- Email: [email protected]
Residue Persistence in Sheep Milk Following Antibiotic Therapy
The Veterinary Journal 2003, 165, 84–89 doi:10.1016/S1090-0233(02)00173-9
Residue Persistence in Sheep Milk Following Antibiotic Therapy A. MOLINA , M.P. MOLINAà, R.L. ALTHAUS§ and L. GALLEGO
Departamento de Ciencia y Tecnologıa Agroforestal, E.T.S.I. Agr onomos, Universidad de Castilla-La Mancha, 02071 Albacete, Spain, Departamento de Ciencia Animal, Universidad Polit ecnica, 46071 Valencia, Spain, §C atedra de Biofısica, Facultad de Ciencias Veterinarias, Universidad Nacional del Litoral, 3080 Esperanza, Argentina à
SUMMARY Drug residues in milk supplies may have public health implications and can interfere in the manufacture of dairy products, such as cheese. In Spain, most ewe milk production is destined for cheese making, often using raw milk. This study analyses the main factors influencing antibiotic depletion time in lactating dairy sheep, 42 Manchega ewes were distributed into three groups, each receiving a different treatment (cephalexin intramammary infusion, penicillin G intramuscular, and oxytetracycline intravenous injections). During and after the recommended withdrawal period, milk samples were taken at each milking. A microbiological inhibition test (Brilliant Black Reduction, BRT) was used to screen all samples and antibiotic withholding times were established using a logistic regression model. The response to the BRT method in milk from individual ewes treated showed that the effect of the milking order was significant ðP < 0:001Þ with the three antibiotics. However the only influence on milk yield was with the intramammary treatment ðP < 0:005Þ. The BRT method was found to be very sensitive, particularly to the two b-lactamic antibiotics. Ó 2002 Elsevier Science Ltd. All rights reserved.
KEYWORDS: Dairy ewes; antibiotic residues; withdrawal period; screening methods. INTRODUCTION The presence of drug residues in milk supplies may have public health implications and are perceived by consumers as undesirable (Bishop & White, 1984; Booth & Harding, 1986; Mor etain, 1986; Brady & Katz, 1988; McEwen et al., 1991; Bencini & Pulina, 1997). Moreover, residues can interfere with the manufacture of dairy products, such as cheese. In Spain, most ewe milk production is used for cheese making, often using raw milk, and such ‘‘pure’’ sheep cheeses carry a Protected Denomination of Origin (DOP) which guarantees quality control. Mastitis is one of the most important problems in dairy cows and the major source of residues found in milk supplies (Booth & Harding, 1986; McEwen et al., 1991). In recent years, due to the intensification of milk production in small ruminants, there has been an increase in the frequency of mastitis cases also in dairy ewes and goats (Buswell & Barber,
Correspondence to: A. Molina, Departamento de Ciencia y Tecnologıa Agroforestal, E.T.S.I. Agronomos, Universidad de Castilla-La Mancha, 02071 Albacete, Spain Tel.: +34-967-599238; E-mail: [email protected]
1989). In sheep, treatment of mastitis and other diseases with pharmacological products is standard practice, and in many cases antibiotic milk contamination may be caused by unnecessary treatments carried out without a veterinary prescription, and with inadequate knowledge of the suitable dosage, administration route or depletion time of the antibiotic. One reason for this is because the majority of available treatments for mastitis are specific for cows, and withdrawal times are unknown in sheep. Indeed, no withdrawal times for milk from dairy ewes treated with antibiotics have been specifically established (Buswell & Barber, 1989). The European Union (EU) determines the limits for the presence of specified veterinary residues in milk. The antibiotic residues are defined by Council Regulation EEC 2377/90 (1990), although a number of amendments have subsequently been made, primarily to extend the list of compounds whose Maximum Residues Limits (MRLs) have been previously set. Detectable concentrations of antibiotic residues in milk supplies higher than the MRLs are illegal. Many drug residue screening tests are currently used for cowsÕ milk (Bishop & White, 1984; Booth & Harding, 1986; Laroque & Neville, 1986;
1090-0233/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved.
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
Brady & Katz, 1988; Seymour et al., 1988; Anderson et al., 1998), however, there are only a few evaluations of antibiotic residue screening tests for ewesÕ milk (Althaus, 1999; Althaus et al., 2001a) or goats’ milk (Zeng et al., 1996; Contreras et al., 1997; Zeng et al., 1998). Residue screening tests are important for daily antibiotic control, as well as for preventing antibioticcontaminated milk from entering the processing of milk products (Heeschen & Suhren, 1996). Persistence of antibiotic residues in milk is related to several variables such as daily milk production at the time of treatment, the type and amount of antibiotic used, and the disease state of the animal (Mor etain, 1981; Laroque & Neville, 1985; Seymour et al., 1988; Oliver et al., 1990; McEwen et al., 1991). The objective of this study was to determine the main factors that influence antibiotic depletion time in lactating dairy ewes receiving various antibiotic treatments using a screening method previously used to test sheep milk (Althaus et al., 2001a), the Brillant Black Reduction (BRT). The BRT method is a microbiological inhibition test widely used for the detection of residues in milk, using Bacillus stearotermophilus var. calidolactis. It is the reference method used in Spain in all quality control dairy laboratories for the detection of residues in cow, sheep, and goat milk. MATERIALS AND METHODS Animals Fifty-five Manchegan bred ewes (body weight range 55–80 kg, age 1–8 years) from the Manchegan herd in the experimental flock of the University of Castilla-La Mancha (Spain) were selected for this study. The ewes were randomly divided into three groups, each containing 14 animals roughly equal in body weight, age, and daily milk yield. The animals did not receive any kind of treatment before the experiment. The ewes were housed and fed the same as the rest of the herd; lambing took place in January and February, and mothers suckled their lambs for 35 days. After weaning, the ewes were machinemilked twice a day. The experiment took place in the fourth week of milking. Antibiotic treatment and sample collection Antibiotics were administered immediately after milking was completed. Each group of animals was treated with a different antibiotic; the first group was given three intramammary infusions per udder of a 5 mL suspension of 350 mg/mL cephalexin monohydrate with 35 mg/mL gentamicin sulphate (Max-
85
