FT-NIR Combined with Chemometric
Analysis as a Means to Distinguish Powdered Goldenseal Root from its
Adulterants
Reviewed: Liu Y, Finley J, Betz JM, Brown PN. FT-NIR
characterization with chemometric analyses to differentiate goldenseal from
common adulterants. Fitoterapia.
2018;127:81-88.
Keywords: adulteration,
Berberis aquifolium, coptis, Coptis chinensis, FT-NIR,
goldenseal, Hydrastis canadensis, Oregon
grape, Rumex crispus, Xanthorhiza simplicissima, yellow dock, yellow root
The roots of goldenseal (Hydrastis canadensis, Ranunculaceae) are an important
medicinal ingredient in North American herbal medicine and elsewhere.
Practitioners appreciate the root as a stand-alone ingredient because of its
antimicrobial activities, but it is also frequently combined with echinacea (Echinacea spp., Asteraceae) and other botanicals in dietary
supplements to support healthy immune function.
Adulteration of goldenseal
has been reported on a number of occasions, and is thought to be due to the
availability of materials that also contain berberine, or similar
benzylisoquinoline alkaloids at a considerably lower price, thus providing an
incentive for economically-motivated fraud.1 Most often, analytical
laboratories use chromatographic methods, e.g., high-performance liquid chromatography
(HPLC) combined with ultraviolet/visible spectrophotometry (UV/Vis) to measure
the goldenseal alkaloids.2-5 Using Fourier-transform near-infrared
(FT-NIR) spectroscopy, a faster and less costly approach was evaluated for its
ability to differentiate powdered crude goldenseal from its known adulterants,
and to determine the level at which these adulterants could be detected.
The main goal of the project
was to determine which statistical assays would represent the best choice for
the evaluation of the FT-NIR spectra. The authors compared partial least square
(PLS) regression analysis alone and with preprocessing methods (external
parameter orthogonalisation [EPO] and generalized least squares weighting
[GLSW]), soft independent modelling with class analogy (SIMCA) alone and with
preprocessing methods (EPO, GLSW, and extended multiplicative scatter
correction [EMSC]), and moving window – principal component analysis (MW-PCA)
without preprocessing of the data.
Only MW-PCA and SIMCA
combined with EMSC provided satisfactory results in distinguishing among the
five species included in the analysis. The model was then tested at theoretical
adulteration levels of 2%-95%. It was able to easily detect yellow root (Xanthorhiza simplicissima, Ranunculaceae), Oregon grape (Berberis aquifolium, Berberidaceae) root, and coptis (Coptis chinensis, Ranunculaceae) root adulteration in goldenseal
at levels of 10% and above. For yellow dock (Rumex
crispus, Polygonaceae), SIMCA-EMSC was only able to detect the
adulteration at 15-20%. Using the MW-PCA model, FT-NIR was able to detect coptis
adulteration levels as low as 2%, and yellow dock and Oregon grape adulteration
between 15% and 20%. Baseline drift issues were observed in yellow root,
leading to much higher detection limits (15%-50%).
Comment: The results provide evidence that FT-NIR in combination
with an optimized chemometric model allows the detection of four common
goldenseal adulterants in cases where materials are purchased in whole, cut, or
powdered root form at levels generally between 5-20%. The low costs and quick
sample turnaround will be appealing features for using this method in quality
control laboratories. Another advantage is the possibility of making
computer-generated spectra of hypothetical sample mixtures, eliminating the
need to weigh out and homogeneously mix goldenseal and adulterants at specific
levels (i.e., 1.5 grams of adulterant in 8.5 grams of goldenseal to make a
sample containing 15% adulterant). As is often the case with statistical models
requiring a large number of samples, one of the bottlenecks is the availability
of authenticated plant material, especially of the alleged adulterants. As
such, the inclusion of only one sample of yellow dock in this project is not
ideal. In addition, many manufacturers may need to detect adulterants at levels
below what FT-NIR was able to provide as the 5% limit of detection exceeds the
proscribed 2% foreign organic matter limit established by most pharmacopoeias.
Interestingly, yellow dock was
one of the more difficult ingredients to distinguish from goldenseal despite
the absence of alkaloids and the presence of anthraquinones such as aloin,
emodin, and physcion.6,7 Most previously-reported chromatographic
methods have relied on the evaluation of alkaloids other than those in
goldenseal, such as palmatine and jatrorrhizine, which is an easy way to detect
adulteration with coptis, yellow root, or Oregon grape, and is suggested as
another appropriate way to determine the authenticity of goldenseal root and
root extracts.2-5
References
- Tims M. Adulteration
of Hydrastis canadensis root and
rhizome – Botanical Adulterants Prevention Bulletin.
Austin, TX: ABC-AHP-NCNPR Botanical Adulterants Prevention Program. 2016;1-6.
- Avula
B, Wang YH, Khan IA. Quantitative determination of alkaloids from the roots of Hydrastis canadensis L. and dietary supplements
using ultra-performance liquid chromatography with UV detection. J AOAC Int. 2012;95(5):1398-1405.
- Weber
HA, Zart, MK, Hodges AE, White KD, Barnes SM, Moody LA, Clark AP, Harris RK,
Overstreet JD, Smith CS. Method validation for determination of alkaloid
content in goldenseal root powder. 2002; J
AOAC Int. 86:476.
- Kamath
S, Skeels M, Pai A. Significant differences in alkaloid content of Coptis chinensis (Huanglian), from its
related American species. Chinese Medicine. 2009;4:17.
- Upton
R, Graff A, Swisher D. Goldenseal root, Hydrastis
canadensis L.: Standards of Analysis,
Quality Control, and Therapeutics. Santa Cruz, CA: American Herbal
Pharmacopoeia; 2001.
- Jeelani
SM, Farooq U, Gupta AP, Lattoo SK. Phytochemical evaluation of major bioactive
compounds in different cytotypes of five species of Rumex
L. Ind Crop Prod. 2017;109:897-904.
- Wegiera M, Smolarz
HD, Wianowska D, Dawidowicz AL. Anthracene derivatives in some species of Rumex L. genus. Acta Soc Bot Pol.
2007;76(2):103-108.