GASTROINTESTINALWATERANDELECTROLYTES. IV. THEEQUILIBRATION OF DEUTERIUMOXIDE (D20) IN GASTRO-

INTESTINAL CONTENTSANDTHE PROPORTIONOFTOTAL BODYWATER(T.B.W.) IN THE GASTRO-

INTESTINAL TRACT1

By F. GOTCH,2 J. NADELL,' AND I. S. EDELMAN4

(From the Department of Medicine, University of California School of Medicine, and the SanFrancisco Hospital, San Francisco, Calif.)

(Submitted for publication July 27, 1956; accepted October 15, 1956)

The concept of the anatomy of body water dis-tribution as a two-compartment system consistingof intracellular and extracellular fluid has beenshown to be inadequate (1-5). The heterogene-ous nature of the extracellular fluid compartmenthas been established by previous studies on bone(2), dense connective tissue (6), and transcellu-lar fluid (3-5). Cizek (3) has demonstrated thatintraluminal gut water is a significant subdivisionof body water in a number of species.

Neglecting the contribution of transcellular fluidto body water results in considerable errors in thederived normal values for body water compart-ments. Furthermore, transcellular fluid, if largeenough in volume, must be considered as poten-tially important in determining the volume andosmolarity of plasma, interstitial and intracellularfluid by ion and water flux in or out of transcellu-lar pools in response to metabolic stimuli.

In the preceding three papers of this series wereported the measured amount of sodium, potas-sium and chloride contained within the lumen ofthe gut in rabbits and in human subjects at postmortem (4, 5, 7). The present study is similarin design to these previous experiments and pre-sents observations on a) the amount of intralumi-nal gastrointestinal water expressed as a fractionof total body water (T.B.W.) and the extent of

iThis work was carried out under grants from theAmerican Heart Association, the United States PublicHealth Service (Grant No. H-1441), the FleischmannFoundation, the San Francisco Heart Association, thePaul and Susan Gardiner Fund, and the Raschen-Tiede-mann Fund.

2Research Fellow of the American Heart Association.8Research Fellow of the National Heart Institute of

the United States Public Health Service.4 Established Investigator of the American Heart As-

sociation.

deuterium oxide (DO) exchange equilibrium ingut contents in rabbits, and b) the amount of in-traluminal water in man at post-mortem examina-tion.

METHODS

A. RabbitsForty adult albino rabbits were studied in pairs con-

sisting of a male and a non-gravid female. The animalswere fasted and thirsted. All urine passed during theperiod of isotope equilibration was collected in a metabo-lism cage. This period varied from 2 to 5 hours. Eachanimal was injected intraperitoneally with 2 ml. of D,0 6from a calibrated syringe. At the end of the equilibrationperiod each animal was anesthetized with 2 ml. of 2 percent sodium pentobarbital injected into a dorsal ear veinand was then weighed to the nearest gram. A blood sam-ple was obtained at this time by cardiac puncture throughthe intact chest wall with a syringe containing dryheparin. The syringe was capped and centrifuged im-mediately after collection, and the separated plasma wasaspirated, sealed in a glass ampoule and stored in afreezer.

The gastrointestinal tract was removed in three seg-ments by cutting between double ligatures placed at thecardia of the stomach, the pylorus, the ileocecal valveand at a position in the transverse colon where therewas a transition point between semi-solid and solid stoolpellets. After removal each segment was washed withdistilled water, dried with towels and weighed to thenearest gram.

The contents of each segment were milked into oneligated end, a small incision was made, and an aliquot ofcontents was expressed into a dried test tube, whichwas quickly stoppered and centrifuged. The supernatantwas then aspirated and sealed in a glass ampoule andstored in a freezer. Each segment was then openedlongitudinally, and the remaining contents were evacu-ated into a clean container by gently stripping and thenwashing the mucosal surface with distilled water. The

5Deuterium oxide, 99.6 per cent pure, was obtainedfrom Abbott Laboratories as a sterile isotonic salinesolution.

