Abstract
Background
Newborn screening (NBS) based on tandem mass spectrometry (MS/MS) is essential for the early identification of inherited metabolic disorders. However, physiological immaturity in preterm and low-birth-weight (LBW) infants frequently alters metabolite concentrations, thereby increasing the risk of false-positive results.
Objective
This study aimed to clarify the independent and interactive effects of gestational age (GA) and birth weight (BW) on neonatal amino acid and acylcarnitine profiles, with the goal of improving screening precision in vulnerable populations.
Methods
We conducted a retrospective analysis of 147,643 neonates screened in Hunan Province, China. The cohort was stratified into seven GA groups and five BW groups. One-way and two-way analyses of variance (ANOVA), together with principal component analysis (PCA), were used to characterize main effects, interaction effects, and global metabolic clustering.
Results
Metabolic profiles followed distinct developmental trajectories. Preterm infants (<37 weeks) showed significantly elevated amino acid concentrations (e.g., tyrosine and arginine) but relative attenuation of long-chain acylcarnitine levels, suggesting an altered balance between protein turnover and fatty acid oxidation. Importantly, significant GA × BW interactions were detected for multiple markers, indicating that the metabolic impact of birth weight varied according to gestational maturity.
Conclusion
Neonatal metabolic maturity is shaped by the synergistic interplay between intrauterine duration and fetal growth. The characteristic “high amino acid/low long-chain acylcarnitine” pattern observed in preterm infants, especially in those with lower birth weight, supports the need for individualized reference intervals in newborn screening.
Keywords: interaction effects, metabolic maturation, newborn screening, reference intervals, tandem mass spectrometry
1. Introduction
Newborn screening (NBS) is among the most successful public health interventions in modern preventive medicine and has fundamentally improved the early detection of inborn errors of metabolism (IEMs) (1). The introduction of tandem mass spectrometry (MS/MS) has greatly expanded screening capacity by enabling simultaneous measurement of amino acids and acylcarnitines from dried blood spots (2, 3).
Nevertheless, diagnostic precision remains suboptimal in premature and low-birth-weight (LBW) infants, a population characterized by physiological immaturity, altered nutrient handling, and rapidly changing metabolic homeostasis (4–7). Compared with term neonates, preterm infants often have immature hepatic enzyme activity, incomplete mitochondrial fatty acid oxidation capacity, and distinct nutritional exposures, all of which may substantially influence screening biomarkers independent of disease status.
Although interest in gestationally tailored interpretation has increased, the literature on metabolic reference intervals in preterm infants remains incomplete (8). Several recent studies have confirmed that gestational age (GA) and birth weight (BW) influence neonatal metabolite concentrations (9–11). However, most prior analyses evaluated these factors separately, whereas in clinical practice GA and BW are biologically intertwined. This distinction matters because the metabolic effect of low birth weight may differ markedly between a preterm infant and a term infant who is small for gestational age. A more clinically useful framework therefore requires explicit assessment of their interaction.
2. Materials and methods
2.1. Study population and ethics
This retrospective study used data obtained from the Hunan Provincial Newborn Screening Center. We reviewed the records of all neonates who underwent expanded newborn screening by tandem mass spectrometry between January 1, 2023, and December 31, 2023. After applying the predefined inclusion and exclusion criteria, 147,643 neonates were enrolled in the final analysis.
Inclusion criteria were as follows: (1) neonates who underwent routine expanded newborn screening during the study period; (2) availability of complete demographic and laboratory information, including gestational age, birth weight, sex, age at sampling, and metabolite concentrations; (3) dried blood spot specimens collected at ≥48 h after birth and after adequate feeding; and (4) specimens that met routine analytical and quality-control requirements for MS/MS-based newborn screening (12, 13).
Exclusion criteria were as follows: (1) confirmed inborn errors of metabolism (IEMs); (2) blood transfusion before sample collection; (3) severe congenital malformations, severe infection, or other critical illnesses likely to substantially alter metabolic homeostasis; and (4) missing key clinical variables or poor-quality specimens unsuitable for analysis. These criteria were applied to reduce potential confounding and to ensure that the final cohort was appropriate for evaluating the effects of gestational age and birth weight on metabolic profiles.
The study protocol was reviewed and approved by the Ethics Committee of the Hunan Provincial Maternal and Child Health Care Hospital (Approval No. 2023-S183). Because this was a retrospective study based on de-identified routine screening data, the requirement for informed consent was waived by the ethics committee.
2.2. Sample collection and laboratory analysis
Heel-prick blood samples were collected at least 48 h after birth, after confirmation of adequate nutritional intake (either breastfeeding at least six times or documented enteral feeding at least six times), in accordance with national newborn screening practice standards and technical guidelines for tandem mass spectrometry-based newborn screening (12, 13). Dried blood spots were prepared on standardized filter paper cards and transported to the central laboratory under controlled conditions.
Metabolic profiling was performed using tandem mass spectrometry (MS/MS). Sample preparation was carried out with the NeoBase™ Non-derivatized MSMS Kit (PerkinElmer, Turku, Finland) according to the manufacturer's protocol, and analytes were measured on a Xevo TQD tandem mass spectrometer (Waters, Milford, MA, USA). Internal standards and calibration materials supplied with the kit were used throughout the analytical workflow.
