Lifestyle Modification in the Pharmacologic Treatment of Obesity: A Pilot Investigation of a Potential Primary Care Approach

Transcription

1 Lifestyle Modification in the Pharmacologic Treatment of Obesity: A Pilot Investigation of a Potential Primary Care Approach Thomas A. Wadden, Robert I. Berkowitz, Renee A. Vogt, Suzanne N. Steen, Albert J. Stunkard, Gary D. Foster Abstract WADDEN, THOMAS A, ROBERT I BERKOWITZ, RENEE A VOGT, SUZANNE N STEEN, ALBERT J STUNKARD, GARY D FOSTER. Lifestyle modification in the pharmacologic treatment of obesity: A pilot investigation of a potential primary care approach. Obes Res. 1997;5:21? This study examined a new method of providing brief, individual lifestyle modification to obese individuals treated by pharmacotherapy. Twenty-six women with a mean (+SD) age of 47.0 * 7.2 years, weight of 97.6 f 13.0 kg, and body mass index of 36.5 f 5.0 kg/m2 were prescribed 60 mg/d of fenfluramine and 15 mg/d of phentermine for one year. In addition, half of the women were randomly assigned to traditional group behavior modification, conducted by a nutritionist, which included minute sessions during the year. The other half were provided lifestyle modification by a physician during minute structured visits. All participants received identical treatment manuals and comparable assignments for behavior change. At the end of one year, patients in the physician group achieved the same highly successful weight losses as those treated by group behavior modification ( kg vs f 7.9 kg, respectively). Treatment was associated with highly significant improvements in lipids and lipoproteins, as well as in mood and several measures of appetite. Weight loss the first four weeks, as well as patient completion of daily food records during the first 18 weeks, correlated positively with weight loss at weeks 18, 26, and 52. Results of this study await replication using larger samples Submitted for publication September 17, Accepted for publication October 3 I, From the University of Pennsylvania School of Medicine, Department of Psychiatry, Philadelphia. PA. Reprint requests to Thomas A. Wadden, University of Pennsylvania, Department of Psychiatry, 3600 Market Street, Suite 738, Philadelphia, PA Copyright NAASO. but strongly suggest that effective lifestyle modification can be provided during brief, structured physician visits. The findings are discussed in terms of their implications for the treatment of obesity in primary care practice. Key words: obesity, behavior modification, pharmacotherapy, anorectic medications, fenfluramine, phentermine Introduction The past five years have witnessed a revolution in the pharmacologic treatment of obesity. Several factors have sparked it, the first of which was Weintraub s report that patients treated by dl-fenfluramine (60 mg/day) and phentermine (15 mg/day) lost 14 kg in 34 weeks and maintained the great majority of this loss for up to three years while on medication (37,39). A second factor was the Food and Drug Administration s (FDA) recent approval of dexfenfluramine for weight loss and the maintenance of weight loss (30). The FDA s action recognized that obesity is a serious, chronic condition that requires long-term management, similar to hypertension, diabetes, and related disorders (8). Approval of this serotonergic agent was based on its generally favorable side-effect profile, which stands in sharp contrast to that of the amphetamines of a generation ago (27), and on findings that it was associated with sustained weight reduction (20). The new generation of obesity medications potentially includes agents that not only reduce food intake (9,12,20) but also modulate nutrient digestion (15) and increase energy expenditure (3,41). The need for such agents could hardly be greater, given recent findings that one of every three Americans is now obese (22). A pressing question is whether the very favorable results obtained in clinical research trials, which provided intensive multidisciplinary treatment, can now be achieved in traditional office prac- 218 OBESITY RESEARCH Vol. 5 No. 3 May 1997

2 tice. Patients treated by Weintraub, for example, participated during the first 34 weeks in 14 2-hour group behavior modification classes devoted to diet and exercise modification (37). Although patients treated by behavior therapy plus medication lost substantially more weight than those treated by the same counseling plus placebo, an important question left unanswered by Weintraub s study is whether group behavior modification is indeed necessary with pharmacotherapy for obesity. Few primary care practitioners can provide such group care, and when encouraged to provide lifestyle counseling during brief office visits, practitioners may be uncertain concerning what to provide and how best to provide it. This pilot study examined the effectiveness of fenfluramine plus phentermine combined with either: 1) frequent group behavior modification; or 2) infrequent but structured physician visits. The structured physician visits were designed to