- Email: [email protected]
Panniculitis due to potassium bromide
Brain & Development 20 (1998) 83–87
Original article
Panniculitis due to potassium bromide Wilfried Diener a ,*, Marco Sorni b, Stefan Ruile c, Peter Rude d, Rolf Kruse e, Eberhard Becker f, Konrad Bork b, Peter A. Berg f a
MDK Baden-Wu¨rttemberg, D-77933 Lahr, Germany Department of Dermatology, University of Mainz, D-55101 Mainz, Germany c ICP2, EPFL, CH-1015 Lausanne, Switzerland d Department of Pediatrics, Hospital Siegen, D-57072 Siegen, Germany e Epilepsy Center Kork, D-77694 Kehl-Kork, Germany f Department of Internal Medicine II, University of Tu¨bingen, D-72076 Tu¨bingen, Germany b
Received 26 June 1997; revised version received 10 September 1997; accepted 10 November 1997
Abstract Potassium bromide again is well known to be surprisingly effective in patients with severe myoclonic epilepsy in infants (SME). Rare side effects on the skin reappeared, such as the febrile nodular panniculitis (Weber–Christian syndrome). In 1993 we described the first three cases of necrotizing panniculitis and introduced the term ‘halogen panniculitis’. It is a systemic disease with crops of subcutaneous nodules, fever, elevated sedimentation rate, hepatosplenomegalia, and abdominal pain. Later severe necrosis of the skin and adipose tissue may happen with deep ulcerations. History and course of five cases, described in this paper, suggest either an allergy or toxic reason. Histologic picture shows inflammation of adipose tissues with infiltrating lymphocytes, but lymphocyte transformation test (LTT) was not reliable in diagnosing the disease. Possibly, bromides act as a chemokine and stimulate inflammatory processes. Bromide can be transformed into a bromine radical/free electron pair under UV irradiation at 228.8 nm in aqueous solution. The bromine radical may have detrimental effects on the tissue. However, despite some research, the origin of halogen panniculitis and similar diseases remains unclear. 1998 Elsevier Science B.V. Keywords: Halogen panniculitis; Bromoderma tuberosum; Potassium bromide; Weber–Christian syndrome; Photochemistry of bromide
1. Introduction The element bromine was discovered in 1826 and bromide salts were introduced as an antiepileptic drug by Locock in 1857 [1]. Soon bromides became the most widely used antiepileptic drugs all over the world. The cutaneous side effects became well known such as acne, bromoderma tuberosum, and bromoderma nodosum [2]. Weber described in 1925 a relapsing nonsuppurative nodular panniculitis and Christian added in 1928 the word ‘febrile’ [3,4]. In 1952 56 cases of ‘Weber–Christian syndromes’ had been reported in the literature and some of them were caused by bromides or iodides [5]. After bromides became obsolete as an antiepi* Corresponding author. C. Robert Dold Strasse 2a, D-77654 Offenburg, Germany. Fax: +49 7821 938200.
0387-7604/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0387-7604 (97 )0 0110-1
leptic drug, no more cases of Weber–Christian syndrome due to bromides have been published. The return of potassium bromide as an antiepileptic drug against generalized tonic-clonic seizures is accompanied by the reemergence of the old side effects [6–9] (Table 1). In 1993 we described the first three cases of necrotizing panniculitis in patients with bromide therapy and named the disease ‘halogen panniculitis’ because of similar generally known reactions to iodides [2]. In the last years some more cases occurred and will be described. Weber–Christian syndrome or febrile relapsing nodular panniculitis is a description of some clinical symptoms which are developing because of different etiologic reasons and not because of a defined disease. Therefore, we introduced the term halogen panniculitis in order to separate a distinct entity from others. Halogen panniculitis is a sys-
84
W. Diener et al. / Brain & Development 20 (1998) 83–87
Table 1 Adverse effects of bromides and bromide poisoning Skin Respiratory tract Gastrointestinal tract Nervous system
Acne, dermatitis, bromoderma tuberosum, panniculitis Lacrimation, conjunctivitis, bronchitis Gastric distress, ulcera, colitis Ataxia, tremor, speech defects, sedation, lethargy, coma, psychosis
temic disease with fever, elevated sedimentation rate, myalgia, hepatosplenomegaly, and abdominal pain. Crops of subcutaneous nodules on any part of the body appear and the overlying skin is red, warm and painful. Severe necrosis may happen with deep ulcerations if the drug is not discontinued. The histologic picture shows inflammation and degeneration of adipose tissues with infiltrating lymphocytes and granulocytes. Bromide is an ion without any severe toxicity but may act as a chemokine and possibly stimulates leucocytes in an unspecific or toxic way [12]. Despite several investigations the definite origin of halogen panniculitis and similar diseases remains still unclear.
