how doctors and families will use the information and how
best to present it. The first indications are that in multiple
endocrine neoplasia type 2 there will be little demand for
.antenatal diagnosis because of the low morbidity of the
condition and the availability ofbiochemical screening leading
to effective surgical treatment. Children predicted to carry the
gene will probably be selected for earlier and more careful
biochemical screening. What is not clear is how accurate the
test must be before the clinician and the family are happy for
biochemical screening to be discontinued if the prediction is
Cancer Research Campaign Human Cancer Genetics Research Group,
University ofCambridge, Cambridge CB2 1QP
I Thakker RV, Plonder BAJ.
2 Mathew CGP, Chin KS, Easton 1)1F et al. A linked genctic marker for multiple endocrine nicoplasia
type 2a on chromosome 10. Nature 1987:328:527-8.
3 Simpson NE, Kidd KK, Goodfellow l'N, ctia. Assignment ot mtiltiple endocrine neoplasia tNpe 2A
to chromosomc 10 by linkage. Nature 1987;328:528-30.
4 Jackson CE, Norum RA, O'Neal LW, Nickolai TE, Delaney JP. Linkage between MNIEN 2B anid
chromosome 10 markers liniked to MNIEN 2A. AmJHum Genet 1988;43:A147.
5 Noll WWV, Bowden DW, Maurer l.H, et a. Gcnetic mapping of familial medullary carcinoma of ith
thyrioid/muiltiple endocrine ncoplasia type 2A with polymorphic loci on chromosomc 1(). Am]
Humt G;ceiit 1988;43:A29.
6 Wu J, (ioodt'cellow PN, Giuffra L.A,
polvimorphisms and linkage mapping studies based on non-CTEPH f'amilies. Cvttogenet Cell Genei
7 Nakaitira Y, Mathew CGP, Sobol H, et al. Linked markers flaliking the gene for multiplc
eiidocrine tieoplasia type 2A. Genomics 1989;5:199-203.
8 Betiham F. Hart K, Crolla J, Bobrow Ml, Francavilla Ml, Goodfellow PN. A method for generating
hy-brids econtaitting non-selected t'ragmcnts of'human chromosomes. Genomics 1989;4:509-17.
9 Telenius H, Mathew CGP, Nakamtira Y, ct al. Application of linked DNA markers to scrcening
families with multiple endocrine neoplasia type 2A. Fur] .Surg O)nciil (in press).
10 Sobol H, Narod SA, Nakamura Y, et al. Screening for multiple ciidocrinie neoplasia type 2A with
DNA-polymorphism analysis. N h'ngl,7 IMed 1989;321:996-11)1)1.
MLItiplIC Ciedocrine neoplasia. Baillieres Clin Endocrinol .Mleoab
et al. Fibronectin rcecptor beta subutiit locus: 1)NA
Antidepressant or cognitive therapy is effective
The eating disorder bulimia nervosa was first described 10
years ago. ' Since then more than 50 studies of its prevalence
have been conducted and there have been over 20 controlled
treatment trials. With a decade ofresearch behind us, it seems
timely to consider what has been learnt.
Initially the criteria used to diagnose the disorder were very
different in North America and in Britain.2 These differences
largely disappeared in 1987 with the introduction in the
United States of the DSM-III-R definition,3 which was
broadly similar to the more restrictive criteria already in use in
Britain. The initial differences undoubtedly contributed,
however, to the divergent views that were held over the
character, prevalence, and treatment of the disorder.
It is now agreed that bulimia nervosa has three key
features.' 2Firstly, there is loss of control over eating with
recurrent episodes of bulimia. Typically these episodes of
overeating occur in secret and are a source of shame and self
disgust. Secondly, there are extreme attempts to control
shape and weight; these include self induced vomiting, strict
dieting, and the misuse of purgatives and diuretics. Thirdly,
there are disturbed attitudes to shape and weight, which
resemble those found in anorexia nervosa. These attitudes are
central to the psychopathlogy of the disorder2 and have been
described as "a morbid fear ofbecoming fat"' or a "persistent
overconcern with shape and weight."3 Commonly associated
clinical features include a high level of general psychiatric
symptoms, particularly those associated with depressive
disorders, and impaired social functioning.
