738
Mental Health NEWER CONCEPTS OF PSYCHIATRIC DIAGNOSIS AND BIOCHEMICAL RESEARCH ON MENTAL ILLNESS
J. P. BLASS J. F. MILNE Departments of Psychiatry and Biological Chemistry and the Mental Retardation Research Center, School of Medicine, University of California, Los Angeles, U.S.A. R. RODNIGHT
Department of Biochemistry, Institute of Psychiatry, University of London Biochemical studies of psychiatric patients have revealed many abnormalities but have failed to identify specific biochemical changes linked to the pathophysiology of the illnesses. Review of these data in the light of newer concepts of psychiatric diagnosis may help to clarify the relationship between biochemical abnormalities and disabilities. These concepts may also suggest useful directions for future research.
Summary
INTRODUCTION
THE identification of biological factors which contribute to the development of mental illness has been actively pursued for many years, and it is now well established that many patients inherit genetic predisposition to the development of psychoses.’ Since genes code for proteins, psychotic patients presumably inherit one or more abnormal enzymes or other proteins which alter their body chemistry in such a way as to contribute to the development of their illnesses. The body chemistry of psychotics has already been studied intensively.2 New discoveries in neurochemistry have been applied promptly to the study of mental illness. A good example is the study of the endorphins and related peptides which have been discovered in the past three years.3 4 Unfortunately, the search for a relationship between biochemical abnormalities and mental illness has been disappointing.2 Many abnormalities have been found in different groups of patients, but almost all of these have occurred in patients with very different clinical syndromes and even in some clinically normal peopled However, reexamination of these studies in the light of newer approaches to diagnosis and classification promises to be
rewarding. DIAGNOSIS AND CLASSIFICATION
classifications in Western medicine have between two poles.6 One view, historically associated with the Platonic school, formulates diseases as discrete entities, with the clinical manifestations in a given patient resembling more or less closely those in a "typical" patient. This view assumes that clinical criteria can define groups of patients who are biologically homogeneous, so that the appropriate physiological, anatomical, or biochemical criteria will each define the same group of patients. This approach was developed in detail by Thomas Sydenham in the 17th century
Diagnostic
tended
to move
and underlies the 19th-century classification of psychiatric disorders. It also underlies the search for "the cause" of schizophrenia-a search which has had little success. The second approach to classifying diseases is historically associated with the name of Hippocrates. It is more phenomenologic, viewing diseases as operational definitions of groups of patients, even though the groups thus created are biologically heterogeneous. This approach assumes that whereas clinical, physiological, anatomical, and biochemical measurements may correlate with each other statistically, the different criteria do not identify identical groups of individuals. The sets may overlap, but they are not congruent. From this viewpoint, a "disease" is an intellectual construct defined by the parameters which are chosen to define it. For instance, "schizophrenia" is a convenient way to refer to a particular type of human behaviour which can result from a number of different factors. The implication is that "schizophrenics" are a biologically heterogeneous group and that biochemical studies of these patients will be expected to turn up a variety of abnormalities, none of which is specific for the entity defined by clinical criteria alone. This is, of course, what has been found. The potential value of the hippocratic approach is supported by its value in other areas of medicine. In genetic diseases a single clinical phenotype can be associated with more than one genotype. Entities such as anaemia and pneumonia also do not have single xtiotogies. Indeed, when Scadding’ and Burrows8 began to apply computers and mathematical rigour to medical diagnosis, they stated explicitly that clinical diagnoses were operational definitions for biologically complex and diverse phenomena. The introduction of the problemoriented record reflects, among other things, the inadequacies of dealing with the practical complexities of diagnosis and management in terms of specific diseases. Over 10 years ago, the American Heart Association recommended diagnostic formulations which required separate statements about aetiology, anatomy, physiology, and disability-for example, rheumatic heart-disease with mitral stenosis, left-sided congestive heart-failure, and moderate disability. It seems unlikely that precise formulations for disorders of the brain will prove simpler than for disorders of the heart. METABOLIC ABNORMALITIES AND MENTAL DISORDER
