Real-world use of letermovir for CMV prevention in allogeneic hematopoietic cell transplantation recipients in France: a national claims database study

Introduction

Cytomegalovirus (CMV) is a human herpesvirus with a worldwide seroprevalence of 45%–100%, notably among older populations (1, 2). It establishes a lifelong latent infection following primary infection that can periodically reactivate. Although typically mild in immunocompetent individuals, it can cause life-threatening complications in immunocompromised individuals, notably in transplanted patients (3, 4). In the case of allogeneic hematopoietic cell transplantation (HCT), CMV infection has been observed in 10%–40% of patients, which is associated with a high risk of CMV-related complications or graft dysfunction, requiring rehospitalization (5, 6).

Before 2018, in France, the main approach to CMV prophylaxis for HCT recipients consisted of preemptive treatment of CMV viremia with ganciclovir, cidofovir, or foscarnet (in cases of antiviral resistance). To date, letermovir is the only drug indicated for prophylaxis against CMV reactivation in CMV-positive HCT recipients, with no hematological toxicity when compared with preemptive therapies and with good overall tolerance (2, 7). It is recommended that letermovir should be started on the day of HCT and no later than 28 days post-HCT, usually around the fifth day, and continued for 100 days post-HCT and potentially beyond. Prolonged letermovir prophylaxis beyond 100 days post-HCT may be of benefit in certain patients at high risk of late CMV reactivation (8, 9). In France, letermovir has been available since March 2020 through the general reimbursement scheme (10).

Due to its recent availability, real-world data on letermovir prophylactic use are scarce, notably in the context of recent studies advocating for an extension of the prophylaxis duration up to 200 days (8, 9).

In order to depict the current use of letermovir in France, this study aimed to i) describe the number of patients with letermovir prophylaxis in France, and their main sociodemographic and clinical characteristics, and ii) describe letermovir prophylaxis modalities among HCT recipients, notably in terms of initiation settings, prophylaxis duration, dosage, discontinuation, and re-exposure, and clinical events of interest during the non-exposed periods.

Methods

Design and patient selection

This is a non-interventional, retrospective cohort study based on claims data from the French National Health Data System (SNDS).

The SNDS is the French national claims database, which gathers all reimbursed healthcare utilization from the inpatient and outpatient databases of nearly 99% of French residents, affiliated with one of the compulsory health insurance plans, in a pseudonymized manner (11). The data collected notably include sociodemographic information on the beneficiaries (e.g., age, sex, area of residence, exemption codes, and vital status), information on the reimbursed ambulatory healthcare services (e.g., date and type of medical and paramedical care, care settings, biological and imaging tests, medical devices, and dispensed drugs), and information on those performed in healthcare facilities [i.e., type of facility, date and type of entry and discharge, diagnoses, procedures, and expensive drugs billed on top of the Diagnosis-Related Group (DRG)].

Data on every patient with evidence of at least one dispensation of letermovir, identified based on the international Anatomical Therapeutic and Chemical classification (ATC) code J05AX18, between 1 March 2020 (when letermovir became available in France) and 31 December 2021 (the latest year of data available at the time of access), were extracted. Data on patients presenting any of the following criteria were excluded: i) absence of healthcare consumption within 12 months before letermovir initiation, to ensure exhaustive capture of patients’ clinical histories and to exclude those with transient residence—and disease management—outside of France; ii) coverage by state medical aid [aide médicale d’état (AME), granted to migrant patients before their inclusion in a compulsory insurance scheme] to avoid double counting when patients were included in the compulsory insurance scheme; iii) letermovir initiation during the temporary use period due to non-exhaustive capture in the SNDS; and iv) identity unicity in the SNDS not guaranteed (e.g., twins), to avoid missing information.

Consistent with letermovir indication and current practices (10), letermovir treatment modalities (i.e., initiation settings, overall treatment duration, and treatment dosage) were described among a subgroup of patients who underwent an HCT in the month before or after letermovir initiation and with at least 6 months of follow-up (or death if earlier), referred to as 6M-HCT+ patients.

