Increased ureagenesis revealing metabolic intolerance to enhanced nutrition in the PICU: a secondary analysis of the PEPaNIC RCT

Abstract

Background

As compared with withholding parenteral nutrition (PN) for one week (late-PN), early supplementation of insufficient enteral nutrition with PN (early-PN) increased infection risk and delayed recovery of critically ill children in the PEPaNIC-RCT, with harm attributed to higher amino acid doses. We hypothesized that this harm could be explained by increased ureagenesis, revealing metabolic intolerance to enhanced feeding.

Methods

This is a secondary analysis of the PEPaNIC-RCT (ClinicalTrials.gov-NCT01536275), in which patients (newborn-17 years) admitted to pediatric intensive care units (PICUs, Leuven-Rotterdam-Edmonton) were randomly allocated to early-PN or late-PN. For patients not receiving renal replacement therapy (RRT), we documented plasma/serum urea and urea/creatinine ratio (UCR) profiles throughout two PICU weeks and investigated whether an increased maximum value in the randomized intervention window (first week) statistically explained harm by early-PN, via multivariable logistic regression and Cox proportional hazards analyses. We studied risk of new infections and duration of PICU dependency as primary outcomes, 90-day mortality as safety outcome, and duration of hospital dependency as secondary outcome.

Results

Of 1440 PEPaNIC-patients recruited, 665 early-PN and 663 late-PN non-RRT patients had UCR data. Urea concentrations and UCRs were higher in early-PN than in late-PN patients throughout the first 11 days. Higher first week maximum UCR independently associated with (OR/HR expressed per 30 age-normalized units) an increased risk of new infection [adjusted OR (95%CI) 1.746 (1.353–2.265), p < 0.0001], lower likelihood of earlier live PICU [adjusted HR 0.745 (0.671–0.824), p < 0.0001] and hospital discharge [adjusted-HR 0.778 (0.703–0.860), p < 0.0001], and increased 90-day mortality [adjusted OR 1.432 (1.029–1.993), p = 0.033], and statistically explained (part of) any effect of the randomized intervention. Similar results were obtained for urea concentrations.

Conclusions

Increased plasma/serum urea or UCR with early-PN partially explained the harm by early-PN, suggesting these may identify critically ill children metabolically intolerant to enhanced nutrition. This opens perspectives for incorporating urea or UCR into a dynamic monitoring tool to guide the appropriate timing and dosing of nutritional support.

Trial registration

ClinicalTrials.gov NCT01536275, registered February 2012.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13054-026-06345-7.

Background

Critically ill children treated in the pediatric intensive care unit (PICU) suffer from anorexia and feeding intolerance. The resulting macronutrient deficit associates with poor clinical outcome [1]. The optimal strategy to provide artificial nutritional support to these vulnerable patients remains debated, not only regarding dosing and composition, but also regarding timing [2, 3]. The Pediatric Early versus Late Parenteral Nutrition in Critical Illness (PEPaNIC) multicenter, randomized controlled trial (RCT) showed that when enteral nutrition failed to reach macronutrient/caloric targets, early supplementation with parenteral nutrition (early-PN) increased infection risk and delayed recovery of critically ill children as compared with omitting PN during the first week in the PICU (late-PN) [4]. Nevertheless, the ideal timing when advancing artificial feeding through the enteral or parenteral route becomes beneficial in an individual patient remains unclear. Therefore, availability of an easily quantifiable biomarker signaling when a critically ill child is ready, or not yet ready, to receive enhanced feeding would provide major progress.

Looking into changes in amino acid/protein metabolism may provide leverage. Secondary analyses of the PEPaNIC-RCT, and the equivalent EPaNIC-RCT in adult ICU patients [5], suggested that it were the higher amino acid doses, and not the higher glucose or lipid doses that caused the harm by early-PN [6–8]. Moreover, several recent trials showed that higher protein delivery did not improve outcome or worsened outcome of adult critically ill patients [9–12]. The extra amino acids administered with early-PN or with higher protein doses alone were not used for protein synthesis, but rather shuttled to futile catabolism, with profoundly increased ureagenesis as revealed by increased plasma urea concentrations and high net nitrogen wasting in the urine [7, 8, 11, 12]. This futile catabolism may impose further stress on the liver where ureagenesis takes place and on the kidneys that are responsible for urea excretion [13]. Studies in adult ICU patients suggested that, apart from being a marker of catabolism, increased ureagenesis may also explain harm by excessive dietary protein intake [14, 15]. Indeed, the increased risk of death with a higher protein dose in patients with acute kidney injury and severe organ failure in the Effect of Higher Protein Dosing in Critically Ill Patients (EFFORT Protein) RCT and with glutamine administration in patients with multiple organ failure in the REducing Deaths due to OXidative Stress (REDOXS) RCT was found to be statistically explained by increased ureagenesis, as revealed by elevations in serum urea or the urea/creatinine ratio (UCR) [16, 17]. Whether increased ureagenesis may signal nutritional harm in critically ill children remains unclear.

