Volume 70 February 1977
0,-1 Il5FEB
67
177
'3y' ,cc~~~~~~'
0~~
Editorials Penicillamine in the Treatment of Rheumatoid Arthritis In 1943, the name penicillamine was given to an amino acid BB-dimethylcysteine found among the acid penicillin hydrolysates (Abraham et al. 1943). It is now 21 years since J M Walshe first gave penicillamine to a patient with Wilson's disease (Walshe 1956). In the intervening years, penicillamine has found a place in the treatment of many diseases, including Wilson's disease, cystinuria (Crawhall et al. 1963), lead poisoning (Boulding & Baker 1957), and morphea (Moynahan 1973). It has been tried in many other diseases, sometimes with promising results, such as chronic active hepatitis (Stern et al. 1976), primary biliary cirrhosis (S Jain 1976, personal communication), and scleroderma (Bluestone et al. 1970). The drug has a variety of biological effects and it is evident that exploration of other applications may prove fruitful., Initially penicillamine was used in rheumatoid arthritis because it dissociates certain macroglobulins in vitro (Deutsch & Morton 1957). Evidence of its clinical efficacy was first published in 1964 (Jaffe 1964). In the next nine years the reports of many workers in uncontrolled studies (Zuckner et al. 1970, Huskisson & Hart 1972), confirmed this initial report. It was not until the Multicentre Trial Group published its results in 1973, that penicillamine was clearly shown to be effective in rheumatoid arthritis. This trial, which used a predetermined and rather high dose of penicillamine, compared its effect in 52 patients with severe uncontrolled rheumatoid disease with that of placebo capsules in 53 similar patients. The results of the trial were remarkably favourable in terms of improvement in clinical features, but the number of side-effects was prohibitive and 16 of the treatment group were withdrawn from the trial because of adverse reactions. The harsh realities of this first controlled trial have been tempered somewhat by later work. Penicillamine and gold have been shown to be equipotent in another Multicentre trial (Huskisson et al. 1974). In this study penicillamine possibly had the edge over gold, because fewer patients on penicillamine were withdrawn permanently. Penicillamine and azathioprine are probably also equally effective (H Berry 1976, personal communication). The introduction of the so-called 'go slow, go low' dosage schedule
has shown that more widely spaced increments and a lower maximum dose lead to fewer sideeffects (Eighth European Rheumatology Congress 1975, abstracts; Day 1974). No dosage regime has, however, lowered the incidence of proteinuria. Wilson's disease and cystinuria are lifethreatening conditions. The benefits of penicillamine in the treatment of these diseases easily outweigh its adverse effects. Rheumatoid arthritis, although incurable, painful, and sometimes crippling, is only rarely a fatal disease. It is clear that the introduction of a new long-term treatment for rheumatoid arthritis, with many toxic and sometimes potentially fatal side-effects, must be slow and cautious. Since the publication of the first controlled trial in rheumatoid arthritis, physicians and scientists interested in penicillamine have met annually to discuss the latest advances in treatment. These meetings have provided a forum for the frank exchange of information about the drug. The pooling of knowledge about side-effects and dosage regimes in particular has meant that penicillamine is still not widely used in the United Kingdom, and, perhaps, that some fatal drug reactions have been prevented. It is essential that this cautious approach to the use of the drug should continue, for pencillamine is not a routine treatment for rheumatoid arthritis, its mode of action is unknown, and there is as yet no published evidence that it prevents long-term changes in affected joints. All rheumatologists would agree that the initial treatment of rheumatoid arthritis should be simple and be based on analgesic, anti-inflammatory drugs, rest and splints. The chance of spontaneous remission in this group of patients is high in the first year or two of the disease and the majority of patients will be comfortable and able to lead a normal life by the end of this time. Where the disease follows a remorselessly aggressive course and encroaches increasingly on the patient's daily life, alternative treatment is indicated. The choice of treatment lies between gold, antimalarials, immunosuppressive drugs, and now penicillamine. Immunosuppressive drugs are on the whole less widely used than the other drugs. Antimalarial drugs are not so effective as gold or penicillamine, but their use is more easily controlled and serious side-effects are few. Some patients will respond to antimalarial drugs very well. Provided that ophthalmological control is available, some clinicians
