long-termmedical management and complications of 'resistant' ascites · gut, 1961, 2. 285...

Gut, 1961, 2. 285

Long-term medical management and complicationsof 'resistant' ascites

WILLIAM H. J. SUMMERSKILL, BERNARD F. CLOWDUS, 11,1AND JOHN W. ROSEVEAR

From the Mayo Clinic and Mayo Foundationi, Rochester, Minnesota

SYNOPSIS This paper reports the experience of treating patients with hepatic cirrhosis and asciteswith an aldosterone inhibitor in addition to conventional therapy. Good results are demonstratedin 13 patients previously resistant to treatment.

Ratnoff and Patek (1942) commented that asciteswas the most frequent and characteristic sign ofhepatic cirrhosis, and the serious prognosis of thiscomplication was evident from the six-monthsurvival of only 50% of their patients. Despiteprogress in therapy, the outlook remains extremelygrave when the onset of ascites is insidious andwhen resistance develops to treatment with sodiumrestriction and standard diuretic agents (Sherlock,1958). Most authorities agree with Blakemore(1952) that intractable ascites represents an end stageof cirrhosis and that such patients have at best buta few months to live. Available measures includeparacentesis abdominis, a temporary expedientwhich may lead to protein and electrolyte depletion,hypotension, and coma (Sherlock, 1958), or variousportal-systemic shunt operations (Blakemore, 1952;Linton, 1956; McDermott, 1960). The last reduceportal pressure and obstruction to hepatic venousor lymphatic flow but affect only mechanicalaspects of the disease. Operative mortality withresistant ascites may exceed 40% (Linton, 1956),while the subsequent development of progressivehepatic failure or neuropsychiatric changes are alsoimportant considerations. Refractory ascites, there-fore, is considered a strong contraindication to thesurgical treatment of portal hypertension (Sherlock,1958; Blakemore, 1952; Mackby, 1960; Walker,1957), although newer techniques may justify somereappraisal (McDermott, 1960).

Since patients with cirrhosis and ascites havefeatures of excess aldosterone activity (Luetscherand Johnson, 1954), the management of refractory

'On assignment to the Mayo Foundation. The contents of this articleare the personal views of the authors and are not to be construedas official Air Force policy nor as Air Force endorsement of anycommercial product described.

fluid retention has been directed toward methodsof combating hyperaldosteronism. Bilateral ad-renalectomy (Marson, 1954; Giuseffi, Werk, Larson,Shiff, and Elliott, 1957) has been effective onlyoccasionally, presumably because of the dangers ofthe procedure in patients with liver disease and thedifficulties involved in immediate and long-termpostoperative management. Medical suppression ofadrenocortical function has been attempted(Summerskill and Crabbe, 1957; Stormont, Crabbe,Fast, Wolfe, and Davidson, 1959) but the toxicityof the drugs available precludes their long-term use.More recently, steroidal lactones (Kagawa, Cella,and van Arman, 1957), which act as chemicalantagonists of aldosterone (Liddle, 1958), have beengiven by mouth without side-effects to patients withcirrhosis and have procured an initial diuresis, evenwhen the ascites was refractory to standard measures(Clowdus, Higgins, Rosevear, and Summerskill,1960; Henley, Streeten, and Pollard, 1960).The progress of patients with resistant ascites

treated medically is rarely specified after they leavethe hospital and depends on many factors associatedwith hepatic disease, but particularly on their abilityto adhere to a strict regimen of diet and drugs. Inthis paper, details of an effective long-term medicalregimen to control ascites previously resistant tomedical management by using an aldosteroneantagonist in combination with standard diureticsare described, and the progress of 13 patients thustreated for periods up to two years is reported.Special attention is given to relaxation of the dietaryrestriction of sodium, which is made possible withtreatment, to the influence of liver function indetermining prognosis, and to the nature andmanagement of complications arising duringtreatment.

285

on October 4, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.2.4.285 on 1 D

ecember 1961. D

ownloaded from

http://gut.bmj.com/

286 William H. J. Summerskill, Bernard F. Clowdus, II, and John W. Rosevear

MATERIALS AND METHODS a factor during the preceding two months in only one(Case 13). In the other patients, the cause of the cirrhosis

Thirteen patients with cirrhosis (Table 1) were treated and its possible relationship to previous hepatitis wasat first in a hospital and were examined subsequently uncertain (cryptogenic). Tense ascites had been presentat intervals of two weeks to three months throughout for periods of one month to three years despite treatmentthe period of follow-up. The diagnosis of cirrhosis and was shown to be refractory for 10 days or longerwas made on clinical grounds, being supported by to treatment with dietary restriction of sodium (lessbiochemical tests of liver function (Table I) and, in than 10 mEq. in 24 hours), and combinations of meral-eight instances, by histological examination of hepatic luride (mercuhydrin), chlorothiazide, or prednisone intissue. Six patients (Cases 4, 5, 8, 9, 11, and 13) had been the hospital. Nine patients had required paracenteseschronic alcoholics, but continued alcoholism had been on two to 12 occasions during the preceding year, and

