original

communications

Ineffectiveness of in weight reduction:

human chorionic a double-blind

gonadotropin study13
William Hinshaw,6M.S.

Ronald E. Julis, E. Sawicki, M.D.,

ABSTRACT chorionic duplicate HCG greater participated placebo. weighed administered examinations placebo groups groups of the two circumference, effectiveness 1976. Each daily had

Our gonadotropin the Asher-Harper a statistically loss in our patient Monday the and weight

investigation (HCG) significant and 32-day decrease study

was on weight (1973) benefit in

designed reduction which over

to retest in obese had found the diet Fifty-one (the

the that and women

hypothesis

of the efficacy setting.

of human to and by 60 was who physical in the versus

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

women placebo

in a clinic combination the

We sought as evidenced

the combination between

of 500 cal diet ages of of 18 and study), initial

hunger. diet and studies

prospective, the same Saturday study.

randomized, was were

double-blind one prescribed counselled performed of 25 in the weight loss, by at HCG percent HCG

comparison in the Asher-Harper one the and of the time of

HCG

was given through

investigators

injections. at the end in number

Laboratory of the 28 injections. of injections regimen

Twenty received, or in hunger for weight

21 of 26 patients difference loss,

completed weight of a rigidly

There

was no statistically ratings. reduction.

significant does
Am.J.

in the means hip and waist to enhance the 29: 940 -948,

of weight not

loss per injection, imposed

appear

Clin.

In 1954, Simeons (I) published his technique for obese persons to achieve rapid weight loss without severe hunger. The program was designed for clinic patients. His method, which has undergone some minor modifications (2), consists of a rigidly structured 500 calorie diet and intramuscular injections of 125 units of human chorionic gonadotropin (HCG) given during clinic weight checks 6 days per week (1-5). Simeons (1-3) has claimed that weight loss with this combination is more successful than weight loss achieved with a 500 calorie diet alone because HCG suppresses appetite and improves mood; thus, the patient complies better with the prescribed diet. He claimed patients will lose an average of 250 to 600 g per day. The reported side-effects were hypoglycemia, increased libido, and gout (1). Numerous reports have been published attempting to support (5-10) or to disprove
The American Journal of Clinical Nutrition

(11-15) the Simeons theory on the usefulness of HCG in weight reduction. However, until Asher and Harpers report (4), there had not been a well-designed, prospective, randomized, double-blind study comparing HCG and
From the Departmept of Medicine, Pharmacy Serv-

ice, and Department of Pathology, Letterman Army Medical Center, and Letterman Army Institute of Research, Presidio of San Francisco, California. 2 The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense. Address reprint requests to Technical Publications Editor, Letterman Army Medical Center, Presidio of Francisco, California 94129. Current address: Department of Medicine, Fitzsimons Army Medical Center, Denver, Colorado 80240. Blood Transfusion Division, Letterman Army Institute of Research. e Current address: 7454 Mill Run, Germantown, Tennessee 38138. Chief, Department of Medicine, Letterman Army Medical Center. San

940

29: SEPTEM

BER

1976,

pp.

940-

948.

Printed

in U.S.A.

HCG

IN

WEIGHT

REDUCTION

941

placebo in the setting designed to test the true Simeons technique. Their study (4) yielded evidence that HCG had a statistically significant benefit over placebo; this was reflected by a significantly greater mean weight loss, mean weight loss as percentage of initial weight, mean weight loss per injection, and decrease in hunger. We reviewed Asher and Harpers original article (4); the analysis of the investigators statistical presentation by Hirsch and Van Itallie (16); the rebuttal by Asher and Harper (17), and the Food and Drug Administrations reports (18, 19) which asserted that HCG has not been demonstrated to be an effective adjunctive therapy in weight reduction programs. We reviewed the original protocol of the Asher-Harper study, (Dr. W. L. Asher kindly provided us with a copy of the original Asher-Harper protocol, entitled ASB Double-blind HCG Study). Then we designed our investigation to retest the hypothesis of the efficacy of HCG and incorporated into our design the essential features of the Asher-Harper study. We hoped that a duplicate study performed by a different group could help to settle the controversy about the role of HCG in the treatment of obesity. Materials and methods

