Pediatric cardiac surgery associated-acute kidney injury: Interleukin 18 as a new biomarker

Abstract

BACKGROUND

Creatinine has been the widely accepted and used marker of cardiac surgery associated (CSA) acute kidney injury (AKI) in both adult and pediatric age groups, the diagnosis of which may take hours to days, leading to delayed recognition of renal dysfunction, and this may be partly responsible for the limited progress in preventing and treating postoperative AKI. This underscores the need for more sensitive markers of AKI.

AIM

To assess the detection of AKI by interleukin 18 (IL-18) in contrast to creatinine among post-operative CSA children in Nigeria.

METHODS

Plasma samples were collected from forty consecutive children aged less than 15 years who had open heart surgery for treatment of congenital heart disease at six different time points (0-, 4-, 8-, 12-, 24-, 48 hours) and were assayed for IL-18 and creatinine using enzyme linked immuno-sorbent assay method and automated Jaffe kinetic method respectively.

RESULTS

The average plasma IL-18 concentrations at six different time points were 12.83 ng/L, 19.09 ng/L, 22.15 ng/L, 20.00 ng/L, 18.11 ng/L and 16.36 ng/L correspondingly with the greatest value at 8 hours. Receiver operating characteristic curve showed that IL-18 portend CSA-AKI with a specificity of 88% and a sensitivity of 100% at a limit point of 22 ng/L and had an area under the curve of 0.926.

CONCLUSION

IL-18 had a better specificity and sensitivity in the detection of CSA-AKI in pediatric age group than creatinine from the findings in this study.

Key Words: Interleukin 18; Creatinine; Pediatric; Acute kidney injury; Cardiac surgery; Receiver operating characteristic curve

Core Tip: In the present study, we have shown that interleukin 18 had an earlier rise than the traditional creatinine in those patients who developed cardiac surgery associated acute kidney injury after open heart surgery, and that it is also a better diagnostic marker with its superior area under the curve value on receiver operating characteristic curve than creatinine, thus, our findings underscore the importance of using this biomarker in the diagnosis of cardiac surgery associated acute kidney injury for earlier diagnosis and improved patient outcome.

INTRODUCTION

Acute kidney injury (AKI) can be clinically defined as an abrupt increase in serum creatinine over hours to days leading to retention of nitrogenous and metabolic waste products[1]. These descriptions depend on the total or comparative serum creatinine values, or the variations in concentration from different baseline values[2,3]. The term AKI was recently introduced to replace the older term of acute renal failure or acute kidney failure[4].

Congenital heart disease accounts for about 30%-40% of all congenital defects worldwide[5], affects about 1% of newborns and it is a significant cause of morbidity and mortality[6-10], and the commonest cause of infant death from birth defect[11], clinical presentation is often late with only 58.1% presenting in infancy[6], 43% having a diagnosis before 5 years[4], and 6.9% progressing to definitive intervention[12].

Alterations in renal function are widespread in serious illness and need quick intervention; though, timely management has been hampered by deficient descriptions of kidney dysfunction and the absence of reliable early biomarkers of renal injury[13]. Following cardiac surgery, multiple factors are responsible for the development of AKI and include pre-operative, intra-operative and post-operative factors, leading to tubular injury whose severity is seen as a rise in serum creatinine and reduced urine output[14].

Cardiac surgery associated (CSA)-AKI begins with alterations in vaso-reactivity and renal perfusion causing cellular adenosine triphosphate depletion and oxidative injury, this leads to activation of pro-inflammatory endothelial cells and renal epithelial cells that adhere to the endothelium of the peritubular capillaries of the outer medulla, causing medullary congestion that further worsens the hypoxic injury to the proximal tubules[14]. The increase in serum creatinine and blood urea nitrogen with a decrease in urine output have largely been used as markers of decreased glomerular filtration rate which defines AKI, though they are insensitive markers of glomerular filtration rate because of their affectation by nutrition, high protein diet, muscle mass, age, sex, gastrointestinal bleeding, corticosteroid therapy and aggressive fluid resuscitation[15]. Serum creatinine is not a sensitive postoperative AKI marker because its rise is delayed for about 1-3 days after cardiac surgery[16].

