Determining Reference Intervals of Serum Biochemical Parameters in Juvenile Hybrid Snakehead Channa argus & C. maculata in Mesocosm

Abstract

Hybrid snakehead (Channa argus × Channa maculata) is a major cultured freshwater fish in China, but standardized health monitoring using serum biochemistry is limited by the lack of species-specific reference intervals. This study established reference intervals for 20 serum biochemical parameters in hybrid snakehead reared under 27 °C for 90 days. The body weights of the sampled fish ranged from 50 g to 160 g and were exempted from diseases by health check. All parameters were measured using an automated analyzer with commercial reagent kits. Most parameters exhibited non-normal, right-skewed distributions, and only total protein (TP) was normally distributed. Smoothed bootstrap resampling and kernel density estimation were applied to extract the main peak distribution and reduce bias from outliers and long tails. Species-specific reference intervals were established based on the main peak data, providing more reliable physiological baselines than conventional percentiles. Correlation analysis revealed coordinated changes among liver function, nutrient metabolism, tissue damage, and digestive enzymes. These results provide a standardized tool for health assessment, subclinical disease diagnosis, and comparative analysis in juvenile hybrid snakehead maintained at an optimal temperature in indoor mesocosm systems.

1. Introduction

Hybrid snakehead, derived from crossing Channa argus (northern snakehead) and Channa maculata (blotched snakehead), has emerged as a commercially important aquaculture species in China [1]. This hybrid exhibits good taste, better growth performance, and higher tolerance to diseases and environmental stressors compared to its parental species, making it an ideal species for fish production [1]. In recent years, the domestic production of hybrid snakehead in China has expanded rapidly to approximately 605,000 tons in 2023 [2]. However, concerns were also raised over animal welfare, disease control, and sustainability with the increasing density [3,4,5].

Health assessment is an indispensable daily routine in aquaculture farms to ensure fish welfare, optimize feeding strategies, and prevent economic losses. Traditionally, fish health status is evaluated through physical examinations, including measurements of body weight and length, targeted monitoring for expected diseases, and visual inspection of organ color and morphology (e.g., liver, intestine, gills), or simply by monitoring changes in feeding rate and behavior [6,7]. While these methods are straightforward and non-invasive, they locate rather late in the adverse outcome pathway. Physical signs, such as reduced feeding, abnormal swimming, or visible organ lesions, typically appear only when significant pathological changes have already occurred. By that time, interventions may be less effective, and mortality or growth retardation may be inevitable. For example, hepatic steatosis (fatty liver) in many fish species may not alter the feeding rate in its early stages but can already impair metabolic function and reduce disease resistance [8,9,10]. Therefore, relying solely on external observations is insufficient for the early detection of subclinical health issues, which are common in aquaculture ponds where chronic and multiple stress (e.g., crowding, poor water quality) predisposes fish to opportunistic infections without obvious external signs.

Serum biochemistry analysis offers a cheap, rapid, and minimally invasive tool for quantifying biochemical biomarkers in the blood, which can be used to assess the functional status of multiple internal organs [11,12]. Blood serum contains a variety of enzymes, metabolites, and proteins that reflect physiological and pathological conditions. For instance, elevated activities of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) in serum are indicative of hepatocellular damage, as these enzymes are normally confined within hepatocytes and leak into the bloodstream upon cell membrane rupture [13]. Similarly, increased total triglyceride (TG) and total cholesterol (TG) levels suggest lipid storage and metabolism [13]; while abnormalities in total protein, albumin, and globulin may reflect nutritional status or inflammatory responses [13]. Unlike physical examination, serum biochemical changes often precede observable clinical symptoms, enabling early diagnosis and timely intervention. The test itself requires only a small blood sample (e.g., 0.5–1 mL) and can be processed within hours using standard laboratory equipment, with a cost per sample typically below CNY 80 in China. In comparison, histopathology, the gold standard of diagnosis, takes much longer in both sampling and post-processing. These advantages have made serum biochemistry a routine screening tool in human and veterinary medicine.

Despite its widespread use in hospital and veterinary clinics, serum biochemistry analysis remains largely at the laboratory stage in aquaculture, with limited application in commercial farms [14,15]. The status quo can be attributed to three major obstacles. First, multiple intrinsic and extrinsic factors are known to jointly shape fish serum biochemical profiles. Biological characteristics, including age, weight, sex, and sexual maturation status, can cause obvious fluctuations in physiological indicators [16,17]. Meanwhile, environmental conditions such as water temperature, dissolved oxygen, photoperiod, and water quality parameters also exert profound impacts on serum metabolites [18,19,20,21]. In addition, rearing-related stressors, including stocking density, feeding regime, handling, and transportation, can alter serum biochemical levels [18,22,23,24]. Second, there is a lack of standardized protocols for blood sample collection and processing in fish. Factors such as the use of anticoagulants (heparin, EDTA, or none), the time between collection and centrifugation, and storage temperature can significantly influence the measured values of biochemical parameters [25,26,27]. Third, inconsistent bioassay methods produce disparate numerical results even for the same parameter in the same fish species. Some researchers use commercial reagent kits, while others employ automated biochemical analyzers that offer higher throughput and precision [28,29,30]. These two approaches yield different absolute values because reagent kits and auto-analyzers differ in their calibration standards, reaction conditions, and detection principles. This lack of comparability hinders meta-analyses and the establishment of universal reference ranges.

