ENDOGENOUSPRODUCTIONOF FIXED ACID ANDTHEMEASUREMENTOF THE NET BALANCEOF ACID

IN NORMALSUBJECTS*

lBY- ARNOLDS. RELMAN, EDWARDJ. LENNONt AND JACOB LEMANN, JR.WITH THE TECHNICAL ASSISTANCE OF HELEN P. CONNORS

(From the Evans Memorial Department of Clinical Research, Massachusetts MemorialHospitals, and the Department of Medicine, Boston University School of Medicine,

Boston, Mass.)

(Submitted for publication January 30, 1961; accepted March 24, 1961)

It is commonly assumed that one of the essen-tial functions of the kidney is to maintain acid-base balance by excreting an amount of acidequivalent to the net quantity of protons releasedfrom fixed (i.e., nonvolatile) endogenous acids.Excretion of acid in the urine is easily measured.Therefore, if the net rate of production of fixedacids could also be measured independently, itmight be possible to evaluate the external acidbalance of the body in the same manner as thebalance of mineral elements or nitrogen. In thenormal steady state, the net excretion of acid inthe urine (titratable acid plus ammonium, minusbicarbonate) would be expected to equal the sumof the nonmetabolized acids ingested plus thoseendogenously produced. Except for clinical sit-uations in which stool losses of alkali or acidmight also require consideration, all metabolicdisturbances of the net acid-base balance of thebody would thus be quantitatively determined bythe difference between the acid ingested or endog-enously produced and that excreted in the urine.

At present, however, there is no clear definitionof all the sources of endogenous fixed acids nordoes there exist any satisfactory method for ac-curately quantitating the net acid load independ-ently of measurements of renal acid excretion.It is sometimes assumed that it is the calculated"acidity" or "alkalinity" of the diet "ash" whichnormally determines the endogenous load of fixedacids. This calculation, based simply on the dis-crepancy between mineral cations and the chlo-ride, phosphorus and sulfur of the diet, has the-oretical as well as practical defects (1-4), anddoes not in fact correlate well with the rate ofacid excretion (5).

Hunt has recently adduced evidence to support* Supported by Public Health Service Grant A-3140.t Trainee of the National Institute of Arthritis and

Metabolic Diseases.

his suggestion that acid derived from the oxida-tion of organic sulfur to sulfate constitutes themajor, if not the sole, source of fixed acid (5).In subsequent experiments from this laboratory(6) a detailed study was made of the acidifyingeffects of large oral methionine loads. It wasfound that the increased oxidation of organicsulfur to sulfate was accompanied by the systemicrelease of approximately equivalent quantities ofprotons, which ultimately were excreted in theurine. In many of Hunt's experiments, however,net acid excretion in the steady state was signifi-cantly greater than the measured excretion ofsulfate. In our own studies, in which a naturalsoy flour, rich in potassium, was employed as thenitrogen source, net excretion of acid in the steadystate prior to loading with methionine always wasfound to be approximately 20 to 30 per cent lessthan the excretion of sulfate (6). It thus ap-peared that the net endogenous production offixed acid could not be equated solely with sulfateproduction.

The present study was therefore designed toinvestigate the normal sources of fixed acid morethoroughly. A new liquid-formula diet was de-vised which utilized a purified soy phosphoproteinvirtually free of mineral cations. This made itpossible to measure the acid contribution from thedietary protein more precisely. It was then foundthat the discrepancy between sulfate productionand the excretion of net acid in the steady statecould be quantitatively accounted for by the sumof the total organic acids in the urine plus the acidderived from the combustible cations neutralizingthe phosphate groups in the protein fed. It wastherefore concluded that the net production ofendogenous fixed acids with the present diet wasdetermined by the sum of: a) the oxidation oforganic sulfur in the fed protein, b) the liberationof excess protons from the cations covering the

1621

12ARNOLDS. RELMAN, EDWARDF. LENNONAND JACOB LE.MANN, JR.

phosphate radicals il the protein, and c) theendogenous formation of unnmetabolized organicacids. The existence of an apparent physiologicbalance between acid excretion and the productionof fixed acid measured in this way was also dem-onstrated in a few relatively prolonged balancestudies in which predictable changes in acid bal-ance were temporarily induced by experimentalalteration of the excretion or net production ofacid.

METHODSAND MATERIALS

Ten balance studies were performed on 9 healthy youngadult males. Liquid-formula diets were used which pro-vided 35 to 40 calories and 0.8 to 1.0 g protein per kgbody weight per day. Corn oil, glucose and dextrinserved as sources of fat and carbohydrate. A pure soy-bean protein was used as the protein source. Accordingto the manufacturer, this protein1 is essentially uniformby electrophoresis. It is virtually free of minerals otherthan phosphorus as shown by the following analyses:sodium, 1 mEq/100 g; potassium, 3 mEq/100 g; calcium,<0.1 mEq/100 g; magnesium,2 <0.1 mEq/100 g; chlo-ride, < 1 mEq/100 g; phosphorus, 17 mmoles/100 g(0.53 per cent).