We tested the hypothesis that the early-PN-evoked harm to critically ill children could be explained by increased ureagenesis, revealed by a higher plasma/serum urea or UCR, and therefore that the plasma/serum urea or UCR may serve as a biomarker of metabolic intolerance to enhanced feeding of these vulnerable patients.

Methods

Study design and patients

This is a secondary analysis of the PEPaNIC-RCT (ClinicalTrials.gov-NCT01536275) on timing of initiating supplemental PN when enteral nutrition of critically ill pediatric patients was insufficient to reach caloric targets. From June 2012 to July 2015, the study enrolled 1440 critically ill infants and children from full-term newborn to 17 years old, with a medium to high risk of malnutrition and a central venous line in place for clinical purposes, all patients who were admitted to the PICUs of the University Hospitals in Leuven (Belgium), the Sophia Children’s Hospital at the Erasmus Medical Center in Rotterdam (the Netherlands), or the Stollery Children’s Hospital in Edmonton (Canada). The study protocol and primary results of the PEPaNIC-RCT have been published [4, 18]. The study protocol was approved by the Institutional Review Boards of the participating centers (ML8052, NL49708.078, Pro00038098). Written informed consent was obtained from the parents or legal guardians. The study was done in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and later amendments.

The patients had been randomly assigned (1:1) to an early-PN or a late-PN group. In both patient groups, enteral nutrition was initiated as soon as possible. In the early-PN group, supplemental PN was initiated within 24 h after PICU admission when enteral nutrition was insufficient, in order to reach local caloric and macronutrient targets early. In the late-PN group, supplemental PN was omitted in the first week in the PICU. Patients in this group received an intravenous mixture of 5% dextrose and saline (60/40 v/v) to match the amount of intravenous fluid administered to the patients in the early-PN group, accounting for the volume of the enteral nutrition. In case on the morning of day 8 enteral nutrition still failed to provide 80% of the caloric target, supplemental PN was initiated to reach that target. Both study groups received intravenous micronutrients (trace elements and vitamins) and minerals early to prevent deficiencies and subsequent refeeding syndrome, as well as blood glucose control with insulin to local target ranges. PN was discontinued when patients received 80% of the caloric target through the enteral or oral route. Outcome assessors and investigators not directly involved in the care for the patients during the PICU stay were not informed about treatment allocation.

For this study, we selected all PEPaNIC-patients with any data available for the plasma/serum UCR in the first week in the PICU. For the primary analyses, we also excluded those who received renal replacement therapy (RRT) in the first week [19].

Urea and creatinine measurements and calculation of the UCR

To study whether harm by early-PN vs. late-PN could be statistically explained by increased ureagenesis, we studied the impact of the intervention on plasma (Leuven) or serum (Rotterdam, Edmonton) urea concentrations and UCR, the latter to correct for changes in kidney function [20]. These analytes are further referred to as urea, creatinine and UCR. Urea and creatinine concentrations were measured daily at 6 ± 2 a.m. by routine clinical chemistry in the participating centers’ central hospital laboratories. In case multiple measurements were available on a day, the maximum value was used.

In view of the age-dependency of normal creatinine ranges, we converted the measured creatinine concentrations to age-normalized values by dividing the measured concentrations by the corresponding age-specific reference value. As reference value, we used the arithmetic mean of the lower and upper limits of the specific reference ranges for each age group (Additional File 1 Table S1). Accordingly, the age-normalized UCR was calculated by dividing the urea concentrations by the age-normalized creatinine. Any further referral to creatinine or UCR depicts the age-normalized values.

