Determination of the Major Bioactive Component of Silybum marianum in Nutricosmetics by a HPLC Method With Amperometric Detection and UAE Pretreatment

ABSTRACT

Introduction

Nutricosmetics derived from Silybum marianum , known for their anti‐inflammatory and hepatoprotective properties, necessitate accurate quantification of silybin, a key bioactive component.

Objectives

This study aims to develop a novel high‐performance liquid chromatography (HPLC) method with amperometric detection (HPLC‐ECD) for the precise determination of silybin. An ultrasound‐assisted extraction (UAE) procedure is also established for solid sample preparation prior to chromatographic analysis.

Materials and Methods

Chromatographic separation of silybin was performed on a C18 column and using methanol–0.035 M potassium phosphate (pH 4.0) at 1.0 mL min⁻¹ flow rate as mobile phase in gradient mode. The electrochemical detection (ECD) of silybin was carried out on a glassy carbon electrode (GCE) at +1.10 V versus Ag/AgCl. The UAE procedure for silybin extraction from solid samples was performed by 15 min sonication in an ultrasonic bath (80 kHz and 100% power) at room temperature.

Results

Under the optimal chromatographic conditions, silybin diastereoisomers (silybin A and silybin B) can be separated from other S. marianum flavonolignans in less than 20 min, with a detection limit (LOD) of 0.060 mg L⁻¹ and a reproducibility (RSD) of 5%. This method was successfully applied to analyze silymarin‐containing products with recoveries close to 100%.

Conclusions

This work presents the first HPLC method for silybin determination using an amperometric detector with a GCE. The LOD is competitive in comparison with previously published HPLC‐DAD methods. This HPLC‐ECD method allows silybin diastereoisomers identification without interferences of other flavonolignans present in silymarin extracts.

1. Introduction

Cosmetic industry has changed in recent years due to the major concern of the consumers about good physical appearance, fitness, and disease prevention. To meet the actual society requirements, new cosmetic products have been commercialized, nutricosmetics being one of the most used [1, 2]. The term nutricosmetics can be related to the oral consumption of products formulated with natural bioactive ingredients that have beauty and health benefits [2]. Among the different bioactive compounds added to nutricosmetics, polyphenols can be highlighted due to their benefits on human health [2, 3]. Polyphenols can be found in fruits, vegetables, and different plants as Silybum marianum (milk thistle). Milk thistle belongs to the Asteraceae family. It is a key source of unique polyphenols known as flavonolignans [4]. The mixture of flavonolignans extracted from seeds and fruits of S. marianum is known as silymarin, and it is mainly composed by silybin, isosilybin, silychristin, and silydianin [5]. Silybin is the major component (constituting of about the 60%–70%) and responsible of the biological activity. Silybin has beneficial properties such as potent antioxidant, anti‐inflammatory, and antifibrotic activity, and it is a potential candidate for the treatment of cancer and/or chronic liver disease [6, 7, 8]. Silybin also possess the ability of reducing the oxidative damage caused by the solar radiation and environmental toxins due to their antioxidant properties, so a more youthful appearance can be observed when silybin‐based cosmetics are used [9, 10]. Recent studies have also demonstrated the potential antiviral activity of silybin against viruses, such as SARS‐CoV‐2 [11, 12]. These potential properties are related to its structure. Silybin is characterized by the presence of chiral centers, so two diastereoisomers called silybin A and silybin B can be differentiated (Figure 1).

FIGURE 1.

Chemical structure of silybin diastereoisomers.