increase the possibility of their implementation in traditional office practice. Methods and Procedures Patients The study included 26 women with a mean (f SD) age of 47.0 f 7.2 years, weight of 97.6 f 13.0 kg, height of f 6.9 cm, and body mass index (BMI) of 36.5 f 5.0 kg/m2. Patients responded to notices in local newspapers describing long-term treatment by medication and behavior modification. Interested persons completed a 1-hour clinic assessment to determine their weight and dieting histories, eating and activity habits, and psychosocial status (32). Thereafter, acceptable candidates were referred to their family physicians, who completed a history and physical examination and determined if there were any contraindications to treatment. These included, in addition to major illnesses (i.e., hepatic, renal, cardiovascular, etc): essential hypertension (> 140/90 mm Hg); primary pulmonary hypertension; glaucoma; Types I or I1 diabetes; pregnancy or lactation; use of antidepressant medications; and the chronic use of nasal decongestants or any medications known to affect weight (37). Family physicians provided written approval for their patients to participate. Results of the history and physical examinations were forwarded to the project physician (RIB), who contacted the family physicians if further information was required, and then gave final approval for the patient to participate. All patients gave their written informed consent and paid $800 for one year of treatment, which covered all costs of medications, blood tests, physician visits, and lifestyle modification (whether provided individually or in group). (This fee represents a substantial reduction from the actual costs of treatment.) Treatment Patients were randomly assigned to treatment conditions, as described below. All completed a 1-week run-in period (i.e., week 0) during which they maintained their usual eating and activity habits and completed assessments of appetite, mood, and other measures. Thereafter, all patients were prescribed 20 mg/day of dl-fenfluramine the first week, 40 mg/day the second week, and 60 mg/day the third week. Fifteen mg/day of phentermine were added the thud week. All patients were instructed to consume a selfselected diet of approximately 1200 kcal/day (with no more than 30% of calories from fat) and to increase their energy expenditure by walking or other aerobic activity. Baseline characteristics of patients in early treatment condition are shown in Table 1. Group Behavior Modification (n=13). Patients in this condition attended weekly group behavior modification classes (34) for the first 18 weeks which were led by either a masters- or doctoral-level nutritionist, each of whom had extensive experience with weight control. Sessions included six to seven participants, lasted 75 minutes, and were conducted using the LEARN Manual for Weight Control (lo), a copy of which was provided to participants. Patients completed weekly homework assignments that recommended: 1) recording food intake; 2) limiting eating to set times and places; 3) slowing the rate of eating; 4) serving and storing foods properly; 5) walking (eventually) four times weekly for 30 or more minutes per bout; 6) increasing lifestyle activity by using stairs rather than escalators, etc.; and 7) modifying self-defeating thoughts related to weight and eating. From weeks 19 to 40, patients attended every-otherweek weight maintenance classes conducted by the same staff member (with the same participants). Sessions were conducted using the Weight Loss Maintenance Survival Guide (1 1) and focused on issues that included: 1) responding to dietary and exercise lapses; 2) identifying high-risk situations for relapse; 3) eliciting social support to maintain healthy eating and exercise habits; and 4) improving body image and self-esteem. From weeks 41 to 52, participants attended three monthly group visits, during which their progress was reviewed. In addition to this total of 32 group meetings, patients in this condition had individual 5-10 minute physician visits at weeks 1, 2, 4, 6, 10, 18, 26, 34, Table 1. Pretreatment characteristics of patients in the two conditions Group behavior Physician visits Variable therapy (n=13) (~13) Age (yr) 47.6 f f 6.1 Weight (kg) 96.7 f k 15.6 Height (cm) f f 7.4 BMI (kg/m2) 36.7 f f 6.2 Values are mean k SD. OBESITY RESEARCH Vol. 5 No. 3 May

