\ Plasma Glucose and Insulin Responses conebusion to Orally Administered Simple R-OK oe and Complex Carbohydrates Phyllis A. Crapo, R.D.,* Gerald Reaven, M.D.,+ and Jerrold Olefsky, M.D.,% Palo Alto We have studied the effects of gucose, sucrose, and various “starches on postprandial plasma glucose and insulin responses in 19 subjects. Allearbohydrateloads were calculated tocontain 0m. of slucase, and the response to each carbohydrate was tested twice: when givenaloneina drink or whengivenin combination with other nutrients asa mal. The data demonstrate: (1) Glucose and scrose ‘cited similar plasmaghicose responce curves butsucroveelcted a somewbat greater 20 per cent) plasma insulin response, (2) Raw starchingestion resulted ina per centlower glucose response and 135.65 per cent lower insulin response than did either glucose or ‘suerase ingestion. (3) When carbohydrate was given as a meal the plasma glucose responses were 40.60 percent lower then when the same carbohydrate was given as 2 drink, while the lsulin responses ‘were generally sillar, and (6) when different cooked starches were compared, the plasma glicose and insulin responses to rice were slgnlficantly lower ($0 per cent) than to potato. Inconchsion, thesize ofthe carbohydrate molecule appears to tafluence the postprandial slucose and lasulln responses such that more complex carbohydrates (tarchesellitlower responses. This effect maybe related teifer- ences In digestion rather than to differences in absorption. Duaweres28:741-47, September, 1976, ‘The carbohydrate content of diets has received con- siderable actention in recent years. Diets restricted in carbohydrate have been recommended for two particu lar groups of patients—rhose with diabetes! and those vith endogenous hypertriglyceridemia.® The rationale for carbohydrate restriction is usually based on the supposition that a decrease in the carbohydrate con- tent of che diet will decrease the postprandial glucose From the Department of Medicine and General Clinical Re- seaich Cencer, Stanford University Medical Cente, and Palo Alto Veterans Admiaistation Hospital, Plo Alto, California 94308, ‘Research Dietitian, General Clinical Research Center ‘Medical Investigator, Veterans Administration (MRIS 0 7363) ‘Clinical Investigator, Veterans Adminstration (MRIS 90, 188) ‘Ades repre requests co Phyllis Crapo, R.D., Gener! Clini= ‘al Reszarch Center. Room E340, Sranford Univesity Modicat Center, Stanford, California 94305. ‘Accepted for pobliation March 19,1976, and/or insulin responses. The usual way of actom- lishing total carbohydrace reduction is to decrease the amount of glucose and refined sugar (sucrose) without necessarily decreasing the starch (complex carbohy- date) content of the diet. This is based on the belief that glucose and sucrose are more readily available for immediate absorption, thereby producing a greater and faster rise in postprandial plasma glucose and in- sulin responses, in contrast t0 the supposedly slower and more gradual digestion and absorption of more complex carbobydrates.*°* However, other lines of investigation suggest that starch and glucose are assimilated at the same rate. Thus, after a test meal in normal man, Dehlqvist and Borgsttom® found that sufficient amounts of in- traluminal amylase were present to rapidly hydrolyze ingested search. This has been confirmed by Fogel and Grey,? who found that absorption, not intraluminal digestion, was the rate-limiting step in over-all starch assimilation. These findings cast doubt on the original may Db [3-5PLASMA RESFONSES TO CARBOHYDRATES assumption that the kind of dietary carbohydrate af- fects the subsequent postprandial serum glucose re sponse. In addition, search tolerance tests have beea used to diagnose amylase deficiency in patients with pancreatic insufficiency,*-#® and these tests show that in normals the increases in blood glucose following starch or glucose ate comparable. However, these studies! were done wich the carbohydrate as @ single agent mixed with water, and