Original Author: Paul Perry, Ph.D, BCPP
Latest Reviser: Vicki Ellingrod, Pharm.D., BCPP
Creation Date: 1996
Peer Review Status: Internally Peer Reviewed
The eating disorders, anorexia nervosa and bulimia nervosa, have received increased scientific and public interest during the past decade. The prevalence of anorexia nervosa is approximately 1%, and bulimia nervosa is 5% to 10% (Crisp et al 1976, Jones et al 1980, Pope et al 1984).
The pharmacological treatment of these disorders is relatively new. There is more pharmacotherapy data about the drug treatment of bulimia nervosa than anorexia nervosa. There are still many unanswered questions about medication treatment of these two entities. For example, are antidepressants the drugs of choice? If so, is one antidepressant better than another? What is the optimal dose of these medications? How long should a therapeutic trial last before switching to another medication? Are combinations of medications helpful? Will the newer serotonin uptake blockers be more effective than the older antidepressants? How long should maintenance therapy last? While many unanswered questions exist, there are some data to assist the clinician in the treatment of these challenging patients.
ANOREXIA NERVOSA
Anorexia nervosa may result from the coalescence of several etiological factors, one of which may be a biological abnormality involving a serotonin. Serotonin plays an important role in the loss of appetite or inhibition of food intake (Breisch et al., 1976; Blundell, 1984a; Blundell, 1984b; Leibowitz and Papadakos). Injection of serotonin into the paraventricular nucleus inhibits feeding in animals (Leibowitz and Papadakos, 1978). Serotonin agonists such as m-chlorophenylpiperazine (m-CPP). a metabolite of trazodone, or agents that may indirectly facilitate serotonergic transmission such as fenfluramine cause a loss of appetite (Blundell, 1984a). In animals, intracerebral injection of p-chlorophenylalanine, an inhibitor of serotonin synthesis, depletes serotonin and increases appetite (Breisch et al. 1976).
Serotonin activity is affected by all of the medications that have shown some therapeutic benefit in treatment of anorexia nervosa (e.g., cyproheptadine, lithium, antidepressants). At present it is not known how the medications that effect improvement in anorexia nervosa work, but their effect on serotonin activity is one possibility.
A variety of other antidepressants have been used to treat anorexia nervosa in open labeled trials. Some of these medications appear to be helpful to some patients (Hudson et al, 1985). Some authors have suggested that anorexia nervosa patients with coexistent depression might be more likely to respond to antidepressants (Hudson et al, 1985). However, other researchers cautioned against using the presence of depression as a criterion for administering antidepressants early in the course of treatment, believing these symptoms often remit with weight gain (Eckert and Mitchell, 1989).
Medications should be part of comprehensive treatment strategy that may include interventions such as cognitive-behavioral therapy, family therapy, educational sessions , and dietary management (Eckert and Mitchell, 1989). Two recent articles review these non-pharmacological treatments for anorexia nervosa (Garner and Garfinkle, 1985; Eckert and Mitchell, 1989).
For the pharmacological treatment of anorexia nervosa, limited data are available that support the use antidepressants, lithium, and cyproheptadine, an antihistamine. Antipsychotics have also been tried in the treatment of anorexia nervosa but without apparent success.
Neuroleptics
Dally et al (1960) studied the use of large doses of chlorpromazine (CPZ) combined with insulin treatment in 20 patients (all but one were female) with anorexia nervosa (average age 18.8 ± 5.5 years and average weight lost 37.1 ± 15.1 lbs). Patients stayed in bed throughout treatment until they regained near normal weight. CPZ was given starting at 150 mg/day po and increased by 75 mg/day to the limit of the patient's tolerance (range =150-1600 mg/d). Modified insulin therapy was also begun. Starting with five units and progressively increasing the dose was increased (mean = 60 units in the AM) until the patient began to experience sweats and become drowsy. At this point the patient was given a "large meal." The caloric intake was initiated at 1500 calories and titrated upwards to 4000 calories/day. No supportive psychotherapy was given until the weight was nearly normal. The CPZ plus insulin patients (treatment group A) were compared to those obtained in 24 similar cases treated at the same hospital with other methods (treatment group B). Group A gained an average of 4.4 ± 1.3 lbs per week, and their average stay in the hospital was only 35 ± 12 days. This contrasted with group B who gained an average of only 1.3 ± 1.6 lbs (p < 0.01) a week during an average stay of 58 ± 6 days (p < 0.01). Most patients began to gain weight almost at once when treated with the combination of CPZ. No complications were encountered. Some of the initial weight increase was due to water retention. It was regarded as "a favorable sign" since it was usually "accompanied by cheerfulness and more ready acceptance of treatment." The authors concluded that the CPZ/insulin treatment more effective than intubation, psychotherapy, and other methods previously used. Of the two pharmacologic treatments, a follow-up study concluded that chlorpromazine was the more important of the two (Dally and Sargant 1966). On the basis of on the mean percentage pre-illness weight regained it was concluded that treatment with insulin alone (84%) was scarcely more effective than bed rest (82%) in contrast to the drug combination (90%).
