experience...titration, the accuracy of the determination surprisingly did not seem to suffer, and...

EXPERIENCEWITH THE CHROMICOXIDE METHODOF FECALMARKING IN 1\1A ETABOILIC BALANCIE INVESTIGATIONS

ON HtTNM1ANS

By L. G. WHITBYANDDAPHNELANG

(From the DepartneWt of Chemitical Pathology, Postgraduate Medical School of London, anldthe Mletabolic Unit Laboratory, Hammniersmith Hospital, London, Entgland)

(Submitted for publication November 2, 1959; accepted February 4, 1960)

In metabolic balance studies the assessment ofthe subject's state of balance is derived from esti-mates of the dietary intake and the excretion inthe urine and feces of the constituent under in-vestigation. It is relatively simple to measure thedietary and urinary components, but estimation ofthe fecal component is more liable to error becauseof difficulties in secturing a quantitative fecal col-lection. Where a significant proportion of thedietary intake is recoverable from the feces, forinstance in studies of calciumii balances, accuratefecal collections are of prime importance.

Many investigators employ the "time-marker"method for obtaining quantitative fecal collectionsgiving, for instance, two doses of carmine as amarker, collecting the feces passed between thetime of appearance in the feces of the first andsecond doses of marker, and relating this collectionto the diet consumled and uirine passed in the in-terval between the giving of the markers. Thesuccess of this method depends upon the clearrecognition of the marker in the feces and theaccurate collection of all feces passed during theperiod of study. In investigations in which nomarker is employed (i.e., simple "time-collection"studies), accuracy also depends upon the subject'shaving a regular pattern of colonic emptying.

The "inert indicator" method provides an al-ternative means of estimating the fecal componentin balance studies, and it has the big advantage ofavoiding the dependence on quantitative fecal col-lections. It was first used in experiments withanimals, and a summarized description of the"Edin indicator method," which used chromiumsesquioxide (Cr9O3), was given by Edin, Kihlenand Nordfeldt (1). The theory of the methodhas been discussed by Stanley and Cheng (2).Various substances have been tried as indicators,but all have to fulfill the following criteria: theyshould neither be absorbed from nor be chemically

altered in the digestive tract (and so be fully re-coverable from the feces); they should be non-toxic and should not interefere with the processesof digestion; they should be uniformly dispersedin the feces, after a stufficient time for attainmentin the bowel of a steady state with respect to theindicator; and they should be readily estimatedquantitatively in feces.

Several indicators have been used in a widerange of investigations on different species of ani-mals (e.g., 3-5), but the principal indicators em-ployed in human studies have been Cr.,O3 (1, 2,6-8) and polyethylene glycol (9, 10). Animalexperimiients had shown that Cr2O3 met all the re-quired criteria, and this has been confirmed forman in the above accounts of its application andin our experience. These earlier accounts all hadshown that its use in man provided a valid basison which to express fecal analytical data, whencompared wvith "time-collection" studies; one in-vestigation (7) also showed that fecal analysesbased on Cr,O3 compared satisfactorily with car-mine "time-marker" studies. However, previouswork on the application of Cr2O3 to humans hadinvolved only small numbers of volunteers or pa-tients: i.e. three volunteers (6), six volunteers(7), three volunteers and two patients (2, 8),while Edin and associates (1) merely stated thatthey had applied the method to humans.

This article describes the more extensive ap-plication of Cr.,O3 as an inert indicator in balancestudies on man. No comparison has been madewith polyethylene glycol, but Cr,O3 is theoreticallymore suitable as an inert indicator, since it is asingle chemical substance, whereas polyethyleneglycol is a mixture of polymers which do not allbehave identically in the method used for theirestimation (11). Polyethylene glycol has beenimiainly used in man as an inert reference sub-stance in studies of intestinal absorption; unlike

854

FECAL MARKINGWITH CHROMICOXIDE

Cr2,O, it is suitable for aspiration via a narrow-lumen gastrointestinal tube (9, 10).

METHODS

The results presented here have been derived fromstudies on 4 volunteers and 36 patients, all of whomtook capsules of Cr2O3 during calcium balance investi-gations (the technique has also been used in conjunc-tion with fat, nitrogen and potassium balance studiesin this metabolic unit).

Regimen. Subj ects received a low calcium diet, andcalcium supplements were provided in tablet form, asrecommended by Reifenstein, Albright and Wells (12).The diet was prepared in a separate kitchen by atrained dietitian and was analyzed periodically. Sub-jects were placed on their particular regimen for a pre-liminary (or pre-investigation) period of 3 days, afterwhich balance investigations commenced. Results werecalculated on the basis of a pair of 3-day collections oneach regimen; if the conditions of balance study werethen altered, there was a transitional period of 3 daysfollowed by another pair of 3-day collections. Purga-tives were used whenever indicated in order to securea bowel action at least on alternate days. Fecal col-lections closed at 8 a.m.; if a subj ect had not had aspontaneous bowel action by this time on the last dayof the balance investigations, an enema was administeredand its result added to the fecal collection for that period.

