Comparison of prophylaxis and preemptive strategy as cytomegalovirus prevention in liver transplant recipients

Prophylaxis (P) or pre‐emptive strategy (PS) in high‐risk liver transplant recipients (LTRs) are either recommended. We compared the results of each strategy.


INTRODUCTION
[4] Seronegative recipients (R-) transplanted with a graft from a seropositive donor (D+) demonstrate the highest risk of CMV infection after transplantation 2,5 and two strategies to prevent consequences of CMV primary infection in those high-risk patients have been implemented 1,4 : antiviral prophylaxis (P) or pre-emptive strategy (PS).
In liver transplantation (LT), both strategies have been recommended 6 and can be used according to the organization of each transplant center.Recently, in a randomized controlled trial called the CAP-SIL study, the use of pre-emptive therapy, compared with antiviral P, resulted in a lower incidence of CMV disease over 12 months. 7The authors concluded that the PS could be the best option for high-risk liver transplant recipients (LTRs).However, they acknowledged that the logistical aspects of PS (i.e., frequent CMV DNAemia monitoring, rapid initiation of antiviral therapy if a result is positive) especially for outpatients, could be an important concern.They also noted that this result needed confirmation during the traditional follow-up of LTRs, without all the logistical benefits that funded randomized controlled trials might provide.Thus, they evaluated the performance of PS through a monocenter retrospective real-world study conducted in their highly trained center, and they confirmed this result. 8However, whether such a result is generalizable to all centers worldwide remains uncertain.
We took advantage of that both strategies have been successively used in our center to assess the generalizability of these findings.The P strategy was applied first, then, based on previous studies reporting a higher rate of late-onset CMV disease 9 associated with higher mortality with the use of P in high-risk D+R-LTRs, 10 the prevention strategy was changed and replaced by PS for those patients.The aim of this study is to compare the results of each preventive strategy in those patients during these 2 consecutive periods.

Definitions
DNAemia was measured in plasma by "in-house" real-time polymerase chain reaction (PCR) based on TaqMan technology 11 and expressed in copies/ml.CMV disease was defined by the association of positive PCR with clinical and/or biological signs.A number of copies ≥400/mL were considered positive.In the case of tissue-invasive disease, histopathological or virological evidence by culture or positive PCR in the tissue was necessary for the diagnostic.
Ganciclovir resistance was suspected by the persistence or the increase in CMV viral load under antiviral therapy, and confirmed by the presence of genotypic resistance mutations in UL97 and UL54 genes.
Neutropenia was defined by a neutrophil count below 500/mm 3 and was managed by repeated injections of lenograstim until the neutrophil count came over 1500/mm 3 .

CMV DNA surveillance
The details of our P and preemptive prevention protocols and the comparison with the CAPSIL study are shown in Table 1.In the P group, no systematic surveillance of quantitative CMV PCR was performed during the P treatment period in the absence of symptoms.When P was stopped, CMV PCR was performed twice a month until 6 months after LT, then every month or two months during the 6 following months, depending on the scheduled follow-up visits.
CMV DNAemia (regardless of the number of copies) or CMV syndrome were the triggers of PS.In this group, blood was sampled weekly during the first 3 months after transplantation, twice a month during the three following months, and once a month until 1 year after LT in the absence of symptoms.Additional samples (on average every 2 weeks) were collected from CMV DNaemic patients.The duration of a CMV DNAemia episode was calculated as the time elapsed between the first positive plasma sample and the first negative sample under treatment.Surveillance was stopped at 1 year following transplantation.

Antiviral regimens
In the P group, valganciclovir 900 mg daily was given from the first postoperative day for 90 days.Abbreviations: G-SCF, granulocyte colony-stimulating factor; NA, non-applicable; ND, no data.
until the viral load was negative and no secondary P was given.Further CMV episodes were treated on the same principles.
In the PS group, valganciclovir was started at a curative dose at the time of the first positive DNAemia and discontinued until two negative PCR at a week interval.

