In Response:
The Biological Basis and Treatmen t
of Bulimia
Jeffrey M. Jonas, M.D.
Th e recent papers in th e J efferson J ournal of Psychia t r y b y Drs. Le vin
("Bulimia as a Masturbatory Equivalent, " Jul y, 198 5) a nd Wil son ("A Discussion
of ' Bu lim ia as a Ma sturbatory Equi valent' ," Winter, 1986) co ntai n a num be r of
th eories and recommendations abo u t th e e tio logy a nd treatm en t of b ulimi a
whi ch shou ld be ad d resse d. In large measure , both papers pr ese nt a d yn ami c
vie w of this disorder. H o we ver , psych odynami c fo r m u la t io ns are bu t one of
man y theories put forward to e xp lain this di sorder. Other t heories of etiolo gy
in clude notions that bulimia a r ises from cu ltu ra l fac tors and an in creased
a tten tio n to body image , from di so rde rs with in th e fa mi ly as a whole , or from
ass o rted behavioral fac to rs wh ich so me how r einforce binge-ea ti ng a nd pu rgm g.
One wa y to e valuat e a theory o f e tio logy is to as k wh ethe r t reatme n ts ba sed
upon a giv en theory ben efit individuals with th e disorde r in questio n . Fo r
e xa m p le , it would be of interest to know what e vidence e xists that individ ua l
psychoth erapy is superior to group psychotherapy for th e treatm en t of b ul im ia .
Similarly it is important to know what e m p ir ica l data underlie th ese theories. In
th e case o f bulimia, as noted above , th ere are a h ost o f id eas a bout wha t causes
the di sorde r. And whil e a review of this literature is be yond th e scope of this
paper, at this time th ere is no systematic, controlle d, non-an ecd ot al evidence
that psych odynamic, beha vioral , o r cu ltura l factors a lo ne ca n lead to b ul im ia
(1,2) .
In terms of treatment, we h av e a sim ila r situa t io n, wh ere co ntro lled st udies
of individual psychoth erapy, group th erapy, and behavior th erapy a re a lso
lacking. There are numerous un contro lled r eports of th e efficacy of th ese
modal ities in bulimic pati ents, but it is important to note wh y co ntro lled studies
are important in assessing treatm ents of this dis order. First , bulimia may remit
sp o n ta ne o us ly ( I), so that a placebo group is needed to eva lua te if a giv en
th erapy is e ffect ive . Second, bulimic pati ents ma y hav e a mel iora tio n of sym ptom s in str uc tu red se tt ings. In assessing th e fir st 40 consecuti ve inpatients
J effrey M. J onas, M.D., is director of the Eating Disord ers Prog ram at Fair Oaks Hospital,
Summit, New J ersey.
78
IN RESPONSE
79
admitted to the Fair Oaks Hospital Eating Disorders program , 33 h ad complete
cessation of binging within the first 10 da ys. Such a response is misl eading . At
the end of two weeks (before treatment is completed) nearly tw o-th irds of the 33
"responders" will binge and purge on their first pass, and nearly all of the 33
"responders" will report urges to binge during the first two weeks. T his kind of
phenomenon points out the need for a control group, as wel l as follow-up
information, before a given treatment can be sa id to be effecti ve.
Given the prevalence of bulimia, there are relativel y few stu dies of psychological and group treatments. One difficulty often cited in d esign in g studies is
that response to therapy cannot be assessed in a systematic ma nner. I n bulimia,
this is not the case. Bulimic patients have a number of ov ert sym p to ms which can
be followed readily, such as the number of binge and purge e p isodes each day ,
severity of binge and purge urges, and se ve r ity of depression . T h us , t here seems
little reason why studies of therapeutic efficacy of various modalities could not
be undertaken . Yet data supporting the use of many p opular th erapi es a re scanty
at best. There are at this time no co n tro lled studies of individual psych othe ra py
in bulimia beyond uncontrolled cas e se r ies . Cogniti ve beha vio ral th era py has
been reported successful in two uncontrolled studies (3 ,4) , wh ereas th e rapy has
been reported useful in several uncontrolled reports (5-9).
