F1Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research Cerebral fuels within the first week of life in very preterm infants: a cohort study Gordon Xin Hua Liu  ‍ ‍ ,1 Loredana Marcovecchio  ‍ ‍ ,2,3 K Beardsall  ‍ ‍ ,2,3 REACT Collaborative To cite: Liu GXH, Marcovecchio L, Beardsall K, et al. Arch Dis Child Fetal Neonatal Ed Epub ahead of print: [please include Day Month Year]. doi:10.1136/ archdischild-2025-328701 ► Additional supplemental material is published online only. To view, please visit the journal online (https://​doi.​org/​ 10.​1136/​archdischild-​2025-​ 328701). 1School of Medicine, University of Auckland—Grafton Campus, Auckland, New Zealand 2Department of Paediatrics, University of Cambridge, Cambridge Biomedical Campus, Cambridge, UK 3Cambridge University Hospitals NHS Foundation Trust, Cambridge Biomedcal Campus, Cambridge, UK Correspondence to Dr K Beardsall; ​kb274@​cam.​ac.​uk Received 25 February 2025 Accepted 23 July 2025 © Author(s) (or their employer(s)) 2025. Re-use permitted under CC BY. Published by BMJ Group. ABSTRACT Background  Ketones and lactate may contribute towards overall cerebral fuel availability in term infants, yet the availability of such cerebral fuels in very preterm infants is unclear. We undertook a prespecified substudy to explore ketone and lactate concentrations in the first week of life in infants recruited to the REACT trial (real- time continuous glucose monitoring in the newborn): an international multicentre randomised controlled trial of 182 very low birth weight infants investigating the use of continuous glucose monitoring in glycaemic care. Methods  Ketone and lactate measurements were prospectively collected over the first week of life using the Nova Biomedical point-of-care meter. A longitudinal analysis was undertaken to explore lactate and ketone concentration trends across time and their relationships with blood glucose, baseline demographics, nutritional support and insulin treatment. Results  Data were available for 168 infants (85 females) including 2902 blood glucose, 2084 ketone and 2017 lactate samples. The mean (SD) gestational age was 27.4 (2.0) weeks. Lactate concentrations were higher initially, with mean (SD) 1.72 (1.26) mmol/L on day 2 and lowered to 1.19 (1.1) mmol/L on day 7. Ketone concentrations remained consistently low at 0.1 mmol/L. Neither simultaneous blood glucose concentrations, macronutrient intake nor receipt of insulin was consistently related to ketone or lactate concentrations. Conclusion  In this cohort of very preterm infants, there were persistently low concentrations of ketones and relatively higher concentrations of lactate throughout the first week of life. Future research should evaluate changes in these metabolites during episodes of acute hypoglycaemia or hyperglycaemia over more prolonged periods of neonatal intensive care. INTRODUCTION The neonatal period is characterised by significant glycaemic transition, driven by a cessation of exog- enous glucose supply following cord cutting and the subsequent activation of glycogenolysis and gluco- neogenesis to maintain blood glucose concentra- tions.1 2 These processes rely on mature enzymatic cascades and sufficient hepatic glycogen stores,3 4 and it can take up to 4 days in healthy term infants before adult glucose concentrations are reached. Ketogenesis and lactogenesis can supplement energy needs, particularly during times of increased energy demand or within hypoxic/glucose-deprived environments,5–9 and both ketones and lactate may contribute towards overall cerebral fuel availability in healthy term infants during the first 4 days of life.7 Prematurity (defined as <37 weeks gestational age) is a well-known risk factor for metabolic dysfunction.3 4 10 Preterm infants are at higher risk for both hypoglycaemia and hyperglycaemia compared with term counterparts secondary to hypoactive gluconeogenic pathways, inadequate glycogen reserves and abnormal insulin profiles.3 4 Moreover, physiological immaturity and instability, concomitant illness and painful procedures may further escalate energy demands, putting preterm infants into energy-deficient states that may adversely impact long-term outcomes.5 Although previous studies have attempted to characterise neonatal ketone and lactate profiles during the first few days of life,7 8 11–15 most are WHAT IS ALREADY KNOWN ON THIS TOPIC ⇒ During times of glycaemic instability, ketones and lactate may act as alternative cerebral fuels in healthy term infants, yet it is unclear if this occurs in preterm infants. We conducted a longitudinal analysis of very low birth weight infants over the first week of life exploring ketone and lactate trends and their association with blood glucose, baseline demographics, nutritional support and insulin treatment. WHAT THIS STUDY ADDS ⇒ We found that, in this cohort of very preterm infants, ketone concentrations remained persistently low while lactate concentrations were comparatively higher, which may reflect insufficient substrate availability and increased metabolic stress. There was no relationship between simultaneous blood glucose concentrations, increasing macronutrient support or receipt of insulin with ketone or lactate concentrations, although greater weight gain across the first week of life was associated with slightly greater lactic and ketogenic responses in later periods. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY ⇒ This study provides insight into the presence of potential alternative cerebral fuels within a vulnerable population, and future studies should evaluate ketone and lactate profiles in preterm infants during periods of glucose instability. http://www.rcpch.ac.uk/ http://fn.bmj.com/ http://orcid.org/0000-0002-4261-9683 http://orcid.org/0000-0002-4415-316X http://orcid.org/0000-0003-3582-183X https://doi.org/10.1136/archdischild-2025-328701 https://doi.org/10.1136/archdischild-2025-328701 https://doi.org/10.1136/archdischild-2025-328701 https://crossmark.crossref.org/dialog/?doi=10.1136/fetalneonatal-2025-328701&domain=pdf&date_stamp=2025-010-14 F2 Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research cross-sectional or have limited sample sizes and few have eval- uated the impact of feeding or other nutritional parameters. Therefore, we aimed to explore glucose, ketone and lactate concentrations in a cohort of very preterm infants requiring intensive care in the first week of life. This was a prespecified substudy of infants recruited to the REACT trial (real-time continuous glucose monitoring in the newborn) with specific aims to explore trends over time, relationship to simultaneous glucose concentrations and impact of baseline demographics and nutritional support on plasma ketone and lactate concentrations. METHODS Population Data were collected as part of the REACT trial, which was an international, parallel-group, open-label randomised controlled trial (RCT). It was conducted in 13 neonatal intensive care units (NICUs) across the UK, Spain and the Netherlands between 2016 and 2019. Included infants had a birth weight of ≤1200 g, gestational age up to 33 weeks plus 6 days and were recruited within 24 hours of birth. Any infants with lethal congenital malformations or congenital metabolic disorders were excluded. The protocol and findings for the REACT trial are available elsewhere.16 17 In brief, infants were randomised to receive glycaemic care that was informed by either real-time continuous interstitial glucose monitoring (CGM) (intervention) or standard care (control) in the first week of life. Infants in the intervention group were managed according to a tailored clinical manage- ment guideline based on the CGM data, which was read and recorded hourly, and advised that blood glucose concentrations should be measured at a minimum of every 12 hours or when there were rapid changes in CGM values or if CGM values fell to <4 mmol/L. Infants in the control group were managed according to local standard clinical practice with intermittent blood glucose concentrations guiding clinical management. All other aspects of clinical care, including nutritional intake, were at the discretion of local staff. All participants also had β-hy- droxybutyrate (BOHB) and lactate concentrations measured using Nova Biomedical (Waltham, Massachusetts, USA) point- of-care (POC) devices at least once in a 24-hour period at the same time as blood glucose monitoring. Data collection A case report form was completed at the end of each study day for the first week of life, which recorded details of clinical care including level of care, respiratory support, medications, nutri- tional intake and safety events. Weight was measured at birth and 7 days follow-up. SD scores (SDS) were calculated using the LLMS method18 according to the UK-WHO growth charts for preterm infants (British 1990 reference data, reanalysed 2009).19 Assay methodology Plasma glucose, BOHB and lactate measurements were performed using POC bedside metres (StatStrip Glucose/ Ketone and StatStrip Lactate; Nova Biomedical). The StatStrip Glucose/Ketone metre requires 1.2 µL and 0.8 µL of blood for glucose and BOHB measurements, respectively, with measure- ment ranges of 0.6–33.3 mmol/L (glucose) and 0.0–7.0 mmol/L (BOHB).20 Coefficients of variability (CV) range from 4% (at 33.3 mmol/L) to 8% (at 2.8 mmol/L) for glucose, and 5% (at 7.0 mmol/L) to 7% (at 3.0 mmol/L) for BOHB. The StatStrip Lactate metre requires 0.6 µL of blood and has a measurement range of 0.3–20.0 mmol/L,21 with CV ranging from 1.7% (at 16.83 mmol/L) to 9.1% (at 0.76 mmol/L). Statistical analysis Data are presented as n (%), mean (SD), and median (IQR) and were analysed in 24 hours epochs. Prior to performing our main analysis, we analysed BOHB and lactate concentrations between the intervention and control arms of the REACT trial to ensure that these values could be combined into a single cohort for analysis (online supplemental table 1). This analysis showed no significant difference across groups. Glucose and lactate values were log(e)-transformed to near-normal distributions. BOHB values were predominantly left-skewed in our data set and so were categorised into ‘low’ (BOHB≤0.1 mmol/L) and ‘high’ (BOHB>0.1 mmol/L) subgroups during analysis. To account for repeated measures, analyses were conducted using generalised linear mixed regres- sion models, which assumed normal distributions for glucose and lactate concentrations and binary logistic distributions for BOHB concentrations. Where appropriate, Holm-Bonferroni adjustment was used for multiple comparisons. When assessing for differences in glucose, BOHB and lactate concentrations across epochs, day two was prespeci- fied as the reference epoch. This was because we considered it to be the closest marker of baseline metabolite concen- trations prior to significant clinical intervention, being the earliest time point when all participants had completed study enrolment and were consistently having metabolite concentrations measured. We assessed the relationship between BOHB and lactate against glucose concentrations by comparing BOHB and Table 1  Maternal and neonatal