351

During the last 3 years of this study a controlled clinicaltrial was undertaken to decide whether anticoagulantswould help the patient with ingravescent infarction (over2 hours) or signs of increasing paralysis. The best resultswere in patients who were conscious and whose lesionwas incomplete when treatment was begun.

It is a pleasure to thank my colleagues, Dr. K. D. Keele andDr. J. E. Royds, for allowing me access to their patients, and for theirvaluable suggestions regarding this article.

REFERENCES

Adams, R. D. (1958) Trans. Conf. Amer. Hear Ass. p. 23.Alajouanine, T. (1957) Rapp. Congr. franç. Méd. 1, 130.Brain, W. R. (1957) Lancet, ii, 857.Carter, A. B. (1957) Quart. J. Med. 26, 335.

— (1959) ibid. 28, 125.de Morsier, G., Tissot, R. (1957) Rapp. Cong. franç. Méd. 1, 110.Elkington, J. St. C. (1958) Lancet, ii, 327.Fisher, C. M. (1958) Neurology, 8, 311.Glynn, A. A. (1956) Brit. med. J. i, 1216.Hausman, L. (1958) Trans. Conf. Amer. Heart Ass. p. 30.Hicks, S. P., Warren, S. (1951) A.M.A. Arch. Path. 52, 403.Hutchinson, E. C., Yates, P. O. (1957) Lancet, i, 2.Kelly, R. E. (1958) Proc. R. Soc. Med. 51, 209.Kety, S. S., Schmidt, C. F. (1948) J. clin. Invest. 27, 484.Leriche, R., Fontaine, R. (1934) Pr. méd. 42, 849.Livingston, K. E., Escorbar, A., Nichols, G. D. (1955) J. Neurosurg. 12, 336.McKissock, W., Richardson, A., Walsh, L. (1959) Lancet, ii, 683.Meyer, J. S., Denny-Brown, D. (1957) Neurology, 7, 447.Moyes, P. D., Millikan, C. H., Wakim, K. G., Sayre, G. P., Whisnant, J. P.

(1957) Proc. Mayo Clin. 32, 124.Murphy, J. P. (1954) Cerebrovascular Diseases. Chicago.Peterman, A. E., Wakim, K. G., Sayre, G. P., Whisnant, J. P., Millikan,

C. H. (1959) J. Neuropath. 18, 263.Sibley, W. A., Morledge, J. H., Lapham, L. W. (1957) Amer. J. med. Sci.

234, 663.Symonds, C. P. (1957) Modern Trends in Neurology. London.Vastola, E. F., Frugh, A. (1959) Neurology, 9, 143.Wells, C. E. (1959) Arch. Neurol. Psychiat. 81, 667.Wright, I. S., McDevitt, E. (1954) Lancet, ii, 825.

CLINICAL COURSE OF UNCOMPLICATED

ACUTE TUBULAR NECROSIS

LAVINIA W. LOUGHRIDGEM.B. Belf., M.R.C.P.REGISTRAR IN MEDICINE

M. D. MILNEM.D., B.Sc. Manc., F.R.C.P.

LECTURER IN MEDICINE

R. SHACKMANM.B. Lond., F.R.C.S.READER IN SURGERY

I. D. P. WOOTTONM.A., M.B. Cantab., Ph.D Lond.

SENIOR LECTURER IN CHEMICAL PATHOLOGY

From the Departments of Medicine, Surgery, and Chemical Pathology,Postgraduate Medical School, Ducane Road, London, W.12

ACUTE oliguric renal failure is present when the volumeof urine produced by the kidneys is insufficient for main-tenance of the normal body internal environment. Inchronic renal failure the clinical and biochemical statusof the patient may change little for long periods, some-times for several months or even years. By contrast, inacute renal failure there is progressive and continuousdeterioration unless diuresis occurs or some form of

dialysis is applied.The natural history of cases of acute oliguric renal

failure is extremely varied because the condition may becaused by many unrelated types of disease; for example,there may be lesions of the major or minor renal arteriesor veins, glomerular lesions of acute glomerulonephritisor polyarteritis nodosa, proximal tubular necrosis dueto nephrotoxic drugs or poisons, ischasmic tubular lesionsincluding acute tubular necrosis, renal cortical necrosis,acute papillary necrosis, or an acute exacerbation ofchronic renal disease or bilateral ureteric obstruction.The commonest cause of acute oliguric renal failure is

