4
102 Proc. roy. Soc. Med. Volume 68 February 1975 Conclusion The analytical method for the determination of NT in plasma here proved to be highly sensitive and specific. The method has been available for controlled trials and in clinical practice. Concerning the effect of TAs during long-term treatment our investigations have shed light on unclarified dosage factors of great importance for the elucidation of the possible prophylactic effect of the TAs. The plasma level determination of NT has simplified the discrimination between true pharmacological side effects, which call for reduction of dosage, and placebo or depressive symptoms. In patients not responding to a certain dosage, a plasma level determination will indicate whether the dosage should be raised or rather lowered. Our results prove this point, and we think we have a reasonable basis for the hypothesis that both a lower and an upper limit for therapeutic effect exist. Next, those patients who do not recover, even on optimal plasma levels, can be identified and receive alternative treatment, instead of undue prolongation of an ineffective therapy on changing doses. Whether a relationship between the plasma concentration and effects will hold also for other TAs remains to be investigated. It is not possible, at the present stage of knowledge, to extrapolate from one drug to another. It should also be pointed out that the curvilinear relationship between plasma level and therapeutic effect has only been demonstrated in patients with endogenous depressions and cannot be generalized to other depressive states without further trials. REFERENCES Alexanderson B, Evans D A P & Sj6qvist F (1969) British Medical Journal iv, 764-768 Alexanderson B & Sjoqvist F (1973) In: Proceedings of the Fifth International Congress on Pharmacology, San Francisco 1972. Karger, Basel; 3, 150-162 Asberg M, Cronholm B, Tuck D & Sjoqvist F (1970) British Medical Journal iv, 18-21 (1971) British Medical Journal iii, 331-334 Bennet J F (1967) Excerpta Medica International Congress Series No. 122, 375-393 Borga 0 & Garle M (1972) Journal ofChromatography 68, 77-88 Braithwaite R A, Goulding R, Theano G, Bailey J & Coppen A (1972) Lancet i, 1297-1300 Burrows G D, Davies B & Scoggins B A (1972) Lancet ii, 619-623 Gram L F & Over0 K F (1972) British Medical Journal i, 463-465 Hammer W & Brodie B B (1967) Journal ofPharmacology and Experimental Therapeutics 157, 503-508 Hammer W. Idestorm C-M & Sj6qvist F (1967) Excerpta Medica International Congress Series No. 122, 301-310 Hammer W & Sjllqvist F (1967) Life Science 6, 1895-1903 Kragh-Sorensen P, Eggert Hansen C & Asberg M (1973) Acta psychiatrica Scandinavica 49,444-456 Kragh-Sorensen P, Eggert Hansen C, Larsen N E, N2stoft J & Hvidberg E (1974) Psychological Medicine 4, 174-180 Sj6qvist F (1971) International Pharmacopsychiatry 6, 147-169 Walther C J S (1971) Proceedings of the Royal Society of Medicine 64, 282-285
Dr E H Reynolds (University Department of Neurology, Institute ofPsychiatry, De Crespigny Park, London SE5)
The Value of Serum Diphenylhydantoin (Phenytoin) Levels in the Management of Epilepsy
Techniques are now available for measuring many of the anticonvulsants in clinical use, but I will confine my discussion to a consideration of diphenylhydantoin (DPH), for which the most information and experience has accumulated. Although the drug has been in use since 1938 (Merritt & Putnam), it was not until the late 1950s and early 1960s that successful attempts were made to measure serum levels, principally in Denmark. With the more recent development of gas chromatographic techniques there has been widespread interest in their possible clinical application (see Buchthal & LennoxBuchthal 1972). Factors Affecting Serum Levels of DPH As for most drugs, it soon became apparent that for any given dose of DPH there is a wide variation in serum levels between individual
patients: (1) The mean serum half life of DPH is 22 hours, but a wide range (7-42 hours) has been noted by various authors, as the half life increases linearly with the serum level (Houghton & Richens 1974). (2) With oral medication a steady state level of DPH is achieved in 2-13 days, depending on the daily dose (Svensmark et al. 1960, Loeser 1961). It is usually unnecessary to administer the drug more than twice daily; indeed, steady state levels throughout the day have been reported with a single daily dose (Haerer & Buchanan 1972, Strandjord & Johannessen 1972). (3) For any individual patient the relationship between dose and serum level is not a linear one, as the latter will rise increasingly steeply as the dose is steadily increased (Bochner et al. 1972a). Within the optimum range of 10-20,tg/ml small incremental changes of dose may lead to a relatively large rise or fall in serum level, perhaps precipitating toxicity or loss of seizure control. (4) The capacity to metabolize (parahydroxylate) the drug varies with age, especially in children, in whom relatively larger doses (in terms of body weight) are required to achieve serum levels comparable to adult patients (Svensmark & Buchthal 1963, Jalling et al. 1970).
