Development and Validation of the Multi-Residue Method for Identification and Quantitation of Six Macrolide Antiparasitic Drugs

Abstract

Objective: This study aimed to develop a robust multi-residue analytical method for the precise identification and quantification of six macrolide antiparasitic agents commonly used in animal husbandry feeds. Method: Feed samples were extracted using a water-saturated acetonitrile solution. The resulting crude extracts were then treated with n-hexane and further purified by HLB solid-phase extraction columns to obtain the test solutions. These prepared samples were analyzed by high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS). The method was validated across six different feed matrices, including pig premix, concentrate, and complete feed, as well as chicken premix, concentrate, and compound feed. The method exhibited average recoveries ranging from 80.07% to 98.80%. The intra-day coefficients of variation (CV) for the first three feed types ranged from 1.98% to 12.84%, while for the latter three, the CVs ranged from 2.43% to 13.69%. Results: The method’s precision led to the quantification limit of avermectin, doramectin, acetyl avermectin, and ivermectin being 25 μg/kg, whereas for moxifloxacin and milbemycin, the limit was 50 μg/kg. These thresholds meet the stringent requirements for trace drug analysis, supporting the method’s suitability for regulatory surveillance and monitoring of these specified antibiotics in animal feeds.

1. Introduction

Antiparasitic drugs are widely used in animal husbandry due to their critical role in maintaining livestock health and productivity. Among these, macrolide antiparasitic drugs, also known as anti-helminthic macrocyclic lactones (MLs), have become the most widely applied class because of their dual efficacy against both endoparasites and ectoparasites. These compounds are not only essential in veterinary medicine but are also used in human healthcare and agricultural pest management (Lespine et al., 2013) [1]. Macrolide antiparasitic agents belong to a broad family of hydrophobic, structurally related compounds. In veterinary medicine, they are primarily classified into two groups: avermectins (AVMs) and milbemycins (MBLs) (Lespine et al., 2013) [1].

Avermectins (AVMs) were first discovered in 1967 and are produced through the fermentation of Streptomyces species. Currently, the most commonly used avermectin derivatives include avermectin (AVM), ivermectin (IVM), doramectin (DOR), and eprinomectin (EPR, also known as acetyl avermectin) (Rúbies et al., 2015) [2]. These compounds have gained widespread application in animal husbandry and agricultural production due to their novel chemical structures, unique mode of action, high safety profile, exceptional efficacy, and low toxicity. Similarly, milbemycins represent another critical class of antiparasitic agents used in livestock production and cultivation. The primary members of this group include milbemycin oxime (MIL) and moxidectin (MOX). While both avermectins and milbemycins share some functional similarities, they exhibit distinct differences in their chemical structures and physical properties, which influence their specific applications and efficacy. The chemical structures of these six compounds are illustrated in Figure 1.

Macrolides, while effective against harmful insects and parasites, demonstrate mammalian toxicity (Merola & Eubig, 2012) [3]. According to World Health Organization (WHO) regulations, avermectins (AVMs) are classified as highly hazardous compounds with significant neurotoxic potential. Animals with elevated AVM concentrations may exhibit central nervous system (CNS) toxicity symptoms, including dyskinesia, respiratory depression, and tremors. The specific toxicological manifestations show slight interspecies variations (Bai & Ogbourne, 2016) [4].

The irrational use of veterinary drugs may lead to drug accumulation in livestock and residues in animal-derived food products, posing direct and indirect risks to human health. In the year 2020, China’s Ministry of Agriculture and Rural Affairs issued Announcement No. 194, which prohibited the direct use of antibiotics in animal feed. This regulation aims to reduce the negative impacts of antibiotic misuse and ensure the safety of animal-derived foods and public health. Macrolide drugs are widely used in animal husbandry and agriculture. However, repeated or improper application, including off-label use, may result in excessive drug levels in feed. Livestock can be exposed to macrolide antiparasitic agents through multiple pathways, such as residues in plant-based feed, feed additives, and veterinary drug administration. Such exposure can lead to bioaccumulation in animal tissues, ultimately entering the human food chain and posing health risks.

This study aimed to develop a robust and sensitive LC-MS/MS method for the simultaneous determination of six macrolide antiparasitic drugs (EPR, AVM, IVM, DOR, MOX, and MIL) in various feed matrices, including premix, concentrate, and complete feed, commonly used in aquaculture. The performance was comprehensively validated by evaluating key parameters such as specificity, linearity, recovery, precision, and sensitivity to ensure compliance with regulatory and analytical standards. This validated method provides a reliable approach for monitoring drug residues in feed samples, offering technical support for promoting safe aquaculture practices and protecting public health.