baseline demographics Demographic n (N=168) Sex—female 168 85 (51) Multiple birth 168 41 (24) Ethnicity 168  � White 121 (72)  � Black 13 (8)  � Asian 17 (10)  � Mixed 10 (6)  � Other 7 (4) Gestational age (weeks) 168  � Mean (SD) 27.4 (2.0)  � Range 23.3–33.7 Received maternal steroids 168 129 (77) Mode of delivery 168  � Spontaneous vaginal delivery 57 (34)  � Caesarean section before onset of labour 74 (44)  � Caesarean section after onset of labour 35 (21)  � Assisted vaginal delivery 2 (1) Birth weight (g) 168 890 (171) Birth weight SD score 168 −0.85 (0.92) Birth length (cm) 115 33.1 (2.5) Head circumference (cm) 153 24.3 (1.8) Apgar score  � 1 min 160 5.39 (2.4)  � 5 min 161 7.64 (1.96) Base excess at birth 165 −6.16 (5.2) Intubation at birth 162 105 (65) Values are n(%), mean (SD) or range. https://dx.doi.org/10.1136/fetalneonatal-2025-328701 F3Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research lactate concentrations when simultaneous plasma glucose concentrations were in the lower, middle or upper tertile. Glucose tertile ranges were assigned based on the 33rd and 67th percentiles of all plasma glucose concentrations across the first week of life. We also compared BOHB and lactate concentrations across prespecified subgroups based on baseline demographics: sex (male vs female), gestation (<27 weeks vs ≥27 weeks), birth weight SDS (<−0.85 vs ≥−0.85), base excess at birth (<−6.16 vs ≥−6.16), intu- bation at birth (yes vs no) and receipt of maternal steroids (yes vs no). To explore the impact of nutritional support, we calcu- lated the cumulative total dose per kilogram of birth weight for carbohydrate, amino acid, lipid and breast milk supple- mentation and change in weight SDS score from birth to 7 days, and compared BOHB and lactate concentrations across cumulative doses of nutritional supplementation and weight change tertiles. The middle tertile was desig- nated as the reference tertile. Preplanned analyses across formula supplementation tertiles were not possible due to the limited number of participants who received formula supplementation. We analysed the effect of exogenous intravenous insulin dose by comparing BOHB and lactate concentrations against cumu- lative insulin dose tertiles, which were calculated by summating the total dose of insulin given across the first week of life and then categorising into lower, middle and high tertiles. We further assessed whether the effect of insulin varied according to age by conducting univariable analyses that compared BOHB and lactate concentrations across participants who did or did not receive insulin within a specific epoch. In epochs where a significant effect was demonstrated, multivariable analyses were then performed, with adjustment for the following pre-specified confounders: gestation, weight SDS from birth to 7 days, base excess at birth, intubation at birth, maternal steroids and simul- taneous plasma glucose concentration. We estimated total potential ATP availability from cerebral fuel metabolism by assigning an ATP equivalent to glucose (31), BOHB (21.5) and lactate (15).7 Analysis was conducted in R (V.4.3.1; R Core Team 2023). Percentile curves were constructed using the purrr and ggplot packages, and regression analysis using the lmerTest and glmer packages. RESULTS Of the 360 infants who were eligible for the REACT study, 180 (51%) were randomised, of which 168 (92%) had glucose, BOHB and/or lactate data recorded (n=167 glucose, 162 BOHB and 161 lactate). In total, we had 2902 glucose, 2084 BOHB and 2017 lactate samples, with the median (IQR) samples per infant being 16 (12–22) for glucose, 12 (8–16) for BOHB and 13 (8–16) for lactate. For this cohort, the mean (SD) gestational age and birth weight were 27.4 (2.0) weeks and 890 (171) g, respectively. Half of our cohort were girls (n=85, 51%) and three-quarters were single- tons (n=127, 76%). Delivery was by Caesarean section in 109 (65%) of infants and two-thirds of infants (n=105, 65%) were intubated at birth. The mean base excess at birth was −6.16. Further details on baseline demographics are provided in table 1. Time course of glucose, BOHB and lactate concentrations The time course of glucose, BOHB and lactate concentration centiles is provided in figure 1. Plasma glucose concentrations were lower on day 2 (mean 6.5 mmol/L) but were consistently higher from days 3 to 7 (p<0.05), with mean concentrations ranging from 7.1 to 7.6 mmol/L (table 2). Only 24/2902 (1%) samples were in the hypoglycaemic range (<2.6 mmol/L). However, hyperglycaemia (defined as plasma glucose concen- tration >8.0 mmol/L) was common, with rates ranging from 21–37% per epoch. Lactate concentrations were higher on day 2 (mean 1.72 mmol/L) before falling steadily from days 3 to 7 (p<0.05), reaching mean values of 1.19 mmol/L on days 6 and 7. BOHB concentrations were low across all epochs, with the proportion of low BOHB readings (defined as BOHB≤0.1 mmol/L) increasing in later epochs (65% vs 75% on days 2 vs 5, respectively, p<0.05). Similar trends in glucose, BOHB and lactate concentrations across epochs were identified when we separated our cohort into intervention and control groups as per the original REACT trial (online supplemental table 2). BOHB and lactate concentrations in relation to glucose concentrations and baseline demographics Lactate and BOHB concentrations were not consistently related to simultaneous glucose concentrations (table  3). Females and infants whose mothers had received steroids had higher propor- tions of low BOHB concentrations compared with males or infants who had