This protein was prepared at its isoelectric pH (5.0 to5.2) and does not disperse well in water. In 7 balancesL-arginine (free-base), in quantities ranging from 25 to57 mmoles per day (0.46 to 0.64 mmoles per g protein),was used to raise the pH of the diet mixture to approx-imately 7.0 and thus effect dispersion of the protein. Inone subject small amounts of ammonium hydroxide wereused instead of arginine. In the remaining 2 balances theprotein was dispersed by the addition of Ca(OH)2 andMg(OH)2, as described in the following paragraph.

In all of the studies except those two in whichCa (OH) 2 and Mg(OH)2 were given, the calculated acidderived from the phosphoprotein was added to the acidproduction measured as urinary inorganic sulfate andorganic acids. This calculation was made on the assump-tion that the phosphate esters present in the protein werecompletely hydrolzed and the covering cations completelymetabolized. Metabolism of an organic monophosphateester does not necessarily yield protons. This result willdepend upon the nature of the cationic group neutralizingthe charge on the phosphate residues. At the isoelectricpH of the protein, phosphate groups are neutralized byprotons or cationic amino groups. When arginine free-base or NH4OHis used to raise the pH of the isoelectricprotein (as was the case in 8 of the present balancestudies), the cations neutralizing the phosphate chargesare combustible and hence metabolism of these cationswill ultimately yield excess protons, just as if the isoelec-tric protein had been fed. In either case, the number of

1 "J-protein-100" kindly made available by the J. R.Short Milling Co., Chicago, Ill.

2 Magnesium analyses through the courtesy of Dr. W.Wacker, Peter Bent Brigham Hospital.

protons released can be calculated by knowing th11e 111it-her of phosphate residues present, the pK' of the phos-p)hate and the final pH of the medium. Thus, withreference to the excellular pH (7.40), each mole of phos-phorus fed would be expected to yield 1.8 equivalents ofacid (pK.' of phosphoric acid = 6.8). If Ca(OH)2 orMg(OH)2 is used to raise the pH of the protein, thecation "covering" the phosphate (Mg'+ or Ca++) is notcombustible, and hence no protons will be released. Inthe two experiments in which Ca(OH)2 and Mg(OH)2were given instead of arginine, the amount of alkali addedwas equal to the calculated excess of H+ which is releasedif a combustible alkali such as ammonium hydroxide orarginine free-base is used to disperse the protein. Underthese circumstances, therefore, the acidifying effect of thearginine phosphate or ammonium phosphate is avoided,and the hydrolysis and metabolism of the phosphateresidues contribute no endogenous acid. It should benoted that, when arginine or ammonium phosphate is fed,it is not the phosphate residues but rather the combustiblecations neutralizing these residues which give rise to theacid. Nevertheless, the quantity of acid so released canbe calculated only from a consideration of the number ofphosphate residues in the protein fed. (Any arginine orammonia added to the phosphoprotein in excess of thisnumber of residues does not give rise to acid.) Any non-combustible anionic residue on the protein would play thesame role here as the phosphate, except that the pK ofthe residue determines what proportion of the protonsyielded by metabolism of the combustible cations willcontinue to be held by the anionic residue at body fluidpH and what proportion will be given up to be handledby the physiological mechanisms for acid excretion.

The soybean protein contained adequate amounts of allthe essential amino acids except methionine, so that asupplement of 1 g of DL- or L-methionine was provideddaily in the basic diets. In addition, 50 to 80 mEq ofpotassium chloride and 90 to 200 mEq of sodium chlorideper day were added to the diets. In 2 experiments 15mmoles per day of calcium phosphate (as CaHPO4) wasalso added. In these 2 studies correction was made forthe alkali thus fed, arbitrarily assuming virtually com-plete absorption. Each subject took 2 multiple vitamincapsules daily and, to prevent diarrhea, they also weregiven small doses of deodorized tincture of opium. Allstools were formed, but soft in consistency.

Each study was preceded by a 2-day period of adapta-tion to the standard diet. The balance then began with acontrol period lasting 3 to 8 days. In 6 experiments onlythese control observations were made. In the other 4studies the control period was followed by an experi-mental period lasting 5 to 7 days during which the fol-lowing changes in the control regimen were made: a) 1subject received sodium bicarbonate, 70 mEq per day individed doses, in addition to the control diet; b) 2 sub-jects received oral loads of neutral (pH 7.40) sodiumorthophosphate, the phosphate mixture replacing a partof the chloride in the control diet, with sodium intakebeing kept at the control level; and c) 1 subject wasgiven acetazolamide, 1.0 g daily in divided doses, in addi-

1692

1623ENDOGENOUSACID PRODUCTIONAND NET ACID BALANCE

CO~~~~~~C

C4

0 C

LO-C- Cr t -r"

0000 0000 000000ooM "

CO CU)c_ 1-o

_e4 o 0 -_ "s 0 es 040 U) -

C_ ) eS - C- U)

U) so U) U U) o0'

U) 0-O _

) C -

CO Nt _NC

000co0000000It It . . .

6 6 6O o

0 '0 .o. .