icef, Vetoquinol), at 24 h intervals. The second group was given two intramuscular injections of 1 mL/10 kg body weight of a suspension of 200.000 IU benzyl penicillin G procaine with 200 mg/mL dihydrostreptomycin (Dihidropen, Iven) at 24 h intervals. The third group was given one intravenous injection of 1 mL/10 kg body weight of 200 mg/mL of oxytetracycline (Oxiciven LA/200, Iven). The products were administered according to manufacturer’s instructions see Table I. After treatment, the ewes were machine-milked twice a day, the milk yield from each ewe measured and samples were taken. Further milk samples were collected at 12 h post treatment, and sampling continued during every milking session until double the time of the withdrawal period recommended by manufacturers. The samples (50 mL/milking session/animal) were collected in disposable plastic containers and kept at 5 °C until they were analysed. Brilliant Black Reduction Test Milk samples were analysed during the 24 h period after collection using the Brilliant Black Reduction Test (BRT), according to manufacturer’s instructions. One hundred microlitres of milk were added to ready-prepared individual BRT cups containing Bacillus stearothermophilus var. calidolactis. Milk shown to be free from antimicrobials was used as ‘‘negative control’’. For the ‘‘positive control’’, milk samples with 4 lg penicillin/kg were used. After a 1 h diffusion time at +4 °C, the BRT tests were floated in a waterbath and incubated at 64 1 °C for 3 h. Visual interpretation was carried out independently by three trained observers and evaluated visually as ‘‘negative’’, ‘‘doubtful’’ or ‘‘positive’’. Statistical analysis To evaluate the influence of the milking order and milk yield on the responses to the BRT inhibition microbiology screening test, the logistic regression model (Agresti, 1990) was used employing Stepwise option from the Logistic procedure of the SAS (SAS, 1998). Variables were analysed using the following logistic model: Lijk ¼ logit½Pijk ¼ b0 þ b1 ½MOi þ b2 ½Pj þ eijk where Lijk is the logit model; [Pijk ] is the probability for the response category (‘‘positive/doubtful and negative’’ or ‘‘doubtful/negative’’); b0 is the intercept; b1 ; b2 are the estimate parameters for the model; [MO]i is the effect of milking order (1, 2, 3, 4, 5, 6, 7, and 16); [P]j is the effect of dairy production and eijk is the residual error.
86
THE VETERINARY JOURNAL, 165, 1
The concordance coefficient was applied as rank correlation between the observed responses and predicted probabilities (SAS, 1998). RESULTS AND DISCUSSION Intramammary cephalexin infusion The application of the logistic regression model considered the effect of the milking order ðv2 ¼ 63:4393; P < 0:0001Þ and milk yield ðv2 ¼ 6:5582; P < 0:05Þ to be significant. The prediction model for calculating the frequencies of positive and doubtful samples, taking into account the milking order [MO] and dairy production [P], was L ¼ b0 0:8330½MO 0:00194½P ðconcordance coefficient : 95:2%Þ where, b0 ¼ 7:0413 to estimate the frequency of ‘‘positive/doubtful + negative’’ cases or b0 ¼ 7:6784 to estimate the frequency of ‘‘positive + doubtful/ negative’’. Fig. 1 shows the effect of the two variables, milking order and dairy production (in this case depending on the milk yield of ewes, 400, 600, and 800 mL/milking, respectively) on the frequency of ‘‘positive’’ and ‘‘doubtful’’ responses to the BRT method. A decrease in the frequency of positive cases was observed as time elapsed after the last application. This fact has also been observed in dairy cows treated with different antibiotics by the intramammary route (Oliver et al., 1990) to such an extent that the number of positive samples decreased after treatment increased. This finding was more pronounced in sheep with high milk production and there was a greater decrease in positive cases with time for ewes with a
Fig. 1. Intramammary cephalexin infusion. Frequency of positive and doubtful cases using Brilliant Black Reduction Test (BRT).
high milk yield. On the other hand, sheep with lower milk production took longer to eliminate the antibiotic. The amount of milk produced could influence responses to the BRT method since milk represents the main elimination route of the antibiotic when administered by the intramammary route (see Table I). Continued synthesis of the milk in the mammary tissues impedes the diffusion of this antibiotic into the blood. The effect of the level of milk production on the elimination of antibiotics has been reported in dairy ewes (Althaus et al., 2001b), although this study dealt with animals at the end of lactation (fifth month), with only one milking session per day. The animals in our study, however, were in the third month of lactation. In dairy cows, the literature is contradictory; Seymour et al. (1988) stated that daily milk production did not contribute significantly to drug persistence in the case of seven different antibiotic preparations; while Anderson et al. (1998) concluded in dairy cows with clinical mastitis, that concentrations of amoxicillin or penicillin G in milk were variable among cows over time and appeared to be influenced by milk production. Manufacturers recommend a withdrawal period of four days/eight milking sessions. As seen in Fig. 1, this is insufficient to guarantee the complete elimination of the drug since positive results to the BRT method oscillate between 23.6% (800 mL/milking session) and 40.1% (400 mL/milking session). The detection of antibiotics after the withholding time has been reported in dairy cattle. For example, Oliver et al. (1990) found positive cases of gentamicin in cow milk 96 h after intramammary treatment and Zeng et al. (1996) needed 120 h or 10 milking sessions to remove all traces of cephapirin administered by an intramammary route. These findings could be related to the fact that the majority of available pharmaceutical preparations are specifically for cows, hence the recommended withdrawal period is for the bovine species. Few studies have been carried out on dairy ewes, although a study by Buswell and Barber (1989) found that one available bovine intramammary preparation, when infused into milking sheep, produced a withholding time approximately three times as long as that recommended for cows. Althaus et al. (2001b) concluded that 10 days seemed to be sufficient to guarantee the elimination of a mix of antibiotics (intramuscularly injected) at the end of the lactation period, when the mean daily milk production was low (228 mL/milking). In our study, as shown in Fig. 1, after the sixth day (12th milking)
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
87
Table I Route of antibiotic administration, number of animals treated, and types of antibiotics used during lactation Route of treatment
No. of ewes
Intramammary
14
Intramuscular
14
Intravenous
14
Antibiotic used Type
Amount
Cephalexin monohydrate + gentamicin sulphate Benzyl penicillin G procaine + dihydrostreptomycin Oxytetracycline
350 mg/mL + 35 mg/mL (1 syringe of 10 mL/udder) 200.000 IU + 200 mg/mL (1 mL/10 kg BW) 200 mg/mL (1 mL/10 kg BW)
No. of treatments
3 2 1
