Received: 23 January 2021 Revised: 4 June 2021 Accepted: 9 July 2021

DOI: 10.1002/pbc.29255 Pediatric

Blood &
PSYCHOSOCIAL AND SUPPORTIVE CARE:
Cancer The American Society of
Pediatric Hematology/Oncology

RESEARCH ARTICLE

Fluid overload and acute kidney injury in children with tumor

lysis syndrome

Kayla Flood1 Jacob Rozmus2 Peter Skippen3 Douglas G. Matsell1

Cherry Mammen1

1
Department of Pediatrics, Division of
Nephrology, BC Children’s Hospital, Abstract
Vancouver, British Columbia, Canada
Aim: Tumor lysis syndrome (TLS) is a common oncologic emergency among patients
2
Department of Pediatrics, Division of
Oncology and Hematology, BC Children’s
with pediatric hematologic malignancies. The mainstay of TLS management is aggres-
Hospital, Vancouver, British Columbia, Canada sive intravenous hydration. However, the epidemiology of fluid overload (FO) and
3
Department of Pediatrics, Division of Critical acute kidney injury (AKI) in this population is understudied. In this study, we aimed
Care, BC Children’s Hospital, Vancouver,
British Columbia, Canada to describe the incidence, severity, and complications of FO and AKI among pediatric
patients with TLS.
Correspondence
Methods: We completed a single-center retrospective cohort study of pediatric
Kayla Flood, Royal University Hospital, 103
Hospital Dr, Saskatoon, SK, Canada. patients with a new diagnosis of hematologic malignancy over a 10-year period.
Email: [email protected]
Patients with TLS were analyzed in two groups based on the severity of AKI and
FO. Charts were reviewed for complications associated with AKI and FO including
hypoxemia, mechanical ventilation, hyponatremia, pulmonary edema, pediatric inten-
sive care (PICU) admission, and need for renal replacement therapy (RRT).
Results: We analyzed 56 patients with TLS for FO and AKI. We found severe FO
(≥10%) occurred in 35.7% (n = 20). PICU admission occurred in 35% of patients with
severe FO compared to 8.3% in those with mild/moderate FO <10% (p = .013). Com-
plications of hypoxemia (30% vs. 5.6%, p = .012) and pulmonary edema (25% vs. 2.8%,
p = .010) were more common among those with severe FO. AKI occurred in 37.5%
(n = 21) patients and resulted in a significant increase in PICU admission and require-
ment for RRT (p = .001 and <.001, respectively).
Conclusion: Our results show FO and AKI are common, and often unrecognized com-
plications of TLS associated with increased morbidity. Prospective, multicenter studies
are needed to further dissect the burden of FO and AKI within this vulnerable popula-
tion.

KEYWORDS
acute kidney injury, fluid overload, leukemia, lymphoma, tumor lysis syndrome

Abbreviations: %FO, percent fluid overload; AKI, acute kidney injury; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; BCCH, British Columbia Children’s Hospital; CRRT,
continuous renal replacement therapy; DLBCL, diffuse large B-cell lymphoma; FO, fluid overload; IHD, intermittent hemodialysis; IQR, interquartile range; IV, intravenous; KDIGO, Kidney Disease
Improving Global Outcomes; LTLS, laboratory tumor lysis syndrome; NS, normal saline; PICU, pediatric intensive care unit; RRT, renal replacement therapy; SCr, serum creatinine; TLS, tumor lysis
syndrome; UO, urine output.

Pediatr Blood Cancer. 2021;68:e29255. wileyonlinelibrary.com/journal/pbc © 2021 Wiley Periodicals LLC 1 of 9

https://doi.org/10.1002/pbc.29255
2 of 9 FLOOD ET AL .