68
Proc. roy. Soc. Med. Volume 70 February 1977
would put all the patients requiring long-term ageusia, and indigestion. Except in a few patients treatment on to these drugs for an initial period of with vomiting, most of the effects will disappear three months. Patients who do not respond would and the patient should be encouraged to soldier on then be considered for gold or penicillamine. with the drug. A variety of other rarer but importThere is no way at present of predicting the ant side-effects has now been encountered with response to either gold or penicillamine. The wider use of penicillamine. Some require imchoice of drug will depend upon local and personal mediate cessation of the therapy, some merely a factors, such as the availability of appropriate lowering of the dose. Such effects include a lupusback-up services, the clinician's personal ex- like reaction (Day & Golding 1974), mammary perience of the drugs, a previous history of drug hyperplasia, hemolytic anemia, bone marrow aprashes, eczema or penicillin sensitivity, and the lasia, mouth ulcers, Goodpasture's syndrome, patient's preference as between tablets and in- myasthenia gravis, late rashes, hirsutism and polyjections. Neither treatment should be embarked myositis. Seven fatalities directly attributable to upon without adequate laboratory facilities for the drug have been reported, and more cases will hmmatological and biochemical monitoring of undoubtedly arise. side-effects. The two most common and potentially serious The optimum dose of penicillamine has not yet side-effects are thrombocytopenia and proteinuria, been determined. The regime followed in the which are met with in about 30 % of patients who Multicentre Trial was an incremental dosage over are able to tolerate penicillamine for more than six ten weeks to a total of 1.5 g. Since then further months. Fortunately there is usually ample warnexperience has led to a gradual lowering of the ing of their development, namely a gradual slide in total daily dose and an increase in the time scale for the platelet count to bel6w 100 000 and a gradual increments. At present most clinicians start the increase in proteinuria to above 2 g per day. Both patient on 250 mg daily and introduce the first these levels are regarded by clinicians as the cut-off increment of 250 mg only after eight weeks. Fur- point for withdrawal of therapy. It is sometimes ther increments are made even more slowly and it is possible to stave off this necessity by cutting the seldom necessary to exceed 750 mg daily, except in daily dose when these side-effects first appear. the severer and more longstanding cases. Some Thrombocytopenia is usually quickly reversed on patients do remarkably well on only 125 mg daily. withdrawal of the drug, which can then sometimes It is sometimes possible to scale down the dose by be introduced at a much lower dosage. Proteinuria similar decrements after one year of continuous disappears more slowly, over the next one to two treatment. The aim of treatment should be to cut years. There have been reports of proteinuria out all other drugs except occasional analgesics by disappearing while the drug was continued, but the end of the first year of treatment. It has not so most clinicians would not advocate this, and few far proved possible to stop treatment without an have so far restarted patients on the drug who have eventual relapse. Further work is continuing to previously experienced this side-effect. determine whether a single daily dose of 250 mg The mode of action of penicillamine is unknown between meals is sufficient for the majority of and its absorption and metabolism poorly underpatients (I A Jaffe 1976, personal communication). stood. There is recent evidence that it is well As with all drugs, frequent monitoring of absorbed from the empty stomach (I A Jaffe 1976, patients is essential, and should be carefully ex- personal communication), and that other drugs plained before embarking on treatment. Weekly such as iron