Gut, 1978, 19, 699-706

Postprandial duodenal function in man'

L. J. MILLER,2 J.-R. MALAGELADA, AND V. L. W. GO3

From the Gastroenterology Unit, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA

SummARY Duodenal function was studied in 11 healthy volunteers after intragastric instillation of amixed semi-elemental meal. The duodenum accepted chyme of varying pH, osmolality, and nutrientconcentration; and, as a result of biliary, pancreatic, and enteric secretion as well as absorption, itdelivered chyme with nearly constant pH, osmolality, and nutrient concentration to the jejunum. Theflow rate and nutrient load ofjejunal chyme varied. The duodenum absorbed more carbohydrate thanlipid and less protein, taking up each nutrient at a constant rate during most of the postprandialperiod. The percentage of nutrient load absorbed was greatest in the late postprandial period, whenflow rate, nutrient load, and concentrations were low.

Duodenal chyme influences all major functions of theduodenum; yet postprandial chyme in normal manhas not been fully characterised. Duodenal hormonaland neural regulation of gastric, pancreatic, andbiliary secretion and of upper gastrointestinal motoractivity is sensitive to chyme nutrient content(Windsor et al., 1969), osmolality (Meeroff et al.,1975), and pH (Johnston and Duthie, 1966).Pancreatic and biliary secretions that are importantto digestion mix with chyme in this segment of boweland are similarly sensitive to the characteristics ofchyme. An example of this is seen in the Zollinger-Ellison syndrome, where duodenal delivery of anacidic chyme inactivates lipase and precipitates bileacids, thus producing steatorrhoca (Go et al., 1970).Also, duodenal absorption and secretion of fluid andelectrolytcs and absorption of nutrients are certainlydependent on the composition of duodenal chyme.This is apparent in considering the dumping syn-drome (Abbott et al., 1960).

Little information has been obtained from normalman to characterise the postprandial gastric contentsdelivered into the duodenum, the modifications ofthis chyme that occur along the duodenum, and thechyme that is delivered into the jejunum. We havetried to develop more thorough knowledge of these

'Supported in part by Research Grant AM-6908 from theNational Institutes of Health, Public Health Service.'Dr Laurence Miller is an NIH trainee supported by GrantAM-7198 from the National Institutes of Health, PublicHealth Service.'Address for reprint requests: Dr V. L. W. Go, Gastro-enterology Unit, Mayo Clinic, Rochester, Minnesota 55901,USA.

Received for publication 6 January 1978

substances and changes and of the nutrient absorp-tion taking place at this level of the bowel afteringestion of a liquid, mixed, semi-elemental meal.Although this meal might not induce the same duo-denal events as a more complex one, it was used tosimplify analytical procedures.

Methods

SUBJECTSEleven healthy volunteers (two female and ninemale, aged 21 to 62 years) participated in 16 studiesafter giving informed consent. All data reported asresults are from the initial study performed in eachof the 11 subjects. The five duplicate studies areused only to provide further independent assessmentof the correlation between emptying of nutrient andof meal marker in a particular study.

MEALA 400-mI standard liquid meal containing about 300calories distributed as 40% carbohydrate, 40% lipid,and 20% protein (similar to their distribution in thenormal American diet) was used. The nutrients weresemi-elemental, in forms normally appearing in thebowel lumen, which could be assimilated easily andwhich permitted simple analysis of intestinal chymefor nutrient composition. The meal was prepared bydissolving in water 30 7 g maltose (0-224 molar), 14 goleic acid (0-124 molar), l 64 g of a complete tryptichydrolysate of casein, and 15 g of a nonabsorbablemarker (polyethylene glycol 4000) and adjusting thepH to 7 0 with a small amount of NaOH. Sonicationfor 10 minutes produced an emulsion with osmola-lity 544 ± 5 mOsm/l which was stable for severalhours, thus longer than the study period.