not received maternal steroids (75% vs 65%, Figure 1  Percentile curves of plasma glucose, β-hydroxybutyrate and lactate concentrations across the first 8 days of life https://dx.doi.org/10.1136/fetalneonatal-2025-328701 F4 Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research respectively, for sex, p=0.027; 72% vs 64%, respectively, for maternal steroids, p=0.038) (online supplemental table 3a). Meanwhile, mean lactate concentrations were higher in infants with lower versus higher birth weight SDS (1.59 vs 1.30 mmol/L, respectively, p=0.001) and a higher versus lower base excess at birth (1.59 vs 1.26 mmol/L, p<0.001) (online supplemental table 3b). Impact of insulin and nutritional supplementation on BOHB and lactate concentrations There was no consistent relationship between lactate and BOHB concentrations against receipt of insulin, both when analysed as cumulative dose tertiles (online supplemental table 4a and b) or when stratified across different epochs (table 4). When compared against the middle tertile, higher cumulative volumes of breast milk were associated with a higher propor- tion of low BOHB concentrations (70% vs 75%, respectively) and higher lactate concentrations (1.28 vs 1.59 mmol/L, respec- tively, p<0.05) (online supplemental table 4a and b). Conversely, greater increases in weight SDS were associated with a lower proportion of low BOHB concentrations when compared with the middle tertile (67% vs 74%, respectively, p<0.05) and higher lactate concentrations (1.57 vs 1.34 mmol/L, p<0.05). There was no relationship between BOHB and lactate concentrations with carbohydrate, amino acid or lipid supplementation dose. ATP contribution or glucose, BOHB and lactate Glucose was the predominant ATP source across all epochs, contributing between 90% and 95% (online supplemental table 5). Lactate initially provided 8–9% of overall ATP during days 1–2, although this halved to 4–5% by days 7–8. BOHB comprised approximately 1% of total ATP across epochs. DISCUSSION This study is the first to characterise the concentrations of BOHB and lactate in very preterm infants across the first week of life. Our findings indicate that very preterm infants do not have significant circulating concentrations of BOHB or lactate during times of significant physiological stress and must there- fore rely predominantly on glucose as their primary energy source. Increasing macronutrient support or receipt of insulin was not associated with different BOHB or lactate concentra- tions, although greater weight gain may influence BOHB and lactate concentrations in later stages. We also found that female sex, lower birth weight, lower base excess at birth and receipt of maternal steroids were associated with higher concentrations of BOHB or lactate. The neonatal brain derives approximately 70% of its energy needs from glucose.22 Ketones and lactates are produced via β-ox- idation of free fatty acids or anaerobic fermentation of glucose respectively23 24 and, alongside other roles as signalling mole- cules,24–27 can be used to supplement energy needs during times of increased demand or insufficient supply by acting as oxida- tive substrates. As such, ketones and lactate may play important roles during times of stress or disturbed glucose metabolism,27–29 such as following traumatic brain injury or periods of fasting.30 31 Previous research has suggested that extremely preterm infants are prone to states of relative insulin deficiency,32 secondary to both reduced secretion of and sensitivity to insulin, particu- larly in the immediate postnatal phase when there is increased catabolic stress.33 These changes are partially demonstrated in our cohort, where the prevalence of hypoglycaemia and hyper- glycaemia was low and high respectively, likely reflecting the care received in NICU. Our finding that receipt of insulin was not associated with significant differences in lactate or BOHB concentrations further reflects the underlying insulin-resistant state that persisted in our cohort across the immediate postnatal period. The metabolic frailty of our cohort is further exemplified when BOHB and lactate concentrations are compared against previously reported data from term counterparts. Overall, our cohort had consistently low concentrations of BOHB with rela- tively higher concentrations of lactate. The Glucose in well babies (GLOW) study, which analysed alternative fuel concentrations in healthy term babies across the first 5 days of life, reported similar median concentrations of lactate (1.9 mmol/L in day 1 to 1.4 mmol/L in day 5) but significantly higher BOHB concen- trations (ranging from 0.1 to 0.7 mmol/L).7 Given our cohort’s extreme prematurity and low birth weight, it is probable that our cohort lacked sufficient energy stores to mount appropriate ketogenic responses (and therefore had low BOHB concen- trations), whereas the relatively higher lactate concentrations Table 2  Plasma concentrations of glucose, BOHB and lactate across different postnatal ages Postnatal age (d) Glucose (mmol/L) BOHB (mmol/L) Lactate (mmol/L) Mean (SD) n (%) of glucose readings <2.6 mmol/L n (%) of glucose readings >8.0 mmol/L N samples Median (IQR) BOHB ≤0.1 mmol/L BOHB >0.1 mmol/L N samples Mean (SD) N samples 1 7.1 (3.1)* 5 (3) 53 (29) 184 0.1 (0.1–0.2) 85 (66) 43 (34) 128 2.13 (1.46)* 126 2† 6.5 (3.0) 15 (3) 128 (23) 548 0.1 (0.1–0.2) 245 (65) 130 (35) 375 1.72 (1.26) 376 3 7.4 (2.8)* 5 (1) 167 (33) 501 0.1 (0.1–0.2) 243 (66) 124 (34) 367 1.51 (1.38)* 363 4 7.6 (2.6)* 2 (0.4) 170 (37) 461 0.1 (0.1–0.2) 233 (69) 104 (31) 338 1.33 (0.66)* 