0'0'0'0'0'

O 6O O o Cs

-4 C"4 %4C "4

U) U) U) U) U)

IU) cm 4 CI m"

r ' -

t- 0 -. - .( 't - 4--4 -4 - -

U)0 j U) U)m e m m 't

00 U)) '0 r-

C, 0 -t -(- ~o t m

C' -00 eq8U) U)O'o

4- -4 -4 t 4

'0 00 00 0t

U) U) U)

,g U) s 00

- eq N 00

00 0 '

Cz- F o 00

0' 0 N C- U) U C"4 %

U 0 U) U) U)

0000'0 in) )0' '0'

C1 C4 N -4 C4 N " c- C4 C

0 0 _ in 0 N uI 00 C 0 - NOC U) n 00- 0 000. 'c"r0 - 0 t- 0' U

v) u) r , 00

r a-0 'e 0 00 0i 0 )" Ue 00 - 0o 0-_ _ " _ _" c4 _- o r- vo - ~o eq -. " C4 M

U t4) toU) 0 000

- C -) e 00 rz V:-4 _-4 _4-

U) '0 r- '0 U).' eU) U) U) ' 0 +U) U) U) U) U) U)'0

CO

0'0U)00000

N 0' 0'0'0' -

U) U))U

%" U)U)-o

v o

00 'o U U) U)C"- - el

wi N6 6 S6- - C4 C4

0 o 00

0- _ -

U) U)C U)0 It 0~ Im m

- m Id 't4 - -

C4 00mU d

0tn C'i t -

V '0 U) U) U)

t-a m -4

0U0 -0-

co0 00U ) 0 0 0C- -I U)

u) Ofi - o0a

"m r- In m

0O 0O )Uu) s 00 m'UU) U)

00 00'00 00 C OU 00 t6

00 U) 0 000-o -0 r U)

0UV)0'00C-C"- 00-U)0 0_ O _.-CU)U )U)oO0000e4C-~' 0'-U)+ 0 000

InlO

000000 0000 00 o0o0000000 00000000000

t U) c0C- 00 -' -U

00

-U'vU

CO)V4

Cd

Z

u

cdu

Cd

OC"

bO ,00C: (n$1-

0):Z19: v)"d

wz*-

E'-°OU.6

*

rt- .b

k..

000

E

CO

.-

U

e

CI

z

r

COz

.~0

0

0E

Cd

- C

Cd

Z

Co 0

oU C

-e

E_

04c

_ CO

*0

Cd M

_ UV

dC

CO C

00

oC V

C0 .-c

O 4

%j 0

-E~

* *-+

E

J0

u

z

COd'-"0

0UHC

H 0>0 CO

5im cla -b

cO NO 14 10 %O so c 1so

0

ARNOLDS. RELMAN, EDWARDJ. LENNONANDJACOB LEMANN, JR.

tion to the control diet. In each of these 4 studies theexperimental periods were followed by recovery periodsof 4 to 6 days' duration.

Further details of the balance technique in use in thislaboratory, as well as descriptions of most of the methodsused for the analyses of diets, urine, feces and blood havepreviously been reported (6, 7). In the present experi-ments "organic acids" in urine were measured by the VanSlyke and Palmer method (8), modified as follows.Phosphate was removed by mechanical shaking for 30minutes with added calcium hydroxide powder. Remain-ing carbonate was then eliminated by addition of HCluntil the pH was exactly at 2.70. A Cambridge model RpH meter with dipping electrodes and a constant tempera-ture bath at 37'C was employed for this purpose. "Or-ganic acids" were then determined by titrating from pH2.70 to pH 7.40, with 0.1 N NaOH. Correction was

made for creatinine. Additional techniques, used in a

few studies, included: a) calcium, by the method of Clarkand Collip (9) and b) citrate, by a pentabromacetonemethod (10).

RESULTS

I. Control observations during the steady state(ten subjects)A. Experimental data. Electrolyte balances

were carried out in ten subjects under steady stateconditions. In one, this period lasted only 3 days,and in another 5 days, but in the others the con-

trol observations lasted 6 to 8 days. Completeanalytical data from one experiment which beganwith a control period of 7 days are shown inTable I. In this study, as well as in the othernine, serum electrolyte concentrations and CO2content and blood pH were essentially constantduring the control periods. As shown in Table I,the urine pH and the daily excretion of all urinaryconstituents were also relatively steady. Fecalexcretion of sodium, potassium and chloride was

low. Fecal calcium varied in six experimentsfrom 10 to 27 mEqper day and fecal phosphorusranged from 6 to 11 mmoles per day.

Cumulative changes in weight were small andvariable (-1.02 to +0.59 kg). Small positivecumulative sodium and chloride balances were

noted in all subjects, consistent with the expectedskin losses. Although not shown in the experi-ment of Table I, nitrogen balances were carriedout in three subjects. Cumulative nitrogen bal-ances ranged from -1.5 to +3.3 g. Cumulativepotassium balances, uncorrected for nitrogen,varied from -13 to -109 mEq. Cumulativecalcium balance varied from -111 to +6 mEq,and cumulative phosphorus balance ranged from

-60 to +2 mmoles, the large negative balancesresulting from stool losses in the subjects whosediets did not contain added calcium and phos-phorus.

