Development of a Sterilized Mixed Herbal Drink Product

INTRODUCTION
As people become increasingly concerned about their health, and aware of the potential role of diet and dietary supplements upon health, then so herbs and
herbal extracts in particular have become more popular due to their reported health benefits. Amongst Thai people herbal extract drinks are one of the
commonly consumed herbal products. Polyphenolic compounds, like flavonoids, are the principal bioactive compounds found in many herbs and provide a strong antioxidant activity (Katalinic et al., 2006). However, the antioxidant activity and total phenolic compounds change during the processing stages and are usually depleted when they encounter heat (Roy et al., 2007) and a relatively strong oxidizing potential. Potassium is one of the major important dietary minerals, and has protective effects against hypertension and kidney stone formation (US Department of Health and Human Services and US Department of Agriculture, 2005). Like most minerals, potassium has a high thermal resistance, so the thermal process does not significantly decrease the amount of potassium in food products (Rojas-Gonzalez et al., 2006).
Imperata cylindrica (L.) P. Beauv. [IC] and Orthosiphon aristatus Miq. [OA] are well-known medical herbal plants because of their diuretic effects (Dat
et al., 1992 and Jiratchariyakul, 1991), whilst Murdannia loriformia (Hassk.) Rolla Rao et Kammathy [ML] and Hedyotis corymbosa Lamk. [HC] are both
relatively well known for their antioxidant activities and high toxicity to cancer cells (Jiratchariyakul et al., 2006; Noiarsa et al., 2008).
In Thailand these herbs are consumed by some cancer patients, and especially by those seeking alternative or folklore medicine, due to their associated anti-cancer therapeutic health claims. However, access to these herbs and the required preparation steps are not easy, limiting the product availability and also
potentially challenging its efficacy through batch to batch variation in the actual bioactivity of the extracts. Therefore, this study aimed to develop a sterilized
mixed herbal drink product formula and process, derived from these four herbs, to make it easier to purchase a consistent bioproduct in a palatable format.

MATERIALS AND METHODS
2.1 Material Preparation
The four dried herbs; IC, ML, HC, and OA, were purchased from a market in China Town, Bangkok, Thailand, and kept in a glass jar at room temperature. The herbs obtained were in several forms as follows: IC was stems about 2-cm in length, ML was leaves about 3-4 cm in length, HC was long stem, and OA was stems and leaves about 1-cm in length.
2.2 Effect of Infusion Time and Temperature on Herbal Infusion
From our preliminary study each infusion was obtained by extracting the herbs with water, at a herb concentration of 2% (w/v), at 70o, 80o or 90oC in a
controlled temperature water bath for 30, 45 or and 60 minutes. The infusions were then filtered through a tea filter and cooled. The net antioxidant activity was determined by the DPPH assay as described (Sakanaka, Tachibana, and Okada, 2005), whilst the total phenolic compound contents and potassium levels were measured as reported by Nawaz et al. (2006) and the AOAC guidelines (2006), respectively. All analyses and experimental treatments were run in triplicate. The optimum condition was determined by response surface method (RSM) as
outlined by Myers and Montgomery (1995).
2.3 Development of Mixed Herbal Drink Formula
2.3.1 Evaluation of the bitterness of the herbal infusion
Each herbal infusion was prepared via the more optimal condition determined as per section 2.2. The samples were sweetened with 6% (w/w) sucrose before the sensory evaluation for bitterness using 30 panelists and a ranking test.
2.3.2 Determination of the formula of mixed herbal drink.
The mildest infusion evaluated from the step 2.3.1 above was fixed at 10% (w/w) of the mixed infusion. The ratio of the three most bitter infusions was set using a linear simplex lattice design, with a fixed level of sucrose at 6% (w/w) beingadded to all the mixtures. Each formula were served at room temperature and evaluated by 30 test panelists using a 9-point hedonic scale. The optimum combination of herbal infusions was determined from the RSM. Next, the
appropriate amount of the mildest infusion in the mixed drink was varied at 5, 10 15 and 20% (w/w), and sensory evaluation of the samples at room temperature was performed as above to select the appropriate level of the mildest infusion.
2.3.3 Determination of the sugar level in mixed herbal drink.
Sucrose, as white table sugar, was dissolved in the herbal drink to a final concentration of 5, 6, 7 and 8% (w/w) of finished herbal infusion, and then
evaluated by the same 30 test panelists using a 9-point hedonic scale as above. The mixed herbal drink was prepared according to the formula and conditions
evaluated in the previous steps above. The drink was sterilized at 130o, 135o and 140oC for 3, 4 or 5 minutes using a UHT/HTST processing system
(Microthermics Bantam-DH, U.S.A.) to obtain the F0 value, and filled into 250-ml polyethylene terephthalate bottles in a ultra-clean fill hood (Microthermics Clean Fill Hood, U.S.A.). The level of antioxidant activity, total phenolic compounds and potassium, and the total plate count (3M PetrifilmTM) and Clostridium botulinum (US Food and Drug Administration, 1992) levels were all analyzed. All analyses were run in triplicates. The optimum heating condition was determined from the RSM.
2.5 Statistical Analysis
Statistical analysis was carried out using SPSS 15.0 for Windows. The Duncan’s new multiple range test was used to compare the difference between means. The optimum conditions were determined by RSM using Design Expert 7.0.

