1443
according to the International Headache Society Classification of Headache, 1988.6 The criteria for analgesic abuse were those of Diener, including onset of secondary chronic daily headache.7 The patients were asked to discontinue analgesics or ergotamine abruptly. They were informed orally and in writing about the phenomenon of medication-induced headache, and were encouraged to achieve withdrawal. A tricyclic antidepressant (amitriptyline, 10 mg) was given at night and a non-steroidal anti-inflammatory (naproxen 500 mg) was given orally for treatment of headache symptoms. Patients kept headache diaries, noting frequency and intensity of headaches and analgesic intake during the 6-month follow-up period. They were invited to follow-up visits 6 weeks, 12 weeks, and 6 months after withdrawal of medication, and all but 1 attended. The mean age of the group was 41-6 (range 22-56) years and the mean duration of headache history 23-7 (6-35) years. Most of the 29 patients taking drugs containing no ergotamine were heavy abusers; the weekly analgesic intake averaged 28-4 tablets (range 16-48). 10 were taking simple analgesics only (paracetamol or aspirin) and 19 were taking medications containing caffeine and/or codeine (’Propain’,
’Solpadeine’, ’Syndol’). 17 patients were abusing ergotamine; most were taking 2 mg/day regularly (mean intake 10.9
mg/week;
range 10-12
mg/week.
All 17
were
taking preparations containing caffeine (MigriF, ’Cafergot’). 11 were taking only these drugs, and 6 were taking analgesics as well. Ergotamine had been used regularly for 2 years on average (range 6 months to 20 years), and analgesics for even longer, although not on a regular basis. Abuse started on average a year after use of the medications began, but the range was very wide (2 months to
10
years).
The only withdrawal symptoms reported were increased headache intensity and nausea and vomiting. These symptoms were most severe on days 2-5, although 4 patients had no headache after drug withdrawal. At the end of the study, 37 patients had experienced relief from chronic daily headache; symptoms had reverted to those of the original occasional migraine attacks. Analgesic intake had been reduced to appropriate intermittent use for these attacks. In 6, headache severity had not changed, despite reduced analgesic intake. 2 patients had not reduced analgesic intake and showed no relief of headache. 1 patient who was taking ergotamine and caffeine did not return after the first visit at which withdrawal was proposed, and no follow-up information is available.
Although patients with medication abuse are said to be difficult to treat and many require hospital admission, we were able to treat 45 of a group of 46 migraine outpatients by abrupt withdrawal with good outcome in the majority. With written and spoken explanation, awareness of the facts, encouragement, the use of a non-steroidal antiinflammatory drug for relief of symptoms, a small dose of a tricyclic antidepressant as a prophylactic,8 and close followup, withdrawal was achieved without hospital admission or great interference in everyday life. As a specialist clinic, we see a large number of patients who have migraine with aura. All but 3 of the study subjects suffered from migraine without aura. Patients with aura are known to have less frequent attacks on average than those without, and may therefore be expected to use fewer analgesics, and be less liable to start abusing medication. Patients without aura may be less certain about the start of attacks and often treat less specific symptoms. Many of our patients embarked on the road to abuse by starting to take the medications in anticipation of a full-blown
development of abuse was apparently of independent the time over which the drugs had been taken, since the time before abuse ranged from as little as 2 months to up to 10 years after the medication was started. 17 of our 46 patients abused ergotamine with caffeine, but a surprisingly large number (10) abused only paracetamol or aspirin, which have been considered "safe". This study shows that such simple analgesics can be the cause of intoxication headaches, which accords with previous findings.9,lo Recognition of this possibility is obviously important, since management appears easy and readily effective if there is no concomitant complicating use of other drugs such as benzodiazepines or barbiturates (more common in North America). Most importantly, this study shows that drug withdrawal in analgesic and ergotamine abuse headaches can be achieved in outpatients. The drugs do not need to be gradually withdrawn-abrupt withdrawal is well tolerated, provided that adequate explanation and support are given. attack.
