380

significantly better than that in the placebo group at 8 and 12weeks (table ll).

Scalirtg.-The plaque scaling had improved in the

maxepa group by 8 weeks (fig 3), with no change in thecontrols. However, the improvement in the treatment groupwas not significantly better than that in the placebo group(table 11).

Surface area.-The percentage surface area affectedshowed a trend towards improvement in the treatmentgroup (fig 3), but this did not attain statistical significance.

Discussion

Psoriasis is an inflammatory dermatosis in whichdisordered arachidonic acid metabolism may play an

important part. We have shown that dietarysupplementation with small amounts of oils from marinefish can reduce the itching, erythema, and scaling inpsoriasis with a trend towards decreasing the area of affectedskin. The improvement occurred in patients with chronicstable psoriasis and was achieved without any other dietarymanipulation.The finding that the improvement obtained with EPA

was significant when compared against placebo suggeststhat EPA has an anti-inflammatory effect in psoriasis. EPAhas already been shown to have an anti-inflammatory effectin rheumatoid arthritis. 12 More recently, dietarysupplementation with EPA has been shown to be of benefitin atopic dermatitis ’13 an inflammatory dermatosis in whichepidermal levels of LTB4 may be_ raised.14 The anti-

inflammatory effect of EPA may be mediated in two ways.Firstly, the leukotrienes and prostaglandins derived fromEPA may be less biologically active than those derived fromAA and thus act as competitive inhibitors. Secondly, it isthought that AA and its derivatives catalyse some stages ofthe cyclooxygenase and lipoxygenase pathways. EPA and itsderivatives may be less effective catalysts, thus slowing theinflammatory pathway reactions.

Since stable plaque psoriasis improves significantly withEPA, more profound effects are likely in the inflammatoryunstable variants such as erythrodermic and generalisedpustular psoriasis. Treatment with EPA would haveconsiderable advantages over the current therapies for thesepatients, such as methotrexate, etretinate, and PUVA,especially in women of childbearing age. The effectivenessof EPA could probably be improved by increasing dietaryintake of oily fish and reducing that of foods with a high AAcontent. Interestingly, only 150 g of oily fish (eg, mackerel)consumed daily would provide the same amount of EPA asused in our study. This controlled study confirms findingsin open trials8.11 that dietary supplementation with EPA isuseful as an adjunctive treatment in psoriasis, particularlywhen itching is troublesome.

Correspondence should be addressed to S. S. B., Department of

Dermatology, Royal Hallamshire Hospital, Sheffield S10 2JF.

REFERENCES

1. Hammerstrom S, Hamberg M, Samuelsson B, Duell EA, Stawiski M, Voorhees JJ.Increased concentrations of nonesterified arachidonic acid, 12-L-hydroxy-5,8,10,14-eicosatetraenoic acid, prostaglandin E2 and prostaglandin F2 in epidermisof psoriasis. Proc Natl Acad Sci USA 1975, 72: 5130-34.

2. Brain S, Camp R, Dowd P, Black AK, Greaves M The release of LTB4-like materialin biologically active amounts from the lesional skin of patients with psoriasis. JInvest Dermatol 1984; 83: 70-73.

3 Camp R, Russel Jones R, Brain S, Wollard P, Greaves M Production ofmicroabscesses by topical application of leukotriene B4. J Invest Dermatol 1984; 82:202-04.

4. Pease CT, Fennell M, Staughton RCD, Brewerton DA. Polymorphonuclearleukocyte function m psoriasis vulgaris: Br J Dermatol 1987; 117: 161-67.

5. Lassus A, Forsstrom S A dimethoxynaphthalene derivative (RS-43179 gel) comparedwith 0·025% fluocinolone acetonide gel in the treatment of psoriasis. Br J Dermatol1985; 113: 103-06.

6. Ellis CN, Fallon JD, Heezen JL, Voorhees JJ. Topical mdomethacm exacerbateslesions of psoriasis J Invest Dermatol 1983; 80: 362 (abstract).

7. Kragballe K, Desjarlais L, Voorhees JJ. Leukotrienes B4, C4 and D4 stimulate DNAsynthesis in cultured human epidermal keratinocytes. Br J Dermatol 1985; 113:43-52.