In metabolic disorders known to be associated with psychiatric abnormalities, there is no simple and con-
relationship between the metabolic disorder and the clinical manifestations. Thus, a small proportion of patients with phenylketonuria have the syndrome of infantile autism, and another small proportion with typical hydroxylase-deficiency phenylketonuria appear clinically normal with normal intelligence.lO Some patients whose -tissues cannot hydrolyse nitrocatechol sulphate (and are therefore diagnosed as having metachromatic leucodystrophy) have hallucinations, ideas of reference, grandiosity, or bizarre behaviour." Although this disorder is now thought to be rare, we know of no systematic study of its frequency in psychiatric populations. When modified Watson-Schwarz tests for porphyria were done routinely in a university psychiatry service in Ohio, 35 patients with porphyria were found among stant
2500 consecutive admissions. 12 This contrasts with 2 among the previous 2000 admis-
porphyrics diagnosed
‘
739 among 1000 consecutive admissions to a nearby general hospital. The admission diagnoses of the 35 were the following: schizophrenia (10), chronic brain
sions and
none
syndrome (10), alcoholism (6), psychotic depressions (2), neuroses, personality disorders, and miscellaneous (7). Unfortunately, more detailed studies to identify the specific mutations of porphyrin metabolism in these patients were not done. Part of the clinical variability in patients with porphyrias may relate to the effects of an independent "modifier" mutation affecting steroid-hormone
breakdown. 13
Furthermore, the psychiatric manifestations of organic brain disease can sometimes be identical to functional psychosis.14 Indeed, the international classification of psychiatric disorders lists the absence of demonstrable brain disease as one of the criteria for the diagnosis of schizophrenia.6 IS The variable psychiatric manifestations of temporal-lobe epilepsy are well known,16 as are the various psychiatric syndromes which follow head injuryl7 and accompany brain damage in children.18 Metabolic or other biological abnormalities of the brain may predispose to the development of psychiatric disorders, but they are only a few of the factors which determine the form of the observed behavioural abnormalities. These are presumably influenced as well by cultural and other experiences of the individual, by epigenetic factors, and by modifier genes.
CRITERIA FOR IDENTIFICATION OF METABOLIC DISEASE
The analysis presented above suggests that techniques and criteria used to identify inherited metabolic abnormalities in medical disorders can be usefully applied in psychiatry. Such criteria have been outlined elsewhere. " 1. Analytical studies should indicate abnormal amounts of compound or group of related compounds much more frequently in psychiatric patients than in normal individuals. However, there need not be a one-to-one correspondence between the analytical abnormalities and a single behavioural diagnosis. (Many analytical abnormalities which would meet this criterion have already been described) 2. The rates of metabolism of these compounds should be abnormal while those of related compounds are normal. Demonstration of the metabolic abnormalities in vitro in biopsy specimens which were histologically normal would indicate that the abnormality was not due simply to tissue damage by a
the disease process. 3. Demonstration of a defect in cultured fibroblasts or other cultured cells would demonstrate that it was genetic rather than due to diet, drugs, or other external factors. After serial passages in culture, the material in the original explant will have been diluted several millionfold by new cell-growth, and cells from the patients and controls are exposed to identical conditions in culture. Thus, abnormalities which persist in culture may be assumed to be coded into the genetic information in the cells. 4. The metabolic abnormality should be shown to be due to aberration in a specific enzyme or other protein, if possible in tissue-culture. 5. Demonstration that the enzymatic abnormality was distnbuted among the patients’ relatives in accordance with the known genetics of the disorder would be a strong indication that the enzymatic aberration was inherited as a primary abnormality. For instance, in a disorder inherited as an autosomal recessive, the enzyme activities in the parents should be intermediate between those of the patient and controls. 6. Demonstration of a reasonable mechanism by which the an
observed abnormality could impair brain function would support its pathophysiological significance. 7. The most important evidence that a metabolic abnormality had a significant role in the pathogenesis of a psychiatric disorder would be if a specific diet or other regimen which normalised the affected patients’ body chemistry also benefited them clinically, in double-blind cross-over trials.