B. Derived data. The daily net excretion ofacid (calculated as titratable acid plus ammoniumminus bicarbonate) was approximately constantin every subject, as exemplified by the data inTable I, in the column headed "net acid." Theexcretion of inorganic sulfate was always less thanhalf the net acid excretion, but it was noted thatin each case the discrepancy between sulfate andacid excretion was approximately equal to thetotal "organic acid" excretion plus the calculatedacid derived from dietary protein phosphorus. Ineach subject there were significant daily fluctua-tions in this relationship, as illustrated by the datain Table I, but on a cumulative basis the excretionof net acid for the entire control period was al-ways very nearly equal to the sum of these threequantities. Table II presents a comparison of thecumulative excretion of net acid with the sum ofexcreted sulfate plus "organic acids" plus the cal-culated acid derived from the cation covering thephosphate residues in the protein. This compar-ison is made for the control periods only, in eachof the ten subjects. The close agreement betweenacid excretion and production is apparent, themean discrepancy being -0.4 ± 8.9 per cent.This result is also illustrated in the data in threeindividual subjects shown in the left-hand sectionsof Figures 1-3. An explanation of the design ofthese Figures is given below.

TABLE II

Comparison of net acid excretion (A) with the sum of urinesulfate and "organic acids," plus calculated acid

phosphorus from diet (B) during steadystate control conditions

Discrepancy

Subject Period A B (A-Bxioo)

days %TW 7 510 552 -7.6RS 6 548 629 -12.9JL 6 460 518 -11.2JB 5 350 342 +2.3GdeF 8 536 566 -5.3RH 3 328 344 -4.7FR 6 466 416 +12.0PVanO, 1 6 591 556 +6.3LS* 8 540 493 +9.5PVanO, 2* 8 633 589 +7.5

Mean4SD 496 +98 500±4101 -0.4 48.9

* Dietary phosphorus neutralized by Ca(OH)2 and Mg(OH)2, helncenot included in calculation of acid production (column B).

1624

ENDOGENOUSACID PRODUCTIONAND NET ACID BALANCE

II. Effects of experimental changes in net acidproduction or acid excretion

If we assume the existence of a steady state ofacid-base balance during the control periods, thedata of Table II seem to imply that the only sig-nificant sources of endogenous fixed acid in theseexperiments were the protons released during:a) the oxidation of organic sulfur to sulfate, b)the generation of organic acids, and c) the com-bustion of the disposable cations covering phos-phate residues in the dietary protein. In anattempt to determine whether acid-base equilib-rium measured in this way could be demonstratedover even longer periods, and to ascertain whetherinduced changes in net acid production or excre-tion would have quantitatively predictable effectson this equilibrium, four additional experimentswere carried out.

A. Sodium bicarbonate. In the first such ex-periment, an amount of sodium bicarbonate wasadministered which was equal to the estimatedrate of acid production during the control period.The essential observations made in this study aresummarized in Figure 1. In the upper half of thefigure are shown the values for serum CO2 con-tent and 24-hour urine pH. In the lower half isshown the "acid" balance, plotted in the conven-tional manner of an electrolyte or nitrogen bal-ance. Acid production, calculated in the manneralready described, is plotted as a heavy line meas-ured downward from the zero axis. Net acidexcretion is measured upward from the line indi-cating acid production (diagonally-shaded area).Thus, clear areas below the zero axis representpositive proton balance, and extensions of theshaded area above the zero axis define a negativeproton balance.

During the control period of this study, theaverage acid production and excretion was ap-proximately 70 mEqper day and, as shown in thefigure, net proton balance was virtually zero.Serum CO2 content and blood pH were essentiallyconstant, as was daily urine pH. After this, 70mEq of NaHCO3was added to the daily diet for7 days, as shown by the stippled area in the figure.Since the measured production of acid and theingestion of protein phosphate remained virtuallyconstant throughout the loading period, as shownby the dashed line, the net effective acid produc-

FIG. 1. NET EXTERNAL ACID BALANCE, SERUM CO2CONTENTAND URINE PH IN A NORMALSUBJECT GIVEN

NAHCO, FOR 7 DAYS. In this and subsequent figures,acid production is plotted downward from the zero lineand net acid excretion in the urine is plotted as a di-agonally-shaded column back up from the heavy lineindicating acid production. Thus, a clear area below thezero line indicates a positive acid balance, and extensionof the shaded area above the zero line indicates a negativeacid balance. During the experimental period acid pro-duction is indicated by the dashed line but, due to theadministration of NaHCO3, net acid production is 70mEq per day less than this, as shown by the heavy solidline. Urine acid excretion is plotted upward from thisline.

tion was zero during this time. This is indicatedin the figure by the heavy line which is nearlysuperimposed upon the zero axis. Acid excretionis plotted upward from this heavy line. Duringthe first 3 days cumulated net proton balance was-85 mEq, because acid excretion did not ceaseimmediately. Thereafter, acid excretion was es-sentially zero during the last 4 days of the experi-mental period. The reduction in net acid excre-tion was mainly accounted for by reductions in theexcretion of ammonium and titratable acid. Al-though urine pH rose to approximately 6.6, nomore than 28 mEq per day of the administeredbicarbonate load was excreted as such in theurine. Serum CO2 content rose slightly.