8. Ziboh VA, Cohen KA, Ellis CN, et al. Effects of dietary supplementation of fish oil onneutrophil and epidermal fatty acids. Arch Dermatol 1986; 122: 1277-1282

9. Lee TH, Mencia-Huerta JM, Shih C, Corey EJ, Lewis RA, Austen KFCharacterisation and biologic properties of 5, 12-dihydroxy derivatives of

eicosapentaenoic acid, including leukotriene B5 and the double lipoxygenaseproduct. J Biol Chem 1984; 259: 2383-89

10. Kragballe K, Voorhees JJ, Goetzl EJ. Leukotnene B5 derived from eicosapentaenoicacid does not stimulate DNA synthesis of cultured human keratinocytes butinhibits the stimulation induced by leukotriene B4. J Invest Dermatol 1985, 84: 349

11. Maurice PDL, Allen BR, Barkley ASJ, Cockbill SR, Stammers J, Bather PC Theeffects of dietary supplementation with fish oil in patients with psoriasis Br JDermatol 1987; 117: 599-606.

12 Kremer JM, Michalek AV, Lininger L, Huych C, Bigauoette J, Timchalk MA, et alEffects of manipulation of dietary fatty acids on clinical manifestations ofrheumatoid arthritis Lancet 1985; i 184-87.

13 Bjorneboe A, Soyland E, Bjomeboe GEA, Rajka G, Drevon CA Effect of dietarysupplementation with eicosapentaenoic acid m the treatment of atopic dermatitisBr J Dermatol 1987; 117: 463-69

14. Ruzicka T, Simmet T, Peskar BA, Braun-Falco O Leukotrienes in skin of atopicdermatitis. Lancet 1984, i: 222-23

SHORT VERSUS STANDARD PREDNISONETHERAPY FOR INITIAL TREATMENT OFIDIOPATHIC NEPHROTIC SYNDROME IN

CHILDREN

ARBEITSGEMEINSCHAFT FÜR PÄDIATRISCHENEPHROLOGIE*

Summary Two regimens of steroid treatment for theinitial attack of idiopathic nephrotic

syndrome in children were compared in a controlledmulticentre study. Short-course prednisone therapyconsisted of 60 mg/m2 per 24 h until proteinuria haddisappeared for 3 days, followed by 40 mg/m2 per 48 h untilcomplete remission had occurred. The standard prednisonetherapy was 60 mg/m2 per 24 h for 4 weeks, followed by 40mg/m2 per 48 h for 4 weeks. 61 children with a first attack ofidiopathic nephrotic syndrome were allocated at random tothese groups. Urinary remission in the short-course groupwas achieved after 14 days of daily prednisone, and completeremission after an additional 16 days of alternate dayprednisone. The cumulative rate of patients with sustainedremissions after two years was significantly lower after theshort course than after standard treatment (19% versus41%, p=0·001). The mean duration of remission in patientswith a relapse was half as long after the short course (79versus 169 days, p = 0·004). Complete initial remission ofsteroid responsive nephrotic syndrome can be obtained withhalf the standard prednisone dose, but the short course isfollowed by a higher rate of relapses, that require repeatedprednisone administrations. In the long term, the standardregimen is preferable.

Introduction

THERE have been several uncontrolledl-3 and controlled4-6trials of corticosteroid treatment in children with minimal

change nephrotic syndrome (MCNS). This condition-previously known as lipoidnephrosis-has been known torespond to corticosteroids for nearly 40 years, but the besttreatment for an initial attack remains uncertain. The most

widely used regimen consists of 4 weeks of daily prednisone,

*Coordinated by Prof J. Brodehl and Prof J. H. H. Ehrich (Hannover)

381

Fig I-Protocol of study.

in a dosage of 60 mg/m2 per 24 h, followed by 4 weeks ofalternate day prednisone, in a single morning dose of 40mg/m2 per 48 h.’

Leisti and coworkers8 suggested that the initial steroidtreatment may cause a state of endocrine and immuneimbalance in the patients, which could induce a tendency formultiple relapses of MCNS. A multicentre controlled studyon the initial prednisone treatment of children with

idiopathic nephrotic syndrome was conducted to comparethe standard initial prednisone regimen with a short initialprednisone regimen.