3 42, and 52 at which their medication status and health were assessed. Structured Individual Physician Visits (n=13). These patients met with a physician (a psychiatrist with extensive experience with weight control; RIB) for minutes at weeks 1,2,4,6, 10, 18,26,34,42, and 52. At each visit, the physician first assessed the physical health of the patient (as in the group treatment condition) and then spent approximately 10 minutes reviewing her progress since the last visit in changing eating and exercise habits. These patients were provided the same treatment manuals as those in the group behavior modification condition and were expected to complete the same homework assignments. They received no further counseling of any kind during the year. Treatment Standardization. The nutritionists and physician conducted treatment sessions using detailed protocols that indicated the topics to be covered from the manuals at each visit and the assignments to be completed before the next meeting. Separate protocols were prepared for each treatment condition. Dependent Measures Weight. Weight was measured at each visit using a balance-beam scale. Patients were measured in light clothing without shoes. Lipids and Lipoproteins. Blood samples were drawn at baseline and weeks 6, 18, 26, and 52 following a 12-hour overnight fast. Triglycerides, total cholesterol, high-density lipoprotein (HDL) cholesterol, and low-density lipoprotein (LDL) cholesterol were determined using methods previously described (2). Symptoms, Mood, and Attention. All participants were asked, at each of their 10 physician visits, to report any adverse side effects, which were noted in their charts. In addition, patients completed a symptom questionnaire at baseline and weeks 4, 10, 18,26, and 52. They indicated on a scale ranging from 1 ( not at all ) to 4 ( great amount ) the extent to which they had been troubled during the previous week by each of 12 symptoms: constipation, diarrhea, dizziness, drowsiness, dry mouth, headache, nausea, nervousness, shortness of breath, sleep disturbance, stomach pain, and tiredness (17,31). Mood was measured by the Beck Depression Inventory (5) at the same intervals as physical symptoms. Scores on this inventory range from 0 to 63, with higher scores indicative of depression. Attention span (i.e., concentration) was assessed at baseline and week 52 using the digit span test of the Wechsler Memory Scale (35). Appetite and Eating Behavior. Food cravings, hunger, preoccupation with eating, and satiety were measured at baseline and weeks 4, 10, 18, 26, and 52 using 100 mm visual analog scales (17,31). Higher scores indicate greater hunger, satiety, etc. Eating behavior was assessed at baseline and week 52 by the Eating Inventory (28) which in- cludes scales measuring cognitive restraint (of eating), dietary disinhibition, and hunger. Behavioral Adherence. Patient completion of daily food records during the first 18 weeks was used as a measure of adherence to homework assignments. Participants in the group condition turned in their records at each weekly meeting, whereas those in the physician condition submitted all records completed since the last office visit. Each set of weekly diaries was scored from zero to seven, reflecting the number of days that food intake was recorded. If the patient did not submit any records for a given week (and did not mail them in or bring them to the next appointment), she received a zero for the week. Patients were encouraged, but not required, to keep records after week 18. Statistical Analyses Changes in weight, lipids, appetite, and the other measures were examined using analysis of variance (ANOVA) with repeated measures which reveals changes over time, as well as differences between conditions. The relationship of weight loss to other variables, including behavioral adherence, was examined using correlation and regression analyses. Attrition. One patient (in the physician condition) discontinued treatment the first week after developing a skin rash and elected not to restart medication. Two additional patients (one group condition, one physician condition) discontinued treatment after week 26 because of scheduling conflicts. A fourth participant (group condition) discontinued the program at week 30 when she underwent a hysterectomy. Weight losses were analyzed in two ways. The first used the last observation carried forward (LOCF) method, in which the last recorded weight for each of the 26 patients (including the one who dropped out the first week) was included in the week 52 analysis. The second method, an end-point analysis, included data from patients only at the times when they remained in treatment. The latter method was used in analyzing findings for the other principal measures. Results Weight Patients in both treatment conditions lost large amounts of weight, as shown in Table 2. At the end of one year, those who received group behavior therapy lost a mean of 15.4 k 7.9 kg whereas those treated by brief physician visits lost 13.9 k 9.6 kg. A repeated measures ANOVA revealed that there were no significant differences between conditions in weight losses at any time. The mean percentage reduction in initial weight (collapsing across the two conditions) was 15.2 * 8.9% (see Figure 1). Of the total sample (n=26), 96% lost 5% or more of initial weight, 69% lost?lo%, 38% reduced 215%, 31% lost 220%, and 19% achieved reductions of 25% or more of initial weight. (Percentages 220 OBESITY RESEARCH Vol. 5 No. 3 May 1997

