The predictive potential and relationship of peripheral non-enzymatic antioxidants with clinical symptoms in first-episode male adolescent-onset schizophrenia patients

Abstract

Background

Oxidative stress is associated with the pathophysiology of schizophrenia. Peripheral non-enzymatic antioxidants can reflect the degree of oxidative stress and antioxidant levels, and can be used as markers of oxidative stress. However, the relationships between these markers and the clinical phenotypes of first -episode adolescent-onset schizophrenia (AOS) is unclear. This study aimed to elucidate the impact of peripheral non - enzymatic antioxidants on male AOS patients and their association with clinical symptoms.

Methods

Serum albumin (ALB), total bilirubin (TBIL), and uric acid (UA) of 59 first-episode AOS patients and 55 healthy controls were measured respectively. Furthermore, schizophrenia-related clinical symptoms were evaluated using the five-factor model of the Positive and Negative Syndrome Scale (PANSS).

Results

In Han Chinese first-episode male AOS patients, levels of TBIL and UA were significantly increased (p < 0.05), while ALB decreased (p < 0.01) compared with healthy controls. Furthermore, multivariate logistic regression showed that UA, ALB, and TBIL (all p < 0.01) were independently associated with AOS. Moreover, correlation analysis revealed that ALB was positively correlated with the PANSS total, cognitive and excited factor scores (all p < 0.05); TBIL was positively correlated with the excited factor score (p < 0.01); UA was positively correlated with the PANSS total, negative, cognitive and excited factor scores (all p < 0.05). In addition, Receiver operating characteristics assessment indicated that ALB and TBIL could effectively distinguish AOS patients from healthy controls.

Conclusions

This study confirms the association of peripheral non-enzymatic antioxidants (ALB, TBIL, and UA) with clinical manifestations and pathophysiology in schizophrenia. ALB and TBIL may serve as potential biomarkers for oxidative stress and symptom severity in AOS, particularly in males, providing insights for diagnostic and therapeutic strategies.

Introduction

Schizophrenia is a chronic psychiatric disorder characterized by high disability and clinically heterogeneous manifestations [1]. Current research has yet to fully elucidate its etiology and pathophysiological mechanisms, and reliable objective biomarkers remain lacking [1, 2]. Schizophrenia with onset between 13 and 18 years is clinically defined as adolescent-onset schizophrenia (AOS) [3, 4]. AOS is observed 5 times less often than adult-type schizophrenia, the incidence of which is considered to be 0.5‒1% [4]. Given that AOS patients are in a critical stage of neurodevelopment, they typically present with more complex, variable, and severe symptomatology compared to adult-onset cases, resulting in greater therapeutic challenges [5]. Therefore, identifying reliable biological markers and investigating their associations with clinical manifestations are crucial for advancing our understanding of the causes and underlying mechanisms of schizophrenia, as well as for improving diagnostic accuracy and therapeutic strategies, particularly for AOS.

Recent studies have increasingly indicated that oxidative stress plays a significant role in the pathogenesis and progression of schizophrenia. An imbalance between oxidative and antioxidant systems is thought to contribute to the onset of the disorder [6, 7]. Specifically, elevated levels of oxidative stress markers have been observed in schizophrenia patients. For instance, Yao et al. [8] demonstrated through post - mortem studies that nitric oxide (NO) levels in the caudate nucleus region were significantly higher in schizophrenia patients compared to healthy controls. Moreover, notable differences in NO levels across various brain regions were identified. Subsequently, Radonjić et al. [9] further validated these findings. They utilized an animal model of schizophrenia by administering phencyclidine (PCP) to Wistar rats during the perinatal period. Their results revealed significantly elevated levels of lipid peroxidation in the hippocampus and thalamus of the modeled rats, thus providing additional support for the role of oxidative stress in schizophrenia.