Patients and Methods

61 patients with a first attack of idiopathic nephrotic syndrome,and who had preserved glomerular function (creatinine clearancegreater than 68 ml/min per 1-73 m2), were entered into the study,after the informed consent of their parents. The children were from2 to 16 years of age, had had no previous treatment withcorticosteroids or immunosuppressive agents, and did not have anycontraindications to corticosteroid therapy. A renal biopsy was notrequired for admission to the study, but all patients known to havelow C3 complement, diabetes mellitus, systemic lupuserythematosus, amyloidosis, postinfectious glomerulonephritis,metabolic or drug-induced nephritis, hereditary glomerulardisease, vasculitis, or Henoch-Schoenlein nephritis were excluded.Definitions and criteria for nephrotic syndrome, remission, andrelapse were the same as those used by the International Study ofKidney Diseases in Children (ISKDC)9 and our own group,toexcept as discussed in the text. Two types of remission weredefined: urinary remission was said to occur when three consecutivedays were without abnormal proteinuria (<4 4 mg/m2 per h).Complete remission was said to occur when the serum albumin hadreached the level of 35 g/1.

’

Patients were randomised into groups receiving short initial

prednisone treatment or the standard dose. The treatment regimensare summarised in fig 1. If a patient in the short-course group did

TABLE I-MEAN DURATION AND DOSAGE OF PREDNISONE THERAPY

FOR THE INITIAL ATTACK OF NEPHROTIC SYNDROME IN CHILDREN

*Mann-Whitney test.

not reach urinary remission after 4 weeks of continuous treatment,the alternate day regimen was started. If there was no response after8 weeks of treatment, the patient was taken out of the study andtreated individually. If a patient relapsed during, or after, the initialtherapy, the relapse was treated as shown in fig 1. In theshort-course group, the same regimen was used for relapse as for theinitial treatment; in the standard group, prednisone was given at 60mg(m2 per 24 h until the urine had been protein-free for 3 days,followed by a 4 week course of alternate day prednisone at 40 mg(m2per 48 h. In both groups, diuretic and antihypertensive drugs couldbe given as required. Urinary protein excretion was assessed by theparents with ’Albustix’. For the short-course group, blood albuminwas checked once a week during the initial treatment, but in thestandard group it was only checked after 4 and after 8 weeks.The study aimed to follow each patient for 2 years to assess the

occurrence of remission and relapse, cumulative doses of

prednisone, and side-effects of steroid treatment. If the full 2 yearfollow-up was not possible, the patient’s data remained registeredup to the date of withdrawal. However, if there was a therapeuticfault, or if the observation period was less than three months, thepatient’s data were withdrawn completely from the final evaluation.

Statistical analysis was performed with the Mann-Whitney test,the paired and unpaired t-test, chi-squared test, and the rank test forlife-table analysis.

Results

Description of Patient Population77 patients were initially recruited into the trial, but 16

had to be removed at an early stage due to steroid resistance(n = 8), or early deviations from the treatment protocol(n = 8). Of the 61 remaining children, 32 were randomlyassigned to the short prednisone therapy, and 29 to thestandard regimen. There were no statistically significantdifferences between the two groups in respect of age, sex,blood pressure, or the serum concentrations of creatinine,cholesterol, total protein, albumin, IgG, IgM, or IgA.

34 patients completed the study for the full 2 years. Datafor the other 27 patients were included for the period that

Fig 2-Length of steroid therapy until urinary and completeremission is reached in short prednisone therapy group.

382

they remained in the study protocol. Of the 27, 5 patients ofthe short-course group and 4 from the standard group wereremoved when they required other immunosuppressiveagents; 2 patients from each group left the couritry duringthe course of the study; 7 children from the short-coursegroup, and 3 from the standard group, were lost to follow-updue to failure of continuous parental cooperation; and latetreatment faults were observed in 3 cases after short-coursetreatment, and in 1 patient after standard therapy. The fullcourse was completed by 15 patients receiving the shortcourse and by 19 receiving standard treatment. The meanlength of follow-up in the short-course group was 17

months, and in the standard group was 20 months, adifference which was not statistically significant.Remissions after the Initial AttackWith short-course prednisone therapy, a urinary

remission was achieved after a mean duration of 14 days ofcontinuous prednisone, and a complete remission after 16days of alternate day prednisone (table I). Mean totalprednisone dose in the short course was 1220 mg/m2 bodysurface-just over half the standard dosage of 2201 mg/m2(p=0001). Serum analysis after completion of prednisonetherapy revealed no significant differences in serum

concentrations of cholesterol, albumin, alpha-2 globulin,IgA, IgM, or IgG in the two groups. In 50% of theshort-course patients, disappearance of pathologicalproteinuria was achieved after 7 days. The serum proteinsshowed a complete remission after 26 days in 50% ofshort-course patients (fig 2).

