CMV Infection After Letermovir Primary Prophylaxis Discontinuation in Allogeneic Hematopoietic Cell Transplant Recipients

Abstract

In this single-center study of 61 allogeneic hematopoietic cell transplant (HCT) recipients receiving letermovir primary cytomegalovirus (CMV) prophylaxis for the first 100 days, we report 23% incidence of clinically significant CMV infection during the first 100 days after letermovir discontinuation, predominately in haploidentical HCT recipients, without any associations with CMV-DNAemia under letermovir.

Graphical Abstract

Graphical Abstract.

This graphical abstract is also available at Tidbit: https://tidbitapp.io/tidbits/cmv-infection-after-letermovir-primary-prophylaxis-discontinuation-in-allogeneic-hematopoietic-cell-transplant-recipients

Since letermovir approval as primary CMV prophylaxis during the first 100 days after an allogeneic hematopoietic cell transplant (HCT), clinically significant CMV (csCMV) infections appear to be a relatively frequent event after letermovir discontinuation [1–6]. This could be attributed to ongoing immunosuppression and/or delayed CMV-specific T-cell immunity reconstitution [3]. Furthermore, the clinical significance of CMV DNAemia during letermovir administration and its potential impact on the development of CMV-specific T-cell immunity and hence protection against postletermovir csCMV infections remain to be defined. We performed a retrospective cohort study to describe the incidence of csCMV infection during the first 3 months after letermovir discontinuation and the potential associations with CMV DNAemia while on letermovir.

METHODS

Study Design, Patient Population, Objectives

This was a retrospective single-center cohort study of consecutive adult (≥18-year-old) allogeneic HCT recipients who received primary CMV prophylaxis with letermovir from May 1, 2019, to May 1, 2022. Adult allogeneic HCT CMV serology-positive (R+) recipients who received letermovir as primary CMV prophylaxis from post-HCT day 1, with a minimum of 3 months of follow-up postletermovir discontinuation, were included. Patients were excluded if they were still on letermovir by post-HCT day 200 or had csCMV infection before letermovir initiation. The study was approved by the local ethics committee, and all participants signed an informed consent form.

Cumulative incidence of csCMV infection during the first 100 days after letermovir discontinuation was described as the primary objective. The following secondary objectives were studied: (i) frequency of CMV DNAemia during letermovir administration and (ii) risk factors associated with csCMV infection during the first 100 days after letermovir discontinuation, including CMV DNAemia events during letermovir administration.

Institutional Practices, Definitions

Our institutional practices have been previously described and are detailed in the Supplementary Data [4–6]. Briefly, letermovir primary CMV prophylaxis was administered to all adult CMV serology donor-negative (D-)/R+ participants from May 1, 2019, to December 31, 2020, and starting January 1, 2021, to all adult CMV R+ allogeneic HCT recipients from post-HCT day 1 to day 100. Primary prophylaxis with letermovir was also administered to all adult CMV R+ with early (during the first 6 months post-HCT) grade ≥2 acute graft-vs-host disease (aGvHD) requiring corticosteroid treatment at a dose of ≥1 mg/kg/d and until tapering to <10 mg/d of prednisone equivalent since May 1, 2019. Breakthrough csCMV infection was defined as CMV DNAemia ≥150 IU/mL while on letermovir between May 1, 2019, and December 31, 2020, and ≥500 IU/mL after January 1, 2021 (based on new published evidence suggesting that low-grade CMV DNAemia in patients treated with letermovir could represent aborted viral replication, rather than effective viral replication), leading to CMV-preemptive treatment initiation [4, 5, 7]. For patients not receiving letermovir, csCMV infection was defined as any CMV DNAemia >150 IU/mL prompting initiation of preemptive treatment, based on institutional protocols. GvHD was defined based on international consensus guidelines [8].

Statistical Analysis

Continuous variables were described as median and range. Categorical and continuous variables were compared with the Fisher exact and a 2-tailed Student t test, as appropriate. Cumulative incidence was calculated for the first episode of csCMV infection, censoring for death. Logistic regression was used to identify risk factors for csCMV infection. Results are presented as odds ratios (ORs) with 95% CIs. Data were analyzed using STATA 16.0 (StataCorp, College Station, TX, USA).

