Epilepsy Research (2009) 85, 65—71
journal homepage: www.elsevier.com/locate/epilepsyres
Hyperammonemia following intravenous valproate loading Jennifer L. DeWolfe a,∗, Robert C. Knowlton a,1, Mark T. Beasley b,2, Stacey Cofield b,2, Edward Faught a,1, Nita A. Limdi a,1 a
University of Alabama at Birmingham Epilepsy Center, Department of Neurology, 1719 6th Avenue South, CIRC 312, Birmingham, AL 35294-0021, United States b University of Alabama at Birmingham School of Public Health, Department of Biostatistics, 1665 University Boulevard, Ryals Public Health Building, Room 327 Birmingham, AL 35294-0022, United States Received 30 September 2008; received in revised form 5 February 2009; accepted 8 February 2009 Available online 18 March 2009
KEYWORDS Hyperammonemia; Valproate; Loading doses
Summary Background: Valproic acid (VPA) has been associated with hyperammonemia with and without encephalopathy. We report the frequent but transient nature of hyperammonemia following intravenous (IV) administration of loading doses of VPA. Methods: Forty participants received a VPA loading dose (20 or 30 mg/kg) at 6 or 10 mg/kg/min. All participants were monitored for signs of systemic and local intolerance. Serum VPA level, ammonia, complete blood count, bilirubin, transaminases, pancreatic enzymes, and level of consciousness were obtained at baseline, 1 and 24 h after administration. Changes in ammonia levels were assessed using repeated-measures ANOVA. Results: Asymptomatic hyperammonemia occurred in 30 of 40 participants at 1 h post-VPA infusion. Majority of the participants (66%) demonstrated decreasing ammonia concentrations at 24 h post-infusion. Multivariable repeated-measures analysis indicates the lack of influence of VPA dose (p = 0.8), VPA levels (p > 0.24, all time points), infusion rate (p = 0.41) and gender (0.68) on ammonia levels across time. Age (p = 0.015), time since dosing (p = 0.017) and co-therapy with enzyme-inducing antiepileptic drugs (p = 0.035) were significant predictors of changes in ammonia levels. Conclusions: Hyperammonemia is a frequent but transient finding following intravenous administration of loading doses of VPA. Hyperammonemia was not associated with alteration in consciousness or hepatic transaminases. Published by Elsevier B.V.
∗
Corresponding author. Tel.: +1 205 934 3866; fax: +1 205 975 6255. E-mail addresses:
[email protected] (J.L. DeWolfe),
[email protected] (R.C. Knowlton),
[email protected] (M.T. Beasley), scofi
[email protected] (S. Cofield),
[email protected] (E. Faught),
[email protected] (N.A. Limdi). 1 Tel.: +1 205 934 3866; fax: +1 205 975 6255. 2 Tel.: +1 205 934 4905; fax: +1 205 975 2540. 0920-1211/$ — see front matter. Published by Elsevier B.V. doi:10.1016/j.eplepsyres.2009.02.012
66
Introduction Valproate (VPA) is widely used in the treatment of epilepsy, migraines and psychiatric disorders and has an increasing role in the management of behavioral problems in elderly participants (Perucca, 2002). Additionally, VPA has emerged as an alternate medication for status epilepticus (Hirsch, 2007; Misra et al., 2006). Adverse effects associated with therapy include tremor, weight gain, coagulopathies, gastrointestinal effects, hepatic toxicity and hyperammonemia with or without encephalopathy that may be potentially fatal in people with urea cycle disorders (Depacon® , Abbott Laboratories, 2006; Sewell et al., 1995; Davis et al., 2000; Perucca, 2002; Nicolai and Carr, 2008; Hirsch, 2007), in addition to idiosyncratic reactions. Hyperammonemia is frequently reported in participants on VPA therapy (Murphy and Marquardt, 1982; Zaccara et al., 1987, 1984b). Although asymptomatic in most participants (Murphy and Marquardt, 1982), valproate-induced hyperammonemia can produce gastrointestinal distress and symptoms of encephalopathy manifested as malaise, lethargy, and mental slowing, which may progress to marked sedation, focal neurological deficits, vomiting, altered mental status, coma, and even death (Zaccara et al., 1984a). Sedation is also a side effect of valproate, making it more difficult to distinguish between the side effects of valproate and underlying hyperammonemia. Hyperammonemia may occur at normal therapeutic valproate blood levels in the absence of any abnormalities in liver enzymes (Altunbasak et al., 1997; Beghi et al., 1990; Marescaux et al., 1985; Murphy and Marquardt, 1982; Wyllie et al., 1983; Zaret and Marini, 1985). VPA-induced hyperammonemia may have a gradual onset and can present after years of stable maintenance therapy (Nicolai and Carr, 2008; McCall and Bourgeois, 2004; Dealberto, 2007). Herein we report the incidence of hyperammonemia among participants receiving large rapid intravenous loading doses of valproate.