Finally, the number of letermovir courses, the course duration, and the clinical events of interest were described in 6M-HCT+ patients whose first course of letermovir was identified as complete (i.e., the discontinuation date was formally identified before the end of the patient’s follow-up, including death). These patients were labeled as complete coverage patients.

Patient characteristics were described among all three populations at letermovir initiation. Medical history was described using a 2-year historical period before the index date. Patients in the 6M-HCT+ and complete coverage groups were followed from the index date to the end of the study (31 December 2021) or death, whichever occurred first (Figure 1).

Outcomes and variables

The first outcome was the description of the sociodemographic and clinical characteristics of patients with letermovir prophylaxis. The sociodemographic characteristics gathered included age, sex, and region of residence. The clinical characteristics encompassed comorbidities of interest (e.g., diabetes, chronic pulmonary disease, cerebrovascular disease, inflammatory disease, cardiovascular disease, solid tumor, psychiatric disease, hepatic impairment, and end-stage renal disease), the Charlson comorbidity index, the underlying disease at the time of letermovir initiation (e.g., leukemia, lymphoma, myeloma, or myelodysplastic syndrome), and a history of HCT (excluding index HCT) or solid organ transplantation (SOT) before letermovir initiation.

The second outcome was the description of letermovir prophylaxis modalities. This included the care setting at letermovir initiation, the time from HCT to letermovir initiation, the dosage, and the first course discontinuation-free survival. Letermovir prophylaxis modalities were assessed among 6M-HCT+ patients.

The third outcome was the description of the letermovir courses. This included the duration of crude prophylaxis (overall and by care setting), the presence of a second course of letermovir, and the duration of the second course. Letermovir courses were described among complete coverage patients to ensure a distinction between courses.

The final outcome was the description of the clinical events of interest. These clinical events were hospitalizations (all-cause and CMV-related), dispensations of CMV antiviral therapy, and medical visits (all specialties). A course was considered complete when discontinuation was observed before the end of follow-up (including death). For patients with only one course, the non-exposed period corresponded to the time between the end of the first course and the end of follow-up. For patients with two courses, the non-exposed period was defined as the time between the end of the first course and the beginning of the second one. Clinical events were described among patients with complete coverage to identify non-exposed periods.

The codes and algorithms for the selection of patients and the definitions of the variables were derived from validated algorithms and are available in the Supplementary Material.

Statistical analyses

Continuous data were summarized as the mean (standard deviation, SD) or the median (interquartile range, IQR), while categorical variables were presented as numbers and frequencies. Analyses were carried out using SAS^® version 9.4 (SAS Institute Inc., Cary, NC, USA).

Sociodemographic characteristics were described at letermovir initiation and the medical history over a 2-year historical period. First course of letermovir discontinuation-free survival and the prophylaxis duration were described among the 6M-HCT+ patients and the complete coverage subgroup, respectively.

The letermovir dosage is one tablet/injection per day, regardless of the treatment strength (240/480 mg). Injectable letermovir cannot be identified in the SNDS as it is only available in the hospital and billed within the DRG. Sensitivity analyses were conducted to account for potential hospital use of the injectable formulation; however, based on internal data and expert opinion, its use could be considered anecdotal. During hospitalizations, letermovir tablets (240/480 mg) were identified by tablet, with a dispensing date for each tablet. In community pharmacies, letermovir is dispensed in 28-tablet boxes, with a dispensing date for each box, which was assumed to cover 28 days.

A course of letermovir was defined as the number of days covered by each tablet/box dispensed with no treatment discontinuation. Discontinuation was defined as the absence of a new letermovir dispensation for a period of at least 1 month after the period covered by the last dispensation of the course. For instance, for a dispensation of one box, discontinuation was considered after 58 days (28 + 30). Thereafter, the course duration spanned from letermovir initiation to the end of the period covered by the last dispensation before discontinuation. In the case of a new dispensation identified after this period, a new course of letermovir was considered (i.e., second course), corresponding to a second letermovir exposure, and the same rules applied. Subsequent courses were not assessed (Figure 2).