Capsules of Cr2O3 were taken 3 times daily, one witheach of the main meals, beginning when the subj ectsstarted their balance regimen. Capsules were preferredto other means of Cr2O3 administration because of theknown highly abrasive properties of the compound (atrial batch of tablets, specially prepared industrially, hadto be rejected because of their friability and marked vari-ation in Cr2O3 content).

Balance studies rejected. Four balances could not beanalyzed. One volunteer developed appendicitis and hadto withdraw from study; one patient did not cooperatein taking Cr203 capsules; another did not adhere to hisdiet; and a third patient was severely constipated.

Purification of Cr2O3. Commercial Cr2O3 containssmall amounts of dichromate. Two hundred g chromicoxide, green (B.D.H.1) was shaken with successive 800ml quantities of 10 per cent hydrochloric acid until thesupernatant was colorless; it was then washed withdistilled water and dried at 6000 C for 24 hours. Cap-sules were filled with 500 ± 2 mg Cr2O3.

Preparation of specimens for analysis. Diets and feceswere processed as described by King and Wootton (13).Samples of homogenized feces were taken as soon asblending had ceased, to avoid sedimenting of Cr2O3which has a high specific gravity (14). The fecalsamples were dried and then ground to a fine powder.Urine collections were made over 24-hour periods andsamples pooled for analysis to correspond with the pe-riods of fecal collections.

1 British Drug Houses.

Standard solutions. Potassium dichromate (AR) washeated for 30 minutes at 1800 C and cooled in a desicca-tor. A 0.1 N solution was prepared in 2 N sulfuricacid.

Ammonium ferrous sulfate: an approximately 0.1 Nsolution was prepared in 2 N sulfuric acid and standard-ized daily against potassium dichromate, using 0.5 mlN-phenyl anthranilic acid solution as indicator (15).

Estintation of Cr2O3 in feces. One-half g samples ofdried feces were digested with 20 ml AR concentratednitric acid until all of the organic matter was oxidized(the flask contents were then a suspension of greenCr2O, in a clear liquid). Digestions were carried out in250 ml quartz Erlenmeyer flasks on a hot plate in a fumecupboard and bauxalite boiling chips were added;further additions of nitric acid were made, if necessary,to avoid evaporation to dryness. Twenty ml AR per-chloric acid (60 per cent; sp gr 1.54) was then added andheating recommenced (behind a screen of Perspex) un-til Cr2O3 had been converted to dichromate. The con-ditions of the digestion of specimens with nitric andperchloric acids conformed to the recommendations ofthe Analytical Methods Committee concerning the useof perchloric acid (16), and the practical directions ofSmith (17) were followed; these articles should beconsulted since there is a risk of serious explosion.

For the final estimation of dichromate, 50 ml 6 N sul-furic acid, 20 ml standardized ammonium ferrous sulfate(more is occasionally necessary) and 0.5 ml indicatorare added, to the flask contents (final volume approxi-mately 150 ml). The excess of ammonium ferrous sulfateis estimated by back-titration with potassium dichromateadded dropwise from a buret. The flask is shaken con-tinuously but the endpoint is not so clear as in the stand-ardization titration; it is marked by the disappearance ofgreen color from the solution and the appearance of aslate-like indeterminate shade which develops, afterstanding for a few minutes, into a cherry color. Theconditions of the titration conform to the recommenda-tions of earlier workers (15, 18, 19), and the reasonfor the slow development of the indicator's change incolor at the endpoint has not been discovered. Furness(18) and Stockdale (19) observed that, although manyfactors interfered with the sharpness of the endpointtitration, the accuracy of the determination surprisinglydid not seem to suffer, and they regarded the choice ofindicator as largely a matter of personal preference;this has been our experience also.

Accuracy of Cr2O3 determiniiationts. Known amounts(10 to 50 mg) of purified Cr2O3 were converted to di-chromate and estimated; the recovery was 98.4 + 2 percent (mean ± SD mean; 21 determinations). Similarquantities of Cr2O3 were added to 0.5 g samples of driedfeces (specimens from patients who had not receivedCr2O,) and the recovery was 98.6 ± 1.9 per cent (mean± SD mean; 25 determinations). Duplicated analysesfor Cr2O3 in fecal specimens from patienits who had re-ceived Cr2O3 agreed with a root mean square differenceof 1.2 per cent (21 duplicated analyses).

The results of these recovery experiments are satis-

855

L. G. WHITBY AND DAPHNELANG

factory, considering that the digestions are carried outin simple apparatus. Gorsuch (20) obtained similarfigures when investigating the recovery of chromiumfrom organic and biological material by a radiochemicaltechnique; he fitted each digestion flask with a splash-head leading to a receiver, recovered 2 per cent of theradioactivity in the distillate, and attributed this lossfrom the digestion flask to the formation of volatilechromyl chloride.