4 Table 2. Weight losses (in kg) of patients in two conditions Group behavior therapy (n=13) Physician visits (n=13) Time LOCF EPA LOCF EPA Week f f f f 1.8 Week f f f f2.7 Week f f f f 4.7 Week f f f f 6.4 Week f f f f 7.0 Week f f f f 8.0 Week f f f f 9.1 LOCF = Last observation carried forward; EPA = End-point analysis. Values are means f SD. reported are cumulative; thus, for example, the 31% who lost 220% includes the 19% who lost 25% or more.) Weight loss the first month correlated strongly with subsequent weight loss at week 18 (r=.68, p<o.ool), week 26 (r=0.62, p<o.ool), and week 52 (r=0.52, p<o.ol). Thus, patients who lost the most weight the first month tended to lose the most later in treatment. No relationship was observed between initial weight and weight loss at subsequent periods. Weight Loss and Medication Dosage. Medication doses were reduced in 9 of 25 patients in response to their reports - $20 I I I I - Brief Physician Visits i, Group Behavior Modification a WEEKS Figure 1: Mean percentage reduction in initial weight for patients in the group-behavior-therapy and physician-visit conditions. Data shown are based upon a last-observationcarried-forward analysis. of negative side effects (see Table 3). No relationship was observed (using multiple regression) between medication dosage and weight loss at any time. Thus, patients on reduced doses of dl-fenfluramine and/or phentermine lost as much weight as those on full doses. Mean weight losses, for example, of these two groups at week 52 were 16.7 f 9.1 and 15.8 f 8.1 kg, respectively. Lipids and Lipoproteins Weight loss was associated with large reductions in triglycerides, total cholesterol, and LDL cholesterol, as shown in Table 4. At the end of 1 year, these values decreased by 22.8 _+ 29.0%, 9.7 f 10.0%, and 12.7 _+ 13.3%, respectively. HDL-cholesterol, by contrast, increased by 10.6 f 13.9% from baseline to one year, which resulted in a 17.1 f 13.6% decline in the total cholesterol/hdlcholesterol ratio. These changes suggest that patients significantly reduced their risk of coronary heart disease. There were no significant differences between conditions on any of these variables at any time. Changes in total and HDL cholesterol did not fully parallel changes in weight. Total cholesterol reached its maximum decline of 15.9 f 1 1.2% after only six weeks of treatment, when weight had declined by only 5.4 f 2.0%. Cholesterol levels then increased significantly (p<o.ool) from week 6 to week 52, despite continued and substantial weight loss during this time. Part of this increase was attributable to an increase in HDL cholesterol, which fell the first 6 weeks by 25.2 f 10.0%, but rose by the end of treatment a significant (p<0.004) 10.6 f 13.9% above baseline. Symptoms, Mood, and Attention The most frequent complaints of the eight patients who eventually required reductions in medication were fatigue, drowsiness, diarrhea, and poor concentration (see Table 3). In addition, phentermine was never started in a ninth patient because of blood pressure readings >140/90 mm Hg. Results of the symptom questionnaire, completed by all 25 patients, showed that complaints of dry mouth increased significantly (p<0.02) from baseline to week 4 and remained significantly Q~0.04) elevated at week 26 (but not week 52). At its highest point (at week lo), patients indicated that this symptom troubled them somewhat to a moderate amount (mean=2.6 f 1.1). Complaints of constipation increased significantly (p<0.02) from baseline to week 10 and remained significantly (p<0.05) elevated at week 26 (but not week 52). No other side effects increased significantly during treatment, although more complaints clearly would have been observed on the symptom questionnaire if medication had not been reduced in eight of 25 patients. Mood and Attention. There were no significant differences between the two conditions in changes in mood, at- OBESITY RESEARCH Vol. 5 No. 3 May

5 ~ ~~~ ~ Behavioral-Pharmacologic Treatment of Obesity, Wadden et al. Table 3. Medication doses and symptom complaints in nine patients on reduced doses of fenfluramine and/or phentermine Case Treatment condition Fenfluramine Phentermine Symptoms 1 Group 60 (60) mg/d' Never started blood pressure > 140/90 mm Hg 2 Individual 60 (40) mg/d 15 mg/d drowsiness, diarrhea 3 Individual 60 (40) mg/d 15 mg/d fatigue, stomach pain 4 Group 60 (40) mg/d 15 mg/d drowsiness, fatigue, poor concentration 5 Group 60 (40) mg/d 15 mg/d dry mouth, dry eye sockets 6 Group 60 (40) mg/d 15 mg/d fatigue, minor forgetfulness 7 Group 60 (40) mg/d 15 (7.5) mg/d light headed, poor concentration 8 Individual 40 (20) mg/d 15 mg/d constipation, diarrhea 9 Individual 40 (20) mg/d 15 mg/d fatigue 1 ' First value shows the dose that was associated with side effects listed under symptoms. The value in parentheses shows the dose prescribed to alleviate side effects. tention, or appetite; thus, scores reported here are collapsed across conditions. The mean baseline score of 7.5 f 7.1 on the Beck Depression Inventory declined significantly (p<o.ool) to 2.7 f 4.7 at the end of treatment. Thus, patients reported fewer symptoms of depression over time. Mean (scaled) scores on the digit span test of the Wechsler