Simultaneous determination of vitamin A and E in infant formula using a modified QuEChERS approach coupled with ultrahigh-performance liquid chromatography-diode array detection

Abstract

The simultaneous determination of vitamin A and E vitamers in complex matrices presents a significant analytical challenge. To address this, we validated a straightforward method based on a QuEChERS pretreatment coupled with ultrahigh-performance liquid chromatography and diode array detection (QuEChERS-UPLC-DAD). The modified QuEChERS pretreatment procedure integrates saponification and extraction processes into a single step, enabling efficient phase separation, reducing analysis time, and eliminating the need for a salting-out step. Notably, this approach simultaneously accomplishes a recovery of 83.2% to 100.5%, a relative standard deviation of less than 4.6%. The accuracy of the proposed QuEChERS-UPLC-DAD method was confirmed through comparison with a UPLC-IDMS/MS method and analysis of a commercial certified reference material (NIST SRM 1869). Importantly, the approach was successfully applied to analyze various complex dairy products, including infant formula, milk, yogurt, and dairy beverages. Therefore, QuEChERS-UPLC-DAD approach is well-suited for routine quality control and nutritional assessment.

Highlights

• •A modified QuEChERS method integrates saponification and extraction into one step.
• •Eliminating the salting-out process and reducing analysis time.
• •Satisfactory recoveries (83.2–100.5%) with excellent precision (RSD < 4.6%).
• •Accuracy validated by UPLC-IDMS/MS and certified reference material (NIST SRM 1869).
• •Successfully applied to analyze various complex dairy products.

1. Introduction

Vitamins A and E are essential nutrients for both adults and infants, playing critical roles in maintaining biological and immunological functions (Liu et al., 2025). For instance, vitamin A is indispensable for vision and immune competence, and its deficiency is recognized as a leading cause of preventable childhood blindness worldwide (Fang et al., 2026). Vitamin E acts as a primary lipid-soluble antioxidant, protecting unsaturated fatty acids from oxidative damage (Levêques et al., 2019). Infant formulas often serve as a principal source of these vitamins for infants. Accordingly, regulations have been established to strictly define the content of vitamin A (ranging from 14 to 50 μg/100 kJ) and vitamin E (ranging from 0.12 to 1.2 mg/100 kJ) in infant formulas (Chávez-Servín et al., 2008). Moreover, the compounds permitted for addition as vitamin A include retinol, retinyl acetate, retinyl palmitate, and β-carotene. Vitamin E comprises four tocopherols, characterized by a saturated isoprenoid side chain and designated as α-, β-, γ-, and δ-tocopherol based on the number and position of methyl groups on the chromanol ring. Thus, vitamins A and E could be analyzed simultaneously due to their structural and chemical similarities. It should be noted that vitamin A is sensitive to light, oxidation, heat, and acidic conditions, while vitamin E is also susceptible to light and oxidation (Zou et al., 2025). Moreover, the aforementioned compounds, whether in their free forms or as fatty acid esters, coexist in various lipid matrices (Wang et al., 2022). These complex matrices typically contain lipids such as glycerides, sterols, and phospholipids, all of which exhibit solubility properties similar to those of vitamins A and E. The presence of co-extracted lipids can interfere with accurate analysis. Consequently, the complexity of the food matrix, coupled with the varying stabilities and endogenous levels of individual vitamers, renders their extraction and isolation from the lipid phase a significant analytical challenge (Konings et al., 2024; Li et al., 2025). Consequently, there is a high demand for efficient detection techniques and robust extraction methods (Fanali et al., 2017).

