Preemptive ganciclovir for mechanically ventilated patients with cytomegalovirus reactivation

Background The effect of cytomegalovirus (CMV) reactivation on the length of mechanical ventilation and mortality in immunocompetent ICU patients requiring invasive mechanical ventilation remains controversial. The main objective of this study was to determine whether preemptive intravenous ganciclovir increases the number of ventilator-free days in patients with CMV blood reactivation. Methods This double-blind, placebo-controlled, randomized clinical trial involved 19 ICUs in France. Seventy-six adults ≥ 18 years old who had been mechanically ventilated for at least 96 h, expected to remain on mechanical ventilation for ≥ 48 h, and exhibited reactivation of CMV in blood were enrolled between February 5th, 2014, and January 23rd, 2019. Participants were randomized to receive ganciclovir 5 mg/kg bid for 14 days (n = 39) or a matching placebo (n = 37). Results The primary endpoint was ventilator-free days from randomization to day 60. Prespecified secondary outcomes included day 60 mortality. The trial was stopped for futility based on the results of an interim analysis by the DSMB. The subdistribution hazard ratio for being alive and weaned from mechanical ventilation at day 60 for patients receiving ganciclovir (N = 39) compared with control patients (N = 37) was 1.14 (95% CI from 0.63 to 2.06; P = 0.66). The median [IQR] numbers of ventilator-free days for ganciclovir-treated patients and controls were 10 [0–51] and 0 [0–43] days, respectively (P = 0.46). Mortality at day 60 was 41% in patients in the ganciclovir group and 43% in the placebo group (P = .845). Creatinine levels and blood cells counts did not differ significantly between the two groups. Conclusions In patients mechanically ventilated for ≥ 96 h with CMV reactivation in blood, preemptive ganciclovir did not improve the outcome.


Trial objective
For patients under mechanical ventilation for at least 96 hours, to show that ganciclovir treatment of patients with CMV replication or acyclovir treatment of those with HSV replication enables to reduce the length of mechanical ventilation and improves prognosis.

Type of trial
Dual interventional multicenter prospective, randomized, placebocontrolled, double-blind trial with a common patient selection process: patients are jointly screened twice a week for CMV replication in blood and oropharyngeal HSV replication, then included in one of the arms (CMV arm or HSV arm) according to the type of reactivation (CMV or HSV respectively).

Primary evaluation criterion (identical for both study subsets)
Number of ventilator-free days and alive at 60 days after randomization.

Main secondary evaluation criteria
-Mortality at 60d, in ICU and in hospital -Total length of mechanical ventilation in survivors -Total length of stay in ICU and in hospital -Incidence of bacterial pneumonia acquired under mechanical ventilation and of bacteremia -Sequential organ failure assessed by SOFA score -Appearance of side effects linked to acyclovir (kidney failure) or to ganciclovir (myelotoxicity) -Disappearance of oropharyngeal HSV reactivation and CMV PCR in blood becoming negative -Onset of herpes simplex pneumonia or of active CMV infection

Selection of patients eligible for the 2 study subsets
Patients hospitalized in ICU, under mechanical ventilation for the past 4 days, with further mechanical ventilation planned for >48 hours and with no exclusion criteria were screened twice per week: for CMV replication in the blood by PCR until extubation, death or inclusion in the HSV arm for oropharyngeal HSV replication by PCR from a throat swab until extubation, death or inclusion in the CMV arm

Inclusion criteria
Patients hospitalized in ICU, under mechanical ventilation for at least the past 4 days and for the CMV arm: first show CMV replication in blood for the HSV arm: first show oropharyngeal HSV replication

Main exclusion criteria
Age <18 years, immunodepression (HIV, with bone marrow or solid organ graft, under immunosuppressive therapy including steroids at a dose >0.5mg/kg/day for over one month), pregnant, under acyclovir, ganciclovir or foscavir treatment within the previous 30 days, morbidity or a decision to limit active treatment, active CMV or HSV infection, severe neutropenia, severe thrombocytopenia, hypersensitivity to ganciclovir (in CMV arm) or to acyclovir (in HSV arm).
Treatments under study For the CMV subset, patients randomized into the ganciclovir arm will be given 5 mg of ganciclovir/kg twice a day (10 mg/kg/d) for 14 days. In the event of kidney failure, doses shall be adapted to renal function. Patients randomized into the placebo arm will be given a placebo with an identical presentation to ganciclovir. For the HSV subset, patients randomized into the acyclovir arm will be given 5 mg acyclovir/kg 3 times a day (15 mg/kg/d) for 14 days. Patients randomized into the placebo arm will be given a placebo with an identical presentation to acyclovir. In the event of kidney failure, doses shall be adapted to renal function.

Number of patients
The hypothesis is that ganciclovir or acyclovir therapy would increase the number of days alive without mechanical ventilation by 8 days calculated over 60 days compared to a placebo. Accordingly, 112 patients must be included in each arm. In order to ensure obtaining 112 patients per arm, we plan to include 240 patients in total for each arm, i.e. a total of 480 patients.

Introduction and justification for research
Viral infections have only very recently become an object of investigation in patients under mechanical ventilation (MV) hospitalized in intensive care unit (ICU) (1-4). It is mainly the progress that has been made in directly detecting these pathogens that has enabled intensive care specialists to evaluate the impact of viruses on their patients more objectively.
This of course includes "community" viral infections (influenza, rhinovirus…), but also viral infections in immunocompetent adults under MV. These are known as "nosocomial" viral infections among which Herpesviridae are the most frequently encountered and the most studied (1-3). Herpesviridae replication in intensive care patients is more often reactivation than a primary infection. This reactivation results from the fact that after a few days of MV there is an onset of immunoparalysis which can pave the way for bacterial or viral infections (5).