In the present investigationi, volunteers and patientshave ingested 1.50 g Cr2O3 daily, but in the calculationswhich follow, this figure has been modified to 1.48 gto correct for the observed incomplete recovery of Cr2O3in the analytical procedure.

Determiinations of calcium. These were made with aflame-spectrophotometer (21). Cr2O3 is unaffected bythe dry-ashing procedure (13) employed in these analy-ses, but its presence in the residue after ashing does notinterfere with the estimation of fecal calcium in theflame-spectrophotometer.

Uniiformity of mixing of Cr2O3 in feces at the timie ofevacuation. Irwin and Crampton (7) made a pilot studyof one volunteer who received a shredded wheat diet;Cr2O, was given in various ways at mealtimes, but un-mixed with the diet. They found that thrice dailyadministration of Cr2O. resulted in uniform mixing ofthe index substance in the feces at the time of evacuation,whereas once daily dosage did not.

In the present investigation duplicate samples of fecesfrom 4 patients have been analyzed; the samples weretaken from the beginning and end of a single fecal evacu-ation and the wet specimens were used directly in theprocedure for estimating Cr2O:,. The results of these 4paired analyses agreed with a root mean square differ-ence of 1.6 per cent, and it was concluded that thrice dailydosage with capsules of Cr2O: gave satisfactory mixingof the indicator in the feces.

Explanation For All Figures

In each figure data are included about fecal output,calcium intake and excretion, and recovery of chromiumsesquioxide (Cr2O3) from the feces. These data arepresented under two headings (A and B) and the indi-vidual periods of balance (indicated by Roman numerals)correspond in the two parts of each figure.

Each balance period is of 3 days' duration, and thesubject begins the balance regimen and the ingestion ofCr2O3 at the start of Period I. Specimens are collectedfor calcium estimation from Period II onward, until thebalance regimen is ended; if the regimen is altered, thereis a transitional period in which no calcium analyses areperformed, followed by further collection periods.

Upper pair of histograms. A: The average daily weightof fecal dry matter in each period (obtained by dryinga representative sample of feces). B: The theoreticaldaily weight of fecal dry matter inl each period, ormean stool flow (2) ; this is derived from measurementof the fecal Cr2O. content, and is the weight of fecal dry

matter that would contain the same quantity of Cr2O3 asthe daily intake (i.e., 1.50 g Cr2O3 per day).

Lowzer pair of histogra(ms. These show the calciumbalance data plotted according to the conventions ofReifenstein and colleagues (12). The dietary intake ismeasured downward from the baseline and is repre-sented by the lowest horizontal line in these histograms.The calcium content of the excreta is built upward fromthis level (fecal component hatched, urinary componentunshaded). The state of balance is shown by the heightof each completed histogram relative to the baseline; apositive balance is represented by a gap between the topof a histogram and the baseline, and a negative balanceexists when the histogram reaches above the level ofthe baseline.

The state of balance for any period is different in thecorresponding parts of each figure; this is due to thedifferent contributions made by the fecal components inthe two methods of expressing the calcium balance data.In A the fecal calcium excretion is the product of thefecal calcium concentration (in milliequivalents pergram dried feces) and the average daily weight of feces.whereas in B it is the product of the fecal calcium coIn-centrationi and the mean stool flow.

Unipaired histogram (bottomii. right). This shows theaverage daily recovery of Cr2O3 from the feces for eachperiod; it is the product of the fecal Cr2O3 content andthe average daily fecal dry weight.

RESULTS

Many of the ways in which use of an inert indi-cator modifies and supplements balance data will beconsidered in relation to the figures, in which theresults of four representative balance studies arepresented. A general explanation for all the fig-ures has been given, and this should be studiedbefore referring to the individual figure legends.

The need for the use of an inert indicator

1) To measure individual irregularities inbowel habit. The degree of variation in fecaloutput is demonstrated by examination of the re-covery of Cr.,O3 from the feces in single balanceperiods. Although the average recovery of Cr2O3from feces during actual periods of balance in-vestigation approximately equals the daily intakeof inert indicator, considerable variation fromthis figure is observed between the values obtainedfor individual collection periods (e.g., Figures 1,3 and 4; bottom right histograms).

The balance studies reviewed here have beenconducted on the basis of paired three-day col-lection periods. and 52 such balance investigations

856


(from 33 patients and 3 volunteers) have beenexamined to assess the recovery of Cr2O3 duringthe periods of balance study (i.e., the pre-investi-gation and post-balance figures for excretion ofCr2O3 are not being considered here). The aver-age recovery of Cr2O3 from the feces for the firstperiod of each pair of balance periods was 1.42(range 0.60 to 2.53) g per day, and for the sec-ond period it was 1.38 (range 0.70 to 2.12) g perday; the theoretical recovery, allowing for lossesof Cr2O3 during the analyses, was 1.48 g per day.