High-performance liquid chromatography (HPLC) possesses the advantages of high resolution and compatibility with various detectors, which enables the determination of vitamers even in some complex matrices, thereby addressing the analytical challenges posed by co-extracted lipids. Conventional analytical methods based on HPLC with ultraviolet (UV) or fluorescence detection have been widely used to determine vitamers in simple matrices (Delgado-Zamarreño et al., 2016; Fatima et al., 2023; Sadrykia et al., 2019; Viñas et al., 2013). More recently, liquid chromatography-tandem mass spectrometry (LC-MS/MS) has enabled the simultaneous determination of multiple fat-soluble vitamins in various complex sample matrices (Ertugrul et al., 2020; Fatima et al., 2023; Gentili et al., 2013; Kim et al., 2022; Plozza et al., 2012). Regarding these detection techniques, hot saponification is commonly employed as an extraction method to remove neutral lipids from lipid-rich foods. This process releases endogenous vitamins from their esterified forms, yielding free vitamins for subsequent analysis (Fanali et al., 2017; Lee et al., 2013; Plozza et al., 2012; Zhang et al., 2021). The typical sample pretreatment workflow involves liquid-liquid extraction using organic solvents (Nimalaratne et al., 2014), followed by a cleanup step via solid-phase extraction (SPE) (Sunarić et al., 2017). However, these procedures are generally labor-intensive, time-consuming, and require large volumes of organic solvents. To address these limitations, dispersive liquid-liquid microextraction (DLLME) has been developed for the isolation of vitamin E from infant formula samples (Sadrykia et al., 2019). Despite these advances, the development of efficient and effective sample pretreatment techniques remains an ongoing demand. Recently, the QuEChERS approach, has gained considerable attention due to its simplicity, low-cost, high throughput, and favorable recovery rates (Decheng et al., 2022; Rodríguez-Cañás et al., 2023; Santana-Mayor et al., 2019). This pretreatment method has since been adapted for the determination of vitamin E and sitosterols in seeds and nuts (Delgado-Zamarreño et al., 2016), as well as vitamins D and K in yogurt (Kim et al., 2022). It should be noted, however, that these previously reported methods are typically limited to the detection of only one or two vitamins per analysis. To the best of our knowledge, a rapid extraction procedure capable of simultaneously processing vitamin A and multiple forms of vitamin E has rarely been reported.

Herein, we developed a modified QuEChERS approach coupled with ultra-performance liquid chromatography with diode array detection (UPLC-DAD) for the simultaneous quantification of vitamin A and various forms of vitamin E in complex matrices. This modified QuEChERS method integrates saponification and extraction into a single step, using isopropanol as both the saponification and extraction solvent. For sample cleanup, an aminopropyl-modified silica sorbent was employed in a dispersive solid-phase extraction (dSPE) procedure. The proposed QuEChERS-UPLC-DAD technique is both time- and solvent-saving. The validated method was successfully applied to the rapid quantification of these vitamins in dairy products, including milk, yogurt, and dairy drinks.

2. Materials and methods

2.1. Chemicals and materials

Vitamin A (purity >99.9%), (±) α-tocopherol (purity >99.9%), (±) β-tocopherol (purity >99.7%), D-γ-tocopherol (purity >99.4%), and D-δ-tocopherol (purity >91.9%) were obtained from First standard (Tianjin, China). Vitamin A-D₆ (97.8%), α-tocopherol-D₆ (99.3%) were obtained from BePure (Beijing, China). Acetonitrile (ACN, HPLC grade), methanol (MeOH, HPLC grade), ethanol (EtOH, HPLC grade) and isopropanol (IPA, HPLC grade) were purchased from Merck (Darmstadt, Germany). Potassium hydroxide (KOH) was purchased from Damo (Tianjin, China) and butylhydroxytoluene (BHT) was purchased from Thermo Fisher (Shanghai, China). Sodium chloride (NaCl) was purchased from J&K Scientific (Beijing, China), and anhydrous magnesium sulfate (MgSO₄) was purchased from Macklin (Shanghai, China). Pure water was supplied by Ultrapure Water Purification System (Milli-Q®) in the study. C₁₈ reversed-phase silica gel (C₁₈), aminopropyl-modified silica sorbent (NH₂), primary secondary amine (PSA) sorbents were purchased from Agela Technologies (Lake Forest, CA, USA), and Z-sep sorbent for dSPE was from ANPLE (Shanghai, China).

