fortyyears of nephrosis childhood* · characteristics, namely oedema, gross proteinuria, lowered...

FORTY YEARS OF NEPHROSIS IN CHILDHOOD*BY

DAVID LAWSON, ALAN MONCRIEFF and WILFRID W. PAYNEFrom The Hospital for Sick Children, Great Ormond Street, and the Institute of Child Health, London

(RECEIVED FOR PUBLICATION JUNE 19, 1959)

This article is an attempt to relate the present-dayoutlook on nephrosis in childhood with the 'naturalhistory' of the disorder as it was seen at The Hospitalfor Sick Children, Great Ormond Street, in the daysbefore modern treatment was available. The firstperiod of the study is the responsibility of one of us(D.L.) and covers the years 1917-38. During thewar years in-patient care of the children attendingthis hospital was widely dispersed and records arenot always available. With the opening up of thehospital to more than a handful of patients in 1945and the development of a new index system thestudy starts again and covers up to 1956-7. Thisperiod falls into two parts: for 1945-51 it hadbecome possible to prevent and treat infection inchildren with nephrosis by means of sulphona-mides and antibiotics; from 1951 steroids becameavailable, at first in limited supply for selected cases,and in 1955 freely available for all cases as required.The Hospital for Sick Children, Great Ormond

Street has a large turn-over of patients and serves awide area. This means that large numbers of caserecords have to be read and details checked beforea child can be accepted as qualifying for inclusionin the series. It also means that follow-up ofchildren, especially after they have passed the agefor attendance at the hospital, presents manydifficulties.

DefinitionsFor the purpose of this study the term nephrosis is

applied to a syndrome presenting four essentialcharacteristics, namely oedema, gross proteinuria,lowered plasma albumin and raised blood choles-terol.The disorder presents, usually in the absence of

any constitutional disturbance or other sympto-matology, with the unheralded onset of generalizedoedema. In severe cases the child becomes univer-sally waterlogged in the course of a few days,

* A summary of this paper was given at the IX InternationalPaediatric Congress, Montreal, in July, 1959.

although a slower but steady increase in oedemaover the course of two or three weeks is commoner.The urine, which has usually been unremarkable inquality, although often moderately diminished inquantity, is found to contain from 500 to 2,000 mg.%of protein but only a moderately increased red cellcontent. Addis counts of up to 10 million are notuncommon but the haematuria seldom amounts tomore than this and is almost never macroscopic.The serum cholesterol is usually raised at the 'onset'but, if not, it steadily rises in the course of a fewweeks to levels between 400 and 2,000 mg. %. Theplasma albumin concentration is always markedlydecreased and may be below 1 %.An infection often precedes the clinical onset but

it is seldom clearly streptococcal and is more oftenin the nature of a common cold or bronchitis.Subsequently, it is often found that such infectionswill precipitate a return of, or an increase in thedegree of, oedema and it is easy to understand thatsuch an episode may first bring the hitherto sub-clinical disorder to light. The oedema may lastfor a few weeks only or may persist over monthsor years with alternating remission and relapse.When onset or relapse can be closely studiedbiochemically, it is found that the oedema is the lastof the signs to appear in relapse and the first todisappear in remission, with the possible exceptionof proteinuria. After such a variable course all themanifestations may clear and the disorder mayappear to be healed. Death may occur in periodsof severe oedema from biochemical and water-balance disturbances or from overwhelming bacterialinfection to which these children are particularlysubject; or there may be a gradual or rapid super-imposition, at any phase, of the features of general-ized renal failure and hypertension with final deathin uraemia. The transient occurrence of nitrogenretention or hypertension, particularly common inperiods of increasing oedema, does not howeveritself presage such a termination.

In the 1945-57 series clinical notes and biochemical115 2

copyright. on O

ctober 8, 2020 by guest. Protected by

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.35.180.115 on 1 April 1960. D

ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOODrecords have been fairly complete; most of the cases

could be followed up and re-examined and all werewithin the memory of authors or colleagues.The extraction and definition of the 1917-38

series was more difficult because (1) records, and inparticular biochemical investigations, were lesscomplete; (2) there were no existing follow-upstudies as out-patient records for the period did notsurvive the Second World War; and (3) the unityof this disease was not recognized at the time andcases were found under a wide range of diagnosticheadings. Biochemical criteria were applied wher-ever possible, but in many cases a subjective elementhad to be introduced.When the question of differentiation between

nephrosis and acute glomerulo-nephritis arose, mainreliance was placed on the following contrastingfeatures in the natural history of acute glomerulo-nephritis: (1) sudden onset with urinary signs, oftenwithin 10-14 days of a streptococcal infection;(2) transient nature of the oedema; (3) rapid rise andfall in any transient initial nitrogen retention andhypertension; and (4) normal blood cholesterol andplasma protein at clinical onset.

ClassificationFor the purpose of analysis the cases have been

placed in one of the four following groups:Inactive. This means the absence of the four

essential criteria already quoted. Albuminuriamust not exceed 50 mg. per 100 ml. The electro-phoretic pattern of the plasma proteins is normal.

Latent. There is no oedema but albuminuria isabove 50 mg. per 100 ml. and blood urea andcholesterol levels may or may not be normal.

Active. This group consists of cases with thecomplete, potentially reversible nephrotic syndromejust described, who may have transient increases intheir blood urea and blood pressure, and also caseswho have progressed to the irreversible uraemicstage, as judged by a persistently high and increasingblood urea and an increasing blood pressure.

