valproate-induced hepatic injury: analyses of 23 fatal cases

0270-9139/82/0205-059 l$OZ.oO/O HEPATOLOGY Copyright 0 lY82 by the American Association for the Study of Liver Diseases

Vol. 2, No. 5, pp. 591-597, 1982 Printed in U.S.A.

Valproate-Induced 23

Hepatic Injury: Analyses of Fatal Cases

HYMAN J. ZIMMERMAN AND KAMAL G. ISHAK

Department of Medicine, George Washington University, School of Medicine (D.C. Registry of Drug- Induced Liver Disease, Veterans Administration Medical Center), Washington, D.C.

20037 and the Armed Forces Institute of Pathology, Washington, D.C. 20306

Analyses of 23 fatal instances of hepatic injury in patients taking valproic acid reveals that all but three were less than 20 years old, and all but four had been taking the drugs for more than 1 month. Convulsions, facial edema, lassitude, and vomiting were prominent clinical features. Hypo- glycemia was recorded in six patients. Rash and eosinophilia were not seen. Values for transami- nases were modestly elevated in most patients. Most levels of SGOT were below 500 IU, and SGPT levels were below 200 IU. Livers showed microvesicular steatosis in most patients, usually accom- panies by necrosis. Four patients had cirrhosis. Overt valproic acid-induced hepatic injury appears to be rare and hence, by definition, idiosyncratic. That it may be an idiosyncratic exaggeration of a much more frequent phenomenon is suggested by the higher incidence of seemingly trivial injury. The idiosyncrasy appears to be metabolic rather than immunologic, and the available information leads to the plausible hypothesis that a metabolite is responsible for the microvesicular steatosis seen in most fatal cases. The steatosis resembles that of Reye’s syndrome and Jamaican vomiting sickness, and there is reason to believe that the metabolite responsible for the steatosis resembles the agent responsible for Jamaican vomiting sickness. A different metabolite is presumably responsible for the necrosis seen in many of the cases.

Valproic acid (VPA) has been in clinical use in the United States as an anticonvulsant since 1978. During that time, it has been incriminated as the cause of a number of instances of severe hepatic injury, some fatal (1-10). The hepatic lesion has been described to resemble that of Reye’s syndrome (RS) and Jamaican vomiting sickness (JVS) (4). The pathogenesis of the hepatic injury remains obscure, although it has been attributed on reasonable hypothetical grounds to the hepatotoxic effect of a metabolite of VPA (4). The present report is based on an analysis of 23 fatal instances of hepatic injury in patients taking VPA. Six of these cases have been re- ported previously (2-5, 8).

MATERIALS AND METHODS There were 26 patients known to have developed fatal

hepatic disease while taking VPA. Summaries were made available on all of them. Three of these were judged to be unrelated to the drug. One of these had been taking phenytoin as well and had a clinical syndrome typical for phenytoin hypersensitivity and distinctly different from that of the other patients. A second patient in this group

Received December 4, 1981; accepted March 31, 1982. Address reprint requests to: Hyman J. Zimmerman, M.D., Depart-

ment of Medicine, George Washington University School of Medicine, Burns Memorial Building, Room 648,2150 Pennsylvania Avenue, N.W., Washington, D.C. 20037.

had no data available other than the statement that VPA had led to the injury. The third patient was excluded because liver disease appears to have been present before the VPA was started.

Clinical summaries, autopsy protocols, and liver sections were made available through the courtesy of Abbott Laboratories (South Pasadena, Calif.) and the individual clinicians and pathologists (see “Acknowledgements”) involved in the respective cases. There were clinical summaries of varying completeness on all 23 cases and autopsy protocols containing a description of hepatic histology on 21 of them. Slides or paraffin-fixed blocks of liver tissue on 16 of the 21 cases also were made available by the respective pathologists. Sections were prepared at the Armed Forces Institute of Pathology (Washington, D.C.). Because the material had been fixed in formalin, assessment of hepatocyte glycogen was not possible.

