In the Matter of the Complaint Against GENERAL NUTRITION CORPORATION GENERAL NUTRITION CENTERS 418 Wood Street at Pittsburgh, PA 15222-1878 and 137 - 6th Street at Pittsburgh, PA 15222-3308 P.S. Docket No. 15/115; 10/31/83 Mason, Randolph D. APPEARANCES FOR COMPLAINANT: Thomas A. Ziebarth, Esq. Kenneth N. Hollies, Esq. Consumer Protection Division Law Department United States Postal Service Washington, DC 20260-1100 APPEARANCES FOR RESPONDENT: Robert Ullman, Esq. Lawrence Roth, Esq. Bass, Ullman & Lustigman 747 Third Avenue New York, NY 10017-2803 BEFORE: Randolph D. Mason Administrative Law Judge
This proceeding was initiated on December 14, 1982, when the General Counsel filed a Complaint against General Nutrition Corporation ("Respondent"). The Complaint alleged that Respondent is engaged in conducting a scheme or device to obtain money or property through the mails by means of false representations in violation of 39 U.S.C. § 3005. Specifically, the Complaint alleges in paragraph 3 that Respondent falsely represents that:
(a) One STARCH BLOCK tablet enables the user to eat 1 1/2 pounds of pasta without digesting or absorbing the starch calories contained therein;
(b) One STARCH BLOCK tablet totally blocks the absorption of 600 starch calories;
(c) STARCH BLOCK tablets are effective in blocking the absorption of all starch, regardless of its source;
(d) The safety and efficacy of STARCH BLOCK tablets has been established by competent scientific and clinical tests.
In its timely filed Answer, Respondent denied any violation of § 3005.
Respondent argues on brief that a decision in the instant proceeding should await the outcome of an action between the Food & Drug Administration (FDA) and Respondent and others in U. S. District Court in the Southern District of New York (American Health Products Co., Inc., et al v. Arthur Hull Hayes, et al., Case No. 92 C 4492). That action will determine whether Starch Block is a "food" or a "new drug" under statutes administered by the FDA. That issue is not present in the instant case and it is clear that a determination by the District Court may have no effect on this case. Moreover, Complainant urges that a decision be rendered herein in any event because waiting for the decision in the other case and the resulting appeals process would unduly delay the instant proceeding. These arguments are sound. Accordingly, Respondent's argument is rejected.
The initial hearing was held in Washington, DC, on March 18, 1983. Thereafter, the hearing was reconvened on April 13-14, 1983, in Madison, WI and on May 3-5, 1983, in New York, NY. Three witnesses testified for Complainant: Jay Stuart Morrow, M.D., Paul Bass, Ph.D., and Sami A. Hashim, M.D. Respondent presented the testimony of Richard F. Wallin, D.V.M., Ph.D.; Rene G. Gibson, Ph.D.; Ali Behforooz, Ph.D.; Thomas D. Darby, Ph.D.; David E. Walsh, Ph.D.; and Ronald W. Thompson, Ph.D. Both parties were represented by Counsel and afforded full opportunity to be heard, adduce relevant evidence and examine and cross-examine witnesses. Both parties filed proposed findings of fact and conclusions of law on July 19, 1983, which have been duly considered. To the extent indicated below, proposed findings and conclusions have been adopted; otherwise they have been rejected as irrelevant or contrary to the evidence. Based on the entire record herein, including my observation of the witnesses and their demeanor, the exhibits and other relevant evidence adduced at the hearing, I make the following findings of fact and conclusions of law:
1. General Nutrition Corporation and General Nutrition Centers ("Respondent") seek remittances of money through the mails by means of advertisements for Starch Block tablets. (Tr. 96, 901; Exh. C-16).
2. Respondent's advertising materials make the representations alleged in paragraph 3 of the Complaint. Each of the subparagraphs (a-d) of paragraph 3 are set forth below. Immediately following each subparagraph are quotations from Respondent's advertisements (Exh. C-1, 2, 3, 4 and 16) showing that these representations are, in fact, made.
(a) One STARCH BLOCK tablet enables the user to eat 1 1/2 pounds of pasta without digesting or absorbing the starch calories contained therein;
STARCH BLOCK
One tablet can block the conversion of the starch in 1 1/2 lbs. of pasta, for instance, into absorbable calories.
So you can cut out 600 calories--right there]
(Exh. C-1)
STARCH BLOCK
Now] Eat the pasta, potatoes, cereals, and breads you want, and... ellipsis in original Block out up to 1800 Starch Calories a day with--Starch Block. The revolutionary new concept in natural weight control is a pasta lover's dream come true] think of it--a single Starch Block tablet can block the conversion of the starch in 1 1/2 pounds of cooked spaghetti, for instance, into absorbable calories.
(Exh. C-3, 16)
(b) One STARCH BLOCK tablet totally blocks the absorption of 600 starch calories;
STARCH BLOCK
One tablet can block the conversion of the starch in 1 1/2 lbs. of pasta, for instance, into absorbable calories.
So you can cut out 600 calories--right there]
(Exh. C-1)
Just one Starch-Block tablet does one simple thing, and does it very well... ellipsis in original it can actually block up to 600 calories from being absorbed by your system. What this means to you is that you can really lose weight, while still enjoying all those delicious starchy foods]
(Exh. C-2, p. 2)
One STARCH BLOCK tablet can block the conversion of 150 grams of starch to absorbable calories--that's 600 calories] Simply take one tablet at the beginning of each meal or two tablets at the beginning of a large starch meal.
(Exh. C-4)
So you cut out 600 calories, right there--without dangerous chemically synthesized drugs, stimulants, or appetite suppressants. STARCH BLOCK is natural protein concentrate that actually blocks up to 600 calories per meal. The higher your diet is in starchy, delicious foods, the more calories STARCH BLOCK can cut out. Order Now]
(Exh. C-3)
(c) STARCH BLOCK tablets are effective in blocking the absorption of all starch, regardless of its source ; 1/
Can you eat pasta, bread, and potatoes--and lose pounds fast?
