Byung In Lee MD
Professor, Department of Neurology, Yonsei University College of Medicine; and Chief, Department of Neurology, Severance Hospital, Seoul, Korea
Zonisamide (ZNS) is a broad-spectrum antiepileptic drug (AED) that was first synthesized by Uno et al. (1). ZNS became commercially available for use in Japan in 1989 and in Korea in 1993. However, its development in the United States and Europe was prematurely terminated because of a high incidence of renal stones (2). In Japan, ZNS was not associated with a higher incidence of renal stones and continued its successful development. Clinical studies of ZNS were reinstituted in the United States in 1992, and ZNS was approved for clinical use in 1999.
Most of the clinical experience with ZNS is from Japan. The European and U.S. experience with ZNS has been limited to a few clinical studies conducted during its early stage of development (2, 3, 4, 5, 6). One of two controlled clinical trials performed in the United States (6,7) have not been published yet, but their results are available through the data files of Dainippon Pharmaceutical Co., Ltd., Osaka, Japan (8).
MOST COMMONLY OBSERVED ADVERSE EFFECTS
In the European double-blind, placebo-controlled study of ZNS add-on therapy (5), the incidence of adverse effects (AEs) was 59% in the ZNS group and 27% in the placebo group. Most common AEs reported in the ZNS group were fatigue (22.5%), dizziness (16.9%), somnolence (14.1%), anorexia (12.7%), psychomotor slowing (11.3%), ataxia (11.3%), nervousness (9.9%), abdominal pain (7.0%), and confusion (5.6%). Among those, dizziness, anorexia, psychomotor slowing, ataxia, and confusion were more common in the ZNS group than the placebo group. In one U.S. trial (6), the incidence of AEs was higher for both ZNS and placebo groups, 92% and 58%, respectively. However, in the other U.S. trial (7), the incidence of AEs was lower and comparable between the two groups, which might be related to the adoption of a slower dose escalation of ZNS in this study.
The incidences of AEs in the pooled data of these three double-blind, placebo-controlled trials (8) were 78.1% for ZNS and 61.3% for placebo. Somnolence, ataxia, anorexia, dizziness, fatigue, nausea and vomiting, and irritability were the most common AEs (Table 95.1). AEs that occurred more than twice as often in the ZNS than in the placebo group were ataxia, anorexia, irritability, diplopia, impaired concentration, insomnia, abdominal pain and discomfort, and depression. Premature dropout from the placebo-controlled trials because of AEs occurred in 31 of 269 patients (11.5%) treated with ZNS and 15 of 230 patients (6.5%) on placebo. Among 84 AEs reported from 31 patients prematurely withdrawn from the trials during ZNS treatment, psychiatric or behavioral symptoms were the most common (28.6%), followed by cognitive impairment (10.7%), fatigability (10.7%), anorexia (9.5%), and headache (7.1%). Skin rash and asymptomatic renal calculi were found in one patient each.
In a comparative trial of ZNS and carbamazepine (CBZ) conducted in Japan (9), the incidence of AEs was 52% in the ZNS group and 57% in the CBZ group. Common AEs with ZNS were similar to those in the placebo-controlled trials, with a significantly higher incidence of anorexia in the ZNS group and ataxia in the CBZ group (Table 95.2).
A double-blind, comparative trial of ZNS and CBZ in newly diagnosed epilepsies was conducted in Korea (10). The design of the study was basically similar to the monotherapy trial of lamotrigine conducted by Brodie et al. (11). The incidence of AEs was 67.1% in the ZNS group and 53.7% in the CBZ group, which was not significantly different. Common AEs in the ZNS group in order of frequency were anorexia, abdominal pain or discomfort, dizziness, weight loss, and fatigue, compared with dizziness, somnolence, skin rash, headache, and fatigue in the CBZ group (Table 95.2). Most AEs were reported during the dose-escalation phase (ZNS, 54.8%; CBZ, 46.3%). After the dose-escalation phase, AEs were reported in 50.0% of the ZNS group and 27.1% of the CBZ group (p = .006), but in only 13.2% of the ZNS group and 8.6% of the CBZ group were these AEs newly reported ones (p = .4). The results suggested that the AEs related to ZNS treatment
lasted longer than the AEs associated with CBZ (p = .02). Premature withdrawal from the study due to AEs occurred in 11 patients from each group. However, the AE-precipitated premature withdrawals from the study were quite different between the two groups: for the ZNS group, AEs were gastrointestinal (GI) symptoms (four patients), anorexia (three patients), renal stone (two patients), psychomotor slowing (one patient), and skin rash (one patient), compared with skin rash (eight patients), somnolence (two patients), and exacerbation of seizures (one patient) in the CBZ group.
