clinical investigation see commentary on page 1019

Transcription

1 & 2013 International Society of Nephrology clinical investigation see commentary on page 1019 Prompt plasma exchanges and immunosuppressive treatment improves the outcomes of anti-factor H autoantibody-associated hemolytic uremic syndrome in children Aditi Sinha 1, Ashima Gulati 1, Savita Saini 1, Caroline Blanc 2, Aarti Gupta 1, Bahadur Singh Gurjar 3, Himanshi Saini 1, Shambhuprasad T. Kotresh 2, Uma Ali 4, Divya Bhatia 1, Alpana Ohri 4, Manish Kumar 5, Indira Agarwal 6, Sanjeev Gulati 7, Kanav Anand 8, M. Vijayakumar 9, Rajiv Sinha 10, Sidharth Sethi 1, Maud Salmona 2, Anna George 3, Vineeta Bal 3, Geetika Singh 11, Amit K. Dinda 11, Pankaj Hari 1, Satyajit Rath 3, Marie-Agnes Dragon-Durey 2 and Arvind Bagga 1 for the Indian HUS Registry 1 Division of Nephrology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India; 2 Laboratoire d Immunologie, Hôpital Européen Georges Pompidou, INSERM UMRS 872, team 13 and Paris Descartes University, Paris, France; 3 National Institute of Immunology, New Delhi, India; 4 Department of Pediatrics, BJ Wadia Hospital for Children, Mumbai, India; 5 Chacha Nehru Bal Chikitsalaya, New Delhi, India; 6 Department of Pediatrics, Christian Medical College, Vellore, India; 7 Department of Nephrology, Fortis Hospitals, New Delhi, India; 8 Division of Pediatric Nephrology, Sir Ganga Ram Hospital, New Delhi, India; 9 Department of Pediatric Nephrology, Mehta Children s Hospital, Chennai, India; 10 Institute of Child Health, Kolkata, India and 11 Department of Pathology, All India Institute of Medical Sciences, New Delhi, India Antibodies to complement factor H are an uncommon cause of hemolytic uremic syndrome (HUS). Information on clinical features and outcomes in children is limited. In order to explore this we studied a multicenter cohort of 138 Indian children with anti-complement factor H antibody associated HUS, constituting 56% of patients with HUS. Antibody titers were high (mean 7054 AU/ml) and correlated inversely with levels of complement C3, but not complement factor H. Homozygous deletion of the CFHR1 gene was found in 60 of 68 patients. Therapies included dialysis in 119 children, 105 receiving plasma exchanges and 26 intravenous immunoglobulin. Induction immunosuppression consisted of 87 children receiving prednisolone with or without intravenous cyclophosphamide or rituximab. Antibody titers fell significantly following plasma exchanges and increased during relapses. Adverse outcome (stage 4-5 CKD or death) was seen in 36 at 3 months and 41 by last follow up, with relapse in 14 of 122 available children. Significant independent risk factors for adverse outcome were an antibody titer over 8000 AU/ml, low C3 and delay in plasma exchange. Combined plasma exchanges and induction immunosuppression resulted in significantly improved renal survival: one adverse outcome prevented for every 2.6 patients treated. Maintenance immunosuppressive therapy, Correspondence: Arvind Bagga, Division of Nephrology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi , India. arvindbagga@hotmail.com Received 1 December 2012; revised 22 July 2013; accepted 1 August 2013; published online 2 October 2013 of prednisolone with either mycophenolate mofetil or azathioprine, significantly reduced the risk of relapses. Thus, prompt use of immunosuppressive agents and plasma exchanges are useful for improving outcomes in pediatric patients with anti-complement factor H-associated HUS. Kidney International (2014) 85, ; doi: /ki ; published online 2 October 2013 KEYWORDS: acute kidney injury; complement; hemolytic uremic syndrome; intravenous immunoglobulin; plasma exchange Hemolytic uremic syndrome (HUS) is an important cause of acute kidney injury in children. The current classification distinguishes shigatoxin-associated HUS from atypical HUS, the latter chiefly mediated by an activated alternative complement pathway. 1 A proportion of patients show mutations in genes encoding regulators of complement activation, commonly complement factor H (CFH). 2 5 In some patients, atypical HUS is secondary to anti-cfh autoantibodies, which bind to multiple epitopes on CFH, impairing its functions. 5,6 HUS associated with anti-cfh autoantibodies is described in a frequency of 5 25% in cohorts from France, Germany, Spain and UK, 7 12 higher proportions being seen in pediatric series. The diagnosis is important, as therapy with plasma exchange and immunosuppression is believed to result in favorable outcomes Prospective information on clinical features and outcome is, however, limited. 12,14 17 Through collaboration between the All India Institute of Medical Sciences (AIIMS) and Laboratoire d Immunologie, Kidney International (2014) 85,

2 clinical investigation A Sinha et al.: Anti-factor H autoantibody-associated HUS Hôpital Européen Georges Pompidou, we set up testing for CFH and anti-cfh antibodies in New Delhi in These tests were later offered to patients from other centers in the country. We report the clinical features and outcome in 138 children with anti-cfh antibody-associated HUS. Almost one-half of these patients were managed at one center; the outcome of 11 patients has been reported previously. 17 RESULTS During 2007 to 2013, of 246 patients with HUS from 26 centers, 138 (56.1%) had anti-cfh antibodies. These included 109 consecutive patients evaluated at AIIMS, of which 62 (56.9%) showed these antibodies. Clinical characteristics at baseline in patients with and without anti-cfh antibodies are shown in Table 1. The mean age at onset of HUS in the former was 8.4±4.1 years (Supplementary Figure S1 online). Patients presented throughout the year, with a peak between November and March, compared to late spring or summer for those without anti-cfh antibodies (Supplementary Figure S2 online). Patients with antibodyassociated HUS had prolonged oligoanuria, and showed lower levels of hemoglobin, platelets, and complement C3. Hepatic or neurological features were present in 59 (60.8%), stage 2 hypertension in 91 (65.9%), and nephrotic range proteinuria in 81 (58.7%). Unless specified, subsequent details refer to patients with anti-cfh antibody-associated HUS. Thirteen patients had history of diarrhea and 67 had preceding febrile illness. Two patients developed peripheral gangrene soon after presentation. 