When the administration of sodium bicarbonatewas stopped, the net effective acid productionimmediately returned to control, but the acidexcretion remained low for 2 more days beforereturning to the control rate. During this periodthe net retention of protons was 78 mEq, anamount closely approximating the net loss of 85mEq of acid which had occurred at the beginningof the period of alkali administration. At the end

1625

ARNOLDS. RELMAN, EDWARDJ. LENNONAND JACOB LENIANN, JR.

of 19 days. when serum CO2 content and urineAH had returned to control, the total cumulativeproduction of fixed acid (corrected for the bicar-bonate fed) was 827 mEq, as compared with acumulative net acid excretion of 806 mEq.

B. Neuttral ortho phosphate. The purpose ofthese two experiments was to increase acid excre-tion without changing acid production. This wasaccomplished by the administration of extra uri-nary buffer in the form of neutral inorganic phos-phate. If measured acid production were to re-main constant, this should result in a negativeproton balance and the development of metabolicalkalosis. Furthermore, we would expect restora-tion of acid-base balance after cessation of phos-phate loading to be accompanied by a (lemon-stral)ly positive proton balance of approximatelyequal magnitude.

Subject TW (Table I) and Subject P VanO(No. 1, Figure 2) were given oral loads of neu-

tral (pH 7.40) sodium orthophosphate for 5 days,the phosphate replacing part of the chloride in thecontrol intake and the sodium intake remainingconstant. After control periods, in which acidproduction and excretion in both studies wereapproximately equal, the phosphate load resultedin an immediate increase in net acid excretion.The rise in acid excretion was due entirely toincreased titratable acid (Table I), which oc-curred despite a rise in urine pH and a reductionin ammonium excretion.

Since there was no significant change in calcu-lated acid production, the initially increased ex-

NEUTRALORTHOPHOSPHATE

94 mM/Day

SERU 10-.0CONTENT 28 __.URINEmM/L 26 NC6.0

J5.0

ACID L- 20BALANCE 0 0

60-60

100 -00

12 13 14 1516 IS IS I0 II 12 113 14 10 10 II?11DAYS

FIG. 2. THIE EFFECT OF FEEDING NEUTRAL SODIUMPHOSPHATEON NET EXTERNAL ACID 13ALANCE, SERUMCO.CONTENTAND URINE PH.

cretion of net acid resulted in a negative acidbalance. Cumulative balance of acid during theloading period was -66 mEq in TWand - 112mEq in P VanO. As would be predicted, therewas a rise in serum CO2 content in both cases,the final level being approximately 3 mmoles perL above control in TWand 5 above control inP VanO.

When phosphate loading was stopped, urinaryexcretion of phosphate immediately returned tocontrol. For the first 2 days of the recoveryperiod urine pH remained somewhat elevatedand ammonium excretion was slightly reduced.Therefore, net acid excretion dropped sharply andacid balance became positive. Cumulative reten-tion of acid during the recovery period was +106mEq in TWNT and +48 mEq in P VanO. SerumCO2 returned to control values and approximateequilibrium between production and excretion ofacid was ultimately restored. For the entireperiod of study, cumulated acid production inTWwas 1,421 mEqand cumulated excretion was

1,342 mEq. In the other subject (Figure 2)total production was 1,618 mEq and total excre-tion 1,693.

C. Acetazolamnide. This experiment was de-signed to reduce renal excretion of acid withoutaltering acid production. After a control period,the subject was given 250 mg of acetazolamidefour times daily by mouth for 7 days. Adminis-tration of the drug produced a large diuresis ofbicarbonate, and therefore net acid excretion was

negative on the first day (Figure 3). This isillustrated in the figure by the dotted bar extend-ing in a downward, or negative, direction from thelevel of acid production. Net retention of acid onthis day was +226 mEq, and during this timeserum CO2 content fell from 31 to 24 mmolesper L. With continuing administration of drug,bicarbonate diuresis waned, and net excretion ofacid became positive. However, acid balance con-tinued to be slightly positive for the next few daysand serum CO2 content continued to fall slightly,reaching a minimal level of 21 mmoles per L.During the last 2 days of treatment the averageacid balance was virtually zero and there was nofurther fall in serum CO2 content.

On cessation of drug, excretion of net acid in-creased and exceeded acid production. Thus,during the last 4 days, acid balance was markedly

1626

ENDOGENOtTSACID PRODtUCTION AND NET ACID BALANCE

7.0URINE

6.0 pH

5.0

SERUMcoz

CONTENTmM/L

ACIDBALANCEmEq/Day

FIG. 3. THE EFFECT OF ADMINISTRATION OF ACETAZOL-

AMIDE ON NET EXTERNAL ACID BALANCE, SERUM CO.CONTENTAND URINE PH. On the first day of drug ad-ministration net acid excretion was negative, due to theloss of urinary bicarbonate. This is plotted as a dashedline downward from the heavy solid line indicating acidproduction. The net positive acid balance on that day isrepresented by the total distance between the dashed lineand the baseline.

negative and serum CO, content was risingsteadily.