CHARACTERISTICS OF THE PATIENTS The 23 patients ranged in age from under 1 to 28 years.

Only six of them were older than 10. All but three were less than 20 years old (Figure 1). Thirteen of the patients were male. All were in the United States and had taken the preparation known as Depakenea. Blood levels of VPA were available in 14 patients. In six of them, they were above the upper limit of the accepted therapeutic range (50 to 100 mg per ml) and in four, below the lower limit a t the time measured. The dose of the drug ranged

59 1

592 HEPATOLOGY ZIMMERMAN AND ISHAK

when the jaundice, ascites, and edema became marked or biochemical abnormalities were encountered. Spider nevi and physical signs of portal hypertension were no- tably absent; hepatomegaly was not described. Cause of death was judged to be liver failure in 20 of the patients and insufficiently described in three. It was accompanied by oliguria and azotemia in 10, by Gram-negative sepsis in four, and by pneumonia of varying extent in some.

LABORATORY FEATURES Blood leucocyte levels ranged from 2,300 to 53,000, and

<1 1 -5 6-10 11-19 20-30 neutrophil levels as high as 80% were noted. Eosinophilia was not noted. Slight anemia was frequent. Hypoglyce- AGE (YEARS)

0- FIG. I . Age distribution of 23 cases of valproate toxicity.

TABLE 1. OTHER ANTICONVULSANTS TAKEN WITH VALPROATE 7-

Anticonvulsant No. of patients (0

5 6- Phenytoin 15 Phenobarbitol 13 w

a a 5- 0

I- Car baniazepine 7 Clonazepam 6 U

Acetazolamide 5 Diazepam 2 Primidone, mesantoin, diazoxide, 1 each

6 4-

ethosuximide 3-

m 3 3

from a low of 13.1 to a high of 71 mg per kg. All but three

addition to VPA (Table 1). In most of the patients, they

2-

of the patients had been taking other anticonvulsants in

included phenytoin and phenobarbital. Only four of the patients had been taking the VPA for less than 1 month at the time of onset of the syndrome. All but two of the others had been taking the drug for periods of 1 to 4 months (Figure 2). One of the two had been taking it for va:E;t; ~ ~ ? ? ~ ~ n Of treatment prior to Onset Of disease in 23 180 days, and the other had taken the drug during two separate periods totalling 241 days.

CLINICAL MANIFESTATIONS Phase of illness Manifestations Frequency"

1-

61 1-2 2 - 4 >4 MONTHS

Of

TABLE 2. CLINICAL MANIFESTATIONS

Delineation of the incidence of the various clinical features was handicapped by the variation in form and completeness of the individual clinical summaries. Also,

merge with the underlying disorder responsible for the convulsive state that was being treated. Nevertheless, a

developed in almost all of the patients. Anorexia, nausea, Syndromal Jaundiceh 15 9

Generalized edema, including facial puffiness, occurred Fever 8

Prodromal Lethargy, lassitude, anor- All or almost all pa- exia, nausea, and vomit- tients ing

the onset of the syndrome of hepatic disease seemed to Edema and facial puffiness 12 Change in pattern of con- Frequent

vulsions clear picture emerged (Table 2). Lethargy and lassitude

and vomiting appeared to mark the onset in nine patients.

in 12 patients. Hypoglycemia 7

Eight patients had fever ranging from slight elevations to very high levels. Rash was recorded in none of the patients. Jaundice was recorded in 15 patients, seven of whom also had ascites. Two other patients had ascites

apparent, severe liver disease at autopsy. Five patients Unclear 3

Ascites'

Hemorrhagic phenomena 5

20 Terminal Coma ~hyperammonemic~~ + Azotemia 10 + Sepsis 4 + Pneumonia Frequent without jaundice and four patients had neither, despite

had hemorrhagic phenomena inciuding gastroin- * Where figures for incidence are ,.Iear, they are indicated. Where testinal bleeding, ecchymoses, and intracranial hemor- rhage- Coma was usually a terminal event but in several patients appeared near the apparent onset of the hepatic syndrome. For the most part, these features evolved slowly; and hepatic injury was usually recognized only

not clear, estimate of relative frequency is stated.

jaundice had ascites.

values.