2. Materials and Methods

2.1. Chemicals and Drugs

All organic solvents and chemical reagents, including HPLC-grade methanol, acetonitrile, n-hexane, and formic acid were purchased from Guoyao Group Chemical Reagent Co., Ltd. (Shanghai, China). C₁₈ and Oasis HLB solid-phase extraction (SPE) cartridges (6 cc, 200 mg) were obtained from Waters Corporation (Milford, MA, USA). Syringe filters (0.22 μm nylon membrane) were acquired from Pall Corporation (Ann Arbor, MI, USA).

Standards of Acetyl Avermectin, Ivermectin, Moxidectin, Milbemycin, Doramectin, and Abamectin were purchased from Guoyao Group Chemical Reagent Co., Ltd. (Shanghai, China). Individual standard stock solutions (1 mg/mL) were prepared by dissolving 10 mg of each drug standard in 10 mL of acetonitrile. These stock solutions were stored at −18 °C and remained stable for up to three months. To prepare the mixed standard solution (0.1 mg/mL), 1 mL of each stock solution was accurately transferred into a 10 mL volumetric flask, diluted with acetonitrile to volume, and thoroughly mixed. This mixed solution was stored at 2–8 °C and remained stable for one month. A series of mixed standard working solutions with concentrations of 0.01 µg/mL, 0.02 µg/mL, 0.05 µg/mL, 0.1 µg/mL, 0.5 µg/mL, 1 µg/mL, and 2 µg/mL were prepared by diluting the 0.1 mg/mL mixed standard solution with 90% methanol to the appropriate volumes. These working solutions were stored at 2–8 °C and were stable for 1 week.

2.2. Sample Preparation

The spiked samples were prepared by accurately weighing 2.0 g of homogeneous blank feed into a 50 mL centrifuge tube and adding a precise volume of mixed standard working solution. After thorough vortex mixing, the samples were protected from light and allowed to equilibrate overnight (12–16 h). For extraction, 10 mL of water-saturated acetonitrile was added to each sample, followed by vigorous vortex mixing for 1 min. The samples were then subjected to extraction at 40 °C with orbital shaking at 300 rpm for 40 min, followed by centrifugation at 8000× g for 10 min at 4 °C. The supernatant was collected, and the extraction process was repeated to ensure complete recovery, with the combined extracts pooled for subsequent cleanup.

For a sample purification, 2.5 mL of the combined extract was transferred to a new tube and mixed with 1.0 mL of deionized water and 1.5 mL of n-hexane. The mixture was vortexed for 30 s and centrifuged at 8000× g for 10 min at 4 °C to separate the phases. The upper n-hexane layer containing lipid components was carefully discarded, and this degreasing procedure was repeated once more to ensure complete lipid removal. The remaining extract was then concentrated to approximately 1 mL under a gentle stream of nitrogen at 65 °C. Finally, the concentrate was reconstituted with 5.0 mL deionized water and thoroughly mixed by vortexing to obtain the final loading solution ready for analysis.

The HLB solid-phase extraction columns were preconditioned by sequential activation with 5 mL of acetonitrile followed by 5 mL of methanol. After loading all prepared sample solutions onto the activated columns, the columns were first washed with 5 mL of 18% (v/v) acetonitrile in water, followed by 5 mL of 18% (v/v) methanol in water to remove interfering matrix components. The target analytes were then quantitatively eluted using 6 mL of pure methanol at a controlled flow rate maintained below 1 mL/min to ensure optimal recovery. The collected eluate was carefully evaporated to complete dryness under a gentle stream of nitrogen gas at 65 °C. For the final preparation, the dried residue was precisely reconstituted in 1.0 mL of 90% (v/v) methanol in water using a volumetric approach, followed by vigorous vortex mixing for 1 min to ensure complete dissolution. Prior to the instrumental analysis, the solution was filtered through a 0.2 μm nylon syringe filter and transferred into a certified LC-MS (Waters Corporation, Milford, MA, USA) vial for analysis.

2.3. Instrument Conditions

The chromatographic separation was performed on an ACQUITY UPLC BEH C18 analytical column (50 mm × 2.1 mm internal diameter, 1.7 μm particle size; Waters Corporation, Milford, MA, USA) maintained at 40 °C. The analytes were detected and quantified using a Waters Xevo TQ-S micro triple quadrupole mass spectrometer coupled with an ACQUITY UPLC system (Waters Corporation, Milford, MA, USA). The mass spectrometer was operated in multiple reaction monitoring (MRM) mode with electrospray ionization (ESI) in positive ion mode.

