The measure of cholesterol removal from homogenized milk is affected by the conditions of mixing and the concentration of β-CD. The application of β-CD does not affect the milk color characteristics.

Commercial milk (3.5% declared fat content) bought in a local market.

β-cyclodextrin (Wacker Chemie AG, Burghausen, Germany, ≥95.0%, HPLC).

Cholesterol (Sigma-Aldrich, Merck KGaA, Darmstadt, Germany, Sigma Grade, ≥99%).

Chloroform (Centralchem s.r.o., Bratislava, Slovakia, p.a.). n-hexane (Centralchem s.r.o., Bratislava, Slovakia, p.a.).

Ethanol (Centralchem s.r.o., Bratislava, Slovakia, p.a.).

Sodium sulfate anhydrous (Centralchem s.r.o., Bratislava, Slovakia, p.a.).

Potassium hydroxide (Mikrochem, Pezinok, Slovakia, p.a.).

Methanol (Fisher Chemical, Loughborough, UK, HPLC grade).

Acetonitrile (Fisher Chemical, Loughborough, UK, HPLC grade).

HPLC system (Agilent Technologies 1260 Infinity, Santa Clara, CA, USA).

Vacuum degasser (Agilent Technologies, Santa Clara, CA, USA).

Quarterly pump (Agilent Technologies, Santa Clara, CA, USA).

Autosampler (Agilent Technologies, Santa Clara, CA, USA).

UV-DAD detector (Agilent Technologies, Santa Clara, CA, USA).

Chromatography column (Zorbax Eclipse Plus C₁₈, 2.1x50 mm, 5 μm particle size, Agilent, Santa Clara, CA, USA).

Guard column (Zorbax SB-C₁₈, 4.6x12.5 mm, 5 μm particle size, Agilent, Santa Clara, CA, USA).

Rotary vacuum evaporator (Heidolph, Germany).

Centrifuge (Hettich Zentrifugen, Germany)

Magnetic stirrer (Arex-6 Connect Pro, Velp Scientifica, Italy).

Analytical balance (Sartorius, Goettingen, Germany)

UV-VIS spectrophotometer (Cary 300, Agilent Technologies, Santa Clara, CA, USA).

Results are expressed as mean ±standard deviation or as a percentage. Statistical analysis was performed using the XLSTAT tool of Microsoft Excel 365 (version 2012, Microsoft, USA). The obtained data from the cholesterol content analysis and the analysis of color were subjected to one-way analysis of variance (ANOVA) and the Student ́s test, and the values were considered significantly different when p <0.05. From the ANOVA and Student´s test, it was able to determine if there are some significant differences between the results obtained by the analysis of the effect of mixing conditions (speed, time, and temperature), CD concentration, and color changes.

The cholesterol removal was significantly affected by mixing speed. It was noticed significant (p <0.05) differences between the results obtained at 480 to 600 rpm, and 720, 840, and 960 rpm. The results were not significantly different at 720, 840, and 960 rpm, suggested that the optimum mixing speed begins at 720 rpm. However, the highest removal efficiency (73.3%) was observed at 840 rpm thus this speed was set as the most suitable one, as shown in Figure 1A. The results are similar to other studies. Kwak, Nam and Ahn (2001) stirred homogenized milk with β-CD at 800 rpm and reached cholesterol removal up to 78.2% (1.0% β-CD). Lee, Ahn and Kwak (1999) reported that mixing speed at 800 rpm is efficient to remove 94.0% of the cholesterol in commercial milk by β-CD (1.0%), also supported by Lee et al. (2007). According to Maskooki et al. (2013), by the application of mixing speed at 700 rpm, it is possible to remove 67.7% of the cholesterol in homogenized milk by β-CD in the same concentration. Lower mixing speed (530 rpm) is reported by Gianni et al. (2020) with the cholesterol removal rate varied from 33 to 89%. According to Kim, Ahn and Kwak (2004), the mixing speed is not an important factor in cholesterol removal in milk by crosslinked β-CD as it was not noticed a significant difference with various mixing speeds. The mixing speed is also influenced by the type of sample. E.g., higher speed is required in cream treatment as Jung, Ko and Kwak (2013) used 1400 rpm, Ahn and Kwak (1999) 800 – 1600 rpm, or Shim, Ahn and Kwak (2003) 400 – 1200 rpm. Tahir, Bokhari and Adnan (2015) extracted the cholesterol from ghee by mixing with β-CD immobilized on glass beads at 170 rpm.