1 INTRODUCTION need for renal replacement therapy (RRT). AKI is also a complication
of TLS and a known risk factor for FO but has not been well studied in
Tumor lysis syndrome (TLS) is a common oncologic emergency among children.8–10 Consequently, we aimed to describe the incidence, sever-
pediatric hematologic malignancies.1–3 TLS results from the rapid ity, and complications of FO and AKI among pediatric patients with TLS.
release of intracellular ions, nucleic acids, and proteins into the extra-
cellular space leading to hyperuricemia, hyperkalemia, hyperphos- 2 METHODS
phatemia, and hypocalcemia.1–3 The release of intracellular contents
can lead to clinically significant sequelae including acute kidney injury We completed a single-center retrospective cohort study of pediatric
(AKI), arrhythmias, seizures, and sudden death.1–3 The incidence of TLS patients aged 1–18 years with a new diagnosis of a hematologic malig-
in pediatric patients varies depending on the underlying cancer. Hema- nancy who were at risk for TLS between January 1, 2009 and January
tologic malignancies, such as leukemia and lymphoma, with large num- 1, 2019 (Figure 1). Patients at risk for TLS included acute lymphoblas-
bers of rapidly growing chemosensitive cells are at the highest risk with tic leukemia (ALL), acute myeloid leukemia (AML), Burkitt lymphoma,
an estimated one in four children developing TLS.4 and diffuse large B-cell lymphoma (DLBCL). Patients with solid organ
There are very few universally accepted clinical practice guide- malignancies were excluded from our cohort as these patients were
lines for the management of TLS in pediatric patients. However, considered at low risk for TLS and typically did not receive aggres-
one of the mainstays of management is aggressive intravenous (IV) sive IV hydration. All patients were identified using a pediatric oncol-
hydration.1,2,5,6 It is hypothesized that large volumes of IV fluids ogy database at British Columbia Children’s Hospital (BCCH) in Van-
drive urinary flow aiding in the excretion of uric acid and phospho- couver, Canada, which is the only pediatric tertiary care hospital in the
rous, thereby preventing formation of uric acid and calcium-phosphate province that provides induction chemotherapy.
deposits in the kidney.7 Use of diuretics to achieve urinary flow All patients were retrospectively evaluated for laboratory tumor
rates is recommended in the absence of obstructive uropathy and/or lysis syndrome (LTLS) based on the 2004 Cairo and Bishop definition.2
hypovolemia.1 LTLS was defined as two or more of the following serum laboratory
While experts recommend vigorous hydration as a pillar of TLS pre- abnormalities: uric acid ≥476 µmol/L or ≥25% increase from base-
vention and management, the burden of fluid overload (FO) in this line, potassium ≥6 mmol/L or ≥25% increase from baseline, phos-
population is unknown. Excess fluid administration can lead to clin- phate ≥2.1 mmol/L or ≥25% increase from baseline, and total cal-
ically significant FO and the development of hypoxemia, pulmonary cium ≤1.75 mmol/L or ≥25% decrease from baseline.2 Baseline values
edema, hypertension, pediatric intensive care unit (PICU) transfer, and were established from initial bloodwork upon presentation to care. We

FIGURE 1 Patient inclusion and exclusion criteria

FLOOD ET AL . 3 of 9

defined the TLS period as 3 days before to 7 days after initiation of TA B L E 1 Characteristics of patients with tumor lysis syndrome
chemotherapy.2 If a patient was admitted less than 3 days prior to the (TLS) (n = 56)
first day of chemotherapy, we used the admission date until 7 days after Age (years; median, IQR) 5 (3, 11)
the first dose of chemotherapy. We excluded patients who had surgery
Sex (male) 60.7% (n = 34)
prior to chemotherapy, with the exception for central line access, and
Diagnosis: ALL 83.9% (n = 47)
whose induction chemotherapy occurred at an out-of-province center.
Diagnosis: AML 5.4% (n = 3)
AKI was defined and staged according to the KDIGO (Kidney
Diagnosis: Burkitt lymphoma 10.7% (n = 6)
Disease Improving Global Outcomes) definition based on the maxi-
mum rise of serum creatinine (SCr) from a baseline within the prior Initial creatinine (µmol/L) 40.5 (34, 55)
2
3 months.11 AKI was staged for severity, in accordance with KDIGO, Initial GFR (ml/min/1.73 m ) 100.6 [30.5]
with stage 1: ≥1.5–1.9 times baseline, stage 2: ≥2–2.9 times baseline, ΔCreatinine from baseline (fold change) 1.4 (1.2, 1.7)
and stage 3: ≥3 times from baseline.11 KDIGO AKI urine output (UO) ΔPotassium from baseline (%) 14.1 (5.9, 21.1)
criteria was not utilized due to lack of hourly accurate UO recording on ΔPhosphate from baseline (%) 30.5 (18.3, 47.1)
inpatients. As none of the patients had SCr in the prior 3 months, we
ΔCalcium from baseline (%) −8.6 (−2.6, −17.7)
imputed a baseline creatinine using the bedside Schwartz formula with
ΔUric acid from baseline (%) 0 (0, 57)
an estimated glomerular filtration rate (GFR) of 120 ml/min/1.73 m2
9
Initial WBC (×10 /L) 11.5 (5.1, 50.6)
and the patient’s height, as used in other AKI studies.12–14 The max-
Peak WBC (×109 /L) 12.1 (5.7, 49.2)
imum cumulative percent fluid overload (%FO) was calculated as the
([peak weight − nadir weight]/nadir weight) × 100. Peak and nadir Peak lactate dehydrogenase (U/L) 2129 (1180, 4698)