We generated profiles of the daily maximum values of urea, creatinine and UCR for the first 14 days in the PICU, complemented with the corresponding maximum values in the first week (day 1 to day 7) and the second week (day 8 to day 14).

Outcomes

As done for the original trial [4], we studied the risk of acquiring a new infection and the duration of PICU-dependency as primary outcomes, mortality at 90 days after randomization as the safety outcome, and the duration of hospital-dependency as secondary outcome. Acquisition of a new infection was defined as need for treatment with antimicrobial agents for more than 48 h, which was not for prophylaxis and not initiated prior to PICU admission or within the first 48 h of PICU admission. A left-truncation approach was applied, meaning that patients who acquired a new infection before the day of the maximum urea or age-normalized UCR in the first week were considered as not having acquired a new infection. We both studied risk of acquiring a new infection as well as time-to-new infection, the latter censoring patients without an infection at 90 days. Duration of PICU dependency and hospital dependency were studied with time-to-event analyses censored at 90 days, with non-surviving patients censored beyond all survivors at 91 days to account for death as a competing risk.

Statistical analyses

Patient characteristics are reported as medians with interquartile ranges or as numbers with percentages. Biomarker profiles are reported as means with standard errors. We used the Wilcoxon Rank Sum test to compare continuous biomarker data between groups. We then investigated whether an effect of early-PN vs. late-PN on the maximum urea and UCR in the first week statistically explained the effect of the randomized intervention on the outcomes of interest. We therefore performed multivariable logistic regression and Cox proportional hazards analyses, adjusting for baseline risk factors (directed acyclic graph presented in Additional File 1 Figure S1). These baseline risk factors included treatment center, age, type of illness (admission diagnosis), severity of illness (pediatric index of mortality (PIM) 3 score and pediatric logistic organ dysfunction (PeLOD) score over the first 24 h), and severity of malnutrition risk (screening tool risk on nutritional status and growth in children (STRONGkids) score dichotomized for medium and high risk) [21–23]. For each outcome, a first model (model A) assessed the independent impact of early-PN vs. late-PN on the outcome, adjusting for the described risk factors. In a second model (model B), the first week’s maximum value of either urea or UCR was added to assess potential statistical explanation of any harm by early-PN. In addition, we performed several sensitivity analyses. We tested whether the results would remain similar when penalizing for risk of death in the PICU in the analyses for new infection (assigning a new infection on the day of death if the patient was not scored as having a new infection before), when not excluding RRT patients, when excluding neonates (≤ 28 days) to avoid potential confounding by maternal renal function, for infants (< 1 year) versus older children (≥ 1 year), for boys versus girls, and for patients with or without hypophosphatemia on day 2 (in view of the previously documented higher vulnerability to harm by early-PN in patients with hypophosphatemia on day 2 [24]. We further aimed to gain insight in whether an increase in urea or UCR as such, or rather an increase in urea or UCR due to higher protein doses, would be most important in determining outcome. Therefore, we added individual average protein intake up to the day of maximum urea or UCR to the models. Results were reported as adjusted odds ratios (aOR) or adjusted hazard ratios (aHR) with 95% confidence intervals (CIs).

Next, we generated receiver operating characteristic (ROC) curves to assess the performance of the first week’s maximum urea and UCR in predicting new infection or 90-day mortality, with determination of the corresponding area under the ROC curve (AUC) and cut-off point that maximizes specificity and sensitivity, further referred to as “threshold”. We determined the percentage of early-PN and late-PN patients who crossed these thresholds in the first week, respectively in the second week. The identified thresholds were further used for the drawing of Kaplan-Meier curves to visualize the impact of exceeding the threshold in the first week on the risk of acquiring a new infection and of 90-day mortality, with differences among groups assessed with the log-rank test and with multivariable logistic regression analyses adjusting for the risk factors described above.

All statistical analyses were done with JMP © 19.0.0 (SAS Institute, Inc., Cary, NC). A p ≤ 0.05 was considered statistically significant.

Results

Urea and creatinine data in the first week of PICU stay were available for 665 early-PN and 663 late-PN patients who did not receive RRT in the first week (Fig. 1). Patients in both groups were comparable for baseline characteristics (Table 1).

Fig. 1.

Flow chart of the study participants. Number of patients with urea/creatinine ratio data available in the first week of PICU stay are shown. Abbreviations: PICU: pediatric intensive care unit

Table 1.

