DRUG EXPERIENCE
Drug Safety 6 (5): 350-360, 1991 0114-5916/91/0009-0350/$05.50/0 © Adis International Limited. All rights reserved. DRS1036
Systemic Lupus Erythematosus (SLE)-Like Syndromes Associated with Carbamazepine Therapy K.K. Jain Central Drug Monitoring, Medical Department, Ciba-Geigy, Basel, Switzerland
Contents 350
351 351 351 352 352 353 353 354 355 355 357 357 358 358 358 358
Summary
Summary 1. A Review of Spontaneous Systemic Lupus Erythematosus 1.1 Pathogenesis and Aetiology 1.2 Clinical Features I.3 Laboratory Diagnosis 2. Drug-Induced Lupus Erythematosus 2.1 Mechanism 2.2 Anticonvulsant Drugs and SLE-Like Syndromes 2.2.1 Case Reports of Carbamazepine and SLE-Like Syndromes 3. Discussion 3.1 Case Reports 3.2 Pharmacoepidemiological Aspects 3.3 Causal Relationship 3.4 Pathogenic Mechanism 1.5 Severity and Outcome 4. Management of Drug-Induced Lupus Erythematosus 5. Conclusions
Records of 80 cases of systemic lupus erythematosus (SLE)-like syndromes associated with use of carbamazepine entered in the single case files of Central Drug Monitoring, Ciba-Geigy, Basel, were reviewed. Included in these were 11 cases reported in the literature. These were assessed in light of the current concepts of both idiopathic and drug-induced forms of SLE. Even allowing for under-reporting of adverse drug reactions, the number of such cases is far below the prevalence rates for SLE. This adverse drug reaction remains allocated to the category ' isolated cases' i.e. frequency is 0.5 g/day or cellular casts) Neurological disorder (seizures or psychosis) Haematological disorder haemolytic anaemia
or leucopenia «4 x 109 /L leucocytes on 2 or more occasions)
or lymphopenia «1.5 x 109/L lymphocytes on 2 or more occasions)
1.1 Pathogenesis and Aetiology
or thrombocytopenia «100 x 109/L thrombocytes in the
Antibodies to double-stranded (ds) DNA found frequently in SLE result from an activation of Bcells by an antigenic stimulus in an c;:nvironment where there is suppression of T -helper cells. Genetic predisposition for SLE is indicated by a higher concordance in monozygotic than in dizygotic twins, a 10% frequency of patients with more than I affected individual in the family and the fact that 6% of SLE patients have inherited deficiencies of complement components (Hahn 1991). Genetic disposition is also an important factor in drug-induced SLE (Stratton 1985). Other factors potentially involved in the development or exacerbation SLE include C-type RNA viruses, estrogens and exposure to UV light (Koffler 1980). 1.2 Clinical Features Clinical manifestations of SLE are very variable. Systemic symptoms are usually prominent and include malaise, fatigue, fever, anorexia and weight loss. The systems most frequently involved are:
absence of offending drugs) Immunological disorder positive LE cell preparation or anti-DNA (antibody against native DNA) in abnormal titres
or anti-SM (antibody against Smith nuclear antigen) in abnormal titres or false-positive test for syphilis Antinuclear antibody in abnormal titres
musculoskeletal (95%), cutaneous (80%), haematological (85%), neurological (60%), cardiopulmonary (50%) and renal (50%) [Hahn 1991]. The diagnostic criteria are shown in table I. SLE may involve one system at the onset and become multisystemic later, or manifest as a multisystem disease from the onset. The intensity of the disease varies, and periods of exacerbation may alternate with periods of relative quiescence. Less than 10% of patients have symptom-free remissions. The natural course ofSLE is incompletely documented,
352
but a study of 100 patients from Southwestern England who were followed for 6 months to 11 years, showed that the 5-year survival was 88%. Joint problems (94%), rash (90%) and haematological abnormalities (89%) were the most common clinical features. Antinuclear antibodies were found in 98% and antiphospholipid antibodies in 38% (Worrall et al. 1990). Involvement of the nervous system in SLE carries a poor prognosis and is the most common cause of death in these patients (Feng et al. 1973; Gibson & Myers 1976). Neurological and psychiatric symptoms are found in 60% of patients with SLE (Trautman et al. 1985). In a clinicopathological study of 24 SLE cases, Johnson and Richardson (1968) found that the CNS was involved in 75% of the cases. Various manifestations were: seizures (54%), cranial nerve disorders (42%), hemiplegia (12%), paraplegia (4%), peripheral neuropathy (8%), mental disorders (33%). Epilepsy may precede development of other signs of SLE by several years (Mackworth-Young & Hughes 1985; Russell et al. 1951). It has been suggested that the disease should be considered in the differential diagnosis of every patient with epilepsy and arthritis (Russell et al. 1951). Neurological manifestations such as peripheral neuropathy (Bloch et al. 1979; Hughes et al. 1982) cerebral vasculitis (Sanders & Hogenhuis 1986), pseudobulbar palsy (Hazelton et al. 1980), hemiplegia and choreiform movements (Siekert & Clark 1955) were reported as initial manifestations of SLE. SLE, therefore, should be considered in the differential diagnosis of a patient with obscure neurological problem: Mavrikakis et al. (1980) reported a patient with cerebral lupus with diabetes inspidus; and transverse myelopathy (Andrianakos et al. 1975) and cerebral atrophy (Gordon et al. 1986) may occur as complications of SLE. A patient with neuromyelitis optica in SLE was reported by April and Vansonnenberg (1976). Cases of SLE resembling multiple sclerosis have been described by Allen et al. (1979) and Fulford et al. (1972). Neuropsychological testing of patients with SLE has been reported to show deficits both with and without involvement of the CNS (Kutner et al. 1988).
