Acute kidney injury (AKI) is a common issue in children with oncological diseases. The reported incident in literature ranges from 16.9% to 52%.^[1,2] Its etiology is multifactorial, with key risk factors including drug-induced nephrotoxicity, tumor lysis syndrome, and infection, including sepsis emerging during periods of aplasia.^[3] The occurrence of AKI during treatment often necessitates dose reduction of the implicated medication or its replacement with a less nephrotoxic alternative. This may have an adverse effect on the overall therapeutic outcome.

The classical diagnosis of AKI relies on monitoring serum creatinine levels and the presence of oliguria. However, these are late markers, changing more than 48 hours after the onset of kidney injury. Furthermore, they are unable to differentiate between prerenal and renal injuries. Serum creatinine levels rise only after more than 50% of nephron function has been lost. This means that mild and moderate impairments in kidney function may go undetected.

According to the KDIGO criteria, the diagnosis of acute kidney injury (AKI) is based on changes in serum creatinine levels and/or estimated glomerular filtration rate (eGFR), as well as the presence of oliguria. In pediatric practice, the Schwartz formula is the most widely applied method for estimating GFR, relying on serum creatinine concentration and patient height. Nevertheless, serum creatinine is strongly influenced by muscle mass, and in the context of cachexia, frequently observed in children with malignancies, its levels may remain deceptively low even in the presence of AKI. As a result, the use of the classical Schwartz formula, a simple and widely accessible method, tends to overestimate glomerular filtration. The literature provides only limited evidence regarding the optimal formula to apply in pediatric oncology patients; however, the available studies consistently emphasize formulas that incorporate body mass. These include the body surface area–adjusted Schwartz formula, as well as the Brandt and the Cockcroft–Gault formulas.

In the context of drug-induced nephrotoxicity, the primary pathogenetic mechanism of AKI is tubular injury, which is typically non-oliguric, and even in cases of severe kidney damage, serum creatinine levels may remain unchanged. This highlights the need for new markers for the early diagnosis of AKI in pediatric oncology. One of them is NGAL, which has been extensively discussed in the literature as a biomarker of acute kidney injury (AKI), with proven sensitivity and specificity in various conditions such as AKI following perinatal asphyxia or after cardiac surgery with cardiopulmonary bypass, among others. Its role in pediatric oncology, and particularly as a marker of nephrotoxicity, remains incompletely elucidated. Other potential nephrotoxicity biomarkers include tubular markers of injury, such as the fractional excretion of phosphate (FeP%), or the renal phosphate threshold.

To compare classical markers of AKI with tubular injury markers and NGAL as early markers of drug-induced nephrotoxicity in children undergoing chemotherapy.

A cross-sectional study was conducted among 40 children undergoing chemotherapy with nephrotoxic medications at the Pediatric Hematology and Oncology Department of the Pediatrics Clinic at St. George University Hospital for the period 2021–2024.

Inclusion criteria: Children aged 0 to 18 years undergoing treatment with nephrotoxic medication (ifosfamide, cisplatin, or carboplatin), with no history of prior kidney injury or renal anomalies, and with signed informed consent.

Exclusion criteria: Presence of a comorbid condition—evidence of sepsis, or treatment with another nephrotoxic agent such as an antibiotic, antiviral, or antifungal medication, dehydration.

For all children, blood and urine samples were collected before the start of each chemotherapy cycle, as well as 12 hours after its completion. The following parameters were analyzed:

In blood, serum creatinine, urea, and electrolytes—including phosphorus and magnesium—were measured using an AU480 automated biochemical analyzer (Beckman Coulter) based on the turbidimetric principle. Estimated glomerular filtration rate (eGFR) was calculated using the original Schwartz formula, the body surface area–adjusted Schwartz formula (Equ. 1), and the Brandt equation (Equ. 2).

GFRESTmL/min=k⋅Ht(cm)SCR×BSA(m2)1.73m2 (Equ. 1)

where k=0.45 for children ≤1 year, 0.70 for pubertal males, and 0.55 for all other children, Scr is serum creatinine level (mg/dL), BSA: body surface area, Ht : the child’s height in cm.

GFRmL/min=k( agemos +6)×wtSCR (Equ. 2)

where k is 1.05 for males and 0.95 for females; agemos is the child’s age in months; wt is the child’s weight in kilograms; Scr is serum creatinine levels (mg/dL).