You can with Starch-Block.
. . .
Nearly 25% of the average American's calories come from starch (carbohydrates). Starch-Block was created in concentrated form to block calories from being absorbed in your system. This natural enzyme blocker prevents the breakdown of starch into simple sugars so that they are not absorbed as calories. With less calorie intake, you begin to lose weight...its just that simple.
(Exh. C-2, p. 2)
Starch is a major dietary complex carbohydrate. If enzymes did not break down the starch to simpler sugars, those calories would not be absorbed to add to total calorie intake.
. . .
BLOCK OUT STARCH CALORIES] EAT BREAD, POTATOES, PASTA] with amazing, incredible, new STARCH BLOCK.
(Exh. C-4)
(d) The safety and efficacy of STARCH BLOCK tablets have been established by competent scientific and clinical tests.
Starch-Block is not an appetite suppressant, dangerous fad diet, or chemically created drug. It's a natural enzyme that was lab tested. One group used Starch-Block and the other group did not. The group using Starch-Block lost more weight and with no undesirable side effects]
(Exh. C-2, p. 1)
Scientists have spent years studying and reporting on naturally occurring protein in legumes (beans) that blocks alpha-amylase, a major body enzyme that breaks down starch to absorbable forms. It was decided that if this protein could prevent or inhibit alpha-amylase from breaking down starch in test tubes, maybe it could do the same in people. The idea was tested in human diet study and it worked] The test group given the inhibitor lost more weight than the control group not given it. No undesirable side effects were reported. These exciting results were put to use in the development of STARCH BLOCK. Now you can put STARCH BLOCK to work for you]
(Exh. C-4)
3. The average American consumes about 350 grams of carbohydrates per day, and 80% of this amount is starch. About half of the calories consumed by Americans are derived from starch. (Tr. 466-468).
4. Starch is digested in two ways: (a) the amylase pathway and, (b) the bacterial pathway. (Tr. 346). Amylase is an enzyme produced by the salivary glands and the pancreas. This enzyme chemically reacts with food starch and breaks it down into its component disaccharide sugars, in order that the body can derive the nutrient value (calories) from the sugars. Amylase is present in the digestive system in several locations. Salivary amylase performs the amylase function in the mouth; salivary amylases are also found in the stomach since saliva is swallowed. In addition, the pancreas produce amylase which is present in the small intestine of man. (Exh. R-11, pp. 2-3).
5. Ten to forty percent of the dietary carbohydrate can be broken down by salivary amylase in the mouth, esophagus and stomach. (Tr. 566-568).
6. As previously stated, starch is also digested by means of the bacterial pathway. This bacterial action begins in the small intestine. (Tr. 347). In this regard, the human excretes 8 billion bacteria per day which produce fermentation and cause most products to be absorbed. (Tr. 348).
7. Respondent's advertisements claim that STARCH BLOCK tablets have the effect of preventing the digestion of starch by inactivating and inhibiting the action of alpha-amylase, the starch-splitting enzyme.
8. It is undisputed that Respondent's product, as well as many other so-called "starch blockers," has the effect of inhibiting the action of alpha-amylase in vitro (in the test tube). (Tr. 27-28). In this regard, it has been known for more than 25 years that certain plant foods, such as kidney beans and wheat contain a substance that inhibits the activity of salivary and pancreatic amylase. In 1975 the amylase inhibitor in the kidney bean, Phaseolus vulgaris, was purified and named phaseolamin. At 37 degrees centigrade this inhibitor forms a complex with amylase, optimally at pH 5.5 in test tube experiments. In vitro, alpha-amylase inhibitors prevent hydrolysis of the starch by noncompetitive binding to the enzyme. (Exhs. C-10, C-12).
9. Over 100 different starch blocker preparations have been placed on the market and widely sold by pharmacists and health food stores. Each starch blocker tablet is alleged to produce enough antiamylase activity to block the digestion and absorption of 100-150 grams of starch (400-600 calories). Thus, people who enjoy eating bread, spaghetti, and potatoes but who still want to lose weight have been urged in advertisements to take starch blocker with their meals. This weight control concept has been so attractive to the public that over 1 million starch blocker tablets were consumed daily in the United States in the first part of 1982. (Exh. C-10).
10. Respondent's tablets are manufactured from a processed bean material purchased from the Sunrise Chemical Company. The product is made by initially processing Great Northern White kidney beans. The beans are flattened by a series of rollers, injected with steam, dried, cooled, and dehulled. The beans are then placed in a mill which reduces the particle size. Then the product is put through an "air classifier" which separates the heavier starch fraction from the protein fraction. This product is purchased by Sunrise Chemical Company. The latter company subjects the powder to an additional heating, filtering, and drying process before shipping it to the Respondent who then manufactures the tablets. (Exh. R-15, pp. 279-281).
11. Before examining several studies that have been conducted to determine whether starch blockers are effective in man, certain facts should be noted which militate against a finding of efficacy:
(a) First, the pancreas produce an abundance of alpha-amylase and only 4% of that which is produced is necessary to cause the digestion of large amounts of starch. 2/ Thus, even if it were assumed that Starch Block was an effective inhibitor of amylase in the human body, the product would have no effect upon the digestion of starch unless it were capable of inhibiting more than 96% of the amylase produced. It is the opinion of Dr. Hashim that such a result is not possible because the pancreas are constantly producing amylase each time a person swallows and the Respondent's product would literally have to "sit there and wait" for all of the amylase to come down from the pancreas. (Tr. 354-356). Dr. Hashim's views are consistent with the consensus of informed medical opinion. (Tr. 398).
(b) It has been noted that the pH optimum for inhibition of amylase is 5.5 in test tube studies and that an incubation period of 20 minutes (for the inhibitor to inactivate the starch-splitting enzyme) is also required. In light of this, the Carlson study (Exh. C-12, p. 394) stated, and I find, that "intraluminal conditions are clearly not favorable for optimum inhibition." (See also Exh. C-10, p. 1415).
