Florfenicol Depletion in Water, Amine From Fillet Tissue After Feeding Aquaflor to Tilapia

Introduction

Streptococcus iniae, a Gram-positive bacterium, causes substantial mortality in tilapia, Oreochromis spp., especially among fish cultured in recirculating or intensive flow-through systems. Worldwide annual economic loss as a result of S. iniae-associated mortality in tilapia has been estimated to be about US $100 million.1 Consequently, MSD Animal Health is seeking US approval of Aquaflor, a feed premix containing the broad-spectrum antibacterial agent florfenicol (50% w/w; Figure 1A), for treatment of S. iniae in tilapia. Use of trade or product names does not imply endorsement by the US government.

Aquaflor was recently approved in the US at a dose rate of 10 mg/kg bodyweight per day (BW/day) administered in feed for 10 days to control mortality due to enteric septicemia in catfish (2005) and coldwater disease and furunculosis in trout (2007); it was conditionally approved (Aquaflor-CA1) to control mortality due to columnaris disease in catfish (2007).

Globally, Aquaflor, which is also marketed as Aquafen and Florocol in some regions, is registered for use in more than 20 countries including Norway (1993); Chile (1995); Canada (1997); the United Kingdom (1999); Ecuador and Venezuela (2005); Colombia (2006); and Brazil, Costa Rica, Vietnam and China (2007) to control various susceptible pathogens in a variety of commercially important freshwater and marine species.

**1A: Florfenicol [R-(R*,S*)-2,2-dichloro-N-[1-fluoromethyl-2-hydroxy-2-(4-methylsulfonylphenyl)] ethyl acetamide] is the active ingredient of Aquaflor.
1B: Florfenicol amine [R*,S*0]-?-(1-amino-2-fluoroethyl)-4-(methylsulfonyl)-benzenemethanol]) is the marker residue of florfenicol.

We evaluated the depletion of florfenicol amine (FFA) a marker of florfenicol (FFC) residue from tilapia fillet and the decline of FFC following FFC-medicated feed administration at a nominal dose rate of 20 mg/kg BW/day to fish reared in a recirculating aquaculture system (RAS). The objective of the study was to develop the marker residue depletion data needed to allow FFC administration at a proposed maximum dose of 15 mg/kg BW/day for 10 consecutive days.

Florfenicol Depletion

FFC distribution, metabolism and depletion following dosing at 10 mg/kg BW/day has been well characterized in a variety of fish. FFC was similarly distributed in freshwater- or seawater-acclimated tilapia with the maximum concentration occurring 2 to 24 hours post-dosing, depending on the tissue. In Atlantic salmon, FFA (Figure 1B) was identified as the primary metabolite of FFC in muscle. Muscle (skin-on fillet), by regulation, is considered the edible tissue of most fish. FFA was subsequently selected as the marker residue of FFC administration because it is the primary FFC metabolite, and other lesser metabolites (and FFC) are converted to FFA through acid hydrolysis.

Monitoring total FFA concentration (FFA + acid-hydrolyzed FFC and metabolites) in the target tissue thus provides a conservative estimate of FFC residues and enables calculation of a conservative withdrawal period. Although FFC metabolism data were not available for tilapia, FFA was assumed to be the marker residue since it is the marker residue in cattle, swine, sheep, poultry, catfish, salmon and trout.

Data from residue depletion studies are used to calculate a withdrawal period for a drug, which is the time required for the animal to deplete the drug residue to a level that is considered safe for human consumption. Regulatory agencies estimate the safe concentration or maximum residue level (MRL) by combining an acceptable daily intake (ADI) level (from toxicology data) with a standard human mass estimate and a consumption factor (an estimate based on the mass of residue-bearing tissue consumed).

For FFC, the 10 ?g/kg ADI is multiplied by a standard human weight (60 kg), then divided by a consumption factor (in fish, a standard mass [300 g] of skin-on fillet [muscle] is used), resulting in a tolerance of 2 ?g/g; the European Agency for the Evaluation of Medical Products and the US Food and Drug Administration have applied an additional safety factor and established the MRL for Europe and the US at 1 ?g/g.