TABLE I

PATIENTS STUDIED AND RESULTS OF TREATMENT

t Previous History' Loss of Weight Chemical Tests of Liver Function Contllepitsin Hospital with (a) Initial and (b) Latest Follow-upInitial Treatment

~~~~~~~~~~~~~0~~~ ~ ~~~~~~~A.Z5 . :

a ..t ~

l2 55 F 32 1 0 3 35 70 (a) 1.9 3-0 >30 23 Transient impending coma with(b) 1-3 3-4 >30 24 chlorothiazide, otherwise well

for 24 mth., taking diureticsdaily and 80-mEq. sodium diet(see Fig. 2)

12 67 F 18 0 0 3 8 22 (a) 1 1 234 >30 22 Diedafter 2mth. Azotaemia(b) 1-0 348 >30 23 andcterminal gastrointestinal

haemorrhage32 51 M 9 0 0 2 23 18 (a) 2-6 3.0 >30 22 Well for 22 mth., taking

(b) 3-3 3-2 >40 23 intermittent diuretic therapyand 10-mEq. sodium diet

4 46 M 6 10 0 12 22 (a) 114 4-2 >30 22 Well for 15 mth. Transient(b) 03747 16 18 hyponatraemia and azotaemian

slow improvement in hepaticfunction, no therapy necessary

52 59 M 36 0 3 24 14 22 (a) 1-2 304 20 21 Well for 20 mth., taking(b) 037 438 20 19 intermittent diuretic therapy and

10-mEq. sodium diet64 69 M 3 0 0 0 65 32 (a) 1-4 348 30 24 Died after 6 mth. Hepatoma

(b) 102 470 30 - invading portal vein with re-accumulation of ascites andterminal gastrointestinalhaemorrhage

7 61 M 2 0 3 1 32 30 (a) 2-3 234 30 22 Died of unrelated cause after(b) 201 246 30 19 2 mth.

6' 61 M 3 0 - - 33 20 (a) 4-1 237 - 21 Died after 1 mth. from(b) 5.9 2t8 - igastrointestinal haemorrhage.

Hepatoma found at necropsy92 64 F 12 1 0 5 16 11 (a) 1-2 3-7 >30 20 Well for 9 mth., taking diuretics

(b) 0.9 3-7 >30 - and 40-80-mEq. sodium diet.Died in hepatic failure withazotaemia

10' 44 M 24 1 0 4 22 19 (a) 0.9 3-2 >30 20 Died after 4 mth.: azotaemia(b) 1-1 3-4 >30 21 and hyperkalaemia (Fig. 4)

A1 40 F 24 0 0 1 17 24 (a) 2-5 2-8 >30 22 Well for 15 mth. Slow improve-(b) 0-6 5.0 16 18 ment in liver function. No

therapy now necessary12' 62 F 7 1 0 6 23 24 (a) 3-6 2-7 >30 20 Died after 3 mth.: azotaemia

(b) 5-3 2-4 - 21 and terminal gastrointestinalhaemorrhage

132 49 M 1 0 0 0 51 32 (a) 9-2 2-8 - 27 Well for 9 mth., taking(b) 1-50 3.9 22 22 20-mEq. sodium diet.

Recovered from gastrointestinalhaemorrhage and azotaemia

Normal values <1-0 > 3.5 < 5 17

'All had been unresponsive to diet and diuretics previously2Patients with gastro-oesophageal varices



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

Long-term medical management and complications of 'resistant' ascites

10 had splenomegaly with oesophageal varices asevidence of portal hypertension. Post-mortem examina-tions were made on all patients who died.During initial hospital treatment patients were

given diets containing on analysis less than 10 mEq.of sodium and between 80 and 100 mEq. of potassiumdaily. Fluid intake was restricted only if azotaemiaoccurred. Weight was determined daily, and thevolume, together with the sodium and potassiumcontents, of 278 specimens of 24-hour urine wasdetermined during the treatment of patients 1 to 9.During the time in hospital and at follow-up appoint-ments, chemical tests of liver function and concentrationsof blood urea, together with serum sodium, potassiumchloride, and bicarbonate, were measured at intervalsby standard procedures. Determinations of ammoniain arterial blood using the method of Seligson andHirahara (1957), and electroencephalographic recordings,which were graded 0 to 4 (Summerskill, Clowdus, andCasey, 1960a), were made on patients with neuro-psychiatric changes. Exchangeable sodium, exchangeablepotassium, and total body water were measured by usingNa24, K42, and D20 as tracer substances, as describedelsewhere (Summerskill et al., 1960a; Clowdus, 11,Summerskill, Casey, Higgins, and Orvis, 1961).

PRELIMINARY ASSESSMENT AND RESPONSE TO INITIALTREATMENT

Initially, all patients had evidence of severe hepaticdisease with ascites and muscle wasting (Table I).

Three patients were clinically jaundiced, one beingdeeply icteric. All had direct-reacting bilirubin in theserum. The concentration of serum albumin initiallywas 3 g. per 100 ml. or less in eight patients, andall who were not jaundiced had grossly abnormalretention ofsulphobromophthalein (bromsulphalein).Serum glutamic oxalacetic transaminase activityand concentrations of globulin were abnormal inevery instance. All patients were hypoprothrombin-aemic, and the prothrombin time was unresponsiveto treatment with vitamin K. Five patients (Cases 4,7, 9, 10, and 13) had initial clinical and electro-encephalographic evidence of impending hepaticcoma, and two (Cases 1 and 10) had previously hadcoma following paracentesis. These considerationswere felt to preclude any patient from considerationfor surgical treatment.

After resistance of the ascites to a low-sodiumdiet with standard diuretic agents (mercuhydrin,chlorothiazide, and prednisone) had been confirmed,each patient was given medication with spirolactone-aldactone, 400 to 1,200 mg. in two to four dosesdaily in combination with chlorothiazide, 500 to1,000 mg. in two doses daily; meralluride (mercu-hydrin), 2 ml. daily intramuscularly; or prednisone,30 mg. daily in three doses (Fig. 1). Adjustments ofthe dose and frequency of administration of eachdrug and of the most satisfactory combinations of

Mean 24 hour weight change (lb.)

0 -0.5 -1.2 -1.3 -1.4 -1.8 -1.8

r5.Q1.c:.Izb

q).Q9h.

-C.,1*

zrti

Control Mercuhydrin + Spirolactone +Prednisone Prednisone Mercuhydrin Chloro- Prednison,e Prednisone

Chlorothiazide thiazide Mercuhy'drin Chloro-thiazide

FIG. 1. Mean change in weight, volume of urine, and urinary excretion ofsodium andpotassium in relationto treatment.

287

+0.5



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

William H. J. Summerskill, Bernard F. Clowdus, 1I, and John W. Rosevear

drugs to produce a sustained diuresis were necessaryin each patient.All responded strikingly to treatment (Table I

and Fig. 1), between 8 and 65 lb. in weight beinglost in 12 to 42 days. The greatest loss of weightby an individual in 24 hours was 5j lb., this beingassociated with a 24-hour excretion of 337 mEq. orsodium. There was usually a delay of a week, andsometimes considerably longer, before the diuresisbegan. With aldactone alone, a moderately increasedamount of sodium was excreted in association withtrivial loss of weight. The addition of prednisone oraldactone doubled loss of weight and increasedexcretion of sodium more than fourfold, while theuse of aldactone in combination with chlorothiazideor mercuhydrin promoted comparable loss inweight but greater excretion of sodium. Two drugswere more effective than one in combination withaldactone; greater loss of weight and excretion ofsodium occurred when either mercuhydrin orchlorothiazide was added to aldactone and pred-nisone. The greatest loss of weight (water) in relationto urinary excretion of sodium and potassiumoccurred when prednisone was administered withaldactone; the addition of thiomerin or chloro-thiazdie to this combination of drugs caused aproportionately greater increase in excretion of thecations than loss of weight. Loss of potassium wasgreatest during regimens in which chlorothiazidewas incorporated, and least when aldactone wasadministered. When aldactone and chlorothiazidewere given together, a negative potassium balanceoften occurred.Apart from improvement in jaundice in patient 13,

