ORIGINAL ARTICLE

Treating severe Henoch-Schönlein and IgA nephritiswith plasmapheresis alone

Mohan Shenoy & Milos V. Ognjanovic &

Malcolm G. Coulthard

Received: 9 November 2006 /Revised: 24 March 2007 /Accepted: 27 March 2007 / Published online: 26 May 2007# IPNA 2007

Abstract The aim of our study was to determine theoutcome of children with severe Henoch-Schönlein nephri-tis (HSN) and immunoglobulin A (IgA) nephritis (IgAN)treated with early plasmapheresis alone. Children withacute renal impairment, heavy proteinuria or both andhistology greater than grade 3 were treated with earlyplasmapheresis alone. Glomerular filtration rate (GFR)estimated from plasma creatinine (eGFR), urine albumin:creatinine ratio (UA/UC) and blood pressure 2 weeks aftertreatment and were measured at the last follow-up. Sixteenchildren (14 HSN, 2 IgAN) had a mean eGFR of 56 (17–136) ml/min per 1.73 m2 and UA/UC of 590 (12–1,379)mg/mmol. Fifteen were referred at presentation and oneafter 2 months, and all commenced plasmapheresis within 6(2–13) days. All had at least nine exchanges of 90 ml/kgover 2 weeks. At 2 weeks, the eGFR had increased by 51(95% CI 34–68; P=0.002), and the UA/UC fell by 457(95% CI 241–673; P=0.0001). At last review after 4 (1–7.5)years, the late-referred child had required a renal transplantbut the other 15 had normal eGFRs (98–142), did not requirehypotensive medication, and had normal or minimallyelevated UA/UC (maximum 42). Children with severeHSN and IgAN recover well if treated with plasmapheresisalone without the need for immunosuppressive therapy. Arandomised trial is needed.

Keywords Plasmapheresis . Henoch-Schönlein disease .

IgA disease . Nephritis . Proteinuria . Renal failure

Introduction

Henoch-Schönlein nephritis (HSN) and immunoglobulin A(IgA) nephropathy (IgAN) share similar disturbances in theIgA system and are indistinguishable by renal histopathol-ogy. HSN has been reported in 15–62% of patients withHenoch-Schönlein purpura. Reports of full recovery fromHSN range from 65% to 95%, and the incidence of renalfailure ranges from 2.5% to 20% [1–6]. A follow-up studyof IgAN in children younger than 15 years showed that 5%developed renal failure [7]. Though a number of signs andsymptoms at presentation have been suggested to be usefulprognostic markers, no specific ones are reliable whenapplied to individual patients [8–10], though either protein-uria [10, 11] or a low glomerular filtration rate (GFR) [8] atfollow-up may indicate a poor outcome. As yet, there is noconsensus on therapy. Several uncontrolled treatmentstudies, often in patients who have had other priortreatments, have reported outcomes considered to be betterthan expected. There are reports of apparent benefit fromsteroids given alone [12, 13], in combination with azathi-oprine [14, 15] or cyclophosphamide [16, 17], with the usecyclosporine and angiotensin converting enzyme (ACE)inhibitors [18, 19], or other multiple combination therapies[20]. There is limited data to suggest that plasmapheresiswith [21, 22] or without [23] other immunomodulators canalter the long-term outcome of HSN and IgAN, especially ifinstituted early in the course of the disease [21–23]. In thisstudy, we retrospectively looked at the immediate and long-term outcome of children with severe HSN and IgAN whowere treated with early plasmapheresis as their sole therapy.

Pediatr Nephrol (2007) 22:1167–1171DOI 10.1007/s00467-007-0498-y

M. Shenoy :M. V. Ognjanovic :M. G. Coulthard (*)Department of Paediatric Nephrology,Royal Victoria Infirmary, Newcastle,Newcastle NE1 4LP, UKe-mail: malcolm.coulthard@nuth.nhs.uk

Methods

We retrospectively analysed the records of children referredto our tertiary paediatric nephrology unit from 1988 to2005, who had either HSN or IgAN, and who had either aGFR of <70 ml/min per 1.73 m2 [estimated from plasmacreatinine and height; estimated GFR (eGFR)] [24], ornephrotic range proteinuria estimated as a urinary albumin:creatinine ratio (UA/UC) of >200 mg/mmol. All patientsalso had at least grade 3 nephritis on renal biopsy [Interna-tional Study of Kidney Diseases in Children (ISKDC)] [11].Since 1988, it had become our policy to treat this group ofchildren with plasmapheresis alone and not to use steroids,other immunomodulators or antiplatelet agents.

