Post-Transplant Complications Renal Function in Long-Term Survivors of Stem Cell Transplantation in Childhood

Bone Marrow Transplantation (2001) 27, 319–327  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt Post-transplant complications Renal function in long-term survivors of stem cell transplantation in childhood. A prospective trial

L Patzer1, F Ringelmann1, K Kentouche1, D Fuchs1, F Zintl1, M Brandis2, LB Zimmerhackl2 and J Misselwitz1

1Department of Paediatrics, Friedrich-Schiller-University, Jena; and 2Department of Paediatrics, Albert-Ludwigs-University, Freiburg i Br, Germany

Summary: Disturbances of renal function caused by chemotherapy or by supportive therapy in the treatment of malignancies are The aim of this prospective study was to assess glomeru- well-known and recognised.1,2 The pathogenesis of acute lar and tubular renal function before, and 1 and 2 years renal dysfunction is very often multifactorial and associated after hematological stem cell therapy (HSCT) in chil- with the nephrotoxic potential of cytostatic agents dren and adolescents. 137 consecutive patients undergo- (cisplatinum, ifosfamide), veno-occlusive disease, hepato- ing HSCT, for malignant diseases, were included in a renal syndrome, sepsis and the corresponding antibiotic prospective trial. Forty-four patients were followed for (aminogylcosides) and antifungal (amphotericin B) treat- up to 1 year after HSCT and 36 for up to 2 years, with- ment. In patients treated with HSCT, acute renal dysfunc- out relapse. Ninety healthy school children were used tion is a frequent complication3,4–10 and the effect of con- as a control group. The following parameters were ditioning therapy for HSCT (high-dose chemotherapy and investigated: inulin clearance (GFR), urinary excretion fractionated total body irradiation) on renal function in chil- 11 of albumin, ␣1-microglobulin (␣1-MG), calcium, ␤-N- dren has been published by our group earlier. Chronic or acetylglucosaminidase (␤-NAG) and Tamm–Horsfall long-term sequels of high-dose chemotherapy or HSCT protein (THP), tubular phosphate reabsorption have been less intensively investigated.12–16 BMT-nephro-

(TP/Clcr) and percent reabsorption of amino acids pathy (see Ref. 17 for review) has been found in up to 25% (TAA). Significantly lower GFR was found 1 and 2 of transplanted adults and may progress to end-stage renal years after HSCT but within the normal range in the disease with a poor outcome in terms of survival on dialy- period before HSCT. There was no correlation between sis. Some renal injury syndromes are probably related to GFR within the first month after HSCT and long-term CyA, especially a hemolytic–uremic-like syndrome as outcome of GFR. Tubular dysfunction was found in 14– firstly described by Shulman et al.18 Radiation, cyclosporin 45% of patients 1 and 2 years after HSCT depending A therapy, late effects of cytostatic treatment and graft- on the parameter investigated. Pathological values 1 and versus-host disease are potential risk factors especially for ␣ 2 years after HSCT were found for 1-MG excretion in renal tubular dysfunction and the development of ifosfam- 19 40% and 39%, respectively, for TP/Clcr in 44% and ide-induced nephrotoxicity is well characterized. It is also 45%, for ␤-NAG in 26% and 19%. Median TP/Clcr was well-known that renal Fanconi syndrome may occur months significantly lower 2 years after HSCT than before. or even years after the end of chemotherapy.20 Neverthe- TAA was mildly impaired in 7/14 patients before, in less, no prospective trial has been published investigating 5/29 one and in 9/29 2 years after HSCT, but median frequency and pattern of tubulopathy in children under- TAA was within normal range at all times. The median going HSCT. excretion of albumin, THP and calcium was within the The aim of this study was to assess glomerular and tubu- normal range at all investigations. No influence of ifosfa- lar renal function 1 and 2 years after successful HSCT in mide pre-treatment on the severity of tubulopathy was children in a prospective trial. found. The investigation of tubular renal function should be part of a long-term follow-up in children after HSCT. Bone Marrow Transplantation (2001) 27, 319–327. Patients and methods Keywords: tubulopathy; nephrotoxicity; phosphate loss; chemotherapy; sequela; total body irradiation Between 1992 and 1998, 137 patients suffering from malignant diseases were treated with a first bone marrow transplant or high-dose chemotherapy with autologous stem cell rescue in a single center. 128 patients were enrolled in a Correspondence: Dr L Patzer, Friedrich-Schiller-University, Department prospective trial to assess renal function before, during and of Paediatrics, Kochstrasse 2, D-07745 Jena, Germany after HSCT. Nine patients were not included because they Received 29 February 2000; accepted 13 October 2000 did not give informed consent or were below the age of 2 Renal function after HSCT in childhood L Patzer et al 320 years or because of likely loss to follow-up. Seventy chil- Patient 2 (CR): Autologous HSCT at 5.6 years of age dren died within 365 days after HSCT, 12 suffered from because of metastatic clear cell sarcoma of the left kidney, relapse and two children were not available for follow-up. pre-treatment according to SIOP-93–01-HR and unilateral Forty-four relapse-free children were investigated 1 year nephrectomy at 4.9 years of age. Cumulative ifosfamide 12 after HSCT and 36 2 years after HSCT. For the analysis, g/m2, carboplatin 1.8 g/m2, no cisplatinum. At referral for patients were divided into two groups: group A, all patients HSCT incomplete Fanconi syndrome was found. Con- with normal renal parameters prior to HSCT (n = 41); group ditioning regimen: VP16 1 × 60 mg/m2, melphalan 4 × 30 B, unilaterally nephrectomized patients (n = 3). Because of mg/m2, fTBI 8 × 1.5 Gy. the small number, group B will be shown only for descrip- tion only and will not be used in any statistical evaluations. Patient 3 (MH): Autologous HSCT at 22.3 years of age At the time of the investigation all children were well, in because of 6th relapse of metastatic nephroblastoma since complete remission and without any acute illnesses. initial diagnosis of nephroblastoma at 8 years of age. Left nephrectomy after first relapse at 10 years of age, pre-treatment, cumulative ifosfamide 43 g/m2, cisplatinum 300 Group A mg/m2, no carboplatin conditioning regimen: VP16 5 × 200 mg/m2, carboplatin 5 × 275 mg/m2, melphalan 1 × 180 One year after HSCT all were taking antibiotic prophylaxis 2 (penicillin or cotrimoxazol), seven patients were receiving mg/m . methotrexate, six CyA, one FK506, three prednisolone, four All three patients were well on investigation. Patient 1 azathioprine, four 6-thioguanine and four 6-mercaptopur- was receiving penicillin and thyroxin; patient 2 penicillin ine. Only two patients were receiving phosphate and one and phosphate 1 and 2 years after HSCT; and patient 3 was calcium supplementation, respectively. Two years after off all medications. HSCT all children were off CyA. Median age at HSCT was 13.6 (3.9–42) years. Ten patients were older than 16 years. Reference values Thirteen had acute lymphoblastic leukemia: six were trans- As a reference group for tubular parameters (excretion of planted in CR1, five in CR2, one in CR3 and one in incom- ␣ ␤ plete remission. Nine had acute non-lymphoblastic leuke- 1-MG, -NAG, calcium and THP) and for albumin mia: six in CR1, two in CR2 and one in first relapse. Four excretion, 90 healthy children, 15 males and 15 females in had chronic myeloid leukemia, four Hodgkin’s lymphoma, each of the three groups (3 to 6, 6 to 12 and 12 to 18 years; two non-Hodgkin’s lymphoma, two Ewing’s sarcoma, two Table 1) were investigated. PNET, one rhabdomyosarcoma, one myelodysplastic syn- Reference values of ClIn were measured in 12 healthy volunteers; pathological GFR was defined below 90 ml/min drome, one osteosarcoma, one severe aplastic anemia and 2 1.73 m . The reference values of TAA and TP/Clcr were one neuroblastoma. Twenty-three patients underwent treat- 21,22 ment before HSCT with protocols containing ifosfamide taken from the literature. (median 10 g/m2 IFO, range 2–86 g/m2, 2–10 g/m2 n = 12, 20–24 g/m2 n = 3, 30–48 g/m2 n = 4, 54–86 g/m2 n = 4), Conditioning regimens three patients had cisplatinum. The following conditioning regimens were used: Group A: Twenty-one children received autologous and 20 allog- = eneical transplants (six matched unrelated donor, 10 combination containing cytostatic medications only, n 27: (1) combination of busulfan (BU, 4 mg/kg for 4 days) and matched related donor, three haploidentical related donor, = one mismatched unrelated donor). Twenty-five patients other cytostatics n 18, (either cyclophosphamide, melphalan, etoposide, thiothepa or antithymocyte globulin); (2) were transplanted using bone marrow and 15 with stem = cells, one patient received both on the same occasion. Dis- others without fTBI, n 9; (cyclophosphamide, etoposide, tribution of kind of transplant in 33 relapse-free survivors melphalan, carboplatin, AraC, carmustin). 2 years after HSCT was similar compared to before HSCT (autologous n = 18, allogeneic n = 15, NS, – chi-square ␣ ␤ Table 1 Reference values for urinary excretion of 1-MG, -NAG test). THP and albumin in three age groups