may interfere with its mode of action. blood and urine tests are required initially, with It is excreted as a number of metabolites fairly full platelet and differential white blood counts. rapidly, but some of the drug is tissue bound very After four weeks such monitoring can be carried strongly and is excreted only slowly over many out at fortnightly intervals. Once the patient is months. The idea that penicillamine acts because stabilized on his maximum dose, tests can be of its action on macroglobulins is now discounted. carried out at monthly visits, provided that the It is not an immunosuppressive drug (Huskisson & patient will undertake to test his own urine weekly Berry 1974). It has no effect on adjuvant arthritis in for blood and protein. Three monthly biochemical the rat (Liyanage & Currey 1972). Penicillamine tests of renal function are also required. It is inhibits certain viruses in vitro (Lodde & Marcialis important that the results of all tests are scrutinized 1974). There is no evidence that its chelating effect personally by the physician and not filed routinely. on various metals bears any relationship to its With such careful monitoring any side-effects clinical effect in rheumatoid arthritis. should be quickly reversible. Although penicillamine now has an established The initial side-effects of penicillamine, apart place in the treatment of rheumatoid arthritis, from rare cases of bone-marrow idiosyncrasy, are more widespread use of the drug in recent years has unpleasant but reversible and are not serious. highlighted rather than diminished its many adThese include nausea, vomiting, rash, dysgeusia or verse effects. It remains a difficult drug to use,
Editorials requiring careful monitoring and frequent visits by the patient to the clinician. There is certainly no place for further extension of its use at present. Any doctor wishing to use it must be prepared to commit himself to the exacting and timeconsuming supervision of the patient which may have to be continued for many years. If there is any doubt about the availability of spegialist back-up care, or the willingness of the patient to attend for regular tests, this treatment should not be started. On the evidence available at present it would be wise to adopt only the treatment regime of a low starting dose with graduated increments to a low maximum dose. Penicillamine treatment for rheumatoid arthritis is undoubtedly an advance but one tempered with many difficulties. A VIRGINIA CAMP
Consultant Physician Wycombe General Hospital High Wycombe, Buckinghamshire
REFERENCES Abraham E P, Chain E, Baker W & Robinson R (1943) Nature, London 151, 107 Bluestone R, Grahame R, Holloway V & Hort P J L (1970) Annals of the Rheumatic Diseases 29, 153 Boulding J E & Baker R A (1957) Lancet ii, 985 Bucknall R C, Dixon A St J, Glick E N, Woodland J & Zutshi D W (1975) British Medical Journal i, 600 CrawhalH J C, Scowen E F & Watts R W E (1963) British Medical Journal i, 588 Day A T (1974) Current Medical Research and Opinion 2, 581 Day A T & Golding J R (1974) Postgraduate Medical Journal Suppl. 2, p 71 Deutsch H F & Morton J I (1957) Science 125, 600 Huskisson E C & Berry H (1974) Postgraduate Medical Journal Suppl. 2, p 59 Huskisson E C, Gibson T J, Balme H W, Berry H, Burry H C, Grahame R, Hart F D, Henderson D R F & Wojtulewski J A (1974) Annals of the Rheumatic Diseases 33, 532 Huskisson E C & Hart F D (1972) Annals of the Rheumatic Diseases 31, 402 Jaffe I A (1964) Annals of Internal Medicine 61, 556 Liyanage S P & Currey H L F (1972) Annals of the Rheumatic Diseases 31, 521 Lodde B & Marcialis M A (1974) Postgraduate Medical Journal Suppl. 2, p 45 Moynahan E J (1973) Lancet i, 428 Multicentre Trial Group (1973) Lancet i, 275 Pass D, Goldfischer S, Steinlieb I & Scheinberg I H (1973) Archives of Dermatology 108, 713 Stern R B, Wilkinson S P & Williams R (1976) Gut 17, 390 Walshe J M (1956) Lancet i, 25 Zuckner J, Ramsey R H, Dorner R W & Gantner G E (1970) Arthritis and Rheumatism 13, 131
69