Plasmapheresis was with 90 ml/kg (approximately twicethe plasma volume) per session, using central venous lineaccess, exchanging the first approximately 70 ml/kg for4.5% human albumin solution and the last 20 ml/kg forfresh frozen plasma to prevent depletion of clotting factors.All children received at least nine exchanges during the first2 weeks, and some had further treatment sessions atincreasingly spaced intervals if their clinical recoveryappeared incomplete. Because of the potential immunosup-pressive effects of plasmapheresis, all children receivedcotrimoxazole 12 mg/kg daily for the duration of thetreatment plus for another 2 months.

Table 1 Clinical and laboratory data of patients with Henoch-Schönlein nephritis (HSN) and immunoglobulin A nephritis (IgAN) treated withplasmapheresis alone

Cases Histology At presentation At last follow-up

Number Gender Age Diagnosis High BP Delay ISKDC grade Crescents eGFR UA/UC Years eGFR UA/UC

% Fibrous

1 F 11.0 HSN 3b 20 46 180 7.5 102 *2 F 6.8 HSN 3a 40 82 768 1.1 127 23 M 5.8 HSN 3b 24 93 979 2.1 98 34 M 15.0 HSN + 3b 20 20 146 2.5 108 385 F 3.7 HSN 3a 0 136 1379 6.2 102 26 F 13.5 HSN 4b 53 28 36 2.6 134 427 F 12.5 HSN 3b 43 61 350 3.1 101 108 M 11.8 HSN + 3b 0 33 245 3.8 142 19 M 12.3 HSN + 3b 10 90 936 1.1 101 710 F 10.1 IgAN 3b 29 + 42 766 4.2 106 211 M 13.1 IgAN 3b 5 17 248 3.4 113 1612 M 9.9 HSN + 3b 14 + 43 1037 5.2 105 913 F 8.4 HSN + 4b 52 64 1000 5.5 121 14.314 F 8.3 HSN + 3a 0 22 454 4.3 121 215 F 8.9 HSN + 3b 0 67 12 5.4 112 316 F 7.7 HSN + Yes 3b 26 + 29 913 6.3 Transplanted

HSN Henoch-Schönlein nephritis, IgAN IgA nephropathy, ISKDC International Study for Kidney Disease in Children, eGFR glomerular filtrationrate ml/min per 1.73 m2 calculated from plasma creatinine and height, UA/UC urinary albumin:creatinine ratio, mg/mmol, High BP systolic bloodpressure persistently above 95th centile for age*Value not available, but urine dipstick for albuminuria negative

0

50

100

150

200

0123

weeks before plasmapheresis

eGF

R m

l/min

/1.7

3m2

Fig. 1 Estimated glomerular filtration rate (eGFR) in the 3 weeksbefore starting plasmapheresis for Henoch-Schönlein or immunoglob-ulin A nephritis in the children who had eGFR values below 80 ml/minper 1.73 m2

1168 Pediatr Nephrol (2007) 22:1167–1171

The effect of treatment was assessed using paired t teststo compare the eGFR (measured at least 36 h after the lastplasmapheresis session to allow recovery from the artifi-cially reduced postexchange level), systolic blood pressure,and UA/UC at the start of plasmapheresis, 2 weeks aftercompleting the first 9 exchanges and at the last follow-upvisit. No child received ACE inhibitors or angiotensin IIblocking medication, and the need for other hypotensiveagents was noted.

Results

We treated 16 children (six boys), median age 10 years(range 3.7–15) during the 18 years from 1988 to 2005(Table 1). Fourteen had HSN and two had IgAN. Atpresentation, the mean eGFR was 56 ml/min per 1.73 m2

(range 17–136), and the mean albumin:creatinine ratio was590 mg/mmol (range 12–1,379). Eight children had both alow eGFR (<80 ml/min per 1.73 m2) and heavy proteinuria

(urine albumin:creatinine ratio >200 mg/mmol), whereasfour had just a low eGFR and four had just heavyproteinuria (this group all had urinary albumin:creatinineratios >700 mg/mmol). Over the short pretreatment moni-toring period, the impression was that the children with loweGFR values had either stable or worsening function(Fig. 1). All patients had haematuria, which was macro-scopic in half of them. Half the children were hypertensiveat presentation, with systolic blood pressures above the95th centile for age.