there was no significant change in hepatic functiontests during the period of initial diuresis. However,four of five patients who initially were in impendinghepatic coma recovered during treatment, whichincluded chlorothiazide, without antibiotic therapyor restriction of protein. No neuropsychiatricchanges developed in the other patients. Complica-tions arising during initial treatment, includingthose involving water and electrolyte metabolism,are discussed later.

RESULTS OF LONG-TERM MANAGEMENT

Long-term programmes of management after thepatients left the hospital were adjusted to subsequentprogress determined by their ability to continue adiet severely restricted in sodium and to abstainfrom alcohol, as well as by alterations in hepaticfunction and other complications incident to liverdisease. Six patients (Cases 1, 3, 4, 5, 11, and 13)are alive and have remained free from significantfluid accumulation during periods of eight months

to two years (mean 16 months). One patient (Case 7)died at home of an unrelated cause without re-accumulating ascites. Six patients died as a resultof complications due to progressive liver disease andportal hypertension (Table I). The mean survivalperiod in these cases was only four months, butgross ascites re-accumulated in only one (Case 6),this being due to occlusion of the portal vein byhepatoma.Two patterns of response were defined during

follow-up studies: (1) that of the patients whoultimately dispensed with diet and drugs, and (2) thatof the patients who required continuous therapy.The second group of patients was subdivided intothose who were able to continue strict sodiumrestriction at home (as shown by analysis of aliquotsof their diets) and those in whom some relaxationof sodium intake was inevitable because of theirinability to prepare or adhere to a correct diet. Atrivial amount of ascites was allowed to accumulatein patients taking diuretics continuously as a pre-caution against sodium depletion.

PATIENTS RETURNING TO NORMAL DIETS WITHOUTDIURETIC THERAPY Three patients (Cases 4, 1 1,and 13) were chronic alcoholics in whom asciteshad developed within the year preceding treatmentbut who ceased drinking one to three months beforeentering the hospital. After initial treatment, liverfunction improved in patient 13 only. All avoidedfurther consumption of alcohol, and a diet providingapproximately 10 mEq. of sodium, a liberal intakeof protein (through low-sodium supplements), andadded vitamins was prescribed. Subsequently, nonereaccumulated fluid, all gained flesh, and hepato-megaly decreased. Chemical tests of liver functionimproved commensurately (Table I). The dailysodium content of the diet was increased cautiouslyby 20 mEq. increments at three-monthly intervals,and all have now reverted to normal diets withoutadverse effects, nine to 18 months having elapsedwithout retention of fluid.

PATIENTS REQUIRING CONTINUOUS TREATMENT Thepatients in this group can be subdivided into twogroups according to whether they required inter-mittent or continuous diuretic therapy.

1 Dietary Sodium of 10 mEq. Daily withIntermittent Diuretic Therapy Two patients (Cases 3and 5) were free from ascites when they left thehospital, and continued a diet providing approxi-mately 10 mEq. of sodium without undue difficulty.Subsequently, fluid re-accumulated, with gains of4 and 6 lb. within three months, and in additionto diet diuretics were necessary. After severaladjustments, one patient (Case 3) has remained

288



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/


DBWlkg.)----2-8-- 5-36 - - -33.77NOE(mEq.)-----------------4280---?280--890-----3099----3000- -- -3001KE(mEq.)---- - - ----2020--2570----2710-2460--2442--- -2383- --- - -2440

NaS(mEqL.)--- ---- - - -- ---- - ---141----140----139------ 143-----140------ --- -- -146KS(EqL)- ------.----4.6----5.4--5.---S.7-----54- ---4.8

EEG(grodeo-4)-- - I-------I----mNH4(pg.Aooml.)--- - - -- - - - --196- 121-- -------242

Impending comaImpending coma

Body 15511weight(lb.) 145

135

125LDietory sodium (mEq.)

Spiroloctone (mg.)Chlorothiazide (mg.)

Mercuhydrin (ml.)Prednisone (mg.)

Potossium supplements(90 mEq. daily)

10%

500 b

10 4080-L6 80-100

_600 ~600 ---------*.0 400

.2n*QOO. ,* 500 ,JOQQ 500. ioootwiceweekl .2.0 toaernate days

. 30 ,

5 10 30 40 50 60 70

Weeks80 90 100 110

FIG. 2. Details of treatment and complications in the long-term management of resistant ascites (patient 1). Diet andmedication are related to body weight, dry body weight (DBW), exchangeable sodium (NaE) and potassium (KE) values;serium sodium (Nas) and potassium (Ks) concentrations, and arterial ammonia concentrations (NH4).

free from ascites for 22 months with treatment forone week every two months with spirolactone,600 mg. daily, to which prednisone, 30 mg. daily,and chlorothiazide, 500 mg. twice daily, has beenadded during the last three days of the week.Patient 5 has again become responsive to chloro-thiazide and has maintained a steady weight for19 months by taking 500 mg. of this drug on alternatedays. On similar regimens, patient 7 remained freefrom significant ascites until his death, but patient 6re-accumulated fluid at home terminally.

2 Liberal Dietary Sodium with ContinuousDiuretic Therapy Five patients (Cases 1, 8, 9, 10,and 12) were unable technically or by inclinationto persevere with strict dietary restriction of sodium.Consequently, fluid began to re-accumulate afterdischarge from the hospital. Studies of sodiumbalance (Clowdus et al., 1960) showed that thedaily administration of spirolactone in doses of300 to 800 mg., together with other diuretics,permitted relaxation of the dietary restriction ofsodium without undue retention or loss of sodium.Careful supervision with frequent readjustments oftreatment was necessary, but with appropriateregimens all remained free from obvious ascites,despite diets containing 50 to 100 mEq. of sodium.Management sometimes involved difficulties, mostof which are illustrated by the progress of patient 1(Fig. 2).

After a good response to treatment in hospital andstabilization of weight on diet alone, this patient wasunabIe to persevere with the diet at home and gained7 lb. within six weeks. Chlorothiazide, prednisoneand then spirolactone were prescribed in varyingdoses. She then lost weight, but an episode of im-pending coma, presumed to be due to chlorothiazide,necessitated withdrawal of drugs. Weight againincreased; a return to treatment with aldactone andchlorothiazide was made with satisfactory diuresis,potassium supplements having been added to theregimen. However, an intermittent mild neuro-psychiatric disorder persisted with minor abnor-malities in the electroencephalogram.An attempt was then made to provide more