(c) Finally, three studies have speculated that "starch blockers" would be inactivated by gastric acid, pepsin or pancreatic proteases. (Exhs. C-10, C-12, and C-13).
12. When amylase has been completely removed from the body, as in the case of a person who has had his pancreas removed, then the ingestion of starch results in abdominal distension, cramping, abdominal pain, diarrhea, foul-smelling stools, and anal irritation. Varying degrees of amylase inhibition in a normal person would produce varying degrees of these symptoms. (Tr. 353).
13. Bo-Linn Studies. Several studies have been published which conclude that starch blocker tablets do not inhibit the digestion and absorption of starch calories in human beings. The first of these studies appeared in the December 2, 1982 issue of The New England Journal of Medicine. The study is entitled "Starch Blockers--Their Effect on Calorie Absorption from a High-Starch Meal" and was conducted by George W. Bo-Linn, M.D., John S. Fordtran, M.D. and two associates. The New England Journal of Medicine is a distinguished, widely disseminated publication and articles appearing therein are subjected to a rigorous peer review system (Tr. 359-360, 178). Dr. Fordtran is one of the outstanding investigators in the field of gastroenterology and has an outstanding laboratory and an international reputation. (Tr. 181, 189). Briefly stated, they used a one-day calorie-balance technique and a high-starch (100 gram) mean (spaghetti, tomato sauce, and bread), that measured the excretion of fecal calories after normal subjects had taken either placebo or starch blocker tablets. If the starch blocker tablets had prevented the digestion of starch, fecal calorie excretion should have increased by 400 calories. However, fecal calorie excretion was the same on the two test days. (Exh. C-10).
14. The one-day calorie-balance technique begins with a preparatory washout in which the entire gastrointestinal tract is cleansed of all food and fecal material by lavage. After four hours, the subject then eats the test meal. Test meals are prepared in duplicate with one consumed in its entirety by the subject and the other analyzed for calorie content in a bomb calorimeter. After eating the test meal, the subject did not eat or drink for fourteen hours. The entire gastrointestinal tract was then cleansed again by lavage. The rectal effluent from the latter process was then tested for calorie content in a bomb calorimeter to reveal the number of calories absorbed.
15. The above procedure was carried out with five subjects on two separate test days --on one with starch blocker tablets and on the other with placebo tablets. The order of the test days was randomized. On the day of the starch blocker test the subject ingested two commercially available starch blocker tablets immediately before beginning to eat the test meal. Each of the tablets contained 500 milligrams of the amylase inhibitor phaseolamin. When the subject had eaten half the meal, he or she took another starch-blocker tablet. The branch of starch blocker tablets chosen for this experiment have been shown by in vitro tests to produce the required antiamylase activity. Placebo tablets were ingested along with the test meal on the other test day. (Exh. C-10).
16. If the starch blocker tablets in the Bo-Linn study had been effective in vivo, they should have increased fecal calorie excretion by 388 calories after the subjects ate the high starch meal. However, the study revealed that fecal calorie output was not higher after ingestion of the starch blocker tablets than after the placebo. It was concluded that the starch blocker tablets did not inhibit the digestion and absorption of starch in vivo. (Exh. C-10).
17. The scientific methodology of the Bo-Linn test was sound. (Tr. 359). Complainant's highly experienced expert, Dr. Hashim, has had a great deal of experience with bomb calorimetery and thinks highly of it. (Tr. 361-362). This is so even though he recognizes that bomb calorimetery does not measure 100% of the calories absorbed since some are converted into gas which cannot be detected by that method. (Tr. 430). He has no problems with this study. (Tr. 363).
(a) The lavage technique is a routine procedure for cleansing the G. I. Tract. (Tr. 179-180). The use of lavage in combination with a bomb calorimeter, however, is new. (Tr. 424). Dr. Hashim recognizes that when food is present in the system the lavage method tends to cause the production of about 5% more amylase than normal (Tr. 427) because the food and electrolyte solution stimulates the vagus nerve. (Tr. 300). But this does not invalidate the study, because the lavage performed in the presence of food was done long after the meal was ingested, and any remaining digestive process would then be occurring in the colon rather than the upper part of the small intestine where amylase activity exists. (Tr. 300-301). This methodology was scientifically sound. (Tr. 359). 3/
18. Subsequently, Dr. Bo-Linn conducted the same study (with two changes discussed below) using Respondent's Advantage Starch Block tablets. (Exh. C-15). As in the previous study, Bo-Linn found that fecal calorie excretions were similar regardless of whether the subjects took Respondent's tablets or a placebo. He concluded that Advantage Starch Block tablets do not inhibit the digestion and absorption of starch calories in human beings.
(a) This study differed from Bo-Linn's first study in two ways: (1) the second study used only three subjects instead of the five previously used; and (2) the second study added ten grams of polyethylene glycol ("PEG") to the meal as a nonabsorbable marker. Dr. Fordtran, a co-author of both Bo-Linn studies, has previously indicated that the use of ten grams of PEG per liter could influence the result of an experiment. (Exh. R-12, p. 38). PEG inhibits the absorption of sodium, and the absorption of glucose in man is dependent on sodium. (Tr. 653). Although the second Bo-Linn study used ten grams of PEG in the meal, there is no evidence of record as to the resulting concentration per liter. (Tr. 661).
(b) There was no statistically significant difference between the fecal calorie excretions from the Starch Block group and the placebo group. (Tr. 363). Also, the fact that the standard deviation from the mean in the second study was much greater than the one in the first study is not meaningful because the later study only had three subjects. (Tr. 438).
19. The Carlson Study. S. C. Johnson & Son, Inc. (Johnson's Wax) of Racine, Wisconsin, was interested in the possibility of marketing a starch blocker preparation. (Tr. 217). This company, in collaboration with the Gastroenterology Laboratory of the Middleton Memorial VA Hospital and the Center for Health Sciences of the University of Wisconsin, sponsored an experiment to determine the efficacy of starch blockers inhibiting starch digestion. The study was conducted by Gerald L. Carlson, Paul Bass, Ph.D., and two other investigators. The study was reported (Exh. C-12) in the January 28, 1983, issue of Science, which is a very highly regarded, peer reviewed publication. (Tr. 221). Dr. Bass teaches at the University of Wisconsin School of Medicine and the School of Pharmacology where he specializes in the gastrointestinal tract. He is an expert with regard to the action of enzymes in the body. (Tr. 219-220).