18 One patient each showed double-stranded DNA antibodies (screened in 82), antineutrophil cytoplasmic antibodies (n ¼ 62) and antibodies to leptospira (n ¼ 33); Plasmodium vivax infection was present in six. Nine patients presented during a relapse, 4 43 months after an episode of HUS, from which they had recovered. During the previous illness, six patients required dialysis and renal histology showed thrombotic microangiopathy in four. Anti-CFH antibody titers Blood samples were initially screened at Laboratoire d Immunologie (n ¼ 11), and later at both centers (n ¼ 24) or at AIIMS (n ¼ 103). At presentation, the mean±s.e.m. titer was ± U/ml. The titer was below 1000 AU/ ml in 34 (24.6%) patients, AU/ml in 61 (44.2%), ,000 AU/ml in 24 (17.2%), 10,000 15,000 AU/ml in 12 (8.7%), and 415, 000 AU/ml in 7 (5.1%) patients. Followup, in 75 patients, showed that antibody titers reduced significantly by 76.1±35.1% and 75.9±35.0% following plasma exchange with and without intravenous (IV) immunoglobulin, respectively (Figure 1a and b). Isotyping on 27 samples from 10 patients showed predominantly IgG3 antibodies, in all except one patient (Supplementary Figure S3 online). Other investigations Seventy five (62.0%) of 121 patients showed complement C3 level below 70 mg/dl. There was negative correlation between antibody titer and C3 (r ¼ 0.24; P ¼ 0.014). Titers of anti- CFH antibody were 11, 622.6± and ±586.2 AU/ ml (P ¼ 0.025) in those with low and normal C3, respectively. Table 1 Comparison of baseline clinical and biochemical features in patients with HUS with and without antibodies to CFH Variable Anti-CFH HUS (n ¼ 138) Other HUS (n ¼ 108) P Boys 100 (72.5%) 61 (56.5%) Age, years 8.4±4.1 (6 10.2) 5.1±4.9 (1 9.2) o Presentation during relapse 9 (6.5%) 5 (4.6%) 0.27 Duration of oligoanuria, days 10.6±10.4 (3 14) 6.0±5.8 (2 10) o Anuria 58 (42.0%) 34 (31.5) Prodrome Febrile illness 67 (48.6%) 59 (54.6%) Diarrhea, dysentery 13 (9.4%) 34 (31.5%) o Jaundice, elevated transaminases 79 (57.3%) 41 (38.0%) Seizures 56 (40.6%) 31 (28.7%) Stage 2 hypertension 91 (65.9%) 39 (36.1%) o Hemoglobin, g/dl 5.5±1.3 ( ) 6.2±1.8 ( ) Platelet count, 10 3 /mm ±37.6 (32 81) 91.3±87.0 ( ) Reticulocyte count, % 9.7±9.3 ( ) 5.0±5.8 ( ) o Hematuria 56 (40.6%) 20 (18.5%) Nephrotic range proteinuria 81 (58.7%) 39 (36.1%) Blood creatinine, mg/dl 5.76±2.60 ( ) 4.76±2.87 ( ) Lactate dehydrogenase, IU/l ± ( ) ± ( ) 0.84 Complement C3, mg/dl 69.0±25.8 ( ) 82.2±36.2 ( ) Anti-CFH antibody, AU/ml a ± ( ) 30.7±4.37 (0 44.6) Abbreviations: AU, arbitrary units; CFH, complement factor H; HUS, hemolytic uremic syndrome. Data are presented as mean±standard deviation (interquartile range). C3 levels were estimated in 121 and lactate dehydrogenase levels in 122 patients with anti-factor H-associated HUS. C3 levels were estimated in 70 and LDH levels in 76 patients with other forms of HUS. a Mean±standard error of mean (interquartile range) Kidney International (2014) 85,

3 A Sinha et al.: Anti-factor H autoantibody-associated HUS clinical investigation a Titers of antibodies to factor H (AU/ml) c Titers of antibodies to factor H (AU/ml) 25,000 20,000 15,000 10, ,000 14,000 12,000 10, ±617.2 AU/ml At onset ± AU/ml At onset Disease remission 888.1±146.2 AU/ml After plasma exchanges ±817.5 AU/ml 844.3± AU/ml At relapse Titers of antibodies to factor H (AU/ml) Titers of antibodies to factor H (AU/ml) 200,000 20,000 15,000 10, , ,000 30,000 25,000 20,000 15,000 10, At onset ± AU/ml After IVIG and plasma exchanges Adverse outcome ± AU/ml Non-adverse outcome ±428.6 AU/ml ±242.5 AU/ml Figure 1 Anti-complement factor H antibody titers (mean±s.e.m.) in relation to therapy and outcome. (a) Antibody titers, repeated in 54 patients at a median of 32 (interquartile range 11 84) days of plasma exchanges, were significantly lower (Po0.0001) compared with those at onset. (b) Following therapy with intravenous immunoglobulin (IVIG) and plasma exchange (n ¼ 21), antibody titers declined significantly (Po0.0001) at a median of 15 (interquartile range 9 28) days. Most patients in both groups were also receiving immunosuppressive medications. (c) Serial samples, in 10 patients, showing that titers declined at remission (P ¼ ) and increased during relapse (P ¼ 0.006). (d) Antibody titers were significantly higher at disease presentation in patients with adverse outcome at 3 months (P ¼ 0.010). b d.. Table 2 Results of multiplex ligation-dependent probe amplification in patients with anti-cfh antibodies and healthy controls Results Anti-CFH antibodies n ¼ 68 Controls n ¼ 84 Homozygous deletion CFHR1, CFHR3 56 (82.3%) 8 (9.5%) Homozygous deletion CFHR1; heterozygous deletion CFHR3 4(5.9%) 0 Heterozygous deletion CFHR1, CFHR3 6 (8.8%) 29 (34.5%) Normal CFHR1; heterozygous deletion CFHR3 0 1 (1.2%) Normal CFHR1, CFHR3 2 (2.9%) 46 (54.8%) Abbreviation: CFH, complement factor H; CFHR, complement factor H related. While levels of CFH were lower in patients (183.8±76.6 mg/l) versus 50 healthy controls (224.0±52.0 mg/l; P ¼ 0.011), there was no correlation between these levels and antibody titers (r ¼ 0.13; P ¼ 0.31). Histology Of 57 biopsies, 29 (50.9%) showed thrombotic microangiopathy involving arterioles, 9 had chiefly glomerular involvement, and 19 had mixed pattern. Organized thrombi were seen in 17 patients, mesangiolysis in 27, and glomerular ischemia with contracted tuft in 19. Acute tubular necrosis and patchy cortical necrosis were seen in 27 and 10 patients, respectively. Immunofluorescence showed fibrin deposits in arterioles in 19 and glomeruli in 14 patients. CFHR1/3 deletion and studies in families and controls Multiplex ligation-dependent probe amplification in 68 patients