The physical features of bulimia nervosa have been less
extensively studied than those of anorexia nervosa. The most
common physical symptoms are fatigue, feeling "bloated,"
"puffy cheeks," irregular or absent menstruation, and tooth-
ache.4" Few abnormalities tend to be found on examination,
and body weight
Occasionally the salivary glands, particularly the parotid
glands, are enlarged, and this accounts for some complaints of
facial swelling. 7The pathophysiology of this hypertrophy is
unclear. In those patients who have vomited frequently for
some years there may be erosion of the dental enamel,
especially ofthe ligual surfaces ofthe teeth.' This accounts for
most complaints oftoothache. In addition, some patients who
have repeatedly induced vomiting by using their fingers to
stimulate the gag reflex have a characteristic distribution of
calluses on the dorsum ofthe hand.'
is usually within the normal range.
Various abnormalities may be found on laboratory testing.
About halfthe patients have electrolyte disturbance, the most
commonabnormalities beinghypochloraemia, hypokalaemia,
hyponatraemia, and a raised bicarbonate concentration.9 Very
occasionally these abnormalities are life threatening. They
result from the self induced vomiting and the misuse of
purgatives and diuretics. The nature of the disturbance
dependsonthepredominant behaviour: vomiting is associated
with metabolic alkalosis, whereas metabolic acidosis
characteristic of purgative misuse. Another abnormality
sometimes found on laboratory testing is a raised serum
isoenzymes. The explanation for this increase is unclear. With
the exception of severe electrolyte disturbance, none of these
physical abnormalities merits direct intervention: most
resolve rapidly in response to treatment directed at the eating
The diagnostic status of bulimia nervosa has been the
subject ofmuch debate. Is it a discrete psychiatric syndrome
or a variant of another psychiatric disorder? Some authors
have argued that bulimia nervosa is closely related to major
depressive disorder." This view is difficult to sustain in the
light of the mounting evidence of differences between the
disorders in phenomenology, distribution, familial transmis-
biological correlates, and course.'2 3 Others have
emphasised the association with substance abuse and "border-
line personality disorder,"'4 but these associations are found
in only a few patients.
undoubtedly a close relation is anorexia nervosa.2 Bulimia
nervosa and anorexia nervosa have very similar clinical
features, and between a third and a half of patients with
bulimia nervosa have met strict diagnostic criteria for anorexia
nervosa in the past, while many others have almost, but not
quite, met these criteria.' A case may be made for viewing
anorexia nervosa and bulimia nervosa as different expressions
of a single psychiatric disorder2; but until data are available
that allow the prognostic and therapeutic implications of the
two diagnoses to becompared (thereby testing their predictive
validity) it seems premature either to amalgamate the two
diagnostic concepts or to give one diagnosis precedence over
the other. This is the position taken in DSM-III-R,3 and as a
result some patients are eligible for both diagnoses-namely,
those who are both seriously underweight and have episodes
activity,'° usually with high
activities of both
16 The disorder with which there is
24 FEBRUARY 1990
Like anorexia nervosa, bulimia nervosa mainly affects
white women in Western societies. Most patients present
when they are in their 20s, though their history of disturbed
eating usually stretches back into adolescence. Typically the
breakdown of control over eating is preceded by a period of
extreme dieting and loss ofweight. Views on the prevalence of
the disorder among the group most at risk-adolescent girls
and young women-have varied over the years, but with
improved methods of case definition and detection most
studies now give a rate ofaround 1 -0%. 18-20 This figure should
not be accepted uncritically, however, as consistency must
not be confused with accuracy: the studies could all be subject
to equivalent sources of error.2' Also unclear is the extent to
which the cases detected in community surveys resemble
those seen in clinics. The few longitudinal studies of
community samples have found that bulimic symptomatology
varies over time with only a few subjects showing the
Views on the treatment ofbulimia nervosa are continuing to
evolve. In the United States, but not in Britain, psychiatrists
have shown enthusiasm for the use of antidepressant drugs,
and controlled trials have shown that these drugs are more
effective than placebo in reducing the frequency ofovereating
and the intensity of some of the other symptoms of the
disorder.2926 Interestingly, antidepressant drugs are equally
effective whether ornot the patient is depressed,2728 suggesting
that the decrease in bulimic symptoms might not be the result
of their antidepressant action. Instead,
expression of a direct effect of these drugs on the brain