332 5 7.3 (2.4)* 0 (0) 148 (33) 457 0.1 (0.1–0.2) 246 (74)* 85 (26) 332 1.22 (0.51)* 332 6 7.2 (2.5)* 0 (0) 129 (31) 418 0.1 (0.1–0.2) 221 (72) 86 (28) 306 1.19 (0.50)* 302 7 7.1 (2.6)* 2 (1) 89 (30) 291 0.1 (0.1–0.2) 162 (74) 56 (26) 216 1.19 (1.1)* 217 8 6.5 (2.0) 0 (0) 9 (21) 42 0.1 (0.1–0.1) 21 (95)* 1 (5) 22 1.13 (0.51)* 23 Values are n (%), mean (SD) or median (IQR). P values are for comparisons of mean glucose and lactate values across postnatal age epochs using univariable linear mixed modelling or comparisons of low and high ketone values across postnatal age epochs using univariable binary logistic mixed modelling, reference epoch = ‘2’. Glucose and lactate values were log-transformed to near-normal distributions. *p<0.05. †Reference epoch. BOHB, β-hydroxybutyrate. https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 https://dx.doi.org/10.1136/fetalneonatal-2025-328701 F5Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research reflected greater degrees of metabolic stress and anaerobic respi- ration. This is supported by our finding that lower birth weight and a more negative base excess at birth, reflective of poorer metabolic condition, were both associated with higher lactate concentrations and is consistent with previous research that has suggested preterm babies as being at higher risk for hypogly- caemic injury secondary to insufficient fat and protein stores.3 6 Interestingly, we found that greater weight gain from birth to 7 days was associated with higher lactate and ketone concentra- tions, which may reflect how our infants had had more time by then to generate sufficient energy reserves and increase their sensitivity towards exogenous insulin. However, we did not find any association between greater macronutrient support and higher concentrations of BOHB or lactate, suggesting that this slight increase in ketogenic and lactic response may not be attrib- utable to a specific nutrient source, but is related to a general accretion of tissue and body mass. Alternatively, our observed low BOHB concentrations and lack of effect of nutritional supplementation may also be related to the relatively immaturity of enzymatic pathways in preterm infants compared with term counterparts, which may limit the rate or capacity for ketogen- esis and lactogenesis to be performed. To our knowledge, only two previous studies have assessed alternative fuel concentrations in preterm infants.8 13 Hawdon et al13 found that, during the first postnatal week, very preterm infants had low concentrations of ketone bodies compared with term counterparts (mean 0.04 vs 0.20 mmol/L, respectively), and that lower gestational age (particularly before 36 weeks) was a significant predictor of reduced ketone body concentrations, even after adjustment for birth size, feeding variables and blood glucose concentrations. In contrast to our findings, they also reported that increased volume of enteral feeding was positively correlated with ketone body concentrations. However, their cross-sectional analysis only included clinically stable preterm participants, who likely had more robust energy reserves and were subjected to less metabolic stress than our cohort of unwell infants. This study employed recent developments in POC tech- nology to allow bedside collection of small samples in a vulner- able population. A recent systematic review assessing the utility of POC BOHB measurements in emergency departments for people without diabetes found that most of the available data relates to children with gastroenteritis, and that there is limited information that is specific for other paediatric populations.34 Nonetheless, these metres are widely used clinically,34 and POC BOHB measurements have been previously validated in unwell preterm and term infants.35 Further, their ease of use and low volume of blood required allows for repeated samples in infants with limited circulating volumes and enables clinicians to serially monitor metabolites that would otherwise only be intermittently measured. Limitations of this study include the limited data at low blood glucose concentrations, which may contribute towards lower BOHB and lactate concentrations, and that we had relatively fewer samples during the first 24 hours of life, which has been recognised as a time of significant metabolic transition.2 It is also important to recognise that our data was collected as part of a larger RCT assessing glycaemic control, which may alter clinician behaviour and restrict generalisability of our findings towards standard clinical care. Strengths of our study include its multicentric nature, longitudinal design spanning several days, large sample size and detailed collection of clinical care and nutritional data. In conclusion, we report that very and extremely preterm infants in NICU have low concentrations of BOHB with Ta bl e 3  Pl as m a co nc en tr at io ns o f B O HB w he n si m ul ta ne ou s gl uc os e co nc en tr at io ns w er e in lo w er , m id dl e or u pp er te rt ile fo r t ha t e po ch A ge (d ) G lu co se w it hi n te rt ile 1 (0 .9 –5 .7  m m ol /L ) G lu co se w it hi n te rt ile 2 (5 .7 –7 .9  m m ol /L )* G lu co se w it hi n te rt ile 3 (7 .9 –2 4. 1  m m ol /L ) Ke to ne s La ct at e (m m ol /L ) Ke to ne s La ct at e (m m ol /L ) Ke to ne s La ct at e (m m ol /L ) BO H B ≤0 .1  m m ol /L BO H B > 0. 1  m m ol /L N s am pl es M ea n (S D ) N s am pl es BO H B ≤0 .1  m m ol /L BO H B > 0. 1  m m ol /L N M ea n (S D ) N s am pl es BO H B ≤0 .1  m m ol /L BO H B > 0. 1  m m ol /L N s am pl es M ea n (S D ) N s am pl es 1 33 (6 7) 16 (3 3) 49 2. 