We are not advocating that individuals with bulimia be d eprived of group or
individual therapy. Rather, we are pointing out that at th is time th e re is litt le to
commend one talking therapy from another. Recommendations for group and
individual therapy should be made on an individual basis, rather than on the
basis of diagnosis alone. In conjunction with such treatment, howe ve r , we must
now address the use of a different therapeutic technique which appears applicable to this diagnostic category as a whole-the use of a n tidep ressa nt med ication.
The rationale for the use of antidepressants in bulimia co mes from th e bod y
of evidence linking bulimia and anorexia nervosa to affecti ve di sorders. T his
evidence stems from three lines of research: studies of phenomenol ogy, family
history , and biological tests. Studies of phenomenology have examined groups
of bulimic patients seeking treatment for their eating disorde r. Su ch studies
have consistently reported that bulimic patients have an in creased life ti me
prevalence of major depression and bipolar disorder, and that in up to 50
percent of cases the affective disorder antedates the onset of th e ea ting di sor de r
(10-12) . Studies of family history also suggest a biological link between affect ive
disorders and bulimia, in that most but not all groups have reported an in creased
incidence of affective disorders in the families of patients with bulimia ( 13- 15).
The third line of evidence linking eating disorders to affecti ve di sor d e rs
arises from studies of the dexamethasone suppression test (DS T) a nd th yrotr opin releasing hormone stimulation test (TRHST). Both the DST (14 ,16 ,17) and
TRHST (14) yield positive results in bulimia at rates co m parab le to those
observed among individuals with depression. These r esults sho u ld be inte rpreted with caution, since metabolic stresses a ttr ib u tab le to binging, p urging ,
IN RESPONSE
79
admitted to the Fair O a ks Hospital Eating Disorders program , 33 had complet e
cessation of binging within the first 10 days. Such a response is mislead ing. At
the end of two weeks (before treatment is completed) nearl y two-thi rds of th e 33
" r espo nd ers " will b inge and purge on their fir st pa ss, and nearly a ll of the 33
"responders" will report urges to binge during the first two week s. T h is kind of
phenomenon points out the need for a control group, as we ll as follow-up
information , before a given treatment can be said to be effec tive.
Gi ven the prevalence of bu limia, there are relativel y fe w studies o f psych olog ical and group treatments. One difficulty often cit ed in d esigning studies is
that re sponse to therapy cannot be assessed in a syste mat ic m ann er. In bu limi a ,
this is not the case. Bul im ic patients have a number of o vert sym ptoms wh ich can
be followed readi ly, such as the number of binge and purge e pisodes ea ch day,
severity of binge and purge urges, and severity of depression. Thus, th ere seems
little reason why stud ies of therapeutic efficacy of various modalities co uld not
be u nd e r tak en . Yet d ata supporting t he use of many popular th e rapies are scanty
at best. There are at this time no controlled studies of individual psych otherap y
in bulimia beyond uncontrolled case series. Cognitive behavi oral th erapy has
been reported successful in two uncontrolled studies (3, 4), wh ereas th erap y ha s
been reported useful in several uncontrolled reports (5-9).
We are not advocating that individuals with bulimia be d epri ved of group o r
individual therapy. Rather, we are pointing out that at this time th ere is little to
co m me nd one talking th erapy from another. Recommendations fo r group and
individual therapy shou ld be made on an indi vidual ba sis, rathe r th an o n th e
ba sis of diagnosis alone. In conjunction with such treatment, how e ve r , we must
now address the use of a different th erapeutic technique whic h a ppears a pp licable to this diagnostic catego r y as a whole-th e use of antidepressa nt medication .