Dead. The cause may not always be directlyrelated to the nephrosis.The second group is the most controversial and

must be further discussed. By oedema in a follow-up series is meant essentially the pre-tibial oedemaof the child, otherwise apparently well, who isattending the out-patient department, not the water-logged condition of the in-patient which fluctuatesso much and not always for obvious reasons. Thereis no simple way of screening patients other than bylooking for the pre-tibial oedema. All such caseswill have some degree of albuminuria. When theoedema goes the child is clinically normal, but there

may or may not be albuminuria of over 50 mg. per100 ml. If there is no such albuminuria then thechild will almost certainly come into the inactivegroup provided that the blood chemistry is normal.Strictly speaking therefore inactive means clinicallynormal (i.e. no oedema) and chemically normal.The latent group seems to contain three sub-divisions: (a) clinically normal but still albuminuriaof over 50 mg. per 100 ml.; (b) clinically normal butboth albuminuria and abnormal chemistry; and(c) clinically normal, no albuminuria but bloodchemistry not quite normal. It is recognized thatthis latent stage may persist, but it only appears in avery small number of the patients whose recordshave been analysed in the 1945-57 series and thetransfer of such patients to either the inactive oractive groups will not affect the main conclusions.

Description of Search (1917-38 Series)The in-patient case records for this period are

bound in volumes at the end of which is a diagnosticindex prepared by the registrar. All these indiceswere searched and every case history examined inwhich there appeared the remotest possibility that itmight be a case of nephrosis. These diagnosticterms included every case in which the word neph-ritis of any kind was used and any in which the wordsoedema or toxic oedema were used. This resulted ina close study of the records of 1,500 cases. Thegreatest care was taken to make this selection asaccurate as possible, but in certain cases recoursehad to be made to something closer to intuition thanto scientific study in view of the paucity of records.The author (D.L.) who carried out this part of thework is however satisfied that the diagnosis wasalmost certainly correct in every case included andany doubtful ones were discarded.

Analysis. The total number of cases found inthis way was 78. Of these 18 had died during thefirst or subsequent admission and 60 had beendischarged and lost from the records (Table 1).

TABLE 1SEARCH AND FOLLOW-UP (1917-38 SERIES)

In-patients' Records Searched .1,500

Accepted as nephrosis .78Died in hospital .18Follow-up attempted 1953 .60Follow-up failed 1953.44Followed up 1953 16

An attempt was made in 1953 to trace these60 cases. Letters were sent to them at the recordedaddresses and if no reply was received a furtherletter was sent. In a number of cases the patientscould be traced at these addresses and in others

116

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

FORTY YEARS OF NEPHROSIS IN CHILDHOOD

information was given by the present occupierwhich led to their successful tracing subsequently.This search was completely unsuccessful in 44 of the60 cases. Sixteen were successfully traced and ofthese five had died (Table 2).

TABLE 2ANALYSIS OF 16 CASES SUCCESSFULLY FOLLOWED UP IN

1953 (1917-38 SERIES)

Died .. .. .. .. 5

Alive and well .. .. 11

Active .0Latent .1Inactive .. . 8Unclassified .. 2 (in perfect health but

not examined)

The remaining 11 were asked to come to theout-patient department for clinical and biochemicalexamination and nine did so. None of these hadhad any clinical relapse since his admission tohospital in childhood. All were in perfect healthand none showed any clinical or biochemical evi-dence of activity or of sequelae of any kind, exceptfor one woman who had borne children withoutmishap but whose urine contained 100 mg.y% ofprotein. This technically places her in the latentrather than the inactive group. The remaining twowere unable to come but sufficient informationcould be obtained to ensure that they were inperfect health.

Causes of Death (1917-38 Series). The 23 knowndeaths in this series have been classified as follows(see Table 3):

TABLE 3ANALYSIS OF CAUSES OF DEATH (1917-38 SERIES)

No. of Cases

Bacterial infection 9 (includes 1 measles)Treatment .. .Uraemia .. 4Water and electrolyte disturb-

ance (includes D. and V.) 3Diphtheria.1Unknown.1

Total .23

(1) INFECTION (nine cases). In all cases this was abacterial infection with the exception of one childwho died in a severe attack of measles. In themajority of cases the infection took the form ofcellulitis, peritonitis and pneumonia, often accom-panied by a septicaemia. There have been excludedfrom this heading infections which were the directresult of treatment.

(2) DISTURBANCES OF WATER AND ELECTROLYTEBALANCE (three cases). This includes children whodied from overwhelming oedema and also those in

whom the cause of death was given as 'gastro-enteritis'.

(3) URAEMiA (four cases). Children who died inchronic uraemia with or without hypertension.

(4) DEATH DUE TO TREATMENT (five cases). Thisincludes two children who died suddenly followingover-rapid paracentesis, one who died of post-operative pneumonia following a renal biopsy, onewho died with convulsions without previous nitrogenretention in the post-operative period following adecapsulation operation, and one who died followingdecapsulation, paracentesis and subsequent woundinfection.

(5) UNCLASSIFIED (two cases). One appears tohave been in perfect health for some years followingdischarge but died of diphtheria, and one died afterdischarge but the cause of death is unknown.Of 21 cases in which the cause of death is known,

therefore, all but four are deaths which couldprobably have been avoided had modern techniquesof fluid and electrolyte balance and modern chemo-therapeutic agents been available, and if methods nowknown to be useless and dangerous had not been used.

Estimate of Mortality (1917-38 Series). Of thetotal of 78 cases 18, or 23%, died during an initialor subsequent admission to The Hospital for SickChildren. In view of the incompleteness of thefollow-up it is not possible to make a very accurateassessment of the subsequent mortality of the caseswho were discharged alive. The first question to beanswered is whether or not it is reasonable to assumethat the 16 cases who could be traced were a repre-sentative sample of the 60 in which the attempt wasmade. This turns on the question of whetherit is likely to be more difficult to trace patients whohave died than to trace those who have survived.