Seven patients with jaundice had ascites. Two patients without

' All nine patients with ammonia levels recorded had increased

Vol. 2, No. 5, 1982 VALPROATE-INDUCED HEPATIC INJURY 593

mia, often severe, was recorded in eight patients in all after onset of overt hepatic disease. Albumin levels were low, ranging from 1.5 to 2.7 gm per dl in 5 of the 6 patients with recorded levels. Globulin levels were normal.

Biochemical evidences of hepatic injury (Figure 3) included moderately elevated aminotransferase levels. In most patients, the SGOT values were less than 500 IU. Six had higher values ranging from 564 IU to a level greater than 12,000 IU. Values for SGPT were available in only 13 of the patients and were lower than the respective SGOT in 10 of them. In those 10 patients, the SG0T:SGPT ratio ranged from 1.5 to 5.4 (Figure 3). In all but two patients, the SGPT values were less than 300 IU. Bilirubin levels were normal in three of the patients. In the others, they ranged from 1.3 to 36 mg per dl. Values above 10 mg per dl were found in only eight patients. When available, the direct fraction was equiv- alently increased. Prothrombin time was prolonged in almost all patients with values that ranged from 1% times the control, to those more than 5 times the control values.

Serum ammonia levels were elevated in all nine patients whose values were recorded. They were above 300 p g per dl in two patients, above 200 pg in four, and between 100 and 200 in the remaining three. Most of the patients with elevated ammonia levels also had azotemia.

HISTOLOGIC DATA Information was available in 21 of the 23 patients

(Figure 4, Tables 3 and 4). The most frequent lesion was microvesicular steatosis, which was recorded in 17 of the 21 cases (Figure 5a). In 10 of these, it was accompanied by necrosis which was centrizonal in seven (Figure 5b) and nonzonal (Figure 5c) in the other three (Table 3).

0000

4000

2000

1000 - 3 =. v) w 3 J 2 500

2 400

300

200

100

c

I . ,

. i

. j . .

Four other patients with steatosis also had cirrhosis which was micronodular in three and macronodular in one (Table 3). Three patients with steatosis had no other significant lesion. In four patients, there was no steatosis but only necrosis which was centrizonal in two and nonzonal in the other two. Liver sections were made available for study on 16 of the 21 patients. They revealed the changes listed in Tables 3 and 4. Cholestasis was prominent in 9 of the 16 and slight in the remainder.

1 2 1 UNKNOWN

STEATOSIS f

CIRRHOSIS

STE A T 0 S IS +

NECROSIS

FIG. 4. Prevalence of hepatic lesions in 21 of the 23 patients with valproate toxicity. Figures in each panel of the bar graph indicate the number of patients with the respective lesion.

36 . 30

20

10

5

1 -!

2.0 3.0 4.0 5 0 6.0 NO ALT ' ' O VALUES RATIO AST/ALT

. 2 5.0 . 0 .

4.0 . . . . . . 3 0

0 . 2.0 . . 1 .o .

$ 8

......

PROTHROMBIN TIME

FIG. 3. Peak values for SGOT (AST) and SGPT (ALT) and bilirubin and peak prolongation of prothrombin time (expressed as patient's value divided by the control value) in 23 patients with fatal valproate toxicity.

594 ZIMMERMAN

TABLE 3. MAJOR HISTOLOGICAL ABNORMALITIES IN PATIENTS" WITH

FATAL VALPROATE TOXICITY Steatosis"

Presenl Absent Total __

Necrosis Zonal 7 (5) 2 (0) 9 (5) Nonzonal 3 (3)' 2 (2)" 5 (5)

No necrosis 3 (2)' 0 3 (2) Cirrhosis' 4 (4) 0 4 (4)

Total 17 (16) 4 (2) 21 (16)

" Main figures refer to all 21 patients on whom autopsy reports were submitted by individual pathologists. Figures in parentheses refer to the 16 patients whose liver sections were made available for study by the authors.