lar, patients with a history of heart disease, renal disease, asthma, gout, hypoglycemia, migraine, genital malignancy, or breast malignancy were excluded. All patients had to be well-motivated, be willing to obtain a postage scale to weigh their food, and be available to come for a clinic visit every morning Monday through Saturday for the entire study period of 32 days. A person was excluded if she I) had been treated with HCG previously; 2) had taken appetite suppressants in the preceding 3 months; 3) had a base-line weight which was greater than or less than 10 lb of what it had been 3 months before the study: 4) was pregnant, within 4 months postpartum, or nursing an infant; or 5) had been taking medications other than estrogens or birth control pills, (exceptions are noted in Table I). Participants had to be between 20 and 60 pounds weight over the ideal weight per was defined as the midpoint given height. Ideal of range at a given to are

height for the appropriate Metropolitan Life Insurance noted in Table I).

frame size according tables, (exceptions

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

Parameters

measured

On -a daily basis, during the morning clinic hours (between 8:00 and 11:00 AM) Monday through Saturday, each patient, wearing only minimal underclothes,

was weighed after voiding. tered the weights throughout

The same two nurses registhe study. At this time, the

patient also indicated her degree of appetite according to the following five-point scale: I) absence of appetite (can hardly eat all the food permitted): 2) normal appetite (content with the food allowed); 3) mildly hungry (would eat more food if permitted); 4) modertately hungry (desire for food is bothersome); and 5) severe hunger (desire more food most of the time and have been tempted or have eaten more). Each patient had her progress reviewed by one of the investigators who was unaware of the patients treatment group.

Patients were solicited for the study by an announcement in the medical centers bulletin. All respondents were screened by a questionnaire and each received a full explanation of the double-blind nature of the study. Each patient selected gave her informed consent to participate. A participant was given a complete physical examination and a battery of laboratory tests including a complete hemogram, chemistry screening tests, thyroid and pituitary function tests, and was then was assigned a patient number. (The sequence in which patients took the physical cation treatment subjects by two injections lasting include United Patient examination arbitrarily determined the identifinumber.) The numbers were randomized into two groups by a registered hospital pharmacist. All were given thorough instructions on the diet plan staff dieticians. The entire study group began the and diet on the same day. The study period, 32 days (April 21 to May 22, 1975), did not any holidays traditionally celebrated in the States. selection

On a weekly

basis, a sitting
were at least

blood

pressure,

and hip
same with

and waist measurements nurse. Each patient had the dietician.

obtained a weekly

by the interview

A the conclusion
physical survey examination was repeated

of the test period,

a second complete
the laboratory

was performed and for each participant.

injections The HCG was purchased from Park-Davis and Company, Detroit, Mich. A registered hospital pharmacist packaged each individual dose in a glass (disposable) syringe, labeled with the corresponding patient number. Each syringe contained a 1 ml volume of either 125 units of HCG or normal saline (the standard diluent for this HCG), and each appeared identical. A fresh supply for each patient was prepared every 10 days and kept under refrigeration. One ml injections of still cold solution were given deep intramuscularly in the upper outer quadrant of the buttocks by a nurse who did not know the contents of the syringe. Patients returned to the clinic, which was conducted on the Metabolic Ward, Letterman Army Institute of Research, 6 days each week for 28 consecutive injections over a 32-day test period.

All patients were women between 18 and 60 years of age who voluntarily responded to the announcement.
Race was not a factor in the selection. No one with more

than minimal

health

problems

was accepted.

In particu-

942 Diet

STEIN

ET AL.

The diet was identical to that used in the protocol of Asher and Harper, (W. L. Asher, M.D., Executive Director, American Society of Bariatric Physicians, personal communication). Patients were encouraged to drink 8 to 10 glasses of water daily and were advised of the need for strict compliance with the diet. Our diet instruction sheets ended with the statement, Any slight change in the above diet rules will result in downright disappointment. The patients were encouraged to keep a food diary and were impressed that a reason must be found if daily weight loss was not achieved. Miscellaneous Patients those were advised to avoid all medications prescribed by one of the three physicians other (M.

than

S., R. J. or J. D.) who would

see them daily.