Relapses after the Initial AttackThe majority of patients in the short-course group (59%)

relapsed during the first three months after cessation ofinitial daily treatment (table II), in contrast to the standardgroup in which only 21 % relapsed. This difference in therate of relapses remained obvious throughout the two yearsobservation period, irrespective of whether the end of dailyprednisone or the end of alternate day prednisone was usedas the starting point for observation. The cumulative rate ofpatients with sustained remission two years after the initialattack was significantly lower in the short-course group thanin the standard group (19% versus 41 %, p = 0-009) (fig 3).

TABLE II-PATIENTS WITH OR WITHOUT RELAPSE IN THE TWO

STUDY GROUPS 3, 6, AND 12 MONTHS AFTER THE END OFCONTINUOUS STEROID THERAPY FOR INITIAL ATTACK

I I I i I ’

*1 additional patient who remained without a relapse was followed for 10months only.

TABLE III-MEAN RELAPSE RATE PER PATIENT AT INTERVALS OF 3,6,AND 12 MONTHS AFTER CESSATION OF DAILY PREDNISONE IN

INITIAL THERAPY IN THE TWO STUDY GROUPS

*Mann-Whitney test.

Fig 3--Cumulative percentage of sustained remissions in childrenwith nephrotic syndrome after short and standard prednisonetreatment.

The number of relapses was higher in the short-course thanin the standard group, as shown by the mean relapse rate perpatient during the first year (table III). Mean duration ofremission after completion of initial daily prednisone, inpatients who had a relapse, was only half as long after the

TABLE IV-MEAN DURATION OF REMISSION IN RELAPSERS

-

*Mann-Whitney test; ns = not significant.

short course than after standard treatment (table IV). Thepercentage of frequent relapsers, according to the definitionof ISKDC,6 was 55% after short-course therapy and 39%after standard. This difference, however, was not

statistically significant.Prednisone Therapy of the Relapses

Duration and total dosage of prednisone therapy for thetreatment of relapses were not significantly different in thetwo study groups (short = 957 mg/m2 per relapser, SD 240;standard = 1156 mg/m2, Sb 321). The mean duration ofremission after the first and second relapse therapy was notsignificantly different in both groups, and ranged between58 and 84 days (table iv). No patient became steroidresistant during the study. The effect of prednisone uponthe disappearance of proteinuria in relapses was seen after amean of 7 days and was not different from the initial attack.

Side-effects of Steroids

Mild sequelae of steroid therapy during the initialtreatment were detected more frequently after standardtherapy, when 13 patients (45%) showed side-effects: mildcushingoid facies was found in 4 patients, obesity with striaein 6, hirsutism in 2, and psychological and emotionaldisturbance in 1. In the short-course group, 8 patients(25%) showed side-effects: mild cushingoid facies or acne in7, and moderate obesity with striae in 1.

Discussion

Only a few reports describe attempts to find the minimumeffective dose of steroids to induce remission. In the earlydays of steroid treatment, ACTH and/or cortisone was

383

administered for a few days to induce diuresis.ll-14 This,however, was followed by a high rate of rapid relapses,15-17 solonger regimens of intensive steroid therapy were

introduced.1-3,16 These could induce long remissions, butthe rate of serious and fatal side-effects increasedconsiderably. 18-22 The International Study of KidneyDiseases in Children (ISKDC) started its multicentrestudies in 1967 and took into account all the previousexperiences with different initial steroid regimens. It

decided, rather arbitrarily, on a standard regimen for theinitial attack of nephrotic syndrome in children whichconsisted of 4 weeks of continuous prednisone (60 mg/m2per day) and 4 weeks of intermittent prednisone (40 mg/m2on 3 out of 7 days). This standard regimen is now usedworld-wide, mostly with the modification of alternate-dayapplication instead of intermittent application for thesecond 4 weeks, since it was shown that the alternate-dayregimen is more effective than the intermittent regimen withthe same total dose of prednisone. 5,7This study showed that 50% of patients in the short-

course group had lost their proteinuria after 7 days, but thatit took 26 days until 50% had achieved a complete remission(fig 2). This time sequence could be studied only in theshort-course patients, in whom the urine was checked everyday and the blood once a week: the interval to achieve acomplete remission may have been slightly shorter thanregistered, due to delay in detecting the normalisation ofserum albumin. The standard group only had bloodalbumin assessed at the end of 4 and 8 weeks. This studycannot answer the question whether the time sequence ofthe normalisation of serum albumin is influenced by thedifferent regimens-in particular whether the normalisationof albumin is merely the consequence of the loss of

proteinuria, or if it may be influenced by the dosage andlength of prednisone treatment.