" Microvesicular. ' Includes two cases with submassive coagulative degeneration and

"Includes one case with massive necrosis and one with diffuse

" Steatosis without necrosis. ' Micronodular in three cases and macronodular in one.

necrosis and one with severe diffuse hepatocellular cholestatic injury.

hepatitis-like hepatocellular injury.

TABLE 4. HISTOLOGICAL FEATURES ASSOCIATED WITH MAJOR

SECTIONS WERE AVAILABLE FOR STUDY H~STOLOCICAL ABNORMALITIES AMONG 16 PATIENTS WHOSE LIVER

Histological features

Major histological abnormalities

_. .

Non- Non- Zonal necrosis and stea- cronis crosis

tosis and stea- without tosis steatosis

ne- zonal ne- Steatosis Cirrhosis

No. of patients Cholestasis

Slight Moderate

Cell ballooning Slight Moderate, marked

Slight Moderate, marked

Cholangiolar proliferation Slight Moderate, marked

Portal inflammation Slight Moderate, marked

Hemosiderin in Kupf. fer cells

Hematopoietic foci Oncocytes (28,29)

Acidophilic bodies

5 3 2 2

I 0 0 0 3 2 1 1

1 0 1 0 0 1 1 0

2 0 1 1 0 1 0 0

0 0 0 2 4 2 1 0

2 0 1 1 3 2 0 0 1 1 0 1

1 1 2 1 1 0 0 0

4

1 2

0 0

2 1

1 3

1 3 1

1 2

Ballooning, degeneration, and sinusoidal acidophilic bodies were seen in the nonnecrotic areas in several of the patients (Table 4). Cholangiolar proliferation was prominent in 10 of the 16 patients. The liver of one patient showed venoocclusive changes (Figure 5d). Sections from three patients (Table 4) showed oncocytes (28,29).

DISCUSSION The observations recorded here strongly suggest that

valproate can lead to a characteristic form of hepatic

AND ISHAK HEPATOLOGY

injury. Microvesicular steatosis was found in 80% of the cases, accompanied in most instances by necrosis or cirrhosis (Figure 4). Most of the instances of necrosis were centrizonal.

The clinical syndrome usually includes jaundice, with or without ascites, progressing to coma, and terminal renal failure. Hemorrhagic phenomena are frequent. It is usually preceded by nausea, vomiting, lassitude, and generalized edema. In some instances, the frst manifes- tation of hepatic injury is overt liver failure. Hypoglyce- mia seems fairly frequent, having been recorded in ?h of patients. Indeed, it is reasonable to conclude that the lesion and syndrome of hepatic injury induced by VPA are readily distinguishable from the hepatic injury induced by phenytoin. Phenytoin leads to nonzonal necrosis, without steatosis and is usually accompanied by inflammation; the syndrome includes overt features of hypersensitivity not seen with VPA. Furthermore, the uniformity of the syndrome and the distinctiveness of the lesion, microvesicular steatosis, suggests that the large variety of other drugs employed cannot be held accountable for the hepatic injury.