␣ ␤ Age 1-MG -NAG THP Albumin Group B (years) (mg/mmol (U/mmol (mg/mmol (mg/mmol creat) creat) creat) creat) Group B consisted of three patients with following features: 3–6 0.2 0.2 6.2 0.9 Patient 1 (RH): Autologous HSCT at 3.9 years of age n = 30 0.1–0.9 0.1–0.3 2.6–10.8 0.5–2.0 because of pulmonary relapse of nephroblastoma, pre-treat- 6–12 0.4 0.2 4.0 0.7 ment according to SIOP-90 and unilateral nephrectomy at n = 30 0.2–1.1 0.1–0.5 1.5–14.7 0.4–2.4 2 12 months of age. Cumulative ifosfamide: 24 g/m , no cis- 12–18 0.2 0.2 3.6 0.4 platinum; carboplatin 3.6 g/m2: conditioning regimen: n = 30 0.1–0.5 0.1–0.3 1.2–6.4 0.2–1.5 × 2 × 2 VP16 5 200 mg/m , carboplatin 4 400 mg/m , melphal- all 0.3 0.1 4.3 0.6 ane 1 × 180 mg/m2, renal replacement therapy from day +6 n = 90 0.1–0.9 0.2–0.3 1.7–10.1 0.3–1.9 until day +27 because of septicemia and multiorgan failure; complete recovery. Values are median, 5th and 95th centile.

