Pediatric acute kidney injury is associated with impairment in nicotinamide adenine dinucleotide (NAD+) metabolism

Abstract

Background

Acute kidney injury (AKI) is highly prevalent among hospitalized children, but there is no treatment. Impaired de novo nicotinamide adenine dinucleotide (NAD+) biosynthesis measured via elevation of the urine quinolinic acid-to-tryptophan ratio (uQ:T) is a feature of AKI that has been described in preclinical models and humans with AKI. Small prospective trials to restore NAD+ abundance with NAD+ precursor supplementation have shown promise in the prevention and treatment of AKI. It is not known whether pediatric patients also develop suppression of NAD+ biosynthesis during AKI, but such information will be critical before children can be included in NAD+ -based clinical trials to treat or prevent AKI.

Methods

An observational cross-sectional study was performed on convenience urine samples from children hospitalized in a tertiary care children’s hospital. Samples were split into five groups: outpatient controls, floor controls, ICU controls, floor AKI, and ICU AKI. Clinical data were collected from the medical record, and metabolites were measured using targeted mass spectrometry. Patients with AKI were compared to their respective controls. A multivariate linear regression was used to assess whether demographic variables were independently associated with uQ:T, and odds of AKI were assessed in serial uQ:T tertiles using multivariate logistic regression models that adjusted for patient variables.

Results

Sixty-nine control patients (39 outpatient, 10 floor, and 20 ICU controls) and 22 AKI patients (12 floor and 10 ICU) were enrolled. uQ:T was elevated in patients with AKI compared to their respective controls. No demographic variables were independently associated with uQ:T, and when adjusting for patient demographic and clinical variables, the odds of AKI increased serially with uQ:T tertile.

Conclusions

Elevated uQ:T is a feature of pediatric AKI. The present results warrant further exploration in observational and potentially interventional studies with NAD+ precursor therapies.

Background

Acute kidney injury (AKI) occurs in 11–31% of hospitalized children [1–3] and up to 54% of critically ill children [2]. As the medical complexity of pediatric inpatients increases, the incidence of AKI is increasing [4, 5]. Despite this, there is no treatment for AKI beyond supportive measures. In adults, there is evidence that episodes of AKI increase the risk of future chronic kidney disease (CKD) in a severity-dependent fashion [6–8]. A small number of trials in children also describe an increased risk of death, CKD, and future episodes of AKI after an index AKI episode [9–12]. Notably, the longest follow-up period in these trials is 25 years [11], a fraction of the lifespan of a pediatric patient.

The kidney is a highly metabolic organ that requires substantial adenosine triphosphate (ATP) for essential solute transport functions. Nicotinamide adenine dinucleotide (NAD+) is required for ATP production. Disruption of kidney NAD+ metabolism is a recently described feature of AKI in preclinical models and adult humans. Specifically, the de novo NAD+ biosynthesis pathway—whereby tryptophan is enzymatically converted through the intermediary metabolite quinolinic acid to NAD+ —is suppressed in AKI (Fig. 1) [13, 14]. Suppression of this pathway leads to a reduction of kidney NAD+ and an accumulation of de novo pathway metabolites [13, 14]. Several studies have confirmed these findings by detecting elevated de novo pathway metabolites in kidney tissue, urine, and plasma of adults with kidney injury [13, 15–20]. The most commonly measured metric is the quinolinic acid-to-tryptophan ratio (uQ:T), which correlates with disease severity and adverse clinical outcomes [13, 15–17]. Suppression of de novo NAD+ biosynthesis may be clinically impactful, as several prospective interventional studies have proposed that amelioration of AKI and downstream adverse outcomes can be achieved via treatment with nutritional NAD+ precursors [13, 21–23].

Fig. 1.

Simplified schematic of NAD+ biosynthesis. QPRT, quinolinate phosphoribosyltransferase; NAMN, nicotinic acid mononucleotide; NAAD, nicotinic acid adenine dinucleotide; NAD+, nicotinamide adenine dinucleotide; NAM, nicotinamide; NMN, nucleotide adenine dinucleotide

To date, no studies have assessed whether NAD+ biosynthesis disruption is a feature of pediatric AKI. This is a critical gap in knowledge that needs to be addressed both to expand the understanding of NAD+ biosynthesis disruption in AKI and to determine whether trials of NAD+ augmentation could be pursued among children at risk of AKI or its adverse outcomes.