The test solution was injected at a volume of 5.0 μL for analysis. Chromatographic separation was achieved using a binary mobile phase system consisting of 0.3% (v/v) formic acid in water (Solvent A) and HPLC-grade methanol (Solvent B). The mobile phase flow rate was maintained at 0.20 mL/min throughout the analysis, and the gradient elution program was applied as detailed in

Table 1. Gradient elution procedure for liquid chromatography separation.

Time (min)	Solvent A (%)	Solvent B (%)
0	15	85
0.5	15	85
2	3	97
4	3	97
5	15	85
6	15	85

.

The mass spectrometer was operated in positive electrospray ionization (ESI+) mode with the following optimized source parameters: capillary voltage 2.8 kV, cone voltage 70 V, source temperature 100 °C, desolvation temperature 320 °C, desolvation gas (N₂) flow 630 L/h, and collision gas (Ar) flow 33 L/h. Detection was performed in multiple reaction monitoring (MRM) mode with compound-specific transitions and collision energies as detailed in

Table 2. Mass spectrometer conditions for determination of the six drugs.

Analyte	Parent Ion (m/z)	Daughter Ion (m/z)	Taper Hole Voltage (V)	Collision Energy (eV)
Acetyl avermectin	936.69	352.13 a/490.27	70	60/60
Abamectin B1a	895.67	182.92 a/327.40	70	60/60
Doramectin	921.61	353.33 a/777.70	70	50/50
Milbemycin	556.60	129.11 a/169.33	40	40/40
Moxidectin	640.30	199.10 a/528.30	30	25/9
Ivermectin H2B1a	897.52	609.35/753.70 a	70	45/45

^a The quantitative ion.

.

2.4. Method Validation

2.4.1. Specificity

The method’s specificity was evaluated by analyzing six different blank feed matrices (pig premix, pig concentrate, pig complete feed, chicken premix, chicken concentrate, and chicken compound feed) to verify the absence of interfering peaks at the retention times corresponding to all target analytes.

2.4.2. Linearity

The method’s linearity was evaluated through matrix-matched calibration using seven concentration levels (0.01, 0.02, 0.05, 0.1, 0.5, 1.0, and 2.0 μg/mL) covering the expected analytical range for all six macrolides. Calibration standards were prepared in blank matrix extracts and analyzed in randomized order to avoid sequence effects. Quantitative analysis was based on the averaged peak areas (y-axis) plotted against nominal concentrations (x-axis, μg/L). A weighted (1/x) least squares regression model was applied to establish the calibration curve, with the resulting regression equation (y = ax + b) and coefficient of determination (R²) used to evaluate linearity. Each calibration level was analyzed in triplicate to verify reproducibility across the working range.

2.4.3. Accuracy and Precision

The method’s accuracy was validated through spike recovery experiments conducted across five feed matrices at three concentration levels each, reflecting expected residue ranges in different feed types: complete feed and vitamin premix (0.05, 0.5, 5.0 μg/g), concentrated feed and bovine supplement (0.1, 1.0, 10.0 μg/g), and premix feed (0.5, 5.0, 50.0 μg/g). For each concentration level, five replicate samples and corresponding blank controls were prepared following the national standard GB 5009.295-2023 [5] for food safety analysis method validation. Following the established sample preparation protocol, spiked samples were analyzed by LC-MS/MS. Measured concentrations were determined using matrix-matched calibration curves, with recovery rates calculated as follows: Recovery (%) = (measured concentration/spiked concentration) × 100%. The method’s precision was evaluated through both intra-day (n = 6 replicates within one analytical batch) and inter-day (n = 6 replicates across three consecutive days) repeatability studies at all concentration levels. The precision was expressed as relative standard deviation (RSD, %) of the recovered concentrations.

2.4.4. Limit of Detection (LOD)

The method’s sensitivity was characterized by determining the limit of detection (LOD) and limit of quantification (LOQ) through matrix-matched calibration. Ten independent blank feed samples (representing all validated matrix types) were processed following the complete sample preparation protocol to establish the baseline noise level. A matrix-matched standard working solution (10 μg/L) was prepared by spiking the blank matrix extract with appropriate dilutions of the mixed standard solution.