In a double-blind 3-week each cross-over study, Vandereycken et al (1982) studied pimozide (PZD), a dopamine blocker, in doses of 4 or 6 mg versus placebo in 18 female patients. Patients had a mean age of 22 years, a duration of illness median of three years, and a mean weight of 38 ± 5 kg. Mean daily weight gain of PZD versus placebo was 135 g versus 80 g after phase one, and 130 g versus 93 g after the crossover in phase two. The cross-over analysis showed that the direct effect of PZD tended to be higher (p= 0.067) than that of placebo. Changes in weight were always positive with PZD irrespective of the treatment sequence. Analysis of behavior showed a superiority of PZD in "attitude towards treatment" (p=0.05) and the behaviors of bingeing and purging (p=0.04). It did not affect the scale items of hyperactivity, preoccupation with eating/weight, bulimic tendencies, and tension at eating. The benefit of PZD to placebo could be important since restoration of normal body weight is one of the primary goal in these patients. However, the authors were disappointed in the results because there was only a minimal effect on the patients eating attitudes and behaviors.
Vandereycken et al (1984) evaluated sulpiride (SPD) in a double-blind placebo-controlled cross-over study in 18 female anorectic patients. After a baseline period of one week, patients started two medication periods of three weeks each, alternating SPD with placebo or the reverse by random allocation. Thirteen patients received a daily dose of 300 mg and five patients received 400 mg. Assessments were carried out on mean daily weight change, the Eating Attitudes Test (EAT), and the Body Attitudes Test (BAT). Unfortunately, the SPD-placebo sequence group 1 (mean age 23 ± 7 years, mean weight 40 ± 5 kg) significantly differed from the placebo-SPD group 2 (mean age 24 ± 10 years, mean weight 39 ± 4 kg) as to the baseline scores on the EAT and BAT. Daily weight change in period 1 was 154 ± 91 g for SPD versus 93 ± 49 g for placebo and in period 2 was 103 ± 48 g for SPD and 98 ±51g for placebo. No difference was found between the 300 mg and 400 mg groups. According to the cross-over analysis, no significant direct effects of SPD on the mean daily weight change or on the behavioral/attitudinal tests were established. Because of limitations with a cross-over design, a carry over effect could not be ruled out. Also, treatment was only approached from the "short-term perspective," as in the previously mentioned study.
Conclusion: Dally and Sargant (1960, 1966) suggested in their uncontrolled study that dopamine blocking agents were useful in the treatment of anorexia nervosa. However, over 20 years later, Vandereycken's group (1982, 1984) in two controlled crossover studies was unable replicate this original finding. Currently available data conclude that neuroleptic agents do not significantly effect weight gain or eating attitudes/behavior in anorectic patients.
Antidepressants
Biederman et al (1985) studied amitriptyline (AMT) in a five week double- blind placebo controlled study in which 11 patients received AMT, 14 received placebo, and 18 received only psychosocial treatment because they refused drug treatment. Dosage was titrated upward at a rate of 50 mg every other day to a maximum dose of 3 mg/kg/d or 175 mg/d which ever was the least or until significant adverse effects developed. Response to treatment was assessed using daily weight gain and the numerous rating scales that included the Eating Attitude Test. The three treatment groups were similar in age, severity of symptoms, weight, percentage below ideal weight, socioeconomic status, and affective illness family history. Improvement was poor overall. There were no significant differences between the treatment groups in the mean weight gain or the EAT scores. Patients did not manifest any tendency toward reduction of depressive symptoms. The majority of AMT treated patients reported adverse effects including drowsiness and dry mouth that were severe for some. It is possible that the negative results were the consequence of inadequate sample size, inadequate dose, or insufficient treatment time, but the consistency of the negative data was discouraging and did not suggest a therapeutic role for amitriptyline in the short-term treatment of anorexia nervosa.
Crisp et al (1987) randomized 16 female anorexic patients to an 8-week trial of clomipramine (CMP) 50 mg/d or placebo in patients matched for mean duration and onset of illness. According to analogue scale scores for eight feelings, patients only reported feeling significantly hungrier while taking the drug. The two groups showed no significant differences in weight gained or in the time required to reach their target weights (76 days for CMP and 72 days for placebo). The analogue appetite scores were elevated in the CMP group for the first 5 weeks of the trial and for hunger for the first 7 weeks. Additionally, on the "free days" for unrestricted calorie intact the CMP ingested more calories at the end of week 1, 2 and 8. Thus the authors suggest that CMP may at least facilitate early in the treatment process the rate at which patients reach their target weight.
