Abstract Background This study aimed to compare the blood metabolism status of newborns with maternal-isolated oligohydramnios (IO) and normal amniotic fluid and investigate the relationship between IO and fetal health. Methods Blood metabolites of 60 neonates with maternal IO and 60 healthy controls (HC) admitted to The Sixth Affiliated Hospital, Sun Yat-sen University (a discovery set), were analyzed using liquid chromatography-mass spectrometry to identify differential metabolites. Pathway analysis was performed to identify the most enriched metabolic pathways in discovery set. The discriminant metabolites were further verified in an independent set consisting of 50 neonates with maternal IO and 60 HC admitted to The Maternal and Child Health Care Hospital of Huadu. Results The blood metabolic profile of newborns with IO was significantly different from that of HC. Differential metabolites were mainly enriched in Phe, Tyr and Trp biosynthesis. Levels of Phe, Tyr and Trp were significantly lower in IO neonates which found in the discovery set, and were verified in the validation set. Conclusions The difference in the blood metabolomics profiles between neonates with maternal IO and HC was mainly in the pathway of Phe, Tyr and Trp biosynthesis. Phe, Tyr, and Trp biosynthesis may be involved in the physiological processes related to the nervous system and digestive system. Continuous attention to changes in blood metabolites in IO neonates and the effect on body growth and development is necessary. Supplementary Information The online version contains supplementary material available at 10.1186/s12884-025-07729-3. Keywords: Tandem mass spectrometry, Isolated oligohydramnios, Metabolomics, Phenylalanine, Tyrosine Background Amniotic fluid volume is an important indicator of fetal health, and its assessment is required during every ultrasound examination in the middle and third trimesters of pregnancy [[44]1, [45]2]. The most commonly used techniques are the maximum vertical pocket (MVP) and amniotic fluid index (AFI) measurements. The MVP, also known as deepest vertical pocket (DVP) or single deepest vertical pocket (SDVP), is the maximal vertical pocket depth, which should not contain fetal limbs or umbilical cord, while the AFI is the summation of the vertical diameter of the largest pocket in each of the four quadrants, using the maternal umbilicus as a central reference point [[46]3–[47]5]. The definition of oligohydramnios has been controversial for decades, including MVP < 2 cm, AFI < 5 cm, AFI < 8 cm, and or AFI < 5th percentile [[48]3]. In general, MVP < 2 cm or an AFI < 5 cm have been widely recognized, among which MVP is better than AFI in identifying oligohydramnios [[49]5]. The prevalence of oligohydramnios is 0.5-5% [[50]6]. However, less than half of the cases with oligohydramnios have no pathological factors and are called isolated oligohydramnios (IO) [[51]7]. Currently, the cause of IO remains unclear. Changes in aquaporin expression may be among the pathophysiological mechanisms, and reduced placental perfusion and chronic hypoxia have been implicated [[52]8, [53]9]. IO is associated with increased labor induction and cesarean section rates [[54]10, [55]11]. A meta-analysis identified increased risks of adverse outcomes at birth in low-risk pregnancies with IO, including meconium aspiration syndrome, cesarean section due to abnormal fetal heart rate, and admission to the neonatal intensive care unit [[56]12]. However, some studies have shown that IO has similar neonatal outcomes to those of normal amniotic fluid pregnancies [[57]11, [58]13]. Therefore, the gestational management of full-term IO remains controversial. Notably, a retrospective cohort study on the long-term outcomes of children born with IO found a significant association between IO and long-term neurological disorders in the offspring, with common diseases such as movement disorders, pervasive developmental disorders [[59]14] and IO is an