weights were identified within the TLS period. Nadir weight was uti- Initial fluid rate (ml/kg/h) 3.7 [1.5]
lized instead of admission weight with the concern that FO was poten- Initial fluid rate (ml/m2/day) 2326 [774]
tially already present upon admission to BC Children’s Hospital as
Note: ( ) Denotes median and interquartile range, and [ ] denotes mean and
many children are transported from various sites across the province standard deviation.
and may have received IV fluid resuscitation already.
Severe FO was defined as ≥10% using the previously described (IQR) including 25th percentile and 75th percentile values (nonpara-
weight-based calculation. Those with severe FO (≥10%) were com- metric data) were used for all study variables. We used t-tests or
pared to those with mild/moderate FO (<10%). Severe AKI was defined Mann–Whitney U tests, as appropriate, to compare continuous vari-
as ≥2 times baseline SCr. Patients with severe AKI were compared to all ables and chi-square tests to compare categorical data between two
other patients (no/mild AKI). Charts were reviewed for potential com- groups. All statistical analysis was performed with current SPSS soft-
plications associated with severe FO (≥10%) and severe AKI (≥2 times ware (version 26). A p-value of ≤.05 was considered statistically signif-
baseline SCr) within the TLS period including the presence of hypox- icant.
emia, pulmonary edema, hyponatremia, mechanical ventilation, need
for RRT, and PICU admission. Hypoxemia was defined as need for sup-
plemental oxygen. Pulmonary edema was defined as radiographic evi- 4 RESULTS
dence of pulmonary edema on chest radiographs as described in the
report. Hyponatremia was defined as at least one serum sodium value A total of 336 eligible patients identified from the pediatric oncology
<135 mmol/L. Mechanical ventilation included continuous positive air- database were considered at risk of TLS based on their hematologic
way pressure (CPAP), bi-level positive airway pressure (BiPAP), and diagnosis between 2000 and 2019 (Figure 1). Of the 324 reviewed,
endotracheal intubation and ventilation. 19.4% (n = 63) met criteria for LTLS. The median age at presenta-
In our analysis of FO and AKI, we excluded patients with a positive tion was 5 years (IQR 3, 11) and 60.7% were male (Table 1). Of these,
blood culture and those with less than two weights recorded within 256 patients (79%) had ALL of whom 21% (n = 54) developed TLS, 31
the defined TLS period. Patients with positive blood cultures were patients (9.6%) had Burkitt lymphoma of which 19.4% (n = 6) devel-
excluded as they may have developed AKI and/or FO due to sepsis- oped TLS, and 23 patients (7.1%) had AML of which 13% (n = 3) devel-
related complications. oped TLS. None of the 14 patients diagnosed with DLBCL developed
This study was approved by the research ethics board at BCCH. TLS. We analyzed 56 patients with TLS (Figure 1) for FO and AKI. We
Patient characteristics and clinical variables were collected in a RED- found 9% (n = 5) met criteria for TLS upon presentation, while the
Cap database. remaining 51 developed TLS following administration of chemother-
apy. All patients in our cohort survived induction chemotherapy.
All patients received allopurinol upon admission, while 30.4%
3 STATISTICAL ANALYSIS (n = 17) received rasburicase in addition to allopurinol. There was
no significant difference in age between those who received rasburi-
Descriptive statistics including proportions (%), mean ± standard devi- case and those who did not (median age 6.5 vs. 5 years, p = .34,
ations (normally distributed data), and median with interquartile range respectively). All patient’s with Burkitt lymphoma and TLS received
4 of 9 FLOOD ET AL .

TA B L E 2 Comparison of patients with no/mild acute kidney injury (AKI) versus severe AKI

No/Mild AKI (SCr <2.0×)n = 47 Severe AKI (Scr ≥2.0×)n = 9 p-value

Age at diagnosis (years) 5 (3, 8.5) 11 (5, 14) .06
Sex (male) 57.4% (n = 27) 77.8% (n = 7) .25
ALL 91.5% (n = 43) 44.4% (n = 4) <.001
AML 4.3% (n = 2) 11.1% (n = 1) .4
Burkitt lymphoma 4.3% (n = 2) 44.4% (n = 4) <.001
Rasburicase use 21.3% (n = 10) 77.8% (n = 7) .001
Diuretic use 40.4% (n = 19) 66.7% (n = 6) .15
Peak creatinine (mmol/L) 45 (38, 58) 122 (99, 144) <.001
ΔCreatinine from baseline (fold change) 1.3 (1.2, 1.5) 2.4 (2.2, 3.2) <.001
Peak potassium (mmol/L) 4.9 (4.6, 5.2) 5.0 (4.6, 5.7) .52
ΔPotassium from baseline (%) 15 (6.1, 29.3) 8.7 (4.8, 29) .55
Peak uric acid (µmol/L) 270 (209, 358) 495 (375, 745) .05
ΔUric acid from baseline (%) 7.2 (0, 58.7) 0.00 (0, 0.5) .18
Peak phosphate (mmol/L) 2.1 (1.89, 2.22) 2.31 (2.25, 3.7) .02
ΔPhosphate from baseline (%) 30.1 (17.6, 45.7) 36.1 (19.3, 154) .22
Trough calcium (mmol/L) 2.02 (1.83, 2.14) 1.8 (1.57, 1.89) .003
ΔCalcium from baseline (%) −8.0 (−15.8, −2.4) −31.1 (−37.4, −5.6) .81
2
Initial fluid rate (ml/m /day) 2322 [776] 2344 [806] .94
Initial fluid rate (ml/kg/h) 3.7 [1.5] 3.4 [1.8] .63
Peak urea (mmol/L) 7 (6.15, 8.35) 11.1 (10.4, 19.7) <.001
ΔUrea from baseline (fold change) 2.7 (1.95, 3.95) 7 (0.4, 14.9) .14
Initial lactate dehydrogenase (LDH) (U/L) 1230 (832, 3069) 3202 (1755, 5381) .04
Peak LDH (U/L) 2001 (1179, 3821) 7335 (2394, 8382) .02
Initial WBC (×109 /L) 11.5 (4.75, 50.7) 11.1 (8.5, 47.2) .66
9
Peak WBC (×10 /L) 11.9 (5.55, 48.1) 15.2 (69, 113) .79