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 321 Table 2 Results of renal function investigation regarding to pre-treatment with ifosfamide

Pre-treatment Pre-treatment Mann–Whitney without IFO containing IFO U-Chi-square test

GFR (ml/min 1.73 m2) Before HSCT 127 (73–175) 131 (94–218) NS No of patients Ͻ90 1/10 0/20 NS 1 year 122 (68–185) 123 (82–157) NS 1/10 1/15 NS 2 years 111 (81–177) 103 (97–177) NS 1/7 1/6 NS

␤-NAG (U/mmol creat) Before HSCT 0.48 (0.16–1.7) 0.45 (0.24–1.49) NS No of patients Ͼ0.4 6/11 11/20 NS 1 year 0.36 (0.09–1.4) 0.27 (0.05–0.67) NS 4/12 4/19 NS 2 years 0.2 (0.08–1.13) 0.24 (0.06–0.44) NS 2/9 3/17 NS

␣ 1-MG (mg/mmol creat) Before HSCT 1.07 (0.23–9.9) 0.95 (0.02–6.75) NS No of patients Ͼ1.0 9/18 9/23 NS 1 year 0.8 (0.03–4.68) 0.66 (0.05–23.2) NS 7/17 9/23 NS 2 years 0.56 (0.04–7.25) 0.63 (0.03–17.12) NS 4/12 9/21 NS

TP/Clcr (mmol/l) Before HSCT 1.2 (0.51–1.75) 1.24 (0.63–1.73) NS No of patients Ͻ1.07 6/18 7/23 NS 1 year 1.16 (0.56–1.57) 1.05 (0.68–1.64) NS 5/16 12/23 NS 2 years 0.97 (0.72–1.34) 1.13 (0.53–1.44) NS 7/12 8/21 NS

Given are median, minimum and maximum as well as number of patients with pathological investigation.

Table 3a Correlation of impaired phosphatre absorption (TP/Clcr) and Table 3b Correlation of impaired ␣1-MG reabsorption and impaired impaired amino acid reabsorption (TAA) at all measurements amino acid reabsorption (TAA) at all measurements

␣ Ͼ TP/Clcr 1-MG 1 mg/mmol Ͻ0.84 mmol/l creat

No Yes No Yes

Impaired amino acid Impaired amino acid reabsorption No 42 8 50 reabsorption No 36 14 50 (TAA Ͻ − 2 s.d. in (TAA Ͻ − 2 s.d. in Ͼ4 amino acids) Yes 18 3 21 Ͼ4 amino acids) Yes 6 15 21 60 11 71 42 29 71

Number of patients is given. Number of patients is given. Chi-square test. Spearman correlation coefﬁcient 0.40, P Ͻ 0.01. Group B: combinations containing fractionated total body irradiation (6 × 2 Gy) n = 14: (1) combination of fTBI and Glomerular function: (1) Single shot injection inulin clear- cytostatics (etoposide, melphalan), n = 8; (2) others with ance done by the method of Gretz et al.23 (2) Albumin fTBI (cyclophosphamide, etoposide or ALG) n = 6. (nephelometric assay, ARRAY 360, Beckmann, Fullerton, CA, USA). Study design and laboratory investigations ␣ Tubular function: (1) excretion of 1-MG: commercially Patients were investigated 20 days before, and 1 and 2 years available solid phase ELISA (Elias, Freiburg, Germany). after HSCT. A spot urine and a fasting blood sample were (2) fractional tubular phosphate reabsorption: TP/Clcr = − × ␤ taken followed by inulin clearance. Urine samples for deter- (mmol/l) PPO4 (UPO4 Pcr)/Ucr. (3) -NAG excretion: ␣ mination of 1-MG, THP and albumin were frozen at commercially available colorimetric assay (Boehringer, −18°C until measurement. The following variables were Mannheim, Germany). (4) urinary calcium excretion: ion- investigated: sensitive electrode indirect (Beckman Synchron CX9). (5)

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 322 Tamm-Horsfall protein: commercially available solid phase P < 0.05 ELISA (Elias). (6) percent reabsorption of amino acids 200 (amino acidanalyzer LC300, Biotronic Eppendorf, Ham- = − × × burg, Germany) TAA (%) (1 (UAA Pcr)/(Ucr PAA)) ) × 100. 2

