468 FARER & HAYES: JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO.

2, 2005

DRUGS, COSMETICS, FORENSIC SCIENCES

Comparison Study of Two Procedures for the Determination of

Emamectin Benzoate in Medicated Fish Feed
LESLIE J. FARER and JOHN M. HAYES
Schering-Plough Research Institute, Analytical Development, 2000 Galloping Hill Rd, Kenilworth, NJ 07033

A new method has been developed for the matrix standards and an internal standard, and to substitute
determination of emamectin benzoate in fish feed. external standards. In the process, a faster wet-extraction

Downloaded from https://academic.oup.com/jaoac/article/88/2/468/5657333 by guest on 02 October 2021

The method uses a wet extraction, cleanup by technique was developed, and improved method precision
solid-phase extraction, and quantitation and was obtained.
separation by liquid chromatography (LC). In this A comparison of these methods was recently undertaken.
paper, we compare the performance of this method Twenty-two medicated salmon feed samples, representing a
with that of a previously reported LC assay for the variety of sources and manufacturers, were assayed by both
determination of emamectin benzoate in fish feed. methods for statistical evaluation of the results.
Although similar to the previous method, the new
procedure uses a different sample pretreatment, Experimental
wet extraction, and quantitation method. The
performance of the new method was compared
Complete descriptions of the original method (Method I)
with that of the previously reported method by
and the revised method (Method II) are outlined in the
analyses of 22 medicated feed samples from
referenced articles (14, 15). Tables 1 and 2 outline the sample
various commercial sources. A comparison of the
preparation and liquid chromatography (LC) of each method.
results presented here reveals slightly lower assay
Briefly, the differences between the methods are as
values obtained with the new method. Although a
follows. For Method I, a calibration curve is constructed from
paired sample t-test indicates the difference in
preparations of fortified control feed over a range of
results is significant, this difference is within the
1–25 ppm emamectin (based on 20 g control feed weight).
method precision of either procedure.
Quantitation is performed by internal standard analysis.
Method II uses external standard calibration solutions over a

E
range of 1–40 ng/mL (Table 1). Although the use of standards
mamectin benzoate is an avermectin, a class of
prepared from fortified control feed and the internal standard
compounds including ivermectin and abamectin (1–3),
improved recovery, the procedure was tedious and time
which are used as pesticides and parasiticides.
consuming.
Emamectin benzoate has been used as a pesticide in crop
protection (4, 5) and more recently is being commercialized in Method II simplifies and shortens the sample preparation
the aquaculture industry as an ectoparasiticide (6–13). The procedure considerably (details presented in Table 1). For
commercial feed additive SliceÒ (emamectin benzoate, 0.2%, Method I, samples are prepared from intact feed pellets,
Schering-Plough Corp., Kenilworth, NJ), a type-A medicated disintegrated by 30 min of vortexing in the presence of glass
premix, was developed for the control of sea lice beads, followed by a 16 h extraction on a wrist-action shaker.
(Lepeophtheirus spp. and Caligus spp.) in Atlantic salmon and For Method II, frozen pellets are ground in a coffee grinder,
trout. The analyte is incorporated into fish feed by a top-dress followed by a 2.5 h extraction using ultrasonication and
procedure at concentrations ranging from 1 to 25 ppm. shaking. The solid-phase extraction (SPE) steps are identical
In 1999, our laboratory originally reported a liquid for both methods.
chromatographic procedure for the quantitation of emamectin The chromatography used in the 2 methods is similar
benzoate in fish feed (14). The original procedure (Method I) (Figures 1 and 2; Table 2). Both use a C8 reversed-phase
used spiked matrix standards and an internal standard to column with detection at 244 nm. Gradient elution (40 to 90%
optimize recovery. A new procedure (Method II) was acetonitrile over 20 min) is used to separate emamectin
developed to conform to regulatory requirements specified by benzoate from endogenous interference, and a column wash,
the Center for Veterinary Medicine of the U.S. Food and Drug acetonitrile–tetrahydrofuran (50 + 50), is required to elute
Administration. Our goal was to eliminate the use of both strongly retained components. Finally, data analysis is
simplified in Method II by eliminating the internal standard
while maintaining adequate accuracy and precision.
Received July 15, 2004. Accepted by JM August 10, 2004. The 22 batches used for this study were obtained from
Corresponding author's e-mail: [email protected].
various commercial suppliers. Emamectin benzoate was
 FARER & HAYES: JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 2, 2005 469

Table 1. Comparison of the 2 sample preparation methods

Method parameter Method I Method II

Sample pretreatment Intact pellets disintegrated with glass beads and Pellets ground in coffee grinder
vortexing in extraction solvent
Sample size 20 g Same as Method I
Extraction solvent Disintegration in 20 mL 0.1% phosphoric acid solution, 100 mL methanol
followed by extraction in 100 mL methanol
Wet extraction Performed in 100 mL methanol with 30 min of Performed in 100 mL methanol with 2 cycles of
vortexing, followed by 16 h of shaking shaking (1 h) and sonication (15 min)
SPE cartridge 2 g C18a Same as Method I
Volume of sample extract 50 mL 25 mL
loaded onto SPE cartridge