Kaye et al (1991) completed an open trial of fluoxetine in 31 patients (all but one were female) with DSM-III-R anorexia nervosa, with a mean age of 20±7 years, and at 89 ± 7% of their average body weight (ABW). Twenty- seven of the anorectics were started on fluoxetine during inpatient treatment while seven began treatment as outpatients. All subjects were started on a dose of 20 mg/day for three weeks, with an increase in dose dependent upon response. All inpatients and some outpatients received psychotherapy. However, others were seen only occasionally for brief supportive therapy. Response was defined as good, partial, or poor with maintaining weight > or = 85% ABW necessary to meet a good or partial response. Ten patients had a good response with a mean maximum dose of 26 ± 9 mg/d of fluoxetine, and their mean weight was 98 ± 7 % ABW. Seventeen patients had a partial response with a mean maximum dose of 41 ± 18 mg/d, and their mean weight was 100 ± 14% ABW. There were 4 poor response and 3 dropouts. The most notable finding was that 29 (94%) of 31 patients maintained a body weight above the anorexic range at the time of follow-up that was after 11 ± 6 months on fluoxetine. However, since there was not a control group this finding requires replication. Only a few patients had a dramatic response characterized by a large reduction in symptoms. For most patients on fluoxetine, symptoms were reduced but continued to be present.
Lithium
Gross et al (1981) evaluated lithium carbonate in a four week random, double-blind, placebo-controlled, parallel group study of 16 anorectic patients. The lithium group had a mean age of 21 ± 2 years versus 19 ± 3 years for placebo, and a mean weight of 36 ± 1 kg versus 33 ± 2 kg for placebo. Patients on lithium gained 6.8 ± 0.4 kg versus 5.2 ± 0.1 kg for placebo with significant differences in weight gain seen at week 3 and week 4. A significant effect on the "denial or minimization of illness" item on the physician-rated Psychiatric Rating Scale suggesting lithium may reduce denial along with increasing weight gain. These results indicate that lithium may augment weight gain in patients with anorexia nervosa also treated with behavior modification.
Antihistamines
Goldberg et al (1979) randomized 81 female anorectic patients to one of four treatment combinations of CPH and placebo with/without behavior therapy. Patients were started on 12 mg/d of CPH liquid concentrate and were increased by 4 mg every fifth day if they did not gain at least 0.5 kg up to a maximum dose of 32 mg/day. The overall weight gain difference between CPH and placebo was small and not significant (CPH mean = 5.1 kg versus placebo mean = 4.3 kg). The largest difference in weight gain favoring CPH was seen in patients who had a history of birth delivery complications (p=0.003), had lost 41-52% of weight from normal (p=0.06), and had a history of prior outpatient treatment failure (p=0.07). This finding is debatable. Only the birth complication finding was at all robust. However, it was entirely a result of three patients with 2 birth complications each having an average weight gain of about 9 kg. Despite their enthusiasm a more unbiased view probably would be skeptical of these findings.
Halmi et al (1982) randomly treated 36 female anorectic patients (mean age 21 ± 6 years) with AMT, CPH, or placebo. Patients were admitted to the hospital for up to three months or until they reached their normal target weight. Patients were weighed every morning and rated on the Hamilton Rating Scale for Depression (HAMD) by two nurses every other week. The CPH dose was titrated from 8 to 32 mg/d over three weeks while the AMT dose was titrated from 40 to 160 mg/d. A significant positive drug effect versus placebo was seen on HAMD scores (mean score of 10 for CPH, 12 for AMT, and 16 for placebo) and on weight gain (mean percentage of normal weight 98% for CPH, 96% for AMT and 88% for placebo). Although the effects on weight seem reasonable, it is doubtful if the change is the HAMD scores are of any clinical significance. It could easily be argued that this HAMD effect originated from the sedative effects of the drugs.