showed homozygous deletion of both CFHR1 and CFHR3 genes in 56 (82.4%) (Table 2). Six patients were heterozygous for this deletion, and 4 had homozygous deletion in CFHR1 and heterozygous deletion in CFHR3. The overall prevalence of CFHR1 deletion in these patients was 88.2% (95% confidence interval (CI) ). The mean antibody titer in patients with homozygous CFHR1 deletion (n ¼ 60) was ± AU/ml, compared with ± AU/ml in heterozygous deletion or normal CFHR1 alleles (n ¼ 8; P ¼ 0.25). Of 84 healthy volunteers, homozygous CFHR1 deletion was present in 8 (9.5%; 95% CI ). The odds of anti-cfh antibody-mediated HUS in the presence of homozygous CFHR1 deletion were 67.4 (95% CI ; Po0.0001). The frequency of abnormal CFHR1 allele was 0.27 among healthy controls and 0.93 among patients with antibody-associated HUS. The observed genotype frequencies for CFHR1 in healthy population were consistent with the Hardy Weinberg principle (P ¼ 0.28), but not in those Kidney International (2014) 85,

4 clinical investigation A Sinha et al.: Anti-factor H autoantibody-associated HUS with antibody-associated HUS (P ¼ ). No copy number variations of CFHR2 and CFHR5 were detected. Table 3 shows the results on multiplex ligation-dependent probe amplification screening in 15 families of patients with HUS. The parents and siblings showed variable abnormalities in CFHR1/3, confirming an autosomal pattern of inheritance. The mean±s.e.m. anti-cfh antibody levels were 59.0± 8.2 AU/ml in 45 parents and 79.8±16.4 AU/ml in 17 siblings, compared with 42.5±2.4 AU/ml in 90 controls (normal o150 AU/ml). Four healthy siblings, aged 3 15 years, showed high antibody titers ranging from to AU/ml in association with homozygous deletions of CFHR1/3. Urinalyses and renal functions in these siblings were normal. Antibody-negative HUS Clinical information for 108 patients is shown in Table 1. The evaluation of these patients was incomplete; stools were not screened for shigatoxin. Absence of CFHR1 protein was found in 7 (20%) of 35 patients screened by western blot. Table 3 Results of multiplex ligation-dependent probe amplification in family members of 15 patients with anti-cfh antibodies and deletion in CFHR1 and/or CFHR3 genes Results Father (n ¼ 11) Mother (n ¼ 15) Siblings (n ¼ 14) Homozygous deletion CFHR1, CFHR3, n ¼ 12 patients a Homozygous deletion CFHR1, CFHR Heterozygous deletion CFHR1, CFHR Normal CFHR1; heterozygous deletion CFHR3 Normal CFHR1, CFHR Homozygous deletion CFHR1, heterozygous deletion CFHR3, n ¼ 2patients Homozygous deletion CFHR1, CFHR Homozygous deletion CFHR1, heterozygous deletion CFHR3 Heterozygous deletion CFHR1, heterozygous deletion CFHR3 Heterozygous deletion CFHR1, normal CFHR Heterozygous deletion CFHR1, heterozygous deletion CFHR3, n ¼ 1patient Heterozygous deletion CFHR1, ND 1 0 heterozygous deletion CFHR3 Homozygous deletion CFHR1, homozygous deletion CFHR3 ND 0 1 Abbreviations: CFH, complement factor H; CFHR, complement factor H related; ND, not done. a Three fathers in this group were not tested. Sequencing of genes encoding CFH, complement factor I, and membrane cofactor protein, done in 10 patients, showed abnormalities in three. Homozygous pathogenic mutations in CFH in two patients were: c3590, T4C, V1197A in C-terminus short consensus repeat 20 and c delA- TAG in exon 23. A heterozygous variation in exon 12 of CFI resulting in amino-acid change (c.1505 G4T, R502L) was present in one; the change was predicted by Alamut (Interactive Biosoftware, Rouen, France) to be pathogenic. Therapy Of 138 patients with antibody-associated HUS, 119 (86.2%) required peritoneal or hemodialysis. The duration of dialysis was 28.2±21.7 days in patients who recovered renal function. Plasma exchange was performed in 105 patients, 19.2±15.2 days from onset of illness, for 29.9±21.6 days. Eighty-three patients received five or more daily sessions of plasma exchange; 15 additional patients received 2 10 plasma infusions. Patients with high antibody titers (n ¼ 14) or delayed hematological remission (n ¼ 12) received IV immunoglobulin 43.5±43.0 days from onset. Induction immunosuppression initiated 27.1±20.5 days from onset in 87 patients, comprised of oral prednisolone with or without additional immunosuppressive agents, including IV cyclophosphamide in 49 and IV rituximab in 18 cases. Subsequently, maintenance immunosuppression, administered to 47 patients, included prednisone alone (21), or combined with mycophenolate mofetil (18) or azathioprine (8). Outcome Information on outcome at 3 months or more was available for 122 patients. Hematological remission was achieved at 31.0±22.0 days from onset. Thirty-six patients were dialysis dependent at 3 months, of which three did not require dialysis beyond 4.5 months (Table 4). Fourteen patients relapsed at mean duration of 9.4±8.8 (range ) months, including one patient on maintenance hemodialysis who had hematological relapse and seizures. Anti-CFH antibody titers fell significantly during remission and increased during relapse (Figure 1c). Four patients were receiving immunosuppressive agents at relapse; 7 patients did not recover renal function and showed dialysis dependence. Relapse-free survival was 88.9% at 6 months, 85.4% at 12 months, and 72.6% at last follow-up. Table 4 Outcome of patients with anti-cfh antibody associated hemolytic uremic syndrome (HUS), n ¼ 122 Outcome At 3 months At last follow-up CKD stage 1 (estimated GFR X90 ml/min per 1.73 m 2 ) With normal urinalysis 10 (8.2%) 13 (10.7%) With hypertension stage 2, hematuria, or proteinuria X2 þ 52 (42.6%) 58 (47.5%) CKD stages 2 3 (estimated GFR ml/min per 1.73 m 2 ) 24 (19.7%) 10 (8.2%) Adverse outcome (CKD stage 4 5; patient death) 36 (29.5%) 41 (33.6%) Abbreviations: CKD, chronic kidney disease; GFR, glomerular filtration rate. Adverse outcome includes 16 and 20 patient deaths at 3 months and at last follow-up, respectively Kidney International (2014) 85,