The human body is equipped with two major types of antioxidant systems: enzymatic and non-enzymatic antioxidants. Enzymatic antioxidants represent the primary intracellular antioxidant defense system, with three key enzymes playing pivotal roles: superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GSH - Px). Conversely, non - enzymatic antioxidants serve as the main extracellular antioxidant system, primarily found in plasma, cerebrospinal fluid, and other extracellular fluids. These include serum albumin (ALB), total bilirubin (TBIL), and uric acid (UA), vitamins A and C, tocopherols, and glutathione. However, the detection processes for enzymatic antioxidants are often labor-intensive and costly, significantly limiting their clinical application. Among non-enzymatic antioxidants, serum ALB, TBIL, and UA are commonly measured markers in plasma, contributing to over 85% of the total antioxidant capacity in plasma. In addition, the three indices exhibit strong antioxidant capabilities in the body, are easily obtained because they are included in the routine bloodwork examinations performed in hospitals, and can effectively reflect the degree of oxidative stress and antioxidant levels, making them valuable biomarkers of oxidative stress [10, 11].

Current studies on oxidative stress in schizophrenia mainly focus on enzymatic antioxidants, while limited studies have been carried out on non-enzymatic antioxidants. Significant heterogeneity exists among studies, and some results are even contradictory. For instance, several studies have shown that the serum UA and TBIL levels were elevated in schizophrenia patients during the acute phase compared to healthy controls [12, 13], whereas serum ALB levels were decreased [12, 14]. Conversely, other studies indicated that ALB, TBIL, and UA levels are significantly lower in schizophrenia patients than in healthy controls [14–16]. Furthermore, existing evidence suggests that antioxidant levels are associated with clinical severity in schizophrenia [17–19], Moreover, greater oxidative stress may result in poorer prognoses in patients with first episodes of early onset psychosis [20, 21]. Notably, most of the aforementioned studies were conducted on adults, and the relevant research on adolescents is extremely scarce. Given the differences in cytokines produced by children, adolescents, and adults, the previous research findings in the adult patient group may not be applicable to children and adolescents [22].

Oxidative stress exhibits significant sex differences. Studies have shown that male patients with schizophrenia generally exhibit higher levels of oxidative stress compared to female patients [23, 24]. Furthermore, under physiological conditions, the normal reference values for uric acid levels differ between males and females, which may be attributed to the fact that estrogen in females promotes the renal excretion of UA, whereas androgens in males encourage renal reabsorption of UA, leading to relatively higher uric acid levels in males [25]. Additionally, differences in ALB and TBIL levels between sexes may be related to disparities in hormone levels, particularly estrogen and androgen levels, as well as variations in liver function during different age stages [26–29]. Current research indicates that in psychiatric disorders, particularly in patients with schizophrenia, male patients exhibit higher ALB and TBIL levels compared to female patients [30]. Moreover, in untreated male schizophrenia patients, UA levels were found to be significantly lower than those in female patients [31]. It is worth noting that menstrual cycles also influence the body’s oxidative stress status [32, 33]. However, previous studies on oxidative stress in schizophrenia have rarely considered the impact of sex and this factor. Beyond sex and menstrual cycles, several other factors may influence antioxidant status. First, age is closely associated with oxidative damage [34]; second, medications [35, 36] can affect oxidative stress status; third, dietary patterns, caloric intake, alcohol, and cigarette consumption can influence antioxidant systems and the production of free radicals [37, 38].

Compared to adult-onset schizophrenia, AOS patients typically present with a more severe neurobiological subtype characterized by greater genetic susceptibility [39, 40]. This biological profile further leads to more severe clinical manifestations, a prolonged disease course, higher relapse rates, poorer prognosis, and impaired social functioning, often requiring higher doses of antipsychotic medications in clinical practice [41]. Notably, first-episode AOS patients are less influenced by environmental factors such as familial stress and social stressors. Therefore, research focusing on AOS provides a valuable opportunity to better elucidate the underlying etiology and pathogenesis of schizophrenia [42].

Therefore, this study focuses on peripheral non-enzymatic antioxidants while controlling for potential confounding factors that may influence oxidative stress. We specifically investigated first-episode, drug-naïve male AOS patients in the Han Chinese population to elucidate the role of peripheral non-enzymatic antioxidants in schizophrenia pathogenesis and their association with symptom severity, as assessed by the PANSS. We hypothesized that peripheral non-enzymatic antioxidants (ALB, TBIL, UA) — which reflect oxidative stress levels— could serve as potential biomarkers to distinguish AOS patients from healthy controls and would correlate with clinical symptom severity. Understanding the relationship between non-enzymatic antioxidants and clinical manifestations of AOS may provide critical insights into the biological mechanisms underlying schizophrenia and potentially reveal novel therapeutic targets.