Serum, urine: The concentration of sodium and potassium was measured by ionsensitive electrodes. Calcium, phos- 100 phate, creatinine and urea were measured on a CX3/CX4 colorimetrically (Beckmann). GFR (ml/min 1.73 m Statistics − 0 All values are presented as median, minimum maximum n = 33 28 16 value. Statistical significance was judged by a P value of Before BMT 365 730 Days after HSCT Ͻ0.05. Statistical procedures were performed using SPSS for Windows version 9.0 (SPSS Inc., Chicago, IL, USA). Figure 1 Glomerular filtration rate (GFR) before, 1 and 2 years after The statistical difference between values of two inde- HSCT. Friedman test, horizontal line indicates lower end of normal range. Box plots represent group A. Group B: , patient 1; ᭜, patient 2; #, pendent groups was done with the Mann–Whitney U-Wil- patient 3. coxon rank sum W test. The Wilcoxon matched-pairs signed-ranks test was used to compare the results before and after HSCT. To compare three matched pair groups no significant differences in GFR 2 years after HSCT with (before, 1 and 2 years after HSCT) the Friedman test was regard to the initial disease, type of conditioning regimen used. Correlation analysis to assess bivariate relationships (containing fTBI or not) or kind of HSCT (data not shown). was achieved by using the Spearman-Rho correlation test. We did not find significant differences in GFR regarding All investigations were made strictly in accordance with the presence of chronic GVHD at the time of investigation. the guidelines of the local ethics committee. Group B, GFR in patients 1, 2 and 3 was 117, 58 and 87 ml/min 1.73 m2 before HSCT. One patient (No. 1) developed acute renal insufficiency and had to be treated Results with renal replacement therapy, patient 2 and 3 did not double their serum creatinine. GFRs 1 year after HSCT were All children were free of urinary tract infection by dip stick 59, 54 and 127 ml/min 1.73 m2 and 2 years after 83, 69 and testing at the time of investigation. Group A: median height 150 ml/min 1.73 m2 (GFR was calculated by the Schwartz standard deviation scores in children transplanted before the formula at 1 and 2 years after HSCT in patient 3 because age of 12 years (n = 18) were +0.39 before, +0.16 1 year he could not tolerate repeated inulin clearance). and −0.19 2 years after HSCT (P Ͻ 0.05, Friedman test). Seven patients (17%), all transplanted allogeneically and Albumin excretion: Group A, median albumin excretion receiving CyA, suffered from acute renal insufficiency was found to be within the normal range and did not differ (doubling of serum creatinine) within 30 days after HSCT. before, 1 and 2 years after HSCT: 1.13, 0.37–15.36 vs 1.02, Two of them had to be treated with renal replacement ther- 0.38–15.4 vs 0.96, 0.29–12.39 mg/mmol creat. apy because of multiorgan failure. Group B, albumin excretion was normal before, 1 and 2 In Group B, one patient (No. 2) was transplanted with years after HSCT in patient 1 (1.01, 1.32 and 0.65 pre-existing renal Fanconi syndrome after 12 g/m2 ifosfam- mg/mmol creat) and in patient 3 (0.6, 0.74, 0.56 mg/mmol ide; tubulopathy persisted but is improving significantly. creat). Patient 2 had markedly elevated, but improving albumin excretion (16.3, 23.7 and 5.54 mg/mmol creat). Glomerular function Tubular function Glomerular filtration rate: Group A, GFR was significantly different before and 1 year after HSCT (130, 73– Phosphate reabsorption: Group A, TP/Clcr was signifi- 217 vs 123, 68–185 ml/min 1.73 m2, P Ͻ 0.05). It was cantly lower 2 years after HSCT than before HSCT (Figure within the normal range but significantly lower 2 years after 2): 1.08, 0.53–1.44 mmol/l vs 1.21, 0.51–1.75 mmol/l, P HSCT compared to before (105, 81–177 ml/min 1.73 m2, Ͻ 0.005. The difference between before and 1 year after P Ͻ 0.01). GFR was lower than 90 ml/min 1.73 m2 in only HSCT was significant at a level of P Ͻ 0.05: 1.21, 0.51– one before, in two 1 year and in two patients 2 years after 1.75 vs 1.11, 0.56–1.64 mmol/l. HSCT (Figure 1). It was below the normal range (Ͻ1.07 mmol/l) in 13 GFR after HSCT did not differ in seven patients who had patients before, 17/39 patients 1 year and in 15/33 2 years Ͻ acute renal failure within 30 days after HSCT compared to after HSCT. TP/Clcr lower than 0.84 mmol/l, ie ( 3 s.d. patients who did not double their serum creatinine: 109 of the mean20) was found in four patients before, in 8/39 ml/min 1.73 m2, 68–147 vs 124 ml/min 1.73 m2, 82–185, 1 year and in 5/32 2 years after HSCT. There was no sig- 1 year after HSCT, 105 ml/min 1.73 m2, 81–127 vs 107 nificant difference in phosphate reabsorption with respect ml/min 1.73 m2, 87–177, 2 years after HSCT. There were to earlier ifosfamide therapy (Table 2), kind of HSCT

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 323 2.0 >15 P < 0.05 10 1.8 NS 9 1.6 8 1.4 7 cr 1.2 6

TP/CL 1.0 5 (mmol/l) 4 0.8 3

0.6 -MG (mg/mmol creat) 1

a 2 0.4 1 n =413933

Before BMT 365 730 Days after HSCT 0 n = 41 40 33

Figure 2 Tubular phosphate reabsorption (TP/Clcr) before, 1 and 2 years Before HSCT 365 730 Days after HSCT after HSCT. Friedman test, horizontal line indicates lower end of normal − ᭜ ␣ range ( 2 s.d.). Box plots represents group A. Group B: , patient 1; , Figure 3 Tubular reabsorption of alpha-1-microglobulin ( 1-MG) patient 2; #, patient 3. before, 1 and 2 years after HSCT. Friedman test, horizontal line indicates upper end of normal range (2 s.d.). Box plots represent group A, ᭺, extreme values. Group B: , patient 1; ᭜, patient 2; #, patient 3. (allogeneic vs autologous), use of radiation in the conditioning regimen, occurrence of acute renal insufficiency during the acute phase of HSCT and presence of chronic GVHD P < 0.001 or CyA therapy 1 year after HSCT. There was a positive 1.0 correlation between TP/Clcr at 1 and 2 years after HSCT (Spearman-Rho coefficient = 0.41, P Ͻ 0.05). However, it was not possible to predict phosphate reabsorption 2 years after HSCT using TP/Clcr at day 365. Impaired amino acid reabsorption 1 year after HSCT was not associated with 0.5 impaired TP/Clcr 2 years after HSCT and did not precede deterioration of tubulopathy (data not shown). Group B, TP/Clcr was normal on all investigations in -NAG (mg/mmol creat) patient 1 and 3. Patient 2, with pre-existing renal Fanconi b syndrome showed, severely impaired phosphate reabsorption up to 1 year after HSCT which improved to normal 0.0 n =312630 TP/Clcr 2 years after (0.44, 0.52, 1.08 mmol/l). Before HSCT365 730 Days after HSCT ␣ ␣ 1-microglobulin excretion: Group A, excretion of 1-MG Ͼ Figure 4 Urinary excretion of ␤-N-acetylglucosaminidase (␤-NAG) was elevated ( 1.0 mg/mmol creat) in 18 patients before, before, 1 and 2 years after HSCT. Friedman test, horizontal line indicates in 16/40 1 year and in 13/33 patients 2 years after HSCT upper end of normal range (2 s.d.). ␣ and median excretion of 1-MG did not differ significantly: 0.98, 0.02–9.9 mg/mmol creat vs 0.66, 0.03–23.2 vs 0.63, ␣ 0.03–17.12 (Figure 3). 1-MG excretion was not influenced stable until 2 years after HSCT: 0.45, 0.16–1.7 vs 0.27, by earlier ifosfamide therapy (Table 2), kind of HSCT 0.05–1.4 (P Ͻ 0.005) vs 0.22, 0.06–1.13 U/mmol creat (allogeneic vs autologous), use of radiation in the condition- (NS). ␤ ␣ ing regimen, occurrence of acute renal insufficiency during A highly significant correlation between -NAG and 1- the acute phase of HSCT and presence of chronic GVHD MG excretion was found (Spearman coefficient 0.461, P Ͻ or CyA therapy 1 year after HSCT 0.001) using all measurements. ␣ ␤ No correlation between TP/Clcr and 1-MG excretion Group B, -NAG was normal on all investigations in was found. patient 1, and severely elevated on all measurements in ␣ Group B, 1-MG-excretion was normal before HSCT patient 23 (4.31, 3.78, 1.86). It was not measured in before, 1 and 2 years after HSCT in patient 1 and 3; it was patient 3. severely impaired on all investigations in patient 2 (43.5, ␣ 54.5, 10.9 mmol/l.). In this patient, improvement of 1- Tamm–Horsfall protein excretion: median THP-excretion MG-excretion correlated with improvement in TP/Clcr and was within normal range and not significantly different with albuminuria. before and after HSCT: 5.2, 0.8–15.2 vs 3.9, 1.5–11.9 vs 4.5, 1.2–9.2 mg/mmol creat. It was normal in the patients ␤-NAG excretion: Group A, urinary excretion of ␤-NAG of group B on all measurements. was investigated in 31 patients. Median excretion and inter- quartile range are depicted in Figure 4. It was significantly Calcium excretion: Group A, median calcium creatinine lower 1 year after HSCT compared to before and remained ratio on spot urine testing was within normal range

