QUANTIFICATION OF RESIDUAL CLOVE OIL, BENZOCAINE AND TRICAINE IN FISH FILLETS USING SPE AND UPLC-DAD

Residual quantification of the anesthetics clove oil (CO) – isoeugenol (ISO), eugenol (EUG) and methyleugenol (MET) –, benzocaine (BZN) and tricaine (MS-222) was made in fillets of two fish species: Nile tilapia (Oreochromis niloticus) and a catfish hybrid, cachadia (Pseudoplatystoma reticulatum x Leiarius marmoratus). Samples (n=4) of each fish were evaluated after submitted to anesthesia in five dosages defined based on the induction time of each species after depuration times (0h, 12h, 24h and 48h). Different methodologies of sample preparation were tested and selected according to the better recovery. The quantification of anesthetics was performed by UPLC-DAD. The variance of residual means among anesthetics, dosages and fish species was compared. After anesthesia (0h) both species, tilapia and cachadia, presented residual anesthetics. Fishes depurated during 12h, 24h and 48h did not present detectable values, it means, values were below the limits of detection. BZN presented the highest mean residual concentration for tilapia and cachadia (p=0.01), while MS-222 presented the lowest residual amounts in tilapias and EUG in cachadias, what may be related to the metabolism and carcass composition of each fish species. There were no significant differences among the five dosages, except the lowest MS-222 concentration in tilapias that resulted in higher residual concentrations because low dosages increase the induction time and consequently the permanence of the fish in anesthesia. Ultimately, mean values of residues in cachadia were higher than in tilapia, and MS-222 and EUG presented the lowest residual values for tilapia and cachadia, respectively.


Introduction
On aquaculture the management, transport, marking, application of hormones and vaccines cause intense stress on animals, and in this context anesthetic or analgesic compounds have been applied trying to ease the negative effect of such practices (Mettam et al., 2011;Sneddon, 2009;Ross et al., 2008).
Nowadays the most used anesthetics on aquaculture are MS-222, BZN, CO or formulations containing its active ingredient: ISO, menthol, quinaldine and metomidate (CCAC, 2011;FDA, 2011;Kiessling et al., 2009). Such compounds may persist on the animal organism even after being metabolized, and then it is necessary to define the adequate depuration period for each species and anesthetics applied, as well as the conditions for application, like concentration, time of permanence in anesthesia and recovery (Mettam et al., 2011;Sneddon, 2009).
The anesthetic residual period depends on factors like temperature, salinity, chemical composition of water, fat content, among others (Burka et al., 1997). Such factors may contribute differently with fishes of temperate and tropical regions, for example, once these last are submitted to higher temperatures, what explains their faster metabolism. Other important aspect to be considered is the physiology of each species since there are more than 30000 fish species (Zahl et al., 2012).
Fishes with scales like Nile tilapia andcatfishes like cachadia, both studied in this work, may absorb and excrete anesthetic compounds differently due to their different body composition and metabolism. Thus it is necessary to evaluate these factors in order to define the depuration time of such compounds on the organism of fishes submitted to anesthesia.
Countries like United States of America, Canada, Norway and United Kingdom regulated the MS-222 use for some fish species and already have commercial products. In these countries, fishes destined to the human consumptionthat have been anesthetized with MS-222 need to pass through a depuration period of 21 days before being liberated for consumption. BZN has the same regulation than MS-222 in Norway (Zahl et al., 2012;Carter et al., 2011).
South Korea, New Zealand, Australia, Costa Rica and Chile have already regulated the use of anesthetics containing the active principle ISO. However in these countries the depuration period is not necessary provided that parameters indicated by producers are respected (Zahl et al., 2012;Tuckey et al., 2012).
The depuration period may be defined through the quantification of anesthetic residues and their metabolites after exposition of fishes to anesthesia (Stehly et al., 2000;Stehly et al., 1995). All these results require methodologies that involve procedures of sample preparation like homogenization, extraction and purification for the definition of residual period, besides evaluating the capacity of anesthetics depuration by fishes (Meinertz et al., 2009;Meinertz et al., 2006;Meinertz et al., 1999).
Sample preparation may represent until 75% of the time spent in an analysis and it consists of an important step for the development of analytical methods. The combination of extraction procedures with purification and preconcentration like solid-phase extraction (SPE) and liquid-liquid extraction are necessary due to the complexity of the involved matrix (Baniceru et al., 2011).
This work aimed to define methodologies of sample preparation and residual quantification of BZN, MS-222 and CO in fillets of two species of tropical fishes, Nile tilapia (Oreochromis niloticus) and cachadia (Pseudoplatystoma reticulatum x Leiarius marmoratus), submitted to anesthesia in different dosages and depuration times using methodologies for extraction and quantification via UPLC-DAD.