Materials and methods

Participants

All the subjects were recruited from Nanyang Fourth People’s Hospital, the only tertiary psychiatric specialty hospital and the largest mental health institution in Nanyang City, Henan Province, serving a population of 10 million residents. The study cohort comprised 59 treatment-naïve male Han Chinese adolescents (aged 12–18 years) hospitalized between December 2023 and January 2025. A control group of 55 demographically matched healthy adolescents (sex-, age-, and ethnicity-matched) was concurrently enrolled during the same period.

Patients meeting the following criteria were included: (1) Diagnosis confirmed by the schizophrenia diagnostic criteria proposed in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition(DSM-5); (2) Total illness duration < 2 years since onset; (3) Antipsychotic-naïve status (no prior exposure to psychotropic medications); (4) Han Chinese ethnicity with ≥ 6 years of formal education; (5) Full-scale IQ > 70 on standardized assessment (assessed via the Wechsler Intelligence Scale, exclusively for this study).

Individuals were excluded if they had: (1) History of substance abuse (alcohol, tobacco, or illicit drugs); (2) Intellectual disability, autism spectrum disorder, or sensorineural deficits (e.g., deafness/mutism); (3) Chronic physical illnesses, including: (a) Systemic inflammatory conditions (acute/chronic infections, autoimmune disorders) (b) Endocrine/metabolic disorders (diabetes, thyroid dysfunction) (c) Febrile episodes within the past 4 weeks.

Data collection

Venous blood samples (5 mL) were collected from all participants between 7:00 and 8:00 a.m. after at least 8 h of overnight fasting. Samples were immediately transported to the clinical laboratory, centrifuged for serum separation, and analyzed for ALB, TBIL, and UA within 2 h of collection using a Mindray BS-2000 M Fully Automated Biochemical Analyzer (Shenzhen Mindray Bio-Medical Electronics Co., Ltd., China) according to the manufacturer’s protocols. This analyzer demonstrates excellent precision, with both within-run and between-run coefficients of variation being less than one-fourth of the allowable total error (TEa) limit defined by CLIA ‘88, thereby meeting the requirements for clinical testing. Furthermore, peripheral non-enzymatic antioxidants (ALB, TBIL, and UA) were assessed and compared between the two cohorts.

Clinical evaluation

Psychotic symptom severity was evaluated using the PANSS, which comprises 30 items [43], with subsequent analyses incorporating the five-factor model of the PANSS as proposed by Wallwork and colleagues [44] and the European Psychiatric Association’s updated guidelines for negative symptom assessment [45]. The model categorizes symptoms into five domains: the positive factor (PANSS-PS: P1, P3, P5, G9), the negative factor (PANSS-NS: N1, N2, N3, N4, N6) [45], the cognitive/disorganized factor (PANSS-CD: P2, N5, G11), the excited factor (PANSS-EX: P4, P7, G8, G14), and the depressed factor (PANSS-DP: G2, G3, G6). All items were scored on a 1–7 severity scale (higher scores indicating greater symptom intensity). To ensure reliability, two independently trained psychiatrists conducted PANSS evaluations, with the same psychiatrist per patient during treatment.

Ethical considerations

The study protocol and informed consent documentation were reviewed and approved by the Ethics Committee of the Fourth People’s Hospital of Nanyang City prior to implementation (Approval No. 2023-SP002). Written informed consent forms, detailing research objectives and procedures, were obtained from all participants aged ≤ 18 years and legal guardians/parents of minors, in accordance with the Declaration of Helsinki.