Th e rationale for the use of antidepressants in bulimia co mes from th e body
of evidence linking bulimia and anorexia nervosa to affec t ive di sorders. This
evidence stems from three lines of research : stud ies of ph en omenology, fam ily
history , and biological tests. Studies of phenomenology have exam ined group s
of b ulim ic patients seeking treatment for their eating di sorder. Su ch st udies
have consistently reported that bu limic patients have an in creased life time
prevalence of maj o r depression and b ipolar disorder, and that in up to 50
percent of cases th e affective disorder an tedates the onset of th e eating di sorder
(10- 12) . Studies offam ily h istor y a lso suggest a biological link between affect ive
di sorders and bul imi a , in that most bu t not all groups have reported an increased
inci dence of affective disorders in the families of patients with bulimia (13- 15).
The third line of evidence lin king eating di sorders to affec tive d iso rd ers
arises from studies of the dexamethasone suppression test (DST) and th yr o tropin releasing hormone stimulation test (T RHST). Both th e DST ( 14 , 16, 17) and
TRHST ( 14) yiel d positive results in bulimia at rates co mpara b le to those
observed among individuals with depression . These results shou ld be int erpreted with caution, since metabolic str esses attr ib u ta ble to bing ing , purg ing,
80
JEFFERSON JO URN AL OF PSYCH IAT RY
and weight loss could produce false positive results (18). H o wever, st udies of
inpatient bulimics have shown that the rate of dexamethasone non-su pp r esio n is
not affected by hospitalization , even after binging and purging are co ntrolled
through milieu treatment (16,19).
The evidence summarized above suggests a biologica l link betwee n
bulimia and the affective disorders. It should be noted that th ere a re aspects of
th e eat ing disorders not explained b y this linkage, suc h as th e rel at ionsh ip of
eating disorders and substance abuse (20). However, th e affecti ve model of
eating disorders has led to an important treatment o f bulimia , th e use of
antidepressant medication.
Numerous open-studies of antidepressants in bulimia hav e a ppeared in the
literature since 1977 (1,2), describing the use of heterocyclic a n tidepressants
and monoamine oxidase inhibitors (MAOI's). Since th en , five pla cebocontrolled, double-blind studies of heterocyclic antidepressants or MAOI' s have
been completed (21-25). One study using low dosages of mianserin had negative
resu lts (21), and one study using amitriptyline had weakly positive results (22) .
Both studies appeared to suffer from the use of inadequate dosages of antidepressant. The other three studies of antidepressants reported strongly positive
findings. Pope et al. (23) observed a 70 percent reduction o f b in ge-eat in g in
patients taking imipramine in dosages up to 200 mg each d ay, as co mpared to
vir t ua lly no change in a placebo-control group. Significant redu ctions in depressio n , food preoccupation , and subjective global improvement we re also noted .
Walsh et al. had similarl y positive re sults using phenel zin e (23), as d id Hugh es et
al. using desipramine (24). The study by Hughes and co lleagues was of particu lar
interest because ·they selected bulimic patients who did no t d isplay major
d epression. Despite this fact , Hughes' group found that 15 (68 %) of 22 subjects
experienced a remission of th eir bulimic sym p to ms within ten wee ks. Another
important finding of the Hughes study was that pla sma le vels of d esip ram ine
comparable to those required in depression were required fo r res po nse, and that
dosages up to 350 mg of desipramine were required to obtain adeq ua te serum
levels (125-275 ug /ml in their laboratory).
Overall, a growing body of evidence supports th e use of a nt idep ressant
medication in the treatment of bulimia, whether or not th e patient suffers from
co nco m itant major depression. Follow-up data from th e study o f Pope e t al.
further support the utility of antidepressants (26). After o ne to two yea rs o n
medication, 50 percent of patients were free of bulimic symptoms, 4 5 percent
had a greater than 50 percent reduction of symptoms, and only one pa tien t, who
had discontinued medication , was unchanged. These promisin g data are tempered by the fact that complicated pharmacologic treatments were requi red in
so me ca ses in order to achieve maximal results. Nevertheless, a majority of these
patients were able to return to full function , and did so without th e aid of other
ancillary therapies.