independent risk factor for long-term gastrointestinal morbidity in offspring, including gastroduodenal diseases and inflammatory bowel diseases (IBD) [[60]15]. The suspected increased risk of long-term neurodevelopmental abnormalities and gastrointestinal disease in these IO offspring suggests that there may be minor abnormalities during pregnancy that may not be detected by routine examination but potentially affect fetal health in the short or long term. Therefore, we applied tandem mass spectrometry and metabolomics analysis, which can accurately identify trace metabolic substances and is widely used in neonatal metabolic disease screening, diagnosis, and monitoring, by comparing the metabolic data profiles of IO and normal amniotic fluid pregnancy to identify specific differential metabolic markers and explore the influence of IO on fetal health. Materials and methods Design and setting This study adopted a retrospective observational design. Participants were recruited from two medical centers between January 2017 and December 2023. The discovery cohort comprised 60 neonates with IO and 60 healthy controls (HC) from The Sixth Affiliated Hospital, Sun Yat-sen University, while the validation included 50 IO and 60 HC from The Maternal and Child Health Care Hospital of Huadu. Participants Inclusion criteria: (1) Ultrasound-based assessment of amniotic fluid volume was performed before the delivery and between 37^+ 0~41^+ 6 gestation weeks, and IO was diagnosed; (2) neonates with no congenital anomalies or inherited metabolic disorders confirmed by prenatal ultrasound and postnatal screening; (3) Parental absence of metabolic disease or substance abuse history (tobacco, alcohol, or illicit drugs). Exclusion criteria: (1) Preterm (< 37 weeks) or post-term (≥ 42 weeks) delivery. (2) Neonatal critical conditions (e.g., respiratory distress syndrome, sepsis, necrotizing enterocolitis). (3) Maternal comorbidities during pregnancy (e.g., diabetes, hypertension, thyroid dysfunction, infectious diseases). (4) Parental refusal to participate. Diagnostic parameters Ultrasound-based assessment of amniotic fluid volume was performed according to standardized protocols: The MVP is defined as the maximal vertical pocket depth, which should not contain fetal limbs or umbilical cord. AFI is the summation of the vertical diameter of the largest pocket in each of the four quadrants, which was divided by the umbilical horizontal line and abdominal white lines, using the maternal umbilicus as a central reference point. IO is diagnosed as MVP < 2 cm or an AFI < 5 cm, validated by two independent sonographers [[61]3]. Data collection A total of 230 participants were recruited, including a discovery sets from The Sixth Affiliated Hospital, Sun Yat-sen University and a validation set from The Maternal and Child Health Care Hospital of Huadu. The demographic characteristics of the participants were collected, including maternal age, weight (before delivery), height, delivery mode, and neonatal gender, birth weight and length. Newborns have adopted mixed feeding methods in order to ensure adequate feeding in early life. Neonatal heel blood samples (three drops, diameter > 0.8 cm) were collected on the third day after birth using standardized filter paper cards (Perkin-Elmer, Turku, Finland). After air-drying the blood spot card at ambient temperature (18–25℃), it was placed in a polystyrene sealed bag and stored in a 4℃ refrigerator for detection within 1 week. Metabolite extraction and analysis Dried blood spots were perforated (3 mm diameter) and transferred to 96-well plates (Millipore Inc., USA). Metabolites were extracted using a methanol-water solution (8:2) containing 100 µL of isotope-labeled internal standards. Samples underwent agitation (700 rpm, 45 ℃, 45 min) and centrifugation (4,000 rpm, 5 min). Following centrifugation, 75 µL of supernatant was aspirated and transferred into a 96-well V-bottom thermally stable microplate, which was subsequently