Statistical analysis

Statistical analysis was performed using SPSS 26.0. The normality of data distribution was assessed using the Kolmogorov-Smirnov test. Normally distributed data were expressed as mean ± standard deviation (mean ± SD), while non-normally distributed data were presented as median (interquartile range, IQR). First, univariate analysis was conducted to compare differences in general clinical characteristics and peripheral non-enzymatic antioxidant levels between male AOS patients and healthy controls. Subsequently, binary multivariate logistic regression analysis was performed to identify independent variables associated with AOS in males, with calculation of odds ratios (ORs) and 95% confidence intervals (CIs). Furthermore, Pearson correlation analysis was employed to examine relationships between peripheral non-enzymatic antioxidant levels and Positive and PANSS scores. Finally, the receiver operating characteristic (ROC) curve was utilized to evaluate the discriminative efficacy of peripheral non-enzymatic antioxidants in distinguishing between AOS patients and healthy controls. All statistical analyses were conducted using two-tailed tests, with a statistical significance threshold set at p < 0.05.

Results

This study enrolled a total of 114 male Han Chinese participants: 59 AOS patients and 55 age-matched healthy controls. The demographic and clinical characteristics of the study population are summarized in Table 1. No patient had a history of smoking and drinking. In the AOS group, there was a mean age of 16 (15, 17) years. Whereas the healthy control cohort comprises with a mean age of 17 (15, 17) years. Statistical analysis revealed no significant differences between the two groups in terms of weight, height, and body mass index (p > 0.05) (Table 1).

Table 1.

Characteristics of male patients with AOS and healthy comparison subjects

Variable	Patients with AOS (n = 59)	Healthy controls (n = 55)	t/z	p
Age (years)	16(15,17)a	17(15,17)a	−0.567	0.571b
Race	Han Chinese	Han Chinese	-	-
Smoker	No	No	-	-
Drinker	No	No	-	-
Family genetic history	20(33.9%)	-	-	-
Disease duration (Months)	2(1,12)	-	-	-
Weight (kg)	60.68 ± 10.13c	62.31 ± 10.54c	−0.842	0.402d
Height (m)	1.70 ± 0.06	1.72 ± 0.07	−1.237	0.218
BMI (kg/m2)	20.95 ± 3.23	21.10 ± 2.89	−2.262	0.794
PNASS-TS	115.10 ± 8.54	-	-	-
PANSS-PFS	16.32 ± 2.83	-	-	-
PANSS-NFS	20.03 ± 4.01	-	-	-
PANSS-CFS	13.98 ± 1.70	-	-	-
PANSS-EFS	17.93 ± 3.08	-	-	-
PANSS-DFS	8.25 ± 2.01	-	-	-

AOS: Adolescent-onset Schizophrenia; BMI: Body mass index; PANSS: Positive and Negative Syndrome Scale; PNASS Total Score: PNASS-TS; PANSS-PFS: Positive Factor Score; PANSS-NFS: Negative Factor Score; PANSS-CFS: Concrete Factor Score; PANSS-EFS: Excited Factor Score; PANSS-DFS: Depress Factor Score; a, Expressed as median (lower–upper quartiles); b, Mann–Whitney U test; c, Expressed as mean ± standard deviation; d, Student’s t-test

Analysis of peripheral non-enzymatic biochemical parameters demonstrated significantly lower serum ALB levels in AOS patients compared to healthy controls (p = 0.003). Conversely, TBIL (p = 0.012) and UA (p = 0.032) levels were significantly elevated in the AOS group (Table 2). Binary multivariate logistic regression analysis identified ALB (OR: 0.724; 95% CI: 0.619–0.847; p < 0.001), TBIL (OR: 1.108; 95% CI: 1.028–1.195; p = 0.007), and UA (OR: 1.007; 95% CI: 1.002–1.012; p = 0.006) as independent factors associated with AOS (Table 3).

Table 2.

Comparison of peripheral levels of ALB, TBIL, and UA between male AOS patients and healthy controls

Variables	Patients with AOS	Healthy controls	t	p
ALB (g/L)	47.05 ± 3.98	49.03 ± 2.79	−3.090	0.003*
TBIL (µmol/L)	17.36 ± 7.01	14.33 ± 5.62	2.540	0.012*
UA (µmol/L)	423.32 ± 136.64	379.21 ± 71.62	2.179	0.032*

ALB: albumin; TBIL: total bilirubin; UA: uric acid; a, Expressed as mean ± standard deviation; b, Student’s t-test; * indicates statistical significance

Table 3.