A number of caveats should be rai sed at this point. First, age nts o ther than
IN RESPONSE
81
antidepressants have been useful in bulimia , including lithium (27), and meth ylamphetamine (28) . Naltrexone has a lso been reported to be of use in these
patients (29), but only on an o pen-label basis. In ge neral, treatment with
th ymoleptics must be syste m atic and care fu l utilizing medi cation in adequate
dosages for adequate pe ri ods of time , ca refu lly monitorin g seru m levels (30).
Second, medication does not interfere with psych otherapy. If anything the
co gn itive improvement whi ch patients exper ie nce wh en d epressio n and food
preoccupation lessen make th em more amenable to th erapy.
Third, the notion th at patients sho uld be allowed to " e xper ie nce " thei r
eat ing di sorder in order to achieve insight should be co ns idered as archa ic and
cr uel as notions that antidepressants sho uld be withheld from d epressed individual s for like reasons. Such an approach is also medicall y d an ge rous, in light o f
the ph ysical damage whi ch may attend constant binging and purging . The role
of treatment is to hasten recovery, and antidepressant th erapy ca n hel p in this
regard.
Finall y, clinicians often misinterpret the use of a nt idep ressa nts as a de facto
state me n t against the use of ps ychotherapy. This is untrue. Many if not most
patients with bulimia will have individual and famil y problem s whic h require
indi vidual , group, and family psychotherapy. In our o wn treatmen t p rogram ,
patients are e ngaged in all of th ese th erapies, in addition to mi lie u, n u tritional,
and exe rc ise therapy. In treating the whole patient, one needs to address
biological as well as psychological needs. In so d oing , t he use of thymoleptics
pl ays an important role. Given th e evidence fo r th e efficacy of antidepressants in
bulimia , th ere is little rea son to withhold th em from pat ie nt s with t his disorder.
[ Ed .: For Dr. Wilson' s reply, see the Letters to the Ed itor section.)
REFER ENCES
I. Pope HG, Hudson JI: Ne w Hope fo r Binge Eat ers: Advances in the Understanding
and T reatmen t of Bulimia . Ne w York , Harper a nd Ro w: 1984
2. Pope HG, Hudson , Jonas JM , et a l: T reatment o f bulimia wit h a nti depressants: A
research update , in Ph ysiol ogy and Treatme n t of Norma l Weigh t Wom en With
Bulimia; Edited b y Kaye W , Gwirtsman HE. Washington D.C.: Ame rica n Psych iatr ic
Association , 1985: 120 -1 31
3. Fa irburn C: A cognitive behavioral a pproac h to th e treatment o f bulimia . Psychol Med
11:707-711 ,1981
4 . Schneid er JA, Agras WS: A co gn itive beh avi oral gro up treatment of buli mia . Br J
Psychiatry 146 :66-69 , 198 5
5 . Boskind-Lodahl M, White WC : T he definition and trea tm ent of bul imarexia in
co llege wom en-a p ilot stud y. JAm Coll H ealth Assoc 27 :85-86, 19 78
6. White WC , Boskind-White M. An experiential-beh avioral app roach to the treatment
ofbulimarexia. Psychothe rapy: Theory, Research , and Practice . 1981 ; 4 :501-507
7 . Johnson CL , Connors M, Stuck ey M: Short-term gro up treatment of bul imia : A
preliminary r eport. Int J Eat Dis 2: 199-208 , 198 3
82
JEFFERSON JO URNAL OF PSYCHIATRY
8 . Lacey JG: Buli m ia nervosa, binge eating, and psychogenic vo mi ti ng: A controlled
treatment study and long-term outcome. Br Med] 286: 1609-161 3, 1983
9 . Conners M, Johnson CL, Stuckey M: Treatment of bulimia with brie f psych oed ucational group therapy. Am] Psychiatry 141 :15 12- 15 16 , 1984
10. Hudson Jl , Pope HG, J o nas JM, et al: Phenomenologic relati onship of eating
disorders to major affective disorder. Psychiatry Res 9:345-354, 198 3
II. Hatsukami DK , Eckert ED, MitchellJE, et al: Affective disorde r and substa nce abuse
in women with bulimia. Psychological Medicine 14 :701-704, 1984
12 . Herzog DB: Are anorexi c and bulimic patients depressed? Am] Psychiatry 141 : 15941597,1984
13 . Hudson JI , Pope HG, Jonas JM , et al: Family history stud y of ano rexia nervosa and
bulimia. Br] Psychiatry 142: 133-138, 1983
14 . Gwirtsman H, Roy-Byrne P, Yager J , et al: Neuroendocrine ab no rmalities in bulimia .