In a randomized double-blind parallel design study, Halmi et al (1986) treated 72 female anorectic patients (mean age of 21 ± 5 years) with either cyproheptadine (CPH), amitriptyline, or placebo. The maximum daily dose for CPH was 32 mg and AMT 160 mg. Patients taking AMT attained their target weight an average of 13 days earlier than patients taking placebo while patients receiving CPH attained their weight 11 days earlier (p = 0.05). CPH specifically was useful for increasing the rate of weight gain and reducing depressed mood in non bulimic anorectics. There were nine treatment failures on placebo, six on AMT and only 4 on CPH. Thus the patients could potentially experience shorter hospital stays if AMT or CPH is utilized as part of their therapy.
|
Table 1. Pharmacotherapy Trials for Anorexia Nervosa |
||||
|
Author |
|
|
Weight Gain Results |
Comment |
|
Neuroleptics |
||||
|
Dally |
vs. nonpharmacologic tx, unblinded (n=24) |
|
vs. 1.3 lb/wk |
CPZ/insulin more effective than intubation, psychotherapy, & other methods previously used |
|
Vandereycken 1982 |
PLB crossover (n=18) |
|
|
weight increase change = NS |
|
Vandereycken 1984 |
PLB crossover (n=13) |
|
|
weight increase change = NS |
|
Antidepressants |
||||
|
Biederman 1985 |
|
< or =175 mg/d x 5 wk |
No change in mean wt increase between 3 txs |
weight increase change = NS |
|
Crisp |
parallel |
|
No change in mean wt increase or time to reach target wt |
Subjectively hunger increased on drug; caloric intake on free days increased on CMP for 3 of 8 wks |
|
Kaye |
open outpatient follow-up study |
|
29/31 maintained wt above the anorexic range @ 11 mo follow-up |
good response =10 |
|
Lithium |
||||
|
Gross |
PLB (n=8) + behavior modification parallel |
|
6.8 kg vs. 5.2 kg total wt increase |
significant effect on the "denial or minimization of illness" |
|
Cyproheptadine |
||||
|
Goldberg 1979 |
± behavioral tx parallel |
|
5.1 kg vs 4.3 kg |
No change in mean wt increase |
|
Halmi |
parallel |
40-160 mg/d x 3 mo or until target wt reached |
98% vs |
HAMD scores decreased (p=0.05) |
|
Halmi |
parallel |
< or =32 mg/d < or =160 mg/d x 3 mo or until target wt reached |
Target wt reached 11 d & 13 d earlier with drugs |
CPH most useful nonbulimic anorexics with dysphoria |
BULIMIA NERVOSA
A variety of antidepressants appear to be effective for bulimia nervosa. Table 2 summarizes those studies that have compared placebo to active medication treatment. The presence of depression is not a prerequisite for medication improvement of bulimic behaviors. Two studies specifically excluded patients with depression and still showed a positive treatment response (Hughes et al, 1986; Horne et al, 1988).
|
Table 2. Placebo Controlled Medication Trials for Bulimia Nervosa |
||||
|
Author |
|
|
|
Comment |
|
|
||||
|
Pope |
parallel, outpatient |
x 6 wk |
|
A 1- to 8-month follow-up of 20 AD tx'd subjects, change 90% decrease in bingeing |
|
Mitchell 1984 |
parallel, outpatient |
x 8 wk |
decrease in bingeing did not reach significance; there was no change in wt |
Dysphoric severity correlated negatively with the change (decrease) in bingeing |
|
Hughes 1986 |
parallel, outpatient |
x 6 wk |
DIP = (91% decreased) PLB = (19% increased) |
dysphoria improved; |
|
Agras |
parallel, outpatient |
x 16 wk |
PLB = (43% decrease) Purging @ 16 wk IMP = (72% decrease) PLB = (35% decrease) |
Dysphoria decreased at wk 6 but not at wk 16 |
|
Barlow 1988 |
|
x 6 wk, 3 wk washout |
DIP = (@ 50% decrease) PLB = (@ 0% decrease) Purges/wk DIP = (58% decrease) PLB = (13% decrease) |
No effect on dysphoria |
|
|
||||
|
Walsh |
PLB (n=27) parallel, outpatient |
x 8 wk |
DIP = (= 55% decrease) PLB = (= 9% decrease) PNZ > PLB FOR EAT |
9/31 (29%) on PNZ dc'd b/c of side effects orthostais (3) and sedation (2) |
|
|
||||
|
Blouin 1988 |
|
60 mg x 6 wk each,3 wk washout |
DIP = (decrease 2.0 --> 1.0) FFA = (decrease 2.0 --> 1.0) |
Both drugs decreased the urge to binge & depressive symptoms |
|
Pope |
parallel, outpatient |
|
TZD = (31% decrease) PLB = (21% increase) Purges (d/wk) TZD = (34% decrease) PLB = (5% increase) |
Results could be biased against TZD because 3 of the 20 TZD completers did not have detectable blood levels. |
|
Fluoxetine Group 1992 |
PLB (n=129) parallel, outpatient |
|
60 mg = (67% decrease) 20 mg = (45% decrease) PLB = (33% decrease) Purges/wk 60 mg = (56% decrease) 20 mg = (29% decrease) PLB = (5% decrease) |
20 mg not effective for decreased bingeing and dysphoria |
|
|
||||
|
Horne |
parallel, outpatient |
|
BPR = (67% decrease) PLB = (2% decrease) Purges (d/wk) BPR = (54% decrease) PLB = (3% decrease) |
no change in dysphoria 5.8% rate (4/69) of grand mal seizures (study terminated after seizure #4) |
As with the treatment of anorexia nervosa, pharmacological treatments for bulimia nervosa are part of an individualized treatment program that may include specific forms of psychotherapy and educational interventions. A review of controlled studies using specific psychotherapies (most often cognitive-behavioral) indicate they are often effective (Mitchell et al, 1989b).
