Maribavir is an oral benzimidazole riboside antiviral which has been in development for many years, but only recently became available as therapy for R/R CMV. It inhibits viral UL97 kinase and thus interferes with multiple pathways including nuclear egress of CMV viral capsids. It has no significant renal, hematologic, or hepatic toxicity; its most common adverse effect is dysgeusia. Early trials for prophylaxis in stem cell transplant [31] and liver transplant recipients [32] failed to show efficacy, likely because the dose selected, 100 mg twice daily, was too low [33]. However, a case series of six patients with R/R CMV treated with compassionate use maribavir at doses of 400–800 mg twice daily showed striking responses in several patients [34]. This, and the toxicity of other agents available for R/R CMV, spurred the performance of a Phase 2 trial of 3 dosing regimens for maribavir (400, 800, and 1,200 mg twice daily) among SOT and HSCT recipients [19]. This study demonstrated clearance of CMV DNAemia at 6 weeks of therapy in 70%, 63%, and 68%, respectively, in this highly treatment-experienced population [19].

Subsequently, a multicenter Phase 3 trial of maribavir versus investigator-assigned therapy (IAT) was performed involving 352 SOT and HSCT recipients in a 2:1 randomization [8]. IAT, which could be ganciclovir, valganciclovir, foscarnet, cidofovir, or a combination of these, was chosen as the comparator because of patients’ varied treatment histories. The primary endpoint, confirmed CMV-DNA clearance at the end of week 8, was achieved by 55.7% in the maribavir arm vs. 23.9% in the IAT arm (p > 0.001). The key secondary endpoint, a composite of CMV-DNA clearance and symptom control at the end of week 8 maintained through week 16, was achieved by 18.7% vs. 10.3% (p = 0.01). Dysgeusia was the most frequent adverse effect in the maribavir group (37.2%); the maribavir group also had significantly less neutropenia than the val/ganciclovir group and less acute kidney injury than the foscarnet group [8]. These results led to the approval of maribavir by the US FDA in 2021 for treatment of post-transplant CMV infection/disease in patients age 12 and older, that is refractory (with or without genotypic resistance) to treatment with ganciclovir, valganciclovir, cidofovir or foscarnet, with a similar authorization by the EMA in 2022. A second Phase 3 randomized double-blinded trial (the AURORA trial, NCT02927067) compared maribavir to valganciclovir for treatment of asymptomatic CMV DNAemia in stem cell transplant recipients. At the time of this writing, full results have not yet been published, but topline results were announced by the study sponsor (Takeda) in December 2022. At week 8, which was the end of study treatment, 69.6% of patients treated with maribavir achieved CMV clearance vs. 77.4% for valganciclovir; this did not meet non-inferiority based on a prespecified margin of 7%. At week 16, 52.7% of patients treated with maribavir achieved maintenance of viremia clearance and symptom control vs. 48.5% for valganciclovir. Similar post-treatment maintenance effect was observed at week 12 (59.3% vs. 57.3%) and week 20 (43.2% vs. 42.3%) time points. Maribavir’s safety profile was confirmed, particularly with regards to neutropenia (21.2% vs. 63.5% for valganciclovir). Despite not meeting the prespecified noninferiority margin, this study demonstrated that maribavir has potential utility for treatment of non-refractory CMV DNAemia, with a lower risk of hematologic toxicity than valganciclovir.

While the approval of maribavir for R/R CMV was long-awaited, questions about optimal use remain. In the Phase 3 R/R CMV trial, subgroup analyses showed that the proportion achieving the primary endpoint was higher when maribavir was initiated at a viral load of <9100 IU/mL than at higher viral loads (62.1% vs. 43.9%), and was higher with documented genotypic resistance vs. refractory CMV without resistance (62.8% vs. 43.8%) [8]. Some experts have proposed that R/R CMV with high viral load might most effectively be treated with an agent such as foscarnet initially, then switch over to maribavir at a lower viral load, to minimize foscarnet toxicity and to maximize the efficacy of maribavir [35]. Another issue, as with all therapies for R/R CMV, is the risk for recurrences. While maribavir achieved the key secondary endpoint significantly more often than IAT, the numbers in both groups were relatively low (who maintained CMV clearance and symptom control out to week 16 after completion of therapy at week [8]). Of note, the Phase 3 R/R maribavir trial [8] did not permit secondary prophylaxis after the defined 8 weeks treatment period, whereas the Phase 2 R/R maribavir study had allowed continuation of maribavir out to 24 weeks [19]. Whether secondary prophylaxis would be of benefit (in terms of decreasing recurrences after CMV DNAemia clearance), and whether that would be offset by potential increases in maribavir resistance, has yet to be studied, but will be important to assess. Although the evidence supporting the use of secondary prophylaxis is mostly lacking, many centers use secondary prophylaxis, and current guidelines recommend considering secondary prophylaxis in high-risk scenarios [10]. Combination therapy with maribavir is also a promising frontier that is yet to be explored. Chou et al. demonstrated that the maribavir/ganciclovir combination is antagonistic, and additive for maribavir + foscarnet or cidofovir or letermovir, but synergistic for maribavir + rapamycin (sirolimus) [36]. The use of an mTOR inhibitor-based immunosuppressive regimen is another strategy in prevention or management of R/R CMV particularly in organ transplant recipients [37]. The maribavir + mTOR inhibitor combination deserves further study.