Patients’ baseline characteristics

	Early-PN	Late-PN
n	665	663
Age (years), median (IQR)	1.4 (0.3–6.2)	1.5 (0.2–7.2)
Neonates (≤ 28 days), no. (%)	90 (13.5)	100 (15.1)
Infants (≤ 1 year), no. (%)	305 (45.9)	299 (45.1)
Older children (> 1 year), no. (%)	360 (54.1)	364 (54.9)
Male sex, no. (%)	376 (56.5)	382 (57.6)
Weight (kg), median (IQR)	10.0 (4.8–20.0)	10.0 (4.5–21.8)
Center, no. (%)
Leuven	364 (54.7)	373 (56.3)
Rotterdam	272 (40.9)	260 (39.2)
Edmonton	29 (4.4)	30 (4.5)
Admission diagnosis, no. (%)
Surgical
Abdominal	47 (7.1)	53 (8.0)
Burns	4 (0.6)	5 (0.8)
Cardiac	263 (39.6)	261 (39.4)
Neurosurgery	61 (9.2)	52 (7.8)
Thoracic	31 (4.7)	26 (3.9)
Transplantation	7 (1.1)	16 (2.4)
Orthopedic surgery–trauma	27 (4.1)	26 (3.9)
Other	16 (2.4)	23 (3.5)
Medical
Cardiac	25 (3.8)	24 (3.6)
Gastrointestinal–hepatic	2 (0.3)	4 (0.6)
Oncologic–hematologic	7 (1.1)	6 (0.9)
Neurologic	44 (6.6)	46 (6.9)
Respiratory	90 (13.5)	82 (12.4)
Other	41 (6.2)	39 (5.9)
Severity of illness, median (IQR)
PIM3 score a	-3.4 (-4.4 to -2.3)	-3.6 (-4.4 to -2.5)
PIM3 probability of death (%) b	3.3 (1.3–9.2)	2.7 (1.2–7.6)
PeLOD score first 24 h in PICU c	21 (11–31)	21 (11–31)
STRONGkids risk, median (IQR) d
Medium	590 (88.7)	594 (89.6)
High	75 (11.3)	69 (10.4)

^a Higher PIM3 scores indicate a higher risk of mortality. ^b PIM3 probability of death, ranging from 0% to 100%, with higher percentages indicating a higher probability of death in PICU. ^c PeLOD scores range from 0 to 71, with higher scores indicating more severe illness. ^d STRONGkids scores range from 0 to 5, with a score of 0 indicating a low risk of malnutrition, a score of 1 to 3 indicating a medium risk, and a score of 4 to 5 indicating a high risk

Abbreviations: IQR: interquartile range, PeLOD: pediatric logistic organ dysfunction score, PN: parenteral nutrition, PICU: pediatric intensive care unit, PIM3: pediatric index of mortality 3 score, STRONGkids: screening tool for risk on nutritional status and growth in children

Patients in the early-PN group had higher urea concentrations than those in the late-PN group on the first 11 days (Fig. 2A). This resulted in higher maximum urea concentrations in the first and second week in the PICU in early-PN as compared with late-PN patients, the latter despite a rise after day 7 in late-PN patients when parenteral nutrition was initiated as needed. Daily maximum creatinine was comparable for both groups, except for slightly lower values on day 2 in early-PN as compared with late-PN patients (Fig. 2B). Consequently, the daily maximum UCR in the early-PN group was significantly elevated as compared with the late-PN group throughout the first 12 days, with higher maximum UCR values in both the first and the second week (Fig. 2C).

Fig. 2.

Profiles of urea, creatinine and the urea/creatinine ratio in the first two weeks in the PICU. Daily maximum urea (A), age-normalized creatinine (B), and age-normalized urea/creatinine ratio (UCR) values (C) for the first 14 days in the PICU, as well as the corresponding maximum values in the first (day 1–7) and in the second week (day 8–14) are shown as mean and standard error of the mean. The blue shaded area indicates the randomized intervention time window (first PICU week). * indicates p < 0.05 between early-PN and late-PN. Abbreviations: PICU: pediatric intensive care unit, UCR: urea/creatinine ratio