RESULTS

Patient Population

From 80 originally identified patients, 61 patients were included in the study (Supplementary Figure 1). The patient baseline characteristics are detailed in Supplementary Table 1. Thirty (49%) patients were CMV D-/R+, 27 (44%) had a haploidentical donor, and most patients had a peripheral blood stem cell transplant (58, 95%) and received reduced-intensity conditioning (43, 70%). Twenty-eight (46%) patients were diagnosed with grade ≥2 aGvHD at a median (interquartile range [IQR]) of 44.5 (23–125) days from HCT, 16/28 (57%) with grade ≥2 gastrointestinal tract aGvHD. The median duration of letermovir administration (IQR) was 103 (96–127) days; letermovir was administered for >150 days in 10 patients. Letermovir administration was continued uninterrupted from day 0 to day ≥100 of prophylaxis in 25/61 (41%) patients, with a median duration (IQR) of 100 (100–107) days. Letermovir administration could be extended after post-HCT day 100 because of grade ≥2 aGvHD treated with high-dose corticosteroids, with a median duration (IQR) of 113.5 (96–160) days. Letermovir prophylaxis was temporarily interrupted either because of breakthrough csCMV (10/61 patients, 16%) or HHV-6 infection requiring preemptive treatment with valganciclovir (7/61 patients, 11%), with a median duration of prophylaxis (IQR) of 54 (35–94) days and 41 (21–54) days, respectively. Prophylaxis with letermovir was restarted in 6/10 patients with breakthrough csCMV infections, none of whom developed csCMV infection after discontinuation of letermovir, for a median follow-up (IQR) of 213.5 (217–301) days. In contrast, among the 4/10 patients who did not restart letermovir prophylaxis, 2/4 developed a second episode of csCMV infection within a median (IQR) of 33.5 (20–47) days after letermovir discontinuation. Letermovir prophylaxis was restarted in 5/7 patients with HHV-6 infection for a median (IQR) of 53 (52–83) days; letermovir was not restarted in 2/7 patients because they died at the time of HHV-6 infection diagnosis. Three of 5 patients who restarted letermovir prophylaxis developed a csCMV infection after letermovir discontinuation at a median follow-up (IQR) of 200 (200–224) days.

csCMV Infection

The overall incidence of csCMV infection within 100 days after discontinuation of letermovir prophylaxis was 23% (14/61) (Figure 1A). The median time between letermovir discontinuation and csCMV infection (IQR) was 54 (36–69) days. All 14 patients with csCMV infection were treated with valganciclovir for a mean duration (range) of 32.5 (13–101) days. None of the patients developed CMV end-organ disease. Cumulative incidence of csCMV infection was 37% (10/27) and 11.7% (4/34) for patients with haploidentical and nonhaploidentical donors, respectively (log-rank P = .02) (Figure 1B). Patients with a haploidentical donor tended to develop csCMV infection sooner (median [IQR], 47.3 [29–64] days) than the rest of the patients (median [IQR], 67.5 [46.5–88.5] days; P = .11) after letermovir prophylaxis discontinuation. Cumulative incidence of csCMV infection was 30% (9/30) and 16% (5/31) for patients in the CMV D-/R+ and D+/R+ groups, respectively (log-rank P = .20). There was no significant difference in csCMV infection incidence after letermovir discontinuation between the first part (January 2019–December 2020: breakthrough csCMV infection defined as CMV DNAemia while on letermovir ≥150 IU/mL) as compared with the second part of the study (after January 2021: breakthrough csCMV infection defined as CMV DNAemia while on letermovir ≥500 IU/mL; log-rank P = .09). The median CMV peak level at infection (IQR) was 1030 (298–1350) IU/mL vs 3195 (510–7670) IU/mL in the first and second parts of the study period, respectively (P = .06).

Figure 1.

A, Cumulative incidence of clinically significant CMV infection within the first 3 months after letermovir discontinuation for the overall patient population and (B) by haploidentical vs other donors (log-rank P = .02). C, Grid plot of CMV DNAemia for all patients during letermovir administration and for a minimum of 3 months of follow-up. In Figure 1C, each line represents 1 patient, and day 0 corresponds to the start of letermovir prophylaxis for all patients. Patients are grouped according to whether they developed a csCMV infection after letermovir discontinuation. Within these 2 groups, patients are organized according to viral activity in terms of CMV blips during 3-month follow-up. Each column represents days after HCT. Colors represent CMV DNAemia as indicated in the legend, and “x” represents the day of discontinuation of letermovir prophylaxis. Shaded orange-colored squares represent a breakthrough csCMV infection. Nonshaded orange-colored squares represent csCMV infection after letermovir discontinuation. Letermovir was restarted in 21 patients after a brief interruption, represented by shaded white squares with an “x” sign at the time of definitive letermovir stop. Abbreviations: CMV, cytomegalovirus; csCMV, clinically significant cytomegalovirus; HCT, hematopoietic cell transplantation; LET, letermovir.

Overall, 67% (41/61) of patients experienced ≥1 episode of detectable and/or quantifiable CMV DNAemia at <150 IU/mL during the study period. While on letermovir, CMV DNAemia remained <25 IU/mL in 42/61 (69%) patients, 19/61 (31%) had ≥1 quantifiable (25–149 IU/mL) CMV DNAemia, and 14/61 (23%) had a CMV DNAemia ≥100 IU/mL (Figure 1C). After letermovir discontinuation, 26/61 (43%) patients had CMV DNAemia ≥25 IU/mL, while in 35/61 (57%) patients CMV DNAemia remained <25 IU/mL during the 3-month follow-up.