Fifteen children were referred from their local hospitalwithout delay, but one girl was not referred for 2 months.During this time, no blood tests were taken because she hada severe needle phobia despite the fact that she was knownto have marked haematuria and heavy proteinuria. Allchildren had a renal biopsy within a mean of 3 days (range1–9) of referral, which showed mesangial IgA on immu-nofluorescence. The ISKDC histological gradings werestage 3a in three, stage 3b in 12, and stage 4b in one. Ninechildren had between 5% and 53% cellular crescents, andthree had fibrous crescents. Two with fibrous crescents(including the late-referred girl) had other evidence ofpermanent renal damage, including tubular atrophy andinterstitial fibrosis.

Plasmapheresis was started 6 (range 2–13) days afterreferral. No child had received steroids, other immunemediators, or antiplatelet agents. All children had nineexchanges in the first 2 weeks, and after that, some childrenhad more exchanges at a lower frequency, according toclinical response. A mean of 17 (range 9–28) sessions wasperformed. The only complication of plasmapheresis wasthat some children sometimes developed an itchy rashwhilst receiving fresh frozen plasma, which requiredtreatment with an intravenous hydrocortisone 1 mg/kg andchlorphenamine 0.2 mg/kg.

At 2 weeks after completing nine plasmapheresissessions (Fig. 2), the eGFR had risen by 51, to 107 ml/min per 1.73 m2 (95% CI for increase, 34–68; P=0.002),and the albumin:creatinine ratio had fallen by 457, to133 mg/mmol (95% CI for fall, 241–673; P=0.0001). Atthe final follow-up assessment, four (range 1–7.5) yearsafter presentation, the late-referred child had required arenal transplant, whereas the 15 children referred promptlywere all well. All 15 had a normal eGFR at 113 (range 98–142) ml/min per 1.73 m2, a normal or only a slightlyelevated UA/UC (maximum 42 mg/mmol) and nonerequired hypotensive medication.

Discussion

A major difficulty in interpreting the impact of treatmentfor aggressive HSN and IgAN, in which children present

0

50

100

150eG

FR, m

l/min

/1.7

3m2

presentation 2 weeks last follow-up

0

500

1000

1500

Uri

nar

y al

bu

min

:cre

atin

ine

rati

o (U

A/U

C),

mg

/mm

ol

transplanted

Fig. 2 Estimated glomerular filtration rate (eGFR) and urinaryalbumin:creatinine ratios at presentation, 2 weeks after startingtreatment and at the last follow-up in 15 children treated for Henoch-Schönlein or immunoglobulin A nephritis by plasmapheresis alonewithout delay (open circles, plain lines) and one child treated after2 months’ delay (solid circles, broken lines)

Pediatr Nephrol (2007) 22:1167–1171 1169

with nephrotic-range proteinuria or with a low GFR orboth, is that it is impossible to know whether some of theindividuals would have had a good outcome withouttreatment. Combined with this is the fact that there hasonly been one randomised controlled trial (RCT) toevaluate a treatment modality [25]. Thus, we are mainlyleft comparing the outcomes for children treated with oneform of intervention in one centre with the outcomes for acompletely different group of children treated in anotherway, possibly with different selection criteria and afterdifferent pretreatment intervals, carried out elsewhere. Thenumbers of cases would dictate that such a trial would needto be multicentre and probably multinational.

Within this context, our relatively small single-centreretrospective study, with a only a moderately short follow-up period, serves to argue the case for an RCT for thisgroup of children that would include a limb where theyreceive prompt treatment with plasmapheresis alone. Thefact that all 15 promptly referred children had at worst mildproteinuria 4 years later is encouraging, even allowing forthe fact that some of the recovery could have occurredspontaneously. Although deterioration in renal function hasbeen described many years after apparent recovery [8, 9],this is extremely unlikely to occur in our children, who allhad a GFR above 70 ml/min per 1.73 m2 [2, 8] and little orno proteinuria at follow-up [10, 11], which both stronglypredict a good outcome. By contrast to the children we sawand treated promptly, the one child who was referred latewith chronic renal injury progressed to end-stage renalfailure. Although this was just one case, the importance ofearly intervention has been stressed previously [26].