5 A Sinha et al.: Anti-factor H autoantibody-associated HUS clinical investigation At 14.5±18.0 (range 3 95) months follow-up, 41 patients had an adverse outcome (Table 4) including 33 patients with dialysis dependence at 3 5 months, 7 with renal failure following relapse and 1 with progressive renal impairment. Twenty patients, all with dialysis dependence, died of complications of renal failure (n ¼ 16) or septicemia (n ¼ 4). These included 16 patients who died within 3 months, of which 10 succumbed at 27±16 days without hematological remission. The mean antibody titer in these 10 patients was higher at 23, AU/ml than others (mean AU/ml; P ¼ 0.016). Renal survival was 68.4% at 6 months, 62.2% at 12 months, and 55.1% at last follow-up. Two patients received live-related and one a cadaveric transplant following 4 6 sessions of pre-transplant plasma exchanges (n ¼ 2), and perioperative IV immunoglobulin (n ¼ 2) and rituximab (n ¼ 2). At 6 15 months post transplantation, patients had satisfactory allograft function with antibody levels ranging between AU/ml. Determinants of outcome Table 5 shows that determinants of adverse outcome at 3 months and last follow-up were peak creatinine, high antibody titer, low C3 at presentation, delayed hematological remission, acute cortical necrosis, and need for prolonged dialysis. The mean antibody titer at onset was 16, 156.4± AU/ml in patients with adverse outcome at 3 months, compared to ±428.6 AU/ml in those who recovered renal function (P ¼ 0.010) (Figure 1d). Early institution of plasma exchange, performance of five or more exchanges, and its combination with induction immunosuppression were associated with recovery of function. On receiver operating characteristic analysis, predictors for adverse outcome were hematological remission X5 weeks from onset (area under the curve (AUC) 82.2%; sensitivity 78.6%, specificity 78.5%), anti-cfh antibody X8000 AU/ml (AUC 62.3%; sensitivity 69.3%, specificity 65.6%), time from onset to plasma exchange X17 days (AUC 73.8%; sensitivity 70.8%, specificity 67.1%), and dialysis for X1 month (AUC 75.0%; sensitivity 71.4%, specificity 62.3%). On multivariate analysis, factors associated with adverse outcome at 3 months were antibody titer X8000 AU/ml (odds ratio (OR) 6.92, 95% CI ; P ¼ 0.002) and time to plasma exchange X17 days (OR 6.11, 95% CI ; P ¼ 0.002), whereas combined therapy with plasma exchange and induction immunosuppression (OR 0.22, 95% CI ; P ¼ 0.016) resulted in favorable outcome. Independent predictors of adverse outcome on longterm follow-up were antibody titer X8000 AU/ml (hazard Table 5 Determinants of adverse outcome at 3 months and at last follow-up At 3 months Last follow-up Parameter Odds ratio (95% CI) P Hazards ratio (95%CI) P Age, years 1.02 (0.91, 1.14) (0.90, 1.08) 0.77 Anuria 3.25 (1.43, 7.39) (1.36, 5.23) Duration of oliguria, days 1.04 (1.00, 1.08) (1.01, 1.05) Jaundice, raised transaminases 0.79 (0.28, 2.21) (0.40, 2.16) 0.86 Seizures 3.01 (1.18, 7.67) (0.85, 3.68) 0.13 Severe hypertension 0.83 (0.51, 1.35) (0.57, 1.22) 0.35 Febrile illness at onset 1.34 (0.49, 3.66) (0.69, 4.37) 0.25 Hemoglobin o5 g/dl 0.48 (0.19, 1.24) (0.28, 1.28) 0.19 Platelets o100, 000/mm (0.31, 3.62) (0.42, 2.85) 0.85 C3 o70 mg/dl 3.75 (1.42, 9.93) (1.07, 5.06) Peak creatinine, mg/dl 1.24 (1.06, 1.46) (1.02, 1.28) CFH antibody X8000 AU/ml 3.29 (1.38, 7.82) (1.27, 4.41) CFH, mg/l 1.00 (1.00, 1.01) (1.00, 1.01) 0.26 Renal histology Mesangiolysis 1.2 (0.33, 4.32) (0.43, 3.38) 0.73 Acute tubular necrosis 0.63 (0.18, 2.20) (0.28, 1.97) 0.55 Acute cortical necrosis (1.19, 109.0) (1.36, 12.54) Tubular atrophy, interstitial fibrosis 1.35 (0.36, 5.01) (0.33, 2.78) 0.94 Hematological remission X5 weeks (3.27, 54.54) o (2.00, 17.58) Dialysis X1 month (3.70, 33.52) o (1.57, 8.02) Plasma exchange X5 sessions 0.05 (0.01, 0.24) o (0.03,0.16) o Time to plasma exchange X17 days 4.96 (1.81,13.57) (1.40, 7.42) Induction immunosuppression 0.28 (0.11, 0.65) (0.17, 0.59) o IV immunoglobulin 2.61 (1.05, 6.48) (0.60, 2.43) 0.60 Combined therapy a 0.21 (0.09, 0.48) o (0.15, 0.34) o Maintenance immunosuppression b 0.15 (0.03, 0.82) Occurrence of relapse b 2.86 (0.92, 8.90) Abbreviations: AU, arbitrary units; CFH, complement factor H; CI, confidence interval; egfr, estimated glomerular filtration rate; IV, intravenous. Adverse outcome was defined by chronic kidney disease stage 4 5 (egfro30 ml/min per 1.73 m 2 ) or patient death. a Plasma exchanges and induction immunosuppression (prednisolone with/without cyclophosphamide or rituximab). b Assessed only in patients with favorable renal outcome at 3 months. Kidney International (2014) 85,

6 clinical investigation A Sinha et al.: Anti-factor H autoantibody-associated HUS Survival free of adverse outcome At risk Combination therapy Follow up (months) Not received such therapy Figure 2 Probability of renal survival in patients with anticomplement factor H antibody-associated hemolytic uremic syndrome. Patients who received combined therapy with plasma exchanges and induction immunosuppression showed 83.0% survival at 6 months, 75.6% at 12 months, and 71.2% at last follow-up (interrupted line). Corresponding renal survival in patients not receiving combined therapy (continuous line) was 46.1, 41.5, and 33.2% (log rank Po0.0001). At risk Survival free of relapses Follow up (months) Maintenance immunosuppression Not received such therapy Figure 3 Probability of disease relapses with respect to maintenance immunosuppression. Relapse-free survival was 95.3% at 6 months, 92.3% at 12 months, and 87.2% during follow-up in patients receiving maintenance therapy (interrupted line), compared with 76.8, 69.1, and 46.1% in patients not receiving such therapy (continuous line) (log rank P ¼ 0.010). ratio (HR) 5.40, 95% CI ; P ¼ 0.020), low C3 (HR 5.91; 95% CI ; P ¼ 0.020), and time to plasma exchange X17 days (HR 10.35, 95% CI ; P ¼ 0.002). Combined therapy was independently protective (HR 0.11; 95% CI ; P ¼ 0.009) and was confirmed on survival analysis (Figure 2). It was determined that 2.6 patients needed to receive combined therapy to prevent one adverse outcome. Relapse-free survival was better among patients receiving maintenance immunosuppression (Figure 3); 4.5 patients needed treatment in order to prevent one relapse. DISCUSSION This multicentric report from India describes the clinical features, therapies, and outcome in a large group of children with anti-cfh antibody-associated HUS. These autoantibodies have been described in a frequency of 5 25% in patient cohorts of HUS from Europe, with higher prevalence reported in children. 