TABLE 95.1. TREATMENT-RELATED EMERGENCY ADVERSE EVENTS OCCURRED IN >5% OF PATIENTS IN DOUBLE-BLIND PLACEBO-CONTROLLED CLINICAL TRIALS (POOLED DATA)
TABLE 95.2. ADVERSE EFFECTS DEVELOPED IN >5% OF PATIENTS IN THE COMPARATIVE CLINICAL TRIALS BETWEEN ZONISAMIDE AND CARBAMAZEPINE
Three large-scale open trials are worth mention. In the Japanese series of 1,008 adult patients (12), AEs were reported by 51.3%, and 185 patients (18%) discontinued ZNS. The most common AEs were drowsiness (24%), ataxia (13%), anorexia (11%), GI symptoms (7%), decrease
in spontaneity (6%), and psychomotor slowing (5%). In 55 patients given ZNS monotherapy, the incidence of AEs was 29%, compared with 41% in patients taking one concomitant drug and 55% in patients taking two or more concomitant drugs. Another Japanese open trial of ZNS that recruited 393 pediatric patients (13) revealed that the incidence of AEs was 43.8%. Somnolence was the most common AE (26.2%), followed by ataxia (10.2%), anorexia (9.2%), decreased mental activity (7.1%), and enervation (6.9%). In the open trial of ZNS conducted in the United States (2), 136 of 167 patients (81.4%) reported adverse events and 21 patients (12.6%) were prematurely withdrawn from the study due to AEs. Common AEs were anorexia, asthenia, ataxia, confusion, fatigue, and psychomotor slowing. Most AEs were reported during the first 4 weeks after introduction.
From these clinical studies, the common AEs related to ZNS treatment could be identified: symptoms related to the central nervous system (somnolence, dizziness, ataxia, fatigue, anorexia), GI symptoms (abdominal discomfort or pain, nausea and vomiting), and cognitive dysfunction (psychomotor slowing, decreased concentration, and memory impairment).
Dose-Dependent Incidence and Prevalence
An earlier clinical study (2) suggested a higher incidence of AEs in patients taking doses above 5 mg/kg or greater than 300 mg/day. AEs that were more common at the highest doses and may be dose related included amnesia, ataxia, visual disturbances, headache, confusion, dysarthria, and somnolence.
From the pooled data of controlled trials (8) (Table 95.1), the incidence of AEs was lower at lower doses (<200 mg/day) and peaked at doses of 400 to 599 mg/day, which were the most common doses used in the trials, and then leveled off at doses greater than 600 mg/day. This leveling-off of AEs might reflect the duration of exposure (adaptation effect). The importance of adaptation effects on AEs was more apparent in the analysis of the blood level-AE relationship of the pooled data (8). The incidence of patients reporting AEs was higher at lower plasma levels of ZNS: 55.7% at a concentration of 0 to 10 µg/mL, 51.1% at 10 to 20 µg/mL, 39.9% at 20 to 30 µg/mL, 35.8% at 30 to 40 µg/mL, and 19.0% at >40 µg/mL. This is consistent with the fact that the highest incidence of AEs was found during the first month of treatment, when initial doses were low and being titrated upward. Mimaki (14) also reported a poor relationship between dose or blood level of ZNS and the emergence of AEs. The effects of ZNS on cognition were assessed in nine patients with refractory epilepsy (15). At minimal steady-state plasma concentrations of >30 µg/mL, ZNS adversely affected acquisition and concentration of new information in the absence of clinical signs and symptoms of toxicity. However, repeated tests after 24 weeks of ZNS therapy revealed improvement in cognitive AEs.
All these findings suggest that most of the common AEs emerge in a dose-related manner at the start of ZNS treatment but that an adaptation process takes place over 4 to 8 weeks. Thus the prevalence of AEs decreases as the duration of exposure increases.
