Mixing and clean-out properties of sulfamethazine and carbadox in swine feed

Two forms of sulfamethazine (pelleted and granular) and one form of carbadox (granular) were evaluated for their mixing and clean-out properties in replicated batches of swine feed prepared at the research feed mill at Kansas State University. Analysis of variance indicated that carbadox was incorporated into the feed more uniformly than sulfamethazine. Additional mixing beyond 1.5 minutes did not improve drug distribution (P>0.05). A paired-comparison T-test was used to compare mixing properties of drugs versus salt (the latter is commonly used to test mixer performance). We found the mixing properties of sulfamethazine and salt to be different (P 0.05). Ground corn was used to flush the mixer and conveying system between feed batches; then the mixer, leg, and sack-off bin were cleaned, and materials were assayed for drug carryover. Sulfamethazine was detected in the mixer cleanout material at concentrations of 12.6 ppm and 8.1 ppm for the granular and pellet forms, respectively. Carbadox carryover was not detected in mixer clean-out material. Detectable concentrations for all three drugs occurred in cleanout material from the leg and sack-off bin.

granular) and one form of carbadox (granular) were evaluated for their mixing and clean-out properties in replicated batches of swine feed prepared at the research feed mill at Kansas State University.Analysis of variance indicated that carbadox was incorporated into the feed more uniformly than sulfamethazine.Additional mixing beyond 1.5 minutes did not improve drug distribution (P>0.05).A paired-comparison T-test was used to compare mixing properties of drugs versus salt (the latter is commonly used to test mixer performance).
We found the mixing properties of sulfamethazine and salt to be different (P < 0.05), whereas those of carbadox and salt did not differ (P > 0.05).Ground corn was used to flush the mixer and conveying system between feed batches; then the mixer, leg, and sack-off bin were cleaned, and materials were assayed for drug carryover.Sulfamethazine was detected in the mixer cleanout material at concentrations of 12.6 ppm and 8.1 ppm for the granular and pellet forms, respectively.Carbadox carryover was not detected in mixer clean-out material.Detectable concentrations for all three drugs occurred in cleanout material from the leg and sack-off bin.