Tricyclic Antidepressants
The efficacy of imipramine in the treatment of bulimia was first studied extensively in 1983 by Pope et al (1983). Twenty-two outpatients meeting DSM-III criteria for bulimia were randomly assigned to either 50 mg of imipramine (n=11) or placebo (n=11). They were then instructed to increase the dose by one capsule every other day until they had reached four capsules or they experience intolerable side effects. Patients were evaluated at the end of 2, 4, and 6 weeks. Three subjects withdrew from the study due to photosensitivity reactions. Of the 9 subjects taking imipramine, four reported marked (more than 75%) and 4 reported moderate (more than 50%) decreases in binge eating, while 1 subject remained unchanged. Among the 10 subjects receiving placebo, one reported a moderate decrease in bingeing, eight were unchanged, and one subject reported getting worse. On a 1- to 8-month follow-up of these treated subjects, 20 of the 22 responded to imipramine or another antidepressant. These data suggested that there may be a link between bulimia and affective disorders, thus uncovering new treatment pathways for the treatment of bulimia nervosa.
In a placebo-controlled, double-blind trial, Mitchell and Groat (1984) examined the activity of amitriptyline in the treatment of 32 female outpatients bulimics (DSM-III). Patients were randomly assigned to amitriptyline (dose: 50 mg at bedtime increased by 50 mg increments every third day to a dosage of 150 mg/day) or placebo for eight week treatment trial. Six subjects dropped out due to complaints that the medication was not working. Only one patient complained of sedation. These six were replaced. Subjects kept weekly records of their eating behavior. The subjects were seen weekly by psychiatrists where they received a minimal behavioral treatment protocol in addition to the drug therapy. The drug was well tolerated and was not associated with weight gain and no subjects required modification of the dosage. The results showed that 150 mg of amitriptyline had significant antidepressant activity in these patients that became apparent after only 2 weeks of drug therapy. Both the active drug and the placebo groups showed considerable improvement in eating behavior. Table 3 below compares the effects on eating behavior between the two treatments.
|
Table 3. Change in eating behavior |
|||
|
Item |
|
|
|
|
bingeing (days) |
|
|
|
|
bingeing (episodes) |
|
|
|
|
bingeing (h/wk) |
|
|
|
This study further emphasizes the need for behavior modification in addition to drug therapy.
To determine if other tricyclic antidepressants were as equally effective as imipramine in the treatment of bulimia, Hughes et al (1986), performed a double-blind controlled study comparing the effects of desipramine versus placebo. Each patient was randomly assigned to receive either desipramine (n=10) or placebo (n=12). The dose for the desipramine was 50 mg increased by 50 mg every third day up to a maximum dose of 200 mg nightly. These patients were then crossed over to the other treatment group. No systematic supportive counseling was provided other than weekly 15 minute interviews at which ratings on four variables were obtained. The study lasted six weeks. Three patients taking desipramine withdrew after 3 weeks because of side effects. Maculopapular rashes developed in two patients and one patient complained of oversedation. None of these patients had major depression. The results demonstrated significant benefit in comparison with placebo and with baseline status. Nineteen of 22 patients attained 80% or more decrease in their binge eating, with 15 (68%) attaining complete abstinence from binges at 10 weeks. This study demonstrates a dramatic initial benefit from antidepressant medication in bulimia. Patients receiving active medication could clearly identify a difference in their preoccupation with food and the urge to binge.
A 16 week controlled trial by Agras et al (1987), compared imipramine to placebo for the treatment of bulimia for. Twenty-two bulimic women participated in the double-blind study. Of the twenty completers, ten had been given 50 mg of imipramine with the instructions to continue to increase the dose up to a maximum of 300 mg (mean = 167 mg/d @ week 16) for the remainder of the study. Patients were assessed at the end of 6 and 16 weeks. The imipramine group showed a significantly greater reduction in purging at each assessment. Depression was also measured and was found to be reduced to a significantly greater extent at the end of six weeks but not at the end of 16 weeks. These findings suggest that imipramine is an effective treatment for bulimia. It was found that only one-third of these patients had stopped purging by the end of the study. Thus, it seems as if adjunct treatments, such as psychotherapy, may need to be added to the management of bulimic patients.