sealed with aluminum foil. The prepared samples were analyzed via ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS; Xevo-TQD, Waters Corp., USA). Quantification utilized manufacturer-specific software with internal standard calibration [[62]16]. Statistical analysis Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) was implemented in MetaboAnalyst 6.0 to identify inter-group metabolic disparities. Model validity was assessed by cumulative R²X (variance explained) and Q² (predictive accuracy), with thresholds > 0.5 indicating robustness. Variable Importance in Projection (VIP) scores > 1 was prioritized for further analysis. T-test and Non-parametric Mann-Whitney U tests (SPSS, v.25.0) compared metabolite levels between groups, adjusted for multiple testing using Benjamini-Hochberg correction (FDR < 0.05). Receiver Operating Characteristic (ROC) curves (MedCalc^® v20.1) evaluated diagnostic potential of differential metabolites, reporting AUC, sensitivity, and specificity. Metabolite Set Enrichment Analysis (MetaboAnalyst 6.0) mapped metabolites to Kyoto Encyclopedia of Genes and Genomes (KEGG) human metabolic pathways to elucidate biological relevance. Results Demographic characteristics The demographic characteristics of the participants (discovery set and validation set) are shown in Table [63]1. Maternal age, weight, height, and delivery mode did not significantly differ between the IO and HC groups. There were no significant differences in gestational age, sex, birth weight, length, or head circumference between infants with IO and HC. Table 1. Comparison of clinical characteristics of pregnant mothers and infants Characteristics Variables Discovery set Validation set FDR between sets IO (n = 60) HC (n = 60) FDR IO (n = 50) HC (n = 60) FDR IO HC Mothers Age(year) 27.85 ± 4.30 28.25 ± 4.46 0.618 28.22 ± 3.91 28.17 ± 3.69 0.86 0.66 0.84 Weight(kg) 64.63 ± 8.52 63.80 ± 7.71 0.578 66.51 ± 7.98 67.88 ± 10.89 0.60 0.25 0.10 Height(cm) 158.34 ± 4.76 159.00 ± 5.26 0.473 159.88 ± 5.01 159.08 ± 4.57 0.68 0.25 0.84 Delivery mode (natural labour/ cesarean section) 36/24 43/17 0.180 41/9 42/18 0.60 0.05 0.84 Infants Gestational age (days) 271.10 ± 8.63 273.55 ± 5.46 0.066 274.90 ± 5.76 274.17 ± 6.16 0.86 0.05 0.84 Gender (male/female) 27/33 32/28 0.363 25/25 35/25 0.60 0.66 0.84 Birth weight(g) 3039.00 ± 361.85 3163.83 ± 356.94 0.060 3150.00 ± 303.72 3062.00 ± 347.66 0.60 0.09 0.41 Length(cm) 49.35 ± 1.30 49.85 ± 1.60 0.063 49.92 ± 1.64 49.50 ± 1.95 0.60 0.11 0.84 [64]Open in a new tab OPLS-DA model building and score plot Blood metabolites of the infants in the IO and HC groups in discovery set were analyzed using the OPLS-DA model. The OPLS-DA score plots for the two groups are shown in Fig. [65]1. The model contained three principal components: R^2Xcum = 0.5355, R^2Ycum = 0.8357, and Q^2 = 0.6261, all of which exceeded 0.5, indicating that the model had good explanatory ability, stability, predictability, and no overfitting. As shown in Fig. [66]1, the metabolites of the IO and HC groups in discovery set were relatively distributed in the left and right regions, suggesting differences in blood metabolites between the groups. Fig. 1. [67]Fig. 1 [68]Open in a new tab The OPLS-DA score plot of the IO and HC groups. Green dots represent IO group (n = 60); Blue dots represent HC group (n = 60). IO, isolated oligohydramnios. HC, healthy controls Contribution analysis of blood metabolites The variable importance plot was analyzed using the OPLS-DA model, which is the most common method for evaluating the contribution of variables. As shown in Fig. [69]2, all metabolites in discovery set were ranked by variable importance in projection (VIP), and the top 10 metabolites with VIP > 1 were identified as important, including phenylalanine (Phe), proline (Pro), methylglutarylcarnitine (C6DC), alanine (Ala), hydroxy isovalerylcarnitine (C5-OH), glycine (Gly), tryptophan (Trp), leucine (Leu), histidine (His), tyrosine (Tyr). Fig. 