12 (1 .4 1) 48 32 (7 0) 14 (3 0) 46 2. 15 (1 .7 6) 42 20 (6 1) 13 (3 9) 33 2. 03 (1 .0 6) 32 2 11 7 (6 4) 66 (3 6) 18 3 1. 36 (0 .6 3) † 18 0 71 (7 0) 30 (3 0) 10 1 1. 67 (0 .9 6) 99 56 (6 4) 32 (3 6) 88 2. 45 (2 .0 0) 91 3 80 (7 3) 29 (2 7) 10 9 1. 31 (0 .7 6) 10 6 84 (6 7) 42 (3 3) 12 6 1. 45 (1 .0 8) 12 5 78 (6 1) 49 (3 9) 12 7 1. 71 (1 .9 6) 12 5 4 47 (7 3) 17 (2 7) 64 1. 19 (0 .5 7) 70 10 2 (7 0) 43 (3 0) 14 5 1. 20 (0 .4 5) 13 9 79 (6 5) 42 (3 5) 12 1 1. 58 (0 .8 5) † 11 6 5 68 (7 9) 18 (2 1) 86 1. 16 (0 .4 6) 92 97 (7 4) 34 (2 6) 13 1 1. 22 (0 .5 6) 12 9 70 (6 9) 32 (3 1) 10 2 1. 29 (0 .5 0) 97 6 81 (7 6) 26 (2 4) 10 7 1. 09 (0 .5 2) † 10 4 69 (6 6) 35 (3 4) 10 4 1. 23 (0 .4 2) 10 1 68 (7 4) 24 (2 6) 92 1. 27 (0 .5 4) 92 7 64 (8 0) 16 (2 0) 80 1. 03 (0 .5 2) 79 45 (7 4) 16 (2 6) 61 1. 41 (1 .8 2) 61 48 (6 9) 22 (3 1) 70 1. 18 (0 .7 1) 67 8 8 (8 9) 1 (1 1) 9 1. 02 (0 .4 8) 10 9 (1 00 ) 0 (0 ) 9 1. 14 (0 .6 4) 9 3 (1 00 ) 0 (0 ) 3 1. 37 (0 .2 1) 3 Va lu es a re n (% ) o r m ea n (S D) . L ac ta te v al ue s w er e lo g- tr an sf or m ed to n ea r-n or m al d is tr ib ut io n du rin g an al ys is. G lu co se te rt ile ra ng es w er e as si gn ed b as ed o n th e 33 rd a nd 6 7t h pe rc en til es o f a ll pl as m a gl uc os e co nc en tr at io ns a cr os s th e fir st w ee k of li fe . P v al ue s ar e fo r c om pa ris on s ac ro ss g ro up s w ith in e ac h sp ec ifi c ep oc h, w ith g lu co se te rt ile 2 b ei ng d es ig na te d as th e re fe re nc e te rt ile a nd w er e ob ta in ed u si ng s ep ar at e un iv ar ia bl e bi na ry lo gi st ic m ix ed m od el s fo r a na ly si s of k et on e co nc en tr at io ns a nd li ne ar m ix ed m od el s fo r a na ly si s of la ct at e co nc en tr at io ns . P v al ue s ar e ad ju st ed fo r m ul tip le c om pa ris on s us in g Ho lm -B on fe rr on i c or re ct io n. *R ef er en ce te rt ile . n re fe rs to th e nu m be r o f B O HB o r l ac ta te s am pl es w ith in th at s pe ci fic e po ch a nd g lu co se te rt ile . †A dj us te d p< 0. 05  w he n co m pa re d ag ai ns t l ac ta te v al ue in g lu co se te rt ile 2 . BO HB , β -h yd ro xy bu ty ra te . F6 Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research relatively higher lactate concentrations, which may reflect insuf- ficient substrate availability and increased metabolic stress. We did not find any effect of increased macronutrient support, breastfeeding or receipt of insulin on lactate or BOHB concentra- tions, although weight gain appeared to increase ketogenic and lactic responses in later periods. Future studies should continue to explore the metabolic profiles of this vulnerable population during periods of glycaemic instability (eg, hypoglycaemias or hyperglycaemias) and whether their trajectories can be altered using clinical interventions. X Loredana Marcovecchio @l_marcovecchio Acknowledgements  We would like to acknowledge Laurentyna Olga for her assistance with calculating the birthweight and 7-day weight standard deviation scores and Professor Ken Ong for his invaluable input during the statistical analysis process. We would also like to thank the REACT investigators and staff as well as the families who took part in the original REACT trial, as well as the clinical teams who worked in the tertiary neonatal units, and would also like to recognise and acknowledge the invaluable contributions and leadership of the late Professor David Dunger as part of the REACT study. Finally, we would like to thank the funding support received for the REACT trial from National Institute for Health Research (NIHR) Efficacy and Mechanisms Evaluation Programme, the NIHR Cambridge Biomedical Research Centre and the Cambridge Clinical Trials Unit (Cambridge, UK) and the NIHR Clinical Trials Network. The study sponsor was the University of Cambridge and Cambridge University Hospitals NHS Trust (Cambridge, UK). We wish to thank both Medtronic and Nova Biomedical who donated equipment to support the study. Collaborators  The members of the REACT Collaborative are: Lynn Thomson, Catherine Guy, Isabel Iglesias-Platas, Mirjam M van Weissenbruch, Zoltan Molnar, Sheula Barlow, Sharon Baugh, Kathryn Johnson, Lindsay Uryn, Collette Spencer, Maria Hubbard, Sateeshkumar Somisetty, Olaitan Adesiyan, Jogesh Kapadia, Yvonne Millar, Kalyana Gurusamy, Lindsay Bibb, Kathryn Jones, Richard Heaver, Priya Muthukumar, Amy Nichols, Mark Johnson, Jenny Pond, Philippa Crowley, Christie Mellish, Divyen D Shah, Mercy Abraham, Presillina Vincent, Suma Anil kumar, Angelina Iringan, Barbara Aninakwa, R A Dalangin-Chalmers, Annemieke de Lange. Contributors  GXHL, LM and KB conceptualised and designed the study. GXHL conducted data cleaning, management and analysis, and wrote the initial draft of the manuscript. LM and KB provided supervision and critical revisions to first and subsequent drafts. All authors approved the final manuscript version for submission and KB acts as the guarantor for the overall content. The REACT Collaborative acquired the data for this study as part of the REACT trial. Funding  Funding for the REACT study was provided by the National Institute for Health Research (NIHR) Efficacy and Mechanisms Evaluation Programme (RG68719) and supported by the NIHR Cambridge Biomedical Research Centre and the Cambridge Clinical Trials Unit (Cambridge, UK), and through the portfolio from NIHR Clinical Trials Network. Medtronic and Nova Biomedical donated equipment. Neither Medtronic nor Nova Biomedical had any role in design of the study, gathering of data, access to data, preparation of the manuscript