Methods

Ethics

This study was approved by the institutional review board (IRB) at the University of Texas Southwestern (STU-2021–0835) and was carried out in accordance with the Declaration of Helsinki. This study was performed on discarded samples with minimal risk to patients; thus, no informed consent was obtained, per a waiver of consent approved by the IRB.

Study population and specimen collection

This is an observational cross-sectional study using discarded patient samples. For a 4-month period, all urine samples discarded from a tertiary care children’s hospital were collected. Samples were collected via multiple methods based on clinical needs, including clean catch, bagged specimens, and catheterization. Using medical record numbers associated with the samples, the electronic medical records were reviewed. Patient samples from children aged 0–18 years that fell into any of five categories were aliquoted and stored at − 80 °C within 2 h of collection from the clinical lab. Samples were maintained on ice until frozen. Categories were (1) ICU admission without acute kidney injury (AKI) (ICU controls); (2) ICU admission with AKI (ICU AKI); (3) floor admission without AKI (floor controls); (4) floor admission with AKI (Floor AKI); and (5) outpatient controls. AKI was defined as doubling of creatinine from documented baseline in the 3 months prior to collection or clinician documentation of AKI diagnosis, which was verified via independent review of the medical record by a pediatric nephrologist. Even though KDIGO guidelines describe AKI as > 50% increase from baseline, we chose to define AKI as doubling (> 100% increase) to ensure our AKI population had true kidney injury and to avoid misclassification of patients with low baseline creatinine, in whom a 50% creatinine increase was within the margin of error for testing. ICU patients include those admitted to the Pediatric Medical ICU (PICU) and the Cardiovascular ICU (CVICU). Outpatient controls were defined as samples collected from healthy patients undergoing some form of pregnancy, drug, trauma, or sexually transmitted infection (STI) screening. Outpatient control patients without baseline creatinine measurements available were assumed to not have AKI as they were described as “well” by the treating physician collecting the urine sample and typically being seen for well-child visits undergoing routine screenings. Any control patients who tested positive for pregnancy, STI, or UTI were excluded. Samples and data from the medical record were deidentified.

Data collection

Patient age, race, ethnicity, sex, height, serum creatinine at the time of urine collection, and medical diagnosis were collected from the medical record. eGFR was calculated using the bedside Schwartz equation.

Targeted mass spectrometry

Isotopic standards

Isotopic standards (L-Tryptophan-d5 and quinolinic acid-d4) were purchased from Toronto Research Chemicals, North York, Ontario, Canada. Isotopic standards were mixed at a concentration of 50 μM in water and mixed 1:1 with urine samples before protein precipitation with acetonitrile (80% w/v final concentration). Non-isotopic calibrators were purchased from Sigma-Aldrich (St. Louis, MO).

Instrumentation

All liquid chromatography-mass spectrometry (LC–MS) analyses were performed using an API 5500 Triple Quadrupole mass spectrometer (AB Sciex) coupled to a Prominence UFLC liquid chromatography system with an autosampler (Shimadzu Scientific Instruments). High-performance chromatography of metabolites was performed using hydrophilic interaction chromatography on Luna HILIC HPLC columns (100 × 4.6 mm, 2.6 μm bead diameter, 100 A pore size; Phenomenex). Instrument control, data acquisition, and quantification were performed using the Analyst version 1.6.2 for Windows software (Sciex).

Urine small molecule analyses

A quantity of 10 μl of urine was mixed with 10-μl isotopic standards (50 μmol diluted in water). Urine proteins were then precipitated by mixing with 80-μl acetonitrile (80% final concentration), and precipitates were cleared by centrifuging at 14,000 rpm for 10 min. The supernatant was transferred into a 96-well microplate for analysis. Analysis was performed by hydrophilic interaction chromatography on a Luna HILIC HPLC column using a gradient elution protocol. Calibration standards for each metabolite were purchased from Sigma-Aldrich, weighed, and diluted into phosphate-buffered saline at 100 mM, and calibrator mixtures were made from stock solutions starting at 200 μM. Peaks consistent with tryptophan, kynurenine, kynurenic acid, quinolinic acid, and creatinine were quantified.