699

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

L. J. Miller, J.-R. Malagelada, and V. L. W. Go

TUBESTwo peroral tubes were used (Fig. 1). For the duo-denum, there was a sump tube that ran to a mercury-weighted tip beyond an occlusive balloon and hadthree small polyvinyl tubes cemented to it. Thisassembly (of total external diameter, except for theballoon, of approximately 6 mm) provided (1) theduodenal perfusion site; (2) an aspiration site with anair channel, to facilitate suction, located 20 cm distalto the perfusion site; (3) an inflatable balloonimmediately distal to the aspiration site; and (4) anaspiration site immediately beyond the balloon.Gastric sampling was done via a separate 14-F sumptube.

PROCEDUREEach study was begun after an overnight fast. Thevolunteers were seated in an upright position through-out the study. Under fluoroscopic control, the duo-denal tube was positioned with the balloon at theligament of Treitz and the gastric tube was positionedwith its tip in the most dependent area of the antrum.Duodenal perfusion with 14C-PEG (polyethyleneglycol, specific activity 0 5 zCi/mg) dissolved in 0-15M NaCl was maintained at 2 ml/min throughout thestudy period. The occlusive balloon was inflated with30 to 45 ml of air until the subject sensed its presence,without having any discomfort. Total occlusion wasconfirmed by demonstrating that neither bile nor14C-PEG was present distal to the balloon. More

Duodenal ( Q -' 77 Aiperfusion uode

(+ '4C-PEG,2 ml/min)

O1

tle

than 92% of marker was recovered proximal to theballoon in all studies. No study was included inwhich duodenal-gastric reflux of duodenal markerexceeded 15 %.

Fasting gastric and duodenal collections were madeby continuous suction (- 25 mm Hg) during two 10-minute intervals.Then the meal was injected via the gastric tube over

eight minutes, and gastric and duodenal sampleswere collected for two hours after the meal. Every 10minutes, 200 ml of gastric contents was aspirated, a10-ml aliquot was taken from it, and the remainderwas returned immediately to the stomach. Thealiquots from each 30-minute interval were pooled.Duodenal samples were aspirated by continuous

suction (-25 mm Hg), collected over ice, and pooledat 30-minute intervals. No duodenal chyme wasreinfused. To correct for transit time, duodenal col-lections were begun five minutes after correspondinggastric collections.At the end of the study period, gastric contents

were aspirated completely; then 200 ml of a normalsaline gastric wash was injected over five minutes;and this was aspirated, to recover as much of themarker as possible.

Determinations of osmolality (Wescor 5100 VaporPressure Osmometer) and pH (Fisher 520 DigitalpH/Ion Meter) were performed immediately on allgastric and duodenal samples. Marker concentrationsalso were measured in all samples (Brunner et al.,

al (+PEG)

Fig. 1 Peroralgastric sump tube andri C sampling multiluminalduodenalassembly, asplacedfor use.

al aspiration

cclusive balloon-Sampling sitedistal to balloon

700

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

Postprandial duodenal function in man

1974), and bilirubin and trypsin concentrations weremeasured in all duodenal samples (Brunner et al.,1974). Bilirubin and trypsin outputs as well as thegastric volume emptied and meal emptied werecalculated as previously reported (Brunner et al.,1974; Malagelada et al., 1976). The formulas weremodified to include the actual volume collected at theligament of Treitz rather than a flow rate previouslycalculated from the duodenal perfusate (Malageladaet al., 1976; Clain et al., 1977). Therefore, charac-terisation of gastric contents emptied into the duo-denum was indirect (based on marker determina-tions), and that of chyme leaving the duodenum wasmeasured directly.

Total protein was determined by the method ofLowry et al. (1951), fatty acid by the method ofCohen et al. (1969), carbohydrate by analysis ofmaltose (Bernfeld, 1955), and glucose by the hexo-kinase method (Bergmeyer et al., 1974). In the mal-tose assay, correction was made for free glucosepresent. Nutrient assays were performed on the meal,all duodenal samples, and the gastric contentsaspirated at the end of the two-hour study period.Although we cannot be sure that we were measuringonly exogenous nutrient, the contribution by endo-genous secretions probably was very small.