TABLE 4ROUGH ESTIMATE OF MORTALITY (1917-38 SERIES)

No. in series .78Death in hospital 18Deaths after discharge 5 out of 16, equivalent to =37

to 19 out of 60 .19

Mortality approximately 37 out of 78=47%

It is in fact the families rather than the individualpatients who are being traced, and there seems noreason to think that it would be more difficult totrace the family of a child who had died than of onewho had survived, although it is possible that thefamilies of children who have died might be lesswilling to reply if they received the letter. Of the44 untraceable cases, in all but four the follow-upletters were returned by the Post Office marked 'Notknown at this address'. There were only four cases

117

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOODin which letters were not so returned by the PostOffice and had therefore presumably been acceptedat the address given. In these circumstances it isconsidered reasonable to assume that the 16 casestraced may be regarded as a representative sampleof the 60 and the approximate overall mortality of theseries is as follows: Five of the 16 are dead; thereforeof the total of 60 it may be assumed that 19 are dead.To these 19 presumed dead must be added the 18who are known to have died while in hospital,giving an overall mortality of 37/78 or approxi-mately 47% (Table 4).

It therefore appears that in the years 1917-38,before the days of antibiotics and steroids, therecovery rate was of the order of 50% and that of21 cases in which the cause of death is known all butfour (or approximately 20%) were due to causeswhich might have been averted by modem methods ofcontrol of infection and water balance disturbance.

Description of Search and Analysis (1945-57 Series)For this period of the study indexed diagnoses

and a special list held by one of us (W.W.P.) wereaccepted in the first instance, and during the 12years or so covered the number of children classifiedas nephrosis was 123. Of these, five were privatepatients under the care of colleagues, and for variousreasons records were not always complete nor wasfollow-up easy to arrange. They have thereforebeen omitted. Some of the remainder were seen asout-patients and transferred to other hospitals atonce or after a short period. Their records havebeen used only in regard to age of onset and ultimatefate, if known. One of us (A.M.) read all the caserecords and reduced the data to a single pro formasheet. This proved a fascinating but formidabletask. Patients were often in hospital for longperiods or were admitted on several occasions.(The record was held by one child admitted 23times.) When no recent information existed thiswas obtained by seeing the child, by writing to thefamily doctor, to another hospital and so forth.As already mentioned, patients come to the hospitalfrom a wide area and the tracing of them afterdischarge has been extremely difficult.* Altogether

* It may be mentioned in passing that successes in tracing includeddetails obtained and contacts re-established in numerous ways. Thepolice traced the address of one family; another patient was foundafter an announcement of her marriage engagement in The Times;the trail of one began in Wales, followed to a wrong part of EastAfrica, eventually finished in Central Africa and the child was seen inOut-patients when the father was next on leave. (She had actuallymade an apparently complete recovery which early records would nothave suggested.) The Service authorities have been most helpful intracing patients overseas, one going as far as Singapore. In otherareas the health authorities have been most helpful and family doctorsmost willing to assist. One child in Malta, who was an in-patient,has been reported on whenever necessary.

failure has to be recorded for four children in the1945-57 series. One is the child of a United StatesAir Force officer who has returned to the U.S.A.and, despite every possible effort on the part of theauthorities, the present whereabouts are unknown.Three others were all going downhill shortly afterthe end of the war and cannot now be traced. Twoof these were in fact rejected on further study of therecords and are included in the 12 rejects to bementioned later. Therefore of the true nephroticsin the series only two have been lost. After deduct-ing lost patients, exclusions and transfers, there are102 left for the purpose of estimating gross resultsand other factors (Table 5).

TABLE 5PERIOD 1945-57

Case records abstracted .123Selected in error 1

122Private patients 5

117Excluded on further scrutiny .12 (2 not traced)

105Transfer and no follow-up 1

104Not traced. 2

102

As already pointed out, this period can be dividedinto two main sections. Up to 1951 chemotherapywas available and then steroids were introduced,at first (since the supply was short) for limited andselected cases and by 1955 for all considered suitable.Detailed analysis as regards treatment will be givenlater. Here the gross results are considered accord-ing to the definitions already set out. In the 102children successfully followed the disorder is con-sidered to be inactive in 49, latent in nine and activein six. The remaining 38 are dead. This gives acrude recovery rate of about 50% or the same as inthe 1917-38 series when neither chemotherapy norsteroids were available. There is however a strongsuggestion, supported by a small number of cases,that maintenance methods of using steroids haveremarkably altered the picture.A histogram (Fig. 1) illustrates the distribution on

the basis of age at onset of the oedema, showing theusually accepted peak between 1 and 5 years of age.The deaths are very much the same proportionatelyfor each age group throughout the series. Thethree children with a recorded onset before 6 monthsof age are the subject of a special note on congenitalnephrosis on page 125.

118

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/


Inactive Latent

Active Dead

-

1 2 3 4 5 60 1 9Age at Onset (year)

FIG. 1.-Histogram showing age at onset in 102 cases of nephrosis(1946-57 series).

The totals for the whole series are shown inTable 6. This shows a higher incidence in boys anda slightly higher mortality rate among boys than

TABLE 6GROSS RESULTS (1945-57 SERIES)

Total Boys Girls

Cases followed .. . . 102 59 43

Inactive .. 49 25 24Latent .. 9 6 3Active .. 6 5 1Dead .38 23 15

girls. The numbers for latent or active cases are, asalready mentioned, small in relation to the clear-cutgroups of a satisfactory or fatal ending.

Table 7 gives details of the records of 12 patientsin the 1945-57 series which, on final scrutiny, havenot been included in the total of examples ofnephrosis. Reasons for the rejection are given inthe final column. The other columns indicate whyat first sight the records appeared to justify inclusion.Since over half are known to be dead, and most ofthe others to be going down-hill, it seemed importantto remove them from the series, otherwise theywould have erroneously weighted the dead, activeor latent groups.