BW: Body Weight
following treatment, low frequencies of positive results were seen: between 1.1% for highly productive _ L/milking) and 2.3% for sheep with sheep (800m lower production (400 mL/milking). The detection limit of the BRT method for sheep milk was 270 ppb (Althaus, 1999; Althaus et al., 2001a), which is close to the MRL (100 ppb) established for cow milk, and recently incorporated on the Annex I list set out by the EU (EC 2728/99). Notice that the BRT is a method which presents a high sensitivity to b-lactamics. In Fig. 1, the evolution of the frequency of doubtful cases can also be seen but this was generally low and <16% for the different levels of milk production (3/4 days after treatment). Intramuscular penicillin G treatment In this treatment the logistic model only considered the effect of milking order to be significant ðv2 ¼ 44:5392; P < 0:0001Þ. The mathematic equation which estimates the frequency of ‘‘positive’’ and ‘‘doubtful’’ cases for the BRT method was L ¼ b0 0:6771½MO ðconcordance coefficient : 86:7%Þ where b0 ¼ 6:0724 to estimate the frequency of ‘‘positive/doubtful and negative’’ cases or b0 ¼ 6:4990 to estimate the frequency of ‘‘positive and doubtful/negative’’ cases. The effect of the milking order on the frequency of ‘‘positive’’ and ‘‘doubtful’’ responses to the BRT method is shown in Fig. 2, with values from 100% in the first milking to 3.2% for the 14th milking. As with cephalexin, the withdrawal period recommended by manufacturers (three days/six milkings) was shown to be insufficient to guarantee a low percentage of positive cases, given that they show 88.2% of positive results at that time. Nevertheless, it should be mentioned that the BRT method is very
Fig. 2. Intramuscular penicillin injection. Frequency of positive and doubtful cases using Brilliant Black Reduction Test (BRT).
sensitive to penicillin G, given that its detection limit in dairy ewes (2 ppb, Althaus et al., 2001a) is lower than the 4 ppb MRL established by the EU (Commission Regulation (EC) No. 2728/99 (1999)). This finding has been commented on by various authors (Laroque & Neville, 1985; Booth, 1986) and suggests that most of the currently available antibiotic products were developed and introduced onto the market many years before comprehensive and specific testing methods had been improved. This could also be why early results showed that milk from cows treated with different antibiotic preparations was positive, indicating contamination beyond the recommended withholding period (Laroque & Neville, 1985; Seymour et al., 1988). The use of liquid chromatography as a method to confirm the screening test has shown better results. Anderson et al. (1998) found concentrations of Penicillin G in cows below the safe level at withholding time (36–60 h) when treated for mastitis by the intramammary route. Goats treated intramuscularly with penicillin G had a superior withholding time of at least 72 h (six milkings) (Zeng et al., 1996).
88
THE VETERINARY JOURNAL, 165, 1
The finding that milk production did not significantly affect the response to the BRT method ðv2 ¼ 0:0624; P ¼ 0:8028Þ could be due to the fact that the main elimination route of the antibiotic following intramuscular injection is through the loss of urine and not milk. These findings coincide with the results of Seymour et al. (1988), who observed that daily milk production did not contribute significantly to variations in the excretion of penicillin in milk administered via the intramuscular route. Intravenous oxytetracycline treatment The logistic regression model showed a significant effect due to milking order ðv2 ¼ 20:3488; P < 0:0001Þ as well as for milk production ðv2 ¼ 4:9519; P < 0:05Þ. The frequency of ‘‘doubtful’’ cases was <7%. For this reason only the variation of the frequency of ‘‘positive’’ cases to the BRT method is shown in Fig. 3 according to the milking order for different dairy production (400, 600, and 800 mL/ milking). The prediction logistic model for calculating the frequencies of positive and doubtful cases was
this analysis. Suhren and Heeschen (1993) pointed out that Bacillus cereus var. mycoides ATCC 9634 is sensitive to concentrations of <100 ppb of different tetracyclines. Oxytetracycline was used in this case because it is a widely used antibiotic in sheep. Moreover, we wanted to verify whether the method was capable of detecting residues. The results show that oxytetracycline cannot be detected and Quality Control Laboratories in Spain should now use other tests which are more sensitive to this type of antimicrobial. The detection of oxytetracycline 48 h after treatment presented an important variability (between 28% and 47%) in dairy cows, depending on the screening test used (Oliver et al., 1990). For the withdrawal period recommended by the manufacturer (three days/six milkings), high percentages of positive cases were found: 40.2% (800 mL/milking) and 53.8 % (400 mL/milking), thus showing that this time interval is not enough to guarantee the absence of oxytetracycline residues in dairy sheep.
L ¼ b0 0:2232½MO 0:00137½P ðconcordance coefficient : 70:6%Þ
CONCLUSIONS
where b0 ¼ 2:0384 is used to estimate the frequency of ‘‘positive/doubtful and negative’’ cases or b0 ¼ 2:3150 to estimate the frequency of ‘‘positive and doubtful/negative’’ cases. A decrease in the frequency of positive cases in the number of milking sessions can be seen in Fig. 3. The BRT method detected <80% of the positive cases in the first control since it will not detect oxytetracycline residues <5500 ppb in dairy ewes (Althaus, 1999; Althaus et al., 2001a). This value is however higher than the MRL established by the EU (100 ppb, Commission Regulation (EC) No. 508/99 (1999)), implying that the test is not appropriate for
The response to the BRT method varied according to the milking order as lactation took place, although each treatment gave different results. The frequency of positive results using the BRT method for the recommended withdrawal period was high in the three treatments used. For this reason it is important to corroborate the findings of this study by using other screening tests to compare results. This study contributes to our knowledge of antibiotic residues in milk sent to the dairy industry. It is vital to reduce processing problems in dairy products, for example the losses in fermented products, when they inhibit the bacterial processes involved in the production of cheese-making. It is equally important to enhance the quality of these products available to the consumer, who need protecting against the presence of antibiotics in dairy products. Further controlled studies using sensitive methods for antibiotic residue detection in ewe milk are needed to clarify withdrawal times necessary to minimize the risk of antibiotic contamination of milk. ACKNOWLEDGEMENTS
Fig. 3. Intravenous oxytetracycline injection. Frequency of positive cases using Brilliant Black Reduction Test (BRT).
