or tissue-invasive disease. Heterogeneity was substantial for DNAemia and infection but absent for tissue-invasive disease, reflecting variability in surveillance intensity, testing thresholds, baseline CMV serostatus, patient populations, CAR T-cell targets, and outcome definitions. Thirty-two studies were rated as good quality and six as fair quality; none were rated poor. CMV-specific prophylaxis and preemptive monitoring strategies were inconsistently defined and reported and could not be quantitatively pooled.
CMV DNAemia occurred in approximately 1 in 14 patients (7.1%) following CAR T-cell therapy, whereas clinically defined CMV infection and tissue-invasive disease were each reported in approximately 1 in 50 patients (2%). The higher frequency of DNAemia compared with invasive disease suggests that CMV detection does not uniformly translate into clinically severe disease; however, substantial heterogeneity limits the precision and generalizability of estimates for DNAemia and infection. These findings highlight the need for prospective studies using standardized definitions, prespecified surveillance windows, and consistent reporting of baseline serostatus, antiviral prophylaxis, corticosteroid exposure, and immune-reconstitution measures. Such studies are needed to identify patients at greatest risk and determine whether risk-adapted CMV monitoring, preemptive therapy, or prophylaxis improves outcomes after CAR T-cell therapy.
DISCLOSURES
Ishan Iyengar Bhatia, MD: No relevant conflicts of interest to declare.