sodium in the diet. Balance studies with an 80-mEq.sodium diet showed a satisfactory sodium balanceand no gain in weight, while treatment with diureticswas continued, but fluid accumulated with greateramounts of sodium. Accordingly, an 80-mEq.sodium diet with continued daily treatment withspirolactone and chlorothiazide was institutedsatisfactorily. Attempts to stop chlorothiazide andlater to reduce the dosage of aldactone slowly by100 mg. were followed by increases in weight.Diuresis was procured by increasing the dose ofchlorothiazide, but another transient episode ofimpending coma occurred. The dose of spirolactonewas increased, but replacement of chlorothiazide by

289



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

William H. J. Summerskill, Bernard F. Clowdus, II, and John W. Rosevear

injections of thiomerin on alternate days failed toretard the gain in weight. With a schedule of spiro-lactone, 400 mg. daily, with chlorothiazide givenonly twice weekly in doses of 500 mg. twice a day,the subsequent course has been satisfactory. Thus a

palatable diet requiring little special preparation hasbeen combined satisfactorily with continuousdiuretic therapy which is less likely to cause com-

plications.Certain observations, which apply also to other

patients, were evident from this study. A con-

siderable proportion of the weight gained afterleaving the hospital was due to increase in flesh,as shown by serial calculations of the dry bodyweight (DBW), which represents body weight lessbody water. Neuropsychiatric complications occurredspecifically in relation to chlorothiazide therapy,but could not be related to potassium depletion,as values for exchangeable body potassium (KE),KE/DBW (78.1, 89.2 and 72-3 mEq./kg.), andconcentrations of serum potassium (Ks) were notstrikingly changed in the presence of clinical or

electroencephalographic evidence of impendingcoma. Moreover, potassium supplements failed toprotect against the onset of coma on two occasionsand were not followed by striking increases in bodypotassium stores as shown by KE measurements.In contrast, concentrations of ammonia in arterialblood were above the normal range (less than100 mcg. per 100 ml.) and became more abnormalin association with the clinical and electro-encephalographic changes of impending coma.

COMPLICATIONS

GASTROINTESTINAL HAEMORRHAGE AND HEPATOMAMassive haemorrhage from gastro-oesophagealvarices occurred in five patients, all of whom hadseverely impaired liver function with defectiveblood coagulation; only one survived (Table I).Bleeding occurred during initial hospital treatmentin three patients (Cases 2, 8, and 13), in two ofwhom hepatoma was found at necropsy, despiteprevious failure to identify malignant cells in theascitic fluid.

OVERHYDRATION, HYPERKALAEMIA, AND AZOTAEMIA

Disorders of water and electrolyte metabolism,sometimes with azotaemia, were prominent, eitherinitially or during subsequent management (Tables I

and II). Before treatment, isotope studies on 11patients confirmed the excess total body water(TBW) and exchangeable sodium (NaE) known tooccur in patients with ascites (Birkenfeld, Liebman,O'Meara, and Edelman, 1958). Paradoxical hypo-natraemia (Talso, Spafford, Ferenzi, and Jackson,1956) was present at this stage in five patientswho had serum concentrations of less than 135 mEq.of sodium per litre. Values for serum potassium(Ks), exchangeable body potassium (KE), and bloodurea were within the normal range, except in one

patient who was hyperkalaemic and mildly azo-

taemic at this stage. Hyponatraemia was associatedwith hypochloraemia and reductions in serumbicarbonate concentrations. The lower initial values

TABLE II

CHANGES IN VALUES FOR DRY BODY WEIGHT (DBW), TOTAL BODY WATER (TBW), BODY EXCHANGEABLE SODIUM

(NaE) AND POTASSIUM (KY), SERUM SODIUM (NaS) AND POTASSIUM (KcS), AND BLOOD UREA (UB) BEFORE AND

AFTER DIURESIS

DBW (kg.) TBW (kg.) NaE (mEq.) Nas (mEq./l.) KE (mEq.) KS (mEq.l. ) UB (mg./lOOml.)

Before After Before After Before After Before After Before After Before After Before AfterCase

No azotaemia13S

68

MeanPercentagechange

Azotaemia249101213

MeanPercentagechange

28.132.825 324-434.3290

28.533.025925-733.929.4

47.049 049.876.149.454.3

4- 1-4

36.317-820.224-723-126.8248

35-118-019-625 223.425.824.5

36.740041 256.640 643-0

20.7

39.448.232948.541*7504943.6

37.646230.339.332.835.4369

5,0003,6454,7205,2704,8804,703

2,7002,4643,6103,7843,7603,263

30-6

3,9304,5502,7513,6173,6165,7564,037

2,2903,2632,3302,7022,4072,3382,555

143132132140132136

137127127134132131

2,6502,940

2,9102,4702,744

3.7

137135140134130138136

108119136115117102116

2,5052,800

2,6802,0822,517

5.44.25.9

5.45.35.2

8-2

2,0502,0501,8512,7501,5742,3842,110

2,1002,0801,9492,6001,2632,1692,027

5.34.55 66 54.03.34.9

5.5

4-65S65.25-65.3

+ 1.9

684.76-58.26 25.76 4

492841473640

403434443638

5.0

59

377854331245

15390132110123111

120

- 154 - 36.7 - 147 - 3.9

290

+ 30 + 166- 1.2



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/


for total body water, NaE, and KE in patients sub-sequently having azotaemia reflected their smallerphysiques as shown by the dry body weight. Other-wise the two groups could not be distinguished beforetreatment.During diuresis, greater hyponatraemia developed,

while three patients became azotaemic at this stage,and three more became azotaemic during subsequentmanagement (Table II). The presence of azotaemiawas associated with symptoms, more strikingmetabolic changes, and a worse prognosis. Theclinical syndrome consisted of apathy, nausea, andanorexia, progressing in more severe instances topersonality changes and stupor. The motor systemdisorder of impending coma (flapping tremor,hyporeflexia, rigidity, and clonus) was not prominent.Mental changes were not specific, and abnormalitiesof the electroencephalogram, arterial pH, andarterial ammonia concentrations, if present, werenot striking and did not become more severe asthe condition deteriorated (Summerskill et al.,1960a). Azotaemia developed in the absence ofprecipitating factors, such as gastrointestinalhaemorrhage or paracentesis abdominis, and withoutevidence of pre-existing renal disease, alterations insystemic blood pressure, and volume of urine(which were recorded daily), or characteristichistological changes in the kidneys at necropsy. Asmaller proportion of body water was lost byazotaemic patients than by the other patients.Losses of sodium were comparable, while relativelylittle potassium was lost by either group. Therewas much greater hyponatraemia in azotaemicpatients in association with the greater waterretention, but concentrations of serum potas-sium became raised without elevation of bodypotassium content (KE). These disorders werenot related to changes in flesh, as little alterationoccurred in dry body weight in either group(Table II).The patients who became azotaemic often had