This work was funded by the project 191/IA-40, from the Consejeria de Agricultura y Medio Ambiente (Junta de Comunidades de Castilla-La Mancha, Spain).
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
REFERENCES AGRESTI, A. (1990). Categorical Data Analysis, 588p. New York: Wiley. Althaus R. (1999). Estudio sobre los m etodos de deteccio´n de inhibidores en leche de oveja. Tesis Doctoral, Universidad Politecnica de Valencia, Spain, 309pp. NDEZ, N. ´GUEZ, M. & FERNA ALTHAUS, R., MOLINA, P., RODRI ODRIGUEZ (2001a). Evaluation of the BRT method for detection of b-lactam antibiotics in ewe milk. Milchwissenschaft 56, 568–72. ALTHAUS, R. L., MOLINA, P., SOUTULLO, H., PERIS, C., NDEZ, N. (2001b). Antibiotic ´NEZ, M. & FERNA MARTI ARTINEZ depletion time and screening methods in milk of Manchega dairy ewes. Options m editerran eennes 46, 141–4. ANDERSON, K. L., MOATS, W. A., RUSHING, J. E. & O’CARROLL, J. M. (1998). Detection of milk antibiotic residues by use of screening tests and liquid chromatography after intramammary administration of amoxicillin or penicillin G in cows with clinical mastitis. American Journal Veterinary Research 59, 1096–9. BENCINI, R. & PULINA, G. (1997). The quality of sheep milk: a review. Australian Journal of Experimental Agriculture 37, 485–504. BISHOP, J. R. & WHITE, C. H. (1984). Antibiotic residue detection in milk – a review. Journal of Food Protection 47, 647–52. BOOTH, J. M. (1986). Intramammary antibiotic preparations and their withholding times. The Veterinary Record 118, 34–5. BOOTH, J. M. & HARDING, F. (1986). Testing for antibiotic residues in milk. The Veterinary Record 119, 565–9. BRADY, M. S. & KATZ, S. E. (1988). Antibiotic/antimicrobials residues in milk. Journal of Food Protection 51, 8–11. BUSWELL, J. F. & BARBER, D. M. (1989). Antibiotic persistence and tolerance in the lactating sheep following a course of intramammary therapy. British Veterinary Journal 145, 552–7. CONTRERAS, A., PAAPE, M. J., DI CARLO, A. L., MILLER, R. H. & RAINARD, P. (1997). Evaluation of selected antibiotic residue screening tests for milk from individual goats. Journal of Dairy Science 80, 1113–8. Commission Regulation (EC) No. 508/99 (1999). 4 March 1999, amending Annexes I to IV of Council Regulation (EEC) No. 2377/90. Official Journal of the European Communities L60, 16–52. Commission Regulation (EC) No. 2728/99 (1999). 20 December 1999, amending Annexes I, II and III of Council Regulation (EEC) No. 2377/90. Official Journal of the European Communities L328, 23–34.
89
Council Regulation (EEC) No. 2377/90 (1990). 26 June 1990. Official Journal of the European Communities L224, 1–8. HEESCHEN, W. H. & SUHREN, G. (1996). Principles of and practical experiences with an integrated system for the detection of antimicrobials in milk. Milchwissenschaft 51, 154–60. LAROQUE, L. & NEVILLE, G. A. (1985). Quantitative evaluation of a bovine antibiotic infusion product by milk residue depletion studies. Journal of Food Protection 48, 611–5. LAROQUE, L. & NEVILLE, G. A. (1986). A practical evaluation of the Delvotest Pâ Multi Plate test in screening raw milk for antibiotics. Journal of Food Protection 49, 868–70. MCEWEN, S. A., ALAN, H. M. & WILLIAM, D. B. (1991). A dairy farm survey of antibiotic treatment practices, residue control methods and associations with inhibitors in milk. Journal of Food Protection 54, 454–9. ´ TAIN, J. P. (1981). Cin MORE ORETAIN etique d’elimination des r esidus d’antibiotiques dans le lait apr es traitement th erapeutique. Recherche Medicine Ve´te´rinaire 157, 199– 204. ´ TAIN, J. P. (1986). Les m MORE ORETAIN edicaments v et erinaires et la qualit e du lait: le probl eme des r esidues d’antibiotiques. Technicien du lait, 10–6. OLIVER, S. P., MAKI, J. L. & DOWLEN, H. H. (1990). Antibiotic residues in milk following antimicrobial therapy during lactation. Journal of Food Protection 53, 693–6. SAS (1998). User Guide Statistics, Version 6.12, Edition 1996. SASâ Institut Cary, NC. SEYMOUR, E. H., JONES, G. M. & MCGILLIARD, M. L. (1988). Persistence of residues in milk following antibiotic treatment of dairy cattle. Journal of Dairy Science 71, 2292–6. SUHREN, G. & HEESCHEN, W. (1993). Detection of tetracyclines in milk by Bacillus cereus microlitre test with indicator. Milchwissenschaft 48, 259–63. ZENG, S. S., ESCOBAR, E. N. & BROWN-CROWDER, I. (1996). Evaluation of screening tests for detection of antibiotic residues in goat milk. Small Ruminant Research 21, 155–6. ZENG, S. S., HART, S., ESCOBAR, E. N. & TESFAI, K. (1998). Validation of antibiotic residue tests for dairy goats. Journal of Food Protection 61, 344–9. (Accepted for publication 15 May 2002)
Residue Persistence in Sheep Milk Following Antibiotic Therapy A. MOLINA , M.P. MOLINAà, R.L. ALTHAUS§ and L. GALLEGO
Departamento de Ciencia y Tecnologıa Agroforestal, E.T.S.I. Agr onomos, Universidad de Castilla-La Mancha, 02071 Albacete, Spain, Departamento de Ciencia Animal, Universidad Polit ecnica, 46071 Valencia, Spain, §C atedra de Biofısica, Facultad de Ciencias Veterinarias, Universidad Nacional del Litoral, 3080 Esperanza, Argentina à
SUMMARY Drug residues in milk supplies may have public health implications and can interfere in the manufacture of dairy products, such as cheese. In Spain, most ewe milk production is destined for cheese making, often using raw milk. This study analyses the main factors influencing antibiotic depletion time in lactating dairy sheep, 42 Manchega ewes were distributed into three groups, each receiving a different treatment (cephalexin intramammary infusion, penicillin G intramuscular, and oxytetracycline intravenous injections). During and after the recommended withdrawal period, milk samples were taken at each milking. A microbiological inhibition test (Brilliant Black Reduction, BRT) was used to screen all samples and antibiotic withholding times were established using a logistic regression model. The response to the BRT method in milk from individual ewes treated showed that the effect of the milking order was significant ðP < 0:001Þ with the three antibiotics. However the only influence on milk yield was with the intramammary treatment ðP < 0:005Þ. The BRT method was found to be very sensitive, particularly to the two b-lactamic antibiotics. Ó 2002 Elsevier Science Ltd. All rights reserved.