TABLE 7NEPHROSIS RECORDS REJECTED ON FINAL SCRUTINY (1945-57)

When First Seen

Urinary Proteins Total Serum BloodName Age Oedema (mg. %) Proteins Cholesterol Reasons for Rejection

(yr.) (mg. %) (mng. %)

Kenneth B. .. 12 Nil 440 4-7 Not done No oedema; persistent haema-turia; probably chr. nephritis

Margaret D. .. 2i +++ 1,500 3 *6 870 Raised blood urea; haematuriatype 2 nephritis from onset; acutenephritis 8 mth. before

Raymond E. .. 71 ++ + 700 3 .3 470 Haematuria + + ;scarlet fever 3 yr.before; chronic nephritis

Daniel H. .. 8 Minimal 700 5*62 208 Hypertension; hepato-splenomeg-aly; chronic nephritis

Peter H. .. 10 + + Nil 3-02 367 Never any albuminuria at G.O.S.;? 'cured' before admission; oedemafor 7 weeks at age 41 yr.; lost tofollow-up

Michael H. .. 3i + 140 4 49 386 Latent stage when first seen; diedchronic nephritis age 10 yr.

Michael K. . .1 +++ Gross 3-62 394 Transferred

Peter L. .. 1 Nil 800 4-6 356 No oedema at any time; persistentred cells in urine

Vera M. .. 9 Slight 800 4-4 - Lost in follow-up; raised bloodpressure; acute nephritis aged 9 yr.

Lothan M. .. 9 Nil 1,000 56 - No oedema; blood urea 148 mg.%;died chronic nephritis

Christopher S. 4 Nil 375 5 9 166 No oedema at first attendance;latent in 1951

Julian S. .. 6i Slight 400 5 3 206 Fibrocystic disease; cirrhosis ofliver; died aged 7* yr.

25 r

201(I)qI,

( 15

10

5

119

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

ARCHIVES OF DISEASE IN CHILDHOOD

Results of Treatment (1945-57 Series)In the 1945-57 period a low salt and high protein

diet was almost universally the basic treatment forcases of nephrosis while in hospital. Sometimes adiet which was almost salt-free was given, butsometimes it was found to produce abnormally lowlevels of serum sodium while oedema still persisted,and it is now rarely used. As well as this basic diet,various other forms of treatment were tried.

Thyroid. It was for a long time considered thatthyroid had a place in the treatment of nephrosis andmany cases were given this substance. Some caseswere given thyroid up to the level of tolerance,shown usually by the onset of diarrhoea. In othersnormal therapeutic doses only were given (i-1 gr.thyroid daily).

Resins. With the introduction of resins, in aform suitable for therapeutic use, a group of childrenwas treated using various types of resin. Theresults have been published in more detail elsewhere(Payne and Wilkinson, 1951).

Diuretics. Several types of diuretic have beenused.UREA. When urea is used as a diuretic it has been

found necessary to give relatively large doses, suchas 15 g. a day, and at the same time to restrict thefluid intake. Until the blood urea reaches approxi-mately 100 mg. %, diuresis does not occur. Whenthis level is reached diuresis follows in some cases.In others, unfortunately, no diuresis occurs. In-creasing the blood urea to a higher level almostinvariably produces such severe discomfort that it isimpractical as a form of treatment. The misery ofthe child in these circumstances is most striking.When successful, the treatment can be maintainedfor many months.WATER. Diuresis will result from forced water

intake. Again this is by no means always possible.In some cases there is an initial retention of waterand then any excess water given is excreted quanti-tatively. In other cases the excess water carrieswith it some of the oedema and a genuine diuresisoccurs with loss of oedema.MERCURIAL DIURETICS. Mersalyl was the most

frequently used and, as with other diuretics, therewas sometimes an initial diuresis which in generalwas not maintained with further doses.NON-MERCURIAL D1URETICS. Usually these have

had no effect but occasional diuresis has occurredwith the first few doses. Subsequent doses haveproved ineffective.

Fever. In a few cases production of fever, eitherby measles or malaria, has been used, but the

numbers are too small for use in this study. Usuallythere has been an initial loss of oedema followed inalmost all cases by a return of oedema at varyingintervals of time.

In Table 8 the results of individual treatments areshown, the results being classified as inactive, latentor active, according to the child's condition at theend of any particular period of treatment. Themajority were on a basic low-salt high-protein dietand antibiotic treatment when necessary. Examina-tion of this Table will, of course, show many morecase treatments than the actual number of children.It will be seen that the number of cases classified asinactive occur mainly in the group which had ratherprolonged treatment on the basic diet alone.

In two of the 36 treatments with urea the inactivestate was reached, but the other diuretics, resin orwater, produced only a transient improvement: fiveof 14 cases treated with water and 10 of 36 casestreated with urea became latent. Of the 23 casesreceiving the basic treatment plus thyroid, fivebecame inactive and six latent, and of the 49 casetreatments in which no specific treatment other thanthe basic diet was given, 16 became inactive and13 latent.

TABLE 8RESULTS OF TREATMENT (1945-57 SERIES)

ResultTreatment Number

Active Latent Inactive

Basic alone .. 49 20 13 16+thyroid 23 12 6 5+resms..13 12 1 0+urea .. 36 24 10 2+water .. 14 9 5 0+other diuretics 19 18 1 0

None I... 10

It will thus be seen that no better results wereobtained when any specific treatment was given thanwhen basic treatment alone was given. In fact,since the duration of the basic treatment wasnormally much longer than that of the specifictreatment, the results appear much better.The total number of cases not receiving steroids

and becoming inactive in this group was 33: 10recovered after being discharged and were presum-ably on no particular treatment at all.

Steroids. These became available in smallamounts in 1951. A few cases were treated in thefirst years, and subsequently the use of steroids asthe main treatment became general. Since 1951 theway in which steroids have been used has changedconsiderably and, largely owing to the more criticalexamination of results, the general aim of individualtreatment has altered. Before steroid treatment

120

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/


there was no known method whereby the albumin-uria could be controlled, and most clinicians plannedthe treatment of their cases on the basis of gettingrid of the oedema and hoped that the albuminuriaand other abnormalities would clear up in time, butthe disappearance of the oedema was regarded as thetherapeutic aim. (This point is discussed at greaterlength later.) It was on the disappearance andre-appearance of oedema that the reputation forfrequency of spontaneous remissions and relapses innephrosis was based. If the stricter criteria forapparent recovery are used, it will be seen that,although there is still a liability to relapse, it is by nomeans as frequent as was previously considered.The records have been examined using these

stricter criteria and, of the 15 children who reachedthe inactive phase and had had steroid treatment,two cases only relapsed, one recovered spontan-eously and the other recovered after further treat-ment. In 22 cases of children not receiving steroids,who reached the inactive stage, only one caserelapsed, and that case also made a spontaneousrecovery. However, in the steroid-treated caseswho have reached the inactive stage relativelyrecently, intercurrent infections may cause atransient return of the albuminuria, lasting from afew days to two or three weeks and disappearingwithout treatment.