more severe liver disease on clinical assessment andmore grossly deranged chemical tests of liverfunction (Table I). Two (Cases 4 and 13) recoveredsimultaneously with improvement in liver function,whereas azotaemia, hyponatraemia, and hyper-kalaemia persisted for periods of two to nine monthsbefore death in three other patients (Cases 9, 10,and 12). Sudden death of three azotaemic patients(Cases 2, 9, and 10) appeared to be primarily dueto hyperkalaemia, as electrocardiograms showedcharacteristic abnormalities and no alternativeexplanation could be found at necropsy. There wasno relationship between the type or amount ofdrugs prescribed and the onset of azotaemia;patients subsequently having this complication had

received smaller quantities of drugs and had aslower diuretic response.Treatment of overhydration and azotaemia was

unsatisfactory. Hyponatraemia without raised con-centrations of serum potassium or blood urea gaverise to no specific symptoms and corrected itselfspontaneously after diuresis. In azotaemic patients,overhydration was controlled temporarily by restric-tion of fluids, but severe thirst made this difficultfor longer-term use, and azotaemia progressednotwithstanding (Cases 2, 9, 10, and 12). Theenhanced water diuresis with prednisone (30 to45 mg. daily) was not sufficient to prevent the onsetor progression of overhydration and azotaemia infive patients, nor did the administration of mannitol(5% solution) by vein cause weight loss in fivepatients already receiving diuretics. An increasedurinary volume followed mannitol infusions but wasaccounted for by the fluid load administered. Fourpatients excreted an increased amount of sodiumin the subsequent 48 hours (mean 20.5 mEq. ofsodium per litre of mannitol administered).

HEPATIC COMA In addition to the neuropsychiatricchanges during azotaemia, hepatic coma occurredas a terminal event following gastrointestinalhaemorrhage in Cases 2, 6, and 8. Personalitychanges, non-specific tumour, and minor electro-encephalographic abnormalities of grade 1 or 2severity appeared to be due primarily to the severityof the liver disease on initial assessment in Cases 4,9, 10, and 13. In Cases 1 (Fig. 2) and 7 transientimpending coma was clearly related to treatmentwith chlorothiazide and responded quickly towithdrawal of the drug. In other cases, diuretictherapy, including chlorothiazide, had no directinfluence on the neuropsychiatric changes, and threepatients recovered during good diuretic responses.

Several problems associated with managementduring the development and course of disorders ofwater and electrolyte metabolism and azotaemiaare illustrated by the progress in Case 10 (Fig. 3).Prolonged treatment with sodium restriction andincreasing doses of diuretics were necessary toinitiate diuresis. After steady loss of weight, apathy,confusion, and anorexia appeared in associationwith more severe hyponatraemia and hyperkalaemia,neither of which could be explained from changes inbody composition as there was no increase inhydration (as indicated by the ratios NaE + KE/TBW) and no evidence of sodium depletion orpotassium retention from the NaB and KB values.The electrocardiogram showed gross changes ofhyperkalaemia. Simultaneously, azotaemia de-veloped, with deterioration in the electroencephalo-gram and abnormal arterial concentrations of

291



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

William H. J. Summerskill, Bernard F. Clowdus, 11, and John W. Rosevear

NaCKE/TB3W(mEqikg.) 131.3-134.9 129.4-1333TBW(kg.) 48.5-39.3-----39.8------ --34.2

NaE(mEq.) 3617--2702----2607--------- -2470KE(mEq.) 2750------ - -------2600----2443-----------------2090

NaS(mEq./L.) 127- - - - - - - - - - - - - - - - - - - - 115-- - --116 ---- - -- 95 - - - - - - 110 - --106- - - - -110-- --117Ks(mEq./L.) 6.5 - - - -8.2----- -70- - - - - 6.7--- -- - -7.5----7.1- -- - --7.9 ---- ---7.

UB(mg.Aoo ml.) 54 - ----------- -- -134 ------- 152 ------- --96 114------95------- 131Hb.(gm./looml.) 12.8----- ------- -13.1- -10.8-- 82

EEG I--- ----- -ii-------170 Fluid restriction160 Potassium restriction

Body 150- Protein restriction Paracentesisweight

(Ib.) 140-_\ Left hospital

130 - Apathy and confusion

120Dietary sodium (mEq.) 10 ; 40-100

Spiroloctone (mg.) ..800 2i 20 - .. 800 1000 800 -Chlorothiazide (mg.) 1000 1000 ___Mercuhydrin (2 ml.) t t t t t

Prednisone(ng.) --- 20 >* 30 45 ..30Mannitol (1-2 liters) t

2 32 3 17 8 9 10 [ 15 16

WeeksFIG. 3. Disorders of water and electrolyte metabolism and azotaemia during treatment of resistant ascites (patient 10).Changes in total body water (TB W), exchangeable sodium (NaE) and exchangeable potassium (KB) values; serum sodium(Nas) and potassium (Ks) concentrations; and haemoglobin (Hb) and blood urea (UB) concentrations, in relation tomanagemenit.

ammonia, but without change in blood pressure oroutput of urine. The neuropsychiatric disorderwas not influenced by withdrawal of protein andchlorothiazide, but restriction of potassium resultedin some improvement in serum potassium con-centration and in the electrocardiographic pattern.Relative overhydration then occurred, with a fallin the ratio, NaE + KE/TBW. Despite restrictionof fluid, the concentration ofserum sodium continuedto fall and azotaemia increased, although improve-ment in the NaE + KEBTBW ratio occurredsubsequently.With improvement in the general condition, an

increased dietary intake of sodium was attemptedin conjunction with higher doses of spirolactoneand chlorothiazide, so as to relax dietary restrictionsbefore discharge from the hospital. The weightremained stable when a 40-mEq. sodium diet wasused, although the concentration of serum sodiumcontinued to fall and that of the blood urea rose,whereas an 80-mEq. sodium diet resulted inaccumulation of fluid. Improvement in serumsodium and blood urea concentrations occurred,and to facilitate discharge from the hospital,paracentesis was performed uneventfully.The further course was satisfactory only in

relation to the fact that no further weight wasgained. Precautions against excess intakes of water

and potassium remained necessary, and serumpotassium concentrations remained dangerouslyhigh, while hyponatraemia persisted and concentra-tions of blood urea fluctuated but never reverted tonormal. On four occasions, mannitol was givenwithout effect. Progressive anaemia occurred inassociation with the azotaemia; tests for occultblood in the stools had been negative on 12 occasions.There was no alteration ofserum bilirubin concentra-tion or other evidence of increased haemolysis, andblood smears, together with leucocyte and plateletcounts, were unremarkable. The anaemia appearedto be associated with the impaired renal function,and blood transfusions were necessary. Long-standing constipation was aggravated by fluidrestriction, and evacuation of faeces under caudalanaesthesia was necessary. Three days later thepatient complained suddenly of progressive paralysisand died within five minutes, the mode of deathbeing ascribed to hyperkalaemia, as there was noother explanation for it at post-mortem examination.