2. [70]Fig. 2 [71]Open in a new tab The variable importance plot of blood metabolites with VIP > 1. The X-axis represents the variable importance plot (VIP) score, and the Y-axis represents the compounds. Red and blue colors represent increased and decreased levels of metabolites, respectively. IO, isolated oligohydramnios. HC, healthy controls Violin plot analysis of important metabolites The levels of 10 important metabolites in the two groups in discovery set were compared using the Mann–Whitney U test and are shown in Table [72]2. A violin plot of the 10 metabolites is shown in Supplementary Fig. [73]1. The levels of Phe, Pro, C6DC, Ala, C5-OH, Gly, Trp, Leu, His, and Tyr were significantly different between the IO and HC groups in discovery set (FDR < 0.001). Table 2. Levels of important metabolites in the two groups in discovery set Material (µmol/L) HC M (P25, P75) IO M (P25, P75) P-value FDR Phe 55.89 (48.68, 62.30) 42.23 (34.68, 48.86) <0.001 <0.001 Pro 152.16 (124.41, 175.54) 100.38 (89.40,126.22) <0.001 <0.001 C6DC 0.02 (0.02, 0.03) 0.01 (0.01, 0.02) <0.001 <0.001 Ala 254.50 (217.02, 294.90) 203.89 (168.54, 232.07) <0.001 <0.001 C5-OH 0.22 (0.17, 0.26) 0.16 (0.13, 0.19) <0.001 <0.001 Gly 364.42 (306.08, 457.77) 296.14 (257.78, 336.29) <0.001 <0.001 Trp 15.90 (12.79, 17.52) 12.50 (10.44, 14.20) <0.001 <0.001 Leu 109.11 (90.63, 128.34) 87.51 (75.66, 102.04) <0.001 <0.001 His 80.54 (55.95, 111.95) 55.30 (41.13, 75.82) <0.001 <0.001 Tyr 62.52 (51.79, 79.98) 43.28 (35.39, 57.03) <0.001 <0.001 [74]Open in a new tab IO, isolated oligohydramnios. HC, healthy controls Metabolite set enrichment analysis Ten important metabolites in discovery set were screened out, and pathway analysis was conducted using MetaboAnalyst 6.0 to identify the metabolic pathways associated with IO. A dot plot was drawn based on Kyoto Encyclopedia Genes and Genomes human metabolic pathways (Fig. [75]3). The match status for the metabolites and pathways were shown in Table [76]3. The closer the pathway was to the upper right, the more reliable it was. Phe, Tyr and Trp biosynthesis was the most significantly enriched pathway in the IO group compared to the HC group (P < 0.01), and Phe, Tyr and Trp were involved in this pathway. Phe, Tyr and Trp biosynthesis and scatter plots of Phe and Tyr in the IO and HC are shown in Fig. [77]4. The Phe and Tyr levels were significantly lower in IO neonates in discovery set and were verified in the validation set. Fig. 3. [78]Fig. 3 [79]Open in a new tab Dot pot of pathway analysis. The X-axis represents the pathway impact value from the pathway topology analysis. Pathway impact is presented by the size of the dot and the larger the point, the more important the pathway is. The Y-axis represents -log10(p) and the p-value was from the pathway enrichment analysis. The color of the point represents -log10(p), and the redder the color, the higher the significance of the pathway (the smaller the P value) Table 3. Match status for the metabolites and pathways Pathway name Match status P -log(P) FDR Impact Phenylalanine, tyrosine and tryptophan biosynthesis 2/4 1.012E-4 3.995 0.008 1.000 Phenylalanine metabolism 2/8 4.684E-4 3.330 0.019 0.357 Valinleucine and isoleucine biosynthesis 1/8 0.035 1.455 0.921 0.000 Histidine metabolism 1/16 0.069 1.161 0.921 0.221 Glycine, serine and threonine metabolism 1/33 0.138 0.860 0.921 0.260 [80]Open in a new tab Fig. 4. [81]Fig. 4 [82]Open in a new tab Phenylalanine (Phe), tyrosine (Tyr) and tryptophan (Trp) biosynthesis and scatter plot of Trp, Phe and Tyr in IO and HC. *** P < 0.001, ** P < 0.01. “C numbers” is the KEGG identifier. IO, isolated oligohydramnios. HC, healthy controls ROC curve analysis of specific metabolites ROC curve analysis of specific metabolites was used to identify IO neonates in validation set (Table [83]4). The ROC curves of Phe and Tyr in the validation set for IO are plotted (Fig. [84]5). The AUC of Phe was greater than 0.8 indicating a good ability to