Too Many Hospital Referrals?' In all systems of health care there are certain inevitable, definable levels of care - self-care, primary professional care (general practice in the NHS), 'general' specialist care at a district hospital level and 'super' specialist care at regional units (such as thoracic surgery, neurology, ophthalmology &c). Each level of care has its interface with the level adjacent to it. It is at these interfaces that referrals of patients take place. If we are to be in a position to make better use of available resources, then the movements of persons seeking care between the four levels is of supreme importance. In the British National Health Service there is a single portal of entry into most medical care, that is through the general practitioner, with whom 98 % of the population are registered. Apart from the accident-emergency and the venereal diseases departments, hospital care can be obtained only through referral of a patient by a general practitioner to a hospital specialist unit. The use of our hospitals has more than doubled since the NHS was created in 1948, and still the trends are for more and more persons being referred by general practitioners to hospitals each year. The most recent information from the DHSS for 1974 shows that almost 12 % of our population are admitted to hospital each year, 16 % are referred as new cases to outpatient departments and almost 20 % take themselves (or are taken) to accident-emergency departments. Thus one-half of the population may be attending our hospitals in a year. There is also scant information on the reasons why general practitioners refer their patients to hospitals. There is a thirteen-fold range of difference, quoted in the Royal College of General Practitioners report 'Present State and Future Needs' (1973). Some general practitioners are quoted as referring 20 per 1000 to outpatient departments in a year and, at the other extreme, some refer as many as 260 per 1000. Why? This we do not know and it is urgently necessary that we find out. It may be that some practitioners refer too few of their patients to hospital but it is much more likely that many more refer too many of their patients. This has to remain an unproven hypothesis until studies are carried out to examine and analyse the hospital referral patterns of general practitioners in an area or district and the reasons for different rates of referrals. In my own practice (Fry J, 1971, Lancet ii, 148) the referral rate has been halved during 25 years. In 1951 I was referring 105 per 1000 of my patients to ' Based on Presidential Address by Dr John Fry to the Section of Medical Education, 10 November 1976
0,-1 Il5FEB
67
177
'3y' ,cc~~~~~~'
0~~
Editorials Penicillamine in the Treatment of Rheumatoid Arthritis In 1943, the name penicillamine was given to an amino acid BB-dimethylcysteine found among the acid penicillin hydrolysates (Abraham et al. 1943). It is now 21 years since J M Walshe first gave penicillamine to a patient with Wilson's disease (Walshe 1956). In the intervening years, penicillamine has found a place in the treatment of many diseases, including Wilson's disease, cystinuria (Crawhall et al. 1963), lead poisoning (Boulding & Baker 1957), and morphea (Moynahan 1973). It has been tried in many other diseases, sometimes with promising results, such as chronic active hepatitis (Stern et al. 1976), primary biliary cirrhosis (S Jain 1976, personal communication), and scleroderma (Bluestone et al. 1970). The drug has a variety of biological effects and it is evident that exploration of other applications may prove fruitful., Initially penicillamine was used in rheumatoid arthritis because it dissociates certain macroglobulins in vitro (Deutsch & Morton 1957). Evidence of its clinical efficacy was first published in 1964 (Jaffe 1964). In the next nine years the reports of many workers in uncontrolled studies (Zuckner et al. 1970, Huskisson & Hart 1972), confirmed this initial report. It was not until the Multicentre Trial Group published its results in 1973, that penicillamine was clearly shown to be effective in rheumatoid arthritis. This trial, which used a predetermined and rather high dose of penicillamine, compared its effect in 52 patients with severe uncontrolled rheumatoid disease with that of placebo capsules in 53 similar patients. The results of the trial were remarkably favourable in terms of improvement in clinical features, but the number of side-effects was prohibitive and 16 of the treatment group were withdrawn from the trial because of adverse reactions. The harsh realities of this first controlled trial have been tempered somewhat by later work. Penicillamine and gold have been shown to be equipotent in another Multicentre trial (Huskisson et al. 1974). In this study penicillamine possibly had the edge over gold, because fewer patients on penicillamine were withdrawn permanently. Penicillamine and azathioprine are probably also equally effective (H Berry 1976, personal communication). The introduction of the so-called 'go slow, go low' dosage schedule