COMMENT

The long-term medical management of asciteshas been emphasized less than immediate hospitaltreatment, although successful therapy dependsparticularly on perseverance with an unpalatable

l

292



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/


and technically difficult diet, while there is under-standably little information available regarding theprogress of patients when ascites is refractory totreatment. In our experience, the use of spirolactonein combination with other diuretics always overcamethe initial resistance of ascites to other medicalmeasures, although there was occasionally a delayof two weeks before diuresis began. Moreover, thistreatment permitted the relaxation of dietaryrestriction of sodium (Clowdus et al., 1960) whennecessary in subsequent management without fluidre-accumulating. The initial diuretic response wasrelated to the number of drugs and the dosesprescribed, but long-term management was achievedwith less medication, the amount usually varyingwith the intake of sodium. Balance studies confirmedearlier observations regarding the metabolic effectsof these various drugs. Spirolactone had a potassiumconserving effect, whereas chlorothiazide increasedthe loss of potassium, and prednisone increased theexcretion of water in relation to sodium andpotassium (Clowdus et al., 1960; Cattan and Vesin,1957; Dingman, Finkenstaedt, Laidlaw, Renold,Jenkins, Merrill, and Thorn, 1958; Shaldon,McLaren, and Sherlock, 1960; Laragh, Heinemann,and Demartini, 1958; Read, Laidlaw, Haslam, andSherlock, 1959; Salassa, Mattox, and Power 1958).Nevertheless, the simultaneous administration ofaldactone and chlorothiazide often led to a negativepotassium balance, while the increased diuresis ofwater associated with prednisone therapy did notnecessarily prevent overhydration. The incidence ofneuropsychiatric complications attributable tochlorothiazide in these and other patients treated byus was much below that reported by Read et al.(1959), and potassium depletion was not proved tobe the major factor. Concentrations of serumpotassium are infrequently reliable indices of bodystores of potassium in patients with ascites (Summer-skill et al., 1960a; Clowdus, II, et al., 1961; Talso etal., 1956), and the direct effect of chlorothiazide onammonia metabolism (Owen, Flanagan, and Tyor,1959; Mackie, Stormont, Hollister, and Davidson,1958; Casey, Summerskill, and Orvis, 1961) seenin one of the patients reported herein was supportedby the demonstration of an interrelationshipbetween ammonia and potassium metabolism inothers (Casey et al., 1961). This may explain thesusceptibility of patients who had previouslyexperienced coma to neuropsychiatric changesduring chlorothiazide therapy and the raised arterialconcentrations of ammonia under such circum-stances, which is evident from earlier data (Readet al., 1959). These considerations were of greatestimportance during periods of rapid diuresis andwhen complications involving water and electrolyte

metabolism occurred, but they seldom influencedlong-term management because of the smaller dosesof drugs necessary for maintenance. A combinationof spirolactone and chlorothiazide was most con-venient for long-term treatment but newer drugs mayhave fewer disadvantages.The constant response to a drug acting primarily

as an antagonist of aldosterone indicates theimportance of hyperaldosteronism in the refractori-ness of ascites to standard treatment. Such resistanceappeared to be closely related to the severity of theliver disease, as patients in whom liver functionultimately improved dispensed with treatment,whereas others required continued diet and therapywith diuretics. A direct relationship betweenhyperaldosteronism and liver function (Summerskilland Crabbe, 1957; Coppage, Island, Cooner, andLiddle, 1961) or other factors important in theformation of ascites, however, remains speculative.

Resistant ascites associated with cirrhosis carrieda more favourable long-term prognosis in thealcoholic than in the non-alcoholic, as greaterrecovery of liver function was possible as a result ofa nutritious diet and abstinence from alcohol(Summerskill, Davidson, Dible, Mallory, Sherlock,Turner, and Wolfe, 1960b). Nevertheless, relaxationof diuretic therapy and sodium restriction had tobe gradual in the alcoholic. In contrast, patientswith cirrhosis of uncertain cause pursued a moreinexorable course (Summerskill et al., 1960b) whichwas reflected by their need for continued treatmentto prevent re-accumulation of ascites. Only onepatient became responsive after treatment to adiuretic which had been ineffective previously.The prognosis is grave in resistant ascites because

of several complications associated with advancedliver disease. Haemorrhage from ruptured gastro-oesophageal varices, hepatoma, and hepatic comaoccurred during the course of follow-up. Disordersof water and electrolyte metabolism of ui¶certainorigin occurred particularly frequently and consti-tuted perhaps the greatest problem in management.Their high incidence may support an often assumedrelationship to the use of a sodium-restricted dietand diuretics. No direct connexion between thevarious components of treatment and thesecomplications could be demonstrated, however,and we were particularly careful to guard againstsodium or potassium depletion. Advances in themanagement of other complications associated withhepatic disease have resulted in the greaterprominence of disorders of water and electrolytemetabolism in hepatic failure (Chalmers, 1960),and their relationship to deterioration in hepaticfunction appears to be most important, especiallyas two patients with severe hyponatraemia and

293



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

William H. J. Summerskill, Bernard F. Clowdus, II, and John W. Rosevear

azotaemia recovered when the liver disease improved.Two uncertainly related phases were distinguished

in this disorder, the first reflecting overhydration(Clowdus, II, et al., 1961; Birkenfeld et al., 1958;Talso et al., 1956) and the second comprising more

severe overhydration in association with azotaemiaand additional metabolic and clinical changes(Clowdus, II, etal., 1961). Hyponatraemia representedprimarily dilution due to a relatively greater excess

