Material and Methods

Animals—The study protocol was approved by the Institutional Animal Care and Use Committee at North Carolina State University. Six healthy dairy cattle (5 Holsteins and 1 Jersey) between 3 and 7 years old that were housed at the university dairy farm were used in the study. An analysis of herd records indicated that none of the cattle had been treated with antimicrobials during the current lactation. Cows were in their second, third, or fourth lactation, with days in milk for the current lactation ranging from 305 to 400 days. Amount of milk produced up to the time of enrollment in the study ranged from 11,757 to 15,832 kg (25,865 to 34,830 lb), and mature equivalent 305-day milk production ranged from 10,727 to 13,604 kg (23,599 to 29,929 lb). Cows produced between 13.6 and 31.8 kg (30 and 70 lb) of milk/d throughout the study, with mean daily production ranging from 16.4 to 31.4 kg/d (36 to 69 lb/d). Cows were housed with the remainder of the herd throughout the study and were fed a total mixed ration containing corn silage, alfalfa silage, whole cottonseed, ground corn, soybean meal, soybean hulls, and minerals. All 6 cows were enrolled in the study at the same time.

Experimental protocol—Baseline (pretreatment) milk samples were collected following the morning milking, and 1,200 mg (4 mL) of tilmicosin^a was then infused into the left front and right rear glands of each cow (total, 2,400 mg of tilmicosin/cow). Approximately 12 hours later, foremilk samples were collected and cows were milked out by machine. An additional 1,200 mg of tilmicosin was then infused into the left front and right rear glands. No tilmicosin was ever infused into the right front or left rear glands. Foremilk samples (10 mL) were subsequently collected from all 4 glands of each cow every 12 hours (ie, immediately prior to milking) for the next 40 days. Milk samples were immediately frozen after collection and stored at −20°C until analyzed. At the end of the study, cows either were allowed to stop lactating as scheduled or were culled from the herd. No milk obtained from any cow in the study during the remainder of the lactation was sold for human consumption.

Determination of tilmicosin concentration—Concentration of tilmicosin in milk samples was determined by means of solid-phase extraction followed by ultraperformance liquid chromatography–mass spectrometry. In brief, milk samples were thawed and vortexed thoroughly, then centrifuged at 4°C in a fixed-rotor centrifuge at 1,500 × g for 20 minutes. Taking care to avoid any fat contamination, approximately 3 mL of the skim milk was removed with a disposable pipette, placed in a clean tube, and frozen until analyzed.

For analysis, skim milk samples were thawed and vortexed, then loaded into a preconditioned solidphase extraction cartridge^b that had been equilibrated with methanol^c (1 mL of methanol/1 mL of water). The cartridge was washed with 1 mL of a 5% methanol:95% water solution and dried under high vacuum for 1 minute and then eluted with 1 mL of methanol. The eluate was evaporated to dryness under nitrogen^d and reconstituted with a known volume of mobile-phase solvent. A 0.25-mL aliquot of the reconstituted eluate was then filtered with a syringeless filter–vial device,^e and 2.0 PL of the final sample was injected onto a C₈ column^f equilibrated with 25% acetonitrile:75% H₂O mobile phase with 0.2% acetic acid in each constituent. Tilmicosin concentration was determined in the positive electrospray mode by quantification of ion 435.^g

Owing to the large range of tilmicosin concentrations, the amount of each sample loaded on the solidphase extraction cartridge ranged from 0.05 to 1.0 mL and the final eluate was reconstituted in either 0.5 or 1 mL of mobile-phase solvent. These variations were needed to maintain the tilmicosin concentration in reconstituted samples within the calibration range of the instrument. It was found that when smaller volumes (ie, < 0.25 mL) were used, it was necessary to dilute the sample with water to obtain more reproducible results, as the added volume ensured better contact with the sorbent bed. Experiments revealed that addition of 0.25 to 10 mL of water to 0.25 mL of a milk sample spiked with tilmicosin did not result in substantial differences in the measured concentration. For samples with the highest tilmicosin concentrations, 0.05 mL of the skim milk sample was diluted with 0.95 mL of water, and the eluate was reconstituted with 1 mL of the mobile-phase solvent.

Positive control samples were prepared by spiking whole milk with a stock solution of tilmicosin in methanol.^h These spiked samples were frozen and analyzed as described for study samples. Positive and negative control samples were analyzed alongside each set of study samples, and study sample concentrations were calculated by extrapolation from results for positive control samples. Initially, positive control samples with tilmicosin concentrations ranging from 0.01 to 5.0 Pg/ mL were analyzed, and results were found to be linear (R² > 0.99 for each data set). Further testing revealed that results were linear for samples with concentrations as high as 1,000 Pg/mL. The limit of quantification was determined to be 0.01 Pg/mL (10 ppb) during initial method validation.

