Currently, as the 5-year survival in children with different malignancies has reached almost 70 %, oncologists aim at offering cancer survivors the best possible quality of life. According to published data, 40 % of cancer survivors develop chronic conditions [
15]. Brain tumour survivors are the most severely affected with late effects of cancer treatment [
16]. The most common sequelae include endocrine dysfunction, cardiotoxicity, neurotoxicity and nephrotoxicity. According to some authors, chronic nephrotoxicity affects 30–70 % of children after previous antineoplastic therapy [
6]. Since CKD leads to end-stage renal failure in many cases, its early diagnosis and appropriate treatment may have a significant impact on its course. However, potential preventive and therapeutic interventions are often delayed. Hence, prompt diagnosis during the prodromal stage is necessary, followed by the long-term observation.
GFR
The eGFR is an estimator of functional nephron count used in clinical practice. It constitutes a basis for CKD staging [
17]. Fitness level conditions endogenous creatinine and estimation of GFR. Furthermore, the determination of GFR using inulin (the gold standard for determining GFR) is associated with many obstacles, such as the need for repeated blood sampling, which may be troublesome particularly in paediatric population. Taking into consideration these reservations, the research for better parameters of GFR is ongoing. Mathematical equations involving additional parameters may increase diagnostic sensitivity of GFR in an individual patient. The National Kidney Foundation (NKF) recommends the Schwartz formula for estimating GFR in children (
https://www.kidney.org/content/creatinine-based-%E2%80%9Cbedside-schwartz%E2%80%9D-equation-2009). In 2009, Schwartz published a new, revised equation to estimate GFR in children with CKD. The new formula is based on the serum concentration of creatinine, CYS C and uric nitrogen [
12]. According to some authors, the simultaneous use of at least two methods (GFR estimated based on serum creatinine and CYS C levels) offers more reliable results [
18].
As a part of our study, renal function in each patient was estimated using both Schwartz formula and revised Schwartz formula. The difference between the mean GFR estimated by Schwartz formula (76 ml/min/1.73 m2) and GFR estimated revised Schwartz formula (63 ml/min/1.73 m2) was statistically significant (p < 0.0001).
In this study, the laboratory signs of subclinical nephrotoxicity (eGFR of 90–60 ml/min/1.73 m
2) were observed in 57.8 % of patients, which corresponds to the findings by other authors [
19]. In our patient cohort, 15.7 % of cancer survivors had eGFR of 60–30 ml/min/1.73 m
2 (including one individual with eGFR of 34 ml/min/1.73 m
2). Skinner et al. found eGFR <60 ml/min/1.72 m
2 in 13 % of their 25-person study group with malignant sarcoma treated with IFA before the end of 10-year follow-up [
20]. Similarly, in a group of 63 patients treated with cisplatin, 11 % of subjects had decreased GFR at the end of the 10-year follow-up [
7]. Oberlin et al. found decreased GFR <60 ml/min/1.73 m
2 in 21 % of patients with GFR being as low as 50.1 ml/min/1.73 m
2 in a single case. Their study group included children treated with IFA for soft tissue sarcomas and bone sarcomas [
6]. In the study by Zubowska et al., 20 % of childhood cancer survivors developed CKD (GFR < 60 ml/min/1.72 m
2) [
19]. Most studies were carried out on mixed case groups including only a few brain tumour survivors [
19,
20]. We have not found any studies focusing on late nephrotoxicity in patient samples consisting only of brain tumour survivors.
We did not find an association between the age at diagnosis and GFR in contrast to other authors who report correlations between CKD and patient age. Skinner [
21] showed more severe proximal tubular toxicity in younger children treated with IFA, as compared to older children. However, the same author later found older age at treatment to be a risk factor for cisplatin nephrotoxicity [
7]. There are also other papers emphasizing that patients younger than 5 years are more likely to develop CKD [
21‐
23]. Perhaps the lack of correlation of GFR with age in our work results from the fact that an average age of children was 9.5 years (consistent with epidemiological data regarding the age of the incidence of malignant CNS tumours), and there were only two children under 5 years of age in the studied group.