acute tubular necrosis; and this is singularly important

because, if the patient survives, recovery is accompaniedby full restitution of renal function. Uncomplicated acutetubular necrosis is usually a disease of predictable course,and, if properly treated, is relatively benign. When the con-dition is complicated by infection or by associated injuries,whether accidental or surgical, it is much less predictableand much more serious. Teschan et al (1955) emphasisedthis difference as a result of their experience of acutetubular necrosis in the injured and wounded in theKorean campaign. The course of the disease in their

patients, particularly those with severe injuries to volun-tary muscle, proved to be much more fulminant than inthe average civilian patient without such complications.A similar difference is apparent in the series of civiliancases reported recently by Parsons and McCracken (1959)and by Bluemle et al. (1959). The proportion of fatalcases was less than 25% in patients in whom the acutetubular necrosis was due to medical or obstetric causes,but more than 70% when it was complicated by trauma.Biochemical deterioration is more rapid in the surgicalgroup. There is frequently a dramatic rise of plasma-potassium, uncontrolled by enteral cation-exchangeresins, and there is a rapid fall of plasma-bicarbonate dueto urxmic acidosis. In a series of 220 cases of acute

oliguric renal failure of all types referred to HammersmithHospital for dialysis by an artificial kidney, the prognosishas been good where acute tubular necrosis was due to amedical or obstetric cause. In contrast, the prognosishas been much worse in those patients in whom acutetubular necrosis was secondary to severe injuries or

surgical operations. A high proportion of such patients

TABLE I-RECOVERED CASES OF ACUTE TUBULAR NECROSIS

352

died of their injuries or from complicating infection

despite carefully controlled conservative treatment and inmany cases repeated hsemodialysis.We describe here the natural history of uncomplicated

acute tubular necrosis in a selected group of 28 patientsin whom the primary cause of the disease was medicalor obstetric (table i). All of these patients recoveredcompletely with or without hxmodialysis. This selectionavoids the difficulties due to the irregular and unpredict-able course when the condition is complicated by uncon-trolled infection or extensive tissue damage, and permitsanalysis of the full course of the disease from onset tofull recovery. The series may be unrepresentative of allexamples of uncomplicated acute tubular necrosis, sinceprobably the more severe cases are referred to centresequipped for haemodialysis. This difficulty of sampling is,however, almost universal in descriptions of diseasebased on the experience of a single hospital. 5 patientswith proximal tubular necrosis have been included sincetheir course seemed to be identical with that of acutetubular necrosis.

During this period deaths in cases diagnosed on

admission as uncomplicated acute tubular necrosisnumbered 11 (table II). 3 of these patients were provedat necropsy to have renal cortical necrosis with almostcomplete destruction of all nephrons. The fatal cases willbe described after the analysis of the 28 patients whorecovered.

ManagementAll patients were treated by a standard conservative

regime. The diet consisted of a 20% solution of lactoseby mouth, a volume of X + 400 ml. daily being given fromNovember to April, and X+500 ml. daily from May toOctober, where X ml. is the total volume of the measurablefluid loss from the patient on the previous day. Whenintractable urxmic vomiting occurred, the same volumeof sterile 40% glucose was given by infusion, preferablyinto the superior vena cava, but at times into the inferior

vena cava, instead of by mouth. Soluble insulin was givenby injection when the blood-sugar was over 200 mg. per100 ml. Sulphonamides or antibiotics were given onlyin cases with established infection, but all patients wereisolated and standard methods of barrier nursing wereused. Infusions of packed cells were given in amountslarge enough to maintain the hxmoglobin concentrationabove 10.2 g. per litre (70% Haldane). Non-virilisingandrogens in the dosage recommended by McCrackenand Parsons (1958) were given to the female patientsduring the latter half of the series. Blood-urea and

plasma potassium and bicarbonate were measured daily,and electrocardiography was performed daily throughoutthe whole of the oliguric stage. Hxmodialysis wascarried out by a rotating-drum artificial kidney whenany one of the following indications was present: (a) blood-urea above 400 mg. per 100 ml.; (b) plasma-bicarbonatebelow 12 mEq. per litre; (c) plasma-potassium above7-5 mEq. per litre, despite therapy with cation-exchangeresins by mouth (Elkinton et al. 1950, Evans et al. 1953);(d) electrocardiographic evidence of intraventricularblock with widening of the QRS complexes from hyper-kalæmia; (e) urxmic stupor, coma, vomiting, hiccough,or fits.