Binary multivariate logistic regression model for the association between AOS and ALB, TBIL and UA

Variables	B	Std. error	Wald	p	95% CI for EXP (B)
					Exp (B)	Lower	Upper
ALB	−0.322	0.080	16.251	< 0.001*	0.724	0.619	0.847
TBIL	0.103	0.038	7.213	0.007*	1.108	1.028	1.195
UA	0.007	0.003	7.446	0.006*	1.007	1.002	1.012

ALB: albumin; TBIL: total bilirubin; UA: uric acid; * indicates statistical significance

In addition, correlation analyses were conducted to evaluate associations between peripheral non-enzymatic antioxidants (ALB, TBIL, and UA) and PANSS scores. Significant positive correlations emerged between ALB levels and total PANSS scores (r = 0.294, p = 0.024), cognitive factor scores (r = 0.324, p = 0.012), and excited factor scores (r = 0.273, p = 0.036). TBIL showed a significant association with excited factor scores (r = 0.357, p = 0.005). UA demonstrated robust correlations with total PANSS scores (r = 0.519, p < 0.001), negative factor scores (r = 0.259, p = 0.048), cognitive factor scores (r = 0.340, p = 0.009), and excited factor scores (r = 0.347, p = 0.007). Comprehensive correlation matrices are provided in Table 4.

Table 4.

Correlations between PNASS scores with ALB, TBIL and UA

Variables	PANSS-TS		PANSS-PFS		PANSS-NFS		PANSS-CFS		PANSS-EFS		PANSS-DFS
	r	p	r	p	r	p	r	p	r	p	r	p
ALB	0.294	0.024*	−0.100	0.453	0.106	0.424	0.324	0.012*	0.273	0.036*	0.017	0.896
TBIL	0.156	0.238	0.178	0.177	−0.132	0.318	0.144	0.275	0.357	0.005*	−0.226	0.085
UA	0.519	< 0.001*	0.004	0.975	0.259	0.048*	0.340	0.009*	0.347	0.007*	0.168	0.203

ALB: albumin; TBIL: total bilirubin; UA: uric acid; PANSS: Positive and Negative Syndrome Scale; PNASS Total Score: PNASS-TS; PANSS-PFS: Positive Factor Score; PANSS-NFS: Negative Factor Score; PANSS-CFS: Concrete Factor Score; PANSS-EFS: Excited Factor Score; PANSS-DFS: Depress Factor Score

Based on binary multifactor regression and ROC curve analysis, the diagnostic performance of ALB, TBIL and their combination in assessing AOS was evaluated. The results revealed that the area under the curve (AUC) for ALB was 0.670 (95% CI: 0.576–0.755, p = 0.001), and for TBIL was 0.625 (95% CI: 0.529–0.714, p = 0.017). The combined model demonstrated superior diagnostic performance with an AUC of 0.762 (95% CI: 0.673–0.837, p < 0.001). Optimal diagnostic thresholds with corresponding sensitivity and specificity were determined: ALB (cut-off = 46.7 g/L, sensitivity = 55.93%, specificity = 89.09%), and TBIL (cut-off = 21.0µmol/L, sensitivity = 28.81%, specificity = 94.55%). The combined model achieved balanced performance metrics (sensitivity = 61.02%, specificity = 81.82%) (Table 5; Fig. 1).

Table 5.

Receiver-operating characteristic (ROC) curve of ALB, TBIL, UA and combined model for differentiating male patients with AOS from healthy controls

Variables	AUC	95%CI	Cut-off	Sensitivity	Specificity	p
ALB	0.670	0.576–0.755	46.7	55.93	89.09	0.001*
TBIL	0.625	0.529–0.714	21.0	28.81	94.55	0.017*
UA	0.568	0.472–0.661	467	30.51	92.73	0.211
Combined	0.762	0.673–0.837	0.547	61.02	81.82	< 0.001*

ALB: albumin; TBIL: total bilirubin; UA: uric acid; Combined: Joint ALB andTBIL; * indicates statistical significance

Fig. 1.