2.2. Standard solutions

The stock solutions of vitamin A and the four tocopherols were gravimetrically prepared in methanol at a concentration of 500 μg/g, respectively. The concentrations of calibration solutions in the UPLC-DAD method were prepared in the range from 0.02 to 20 μg/g for vitamin A and 0.2–60 μg/g for vitamin E. The concentrations of calibration solutions were prepared at 0.05 μg/g for vitamin A and 0.6 μg/g vitamin E, respectively. All the solutions in the brown vials were stored at −20 °C for further use.

2.3. Pretreatment of sample (QuEChERS process)

An aliquot of infant formula (1.000 g) was added to 5 mL of water in a 50 mL centrifuge tube, after which the mixture was shaken until completely dissolved. Subsequently, 15 mL of isopropanol containing BHT with a concentration of 0.2 g /100 mL, and 5 mL of KOH solution (1.00 g/mL) were added into the tube, and they were further vortexed for 1 min. After that, the sample was undergone the hot saponification in a water bath shaker at 80 °C for 30 min (Fanali et al., 2017; Karrar et al., 2022). Then, the solution was naturally cooled to room temperature, and centrifuged at 3000 rpm for 3 min. Isopropanol layer (4 mL) in the tube was then transferred to a dispersive solid phase extraction (d-SPE) vial, which contained 150 μL of HCl solution (2.5 mol/L), 1 g of MgSO₄, and 150 mg of aminopropyl-modified silica sorbent. The pH was in the range from 8.00 to 8.50. After vortexed for 30 s, the resultant supernatant was filtered through an ANPLE PP filter (0.22 μm), and they were injected into the UPLC-DAD analysis system. For comparison, UPLC-MS/MS detection was also performed. In this case, a certain amount of vitamin A-D₆ or α-tocopherol-D₆ were spiked into infant formula sample (1.000 g). The following pretreatment processes were identical with those in UPLC-DAD method. 1 mL of isopropanol extract was transferred into the d-SPE vial, which contained 15 mL acetonitrile and 10 mL water, 1.000 g NaCl, 1.000 g MgSO₄, and 150 mg aminopropyl-modified silica sorbent. After vortexed for 30 s, the supernatant was filtered using ANPLE PP filter (0.22 μm) prior to injection into the UPLC-MS/MS analysis system.

2.4. Instrument and apparatus

UPLC-DAD was performed on an Agilent 1290 Infinity II UPLC system equipped with a diode array detector (DAD). The DAD detection wavelength was set at 325 nm for vitamin A and at 294 nm for four tocopherols, respectively. The chromatographic column was 120 PFP (InfinityLab Poroshell, 2.1 mm × 100 mm, 1.9 μm) (Agilent, USA). The mobile phases consisted of H₂O (phase A) and CH₃OH (phase B). The column temperature was 35 °C. The injection volume was 2 μL at a flow rate of 0.4 mL/min. The gradient elution program started from 75% B and kept for 3 min, and then, the proportion of B increased linearly to 85% within 2 min, and kept increasing to 90% B within 7 min. After that, the gradient elution changed to 75% B at 12.1 min and held isocratic for 2 min. UPLC-MS/MS detection was conducted on a Shimadzu LC-30 CE UPLC system connected to a SCIEX QTRAP AB5500 mass spectrometer. The mobile phase was composed of water (phase A) and methanol (phase B); both phases were supplemented with 0.05% formic acid and 5 mM ammonium formate. The chromatographic columns and gradient conditions were identical with those in the above-mentioned UPLC-DAD method. The parameters for AB5500 triple quadrupole mass spectrometry: the ion source was an ESI⁺ source and the analytes were monitored using multiple reaction monitoring (MRM) mode. The analytes were quantified in multiple reaction monitoring (MRM) mode for vitamin A, tocopherol and their isotopic internal standards. The typical parent ion was m/z 269.3 (vitamin A), m/z 275.1 (vitamin A-D₆), m/z 431.8 (α-tocopherol), m/z 437.4 (α-tocopherol-D₆), m/z 417.4 (β-tocopherol), m/z 417.4 (γ-tocopherol), m/z 403.4 (δ-tocopherol) and corresponding product ions were m/z 93.1 and 213.2 (vitamin A), m/z 96.0 and 84.1 (vitamin A-D₆), m/z 165.1 and 111.1 (α-tocopherol), m/z 171.4 and 211.4 (α-tocopherol-D₆), m/z 150.9 and 97.1 (β-tocopherol), m/z 150.9 and 97.1 (γ-tocopherol), m/z 137.1 and 177.3 (δ-tocopherol).

