A UV/Vis spectrophotometric methodology for quality control of stevia-based extracts in the food industry

Abstract

This study established a methodology for quality control of stevia extracts in food industry where one, or a mixture of major stevia glycosides, rebausioside A and stevioside, are present. The methodology relies on pre-treatment to eliminate glucose that may be already present, prior to acid hydrolysis of stevia glycosides toward isosteviol plus glucose and ensuing released glucose quantification using the glucose oxidase/peroxidase coupled assay. Subsequent isosteviol isolation and concentration assessment using UV spectroscopy, permits the quantification of an average number of glucose molecules released per a generated molecule of isosteviol. The latter increases the assay specificity, allowing confirmatory assessment of stevia glycoside identity and purity. Favorable technical characteristics, satisfying international specifications for method validation, were determined for each individual assay (glycoside quantification by acid hydrolysis and isosteviol quantification) for rebaudioside A and stevioside: Standard curves were linear (R2 ≥ 0.993 [R2 ≥ 0.984 for the glucose/isosteviol quantification assay]) over a concentration range from approximately 10−1 mg/mL for the acid assay and 3.0 mg/mL for the glucose/isosteviol assay up to at least 10 mg/mL glycoside. The CV% on all quality control levels was ≤6.33% for repeatability and ≤12.03% for intermediate precision. Bias was ≤±5.00% in all cases. Upon comparison with the gold standard LC/UV method for rebA quantification, we demonstrated a very good agreement and acceptable bias (slope: 1.02, y-intercept: −0.73, R2: 0.994). Acid hydrolysis was tested successfully on commercial products, stevia extracts and commodity sweeteners, containing either of the two major glycosides. We demonstrated that the proposed methodology was able to distinguish between commercial samples containing pure rebaudioside A or stevioside alone and samples containing additional steviol glycosides in combination. When verifying glycoside content in binary mixtures of rebaudioside A and stevioside, bias was sufficiently low (<10%). The proposed methodology requires minimal equipment and expertise and is applicable to a significant percentage (approximately 50%) of all commercial stevia extracts and sweeteners. Our findings should provide a useful tool for stevia quality control, thus allowing an increase in the frequency and extent of testing, even in minimally equipped settings, while containing cost.