The
REFERENCES 1. Edmeads J. Analgesic-induced headache: an unrecognized epidemic. Headache 1990; 30: 614-15. 2. Tfelt-Hansen P, Krabbe EA. Ergotamine abuse. Do patients benefit from withdrawal? Cephalalgia 1981; 1: 29-32. 3. Mathew NT. Amelioration of ergotamine withdrawal symptoms with naproxen. Headache 1987; 27: 130-33. 4. Baumgartner C, Wessely P, Bigol C, Maly J, Holzner F. Longterm prognosis of analgesic withdrawal in patients with drug-induced headaches. Headache 1989; 29: 510-14. 5. Bernstein AL. More on ergotamine withdrawal. Headache 1987; 27: 458. 6. Headache Classification Committee of the International Headache Society. Classification and diagnosis criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988; 8 (suppl 7): 1-96. 7. Diener HC. Klinik des Analgetikakopfschmerzes. Deutsch med Wschr 1988; 113: 472-74. 8. Holland JV, Holland CV, Kudrow L. Low-dose amitriptyline prophylaxis in chronic scalp muscle contraction headache. Proceedings 1st International Headache Congress, Munich, 1983: 134. 9. Rapoport AM, Weeks RE, Sheftell FD, Baskin SM, Verdi J. Analgesic rebound headache: theoretical and practical implications. Cephalalgia 1985; 5 (suppl 3): 448-49. 10. MacGregor EA, Vohrah C, Wilkinson M. Analgesic use: a study of treatments used by patients for migraine prior to attending the City of London migraine clinic. Headache 1990; 30: 571-74.
ADDRESSES The Princess Margaret Migraine Clinic, Charing Cross Hospital, London W6 8RF, UK (R. Hering, MD, T J. Steiner,
PhD). Correspondence to
Dr R.
Hering
Glutathione peroxidase deficiency and childhood seizures
with intractable seizures, repeated and intolerance to anticonvulsants had infections, evidence of glutathione peroxidase deficiency. 2 had low intracellular enzyme activity but normal blood selenium and high plasma glutathione peroxidase concentrations. The other 2 had low intracellular glutathione peroxidase activity with low circulating glutathione peroxidase and selenium concentrations. The clinical state of the children improved after discontinuation of anticonvulsant medication and selenium substitution. 4 children
1444
Time-course of
neutrophil oxygen burst after phorbol-12myristate-13-acetate stimulation in 1 patient before (————W) and after (8 - - -8) selenium substitution, in healthy adults (0 - - - 0), and in healthy infants (x).
Glutathione peroxidase deficiency is chronic granulomatous disease1 and of
a
known
cause
of
idiosyncratic drug reactionsparticularly acute pancreatitis and increased serum lipid peroxides after therapy with valproic acid. Peroxidation stress, based on increased generation of active oxygen metabolites or reduced activity of antioxidative protection mechanisms, may increase the risk of recurrent seizures. We describe reduced glutathione peroxidase activity in 4 children with intractable seizures, repeated infections, and intolerance to anticonvulsant treatment. Concentrations of glutathione peroxidase and glutathione reductase were measured as previously described,4.5 and circulating selenium was measured by atomic absorption spectroscopy. The time-course of the neutrophil oxygen burst was measured as the rate of cytochrome C reduction by absorbance reading at 550 nm against a reference that contained superoxide dismutase, as described by Baker and Cohen.6 When an evaluable rate was detected, catalase was added to the cuvettes to distinguish between reoxidation of cytochrome C by accumulated H202 and inactivation of NADPH oxidase. All 4 patients had seizures during the first 6 months of life. Although the clinical pattern varied, electroencephalograms all showed multifocal activity and computed tomography and magnetic resonance imaging showed progressive cerebral atrophy. There was no evidence of inborn errors of metabolism (normal copper, caeruloplasmin, lactate, and pyruvate in serum and cerebrospinal fluid; normal oligosaccharides, mucopolysaccharides, and sialic acid in urine; normal aminoacids in plasma and urine; and normal lysosomal enzyme activities) or of central-nervous-system infections (no evidence of serum HIV antigen; negative for CMV and EBV antibodies in serum and cerebrospinal fluid; and normal immunoglobulins in serum) that are known to cause recurrent seizures. The karyotypes of all 4 patients were normal. 2 patients had low intracellular glutathione peroxidase
