Factors limiting usefulness of serum and urinary NGAL as a marker of acute kidney injury in preterm newborns

Abstract

Background: Neutrophil gelatinase-associated lipocalin (NGAL) is postulated to be a highly sensitive and specific marker of acute kidney injury (AKI). The aim of this study was to assess the factors affecting serum and urine total NGAL in preterm newborns, limiting the role of this new potential marker of AKI. Methods: Serum and urinary total NGAL concentrations were determined in 57 preterm infants admitted to the Neonatal Intensive Care Unit in the following points of time: first week of life, between 8 and 14 days of life, and after the fourth week of life. Patients’ clinical conditions were evaluated based on NTISS (Neonatal Therapeutic Intervention Scoring System). Two gestational age subgroups were distinguished: ≤29 and 30 to 35 weeks of gestation. We sought correlation between total NGAL values and gestational age, birth weight, Apgar score and severity of clinical condition, with particular interest in inflammatory status. Results: Serum and urinary total NGAL concentration correlated with inflammatory markers, such as CRP and procalcitonin, as well as with NTISS values. Birth weight and gestational age influence urinary NGAL (uNGAL) values in the first two weeks of life. In AKI (N = 8) patients uNGAL values were significantly higher than in non-AKI newborns. Conclusions: We conclude that inflammatory status and prematurity limits the specificity of total NGAL measurement as a marker of AKI.

• Acute kidney injury
• creatinine
• kidney function
• NGAL
• perinatal asphyxia
• preterm newborns
• sepsis

Introduction

Neutrophil gelatinase-associated lipocalin (NGAL), known as lipocalin 2, is a siderophore with bacteriostatic features, expressed by multiple human cells, including epithelial cells and neutrophiles.1–3 The major biological role of this 24 kDa protein is iron-chelating, preventing its uptake by microorganisms.1 NGAL has short half-life, estimated at 10–20 min due to rapid elimination from circulation by the kidney.3 It is postulated that increased serum NGAL (sNGAL) and urinary NGAL (uNGAL) is a sensitive marker of acute kidney injury (AKI), predominantly of tubular cell damage, not related to glomerular filtration decline.4^,5 Additionally, it was shown that NGAL is a marker of neutrophil activation and the severity of sepsis in populations other than newborns.6^,7

Few studies have analyzed total sNGAL and uNGAL levels in preterm newborns. Huynh et al.8 revealed that uNGAL levels are declining with gestational age (GA) by 17.8% by each week in very low birth weight infants (VLBW) with birth weight (BW) ranges from 790 to 1490 g. Similar results were obtained by Lavery et al.9 who showed uNGAL to be inversely related to BW and GA in 20 infants with BW range from 500 to 1500 g. Contradictory data has also been published. Inoue et al.10 reported a strong positive correlation between GA and sNGAL in 52 infants (mostly preterm) at birth, admitted to a tertiary neonatal intensive care unit (NICU). The authors demonstrated that sNGAL is related to neutrophil count, and predicts the development of bronchopulmonary dysplasia (BPD).10 Additionally, a known factor that increases both sNGAL and uNGAL levels is asphyxia.11 Congenital obstructive urinary tract anomalies account for an increase in uNGAL.12

Sparse, and partially contradictory, data justifies performance of a study that aims to assess the factors affecting total sNGAL and uNGAL in preterm newborns, and also limits the role of this new potential marker of AKI.

Materials and methods

The study was conducted on a group of 57 neonates in the first month of life, who were all out-born and transferred to the NICU from other hospitals. Prematurity, respiratory insufficiency or suspicion of early sepsis were all the reasons for admission. Based on gestational age, the study group was divided into two subgroups: the first, GA ≤29 weeks, and the second, GA between 30 and 35 weeks. GA was calculated based on the Naegele formula and confirmed by clinical evaluation using the Ballard score. Cases of intrauterine growth restriction (IUGR) were also diagnosed. Clinical condition was estimated by means of Neonatal Therapeutic Intervention Scoring System (NTISS). Demographic data and clinical evaluation of the study group are presented in Table 1. There were both uncomplicated and complicated clinical courses in the study group, and the majority of the patients had risk factors for AKI.

Table 1. Characteristics of all premature infants and distinguished subgroups according to the grade of maturity (means with 95% confidence intervals).