Experimental design
Two different analytical methodologies were performed in order to evaluate three anesthetics on fish species. For the samples composition a factorial model with 30 treatments (2x3x5) was adopted. The factorial consisted in the interaction of anesthesia of the two studiedspecies with three anesthetics (CO, BZN and MS-222) used in five dosages (Table 1). Each treatment was applied on four repetitions (n=4), which were evaluated in four different depuration periods (0 h, 12 h, 24 h and 48h). S e p t e m b e r 0 5 , 2 0 1 4

Samples
The samples were obtained by slaughtering the fish through a section of the spinal cord after being stunned on ice. All experiments were submitted to the Ethics Committee on Animals Usage (CEUA-UFLA) and approved under protocol 025/2013. Composed of fillets of two fish species, Nile tilapia (Oreochromis niloticus) with mean weight of 456.9 ± 113.1 g and cachadia (Pseudoplatystoma reticulatum x Leiarius marmoratus) with mean weight of 345.6 ± 80.98 g. Fishes were submitted to five dosages of the anesthetics CO, BZN and MS-222 (Table 1) following procedures described by Okamura et al. (2010) in which dosages and time of permanence in anesthesia were optimized according to the ideal time of induction and recovery. Four depuration periods were adopted (0 h, 12 h, 24 h and 48h), before the fish slaughter and filleting. Fillets were frozen (-20 ºC) and kept this way until the sample preparation.

Extraction
For the evaluation of analytes three procedures of homogenization were used after fillets defrosting using grinder, dry ice and maceration besides two extractive solutions, acetonitrile (100%) and solution 50:50 (v/v) methanol:bufferMcIllvaine pH 4.4 (11.823 g of citric acid monohydrate and 16.4312 g of sodium phosphate dibasic dihydrate in 1 L of water type 1).  (2006,1999) procedure was modified: 5 g of macerated fillets were put in tubes of 50 mL and 5 mL of extractive solution were added. It was stirred during 5 min in a Vortex Mixer (Vixar ® ) and the extract was centrifuged at 3400 rpm during 5 min. This procedure was repeated three times. Aliquots were grouped, concentrated by evaporation until half of the volume and diluted with water at 25 mL in order to perform the SPE procedure.
For BZN and MS-222 the Scherpenisse and Bergwerff (2007) method was modified: 5 g of fillet were stirred mechanically (Quimis ® -2501) for 10 min using 5 mL of extractive solution and centrifuged at 340 rpm during 5 min. This procedure was repeated three times. Aliquots were grouped, concentrated by evaporation until half of the volume and diluted with water in order to perform the SPE procedure.

Solid-phase extraction
The SPE procedure was performed in Manifold Vac Elut 12 (Agilent ® ) coupled with a vacuum pump (Prismatec) using SPE cartridges Discovery® DSC-18 (3 mL; 500 mg). Cartridges were previously conditioned with 6 mL of methanol for CO samples and with 3 mL of methanol followed by McIllvaine buffer for samples containing BZN and MS-222.
After conditioning, samples were submitted to extraction under a flow of two drops per second. Analytes were eluted from cartridges by adding 1 mL of acetic acid (1%) followed by 1 mL of methanol. Aliquots (2.5 mL) were filtered in membrane PTFE (13 mm d; 0.45 µm; Millex, Millipore).

Selectivity and linearity
The selectivity of method was evaluated by means of the analysis of tilapia and cachadia samples lacking the analytes of interest. Matrices (n=6) of both species without analytes were fortified with analytical standards in five concentrations (0.0; 0.5; 1.0; 1.5 and 2.0 times the LD). [21] S e p t e m b e r 0 5 , 2 0 1 4 The selected wave lengths for analysis were: BZN (280 nm), MS-222 (315 nm), ISO 261 nm and EUG and MET in 275 nm. Analytical curve was prepared from dilutions of stock solution (1 g L -1 ) in methanol.

Limits of Detection and Quantification
Limits of detection (LD) and quantification (LQ) for each compound were obtained from parameters of analytical curve and the equations (1) and (2).
in which s is the estimate of standard-deviation of the regression equation and S is the slope or angular coefficient of analytical curve (Snyder et al., 1997).

Accuracy and recovery
The

Statistical analysis
Residual means obtained on repetitions of each treatment were submitted to variance analysis through the software SISVAR (5.1 Build 72). Variables that presented significant differences were compared through SNK (Student-Newman-Keuls) test considering α = 0.01 (Ferreira, 2011).