Memory Scale increased significantly (~~0.05) during the year from 10.3 f 2.3 to 11.1 f: 2.6. The change indicates that attention span (i.e., concentration) increased over time, although, once again, different results might have been ob- tained if medication doses had not been reduced in eight patients. Appetite and Eating Behavior Patients reported reductions in food cravings, hunger, and preoccupation with eating, which reached their nadir the 10th week of treatment (see Table 5). Only the decline in preoccupation with eating differed significantly from baseline, which it did at weeks 4, 10, and 52 (allp values<0.05). Reports of increased satiety approached significance at Table 4. Change in lipids and lipoproteins in patients in the two conditions combined Variable Baseline Week 6 Week 18 Week 26 Week 52 Weight (kg) 97.6 f 13.0 Triglycerides (mg/dl) f 67.3 Total Cholesterol (mg/dl) f 32.5 LDL Cholesterol (mg/dl) f 28.1 HDL Cholesterol (mg/dl) 50.3 f 10.2 Cholesterol/HDL Ratio 4.6 f f 13.0' -5.4 f f: 49.3" f f 37.9' f f 31.3" f f 9.2b f f 1.3" 13.9 f f 13.4' f f 33.6" f f 31.6' f f 27.4' f f 9.5' f f f f 14.2' 14.6 f f 44.1' f f 32.9' -9.0 f f 27.6" -9.0 f f f f 1.1" -7.7 f f 15.7' 16.8 f f f 29.0" f 34.3' -9.7 f f 26.8' f: f 12.4" 10.6 f f:.89' f 13.6 Values are means f SD. "p%o.ol for comparison with baseline. bp<o.oo1 for comparison with baseline. 222 OBESITY RESEARCH Vol. 5 No. 3 May 1997

6 ~ ~~ ~~ Behavioral-Pharmacologic Treatment of Obesity, Wadden et al. Table 5. Changes in appetite and Eating Inventory scores in patients in the two conditions combined V a r i a b 1 e Baseline Week 4 Week 10 Week 18 Week 26 Week 52 Appetite Craving 28.3 f 16.0 Hunger 48.8 f 20.3 Preoccupation 41.1 f 15.9 Satiety 60.9 f 12.0 Eating Inventory Cog. Restraint 9.3 f 4.2 Disinhibition 10.5 f 2.8 Hunger 6.7 f f f f f f f f f f f f 17.0b 25.0 f 14.1b 29.5 f f f 18.9" 67.9 f f 14.2" 62.5 f f f f 3.1' 7.1 f 3.5' 3.8 f 3.2" Values are means?r: SD. "p%0.05 for comparison with baseline. 'ps0.01 for comparison with baseline. 'p%o.ool for comparison with baseline. week 4 (~~0.06) and were significantly different from baseline at week 10 (p<0.02); there were no significant differences thereafter. Thus, during the early but not later stages of treatment, patients reported feeling fuller after eating. Eating Behavior. Highly significant changes (all ps<o.ool) were observed from baseline to week 52 on all three scales of the Eating Inventory (see Table 5). Cognitive restraint increased, whereas both disinhibition and hunger decreased. Together, these findings indicate that patients experienced greater cognitive control over their food intake, were less troubled by impulsive eating, and reported fewer complaints of hunger. Neither baseline scores nor changes on these three scales correlated significantly with weight loss at any time. Behavioral Adherence There were no significant differences between the two conditions in patients' completion of food records. As shown in Table 6, adherence to this task was generally favorable, although the percentage of possible records completed during weeks was significantly lower than that during weeks and that during weeks 5-10 (p<o.ool). Patient completion of food records was positively related to weight loss. Thus, the number of records completed during the first 10 weeks correlated with weight loss at week 10(r=0.45, p<0.03), week 18 (r=0.52, p<0.008), week 26 (r=0.47, p<0.02), and week 52 (r = 0.42, p<0.05). Similarly, the total number completed during the first 18 weeks correlated with weight loss at week 18 (r=0.55, p<0.004), week 26 (r=0.53, p<0.007), and week 48 (r= 0.48, p<0.02). Results of multiple regression analyses showed that the combination of early weight loss and completion of food records did not predict later weight loss (at weeks 26 or 52) significantly better than either of these variables alone. Discussion Obese women treated by the fenfluramine-phenterine combination in 10 brief but structured office visits achieved the same highly successful results at one year as those who received medication combined with 32 group behavior modification sessions. Thus, intensive group behavior modification, as provided in this and Weintraub's (37) study, does not appear to be necessary for successful weight loss with the fenfluramine-phentermine combination. Instead, our results strongly suggest that primary care providers could effectively treat obese individuals Table 6. Completion of diet diaries by patients in the two conditions Condition Week 4 Week 10 Week 18 Group Treatment 27.1 f 12.1" 59.6 f 27.5' 95.1 f 47.0' % completed 77.4 f f f 35.4 Physician Visits 28.0 f f f 42.9 % completed 73.8 f f f 35.9 Mean (both groups) 27.5 f f f 44.3 % completed 75.6 f f k 35.3 "Of a possible total of daily 35 diaries. bof a possible cumulative total of 77 daily diaries. "Of a possible cumulative total of 133 daily diaries. OBESITY RESEARCH Vol. 5 No. 3 May