2.5. Method validation

QuEChERS-UPLC-DAD technique for determining vitamin A and four tocopherols were validated by means of the linearity, limit of detection (LOD), limit of quantification (LOQ), precision, and accuracy. LOD and LOQ were defined as concentrations with signal-to-noise ratio (S/N) greater than 3 and 10, respectively (Ma et al., 2025). The corresponding calibration solutions for vitamin A were prepared in methanol at concentrations of 0.2, 1.0, 2.5, 5.0, 10.0, 20.0 μg/g. The concentrations of 1.0, 5.0, 10.0, 20.0, 40.0, 60.0 μg/g were for the four tocopherols, accordingly. As for the UPLC-IDMS/MS method, the quantification was evaluated using an external standard method. The concentrations of vitamin A calibration solution were 6.0, 15.0, 50.0, 200.0, 500.0 ng/g with vitamin A-D₆ of 50.0 ng/g as reference, and the four tocopherols of vitamin E were 2.0, 10.0, 50.0, 200.0, 600.0 ng/g with α-tocopherol-D₆ 100.0 ng/g as reference. The accuracy of UPLC-IDMS/MS method was determined by spiking at three different concentrations of vitamin A (3, 10, 20 μg/g) and vitamin E (20, 60, 150 μg/g) standard solutions into infant formula, and the spiked recoveries were calculated accordingly. Additionally, its precision was evaluated by means of relative standard deviation (RSD) and matrix effect (ME) according to previous reports (Ma et al., 2025; Zhao et al., 2024).

3. Results and discussion

3.1. Optimization of UPLC-DAD and UPLC-IDMS/MS

Chromatographic and mass spectrometric parameters were systematically optimized to achieve optimal separation and detection. In the UPLC-DAD system, key variables, including stationary phase, mobile phase composition, and injection volume, were evaluated to improve peak resolution. Three columns with different stationary phases were compared: an Agilent PFP (2.1 mm × 100 mm, 1.9 μm), an Agilent PAH (2.1 mm × 100 mm, 1.8 μm), and a Waters BEH C18 (2.1 mm × 100 mm, 1.7 μm). The PFP column provided the best separation, whereas the PAH and BEH C18 columns failed to fully resolve β- and γ-tocopherol (Fig. 1a). The enhanced selectivity of the PFP phase can be attributed to its pentafluorophenyl group, which contains electron-withdrawing fluorine atoms that promote specific interactions with electron-rich analytes such as β- and γ-tocopherol (Delgado-Zamarreño et al., 2016; Karrar et al., 2022). Based on instrument software analysis, the resolution of β-tocopherol and γ-tocopherol was 2.4. The theoretical plate numbers for β and γ-tocopherols were 6.1 × 10⁴ and 4.7 × 10⁴, respectively. The two parameters indicated that good separation efficiency was achieved for the two compounds, meeting the requirements of the quantification. Methanol and water were selected as the mobile phase components, and a representative chromatogram for all analytes is shown in Fig. 1b. Detection wavelengths were set at 325 nm for vitamin A and 294 nm for vitamin E, respectively. In the UPLC-MS/MS system, precursor and product ions were optimized for each target compound and isotopically labeled internal standard through infusion experiments conducted in positive ion mode. Full-scan mass spectra were acquired over an m/z range of 100–400, and the most abundant ions were selected for fragmentation to obtain the corresponding MS/MS transitions. The selected mass transitions are consistent with those previously reported in the literature (Arachchige et al., 2021; Rubió et al., 2020). The composition of the aqueous mobile phase was also optimized, and the addition of 0.05% formic acid and 5.0 mmol/L ammonium formate resulted in satisfactory signal intensity and stability. Fig. 1c presents a representative chromatogram of the target analytes and isotopically labeled internal standards obtained under the finalized conditions.