Immunosuppression and general anti-infective prophylaxis
The immunosuppression regimen was mostly tacrolimus, mycophenolate, and low-dose steroids as previously reported. 12Basiliximab was given when the estimated glomerular filtration rate was 50 mL/min or lower.The initial dose of prednisone was 20 mg per day.The dose was tapered progressively and discontinued by 6 months.Acute cellular rejection was treated by increasing transiently the dose of (or initiating) tacrolimus (targeted trough concentration 15-20 ng/ml), methylprednisolone boluses (10-15 mg/kg), or increasing the daily oral dose of prednisone.Other anti-infective prophylaxis measures are detailed in Table S1.

Endpoints
The primary endpoints were the onset of PCR-based CMV DNAemia (expressed as a proportion of patients and cumulative hazard) and the proportion of patients who developed CMV disease in both groups.The number of episodes of CMV infection, delay of onset, quantitative viral load, the total number of samples performed, antiviral therapy (nature, duration, and cost), episodes of neutropenia, ganciclovir resistance, infectious and immunological complications, cost of both strategies and survival (1-, 5-and 10-year) was analyzed as secondary endpoints.
We evaluated the drug-associated cost of CMV prevention by adding the total costs of antivirals in the first year regardless of the indication, the cost of samples for CMV DNAemia, and the cost of lenograstim to treat episodes of neutropenia.Direct costs of hospital stays were not assessed.

Statistical analysis
Continuous variables were expressed by the median (25th percentile-75th percentile) and compared using nonparametric tests.Categorical variables were compared by chi-square test or Fisher's exact test when appropriate.Kaplan-Meier Survival curves were compared using the log-rank test and hazard ratios and their 95% confidence interval (95% CI) was calculated using the Cox model after adjustment for covariates.
Because this was an exploratory retrospective study, no sample size was calculated.The sample size in each group was sufficient to detect a reduction from 20% to 7% in the rate of CMV disease in the P and PS groups, respectively (95% CI for the difference 0.09% to 26.98%, p <.05).

Patients
Among the 661 patients who underwent LT during the study period, 108 (16%) had the CMV D+/R-serological status.The number of patients who were transplanted per month was similar in the P and PS groups (1 ([0-2] and 1 [0.5-2], respectively, p = .96).Underlying liver diseases were not similar between groups: in the P group, severe cirrhosis was more frequent, and the median MELD score was higher.Other characteristics and treatments did not differ significantly between groups except for a more frequent use of cyclosporine in the P group (Table 2).

Characteristics of virological episodes
More patients experienced two or more episodes of CMV DNAemia in the PS group (52 % vs. 17 %; p = 0.01).In this group, 19, seven, and three patients experienced two, three, and four episodes, respectively, and the median time between the first and the second episode was 26 (14-42) days.The median number of PCR drawn was higher in the PS group (18) than in the P group (seven) (p <.0001) (Table 3).

Antiviral therapy
The use of antivirals was similar in both groups, except for the more frequent use of foscavir in the PS group (p = .02).In this group, antiviral therapy was started with a median delay of 6 (1-12) days after  F I G U R E 1 Cumulative incidence of cytomegalovirus (CMV) DNAemia with death as a competing risk.
the first CMV DNAemia and 7 (2-15) days after the onset of the second episode.The delay between the first negative PCR during the first treated episode and a second episode was 26 (14-42) in 24 PS patients.
The cumulative number of days of antiviral treatment did not significantly differ between groups but the duration of curative therapy was significantly longer in the PS group (91 vs. 16 days; p <.0001) (Table 3).

Other endpoints
Acute cellular biopsy-proven rejection was less frequent in the PS group (Table 3).More patients received lenigrastim, with more injections in treated patients, in the PS group.Of note, more patients experienced an episode of ganciclovir-resistant CMV in the PS group (10% vs. 0, p = .03).
One-year survival was 93.3% (87.0%−99.6%)and 83.3% (72.7%− 93,9%) in the PS and P groups, respectively.The cumulative survival curve at 1 year did not significantly differ between groups (p = .09,Figure S1).After adjustment for variables that could influence survival and were not balanced between groups (diagnosis, MELD score, post-operative bacterial infection), the hazard ratio for death at 1 year was not significantly lower in the PS group (HR = 0.34 [0.08-1.40],p = .14).Long-term survival was similar at 5 and 10 years (Figure S1).