7 with a combination of medication and brief, protocolguided lifestyle modification. Clearly, these results await replication using larger patient samples, which include men, and using physicians who do not specialize in the treatment of obesity, as did the psychiatrist who participated in this investigation. Such studies should compare the results of brief, protocol-guided therapy not only with group behavior modification but also with treatment by pharmacotherapy alone, as is likely to be dispensed in busy primary care practices. We, like others (4,19,24,25,26), believe that the best results will be obtained when patients combine medication with their own efforts to modify eating and exercise habits. This belief is supported by the finding in the present study that the more food records patients completed during the first 18 weeks of treatment, the greater their short- and long-term weight losses. By contrast, the poorest results are likely to be obtained by individuals who mistakenly believe that they can eat whatever they want simply because they are on a weight loss medication. Medications such as dexfenfluramine and the fenfluramine-phentermine combination may well facilitate patients efforts to change their eating habits by increasing satiation at mealtime (7), as shown in this study, and by decreasing preoccupation with eating. Comments of one patient echoed those expressed by several: For the first time in my life, I feel normal. I m not always thinking about what I m going to eat next or when. The highly significant changes on the cognitive restraint, disinhibition, and hunger scales of the Eating Inventory similarly indicate that patients felt more in control of their eating. Weight loss in this study was associated with marked improvements in lipids and lipoproteins, as reported by other investigators using diet alone (6,40), or diet combined with the fenfluramine-phentermine protocol (4,38). Our results confirm that weight losses of 10%-15% of initial weight are frequently sufficient to normalize lipid levels, despite patients remaining significantly obese after weight loss (6,18,21). We also replicated previous findings in obese women that HDL cholesterol may decline sharply in the initial stages of treatment before patients have lost substantial amounts of weight (2). Practitioners, however, should not be overly alarmed by these short-term reductions; by the end of treatment, HDL cholesterol had rebounded to a level 10.6% greater than baseline, while LDL cholesterol had fallen 12.7% below pretreatment levels. Further studies are needed to determine the factors which contribute to short- as compared with long-term changes in lipids and lipoproteins (16,36). The present results suggest that doses of anorectic medications can be adjusted slightly to control adverse side effects, without sacrificing weight loss. After fenfluramine was reduced in eight patients, the 25 participants, as a group, reported increased complaints only of dry mouth and constipation, which at their highest levels, troubled patients less than a moderate amount. One patient complained of minor forgetfulness and two of poor concentration. Results of the digit span testing revealed significantly improved, rather than worsened, concentration. Clearly, these findings might have been different if medications had not been adjusted in patients. Further studies using more sophisticated measures of cognitive function are needed to assess complaints of memory loss, which were reported by about 10% of persons treated in a large private practice setting (4). Patients in the present study also reported significant improvements in mood with weight loss, thus, confirming previous findings (33). We note that the baseline scores of patients on both the digit span and depression scales fell within the normal range. Thus, although the observed changes were statistically significant, they were not, on average, clinically significant. This study did not address the very important issue of whether lifestyle counseling, provided either in brief structured visits or group sessions, improves the treatment of obesity by pharmacotherapy alone. Two studies, however, suggest that it does (13,14). Craighead and colleagues (14), for example, found that persons treated by 120 mg/day of fenfluramine, and who were examined only once a month in routine office visits, lost an average of 6.0 kg during 26 weeks. By contrast, those treated by the identical medication, but who also attended weekly group behavior modification sessions, lost a significantly greater 15.3 kg. Thus, the addition of lifestyle modification dramatically improved the results of treatment by pharmacotherapy alone. Patients in a third condition, who received medication combined with weekly group social support, rather than behavior modification, also achieved a large weight loss (of 14.5 kg). This latter finding suggests that frequent weigh-ins and group support may improve the results of pharmacotherapy for obesity, independent of the specific content of group meetings. Similarly, results of the Craighead (14) study, taken with those from the present investigation, suggest