Perhaps the most important questions regarding its future utility relate to the risk for development of resistance to maribavir. An impressive body of work by Chou has addressed this issue for nearly 20 years, now utilizing updated sequencing technology [38]. Chou et al. analyzed resistance mutations from the Phase 2 maribavir trials, and found known UL97 maribavir resistance mutations after 46–166 days of maribavir therapy (T409M or H411Y) in 17 of 23 who had had CMV recurrences while on maribavir [39]. Moreover, they identified the mutation UL97 C480F in six patients, which confers high-level maribavir resistance and low-level ganciclovir resistance [39]. A recent real-world case series described maribavir resistance in 4 of 13 patients treated for R/R CMV (with H411Y in 2, T409M in 1, and C480F in 1) [40]. Another report described two patients refractory to maribavir, one with H411Y and one without known maribavir resistance mutations [41].

Maribavir has far less toxicity than other agents for R/R CMV, and is a major advance in treatment of this entity. However, we still have much to learn about optimizing its use and preventing recurrences and resistance.

Letermovir is a 3,4-dihydroquinazoline derivative and is an inhibitor of the viral terminase complex, mainly at the pUL56 subunit. Terminase inhibition leads to compromised viral replication by inhibiting the cleavage of genome particles to units of proper length and accumulation of immature viral DNA [29]. Based on the mechanism of action, letermovir is selectively active only against CMV, and mechanism-derived adverse effects are unlikely. Letermovir was approved in 2017 for prophylactic use in adult CMV-seropositive allogeneic hematopoietic stem cell transplant (HCT) recipients, where it has shown good efficacy in the placebo-controlled phase III trial [42] and as of 6 June 2023, the US FDA approved letermovir for the new indication of CMV prophylaxis in D+/R− kidney transplant recipients, based on the results of the Phase 3 trial [43]. No statistically significant differences were seen in the frequency or severity of any adverse events between letermovir and placebo, although gastrointestinal adverse events (such as nausea) were slightly more common in the letermovir group. It is available in both peroral (PO) and intravenous (IV) formulations. The standard dose is 480 mg daily (IV/PO) when used as prophylaxis. However, due to interaction via the hepatic drug transporter organic-anion-transporting polypeptide (OATP), cyclosporine increases bioavailability of letermovir, and dose reduction to 240 mg daily is recommended [43].

In the phase 3 trial, 601 CMV D+/R− adult kidney transplant recipients were randomized to receive prophylaxis with either valganciclovir or letermovir 480 mg once daily (240 mg if used with cyclosporine) until week 28 after transplantation. Primary efficacy endpoint of the study was met, as letermovir was non-inferior to valganciclovir in preventing CMV disease (frequency 10.4% in the letermovir vs. 11.8% in the valganciclovir group). Importantly, letermovir resulted in lower toxicity compared to valganciclovir, especially lower rate of leukopenia (11.3% vs. 37%) or neutropenia (2.7% vs. 16.5%), and lower rate of drug discontinuation due to adverse events (4.1% vs. 13.5%) [43]. The study results are very convincing for the good efficacy of letermovir also in the SOT setting, when used as prophylaxis, and have recently led to the expanded indication mentioned above, by the US FDA.