it is conceivable that some interplay between food constituents (pro- tein, fat, carbohydrate) occurs that could result in different rates of digestion and absorption for different catbohydraces. Furthermore, since che postprandial plasma insulin response is partially dependent on the ability of food constituents to stimulate the secretion of a number of gastrointestinal insulinogenic hormones," it is possible chat when mixed with other nattients (peotein and ft) different kinds of cazbohyd- rate differ in theie ability to stimulate these gastcoin~ testinal insulinogenic factors. ‘We have therefore studied the effects of different kinds of dietary carbohydrate on the postprandial plasma glucose and insulin responses when given alone and in combination with other nutrients in @ test meal. MATERIALS AND METHODS Nineteen normat volunteers, eleven women and eight men, were seudied. Mean age of the 19 subjects was 34 (range: 25-48) years. Mean weight was 147 pounds, and mean relative weight according to the Metropolican Life Tables was 0.92, with a range of 0.78 t0 1.15. No drugs known to affect glucose or insulin metabolism were being used by any of the subjects. During the course of the studies, each sub- ject consumed a weight-maintenance solid-food diet that included at least 200 gm. of carbohydrate each day. Three different sources of carbohydrate (glucose, sucrose, and starch) were studied. Each was stu: alone and in combination with other nutrients. Since starch is normally eaten in its cooked form, and our starch solution was uncooked, we also studied potato ‘nd rice. The composition of these test carbohydrate loads is outlined in table 1. Ie should be noted that we used 100 gm. of sucrose, 50 gm. of glucose, and 50 ‘gm. of starch. The test loads were calculated 50 that the {glucose load in all chree carbohydrates would be equal ‘Throughout the rest of che paper the carbohydrates when given alone will be referred to as “drinks” and when given with che other nutrients will be referred to as “meals.” All tests were conducted following an overnight fast, and their order was randomized. AC Ba.m., the subject was given one of the solutions or the potato of rice to drink or eat. They consumed the test load in 15 minutes. Blood samples were drawn for ‘measurement of plasma glucose and insulin at time 0 and at 30, 45, 60, 120, and 180 minutes following the beginning of the period of consumption. ‘There was at least a one-day interval between each study. ANALYTIC METHODS Samples for plasma glucose were collected in EDTA tubes and measured by a Beckman Glucose Analyzer by the glucose oxidase method of McComb and ‘Yushok.'? Plasma immunoreactive insulin was meas- tured by the method of Desbuquois and Aurbach." Statistical analysis was carried out by the use of the paired f-test for dependent means. TABLE Composition of tolerance tests Soluble Powdered Com Russet Long-grain Givcose Sucrose stach* egg albumin olla white rice Lemon Total pad ee a a) ee Gd 1A Glucose (drinks) ‘0 = — ~ = Smt. 500m. Bi, Sucrose (inks) St = = Smt Soom: C! Stareh (inks) = S = = = Smt Soom, D. Glucose (meals) 0 = 2 2 = = Smt om E, Sucrose (meals) 00 » 2 = = Smt omt F. Starch (meals) - » 0 = = Sm Soom, G. Pouto (baked) - = = = wR = soo mit Hi. Rice (oiled) = = = = = re “Amylopectn, Sigma $100 gm. of suerose was used 1 give equivalent $0 gn. lox of glvcose {85 in. of sav Rect potato mas aed and the skins were removed. Mean carbohydrate ingested was $1 gm 61.3 gm. of dry rice contains $0 gm, of exrbohydre, [Subjects dea $00 oe. water while eating the potato or rice m2 DIABETES, VOL. 25, NO.9PHYLLIS A. ERAPO, R.D, AND ASSOCIATES RESULTS ‘The plasma glucose and insulin responses co the slucose, sucrose, and starch drinks are summarized in figure I. It can be seen (figure 1A) that che plasma glucose curve following a 50-gm. glucose load is not statistically different ffom the plasma glucose curve following 4 100-gm. sucrose load (50 gm. glucose) The plasma glucose response curve co starch is flat, Tic