of body water than sodium, rather than sodiumdepletion; hypokalaemia, when present, may reflectthe same phenomenon as well as deficiency ofpotassium. However, potassium supplements infre-quently lead to replenishment of potassium storesas judged by changes in body potassium (Caseyet al., 1961), nor were they always effective inpreventing coma during treatment with chloro-thiazide. In addition to sodium retention, therelative excess of body water may result fromdiminished free water clearance (Schedl and Bartter,1960) or excess ADH activity (Ralli, Robson,Clarke, and Hoagland, 1945) and it was particularlyprominent when azotaemia developed. Althoughprednisone (Dingman et al., 1958) and mannitol(Shaldon et al., 1960; Schedl and Bartter, 1960) mayenhance free water clearance in some patients, therewas no significant clinical effect in our azotaemicgroup, perhaps beoause sodium diuresis was alreadyin progress. Restriction of fluids limited over-

hydration most effectively but failed to halt azot-aemia and caused great thirst.

Despite overhydration and consequent dilution,hyperkalaemia occurred with azotaemia andappeared important as a cause of death. It couldnot be related to retention of potassium, thus raisingthe possibility that redistribution of potassiumfrom the intracellular to the extracellular compart-ments resulted in hyperkalaemia and contributed tohyponatraemia. Potassium restriction was not fullyeffective, and the use of insulin, glucose, and sorbitolin the treatment ofhyperkalaemia must be consideredalso.The impairment of renal function leading to

azotaemia in our cases occurred without clinical or

histological evidence of renal disease. A disorder ofthe renal circulation has been inferred by Papper,Belsky, and Bleifer (1959) and Hecker and Sherlock(1956), although oliguria and hypotension were notinvariably present in their patients. In our cases,azotaemia preceded detectable oliguria and peri-pheral vascular collapse, but renal plasma flowwas low (Feichter, Summerskill, and Wakim,unpublished data), as in three cases reported byPapper et al. (1959). This appeared the mostimportant cause of a diminished glomerular filtrationrate, but the anatomical and pathological basis is

conjectural. The occurrence of progressive anaemiain one azotaemic patient was of interest in view ofthe relationships between erythropoiesis and renalfunction. The poor immediate prognosis of severehyponatraemia with azotaemia stressed by others(Papper et al., 1959; Hecker and Sherlock, 1956)was not altogether supported in our group, as thesechanges sometimes persisted for weeks or monthsand two patients recovered.Secondary impairment of renal function may

modify the clinical picture. The neuropsychiatricchanges in azotaemic patients were not characteristicof hepatic coma and did not respond strikingly torestriction of protein and antibiotics. Abnormalitiesof the electroencephalogram, arterial pH, and bloodammonia concentration may be influenced bydisorders of water and electrolyte metabolism andazotaemia (Chalmers, 1960; Webster and Gabuzda,1959), thus making separation of clinical patternsdifficult in patients with severe liver disease.Our patients had hepatic disease of a severity

which precluded initial consideration of surgicaltreatment and militated against the successfulmedical or surgical management of any subsequentgastrointestinal haemorrhage. The therapeuticapproach of choice to resistant ascites, therefore,must be evaluated in relation to improvements inconservative management, as well as in surgicaltechniques and results (McDermott, 1960; Welch,Welch, and Carter, 1959; Barker and Reemtsma,1960). The simultaneous improvement in liverfunction and portal hypertension (Reynolds, Geller,Kuzma, and Redeker, 1960) in cirrhosis of thealcoholic suggests that operation may be deferredfor six months, so that response to medical treatmentand the sincerity of abstention from alcohol can beassessed. Post-operative improvement in ascites mayon occasion be related to abstinence from alcoholrather than to the operation itself, whereas hazardousoperations are seldom warranted when continuedalcoholism leads to hepatic failure. Until the variousfactors responsible for resistance of ascites indifferent patients are more clearly defined, surgicaltreatment would appear to be most clearly indicatedfor the relatively uncommon patient in whomresistant ascites is associated with apparently goodliver function, the latter being judged particularlyby general nutrition, the absence of jaundice orprevious coma, and satisfactory blood coagulation.The demonstration of gastro-oesophageal variceswould favour a portal-systemic shunt, particularly ifgastrointestinal haemorrhage had occurred pre-viously, whereas adrenalectomy might be consideredfor the other patients. In regard to paracentesis,Copland (1844-58) commented a century ago thatit 'seems calculated to increase the mischief, and to

294



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

Long-term medical management and complications of 'resistant' ascites 295

diminish the chances of a complete cure. It is thelast means to which recourse should be had'.

SUMMARY

Resistance of ascites associated with hepatic cirrhosisto standard treatment was demonstrated in 13patients, all of whom responded to treatment in ahospital with dietary restriction of sodium and acombination of diuretic agents, which includedspirolactone. The effects of various combinationsof drugs on excretion of water, sodium, and potas-sium were studied. Medical supervision and treat-ment were continued for periods of up to two years,and the following patterns of response in relationto the long-term management of resistant asciteswere found.

Patients in whom alcoholism had been a factor inthe previous six months were ultimately able todispense with treatment, but only after severalmonths. This was associated with continuedabstention from alcohol and slow improvement inliver function. In patients with cirrhosis of uncertaincause, liver function either deteriorated or remainedunchanged, and continuous or intermittent treatmentwas necessary. The amount of drugs required de-pended on patients' ability to adhere to a low-sodiumdiet. In most instances, it was necessary to relax thedegree of sodium restriction. Continuous treatmentwith spirolactone and other drugs permitted amore liberal intake of sodium when necessary.The details and complications of management aredescribed with particular emphasis on those involvingwater and electrolyte metabolism.

Six patients have remained well and free fromdisabling ascites, and six have died from complica-tions associated with progressive liver disease; onepatient died from an unrelated cause. Now that itis feasible to treat resistant ascites for long periodson an out-patient basis with a medical programme,the management of this condition, especially inregard to surgical procedures, is re-evaluated.

REFERENCES

Barker, H. G., and Reemtsma, K. (1960). The portacaval shuntoperation in patients with cirrhosis and ascites. Surgery,48, 142-152.

Birkenfeld, L. W., Leibman, J., O'Meara, M. P., and Edelman, I. S.(1958). Total exchangeable sodium, total exchangeablepotassium, and total body water in edematous patients withcirrhosis of the liver and congestive heart failure. J. clin.Invest., 37, 687-698.

Blakemore, A. H. (1952). Portacaval shunting for portal hypertension.Surg. Gynec. Obstet., 94, 443-454.

