Journal of Hepatology, 1993; 17:227-235 227 ~ 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0168-8278/93/$06.00

HEPAT 01371

The results of a randomized double blind controlled trial evaluating malotilate in primary biliary cirrhosis

A European Multicentre Study Group 1 (Received 14 January 1992)

One hundred and one patients were included in a double-blind controlled trial to determine whether malotilate (diisopropyl 1,3-dithiol-2-ylidene malonate) is therapeutically effective in primary biliary cirrhosis. Fifty-two patients received malotilate (500 mg three times a day) and 49 patients placebo. The mean follow-up time was 28 months

(range 6-46 months). The large majority of patients did not have advanced liver disease since only ten patients were in Child-Pugh class B and none in class C, and the median bilirubin and albumin at entry were normal. Malotilate had no clear effect on pruritus. In malotilate recipients the following statistically significant biochemical changes occurred: alkaline phosphatase decreased 21%, AST 20%, ALT 40%, IgA 12% and IgM 26%. In the placebo group no significant changes occurred. Evaluation of entry and 2-year liver biopsies indicated that malotilate diminished plasma cell and lymphocytic infiltrate and piece-meal necrosis, but had no effect on liver fibrosis. There was no difference in survival or in disease progression according to Child-Pugh criteria. In six patients receiving malotilate, but in none

on placebo, treatment was discontinued due to suspected side effects. All patients recovered completely. We conclude that malotilate has an immune-modulating, anti-inflammatory but not anti-fibrotic effect in primary biliary cirrhosis.

The clinical relevance of the observed benefits, however, appears too.slight to recommend malotilate as single drug therapy in primary biliary cirrhosis.

K e y words: Primary biliary cirrhosis; Malotilate; Double-blind randomized controlled trial; Medical treatment

Primary biliary cirrhosis (PBC) is a chronic, slowly progressive cholestatic liver disease of unknown cause. The disease has many features indicating an underlying

autoimmune disorder, and is histologically characterized by ongoing non-suppurative bile duct destruction, leading to cholestasis, secondary copper overload, progressive fibrosis and eventually cirrhosis (1).

To date no drug has been shown to induce complete remission of the disease process (2,3), and the only treatment that prolongs survival for patients who pro- gress to end-stage disease is liver transplantation (4).

At least eighteen controlled trials have been published to evaluate immunosuppressive, cupruretic and antifi- brotic therapies (5-23). Trials using d-penicillamine (5-12) and azathioprine (13,14) have failed to show significant benefits (3,24), although re-analysis of one of these studies showed that azathioprine was associated with slightly improved survival (14). Promising results have been reported using chlorambucil (18), cyclosporin (19,20), prednisolone (21) and methotrexate (22,23), but no studies have yet shown these drugs to be safe and effective in long-term use. Three recent reports indicate

Correspondence to: H.R. van Buuren, M.D., Department of Internal Medicine II, Room Ca 326, Dijkzigt University Hospital, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands. J The Jbllowing investigators (all M.D. except when indicated otherwise) and institutions (with number of patients randomized) participated in the Stt~dy Group: United Kingdom: Harriet C. Mitchison, David J. Mutimer, Oliver F.W. James, Freeman Hospital, Neweastle-upon-Tyne (21); David R. Triger, Royal Hallamshire Hospital, Sheffield (21). Germany: Bernd M611er, Uwe Hopf, Universit/itsklinikum Rudolf Virchow, Standort Charlottenburg, Berlin (12); Eckhardt G. Hahn, Universit/itsklinikum Steglitz, Berlin (8); Richard Raedsch, Karl Gmelin, Burkhard Kommerell, Medizinische Klinik, Universit/it Heidelberg, Heidelberg (7); Heinrich Liehr, Kliniken der Stadt Saarbrfieken, Saarbrficken (2). The Netherlands: Solko W. Schalm, Henk R. van Buuren, Academisch Ziekenhuis Dijkzigt, Rotterdam (12); Karl-H. Brandt, Ziekenhuis Rijnstate, Arnhem, Gerard P. van Berge Henegouwen, Academisch Ziekenhuis, Utrecht (6). France: Christian Tr6po, H6tel-Dieu, Lyon (4). Belgium: Michael Adler, H6pital Erasme, Brussels (5); Andr6 P. Geubel, Universit6 St. Luc, Brussels (I); Johan Fevery, UZ Gasthuisberg, Leuven (l). Study coordinators: Solko W. Schalm, Henk R. van Buuren. Writing Committee: Henk R. van Buuren, Solko W. Sehalm, David R. Triger, Oliver F.W. James. Histological analysis: Fiebo J.W. ten Kate, Academisch Ziekenhuis Dijkzigt, Rotterdam. Statistical analysis: Henk R. van Buuren, Wire C.J. Hop, ir., Erasmus University, Rotterdam.

228 A EUROPEAN MULTICENTRE STUDY GROUP

that colchicine may have modest benefits (15-17), but other studies must be performed before this drug can be considered standard therapy (25). Favorable results have recently been reported on the use of ursodeoxycholic acid (26,27), but there are limited data on liver histology and no long-term data on survival.