Fig. 1.

(a) Column optimization in the UPLC-DAD system; (b) typical chromatograms of vitamin A and vitamin E (α-, β-, γ-, and δ-tocopherol) at 325 nm and 294 nm; (c) Chromatograms of UPLC-MS/MS including, vitamin A, vitamin E and corresponding isotopically labeled internal standards.

3.2. QuEChERS pretreatment procedures

Saponification is a standard pretreatment step used to liberate fat-soluble vitamins from the lipid matrix and hydrolyze their esterified forms into free compounds (Fanali et al., 2017; Karrar et al., 2022). In this study, saponification was evaluated as a simultaneous extraction procedure for vitamins A and E. Various organic solvents were compared (Fig. 2a), including ethanol (EtOH), acetonitrile (ACN), isopropanol (IPA), and methanol (MeOH). When methanol or ethanol was used, the recovery of vitamin A was below 30% or 65%, respectively. In contrast, recovery increased to over 80% with acetonitrile and isopropanol. Furthermore, among the four tocopherols (α-, β-, γ-, and δ-tocopherol), the highest recoveries were also achieved using isopropanol. Therefore, isopropanol was selected as the saponification solvent due to its satisfactory extraction efficiency for all target vitamers (Fig. 2a). Subsequently, saponification time was optimized at 80 °C for durations of 5, 15, 30, and 60 min. Optimal recoveries exceeding 80% for both vitamin A and the four tocopherols were achieved, with recoveries reaching up to 88% at a saponification time of 30 min (Fig. 2b). Finally, the volume of isopropanol was optimized, and 15 mL provided the highest recoveries for all five vitamers, as shown in Fig. 2c. It should be noted that conventional QuEChERS protocols for the detection of fat-soluble vitamins in fortified nanoemulsions and nuts typically employ methanol or acetonitrile as the extraction solvent, followed by the addition of anhydrous MgSO₄ and NaCl to induce phase separation (Gao et al., 2021; Kumari et al., 2023; Santana-Mayor et al., 2019). In contrast, the present study introduces isopropanol as the saponification solvent for the first time. Notably, after the saponification process, phase separation between isopropanol and the aqueous layer occurred spontaneously, eliminating the need for salt addition.

Fig. 2.

Optimization of QuEChERS pretreatment procedures: Various solvents (a), saponification time (b), volume of isopropanol (c) and different sorbents in dSPE (d) on the recoveries of the five analytes (vitamin A and α-, β-, γ-, and δ-tocopherol).

As a result, the isopropanol layer could be directly collected for the subsequent dispersive solid-phase extraction (dSPE) cleanup step. It is worth emphasizing that this modified QuEChERS approach integrates saponification, extraction, and phase separation into a single step, without the use of added salts. Thereby, the modified QuEChERS approach simplifies the pretreatment workflow while maintaining high recovery efficiencies. The saponification pretreatment step of the QuEChERS method using KOH hydrolyzed the fats in infant formula, producing large amounts of potassium fatty acid salts. Moreover, the resulting salts could induce phase separation between the aqueous and polar organic solvent layers. In this study, isopropanol was used as the saponification solvent in infant formula. Notably, under the optimal conditions, phase separation between isopropanol and the aqueous layer occurred spontaneously after the saponification process, eliminating the need for salt addition. However, phase separation was not observed for acetonitrile, ethanol, or methanol. This difference was attributed to their distinct dielectric constants and polarities. (Kalinke et al., 2025).