Due to the differences in the 3D structure of silybin A and silybin B, the stereoisomers present different metabolic profile and biological and chemical activity. Recent papers have evaluated the different biological activities of silybin A and silybin B [13, 14, 15]. García‐Viñuales et al. [14] studied the possible benefits of silybin stereoisomers in diabetes prevention. This action is associated to the potential of silybin A and silybin B to avoid the aggregation of islet amyloid polypeptide in pancreatic β‐cell, silybin B isomer being the most efficient [14]. Dobiasová et al. [15] evaluated the potential of silybin stereoisomers to modulate the transport pumps from P‐glycoprotein, with silybin B being the only compound to act directly on P‐glycoprotein. Despite the differences in activity of the stereoisomers, both have beneficial effects on human health and an increase in the use of S. marianum –based nutricosmetics has been observed in recent years [16]. Thus, the development of analytical methods to determine the major bioactive component of S. marianum is of great interest for nutricosmetic industry and in studies related to the effects of natural products on health. Spectroscopic techniques can be useful for quality control of silymarin‐based nutricosmetics, although a selective determination of the main bioactive ingredient is not possible [17]. As silymarin is composed of a mixture of different flavonolignans with similar spectroscopic properties; using spectroscopic techniques, only information on the total amount of flavonolignans can be obtained. Due to the advantages of electrochemical techniques, such as possibilities for in situ and real‐time analysis, sensitivity, and selectivity, methods based on different voltammetric techniques and using diverse electrodes have been developed for the determination of silybin [18, 19, 20]. Using these electrochemical methods, the total concentration of silybin was calculated, but the signal of each stereoisomer cannot be distinguished. In addition, other electroactive flavonolignans present in silymarin extracts can interfere in silybin determination. Separation techniques are widely used for the determination of total silybin and both diastereoisomers. Methods based on high‐performance liquid chromatography (HPLC) coupled to diode‐array and mass spectrometer detectors are the most used [21, 22, 23, 24]. In fact, the different diastereoisomers of silybin can be separated using conventional reverse stationary phases and methanol or acetonitrile as organic modifiers [21, 24]. The combination of a separation technique as HPLC and a detection mode based on the electrochemical properties (HPLC‐ECD) is a promising alternative for the selective determination of silybin due to the advantages of this type of detection such as good sensitivity and selectivity in the measurement.

Sample treatment procedure is a critical step in the development of analytical methods. These procedures must be simple to ensure its reproducibility and non‐time consuming. Silybin from seeds, plants, or solid nutricosmetics must be extracted prior the analytical measurement, and different extraction procedures based on microwave‐assisted, supercritical CO₂, or ultrasound‐assisted extraction (UAE) have been described [25, 26, 27]. UAE is an extraction procedure widely used due to its simplicity, low cost, and short extraction times with a low consumption of solvents. The aim of this work is to develop a HPLC method coupled to ECD for the quick and selective determination of silybin in several types of nutricosmetics. As far as our knowledge, there are no previous works based on HPLC‐ECD methods for the determination of silybin. An UAE method is also developed to analyze nutricosmetics with different composition and characteristics.

2. Experimental

2.1. Reagents

All reagents used throughout this work were of analytical grade and used as received. Commercial silybin standard (≥ 98%, with a 1:1 diastereomeric ratio) and silymarin standard (≥ 30% silybin) were purchased from Sigma‐Aldrich (USA). Stock solutions of the target analytes at a concentration of 1000 mg L⁻¹ were prepared in methanol and stored at 4°C in the darkness. Standard solutions for HPLC analysis were prepared by diluting the stock solutions in the mobile phase.

Methanol (HPLC‐grade) used as organic modifier was purchased from Scharlab (Spain). Ortho‐phosphoric acid (85%, w/w, Scharlab, Spain), potassium hydroxide (Sigma‐Aldrich, USA), and potassium and sodium phosphate salts (Fluka, Switzerland) were used to prepare buffer solutions at different pH values. Ultrapure water with resistivity ≥ 18.2 MΩ cm at 25°C used throughout this work was obtained using a Milli‐RO‐Milli‐Q water purification system.

2.2. Instruments

Voltammetric measurements were performed using the potentiostat/galvanostat Autolab PGSTAT302N (EcoChemie, Metrohm, Switzerland) controlled by the General‐Purpose Electrochemical System software (GPES, 4.9007 version). The electrochemical cell consisted of a Ag/AgCl/KCl 3 M reference electrode (Metrohm, Switzerland), a platinum counter electrode (Metrohm, Switzerland), and a glassy carbon working electrode (3 mm of diameter, CHI104 model, CH Instruments, USA).