VALIDATION STUDYIn an attempt to determine the maximal potentialinterference, a validation study was performed inwhichpancreaticand biliary secretions aspirated fromfive normal subjects at the time of maximal chole-cystokinin stimulation were analysed for carbo-hydrate, lipid, and protein by the same techniquesmentioned above. As proportions of the mean post-prandial concentrations at the ligament of Treitz inthe main study, the highest concentrations in thevalidation study were carbohydrate 3-5 %, lipid9 5%, and protein 15-1 %. Adibi and Mercer (1973)also have shown that dietary protein makes up themajor portion of intraluminal amino acids and pep-tides after a meal.

STATISTICAL METHODSPaired sets of data from individuals were analysed bythe paired t test (Dixon and Massey, 1969).

Results

pH AND OSMOLALITYAfter ingestion of the meal (pH 7-0), the pH ofgastric contents-and therefore of chyme deliveredinto the proximal duodenum-decreased progres-sively (each point different from preceding point,p < 0-01). The pH at the distal end of the duo-denum (aspirated at the ligament of Treitz) was quite

701

stable, however, close to neutrality (Fig. 2). Osmo-lality of the chyme behaved similarly: after the meal(osmolality 544), gastric osmolality progressivelydecreased toward the osmolality of blood (each pointdifferent from preceding point, P < 0-01). This steadydecline of the osmolality of gastric contents enteringthe duodenum was not reflected by chyme at theligament of Treitz, where osmolality remained stablenear isotonicity (Fig. 2).

876

Z 5Q 4

32l

\

E

3;

C)

600

500

400

300

200

100

Meal

'WChyme at ligament of Treitz

Gastric contentsentering duodenum

I I I

Gastric contentsenering duodenum

.__ _ --__________Chyme at ligament of Treitz

I Mean± SE

90 l I

0 30 60 90 120

Time postprandial, min

Fig. 2 Simultaneously measuredpostprandialpHandosmolality ofgastric contents entering duodenum( ) andchyme leaving duodenum at ligament ofTreitz- -). Pointsplotted at zero time representpHand

osmolality ofmeal.

VOLUME FLOWFigure 3 demonstrates volume flows. The totalvolume of gastric contents emptied into the duo-denum during each 30-minute interval was constantthroughout the two-hour postprandial period (nopoint different from any other at P < 0 05 level). Theactual meal volume emptied into the duodenum wasgreatest in the first 30 minutes, then progressively lessin each interval thereafter as diluting gastric secretionbecame a greater proportion of the gastric volumeemptied (each point different from preceding point,p < 0-01). Throughout the study period, the volumeflow at the ligament of Treitz was greater than thatentering the duodenum (p < 0-01). The net change ofchyme volume along the duodenum is represented by

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

L. J. Miller, J.-R. Malagelada, and V. L. W. Go

2 - Meal ~ protein, lipid, and carbohydrate were emptied fromt the stomach in the same proportions as administered

30 T Mean -and in stable proportion to the meal marker.' ± SE Figure 4 demonstrates the correlations among the

40 meal marker, protein, lipid, and carbohydrate

10 //X/,S,, emptied over two houm. expressed as percentages ofVolume flow at the marker or nutrient administered. (Allvalues were

50 7volame>'/ ligament of Treitz calculated from actual measurements of residualNet' volumechange sZ gastric volume and of meal marker or nutrient

20 of doodeoal chyme / , concentrations.) These correlations validate our useof meal marker to calculate nutrient loads and con-

30 _.. Volume deliveed centrations entering the duodenum..t0.duodeoum All nutrient loads and concentrations delivered

10 - VMeal deliverued ..into the duodenum were maximal in the early post-O ___________________intduodenum , prandial period and decreased progressively (Fig. 5;

0 30 60 90 120 each point different from preceding point, P < 005).The duodenum handled the three nutrients dif-

Time postprandial, min ferently, however, absorbing more carbohydrate thanlipid and less protein, and therefore delivering less