On the fifth day of recovery the subject devel-oped a gastroenteritis, and the study had to beterminated before the control state of acid-baseequilibrium could be restored. At this time thecumulative acid production for the total balancewas 1,340 mEq, as compared with a cumulativeacid excretion of 1,259 mEq. The net balance ofacid was therefore +81 mEq; the final serum CO2content was 28.2, as compared with the two finalcontrol values of 31 mmoles per L.

DISCUSSION

Most authorities assume that endogenous acidproduction results largely from the oxidation oforganic sulfur- and phosphorus-containing com-

pounds to inorganic sulfate and phosphate (11,12), but until the recent study by Hunt (5) therehad been no attempts at quantitative verificationof any part of this hypothesis. Using various

types of inattiral dliets, [htint found tiat ilet excrc-tioii of acid could be mainly accounted for byinorganic sulfate but observed no correlation be-tween acid excretion and dietary phosphorus. Hetherefore suggested that endogenous acids arosealmost entirely from the oxidation of sulfur; how-ever, in several of his experiments sulfate wassignificantly less than net acid excretion. Thisdiscrepancy would have been even greater had thedetermination of "titratable acidity" been carriedto the usual endpoint of pH 7.4 instead of to pH7.0. A significant fraction of acid productiontherefore remained to be identified.

Recent experiments in this laboratory (6) haveprovided more direct quantitative evidence for therelationship of the oxidation of organic sulfur tofixed acid production. The results differed fromthose of Hunt in revealing the excretion of sulfatein the steady state to be significantly greater thanthe net excretion of acid. The explanation prob-ably resided in the fact that the natural soybeanflour used in those initial experiments containedpotential alkali in the form of potassium, calciumand magnesium salts of the carboxyl groups in theprotein. Since virtually all natural foods can beexpected to contain variable and probably indeter-minate quantities of "hidden alkali" in the form ofsuch salts of metabolizable anions, it was decidedto restudy the problem in the experiments re-ported here, with a synthetic diet and a proteinsource essentially free of fixed cation.

The present results indicate that less than halfof net acid excretion in the steady state was dueto oxidation of sulfur. The discrepancy could bealmost quantitatively accounted for by the totalorganic acids in the urine, plus the calculated acidderived from the protein phosphorus in the diet.This seems to imply that under these experimentalconditions the only significant sources of endog-enous fixed acid were the protons released during:a) the oxidation of organic sulfur to sulfate (6),b) the generation of nonmetabolized organic acids,and c) the combustion of the disposable cationneutralizing phosphate residues in the dietaryprotein.

In the classical descriptions of the "acid-base"balance of the urine, organic acid anions were con-sidered to represent "acid," but only in the sensethat all anions were defined as "acids" and allcations as "bases" (13, 14).- Such an approach

162 /

fR 'AAWM

ARNOLDS. RELMAN, EDWARDJ. LENNONAND JACOB LEMANN, JR.

does not permit calculation of the actual net rateof fixed acid production and cannot logically beapplied to a description of the acid balance of thebody (3, 4, 15). It is well recognized that incertain abnormal states, such as diabetic ketosis,endogenous production of organic acids consti-tutes a major source of the acid load. In thepresent experiments, however, the special signifi-cance of urinary organic acids as one of the majorsources of normal endogenous release of protonsis clearly demonstrated. It seems reasonable thatthey should have this significance, for the organicacid anions normally appearing in the urinemust represent acids produced during metabolism,which escape complete oxidation to CO2 and H20and which therefore must have donated theirprotons to body buffers.

It should be noted that the relatively small frac-tion of excreted organic acids which appears inthe urine as free acid (i.e., that fraction which istitrated between the pH of urine and the pH ofplasma) does not contribute its protons to thebody, and properly should be subtracted from thedetermination of "organic acids" in the calcula-tion of acid production; but, since this free acidfraction is also measured in the determination oftitratable acidity, net acid excretion by the kidneyis overestimated by the same amount, and thus thecalculation of net proton balance is unaffected.

The Van Slyke and Palmer titration is not aprecise method and certainly does not measureorganic acids alone (8). It includes not onlycreatinine (which must be subtracted from thetotal measured value) but also small quantities ofamino acids and sulfate. The apparent pK forthe carboxyl groups of the various amino acidsranges from approximately 1.8 to 2.4. Assumingnormal urinary excretion of amino acids (up to15 mmoles), partial titration of these carboxylgroups might overestimate organic acids by asmuch as 5 mEqper day. Since the apparent pK,for sulfuric acid is approximately 1.7, some 10 percent of the urinary sulfate would also be titratedbetween 2.7 and 7.4, and in the present studiesthis would tend to overestimate organic acids byanother 2 to 4 mEqper day. On the other hand,it has been reported (and we have confirmed)that nearly half of the urinary citrate may be lostduring the "organic acid" titration, as a result ofprecipitation with calcium (16). During two of

the present studies urinary citrate was found torange from 5 to 8 mEqper day, and it can there-fore be estimated that the Van Slyke and Palmertitration probably underestimated citrate by about2 to 4 mEqper day.