2. Case reports After the index patient (Case 1) two more children from the Epilepsy center of Kork were identified retrospectively (Cases 2 and 3). Results have been published in Germany in 1993 [2]. Since 1992 side effects during bromide therapy in humans and in dogs were collected at an information center. Most side effects on the skin dealt with acne, unspecific rashes, and bromoderma tuberosum, but some more children developed signs of a nodular panniculitis with fever. In two patients (Cases 4 and 5) we were allowed to collect the clinical data and to demonstrate photographs of the skin lesions.
changed to deep painful infiltrations around the legs and soles without fluctuation, some of them more than 10 cm in diameter. At the same time new nodules appeared on the upper extremities and the trunk. Phenytoin was discontinued without improvement. Tuberculin skin test (5 TU and 10 TU) and X-ray films of the chest were negative. Repeated ultrasonography revealed signs of colitis and an enlarged spleen, but no pancreatitis. Investigations for bacterial, viral, and fungal infections with repeated serologic and stool examinations remained negative. Tests for antinuclear antibodies, rheumatoid factors, immunoglobulins, and immune complexes were negative. No a-1-antitrypsin deficiency, hepatitis, or lymphadenopathy was found. A biopsy revealed the diagnosis of necrotizing panniculitis. Bromide therapy was discontinued, and only a few days later all symptoms improved. After 2 weeks the girl could walk again. The nodules regressed, leaving pigmented depressions of the skin. A lymphocyte transformation test (LTT) to potassium bromide was performed and the stimulation index was elevated above 3. A second LTT several months later indicated a stimulation index to potassium bromide below 2. After discontinuation of potassium bromide the epilepsy became worse, and more and long lasting tonic clonic seizures occurred. During a cluster of generalized tonic-clonic seizures the girl died. After this index-case, the diagnosis of halogen-panniculitis could be established retrospectively in two other patients of the Epilepsy-center Kork. 2.2. Case 2 An 8-year-old mentally retarded girl with a severe generalized epilepsy with generalized tonic clonic seizures of unknown origin. Four months after starting a combined
2.1. Case 1 A handicapped 15-year-old girl with partial duplication of the short arm of chromosome 3. She suffered from severe generalized tonic-clonic seizures and received a combined therapy of bromide and phenytoin (bromide plasma level 1500 mg/l). One month after introduction of potassium bromide, the girl developed abdominal pain, vomiting, and fever. Serum amylase and lipase activities were elevated, and pancreatic ultrasound examination showed an enlarged pancreas of low density. Etiology of the pancreatitis remained unclear. Trauma, viral illness, cholelithiasis, hyperlipemia, and hypercalcemia could be excluded. Symptoms subsided spontaneously after 1 week, but light colored stools without evidence of cholestasis persisted. Three months later the girl again complained about diarrhea, abdominal pain, fever, and cutaneous eruptions similar to erythema nodosum with painful red elevated nodules 0.5–5 cm in diameter in the lower extremities. The nodules
Fig. 1. Painful single cutaneous lesion on the calf, 4 months after starting a combined therapy with potassium bromide and methsuximide (Case 2).
W. Diener et al. / Brain & Developlment 20 (1998) 83–87
Fig. 2. Histology of a cutaneous lesion with a severe inflammatory reaction, predominantly of round cells at ×200 magnification, PAS (Case 2) (Professor Weidner, Department of Dermatology, Bad Cannstatt).
therapy with potassium bromide (plasma bromide level 1650 mg/l) and methsuximide, she became ill with fever, diarrhea, and a painful lesion on her right calf (Fig. 1). Clinically a diagnosis of bromoderma tuberosum was made, but histologically it was a necrotizing panniculitis (Fig. 2). After some weeks, administration of potassium bromide was stopped as an experiment and the skin lesion disappeared rapidly, leaving a pigmented depression.
85
After many frustrating attempts he received a combined therapy with valproate, potassium bromide, and barbiturates. Thereafter the frequency of seizures declined. 10 years later, after the dosage of potassium bromide had been elevated, a disseminated folliculitis developed, and thereafter painful abscesses with deep ulcerations all over the skin (Fig. 3). The child was ill with fever and in a bad condition. Histology showed a nodular panniculitis, and the diagnosis Weber–Christian syndrome was established. Antiepileptic drug therapy was continued without any change because of the severe epilepsy. During the following 8 years the disease became chronic with repeated appearance of deep nodules and painful abscesses. Several surgical interventions became necessary and deep scars developed. The course of the disease was identical in regard with the older descriptions of Weber and Christian. After the first ‘modern’ paper about cases of halogen panniculitis due to potassium bromide the drug was discontinued. Again the boy suffered from a lot of severe seizures and he was in a bad condition. Now the situation has improved, but the frequency of tonic-clonic seizures is still higher than before. The cutaneous lesions have disappeared. 2.5. Case 5 A 10-year-old mentally retarded girl with a severe gen-
2.3. Case 3 A 15-year-old severely retarded boy with a symptomatic focal epilepsy and drug-resistant secondarily generalized tonic-clonic seizures. He had been treated with potassium bromide, phenobarbital, and methsuximide for more than 10 years. Plasma bromide levels were high, as were levels of the barbiturate. He became ill after a slight change in bromide dosage with subcutaneous infiltrations on his buttocks up to 7 cm in diameter. He suffered from high grade fever, diarrhea, and was unable to walk. Extensive laboratory investigations could not explain the symptoms. After discontinuation of potassium bromide the skin lesions improved quickly, leaving again depressions and pigmentations of the skin above the lesions. The boy died some years later after a tonic-clonic seizure of unknown duration. 2.4. Case 4 A patient with a very severe myoclonic epilepsy in infants (SME) since the third month of his life. He suffered from numerous tonic clonic seizures and had been hospitalized for 1 year in an epilepsy center when he was 4-years-old.