One might consider as an example how these criteria could be applied to the recent observation that a significant proportion of psychotic patients excrete abnormally low amounts of phenylethylamine, as do a small proportion of clinically normal people .20 Experimental studies have indicated that phenylethylamine may act as a neurotransmitter, and pathophysiological models have been
suggested for linking phenylethylamine deficiency
to
psychoses.21Thus,
criteria 1 and 6 listed above have been fulfilled. Several questions arise. Is the metabolism of phenylethylamine and phenylalanine abnormal in tissues of the affected individuals, and is the metabolism of other amines and aminoacids normal ? Can an abnormality in the metabolism of phenylethylamine be demonstrated in cultured cells from these patients? Is there an abnormality in phenylalanine decarboxylase, the enzyme which synthesises phenylethylamine from phenylalanine, or in some other enzyme of phenylethylamine metabolism? Can an abnormality be demonstrated in tissues from patients’ relatives? Finally, and most important, can the excretion of phenylethylamine be normalised-for instance, by a diet enriched in phenylalanine or phenylethylamine? If so, would such a diet benefit specifically those patients who have the abnormality in phenylethylamine metabolism, in double-blind cross-over studies? Similar questions could be asked about other metabolites found to be abnormal in other patients.5These questions can be answered with -the help of well-known techniques of clinical and biochemical in-
vestigation. AN INTEGRATED APPROACH TO MENTAL ILLNESS
The search for hereditary metabolic abnormalities which predispose to mental illness does not deprecate at all the value of other approaches to studying psychiatric disease. For instance, psychopharmacological studies are obviously important clinically. The responses of particular "target symptoms" to specific drugs may provide critical clues to pathophysiological mechanisms, even if not necessarily to aetiology. (It is worth noting that the response to antiseizure medication provides relatively little information about the origin of a seizure disorder, and the response to aspirin little about the aetiology of a fever.) Furthermore, it should be emphasised that the search for biological abnormalities in psychiatric patients in no way implies that all psychiatric problems derive from biological abnormalities of the brain. Nor does it minimise the importance of psychological treatment techniques even in patients with demonstrable organic disease. Psychological mechanisms clearly continue to operate in people with biologically abnormal brains. The search for metabolic abnormalities is only an attempt to define more precisely certain biological factors which limit function and adaptability in some psychiatric patients. Both genetic1 and biochemical5 studies suggest that such abnormalities are important in a large proportion of patients with psychoses. One may hope that eventually information from different types of studies can be
740
integrated into coherent descriptions of pathophysiological mechanisms in patients with psychiatric disorders. The approach suggested in this paper is in essence a restatement of Claude Bernard’s proposals for an "experimental medicine", phrased in more modern terminology and in relation specifically to psychiatry.22 It implies the division of patients with similar behavioural disorders into biological subgroups and explicitly rejects a search for single causes or mechanisms for such complex clinical phenomena as "schizophrenia" or "depression". Analogous studies have given very useful information in other areas of medicine. The usefulness of this approach in neuropsychiatry can be determined only by further observations. This research was supported in part by grants NF 6-76-50 from the National Foundation March of Dimes and HD-06576 from the N.I.C.H.D.
Requests for reprints should be addressed to J. P. B., Mental Retardation Research Center, U.C.L.A. School of Medicine, Los Angeles, California 90024, U.S.A.
REFERENCES 1. 2. 3. 4. 5.
Slater, E., Cowie, V. Genetics of Mental Disorders. London, 1971. Rodnight, R. Psychol. Med. 1971, 1, 353. Bloom, F., Segal, D., Ling, N., Guillemin, R. Science, 1976, 194, 632. Jacquet, Y., Marks, N. ibid. p. 634. Weil-Malherbe, H., Szara, S. I. Biochemistry of Functional and Experimental Psychoses. Springfield, Illinois, 1971. 6. Kendell, R. E. Role of Diagnosis in Psychiatry. Oxford, 1975. 7. Scadding, J. G. Lancet, 1967, ii, 877. 8. Burrows, B. Ann. intern. Med. 1975, 83, 419. 9. Ritvo, E. R., Ornitz, M., LaFranchi, S. Archs gen. Psychiat. 1968, 19, 341. 10. Brown, E. S., Waisman, H. A. Pediatrics, Springfield, 1967, 40, 247. 11. Müller, D., Pilz, H., Meulen, V. T. J. neurol. Sci. 1969, 9, 567. 12. Kaebling, R., Craig, J., Pasamanik, B. Archs gen. Psychiat. 1961, 5, 494. 13. Kappas, A., Sassa, A., Granick, S., Bradlow, H. L. in Brain Dysfunction in Metabolic Disorders (edited by F. Plum); p. 225. New York, 1974. 14. Slater, E., Roth, M. Clinical Psychiatry; p. 483. Baltimore, 1969. 15. International Statistical Classification of Diseases, Injury, and Death. W.H.O., 1967. 16. Baldwin, M., Bailey, P. (editors) Temporal Lobe Epilepsy. Springfield, Illinois, 1958. 17. Hillbom, E. Acta psychiat. neurol. scand. 1960, 35, suppl. 142, 128. 18. Graham, P., Rutter, M. Br. med. J. 1968, iii, 698. 19. Blass, J. P., Steinberg, D. in Biology of Brain Dysfunction (edited by G. Gaull); vol. 2, p. 239. New York, 1973. 20. Sandier, M., Reynolds, G. P. Lancet, 1976, i, 70. 21. Giardina, W. J., Pedemonte, W. A., Sabelli, H. C. Life Sci. 1973, 12, 153. 22. Bernard, C. Introduction à l’étude de la médicine experimentale. Pans, 1952.