discriminate IO. These results suggest that Phe could be the specific metabolite in IO but a larger sample size is needed to further validate it before clinical application. Table 4. ROC analysis of metabolites Metabolites AUC 95%CI P-value Sensitivity (%) Specificity (%) Associated criterion (µmol/L) Phe 0.814 0.733–0.895 < 0.001 78.00 81.70 ≤ 54.55 Tyr 0.638 0.534–0.741 0.013 46.00 80.00 ≤ 59.93 Trp 0.618 0.513–0.724 0.033 92.00 38.30 ≤ 18.43 [85]Open in a new tab Fig. 5. [86]Fig. 5 [87]Open in a new tab ROC curve of Phe, Tyr and Trp for IO. The closer the curve is to the upper left corner, the larger the area under curve (AUC), indicating a higher prediction accuracy. Phe shows a better ability to identify IO with greater AUC Discussion Currently, there are few studies on the association between IO and neonatal disease, and whether IO is strongly associated with fetal health is unclear. In obstetrics, IO may cause concerns, such as intrauterine hypoxia, and to some extent, may increase interventions, such as cesarean Sects. [[88]10, [89]11]. Therefore, there is an urgent need to determine the relationship between the IO and fetal health. Currently, there are no studies on IO neonatal metabolism. Our data point to the differential metabolites of IO neonates, thus providing clues for studying the effects of IO on fetal health. In this study, all samples were tested on the experimental platform of the Sixth Affiliated Hospital, Sun Yat-sen University, to ensure data comparability. The clinical characteristics showed no significant differences between the IO and HC groups in either the discovery or validation sets, suggesting that the metabolites were comparable and had reliable results. We found that blood metabolic profiles could significantly distinguish IO neonates from HC neonates. Ten significantly different metabolites were identified, and that were mainly enriched in the Phe, Tyr and Trp biosynthesis. Phe, Tyr and Trp, which were significantly lower in the IO group than in the HC group in both the discovery and validation sets, are involved in the Phe, Tyr and Trp biosynthesis. After verification by AUC in the validation set, Phe could be used as a specific biomarker for IO neonates. IO may be related to reduced placental perfusion, which may affect the exchange of nutrients between mother and fetus, and affect the fetal acquisition of adequate essential amino acid, such as Phe, Tyr and Trp [[90]9]. As is well known, Trp, Tyr, and Phe (TTP) are important precursors of monoaminergic neurotransmitters, including serotonin (5-HT), dopamine (DA), norepinephrine (NE), and epinephrine. Monoamines are involved in a wide range of physiological processes, and are especially important for behavioral and emotional regulation [[91]17–[92]23]. In humans, mood is indirectly associated with the serotonin, norepinephrine and dopamine levels [[93]24]. Low concentrations of TTP were closely associated with depressive symptoms and behaviors in humans [[94]25–[95]27]. In the brain tissue and feces of depressed rats, TTP levels were also found to be reduced [[96]28, [97]29]. The intake of an essential amino acid mixture lacking catecholamine precursors, Phe and Tyr, reduces serotonin and/or catecholamine neurotransmission and increases vulnerability to lowered mood in healthy females [[98]30]. In animal experiments, mice receiving dietary TTP deprivation showed reduced food intake and body weight and increased locomotor activity (“hyperactivity”) [[99]31, [100]32]. TTP-deprived mice have decreased serum TTP levels with a concomitant reduction in hippocampal Phe and Tyr levels and decreased 5-HT, DA, and NE concentrations in some brain regions [[101]31, [102]33]. Prolonged TTP depletion may have a profound effect on the catecholaminergic system in the brain. In the present study, Phe, Tyr and Trp biosynthesis was the most prominent metabolic pathway in neonates with IO. Phe, Tyr and Trp levels were significantly lower in neonates with IO than in