Another major hindrance to implementing serum biochemistry in fish aquaculture is the absence of reliable reference values or benchmarks for most cultured fish species. In human and veterinary medicine, reference intervals are well-established based on large cohorts of healthy individuals, accounting for variations due to age, sex, and physiological status [31]. In recent years, a growing number of studies have established serum biochemical reference intervals for commercially important farmed fish [32,33,34,35,36,37]. Atlantic salmon farming in Scandinavia has set a benchmark in this field, with serum biochemical analysis provided for fish farmers as a standard health management tool [32]. This standardized monitoring system has proven effective for modern intensive aquaculture. Nonetheless, such mature implementation remains scarce for other cultured species. Without reference intervals, veterinarians cannot interpret individual test results, and researchers find it difficult in comparative analysis across different studies. Consequently, the test results remain underutilized in decision-making. Importantly, hybrid snakehead, as an interspecific hybrid, may exhibit biochemical profiles distinct from either parent species [38]. Thus, species-specific reference intervals generated under standardized conditions are urgently needed to bridge the gap between laboratory research and on-farm application.

Given the above background, the aim of this study is to determine the reference values of serum biochemical parameters in hybrid snakehead. Blood samples were collected from fish fed in mesocosms for 90 days. Health status was confirmed by physical examination and by gross inspection of internal organs upon sampling. Parameters to be measured include liver function markers, nutritional, and metabolic markers. Depending on the distribution of each parameter, reference intervals are calculated using either nonparametric methods. The resulting benchmarks will provide a practical tool for health assessment, detect subclinical disorders, and guide therapeutic decisions in fish aquaculture.

2. Materials and Methods

2.1. Fish and Sample Collection

Hybrid snakehead (Channa maculata × C. argus) were used as the experimental model. Fish fry (2.9 ± 0.3 g, n = 1200) were purchased from a commercial breeding farm (Guangdong Bairong Seed Group, Zhaoqing, China) and transported to the Haid Aqua R&D Base (Guangzhou, China). Fish were evenly distributed into 4 round tanks (diameter 3 m; depth 1 m; n = 300) with thorough aeration. The tanks were built indoors and only equipped with aeration and water-exchange facilities. Following a 2-day adaptation, fish were fed an average of 4% body weight with a commercial pellet diet (

Table 1. Feed formulation and proximate nutritional composition of the experimental diet.

Ingredients	Content (g/kg)	Nutritional Level	Calculated Value
Steam Fish Meal	400	Crude Protein	49.13%
Casein	245	Crude Fat	10.90%
Corn Starch	210	Crude Ash	9.59%
Squid Visceral Meal	17	Crude Fiber	4.56%
Fish Oil	61.6	Total Phosphorus	1.67%
Lecithin	5	Lysine	3.70%
Choline Chloride	0.4	Moisture	9.14%
Vitamin C (Ascorbic Acid)	1
Vitamin Premix	3.5
Mineral Premix	3
Ethoxyquin (antioxidant)	0.1
Calcium Dihydrogen Phosphate	10
Microcrystalline Cellulose	43.4
Total	1000

Note: All values are expressed as g per kg of diet. Nutritional levels were calculated based on the nutrient contents of individual ingredients.

) at 8:00 and 16:00 daily. Water quality parameters were as follows: temperature 27.3 ± 2.2 °C, pH 7.7 ± 0.5, dissolved oxygen 6.5 ± 0.9 mg/L, and ammonia nitrogen < 0.50 mg/L.

After 90 days of culture, 40 clinically healthy fish with normal feeding and swimming activity were sampled from each tank before feeding. The fish were euthanized by cranial stunning and central nervous system pithing [39]. Caudal vein blood was collected using anticoagulant-free sterile syringes, after which individual body and liver weights were measured. Gross anatomical examination of the liver, intestine, gallbladder, and spleen was conducted to rule out visible diseases. Whole blood was transferred to RNase-free centrifuge tubes and allowed to clot at room temperature for 30 min. Samples were centrifuged at 3000× g for 5 min, and the upper serum was gently transferred and preserved at −80 °C. All samples were frozen within 6 h of collection, and all analyses were finished within 72 h.

2.2. Biochemical Analyses of Serum

Serum samples were analyzed for 20 selected biochemical biomarkers (

Table 2. Information on serum biochemical detection reagents and methods used in this study.