(a) It is undisputed that effective inhibition of starch digestion in vivo would diminish glucose formation and absorption by the small intestine and increase the amount of undigested starch reaching the colon. The Carlson study tested the effect of starch blockers made from Great Northern White kidney beans on glucose formation and absorption by measuring changes in the concentration of glucose and insulin in the blood serum of human volunteers after they had consumed a starch meal. It simultaneously measured breath hydrogen levels --these would rise if greater than 6 to 10 grams of unabsorbed carbohydrates reached the colon. If the product was effective, the authors would have expected a reduced increase in glucose and insulin and elevated breath hydrogen production.
(b) Six subjects were tested on two occasions. The study was initiated in the early morning after a ten hour overnight fast. Baseline breath hydrogen samples and blood samples for determination of glucose and insulin concentrations were taken prior to a prescribed high starch test meal. The alpha-amylase inhibitor product or the placebo was crushed to a powder and homogenized with the meal prior to consumption. After the meal, blood samples were taken at 15, 30, 60, 90, 120, 150, 180, and 240 minutes. Breath hydrogen samples were taken every 15 minutes for six hours after meal consumption. The process was repeated seven days later except that those who received starch blocker in the first meal received placebo on the second day and vice versa. In this way each subject acted as his own control.
(c) Using standard statistical methods, the study reported that there were no significant differences between the starch and placebo subjects. Breath hydrogen results were similarly analyzed and revealed no inhibitory effect. The study concluded that formulations of the kidney bean-derived alpha-amylase inhibitor do not alter the digestion of cooked starch in humans. (Exh. C-12).
(d) The Carlson study represents a scientifically valid approach to determine the efficacy of starch blockers. (Tr. 368). The six subjects who were tested represented a sufficient number for the study. In this regard the number of subjects required depends upon the variability of the measurements one is making; this study had very objective measurements. (Tr. 229). It is also noted that the use of different time intervals for the taking of blood samples would have yielded results very similar to those obtained by the authors. (Tr. 249).
(e) The fact that the product was ground up and blended with the diet, rather than administered in tablet form, may have made it easier for the enzymes in the stomach to deactivate the amylase inhibitor. (Tr. 798-799). However, this is counterbalanced by the fact that when the ground tablet is well-mixed with food, then the food serves as a capsule to delay interaction with stomach acid. (Tr. 820).
20. The Garrow Study. The efficacy of starch blockers has been tested by another method by Dr. J. S. Garrow. The latter has a reputable nutrition laboratory under the auspices of the National Research Council of Britain, and is personally known and respected by Complainant's expert, Dr. Hashim. (Tr. 369). Dr. Garrow's study was published as a letter to the editor in the January "1/8," 1983, issue of The Lancet. (Exh. C-13). The Lancet is a clinical journal published in England which is highly respected by general practitioners; however, its letters to the editor are not subjected to peer review. (Tr. 775-776, 202).
(a) Dr. Garrow's study tested starch blockers made from red kidney beans which had been shown to inhibit amylase activity in the test tube. The digestion of starch can be measured by analyzing the amount of carbon dioxide in breath samples taken before and after a test meal. This is done by "labeling" the test meal of starch with Carbon 13. Normally the carbon dioxide in expired air is maximally labeled with Carbon 13 about three hours after the meal, and about 30% of the ingested Carbon 13 is oxidized to carbon dioxide within six hours of the meal. If starch blockers totally inhibit the digestion of starch, breath samples would not contain Carbon 13, i.e., no labeling of expired carbon dioxide with Carbon 13 should be observed.
(b) Dr. Garrow tested five obese women on three occasions after an overnight fast with a meal containing "corn flour in the form of a blancmange flavoured with lemon." 4/ Ten minutes before the meal, in random sequence, a starch blocker tablet or a placebo tablet (Gelusil) was ingested. For 30 minutes before the meal and six hours after the meal, the metabolic rate and rate of carbon dioxide production were measured. Breath samples were taken for analysis of the labeled carbon dioxide on three occasions before the meal and every half hour after the meal for six hours. It was found that the average percentage of carbohydrate meal oxidized was not significantly different when the placebo group was compared with the two starch blocker groups. Moreover, there was no significant difference in the rates of oxidation between the two groups. It was concluded that starch blocker tablets failed to delay either the digestion or further metabolism of the starch to any measurable extent. 5/
(c) Since Gelusil is a pharmaceutically active substance which affects digestion, it would have been preferable for Dr. Garrow to have used another kind of placebo. (Tr. 296). However, there is no evidence of record indicating what kind of effect, if any, Gelusil might have had on the findings of this experiment.
21. Another study was conducted by Dr. Bjontorp, a professor of medicine with an obesity clinic in Sweden. He gave starch blockers to a series of obese human subjects and compared them with a placebo group on the basis of weight alone. He found no difference between the starch blockers and the placebo on body weight. (Tr. 375).
22. The Walsh Study. David E. Walsh is Vice President of Research for Respondent General Nutrition Corporation. He received a Ph.D. in cereal chemistry and technology in 1969 at North Dakota State University. He has worked at GNC for eight years. (Exh. R-21; Tr. 828). He designed and supervised a study to determine the efficacy of Respondent's Advantage Starch Block tablets. (Tr. 829). GNC obtained the processed bean material for these tablets from the Sunrise Chemical Company. (Tr. 830). This study attempted to measure the amount of starch digestion with and without Starch Block tablets by comparing the blood glucose levels after test meals.