Downloaded from https://academic.oup.com/jaoac/article/88/2/468/5657333 by guest on 02 October 2021

SPE Wash steps: methanol–water (70 + 30) followed by Wash steps: same as Method I; elution step: same as
methanol–water (40 + 60); elution step: 2% ammonium Method I; reconstitution: methanol (2.0 mL)
hydroxide in ethanol; reconstitution: methanol (1 mL)

a
Isolute C18EC (Jones Chromatography Inc., Lakewood, CO).

top-dressed onto fish feed pellets at levels of 1–25 ppm. Pellet The results for 22 samples analyzed by each method are
sizes ranged from 2.5 to 12 mm. Batch samples of ³500 g were outlined in Table 3. In general, results generated by using
processed through a sample splitter until portions weighing Method I were higher than those of Method II. The means of
approximately 250 g were obtained. Smaller batch samples 22 sample assays performed by using Methods I and II were
(<500 g) were split manually. Unground pellets were stored in a 8.00 and 7.75 ppm, respectively, with a mean difference of
freezer until analysis. Duplicate 20 g weighings of intact pellets 0.25 ppm. However, Method II showed somewhat better
were treated according to Method I. For Method II, pellets were agreement between replicates, with an average difference of
ground by using a coffee grinder, and duplicate 20 g samples 0.25 ppm versus 0.47 ppm for Method I. The difference
were weighed, processed, and chromatographed. between the means obtained by Methods I and II for the
22 samples ranged from 0.6 to 19%.
Results and Discussion A paired-sample t-test was used to compare the result pairs
(JMP 4.0.4, The SAS Institute, Cary, NC) to evaluate whether
Both methods were validated to assay medicated feed over the difference in results obtained by Methods I and II was
a range of 1–25 ppm. The limit of quantitation for both significant. Data were evaluated on the basis of the null
methods is 0.5 ppm. hypothesis that the difference between data pairs was zero, at a

Table 2. Comparison of liquid chromatograph

LC parameter Method I Method II

Mobile phase Mobile phase A: 0.1% phosphoric acid Mobile phase A: same as Method I; mobile phase B:
solution–acetonitrile (60 + 40); mobile phase B: acetonitrile 0.1% phosphoric acid solution–acetonitrile (10 + 90)
Gradient 0 to 50% mobile phase B in 20 min, followed by column 0 to 55% mobile phase B in 20 min, followed by column
wash of acetonitrile–tetrahydrofuran (50 + 50) wash of acetonitrile–tetrahydrofuran (50 + 50)
Flow rate 1.0 mL/min Same as Method I
Detection UV, 244 nm Same as Method I
Injection volume 10 mL 20 mL
Column temp. 50° ± 5°C Same as Method I
Analytical column Phenomenex Primesphere C8, 150 ´ 4.6 mm Zorbax RX-C8, 250 ´ 4.6 mm
Run time 40 min Same as Method I
Quantitation method Internal standard analysis using fortified controls as External standard analysis versus a set of
calibration standards standard solutions
Range 1–25 ppm Same as Method I
Limit of quantitation 0.5 ppm Same as Method I
470 FARER & HAYES: JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 2, 2005

Downloaded from https://academic.oup.com/jaoac/article/88/2/468/5657333 by guest on 02 October 2021

Figure 2. Liquid chromatogram of a salmonid feed
Figure 1. Liquid chromatogram of a salmonid feed
sample medicated with Slice at a level of 5 ppm,
sample medicated with Slice at a level of 5 ppm,
obtained by using Method II.
obtained by using Method I.

Table 3. Emamectin benzoate assay results for 22 samples analyzed by Methods I and II
Method I assay results, ppm Method II assay results, ppm

Difference
between
Sample Pellet size, mm Prep. 1 Prep. 2 Diff. Mean Prep. 1 Prep. 2 Diff. Mean means, %