Adjusting for baseline risk factors, randomization to early-PN as compared with late-PN was independently associated with a higher risk of acquiring a new infection in the PICU (p = 0.0063) and a lower likelihood of earlier live discharge from the PICU (p = 0.0068) and the hospital (p = 0.023), but not with 90-day mortality (p = 0.22) (Fig. 3, Additional File 1 Table S2). Adding the maximum urea concentration in the first week to these basic models revealed that a higher maximum urea concentration was independently associated with a higher risk of acquiring a new infection (p = 0.0011), a lower likelihood of earlier live PICU (p < 0.0001) and hospital discharge (p < 0.0001), and a higher likelihood of 90-day mortality (p < 0.0001). The higher maximum urea concentration hereby rendered the effect of randomization to early-PN versus late-PN insignificant for likelihood of earlier live PICU (p = 0.25) and hospital discharge (p = 0.33) and attenuated the effect size for risk of acquiring a new infection. The loss of significance for the randomized intervention, or the attenuation hereof, means that the stronger rise in urea with early-PN statistically explained (part of) the clinical harm induced by early-PN. When the maximum UCR in the first week was added to the basic models, a higher maximum UCR was independently associated with a higher risk of acquiring a new infection (p < 0.0001), a lower likelihood of earlier live PICU (p < 0.0001) and hospital discharge (p < 0.0001), and a higher likelihood of 90-day mortality (p = 0.033). The higher maximum UCR hereby rendered the effect of randomization to early-PN versus late-PN insignificant for risk of acquiring a new infection (p = 0.10), and likelihood of earlier live PICU (p = 0.44) and hospital discharge (p = 0.36). This thus also indicates that the increase in age-normalized UCR with early-PN statistically explained (part of) the clinical harm induced by early-PN. For development of new infection, similar results were obtained when its likelihood was studied in time-to-event analysis, and when penalizing for death in the PICU (Additional File 1 Table S2-3).

Fig. 3.

Increased ureagenesis as a (partial) explanation of the harm by early-PN vs. late-PN to clinical outcome. Stepwise multivariable logistic regression and Cox proportional hazard analyses were performed. All analyses were adjusted for center, age, type of illness, severity of illness (PIM3 and PeLOD scores) and severity of malnutrition risk (dichotomized STRONGkids score). Model A shows the independent impact of early-PN vs. late-PN on each of the indicated clinical outcomes. Model B additionally includes the first week’s maximum urea or UCR to assess potential statistical explanation of any harm by early-PN. Diamonds represent odds ratios or hazard ratios, wheras whiskers represent the corresponding 95% confidence intervals. Abbreviations: PeLOD: pediatric logistic organ dysfunction score, PICU: pediatric intensive care unit, PIM3: pediatric index of mortality 3 score, PN: parenteral nutrition, STRONGkids: screening tool for risk on nutritional status and growth, UCR: urea/creatinine ratio

Sensitivity analyses not excluding RRT patients (Additional File 1 Table S4) or excluding neonates (Additional File 1 Table S5) yielded similar results. Likewise, qualitatively similar results were obtained when the analyses were repeated in infants and older children (Additional File 1 Table S6), in boys and girls (Additional File 1 Table S7), and in patients with and without hypophosphatemia on day 2 (Additional File 1 Table S8), separately, though statistical power was reduced.

The average protein dose up to the day of maximum first week urea or UCR was associated with a higher risk and faster development of new infection and a lower likelihood of earlier live PICU and hospital discharge, and rendered the effect of early-PN vs. late-PN insignificant or inversed it (Additional File 1 Table S9, models C). When added to these models, the maximum first week urea or UCR was again independently associated with worse outcome (as in models B), but addition of this variable only slightly reduced the effect size of the harm associated with higher protein doses (Additional File 1 Table S9, models D). This suggests that the relative contribution of increased urea or UCR in explaining the adverse outcomes observed with early-PN may be small.

The maximum urea concentration in the first week showed an AUC of 0.734 for predicting a new infection (Additional File 1 Figure S2A) and of 0.794 for predicting 90-day mortality (Additional File 1 Figure S2B), both with a 39 mg/dL threshold. The maximum UCR in the first week showed an AUC of 0.652 for predicting a new infection, with a threshold of 35.8481 (Additional File 1 Figure S2C), and 0.573 for 90-day mortality, with a corresponding threshold of 49.3704 (Additional File 1 Figure S2D).