Appropriate supplements of sodium and potassium saltswere given in the diuretic stage, usually in amounts equalto the urinary losses of the previous day. A diet of

carbohydrate alone was continued until the blood-ureafell below 100 mg. per 100 ml., unless the patient becameextremely hungry and demanded solid food.

StagesThe course of the disease, as suggested by Bull et al.

(1950), has been divided into an onset stage, oliguricstage, early diuretic stage, late diuretic stage, and conva-lescent stage; but the definitions of the five stages adoptedhere are more precise and better adapted for clinical use:

Onset stage.—This is the short period between the acuteevent causing the renal damage and the development of

TABLE II-FATAL CASES OF ACUTE TUBULAR NECROSIS AND RENAL CORTICAL NECROSIS

353

Fig. 1-The rise of blood-urea during the oliguric stage of acutetubular necrosis before haemodialysis or in undialysed patients.The rate of rise of blood-urea is greater in the first two days. The

regression line is calculated from days 2 to 7 inclusive, as after day 7many of the higher blood-urea values have been reduced byhxmodialysis.

established oliguria with a urine output in an adult patient ofless than 400 ml. per day.

Oliguric stage.-This is the interval between the onset stageand the first day when the urine volume is consistently morethan 400 ml. per day.Early diuretic stage.-This is the period between the end of

the oliguric stage and the first day when there is an unequivocaland sustained fall of the blood-urea.

Late diuretic stage.-°This is the period between the end ofthe early diuretic stage and the first day on which the blood-urea is consistently lower than the upper limit of normal-i.e45 mg. per 100 ml.

Convalescent stage.—This is the period in which, althoughthe blood-urea is normal, the patient is still unable to engagein normal activities or duties.

In this account of the clinical course of uncomplicatedacute tubular necrosis, the daily urine volume and thevalues of blood-urea are considered to be the two most

important biochemical indices. Although concentrationsof plasma potassium and bicarbonate are of especialvalue in assessing the need for hæmodialysis, they mayboth be influenced by relatively minor therapeuticprocedures such as the administration of cation-exchangeresins or infusions of sodium lactate. Their values do not

necessarily therefore reflect the course of the disease. In

contrast, the blood-urea will rise continuously unlessspontaneous diuresis occurs or hasmodialysis is applied.Onset Stage

Since all patients were referred from other hospitalsearly in the oliguric stage there was unfortunately noopportunity to study the patients in this important stageof the disease.

Oliguric StageComplete anuria is extremely rare in acute tubular

necrosis. Urethral catheterisation was avoided in orderto prevent infection, and it is therefore impossible togive the precise urinary volume produced during eachday of the oliguric stage. The calculated mean dailyvolume during the first six days was 75 ml. and the meanof the whole oliguric stage was 150 ml., suggesting a slowincrease of urinary volume during this stage of the disease.The mean duration of the oliguric stage was 12.5 days

(s.D. 5-0 days), which is similar to the mean durationreported by Bull et al. (1950) and Swann and Merrill

(1953). The rate of rise of blood-urea was greater duringthe first two days of the oliguric stage than in the suc-ceeding days (fig. 1), the mean blood-urea rising to