	All [N = 57]	≤29 hbd [N = 25]	≥30 hbd [N = 32]	Statistical significance
Gender (M/F)	35/22	18/7	17/15	0.15
Gestational age (weeks)	30 (29–31)	27 (26–28)	33 (32–34)	<0.001
Birth weight (g)	1536 (1365–1708)	1029 (895–1163)	1932 (1735–2131)	<0.001
Apgar 5′ (pts)	6.3 (5.7–7.0)	4.8 (4.0–5.5)	7.5 (6.7–8.3)	<0.001
Day of life	5 (3–8)	4 (2–5)	7 (3–11)	0.19
NTISS 1st week (pts)	20 (19–23)	24 (21–26)	18 (15–22)	0.02
NTISS 2nd week (pts)	16 (14–19)	21 (18–23)	13 (9–16)	<0.001
NTISS 4th week (pts)	9 (6–12)	9 (6–14)	9 (4–13)	0.72
WBC 1st week (10^3/μL)	14.9 (12.3–17.5)	17.6 (13.4–21.8)	12.6 (9.4–15.7)	0.05
WBC 2nd week (10^3/μL)	14.5 (12.9–16.2)	16.7 (15.5–19.9)	12.7 (11.3–14.2)	0.02
WBC 4th week (10^3/μL)	11.3 (10.0–12.5)	10.6 (8.3–12.9)	11.7 (10.2–13.3)	0.38
PLT 1st week (10^3/μL)	181 (154–209)	152 (115–189)	206 (166–247)	0.05
PLT 2nd week (10^3/μL)	289 (245–333)	227 (169–285)	340 (278–402)	0.01
PLT 4th week (10^3/μL)	344 (292–396)	278 (222–334)	394 (316–474)	0.02
CRP 1st week (mg/dL)	9.4 (3.5–15.3)	8.7 (0.3–17.6)	10.7 (0.5–21.0)	0.79
CRP 2nd week (mg/dL)	16.0 (4.5–27.5)	14.8 (2.7–26.9)	16.8 (0.5–34.6)	0.87
CRP 4th week (mg/dL)	10.5 (1.0–19.9)	5.6 (1.6–9.5)	14.4 (0.1–31.3)	0.36
PCT 1st week (ng/mL)	10.6 (4.6–16.6)	10.8 (1.6–20.1)	10.3 (2.1–18.6)	0.93
PCT 2nd week (ng/mL)	3.4 (0.1–7.7)	2.4 (0.1–6.1)	4.1 (0.2–0.9)	0.71
PCT 4th week (ng/mL)	1.1 (0–2.6)	0.4 (0.1–0.6)	1.5 (0–3.9)	0.44
Urea 1st week (mg/dL)	40.8 (34.5–47.1)	53.0 (42.5–63.4)	31.4 (25.2–37.5)	0.003
Urea 2nd week (mg/dL)	24.2 (19.4–28.9)	30.2 (20.7–39.5)	19.6 (15.6–23.6)	0.02
Urea 4th week (mg/dL)	19.3 (14.6–29.3)	15.6 (10.7–20.6)	21.9 (14.5–29.3)	0.19

Note: M: male; F: female; NTISS: Neonatal Therapeutic Intervention Scoring System; WBC: white blood cell; PLT: platelet count; CRP: C-reactive protein; PCT: procalcitonin.

Inclusion criteria consisted of both GA equal or lower than 34 + 6 weeks and signed formal consent by parents or legal guardians. Patients with malformations of the urinary system were not included in the observation. The study protocol was approved by the Local Bioethics Committee (KNW/0022/KB1/120/11) and performed in a single third level nursery (NICU) of the university children’s hospital. The study protocol did not interfere with the basic diagnostic and treatment algorithms commonly accepted in the unit. All the patients included were treated within the standards of care and the preservation of serum and urinary samples for further biochemical analysis was the only deviation.

On admission, all study patients underwent a basic septic screen, including complete blood count (CBC) with blood smear, C-reactive protein (CRP), procalcitonin (PCT) and blood cultures.