Data analysis—Tilmicosin concentrations were reported on the basis of time after administration of the second 1,200-mg dose of tilmicosin. No significant correlation was found between the rates of depletion from the 2 treated mammary glands of an individual cow (R² = 0.017). Therefore, samples from each treated mammary gland were considered individually (n = 12). Descriptive statistics and graphs of tilmicosin concentration versus time were prepared with standard software._i Concentrations of tilmicosin in milk from treated and untreated glands were analyzed by means of noncompartmental methods.^j Data from treated glands were analyzed by use of an IV bolus model, and data from untreated glands were analyzed by use of an extravascular input model. Area under the concentration versus time curve from time 0 to infinity was calculated by means of the log-linear trapezoidal method; the terminal portion of the curve was estimated as Cn/O_Z, where Cn represents the last measured concentration and O_Z represents the slope of the terminal portion of the curve. Values for volume of distribution and clearance are reported only for the treated glands. Values for maximum concentration of tilmicosin in milk and time to maximum concentration were obtained directly from the concentration versus time curves. The half-life of tilmicosin in milk was calculated as 0.693/O_Z. Owing to the small sample size, the 95% confidence interval was also calculated for the half-life of tilmicosin in milk from each individual gland. Concentrations less than the limit of quantification were not included in the analysis.

To determine a potential withdrawal interval following intramammary infusion of tilmicosin in lactating dairy cows, the tolerance limit method was used in accordance with the FDA's guidelines for determining a milk discard time.² Tolerance limit calculations were performed only on pharmacokinetic data from treated glands.

Results

Intramammary infusion of tilmicosin was well tolerated and did not result in any visible irritation or inflammation in the teats. Concentration of tilmicosin in milk from treated glands was quite high initially (Table 1; Figure 1), although there was considerable variation among glands. High concentrations of tilmicosin were also detected in milk from untreated glands. The half-life of tilmicosin in milk from treated and untreated glands was long, and clearance of tilmicosin from treated glands was low. Tilmicosin was detected in all treated glands for the entire 40-day study period and was detected in untreated glands for 30 to 35 days. The recommended milk withdrawal interval following intramammary infusion of tilmicosin, calculated on the basis of the tolerance limit method, was 82 days.

Concentration of tilmicosin in milk following intramammary infusion of 2 doses (1,200 mg each) in the left front and right rear glands of 6 lactating dairy cattle. Data represent mean concentrations for all 12 treated glands; error bars represent SD. For clarity, data were divided into the initial rapid phase of tilmicosin elimination (0 to 120 hours after administration of the second dose; A) and the slow phase of tilmicosin elimination (B).

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.245

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Concentration of tilmicosin in milk following intramammary infusion of 2 doses (1,200 mg each) in the left front and right rear glands of 6 lactating dairy cattle. Data represent mean concentrations for all 12 treated glands; error bars represent SD. For clarity, data were divided into the initial rapid phase of tilmicosin elimination (0 to 120 hours after administration of the second dose; A) and the slow phase of tilmicosin elimination (B).

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.245

• Download Figure
• Download figure as PowerPoint slide

Concentration of tilmicosin in milk following intramammary infusion of 2 doses (1,200 mg each) in the left front and right rear glands of 6 lactating dairy cattle. Data represent mean concentrations for all 12 treated glands; error bars represent SD. For clarity, data were divided into the initial rapid phase of tilmicosin elimination (0 to 120 hours after administration of the second dose; A) and the slow phase of tilmicosin elimination (B).

Citation: Journal of the American Veterinary Medical Association 234, 2; 10.2460/javma.234.2.245

• Download Figure
• Download figure as PowerPoint slide

Discussion

Results of the present study suggested that intramammary infusion of tilmicosin as an extralabel treatment for mastitis in lactating dairy cattle would result in milk residue violations for extended periods. Therefore, we believe that this practice should be strongly discouraged. If tilmicosin has been given by intramammary infusion in a cow, milk should be discarded for at least 82 days following administration of the last dose.