103
5
Section ofPsychiatry
(5) There has been relatively little investigation of the pharmacogenetics of DPH, but the family studies of Kutt, Wolk, Sherman & McDowell (1964) suggest that it may be important. A few patients develop very high serum levels of the drug on modest doses (e.g. 300 mg daily), possibly because of a genetically determined incapacity to adequately hydroxylate this dose (Lascelles et al. 1970). (6) Studies in Denmark (Lund et al. 1964) and this country (Gibberd et al. 1970) have confirmed that many patients fail to take the prescribed dose of the drug. (7) Variation in absorption, metabolism or excretion of the drug due to gastrointestinal, hepatic or renal disease, has been identified in some patients (Kutt, Winters, Sherman & McDowell 1964, Letteri et al. 1971). (8) Numerous interactions between DPH and other anticonvulsant or nonanticonvulsant drugs have been reported (Table 1) (see Kutt 1972), and this constitutes a serious problem for many epileptic patients who have to take the drug for most of their life. (9) DPH is 80-90 % protein bound. As the biological activity of the drug depends on its free concentration and as the degree of protein binding may be altered, for example, by other drugs (Kutt 1972) or in uremia (Reidenberg et al. 1971), it may be more appropriate in some circumstances to measure the free rather than the total drug concentration (Booker & Darcey 1973). (10) Variation in the formulation of DPH capsules may lead to alteration in the bioavailability of the drug as, for example, when an outbreak of DPH toxicity occurred in Australia when the excipient was changed from calcium sulphate to lactose (Bochner et al. 1972b).
et al. 1972). Recent experimental studies (Sherwin, Eisen & Sokolowski 1973) and studies of resected human temporal lobe specimens (Vajda et al. 1974) suggest that there is also a good correlation with brain levels.
Relation of Serum to Central Nervous System Levels of DPH For meaningful interpretation of serum levels, it is important to know that they reflect levels in the nervous system. Certainly there is a good correlation between serum and cerebrospinal fluid levels (which are equivalent to the unbound fraction in serum) (Triedman etal. 1960, Reynolds Table 1 Drugs which may influence serum DPH levels (see Kutt 1972) Increase of DPH
Sulthiame, ethosuximide, diazepam, chlordiazepoxide, isoniazid, bishydroxycoumarin, disulfiram, phenyramidol, chloramphenicol, sulfaphenazole, methylphenidate, phenylbutazone, chlorpromazine, prochlorperazine, cestrogens, propoxyphene, halothane Decrease of DPH Phenobarbitone, carbamazepine, ethanol
Relation ofSerum Levels to Seizure Control It is apparent that there is no clear relationship applicable to all patients. Buchthal et al. (1960), who found that 86% of patients with levels above 15 ,ug/ml were significantly improved whereas this was true in only 25 % of those with levels below 10 ,ug/ml, suggested that the optimal 'therapeutic' range is 10-20 ,g/ml. Although this is a useful practical objective it is already clear that many patients within this range remain uncontrolled, and others are controlled with levels below 10 ,ug/ml. The severity of the underlying seizure disorder in individual patients is an important factor in determining drug requirements and response to therapy. In large series of patients there may be no significant difference in DPH levels between controlled and uncontrolled epileptics, or the latter group may have relatively higher levels due to the administration of greater quantities of the drug in an unsuccessful attempt to control severe epilepsy completely (Buchanan & Allan 1971, Travers et al. 1972, Borofsky et al. 1972). However, Lund (1973) found that patients who had no seizures in the previous two months had significantly higher DPH levels than those with continuing attacks, although the mean dose of the drug was similar in the two groups. Paradoxically, toxic levels of DPH may sometimes aggravate seizures (Lascelles et al. 1970). The critical question is to what extent patients with continuing seizures and low serum DPH levels will be improved by increasing the drug level into the 10-20 ,ug/ml range, and this has not yet been satisfactorily assessed in a controlled manner with adequate follow up. The answer to this and other questions of the value of serum DPH levels will not be available until long-term studies have been completed, but a two-year study demonstrating the value of serum ethosuximide levels in improving petit mal control provides some grounds for optimism (Sherwin, Robb & Lechter 1973). Relation of Serum Levels to Toxicity Earlier reports of a very precise relationship between acute toxic effects and serum levels (Kutt, Winters, Kokenge & McDowell 1964) have not been confirmed and again considerable individual variation is apparent. Above 15-20 ,ug/ml there is a rising prevalence of clearly recognizable toxic effects such as nystagmus, ataxia and confusion, but some patients seem to tolerate high levels without obvious immediate