In the first week in the PICU, a higher proportion of early-PN than late-PN patients exceeded the maximum urea threshold for increased risk of new infection and 90-day mortality (37.7% vs. 27.6%, p < 0.0001), as well as the UCR threshold for increased risk of new infection (55.7% vs. 31.6%, p < 0.0001) and 90-day mortality (24.5% vs. 14.1%, p < 0.0001, Fig. 4A-C). These proportions rose further in the second week, remaining higher in early-PN than in late-PN patients with 49.4% vs. 35.5% (p = 0.012) for the urea threshold, 72.1% vs. 55.7% (p = 0.0025) for the UCR threshold for new infection, and 50.8% vs. 34.2% for the UCR threshold for mortality (p = 0.0034).

Fig. 4.

ROC curve-derived thresholds for maximum urea or UCR thresholds and risk of new infection or 90-day mortality of early-PN and late PN patients. ROC curve analyses determined cut-off points (“thresholds”) for urea and the UCR that predicted new infection or 90-day mortality with optimized specificity and sensitivity (Additional File 1 Table S2). Panels A-C show the proportions of patients randomized to early-PN or late-PN who exceeded the corresponding threshold for urea in relation to likelihood of new infection or 90-day mortality (A), or for the UCR in relation to likelihood of new infection (B) or 90-day mortality (C). Panels D-G show Kaplan–Meier curves of the likelihood of acquiring a new infection (D, F) and of 90-day mortality (E, G) according to these maximum urea (D, E) or UCR thresholds (F, G), in combination with randomization to early-PN or late-PN, with differences among groups assessed with the log-rank test and with multivariable analyses adjusting for risk factors. ^a Threshold for maximum urea in the first week T_urea=39 mg/dL for the likelihood of acquiring a new infection or 90-day mortality (A, D,E). ^b Threshold for maximum UCR in the first week T_UCR=35.8481 for likelihood of acquiring a new infection (B, F) and T_UCR=49.3704 for likelihood of 90-day mortality (C, G). Abbreviations: d1-7: day 1 to day 7, EPN: early parenteral nutrition, LPN: late parenteral nutrition, PICU: pediatric intensive care unit, T: threshold, UCR: urea/creatinine ratio. TUCR: threshold optimising speficity and sensitivity for maximum UCR in the first week, Turea: threshold optimizing speficity and sensitivity for maximum urea in the first week

Early-PN and late-PN patients who exceeded the first-week urea threshold showed a higher likelihood of acquiring a new infection and of 90-day mortality as compared with patients in the same group who did not exceed the threshold (overall multivariable p < 0.0001, Fig. 4D-E). Analyses for the UCR threshold revealed a similar pattern, both for likelihood of acquiring a new infection and likelihood of 90-day mortality (overall multivariable p < 0.0001 for new infection and p = 0.015 for 90-day mortality, Fig. 4F-G).

Discussion

In this secondary analysis of the PEPaNIC-RCT, early-PN as compared with late-PN increased the maximum urea concentration and UCR in the first and the second week in the PICU. A higher first week urea concentration or UCR with early-PN statistically explained (part of) the increased risk of new infection, and the prolonged PICU and hospital dependency observed with use of early-PN. The observed associations were consistent in sensitivity analyses not excluding RRT patients or excluding neonates, in infants and older children, in boys and girls, and in patients with or without early hypophosphatemia. Both in the first and second week, a higher proportion of early-PN than late-PN patients exceeded the identified thresholds of urea or UCR that predicted new infection risk or 90-day mortality.

Our findings are in line with studies performed in adult critically ill patients, which consistently found increased urea and UCR by early high protein doses [7, 16, 17]. As creatinine remained comparable between both groups, the increase in urea and the UCR reflected increased ureagenesis rather than impaired renal function. These findings point towards anabolic resistance, whereby increased amino acid doses are not used for net protein synthesis, but rather degraded into urea. Interestingly, also in the second PICU week, urea and the UCR remained higher in the early-PN than in the late-PN group. This may indicate that, even beyond the randomized intervention window, early-PN patients remain more resistant to anabolism than late-PN patients. Alternatively, the higher protein intake at the time of late initiation of PN may take a few days to affect urea or the UCR. Hence, a later advantage of omitting early-PN requires further study.