or decision to publish the results. Competing interests  None declared. Patient consent for publication  Not applicable. Ethics approval  Ethical and Health Research Authority (HRA) approval was obtained in the UK (IRAS ID 168042), Research Ethics Committee (reference 15/ EE/0158) and from the Medicines and Healthcare products Regulatory Agency (reference CI/2016/0011), and from all local regulatory boards. International regulatory approvals were obtained for non-UK sites. Written parental informed consent was obtained for all study participants at the time of recruitment. Provenance and peer review  Not commissioned; externally peer reviewed. Data availability statement  Data are available upon reasonable request. Anonymised data collected for the study, including individual participant data and a data dictionary defining each field in the set, will be made available 1 year after publication, after approval of a proposal and with a signed data access agreement. All data requests should be submitted to the corresponding author for consideration. Supplemental material  This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those Table 4  Regression analyses assessing impact of insulin on plasma concentrations of BOHB and lactate within specific epochs Postnatal age (d) Insulin received Insulin not received P BOHB BOHB ≤0.1 mmol/L BOHB >0.1 mmol/L N babies N samples BOHB ≤0.1 mmol/L BOHB >0.1 mmol/L N babies N samples Univariable Multivariable*† 1 22 (71) 9 (29) 17 31 63 (65) 34 (35) 61 97 0.664 N/A 2 67 (61) 43 (39) 37 110 172 (66) 87 (34) 108 259 0.664 N/A 3 87 (61) 56 (39) 48 143 153 (69) 68 (31) 99 221 0.497 N/A 4 97 (65) 52 (35) 51 149 132 (72) 52 (28) 90 184 0.664 N/A 5 122 (71) 50 (29) 55 172 121 (78) 34 (22) 74 155 0.664 N/A 6 100 (64) 56 (36) 51 156 118 (80) 30 (20) 73 148 0.198 N/A 7 67 (64) 38 (36) 38 105 93 (84) 18 (16) 65 111 0.027 0.001 8 9 (90) 1 (10) 10 10 12 (100) 0 (0) 11 12 0.052 N/A Lactate Mean (SD) N babies N samples Mean (SD) N babies N samples Univariable Multivariable*† 1 1.99 (0.860) 18 31 2.18 (1.60) 57 95 1.00 N/A 2 2.19 (1.85) 38 113 1.52 (0.821) 106 257 0.020 0.403 3 1.75 (2.01) 49 144 1.35 (0.696) 97 216 0.253 N/A 4 1.51 (0.798) 48 139 1.21 (0.509) 89 188 0.005 0.209 5 1.29 (0.556) 55 168 1.15 (0.449) 76 159 0.750 N/A 6 1.23 (0.504) 51 153 1.15 (0.493) 73 147 0.750 N/A 7 1.19 (1.33) 41 104 1.20 (0.841) 66 112 1.00 N/A 8 1.29 (0.593) 9 9 1.02 (0.439) 13 14 0.750 N/A Values are n (%) or mean (SD). P values are for comparisons across groups within each specific epoch and were obtained using separate binary logistic mixed models for analysis of BOHB concentrations and separate linear mixed models for analysis of lactate concentrations. P values for univariable analyses are adjusted for multiple comparisons using Holm-Bonferroni correction. *Epochs that indicated a significant effect of insulin on plasma BOHB or lactate concentrations were subsequently selected for multivariable analysis. †Multivariable analyses were adjusted for the following confounders: gestation, weight SD score from birth to 7 days, base excess at birth, intubation at birth, maternal steroids and simultaneous plasma glucose concentration. BOHB, β-hydroxybutyrate; N/A, not assessed. https://x.com/l_marcovecchio F7Liu GXH, et al. Arch Dis Child Fetal Neonatal Ed 2025;0:F1–F7. doi:10.1136/archdischild-2025-328701 Original research of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise. Open access  This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/​ licenses/by/4.0/. ORCID iDs Gordon Xin Hua Liu http://orcid.org/0000-0002-4261-9683 Loredana Marcovecchio http://orcid.org/0000-0002-4415-316X K Beardsall http://orcid.org/0000-0003-3582-183X REFERENCES 1 Harding JE, Harris DL, Hegarty JE, et al. An emerging evidence base for the management of neonatal hypoglycaemia. Early Hum Dev 2017;104:51–6. 2 Tas E, Garibaldi L, Muzumdar R. Glucose Homeostasis in Newborns: An Endocrinology Perspective. Neoreviews 2020;21:e14–29. 3 Mitanchez D. Glucose regulation in preterm newborn infants. Horm Res 2007;68:265–71. 4 Hume R, Burchell A, Williams FLR, et al. Glucose homeostasis in the newborn. Early Hum Dev 2005;81:95–101. 5 Tan JBC, Boskovic DS, Angeles DM. The Energy Costs of Prematurity and the Neonatal Intensive Care Unit (NICU) Experience. Antioxidants (Basel) 2018;7:37. 6 Ramel SE, Brown LD, Georgieff MK. The Impact of Neonatal Illness on Nutritional Requirements-One Size Does Not Fit All. Curr Pediatr Rep 2014;2:248–54. 7 Harris DL, Weston PJ, Harding JE. Alternative Cerebral Fuels in the First Five Days in Healthy Term Infants: The Glucose in Well Babies (GLOW) Study. J Pediatr 2021;231:81–6. 8 Harris DL, Weston PJ, Harding JE. Lactate, rather than ketones, may provide alternative cerebral fuel in hypoglycaemic newborns. Arch Dis Child Fetal Neonatal Ed 2015;100:F161–4. 9 Steiner P. Brain Fuel Utilization in the Developing Brain. Ann Nutr Metab 2019;75 Suppl 1:8–18. 10 Sharma A, Davis A, Shekhawat PS. Hypoglycemia in the preterm neonate: etiopathogenesis, diagnosis, management and long-term outcomes. Transl Pediatr 2017;6:335–48. 11 Bougneres PF, Lemmel C, Ferré P, et al. Ketone body transport in the human neonate and infant. J Clin Invest 1986;77:42–8. 12 Hawdon JM, Ward Platt MP. Metabolic adaptation in small for gestational age infants. Arch Dis Child 1993;68:262–8. 13 Hawdon JM, Ward Platt MP, Aynsley-Green A. Patterns of metabolic adaptation for preterm and term infants in the first neonatal week. Arch Dis Child 1992;67:357–65. 