This wide variation in daily recoveries of Cr2O3is a measure of the inconstancy of human fecaloutput when collection periods are of three days'length. Many investigators use neither carminemarker nor inert indicator, but employ timed col-lections of feces extending over six days, in orderto reduce the effect of individual variation in dailyfecal output. Output variation may still be con-siderable, even when feces are collected over six-day periods, as is shown by calculation of theaverage daily recovery of Cr2O3 from the fecesin the above 52 balance investigations, treatingeach pair of three-day collections as one six-day collection; the average recovery was 1.40(range 0.70 to 1.96) g Cr2O3 per day during thesix-day periods.

2) To correct for the incompleteness of fecalcollections. The completeness of fecal collectionsin this unit has been assessed by determining thetotal recovery of Cr2O3 from the feces of six pa-tients who were kept in the unit for at least threedays after their balance investigations had endedand after they had stopped taking Cr2O3. Fecespassed throughout the period of Cr2O3 administra-tion and for three days after its cessation werecollected as completely as possible and analyzedfor Cr2O3. The total recoveries of Cr2O3 fromthese six patients were 88, 91, 92, 93, 95 and 98per cent of the administered dose; the graph ac-companying the bottom right histogram in Figure1 shows the progress of Cr2O3 recovery from thefeces of the last of these patients.

Stanley and Cheng (2, 8) used a liquid diet inwhich Cr,O3 was uniformly dispersed; in one oftheir subjects excretion of the indicator was al-most complete 72 hours after the last dose ofCr2O3. Irwin and Crampton (7) gave their pilotsubject a diet of shredded wheat and adminis-tered Cr.O3 separately but at mealtimes; they

found that passage of Cr2O3 in the feces had endedwithin 48 hours of the last ingestion. Two of theabove six patients were retained for six days afterthe last capsule of Cr2O3 had been administered;

A3ALANCEPERIODS I a m x

4

E

V. io-

IO-

I-a

A 10

ig

S 20-u

B

I inrx

II I

lizLJm

'1AW'~~~~~10,O>1 L~~~~~~0 20-.'I-20

FIG. 1. RESPONSEOF A PATIENT TO A LOW CALCIUM

INTAKE. Cr2O3 was administered during Periods I toIII inclusive; the actual times of balance study werePeriods II and III, but feces were collected for Cr203analysis for 3 days before and 3 days after the balancestudy proper. The mean stool flow (top histogram, B)has real meaning only while Cr2O3 is being administered,and once the steady state has been attained (Periods IIand III), when it should be compared with the averagedaily fecal dry weight. The histograms for the meanstool flow follow the pattern shown in this figure in allbalance investigations, falling to a relatively constantvalue with attainment of the steady state, and risingagain when dosage with Cr203 ceases. The graph ac-companying the bottom right histogram shows the cum-ulative recovery of Cr,O, from the feces during and afteradministration of Cr..O,. See "Explanation For All Fig-ures" in text.

857


A

I A 1 3r Y 21 x1. .

B

I I. 2.M. Y.M 1

CI]


INTAKE. Two different balance regimens have been em-

ployed; in Period IV dosage with 9a-fluorocortisol com-

menced. No calcium analyses were made in Periods Ior IV, during which the patient was adjusting to eachregimen. Dosage with Cr2O3 ceased after period VI.Three of the histograms are incomplete. In Period Vfeces were collected and processed routinely (i.e., a

"complete" collection was made); a representative samplewas taken, but the data on weighings of the wet feceswere lost, accounting for the gaps in A and in the bot-tom histogram of B. Introduction of 9a-fluorocortisolresulted in a marked change in the mean stool flow(compare Periods II and III of the top histogram, B,with Periods V and VI of the same histogram) a new

steady state has been attained in Period V. This figureshows how the use of an inert indicator can act as an

insurance policy against accidental loss of data; the cal-cium balance data in B are complete, anid the subject'sstate of balanice is accurately (leterminable.

the feces of these l)ost-l)alallce l)erio(ls were proc-esse(l as two three-day collections in the uistialway. No trace of Cr,O., was (letecte(I in the fecesof these later periods. This suiggests that three

days was a sufficiently long time (with the dietaryand nursing regimen employed here) for feces tobe collected after stopping Cr2O3 administrationin order to assess the total recovery of indicatorfrom the stools.

It may be concluded that in this metabolic unita "quantitative" fecal collection is only 93 percent complete. Assuming that "quantitative" fe-cal collections in other metabolic units are simi-larly incomplete, these studies reveal a minorsource of error in balance investigations for whichan inert indicator can correct, but for which nocorrection is possible in simple "time-collection"or "time-marker" studies.