C
oncern over the safety of the food supply in the United Statesis paramount among consumers.The current good manufacturing practices (cGMPs)used to regulate animal feed production outline procedures to help assure that meat, milk, and eggsproduced from animals receiving medicatedfeeds contain no violative drug residues.Food and Drug Administration (FDA) cGMPsspecifythat "equipment shall be capableof producing medicatedfeed of intended purity and potency"!; this includes proper mixer performance.Mixer testing procedures are outlined by the American Societyof Agricultural Engineers (ASAE). 2 This procedure entails describing feed uniformity by calculating the coefficient of variation (CV) using salt assaysfrom 10 feed samples collected from the mixer.The cGMPsalso specify that "adequateprocedures shall be establishedand usedfor all equipment used in the production and distribution of medicatedfeeds to avoid unsafe contamination of medicated and nonmedicated feeds."! Sulfamethazineand carbadox are two antibacterial drugs widely used in swine production.Residuetolerances for thesetwo products in uncooked tissue are 0.1 ppm and 0.0 ppm, respectively.3Both products are classified as category-II drugs under the cGMPs;withdrawal times are 15 daysfor sulfamethazineand 10 weeks for carbadox.Both products are used to improve weight gain and feed efficiency,as well as to control or prevent bacterial diseases.4 The high rate of violations for tissue residues of sulfamethazine has concerned FDApersonnel for years.5TheFDAhas identified that a lack of sequencing,flushing, and cleaning of mixer equipment accounted for 25% of sulfamethazineviolations.As little as 1 ppm of sulfamethazine in feed, or 1/4 teaspoon of sulfa in a 908-kg batch of feed, can causeviolative sulfa residues.6Evaluatinghow these two medicated feed additives perform when flushing and cleaning out feed manufacturing equipment may help explain how feed becomescross-contaminated.
Studies examining the cause of cross-contamination in feed manufactured on-farm revealedthat powdered sulfamethazineincreased this risk compared to the granular form of the drug.7,8 The drug manufacturing industry developedgranular and pelleted forms of sulfamethazine'tohelp reduce cross-contamination.This effort, combinedwith a strongeducationcampaign by theUSDA and the FDA,reduced the violation rate in pork from 13%prior to 1978 to about 5% between 1980 to 1987.5Forsulfamethazinein swine, the current residue violation rate is less than 1%.9 Improper mixing and incorrect inclusion rates of medicatedfeed additives create the potential of tissue residue violations.The FDA has established acceptable assayerror ranges of 20% and 25% for complete feed containing sulfamethazine and carbadox, respectively.!OExceeding these error ranges presents a potential source of violative tissue residue, whereas inclusion rates below the establishederror range may reduce the efficacyof the drug to control diseaseand allow microbial resistanceto develop.6 This study was conducted to examine,the mixing and clean-out properties of two forms of sulfamethazine and one carbadox product to better understand the role that product form, mixing performance, and flushing/clean-out properties may play in producing quality feed.
Sulfamethazinewas included in the feed at a rate of 110 mg per kg (110 ppm) of feed, and carbadox was used at 55 mg per kg (55 ppm) of feed.Treatmentswere arranged in a completelyrandomized design with repeated measures taken at three mixing times and after transferring feed to 22.7-kg sacks.
Corn conforming to United States Grain Grading Standards for number 2 yellowcorn was ground to a particle size ranging between 550 and 700 microns using a Jacobson hammermillwith a 0.317 em (1/8 inch)-diameter screen.A 182-kgground corn placebo was passed through the mixing and sack-off system.Then, the mixer,leg, and sack-offbin were cleaned prior to mixingfeed for the study.The mixer and leg boot were cleaned using separate shop vaccuums.Adifferentvacuum filter was used for each drug treatment.Thesack-offbin was cleaned and the sample collected from this area by striking it with a rubber mallet and collecting the material that was knocked off by the resultingvibration.Feed consisting of corn (73.5% by weight) and soybean meal (22.2% by weight) was batched with a WisconsinElectric Manufacturing, Inc., system and emptied into a Sprout Waldron horizontal double-ribbon mixer.The micro-ingredients (monocalcium phosphate, limestone, lysine, vitamins, trace minerals, and salt) were added to the mixer by an Ablemicro-ingredientsystem.The medicated feed additives were applied by hand after the grain, protein, mineral, and micro-ingredientswere added to the mixer.
Mixing properties of the medicated feed additives were evaluated by sampling the mixer using a Seedburo Grain Probe (Chicago, Illinois) after 1.5, 2.5, and 4 minutes of mixing time.In order to reduce costs to incorporate replications, the CVswere computed based on eight rather than 10 samples.Following the mixing treatment, feed was conveyed to the sack-off bin and packaged into 22.7-kg capacity sacks, of which eight were sampled.Two flush treatments with 91 kg of corn followed each batch of feed.
The feed system was cleaned by the same procedures used in mill preparation.
Samples from the mixer, packaging, flush, and clean-out were split using a riffler and analyzed separately for salt and drug content.Salt analyses were performed using Quantab titrators (Elkart, Indiana).Drug assays for sulfamethazine and carbadox were performed by a commercial lab following Association of Analytical Chemists (AOAC)methods.12 The lowest detection limits in feed samples for these two assays are 5 ppm and 2 ppm, re-spectively.Triplicate assays were performed on all samples that were 30% outside the desired me"dication concentration following the first assay results.

Statistical analysis
Coefficient of variation, standard deviation, and mean measurements taken across the locations were calculated for each drug, replication, and mixing time using the Univariate procedure in SAS.12 Drug concentrations were analyzed on a proportional basis because the carbadox inclusion rate was half of the sulfamethazine inclusion rate.The GLM procedure in SASwas used to evaluate treatment effects for both the mixing and clean-out portions of the study.Main effects were separated using Fisher's least-signifIcant-difference technique, and interactions were analyzed using the least-signifIcant-difference among the least squares means produced by the GLM procedure.Variance components in the general linear model were evaluated using the VARCOMP procedure of SAS.A paired-comparison T-test was performed on the mean difference between drug mixing uniformity and salt mixing uniformity.