Among the 6M-HCT+ patients, the discontinuation-free survival for the first course was estimated using the Kaplan–Meier method, with administrative censoring on 31 December 2021, and was displayed with 95% confidence intervals (95%CI). In the complete coverage group, prophylaxis duration was assessed as the crude number of days between letermovir initiation and the discontinuation of each course among patients whose first course was identified as completed. A course was considered complete when the end of the discontinuation period was observed before the end of follow-up (including death) (Figure 2). Clinical events of interest were described as the number and proportion of patients presenting at least one event of interest, along with the monthly number of events (i.e., the number of events divided by the duration of the non-exposed period).

Ethics

The study was designed in accordance with the International Society for Pharmacoepidemiology (ISPE) guidelines and with the principles set out in the Declaration of Helsinki.

As part of the regulatory submission for claims studies based on French SNDS data, this study was approved by the French Ethics and Scientific Committee (Comité Éthique et Scientifique pour les Recherches, les Études et les Évaluations dans le domaine de la Santé, CESREES) on 8 July 2021, dossier 4698866; by the French Data Protection Agency (Commission Informatique et Libertés, CNIL), dossier MLD/MFI/AR2110215 pursuant to Article 13 of the law and Resolution no. 2019–021 of 28 February 2019 (authorization 921311); and by the French National Health Insurance System (Caisse Nationale d’Assurance Maladie, Cnam) on 15 September 2021.

Furthermore, an independent scientific committee, composed of Dr. Ducastelle-Lepretre, Pr Durand-Zaleski, and Pr Peffault de Latour, was involved from the protocol’s development to ensure scientific and methodological robustness of this study, along with its clinical relevance and its results.

Results

Among the 1,706 patients who initiated letermovir between 1 March 2020, and 31 December 2021, there were 57 patients who presented an exclusion criterion (i.e., patients with AME coverage were identified). The study population (i.e., SNDS patients treated with letermovir) included 1,649 patients. Among them, 511 (31.0%) were identified as 6M-HCT+ patients (i.e., patients treated with letermovir with an index HCT and at least 6 months of follow-up) for the description of letermovir prophylaxis modalities. Finally, 304 patients (18.4%) were identified as patients with complete coverage (i.e., 6M-HCT+ patients for whom a discontinuation was formally identified) for the description of letermovir courses and clinical events (Figure 3).

Sociodemographic and clinical characteristics

Among the 1,649 patients in the study population, the median (IQR) age at letermovir initiation was 57.0 years (44.0–64.0). More than half (54.9%) of the patients were men, leading to a sex ratio of 1.2. The geographical regions where letermovir was most frequently used were Ile-de-France with 361 patients (22.5%), Provence-Alpes-Côte-d’Azur, with 185 patients (11.5%), and Auvergne-Rhône-Alpes, with 183 patients (11.4%). Other regions accounted for less than 10% of the total each. Fewer than 15 patients came from overseas territories.

At the time of letermovir initiation, the most frequent underlying diseases identified were leukemia, myelodysplastic syndrome, and lymphoma, accounting for 69.6% (n = 1,148), 23.3% (n = 385), and 12.7% (n = 209) of patients, respectively. Over the historical period, the majority (89.4%, n = 1,475) of the patients presented with at least one of the comorbidities of interest. The most frequent comorbidity was cardiovascular disease, affecting 1,098 (66.6%) patients. Mental and behavioral conditions were identified in 40.9% of the patients. Hepatic impairment and pulmonary diseases present in 28.0% and 21.6% of the patients, respectively. The median (IQR) Charlson comorbidity index was 6.0 (5.0–7.0). The majority (82.2%, n = 1,356) of patients had a history of HCT, and 81 patients (4.9%) had a history of SOT (Table 1).

Table 1. Sociodemographic and clinical characteristics of the study population (N = 1,649).

The codes and algorithms used for the definition of comorbidities of interest are available in the Supplementary Material.

HCT, hematopoietic stem cell transplantation; IQR, interquartile range; SD, standard deviation; SOT, solid organ transplantation.

6M-HCT+ patients correspond to patients with an HCT within 30 days before or after letermovir initiation and with at least 6 months of follow-up.

Complete coverage patients correspond to 6M-HCT+ patients for whom a discontinuation was formally identified.

Patients may present with several comorbidities or underlying diseases.

History of HCT includes HCTs identified within 2 years before the index date, excluding the index HCT.

