Adverse neuropsychiatric effects of cytomegalovirus prophylaxis with valaciclovir in renal transplant recipients

Nephrology Dialysis Transplantation, May 2006

Background. Valaciclovir (VACV) has been reported to induce adverse neuropsychiatric effects (ANE), especially in patients with renal failure, but few data are available for renal transplant recipients (RTR). Methods. We conducted a retrospective study in RTR given VACV as cytomegalovirus prophylaxis, from January 1999 to December 2000, to define the incidence rate, type and outcome of VACV-induced ANE, and to identify risk factors for ANE. The VACV-induced ANE were defined as neuropsychiatric disorders justifying VACV dose reduction or withdrawal. Patients with and without VACV-induced ANE were compared by univariate and multivariate analysis. Results. In all, 167 RTR were included, of whom 25 (15%) displayed VACV-induced ANE (mainly hallucinations and confusion), which occurred with a mean of 4 days after the start of VACV. ANE were reversible in all cases. Multivariate analysis showed that delayed graft function (DGF) was the main risk factor for VACV-induced ANE [Odds ratio (OR): 12.1; 95% CI = 3.4–43.4; P = 0.0001]. All VACV doses given to patients with ANE were in accordance with the current recommended adaptation to estimated glomerular filtration rate (GFR). Conclusion. In RTR, VACV-induced ANE are significantly frequent but reversible. DGF occurrence is the main risk factor for these ANE. In RTR with DGF, the recommended doses for GFR below 10 ml/min might be too high. Several strategies, in RTR with DGF, might lower the risk of ANE, including reduction of the currently recommended VACV dosage, delayed VACV introduction until improvement of renal function, or use of another anti-cytomegalovirus drug.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://ndt.oxfordjournals.org/content/21/5/1395.full.pdf

Adverse neuropsychiatric effects of cytomegalovirus prophylaxis with valaciclovir in renal transplant recipients

Vincent Das 1 2 Marie-Noelle Peraldi 0 1 Christophe Legendre 1 2 0 Service de Ne phrologie, Ho pital Saint-Louis , 1 avenue Claude Vellefaux, 75010 Paris , France 1 Service de Transplantation Re nale, Ho pital Necker - Enfants 2 Service de Transplantation R e nale, Ho pital Necker-Enfants Malades , 149 rue de Se` vres, 75015 Paris , France Background. Valaciclovir (VACV) has been reported to induce adverse neuropsychiatric effects (ANE), especially in patients with renal failure, but few data are available for renal transplant recipients (RTR). Methods. We conducted a retrospective study in RTR given VACV as cytomegalovirus prophylaxis, from January 1999 to December 2000, to define the incidence rate, type and outcome of VACV-induced ANE, and to identify risk factors for ANE. The VACVinduced ANE were defined as neuropsychiatric disorders justifying VACV dose reduction or withdrawal. Patients with and without VACV-induced ANE were compared by univariate and multivariate analysis. Results. In all, 167 RTR were included, of whom 25 (15%) displayed VACV-induced ANE (mainly hallucinations and confusion), which occurred with a mean of 4 days after the start of VACV. ANE were reversible in all cases. Multivariate analysis showed that delayed graft function (DGF) was the main risk factor for VACV-induced ANE [Odds ratio (OR): 12.1; 95% CI 3.4-43.4; P 0.0001]. All VACV doses given to patients with ANE were in accordance with the current recommended adaptation to estimated glomerular filtration rate (GFR). Conclusion. In RTR, VACV-induced ANE are significantly frequent but reversible. DGF occurrence is the main risk factor for these ANE. In RTR with DGF, the recommended doses for GFR below 10 ml/min might be too high. Several strategies, in RTR with DGF, might lower the risk of ANE, including reduction of the currently recommended VACV dosage, delayed VACV introduction until improvement of renal function, or use of another anti-cytomegalovirus drug. Introduction Cytomegalovirus (CMV) infection remains a major issue in renal transplantation, despite recent advances in prophylactic and curative treatments. CMV infection carries an increased risk of morbidity and mortality linked with CMV disease, and might increase the risk of acute rejection and chronic allograft dysfunction [1]. Nowadays, two drugs are commonly used as prophylaxis against CMV in renal transplant recipients (RTR): valaciclovir (VACV) and valganciclovir (VGCV). In a large-scale multicentre prospective randomized study, VACV was shown to prevent CMV infection in RTR, and was given as soon as possible after renal transplantation if either the donor or the