A significant difference (p <0.05) was noticed between the mixing time of 5 min and 10, 15, and 20 min. The highest measure of cholesterol removal was observed after 20 min of mixing (79.8%) but similar values were monitored in 10 and 15 min (79.3 and 79.2%, respectively). The lowest value of 73.0% was noticed at a mixing time of 5 min (Figure 1B). As found, a mixing time over 10 min is enough for sufficient removal. In comparison to this, slightly different results were reported by Lee, Ahn and Kwak (1999). Based on their results, the cholesterol removal was not affected by a mixing time of 5 to 20 min, but it was noticed a reduction at 25 min of mixing. A similar trend was also mentioned by Alonso et al. (2009). A slight reduction was observed when the mixing time increased to >20 min for 0.8 and 1.0% β-CD. The above results suggest that at longer mixing times, the inclusive complex between β-CD and cholesterol could be unstable. Therefore, from the other studies, it can be concluded that the mixing time between 10 and 20 min is the most suitable for the sufficient removal of cholesterol from milk. Kwak et al. (2003) reported that by the mixing time of 10 min it is possible to remove 79.3% of cholesterol content in milk by β-CD addition (1.0%), which is similar to our study. Lee, Ganesan and Kwak (2012) used 10 min mixing of milk with crosslinked β-CD and reached the removal of cholesterol up to 93%. Higher mixing times are reported in the case of cream. Ahn and Kwak (1999) achieved the highest removal of cholesterol in cream by the mixing time of 30 min. This time was also used in other studies (Shim, Ahn and Kwak, 2003; Jeon et al., 2012; Galante et al., 2017).

Besides the importance of mixing speed and duration, temperature is reported as another factor affecting the measure of cholesterol removal. According to Matencio et al. (2020), the most important factors to consider in the mechanism of CD/guest reaction are the structure of the guest molecule, solvent, and temperature. The various results are noticed among published reports. According to this study (Figure 1C), no significant differences (p <0.05) between the mixing temperatures of 25, 40, and 50 °C, were observed, when the measure of cholesterol removal varied at 79.3, 78.2, and 77.9%, respectively. Similar results were mentioned by Lee, Ahn and Kwak (1999) as they also found no differences in cholesterol removal at different temperatures but suggested that mixing at 10 and 25 °C tends to increase cholesterol removal compared to lower values of temperature. A different mixing temperature was used in studies by Alonso et al. (2009), Alonso et al. (2018), and Alonso, Calvo and Fontecha (2019). According to them, the mixing in at 4 °C is sufficient for cholesterol removal up to 95%, which is effective in maintaining the quality of milk. Maskooki et al. (2013) reported that the mixing temperature of 20 °C is more suitable than 8 °C because the aggregations of milk lipoproteins at higher temperatures are lower, and consequently there is a better chance for the entrapment of cholesterol molecules in β-CD cavities.

The effect of milk lipoproteins aggregation may play a greater role by the processing of cream as the cream is mixing with β-CD usually at temperatures between 40 to 50 °C. E.g., Shim, Ahn and Kwak, 2003, Ha et al. (2010), or Lee, Ganesan and Kwak (2012) used the temperature of mixing at 40 °C and achieved effective removal of cholesterol content in cream.

After the optimization of mixing conditions, the influence of β-CD concentration was also studied, as shown in Table 1. The chromatographic records of the analysis of cholesterol in control milk without the addition of β-CD and with 2.0% β-CD addition are shown in Figure 2. As follows from Table 1, the lowest value was observed at 0.5% of β- CD (6.2%) and the highest at 2.0% (98.1%). A slight decrease was noticed at 2.5% of β-CD (97.4%). The results were significantly (p <0.05) different at 0.5 and 1.0% of β- CD addition and did not differ at the concentrations of 1.5, 2.0, and 2.5%. Therefore, the optimal β-CD concentration was thus determined at 1.5%. Similar results are reported by Kwak, Nam and Ahn (2001). According to them, a high increase of cholesterol removal was also found between 0.5 and 1.0% β-CD addition. They suggested that 1.0% of β-CD addition may be effective enough for cholesterol removal. However, lower optimal concentrations are published by other authors. Lee, Ahn and Kwak (1999) noticed that the least efficiency of removal showed the addition of β-CD at 2.0%. According to Alonso et al. (2009), no significant differences in cholesterol reduction were observed at concentrations of 0.6, 0.8, and 1.0%. According to Maskooki et al. (2013), the measure of cholesterol adsorption was reduced at 1.5% β-CD due to excess of β- CD, which could compete with itself to bind to cholesterol molecules, and consequently reduced the cholesterol adsorption effectivity (Lee, Ahn and Kwak, 1999; Maskooki et al., 2013). Besides that, from the published results, it can be stated that the variation of the optimal β- CD concentrations may be caused by different operational conditions, e.g., mixing temperature or mixing and centrifugal speeds. The β-CD concentration depends also on the fat/cholesterol content of food items. E.g., higher additions are required by the processing of cream. Shim, Ahn and Kwak (2003) used 10% of β-CD in the extraction of cholesterol in cream while Dias et al. (2010) reported that 20% of β-CD addition showed the best removal rate of cholesterol in commercial butter.

Since the color characteristics of treated milk are scarce, these were studied by analyzing measured values in CIEL*a*b* color space. As follows from Table 1, treatment with β-CD did not affect the lightness values (L*), which is in accordance with results reported by Gianni et al. (2020). Slight differences were noticed in a* coordinate values between the control sample and the sample treated with 1% of β-CD. The significant differences (p <0.05) were shown between the control and 1.5, 2, and 2.5% of β-CD in b* coordinate values. In previous studies, there is no evidence about the influence of β-CD on the yellowness factor of milk. Despite higher values of this coordinate, it was not noticed any significant changes in the visual appearance of treated milk, which was also proven by the calculation of ΔE values. According to Bhatia et al. (2019), the process of cholesterol removal by β-CD does not affect the typical color of cow ghee so it has no impact on the visual consumer perception of the product.