It seems quite clear that the occurrence of severe hepatic disease in a small proportion of the large number of patients who have taken VPA represents idiosyncratic susceptibility rather than potent intrinsic toxicity of the drug. Furthermore, the character of the idiosyncrasy is suggestive of metabolic aberration with production of toxic metabolites rather than hypersensitivity. This may be deduced from the lack of clinical hallmarks of hypersensitivity and the histologic changes which are not those of hypersensitivity. The syndrome does not include rash or eosinophilia, and steatosis is not a lesion associated with hypersensitivity. Delay in onset of the illness until after there had been exposure to the drug for more than 1 month in over 80% of the patients also is much more consistent with metabolic idiosyncrasy than with hypersensitivity as the basis for the idiosyncratic hepatic injury (11). Indeed, the evidence seem compelling that the severe hepatic injury represents idiosyncratic susceptibility implemented by the production of hepatotoxic metabolites able to produce microvesicular steatosis and/ or zonal necrosis. In that connection, it is intriguing to note that the microvesicular steatosis seen in 80% of these patients resembles the lesion seen in JVS and in RS (4, 12, 13). Furthermore, the lesions can be repro- duced by administration of 4-pentanoic acid (Figure 6) (13, 14). It has been demonstrated that VPA is metabo- lized along four main pathways (15, 16): (a) conjugation with glucuronic acid (b) P-oxidation; (c) w-oxidation, and (d) wl-oxidation. The w pathway leads to the formation of 2-n-propylglutaric acid with 4-en-VPA and 5-hydroxy- VPA as intermediates and the w1 pathway to the formation of 4-hydroxy-VPA via 3-en-VPA (Figure 6).

The nature of the metabolic lesion responsible for the hepatic injury remains to be defined. Indeed, a large number of metabolic abnormalities have been induced by VPA in experimental animals or observed in humans taking the drug (17-21). Defects in carbohydrate metabolism have appeared to include inhibition of glycolytic or citric acid cycle activity and depletion of hepatic glycogen stores (18); impaired fat metabolism seems reflected in


FIG. 5. Section from cases of valproate toxicity showing: (a) microvesicular steatosis (H & E, X 400); (b) centrizonal necrosis and surviving hepatocytes that are large and pale due to marked steatosis (H & E, X 160); (c) massive hepatocellular necrosis (H & E, x 60), and (d) oncocytes darkly stained cells in the area on the right (H & E, x 160).

depression of plasma hydroxybutyrate and ketone levels (18); distorted amino acid metabolism is reflected in the hyperglycinemia and glycinuria ( 19) and proprionic acid acidemia (ZO), and impaired urea cycle activity has been

assumed to be responsible for the hyperammonemia (17, 21) seen in some patients. Indeed, the observation of elevated ammonia levels even without concomitant overt hepatic disease in patients taking VPA (17, 21), which

HEPATOLOGY 596 ZIMMERMAN AND ISHAK

CH, / \

A. CHp=C-CH-CHt-CH-COOH HYPOGLYCIN A I I .) NH,

C H, / \

CH,=C-CH-CH.-COOH TOXIC METABOLITE

0 . CH,=C H-C HI-CH 2-COOH

I CaH, I I 4-EN-VPA I I

L - - - - - - - 7--------1

CHs-CH,-CHZ-CH-COOH I I OH CaH,

5-HYDROXY-VPA

4 C-CH2-CHl-CH-COOH 0 I

CsHr 0 H

1 I CHa-CH=CH-CH-COOH I

I I

- - - - - - - - - - - - - - - I I I CaH, I I 3-EN-VPA I

3-- - - - - -J

i

L - - - - - - - CH,-C H-CHz- CH -COOH

I I OH CaH,

4-HY DROXY-VPA

CH,-C-CHI-CH-COOH II I 0 CsH,

2-n-PROPYLGLUTARIC ACID

FIG. 6. Structure of (A) hypoglycin A and its presumed toxic metabolic product, (B) 4-pentanoic acid, and (C) valproate and the metabolic products formed by metabolism along the w and wz pathways. Note structural similarity of metabolite of valproate (4-en VPA) to the toxic product of hypoglycin and 4-pentanoic acid.

has been attributed to suppression of urea cycle activity by the drug, provides another link to the lesion induced by 4-pentanoic acid and that described in RS (13, 14). Induction of overt hepatic injury and hyperammonemia seem separable phenomena (17). Nevertheless, the two may occur together and are probably linked, although separable, biochemical defects.