recipient was seropositive for CMV [2]. After absorption, VACV is rapidly metabolized into aciclovir (ACV). After phosphorylation in CMV infected cells by the viral kinase UL97, ACV inhibits the replication of herpetoviridae. The bioavailability of ACV after VACV ingestion is much higher than after ACV ingestion, which explains VACV effectiveness against CMV. VACV is mainly eliminated by the kidneys either as ACV or as 9-(carboxymethoxymethyl)guanine, its metabolite [3]. Several cases of VACV-induced adverse neuropsychiatric effects (ANE) have been reported, mostly in the general or haemodialyzed population, and sometimes as case reports of RTR [47]. ACV-induced ANE have also been frequently reported [8]. Most patients with VACV- or ACV-induced ANE were elderly or had renal failure [68]. Many patients with VACV-induced ANE had ingested VACV doses that had not been adjusted to their renal function [6]. To date, no study has specifically dealt with VACV-induced ANE in RTR or tried to identify specific risk factors for such ANE. We therefore conducted a retrospective study in RTR, firstly, to describe the incidence, type, and outcome of VACV-induced ANE, and secondly, to identify risk factors for VACV-induced ANE, by univariate and multivariate analysis. Subjects and methods We conducted a retrospective study in the renal transplantation units of two teaching hospitals in Paris, France. All the patients who had received a renal allograft from January 1999 to December 2000 and had been given VACV as prophylaxis against CMV infection were eligible. Patients were excluded if they had received transplants other than kidney (e.g. liver, heart, pancreas, etc.) or if their medical and nursing charts were not available for at least 12 days after transplantation. Charts were studied until the 21st day after transplantation or until patient discharge. VACV prophylaxis against CMV infection was given if either the graft donor or the recipient was serologically positive for CMV. VACV was given after transplantation, as soon as oral intake was possible, until the end of the third month. VACV doses were adjusted to the estimated glomerular filtration rate (GFR), according to the usual guidelines, as follows: 1500 mg once a day for patients with GFR below 10 ml/min or dialyzed patients, 1500 mg twice a day for GFR between 10 and 25 ml/min, 1500 mg 3 times a day for GFR between 25 and 50 ml/min, 1500 mg 4 times a day for GFR between 50 and 75 ml/min, and 2000 mg 4 times a day for GFR above 75 ml/min. The following data were collected: Demographic data and medical history (i.e. age, gender, neuropsychiatric history, initial nephropathy and duration of dialysis before transplantation). Characteristics of the transplantation procedure: origin of the allograft (deceased donor or living related donor), donor age and cold ischaemia time. Post-transplantation outcome: occurrence of delayed graft function (DGF), defined as at least one haemodialysis session required during the 7 days after transplantation, time required to obtain an estimated GFR above 10 ml/ min, biological data (levels of serum creatinine, blood urea nitrogen, sodium and calcium at days 3, 7 and 15), presence or absence of oliguria (defined as diuresis of 500 ml/day or less) at days 3, 7 and 15, and the need for new general anaesthesia (yes or no). GFR was estimated from creatininaemia using the Cockcroft and Gault formulae. For statistical analysis, GFR was arbitrarily set at 10 ml/min for regularly dialyzed patients or patients who had had a dialysis session in the 24 h before creatininaemia measurement. The time of initiation and dosage of the following drugs were also recorded: VACV, immunosuppressors, antibiotics, benzodiazepines, neuroleptics and analgesics. VACV-induced ANE were defined as neuropsychiatric symptoms which occurred after the beginning of VACV treatment and induced dose reduction or withdrawal of VACV, with no other known cause of neuropsychiatric disorders. For patients with VACV-induced ANE, the following additional data were recorded: date of occurrence and type of effect(s), VACV dosages received from the day of transplantation until the day the effects occurred, time between VACV initiation and the occurrence of the effect(s), any withdrawal of VACV treatment or dose reduction, time from appearance of the effects to [1] improvement and [2] complete disappearance of the effects, whether or not VACV was restarted, and if it was, whether or not ANE recurred. When available, the level of plasma ACV at the time of ANE occurrence was also ascertained. Statistics To identify independent risk factors for ANE, patients with