Initially, short courses of steroids were given,either ACTH or cortisone. ACTH was usuallygiven for eight to 10 days in varying doses, usually20 to 40 mg. daily, and cortisone in doses from 100 to200 mg. daily for approximately 10 days. In bothcases at the end of that period treatment wasabruptly stopped. This was called the shorttreatment. In most cases diuresis occurred andgenerally oedema disappeared but albuminuria,although frequently lessened, was generally still pres-ent and in most cases the oedema returned later on.

The next phase was to give larger doses, 200 to300 mg. of cortisone, or the equivalent of predniso-lone, or if ACTH was used, 80 mg. daily using thegel in divided doses. This treatment was maintainedfor approximately one month and then the steroidswere tailed off fairly quickly, usually within two tothree weeks. This type of treatment has beencalled long-term. In general in these cases thealbuminuria disappeared only to reappear as thedose was being tailed off or very shortly afterwards.

In the next form of treatment, called maintenancetreatment, steroid therapy has been maintained at aslowly decreasing level for much longer periods.There is much difference of opinion as to the idealway of doing this. In the majority of the main-tenance cases an initial month's treatment is given

similar to that in the long-term, by which time thereis usually no albuminuria and the plasma proteinsare approaching normal while the cholesterol isfalling. The dose is then cautiously reduced by10 to 15% every week at first and later every fort-night. This slow reduction is continued as long asthere is no albuminuria. Usually within two tothree months the plasma proteins are chemically andelectrophoretically normal and the cholesterol alsohas returned to a normal level. If these values areslow in returning to normal the therapy is continuedfor a longer period and the intervals betweenreductions are increased. If albuminuria returnsduring the period of reduction, the dose of steroidis increased until it disappears again. Reduction isthen carried out more slowly.

In a few cases the intermittent high dosage treat-ment was used, 400 mg. cortisone three days insuccession, followed by four days without steroid.

In assessing the final outcome, the cases have beendivided into two groups, those who have receivedsteroids and those who have not. The results aregiven in Table 9. The results in this Table are

TABKE 9RESULTS OF TREATMENT WITH STEROIDS

No Steroids Steroids

(no) (%) (no) (%)

Total cases .62 40

Inactive .33 53 16 40Latent .5 8 4 10Active .4 7 2 5Dead .20 32 18 45

somewhat unexpected. It appears that the resultsfrom steroid treatment are definitely inferior to thosewhen no steroids are given. One possible explana-tion for this is the inclusion in the steroid group ofcases carried over from the previous group whichwere already doomed to failure.

Table 10 is an analysis of the deaths in the twogroups. Uraemia accounted for more than half thedeaths but, whereas 10 cases of uraemic death

TABLE 10CAUSES OF DEATH

1917-1938 1945-1957

No Steroids Steroids

Uraemia .. . 4 10 10Heart failure - 1 2Infections .. . 10 3 2Cystinosis - 2 -

Pulmonary embolism - - 2Electrolyte imbalance 3 - -

Treatment 5 2 2Notknown 1 2 -

Total .. .. 23 20 18

121

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/


No stroids (62 cass)

Totlic2d(0 )

I h r~~~otal 20I.

7

61.(05

4Lt-%3

Steroids (40 caes)Uramicdeith (10)

Total 18

?c 32~~~~~~~~~~~~~~~~A -] iI 2 *oe-62 I - 2 - 3 - 4 - 5--over5 O0- 1I 2-3-4- 5-ov.r5---

YEARS YEARSFIG. 2.-Histogram showing duration to death in 38 cases (1945-57 series).

occurred out of 62 non-steroid cases, the samenumber occurred in only 40 cases given steroids(Fig. 2). Although this suggests a slight increase inuraemic death, it is hardly sufficient to account forthe difference in the two groups. A more detailedanalysis of the effects of steroid treatment based onindividual treatments using the same criteria as inTable 8 is shown in Table 11. In this Table areincluded cases treated in 1957-58 which were notincluded in the 102 cases under survey. This Tablebrings out the difference in ultimate results betweenthe short, long and maintenance types of treatment,and largely explains the apparent inactivity ofsteroids shown in Table 9.

TABLE I 1RESULTS OF TREATMENT WITH STEROIDS

1951-57 1957-58Short Long Main- Short Long Main- No

tenance tenence Treatment

Totaltreat-ments 35 36 6 3 8 4 1

Inactive 0 10 6 0 2 4 1Latent 10 13 0 0 0 0Active 25 13 0 3 6 0

'Cure' % 0 28 100 0 25 100

The numbers treated on maintenance are notsignificant and the 100% cure reached is certainlyaccidental and is partly due to the fact that onlycases whose treatment had been completed wereincluded in this Table. In order to appreciate morefully the effect of steroid treatment a study ofTable 8 (individual treatments, non-steroid) isnecessary. It is apparent from this that no indi-vidual treatment produces as good results as simpledietary measures alone, or even no treatment at all.It is against this background of spontaneous remis-sion that the results of steroid treatment should beassessed.

In Table 12A the length of time elapsed betweenonset of disease and cure, in the steroid and non-steroid cases, is compared, and it will be seen thatsteroid treatment definitely produces more rapidcure.

In Table 12B the time relationship between onsetand death shows that in the fatal cases there is nomaterial difference.