7 12,16 In comparison, 56.1% patients in the present series had circulating anti-cfh antibodies. Although there is possibility of an ascertainment bias, the finding that a similar proportion of 109 consecutive patients who presented to one center showed antibodies suggests an increased prevalence of the condition among Indian children. In conformity with others, we found that 88.2% patients with anti-cfh antibodies had homozygous deletions in CFHR1 gene; the majority had associated homozygous deletion of CFHR ,14,19 While generation of antibodies appears specifically related to CFHR1 deficiency, patients may rarely show other abnormalities, including risk allotypes of CFHR1 10 or rearrangements in CFHR1/CFHR4. 10,11 The frequency of homozygous CFHR1 deletion in controls was 9.5%, similar to the 2 9% reported in other populations. 10,11,19,20 As the population frequency of CFHR1 deletion in healthy controls was similar to that reported elsewhere, the high prevalence of antibody-associated HUS in Indian children remains unexplained. Recent reports suggest that a small proportion of patients with atypical HUS may have mutations in more than one complement genes that might influence the clinical phenotype. 21 Data from four studies shows that 6 of 54 patients with anti-cfh antibodies and CFHR1 deletion had an abnormality in the CFH, MCP, CFI, and C3 genes ,17 Their functional significance was not established and most of these abnormalities are now classified as polymorphisms. In another report, none of the 27 patients with combined-complement mutations showed anti-cfh antibodies or homozygous CFHR1-3 deletion. 21 While anti- CFH antibodies were not present in healthy controls in this and previous reports, 11,12 4 of 21 asymptomatic siblings with homozygous deletion of CFHR1 showed high antibody titers. Further studies are required to define the role of additional genes and/or environmental influences in the pathogenesis of autoantibody-associated HUS. The age and seasonal prevalence, and history of diarrhea or fever in one half of the patients suggest that antibodymediated HUS may be triggered by infections. 5,16,17,19 We recently reported the association of antibody-negative HUS and P. vivax infection in seven patients. 22 Patients with shigatoxin-associated HUS show activation of the alternative complement pathway, as evidenced by elevated levels of factor Bb and soluble C5b-9, 23 and recovery of illness following therapy with eculizumab. 24 However, only one case each of shigatoxin and norovirus-related HUS has been reported in association with anti-fh antibodies. 17 Most patients (87.7%) with antibody-associated HUS in the present series were more than 5 years old, confirming 1156 Kidney International (2014) 85,

7 A Sinha et al.: Anti-factor H autoantibody-associated HUS clinical investigation that the condition affects older children and adolescents. 7,11,12,14,16,17 In contrast, post-dysenteric HUS, in south Asia, chiefly affects preschool children. 25,26 Patients with antibody-associated HUS had severe hemolysis, thrombocytopenia, prolonged anuria, and significant hypertension (Table 1). Hepatic or neurological features are reported in 17 66% cases, as was also seen in our patients. 12,15 17,27 Titers of anti-cfh antibodies in the present study were comparable to those reported previously and were not correlated with levels of CFH. 12,17 More than one-half of the patients showed low C3, confirming activation of the alternative pathway. 5,7,12,17 Titers of anti-cfh antibodies were fourfold higher in patients with hypocomplementemia, suggesting an inverse relationship. Furthermore, this study confirms that high antibody levels are associated with adverse outcomes 17 with highest titers in those dying during the acute illness. Antibody titers exceeding 8000 AU/ml were an independent risk factor for adverse outcome. Although therapies were heterogeneous, 62 patients were managed at a single center with uniform protocol. Based on anecdotal reports, experts advise that patients with atypical HUS should receive prompt and prolonged plasma exchanges, which result in decline in antibody titers and provision of complement factors. 3,12,17,28 Our findings support these recommendations and show that delayed initiation of plasmapheresis, 2 3 weeks beyond onset, predicted adverse outcomes. While their relative impact on antibody levels needs to be clarified, a significant reduction in titers and satisfactory outcome was observed in patients treated with plasmapheresis and immunosuppressive agents. The effect of acute management was confirmed on risk estimates, which showed that one adverse outcome was prevented for approximately three patients receiving combined treatment. Given that many patients were treated rather late after onset, the outcomes are likely to improve with prompt diagnosis and rapid initiation of treatment. Boyer et al., reported that therapy with high-dose cyclophosphamide and plasma exchanges, but not the latter alone resulted in sustained reduction in anti-cfh antibodies. 13 While 49 patients in the present report received IV cyclophosphamide, 18 received rituximab based on experience in patients with anti-adamts13 antibodies. 29 However, the benefits of these therapies were difficult to distinguish from those of plasma exchange and corticosteroids. Prospective studies are necessary to compare the efficacy and safety of rituximab with those of cyclophosphamide, and clarify the indications for use of IV immunoglobulin. Outcomes in 122 patients, in the present study, showed improved renal function in almost one-half of the children. These findings are comparable to reports showing renal recovery in % of patients. 