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 324 (Ca/creat Ͻ0.7 mmol/mmol) and not significantly different a given patient. Some studies have been published dealing before and after HSCT: 0.3, 0.02–2.7 mmol/mmol creat vs with renal function after BMT, but for the first time we 0.21, 0.03–1.05 vs 0.22, 0.04–1.02 mmol/mmol creat. Eight combined a prospective study design, the use of inulin patients had hypercalciuria before, 3/40 1 year after and clearance as measurement for GFR and investigation of a 1/33 2 years after HSCT on the basis of spot urine testing broad spectrum of tubular markers. Blood pressure, serum Ͼ (Ca/creat 0.7 mmol/mmol creat). HCO3 and response of urine osmolality to DDAVP were Group B, calcium/creatinine was normal in all patients not tested within this study. on all investigations. The three unilaterally nephrectomized children were not included in the statistical analysis, but their individual Reabsorption of amino acids: Group A, reabsorption of courses are shown. amino acids was mildly impaired (reabsorption Ͻ2 s.d. of We found the median GFRs to be within the normal more than four amino acids) in 7/14 patients before, 5/29 range in all investigations but it was significantly higher 1 year and in 9/29 2 years after HSCT. We did not find a before HSCT compared to 1 and 2 years after. This indi- correlation between reabsorption of amino acids and phos- cates some degree of hyperfiltration prior to HSCT. Similar phate (Table 3a). However, there was a significant corre- findings have been reported by Leblond et al,15 who found lation between frequency of impaired reabsorption of significantly lower GFRs after HSCT compared to a group ␣ amino acids and 1-MG (Table 3b). of adult patients before HSCT, in a non-prospective study. The median TAA 1 and 2 years after HSCT was within In contrast to our study, they found significantly lower the normal range (Figure 5). Two years after HSCT there GFRs in patients who received TBI. Lonnerholm et al24 was no significant difference in TAA of each amino acid described normal GFR (Cr-EDTA clearance) 2 years after between the group of patients with IFO pre-treatment and autologous HSCT in 15 children and found single fraction the group without treatment (data not shown). TBI to be a risk factor for deterioration. In contrast, Berg Group B, TAA was severely impaired on all investi- and Bolme14 described decreased GFR (GFR estimation by gations in patient 2 (number of amino acids reabsorbed Ͻ2 the formula of Morris) before and at all yearly examinations s.d.: 9, 10, 4). TAA was normal 2 years after HSCT in after HSCT in children with ALL and AML. Kumar and patients 1 and 3. co-workers12 did not find significant long-term renal dysfunction in a retrospective study of 17 survivors who underwent BMT for acute lymphoblastic leukemia 2–10 Discussion years after transplantation, but they report hypertension, glucosuria and hematuria in two of them. Similar to our The aim of this study was to investigate frequency and pat- study, they found hyperfiltration in 10 patients, but the GFR tern of renal tubulopathy 1 and 2 years following HSCT in was estimated by the Schwartz formula. Tubular parameters children. Only children free of disease relapse were were not investigated in this study. included in the analysis, resulting in a small group of In our study, there was no correlation between GFR patients treated in a single center. It is obvious that this within the first month after HSCT and long-term outcome kind of study is hampered by the non-homogeneous study of GFR in long-term survivors. This is in contrast to the group and the large number of independent influencing retrospective study from Kist et al10 who found chronic variables. The results presented here show important fea- renal insufficiency (GFR Ͻ85 ml/min 1.73 m2 using tures concerning renal function during the course of high- Schwartz formula) in 25/90 children 1 year after BMT and dose chemotherapy, obtained prospectively but they are not a correlation of serum creatinine within 30 days after BMT, appropriate for defining an individual renal risk index for and renal insufficiency 1 year after BMT. The authors report acute renal insufficieny in 34% of patients compared 100 to 17% in our study, and no patients who needed dialysis compared to two patients in our group A. Surprisingly, they 98 found chronic renal failure in 28% of their children 1 year after BMT (7% in our group A). This result is difficult to 96 interpret because no information about medication, especially CyA, on investigation 1 year after BMT was 94 given by Kist et al. Van Why and co-workers9 report renal insufficiency later than 60 days post BMT in 28% of 64 92