Drug Safety 6 (5) 1991
1.3 Laboratory Diagnosis A number of antibodies are found in patients with SLE (for a detailed list see Hahn 1991). Antinuclear antibodies (ANA) are the best screening test, as they are found in more than 95% ofSLE patients (Hahn 1991). A positive ANA test, however, is not specific for SLE; low titres may be found in healthy individuals, those with other autoimmune diseases, acute viral infections and chronic inflammatory processes, and in patients on various medications. A negative ANA test does not rule out SLE; antibodies to dsDNA are relatively specific for SLE.
2. Drug-Induced Lupus Erythematosus Several drugs can cause a syndrome resembling SLE in individuals without any obvious predisposition to the disease. Hydralazine was the first drug reported to produce drug-induced lupus erythematosus (DlLE) [Henn et al. 1955]. The most common offender is procainamide which induces ANA in 50 to 75% of individuals within 1 to 2 months, and 20% of patients develop clinical DILE (Blomgren 1973). Drugs implicated as activators of SLE are listed in table II. There is no common feature in the chemical structure of the drugs listed. Definitive criteria have not yet been established for distinguishing idiopathic and drug-induced lupus erythematosus, but a number of requirements are agreed by most authors (Cohen & Prowse 1989). These are: (a) the appearance of SLE-like symptoms and signs during the course of drug therapy; (b) no prior clinical of laboratory features of SLE - a drug may, however, unmask subclinical SLE; (c) resolution of clinical features on withdrawal of the drug within days to weeks - persistence of symptoms after withdrawal suggests activation of idiopathic SLE; and (d) the presence of antihistone antibodies in the absence of high titres of antidsDNA antibodies usually suggests the diagnosis of DlLE (Meyer et al. 1984). In idiopathic SLE, ANA are of heterogenous specificity and may consist of antibodies to native DNA, histones or nonhistone proteins. ANA to DILE are less heterogeneous and consist mainly of
353
SLE-Like Syndromes and Carbamazepine
Table II. Drugs implicated as activators of drug-induced lupus erythematosus (from Dubois & Wallace 1987, with permission) Antlconvulsants Carbamazepine Phenytoin 8 Ethosuximide Mephenytoin Phenylethylaceturea Primidone Trimethadione Antipsychotics Chlorpromazine8 Chlorprothixene Levomepromazine or methotrimeprazine Lithium carbonate Perazine Perphenazine Phenelzine Promethazine Thioridazine
NSAIDs Benoxaprofen Ibuprofen Phenylbutazone Antihypertensives Acebutolol Captopril Clonidine Guanoxan Hlldralazine8 MethyldopaB Pindolol Practolol Prazosin
Antibiotics Griseofulvin Nalidixic acid Nitrofurantoin Penicillin Sulfadimethoxine Sulfamethoxypyridazine Tetracycline
Antituberculosis agents Isoniazid Para-aminosalicylic acid Streptomycin Antiarrhythmics Disopyramide Procainamide8 Quinidine Hormonal drugs Danazol Methimazole Methylthiouracil Oral contraceptives Propylthiouracil Miscellaneous Amoproxan Anthiomaline Cinnarizine Gold salts Levodopa Methysergide Oxyphenisatin D-Penicillamine8 Phenopyrazone SulfasalazineB Tolazamide
a
Drugs the authors consider as well-documented provocateurs of SLE. Abbreviations: NSAIDs '" nonsteroidal anti·inflammatory drugs.
antibodies to histones (Fritzler & Tan 1978). Presence of antihistone antibodies is strong evidence that the symptoms of a given patient are due to DILE (Epstein & Barland 1985). Differences between idiopathic SLE and DILE are shown in table III.