200 mg. per 100 ml. by the end of the second day. This

probably reflects a temporary increase of catabolism witha greater production of urea by breakdown of tissueproteins due to the initial insult. In addition, the rateof loss of urea in the small amounts of urine producedand in sweat, vomit, and faeces is roughly proportional tothe concentration of urea in the blood. After the second

day, the rate of rise of blood-urea was reduced by treat-ment, and averaged 34 mg. per 100 ml. per day. In some

patients the rate of rise of blood-urea was not greaterthan 25 mg. per 100 ml. per day, which represents thetheoretical minimum assuming a low endogenous proteinbreakdown of 40 g. per day (Gamble 1951). In 3 patientsthe rate of rise of blood-urea was unusually great andhxmodialysis was required early in the oliguric stage.1 patient had an acute haemolysis due to arsine poison-ing, another had severe staphylococcal pneumonia, whilstthe 3rd had leptospirosis with considerable liver damage.In all 3 cases there was probably an increased rate ofprotein breakdown from the associated condition.Of the 28 patients 13 were treated by hxmodialysis:

11 had a single treatment, and 2 were dialysed on twooccasions. After the fifteen dialyses (twelve in the

oliguric stage and three in the early diuretic stage) thesubsequent daily rise of blood-urea continued unchanged(fig. 2). Hæmodialysis causes the biochemical changes torevert to those found earlier in the oliguric stage, and hasno other influence on the course of the disease. Therate of rise of blood-urea on the first day after hæmo-

dialysis is usually greater than on succeeding days(fig. 2). This probably reflects equilibration of therelative excess of urea in the intracellular compartmentwith that in the extracellular fluid and plasma. The arti-ficial kidney clears urea first from the blood andextracellular fluid.

Early Diuretic StageDivision of the diuretic stage into an early and late

stage is of great clinical value because the amount of ureaexcreted during the early diuretic stage is approximatelyequal to that formed from endogenous protein breakdown,and there is usually no fall of the blood-urea. Indeed,during the early diuretic stage the clinical and biochemical

Fig. 2—The rise of blood-urea during the oliguric stage of acutetubular necrosis after haemodialysis.The rate of rise is greater on the first day after dialysis. Sub-

sequently the mean rise is identical with that shown in fig. 1.

354

Fig. 3--Urinary volume per day in the diuretic stage of acute tubularnecrosis. Continuous lines show the results from the presentseries, interrupted lines the results of a similar series treated byBull et al. with a greater fluid intake during the oliguric stage,with mean and standard deviations in each case.

The mean volume in the series of Bull et al. is consistently greateruntil the 17th day of the diuretic stage.

status of the patient does not improve and frequentlycontinues to deteriorate. In the late diuretic stage, how-

ever, when the blood-urea falls, there is steady andprogressive improvement. Haemodialysis may have to becarried out in the early diuretic stage, but is rarely, if

ever, indicated in the late diuretic stage. As soon as thedaily urinary volume consistently exceeds 400 ml. thereis usually a steady daily increase at a mean rate of 280 ml.per day until a volume of about 2000 ml. per day isreached after five or six days (fig. 3). Thereafter the mean

daily urinary volume is stabilised at 2000-2500 ml. Themean duration of the early diuretic stage was 4-7 days(S.D. 3.1), and stabilisation of the urinary volume roughlycoincided with the day on which the blood-urea startedto fall.

Late Diuretic Stage .

Urine volumes in the late diuretic stage are now lessthan in cases treated ten to twelve years ago (Bull et al.1950). It was then customary to give patients morefluid during the oliguric stage-usually 1000 ml. morethan the total fluid loss on the previous day. In fig. 3the mean urinary volumes in the present series are

compared with those in a similar series of cases of acutetubular necrosis reported by Bull et al. (1950). Themean urinary volumes are consistently higher in the latterseries until the seventeenth day of the diuretic stage(X2=95; p < 0001). If it is assumed that the patients ofBull et al. received an excess fluid intake of 500 ml. each

Fig. 4-The actual fall of blood-urea in 20 cases of acute tubularnecrosis during the late diuretic stage. Crosses indicate casestreated by haemodialysis.