Data processing and statistical analysis

Urine quinolinic acid was expressed as a ratio with urine tryptophan based on previous studies [13, 16, 18, 24]. For statistical comparison, demographic variables were compared using a Kruskal–Wallis test, continuous variables were compared with a Mann–Whitney test, and correlations were assessed with a coefficient of determination. A multivariate linear regression was used to assess whether demographic variables were independently associated with uQ:T. For the most robust analysis, the odds of AKI were assessed in serial uQ:T tertiles using multivariate logistic regression models that adjusted for patient variables. Statistical outliers were identified and removed using the ROUT method with a Q = 0.1% to only remove definite outliers. Removed outliers are described in the figure legends. All statistical testing was completed using Graphpad Prism or R version 4.1.2, 2021 [25].

Results

Demographic data of patients from each group are displayed in Table 1. There was no difference in age, sex, race, or Hispanic ethnicity between groups. To test our hypothesis that uQ:T is elevated in pediatric patients with AKI, we first compared uQ:T among three groups: floor AKI vs. floor controls (Fig. 2A), ICU AKI vs. ICU controls (Fig. 2B), and all AKI vs. all controls (Fig. 2C). All three comparisons showed increased uQ:T in the AKI groups (p = 0.0043, 0.035, and 0.0014, respectively). Among all AKI and control patients, uQ:T negatively correlated with eGFR (R² = 0.14, p = 0.013) (Fig. 2D).

Table 1.

Demographic characteristics and eGFR of patients. eGFR calculated based on serum creatinine and the bedside Schwartz equation.

	Number of patients	Mean age (years)	Female, n (%)	White, n (%)	Hispanic, n (%)	Mean eGFR (ml/min/1.72 m2)
All controls	69	10.1	34 (49)	45 (65)	35 (51)	115
Outpatient controls	39	10.7	18 (45)	24 (60)	23 (58)	96
Floor controls	10	11.3	4 (40)	8 (80)	2 (20)	143
ICU controls	20	8.2	12 (60)	13 (65)	10 (50)	110
All AKI	22	10.4	12 (52)	10 (43)	8 (35)	66 *
Floor AKI	12	12.1	5 (42)	5 (42)	5 (42)	60
ICU AKI	10	8.4	7 (70)	5 (50)	3 (30)	78*

^*Excludes patients receiving kidney replacement therapies

AKI acute kidney injury, ICU intensive care unit, eGFR estimated glomerular filtration rate

Fig. 2.

Quinolinic acid is elevated in the urine of pediatric patients with AKI. A Urine quinolinic acid in floor AKI patients vs. floor controls. Expressed as a ratio with urine tryptophan (uQ:T). One outlier was removed from the floor control group, as described in methods (uQ:T = 10.6). B uQ:T in ICU AKI vs. ICU controls. One outlier was removed from the ICU AKI group (uQ:T = 25.5), and two outliers were removed from the ICU control group (uQ:T = 48.2 and 17.4). C uQ:T in all AKI patients vs. all controls. The same four outliers described in A and B were removed. D Correlation of uQ:T with eGFR estimated using serum creatinine and the bedside Schwartz equation. Four patients with artificially elevated eGFR while receiving kidney replacement therapy are excluded. E Results of multivariate linear regression assessing the association of demographic variables with uQ:T among all AKI and control patients. No outliers were removed. F Odds ratios of AKI diagnosis from multivariate regression models using data from all AKI and control patients. Error bars represent 95% confidence intervals. Model 1 controls for demographic variables described in 2E. Model 2 controls for demographic variables plus illness severity markers like patient location (ICU, inpatient, outpatient) and whether patients required advanced support. No outliers were removed. *p < 0.05, **p < 0.01. AKI, acute kidney injury; eGFR, estimated glomerular filtration rate; uQ:T, urine quinolinic acid-to-tryptophan ratio; ICU, intensive care unit

To ensure no demographic variables are independently associated with uQ:T, we performed a multivariate linear regression, which showed no significant association between uQ:T and any demographic variable while controlling for all other variables (Fig. 2E). For the most robust assessment of uQ:T in this diverse cohort, patients were assessed in serial tertiles of uQ:T using multivariate models that adjusted for patient variables. Model 1 adjusted for all demographic variables (age, sex, race, ethnicity). Model 2 adjusted for demographic variables plus patient location (outpatient, inpatient floor, or ICU) and the requirement for any advanced support (pressor support and/or positive pressure ventilation). In these models, the odds ratio (OR) of AKI diagnosis significantly increased with increased uQ:T (Fig. 2F). To ensure critical illness alone was not associated with elevated uQ:T, we replicated Models 1 and 2 to calculate odds ratios of ICU admission, adjusting for all variables listed above plus AKI diagnosis status. All three uQ:T tertiles in both models did not demonstrate a significant OR > 1 (not shown).