Occasional
Survey
CAN RIFAMPICIN USE BE SAFELY EXTENDED?
Evidence for Non-emergence of Resistant Strains of Mycobacterium tuberculosis G. ACOCELLA*
J. M. T. HAMILTON-MILLER W. BRUMFITT
Royal Free Hospital,
Pond Street, London NW3 2QG
the incidence of primary resisrifampicin in Mycobacterium strains have been collected from various tuberculosis countries. Strains isolated from those countries where rifampicin is used for both tuberculous and non-tuberculous conditions (Italy, Argentina, Brazil, and Spain) did not show a higher incidence of primary resistance than did strains from other countries (France, U.K., and U.S.A.) where rifampicin use is confined to tuberculosis. It is concluded that there is no evidence to justify fears of an increased incidence of resistance to rifampicin in M. tuberculosis if rifampicin were used discreetly for treating non-tuberculous infections.
Summary
Data
on
tance to
INTRODUCTION
Kerry, Hamilton-Miller, and Brumfittl reported on the antibacterial activity of the combination rifampicin/trimethoprim. They found at least an additive action against a wide range of pathogens, as well as suppression of the emergence of rifampicin-resistant mutants. These results, together with those obtained in pharmacokinetic experiments23 and with an animal-infection model4 suggest that this combination might be useful for treating a wide variety of infections in man. However, some authorities’6 have declared themselves against rifampicin therapy for any non-tuberculous condition, because *Visiting Milan.
research
fellow,
on
release from
Gruppo Lepetit,
of the danger that this would encourage the emergence of rifampicin-resistant strains of Mycobacterium tuberculosis. There is reason to believe that this danger may have been exaggerated; for instance, during monotherapy of tuberculosis with rifampicin, resistant variants of this organism took at least a month to arise.’ Further, there is no evidence that the use of streptomycin during the 1950s for non-tuberculous conditions caused an increase in the frequency of strains of M. tuberculosis resistant to this antibiotic. In this paper we compare the development of primary resistance to rifampicin over the past few years, in strains of M. tuberculosis isolated in countries where the antibiotic is in use for non-tuberculous conditions, with figures from other countries where rifampicin is reserved solely for tuberculous infections. We have also made an alternative approach to the problem by searching for resistance among gram-negative bacteria from a hospital where rifampicin is widely used for the treatment of tuberculosis. METHODS
Resistance in M. tuberculosis Information on the occurrence of primary resistance (as defined, diagnosed, and described by Cannetti et al. 8) was collected from Italy, Argentina, Brazil, and Spain. Clearly the amount of rifampicin used is important, and in order to assess its usage we have defined the total consumption of rifampicin as the "use ratio". This means the use in tuberculosis, divided by the use for non-tuberculous conditions. The calculation has been made from information for the individual countries.Y
Resistance in
Gram-negative Bacteria organisms isolated from the Royal .