The sociodemographic and clinical characteristics of the 6M-HCT+ patients (n = 511) and the patients with complete coverage (n = 304) were similar to those of the study population. Leukemia was the most frequent underlying disease (59.3% and 61.2%, respectively), followed by myelodysplastic syndrome (27% and 25.7%, respectively). The median Charlson score was similar to that of the study population, and a smaller proportion of patients had a history of SOT (<0.4%) (Table 1).

Prophylaxis modalities

The median follow-up for the 511 6M-HCT+ patients was 6 (3–9) months. In this population, the initiation of letermovir was almost exclusively (95.9%, n = 490) done during a hospital stay. The majority of these patients (80.8%, n = 413) initiated letermovir after HCT, with a median (IQR) time of 6.0 (1.0–14.0) days between HCT and initiation. The remaining 98 patients (19.2%) who initiated letermovir before HCT had a median time of 5.5 (1.0–8.0) days between the two events. Over the entire exposure, patients primarily received courses with the 240-mg formulation only, accounting for 390 (76.4%) patients. A dose change between the 240- and 480-mg formulations was observed in 103 patients (20.1%), while courses with the 480-mg formulation only applied to 18 (3.5%) patients. Of note, letermovir at a 240-mg dosage is only used when associated with cyclosporin (due to pharmacokinetic concerns). According to the label, 240 mg when associated with cyclosporin is equivalent to 480 mg without cyclosporin. The letermovir discontinuation-free survival function (95%CI) showed that 91.9% (89.1%–94.0%) of the patients were still being treated with letermovir after 30 days. This rate decreased during follow-up:80.8% (76.8%–84.1%) at 90 days and 53.4% (48.4%–58.2%) at 180 days. The median (IQR) time to discontinuation was 192.0 (173.0–204.0) days (Figure 4).

Letermovir courses

Among the 304 patients with complete coverage, the total mean (SD) prophylaxis duration was 152.0 (96.4) days, regardless of the number of courses, of which 76.9% were dispensed in outpatient settings. The proportion of patients without a second course was 65.2% (n = 195). For these patients, the mean (SD) duration of the initial course was 180.1 (94.9) days, primarily (81.3%) dispensed in outpatient settings. The proportion of patients with a second course was 35.8% (n = 109). For these patients, the mean (SD) duration of the initial course was 101.7 (77.1) days, mainly (62.8%) dispensed in outpatient settings. The mean (SD) time between the two courses was 81.3 (73.1) days. The second course lasted for 103.2 (81.8) days, also primarily dispensed in outpatient care. The cumulative duration of the two courses was 204.9 (154.3) days (Figure 5).

Clinical events of interest

Clinical events were described during the non-exposed periods among 304 patients with complete coverage with either one (n = 195) or two (n = 104) letermovir courses. The proportion of patients with at least one all-cause hospitalization was equivalent regardless of the number of courses (50.8% for one course vs. 49.5% for two courses). Conversely, the mean (SD) monthly number of all-cause hospitalizations among hospitalized patients was 0.4 (0.3) for patients with one course and 0.7 (0.5) for patients with two courses. The proportion of patients with at least one CMV-related hospitalization was 0.5% vs. 7.3%, respectively, while the rates of dispensation of antiviral therapy were 8.7% and 27.5%, respectively. The mean number of medical visits and antiviral therapy dispensations is displayed in Table 2.

Table 2. Cytomegalovirus (CMV)-related events during periods not covered by letermovir according to the number of letermovir courses.

Clinical events were described during periods not exposed to letermovir. for patients with only one course, the non-exposed period corresponded to the time between the end of the first course and the end of follow-up. For patients with two courses, the non-exposed period was defined as the time between the end of the first course and the beginning of the second one. CMV-related hospitalization refers to hospitalizations with a diagnosis of CMV as the main or related diagnosis. The following ICD-10 codes were used: B25.0, B25.1, B25.2, B25.8, and B25.9.

Results are displayed by month among patients presenting with at least one healthcare resource use of interest.

Discussion

To the best of our knowledge, this is the first study to provide real-world data on letermovir prophylaxis modalities in France in terms of overall patient characteristics and letermovir use.