(a) The same seven subjects were used for both the experimental group and the control group. In each case blood glucose levels were measured after 12 hours of fasting. Each subject was then directed to ingest a measured quantity of instant mashed potatoes (one gram of starch per kilogram of ideal body weight). When Starch Block was being tested, the subjects ingested two 500 milligram tablets immediately prior to the test meal. In all cases, following the completion of the starch meal blood glucose levels were measured at 30, 60, 90, 120, and 180 minutes. Each treatment was repeated after five days on the same seven subjects. Thus the seven subjects were tested twice with Starch Block and twice without Starch Block. (Exh. R-17a; Tr. 833).
(b) Dr. Walsh reported that in all cases, with or without Starch Block, blood glucose levels rose to a peak at 30 minutes and slowly fell over the remaining 150 minutes. The peak glucose levels, in the presence and absence of Starch Block, were approximately equal in all subjects. The study also claims, however, that at 60 and 90 minutes the glucose levels of subjects treated with Starch Block were "significantly lower" than without the tablets. According to the calculation reported in Respondent's study, the Starch Block tablets caused a 38% reduction in blood glucose. Thus, the study concluded that the Starch Block meal resulted in 38% less starch digestion than the control meal. (Exh. R-17a).
(c) Dr. Hashim testified that the claim of a 38% reduction is "absolutely incredible." I find that Respondent's claim is unwarranted by the data furnished. (Tr. 395). The study specifically stated that no adverse effects or digestive problems such as gas, diarrhea, nausea, or pain were observed. If Dr. Walsh's subjects had actually experienced a 38% reduction in the amount of starch digested and absorbed in the blood stream, they would have experienced havoc in their intestines. The undigested starch would have been taken over by the bacteria and would have produced gas, distention, borborygmi (audible bowel noises), diarrhea, acidity, lactic acid, and other acidic material that would irritate the colon. (Tr. 392-393).
(d) As an afterthought, Dr. Walsh gave each of the subjects a test meal containing 38% less starch than previously given. The resulting blood glucose levels were similar to those recorded after the original higher starch meals. This casts additional doubt on the validity of the experiment. (Tr. 317). It is also remarkable that the Walsh report repeatedly refers to this final meal as containing only one-half as much starch as the previous 1-gram per kg. meal. Dr. Walsh changed this during his testimony from .5 gram to .6 gram (i.e., one gram less 38%) of starch per kilogram of body weight. This was such an obvious, nontypographical error that it is reasonable to infer that the report was drafted by someone other than those actually conducting the experiment. 6/
(e) The Walsh report is not in publishable form. (Tr. 390). Also, since it has not been peer-reviewed, its conclusions are at best "preliminary" or "tentative". (Tr. 405).
23. Gibson Dog Study. Respondent's Starch Block tablets are manufactured from processed white kidney bean material obtained from the Sunrise Chemical Company. (Tr. 830). Sunrise arranged for Dr. Rene G. Gibson, Ph.D., to conduct a study concerning the efficacy of amylase inhibitors derived from white kidney beans. (Tr. 495-496). Dr. Gibson is a gastrointestinal physiologist, i.e., one who studies the regulatory and controlled mechanisms of the G.I. tract, including the mouth, esophagus, stomach, entire intestinal tract, liver, pancreas, associated small glands, exocrine and endocrine glands. (Tr. 491). He has written almost 70 articles in the field of G. I. physiology and G. I. medicine (Tr. 493) and has been a teacher of those subjects. (Tr. 492). Sunrise Chemical paid for a portion of the cost of the study described below, and while he was writing his report on the study for the company, he was contacted and asked if he would like to be a consultant for the company. (Tr. 638). Following the completion of the study, Dr. Gibson became a consultant for Sunrise.
(a) Dr. Gibson studied three dogs with surgically implanted gastric and pancreatic Thomas cannulas (metal tubes). (Tr. 499). After recovering from their surgery, the dogs were fasted 12 to 16 hours before the study. Then a starch meal with or without the starch blocker material was introduced into the stomach through the gastric cannula. (Tr. 503). Thereafter, the blood sugars were measured every 15-30 minutes for 4-6 hours. Dr. Gibson reported that when the starch blocker material was added, blood sugar levels were depressed (Tr. 505) and fecal carbohydrates were increased. (Tr. 509). (Exh. R-10, ?III).
(b) Dr. Gibson did not perform any tests for statistical significance and did not try to draw a statistical inference. He does not consider his report (Exh. R-10, ?III) to be a scientific article written for his peers. He characterizes the study as a "bit of information submitted to the FDA for their use" and is admittedly not up to his usual standards. (Tr. 610-611). Clearly, the study is not of publishable quality. (Tr. 320). Moreover, the study is equivocal since only three dogs were used. (Tr. 319-320).
(c) The average human produces four to eight times the amount of amylase enzyme activity produced by the average dog. (Tr. 542-544).
24. Dr. Gibson also conducted a one-time test on himself and two friends in which blood samples were taken after eating a sandwich and a starch blocker tablet. Although he found no rise in their blood sugars, he admits that there was no attempt to perform any "real study or to draw any real analysis". (Tr. 632; Exh. R-10a, ?IV).
25. Although not introduced as an exhibit, it has been reported that one alpha-amylase inhibitor derived from wheat appears to be effective in vivo in diminishing the serum glucose increase that occurs in rats and dogs receiving a starch load, although results in humans were equivocal. W. Puls and U. Keup, Diabetologica 9, 97 (1973) (Exh. C-12, p. 395, f.n. 2; Tr. 284-285).
26. Another study was reported in a letter to the editor of Lancet. It stated that Tendamistate (Hoe 467) is an alpha-amylase inhibitor which inhibits the gastrointestinal absorption of starch. Hoe 467 is a bacterial product which has no relation to the bean-derived product in issue herein. The study reported that blood glucose levels decreased as the dosage of Hoe 467 was increased in conjunction with a starch meal. Several of the subjects reported flatulence. (Exh. R-20; The Lancet, April 23, 1983, p. 934).