Methods This was an open-label, prospective trial of rapid intravenous loading of valproate sodium in participants with epilepsy. Participants over 19 years of age taking valproate to control seizures or had an epilepsy indication for its use were eligible to participate in the study. Patients with known allergy to VPA, liver disease, status epilepticus (SE) defined as continuous seizure activity >5 min, or recurrent seizures > 30 min without return to baseline in between seizures, or concurrent high-dose lamotrigine therapy (>200 mg/day) were excluded. Due to study time constraints, the possibility of development of hyperammonemia and liver failure with status epilepticus (Decell et al., 1994), and the possibility of VAP associated hyperammonemia-inducing SE (Velioglu and Gazioglu, 2007), patients in SE were excluded. The absence of liver disease was confirmed by assessment of admission laboratory values and medical history. All subjects provided informed consent before entering the study. The study protocol necessitated a 24 h observation period after administration of VPA loading dose (Limdi et al., 2007). The study was approved by the Institutional Review Board at the University of Alabama at Birmingham (UAB). The study protocol was developed with input from the FDA (IND # 62165).
J.L. DeWolfe et al. Dose administration Valproate sodium injection (Depacon® , Abbott Laboratories 2006) was supplied in 5-ml single-use vials. Each milliliter contained valproate sodium equivalent to 100 mg of valproic acid. All doses were decided by the treating physician. Twenty participants received a VPA loading dose of 20 mg/kg and 20 participants 30 mg/kg. Within each dose group, the participants were equally divided into two cohorts of 10 to receive VPA infusion at the rates of 6 or 10 mg/kg/min. All doses were administered undiluted by manual injection using a stop watch to time administration rate. The patency of IV access was confirmed with a saline flush before and after dose administration. Valproate maintenance dosing was initiated within 8 h post-infusion as directed by the treating physician.
Determination of valproate plasma concentrations Plasma concentrations of VPA (total and unbound) were measured at baseline, 5, 10, 15, 30, 60 and 240 min post-infusion in the first 20 participants. Based on pharmacokinetic analysis from the first 20 participants, the frequency for drug concentration monitoring was reduced to the baseline, 30, 60 and 240 min time points. All participants had VPA plasma concentrations at 24 h. All VPA concentrations were analyzed by MedTox laboratories (St. Paul, MN) by ultrafiltration immunoassay. The lower limits of quantification for both total and unbound VPA in plasma were 0.7 g/ml.
Monitoring for adverse effects Cardiovascular parameters (heart rate, blood pressure, respiratory rate, oxygen saturation and electrocardiogram) and signs of local irritation at the infusion site were monitored at 2.5 min intervals for the first 20 min, then at 30, 45, 60 min and 4 h. The level of consciousness (LOC) tool of the NIH Stroke Scale was used to assess level of consciousness, orientation, and comprehension at baseline, then between 30 and 60 min after VPA administration, and at 24 h to assess for change in LOC from baseline.
Laboratory monitoring Ammonia, complete blood count with platelets, total bilirubin, direct bilirubin, calculated indirect bilirubin, hepatic transaminases (AST, ALT), alkaline phosphatase, amylase and lipase were measured at baseline (before infusion), 1 and 24 h after rapid infusion of valproate. All samples were collected from a second indwelling peripheral venous catheter (not used for drug administration). Special attention was paid to specimen handling for analysis of ammonia levels. All samples were placed on ice immediately after collection and processed within 20 min as per laboratory protocol at the UAB (NW, 1994). Plasma ammonia levels were determined using the Ammonia Kit in conjunction with SYNCHRON® systems (Beckman Coulter Inc.). In the assay reaction, glutamate dehydrogenase catalyzes the condensation of ammonia and ketoglutarate to glutamate with the concomitant oxidation of reduced P-nicotinamide adenine dinucleotide phosphate (NADPH) to P-nicotinamide adenine dinucleotide phosphate (NADP+ ). The amount of NADPH oxidized is directly proportional to the amount of ammonia in the sample. The normal reference range of ammonia levels at the UAB laboratories is 15—54 mol/l (Ratliff and Hall, 1982). Hyperammonemia was defined as ammonia levels >54 mol/l.
Statistical analysis Chi-square tests were used to assess differences in the frequencies in categorical variables and one-way ANOVA was used to assess differences in the continuous variables and in the 2 dose groups (20 and 30 mg/kg). Differences in continuous variables exhibiting
Hyperammonemia following intravenous valproate loading unequal variances were evaluated using Welch-ANOVA. The ammonia levels in participants receiving VPA were compared across time (baseline, 1 and 24 h) using a repeated-measures ANOVA with dose (mg/kg), rate (mg/kg/min), time, VPA serum levels, age, gender, and concurrent enzyme-inducing AEDs as predictors. All analyses were performed using JMP version 5.1 (SAS Institute, Cary, NC) at a non-directional ˛ level of 0.05.