Note: ( ) Denotes median and interquartile range, and [ ] denotes mean and standard deviation.

rasburicase in comparison to 67% with AML and 19% with ALL. The Anti-hypertensive medications were prescribed at least once in 28.6%
median initial serum uric acid level (478 vs. 172 µmol/L, p < .001) and (n = 16) patients, with all medications being a calcium channel blocker,
peak uric acid level (478 vs. 249 µmol/L, p < .001) were significantly either nifedipine or amlodipine.
higher in the rasburicase group compared to the no-rasburicase group, FO was defined as mild (<5%) in 25% (n = 14), moderate (≥5%
respectively. to <10%) in 39.3% (n = 22), and severe (≥10%) in 35.7% (n = 20) of
The most common initial fluid prescribed on admission was 5% dex- patients with TLS. Of those patients with severe FO, 35% (n = 7) had
trose in normal saline (D5NS) in 45 patients (80.4%), 0.9% saline (NS) ≥20% FO. The median number of weight measures was six (IQR 4,
in seven patients (12.5%), 5% dextrose in 0.45% saline (D5 ½ NS) in 8) during the described TLS period. Among the patients with severe
two patients (3.6%), and other fluid types in two patients (3.6%). The FO, 10% (n = 2) presented with an AKI. We compared clinical char-
average initial fluid rate prescribed was 3.7 ± 1.5 ml/kg/h or 2326 acteristics between those with severe (≥10%) and mild/moderate FO
± 774 ml/m2 /day. With estimated maintenance fluid requirements of (<10%) (Table 2). Those with severe FO were younger 3.5 years (IQR
1500 ml/m2 /day, the average initial fluid rate prescribed was 1.6 ± 0.5 2, 6.3) versus 6.5 years (IQR 4.8, 11), p = .004, more likely to have a
times maintenance. primary diagnosis of Burkitt lymphoma, (p = .01), more likely to have
Diuretics were prescribed at least once during the course of the received rasburicase (45% vs. 22.2%, p = .08), more likely to have
TLS period in 44.6% (n = 25) of patients. Intermittent dosing of IV received diuretics (70% vs. 30.6%, p = .004), and had higher IV fluid
furosemide was the only formulation of diuretics prescribed. Among rates on admission (4.4 ± 2 vs. 3.3 ± 1.1 ml/kg/h, p = .01) compared
all patients who received diuretics, the median time to administration to those patients with mild/moderate FO, respectively (Table 3). The
of the first dose was 2 days (IQR 2, 3) from admission. Among those frequency of peak %FO increased daily from admission, with the peak
with severe FO, 70% (n = 14) received at least one dose of diuretic %FO occurring most frequently on day 2 postchemotherapy initiation
with the median time to first dose from admission of 1 day (IQR 1, 2). in 25% (n = 14) of patients (Figure 2A).
FLOOD ET AL . 5 of 9

TA B L E 3 Comparison of patients with mild/moderate fluid overload (FO) and severe FO

Mild/moderate FO Severe FO
(<10%)n = 36 (≥10%)n = 20 p-value
Age at diagnosis (years) 6.5 (4.8, 11) 3.5 (2, 6.3) .004
Sex (male) 58.3% (n = 21) 65% (n = 13) .63
ALL 88.9% (n = 32) 75% (n = 15) .18
AML 8.3% (n = 3) 0% (n = 0) .18
Burkitt lymphoma 2.8% (n = 1) 25% (n = 5) .01
Rasburicase use 22.2% (n = 8) 45% (n = 9) .08
Diuretic use 30.6% (n = 11) 70% (n = 14) .004
Peak creatinine (mmol/L) 57 (43, 76.25) 43 (38, 59.5) .11
ΔCreatinine from baseline 1.35 (1.19, 1.70) 1.38 (1.23, 1.59) .81
(fold change)
Peak potassium (mmol/L) 4.9 (4.6, 5.0) 5.1 (4.7, 5.6) .08
ΔPotassium from baseline (%) 19.3 (9.3, 31) 8.7 (3.1, 22.1) .05
Peak uric acid (µmol/L) 264 (204, 347) 413 (241, 639) .04
ΔUric acid from baseline (%) 0 (0, 58) 4 (0, 47) .67
Peak phosphate (mmol/L) 2.06 (1.85, 2.25) 2.16 (1.98, 2.49) .12
ΔPhosphate from baseline (%) 33.6 (18.35, 47.16) 23.20 (18.58, 44.08) .63
Trough calcium (mmol/L) 2 (1.88, 2.14) 1.86 (1.61, 2.05) .02
ΔCalcium from baseline (%) −8.1 (−16.5, −5.5) −10.0 (−29.8, −0.7) .98
Initial WBC (×109 /L) 10.4 (5.7, 24) 13.4 (4.6, 54.2) .64
9
Peak WBC (×10 /L) 11.4 (6.1, 24) 15.2 (5.6, 58.1) .42
Initial fluid rate (ml/m2 /day) 2202 [604] 2548 [990] .11
Initial fluid rate (ml/kg/h) 3.3 [1.1] 4.4 [2.0] .01