Furthermore, it is well known that the pH of the isopropanol solution plays a critical role in the dispersive solid-phase extraction (dSPE) step. Given that the saponification process generates an alkaline medium, which may interfere with analyte–sorbent interactions, the extract was neutralized prior to cleanup by adding an appropriate volume of hydrochloric acid (2.5 mol/L) to adjust the pH for subsequent UPLC-DAD analysis. Various sorbents for dSPE were evaluated, including PSA, PSA/C18, NH₂, and Z-Sep (Fig. 2d). For all sorbents tested, the recoveries of vitamin A exceeded 78%, indicating the overall effectiveness of the optimized QuEChERS protocol. However, when the PSA/C18 sorbent was used, the recoveries of the four tocopherols were below 43%, which were considerably lower than those achieved with PSA alone. The poorer performance can be attributed to the strong retention of tocopherols on the C18 phase, which possesses a high affinity for lipophilic compounds (Santana-Mayor et al., 2019). However, recoveries for the four tocopherols ranged from 55% to 70% when PSA or Z-Sep was employed. Notably, the NH₂ sorbent exhibited the highest recoveries and the best overall extraction performance. More importantly, it delivered consistently high recoveries across all five analytes simultaneously, with values ranging from 70% to 90%. To further optimize the NH₂-based cleanup, sorbent amounts of 50, 100, 150, and 200 mg were tested. Considering both good repeatability (as indicated by small standard deviations) and cost efficiency, a sorbent dosage of 100 mg was selected for subsequent experiments (Fig. S1).

3.3. Optimization of UPLC-MS/MS method

To validate the UPLC-DAD analysis, a UPLC-MS/MS method employing electrospray ionization (ESI) was simultaneously performed. Matrix effects (ME) were first evaluated, as they can adversely affect detection sensitivity and precision (Rodríguez-Cañás et al., 2023; Li et al., 2023). The ME values (ME_α and ME_β) and the correction factor θ (θ = ME_α / ME_β) were calculated accordingly. The ME values for vitamin A and α-, β-, γ-, and δ-tocopherol ranged from 0.93 to 0.98 (Table S1), indicating negligible ion suppression or enhancement for all target analytes. These results further demonstrate that the proposed modified QuEChERS procedure effectively removes interfering matrix components. The corresponding θ values for the five analytes were 0.99, 1.01, 1.01, 1.03, and 0.97, respectively. Additionally, the RSDs are lower than 6.0%. Notably, the consistency of these θ values across all target compounds further validates the suitability of isotopically labeled internal standards for the accurate quantification of vitamins A and E in complex matrices.

3.4. Method validation

The proposed QuEChERS method was validated in terms of linearity, limits of detection (LOD) and quantification (LOQ), recovery, and precision. The corresponding data for the UPLC-DAD and UPLC-IDMS/MS methods are summarized in Table 1 and Table S2, respectively. Calibration curves were constructed using six concentration levels for both analytical methods. In the UPLC-DAD method, the concentration ranged from 0.2 to 20 μg/g for vitamin A and from 1.0 to 60 μg/g for vitamin E. In contrast, the concentration range for the UPLC-IDMS/MS method was 2.0 to 600 ng/g. Excellent linearity was achieved for all analytes in both methods, with excellent correlation coefficients (R²). For the UPLC-DAD method, recoveries for all analytes ranged from 83.2% to 100.5%, with relative standard deviations (RSD) below 4.6%.

Table 1.

Validation parameters for Vitamin A and Vitamin E vitamers detected using UPLC-DAD method.