From the pooled data of placebo-controlled trials (8), the prevalence of AEs was highest (60.2%) during the first month of ZNS treatment and decreased to 34.4% during the third month of treatment, which was a gradual reduction rather than an abrupt change. However, the incidence of first occurrence of most common AEs dropped sharply after the first month to become similar in both ZNS- and placebo-treated patients from the second month, with a few exceptions. Anorexia and GI symptoms became less frequent during the second month, but still were more frequent than with placebo. The incidence of first occurrence of depression and anxiety did not decrease during the first 2 months. These features also were noticed in the ZNS monotherapy trial in newly diagnosed patients (10), in which the prevalence of AEs was significantly higher in ZNS than CBZ group during the maintenance phase, but the number of patients reporting the first occurrence of AEs was comparable between the two groups after the first 4 weeks of dose escalation. Therefore, common AEs usually appear during the first month of ZNS therapy, but patients appear to adapt to most untoward side effects over 1 to 2 months of treatment, and they do not represent a major clinical concern after prolonged treatment. According to the pooled data of Dainippon Pharmaceutical Co. (8), the incidence of first occurrence of AEs after 6 months of ZNS treatment did not exceed 5% for any specific AE except rhinitis.
Analysis of clinical data revealed that the incidence of common AEs was higher in cases of rapid dose escalation, higher doses of ZNS, and two or more concomitant AEDs. From the pooled data of controlled trials, the incidence of AEs was slightly higher in patients aged 40 to 65 years than in patients younger than 40 years of age: 85.5% and 74.9%, respectively. The number of patients aged 65 years and older was too small to be adequately analyzed; however, the trend of higher incidences of AEs in patients older than 40 years of age certainly suggests a need for caution with ZNS therapy in older patients. Because ZNS is metabolized in the liver and mainly excreted through the kidneys, patients with hepatic or renal diseases may require special caution. No specific studies have been conducted to evaluate the safety of ZNS in these groups of patients.
Advice Concerning Precaution and Management
ZNS is available in tablet (100 mg/tablet) and powder (200 mg/g) forms. The recommended initial dosage is 100 to 200 mg/day for adults and 2 to 4 mg/day for children. Because the elimination half-life is long (60 hours or 27 to 36 hours in patients taking enzyme-inducing AEDs), the steady-state blood level is reached 7 to 14 days after drug administration, and the peak-to-trough difference in plasma concentration is small (27% in once-daily dosing and 14% in twice-daily dosing) (16). Therefore, ZNS is administered once or twice a day and doses should be titrated at 2-week intervals while efficacy and tolerability are carefully monitored. Maintenance dosages of ZNS in adults and children are 200 to 400 mg/day and 4 to 8 mg/day, respectively, and the usual maximum dosage is 600 mg/day for adults and 12 mg/day for children. However, in the author's opinion, the current recommendation for ZNS titration may be too fast to avoid the common AEs that usually occur during the first 4 to 8 weeks of drug initiation. Given a slow process of adaptation to most of the common AEs, it seems preferable to start with an initial dosage of 50 mg/day and escalate the dose by 50 mg every 2 weeks until the initial target dosage of 200 mg/day.
It usually is recommended to monitor the blood level of ZNS. Therapeutic blood level ranges have been reported to be 20 to 30 µg/mL, and the plasma concentration of >30 µg/mL has been associated with the emergence of AEs (3,15,17). However, a recent investigation (14) and the analysis of pooled data (8) did not demonstrate any clear concentration-response relationships. Therefore, dose adjustment and therapeutic blood level monitoring should be based on clinical judgment. Further clinical studies on the schedules of dose escalation and dose-response or concentration-response relationships are needed. Withdrawal of ZNS may cause aggravation of seizures (5.1%) or status epilepticus (1.3%) (8). Among nine patients who were confirmed to have withdrawal effects, status epilepticus developed in two after abrupt discontinuation of ZNS, but in none after gradual tapering. Therefore, a gradual tapering of ZNS is recommended.