Malotilate (diisopropyl 1,3-dithiol-2-ylidene maionate) is a drug discovered in Japan (28). In animal studies malotilate has anti-necrotic and anti-steatotic effects in models of experimental liver injury using carbon tetra- chloride (29-31), paracetamol (32), galactosamine (33), orotic acid (34), egg yolk sensitization (31), and other toxins (28,35). In several studies malotilate enhanced protein synthesis (31,36) and hepatocyte regeneration (37,38), and decreased liver collagen synthesis (29,31). Previous controlled observations in a heterogenous group of 681 patients with chronic hepatitis and cirrhosis indicated that malotilate treatment decreased transami- nases and increased albumin, cholinesterase and choles- terol, and was free from major side-effects (36).

Due to the results of these preliminary studies and the lack of effective medical therapy for PBC, a multi- centre double-blind controlled trial was initiated comparing malotilate treatment with placebo.

Patients and Methods

Patient selection The criteria for entry were that the patient was

symptomatic, not older than 65 years and had not received corticosteroids, azathioprine or any other potentially active drug in the 6 months preceding the study. A patient was qualified symptomatic when one or more of the following symptoms or signs were present:

pruritus, jaundice, fatigue, arthralgia, abdominal pains

or previous esophageal variceal bleeding.

A diagnosis of PBC was accepted if there was no evidence of extrahepatic biliary obstruction, serum alka-

line phosphatase levels were at least twice the upper normal limit, serum mitochondrial antibodies were pre-

sent and liver histology was diagnostic of or compatible with PBC (39). Patients with renal insufficiency (serum

creatinine >120~mol/1) and very advanced disease or

evidence of recent marked deterioration (e.g., massive ascites, encephalopathy, more than 5 kg weight loss in

6 months, bilirubin > 300/~mol/1) were excluded. At entry a radiological or endoscopic assessment for the presence

of esophageal varices was required.

Randomization and informed consent The participating centers all had one or more series

of four consecutively numbered boxes with trial medica- tion. Within each series of four, and in a random way, two contained malotilate and two placebo tablets. A new patient entering the study was assigned to the next randomization number and thus therapy. All patients were given oral and written information on the nature and goals of the study. Patients entering the study gave informed consent. The study was approved by the local ethical committees in each of the participating hospitals.

Treatment Patients were treated with 500 mg malotilate or pla-

cebo tablets of identical appearance three times a day. Both patients and doctors were unaware of the nature of the tablets. In case of suspected side-effects, the dose was lowered or stopped, after which attempts were made to reinstitute the full dose, without breaking the code. Patients were allowed to continue with support medica- tion such as vitamins, calcium and cholestyramine.

Criteria for stopping treatment included progression of the disease, defined as a rise in ALT to more than twice the baseline value or to 10 times the upper limit of normal, or a rise in bilirubin of 30 /~moi or more above the baseline value.

The study started on I September 1983. Patients were recruited up to 1 April 1986. The study and follow-up ended on 31 December 1987, after which the trial medication was discontinued and the randomization code broken.

Follow-up, biochemical and histologic testing During the first 6 weeks, patients were seen at 2-week

intervals for symptomatic assessment as well as hemato- logical and biochemical testing, mainly to detect possible adverse effects of treatment. Thereafter, clinical and laboratory evaluations were conducted every 3 months. This included medical history, physical examination, complete blood count, urinalysis, creatinine, total pro- tein, albumin, cholesterol, bilirubin, alkaline phospha- tase, alanine and aspartate aminotransferases (ALT and AST), and immunoglobulins IgA, IgG and IgM.

Since 13 centers contributed to the trial, using non- uniform laboratory methods, biochemical units and nor- mal ranges, the biochemical variables of the parameters that were studied in detail were converted to reference values. The method used was to divide the actual test result of a variable by the upper or lower limit of the normal range for that particular center, and to multiply the quotient by the normal upper or lower limit of the reference center, the University Hospital Rotterdam. For

MALOTILATE IN PRIMARY BILIARY CIRRHOSIS 229

variables with a critical upper limit, e.g., bilirubin, AST and IgM, the upper limit of normal was used for the conversion. For variables with a critical lower limit, e.g., albumin, the lower limits were used.

A liver biopsy (required size 2 cm) was performed not more than 6 months before entry and was repeated after 2 years of treatment. Some patients had additional biopsies at 1 or 3 years. The biopsy specimens were reviewed by a single histopathologist according to accepted criteria (39) and without knowledge of the clinical condition or the treatment. The following fea- tures were assessed: stage of disease and grade of fibrosis on a scale of 1 to 4, plasma cell and lymphocytic infiltrate and piece-meal necrosis as 0, 1 or 2, cholestasis (bile thrombi) 0 (absent) or 1 (present).

Statistical analyses When the trial was designed no data were available

to predict expected responses thus precluding meaningful calculations of the required number of patients. How- ever, it was considered necessary to include at least 80 patients and to start analysis only when the mean follow- up time was at least 2 years. Although this period was too short to expect differences in survival, it was consid- ered sufficient to establish the effects of malotilate on liver blood tests and histology.

