Long-term kidney complications in childhood leukemia survivors: a study from the Childhood and Adolescent Leukemia (LEA) project

Abstract

Acute leukemias represent the first cause of cancer in children. Their prognosis has improved significantly due to remarkable advances in therapeutic management, despite the risk of long-term consequences, especially for patients who underwent allogenic hematopoietic stem cell transplantation (aHSCT). Through the Leukemia in Children and Adolescents (LEA) long-term follow-up cohort (clinicaltrials gov. Identifier: NCT01756599), we conducted a French national multicenter prospective study on the occurrence and risk factors of chronic kidney disease (CKD), differentiating glomerular and tubular dysfunctions, corresponding to the NephroLEA project. Among the 1,676 patients included, the median age at evaluation was 15.8 (interquartile range [IQR], 11.3-20.5) years, with a median follow-up of 9.2 (IQR, 5.8-13.9) years. aHSCT was performed on 343 (20.6%) patients, half of whom have undergone the procedure after achieving second or greater remission. A higher percentage of children among transplanted patients had diastolic and systolic blood pressure above the 95th, with 13.7% versus 5.2% (P=3x10-3) and 15% versus 6.5% (P=9x10-2), respectively. A total of 187 patients (11.1%) had a mild CKD (i.e., eGFR between 75 and 90 mL/min/1.73 m2), while 3% (N=50) exhibited mild to severe CKD (eGFR <75 mL/min/1.73 m2). Notably, no patient reached kidney failure. Twenty-one patients (1.3%) had decreased glomerular filtration rate associated with tubular impairment. The principal risk factors for developing CKD were aHSCT and leukemia relapse. In conclusion, CKD represents a long-term risk for patients who relapsed and/or underwent aHSCT. These patients could benefit from nephroprotection advice to further improve their long-term outcomes, which is becoming a public health issue.

Introduction

Acute leukemias (AL) are the leading cause of cancer in children in developed countries.¹ Thanks to improved stratification of these hematologic diseases, based on enhanced clinical, biological and therapeutic knowledge, remarkable progress has been made in their therapeutic management. The overall survival rate has risen to approximately 85-90% for acute lymphoblastic leukemia (ALL)² and 60-70% for acute myeloid leukemia (AML).³ These favorable outcomes have been achieved despite the cumbersome and toxic nature of current chemotherapy regimens. Furthermore, although dramatic overall result enhancement over decades, allogenic hematopoietic stem cell transplantation (aHSCT) remains the treatment of choice for (very) high-risk hematological malignancies.^4,5 All these treatments, however, can induce various toxicities and complications in the short, medium or long term. With current high cure rates, long-term survivors of childhood leukemia now represent a growing population, and these long-term side-effects have become a major public health issue. One significant post-treatment complication is kidney dysfunction.⁶ While short-term acute renal failure has been extensively studied,^7-15 there is a lack of data from large, homogeneous cohorts of patients treated for childhood leukemia, particularly regarding the risk of developing chronic kidney disease (CKD) and/or tubular impairment.^16-21 Over time, kidney impairment may develop in patients without predisposing kidney failure, impacting long-term outcomes and subsequent quality of life. It may also represent an additional cardiovascular risk factor,^22-24 alongside the well-documented occurrence of metabolic syndrome in this population.^25,26 Therefore, these patients may benefit from education on general nephroprotective measures.

According to the international definitions, CKD is characterized by a reduced glomerular filtration rate (GFR) under 90 mL/min/1.73 m² in children and below 60 mL/min/1.73 m² in adults, and/or a urinary albumin-to-creatinine ratio (ACR) above 3 mg/mmol, and/or other morphological or electrolyte abnormalities due to tubular disorders.²⁷ Moreover, glomerular lesions and tubular impairment result from different toxic mechanisms and may lead to non-comparable long-term sequelae. Thus, we decided to systematically study the long-term renal function in a significant and homogeneous cohort of patients treated during childhood for acute lymphoblastic or myeloblastic leukemia, either with chemotherapy alone or intensified with aHSCT. This study was made possible by the French LEA (Leukemia in Children and Adolescents) patient cohort, established in 2004. The aim of this national multicenter prospective cohort is to evaluate the medium- and longterm outcomes of patients treated for acute leukemia in childhood, since January 1980. The participation rate in the LEA cohort, comprising patients treated for childhood leukemia across 18 French pediatric hematology centers, is around 80%.²⁷ Our main objective was to report the occurrence of CKD and its risk factors in the NephroLEA cohort, based on the presence of glomerular and/or tubular abnormalities.