The similarity of the hepatic injury (microvesicular steatosis) and some metabolic lesions of VPA toxicity to those of JVS (Table 5 ) and the similarity between the chemical structures of hypoglycin, 4-pentanoic acid and the metabolic products of VPA (Figure 6) offer strong theoretical support for the view that the hepatic injury in the patients taking VPA is caused by toxic metabolites. The amount of such toxic metabolites formed, presumably, is negligible in most patients taking VPA, otherwise, the incidence of abnormality would be higher. Neverthe- less, studies (22-26) of patients taking VPA demonstrate a higher incidence of abnormality, albeit minor, than that indicated by the low incidence of severe injury (Table 6). This is reminiscent of the experience with isoniazid which leads to a 10 to 20% incidence of minor injury, but to an approximately 1% incidence of severe injury (11). One may speculate that the patients who develop minor abnormalities while taking isoniazid or VPA and then appear to become tolerant to the adverse effect, develop metabolic pathways that lead to decreased production of toxic intermediates of the drug or improved means of dealing with them. According to this hypothesis, those

who develop severe disease are unable to make the adjustments.

Young individuals appear to be disproportionately vul- nerable to hepatic injury from VPA. Over 80% of the patients with severe VPA toxicity are less than 20 years old (Figure 1); yet only 40% of the users of VPA are in this group, (J. Page, M.D., Abbott Labs., personal com- munication). Accordingly, it is intriguing that the syndromes resembled by VPA toxicity, namely, JVS and RS, also tend to affect very young individuals (Table 5). The basis for an enhanced susceptibility of young individuals to this form of injury is not clear.

An additional factor that may have a bearing on susceptibility is the administration of other drugs with VPA. Since the number of patients not taking other drugs is small (three patients) and the combinations taken are extremely varied, elucidation of the effects of concurrent drugs is not possible a t this time. Nevertheless, the possibility that other drugs enhance the toxicity of VPA remains a possibility. The putative toxic product of VPA is produced by w oxidation (Figure 6) which is affected by the cytochrome P-450 system of the smooth endo- plasmic reticulum (27). Enzymes of this system can be enhanced by administration of enzyme inducers such as phenytoin or phenobarbital (16). Accordingly, there is a hypothetical basis for the suspicion that concomitant use of other anticonvulsants can contribute to VPA toxicity by enzyme induction.

With regard to the actual pathogenesis of the lesions, presumably the 4-pentanoic acid-like metabolites of VPA

TABLE 5. COMPARISON OF CHIEF SYNDROMAL, HISTOLOGICAL, AND

BIOCHEMICAL FEATURES OF JVS, RS, VPA, AND EXPERIMENTAL EFFECTS OF 4-PENTANOIC ACID

Features J V S RS VPA 4-Pentanoic acid

Syndromal Encephalopathy Hypoglycemia Death

Histological Microvesicular

fat Necrosis Mitochondria1 in-

jury Biochemical

SGOT, SGPT t Bilirubin t NH:, T

+ + + + * +

2-3+ I+ +

+ + + + i- +

2-3+ 1-4+

+

TABLE 6. INCIDENCE OF MINOR BIOCHEMICAL EVIDENCE OF HEPATIC INJURY

Reference Test No. of

ties abnormali- Incidence

( % I )

Willmore et al. (22) t SGOT and 4/25 16 SGPT

zyme” Vining et al. (23) t “Liver en- 18/27 67

Sherard et al. (24) t SGOT 7/100 7 Sussman and McLain SGOT and 4/9 44

(25) SGPT - -1 Fibrinogen 9/9

Coulter et al. (26) T SGOT 44/100 44

“ Abnormalities subsided with reduction of dose.