TAA (%) TAA pediatric patients. In this investigation amphotericin B, Mean of ref value CyA and TBI were found to be predictors for renal insuf- 90 Mean 2 years after BMT ficiency. In accordance with our study and in contrast to Mean 1 year after BMT Kist et al, renal insufficiency within 60 days after BMT was 88 -2 s.d. of ref value not predictive of late insufficiency in the study of van Why. The reason for the different conclusions about long-term 86 T S P G A V C M I L T P O L H A prognosis of GFR may be attributed to the method used . H E R L L A Y E S E Y H R Y I R As showed by Womer et al,25 creatinine clearance is an R R O Y A L S T O U R E N S S G unreliable measure of GFR in children but there are only Figure 5 Percent of amino acid reabsorption (TAA) 1 and 2 years after a few reports using inulin- or Cr-EDTA clearance assessing HSCT; references values from Rossi et al.21 GFR.14,15,24,26 Reporting from our own experience, there is

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 325 an extremely weak correlation between inulin clearance and excretion 2 years after HSCT. Decreased phosphate reab- GFR estimation from serum creatinine (Schwartz formula), sorption was found in 44% 1 year and in 45% 2 years after especially in a group of patients with malignant disease. It HSCT and median phosphate reabsorption worsened over is well-known that the Schwartz formula may overestimate time. GFR in patients with reduced muscle mass27 and no This indicates that the pattern of tubular damage changes methods estimating creatinine clearance from serum creati- over time. Early after HSCT, non-specific damage of proxi- nine provide reliable methods for estimating GFR in pedi- mal tubular cells is present as indicated by high excretion atric patients about to undergo BMT.28 of the brush border enzyme ␤-NAG and by insufficiency Risk factors for deteriorating GFR seem to be single frac- of the different reabsorption systems (low molecular weight tion TBI and BMT nephropathy as stated by Cohen et al proteins, phosphate, amino acids). At a later stage, the signs in adults.29,30 BMT nephropathy particularly seems to bear of non-specific cell damage become less prominent but dis- a significant risk for progression into end-stage renal distinct damage, especially of phosphate reabsorption, persists. ease31 and a poor survival on dialysis. We did not observe It is possible that the overall proximal cell injury caused development of BMT nephropathy in our patient group. by the conditioning therapy is being repaired within weeks The reason for the low incidence remains unclear, the or months after HSCT but only with defective healing of results of our study do not explain it. It may be possible the phosphate transporter (most likely the sodium phos- that BMT nephropathy will develop later than 2 years after phate cotransporter type 2). HSCT especially in a mainly pediatric population. We did This hypothesis resembles the situation of impaired not find any influence of chronic GVHD on GFR 1 and 2 TP/Clcr in patients with generalized tubulopathy after IFO years after HSCT and this is similar to the work of Leblond published by Rossi et al.20 Nonetheless, we did find higher et al15 who suggested that GVHD is not a major contributor cumulative probabilities for impaired phosphate reabsorp- to BMT nephropathy. However, in contrast, Miralbell et tion (44% vs 8% 1 year and 45% vs 14% 2 years after al13 found a significant correlation between presence of HSCT). Skinner and co-workers19 report exactly the same

GVHD and probability of renal dysfunction 18 months incidence of pathological TP/Clcr (44%) 1–47 (median 6) after BMT. months after ifosfamide treatment in 103 children. In our In the nephrectomized patients, the GFR was impaired study, not all patients with impaired TP/Clcr 1 year after before HSCT only in the patient with renal Fanconi syn- HSCT had impaired TP/Clcr 2 years after HSCT. There drome, but it remained stable. One patient required dialysis was, however, a significant correlation between phosphate because of extrarenal problems but kidney function showed reabsorption 1 and 2 years after HSCT, indicating that marked recovery. Kapinga et al32 report a retrospective impaired phosphate reabsorption 1 year after HSCT persists analysis of 23 children treated with nephrectomy and high- within the next year. We could not find any evidence that dose chemotherapy and abdominal radiotherapy. Ten tubulopathy in our patient group was attributable to ifosfa- patients developed impairment of creatinine clearance and mide pre-treatment. There was no correlation between ifos- five of seven survivors did not improve. famide treatment and degree of tubulopathy 1 or 2 years The bad prognosis of renal function in patients unilat- after HSCT. The reason could be the relatively low dose erally nephrectomized for Wilms’ tumour was also shown of ifosfamide in our group of patients who predominantly by Burk et al.33 had leukemia, as compared to patients receiving high-dose Thus, although reports from the literature show inconsist- ifosfamide therapy for osteosarcoma or rhabdomyosar- ent findings, the prognosis of glomerular renal function coma. Rossi et al34 also failed to show a linear correlation seems to be good in long-term survivors of HSCT, inde- between cumulative IFO dose and phosphate reabsorption. pendent of the initial disease, kind of HSCT and problems Low TP/Clcr values were found in this study only with IFO in the acute phase of HSCT. Our patients from group B doses above 24 g/m2, but in a number of patients it was show that unilateral nephrectomy and pre-existing renal present even after high IFO doses. We also found impaired