Only one other group has used plasmapheresis alone inchildren with HSN [23]. They treated nine very severelyaffected children (mean eGFR 46.5 ml/min per 1.73 m2;proteinuria 4.9 gm/m2 per day; biopsy ISKDC grade 5), ofwhom six had a good outcome a decade later. The authorsalso concluded that early treatment was important inachieving success. Similarly, plasmapheresis being com-bined with immunosuppressive drugs appeared to be moreeffective after early referral in an uncontrolled series ofchildren with IgAN [21], and this relationship was con-firmed statistically in children with HSN [22].

On the basis of crude comparisons, the outcomes afterplasmapheresis appear to be at least as good as afterimmunosuppressive drug treatments, but it is relativelyinvasive, requiring good vascular access, considerablenursing time, and exposure to blood products from multipledonors. Despite this, we have not seen any serious clinicalside effects from it in our children. Some had transientsymptoms and rashes during infusion with some units offresh frozen plasma, requiring small doses of cortico-steroids and antihistamines (which we feel were unlikelyto have had any influence on clinical outcomes). Nonethe-

less, theoretical risks exist, and another child in ourdepartment acquired hepatitis C from plasmapheresis foranother form of glomerulonephritis shortly before screeningfor the virus was introduced for blood donations. No patienthad a significant infective illness during their treatment.The mechanism of benefit from plasmapheresis is unclear,though removal of circulating IgA complexes or proin-flammatory mediators is widely assumed. It is interesting tonote that there is a compelling report of plasmapheresisapparently reversing a case of acute Henoch-Schönleincerebral vasculitis [27].

As with plasmapheresis, reports of immunosuppressivedrug treatments have been mainly from uncontrolled seriesand can therefore only suggest that there may be benefit.Although reports of treatments with steroids alone [12, 13],with azathioprine [14, 15] or cyclophosphamide [16, 17]and other drug combinations [20] do suggest some benefit,all of these medications have potential serious side effects.In the case of cyclosporine [18, 19], some children alsoremained cyclosporin dependent. There has been onerandomised controlled trial of treating HSN, comparingsupportive treatment alone to cyclophosphamide, but thesechildren did not do well, with half of those presenting withheavy proteinuria failing to recover [25]. This may beentirely or in part because the study design dictated thatchildren had to have heavy proteinuria, with or without apoor GFR, for at least a month before enrolment, prevent-ing the early intervention that uncontrolled studies haveconsistently noted to be important.

In conclusion, this uncontrolled study gives the impres-sion that plasmapheresis may have had a beneficial effecton severe HSN and IgAN without risking the potential sideeffects of immunosuppressive therapies. There is a need fora large multicentre RCT in which one limb is promptplasmapheresis alone.

References

1. Meadow SR, Glasgow EF, White RHR, Moncrief MW, CameronJS, Ogg CS (1972) Schönlein-Henoch nephritis. Q J Med 41:241–258

2. Counahan R, Winterborn MH, White RH, Heaton JM, MeadowSR, Bluett NH, Swetschin H, Cameron JS, Chantler C (1977)Prognosis of Henoch-Schonlein nephritis in children. BMJ 2:11–14

3. Koskimies O, Mir S, Rapola J, Vilska J (1981) Henoch-Schonleinnephritis: long-term prognosis of unselected patients. Arch DisChild 56:482–484

4. Stewart M, Savage JM, Bell B, McCord B (1988) Long term renalprognosis of Henoch-Schonlein purpura in an unselected child-hood population. Eur J Pediatr 147:113–115

5. Coppo R, Mazzucco G, Cagnoli L, Lupo A, Schena FP (1997)Long-term prognosis of Henoch-Schonlein nephritis in adults andchildren. Italian Group of Renal Immunopathology CollaborativeStudy on Henoch-Schonlein purpura. Nephrol Dial Transplant12:2277–2283

1170 Pediatr Nephrol (2007) 22:1167–1171

6. Narchi H (2005) Risk of long term renal impairment and durationof follow up recommended for Henoch-Schönlein purpura withnormal or minimal urinary findings: a systematic review. Arch DisChild 90:916–920

7. Yoshikawa N, Ito H, Nakamura H (1992) Prognostic indicators inchildhood IgA nephropathy. Nephron 60:60–67

8. Bunchman TE, Mauer SM, Sibley RK, Vernier RL (1988)Anaphylactoid purpura: characteristics of 16 patients who pro-gressed to renal failure. Pediatr Nephrol 2:393–397

9. Ronkainen J, Nuutinen M, Koskimies O (2002) The adult kidney24 years after childhood Henoch-Schönlein purpura. A retrospec-tive cohort study. Lancet 360:666–670