The KDIGO 2022 classification was used to stage acute kidney injury.^[4] According to the KDIGO criteria, AKI is defined as an increase in serum creatinine levels or a decline in GFR of 25% for stage 1, 50% for stage 2, and 75% or an eGFR less than 35 mL/min/1.73 m² for stage 3 (Table 1).

In urine, neutrophil gelatinase-associated lipocalin (NGAL) was assessed for each chemotherapy cycle. Urine NGAL levels were measured 12 hours before and after treatment using a chemiluminescent microparticle immunoassay for quantitative determination of NGAL. NGAL was considered elevated if its value exceeded 150 ng/mL, the reference value provided by the manufacturer. The NGAL/creatinine ratio was calculated to account for the diluting effect of hyperhydration administered during each chemotherapy cycle. In addition, the fractional excretion of phosphate (FeP%) was measured (Equ. 3), and the renal phosphate threshold was calculated (Equ. 4).

FeP%=PO4(U)×Cr(S)PO4(S)×Cr(U)×100 (Equ. 3)

where PO4: phosphate (mmol/L), Cr: creatinine (µmol/L), U: urine concentration, S: serum concentration.

TmpGFR=Pp−(Up×PcrUcr) (Equ. 4)

where P is the phosphate level, Pp is plasma phosphate concentration, Up is Urine phosphate concentration, Pcr is plasma creatinine concentration, and Ucr is the urine creatinine.

Under conditions of normal serum phosphate levels, the fractional excretion of phosphate (FeP) is normally up to 20%. In the presence of hypophosphatemia with proximal tubular dysfunction, FeP increases above 20%, whereas in intact tubular function it is expected to decrease, even approaching 0%.

The analysis included 116 cycles, which were divided into four categories based on their cycle sequence: “first,” “second,” “third,” and “from the fourth to the tenth cycle.” These categories include all of the analysis’s first, second, and third cycles, respectively, while the final category includes all of the fourth to tenth cycles.

The largest percentage of children were included in the “4–10 cycles” group. This is because, during the study period, in addition to newly diagnosed patients, children with disease relapse and those admitted for a new course of chemotherapy involving nephrotoxic medications were also included. Regarding the use of chemotherapy agents with high (ifosfamide, cisplatin) and low (carboplatin) nephrotoxicity, the children were divided into two main groups: a high nephrotoxicity group and a low nephrotoxicity group (Fig. 1).

A higher percentage of children were treated with highly nephrotoxic medications—92 out of 116 cycles, or 79.3% of all cycles, involved a high-toxicity drug. Fig. 2 presents a summary of the NGAL values at baseline (0 hour) and at 12 hours post-treatment, distributed by cycles.

The figure demonstrates that NGAL levels remain within the reference range and that, even at 12 hours, lower values are observed compared with those measured before the initiation of chemotherapy.

To determine whether there was a statistically significant difference between NGAL levels before each chemotherapy cycle and 12 hours after, the Wilcoxon signed ranks test was used (Table 5).

A statistically significant difference in NGAL levels before and 12 hours after the chemotherapy cycle was observed in patients who completed their second cycle (Z=−3.11, p=0.003) and third cycle (Z=−2.344, p=0.019), as well as in patients who underwent four or more chemotherapy cycles (Z=−0.856, p=0.004). However, NGAL levels remained within their reference values. Using KDIGO criteria, two episodes of AKI were diagnosed following treatment with a nephrotoxic agent. In both episodes, NGAL levels remained unchanged.

To avoid the dilution effect resulting from hyperhydration, which is part of the chemotherapy protocol for these children, NGAL levels were compared to creatinine levels in a urine sample collected 12 hours after the chemotherapy cycle. The results are presented in Table 6.

A statistically significant marker increase was found only in the “4–10 cycles” group, that is, in patients who underwent multiple courses of nephrotoxic chemotherapy.

One of the main KDIGO criteria for diagnosing acute kidney injury is the patient’s GFR. Using the original Schwartz formula, GFR values were calculated before and after each chemotherapy cycle. Normal values were considered between 90 and 160 mL/min/1.73 m², hyperfiltration was defined as values above 160 mL/min/1.73 m^{2 [5]}, and reduced GFR as values below 90 mL/min/1.73 m²(Fig. 3).

What stands out is the high percentage of hyperfiltration—almost 50% of all chemotherapy cycles—as well as the diagnosis of only one child with AKI before and two children after chemotherapy.

To determine whether there was a statistically significant difference in eGFR before and 12 hours after chemotherapy for each cycle, the Wilcoxon signed ranks test was used (Table 7).