Fanconi syndrome are not necessarily risk factors for TP/Clcr in children who had never had IFO or cisplatinum. developing chronic renal insufficiency after HSCT. It might We did not investigate the sequelae of phosphate loss be possible that some degree of hyperfiltration compensates on growth and bone metabolism in this study. However, for the reduction in nephrons for some years after HSCT. disturbances in bone turnover in children after completion Because hyperfiltration by itself is clearly a risk factor for of chemotherapy and impaired final height after HSCT in progression into renal failure, we advocate thorough fol- childhood are reported35–37 and it is possible that renal low-up of kidney function in these children for a long time. phosphate loss may contribute to disturbed bone metab- In contrast to glomerular function, we were able to find olism and to the well-documented growth impairment tubular dysfunction in 14–45% of patients 1 and 2 years after HSCT. after HSCT depending on the parameter investigated. Amino acid reabsorption was impaired in 17% and in ␣ Although median 1-MG excretion decreases over time 31% patients 1 and 2 years after HSCT, respectively. In after HSCT, 40% of patients regarded as healthy showed contrast to Rossi et al20 and Caron et al38 we were not ␣ enhanced 1-MG excretion 1 year after HSCT and 39% able to prove that early impaired amino acid reabsorption after 2 years, respectively. Pathological ␤-NAG excretion progresses to generalized tubulopathy or renal Fanconi syn- was found in 26% of patients 1 year and in 19% 2 years drome. We did not observe the development of complete after HSCT. Median excretion of the brush border enzyme Fanconi syndrome within 2 years after HSCT in a patient ␤-NAG was already normal 1 year after HSCT and only who had not received IFO. 5/26 patients showed significantly elevated ␤-NAG From the normal excretion of THP, it seems that the dis-

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 326 tal tubule and thick ascending loop of Henle´ are not affec- References ted at all after HSCT. Our investigations of tubular function indicate significant 1 Donckerwolcke RA, Coppes MJ. Cytotoxic treatment and residual tubulopathy 1 and 2 years after HSCT. Reasons alterations in kidney function. Med Pediatr Oncol 1998; 30: for this tubulopathy are probably multifactorial. Because of 178–179. the small number of patients and the heterogeneity of the 2 Rossi R, Kleta R, Ehrich JH. Renal involvement in children with malignancies. Ped Nephrol 1999; 13: 153–162. group with regard to diagnosis, pre-treatment, conditioning 3 Fink JC, Cooper MA, Burkhart KM et al. Marked enzymuria regimens, type of HSCT, etc, it was not possible to define after bone marrow transplantation: a correlate of veno-occlus- independent causes for tubulopathy. Cumulative nephrotox- ive disease-induced ‘hepatorenal syndrome’. J Am Soc icity during conventional chemotherapy (ifosfamide, Nephrol 1995; 6: 1655–1660. cisplatinum), conditioning regimen prior to HSCT, septice- 4 Lane PH, Mauer SM, Blazar BR et al. Outcome of dialysis mia and its supportive therapy, GVHD and medication at for acute renal failure in pediatric bone marrow transplant the time of our investigation (CyA, cotrimoxazol) may all patients. Bone Marrow Transplant 1994; 13: 613–617. have contributed to the impairment of renal proximal tubu- 5 Zager RA. Acute renal failure in the setting of bone marrow lar function. It is possible, that the different nephrotoxic transplantation. Kidney Int 1994; 46: 1443–1458. events are cumulative, but by the methods used it is not 6 Zager RA. Acute renal failure syndromes after bone marrow transplantation. Adv Nephrol Necker Hosp 1997; 27: 263–280. possible to distinguish between tubulopathy caused by 7 Cooper BW, Creger RJ, Soegiarso W et al. Renal dysfunction nephrotoxic drugs from tubulopathy caused by GVHD. during high-dose cisplatin therapy and autologous hematopo- Our three unilaterally nephrectomized children indicate ietic stem cell transplantation: effect of aminoglycoside ther- that neither nephrectomy nor pre-existing renal Fanconi apy. Am J Med 1993; 94: 497–504. syndrome by itself are risk factors for HSCT. 8 Herget-Rosenthal S, Uppenkamp M, Beelen D et al. Renal complications of high-dose chemotherapy and peripheral blood stem cell transplantation. Nephron 2000; 84: 136–141. Conclusion 9 Van Why SK, FriedmanAL, Wei LJ, Hong R. Renal insufficiency after BMT in children. Bone Marrow Transplant The prognosis of glomerular renal function in healthy long- 1991; 7: 383–388. term survivors of HSCT is good. In contrast, about 40% of 10 Kist-van HJ, van-Zwet JM, Brand R et al. Bone marrow trans- patients including those without prior exposure to IFO plantation in children: consequences for renal function shortly after and 1 year post-BMT. Bone Marrow Transplant 1998; show significant signs of tubulopathy, especially phosphate 22: 559–564. loss. Whereas phosphate and bicarbonate loss are likely to 11 Patzer L, Hempel L, Ringelmann F et al. Renal function after result in clinically relevant problems, pathological conditioning therapy for bone marrow transplantation in child- ␤ ␣ excretion of -NAG, 1-MG and amino acids will not hood. Med Pediatr Oncol 1997; 28: 274–283. cause problems per se. Nevertheless, they are important 12 Kumar M, Kedar A, Neiberger RE. Kidney function in long- indicators of the integrity and function of the proximal term pediatric survivors of acute lymphoblastic leukemia fol- tubular cells. It is possible that glucosuria, hypercalciuria, lowing allogeneic bone marrow transplantation. Pediatr Hem- enzymuria or microproteinuria are the first signs of pro- atol Oncol 1996; 13: 375–379. gressive renal damage as known from distinct renal dis- 13 Miralbell R, Bieri S, Mermillod B et al. Renal toxicity after eases (Fanconi syndrome in cystinosis, Dent’s disease, allogeneic bone marrow transplantation: the combined effects of total-body irradiation and graft-versus-host disease. J Clin tyrosinemia). Unfortunately, no predictors have been found Oncol 1996; 14: 579–585. to differentiate children with ongoing renal damage who 14 Berg U, Bolme P. Renal function in children following bone need follow-up from patients with harmless pathological marrow transplantation. Transplant Proc 1989; 21: 3092– values. 3094. The significance of isolated renal phosphate loss on 15 Leblond V, Sutton L, Jacquiaud C et al. Evaluation of renal growth and bone metabolism needs to be determined by function in 60 long-term survivors of bone marrow transplan- further prospective, multicenter studies. Based on our tation. J Am Soc Nephrol 1995; 6: 1661–1665. investigations and reports from the literature, investigation 16 Tarbell NJ, Guinan EC, Niemeyer C et al. Late onset of renal of tubular renal function – especially phosphate reabsorp- dysfunction in survivors of bone marrow transplantation. Int tion – should be part of the long-term follow-up in children J Rad Oncol Biol Phys 1988; 15: 99–104. 17 Cohen EP. Radiation nephropathy after bone marrow trans- after HSCT especially in those with pre-existing renal prob- plantation. Kidney Int 2000; 58: 903–918. lems, such as unilateral nephrectomy or renal Fanconi 18 Shulman H, Striker G, Deeg HJ et al. Nephrotoxicity of cyclo- syndrome. sporin A after allogeneic marrow transplantation – glomerular thromboses and tubular injury. New Engl J Med 1981; 305: 1392–1395. Acknowledgements 19 Skinner R, Cotterill SJ, Stevens MCG. Risk factors for nephrotoxicity after ifosfamide treatment in children: a This work was supported by a grant from the ‘Deutsche Krebsh- UKCCSG Late Effects Group study. Br J Cancer 2000; 82: ilfe–Mildred Scheel Stiftung’ M107/91/Br 1; No. 70482. We are 1636–1645. grateful for their financial support. The authors also thank Mrs 20 Rossi R, Pleyer J, Schafers P et al. Development of ifosfam- Isolde Bernhardt for her excellent technical assistance and data ide-induced nephrotoxicity: prospective follow-up in 75 management, Dr Bellstedt for the amino acid analysis and the patients. Med Pediatr Oncol 1999; 32: 177–182. nursing staff of the HSCT unit for providing outstanding care of 21 Rossi R, Danzebrink S, Linnenburger K et al. Assessment of the patients. tubular reabsorption of sodium, glucose, phosphate and amino

