alteration - dm5migu4zj3pb.cloudfront.netdm5migu4zj3pb.cloudfront.net/manuscripts/102000/102529/JCI51102529.pdf · Balance data were calculated as net intake, minus com-bined output

ELECTROLYTEEQUILIBRIUM DURINGMERCURIALDIURESIS 1, 2

By WILLIAM B. SCHWARTZ3 AND WILLIAM M. WALLACE

(From the Medical Serzices of the Peter Bent Brigham Hospital and the Children's MedicalCenter and the Departments of Medicine and Pediatrics, Harvard Medical School,

Boston, Mass.)

(Submitted for publication May 23, 1951; accepted July 30, 1951)

INTRODUCTION

Earlier workers have shown that the amountof fixed cation contained in the extra urine voidedduring the course of mercurial diuresis corre-sponded to that contained in an equivalent amountof body fluid, while the quantity of chloride wasgreater than that which would be expected to befound in the fluid lost. This differential loss ofchloride was shown to result in hypochloremic al-kalosis. It was further observed that the adminis-tration of acidifying salts prevented the occurrenceof alkalosis and at the same time enhanced diuresis(1,2).

The observations reported in this communicationindicate that the alteration in the concentration ofchloride and bicarbonate in the serum that followsrepeated diureses with a mercurial is usually ac-companied by unresponsiveness to further adminis-tration of the drug. The various patterns of elec-trolyte loss observed during mercurial diuresis,their significance in relation to body composition,and their relation to certain clinical and physiologicproblems will be discussed.

METHODSAND CALCULATIONS

Ten patients in congestive heart failure as the resultof rheumatic or hypertensive heart disease were studiedon the metabolism ward of the Peter Bent Brigham Hos-pital. All were chosen on the basis of their willingnessand ability to cooperate in the collection of specimensand the consumption of a constant diet. Six were males

1 This investigation was supported by research grantsfrom the National Heart Institute of the National Insti-tutes of Health, U. S. Public Health Service, the GrantFoundation, Mead Johnson and Company, and LakesideLaboratories, Inc.

2 Presented in abstract form at the 42nd Annual Meet-ing of the American Society for Clinical Investigation,Atlantic City, N. J., May 1, 1950.

8 Present address: New England Center Hospital andTufts College Medical School, Boston, Mass.

and four were females. Control periods were obtainedwhen the clinical situation allowed. In most instances,however, it was deemed necessary to begin therapy atonce. Each patient was placed on a constant weigheddiet that contained minimal sodium and approximatedthe usual ward diet with regard to nitrogen and caloricvalue. Every third day a duplicate diet was analyzed forsodium, chloride, potassium, and nitrogen. All food refusedwas saved for analysis and the intake figures correctedaccordingly. Water intakes were kept as constant aspossible.

The mercurial compound (Mercuhydrin) was adminis-tered intramuscularly in a standard dose of 2 cc. at thebeginning of each day of treatment.

The urine was collected for 24-hour periods, with thy-mol in chloroform used as preservative. The feces werecollected from eight of the ten patients and were pooledin four to six-day periods. Carmine was used to markthe beginning and end of each period. The urine andstools were analyzed for sodium, chloride, potassium, andnitrogen. In addition, titratable acidity, pH, ammonia,and total phosphorus were determined in the urine.

Venous blood was drawn without stasis in oiled syr-inges at the beginning of each daily collection period whilethe patient was fasting. At the time of blood samplingthe patient was weighed to the nearest 50 grams.

The methods used for determination of sodium, potas-sium, chloride, nitrogen, pH, and carbon dioxide have beendescribed in a previous paper from this laboratory (3).Urine titratable acidity was determined by the methodof Henderson and Palmer (4) and ammonia was deter-mined by the method of Folin (5).

Balance data were calculated as net intake, minus com-bined output in urine and stool. The quantity of sub-stance in the pooled stool collections was divided by thenumber of days in the collection period to obtain theaverage figure for the daily stool output. No unusuallylarge stools were passed by any of the subjects reportedin this study and daily losses were assumed to be equal.Changes in the volume of the chloride space and shiftsof sodium into and out of the intracellular fluid werecalculated by methods previously described and used byvarious authors, (6-8). The form of the calculation andthe factors used have been detailed in a previous paperfrom this laboratory (3). In calculation of the presentdata the changes in chloride space were calculated back-ward from the balance of chloride using the assumptionthat chloride was distributed through 20% of the bodyweight at the end of the period of diuresis.

1089

WILLIAM B. SCHWARTZAND WILLIAM M. WALLACE

RESULTS

The ten patients with congestive heart failurewere observed for a combined total of 22 periodsduring which Mercuhydrin was administered. Theessential features with regard to the ratio of sodiumto chloride (Na/Cl ratio) of the negative balance,the potassium loss or absence thereof, and thedevelopment of unresponsiveness to the diureticare summarized in Table I. A patient was con-

sidered to be unresponsive if he was still edematousand his weight did not change after administrationof the diuretic. The data are too few to allow a

categorical statement about the relationship ofthe character of the balance losses to the develop-ment of unresponsiveness but a certain groupingof patients appears possible. A review of therapyemployed prior to the period of study with respectto diuretics, digitalis and salt restriction showedno apparent correlation with the urine Na/Cl ratioduring diuresis. Two of the patients (V. S., B. C.)responded to the diuretic by the loss of body fluidthe composition of which was similar to that ofextracellular fluid. In these, the Na/Cl ratio of thenegative balance was above unity; neither alkalosisnor negative potassium balance developed and thepatients were continually responsive until approxi-mately normal weight was reached. In eight ofthe patients the Na/Cl ratio of the negative balancewas less than 1. In these subjects there was a fallin serum chloride and a rise in serum bicarbonateconcentrations, often accompanied by a negativebalance of potassium and a reduction in serum po-tassium concentration. In all but one (S. S.) of

these patients unresponsiveness appeared. Theadministration of ammonium chloride in amountssufficient to return the serum chloride and bicarbo-nate to normal in the refractory alkalotic patientsalso restored responsiveness in every instance.