The ROC curves of ALB, TBIL and combined model for differentiating male adolescents with schizophrenia from healthy controls. ALB is presented with a blue line, with an AUC of 67.0%, a sensitivity of 55.93%, and a specificity of 89.09%. TBIL is presented with a green line, with an AUC of 62.5%, a sensitivity of 28.81%, and a specificity of 94.55%. Combined model is presented with a orange line, with an AUC of 76.2%, a sensitivity of 61.02%, and a specificity of 81.82%

Discussion

This research revealed that (1) First - episode, non - medicated male Han Chinese AOS patients had lower ALB levels and higher TBIL and UA levels than the healthy control group; (2) Binary multivariate logistic analysis showed ALB, TBIL and UA were independently related to AOS; (3) Correlation analysis revealed a positive association of ALB with total PANSS, cognitive and excited factor scores; positive association of TBIL with PANSS excited factor scores; positive association of UA with total PANSS, negative, cognitive and excited factor scores; (4) ROC curve analysis indicated ALB, TBIL and the combined model could independently distinguish AOS patients from healthy adolescents.

ALB, as an important endogenous antioxidant in the human body, plays a critical role in scavenging free radicals, binding metal ions, and regulating reactive oxygen metabolism. Its serum level is considered a key biomarker reflecting the body’s oxidative stress status [46]. Our study found that the serum ALB levels of male adolescents with schizophrenia were significantly lower than those of the healthy controls, yet remained within the normal physiological range. This result is consistent with previous findings from multiple studies on adult patients [13, 31]. Notably, Lu, Z., et al. [13] also observed through longitudinal studies that ALB levels in patients with schizophrenia remained consistently lower than those in healthy individuals, whether in the acute or remission phase, and further decreased after antipsychotic drug treatment. This suggests that there may be continuous oxidative stress consumption during the disease process, while compensatory synthetic functions of the liver may maintain the relative stability of ALB levels. Although most studies support the association between decreased ALB levels and schizophrenia, notable discrepancies persist among different studies that warrant further investigation. For instance, Yin, X.Y., et al. [47] found a significant negative correlation between serum ALB levels and depressive symptoms in patients, which differs from the findings of the current study. It is speculated that these discrepancies may arise from differences in study populations, as their research focused on adult patients without stratification by sex and incomplete exclusion of medication effects. Additionally, evidence from studies focusing on adolescent populations remains scarce. Apart from the present study, only Zhai, D., et al. [48] have reported decreased ALB levels in adolescent schizophrenia patients; however, that study did not control for confounding factors such as sex and ethnicity, nor did it explore the correlation between ALB levels and clinical symptoms. Nevertheless, these results collectively underscore the critical role of oxidative stress in the pathogenesis of schizophrenia and further reinforce the hypothesis that reduced serum albumin levels may contribute to disease progression by impairing antioxidant defense mechanisms [49, 50].

It is noteworthy that during acute inflammatory states, the plasma concentration of albumin, a typical negative acute-phase reactant, often decreases [51]. The underlying reasons may be as follows: on one hand, the acute inflammatory response can lead to a relative reduction in albumin mRNA levels in the liver, resulting in a corresponding reduction in the liver’s ability to synthesize and secrete albumin [52]. On the other hand, systemic inflammation may contribute to malnutrition, as proinflammatory cytokines can reduce food intake and induce anorexia, which subsequently affects the rate of protein breakdown [51]. Furthermore, oxidative stress and immune-inflammatory responses interact and reinforce each other [53]. Accumulating evidence has indicated that immune inflammation plays a role in the pathogenesis and progression of schizophrenia [52]. Nevertheless, it is well-established that plasma albumin levels are reduced in patients with schizophrenia. Our findings further support this conclusion, although the precise mechanisms warrant further investigation.