(12-7 and 20-02 IU/g haemoglobin; normal range 31 °Cr394) with normal or nearly normal whole-blood selenium (8-4 and 10-4 Ilg/d1, respectively; normal range 10-3-19-1) and high plasma glutathione peroxidase activity (55-09 and 50-44 IU/dI, respectively; normal range 32.447.6); their parents had reduced red-blood-cell glutathione peroxidase activity. The other 2 patients had low red-bloodcell glutathione peroxidase activity (13-50 and 8-25 IU/g haemoglobin) and low plasma enzyme activity (18-99 and 24-82 IV/d1, respectively), and low whole-blood selenium (5-1 and 3-4 u.g/dl, respectively); only the fathers had reduced red-blood-cell glutathione peroxidase activity, but glutathione reductase activity was low in all parents. concentrations
Intracellular glutathione reductase activity was in the high normal or above the normal range in all 4 patients (6-5-9 89 IU/g haemoglobin; normal range 5-4-7-4). Discontinuation of anticonvulsant therapy and supplementation of selenium, the metal in the active centre of glutathione peroxidase, led to clinical improvement in all 4 patients. The neutrophil oxygen burst has a duration of about 20 min in healthy children and adults. In the neutrophils of our 4 patients, the initial rate of superoxide generation was above normal for children of that age, which may reflect a primed state after repeated infections, but stopped after about 10 min. This shortened activity may be attributable to H20z-mediated damage to NADPH oxidase because of the lack of protection by glutathione peroxidase6 (see figure). 1 of our patients had a low overall rate of superoxide generation and was also deficient in glucose-6-phosphate dehydrogenase (0-04 U /1010 leucocytes, normal range 0-08-0-12, 4 U /1010 erythrocytes, normal range 18-34-8). Glucose-6-phosphate dehydrogenase regenerates NADPH, the cofactor for the free-radical generating enzyme NADPH oxidase. Muscle biopsies from all 4 patients showed increased interfascicular fibrous tissue and sarcolemmal accumulation of lipids. 1 patient died at 26 months; necropsy findings indicated white-matter disease with reduced brain weight, gliosis in the putamen, numerous Alzheimer type II astrocytes prominent in the globus pallidus, and Bergmann astrocytosis in the cerebellum; neurons were generally well
preserved. Our findings indicate that glutathione peroxidase deficiency may be a cause of childhood seizures. Our first 2 patients appear to have a primary deficiency of the intracellular enzyme, whereas the other 2 may have a disturbance of selenium resorption or transport. Enzyme defects which affect neutrophil-mediated immune defence, as well as the central nervous system, may need to be excluded in children with seizures, especially if combined with repeated infections and intolerance to anticonvulsants. This investigation was supported in part by a grant (to G. F. W.) from the Deutsche Forschungsgemeinschaft.
REFERENCES 1. Holmes B, Parks BH, Malawista SE, Quie PG, Nelson DL, Good RA. Chronic granulomatous disease in females. A deficiency of leukocyte glutathione peroxidase. N Engl J Med 1970; 283: 217-21. 2. Pippenger CE, Meng X, van Lente F. Alternative approaches to the prediction of antiepileptic, idiosyncratic, or drug-drug interactions. In: Pitlick WH, ed. Antiepileptic drug interactions. New York: Demos, 1989: 293-307. 3. Wilmore LJ, Rubin JJ. Antiperoxidant pretreatment and iron-induced epileptiform discharges in the rat: EEG and histopathologic studies. Neurology 1981; 31: 63-69. 4. Paglia DE, Valentine WN. Studies on the quantitative and qualitative characterization of glutathione peroxidase. J Lab Clin Med 1967; 70: 158-69. 5. Carlberg I, Mannervik B. Glutathione reductase. In: Meister A, ed. Methods in enzymology, vol 113. Orlando: Academic Press, 1985: 484-90. 6. Baker SS, Cohen HJ. Altered oxidative metabolism in selenium-deficient rat granulocytes. J Immunol 1983; 130: 2856—60.