Analytical data from four representative sub-jects (V. S., F. C., N. B., and S. S.) are presentedin Tables II-V and Figures 1-6. Figures 1-4

show the daily values for the serum electrolytes,the course of the cumulative balance of electrolytes,the cumulative course of the change of body weight,and an estimate of the shifts of sodium and potas-sium between extracellular and intracellular fluid.The cumulative nitrogen balance is plotted as grams

times 2.7 to indicate the degree of deviation of po-

tassium and nitrogen from the theoretical lossesthat would occur if body tissue as a whole were

being lost or gained (9). The ordinate indicatingthe cumulative weight loss is so chosen that one kg.equals a value on the mEq. ordinate equivalent tothe mean calculated value for extracellular fluid so-

dium during the period of observation.In the theoretical situation in which only extra-

cellular fluid is being lost these two cumulativebalances should coincide. Deviation of the twocurves from each other indicates unmeasured lossesthrough the skin, losses of intracellular fluid, or

shifts of sodium into or out of the intracellularcompartment. Minimal sweating occurred duringthese experiments so that losses of sodium throughthe skin were probably about 5 to 10 mEq./day(10, 11).

The calculated shifts of sodium between the ex-

TABLE I

Interrelationships of Na/CI ratio of balance to K loss and unresponsiveness

NaNa of balance K loss

Subject Diuretic period Diuretic period Remarks

1 2 3 1 2 3

V. S. 1.21 1.33 1.14 0 0 0 Continually responsiveB. C. 1.39 0 Continually resonsiveF. C. .73 .51 .53 + + + Refractory at the end of each periodN. B. .66 .50 + + Refractory at the end of each periodS. S. .44 .57 + 0 Continually responsiveJ. C. .29 .78 .56 + 0 0 Refractory at the end of each periodC. A .62 .90 0 0 Refractory at the end of each periodS. H. .94 .57 0 0 Refractory at the end of each periodJ. B. .60 .61 + + Refractory at the end of each periodC. G. .91 .68 0 0 Refractory at the end of each period

1090

ELECTROLYTEEQUILIBRIUM DURING MERCURIALDIURESIS

0 0 Go co'-4 N if

) qU)*. UW U)* o.U U) U) 00 00ooo00%in in N i co % 0o 0%0U)O

$ ~ ~ ~~ m m m vmOm "f fimimOOfi 0 o

%Z b 0 %00W '0'0U U N U) 0 0 '0 00 0'0'I i if I nf) i)ne .f)

6 ~ ~ ~ U)I*0*U) f N6Ilq q q 6 I.U) q i000 0if) m m m ' mme '0 U) mm

S4~~~ ~ ~ ~ ~ ~~~~~~f ifif i)f) fif i)

N- 0- 1 t-.t &- bz 0-.

0 '0 0%0C 0 N 0 N Nin~ 4' ifi if 4'* if4' 4'*1N- N. 0-NN:N. 0-NN.lz

I-. N^ N NNNr N N N 0% 0%%000%o 00 00000000

~~~ 4'~~~1 4' 00 00 00 0000 00 '0 '40'10'10'10'

N N NN N NNNN N

'0 0-. 0- I~'0N N- N- N- N- NO -%'0 ' '0 '0 W)U) W) U) U)

N -' OU N ' 0 - '0 0N000 0 4 4''4'

0Z0 0 0 0 0 0 t0 W) N 0 "0"ifN4'00

WNW cN 0%4') 0% 00 i) '0 4 *4'"u f 0~''..%UC4 if(4 N 4~ N if i) N NNN N 0NN

x ~U 0 00N0 +-"0 0% 0 0 if) e0%i) 0 ONNN

U)_ '4 ' N04 0 N if UU N N4'004'00U)

)0-i) 0%G ,R Noif)nNN

U) if) N_I No m 0

if) 4N48) 8N I-if) k-W0"0O'

I'mm!$- OO-W eetO N N N if if NNN4 U) '*b_w 0 0 if)U)ifg

-~~~~~~~k W)-O

*EX tX W o -o bO~~~t-W) - t- fi-o t- t-t tt-"t-o t- t.al t b t

U) U ' N N N "'0U)N N 0% _

~~~ ~~~~~8 8 8 0 888C4 8 8 88888f

inW) In ) 0 C) C) C0_

,qS-e q -- -I.N -t n q N|

8

-O Net0 0% co co co 00 o 0 co C__^_

"0if if) if if if '_N_N N N _-

* Bi YN N " N ) NNW N N NNSN N NN N N N N N

- u NU) U) U) 0UQU00U00uZvZuvwU )U)U) ooooo-' '0E N_N fj'0 N0'0'0'00%000 N N NoNoN

m000r 30~o 00o0 000 ;

o :g,{88i '0'0'0Sii8

. 1 UNXSNXNjNN gGNg {NNN ENlN N

.~~~~~~~~- N ) 4U)' N 00 0% 0 - NU" 4U) '0 Nt 00%0-- 4N

__4 _____4_4 NNNN

0mO

r 6._

;' Z

Il

1091

0)

co

C-)

U)

-0

0


0

I ~ ~ U4ou)o.Co'ooo i U) 0% 0 W-

Z 0 q Uui 00y"0* a 01 I)uot aq q -I -I tt qq q a Go Moo Co 'n C)

C(4 Ns-'0' 0- ('0%U '06d0% 0%U)6V~ ('I('4 if) if) if) )U) 4if)NS U) U 4MNN N N N N

'00% N N-(' 4 0 U) - 0%

2- t- t- t- t-e

t. tp thet- 1- t- kl.V

e4b 4b 4b b O r 4b

V V co t.Ndrz 066N 4 ot oc o c

o £Ut M OoO00 U) U) U)U))'-- -_ ))UU)e U) U

. -, Cs _ _q _4 __o _ _N__Oj= %('4C '4 No .0. %O (;I lot0 N ) U)o. 0A ad adU .'0NoU)U)~' N 0%C)wONGo U) U) U)