Interleukin 18 (IL-18) being one of the pro-inflammatory ILs produced from the intercalated units of the collecting duct, connecting tubule as well as the late distal convoluted tubule of normal renal organs that is a mediator of inflammation measurable in the urine after being induced in the proximal tubule as an early biomarker of AKI, rising about 6 hours after initiation of cardiopulmonary bypass (CPB) in those subsequently diagnosed with AKI[17]. Study in children who developed AKI after CPB showed an increase in IL-18 after 4-6 hours, peaking at 12 hours with a 25 fold increase and remaining elevated for up to 48 hours[17]. The performance of IL-18 in the diagnosis of AKI after CPB using area under the receiver operating characteristic (ROC) curve at 4 hours, 12 hours and 24 hours was 61%, 75%, and 73% respectively[17].

MATERIALS AND METHODS

This study was conducted in a tertiary hospital that carries out open heart surgery for the treatment of congenital heart diseases for patients across the country. It was a prospective longitudinal study with the aim of evaluating the usefulness of IL-18 in the diagnosis of CSA-AKI.

Ethical approval for this study was gotten from the hospital Ethical Committee (ERC/2020/03/16). The inclusion criteria for the study participants were age less than 15 years, diagnosed with congenital heart diseases, a normal C-reactive protein value, while those with chronic kidney diseases were excluded, forty consecutive patients were enrolled.

Informed written consent was obtained from the parents. Anthropometric parameters (height and weight in centimeter and kilogram) were obtained from study participants’ case files. The haemodynamics parameters of the study participants were optimized before surgery and a normal C-reactive protein value was a requirement, they also had invasive and non-invasive monitoring using multi-parameters monitors peri-operation and post operation. The CPB time ranged from 50-64 minutes for those with atrial and ventricular septal defects, 66-142 minutes for those with tetralogy of Fallot, and weaning off mechanical ventilation occurred from immediate post operation to 6 hours post operation in those with atrial and ventricular septal defect and 12-24 hours in those with tetralogy of Fallot. Two milliliters (2 mL) of blood, was collected from the central venous line of each participant, over six time points measurements for IL-18 and creatinine assays. The time points were baseline zero hour (pre-operative blood collected immediately after passing the central line in the operating room), 4 hours, 8 hours, 12 hours, 24 hours and 48 hours post operation. Collected blood specimens were discharged into heparinized bottles and centrifuged at 1500 g for 10 minutes. Subsequently, separated supernatant were harvested by Pasteur’s pipette and dispensed into cryovials for storage at -70 °C. The aliquoted plasma were thawed at room temperature immediately prior to analysis. Determination of plasma concentrations of creatinine and IL-18 were done using an auto-analyzer, Cobas c311 and enzyme linked immunosorbent assay using BT Lab Human IL-18 enzyme linked immunosorbent assay kits respectively. Participants were stratified according to the Kidney Disease: Improving Global Outcomes criteria using a 50% rise in creatinine in 48 hours above baseline into those with and without CSA-AKI.

Statistical analysis

Data was analyzed by IBM SPSS version 24.0. Comparison of the sensitivity and specificity of the different biomarkers in predicting CSA-AKI was done statistically by the ROC curve, P ≤ 0.05 was statistically significant and an area under the curve (AUC)-ROC of 0.900 or greater was considered to represent an excellent biomarker.