that lifestyle counseling provides an incremental benefit over treatment by medication alone, whether such counseling is delivered under conditions of intensive or minimal contact. These hypotheses await further testing. Practitioners must carefully select patients for pharmacotherapy and assess with each the benefits and risks of treatment; these risks include primary pulmonary hypertension (1,23). In cases in which pharmacotherapy is elected, the results of this pilot study demonstrate the feasibility of providing effective weight management in primary care practice by the use of structured office visits. Given demands on physician time, however, which are unlikely to allow a patient with obesity multiple minute office visits over the course of a year, future studies should examine whether brief lifestyle counseling can be provided in primary care settings by dietitians or other nonphysician personnel. The use of such staff should reduce the cost of 224 OBESITY RESEARCH Vol. 5 No. 3 May 1997

8 treatment even further from that achieved by using infrequent, brief lifestyle counseling as compared with traditional group behavior modification. In an exploratory analysis, we calculated the cost of each 1-kg reduction in body weight for patients in the two conditions who completed our 52-week program. When the costs of medications, physician visits, blood tests, group treatment sessions, and patients time (i.e., missed work and leisure time) were included, we found that each 1-kg reduction in body weight cost $182 for patients in the individual physician condition, as compared with $246 for participants in the group behavior modification condition (a difference that was not statistically significant because of the limited power of the study, p<0.20). Costs were $169 and $187 per kilogram of weight loss for the two conditions, respectively, when the value of patient time was not included. These admittedly crude estimates did not include costs for such factors as child care and parking or taxi fees, which were greater in the traditional group behavior modification condition and, thus, increase the relative cost-effectiveness of treatment by brief physician visits. Future studies should include cost-benefit analyses that calculate not only the costs of treatment but also the savings that result from weight loss and improvements in obesity-related health complications. Such savings include reductions in sick days (and lost productivity), medication use, and doctor visits and hospitalizations to treat obesity-related comorbidities. Further studies are needed to determine whether the favorable results of brief physician visits, obtained in this study with the fenfluramine-phentermine combination, can be achieved with other medications, such as dexfenfluramine, whether used singularly or in combination. Moreover, long-term studies of three to five years duration are needed to determine whether medications must be taken continuously in order to maintain weight losses and improvements in health (29). The need for continuous therapy is suggested by the results of several studies which found that weight was regained rapidly when fenfluramine (13,14), as well as the fenfluramine-phentermine combination (39), was terminated. The benefits of chronic pharmacotherapy for obesity must be carefully weighed against its risks. Acknowledgments We thank Wyeth-Ayerst Laboratories for providing dlfenfluramine (i.e., Pondimin) and Fissons Pharmaceuticals and Medeva Pharmaceuticals for providing phentermine (i.e., Ionamin). Sharon Sasanow, M.S., R.D., is acknowledged for providing group behavior therapy. Copies of the protocol used to implement the brief physician visits may be obtained from the first author. The protocol describes how to implement treatment using the LEARN Program for Weight Control (Dallas, TX: American Health Publishing Co.). This research was supported by an NIMH Research Scientist Development Award (to Dr. Wadden), an NIH Physician Scientist Award (to Dr. Berkowitz), and by NIH grant DK (to Dr. Wadden). Drs. Wadden, Berkowitz, Stunkard, and Foster have spoken at conferences which were supported in part, or in full, by Wyeth-Ayerst Laboratories, which markets dlfenfluramine (Pondimin) and d-fenfluramine (Redux). References 1. Abenhaim L, Moride Y, Brenot F, et al. Appetitesuppressant drugs and the risk of primary pulmonary hypertension. N Engl J Med. 1996;335: Andersen RE, Wadden TA, Bartlett SJ, Vogt RA, Weinstock R. Weight loss and changes in serum lipids and lipoproteins in obese women. Am J Clin Nutr. 1995;62: Astrup A, Toubro S, Cannon S, Hein P, Madsen J. Thermogenic synergism between ephedrine and caffeine in healthy volunteers: A double-blind, placebo controlled study. Metabolism. 