333 such Free Hospital and 100 consecutive isolates from the Brompton Hospital were tested. The disc technique (with a 30 jj.g rifampicin disc) as well as a precise determination of minimum inhibitory concentration was used to define resistance. RESULTS
M. tuberculosis Tableshows the frequency of M. tuberculosis with primary resistance to rifampicin. The data from Italy originate from six different regions widely separated, so
Mental Health NEWER CONCEPTS OF PSYCHIATRIC DIAGNOSIS AND BIOCHEMICAL RESEARCH ON MENTAL ILLNESS
J. P. BLASS J. F. MILNE Departments of Psychiatry and Biological Chemistry and the Mental Retardation Research Center, School of Medicine, University of California, Los Angeles, U.S.A. R. RODNIGHT
Department of Biochemistry, Institute of Psychiatry, University of London Biochemical studies of psychiatric patients have revealed many abnormalities but have failed to identify specific biochemical changes linked to the pathophysiology of the illnesses. Review of these data in the light of newer concepts of psychiatric diagnosis may help to clarify the relationship between biochemical abnormalities and disabilities. These concepts may also suggest useful directions for future research.
Summary
INTRODUCTION
THE identification of biological factors which contribute to the development of mental illness has been actively pursued for many years, and it is now well established that many patients inherit genetic predisposition to the development of psychoses.’ Since genes code for proteins, psychotic patients presumably inherit one or more abnormal enzymes or other proteins which alter their body chemistry in such a way as to contribute to the development of their illnesses. The body chemistry of psychotics has already been studied intensively.2 New discoveries in neurochemistry have been applied promptly to the study of mental illness. A good example is the study of the endorphins and related peptides which have been discovered in the past three years.3 4 Unfortunately, the search for a relationship between biochemical abnormalities and mental illness has been disappointing.2 Many abnormalities have been found in different groups of patients, but almost all of these have occurred in patients with very different clinical syndromes and even in some clinically normal peopled However, reexamination of these studies in the light of newer approaches to diagnosis and classification promises to be
rewarding. DIAGNOSIS AND CLASSIFICATION
classifications in Western medicine have between two poles.6 One view, historically associated with the Platonic school, formulates diseases as discrete entities, with the clinical manifestations in a given patient resembling more or less closely those in a "typical" patient. This view assumes that clinical criteria can define groups of patients who are biologically homogeneous, so that the appropriate physiological, anatomical, or biochemical criteria will each define the same group of patients. This approach was developed in detail by Thomas Sydenham in the 17th century
Diagnostic
tended
to move
and underlies the 19th-century classification of psychiatric disorders. It also underlies the search for "the cause" of schizophrenia-a search which has had little success. The second approach to classifying diseases is historically associated with the name of Hippocrates. It is more phenomenologic, viewing diseases as operational definitions of groups of patients, even though the groups thus created are biologically heterogeneous. This approach assumes that whereas clinical, physiological, anatomical, and biochemical measurements may correlate with each other statistically, the different criteria do not identify identical groups of individuals. The sets may overlap, but they are not congruent. From this viewpoint, a "disease" is an intellectual construct defined by the parameters which are chosen to define it. For instance, "schizophrenia" is a convenient way to refer to a particular type of human behaviour which can result from a number of different factors. The implication is that "schizophrenics" are a biologically heterogeneous group and that biochemical studies of these patients will be expected to turn up a variety of abnormalities, none of which is specific for the entity defined by clinical criteria alone. This is, of course, what has been found. The potential value of the hippocratic approach is supported by its value in other areas of medicine. In genetic diseases a single clinical phenotype can be associated with more than one genotype. Entities such as anaemia and pneumonia also do not have single xtiotogies. Indeed, when Scadding’ and Burrows8 began to apply computers and mathematical rigour to medical diagnosis, they stated explicitly that clinical diagnoses were operational definitions for biologically complex and diverse phenomena. The introduction of the problemoriented record reflects, among other things, the inadequacies of dealing with the practical complexities of diagnosis and management in terms of specific diseases. Over 10 years ago, the American Heart Association recommended diagnostic formulations which required separate statements about aetiology, anatomy, physiology, and disability-for example, rheumatic heart-disease with mitral stenosis, left-sided congestive heart-failure, and moderate disability. It seems unlikely that precise formulations for disorders of the brain will prove simpler than for disorders of the heart. METABOLIC ABNORMALITIES AND MENTAL DISORDER