An instrumental analysis was performed on both the processed blanks and matrix-matched standards. The signal-to-noise ratios (S/N) for all six target analytes were calculated at their respective retention times using the following criteria: LOD: defined as the analyte concentration yielding a signal 3 times the standard deviation of blank responses (3× S/N); LOQ: established as the lowest concentration meeting method accuracy (80–120% recovery) and precision (RSD < 20%) criteria with a signal 10 times the blank noise level (10× S/N). The final reported LODs and LOQs were verified through an analysis of samples spiked at these threshold concentrations, with confirmation across three separate analytical batches.

Blank samples of compound feed, concentrated feed, premix, and bovine concentrate were spiked with appropriate amounts of standard solution to achieve target drug concentrations of 25 μg/kg and 50 μg/kg, respectively. For each concentration level, three to five replicates were prepared. The spiked samples were processed according to the established sample extraction and purification protocols and subsequently analyzed using UPLC-MS/MS.

3. Results and Discussion

3.1. Optimization of Mass Conditions

The identification and quantification were performed using multiple reaction monitoring (MRM). The data collection and optimization of mass spectrometry conditions in positive ion electrospray ionization (ESI+) mode were conducted to enhance the selectivity and sensitivity of the method (Inoue et al., 2009 [6]; Wang et al., 2011 [7]; Zhang et al., 2017 [8]). Initially, electrospray ionization (ESI) was employed to perform a full mass spectrum scan of each drug at a high concentration level (typically 1 mg/mL) in positive ion mode, with the ion source temperature optimized to 320 °C. Previous studies have shown that this temperature ensures the effective desolvation and ionization of the analytes (Holčapek et al., 2012 [9]), while scanning at a high concentration level guarantees sufficient signal intensity for accurate identification (Korfmacher, 2005) [10]. However, molecular ion peaks for the fragmented molecules of MOX and MBL were difficult to detect.

Subsequently, the ion source temperature was systematically optimized and adjusted to 100 °C (Taylor, 2005) [11], a condition that significantly enhanced the detection of molecular ion peaks for MOX and MBL. This lower temperature not only improves the stability of molecular ions but also increases the response values of their respective product ions in both product ion scanning and MRM (multiple reaction monitoring). Additionally, the other four target analytes exhibited improved ionization efficiency and signal response under these optimized conditions. These findings suggest that a lower ion source temperature reduces thermal degradation and enhances the overall ionization efficiency of thermally unstable compounds. Consequently, based on these results, the ion source temperature for the final method was set to 100 °C to ensure optimal sensitivity and reproducibility for the simultaneous quantification of all target analytes.

The product ions of the six target drugs were further analyzed through daughter ion scanning. Through the systematic optimization of the cone voltage and collision energy, the experimental conditions were fine-tuned to enhance the selectivity of the method. This optimization aimed to reduce interference from precursor ions by achieving a balance between minimizing the intensity of precursor ion peaks and maximizing the signal intensity of the target analytes (Van De Steene & Lambert, 2008) [12]. Precursor ions refer to the parent ions that undergo fragmentation in mass spectrometry analysis, generated by the electrospray ionization source (ESI). Their mass-to-charge ratio (m/z) is the molecular ion or product ion of the compound to be analyzed. They serve as the starting point for fragmentation and are used to determine the molecular weight and structural features of the target compound. Meanwhile, product ions are the fragment ions produced by the precursor ions through collision-induced dissociation or other fragmentation techniques. Their mass-to-charge ratios and intensities are used to infer the structural information of the precursor ions.

These precursor and product ions were subsequently used for multiple reaction monitoring (MRM) analysis. The daughter scans of the determined precursor ions are shown in Figure 2: for acetyl avermectin, m/z936.69 produced m/z352.13, m/z490.27, m/z751.51, and m/z368; for abamectin B1a, m/z895.67 produced m/z182.86 and m/z327.12; for Moxidectin, m/z198.53 and m/z528.50 was produced by m/z640.30; for doramectin, m/z921.61 produced m/z353.15 and m/z777.47; for ivermectin H2B1a, m/z897.52 produced m/z609.35 and m/z753.30. Milbemycin continued to be optimized to produce m/z129.11 and m/z169.33. And the numbers after the decimal point of the m/z value were adjusted based on the instrument detection statutes in the actual sample determination. In subsequent analyses of actual samples, the method demonstrated excellent selectivity and high sensitivity, enabling reliable qualitative and quantitative analyses.