Fig. 3. Relapsing deep ulcerations all over the skin in a 12-year-old boy with fever, 10 years after introduction of potassium bromide and soon after elevation of the dosage (Case 4).
86
W. Diener et al. / Brain & Development 20 (1998) 83–87
Fig. 4. Subcutaneous nodules such as erythema nodosum in a 10-year-old child several months after introduction of potassium bromide (Case 5).
eralized epilepsy with generalized tonic-clonic seizures. After therapy with Valproate and potassium bromide, frequency and duration of seizures declined. During slight infections the child developed subcutaneous nodules such as erythema nodosum. Then, a more severe infection happened with high fever. A lot of subcutaneous nodules appeared on the whole skin (Fig. 4). A pneumonia was diagnosed without any improvement during 1 week of antibiotic therapy. Investigations for infections remained negative as did tests for antinuclear antibodies, immune complexes and a-1-antitrypsin deficiency. The diagnosis of halogen panniculitis was established and therapy with potassium bromide discontinued. Without any more treatment, the disease improved soon. Lung disease and skin lesions both disappeared. A LTT was performed, but the stimulation index was not elevated.
3. Discussion Recent literature did not mention new cases of necrotizing panniculitis due to bromides since 1952. To our knowledge Sandford was the last to describe cases of panniculitis due to bromides and iodides [5]. Some decades later, the ancient drug potassium bromide had a come back as an antiepileptic drug and caused a kind of reply of the drug’s special side effects [10]. The old information about panniculitis had been lost, and a lot of laboratory testing was performed before the first new case of panniculitis reached the correct diagnosis. Now the disease is named halogen panniculitis and not any longer
Weber–Christian syndrome, because Weber–Christian syndrome represents no etiologic entity. The clinical symptoms of halogen panniculitis are summarized in Table 2. History and course of the five cases presented above suggests either an allergy or toxic reasons. We hoped to answer the questions about the pathophysiology by performing lymphocyte transformation tests (LTT) [10,11]. But the test was found to be not reliable and in case five there was no stimulation of lymphocytes after exposition to therapeutic concentrations of bromide ions. Furthermore type IV allergy seems to be possible with T cell proliferation. Alegre et al. described some cases of panniculitis and emphasized the role of T-cells in generating the disease, but no case of halogen panniculitis was described [12]. Epilepsy has shown to be associated with different immune abnormalities like increased CNS- and phospholipid antibodies [13]. Some of these changes may promote halogen panniculitis and may be seen in LE panniculitis too. Familial disposition, sunlight and trauma are able to induce panniculitis in SLE also [14]. Distribution of the lesions in LE panniculitis is similar to patients with halogen panniculitis, and reappearance of lesions in scars is well known in both diseases. Pulmonar involvement is possible in LE panniculitis, and may be in halogen panniculitis, too. We are informed about three cases of long-lasting pulmonary infiltrations in patients with potassium bromide (unpublished data). SLE with bullae is an extremely rare disorder, but still possible [15]. In 1941 Gillman and Goodman mentioned the possibility of pemphigus-like bromoderma with vesicles [16]. But antinuclear antibodies were never present and halogen panniculitis is not related to druginduced LE. a-1-Antitrypsin deficiency could be ruled out in every patient. a-1-Antitrypsin is a proteinase inhibitor, and severe deficiency is associated with cutaneous vasculitis due to activated inflammatory mediators such as complement and Table 2 Clinical symptoms of halogen panniculitis Immediately, soon, or many years after introduction of bromides (or iodides) the disease starts, possibly after increase of drug dosage First symptoms may be completely unspecific, such as fever, signs of inflammation, diarrhea, and an unspecific rash Subcutaneous nodules appear and diagnosis of an erythema nodosum is possible. The nodules are painful, red, warm, and elevated with up to 10 cm in diameter. Later, abscesses may develop with deep ulcerations and scarring The patient’s condition may become worse with myalgia, abdominal pain and high grade fever Severe diarrhea with signs of colitis or pancreatitis may develop. (In dogs haemorrhagic forms of colitis occurred, as did some cases of pancreatitis) Enlarged spleen, leucocytosis, leucopenia, and anaemia may be found Histologically, there are cellular infiltrations with vascular involvement, necrosis, and degeneration of adipose tissue After discontinuation of potassium bromide, all symptoms will improve soon but reexposition in iodide or bromide is able to reactivate the disease with new nodules
W. Diener et al. / Brain & Developlment 20 (1998) 83–87
kinine. Some mediators also activate lymphocytes. a-1Antitrypsin modulates migration of neutrophils, chemotaxis, and complement-dependent rosette formation [17]. Possibly bromides act as a chemokine and stimulate inflammatory processes [18]. A theory has been postulated, that elemental bromine might be produced in the body [19]. But it is very doubtful that the human organism can provide the high amount of energy needed to perform the chemical reaction. Only sunlight would be able to provide enough energy to transform bromide ions into elementary bromine. Indeed, most cases of bromoderma tuberosum are located on areas exposed to the sun and bromide ions in the superficial layers of the skin could absorb light energy and react to aggressive elemental bromine. Possibly bromide in surface layers of the skin is transformed photochemically to bromine radicals. These radicals would react immediately with organic tissue. The photochemistry of bromide and the other halides has been studied extensively by Jortner, 1963 [20]. Evidence was given that bromide can be transformed into a bromine radical/free electron pair under UV irradiation at 228.8 nm in aqueous solution. Quantum yields of that process were shown to depend on solution pH, irradiation wavelength, temperature, but not salt concentration. Photoexcitation of bromide in the near-UV and visible regions requires sensitization with dyes (or semiconductors). Such sensitization in aqueous solution will produce an excited sensitizer, which may react with water or OH − ions to produce OH radicals. These radicals might oxidize bromide yielding bromine and hydroxide [21]. This type of reaction has been studied extensively as it is attractive for an application in hydrogen-bromine fuel cells. As sensitizers titanium dioxide, ferric bromide, anthraquinone derivates and phthalic acid were reported [22]. Assuming that the same process occurred in the body, there would be a free electron, which is a reducing agent, and the bromine radical, which may have detrimental effects on the tissue. Panniculitis in children has been reviewed by Schuval [23]. Bromides are no longer mentioned as a possible reason. We would like to add halogen panniculitis again into diagnostic considerations. However, despite some research, the origin of halogen panniculitis and similar diseases remains unclear.