Parameter	Abbreviation	Lot Number	Detection Method
Alanine Aminotransferase	ALT	140125005	IFCC Method
Aspartate Aminotransferase	AST	140225015	IFCC Method
De Ritis Ratio	AST/ALT	—	Calculation
Alkaline Phosphatase	ALP	140325002	AMP Buffer Method
γ-Glutamyl Transferase	γ-GT	140425013	IFCC Method
Total Protein	TP	148625004	Biuret Method
Albumin	ALB	148325012	Bromocresol Green Method
Globulin	Glo	—	Calculation
Albumin/Globulin	A/G	—	Calculation
Total Bilirubin	T-Bil-V	140625012	Vanadate Oxidation Method
Direct Bilirubin	D-Bil-V	140724026	Vanadate Oxidation Method
Indirect Bilirubin	IBIL-V	—	Calculation
Lipase	LIP	044825014	Enzyme Colorimetric Method
α-Amylase	α-AMY	142824008	Continuous Monitoring Method
Creatine Kinase	CK	140925018	IFCC Method
Triglyceride	TG	141725007	Oxidase Method
Total Cholesterol	TC	141625004	Oxidase Method
Low-Density Lipoprotein Cholesterol	LDL-C	142025007	Direct Method
High-Density Lipoprotein Cholesterol	HDL-C	142125008	Direct Method
Glucose	Glu-G	141525010	Oxidase Method

Note: AST/ALT, A/G, Glo, and IBIL-V are calculated values from detected indicators; no separate reagents are required. All reagents are matched with Mindray BS series biochemical analyzers, and the lot numbers can be replaced with actual batches according to experimental records. All detection methods are standard methods specified in the official registration certificates of Mindray reagents, complying with clinical laboratory quality requirements.

). The analyses were conducted using a fully automatic biochemical analyzer (BS-240Vet, Mindray Bio-Medical Electronics Co., Shenzhen, China). All assays were performed according to the manufacturer’s instructions using commercial reagent kits produced by Mindray (Table 2). Quality control and calibration were ensured by strictly following the standard operating procedures of both the analyzer and reagent kits.

2.3. Statistics

The specific growth rate (SGR), feed conversion ratio (FCR), and hepatosomatic index (HSI) were calculated as follows:

SGR (%) = 100 × ln(final weight/initial weight)/days,

(1)

FCR = dry diet fed/wet weight gain

(2)

HSI = 100 × liver weight/body weight

(3)

All statistical analyses and visualizations were performed using R software (version 4.5.2) [40]. Raw data were imported and restructured into wide format for correlation analysis. Spearman’s rank correlation was used to analyze the relationships among body weight, liver weight, HSI, and all serum biochemical indicators, and visualized as a lower-correlation matrix plot [41].

Distribution characteristics, including the mean, median, skewness, kurtosis, and Shapiro–Wilk normality test, were calculated for each biochemical parameter. As most parameters exhibited non-normal, right-skewed distributions, conventional parametric or percentile-based methods were deemed inappropriate. To improve reliability and reduce bias from extreme values, a smoothed bootstrap resampling approach was applied prior to reference interval establishment. Bootstrap resampling was performed with 500 iterations to stabilize the distribution shape and reduce random sampling variation, and the bandwidth for smoothing was determined using Silverman’s rule of thumb [42,43].

h = 1.06 × min (SD, IQR/1.34) × n^−1/5

(4)

where:

h = The optimal bandwidth for kernel density smoothing;

SD = Standard deviation of the original dataset;

IQR = Interquartile range, calculated as the difference between the 75th percentile and the 25th percentile of the dataset;

n = Total sample size of the corresponding biochemical parameter.

Kernel density estimation (KDE) was subsequently performed to identify the main peak of each smoothed distribution. The main peak was defined as the continuous central region where the estimated density exceeded 5% of the maximum density value (peak threshold = 0.05), which represents the physiological baseline of dominant healthy individuals. All data points outside the main peak were regarded as long-tailed outliers arising from non-pathological individual differences and were directly excluded to avoid distorting reference interval results. Based on the screened data within the main peak, 95% reference intervals were calculated as the 5th to 95th percentiles. Histograms of the original and smoothed datasets were generated using the ggplot2 package (ver. 4.0.1) [44]. Source codes and raw data file are provided in Supplementary Files.

3. Results

3.1. Growth Performance

After 90 days, the survival was 97.3 ± 0.9%, with an average weight of 96.3 ± 18.7 g and HSI of 2.7 ± 0.5. The SGR and FCR were 3.9% and 1.12, respectively. No external injuries, parasites, or abnormal behaviors were found in the sampled fish. Post-mortem anatomical inspection showed that all visceral organs were intact and free from lesions, nodules, or obvious necrosis. The overall health status of the fish met the requirements for subsequent serum biochemical analysis.

3.2. Correlation Analysis

Body weight showed a strong positive correlation with liver weight but was independent of HSI (Figure 1). Body metrics were not strongly correlated with the selected serum biochemical parameters in general. Significant positive correlations were found between AST and ALT, among TP, ALB, and Glo, and between IBIL-V and T-Bil-V. Moderate positive correlations were found between CK and apoptosis biomarkers (AST and ALT) and between bilirubin (T-Bil-V and IBIL-V) and protein synthesis (TP, ALB, and Glo) biomarkers. A moderate negative correlation was found between pancreatic biomarkers (LIP and α-AMY) and AST/ALT, TP, Glo, and D-Bil-V. Slight positive correlations were found between TG and body metrics (body weight, liver weight, and HSI) and between TC and protein synthesis (TP, ALB, and Glo) and bilirubin biomarkers (T-Bil-V and D-Bil-V).