3.2. Optimization of Sample Preparation

3.2.1. Optimization of Sample Extraction Solution

Sample preparation plays a critical role in determining the reliability of test results. Therefore, the initial step involved screening extraction solvents capable of effectively extracting the six macrolides from feed matrices. Given the high lipophilicity of macrolide antiparasitic drugs, two extraction methods using acetonitrile (Ozdemir & Kahraman, 2016) [13] and methanol were designed, based on relevant studies (Zhan et al., 2013 [14]; Rübensam et al., 2013 [15]; Wang et al., 2012 [16]; de Oliveira Ferreira et al., 2016) [17].

In the preliminary experiments, the premix was selected as a representative matrix to optimize the sample processing procedure. This choice was based on its similarity to biological matrices commonly encountered in analytical studies, such as plasma, serum, and tissue homogenates (Matuszewski et al., 2003) [18]. The premix matrix, composed of a balanced mixture of proteins, lipids, and other endogenous compounds, effectively simulates the complexity of real-world samples. This approach facilitates a comprehensive evaluation of potential matrix effects during method development. Two grams of blank feed and premix feed spiked with the target analytes (at a concentration of 1000 μg/kg) was weighed into 50 mL centrifuge tubes. Each sample was extracted twice with 10 mL of water-saturated acetonitrile and 10 mL of water-saturated methanol, respectively, to compare the extraction efficiency of the two solvents. The results indicated that both extraction solvents exhibited similar efficiency in extracting the six macrolides (the extraction rates of six drugs using two solvents are shown in Figure 3), meeting the detection requirements for all target analytes. However, the recovery rates for all drugs except AVM were slightly higher when using acetonitrile. This improvement is likely attributed to the reduced interference from co-extracted impurities with acetonitrile. Furthermore, research demonstrates that acetonitrile is selected over methanol as the extraction solvent primarily due to its enhanced ability to precipitate proteins in feed matrices (Boti et al., 2024) [19]. Additionally, a 50% acetonitrile solution is chosen instead of pure acetonitrile as the extraction solvent, as the extraction efficiency of pure acetonitrile in fish feed has been found to be comparatively lower.

3.2.2. Optimization of Sample Extraction Temperature

The average extraction efficiency for IVM, MBL, and AVM using acetonitrile was approximately 80%, indicating room for improvement. Studies suggest that increasing the extraction temperature can enhance the extraction efficiency of these drugs when using the same solvent (Tao et al., 2012) [20]. Based on preliminary findings, the extraction temperature was systematically increased from room temperature (25 °C) to 40 °C to assess its impact on extraction efficiency (Ramos et al., 2002) [21]. As illustrated in Figure 4, the extraction efficiency for five of the six target drugs improved significantly at 40 °C compared to room temperature. This enhancement can be attributed to the increased solubility of analytes and accelerated mass transfer kinetics at elevated temperatures. However, the extraction efficiency of DOR remained largely unaffected by temperature changes, likely due to its unique physicochemical properties, such as its higher thermal stability and lower solubility in the extraction solvent. These results highlight that temperature optimization is a critical factor for maximizing the extraction efficiency of thermally stable compounds. In contrast, alternative strategies may be required for compounds like DOR that exhibit temperature-independent behavior. Consequently, the extraction temperature for the final method was set to 40 °C to ensure optimal recovery for the majority of target analytes.

3.2.3. Optimization of Purification Conditions of Solid-Phase Purification Column

Both the C18 purification column and the HLB (Hydrophilic–Lipophilic Balance) are classified as non-polar adsorption columns, designed to retain non-polar target analytes in polar aqueous solvents through non-polar interactions (Hennion, 1999) [22]. The HLB purification column was developed as an advanced alternative to the traditional silica gel C18 column. In previous studies, the purification of macrolides predominantly utilized C18 columns (Senta et al., 2017) [23]. However, the HLB column, which incorporates a styrene-divinylbenzene copolymer, offers several advantages over the C18 column, including a higher tolerance to loading solutions, simpler activation procedures, and greater resistance to drying (Anastassiades et al., 2003) [24]. While both columns serve similar functions, the HLB column is more user-friendly and efficient. Therefore, the HLB purification column was selected as the solid-phase extraction (SPE) column for this method.

In previous studies, acetonitrile, methanol, and ethyl acetate were evaluated as elution solvents for the HLB purification column (Subirats et al., 2005) [25]. While ethyl acetate demonstrated the strongest elution capability and yielded the highest recovery rates, it also introduced significant matrix interference and extraneous peaks during sample analysis (González-Ruiz et al., 2011) [26]. In contrast, acetonitrile and methanol exhibited comparable elution efficiency with minimal differences in performance. Based on these findings, methanol was selected as the elution solvent for this method. To ensure the complete elution of the target analytes, 6 mL of methanol was used for the elution process.