DISCUSSION
Our study highlights important points: i) there was more frequent CMV infection in the PS group but no more CMV disease; ii) ganciclovirresistant virus emerged in the PS group, iii) 1-, 5-and 10-year survival was not significantly different between groups and iv) the drug-associated cost with PS was much higher.
The impact of PS on the incidence of CMV disease has been debated.
In their retrospective study, Bodro et al. reported CMV disease in 33.3% and 8.6% of high-risk LTRs with pre-emptive treatment and P, respectively. 13On the opposite, the incidence of CMV disease was lower with PS (9% versus 19%, p = .048)in the recent trial of Singh et al., 7 and these authors further confirmed the low incidence of CMV diseases in high-risk LTRs followed with a PS through a subsequent real-world study. 8Thus, they stated that PS was probably the best option for high-risk LTRs and that this strategy could easily be implemented during a real-life follow-up.In our study, we did not find any difference in the frequency of CMV diseases between the two groups, in line with the results of Mumtaz et al.'s meta-analysis. 14wever, patients in the PS group experienced more frequent CMV DNAemia, occurring earlier after transplantation, with more frequent recurrences than in the P group, despite curative antiviral treatment.In the Doss study, the turnaround time from CMV PCR sample collection to patient notification, and probably antiviral prescription was short: approximately 2-3 days.This high level of efficiency was likely a result of a well-established procedure, including a central laboratory with CMV PCR testing 6 days a week, and a multidisciplinary team (transplant coordinator, nurses, transplant pharmacist) that was consistently available to notify patients of results and to order and administer the medications.In our referral center, without dedicated staff members available for this task, antiviral therapy was initiated with a median (IQR) delay of 6 (1-12) days and 5 (1-8) days in the PS and P groups, respectively.This extended delay may account for the less effective outcome we observed with the PS.Therefore, it appears that a preemptive strategy could be effective for high-risk LTRs, but only if a rapid administration of antiviral therapy following a positive PCR is consistently feasible.Otherwise, P is likely the preferable approach.
Whereas CMV D+/R-status is a well-established risk factor of ganciclovir-resistant virus, [15][16][17] with an incidence estimated between 5 and 10%, 4 the impact of the type of prevention is unknown.In our study, genotypically confirmed antiviral resistance was observed in the pre-emptive group exclusively.Indeed, prolonged DNaemia with increasing viral load under antiviral treatment longer than 21 days was only observed in the PS group.The main reasons for CMV resistance could also be the longer duration of curative treatment due to longer viral episode duration and more recurrence in the PS group.
In a case-control study, Fisher et al. 17 established that ganciclovir resistance was associated with increased ganciclovir exposure.Similar findings were reported by Couzi et al. in D+R-kidney transplant R receiving pre-emptive therapy. 18The absence of routine ganciclovir dosage raises the question of inadequate ganciclovir plasma concentration with oral valganciclovir.We believe that this hypothesis was unlikely because intravenous ganciclovir and valganciclovir were found to deliver equivalent ganciclovir exposure in LTRs 19 and have similar efficacy in treating CMV infection or disease in transplant R, regardless of their CMV serostatus. 20Moreover, viral clearance may be variable despite adequate ganciclovir plasma levels. 21e PS proved to be more expensive due to several factors.First, because of the greater number of viral episodes and the longer duration of first and recurrent episodes in the PS group, there was a markedly higher duration of curative valganciclovir therapy responsible for the extra cost associated with the PS (i.e.1800 mg bid) compared with P (900 mg bid).Second, the cost of antiviral therapy in patients with ganciclovir-resistant CMV (PS group) was much higher due to both expensive and prolonged second-line treatment (foscavir).
Third, the greater number of samples required for surveillance and the cost associated with the use of lenograstim to treat episodes of neutropenia contributed to a lesser extent to the extra cost of the PS.
Our study had some limitations.First, it was conducted during the 2006-2012 period when CMV DNaemia was mostly performed at University centers.More recently, CMV DNAemia became available in local laboratories, although turnaround time may still reach 7-14 days.
Second, it was a retrospective single-center trial with an intermediate but not large sample size.Because only one-fifth of LTRs have the D+R-CMV serostatus in France, we believe that our study, evaluating more than 100 patients and assessing both strategies for high-risk LTRs in the real world could still provide important information.Third, change in the management of LTRs could have occurred during a 7-year period.
However, except for the CMV prevention strategy, there was no significant change in medical care and immunosuppression which could have influenced the outcome over time.Fourth, total hospitalization costs could not calculated because direct costs of hospital stays were too complex to assess.We could only provide hospital stay days and drug-associated costs.
In conclusion, a PS was associated with more CMV infections but did not result in fewer cases of CMV disease or significantly improved survival.In contrast, the number of samples required for surveillance was greater and the strategy was associated with overall higher cost especially due to the higher cost of antiviral therapy.More frequent episodes of ganciclovir-resistant CMV were observed.PS was not a suitable prevention method in our center, suggesting that the conclusions of the CAPSIL study could not be generalizable to all centers. 7e to simplicity and cost-effectiveness, P in D+/R-LTRs could be more appropriate in centers with limited resources or in those with a long turnaround time for CMV DNAemia.
This is a single-center retrospective study including all the CMVmismatched (D+/R-) LTRs at the Rennes University Hospital, a highvolume LT French center performing more than 100 LTs per year.Patients were identified from the transplant databases and clinical and microbiological data were collected in the computerized patient's files.Two groups of LTRs transplanted during two consecutive periods were compared.P was given to the first group of patients (n = 48) who were transplanted from March 2006 to December 31, 2008, and PS was the prevention strategy for the second group of patients (n = 60) who were transplanted from January 1, 2009 to December 31, 2012.Patients who underwent combined organ transplantation and patients with other CMV serostatus (D+/R+, D-/R+, and D-/R-) were excluded.The study was approved by the institutional review committee.