(e.g., 4-en-VPA) lead to microvesicular steatosis by: (a) interfering with the exit of lipid from the liver through depressing of synthesis of the apoprotein of very low density lipoproteins; (b) depressing oxidation of fatty acids through mitochondrial injury, or (c) enhancing mobilization to the liver and other organs of fatty acids from the fat depots. The mechanism for the production of the centrizonal hepatic necrosis is presumably the formation of reactive metabolites in the central zone (1 1). One may assume that the metabolite(s) responsible for centrizonal necrosis differs from that (those) responsible for the microvesicular steatosis, since centrizonal necrosis is not seen in the other lesions associated with microvesicular steatosis (JVS and RS). It is also possible that the centrizonal necrosis is caused by metabolites of other drugs that had been taken with VPA or that the necrosis might, in some instances, have been the product of terminal shock and hepatic anoxia.

The long delay in the onset of the rarely observed overt injury and the high incidence of dose-related minor abnormality suggest guidelines for use of the effective anticonvulsant. The long delay permits assessment of therapeutic benefit while monitoring for enzymologic evidence of hepatic injury. Patients in whom the anticonvulsant effect is adequate may be continued on the drug. Those who develop abnormal levels of SGPT and SGOT might have the dose of drug lowered until the values are normal. Most instances of minor elevation of aminotrans- ferases in recipients of this drug appear not to be harbin- gers of severe hepatic injury; and it is possible that the severe injury might develop abruptly without forewarn- ing provided by trivial hepatic abnormalities. Neverthe- less, the approach suggested would appear to be a reasonable one for the use of this drug until the issue has been clarified.

It is also tempting to speculate that VPA-induced hepatic injury might be heralded by evidence of metabolic defects of the type known to be produced by 4- pentanoic acid and hypoglycin A (12, 13), since injury and presumed chemical causation are similar. These agents lead to the excretion of isovaleric and other amino acids as well as other organic acids. Monitoring for these substances might provide clues to impending significant injury, and the effort seems warranted.

ADDENDUM A brief summary of the Abbot Laboratory data in

these 23 cases was presented at the Child Neurology Society and published in abstract form. [Rathner AD. Valproic acid, a review of 23 fatal cases. Ann Neuroll981; 10:287 (A).]

Acknowledgments: We are grateful to the following physicians who participated in the care or postmortem study of the patients in this study and who cooperated generously in making information and liver sections available for study.

Drs. F. J. Allen, J. Antonius, D. Bakker, S. Bates, C. Bauserman, R. Brumund, B. A. Burke, R. Brent Clark, R. K. Cochran, W. Corrie, D. Coulter, D. J. Dove, W. E. Dodson, J. F. Donat, N. Futch, D. L. Gang, E. Gonzalez, C . Green, T. Green, R. Hardman, J. H. Harris, S. Harris, K. P. Heidelberger, R. T. Hoss, J. Ison, H. Krous, T. Kunczynski, D. Linder, S. Linder, L. A. Lochman, E. Marquardt, W. May, J. Melson, F. J. Monrael, P. A.

OShea, M. Petrelli, D. L. Porter, G. Prentice, F. 0. Raasch, E. P. Richardson, J. R. Schimschock, R. E. Schmidt, R. E. Shenenfelt, P. Singla, G. Vandenvorker, F. H. Watson, C. West, and R. S. K. Young.

We are also grateful to Abbott Labs. and particularly to Dr. J. Page for assistance.

We thank M. Varoff for secretarial assistance.

REFERENCES

1. Browne TR. Valproic acid. N Engl J Med 1980 302561. 2. Bowdle TA, Patel IH, Wilensky AJ, et al. Hepatic failure from

valproic acid. N Engl J Med 1979; 301:435. 3. Donat JF, Bocchini JA, Gonzalez E, et al. Valproic acid and fatal

hepatitis. Neurology 1979; 29273. 4. Gerber N, Dickinson RG, Harland RD, et al. Reye-like syndrome

associated with valproic acid therapy. J Pediatr 1979 95:142. 5. Suchy FJ, Balistreri WF, Buchino JJ , et al. Acute hepatic failure

associated with the use of sodium valproate: report of two fatal cases. N Engl J Med 1979; 300:962.