Materials and Methods

Commercial tilapia culture is principally focused on the rearing of phenotypic males produced by the administration of feed containing 17?-methyltestosterone (MT) to tilapia fry. Tilapia used in the study were a mix of MT gender-reversed females (phenotypic males) and genetic males of the two most commonly cultured tilapia strains, pure Nile tilapia (O. niloticus x O. niloticus) and hybrid tilapia (O. niloticus x O. aureus).

Aquaflor-medicated premix was used to prepare the medicated feeds. The feeds used were assayed for FFC content by high-performance liquid chromatography15 (HPLC) before and after dosing (Table 1). The non-medicated feed was analyzed to determine proximate nutrient content as well as inorganic or organic contaminants (Eurofins Scientific Inc., Des Moines, Iowa, USA). Non-medicated feed and feed medicated with Aquaflor (2.667 g FFC/kg) for the residue depletion study (extruded 4.8 mm floating pellets) were prepared according to standard procedures at Delta Western Research Center (Indianola, Mississippi, USA). Aquaflor premix was added to the medicated feed during mixing, prior to extrusion.

Nile and hybrid tilapia (mean weight = 447 56 g) were obtained from a commercial farm. Tilapia were held in a commercial RAS (Aquatic Eco-Systems Fish Farm II) consisting of twin ~1,900 L (500 gal.) polyethylene tanks, mechanical filters (clarifier and suspended solids filter) and a biological filter; there was ~3,350 L (885 gal.) total system water.

The RAS biofilter was inoculated with commercial biofilter bacterial inoculum and allowed to operate for ~6.5 months with fish present before FFC administration. These fish were removed, and the tilapia used for testing were stocked into the RAS 38 days before FFC administration. Temperature was maintained at 27 C to 27.6 C (80.6 F to 81.7 F). Waste solids were removed once daily ~1 hour before feeding, and concurrent with solids removal, a portion of the RAS water was removed (acclimation 6-11%; dosing and post-dosing 5-8%) and replaced with temperature-adjusted well water (at ~22 C/71.6 F). Water chemistry (temperature, dissolved oxygen, pH, total ammonia, nitrite and nitrate) was determined once daily prior to tank cleaning. Water hardness and alkalinity were determined weekly. Alkalinity was maintained at >150 mg/L as CaCO3 by occasional addition of sodium bicarbonate. A single water sample was analyzed for metals and volatile and semi-volatile organics (Davy Laboratories, La Crosse, Wisconsin, USA). No contaminants at levels of concern were identified.

Non-medicated feed was offered at a rate of 0.25%-1% BW/day during the 38-day acclimation period; the feed rate was 0.75% BW/day for the last 11 acclimation days and remained constant through the remainder of the study, including the dosing and post-dosing periods. Three equal feed portions were offered each day with ~4 hours between each feed administration. Daily feed consumption was estimated during the dosing period.

¹ Bodyweight
² LOQ = limit of quantitation (0.0002 g/kg)

*Determined by high-performance liquid chromatography in feed samples collected at the start and end of dosing for residue depletion studies

Five fish from each tank were sampled 4 days prior to dosing to obtain control fillet tissue. Fish were indiscriminately removed and sacrificed; the fish were then scaled and skin-on fillets were collected, individually bagged and stored at <-70 C (-94 F).

Tilapia (n = 209) were offered feed medicated with Aquaflor at 0.75% BW/day (nominal dose = 20 mg FFC/kg BW/day) for 10 consecutive days. The estimated delivered dose was calculated from the estimated feed mass consumed, feed FFC concentration and the total fish mass at terminal sampling. Twenty fish (10 per tank) were indiscriminately removed from the RAS tanks on days 0.04, 0.5, 1, 1.5, 2, 3, 4, 5 and 10 post-dosing, and skin-on fillets were collected as previously described.