289

F. GOTCH, J. NADELL, AND I. S. EDELMAN

TABLE I

Effect of equiJibration time on measured total body uwatr

Number of rabbits Equilibration Body weight Total body watertime mean 4 s.d.* as %body wt.

Male Female Total (hours) (Kg.) mean I s.d.* t Value p Valuet

5 5 10 2 2.46 4 .27 71.0 4 7.04 5 9 3 2.05 i: .25 75.1 5.5 1.44 >0.15 5 10 4 2.16 4 .23 74.4 4 3.5 1.29 >0.25 5 10 5 2.20 4 .20 72.0 4 4.4 0.39 >0.5

*s.d. = -.t Compared with 2-hour value for T.B.W. in each instance.

evacuated contents were quantitatively transferred withmultiple distilled water rinses into a graduated cylinderto which the remainder of the previously centrifuged ali-quot was added. The diluted contents were thoroughlymixed, the volume recorded and an aliquot taken for de-termination of solids. The wall of each segment wasdried and reweighed.

Ten-ml. aliquots of the diluted contents were pipettedin duplicate into tared weighing bottles. The solidcontent of each sample was estimated gravimetricallyafter drying at 105° C for 72 hours. All duplicateschecked within 10 per cent

The concentration of D.O in blood was determined induplicate by the falling-drop method (8). The maximalacceptable difference between duplicates was 0.006 vol-ume per cent. Only single analyses were carried out onurine samples since less than 1 per cent of the tracer wasexcreted during the equilibration periods. The concen-tration of DOin intraluminal water was determined bymass spectrometer analysis (9).6 Each of these analyseswas done in duplicate, and all duplicates checked within0.006 atom per cent D.

B. Human subjectsThe gastrointestinal tracts were removed and the con-

tents analyzed in 13 human subjects at autopsy. The tech-nique of sample collection was described previously (4).Water content was estimated as described for the rab-bits. Subjects selected for study had no gastrointestinaldisease and minimal clinical abnormality of fluid andelectrolyte metabolism. The body weight and height ofeach subject were measured. The appendix to the firstreport in this series (4) describes the pertinent clinicaland pathological findings in these cases.

CALCULATIONS

RabbitsTotal body water was calculated from the well-known

isotope dilution formula (1). The amount of tracer re-

6 These analyses were performed by Mr. Brad Pearsonunder the direction of Dr. A. K Solomon at the Bio-physics Laboratory of the Harvard Medical School inBoston, for which we are most grateful.

tamed was assumed to be the amount injected minus theamotmt excreted in the urine during the equilibration pe-riod. Isotope excretion varied between 0.1 per cent to0.6 per cent of the injected D20. The maximum com-bined errors in this calculation are less than 5 per cent.

The volume of intraluminal water in each gut segmentwas calculated in the following manner:

a) Weight of intraluminal contents =Weight intact gut segment-Weight of segmentwall

b) Total solids in intraluminal contents =(Volume diluted contents) (Weight solids ingrams per ml.)

c) Volume of intraluminal water =Weight of intraluminal contents - Total solids inintraluminal contents

Accumulated errors in the final calculation are 10 percent or less.

Completeness of DOexchange between plasma waterand intraluminal water was determined from the ratio ofDOconcentration of intraluminal water to plasma water.This ratio is designated the specific activity ratio(S.A.R.) (2, 4).

When exchange is complete, S.A.R. is unity since theconcentration of DOwill be the same in both intralumi-nal water and plasma water. The value of S.A.R. is as-sumed to represent the fraction of gut water exchangedwith extracellular water (2, 4).

Human subjectsTotal body water was estimated in these subjects from

data derived from D20 dilution studies on normal sub-jects in this age group (10). The predicted T.B.W. wascalculated, using the values 22.2 liters per square meterin males and 17.1 liters per square meter in females.

Volume of intraluminal water was calculated by themethods outlined above for the rabbit.