RESULTS

The anthropometric parameters of the study participants are presented in Table 1. Mean plasma IL-18 concentrations at 0 hour, 4 hours, 8 hours, 12 hours, 24 hours, and 48 hours were 12.83 ± 3.04 ng/L, 19.09 ± 4.07 ng/L, 22.15 ± 5.86 ng/L, 20.00 ± 5.38 ng/L, 18.11 ± 5.13 ng/L, and 16.36 ± 4.61 ng/L, respectively, with a peak concentration observed at 8 hours. Corresponding creatinine levels were 48.98 ± 11.6 μmol/L, 59.65 ± 13.06 μmol/L, 63.00 ± 16.53 μmol/L, 64.90 ± 17.65 μmol/L, 68.50 ± 19.99 μmol/L, and 70.78 ± 21.86 μmol/L, peaking at 48 hours (Table 2). The mean IL-18 and creatinine levels at different time points in participants with and without CSA-AKI are compared in Tables 3 and 4. Figure 1 shows the ROC curves of IL-18 and creatinine at the six time points; no statistically significant difference was observed at 0 hour, whereas significant differences were seen at the other time points. The statistical significance of the AUC for IL-18 and creatinine is summarized in Table 5. ROC curve analysis demonstrated that IL-18 predicted CSA-AKI with a sensitivity of 100% and a specificity of 88% at a cut-off level of 22 ng/L, with an AUC of 0.926 (Figure 2).

Figure 1 Receiver operating characteristic curve of interleukin-18 and creatinine. A: At 0 hours; B: At 4 hours; C: At 8 hours; D: At 12 hours; E: At 24 hours; F: At 48 hours. ^aP ≤ 0.05 significant. ROC: Receiver operating characteristic; AUC: Area under the curve; IL: Interleukin; CR: Creatinine.

Figure 2 Receiver operating characteristic curve of mean values of interleukin 18 and creatinine. IL: Interleukin; CR: Creatinine; ROC: Receiver operating characteristic.

Table 1 Anthropometric parameters of study participants.

Variables	n (%)/mean ± SD
Male	22 (55)
Female	18 (45)
Weight (kg)	17.66 ± 9.05
Height (m)	0.99 ± 0.31
BMI (kg/m2)	14.74 ± 3.12

BMI: Body mass index.

Table 2 Mean concentration of measured biomarkers over time, mean ± SD.

	Time (hour)
Biomarkers	0:00	4:00	8:00	12:00	24:00	48:00
IL-18 (ng/L)	12.83 ± 3.04	19.09 ± 4.97	22.15 ± 5.86	20.00 ± 5.38	18.11 ± 5.13	16.36 ± 4.61
Creatinine (μmol/L)	48.98 ± 11.6	59.65 ± 13.06	63.00 ± 16.53	64.90 ± 17.65	68.50 ± 19.99	70.78 ± 21.86

IL: Interleukin.

Table 3 Comparison of interleukin 18 at different time points in those with and without cardiac surgery associated acute kidney injury.

Renal biomarker at time points (hours)	Cardiac surgery associated acute kidney injury		P value
	Present (n = 7)	Absent (n = 33)
0	13.36 ± 4.05	12.72 ± 5.27	0.727
4	26.73 ± 3.44	17.47 ± 6.45	0.001a
8	40.02 ± 6.52	18.35 ± 7.16	0.000a
12	29.61 ± 5.37	17.96 ± 6.06	0.002a
24	24.68 ± 4.50	16.72 ± 5.54	0.000a
48	22.98 ± 3.27	14.95 ± 4.98	0.000a

^aP ≤ 0.05 significant.

Table 4 Comparison of creatinine at different time points in those with or without cardiac surgery associated acute kidney injury.

Renal biomarker at time points (hours)	Cardiac surgery associated acute kidney injury		P value
	Present (n = 7)	Absent (n = 33)
0	46.57 ± 9.54	50.09 ± 12.03	0.579
4	55.57 ± 13.21	60.52 ± 14.27	0.511
8	68.00 ± 50.02	60.15 ± 19.43	0.473
12	64.29 ± 47.37	65.21 ± 14.30	0.349
24	96.14 ± 85.36	64.52 ± 28.47	0.277
48	95.29 ± 64.89	65.58 ± 29.86	0.066

Table 5 Statistical significance of area under the curve of measured biomarkers.

Biomarkers	AUC	SE	P value
IL-18	0.926	0.043	0.000a
CR	0.429	0.148	0.557

^aP < 0.05 significant.