3.3. Distribution Characteristics

Most serum biochemical parameters exhibited right-skewed, non-normal distributions except for TP (

Table 3. Descriptive statistics, distribution characteristics, normality tests, and suggested reference intervals of serum biochemical parameters in hybrid snakehead.

Para	Mean	Median	95% CI	Skewness	Kurtosis	Normal?	Smoothed 95% CI	Reference Interval
ALT (U/L)	11.0	7.3	0.4–52.7	4.8	30.7	N	0–51.3	0–14
AST (U/L)	72.7	54.0	27.8–225.3	3.5	16.6	N	18.3–220.4	24–71
AST/ALT	12.3	7.8	1.4–46.5	5.7	43.7	N	0–46.1	2–12
ALP (U/L)	19.1	16.9	7.3–52.4	1.9	4.0	N	4–55.3	6–25
γ-GT (U/L)	5.4	4.9	3–10.1	1.9	5.7	N	3–10.5	4–6
TP (g/L)	23.3	22.5	9–39.5	0.2	−0.4	Y	7.5–40.6	10–36
ALB (g/L)	15.4	15.6	10.2–19.3	−0.5	0.4	N	9.9–19.1	12–18
Glo (g/L)	7.9	6.4	−5.3–21.4	0.1	−0.8	N	0–21.6	0–18
A/G	1.4	1.8	−5.2–11.5	−8.8	95.6	N	0–11.9	0–4
T-Bil-V (μmol/L)	7.1	4.3	−2.3–20.2	1.1	−0.1	N	0–20.1	3–5
D-Bil-V (μmol/L)	−0.2	−0.2	−1–0.4	−6.6	50.5	N	0–0.5	0–0.3
IBIL-V (μmol/L)	7.6	4.7	1.3–20.3	1.1	−0.3	N	1–20.8	2–6
LIP (U/L)	26.0	12.9	−1.7–175.7	3.3	11.7	N	0–238.9	0–30
α-AMY (U/L)	336.1	161.5	12.2–1532.6	4.9	27.7	N	0–1833.9	0–367
CK (U/L)	1396.2	824.4	246.6–6099.9	3.1	10.9	N	0–6406.3	0–1682
TG (mmol/L)	3.1	2.3	0.7–9.3	1.6	2.6	N	0.2–9.1	0–4
TC (mmol/L)	4.0	3.9	2.8–5.3	0.6	1.8	N	2.8–5.3	3–5
LDL-C (mmol/L)	1.3	0.3	−0.1–9	4.5	24.7	N	0–9.1	0–1
HDL-C (mmol/L)	2.3	2.3	1.5–3.1	−0.6	3.7	N	1.4–3.2	2–3
Glu-G (mmol/L)	6.1	3.5	1.6–19.1	1.3	1.2	N	0–19.1	0–12

Note: CI = confidence interval; Y = normally distributed (p > 0.05); N = non-normally distributed (p < 0.05) by the Shapiro–Wilk test. Suggested reference intervals were established using kernel density estimation of the main peak after smoothed bootstrap.

; Figure 2). Shapiro–Wilk tests confirmed that only TP conformed to a normal distribution (p > 0.05), while all other 19 indicators showed significant deviations from normality (p < 0.05). Parameters including ALP, γ-GT, Glo, T-Bil-V, IBIL-V, TG, and Glu-G presented apparent bimodal distribution. High skewness and kurtosis values were observed in ALT, AST, CK, α-AMY, and LDL-C, indicating long-tailed distributions. In this case, the quantiles of 2.5% and 97.5% could fall in extreme values.

For this reason, we conducted smooth bootstrapping on all parameters and expanded the number of samples from 160 to 500. The adjusted data distribution is shown in Figure 3, and the smoothed 95% confidence intervals are shown in Table 3.

3.4. Smoothed Distribution, Main Peak Extraction, and Established Reference Intervals

After smoothed bootstrap augmentation, the distribution of each indicator became more stable and continuous (Figure 3). Kernel density estimation successfully identified the main peak for each parameter, eliminating interference from extreme tail values. The adjusted main peak distributions were more concentrated and representative of the majority.

Based on the main peak data, species-specific reference intervals were established for 20 serum biochemical indicators by calculating the 95% confidence intervals (Table 3). These intervals were narrower than raw percentiles and could cover the major group of fish. In addition, we listed the indication of either elevation or decrease (if the left boundary is above 0) in these parameters and the potential initiating factors.

4. Discussion

Serum biochemistry represents a sensitive and efficient approach for early health evaluation in human and veterinary science for diagnosis. It has also been widely used for assessing liver health status in aquaculture research, but its on-farm application is severely limited by the lack of standardized detection procedures and species-specific reference intervals. The present study attempted to surmount these obstacles by determining the distribution characteristics and constructing robust reference intervals under a relatively controlled mesocosm with stable rearing conditions and no history of disease or medication. By adopting standardized detection protocols and advanced statistical methods, this study provides reliable baseline data for health assessment in hybrid snakehead aquaculture.