Tables 12C and D show the ultimate result andthe interval of time between onset of disease and thestart of steroid therapy. There is no suggestion inthese Tables that better results are obtained whensteroid therapy is started early.

TABLE 12

C Time from Start of Steroid Therapy to Cure

<6 6-12 1-2 2-5 >5 Totalmth. mth. yr. yr. yr.

7 1 2 6 16

D Interval between Onset of Disease andSteroid Therapy l . ~~~Total

<6 6-12 1-2 2-5 >5mth. mth. yr. yr. yr.

Cure .. 8 2 3 3 16Death .. 9 3 4 2 18Latent or

active 2 1 1 2 6

7

6C

4

3

2

122

piI

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

FORTY YEARS OF NEPHROSIS IN CHILDHOODIn analysing the causes of death the hazards of

treatment have to be considered. In the period1917-38 it will be observed that a fair number ofdeaths can be attributed directly to treatment. It isnot quite so obvious in the period 1945-57 but asuspicion by one of us (A.M.) that mersalyl was adangerous treatment has been confirmed. Theresults in 12 cases treated with mersalyl are: inactive(1), latent (1), dead (10). These results stronglysuggest that mersalyl in itself is dangerous. That itwas used merely as a last resort was not found to bethe case on examining the notes, since in most casesit was given in the first 12 months from the onsetof disease.

It must be realized that steroid treatment ispotentially dangerous and the group of cases wehave treated shows the following results:

Died of overwhelming infection . 1Sudden cardiac failure and death on the fourth day of

cortisone treatment IDecalcification of vertebrae. 1Acute cerebral crisis with unconsciousness and fits (? cere-

bral oedema), with ultimate recovery ..1Adrenal crisis, low blood pressure, low serum sodium and

collapse. 2

Thus, of the 18 steroid deaths, two can be attribu-ted to steroidtherapy, and both the cases of adrenalcrisis could quite well have been fatal if they had notfortunately occurred while in hospital.

DiscussionThe literature on nephrosis threatens to become

unmanageable. This paper is essentially concernedwith a presentation of facts as ascertained from astudy of the disorder in childhood in hospitalrecords for over 40 years. It is not proposed todiscuss the pathology or pathogenesis. The defini-tions already set out indicate what sort of diseaseprocess is being considered. The results of treat-ment, with special reference to those obtained by theuse of steroids, can usefully be compared with thosepresented by McCrory and Fleisher (1958), Metcoff(1958) and Riley and Scaglioni (1959). A usefulhistorical survey was given by Mann (1958) on theoccasion of the Bright centenary celebrations.

If the assumptions made in the section whichconsiders the 1917-38 series are accepted as correct,then about half the patients are dead. In the 1945-57 series as a whole about half are either dead or inan active or latent phase. Some of those classifiedin the last two categories at the time of assessmentmay by now have moved or will move into theinactive group, so that broadly speaking theseresults probably show a slight improvement overthe results of 1917-38, but certainly not as great an

improvement as might have been expected with thebetter control of infection and, in the later cases, anactive agent to attack what is probably the funda-mental lesion which leads to the proteinuria. Again,if the 1945-57 group is broken down into those whodid not receive steroids and those who did, there isno dramatic improvement with the new drugs, at anyrate in the early years before the best method ofusing steroids was worked out. Indeed, taking theoverall figures as shown in Table 9, there is a 45%mortality rate in the steroid-treated cases comparedwith only 32% in those not receiving steroids, with15% in each group still classified as latent or active.A review of the causes of death in each of the

three periods shows that 17% died of uraemia in thepre-antibiotic era, the remainder dying from causessuch as infections, which now would be regarded ascurable. Since the introduction of antibiotics thecause of death has changed and is now uraemia inover 50% of cases (Table 10).However, deaths which would have been theoreti-

cally preventable still occur and in the final analysisuraemia and heart failure seem to be the expectedcauses of death in nephrotic children. A smallunknown proportion of these uraemic deaths will bedue to unavoidable errors in initial diagnosis, suchas in those cases shown in Table 7, which wereretrospectively rediagnosed. Until more certainmeans of diagnosis are utilized (such as renalbiopsy), this error will prevent an accurate prognosis,and therefore a proper evaluation of treatment,being made. From studies already published(Vernier, Farquhar, Brunson and Good, 1958) it islikely that in the majority of cases the correctdiagnosis can be made in the early stages of thedisease. Thus, a death rate from uraemia and heartfailure of 23% of all recent cases (and it will becomehigher, since many of the still active cases arealready uraemic) indicates that a proportion of 'purenephrotic' children will die of uraemia. A smallproportion will still die of intercurrent accidents,preventable or not; the remainder, if they can bekept alive, should recover. It is impossible toassess this figure accurately, but it is over 50% andprobably about 60-70%, as judged by the non-uraemic causes of death. Any treatment must havea 'cure' rate of about 80-90% before it can be regar-ded as of real value. Table 12 shows that themajority of deaths occur within the first five yearsof the disease, so it should be possible to determineif a given treatment has successfully prevented thedevelopment of uraemia when enough cases havebeen treated for over five years. At the presentmoment the evidence is only sufficient to indicate apromising trend.

123

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/


In considering the value of many of the remediesused, the aim of the treatment at the time must beborne in mind. Until quite recently the physician'sprimary aim was the cure of the oedema, with thehope that in due course the remaining pathologicalchanges would cure themselves. With such animperfect criterion of cure it is not surprising thatrelapses were many and frequent. Thus, whenexamining Table 8, all the latent results would havebeen called cures. Even with this proviso, it will beseen that the more common methods of therapywere quite ineffective when compared with the basictreatment (low-salt, raised-protein diet). In a smallseries of eight cases treated with measles and twowith malaria, seven reached the latent stage, but onlyone became inactive without further relapses.