6
104 Proc. roy. Soc. Med. Volume 68 February 1975 Table 2 Indications for occasional serum DPH estimation (1) Uncontrolled seizures (2) Clinical toxicity (3) New neurological or psychiatric symptoms
Serum DPH may be too low or too high If on more than one drug To exclude unrecognized
toxicity
Special risks (1) Children (2) Brain damage or
Variation in metabolism Unreliable intake, unrecognized toxicity psychiatric illness Impaired metabolism (3) Hepatic or renal and excretion disease (4) Multiple drug therapy Drug interaction
ill effects (Buchthal et al. 1960). However, the problem is occasionally complicated by unusual neurological or psychiatric manifestations of toxicity, sometimes in the absence of the more classical signs, especially if the patient is already brain damaged (Reynolds 1970, Glaser 1972, Reynolds & Travers 1974). An inverse relationship between DPH (and phenobarbitone) and folate levels has been reported (Reynolds et al. 1972) but some of the insidious chronic effects of the drugs are not necessarily clearly related to serum levels, and duration of therapy may be equally, if not more, important (Reynolds 1975).
REFERENCES Bochner F, Hooper W D, Tyrer J H & Eadie M J (1972a) Journal ofNeurology, Neurosurgery andPsychiatry 35, 873 (1972b) Journal of the Neurological Sciences 16, 481 Booker H E & Darcey B (1973) Epilepsia 14, 177 Borofsky L G, Louis S, Kutt H & Roginsky M (1972) Journal ofPediatrics 81, 995 Buchanan R A & Allan R J (1971) Neurology (Minneapolis) 21, 866 Buchthal F & Lennox-Buchthal M A (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 193 Buchthal F, Svensmark 0 & Schiller P J (1960) Archives ofNeurology 2, 624 Gibberd F B, Dunne J F, Handley A J & Hazelman B L (1970) British MedicalJournal i, 147 Glaser G H (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 219 Haerer A F & Buchanan R A (1972) Neurology (Minneapolis) 22, 1021 Houghton G W & Richens A (1974) British Journal of Clinical Pharmacology 1, 155 Jalling B, Boreus L 0, Rane A & Sj6qvist F (1970) Pharmacologia Clinica 2, 200 Kutt H (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 169 Kutt H, Winters W, Kokenge R & McDowell F (1964) Archives ofNeurology 11, 642 Kutt H, Winters W, Scherman R & McDowell F (1964) Archives of Neurology 11, 649 Kutt H, Wolk M, Scherman R & McDoweli F (1964) Neurology (Minneapolis) 14, 542 Lascelles P T, Kocen R S & Reynolds E H (1970) Journal ofNeurology, Neurosurgery and Psychiatry 33, 501 Letteri J M, Mellik H, Louis S, Kutt H, Durante P & Glazko A (1971) New England Journal of Medicine 285, 648 Loeser E W (1961) Neurology (Minneapolis) 11, 424 Lund L (1973) In: Biological Effects of Drugs in Relation to Their Plasma Concentrations. Ed. D S Davis & B N C Prichard. Macmillan, London; p 227 Lund M, J0rgensen R S & KUMh V (1964) Epilepsia 5, 51 Merritt H H & Putnam T J (1938) Journal of the American Medical Association 111, 1068 Reidenburg M M, Odar-Cederlf I, Von Bahr C, Borga A & Sjoqvist F (1971) New England Journal of Medicine 285,264 Reynolds E H (1970) In: Modern Trends in Neurology. Ed. D Williams. Butterworths, London; 5,271 Reynolds E H (1975) Epilepsia (in press) Reynolds E H, Mattson R H & Gallagher B B (1972) Neurology (Minneapolis) 22, 841 Reynolds E H & Travers R D (1974) British Journal ofPsychiatry 124, 440 Sherwin A L, Eisen A A & Sokolowski L D (1973) Archives of Neurology 29, 73 Sherwin A L, Robb J P & Lechter M (1973) Archives ofNeurology 28, 178 Strandjord R E & Johannessen S I (1972) Acta neurologica Scandinavica Suppl. 51, 499 Svensmark 0 & Buchthal F (1963) Danish Medical Bulletin 10, 234 Svensmark 0, Schiller P J & Buchthal F (1960) Acta pharmacologica et toxicologica 16, 331 Travers R D, Reynolds E H & Gallagher B B (1972) Archives of Neurology 27, 29 Triedman H M, Fishman R A & Yahr M D (1960) Transactions of the American Neurological Association 85, 166 Vaida F, Williams F M, Davison S, Falconer M A & Breckenridge A (1974) Clinical Pharmacology and Therapeutics 15, 597