4.1. Data Distribution of Serum Biochemical Parameters

In the present study, almost all of the selected serum biochemical parameters, except for TP, exhibited non-normal distributions with obvious right skewness and high kurtosis, which deviated significantly from the Gaussian distribution. This observation is consistent with previous reports in fish and humans [31,32], suggesting that non-normality is a common and inherent feature of fish physiological data. It also unveiled a fact that even under controlled rearing conditions and strict health screening, a small proportion of individuals displayed marginally elevated or reduced values that fell outside the central core of the distribution. Such skewed distributions or ‘outliers’ may not be from disease, but from individual variation, moderate environmental stress, transient metabolic fluctuations, feeding rhythms, and sexual maturation [45,46].

To address this problem, previous studies reporting reference intervals have followed the ASVCP Quality and Laboratory Standards Committee (QALS) Guidelines [31,35,36,47]. In brief, reference intervals are established using either the non-parametric percentile method (2.5th–97.5th percentiles of all samples) or data transformation to achieve normality. Considering the distinct right-skewed characteristics of our dataset, we adopted a combined approach of smoothed bootstrap resampling, kernel density estimation (KDE), and main peak extraction to construct reference intervals. Unlike the conventional full-sample percentile method, this strategy first stabilizes the overall data distribution via bootstrap smoothing, then identifies the dominant population using KDE and retains data within the main distribution peak. This processing approach aligns well with the skewed distribution of serum biochemical data in hybrid snakehead. Long-tailed values attributed to non-pathological individual variation are excluded, such that the derived physiological baselines more accurately represent the characteristics of the dominant healthy population.

The pervasive non-normality observed in this study carries important statistical implications. Conventional parametric statistical methods, including the t-test, one-way ANOVA, and Pearson correlation, rely on the assumption of normality and homogeneous variance; their direct application to skewed data may lead to inflated Type I error, reduced statistical power, and contradictory conclusions [48]. In future comparative studies on biochemical data, nonparametric methods such as the Wilcoxon test, Kruskal–Wallis test, or Spearman rank correlation should be strongly recommended [49].

From another perspective, the high degree of skewness and variability also highlights the requirement for adequately large sample sizes in fish physiological studies, preferably above 100 [50]. Larger cohorts improve the stability of density estimation, reduce the influence of random outliers, and enhance the reproducibility of reference intervals.

4.2. Interpretation of the Parameters

Interpretation of the serum or plasma biochemical data is often carried out by comparing the treated group with the control group, or between two sampling points. The serum biochemical parameters measured in this study can help determine the extent of liver injury, and the elevation pattern can provide information on pathogenesis (see

Table 4. Clinical indications and main inducing factors of serum biochemical parameters in hybrid snakehead.