Since the oedema most frequently causes thepatient to consult a doctor, its disappearance musthave a good effect on the outlook of the child and hisparents, although there is no direct evidence thatremoving oedema improves the prognosis. Thus itis reasonable to take non-specific measures whichwill help in its removal, providing that they areharmless. For example, urea, resins, water andoccasionally non-mercurial diuretics are useful. Forthe same reason a low-salt diet is given.Oedema, if severe, can be very distressing and can

even endanger life. In these cases acupuncture ofthe dependant legs and abdominal paracentesis,both under antibiotic cover, have been used andoccasionally a complete remission has followed.

Steroid treatment has undoubledly made a bigchange in the method of treating nephrosis. In theearly days, when removal of oedema was thetherapeutic aim, short eight to 10 day treatmentswere used with an initial gratifying effect but, aswith previous methods, relapses occurred, whichwere more resistant to further treatment withsteroids. The aim of treatment was altered tocomplete reversal of all pathological changes. Muchlonger terms of treatment were used and the resultsbegan to be more encouraging. While, as yet, nouniform method has been arrived at, the generalprinciple is to give adequate dosage of a steroidpreparation, the equivalent of 240 to 300 mg.cortisone daily. It does not appear to make muchdifference which of the new, less sodium-retainingsteroids is used. Treatment is continued for six to12 months according to the plan for reductionalready mentioned. Clinically it seems that itneeds many months to reverse completely thepathological changes in the glomerular basementmembrane, if indeed this can be done. Since it isprobable that some 20-30% of untreated patientswill progress to irreversible renal damage, the

essential fact to determine is whether in these casesit is possible to prevent this.As has been pointed out, this can only be done

when sufficient time has elapsed for such changes todevelop-at least five years. This delay in makinga judgment of cure is necessary since steroid therapycan remove the nephrotic symptoms for a time incases in which the diagnosis is known not to besimple nephrosis. Until an adequate period haselapsed it is possible only to say that the latestresults are promising in that a high percentage arereaching the inactive stage.

It has frequently been stated that liability to areturn of nephrosis can never be eliminated. It isclear that clinical relapse is common if only thelatent stage is reached. When the inactive stage isreached relapse is less frequent but may still occur.Many patients, whose nephrosis has been success-fully treated with steroids, for some years tend toget transient albuminuria during fairly mild inter-current infections. In a few this proceeds to a fullrelapse which may need a further course of steroids.

This raises the fundamental question of theaetiology of nephrosis. Should the condition beregarded as an inborn liability of the kidney to aspecial type of pathological response to certaininsults, or should the pathological response beregarded as due to a specific insult, such as isbelieved to occur in acute nephritis followinginfection with certain types of streptococci? If theformer view is held, it is unlikely that any form oftreatment will prevent recurrence of symptoms evenmany years later. In the second possibility, oncea cure is complete, relapse would only occur with afurther specific insult. Such clinical evidence asexists suggests the former view as being possiblythe more likely since occasionally relapses haverecurred after manyyears of freedom from symptoms.To sum up, the natural course of nephrosis, if

intercurrent infections are adequately dealt with, isfor at least 50% and probably nearer 60-70% of thepatients to be cured, but there is also an inbornliability to relapse. There is in every series, basedon clinical diagnosis alone, a small group with adifferent pathological basis and a bad prognosis.No treatment has yet been shown to improve theultimate prognosis, although some treatments haveworsened it. The most recent method of steroidtherapy, while very promising, has undoubtedhazards which must be balanced against the presentuncertainty that permanent improvement in prog-nosis will occur.

Since delaying the start of steroid therapy doesnot seem to affect the ultimate result (Table 12D)and since a proportion of untreated cases remit

124

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

FORTY YEARS OF NEPHROSIS IN CHILDHOODspontaneously within a few months (Table 12A),delay of up to three months in starting steroidtherapy is suggested to allow time to see if there aresigns of spontaneous improvement. There is noevidence that any adjuvant treatment has a materialeffect, apart from the need to give an adequateprotein diet and to keep the salt content low whenthere is oedema, especially during steroid therapy.Complete bed rest, except when there is grossoedema, is not necessary. A change from prolongedbed rest to reasonable activity at The Hospital forSick Children caused no worsening of the prognosisand a decided improvement in the children's morale.Another advantage of permitting reasonable activityis that the child can be sent home and attend schoolwhen desired while still on steroid or other therapy.

It is difficult to assess the value of the routine useof antibiotics. Oral penicillin has been used in thepresent series ever since it became available, rein-forced by other antibiotics if an intercurrent infec-tion occurred. This has been quite successful, andit has not been necessary to reduce steroid dosagebecause of infection. Prophylactic antibiotics orsulphonamides have been used to try to preventintercurrent infections, but they have not beensuccessful.

Note on Congenital NephrosisThree of the children in the 1945-57 series mani-

fested the nephrotic syndrome at or within a fewweeks of birth and all died. Two of these weresiblings (boy and girl) whose parents were firstcousins and they have been reported elsewheretogether with a third case (not in the Great OrmondStreet series) whose parents were also cousins(Giles, 1957). Briefly, the post-mortem changeswere mainly in the proximal tubules, and doublyrefractile crystals were present. It is likely that thewhole condition was related to cystinosis.The third child in the present series was the first

child of unrelated parents who have since had anormal child. Her eyes were reported as swollenat birth and there was a sudden weight gain aboutthe tenth day. Generalized oedema appeared atabout the fourth month. She was given mersalyl atanother hospital and later referred here because ofa persistent urinary tract infection. Her conditionsatisfied the four criteria for nephrosis and noimprovement was achieved by the use of thyroidextract, blood transfusion, urea or decapsulationof the kidneys. The kidneys at operation were paleand soft with no scarring. No biopsy was done.A year later on re-admission she had no oedema butthe blood pressure was raised to 140 mm. Hg, theblood urea was 340 mg.% and there was an albumin-

uria of 650 mg. % She died at home at the age of2- years. No autopsy was performed.Another infant admitted to The Hospital for Sick