AUC: Area under the curve; IL: Interleukin; CR: Creatinine.

DISCUSSION

In this study, mean plasma IL-18 levels increased at 4 hours post-operation compared to baseline (0 hour), consistent with another study showing that urinary IL-18 levels in patients with AKI after CPB rose within 4-6 hours, peaked at 12 hours, and remained elevated up to 48 hours post-surgery. However, the rise in urinary IL-18 in AKI occurs more slowly than the increase observed in urinary neutrophil gelatinase associated lipocalin[18].

In this study, the mean peak plasma level of IL-18 was reached at 8 hours post-operation, whereas a study in Poland reported that urinary IL-18 levels increased 4-6 hours after cardiac surgery with CPB, peaked at 12 hours, and returned to normal by 48 hours[19], and this difference can be attributed to the different sample matrixes used in the studies. In this study, a steady decline in mean plasma IL-18 levels was observed from 12 hours to 48 hours postoperatively, in contrast to findings from another study in which urinary IL-18 increased within 4-6 hours after CPB, peaked at more than a 25-fold increase at 12 hours, and remained markedly elevated up to 48 hours post-CPB[20] because secretion of IL-18 into the urine occurs earlier in the urine than into the plasma with slower clearance rate from the urine. In one study, the diagnostic performance of IL-18, as assessed by the area under the ROC curve, was 61%, 75%, and 73% at 4 hours, 12 hours, and 24 hours after CPB, respectively[21] in which urine samples were used after pediatric cardiac surgery because urinary IL-18 peaks at 12 hours thereby giving a corresponding highest AUC at 12 hours unlike what was observed in this study where plasma IL-18 was assayed and the peak was at 8 hours.

Studies have demonstrated that urinary IL-18 is an early predictive biomarker in children undergoing CPB, with levels rising at 6 hours post-CPB and an AUC of 0.750 for AKI prediction. Subsequent pediatric studies have further confirmed that urinary IL-18 measured between 6 hours and 12 hours post-CPB provides moderate predictive accuracy for AKI, with reported AUC values ranging from 0.720 to 0.820[22] due to the rate of secretion of IL-18 into the tubules following AKI. It was observed from the statistically significant findings in the comparison of the means at different sampling times in this study that the ideal sampling of plasma IL-18 for the prediction of CSA-AKI can be done between 4 hours and 8 hours after open heart surgery in children with congenital heart diseases.

The mean values of IL-18 assay in this study had an AUC of 0.927 (at 95% confidence interval) from the ROC curve with 100% sensitivity and 97% specificity at a cut off value of 22 ng/L as compared to an AUC of 0.792 (95% confidence interval: 0.685-0.898)[23], which is different from the AUC of assayed IL-18 gotten from a study of critically ill adult patients which is 0.589 (95% confidence interval)[24] and this reduced diagnostic performance is attributable to the difference in the study population of children with congenital heart disease who had open heart surgery vs critically ill adult patients with varied background renal state and co-morbidities.

A systematic review of studies following cardiac surgery reported that urinary IL-18 demonstrated a sensitivity of 58% and a specificity of 75% for the determination of AKI, with the area under the ROC curve indicating moderate prognostic performance was 0.700 which was similar to what was observed in another meta-analysis, in which the area under the ROC curve was 0.770 and concluded that urinary IL-18 is a biomarker of AKI with moderate diagnostic value[18]. Furthermore, a meta-analysis of previously conducted research in adults with the search engine using results with non-specific sample matrixes like urine, serum and plasma in adult patients showed that IL-18 has 58% sensitivity and 75% specificity in detecting CSA-AKI[20].

CONCLUSION

In the present study, we have shown that IL-18 had an earlier rise than the traditional creatinine in those patients who developed CSA-AKI after open heart surgery, and that it is also a better diagnostic marker with its superior AUC value on ROC curve than creatinine, thus, our findings underscore the importance of using this biomarker in the diagnosis of CSA-AKI for earlier diagnosis and improved patient outcome.