Para	Indication	Inducing Factor
ALT	Elevated: Hepatocyte necrosis, inflammation, or toxicosis.	Fish meal replacement [21,51,52,53,54,55,56,57,58,59,60,61,62,63,64], fish oil replacement [65,66,67,68], herbicides [69,70], pharmaceuticals [71,72,73], microplastics [71,74], heavy metals [75,76,77,78,79,80,81], aromatic hydrocarbons [82,83,84], pesticides [85,86,87,88,89,90,91,92], acidifier [93], microbes [93,94,95,96], heat stress [97], plant extracts [70,98,99,100,101,102], hypoxia [103], algae [90], malnutrition [101], hemolysis [104]
AST	Elevated: Liver injury, heart injury, skeletal muscle, or general stress response.	Fish meal replacement [21,51,52,53,54,55,56,57,59,60,61,62,63,64,105,106,107,108], fish oil replacement [65,66,67,68,109], herbicides [69,70], pharmaceuticals [71,72,73], microplastics [71,74], heavy metals [75,76,77,78,79,80,81], pesticides [85,86,88,89,90,91,92], acidifier [93], microbes [93,94,95], heat stress [97], plant extracts [70,98,99,100,101,102,110], hypoxia [103], aromatic hydrocarbons [83,84], sampling methods [111], algae [90], malnutrition [101], light intensity [112], season [113], hemolysis [104]
AST/ALT	Elevated ratio: Severe liver damage, muscle, or cardiac injury. Decreased ratio: Mild hepatocellular damage.	Fish meal replacement [106], parasite [114], heavy metals [115], water quality [116].
ALP	Elevated: Biliary stasis, liver disorder, bone metabolism abnormality. Decreased: Malnutrition.	Fish meal replacement [21,51,53,61,64,106,109,117,118,119], fish oil replacement [66,67,68], pharmaceuticals [72], herbicides [120], plant extracts [98,101,102,121], heavy metals [76,77,78], fungicides [122], aromatic hydrocarbons [83,123], pesticides [89,90,92,124], algae [90], malnutrition, light intensity [112], season [113], hemolysis [104], microplastics [74]
γ-GT	Elevated: Biliary epithelium injury, intrahepatic cholestasis, toxin-induced liver damage.	Fish meal replacement [54], aromatic hydrocarbons [123], plant extracts [110]
TP	Elevated: Dehydration, chronic inflammation, immune activation. Decreased: Liver dysfunction, malnutrition, protein loss.	Fish meal replacement [51,52,53,54,55,56,59,63,64,106,108,117,125,126], fish oil replacement [66,68,127], herbicides [69,70,120], plant extracts [70,98,102,121,128,129], heavy metals [75,76,79,130,131,132,133], pesticides [85,86,89,90,134,135], pharmaceuticals [72], acidifier [93], microbes [94,95,96,136], season [137], fungicides [122], sampling methods [111], algae [90], hemolysis [104], microplastics [74]
ALB	Elevated: Severe dehydration. Decreased: Liver synthetic failure, malnutrition, protein-losing syndrome.	Fish meal replacement [51,52,53,54,55,59,63,106,108,125], fish oil replacement [66,68], herbicides [69,120], plant extracts [98,128,129,138], pharmaceuticals [71], microplastics [71,74], heavy metals [75,76,130], microbes [94,95,96,136], sampling methods [111], pesticides [90], algae [90], light intensity [112], hemolysis [104]
Glo	Elevated: Immune activation, chronic infection, inflammation. Decreased: Immunosuppression, severe liver disease.	Fish meal replacement [51,52,53,54,55,59,63,64,108,117], fish oil replacement [68,127], plant extracts [98,128], heavy metals [75,130], pesticides [85,90], herbicides [120], microbes [94,95,136], fungicides [122], sampling methods [111], light intensity [112]
A/G	Decreased ratio: Chronic inflammation, liver injury, immune hyperactivation. Elevated ratio: Dehydration.	Fish meal replacement [55], plant extracts [128]
T-Bil-V	Elevated: Hemolysis, hepatocellular injury, or biliary obstruction (jaundice).	Fish meal replacement [53,64], pesticides [89], parasites [139]
D-Bil-V	Elevated: Post-hepatic cholestasis, biliary obstruction.	Pesticides [89], parasites [139]
IBIL-V	Elevated: Pre-hepatic hemolysis or hepatocellular jaundice.	Parasites [139]
LIP	Elevated: Pancreatic injury, intestinal inflammation. Decreased: Pancreatic insufficiency.	Fish meal replacement [140], light intensity [112]
α-AMY	Elevated: Pancreatic damage, intestinal stress, acute stress response.	Parasites [96], light intensity [112]
CK	Elevated: Skeletal muscle injury, capture stress, hypoxia, exercise exhaustion.	Hypoxia [141], sampling methods [111,142], heavy metals [143]
TG	Elevated: Hyperlipidemia, fatty liver, metabolic disorder. Decreased: Malnutrition.	Fish meal replacement [21,52,54,55,56,106,118,125,144,145], fish oil replacement [67,68,109,127,146], heat stress [97], hypoxia [141,147], herbicides [120], heavy metals [76,80,130,133], plant extracts [100,101,110,121,129,138,148,149], microbes [95], season [137], malnutrition [101], light intensity [112], stock density [22], microplastics [74]
TC	Elevated: Hyperlipidemia, fatty liver, metabolic syndrome. Decreased: Liver dysfunction, malnutrition.	Fish meal replacement [21,52,53,54,55,58,59,64,106,107,118,144], fish oil replacement [67,68,146,150,151,152], water quality [30], hypoxia [103,141,147], herbicides [120], heavy metals [76,77,80,130], plant extracts [98,101,110,121,129,148,149], microbes [95], pharmaceuticals [73], season [113,137], pesticides [90], algae [90], malnutrition [101], microplastics [74]
LDL-C	Elevated: Lipid metabolism disorder, liver dysfunction.	Fish meal replacement [52,106], fish oil replacement [67,109,146,150,151], plant extracts [98,121,148,149]
HDL-C	Decreased: Metabolic disorder, liver injury, chronic stress.	Fish meal replacement [52,153], fish oil replacement [67,146,150], plant extracts [98,100,148,149]
Glu-G	Elevated: Acute stress, hyperglycemia, glycogenolysis. Decreased: Starvation, liver glycogen depletion.	Fish meal replacement [53,54,56,59,64,106,145,154], fish oil replacement [66,67,68,127,150,151], herbicide [69,70,120], pharmaceutical [71,72,73], microplastic [71], water quality [30], hypoxia [103,141,147], heavy metals [76,130,131,132,133], plant extracts [70,101,102,110,149,155,156], microbes [95,96], pesticides [87,89,91,134,135,157], sampling methods [142], malnutrition [101], light intensity [112], season [113], stock density [22]

Note: References in this table were retrieved from the Web of Science Core Collection via advanced search. For each serum biochemical parameter, combined searches with fish/teleost/aquaculture terms were performed. Only studies reporting significant parameter changes under relevant inducing factors were included.

).