The fact that our study failed to show relationship between specific factors such as age and total dose of administered drugs could also be related to the modest sample size.
We did not observe an association between the time interval from treatment completion and GFR. Similarly, Zubowska et al. did not observe the association between the follow-up time and nephropathy in a group observed for the mean period of 4.6 years [
19]. Furthermore, other authors who carried out their studies at 10 years following chemotherapy completion did not observe kidney function deterioration with time [
20]. However, Oberlin in her study on 138 patients who had completed chemotherapy at least 5 years earlier found a longer follow-up period to be an independent risk factor for abnormal GFR [
6]. On the other hand, there are papers reporting spontaneous recovery of renal function after chemotherapy [
25,
26].
Many authors indicate an association between kidney damage and cumulative dose of chemotherapeutics. It is true for both such platinum derivatives as carboplatin, cisplatin [
25,
26] and IFA [
6,
24].
We did not confirm an association between the total dose of chemotherapeutics (cisplatin, carboplatin, IFA and CPA) and GFR, which might potentially be attributed to the fact that all children in our study group received chemotherapy in line with approved treatment protocols involving simultaneous use of IFA and/or CPA as well as cisplatin and/or carboplatin. Multidrug chemotherapy may constitute a confounding factor. All quoted papers assessed nephrotoxicity after monotherapy. Other authors reported a significant effect of IFA on the kidney function when a cumulative dose of the drug exceeded 60 g/m2. Children in our study group did not receive such a high total dose of IFA. In the literature, there is no conclusive data on the toxic dose of cisplatin. Skinner showed toxic effect of a daily dose of cisplatin exceeding 40 mg/m2. There is the lack of data on toxic total doses of carboplatin.
There was no statistically significant difference in GFR between the subgroup of children treated with craniospinal radiation therapy and those who did not receive radiotherapy. In Zubowska paper, the higher frequency of nephrotoxicity was observed among the children after abdominal radiotherapy [
19].
Markers of kidney injury
There are many studies assessing new markers of early kidney damage, immediately after activation of a damaging factor [
10]. Very few biomarkers exist for monitoring CKD.
Neutrophil gelatinase-associated lipocalin (NGAL) is a protein secreted to urine by the cells of the thick ascending limb of loop of Henle and a connecting tubule [
27]. Based on the available published data NGAL is thought to be a novel, sensitive and specific early marker of acute kidney injury (AKI) [
11]: in ischemic acute kidney injury [
28,
29], in a septic shock [
30], in contrast-induced nephropathy [
31]. However, the role of measurement of NGAL in CKD is still unclear. Several recent studies have showed increased serum NGAL levels in cases with CKD [
32‐
34]. Mitsnefes et al. showed that serum NGAL significantly correlated with cystatin C and both NGAL and cystatin C significantly correlated with eGFR in children with CKD stages 2–4 [
35]. In our study, no significant association between plasma NGAL levels and eGFR was found, and NGAL levels in children with kidney injury were comparable to those in children with normal eGFR. In line with Forest Fire Theory [
35], we consider NGAL to be a marker of an early as well as AKI. An initial elevation of NGAL level directly after the exposure to an insulting agent (“burning trees”) is followed by its decrease to the values comparable with those of healthy individuals (“burnt out trees”). Our findings correspond to those of Nikolas in a sample of 635 patients: NGAL levels in patients with CKD and those with normal kidney function were comparable [
36]. Plasma NGAL measurements may be influenced by a number of coexisting variables such as chronic hypertension [
37], systemic infection [
38] and neoplasms [
39]. Recent studies have showed that in the primary brain neoplasms, NGAL is over-expressed in tumours, which correlates with elevated serum and urine NGAL [
40,
41]. Thus, it seems that the concentration of serum NGAL in children with cancer can be caused by many factors and not only resulting from kidney damage.