They

were

advised to avoid laxatives but were permitted a Fleets enema, if needed. For more persistent constipation, dioctyl sodium sulfosuccinate was prescribed. Patients were advised to avoid the use of all cosmetics containing fats or oils, as well as any skin contact with fats or oils. Statistical methods

Outcome measures of weight reduction (weight loss, percentage of initial weight loss, weight loss per injection) and comparability of treatment group statistics (age, height, initial weight, initial blood pressure, race) were all tested (untransformed) for treatment differences using the Students t test (20) or Fishers exact test (21). In addition, the arcsine transformation (20) was applied to the fractional weight loss and fractional weight loss per injection in an attempt to stabilize the variance of the proportions: again students t test was applied. In order to improve upon the lack of robustness of Students t test to departures of the data from normality, the MannWhitney U test was applied to all untransformed outcome measures (21). The hunger ratings for each treatment group were tested by using the Mann-Whitney U test for two independent samples making the correction for ties suggested by Hollander and Wolfe (22). The blood test results were entered into a two-way analysis of variance (ANOVA), repeated measures design, and the analysis of variance tables were examined for timedependent differences, treatment group differences, as well as treatment-time interactions (BMD-08V (23) implemented via remote terminal on CDC 7600 Lawrence Berkeley Laboratories, St. Paul, Minn.). In order to provide the required equal ns for the program for this analysis, one case in the placebo group (patient 20) was randomly deleted from the 23 complete sets of biochemical data. Differences significant if the one-tailed when the differences were found by used when ences were were considered statistically P value was less than 0.05 in the same direction as those

group and 21 of the 26 in the placebo group completed 28 injections; there was no significant difference between the groups in the mean number of injections received (Table 1). Data on all starting patients were included in the final analysis whenever possible. Patients 12 and 28 lost their motivation and stopped participation. Patient 55 received transfer orders. Patient 19 left on emergency leave, and patient 26 left on an unanticipated vacation. Laboratory studies on patient 39 showed evidence of hypothyroidism and she was started on thyroid replacement after receiving her twentieth injection. Because of this, she was listed dropped from study. Other patients who missed an occasional injection claimed their absence was due to either illness or business responsibilities. Table 2 summarizes the measures used to satisfy the requirement of comparability of groups and adequacy of randomization. No statistically significant differences emerged. Weight reduction data are listed in Table 1. Raw weight loss (column 6) was used to compute other weight reduction measures, which take into account initial body weight and number of injections. These appear separately in columns 7 and 8 and are combined in column 9. All analyses outlined in the Methods section were performed; no statistically significance differences were found between the two treatment groups on any test. Although the direction of the differences favored the HCG group, the differences were small and in no instance was the P value (one-tailed) less than 0.25; neither were there statistically significant differences between the two treatment groups when all of the patients who did not meet all the study criteria were omitted from the analyses and the analyses were repeated. A somewhat heterogeneous distribution of the proportion of responses in each hunger category emerged between the treatment groups (Table 3). However, no statistically significant difference could be found using the Mann-Whitney U test (corrected for ties), a test that specifically takes into account the ordering implied by the response rankings (1+ to 5+) (21; G. Noether, Professor, University of Connecticut; personal communication with Carl C. Peck, M.D.). The mean percentage changes in waist and

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

Asher and Harper (4): a two-tailed test was no notion as to the direction of possible differagreed upon before the trial.

Results Of sians) the 51 patients (9 blacks and 42 caucastarting the study, 20 of 25 in the HCG

HCG TABLE
Summary

IN WEIGHT

REDUCTION

943

I
of patients and weight reduction data No. of injections Starting weight (Ib) Percent body weight loss Loss, lb per injection Percent weight loss per injection

Patient

Age (yrs)

Height (in)

Loss (Ib)