z4 U4O ' 0 0 0 ' U0oo f 0%U N ) -

b-4 *.4) 4U) ('4 - - 4(1U)>4O U)NO

~~~~44 U) ~U) 0% GON0GO(ao (4t- .44

M NO

U) U) c -S4- - 0

*.4U) oo '4(1'm U~~4 U~)N '

N_'No0 _ _ S__ G N _

> N

co_- -co -ea ooa -c'4o a -

O')%U)U)- - U) iU)'0 4 n in in4 0nIn in

z e44)4 - U) U 4'4U)) C4( 4 C4N N N

U

4- N NN N (4 ('4 (4 -4

-4 W) U) U) 4U)4Ni U) U) "W) __4 in '0_W4 N U) GO'(4U0 4UU (40%S('4Wg g

0o * ot 8 °2 +N00 °- 8 8Kwe _ ao e o o -U)0%'0'_ U)o('4V4) (4 (4 4 OTT 4 ('4 (4 .- 4 ('4 ('

4;)U ) U) U ) IL)U)UU) U)U U)U))U UUL)U)

o,e--~~~~ - -) - __ _

3 4M N0 0 000000 00 000 00 0

o 88 8 8 88888 8 8 w°88 8z~~~U 00 U)) U) 0U)) V>>>>aV WouvUX)U

. Nt-0 U) U) U) Ue+)U)U (1 (' ('4'U 4 ( - -m)U U) U) U)UUUUU ) U) U)U U) U) U) U)

*.*.*. *.*.*. *. *.*.

I :oIc 8eX g EdE,E

4

1092

co

b~p 00

*D z

C;

04.) Cdu

co

'0

co

~U

':

* _

II


'o

I ,q ,V qo uto qo uo uu4 i

I- O )0% C%ok 0ooo oo

I O i) t U U, ifU* eq eq in u

mW)W)$ %s %

if) 1N0° v'IOf 0

_if)100 )0 ) )0 000 ~~0d ~ ~ ~ ~ ' N U)o.Ow '0.w .oo .o

UD if if)e,e, N NNS N if if if

- o -4 M co - $o0. I- -

N . N. N. N N N. N

I *0 am C i C it- r_ tN -N

000'e''-'~~~t00 000

U -

04 Nr~OW) '0' 0' -' 4 "'O 0'tQ~~ ~~~ MN&1§WoWoo W

I; $

z I 't.o. if N Oe if o. ' NoU) N N

> | CX, $ + ff N

, NO NO if)NONS NM V)VON 0 ON ON

z t N N" )N N - _

B _ 0 o NWv) iI W U W) WU)W)0W 0 00

co 1 s+ eo - -e0 +

I'% if) '0 if)00 if) NOenNs0

z IAs NbW*

. W In o ) oo oin i If )o o In

eWuC1NN UN gni) No e'4% I 3 m

N U S N_ if0__N

_ NN 00----E 2 2o oU) oU, U,U, U, if U, U,,U U, U, U,o

0 0 000 0 U, U, U, 0 U, U, 0 0 0

* *.*. *, .~ .~ .;5 aN{ "a E zE e X

_ @@_NNN@@

X ZNZNZ_U_U_ >NXN>NXN e 0 - e,P4"MW) %O - W Q~ - N M " in

-

0

.~ 0%

-e11

'Q =.

1093

4;

U)

0

::

0

0

Cd

CO)0

,0

o

4)

*

u* -

04-

* 4

I.

ig

j


0% eq N " _ 0% " _0N ) 4 ) eq MM")

In 0 0% ) '0 t )a o o

,. U e el0. Un U) e 'q coeq

og oo) N U)0 o 0M ~ ~~t coZ fi w 0o 0 Cso

_) U U U) () (4 N q

Y ~ ~ eI) N W

NNN N NU) t N t- N N QNNN

t- tz C. Z.: I-Z R-Z t- I- IZt t v: :t:v

I$ ~ ~ ~ ~ 0 eq eqN qN N

- - -I-$i|*to '0oo N N N --

so C-4 ein) coU)) U)a ) U) '0t ) U) U) '0co 0

I nw 0 0o eo Oq N

t It4ute WWooIn C4 C4

Pg~U)'0U) '0 U) N '0 N U) U) __

t $>eS °~42. o m000 u u

' -~~0%q0% 0 0) U)W -- U)

.0 > Z -- eq e fi _-

N'04N- U) U) U)

(jeqU)4 N0N 0 U) -° U) N N NU)

- U 0% -40 -_ U)U)_-It

- eqUeI N 0.o '0 b ' U

tnWeU)) U)C - U)UC,;1 4 eq14t

4 4 N - '0 U) U)

t) U) U) - U N

eqW N) U) 0%InW) U)'0 ) 4' U) U-)

21e^°-_q -2 - 2

C4 U) N N*N N NC0

m400 ' 0 0 U) U) U) U

.I

X; i--- - - - --i

Ia

1094

If)U)

a gb o)

*e cn5*2 0

e~6,

._

04

._

._0

0

a)

4-

._

Ca

-_

CI)

C:U)

'0

._O

4- 4-'

._

'0

co

C4

U4a)

00

* .

s

i

" NM V W) %O 1-- 0 Q. 0 - NM.4 W" W" 1-4


tra- and intracellular compartments, as well asthe loss or gain of intracellular potassium, are in-dicated at the bottom of the chart. These intra-cellular balances are cumulative only for the pe-riod indicated. For example, in Figure 3, theshifts during the first four days are indicated at the

end of the fourth day; the shift on day 11 iscumulative for days five through 11, and not forthe entire 11-day period.

Graphic presentation of the intracellular shiftshas been made in terms of the percentage of esti-mated total extracellular fluid sodium on the day

FIG. 1. SERUMELECTROLYTECONCENTRATIONS,EXTERNALAND INTERNAL BAL-ANCES IN RELATION To THERAPYFOR PATIENT V. S. (TABLE II) SHOWINGCON-TINUED RESPONSETO THE DIuRETIc WITHOUT PROGRESSIVECHANGESIN SERUMCONCENTRATIONS

In this and the succeeding three charts the coordinate indicating weight loss ischosen to predict the theoretical sodium loss as calculated from mean extracellularfluid sodium concentration (see text).

1095


SERUMINRUb PT. F,C. PBBH 8A48MEO/L.-

N. Cl CO)N

130 N0025X

K C7-90 i5"6 - KCONTENT Ka480

_70_

DAYF I F. C. 9(TABL III) SN2131T4H1R51617 18

120i~~~~~~~~~~~~~1

IN 200 - O

FIG.LRI2. ISER MIELECTROYECNNTATION SE-

ThoeRnA AND ItheRA BAlclaNCES Ind RheLTOpretOgTHER

chosng forete calculationvandeithe prpercpentageof

tive with the total amounts involved. Total bodypotassium was estimated from the body weight atthe end of the period of observation, based on thevalue 68 mEq./kg. given by Shohl (12).

The necessity for presentation in this way isobvious when the method of calculation is ex-amined. It is evident that analytical errors inthe determination of serum sodium and chlorideconcentrations can produce large changes in theapparent quantity of sodium shifting between cellsand extracellular fluid. In the present study theanalyses for serum sodium and chloride are as-

sumed to carry a possible error of + 2%, the tech-nique of the balance, an error of + 5%o, and the de-termination of weight, an error of + 1 %. Whenthe analytical values are applied in the equationsused, calculation of the summation of these pos-sible errors 3 indicates that at least 6% of the ex-tracellular sodium must shift into or out of cellsbefore the shifts can have even statistical sig-nificance.

3 The equations used in calculation of the summation oferrors were:

1. (A d a) - (B i b) = A + B d Nra2 + b2,2. (A 4 a) X (B ± b) = AB i 4(AB)' + (Ba)2,3. (A i a)(B ± b)(C 1 c)

= ABC 4(aBC)2 + (bAC)2 + (cAB)2,B+ b B /( Ba )2A a~A A\ AJ 2

FIG. 3. SERUM ELECTROLYTE CONCENTRATIONS, Ex-TERNAL AND INTERNAL BALANCESIN RELATION TO THER-APY FOR PATIENT N. B. (TABLE IV) SHOWINGA PAT-TERN SIMILAR TO THAT FOR PATIENT F. C.

1096


The data on a massively edematous patient (Vr.S.), who lost 33 kg. of weight during a 22-day pe-riod of observation, are presented in Table II.Figure 1 is a graphic presentation of part of thesedata. Serum analyses showed an elevation of se-

rum bicarbonate concentration and a reduction inchloride concentration prior to diuresis. On thefirst day of diuresis an additional elevation of thebicarbonate concentration occurred. After thisinitial rise, however, patient V. S. continued todiurese without progression of the serum changes.These variations from normal concentrations were

not seen in the other patient (B. C.), whose diu-resis was of the same type as that described inFigure 1.