RESULTS AND DISCUSSION
3.1 Effect of Infusion Time and Temperature on Herbal Infusion
Nine conditions of four herbal infusions were analyzed and are summarized in Table 1. Increasing the infusion time and temperature decreased the antioxidant
activity (p < 0.05). This may be because most of the relevant bioactive compounds are relatively unstable to heat and or mild oxidation (Polydera et al.,

In contrast, the total phenolic compounds increased with as either the infusion time or the temperature of the extraction process was increased. The
extract from HC had the highest amount of phenolic compounds, followed by OA, ML and IC, respectively. Indeed, the level of phenolic compounds in HC is
significantly more (2.6 to 6.1 fold higher) than the other three herbal extracts. That the total level of phenolic compounds was mainly affected by extraction time is consistent with the notion that they would leach out from the tissue during soaking (Katalinic et al., 2006), and that this was more marked for the total
phenolics than potassium due to the weaker hydrophilic nature of the former. The potentially contradictory results of decreasing antioxidant activity with increasing total phenolic compounds over increased extraction time and temperature, given that phenolic compounds are themselves antioxidants, may in
part be due to the method of determination. The DPPH assay determines free antioxidants in the samples and is sensitive to various antioxidants having fast or
intermediate reaction kinetics, whilst the Folin-Ciocalteau method determines both free and bound phenolic compounds and is sensitive to a wide range of
substrates being easily oxidized. Some phenolic compounds, such as gallic and tannic acids, have slow reaction kinetics, so they cannot be reliably detected in a DPPH assay. In addition, some phenolic antioxidants, such as phenol and coumaric acid, react strongly with the Folin-Ciocalteau reagent but do not react
with DPPH. Finally, other compounds with an absorbance at 517 nm may interfere with the DPPH assay and lead to an underestimation of the actual
antioxidant activity, because the DPPH method determines the free radicals left after being oxidized with antioxidant (Yang et al., 2007). In accord to the total phenolics and diffusion mentioned above, the higher extract infusion time and temperature allowed potassium to diffuse from the herbal tissue into the fluid (Katalinic et al., 2006). In general, potassium as a mineral has a high thermal resistance (Rojas-Gonzalez et al., 2006). The correlations between each chemical constituent and their infusion conditions are shown for each herb in Table 2. From RSM, the optimum extraction conditions giving the high amount of
chemical constituents were found to be the extraction temperature and time of 71o - 80oC and 35 - 47 minutes (Table 3). For the practical reason of suitability for industrial preparation, the average temperature and time of 75oC and 40 minutes were chosen for the next step.