Chromatographic analyses were carried out in a HPLC system (Jasco, Japan) formed by a PU‐2089 Plus Quaternary Gradient Pump, an AS‐2055 Plus Intelligent Sampler, and a CU‐2067 Plus Intelligent Column Oven and a MD‐2010 Plus Multiwavelength Detector. A LC‐Net ll/ADC interface with the Jasco ChromPass chromatography data system was used to control the HPLC system and chromatographic data acquisition and handling. For amperometric detection, a 614 A‐Detector (Metrohm, Switzerland) equipped with a wall‐jet cell (model 656, Metrohm, Switzerland) was used. This cell was equipped with a glassy carbon working electrode (model 6.1204.110 Metrohm, Switzerland) with an area of 7 mm², an Ag/AgCl/KCl 3 M reference electrode (model 6.0727.000 Metrohm, Switzerland), and a gold counter electrode (model 6.530.320 Metrohm, Switzerland). Prior to the HPLC measurements, the working electrode surface was mechanically polished for 60 s with 1 μm diamond paste (Buehler, USA) on a polishing cloth, rinsed with ultrapure water, and sonicated with ethanol to remove any impurity.

The amperometric detection was optimized using a flow injection analysis (FIA) system composed by a peristaltic pump (Miniplus 3 model, Gilson, USA), an injection valve with a 20‐μL sample loop (Rheodyne 7125, Merck, Germany) and the μAUTOLAB II (EcoChemie, Metrohm, Switzerland) potentiostat controlled by the GPES software.

An ultrasonic bath (P30HS model ELMA) and a centrifuge model Universal 16 (Hettich, Germany) were used in solid sample extractions. All pH measurements were performed with a GLP22 pH meter (Crison, Spain).

2.3. Voltammetric Measurements

To evaluate the electrochemical behavior of silybin on glassy carbon electrode (GCE), cyclic voltammograms of 200.0 mg L⁻¹ silybin solutions in 0.10 M phosphate buffer at different pH values in the presence of 20% methanol were registered between −0.50 V and +1.0 V at a scan rate of 100 mV s⁻¹.

To obtain the hydrodynamic voltammograms in the FIA system, a 10.0 mg L⁻¹ silybin solution was injected by triplicate at different potential values between +0.60 V and +1.20 V (vs. Ag/AgCl) and using 50% methanol‐50% 0.035 M phosphate solution at pH 4 as carrier.

2.4. Chromatographic Conditions

The chromatographic separation was carried out on a Kromasil 5 μm C18 (150 mm length × 4.6 mm inner diameter) column (Scharlab, Spain), at 40°C, and using methanol (Solvent A) and 0.035 M potassium phosphate buffer pH 4.0 (Solvent B) at 1.0 mL min⁻¹ in gradient mode. A linear gradient of solvent A from 30% to 60% in 20 min, followed by the decrease in percentage of A to the initial conditions in 1 min was used. To assure a stable base line, the initial conditions were kept during 5 min between injections. The amperometric determination was performed by applying +1.10 V versus Ag/AgCl as detection potential. To carry out the measurements, 20 μL was chosen as the injection volume.

2.5. Samples and Sample Preparation

2.5.1. Description of Samples

The proposed HPLC‐ECD method was applied to different S. marianum commercial preparations acquired from several local supermarkets. Specifically, the content of silybin was determined in four different nutricosmetic preparations: milk thistle tablets, milk thistle capsules, a milk thistle hydroglycerinated extract, and S. marianum seeds.

Milk thistle tablets consist in dry extract of S. marianum seeds (about 60 mg per tablet) and different excipients (corn starch, cellulose, magnesium stearate, and silicon oxide). Milk thistle capsules contain S. marianum crushed seeds (400 mg). The liquid nutricosmetic is a concentrated extract of S. marianum seeds in water and glycerin and contains potassium sorbate as preservative.