Analyte	Calibration curve	R2	LOD (μg/g)	LOQ (μg/g)	Recovery	RSD
Vitamin A	y = 26.025× - 0.4583	0.9999	0.06	0.19	89.0–100.5%	˂ 4.6%
α-tocopherol	y = 1.3257× - 0.5093	0.9999	0.63	2.10	87.7–92.6%	˂ 3.4%
β-tocopherol	y = 1.9894× - 0.2099	0.9999	0.36	1.19	87.1–88.5%	˂ 4.0%
γ-tocopherol	y = 2.0085× - 0.1538	0.9999	0.37	1.23	86.4–90.2%	˂ 3.4%
δ-tocopherol	y = 1.9169× - 0.0587	0.9999	0.36	1.18	83.2–87.4%	˂ 3.0%

In comparison, the UPLC-IDMS/MS method yielded recoveries between 95.6% and 107.8%, with RSD values below 4.0%. The limits of detection (LOD) and quantification (LOQ) for the UPLC-DAD method were 0.06–0.63 μg/g and 0.19–2.10 μg/g, respectively. For the UPLC-IDMS/MS method, the corresponding LOD and LOQ ranged from 0.81 to 2.07 ng/g and from 2.70 to 6.90 ng/g, respectively. Furthermore, the reproducibility of the method was assessed using infant formula samples. For both analytical methods, the intra-day and inter-day precision (RSD) were less than 4.8%, demonstrating good reproducibility and confirming the method's suitability for routine analysis of real samples.

The accuracy of the proposed QuEChERS-UPLC-DAD method was further verified by analyzing a NIST commercial reference material (SRM 1869), which contains the five targets. The reference and measured values for these targets, obtained by both UPLC-DAD and UPLC-MS/MS methods, are summarized in Table 2. These results strongly demonstrate that the proposed QuEChERS-UPLC-DAD method provides reliable and accurate quantification of vitamins A and E in infant formula. Furthermore, the E_n  value was used to assess the consistency of measurement results obtained from different analytical methods (Steele & Douglas, 2006). The E_n value can be calculated using Eq. (1),

$$E_n=\frac{x_{\mathit{mean}}-x_{\mathit{ref}}}{\sqrt{U_{\mathit{mean}}^2+U_{\mathit{ref}}^2}}$$ (1)

where $x_{\mathit{mean}}$, $x_{\mathit{ref}}$, $U_{\mathit{mean}}$, and $U_{\mathit{ref}}$ represent the value detected by the proposed method, the certified value of the NIST reference material, the expanded uncertainty of the proposed method, and the expanded uncertainty of the NIST reference material, respectively. The corresponding E_n values are summarized in Table S3. The results show that the E_n values for UPLC-DAD and UPLC-IDMS/MS methods ranged from 0.03 to 0.84. All these values are below 1, indicating good agreement between the measured values obtained using two methods and the reference values (NIST SRM).

Table 2.

Comparison between the reference values and the measured values of Vitamin A and Vitamin E vitamers in the NIST SRMs, measured using UPLC-DAD and UPLC-MS/MS methods.

Analytes	NIST (mg/kg)	UPLC-MS/MS (mg/kg)	UPLC-DAD (mg/kg)
Vitamin A	19.27 ± 0.32	19.55 ± 0.48	19.32 ± 0.39
α-tocopherol	217.2 ± 6.2	217.60 ± 7.55	217.53 ± 8.43
β-tocopherol	4.22 ± 0.69	4.845 ± 0.46	4.639 ± 0.51
γ-tocopherol	99.4 ± 5.1	100.22 ± 5.33	99.82 ± 5.79
δ-tocopherol	32.5 ± 2.9	33.43 ± 2.82	31.89 ± 2.74

As summarized in Table S4, conventional methods typically rely on saponification coupled with liquid-liquid extraction (LLE), which are labor-intensive and consume large volumes of organic solvent (30–150 mL per sample). These procedures require extended processing times (1–15.5 h) due to lengthy saponification, overnight drying, or low-temperature storage, and are often limited to single matrices. In contrast, the proposed QuEChERS-UPLC-DAD approach streamlines sample preparation to under 30 min and reduces solvent usage to approximately 15 mL. It demonstrates broad applicability across diverse dairy products, including milk, yogurt, dairy beverages, and infant formulas. Moreover, it enables simultaneous determination of vitamin A and multiple vitamin E vitamers, providing a rapid, cost-effective, and environmentally sustainable solution for routine monitoring in dairy products.