The attainment of a steady state for excretion ofCr203

When a subject receives Cr2,O the fecal con-tent rises from zero to a level which becomesfairly constant for a given individual on a par-ticular balance regimen; expressed in anotherway, the subject's mean stool flow falls, with ap-pearance of Cr2O3 in the feces, and becomes con-stant, to within 15 per cent of a mean value (seebelow) for the particular regimen. When dosagewith Cr,O3 ceases, it continues to be excreted fora time, and a value for the "mean stool flow" canstill be calculated although it has ceased to haveany real meaning. The general pattern followedby the mean stool flow in any balance investigationis illustrated in Figures 1 and 2 (top right histo-grams).

The time needed for fecal excretion of Cr2O3to become steady depends partly upon the dietaryregimen and partly on the individual. Stanleyand Cheng (2, 8), using Cr2O3 dispersed uni-formly in a liquid diet, found that their subjectsattained steady levels of fecal Cr203 excretionwithin 48 hours, and Irwin and Crampton (7)found that the steady state was attained with equalrapidity under their conditions. With our sub-jects a steady level of Cr.,O. excretion has failedto be attained within 72 hours of starting Cr2O:,dosage in approximately one qtuarter of the bal-ances. To reach this conclusion, the 52 balanceinvestigations listed above were examiinie(d to com-pare the percentage difference between fecal (lryweight figures for the in(lividual perio(Is in eaclof a pair of balance periods, before and after cor-

BALANCEPERIODS

i 40-

E

nW 20-U.

0-

201

a-Vr0

E

zI

m

-

0-

20-

858


rection of these figtures in terms of the Cr,O, out-ptut of the feces (i.e., the comparison which ismade between the upper pair of histogramiis ineach of the figures). With the uncorrected datathe mean difference in fecal dry weight betweenperiods was 31.4 per cent (difference betweenpairs of data expressed as a percentage of themean value; range 0 to 84 per cent), and with thecorrected data the difference was 12.8 per cent(range 0 to 37.3 per cent). With extended bal-ance investigations it was noted that the meanstool flow stayed constant, to within 15 per centof the mean value (e.g., Figure 3), unless thebalance regimen was altered (e.g., Figure 2); achange of regimen sometimes caused a temporarylarge change in the mean stool flow, but a newsteady state was always attained within 72 hours,i.e., by the end of the transitional period.

A degree of agreement of 15 per cent about theaverage value of the mean stool flow for pairs ofbalance periods on a given regimen is now ap-plied retrospectively as a criterion of whether apatient had reached a steady state with respect toCr,O, in the first period of that balance investi-gation. As already noted, with a three-day pre-investigation period several of our patients didnot meet this criterion, and latterly these pre-investigation periods have been lengthened to fivedays; so far ten patients have been investigatedwith this modified regimen, and the variation inmean stool flow has satisfied the above arbitrarycriterion in all cases. Failure of the mean stoolflow to attain a steady value may reflect an in-complete adjustment of the bowel to a balanceregimen (e.g., because a new diet has caused achange in the intestinal flora) or failure of thefecal output of Cr9O3 to reach a steady value; ineither case it means that the fecal component ofthe calcium balance data from such periods can-not be satisfactorily corrected in terms of themean stool flow, since it is unsound to correctanalytical data by the inert indicator techniqueuntil the patient has become adjusted both to theindicator and to the particular balance regimen.

Applying this criterion for the attainment of asteady level of Cr2O3 excretion to the above 52balance investigations, 37 were acceptable. Inthese 37 selected balances the mean difference inuncorrected fecal weight data between periodswas 32.2 (range 1.1 to 84) per cent, which com-

pared with a difference of 6.4 (range 0 to 13) percent between the mean stool flow for individualperiods.

Calcium balance data

The upper pairs of histograms in the figuresshow how the use of Cr2O3 has an "averaging" ef-fect on the fecal weight data, and the lower pairsof histograms (e.g., Figures 1 and 3) show howapplication of this averaging effect improves theagreement between the results of individual bal-

ABALANCEPERIODS

20-

' 0-I

20

4

o 40-

co

80-

U 120-

160-

I ] m ME 7 z

JILEbL I

FX =I Fx ]

2

2.

7

/

/

/

/

B

I x x E 7I

Ihi

B I-

I,

FIG. 3. RESPONSEOF A PATIENT TO A HIGH CALCIUMINTAKE. The average daily fecal dry weight (top histo-gram, A) shows considerable variation from period toperiod when compared with the mean stool flow (tophistogram, B). This variation is responsible for themarked fluctuation in the state of calcium balance (lowerhistogram, A), whereas in the corresponding histogram(B) there is little variation between periods, and thetrue state of calcium balance is accurately determinable.No data are available for fecal Cr2O3 content either be-fore or after the actual periods of balance investigation.