Bone Marrow Transplantation Renal function after HSCT in childhood L Patzer et al 327 acids based on spot urine samples. Acta Paed 1994; 83: 31 Cohen EP, Piering WF, Kabler BC, Moulder JE. End-stage 1282–1286. renal disease (ESRD) after bone marrow transplantation: poor 22 Brodehl J, Gellissen K, Weber HP. Postnatal development of survival compared to other causes of ESRD. Nephron 1998; tubular phosphate reabsorption. Clin Nephrol 1982; 17: 79: 408–412. 163–171. 32 Kapinga E, Kardos G, van der Deure J, van Wijk JAE. 23 Gretz N, Ecker TK, Kuhnle HF et al. Practicability of the Impaired renal function in children with renal tumors treated inulin plasma single-shot clearance. Contrib Nephrol 1990; with nephrectomy, high dose chemotherapy and abdominal 81: 220–228. radiotherapy. Pediatr Nephrol 2000; 13: C40 (Abstr.). 24 Lonnerholm G, Carlson K, Bratteby LE et al. Renal function 33 Burk CD, Restaino I, Kaplan BS, Meadows AT. Ifosfamide- after autologous bone marrow transplantation. Bone Marrow induced renal tubular dysfunction and rickets in children with Transplant 1991; 8: 129–134. Wilms’ tumour. J Ped 1990; 117: 331–335. 25 Womer RB, Pritchard J, Barratt TM. Renal toxicity of cispla- 34 Rossi R, Danzebrink S, Hillebrand D et al. Ifosfamide-induced tin in children. J Ped 1985; 106: 659–663. subclinical nephrotoxicity and its potentiation by cisplatinum. 26 Brandt LJ, Broadbent V. Nephrotoxicity following carboplatin Med Pediatr Oncol 1994; 22: 27–32. use in children: is routine monitoring of renal function neces- 35 Arikoski P, Komulainen J, Riikonen P et al. Reduced bone sary? Med Pediatr Oncol 1993; 21: 31–35. density at completion of chemotherapy for a malignancy. Arch 27 Schwartz GH, Birn LP, Spitzer AL. The use of plasma creatin- Dis Child 1999; 80: 143–148. ine concentration for estimating glomerular ﬁltration rate in 36 Arikoski P, Kroger H, Riikonen P et al. Disturbance in bone infants, children and adolescents. Pediatr Clin North Am 1987; turnover in children with malignancy at completion of chemo- 34: 571–590. therapy. Med Pediatr Oncol 1999; 33: 455–461. 28 Jacobson P, West N, Hutchinson RJ. Predictive ability of crea- 37 Clement De BA, Oostdijk W, Van Weel Sipman MH et al. tinine clearance estimate models in pediatric bone marrow Final height and hormonal function after bone marrow trans- transplant patients. Bone Marrow Transplant 1997; 19: 481– plantation in children. J Ped 1996; 129: 544–550. 485. 38 Caron HN, Abeling N, van Gennip A et al. Hyperaminoacidu- 29 Lawton CA, Cohen EP, Barberderus SW et al. Late renal dys- ria identiﬁes patients at risk of developing renal tubular tox- function in adult survivors of BMT. Cancer 1991; 67: icity associated with ifosfamide and platinate containing regi- 2795–2800. mens. Med Pediatr Oncol 1992; 20: 42–47. 30 Cohen EP, Lawton CA, Moulder JE et al. Clinical course of late-onset BMT nephropathy. Nephron 1993; 64: 626–635.

Bone Marrow Transplantation