5,11,12,17 The risk of relapse in the present patients was 16.7%, compared to the % reported previously. 3,7,12 The late occurrence and lower frequency of relapses may relate to institution of maintenance therapy, which reduced the risk of relapses by 91%, such that one relapse was prevented for 4.5 patients receiving immunosuppression. Similar to other reports, most relapses in our patients were associated with increase in antibody titers. 12,17 While the utility of serial monitoring is unclear, increased vigilance is advisable for patients with high titers, particularly following exogenous triggers. Maintenance immunosuppression appears a more feasible strategy for preventing relapses, than prolonged plasma exchanges. 30,31 An important limitation was the inability to characterize the proximate trigger for the disease. Patients usually presented late, after having received one or more broad spectrum antibiotics and facilities for testing for fecal verotoxin or anti-lipopolysaccharide antibodies were limited. Detailed biochemical and genetic exploration for complement components, especially in patients with antibodynegative HUS, was lacking. 3 Finally, variable therapeutic practices across centers limit conclusions regarding the precise effect of therapies and outcome. Despite these concerns, this clinical study highlights the efficacy of therapy with plasma exchange and immunosuppressive medications in inducing favorable outcomes. This large multicentric cohort provides important information on clinical features and outcomes in patients with anti-cfh antibody-associated HUS. Genetic evaluation in patients, their families, and healthy controls enabled assessment of CFHR1 and CFHR3 allele frequencies. Although eculizumab might be effective during the acute illness, the present report underscores the long-term benefits of early implementation of antibody-lowering strategies, chiefly plasma exchange and immunosuppressive agents. Prospective studies are necessary to examine the relative role of therapies in enabling renal recovery and preventing relapses. MATERIALS AND METHODS We reviewed the case records of patients with HUS, below 18 years old, that were screened for anti-cfh antibodies from March 2007 to February While consecutive patients of HUS at the AIIMS were included, specimens from other centers were chiefly from patients with suspected atypical HUS or in the absence of diarrheal prodrome. The diagnosis of HUS was based on the presence of microangiopathic hemolytic anemia (fragmented red cells, reticulocytosis, lactate dehydrogenase 4400 U/l), thrombocytopenia (platelets o150, 000/mm 3 ), and acute kidney injury. 1,32 A history of diarrhea in the preceding 2 weeks constituted a diarrheal prodrome. Hematological remission was the absence of microangiopathic anemia and thrombocytopenia for 2 weeks. 32 Disease relapse was defined as a new episode of illness, presenting more than 4 weeks after remission in a patient with previous HUS. Patients with septicemia, disseminated intravascular coagulation, or thrombotic microangiopathy secondary to medications and HIV infection were excluded. Anti-CFH antibodies assay Following approval of the Ethics Committee, plasma samples, obtained before initiation of plasma exchange or infusion of blood products, were screened for anti-cfh IgG antibodies using an enzyme-linked immunosorbent assay. 7,17 When possible, samples were also collected during disease remission and relapse. Briefly, Kidney International (2014) 85,

8 clinical investigation A Sinha et al.: Anti-factor H autoantibody-associated HUS 96-well solid plates (Nunc-Immuno Micro Well, Sigma-Aldrich, MO), coated with purified factor H (Sigma-Aldrich) in 0.1 M carbonate buffer, were incubated overnight at 4 1C. Plates were blocked with 1% bovine serum albumin for 2 h at 37 1C. Plasma samples (100 ml) were added in serial dilutions and incubated for 1 h at room temperature. Following color development using goat antihuman IgG horse radish peroxidase (Sigma-Aldrich) and tetramethylbenzidine, optical density was read at 450 nm. Antibody titer of plasma at 1:50 dilution was expressed as arbitrary units (AU)/ml and calculated using a calibration curve obtained with serial dilutions of reference plasma. Following screening for anti-cfh antibody in plasma from 90 healthy donors, the positive threshold was established at 150 AU/ml, corresponding to five standard deviations above the mean. Samples from 10 patients and positive controls were assayed in duplicate in three different experiments. Inter-assay and intra-assay variation were 6% and 8%, respectively. Results on 22 samples, tested simultaneously at New Delhi and Paris, showed a variation of o15% between the titers. The isotype of anti-fh antibodies was determined using mouse monoclonal anti-igg1, IgG2, IgG3, and IgG4 antibody at the second step of ELISA, as described. 17 After washing, plates were incubated with anti-mouse IgG labeled with horseradish peroxidase, followed by color development as described above. Other investigations Blood levels of complement C3, antinuclear antibody, and antineutrophil cytoplasmic antibody were determined. Where suspected, evaluation included testing for malaria, leptospirosis, and stool culture for Shigella dysenteriae. Factor H concentrations in plasma were measured by ELISA 7 ; levels in 50 healthy children ranged between ng/ml. Renal biopsy was performed at the discretion of the managing physician and was examined by light and immunofluorescence microscopy. Multiplex ligation-dependent probe amplification Rearrangements in the CFH-CFHR1-5 genomic region were analyzed by multiplex ligation-dependent probe amplification in 68 patients with anti-cfh antibodies, families of 15 patients and 84 controls, using P236-A3 ARMD mix 1 from MRC-Holland. 