HCG

group
42 42 27 30 21 25 28 25 42 33 26 27 53 64 26 32 35 48 52 21 34 25 26 19 34 33.5 2.3 group 42 34 38 25 25 52 32 34 39 20 31 22 20 50 45 51 26 23 51 36 38 20 24 24 30 27 67.0 57.0 68.0 62.5 64.0 61.0 65.0 62.5 62.8 70.0 64.0 53.0 67.0 65.0 59.5 60.0 64.0 71.0 62.0 71.0 66.0 66.0 68.0 70.0 63.0 62.0 64.3 28 28 28 28 28 28 27 28 28 26 28 21 6 28 27 28 28 28 28 28 28 28 28 28 28 28 26.8 172.0 144.0 182.8 168.8 159.8 152.2 154.5 149.4 170.6 174.0 147.6 194.4 156.8 142.0 141.6 136.2 188.8 182.4 167.4 247.2 205.0 157.8 169.2 181.8 147.0 165.8 167.7 18.9 15.2 23.4 21.4 13.7 17.8 15.7 11.0 17.7 9.2 12.7 16.6 5.2 14.3 13.3 12.7 13.6 15.4 12.4 22.2 25.4 10.0 20.0 16.9 18.2 10.7 15.5 11.0 10.6 12.8 12.6 8.6 11.7 10.2 7.4 10.4 5.3 8.6 8.5 3.3 10.1 9.4 9.3 7.2 8.4 7.4 9.0 12.4 6.3 11.8 9.3 12.4 6.5 9.3 0.675 0.543 0.836 0.764 0.489 0.636 0.581 0.393 0.632 0.341 0.454 0.755 0.867 0.511 0.493 0.454 0.486 0.550 0.443 0.793 0.907 0.357 0.714 0.604 0.650 0.382 0.59 0.392 0.377 0.457 0.453 0.306 0.418 0.376 0.263 0.371 0.196 0.307 0.388 0.553 0.360 0.348 0.333 0.257 0.302 0.265 0.321 0.443 0.226 0.422 0.332 0.442 0.231 0.35 61.5 64.5 60.5 70.0 61.0 61.0 63.0 66.0 61.0 65.0 64.0 65.0 65.0 65.0 64.0 61.0 68.0 64.0 65.0 63.0 67.0 61.0 66.0 65.0 68.0 64.2 0.5 28 28 28 28 28 23 28 27 28 23 28 28 28 28 28 28 20 28 28 28 28 28 28 9 28 26.5 0.8 172.8 152.0 160.0 164.8 142.0 160.0 153.4 181.4 158.4 175.8 178.2 161.8 145.0 159.4 220.4 176.0 196.2 159.8 184.0 167.6 168.8 155.0 143.0 147.0 165.8 165.9 3.5 20.1 14.3 17.2 19.4 13.7 9.8 16.0 15.0 17.4 10.8 19.6 19.3 15.4 11.8 30.7 15.2 12.6 17.3 17.9 19.6 14.2 12.3 14.5 9.9 10.7 15.8 0.9 11.6 9.4 10.8 11.8 9.6 6.1 10.4 8.3 11.0 6.1 11.0 11.9 10.6 7.4 13.9 8.6 6.4 10.8 9.7 11.7 8.4 7.9 10.1 6.7 6.5 9.5 0.4 0.718 0.511 0.614 0.693 0.489 0.426 0.571 0.556 0.621 0.470 0.700 0.689 0.550 0.421 1.096 0.543 0.630 0.618 0.639 0.700 0.507 0.439 0.518 1.100 0.382 0.61 0.04 0.415 0.336 0.384 0.420 0.345 0.278 0.373 0.306 0.392 0.279 0.393 0.426 0.379 0.264 0.498 0.308 0.338 0.387 0.347 0.418 0.300 0.283 0.362 0.748 0.231 0.37 0.02

3 4 6 8 9 12 13 17 18 19 22 27 31 32 35#{176} 38 39 41
450

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

46 47 51 52 55 58 Mean SEM Placebo 2 7 10 II 14 15 16 20 21 23 26#{176} 28 29 30 33

36 37 42#{176}
430

49 50 53 54 56

Mean SEM
#{176} Patients (chlorpromazine);

33.0 2.1
weight

0.8

0.9

4.7
given

0.9

0.5

0.03
20 was taking tolbutamide.

0.02
Thoraiine

exceeded 60 lb over the ideal weight per patient 33 was taking Synthroid (levothyroxine):

height. Patient patient 37 was taking

944

STEIN

ET AL.

hip circumferences are summarized in Table 4. Although both treatment groups experienced significant decreases in these circumferences, the HCG group did not have a significantly greater reduction compared to the placebo group. Complete sets of pretreatment and posttreatment biochemical data were obtained from 22 patients in the HCG group and 23 patients in the placebo group. Table 5 summarizes the changes in blood studies as well as the two-way repeated measures analysis of variance. The mean changes (column 4) include all complete sets of data, whereas, the F ratios with P values (columns 5 and 6) are based on 22 sets of data in each test group

(see Statistical methods). The independent effect of HCG versus placebo on the change in each measurement is reflected by the time-treatment interaction F-ratio (column 6). During the treatment period, both groups showed significant decreases (<0.005) in hematocrit, white blood cell counts, blood urea nitrogen, cholesterol levels, triglyceride values, and total protein levels. Significant increases occurred in the fasting glucose and uric acid levels. The increase in fasting glucose appeared to be significantly greater in the placebo group than in the HCG group. No other significant time-treatment interactions were discovered. Both groups experienced significant down-