All randomized patients were analyzed on an intent- ion-to-treat basis, irrespective of any event during follow- up or protocol deviation. No interim analyses were performed. Clinical and biochemical characteristics at entry were compared using the Wilcoxon rank sum test for continuous variables. Fisher's exact test (for percent- ages) and the Z2-test were used to compare other discrete variables in the two groups.

The biochemical changes during follow-up were analyzed and compared using two statistical methods:

(i) Biochemical variables in the two patient groups both at entry and at 6, 12 and 18 months were compared using the Wilcoxon two-sample rank sum test. The values within the treatment groups at entry and at 6, 12 and 18 months were compared using the Wileoxon signed-rank test. Non-parametric tests were used since the biochemical values did not conform to a normal distribution. The unpredictable influence of patients with incomplete data or follow-up was eliminated by restricting this analysis to patients followed and treated for at least 18 months, and for whom complete values until 18 months were available. After 18 months the number of patients meeting these criteria dropped rela- tively sharply and, therefore, values obtained at later times were not used for this part of the analysis.

(ii) In order to include incomplete data and follow-

up (e.g., due to interruption of treatment or death) the biochemical tests were analyzed using the Kaplan-Meier actuarial method. For this purpose two events were defined for each biochemical parameter: when at follow- up time 6, 12, 18, 24 or 30 months a decrease or increase of the entry value of a certain percentage was established, an event was considered to have occurred. The percent- ages were arbitrarily chosen based on clinical experience. The occurrence of these events (percentage of decrease or increase of biochemical value) was analyzed by the life-table method, using the log-rank test for the compari- son of groups. Biochemical values subsequent to an event were not used for this analysis. The percentage changes defined were: 50% for bilirubin, alkaline phos- phatase 25%, AST and ALT 50%, albumin 12.5%, choles- terol 25%, IgA and IgG 25% and IgM 50%. Patient survival was calculated using the Kaplan-Meier analysis. An analysis of disease progression was performed using the same method. For this purpose, the disease was considered progressive when the total score for the Child-Pugh classification (40) increased by one or more points.

To analyze the liver biopsies the scores obtained in both groups for the stage of PBC, fibrosis, piece-meal necrosis, lymphocytic and plasma cell infiltrates and cholestasis were compared using the Wilcoxon rank sum test. The number of patients in each group with histological progression or a decrease in piece-meal necrosis, lymphocytic and plasma cell infiltration was compared using the ;(Z-test. Analysis was only performed on entry and 2-year biopsies.

Statistical significance was defined by p values (two- sided) <0.05.

Results

Clinical characteristics randomization and follow-up One hundred and four patients entered the study. One

patient on malotilate and two on placebo had to be excluded from analysis since elementary data were not available. Of the remaining 101 patients, 52 were ran- domized to malotilate treatment and 49 to placebo. Patients were well matched for clinical and laboratory characteristics (Table 1). None of the characteristics at entry was significantly different. The mean age of all participants was 54_+9 years. The mean follow-up time was 28 _+ 8 months (median 25 months, range 6-46). The mean duration of treatment in the malotilate and placebo group was 23 and 28 months, respectively.

Four patients, all in the placebo group, dropped out. Two failed to return to the hospital, one was unwilling

230

TABLE 1

Clinical and laboratory characteristics at entry into the study

Normal range Malotilate Placebo (n = 52) (n = 49)

Sex (female) Median age (year) Median duration of

disease (years) Pruritus Fatigue Arthralgia Ascites Splenomegaly Child-Pugh class

A B

Esophageal varices present no varices not investigated/unknown

Liver histology stage: I II III IV

Bilirubin (/lmol/l) Alkaline phosphatase (U/I) ALT (U/I) Albumin (g/l) Cholesterol (mmol/l} IgG (g/I) lgM (g/l)

<16 25-75

<30 38-48 3.3-7.2 8-18 0.6-2.8

50 47 54.5 54.3

1.8 2.6 24 26 36 34 22 20 5 2 9 5

45 46 7 3

21% 22% 54% 45% 25w~ 33%

29% 22% 26% 30% 16% 30% 29% 18% 10 11

343 333 75 57 43 44 8.1 7.6

18.1 17.9 5.2 5.2

Biochemical variables expressed as median values. No significant differences for any of the characteristics between the groups.

to continue with the drug and for one patient the reason

was unknown.

Effect on pruritus At entry, 46% of the patients in the malotilate group

and 53% of the patients in the placebo group had

pruritus. After 1 year the percentages were 30% and

40%, and after two years 47% and 55%, respectively

(NS).