Water samples for FFC analysis were collected: (1) concurrent with the control fillet collection, (2) prior to tank cleaning during the dosing period (~1 hour prior to the first daily feeding), (3) just prior to the second and third daily feeding intervals, (4) 4 hours after the third daily feed interval and (5) concurrent with tissue collection during the post-dosing period, except the 0.04-day post-dosing fillet collection. At each collection interval, one water sample (~50 mL) was taken from the RAS clarifier and from the RAS suspended solids filter. Each sample was hand-mixed and syringe-filtered (Durapore [PVDF, 0.45 ?m] membrane, Millipore, Billerica, Massachusetts, USA) in ~2-mL aliquots into HPLC vials then stored at ?-20 C (?-4 F) until analyzed.

FFA concentrations were determined using a method validated for FFA in tilapia fillet tissue at MPI Research, Inc. (State College, Pennsylvania, USA). The method involved converting all FFC residues to FFA by acid-catalyzed hydrolysis. Fillet tissue was hydrolyzed by adding 6N hydrochloric acid then held for approximately 2 hours at 95 C to 100 C (203 F to 212 F). The tissue-hydrolysate was extracted with ethyl acetate and centrifuged. The aqueous hydrolysate was retained and adjusted to pH 12.5 or greater with 30% (w/w) sodium hydroxide solution. The pH-adjusted solution was adsorbed from 45 to 60 minutes onto a Varian Chem Elut CE120 sorbent column (Varian, Inc., Palo Alto, California, USA) then eluted with methylene chloride. The methylene chloride eluates were evaporated to dryness, dissolved in 10 mM potassium phosphate buffer (pH 4.0, 1% [v/v] acetonitrile), filtered (0.2 ?m) and then analyzed by HPLC using UV detection at 220 nm. The method quantitation limit (LOQ) was 0.05 ?g/g.

FFC concentration was determined in water samples using a validated determinative procedure capable of quantifying FFC from 10 to 5,000 ng/mL and up to 20,000 ng/mL after dilution. FFC concentration was determined by ultra-pressure liquid chromatography with mass spectrometric detection on an atmospheric pressure ionization interface. Water samples were fortified with FFC-d4 as an internal standard and analyzed directly. Ionic transitions of 356 to 185 m/z and 360 to 189 m/z were monitored for FFC and the internal standard, respectively. The method LOQ was 10 ng/mL.

The residue depletion profile of FFA in skin-on fillet of tilapia following withdrawal from the medicated diet was estimated by log-linear regression.16 The fitted loglinear regression model ([Ln (ppm)] = 2.3215 Ln (ppm) + 0.0151 Ln (ppm)/hour x hours post-dosing) R2 was 0.8271. The withdrawal period was defined as the time when the tolerance limit of the residue concentration was at or below the 1 ?g/g MRL. The tolerance limit was set as the 99th percentile of the potential residue level at 95% confidence.16 The withdrawal period was, therefore, equivalent to the time at which the tolerance limit was less than 1 ?g/g. Analyses were considered significant if P < 0.05.

Results

The mean minimum daily delivered doses were 19.4 mg/kg BW/day for tilapia in tank 1 (range 19.3 to 19.57) and 19.8 mg/kg BW/day for tilapia in tank 2 (range 19.7 to 20.0) or 97% to 99% of the target dose. Fish consumed 100% of the feed medicated with Aquaflor that was offered during the 10-day dosing period, similar to those of non-medicated feed during the acclimation and post-dosing periods. FFA concentrations in tilapia fillets are summarized in Table 2. Representative analytical standard, control and treated tissue chromatograms are presented in Figure 2.

Mean FFA concentration in tilapia fillets exceeded the MRL during the post-dosing period until the last post-dosing fillet collection (10 days post-dosing) when all fillet concentrations were below the MRL. Mean FFA concentrations steadily declined during the post-dosing period. Fillet FFA concentrations were similar between the two study tanks throughout the postdosing period. FFA level depleted rapidly to below the MRL, and the calculated tolerance limit was below the MRL at 10 days after withdrawal of the medicated feed (Figure 3).

* FFA concentration in fillet tissue from Nile and hybrid tilapia following administration of feed medicated with Aquaflor as the sole ration for 10 consecutive days. Standard deviations are in parentheses and the range is in brackets. Only samples above the FFA quantitation limit of 0.05 ?g/g were included in summary calculations.

