Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J AOAC Int. Author manuscript; available in PMC 2010 April 15.
Published in final edited form as:
PMCID: PMC2855185

Determination of Aconitum Alkaloids in Dietary Supplements and Raw Botanical Materials by Liquid Chromatography/UV Detection with Confirmation by Liquid Chromatography/Tandem Mass Spectrometry: Collaborative Study


An interlaboratory study was conducted to evaluate a method for the determination of 3 Aconitum alkaloids, viz., aconitine, mesaconitine, and hypaconitine, in raw botanical material and dietary supplements. The alkaloids were extracted with diethyl ether in the presence of ammonia. After cleanup by solid-phase extraction to remove matrix interferences, the alkaloids were determined by reversed-phase liquid chromatography (LC)/UV detection at 235 nm with confirmation by LC/tandem mass spectrometry (MS/MS). A total of 14 blind duplicates were successfully analyzed by 12 collaborators. For repeatability, the relative standard deviation (RSDr) values ranged from 1.9 to 16.7%, and for reproducibility, the RSDR values ranged from 6.5 to 33%. The HorRat values were all <2 with only one exception at 2.3. All collaborating laboratories had calibration curves with correlation coefficients of >0.998. In addition, 6 collaborators performed the confirmation and were able to verify the identities of the alkaloids by using LC/MS/MS.

The diester-diterpene Aconitum alkaloids, viz., aconitine, hypaconitine, and mesaconitine, are highly toxic compounds commonly present in aconite roots such as R. aconiti (Chuanwu), R. aconiti kusnezoffii (Caowu), and R. aconiti lateralis (Fuzi). Many dietary supplements for enhancing sexual ability and circulation, restoring health, and relieving pain contain processed aconite roots and thus the Aconitum alkaloids. Proper processing by heating, steaming, and soaking the aconite roots can hydrolyze the highly toxic diester-diterpene Aconitum alkaloids to compounds of much lower toxicity, e.g., benzoylaconine, benzoylmesaconine, and benzoylhypaconine (1). Pharmacological studies indicated that the diester-diterpene Aconitum alkaloids have the same or similar anti-inflammatory and analgesic actions as their hydrolyzed analogs (2). The absence of standardized methods for processing aconite roots has resulted in drastic variation in the alkaloid content, and thus in the safety, of the supplement products containing aconite roots. Intoxication cases arising from consumption of improperly processed aconite roots have been reported in many countries.

Because of this significant public health impact, a single-laboratory validation (SLV) study was conducted for the determination of the 3 Aconitum alkaloids in 6 representative matrixes of aconite root products, including processed raw material (Fuzi), single-ingredient dry powder extract, multi-ingredient dry powder extract, pills, and capsules found in the marketplace (3). This method will facilitate the accurate determination of the quality of botanicals and dietary supplements with respect to the 3 Aconitum alkaloids. In addition, the use of this method may allow dietary supplement manufacturers to set quality standards, and regulatory agencies to monitor safety in the use of dietary supplements containing aconite roots.

Collaborative Study

Study Design

The liquid chromatography/UV detection (LC/UV) phase of this study was conducted with 7 materials as blind duplicates. Some of the materials contained known concentrations of naturally occurring or purposely added (formulated or fortified) aconitine, mesaconitine, and hypaconitine. Two of the materials were negative controls. In addition, each participant was supplied with sufficient quantities of reference standards to conduct the study. Random identification numbers were assigned to each of the blind duplicate materials. A practice sample of dietary supplement material of known concentration was provided for participants. Also, a blank of botanical raw material was provided for the recovery study. These samples were used to ensure that each laboratory could successfully follow the method and to optimize each participant’s instruments and chromatography before the collaborative study samples were analyzed.


Thirteen laboratories participated in this study and received collaborative study materials. From these 13 laboratories, 12 sets of LC/UV data and 6 sets of confirmation data were generated for this collaborative study. The locations of the 13 laboratories were as follows: 4 in the United States, 4 in mainland China, 1 in the United Kingdom, 1 in Korea, and 3 in Hong Kong.