References [1] Locock C. Discussion to Sieveking E.H. Analysis of fifty-two cases of epilepsy observed by the author. Lancet 1857;1:614–618.
87
[2] Diener W, Kruse R, Berg P. Halogenpanniculitis auf Kaliumbromid. Monatsschr Kinderheilkd 1993;141:705–707. [3] Weber FP. Relapsing non-suppurative nodular panniculitis, showing phagozytosisn of subcutaneous fat cells by macrophages. Br J Dermatol 1925;37:301. [4] Christian HA. Relapsing febrile nodular nonsuppurative panniculitis. Arch Intern Med 1928;42:351–388. [5] Sanford HN, Eubank DF, Stenn F. Chronic panniculitis with leucopenia (Weber–Christian Syndrome). Am J Dis Child 1952;83:156– 163. [6] Boenigk HE, Lorenz H, Jurgens U. Aktuelle Erfahrungen mit Bromiden zur Behandlung generalisierter Epilepsien. In: Kruse R, editor. Epilepsie 84 Reinbeck: Einhorn Verlag, 1984:316–325. [7] Ernst JP, Doose H, Baier WK. Bromides were effective in intractable epilepsy with generalized tonic-clonic seizures and onset in early childhood. Brain Dev 1988;10:385–388. [8] Steinhoff BJ, Kruse R. Bromide treatment of pharmaco-resistant epilepsies: a clinical study. Brain Dev 1992;14:144–149. [9] Oguni H, Hayashi K, Oguni M, Mukahira A, Uehara T, Fukuyama Y, et al. Treatment of severe myoclonic epilepsy in infants with bromide and its borderline variant. Epilepsia 1994;35:1140–1145. [10] Diener W, editor. Das a¨lteste moderne Antiepileptikum: Kaliumbromid. Offenburg: Davidis Science, 1997. [11] Berg PA, Becker W, Holzschuh J, Brattig N, Daniel PT. Lymphozyten-Transformations-Test fu¨r die Diagnose der medikamento¨sen ¨ rztebl 1988;42:2917–2922. Allergie. Dt A [12] Alegre VA, Winkelmann RK. Histiocytic cytophagic panniculitis. J Am Acad Dermatol 1989;20:177–185. [13] Diener W, Klein R, Kast K, Mohrmann M, Berg AP, Dambinova S, et al. Increased CNS-and Phospholipid Antibodies in epileptic syndromes of childhood. 11th ESN Meeting, Groningen, The Netherlands, 1996. [14] Winkelmann RK. Panniculitis in connective tissue disease. Arch Dermatol 1983;119:336–344. [15] Kind P, Schuppe HC, Goerz G. Kutaner Lupus erythematodes. Dt ¨ rztebl 1992;89:1306–1311. A [16] Goodman L, Gilman, A, editors. The pharmaceutical basis of therapeutics. New York: Macmillan, 1941:155–165. [17] Smith KS, Pittelkow MR, Su D. Panniculitis associated with severe a-1-antitrypsin deficiency. Arch Dermatol 1987;123:1655–1661. [18] Steele WO, Woody RC, Steele WO. Enhanced neutrophil function in children on bromide therapy. Am J Med Sci 1991;302:145–147. [19] Pfeifle J, Grieben U, Bork K. Bromoderma tuberosum durch antikonvulsive Behandlung mit Kaliumbromid. Hautarzt 1992;302:792– 794. [20] Jortner J, Ottolenghi M, Stein G. On the photochemistry of aqueous solutions of chloride, bromide, and iodide ions. J Phys Chem 1963;68:247–255. [21] Wagner I, Strehlow H. On the flash photolysis of bromide ions in aqueous solutions. Ber Bunsenges 1987;91:1317. [22] Gra¨tzel M, Halmann M. Photosensitized oxidation of bromide to bromine with phthalic acid derivatives in aqueous solutions. Solar Energy 1990;20:117–129. [23] Schuval SJ, Frances A, Valderrama E, Bonagura R, Ilowite N. Panniculitis and fever in children. J Pediatr 1993;122:372–378.