RESULTS

Rabbits

Uniform dispersal of a tracer theoretically oc-curs after an infinite period of time (11). A prac-

290

GASTROINTESTINAL WATER

tical means for judging equilibrium of distribution,however, is to determine the time required forconstant volumes of dilution, in this case T.B.W.The effect of the time allowed for equilibration onthe apparent T.B.W. is summarized in Table I.Mean T.B.W. values at 3, 4 and 5 hours of equili-bration are compared to the 2-hour values. Nosignificant differences were found among thesegroups, nor was there a trend in either direction.Two hours is probably an adequate period forD20 equilibration in this species.

Data on T.B.W. and "total" intraluminal gas-trointestinal water 7 in animals matched care-fully for weight, age and the period of isotopeequilibration and compared on the basis of sex aretabulated in Table II. The differences in T.B.W.and the gut water content of male and femalerabbits were not statistically significant.

The absolute magnitude of intraluminal watercontent and its isotope exchange characteristics ineach segment of the tract are given in Tables IIIand IV.

The mean volume of intraluminal water, ex-pressed as per cent of T.B.W., in the stomach,

7 "Total" intraluminal gastrointestinal water refers towater contained in the gut from the cardia of thestomach to the mid-transverse colon.

small intestine and large intestine is 4.1, 2.0 and6.0 per cent, respectively (cf. Table III). Thesum of these, or "total" intraluminal water, is 12.1per cent of T.B.W. The coefficients of variationfor these data are large, varying between 22 and50 per cent, and reflect the considerable variationin the volume of intraluminal gut water found indifferent animals in spite of rigidly controlledexperimental conditions.

The specific activity ratios of gut water toplasma water at 2, 3, 4 and 5 hours' equilibrationfor stomach, small intestine and large intestineare listed in Table IV. The p values in each in-stance refer to comparison of the measuredS.A.R. with unity, the theoretical value for com-plete equilibration. Equilibration is complete inthe large bowel segment at 2 hours. In the smallintestine, equilibration is nearly complete at 2hours, with an S.A.R. of 0.95 + 0.05, and is com-plete at 3 hours, with an S.A.R. of 0.99 0.03.There is a significantly slower rate of exchange ofwater in the stomach. The S.A.R. of + 0.86 at 2hours differs significantly from 1.00 (p < 0.001).At 3 hours only 93 per cent of water in the stomachhas exchanged with plasma water, but at 4 hourscomplete exchange is demonstrated (S.A.R.0.98 + 0.05).

TABLE II

The total body water and "total" intraluminal gastrointestinal water* of male versus female rabbitst

Male Female t Value p Value

Number 13 13Body weight in Kg. 1 s.d. 2.15 A= .18 2.14 4: .16T.B.W. as %body weight a s.d. 73.3 1: 2.6 74.8 : 3.5 1.25 >0.2"Total" G.I. water as %T.B.W. 11.6 ± 1.9 12.5 d: 3.2 0.82 >0.4

* "Total" intraluminal gastrointestinal water refers to water contained in the gut from the cardia of the stomachto the mid-transverse colon.

t Each pair of rabbits was matched for weight, age and period of isotope equilibration.

TABLE III

The intraluminal gastrointestinal waer content in the rabbit

Cecum and proximalStomach Small intestine transverse colon "Total" G.I.

(mi.) (ml./Kg.) (% T.B.W.) (mi.) (mf./Kg.) (% T.B.W.) (mi.) (ml.lKg.) (% T.B.W.) (mi.) (ml./Kg.) (% T.B.W.)

Mean 65 30 4.1 31 18 2.0 93 44 6.0 189 90 12.1s.d. :i:15 d7 =10.9 412 =19 1.0 =122 4113 A-1.6 -140 4124 1:2.7Coefficient of

variation 23% 23% 22% 39% 50% 50% 24% 30% 27% 21% 27% 22%Number of

rabbits* 29 29 28 29 29 28 27 27 26 26 26 25* All animals were allowed 3 to 5 hours of isotope equilibration and thirsted 3 to 5 hours as summarized in Table I.