Conclusion Although the place of regular monitoring of serum DPH levels has yet to be evaluated by long-term studies, the value of occasional measurements in individual patients who are difficult management problems has been emphasized by many. Table 2 lists some of the clinical situations in which such measurements may be helpful: (1) For any patient with continuing seizures, it is useful to make sure that an optimum serum level has been achieved before considering the addition of or substitution by another drug. (2) Serum levels will also indicate which drug may be appropriately reduced if a patient on multiple drug therapy presents with signs of toxicity. (3) As DPH toxicity may sometimes present with a variety of atypical neurological or psychiatric syndromes, the serum level should be measured in any patient on the drug presenting with new neuropsychiatric symptoms. With the increasing availability of facilities for measuring serum levels of DPH and other anticonvulsants, there is now much less risk of under or overdosage with these drugs. This is particularly important in patients who may have to take anticonvulsant therapy for many years of their life. Studies currently in progress will define more The following paper was also read: precisely the indications for and frequency of such measurements, and will also assess whether Concentrations of Drugs at Their Sites of Action this has any impact on the long-term prognosis Dr R H Nimmo-Smith (for Dr Miles Weatherall) of epilepsy. (Wellcome Research Laboratories, Kent)
102 Proc. roy. Soc. Med. Volume 68 February 1975 Conclusion The analytical method for the determination of NT in plasma here proved to be highly sensitive and specific. The method has been available for controlled trials and in clinical practice. Concerning the effect of TAs during long-term treatment our investigations have shed light on unclarified dosage factors of great importance for the elucidation of the possible prophylactic effect of the TAs. The plasma level determination of NT has simplified the discrimination between true pharmacological side effects, which call for reduction of dosage, and placebo or depressive symptoms. In patients not responding to a certain dosage, a plasma level determination will indicate whether the dosage should be raised or rather lowered. Our results prove this point, and we think we have a reasonable basis for the hypothesis that both a lower and an upper limit for therapeutic effect exist. Next, those patients who do not recover, even on optimal plasma levels, can be identified and receive alternative treatment, instead of undue prolongation of an ineffective therapy on changing doses. Whether a relationship between the plasma concentration and effects will hold also for other TAs remains to be investigated. It is not possible, at the present stage of knowledge, to extrapolate from one drug to another. It should also be pointed out that the curvilinear relationship between plasma level and therapeutic effect has only been demonstrated in patients with endogenous depressions and cannot be generalized to other depressive states without further trials. REFERENCES Alexanderson B, Evans D A P & Sj6qvist F (1969) British Medical Journal iv, 764-768 Alexanderson B & Sjoqvist F (1973) In: Proceedings of the Fifth International Congress on Pharmacology, San Francisco 1972. Karger, Basel; 3, 150-162 Asberg M, Cronholm B, Tuck D & Sjoqvist F (1970) British Medical Journal iv, 18-21 (1971) British Medical Journal iii, 331-334 Bennet J F (1967) Excerpta Medica International Congress Series No. 122, 375-393 Borga 0 & Garle M (1972) Journal ofChromatography 68, 77-88 Braithwaite R A, Goulding R, Theano G, Bailey J & Coppen A (1972) Lancet i, 1297-1300 Burrows G D, Davies B & Scoggins B A (1972) Lancet ii, 619-623 Gram L F & Over0 K F (1972) British Medical Journal i, 463-465 Hammer W & Brodie B B (1967) Journal ofPharmacology and Experimental Therapeutics 157, 503-508 Hammer W. Idestorm C-M & Sj6qvist F (1967) Excerpta Medica International Congress Series No. 122, 301-310 Hammer W & Sjllqvist F (1967) Life Science 6, 1895-1903 Kragh-Sorensen P, Eggert Hansen C & Asberg M (1973) Acta psychiatrica Scandinavica 49,444-456 Kragh-Sorensen P, Eggert Hansen C, Larsen N E, N2stoft J & Hvidberg E (1974) Psychological Medicine 4, 174-180 Sj6qvist F (1971) International Pharmacopsychiatry 6, 147-169 Walther C J S (1971) Proceedings of the Royal Society of Medicine 64, 282-285