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

TABLE
Comparability

2
of treatment Measure groups Group No.of patients Mean SEM Test statistic value
P

Age(yr) Height Initial Initial Initial Ratio: (in) weight systolic diastolic blacks/total (Ib) blood blood pressure, pressure, (mmHg) (mm Hg)

HCG Placebo HCG Placebo HCG Placebo HCG Placebo I-ICG Placebo HCG

25 26 25 26 25 26 25 26 25 26 25

33.5 33.0 64.2 64.3 165.9 167.7 123.9 123.4 77.0 77.8 6/25

2.3 2.1 0.5 0.8 3.5 4.7 2.8 2.9 2.2 2.2

0.14 0.08 0.29 0.11 0.27 exact 0.44

NS#{176} NS NS NS NS test:

Placebo
#{176}

26
P < 0.05 level.

3/26

Fishers P

NS,

not

statistically

significant

at the two-tailed

TABLE Hunger

3 responses None Little 392 (43.6%)#{176} 410 (56.8%)#{176} Some 181(26.2%) 155 (2l.5%)#{176} hunger category. Much 52 (7.5%)#{176} 74 (lO.2%)#{176} Uncontrollable 17(2.4%) 9(1.2%)

HCG Placebo
#{176}

149 (20.2%)#{176} 74(10.2%)

Percentage of HCG

or placebo group responses in each

TABLE 4 Percentage decrease

in waist and hip circumferences

value

Waist circumference Hipcircumference

HCG Placebo HCG Placebo

22 24
22

7.21 (0.48) 6.68(0.44)

0.81
0.45

NS#{176}
NS

24
P < 0.05

5.53(0.36)
level.

NS,

not

statistically

FICG

IN

WEIGHT

REDUCTION

945

TABLE
Changes

5
in blood measurements Treatment group No. of patients Mean change (SEM)

F,#{176} (P value)

F,,#{176} (P value)

Hematocrit(%) White blood cell count (cells x l0) Lymphocytes (total white

HCG Placebo HCG Placebo HCG

22 23 22 23 22

blood cells)
Fasting glucose (mg/l00 ml)

Placebo
HCG Placebo HCG Placebo HCG Placebo HCG Placebo HCG Placebo HCG Placebo

23
22 23 22 23 22 23 22 23 22 23 22 23

Blood urea nitrogen (mg/lOOml) Uric acid (mg/ 100 ml) Cholesterol(mg/lOOml) Triglyceride Albumin (mg/l00 (g/lOO ml) ml)

-2.00(0.55) -2.74(0.55) 1.74 (0.29) -l.74(0.65) + 1.00 ( 1.80) +0.35 ( 3.00) +5.80 (3.70) +l6.30(2.80) -3.73 (0.72) -3.44(0.77) +0.63 (0.26) +0.90(0.3I) -36.6( 8.1) -43.7(9.7) 17.6 ( 11.4) -26.2(12.2) +0.1 (0.1) +0.1 (0.1)

21.24 15.55 0.14 21.44 44.31 13.94 37.82 6.61 2.44 14.83

(<0.005) (<0.005) (NS) (<0.005) (<0.005) (<0.005) (<0.005) (<0.025) (NS) (<0.005)

0.35 0.56 0.04 4.63 0.06 0.49 0.26 0.25 0.00 0.13

(NS)5 (NS) (NS) (<0.05) (NS) (NS) (NS)

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

(NS) (NS) (NS)

Total

protein

(g/l00

ml)

HCG
Placebo

22
23

-0.3

(0.1)

-0.3(0.l)

#{176}

F1 and
b

F ratios for
NS, not statistically

ANOVA).

effects of time (t) and time-treatment significant at the two-tailed P <

0.05

interaction level.