Methods

Patients

From March 2019 to February 2022, we included all patients who underwent a long-term follow-up evaluation in the LEA program (clinicaltrials gov. Identifier: NCT01756599) (Online Supplementary Figure S1). The LEA project is a national French cohort initiated in 2004 to assess long-term health of patients treated during childhood or adolescence for AL after 1980. Follow-up begins 1 year after completion of chemotherapy for AML or aHSCT, and 2 years after the completion of chemotherapy for ALL. Clinical examinations are conducted at predefined intervals: every 2 years until the age of 20 and at least 10 years of complete remission (CR), and every 4 years thereafter. Late effects are detected through medical examination and adequate additional tests (biological and radiological) by physicians.²⁸

Clinical data related to AL were obtained from the LEA database, including patient demographics, AL subtype, CR status (including any relapse), details of aHSCT procedures (i.e., conditioning regimen, donor type, source of stem cells, and graft-versus-host disease (GVHD) grading, if any. All patients (or their parents or legal guardians) provided written informed consent. The study was approved by the Comité de Protection des Personnes Sud Méditerranée V Ethics Review Board (opinion n° 2012-A00984-39), and conducted in compliance with the General Data Protection Regulation.

Clinical definition and specimen collection

Blood pressure was measured three times on the right arm, with the patient lying down or seated calmly. The lowest systolic and diastolic results were recorded, as recommended.²⁹ Arterial hypertension was defined as a systolic or diastolic blood pressure (SBP or DBP, respectively) above the 95^th percentile for children under 15 years, or SBP >140 mmHg and/or DBP >90 mmHg for older children and adults.²⁹

Blood samples for creatinine (Cr), sodium (Na), potassium (K), bicarbonates (HCO₃), calcium (Ca), phosphate (Ph), magnesium (Mg), uric acid, glucose, as well as urine samples for Cr, Na, K, Ca, Ph, Mg, uric acid, glucose, |32-microglobulinuria, and albuminuria were collected either at time of medical examination or from local laboratories. This allowed evaluation of glomerular (estimated GFR [eGFR] and ACR) and tubular functions. Tubular function was assessed using formulas for electrolytes (E) reabsorption rates: (E_urine/E_plasma)/(Cr urine/Cr plasma) x100 (%), as well as tubular maximum Ph reabsorption per glomerular filtration rate (TmP/GFR), and urinary Ca/ Cr and Mg/Cr ratios.³⁰ For patients under 25 years old, eGFR was calculated using the Chronic Kidney Disease in Children U25 equation (eGFR-CKiDU25).³¹ For patients older than 25 years at evaluation, the CKD-EPI equation was employed.^32,33 Plasma Cr measurements were standardized according to National Institute of Standards and Technology recommendations.

According to Kidney Disease Improving Global Outcomes (KDIGO) guidelines,^27,34 CKD was defined as GFR <90 mL/ min/1.73 m². CKD staging followed classical GFR stratification: (i) stage 1 for normal renal function (GFR ≥90 mL/ min/1.73 m²), (ii) stage 2 for mild renal dysfunction (60 ≤ GFR <90 mL/min/1.73 m²), (iii) stage 3 for moderate renal dysfunctions, subdivided into 3a (45 ≤ GFR <60 mL/ min/1.73 m²) and 3b (30 ≤GFR <45mL/min/1.73 m²), and (iv) stage 4 for severe renal dysfunction (GFR <30 mL/min/1.73 m²). To refine the definition of CKD in younger patients (<40 years), we further subdivided the stage 2 into two subclasses: stage 2a (75≤ GFR <90 mL/min/1.73 m²) and stage 2b (60 ≤ GFR <75 mL/min/1.73 m²), as proposed by Delanaye et al.³⁵