Risk Factors

Univariable analyses were performed with clinically significant variables, including CMV DNAemia during letermovir prophylaxis, in order to identify risk factors for csCMV infection after stopping letermovir prophylaxis. There were no associations between CMV DNAemia, neither in frequency nor in value, and csCMV infections (Supplementary Table 2). Haploidentical donor was identified as the only significant predictor of csCMV infection after letermovir discontinuation (OR, 4.41; 95% CI, 1.2–16; P = .03) in univariable analyses. Multivariable analyses were not performed considering the limited number of cases.

DISCUSSION

Our data suggest that 1 in 4 allogeneic HCT recipients may develop csCMV infection within the first 100 days after letermovir prophylaxis is discontinued, predominately observed in patients with haploidentical donors. Clinically significant CMV infection after letermovir discontinuation was reported in 19% of allogeneic HCT recipients who received primary CMV prophylaxis with letermovir in the pivotal clinical trial [1]. Our findings are consistent with those of the clinical trial and with other real-life data showing that CMV DNAemia after letermovir discontinuation may be a relatively frequent problem in clinical practice [4, 9, 10]. This may, in part, be attributed to the delayed development of CMV-specific T-cell immunity under letermovir, resulting from lower CMV antigenemia and early CMV preemptive treatment initiation while on letermovir. The importance of CMV-specific T-cell immunity to control CMV replication after an allogeneic HCT has been well demonstrated, and its role in the management of CMV DNAemia and infection in this patient population will need to be better defined and incorporated into clinical practice in the near future [3, 11, 12]. It is likely that the low CMV DNAemia cutoffs used at our center to initiate preemptive treatment might have contributed to early and effective suppression of CMV replication and hence limited development of CMV-specific T-cell imunity [3]. Initially, the threshold for CMV preemptive treatment initiation at our center was not adjusted after introduction of letermovir prophylaxis and was maintained at plasma CMV DNAemia >150 IU/mL. However, a single-center noncontrolled prospective study suggested that plasma CMV DNAemia as high as 1000 IU/mL may merely represent aborted CMV replication rather than infectious virion production [7]. Following those findings, we increased the plasma CMV DNAemia cutoff for preemptive CMV treatment initiation to >500 IU/mL in the latter part of the study period [4, 5]. Nevertheless, this might have remained a low enough cutoff for preemptive treatment initiation to allow for CMV-specific T-cell immunity stimulation, which could potentially explain why we were not able to demonstrate any potential associations between any level of CMV DNAemia while on letermovir and postletermovir csCMV infection rates. Different thresholds of CMV DNAemia above which preventive treatment should be started have been described in the literature, ranging from 300 to 1000 copies/mL (∼274–909 IU/mL) for low-risk patients and ∼150 copies/mL (137 IU/mL) for high-risk patients, although this decision is a combination of assessment of patient risk and close monitoring of CMV DNAemia kinetics [1, 13–17]. The cutoff of preemptive treatment initiation in patients with CMV DNAemia while on letermovir remains an open question in the field.

The cumulative incidence of csCMV infection was >3 times higher in patients with a haploidentical donor. Receiving an HCT from a haploidentical donor is a well-known risk factor for csCMV infection in allogeneic HCT recipients [18–21]. This may, in part, be associated with the GvHD prophylaxis these patients receive, namely post-transplant cyclophosphamide administration, and higher rates of GvHD. More data are required to assess whether letermovir prophylaxis prolongation in higher-risk patients, such as HCT recipients from haploidentical donors, may be an effective and safe approach. Extending letermovir prophylaxis beyond post-HCT day 100 as a universal approach in all patients remains to be proven. While waiting for the results of the 100 vs 200 days of letermovir primary prophylaxis clinical trial in allogeneic HCT recipients to be published, our data show through a very intensive monitoring that more than half of patients had CMV DNAemia <25 IU/mL for 3 months after letermovir discontinuation, suggesting that perhaps not all patients benefit from more prolonged prophylaxis courses (Figure 1C). Those findings may initiate the discussion for risk stratification rather than a blanket approach of universal prophylactic strategy.

Despite its limitations—including its small size, retrospective cohort study design, limitations in the identification of risk factors for csCMV infection due to the small event numbers, and lack of relevant immunology data—our findings complement the existing literature, showing a relatively high proportion of allogeneic HCT recipients with CMV DNAemia after prophylaxis discontinuation, predominately in patients with haploidentical donors. The optimal duration of primary CMV prophylaxis with letermovir and the CMV DNAemia cutoff for preemptive CMV treatment initiation in patients on letermovir remain open questions urgently requiring answers.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.