Results 40 participants (27 men) mean age 39 years (SD ± 14.6 range 19—77 years) received VPA loading doses as part of their treatment. The majority of the participants were white (n = 28) or black (n = 9). Two participants were Hispanic and one Asian. The average weight was 81.6 (±23.1) kg and average dose was 1993.8 (±526.7) mg. The two dose groups (20 and 30 mg/kg) did not differ with regard to age (p = 0.44) or gender (p = 0.73). Rapid administration of VPA was well tolerated with no significant changes in vital signs. Details on tolerability and safety of rapid administration of VPA loading doses are presented in a recent publication (Limdi et al., 2007). Table 1 summarizes VPA dosing, VPA plasma levels and ammonia levels by dose group (20 and 30 mg/kg). Participants on co-therapy enzyme-inducing antiepileptic drugs (EIAEDs) such as phenytoin, carbamazepine and phenobarbital were more likely to receive a higher loading dose
Table 1
67 (30 mg/kg) of VPA. Fifteen participants (11 in the 20 mg/kg dose group) were on prior VPA therapy. VPA concentrations did not differ at baseline (Table 1) although the large variance and small sample size make this comparison tenuous. Loading dose of 30 mg/kg produced higher VPA concentrations (total and unbound) at 30 and 60 min compared to dose of 20 mg/kg. The difference in VPA concentration was not significant at 4 h. Dutta et al. (2007) present a detailed pharmacokinetic report in a recent publication. Participants in the two dosing groups did not differ with regard to laboratory parameters at baseline. There were no significant changes in platelets, total bilirubin (indirect and direct), hepatic transaminases (ALT, AST), alkaline phosphatase, amylase or lipase following VPA administration (Table 2). Baseline ammonia levels were similar in both dose groups (p = 0.57, Table 1). Sixteen participants had hyperammonemia at baseline. There was no correlation between baseline ammonia levels and baseline VPA levels (total or unbound, p = 0.18 and 0.38 respectively) or prior VPA therapy. Four participants receiving 30 mg/kg and eleven participants receiving 20 mg/kg dose were on prior VPA therapy. Thirty participants had hyperammonemia at 1 h postinfusion, of which 10 had elevated ammonia levels at baseline. Therefore 20 of 30 patients (66%) developed new hyperammonemia. Concurrent therapy with EIAEDs was more frequent among patients with elevated baseline
VPA dosing, VPA plasma levels, and ammonia levels by dose group (20 and 30 mg/kg).
Parameter
20 mg/kg (n = 20)
30 mg/kg (n = 20)
p
Valproate dose (mg)
1796.1 (551.7)
2191.6 (427.6)
0.016
Co-therapy with enzyme-inducing AED No Yes
17 (85.0%) 3 (15.0%)
10 (50.0%) 10 (50.0%)
0.016
VPA concentration (50—100 g/ml) Baseline (n = 15)a 30 min 60 min 4h
12.0 129.2 118.0 75.7
(12.7) (36.8) (20.6) (19.3)
28.8 165.3 141.9 79.2
(35.9) (41.3) (35.1) (26.3)
0.18 0.01 0.013 0.65
0.9 23.3 21.5 10.1
(0.5) (8.5) (7.1) (5.6)
7.7 50.2 39.1 17.5
(12.7) (23.5) (27.8) (18.5)
0.08 0.0002 0.012 0.1
Unbound VPA concentration (6-20 g/ml) Baseline (n = 15)a 30 min 60 min 4h Albumin (3.4—5 g/dl) Ammonia (18—54 mol/l) Baseline 1 hb 24 hc
3.6 (0.3)
3.4 (0.6)
0.21
53.9 (22.8) 68.8 (24.2) 56.2 (23.3)d
58.6 (27.2) 92.5 (38.2) 67.6 (21.5)
0.57 0.03 0.12
Continuous variables are denoted as mean and SD (in parenthesis). Enzyme-inducing AEDs: Phenytoin, Phenobarbital, Carbamazepine, Primidone. Reference ranges for all laboratory parameters denoted in parenthesis. VPA levels were determined by MedTox laboratories (St. Paul, MN). Reference ranges denoted in parenthesis. Participants with missing laboratory measurements at specific time points were excluded from analysis at that specific time point. Significant p-values are given in bold. a 4 participants in the 30 mg/kg and 11 participants in the 20 mg/kg group were on prior VPA therapy. b 30 of 40 participants developed hyperammonemia at 1 h post-infusion. c 23 of 30 participants with hyperammonemia 1 h post-infusion had decreased levels at 24 h post-infusion. d Excludes participant with ammonia level of 326 mol/l at 24 h where the sample processing was delayed.
68
J.L. DeWolfe et al.