Note: ( ) Denotes median and interquartile range, and [ ] denotes mean and standard deviation.

PICU admission occurred in 35% (n = 7) of patients with severe FO nificant, we observed a trend toward higher %FO among those with
compared to 8.3% (n = 3) in those with mild/moderate FO (p = .013) severe AKI compared to no/mild AKI (9.9% vs. 7.4%, p = .06). During
(Table 4). With regards to complications, patients with severe FO were the TLS period, the peak SCr was observed most commonly on the day
more likely to experience hypoxemia (30% vs. 5.6%, p = .01) and pul- chemotherapy was first administered (33.9%, n = 19), with the second
monary edema (25% vs. 2.8%, p = .01) compared to those patients most frequent peak occurring the following day (Figure 2B).
with mild/moderate FO, respectively. There was no statistical differ- When compared to patients with no/mild AKI, patients with severe
ence between the proportions of patients needing mechanical ventila- AKI were more likely to require PICU admission (55.6% vs. 10.6%,
tion (15%, n = 3) with severe FO compared to mild/moderate FO (2.8%, p = .001), antihypertensives (55.6% vs. 23.4%, p = .05), and RRT (44.4%
n = 1), p = .089. Similarly, no difference was seen in those patients with vs. 0%, p < .001) (Table 4). In patients requiring RRT (n = 4), two
severe FO requiring RRT (15%, n = 3) compared to mild/moderate FO received intermittent hemodialysis (IHD) and two received continuous
(2.8%, n = 1), p = .089. renal replacement therapy (CRRT). Of the patients who received IHD,
Among patients with TLS, AKI incidence was 37.5% (n = 21), of one patient received a single treatment, and the other patient required
which 57.1% (12/21) had Stage 1 AKI, 28.6% (6/21) had stage 2 AKI, two consecutive treatments. Both patients who received CRRT were
and 14.3% (3/21) had stage 3 AKI. Among those with TLS, 4% (n = 2) prescribed the continuous veno-venous hemodiafiltration (CVVHDF)
met criteria for AKI at initial presentation. We found 83.9% (n = 47) had modality.
no/mild AKI (SCr <2 × baseline) compared to 16.1% (n = 9) with severe
AKI (SCr ≥2 × baseline). Patients with Burkitt lymphoma were more
likely to develop severe AKI (44.4%, n = 4) than no/mild AKI (4.3%, 5 DISCUSSION
n = 2, p < .001) (Table 2).
When compared to patients with no/mild AKI, patients with severe In one of the first pediatric studies analyzing the epidemiology of FO
AKI were more likely to have received rasburicase (77.8% vs. 21.3%, in TLS, we demonstrated a high burden of FO in our TLS patients, with
p = .001), had a higher median initial serum uric acid (495 vs. 39.2% having peak FO ≥5% and 35.7% having FO ≥10%. We defined
223 µmol/L, p = .04), and higher peak lactate dehydrogenase (7335 vs. severe FO as ≥10%, which has been strongly correlated with poor out-
2001 U/L, p = .02), respectively (Table 2). While not statistically sig- comes including mortality.15,16 As opposed to using strict fluid intake
6 of 9 FLOOD ET AL .