A total of 43 metabolic markers were quantified, comprising 11 amino acids (including Alanine, Arginine, Citrulline, etc.), 31 acylcarnitines (ranging from free carnitine C0 to long-chain acylcarnitines), and Succinylacetone.
To ensure analytical accuracy and reproducibility, rigorous quality-control procedures were implemented. Each 96-well plate included two blank controls, two low-level quality-control samples, and two high-level quality-control samples. Internal and external quality-assurance procedures were performed in accordance with the requirements of the National Center for Clinical Laboratories, the provincial newborn screening quality-control program, and CLSI recommendations for newborn screening by tandem mass spectrometry (12).
2.3. Grouping strategy
To systematically assess the independent and interactive effects of gestational age (GA) and birth weight (BW) on metabolic profiles, the study population was stratified into predefined subgroups. Based on gestational age, neonates were classified into seven categories: <37 weeks, 37–37 + 6 weeks, 38–38 + 6 weeks, 39–39 + 6 weeks, 40–40 + 6 weeks, 41–41 + 6 weeks, and ≥42 weeks. Based on birth weight, neonates were classified into five categories: <2,500 g, 2,500–2,999 g, 3,000–3,499 g, 3,500–3,999 g, and ≥4,000 g.
2.3.1. Statistical analysis
Data management and statistical analyses were conducted using R software (version 4.4.3; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are presented as mean ± standard deviation (SD), whereas categorical variables are expressed as counts and percentages. Differences in sex distribution across GA and BW subgroups were evaluated using the Pearson chi-square test.
To evaluate metabolic differences across developmental stages, one-way analysis of variance (ANOVA) was used to compare metabolite concentrations among the seven gestational-age groups and among the five birth-weight groups. post-hoc pairwise comparisons were performed when the overall test was significant.
To further dissect the interplay between gestational age and fetal growth, two-way ANOVA models were fitted with GA and BW as fixed factors. These models were used to estimate both the main effects of each factor and their interaction effects on individual metabolites.
Given the large number of simultaneous hypothesis tests, Bonferroni correction was applied to post-hoc pairwise comparisons to control the family-wise error rate. Two-sided P values < 0.05 were considered statistically significant unless otherwise specified.
Finally, principal component analysis (PCA), an unsupervised multivariate technique, was used to reduce the dimensionality of the high-throughput MS/MS dataset and to visualize global metabolic clustering patterns across developmental subgroups.
3. Results
3.1. Demographic characteristics
A total of 147,643 neonates who met the inclusion criteria were enrolled in the final analysis. Baseline demographic characteristics stratified by gestational age (GA) and birth weight (BW) are summarized in Table 1.
Table 1.
Demographic characteristics of the newborn population stratified by gestational age and birth weight (N = 147,643).
| Variable | Total | Male | Female | P valuea |
|---|---|---|---|---|
| No. (%) | No. (%) | No. (%) | ||
| Total population | 147,643 (100.0) | 78,841 (53.4) | 68,802 (46.6) | |
| Gestational age (weeks) | <0.001 | |||
| <37 (Preterm) | 6,363 (10.4) | 3,677 (57.79) | 2,686 (42.21) | |
| 37–37 + 6 | 15,303 (26.3) | 8,636 (56.43) | 6,667 (43.57) | |
| 38–38 + 6 | 38,874 (19.3) | 21,288 (54.76) | 17,586 (45.24) | |
| 39–39 + 6 | 48,726 (33.0) | 25,123 (51.56) | 23,603 (48.44) | |
| 40–40 + 6 | 31,848 (10.0) | 15,562 (48.86) | 16,122 (50.62) | |
| 41–41 + 6 | 6,675 (0.9) | 3,126 (46.83) | 3,549 (53.17) | |
| ≥42 (Post-term) | 44 (0.03) | 26 (59.09) | 18 (40.91) | |
| Birth Weight (g) | <0.001 | |||
| <2,500 (LBW) | 3,842 (2.6) | 1,755 (45.68) | 2,087 (54.32) | |
| 2,500–2,999 | 29,477 (20.0) | 13,231 (44.89) | 16,246 (55.11) | |
| 3,000–3,499 | 68,829 (46.6) | 34,762 (50.50) | 34,607 (49.50) | |
| 3,500–3,999 | 38,125 (25.8) | 22,762 (59.70) | 15,368 (40.31) | |
| ≥4,000 (Macrosomia) | 7,396 (5.0) | 4,928 (66.63) | 2,468 (33.37) |
LBW, Low Birth Weight.
P values were calculated using the Pearson Chi-square test to compare gender distribution across subgroups.
Bold values indicates statistically significant difference at P<0.05.
When categorized by gestational age, the study population was divided into seven subgroups. The largest cohort consisted of infants born at 39–39 + 6 weeks of gestation (n = 48,726, 33.0%), whereas preterm infants (<37 weeks) represented 10.4% of the study population. Overall, the cohort was predominantly composed of term neonates.
Regarding gender distribution, there were slightly more males (53.4%) than females (46.6%) in the overall cohort. Statistical analysis revealed significant differences in gender composition across both gestational age (χ² test, P < 0.001) and birth weight groups (P < 0.001). Specifically, the proportion of males was generally higher in the heavier birth weight categories, consistent with known biological growth patterns.
3.2. Main effects of gestational age and birth weight on metabolic profiles
One-way ANOVA demonstrated that neonatal metabolic profiles were strongly influenced by developmental maturity. As summarized in Table 2, gestational age significantly affected the concentrations of 40 of the 43 measured biomarkers, indicating a broad maturational effect on both amino acid and acylcarnitine metabolism.