In this study, a short regimen was compared with thestandard regimen for the initial steroid treatment of a firstattack of MCNS in children. A renal biopsy was notmandatory for this study, which is usual in such patients :23the continuing responsiveness to prednisone, both at thestart of treatment and during relapses, makes it likely that allof the patients in this study belonged to the minimal changecategory. The results are not likely to be significantly biasedby a heterogeneity of patients.The results show that remission from the initial attack of

nephrotic syndrome can be achieved by using just over halfthe standard steroid dose. In most cases, 2 weeks ofcontinuous prednisone, followed by 2 weeks of alternate daytherapy, effectively induced complete remission. However,the short regimen was followed by a significantly higher rateof relapse. As each relapse required further treatment withsteroid, the mean cumulative dose of prednisone by the endof one year was 4-5 g/m2 in the short-course group,compared with 5-0 g/m2 in the standard group. There wasno significant difference between these figures. Each relapsehas a risk of increased susceptibility to infection,24 and tothromboembolism 2s and the aim of the initial treatmentmust be to minimise the risk of relapse. These data stronglyfavour the use of the standard regimen over the shortregimen.The higher rate of relapses and the shorter duration of

remissions after the short initial therapy found in this studyis against a role for poststeroidal adrenocortical suppressionin relapses of MCNS, which has been suggested.826

Earlier controlled studies 10,27 have found that prolongedand intensified steroid treatment of a relapse did not change

the relapse rate thereafter. This is again shown in this study,where there was no difference between success in treatmentof a relapse, once the relapse had occurred. This studysuggests that the initial immunosuppressive attackdetermines the length of benefit from corticosteroidtreatment in MCN S, and that further attempts to reduce theinitial steroid treatment are unjustified.

Contributing investigators and centres were: Prof K. J. van Acker(Antwerp); Prof F. R. Egli (Basel); Dr T. Lennert (Berlin); Dr R. Mallmann(Bonn); Prof G. Mau (Braunschweig); Dr D. Michalk (Erlangen); Prof H.Olbing and Prof H. Bachmann (Essen); Prof K. Scharer (Heidelberg); ProfH. P. Weber (Liidenscheid); Dr R. Eife (Munchen); Dr K. E. Bonzel(Munster); Dr K. Gellissen (Neuwied); Prof K. E. von Miihlendahl

(Osnabruck); Prof W. Hagge (Stuttgart); Prof D. Gekle (Wurzburg).

Supported by grant Ez. 1-34844 from the VW-Foundation.

We thank Prof B. Schneider and Mr H. Geerlings (Hannover) for

statistical analyses, and C. Menzel (Hannover) for secretarial assistance.

Correspondence should be addressed to J. B., Kinderklinik, MedizinischeHochschule, D-3000 Hannover 61, West Germany.

REFERENCES

1. Lange K, Wasserman E, Slobody LB. Prolonged intermittent steroid therapy fornephrosis in children and adults. JAMA 1958; 168: 377-81.

2. Soyka LF, Saxena KM. Alternate-day steroid therapy for nephrotic children JAMA1965; 192: 225-30

3. Arneil GC. Management of the nephrotic syndrome Arch Dis Childh 1968; 43:257-62.

4. Abramowicz M, Ameil GC, Bamett HL, et al Controlled trial of azathiopnne inchildren with nephrotic syndrome. Lancet 1976; i: 959-61.

5. Arbeitsgemeinschaft fur Padiatnsche Nephrologie. Alternate-day versus intermittentprednisone in frequently relapsing nephrotic syndrome. Lancet 1979; i: 401-03.