average total excess given to each patient was 6 litres.Comparison of the urinary volumes in fig. 3 shows thatthe patients of Bull et al., on an average, each excreted6-5 litres more urine than the patients of the presentseries during the first seventeen days of the diuretic stage.This suggests that the greater diuresis in the patients ofthe earlier series was largely due to the higher fluid intakeduring the oliguric stage. It has been claimed (Black1957) that high urinary volumes in the diuretic stage ofacute tubular necrosis are chiefly due to osmotic diuresis,the osmotic load per single functioning nephron beingabnormally high. The observations recorded here

suggest, however, that at least part of the diuresis in

patients treated in the past was due to water diuresisfrom prior over-hydration. Whereas an individual withnormal renal function would excrete an excess body-watercontent of 6 litres within twenty-four hours, a patientin the diuretic stage of acute tubular necrosis has grossimpairment of renal function (Swann and Merrill 1953)and requires over two weeks to excrete excess water which

Fig. 5-The fall of blood-urea during the late diuretic stage of acutetubular necrosis. The logarithmic regression-line and limits attwice the standard deviation are shown.

could be eliminated by healthy kidneys within a singleday. An excess body water content of 6 litres is equivalentto 13% of the normal total body-water and could causeconsiderable water intoxication.

During the late diuretic stage there is a progressiveexponential fall in the blood-urea (figs. 4 and 5). On an

average, the blood-urea becomes normal within seventeendays (fig. 5).Convalescent StageBy the end of the late diuretic stage the patient has

passed through a long and serious illness which has lastedon an average five weeks (fig. 6). During this periodthere has been little or no intake of protein. Althoughmodern conservative therapy reduces endogenous proteinbreakdown, often to a value as low as 40 g. per day, fiveweeks of protein starvation causes a loss of at least 1-5 kg.of body-protein-about an eighth of the total body con-tent. The majority of this endogenous protein is derivedfrom voluntary muscle. The patient at the start of theconvalescent stage is still an invalid, characteristicallyweak and often unable to stand or walk unaided. Further

recovery, however, is very rapid; the patient usuallydevelops an insatiable appetite and quickly gains weightwith increase of muscle mass and restoration of muscle

355

Fig. 6-Graphical representation of an average case of acute tubularnecrosis which summarises the previous charts.The regression-lines of blood-urea and the upper limit at twice the

standard deviation for the oliguric and late diuretic stages are notcontinuous in this series, because many of the patients were treatedby h&aelig;modialysis in the last few days of the oliguric stage.

strength. Usually it is at least two or three months fromthe start of the illness before the patient is well enough toresume full activities and duties.

CLINICAL COURSE OF FATAL CASES

Of the 11 fatal cases (table II) originally diagnosed asuncomplicated tubular necrosis, all were examined at

necropsy. 3 cases, all due to concealed accidental h&aelig;mor-rhage of pregnancy, were found to have had severe

bilateral renal cortical necrosis causing irreversible renaldamage. We prefer to avoid percutaneous renal biopsyduring the oliguric stage of acute renal failure, and there-fore have no means of differential diagnosis betweenrenal tubular necrosis and renal cortical necrosis exceptthe response to therapy. The apparent mortality in thisseries was therefore 28%, but the true mortality excludingthe cases of renal cortical necrosis was 22%-a figuresimilar to those in other reported series.Of the 8 fatal cases of acute tubular necrosis, 4 died

in the oliguric stage, 2 in the early diuretic stage, and 2 inthe late diuretic stage. They showed no obvious bio-chemical differences from the patients who survived, andthe oliguric stage of the 4 patients who died in thediuretic stage was not unusually long. Table II shows thatsepsis, usually from organisms resistant to most availableantibiotics, was the main cause of death.Of the 3 fatal cases of renal cortical necrosis 2 also

died as a direct result of infection, although the severityof their renal lesion would have precluded final recovery.2 of these 3 patients died in the oliguric stage, on theninth and eighteenth days. The 3rd patient passed intoan early diuretic stage on the twentieth day of her illness,and the urine output rose to a maximum of 970 ml. perday before death on the twenty-ninth day. This showsthat a guarded prognosis should be given in all cases ofacute oliguric renal failure due to concealed accidentalhemorrhage of pregnancy, unless the patient has passedinto the late diuretic stage with a steadily falling blood-urea. Even if percutaneous renal biopsy in such casesshowed only acute tubular necrosis, this would not precluderenal cortical necrosis in other parts of the kidneys.