1991;40: Atkinson RL, Blank RC, Loper JF, Schumacher D, Lutes RA. Combined drug treatment of obesity. Obes Res. 1995;3: 497s-500s. 5. Beck AT, Steer R. Manual for the Beck Depression Inventory. New York: Psychological Corp, Blackburn GL. Effect of degree of weight loss on health benefits. Obes Res. 1995;3:211S-216S. 7. Blundell JE, Lawton CL. Pharmacological aspects of appetite regulation. In: Stunkard AJ, Wadden TA, eds. Obesity: Theory and Therapy. New York: Raven Press; 1993; Bray GA. Use and abuse of appetite-suppressant drugs in the treatment of obesity. Ann Intern Med. 1993;119: Bray GA, Ryan DH, Gordon D, Heidingsfelder S, Cerise F, Wilson K. A double-blind randomized placebo controlled trial of sibutramine. Obes Res. 1996;4: Brownell KD. The LEARN Program for Weight Control (6th edition). Dallas: American Health Publishing; Brownell KD, Rodin J. The Weight Maintenance Survival Guide. Dallas: American Health Publishing; Campfield L, Smith F, Guisez Y, Devos R, Burn P. Recombinant mouse OB protein: Evidence for a peripheral signal linking adiposity and central neural networks. Science. 1995; 269: Craighead LW. Sequencing of behavior therapy and pharmacotherapy for obesity. J Consult Clin Psychol. 1984;52: 19G Craighead L, Stunkard AJ, O Brien R. Behavior therapy and pharmacotherapy for obesity. Arch Gen Psychiatry. 1981; 38: Drent ML, van der Veen EA. Lipase inhibition: a novel concept in the treatment of obesity. Int J Obes. 1993;17: Ellis RW, Darga LL, Lucas CP. The short- and long-term effect of a low-fat, cholesterol-free, hypocaloric diet on serum OBESITY RESEARCH Vol. 5 No. 3 May

9 triglyceride and cholesterol distribution in severely obese humans. Int J Obes. 1987; 11: Foster GD, Wadden TA, Peterson FJ, Letizia KA, Bartlett SJ, Conill AM. A controlled comparison of three vev-lowcalorie diets: Effects on weight, body composition, and symptoms. Am J Clin Nutr. 1992;53: Goldstein D. Beneficial effects of modest weight losses. Int J Obes. 1992;16: Goldstein D, Potvin J. Long-term weight loss: The effects of pharmacologic agents. Am J Clin Nutr. 1994;60: Guy-Grand B, Apfelbaum M, Crepaldi G, Greis A, Lefebvre P, Turner P. International trial of long-term dexfenfluramine in obesity. Lancer. 1989;i: Institute of Medicine, National Academy of Sciences. Weighing the Options: Criteria for Evaluating Weight Management Programs. Washington, D.C.: Government Printing Office; Kuczmarski J, Flegal K, Campbell S, Johnson C. Increasing prevalence of overweight among US adults: The National Health and Nutrition Examination Surveys, I. JAMA. 1994;272: Manson JE, Faich GA. Pharmacotherapy for obesity-do the benefits outweigh the risks? N Engl J Med. 1996;335: Marcus MD, Wing RR, Ewing L, Kern E, McDermott M, Gooding W. A double-blind, placebo-controlled trial of fluoxetine plus behavior modification in the treatment of obese binge-eaters and non-binge-eaters. Am J Psychiatry. 1990; 147: National Task Force on the Prevention and Treatment of Obesity. Long-term pharmacotherapy in the management of obesity. JAMA. 1996;276: Pi-Sunyer XF. Guidelines for the approval and use of obesity drugs. Obes Res. 1995;3: Scoville BA. Review of amphetamine-like drugs by the Food and Drug Administration: Clinical data and value judgements. In: Bray GA, ed. Obesity in Perspective. Bethesda, MD: DHEW Pub1 NO (NIH) ; 1975; Stunkard AJ, Messick S. Manual for the Eating Inventory. New York: Psychological Corp; Stunkard AJ. Minireview: Anorectic agents lower a body weight set point. Life Sci. 1982;30: US. Food and Drug Administration. FDA Talk Paper: FDA approves dexfenfluramine for obesity treatment. April 29, Wadden TA, Stunkard AJ, Day SC, Gould RA, Rubin CJ. Less food, less hunger: Reports of appetite and symptoms in a controlled trial of a protein-sparing modified fat. Int J Obes. 1987;11: Wadden TA, Foster GD. Behavioral assessment and treatment of markedly obese patients. In: Wadden TA, VanItallie TB (eds.). Treatment of the Markedly Obese Patient. New York: Guilford Press 1992; Wadden TA, Steen SN, Wingate BJ, Foster GD. Psychosocial consequences of weight reduction: how much weight loss is enough? Am J Clin Nutr. 1996;63: Wadden TA, Sarwer DB. Behavioral treatment of obesity: New approaches to an old disorder. In: Goldstein D (ed.). The Management of Eating Disorders. Totowa, NJ.: Humana Press; in press. 35. Wechsler D. Wechsler Memory Scale-revised manual. San Antonio, Tx.: Psychological Corp; Weinsier RL, James DL, Darnell BE, et al. Lipid and insulin concentrations in obese postmenopausal women: separate effects of energy restriction and weight loss. Am J Clin Nutr. 1992;56: Weintraub M, Sundaresan P, Madan M, et al. Long-term weight control study I (weeks 0-34). Clin Pharm Ther. 1992; 51:58& Weintraub M, Sundaresan PR, Schuster B. Long-term weight control study VII (weeks 0 to 210): Serum lipid changes. Clin Pharm Ther. 1992;51 : Weintraub M, Sundaresan P, Schuster B, Moscucci M, Stein E. Long-term weight control: Study I11 (weeks ). Clin Pharm Ther. 1992;51: Wing RR, Jeffery RW. Effect of modest weight loss on changes in cardiovascular risk factors: Are there differences between men and women or between weight loss and maintenance? Int J Obes. 1995;19: Yen TT. B-agonists as antiobesity, antidiabetic, and nutrient partitioning agents. Obes Res. 1995;3: OBESITY RESEARCH Vol. 5 No. 3 May 1997