Considered together, all of these small errors inboth directions tend approximately to cancel eachother out. Although there are probably othersources of minor error in the method which havenot been discussed, it appears that under the pres-ent conditions the Van Slyke and Palmer titrationprovides an approximate but reasonably satisfac-tory estimate of the contribution of organic acidsto the generation of fixed acid. It is worth notingin this context that when the diet consists ofnatural foods containing organic acids the rela-tionship of these compounds to net acid produc-tion may become considerably more complex thanthat described here. Neutral salts of metaboliz-able acids would cause removal of hydrogen ionsfrom the body and thus would reduce net acidproduction without affecting the organic acid out-put in the urine. On the other hand, salts ofnoncombustible acid (e.g., benzoic or tartaric)would not affect acid production although theymight well increase the total measured excretionof organic acids in the urine. Such considerationsemphasize the importance of the synthetic diet inproviding the conditions necessary for the accu-rate determination of net endogenous acid pro-duction.

In the present studies the presence of phosphateresidues in the dietary protein appeared to beresponsible for the third component of endogenousacid production. Previous opinion on the role ofphosphorus in acid production has varied. Mostof the earlier authorities (1, 2, 11-14) had as-sumed that all dietary phosphorus contributed toacid production. More recently, however, Hunt(5), Christensen (4, 15) and others have pointedout that, since most phosphorus in natural foodsexists already oxidized as the potassium, mag-nesium, calcium or sodium salts of organic phos-phoric acid monoesters, hydrolysis of such esterswould result in the formation of essentially neutralinorganic phosphate with no liberation of protons.3

3 Since organic phosphate monoesters in proteins prob-ably have slightly lower pK'2 values than inorganic phos-phate (17) the cleavage of such neutral esters wouldprobably have a slightly alkaline effect.

1628

ENDOGENOUSACID PRODUCTIONAND NET ACID BALANCE

The metabolism of phosphodiesters might be ex-

pected, on the other hand, to yield phosphoricacid. However, in most diets diester linkageswould occur relatively infrequently and would not

often make a significant contribution to acid pro-

duction. In support of such reasoning, Hunt re-

ported that he was unable to observe any constant

relationship between dietary phosphorus and uri-nary excretion of acid.

In the experiments reported here, the onlysource of dietary organic phosphorus was thepurified soy phosphoprotein which was virtuallyfree of inorganic electrolytes. In contrast to thesituation prevailing in most natural foods, thephosphate residues in the soyprotein used in eightof the ten balances were neutralized by combus-tible cations in the form of arginine (R-NH3+) or

ammonium. Such cations, when metabolized, ul-timately must yield protons (see discussion inMethods). In calculating the acid contributionfrom this source, it was assumed that metabolismand hydrolysis of dietary protein was completeand that the liberated phosphate residues were

titrated to a pH of 7.4. That such assumptionswere at least approximately true is indicated bythe reasonably close agreement of acid excretionwith the total acid production calculated in thismanner (Table II).

The observations made during the course of thefour relatively prolonged balance studies meritsome final comment. In the bicarbonate experi-ment, the prompt fall in renal acid excretion tovery close to zero levels (Figure 1) providesstrong support for the idea that Subject GdeFwas in fact in acid-base equilibrium prior to thebicarbonate load. In the two phosphate experi-ments a similar conclusion may be drawn from thefact that relatively small increments in net acidexcretion were promptly followed by rises in

serum CO2 content. The fact that adjustmentsin renal acid excretion following the bicarbonateor neutral phosphate load resulted in a final cum-

ulative equality between calculated acid produc-tion and acid excretion in the urine is also of greatinterest, for it tends to indicate that administeredor endogenous acid or alkali is not ordinarilyexcreted in significant quantities by any extra-

renal route, such as the stool. A similar conclu-sion follows from the fact that essentially all of the

rextra endogenous acid derived from methioninlccould be recovered in the urine (6).

In all of the prolonged balance studies, the ob-served relationships between acid balance and theserum CO, content were entirely consistent withwhat would be expected from the interplay of twophysiological variables such as acid productionand excretion. Thus, whenever the net excretionof acid exceeded the calculated rate of production,serum CO, content tended to rise and, conversely,whenever production exceeded excretion, the CO2content fell.

In an earlier study of normal subjects given anacute acid load (18), it was demonstrated thatcalculation of acute changes in net acid balancewas feasible, assuming a constant, though un-known, rate of endogenous acid production. Thepresent results appear to offer a technique formeasuring the actual net balance of protons andtherefore may permit study of the absolute gainor loss of hydrogen ions under a variety of experi-mental and clinical situations. The method re-quires, however, that the subjects be in approx-imate nitrogen equilibrium, because catabolism oftissue protein would present difficulties in calcu-lation of acid production similar to those alreadydescribed with the use of natural dietary protein.