Cytomegalovirus
Cytomegalovirus (CMV) is a DNA virus belonging to the Herpesviridae family beginning with a primary infection during childhood which is usually totally or partially asymptomatic (6). Afterwards, the virus remains latent in cells of the immune system, however a drop in the host's cellular immunity (transplantation, AIDS, immunosuppressive therapy…) can cause an endogenous recurrence of viral infection.
CMV seroprevalence in adults is higher than 60% in industrialized countries (6), and from our experience, 75% on hospital admission (1). Patients in intensive care essentially have a reactivation of endogenous CMV (10), however infection by exogenous CMV remains possible (ex: transfusion) though rare, especially since systematic leukodepletion of blood products in France.

Incidence in intensive care and methods of detection
The incidence of active CMV infection in ICU patients is between 15 and 20% (1, 3, 7-11). Patients hospitalized for severe septicemia are probably at a higher risk of subsequent CMV reactivation (8,12), as are patients who are CMV seropositive on hospital admission (1). Incidence variability also depends on the method chosen to screen CMV. Indeed, some methods such as PCR or pp65 antigenemia of the blood are more sensitive than the culture of respiratory secretions and/or the presence of a cytopathic effect in bronchoalveolar lavage Finally, considering all of the patients admitted to intensive care unit (ICU) and monitored by weekly antigenemia assays, CMV infection concerns at least 1 out of 6 patients (1).

Manifestations: virus does not mean "viral infection"
On the contrary to immunodepressed patients (6), extra-pulmonary clinical manifestations (colitis, retinitis) are very rarely reported in ICU patients (13). It was a study on lung histology that revealed the reality of CMV infection in immunocompetent patients, this virus could be considered as a pathogen responsible for ventilator-associated pneumonia (VAP) (12). In a study on surgical lung biopsies from 86 patients with unexplained respiratory distress, 25 had histological signs of CMV lung infection. The same group then identified the same signs in 30 out of 100 patients who had had a lung biopsy for persistent ARDS (14). Furthermore, patients with CMV infection had a much higher rate of morbidity (length of MV and length of ICU stay) and/ or of mortality than other patients, independently of other prognostic factors (3). Finally, some experimental arguments also suggest that CMV has a genuine pathogenicity. Thus, in mice pre-infected with CMV, sepsis induced CMV reactivation in the lungs (7).
Nevertheless, CMV reactivation detected by PCR or antigenemia in intensive care patients is not necessarily synonymous with CMV disease. Indeed a certain number of these reactivations clear without any antiviral treatment, suggesting that CMV may only be a marker of the degree of immunodepression acquired in intensive care. Only a therapeutic trial testing the impact of an antiviral drug could provide undeniable evidence of the pathogenicity of CMV for these patients.

Antiviral treatment of CMV: with what and for whom?
Antiviral drugs have adverse effects preventing an empirical proposal of systematic prophylaxis on admission, whereas only 1/3 patients at the most are concerned. Apart from the cases where a lung biopsy is performed (curative treatment), ICU specialists sometimes start a "preemptive" treatment as soon as early reactivation markers become positive (antigenemia or PCR), and even more so in cases where there is no other explanation for correctly investigated respiratory distress and/or when steroid therapy becomes necessary. In these cases ganciclovir is used. Up until now there have been no studies on the pertinence of a preemptive strategy with regards to CMV. Only a large-scale, interventional multicenter trial could evaluate the pertinence of such a therapeutic strategy (efficiency, tolerance).

Herpes simplex virus type 1
Herpes simplex virus (HSV)-1 is a DNA virus belonging to the Herpesviridae family.
Like CMV, it infects humans during their early childhood: primary infection can be asymptomatic (the majority of cases) or symptomatic, in which case it appears as herpetic gingivostomatitis. The virus remains latent in the sensory ganglia of the spinal cord (the most often in trigeminal ganglia), and in the event of trauma, endocrine stimulation or immunodepression, it can be reactivated and either induces an asymptomatic saliva excretion, or lip sores ("cold sores"). By the age of 50, 90% of adults are thought to be HSV seropositive.

Herpes simplex virus type 1 in ICU
HSV-1 was isolated in the saliva of 1-5 % of the general population. In  with HSV bronchopneumonitis could develop nosocomial bacterial pneumonia more frequently than others, but with an identical rate of mortality (2).
More recently, Linssen et al. found that patients with an HSV viral load above 10 5 copies/ml in BAL had an adjusted mortality 14 days higher than those with a viral load <10 5 copies/ml (23). Yet the viral load in BAL was shown to be associated with the development of HSV pulmonary damage (2).
If the replication of HSV in the lower respiratory tract (and even more so in cases of genuine HSV bronchopneumonitis) is associated with increased morbidity and maybe mortality, this is only an association and there will still be a controversy about the actual pathogenicity of HSV in ventilated patients: is it a genuine parenchymal disease with its own morbidity/mortality (autonomous pulmonary damage paving the way for bacterial secondary infections), or is it just a marker for the severity of an underlying disease? There are no data in literature to currently answer this question. The only randomized study was performed by Tuxen in 1987, who randomly allocated 45 patients with ARDS for the preventive administration of acyclovir or a placebo. The results of this study showed that preventive therapy enabled to reduce the incidence of HSV reactivation but had no effect on the length of ventilation or on mortality (24). However, this study was small and it was not possible to formally conclude that the treatment had no efficacy for these patients.