6. International Study of Kidney Disease in Children. The primary nephrotic syndromein children. Identification of patients with minimal change nephrotic syndromefrom initial response to prednisone. J Pediatr 1981; 98: 561-64

7. Brodehl J, Krohn HP, Ehrich JHH. The treatment of minimal change nephroticsyndrome (lipoidnephrosis). cooperative studies of the Arbeitsgemeinschaft furPadiatrische Nephrologie (APN). Klin Padiat 1982, 194: 162-65.

8. Leisti S, Koskimies O. Risk of relapse in steroid-sensitive nephrotic syndrome. effectof stage of post-prednisone adrenocortical suppression. J Pediatr 1983; 103:

553-57.

9. International Study of Kidney Disease in Children. Primary nephrotic syndrome mchildren- clinical significance of histopathologic variants of minimal change and ofdiffuse mesangial hypercellularity. Kidney Int 1981; 20: 765-71.

10. Arbeitsgemeinschaft fur Padiatrische Nephrologie. Alternate-day prednisone is moreeffective than intermittent predisone in frequently relapsing nephrotic syndrome.Eur J Pediatr 1981; 135: 229-37.

11. Farnsworth EB Studies on influence of adrenocorticotrophin in acute nephntis, insimple nephrosis, and in nephrosis with azotemia. In: Mote JR ed. Proceedings ofthe first clinical ACTH conference. Philadelphia: Blakiston, 1950: 297.

12. Barnett HL, Forman CW, McNamara W, et al. The effect of adrenocorticotropichormone in children with the nephrotic syndrome. II. Physiologic observations ondiscrete kidney functions and plasma volume. J Clin Invest 1951; 30: 227-35.

13. Luetscher JA Jr, Deming QB, Harvey J, Lew W, Pool LJ. Treatment of nephrosiswith cortisone. J Clin Invest 1950; 29: 1576-87.

14 Metcoff J, Rance CP, Kelsey WM, Nakasone N, Janeway CA. Adrenocorticotrophichormone (ACTH) therapy of the nephrotic syndrome in children. Pediatrics 1952;10: 543-66.

15. Lange K, Slobody L, Strang R Treatment of nephrotic syndrome with interruptedACTH or oral cortisone therapy. Proc Soc Exper Biol Med 1953, 82: 315-17.

16. Lange K, Strang R, Slobody LB, Wenk EJ The treatment of the nephrotic syndromewith steroids in children and adults. Arch Intern Med 1957; 99: 760-70.

17. Comfeld D, Schwartz NW. Nephrosis: a long-term study of children treated withcorticosteroids. J Pediatr 1966, 68: 507-15.

18. Good RA, Vernier RL, Smith RT Serious untoward reactions of therapy withcortisone and adrenocorticotrophin in pediatric practice. Pediatrics 1957; 19:

95-118.19. Saxena KM, Crawford JD. The treatment of nephrosis N Engl J Med 1965; 272:

522-26

20. Lieberman E, Heuser E, Gilchrist GS, Donnell GN, Landing BH. Thrombosis,nephrosis and corticosteroid therapy. J Pediat 1968; 73: 320-28.

21. Lam CN, Arneil GC. Long-term dwarfing effects of corticosteroid treatment ofchildhood nephrosis. Arch Dis Childh 1968; 43: 589-94

22. Brodehl J. Nephrotic syndrome in children- diagnosis and treatment World PediatrChild Care 1986; 9-18.

23 Barnett HL, Schoeneman M, Bernstein J Edelmann CM Minimal change nephroticsyndrome. In: Edelmann CM ed Pediatric kidney disease Boston: Little, Brown,1978. 695-711.

24. Krensky AM, Ingelfinger JR, Grupe WE. Peritonitis in chilhood nephrotic syndrome.Am J Dis Child 1982; 136: 732-36.

25. Hoyer PF, Gonda S, Barthels M, Krohn HP, Brodehl J. Thromboemboliccomplications in children with nephrotic syndrome Risk and incidence. ActaPaediatr Scand 1986; 75: 804-10.

26 Leisti S, Koskimies O, Perheentupa J, Vilska J, Hallman N Idiopathic nephroticsyndrome prevention of early relapse Br Med J 1978; i: 892.

27. International Study of Kidney Disease in Children Nephrotic syndrome in children.Randomized trial comparing 2 prednisone regimens in steroid-responsive patientswho relapse early. J Pediatr 1979; 95: 239-43.