Routine samples were collected for microbiological evaluation of colonization status. The remaining tests included glucose, electrolytes, serum creatinine (sCr), total serum protein, serum albumin, bilirubin, and also urinalysis and urine cultures. The standard protocol included routine daily or more often if required, arterial blood gases (ABG). CBC and biochemical screening (glucose, sCr, aminotranspherases, CRP and PCT) were performed on the following days according to the clinical requirements. Serum and urine samples were collected after the routine tests had been completed, and frozen to −70 °C in polypropylene tubes.

Additionally, samples of serum and urine were used to evaluate s/u NGAL (monomeric and dimeric) and serum cystatin C (CystC). The first day of sampling was during the first week of life (as soon as clinical condition allow for blood sampling, which was particularly important in extremely low birth weight (ELBW) babies), possibly within the first three days after birth. The second evaluation was made by the end of the second week of life, and the last one, when the patient completed the first month of life.

Biophysical parameters and interventions were noted on daily charts with particular interest in the measurements of blood pressure, urine output (UO), fluid input, and amount of supplied protein (both intravenous and enteral). The final evaluation included end-point result (survival vs. death) and the presence of morbidities typical for preterm babies such as respiratory distress syndrome (RDS), bronchopulmonary dysplasia, necrotizing enterocolitis (NEC), persistent arterial duct (PDA), retinopathy of prematurity (ROP), anemia of prematurity (AOP), intraventricular hemorrhage (IVH), and periventricular leukomalacia (PVL).

Laboratory measurements

Serum and urinary total NGAL measurements were performed in duplicate using commercially available ELISA kits (BioPorto Diagnostics, Gentofte, Denmark) with intra- and interassay coefficients of variation of 3.6 and 7.9%, respectively (the sensitivity below 0.01 ng/mL).

Cystatin C was assessed by ELISA (R&D Systems, Minneapolis, MN), with intra- and inter-assay coefficients of variability <5.9%.

Blood count (including white blood cells, WBC) and biochemical parameters were evaluated using routine laboratory methods and measurements.

Data analysis

Clinical condition was evaluated at all three time points using NTISS.

Sepsis was defined as Systemic Inflammatory Response Syndrome (SIRS) and evidence of infection (positive microbiological culture, clinical symptoms). Severe sepsis was identified, when the course of sepsis was complicated by dysfunction of two or more organs or systems. Early onset sepsis (EOS) was recognized in the first week of life, while late onset sepsis (LOS) was diagnosed after the first week of life.

Estimated glomerular filtration rate (eGFR) was calculated according to the Schwartz formula. The estimates were not used for analysis of GFR changes as they were frequently calculated in unstable conditions in the first days of observation. Only sCr values at discharge fulfilled steady-state conditions for GFR estimation.

AKI was defined as persistently increased serum creatinine (≥1.5 mg/dL) for at least 24 h or rising values >0.3 mg/dL from the baseline [13].

Statistical analysis

Analyses were performed using the STATISTICA 10.0 (StatSoft Polska, Kraków, Poland) software. Normality of the distribution was tested with the Kolmogorov–Smirnov test. The data presented were expressed as means with 95% confidence intervals or median values with 1st and 3rd quartiles. Mann–Whitney U test for comparison of independent variables and Friedman ANOVA for serial measurements were used. Chi2 test with Yates’s correction were used to compare distribution between the groups. Correlation coefficients were calculated according to Spearman’s method.

Multivariate regression analysis was performed for sNGAL and uNGAL including potential explanatory variables: GA, BW, Apgar score at the 5th min of life (5′ Apgar), and alternatively CRP or PCT values. A model with backward selection of explanatory variables was applied.

p Values <0.05 were considered as statistically significant, and <0.1 was interpreted as borderline statistical significance.

Results

The study included 25 premature babies of 29 and less weeks of GA, whose BW and 5′ Apgar scores were significantly lower than in the compared subgroup (N = 32) of more mature babies (30 weeks of gestation and above)—Table 1. The clinical condition evaluated by NTISS was significantly worse during the first and the second week of life, improving over time (at the end of the first month of life), and was comparable with more mature patients. The subgroup of less mature newborns was characterized by higher WBC count during the first two weeks of life, but similar CRP and PCT values. There was no clear trend in CRP values. Contrary, PCT levels were declining over time. UO, sCr, and serum CystC values were nearly similar (Table 2). As expected, sCr values were declining over time, and corresponding to increasing eGFR values. Both subgroups did not differ in sNGAL and uNGAL during all time points. Higher values of NGAL were found in serum than in parallel urine samples. In 22 newborns, perinatal asphyxia was diagnosed. Both sNGAL and uNGAL levels in those newborns assessed during the first week of life did not differ significantly from values in newborns without a history of asphyxia (data not shown).