Original article
Panniculitis due to potassium bromide Wilfried Diener a ,*, Marco Sorni b, Stefan Ruile c, Peter Rude d, Rolf Kruse e, Eberhard Becker f, Konrad Bork b, Peter A. Berg f a
MDK Baden-Wu¨rttemberg, D-77933 Lahr, Germany Department of Dermatology, University of Mainz, D-55101 Mainz, Germany c ICP2, EPFL, CH-1015 Lausanne, Switzerland d Department of Pediatrics, Hospital Siegen, D-57072 Siegen, Germany e Epilepsy Center Kork, D-77694 Kehl-Kork, Germany f Department of Internal Medicine II, University of Tu¨bingen, D-72076 Tu¨bingen, Germany b
Received 26 June 1997; revised version received 10 September 1997; accepted 10 November 1997
Abstract Potassium bromide again is well known to be surprisingly effective in patients with severe myoclonic epilepsy in infants (SME). Rare side effects on the skin reappeared, such as the febrile nodular panniculitis (Weber–Christian syndrome). In 1993 we described the first three cases of necrotizing panniculitis and introduced the term ‘halogen panniculitis’. It is a systemic disease with crops of subcutaneous nodules, fever, elevated sedimentation rate, hepatosplenomegalia, and abdominal pain. Later severe necrosis of the skin and adipose tissue may happen with deep ulcerations. History and course of five cases, described in this paper, suggest either an allergy or toxic reason. Histologic picture shows inflammation of adipose tissues with infiltrating lymphocytes, but lymphocyte transformation test (LTT) was not reliable in diagnosing the disease. Possibly, bromides act as a chemokine and stimulate inflammatory processes. Bromide can be transformed into a bromine radical/free electron pair under UV irradiation at 228.8 nm in aqueous solution. The bromine radical may have detrimental effects on the tissue. However, despite some research, the origin of halogen panniculitis and similar diseases remains unclear. 1998 Elsevier Science B.V. Keywords: Halogen panniculitis; Bromoderma tuberosum; Potassium bromide; Weber–Christian syndrome; Photochemistry of bromide
1. Introduction The element bromine was discovered in 1826 and bromide salts were introduced as an antiepileptic drug by Locock in 1857 [1]. Soon bromides became the most widely used antiepileptic drugs all over the world. The cutaneous side effects became well known such as acne, bromoderma tuberosum, and bromoderma nodosum [2]. Weber described in 1925 a relapsing nonsuppurative nodular panniculitis and Christian added in 1928 the word ‘febrile’ [3,4]. In 1952 56 cases of ‘Weber–Christian syndromes’ had been reported in the literature and some of them were caused by bromides or iodides [5]. After bromides became obsolete as an antiepi* Corresponding author. C. Robert Dold Strasse 2a, D-77654 Offenburg, Germany. Fax: +49 7821 938200.
0387-7604/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0387-7604 (97 )0 0110-1
leptic drug, no more cases of Weber–Christian syndrome due to bromides have been published. The return of potassium bromide as an antiepileptic drug against generalized tonic-clonic seizures is accompanied by the reemergence of the old side effects [6–9] (Table 1). In 1993 we described the first three cases of necrotizing panniculitis in patients with bromide therapy and named the disease ‘halogen panniculitis’ because of similar generally known reactions to iodides [2]. In the last years some more cases occurred and will be described. Weber–Christian syndrome or febrile relapsing nodular panniculitis is a description of some clinical symptoms which are developing because of different etiologic reasons and not because of a defined disease. Therefore, we introduced the term halogen panniculitis in order to separate a distinct entity from others. Halogen panniculitis is a sys-
84
W. Diener et al. / Brain & Development 20 (1998) 83–87
Table 1 Adverse effects of bromides and bromide poisoning Skin Respiratory tract Gastrointestinal tract Nervous system
Acne, dermatitis, bromoderma tuberosum, panniculitis Lacrimation, conjunctivitis, bronchitis Gastric distress, ulcera, colitis Ataxia, tremor, speech defects, sedation, lethargy, coma, psychosis
temic disease with fever, elevated sedimentation rate, myalgia, hepatosplenomegaly, and abdominal pain. Crops of subcutaneous nodules on any part of the body appear and the overlying skin is red, warm and painful. Severe necrosis may happen with deep ulcerations if the drug is not discontinued. The histologic picture shows inflammation and degeneration of adipose tissues with infiltrating lymphocytes and granulocytes. Bromide is an ion without any severe toxicity but may act as a chemokine and possibly stimulates leucocytes in an unspecific or toxic way [12]. Despite several investigations the definite origin of halogen panniculitis and similar diseases remains still unclear.