17 Briefly, 125 ng DNA was incubated with probes that hybridized targets in exons of CFH, CFHR3, CFHR1, CFHR2, and CFHR5. Amplified products were run on an ABI PRISM 3130 Genetic Analyzer capillary electrophoresis system (Applied Biosystems, Foster City, CA). Peaks and areas for samples were determined by Genemapper v4.0 Software (Applied Biosystems) and dosage quotients calculated using Coffalyser NET (MRC-Holland); peak heights/areas between 50 70% were considered heterozygous deletions. Western blot for CFHR1 Proteins extracted from sera on 10% Nupage gel (Novex, Life Technologies, NY) were transferred to nitrocellulose membrane using iblot western blotting system (Invitrogen, Carlsbad, CA). Following blocking with 1% bovine albumin, membranes were incubated with polyclonal goat sera against CFH (Quidel, San Diego, CA) and rabbit anti-goat HRP-conjugated IgG (Santa Cruz Biotechnology, Dallas, TX); blots were developed using enhanced chemiluminescence system (PerkinElmer, Waltham, MA). Genetics Genomic DNA was extracted from the peripheral blood of 10 patients with HUS without auto-antibodies, and amplified by polymerase chain reaction using oligonucleotides flanking each exon of CFH, CFI, and MCP genes. Details of the procedure have been described previously. 4 Therapy Supportive care for acute kidney injury and peritoneal or hemodialysis was provided. Specific management varied across centers, and included plasma exchanges and/or immunosuppressive therapies. Plasma exchange was initiated at disease onset or relapse, using membrane filtration. During each session, ml/kg of plasma was replaced with fresh frozen plasma. Exchanges were performed daily for 5 7 days or until hematological remission, followed by exchanges on alternate days, twice weekly and then weekly. In patients with difficult vascular access or limited facility for plasma exchange, fresh frozen plasma was infused at 10 ml/kg/ day for 5 10 days. Patients with high antibody titers or lack of hematological remission received IV immunoglobulin (2 g/kg over 2 days). Induction immunosuppression comprised prednisolone given at a dose of 1 mg/kg daily for 2 weeks and on alternate days for 4 weeks; the dose was tapered by 0.25 mg/kg every 2 weeks to mg/kg for 12 months. Following 7 10 plasma exchanges, some patients received IV cyclophosphamide (500 mg/m 2 every 3 4 weeks for 5 doses) or IV rituximab (375 mg/m 2 once a week for two doses). Subsequently, patients with estimated GFR 430 ml/min per 1.73 m 2 received maintenance therapy with tapering doses of prednisolone with or without azathioprine (1 2 mg/kg/day) or mycophenolate mofetil (750 mg/m 2 /day), beginning at 3 4 months and continued for 1 2 years after disease remission. Outcomes Outcomes, at 3 months and at last follow-up, were assessed in terms of hypertension, 33 proteinuria (urine dipstick 42 þ ), and estimated GFR. 34 Adverse outcome was defined as estimated GFR o30 ml/min per 1.73 m 2 at 3 months or subsequent follow-up, or patient death. Statistical analysis Data were analyzed using Stata version 12.0 (Stata Statistical Software: Release 12; College Station, TX). Continuous data were expressed as mean±standard deviation. Anti-CFH antibody titers were analyzed following log transformation and reported as mean±standard error. Tests for significance included chi-square or Fisher exact tests, and paired or unpaired t-tests; strength of linear relationships was measured using Pearson s correlation coefficient. Two-tailed Po0.05 was considered significant. Determinants of adverse outcome were estimated as odds or hazards ratios, by univariate and by multivariate analyses. Patients with adverse outcome at 3 months were excluded from analyses for determinants of relapse. Receiver operating characteristic curves were drawn with adverse outcome at 3 months as the dichotomous variable, and time to hematological remission, antibody titer, time to plasma exchange, and duration of dialysis as continuous variables. Survival estimates compared the time to adverse outcome and relapses in relation to therapy. The number of patients needed to be treated to prevent an adverse outcome or relapse was calculated. 35 DISCLOSURE All the authors declared no competing interests Kidney International (2014) 85,

9 A Sinha et al.: Anti-factor H autoantibody-associated HUS clinical investigation ACKNOWLEDGMENTS We thank the physicians who contributed patients involved in this study. This study was supported by funding from the Department of Biotechnology, Government of India (102/IFD/SAN/PR2624/ ). Support from the Indo-French Center for Proposal for Advanced Collaborative Research for enabling collaborative work between Georges Pompidou European Hospital, Paris and AIIMS, New Delhi is acknowledged. CONTRIBUTIONS K. Afzal, Jawahar Lal Nehru Medical College, Aligarh; I. Agarwal, Christian Medical College, Vellore; V. Aggarwal, B.L. Kapoor Hospital, New Delhi; J Ahmed, Grant Medical College, Mumbai; U. Ali, BJ Wadia Hospital for Children, Mumbai; K. Anand, Sir Ganga Ram Hospital, New Delhi; S. Kishore Babu, Manipal Hospital, Bangalore; A. Bagga, All India Institute of Medical Sciences, New Delhi; H. Doshi, Balabhai Nanavati Hospital, Mumbai; N. Gopalakrishnan, Madras Medical College, Chennai; S. Gulati, Fortis Hospitals, New Delhi; A. Gupta, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow; M. Kanitkar, Armed Forces Medical College, Pune; Manish Kumar, Chacha Nehru Bal Chikitsalaya, New Delhi; N. Parameswaran, Jawaharlal Institute of Postgraduate Medical Education and Research, Puducherry; M. Patil, Jawahar Lal Nehru Medical College, Belgaum; K. Ravishankar, Sowmya Children Hospital, Hyderabad; V.K. Sairam, Kanchi Kamacoti Child Trust Hospital,Chennai;S.Reddy,RainbowChildren shospital,hyderabad; M.A. Shah, Apollo Health City, Hyderabad; F. Shah, Child Kidney Care Center, Surat; R. Sinha, Institute of Child Health, Kolkata; N.H. Thaker, Ahmedabad; A. Udani, Lokmanya Tilak Municipal Medical College, Mumbai; A.S. Vasudev, Indraprastha Apollo Hospital, New Delhi; M. Vijayakumar, Mehta Children s