KEYWORDS: Dairy ewes; antibiotic residues; withdrawal period; screening methods. INTRODUCTION The presence of drug residues in milk supplies may have public health implications and are perceived by consumers as undesirable (Bishop & White, 1984; Booth & Harding, 1986; Mor etain, 1986; Brady & Katz, 1988; McEwen et al., 1991; Bencini & Pulina, 1997). Moreover, residues can interfere with the manufacture of dairy products, such as cheese. In Spain, most ewe milk production is used for cheese making, often using raw milk, and such ‘‘pure’’ sheep cheeses carry a Protected Denomination of Origin (DOP) which guarantees quality control. Mastitis is one of the most important problems in dairy cows and the major source of residues found in milk supplies (Booth & Harding, 1986; McEwen et al., 1991). In recent years, due to the intensification of milk production in small ruminants, there has been an increase in the frequency of mastitis cases also in dairy ewes and goats (Buswell & Barber,
Correspondence to: A. Molina, Departamento de Ciencia y Tecnologıa Agroforestal, E.T.S.I. Agronomos, Universidad de Castilla-La Mancha, 02071 Albacete, Spain Tel.: +34-967-599238; E-mail: [email protected]
1989). In sheep, treatment of mastitis and other diseases with pharmacological products is standard practice, and in many cases antibiotic milk contamination may be caused by unnecessary treatments carried out without a veterinary prescription, and with inadequate knowledge of the suitable dosage, administration route or depletion time of the antibiotic. One reason for this is because the majority of available treatments for mastitis are specific for cows, and withdrawal times are unknown in sheep. Indeed, no withdrawal times for milk from dairy ewes treated with antibiotics have been specifically established (Buswell & Barber, 1989). The European Union (EU) determines the limits for the presence of specified veterinary residues in milk. The antibiotic residues are defined by Council Regulation EEC 2377/90 (1990), although a number of amendments have subsequently been made, primarily to extend the list of compounds whose Maximum Residues Limits (MRLs) have been previously set. Detectable concentrations of antibiotic residues in milk supplies higher than the MRLs are illegal. Many drug residue screening tests are currently used for cowsÕ milk (Bishop & White, 1984; Booth & Harding, 1986; Laroque & Neville, 1986;
1090-0233/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved.
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
Brady & Katz, 1988; Seymour et al., 1988; Anderson et al., 1998), however, there are only a few evaluations of antibiotic residue screening tests for ewesÕ milk (Althaus, 1999; Althaus et al., 2001a) or goats’ milk (Zeng et al., 1996; Contreras et al., 1997; Zeng et al., 1998). Residue screening tests are important for daily antibiotic control, as well as for preventing antibioticcontaminated milk from entering the processing of milk products (Heeschen & Suhren, 1996). Persistence of antibiotic residues in milk is related to several variables such as daily milk production at the time of treatment, the type and amount of antibiotic used, and the disease state of the animal (Mor etain, 1981; Laroque & Neville, 1985; Seymour et al., 1988; Oliver et al., 1990; McEwen et al., 1991). The objective of this study was to determine the main factors that influence antibiotic depletion time in lactating dairy ewes receiving various antibiotic treatments using a screening method previously used to test sheep milk (Althaus et al., 2001a), the Brillant Black Reduction (BRT). The BRT method is a microbiological inhibition test widely used for the detection of residues in milk, using Bacillus stearotermophilus var. calidolactis. It is the reference method used in Spain in all quality control dairy laboratories for the detection of residues in cow, sheep, and goat milk. MATERIALS AND METHODS Animals Fifty-five Manchegan bred ewes (body weight range 55–80 kg, age 1–8 years) from the Manchegan herd in the experimental flock of the University of Castilla-La Mancha (Spain) were selected for this study. The ewes were randomly divided into three groups, each containing 14 animals roughly equal in body weight, age, and daily milk yield. The animals did not receive any kind of treatment before the experiment. The ewes were housed and fed the same as the rest of the herd; lambing took place in January and February, and mothers suckled their lambs for 35 days. After weaning, the ewes were machinemilked twice a day. The experiment took place in the fourth week of milking. Antibiotic treatment and sample collection Antibiotics were administered immediately after milking was completed. Each group of animals was treated with a different antibiotic; the first group was given three intramammary infusions per udder of a 5 mL suspension of 350 mg/mL cephalexin monohydrate with 35 mg/mL gentamicin sulphate (Max-
85
icef, Vetoquinol), at 24 h intervals. The second group was given two intramuscular injections of 1 mL/10 kg body weight of a suspension of 200.000 IU benzyl penicillin G procaine with 200 mg/mL dihydrostreptomycin (Dihidropen, Iven) at 24 h intervals. The third group was given one intravenous injection of 1 mL/10 kg body weight of 200 mg/mL of oxytetracycline (Oxiciven LA/200, Iven). The products were administered according to manufacturer’s instructions see Table I. After treatment, the ewes were machine-milked twice a day, the milk yield from each ewe measured and samples were taken. Further milk samples were collected at 12 h post treatment, and sampling continued during every milking session until double the time of the withdrawal period recommended by manufacturers. The samples (50 mL/milking session/animal) were collected in disposable plastic containers and kept at 5 °C until they were analysed. Brilliant Black Reduction Test Milk samples were analysed during the 24 h period after collection using the Brilliant Black Reduction Test (BRT), according to manufacturer’s instructions. One hundred microlitres of milk were added to ready-prepared individual BRT cups containing Bacillus stearothermophilus var. calidolactis. Milk shown to be free from antimicrobials was used as ‘‘negative control’’. For the ‘‘positive control’’, milk samples with 4 lg penicillin/kg were used. After a 1 h