291

F. GOTCH, J. NADELL, AND I. S. EDELMAN

TABLE IV

The equilibration of DObetwveen plasma and gastrointestinal contents

Equili- Stomach Small intestine Cecum and transverse colonbration

time No. of S.A.R. No. of S.A.R. No. of S.A.R.(hours) animals mean ±s.d. t* p animals mean+s.d. t* p animals mean 4s.d. t* p

2 10 0.864.08 5.60 0.01

4 10 0.984.05 1.25 >0.2 10 1.004.03 0 1.0 10 1.014.03 1.11 0.35 10 1.014.02 1.67 >0.1 10 1.02-.03 2.22 >0.05 10 1.014.03 1.11 0.3

GASTROINTESTINAL WATER

U) 0~. t.. N4 N4 @0 ) .46U)E' o0 0o N - - c',." - '., - +- oe@

Go a' - e 4 04 m' N 0% 1 46 10a4 pi U) Nd iw; ) 6 ~ 0

-o 0 U U_ 4 U4ooa IN aN Nooo o o 0 '046

Q - co I -

O n 0|E|o_-

460% @0 U) - W) 0 W U) N U) o

0WU - 0 00 0 0'

t; 0 00e$e e o 46@o0A

X~~~~~~~~~~C14Mo-oooooooooOo<

g E , _ ~~~~~~eq C4 _ > Wm C o-E !:Ov , °+,aN°OO*U ,@11.1-. 0 @'0 W4e. ° Ico

0

U) !@0 @ 0 N .0 ) o o '0 0 4

4, E'i 0 0 0> 0 0 00 0 0 0 0U) tW) W) 04 0' ) U 0 U -4

vb Wi)' co) U)C) U

U)I)I -) 0 4 0'° '° N 0 ) 0c

E U_ - N U N W) N U) N U)

N @0 - 0 ) N @0 irk 4)

SA v v z ~~~v:NO co C4o C)"

U) 0n U) 4N0 U)N 0) NW ) O

NN'.4 N ~~~~~~~~~~~. o-

4')~~~~ ~~~~~~~~~A4

46 46 t'.@0 0' @0 0 a % ) '

N U) N~~=N . N E N ) N:) U U) U4'.~~~ ~~ ~ '0 N '~~0 co4 N t0ic U'~~~~O*~~~~~~v @0 4;~~~~~~~~~~ >

a'.. ~~ ~ ~ i46 4 ' 0 % a" 0 0 '0U)U'0 U) 0 ' '0 ) a. Uco

293

F. GOTCH, J. NADELL, AND I. S. EDELMAN

TABLE VI

Summary of intraluminal gastrointestinal sodium, potassium, chloride, and water content in the rabbit

Calculated concentrationSodium Potassium Chloride Water in intraluminal water'

M4s.d. M4s.d. M4s.d. M4s.d. M4s.d. M±s.d. M ts.d. M-s.d. Sodium Potassium Chloride(mEq.) (% Na.) (mEq.) (% K.) (mEq.) (% Clo) (ml.) (% T.B.W.) (mEq./L.) (mEq.IL.) (mEq./L.)

Stomach 0.8 0.9 0.7 0.7 8.7 11.7 65 4.1 12 11 13440.4 ±0.4 40.3 40.3 ±3.4 44.4 ±+15 ±0.9

Small 3.1 3.2 1.2 1.2 1.8 2.5 31 2.0 100 39 58intestine 41.1 ±1.2 ±0.4 ±0.4 ±0.4 ±0.7 412 ±1.0

Cecum and 10.0 10.2 4.5 4.5 1.2 1.7 93 6.0 108 48 13transverse -4:2.0 42.1 :4-3.1 :+2.7 ±0.3 40.4 ±22 ±1.6colon

"Total"G.I. 13.7 14.2 7.1 7.2 11.8 16.0 189 12.1content ±2.4 ±2.4 42.8 ±2.2 43.6 ±4.5 440 ±2.7

* Calculated concentrations are derived by dividing intraluminal sodium, potassium, and chloride contents by intra-luminal water content. Each quantity was determined in separate series of animals.