Rationale for a randomized trial evaluating acyclovir
The fact that HSV reactivation is potentially pathogenic for ICU patients (via HSV pneumonia, itself responsible for morbi-mortality) and the lack of randomized studies in scientific literature assessing the efficacy of preemptive acyclovir treatment for patients, justifies carrying out a randomized study in order to answer the question of whether preemptive treatment of these patients is useful.
The choice of the population studied, i.e. patients with oropharyngeal HSV reactivation is linked to the pathophysiology of lung damage: first a primary reactivation of HSV in the oropharynx, then a secondary infection of subglottic airways via microinhalations. In the majority of cases, there is viral reactivation in the distal airways but in certain cases this replication leads to actual histological lung damage (pneumonia). Treating patients at the beginning of the cycle, during oropharyngeal replication of HSV would prevent the contamination of distal airways and thus prevent HSV pneumonia.
The choice of preemptive rather than curative treatment is justified by the fact that when lung damage is declared, it is probably too late and treatment is probably less beneficial: by analogy, acyclovir or valacyclovir treatment is less efficient on herpetic lip sores once the sores are visible. Furthermore, HSV bronchopneumonitis is difficult to diagnose in ICU patients, requiring histological confirmation or quantification of the viral load, 2 exams which are difficult to implement in daily practice, and even more so as screening for a multicenter trial (2). On the contrary, preventive (prophylactic) treatment of all intubated patients under MV for longer than 4 days with acyclovir could be discussed; however, this would expose patients without HSV reactivation to a potentially toxic treatment.

A project in common to pool efforts
There is a strong rationale to submit a project in common including both study subsets

Task-sharing to optimize response time for the questions asked
Each respective arm will be managed by the following separate teams

Hypothesis tested
Our hypothesis is that the administration of an anti-viral drug (ganciclovir for CMV, acyclovir for HSV) as soon as PCR results are positive (blood for CMV, oropharyngeal for HSV) enables to increase the number of ventilator-free days and alive (VFD) after 60 days (the number of days for which the patient is alive and weaned off MV) (d1 = the day ganciclovir, acyclovir or a placebo treatment is started) for patients treated with ganciclovir compared to the placebo on the one hand (CMV arm), and for patients treated with acyclovir compared to the placebo on the other hand (HSV arm).

Primary objective
An increase by at least 8 days of the number of days the patients are weaned off MVand alive at 60d (VFD D60) post-inclusion:  with a 14d treatment of ganciclovir following positive CMV PCR results in blood  with a 14d treatment of acyclovir following positive oropharyngeal HSV PCR results.
The number of ventilator-free days is calculated after the absence of any invasive MV for at least 48h.

Secondary objectives
The following will be studied for each arm:

Experimental plan
It is a prospective trial split into two arms, randomized, comparative and doubleblind carried out on 2 parallel groups of 2.
 CMV arm: one experimental group using ganciclovir and one group with standard care and a placebo in case of positive CMV PCR results in blood.

Multicenter recruitment
Given the number of patients to be included and the units' capacity of recruitment, the centers participating in this trial should make it possible to answer the questions being asked.
Each coordinator will have to classify the patients susceptible of being included, and if appropriate, the reasons why they are excluded. All of the screened patients will therefore be classified, regardless of virology results. The list of centers is displayed in Annex 1. We count on 15 inclusions per center and per year, i.e. approximately one inclusion every three weeks.

Patient selection criteria
Selection of patients eligible for both arms of the study: All patients hospitalized in ICU, under MV for the past 4 days, with ventilation planned to last at least 48 hours longer and without any exclusion criteria will be screened twice a week: for CMV replication by PCR on blood until 30 full days under MV, leaving ICU or death.
-For oropharyngeal HSV replication by PCR on a throat swab until 30 full days under MV, leaving ICU or death.

Inclusion criteria
Patients with all of the following criteria will be eligible for inclusion in either arm of the study depending on the type of positive PCR results: -invasive MV for at least the past 96 hours and planned to last for at least 48 hours longer -PCR positive (see below for PCR positivity criteria) -aged at least 18 -with written consent from the patient, from a close relative or from the person of trust previously appointed.
-under social security cover -negative pregnancy test for women of childbearing age

Exclusion criteria
Patients with at least one of the following criteria will not be eligible:  The product is prepared in compliance with the method described in the SmPC. This is carried out by the nurse. As of February 14, 2016, ganciclovir will be reconstituted and dispensed with the label "clinical trial" by the pharmacy of each site which is authorized to carry out preparations made necessary for biomedical research including experimental drugs. In the event of it being impossible for the PUI to prepare the treatment under study (too many demands for clinical study preparations, less staff on weekends and bank holidays), the latter will dispense the product labeled "clinical trial" to a third-party unit not involved in the PTH trial. The verum treatment labeled "clinical trial", will be prepared according to the SmPC of the product (ganciclovir). The list of "third-party" units is provided in section 20.1 of this protocol. A specific prescription is provided for when a "third-party" unit is called for.

Treatments under study
Length of treatment: treatment shall be administered for 14 days

Concomitant treatment with Ganciclovir
COMBINATION ADVISED AGAINST didanosine: on the one hand, a risk of increasing didanosine side effects due to a significant increase of its serum concentration; on the other hand, a risk of reducing ganciclovir efficacy by reducing its serum concentration, if both drugs are administered within 2 hours of each other.