After determining the drug elution solvent for the HLB purification column, the elution conditions were further optimized. The HLB column was eluted with pure water, 0.1% formic acid, and methanol solutions at concentrations of 5%, 10%, 15%, and 20% to identify the optimal elution conditions. Preliminary results indicated that the average recovery rates for drugs eluted with 20% methanol ranged between 50% and 90%, with significant losses observed for certain individual analytes.

To further optimize the elution conditions, 17% methanol, 17% acetonitrile and 18% methanol, 18% acetonitrile were evaluated as eluents. The effects of these elution conditions on the recovery rates of the target drugs and the interference from feed matrix impurities were investigated. The average recovery rates under the two elution conditions are illustrated in Figure 5a, while the matrix effect rates are shown in Figure 5b. The results demonstrated that both elution methods had a minimal impact on the recovery rates of the target drugs when using the HLB purification column. However, the 18% methanol/18% acetonitrile eluent significantly improved the average recovery rate of IVM (ivermectin). This enhancement is likely attributed to the more efficient elution, better removal of impurities, and improved chromatographic response at this concentration. Additionally, the matrix effect observed in chromatograms treated with the 18% eluent was lower compared to those treated with the 17% eluent, indicating superior sample purification efficiency. To ensure the complete removal of impurities, the HLB purification column was sequentially eluted with 18% methanol followed by 18% acetonitrile.

3.2.4. Improvement of Sample Preparation Conditions

During the analysis of multiple samples using the established sample pretreatment and instrument detection methods, significant interference was observed in the chromatographic peaks of IVM and MOX across various feed formulations. This interference prevented the proper integration of the drug peaks in spiked samples at the quantitative limit (QL). Additionally, the recovery rates of EPR (eprinomectin) in some concentrated feed samples were notably low.

To address issues related to matrix interference and extraneous peaks, an additional n-hexane extraction step was incorporated into the purification process for formula feed and concentrated feed samples. This step is specifically designed to remove common non-polar interfering substances, such as lipids and pigments, from the feed matrix (Maštovská & Lehotay, 2004) [27]. The average recovery rates of the target drugs in formulated feed, with and without n-hexane treatment, are presented in Figure 6a, while the corresponding results for concentrated feed are shown in Figure 6b. The results demonstrated that the recovery efficiency for both feed types improved significantly after n-hexane treatment. This step successfully resolved the interference issues affecting the analysis of IVM and MOX in formulated feed. Furthermore, all target drugs in the concentrated feed samples exhibited satisfactory recovery rates. These findings suggest that for samples with complex matrices, where conventional methods fail to yield reliable results, the inclusion of an n-hexane extraction step during sample pretreatment can effectively purify the samples and meet analytical requirements.

3.3. Method Validation

To further validate the accuracy and precision of the method, tests were conducted based on the potential drug concentrations typically found in various feed categories (Anastassiades et al., 2003) [24]. The spiking levels for formulated feed and vitamin feed were set at 0.05 μg/g, 0.5 μg/g, and 5.0 μg/g. For concentrated feed and bovine concentrate, the concentrations were 0.1 μg/g, 1.0 μg/g, and 10.0 μg/g, respectively. The premixed feed was spiked at three levels: 0.5 μg/g, 5 μg/g, and 50 μg/g. At each concentration level, five replicate samples were prepared, along with corresponding blank controls. The samples were processed according to the established sample pretreatment protocol and analyzed using LC-MS/MS. The concentration and recovery rates were calculated based on the peak areas of the target analytes.

The results of the recovery tests are presented in

Table 3. Recovery rate and coefficient of variation in acetamiprid B1a added to feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	91.27	2.09	2.43
	0.5	90.38	3.51	4.75
	5.0	93.19	3.32	4.70
Concentrated chicken feed	0.1	92.44	7.74	8.03
	1.0	90.06	1.98	3.94
	10.0	95.25	6.08	6.52
Chicken premix feed	0.5	91.77	5.55	6.18
	5.0	95.26	4.70	5.35
	50.0	97.72	3.80	4.90
Pig compound feed	0.05	87.13	4.95	5.17
	0.5	82.75	4.67	4.85
	5.0	85.48	2.88	4.66
Concentrated pig feed	0.1	81.17	9.97	8.74
	1.0	83.76	4.85	11.76
	10.0	88.45	5.72	9.28
Pig premix feed	0.5	98.07	10.14	10.59
	5.0	96.82	7.24	8.82
	50.0	95.44	5.08	6.75
Vitamin supplements	0.05	90.50	4.37	5.55
	0.5	91.73	6.42	7.12
	5.0	88.42	3.79	4.71
Cattle concentrate	0.1	85.78	7.90	10.76
	1.0	83.12	11.26	12.55
	10.0	81.94	9.97	10.87