In a double-blind crossover study, Barlow et al (1988) compared the effects of desipramine and placebo in a sample of bulimics containing both depressed and non-depressed subjects. Forty-seven patients were randomly assigned to receive either desipramine (dose: 50 mg/day increased by 50 mg increments every third day to a dosage of 150 mg/day) for six weeks, followed by a three week washout period, and then six weeks of identical placebo, or the reverse. Twenty-four subjects completed the entire fifteen week protocol. The remaining twenty-three dropouts were due to adverse effects e.g., dry mouth, dizziness, etc. In comparison to placebo, desipramine reduced weekly bingeing and weekly vomiting (p < 0.05). The pattern of change indicated that the antibulimic benefits appear as early as week one. Eleven subjects showed a greater that 50% reduction in bingeing, while three of these improved more than 80%. Only one subject stopped bingeing completely. No significant changes were detected on the depression scales of the two rating instruments that were utilized. Blood levels were obtained for 12 out of 24 subjects. Only three had blood levels greater than 125 ng/ml.
Monamine Oxidase Inhibitors
In a trial conducted by Walsh et al (1984) the effects of phenelzine compared to placebo was studied in 50 patients with bulimia. Eighty-two patients were admitted to a 10 week single-blind placebo washout phase that lasted two weeks. The remaining 50 subjects were then randomly assigned phenelzine or placebo in a double-blind trial. The dose of phenelzine was increased to 60 mg by week 2. Patients who had not responded by week 4 were increased to 90 mg by week 6, and maintained there until the study ended at week 8. For phenelzine the frequency of bingeing decreased by 55% which was significantly than the 9% decrease for placebo. Also, the phenelzine patients showed greater decreased on the EAT scale. Both depressed and non depressed patients experienced significant improvement. There were also significant changes in measures of psychological state noted in the phenelzine group. However, side effects posed a major problem for the use of phenelzine. Twenty-nine percent (9/31) of the patients randomized to phenelzine discontinued the drug because of side effects, usually orthostatic hypotension (3) and sedation (2). On follow-up, 24 of 27 patients who had responded to phenelzine eventually discontinued the drug after a median course of 16 weeks. Thirteen discontinued the drug because of side effects. who responded to phenelzine were unable to remain on the drug regimen because of side effects. The most frequently reported of these were orthostatic hypotension, weight gain, and sedation. No patients experienced a hypertensive crisis during the study. However, because of the strict dietary regulations to avoid a hypertensive crisis and the high incidence of side effects, it is very likely that compliance will be a major deterrent in the future use of these agents and the authors suggest that the role of phenelzine in the bulimic population will be limited.
Selective Serotonin Reuptake Inhibitors
In a double-blind, placebo-controlled, crossover study comparing desipramine, 150 mg/d, and fenfluramine, 60 mg/d, twelve subjects were randomly allocated to fenfluramine and 10 subjects received desipramine (Blouin et al 1988). Half the subjects in each group received the active drug during the first six weeks and the other half received placebo. A 3 week washout period followed and then the patients were crossed over for the remaining 6 weeks. The EDI, PMS, and HSCL were administered at weeks 0, 2, 4, 6, 9, 11, 13, and 15. Subjects maintained a daily record of bingeing and purging activities. Results showed that desipramine and fenfluramine had significantly greater effects on bingeing and purging frequency than placebo, although a greater percentage of patients responded to fenfluramine than desipramine. Both drugs were also effective in reducing the psychological symptoms of bulimia, e.g., urge to binge, feelings of depression, etc. This study shows that these drugs may act on different physiological mechanisms involved in feeding. Future work will be directed toward defining bulimic subtypes who may respond best to particular psychopharmacological agents.
In a placebo-controlled double-blind study of trazodone by Pope et al (1989), 42 women who met the new DSM-III-R criteria for bulimia were given either 50 mg of trazodone or placebo. They were instructed to raise the dose by one tablet every second day to a maximum of eight tablets (trazodone 400 mg). During the active phase, subjects were evaluated at weeks 0, 2, 4, and 6. Trazodone was found to be significantly superior to placebo in doses up to 400 mg/day, in both reducing the frequency of binge eating and vomiting and in several of the subjects' subjective ratings of improvement that included fear of eating and self-control. Side effects reported were minimal with only one subject dropping out due to headaches. This study was the first study to use subjects meeting the newer narrowed criteria of DSM-III-R for bulimia nervosa. Therefore, the results of this study are able to prove that antidepressants are effective in true cases of bulimia nervosa since cases of hyperphagic depressions were excluded.