TBIL, a terminal product of heme metabolism in the body, functions as an endogenous antioxidant with anti-inflammatory properties [53]. Although elevated serum concentrations may induce severe neurotoxicity, mildly increased bilirubin levels exert protective effects by enhancing systemic antioxidant capacity, primarily through scavenging reactive oxygen species (ROS) [54]. Abnormal levels of TBIL were also found in mental disorders [18]. Our findings revealed elevated TBIL levels in male adolescents with schizophrenia, consistent with the observations by Lu, Z., et al. [13]. Notably, their study identified a sex disparity in serum TBIL levels, with male schizophrenia patients exhibiting more pronounced elevations compared to females. However, their cohort comprised adult patients without strict control for confounding factors such as antipsychotic medication. Elevated serum TBIL may result from increased erythrocyte membrane fragility under oxidative stress [55], suggesting heightened oxidative stress in male adolescent schizophrenia patients. Nevertheless, conflicting findings have been reported in other studies [16, 56, 57], potentially attributable to sample heterogeneity or inadequate control of confounders such as age, sex, and medication. Further investigation is warranted to clarify these discrepancies. Correlation analyses demonstrated that TBIL levels in male AOS patients were positively correlated with the excited factor scores of the PANSS, but showed no significant association with other PANSS factor scores. This contrasts with findings from a Chinese study of adult schizophrenia patients, which reported an independent positive correlation between TBIL levels and cognitive function in males [58]. The underlying mechanisms for this divergence remain unclear and merit further exploration. To date, no studies have investigated the relationship between TBIL and clinical symptoms in AOS patients.

UA, as the end product of purine metabolism, plays a role in scavenging reactive free radicals. It inactivates reactive intermediates such as peroxynitrite, nitric oxide, and peroxyl radicals, thereby protecting erythrocyte membranes from lipid peroxidation [59]. This study found that serum UA levels were significantly elevated in first-episode male AOS patients compared to healthy controls, consistent with previous findings in adult first-episode schizophrenia patients [13] and male schizophrenia patients [60]. These observations suggest that increased UA levels in the acute phase may reflect a compensatory response to counteract oxidative stress, indicating its potential neuroprotective role [61], whereas persistent UA elevation could also serve as a biomarker of chronic oxidative stress [11]. Experimental studies in murine models demonstrated that exogenous UA administration suppresses neuroinflammation and oxidative stress in the central nervous system [62].However, a meta-analysis incorporating 61 studies reported significantly reduced UA levels in first-episode schizophrenia patients [6], while a cross-sectional study found no significant differences in UA levels between schizophrenia patients and controls [63], contrasting with our findings. Potential explanations for these discrepancies include uncontrolled confounding factors (e.g., sex, age, diet, smoking) and sample heterogeneity. Notably, a study focusing on first-episode adult schizophrenia identified significantly lower UA levels in antipsychotic-naïve male patients [31], further suggesting age-dependent variations in UA profiles among schizophrenia populations. Correlation analysis revealed positive associations between UA levels and total PANSS, cognitive factor, and excited factor scores. A recent study reported that UA positively correlated with cognitive function while negatively associating with negative symptoms in schizophrenia [64]. However, Gültekin, B.K., et al. found no significant correlations between UA and PANSS subscale scores [60], possibly due to unadjusted age or medication effects. Importantly, to date, no prior studies have specifically investigated UA levels in male adolescent schizophrenia populations.

In the multivariate regression analysis, ALB, TBIL, and UA demonstrated effectiveness in differentiating first-episode AOS patients from healthy controls. However, subsequent ROC curve analysis revealed that UA failed to achieve statistical significance (p > 0.05) in discriminative efficacy, suggesting that ALB and TBIL may hold potential diagnostic and predictive value for male AOS, whereas the clinical utility of UA appears limited. It is crucial to interpret these findings with caution: The heterogeneity of schizophrenia, individual variations in treatment response, and the functional limitations of peripheral non-enzymatic antioxidants preclude reliance on single biomarkers for diagnosis. Comprehensive clinical evaluation should integrate psychopathological manifestations, family genetic history, neuroimaging findings, and multidimensional biomarker profiling to ensure scientifically robust diagnostic decisions.