Simultaneously measuredpostprandialvolume carbohydrate than lipid and more protein to the

ueliveredginto duodenum ( ) and volume leaving jejunum. Consequently, separate lines are drawn to

um at ligament ofTreitz (- - -)~with shaded area rersn th'ifrn uretlasadcneta!n these curves representing net volume change ofalchyme. Alsoplotted(. ) is portion of volume tions at the level of the ligament of Treitz. Theed into duodenum which represents meal volume nutrient loads arriving at the ligament of Treitz werethan gastric secretions. largest in the early postprandial period and diminish-

ed progressively (each point different from precedingrea between the curves representing volume point, except lipid and carbohydrate at 75 and 105red into the duodenum and volume at the min, P < 005). The nutrient concentrations, how-ent of Treitz. This net difference was greatest ever, had stabilised; and there was no change ofin the postprandial period and decreased pro- protein or lipid concentration in chyme between anyvely (each point different from preceding point, 30-minute intervals in the study period (no point)01). different from any other, at P < 005 level). The

carbohydrate concentration in chyme at the ligament,TENT LOADS AND CONCENTRATIONS of Treitz, although much more stable than at theneal administered is a stable emulsion; and the pylorus, gradually decreased (each point different

Percentage of meal nutrient or meal marker emptied over 2 h

20 40 60 80 100

Carbohydrate

lb..q)tQ3Ei(t

Lipid

(30

20 40 60 80 100

Lipid

.A.IwIwII Fig. 4 Correlations between mealCarbohydrate marker, protein, lipid, and carbohydrate

emptiedfrom stomach over two-hourpostprandial study period, expressedaspercentage ofmeal markerorcorresponding meal nutrient. Each pointrepresents a single study. Lines drawn arethe lines ofidentity. Correlationcoefficients rangefrom 0-92 to 0-99, allbeing significant (p < 0 001).

20 40 6 80 100

Lipid

702

32

28

24

2(

1U

12

Fig. 3being dduodenbetweeduodendeliverrather,

the a]deliveiligameearly igressi'P <C

NUTRThe n

10080604020

Protein

10080604020

0Q..

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

Postprandial duodenal function in man

Nutrient concentrationelivered into duodenum

-_, Protein;I -- Lipid Nutrient conc.

t- I _ 1 - -ij - - - -i at ligament'1-

Itof Treitz

CarbohydrateA

I I I I

0 30 60 90 1 20

Time postprandial, minFig. 5 Postprandial nutrient loads (left)and concentrations (right)in chyme entering andleaving duodenum. Solid linesrepresentpercentage ofeach meal nutrient volume or concentration being delivered into the duodenum. Only one line isdrawn, sinceprotein, lipid, and carbohydrate are emptiedfrom stomach in stableproportion to one another after this meal.Dashedlines representpercentage ofeach mealnutrient or concentration in meal reaching ligament ofTreitz.Differences in duodenalhandling ofnutrients necessitates drawing ofthree lines.

from preceding point, P < 0 05) because the duo-denum absorbed a much greater proportion of thisnutrient than of protein or lipid. Actual concentra-tions of each nutrient in chyme at the ligament ofTreitz after this meal were: protein 9-71 ± 0 53 mg/ml, lipid 5 79 ± 0-56 mg/ml (20 mM), carbo-hydrate 7T22 ± 0-82 mg/ml (21 mM). During thetwo-hour study period, mean absorptions were23 ± 6% of the protein load, 41 + 9% of the lipid,and 62 ± 9% of the carbohydrate (differencessignificant, P < 0 01).

NUTRIENT ABSORPTION

Duodenal nutrient handling is further demonstratedin Fig. 6. During the first 90 minutes postprandially,the amounts of each nutrient absorbed and, there-fore, the rates of absorption (protein 0-68 + 010,lipid 0-95 ± 0-09, carbohydrate 2-81 ± 0'24 g/30min) were constant despite changes in nutrient loads,nutrient concentrations, and flow rates (no pointdifferent from any other at P < 0 05 level). Thequantities of each nutrient absorbed decreased in thefinal 30 minutes as the nutrient load decreased(p < 0-05). The percentage of the nutrient loadabsorbed each 30 minutes, a measure of absorptiveefficiency, increased through the first three 30-

minute postprandial intervals (each point differentfrom preceding point, P < 0-05).

PANCREATIC AND BILIARY OUTPUTSTrypsin and bilirubin outputs are demonstrated inFig. 7. Pancreatic enzyme output was maximal in thefirst 30 minutes postprandially and declined steadilythereafter. Bilirubin output, reflecting gallbladdercontraction, also was maximal in the first 30 minutes;but thereafter it fell off faster than trypsin output.