Biochemical analysis Table 2 shows the effect of treatment on a number of

biochemical parameters of liver disease. Bilirubin showed

no change, even when patients with abnormal entry

levels were seperately evaluated. Alkaline phosphatase

decreased by 21% in the malotilate group, and the values

at 6, 12 and 18 mo were significantly different from entry

values (p<0.0001). In the placebo group no significant

changes occurred. The effect on alkaline phosphatase

was confirmed by actuarial analysis. In the malotilate

group AST and ALT at 6, 12, and 18 months decreased

significantly (p < 0.005). As illustrated in Fig. 1, actuarial

analyses indicated that significantly more patients taking

malotilate had a response as defined by a decrease of

A EUROPEAN MULT1CENTRE STUDY GROUP

O' - -

I placebo

p : 0 . 002

50 malotilate . . . . . . placebo

t ................. I

~ . . . . a I

25 I r . . . . . . . . . ~ P = 0 . 0 2

. . . . J

malotilate

0 I . . . .

12 2q 36 months

Fig. 1. Actuarial percentage of patients with primary biliary cirrhosis with either a 50% decrease (solid lines) or increase (dotted lines) of

serum ALT, during treatment with malotilate or placebo.

50% or more in ALT (p = 0.002). Moreover, significantly

fewer patients on malotilate had a 50% ALT increase

(p=0.02). Although the entry values for transaminases

were higher in the malotilate group, these differences

did not reach statistical significance. In order to investi-

gate whether malotilate only affected patients with high

transaminases and to exclude the possibility of a decrease

by chance, e.g., due to high transaminases at entry with

a 'spontaneous' decrease, the statistical analyses were

repeated excluding all patients with initial AST and ALT

>100 U/I. The results of these subgroup analyses were

comparable to the results for the whole population.

Serum albumin did not significantly change in either

treatment group. Malotilate treatment resulted in a

significant rise in serum cholesterol, as shown by both

methods of analysis, with a median increase at 18 months

of 1.5 mmol/l. Patients treated with malotilate showed

a slight but significant decrease in IgA and a marked

decrease in IgM. At entry only four patients in the

malotilate group and two in the placebo group had a

prolonged prothrombin time, which was repeatedly nor-

mal in two during follow-up. In the remaining 95 patients

no consistent prolongation outside the normal range

was observed during follow-up.

Histology Histological assessment was possible for 61 biopsies

from the malotilate group and 66 from the placebo

group (64 entry and 63 2-year biopsies). At entry, the

histologic stage of PBC, the degree of fibrosis, the

intensity of lymphocytic and plasma cell infiltrates and

the severity of piece-meal necrosis and cholestasis were

comparable. After 2 years, the average scores for the

stage of PBC and the amount of fibrosis increased to

the same extent in both groups. The average scores for

lymphocytic and plasma cell infiltrate and piece-meal

necrosis decreased in the malotilate but not in the

MALOTILATE IN PRIMARY BILIARY CIRRHOSIS

TABLE 2

Biochemical variables at entry, 6 months and 18 months"

231

Variable (normal value) n Entry 6 months 18 months Change 0-18 months

pb median c % pd

Bilirubin (< 16/tmol/l) Malotilate 49 I0 (5-76) 10 (4-31) 9 (4-50) NS - 2 - 2 0 Placebo 47 11 (3-39) 10 {5-67) 8 (3.5-23.5) NS 0 0

Alkaline phosphatase (< 75 U/I) Malotilate 43 364 (121-566) 303 (121-566) 262 (121-545) 0.0001 - 7 6 - 2 1 Placebo 43 333 (90-697) 303 (131-727) 282 (121-6571 NS - 1 0 - 3

AST (<30 U/I) Malotilate 43 64 (28-135) 45 121-I 14) 50 (20-I 16) 0.003 - 1 3 - 2 0 Placebo 43 54 (31-84) 52 (31-84) 53 (30-93) NS 0 0

ALl" (<30 U/I) Malotllate 34 75 (29-196) 54 (20-I 15) 49 (21-80) 0.0001 - 3 0 - 4 0 Placebo 32 57 (34-110) 59 (32-991 61 (32-115) NS - 3 - 5

Albumin (38-48 g/I) Malotllate 43 43 (35-49) 44 (37-50) 43 (36-511 NS + l +2 Placebo 41 44 (37-49) 44 (37-49) 43 (37-49) NS 0 0

Cholesterol (3.3-7.2 mmol/l) Malotilate 39 8.2 (5.6-12.4) 10 (7.6-12.4) ' 10.1 (7.4-13.8) ~ 0.001 +1.5 +18 Placebo 38 7.6 (5.8-1 I) 7.4 (6.1-10.71 7.9 (6-10.6) NS 0 0

IgA (0.9-4.5 g/I) Malotilate 41 3.4 (I.7-5.91 3 (1.1-6.3) 2.7 (1.2-5.6) 0.004 - 0 . 4 - 1 2 Placebo 42 3.1 (I.7-5.8) 3.2 (1.4-5.6) 2.7 (1.4-5.8) NS 0 0

lgG (8-18 g/I) Malotilate 40 17.5 111.6-27) 16.6 (10.9-21.9) 17.6 (10.6-26.7) NS -0.1 0 Placebo 40 17.6 (11.9-28.7) 17 (I0.5-26.4) 18 (10.8-27.7) NS +0.4 +2

lgM (0.6-2.8 g/I) Malotllate 41 5.4 (2.5-12.9) 3.9 (I.I-7.6) t 4.1 (I.3-8.7) 0.0001 - I . 4 --26 Placebo 42 5.2 (2.7-12) 6.1 (2.5-14.4) 5 (2.2-13) NS 0 0