Casey, T. H., Summerskill, W. H. T., and Orvis, A. L. (1961).Disorders of potassium metabolism in liver disease. (Abstr.)J. clin. Invest., 40, 1028-1029.

Cattan, R., and Vesin, P. (1957). ttat actuel du traitement descirrhoses ascitiques par la delta-cortisone. Sem. Hop. Paris,33, 76-79.

Chalmers, T. C. (1960). Pathogenesis and treatment of hepaticfailure. New Engl. J. Med., 363, 23-30; 77-82.

Clowdus, B. F., Higgins, J. A., Rosevear, J. W., and Summerskill,W. H. J. (1960). Treatment of 'refractory' ascites with a newaldosterone antagonist in patients with cirrhosis. Proc. MayoClin., 35, 97-105.II, Summerskill, W. H. J., Casey, T. H., Higgins, J. A., andOrvis, A. L. (1961). Isotope studies of the development ofwater and electrolyte disorders and azotaemia during thetreatment of ascites. Gastroenterology, 41, 360-370.

Copland, James (1844-58). A Dictionary of Practical Medicine.Longman, Brown. Green, Longmans and Roberts, London.

Coppage, W. S., Island, D., Cooner, A. E., and Liddle, G. W. (1961).Metabolism of aldosterone in normal subjects and patientswith cirrhosis (Abstr.) Clin. Res., 9, 177.

Dingman, J. F., Finkenstaedt, J. T., Laidlaw, J. C., Renold, A. E.,Jenkins, D., Merrill, J. P.. and Thorn, G. W. (1958). Influenceof intravenously administered adrenal steroids on sodiumand water excretion in normal and Addisonian subjects.Metabolism, 7, 608-623.

Feichter, R. A., Summerskill, W. H. J., and Wakim, K. G. Un-published data.

Giuseffi, J., Werk, E. E., Jr., Larson, P. U., Shiff, L., and Elliott,D. W. (1957). Effect of bilateral adrenalectomy in a patientwith massive ascites and postnecrotic cirrhosis. New Engi.J. Med., 257, 796-803.

Hecker, R., and Sherlock, Sheila (1956). Electrolyte and circulatorychanges in terminal liver failure. Lancet, 2, 1121-1125.

Henley, K. S., Streeten, D. H. P., and Pollard, H. M. (1960).Hyperaldosteronism in liver disease: The treatment ofrefractory ascites by adrenalectomy and by the administrationof spirolactones. Gastroenterology, 38, 681-689.

Kagawa, C. M., Cella, J. A., and van Arman, C. G. (1957). Actionof new steroids in blocking effects of aldosterone anddeoxycorticosterone on salt. Science, 126, 1015-1016.

Laragh, J. H., Heinemann, H. 0., and Demartini, F. E. (1958).Effect of chlorothiazide on electrolyte transport in man:Its use in the treatment of edema of congestive heart failure,nephrosis, and cirrhosis. J. Amer. med. Ass., 166, 145-152.

Liddle, G. W. (1958). Aldosterone antagonists. A.M.A. Arch. intern.Med., 102, 998-1004.

Linton, R. R. (1956). Portal hypertension. In Schiff, Leon: Diseasesof the Liver, pp. 195-233. Lippincott, Philadelphia.

Luetscher, J. A., Jr., and Johnson, B. B. (1954). Observations onthe sodium-retaining corticoid (aldosterone) in the urine ofchildren and adults in relation to sodium balance and edema.J. clin. Invest., 33, 1441-1446.

Mackby, M. J. (1960). Techniques of portal decompression. Int.Abstr. Surg., 110, 521-526.

Mackie, J. E., Stormont, J. M., Hollister, R. M., and Davidson,C. S. (1958). Production of impending hepatic coma bychlorothiazide and its prevention by antibiotics. New Engl.J. Med., 259, 1151-1156.

Marson, F. G. W. (1954). Total adrenalectomy in hepatic cirrhosiswith ascites. Lancet, 2, 847-848.

McDermott, W. V., Jr. (1960). The double portacaval shunt in thetreatment ofcirrhotic ascites. Surg. Gynec. Obstet. ,110, 457-469.

Owen, E. E., Flanagan, J. F., and Tyor, M. P. (1959). Kidney as asource of blood ammonia: Effect of chlorothiazide. Proc.Soc. exp. Biol. (N.Y.), 102, 696-697.

Papper, S., Belsky, J. L., and Bleifer, K. H. (1959). Renal failure inLaennec's cirrhosis of the liver. I. Description of clinical andlaboratory features. Ann. intern. Med., 51, 759-773.

Ralli, Elaine P., Robson, J. S., Clarke, Delphine, and Hoagland,C. L. (1945). Factors influencing ascites in patients withcirrhosis of the liver. J. clin. Invest., 24, 316-325.

Ratnoff, 0. D., and Patek, A. J., Jr. (1942). The natural history ofLaennec's cirrhosis of the liver. Medicine (Baltimore), 21,207-268.

Read, A. E., Laidlaw, J., Haslam, R. M., and Sherlock, Sheila(1959). Neuropsychiatric complications following chloro-thiazide therapy in patients with hepatic cirrhosis: Possiblerelation to hypokalaemia. Clin. Sci., 18, 409423.

Reynolds, T. B., Geller, H. M., Kuzma, 0. T., and Redeker, A. G.(1960). Spontaneous decrease in portal pressure with clinicalimprovement in cirrhosis. New Engl. J. Med., 263, 734-739.

Salassa, R. M., Mattox, V. R., and Power, M. H. (1958). Effect ofan aldosterone antagonist on sodium and potassium excretionin primary hyperaldosteronism. (Letter to the Editor.) J. clin.Endocr., 18, 787-789.



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

296 William H. J. Summerskill, Bernard F. Clowdus, II, and John W. Rosevear

Schedl, H. P., and Bartter, F. C. (1960). An explanation for andexperimental correction of the abnormal water diuresis incirrhosis. J. clin. Invest., 39, 248-261.

Seligson, David, and Hirahara, Katsugi. (1957). The measurementof ammonia in whole blood, erythrocytes, and plasma. J. Lab.clin. Med., 49, 962-974.

Shaldon, S., McLaren, J. R., and Sherlock, S. (1960). Resistantascites treated by combined diuretic therapy (spironolactone,mannitol, and chlorothiazide). Lancet, 1, 609-613.

Sherlock, Sheila (1958). Diseases of the Liver and Biliary System,2nd ed. Blackwell, Oxford.

Stormont, J. M., Crabbb, J., Fast, B. Wolfe, S. J., and Davidson,C. S. (1959). The effect of prednisone and amphenone onfluid and electrolyte balance and on aldosterone excretion ofpatients with cirrhosis and ascites. J.Lab. clin. Med. ,53, 396-416.

Summerskill, W. H. J., and Crabb6, J. (1957). Effect of amphenonetherapy on urinary excretion of aldosterone and sodium inhepatic cirrhosis with ascites. Lancet, 2, 1091-1095.

Clowdus, B. F., and Casey, T. H. (1960a). Clinical and metabolicchanges during the development of azotemia in hepatic failurewith ascites. Proc. Mayo Clin., 35, 783-791.Davidson, C. S., Dible, J. H., Mallory, G. K., Sherlock, S.,Turner, M. D., and Wolfe, S. J. (1960b). Cirrhosis of theliver: A study of alcoholic and non-alcoholic patients in Bostonand London. New Engl. J. Med., 262, 1-9.

Talso, P. J., Spafford, Norma, Ferenzi, G., and Jackson, Helen 0.(1956). Paradoxical hyponatremia associated with congestiveheart failure and with cirrhosis of the liver. Metabolism, 5,58-69.

Walker, R. M. (1957). Portacaval anastomosis. Lancet, 1, 57-59.Webster, L. T., Jr., and Gabuzda, G. J. (1959). Relation of azotemia

to blood "ammonium" in patients with hepatic cirrhosis.A.M.A. Arch. intern. Med., 103, 15-22.

Welch, C. S., Welch, H. F., and Carter, J. H. (1959). The treatnmentof ascites by side to side portacaval shunt. Ann. Sierg., 150.428-440.



j.com/


ecember 1961. D

ownloaded from

http://gut.bmj.com/

long-termmedical management and complications of 'resistant' ascites · gut, 1961, 2. 285...

Documents