(txT),

independently

(2-way

TABLE
Blood

6
pressure changes No. of patients Initial 25 Final 22 Mean pr essure Initial l24(3) Final l08(2) l04(2) 66(l)

Pressure (mm Hg)

Treatmen! group

Systolic Diastolic

HCG

Placebo
HCG

26
25

24
22

123(3)
77(2)

cal power analysis (24) before the onset of the study indicated that our study had a 99+ percentage chance of detecting treatment differences as large or larger than found by Asher and Harper (4) (at the one-tailed, P 0.05 level). The differences in the results of these two studies requires examination. Comparison with Asher-Harper study

Placebo

26

24

78(2)

67(l)

ward shifts 0.005), but were noted. Discussion We

in blood pressure (Table 6, P < no treatment group differences

anticipated

that

our

study

should

yield

results similar to Asher and Harpers findings (4); that is, that the HCG-treated group would lose a significantly greater amount of weight (actually about twice as much) than the placebo-treated group. We followed a comparable protocol. However, careful statistical analyses of our data did not reveal a significant difference for any of the parameters of weight loss or hunger between the two treatment groups. With the size of the groups tested, statisti-

Treatment groups. Our study contained 51 participants while Asher-Harpers contained 40 patients. In our study, the mean ages were 33.5 years in the HCG group and 33.0 years for the placebo group; theirs were 37.8 and 38.4 years of age, respectively. Mean heights were almost identical. Starting mean weights in our study were 165.9 lb for the HCG group and 167.7 lb for the placebo group; theirs were 171.7 and 165.4 lb. respectively. In our study, patients had a lower mean age due to imposed limitations (the announcement stated between 18 and 60 years of age) and patients had weight range restrictions (with exceptions noted in Table 1). These limitations were not imposed in the Asher-Harper study. Study period, visits with physician, injec-

946

STEIN

ET AL.

tions. Perhaps the most critical differences between the two studies were the time period chosen for the investigation (i.e., 32 days without intervening holidays) and a patientphysician consultation on each visit. In the Asher-Harper study (4), a period of August to February was reported and the patients were seen by a physician only at the time of the initial and final visits. The patients in our study received 28 injections on consecutive days, except Sundays; the Asher-Harper study required 36 injections, but none were required on holidays or at the time of heavy menstrual flow (4, 17). Therefore, this implies that in their study not every patient was on the diet for the same period of time while completing the 36 injections. In comparing the weight reduction results, it is necessary to remember that patients in our study received 28 injections over a 32-day period of time while their patients received 36 injections over a variable number of days (exact information not published) (4). Their report (4) makes assessment of compliance with diet and comparable weight loss difficult for any patient within one treatment group as well as the comparison of the two treatment groups. Another point of difference between these two studies is that all the patients in our study group had the added reinforcement to comply by peer pressure as they all reported daily to the clinic at the same time and developed a certain group spirit which the Asher-Harper patients did not have. The importance of this in overriding any small effect of HCG without this added force in a control group can not be answered. Comparison of data. The mean total weight loss in our study was 15.79 lb for the HCG group and 15.52 lb for the placebo group; theirs was 19.96 and 11.05 Ib, respectively. The percentage body weight loss in this study was 9.48 for the HCG group and 9.25 in the placebo group; theirs was 11.47 and 6.77, respectively. However, the average raw weight loss and percentage weight loss is virtually identical in the two study groups: raw weight loss, Asher and Harper 15.51 6.5 lb and Stein et al. 15.65 4.5 lb. percentage weight loss, Asher and Harper 9.12 4.5 lb and Stein et al. 9.35 3.2 lb. Weight loss per injection in this study was
= = = =

0.61 lb for the HCG group and 0.59 for the placebo group; theirs was 0.585 and 0.403 lb. respectively. Our placebo group achieved the same weight loss in pounds per injection as their HCG group. Clearly, Asher and Harpers placebo group stands alone in regard to weight loss compared to their HCG group and our two groups. In our study, none of the weight loss measures showed a statistically significant difference. The mean number of injections per patient in this study was 26.5 in the HCG group and 26.8 in the placebo group; theirs was 33.85 and 29.05, respectively. We have accounted for all missed injections (see Results first paragraph). Asher and Harper, in their letter of rebuttal (17), were only able to provide a reason for the missed injections in 11 of 18 patients who failed to receive the full 36 injections. In the group of 7 patients without a reason listed for missed injections, five of these patients were in the placebo group (4, 17). As Hirsch and Van Itallie (16) noted, One must ask why the placebo group received fewer injections on the average than those in the treatment group? If all patients who did not receive 36 injections were deleted from their study, their P value for mean percentage weight loss would fall from less than 0.001 to a value close to 0.02 (16, 17). With that change, the probability of this occurring by chance alone drops from less than one in 1,000 to close to 2 in 100. It would appear that a number of uncontrolled variables in their study may have resulted in a disadvantaged placebo group and an unintentional bias in their results. Asher and Harper (4) noted that, The results of our study were quite unexpected by the author responsible for study design because the results of our initial study were negative, as have been most double-blind studies reported in the literature. Additional cal Center findings study) (Letterman Army Medi-