LESS COMMON BUT CLINICALLY RELEVANT ADVERSE EFFECTS
Earlier ZNS development in the United States and Europe was stopped because of a high incidence of renal stones (2). Overall, 13 of 505 subjects (2.6%) in the U.S. and European series had symptomatic kidney stones. Four stones have been analyzed; one was mostly urate and the others were primarily calcium oxalate and calcium phosphate (18). Yagi and colleagues (19) measured 24-hour urine calcium, magnesium, citrate, and phosphate excretion and found a significant decrease in urine citrate excretion, but no change in the other substances. Compared with the Western trials, the incidence of renal stones was very low in Japan. Only 2 of 1,008 patients (0.2%) treated with ZNS were found to have renal stones, and both patients had a history nephrolithiasis in one parent (12).
In 1992, Dainippon Pharmaceutical Co. restarted clinical trials of ZNS in partnership with the Institute of Biological Research and Development-Rostrum Global (IRG). Renal ultrasound examinations were conducted at baseline, after 12 weeks of treatment, annually, and at discontinuation of therapy to determine whether ZNS is associated with an increased incidence of renal calculi. Thirteen of 429 (3.0%) ZNS-treated patients and 2 of 85 (2.4%) placebo-treated patients developed calculi (either symptomatic or asymptomatic) in the IRG studies (8). The duration of ZNS therapy was ≤6 months in four patients, 7 to 12 months in four patients, and ≥12 months in five patients. ZNS dosages for these patients were 400 to 600 mg/day in 11 patients and >600 mg/day in 2 patients. One additional healthy volunteer who did not have an ultrasound study had symptoms of renal colic after 24 days of ZNS administration while participating in a clinical pharmacology study. Follow-up assessment of seven patients who remained in the study despite the occurrence of renal calculi revealed that three had no evidence of calculi, two patients had evidence of calculi on follow-up examination, and two other patients developed symptoms of renal colic and passed calculi. In summary, the incidence of renal calculi in ZNS-treated patients was 3.3%, compared with 2.4% in placebo-treated patients (8).
Reasons for the lower incidence of real calculi related to ZNS treatment in Japan still are unclear but may be related to racial, dietary, or other environmental factors. In fact, the incidence of renal calculi in the general population seems lower in Japan than in Western countries (20, 21, 22), which might be related to the lower incidence of ZNS-induced renal calculi in Japan. However, there is evidence to suggest a causal relationship between ZNS and renal calculi even in Japan. Recently, Kubota et al. (23) reported that renal calculi developed in three patients, and the ZNS monotherapy trial in Korea found development of renal calculi in two patients (2.7%) (10). Kubota et al. (23) suggested alkaline urine and hypercalciuria were risk factors for development of renal calculi during ZNS treatment. It is reasonable to perform routine urinalysis before and during ZNS treatment in daily practice.
From the analysis of pooled data of controlled trials, ZNS was found to cause cognitive dysfunctions more frequently than placebo: difficulty concentrating (8.2% versus 0.9%), slowed thought (3.3% versus 1.7%), and forgetfulness (7.1% versus 2.2%). Berent et al. (15) conducted serial neuropsychological assessments in nine patients with refractory
epilepsies under polypharmacy. ZNS adversely affected memory quotient (MQ) scores at the end of 12 weeks of treatment, and there was a linear relationship between the change in MQ and plasma levels of ZNS. However, repeated tests at the end of 24 weeks of ZNS treatment revealed a significant recovery of MQ and a loss of relationship between MQ and plasma levels of ZNS. In the subtest analysis, ZNS appeared to affect specific cognitive functions such as acquisition and consolidation of new information and verbal learning at 12 weeks of ZNS treatment. Previously learned material, psychomotor performance, and visual-perceptual learning were not affected.
These results suggest that ZNS adversely affects specific cognitive dysfunctions, especially verbal learning, in a dose- or concentration-dependent manner during the acute stage, but an adaptation process occurs subsequently with a gradual recovery of cognitive function. The cognitive effects of ZNS require further investigation, and it seems likely that a lower starting dose and slower dose escalation of ZNS may help decrease the emergence of cognitive dysfunction.