3.5. Determination of vitamers in real dairy products

The proposed QuEChERS-UPLC-DAD method was subsequently applied to quantify vitamin A and E tocopherols in commercial dairy products, owing to its robustness, efficiency, and accuracy. A total of 15 samples, comprising various brands of infant formula, goat milk, cow milk, yogurt, and dairy beverages, were purchased from local supermarkets in Beijing. Representative UPLC-DAD chromatograms of vitamin A and E (α-tocopherol) for five selected samples are shown in Fig. 3, and the measured concentrations along with their relative standard deviations (RSDs) are summarized in Table 3.

Fig. 3.

Typical UPLC-DAD chromatogram of selected real samples: (a) Vitamin A; (b) Vitamin E, including α-, β-, γ-, and δ-tocopherol.

Table 3.

The measured values of Vitamin A and E vitamers in various real samples collected from market, detected using UPLC-DAD method.

Sample No.	Sample information	Vitamin A		Vitamin E
		Value (μg/100 g)	RSD (%)	Value (mg/100 g)	RSD (%)
S1	Milk 1#	72.65	5.77	2.79	1.79
S2	Milk 2#	144.70	1.57	–	–
S3	Milk 3#	130.10	8.11	–	–
S4	Dairy beverage 1#	102.34	4.01	3.72	6.18
S5	Dairy beverage 2#	58.37	2.35	0.84	1.19
S6	Dairy beverage 3#	42.53	4.04	–	–
S7	Dairy beverage 4#	57.69	3.35	–	–
S8	Dairy beverage 5#	14.41	2.58	–	–
S9	Yogurt 1#	149.81	2.41	–	–
S10	Yogurt 2#	47.21	2.63	–	–
S11	Infant formula 1#	883.91	6.11	13.50	6.09
S12	Infant formula 2#	977.09	7.66	20.27	8.48
S13	Infant formula 3#	858.48	3.98	16.77	6.30
S14	Infant formula 4#	617.73	5.52	9.56	4.58
S15	Infant formula 5#	457.36	4.94	9.03	4.37

Vitamin A was detected in all 15 samples, with the highest concentration found in infant formula samples. Notably, α-tocopherol was also present in the one milk, five infant formula, two dairy beverage samples. In contrast, and as expected, no vitamin E was detected in several dairy products, such as, yogurt. These findings confirm that infant formula sample serve as a primary dietary source of these vitamins. To further validate the method, all samples were also analyzed using the UPLC-IDMS/MS method, with results summarized in Table S5. Importantly, the values obtained by the UPLC-DAD method showed good agreement with those from UPLC-IDMS/MS across all sample matrices. This strong correlation further validates the reliability and broad applicability of the QuEChERS-UPLC-DAD method for the simultaneous determination of vitamins A and E in diverse dairy products.

4. Conclusion

A modified QuEChERS-UPLC-DAD method was developed for simultaneous determination of vitamin A and four tocopherols in infant formula. Using isopropanol as a combined saponification and extraction solvent enabled a streamlined one-step pretreatment. The method exhibited excellent performance, including good linearity, low LODs/LOQs, high recoveries, and satisfactory reproducibility, which were further validated by analyzing NIST SRM 1869. Results showed strong agreement with those obtained by UPLC-IDMS/MS, confirming method robustness and accuracy. Successfully applied to various dairy products, this approach provides a reliable and practical tool for the routine quality control and nutritional assessment of fat-soluble vitamins in complex dairy matrices.

CRediT authorship contribution statement

Xiaotong Zhao: Validation, Methodology, Formal analysis, Data curation. Xiaomin Li: Writing – review & editing, Writing – original draft, Resources, Project administration, Methodology, Conceptualization. Shuangxia Luo: Writing – original draft, Validation, Formal analysis, Data curation. Xuanping Zhang: Validation, Investigation, Data curation. Qinghe Zhang: Supervision, Resources, Project administration. Xiuqin Li: Resources, Project administration. Xiong Yin: Writing – review & editing, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.