Blood Amino Acid Fingerprints Reveal Hidden Metabolic Shifts in First-Episode Schizophrenia

Schizophrenia has long been studied through the lens of brain circuits, dopamine, and genetics, but a growing body of evidence suggests that the answers to some of psychiatry’s most stubborn questions may also be circulating in the blood. A new case-control study published in the journal Schizophrenia has mapped the plasma amino acid profiles of people experiencing schizophrenia for the first time, and the results paint a striking picture of systemic metabolic disruption. Crucially, the researchers studied patients who had never yet received antipsychotic medication, removing one of the most persistent confounders in biological psychiatry research: the drugs themselves can reshape metabolism, making it difficult to know whether observed abnormalities are part of the illness or a consequence of its treatment.

The research team, led by Yong Xue, Li Wang, and Xin Wang of Huai’an No. 3 People’s Hospital, together with colleagues at several Chinese institutions and corresponding author Bo Zhang of Shandong Provincial Hospital, recruited 80 antipsychotic-naïve patients with first-episode schizophrenia and 40 healthy controls. Using ultra-performance liquid chromatography coupled with tandem mass spectrometry, one of the most sensitive and precise analytical platforms available for small-molecule quantification, they measured the plasma concentrations of 31 amino acids. This technology allows researchers to separate and quantify dozens of chemically related compounds simultaneously, providing a comprehensive snapshot of nitrogen metabolism, neurotransmitter precursor availability, and detoxification pathways in a single blood draw.

The findings were unambiguous. After applying the Benjamini–Hochberg procedure to control the false discovery rate across multiple statistical comparisons, eleven amino acids differed significantly between patients and controls. Patients showed lower levels of kynurenine, arginine, homoarginine, aminoadipic acid, and methionine, alongside elevated levels of ornithine, alanine, glutamate, serine, glycine, and gamma-aminobutyric acid, better known as GABA. Each of these molecules participates in biochemical pathways with plausible and well-studied connections to brain function, which is precisely what makes the pattern so compelling to neuroscientists.

Consider what each cluster represents. Arginine and its relatives sit at the center of nitric oxide signaling, a pathway that regulates vascular tone, neurotransmission, and neuroinflammation. The depletion of arginine and homoarginine, combined with elevated ornithine, suggests a shift in how the urea cycle and nitric oxide pathways are balanced in first-episode patients. The researchers captured this by computing amino acid-derived ratios, and several of them differed clearly between groups: the arginine/ornithine ratio, the citrulline/arginine ratio, and the tyrosine/phenylalanine ratio all showed significant between-group differences. The tyrosine/phenylalanine ratio is particularly interesting because tyrosine is the direct precursor of dopamine, the neurotransmitter most closely implicated in psychotic symptoms, and this ratio reflects how efficiently the body converts dietary phenylalanine into that critical precursor.

Equally notable was the finding of reduced kynurenine alongside, by implication, altered kynurenine/tryptophan dynamics. The kynurenine pathway is the major route by which tryptophan, the amino acid precursor of serotonin, is broken down in the body, and its intermediates include compounds that are themselves neuroactive, capable of modulating glutamate receptors and contributing to oxidative stress. Disturbances in this pathway have been repeatedly linked to psychiatric illness, but demonstrating them in medication-naïve patients at the very onset of psychosis strengthens the argument that they are tied to the underlying disease process rather than to years of pharmacological exposure.

Perhaps the most eye-catching result concerns the neurotransmitter-related amino acids. Patients displayed elevated glutamate and elevated GABA, together with higher serine and glycine, both of which act as co-agonists at the NMDA glutamate receptor. The glutamate hypothesis of schizophrenia, which holds that dysfunction at NMDA receptors may underlie positive, negative, and cognitive symptoms, is one of the most influential frameworks in contemporary schizophrenia research. A peripheral signature involving both glutamate and GABA in untreated, first-episode patients offers a measurable biochemical correlate of that theoretical framework, one that could eventually be tracked with a routine blood test rather than an invasive procedure.