2. Case reports After the index patient (Case 1) two more children from the Epilepsy center of Kork were identified retrospectively (Cases 2 and 3). Results have been published in Germany in 1993 [2]. Since 1992 side effects during bromide therapy in humans and in dogs were collected at an information center. Most side effects on the skin dealt with acne, unspecific rashes, and bromoderma tuberosum, but some more children developed signs of a nodular panniculitis with fever. In two patients (Cases 4 and 5) we were allowed to collect the clinical data and to demonstrate photographs of the skin lesions.
changed to deep painful infiltrations around the legs and soles without fluctuation, some of them more than 10 cm in diameter. At the same time new nodules appeared on the upper extremities and the trunk. Phenytoin was discontinued without improvement. Tuberculin skin test (5 TU and 10 TU) and X-ray films of the chest were negative. Repeated ultrasonography revealed signs of colitis and an enlarged spleen, but no pancreatitis. Investigations for bacterial, viral, and fungal infections with repeated serologic and stool examinations remained negative. Tests for antinuclear antibodies, rheumatoid factors, immunoglobulins, and immune complexes were negative. No a-1-antitrypsin deficiency, hepatitis, or lymphadenopathy was found. A biopsy revealed the diagnosis of necrotizing panniculitis. Bromide therapy was discontinued, and only a few days later all symptoms improved. After 2 weeks the girl could walk again. The nodules regressed, leaving pigmented depressions of the skin. A lymphocyte transformation test (LTT) to potassium bromide was performed and the stimulation index was elevated above 3. A second LTT several months later indicated a stimulation index to potassium bromide below 2. After discontinuation of potassium bromide the epilepsy became worse, and more and long lasting tonic clonic seizures occurred. During a cluster of generalized tonic-clonic seizures the girl died. After this index-case, the diagnosis of halogen-panniculitis could be established retrospectively in two other patients of the Epilepsy-center Kork. 2.2. Case 2 An 8-year-old mentally retarded girl with a severe generalized epilepsy with generalized tonic clonic seizures of unknown origin. Four months after starting a combined
2.1. Case 1 A handicapped 15-year-old girl with partial duplication of the short arm of chromosome 3. She suffered from severe generalized tonic-clonic seizures and received a combined therapy of bromide and phenytoin (bromide plasma level 1500 mg/l). One month after introduction of potassium bromide, the girl developed abdominal pain, vomiting, and fever. Serum amylase and lipase activities were elevated, and pancreatic ultrasound examination showed an enlarged pancreas of low density. Etiology of the pancreatitis remained unclear. Trauma, viral illness, cholelithiasis, hyperlipemia, and hypercalcemia could be excluded. Symptoms subsided spontaneously after 1 week, but light colored stools without evidence of cholestasis persisted. Three months later the girl again complained about diarrhea, abdominal pain, fever, and cutaneous eruptions similar to erythema nodosum with painful red elevated nodules 0.5–5 cm in diameter in the lower extremities. The nodules
Fig. 1. Painful single cutaneous lesion on the calf, 4 months after starting a combined therapy with potassium bromide and methsuximide (Case 2).
W. Diener et al. / Brain & Developlment 20 (1998) 83–87
Fig. 2. Histology of a cutaneous lesion with a severe inflammatory reaction, predominantly of round cells at ×200 magnification, PAS (Case 2) (Professor Weidner, Department of Dermatology, Bad Cannstatt).
therapy with potassium bromide (plasma bromide level 1650 mg/l) and methsuximide, she became ill with fever, diarrhea, and a painful lesion on her right calf (Fig. 1). Clinically a diagnosis of bromoderma tuberosum was made, but histologically it was a necrotizing panniculitis (Fig. 2). After some weeks, administration of potassium bromide was stopped as an experiment and the skin lesion disappeared rapidly, leaving a pigmented depression.
85
After many frustrating attempts he received a combined therapy with valproate, potassium bromide, and barbiturates. Thereafter the frequency of seizures declined. 10 years later, after the dosage of potassium bromide had been elevated, a disseminated folliculitis developed, and thereafter painful abscesses with deep ulcerations all over the skin (Fig. 3). The child was ill with fever and in a bad condition. Histology showed a nodular panniculitis, and the diagnosis Weber–Christian syndrome was established. Antiepileptic drug therapy was continued without any change because of the severe epilepsy. During the following 8 years the disease became chronic with repeated appearance of deep nodules and painful abscesses. Several surgical interventions became necessary and deep scars developed. The course of the disease was identical in regard with the older descriptions of Weber and Christian. After the first ‘modern’ paper about cases of halogen panniculitis due to potassium bromide the drug was discontinued. Again the boy suffered from a lot of severe seizures and he was in a bad condition. Now the situation has improved, but the frequency of tonic-clonic seizures is still higher than before. The cutaneous lesions have disappeared. 2.5. Case 5 A 10-year-old mentally retarded girl with a severe gen-
2.3. Case 3 A 15-year-old severely retarded boy with a symptomatic focal epilepsy and drug-resistant secondarily generalized tonic-clonic seizures. He had been treated with potassium bromide, phenobarbital, and methsuximide for more than 10 years. Plasma bromide levels were high, as were levels of the barbiturate. He became ill after a slight change in bromide dosage with subcutaneous infiltrations on his buttocks up to 7 cm in diameter. He suffered from high grade fever, diarrhea, and was unable to walk. Extensive laboratory investigations could not explain the symptoms. After discontinuation of potassium bromide the skin lesions improved quickly, leaving again depressions and pigmentations of the skin above the lesions. The boy died some years later after a tonic-clonic seizure of unknown duration. 2.4. Case 4 A patient with a very severe myoclonic epilepsy in infants (SME) since the third month of his life. He suffered from numerous tonic clonic seizures and had been hospitalized for 1 year in an epilepsy center when he was 4-years-old.