Test Sample Preparation

  1. Source of materials.—The test materials used in this study were obtained form several local health-product shops. In addition, the reference materials were purchased from Sigma (St. Louis, MO; aconitine) and Wako (Osaka, Japan; mesaconitine and hypaconitine). The following materials were used in this collaborative study: processed R. aconiti; a selected dietary supplement product containing processed R. aconiti kusnezoffii; processed R. aconiti and 10 other common herbs; Panax ginseng C.A. Mey as the botanical raw material negative control; 1 mg/kg fortification with P. ginseng C.A. Mey; 10 mg/kg fortification with P. ginseng C.A. Mey; a selected dietary supplement containing 5 common herbs not including Aconitum or related species as the dietary supplement negative control; 5 mg/kg fortification with the dietary supplement negative control.
  2. Preparation, shipment, and storage.—Individual prepared samples and the calibration standard were provided to each collaborator. The samples, as well as the standard, were delivered at ambient temperature with a return receipt document. Collaborators were directed to store the samples at room temperature and the standard at refrigerated temperature until analysis. The prepared calibration solutions were also stored at refrigerated temperature. Following the method provided, the collaborators optimized instrumentation, prepared calibration solutions, analyzed samples, and calculated the results. Before delivery to the collaborators, the Government Laboratory analyzed all of the materials to establish that they were prepared correctly and were homogeneous. Also, the stability of the calibration standard was also checked before delivery.
  3. Practice sample.—One selected dietary supplement containing low levels of the 3 Aconitum alkaloids was provided to each collaborator. This practice sample was used to optimize each collaborator’s instrument and allowed the analysts to become familiar with the testing procedures before proceeding with the full study.

SLV Method Performance

  1. Concentration range.—The calibration curves had a range of 0.5–20 μg/mL for all 3 Aconitum alkaloids.
  2. Validation data.—The responses for the 3 Aconitum alkaloids were linear from 0.05 to 100 μg/mL with a correlation coefficient (r) >0.9999. The slopes of 5 replicates of the calibration curve covering the analytical range for each Aconitum alkaloid standard varied by ≤1.5% in terms of relative standard deviation (RSD) over a period of 2 weeks.
  3. SLV data: precision summary.—Six samples comprising 1 raw material and 5 dietary supplements containing aconite roots or their extracts in the forms of single-ingredient dry powder extract, multi-ingredient dry powder extract, pills, and capsules were chosen to represent products commonly available in the marketplace. For all materials tested in duplicate over a 5-day period, the repeatability RSD (RSDr) values ranged from 1.98 to 3.72% and from 3.57 to 6.94% for hypaconitine and mesaconitine, respectively. All calculated HorRat scores were within the limits for performance acceptability, ranging from 0.35 to 0.61 and from 0.46 to 0.91 for hypaconitine and mesaconitine, respectively, indicating satisfactory precision of the method Because all samples were found to contain aconitine levels lower than the quantitation limit of 1 μg/g, the corresponding precision could not be evaluated. The precision was estimated with reference to the negative control samples spiked with aconitine. Each of the 3 control samples spiked with reference standards at levels of about 2.5, 5, and 10 μg/g was analyzed in triplicate over 3 different days. Satisfactory precision of the method for aconitine was also demonstrated by its RSDr values of 2.68–4.82% and HorRat values of 0.47–0.68.
  4. SLV data: recovery summary.—1Negative control recovery was determined by adding all 3 reference compounds in triplicate to 3 blank matrixes at 3 concentrations (low at approximately 50% and high at 200%) of the normal concentrations (about 5 μg/g) encountered. Negative control 1 was a ginseng sample containing the 3 Aconitum alkaloids. Negative control 2 was a multi-ingredient powder extract dietary supplement. Negative control 3 was a dietary supplement capsule. The recoveries for the 3 Aconitum alkaloids ranged from 86 to 92, 90 to 95, and 91 to 99% for the low, intermediate, and high spiking levels, respectively, indicating satisfactory performance.