In a multicenter, placebo-controlled, double-blind trial, the efficacy of fluoxetine was studied (Fluoxetine Bulimia Nervosa Collaborative Group 1992). Women meeting the DSM-III-R criteria for bulimia nervosa were enrolled at 13 centers throughout the US and Canada. Eligible patients were randomized to one of three groups: placebo, 20 mg/d of fluoxetine once daily, or 60 mg/d of fluoxetine once daily. Patients were seen weekly. Weight and vital signs were recorded. Patients were required to keep a daily diary. Patients were also assessed by the HDRS, EDI, EAT, and scales to measure bulimic intensity. Blood samples were obtained at weeks 4 and 8 to check serum concentrations of fluoxetine and norfluoxetine. The total number of completers was 387. The results showed that the groups receiving 60 mg/d of fluoxetine, 20 mg/d of fluoxetine, and placebo had median reductions of 56%, 29%, and 5%, respectively, in vomiting episodes per week, and median reductions of 67%, 45%, and 33%, respectively, in binge-eating episodes per week. The large dose was more effective than the lower dose which in turn was more effective than the placebo. The modest incidence of adverse effects observed with fluoxetine in this study and the low discontinuation rate as a result suggests that fluoxetine may offer some advantage over earlier antidepressants in the treatment of bulimia nervosa.
Bupropion
In a multicenter controlled trial (Horne et al 1988), the treatment of bulimia with bupropion compared to placebo showed favorable results. Of the thirty-seven bupropion completers of the 8 week active treatment phase, thirty-four had taken a maximum dosage of 450 mg; two had taken 300 mg and one had taken 225 mg. Bupropion was found to be significantly superior to placebo in reducing the frequency of eating binges and purging behavior. No change occurred on the HAMD. However, one of the exclusion criteria for the study was major depression. However, four cases of grand-mal seizures occurred leading to the termination of the study. The etiology of these bupropion associated seizures is unknown. As a result, the presence of bulimia nervosa is considered a contraindication to the use of this drug.
Miscellaneous
Medications other than traditional antidepressants may also be effective for bulimia nervosa. Lithium reduced bulimia nervosa symptoms in the majority of patients in a non-blind study (Hsu, 1984). Phenytoin produced improvement in a placebo crossover study. However, the results appeared to be confounded by extension of active medication effects into the placebo phase of the trial (Wermuth et al, 1977). An open label study of naltrexone suggested a positive therapeutic response (Jonas and Gold 1987), but this was not confirmed in a small (N=14) placebo controlled trial that showed only a trend favoring naltrexone (Mitchell et al 1989a).
It has been shown through the results of many controlled trials that antidepressant drugs clearly are the best choice if pharmacotherapy is to be instituted in bulimic patients. However, it is important to note that many of these studies noted greater improvement in patients on a behavioral therapy protocol and pharmacotherapy, rather than either treatment used alone. Some studies are underway to determine if any antidepressant possesses an advantage over the others in regards to efficacy. No solid data has yet been revealed.
Most of the outcome research for treatments of anorexia nervosa and bulimia nervosa has been short term (i.e., weeks to months). Much less is known about the long-term efficacy of various treatments. In view of the chronicity of eating disorders, medications will probably need to be administered for long periods. Optimal dosages or treatment times for longitudinal therapy have not been established.
REFERENCES
Agras WS, Dorian B, Kirkley BG, et al (1987). Imipramine in the treatment of bulimia: double-blind controlled study. Int J Eating Dis. 6:29-38.
Barlow J, Blouin J, Blouin A, et al (1988). Treatment of bulimia with desipramine: a double blind crossover study. Can J Psychiatry 33:129-32.
Biederman J, Herzog DB, Rivinus TM (1985). Amitriptyline in the treatment of anorexia nervosa: A double-blind placebo-controlled study. J Clin Psychopharmacol 5: 10-16.
Blouin AG, Blouin JH, Perez EL, et al (1988). Treatment of bulimia with fenfluramine and desipramine. J Clin Psychopharmacol 8:261-69.
Blundell JE (1984a). Serotonin and appetite. Neuropharmacol 23:1537-1551.
Blundell JE (1984b). Systems and interactions: An approach to the pharmacology of eating and hunger. In Stunkard AJ, Stellar E:(eds.) Eating and Its Disorders. Raven Press, New York, 39-65.
Breisch ST, Zemlan FP, Hoebel BG (1976). Hyperphagia and obesity following serotonin depletion by intraventricular parachlorophenylalanine Science 192:382-3845.
Crisp AH, Palmer RL, Kalucy RS, et al (1976). How common is anorexia nervosa? A prevalence study. Br J Psychiatry 128:549-54.
Crisp AH, Lacey JH, Crutchfield M (1987). Clomipramine and 'drive' in people with anorexia nervosa: An inpatient study. Br J Psychiatry 1987; 150:355-358.