SUMMARYAND CONCLUSIONS

The purpose of this study was to determinewhether it was possible to measure the totalendogenous production of fixed acid and, by com-paring this independently determined quantitywith the net renal excretion of acid, to measurethe net external acid balance of the body.

In a series of ten balance experiments, normalsubjects in an apparently steady state were fed aliquid-formula diet which had as the nitrogensource a purified soy phosphoprotein essentiallyfree of mineral cations. It was demonstrated thatthe net renal excretion of acid was closely matchedby the sum of: a) inorganic sulfate in the urine,b) total "organic acids" in the urine, and c) theacid calculated to be released by the metabolism ofthe combustible cationic groups covering the phos-phate residues in the dietary phosphoprotein. Inthree experiments net production or excretion ofacid was temporarily altered and, at the end, whena steady state had been restored, the equivalence

1629)

ARNOLDS. RELMAN, EDWARDJ. LENNONAND JACOB LEMANN, JR.

of cuIllll11tesl acicl excretion with the cumulatedplroduction of acid could again le denlonstrate(d.

These observations are taken to mean that,under the present conditions, the total net produc-tion of fixed acid arose from the oxidation oforganic sulfur, plus the production of nonmetab-olized organic acids from neutral precursors, andthe release of protons during the metabolism ofthe organic cations neutralizing phosphate estersor during the hydrolysis of phosphate diesters inthe dietary phosphoprotein, or both. However,when noncombustible cations neutralize the phos-phate residues, no acid production can be at-tributed to the presence of phosphate monoestersin the protein.

The relationship between acid production meas-ured in this way and the renal excretion of acidappears to be that of a true physiological equilib-rium. Provided that body composition remainsrelatively constant during the period of study, itshould be possible to apply this technique to thedirect measurement of the external balance ofacid under a variety of experimental and clinicalconditions.

REFERENCES

1. Shohl, A. T. Mineral Metabolism. New York,Reinhold, 1939, pp. 293-295.

2. Sherman, H. C. Chemistry of Food and Nutrition,7th ed. New York, Macmillan, 1946, pp. 248-253.

3. Relman, A. S. What are "acids" and "bases"?Amer. J. Med. 1954, 17, 435.

4. Christensen, H. N. Diagnostic Biochemistry; Quan-titative Distributions of Body Constituents andTheir Physiological Interpretation. New York,Oxford Univ. Press, 1959, chap. 10.

5. Hunt, J. N. The influence of dietary sulphur on theurinary output of acid in man. Clin. Sci. 1956, 15,119.

6. Lemann, J., Jr., and Relman, A. S. Tlhe relation ofsulfur lcetab)olism to acid-base balance and electro-lyte excretion: The effects of DL-methionine innormal man. J. clin. Invest. 1959, 38, 2215.

7. Relman, A. S., and Schwartz, W. B. The effect ofDOCAon electrolyte balance in normal man andits relation to sodium chloride intake. Yale J. Biol.Med. 1952, 24, 540.

8. Van Slyke, D. D., and Palmer, W. W. Studies ofacidosis. XVI. The titration of organic acids inurine. J. biol. Chem. 1920, 41, 567.

9. Clark, E. P., and Collip, J. B. A study of the Tisdallmethod for the determination of blood serum cal-cium with a suggested modification. J. biol. Chem.1925, 63, 461.

10. Ettinger, R. H., Goldbaum, L. R., and Smith, L. H.,Jr. A simplified photometric method for the deter-mination of citrate in biological fluids. J. biol.Chem. 1952, 199, 531.

11. Peters, J. P., and Van Slyke, D. D. QuantitativeClinical Chemistry. Vol. I, Interpretations. Balti-more, Williams & Wilkins, 1931, p. 964.

12. Smith, H. W. Principles of Renal Physiology. NewYork, Oxford Univ. Press, 1956, p. 134.

13. Gamble, J. L., Ross, G. S., and Tisdall, F. F. Themetabolism of fixed base during fasting. J. biol.Chem. 1923, 57, 633.

14. Albright, F., and Reifenstein, E. C., Jr. The Para-thyroid Glands and Metabolic Bone Disease. Bal-timore, Williams & Wilkins, 1948, pp. 241-247.

15. Christensen, H. N. Anions versus cations? Amer.J. Med. 1957, 23, 163.

16. Sherman, C. C., Mendel, L. B., and Smith, A. H.The metabolism of orally administered citric acid.J. biol. Chem. 1936, 113, 265.

17. Folsch, G., and Osterberg, R. The apparent acidionization constants of some O-phosphorylatedpeptides and related compounds. J. biol. Chem.1959, 234, 2298.

18. Schwartz, W. B., Jenson, R. L., and Relman, A. S.The disposition of acid administered to sodium-depleted subjects: The renal response and the roleof the whole body buffers. J. clin. Invest. 1954, 33,587.

1630