has shown that more widely spaced increments and a lower maximum dose lead to fewer sideeffects (Eighth European Rheumatology Congress 1975, abstracts; Day 1974). No dosage regime has, however, lowered the incidence of proteinuria. Wilson's disease and cystinuria are lifethreatening conditions. The benefits of penicillamine in the treatment of these diseases easily outweigh its adverse effects. Rheumatoid arthritis, although incurable, painful, and sometimes crippling, is only rarely a fatal disease. It is clear that the introduction of a new long-term treatment for rheumatoid arthritis, with many toxic and sometimes potentially fatal side-effects, must be slow and cautious. Since the publication of the first controlled trial in rheumatoid arthritis, physicians and scientists interested in penicillamine have met annually to discuss the latest advances in treatment. These meetings have provided a forum for the frank exchange of information about the drug. The pooling of knowledge about side-effects and dosage regimes in particular has meant that penicillamine is still not widely used in the United Kingdom, and, perhaps, that some fatal drug reactions have been prevented. It is essential that this cautious approach to the use of the drug should continue, for pencillamine is not a routine treatment for rheumatoid arthritis, its mode of action is unknown, and there is as yet no published evidence that it prevents long-term changes in affected joints. All rheumatologists would agree that the initial treatment of rheumatoid arthritis should be simple and be based on analgesic, anti-inflammatory drugs, rest and splints. The chance of spontaneous remission in this group of patients is high in the first year or two of the disease and the majority of patients will be comfortable and able to lead a normal life by the end of this time. Where the disease follows a remorselessly aggressive course and encroaches increasingly on the patient's daily life, alternative treatment is indicated. The choice of treatment lies between gold, antimalarials, immunosuppressive drugs, and now penicillamine. Immunosuppressive drugs are on the whole less widely used than the other drugs. Antimalarial drugs are not so effective as gold or penicillamine, but their use is more easily controlled and serious side-effects are few. Some patients will respond to antimalarial drugs very well. Provided that ophthalmological control is available, some clinicians
68
Proc. roy. Soc. Med. Volume 70 February 1977
would put all the patients requiring long-term ageusia, and indigestion. Except in a few patients treatment on to these drugs for an initial period of with vomiting, most of the effects will disappear three months. Patients who do not respond would and the patient should be encouraged to soldier on then be considered for gold or penicillamine. with the drug. A variety of other rarer but importThere is no way at present of predicting the ant side-effects has now been encountered with response to either gold or penicillamine. The wider use of penicillamine. Some require imchoice of drug will depend upon local and personal mediate cessation of the therapy, some merely a factors, such as the availability of appropriate lowering of the dose. Such effects include a lupusback-up services, the clinician's personal ex- like reaction (Day & Golding 1974), mammary perience of the drugs, a previous history of drug hyperplasia, hemolytic anemia, bone marrow aprashes, eczema or penicillin sensitivity, and the lasia, mouth ulcers, Goodpasture's syndrome, patient's preference as between tablets and in- myasthenia gravis, late rashes, hirsutism and polyjections. Neither treatment should be embarked myositis. Seven fatalities directly attributable to upon without adequate laboratory facilities for the drug have been reported, and more cases will hmmatological and biochemical monitoring of undoubtedly arise. side-effects. The two most common and potentially serious The optimum dose of penicillamine has not yet side-effects are thrombocytopenia and proteinuria, been determined. The regime followed in the which are met with in about 30 % of patients who Multicentre Trial was an incremental dosage over are able to tolerate penicillamine for more than six