0.31

0.005

0.001

0.0005

0.47

0.0005

0.01

0.92

0.0004

~Analysis is based on patients who received malotilate or placebo treatment for at least 18 months and for whom data until at least were available. Data are expressed as medians with 10th and 90th percentiles.

bSigned rank test of values at entry and 18 months. CMedian of differences of values at 18 months and at entry. dComparison of median decrease in malotilate and placebo group using rank sum test. ~Comparison of both treatment groups using two-sample rank sum test; p<0.0005. fComparison of both treatment groups using two-sample rank sum test; p =0.005.

18 months

placebo group. The difference at 2 years was significant (p<0.05, Wilcoxon rank-sum test), as shown in Fig. 2. Only 44 paired (entry vs. 2 year) biopsies (20 malotilate group) were available. The main reason for this low number was the failure of two major centers to forward biopsies to the evaluation center. Other reasons for the failure of paired evaluation were: inadequate biopsy size (8) initial biopsy taken more than 6 months before start of the trial (7), and lack of a second year biopsy in patients who dropped out due to adverse effects, death or refusal. In entry biopsies no differences were found and at 2 years average scores for cellular infiltrate and piece-meal necrosis decreased in the malotilate group. For plasma cell infiltrate the difference was significant (p--0.03), while the differences in lymphocytic infiltrate and piece-meal necrosis approached significance (p= 0.06 and 0.07, respectively). The number of patients in each treatment group with histological progression,

according to the grade of fibrosis and stage of disease, was not significantly different.

Progression of disease and sm'vival Only one patient in each group discontinued treatment

due to progression of the disease. One of these patients subsequently received a liver transplant. Disease pro- gression, defined as an increase of the total Child-Pugh score with one or more points during follow-up (see above), was the same in both treatment groups (Fig. 3a).

Of the nine patients who died, six were in the maloti- late group. In each group two patients died from liver insufficiency, two patients (receiving malotilate) died from variceal bleeding, one patient from each group died from myocardial infarction and one patient on malotilate died from hepatocellular carcinoma. Actuarial analysis indicated no difference in survival between the groups: I-year survival in the malotilate group was 96% and in

232 A EUROPEAN MULT1CENTRE STUDY G R O U P

IV stage 4 -

'1 0 ! !

o 2

2 piece-meal ne(:~eds 2

flbeo~5

maloMale I .

°"t 02

0 ! ! o 2

year

pUerna c ~ v~ma~e 2

cl.lo~slam

! 2

~ a r

Mnl~ocyUc Inemam

1.5

05

15

control !

malolilate, 0.5

1 5

conlrol 1

melo~ale

coralol 0 13

malo~iale

1 I I i 0 I ! 0 2 0 2 0 2

year year year

Fig. 2. Average degree of stage of PBC, fibrosis, cholestasis, piece-meal necrosis, plasma cell and lymphocytic infiltrate. Analysis based on 64 entry and 63 2-year liver biopsies.

the placebo group 98%, 2-year survival was 92% and 94%, respectively (Fig. 3b). Seven out of 10 Child-Pugh B patients died, compared with only 2 of 91 Class A patients, who both died from myocardial infarction (log- rank test, p < 0.001 ).

Adverse effects Malotilate treatment was associated with more sus-

pected adverse effects (Table 3). Treatment had to be discontinued in six patients receiving maiotilate, five of

these patients coming from one center (Rotterdam). The most severe suspected adverse effect occurred in a 38- year-old woman who showed a gradual increase in disease activity after 9 months and above normal biliru- bin after 16 months. After 20 months severe hepatitis developed with AST reaching a maximum of 745 U/l, bilirubin 393/amoi/i and a normotest falling to 21%. Tests for viral hepatitis (A, B, CMV and EBV) were negative. Liver biopsy showed liver cell necrosis and focal collapse. Besides toxic liver damage an exacerba-

TABLE 3

Suspected adverse effects

Malotilate Placebo No. No.

Continuation of drug

No. Same dose Lower dose

Toxic hepatitis 1 Neuropathy 1 ° Allergic reaction 2 Nausea 4 Headache 1 Diarrhea 2 Bad taste/joint pain 1 Alopecia 1

~Patient also had a suspected allergic reaction, however, symptoms did not recur after reinstitution of treatment.