2.1 Mechanism DILE is usually considered to be an idiosyncratic drug reaction, i.e. an adverse effect not expected on the basis of known pharmacological properties of the responsible medicine. Proposed mechanisms of DILE have been reviewed by Dubois and Wallace (1987) and include: (a) interactions of drugs and nuclear antigens; (b) interactions of drugs with lymphocytes; (c) photosensitivity damage to epidermal DNA; and (d) immunogenic metabolites. Although DILE is associated with autoantibody production, it is unclear what role the autoantibodies play in the clinical manifestation of the disease (Uetrecht 1990). An association between slow acetylator status and increased risk of DILE has been suggested (Uetrecht & Woosley 1981), but Chinese of Southeast Asian descent have a relatively high prevalence of rapid acetylators and also a high incidence of SLE (Kumana et al. 1990). These authors found a lack of association between slow acetylator status and SLE in study subjects from this population. 2.2 Anticonvulsant Drugs and SLE-Like Syndromes SLE-like syndromes induced by anticonvulsants are cQnsidered to be caused by the potential of these drugs to induce ANA (Alarcon-Segovia et al. 1972). Anticonvulsant drug-induced SLE-like syndromes in children are commonly associated with administration of phenytoin, ethosuximide and trimethadione (Benton et al. 1962; Livingston et aL 1968; Wilske et al. 1965). Singsen et al. (1976) tested a group of 101 children from a seizure clinic for the presence of ANA; antibodies were found in 14 of the 70 children receiving ethosuximide and/or phenytoin. Only 5 of these had symptoms resembling LE. After discontinuation of the medication in all children, 4 recovered while the fifth continued to have SLE with nephritis despite continuation of therapy. A follow-up after 10 months showed that no asymptomatic child with ANA had developed clinical evi-
Drug Safety 6 (5) 1991
354
Table III. Differences between idiopathic and drug-associated systemic lupus erythematosus (SLE)
Clinical features
Idiopathic SLE
Drug-associated SLE
Cutaneous manifestations CNS involvement Renal involvement Antinuclear antibodies Antibodies to 'native' (double-stranded) DNA Antihistone antibodies
80% Common Common Present in > 85% Present in 85% Present in 50%
Less frequent (50%) Rare Rare Always present Almost always absent Present in 95%
dence of SLE. These results suggest that asymptomatic children who develop ANA should be observed carefully, but need not have their anticonvulsant medication discontinued. 2.2.1 Case Reports of Carbamazepine and SLE-Like Syndromes There were 80 cases of SLE-like syndromes entered in single-case files at Ciba-Geigy Headquarters, Basel. These reports covered all countries, from introduction of the drug in 1963 until October 1990. Spontaneous reports forwarded by clinicians or published were evaluated and entered into a database. Adverse drug reactions were recorded as reported, even if supporting evidence was missing. The diagnostic terms used in the reports submitted
Table IV. Distribution of systemic lupus erythematosus associated with carbamazepine by age and sex (neither were stated in 11 patients)
Age at onset of unwanted effect (y)
Number of patients male
Not stated ~10
11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 >90 Total
3 3 3
female
total
5 2 13 19 7 3 6
16 5 16 22 8 3 6
56
80
2
2 13
were vague in some cases. Patients with skin rash described as lupus-like rash were included. Two patients with only pulmonary lesions, referred to as lupus-like in 1 and lupoid drug-reaction in the other, have been excluded from the SLE analysis because of lack of evidence ofSLE. Reports of vasculitis alone were also excluded and assigned to the category of hypersensitivity reactions involving blood vessels. No information was available on antiphospholipid antibodies. Two patients had associated lupus anticoagulant but no details were available. These 80 cases originated from 19 countries. Age and Sex of the Patients (Table IV): Females predominated (56 women versus 13 men) in the 69 reports in which the sex was stated. The majority of the patients (43 of 64 in whom age was stated) were below the age of 30. Causality and Severity of Reactions (Table V): Causality assessment is an estimate of the relationship of the drug to the adverse effect. The term 'definite' is used only if there is objective evidence ·of the relationship, such as recurrence of the adverse effect after re-administeration of the drug. 