Discussion

Isolation of the gastroduodenal field permitted studyof the coordinated upper gastrointestinal events thatoccur in the postprandial period in normal man,

making possible the determination of proximal anddistal duodenal nutrient loads, amount of nutrientabsorbed, and percentage of nutrient load absorbedfrom duodenal chyme in its normal postprandialform. Previous studies of duodenal absorption haveutilised either isolated nutrient infusions (Abbott etal.,1940; DiMagno et al., 1971) or comparison ofnutrient concentration with nonabsorbable markerconcentration, a technique permitting calculation of

100l

80 -

I.z

qV)

q)

'4..0)

0..

60 -

40 -

201-

703

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

L. J. Miller, J.-R. Malagelada, and V. L. W. Go

Amount absorbed Percent of load absorbed

100

30 60 90 1 20

Qd)

SC.)

C.)

1.

Q.c

CZ t.

c4 Q

80

60-

40

20h

I I

0 30 60 90 120

Time postprandial, minFig. 6 Postprandial duodenal absorption ofeach nutrient expressedaspercentage ofthatnutrient in meal administered (left). Efficiency ofabsorption ofeach nutrient is expressedaspercentage ofthat nutrient load absorbed (right).

1.0

0.8

0.6

0.4~

0.2

.cZE

a

-C.c

Time postprandial, min

Fig. 7 Postprandialtrypsin and bilirubin outputs.

percentage absorption but not nutrient load(Borgstrom et al., 1957). Although permitting thesefurther observations, this study may not be directlycomparable with studies in which chyme either wasnot diverted or was reintroduced, because of therecognised jejunal phase of gastric function (Clain eta!., 1977).

Gastric contents entering the duodenum after a

meal have progressively decreasing pH, osmolality,and nutrient concentration as the stomach dilutes themeal with acidic, near-isotonic gastric secretion and,at the same time, meal buffer is being emptied.Mechanisms to modify the chyme between its empty-ing from the stomach and its delivery to the jejunuminclude enteric, biliary, and pancreatic secretion aswell as duodenal absorption. Consequently, thechyme delivered to the ligament of Treitz after thismeal has constant pH near neutrality and constantosmolality near isotonicity. Fordtran and Locklear(1966) reported similar findings with a solid, com-plex, meal. Individual nutrient concentrations alsobecome fairly constant before reaching the ligamentof Treitz.

In contrast, nutrient and volume loads at the liga-ment of Treitz change during the postprandialperiod. This reflects nutrient rather than volumedelivery into the duodenum, since the rate of gastricvolume emptying is constant throughout the studyperiod.Throughout the study period, the volume flow at

the ligament of Treitz is greater than that enteringthe duodenum, with its greatest net increase occurringearly postprandially. Pancreatic and biliary secretionscontribute significantly to this increase of chymevolume. Pancreatic and biliary secretion is maximalearly, when maximal nutrient loads and concentra-tions are being delivered into the duodenum, anddecrease as the nutrient loads and concentrations

Cd)

-Q

-Q

0

*QC.,

q)

704

1'0

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

Postprandial duodenal function in man 705

decrease. It is of interest that both trypsin and bili-rubin outputs progressively decrease, while theamount of each nutrient absorbed remains constant.Although the early peaks of apparent output mayrepresent a washout phenomenon, the output curvesdo not stabilise as would be expected if they werecontrolled only by absorbed nutrient.

Another reason for the large early net increase induodenal volume is the limited amount of nutrientabsorption. The greatest percentage of the nutrientload is absorbed in the third 30-minute postprandialinterval, when flow rates, loads, and concentrationsare less. Also, little duodenal fluid absorption can beexpected early in the postprandial period, whenchyme from the stomach is hypertonic and flowsthrough the duodenum rapidly.

Despite the changes of nutrient loads, nutrientconcentrations, and flow rates, the amount of eachnutrient absorbed by the duodenum per 30 minutesis constant through two hours postprandially. Morecarbohydrate than lipid is absorbed, and less protein.