We calculated the cut-off values for non-AKI patients (95th percentiles) in the first week of life as 279 ng/mL for sNGAL and 187 ng/mL for uNGAL.

On the basis of sCr levels and changes, the diagnosis of AKI was made in only eight neonates, six ELBW in the first two weeks of life and two in more mature patients, who developed severe LOS. In AKI neonates sNGAL median values with 1st and 3rd quartiles: 238 (94–265) versus 89 (55–172); p = 0.08 and especially uNGAL levels (187 (160–404) vs. 59 (35–92), p = 0.006) were increased. Serum CystC levels were insignificantly increased (median values and 1st and 3rd quartiles: 1.91 (1.78–2.08) versus 1.75 (1.41–1.99, NS). In consecutive measurements, a mild, 22% decline of serum CystC was noticed. The decline was markedly lower than those observed for sCr (58%).

The analysis of a small and heterogeneous subgroup of AKI patients was difficult and underpowered. Investigation of clinical course of patients with AKI (data not showed) revealed a tendency towards higher values of sNGAL (however, not always reaching the cut off value of 95th percentiles for non-AKI patients) consistent with the elevated values of CRP and/or PCT, but in the presence of decreasing sCr. uNGAL values were elevated. In one of the patients, severe sepsis was recognized in the second month of life (beyond the period of observation) and both sNGAL and uNGAL values were elevated (above 95th percentiles) with normal sCr (0.63 mg/dL). In other patients, elevated CRP on the second week of life correlated with increased sNGAL and uNGAL despite moderately increased sCr values (0.9 and 1.02 mg/dL). Very high NGAL values were observed in AKI septic patients (late onset sepsis) with very severe clinical condition and fatal outcome.

The small number of subjects prevents us from incorporating statistical methods in the analysis, however, attention to this phenomenon is required.

IUGR was diagnosed in eight newborns and the diagnosis was associated with significantly increased uNGAL during the first assessment (median values with 1st and 3rd quartiles: 88 (60–306) vs. 59 (35–92), p = 0.007). RDS, BPD and PDA did not affect sNGAL or uNGAL assessment in any time point.

During the 3-month period of observation, four patients died. The low number of fatal outcomes precluded the analysis of predictive value of NGAL.

Correlation analyses

Analysis of demographic factors revealed a lack of correlation between sNGAL, GA, BW and 5′Apgar. uNGAL on the other hand correlated with BW and GA on the first and second week of life. Both sNGAL and uNGAL levels were correlated with NTISS values on the second week of life, and additionally uNGAL during the first point in time (Table 3).

Table 2. Characteristics of renal parameters in all premature infants and distinguished subgroups according to the grade of maturity (means with 95% confidence intervals).