'Probable' means that the symptoms were relieved after discontinuation of the drug. 'Possible' simply means that the role of the suspected drug cannot be excluded even though other factors, such as comedications, known to cause the adverse effect were present. The causality rating in the 80 cases examined was as follows: definite (l), probable (lO), possible (32), unlikely (3), no relation (5), insufficient data for assessment (23) and causality not assessed (6). In terms of severity, the reported adverse effects
355
SLE-Like Syndromes and Carbamazepine
were classed as: severe/serious (24); mild/moderate (39); and unknown {I 7). There were no deaths in this group. Dose-Effect Relationships: Most of the reported cases occurred at the average therapeutic dose of 400 to 600mg (fig. I), suggesting that carbamazeTable V. Frequency of symptoms/findings of SLE reported in 80 cases of SLE-Iike syndromes associated with Tegretol'
(carbamazepine) therapy. The number of symptoms listed varies from 0 to 5 per case
20
15 (I)
5:
~o 10 o z 5
20'·400
40,·600
601-800
80HOOO
>1000
Daily dose (mg) No. of cases
%
Only listing of SLE-like syndrome without details except laboratory
25
31 .25
findings reactivation of SLE Systemic manifestations: fatigue, malaise, weight loss Musculoskeletal: arthralagia, myalgia Cutaneous manifestations malar or butterfly rash other types of skin rashes photosensitivity alopecia cutaneous vasculitis oral ulcers Haematological effects anaemia of chronic disease haemolytic anaemia leucopenia thrombocytopenia lymphopenia lymphadenopathy Neurological effects ataxia (associated with toxic carbamazepine levels) headache (no specific cause) Pericarditis (pericardial effusions) Pleurisy (or pleural effusions) Renal abnormalities (proteinuria/ haematuria) Gastrointestinal (nonspecific) diarrhoea Thrombophlebitis (associated with circulating anticoagulant) Eye [conjunctivitis, scleritis, swelling (1 case each)] Thyroiditis No clinical symptoms (only positive laboratory tests)
2 17
2.50 21.25
19
23.75
19 14 3 4
23.75 17.50 3.75 5.00 1.25 1.25
2 2 8 4 1 2
2.50 2.50 10.00 5.00 1.25 2.50 25.00
4 3 2
1.25 5.00 3.75 2.50
3
3.75
2
3.75
3
1.25 1.25 1.25
Fig. 1. Relation of daily dose of'Tegretol' (carbamazepine) to frequency of SLE-like syndromes in 45 cases. Note most cases occurred at less than the average daily dose of 600mg.
pine-associated SLE is not dose related. Information on the relationship between duration of treatment and onset of symptoms was available in 40 patients (fig. 2). Most of the cases occurred after treatment for 6 months or longer. Concomitant Drugs: Information about concomitant medications was available in only 45 cases. Of the 40 patients in whom various drugs were used, 15 were taking other agents which can themselves induce SLE-like syndrome. These were: phenytoin (7 patients), primidone (2), ethosuximide (1), promethazine (2), acebutolol (1) and oral contraceptives (2). Outcome: This was assessed as follows: recovery (21 patients), no recovery at time of reporting (21), unknown (38). Cases from the Literature: Of the 80 cases in the database, II had been described in the literature: these are summarised in table VI. Only 6 can be confirmed as DILE; 4 were considered to be idiopathic SLE; and insufficient information was available in I case (abstract only) [Rabinowicz et al. 1990].
3. Discussion 3.1 Case Reports Discussion of the 80 case reports is limited by the lack of full information in most cases. The limitations of this data are: (a) not all cases of adverse
....