6. Ware SE, Milward-Sadler GH. Acute liver disease associated with sodium valproate. Lancet 1980 2: 110.

7. Jacobi G, Thorbeck R, Ritz A, et al. Fatal hepatotoxicity in child on phenobarbitone and sodium valproate. Lancet 1980; 1:712.

8. Young RSK, Bergman I, Gang DI, et al. Fatal Reye-like syndromes associated with valproic acid. Ann Neurol 1980; 7:389.

9. Le Bihan G, Bourreille J, Sampson M, et al. Fatal hepatic failure and sodium valproate. Lancet 1980; 2:1298.

10. Trim JH. Harereaves T. Anthonv PP. et al. Sodium valoroate and

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

.. I Y

ornithine carbarnyl transferase deficiency. Lancet 1981, 1: 1165. Zirnmerman HJ. Hepatotoxicity: the adverse effects of drugs and other chemicals on the liver. New York: Appleton-Century-Crofts, 1978. Tanaka K, Kean EA, Johnson B. Jamaican vomiting sickness: biochemical investigation of two cases. N Engl J Med 1976; 295:461. Tanaka K. Jamaican vomiting sickness. In: Vinken PJ, Bruyn GW, eds. Handbook of clinical neurology, Chap. 17, Vol 37. Amsterdam: Elsiever/North Holland, 197951 1-539. Glasgow AM, Chase PH. Production of the features of Reye's syndrome in rats with 4-pentanoic acid. Pediatr Res 1975; 9:133. Loscher W. Concentration of metabolites of valproic acid in plasma of epileptic patients. Epilepsia 1981; 22:169. Schafer H, Luhrs R. Metabolic pattern of valproic acid. I. Gas chromatographic determination of the valproic acid metabolite artifacts, hepatanone-3, 4-, and 5-bydroxyvalproic acid lactone. Arzneim Forsch 1978; 28657. Coulter DL, Allen RJ. Secondary hyperammonaemia: a possible mechanism for valproate encephalopathy. Lancet 1980 1:1310. Thurston JH, Hauhart RE, Schulz DW, et al. Chronic valproate administration produces hepatic dysfunction and may delay brain maturation in infant mice. Neurology 1981; 31:1063. Jaeken J , Corbell L, Casaer P, et al. Dipropylacetate (valproate) and glycine metabolism. Lancet 1977; 2517. Schmid RD. Propionic acid and dipropylacetic acid in the urine of patients treated with dipropylacetic acid. Clin Chim Acta 1977; 74:39. Rawat S, Borkowski WJ Jr, Swick HM. Valproic acid and secondary hyperammonemia. Neurology 1981; 31:1173. Willmore LJ, Wilder BJ, Bruni J , et al. Effect of valproic acid on hepatic function. Neurology 1978; 28:961. Vining EPG, Botsford E, Freeman JM. Valproate sodium in refrac- tory seizures. Am J Dis Child 1979; 133:274. Sherard ES, Steinman GS, Couri D, et al. Treatment of childhood epilepsy with valproic acid. Neurology 1980; 3031. Sussman NM, McLain LW. A direct hepatotoxic effect of valproic acid. JAMA 1979; 242:1173. Coulter DL, Wu H, Allen RJ. Valproic acid therapy in childhood epilepsy. JAMA 1980; 244:785. Patel IH, Levy RH, Cutler RE. Phenobarbital-valproic acid inter- action. Clin Pharmacol Ther 1980; 27:515. Gerber MA, Thung SN. Hepatic oncocytes: incidence, staining characteristics, and ultrastructural features. Am J Clin Patholl981; 79:498. Lefkowitch JH, Arborgh BAM, Scheuer PJ. Oxyphilic granular hepatocytes: mitochondrion-rich liver cells in hepatic disease. Am J Clin Pathol 1980; 74:432.

valproate-induced hepatic injury: analyses of 23 fatal cases

Documents