Am] Psychiatry 140 :559-563 , 198 3
15 . St eern SL, Dixon KN, Nezmer E, et al : Affective disorder in th e fa milies of wom en
with normal-weight bulimia. Am] Psychtry 141: 1224-1227 , 1984
16 . HudsonJl , Pope HG,JonasJM, et al: H ypothalam ic-pituitary-ad renal axis hyperactivity in bulimia. Psychiatry Res 8 :111-117 , 1983
17 . Mit ch ell JE, Pyle RL, Hatsukarni D, et al: The d examethasone sup pression te st in
pati ents with th e bulimic syndrome.] Clin Psychiatry 45:508- 511 , 1984
18 . Edelstein CK, Roy -Birne P, Faw zy FI, e t al: Effect s of we ight loss on the dexamethason e suppression te st. Am] Psychiatry 140:3 38-341 , 198 3
19. Kay WH, Gwirtsman HE, Ebert MH, et al: Neurobiol ogy of b ingin g and vomiting
behavior. Presented at th e Annual Meeting, Am eri can Psych iatr ic Association, May
10 ,1984
20. Jonas JM , Gold MS: Naltrexone treatment o f bulimia: Clinical and theoretical
find ings linking eating disorders and substance abuse. Adv Alcohol and Substance
Abuse (in press).
2 1. Sabine EJ, Yonace A , Farrington AJ , et al : Bulimia nervosa : A placebo-con trolled
double-bl ind th erapeutic trial of mianserin. Br] Clin Pharmacol 15: 195s- 20 2s, 1983
22. Mit ch ell JE, Groat R: A placebo-controlled, double-blind tri al of a mi triptyline in
bulimia.] Clin Psychopharm 4 :186-193 , 1984
23 . Pope HG, Hudson Jl, Jonas JM , et al: Bulimia treated with im ipramine : A place boco ntrolled double-b lind study. Am] Psychiatry 140: 554- 558 , 198 3
24 . Walsh BT, St ewart JW, Roose SP , et al: Treatment of bulimia with ph enel zine : A
double-b lind, placebo-controlled study. Arch Gen Psychiatry 41 : II 0 5-11 09 , 1984
25 . Hughes PL, Wells LA , Cunningham CJ , et al: Treating bulimia with d esipr amine : A
placebo-controlled double-blind study. Arch Gen Psychiatry 4 3: 182-186 , 1986
26. Pope HG, Hudson J l , Jonas JM , et al : Antidepressant treatment o f bulim ia: A
two-year follow-up.] Clin Psychopharm 5:320-327, 1985
27 . Hsu LKG: Treatment of bulimia with lithium . Am] Psychiat ry 141 : 1260-1 26 2 , 1984
28 . Ong YL , Checkely SA , Russell GFM : Suppression of bulim ia sym ptoms with
methylamphetamine. Br] Psychiatry 143 :288-293 , 1983
29 . J onas JM , Gold MS: Naltrexone reverses bulimic sym ptoms . Lancet 1:807 , 1986
30 . Pope HG , Hudson J l , Jonas JM : Antidepressant treatment of bul imi a : Preliminary
exper ience and practical recommendations.] Clin Psychopharmacology 3:274-28 1,
1983