FFC levels in RAS water before dosing were
Unionized ammonia-nitrogen levels were <0.02 mg/L NH3-N during the acclimation, dosing and post-dosing periods. Nitratenitrogen levels fluctuated in RAS from between 7 to 123 mg/L with mean concentrations of 67, 81 and 75 mg/L during the acclimation, dosing and postdosing periods. Nitrite-nitrogen levels occasionally exceeded the safe upper limit of 2.0 mg/L but only during the acclimation period; nitrite levels were <0.9 mg/L during the 8 days prior to dosing. Nitrite levels steadily increased during the dosing period until peaking on dosing day 8 at 1.76 mg/L (Figure 5). This concentration increase was apparently not associated with any effect on the RAS biofilter but rather the inadvertent buildup of biofilm material in the water supply lines to the biofilter from the RAS tanks as nitrite levels rapidly decreased during the remainder of the dosing period and during the postdosing period after the water supply lines were flushed. Nitrite levels again spiked during the post-dosing period then dropped after the RAS biofilter supply lines were flushed (Figure 5).

* 2A: Extract from control tilapia fillet tissue sample
2B: Extract from control tilapia fillet sample fortified with florfenicol amine (FFA) at 2 ug/g
2C: Extract (2.33 ug/g FFA found) from fillet tissue taken from a fish at 5 days post-dosing

Discussion

Fish readily consumed feed medicated with Aquaflor. There was no reduction in feed consumption during the dosing periods. There do not appear to be any palatability concerns regarding feed medicated with Aquaflor.

FFC concentration increased in RAS water during the dosing period then gradually decreased during the post-dosing period. The apparent concomitant increase in nitrite concentration in the RAS does not appear to correlate to RAS-water FFC concentration. Rather, the increase in RAS nitrite concentration was apparently due to microbial growth in the biofilter water supply lines restricting water flow to the RAS biofilter, as flushing those supply lines was followed by a rapid decline in RAS nitrite concentration. Some antibiotics (erythromycin, oxytetracycline) have been found to negatively affect denitrification in aquatic systems, albeit at levels much higher than observed in the present study.

FFC is known to be rapidly and completely distributed in fish4-5,7-8,10-11,18 during dosing and to rapidly deplete from tissues6,9,17,19 after withdrawal from medication, which are excellent traits for use in food fish to control susceptible bacterial infections. However, its rapid elimination means that tissue FFC levels will rapidly decrease after fish are withdrawn from the medicated feed. While minor differences in study design (e.g., feeding procedure, test temperature, feed rates) preclude direct comparison between FFA-residue depletion studies conducted with tilapia fed FFCmedicated feed while held in flow-through systems17,19 and this study, there do not appear to be substantial differences between the depletion of FFA from skinon fillet of tilapia fed FFC-medicated feed whether fed in a RAS or in a flow-through tank.

This rapid clearance indicates that there is likely to be only a short post-treatment therapeutic effect associated with Aquaflor administration, such as when FFC levels are at or above the bacteria MIC. Without concomitant steps by the farmer to reduce disease transmission (e.g., water disinfection, improved husbandry, vaccination), re-infection of treated fish is possible.

As with other antibiotics, Aquaflor should not be expected to eliminate the need for proper husbandry practices but should be regarded as one effective tool in the management of disease in aquaculture.

Conclusions

FFA, a marker for FFC residue, depletes rapidly in the skin-on fillet tissue of tilapia following the withdrawal of feed medicated with Aquaflor. When dosed at 19.62 mg FFC/kg BW/day for 10 days, FFA detected in the skin-on fillet of tilapia depleted to less than the MRL in all samples collected 10 days after the end of dosing. Based on these residue data and our interpretation of published regulatory guidance, FFA should deplete from tilapia treated with Aquaflor at 1.31 times the maximum proposed dosage (15 mg/kg BW/day for 10 consecutive days) to a level safe for human consumption 11 days after treatment.

October 2012