Albuminuria was expressed as ACR in a spot urine sample. Microalbuminuria was defined as ACR between 3-30 mg/mmol, and macroalbuminuria as ACR >30 mg/mmol. Tubular renal disorders were characterized by metabolic acidosis (bicarbonates <18 mmol/L) and/or a renal phosphate wasting (hypophosphatemia for age 36 and a low age-adjusted TmP/GFR).³⁷

Statistical analysis

Statistical analysis was performed using SPSS 20.0 ® (SPSS Inc., Chicago, IL, USA). Qualitative data were expressed as counts and percentages. Quantitative data were presented as median (minimum-maximum). Continuous variables, such as age at leukemia diagnosis, age at last evaluation, and follow-up duration, were categorized at the median and interquartile range (IQR). Fisher’s exact tests were used to compare qualitative variables, while the Student’s t test or the Mann-Whitney U test was employed for quantitative variables. A P value <0.05 was considered statistically significant.

Results

Patients’ demographics

The NephroLEA cohort included a total of 1,772 patients, of whom 1,676 patients were eligible with available Cr-based eGFR data. Demographic and clinical features are detailed in Table 1. The median age at AL diagnosis was 4.5 years (IQR, 2.7-8.7), with a median follow-up of 9.2 years (IQR, 5.8-13.9). The cohort consisted of 52.8% (N=885) male patients. The most common diagnosis was ALL (84.5%, N=1,416), with 90% (N=1276) achieving first CR. aHSCT was performed in 343 patients (20.6%), half of whom underwent the procedure in CR2 or beyond. Among patients receiving HSCT, 219 (63.8%) were treated for ALL and 124 (36.2%) for AML. One-third of transplanted individuals had sibling-related donors. All aHSCT recipients underwent myeloablative conditioning, with total body irradiation (TBI) included in 58% (N=199) of the regimens for children. The median age at NephroLEA screening was 15.8 years (IQR, 11.3-20.5). Clinical abnormalities during physical examination were observed in 1.7% (N=28) of patients, including edema and/ or elevated blood pressure. A small proportion of patients received medications: 23 patients received antihypertensive treatment, ten were on anti-diabetic drugs, four were treated with lipid-lowering medications, and three received nephroprotection therapy (angiotensin-converting enzyme inhibitor or angiotensin receptor antagonists). Patients who underwent aHSCT exhibited a trend toward a higher median DBP compared with non-aHSCT (69 mmHg [IQR, 61-78] vs. 68 mmHg [IQR, 60-74]; P=0.1) (Table 2), although no significant difference in the median SBP was observed (P=0.6). These results were supported by an increased percentage of children with DBP above the 95^th in the aHSCT group (13.7% vs. 5.2%; P=0.003) (Table 2). Furthermore, aHSCT patients were older (17 vs. 15.6 years; P=0.002), lighter (48 vs. 52.5 kg; P=0.006), and had a trend toward shorter height (157 vs. 160 cm; P=0.07). These findings were corroborated by significantly smaller body surface area (BSA) in aHSCT patients compared with others (1.45 m² [IQR, 1.2-1.7] vs. 1.53 m² [IQR: 1.2-1.7]; P=0.001) (Table 2).

Table 1.Patient demographic data and clinical characteristics at baseline.

Evaluation of long-term glomerular function after acute leukemia treatment during childhood

In the entire cohort, 85.9% (N=1,439) of patients exhibited normal eGFR (CKD stage 1), with a mean eGFR of 111.8 mL/min/1.73 m² [IQR, 102.3-123.9]). Additionally, 11.1% of patients displayed a slight decrease in eGFR (CKD stage 2a), while 3% (N=50) were categorized as CKD stage 2b-4. Notably, only one patient was classified as CKD stage 4, with an eGFR of 26.3 mL/min/73 m² (Table 2). No cases of kidney failure were observed. Among patients classified in CKD stage 1, 10.5% (N=124) demonstrated abnormal albuminuria. Of these, 9.6% (N=113) had only microalbuminuria, while 0.9% (N=11) presented with macroalbuminuria.