because the glucose and sucrose curves decrease below fasting values at later time points, it exceeds the curves for glucose and sucrose at 60, 120, and 180 minutes} The plasma insulin curve (figure 1B) following the sucrose load is greater at all points than the plasma insulin curve after glucose, and these dif- ferences are statistically significant at 30 (p < 0.01) aod 120 (p< 0.01) mines, indicating thafnerone results in a greater insulin response than does gTucose on a molar bas insulin response curve following the starch is fatybut ie exceeds the glucose and su- jerose curves at 120 and 180 minutes. Figure 2 compares the plasma glucose and insulin 150 vo} (A) PLASMA GLUCOSE. 130 v0 0 mt 100 T (@) PLASMA INSULIN © atucose prINK 1B SUCROSE ORINK 6 STARCH DRINK pont 304560 70 TIME (MINUTES) 780 HIG. 1. Meon (= SE) plasma glucor, (A) and inulin (8) 16> ‘fone to potdh, nur, end str ds serrewsen, 1976 180 11 T 7 (a) PLASMA GLUCOSE r i" (8) PLASMA INSULIN Ni Sy sor @ GLUCOSE MEAL sok TS Suenos Meat z a Srancn meat eS 30 é a f. 4 20 7] 0 +3 7 30 45 60 120 180 TIME (MINUTES) HG. 2. Mean (2 S.E) plasma glucose (A) and insulin (8) x Sponves to gluco, rerose, and strch mechs. responses to the glucose, sucrose, and starch meals. There is no difference in the glucose response £0 a Jglucose or sucrose meal (figure 2A). The glucose re- sponse to a starch meal is flattened in the firse hour [compared with glucose and sucrose meals but, again, is higher at 60, 120, and 180 minutes. JThe insulin response (figure 2B) following the sirose meal is statistically greater than following the glucose meal at 120 (p < 0.005) and 180 (p < 0.005) minutes, and again it can be seen that, on a molar basis, sucrose results in a greater insulin response than does glucose. Both the glucose and sucrose meals have significantly greater insulin responses than do starch meals. Tn figures 3 and 4 one can see that although identi- cal amounts of carbohydrate were given in the drinks and meals,lthe plasma glucose response (figure 3) £0 US ARs tn ehan ees sins, leche insulin responses (Figure 4) were slightly higher with the meals. We have reported similar findings previously, and this indicates that the interplay of food constituents will attenuate glucose responses without affecting over-all insulin responses. The plasma glucose and insulin response curves £0 che starch drinks and starch meals are quite flat, and m3PLASIA GLUCOSE © GLUCOSE DRINK GLUCOSE meAL met 3 1 SUCROSE ORINK '& SUCROSE MEAL of {4 STARCH DRINK (A STARCH MEAL 100 <- “Ae ma 4 ob ros 3 a0 cy 780 1 ‘Time (wanures) FG. 3. Maan (+ S.£) plosme glucose response tucrave (8), ond strch €) given or meal we fele that chis could possibly be related to using an uncooked starch. Although amylases can attack raw starch, che gelatinization of starch during cooking may make the starch more accessible to the enzymes and the action much more rapid. Therefore, we de- cided to try some sources of starch thac were cooked snd were mote physiol in Homans Basa lacse tnd insulin responses were determined when starch was given as cooked, potato or rice. As demonserated in figure 5, the glucose and insulin responses to potato were significantly greater ehan the glucose and insulin responses to rice,§ although on comparing figures 1 Jand 5 (also table 2), one can see that the glucose and insulin responses t0 potato are still less than the re sponses to the glucose drink.JThese results were sur- prising and indicate a diffeFence in the ability of dif- ferent starches to elicit glucose and insulia responses. Since potatoes are active metabolically and convert starch to sugar during storage, it seemed passible that small amounts of fce glucose ingested with the potato starch could affect the postprandial glucose and insu- lin responses. However, this was not the case, since Sle ehis case, 13 patients had rice ests and 17 patients had poraco tests, and the data in Sigur 5 represent the mean (+ S.E.) OF all tests. However, forthe pated statistical analysis only the