diffusion time at +4 °C, the BRT tests were floated in a waterbath and incubated at 64 1 °C for 3 h. Visual interpretation was carried out independently by three trained observers and evaluated visually as ‘‘negative’’, ‘‘doubtful’’ or ‘‘positive’’. Statistical analysis To evaluate the influence of the milking order and milk yield on the responses to the BRT inhibition microbiology screening test, the logistic regression model (Agresti, 1990) was used employing Stepwise option from the Logistic procedure of the SAS (SAS, 1998). Variables were analysed using the following logistic model: Lijk ¼ logit½Pijk ¼ b0 þ b1 ½MOi þ b2 ½Pj þ eijk where Lijk is the logit model; [Pijk ] is the probability for the response category (‘‘positive/doubtful and negative’’ or ‘‘doubtful/negative’’); b0 is the intercept; b1 ; b2 are the estimate parameters for the model; [MO]i is the effect of milking order (1, 2, 3, 4, 5, 6, 7, and 16); [P]j is the effect of dairy production and eijk is the residual error.
86
THE VETERINARY JOURNAL, 165, 1
The concordance coefficient was applied as rank correlation between the observed responses and predicted probabilities (SAS, 1998). RESULTS AND DISCUSSION Intramammary cephalexin infusion The application of the logistic regression model considered the effect of the milking order ðv2 ¼ 63:4393; P < 0:0001Þ and milk yield ðv2 ¼ 6:5582; P < 0:05Þ to be significant. The prediction model for calculating the frequencies of positive and doubtful samples, taking into account the milking order [MO] and dairy production [P], was L ¼ b0 0:8330½MO 0:00194½P ðconcordance coefficient : 95:2%Þ where, b0 ¼ 7:0413 to estimate the frequency of ‘‘positive/doubtful + negative’’ cases or b0 ¼ 7:6784 to estimate the frequency of ‘‘positive + doubtful/ negative’’. Fig. 1 shows the effect of the two variables, milking order and dairy production (in this case depending on the milk yield of ewes, 400, 600, and 800 mL/milking, respectively) on the frequency of ‘‘positive’’ and ‘‘doubtful’’ responses to the BRT method. A decrease in the frequency of positive cases was observed as time elapsed after the last application. This fact has also been observed in dairy cows treated with different antibiotics by the intramammary route (Oliver et al., 1990) to such an extent that the number of positive samples decreased after treatment increased. This finding was more pronounced in sheep with high milk production and there was a greater decrease in positive cases with time for ewes with a
Fig. 1. Intramammary cephalexin infusion. Frequency of positive and doubtful cases using Brilliant Black Reduction Test (BRT).
high milk yield. On the other hand, sheep with lower milk production took longer to eliminate the antibiotic. The amount of milk produced could influence responses to the BRT method since milk represents the main elimination route of the antibiotic when administered by the intramammary route (see Table I). Continued synthesis of the milk in the mammary tissues impedes the diffusion of this antibiotic into the blood. The effect of the level of milk production on the elimination of antibiotics has been reported in dairy ewes (Althaus et al., 2001b), although this study dealt with animals at the end of lactation (fifth month), with only one milking session per day. The animals in our study, however, were in the third month of lactation. In dairy cows, the literature is contradictory; Seymour et al. (1988) stated that daily milk production did not contribute significantly to drug persistence in the case of seven different antibiotic preparations; while Anderson et al. (1998) concluded in dairy cows with clinical mastitis, that concentrations of amoxicillin or penicillin G in milk were variable among cows over time and appeared to be influenced by milk production. Manufacturers recommend a withdrawal period of four days/eight milking sessions. As seen in Fig. 1, this is insufficient to guarantee the complete elimination of the drug since positive results to the BRT method oscillate between 23.6% (800 mL/milking session) and 40.1% (400 mL/milking session). The detection of antibiotics after the withholding time has been reported in dairy cattle. For example, Oliver et al. (1990) found positive cases of gentamicin in cow milk 96 h after intramammary treatment and Zeng et al. (1996) needed 120 h or 10 milking sessions to remove all traces of cephapirin administered by an intramammary route. These findings could be related to the fact that the majority of available pharmaceutical preparations are specifically for cows, hence the recommended withdrawal period is for the bovine species. Few studies have been carried out on dairy ewes, although a study by Buswell and Barber (1989) found that one available bovine intramammary preparation, when infused into milking sheep, produced a withholding time approximately three times as long as that recommended for cows. Althaus et al. (2001b) concluded that 10 days seemed to be sufficient to guarantee the elimination of a mix of antibiotics (intramuscularly injected) at the end of the lactation period, when the mean daily milk production was low (228 mL/milking). In our study, as shown in Fig. 1, after the sixth day (12th milking)
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
87
Table I Route of antibiotic administration, number of animals treated, and types of antibiotics used during lactation Route of treatment
No. of ewes
Intramammary
14
Intramuscular
14
Intravenous
14
Antibiotic used Type
Amount
Cephalexin monohydrate + gentamicin sulphate Benzyl penicillin G procaine + dihydrostreptomycin Oxytetracycline
350 mg/mL + 35 mg/mL (1 syringe of 10 mL/udder) 200.000 IU + 200 mg/mL (1 mL/10 kg BW) 200 mg/mL (1 mL/10 kg BW)
No. of treatments
3 2 1
BW: Body Weight