Table 2.
Concentrations of amino acids and acylcarnitines across different gestational age groups.
| Analyte | <37 weeks | 37–37 + 6 weeks | 38–38 + 6 weeks | 39–39 + 6 weeks | 40–40 + 6 weeks | 41–41 + 6 weeks | ≥42 weeks | F valuea | P adjb |
|---|---|---|---|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |||
| Amino Acids (μmol/L) | |||||||||
| Alanine (ALA) | 304.72 ± 91.69 | 319.57 ± 90.42 | 319.77 ± 88.27 | 325.01 ± 88.71 | 332.01 ± 90.35 | 335.69 ± 89.73 | 362.26 ± 98.41 | 134.44 | <0.001 |
| Arginine (ARG) | 10.83 ± 9.44 | 7.55 ± 6.04 | 7.51 ± 5.75 | 7.41 ± 5.67 | 7.18 ± 5.53 | 6.78 ± 5.30 | 9.30 ± 7.04 | 373.93 | <0.001 |
| Citrulline (CIT) | 13.66 ± 4.87 | 13.04 ± 3.52 | 13.39 ± 3.44 | 13.73 ± 3.48 | 13.95 ± 3.57 | 13.92 ± 3.47 | 14.60 ± 3.67 | 157.38 | <0.001 |
| Glycine (GLY) | 407.86 ± 105.52 | 450.93 ± 105.52 | 451.43 ± 98.46 | 450.67 ± 98.40 | 454.06 ± 99.84 | 458.38 ± 103.83 | 449.82 ± 115.74 | 206.96 | <0.001 |
| Leucine (LEU + ILE + PRO-OH) | 136.11 ± 33.86 | 131.83 ± 31.40 | 133.64 ± 32.44 | 135.31 ± 32.32 | 137.18 ± 32.93 | 139.38 ± 33.58 | 143.05 ± 37.87 | 81.40 | <0.001 |
| Methionine (MET) | 20.30 ± 5.72 | 20.00 ± 5.37 | 20.01 ± 5.20 | 19.81 ± 5.19 | 19.78 ± 5.19 | 19.68 ± 5.01 | 20.77 ± 5.59 | 17.07 | <0.001 |
| Ornithine (ORN) | 79.91 ± 27.44 | 76.73 ± 22.28 | 80.49 ± 22.81 | 82.95 ± 22.55 | 84.67 ± 22.83 | 85.55 ± 24.97 | 86.37 ± 24.38 | 283.16 | <0.001 |
| Phenylalanine (PHE) | 44.96 ± 10.75 | 46.41 ± 10.67 | 46.05 ± 10.20 | 46.07 ± 9.92 | 46.47 ± 11.27 | 47.10 ± 9.91 | 45.73 ± 8 | 30.70 | <0.001 |
| Proline (PRO) | 151.15 ± 39.58 | 158.17 ± 39.53 | 163.98 ± 39.83 | 167.62 ± 41.50 | 169.70 ± 41.50 | 170.93 ± 42.11 | 175.19 ± 42.13 | 319.61 | <0.001 |
| Tyrosine (TYR) | 100.35 ± 49.72 | 100.38 ± 42.99 | 95.43 ± 37.95 | 90.77 ± 34.07 | 87.01 ± 30.67 | 85.78 ± 28.90 | 84.21 ± 33.65 | 394.98 | <0.001 |
| Valine (VAL) | 108.60 ± 28.61 | 110.92 ± 28.14 | 115.31 ± 28.62 | 118.73 ± 29.34 | 121.63 ± 30.17 | 123.84 ± 31.12 | 124.88 ± 35.72 | 434.50 | <0.001 |
| Acylcarnitines (μmol/L) | |||||||||
| Free carnitine (C0) | 24.47 ± 9.96 | 21.63 ± 7.96 | 20.20 ± 7.14 | 19.93 ± 6.95 | 19.96 ± 7.10 | 19.91 ± 7.24 | 22.17 ± 6.66 | 465.22 | <0.001 |
| Acetylcarnitine (C2) | 17.63 ± 8.11 | 19.47 ± 7.09 | 18.74 ± 6.43 | 18.63 ± 6.43 | 19.10 ± 6.82 | 20.27 ± 7.22 | 20.32 ± 7.12 | 124.02 | <0.001 |
| Propionylcarnitine (C3) | 1.73 ± 0.90 | 1.86 ± 0.76 | 1.73 ± 0.68 | 1.64 ± 0.63 | 1.60 ± 0.62 | 1.58 ± 0.60 | 1.67 ± 0.56 | 351.10 | <0.001 |
| Isobutyrylcarnitine (C4) | 0.25 ± 0.07 | 0.27 ± 0.07 | 0.26 ± 0.07 | 0.25 ± 0.07 | 0.25 ± 0.07 | 0.25 ± 0.07 | 0.25 ± 0.07 | 157.71 | <0.001 |
| Isovalerylcarnitine (C5) | 0.15 ± 0.06 | 0.13 ± 0.05 | 0.12 ± 0.04 | 0.12 ± 0.04 | 0.11 ± 0.05 | 0.11 ± 0.04 | 0.13 ± 0.06 | 817.20 | <0.001 |
| Hexanoylcarnitine (C6) | 0.069 ± 0.024 | 0.071 ± 0.025 | 0.071 ± 0.024 | 0.070 ± 0.023 | 0.070 ± 0.025 | 0.070 ± 0.025 | 0.072 ± 0.025 | 10.31 | <0.001 |