AOAC Official Method 2008.11 Aconitum Alkaloids in Dietary Supplements and Raw Botanical Materials

Liquid Chromatography with UV Detection First Action 2008

See Table 2008.11A for the results of the interlaboratory study supporting acceptance of the method.

Table 2008.11A
Interlaboratory study results for Aconitum alkaloids in dietary supplements (LC/UV)

A. Principle

The Aconitum alkaloids are extracted from the various matrixes with diethyl ether in the presence of ammonia, the extract is cleaned up by solid-phase extraction (SPE), and the analytes (aconitine, mesaconitine, and hypaconitine) are quantitated by liquid chromatography/UV detection (LC/UV) at 235 nm. Identities are confirmed by liquid chromatography/tandem mass spectrometry (LC/MS/MS).

B. Apparatus

  1. LC/UV system.—Equipped with UV-Vis detector, quaternary pump, autosampler, column heater, and ChemStation for data acquisition and processing (Agilent Technologies, Wilmington, DE).
  2. LC/MS/MS system.—Triple-quadrupole mass spectrometer equipped with turbo ion spray source and LC system (Applied Biosystems, Foster City, CA).
  3. LC column (for LC/UV system).—Zorbax Eclipse XDB-C8, 5 μm, 4.6 × 150 mm with guard cartridge (Agilent Technologies).
  4. LC column (for LC/MS/MS system).—Xterra MS C8, 3.5 μm, 2.1 × 150 mm (Waters Corp., Milford, MA).
  5. SPE.—Gilson Inc. (Middleton, WI).
  6. Extraction cartridges.—Oasis MCX SPE, 3 mL, 60 mg (Waters Corp.).
  7. Analytical balance.—Readability of 0.1 mg.
  8. Ultrasonic bath.
  9. Centrifuge.—Capable of maximum of 5000 rpm with 15 cm radius of rotor.
  10. Centrifuge tubes.—50 mL.
  11. Pipettor.—Dispensing 100–1000 μL.
  12. Membrane filters.—Puradisc, 13 mm, syringe type, 0.45 μm, made of poly(vinylidene fluoride).

C. Reagents

  1. Deionized water.
  2. Methanol.—LC grade.
  3. Acetonitrile.—LC grade.
  4. Acetic acid.—AR grade.
  5. Formic acid.—AR grade.
  6. Diethyl ether.—AR grade. (Caution: Diethyl ether is a highly flammable solvent.)
  7. Triethylamine.—AR grade.
  8. Ammonium hydroxide.—AR grade.
  9. Ammonium formate.—AR grade.
  10. SPE washing solution.—Mix 70 mL 5% ammonium hydroxide solution with 30 mL methanol.
  11. SPE elution solution.—Mix 5 mL 5% ammonium hydroxide solution with 95 mL methanol.

D. Reference Standards

  1. Aconitine.—95% (Sigma-Aldrich, St. Louis, MO; Store in dark at 2–10°C.
  2. Hypaconitine. —98% (Wako, Osaka, Japan; Store in dark at <25°C.
  3. Mesaconitine. —98% (Wako). Store in dark at <25°C.

E. Preparation of Standard Solutions

  1. Mixed stock standard solution (ca 1000 mg/L).—Accurately weigh about 10 mg each of aconitine, mesaconitine, and hypaconitine into a 10 mL volumetric flask. Add 6 mL acetonitrile–0.1% acetic acid (1 + 1) to the flask and dissolve the solids by shaking. Dilute to volume with the same solvent.
  2. Mixed intermediate standard solution (ca 100 mg/L).—Pipet 1.0 mL mixed stock standard solution into a 10 mL volumetric flask, and dilute to volume with acetonitrile–0.1% acetic acid (1 + 1).
  3. Working standard solutions (ca 0.5–20 mg/L).—Freshly prepare ≥5 working standard solutions (including a calibration blank), within the range stated above by dilution of the mixed intermediate standard solution with acetonitrile–0.1% acetic acid (1 + 1) according to Table 2008.11B. Pipettors of appropriate range should be used to transfer the standard solution. The volume of the prepared stock standard solution should be enough for the preparation of 2 sets of working standard solutions.
    Table 2008.11B
    Preparation of working standard solutions

(Note: The 3 Aconitum alkaloids are susceptible to hydrolysis in methanol, especially under alkaline conditions. The half-life of aconitine, hypaconitine, and mesaconitine in methanol with 5% ammonia is about 4–5 days. However, there were no signs of degradation of the alkaloids when stored in acetonitrile or acidic media even for a period of 6 months.)