and without VACV-induced ANE were compared by univariate analysis and then by multivariate analysis. For univariate analysis, quantitative variables were compared using the Student t-test or Wilcoxon test, and qualitative variables by the Chi-square test or Fishers exact test, according to sample sizes. Variables for which univariate analysis showed a difference between patients with and without ANE, with a P-value below 0.10, were selected for multivariate analysis by logistic regression. We excluded variables with numerous missing data and those that could not be used to predict the occurrence of ANE (for example, variables for which data were collected after the ANE occurrence). The variables providing possible explanation for ANE occurrence were then selected by extraction from the model and tested against the initial model, using the 2logL value, where L is the model likelihood. The variables that significantly increased the maximum likelihood were kept in the model. This stage was repeated until no variable could be excluded from the model. The variables discarded after univariate analysis because their P-value was above 0.10 were then reintegrated into the model. If the inclusion of one of these variables significantly reduced 2logL, it was kept in the model. Statistical analyses were carried out using the 8.2 version of the SAS software. Results Patients vascular nephropathy (7%), secondary glomerulonephritis (7%), and diabetic nephropathy (1%). Incidence, type and outcome of VACV-induced ANE The mean actual duration of follow-up was 20.4 (2.1) days. VACV treatment was started a mean 3.80.9 days after transplantation. Of the 167 patients studied, 25 (15%) displayed VACV-induced ANE, which occurred after 4.43.4 days of VACV treatment (range: 117). The types of ANE observed are listed in Table 2. The most frequent were hallucinations and confusion. Only two patients had VACV dose reduction or withdrawal because of isolated insomnia (one patient) or isolated anxiety (one patient) without hallucination or confusion. In 19 cases, VACV was withdrawn. In the remaining six cases, VACV doses were reduced, but three of the patients concerned later required VACV withdrawal. All ANE improved after VACV withdrawal or dose reduction and 23 patients recovered completely. For the two others, complete recovery could not be established because psychiatric disorders were already present before transplantation. After VACV withdrawal or dose reduction, the mean time to complete recovery was 3.7 days (extremes: 26). VACV-induced ANE required treatment by benzodiazepine in 17 cases (68%) and neuroleptics in four (16%). Note that two patients required physical restraint for aggressiveness and agitation. VACV prophylaxis was resumed in nine patients, and ANE recurred in four of them. During the 6 months post-transplantation, nine patients (5%) displayed CMV infection requiring intravenous ganciclovir: seven patients had not displayed VACV-induced ANE (5%) and two had displayed ANE (8%). Risk factors for VACV-induced ANE Univariate analysis. The risk factors for VACVinduced ANE disclosed by univariate analysis are listed in Table 3. The following factors were not tested: cyclosporine and tacrolimus treatment, because they were started at different dates, depending on the patients GFR and on each centres specific practices; treatments given to all patients (such as trimethoprimsulfamethoxazole for prophylaxis against pneumocystis infection); and lastly, interactions between antibiotics, Type of ANE Number of patients (%) psychotropics, analgesics and VACV, because of interindividual differences between doses, molecules and the dates at which administration of the drug was started. Factors significantly associated with ANE occurrence after univariate analysis were: mean cold ischaemia time, DGF occurrence, mean time from transplantation to estimated GFR >10 ml/min, mean estimated GFR at day 3, mean blood urea nitrogen (BUN) level at day 3, presence of oliguria at day 3, and sirolimus use. Multivariate analysis. In the first step, variables for which univariate analysis showed a difference between patients with and without ANE with a P-value below 0.10 were selected for multivariate analysis. These variables were: cold ischaemia time, occurrence of delayed graft function, azathioprine use, sirolimus use, estimated GFR at day 3, and BUN level at day 3. Diuresis (i.e. the presence or absence of oliguria) was excluded from multivariate analysis because too many data were missing. Variables for which univariate analysis showed a difference with a P-value above 0.10 were then reintegrated into the model. Finally, three