| Octanoylcarnitine (C8) | 0.076 ± 0.026 | 0.081 ± 0.091 | 0.079 ± 0.029 | 0.079 ± 0.027 | 0.079 ± 0.058 | 0.079 ± 0.026 | 0.079 ± 0.026 | 9.67 | <0.001 |
| Decanoylcarnitine (C10) | 0.087 ± 0.036 | 0.098 ± 0.041 | 0.098 ± 0.039 | 0.098 ± 0.039 | 0.098 ± 0.039 | 0.100 ± 0.040 | 0.099 ± 0.038 | 94.00 | <0.001 |
| Lauroylcarnitine (C12) | 0.089 ± 0.052 | 0.11 ± 0.057 | 0.11 ± 0.054 | 0.11 ± 0.053 | 0.11 ± 0.053 | 0.11 ± 0.056 | 0.12 ± 0.064 | 181.52 | <0.001 |
| Myristoylcarnitine (C14) | 0.17 ± 0.085 | 0.21 ± 0.077 | 0.20 ± 0.071 | 0.19 ± 0.068 | 0.19 ± 0.067 | 0.20 ± 0.069 | 0.20 ± 0.085 | 193.37 | <0.001 |
| Palmitoylcarnitine (C16) | 2.31 ± 1.37 | 3.00 ± 1.20 | 3.04 ± 1.15 | 3.04 ± 1.15 | 3.09 ± 1.17 | 3.24 ± 1.22 | 3.04 ± 1.39 | 448.61 | <0.001 |
| Stearoylcarnitine (C18) | 0.75 ± 0.35 | 0.90 ± 0.30 | 0.90 ± 0.29 | 0.89 ± 0.29 | 0.89 ± 0.29 | 0.92 ± 0.30 | 0.84 ± 0.32 | 261.22 | <0.001 |
| Malonylcarnitine/3-Hydroxybutyrylcarnitine (C3DC + C4OH) | 0.15 ± 0.069 | 0.17 ± 0.072 | 0.17 ± 0.070 | 0.17 ± 0.069 | 0.17 ± 0.071 | 0.18 ± 0.075 | 0.16 ± 0.073 | 130.79 | <0.001 |
| Methylmalonylcarnitine/3-Hydroxyisovalerylcarnitine (C4DC + C5OH) | 0.20 ± 0.057 | 0.21 ± 0.058 | 0.21 ± 0.058 | 0.21 ± 0.058 | 0.22 ± 0.066 | 0.22 ± 0.060 | 0.22 ± 0.067 | 154.19 | <0.001 |
| Tiglylcarnitine (C5:1) | 0.036 ± 0.017 | 0.036 ± 0.017 | 0.036 ± 0.017 | 0.035 ± 0.016 | 0.034 ± 0.016 | 0.033 ± 0.015 | 0.037 ± 0.018 | 56.61 | <0.001 |
| Glutarylcarnitine/3-Hydroxyhexanoylcarnitine (C5DC + C6OH) | 0.18 ± 0.059 | 0.19 ± 0.059 | 0.19 ± 0.059 | 0.19 ± 0.058 | 0.18 ± 0.058 | 0.18 ± 0.054 | 0.18 ± 0.057 | 45.50 | <0.001 |
| Adipylcarnitine (C6DC) | 0.18 ± 0.071 | 0.19 ± 0.073 | 0.19 ± 0.074 | 0.19 ± 0.073 | 0.19 ± 0.071 | 0.18 ± 0.067 | 0.20 ± 0.075 | 40.16 | <0.001 |
| Octenoylcarnitine (C8:1) | 0.16 ± 0.060 | 0.17 ± 0.058 | 0.17 ± 0.057 | 0.17 ± 0.057 | 0.17 ± 0.057 | 0.17 ± 0.056 | 0.17 ± 0.053 | 15.71 | <0.001 |
| Decenoylcarnitine (C10:1) | 0.093 ± 0.036 | 0.099 ± 0.036 | 0.098 ± 0.034 | 0.097 ± 0.034 | 0.097 ± 0.033 | 0.095 ± 0.030 | 1.00 ± 0.028 | 33.50 | <0.001 |
| Decadienoylcarnitine (C10:2) | 0.030 ± 0.012 | 0.030 ± 0.012 | 0.030 ± 0.012 | 0.029 ± 0.015 | 0.029 ± 0.011 | 0.028 ± 0.011 | 0.030 ± 0.011 | 35.46 | <0.001 |
| Dodecenoylcarnitine (C12:1) | 0.092 ± 0.050 | 0.11 ± 0.054 | 0.12 ± 0.052 | 0.11 ± 0.051 | 0.11 ± 0.051 | 0.12 ± 0.051 | 0.11 ± 0.049 | 134.97 | <0.001 |
| Tetradecenoylcarnitine (C14:1) | 0.12 ± 0.055 | 0.14 ± 0.058 | 0.14 ± 0.055 | 0.13 ± 0.052 | 0.13 ± 0.051 | 0.13 ± 0.053 | 0.13 ± 0.053 | 148.37 | <0.001 |
| Tetradecadienoylcarnitine (C14:2) | 0.311 ± 0.011 | 0.033 ± 0.011 | 0.032 ± 0.011 | 0.032 ± 0.010 | 0.031 ± 0.010 | 0.031 ± 0.0098 | 0.032 ± 0.0091 | 122.95 | <0.001 |