Table 2 Laboratory parameters including hepatic aminotransferase and ammonia levels on study participants. Parameter
20 mg/kg (n = 20)
30 mg/kg (n = 20)
p-Value
227.9 (62.4) 236.8 (56.2) 233.9 (51.4)
0.61 0.57 0.68
1.05 (0.6) 1.05 (0.8) 0.83 (0.3)
0.66 0.29 0.97
AST (0—37 units/l) Baseline 29.8 (12.0) 1h 27.0 (8.3) 24 h 27.7 (13.4)
29.8 911.4) 30.3 (12.9) 28.6 (16.4)
1.00 0.33 0.86
ALT (6—45 units/l) Baseline 26.3 (19.1) 1h 27.3 (18.8) 24 h 30.0 (32.2)
29.7 (23.7) 29.3 (24.4) 29.4 (24.3)
0.61 0.77 0.94
Alk-Phos (39—117 units/l) Baseline 67.6 (24.9) 1h 65.9 (24.3) 24 h 67.0 (28.4)
84.1 (39.2) 28.4 (38.0) 81.6 (34.2)
0.12 0.11 0.15
Amylase (14—151 units/l) Baseline 85.6 (32.8) 1h 82.0 (27.9) 24 h 89.0 (35.8)
96.6 (40.5) 96.8 (38.8) 102.2 (60.5)
0.35 0.17 0.42
Lipase (0—61 units/l) Baseline 28.3 (7.3) 1h 27.3 (6.6) 24 h 29.9 (7.4)
29.7 (10.1) 28.7 (9.8) 29.5 (8.7)
0.62 0.60 0.89
Platelets (150—400 × 103 ) Baseline 237.9 (61.6) 1h 226.4 (58.0) 24 h 227.1 (53.9) Total bilirubin (0.0—1.0 mg/dl) Baseline 0.97 (0.5) 1h 0.85 (0.3) 24 h 0.83 (0.6)
Figure 1 Ammonia levels by dose group across time (baseline through 24 h). Univariate repeated-measures analysis indicate a transient increase (at 60 min) in ammonia levels in both dose groups, with a greater increase in the 30 mg/kg group (p = 0.04). The 24 h ammonia levels did not differ from baseline levels (p = 0.16) despite inclusion of an extremely high outlying value (326 mol/l).
administration. Ammonia levels were not influenced by rate of administration (p = 0.41) or gender (p = 0.69). Age (p = 0.015), time since VPA loading dose (p = 0.017) and cotherapy with EIAEDs (p = 0.035) were the only significant predictors. Participants on EIAEDs exhibited higher ammonia levels (difference 15.6, 95% CI [1.1, 30.1]) compared to participants without EIAEDs (Fig. 2). The small sample size did not allow analyses of the effect on individual AEDs on VPA-induced changes in ammonia levels. Among participants on EIAEDs the ammonia levels did not differ by VPA dose (20 mg/kg versus 30 mg/kg, p = 0.34).
Continuous variables are denoted as mean and SD (in parenthesis). Reference ranges for all laboratory parameters denoted in parenthesis. Participants with missing laboratory measurements at specific time points were excluded from analysis at that specific time point.
ammonia levels (58.3%) compared to those with normal ammonia levels (47.7%) although this difference was not statistically significant (p = 0.12). At 24 h post-infusion, 24 (80%) participants (with normal baseline levels) had decreased ammonia levels compared to 1 h post-infusion. Univariate analysis indicates that administration of VPA increased ammonia levels transiently in both dose groups. At 1 h postloading dose participants receiving a 30 mg/kg dose had significantly higher ammonia levels than those receiving 20 mg/kg (p = 0.03, Table 1; multivariable analysis p = 0.04, Fig. 1). Ammonia levels at 24 h were less than at 1 h in both groups. The ammonia level at 24 h did not differ between the two dose groups (p = 0.86) indicating transience of its elevation. Multivariable repeated-measures analysis was used to evaluate factors influencing ammonia levels across time (baseline through 24 h). Changes in ammonia levels were independent of VPA dose (p = 0.86) and VPA levels at 30 min (p = 0.35), 1 h (p = 0.83) and 4 h (p = 0.24) post-
Figure 2 Ammonia levels across time stratified by co-therapy with EIAEDs. Multivariable repeated-measures analysis indicate ammonia levels were independent of VPA dose (p = 0.86) and VPA levels at 30 min (p = 0.35), 1 h (p = 0.83) and 4 h (p = 0.24) postadministration. Ammonia levels were not influenced by rate of administration (p = 0.41) or gender (p = 0.69). Age (p = 0.015) and time since dosing (p = 0.017) and co-therapy with EIAEDs (p = 0.035) were the only significant predictors. The 24 h ammonia levels did not differ from baseline levels (p = 0.16) despite inclusion of an extremely high outlying value (326 mol/l).
Hyperammonemia following intravenous valproate loading Although there was a transient surge in ammonia levels 1 h post-infusion, most ammonia levels decreased at 24 h post-infusion and lacked statistical significance compared to baseline (p = 0.16). The 24 h ammonia level in one participant (who received 20 mg/kg) was 326 mol/l. We could not rule out laboratory error and therefore repeated hepatic transaminases and ammonia levels which were within normal limits. The participant did not complain of adverse effects or alterations in level of consciousness. All analyses included this extreme outlying value. Ammonia levels were not associated with changes in LOC. Two participants with fluctuating LOC at baseline (brain tumor) also exhibited similar fluctuation in LOC postdosing despite transient elevation of ammonia in one. In 12 participants ammonia levels at 24 h exceeded baseline and 54 mol/l without change in LOC. One participant had insignificant elevation in LOC at 1 h post-infusion that resolved at 24 h post-infusion. All 40 participants had normal NIH Stroke Scale LOC measurement scores at 24 h postinfusion.