4.2.1. Liver Injury Markers

Liver function and hepatocellular integrity were assessed using a panel of indicators, including ALT, AST, AST/ALT, ALP, and γ-GT. ALT is only present in the hepatocyte cytoplasm, whereas AST is located in both the hepatocyte cytoplasm and mitochondria. The activity of AST in the liver is 2.5 times higher than ALT, and therefore increases faster upon liver injury. Human medical studies revealed that AST is also abundantly present in heart tissue [158]. When isolated elevation of AST occurs, heart damage may be considered, and CK can be measured [159]. These two parameters are the primary liver injury indicators and are extensively used in health assessment (Table 4). The AST/ALT ratio (De Ritis ratio) further improves diagnostic specificity, with elevated ratios commonly associated with severe hepatic injury, while reduced ratios may indicate milder hepatocellular damage [84,114]. This parameter is rarely used but also studied in fish. A decrease in the De Ritis ratio was observed in northern pike Esox lucius and European perch Perca fluviatilis infected with the cestode Triaenophorus nodulosus, indicating chronic liver damage [114]. Significant elevation was recorded in spotted snakehead Channa punctatus exposed to thermal power plant effluent contaminated with heavy metals, reflecting acute liver injury [115].

ALP is a multifunctional enzyme and exists in many organisms, from bacteria to mammals. Serum ALP activity is initially used to indicate damage to hepatocytes and biliary epithelial membranes in human diagnosis due to its high sensitivity (>4 times normal values). However, later studies on dogs found that the specificity of serum ALP activity for hepatobiliary disease could be as low as 51% [160]. γ-GT is primarily present in cells with secretory or absorptive functions, and therefore its activity in serum is a more specific indicator of biliary injury [161,162]. Elevated levels are medically associated with cholestasis, biliary obstruction, or metabolic disturbance [163,164]. Combined tests of both ALP and γ-GT are highly recommended as their specificity for cholestasis could be elevated [160]. Nutritional alterations, chemical exposure, and malnutrition could induce the elevation of ALP and γ-GT (Table 4).

4.2.2. Nutritional, Immune, and Metabolic Status

Serum total protein is generally the sum of serum albumin and globulin. From the experience of human diagnosis, a decrease in serum TP is detected in pancreatitis, enterocolitis, cirrhosis, hepatitis, etc., whereas an increase in serum TP leads to infectious diseases, arthritis, nephritis, and others [165]. Albumin is capable of transporting various nutrients, such as fatty acids, bilirubin, hormones, metals, and drugs due to its negative charge and large surface area [165]. Serum albumin levels may decrease in cases of fish meal replacement, fish oil replacement, or toxin exposure [52,75,129], indicating reduced function of protein synthesis. Globulin is closely associated with immune activation and inflammatory response, and the A/G ratio provides additional information on immune balance and chronic stress [165]. TP and ALB were found to be susceptible to nutritional alterations, chemical compounds, and environmental changes [58,109,113]. Glo was found to be more sensitive to hazardous chemicals and microbes [94,95,129,136].

Glucose is a key indicator of an acute stress response and short-term energy mobilization; elevated glucose is a classic physiological hallmark of handling, hypoxia, or acute disturbance in fish [166]. For this reason, serum glucose has been widely used to evaluate the stress level of fish exposed to suboptimal environments, heavy metals, stock density, and toxic compounds [22,103,141]. Lipid parameters, including TG, TC, HDL-C, and LDL-C, reflect lipid transport, energy storage, and metabolic homeostasis; abnormal levels are associated with fatty liver, dietary imbalance, or metabolic disorders [167]. Total cholesterol and triglycerides are widely used to assess the adverse effects of fish meal or fish oil replacement and the potential of plant extracts [52,117]. LDL-C and HDL-C are important indicators of lipoprotein synthesis, transport, and utilization and were found to only respond to nutritional alterations [52,67,106,146,150].

4.2.3. Bilirubin Metabolism and Biliary Function

Bilirubin metabolism indices, including T-Bil-V, D-Bil-V, and IBIL-V, reflect erythrocyte turnover, hepatic uptake and conjugation, and biliary tract patency [168,169]. However, these parameters are rarely investigated in fish. Few studies found them responsive to specific pesticides and parasites [89,139]. From the perspective of human studies, elevated total bilirubin may arise from pre-hepatic hemolysis, hepatocellular dysfunction, or post-hepatic biliary obstruction [170]. Direct bilirubin is particularly indicative of cholestasis and biliary blockage, which could be induced by antinutritional factors, antibiotics, or contaminants [171], whereas indirect bilirubin elevation is associated with hemolytic processes. The low and stable bilirubin levels observed in this study confirm normal erythrocyte integrity and unobstructed biliary function in healthy hybrid snakehead.

4.2.4. Pancreatic Exocrine Function

Pancreatic exocrine function was assessed using LIP and α-AMY. These enzymes are synthesized in the exocrine pancreatic islet in the hepatopancreas gland in fish and excreted to facilitate lipid and carbohydrate digestion [172]. Abnormal elevation may indicate pancreatic damage, intestinal inflammation, or acute stress, while reduced activity suggests pancreatic insufficiency or malnutrition. In a few cases, exogenous protein, light intensity, and parasites could induce the elevation of LIP and α-AMY [96,112,140], possibly via different pathways.