COMBINATIONS REQUIRING USE OF PRECAUTION
-zidovudine: increased hematologic toxicity (added effect of bone marrow toxicity). Temporarily stop taking zidovudine; monitor blood count and, if possible, reintroduce zidovudine at a low dose.
-zalcitabine: increased risk of peripheral neuropathy by adding adverse effects. Regular clinical monitoring.
-more generally, some prudence is recommended when concomitantly using ganciclovir and potentially hematotoxic products such as linezolide for example, which requires blood count monitoring every other day as the ANSM (French National Agency for Drug Safety) recommend for this protocol.

Acyclovir
Rationale: Acyclovir is the injectable drug of reference in HSV treatment. The product is prepared in compliance with the method described in the SmPC. This is carried out by the nurse. As of November 16, 2016, acyclovir will be reconstituted and dispensed with the label "clinical trial" by the pharmacy of each site which is authorized to carry out preparations made necessary for biomedical research including experimental drugs. In the event of it being impossible for the PUI to prepare the treatment under study (too many demands for clinical study preparations, less staff on weekends and bank holidays), the latter will dispense the product labeled "clinical trial" to a third-party unit not involved in the PTH trial. The verum treatment labeled "clinical trial", will be prepared according to the SmPC of the product (acyclovir). The list of "third-party" units is provided in section 20.1 of this protocol. A specific prescription is provided for when a "third-party" unit is called for.
Length of treatment: treatment shall be administered for 14 days

Placebos
Placebos are prepared in exactly the same way as ganciclovir (CMV arm of the study) and as acyclovir (HSV arm of the study), thus requiring a powder to be reconstituted. A pharmaceutical processing company authorized by the Agence Nationale de la Sécurité du Médicament (French National Agency for Drug Safety) will be in charge of manufacturing, packaging and labeling these placebos. Given that the study is double-blind, dosage, methods of administration and length of treatment are identical to those of the respective verum.
As of February 14, 2016, the placebo for ganciclovir will be reconstituted and dispensed with the label "clinical trial" by the pharmacy of each site which is authorized to carry out preparations made necessary for biomedical research including experimental drugs.
In the event of it being impossible for the PUI to prepare the treatment under study (too many demands for clinical study preparations, less staff on weekends and bank holidays), the latter will dispense the product labeled "clinical trial" to a third-party unit not involved in the PTH trial The placebo will be made up of a pouch of 5% glucose solution (50 or 100 ml) labeled "clinical trial". The list of "third-party" units is given in section 20.1 of this protocol. A specific prescription is provided for when a "third-party" unit is called for.
As of November 16, 2016, the placebo for acyclovir will be reconstituted and dispensed with the label "clinical trial" by the pharmacy of each site which is authorized to carry out preparations made necessary for biomedical research including experimental drugs.
In the event of it being impossible for the PUI to prepare the treatment under study (too many demands for clinical study preparations, less staff on weekends and bank holidays), the latter will dispense the product labeled "clinical trial" to a third-party unit not involved in the PTH trial The placebo will be made up of a pouch of 5% glucose solution (50 or 100 ml) labeled "clinical trial". The list of "third-party" units is given in section 20.1 of this protocol. A specific prescription is provided for when a "third-party" unit is called for.

Packaging, labeling and storage conditions
The trial products (verum and placebos) will be packaged double-blind by the same processing pharmaceutical company in charge of making the placebo. For the economical and ergonomic management of the drugs, they will be presented in packs of 7 days treatment. A total of 1000 kits for 7 days will be made (2 kits per patient for the full 14 days of treatment) (500 of verum, 500 of placebo). The fact that there is a higher number of kits than the number of patients included can be explained by the number of kits "immobilized "in each center. As of February 14, 2016, with regards to the CMV study subset, reconstituted ganciclovir or its placebo will be issued by the pharmacy of each site. The type of product resulting from randomization (verum or placebo) as well as the randomization number will be given to the on-site pharmacy. Ganciclovir will be prepared according to recommendations given in the Vidal: inject 10 ml water for injectable preparation into the bottle of ganciclovir -shake thoroughlyremove the required quantity and dilute it in a pouch or flask of either 50 or 100 ml of 5% glucose solution. Reconstituted mix can be stored for 24h maximum in the fridge. The placebo will be prepared using a pouch or flask of 50 or 100 ml 5% glucose solution with nothing added. Each site will define their own common strategy for the preparation of verum or placebo (type of solvent: 5% glucose solution, flask or pouch of 50 or 100 ml). A label will be stuck on each reconstituted product including:  Name, address and telephone number of the Sponsor As of November 16, 2016, with regards to the HSV study subset, reconstituted aciclovir or its placebo will be issued by the pharmacy of each site. The type of product resulting from randomization (verum or placebo) as well as the randomization number will be given to the on-site pharmacy. Acyclovir will be prepared according to recommendations given in the Vidal: inject 20 ml water for injectable preparation into the bottle of 500 mg acyclovirshake thoroughlyremove the required quantity and dilute it in a 100 ml pouch or flask of saline solution. The reconstituted mix can be stored at room temperature (temperature < 25°C) for 12 hours maximum. Placebo will be prepared using a pouch or flask containing 100 ml of saline solution with nothing added. A label will be stuck on each reconstituted product including:  Name, address and telephone number of the Sponsor

Distribution
Distribution and stock management of treatments under study in the different pharmacies for internal use will be performed by the Hôpitaux Sud pharmacy of the AP-HM in Marseille.
To enable economical stock management, treatments will be distributed to the different centers in several phases depending on available stock.
As of February 14, 2016, each investigation site will reconstitute ganciclovir or issue the placebo under an identical form, for the CMV study subset. The investigator-clinician will draw up a prescription and transmit it to the pharmacy. This will include the identity of the patient, their inclusion number, the words PTH-CMV trial, that day's serum creatinine results, the dosage prescribed.
As of November 16, 2016, each investigation site will reconstitute acyclovir or issue the placebo under an identical form, for the HSV study subset. The investigator-clinician will draw up a prescription and transmit it to the pharmacy. This will include the identity of the patient, their inclusion number, the words PTH-HSV trial, that day's serum creatinine results, the dosage prescribed.