Children in February, 1958, born after a normalpregnancy and delivery and weighing 51 lb., deve-loped generalized oedema on the third day of life.(He was the first child of the present marriage ofunrelated parents; the mother had had two normalchildren by a former marriage.) The oedemasubsided after two days and recurred at the age offour weeks, leading to admission under the care of acolleague (Dr. A. P. Norman). The infant satisfiedthe four criteria of nephrosis and electrophoresisshowed a typical nephrotic pattern. Urinaryamino-acid excretion was within normal range. Hewas treated with steroids and antibiotics. Verytroublesome diarrhoea caused great difficulties inoral feeding and he died after a sudden collapse atthe age of 3j months. Autopsy (Dr. M. Bodian)showed bilateral renal vein thrombosis of longstanding with some extension into the inferior venacava. The changes in the kidneys were thoseassociated with nephrosis.These five cases illustrate various ways in which

nephrosis may be caused very early in life. Poison-ing by mercury is another possible factor. Theposition briefly is that congenital nephrosis orneonatal nephrosis is likely to be due to mechanismsother than those causing the more usual type ofnephrosis such as is dealt with in the present report.

Summary and ConclusionsA study of nephrosis as seen at The Hospital for

Sick Children during the last 40 years has beenmade. The cases have been divided into threegroups: (1) pre-antibiotic 1917-38; (2) antibiotic,and (3) steroid. The recovery rate in the first groupwas approximately 50% and much of the mortalitywas due to infection. In those survivors followed-upin this group there was no instance of relapse afterthe final hospitalization period. The recoveryrate in the second group was 53% and in the third40%, and in these last two groups the mortality fromuraemia was very much higher.While the control of infection and of water balance

has clearly improved the short-term prognosis, noneof the other methods of treatment, including short-term steroid therapy, appears to have had any verystriking effect on the mortality.

Before there was any method of treatment whichstrikingly affected the basic sign of proteinuria andits associated biochemical disturbances, attentionwas focused on the control of infection and on thetreatment of oedema. Experience has shown thatthe treatment of oedema per se is not particularly

125

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

126 ARCHIVES OF DISEASE IN CHILDHOODimportant except when the oedema is so severe as tobe disabling in itself.The observation that steroids could bring about

a complete biochemical remission in the largeproportion of cases has switched attention fromprotective and palliative treatment to the idea ofcurative treatment. With the advent of steroidtherapy therefore the aim has become the completereversal of all the pathological changes.The results of various types of therapy have been

compared. Until the technique of prolonged (six-12months) steroid therapy was introduced, no specifictherapy appears to have shown any advantage oversimple basic dietary therapy and the control ofinfection with antibiotics.

Results in a small number of cases in whom longcontinued cortisone therapy has been given suggest,however, that permanent remissions may be obtainedby this means, and raise the hope that the suppres-sion of the biochemical disease in this way mayreduce the incidence of that secondary nitrogenretention and hypertension which have become themajor cause of mortality in recent years.There is a high natural tendency to recovery. If

all avoidable intercurrent incidents are excluded thenatural remission rate is between 60 and 70 %. In areasonable proportion of cases of nephrosis spon-taneous remission occurs within the first few monthsof the disease. As there is no evidence in this seriesthat early treatment gives a better prognosis thanthat obtained when steroid treatment is delayed forsome months from the onset of the disease, and sincethere are undoubted hazards from the use of steroids,it is suggested that steroid treatment of a new caseshould be delayed for a month or so to exclude thepossibility of an early natural remission.

In our view the present aims of treatment shouldbe as follows:

(1) The control of infection with antibiotics

(2) The maintenance of nutrition as far as pos-sible with a moderately high protein intake

(3) The partial control of oedema with a low-sodium (but not a sodium-free) diet

(4) Long-term treatment with steroids in variabledosage sufficient to maintain urinary proteinexcretion at the lowest possible level.

By these methods it should be possible to maintainaffected children in as good health as possibleduring the active phase of the disease, and to keepthem out of hospital for most of the time, thusinterfering with their lives as little as possible.The success of long-term steroid therapy must be

judged by the extent to which it reduces the incidenceof renal failure, which has become the major causeof mortality.

Five fatal cases of congenital nephrosis arereported, with varying underlying pathologies.

Thanks are cordially expressed to colleagues on thestaff of The Hospital for Sick Children, Great OrmondStreet, for permission to include the records of theirpatients and in some instances to follow up the courseof the disease; also to the Department of MedicalIllustration for the histograms, and to secretarial stafffor much patient typing and re-typing of drafts. Theinvestigations recorded were undertaken under theauspices of the Research Committee of The Hospitalfor Sick Children.

REFERENCESGiles, H. Mc., Pugh, R. C. B., Darmady, E. M., Stranack, F. and

Woolf, L. I. (1957). The nephrotic syndrome in early infancy-a report of three cases. Arch. Dis. Childh., 32, 167.

Mann, W. N. (1958). Bright's disease: the changing concept of acentury. Guy's Hosp. Rep., 107, 323.

McCrory, W. W. and Fleisher, D. S. (1958). The nephrotic syn-drome. In Recent Advances in Pediatrics, ed. Gairdner, D.2nd ed., p. 227. Churchill, London.

Metcoff, J. (1958), Editor. Proceedings of the Ninth Annual Con-ference on the Nephrotic Syndrome. New York.

Payne, W. W. and Wilkinson, R. H. (1951). Nephrotic oedematreated with an ion-exchange resin. Lancet, 2, 101.

Riley, C. M. and Scaglione, P. R. (1959). Current management ofnephrosis. Pediatrics, 23, 561.

Vernier, R. L., Farquhar, M. G., Brunson, J. G. and Good, R. A.(1958). Chronic renal disease in children-correlation ofclinical findings with morphological characteristics seen by lightand electron microscopy. A.M.A. J. Dis. Child., 96, 306.

copyright. on O


http://adc.bmj.com

/A

rch Dis C


ownloaded from

http://adc.bmj.com/

fortyyears of nephrosis childhood* · characteristics, namely oedema, gross proteinuria, lowered...

Documents