Hospital, Chennai. SUPPLEMENTARY MATERIAL Figure S1. Frequency of patients with onset of anti-complement factor H antibody associated HUS according to age at presentation. Figure S2. Frequency of patients with onset of (a) anti-complement factor H antibody associated HUS and (b) autoantibody negative HUS in relation to month of the year. Figure S3. Ratio of IgG2, IgG3 and IgG4 isotypes to IgG1 in ten patients with antibody associated HUS. Supplementary material is linked to the online version of the paper at REFERENCES 1. Besbas N, Karpman D, Landau D et al. European Pediatric Research Group on HUS. A classification of hemolytic uremic syndrome and thrombotic thrombocytopenic purpura and related disorders. Kidney Int 2006; 70: Fremeaux-Bacchi V, Fakhouri F, Garnier A et al. Genetics and outcome of atypical hemolytic uremic syndrome: a nationwide French series comparing children and adults. Clin J Am Soc Nephrol 2013; 8: Loirat C, Frémeaux-Bacchi V. Atypical hemolytic uremic syndrome. Orphanet J Rare Diseases 2011; 6: Sellier-Leclerc AL, Fremeaux-Bacchi V, Dragon-Durey MA et al. Differential impact of complement mutations on clinical characteristics in atypical hemolytic uremic syndrome. J Am Soc Nephrol 2007; 18: Noris M, Caprioli J, Bresin E et al. Relative role of genetic complement abnormalities in sporadic and familial ahus and their impact on clinical phenotype. Clin J Am Soc Nephrol 2010; 5: Blanc C, Roumenina LT, Ashraf Y et al. Autoantibodies in the acute phase of the autoimmune form of atypical hemolytic uremic syndrome. J Immunol 2012; 189: Dragon-Durey MA, Loirat C, Cloarec S et al. Anti-factor H autoantibodies associated with atypical hemolytic uremic syndrome. J Am Soc Nephrol 2005; 16: Dragon-Durey MA, Blanc C, Marliot F et al. The high frequency of complement factor H related CFHR1 gene deletion is restricted to specific subgroups of patients with atypical hemolytic uremic syndrome. J Med Genet 2009; 46: Joszi M, Licht C, Strobel S et al. Factor H autoantibodies in atypical hemolytic uremic syndrome correlate with CFHR1/CFHR3 deficiency. Blood 2008; 111: Abarrategui-Garrido C, Martinez-Barricarte R, Lopez-Trascasa M et al. Characterization of complement factor H-related (CFHR) proteins in plasma reveals novel genetic variations of CFHR1 associated with atypical hemolytic uremic syndrome. Blood 2009; 114: Moore I, Strain L, Pappworth I et al. Association of factor H autoantibodies with deletions of CFHR1, CFHR3, CFHR4, and with mutations in CFH, CFI, CD46, and C3 in patients with atypical hemolytic uremic syndrome. Blood 2010; 115: Hofer J, Janecke AR, Zimmerhackl LB et al. German-Austrian HUS Study GroupComplement factor H-related protein 1 deficiency and factor H antibodies in pediatric patients with atypical hemolytic uremic syndrome. Clin J Am Soc Nephrol 2013; 8: Boyer O, Balzamo E, Charbit M et al. Pulse cyclophosphamide therapy and clinical remission in atypical hemolytic uremic syndrome with anti-complement factor H autoantibodies. Am J Kidney Dis 2010; 55: Lee BH, Kwak SH, Shin JI et al. Atypical hemolytic uremic syndrome associated with complement factor H autoantibodies and CFHR1/CFHR3 deficiency. Pediatr Res 2009; 66: Strobel S, Hoyer PF, Mache CJ et al. Functional analyses indicate a pathogenic role of factor H autoantibodies in atypical hemolytic uremic syndrome. Nephrol Dial Transplant 2010; 25: Geerdink LM, Westra D, van Wijk JAE et al. Atypical hemolytic uremic syndrome in children: complement mutations and clinical characteristics. Pediatr Nephrol 2012; 27: Dragon-Durey MA, Sethi SK, Bagga A et al. Clinical features of anti-factor H autoantibody associated hemolytic uremic syndrome. JAmSoc Nephrol 2010; 21: Malina M, Gulati A, Bagga A et al. Peripheral gangrene in children with atypical hemolytic uremic syndrome. Pediatrics 2013; 131: e331 e Zipfel PF, M., Heinen ES, Józsi M et al. Deletion of complement factor H-related genes CFHR1 and CFHR3 is associated with atypical hemolytic uremic syndrome. PLoS Genet 2007; 3: e Hageman GS, Hancox LS, Taiber AJ et al. Extended haplotypes in the complement factor H (CFH) and CFH-related (CFHR) family of genes protect against age-related macular degeneration: characterization, ethnic distribution and evolutionary implications. Ann Med 2006; 38: Bresin E, Rurali E, Caprioli J et al. European Working Party on Complement Genetics in Renal Diseases. Combined complement gene mutations in atypical hemolytic uremic syndrome influence clinical phenotype. J Am Soc Nephrol 2013; 24: Sinha A, Singh G, Bhat AS et al. Thrombotic microangiopathy and acute kidney injury following vivax malaria. Clin Exp Nephrol 2013; 17: Thurman JM, Marians R, Emlen W et al. Alternative pathway of complement in children with diarrhea-associated hemolytic uremic syndrome. Clin J Am Soc Nephrol 2009; 4: Lapeyraque AL, Malina M, Fremeaux-Bacchi V et al. Eculizumab in severe Shiga-toxin-associated HUS. N Engl J Med 2011; 364: Srivastava RN, Moudgil A, Bagga A et al. Hemolytic uremic syndrome in children in northern India. Pediatr Nephrol 1991; 5: Raghupathy P, Date A, Shastry JC et al. Haemolytic-uraemic syndrome complicating shigella dystentery in south Indian children. Br Med J 1978; 1: Dragon-Durey MA, Blanc C, Garnier A et al. Anti-factor H autoantibody associated hemolytic uremic syndrome: review of literature of the autoimmune form of HUS. Semin Thromb Hemost 2010; 36: Lionet A, Provôt F, Glowacki F et al. A case of adult atypical haemolytic uraemic syndrome related to anti-factor H autoantibodies successfully treated by plasma exchange, corticosteroids and rituximab. Nephrol Dial Transplant Plus 2009; 2: Ling HT, Field JJ, Blinder MA. Sustained response with rituximab in patients with thrombotic thrombocytopenic purpura: a report of 13 cases and review of the literature. Am J Hematol 2009; 84: Kwon T, Dragon-Durey MA, Macher MA et al. Successful pre-transplant management of a patient with anti-factor H autoantibodies-associated haemolytic uraemic syndrome. Nephrol Dial Transplant 2008; 23: Kidney International (2014) 85,