Evaluation of long-term tubular function after childhood acute leukemia treatment

Plasma electrolyte levels were within normal ranges and comparable across all patients, except for chloride, uric acid and Ph levels. aHSCT patients had significantly lower TmP/GFR (1.03 [IQR, 0.89-1.23] mmol/L compared with others (1.15 [IQR, 0.97-1.33] mmol/lL; P<0.001) and higher uric acid levels (283 [IQR, 230.2-369.5] μmol/L vs. 258 [IQR, 217-304.2] μmol/L; P<0.001) (Table 3). A greater proportion of aHSCT patients exhibited abnormal uric acid values (17.6% vs. 6.1%; P<0.001) (Table 3).

Table 2.Long-term glomerular function after acute leukemia treatment during childhood.

Table 3.Long-term electrolytes, and uric acid rates, and tubular function after acute leukemia treatment during childhood.

Similarly, Ph levels were slightly but significantly lower in the aHSCT group, with a higher percentage of decreased TmP/GFR (13.9% vs. 7.6%; P=0.004) (Table 3). aHSCT patients also had significantly elevated β2-microglobulin/ Cr ratios (12.6 [IQR, 7.8-21.7] mg/mol vs. 10 [IQR, 6.9-16.4] mg/mol; P<0.001).

There was a trend toward more frequent “tubular signs” (low bicarbonatemia [<18 mmol/L] and/or a renal Ph wasting) among transplanted patients (24.5 % vs. 19.5%, P=0.08). Overall, 19.1% (N=320) of the cohort exhibited at least one tubular or glomerular abnormality, whereas 80.9% (N=1356) were classified as CKD stage 1 without tubular dysfunction. Among the remaining patients, 1.3% (N=21) were classified as CKD stage 2a-4 with tubular dysfunction, 4.9% (N=83) displayed normal GFR but exhibited tubular impairment, and 12.9% (N=216) were classified as CKD stage 2a-4 without tubular dysfunction (Tables 2, 3).

Risk factors for developing chronic kidney disease in all NephroLEA cohort

The analysis identified leukemia relapse (P<0.0001) and aHSCT (P<0.0001) as significant risk factors for long-term CKD (Table 4). Older age at evaluation (≥15.8 years; P<0.0001) was associated with higher incidence of CKD ≥2a. TBI and older patients were significant risk factors for tubular dysfunction (P=0.003 and P<0.0001, respectively) (Table 5).

Chronic kidney disease risk factors in subgroups

Among aHSCT patients, TBI was not a significant risk factor for CKD (P=0.4). However, relapse (P=0.03) and older age at evaluation remained significant (P=0.03 and P=0.04, for thresholds of 15.8 and 4.6 years, respectively) (Table 6). In patients who did not undergo aHSCT, age at evaluation (P=0.01) and sex (P<0.0001) were identified as significant risk factors for CKD ≥2a (Table 7). Notably, only 24 non-aHSCT patients experienced relapse, with three classified as CKD ≥2a.

Table 4.Chronic kidney disease risk factors for all the NephroLEA cohort according to risk factors.

Discussion

The increasing survival rates among children with AL have raised concerns regarding the long-term effects of the disease and its treatments, particularly for patients who have undergone aHSCT. Acute nephrotoxicity is a known side effect of several treatments used for malignant hemopathies, such as high-dose methotrexate, aminoglycoside or vancomycin as empirical antibiotherapy, and anti-calcineurin for GVHD prevention in grafted patients. However, evidence regarding the long-term effects of these treatments on kidney function in a homogeneous, large cohort of children treated during childhood for malignant hemopathies remains unknown. Consequently, the Cochrane meta-analysis recommends enrolling this category of patients in long-term follow-up programs to monitor blood pressure and renal function.²⁷,³⁴,³⁸ To our knowledge, the NephroLEA study is the largest and most comprehensive national prospective study to evaluate the prevalence of CKD among long-term survivors (>9 years) of pediatric ALL and AML treated with chemotherapies alone and/or with aHSCT. This study uniquely differentiates between glomerular and tubular dysfunctions and identifies risk factors for kidney function impairment in these specific long-term survivors.