Results and discussion
The methodologies of quantification proposed by Scherpenisse and Bergwerff (2007) and Meinertz et al. (2009;2006) involves homogenization of fillets using a grinder and dry ice, respectively. Maceration resulted on a satisfactory performance in relation to the recovery of analyzed analytes ( Table 2) on matrices applied to this work, besides it has decreased costs with reagents, equipment and time spent per analysis, since a total of 30 samples per day were processed comparing with 17 made by Meinertz et al. (1999).

Table 2. Analytical parameters for anesthetics quantification
Anesthetics Linearity   The method selectivity for anesthetics was similar for both fish species. CO was analyzed in two wave lengths in which each compound presented higher absorbance, because the residual quantification requires the selection of wave lengths with higher responses in order to ensure the correct quantification. Thereby, 275 nm was selected for EUG and MET and 261 nm for ISO.This method selectivity allowed the simultaneous determination of all used anesthetics and consequently provided the possibility to apply such technique in residual analyses of matrices submitted to anesthesia compounded by more than one active principle. Nowadays the combination of different anesthetics and analgesics has been used in order to decrease toxic effects, therefore studies regarding the residual period of such anesthetics are needed (Mettam et Table 3. S e p t e m b e r 0 5 , 2 0 1 4  On samples submitted to treatments with CO it was possible to determine EUG in both fish species, while ISO only was detected. This is due to the mean composition of CO (~85% EUG), thus EUG suppresses other constituents in lower proportions, which did not attain quantifiable and detectable values After anesthesia (0 h) the mean of residues of the five BZN dosages was 1,46 µg g -1 on tilapias, value significantly superior to EUG (0.514 µg g -1 ) and to MS-222 (0.397 µg g -1 ). Different results were observed for cachadias, in which BZN (2.6 µg g -1 ) and MS-222 (2.17 µg g -1 ) presented residual means superior to EUG (0.987). BZN provided higher residual values right after anesthesia both for tilapia and cachadia. This may be explained by the rapid absorption and low solubility in water (approximately 250 times lower than MS-222), thus justifying its prohibition by Food and Drug Association (FDA) that allows using the MS-222 isomer (FDA, 2011;Burka et al.,., 1997). CO (EUG) did not present differences in relation to MS-222 in tilapias, while in cachadia values were inferior to MS-222 and BZN.
When comparing residual means among the five dosages applied, there was no significant difference for most of the treatments, except MS-222 in tilapias for which the lowest dosage resulted in higher residual concentration. This is associated to the time that the fish was exposed to the anesthetics, once when lower dosages are used more time is required until the fish becomes anesthetized, what increases the time of exposition and consequently the absorption of anesthetics by the fish.
It is important to highlight that, contrarily to other studies involving residual analysis fishes evaluated in the present work kept less than 3 minutes immerse in anesthetics bath (considered the ideal time for anesthesia) ( When comparing mean values of residues between species, overall cachadia retained higher concentrations than tilapia, presenting higher concentrations of MS-222 on the fillets, while tilapias retained more BZN. This is probably associated to differences on the species biology, tilapias present scales and cachadia has leather skin, furthermore they have different food habits, metabolism and fillet compositions, what may have hampered the residue elimination. After 12 h no more quantifiable values were obtained, except for the dosage [1] of BZN in cachadia. This concentration may have presented residual mean above LQ due to the longer time of induction, thus increasing the time of fish permanence in contact with the anesthetic. S e p t e m b e r 0 5 , 2 0 1 4 Samples related to periods of depuration did not present detectable values from 24 h on and this is the period recommended by Stehly et al. (2000) for BZN in rainbow trout (Oncorhynchus mykiss). Kildea  Brazilian laws still do not regulate the depuration time required by fishes anesthetized on aquiculture. However producers and researchers need to attempt in relation to the use of international parameters without considering specific demands of the animal. This evidences the importance of determining the depuration times required by fishes submitted to ideal anesthesia, thus allowing defining an operational protocol that ensures safety for the fish farmer, the fish and finally for the consumer.

Conclusion
Methodologies proposed in the present work were efficient on quantifying residues of the anesthetics CO, BZN and MS-222 in Nile tilapia and cachadia, with BZN presenting the highest mean residual value for tilapia and CO with the lowest mean value for cachadia. Between the species, cachadia (in which BZN and MS-222 values were higher than MS-222) accumulated more residual concentrations than tilapia. Nile tilapia accumulated residues of BZN on fillets, even after 12 h of depuration, as well as cachadia. There was no difference among dosages, except MS-222 for tilapias. Low dosages of anesthetics increases time of induction and consequently the time of fish permanence immerse on the anesthetics, thus contributing with higher residual concentrations on fillets.