Discussion Hyperammonemia, ammonia levels >54 mol/l, is a common finding in participants receiving VPA (Murphy and Marquardt, 1982; Wyllie et al., 1983; Zaret et al., 1982). This is corroborated by the frequent but asymptomatic hyperammonemia following VPA dosing in 30 participants (75%) in our study. The average increase in ammonia level at 1 h was 72% over baseline levels. At 24 h the average increase of ammonia levels over baseline was 15%. To our knowledge this is the first report of the incidence of hyperammonemia following intravenous administration of loading dose (20 or 30 mg/kg) of VPA in adult epileptic participants with normal hepatic function. Although several publications have documented the prevalence of hyperammonemia in epileptic participants on chronic VPA therapy, direct comparisons cannot be drawn since most data are derived from case reports, case series and small observational studies representing a heterogeneous participant population with wide variation in age, baseline comorbidity and concomitant AED therapy (Murphy and Marquardt, 1982; Rao et al., 1993; Ratnaike et al., 1986; Rawat et al., 1981; Wyllie et al., 1983; Zaccara et al., 1987, 1984a,b, 1985; Zaret et al., 1982). However, two publications addressing the acute hyperammonemic effects of VPA loading are pertinent in approach and comparable in study design. The contribution of renal ammoniagenesis in elevating arterial ammonia levels following intravenous administration of 1500 mg loading doses was evaluated in fifteen participants (Warter et al., 1983b). The significant rise in renal vein ammonia concentrations against the modest rise in arterial and hepatic vein concentrations indicated the lack of effect of VPA on hepatic detoxification of ammonia. A subsequent study demonstrated the significant contribution of both renal ammoniagenesis and altered hepatic ammonia metabolism to elevated ammonia levels in participants on concomitant phenobarbital therapy (Warter et al., 1983c). Zacarra et al. studied short-term adverse effects (alteration in consciousness and drowsiness) associated with venous ammonia levels following an 800 mg oral
69 VPA loading dose in 24 epileptic participants on concurrent AED therapy. Ammonia levels increased in all participants, although this increase was independent of VPA levels and was not associated with adverse effects. Seven participants with more significant increases in ammonia levels (following loading doses) reported adverse effects following 3—9 days of chronic therapy indicating that excessive increase in ammonia levels may predict later VPA-induced alteration in consciousness (Zaccara et al., 1987). We understand that the method of blood sample collection and rapidity of analytic determination can significantly influence ammonia levels (Warter et al., 1983a; Zaccara et al., 1984a). Although an arterial blood sample is preferred, collection is invasive and not routinely performed in clinical practice. Venous sampling reflects routine clinical practice. However, we recognize that venous ammonia levels may overestimate the degree of hyperammonemia (Warter et al., 1983b). Although we paid close attention to the laboratory protocols, we recognize that ammonia levels may not have been measured within 20 min for all samples and acknowledge delay in processing of samples for two participants. Although we recognize that even minor delays in analysis could potentially erroneously raise ammonia levels, all levels available were included in the analysis since this reflects situations we as clinicians face. Second, it is possible that ammonia levels may have peaked after 1 h post-infusion and frequent monitoring may have more accurately identified the time at which levels peak. However, the decline in ammonia concentrations at 24 h indicates that the levels peak early (before 24 h). Third, as patients were only monitored for 24 h post-administration, we cannot comment on alterations in laboratory parameters or level of consciousness that may have occurred after discharge. Because of study time limitations and possible complications of hyperammonemia and liver failure in patients with status epilepticus (Velioglu and Gazioglu, 2007) the possibility of development of status epilepticus secondary to VPA-induced hyperammonemia (Decell et al., 1994), patients in SE were excluded. We acknowledge that use of VPA in SE is a current clinical practice (Hirsch, 2007) and the focus of VPA-induced hyperammonemia in this patient population needs further study. Finally we recognize our sample size was inadequate to detect significant differences in ammonia levels during the observation period as this was not the primary aim of the study. Ammonia levels in participants taking VPA can be influenced by physical exertion (seizures) (Maclean et al., 1994, 1996) and timing and protein content of meals (Gidal et al., 1997; Laub, 1986). Since participants were being monitored in the Epilepsy Monitoring Unit, physical exertion was minimal except for participants in whom therapy was rapidly initiated due to seizures. Since participant enrollment into the study was determined by clinical necessity, we could not regulate the timing of VPA administration with regard to time since seizure(s), chronologic time or with regard to ingestion of meals. Moreover study participants did not receive standardized meals. Participant characteristics such as presence of mental retardation (Kane et al., 1992; Williams et al., 1984), urea cycle disorders (Leao, 1995; Oechsner et al., 1998; Sewell et al., 1995; Depacon® , Abbott Laboratories, 2006), VPA dose and plasma levels (Beghi et al., 1990), poor nutri-