859


B

21 K N T

r--=fii

I7 ,

2


INTAKE, FOLLOWEDBY RESPONSETO A HIGH CALCIUM IN-

TAKE. No analyses for calcium were made in PeriodsI or IV, during which the patient was adjusting to thetwo balance regimens. Change of regimen in this pa-

tient has resulted in only a small change of mean stoolflow (compare Periods II and III of the top histogram,B, with Periods V and VI). The patient was moderatelyconstipated during Periods V and VI, which accountedfor the small fecal calcium excretion in A; use of theinert indicator completely altered the interpretation ofthis part of the balance investigation (B). A similarresult would have been obtained if the indicator had beenused to assess the fecal calcium component of a randomsample from each of these last two periods. No dataare available for fecal Cr2O3 content after Period VI,when the balance investigation ceased.

ance periods. Considering the same 52 unselectedbalance investigations, the mean difference be-tween uncorrected estimates of fecal calcium ex-

cretion in the pairs of balance periods was 38.4(range 0 to 136) per cent, compared with a mean

difference of 18.9 (range 1 to 78.2) per cent ob-

tained when correction was applied for Cr2O3 ex-cretion. If the 15 balances are rejected in whichexcretion of Cr.,O3 does not meet the standard ofsteadiness proposed above, these values for themean differences become 32.8 (range 0 to 136)per cent and 15.3 (range 1 to 70.6) per cent foruncorrected and corrected fecal calcium estima-tions, respectively. The agreement between cor-rected fecal calcium estimations is less than thatbetween the mean stool flow values for pairs ofperiods, but this is due to the presence of addi-tional sources of error in the former comparison,which is dependent upon the accuracy of thedietary intake and of the fecal calcium estimations,in addition to the accuracy of the analyses forCr2O3.

Random sampling

In animal investigations an inert indicator isoften used to enable fecal analyses to be performedon "grab" or random, samples. Cr2O3 has notbeen used for this purpose in the present work,but Figures 2 and 4 illustrate two ways in whichthe use of the inert indicator technique can com-pensate for incompleteness of fecal data. In Fig-ure 3, Period V, the gaps in the histograms aredue to loss of weight data concerned with thepreparation of the specimen of dried feces; thecalcium excretion data in Figure 2B are com-plete, however because the only figures necessaryfor the fecal calculations are estimates of the cal-cium and Cr2O. content of portions of the samespecimen of feces, and a knowledge of the dailyintake of Cr2O1 (2). This type of calculation un-derlies the use of the inert indicator technique inanalyses of random samples, whenever feces can-not be collected quantitatively, but the major as-sumption then made is that the sample whichhas been "grabbed" for analysis is representativeof the entire fecal output of a balance period.

In Figure 4 the calcium balance data for Pe-riods V and VI have been modified from astrongly positive state of balance (in part A)into a steady, slight negative balance (in part B).The recovery of Cr203 in each of these periodswas 0.70 g per day; this low value could havebeen due to accidental loss of approximately 50per cent of the fecal collection from each period,but was in fact due to the patient's becoming

ABALANCEPERIODS I Hx Nm xr I

. . . *

'.' 40,

DEL%.J 201

20

' 0O

UT

< 20

<

S 0

40

-10

-4

860


moderately constipated. In a time-marker studythe feces passed in Period VI would have been(correctly) allocated to the second half of Pe-riod V, and the patient would have been retainedin hospital until the feces belonging to Period VIhad been collected; use of Cr2O3 avoided the nec-essity for an indeterminable stay in hospital, andthe agreement between the values for the meanstool flow in Periods V and VI showed that thepatient had adjusted to the changed regimen.The low recoveries of Cr2O3 in these two periodscould alternatively have been due to loss of fecalmaterial, but the use of the inert indicator wouldstill have produced similar results (with the as-sumption that the feces analyzed were "representa-tive"), whereas balance studies conducted on asimple time-collection or a time-marker basiswould have been incomplete for these periods,or would have had to be calculated on a theoreticalbasis (12).

Side effects

No side effects attributable to the administra-tion of Cr2O3 have been observed. It is possible,however, that Cr2O3 was responsible for the de-velopment in one volunteer of symptoms and signssuggesting a diagnosis of pelvic appendicitis. Atlaparotomy the appendix appeared normal, andhistologically it was normal. The lumen of theappendix contained a fecalith which, unfortu-nately, was not preserved for analysis. The in-terest of this case lies in the fact that Cr2O3 (sp gr5.21) is more dense than barium sulfate (sp gr4.50) (14), and there have-been rare examples ofacute appendicitis directly attributed to the use ofbarium contrast materials in X-ray examinations(22, 23).