The majority of patients in the NephroLEA cohort were treated for ALL with chemotherapies alone, including high-dose methotrexate and anthracyclines. Most patients also received empirical antibiotic therapy during the intensive treatment phase, such as aminoglycosides and/or vancomycin, to manage febrile neutropenia. These patients generally maintained normal long-term kidney function (CKD stage 1), with a median eGFR of 112 mL/min/1.73 m² (IQR, 102-124), without evidence of tubular dysfunction. Consequently, routine long-term nephrology follow-up may not be necessary for these patients. These findings align with those of Stotter et al.,¹⁹ who reported no significant association between high-dose methotrexate and a decline in eGFR over time (P=0.07).

For patients with high-risk disease or relapse, aHSCT remains the treatment of choice following complete remission.^39,40 Our team recently reported that this procedure is associated with long-term CKD in a retrospective tertiary monocentric study.⁶ These findings are confirmed by the NephroLEA cohort, where aHSCT emerged as the principal risk factor for CKD development CKD (P<0.0001), including glomerular dysfunction and a tendency for increased tubular dysfunction. Similarly, Pelletier et al. reported an adverse long-term prognosis for adult aHSCT patients with CKD, including an increased risk of mortality.¹⁷

TBI is mandatory in the conditioning regimen for aHSCT to increase treatment success for ALL patients.^41,42 It has been identified as a risk factor for kidney injury in various pediatric studies.^13,14,43-45 However, within the aHSCT population only, TBI was not a significant risk factor (P=0.4). This observation aligns with our previous publication⁶ and contrasts with findings from Abboud et al.,⁴⁴ who reported a higher cumulative incidence of CKD for adults than children in a mixed cohort of 148 patients, with TBI being an associated risk factor. Conversely, Frisk et al. found no difference in eGFR between individuals who underwent TBI and controls.⁴⁶ In the NephroLEA cohort, leukemia relapse (P=0.03), the use of “conventional” second-line chemotherapies (or more), calcineurin inhibitors for aGVHD prevention, and repeated infections (including viral infections) had a greater impact on long-term kidney function than the specific type of conditioning regimen used.

One of the major long-term complications of aHSCT reported by the LEA cohort is the occurrence of metabolic syndrome.^25,26,47,48 This complication, associated with long-term kidney dysfunction, could amplify the risk of cardiovascular disease in these long-term cancer survivors.^22-24 Florido et al.⁴⁹ reported a heightened risk of cardiovascular disease in adult cancer survivors (hazard ratio (HR)=2.7 [IQR, 2.04-3.59]; mean age: 54.5 years) compared to controls. This risk is exacerbated by hypertension, as observed with elevated DBP in aHSCT patients in NephroLEA cohort. In line with our results, Hsiao et al. reported that arterial hypertension was most common among patients who received aHSCT, with or without TBI.⁵⁰

In our cohort, no patient had progressed to kidney failure at a median age of 15 years. However, Calderon-Margalit et al. highlighted that a history of pediatric CKD, regardless its aetiology, significantly increases the risk of kidney failure in later life.⁵¹ Dieffenbach et al.,¹⁸ identified key risk factors for kidney failure, including nephrectomy, irradiation exceeding 15 Gy, and high doses of anthracyclines, or ifosfamide. Effective management of hypertension and diabetes, when present, can decelerate the progression of kidney dysfunction.