The data show no statistically significant difference in eGFR before and after the chemotherapy cycle.

Given the fact that creatinine levels depend on muscle mass, in pediatric oncology their values are often pseudonormal, even in the presence of kidney injury. The standard Schwartz formula relies primarily on serum creatinine levels and the child’s height, which can lead to an overestimation of eGFR. The literature discusses the use of alternative formulas adapted for pediatric oncology patients^[6], two of which are the Schwartz formula corrected for the child’s body surface area, and the formula proposed by Brandt et al., which incorporates the child’s weight and age. The values obtained from these formulas have been compared with those derived using a steady-state methodology involving non-radioactive iothalamate (with a renal clearance similar to that of inulin), measured by high-performance liquid chromatography.^[6]Table 8 presents the comparison of GFR values calculated using the classical Schwartz formula and the body surface area – the adjusted Schwartz formula.

As demonstrated in Table 8, the data indicates a substantial increase in the number of chemotherapy cycles with reduced eGFR: from 2 at 12 hours using the standard Schwartz formula to 41 cycles when using the corrected formula. Furthermore, there is a decline in the percentage of cycles with hyperfiltration – to 31, which corresponds to 28% of all cycles, a percentage comparable to the results obtained using the radioisotope method for determining eGFR.

Using the adjusted Schwartz formula and the Brandt formula, a statistically significant number of children (p<0.001) were found to have reduced GFR before and after chemotherapy cycles. The number of affected children was nearly the same for both formulas, while the cycles with hyperfiltration were significantly fewer compared to the original formula. Table 9 presents a comparative overview of the differences in eGFR calculation using the three formulas.

A comparison was made between the proposed formulas to determine which formula detects the most deviations (reduced eGFR as well as hyperfiltration), with the adjusted Schwartz formula standing out as the one identifying the most pathology (Table 10).

Since the corrected Schwartz formula identified the highest number of pathological cases, it was selected for use in subsequent analyses.

In the course of the study, the elevated rate of hyperfiltration raised the question of whether this finding could be attributed to tubular creatinine loss in the context of tubular injury, which is recognized as the principal pathogenic mechanism of drug-induced nephrotoxicity.

The main markers for tubular injury are the fractional excretion of phosphate and the more accurate indicator relative to the patient’s GFR — the renal phosphate threshold. To clarify the etiology of hyperfiltration, these two indicators — eGFR and fractional excretion of phosphate/renal phosphate threshold — were compared. Using the classical Schwartz formula for calculating eGFR, children with reduced eGFR were detected only in the “4–10 cycles” group. The comparison between eGFR and fractional phosphate excretion data is presented at Table 11.

The comparison revealed no positive correlation between hyperfiltration and increased fractional excretion of phosphate, which is particularly evident during the first chemotherapy cycle. In all children with reduced eGFR, an elevated FeP% was observed.

A comparison was again made between the 12-hour eGFR calculated using the body surface area – corrected Schwartz formula and the fractional excretion of phosphate (FeP%) (Fig. 4).

Once again, no positive correlation was found between the cycles with hyperfiltration and elevated FeP% (>20%). Using the adjusted formula, it was found that elevated FeP% was mainly observed in children from cycles with reduced eGFR. Only in the “4–10” cycle it was equally represented across all eGFR groups, unlike the classic Schwartz formula, where increased phosphate excretion was mainly seen in children with normal eGFR.

One of the widely discussed in the literature early biomarkers for diagnosing subclinical AKI, as well as a predictive marker for the development of clinically manifested AKI, is NGAL. This marker has been shown to have high sensitivity and specificity for detecting acute kidney injury following an episode of hypoxia (e.g., newborns with perinatal asphyxia, children after cardiac surgery under cardiopulmonary bypass).^[7-12] Regarding its role as a marker of drug-induced nephrotoxicity, and specifically following chemotherapy, there is very little and conflicting data.^[13-17]Table 17 presents the most frequently cited studies on NGAL and its role as an early marker of AKI in pediatric oncology.^[18,19]

In the study by Mohamed Almalky et al., conducted among 30 children with solid tumors undergoing chemotherapy, the authors found a statistically significant difference between the number of children diagnosed with AKI based on urinary NGAL levels (10 out of 30) and those diagnosed using serum creatinine levels (0 out of 30). However, they used an NGAL cutoff value of >25 ng/mL for predictive significance^[13], rather than the >113 ng/mL threshold commonly discussed in the literature.