,

Table 4. Recovery rate and coefficient of variation in avermectin B1a added to feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	89.33	3.42	4.57
	0.5	87.36	4.18	5.90
	5.0	86.10	4.07	5.82
Concentrated chicken feed	0.1	91.52	8.46	9.26
	1.0	87.77	3.70	4.51
	10.0	90.45	7.85	7.95
Chicken premix feed	0.5	89.08	7.17	8.62
	5.0	93.26	5.63	6.38
	50.0	95.36	4.85	6.07
Pig compound feed	0.05	85.37	5.76	8.42
	0.5	80.42	8.02	9.65
	5.0	83.55	3.94	5.74
Concentrated pig feed	0.1	82.73	10.11	11.66
	1.0	80.96	7.05	8.09
	10.0	85.77	7.29	8.75
Pig premix feed	0.5	96.64	11.33	10.93
	5.0	95.28	5.82	7.80
	50.0	90.78	6.78	9.48
Vitamin supplements	0.05	87.72	5.46	6.75
	0.5	86.66	7.39	8.82
	5.0	84.53	4.74	5.46
Cattle concentrate	0.1	83.18	8.29	9.73
	1.0	80.07	10.78	11.16
	10.0	80.22	10.32	11.49

,

Table 5. Recovery rate and coefficient of variation in milbemycin added to feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	92.72	5.48	6.68
	0.5	92.99	9.12	10.62
	5.0	95.58	7.06	8.37
Concentrated chicken feed	0.1	94.09	5.55	6.90
	1.0	95.51	8.17	9.54
	10.0	97.76	5.92	6.95
Chicken premix feed	0.5	93.82	6.34	8.49
	5.0	97.89	7.11	8.82
	50.0	98.80	4.65	5.79
Pig compound feed	0.05	88.48	5.85	7.37
	0.5	85.72	5.90	7.80
	5.0	87.97	5.36	7.72
Concentrated pig feed	0.1	84.79	11.32	10.09
	1.0	85.65	5.76	6.66
	10.0	90.29	6.80	7.82
Pig premix feed	0.5	97.02	11.27	12.52
	5.0	97.77	8.03	10.00
	50.0	96.64	6.22	8.43
Vitamin supplements	0.05	92.70	5.72	6.07
	0.5	90.84	7.94	8.86
	5.0	90.56	5.59	7.54
Cattle concentrate	0.1	89.63	8.06	7.98
	1.0	88.45	10.47	11.36
	10.0	83.74	10.12	13.38

,

Table 6. Recovery rate and coefficient of variation in moxifloxacin in feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	84.52	3.76	4.76
	0.5	80.67	4.87	5.59
	5.0	83.99	5.98	6.80
Concentrated chicken feed	0.1	87.75	9.18	10.03
	1.0	85.08	3.75	5.59
	10.0	91.22	7.72	8.11
Chicken premix feed	0.5	86.64	7.67	8.76
	5.0	90.90	6.08	7.95
	50.0	95.57	9.92	11.04
Pig compound feed	0.05	85.70	7.67	8.88
	0.5	84.49	7.76	8.92
	5.0	86.19	5.61	6.64
Concentrated pig feed	0.1	85.53	10.85	11.09
	1.0	84.74	6.76	7.80
	10.0	85.07	6.70	7.92
Pig premix feed	0.5	91.82	11.55	10.61
	5.0	94.76	9.13	10.07
	50.0	93.80	6.60	7.53
Vitamin supplements	0.05	88.56	7.34	8.57
	0.5	87.79	7.89	8.86
	5.0	87.47	5.63	7.90
Cattle concentrate	0.1	82.80	8.22	9.49
	1.0	81.90	12.84	13.69
	10.0	80.93	11.48	12.25