Oligohidrosis or a heat stroke-like episode was found to occur in children taking ZNS. Shimizu et al. (24) reported a child who presented with a heat stroke-like episode during ZNS treatment. A sweating test using pilocarpine iontophoresis revealed a marked reduction in the sweat response, which suggested a postganglionic sweating dysfunction. A skin biopsy did not show any abnormalities of sweat glands, and oligohidrosis resolved within 2 weeks of withdrawal from ZNS. Okumura et al. (25) reported that 12 of 70 epileptic children treated with ZNS had oligohidrosis. Although none of them presented with heat stroke-like episodes, reversible oligohidrosis was confirmed by either heat-loading tests or acetylcholine-loading tests. Although oligohidrosis has not been reported in adults yet, the a presence of hyperpyrexia in patients taking ZNS needs to be carefully investigated for the possibility of oligohidrosis.
In the placebo-controlled trials, weight loss was found in only 3.3% (8); however, a greater than 5% loss of baseline body weight has been shown in 15% of patients taking ZNS in the monotherapy trial conducted in Korea (10), and a U.S. controlled trial (7) reported that 21.6% of ZNS-treated patients lost more than 2.3 kg. It is likely that weight loss is related to anorexia, which is one of the most common AEs related to ZNS treatment.
ZNS has teratogenic effects in mice, rats, and dogs at a higher dosage than the human maximal daily dosage (26), and it causes a higher incidence of spontaneous abortions at the maximal human daily dose (10 mg/kg) (27).
Kondo et al. (28) assessed the risk of teratogenicity of ZNS in humans. They found 26 pregnancies from 22 mothers between 1989 and 1994; 4 were exposed to ZNS alone and the others to multiple drugs. Artificial abortions were performed in 4 mothers and congenital malformations were found in 2 of 24 offspring, anencephaly in one and atrial septal defect in the other. All mothers exposed to ZNS alone delivered normal infants. The authors concluded that the risk of ZNS teratogenicity is no greater than that of other conventional AEDs. However, such a risk cannot be neglected because of the small sample size of this study. ZNS currently is classified as category C.
ZNS is recovered in breast milk in an almost equal concentration to that in plasma, with the ratio of 0.93 ± 0.09 (29), which needs to be considered in the nursing mother. Interactions between ZNS and oral contraceptive pills have not been investigated.
Abnormal Laboratory Test Results
ZNS usually does not alter clinical laboratory test results, and no patients were withdrawn from the controlled clinical trials because of abnormal results on such tests. In one Japanese open trial (13), abnormal values, based solely on deviation from the normal range, were found in 5% to 10% of patients during the 2-year study period, but all were insignificant, and ZNS was discontinued in only 5 of 393 patients (1.3%); they consisted of anemia in 1 patient, elevated alanine aminotransferase (ALT) in 3, and neutropenia in 1. In the other open trial (9), clinical laboratory testing conducted in 909 patients revealed elevated values for r-gamma glutamyl transpeptidase (4.9%), alkaline phosphatase (2.6%), ALT (2.3%), and aspartate aminotransferase (1.4%). A mild degree of hypocalcemia was found in 5.9% of patients in the pooled data of Dainippon Pharmaceutical Co. (8).
Decreased immunoglobulin A (IgA), and sometimes of other immunoglobulins as well, was reported in 14 patients during postmarketing surveillance in Japan (8). Hypoimmunoglobulinemia is well known to occur in patients taking other conventional AEDs, especially phenytoin (30,31). Maneoka et al. (32) reported a patient in whom IgA and IgG2 subclass deficiency developed associated with ZNS therapy. After cessation of ZNS, the serum IgA level was rapidly recovered and the IgG2 level gradually increased but remained subnormal. They recommended that serum immunoglobulins be checked in patients having recurrent infections.
POTENTIALLY LIFE-THREATENING ADVERSE EFFECTS
Life-threatening AEs related to ZNS treatment were rare, and their causal relationships have not been well established.
Serious, irreversible hematologic disorders have not been reported. However, Leppik et al. (2) reported a case of severe leukopenia that resolved after withdrawal of ZNS. In a Japanese open trial (9), leukopenia was found in 18 patients (2%). According to the Dainippon serious AE reports (8), 22 hematologic adverse events were reported in 19 patients: pancytopenia in 2, leukopenia in 11, and thrombocytopenia in 5. In 7 of 19 patients, clinicians considered it unlikely that ZNS was causally related to the AEs. Aplastic anemia was reported in two patients; one had multiple myeloma and the other recovered after withdrawal of ZNS.