mechanisms that control eating. Another possibility is that
the anxiolytic effects of antidepressant drugs help patients
resist the urge to overeat, at least in the short term.28 Two
further points about the effectiveness of these drugs are
important. Firstly, there is no evidence that they affect the
patients' disturbed attitudes to shape and weight or their
extreme attempts to diet.2930 Secondly, there have been no
studies of the maintenance of change after treatment with
antidepressant drugs, and the one anecdotal report was far
from encouraging.9' Until it has been shown that antidepres-
sant drugs have more than a selective and transitory effect,
their place in the treatment of bulimia nervosa must be
The main alternative approach to treatment is a specific
form of short term psychotherapy. This treatment is based
upon the cognitive view of the disorder, namely that the
characteristic attitudes to shape and weight of patients with
bulimia nervosa are ofprimary importance in maintaining the
condition. The treatment, a form of cognitive behaviour
therapy, is designed not just to change the patients' behaviour
but also to modify the disturbed attitudes to shape and
weight.32 It has been evaluated in a series of controlled trials,
and the results suggest that the immediate effects oftreatment
compare favourably with those obtained with antidepressant
drugs, with there being a substantial improvement in both
eating habits and attitudes to shape and weight.33 3 Psycho-
social functioning in general also improves. Significantly,
these changes seem to be maintained for at least the first year
after treatment.35 36 Several direct comparisons of antidepres-
sant drugs and cognitive behaviour therapy are currently
under way. The short term results of one of these studies are
available: group cognitive behaviour therapy was found to be
more effective than treatment with the antidepressant drug
imipramine, and adding imipramine to group cognitive
behaviour therapy did not significantly improve outcome.37
This decade of research on bulimia nervosa has seen
developments that are also relevant to the understanding and
treatmentofothereatingproblems. Forexample, thecognitive
it could be the
view of bulimia nervosa applies equally well to anorexia
nervosa38 and, with certain modifications, the cognitive
behavioural treatment for bulimia nervosa may be applied to
patients with that disorder.3 Interest in evaluating treatments
for bulimia nervosa has led to a more precise characterisation
of the behavioural and cognitive disturbances seen in these
patients and to the development of more refined assessment
techniques.4"4' These techniques could be applied to studies
ofthe clinical features and treatment ofother eating disorders.
It is to be hoped that over the next 10 years the intense interest
ofresearchers in bulimia nervosa and its treatment will extend
to the sister condition, anorexia nervosa, and to other related
eating and weight problems.
CHRISTOPHER G FAIRBURN
Wellcome Trust Senior Lecturer,
University Department ofPsychiatry,
Oxford OX3 7JX
I Russell GFM. Bulimia nervosa: an ominous variant of anorexia nersosa. Psychol Med 1979;9:429-
2 Fairburn CG, Garner DM. The diagnosis of bulimia nersosa. International Journal of Eating
3 American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed.
Revised.(DSM-III-R).Washington, DC: American Psychiatric Association, 1987.
4 Abraham SF, Beumont PJV. How patients describe bulimia or binge-eating. Psychol Med
5 MitchelliJE, Hatsukami D, Eckert ED, Pyle RL. Characteristics of275 patients with bulimia. Amj
6 Levin PA, Falko J, Dixon K, Gallup EM, Saunders W. Benign parotid enlargement in bulimia.
Ann Intern Med 1980;93:827-9.
7 Burke RC. Bulimia and parotid enlargement-case report and treatment. J Otolarvngol 1986;15:49-
8 SimmonsMS,Grayden SK, Mitchell JE. The need for psychiatric-dental liaison in the treatment of
btulimia. AmjPsychiatry 1986;143:783-4.
9 Mitchell JE, Pyle RL, Eckert ED, Hatsukami D, Lentz R. Electrolyte and other physiological
abnormalities in patients with bulimia. Psychol Med 1983;13:273-8.
10 Gwirtsman HE, Kave WH, George DT, Carosella NW, Greene RC, Jimerson DC. Hyper-
amylasemia and its relationship to binge-purge episodes: development of a clinically relevant
laboratory test. J Clin Psychiatrv 1989;50:196-204.
11 Pope HG Hudson JI. Neu, hope for btnge eaters. New York: Harper and Row, 1984.
12 Strober M, Katz JL. Do eating disorders and affective disorders share a common etiology? A
dissenting opinion.International.7ournaloj'EatingDisorders 1987;6:171-80.
13 Levy AB, I)ixon KN, Stern SL. How are depression and bulimia related? Am J Psvchiatrv
14 Lacey JH, Evans CDH. The impulusivist: a multi-impulsive personality disorder. Br j Addict
15 Turnbull J, Freeman CPL, Barry F, Henderson A. The clinical characteristics of bulimic women.
Internationaljournal ofEating Disorders 1989;8:399-409.
16 Pope HG, Frankenburg FR, Hudson JI, Jones YM, Yurgelun-Todd D. Is bulimia associated with
borderline personality disorder? A controlled study.J Clin Psychiatry 1987;48:181-4.