3.2 Formulation of Mixed Herbal Drink
3.2.1 Evaluation of the bitterness of the infusions
The sensory test for bitterness of the herb infusion ranked the IC extract based drink as the mildest being hardly bitter at all, with ML and OA essentially equally slightly bitter. In contrast, the phenolic compound rich HC extract based drink was significantly the most bitter (p ≤ 0.05) (Table 4). Table 4 Average bitterness score of herbal drinks with 6% (w/w) sugar added Herbal Drink Average Score Imperata cylindrica (L.) P. Beauv. (IC) 1.1 + 1.3 a
Murdannia loriformia (Hassk.) Rolla Rao et Kammathy (ML) 3.8 + 2.6 b Hedyotis corymbosa Lamk. (HC) 9.1 + 1.1 c Orthosiphon aristatus Miq. (OA) 4.0 + 2.4 b a- c Means with different lowercase superscript letters are significantly different (p < 0.05) Scale ranges from 0 meaning no bitter taste to 10 meaning extremely bitter.
3.2.2 Determination of the formula of the mixed herbal drinks
Since the IC extract based drink had the mildest taste, the amount of this infusion was fixed at 10% (w/w). The remaining 90% (w/w) was made up of a mixture of the other three more bitter infusions which were varied using a simplex design (Table 5), including sucrose to a constant final level of 6% (w/w). Of the tested compositions, only formulae 4 scored above a liking score index of 6. The relationship of these three herbal infusions with the liking score is as shown in equation (1). score = 4.8A + 2.3B + 5.6C + 32AB - 3.3AC - 32BC :R2 = 0.8868 (1) where A, B and C represent the %ML, HC and OA drink, respectively Using RSM, with the criteria of (i) the liking score to be higher than 5, (ii) a maximal amount of HC (highest total phenolic compounds), and (iii) the least amount of ML (most expensive herb), the optimum combination of the three more bitter infusions was determined to be 40% (w/w) ML, 31% (w/w) HC and 19% (w/w) OA for the total 90% (w/w) of the bitter infusions part. After varying the ratio of the mildest infusion, (IC) and keeping the proportion of the more bitter infusions constant, the result showed that the mixed herbal drink containing 15% (w/w) IC infusion and 85% (w/w) three herbal mixture got the highest score (Table 6).

Table 7 Effect of sugar level on average liking score of mixed herbal drinks Formula Herbal Infusion Sugar Liking score
1 95 5 3.9 + 1.4 a
2 94 6 5.7 + 1.4 c
3 93 7 6.2 + 0.6 d
4 92 8 4.9 + 1.0 b
a- d Means with different lowercase superscript letters are significantly different (p ≤ 0.05) Scale ranges from 0 meaning strongly dislike the taste to 10 meaning extremely like the taste.
3.3 Determination of Sterilization Conditions
For both prolonged shelf life and customer safety, making the product obtainable and affordable, it would be required to sterilize it, and preferably by the
economically viable rapid heat-sterilization process. However, given the thermal lability of the bioactive components, a least harsh but suitable heat sterilization
was investigated by using three temperatures (130, 135 and 140 oC), each for three F0 durations (3, 4 and 5 minutes) and then assaying the products for sterility and for residual antioxidative and total phenolics, amongst other components. The results are summarized in Table 8, where the antioxidant activity especially, but also the phenolic compounds, decreased as the duration of heating (value of F0) increased from three to four or five minutes. This is as expected given the thermal lability, and thus destruction, of some phenolic compounds and antioxidant activities (Polydera et al., 2004; Roy et al., 2007; Xu et al., 2007). In contrast, the sterilization temperature used did not significantly affect either the amount of phenolic compounds or the antioxidant activity because time was changed with temperature to obtain the same F0 value. As expected, the level of potassium was not affected by the sterilization temperature or time, given the thermal resistance and low volatility of this metal ion (Rojaz-Gonzalez et al., 2006), and so acts more as an internal process control for concentration of extracts. Importantly, the microbiological tests revealed that no Clostridium botulinum or clonally cultivatable microbes (total plate count) were detected in all the sterilization samples, in contrast to the unsterilized samples (260+ 30 cfu/ml), confirming that they are all sufficient. Because heat denaturation can change the taste of products, the different heat treated samples were assessed for taste using the tasting panel as before. From the liking score (Table 8), the sample heated at 130 oC with a F0 of 5 minutes had the highest preference score, but a rather low anti-oxidant and total phenolic compounds content, compared to the next most liked sample (135oC and F0 of 3 minutes). At the other extreme, the sample heated at 140oC and F0 of 4 minutes had the lowest satisfactory taste preference. From the overlay plot (not shown, but see Table 8) of antioxidant activity, total phenolic compounds, and liking score, the optimum sterilization condition was found to be heating at 135oC to obtain an F0 of 3 minutes. After this heat- sterilization the mixed herbal drink had an antioxidant activity of 9.45 mg Trolox eq/ml, total phenolic compounds content of 43.25 mg gallic acid eq/ml, and an average liking score of 5.7. So from Table 3, the predicted formulae as stated compared to actual unautoclaved (Table 8) are 12.27 vs 10.87 mg Trolox eq/ml, 47.34 vs. 45.1 mg gallic acid eq/ml and 338.9 vs 330.6 mg/l which compares relatively well, just a bit higher in all three parameters.