Among French patients with letermovir prophylaxis, the median age at letermovir initiation was 57 years, with 54.9% of the patients being men. In addition, almost every patient in this study presented a hematological malignancy, with myeloid leukemia being the most frequent. According to the annual report from the French National Biomedicine Agency (Agence Française de Biomédecine), 1,878 and 1,989 allogeneic HCTs were performed in 2020 and 2022, respectively, among patients with a median age of 57 years, primarily for acute leukemia (>50%), in line with current findings (12). Furthermore, the distribution of letermovir use by region is similar to that of HCT. Moreover, as no HCT-authorized centers exist in overseas territories, the very low proportion of patients treated with letermovir was expected (13). Among the 6M-HCT+ patients, letermovir was almost exclusively initiated as inpatients and mainly (80.8%) within a week after transplantation, in line with current guidelines recommending an initiation within 28 days after transplantation. Conversely, 19.2% appeared to have initiated letermovir before HCT, usually within a week before the graft (10, 14). This difference could be due to some small inaccuracies in the recorded dates of letermovir initiation in the PMSI, leading to some initiations being wrongly identified as occurring before HCT. Another reason could be that patients initiating letermovir before HCT might be considered at high risk of CMV reactivation, potentially due to the use of corticosteroids or conditioning chemotherapy. Also, it cannot be excluded that patients could have presented with CMV reactivation/disease before HCT, requiring an anticipated use of letermovir as secondary prophylaxis.

Almost every patient in the 6M-HCT+ group was treated with the 240-mg dose (alone or followed by 480 mg), with only 3.5% of the patients treated with 480 mg alone. The 240-mg dosage is recommended when associated with cyclosporin to avoid a mutual pharmacokinetic interaction that increases the letermovir and cyclosporin concentrations (15). While not directly assessed in our study, the fact that the majority of our HCT+ patients received this lower dosage advocates for the concomitant use of cyclosporin. The remaining 3.5% of patients treated only with the 480-mg formulation may have received another immunosuppressant (e.g., mycophenolic acid or tacrolimus, among others), for which the interaction is less significant.

Using the Kaplan–Meier method, this study showed that more than 80% of the patients continued their prophylaxis with letermovir after 90 days, and still more than 50% after 180 days, implying that the prophylaxis duration exceeds the currently recommended 100 days post-HCT among a majority of patients. Of note, this study does not provide information on the clinical status of the patients for whom an extended prophylaxis might be beneficial, as suggested in the guidelines (10).

A subsequent analysis was performed on patients for whom a first letermovir course was identified as completed (i.e., a treatment discontinuation was identified before the end of the study period or death). In this population, the mean prophylaxis duration reached 152 days. These two analyses highlight the overall tendency to increase the prophylaxis duration after 100 days post-HCT among 6M-HCT+ patients, also observed in the literature, with half of the patients covered for 190 days (16, 17). Overall, almost 36% of the patients presented a second course of letermovir. In the case of a second course of letermovir, the initial course tended to be shorter, with 102 vs. 180 days in the case of a single course (8, 9, 18–21). Furthermore, patients with only one course of letermovir tended to have shorter total letermovir exposure, with mean prophylaxis durations of 180 vs. 205 days when a second course was initiated. Hence, longer initial courses of letermovir may be associated with a decrease in the total letermovir exposure of approximately 20 days. These results are in line with several recent studies, which analyzed the potential benefits of extending the duration of the letermovir course beyond 100 days post-HCT and up to 200 days (9, 22, 23). Based on these data, and considering the sole cost of letermovir, the fact that longer initial courses may be associated with shorter total exposure could also imply a decrease in reimbursed costs. In fact, this 20-day decrease observed in our study could translate to savings of almost one 28-tablet box of letermovir, which costs approximately €4,000 for the 240-mg dose. Of note is that, for this analysis, courses where death was observed before discontinuation were considered as incomplete, implying a potential survival bias by analyzing healthy survivors. Based on the EBMT registry, the overall survival in HCT patients is estimated at approximately 80% after 6 months and approximately 66% after 1 year (24). Considering that letermovir courses appeared to last less than 6 months, the significant proportion of patients who are still alive mitigates the impact of this bias.