1 10 10.39 10.59 0.20 10.49 9.13 8.20 0.93 8.67 19.0

2 4 9.74 8.51 1.23 9.13 8.53 8.56 0.03 8.55 6.6
3 3 4.17 4.51 0.34 4.34 4.03 4.08 0.05 4.06 6.7
4 4 4.99 4.93 0.06 4.96 4.42 4.46 0.04 4.44 11.1
5 10 8.28 8.20 0.08 8.24 8.14 8.19 0.05 8.17 0.9
6 10 11.39 10.07 1.32 10.73 10.12 10.23 0.11 10.18 5.3
7 10 9.23 9.30 0.07 9.27 8.69 8.50 0.19 8.60 7.5
8 10 9.97 8.40 1.57 9.19 10.07 9.85 0.22 9.96 8.0
9 10 9.22 8.97 0.25 9.10 9.67 9.32 0.35 9.50 4.3
10 10 10.22 9.76 0.46 9.99 9.68 8.82 0.86 9.25 7.7
11 10 9.43 10.24 0.81 9.84 10.30 9.70 0.60 10.00 1.6
12 10 9.31 9.76 0.45 9.54 9.96 9.58 0.38 9.77 2.4
13 5 5.57 5.14 0.43 5.36 4.69 4.71 0.02 4.70 13.1
14 5 5.15 5.48 0.33 5.32 4.60 4.60 0.00 4.60 14.5
15 5 4.76 5.96 1.20 5.36 5.20 4.91 0.29 5.06 5.8
16 5 5.17 5.25 0.08 5.21 5.27 5.29 0.02 5.28 1.3
17 5 5.20 5.24 0.04 5.22 5.43 4.94 0.49 5.19 0.6
18 5 5.01 5.02 0.01 5.02 4.90 4.76 0.14 4.83 3.9
19 10 12.81 12.48 0.33 12.65 11.54 11.37 0.17 11.46 9.9
20 3 5.41 5.08 0.33 5.25 4.47 4.50 0.03 4.49 15.6
21 12 20.26 21.00 0.74 20.63 22.96 22.42 0.54 22.69 9.5
22 2.5 1.15 1.08 0.07 1.12 1.04 1.02 0.02 1.03 8.4
a
Average 0.47 8.00 0.25 7.75

a
The difference between the averages was 0.25 ppm.
 FARER & HAYES: JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 2, 2005 471

significance level (a) of 0.05. Assay values for Method I were References
higher by 0.25 ppm, on average (8.00 ppm for Method I and
7.75 ppm for Method II). By using the paired-sample t-test, (1) Ivermectin and Abamectin (1989) W.C. Campbell (Ed.),
the calculated a value, 0.03, was less than the significance Springer-Verlag, New York, NY, p vii
level of 0.05, indicating that the difference between the results (2) Fisher, M.H. (1997) in Phytochemicals for Pest Control,
obtained by the 2 methods is significant. Americal Chemical Society, Washington, DC, pp 20–238
However, as a percentage of the mean assay value of the (3) Fisher, M.H., & Mrozik, H. (1984) Chemistry, Biology and
22 samples, this difference between means is approximately Practice, Academic Press, New York, NY, pp 553–606
3%, consistent with the method variability for either (4) Cvetovich, R.J., Kelly, D.H., DiMichele, L.M., Shuman,
procedure and less than the quantitation limit for either R.F., & Grabowski, E.J. (1994) J. Org. Chem. 59, 7704–7708
procedure (0.5 ppm). Furthermore, the difference between (5) Putter, I., MacConnell, J.G., Preiser, F.A., Haidri, A.A.,
means, 0.25 ppm, is similar to the average difference between Ristich, S.S., & Dybas, R.A. (1981) Experientia 37, 963–964
duplicate assays. (6) Stone, J., Sutherland, I.H., Sommerville, C.S., Richards,
R.H., & Varma, K.J. (1999) J. Fish Dis. 22, 261–270

Downloaded from https://academic.oup.com/jaoac/article/88/2/468/5657333 by guest on 02 October 2021

(7) Stone, J., Sutherland, I.H., Sommerville, C.S., Richards,
Conclusions
R.H., & Varma, K.J. (2000) Aquaculture 186, 205–219
(8) Stone, J., Sutherland, I.H., Sommerville, C.S., Richards,
A new liquid chromatographic assay method for the
R.H., & Varma, K.J. (2000) Dis. Aquat. Organisms 41,
quantitation of emamectin benzoate in fish feed, Method II, 141–149
was compared with a previously published method, Method I.
(9) Stone, J., Sutherland, I.H., Sommerville, C.S., Richards,
The new method introduces several significant changes in R.H., & Endris, R.G. (2000) J. Fish Dis. 23, 185–192
response to current regulatory issues, while improving (10) Stone, J., Roy, W.J., Sutherland, I.H., Ferguson, H.W.
precision and analysis time. Sommerville, C., & Endris, R. (2002) Aquaculture 210,
The performance of the 2 methods was compared by 21–34
analyzing 22 samples of emamectin benzoate-medicated fish (11) Armstrong, R., MacPhee, D., Katz, T., & Endris R. (2000)
feed. Although results were generally comparable, a Can. Vet. J. 41, 607–612
paired-sample t-test indicated that the assay difference (12) Roy, W.J., Sutherland, I.H., Rodger, H.D. & Varma, K.J.
between the methods was significant. However, the mean (2000) Aquaculture 184, 19–29
difference, 0.25, is approximately 3% of the mean assay value (13) Ramstad, A., Colquhoun, D.J., Nordmo, R., Sutherland, I.H.,
for 22 samples, within the level of precision for either method. & Simmons, R. (2002) Dis. Aquat. Organisms 50, 29–33
Thus, Method II obtained results that, although statistically (14) Farer, L.J., Hayes, J., Rosen, J., & Knight, P. (1999) J. AOAC
different, are adequately accurate and precise in comparison Int. 82, 1281–1287
with those obtained by using Method I. (15) Farer, L.J. (2005) J. AOAC Int. 88, 462–467