12 Viremia*
The daily dose was adapted to the estimated glomerular filtration rate.CMV DNAemia was treated with valganciclovir, based on a curative dose of 900 mg twice daily adapted to kidney function for 3-4 weeks.Antiviral treatment was continued TA B L E 1 Comparison of cytomegalovirus (CMV) surveillance protocol, immunosuppression, and side effects between the CAPSIL study and our study.laboratory, PCR performed 6 days a week by transplant coordinator, dedicated nurses, and transplant pharmacist No dedicated staff, PCR performed 2-3 days a week Prophylaxis duration (days) 100 90 (Val)ganciclovir therapy during the first ten days Valganciclovir or iv ganciclovir was allowed at any dose during the first ten days (before randomization) detection of CMV DNAemia Highly sensitive real-time PCR In-house real-time PCR Limit of detection 20 UI/mL 400 copies/mL Surveillance protocol timing over 12 months Not during prophylaxis, then once a week Once a week Not during prophylaxis, then twice a month from months 4 to 6, once a month for months 6-12Once a week during the first 3 months, then twice a month from months 4 to 6, once a month for months 6-Percent of the number of patients in each column.
This resulted in a much longer duration of curative antiviral treatment, more adverse effects, and significant extra costs.At the time of our study (January 2006-December 2012), quantitative CMV DNAemia in blood was not yet part of the nomenclature of biological acts.This test could not be taken in external laboratories and was only performed at university centers.Because many patients lived far from our transplant center, having them come back once a week to monitor CMV was not possible.CMV surveillance was spaced out, and patients were sampled during their hospitalization stay or during visits according to the planned schedule of patient's follow-up visits at the transplant center.Following our results, the preemptive strategy was abandoned and P was reused for the prevention of CMV in high-risk LTR in our center.Other reasons explaining the very different results between our work and the CAPSIL study may also include:1) the use of a more sensitive real-time PCR technique; 2) a very short delay between the detection of CMV viremia and initiation of antiviral therapy (2 days vs. 6 days); 3) resulting in early detection of CMV DNAemia at a lower viral load.
Characteristics of cytomegalovirus infection, treatments, and complications.
TA B L E 2 Characteristics of 108 study patients.*Calculated as the time between the date of the first negative CMV PCR at the end of the first episode and the date of a new positive PCR (second episode).**Not tested.Only two patients in the prophylaxis group had a second episode.***Not tested, too few patients in each group.§ In 75 patients: 20 patients (two patients were treated despite undetectable DNAemia) in the prophylaxis group and 55 patients in the preemptive group.§ § Including the cost of antivirals, lenograstim, and virological PCR samples.