Table VI. Published cases of SLE-like syndromes in patients receiving 'Tegretol' [carbamazepine (CBZ) therapy]
VI
0-
Reference
Laboratory findings Duration of therapy Anti-dsDNA before onset ANA
Agel sex
Clinical features
Alballa et al. (1987) Bateman (1985)
30/F
Arthritis, rash
1 year
+
30/F
1 year
+
Beurey et al. (1971)
42/F
Arthralgia, rash, pleuritis, vasculitis, leucopenia Rash, arthralgia
4 years
Drory et al. (1983)
181M
Arthralgia, pleuritis, perlcorditis, rash
Kolstee (1982) Livingston et al. (1974) Maravic et al. (1990)
251M 46/F
McNichol (1990)
Rabinowicz et al. (1990) Simpson (1966) Takigawa et al. (1976)
Antihistone LE cells
Results of stopping CBZ
Remarks
Recovery in 18 months
First fully documented case Grayson (1985) stated this case probably spontaneous SLE not induced by CBZ Patient became worse after discontinuation of treatment of SLE Patient met 6 SLE criteria of AMA. Presence of + anti-DNA against diagnosis of DILE
+ 1640
NA
+ 94
U/ml (normal
Drug Safety 6 (5): 350-360, 1991 0114-5916/91/0009-0350/$05.50/0 © Adis International Limited. All rights reserved. DRS1036
Systemic Lupus Erythematosus (SLE)-Like Syndromes Associated with Carbamazepine Therapy K.K. Jain Central Drug Monitoring, Medical Department, Ciba-Geigy, Basel, Switzerland
Contents 350
351 351 351 352 352 353 353 354 355 355 357 357 358 358 358 358
Summary
Summary 1. A Review of Spontaneous Systemic Lupus Erythematosus 1.1 Pathogenesis and Aetiology 1.2 Clinical Features I.3 Laboratory Diagnosis 2. Drug-Induced Lupus Erythematosus 2.1 Mechanism 2.2 Anticonvulsant Drugs and SLE-Like Syndromes 2.2.1 Case Reports of Carbamazepine and SLE-Like Syndromes 3. Discussion 3.1 Case Reports 3.2 Pharmacoepidemiological Aspects 3.3 Causal Relationship 3.4 Pathogenic Mechanism 1.5 Severity and Outcome 4. Management of Drug-Induced Lupus Erythematosus 5. Conclusions
Records of 80 cases of systemic lupus erythematosus (SLE)-like syndromes associated with use of carbamazepine entered in the single case files of Central Drug Monitoring, Ciba-Geigy, Basel, were reviewed. Included in these were 11 cases reported in the literature. These were assessed in light of the current concepts of both idiopathic and drug-induced forms of SLE. Even allowing for under-reporting of adverse drug reactions, the number of such cases is far below the prevalence rates for SLE. This adverse drug reaction remains allocated to the category ' isolated cases' i.e. frequency is 0.5 g/day or cellular casts) Neurological disorder (seizures or psychosis) Haematological disorder haemolytic anaemia
or leucopenia «4 x 109 /L leucocytes on 2 or more occasions)
or lymphopenia «1.5 x 109/L lymphocytes on 2 or more occasions)
1.1 Pathogenesis and Aetiology
or thrombocytopenia «100 x 109/L thrombocytes in the
Antibodies to double-stranded (ds) DNA found frequently in SLE result from an activation of Bcells by an antigenic stimulus in an c;:nvironment where there is suppression of T -helper cells. Genetic predisposition for SLE is indicated by a higher concordance in monozygotic than in dizygotic twins, a 10% frequency of patients with more than I affected individual in the family and the fact that 6% of SLE patients have inherited deficiencies of complement components (Hahn 1991). Genetic disposition is also an important factor in drug-induced SLE (Stratton 1985). Other factors potentially involved in the development or exacerbation SLE include C-type RNA viruses, estrogens and exposure to UV light (Koffler 1980). 1.2 Clinical Features Clinical manifestations of SLE are very variable. Systemic symptoms are usually prominent and include malaise, fatigue, fever, anorexia and weight loss. The systems most frequently involved are:
absence of offending drugs) Immunological disorder positive LE cell preparation or anti-DNA (antibody against native DNA) in abnormal titres
or anti-SM (antibody against Smith nuclear antigen) in abnormal titres or false-positive test for syphilis Antinuclear antibody in abnormal titres
musculoskeletal (95%), cutaneous (80%), haematological (85%), neurological (60%), cardiopulmonary (50%) and renal (50%) [Hahn 1991]. The diagnostic criteria are shown in table I. SLE may involve one system at the onset and become multisystemic later, or manifest as a multisystem disease from the onset. The intensity of the disease varies, and periods of exacerbation may alternate with periods of relative quiescence. Less than 10% of patients have symptom-free remissions. The natural course ofSLE is incompletely documented,
352
but a study of 100 patients from Southwestern England who were followed for 6 months to 11 years, showed that the 5-year survival was 88%. Joint problems (94%), rash (90%) and haematological abnormalities (89%) were the most common clinical features. Antinuclear antibodies were found in 98% and antiphospholipid antibodies in 38% (Worrall et al. 1990). Involvement of the nervous system in SLE carries a poor prognosis and is the most common cause of death in these patients (Feng et al. 1973; Gibson & Myers 1976). Neurological and psychiatric symptoms are found in 60% of patients with SLE (Trautman et al. 1985). In a clinicopathological study of 24 SLE cases, Johnson and Richardson (1968) found that the CNS was involved in 75% of the cases. Various manifestations were: seizures (54%), cranial nerve disorders (42%), hemiplegia (12%), paraplegia (4%), peripheral neuropathy (8%), mental disorders (33%). Epilepsy may precede development of other signs of SLE by several years (Mackworth-Young & Hughes 1985; Russell et al. 1951). It has been suggested that the disease should be considered in the differential diagnosis of every patient with epilepsy and arthritis (Russell et al. 1951). Neurological manifestations such as peripheral neuropathy (Bloch et al. 1979; Hughes et al. 1982) cerebral vasculitis (Sanders & Hogenhuis 1986), pseudobulbar palsy (Hazelton et al. 1980), hemiplegia and choreiform movements (Siekert & Clark 1955) were reported as initial manifestations of SLE. SLE, therefore, should be considered in the differential diagnosis of a patient with obscure neurological problem: Mavrikakis et al. (1980) reported a patient with cerebral lupus with diabetes inspidus; and transverse myelopathy (Andrianakos et al. 1975) and cerebral atrophy (Gordon et al. 1986) may occur as complications of SLE. A patient with neuromyelitis optica in SLE was reported by April and Vansonnenberg (1976). Cases of SLE resembling multiple sclerosis have been described by Allen et al. (1979) and Fulford et al. (1972). Neuropsychological testing of patients with SLE has been reported to show deficits both with and without involvement of the CNS (Kutner et al. 1988).