REFERENCES
AOAC. 2006. Official Methods of Analysis of AOAC International. 18th ed. Maryland: AOAC International.
Dat, D.D., Ham, N.N., Khac, D.H., Lam, N.T., Son, P.T., van Dau, N., Grabe, M.,
Johansson, R., Lindgren, G., and Stjernström, N.E. 1992. Studies on the individual and combined diuretic effects of four Vietnamese traditional herbal
remedies (Zea mays, Imperata cylindrica, Plantago major and Orthosiphon stamineus). Journal of Ethnopharmacology. 36, 225-231.
Jiratchariyakul, W. 1991. Medicine and Natural Products. 1st ed. Bangkok:
Department of Pharmacognosy, Faculty of Pharmacy, Mahidol University.
Jiratchariyakul, W., Vongsakul, M., Sunthornsak, L., Moongkamdi, P., Narintorn, A., Somanabandhu, A., Okabe, H., and Frahm, A.W. 2006. Immunomodulatory effect and quatitation of a cytotoxic glycosphingolipid from Murdannia loriformis. Journal of Natural Medicines. 60, 210-216.
Katalinic, V., Milos, M., Kulistic, T., and Jukic, M. 2006. Screening of 70 medicinal plant extracts for antioxidant capacity and total phenols. Food
Chemistry. 94, 550-557.
Myers, R.H. and Montgomery, D.C. 1995. Response surface methodology. John Willey and Sons, Inc.
Nawaz, H., Shi, J., Mittal, G. S., and Kakuda, Y. 2006. Extraction of polyphenols from grape seeds and concentration by ultrafiltration. Separation
and Purification Technology. 48, 176-181.
Noiarsa, P., Ruchirawat, S., Otsuka, H., and Kanchanapoom, T. 2008. Chemical constituents and concentration by ultrafiltration. Separation and Purification
Technology. 48, 176-181.
Polydera, A.C., Stoforos, N.G., and Taoukis, P.S. 2004. The effect of storage on the antioxidant activity of reconstituted orange juice which had been
pasteurized by high pressure or heat. International Journal of Food Science and Technology. 39, 783-791.
Rojas-Gonzalez, J.A., Avallone, S., Brat, P., Trystram, G., and Bohuon, P. 2006. Effect of deep-fat frying on ascorbic acid, carotenoids and potassium contents of plantain cylinders. International Journal of Food Sciences and Nutrition. 57, 123-136.
Roy, M.K., Takenaka, M., Isobe, S., and Tsushida, T. 2007. Antioxidant potential, anti-proliferative activities, and phenolic content in water-soluble
fractions of some commonly consumed vegetables: Effect of thermal treatment. Food Chemistry. 103, 106-114. Sakanaka, S., Tachibana, Y., and Okada, Y. 2005. Preparation and antioxidant properties of extracts of Japanese persimmon leaf tea (kakinoha-cha). Food
Chemistry. 89, 569-575. 11 US Department of Health and Human Services and US Department of
Agriculture. 2005. Dietary Guidelines for Americans 2005. 6th ed. Washington, DC: US Government Printing Office.
U.S. Food and Drug Administration. 1992. Bacteriological Analysis Manual. 7th ed. Virginia: AOAC International.
Xu, G., Ye, X., Chen, J., and Liu, D. 2007. Effect of heat treatment on the phenolic compounds and antioxidant capacity of citrus peel extract. Journal
of Agricultural and Food Chemistry. 55, 330-335.
Yang, J., Paulino, R., Janke-Stedronsky, R., and Abawi, F. 2007. Free-radicalscavenging activity and total phenols of noni (Morinda citrifolia L.) juice and
powder in processing and storage. Food Chemistry. 102, 302-308. 12