The diuretic was administered at intervalsthroughout the period of study. At each inter-ruption of treatment with mercury, loss of electro-lyte and of body weight quickly stopped. Resump-tion of treatment resulted in renewed diuresis. Atno time was the patient unresponsive. The nega-tive balance of sodium was always greater thanthat of chloride and the ratio of the losses was ap-proximately the same as the ratio of sodium tochloride in extracellular fluid. The course of thecumulative nitrogen and potassium balance indi-cated that intracellular fluid remained undisturbed.The curves for weight loss and sodium loss nearlycoincided, a further indication of the extracellularnature of the fluid loss. The slight deviation ofthe two curves, as well as the indicated shift of so-

dium into cells, could well be accounted for by un-

measured losses. The pH of the urine remainedbetween 6.4 and 7.7 during the course of the diu-resis. No significant changes occurred in the dailytotal excretion of titratable acid or ammonia.

The data on two patients (F. C., N. B.), whoat first responded to the administration of the diu-retic but quickly became unresponsive, are given inTables III and IV and are graphically shown inFigures 2 and 3. In both of these patients the nega-

tive balance of chloride exceeded the negative bal-ance of sodium. Hypochloremic alkalosis devel-oped without change in serum sodium concentra-tion. Unlike the subject (V. S.) previously de-scribed, a negative balance of potassium occurredin the initial phase of diuresis and the concentrationof serum potassium decreased on the initial day.As serum chloride concentration fell and bicarbo-

MEQ/L. PT. S.S. PBBH L5455Na Cl CO

150 120 4'A'Nal130 11O 35

AY _

t-O

I90 1

NH4C1 300

MEO/DAY 1001

I0

CELL .9 Na

CONEN 4o0. Na3.

FIG. 4. SERUM ELECTROLYTE CONCENTRATIONS, EX-TERNAL AND INTERNAL BALANCESIN RELATION TO THER-APY FOR PATIENT S. S. (TABLE V) SHOWINGCONTINUEDRESPONSIVENESSTO THERAPYDESPITE DEVELOPMENTOFHYPOCHLOREMICALKALOSIS

nate concentration rose both subjects became re-fractory to further administration of the diuretic.

The administration of ammonium chloride orhydrochloric acid returned the serum concentra-tions of bicarbonate and chloride to normal.Both subjects once again responded to treatmentwith mercury but further diuresis resulted in repe-tition of essentially the same cycle.

The potassium loss noted in the two patients

1097


ANION - CATION PARTITION OF URINE. Pt. $4

URINEVOL.ML.

MERCUHYDRIN

URINEpH

x x x x x x x

MEQ./DA.

1 2 3 4 5 6 7 8 9 10 II 12DAY

FIG. 5. DAILY URINE VOLUME, URINE PH AND THE ANION-CATION PARTITION OF THEURINE DURINGTwo DIuRETIC PERIODS FOR S. S. (TABLE V)

Note that chloride excretion is always greater than the sum of excretions of sodiumand potassium.

described in Figures 2 and 3 did not invariablyaccompany the development of alkalosis. In threesubjects (shown in Table I) alkalosis occurredwithout any loss of potassium. It will be noted,however, that during all four diuretic periods de-scribed for F. C. and N. B., potassium was lost inexcess of nitrogen and in excess of the smallamounts that would be expected from the simpleloss of extracellular fluid. The physiologic signifi-cance of this loss is difficult to evaluate but may beassessed in terms of the percentage loss of esti-mated total body potassium. Patient N. B., at theend of the period of study, had lost approximately6% of her estimated total body potassium. In herthe potassium losses were repaired from a fixedintake during the period of chloride administra-tion. In F. C., however, the potassium losses dur-ing each period of diuresis were additive and thenet loss at the end of the experiment was more than

10%o of the estimated total body potassium, yetthere were no clinical or electrocardiographic evi-dences of potassium deficiency and the serum po-tassium level was within normal limits.

Examination of the relation of the actual weightloss to theoretical weight loss calculated from so-dium showed that intracellular fluid as well asextracellular fluid was lost. During each periodof diuresis in both subjects the weight loss calcu-lated from sodium was seen to be less than theobserved weight loss, the difference roughly ap-proximating the amount that might be calculatedfrom the loss of potassium associated with the lossof intracellular water. The difficulties that arisefrom attempts to predict shifts in water by calcu-lation from observed body weights and balancemeasurements for electrolytes have been stressedby Yannet and Darrow (13) and Elkinton and co-workers (8). For this reason no attempt was

1098


made to quantitatively assess the relationship ofpotassium loss to intracellular fluid loss. The in-consistencies that would be encountered are par-ticularly evident in Figure 2, where it is seen thatduring the periods when hydrochloric acid was ad-ministered, increments in body weight occurredwithout simultaneous retention of either sodiumor potassium. It may be postulated that changesof this nature are related to an "activation" ofcellular cation, but they can probably be morereasonably explained as resulting from changes indissociation of intracellular anion (phosphate) anda consequent increase in the number of osmoticallyactive particles (14). The additional possibilitythat there is a primary disturbance in the excretionof water can not be ruled out. The degree of re-tention would be expected to produce only a slightchange in tonicity, difficult to identify by measure-ments of serum sodium concentration.

The intracellular balances seen in Figures 2 and3 show that during the development of alkalosis,sodium regularly tended to shift into the cells andpotassium tended to leave them. In most in-stances the reverse occurred during administra-tion of ammonium chloride or hydrochloric acid.These shifts are barely significant numerically, butthe uniformity of their direction suggests thatsome such redistribution is taking place. Themeaning of such shifts is not clear but they may

PATIENTSS,

CHANGEABOVEOR BELOWFORE PERIOD VALUESRECORDEDADOITnIELY

a200 k'

CaY 2 3 4 5 6 7 8 9 0

MERCUKYRINX x x x x x x

FIG. 6. CUMULATI URINARY EXCRETIONS ABOVE

FORODEXCRETIONS FOR PATIENT S. S. (TABLE V)PLOMDTO INDICATE THE QUANTITIES OF POTASSIUMAND

AMMONIUMCOVERINGTHE SURPLUSOF CHLORIDE ABOVE

SODIUMDURING THE SuccEssiVE DAYS OF THE DIuRESIS

serve the purpose of limiting the degree of rela-tive acidosis or alkalosis in the extracellular fluid(3).