Apart from indicating anabolic resistance, we found that increased urea and UCR also statistically explained part of the harm by early-PN versus late-PN on outcome. These findings corroborate earlier data obtained in critically ill adults [7, 16, 17]. Indeed, in the adult EPaNIC-RCT, early-PN increased urea concentrations and the UCR as compared with late-PN from the first day of amino acid infusion onwards, with no, or only minor impact on creatinine or creatinine clearance profiles [7]. In EPaNIC, the rise in urea had been implicated in the prolonged need of RRT with early-PN. In the REDOXS trial, an increased UCR signaled deleterious effects of glutamine administration on mortality of adult ICU patients [16]. Similarly, a secondary analysis of the EFFORT Protein trial implicated rises in urea and the corresponding UCR in the increased risk of mortality that was observed in the subgroup of patients with acute kidney injury and high organ failure scores who received higher protein doses [17]. Altogether, these data put forward increased urea and UCR in response to enhanced amino acid/protein doses during critical illness as a potential marker of metabolic intolerance which could guide the timing and titration of nutrition. The exact mechanisms underlying the observed harm by early full feeding or high amino acid/protein doses as reflected by increased ureagenesis remain to be elucidated. In that regard, it remains unclear whether increased ureagenesis causes direct damage through the production of toxic metabolites, or whether an increase in urea or UCR is merely an indirect readout of metabolic intolerance whereas harm may be mediated by other concomitantly affected pathways such as suppressed autophagy [1, 25–27].

Harm by early enhanced feeding with early-PN has been documented across all age ranges, also in children who were malnourished upon PICU admission, which has led to more restrictive nutritional support in acutely ill patients [4, 5, 28, 29]. However, prolonged fasting likely becomes harmful at some point. It seems likely that the time when anabolic resistance ceases and prolonged fasting becomes maladaptive varies among patients. Therefore, reliable markers able to signal the presence of metabolic intolerance to nutrition are essential. The current study puts forward increased ureagenesis, as reflected by increases in urea concentrations and UCR upon increasing macronutrient doses, as such a biomarker for use in critically ill children, on top of the previously documented early hypophosphatemia upon increasing macronutrient doses [24]. Since urea and the UCR gave similar signals, it may be sufficient to monitor the crude urea levels in clinical practice, which is easier than monitoring the UCR, the latter requiring measurement of two analytes with age-normalization. If urea rises upon initiation or increase of nutrition, this may indicate that the exogenous amino acid supply exceeds the patient’s anabolic capacity, resulting in futile amino acid catabolism that further disrupts homeostasis. A rise in urea may thus require reducing the nutritional intake or at least advise against further increasing macronutrient doses. Urea and UCR are mostly readily available from routine measurements, which is a major advantage. However, they can also be affected by dehydration, gastrointestinal bleeding, and kidney and liver dysfunction [15], and validated thresholds for defining metabolic intolerance are lacking. Moreover, endogenous catabolism due to prolonged unnecessary fasting may also increase urea and UCR. Hence, urea and UCR are to be seen as reactive rather than predictive biomarkers of metabolic intolerance, which implies that dynamic changes are more informative than a focus on fixed thresholds. Ideally a predictive marker would be even more helpful to avoid premature exposure to excessive amino acid/protein provision as well as reduce risks of prolonged underfeeding. We here identified urea and UCR thresholds above which risk of new infection and 90-day mortality increased, irrespective of randomization to early-PN or late-PN, though discriminative power was very limited. This suggests that the urea/UCR response to nutritional support may need to be integrated in conjunction with other potential biomarkers, requiring further study [24].