4.2.5. General Tissue Damage

Creatine kinase (CK) is a highly sensitive marker of muscle damage and also serves an essential role in regulating cellular energy metabolism [173,174]. In aquaculture, CK is strongly induced by handling stress, netting, hypoxia, exhaustive swimming, and physical injury [175]. Therefore, CK serves as a reliable indicator of physical disturbance and tissue integrity in hybrid snakehead.

4.3. Intrinsic Relation Between Serum Biochemical Parameters

Serum biochemical parameters do not function independently but are tightly interrelated within a coordinated physiological network. In our study, Spearman correlation analysis revealed meaningful interrelationships among multiple biochemical parameters. Significant correlations were detected between bilirubin metabolism and protein synthesis [176]. Associations between cholesterol and protein synthesis suggest close links between lipid metabolism and hepatic synthetic function [177]. Strong correlations were also observed between CK and liver injury indicators (AST and ALT), reflecting concurrent cell damage at the organ level and cellular metabolic disorder under systemic disturbance [175]. Furthermore, pancreatic exocrine markers were correlated with protein and bilirubin metabolism, demonstrating functional coupling between digestive function, hepatic processing, and nutrient utilization [172].

Our findings are in line with previous studies. For instance, acute exposure to pharmaceutical waste triggers a cascade of responses, including elevated AST, ALT, ALP, glucose, and protein levels, indicating simultaneous occurrence of hepatocyte apoptosis, bile duct damage, and energetic disturbance [72]. Partial replacement of fish meal or fish oil altered not only protein- or lipid-related parameters like TP, ALB, TC, or TG, but sometimes elevated AST and ALT, indicating the potential hepatotoxicity of alternative feed ingredients [59,109]. These interrelationships highlight the integrated nature of physiological regulation and support the use of a comprehensive biochemical panel for health monitoring in hybrid snakehead. Omics studies are warranted to further explain the underlying mechanisms.

From a diagnostic perspective, this intrinsic connection implies that accurate health assessment and disease diagnosis require integrated interpretation of a panel of biomarkers rather than reliance on a single parameter. In clinical practice, over-interpretation of individual values may lead to misdiagnosis; instead, a combination of liver enzymes, protein markers, lipid metabolites, pancreatic enzymes, and stress indices provides a more holistic and credible evaluation [178].

4.4. Limitations of the Present Study

Despite the robust statistical methods and standardized procedures employed, this study has several limitations that should be acknowledged and addressed in future investigations.

First, the representativeness and generalizability of the established reference intervals are restricted by the sampling framework. All fish were maintained in a relatively controlled mesocosm under stable environmental conditions, dietary regime, stocking density, and management practices. Geographical location, water quality, temperature, photoperiod, feed composition, culture system, and health history can all influence serum biochemical profiles [21,32,179]. In addition, the present study did not explicitly account for variables such as sexual maturity, starvation status, postprandial interval, acute stress prior to sampling, age, or body size range, all of which may introduce physiological variation [32,180,181,182]. Furthermore, samples from clinically diseased fish were not included, limiting the direct diagnostic contrast between healthy and pathological states. Future studies should expand the sampling scope to include multiple geographical regions, cultural models, nutritional backgrounds, environmental gradients, and physiological states to establish more universal and robust reference intervals.

Second, methodological variability remains an important source of heterogeneity in fish serum biochemistry. Pre-analytical factors, including anesthetic type and concentration, anticoagulant, storage temperature, freeze–thaw cycles, and hemolysis, can significantly alter the measured values [25,27,104,111]. Analytical variability may also arise from different reagent kits, calibration standards, detection principles, and automated biochemical analyzers. Generally, we found that values tested by reagent kits were lower than those obtained by automated biochemical analyzers. For instance, the serum AST and ALT of Micropterus salmoides were around 25 U/L and 17 U/L in the control when using a commercial reagent kit [60], but the AST and ALT were found at around 84 U/L and 2 U/L when using a Roche automated biochemical analyzer [183]. In the present study, strict standardization was applied to minimize pre-analytical and analytical variation; however, reference intervals are inherently method-specific and may not be directly transferable to laboratories using different protocols or platforms. Standardization of sampling, processing, and detection across institutions is strongly encouraged to improve comparability and facilitate the wider application of serum biochemistry in aquaculture health management.

5. Conclusions

In conclusion, the present study established the first comprehensive set of serum biochemical reference intervals for juvenile hybrid snakehead (Channa argus × C. maculata) under a mesocosm experiment (27 °C for 90 days) using smoothed bootstrap resampling and kernel density estimation. Most parameters exhibited non-normal, right-skewed distributions; comparative analyses of these indicators should rely on non-parametric statistical methods rather than normality-dependent approaches such as ANOVA or t-tests. The established reference intervals provide robust and method-specific benchmarks for evaluating liver function, nutritional status, lipid metabolism, tissue damage, and stress responses under a relatively controlled environment. These findings fill a critical knowledge gap for this economically important aquaculture species and support comparative investigations into metabolic disorders, toxin and contaminant toxicity, feed additive evaluation, infection, and disease prevention. This study also highlights the importance of standardized detection and statistical optimization for the wider application of serum biochemistry in fish aquaculture.