F I G U R E 2 The day of peak percent fluid overload (%FO) (A) and peak serum creatinine (SCr) (B) in relation to day 0, which is defined as the
first day of chemotherapy

and output from charts, we used a weight-based determination of FO, also due to the fact that younger children in our cohort were prescribed
as accurate fluid intake and output was not available in all patients. higher IV fluid rates, per body surface area. Patients with Burkitt lym-
However, weight-based determination of FO has been shown to cor- phoma were also more likely to develop severe FO, which likely reflects
relate well with traditional daily fluid in–fluid out methods in predict- that these patients are considered very high risk for TLS and may
ing morbidity and mortality.17–19 We found severe FO was more likely have prompted more aggressive fluid hydration early in their treatment
to occur in younger children, which could be due to challenging assess- course. Peak FO ≥15% has been found to correlate significantly
ment and identification of clinical FO compared to older children, and with peak oxygenation index in the PICU.20 Similarly, in our cohort,
FLOOD ET AL . 7 of 9

TA B L E 4 Comparison of patient outcomes associated with percent fluid overload (%FO) and acute kidney injury (AKI)

%FO AKI

<10% (n = 36) ≥10% (n = 20) p-Value <2.0 × SCr (n = 47) ≥2.0 × SCr (n = 9) p-value
PICU admission 8.3% (n = 3) 35% (n = 7) .013 10.6% (n = 5) 55.6% (n = 5) .001
Hypoxemia 5.6% (n = 2) 30% (n = 6 .012 12.8% (n = 6) 22.2% (n = 2) .46
Pulmonary edema 2.8% (n = 1) 25% (n = 5) .010 8.5% (n = 4) 22.2% (n = 2) .22
Mechanical ventilation 2.8% (n = 1) 15% (n = 3) .089 4.3% (n = 2) 22.2% (n = 2) .06
Renal replacement therapy 2.8% (n = 1) 15% (n = 3) .089 0% (n = 0) 44.4% (n = 4) <.001
Hyponatremia 22.2% (n = 8) 20% (n = 4) .846 23.4% (n = 11) 11.1% (n = 1) .41
Hypertension 22.2% (n = 8) 40% (n = 8) .158 23.4% (n = 11) 55.6% (n = 5) .05

mechanical ventilation and RRT were both more common in severe FO, Our study had some limitations owing to the retrospective design
though not statistically significant, likely due to the small sample size. and small sample size from a single center. Many of our statistical com-
Interestingly, while diuretic use was more common in those with severe parisons between groups appeared clinically significant, but did not
FO, a relatively large proportion of patients with severe FO (30%) achieve statistical significance, likely due to the small size. In addition,
did not receive any diuretics, likely due to unrecognized FO in some 11.1% (n = 7) of patients with TLS were excluded from the analysis due
patients. to missing weights (n = 5), missing chart (n = 1), and a positive blood
AKI, as defined and staged by SCr KDIGO criteria, was identified culture (n = 1). This may have introduced bias into the study. As we
in 38% of patients overall with the majority being milder in severity excluded nonhematologic malignancies in our cohort, the generalizabil-
(KDIGO stage 1) and only four (7.1%) patients requiring RRT. We found ity to patients who develop TLS due to solid organ malignancy is lim-
severe AKI was associated with increased incidence of antihyperten- ited. Inaccuracies in weight measurements may have over- or under-
sive use, RRT, and PICU admission compared to those with no/mild AKI. estimated the %FO due to variability in the number of measurements,
There were no deaths in our population, but previous literature has methods, and scales used. The application of an imputed baseline SCr
shown AKI among pediatric oncology patients is associated with sig- may have also altered the AKI incidence and staging. Lastly, we were
nificantly higher mortality when compared to patients without AKI.21 unfortunately unable to use KDIGO AKI UO criteria due to inaccu-
There were no statistically significant differences in complication rates rate hourly UO reporting from the charts. Inclusion of UO-based AKI
including need for mechanical ventilation, hypoxemia, or pulmonary may have yielded a tighter relationship with FO and also altered the
edema, possibly due to a low number of severe AKI cases overall in our associations with outcomes and complications.13,31,32 However, this is
cohort.15,22,23 challenging in non-ICU populations without the regular use of Foley
We did not find FO to be strongly associated with AKI. This is in catheters for accurate hourly UO recording.
contrast to literature that demonstrates a positive correlation between All patients included in our cohort survived their induction treat-
AKI and FO.15,17,22,23 This could be because we did not include the ment; however, this study has clearly demonstrated a high incidence
KDIGO UO criteria to define and stage AKI, and/or the rise of SCr may of FO and a positive association between severe FO and morbidity,
have been blunted by hemodilution effects, thus leading to the under- regardless of AKI status. This is especially important as FO is a poten-
diagnosis of AKI. In the cardiac surgery population, Basu et al. revealed tially modifiable risk factor for outcomes including need for respi-
an increase in incidence and staging of AKI in infants after correcting ratory support and RRT. These findings highlight the importance of
SCr for the effects of FO and hemodilution.24 close monitoring of fluid balances with daily weights and early use of
In our cohort, 19.4% of patients with high-risk oncologic diagnoses diuretics to prevent severe FO. This is relevant as less than half of
developed TLS, similar to other pediatric studies reporting an inci- our patients (n = 25, 44.6%) were prescribed diuretics during their
dence of 19% and 32%, respectively.25–28 Differences in reported inci- TLS course. The metric of %FO has not been utilized on our oncol-
dences is likely related to the type and stage of malignancy, manage- ogy wards, but the results of this study have sparked a potential qual-
ment interventions, and timing of initial presentation. Furthermore, ity improvement initiative to include %FO on the daily ward flow
variability in patient characteristics in addition to the applied definition sheets. According to one recommendation, targeted IV fluid rates in
of TLS contributes to differences in reported incidences.29 Addition- patients with hyperuricemia are recommended to be between 2 and
ally, the requirement for dialysis and incidence of AKI has decreased 3 L/m2 /day (or 200 ml/kg/day if ≤10 kg) to obtain a UO range of 80–
since the introduction of rasburicase, which was used in 30% (n = 17) 100 ml/m2 /day (4–6 ml/kg/h if ≤10 kg).5 However, large variation in IV
of patients in our cohort. This was shown in an international pediatric fluid administration is observed amongst studies, including ours. Our
study where the incidence of AKI, need for dialysis, and TLS were sig- study suggests IV fluids should be used, but with more consideration of
nificantly lower among those who received rasburicase compared to earlier and more frequent use of diuretics to drive UO. Historic con-
those who received only allopurinol.30 cerns of diuretics promoting tubular acid deposition and thereby
8 of 9 FLOOD ET AL .