Regarding the non-significant discriminative performance of UA in ROC analysis, our research team proposes the following multidimensional explanations: First, the restricted sample size may have compromised statistical power. Second, potential confounding factors including population heterogeneity, dietary influences, and physical activity variations might have contributed. Third, the intrinsic association between uric acid and male adolescent schizophrenia could be insufficient to support its diagnostic relevance. Genetic evidence corroborates this observation- Mendelian randomization analysis of genome-wide association studies (GWAS) data from the Global Urate Genetics Consortium (GUGC) and Psychiatric Genomics Consortium (PGC) demonstrated no causal relationship between serum UA levels and schizophrenia pathogenesis. This indicates UA may neither function as a direct etiological factor nor a therapeutic target, though its potential role as a dynamic monitoring indicator for antipsychotic efficacy warrants exploration in prospective cohort studies [65]. Guided by this evidence, we optimized the composite diagnostic model by excluding UA. The refined ALB-TBIL combination model exhibited enhanced discriminative performance (sensitivity 61.02%, specificity 81.82%, AUC = 0.762). This strategic refinement not only improved model stability but also underscored the necessity of multidimensional biomarker screening for precision diagnostics in schizophrenia. Our findings emphasize that biomarker panel optimization through rigorous statistical validation is essential for developing clinically actionable diagnostic tools in complex psychiatric disorders.

Moreover, studies have demonstrated that antioxidant therapy may exert positive effects across various stages of schizophrenia, spanning prenatal and postnatal development to acute and chronic phases of the illness, while improving patients’ psychiatric symptoms [66]. Currently, well-researched antioxidants include N-acetylcysteine (NAC), Ginkgo biloba extract, selegiline, allopurinol, and vitamin E [67–69]. Our data provide evidence for the importance of monitoring and managing abnormal oxidative stress states, further validating the results of this study. Nevertheless, the specific clinical applications of antioxidants, particularly in AOS, still require in-depth understanding and verification by more studies and clinical trials.

Notably, correlation analysis revealed a significant association between peripheral non-enzymatic antioxidant levels and PANSS excited factor score in male AOS patients. This finding further supports existing evidence of sex differences in clinical manifestations of schizophrenia, where male patients exhibit greater impulsivity and aggressive behaviors compared to their female counterparts [70]. However, the precise mechanisms through which peripheral non-enzymatic antioxidants mediate these effects, as well as their role in oxidative stress pathways, remain to be fully elucidated.

Several limitations should be acknowledged in this study. First, the relatively small sample size necessitates that our findings be considered preliminary, warranting future validation in larger cohorts. Second, the cross-sectional design precludes determination of causal relationships between serum non-enzymatic antioxidants and PANSS scores in schizophrenia patients. Therefore, longitudinal and prospective follow-up studies are required to further validate these associations. Third, despite rigorous control of potential confounding factors affecting non-enzymatic antioxidant status, certain variables such as physical activity, sleep patterns, and dietary habits could not be fully standardized, potentially introducing bias into the experimental outcomes, suggesting more rigorous protocols to standardize these confounding variables and minimize experimental bias as much as possible. Finally, although sex differences may significantly affect the body’s antioxidant system function, the exclusion of female patients represents an undeniable limitation—particularly given the potential role of antioxidants as biomarkers and possible sex-specific variations. Future studies should include both sexes to explore potential sex-specific effects and better evaluate the biomarker potential of antioxidants in schizophrenia.

In conclusion, peripheral non-enzymatic antioxidants (ALB, TBIL, and UA) demonstrate independent associations with male AOS and its clinical symptoms, particularly those related to excitement. Furthermore, ALB and TBIL may serve as potential biomarkers for oxidative stress status and symptom severity in AOS. This study reinforces the critical relationship between oxidative stress and schizophrenia pathogenesis. These findings provide novel insights for the diagnosis, therapeutic management, and severity assessment of AOS, while offering empirical support for the clinical application of antioxidant therapies in schizophrenia treatment, particularly for male AOS patients.

Author contributions

LZH, MXB, and MJ designed and supervised the experiments; DBL, CHF, ZK, WZS, ZZ, and LHG performed the experiments; MJ and LZH analyzed the data. DBL, MJ, and LZH wrote and revised the manuscript; All authors reviewed the manuscript.

Funding

This study was funded by the Joint Construction Project of Medical Science and Technology Research Program in Henan Province, China (No.LHGJ20230963).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.