The incidence of skin rash from the pooled data of placebo-controlled trials (8) of ZNS was 3.0% in the ZNS group and 1.3% in the placebo group, with only one patient taking ZNS being withdrawn from the controlled trials. In Japanese open trials, skin rash/itching occurred in 2% (9) and 1.3% (13) of patients. Twenty-six instances of serious skin disorders were reported during postmarketing surveillance in Japan (8). Nineteen were Stevens-Johnson syndrome and seven were toxic epidermal necrolysis. However, most patients were taking multiple drugs, and a causal relationship between the skin disorder and ZNS theory could be established in only one case.
Psychiatric symptoms such as depression, irritability, mania, paranoia, hallucinations, or psychosis have been reported in association with ZNS treatment. In the pooled data of placebo-controlled trials (8), the incidences of irritability (11.5% ZNS versus 5.2% placebo), depression (7.4% versus 3.0%), anxiety (5.6% versus 2.6%), paranoia (1.9% versus 0.4%) and hallucination (1.5% versus 0.0%), were more than twice those with placebo. Psychiatric symptoms were reported as serious AEs in four patients (1.5%) in ZNS groups, compared with two patients (0.9%) in placebo groups.
Several authors in Japan reported psychotic episodes or behavioral abnormalities in patients taking ZNS (33,34). One case each of ZNS-induced paranoia (4) and mania (35) were reported from the U.S. trial. Kimura (36) reported two children in whom ZNS-induced severe behavioral disturbances developed that resolved after withdrawal of ZNS. Recently, Miyamoto et al. (37) reported 14 patients in whom psychotic episodes developed that met the International Classification of Diseases, 10th edition criteria for organic delusional disorder and organic hallucinosis. Most patients made a good recovery with antipsychotic medications and discontinuation of ZNS. They estimated that the incidence of psychotic symptoms in their patient population was 13%, which was higher than the previously reported prevalence of epileptic psychosis (38). Five of the 14 patients who experienced psychosis were mentally retarded.
These findings suggest that ZNS may be implicated in the genesis of various psychiatric disturbances. However, most patients who experienced psychiatric disturbances were taking multiple drugs because of refractory epilepsies, and clear documentation of their causal relationship to ZNS was difficult except in a few cases. ZNS affects the synthesis and degradation of monoamine neurotransmitters (39), especially dopamine and norepinephrine, and its binding sites are related to the γ-aminobutyric acid/benzodiazepine receptor ionophore complex (40), which might be responsible for the induction of psychiatric disturbances in some vulnerable patients.
Sudden Unexplained Death in Epilepsy
Review of sudden unexplained death in epilepsy (SUDEP) associated with ZNS revealed an incidence of approximately 3.1 deaths per 1,000 patients (8), which was similar to the incidence of SUDEP of 4 per 1,000 associated with lamotrigine and gabapentin. The incidence of SUDEP in patients with epilepsy varied widely, but was higher in patients with refractory epilepsies. The incidence of SUDEP in ZNS trials was considered within the range described for the epileptic patient population as a whole.
MANIFESTATIONS AND MANAGEMENT OF OVERDOSE
Naito et al. (41) reported a patient who attempted suicide by taking 7,400 mg of ZNS, 126 mg of clonazepam, and 4,000 mg of CBZ. She was found comatose and was hospitalized. In the hospital, she was comatose with bradycardia and depressed respirations. A mild mydriasis, sluggish light reflexes, and mild hypotension were observed. Gastric lavage, saline infusion, and oxygen inhalation were performed, with a gradual recovery of respiration and blood pressure. She became conscious 10 hours after drug ingestion. Focal myoclonus and nystagmus appeared approximately 13 hours after the ingestion and lasted for approximately 8 hours, with spontaneous improvement. The blood concentration of ZNS was 100.1 µg/mL, that of clonazepam, 376.3ng/mL, and that of CBZ, 3.6 µg/mL at 31 hours after the drug ingestion. Considering that the elimination half-life of ZNS is 36 hours in the presence of CBZ, the peak serum concentration of ZNS might have been approximately 200 µg/mL in this patient. She eventually recovered fully without any sequelae. Based on this case report, ZNS overdose may be managed by gastric lavage and general supportive care, adequate hydration, respiratory support, and frequent monitoring of vital signs. Because of the low protein binding of ZNS (40% to 50%), renal dialysis may not be effective.