17 Russell GFM. The diagnostic formulation in bulimia nervosa. In: Garner DM, Garfinkel PE, eds.
Diagnostic issues in anorexia nervosa and bulimia nervosa. New York: Brunner/Mazel, 1988:3-25.
18 King MB. Eating disorders in general practice. BrMedj 1986;293:1412-4.
19 Schotte DE, Stunkard AJ. Bulimia vs bulimic behaviors on a college campus. JAMA 1987;
258: 12 13-5.
21) Johnson-Sabine E, Wood K, Patton G, Mann A, Wakeling A. Abnormal eating attitudes in London
schoolgirls-a prospective epidemiological study: factors associated with abnormal response on
screening questionnaires. Psvchol Med 1988;18:615-22.
21 Fairbiirn CG, Beglin SJ. The studies of the epidemiology of bulimia nervosa. Am J PsVchiatriy in
22 Drewnowski A, Yee DK, Krahn DD. Bulimia in college women: incidence and recovery rates. Am
J PsYchiatry 1988;145:753-5.
23 King MB. Eating disorders in a general practice population. Prevalence, characteristics and follow-
up at 12 to 18 months. Psvchol Med 1989;Monograph Suppl 14:1-34.
24 Striegel-Moore RH, Silberstein LR, Frensch P, Rodin J. A prospective study of disordered eating
among college students. Inernationaljournal ofEating Disorders 1989;8:499-5 10.
25 Freeman CPL, MunroiKM. Drug and group treatments for bulimia/bulimia nervosa.J Psychosom
26 Agras WS, McCann U. The efficacy and role ofantidepressants in the treatment ofbulimia nervosa.
Annals of Behavioral Medicine 1987;9:18-22.
27 Hughes PL, Wells LA, Cunningham CJ, llstrup DM. Treating bulimia with desipramine: a
double-blind, placebo-controlled study. Arch Gent Psychiatry 1986;43:182-6.
28 Walsh BT, Gladis M, Roose SP, Stewart JW, Stetner F, Glassman AH. Phenelzine vs placebo in 50
patients with bulimia. Arch Gen Psychiatry 1988;45:471-5.
29 Rossiter EM, Agras WS, Losch M. Changes in self-reported food intake in bulimics as a
consequence of anti-depressant treatment. International journal ofEating D)isorders 1988;7:779-
30 Mitchell JE, Fletchcr L, IPyle RL, Eckert ED, Hatsukami DK, Pomeroy C. Ihe impact of
treatment on meal patterns in patients with bulimia nervosa. International Journal of Easing
31 Pope HG, Hudson JI, Jonas JM, Yurgelun-Todd D. Antidepressant treatment of bulimia: a two-
year follosv-up study.JClin Psychopharmacol 1985;5:320-7.
32 Fairburti CG. A cognsitive behavioural approach to the managemcnt of bulimia. Psvchol Med
33 Wilson GT, Smith 1). Cognitive-behavioral treatment of bulimia nervosa. Annals of'Behavioral
34 Fairburn C(i. TIhe cturrent status of the psychological treatments for bulimia nervosa. J Psvchosom
35 Wilson GT, Rossiter E, KIcifield El, Lindholm L. Cognitive-behavioral treatment of bulimia
nervosa: a controlled evaluation. Behaz Res Ther 1986;24:277-88.
36 Fairburn CG, Kirk J, O'Connor M, Cooper l'J. A comparison of two psychological trcatmcnts for
bulimia nervosa. Behav Res I'her 1986;24:629-43.
37 Mitchell JE, Pyle RL, Eckert ED, Hatsukami 1), Poomeroy C, Zimmerman R. A comparison study
of antideprcssants and intensise structtired group psychotherapy in the treatment of butlimia
ncrvosa. Arch Gen PsYchiatrv (in press).
24 FEBRUARY 1990
38 Garner DMi, Bemis KMl. A cognitive-behavioral approach to anorexia nervosa. CognitiVe Therapv
and Research 1982;6: 123-50.
39 Garner DM, Bemis KM. A cognitive-behavioral approach to anorexia nervosa. In: Garner DM\,
Garfinkel PE, eds. Handbook of psychotherapy for anorexia nervosa and bulimia. New York:
40 Cooper Z, Fairburn CG. The eating disorder examination: a semi-structured interview for the
assessment of the specific psychopathology of eating disorders. International Journal of Eating
41 Cooper Z, Cooper PJ, Fairburn CG. The validity of the eating disorder examination and its
subscalcs. Brr Psychiatry 1989;154:807-12.