Discussion

This cross-sectional assessment of NAD+ -related metabolites in the urine of children with AKI demonstrates evidence of impaired de novo NAD+ biosynthesis, as has previously been reported in preclinical studies and adult humans. Our findings suggest that de novo NAD+ biosynthesis impairment may be a conserved feature of kidney stress. To our knowledge, this is the first study to document evidence of impaired NAD+ biosynthesis in children with kidney injury.

In adults, NAD+ -related metabolites have been elevated in patients with AKI after cardiac bypass surgery, with levels correlating to AKI severity [24]. Likewise, elevated urinary NAD+ metabolites are associated with the development of AKI after methotrexate therapy or liver transplant [26]; changes in these metabolites may also predict kidney recovery after AKI in critical illness [27]. The present results suggest that uQ:T may be a useful, non-invasive clinical biomarker of kidney stress in children. Serum creatinine is not sufficiently accurate to diagnose AKI for several reasons: it is a late marker of GFR; it fluctuates with muscle mass; its production and/or release into the blood may be affected by critical illness; and normal values can be low in small children, thus making detection of incremental increases difficult. Future prospective studies will need to determine how uQ:T compares to other biomarkers of AKI, such as urine neutrophil gelatinase associated lipocalin (NGAL) or kidney injury marker-1 (KIM1).

In preclinical models, elevation of de novo NAD+ metabolites in the urine associates with reduced NAD+ in kidney tissue [13, 14]. Replenishing NAD+ via supplementation of NAD+ precursors in the form of vitamin B3 has been shown to reduce AKI severity in pilot prospective clinical trials [13, 21–23]. No such trials have been undertaken in children, but the present results propose that pediatric AKI may also be a setting amenable to evaluation of therapeutic NAD+ augmentation, either as a preventative strategy in high-risk settings or as a therapeutic strategy in established AKI. Further supporting the rationale for such trials, vitamin B3 supplementation has an extensive safety profile in children [28]. Additional prospective studies will be necessary to identify patient populations and clinical scenarios that could be amenable to vitamin B3-based therapies.

One limitation to this study is the non-prospective sampling method using discarded patient samples. This limited the ability to compare uQ:T performance to other biomarkers, which are not stable for measurement after these collection conditions. Second, we collected all discarded urine samples in a 4-month period that were sent as part of routine clinical care. This may have limited collection from patients with small urine volumes that left no remaining urine for discard after testing, such as small babies or oliguric patients. Third, our study populations are heterogeneous: stages of AKI varied; implementation of kidney replacement therapy varied; samples were collected from different time points in participants’ hospitalizations. However, the ability to detect a robust uQ:T elevation in patients with AKI despite this heterogeneity highlights the prominence of uQ:T as a biomarker of kidney stress. Finally, the small number of patients enrolled in our study limited the power to complete more in-depth sub-group analyses.

More research is needed to understand the physiologic significance of elevated quinolinic acid in AKI. While uQ:T is assessed as a biomarker of perturbed NAD+ biosynthesis, quinolinic acid is, itself, a potentially toxic metabolite [29]. Numerous theories suppose that the accumulation of these metabolites may serve an anti-inflammatory role [30]. NAD+ boosting therapies have shown promise in adults, though boosting NAD+ via alternative biosynthetic pathways does not alleviate the accumulation of these metabolites. Likewise, the effect of individuals’ microbiomes on uQ:T in the setting of different ages, diets, and illnesses may warrant additional further study. Thus, more studies are needed to understand the physiologic significance of de novo biosynthesis metabolites and whether therapies to reduce metabolite accumulation may also be beneficial.

In summary, this observational study demonstrates evidence of de novo NAD+ biosynthesis impairment assessed via elevated uQ:T in children with AKI. Our results add to the growing body of evidence that de novo NAD+ biosynthesis impairment is a conserved feature of kidney stress. To our knowledge, this is the first study to document evidence of impaired NAD+ biosynthesis associated with kidney injury in a pediatric population. Larger prospective studies in children are needed to confirm and expand these findings.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s00467-025-06791-5.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.