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

Laboratory data. The laboratory studies were originally designed to help explain a mechanism for the effects of HCG on weight reduction, in addition to detecting possible adverse effects of the treatment program. Table 5 shows that both treatment groups

HCG

IN WEIGHT

REDUCTION

947

had a significant decrease in hematocrit, white blood cell count, blood urea nitrogen, cholesterol, triglyceride, and total protein. These findings are expected changes due to the diet and coincident weight reduction. Other semistarvation diets have been associated with a loss of vitamins and minerals, loss of body protein, and a lowering of serum lipids (25). Loss of protein, vitamins, or essential minerals may account for the mild drop in hematocrit and white blood cell count through any of a number of conceivable combinations of deficiencies (25-27). The findings of an increase in uric acid has been noted previously and is associated with the competition between ketoacids and uric acid for renal tubular transport (28, 29). The significant increases in fasting glucose and the difference in the increase between the treatment groups are unexplained. Both the prestudy and poststudy glucose determinations were performed by the same method. A gradual fall in blood sugar in the absence of exogenous carbohydrate has been reported (25); symptoms of hypoglycemia as a complication were noted by Simeons (I). Both of our groups experienced a decrease in blood pressure which would be expected with the weight loss and this type of diet (25). The changes in hip and waist measurements are also expected with this amount of weight loss. Simeons (1) has claimed that his technique leads to a more significant loss of fat in these two areas. The results in Table 4 shows no significant difference between the two groups in terms of percentage change in these two measurements. Complications. Throughout our study, patients were observed for possible complications. A large number of patients suffered from severe headaches of brief duration in the first week of treatment. The two most common complaints were constipation and fatigue. Delayed onset of menses was also seen. These problems were experienced equally by members in both groups. Two unexpected complications occurred in the HCG group. Patient 9, during the last week of treatment, developed a severe gastritis which was documented by gastroscopy and required a brief hospitalization. Patient 35, who had a problem with infertility before entering the study, did not have her usual

expected menses during the study period and was subsequently found to be pregnant. Patient 42 in the placebo group developed acute cholecystitis 2 days following termination of the study and diet; she subsequently had a cholecystectomy. In addition to these possible complications of therapy, there were other minor problems requiring treatment that arose during the study. These included colds, sinusitis, and gastroenteritis. Conclusions We attempted to answer one question in this study-is HCG an effective adjunct to a rigidly imposed dietary regimen for weight reduction? On the basis of our results, the answer is negative. The milieu under which this program was administered to a highly motivated group of women proved effective in achieving the desired result-rapid weight reduction. In our clinical setting, however, HCG per se offered no advantage over placebo injections in regard to weight loss, distribution of fat lost, or hunger index during weight reduction. Our data to date can not answer the question of the safety of HCG. Additional studies are underway to address this question. We also have not resolved completely why our data refute the findings reported by Asher and Harper (4), but we have proposed several areas which may account for the differences. Our study, performed in a manner similar to current Simeons-type programs, casts serious doubt of any inherent efficacy that HCG may have as an adjunct to a weight reduction program.
We wish to express our sincere appreciation for the

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

assistance in conducting staff of the Metabolic of Research. Specifically, Miller (Dietitian), CPT MAJ Susie M. Sherrod (Nurse).

this Ward,

study to Letterman

the professional Army Institute

we acknowledge LTC Clara L. Barbara B. McCarty (Dietitian), (Nurse), CPT Susan P. Nichols

References
SIMEONS, A. T. W. The action of chorionic gonadotrophin in the obese. Lancet 2: 946, 1954. 2. SIME0Ns, A. T. W. Pounds and Inches. Los Angeles: Medical Weight Control, 1974. 3. SIMEONS, A. T. W. Chorionic gonadotropin in

I.

geriatrics.

J. Am.

Geriat.

Soc.

4: 36, loss,

1956.

4,

ASHER,

W. L., AND
gonadotropin

H. W.

HARPER.