independent risk factors for the occurrence of VACV-induced ANE were identified by multivariate analysis (Table 4). The independent risk factor most significantly associated with the occurrence of ANE was DGF (Odds ratio 12.1, 95% CI 3.443.4, P 0.0001). The two other factors (use of rapamycin or induction therapy with antibodies) may also have increased the risk of ANE but did not reach significance. VACV doses given to the patients with ANE and levels of plasma ACV Medical and nursing charts were carefully examined to check all the VACV doses received by RTR with ANE from the day of transplantation until the day of ANE occurrence. All the doses given were in perfect accordance with those recommended for the estimated GFR of the patients, except for the two who had received lower doses than those recommended. Patients with ANE (n 25) Patients without ANE (n 142) Males* Mean recipient age (years, SD) Neuropsychiatric history* Psychotropic treatment before transplantation* First transplantation* Deceased donor* Mean donor age (years, SD) Mean cold ischaemia time (hours, SD) Delayed graft function* Mean time from transplantation to an estimated GFR >10 ml/min (days, SD) Mean estimated GFR at day 3 (ml/min, SD) Mean BUN level at day 3 (mmol/l, SD) Oliguria at day 3* Mean natraemia at day 3 (mM) Mean calcaemia at day 3 (mM) Corticosteroids* Induction therapy with antibodies* Azathioprine* Mycofenolate mofetil* Sirolimus* *Percentage of patients. DGF occurrence Sirolimus treatment Immunosuppressive induction therapy with antibodies Due to the retrospective design of our study, levels of plasma ACV, the VACV metabolite were known in five patients only. Four of these patients (two with hallucinations, one with headache and one with panic attack and anxiety) had peak level values higher than the 5 mg/l recommended for peak plasma ACV (12.2, 10.8, 8.3 and 5.7 mg/l, respectively). One patient, who displayed hallucinations, had a residual ACV level 24 h after last ingestion, at 4.9 mg/l, which was higher than the recommended value of 0.8 mg/l. Discussion Incidence, type and outcome of VACV-induced ANE To characterize VACV-induced ANE in an RTR population, we performed a retrospective study of a cohort of RTR representative of published RTR cohorts in Western Europe: most donors for such cohorts are deceased donors, and the high incidence of DGF reflects the decline in the quality of grafts due to donor scarcity. In the present series, the initial cause of end-stage renal disease (ESRD) was a hereditary or congenital nephropathy in a large number of cases, whereas diabetic nephropathy was rare. This probably resulted from: (1) the presence, in one of the two centres where this study was carried out, of a pediatric nephrology unit that usually sends its patients to the adult unit when they reach adult age; (2) the exclusion of recipients of kidney and pancreas allograft. However, it is unlikely that this recruitment bias could have affected ANE occurrence in our RTR series. Under these conditions, the incidence of VACVinduced ANE was significant (15%), and similar to the incidences reported in the literature. Indeed, Lowance et al. [2] conducted a prospective randomized study in 616 RTR to compare VACV vs placebo as prophylaxis against CMV infection. In their study, the incidence of ANE in the RTR receiving VACV was higher than in our series (31 vs 15%), but the prospective design of their study allowed all cases of ANE to be included, even those that were minor. Our retrospective definition of VACV-induced ANE, based on the consequence of ANE on VACV treatment, only identified clinically serious ANE, and therefore probably failed to identify all such minor effects as nightmares. Isolated minor ANE (insomnia, anxiety, nightmares) requiring VACV dose reduction or withdrawal were very rare. This is consistent with the fact that most of the VACV-induced ANE in our series were serious enough to require drug treatment by sedatives, or even physical restraint. The follow-up in our study was until day 21 after transplantation and not until the end of VACV prophylaxis (day 90). ANE occurring later than day 21 could have been missed, and the actual incidence of ANE during the entire prophylaxis duration may have been higher. However, in the literature, most VACV or ACV-induced ANE occurred in 17 days after the beginning of the treatment. As reported by others in the RTR or non-RTR populations [2, 58], most of the VACV-induced ANE in our series consisted of hallucinations and confusions, and occurred after about 4 days of VACV treatment. No patient experienced coma or convulsions, both of which have