Strengths of this study include the large sample size, its embedding in a multicenter RCT with detailed prospective data collection, and use of age-normalized UCR values. Furthermore, it provides the first evidence in critically ill children to support nutrition-induced increases in ureagenesis as a biomarker of metabolic intolerance. Nevertheless, the study has limitations. Although we adjusted for baseline risk factors, we cannot exclude residual confounding. For the primary analyses, we excluded patients receiving RRT as this can modulate urea, creatinine, and UCR trajectories. A sensitivity analysis including these patients yielded similar results. However, the number of RRT patients was very low and hence the results may not be extrapolated to this population. Splitting up the patients in subgroups for sensitivity analyses regarding age, sex and hypophosphatemia yielded qualitatively similar results as for the total patient population but reduced statistical power. Although data were missing for only 6% of patients, we cannot fully exclude potential selection bias. We studied increased urea and UCR as marker of increased ureagenesis. It remains unclear whether increased urea cycle activity by itself explains the harm and (in such case) which metabolites are the culprit, or whether increased ureagenesis is a proxy for more prolonged or more severe organ failure. Moreover, the relative contribution of increased urea or UCR in explaining the adverse outcomes observed with early-PN appeared to be rather small, as suggested by the analyses including average protein doses in the models and by the small difference in outcome between early-PN patients exceeding the identified urea or UCR thresholds for harm and late-PN patients exceeding these thresholds. Nevertheless, although a higher protein dose likely causes harm by other mechanisms as well and high doses of glucose and lipids could theoretically also be harmful, this does not preclude the use of urea or UCR as a marker of metabolic intolerance to enhanced artificial nutrition. The mediation, ROC and survival analyses focused on the maximum urea and UCR in the first week in the PICU, precluding extrapolation to individual time points within and beyond this time window. Also, the ROC-derived thresholds were tested for association with outcomes in the same dataset, warranting external and prospective validation and should not yet be interpreted as clinically actionable. Finally, the use of the first week’s maximum urea or UCR ignores information on the trajectories of urea or the UCR over time in the PICU, whereas their evolution over time was previously shown to also be a determinant of outcome [17].

Conclusions

In this secondary analysis of the PEPaNIC RCT, an increase in plasma/serum urea and UCR with early-PN statistically explained (a small) part of the harm by early-PN as compared with late-PN, suggesting that these markers may identify critically ill children who are metabolically intolerant to enhanced artificial nutrition. This finding opens perspectives for further developing a metabolic monitoring strategy to help defining optimal timing and dosing of nutritional support for critically ill children, the clinical value of which requires future prospective validation studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

aHR

Adjusted hazard ratio

aOR

Adjusted odds ratio

CI

Confidence interval

EPaNIC

Early versus Late Parenteral Nutrition in Critical Illness

ICU

Intensive care unit

PeLOD

Pediatric logistic organ dysfunction score

PEPaNIC

Pediatric Early versus Late Parenteral Nutrition in Critical Illness

PICU

Pediatric intensive care unit

PIM3

Pediatric index of mortality 3 score

PN

Parenteral nutrition

RCT

Randomized controlled trial

STRONGkids

Screening tool for risk on nutritional status and growth

UCR

Urea/creatinine ratio

Author contributions

IV, GVdB and JG designed the study. All authors gathered data. IV and NUS analyzed the data. IV, NUS, GVdB and JG interpreted the data. NUS, IV, GVdB and JG wrote the manuscript, which was reviewed and approved by all authors.

Funding

This work was supported by European Research Council Advanced Grants (AdG-2012-321670 from the Ideas Program of the European Union 7th framework program, AdG-2017-785809 from the Horizon 2020 Program and AdG-2023-101133276 from the Horizon Europe Program to GVdB - Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them); by the Methusalem Program of the Flemish government (through the University of Leuven to GVdB and IV, METH14/06); by the Institute for Science and Technology, Flanders, Belgium (through the University of Leuven to GVdB, IWT-TBM150181 and IWT-TBM110685); by the Research Foundation-Flanders (senior clinical investigator fellowship 1842724 N to JG, and G029525N research grant to JG); by KU Leuven (starting grant STG/23/032 to JG); by the Sophia Research Foundation (SSWO to SV); by the Stichting Agis Zorginnovatie (to SV); by the Erasmus Trustfonds (to SV); and by an European Society for Clinical Nutrition and Metabolism (ESPEN) research grant (to SV).

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due ethical constraints. Data sets and other documents may be shared on reasonable request under the format of future collaborative projects that further elaborate on this specific research topic. Proposals for collaborative projects must be directed to the corresponding author and will be considered for approval. The dataset required to address the research question will only be shared after signing of a data access agreement. Data are located in controlled access data storage at KU Leuven.

Declarations

Ethics approval and consent to participate

The participating sites’ institutional review boards approved the original and follow-up studies (Ethical Committee Research UZ/KU Leuven: ML8052; Medische Ethische Toetsingscommissie Erasmus MC: NL49708.078; Ethical Committee Stollery’s Children Hospital: Pro00038098), which were performed in accordance with the 1964 Declaration of Helsinki and its amendments. Written informed consent was acquired from parents or legal guardians.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.