MALOTILATE IN PRIMARY BILIARY CIRRHOSIS 233

1 0 0

7 5

5 0

2 5

- - " 1 . . . . . ~ . . . . . placebo

I ~ malotllate n o S o

i i 1 2 2 4 months

1 0 0

7 5

5 0

2 5

t - ' l . . . . . . . . _ _ r _ _ , _ _ placebo i

malotilate

12 2 q months

Fig. 3. (a) Actuarial percentage of patients without progression of disease, who received treatment with malotilate (solid line) or placebo (dotted line). (b) Actuarial survival percentage of patients with primary biliary cirrhosis treated with malotilate (solid line) or placebo (dotted

line).

tion of autoimmune activity was considered (IgG was markedly increased, circulating immune complexes and anti-smooth muscle antibodies were detected), and a few days after malotilate was discontinued treatment with prednisone was instituted. There was a rapid and com- plete recovery. Corticosteroid treatment was discontin- ued after 5 months. All other patients with suspected adverse drug effects also recovered completely.

Discussion

The results of this controlled trial evaluating maloti- late in primary biliary cirrhosis suggest that malotilate improves several of the biochemical and immunological parameters of disease activity and reduces liver inflam- mation. The most pronounced biochemical effects were decreases in serum alkaline phosphatase, transaminases and IgM.

Although the exact mode(s) of action still need to be elucidated, malotilate does not seem to inhibit free radical formation or directly influence collagen biosyn- thesis (28,31). Malotilate may prevent the accumulation

of fat in the liver by accelerating mitochondrial fatty acid oxidation and may prevent necrosis and fibrosis by mitigating fibrinogenic stimuli. In view of this hypothesis the findings of a study on eicosanoid formation using human ascitic cells may be relevant. In this study malotilate selectively inhibited the 5-1ipoxygenase path- way, and thus the formation of leukotriene B4 (LTB4), and stimulated the 12- and 15-1ipoxygenase pathways, a unique differential drug effect not previously reported (41). Since LTB4 and other eicosanoids through a variety of effects play a major role in the regulation of inflamma- tory and immune responses on several cell types, these findings may provide an explanation for the effect of malotilate in inflammatory processes.

This study indicates that malotilate interferes not only with inflammatory reactions, but also modulates meta- bolic processes. The mechanism by which malotilate increases cholesterol, an effect previously reported (36,42), is still speculative. Malotilate-induced elevations of serum cholesterol levels may increase the risk of atherosclerotic vascular disease, but probably only if LDL-cholesterol increases. A study in rats showed that malotilate increased total serum cholesterol by selec- tively increasing HDL-cholesterol (42). The decreases in inflammatory cell infiltrate and in immuno'globulin sug- gest that malotilate interferes with the inflammatory process by an immune-modulating effect, resembling that of prednisone.

It must be emphasized that all significant biochemical changes reported in this study were demonstrated using two different methods of statistical analysis. Not only was a modification of the more conventional method of comparing values at different times between and within groups used, but also the data were analyzed using the life-table method. As has been argued before (10), the latter method is not only the standard method to analyze death or other clinical events, but may also be a tool to analyze biochemical data. The main advantage of the actuarial method is that it provides a more reliable assessment of patients with incomplete follow-up. A disadvantage is that continuous variables must be regarded as discrete, thus necessitating the definition of arbitrary 'events'. Another problem is that biochemical values show variation in time, implying that criteria for an event may no longer be present at a later time. For these reasons the similar results using these two comple- mentary statistical methods strengthens the reliability of our observations.

Although malotilate seemed to diminish liver inflam- mation, which is in agreement with the results of a previous controlled trial (36), we did not detect a benefi- cial effect on fibrogenesis. Malotilate had no marked

234 A EUROPEAN MULTICENTRE STUDY GROUP

effect on prur i tus and did not induce remiss ion of the

disease. Also, as expected there was no difference in

survival between the two pa t ien t groups. The na tu ra l

h is tory of PBC, the length of the tr ial and the relat ively

mild degree of liver disease of the pat ients in this s tudy

a pr ior i excluded such a finding. F o r the same reasons

it was no surprise that a difference in disease progress ion,

as reflected by changes in the Ch i ld -Pugh score, was not

detected. To establ ish a possible beneficial effect on the

clinical course of PBC, studies of much longer du ra t i on

would be necessary.

Malo t i l a t e t rea tment seems to be associa ted with a

n u m b e r of adverse effects. The most c o m m o n are allergic

symptoms , nausea and diar rhea . All adverse effects com-

pletely d i sappea red ei ther spon taneous ly after the dose

was lowered or after t r ea tment was discont inued. The

number of suspected adverse effects may have been

related to the dose of 1500 mg malo t i la te daily. In

previous studies, which usually used lower doses, the

incidence of side effects was lower (36). Therefore it is

pe rhaps advisable to use lower doses in future studies.

(Malo t i l a te will remain avai lable on the Japanese marke t

but is only suppl ied for non-cl inical studies by the

manufacturer , personal communica t ion) .

Ursodeoxycho l i c acid now seems the most a t t rac t ive

first line t rea tment op t ion in PBC, but a beneficial effect

on fibrosis and survival awai ts conf i rmat ion. Also this

d rug does not a p p e a r to induce comple te inac t iva t ion

of the disease in most pat ients . Fu r the r s tudies which

evaluate a c o m b i n a t i o n of bile acid- and immune-

modu la t i ng t rea tments seem warran ted .