HC. Given the effects of low-level Phe, Tyr and Trp on the brain catecholaminergic system, nervous system development in newborns with IO, especially in the behavior and emotion, may therefore be involved and should receive more attention during long-term follow-up. Evidence suggests that the gut microbiota is actively involved in aromatic amino acids metabolism and microbial metabolites of that serve as signaling molecules linking the gut with distant organs [[103]34]. Aromatic amino acids may influence host behavioral properties by regulating the “microbe-gut-brain axis“ [[104]35, [105]36]. Serum Trp concentration was decreased in patients with Crohn’s disease and ulcerative colitis, which was negatively correlated with disease activity [[106]37]. Dietary tryptophan was proved to alleviate dextran sodium sulfate-induced colitis in mice [[107]38]. The experiments revealed a significant reduction in the Roseburia genus in TTP-deprived mice in terms of the relative abundance of bacterial composition [[108]31]. Notably, Roseburia intestinalis is a potential biomarker of several diseases (including irritable bowel syndrome [[109]39], type 2 diabetes [[110]40], immune-mediated inflammatory disease [[111]41], and nervous system condition [[112]42, [113]43]. Our study showed lower levels of Phe and Tyr in neonates with IO, which may reduce the bacterial composition in the gut and become a risk factor for gastrointestinal problems during long-term growth and development. Metabolic profiling revealed a significant difference in the blood metabolism between newborns with IO and HC. Phe, Tyr and Trp biosynthesis were the most prominent metabolic pathways in IO neonates with low levels of of Phe and Tyr, which was verified in the validation set. Phe can be used as a biomarker specific for IO neonates with good sensitivity and specificity in ROC curve analysis. Phe, Tyr and Trp biosynthesis and low levels of Phe and Tyr in neonates with IO are the major difference in blood metabolomics profiles compared to HC. Phe, Tyr, and Trp biosynthesis may be involved in the physiological processes related to the nervous system and digestive system. Thus, continuous attention to changes in blood metabolites in IO neonates and the effect on body growth and development is necessary. In this study, we revealed the clinical phenomenon of differential metabolic markers between IO and HC. However, it is not determined how these different metabolites and pathway affect the organism, such as the nervous system and the digestive system. In order to ensure adequate feeding in early life, all newborns have adopted mixed feeding methods, but the ratio of breast milk to formula milk is not accurately uniform, which may bring about some subtle differences in metabolic substances. However, there was a study that compared the urine metabolic fingerprint of healthy neonates exclusively breastfed with that of neonates fed with formula and found no differences at the 3rd and 15th days of life [[114]44]. Also, in validation set, we confirmed the metabolic profile differences found in the discovery set. Further studies should be conducted to determine how these alterations in blood metabolites affect the human body and even cause disease. Conclusions In neonates with IO, 10 differential metabolites were identified using metabolic analysis; these metabolites were mainly enriched in one metabolic pathway, Phe, Tyr and Trp biosynthesis, with low levels of Phe and Tyr, which was verified in the validation set. Phe, Tyr, and Trp biosynthesis may be involved in the physiological processes related to the nervous system and digestive system. Metabolic profile changes in neonates with IO can provide valuable information and help children with monitor their health during development. Electronic supplementary material Below is the link to the electronic supplementary material. [115]Supplementary Material 1^ (54.2KB, xlsx) [116]Supplementary Material 2^ (1.1MB, png) Acknowledgements