worsening AKI can be mitigated by avoiding the use of loop diuretics 12. Schwartz GJ, Work DF. Measurement and estimation of GFR in chil-
in hypovolemic patients.6,7 Accurate ins and outs, fluid balances, and dren and adolescents. Clin J Am Soc Nephrol. 2009;4(11):1832-1843.
13. Kaddourah A, Basu RK, Bagshaw SM, Goldstein SL. Epidemiology of
daily weights are suggested to help guide assessment and management
acute kidney injury in critically ill children and young adults. N Engl J
of FO in patients both with and at risk of TLS.7 Our goal is to develop Med. 2017;376(1):11-20.
evidence-based clinical care pathways for TLS management including 14. Hursh BE, Ronsley R, Islam N, Mammen C, Panagiotopoulos C. Acute
standardization of initial fluid rates and FO-based prompts for clini- kidney injury in children with type 1 diabetes hospitalized for diabetic
ketoacidosis. JAMA Pediatr. 2017;171(5):e170020.
cians to evaluate and modify management with fluid restriction and/or
15. Li Y, Wang J, Bai Z, et al. Early fluid overload is associated with acute
diuretics. However, prospective multicenter studies with larger sample kidney injury and PICU mortality in critically ill children. Eur J Pediatr.
sizes are still needed to further dissect the burden of FO and AKI in this 2015;175(1):39-48.
vulnerable population. 16. Soler YA, Nieves-Plaza M, Prieto M, Jesús RG-D, Suárez-Rivera M.
Pediatric risk, injury, failure, loss, end-stage renal disease score iden-
tifies acute kidney injury and predicts mortality in critically ill children.
ACKNOWLEDGMENTS
Pediatr Crit Care Med. 2013;14(4):e189-e195.
We thank Dr. Rachel Li for help with project conceptualization, Boris 17. Askenazi DJ, Koralkar R, Hundley HE, Montesanti A, Patil N,
Kuzeljevic for support with statistical analysis, and Alexandra Roine for Ambalavanan N. Fluid overload and mortality are associated with
data collection. acute kidney injury in sick near-term/term neonate. Pediatr Nephrol.
2012;28(4):661-666.
18. Selewski DT, Cornell TT, Lombel RM, et al. Weight-based determina-
CONFLICT OF INTEREST tion of fluid overload status and mortality in pediatric intensive care
The authors declare that there is no conflict of interest to disclose. unit patients requiring continuous renal replacement therapy. Intensive
Care Med. 2011;37(7):1166-1173.
DATA AVAILABILITY STATEMENT 19. Diaz F, Benfield M, Brown L, Hayes L. Fluid overload and outcomes
in critically ill children: a single center prospective cohort study. J Crit
Data are available on request from the authors.
Care. 2017;39:209-213.
20. Arikan AA, Zappitelli M, Goldstein SL, Naipaul A, Jefferson LS, Loftis
ORCID LL. Fluid overload is associated with impaired oxygenation and mor-
Kayla Flood https://orcid.org/0000-0002-1232-9474 bidity in critically ill children. Pediatr Crit Care Med. 2012;13(3):253-
Peter Skippen https://orcid.org/0000-0001-6716-0381 258.
21. Elsabbagh EM, Hashmat S. Acute kidney injury in childhood
Douglas G. Matsell https://orcid.org/0000-0003-4530-8554
cancer: an insight from Kids’ Inpatient Database. J Clin Oncol.
Cherry Mammen https://orcid.org/0000-0002-8389-8746 2019;37(15_suppl):e18379.
22. Vaewpanich J, Akcan-Arikan A, Coss-Bu JA, Kennedy CE, Starke JR,
REFERENCES Thammasitboon S. Fluid overload and kidney injury score as a predic-
tor for ventilator-associated events. Front Pediatr. 2019;7:204.
1. Russell TB, Kram DE. Tumor lysis syndrome. Pediatr Rev. 2020;
23. Wang N, Jiang L, Zhu B, Wen Y, Xi X-M, BAKIT Workgroup. Fluid
41(1):20-26.
balance and mortality in critically ill patients with acute kidney
2. Cairo MS, Bishop M. Tumour lysis syndrome: new therapeutic strate-
injury: a multicenter prospective epidemiological study. Crit Care.