In metabolic disorders known to be associated with psychiatric abnormalities, there is no simple and con-
relationship between the metabolic disorder and the clinical manifestations. Thus, a small proportion of patients with phenylketonuria have the syndrome of infantile autism, and another small proportion with typical hydroxylase-deficiency phenylketonuria appear clinically normal with normal intelligence.lO Some patients whose -tissues cannot hydrolyse nitrocatechol sulphate (and are therefore diagnosed as having metachromatic leucodystrophy) have hallucinations, ideas of reference, grandiosity, or bizarre behaviour." Although this disorder is now thought to be rare, we know of no systematic study of its frequency in psychiatric populations. When modified Watson-Schwarz tests for porphyria were done routinely in a university psychiatry service in Ohio, 35 patients with porphyria were found among stant
2500 consecutive admissions. 12 This contrasts with 2 among the previous 2000 admis-
porphyrics diagnosed
‘
739 among 1000 consecutive admissions to a nearby general hospital. The admission diagnoses of the 35 were the following: schizophrenia (10), chronic brain
sions and
none
syndrome (10), alcoholism (6), psychotic depressions (2), neuroses, personality disorders, and miscellaneous (7). Unfortunately, more detailed studies to identify the specific mutations of porphyrin metabolism in these patients were not done. Part of the clinical variability in patients with porphyrias may relate to the effects of an independent "modifier" mutation affecting steroid-hormone
breakdown. 13
Furthermore, the psychiatric manifestations of organic brain disease can sometimes be identical to functional psychosis.14 Indeed, the international classification of psychiatric disorders lists the absence of demonstrable brain disease as one of the criteria for the diagnosis of schizophrenia.6 IS The variable psychiatric manifestations of temporal-lobe epilepsy are well known,16 as are the various psychiatric syndromes which follow head injuryl7 and accompany brain damage in children.18 Metabolic or other biological abnormalities of the brain may predispose to the development of psychiatric disorders, but they are only a few of the factors which determine the form of the observed behavioural abnormalities. These are presumably influenced as well by cultural and other experiences of the individual, by epigenetic factors, and by modifier genes.
CRITERIA FOR IDENTIFICATION OF METABOLIC DISEASE
The analysis presented above suggests that techniques and criteria used to identify inherited metabolic abnormalities in medical disorders can be usefully applied in psychiatry. Such criteria have been outlined elsewhere. " 1. Analytical studies should indicate abnormal amounts of compound or group of related compounds much more frequently in psychiatric patients than in normal individuals. However, there need not be a one-to-one correspondence between the analytical abnormalities and a single behavioural diagnosis. (Many analytical abnormalities which would meet this criterion have already been described) 2. The rates of metabolism of these compounds should be abnormal while those of related compounds are normal. Demonstration of the metabolic abnormalities in vitro in biopsy specimens which were histologically normal would indicate that the abnormality was not due simply to tissue damage by a
the disease process. 3. Demonstration of a defect in cultured fibroblasts or other cultured cells would demonstrate that it was genetic rather than due to diet, drugs, or other external factors. After serial passages in culture, the material in the original explant will have been diluted several millionfold by new cell-growth, and cells from the patients and controls are exposed to identical conditions in culture. Thus, abnormalities which persist in culture may be assumed to be coded into the genetic information in the cells. 4. The metabolic abnormality should be shown to be due to aberration in a specific enzyme or other protein, if possible in tissue-culture. 5. Demonstration that the enzymatic abnormality was distnbuted among the patients’ relatives in accordance with the known genetics of the disorder would be a strong indication that the enzymatic aberration was inherited as a primary abnormality. For instance, in a disorder inherited as an autosomal recessive, the enzyme activities in the parents should be intermediate between those of the patient and controls. 6. Demonstration of a reasonable mechanism by which the an
observed abnormality could impair brain function would support its pathophysiological significance. 7. The most important evidence that a metabolic abnormality had a significant role in the pathogenesis of a psychiatric disorder would be if a specific diet or other regimen which normalised the affected patients’ body chemistry also benefited them clinically, in double-blind cross-over trials.