F. Preparation of Reagents

  1. Mobile phase A for LC/UV.—20 mM triethylamine (in water) adjusted to pH 3 with phosphoric acid–methanol (95 + 5).
  2. Mobile phase B for LC/UV. —Methanol.
  3. Mobile phase for LC/MS/MS (isocratic).—Methanol– 20 mM ammonium formate in 0.1% formic acid (45 + 55).

G. Procedure

  1. Extraction.—Accurately weigh approximately 1.0 g sample into a 50 mL centrifuge tube. Add 1 mL 10% ammonium hydroxide solution. Stir well to ensure that the entire sample is wet. Then add 25 mL diethyl ether to the tube. Shake the tube on a platform shaker for 1 h at a speed of 300 rpm. Centrifuge the sample at 4000 rpm for 10 min to settle the solids. (Caution: To avoid loss of any ether vapor and risk of explosion, the tube must be capable of withstanding the forces involved and must be sealed to avoid leaks.) Decant the diethyl ether into another centrifuge tube. Repeat the sample extraction twice each time with 10 mL diethyl ether, and shake the tube for 30 and 10 min, respectively. Combine the extracts, and evaporate to complete dryness at ≤40°C under a stream of nitrogen. Add 5 mL acetonitrile–0.1% acetic acid (1 + 1) and ultrasonicate for 20 min to redissolve the residues The solution is ready for SPE cleanup after it is passed through a 0.45 μm membrane filter If time does not allow the following SPE cleanup step store the solution at 4°C However the cleanup step must be completed within 2–3 days
  2. MCX SPE cleanup.—Condition the SPE cartridge with 1 mL methanol and 1 mL water After loading 4 mL sample solution onto the SPE cartridge wash cartridge sequentially with 1 mL 0.1% acetic acid 1 mL methanol 1 mL water and 1 mL SPE washing solution Dry the cartridge and elute the analytes from the cartridge with 2 mL SPE elution solution The flow rate for loading the sample solution should be 0.2 mL/min whereas the flow rate for the other steps is 0.5 mL/min Collect the elution solution in test tubes of appropriate size provided with the SPE system Evaporate the solvent to dryness at about but not exceeding 40°C under a stream of nitrogen (Note: The analytes will decompose gradually with a half-life of 4–5 days when stored in the SPE elution solution) Reconstitute the residue in 1 mL acetonitrile–0.1% acetic acid (1 + 1) and ultrasonicate for 20 min to redissolve the residues The solution is ready for chromatographic analysis after it is passed through a 0.45 μm membrane filter Appropriate dilution using acetonitrile–0.1% acetic acid (1 + 1) must be made if the responses of the analytes in the sample extracts are out of the calibration ranges
  3. LC/UV operating conditions.—Column temperature: 25°C flow rate: 1.0 mL/min; injection volume: 10 μL; detection wavelength: 235 nm; run time: 56 min The gradient program is shown in Table 2008.11C A representative chromatogram of the standard solutions is shown in Figure 2008.11
    Figure 2008.11
    Mixed standard (10 μg/mL).
    Table 2008.11C
    Gradient program for LC/UV determination
  4. LC/MS/MS operating conditions.—MS settings: ionization mode: electrospray positive ionization; curtain gas: 8 units; nebulizer gas: 15 units; ion spray voltage: 5000 V; temperature: 350°C The flow rate of the mobile phase was set at 200 μL/min.