Furthermore, in addition to presenting a shorter overall exposure to letermovir, analyses of the medical events that occurred during periods not covered by letermovir tended to show that patients with a single course of letermovir tended to have a less frequent need for hospitalizations, antiviral therapies, and medical visits. With a relatively short follow-up period, graft-versus-host diseases (GVHD) were not searched for, as only the most severe would have been identifiable. Given the observational nature of this study, it is not possible to conclude that longer courses of treatment lead to a reduced risk of CMV infections during follow-up, despite the trends identified in our results. In fact, this trend could be due to a selection bias, as the multiplication of letermovir courses may be due to specific risk profiles, other complications (notably GVHD), or treatment center policies. Nonetheless, recent studies have emphasized the positive impact of extending letermovir courses on CMV infections, even in the first months after letermovir discontinuation (20, 25).

The SNDS is a medico-administrative database that contains little clinical information. For example, the results of biological and imaging examinations are not available. Consequently, the results of this study did not take into account potential differences in patients’ clinical presentation, severity, and other clinical factors that could have had an impact on CMV disease or letermovir use. In addition, the description of comorbidities of interest is only based on treatments, hospitalizations, and chronic disease administrative status, with a potential for underestimation. For these reasons, this study focused on the description of letermovir treatment modalities and did not assess treatment effectiveness or factors associated with longer duration. Further analyses adjusting for such clinical information should be carried out to corroborate the trends identified in this study.

Linkage of SNDS with cohorts or registries could improve the granularity and clinical validity of these findings. However, comprehensive identification of hospitalizations, dispensations of letermovir, and other antiviral drugs, and medical visits provides a robust approximation of the overall CMV prophylaxis and management.

The results presented in this study should be interpreted in light of a global organizational and economic spectrum. While the cost-effectiveness of letermovir has been highlighted in several studies when compared with preemptive CMV therapy, no data currently support the organizational importance of longer courses of letermovir (26–29). Although our study did not assess this, administering a second course of letermovir may impact healthcare resource use in both hospital and community care, along with the quality of life of patients receiving repeated courses. Further investigation is necessary to assess the organizational and economic impacts of two letermovir courses compared with a single longer course. Moreover, the possible trade-off between a longer first course and a second course needs to be investigated further, notably to identify patients who might benefit most from a longer course.

Overall, this study provided valuable pieces of data on the typology of HCT+ patients with letermovir prophylaxis and on letermovir prophylaxis modalities, in terms of both exposure and the need for a second course.

Conclusion

This is the first real-world study to describe the use of letermovir in HCT+ patients in France. The mean duration of a first course (with or without a second one) of letermovir among patients with an HCT and at least 6 months of follow-up (or less if death occurred) is nearly 152 days, exceeding the 100 days post-HCT currently recommended. More than one-third of the patients required a second course of letermovir during follow-up. Patients needing a second course of letermovir tended to have a shorter first course compared with those treated with only one course of letermovir. Due to the descriptive nature of this study, further comparative investigations are needed to validate the clinical and economic impacts of extending letermovir exposure up to 200 days post-HCT.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fviro.2026.1733228/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Sreekanth Gopinathan Pillai, Manipal Institute of Virology, India

Reviewed by: Sudheesh N, Manipal Institute of Virology, India

Gayathri Arumughan, University of Sharjah, United Arab Emirates

Footnotes

95%CI, 95% confidence interval; ATC, Anatomical Therapeutic and Chemical classification; AME, State medical aid (aide médicale d’état); CESREES, French ethics and scientific committee (Comité Éthique et Scientifique pour les Recherches, les Études et les Évaluations dans le domaine de la Santé); CI, Confidence interval; CMV, Cytomegalovirus; Cnam, French National Health Insurance System (Caisse National d’Assurance Maladie); CNIL, French Data Protection Agency (Commission Informatique et Libertés); DRG, Diagnosis-related group; HCT, Hematopoietic cell transplantation; IQR, Interquartile range; ISPE, International Society for Pharmacoepidemiology; SD, Standard deviation; SNDS, French National Health Data System (Système national des données de santé); SOT, Solid organ transplantation.