Figure 4 and Table V (S. S.) depict the courseof the balance study in a patient who also re-sponded with a loss of chloride in excess of so-dium but who despite progressive alkalosis did notbecome refractory. The changes in the serumconcentrations are similar to those occurring inpatients F. C. and N. B. (see Figures 2 and 3).Administration of ammonium chloride on thesixth day resulted in a partial return toward nor-mal of the serum values, as well as remission inthe symptoms of lethargy and anorexia which hadappeared as diuresis proceeded. During the initialdiuresis approximately 5%o of the estimated bodypotassium was lost in the urine. During the sec-ond diuretic period no loss of potassium occurredthat could not be accounted for in the extracellularfluid which had been lost. Alkalosis continued toincrease without loss of potassium as diuresis pro-ceeded.

Although the signs of congestive failure wererapidly diminishing as diuresis continued, the pa-tient became anorexic and lethargic once againand gradually lapsed into coma during the final twodays of observation. Several hours before death,intermittent ventricular tachycardia and shock de-veloped. The patient died on the 13th hospital daywith no evidence of congestive failure other thanmild peripheral edema. Postmortem examinationdid not provide an explanation of the death. Itseems possible that the fatal outcome may havebeen related to the severe imbalance in the com-position of the body fluid. Although 5%o of theestimated body potassium had been lost in excessof nitrogen and there was an abnormally low se-rum concentration of potassium, there were nodistinct clinical evidences of potassium deficiencyand the electrocardiogram did not show signs ofhypokalemia. Another factor contributing to theterminal shock-like state in this patient may havebeen the rapid and severe depletion of the bodyfluid volume. In the end phases of observationwater retention appeared with a progressive de-crease in the concentration of serum sodium. Al-though clinical deterioration was observed in thispatient and in one other patient with alkalosis (J.C.) several patients, not included in the presentdata, have been observed to have values for se-

1099


rum pH of from 7.55 to 7.57 and total carbon di-oxide concentrations of from 50 to 60 mEq./l.without the development of untoward symptomsand signs (15).

In Figure 5 a graphic analysis of the factors ofanion-cation excretion in the urine of subject S. S.is presented. In the upper half of the chart thedaily volumes and measurements of pH are shown.In the lower section the uppermost line records thedaily fixed cation excretion plus urine ammoniumand titratable acidity. Although the two relativelysmall components of fixed cation outgo in the urine,calcium and magnesium, are not included, sodiumplus potassium plus ammonium plus titratable acidmay be taken as approximately measuring the totalcation excretion. By reference to the top line theanion excess over sodium plus potassium which iscovered by ammonium plus titratable acid is de-fined. The relatively very small value for the un-measured anions, phosphate, sulfate and organicacids, is shown by the proximity of the chloridemeasurements to this line.

In both periods of administration of Mercuhydrinurine volume is seen to increase greatly on the firstday and then decline progressively, although aquite large diuretic effect continued to be obtained,and there was roughly corresponding change inthe outgo of chloride and sodium in the urine. Inthe first period as may be seen by reference to thevalues before diuresis (days 1 and 2) removal ofpotassium was extensively increased whereas inthe second period urine potassium was not abovethe initial value. This is explained by the largervalues for ammonium and titratable acid over thesecond period.

The relation of potassium removal, ammoniumproduction and urine acidity to the huge excess ofchloride over sodium in the urine, which is theoutstanding feature of the situation, is shown againin Figure 6. In this figure the daily values forchange above or below the foreperiod measure-ments are recorded additively. The solid line de-fines, on this basis, anion excess composed ofchloride plus phosphate over sodium. Measure-ments of sulfate and organic acids were not ob-tained. There was, as shown by the measurementsin Table V, an increase of urine phosphate dur-ing the first mercuhydrin period along with the in-creased removal of potassium. The values for po-tassium, ammonium and titratable acid are re-

corded separately. Taken together, as shown bythe uppermost broken line, they quite closely fol-low the anion excess over the sodium. The largeremoval of potassium above the foreperiod levelduring the first period of diuresis is quite clearlyreferable to the lag in ammonium production withthe result that intracellular fixed cation is requiredto complete the coverage of anion excess over so-dium. After the third day the increase of am-monium plus titratable acid permits the excretionof potassium to fall below the foreperiod value andthere is gradual recovery of the potassium lostover the first three days. It should be noted thatin some patients the discrepancy between chlorideand sodium may be made up entirely by potassiumwithout significant change in excretion of am-monium or titratable acid during the period of diu-resis. This is illustrated particularly well in thedata shown in Table IV (N. B.).

DISCUSSION

By means of direct observations on the kidneyand short-term clearance studies investigators haveidentified many of the renal effects of mercurialcompounds. The diuretic action of these drugsmay be accounted for by a direct renal action (16,17). Diuresis after single doses of mercurial com-pounds is not associated with an increase in therenal blood flow or in glomerular filtration rate butis apparently the result of a decreased tubular re-absorption of certain electrolytes (18-20). It isnot clear whether the depression of reabsorptionoccurs in the proximal or in the distal tubule.Conclusions on this point have differed, dependingon the conditions under which observations weremade (21, 22). In addition to the effects on re-absorption, mercurial compounds affect tubularsecretory function. In man, tubular secretion ofpara-amino-hippurate is diminished (23); in thedog, tubular secretion of potassium is inhibitedduring potassium loading (24). A specific actionof mercurial compounds on tubular reabsorptionof water has not been described. Administrationof free sulfhydryl groups in compounds such as 2,3dimercaptopropanol (British Anti-Lewisite) willinhibit mercurial diuresis (25-27).