In another large prospective study by McMahon et al., involving 159 children undergoing chemotherapy with cisplatin, urinary NGAL levels were found to have low predictive value for AKI. Higher NGAL values were observed during later infusions of the drug compared to the initial cycles of therapy (AUC-ROC 0.54–0.75 for late infusion vs. AUC-ROC 0.48–0.65 for early infusion).^[14]

In a study by Miloševski-Lomić et al., published in 2023, emphasis is placed on the higher predictive value of using a combination of biomarkers (NGAL, KIM-1, L-FABP) for the diagnosis of AKI following the administration of nephrotoxic drugs in children, compared to the use of each biomarker individually.^[20]

In our study, using the KDIGO criteria, two episodes of AKI were diagnosed following courses of nephrotoxic medication. In both episodes, NGAL levels remained unchanged. In two children with disease relapse, elevated NGAL levels were observed both before and after the chemotherapy course; all of these courses were within the “cycle 4–10” group, meaning the patients had undergone previous chemotherapy cycles. For the rest of the children, especially for the second, third, and “4–10” cycles, a statistically significant difference in NGAL levels was observed; however, the values remained within the reference range and, in most cases, were lower than those measured prior to chemotherapy. One of the questions that arises is whether this could be due to urine dilution as a result of the hyperhydration administered to all children after each chemotherapy cycle. To account for this dilution effect, a NGAL-to-creatinine ratio was calculated from a single urine sample collected 12 hours after the chemotherapy cycle. A statistically significant increase of this ratio was observed only in the “4–10 cycle“ group, indicating that NGAL levels rise as a result of the cumulative nephrotoxic effect—an observation consistent with the findings of Kelly R. McMahon. This identifies NGAL as a marker that is not reliable for diagnosing AKI induced by nephrotoxic drugs, particularly chemotherapeutic agents.

Regarding the classical indicators for diagnosing AKI, the more accurate measure is the estimation of eGFR. Its determination is extremely important in oncology, especially when nephrotoxic chemotherapy regimens are to be administered, in order to optimize drug dosing according to renal function. There are very few studies in the literature concerning GFR in pediatric oncology patients. The method used to determine GFR is also widely discussed. The most accurate methods for measuring absolute GFR are radioisotope techniques (such as [51Cr]-EDTA). Unfortunately, the need for multiple blood samples or the lack of access to the respective radioisotope makes these methods rarely used and less preferred in clinical practice. One of the proposed methods for estimating GFR is the classical Schwartz formula, which, given the low body weight of pediatric oncology patients, is recommended to be adjusted for body surface area.^[6] An interesting finding across all these studies is the high prevalence of hyperfiltration, regardless of the chosen GFR measurement method, both before the start of treatment and after the initial chemotherapy cycles. Such is the study by Kwatra et al., which found a 43.5% rate of hyperfiltration among 177 children, measured by DTPA (diethylenetriaminepentaacetic acid) clearance during at least one chemotherapy cycle. In a study by Hjorth et al., also using a radioisotope method (iohexol clearance), a hyperfiltration rate of 31% was reported.^[6,21,22] In the follow-up of eGFR levels in our study, a high rate of hyperfiltration was also observed, comparable to the cited studies, both at disease onset (55 cycles, 47.8%), and after chemotherapy (58 cycles, 50.43%) measured at 12 hours post-cycle. Hyperfiltration was defined as eGFR values >160 mL/min/1.73 m². The pathophysiology of hyperfiltration is widely discussed, with the most accepted theory being the hypermetabolic state in which children with oncological diseases find themselves, especially during chemotherapy.^[6,21,22] The hypothesis of increased creatinine excretion due to tubular injury during chemotherapy cycles as an etiology for hyperfiltration was not confirmed, since no statistically significant correlation was found between children with hyperfiltration and fractional excretion of phosphate or the renal phosphate threshold.

The use of the classical Schwartz formula is an easy and widely accessible method but tends to overestimate glomerular filtration. There are few studies in the literature regarding which formula to use for pediatric oncology patients, but all recommend formulas based on body weight. These include the body surface area–corrected Schwartz formula, the Brandt formula, or the Cockcroft–Gault formula.^[6,21-25] Our study used two of these formulas, the adjusted Schwartz formula and the Brandt formula. With both formulas, the percentage of hyperfiltration observed with the classical Schwartz formula decreased significantly but remained high—up to 28%. This result is comparable to those obtained in studies measuring GFR in pediatric oncology patients using radioisotope methods; however, an increase in cases with reduced eGFR was also observed. Unlike the classical formula, where children with increased fractional excretion of phosphate and correspondingly low renal phosphate threshold mostly have normal GFR, the formulas based on body weight tend to show these children predominantly have reduced GFR.