equilibration in stomach water, where 4 hourswas required for distribution equilibrium comparedwith 1 and 2 hours for colon and small bowel,respectively, may be a result of at least three fac-tors. The ratio of membrane surface area to in-traluminal volume may be smaller in stomach thanin either large or small bowel. The ratio of sur-face to volume has been proposed as the basis forD20 exchange rates in other transcellular pools(14). The delay of equilibration of labelled waterin small bowel contents compared to large bowelcontents cannot, however, be explained on com-parative ratios of surface area to volume. A sec-ond possible explanation is that mucosal bloodflow, and consequently the rate of delivery of iso-tope in proportion to the volume of intraluminalwater, may be highest in the large bowel and leastin the stomach (15). Finally, active transportof water across gut mucosa may occur and ac-count for some of these differences (16).

In the course of these studies it was noted thatthe contents of the cecum and the proximal trans-verse colon were semiliquid and that in the mid-transverse colon there was a sharp transition zone,1 to 2 cm. in length, where the contents weretransformed into hard, dry pellets of stool. Thiswould suggest that the mucosa of the mid-trans-verse colon acts to conserve water efficiently.

Data from previous experiments in which theintraluminal content of sodium, potassium, andchloride were determined are summarized in

Table VI (4, 5, 7). The last three columns inTable VI show the calculated concentrations foreach of these ions in intraluminal water of stomach,small bowel and large bowel. It is apparent fromthese values that the concentration of sodium, po-tassium and chloride maintained in intraluminalwater bears no direct relation to the electrolytestructure of extracellular fluid. It would seemthat their concentration and abundance in intra-luminal water are determined by autonomousmechanisms in the gastrointestinal tract.

The presence of 14 per cent of Nae, 7 per centof Ke and 16 per cent of Cle in the contents of thegastrointestinal tract in rabbits has important im-plications in body partition studies where the nor-mal anatomy of ion distribution or of ion shifts ismeasured (4, 5, 7). Calculating the extracellularfluid volume from a chloride space and assumingthat all chloride exists in the same concentrationas in plasma will lead to significant errors.Changes in the extracellular space inferred fromchanges in chloride concentration in plasma andexternal chloride balance may also be misleadingsince these calculations are based on the assump-tion that all, or nearly all, of the body chloride isin the plasma-interstitial fluid volume.

The volume of intraluminal water found in thegastrointestinal tract of man was a much smallerfraction of T.B.W. than in the rabbit. The signifi-cance of this species difference cannot be deter-mined from our data since the observations on

294

GIASTR0I9TESTINAL WATER

human subjects must be evaluated cautiously forseveral reasons. Total body water van predictedfrom data on normal subjects; in- conrt, a>;curate measurements were made in the; rabbits.Significant migration of water from the guit mayoccur in critically ill patients. 'Post-mortem,changes in intraluniinal volume may have takenplace during the 6 to 22 hours that elapsed be-tween death and autopsy in these subjects. Themeasurements made in the human subjects con-sequently are not reliable, and further studies areneeded to establish the amounts of intraluminalwater and electrolytes in normal man.

SUMMARY

The volume of intraluminal gastrointestinal wa-ter was measured in rabbits and in human sub-jects studied post mortem. In rabbits this volumewas referred to T.B.W. as determined by D20dilution. In man the intraluminal gut water wasreferred to predicted T.B.W. values.

Total body water averaged 75 per cent of thebody weight in rabbits; 12 per cent of T.B.W.was contained in the lumen of the "total" gastro-intestinal tract, with 4 per cent in the stomach, 2per cent in the small intestine and 6 per cent in thelarge intestine. No significant difference betweensexes was noted in either total body water or thevolume of intraluminal gut water. Deuteriumoxide equilibration was complete in large bowelwater and nearly complete in small bowel waterin 2 hours, but required 4 hours for completionin stomach water. The significance of delayedD20 equilibration in stomach water comparedwith more distal segments of bowel was discussedwith respect to the sites and mechanisms of D20exchange across gastrointestinal membranes.

The gastrointestinal tract of man at post-mortemexamination contained approximately 1.5 per centof the predicted T.B.W. The mean values were0.4 per cent for stomach, 0.7 per cent for smallbowel and 0.3 per cent for proximal large bowel.These values cannot be considered to representthe volume of intraluminal gut water to be foundin the normal living human subject.