The results in the present studies, as well asthose of Blumgart and associates (1, 19) in nor-mal subjects and patients with congestive heartfailure may be interpreted to indicate that mercury

1100


specifically depresses the tubular reabsorption ofchloride and that cation excretion follows passively.The quantity of body fluid removed, however, ismore closely related to fixed cation removal thanto anion removal. In each period of diuresis theloss of sodium plus potassium is nearly equal tothat predicted from the weight loss and concentra-tion of these ions in the body fluids.

In a minority of the patients the explanationthat sodium excretion is secondary to chloride ex-cretion apparently does not account for the findingsduring diuresis. In these, sodium and chloridewere lost in the proportion found in extracellularfluid. It should also be noted that, even in thepatients who lost more chloride than sodium, theratio of these losses varied widely. The factorsresponsible for the variation are not clear and re-quire further study before a generally applicableconcept describing the action of mercurial diureticscan be stated. It is possible that in every instancethe effect of mercury is primarily on the reabsorp-tion of chloride and that under optimal circum-stances sodium will be lost with chloride in propor-tions appropriate to the maintenance of normalbody fluid composition. However, if the stimulithat lead to sodium retention are intense becauseof hormonal factors or disturbances in renal he-modynamics, or both, then the factors that controlnormal cation-anion balance might become sec-ondary to the demands of sodium retention, andchloride would be lost in excess of sodium. It hasrecently been suggested that changes in plasmavolume may exert controlling effects on the excre-tion of salt and water (28). According to thisview, a decrease in plasma volume leads to reten-tion of sodium and water. In some of the presentgroup of patients, the intensity of the sodium-re-taining response may, in part, have resulted froma decrease in plasma volume consequent uponthe initial diuretic response to the mercurial. Un-der these circumstances, the rate at which inter-stitial fluid is withdrawn into the vascular com-partment would play an important role in deter-mining the Na/Cl loss.

The tentative hypothesis that mercurial diureticshave their primary action on the tubular reabsorp-tion of chloride is inferred from the observationspresented here which indicate that, over the courseof days, mercurials lead to a disproportionate re-moval of chloride from the body fluids. Other

workers have suggested that the action of mercuryma-y be primarily on proximal (22) or distal (21)tubular reabsorption of sodium. Such hypothesesfind neither support nor refutation in the presentdata and their final evaluation must await furtherinvestigation.

Whatever the mechanism may ultimately proveto be, the data presented here indicate a closetemporal relationship between the fall in serumchloride and rise in serum bicarbonate concentra-tions and the development of unresponsiveness tothe mercurial. Repair of the chloride deficit andthe concomitant alkalosis has regularly been ac-companied by the return of responsiveness.

Although the evidence presented here suggeststhat in the unresponsiveness to mercurial diureticsthe electrolyte disturbance is of critical importance,data of others suggest that the essential factormay be a reduction in glomerular filtration (21,29). It is possible that reductions in glomerularfiltration rate and serum chloride concentrationsboth produce unresponsiveness by diminishingthe rate at which chloride is delivered to thetubules.

In the two patients whose losses during diuresiswere of the same composition as extracellular fluid,responsiveness continued until edema had disap-peared. One of them initially demonstrated hy-pochloremia and an elevated bicarbonate concen-tration of a degree often found in the refractorygroup. These abnormalities in this patient per-sisted throughout the period of study, suggestingthat unresponsiveness may be more related tochanges in concentration than to absolute levels.In this minority group of responsive patients nosignificant changes occurred in the rate of excre-tion of ammonium, potassium, or titratable acid.

In those patients who excreted chloride in ex-cess of sodium during diuresis, the need for ad-ditional cation was met by the loss of potassium andammonium in reciprocally variable quantities, andin some instances indirectly by an increase in theexcretion of titratable acid. In many patients largeamounts of titratable acid and ammonium con-tinued to be excreted in the urine during the pro-gression of the alkalosis. The production of an acidurine in the presence of alkalosis has previously beenobserved in conditions characterized by dehydrationand sodium depletion and has been considered torepresent urgent need for cation conservation by

1101


the body (30, 31). During the production of ex-perimental potassium deficit in animals metabolicalkalosis develops (32). One explanation of thisevent could be the loss of acid in the urine, as issuggested by the observations made during pre-sumed potassium deficit in man by Kennedy,Winkley, and Dunning (33) and by Broch (34).Other mechanisms that might account for the pro-duction of alkalosis are (a) the exchange of hydro-gen and potassium ions between extra- and intra-cellular fluid, or (b) a loss from the intracellularspace of potassium and the anion of some weakacid. That one of the latter two mechanisms maybe operative is suggested by the data describingthe course of patient N. B. (Table IV). Elevationof serum pH and a progressive rise in serum bi-carbonate accompanied by a fall in serum chlorideconcentration occurred without an appreciable in-crease in the excretion of ammonia or titratableacid. The urine remained close to neutrality asthe chloride lost in excess of sodium was excretedwith an almost equivalent amount of potassium.The observation that the intracellular pH of muscletends to fall during imposed alkalosis likewise sup-ports the possibility of internal transfers (35).

The loss of potassium finds its most ready ex-planation in the discrepancy between the excre-tion of sodium and chloride. In certain patients,as has been shown, potassium appears to act as atemporary source of cation until ammonium pro-duction is sufficiently increased to conserve fixedcation. In other patients potassium continues tofurnish a large portion of the cation throughout theperiod of diuresis.