Larger industry-driven studies have all addressed the use of letermovir only as CMV prophylaxis, but due to lack of suitable alternatives for treating resistant CMV infections until recently, there has similarly been interest on using letermovir for treatment of CMV infections. However, as the drug does not block viral DNA synthesis, but inhibits events later in the viral cycle, some concerns have been raised about the potential to promote resistant viral strains, especially when used in case of high-level viremia. Indeed, several mutations in the pUL56 subunit of the terminase complex have been described after exposure to letermovir, potentially causing resistance to the antiviral action of the drug [44]. Interestingly however, in the phase 3 kidney transplant trial, no letermovir resistance-associated substitutions/mutations were detected in the letermovir arm, in comparison to nine patients in the valganciclovir arm, who developed ganciclovir resistance-associated mutations [45].

Table 2 briefly summarizes published case series of studies using letermovir as treatment of CMV infections. Most common dose has been 480 mg once daily PO, but also higher doses (up to 960 mg daily) have been used. In these studies, 76% of the cases with CMV infection treated with letermovir resulted in either viral clearance or decrease to viremia <200 IU/mL, and treatment failure was seen in 24% of cases. Although letermovir was mainly effective and resulted in lowering of viremia or viremia clearance, recurrent infections were common. In the multicenter retrospective study by [46], viral suppression was more likely when letermovir was started at a viral load of <1000 IU/mL. Therefore, another option worth considering would be to treat the viral load to low levels with another agent such as foscarnet, and then switch to letermovir to maximize the chance of clearance and minimize foscarnet toxicity.

Significant interaction with tacrolimus was noted, and tacrolimus dose needed to be adjusted (reduced significantly) in many cases. Letermovir is a moderate inhibitor of CYP3A in vivo [54], and therefore leads to increase in tacrolimus and cyclosporine (and sirolimus) concentrations. In phase 1 studies, coadminstration of letermovir with tacrolimus or cyclosporine resulted in 2.4- and 1.7-fold increases in area under the plasma concentration-time curves, and 1.6- and 1.1-fold increases in maximum plasma concentrations, respectively [55].

The use of letermovir as an antiviral agent in preemptive therapy after solid-organ transplantation has been so far addressed in only one early proof-of-concept phase 2a study, in which antiviral efficacy was shown despite using much lower doses than the current recommendation (only 80 mg/day) [56]. Some more experience of successful use of letermovir as preemptive therapy after HSCT has been described [57].

Combination therapy with letermovir and (val)ganciclovir or CMV IvIG has also been reported. In the largest study reporting combination therapy so far, eight kidney or kidney-pancreas recipients with persisting low-level viremia despite >90 days of valganciclovir were treated with valganciclovir 900 mg twice daily together with letermovir 480 mg once daily. In this study, the use of adjunctive letermovir did not result in viral clearance, and median viral load did not change during 12 weeks of follow-up.

Suggested or confirmed genotypic resistance to letermovir was described in some of the case series, and in addition in case reports. In total at least seven genotypically resistant cases have been published to date after solid-organ transplantation, with mutations seen in UL56 gene [46, 47, 51, 58]. Similarly, mutations in UL56 have been described in patients who received letermovir prophylaxis after HSCT [59]. However, the vast majority of CMV infections treated with letermovir have not resulted in resistance concerns.

Based on the published experience so far and our own clinical experience, letermovir can be considered for treatment of R/R CMV infections. Favorable results will more likely be reached if treatment is initiated at low-level viremia, but recurrence and development of resistance are remaining concerns. In cases of poor tolerance to valganciclovir due to leukopenia or neutropenia, the potential to use letermovir as secondary prophylaxis after clearance of viremia could be further explored. However, some concerns about breakthrough infections and emergence of letermovir resistance have been raised in small case series [52, 60].

T cell immunity is essential for CMV control [61, 62]. In SOT recipients, T cell immunity is weakened by immunosuppressive drugs, making direct restoration of immunity by infusion of CMV-specific T cells (“adoptive” T cell therapy) attractive [63].

To date, most clinical data on CMV-specific T cell therapies derive from phase 1/2 studies in allogeneic HCT recipients in which cells were infused for CMV-prophylaxis or treatment of R/R CMV [64]. Different protocols for T cell generation and application including intrathecal administration [65] were demonstrated to be safe and treatment for R/R CMV was successful in around 70% [64, 66]. Despite these promising data, the safety and efficacy still need to be confirmed in phase 3 studies. Additionally, there is very little data on SOT recipients.