2.5.2. Sample Preparation Procedures

For the analysis of the liquid hydroglycerinated extract sample, a simple 1:5 dilution with the mobile phase composed by methanol and 0.035 M potassium phosphate at pH 4.0 (30/70, v/v) was required. Solid nutricosmetic samples were processed prior to analysis: Seeds and tablets were first ground in a grinding mill, and milk thistle capsules were emptied and their content was thoroughly mixed. Silybin was extracted from solid samples using the following UAE procedure: 100.0 mg of the sample (ground tablets and seeds or the content of capsules) was weighed, and 5.00 mL of methanol were added. The mixture was sonicated 15 min in an ultrasonic bath at a frequency of 80 kHz and 100% power at room temperature. Then, the mixture was centrifuged for 5 min at 3500 rpm. The extract was filtered through a 0.45‐μm nylon syringe filter and diluted 1:5 with the mobile phase prior the injection in the chromatographic system. All the analyses were performed in triplicate.

3. Results and Discussion

3.1. Electrochemical Studies

The electrochemical behavior of silybin on GCE, at different pH values, was evaluated by cyclic voltammetry. As can be seen in Figure 2A, the voltammogram shows an anodic peak at +0.547 V, an oxidation wave at +0.95 V, and a cathodic peak at +0.008 V. The anodic signals observed correspond to the oxidation of different phenolic groups presented in silybin structure, so both silybin diastereoisomers presented the same electrochemical response and cannot be differentiated using cyclic voltammetry [28]. Figure 2B presents the effect of the pH value in the potential and current of the anodic peak of silybin. As can be seen, the potential linearly decreases with a slope close to 59 mV/pH as the pH increases in the range between 2.5 and 5.2. This behavior suggested an exchange of the same number of protons and electrons in the oxidation process. For higher pH values, the pronounced decrease of the anodic peak potential can be related to the acid–base properties of silybin (pK_a 6.5–7). According to previous studies [29], the oxidation process of silybin occurs in two steps with the exchange of one proton for each electron transferred. The anodic current also varied with the pH value, the maximum currents being observed at pH values between 4 and 5 (Figure 2B). According to these results, the anodic response of silybin can be useful in the amperometric detection of this compound.

FIGURE 2.

(A) Cyclic voltammograms obtained for a 200.0 mg L⁻¹ of silybin (―) in 0.10 M phosphate solution at pH 4 in the presence of 20% methanol. Response of supporting electrolyte (‐‐‐). (B) Effect of the pH on the first anodic peak potential (▪) and current (●) of silybin.

3.2. Optimization of Amperometric Detection and Chromatographic Conditions

To establish the detection potential, the hydrodynamic voltammogram of silybin was evaluated using the FIA system described in Section 2.3. For this purpose, a 10.0 mg L⁻¹ silybin solution and a carrier composed by methanol and 0.035 M phosphate solution (50/50, v/v) were used. According to the results obtained in the electrochemical studies (Figure 2B), 4 was selected as optimal pH value. As expected, FIA registers show a unique signal corresponding to the sum of both silybin diastereoisomers. Two current waves from +0.60 to +0.95 V and from +1.00 to +1.20 V, respectively, corresponding to the oxidation of the two phenolic groups of silybin can be observed (Figure 3). An increase in the current was observed with the applied potential, up to a potential value of +1.10 V, decreasing at higher potentials. The maximum measurement sensitivity (highest signal‐to‐background ratio) was achieved at +1.10 V, so this value was chosen as working potential.

FIGURE 3.

Hydrodynamic voltammogram of a 10.0 mg L⁻¹ silybin solution obtained using a 50% mixture of methanol and 0.035 M phosphate solution at pH 4 at 1.0 mL min⁻¹ as carrier (n = 3).