27. Many experts agree that starch blocker products do not represent an immediate or serious health hazard. (Tr. 307, 64; Exh. R-3, R-2). However, adverse side effects have been reported primarily involving the gastrointestinal tract. (Exh. R-1, R-2, R3). There is some degree of concern and controversy regarding the presence of hemagglutinins (lectins) and trypsin inhibitors in starch blocker products. (Tr. 33). Some of these concerns are set forth below:
(a) Although there has been no direct diagnosis of hemagglutinins and toxicity, there are many cases where people have ingested raw beans and have gotten violently ill with severe abdominal cramps, nausea, vomitting, and diarrhea. These symptoms are similar to the more severe reactions reported with starch blockers. There is also concern that hemagglutinin substances may effect clumping of red blood cells and cause them to burst. (Tr. 174-175). The long-term use of these substances is totally unknown. (Tr. 87). On the other hand, one respected author in this field, Dr. Irvin Liener, takes the position that the hemagglutinins found in Northern White kidney beans are nontoxic. (Tr. 309-311).
(b) Members of the medical community are also concerned about the presence of trypsin inhibitors in starch blockers. These are proteins found in the kidney beans that inhibit the enzyme trypsin from the pancreas. The effect of ingesting trypsin inhibitors by humans over time is unknown; however, there is some evidence in experimental animals that the ingestion of trypsin inhibitors causes enlargement of the pancreas. (Tr. 73-74). On the other hand, some studies have shown no problems in this regard with certain larger animals, and Dr. Liener has predicted that this would be so with humans. (Tr. 75; Exh. R-5). Dr. Leiner's prediction is conjectural and has not been tested. (Tr. 77).
28. "LD-50" Study. This laboratory test was performed on Sunrise Chemical Company's AMX tablets, which are similar to, if not the same as, Respondent's product. (Tr. 108). This was an acute toxicity study to find the median lethal dose that is calculated to kill 50 percent of a group of rats. (Tr. 109). Following range-finding studies in which no animals died, five male and five female rats were orally intubated at a dose of 15 grams of starch blocker per kilogram of body weight. The animals were observed for signs of toxicity and death for up to 14 days. Based on the lack of mortality during the study, the oral LD-50 of the product was determined to be greater than 15 g/kg. No adverse effects were found. (Exh. R-7).
29. 90-Day Toxicity Study. The same laboratory also conducted a 13-week dietary study for Sunrise Chemicals. Thirty rats were fed increasing amounts of amylase inhibitor in their diet. Dietary levels of starch blocker were one percent for four weeks, 2 1/2 percent for five weeks and four percent for the final four weeks of the study. At each diet increase interval, blood samples were collected for routine hematology and blood chemistry and ten rats were killed. All major organs were sampled and the tissues were processed routinely and examined by a pathologist. No adverse effects were detected from the doses administered. (Exh. R-8; Tr. 119-120). Test and control animals had the same body weight gain, food consumption, general appearance, hematology values, blood chemistry values, and organ weights at autopsy. (Tr. 120). The 13-week study was not done with proper protocol by FDA standards. In this regard, there were no animals that were treated with a single dose for 90 days; on each occasion the dose was increased and some of the rats were sacrificed. (Tr. 42-43). Also the number of animals was too small. (Tr. 43).
30. In addition to his efficacy study (see Finding No. 23), Rene Gibson, Ph.D., conducted a four-week toxicity study on the same three dogs. (Exh. R-10, Part II). At the initiation of the study and for the first seven days, each dog was orally administered 2.1 grams of the starch blocker product daily. The tablets were mixed with dog food. For the second and third weeks of the toxicity trial the dose was doubled to 4.2 grams per day. During the final week, the dosage was increased to 8.4 grams per day. Dr. Gibson reported no observable significant alterations in the weight or bowel habits of the dogs during the trial. In addition, he reported that neither the hematology nor the serum chemistries of the dogs, which were taken at 10-day intervals, demonstrated statistically significant alterations from acceptable canine values. The dogs were terminated following the observation, and the necropsy revealed no gross morphologic changes in the thoracic and abdominal cavities. Other organs showed no pathological alterations after microscopic examination.
(a) The above 4-week study is not scientifically valid since only three dogs were tested; usually, the medical community requires 15-20 dogs for such a study. In addition, the time of exposure to the product was too short. (Tr. 173).
31. The LD-50 and 90-day studies mentioned previously are the type of animal studies designed to make a decision on whether one can begin to use the product in humans, not whether the product is safe in humans. The only way to establish safety in humans is to perform well-designed clinical trials with human subjects. (Tr. 24). Thus the LD-50 and 90-day studies are only a beginning, and many more studies are needed to establish safety in humans. (Tr. 378). One of Respondent's experts admits, and I find, that there are no long-term studies in existence that would tend to show that starch blockers are safe for sustained and continued human use. (Tr. 897). At the present time, no competent, scientific, or clinical tests have been performed that would establish the safety of starch blocker tablets. (Tr. 21, 38).
32. Based upon Findings 11-21, it is concluded that Respondent's Starch Block tablets are ineffective for preventing the digestion of starch. As alleged in the Complaint, Respondent falsely represents (a) that one tablet blocks the digestion or absorption of 1 1/2 pounds of pasta, (b) that a tablet totally blocks the absorption of 600 starch calories, and (c) that the tablets are effective in blocking the absorption of all starch regardless of its source.
33. Based on Findings 11-24, Respondent falsely represents that the efficacy of Starch Block tablets has been established by competent, scientific and clinical tests. On the contrary, competent, scientific and clinical tests have proved that such tablets are ineffective in man. (Findings 11-21).
34. Based upon Findings 27-31, Respondent falsely represents that the safety of Starch Block tablets has been established by competent, scientific and clinical tests.
35. These findings are in conformity with the informed scientific and medical consensus. (Tr. 35, 398).
36. The false representations are material because they tend to induce the purchase of Respondent's product.
1. The first issue presented for decision is whether the advertisements for Starch Block actually make the representations alleged in the Complaint. In determining whether representations are made, the advertisement is to be read in its entirety and in light of its effect on the ordinary mind. Donaldson v. Read Magazine, 333 U.S. 178 (1948). The issue of falsity of any such representations will be discussed separately. The burden of proof is upon Complainant. Mark Eden v. Lim P. Lee, 433 F.2d 1077 (9th Cir. 1970).