| 3-Hydroxymyristoylcarnitine (C14OH) | 0.023 ± 0.011 | 0.026 ± 0.011 | 0.026 ± 0.011 | 0.026 ± 0.011 | 0.026 ± 0.011 | 0.026 ± 0.011 | 0.026 ± 0.012 | 102.42 | <0.001 |
| Palmitoleoylcarnitine (C16:1) | 0.14 ± 0.093 | 0.18 ± 0.087 | 0.18 ± 0.081 | 0.18 ± 0.080 | 0.18 ± 0.081 | 0.19 ± 0.083 | 0.18 ± 0.092 | 269.34 | <0.001 |
| 3-Hydroxypalmitoleoylcarnitine (C16:1OH) | 0.041 ± 0.015 | 0.047 ± 0.014 | 0.046 ± 0.014 | 0.047 ± 0.014 | 0.047 ± 0.014 | 0.047 ± 0.013 | 0.049 ± 0.013 | 188.56 | <0.001 |
| 3-Hydroxypalmitoylcarnitine (C16OH) | 0.026 ± 0.012 | 0.030 ± 0.012 | 0.030 ± 0.012 | 0.029 ± 0.012 | 0.029 ± 0.012 | 0.030 ± 0.012 | 0.029 ± 0.013 | 111.87 | <0.001 |
| Oleoylcarnitine (C18:1) | 1.29 ± 0.51 | 1.43 ± 0.43 | 1.37 ± 0.40 | 1.35 ± 0.39 | 1.36 ± 0.40 | 1.40 ± 0.40 | 1.36 ± 0.41 | 111.25 | <0.001 |
| 3-Hydroxyoleoylcarnitine (C18:1OH) | 0.028 ± 0.011 | 0.031 ± 0.010 | 0.031 ± 0.010 | 0.031 ± 0.010 | 0.031 ± 0.010 | 0.031 ± 0.010 | 0.029 ± 0.010 | 70.87 | <0.001 |
| Linoleoylcarnitine (C18:2) | 0.25 ± 0.11 | 0.21 ± 0.093 | 0.19 ± 0.087 | 0.19 ± 0.084 | 0.19 ± 0.083 | 0.18 ± 0.083 | 0.22 ± 0.10 | 534.17 | <0.001 |
| 3-Hydroxystearoylcarnitine (C18OH) | 0.019 ± 0.0089 | 0.022 ± 0.0091 | 0.022 ± 0.0089 | 0.022 ± 0.0086 | 0.022 ± 0.0086 | 0.022 ± 0.0085 | 0.022 ± 0.0084 | 138.13 | <0.001 |
| Others (μmol/L) | |||||||||
| Succinylacetone (SA) | 0.86 ± 0.32 | 0.88 ± 0.32 | 0.87 ± 0.32 | 0.87 ± 0.32 | 0.87 ± 0.32 | 0.88 ± 0.31 | 0.85 ± 0.30 | 2.21 | 1.00 |
Data are expressed as mean (standard deviation). SD, standard deviation.
F values derived from one-way analysis of variance (ANOVA).
P values were adjusted using the Bonferroni correction for multiple comparisons.
A clear developmental gradient was observed: concentrations of most amino acids decreased with advancing gestational age. For example, tyrosine (TYR) levels were markedly elevated in preterm infants (<37 weeks) compared with term infants, a pattern consistent with relative hepatic immaturity and reduced enzymatic clearance (Figure 1).
Figure 1.
Metabolic profiles across gestational age. (A) Tyrosine; (B) Arginine; (C) Free Carnitine (C0); (D) C16 (Palmitoylcarnitine).
In contrast, fatty acid oxidation markers showed distinct maturational trajectories. Free carnitine (C0) levels declined progressively with advancing gestational age, whereas several long-chain acylcarnitines displayed gestation-dependent increases, suggesting progressive maturation of mitochondrial lipid oxidation pathways.
3.2.1. Interaction effects of gestational age and birth weight
Beyond the independent main effects, two-way ANOVA revealed widespread interaction effects between gestational age and birth weight, supporting a synergistic interplay between intrauterine duration and fetal growth in shaping neonatal metabolic phenotypes (Table 3).
Table 3.
Results of two-way ANOVA evaluating the independent and interactive effects of gestational age and birth weight.