We conclude that malo t i l a te has an an t i - i n f l ammato ry

effect as well as a number of o ther we l l -documented

effects and remains a d rug of po ten t ia l scientific interest.

However , also cons ider ing the poten t ia l adverse effects

of malot i la te , the clinical relevance of the effects observed

in this s tudy a p p e a r ra ther l imited. Based on the results

of this trial, malo t i la te canno t be r ecommended as a

s ingle-drug therapy for p r imary bi l iary cirrhosis.

Acknowledgements

We are indebted to Dor ien G e ra rd s for excellent

secretar ia l assis tance and to Jan Boot for exper t help

with the compute r i zed da ta m a n a g e m e n t and analysis;

to Z y m a S.A. for supply ing and d i s t r ibu t ing the tr ial

medicat ion; to Ton Janssen and G i a n - P a o l o Ravelli for

o rgan iza t iona l assis tance and to Rudo l f Preisig (Berne,

Switzerland), who was the official adverse d rug moni tor .

The s tudy was suppor t ed in par t by Z y m a S.A., Nyon ,

Switzer land, and by N i h o n N o h y a k u , Tokyo, Japan.

References

l Kaplan MM. Primary biliary cirrhosis. N Engl J Med 1987; 316: 521-8.

2 Wiesner RH, Grambsch PM, Lindor KD, Ludwig J, Dickson ER. Clinical and statistical analyses of new and evolving therapies for primary biliary cirrhosis. Hepatology 1988; 8: 668-76.

3 Triger DR. Autoimmune chronic active hepatitis and primary biliary cirrhosis. In: Davis M, ed. Therapy of Liver Disease. Clinical Gastroenterology. London: Bailli~re Tindall, .1989; 3: 21-38.

4 Markus BH, Dickson ER, Grambsch PM, et al. Efficacy of liver transplantation in patients with primary biliary cirrhosm. N Engl J Med 1989; 320: 1709-13.

5 Epstein O, Jain S, Lee RG, Cook DG, Boss AM, Scheuer PJ, Sherlock S. D-penicillamine treatment improves survival in primary biliary cirrhosis. Lancet 1981; i: 1275-7.

6 Matloff DS, Alpert E, Resnick RH, Kaplan MM. A prospective trial of d-penicillamine in primary biliary cirrhosis. N Engl J Med 1982; 306: 319-26.

7 Bassendine MF, Macklon AF, Mulcahy R, James OFW. Controlled trial of high and low dose D-penicillamine in primary biliary cirrhosis. Gut 1982; 23:A909 (Abstr.).

8 Taal BG, Schalm SW, Ten Kate FJW, Van Berge Henegouwen GP, Brandt KH. Low therapeutic value of D-penicillamine in a short-term prospective trial m primary biliary cirrhosis. Liver 1983; 3: 345-52.

9 Dickson ER, Fleming TR, Wiesner RH, et al. Trial of penicillamine in advanced primary biliary cirrhosis. N Engl J Med 1985; 312: I01 I-5.

10 Neuberger J, Christensen E, Portmann B, et al. Double blind

controlled trial of d-penicillamine in patients with primary biliary cirrhosis. Gut 1985; 26:114-9.

I 1 Bodenheimer HC Jr, Schaffner F, Sternlieb 1, Klion FM, Vernace S, Pezullo J. A prospective clinical trial of D-penicillamine in the treatment of primary biliary cirrhosis. Hepatology 1985; 5:1139-42.

12 Triger DR, Underwood JCE. D-penicillamine in primary biliary cirrhosis. Dig Dis Sci 1986: 31: 130S.

13 Heathcote J, Ross A, Sherlock S. A prospective controlled trial of azathioprine in primary biliary cirrhosis. Gastroenterology 1976; 70: 656-60.

14 Christensen E, Neuberger J, Crowe J, et al. Beneficial effect of azathioprine and prediction of prognosis in primary biliary cirrho- sis. Final results of an international trial. Gastroenterology 1985; 89: 1084-91.

15 Kaplan MM, Ailing DW, Zimmerman HJ, et al. A prospective trial ofcolchicine for primary biliary cirrhosis. N Engl J Med 1986; 315: 1448-54.

16 Warnes TW, Smith A, Lee FI, Haboudi NY, Johnson PJ, Hunt L. A controlled trial of colchicine in primary biliary cirrhosis - trial design and preliminary report. J Hepatol 1987; 5: 1-8.

17 Bodenheimer H Jr, Schaffner F, Pezzullo J. Evaluation ofcolchicine therapy in primary bdiary cirrhosis. Gastroenterology 1988; 95: 124-9.

18 Hoofnagle JH, Davis GL, Schafer DF, et al. Randomized trial of chlorambucil for primary biliary cirrhosis. Gastroenterology 1986; 91: 1327-34.

19 Wiesner RH, Ludwig J, Lindor KD, et al. A controlled trial of cyclosporine in the treatment of primary biliary cirrhosis. N Engl J Med 1990; 322: 1419-24.