10. Coppo R, Andrulli S, Amore A, Gianoglio B, Conti G, Peruzzi L,Locatelli F, Cagnoli L (2006) Predictors of outcome of Henoch-Schonlein nephritis in children and adults. Am J Kidney Dis47:993–1003

11. Farine M, Poucell S, Geary DL, Baumal R (1986) Prognosticsignificance of urinary findings and renal biopsies in children withHenoch-Schonlein nephritis. Clin Pediatr (Phila) 25:257–259

12. Niaudet P, Habib R (1998) Methylprednisolone pulse therapy inthe treatment of severe forms of Schönlein-Henoch nephritis.Pediatr Nephrol 12:238–243

13. Kawasaki Y, Suzuki J, Nozawa R, Suzuki S, Suzuki H (2003)Efficacy of methylprednisolone and urokinase pulse therapy forsevere Henoch-Schönlein nephritis. Pediatrics 111:785–789

14. Bergstein J, Leiser J, Andreoli SP (1998) Response of crescenticHenoch-Schonlein purpura nephritis to corticosteroid and azathi-oprine therapy. Clin Nephrol 49:9–14

15. Shin JI, Park JM, Shin YH, Kim JH, Lee JS, Kim PK, Jeong HJ(2005) Can azathiaprine and steroids alter the progression ofsevere Henoch-Schonlein nephritis in children? Pediatr Nephrol20:1087–1092

16. Flynn JT, Smoyer WE, Bunchman TE, Kershaw DB, Sedman AB(2001) Treatment of Henoch-Schonlein purpura glomerulonephri-tis in children with high-dose corticosteroids plus oral cyclophos-phamide. Am J Nephrol 21:128–133

17. Kawasaki Y, Suzuki J, Suzuki H (2004) Efficacy of methylprednis-olone and urokinase pulse therapy combined with or without cyclo-phosphamide in severe Henoch-Schonlein nephritis: a clinical andhistopathological study. Nephrol Dial Transplant 19:858–864

18. Ronkainen J, Autio-Harmainen H, Nuutinen M (2003) Cyclo-sporin A for the treatment of severe Henoch-Schonlein glomer-ulonephritis. Pediatr Nephrol 18:1138–1142

19. Shin JI, Park JM, Shin YH, Kim JH, Kim PK, Lee JS, Jeong HJ(2005) Cyclosporin A therapy for severe Henoch-Schonleinnephritis with nephrotic syndrome. Pediatr Nephrol 20:1093–1097

20. Iijima K, Ito-Kariya S, Nakamura H, Yoshikawa N (1998)Multiple combined therapy for severe Henoch-Schonlein nephritisin children. Pediatr Nephrol 12:244–248

21. Coppo R, Basolo B, Roccatello D, Giachino O, Lajolo D, MartinaG, Rollino C, Amore A, CostaM, Piccoli G (1985) Plasma exchangein progressive primary IgA nephropathy. Int J Artif Organs 8:55–58

22. Gianviti A, Trompeter RS, Barratt TM, Lythgoe MF, Dillon MJ(1996) Retrospective study of plasma exchange in patients withidiopathic rapidly progressive glomerulonephritis and vasculitis.Arch Dis Child 75:186–190

23. Hattori M, Ito K, Konomoto T, Kawaguchi H, Yoshioka T, KhonoM (1999) Plasmapheresis as the sole therapy for rapidly progres-sive Henoch-Schönlein purpura nephritis in children. Am J KidneyDis 33:427–433

24. Schwartz GJ, Haycock GB, Edelmann CMJ, Spitzer A (1976) Asimple estimate of glomerular filtration rate in children derivedfrom body length and plasma creatinine. Pediatrics 58:259–263

25. Tarshish P, Bernstein J, Edelmann CMJ (2004) Henoch-Schönleinpurpura nephritis: course of disease and efficacy of cyclophos-phamide. Pediatr Nephrol 19:51–56

26. Shin JI, Park JM, Kim JH, Lee JS, Jeong HJ (2006) Factorsaffecting histological regression of crescentic Henoch-Schonleinnephritis in children. Pediatr Nephrol 21:54–59

27. Wen Y-K, Yang Y, Chang C-C (2005) Cerebral vasculitis andintracerebral hemorrhage in Henoch-Schonlein purpura treatedwith plasmapheresis. Pediatr Nephrol 20:223–225

Pediatr Nephrol (2007) 22:1167–1171 1171