| Analyte | Main effect: GA | Main effect: BW | Interaction: GA × BW |
|---|---|---|---|
| F value | F value | F value (P value) | |
| Amino Acids | |||
| ALA | 134.80 | 217.34 | 30.46 (<0.001) |
| ARG | 375.93 | 12.61 | 130.59 (<0.001) |
| CIT | 158.51 | 902.76 | 27.59 (<0.001) |
| GLY | 207.97 | 216.20 | 85.88 (<0.001) |
| LEU + ILE + PRO − OH | 81.42 | 19.99 | 3.965 (0.031) |
| MET | 17.08 | 57.82 | 7.90 (<0.001) |
| ORN | 284.45 | 556.68 | 20.34 (<0.001) |
| PHE | 30.74 | 63.52 | 25.31 (<0.001) |
| PRO | 322.92 | 1,230.57 | 51.42 (<0.001) |
| SA | 2.21 | 0.12 | 1.34 (1.00) |
| TYR | 396.08 | 33.93 | 63.95 (<0.001) |
| VAL | 434.86 | 14.27 | 19.14 (<0.001) |
| Acylcarnitines | |||
| C0 | 467.81 | 694.36 | 22.44 (<0.001) |
| C2 | 124.38 | 11.23 | 71.08 (<0.001) |
| C3 | 359.94 | 3,250.92 | 79.05 (<0.001) |
| C3DC_C4OH | 131.15 | 11.32 | 66.47 (<0.001) |
| C4 | 158.05 | 86.25 | 40.03 (<0.001) |
| C4DC_C5OH | 154.36 | 38.74 | 21.59 (<0.001) |
| C5 | 817.97 | 60.26 | 14.34 (<0.001) |
| C5_1 | 56.62 | 15.41 | 1.96 (1.00) |
| C5DC_C6OH | 45.57 | 136.24 | 19.77 (<0.001) |
| C6 | 10.32 | 58.61 | 4.92 (0.002) |
| C6DC | 40.17 | 0.77 | 9.80 (1.00) |
| C8 | 9.68 | 52.83 | 3.61 (0.06) |
| C8_1 | 15.71 | 22.62 | 6.11 (<0.001) |
| C10 | 94.30 | 202.61 | 46.03 (<0.001) |
| C10_1 | 33.59 | 268.52 | 16.91 (<0.001) |
| C10_2 | 35.46 | 0.54 | 1.83 (1.00) |
| C12 | 182.41 | 165.94 | 94.94 (<0.001) |
| C12_1 | 135.35 | 52.61 | 60.89 (<0.001) |
| C14 | 194.56 | 12.24 | 150.93 (<0.001) |
| C14_1 | 148.87 | 17.25 | 82.32 (<0.001) |
| C14_2 | 123.27 | 209.01 | 29.59 (<0.001) |
| C14OH | 102.58 | 10.66 | 37.28 (<0.001) |
| C16 | 452.52 | 191.98 | 183.74(<0.001) |
| C16_1 | 271.11 | 95.75 | 146.63 (<0.001) |
| C16_1OH | 189.47 | 263.69 | 76.14 (<0.001) |
| C16OH | 112.11 | 10.93 | 52.46 (<0.001) |
| C18 | 262.83 | 20.27 | 149.65 (<0.001) |
| C18_1 | 111.78 | 5.74 | 117.42 (<0.001) |
| C18_1OH | 70.99 | 4.95 | 42.31 (<0.001) |
| C18_2 | 537.27 | 689.34 | 29.03 (<0.001) |
| C18OH | 138.51 | 130.60 | 47.10 (<0.001) |
Bold values indicates statistically significant difference at P<0.05.
The non-uniform impact of birth weight across gestational age groups is visualized in Figure 2. For long-chain acylcarnitines, Palmitoylcarnitine (C16) demonstrated the strongest interaction effect (F = 183.74, P < 0.001). While C16 concentrations generally increased with gestational maturity (as seen in the main effects), the disparity between low birth weight (<2,500 g) and normal weight infants was disproportionately amplified in the preterm cohort compared to term infants (Figure 2A). This suggests that the metabolic deficit in fatty acid oxidation markers associated with low birth weight is exacerbated by physiological prematurity.
Figure 2.
Interaction effects of gestational age and birth weight. (A) C16 (Palmitoylcarnitine); (B) Arginine (ARG).
Similarly, amino acid profiles, particularly arginine (ARG), exhibited distinct interaction patterns (F = 130.59, P < 0.001; Figure 2B). The elevation of arginine levels typically observed in lower-birth-weight infants was substantially more pronounced in preterm neonates, further indicating that fetal growth restriction and developmental immaturity jointly influence amino acid homeostasis.
3.3. Multivariate pattern recognition and metabolic clustering
To visualize the global metabolic architecture and identify systemic patterns extending beyond individual biomarkers, principal component analysis (PCA) was performed. As illustrated in the PCA biplot (Figure 3), the first two principal components captured a biologically interpretable gradient of metabolic maturation.
Figure 3.
PCA biplot of metabolic profiles.
The first principal component (PC1), accounting for the majority of the variation (26.28%), was heavily and positively loaded by long-chain and medium-chain acylcarnitines—specifically C14, C16, C2, and C18—serving as a proxy for fatty acid oxidation capacity. The second principal component (PC2, 19.65%) was primarily driven by branched-chain amino acids (LEU + ILE + PRO-OH, VAL) and urea cycle metabolites (ORN, PRO), reflecting protein turnover status.
The score plot revealed a distinct developmental gradient along these orthogonal axes. Preterm infants (<37 weeks) clustered predominantly in the quadrant characterized by negative PC1 scores and positive PC2 scores, whereas term infants were distributed toward regions consistent with greater metabolic maturity and more coordinated fatty acid oxidation.
To further corroborate these inter-metabolite relationships, a correlation heatmap was generated (Figure 4). The analysis highlighted a structured modularity within the metabolome. A robust, tightly interconnected cluster was observed among acylcarnitines, particularly long-chain species, with Palmitoylcarnitine (C16) exhibiting a near-linear correlation with Stearoylcarnitine (C18) (r = 0.90) and Myristoylcarnitine (C14) (r = 0.81).