Mixing properties
Mixing properties compared among drugs were different (P<O.OI), whereas mixing time compared among drugs did not differ (P>0.05).Carbadox mixed well, as indicated by an average CVof 11.4% (Table 2).THe CV for pelleted sulfa was 30.4%, while the CVfor granular sulfa was 25.6%.
Increased mixing time after 1.5 minutes did not improve the uniformity of drug distribution in the swine feed (P > O.I).This suggests that some factor other than mixing time hindered sulfamethazine distribution in the feed.
Electrostatic properties of feed ingredients are reported to occur;13,14however, a paucity of information is available regarding the influence of static charge on mixing properties of feed ingredients.Ingredient carriers, oil, and grounding the mixer are used to reduce static cling.However, ingredients not directly in contact with the mixer may possess electrostatic charge.If static charge was the cause for non-uniform distribution of sulfamethazine in the feed, additional mixing would not rectify this problem.Further investigation to explain the cause for poor mixing performance should include measuring various physical properties of sulfamethazine, salt, and corn such as density, particle size, hygroscopicity, conductivity, and static charge during mixing or movement.
Mean assay values for each drug x mixing time combination (Table 2) indicate that the pellet form of sulfamethazine was present at a lower concentration (87.1 ppm) than the granular form (108.4 ppm) in the complete feed.Both sulfamethazine products were packaged as a Type-Bpremix at a concentration of 22 g per kg (10 g per lb), and assays of the premixes for drug concentration indicated that the granular and pellet forms contained 114%and 104%of the label, respectively.Thehigher mean for the granular sulfamethazineexplains why its CVwas smaller than that of the pellet form.The standard deviations for both products were similar, and the range between assays was about 27 ppm greater for the granular product.
The statistical components of variability for the two sulfamethazineproducts and carbadox were analyzedusing data from samples subjected to triplicate drug assays.Assayvariability was small relative to other components of variability in the experiment.The greatest variabilityoccurred between replications for the same sample site within each drug treatment.
The paired-comparison T-testbetween salt and drug CVsrevealed that carbadox did not differ (P>0.05) from salt with respect to distribution uniformityin the feed, whereas both forms of sulfamethazine displayed mixing properties that differed (P<0.05) from those of salt.Hasselberger15reported a similar observation with regard to mixing properties of chlortetracycline, penicillin, and sulfamethazinein feed concentrates.
Swine Health and Production-Volume 3. Number 5

Clean-out properties
We observed a drug x location interaction in feedclean-out/flush material(P<O.OI) (Figure 2).Drug concentrations in both the ground-corn flush treatments did not vary (P>0.05)among products.A trend, in which sulfamethazine was present in flush material whereas carbadox was not present at detectable concentrations, was established for both corn flush treatments.
The mixer clean-out samples displayed a similar trend as the ground corn flush with respect to drug carryover (Figure 2).Sulfamethazinecontent in mixer clean-out samples did not differ (P>0.05) between the pelleted (8.1 ppm) and granular (12.6 ppm) product, whereas carbadox «2.0 ppm) differed from the granular form of sulfamethazine (P<0.05).The highest sulfamethazine concentration (16.2 ppm) found in 1.1 kg of mixer clean-out material could result in a contamination of 32 parts per billion in the subsequent 454-kg batch of feed.This is below the 1 ppm concentration that can lead to violativetissue residues.s The feed collected from the boot of the leg contained higher drug concentrations than the flush and mixer clean-out material (P<0.05).No difference was observed among the three drug products.The highest concentration of sulfamethazinecarryover (37.8 ppm pellet and 39.9 ppm granular) occurred in the material collected from the boot of the leg.Sincethis is a dead spot in the feed conveyingsystem,the onlywayto remove carry-overmaterial is to clean the boot (physical removal).A high concentration of drug at this location is not undesirable, since the pellet and granular products were designedto flush from the system.
The concentration of drug in material collected from the sack-off bin varied dramatically between products.The concentration of carbadox in the sack-offbin was approximately86% of the inclusion rate (43 ppm) compared to the sulfamethazine products, which were present at 7 ppm (6.4%) for the pelleted form and 23.5 ppm (21.4%) for the granular form.Clean-out material from the sack-offbin consists of fine, dust-likeparticles.Perhaps carbadox possesses similar dust-like properties and separates from the feed at the sack-off bin.The presence of a high drug concentration in the sack-offbin appears particularly hazardous as it is likely to result in product cross-contamination.The high carbadox concentration in the sack-offbin may also explain why it was not present in the ground corn flush.
Veterinarians should be aware of the different properties that medicated feed additives possess with respect to mixing and clean-out performance.In light of these results, it is imperative that the cGMPsare followed to avoid cross-contamination and violativetissue residue.Veterinarianscan play an integral role in cautioning producers who mix their own feed of the potential hazards associated with medicated feed additivesand the importance of good manufacturingpractices.

Im~lications
. The two medicated feed additives containing sulfamethazine did not incorporate uniformlyin the feed.The cause of the poor mixingperformance of sulfamethazinewas not discovered; however,assay variabilitywas eliminated as a primary source of variation.
. Salt is not necessarily a good model for mixing and clean-out properties of in-feed medications.A paired-comparison T-testof the CVsfrom salt assays and CVsof the drug assays showedthat sulfamethazinedid not perform in a comparable manner to salt during the mixingprocess.
. Flushing the feed mixing, conveying, and sack-off systems twice with ground corn did not eliminate drug carryover.The granular form of sulfamethazine was present in mixer cleanout material at an average concentration of 13 ppm.Carbadox was present in the clean-out material from the sack-off bin at a concentration nearly equal to the inclusion rate.
TJH, KCB: Department of Grain Science and Industry; TML: Department of Statistics, Kansas State University, Manhattan, Kansas, 66506-220 I, Paper No. 94-572-} of the Kansas Agricultural Experiment Station.