results fom the 13 patients who had boeh tests were use. 44 we measured fre glucose in our potato samples and found ic to be negligible. To facilicace cross-comparisons of the glucose and insulin responses to the various stimuli, the responses (o each of the oral carbohydrate loads ate summarized in table 2 DISCUSSION ‘We have studied the effects of orally administered. simple and complex carbohydrates on the postprandial lucose and insulin responses when given alone in a deink and when given in combination wich other nut- rients in a eest meal. The results show chat thefplasma ‘glucose and insulin responses following oral glucose, either drinks or meals, are higher than the responses following starch ingestion. These results seem to sup- port carlier studies showing that simple carbohydraces, cause a greater and faster rise in postprandial glucose and insulin responses than did ehe more complex car- bohydrates. Thus, Allen’ found that starches caused less glycosuria in diabetic dogs than did glucose, while Conn and Newburgh* demonstrated a higher blood glucose response to oral glucose than oan equiv alent load of starch given as potato or bread in diabet- ic subjects. Furthermore, Swana et al.® demonstrated that the blood glucose and insulin kesponses were PLASHNA INSULIN 9 GLUucose DRINK cLucose meaL puto 1b SUCROSE DRINK z of f Sucnose meat 2 0th ta SS Ec 7° ‘8 STARCH ORINK I 2 sranch meat Roe ets 20 es 30 a5 60 wo 780 “TME (miNTES) FIG. 4. Mean (= 8.) plasma insulin responses to glucose (AD, ‘Sucrose (8), and serch (@) given a meals of drink DIABETES, VoL. 25, NO.9she 961 wanasass read reso 9159 oF 0 #95 61 TLS ae S39 r6er@ seat rs ous orl Tesst res) cise 69255 Cray) myosuy easel cise | gtsoe un ces res) 98 Tes | LES TS 9S 5 96 stzor | SesuL 79.5 0 eisa | tess 09 = 501 Lesa S506 £5 = 90 cise | py sel | SOS TRF Tet Tree LSs08 69H | ENF IET “oar “oz se 0 (5 a Buy as0oe8 Buse oe The 3 rt res) (1=N) 2 t 66 6 | STH6 (=n) autem 9 sero rss 1) feu aon “¢ gorse = ¥ISe (61=N) Row 25005 “> lovesiicn = Iola (=n) eu WMS *¢ bosu tis9 (6t=N) 10u 280005 Z B96 LEO (61=N) nou 2802019 “1 tsa cLst6 (c1=W) omoa '¢ Ty Fool ozsas (E=N) AIL Se sO © 60a (=N) Fa pms "> Les | thse (61=N) pou astieg ¢ ssset | Thse (6130 Row 2500015 °F 8255 LT eG W=N) Tou WMS "E cL ser clea z VS SHE sPL= 05 t oe a Tara unos apgoqi> Ju ayy Hunp ssuodias ps! pow DOOD zanave soar ewisejd somo, uy sofnsas ayno2jous o1espAyoqre> a] & jo vonseBus rey awe2xpus sajnsas sno 2uIS swoiay oq Avws sowkeu> sid jospay aya 01 ypsers ays Jo Aaypqeyreaw porsdyd aya pur sup Surfadws ounse# ysnoypye ‘se9)2 200 1 sauas9jJIp asaya 29} usspUEyIOW 9Yy, “Pastape sf VOI -punsar aaespAyoqre> woys ur squsfaed 305 suresSoxd punadvrsqa-Ase21p [ouope: Zupeppwso} vp 29801 “todury [enuaied snoago seq sip pur “ysze2s oxei0d Ip wey asvodsas asoonj® pur uy[nsuy s99%0] yonut 1 pouoyja reas a27EPS aznsyy ur prauasosd vwep oy ‘woy aoddns s9qs9001 S9]na0j0ur y>teas [pe 30) uses at. 9q 200 our vondiosqe pus uonsoSip wus vapr ay, ‘apex peurasovuronsed apy ur ane $21 Gy 390333) “1p [e2;aBD v soxeur arespAyoqre> axp Jo Bury{o09 oy sey afqyssod st a1 pur ‘savas p3yo09 w ut sayespAyoqe> xojduio 20 ypins sp 2e9 Ayjeunow sueuMpy “r220%6| Suyp!0q WH poxzu JF saain> asoonpF yruou ang soqm po? Ypyss paxrul Jf SoAsn asoon|s wey uF paa]Ns a2 ,Noaeas a} qnyos,, 34 parsodas ancy goweakn) pue ‘upsneypiy aours ‘s3ynsaL ayp UF Jor2y v Sem syeaus pur ‘sup tpreas axp UE pasoduio aa yoreas a4 Jo ores payoodun aqp aexp ayqissod st af “uonuppe uy “uonsodur pms Supnojigy sasuodsos urynsuy pue 21w1024)8 29.0], 2p 395 a]qysuodsax aq ospe ews ‘ousoiur yews 243 aoqus saonpord 24)28984p say 30 yoseas yprym ae ares 23 S50 ypns ‘339y soto ows sours ‘sSuIpuyy son0¥] 25242 ‘yaya auaasysuo>uy Aqps8s5900u 10U ase saINs—3 IMG “HET ures 243 a8 as0any se poquosqe aye Ypres puv asomn]F ae pur davs Bujuzwnorop-o1s aya 200 1 voRsesIp dysewikeus awyp PaphPUS> davq] SION 9s9H]F, “TPL 3802 © BuLao]]9y pozdjospAy Ayprdes sx yoseas possoSur aeqp pur ssa0x9 ur auasead st asrure jeurumnpenuy ay