following treatment, low frequencies of positive results were seen: between 1.1% for highly productive _ L/milking) and 2.3% for sheep with sheep (800m lower production (400 mL/milking). The detection limit of the BRT method for sheep milk was 270 ppb (Althaus, 1999; Althaus et al., 2001a), which is close to the MRL (100 ppb) established for cow milk, and recently incorporated on the Annex I list set out by the EU (EC 2728/99). Notice that the BRT is a method which presents a high sensitivity to b-lactamics. In Fig. 1, the evolution of the frequency of doubtful cases can also be seen but this was generally low and <16% for the different levels of milk production (3/4 days after treatment). Intramuscular penicillin G treatment In this treatment the logistic model only considered the effect of milking order to be significant ðv2 ¼ 44:5392; P < 0:0001Þ. The mathematic equation which estimates the frequency of ‘‘positive’’ and ‘‘doubtful’’ cases for the BRT method was L ¼ b0 0:6771½MO ðconcordance coefficient : 86:7%Þ where b0 ¼ 6:0724 to estimate the frequency of ‘‘positive/doubtful and negative’’ cases or b0 ¼ 6:4990 to estimate the frequency of ‘‘positive and doubtful/negative’’ cases. The effect of the milking order on the frequency of ‘‘positive’’ and ‘‘doubtful’’ responses to the BRT method is shown in Fig. 2, with values from 100% in the first milking to 3.2% for the 14th milking. As with cephalexin, the withdrawal period recommended by manufacturers (three days/six milkings) was shown to be insufficient to guarantee a low percentage of positive cases, given that they show 88.2% of positive results at that time. Nevertheless, it should be mentioned that the BRT method is very
Fig. 2. Intramuscular penicillin injection. Frequency of positive and doubtful cases using Brilliant Black Reduction Test (BRT).
sensitive to penicillin G, given that its detection limit in dairy ewes (2 ppb, Althaus et al., 2001a) is lower than the 4 ppb MRL established by the EU (Commission Regulation (EC) No. 2728/99 (1999)). This finding has been commented on by various authors (Laroque & Neville, 1985; Booth, 1986) and suggests that most of the currently available antibiotic products were developed and introduced onto the market many years before comprehensive and specific testing methods had been improved. This could also be why early results showed that milk from cows treated with different antibiotic preparations was positive, indicating contamination beyond the recommended withholding period (Laroque & Neville, 1985; Seymour et al., 1988). The use of liquid chromatography as a method to confirm the screening test has shown better results. Anderson et al. (1998) found concentrations of Penicillin G in cows below the safe level at withholding time (36–60 h) when treated for mastitis by the intramammary route. Goats treated intramuscularly with penicillin G had a superior withholding time of at least 72 h (six milkings) (Zeng et al., 1996).
88
THE VETERINARY JOURNAL, 165, 1
The finding that milk production did not significantly affect the response to the BRT method ðv2 ¼ 0:0624; P ¼ 0:8028Þ could be due to the fact that the main elimination route of the antibiotic following intramuscular injection is through the loss of urine and not milk. These findings coincide with the results of Seymour et al. (1988), who observed that daily milk production did not contribute significantly to variations in the excretion of penicillin in milk administered via the intramuscular route. Intravenous oxytetracycline treatment The logistic regression model showed a significant effect due to milking order ðv2 ¼ 20:3488; P < 0:0001Þ as well as for milk production ðv2 ¼ 4:9519; P < 0:05Þ. The frequency of ‘‘doubtful’’ cases was <7%. For this reason only the variation of the frequency of ‘‘positive’’ cases to the BRT method is shown in Fig. 3 according to the milking order for different dairy production (400, 600, and 800 mL/ milking). The prediction logistic model for calculating the frequencies of positive and doubtful cases was
this analysis. Suhren and Heeschen (1993) pointed out that Bacillus cereus var. mycoides ATCC 9634 is sensitive to concentrations of <100 ppb of different tetracyclines. Oxytetracycline was used in this case because it is a widely used antibiotic in sheep. Moreover, we wanted to verify whether the method was capable of detecting residues. The results show that oxytetracycline cannot be detected and Quality Control Laboratories in Spain should now use other tests which are more sensitive to this type of antimicrobial. The detection of oxytetracycline 48 h after treatment presented an important variability (between 28% and 47%) in dairy cows, depending on the screening test used (Oliver et al., 1990). For the withdrawal period recommended by the manufacturer (three days/six milkings), high percentages of positive cases were found: 40.2% (800 mL/milking) and 53.8 % (400 mL/milking), thus showing that this time interval is not enough to guarantee the absence of oxytetracycline residues in dairy sheep.
L ¼ b0 0:2232½MO 0:00137½P ðconcordance coefficient : 70:6%Þ
CONCLUSIONS
where b0 ¼ 2:0384 is used to estimate the frequency of ‘‘positive/doubtful and negative’’ cases or b0 ¼ 2:3150 to estimate the frequency of ‘‘positive and doubtful/negative’’ cases. A decrease in the frequency of positive cases in the number of milking sessions can be seen in Fig. 3. The BRT method detected <80% of the positive cases in the first control since it will not detect oxytetracycline residues <5500 ppb in dairy ewes (Althaus, 1999; Althaus et al., 2001a). This value is however higher than the MRL established by the EU (100 ppb, Commission Regulation (EC) No. 508/99 (1999)), implying that the test is not appropriate for
The response to the BRT method varied according to the milking order as lactation took place, although each treatment gave different results. The frequency of positive results using the BRT method for the recommended withdrawal period was high in the three treatments used. For this reason it is important to corroborate the findings of this study by using other screening tests to compare results. This study contributes to our knowledge of antibiotic residues in milk sent to the dairy industry. It is vital to reduce processing problems in dairy products, for example the losses in fermented products, when they inhibit the bacterial processes involved in the production of cheese-making. It is equally important to enhance the quality of these products available to the consumer, who need protecting against the presence of antibiotics in dairy products. Further controlled studies using sensitive methods for antibiotic residue detection in ewe milk are needed to clarify withdrawal times necessary to minimize the risk of antibiotic contamination of milk. ACKNOWLEDGEMENTS
Fig. 3. Intravenous oxytetracycline injection. Frequency of positive cases using Brilliant Black Reduction Test (BRT).