	All [N = 57]	≤29 hbd [N = 25]	≥30 hbd [N = 32]	Statistical significance
UO1st week (mL/kg/h)	5.4 (4.7–6.1)	6.0 (4.9–7.2)	4.8 (4.0–5.7)	0.08
UO 2nd week (mL/kg/h)	5.7 (5.0–6.5)	6.3 (5.1–7.5)	5.1 (4.4–5.9)	0.09
UO4th week (mL/kg/h)	4.6 (4.0–5.3)	5.0 (3.9–6.0)	4.1 (3.2–5.1)	0.22
sCr 1st week (mg/dL)	0.98 (0.92–1.05)	1.05 (0.95–1.14)	0.93 (0.84–1.02)	0.07
sCr 2nd week (mg/dL)	0.75 (0.66–0.84)	0.81 (0.68–0.95)	0.70 (0.58–0.81)	0.16
sCr 4th week (mg/dL)	0.53 (0.48–0.59)	0.53 (0.42–0.64)	0.54 (0.47–0.60)	0.93
eGFR 1st week (mL/min)	14.9 (13.7–16.1)	12.4 (11.1–13.7)	17.0 (15.5–18.5)	<0.001
eGFR 2nd week (mL/min)	20.9 (19.0–22.7)	17.4 (14.9–20.0)	23.6 (21.4–25.9)	<0.001
eGFR 4th week (mL/min)	28.5 (25.7–31.3)	26.3 (22.7–29.9)	30.1 (26.0–34.3)	0.17
CystC 1st week (mg/L)	1.80 (1.65–1.95)	1.79 (1.53–2.04)	1.81 (1.61–2.01)	0.85
CystC 2nd week (mg/L)	1.82 (1.65–2.00)	1.76 (1.45–2.07)	1.87 (1.64–2.09)	0.57
CystC 4th week (mg/L)	2.01 (1.70–2.32)	1.75 (1.50–2.00)	2.11 (1.69–2.54)	0.28
sNGAL 1st week (ng/mL)	144 (107–180)	148 (90–207)	140 (90–190)	0.81
sNGAL 2nd week (ng/mL)	118 (82–154)	127 (69–185)	110 (61–160)	0.66
sNGAL 4th week (ng/mL)	100 (43–157)	76 (15–138)	114 (27–200)	0.53
uNGAL 1st week (ng/mL)	106 (59–153)	106 (66–146)	106 (24–188)	0.99
uNGAL 2nd week (ng/mL)	111 (73–149)	134 (73–194)	91 (40–142)	0.27
uNGAL 4th week (ng/mL)	83 (32–135)	63 (32–95)	96 (11–180)	0.54

Note: UO: urinary output; sCr: serum creatinine; eGFR: estimated glomerular filtration rate; CystC: cystatin C; sNGAL: serum NGAL; uNGAL: urinary NGAL.

Table 3. Univariate correlations between sNGAL and uNGAL values on the following points of time and selected parameters.

	NGAL
	1st Week		2nd Week		4th Week
	Serum	Urinary	Serum	Urinary	Serum	Urinary
NTISS	NS	R = 0.354	R = 0.383	R = 0.352	NS	NS
		p = 0.02	p < 0.01	p = 0.02
sCr	NS	NS	R = 0.311	NS	NS	NS
			p = 0.03
eGFR	NS	NS	R = −0.304	R = −0.259	NS	NS
			p = 0.03	p = 0.08
UO	NS	NS	NS	NS	NS	NS
WBC	R = 0.229	NS	R = 0.291	NS	NS	NS
	p = 0.08		p = 0.004
PLT	R = −0.268	NS	R = −0.350	R = −0.336	NS	NS
	p = 0.06		p = 0.005	p = 0.02
CRP	R = 0.311	R = 0.373	R = 0.394	R = 0.302	NS	NS
	p = 0.03	p = 0.011	p < 0.01	p = 0.05
PCT	R = 0.285	NS	R = 0.542	R = 0.478	NS	NS
	p = 0.07		p < 0.001	p = 0.003
BW	NS	R = −0.425	NS	R = −0.340	NS	R = −0.366
		p = 0.003		p = 0.02		p = 0.04
Apgar 5′	NS	NS	NS	NS	NS	NS
GA	NS	R = −0.278	NS	R = −0.289	NS	NS
		p = 0.05		p = 0.04
CystC	R = 0.285	NS	NS	NS	NS	NS
	p = 0.09

Note: NTISS: Neonatal Therapeutic Intervention Scoring System; sCr: serum creatinine; eGFR: estimated glomerular filtration rate; UO: urinary output; WBC: white blood cell; PLT: platelet count; CRP: C-reactive protein; PCT: procalcitonin; BW: birth weight; GA: gestational age; CystC: cystatin C.

There was a weak correlation between sNGAL and sCr on the second week of observation. CRP positively correlated with both sNGAL and uNGAL on the first and second week of life; PCT with sNGAL and uNGAL on the second week of life and only sNGAL on the first week of life. WBC positively correlated with sNGAL and uNGAL on the first week of life; while platelet count (PLT) correlated negatively with both parameters on the second week of life and only with sNGAL on the first week of life. After the first month of life, there was no correlation between NGAL and inflammatory markers which could potentially be explained by a lack of severe inflammatory processes in the majority of observed patients.

Multiple regression analysis (backward model)

Multiple regression analysis including sNGAL and uNGAL as independent variables was performed separately for the three points of time. As explanatory variables GA, BW, 5′Apgar, sCr and alternatively CRP or PCT were introduced (Table 4).