been reported with intravenous ACV [9], but very rarely with oral ACV or VACV. It is worth noting that most of the patients with ANE in our series had DGF and consequently were treated by haemodialysis, which may have limited the severity of ANE [8,10]. Note also that although our 25 cases of ANE were clinically serious enough to require treatment modification, ANE completely disappeared in all but two patients. However, these two patients, who did not recover completely, had developed psychiatric disorders before transplantation, so that retrospective assessment of complete recovery or its absence was not possible. Finally, clinically significant VACVinduced ANE seemed to be reversible in most cases, although we cannot exclude the possibility that they may have aggravated psychiatric disorders in predisposed patients. Risk factors for VACV-induced ANE Several risk factors for VACV- or ACV-induced ANE have been suggested by retrospective studies of nonRTR populations. Most authors have underlined the importance of renal failure as a predisposing condition. Thus in a retrospective study of 30 cases of ACVinduced ANE, two-thirds of the patients had renal failure [8]. Renal failure was also present in 21 of 43 cases of VACV-induced ANE that were notified to the French health authorities [6]. However, because ACV- and VACV-induced ANE have been reported in patients with normal renal function, other additional risk factors have been proposed, including older age, dyscalcaemia or dysnatraemia, uraemic encephalopathy, a state of immunosuppression and drug interactions [6,9], but none of them has been specifically tested. In our series of RTR, the most important independent risk factor for VACV-induced ANE was DGF. This result allowed precise identification of a group of RTR at a particularly high risk of ANE, who might benefit from a strategy for the prevention of VACV-induced ANE. Note that the presence of oliguria in our series was significantly associated with the occurrence of ANE by univariate analysis but was not tested as an independent risk factor by multivariate analysis. We cannot therefore exclude the possibility that patients with DGF and oliguria may be at greater risk of ANE than those with DGF without oliguria, due to differences in residual renal function and ACV elimination. Renal failure associated uraemia, dysnatraemia, or dyscalcaemia did not increase the risk of ANE occurrence. Older age has been proposed as a risk factor for ACV or VACV-induced ANE [6], but was not significantly associated with the occurrence of ANE in our study. However, a selection bias was obvious, since, in our series, the mean age of graft recipients was 44 years. We also tested whether the use of certain drugs was associated with VACV-induced ANE. Mycofenolate mofetil has been reported to compete with tubular secretion of acyclovir [11], but in our series its use was not associated with ANE occurrence. Induction therapy with polyclonal or monoclonal antibodies, and sirolimus, correlated with the occurrence of ANE by multivariate analysis, but these correlations did not reach significance and were not easy to explain. Once again, our study has several limitations: first, our retrospective definition of VACV-induced ANE probably failed to identify minor effects. Second, because of the limited duration of follow-up, later ANE may have been missed. Lastly, due to interindividual differences in dosages, medication and the date of initiation of nephrotoxic calcineurin inhibitors, antibiotics, psychotropics and analgesics, interactions between these drugs and VACV could not be tested as risk factors for ANE. These methodological limitations could have favoured the identification of DGF as the main risk factor for ANE. Other risk factors for early (before day 21) or late (after day 21) ANE should be tested in a prospective study with well-defined protocols and extended follow-up duration. The mechanisms of ACV and VACV neurotoxicity are not fully understood. In the literature, the association of these drugs with renal failure, and the reported improvement of symptoms by haemodialysis [8] suggested a dose-dependent mechanism, which was confirmed by the correlations reported between ANE occurrence and high levels of plasma ACV [12]. However, many authors concluded that a clear relationship between ANE and the level of plasma ACV could not be established [13]. This was proposed to result from the difference between blood and brain equilibration rates. On the basis of these observations, some authors questioned the value of measuring plasma ACV. Because of the retrospective design of our study, these values were