DISCUSSION

The inert indicator method has been used hereprimarily to correct for the variation in the amountof feces excreted from day to day by humans dur-ing metabolic balance studies, and to correct forany incompleteness in "complete" fecal collections.The data for the recovery of Cr203 from the fecesshow the magnitude of the variability in fecal out-put, both when calculated over three-day and six-day periods of balance study, and the ability tocorrect for this variation is one of the chief ad-

vantages of the inert indicator method over analy-ses based on simple time-collections of feces. Byimproving the agreement between individual pe-riods in a balance investigation (e.g., Figure 3),it enables a better assessment to be made of theaccuracy with which the subject's state of balancehas been determined.

Use of the indicator to correct for incompletefecal collections has amounted to an insuranceagainst certain mistakes that occasionally occurand which can ruin a balance investigation basedon simple "time-collection" or on "time-marker"studies. Two examples have been considered(Figures 2 and 4), but no detailed assessment hasbeen made of the use of Cr2O3 to enable the fecalcomponent in human balance investigations to beestimated by analysis of random samples. Inanimal work Cr2O3 has often been used as thebasis for random analyses, but Kane, Jacobsonand Moore (3) stressed that it can be difficultto obtain a truly representative random sampleand some species of animal are known to havedifferent patterns of diurnal and nocturnal fecalexcretion (24). Nevertheless, an assessment ofrandom samples for accurate balance work inman will be worthwhile, if the inert indicatortechnique becomes widely adopted, since randomsamples might enable some balance investigationsto be accurately performed outside of metabolicunits, even though fecal collections in generalhospital wards are so often incomplete.

Another advantage, which the inert indicatormethod has over both "time-collection" and"time-marker" studies, is the shorter time that apatient needs to be kept on a particular balanceregimen. Because of the known variability inthe frequency and amount of human daily fecalexcretion, six-day periods of balance studyhave been recommended (12). This means thatthe minimum length of hospital stay for in-vestigation of a single variable (e.g., the responseof a patient to a low calcium diet) is at least 18days, and with the "time-marker" method theduration of hospital stay will be further pro-longed, while awaiting the passage of the seconddose of marker. Our experience with Cr2O3 hasconvinced us that, after an adequate preliminaryperiod on the balance regimen, it is practicableto base each investigation on two collection pe-riods, each of three days, although a third three-

861


day period would act as a safeguard by revealingthe presence of a trend due to some unforeseen anduncontrolled variable in the experiment (12).Further, in contrast to "time-marker" studies,when delays in the passage of the marker oftenoccur, the inert indicator technique has beenfound to help administratively, since it has provedpossible to plan each metabolic study and fore-cast accurately the stay of each patient in theward.

The main disadvantage of this technique hasbeen the labor involved in the accurate filling ofcapsules with Cr2O3. With the inert indicatortechnique it is, however, essential that the intakeof indicator be regular and accurately known, andpatients' cooperation has been obtained by ex-plaining that the capsules assist in the laboratory'sanalyses; nevertheless, one patient failed to co-operate in taking the capsules.

The requirements of an ideal fecal inert indi-cator were listed in the introduction and the pres-ent study supports the conclusion of earlier work-ers that Cr2O3 fulfills these requirements. Cri-teria for ensuring restropectively that a patienthad attained a steady state with respect to dosagewith Cr2O3 have been proposed, and the im-portance of preliminary periods of adequate lengthon any balance regimen has been emphasized. Anote of caution should, however, be added at thispoint; although several cases of steatorrhea havebeen investigated (some with an average dailyfecal fat excretion of 25 to 30 g), none of the pa-tients studied has been suffering from a gross dis-order of intestinal motility. When intestinal mo-tility is abnormal it is possible that "streamlining"of intestinal contents may occur; under suchconditions the density of Cr2O3 might cause it tosediment and to be carried irregularly along theintestinal lumen. If this occurs, a steady valuefor the mean stool flow may never be obtained.In such patients (sprue is a possible example) anadditional requirement for an ideal fecal inert in-dicator will have to be included, namely that itshall be uniformly dispersed with the intestinalcontents throughout their passage in the gut, andpolyethylene glycol might be more satisfactorythan Cr2O3 under these circumstances. Extracare will be necessary in such patients, if fecalanalyses are based on analysis of random samples,

to ensure that the sample analyzed is truly repre-sentative.

SUMMARY

This article describes the application of chro-mium sesquioxide as an inert indicator in meta-bolic balance studies on 36 patients and 4 normalvolunteers. The information gained by this tech-nique is mainly considered in relation to fourrepresentative calcium balance investigations.

The inert indicator reveals the magnitude offluctuations in the daily output of feces, when col-lections are made over both three-day and six-day periods; use of chromium sesquioxide cor-rects for these fluctuations. The technique alsocorrects for small losses of feces which occur evenin "complete" fecal collections, and inclusion ofan inert indicator in balance investigations actsas a safeguard against the major losses whichoccasionally occur in metabolic work.

The advantages of the inert indicator techniqueare considered in relation to simple "time-collec-tions" and "time-marker collections" of feces, andit is proposed that balance investigations can beconsiderably shortened by its use. The wider ap-plication of inert indicators in connection with theanalysis of random samples of human feces isconsidered.