Dally PJ, Sargent W (1960). A new treatment of anorexia nervosa. Br Med J 1: 1770-3.
Dally PJ, Sargent W (1966). Treatment and outcome of anorexia nervosa. Br Med J 2: 793-5 .
Eckert ED, Mitchell JE : An overview of the treatment of anorexia nervosa. Psychiatric Med 7: 293-315, 1989.
Fluoxetine Bulimia Nervosa Collaborative Group (1992). Fluoxetine in the treatment of bulimia nervosa. Arch Gen Psychiatry 49:139-47.
Garner DM, Garfinkel PE (eds) (1985). Handbook of Psychotherapy for Anorexia and Bulimia. New York: Guilford Press, 107-47.
Goldberg SC, Halmi KA, Eckert ED, et al (1979). Cyproheptadine in anorexia nervosa. Br J Psychiatry 134: 67-70.
Gross HA, Ebert MH, Faden VB, et al (1981). A double-blind controlled trial of lithium carbonate in primary anorexia nervosa. J Clin Psychopharmacol 1: 376-381.
Halmi KA, Eckert ED, Falk JR (1982). Cyproheptadine for anorexia nervosa. Lancet 1: 1357-1358.
Halmi KA, Eckert E, Ladu T, et al (1986). Anorexia nervosa: Treatment efficacy of cyproheptadine and amitriptyline. Arch Gen Psychiatry 43:177-81.
Horne RL, Ferguson JM, Pope HG, et al (1988). Treatment of bulimia with bupropion: multicenter controlled trial. J Clin Psychiatry 49:262-6.
Hsu LKG (1984). Treatment of bulimia with lithium. Am J Psychiatry 141:1260-2.
Hudson JI, Pope HG, Jonas JM, et. al.(1985). Treatment of anorexia nervosa with antidepressants. J Clin Psychopharmacol 5:10-6.
Hughes PL, Wells LA, Cunningham CJ, et al (1986). Treating bulimia with desipramine. Arch Gen Psychiatry. 42:182-6.
Jonas JM, Gold MS (1987). Treatment of bulimia with the opiate antagonist naltrexone: Preliminary data and theoretical implications. In: Hudson Jl, Pope HG, eds., Psychobiology of Bulimia, APA Press.
Kaye WH, Weltzin TE, George Hsu LK, et al (1991). An open trial of fluoxetine in patients with anorexia nervosa. J Clin Psychiatry 1991; 52(11):464-71.
Leibowitz SF, Papadakos PJ (1978). Serotonin-norepinephrine interactions in the paraventricular nucleus: antagonistic effects on feeding behavior in the rat. Neurosci Abstr 4:452.
Mitchell JE, Groat R (1984). A placebo-controlled, double-blind trial of amitriptyline in bulimia. J Clin Psychopharmacol 4:186-93.
Mitchell JE, Christenson G, Jennings J, et al (1989a). A placebo-controlled double blind crossover study of naltrexone hydrochloride in outpatients with normal weight bulimia. J Clin Psychopharmacol 9:94-7.
Mitchell JE, Hoberman H, Pyle RL (1989b). An overview of the treatment of bulimia nervosa. Psychiatric Med 7: 317-332.
Pope HG, Hudson JI, Jonas JM, et al (1983). Bulimia treated with imipramine: A placebo-controlled, double-blind study. Am J Psychiatry. 140:554-558.
Pope HG, Keck PE, McElroy SL, et al (1989). A placebo-controlled study of trazodone in bulimia nervosa. J Clin Psychopharmacol 19: 254-59.
Sabine EJ, Yonace A, Farrington AJ, et al (1983). Bulimia nervosa: A placebo-controlled double-blind therapeutic trial of mianserin. Br J Clin Pharmacol 15:195S-202S.
Vandereycken W, Pierloot R (1982). Pimozide combined with behavioral therapy in the short-term treatment of anorexia nervosa. Acta Psychiatr Scand 66: 445-450.
Vandereycken W (1984). Neuroleptics in the short-term treatment of anorexia nervosa: A double-blind placebo-controlled study with sulpiride. Br J Psychiatry 144: 288-92.
Vigersky RA, Loriaux DL (1977). The effect of cyproheptadine in anorexia nervosa: a double blind trial. In Vigersky RA (Ed.) Anorexia Nervosa. Raven Press, New York, 349-56.
Walsh BT, Gladis M, Roose SP, et al (1988). Phenelzine vs. placebo in 50 patients with bulimia. Arch Gen Psychiatry 45:471-5.
Wermuth BM, David KL, Hollister LE, et al (1977). Phenytoin treatment of the binge-eating syndrome. Am J Psychiatry 134:1249-54.