To test whether these metabolic differences could actually distinguish patients from healthy individuals, the team built a multivariable model based on the amino acid measurements. On the full dataset, the model achieved an apparent area under the receiver operating characteristic curve, or AUC, of 0.949, a value that would suggest near-perfect discrimination. But the researchers went further, applying a repeated stratified 10-fold cross-validation scheme in which the discriminatory features were reselected independently within each training subset, a rigorous design that guards against the optimism and overfitting that plague many biomarker studies. Under this stricter evaluation, the mean out-of-fold AUC settled at 0.869 with a standard deviation of just 0.011, indicating strong, stable, and honest discriminatory performance rather than a statistical artifact.

The clinical correlates added an exploratory but provocative dimension. Among the 26 patients for whom paired symptom data were available, the researchers examined whether baseline amino acid measures predicted improvement on the Positive and Negative Syndrome Scale, or PANSS, following treatment. The baseline glutamate/GABA ratio emerged as negatively associated with improvement in positive symptoms, with a Spearman correlation coefficient of −0.618 that survived false discovery rate correction, yielding a corrected q value of 0.015. In other words, patients whose blood showed a relatively higher glutamate-to-GABA balance before treatment tended to show less improvement in hallucinations, delusions, and other positive symptoms afterward. This hints at the possibility that a simple metabolic ratio measured on day one could eventually help clinicians anticipate how a patient will respond, a goal that psychiatry has pursued for decades.

The researchers were careful, however, to signal the limits of that finding. When they additionally adjusted the association for each patient’s baseline PANSS positive symptom score, the relationship was attenuated and no longer statistically significant. This means the glutamate/GABA ratio may partly reflect the severity of positive symptoms at the outset rather than independently predicting the trajectory of recovery, and the authors are transparent about that ambiguity. It is a level of statistical candor that stands out in a field where exploratory correlations are often oversold, and it underscores why the team frames its clinical findings as hypotheses for validation rather than as ready-made clinical tools.

That caution extends to the study as a whole. The case-control design cannot establish whether the observed amino acid abnormalities play a causal role in psychosis, are consequences of the illness state, or reflect shared upstream factors such as diet, stress physiology, gut microbiome composition, or inflammation. The sample was drawn from a single region of China, and replication in independent cohorts, ideally across different populations and with larger numbers of patients providing paired symptom data, will be essential before the findings can inform practice. The study received no dedicated funding, and the authors declare no competing interests. The article was published open access on 19 October 2026, with the DOI 10.1038/s41537-026-00817-z, allowing researchers anywhere to scrutinize the data and methods in full.

Even with those caveats, the significance of the work is hard to overstate. By studying patients at the first episode and before any antipsychotic exposure, the researchers have captured a metabolic snapshot of schizophrenia in something close to its raw form, and that snapshot reveals coordinated abnormalities spanning arginine-related metabolism, kynurenine pathway activity, aromatic amino acid conversion, and glutamate/GABA balance. The cross-validated biomarker model demonstrates that this information is not merely statistically detectable but diagnostically informative. If future studies replicate these profiles and clarify their relationship to symptom trajectories, a routine amino acid panel could one day complement clinical assessment, helping to identify schizophrenia earlier, stratify patients by metabolic subtype, and perhaps even point toward interventions that target the peripheral chemistry of the illness. For a disorder that has resisted decades of biological investigation, the idea that a simple blood test might carry meaningful clues represents exactly the kind of fresh angle that psychiatry needs.

Subject of Research: Plasma amino acid profiles in antipsychotic-naïve first-episode schizophrenia and their clinical correlates

Article Title: Pretreatment plasma amino acid profiles in first-episode schizophrenia: a case-control study with exploratory clinical correlates

Article References: Xue, Y., Wang, L., Wang, X., Qiu, Y. Z., Ren, Z. X., & Zhang, B. (2026). Pretreatment plasma amino acid profiles in first-episode schizophrenia: a case-control study with exploratory clinical correlates. Schizophrenia. https://doi.org/10.1038/s41537-026-00817-z

Image Credits: AI Generated

DOI: 10.1038/s41537-026-00817-z

Keywords: schizophrenia, amino acids, first-episode psychosis, metabolomics, glutamate, GABA, kynurenine pathway, arginine, biomarkers, PANSS, case-control study, liquid chromatography-mass spectrometry