Two other possible mechanisms suggest them-selves as factors in the potassium loss, thoughneither of these explains the data in all instances.Depression of the tubular reabsorption of potas-sium by mercurials has been shown to occur inthe dog (24). If this occurs in man the effect mustbe transitory or irregular since 24-hour periods ofobservation in the present group of patients oftenindicated no loss of potassium. A third explana-tion, as suggested by the work of Darrow and as-sociates (32), might be that the loss of potassium isthe result of a biological equilibrium attained whennormal renal function is carried out in the pres-ence of deficit of certain ions in the body fluids.However, in the present group of patients, renalfunction cannot be considered normal and the defi-

cits of potassium either did not occur or were of amuch smaller degree than those demonstrated bythese authors.

It should be pointed out that, even when thereis a loss of potassium in excess of nitrogen, thedeficits rarely appeared to be of clinical significance.Electrocardiograms were obtained on all patientsduring successive phases of the observations andwere at no time suggestive of potassium deficit.Except for one instance, as long as the intake offood was adequate, the deficits ot potassium werequickly repaired when mercurial administrationwas stopped. The possibility remains, however,that during a prolonged period of diuresis, a pa-tient may become potassium deficient, particularlyif his intake of food is low.

It has long been known that acidifying salts po-tentiate, and alkalinizing salts depress mercurialdiuresis (2, 36, 37). The potentiation with acidi-fying salts has been found to be synergistic ratherthan additive. Recent studies in dogs by Axelrodand associates suggest that a change in pH is notthe leading factor (38). Reduction in pH by theadministration of ammonium chloride is followedby a synergistically potentiated response to mer-curials but a comparable reduction in pH inducedby breathing 7% carbon dioxide does not enhancethe diuretic effect (38). It would appear from thepresent data that it is the anion pattern of the se-rum that is important in determining the diureticresponse. That acidifying salts containing anionsother than chloride potentiate diuresis may be ex-plained by postulating that these anions either (1)substitute for chloride, or (2) allow for an increasein the concentration of chloride in the body fluids.Evidence for the second viewpoint is provided bythe studies of Gamble and associates (39). Theseauthors have shown that the administration of am-monium sulfate leads to elevation of the serumchloride concentration. Review of their data indi-cates the explanation of this phenomenon is thatthe diuresis induced by sulfate leads to an obliga-tory removal of cation and water without propor-tional loss of chloride.

It should be noted that in the present group ofpatients the reduction in serum chloride concen-tration was in no instance accompanied by a fallin serum sodium concentration while diuresis wasoccurring. The occurrence of hypotonicity aftermercurial diuresis has, however, been reported by

1102


many authors (40-42). The hypochloremic alka-losis seen in the present patients must be differ-entiated from the true "low salt" syndrome. Hy-potonicity failed to develop in the present patientsbecause water and cation were removed in pro-portional amounts. The development of hypo-tonicity requires the removal of cation withoutproportional quantities of water, the retention ofwater alone, or changes in the state of dissociationof certain intracellular compounds. Although hy-pertonic salt solution has been suggested for use inthe management of hypotonicity associated withmercurial diuresis, there appeared to be no indica-tion for its use in the group of patients here de-scribed.

SUMMARYANDCONCLUSIONS

The balance of weight, nitrogen, and electro-lytes, the electrolyte changes in the serum, and thecation-anion pattern of the urine were studied in10 patients with congestive heart failure who wereundergoing mercurial diuresis.

The pattern of the balances was interpreted interms of body composition and over-all functionof the kidneys.

A minority of the patients lost water and elec-trolytes in a manner indicating that the only sig-nificant change in the body fluids was loss of ex-tracellular fluid of usual composition. This smallgroup of patients never developed mercurial un-responsiveness.

The majority of patients lost electrolytes andwater in proportions that led to chloride depletionand alkalosis without change in the concentrationof sodium in the serum. This type of responsewas often associated with a loss of intracellularfluid and potassium as well as with reduction ofserum potassium concentration. In this group un-responsiveness to the diuretic developed in all butone instance. Repair of the chloride deficit withammonium chloride or hydrochloric acid restoredmercurial responsiveness. The serum concentra-tions of chloride and bicarbonate associated withunresponsiveness varied from patient to patient.In those patients in whom chloride excretion ex-ceeded sodium excretion the surplus of chloridewas covered by potassium and ammonium.

It has been suggested that the usual renal re-sponse to mercurial diuretics may result from in-terference with tubular reabsorption of chloride

and that the loss of fixed cation is secondary tothe effect on anion excretion. Possible factors de-termining the ratio of sodium to chloride loss havebeen discussed.

ACKNOWLEDGMENTS

The authors wish to express their appreciation to Dr.Samuel A. Levine for his advice and guidance in carry-ing out this study and to Dr. James L. Gamble for hiscontributions to the form of presentation and analysis ofthe data. They also wish to thank Miss ElizabethBradley and Mr. Joseph Greaney for their valuable tech-nical assistance.

REFERENCES1. Blumgart, H. L., Gilligan, D. R., and Volk, M. C.,

Action of diuretic drugs. II. Effect of diureticdrugs on the acid-base equilibrium of the blood inpatients with cardiac edema. Medical papers dedi-cated to Henry Asbury Christian. Waverly Press,Inc., Baltimore, 1936, p. 191.