Figure 4.
Metabolic correlation heatmap (pheatmap).
Distinct from this lipid oxidation module, amino acids formed a separate correlation block. Specifically, branched-chain amino acids showed strong internal consistency, exemplified by the high correlation between Valine and Leucine + Isoleucine (r = 0.84). Notably, the cross-correlation between the acylcarnitine and amino acid clusters was generally weak. This modular architecture confirms that physiological maturation manifests as coordinated shifts in specific metabolic pathways rather than isolated biomarker fluctuations.
4. Discussion
This large retrospective study of more than 147,000 neonates provides a comprehensive overview of metabolic maturation across the gestational spectrum. Our findings demonstrate that neonatal metabolite profiles are influenced not only by gestational age and birth weight individually, but also by their interaction, underscoring the limitations of applying uniform newborn screening thresholds across heterogeneous neonatal subgroups.
The inverse association between amino acid concentrations and gestational age is biologically plausible and consistent with the physiological immaturity of hepatic and renal systems in premature infants (14). The marked elevation of tyrosine in preterm neonates may reflect delayed maturation of tyrosine catabolic pathways, whereas higher arginine concentrations may be related to developmental differences in nitrogen handling and enteral nutritional exposure (15, 16).
A notable finding of this study is the fatty acid oxidation profile observed in preterm infants: relatively elevated free carnitine (C0) accompanied by reduced long-chain acylcarnitines such as C16. This pattern may reflect incomplete mitochondrial transport and β-oxidation capacity, together with altered substrate availability in early life (17–19).
Perhaps the most clinically meaningful result is the significant interaction between gestational age and birth weight. Our data indicate that the metabolic effect of birth weight is context dependent rather than uniform (20, 21). In practical terms, a low birth weight in a preterm infant does not carry the same biochemical implications as a similar birth weight in a term infant, reinforcing the need for stratified and clinically interpretable screening cutoffs.
The systemic metabolic shifts identified by PCA and correlation analyses further challenge the traditional “one-size-fits-all” approach to newborn screening (22). The modular behavior of acylcarnitines and amino acids suggests that metabolic immaturity affects coordinated biochemical pathways rather than isolated markers alone.
The strengths of this study include its large sample size and the explicit statistical evaluation of interaction effects, which have often been overlooked in prior single-factor analyses. Several limitations should also be acknowledged. First, because the study was retrospective, detailed individual-level data on parenteral nutrition, specific enteral supplements, and concurrent clinical conditions were not available for all infants. Second, the cohort was derived from a single provincial screening center, which may limit generalizability to other populations. Third, although the findings support refinement of screening interpretation, disease-confirmed outcome analyses were beyond the scope of the present work.
5. Conclusion
In conclusion, this study delineates the metabolic landscape of the neonatal period and shows that metabolic maturity is jointly shaped by gestational age and birth weight. The observed subgroup-specific patterns provide supportive evidence for more individualized and developmentally informed interpretation of newborn screening results.
Funding Statement
The author(s) declared that financial support was received for this work and/or its publication. Hunan Provincial High-level Health Talents Major Research Special Fund (R2023120).
Footnotes
Edited by: Yongmei Shen, Tianjin Central Hospital for Gynecology and Obstetrics, China
Reviewed by: Tommolini Maria Lucia, University of Studies G. d'Annunzio Chieti and Pescara, Italy
Tianqiao Zhang, The First Affiliated Hospital of University of South China, China
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.
Ethics statement
This study was approved by the Ethics Committee of Hunan Maternal and Child Health Hospital (Protocol No. 2023-S183). The requirement for informed consent was waived because of the retrospective study design.
Author contributions
HT: Methodology, Writing – review & editing, Supervision, Investigation, Writing – original draft, Formal analysis, Software, Data curation, Resources, Conceptualization, Validation, Project administration, Visualization. YM: Project administration, Visualization, Data curation, Methodology, Conceptualization, Writing – original draft, Writing – review & editing, Formal analysis. YZ: Validation, Supervision, Visualization, Data curation, Software, Writing – original draft, Writing – review & editing. HY: Visualization, Resources, Validation, Writing – review & editing, Investigation, Writing – original draft, Formal analysis, Software, Supervision. HZ: Methodology, Writing – review & editing, Validation, Project administration, Supervision, Writing – original draft, Conceptualization, Visualization, Resources.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
- 1.Mori M, Chaudhari BP, Ream MA, Kemper AR. Promises and challenges of genomic newborn screening (NBS)—lessons from public health NBS programs. Pediatr Res. (2025) 97(4):1327–36. 10.1038/s41390-024-03689-0 [DOI] [PubMed] [Google Scholar]