Table 4. Multiple regression analysis including sNGAL and uNGAL as independent variables and gestational age, birth weight, 5 min Apgar score, serum creatinine, and alternatively CRP or PCT as an explanatory variables.

	sNGAL			uNGAL
	1st Week	2nd Week	4th Week	1st Week	2nd Week	4th Week
BW	NS	NS	NS	NS	NS	NS
GA	NS	NS	NS	NS	β = −0.211	NS
					p = 0.08
Apgar 5′	NS	NS	NS	NS	NS	NS
CRP	β = 0.301	β = 0.683	β = 0.638	NS	β = 0.659	β = 0.753
	p = 0.04	p < 0.001	p < 0.001		p < 0.001	p < 0.001
BW	NS	NS	NS	β = −0.341	β = −0.193	NS
				p = 0.08	p = 0.09
GA	NS	NS	NS	NS	NS	NS
Apgar 5′	NS	NS	NS	NS	NS	NS
PCT	β = 0.307	β = 0.817	β = 0.696	β = 0.293	β = 0.672	β = 0.759
	p = 0.06	p < 0.001	p < 0.001	p = 0.08	p < 0.001	p < 0.001

Note: sNGAL: serum NGAL; uNGAL: urinary NGAL; BW: birth weight; GA: gestational age; CRP: C-reactive protein; PCT: procalcitonin.

First week of life analysis revealed that uNGAL variability was explained by BW (β = −0.341) and PCT (β = 0.293).

Serum NGAL variability was explained by CRP (β = 0.301) and PCT (β = 0.307).

Introducing the following (second week of life) values of sNGAL and uNGAL into the analysis showed CRP (β = 0.659), PCT (β = 0.672), GA (β = −0.211) and BW (β = −0.193), to be an explanatory variables for uNGAL variability, and CRP (β = 0.683) and PCT (β = 0.817) for sNGAL.

Analysis performed at the last point of time did not show any significant impact of BW and GA on uNGAL and sNGAL variability. It was only explained by inflammatory markers.

Discussion

This study demonstrates that BW, immaturity, and inflammatory status are the major factors restricting the usefulness of NGAL (decreasing its specificity) as the marker of AKI (renal tubules injury). sNGAL values are dependent on inflammatory status, while uNGAL – on both immaturity and inflammatory status.

The diagnosis of AKI in preterm newborns is a great challenge.13–15 sCr, regardless of its limitations, in the group of newborns related to a decreasing influence of maternal-fetal transplacental transfer in the first days of life, body mass, and maturation variability, is still the most valuable marker of kidney function, routinely used in daily clinical practice.13 The second determinant of AKI diagnosis: UO is not very useful in neonates as non-oliguric AKI was often observed in this group of patients.14 A newer parameter of eGFR is CystC.16 Its usefulness was attracted by the lack of dependence from muscle mass, however, evidence supporting superiority over sCr are still elusive.16–21 A new potential biomarker postulated as a more sensitive parameter of AKI is NGAL.22 The introduction of this parameter into routine diagnostic workup requires establishing the reference values and knowledge of methodological limitations. Our study gives rise to the need of determining normal ranges of age and BW, especially of uNGAL. It should be stressed that inflammatory status has significant impact on both sNGAL and uNGAL.7^,23 Thus, increased sNGAL and uNGAL values are not just reflecting tubular injury, but probably injuries of other epithelial cells as well. We showed a strong correlation between inflammatory markers such as CRP and PCT and lipocalin-2 assessed both in urine and serum. Similar results were previously obtained in term newborns by us24 and other researchers.22^,23^,25 In our project, sNGAL and uNGAL values were checked as a surrogate, hence possibly more sensitive parameter of kidney injury. To make the study group more homogenous in respect to the stage of maturity, we distinguished two subgroups based on GA. These two subgroups showed similar values of sNGAL and uNGAL, and did not differ significantly in parameters of inflammatory status (CRP, PCT). The cause of inflammation could be of different origin (prolonged respiratory insufficiency and infection vs. infection solely).

In line with previous studies8^,9 we found in univariate analyses a negative association between BW and GA, and uNGAL, but not sNGAL. This finding has been confirmed by multiple regression models on the first and second week of life, but not a month after birth.