Borgstrom et al. (1957) also have investigatednutrient absorption from a mixed meal containingskim milk, dextrose, corn oil, and albumin. Nutrientabsorption was found to be complete in the proximal50 to 100 cm of jejunum, with lipid absorbed moreproximally than protein or carbohydrate. Amounts ofprotein and lipid absorbed proximal to the ligamentof Treitz in that study were similar to the amountsabsorbed in ours, but carbohydrate absorption wasquite different.The reason for this difference is uncertain. The

two studies used different forms of carbohydrate.Maltose, used in our study, is well absorbed-at leastin the jejunum: 60% of a 72-mM/h infusate isabsorbed by a 30-cm segment ofjejunum (Gray andSantiago, 1966). Cook (1973) found greater absorp-tion of carbohydrate from maltose than from glucoseperfused in the human jejunum. Dahlqvist andBorgstrom (1961), however, found little absorptionor hydrolysis of maltose in the duodenum. In thestudy of Borgstrom et al. (1957), 27% of the carbo-hydrate was in the form of lactose, a disacchariderequiring hydrolysis before absorption (Gray andSantiago, 1966). But maltose can be absorbed intact,even though its rate of hydrolysis is about twice thatof lactose (Gray and Santiago, 1966).

In our study, the rate of lipid absorption wasbetween the rates of carbohydrate and proteinabsorption. Conditions should have been ideal forabsorption, since bile was permitted to mix withduodenal chyme in a physiological manner to formmicelles. Pancreatic lipase was not necessary fordigestion, because the source of lipid used was a fattyacid.The nutrient absorbed most slowly was the pro-

tein, despite its presentation as small peptides, a formthat should maximise its absorption rate (Adibi,1971; Crampton et al., 1971; Adibi et al., 1975).Although actual analysis of amino acids and pep-tides was not performed, characterisation of these insimilar tryptic hydrolysates demonstrates an averagepeptide length of 2-2 amino acid units (Crampton etal., 1971). This is similar to the form of proteinnormally found in the intestinal lumen (Adibi andMercer, 1973). Normally, in fact, meal protein isfound as far distally as the mid-ileum (Adibi andMercer, 1973).

In summary, in the postprandial period after aliquid semi-elemental meal, the normal human duo-denum receives chyme of varying pH, osmolality, andnutrient concentration; and as a result of biliary,pancreatic, and enteric secretion as well as absorp-tion, it delivers chyme with near constant pH,osmolality, and nutrient concentration to the jeju-num. Jejunal chyme varies in flow rate and nutrientload. Duodenal conditions permit maximal absorp-tion of each nutrient over the first 90 minutes post-prandially, and more carbohydrate than lipid andless protein is absorbed. The percentage of nutrientload absorbed is greatest in the late postprandialperiod when flow rates, nutrient load, and concen-trations are low. How these results are modified bymeals of different size, composition, and physicalstate will need to be evaluated in the future.

The authors are grateful to Judith A. Duenes,Brenda A. Marben, Larry R. Stokes, and RichardTucker for their excellent technical assistance.

References

Abbott, W. O., Karr, W. G., Glenn, P. M., and Warren, R.(1940). Intubation studies of the human small intestine.XIV. The absorption of glucose from the duodenum.American Journal of the Medical Sciences, 200, 532-536.

Abbott, W. E., Krieger, H., Levey, S.,and Bradshaw, J. (1960).The etiology and management of the dumping syndromefollowing a gastroenterostomy or subtotal gastrectomy.Gastroenterology, 39, 12-26.

Adibi, S. A. (1971). Intestinal transport of dipeptides in man:relative importance of hydrolysis and intact absorption.Journal of Clinical Investigation, 50, 2266-2275.

Adibi, S. A., and Mercer, D. W. (1973). Protein digestion inhuman intestine as reflected in luminal, mucosal, andplasma amino acid concentrations after meals. Journal ofClinical Investigation, 52, 1586-1594.

Adibi, S. A., Morse, E. L., Masilamani, S. S., and Amin,P. M. (1975). Evidence for two different modes of tripeptidedisappearance in human intestine: uptake by peptide car-rier systems and hydrolysis by peptide hydrolases. Journalof Clinical Investigation, 56, 1355-1363.