Early treatment
Treating early with antiviral drugs is a potentially important factor in limiting morbidity/mortality associated with Herpesviridae. The introduction of the verum (ganciclovir or acyclovir) or of the placebo must not be later than 24h after PCR results are known.

Responsibility concerning the treatment under study
The pharmacist is responsible for checking the products received by signing an acknowledgment of receipt for every product delivery. All of the products must be stored in locked premises with access strictly limited to staff appointed for the trial. Products will be stored according to the Sponsor's specifications, following the manufacturer's instructions and in compliance with applicable regulations.
The pharmacist is responsible for maintaining and keeping an accurate register of the quantities of product dispensed, used and returned for each individual. All of the treatments under study must be dispensed according to the investigator's prescription.
Any quality issue observed upon receipt or when using the treatment under study (defect of the state, aspect, documentation, labeling, date of expiry, etc….of the treatments under study) must be immediately reported to the sponsor, who will initiate a complaints process. In the event of the identification of a potential quality defect of the treatments under study, the sponsor can decide to initiate a recall process. In this case, the investigator or the pharmacist will be responsible for immediately treating the sponsor's demand, in order to recall the experimental drug and to eliminate any potential risks.
The Investigator may not in any case, provide the treatment under study to a third party, nor enable the treatment under study to be used outside the clinical trial protocol, nor dispose of the drug under study by any other means. -Treatment with ganciclovir or its placebo will be stopped in the event of:

Unblinding
Unblinding is the exceptional procedure consisting of revealing the treatment code for one of the patients. This is usually done once the trial has finished. Nevertheless, some specific circumstances require unblinding while the trial is going on: mainly in the event of an adverse effect for patient safety and to be able to set up appropriate care for this event by having all the facts. Unblinding before the end of the trial will only apply to the patient concerned. It will be obtained via the eCRF.

Compliance
Treatment compliance will be monitored by the nurse in charge of each patient, he/she will fill in an administration sheet, keep the empty vials and any potentially non-allocated vials.
A register accounting for products returned one by one will be held by the pharmacist at the investigation sites. The study monitor will periodically check the quantity of products held by the investigator or by the pharmacist, to check that the products used are being correctly managed.
At the end of the study, the CRA in charge of quality control will count the number of therapeutic units used. This data will be used to calculate observance and to identify deviations from the protocol.

Treatment returns and destruction
At the end of the trial, all of the non-used or partially used products and all of the packaging will be returned to the PUI at the hôpital de la Timone of the AP-HM.
Destruction will be performed there under the pharmacist's responsibility with the sponsor's permission.

Main judgment criterion: number of days alive and weaned off MV defined at D60
The main objective of either arm is to show a significant reduction in morbi-mortality

Evaluation of tolerance
The occurrence of side effects linked to acyclovir or to ganciclovir shall be investigated.
Blood count results will be noted (for the CMV arm: every other day) and serum creatinine levels on D3, D5, D7 and D14. This will not be specific monitoring since it is usual in intensive care. µmol.L -1 or chronic hemodialysis), recent blood transfusion (< 1 month), recent steroid treatment (less than a month before).

Other parameters to be noted
-The day of inclusion: SOFA, ventilator parameters, prior transfusions, Glasgow coma scale score, temperature, blood gas tests, vasopressors, antibiotics, steroids, dialysis, CPIS.
Clinical Pulmonary Infection Score (CPIS) will also be assessed on D3, D7 and D14. We

Randomization
We will be using an electronic randomization process using internet. A telephone hotline open 24/7 will be operational for any issue related to the protocol. We shall use randomization by minimization, stratified by site, length of MV upon inclusion and number of existing organ failures upon inclusion. During randomization, the presence of inclusion criteria and the absence of exclusion criteria will be checked.

Virology monitoring from D1 to D30
PCR will be performed twice per week, the same frequency as during screening. The results will no longer be transmitted to the clinicians. HSV and CMV PCR will be stopped when leaving intensive care or on D30 post-inclusion at the latest.

Visit on D60
Vital stats and MV status. Inclusion criteria X X SOFA/CPIS X X X X X X X X

Treatment of declared HSV or CMV infection
If, following inclusion, there is an onset of HSV bronchopneumonitis and/or active CMV infection, curative therapy may be prescribed by the clinician in charge of the patient. If this infection occurs during the 14 days with the treatment under study and the investigator considers that it requires treating, the treatment under study will be stopped.

Early withdrawal from the trial
The conditions for early withdrawal from the trial are as follows: -withdrawal of consent -adverse event related to one of the products under study (see chapter 12.5) -occurrence of active CMV infection and/or HSV pneumonia (see chapter 7.9) In the event of early withdrawal from the trial, a full final assessment must be carried out for each patient and the patients must have an end of trial visit within 30 days of stopping the treatment.
Adverse Effects occurring within 30 days from stopping treatment in patients who withdrew from the trial early must be documented, once the investigator has been informed.

Case report form
Data concerning patients included in the trial will be grouped in an eCRF filled in by the investigators.
The eCRF will be set up via a partnership with the network Recherche En Ventilation Artificielle (REVA) who has developed a standard eCRF adapted to intensive care that is already used in several PHRC including OVNI.
Each eCRF includes the center letter (one letter), the inclusion number (4 digits), the first three letters of the patient's last name, the first two letters of the patient's first name and the date of inclusion.