Drug Safety 6 (5) 1991
1.3 Laboratory Diagnosis A number of antibodies are found in patients with SLE (for a detailed list see Hahn 1991). Antinuclear antibodies (ANA) are the best screening test, as they are found in more than 95% ofSLE patients (Hahn 1991). A positive ANA test, however, is not specific for SLE; low titres may be found in healthy individuals, those with other autoimmune diseases, acute viral infections and chronic inflammatory processes, and in patients on various medications. A negative ANA test does not rule out SLE; antibodies to dsDNA are relatively specific for SLE.
2. Drug-Induced Lupus Erythematosus Several drugs can cause a syndrome resembling SLE in individuals without any obvious predisposition to the disease. Hydralazine was the first drug reported to produce drug-induced lupus erythematosus (DlLE) [Henn et al. 1955]. The most common offender is procainamide which induces ANA in 50 to 75% of individuals within 1 to 2 months, and 20% of patients develop clinical DILE (Blomgren 1973). Drugs implicated as activators of SLE are listed in table II. There is no common feature in the chemical structure of the drugs listed. Definitive criteria have not yet been established for distinguishing idiopathic and drug-induced lupus erythematosus, but a number of requirements are agreed by most authors (Cohen & Prowse 1989). These are: (a) the appearance of SLE-like symptoms and signs during the course of drug therapy; (b) no prior clinical of laboratory features of SLE - a drug may, however, unmask subclinical SLE; (c) resolution of clinical features on withdrawal of the drug within days to weeks - persistence of symptoms after withdrawal suggests activation of idiopathic SLE; and (d) the presence of antihistone antibodies in the absence of high titres of antidsDNA antibodies usually suggests the diagnosis of DlLE (Meyer et al. 1984). In idiopathic SLE, ANA are of heterogenous specificity and may consist of antibodies to native DNA, histones or nonhistone proteins. ANA to DILE are less heterogeneous and consist mainly of
353
SLE-Like Syndromes and Carbamazepine
Table II. Drugs implicated as activators of drug-induced lupus erythematosus (from Dubois & Wallace 1987, with permission) Antlconvulsants Carbamazepine Phenytoin 8 Ethosuximide Mephenytoin Phenylethylaceturea Primidone Trimethadione Antipsychotics Chlorpromazine8 Chlorprothixene Levomepromazine or methotrimeprazine Lithium carbonate Perazine Perphenazine Phenelzine Promethazine Thioridazine
NSAIDs Benoxaprofen Ibuprofen Phenylbutazone Antihypertensives Acebutolol Captopril Clonidine Guanoxan Hlldralazine8 MethyldopaB Pindolol Practolol Prazosin
Antibiotics Griseofulvin Nalidixic acid Nitrofurantoin Penicillin Sulfadimethoxine Sulfamethoxypyridazine Tetracycline
Antituberculosis agents Isoniazid Para-aminosalicylic acid Streptomycin Antiarrhythmics Disopyramide Procainamide8 Quinidine Hormonal drugs Danazol Methimazole Methylthiouracil Oral contraceptives Propylthiouracil Miscellaneous Amoproxan Anthiomaline Cinnarizine Gold salts Levodopa Methysergide Oxyphenisatin D-Penicillamine8 Phenopyrazone SulfasalazineB Tolazamide
a
Drugs the authors consider as well-documented provocateurs of SLE. Abbreviations: NSAIDs '" nonsteroidal anti·inflammatory drugs.
antibodies to histones (Fritzler & Tan 1978). Presence of antihistone antibodies is strong evidence that the symptoms of a given patient are due to DILE (Epstein & Barland 1985). Differences between idiopathic SLE and DILE are shown in table III.