Effect

of human
and

chorionic

on weight

hunger,

948 feeling of well-being. Am. 1973. 5. GUSMAN, H. A. Chorionic J. Clin. Nutr. 26:

STEIN 211,

ET

AL.

Further
6. 7.

clinical

observations.

gonadotropin in obesity. Am. J. Clin. Nutr. 22:

onic gonadotropin treatment for obesity: a rebuttal. Am. J. Clin. Nutr. 27: 450, 1974 15. Federal Register. Vol. 39, No. 235. Chorionic go-

nadotropin:

Follow-up

notice

and opportunity

for

8.

9.

10.

686, 1969. LEBON, P. Action of chorionic gonadotrophin in the obese. Lancet 2:268, 1961. LEBON, P. Treatment of overweight patients with chorionic gonadotropin. J. Am. Geriat. Soc. 9: 998, 1961. LEB0N, P. Treatment of overweight patients with chorionic gonadotropin. Follow-up study. J. Am. Geriat. Soc. 14: 116, 1966. HUTTON, J. H., AND A. P. CRETICOS. Human chorionic gonadotropin in treatment of obesity. Ill. Med. J. 132: 693, 1967. VEILLEUX, H., AND Z. FORTIN. Effects gonadiques et extragonadiques, chez lhumains de 3,500 UI. de

19.

20.

21. 22.

23.

HCG

en doses fraction#{234}es.Vie

Med.

Can.

Fra.

I: 24.
25.

862, 1972. I 1. SOHAR, E. A forty-day-550 calorie diet in the treatment of obese outpatients. Am. J. Clin. Nutr. 7: 514, 1959. 12. HA5TRUP, B., B. NIELSEN AND A. P. SKOUBY. Chorionic gonadotropin and the treatment of obesity. Acta Med. Scand. 168: 25, 1960. 13. CARNE, S. The action of chorionic gonadotrophin in the obese. Lancet 2: 1282, 1961. 14. CRAIG, L. S., R. E. RAY, S. H. WAXLER AND H. MADIGAN. Chorionic gonadotropin in the treatment of obese women. Am. J. Clin. Nutr. 12: 230, 1963. 15. FRANK, B. W. The use of chorionic gonadotropin hormone in the treatment of obesity. A double-blind study. Am. J. Clin. Nutr. 14: 133, 1964. 16. HIRSCH, J., AND T. B. VAN ITALLIE. The treatment of obesity. Am. J. Clin. Nutr. 26: 1039, 1973. 17. ASHER, W. L., AND H. W. HARPER. Human chori-

hearing. Washington, D.C.: Government Printing Office, 1974. Federal Drug Administration Bulletin. Vol. 5, No. 2. HCG not approved for use in obesity. Washington, D.C., 1975. SNEDECOR, G. W., AND W. G. COCHRAN. Statistical Methods. Ames, Iowa: The Iowa State University Press, 1967. CONOVER, W. J. Practical Nonparametric Statistics. New York: John Wiley and Sons, 1971, p. 153. HOLLANDER, M., AND D. A. WOLFE. Nonparametnc Statistical Methods. John Wiley and Sons, New York, 1973. DIxoN, W. J. BMD, Biomedical Computer Programs. Berkeley, California: University of California Press, 1974. COHEN, J. Statistical Power Analysis for the Behav-

Downloaded from ajcn.nutrition.org by guest on June 13, 2013

ioral Sciences.
BRAY,

New York:

Academic

Press, 1969.

26.

27.

G. A., M. B. DAVIDSON AND E. J. DRENICK. Obesity: a serious symptom. Ann. Internal Med. 77: 779, 1972. JACOBS, H. S., AND T. AMSDEN. Acute hemolytic anemic with rigid red cells in hypophosphatemia. New Engl. J. Med. 285: 1446, 1971. DUNLAP, W. M., G. W. JAMES, III AND D. M.
HUME. Anemia and neutropenia caused by copper

deficiency.
28.
LECOCQ, of starvation,

Ann.

Internal

Med.

80: 470, 1974.

29.

F. R., AND MCPHAUL J. J. JR. The effects high fat diets and ketone infusions on uric acid balance. Clin. Res. 12: 274, 1964 (Abstr.). DRENICK, E. J. Hyperuricemia, acute gout, renal insufficiency and urate nephrolithiasis due to starvation. Arthritis Rheumat. 8: 988, 1965.