Role of CRA coordinators
Two CRA coordinators will be placed under the responsibility of the principal investigators. One CRA coordinator (0.33 FTE) is linked to the La Pitié site (Paris). One CRA coordinator (0.66 FTE) is linked to the Hôpital Nord site (Marseille).
They will be assigned the following tasks:

Quality control
Quality control will be performed on all of the case report forms in the study (100% of data) by the sponsor.

Data management during the trial
A protocol for database transfer will be set up before the beginning of data collection.
This will describe the database structure and function: tables, variables, labels, codes and formats used, reference data, checks carried out during data entry, as well as the means of transfer. A transfer test will be carried out in order to validate the correct transmission of all of the data.

Data coding
Concomitant treatments will be coded according to ATC classification. Medical history, concomitant illnesses and adverse events will be coded using the MedDRA dictionary.

Data validation
Electronic consistency checks will be performed according to a data management plan which will be drawn up for the protocol. If inconsistencies are detected, correction requests (DCF) will be sent to the investigators and inconsistency removal will be continued until all of the corrections have been entered and validated.

Review of data validation
A meeting to review blind data will be held for each study arm (HSV/CMV).

Freezing the database
The database will be frozen once all of the corrections requested from the investigators and the decisions made during the data review meetings have been integrated into the data files. The database will be frozen according to CIC-CPCET procedures with audit trails, and documented by a declaration of database freeze. The database will be archived before statistical analyses begin. Any further modification will be subject to an unfreeze request.

Data storage
The trial database as well as any change history (audit trails) will be saved and held by the CIC-CPCET for the legal archiving period. The CIC-CPCET will ensure archiving of the frozen database in SAS format.

Planned number of patients and justification
We hypothesize that a 14-day treatment with acyclovir immediately after positive HSV PCR results would enable to increase the number of days alive and ventilator-free by 8 days compared to the placebo group. According to data in literature, the variability of the number of days patients stay alive and ventilator-free corresponding to HSV inclusion criteria has a standard deviation of 20 days (2). Concerning the CMV arm, we based it on the same hypotheses: a standard deviation of 20 days and an expected difference of 8 days between both groups (ganciclovir vs placebo).
Considering these hypotheses, for a bilateral study and with an expected power defined at 80%, the number of individuals to include in each group is 97 patients. However, this number of individuals needs to be adjusted by the asymptotic relative efficiency factor for the Mann-Whitney test (non-Gaussian distribution) (Hollander M, Wolfe D. Nonparametric statistical methods, 2 nd edition. New York NY, John Wiley and Sons, Inc, 1999). The correction factor applied is 0.864. The number of individuals per group is therefore 112 (97 / 0.864). In order to be sure of having 112 exploitable observations per group and to account for any secondary attrition, 120 patients will be included in each treatment group.
A total of 240 patients will be included in the CMV arm (120 will receive ganciclovir, 120 will receive the placebo) and 240 patients will be included in the HSV arm (120 will receive acyclovir, 120 will receive the placebo). results for CMV or HSV. Therefore a total of at least 1200 patients will need to be screened.

Randomization method
Randomization will be performed independently for each arm of the study (CMV and HSV) following the same procedure.
Within each arm, included patients will be randomly allocated by draw to either treatment group: placebo or verum (ganciclovir or acyclovir). Randomization will be stratified according to the center, the length of MV prior to inclusion and the presence of organ failure.
The length of MV prior to inclusion is split into 2 levels: between 4 and 14 days, and between 15 and 30 days. This stratification aims to limit the risk of non-comparability between groups, given that the incidence of positive CMV and HSV results is linked to the length of ventilation. Organ failure is also split into 2 levels: < 2 or ≥ 2 organ failures according to the SOFA score.
Minimization will be used to attribute patients into a group. Randomization will be electronic. A telephone hotline will be available 24/7.

Statistical analysis
Statistical analysis will start after the database has been frozen, the report from the data review meeting has been approved, and the Statistical Analysis Plan (SAP) has also been approved. The methods given in this section may be updated during the study, justified and finalized in the SAP.
Analysis of joint trial data (screening phase) and of the CMV arm will be carried out by Major protocol deviations will be described in detail. The patients for whom there is a protocol deviation will be identified and classified. Only the cases of patients with major protocol deviations will be discussed in the clinical study report. Major deviations will be used to define the per protocol population. In this hypothesis, analysis of the intention-to-treat population prevails over analysis of the per protocol population which will be complementary.

Populations analyzed
In each study arm, the following populations will be analyzed:  For tolerance: all of the patients treated (T), i.e. all of the patients who received at least one dose of the treatment under study and for whom at least one measurement of tolerance data is available.

 For efficiency:
o Full Analysis Set (FAS), all of the patients randomized except patients who withdrew their consent. This is the "intention-to-treat" population.
o Per Protocol (PP), i.e. patients from the tolerance population who participated in the whole trial without any major deviation.

General statistical approach
Descriptive statistics will be presented per treatment group and per visit, for measured values with absolute variations (differences) and/or relative variations (%) in comparison with inclusion. Quantitative variables will be given as mean, standard deviation, median, quartiles Q1 (25 th percentile) and Q3 (75 th percentile), minimum and maximum. Qualitative variables will be given as relative frequency and percentage.
Reference values (baseline) are defined as being the last measurement available prior to randomization.
Missing data will not be replaced.
Statistical tests used to compare both of the treatment groups will be defined according No sub-group analysis is planned in the protocol to analyze efficiency. We will privilege use of multivariate analysis which accounts for adjustment factors if it seems necessary to take concomitant parameters into account (potential bias): models for analysis of variance or of covariance for quantitative parameters, logistic regression model for qualitative parameters.
No intermediate statistical analysis is planned.
The threshold for significance is set at p < 0.05.
Analysis will be carried out using SAS software.