ADDRESSES Department of Biological Chemistry and Molecular Pharmacology and Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115 (G F Weber, DM); Department of Neurology, University of South Alabama Medical Center, Mobile, AL 36617 (P Maertens, MD); and Department of Biochemistry, Cleveland Clinic Foundation, Cleveland, OH 44195, USA (X Meng, PhD, C E Pippenger, MD) Correspondence to Dr G. F. Weber, Department of Biological Chemistry and Molecular Pharmacology, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115, USA.
according to the International Headache Society Classification of Headache, 1988.6 The criteria for analgesic abuse were those of Diener, including onset of secondary chronic daily headache.7 The patients were asked to discontinue analgesics or ergotamine abruptly. They were informed orally and in writing about the phenomenon of medication-induced headache, and were encouraged to achieve withdrawal. A tricyclic antidepressant (amitriptyline, 10 mg) was given at night and a non-steroidal anti-inflammatory (naproxen 500 mg) was given orally for treatment of headache symptoms. Patients kept headache diaries, noting frequency and intensity of headaches and analgesic intake during the 6-month follow-up period. They were invited to follow-up visits 6 weeks, 12 weeks, and 6 months after withdrawal of medication, and all but 1 attended. The mean age of the group was 41-6 (range 22-56) years and the mean duration of headache history 23-7 (6-35) years. Most of the 29 patients taking drugs containing no ergotamine were heavy abusers; the weekly analgesic intake averaged 28-4 tablets (range 16-48). 10 were taking simple analgesics only (paracetamol or aspirin) and 19 were taking medications containing caffeine and/or codeine (’Propain’,
’Solpadeine’, ’Syndol’). 17 patients were abusing ergotamine; most were taking 2 mg/day regularly (mean intake 10.9
mg/week;
range 10-12
mg/week.
All 17
were
taking preparations containing caffeine (MigriF, ’Cafergot’). 11 were taking only these drugs, and 6 were taking analgesics as well. Ergotamine had been used regularly for 2 years on average (range 6 months to 20 years), and analgesics for even longer, although not on a regular basis. Abuse started on average a year after use of the medications began, but the range was very wide (2 months to
10
years).
The only withdrawal symptoms reported were increased headache intensity and nausea and vomiting. These symptoms were most severe on days 2-5, although 4 patients had no headache after drug withdrawal. At the end of the study, 37 patients had experienced relief from chronic daily headache; symptoms had reverted to those of the original occasional migraine attacks. Analgesic intake had been reduced to appropriate intermittent use for these attacks. In 6, headache severity had not changed, despite reduced analgesic intake. 2 patients had not reduced analgesic intake and showed no relief of headache. 1 patient who was taking ergotamine and caffeine did not return after the first visit at which withdrawal was proposed, and no follow-up information is available.
Although patients with medication abuse are said to be difficult to treat and many require hospital admission, we were able to treat 45 of a group of 46 migraine outpatients by abrupt withdrawal with good outcome in the majority. With written and spoken explanation, awareness of the facts, encouragement, the use of a non-steroidal antiinflammatory drug for relief of symptoms, a small dose of a tricyclic antidepressant as a prophylactic,8 and close followup, withdrawal was achieved without hospital admission or great interference in everyday life. As a specialist clinic, we see a large number of patients who have migraine with aura. All but 3 of the study subjects suffered from migraine without aura. Patients with aura are known to have less frequent attacks on average than those without, and may therefore be expected to use fewer analgesics, and be less liable to start abusing medication. Patients without aura may be less certain about the start of attacks and often treat less specific symptoms. Many of our patients embarked on the road to abuse by starting to take the medications in anticipation of a full-blown
development of abuse was apparently of independent the time over which the drugs had been taken, since the time before abuse ranged from as little as 2 months to up to 10 years after the medication was started. 17 of our 46 patients abused ergotamine with caffeine, but a surprisingly large number (10) abused only paracetamol or aspirin, which have been considered "safe". This study shows that such simple analgesics can be the cause of intoxication headaches, which accords with previous findings.9,lo Recognition of this possibility is obviously important, since management appears easy and readily effective if there is no concomitant complicating use of other drugs such as benzodiazepines or barbiturates (more common in North America). Most importantly, this study shows that drug withdrawal in analgesic and ergotamine abuse headaches can be achieved in outpatients. The drugs do not need to be gradually withdrawn-abrupt withdrawal is well tolerated, provided that adequate explanation and support are given. attack.