2. In its advertisements, Respondent clearly makes the representations alleged in subparagraphs (a) through (c) 7/ of paragraph 3 of the Complaint. See Finding of Fact No. 2. Respondent does not seriously argue to the contrary. Respondent does argue, however, that its advertisements do not make the following representation in subparagraph (d):
"(d) The safety and efficacy of STARCH BLOCK tablets has been established by competent scientific and clinical tests."
Respondent contends that the advertisements refer only to one experiment which purportedly found starch blockers effective and safe. 8/ However, a person of ordinary mind would get the impression that many scientific tests had been performed which prove the efficacy and safety of starch blockers. One advertisement stated:
Scientists have spent years studying and reporting on naturally occurring protein in legumes (beans) that blocks alpha-amylase, a major body enzyme that breaks down starch to absorbable forms. Exh. C-4
The ad then speaks of testing "in human diet study" sic --in light of the "years studying and reporting" language above, many ordinary readers would think "study" referred to research in general rather than any single experiment. More importantly, the ads (Exh. C-2, C-4) give the overall impression that the product has generally been proven effective and safe by scientific methods. The ordinary reader is, in effect, being asked to rest assured about the product because it is supported by scientific research. He reasonably infers that the "lab" testing fairly represents the body of medical knowledge on the subject, and he would not suspect that several scientific studies had found the product ineffective. Accordingly, I find that Respondent makes the representation in subparagraph (d).
3. The next issue for consideration is whether the representations made by Respondent are false. The representations set forth in subparagraphs (a)-(c) of paragraph 3 of the Complaint all deal with the question of the efficacy of Starch Block and will be considered together. Complainant produced convincing evidence that Respondent's product and other bean-derived starch blockers fail to prevent the digestion of starch in humans. Complainant's primary witness, Dr. Hashim, is extremely well-qualified and testified in a forthright and credible manner. Respondent's claims of efficacy were originally founded in test tube studies showing that starch blockers inhibited the action of alphaamylase, the enzyme responsible for breaking down starch and thereby causing its digestion. However, Dr. Hashim explained that the pancreas produce an abundance of this enzyme and only four percent of that which is produced is necessary to cause the digestion of large amounts of starch. (Finding 11a). Thus the product would have to inhibit more than 96 percent of the body's amylase before it would be even minimally effective; he considered such a result to be impossible. In addition, other factors have been listed in Finding of Fact No. 11 that militate against a finding of efficacy. Dr. Hashim further testified, and I have found, that the informed medical and scientific consensus is that starch blockers are ineffective in man.
4. In addition, Complainant relies upon four scientific studies which found starch blockers ineffective. Respondent first argues that three of these --two by Bo-Linn and one by Garrow --are inadmissible hearsay evidence since the authors failed to testify in this proceeding. Complainant introduced these studies through the testimony of Drs. Bass and Hashim, respectively, both of whom are well-known experts in this field. The Rules of Practice specifically permit the admission of authoritative medical writings through the testimony of expert witnesses. 39 CFR § 952.18(e). Moreover, although the undersigned is not strictly bound by the Federal Rules of Evidence (39 CFR § 952.13(a)), the writings in question would have also been admissible under those rules as an exception to the hearsay rule. Rule 803 (18) provides the following exception:
(18) Learned treatises. --To the extent called to the attention of an expert witness upon cross-examination or relied upon by him in direct examination, statements contained in published ... periodicals ... on a subject of ... medicine ... established as a reliable authority by the testimony or admission of the witness or by other expert testimony or by judicial notice. If admitted, the statements may be read into evidence but may not be received as exhibits.
The Second Circuit has recognized that where charts or graphs are contained in the exhibit, "good sense would seem to favor its admission into evidence" since they cannot be "read into evidence." United States v. Mangan, 575 F.2d 32 (2d Cir. 1978). Both of the studies in question contained detailed graphs. Moreover, the expert witnesses testified in detail with respect to the significance of these studies. Thus I reaffirm my ruling at the hearing that the documents were admissible. 9/
5. Respondent argues that the Bo-Linn and Carlson studies are not statistically meaningful. To this end Respondent produced the testimony of a computer science teacher who had only limited experience with statistical analysis. The only medical studies he had ever attempted to validate were those introduced in the instant case. On the other hand, the studies in question were subjected to statistical analysis by their authors; the latter have excellent reputations and experience, and statistical analysis of their medical and scientific studies is a routine practice. (Tr. 277). Therefore, I have accorded the testimony of Respondent's expert very little weight. Moreover, the latter's primary contention, even if accepted, would have no effect on the outcome of this case. He is essentially saying that the number of subjects tested was too small to statistically rule out the possibility that the starch blockers were minimally effective. 10/ However, Respondent advertises that the product is very effective and blocks the digestion of 150 out of 200 grams of starch ingested. (Tr. 888; Exh. C-1). Even Respondent's witness would agree that the studies in question adequately refute this claim from a statistical point of view.
6. Respondent has also raised a number of other questions about the studies relied upon by Complainant in an effort to challenge their scientific validity. These questions have been resolved, to the extent necessary, in the Findings of Fact and need not be addressed again at this point. Suffice it to say, I have concluded that the studies showing inefficacy are scientifically valid.
7. I must conclude that Complainant has made a convincing case for the inefficacy of Starch Block and has clearly established that the informed medical and scientific consensus is that starch blockers are ineffective in man. Under these circumstances, it is incumbent on Respondent to produce persuasive evidence proving the consensus to be in error. Standard Research Laboratories, P.S. Docket No. 7/48 (P.S.D. 1980).
8. The evidence presented by Respondent has been carefully analyzed and is set forth, for the most part, in Findings of Fact 22-26. This evidence fails to persuade that the consensus is in error.