Some of the medications used in pediatric oncology, especially in children with solid tumors, are nephrotoxic and primarily cause tubular injury. The proximal tubule, being the most metabolically active part of the tubular apparatus, is the most affected. This leads to the manifestation of partial or generalized proximal tubular dysfunction—renal Fanconi syndrome. The clinical presentation is characterized by aminoaciduria, hypophosphatemia, low-molecular-weight proteinuria (β2-microglobulin, retinol-binding protein), normoglycemic glucosuria, metabolic acidosis, hypokalemia, and hypouricemia. While tubular proteinuria and glucosuria do not have a significant impact on the body, increased phosphate excretion leading to hypophosphatemia is one of the main clinically important tubular dysfunctions as it leads to skeletal demineralization and osteomalacia, proximal muscle weakness, bone pain, spontaneous fractures, or the development of hypophosphatemic rickets. Due to preserved distal tubular function, polyuria is a rare clinical finding, and metabolic acidosis is generally mild (HCO₃>15).^[25-28]

By using only standard methods to assess renal function—serum creatinine levels and oliguria—this type of injury often goes unrecognized, since glomerular function is usually preserved, except in cases where the distal tubule is also affected and creatinine secretion is impaired. In the conducted study, increased phosphate excretion was observed before the chemotherapy cycle in only 10% of all 116 cycles, while 12 hours after the treatment course, the percentage increased significantly to 35.5%. Furthermore, as the cycle number increased, so did the number of children with tubular injury, indicating a cumulative nephrotoxic effect. A similar result was observed using the more precise marker, the renal phosphate threshold. A low renal phosphate threshold was present before the start of therapy in only 17 cycles (14.7%), and 12 hours after the cycle in 44 cycles (37.0%). Here too, an increase in the percentage of low renal phosphate threshold cases was observed with the increasing number of chemotherapy cycles administered.

Monitoring serum electrolyte levels of phosphorus and magnesium shows that with the cumulative nephrotoxic effect over successive cycles, clinically manifest hypophosphatemia and hypomagnesemia develop. The frequency of hypophosphatemia reported in the literature varies from 0 to 36.8%, and tubular phosphate reabsorption disorders range from 0 to 62%.^[29] Our data are consistent with the ones reported in the literature. As a rule, upon the manifestation of tubular toxicity, the medication should be discontinued, or if it must be continued (with no alternative), its dose should be reduced according to the severity of acute kidney injury. In most cases, tubulopathy is reversible and transient, lasting a few months, but in some children, it becomes chronic, with a reported incidence of tubular injury reaching up to 20–25%.^[30] This necessitates the development of a monitoring algorithm for children who have undergone chemotherapy, including for late side effects related to kidney function.

We present an algorithm for monitoring renal function in children undergoing chemotherapy with nephrotoxic drugs (Fig. 8).

Our study has certain limitations, most notably the small sample size and its single-center design, which restrict the ability to draw more robust conclusions.

Acute kidney injury (AKI) is a common complication in the treatment of pediatric oncology patients. One of the leading risk factors for its development is drug-induced nephrotoxicity, particularly associated with nephrotoxic chemotherapeutic agents. This necessitates careful monitoring of renal function both during therapy and after its discontinuation.

The widely discussed biomarker NGAL, although considered an early marker of AKI, has limitations when used as a standalone diagnostic tool in children receiving nephrotoxic chemotherapy. While its role has been better established in hypoxic conditions and certain surgical settings, its diagnostic and predictive value in pediatric oncology remains controversial. It is influenced by multiple factors, including the cumulative effect of treatment cycles and the impact of hyperhydration.

The measurement of eGFR remains a key parameter in evaluating kidney function in pediatric oncology patients. The use of formulas adjusted for body surface area or body mass provides more accurate assessments compared to the classical Schwartz formula.

Markers of tubular injury, such as FeP% and Tmp/GFR (renal phosphate threshold), can serve as valuable tools for the early detection of drug-induced nephrotoxicity in pediatric oncology. The observed cumulative toxicity highlights the urgent need for the development of a standardized algorithm for monitoring renal function both during and after the completion of oncological treatment in order to ensure timely intervention and reduce the risk of progression to chronic kidney disease.