Background
Acute kidney injury (AKI) is an important complication in severe malaria that is associated with increased mortality in both paediatric [
1‐
6] and adult populations [
7‐
10]. However, until recently, the incidence of AKI in paediatric malaria was under-appreciated as few children develop signs of overt renal failure and creatinine testing is not routinely available in resource-constrained settings. Despite significant progress in reducing malaria incidence and mortality [
11], case fatality rates in severe disease remain high, and the identification of an effective adjunctive therapy is a research priority.
Decreased bioavailable nitric oxide is a common feature in both children and adults with severe malaria, and clinical trials to increase nitric oxide (NO) have been conducted [
12‐
14]. Although there has been no conclusive benefit demonstrated in human trials designed to increase bioavailable NO, infusion of
l-arginine in adults with severe malaria improved endothelial recovery [
12], and iNO delivered at 80 ppm was associated with reduced risk of fine motor impairment in children under 5 years of age at 6 month follow up [
15].
Despite early reports that inhaled NO (iNO) was associated with increased splanchnic and renal blood flow [
16], a meta-analysis of adults with acute respiratory distress syndrome found that iNO treatment was associated with a 50% increased risk of developing AKI (relative risk, 95% CI 1.50, 1.11–2.02) [
17]. This has been confirmed in another meta-analysis including non-ARDS patients [
18], but the effect was strongest in patients with ARDS with prolonged exposure and a high cumulative dose. There are limited reports evaluating iNO and renal safety in children. Previously, an increase in the overall incidence of AKI (relative risk, 95% CI 1.36, 1.03–1.90, p = 0.026) was reported in children treated with iNO [
6]. However, there was no association between iNO and AKI when restricting the analysis to children who developed AKI after treatment was initiated.
Chitinase-3-like 1 protein (CHI3L1) is a 39 kDa secreted glycoprotein produced by a variety of cell types in response to inflammation, including activated macrophages, neutrophils, and fibroblasts. CHI3L1 is highly expressed in healthy kidney tissue [
19] and is freely filtered by the glomerulus. CHI3L1 is also secreted by activated macrophages in the kidney upon stress or damage [
20]. Elevated levels of CHI3L1 have been reported in Ugandan children with severe malaria, and further elevated in fatal malaria [
21]. Co-culture of human peripheral blood mononuclear cells with
Plasmodium falciparum-infected erythrocytes in vitro induced CHI3L1 transcription and secretion of CHI3L1 protein [
21]. While CHI3L1 was elevated by day 5 infection in an experimental model of cerebral malaria, genetic disruption of
Chi3l1 did not affect inflammatory responses or outcome [
21].
In this secondary analysis of an iNO intervention trial, CHI3L1 was investigated as a biomarker of morbidity and mortality in paediatric severe malaria, the longitudinal kinetics of CHI3L1 were explored in children hospitalized with severe malaria, and the impact of iNO therapy on CHI3L1 normalization was evaluated.
Discussion
Acute kidney injury has recently become recognized as a common complication of paediatric severe malaria, but its pathogenesis is not well understood. In this study, elevated CHI3L1 levels at admission were associated with the severity of AKI. Using linear regression we explore the relationship between CHI3L1 levels and biomarkers of kidney function and found that CHI3L1 was significantly associated with increased Cystatin C following correction for age, sex and creatinine. CHI3L1 levels were associated with an increased risk of both in-hospital and long-term mortality independent of treatment arm and AKI. Importantly, using LME models to evaluate the longitudinal course of CHI3L1, administration of iNO was associated with prolonged elevation of CHI3L1 levels over the first 4 days of hospitalization. Further, there was a strong relationship between CHI3L1 levels and the severity of AKI that was independent of iNO treatment. These results suggest that CHI3L1 is an important biomarker of disease severity and mortality in paediatric severe malaria that is associated with kidney dysfunction as well as endothelial activation, inflammation and haemolysis. Additional studies are required to validate the relationship between CHI3L1 and AKI in severe malaria and investigate the association between CHI3L1 and established biomarkers of AKI.
CHI3L1 is a relatively new biomarker of AKI or altered renal function that has been investigated in the context of critical illness [
32,
33], and sepsis [
34]. However, CHI3L1 has been predominantly investigated as a biomarker of AKI in urine rather than blood. CHI3L1 was identified in a urine proteomic screen in mice with kidney ischaemic reperfusion injury where a direct correlation was observed between the severity of kidney injury, CHI3L1 expression in the kidney, and levels in the urine [
20]. Studies of urinary CHI3L1 from donor kidneys suggested that CHI3L1 is a repair phase protein produced in response to tubular injury, and associated with recovery from AKI and delayed graft function [
35]. CHI3L1 expression has been reported on the surface of tubular epithelial cells, consistent with either uptake of filtered CHI3L1 or tubular cell secretion, and urinary CHI3L1 levels correlate with the severity of acute tubular necrosis [
35].