,

Table 7. Recovery rate and coefficient of variation in doramectin added to feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	90.01	3.82	4.09
	0.5	90.42	4.75	6.74
	5.0	95.55	5.78	6.92
Concentrated chicken feed	0.1	90.18	9.07	10.38
	1.0	83.64	5.57	7.72
	10.0	91.16	9.82	10.52
Chicken premix feed	0.5	89.35	5.46	7.28
	5.0	92.73	7.78	8.95
	50.0	95.00	9.12	10.41
Pig compound feed	0.05	83.96	8.33	9.90
	0.5	81.92	6.78	7.92
	5.0	83.27	5.29	6.67
Concentrated pig feed	0.1	80.18	8.16	9.76
	1.0	82.79	6.54	7.88
	10.0	95.55	6.87	9.04
Pig premix feed	0.5	99.65	9.95	10.56
	5.0	97.38	8.61	9.90
	50.0	97.01	6.72	8.75
Vitamin supplements	0.05	90.22	9.14	10.37
	0.5	89.67	7.60	8.65
	5.0	85.45	6.85	7.74
Cattle concentrate	0.1	83.76	6.69	7.80
	1.0	82.39	10.75	11.13
	10.0	82.72	9.17	10.45

and

Table 8. Recovery rate and coefficient of variation in ivermectin H2B1a added to feed.

Feed	Added Concentration (μg/g)	Recovery Rate (%)	Intra Batch Coefficient of Variation (%)	Inter Batch Coefficient of Variation (%)
Chicken compound feed	0.05	89.90	6.57	7.09
	0.5	89.64	5.32	6.18
	5.0	91.45	5.56	6.65
Concentrated chicken feed	0.1	90.35	8.14	9.53
	1.0	89.56	5.42	6.92
	10.0	91.82	7.73	9.94
Chicken premix feed	0.5	90.47	6.05	7.36
	5.0	93.25	5.47	7.77
	50.0	95.83	4.43	6.98
Pig compound feed	0.05	85.28	5.96	9.04
	0.5	81.14	5.61	7.82
	5.0	83.07	3.82	7.67
Concentrated pig feed	0.1	82.64	10.04	11.12
	1.0	81.19	7.76	7.85
	10.0	85.75	6.80	7.96
Pig premix feed	0.5	95.55	11.17	10.99
	5.0	93.42	8.62	9.92
	50.0	94.74	6.87	7.08
Vitamin supplements	0.05	87.05	6.65	7.85
	0.5	89.67	7.48	9.74
	5.0	86.83	6.70	9.97
Cattle concentrate	0.1	83.48	8.23	9.43
	1.0	80.92	10.96	11.16
	10.0	80.78	10.51	11.82

. The results demonstrated that the average recovery rates for the different feed categories ranged from 60% to 120%, with coefficients of variation (CV) within 15% at all spiking levels. These findings confirm that the method meets the required standards for accuracy and precision in drug residue analysis in feed (European Commission, 2002) [28].

To clearly evaluate the interference of matrix impurities on the target analytes, the chromatogram of a 250 μg/kg spiked feed sample was compared with that of a blank matrix spiked with standard solution, as illustrated in the figures below (Figure 7, Figure 8, Figure 9, Figure 10, Figure 11, Figure 12 and Figure 13). The results demonstrated that the recovery rates for compound feed, concentrated feed, premixed feed, and cattle concentrate met the required standards at the retention times corresponding to acetyl avermectin B1a, avermectin B1a, milbemycin, moxidectin, doramectin, and ivermectin H2B1a.

4. Conclusions

This study developed a highly efficient and selective HPLC-MS/MS method for the simultaneous quantification of six macrolide antiparasitic drugs in aquaculture feed. These compounds are widely used in aquaculture and livestock production; however, their residual presence may pose significant risks to animal health, food safety, and ecological integrity. Consequently, the establishment of a rapid and precise detection methodology is crucial for monitoring drug residue levels, ensuring food safety, and advancing sustainable aquaculture practices.

The validation results demonstrate that this method exhibits exceptional specificity, accuracy, and practicality, aligning with internationally recognized standards such as ICH Q2(R1) and EU Directive 2002/657/EC. As such, it is well-suited for routine detection and analysis. This approach not only offers robust technical support for feed manufacturers and regulatory agencies but also provides a scientific foundation for the regulated use of macrolides, thereby mitigating issues related to drug misuse and the emergence of antimicrobial resistance.

Looking ahead, we aim to further refine this method by developing faster, more efficient, and more selective HPLC-MS/MS technologies to enhance its analytical performance and broaden its applicability. Potential future directions include extending the method’s application to other complex matrices, such as food and environmental samples, as well as integrating it with advanced technologies like artificial intelligence-assisted data analysis to enable more intelligent and automated detection workflows. Through ongoing innovation and optimization, we aspire to contribute significantly to global efforts in ensuring food safety and improving animal health management.