CYS C undergoes glomerular filtration and a complete reabsorption in proximal tubules. However, it is not involved in tubular secretion. The stability of the molecule of CYS C and the fact that its concentration in blood depends only on the glomerular filtration rate affects its high diagnostic efficiency [
42]. CYS C has been studied intensively as a marker of kidney function in adults and in children. A number of cross-sectional studies have been published showing that serum CYS C concentration is more sensitive and correlate better to GFR than creatinine [
43,
44]. Many reports emphasize that serum cystatin C is a better indicator of GFR than serum creatinine concentration both in patients with chronic kidney damage [
45] including chemotherapy [
46]) and in patients with acute kidney injury: in papers by Liang [
47] and Krawczewski [
48]. Cys C level increased as early as 12 h following the exposure to a nephrotoxic agent and preceded the elevation of creatinine level by 1–2 days. Similarly, we showed a statistically significant negative correlation between the eGFR estimated by Schwartz formula and CYS C levels (
p < 0.003) across the entire study group. Children with CKD (GFR < 60 ml/min/1.73 m
2) had significantly higher levels of CYS C as compared to the remaining subjects (
p = 0.001).
Beta-2 microglobulin (B2MG) is a subunit of the major histocompatibility class I molecule produced by all nucleated cells [
49]. Its small size (11.8 kDa) allows beta-2-M to pass through the glomerular membrane, but it is almost completely reabsorbed in the proximal tubules. Serum beta-2-M levels are elevated in diseases associated with the increased cell turnover. This assay offers improved diagnostic sensitivity for the detection of altered GFR as compared to serum creatinine [
50]. B2MG level is also elevated in several benign conditions such as chronic inflammation, liver disease, some acute viral infections and a number of malignancies, especially haematologic malignancies associated with the B cell lineage [
51]. Many studies showed a very strong association between plasma B2MG levels and GFR [
52]. However, there are few reports evaluating the B2MG in children after chemotherapy. In our study, we observed a statistically significant negative correlation between GFR and B2MG levels (
p < 0.02). Additionally, we observed a strong positive correlation between serum CYS C level and B2MG (
p < 0.001). However, we did not observe an association between CYS C and NGAL (
p = 0.8) or B2MG and NGAL (
p = 0.4).
Tubulopathy
Fanconi syndrome is characterized by a global transport defect in the proximal tubules of the kidney. The spectrum of tubular dysfunction varies in different patients, ranging from a generalized proximal tubulopathy to partial reabsorption defects in phosphorus, calcium, glucose, amino acids and bicarbonate. The mechanism by which IFA induces Fanconi syndrome has not been identified [
8]. Distant effects of IFA tubulopathy in children may be growth retardation [
24] and the bone disease rickets [
53]. CPA, an isomer of IFA, displays only side effects in the form of haemorrhagic cystitis but not other nephrotoxicities.
In our study, 36 % of children presented with hypophosphatemia, which corresponds with the remaining signs of tubulopathy. Decreased fractional TRP (TRP < 85 %) and the decreased ratio of tubular maximum reabsorption of phosphate to GFR (TmP/GFR < 1.15) were found in 34 % of patients. The patients with hypophosphatemia also showed decreased tubular reabsorption and TmP/GFR. All subjects had normal serum and urine sodium, potassium and calcium levels. Among all patients, the osmolality of urine and plasma was normal. For this reason, we did not measure tubular handling of sodium, calcium and potassium. All urine samples tested were negative for the presence of glucose and protein. Three patients were found to have mild metabolic acidosis, which did not require intervention. Five children had microalbuminuria (including one patient with other symptoms of tubulopathy). In summary, 1/3 of children presented partial Fanconi syndrome, which clinically manifested itself mainly with hypophosphatemia and hyperphosphaturia. These findings confirm results obtained by other authors, who observed persistent dysfunction of renal proximal tubule cells in 5–10 % of children treated only with IFA [
6,
23] and in 30–40 % of children treated with IFA and concurrent administration of cis or carboplatinum [
24]. Patients included in these studies were mainly children after treatment of bone and soft tissue sarcomas, neuroblastoma and Wilms’ tumour. We have not found any studies focusing on tubular function assessment in patients after brain tumour therapy in children.
None of our patients experienced hemorrhagic cystitis caused by the CPA, which was probably associated with prophylactic administration of the uroprotectant mesna and intensive hydration during and after chemotherapy.