ten weeks to a total of 1.5 g. Since then further months. Fortunately there is usually ample warnexperience has led to a gradual lowering of the ing of their development, namely a gradual slide in total daily dose and an increase in the time scale for the platelet count to bel6w 100 000 and a gradual increments. At present most clinicians start the increase in proteinuria to above 2 g per day. Both patient on 250 mg daily and introduce the first these levels are regarded by clinicians as the cut-off increment of 250 mg only after eight weeks. Fur- point for withdrawal of therapy. It is sometimes ther increments are made even more slowly and it is possible to stave off this necessity by cutting the seldom necessary to exceed 750 mg daily, except in daily dose when these side-effects first appear. the severer and more longstanding cases. Some Thrombocytopenia is usually quickly reversed on patients do remarkably well on only 125 mg daily. withdrawal of the drug, which can then sometimes It is sometimes possible to scale down the dose by be introduced at a much lower dosage. Proteinuria similar decrements after one year of continuous disappears more slowly, over the next one to two treatment. The aim of treatment should be to cut years. There have been reports of proteinuria out all other drugs except occasional analgesics by disappearing while the drug was continued, but the end of the first year of treatment. It has not so most clinicians would not advocate this, and few far proved possible to stop treatment without an have so far restarted patients on the drug who have eventual relapse. Further work is continuing to previously experienced this side-effect. determine whether a single daily dose of 250 mg The mode of action of penicillamine is unknown between meals is sufficient for the majority of and its absorption and metabolism poorly underpatients (I A Jaffe 1976, personal communication). stood. There is recent evidence that it is well As with all drugs, frequent monitoring of absorbed from the empty stomach (I A Jaffe 1976, patients is essential, and should be carefully ex- personal communication), and that other drugs plained before embarking on treatment. Weekly such as iron may interfere with its mode of action. blood and urine tests are required initially, with It is excreted as a number of metabolites fairly full platelet and differential white blood counts. rapidly, but some of the drug is tissue bound very After four weeks such monitoring can be carried strongly and is excreted only slowly over many out at fortnightly intervals. Once the patient is months. The idea that penicillamine acts because stabilized on his maximum dose, tests can be of its action on macroglobulins is now discounted. carried out at monthly visits, provided that the It is not an immunosuppressive drug (Huskisson & patient will undertake to test his own urine weekly Berry 1974). It has no effect on adjuvant arthritis in for blood and protein. Three monthly biochemical the rat (Liyanage & Currey 1972). Penicillamine tests of renal function are also required. It is inhibits certain viruses in vitro (Lodde & Marcialis important that the results of all tests are scrutinized 1974). There is no evidence that its chelating effect personally by the physician and not filed routinely. on various metals bears any relationship to its With such careful monitoring any side-effects clinical effect in rheumatoid arthritis. should be quickly reversible. Although penicillamine now has an established The initial side-effects of penicillamine, apart place in the treatment of rheumatoid arthritis, from rare cases of bone-marrow idiosyncrasy, are more widespread use of the drug in recent years has unpleasant but reversible and are not serious. highlighted rather than diminished its many adThese include nausea, vomiting, rash, dysgeusia or verse effects. It remains a difficult drug to use,
Editorials requiring careful monitoring and frequent visits by the patient to the clinician. There is certainly no place for further extension of its use at present. Any doctor wishing to use it must be prepared to commit himself to the exacting and timeconsuming supervision of the patient which may have to be continued for many years. If there is any doubt about the availability of spegialist back-up care, or the willingness of the patient to attend for regular tests, this treatment should not be started. On the evidence available at present it would be wise to adopt only the treatment regime of a low starting dose with graduated increments to a low maximum dose. Penicillamine treatment for rheumatoid arthritis is undoubtedly an advance but one tempered with many difficulties. A VIRGINIA CAMP