Dr E H Reynolds (University Department of Neurology, Institute ofPsychiatry, De Crespigny Park, London SE5)
The Value of Serum Diphenylhydantoin (Phenytoin) Levels in the Management of Epilepsy
Techniques are now available for measuring many of the anticonvulsants in clinical use, but I will confine my discussion to a consideration of diphenylhydantoin (DPH), for which the most information and experience has accumulated. Although the drug has been in use since 1938 (Merritt & Putnam), it was not until the late 1950s and early 1960s that successful attempts were made to measure serum levels, principally in Denmark. With the more recent development of gas chromatographic techniques there has been widespread interest in their possible clinical application (see Buchthal & LennoxBuchthal 1972). Factors Affecting Serum Levels of DPH As for most drugs, it soon became apparent that for any given dose of DPH there is a wide variation in serum levels between individual
patients: (1) The mean serum half life of DPH is 22 hours, but a wide range (7-42 hours) has been noted by various authors, as the half life increases linearly with the serum level (Houghton & Richens 1974). (2) With oral medication a steady state level of DPH is achieved in 2-13 days, depending on the daily dose (Svensmark et al. 1960, Loeser 1961). It is usually unnecessary to administer the drug more than twice daily; indeed, steady state levels throughout the day have been reported with a single daily dose (Haerer & Buchanan 1972, Strandjord & Johannessen 1972). (3) For any individual patient the relationship between dose and serum level is not a linear one, as the latter will rise increasingly steeply as the dose is steadily increased (Bochner et al. 1972a). Within the optimum range of 10-20,tg/ml small incremental changes of dose may lead to a relatively large rise or fall in serum level, perhaps precipitating toxicity or loss of seizure control. (4) The capacity to metabolize (parahydroxylate) the drug varies with age, especially in children, in whom relatively larger doses (in terms of body weight) are required to achieve serum levels comparable to adult patients (Svensmark & Buchthal 1963, Jalling et al. 1970).
103
5
Section ofPsychiatry
(5) There has been relatively little investigation of the pharmacogenetics of DPH, but the family studies of Kutt, Wolk, Sherman & McDowell (1964) suggest that it may be important. A few patients develop very high serum levels of the drug on modest doses (e.g. 300 mg daily), possibly because of a genetically determined incapacity to adequately hydroxylate this dose (Lascelles et al. 1970). (6) Studies in Denmark (Lund et al. 1964) and this country (Gibberd et al. 1970) have confirmed that many patients fail to take the prescribed dose of the drug. (7) Variation in absorption, metabolism or excretion of the drug due to gastrointestinal, hepatic or renal disease, has been identified in some patients (Kutt, Winters, Sherman & McDowell 1964, Letteri et al. 1971). (8) Numerous interactions between DPH and other anticonvulsant or nonanticonvulsant drugs have been reported (Table 1) (see Kutt 1972), and this constitutes a serious problem for many epileptic patients who have to take the drug for most of their life. (9) DPH is 80-90 % protein bound. As the biological activity of the drug depends on its free concentration and as the degree of protein binding may be altered, for example, by other drugs (Kutt 1972) or in uremia (Reidenberg et al. 1971), it may be more appropriate in some circumstances to measure the free rather than the total drug concentration (Booker & Darcey 1973). (10) Variation in the formulation of DPH capsules may lead to alteration in the bioavailability of the drug as, for example, when an outbreak of DPH toxicity occurred in Australia when the excipient was changed from calcium sulphate to lactose (Bochner et al. 1972b).
et al. 1972). Recent experimental studies (Sherwin, Eisen & Sokolowski 1973) and studies of resected human temporal lobe specimens (Vajda et al. 1974) suggest that there is also a good correlation with brain levels.