This work was funded by the project 191/IA-40, from the Consejeria de Agricultura y Medio Ambiente (Junta de Comunidades de Castilla-La Mancha, Spain).
ANTIBIOTIC RESIDUE PERSISTENCE IN EWE MILK
REFERENCES AGRESTI, A. (1990). Categorical Data Analysis, 588p. New York: Wiley. Althaus R. (1999). Estudio sobre los m etodos de deteccio´n de inhibidores en leche de oveja. Tesis Doctoral, Universidad Politecnica de Valencia, Spain, 309pp. NDEZ, N. ´GUEZ, M. & FERNA ALTHAUS, R., MOLINA, P., RODRI ODRIGUEZ (2001a). Evaluation of the BRT method for detection of b-lactam antibiotics in ewe milk. Milchwissenschaft 56, 568–72. ALTHAUS, R. L., MOLINA, P., SOUTULLO, H., PERIS, C., NDEZ, N. (2001b). Antibiotic ´NEZ, M. & FERNA MARTI ARTINEZ depletion time and screening methods in milk of Manchega dairy ewes. Options m editerran eennes 46, 141–4. ANDERSON, K. L., MOATS, W. A., RUSHING, J. E. & O’CARROLL, J. M. (1998). Detection of milk antibiotic residues by use of screening tests and liquid chromatography after intramammary administration of amoxicillin or penicillin G in cows with clinical mastitis. American Journal Veterinary Research 59, 1096–9. BENCINI, R. & PULINA, G. (1997). The quality of sheep milk: a review. Australian Journal of Experimental Agriculture 37, 485–504. BISHOP, J. R. & WHITE, C. H. (1984). Antibiotic residue detection in milk – a review. Journal of Food Protection 47, 647–52. BOOTH, J. M. (1986). Intramammary antibiotic preparations and their withholding times. The Veterinary Record 118, 34–5. BOOTH, J. M. & HARDING, F. (1986). Testing for antibiotic residues in milk. The Veterinary Record 119, 565–9. BRADY, M. S. & KATZ, S. E. (1988). Antibiotic/antimicrobials residues in milk. Journal of Food Protection 51, 8–11. BUSWELL, J. F. & BARBER, D. M. (1989). Antibiotic persistence and tolerance in the lactating sheep following a course of intramammary therapy. British Veterinary Journal 145, 552–7. CONTRERAS, A., PAAPE, M. J., DI CARLO, A. L., MILLER, R. H. & RAINARD, P. (1997). Evaluation of selected antibiotic residue screening tests for milk from individual goats. Journal of Dairy Science 80, 1113–8. Commission Regulation (EC) No. 508/99 (1999). 4 March 1999, amending Annexes I to IV of Council Regulation (EEC) No. 2377/90. Official Journal of the European Communities L60, 16–52. Commission Regulation (EC) No. 2728/99 (1999). 20 December 1999, amending Annexes I, II and III of Council Regulation (EEC) No. 2377/90. Official Journal of the European Communities L328, 23–34.
89
Council Regulation (EEC) No. 2377/90 (1990). 26 June 1990. Official Journal of the European Communities L224, 1–8. HEESCHEN, W. H. & SUHREN, G. (1996). Principles of and practical experiences with an integrated system for the detection of antimicrobials in milk. Milchwissenschaft 51, 154–60. LAROQUE, L. & NEVILLE, G. A. (1985). Quantitative evaluation of a bovine antibiotic infusion product by milk residue depletion studies. Journal of Food Protection 48, 611–5. LAROQUE, L. & NEVILLE, G. A. (1986). A practical evaluation of the Delvotest Pâ Multi Plate test in screening raw milk for antibiotics. Journal of Food Protection 49, 868–70. MCEWEN, S. A., ALAN, H. M. & WILLIAM, D. B. (1991). A dairy farm survey of antibiotic treatment practices, residue control methods and associations with inhibitors in milk. Journal of Food Protection 54, 454–9. ´ TAIN, J. P. (1981). Cin MORE ORETAIN etique d’elimination des r esidus d’antibiotiques dans le lait apr es traitement th erapeutique. Recherche Medicine Ve´te´rinaire 157, 199– 204. ´ TAIN, J. P. (1986). Les m MORE ORETAIN edicaments v et erinaires et la qualit e du lait: le probl eme des r esidues d’antibiotiques. Technicien du lait, 10–6. OLIVER, S. P., MAKI, J. L. & DOWLEN, H. H. (1990). Antibiotic residues in milk following antimicrobial therapy during lactation. Journal of Food Protection 53, 693–6. SAS (1998). User Guide Statistics, Version 6.12, Edition 1996. SASâ Institut Cary, NC. SEYMOUR, E. H., JONES, G. M. & MCGILLIARD, M. L. (1988). Persistence of residues in milk following antibiotic treatment of dairy cattle. Journal of Dairy Science 71, 2292–6. SUHREN, G. & HEESCHEN, W. (1993). Detection of tetracyclines in milk by Bacillus cereus microlitre test with indicator. Milchwissenschaft 48, 259–63. ZENG, S. S., ESCOBAR, E. N. & BROWN-CROWDER, I. (1996). Evaluation of screening tests for detection of antibiotic residues in goat milk. Small Ruminant Research 21, 155–6. ZENG, S. S., HART, S., ESCOBAR, E. N. & TESFAI, K. (1998). Validation of antibiotic residue tests for dairy goats. Journal of Food Protection 61, 344–9. (Accepted for publication 15 May 2002)