2. Ethridge, C. B., Myers, D. B., and Fulton, M. N.,The modifying effect of various inorganic saltson the diuretic action of salyrgan. Medical papersdedicated to Henry Asbury Christian. WaverlyPress, Inc., Baltimore, 1936, p. 223.

3. Gamble, J. L., Wallace, W. M., Eliel, L., Holliday,M. A., Cushman, M., Appleton, J., Shenberg, A.,and Piotti, J., Effects of large loads of electro-lytes. Pediatrics, 1951, 7, 305.

4. Henderson, L. J., and Palmer, W. W., On the severalfactors of acid excretion. J. Biol. Chem., 1913-14,17, 305.

5. Folin, O., Laboratory Manual of Biological Chem-istry. Appleton-Century, New York, 1934, 5thed., p. 141.

6. Lavietes, P. H., D'Esopo, L. M., and Harrison, H. E.,The water and base balance of the body. J. Clin.Invest., 1935, 14, 251.

7. Darrow, D. C., da Silva, M. M., and Stevenson, S. S.,Production of acidosis in premature infants byprotein milk. J. Pediat., 1945, 27, 43.

8. Elkinton, J. R., Winkler, A. W., and Danowski, T. S.,Inactive cell base and the measurement of changesin cell water. Yale J. Biol. & Med., 1944, 17, 383.

9. Reifenstein, E. C., Jr., Albright, F., and Wells, S. L.,The accumulation, interpretation, and presentationof data pertaining to metabolic balances, notablythose of calcium, phosphorus, and nitrogen. J. Clin.Endocrinol., 1945, 5, 367.

10. Am, K. D., and Reimer, A., Minimal sodium lossesthrough the skin. J. Clin. Invest., 1950, 29, 1342.

11. Freyberg, R. H., and Grant, R. L., Loss of mineralsthrough the skin of normal humans when sweatingis avoided. J. Clin. Invest., 1937, 16, 729.

12. Shohl, A. T., Mineral Metabolism. Reinhold, NewYork, 1939 (American Chemical Society Mono-graph Series No. 82), p. 20.

1103


13. Yannet, H., and Darrow, D. C., The effect of deple-tion of extracellular electrolytes on the chemicalcomposition of skeletal muscle, liver, and cardiacmuscle. J. Biol. Chem., 1940, 134, 721.

14. Gamble, J. L., Companionship of Water and Electro-lytes in the Organization of Body Fluids, LaneMedical Lectures. Stanford University Press, Stan-ford, California, 1950.

15. Relman, A. S., and Schwartz, W. B., Unpublisheddata.

16. Govaerts, P., Origine renale ou tissulaire de la diuresepar un compose mercurial organique. Compt. rend.Soc. de biol., 1928, 99, 647.

17. Bartram, E. A., Experimental observations on theeffect of various diuretics when injected directlyinto one renal artery of the dog. J. Clin. Invest.,1932, 11, 1197.

18. Walker, A. M., Schmidt, C. F., Elsom, K. A., andJohnston, C. G., Renal blood flow of unanesthetizedrabbits and dogs in diuresis and antidiuresis. Am.J. Physiol., 1937, 118, 95.

19. Blumgart, H. L., Gilligan, D. R., Levy, R. C., Brown,M. G., and Volk, M. C., Action of diuretic drugs;action of diuretics in normal persons. Arch. Int.Med., 1934, 54, 40.

20. Schmitz, H. L., Studies on the action of diuretics. I.The effect of euphyllin and salyrgan upon glomeru-lar filtration and tubular reabsorption. J. Clin. In-vest., 1932, 11, 1075.

21. Pitts, R. F., and Duggan, J. J., Studies on diuretics.II. The relationship between glomerular filtrationrate, proximal tubular absorption of sodium anddiuretic efficiency of mercurials. J. Clin. Invest.,1950, 29, 372.

22. Mudge, G. L., Foulks, J., and Gilman, A., Effect ofurea diuresis on renal excretion of electrolytes.Am. J. Physiol., 1949, 158, 218.

23. Berliner, R. W., Kennedy, T. J., Jr., and Hilton, J. G.,Salyrgan and renal tubular secretion of para-amino-hippurate in the dog and man. Am. J. Physiol.,1948, 154, 537.

24. Mudge, G. H., Ames, A. III, Foulks, J., and Gilman,A., Effect of drugs on renal secretion of potassiumin the dog. Am. J. Physiol., 1950, 161, 151.

25. Handley, C. A., and LaForge, M., Effect of thiols on

mercurial diuresis. Proc. Soc. Exper. Biol. & Med.,1947, 65, 74.

26. Farah, A., and Maresh, G., The influence of sulfhydrylcompounds on diuresis and renal and cardiac circu-latory changes caused by mersalyl. J. Pharmacol.& Exper. Therap., 1948, 92, 73.

27. Earle, D. P., Jr., and Berliner, R. W., Effect of 2,3-dimercaptopropanol on diuresis. Am. J. Physiol.,1947, 151, 215.

28. Peters, J. P., Sodium, water and edema. J. Mt. SinaiHosp., 1950, 17, 159.

29. Weston, R. E., and Escher, D. J. W., An analysis ofthe unresponsiveness to mercurial diuretics ob-served in certain patients with severe chronic con-gestive failure. J. Clin. Invest., 1948, 27, 561.

30. Gamble, J. L., and Ross, S. G., The factors in the de-hydration following pyloric obstruction. J. Clin.Invest., 1924, 1, 403.

31. Van Slyke, K. K., and Evans, E. I., The paradox ofaciduria in the presence of alkalosis caused by hy-pochloremia. Ann. Surg., 1947, 126, 545.

32. Darrow, D. C., Schwartz, R., Iannucci, J. F., andCoville, F., The relation of serum bicarbonate con-centration to muscle composition. J. Clin. Invest.,1948, 27, 198.

33. Kennedy, T. J., Jr., Winkley, J. H., and Dunning,M. F., Gastric alkalosis with hypokalemia. Am.J. Med., 1949, 6, 790.

34. Broch, 0. J., Low potassium alkalosis with acid urinein ulcerative colitis. Scandinav. J. Clin. & Lab.Invest., 1950, 2, 113.

35. Wallace, W. M., and Hastings, A. B., The distributionof the bicarbonate ion in mammalian muscle. J.Biol. Chem., 1942, 144, 637.

36. Keith, N. M., Whelan, M., and Bannick, E. G., Theaction and excretion of nitrates. Arch. Int. Med.,1930, 46, 797.

37. Lyons, R. H., Jacobson, S. D., and Avery, N. L.,The change in plasma volume and body weight innormal subjects after a low-salt diet, ammoniumchloride and mercupurin. Am. J. M. Sc., 1946, 211,460.

38. Axelrod, D. R., Capps, J. N., and Pitts, R. F., Po-tentiation of diuretic action of salrygan by ammo-nium chloride. Federation Proc., 1950, 9, 6.

39. Gamble, J. L., Blackfan, K. D., and Hamilton, B.,A study of the diuretic action of acid producingsalts. J. Clin. Invest., 1924, 1, 359.

40. Soloff, L. A., and Zatuchni, J., Syndrome of salt de-pletion, induced by a regimen of sodium restrictionand sodium diuresis. J. A. M. A., 1949, 139, 1136.

41. Schroeder, H. A., Renal failure associated with lowextracellular sodium chloride; the low salt syn-drome. J. A. M. A., 1949, 141, 117.

42. Citron, D., Bercu, B., Lemmer, R., and Massie, E.,Congestive heart failure and hyponatremia: un-toward effects of mercurial diuresis. Ann. Int.Med., 1951, 34, 872.

1104

Documents

alteration - dm5migu4zj3pb.cloudfront.netdm5migu4zj3pb.cloudfront.net/manuscripts/102000/102529/JCI51102529.pdf · Balance data were calculated as net intake, minus com-bined output