porensuowsp axey ,fo15 pur foFoy pur ,wonstiog pu astabjyeq ‘sjduiexe sog “sassaz0xd aa ~daosqe pue aa13s98%p Jo soxpmas 29931 20932 JO 821A ‘ur a1xopezed wey ausos aq 02 ws99s sy]Ns9H axoge OY, “pool ymmas e Surmofioy ueyr peo] asoonjF e Furmoyjoy soySrK, sot 20 ojo 20 voai6 arson jo wl os 04 sosueds 8) wyjosur pun (y) e2020)8 Busoid (3'5 =) OOWY “s “O1L (sana av os_sr_ of o1vi0d © Nni9sM vasa (8) SS oe oot L ou aso0nT9 viNsyd (v) \Y a : . oct ozPLASMA RESPONSES TO CARBOHYDRATES and insulin responses, one might expect sucrose inges- tion to result in lower responses chan glucose inges- tion, since suceose is a disaccharide and requires en- zymatic digestion prior to absorption. How- ever, it has been shown that disaccharide di- gestive processes provide monosaccharides in sufficient quantities for transport to occur at maxi- mal races, and in both man" and rats'® the rate of absorption of glucose and fructose components did not differ whether sugar was fed as sucrose or a5 its monosaccharide mixture. Our results are compatible with these laeter findings, since we found that when given in equimolar amounts, glucose and sucrose re- sult in equal glycemic responses. However, sucrose did elicit « greater insulin response than glucose; the reason for this is not clear, since sucrose consists of glucose and fructose, and [fructose alone does not stimulate insulin secretion.'7JOn the other hand, Curry etal." have shown that fructose can potentiate insulin release in the presence of glucose; Fhis could possibly account for the increased insulii response €0 sucrose seen in figures 1 and 2, In addition, it is possible that some conversion of fructose to glucose may occur within the intestinal mucosa, resulting in a larger absolute load of glucose in the plasma with the sucrose than with the glucose.!®2° This incteased load could account for the increased insulin secretion. ?* ‘Our findings that starches generally give lower glu- cose responses than pure glucose while the disac- charide (sucrose) gives an equivalent response might be explained by examining the digestive processes Search digestion proceeds by a two-step process: the starch molecule is hydrolyzed (amylase) t0 disac- charides (maltose) and dextrins in the lumen of the ‘gut; these products are then converted to monosac- ‘charides by disaccharidases located on the brush border ‘of mucosal cells. Thus, our results suggest that the brush-border phase of digestion does not limie glucose absorption whereas the luminal phase might. Clearly, slowed luminal digestion could result from delayed gastric transit time, differences in fiber content, or anything that causes the starch molecule to be physi- «ally unavailable to the luminal enzymes and need not imply deficient enzyme activity per se. Lastly, when glucose, sucrose, and starch were -ivea as meals, lower glucose but similar insulin re- sponses resulted than after plain drinks. We have re- ported similar findings previously," and Eserich et al.22 have also shown that in adult diabeties the addi- tion of protein and fat to a carbohydrate load leads to flactening of the glucose response curve. It is possible 46 that these differences are due to stimulation of gas- trointestinal insulinogenic hormones by protein®® or that the amino acids derived from protein have poten- tiated beta-cell secretion of insulin. Additionally, ie has been well demonstrated that fat has an inhibitory effect on gastric emptying. Thus, since the glycemic response to a meal depends on the rate of entry and removal of glucose from plasma, che lower glycemic response to meals could be due to slower absorption, as a result of delayed gastric emptying, or enhanced removal due to insulin’s stimulatory effect on glucose uptake. In conclusion, we find that complex carbohydrates (starches) result Ta lower glucose and insulin responses than equivalent amounts of glucose as mono- oF disae- charides. The mechanisms for this effect are not clear, although delayed luminal digestion seems likely. Fix nally, the