ACKNOWLEDGMENTS

The authors wish to acknowledge the interest andadvice of Professor Russell Fraser, and to thank himfor permission to publish details of his patients. Theyare also very grateful to Dr. M. T. Harrison and Dr.I. D. P. Wootton for helpful discussions, and to Sister S.Cooper and the staff of the Metabolic Unit.

REFERENCES

1. Edin, H., Kihlen, G., and Nordfeldt, S. A sum-marized description of "Edin's indicator method"for the determination of the digestibility of feedsand feed mixtures. LantbrHogsk. Ann. 1944, 12,166.

2. Stanley, M. M., and Cheng, S. H. Excretion fromthe gut and gastrointestinal exchange studied bymeans of the inert indicator method. Amer. J.dig. Dis. 1957, 2, 628.

3. Kane, E. A., Jacobson, W. C., and Moore, L. A. Acomparison of techniques used in digestibilitystudies with dairy cattle. J. Nutr. 1950, 41, 583.

4. Fournier, P. De l'emploi de l'oxyde de titane pourl'etude quantitative de l'absorption intestinale.C. R. Acad. Sci. (Paris) 1950, 231, 1343.

862


5. Sperber, I., Hyden, S., and Ekman, J. The use ofpolyethylene glycol as a reference substance inthe study of ruminant digestion. LantbrH6gsk.Ann. 1953, 20, 337.

6. Kreula, M. S. Absorption of carotene from carrotsin man and the use of the quantitative chromicoxide indicator method in the absorption experi-ments. Biochem. J. 1947, 41, 269.

7. Irwin, M. I., and Crampton, E. W. The use ofchromic oxide as an index material in digestiontrials with human subjects. J. Nutr. 1951, 43, 77.

8. Stanley, M. M., and Cheng, S. H. Cholesterol ex-change in the gastrointestinal tract in normal andabnormal subj ects. Gastroenterology 1956, 30,62.

9. Borgstrom, B., Dahlqvist, A., Lundh, G. and Sjovall,J. Studies of intestinal digestion and absorptionin the human. J. clin. Invest. 1957, 36, 1521.

10. Lundh, G. Intestinal digestion and absorption aftergastrectomy. Acta chir. scand. 1958, suppl. 231,15.

11. Hyden, S. A turbidimetric method for the determi-nation of higher polyethylene glycols in biologi-cal materials. LantbrH6gsk. Ann. 1956, 22, 139.

12. Reifenstein, E. C., Jr., Albright, F., and Wells, S. L.The accumulation, interpretation, and presenta-tion of data pertaining to metabolic balances, no-tably those of calcium, phosphorus, and nitrogen.J. clin. Endocr. 1945, 5, 367.

13. King, E. J., and Wootton, I. D. P. Microanalysis inMedical Biochemistry, 3rd ed. London, J. and A.Churchill, 1956, pp. 135, 183.

14. Hodgman, C. D., Weast, R. C., and Selby, S. M.Handbook of Chemistry and Physics, 40th ed.

Cleveland, Chemical Rubber Publishing Co., 1958,pp. 541, 563.

15. Vogel, A. I. A Text-Book of Quantitative InorganicAnalysis: Theory and Practice. London, Long-mans, Green, 1939, p. 366.

16. Report by the Metallic Impurities in Organic MatterSubcommittee (Analytical Methods Committee).Notes on perchloric acid and its handling in ana-lytical work. Analyst 1959, 84, 214.

17. Smith, G. F. The dualistic and versatile reactionproperties of perchloric acid. Illustrative experi-mental demonstration procedures. Analyst 1955,80, 16.

18. Furness, W. The determination of chromium inchromite. Part I. Volumetric determination ofchromate with special reference to the effects ofvanadate and arsenate. Analyst 1950, 75, 2.

19. Stockdale, D. The estimation of iron by dichromate.Analyst 1950, 75, 150.

20. Gorsuch, T. T. Radiochemical investigations on therecovery for analysis of trace elements in organicand biological materials. Analyst 1959, 84, 135.

21. MacIntyre, I. The flame-spectrophotometric de-termination of calcium in biological fluids and anisotopic analysis of the errors in the Kramer-Tis-dall procedure. Biochem. J. 1957, 67, 164.

22. Gubler, J. A., and Kukral, A. J. Barium appendi-citis. J. int. Coll. Surg. 1954, 21, 379.

23. Young, M. 0. Acute appendicitis following retentionof barium in the appendix. Arch. Surg. (Chicago)1958, 77, 1011.

24. Mueller, W. J. Feasibility of the chromic oxide andthe lignin indicator methods for metabolism experi-ments with chickens. J. Nutr. 1956, 58, 29.

863

experience...titration, the accuracy of the determination surprisingly did not seem to suffer, and...

Documents