As commented before, silymarin is composed by different flavonolignans (isosilybin, silychristin, and silydianin) that must be separated from silybin for its adequate determination. Thus, chromatographic separation is required to adequate silybin quantification. The chromatographic conditions were evaluated using the EC detection at +1.10 V (vs. Ag/AgCl). The separation was carried out under gradient mode using methanol and 0.035 M phosphate solution at pH 4 at 1.0 mL min⁻¹ flow rate as mobile phase. Different percentages of methanol and gradient slopes were investigated injecting a 20.0 mg L⁻¹ silymarin solution. The best resolution and analysis times were obtained using a linear gradient from 30% to 60% of methanol in 20 min, and then the initial percentage was restored in 1 min. Before the next injection, the HPLC system was maintained under the initial conditions (30% of methanol) for 5 min. Under these chromatographic conditions, the separation of silymarin components was achieved. The effect of the concentration and pH value of the electrolyte solution in the mobile phase on the chromatographic separation was evaluated. Concentrations ranging between 0.01 and 0.05 M and pH values from 1.65 and 5.01 were studied. No significant differences in the analysis times and resolutions were observed using the different concentrations and pH values of the electrolyte solution. Thus, 0.035 M phosphate solution at pH 4 was selected as supporting electrolyte in the mobile phase.

Figure 4 shows the chromatograms obtained under the optimal chromatographic and detection conditions for silymarin (Figure 4A) and silybin (Figure 4B) solutions. As can be seen in Figure 4A, six well‐resolved signals due to silymarin flavonolignans are differentiated and do not interfere with silybin stereoisomers signals, indicating the suitability of the chromatographic method to detect the major component of silymarin with good resolutions and analysis times. According to the elution order proposed for silymarin flavonolignans in the literature [21, 23], peaks in the chromatograms in Figure 4A,B can be assigned to silychristin (Peak 1, at 11.0 ± 0.4 min), silydianin (Peak 2, at 12.3 ± 0.3 min), silybin A (Peak 3, at 16.5 ± 0.2 min), silybin B (Peak 4, at 17.2 ± 0.2 min), isosilybin A (Peak 5, at 18.9 ± 0.1 min), and isosilybin B (Peak 6, at 19.4 ± 0.1 min). The reproducibility of the retention times for all flavonolignans, expressed in terms of relative standard deviation (RSD), was lower than 1.5%. Resolution of silybin diastereoisomers (Figure 4B) was higher than 1.5 under the optimal conditions.

FIGURE 4.

Chromatograms obtained using methanol and 0.035 M phosphate solution at pH 4 at 1.0 mL min⁻¹ as mobile phase in gradient mode and applying +1.10 V versus Ag/AgCl as detection potential. Chromatograms of (A) 20.0 mg L⁻¹ silymarin and (B) 10.0 mg L⁻¹ commercial silybin. Peaks: 1: silychristin, 2: silydianin, 3: silybin A, 4: silybin B, 5: isosilybin A, 6: isosilybin B.

3.3. Analytical Parameters of the HPLC‐ECD Method for Silybin Determination

The analytical parameters of the proposed HPLC‐ECD method for the determination of silybin were calculated. Considering that, in general, nutricosmetics indicate in the label the total amount of the bioactive compound, in this work, the sum of the areas of silybin diastereoisomers, silybin A and silybin B, was used as analytical signal. The calibration plot of silybin was obtained at concentrations ranging between 0.10 and 50.0 mg L⁻¹. Good relationship with a correlation coefficient of 0.9981 between peak areas and silybin up to a concentration of 38.0 mg L⁻¹ (upper limit of the linear range) was obtained.

Table 1 shows the analytical features of the HPLC‐ECD method to determine silybin. The limits of detection (LOD) and quantification (LOQ) were calculated as the signal‐to‐noise ratio of 3:1 and 10:1, respectively, using the standard deviation of the sum of peak areas of a standard solution of silybin diastereoisomers at a concentration of 0.10 mg L⁻¹. The repeatability of the method was determined by successive injections in the same day (n = 4) of a 5.00 mg L⁻¹ silybin standard solution. Reproducibility was calculated by injecting 5.00 mg L⁻¹ silybin standard solution in four different days (n = 4). As can be seen in Table 1, the HPLC‐ECD method shows good precision in terms of reproducibility and repeatability.

TABLE 1.