One might consider as an example how these criteria could be applied to the recent observation that a significant proportion of psychotic patients excrete abnormally low amounts of phenylethylamine, as do a small proportion of clinically normal people .20 Experimental studies have indicated that phenylethylamine may act as a neurotransmitter, and pathophysiological models have been
suggested for linking phenylethylamine deficiency
to
psychoses.21Thus,
criteria 1 and 6 listed above have been fulfilled. Several questions arise. Is the metabolism of phenylethylamine and phenylalanine abnormal in tissues of the affected individuals, and is the metabolism of other amines and aminoacids normal ? Can an abnormality in the metabolism of phenylethylamine be demonstrated in cultured cells from these patients? Is there an abnormality in phenylalanine decarboxylase, the enzyme which synthesises phenylethylamine from phenylalanine, or in some other enzyme of phenylethylamine metabolism? Can an abnormality be demonstrated in tissues from patients’ relatives? Finally, and most important, can the excretion of phenylethylamine be normalised-for instance, by a diet enriched in phenylalanine or phenylethylamine? If so, would such a diet benefit specifically those patients who have the abnormality in phenylethylamine metabolism, in double-blind cross-over studies? Similar questions could be asked about other metabolites found to be abnormal in other patients.5These questions can be answered with -the help of well-known techniques of clinical and biochemical in-
vestigation. AN INTEGRATED APPROACH TO MENTAL ILLNESS
The search for hereditary metabolic abnormalities which predispose to mental illness does not deprecate at all the value of other approaches to studying psychiatric disease. For instance, psychopharmacological studies are obviously important clinically. The responses of particular "target symptoms" to specific drugs may provide critical clues to pathophysiological mechanisms, even if not necessarily to aetiology. (It is worth noting that the response to antiseizure medication provides relatively little information about the origin of a seizure disorder, and the response to aspirin little about the aetiology of a fever.) Furthermore, it should be emphasised that the search for biological abnormalities in psychiatric patients in no way implies that all psychiatric problems derive from biological abnormalities of the brain. Nor does it minimise the importance of psychological treatment techniques even in patients with demonstrable organic disease. Psychological mechanisms clearly continue to operate in people with biologically abnormal brains. The search for metabolic abnormalities is only an attempt to define more precisely certain biological factors which limit function and adaptability in some psychiatric patients. Both genetic1 and biochemical5 studies suggest that such abnormalities are important in a large proportion of patients with psychoses. One may hope that eventually information from different types of studies can be
740
integrated into coherent descriptions of pathophysiological mechanisms in patients with psychiatric disorders. The approach suggested in this paper is in essence a restatement of Claude Bernard’s proposals for an "experimental medicine", phrased in more modern terminology and in relation specifically to psychiatry.22 It implies the division of patients with similar behavioural disorders into biological subgroups and explicitly rejects a search for single causes or mechanisms for such complex clinical phenomena as "schizophrenia" or "depression". Analogous studies have given very useful information in other areas of medicine. The usefulness of this approach in neuropsychiatry can be determined only by further observations. This research was supported in part by grants NF 6-76-50 from the National Foundation March of Dimes and HD-06576 from the N.I.C.H.D.
Requests for reprints should be addressed to J. P. B., Mental Retardation Research Center, U.C.L.A. School of Medicine, Los Angeles, California 90024, U.S.A.
REFERENCES 1. 2. 3. 4. 5.
Slater, E., Cowie, V. Genetics of Mental Disorders. London, 1971. Rodnight, R. Psychol. Med. 1971, 1, 353. Bloom, F., Segal, D., Ling, N., Guillemin, R. Science, 1976, 194, 632. Jacquet, Y., Marks, N. ibid. p. 634. Weil-Malherbe, H., Szara, S. I. Biochemistry of Functional and Experimental Psychoses. Springfield, Illinois, 1971. 6. Kendell, R. E. Role of Diagnosis in Psychiatry. Oxford, 1975. 7. Scadding, J. G. Lancet, 1967, ii, 877. 8. Burrows, B. Ann. intern. Med. 1975, 83, 419. 9. Ritvo, E. R., Ornitz, M., LaFranchi, S. Archs gen. Psychiat. 1968, 19, 341. 10. Brown, E. S., Waisman, H. A. Pediatrics, Springfield, 1967, 40, 247. 11. Müller, D., Pilz, H., Meulen, V. T. J. neurol. Sci. 1969, 9, 567. 12. Kaebling, R., Craig, J., Pasamanik, B. Archs gen. Psychiat. 1961, 5, 494. 13. Kappas, A., Sassa, A., Granick, S., Bradlow, H. L. in Brain Dysfunction in Metabolic Disorders (edited by F. Plum); p. 225. New York, 1974. 14. Slater, E., Roth, M. Clinical Psychiatry; p. 483. Baltimore, 1969. 15. International Statistical Classification of Diseases, Injury, and Death. W.H.O., 1967. 16. Baldwin, M., Bailey, P. (editors) Temporal Lobe Epilepsy. Springfield, Illinois, 1958. 17. Hillbom, E. Acta psychiat. neurol. scand. 1960, 35, suppl. 142, 128. 18. Graham, P., Rutter, M. Br. med. J. 1968, iii, 698. 19. Blass, J. P., Steinberg, D. in Biology of Brain Dysfunction (edited by G. Gaull); vol. 2, p. 239. New York, 1973. 20. Sandier, M., Reynolds, G. P. Lancet, 1976, i, 70. 21. Giardina, W. J., Pedemonte, W. A., Sabelli, H. C. Life Sci. 1973, 12, 153. 22. Bernard, C. Introduction à l’étude de la médicine experimentale. Pans, 1952.