unexpected finding that rice starch resulted in lower plasma glucose and insulin responses than potato starch indicates that all starches are not treated identically by gastrointestinal digestive and absorp- tive processes. Clearly, further studies will be neces- sary to delineate the mechanism for this latter finding. ACKNOWLEDGMENTS The work was supported in part by National Heart and Lung Institute Grant HL08506 and General Clin- ical Research Centers Grant FR 70 ftom the National Institutes of Health, Bethesda, Maryland 20014. REFERENCES aja, $.S: Current unsolved problems in diabetes manage- ment. Diabetes 21(Suppl. 2:678-81, 1972. FPredrikson, D.S., Levy, R-I., Boone, M., and Erase, Nt Diecary management of hypedipoproteinemia, A Handbook fr Physicians and Dietitians. U.S, Depe. of Health, Education, and Welfce, Public Health Secvies, Washington, D.C., U.S. Gove Printing Office, 1973. “Allen, F.M.: Experimeneal studies on diabetes. J. Exp. Med. 3139-402, 1920. ‘Conn, J.W., and Newburgh, LH.: The glycemic response co iaoglacogenic quanciies of protein and carbohydrate. J. Clin Invest. 13:665-71, 1936. "Swann, D.C., Davidson, P., and Allrik, MJ: Eflece of| simple and complex carbohydrates on plasma nonesterified fic acs, plsma-sugar, and plasma-insulin dicing oral carbohydrate tolerance tests, Lancet 1:60-63, 1966 “Dahlgvse, A., and Borgsttom, B.: Digestion and absorption of disaccharides in man. Biochem. J. 81:411-18, 1961 Fogel, M.R., and Gray, GM: Starch hydrolysis in. man: an intealuminal process aoe requiring membrane digestion. J. Appl Phys. 35:263-67, 1973, *Alchauien, TL, nd Uyeyama, K. A new tes of panceatc funccion bared on sarc rolerance. Ann. Iovern. Med. 4156375, 1934 DIABETES, VOL 25, NO. 9y "Weijts, HLA., Van De Kames, J.ML, Dicke, W.K., and Tiseling, J Disethoca cased by deficiency of sugar splicing cemmymes. . Acta Paedistt, 50°55-71, 1961 Nugent, F.A., and Millhon, W.A. Clinical evaluation ofthe starch tolerance test. J.A-M.A. 168 2260-62, 1938, "Young, G.: Hormonal control of pancrestic endocrine and exocrine secretion. Gut 13:154-61, 1972 "McComb, R.B., and Yeshok, WD Colotimetie estima ‘ion of D-ghicose and 2-deoxy-d-glacote with glucose oxidase. J, Franklin Inst, 265:417-24, 1958. “*Desbuquois, B., and Aurbach, G.D.: Use of polyethylene _alycol to separate fice and antibody-bound peptide hormones 2 fadioimmunosssays. J. Clin, Endocrinol. Metab. 33:732.38, wri. “Reaven, G.M., Oley, J.M., and Farquhar, J.W.: Does laypeeglycemia or hyperinsulinemia characterise the patent with chemical diabetes? Lancet I:1247-49, 19972, "Gay, GM. and Ingelfinger, FJ; Intestinal absorption of secrose in man: Inteelation of hydkolysis and monosaccharide ‘product absorpcion, J. Clin. Invest, 45:388-98, 1966. MDohigvist, A., and Thomson, D.L.; The digestion and ab- sorption of sucrose by the inace rat. J. Physiol. 167-193-209, 1963, sevrewpen, 1976 MHuctunen, J.K.: Fructose in medicine, Postgrad. Med. J. 47:654-59, 1971. "Cuny, DL, Cuery, K.P., and Gomer, M.: Fruccse potens tiation of insulin seretion, Endocrinology 91:1493-98, 1972, "Holdsworth, C.D., and Dawson, A.M: Absorption of ce tose in man. Prec. Soc. Exp. Biol. Med, 118:142-45, 1965, "Cook, G.C.: Absorption products of D{-)froctese in ma, Clin. Sci, 37:675-87, 1969, *Pecerson, D.T., and Reaven, G.M.: Ryidence thar glucose fod is an important determinant of plasm insulin response in ‘normal subjects. Diabetes 20:929-35, 1971 MEsurich, D., Rank, A., Schlet, G., Fukayama, G., and Kinsell, L.i EMlects of co-ingestion of fit and_ protein upon ‘icbohydrate-induced hyperglycemia. Disbetes16:232-37, 1967, "Rabinowitz, D., Metimee, T,J., Maffeszoli, R., and Burgess, J.A.: Patterns of hormonal release afer glucose, protein, and ghacose plas protein. Lancet 2:45456, 1966 MFajans, S.S., Floyd, J.C. Je, Knopf, RE., and Conn, J.W. foc of amino acids and proteins on insulin secretion ia-man. Recent Prog. Horm. Res. 23:617-62, 1967. *Thomas, J.B! Mechanics and regulation of gasticempeying. Physiol. Rev. 37:453-74, 1957. mr