Studies in mice have shown that CHI3L1 plays a critical role in tissue repair and remodelling following pulmonary insult by limiting oxidative damage, stimulating alternative (M2) macrophage activation, and inhibiting apoptosis [
36‐
38]. In the context of malaria, an increase of M2 monocytes in peripheral blood has been reported in children with severe malaria compared to healthy controls [
39]. Further, M2 monocytes were associated with increased expression of arginase 1, lower NOS2 mRNA, and lower plasma arginine [
39]. Additional studies are needed to delineate the role between CHI3L1, alternative macrophage activation in severe malaria and its relationship with AKI and NO bioavailability.
There are limited, and conflicting, data related to the relationship between CHI3L1 and NO. CHI3L1 has been positively correlated with NO levels in tissue culture supernatant from herniated lumbar discs [
40], and exhaled NO in children with severe therapy-resistant asthma [
41]. However, in patients with type 1 diabetes reduced NO in the blood correlated with elevated CHI3L1 [
42], and plasma CHI3L1 was negatively associated with nitric-oxide mediated vasodilatory capacity in adults with obstructive sleep apnea [
43]. In this study there was delayed CHI3L1 recovery in children receiving iNO independent of AKI-associated changes in CHI3L1, with CHI3L1 the only biomarker identified that has been shown to differ in response to iNO therapy [
27]. While there were differences in the rate of CHI3L1 recovery over the first 4 days of hospitalization associated with iNO therapy, the effect was not significant by day 14 suggesting that iNO results in a transient delay in the normalization of CHI3L1 levels.
Acute kidney injury is a common complication in severe malaria but its pathogenesis is unclear. Peripheral parasitaemia is not associated with worsening kidney function in children or adults with severe malaria [
6,
10], but plasma HRP2 levels (reflecting total parasite biomass) are associated with worsening renal function, suggesting AKI is associated with the sequestered parasite biomass [
10]. Plasma suPAR—a marker of mononuclear cell activation—was elevated in adults with malaria-associated AKI [
10]. These results are consistent with post-mortem studies showing parasite sequestration and mononuclear cell infiltration in glomerular and peritubular capillaries [
44]. Neutrophil gelatinase-associated lipocalin (NGAL), an early marker of renal tubular damage, was elevated in adults with malaria-associated AKI [
10]. Oxidative stress and injury due to cell-free haemoglobin and haem from malaria-induced haemolysis may also contribute to tubular damage in malaria. In children with severe malaria, an increase in the haem to haemopexin ratio was observed over hospitalization in children with severe AKI (Stage 3 AKI) [
26]. In adults with severe malaria, reduced red blood cell deformability, and increased cell-free haemoglobin and lipid peroxidation (indicative of oxidative stress) were associated with AKI [
45].
While the mechanisms leading to severe malaria-associated AKI are not well understood, it is likely a heterogeneous syndrome characterized by reduced renal blood flow due to dehydration, hypotension, and impaired microcirculatory function as a result of parasite sequestration and reduced bioavailable nitric oxide. A subset of children may be at risk of developing direct tubular damage as a result of prolonged ischaemia, endothelial activation, inflammation, and cell-free haemoglobin and haem-mediated injury. In this cohort, an increase in CHI3L1 was associated with significant increases in creatinine, Cystatin C, and BUN by linear regression, and the relationship between CHI3L1 and Cystatin C remained significant following adjustment for creatinine. CHI3L1 levels also correlated with markers of immune activation (CRP, sTREM-1, CXCL10/IP-10), endothelial activation (Ang-2, sICAM-1), and haemolysis (LDH, haem, haemopexin), pathways of injury that are well described in paediatric severe malaria [
26‐
28,
46‐
48]. As CHI3L1 is produced by tubular cells in response to injury and remodelling [
35], it may represent a novel biomarker of AKI in pediatric severe malaria. Additional studies are needed to delineate between CHI3L1 as a marker of inflammation versus AKI by comparing CHI3L1 levels to other established biomarkers of kidney injury (e.g. NGAL) that are well characterized in association with changes to kidney function. Further, additional research is needed to evaluate CHI3L1 over time as it relates to renal recovery and repair in both plasma and urine.
In this study elevated CHI3L1 levels at admission were a risk factor for in-hospital and all-cause 6 month mortality independent of kidney function and treatment group. These results are consistent with reports from adults where CHI3L1 is an independent predictor of all-cause mortality in type II diabetes [
49,
50], heart failure [
51], and sepsis [
52]. Although CHI3L1 was strongly associated with AKI, which is an established risk factor for mortality in severe malaria, the relationship between CHI3L1 and increased risk of death was independent of AKI status suggesting CHI3L1 is not simply a biomarker of kidney function.
This study has several strengths including a randomized trial design with detailed clinical follow up and daily assessment of renal function and plasma CHI3L1 levels. Further, the majority of children were followed up to 6 months allowing us to evaluate the association between CHI3L1 and post-discharge mortality in children with severe malaria. Limitations include a lack of data on renal recovery and long-term renal function in the children. Further, urine was not collected to evaluate plasma versus urine levels of CHI3L1.
Authors’ contributions
ALC participated in the study design, collection of human data, laboratory testing, statistical analysis and wrote the first draft of the manuscript. MTH participated in the study design, collection of human data, and statistical analysis. RE participated in the laboratory testing, statistical analysis and was a major contributor in writing the manuscript. ROO, SN, CM, CCJ and KCK participated in the study design, and collection of human data. All authors read and approved the final manuscript.