Discussion and SummaryThe average course of uncomplicated acute tubular

necrosis is summarised in fig. 6. Average durations of the

first four stages of the disease are plotted on the abscissa,and the blood-urea and mean urinary volumes on the ordi-nate. The regression-lines for blood-urea and the upperlimits at twice the standard deviation for the oliguricand diuretic stages are not continuous because almosthalf the patients were dialysed at the end of the oliguricstage. Haemodialysis artificially reduces the blood-ureaat the start of the early diuretic stage. It is almostcertain that some of the patients would not have survivedhad h&aelig;modialysis not been applied. Since the rate ofclimb of the blood-urea during the oliguric stage isidentical before and after dialysis (figs. 1 and 2), it is

permissible to extrapolate the rise of blood-urea after

dialysis to that before treatment in order to estimate theeventual level had dialysis not been carried out. In the 2

patients who required dialysis on two separate occasions,the blood-urea would have been expected to rise to

1020 and 1110 mg. per 100 ml. Although in isolatedcases recovery has followed the recording of similar levelsof blood-urea, most such patients will die without

h&aelig;modialysis.This paper may prove useful to those responsible for

the decision to transfer cases of acute tubular necrosisto a centre with facilities for h&aelig;modialysis. Although theduration of the oliguric stage in any single case remainsunpredictable, the course of the oliguric stage may beaccurately predicted after only a few days’ carefulobservation. The course of cases with severe tissue

damage cannot be predicted. Deterioration may 15e

extremely rapid, and in these cases a high plasma-potassium or a low plasma-bicarbonate may provideurgent reasons for haemodialysis before the blood-ureaexceeds 400 mg. per 100 ml.Thanks are expressed to Prof. G. M. Bull for permission to publish

urinary volumes of patients under his care, and to Prof. J. McMichaeland Prof. E. J. King for advice.

REFERENCES

Black, D. A. K. (1957) Essentials of Fluid Balance. Oxford.Bluemle, L. W., Jr., Webster, G. D.,Jr., Elkinton, J. R. (1959) A.M.A. Arch.

intern. Med. 104, 180.Bull, G. M., Joekes, A. M., Lowe, K. G. (1950) Clin. Sci. 9, 379.Elkinton, J. R., Clark, J. K., Squires, R. D., Bluemle, L. W., Jr., Crosley,

A. P. (1950) Amer. J. med. Sci. 220, 547.Evans, B. M., Jones, N. C. H., Milne, M. D., Yellowlees, H. (1953) Lancet,

ii, 791.Gamble, J. L. (1951) Chemical Anatomy, Physiology, and Pathology of

Extracellular Fluid. Cambridge, Mass.McCracken, B. H., Parsons, F. M. (1958) Lancet, ii, 885.Parsons, F. M., McCracken, B. H. (1959) Brit. med. J. i, 740.Swann, R. C., Merrill, J. P. (1953) Medicine, Baltimore, 32, 215.Teschan, P. E., Post, R. S., Smith, L. H., Jr., Abernathy, R. S., Davis, J. H.,

Gray, D. M., Howard, J. M., Johnson, K. E., Klopp, E., Mundy, R. L.,O’Meara, M. P., Rush, B. F., Jr. (1955) Amer. J. Med. 18, 172.

CRYPTOGENIC LIVER ABSCESS

J. F. STOKESM.D. Cantab., F.R.C.P.

PHYSICIAN TO UNIVERSITY COLLEGE HOSPITAL, LONDON, W.C.1

LIVER abscess is uncommon in Great Britain. Between1950 and 1958 only 8 cases have occurred in 14,500admissions to the Hospital for Tropical Diseases (U.C.H.),giving an incidence of 0-055%. All these were shown tobe amoebic in origin.With a chronic abscess in a patient at risk for amoebiasis

it is sometimes difficult to be certain whether am&oelig;b&aelig;were initially responsible for it and whether their presencehas been obscured by superadded bacterial infection.There is a growing feeling, however, that pyogenic infec-tion is becoming more important as a primary cause evenin tropical practice (Worms 1956), and this is supportedby McFadzean et al.’s (1953) figures from Hong Kong.During the decade 1949-58, 6 cases of liver abscess have