Description of patients and comparability between groups
This analysis will be carried out independently for each study arm: CMV and HSV.
The populations studied will be described according to the main parameters measured upon inclusion.
All of the parameters collected during the screening phase and the day PCR results became positive will be described per studied population and per treatment group.
Comparability between treatment groups will be checked with the following parameters: This analysis will identify any potential adjustment factors.

Efficiency analysis
This analysis will be carried out independently for each study arm: CMV and HSV.

Primary analysis
The main judgment criterion is the number of days alive and weaned off MV on D60 right-censored: ventilator-free days and alive (VFD). (definition in 6.1) The number of VFD will be given on D60 and compared between groups using a method of means comparison which will be chosen according to: whether adjustment is necessary or not, special attention will be paid to the state of severity of patients on randomization; the distribution of this variable, supposedly not normally distributed, and therefore the possible use of a normalizing transformation.
According to the case, the strategy for analysis will be the following:

Secondary analyses
All of the secondary judgment criteria will be given per treatment group. Parameters measured at several monitoring points will be given for each point with their variation compared to inclusion values.
The following will be compared between both treatment groups, using the Chi2 test (or

Tolerance analysis
Clinical treatment tolerance is evaluated by analyzing adverse events and biological data.

Analysis of adverse events
All adverse events will be given per "organ system" and per "preferred term" (MedDRA) in summary tables. Adverse event frequency (%) will be calculated for the total number of events (%N EI ) and for the number of patients (%N P ).
Criteria in terms of adverse events are the number of patients who presented: at least one adverse effect, -one adverse event for each identified preferred term, -one adverse event for each identified organ system.
These criteria apply to any adverse event whatever its relationship with the product under study. They will be analyzed for the tolerance population, presented as a whole and by treatment group.
Serious adverse events (whether during treatment or not, and whatever their supposed causal relationship with the products under study) will be listed and summarized.
Adverse events leading to a halt by administration will be listed and summarized for each individual, whether they took place during treatment or not, and whatever their supposed causal relationship with the product under study.

Analysis of biological data
Blood count and serum creatinine levels measured on D3, D7 and D14 will be summarized per treatment group as values and variation from baseline.
Altered renal function will be described at each assessment by: -the number of patients per level of serum creatinine in the CMV arm.
-the number of patients per level of serum creatinine in the HSV arm.
(see paragraph 5.5 criteria for modifying or stopping the treatment under study).

Treatment under study
Observance for products under study will be provided as a percentage: number of days treated / number of days of trial.

Concomitant treatment
Concomitant treatment throughout the trial will be given according to its ATC code.

Study length
The inclusion phase will last for 30 months (progressive opening and closing of the 14 centers depending on the sponsor's availability). The trial will last for 36 months.

Length of participation of each person
Each patient will participate for 60 days.

Date Action
July 2012 Application

General Definition
A Serious Adverse Event (SAE) is any event that: -Results in death, -Is life-threatening, -Causes prolongation of existing hospitalization, -Results in medically significant persistent disability or serious temporary incapacity.
The terms disability and incapacity correspond to any temporary or permanent physical or mental disability, clinically significant and having an effect on the patient's physical activity and/or quality of life.
The term medically significant is any clinical event or laboratory result that the investigator considers to be serious and does not correspond to the severity criteria defined above. It may put the patient at risk and require a medical intervention to avoid one of the outcomes corresponding to the severity criteria given above.

Definition of an expected serious adverse event
An E-SAE is one of the events already mentioned in the summary of characteristics of the product under study which already has a marketing authorization.

Definition of an unexpected serious adverse event
A U-SAE is an event that is not mentioned in or differs from the summary of characteristics of the product under study by its type, intensity, or progression.

Intensity criteria
Intensity criteria must not be confused with severity criteria which serve as a guide to define report obligations.
The intensity of events will be estimated using the NCI-CTC grading scale version 3.0 (toxicity graded from 1 to 5). The intensity of adverse events not listed in this scale will be graded according to the following qualifiers: For each event the investigator will note:  Its description as clearly as possible using medical terminology, The investigator must also enclose the following with the SAE report form, whenever it is possible:  A copy of the hospitalization or prolongation of hospitalization report,  A copy of the autopsy report,  A copy of all the supplementary exam results carried out, including pertinent negative results, providing laboratory reference values,  Any other document that he/she considers useful and pertinent.
All of these documents must be anonymized.
Extra information may be requested (by fax, by phone or during a visit) from the investigator by pharmacovigilance. Within 48 hours the latter must send the sponsor the "supplementary information for an unexpected SAE" form (inserted in the patient observation form) correctly filled in.

SAE follow-up
The investigator is responsible for adapted medical follow-up of patients until the event is resolved or stabilized or until the patient dies. Sometimes this can imply that follow-up will be pursued after the patient has left the trial.
He/she transmits supplementary information to Pharmacovigilance (PV) of the DRC at the AP-HM using the SAE report form (ticking the box Follow-up n° X to show that it is a follow-up and not a primary report) within 48 hours of receiving it; He/she also transmits the final follow-up when the SAE is resolved or stabilized.
He/she keeps all of the documents concerning the adverse effect in order to, if necessary, complete previously transmitted information. He/she answers requests made by PV of the DRC in order to document the primary observation.