Relation of Serum to Central Nervous System Levels of DPH For meaningful interpretation of serum levels, it is important to know that they reflect levels in the nervous system. Certainly there is a good correlation between serum and cerebrospinal fluid levels (which are equivalent to the unbound fraction in serum) (Triedman etal. 1960, Reynolds Table 1 Drugs which may influence serum DPH levels (see Kutt 1972) Increase of DPH
Sulthiame, ethosuximide, diazepam, chlordiazepoxide, isoniazid, bishydroxycoumarin, disulfiram, phenyramidol, chloramphenicol, sulfaphenazole, methylphenidate, phenylbutazone, chlorpromazine, prochlorperazine, cestrogens, propoxyphene, halothane Decrease of DPH Phenobarbitone, carbamazepine, ethanol
Relation ofSerum Levels to Seizure Control It is apparent that there is no clear relationship applicable to all patients. Buchthal et al. (1960), who found that 86% of patients with levels above 15 ,ug/ml were significantly improved whereas this was true in only 25 % of those with levels below 10 ,ug/ml, suggested that the optimal 'therapeutic' range is 10-20 ,g/ml. Although this is a useful practical objective it is already clear that many patients within this range remain uncontrolled, and others are controlled with levels below 10 ,ug/ml. The severity of the underlying seizure disorder in individual patients is an important factor in determining drug requirements and response to therapy. In large series of patients there may be no significant difference in DPH levels between controlled and uncontrolled epileptics, or the latter group may have relatively higher levels due to the administration of greater quantities of the drug in an unsuccessful attempt to control severe epilepsy completely (Buchanan & Allan 1971, Travers et al. 1972, Borofsky et al. 1972). However, Lund (1973) found that patients who had no seizures in the previous two months had significantly higher DPH levels than those with continuing attacks, although the mean dose of the drug was similar in the two groups. Paradoxically, toxic levels of DPH may sometimes aggravate seizures (Lascelles et al. 1970). The critical question is to what extent patients with continuing seizures and low serum DPH levels will be improved by increasing the drug level into the 10-20 ,ug/ml range, and this has not yet been satisfactorily assessed in a controlled manner with adequate follow up. The answer to this and other questions of the value of serum DPH levels will not be available until long-term studies have been completed, but a two-year study demonstrating the value of serum ethosuximide levels in improving petit mal control provides some grounds for optimism (Sherwin, Robb & Lechter 1973). Relation of Serum Levels to Toxicity Earlier reports of a very precise relationship between acute toxic effects and serum levels (Kutt, Winters, Kokenge & McDowell 1964) have not been confirmed and again considerable individual variation is apparent. Above 15-20 ,ug/ml there is a rising prevalence of clearly recognizable toxic effects such as nystagmus, ataxia and confusion, but some patients seem to tolerate high levels without obvious immediate
6
104 Proc. roy. Soc. Med. Volume 68 February 1975 Table 2 Indications for occasional serum DPH estimation (1) Uncontrolled seizures (2) Clinical toxicity (3) New neurological or psychiatric symptoms
Serum DPH may be too low or too high If on more than one drug To exclude unrecognized
toxicity
Special risks (1) Children (2) Brain damage or
Variation in metabolism Unreliable intake, unrecognized toxicity psychiatric illness Impaired metabolism (3) Hepatic or renal and excretion disease (4) Multiple drug therapy Drug interaction
ill effects (Buchthal et al. 1960). However, the problem is occasionally complicated by unusual neurological or psychiatric manifestations of toxicity, sometimes in the absence of the more classical signs, especially if the patient is already brain damaged (Reynolds 1970, Glaser 1972, Reynolds & Travers 1974). An inverse relationship between DPH (and phenobarbitone) and folate levels has been reported (Reynolds et al. 1972) but some of the insidious chronic effects of the drugs are not necessarily clearly related to serum levels, and duration of therapy may be equally, if not more, important (Reynolds 1975).