Fig. 3. Relapsing deep ulcerations all over the skin in a 12-year-old boy with fever, 10 years after introduction of potassium bromide and soon after elevation of the dosage (Case 4).
86
W. Diener et al. / Brain & Development 20 (1998) 83–87
Fig. 4. Subcutaneous nodules such as erythema nodosum in a 10-year-old child several months after introduction of potassium bromide (Case 5).
eralized epilepsy with generalized tonic-clonic seizures. After therapy with Valproate and potassium bromide, frequency and duration of seizures declined. During slight infections the child developed subcutaneous nodules such as erythema nodosum. Then, a more severe infection happened with high fever. A lot of subcutaneous nodules appeared on the whole skin (Fig. 4). A pneumonia was diagnosed without any improvement during 1 week of antibiotic therapy. Investigations for infections remained negative as did tests for antinuclear antibodies, immune complexes and a-1-antitrypsin deficiency. The diagnosis of halogen panniculitis was established and therapy with potassium bromide discontinued. Without any more treatment, the disease improved soon. Lung disease and skin lesions both disappeared. A LTT was performed, but the stimulation index was not elevated.
3. Discussion Recent literature did not mention new cases of necrotizing panniculitis due to bromides since 1952. To our knowledge Sandford was the last to describe cases of panniculitis due to bromides and iodides [5]. Some decades later, the ancient drug potassium bromide had a come back as an antiepileptic drug and caused a kind of reply of the drug’s special side effects [10]. The old information about panniculitis had been lost, and a lot of laboratory testing was performed before the first new case of panniculitis reached the correct diagnosis. Now the disease is named halogen panniculitis and not any longer
Weber–Christian syndrome, because Weber–Christian syndrome represents no etiologic entity. The clinical symptoms of halogen panniculitis are summarized in Table 2. History and course of the five cases presented above suggests either an allergy or toxic reasons. We hoped to answer the questions about the pathophysiology by performing lymphocyte transformation tests (LTT) [10,11]. But the test was found to be not reliable and in case five there was no stimulation of lymphocytes after exposition to therapeutic concentrations of bromide ions. Furthermore type IV allergy seems to be possible with T cell proliferation. Alegre et al. described some cases of panniculitis and emphasized the role of T-cells in generating the disease, but no case of halogen panniculitis was described [12]. Epilepsy has shown to be associated with different immune abnormalities like increased CNS- and phospholipid antibodies [13]. Some of these changes may promote halogen panniculitis and may be seen in LE panniculitis too. Familial disposition, sunlight and trauma are able to induce panniculitis in SLE also [14]. Distribution of the lesions in LE panniculitis is similar to patients with halogen panniculitis, and reappearance of lesions in scars is well known in both diseases. Pulmonar involvement is possible in LE panniculitis, and may be in halogen panniculitis, too. We are informed about three cases of long-lasting pulmonary infiltrations in patients with potassium bromide (unpublished data). SLE with bullae is an extremely rare disorder, but still possible [15]. In 1941 Gillman and Goodman mentioned the possibility of pemphigus-like bromoderma with vesicles [16]. But antinuclear antibodies were never present and halogen panniculitis is not related to druginduced LE. a-1-Antitrypsin deficiency could be ruled out in every patient. a-1-Antitrypsin is a proteinase inhibitor, and severe deficiency is associated with cutaneous vasculitis due to activated inflammatory mediators such as complement and Table 2 Clinical symptoms of halogen panniculitis Immediately, soon, or many years after introduction of bromides (or iodides) the disease starts, possibly after increase of drug dosage First symptoms may be completely unspecific, such as fever, signs of inflammation, diarrhea, and an unspecific rash Subcutaneous nodules appear and diagnosis of an erythema nodosum is possible. The nodules are painful, red, warm, and elevated with up to 10 cm in diameter. Later, abscesses may develop with deep ulcerations and scarring The patient’s condition may become worse with myalgia, abdominal pain and high grade fever Severe diarrhea with signs of colitis or pancreatitis may develop. (In dogs haemorrhagic forms of colitis occurred, as did some cases of pancreatitis) Enlarged spleen, leucocytosis, leucopenia, and anaemia may be found Histologically, there are cellular infiltrations with vascular involvement, necrosis, and degeneration of adipose tissue After discontinuation of potassium bromide, all symptoms will improve soon but reexposition in iodide or bromide is able to reactivate the disease with new nodules
W. Diener et al. / Brain & Developlment 20 (1998) 83–87
kinine. Some mediators also activate lymphocytes. a-1Antitrypsin modulates migration of neutrophils, chemotaxis, and complement-dependent rosette formation [17]. Possibly bromides act as a chemokine and stimulate inflammatory processes [18]. A theory has been postulated, that elemental bromine might be produced in the body [19]. But it is very doubtful that the human organism can provide the high amount of energy needed to perform the chemical reaction. Only sunlight would be able to provide enough energy to transform bromide ions into elementary bromine. Indeed, most cases of bromoderma tuberosum are located on areas exposed to the sun and bromide ions in the superficial layers of the skin could absorb light energy and react to aggressive elemental bromine. Possibly bromide in surface layers of the skin is transformed photochemically to bromine radicals. These radicals would react immediately with organic tissue. The photochemistry of bromide and the other halides has been studied extensively by Jortner, 1963 [20]. Evidence was given that bromide can be transformed into a bromine radical/free electron pair under UV irradiation at 228.8 nm in aqueous solution. Quantum yields of that process were shown to depend on solution pH, irradiation wavelength, temperature, but not salt concentration. Photoexcitation of bromide in the near-UV and visible regions requires sensitization with dyes (or semiconductors). Such sensitization in aqueous solution will produce an excited sensitizer, which may react with water or OH − ions to produce OH radicals. These radicals might oxidize bromide yielding bromine and hydroxide [21]. This type of reaction has been studied extensively as it is attractive for an application in hydrogen-bromine fuel cells. As sensitizers titanium dioxide, ferric bromide, anthraquinone derivates and phthalic acid were reported [22]. Assuming that the same process occurred in the body, there would be a free electron, which is a reducing agent, and the bromine radical, which may have detrimental effects on the tissue. Panniculitis in children has been reviewed by Schuval [23]. Bromides are no longer mentioned as a possible reason. We would like to add halogen panniculitis again into diagnostic considerations. However, despite some research, the origin of halogen panniculitis and similar diseases remains unclear.