20 Minuk GY, Bohme CE, Burgess E, et al. Pilot study ofcyclosporin A in patients with symptomatic primary biliary cirrhosis. Gastro- enterology 1988; 95: 1356-63.

MALOTILATE IN PRIMARY BILIARY CIRRHOSIS 235

21 Mitchison HC, Bassendine MF, Malcolm A J, Watson A J, Record CO, James OFW. A pilot double-blind controlled l-year trial of prednisolone treatment in primary biliary cirrhosis: hepatic improvement but greater bone loss. Hepatology 1989; 10: 420-9.

22 Kaplan MM, Knox TA, Arora S. Primary biliary cirrhosis treated with low-dose oral pulse methotrexate. Ann Intern Med 1988; 109: 429-3 I.

23 Kaplan MM, Knox TA. Treatment of primary biliary cirrhosis with low-dose weekly methotrexate. Gastroenterology 1991; 10h 1332-8.

24 James OFW. D-penicillamine for primary biliary cirrhosis. Gut 1985; 26: 109-13.

25 Boyer JL. Definitive therapy for primary biliary cirrhosis - fact or fiction? Gastroenterology 1988; 95: 242-5.

26 Poupon RE, Balkau B, Eschw6ge E, Poupon R and the UDCA- PBC Study Group. A multicenter, controlled trial of ursodiol for the treatment of primary biliary cirrhosis. N Engl J Med 1991; 324:1548-54.

27 Leuschner U, Fischer H, Kurtz W, et al. Ursodeoxycholic acid in primary biliary cirrhosis: results of a controlled double-blind trial. Gastroenterology 1989; 97: 1268-74.

28 Sugimoto T. Malotilate: an overview. In: Oda T, Tygstrup N, eds. Hepatotrophic Agent Malotilate. Tokyo: Excerpta Medica, 1983: 1-8.

29 Dumont JM, Maignan MF, Janin B, Herbage D, Perissoud D. Effect of malotilate on chronic liver injury induced by carbon tetrachloride in the rat. J Hepatol 1986; 3: 260-8.

30 Siegers CP, Pauh V, Korb G, Younes M. Hepatoprotection by malotilate against carbon tetrachlorid~-alcohol-induced liver fibrosis. Agents Actions 1986; 18: 600-3.

31 Monna T. Effect of malotilate on experimental hver fibrosis (cirrho- sin) induced by carbon tetrachloride or egg yolk sensitization. In: Oda T, Tygstrup N, eds. Hepatotrophic Agent Malotllate. Tokyo: Excerpta Medica, 1983: 44-53.

32 Younes M, Siegers CP. Effect of malotilate on paracetamol-induced hepatotoxicity. Toxicol Lett 1985; 25: 143-6.

33 Dumont JM, Maignan MF, Perissoud D. Protective effect of malotilate against galactosamine intoxication in the rat. Arch Int Pharmacodyn Ther 1987; 289: 296-310.

34 Fujisawa K, Kimura K, Yamauchi M, et al. Studies on the lipotropic effect of malotilate on the development of fatty liver induced by orotic acid. In: Oda T, Tygstrup N, eds. Hepatotrophic Agent Malotilate. Tokyo: Excerpta Medica, 1983: 29-43.

35 Katoh M, Notaka M, Sugimoto T, Kasai T. Protective effect of diisopropyl-l,3-dithiol-2-ylidene malonate (NKK-105) on chemi- cally induced liver injury. Jpn J Pharmacol 1978; 28: P125.

36 Suzuki H, Ichida F, Takini T, et al. Therapeutic effects of malotilate on chronic hepatitis and liver cirrhosis: a double blind, controlled multicenter trial. In: Oda T, Tygstrup N, eds. Hepatotrophic Agent Malotilate. Tokyo: Excerpta Medica, 1983: 54-68.

37 Takada A, Nei J, Tamino H, Takase S. Effects of malotilate on ethanol-inhibited hepatocyte regeneration in rats. J Hepatol 1987; 5: 336-43.

38 Niwano Y, Katoh M, Uchida M, Sugimoto T. Acceleration of liver regeneration by malotilate in partially hepatectomized rats. Jpn J Pharmacol 1986; 40:411-5.

39 Ludwig J, Dickson ER, McDonald GSA. Staging of chronic nonsuppurative destructive cholangitis (syndrome of primary biliary cirrhosis). Virchows Arch I-A-I 1978; 379: 103-12.

40 Pugh RNH, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the esophagus for bleeding esophageal varices. Br J Surg 1973; 60: 646-9.

41 Zijlstra FJ, Wilson JHP, Vermeer MA, Ouwendijk RJT, Vincent JE. Differential effects of malotilate on 5-, 12- and 15-1ipoxygenase in human ascites cells. Eur J Pharmacol 1989; 159: 291-5.

42 Kitagawa H. Effects of malotilate on cholesterol metabolism and micrososmal electron transport system in rat livers. In: Oda T, Tygstrup N, eds. Hepatotrophic Agent Malotilate. Tokyo: Excerpta Medica, 1983: 21-8.