The
REFERENCES 1. Edmeads J. Analgesic-induced headache: an unrecognized epidemic. Headache 1990; 30: 614-15. 2. Tfelt-Hansen P, Krabbe EA. Ergotamine abuse. Do patients benefit from withdrawal? Cephalalgia 1981; 1: 29-32. 3. Mathew NT. Amelioration of ergotamine withdrawal symptoms with naproxen. Headache 1987; 27: 130-33. 4. Baumgartner C, Wessely P, Bigol C, Maly J, Holzner F. Longterm prognosis of analgesic withdrawal in patients with drug-induced headaches. Headache 1989; 29: 510-14. 5. Bernstein AL. More on ergotamine withdrawal. Headache 1987; 27: 458. 6. Headache Classification Committee of the International Headache Society. Classification and diagnosis criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988; 8 (suppl 7): 1-96. 7. Diener HC. Klinik des Analgetikakopfschmerzes. Deutsch med Wschr 1988; 113: 472-74. 8. Holland JV, Holland CV, Kudrow L. Low-dose amitriptyline prophylaxis in chronic scalp muscle contraction headache. Proceedings 1st International Headache Congress, Munich, 1983: 134. 9. Rapoport AM, Weeks RE, Sheftell FD, Baskin SM, Verdi J. Analgesic rebound headache: theoretical and practical implications. Cephalalgia 1985; 5 (suppl 3): 448-49. 10. MacGregor EA, Vohrah C, Wilkinson M. Analgesic use: a study of treatments used by patients for migraine prior to attending the City of London migraine clinic. Headache 1990; 30: 571-74.
ADDRESSES The Princess Margaret Migraine Clinic, Charing Cross Hospital, London W6 8RF, UK (R. Hering, MD, T J. Steiner,
PhD). Correspondence to
Dr R.
Hering
Glutathione peroxidase deficiency and childhood seizures
with intractable seizures, repeated and intolerance to anticonvulsants had infections, evidence of glutathione peroxidase deficiency. 2 had low intracellular enzyme activity but normal blood selenium and high plasma glutathione peroxidase concentrations. The other 2 had low intracellular glutathione peroxidase activity with low circulating glutathione peroxidase and selenium concentrations. The clinical state of the children improved after discontinuation of anticonvulsant medication and selenium substitution. 4 children
1444
Time-course of
neutrophil oxygen burst after phorbol-12myristate-13-acetate stimulation in 1 patient before (————W) and after (8 - - -8) selenium substitution, in healthy adults (0 - - - 0), and in healthy infants (x).
Glutathione peroxidase deficiency is chronic granulomatous disease1 and of
a
known
cause
of
idiosyncratic drug reactionsparticularly acute pancreatitis and increased serum lipid peroxides after therapy with valproic acid. Peroxidation stress, based on increased generation of active oxygen metabolites or reduced activity of antioxidative protection mechanisms, may increase the risk of recurrent seizures. We describe reduced glutathione peroxidase activity in 4 children with intractable seizures, repeated infections, and intolerance to anticonvulsant treatment. Concentrations of glutathione peroxidase and glutathione reductase were measured as previously described,4.5 and circulating selenium was measured by atomic absorption spectroscopy. The time-course of the neutrophil oxygen burst was measured as the rate of cytochrome C reduction by absorbance reading at 550 nm against a reference that contained superoxide dismutase, as described by Baker and Cohen.6 When an evaluable rate was detected, catalase was added to the cuvettes to distinguish between reoxidation of cytochrome C by accumulated H202 and inactivation of NADPH oxidase. All 4 patients had seizures during the first 6 months of life. Although the clinical pattern varied, electroencephalograms all showed multifocal activity and computed tomography and magnetic resonance imaging showed progressive cerebral atrophy. There was no evidence of inborn errors of metabolism (normal copper, caeruloplasmin, lactate, and pyruvate in serum and cerebrospinal fluid; normal oligosaccharides, mucopolysaccharides, and sialic acid in urine; normal aminoacids in plasma and urine; and normal lysosomal enzyme activities) or of central-nervous-system infections (no evidence of serum HIV antigen; negative for CMV and EBV antibodies in serum and cerebrospinal fluid; and normal immunoglobulins in serum) that are known to cause recurrent seizures. The karyotypes of all 4 patients were normal. 2 patients had low intracellular glutathione peroxidase