Analytical properties of the HPLC‐ECD method to determine silybin.

Sensitivity (mV min L mg−1)	0.241 ± 0.006
Upper limit of the linear range (mg L−1)	38
LOD (mg L−1)	0.060
LOQ (mg L−1)	0.20
Repeatability (%, RSD)	3.4
Reproducibility (%, RSD)	5.0

Some of the analytical properties of the developed HPLC‐ECD method are compared in Table 2 with those reported in previous works. As can be seen, in general the LOD of the proposed method with amperometric detection is better than those obtained in HPLC methods coupled to diode array (DA) or even MS detection. Therefore, the electrochemical detector can be considered a good alternative to other detection systems such as MS, as it presents good LODs and is cheaper and easier to use and data evaluation is simpler.

TABLE 2.

Comparison of some of the analytical properties of the developed HPLC‐ECD method for the analysis of nutricosmetics with those obtained in previous works.

Method	LOD (mg L−1)	Linear range (mg L−1)	Sample	Total analysis time a (min)	Reference
HPLC‐DAD	1.4 (silybin A) 1.0 (silybin B)	Not indicated	Silymarin preparations	135	[21]
HPLC‐MS/MS	0.00017–0.0003	0.0005–1	Human plasma, urine, and breast tissue	171	[22]
HPLC‐DAD	0.02–0.03	0.2–100	Fresh and commercial turmeric	125	[24]
LC‐MS	Not indicated	Not indicated	Milk thistle fruits	120	[27]
ce‐UV	1.1 (silybin A) 1.3 (silybin B)	2.5–50	Dietary supplements	55	[30]
HPLC‐DAD	3.1	150–350	Pharmaceutical formulations	55	[31]
LC‐MS	0.010	0.050–5.0	Food supplements	194.5	[32]
HPLC‐DAD	0.08	Not indicated	Milk thistle extracts and seeds	100–170	[33]
HPLC‐ECD	0.060	0.20–38	Silybum marianum –based nutricosmetics	40	This work

^a Extraction procedure + HPLC separation.

3.4. Analysis of S. marianum Commercial Preparations

Prior the analysis of real samples, silybin was extracted from the commercial preparations using an UAE method. The effect of the extraction time and the sample weight‐to‐volume of extraction solvent ratio in the extraction efficiency were studied using methanol as extraction solvent. The commercial preparation denoted as milk thistle tablets was used as sample to obtain the best conditions for silybin extraction.

Different extraction times between 15 and 60 min were assayed. For this purpose, 10 tables (0.5 g per tablet) were grinded in a mill. Portions of 100.0 mg of the ground sample were weighed, and 5.00 mL of methanol was added and sonicated at 80 kHz of frequency and 100% of power at the different studied times. The different extracts obtained after centrifugation were injected by triplicate in the HPLC‐ECD system. The results were expressed as amount of total silybin per tablet. No significant differences in the amount of silybin were observed at the studied extraction times. Thus, 15 min was selected as the extraction time.

Sample‐to‐volume ratios of 100.0 mg/5.00 mL, 200.0 mg/5.00 mL, and 500.0 mg/5.00 mL were studied. The UAE procedure previously indicated was applied using 15 min as extraction time. A diminution in amount of silybin was observed as the sample‐to‐volume ratio increases, indicating worse extraction efficiency; thus, 100.0 mg of sample/5.00 mL of methanol was chosen as the sample‐to‐volume ratio for the extraction.

The content of silybin in this nutricosmetic is not indicated by the manufacturer, and reference materials were not available; thus, to establish the trueness of the UAE method, the recovery was evaluated [34, 35] using the milk thistle tablets spiked with silybin. For this purpose, 0.100 mg of silybin was added to an aliquot of 100.0 mg of ground tablets, and then they were thoroughly mixed prior to carry out the UAE extraction procedure. The fortified extract obtained was analyzed using the proposed HPLC‐ECD method. A recovery of 95 ± 1% was obtained, indicating that the UAE method is adequate to extract silybin from nutricosmetics. The here‐proposed UAE followed by HPLC‐ECD method was compared to those previously described by other authors. As can be seen in Table 2, the total analysis time including the sample treatment procedure and the chromatographic separation of the proposed work is lower than those obtained in previous works analyzing samples with similar matrices [21, 30, 31, 32, 33]. The total analysis time in most cases is higher than 1 h compared to the 40 min required for the proposed method in this work.