H. Calculations

Establish a calibration curve of peak area of each analyte in the working standard solutions versus concentration of analyte Determine the concentration of each analyte in the sample extract solution from the calibration curve If the concentration of an analyte in the sample extract exceeds the calibration range make appropriate dilution and repeat the determination Calculate the concentration of each analyte in the sample, C, in mg/kg, by using the following equation:


where A = concentration of the analyte found in the sample solution mg/L; W = weight of sample used g; V = final volume of the sample solution, mL; D = dilution factor if any; and 5/4 = factor that accounted for the portion of extract taken for the cleanup.

I. Confirmation by LC/MS/MS

When an Aconitum alkaloid is detected at ≥1 mg/kg as determined by LC/UV, the identity is confirmed by LC/MS/MS. For confirmation, the intensity of the respective multireaction monitoring ion pair signal should be ≥400 cps (Table 2008.11D).

Table 2008.11D
MRM ion pairs for confirmation

References: J. AOAC Int. 89, 1496 (2006); 92, 111 (2009).

Results and Discussion

Thirteen laboratories participated in the collaborative study. One laboratory was not able to finish the study because of lack of time. The remaining laboratories were able to submit data before the submission deadline.

Collaborative Study Results

Results, in mg aconitine, mesaconitine, and hypaconitine/kg product, for each of the 7 blind replicates are presented in Table 1. Test samples were given codes before shipment to the collaborators and then decoded when the results were returned. To allow the collaborators to check their competence in conducting the analysis, the procedure and blank botanical material were provided for them to conduct a recovery study before the analysis of the test materials. According to the data received, 11 laboratories were able to obtain average recoveries for the 3 Aconitum alkaloids that were very close to or within the acceptable range of 80–100%, whereas the recoveries obtained by the remaining laboratory (Laboratory L) were only about 35%. This laboratory had discussed the problem with the organizing laboratory, but it still failed to identify the causes for the poor recovery. It was also noted that most of the results of Laboratory L were inconsistent with those of the other collaborators. Therefore, it was decided not to include the results of Laboratory L in the evaluation of the performance characteristics for this study.

Table 1
Interlaboratory study results for the determination of Aconitum alkaloids

Laboratory B reported that samples A1 and C1 were lost during the sample preparation. As a result, no data were provided for these 2 samples. Laboratory H observed that, during the sample extraction for samples A1 and A2, sample at the bottom of the centrifuge tube lumped together and was not able to be wetted by the extraction solvent. Because this observation implied that the extraction for these 2 samples might not be complete, the respective data entries were thus not included in the statistical analysis. Judging from the data, it was suspected that Laboratory I and Laboratory J might have mixed up some of their results. For example, Laboratory I may have mixed up the results for samples C1 and C2 with those for samples F1 and F2, and Laboratory J may have mixed up the results for samples D2, E2, and F2. The organizing laboratory raised this issue with both laboratories but the issue was not resolved by either laboratory. In order not to affect the evaluation of the performance characteristics for this study, the data entries in question were not included in the statistical analysis. In addition, Laboratory E reported that mesaconitine was detected in samples G1 and G2, and hypaconitine, in samples F1 and F2, which were the dietary supplement negative control replicates. These results might be due to contamination or to carry-over peaks in the LC/UV analysis.

Table 2008.11A presents statistical summaries of the results from Laboratories A–K. Statistical analysis to determine repeatability and reproducibility was performed by using the AOAC INTERNATIONAL Statistical Program (Version 2.0) for Blind Replicates. The values for repeatability standard deviation (Sr), reproducibility standard deviation (SR), RSDr, reproducibility relative standard deviation (RSDR), and number of statistical outliers are presented. HorRat values are also presented and are calculated as RSDR (observed)/RSDR(predicted), where the RSDR(predicted) is calculated by using the following equation:


where C is the measured analyte concentration in decimal mass units (4). The Cochran, Grubbs, and Double Grubbs tests were used to remove statistical outliers where appropriate. Data from laboratories reporting values for individual alkaloids as greater-than or less-than values were not included in the statistical analysis.