Role
Its role will be to check that the trial is run correctly, especially the incidence of adverse events. It will be regularly consulted as serious adverse events are reported.

Composition
The surveillance committee will be made up of two independent university-hospital intensive care physicians not taking part in the trial (Pr Eric Maury -Hôpital Saint-Antoine, Paris, Pr Elie Azoulay -Hôpital Saint-Louis, Paris) and one methodologist (Pr Christian Mélot -Hôpital Erasme de Bruxelles). Drug Safety) for their favorable opinion. An information sheet will be given to patients and informed consent must be obtained. It will be written in compliance with regulatory recommendations, notably providing the study objective, the advantages and the risks linked to the trial, the trial course and all of the legal provisions to which patients are entitled.

Sponsor's role
This research study will be performed in compliance with good clinical practice, representing a group of quality requirements in ethical and scientific fields, which must be respected during planning, implementation, conduct, follow-up, quality control, auditing, data collection, analysis and expression of the results. Respecting this good clinical practice will guarantee the protection of rights, protection and safety of the individuals taking part in this research study and preservation of their anonymity as well as the credibility (integrity, authenticity, verifiability) and accuracy of data and results in this research study.

Quality control and quality assurance
Each patient will be given a unique identification number, which will be generated on randomization and transmitted to the physician in charge of inclusion. This number will be copied in the eCRF. Quality control will be performed on-site by a clinical research associate.
This study will be conducted following the AP-HM standardized procedures when our Institute is the research sponsor. The course of research on all of the investigation sites as well as patient care will be performed in compliance with Helsinki recommendations and with Good Clinical Practice.
The sponsor's representatives will visit the investigation centers according to the defined procedure. Thus, a set-up meeting will be held in the presence of the sponsor at the opening of each center. The same thing will take place for the closure of each center.
Monitoring visits will take place for every 5 inclusions or every semester at the least if there have been less than 5 inclusions.

Submission to the ethics committee
This study will be submitted to the Comité de Protection des Personnes Sud-Méditerranée V (South Mediterranean Ethics Committee) following the sponsor's approval and obtaining insurance cover. This committee's opinion will be reported in the form sent to the appropriate authority by the sponsor before the research study commences.

Data processing -Confidentiality
Patients and/or their families will be individually informed in writing about the creation of a computer file for the management of collected data. People will be informed that they have the possibility to apply to the person in charge of the study for a right of access or modification of data concerning them.
Clinical and biological data will be stored in a database, the access to which will be controlled by an individual identification system for users (passwords). Computer hardware on which the data is stored will not be linked to a transmission network.
Concerning data processing for this project, which has a research objective in the health sector, it is within the framework of legal requirements, especially the law of 9 August Baseline methodology MR-001).
Medical and non-medical staff involved in this research are sworn to medical and occupational secrecy with relation to the data concerning patients gathered during the study.
Information gathered from patients will remain strictly confidential. It will be kept in paper format inside locked premises. It will be entered into a computer and automatically processed.
This computer processing will not allow direct or indirect identification of the patients. The participates must be informed about the type of data being processed, its finality, the identity of individual and legal entities who will be sent this data. He/she retains the right to access and rectify this data via the physician of his/her choice, as well as the right to object. In accordance with the law of 4 March 2002 relating to patients' rights and the quality of the health system, the global results of the trial can be transmitted to the patients and their representatives at their request, directly or via the physician of their choice.

Criteria for stopping the trial early
The sponsor and the investigator are entitled to stop the trial at any time before all of the planned subjects are included, for valid scientific, safety or administrative reasons. In this event, both parties agree to procedures based on individual evaluation per patient after critical analysis.
The trial can be discontinued and/or be analyzed before all of the subjects are included in the following cases: -If the trial is not led in accordance with the procedures planned in the protocol such as it was approved by administrative authorities.
-In the event of a suspected modification of the benefit-risk ratio (management of the number of adverse events) -At the sponsor's discretion if he/she thinks it necessary; furthermore the sponsor is entitled to request the exclusion of a subject in the event of a (major) breach of protocol, for administrative reasons, or for any other valid or ethical reason.

Publication rules
The first, fourth, one before last and last places of both of the two main articles (one for the CMV arm, one for the HSV arm) will be reserved for the principal investigators. The second place will be assigned to the center which included the most patients in that given arm.
The third, fifth, sixth, seventh and eighth places will be assigned to the five centers which included the most patients (after the first center).

Cost and additional costs of research
The sponsors ensures trial organization and covers the supply of elements such as the protocol, investigation site files, experimental drugs…Potential additional costs named in article R.2038 of the Code de la Santé Publique (French Public Health Regulations) will be subject to an agreement negotiated between the AP-HM and the representative of the investigation sites, taking into account the financial means available to the sponsor within the frame of promotion activities.
Two CRA coordinator/logistic assistance places (0.33 FTE and 0.66 FTE) will be credited over 30 months to work in close collaboration with the principal investigators of the 2 coordinator sites: Paris (HSV arm) and Marseille (CMV arm). They will be en liaison with the different centers participating in the study during the trial.
In order to help each clinical team to gather data, a flat rate of 200 euros will be granted per patient included in the interventional study. This will enable the sponsor's quality control CRA to limit his/her monitoring visits.
In order to reimburse the virology departments involved in the study, a flat rate of 150 euros will be granted for each patient screened.

Responsibilities of the sponsor / Insurance
In compliance with the law Huriet, the sponsor will take out insurance to cover its civil liability in the event of harmful consequences of research for the person taking part in it (art. L.1121-7),

Clinical pulmonary infection score (CPIS)
Modified