2.1 Mechanism DILE is usually considered to be an idiosyncratic drug reaction, i.e. an adverse effect not expected on the basis of known pharmacological properties of the responsible medicine. Proposed mechanisms of DILE have been reviewed by Dubois and Wallace (1987) and include: (a) interactions of drugs and nuclear antigens; (b) interactions of drugs with lymphocytes; (c) photosensitivity damage to epidermal DNA; and (d) immunogenic metabolites. Although DILE is associated with autoantibody production, it is unclear what role the autoantibodies play in the clinical manifestation of the disease (Uetrecht 1990). An association between slow acetylator status and increased risk of DILE has been suggested (Uetrecht & Woosley 1981), but Chinese of Southeast Asian descent have a relatively high prevalence of rapid acetylators and also a high incidence of SLE (Kumana et al. 1990). These authors found a lack of association between slow acetylator status and SLE in study subjects from this population. 2.2 Anticonvulsant Drugs and SLE-Like Syndromes SLE-like syndromes induced by anticonvulsants are cQnsidered to be caused by the potential of these drugs to induce ANA (Alarcon-Segovia et al. 1972). Anticonvulsant drug-induced SLE-like syndromes in children are commonly associated with administration of phenytoin, ethosuximide and trimethadione (Benton et al. 1962; Livingston et aL 1968; Wilske et al. 1965). Singsen et al. (1976) tested a group of 101 children from a seizure clinic for the presence of ANA; antibodies were found in 14 of the 70 children receiving ethosuximide and/or phenytoin. Only 5 of these had symptoms resembling LE. After discontinuation of the medication in all children, 4 recovered while the fifth continued to have SLE with nephritis despite continuation of therapy. A follow-up after 10 months showed that no asymptomatic child with ANA had developed clinical evi-
Drug Safety 6 (5) 1991
354
Table III. Differences between idiopathic and drug-associated systemic lupus erythematosus (SLE)
Clinical features
Idiopathic SLE
Drug-associated SLE
Cutaneous manifestations CNS involvement Renal involvement Antinuclear antibodies Antibodies to 'native' (double-stranded) DNA Antihistone antibodies
80% Common Common Present in > 85% Present in 85% Present in 50%
Less frequent (50%) Rare Rare Always present Almost always absent Present in 95%
dence of SLE. These results suggest that asymptomatic children who develop ANA should be observed carefully, but need not have their anticonvulsant medication discontinued. 2.2.1 Case Reports of Carbamazepine and SLE-Like Syndromes There were 80 cases of SLE-like syndromes entered in single-case files at Ciba-Geigy Headquarters, Basel. These reports covered all countries, from introduction of the drug in 1963 until October 1990. Spontaneous reports forwarded by clinicians or published were evaluated and entered into a database. Adverse drug reactions were recorded as reported, even if supporting evidence was missing. The diagnostic terms used in the reports submitted
Table IV. Distribution of systemic lupus erythematosus associated with carbamazepine by age and sex (neither were stated in 11 patients)
Age at onset of unwanted effect (y)
Number of patients male
Not stated ~10
11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 >90 Total
3 3 3
female
total
5 2 13 19 7 3 6
16 5 16 22 8 3 6
56
80
2
2 13
were vague in some cases. Patients with skin rash described as lupus-like rash were included. Two patients with only pulmonary lesions, referred to as lupus-like in 1 and lupoid drug-reaction in the other, have been excluded from the SLE analysis because of lack of evidence ofSLE. Reports of vasculitis alone were also excluded and assigned to the category of hypersensitivity reactions involving blood vessels. No information was available on antiphospholipid antibodies. Two patients had associated lupus anticoagulant but no details were available. These 80 cases originated from 19 countries. Age and Sex of the Patients (Table IV): Females predominated (56 women versus 13 men) in the 69 reports in which the sex was stated. The majority of the patients (43 of 64 in whom age was stated) were below the age of 30. Causality and Severity of Reactions (Table V): Causality assessment is an estimate of the relationship of the drug to the adverse effect. The term 'definite' is used only if there is objective evidence ·of the relationship, such as recurrence of the adverse effect after re-administeration of the drug. 'Probable' means that the symptoms were relieved after discontinuation of the drug. 'Possible' simply means that the role of the suspected drug cannot be excluded even though other factors, such as comedications, known to cause the adverse effect were present. The causality rating in the 80 cases examined was as follows: definite (l), probable (lO), possible (32), unlikely (3), no relation (5), insufficient data for assessment (23) and causality not assessed (6). In terms of severity, the reported adverse effects