Collaborators’ Comments

The collaborators were able to follow the method with very few difficulties. Five laboratories reported difficulties in mixing the entire sample inside the centrifuge tube with the 1 mL 10% ammonium hydroxide solution before the extraction. One collaborator suggested using 2 mL 5% ammonium hydroxide solution instead. Another laboratory performed the extraction by using a 50 mL conical flask instead of the centrifuge tube.

Performance Characteristics of the Method

The method performed well in the collaborative study for both detection and confirmation of the 3 Aconitum alkaloids. For repeatability, the RSDr values ranged from 1.9 to 16.7% and for reproducibility, the RSDR values ranged from 6.5 to 33%. Five materials had acceptable HorRat values except for the aconitine determined in the processed R. aconiti, for which the HorRat value was 2.37. For the 2 negative control materials, the HorRat values were not applicable. The number of results identified as outliers and disregarded was ≤4 out of 21 or 22 for all samples except for hypaconitine determined in the spiked dietary supplement negative control for which 2 pairs of Cochran and 1 pair of Grubbs outliers were identified. The number of outlier laboratories identified, excluding those suspected of having results mixed up with those of other samples, was <2 out of 11, except for the determination of hypaconitine in the spiked dietary supplement negative control, for which 3 laboratories were identified as outliers.

In addition, all collaborating laboratories had correlation coefficients of >0.998 for the calibration curves generated. Among the participating laboratories, 6 verified the identity of the Aconitum alkaloids found in the samples by using LC/MS/MS. The reason that the other laboratories failed to do so might be the unavailability of the required equipment.


On the basis of the results of this collaborative study, it is recommended that the method be adopted Official First Action for the determination of Aconitum alkaloids in dietary supplements and raw botanical materials.


We thank T.L. Ting, Government Laboratory, for his encouragement and support throughout the course of this study. We would also like to thank the following collaborators for their participation in this study:

Darryl Sullivan, Covance Laboratories, Madison, WI

Brian T. Schaneberg, ChromaDex Co., Boulder, CO

Shen Ji, Shanghai Institute for Drug Control, People’s Republic of China

Kelvin S.Y. Leung, Baptist University, Hong Kong SAR, China

C.T. Che, Chinese University of Hong Kong, Hong Kong SAR, China

Alan Richards, Scientific Services, Durham, UK

Jong-Hwan Kim, Korean Food and Drug Administration, Korea

Sidney Sudberg, Alkemists Pharmaceticals, Costa Mesa, CA

Wendi Wang, Advanced Botanical Consulting & Testing, Inc., Tustin, CA

Hong-Zhu Guo, Beijing Institute for Drug Control, People’s Republic of China

Li-Xin Zhang, Chengdu Institute for Drug Control, People’s Republic of China

Feng-Yi Zheng, Guangdong Institute for Drug Control, People’s Republic of China


Collaborators: C.T. Che; H.-Z. Guo; S. Ji; J.-H. Kim; K.S.Y. Leung; C.K. Lo; A. Richards; B.T. Schaneberg; S. Sudberg; D. Sullivan; W. Wang; Y.C. Wong; L.-X. Zhang; F.-Y. Zheng

The recommendation was approved by the Methods Committee on Dietary Supplements as First Action. See “Official Methods Program Actions,” (2008) Inside Laboratory Management, November/December issue.


1. Ohta H, Seto Y, Tsunoda N. J Chromatogr B. 1998;714:215–221. [PubMed]
2. Ito K, Tanaka S, Konno S, Konishi Y, Mizugaki M. J Chromatogr B. 1998;714:197–203. [PubMed]
3. Tang W-T, Wong S-K, Law T-Y, Pang K-C, Sin D, Tam Y-K. J AOAC Int. 2006;89:1496–1514. [PMC free article] [PubMed]
4. AOAC Guidelines for Single-Laboratory Validation of Chemical Methods for Dietary Supplements and Botanicals. AOAC INTERNATIONAL; Gaithersburg, MD: Apr, 2003.