Within our homogeneous hematological cohort, 3% of patients had an eGFR <75 mL/min/1.73 m², similar to findings from the Dutch Childhood Cancer Survivor Study, where 3.7% of survivors had an eGFR <60 mL/min/1.73 m² at a median age of 32 years, 5 years after diagnosis.¹⁶ The cumulative impact of previous treatment and aging likely contribute to glomerular dysfunction over time, explaining why older patients in our cohort exhibited higher risk of developing CKD. With aging, the reduction in nephron numbers induces compensatory hypertrophy of the remaining nephrons, leading to glomerulosclerosis and a progressive decline in the eGFR. This process could partially explain the increased CKD risk observed in older patients within our cohort. Consequently, lifelong monitoring of glomerular function is strongly recommended for survivors exposed to high-risk factors, alongside adherence to standard nephroprotective measures.

Table 5.Tubular disease risk factors for all the NephroLEA cohort. Tubular dysfunction is defined by metabolic acidosis and/ or phosphate wasting.

Tubular dysfunction remains an under-researched area in the long-term follow-up of pediatric AL survivors, including aHSCT recipients. In our study, “tubular impairment” was defined as low bicarbonate levels (<18 mmol/L) and/ or renal Ph wasting. Similar to glomerular dysfunction, we observed a trend towards a higher prevalence of tubular disease among transplanted patients compared to non-transplanted patients (24.5% vs. 19.5%; P=0.08). While most plasma electrolyte levels were normal, abnormalities in uric acid, Ph and Mg were noted, particularly among aHSCT recipients. This contrasts with our earlier findings, where Ca reabsorption rated differed between subgroups.⁶ In the NephroLEA cohort, 83 patients (5.3%) exhibited tubular impairement with normal eGFR, while 197 patients (12.5%) were classified with CKD stage 2a-4 without tubular dysfunction.

Table 6.Chronic kidney disease risk factors for the allogenic hematopoietic stem cell transplantation patents of the NephroLEA cohort.

These findings suggest a shift in the pattern of tubular damage over time. Subclinical tubulopathy may have a more insidious impact than CKD, as it is harder to detect and interpret. It may also play a putative role in the complex process of growth and bone metabolism for these growing children and adolescents, especially post-aHSCT.

In conclusion the NephroLEA cohort highlights that the aHSCT-specific population is at a significantly higher risk of developing CKD, particularly among patients older than 15 years at time of evaluation. This major long-term side effect may exacerbate cardiovascular risk factors, compounded by the well-documented prevalence of metabolic syndrome in this population of long-term transplanted survivors. Kidney-related side effects, including glomerular dysfunction or more insidious tubular dysfunction, potentially impact growth and overall health. To mitigate these risks, targeted education on general nephroprotective measures is essential. Recommendations include maintaining a diet controlled in sodium and proteins, achieving a normal BMI, engaging in regular physical activity, avoiding non-steroidal anti-inflammatory medications, and abstaining from tobacco exposure. These measures should be supported by regular medical monitoring, including annual blood pressure screenings with prompt therapeutic management in case of elevated SBP and/or DBP. Additionally, urinary ACR should be assessed at least every 5 years, as per nephrology guidelines. Implementing these nephroprotective measures could improve global outcomes of post-leukemia survivors and address the growing public health challenge posed by the increasing number of cancer survivors, particularly those who have undergone aHSCT.

Table 7.Chronic kidney disease risk factors for the patients of the NephroLEA cohort without allogenic hematopoietic stem cell transplantation.

Footnotes

• Received November 18, 2024
• Accepted March 26, 2025

Correspondence

^#GM and LD-D contributed equally as senior authors

C. Domenech carine.halfondomenech@ihope.fr

Disclosures

No conflicts of interest to disclose.

Contributions

CD, GM and LD designed research. CD, PS, MDT, MP, VG, SD, MP, DP, JHD, PC, ST, JK, AT, IP, YR, AB, PA and GM performed research. JB, PA and LD contributed analytical tools. CD, PS, ZH, JB, GM and LD analyzed data. CD, PS, ZH, MDT, MP, VG, SD, MP, DP, JHD, PC, ST, JK, AT, IP, YR, AB, PA, JB, GM and LD wrote the paper. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Funding

The study was funded by Ligue National contre le Cancer.

Acknowledgments

We thank the Dr Erika Vacchelli for language editing.