355
SLE-Like Syndromes and Carbamazepine
were classed as: severe/serious (24); mild/moderate (39); and unknown {I 7). There were no deaths in this group. Dose-Effect Relationships: Most of the reported cases occurred at the average therapeutic dose of 400 to 600mg (fig. I), suggesting that carbamazeTable V. Frequency of symptoms/findings of SLE reported in 80 cases of SLE-Iike syndromes associated with Tegretol'
(carbamazepine) therapy. The number of symptoms listed varies from 0 to 5 per case
20
15 (I)
5:
~o 10 o z 5
20'·400
40,·600
601-800
80HOOO
>1000
Daily dose (mg) No. of cases
%
Only listing of SLE-like syndrome without details except laboratory
25
31 .25
findings reactivation of SLE Systemic manifestations: fatigue, malaise, weight loss Musculoskeletal: arthralagia, myalgia Cutaneous manifestations malar or butterfly rash other types of skin rashes photosensitivity alopecia cutaneous vasculitis oral ulcers Haematological effects anaemia of chronic disease haemolytic anaemia leucopenia thrombocytopenia lymphopenia lymphadenopathy Neurological effects ataxia (associated with toxic carbamazepine levels) headache (no specific cause) Pericarditis (pericardial effusions) Pleurisy (or pleural effusions) Renal abnormalities (proteinuria/ haematuria) Gastrointestinal (nonspecific) diarrhoea Thrombophlebitis (associated with circulating anticoagulant) Eye [conjunctivitis, scleritis, swelling (1 case each)] Thyroiditis No clinical symptoms (only positive laboratory tests)
2 17
2.50 21.25
19
23.75
19 14 3 4
23.75 17.50 3.75 5.00 1.25 1.25
2 2 8 4 1 2
2.50 2.50 10.00 5.00 1.25 2.50 25.00
4 3 2
1.25 5.00 3.75 2.50
3
3.75
2
3.75
3
1.25 1.25 1.25
Fig. 1. Relation of daily dose of'Tegretol' (carbamazepine) to frequency of SLE-like syndromes in 45 cases. Note most cases occurred at less than the average daily dose of 600mg.
pine-associated SLE is not dose related. Information on the relationship between duration of treatment and onset of symptoms was available in 40 patients (fig. 2). Most of the cases occurred after treatment for 6 months or longer. Concomitant Drugs: Information about concomitant medications was available in only 45 cases. Of the 40 patients in whom various drugs were used, 15 were taking other agents which can themselves induce SLE-like syndrome. These were: phenytoin (7 patients), primidone (2), ethosuximide (1), promethazine (2), acebutolol (1) and oral contraceptives (2). Outcome: This was assessed as follows: recovery (21 patients), no recovery at time of reporting (21), unknown (38). Cases from the Literature: Of the 80 cases in the database, II had been described in the literature: these are summarised in table VI. Only 6 can be confirmed as DILE; 4 were considered to be idiopathic SLE; and insufficient information was available in I case (abstract only) [Rabinowicz et al. 1990].
3. Discussion 3.1 Case Reports Discussion of the 80 case reports is limited by the lack of full information in most cases. The limitations of this data are: (a) not all cases of adverse
....
Table VI. Published cases of SLE-like syndromes in patients receiving 'Tegretol' [carbamazepine (CBZ) therapy]
VI
0-
Reference
Laboratory findings Duration of therapy Anti-dsDNA before onset ANA
Agel sex
Clinical features
Alballa et al. (1987) Bateman (1985)
30/F
Arthritis, rash
1 year
+
30/F
1 year
+
Beurey et al. (1971)
42/F
Arthralgia, rash, pleuritis, vasculitis, leucopenia Rash, arthralgia
4 years
Drory et al. (1983)
181M
Arthralgia, pleuritis, perlcorditis, rash
Kolstee (1982) Livingston et al. (1974) Maravic et al. (1990)
251M 46/F
McNichol (1990)
Rabinowicz et al. (1990) Simpson (1966) Takigawa et al. (1976)
Antihistone LE cells
Results of stopping CBZ
Remarks
Recovery in 18 months
First fully documented case Grayson (1985) stated this case probably spontaneous SLE not induced by CBZ Patient became worse after discontinuation of treatment of SLE Patient met 6 SLE criteria of AMA. Presence of + anti-DNA against diagnosis of DILE
+ 1640
NA
+ 94
U/ml (normal