Consultant Physician Wycombe General Hospital High Wycombe, Buckinghamshire
REFERENCES Abraham E P, Chain E, Baker W & Robinson R (1943) Nature, London 151, 107 Bluestone R, Grahame R, Holloway V & Hort P J L (1970) Annals of the Rheumatic Diseases 29, 153 Boulding J E & Baker R A (1957) Lancet ii, 985 Bucknall R C, Dixon A St J, Glick E N, Woodland J & Zutshi D W (1975) British Medical Journal i, 600 CrawhalH J C, Scowen E F & Watts R W E (1963) British Medical Journal i, 588 Day A T (1974) Current Medical Research and Opinion 2, 581 Day A T & Golding J R (1974) Postgraduate Medical Journal Suppl. 2, p 71 Deutsch H F & Morton J I (1957) Science 125, 600 Huskisson E C & Berry H (1974) Postgraduate Medical Journal Suppl. 2, p 59 Huskisson E C, Gibson T J, Balme H W, Berry H, Burry H C, Grahame R, Hart F D, Henderson D R F & Wojtulewski J A (1974) Annals of the Rheumatic Diseases 33, 532 Huskisson E C & Hart F D (1972) Annals of the Rheumatic Diseases 31, 402 Jaffe I A (1964) Annals of Internal Medicine 61, 556 Liyanage S P & Currey H L F (1972) Annals of the Rheumatic Diseases 31, 521 Lodde B & Marcialis M A (1974) Postgraduate Medical Journal Suppl. 2, p 45 Moynahan E J (1973) Lancet i, 428 Multicentre Trial Group (1973) Lancet i, 275 Pass D, Goldfischer S, Steinlieb I & Scheinberg I H (1973) Archives of Dermatology 108, 713 Stern R B, Wilkinson S P & Williams R (1976) Gut 17, 390 Walshe J M (1956) Lancet i, 25 Zuckner J, Ramsey R H, Dorner R W & Gantner G E (1970) Arthritis and Rheumatism 13, 131
69
Too Many Hospital Referrals?' In all systems of health care there are certain inevitable, definable levels of care - self-care, primary professional care (general practice in the NHS), 'general' specialist care at a district hospital level and 'super' specialist care at regional units (such as thoracic surgery, neurology, ophthalmology &c). Each level of care has its interface with the level adjacent to it. It is at these interfaces that referrals of patients take place. If we are to be in a position to make better use of available resources, then the movements of persons seeking care between the four levels is of supreme importance. In the British National Health Service there is a single portal of entry into most medical care, that is through the general practitioner, with whom 98 % of the population are registered. Apart from the accident-emergency and the venereal diseases departments, hospital care can be obtained only through referral of a patient by a general practitioner to a hospital specialist unit. The use of our hospitals has more than doubled since the NHS was created in 1948, and still the trends are for more and more persons being referred by general practitioners to hospitals each year. The most recent information from the DHSS for 1974 shows that almost 12 % of our population are admitted to hospital each year, 16 % are referred as new cases to outpatient departments and almost 20 % take themselves (or are taken) to accident-emergency departments. Thus one-half of the population may be attending our hospitals in a year. There is also scant information on the reasons why general practitioners refer their patients to hospitals. There is a thirteen-fold range of difference, quoted in the Royal College of General Practitioners report 'Present State and Future Needs' (1973). Some general practitioners are quoted as referring 20 per 1000 to outpatient departments in a year and, at the other extreme, some refer as many as 260 per 1000. Why? This we do not know and it is urgently necessary that we find out. It may be that some practitioners refer too few of their patients to hospital but it is much more likely that many more refer too many of their patients. This has to remain an unproven hypothesis until studies are carried out to examine and analyse the hospital referral patterns of general practitioners in an area or district and the reasons for different rates of referrals. In my own practice (Fry J, 1971, Lancet ii, 148) the referral rate has been halved during 25 years. In 1951 I was referring 105 per 1000 of my patients to ' Based on Presidential Address by Dr John Fry to the Section of Medical Education, 10 November 1976