available only for five patients with ANE, and of course were not measured in patients without. However, in the RTR populations, plasma ACV levels could be tested prospectively as a means of detecting increased risks of ANE, although this strategy would require real-time measurement and adaptation of VACV dosage. Has VACV been correctly used for RTR with ANE? Numerous cases of VACV or ACV-induced ANE have been reported to result from misuse of these drugs, and failure to adapt their dosage to patients reduced renal function [6]. A careful review of the doses given to the 25 RTR with ANE in our series established that none of them had been given a dose higher than those recommended for their renal function. DGF had occurred in 22 of them (88%), suggesting that the recommended doses for an estimated GFR of 10 ml/min or less might be too high, at least for RTR with DGF. If so, this probably resulted from decreased renal elimination of ACV, although we cannot exclude the possibility that increased absorption by the gut in patients with ESRD and/or DGF might help to increased ACV bioavailability. Strategies for VACV-induced ANE prevention The main conclusion of our study is that the RTR at high risk of future VACV-induced ANE are those with DGF. These RTR, who run a 12-fold higher risk of VACV-induced ANE than those without DGF, might benefit from prevention strategies designed to reduce the risk of ANE. In our series, the recipients of a graft from a living donor had neither DGF nor VACV-induced ANE. Consequently, the promotion of transplantation with grafts from living donors would reduce the occurrence of DGF and therefore the risk of VACV-induced ANE. In RTR with DGF, VACV could also be replaced by an alternative drug, for example VGCV, which is the most commonly prescribed antiviral medication for CMV prophylaxis. However, prophylactic treatments with VGCV and VACV have not been directly compared, and oral or intravenous ganciclovir have not proved to be more effective or safer than VACV [14,15]. Furthermore, in RTR, early discontinuation of VGCV due to VGCV-induced adverse effects has also been reported [16]. Concerns have also been raised about the emergence of ganciclovir-resistant CMV after prolonged use of oral ganciclovir [17], VGCV [18], and VACV [19]. The aim of our study was not to promote the use of any drug, but to help the physicians who commonly use VACV in RTR to better control the safety profile of VACV. Therefore, several other strategies could be proposed to reduce the risk of VACV-induced ANE. Thus, in RTR with DGF, the start of VACV treatment could be postponed until renal function improves. Alternatively, VACV doses could be reduced, because the recommended dose for estimated GFR of 10 ml/min or less (1500 mg per day) seemed too high for our RTR with DGF. However, the effectiveness of these two strategies should first be tested to ensure that they would not weaken VACV effectiveness in combating CMV infection. Two recent studies suggested that lower VACV doses than usually recommended (i.e. up to 23 g/day instead of 8 g) might be sufficient to prevent CMV infection in RTR [20,21]. Other alternative strategies, based on increased surveillance of RTR with DGF receiving VACV, could also be undertaken. For example, the levels of plasma ACV could be measured regularly and its dosage reduced if the level exceeds a certain threshold, which remains to be determined; or more simply, increased attention could be paid to the RTR with DGF receiving VACV, and the dosage could be rapidly adjusted or the drug withdrawn if the patient exhibited minor neuropsychiatric symptoms such as anxiety, or nightmares. In conclusion, this retrospective study showed that reversible but clinically significant VACV-induced ANE frequently occur in RTR. DGF is the main risk factor for ANE, and therefore RTR with DGF might benefit from prevention strategies, which, however, remain to be defined. Such strategies could include VACV dose reduction, delayed introduction of VACV until the improvement of renal function, more intensive surveillance, or VGCV use instead of VACV. In addition, the interactions between VACV and other drugs commonly used for RTR could be more thoroughly examined in a future prospective study. Conflict of interest statement. None declared.


This is a preview of a remote PDF: https://ndt.oxfordjournals.org/content/21/5/1395.full.pdf

Vincent Das, Marie-Noelle Peraldi, Christophe Legendre. Adverse neuropsychiatric effects of cytomegalovirus prophylaxis with valaciclovir in renal transplant recipients, Nephrology Dialysis Transplantation, 2006, 1395-1401, DOI: 10.1093/ndt/gfk031