14 Costello AM de L, Pal DK, Manandhar DS, et al. Neonatal hypoglycaemia in Nepal 2. Availability of alternative fuels. Arch Dis Child Fetal Neonatal Ed 2000;82:F52–8. 15 Devaskar UP. Point of Care Blood Ketone and Glucose Concentrations in Full Term Healthy Neonates with in First 48 Hours of Life: The POCK Study. BJSTR 2021;36. 16 Beardsall K, Thomson L, Guy C, et al. Protocol of a randomised controlled trial of real- time continuous glucose monitoring in neonatal intensive care “REACT”. BMJ Open 2018;8:e020816. 17 Beardsall K, Thomson L, Guy C, et al. Real-time continuous glucose monitoring in preterm infants (REACT): an international, open-label, randomised controlled trial. Lancet Child Adolesc Health 2021;5:265–73. 18 Cole TJ, Green PJ. Smoothing reference centile curves: the LMS method and penalized likelihood. Stat Med 1992;11:1305–19. 19 Cole TJ, Wright CM, Williams AF, et al. Designing the new UK-WHO growth charts to enhance assessment of growth around birth. Arch Dis Child Fetal Neonatal Ed 2012;97:F219–22. 20 Nova Biomedical. StatStrip and statstrip xpress2 glucose/ketone meters. 21 Nova Biomedical. StatStrip lactate and statstrip xpress lactate systems. 22 De Angelis LC, Brigati G, Polleri G, et al. Neonatal Hypoglycemia and Brain Vulnerability. Front Endocrinol (Lausanne) 2021;12:634305. 23 Watson N, McClelland TJ, Puthucheary Z. Is there a role for ketones as alternative fuel in critical illness? Curr Opin Crit Care 2023;29:300–5. 24 Rabinowitz JD, Enerbäck S. Lactate: the ugly duckling of energy metabolism. Nat Metab 2020;2:566–71. 25 García-Rodríguez D, Giménez-Cassina A. Ketone Bodies in the Brain Beyond Fuel Metabolism: From Excitability to Gene Expression and Cell Signaling. Front Mol Neurosci 2021;14:732120. 26 Newman JC, Verdin E. Ketone bodies as signaling metabolites. Trends Endocrinol Metab 2014;25:42–52. 27 Li X, Yang Y, Zhang B, et al. Lactate metabolism in human health and disease. Sig Transduct Target Ther 2022;7:305. 28 Manosalva C, Quiroga J, Hidalgo AI, et al. Role of Lactate in Inflammatory Processes: Friend or Foe. Front Immunol 2021;12:808799. 29 Byrne NJ, Soni S, Takahara S, et al. Chronically Elevating Circulating Ketones Can Reduce Cardiac Inflammation and Blunt the Development of Heart Failure. Circ Heart Fail 2020;13:e006573. 30 Jensen NJ, Wodschow HZ, Nilsson M, et al. Effects of Ketone Bodies on Brain Metabolism and Function in Neurodegenerative Diseases. Int J Mol Sci 2020;21:8767. 31 Omori NE, Woo GH, Mansor LS. Exogenous Ketones and Lactate as a Potential Therapeutic Intervention for Brain Injury and Neurodegenerative Conditions. Front Hum Neurosci 2022;16:846183. 32 Beardsall K. Hyperglycaemia in the Newborn Infant. Physiology Verses Pathology. Front Pediatr 2021;9:641306. 33 Beardsall K, Acerini C, Dunger DB. Physiological and clinical role of insulin in the neonate. Expert Rev Endocrinol Metab 2010;5:197–207. 34 Hui SA, Chiew AL, Depczynski B. What is the utility of blood beta-hydroxybutyrate measurements in emergency department in patients without diabetes: a systematic review. Syst Rev 2023;12:71. 35 Harris DL, Weston PJ, Harding JE. Point-of-care measurements of blood ketones in newborns. Arch Dis Child Fetal Neonatal Ed 2019;104:F544–6. https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ http://orcid.org/0000-0002-4261-9683 http://orcid.org/0000-0002-4415-316X http://orcid.org/0000-0003-3582-183X http://dx.doi.org/10.1016/j.earlhumdev.2016.12.009 http://dx.doi.org/10.1542/neo.21-1-e14 http://dx.doi.org/10.1159/000104174 http://dx.doi.org/10.1016/j.earlhumdev.2004.10.005 http://dx.doi.org/10.1016/j.earlhumdev.2004.10.005 http://dx.doi.org/10.3390/antiox7030037 http://dx.doi.org/10.1007/s40124-014-0059-3 http://dx.doi.org/10.1016/j.jpeds.2020.12.063 http://dx.doi.org/10.1136/archdischild-2014-306435 http://dx.doi.org/10.1159/000508054 http://dx.doi.org/10.21037/tp.2017.10.06 http://dx.doi.org/10.1172/JCI112299 http://dx.doi.org/10.1136/adc.68.3_spec_no.262 http://dx.doi.org/10.1136/adc.67.4_spec_no.357 http://dx.doi.org/10.1136/fn.82.1.F52 http://dx.doi.org/10.26717/BJSTR.2021.36.005871 http://dx.doi.org/10.1136/bmjopen-2017-020816 http://dx.doi.org/10.1016/S2352-4642(20)30367-9 http://dx.doi.org/10.1002/sim.4780111005 http://dx.doi.org/10.1136/adc.2010.205864 http://dx.doi.org/10.3389/fendo.2021.634305 http://dx.doi.org/10.1097/MCC.0000000000001061 http://dx.doi.org/10.1038/s42255-020-0243-4 http://dx.doi.org/10.1038/s42255-020-0243-4 http://dx.doi.org/10.3389/fnmol.2021.732120 http://dx.doi.org/10.3389/fnmol.2021.732120 http://dx.doi.org/10.1016/j.tem.2013.09.002 http://dx.doi.org/10.1016/j.tem.2013.09.002 http://dx.doi.org/10.1038/s41392-022-01151-3 http://dx.doi.org/10.1038/s41392-022-01151-3 http://dx.doi.org/10.3389/fimmu.2021.808799 http://dx.doi.org/10.1161/CIRCHEARTFAILURE.119.006573 http://dx.doi.org/10.1161/CIRCHEARTFAILURE.119.006573 http://dx.doi.org/10.3390/ijms21228767 http://dx.doi.org/10.3389/fnhum.2022.846183 http://dx.doi.org/10.3389/fnhum.2022.846183 http://dx.doi.org/10.3389/fped.2021.641306 http://dx.doi.org/10.1586/eem.09.58 http://dx.doi.org/10.1186/s13643-023-02203-7 http://dx.doi.org/10.1136/archdischild-2018-316293 Cerebral fuels within the first week of life in very preterm infants: a cohort study Abstract Introduction Methods Population Data collection Assay methodology Statistical analysis Results Time course of glucose, BOHB and lactate concentrations BOHB and lactate concentrations in relation to glucose concentrations and baseline demographics Impact of insulin and nutritional supplementation on BOHB and lactate concentrations ATP contribution or glucose, BOHB and lactate Discussion References