70 tional status, EIAEDs (Haidukewych et al., 1985; Kondo et al., 1992; Murphy and Marquardt, 1982; Williams et al., 1984; Zaccara et al., 1984a, 1985) can also affect the variation in ammonia levels. None of our participants had mental retardation and were fairly well nourished as indicated by albumin levels. Our findings lend support to several previous reports with regard to the influence of EIAEDs on the increased occurrence of hyperammonemia. This held true in both univariate and multivariate analysis. However, we did not have a sample size sufficient to detect differences in hyperammonemia by individual EIAEDs. Ammonia levels in participants receiving 30 mg/kg VPA dose were higher than in those receiving 20 mg/kg VPA dose at 60 min postadministration. This increase was transient as indicated by the lack of difference in ammonia levels at 24 h compared to baseline. Contrary to some reports (Beghi et al., 1990), we found no correlation between ammonia levels and VPA levels (total and unbound). There was no correlation between elevated ammonia levels and other indices or hepatic or pancreatic function; however, the follow-up of 24 h may have been too short to detect changes. Two participants deserve special mention. One subject with baseline elevated bilirubin indices and liver transaminases tested positive for hepatitis C 5 weeks after enrollment. This subject had developed a transient hyperammonemia that returned to baseline within 24 h. In one previous report (Felker et al., 2003), a possible association of VPA with elevations in liver transaminases in hepatitis C positive participants was noted. However, to our knowledge, no study has addressed the association with hyperammonemia. Another participant with elevations in baseline total and indirect bilirubin exhibited a transient increase in ammonia with no associated alteration in consciousness and is suspected to have Gilbert’s disease. There are no data on the use of VPA in participants with Gilbert’s disease. VPA, through the interplay of several mechanisms, can cause hyperammonemia with or without encephalopathy and encephalopathy with or without hyperammonemia (Chen et al., 2001; Coulter and Allen, 1980; Verrotti et al., 2002; Vossler et al., 2002). The absence of change in level of consciousness in hyperammonemic participants up to 24 h post-infusion in our study is in contrast to these reports (Chen et al., 2001; Coulter and Allen, 1980; Paganini et al., 1984; Verrotti et al., 2002; Vossler et al., 2002). These discordant findings could be explained by factors such as varying age, concomitant therapy with multiple EIAEDs, poorly controlled seizures, determination of prevalence versus incidence, etc. Since hyperammonemic encephalopathy may develop several hours, even after chronically on stable VPA doses (Dealberto, 2007; McCall and Bourgeois, 2004; Nicolai and Carr, 2008), after VPA is initiated, it is possible that the development of encephalopathy may have gone undetected after the participant was discharged at 24 h post-infusion. Participants did not develop other hyperammonemia associated symptoms such as gastrointestinal complaints (Dealberto, 2007; McCall and Bourgeois, 2004; Nicolai and Carr, 2008) during the observation period. Our results suggest ammonia levels should not serve as a sole determinant of regulating VPA therapy in asymptomatic participants with normal hepatic function. Measurement of ammonia levels is recommended to ensure transience and facilitate etiologic determination in participants who
J.L. DeWolfe et al. develop decreasing levels of consciousness, experience worsening seizures, or develop gastrointestinal symptoms. If alteration in consciousness does occur in VPA treated participants (with or without concurrent hyperammonemia), a comprehensive evaluation (for possible occult liver disease, urea cycle disorder, dietary contribution, etc.) must be conducted regardless of the ammonia levels. We advise caution when using VPA therapy in participants with a history of unexplained encephalopathy.
Conclusion Hyperammonemia is a frequent but transient (lasting less than 24 h in most cases) finding following rapid intravenous administration of loading doses of VPA. Age, time post-administration, and co-therapy with EIAEDs influence changes in ammonia levels after rapid administration of loading doses. This transient increase was not influenced by size of the loading dose, serum VPA level, or rapidity of administration. Hyperammonemia was not associated with alteration in consciousness or elevation of hepatic transaminases up to 24 h post-infusion. In our experience the increase in ammonia levels following VPA loading is usually transient. However, since our participants were free of liver disease and encephalopathy we caution against extrapolation of our findings to participants with these conditions.
Acknowledgments We are grateful to all the participants that participated in the study. We thank the attending physicians on the neurology inpatient services and the nursing staff of the Neurology patient care unit, especially Suzanne Miller RN, BSN, for their help during the study.