9. Respondent argues that its "most compelling" evidence of efficacy is the Walsh study. (Finding 22). That study was conducted by Respondent and purportedly found that starch blocker tablets caused a 38% reduction in the amount of starch digested. The study further stated that no adverse effects or digestive problems such as gas, diarrhea, nausea, or pain were observed. However, if the product had really worked as reported, the undigested starch would have been taken over by bacteria and would have caused the aforementioned adverse effects and digestive problems. Other problems casting doubt on this study are mentioned in Finding of Fact 21(d). Finally, it is noted that, unlike three of Complainant's studies, the Walsh study was not peer-reviewed.
10. The Gibson dog study is also unpersuasive. This study was conducted for one of Respondent's suppliers. First, the findings were equivocal because only three dogs were tested. Dr. Gibson did not perform any tests for statistical significance and did not consider his report to be a scientific article. He admits that it was not up to his usual standards. Finally, it does not appear that the dog is a good subject for this particular study since humans produce far more amylase than dogs.
11. Finally, Respondent points to two studies which tested the amylase inhibiting characteristics of wheat and a bacterial agent, respectively. The findings of the first study were admittedly equivocal. In the second study, the bacterial substance was entirely different from the bean-derived product in issue. Neither study constitutes persuasive evidence of the efficacy of Respondent's Starch Block.
12. It is concluded that Respondent's product is ineffective for preventing the digestion of starch. Thus Respondent falsely represents that (a) one Starch Block tablet blocks the digestion or absorption of 1 1/2 pounds of pasta, (b) that a tablet totally blocks the absorption of 600 starch calories, and (c) that the tablets are effective in blocking the absorption of all starch regardless of its source.
13. The final issue is whether Respondent falsely represents that the safety and efficacy of Starch Block tablets have been established by competent, scientific and clinical tests. With respect to efficacy, it is clear from the previous discussion that the representation is false. The competent scientific tests that have been performed establish that Starch Block is ineffective in man.
14. With respect to safety, it should be emphasized that Complainant does not allege that Starch Block is unsafe. Complainant merely contends that there has been insufficient testing of the product to warrant the claim that the product's safety has been established by competent clinical tests.
15. Respondent first argues that the product is safe, and to this end introduced evidence which tends to diminish the concerns raised by Complainant's witnesses. However, the evidence of record fails to dispel those concerns. (Finding 27). Next, Respondent points to animal toxicity studies performed with starch blocker substances. (Findings 28-30). Two studies involved short-term tests with rats showing no adverse effects with large doses. 11/ These animal studies are the type designed to make a decision about whether to begin to use the product in humans, not whether the product is safe in humans. The only way to establish safety in humans is to perform well-designed clinical trials with human subjects. It is clear that no such studies have been performed on the question of safety. (Finding 31). Therefore I must conclude that Respondent falsely represents that the safety of Starch Block tablets has been established by competent, scientific and clinical tests.
16. Complainant has proved by a preponderance of the evidence that Respondent made the false representations alleged in the Complaint. Accordingly, I am constrained to conclude and hold that Respondent is engaged in conducting a scheme for obtaining money through the mail by means of false representations regarding the product Starch Block in violation of 39 U.S.C. § 3005. The attached order should be issued.
1/ Contrary to Respondent's assertion on brief (p. 19), Complainant does not take the position that the product effectively blocks the digestion and absorption of some types of starch but does not have that effect on other types. Complainant generally denies its efficacy for all kinds of starch.
2/ Complainant's highly qualified expert cited the following as an "elegant" study establishing this point: M. R. Fogel and G. M. Gray, J. Appl. Physiol. 35, 263 (1973). (Tr. 354-356; Exh. C-12, f.n. 14). This overabundance of amylase is also mentioned in the Carlson (Exh. C-12), Garrow (Exh. C-13), and Bo-Linn (Exh. C-10) studies.
3/ Respondent's expert in pharmacology, Dr. Darby, speculated that the lavage method could remove enzymes attached to the lining areas and bias the study in favor of starch blockers. On the other hand he thought that by initially washing out food, gastric juices would be more likely to inactivate the tablets and pH conditions might be altered. (Tr. 795-797). Such speculation fails to rebut the evidence in favor of Bo-Linn's use of lavage.
4/ The article does not indicate whether the blancmange contained any sugar and there is no reason to assume that the author would fail to mention this ingredient; however, even if it did, it would have had only a minimal effect on the outcome of the study since it was consumed by both the placebo group and the starch blocker group. (Tr. 294).
5/ The article indicates that the weight of starch in the test meals ranged from 44 to 64 grams. Although not specifically stated, it is reasonable to infer that Dr. Garrow followed the standard practice of varying the size of the meals in accordance with the body weight of each subject. (Tr. 411-412).
6/ The Walsh Report was initially introduced as Exh. R-17; however, when it became apparent that Dr. Walsh was testifying from another version of that report, the latter version was received as Exh. R-17a. Although the author of Exh. R-17a is unknown, it is noted that a portion of the first version (Exh. R-17) was drafted from handwritten notes by one of Respondent's attorneys. (Tr. 853).
7/ See footnote 1, supra.
8/ The "study" was introduced as Exh. R-23 only to identify what Respondent was alluding to in the ad and not for the substantive value of the study. (Tr. 885-887).
9/ It is also noted that Respondent had previously cross-examined Dr. Bo-Linn in the U. S. District Court proceeding with respect to his primary study. (Exh. C-10). In addition, Complainant notified Respondent well in advance of the hearing that the document would be introduced through another expert. See Fed. Rule of Evid. 804(b)(5) and 803(24).
10/ Another of Respondent's experts, Dr. Darby, also testified that the number of subjects in the first Bo-Linn study was too small; however, this was based on his erroneous perception that Dr. Bo-Linn used two separate groups of subjects for the treatment and control groups. He argued that the resulting variations in the populations made it necessary to increase the number of subjects. In fact, however, Dr. Bo-Linn used the same subjects for each group.
11/ A third test involving three dogs was clearly not scientifically valid. (FOF 30).