This likely reflects growing maturity with GA of epithelial tissue in the renal tubules, and suggests a necessity of higher cut off values in preterm neonates than in term neonates for the diagnosis of AKI. Similarly, to avoid misdiagnosis of AKI in septic patients, different cut off values for both sNGAL and uNGAL have to be established. We have demonstrated a strong correlation between markers of inflammatory status (WBC, CRP, PCT) and both sNGAL and uNGAL. The fact that the correlation between NGAL levels and CRP or PCT was stronger than with WBC values, indicates damage to endothelial cells other than of tubular origin6^,7^,22^,23^,25 plays a greater role in the release of NGAL than the activation of neutrophils. It is well known that immunological response of immature newborns is weaker and neutrophil activation limited.26

Contrary to the findings made by some authors,10 we did not find increased NGAL values in neonates with perinatal asphyxia. It can be expected that the influence of perinatal asphyxia on the release of NGAL is subsiding with time. We cannot exclude that the first assessment performed in our study was too late to demonstrate the effect of asphyxia, or the study group was not appropriately large.

Based on sCr value (absolute number and increase in trend), AKI was recognized in eight newborns. In our opinion it is too soon to relay on s/uNGAL values solely in early recognition of AKI, but we definitely should not neglect this marker. Inflammatory conditions such as BPD10 and sepsis, and especially urinary tract infection24^,27 are probably the most important limitation for the interpretation of NGAL values, decreasing its specificity in the diagnosis of AKI. However this does not diminish uNGAL value in detection of AKI of non-septic origin in premature infants. NGAL usefulness was shown in cardiac surgery, where the start of injury and duration of the process was relatively easy to estimate, allowing measurement of the time between the incitement of injury and the rise in NGAL urine concentration.28^,29 In newborns, particularly premature ones, the moment when the damaging factor(s) initiate kidney injury is difficult to determine. This is one of the most important reasons why the relation between an increase in uNGAL and the risk factors is not clear and varies in time. It may be possible that more frequent measurements help to detect the moment when kidney injury starts.

Regarding reference values, there is a lack of precise numbers. Huynh suggested 5 ng/mL, 50 ng/mL, 120 ng/mL, and 2–150 ng/mL as the median, 95th and 99th percentiles, and range of pooled uNGAL values, respectively.8 In comparison with data presented by Huynh, our cut off values (for 95th percentiles) is much higher, which might be caused by the less stable clinical condition of our patients caused by immaturity, sepsis, and SIRS. Incorporating cut-off values proposed by Huynh into the analysis of our study group revealed many more cases of elevated uNGAL than would result from the application of our cut off values. On the other hand, in our analyzed group, 60% of non-AKI cases presented increased (higher than 50 ng/mL) levels of uNGAL. Therefore, the suggested cut off value is highly sensitive, but with low specificity. It implies that establishing universal reference ranges useful for AKI detection will be problematic.

Neutrophil origin of non-monomeric NGAL seriously limits current easily available total sNGAL measurements and its value in evaluation of AKI.30 Using molecular methods for distinguishing the origin of NGAL (neutrophile vs. tubular) might be helpful in solving this problem,30 but it would be difficult to introduce into everyday clinical practice; however, even this highly sophisticated method would not overcome limitations related to functional immaturity of renal tubules.

As prematurity affect literally all organs and systems, set parameters and markers are required for control, but unfortunately that is very difficult in practice. The organism gives signals which are not specific and the range of organ/system specific markers is limited. NGAL is a marker which cannot be omitted in search for a useful monitoring parameter of kidney dysfunction; however, its current implementation into the box of tools for AKI recognition requires further investigation.

To summarize, the results of the study revealed that the measurement of NGAL cannot be solely evaluated as a marker of AKI in premature infants, and serum and urinary values of NGAL are influenced by variety of factors, including inflammatory status and prematurity, that limits the specificity of total NGAL measurement as the marker of AKI.

Declaration of interest

The authors declare that there is no conflict of interests regarding the publication of this paper. This study was supported by the grant from Medical University of Silesia (KNW–1-057/P/2/0). None of the authors obtained support from kit manufacturers (R&D, BioPorto and Immunodiagnostic).