gies and classification. Br J Haematol. 2004;127(1):3-11.
2015;19(1):371.
3. Howard SC, Jones DP, Pui C-H. The tumor lysis syndrome. New Engl J
24. Basu RK, Andrews A, Krawczeski C, Manning P, Wheeler DS, Goldstein
Med. 2018;379(11):1094.
SL. Acute kidney injury based on corrected serum creatinine is associ-
4. Wilson FP, Berns JS. Tumor lysis syndrome: new challenges and recent
ated with increased morbidity in children following the arterial switch
advances. Adv Chronic Kidney Dis. 2014;21(1):18-26.
operation. Pediatr Crit Care Med. 2013;14(5):e218-e224.
5. Coiffier B, Altman A, Pui C-H, Younes A, Cairo MS. Guidelines for the
25. Bagshi MA, Sadek AO, Hassan ES, Abbas AA. Tumor lysis syndrome in
management of pediatric and adult tumor lysis syndrome: an evidence-
children with acute leukemia: incidence and outcome. J Appl Hematol.
based review. J Clin Oncol. 2008;26(16):2767-2778.
2013;4(3):100-103.
6. Jones GL, Will A, Jackson GH, Webb NJA, Rule S, British Committee for
26. Micho H, Mohammed Y, Hailu D, Genet S. Evaluation and charac-
Standards in Haematology. Guidelines for the management of tumour
terization of tumor lysis syndrome before and after chemotherapy
lysis syndrome in adults and children with haematological malignan-
among pediatric oncology patients in Tikur Anbessa specialized hos-
cies on behalf of the British Committee for Standards in Haematology.
pital, Addis Ababa, Ethiopia. BMC Hematol. 2018;18(1):22.
Br J Haematol. 2015;169(5):661-671.
27. Galardy PJ, Hochberg J, Perkins SL, Harrison L, Goldman S, Cairo MS.
7. Jones DP, Mahmoud H, Chesney RW. Tumor lysis syndrome: patho-
Rasburicase in the prevention of laboratory/clinical tumour lysis syn-
genesis and management. Pediatr Nephrol. 1995;9(2):206-212.
drome in children with advanced mature B-NHL: a Children’s Oncol-
8. Wilson FP, Berns JS. Onco-nephrology: tumor lysis syndrome. Clin J
ogy Group report. Br J Haematol. 2013;163(3):365-372.
Am Soc Nephrol. 2012;7(10):1730-1739.
28. Wössmann W, Schrappe M, Meyer U, Zimmermann M, Reiter A.
9. El-Husseini A, Sabucedo A, Lamarche J, Courville C, Peguero A. Acute
Incidence of tumor lysis syndrome in children with advanced stage
kidney injury associated with tumor lysis syndrome: a paradigm shift.
Burkitt’s lymphoma/leukemia before and after introduction of prophy-
Am J Emerg Med. 2012;30(2):390.e3-6.
lactic use of urate oxidase. Ann Hematol. 2003;82(3):160-165.
10. Abu-Alfa AK, Younes A. Tumor lysis syndrome and acute kidney injury:
29. Howard SC, Pui C-H. Pitfalls in predicting tumor lysis syndrome. Leuk
evaluation, prevention, and management. Am J Kidney Dis. 2010;55(5
Lymphoma. 2009;47(5):782-785.
Suppl 3):S1-S13. Quiz S14-9.
30. Jeha S, Kantarjian H, Irwin D, et al. Efficacy and safety of rasburicase, a
11. KDIGO. KDIGO clinical practice guideline for acute kidney injury. Kid-
recombinant urate oxidase (Elitek), in the management of malignancy-
ney Int Suppl. 2012;2(1):1-141.
associated hyperuricemia in pediatric and adult patients: final results
FLOOD ET AL . 9 of 9

of a multicenter compassionate use trial. Leukemia. 2004;19(1):

34-38.
How to cite this article: Flood K, Rozmus J, Skippen P,
31. Jetton JG, Boohaker LJ, Sethi SK, et al. Incidence and outcomes
of neonatal acute kidney injury (AWAKEN): a multicentre, multi- et al. Fluid overload and acute kidney injury in children with
national, observational cohort study. Lancet Child Adolesc Health. tumor lysis syndrome. Pediatr Blood Cancer. 2021;68:e29255
2017;1(3):184-194. https://doi.org/10.1002/pbc.29255.
32. Goldstein SL. Urine output assessment in acute kidney injury: the
cheapest and most impactful biomarker. Front Pediatr. 2020;7:565.