Four different samples were analyzed using the proposed HPLC‐ECD method: milk thistle tablets, milk thistle capsules, a milk thistle hydroglycerinated extract, and seeds. As indicated in Section 2.5.2, prior the analysis, the liquid sample was 1:5 diluted with the mobile phase. Silymarin major component was extracted from the solid samples (tablets, capsules, and seeds) using the developed UAE method and a dilution 1:5 with the mobile phase was required before the HPLC determination. Figure 5 shows the chromatograms obtained for the analyzed samples. As can be seen, silybin B is in a higher proportion in all the analyzed samples, being the silybin A and B ratios similar to that found in this type of samples [21]. According to that obtained in Figure 4A, the other signals observed correspond to the other flavonolignans present in silymarin extracts.

FIGURE 5.

Chromatograms of milk thistle (A) tablets (B) capsules, (C) hydroglycerinated extract, and (D) seeds obtained in the same separation and detection conditions as Figure 3B. Peak 3: silybin A; Peak 4: silybin B.

The amount of silybin in the analyzed samples is presented in Table 3. For verification purposes, the results obtained with ECD were compared with those calculated using the DAD. As can be seen in Table 3, there are no significant differences in the amount of silybin calculated using the ECD and DAD. To verify the adequacy of the proposed method, the extracts obtained after applying the UAE procedure were fortified with silybin standard solution at a concentration ranging between the 50% and 100% of the measured concentration in the sample. The extracts were analyzed by the HPLC‐ECD method, and the recovery values were calculated. Recovery values between 95 ± 1% and 107 ± 7% were obtained for all the studied samples. These results demonstrate that the compounds and/or excipients present in silymarin‐based products do not interfere in the determination of silybin, demonstrating the specificity of the proposed method.

TABLE 3.

Amount of silybin in the analyzed samples (n = 3).

		Milk thistle tablets (mg per tablet)	Milk thistle capsules (mg per capsule)	Milk thistle hg extract a (mg L−1)	Milk thistle seeds (mg g−1)
Silybin	HPLC‐ECD	0.44 ± 0.03	0.42 ± 0.03	11.8 ± 0.7	7.5 ± 0.5
	HPLC‐DAD	0.46 ± 0.02	0.43 ± 0.03	10.9 ± 0.6	8.4 ± 0.6

^a Hydroglycerinated extract.

4. Conclusions

This work presents the first HPLC‐ECD method for the determination of silybin using a GCE as working electrode. The potential of the electrochemical detector above other detection modes as DA or even MS has been demonstrated in this work, obtaining a better LOD and sensitivity for the determination of silybin. A good specificity is observed using this detector, and no interferences due to other flavonolignans present in silymarin extracts were observed. In addition, it is highlighted the simplicity of ECD over MS detection mode. Under the developed separation and detection conditions, the milk thistle flavonolignans (silychristin, silydianin, isosilybin A, and isosilybin B) can be also determined. Using the here‐proposed UAE procedure, silybin can be extracted in a short time from nutricosmetics with different characteristics. A significant reduction in the total analysis time including extraction procedure and HPLC measurement is observed using the developed UAE followed by HPLC‐ECD method, achieving the extraction and separation of silymarin flavonolignans in 40 min, whereas the total analysis time in previous works is higher or close to 1 h. Recovery values close to 100% indicate the adequacy and specificity of the method for the determination of silybin in nutricosmetics. The proposed HPLC‐ECD can be also used to determine silybin diastereoisomers because the separation and detection of both silybin A and silybin B occurred under the developed conditions. This is of a great interest in the biological and clinical field due to the different chemical activity and metabolic profile of silybin stereoisomers.