- 2.Millington DS. How mass spectrometry revolutionized newborn screening. J Mass Spectrome Adv Clin Lab. (2024) 32:1–10. 10.1016/j.jmsacl.2024.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Djordjevic D, Tsuchiya E, Fitzpatrick M, Sondheimer N, Dowling JJ. Utility of metabolic screening in neurological presentations of infancy. Ann Clin Transl Neurol. (2020) 7(7):1132–40. 10.1002/acn3.51076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Langeveld M, Sirrs S, Schoenmakers DH, Fazio T, van der Klauw MM, Maillot F, et al. Screening for life: perspectives from adult metabolic specialists on newborn screening for inherited metabolic diseases. J Inherit Metab Dis. (2025) 48(4):e70057. 10.1002/jimd.70057 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Beken S, Abali S, Saral NY, Guner B, Dinc T, Albayrak E, et al. Early postnatal metabolic profile in neonates with different birth weight Status: a pilot study. Front Pediatr. (2021) 9:27–29. 10.3389/fped.2021.646860 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Messina M, Arena A, Iacobacci R, Spina LL, Meli C, Raudino F, et al. Butyrylcarnitine elevation in newborn screening: reducing false positives and distinguishing between two rare diseases through the evaluation of new ratios. Biomedicines (2023) 11(12):3247. 10.3390/biomedicines11123247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.World Health Organization. (2023). Born too Soon: Decade of Action on Preterm Birth. Geneva: WHO. Available online at: https://www.who.int/publications/i/item/9789240073890 [Google Scholar]
- 8.He F, Xie T, Huang X, Zhang J, Tang T. Retrospective analysis of reference intervals for dried blood spot based ms/ms newborn screening programs in Chinese preterm neonates: a nationwide study. BMC Pediatr. (2024) 24(1):424. 10.1186/s12887-024-04865-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Viera CS, Barreto GMS, Silveira DRC, Oliveira HR, Toso DBORG, Rover MS, et al. Biochemical predictors for metabolic syndrome in preterm infants according to weight ratio. Arch Endocrinol Metab. (2020) 64(5):567–74. 10.20945/2359-3997000000237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zhai X, Liu J, Yu M, Zhang Q, Li M, Zhao N, et al. Nontargeted metabolomics reveals the potential mechanism underlying the association between birthweight and metabolic disturbances. BMC Pregnancy Childbirth. (2023) 23(1):14. 10.1186/s12884-023-05346-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Qiu C, Liao J, Gheissari R, Li C, Kapai A, Conti DV, et al. Dysregulated maternal and newborn fatty acid, sugar and amino acid metabolism associated with high birth weight. Int J Obes. (2025) 49:1345–53. 10.1038/s41366-025-01775-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.De Jesús VR, Kuracina M, Morrissey MA, La Marca G, Chace DH, Manning A, et al. Clinical and Laboratory Standards Institute (CLSI). (2022). Newborn Screening by Tandem Mass Spectrometry; Approved Guideline. 2nd ed. Wayne, PA: CLSI. CLSI document NBS04-A2. [Google Scholar]
- 13.National Health Commission of the People’s Republic of China. Technical Specification for Neonatal Disease Screening (WS/T 790-2021). Beijing: Standards Press of China; (2021). [Google Scholar]
- 14.Ernst M, Rogers S, Lausten-Thomsen U, Bjorkbom A, Laursen SS, Courraud J, et al. Gestational age-dependent development of the neonatal metabolome. Pediatr Res. (2021) 89(6):1396–404. 10.1038/s41390-020-01149-z [DOI] [PubMed] [Google Scholar]
- 15.Qasim A, Mehdi MQ, Bhatia S, Franco-Fuenmayor ME, Jain SK. Enteral L-citrulline supplementation in preterm infants is safe and effective in increasing plasma arginine and citrulline levels—a pilot randomized trial. J Perinatol. (2024) 44(5):702–8. 10.1038/s41372-023-01761-x [DOI] [PubMed] [Google Scholar]
- 16.Trivedi A, Jatana V, Sinn JK. Early versus late administration of amino acids in preterm infants receiving parenteral nutrition. Cochrane Database Syst Rev. (2024) 2024(1):CD008771. 10.1002/14651858.cd008771.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lund KC, Scottoline B, Jordan BK. Carnitine-Acylcarnitine translocase deficiency masked by extreme prematurity. J Pediatr Genet. (2021) 12(2):179–83. 10.1055/s-0041-1723960 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Koves TR, Zhang G-F, Davidson MT, Chaves AB, Crown SB, Johnson JM, et al. Pyruvate-supported flux through medium-chain ketothiolase promotes mitochondrial lipid tolerance in cardiac and skeletal muscles. Cell Metab. (2023) 35(6):1056.e8–38. 10.1016/j.cmet.2023.03.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Manta-Vogli PD, Schulpis KH, Dotsikas Y. Quantitation and evaluation of perinatal medium-chain and long-chain acylcarnitine blood concentrations in 12,000 full-term breastfed newborns. J Pediatr Endocrinol Metab. (2021) 34(8):1023–30. 10.1515/jpem-2020-0741 [DOI] [PubMed] [Google Scholar]
- 20.Colicino E, Ferrari F, Cowell W, Niedzwiecki MM, Pedretti FN, Joshi A, et al. Non-linear and non-additive associations between the pregnancy metabolome and birthweight. Environ Int. (2021) 156:106750. 10.1016/j.envint.2021.106750 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Vidarsdottir H, Thorkelsson T, Halldorsson TI, Bjarnason R, Geirsson RT, Rinaldo P, et al. Does metabolomic profile differ with regard to birth weight? Pediatr Res. (2021) 89(5):1144–51. 10.1038/s41390-020-1033-0 [DOI] [PubMed] [Google Scholar]
- 22.Wang X, Wang Y-Y, Xu Z-F. Combined genetic screening and traditional biochemical screening to optimize newborn screening systems. Clin Chim Acta. (2022) 528:44–51. 10.1016/j.cca.2022.01.015 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.