References Abbott Laboratories, 2006. Depakote® Tablets Package Insert, revised ed. Abbott Park, IL. Altunbasak, S., Baytok, V., Tasouji, M., Herguner, O., Burgut, R., Kayrin, L., 1997. Asymptomatic hyperammonemia in children treated with valproic acid. J. Child Neurol. 12, 461—463. Beghi, E., Bizzi, A., Codegoni, A.M., Trevisan, D., Torri, W., 1990. Valproate, carnitine metabolism, and biochemical indicators of liver function. Collaborative Group for the Study of Epilepsy. Epilepsia 31, 346—352. Chen, W.T., Yen, D.J., Yu, H.Y., Liao, K.K., 2001. Valproate-induced encephalopathy. Chung Hua i Hsueh Tsa Chih—–Chin. Med. J. 64, 474—478. Coulter, D.L., Allen, R.J., 1980. Secondary hyperammonaemia: a possible mechanism for valproate encephalopathy. Lancet 1, 1310—1311. Davis, L.L., Ryan, W., Adinoff, B., Petty, F., 2000. Comprehensive review of the psychiatric uses of valproate. J. Clin. Psychopharmacol. 20, 1S—17S. Dealberto, M.J., 2007. Valproate-induced hyperammonaemic encephalopathy: review of 14 cases in the psychiatric setting. Int. Clin. Psychopharmacol. 22, 330—337. Decell, M.K., Gordon, J.B., Silver, K., Meagher-Villemure, K., 1994. Fulminant hepatic failure associated with status epilepticus in children: three cases and a review of potential mechanisms. Intensive Care Med. 20, 375—378.
Hyperammonemia following intravenous valproate loading Dutta, S., Faught, E., Limdi, N.A., 2007. Valproate protein binding following rapid intravenous administration of high doses of valproic acid in patients with epilepsy. J. Clin. Pharm. Ther. 32, 365—371. Felker, B.L., Sloan, K.L., Dominitz, J.A., Barenes, R.F., 2003. The safety of valproic acid use for patients with hepatitis C infection. Am. J. Psychiatry 160, 174—178. Gidal, B.E., Inglese, C.M., Meyer, J.F., Pitterle, M.E., Antonopolous, J., Rust, R.S., 1997. Diet- and valproate-induced transient hyperammonemia: effect of L-carnitine. Pediat. Neurol. 16, 301—305. Haidukewych, D., John, G., Zielinski, J.J., Rodin, E.A., 1985. Chronic valproic acid therapy and incidence of increases in venous plasma ammonia. Therap. Drug Monit. 7, 290—294. Hirsch, L.J., 2007. The status of intravenous valproate for status. Epilepsy Curr. 7, 96—98. Kane, R.E., Kotagel, S., Bacon, B.R., Vogler, C.A., 1992. Valproate use associated with persistent hyperammonemia and mitochondrial injury in a child with Down’s syndrome. J. Pediat. Gastroenterol. Nutr. 14, 223—227. Kondo, T., Ishida, M., Kaneko, S., Hirano, T., Otani, K., Fukushima, Y., Muranaka, H., Koide, N., Yokotama, M., Nakata, S., et al., 1992. Is 2-propyl-4-pentenoic acid, a hepatotoxic metabolite of valproate, responsible for valproate-induced hyperammonemia? Epilepsia 33, 550—554. Laub, M.C., 1986. Nutritional influence on serum ammonia in young patients receiving sodium valproate. Epilepsia 27 (January—February), 55—59. Leao, M., 1995. Valproate as a cause of hyperammonemia in heterozygotes with ornithine-transcarbamylase deficiency. Neurology 45, 593—594. Limdi, N.A., Knowlton, R.K., Cofield, S.S., Ver Hoef, L.W., Paige, A.L., Dutta, S., Faught, E., 2007. Safety of rapid intravenous loading of valproate. Epilepsia 48, 478—483. Maclean, D.A., Graham, T.E., Saltin, B., 1994. Branched-chain amino acids augment ammonia metabolism while attenuating protein breakdown during exercise. Am. J. Physiol. 267, E1010—E1022. Maclean, D.A., Graham, T.E., Saltin, B., 1996. Stimulation of muscle ammonia production during exercise following branchedchain amino acid supplementation in humans. J. Physiol. 493, 909—922. Marescaux, C., Warter, J.M., Brandt, C., Rumbach, L., Micheletti, G., Chabrier, G., Imler, M., 1985. Adaptation of hepatic ammonia metabolism after chronic valproate administration in epileptics treated with phenytoin. Eur. Neurol. 24, 191—195. McCall, M., Bourgeois, J.A., 2004. Valproic acid-induced hyperammonemia: a case report. J. Clin. Psychopharmacol. 24, 521—526. Misra, U.K., Kalita, J., Patel, R., 2006. Sodium valproate vs phenytoin in status epilepticus: a pilot study. Neurology 67, 340—342. Murphy, J.V., Marquardt, K., 1982. Asymptomatic hyperammonemia in patients receiving valproic acid. Arch. Neurol. 39, 591—592. Nicolai, J., Carr, R.B., 2008. The measurement of ammonia blood levels in patients taking valproic acid: looking for problems where they do not exist? Epilepsy Behav. 12, 494—496 (discussion 497—8). NW, T., 1994. Specimen Collection and Processing: Sources of Biological Variation. W.B. Saunders, Philadelphia, PA. Oechsner, M., Steen, C., Sturenburg, H.J., Kohlschutter, A., 1998. Hyperammonaemic encephalopathy after initiation of valproate therapy in unrecognised ornithine transcarbamylase deficiency. J. Neurol. Neurosurg. Psychiatry 64, 680—682. Paganini, M., Zaccara, G., Campostrini, R., Valenza, T., Angelastro, R., Bartelli, M., Arnetoli, G., Zappoli, R., 1984. Venous blood ammonia concentrations in adult epileptic patients are increased by treatment with valproic acid. Acta Neurol. 6, 442—446.