References [1] Locock C. Discussion to Sieveking E.H. Analysis of fifty-two cases of epilepsy observed by the author. Lancet 1857;1:614–618.
87
[2] Diener W, Kruse R, Berg P. Halogenpanniculitis auf Kaliumbromid. Monatsschr Kinderheilkd 1993;141:705–707. [3] Weber FP. Relapsing non-suppurative nodular panniculitis, showing phagozytosisn of subcutaneous fat cells by macrophages. Br J Dermatol 1925;37:301. [4] Christian HA. Relapsing febrile nodular nonsuppurative panniculitis. Arch Intern Med 1928;42:351–388. [5] Sanford HN, Eubank DF, Stenn F. Chronic panniculitis with leucopenia (Weber–Christian Syndrome). Am J Dis Child 1952;83:156– 163. [6] Boenigk HE, Lorenz H, Jurgens U. Aktuelle Erfahrungen mit Bromiden zur Behandlung generalisierter Epilepsien. In: Kruse R, editor. Epilepsie 84 Reinbeck: Einhorn Verlag, 1984:316–325. [7] Ernst JP, Doose H, Baier WK. Bromides were effective in intractable epilepsy with generalized tonic-clonic seizures and onset in early childhood. Brain Dev 1988;10:385–388. [8] Steinhoff BJ, Kruse R. Bromide treatment of pharmaco-resistant epilepsies: a clinical study. Brain Dev 1992;14:144–149. [9] Oguni H, Hayashi K, Oguni M, Mukahira A, Uehara T, Fukuyama Y, et al. Treatment of severe myoclonic epilepsy in infants with bromide and its borderline variant. Epilepsia 1994;35:1140–1145. [10] Diener W, editor. Das a¨lteste moderne Antiepileptikum: Kaliumbromid. Offenburg: Davidis Science, 1997. [11] Berg PA, Becker W, Holzschuh J, Brattig N, Daniel PT. Lymphozyten-Transformations-Test fu¨r die Diagnose der medikamento¨sen ¨ rztebl 1988;42:2917–2922. Allergie. Dt A [12] Alegre VA, Winkelmann RK. Histiocytic cytophagic panniculitis. J Am Acad Dermatol 1989;20:177–185. [13] Diener W, Klein R, Kast K, Mohrmann M, Berg AP, Dambinova S, et al. Increased CNS-and Phospholipid Antibodies in epileptic syndromes of childhood. 11th ESN Meeting, Groningen, The Netherlands, 1996. [14] Winkelmann RK. Panniculitis in connective tissue disease. Arch Dermatol 1983;119:336–344. [15] Kind P, Schuppe HC, Goerz G. Kutaner Lupus erythematodes. Dt ¨ rztebl 1992;89:1306–1311. A [16] Goodman L, Gilman, A, editors. The pharmaceutical basis of therapeutics. New York: Macmillan, 1941:155–165. [17] Smith KS, Pittelkow MR, Su D. Panniculitis associated with severe a-1-antitrypsin deficiency. Arch Dermatol 1987;123:1655–1661. [18] Steele WO, Woody RC, Steele WO. Enhanced neutrophil function in children on bromide therapy. Am J Med Sci 1991;302:145–147. [19] Pfeifle J, Grieben U, Bork K. Bromoderma tuberosum durch antikonvulsive Behandlung mit Kaliumbromid. Hautarzt 1992;302:792– 794. [20] Jortner J, Ottolenghi M, Stein G. On the photochemistry of aqueous solutions of chloride, bromide, and iodide ions. J Phys Chem 1963;68:247–255. [21] Wagner I, Strehlow H. On the flash photolysis of bromide ions in aqueous solutions. Ber Bunsenges 1987;91:1317. [22] Gra¨tzel M, Halmann M. Photosensitized oxidation of bromide to bromine with phthalic acid derivatives in aqueous solutions. Solar Energy 1990;20:117–129. [23] Schuval SJ, Frances A, Valderrama E, Bonagura R, Ilowite N. Panniculitis and fever in children. J Pediatr 1993;122:372–378.