The amounts of intraluminal gut sodium, po-tassium, chloride and water in the rabbit are sum-marized in tabular form.

ACKNOWLEDGMENT

It is a pleasure to aclmowl dethe technical assistanceof Miss Mary Rose Halligan and Miss Mary FrancesMorrilli

REFERENCES

1. Edelman I. S.; Olney, J. M., James, IA. H., andMoore, F. D., Body composition.:: stdies in thehuman being by the dilution principle. Science,1952, 115, 447.

2. Edelman, I. S., James, A. H., Baden, H., and Moore,F. D., Electrolyte composition of bone and thepenetration of radiosodium and deuterium oxideinto dog and human bone. J. Clin. Invest., 1954,33, 122.

3. Cizek, L. J., Total water content of laboratory ani-mals with special reference to volume of fluidwithin the lumen of the gastrointestinal tract.Am. J. Physiol., 1954, 179, 104.

4. Edelman, I. S., and Sweet, N. J., Gastrointestinalwater and electrolytes. I. The equilibration ofradiosodium in gastrointestinal contents and theproportion of exchangeable sodium (Na0) in thegastrointestinal tract. J. Clin. Invest., 1956, 35,502.

5. Nadell, J., Sweet, N. J., and Edelman, I. S., Gastro-intestinal water and electrolytes. II. The equili-bration of radiopotassium in gastrointestinal con-tents and the proportion of exchangeable potas-sium (K.) in the gastrointestinal tract. J. Clin.Invest., 1956, 35, 512.

6. Nichols, G., Jr., Nichols, N., Weil, W. B., and Wal-lace, W. M., The direct measurement of the extra-cellular phase of tissues. J. Clin. Invest., 1953, 32,1299.

7. Sweet, N. J., Nadell, J., and Edelman, I. S., Gastro-intestinal water and electrolytes. III. The equili-bration of radiobromide in gastrointestinal con-tents and the proportion of exchangeable chloride(Cl.) in the gastrointestinal tract. J. Clin. In-

vest., 1957, 36, 279.8. Schloerb, P. R., Friis-Hansen, B. J., Edelman, I. S.,

Sheldon, D. B., and Moore, F. D., The measure-ment of deuterium oxide in body fluids by thefalling drop method. J. Lab. & Clin. Med., 1951,37, 653.

9. Solomon, A. K, Edelman, I. S., and Soloway, S.,The use of the mass spectrometer to measuredeuterium in body fluids. J. Clin. Invest., 1950, 29,1311.

10. Edelman, I. S., Haley, H. B., Schloerb, P. R., Shel-don, D. B., Friis-Hansen, B. J., Stoll, G., andMoore, F. D., Further observations on total bodywater. I. Normal values throughout the life span.Surg., Gynec. and Obst., 1952, 95, 1.

295

F. GOTCH, J. NADELL, AND I. S. EDELMAN

11. Edelman, I. S., Exchange of water between bloodand tissues. Characteristics of deuterium oxideequilibration in body water. Am. J. Physiol., 1952,171, 279.

12. Friis-Hansen, B. J., Holiday, M., Stapleton, T., andWallace, W. M., Total body water in children.Pediatrics, 1951, 7, 321.

13. Kruhsffer, P., Inulin as an indicator for the extra-cellular space. Acta physiol. Scandinav., 1946, 11,16.

14. Bering, E. A., Jr., Water exchange of central nerv-ous system and cerebrospinal fluid. J. Neurosurg.,1952, 9, 275.

15. Johnson, J. A., and Cavert, H. M., Kinetics of D20exchange in perfused organs. Federation Proc.,1952, 11, 79.

16. Visscher, M. B., Fetcher, E. S., Jr., Carr, C. W.,Gregor, H. P., Bushey, M. S., and Barker, D. E.,Isotopic tracer studies on the movement of waterand ions between intestinal lumen and blood. Am.J. Physiol., 1944, 142, 550.

296