The presence of drug residues for such an extended period following intramammary infusion of tilmicosin was not unexpected. A previous study³ found that SC administration of tilmicosin to cows at the label dose (10 mg/kg [4.5 mg/lb]) resulted in detectable concentrations of tilmicosin in milk for 19 to 31 days; the limit of detection for the testing method used in that study was 25 ppb. Data from the European Union revealed that tilmicosin was rapidly distributed to the milk following SC administration, with concentrations in milk peaking at 6 to 8 ppm even though concentration in plasma peaked at only 0.13 ppm.⁴ The elimination half-life for milk following SC administration was 34 hours.⁴ Similar findings have been published for goats, with a milk-to-plasma tilmicosin concentration ratio of 12:1 following SC administration and drug detectable in milk for 11 days after administration of a single dose of 10 mg/kg, SC.⁵ In sheep, the half-life of the drug in milk (26.2 hours) is almost twice the half-life of the drug in plasma (15.4 hours).⁶ To our knowledge, there are no previous studies of the elimination kinetics of tilmicosin following intramammary infusion in lactating cattle; however, the European Union has minimal data on tilmicosin concentrations following intramammary infusion in cows immediately prior to the end of lactation, suggesting that tilmicosin could still be detected for at least 9 milkings after the end of the nonlactating period, which ranged from 38 to 72 days.⁴

The present study had to be terminated after 40 days even though tilmicosin was still detectable in milk from all 6 cows. Cows late in the lactation period were used in the study to minimize the volume of milk that had to be discarded. Because we did not anticipate that drug concentrations would persist for such extended periods, most of the cows in the study were either at or past the time that lactation typically would have ended by the end of the 40-day study period. Therefore, the tolerance limit method was used to calculate a withdrawal interval. The tolerance limit method provides an interval within which a given percentile of the population is expected to lie with a given degree of confidence. The FDA guidelines mandate the use of the 99th percentile for the population of interest and the 95% confidence level, and the withdrawal period is calculated as the time when the residue concentration can be expected, with 95% confidence, to be at or below the permitted concentration for 99% of the population. For purposes of the present study, the permitted concentration was set at 0.01 Pg/mL (10 ppb), which was the limit of quantification for the analytic procedure used and was well below concentrations of tilmicosin permitted in muscle (100 ppb) and liver (1,200 ppb). Drug elimination was assumed to be linear over time, and the variation in the natural logarithm of drug concentrations was assumed to be constant over time.

In the European Union, there is a maximum residue level of 50 ppb for tilmicosin in milk,⁴ and if we had used a permitted concentration of 50 ppb, our calculated withdrawal interval would have been 75 days. Currently in the United States, there is no residue tolerance for tilmicosin in milk, which means that any detectable concentration would constitute a residue violation. The commercial analytic system^k commonly used by milk processing plants can detect tilmicosin concentrations in milk as low as 20 ppb, and most liquid chromatography methods similar to the method used in the present study can detect tilmicosin concentrations as low as 10 ppb.^7,8 Therefore, we elected to use a permitted concentration at 0.01 Pg/mL (10 ppb) in our tolerance limit calculation.

The dosage of tilmicosin used in the present study was developed on the basis of extrapolations from dosages reported to FARAD by various practitioners. Intramammary infusion of tilmicosin constitutes an extralabel use of this drug and is not recommended. However, some veterinarians have suggested that extralabel use of this drug can be justified for treating Staphylococcus aureus mastitis in accordance with the principles of the Animal Medicinal Drug Use Clarification Act. Importantly, the first step in determining whether extralabel drug use is acceptable in food animals is to determine whether there currently is a drug labeled for that particular use that is effective for the given problem.⁹ Several antimicrobials have been approved for the treatment of S aureus mastitis in lactating dairy cattle, including amoxicillin, cephapirin, cloxacillin, erythromycin, hetacillin, and pirlimycin. Although it might be argued that none of these antimicrobials are consistently successful in curing clinical mastitis caused by S aureus, tilmicosin has not been proven to be effective in lactating cows either.¹⁰ In a previous study,¹⁰ for instance, cows with naturally occurring mastitis caused by S aureus were assigned to groups that received intramammary infusions of tilmicosin (300 mg/milking for 6 consecutive milkings) or parenteral injections of tilmicosin (3,000 mg, SC, q 24 h for 3 days). Almost all treated glands remained positive for S aureus 28 days after the end of the treatment period, and the authors concluded that tilmicosin was not effective for treating clinical mastitis caused by S aureus. Therefore, extralabel use of tilmicosin to treat mastitis in dairy cattle cannot be easily justified. Furthermore, the slow elimination of tilmicosin from the udder when administered by means of intramammary infusion is likely to result in milk residue violations. Therefore, we believe that use of tilmicosin in lactating dairy cattle should be strongly discouraged.