(12-7 and 20-02 IU/g haemoglobin; normal range 31 °Cr394) with normal or nearly normal whole-blood selenium (8-4 and 10-4 Ilg/d1, respectively; normal range 10-3-19-1) and high plasma glutathione peroxidase activity (55-09 and 50-44 IU/dI, respectively; normal range 32.447.6); their parents had reduced red-blood-cell glutathione peroxidase activity. The other 2 patients had low red-bloodcell glutathione peroxidase activity (13-50 and 8-25 IU/g haemoglobin) and low plasma enzyme activity (18-99 and 24-82 IV/d1, respectively), and low whole-blood selenium (5-1 and 3-4 u.g/dl, respectively); only the fathers had reduced red-blood-cell glutathione peroxidase activity, but glutathione reductase activity was low in all parents. concentrations
Intracellular glutathione reductase activity was in the high normal or above the normal range in all 4 patients (6-5-9 89 IU/g haemoglobin; normal range 5-4-7-4). Discontinuation of anticonvulsant therapy and supplementation of selenium, the metal in the active centre of glutathione peroxidase, led to clinical improvement in all 4 patients. The neutrophil oxygen burst has a duration of about 20 min in healthy children and adults. In the neutrophils of our 4 patients, the initial rate of superoxide generation was above normal for children of that age, which may reflect a primed state after repeated infections, but stopped after about 10 min. This shortened activity may be attributable to H20z-mediated damage to NADPH oxidase because of the lack of protection by glutathione peroxidase6 (see figure). 1 of our patients had a low overall rate of superoxide generation and was also deficient in glucose-6-phosphate dehydrogenase (0-04 U /1010 leucocytes, normal range 0-08-0-12, 4 U /1010 erythrocytes, normal range 18-34-8). Glucose-6-phosphate dehydrogenase regenerates NADPH, the cofactor for the free-radical generating enzyme NADPH oxidase. Muscle biopsies from all 4 patients showed increased interfascicular fibrous tissue and sarcolemmal accumulation of lipids. 1 patient died at 26 months; necropsy findings indicated white-matter disease with reduced brain weight, gliosis in the putamen, numerous Alzheimer type II astrocytes prominent in the globus pallidus, and Bergmann astrocytosis in the cerebellum; neurons were generally well
preserved. Our findings indicate that glutathione peroxidase deficiency may be a cause of childhood seizures. Our first 2 patients appear to have a primary deficiency of the intracellular enzyme, whereas the other 2 may have a disturbance of selenium resorption or transport. Enzyme defects which affect neutrophil-mediated immune defence, as well as the central nervous system, may need to be excluded in children with seizures, especially if combined with repeated infections and intolerance to anticonvulsants. This investigation was supported in part by a grant (to G. F. W.) from the Deutsche Forschungsgemeinschaft.
REFERENCES 1. Holmes B, Parks BH, Malawista SE, Quie PG, Nelson DL, Good RA. Chronic granulomatous disease in females. A deficiency of leukocyte glutathione peroxidase. N Engl J Med 1970; 283: 217-21. 2. Pippenger CE, Meng X, van Lente F. Alternative approaches to the prediction of antiepileptic, idiosyncratic, or drug-drug interactions. In: Pitlick WH, ed. Antiepileptic drug interactions. New York: Demos, 1989: 293-307. 3. Wilmore LJ, Rubin JJ. Antiperoxidant pretreatment and iron-induced epileptiform discharges in the rat: EEG and histopathologic studies. Neurology 1981; 31: 63-69. 4. Paglia DE, Valentine WN. Studies on the quantitative and qualitative characterization of glutathione peroxidase. J Lab Clin Med 1967; 70: 158-69. 5. Carlberg I, Mannervik B. Glutathione reductase. In: Meister A, ed. Methods in enzymology, vol 113. Orlando: Academic Press, 1985: 484-90. 6. Baker SS, Cohen HJ. Altered oxidative metabolism in selenium-deficient rat granulocytes. J Immunol 1983; 130: 2856—60.
ADDRESSES Department of Biological Chemistry and Molecular Pharmacology and Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115 (G F Weber, DM); Department of Neurology, University of South Alabama Medical Center, Mobile, AL 36617 (P Maertens, MD); and Department of Biochemistry, Cleveland Clinic Foundation, Cleveland, OH 44195, USA (X Meng, PhD, C E Pippenger, MD) Correspondence to Dr G. F. Weber, Department of Biological Chemistry and Molecular Pharmacology, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115, USA.