Occasional
Survey
CAN RIFAMPICIN USE BE SAFELY EXTENDED?
Evidence for Non-emergence of Resistant Strains of Mycobacterium tuberculosis G. ACOCELLA*
J. M. T. HAMILTON-MILLER W. BRUMFITT
Royal Free Hospital,
Pond Street, London NW3 2QG
the incidence of primary resisrifampicin in Mycobacterium strains have been collected from various tuberculosis countries. Strains isolated from those countries where rifampicin is used for both tuberculous and non-tuberculous conditions (Italy, Argentina, Brazil, and Spain) did not show a higher incidence of primary resistance than did strains from other countries (France, U.K., and U.S.A.) where rifampicin use is confined to tuberculosis. It is concluded that there is no evidence to justify fears of an increased incidence of resistance to rifampicin in M. tuberculosis if rifampicin were used discreetly for treating non-tuberculous infections.
Summary
Data
on
tance to
INTRODUCTION
Kerry, Hamilton-Miller, and Brumfittl reported on the antibacterial activity of the combination rifampicin/trimethoprim. They found at least an additive action against a wide range of pathogens, as well as suppression of the emergence of rifampicin-resistant mutants. These results, together with those obtained in pharmacokinetic experiments23 and with an animal-infection model4 suggest that this combination might be useful for treating a wide variety of infections in man. However, some authorities’6 have declared themselves against rifampicin therapy for any non-tuberculous condition, because *Visiting Milan.
research
fellow,
on
release from
Gruppo Lepetit,
of the danger that this would encourage the emergence of rifampicin-resistant strains of Mycobacterium tuberculosis. There is reason to believe that this danger may have been exaggerated; for instance, during monotherapy of tuberculosis with rifampicin, resistant variants of this organism took at least a month to arise.’ Further, there is no evidence that the use of streptomycin during the 1950s for non-tuberculous conditions caused an increase in the frequency of strains of M. tuberculosis resistant to this antibiotic. In this paper we compare the development of primary resistance to rifampicin over the past few years, in strains of M. tuberculosis isolated in countries where the antibiotic is in use for non-tuberculous conditions, with figures from other countries where rifampicin is reserved solely for tuberculous infections. We have also made an alternative approach to the problem by searching for resistance among gram-negative bacteria from a hospital where rifampicin is widely used for the treatment of tuberculosis. METHODS
Resistance in M. tuberculosis Information on the occurrence of primary resistance (as defined, diagnosed, and described by Cannetti et al. 8) was collected from Italy, Argentina, Brazil, and Spain. Clearly the amount of rifampicin used is important, and in order to assess its usage we have defined the total consumption of rifampicin as the "use ratio". This means the use in tuberculosis, divided by the use for non-tuberculous conditions. The calculation has been made from information for the individual countries.Y
Resistance in
Gram-negative Bacteria organisms isolated from the Royal .
333 such Free Hospital and 100 consecutive isolates from the Brompton Hospital were tested. The disc technique (with a 30 jj.g rifampicin disc) as well as a precise determination of minimum inhibitory concentration was used to define resistance. RESULTS
M. tuberculosis Tableshows the frequency of M. tuberculosis with primary resistance to rifampicin. The data from Italy originate from six different regions widely separated, so