REFERENCES Bochner F, Hooper W D, Tyrer J H & Eadie M J (1972a) Journal ofNeurology, Neurosurgery andPsychiatry 35, 873 (1972b) Journal of the Neurological Sciences 16, 481 Booker H E & Darcey B (1973) Epilepsia 14, 177 Borofsky L G, Louis S, Kutt H & Roginsky M (1972) Journal ofPediatrics 81, 995 Buchanan R A & Allan R J (1971) Neurology (Minneapolis) 21, 866 Buchthal F & Lennox-Buchthal M A (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 193 Buchthal F, Svensmark 0 & Schiller P J (1960) Archives ofNeurology 2, 624 Gibberd F B, Dunne J F, Handley A J & Hazelman B L (1970) British MedicalJournal i, 147 Glaser G H (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 219 Haerer A F & Buchanan R A (1972) Neurology (Minneapolis) 22, 1021 Houghton G W & Richens A (1974) British Journal of Clinical Pharmacology 1, 155 Jalling B, Boreus L 0, Rane A & Sj6qvist F (1970) Pharmacologia Clinica 2, 200 Kutt H (1972) In: Antiepileptic Drugs. Ed. D M Woodbury, J K Penry and R P Schmidt. Raven Press, New York; p 169 Kutt H, Winters W, Kokenge R & McDowell F (1964) Archives ofNeurology 11, 642 Kutt H, Winters W, Scherman R & McDowell F (1964) Archives of Neurology 11, 649 Kutt H, Wolk M, Scherman R & McDoweli F (1964) Neurology (Minneapolis) 14, 542 Lascelles P T, Kocen R S & Reynolds E H (1970) Journal ofNeurology, Neurosurgery and Psychiatry 33, 501 Letteri J M, Mellik H, Louis S, Kutt H, Durante P & Glazko A (1971) New England Journal of Medicine 285, 648 Loeser E W (1961) Neurology (Minneapolis) 11, 424 Lund L (1973) In: Biological Effects of Drugs in Relation to Their Plasma Concentrations. Ed. D S Davis & B N C Prichard. Macmillan, London; p 227 Lund M, J0rgensen R S & KUMh V (1964) Epilepsia 5, 51 Merritt H H & Putnam T J (1938) Journal of the American Medical Association 111, 1068 Reidenburg M M, Odar-Cederlf I, Von Bahr C, Borga A & Sjoqvist F (1971) New England Journal of Medicine 285,264 Reynolds E H (1970) In: Modern Trends in Neurology. Ed. D Williams. Butterworths, London; 5,271 Reynolds E H (1975) Epilepsia (in press) Reynolds E H, Mattson R H & Gallagher B B (1972) Neurology (Minneapolis) 22, 841 Reynolds E H & Travers R D (1974) British Journal ofPsychiatry 124, 440 Sherwin A L, Eisen A A & Sokolowski L D (1973) Archives of Neurology 29, 73 Sherwin A L, Robb J P & Lechter M (1973) Archives ofNeurology 28, 178 Strandjord R E & Johannessen S I (1972) Acta neurologica Scandinavica Suppl. 51, 499 Svensmark 0 & Buchthal F (1963) Danish Medical Bulletin 10, 234 Svensmark 0, Schiller P J & Buchthal F (1960) Acta pharmacologica et toxicologica 16, 331 Travers R D, Reynolds E H & Gallagher B B (1972) Archives of Neurology 27, 29 Triedman H M, Fishman R A & Yahr M D (1960) Transactions of the American Neurological Association 85, 166 Vaida F, Williams F M, Davison S, Falconer M A & Breckenridge A (1974) Clinical Pharmacology and Therapeutics 15, 597
Conclusion Although the place of regular monitoring of serum DPH levels has yet to be evaluated by long-term studies, the value of occasional measurements in individual patients who are difficult management problems has been emphasized by many. Table 2 lists some of the clinical situations in which such measurements may be helpful: (1) For any patient with continuing seizures, it is useful to make sure that an optimum serum level has been achieved before considering the addition of or substitution by another drug. (2) Serum levels will also indicate which drug may be appropriately reduced if a patient on multiple drug therapy presents with signs of toxicity. (3) As DPH toxicity may sometimes present with a variety of atypical neurological or psychiatric syndromes, the serum level should be measured in any patient on the drug presenting with new neuropsychiatric symptoms. With the increasing availability of facilities for measuring serum levels of DPH and other anticonvulsants, there is now much less risk of under or overdosage with these drugs. This is particularly important in patients who may have to take anticonvulsant therapy for many years of their life. Studies currently in progress will define more The following paper was also read: precisely the indications for and frequency of such measurements, and will also assess whether Concentrations of Drugs at Their Sites of Action this has any impact on the long-term prognosis Dr R H Nimmo-Smith (for Dr Miles Weatherall) of epilepsy. (Wellcome Research Laboratories, Kent)