Bergmeyer, H. U., Bernt, E., Schmidt, F., and Stork, H.(1974). Determination with hexokinase and glucose-6-phosphate dehydrogenase. In Methods of EnzymaticAnalysis, 2nd English edn., vol. 3, pp. 1196-1201. Edited byH. U. Bergmeyer. Academic Press: New York.

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from

706 L. J. Miller, J.-R. Malagelada, and V. L. W. Go

Bernfeld, P. (1955). Amylases, a and ,B In Methods inEnzymology, vol. 1, pp. 149-158. Edited by S. P. Colowickand N. 0. Kaplan. Academic Press, New York.

Borgstrom, J., Dahlqvist, A., Lundh, G., and Sjovall, J.(1957). Studies of intestinal digestion and absorption in thehuman. Journal of Clinical Investigation, 36, 1521-1536.

Brunner, H., Northfield, T. C., Hofmann, A. F., Go, V. L. W.,and Summerskill, W. H. J. (1974). Gastric emptying andsecretion of bile acids, cholesterol, and pancreatic enzymesduring digestion: duodenal perfusion studies in healthysubjects. Mayo Clinic Proceedings, 49, 851-860.

Clain, J. E., Malagelada, J.-R., Go, V. L. W., andSummerskill, W. H. J. (1977). Participation of the jejunumand ileum in postprandial gastric secretion in man.Gastroenterology, 73, 211-214.

Cohen, M., Morgan, R. G. H., and Hofmann, A. F. (1969).One-step quantitative extraction of medium-chain andlong-chain fatty acids from aqueous samples. Journal ofLipid Research, 10, 614-616.

Cook, G. C. (1973). Comparison of absorption rates ofglucose and mnaltose in man in vivo. Clinical Science, 44,425-428.

Crampton, R. F., Gangolli, S. D., Simson, P., and Matthews,D. M. (1971). Rates of absorption by rat intestine ofpancreatic hydrolysates ofproteins and their correspondingamino acid mixtures. Clinical Science, 41, 409-417.

Dahlqvist, A., and Borgstr6m, B. (1961). Digestion andabsorption of disaccharides in man. Biochemical Journal,81, 411-418.

DiMagno, E. P., Hernmon-Taylor, J., Go, V. L. W., Lillehei,R. C., and Summerskill, W. H. J. (1971). Functions of apancreaticoduodenal allograft in man. Gastroenterology,

61, 363-368.Dixon, W. J., and Massey, F. J., Jr. (1969). Introduction to

Statistical Analysis, 3rd edn., pp. 638. McGraw-Hill BookCompany: New York.

Fordtran, J. S., and Locklear, T. W. (1966). Ionic con-stituents and osmolality of gastric and small-intestinalfluids after eating. American Journal of Digestive Diseases,n.s. 11, 503-521.

Go, V. L. W., Poley, J. R., Hofmann, A. F., andSummerskill, W. H. J. (1970). Disturbances in fat digestioninduced by acidic jejunal pH due to gastric hypersecretionin man. Gastroenterology, 58, 638-646.

Gray, G. M., and Santiago, N. A. (1966). Disaccharideabsorption in normal and diseased human intestine.Gastroenterology, 51, 489-498.

Johnston, D., and Duthie, H. L. (1966). Inhibition ofhistamine-stimulated gastric secretion by acid in the duo-denum in man. Gut, 7, 58-68.

Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall,R. J. (1951). Protein measurement with the Folin phenolreagent. Journal ofBiological Chentistry, 193, 265-275.

Malagelada, J.-R., Longstreth, G. F., Summerskill, W. H. J.,and Go, V. L. W. (1976). Measurement of gastric functionsduring digestion of ordinary solid meals in man. Gastro-enterology, 70, 203-210.

Meeroff, J. C., Go, V. L. W., and Phillips, S. F. (1975).Control of gastric emptying by osmolality of duodenalcontents in man. Gastroenterology, 68, 1144-1151.

Windsor, C. W. O., Cockel, R., and Lee, M. J. R. (1969).Inhibition of gastric secretion in man by intestinal fatinfusion. Gut, 10, 135-142.

on Novem

ber 18, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.19.8.699 on 1 August 1978. D

ownloaded from