Traditionally, CMV-specific T cells were harvested from the HCT donor. This limited the treatment to HCT recipients with CMV seropositive donors. More recently, peripheral blood cells from only partially HLA-matched CMV seropositive third-party donors were also successfully used [67]. This enabled therapy also in SOT recipients. Third-party cells were either collected prior and stored for “off-the-shelf” use [67] or collected upon request from pre-screened individuals in donor registers [68, 69]. Despite concerns about limited proliferative capacity due to continued immunosuppression, studies have shown successful expansion of autologous virus-specific T cells [70–73].

Ex vivo steps are required to exclusively select CMV-specific T cells from the original donor product [64]. Complex and time intensive laboratory expansion protocols of minimum 10 days but up to 30 days are used to obtain high numbers of specific T cells [72, 74]. Alternatively, CMV-specific donor-derived white blood cells are directly isolated ex vivo using immunomagnetic methods (e.g., direct sorting using peptide-HLA multimers, cytokine-capture system or based on T cell activation molecules) [75–77].

Adoptive T cell therapies are still mainly restricted to specialized academic centers and few commercial companies due to the complexity of donor search and selection and the requirement of “good manufacturing practice”-accredited laboratories to prepare the cells in vitro. However, in recent years, increasing number of centers were able to offer “off-the-shelf” products to their patients as part of multicentric trials (e.g., NCT04390113 and [67]).

Virus-specific adoptive T cell therapies are generally reported to be safe. For allogeneic products, graft-versus host disease is a potential concern despite viral-specificity of most cells and was reported in around 5%–16% [64]. Independent of cell source, cytokine release syndrome and graft failure due to T cell mediated inflammation may occur but have rarely been reported [73, 78]. An open issue is the co-administration of immunosuppressive drugs, which affects the expansion and function of T cells in vivo after infusion into the patient. The optimal timing and composition of immunosuppression at the time of virus-specific T cell infusion remains to be determined.

At this time, data from 19 SOT recipients treated with CMV-specific T cells have been reported, including one pediatric patient of 16 years of age, 11 lung, 6 kidney, 1 heart, and 1 liver transplant recipient (Table 3, including one unpublished case from our institution) [69–73, 79]. All recipients were treated for R/R CMV infection (n = 5) or disease (present or recent, n = 14). Anti-CMV drug resistance was reported in 12 cases. All protocols collected T cells from peripheral blood and most used ex vivo expanded cells. At our institution, we have successfully used the cytokine-capture system to isolate CMV-specific T cells.

Sixteen patients received autologous T cells and interestingly, it was possible to harvest CMV-specific T cells from patients with CMV D−/R− and D+/R− serostatus at time of transplantation [72]. In one patient, the immunosuppressive treatment was reduced specifically for cell harvesting, and the authors recommended this measure 2–3 weeks prior to cell collection [70].

Three patients received fully or partially HLA-matched third-party allogeneic T cells; our patient received the cells from his HLA-matched daughter, the pediatric patient received cells from his mother who was not the SOT donor [79], and another patient received cells from a third-party donor who was selected from a donor registry [69].

One to six doses of CMV-specific T cells were infused per patient with single doses between 0.24 × 10⁷ and 3 × 10⁷ cells. After infusion, some trials observed rapid in vivo expansion of CMV-specific T cells with simultaneous drop in viral load [73], however, other protocols could not observe these dynamics [70].

Infusions were generally well tolerated. Smith et al observed in their case series only grade 1 and 2 adverse events with potential association to the T cell infusion [72]. No graft-versus-host disease was observed with the allogeneic products, however, one patient had a mild fever following infusion which was potentially associated with cytokine release [69]. Of note, in the very first reported case, a lung transplant recipient with a drug-resistant CMV pneumonia on mechanical ventilation initially responded clinically and virologically after a first infusion of autologous CMV-specific T cells, could be discharged, and received a second infusion for prophylaxis, however, he subsequently died few weeks later from CMV-negative graft failure and it was not possible to fully exclude an association with the T cell therapy [73]. No changes in graft status were observed in the other cases.

As cases were not controlled and concomitant antiviral-drug regimen were often present, larger and controlled studies are necessary to estimate and prove treatment efficacy (e.g., as for BK virus in kidney transplantation, NCT04605484).

In summary, CMV-specific adoptive T cell therapy is an appealing option for R/R CMV in SOT. However, safety and efficacy need to be confirmed in controlled trials. Additional data is needed to identify the best protocols in terms of T cell generation and optimal time point of application and the influence of different immunosuppressive therapies on treatment efficacy should be investigated. At this point, we recommend that CMV-specific T cell therapies should be preferentially offered within clinical trials in order to close the knowledge gaps.