71 Perucca, E., 2002. Pharmacological and therapeutic properties of valproate: a summary after 35 years of clinical experience. CNS Drugs 16, 695—714. Rao, M.L., Stefan, H., Scheid, C., Kuttler, A.D., Froscher, W., 1993. Serum amino acids, liver status, and antiepileptic drug therapy in epilepsy. Epilepsia 34, 347—354. Ratliff, C.R., Hall, F.F., 1982. Ammonia in plasma enzymatic procedure. Selected Methods Clin. Chem. 9, 85—90. Ratnaike, R.N., Schapel, G.J., Purdie, G., Rischbieth, R.H., Hoffmann, S., 1986. Hyperammonaemia and hepatotoxicity during chronic valproate therapy: enhancement by combination with other antiepileptic drugs. Br. J. Clin. Pharmacol. 22, 100—103. Rawat, S., Borkowski JR., W.J., Swick, H.M., 1981. Valproic acid and secondary hyperammonemia. Neurology 31, 1173—1174. Sewell, A.C., Bohles, H.J., Herwig, J., Demirkol, M., 1995. Neurological deterioration in patients with urea cycle disorders under valproate therapy—–a cause for concern. Eur. J. Pediat. 154, 593—594. Velioglu, S.K., Gazioglu, S., 2007. Non-convulsive status epilepticus secondary to valproic acid-induced hyperammonemic encephalopathy. Acta Neurol. Scand. 116, 128—132. Verrotti, A., Trotta, D., Morgese, G., Chiarelli, F., 2002. Valproateinduced hyperammonemic encephalopathy. Metab. Brain Dis. 17, 367—373. Vossler, D.G., Wilensky, A.J., Cawthon, D.F., Kraemer, D.L., Ojemann, L.M., Caylor, L.M., Morgan, J.D., 2002. Serum and CSF glutamine levels in valproate-related hyperammonemic encephalopathy. Epilepsia 43, 154—159. Warter, J.M., Brandt, C., Marescaux, C., Rumbach, L., Micheletti, G., Chabrier, G., Krieger, J., Imler, M., 1983a. The renal origin of sodium valproate-induced hyperammonemia in fasting humans. Neurology 33, 1136—1140. Warter, J.M., Imler, M., Marescaux, C., Chabrier, G., Rumbach, L., Micheletti, G., Krieger, J., 1983b. Sodium valproate-induced hyperammonemia in the rat: role of the kidney. Eur. J. Pharmacol. 87, 177—182. Warter, J.M., Marescaux, C., Brandt, C., Rumbach, L., Micheletti, G., Chabrier, G., Imler, M., Kurtz, D., 1983c. Sodium valproate associated with phenobarbital: effects on ammonia metabolism in humans. Epilepsia 24, 628—633. Williams, C.A., Tiefenbach, S., Mcreynold, J.W., 1984. Valproic acid-induced hyperammonemia in mentally retarded adults. Neurology 34, 550—553. Wyllie, E., Wyllie, R., Rothner, A.D., Erenberg, G., Cruse, R.P., 1983. Valproate-induced hyperammonemia in asymptomatic children. Cleveland Clin. Q. 50, 275—277. Zaccara, G., Campostrini, R., Paganini, M., Messori, A., Valenza, T., Arnetoli, G., Zappoli, R., 1987. Long-term treatment with sodium valproate: monitoring of venous ammonia concentrations and adverse effects. Therap. Drug Monit. 9, 34—40. Zaccara, G., Campostrini, R., Paganini, M., Moroni, F., Valenza, T., Targioni, G., Arnetoli, G., Zappoli, R., Baruzzi, A., 1984a. Acute changes of blood ammonia may predict short-term adverse effects of valproic acid. Neurology 34, 1519—1521. Zaccara, G., Paganini, M., Campostrini, R., Arnetoli, G., Zappoli, R., Moroni, F., 1984b. Hyperammonemia and valproate-induced alterations of the state of consciousness. A report of 8 cases. Eur. Neurol. 23, 104—112. Zaccara, G., Paganini, M., Campostrini, R., Moroni, F., Valenza, T., Messori, A., Bartelli, M., Arnetoli, G., Zappoli, R., 1985. Effect of associated antiepileptic treatment on valproate-induced hyperammonemia. Therap. Drug Monit. 7, 185—190. Zaret, B.S., Beckner, R.R., Marini, A.M., Wagle, W., Passarelli, C., 1982. Sodium valproate-induced hyperammonemia without clinical hepatic dysfunction. Neurology 32, 206—208. Zaret, B.S., Marini, A.M., 1985. Standardization in VPA-induced hyperammonemia. Neurology 35, 136—137.