Research studies are essential
Over the next decade the organisation and delivery of health
care in the community will present the National Health
Service with one of its most demanding tasks. If it is to
succeed then studies and demonstration projects will need to
be organised and carried out in all parts ofthe country.
The 1990s will see growing demands for comprehensive
care and support of elderly, mentally ill, mentally handi-
capped, and physically disabled people either in their own
homes or in residential accommodation in the community. A
change is unavoidable-partly because of the increasing
numbers ofvery elderly people' and partly because of the fall
in the number ofschool leavers, fromwhom theNHS recruits
most of its staff.2 Only within the community are to be found
the involuntary and voluntary carers and helpers needed to
supplement scarce statutory services.
In November last year, the government published its long
awaited proposals for community care.3 The white paper
contained welcome emphasis on enabling people to live in
their own homes, support for carers, proper assessment of
need and good case management; and a real attempt to abolish
current confusion and to clarify the responsibilities of
agencies. Certainly it will help to remove several uncertainties
and allow planning to proceed. No one, however, should
minimise the task facing those authorities in making effective
arrangements for delivering the services.4
There will still be an enormous array of agencies and
staff with separate responsibilities and different loyalties.5
These include local social services, housing, and education;
health authorities; family practitioner committees; general
practitioners, both budget and non-budget holding; and
nurses working for health authorities, general practitioners,
hospitals providing outreach services, voluntary services, and
residential and nursing homes. They will have to collaborate
to ensure the best use ofresources and the effective delivery of
suitably integrated services.
Those in need of care also have to be identified. Their
problems and capabilities must be assessed and progress
predicted; and the frail elderly, those with chronic de-
generative disorders, and many others will need regular
reassessment. Welding together the various agencies and
disciplines to reach joint decisions and to collaborate in
exercising their responsibilities will require managerial skill
and diplomacy. And this will be against local backgrounds of
immense diversity in terms of geography, demography,
and the ways in which services have been organised and
Health and local authorities face two main tasks. Firstly,
they will need to establish new management structures
and mechanisms for coordinating the work of the agencies
providing (or purchasing) and delivering services. These will
need to be monitored, evaluated, amended, and developed.
This will be achieved only with the help of research,
underpinned by information systems. A recent innovative
study in Newcastle upon Tyne examined the provision ofcare
for physically handicapped and frail elderly people in the
agencies and staff.6 The study team concluded that even in a
city with excellent and well coordinated services the problems
were sufficiently complex that responsibility should be
devolved to a new, free standing, publicly accountable
consortium formed by the health authority and the city
council. Such a scheme might be regarded as at one end of a
spectrum of options open to authorities in the light of the
white papers on the health service7 and on community care.3
Another different and simpler approach adopted by some
district health authorities in trying to unify services has been
to introduce "advocates," whose role is to secure services and
coordinate their delivery directly on behalf of individual
The second main requirement is to develop outcome
measures in community care, which at present are seriously
deficient; among these must be assessment of the quality of
life and patient satisfaction. Only in this way can the
effectiveness of services be assessed.
If we are to achieve soundly based developments in
community care-bearingin mind thewide local diversities-
the prime requirement is for district health authorities
and local social service authorities to conduct studies and
experiments. Such studies must flourish independently of
priorities determined regionally or centrally. The authorities
must be encouraged and helped, and in particular they must
allocate or obtain sufficient finances to fund research. Such
studies may be undertaken by their own staff, by academic
departments, or by management consultants. There will be
problems (in care and service delivery, for example) that are
better examined on a multidistrict or regional basis, so a
regional research capability will still be needed. And there will
be problems with multiregional or national dimensions.
Because government departments must respond to the
requirements of ministers, the final tier needed may be
an independent authority on the lines recommended by
the Lords Select Committee9 or an institute of health as
recommended by Sir Kenneth Stowe.'° Though the govern-
ment's response to the former is less than encouraging."I
We should, then, expect to see a growth of monitoring of
developments, exchange ofinformation, and dissemination of
good practice. But the most urgent need is to ensure that the
authorities directly concerned with the development and
delivery of local services can command studies and research
intotheirown locallydefined problems. Thisposes aparticular
challenge for public health doctors. A second pressing need
is a much closer relation between the health service, academic
departments, and funding agencies to develop research
methods and measures and to train researchers.
Nuffield Provincial Hospitals Trust,
London NW1 7SP
24 FEBRUARY 1990