Pharmacokinetics of azithromycin in lactating dairy cows with subclinical mastitis caused by Staphylococcus aureus

Introduction

Azithromycin is a macrolide belonging to a group of semi-synthetic antimicrobial agents called azalides which were derived from erythromycin by adding an endocyclic nitrogen atom in position 9 (Mulazimoglu et al., 2005). Azithromycin formulation is not available for use in production animals. It has a particular pharmacokinetic profile (Biskri & Mazel, 2003) and it has a clear in vivo postantibiotic effect (PAE) (Van Bambeke & Tulkens, 2001). According to the pattern of antimicrobial activity, azithromycin is a time-dependant antimicrobial agent with prolonged persistent effect. The PK/PD measure that best correlates with azithromycin′s clinical efficacy is the area under the curve/minimum inhibitory concentration (AUC_0-24/MIC) (Ambrose et al., 2004).

Azithromycin offers a number of advantages that differentiate it from conventional macrolides. As a consequence of the structural changes above mentioned, this drug is more potent against gram-negative bacteria than other conventional macrolides such as erythromycin, while its action remains the same against Gram-positive microorganisms (Retsema et al., 1987). In addition, it has higher resistance against acid degradation in comparison with erythromycin (Fiese & Steffen, 1990); it has wider distribution and a longer elimination half-life.

Like traditional macrolides, azithromycin is capable of concentrating in phagocytic cells (McConnell, 1999) where it penetrates by nonionic diffusion (Mulazimoglu et al., 2005). In cells, it is found in lysosomes, where it concentrates by ion trapping (Gómez-Lus et al., 2005). It was postulated that macrolide accumulation inside leukocytes may increase the activity of cell immunity-related mechanisms. Leukocytes, mainly neutrophils, macrophages and monocytes, accumulate antibiotic molecules in their lysosomes and transport them by chemotactic response to the infection site (Labro, 1998; Shryock et al., 1998).

The above mentioned properties suggest that azithromycin may be effective for treating infections caused by Gram-positive aerobic microorganisms which survive inside lysosomes, as the case of S. aureus bovine mastitis.

The objective of this study was to investigate the azithromycin pharmacokinetic profile in lactating dairy cows with subclinical mastitis caused by S. aureus, and calculate the corresponding PK/PD parameters.

Materials and methods

Results

According to the interpretative criteria set by CLSI (2008), the 83.6% of the S. aureus evaluated would be classified as sensitive, 9.1% showed intermedia sensitivity and 7.3% would be classified as resistant. Out of a total of 51 isolations assessed, an MIC₉₀ of 1 μg/mL and an MIC₅₀ of 0.5 μg/mL were determined. The MIC in the S. aureus ATCC 29213 strain was 0.5 μg/mL.

The mean concentrations of azithromycin reached in quarter milk (n = 24) and plasma(n = 6) of the six experimental cows treated are shown in Fig. 1. Pharmacokinetic parameters determined by means of a noncompartmental analysis of milk and plasma concentrations are shown in Table 1. A high C_max and a large AUC demonstrate major azithromycin penetration into the mammary gland.

The mean concentrations of azithromycin determined in milk of mastitic quarters (n = 8) and in milk of healthy quarters (n = 16) are shown in Fig. 2. It can be observed that the concentrations in mastitic quarters are higher than that in healthy ones. PK parameters calculated in mastitic quarters and healthy quarters are shown in Table 2.

The mean concentrations of azithromycin determined in plasma of high-producing (n = 3) and low-producing cows (n = 3) and in quarters of high-producing cows (n = 12) and low-producing cows (n = 12) are shown in Fig. 3. The pharmacokinetic parameters of azithromycin in quarter milk of high-producing cows, in quarter milk of low-producing cows, in plasma of high-producing cows and in plasma of low-producing are shown in Table 3.

PK/PD parameters such as AUC_0-24h/MIC₉₀ and T > MIC₉₀ are shown in Table 4.

The statistical results are shown in Table 5. There was a significant difference (P = 0.043) between mastitic quarters and healthy quarters regarding T_max post-2^nd administration. The health status factor also had a significant effect on the AUC_0-∞ (P = 0.0331) and Cl_mam (P = 0.0331). The productive level factor appeared to have some influence on the T > MIC₉₀ as well as the health status over the F_milk. However, in these cases no statistically significant differences were observed.

Bacteriological cure control showed that seven of the eight infected quarters were negatives. One mammary quarter of one cow was positive at 20 and 30 days post-treatment. At 20 days post-treatment, the SCC in all cows ranged between 58 000 and 223 000, which represented a statistically significant diminution tested by a nonparametric test (Mann–Whitney).

Discussion

There are no azithromycin susceptibility studies on S. aureus isolated from bovine mastitis cases. However, there are data for other macrolides. Usually when running antibiograms, an erythromycin disk is used to evaluate susceptibility to the macrolide family. Variable percentages of macrolide resistant S. aureus ranging from 1.9% to 26.3% have been reported in several studies (Ziv., 1980; Owens et al., 1997; Teagle & David, 1999; Andrade et al., 2000). At present there are highly specific methods for susceptibility testing of veterinary pathogens (CLSI, 2008). The methods and interpretive criteria used to evaluate the antimicrobial susceptibility are determinant factors and might have strong influence on the results. However, the same techniques are not always used and the same study protocols are not strictly followed. This is a common finding in the literature, so, the variability factors related to the used method should be considered.

The MIC₅₀ calculated for the 51 S. aureus isolations was 0.5 μg/mL and the MIC₉₀ was 1 μg/mL. Although it is not advisable to compare MICs of different antimicrobial agents, it is worth stating that erythromycin MIC₉₀ for bovine isolated S. aureus was reported as 0.5 μg/mL in several publications (Watts et al., 1995; Watts & Salmon, 1997; Salmon et al., 1998; Ruiz et al., 2001; Pitkäläet al., 2004). A study made in Argentina reported an erythromycin MIC₉₀ of 0.75 μg/mL (Gentilini et al., 2000).

Comparison between bioassay and HPLC methods of analysis of azithromycin have also been reported by Riedel et al. (1992), showing no antimicrobial activity from azithromycin metabolites. In human, up to 10 metabolites of azithromycin have been identified and all were microbiologically inactive (Ballow & Amsden, 1992; Lalak & Morris, 1993).

The most prominent pharmacokinetic characteristic of azithromycin is the presence of high tissue concentrations which are maintained a long time after serum concentrations decline to very low levels. This characteristic was demonstrated by the same authors in lactating dairy cows (Turic et al., 2003a), in goats (Cárceles et al., 2005) and humans (Foulds et al., 1990). Azithromycin T½λ was long, an expected finding, according to the characteristics of this antimicrobial.

Milk azithromycin levels resulted much higher than those found in plasma (between 40 and 70 times in milk that in plasma concentrations), which is a logical finding according to the lipophilicity and wide distribution of the drug. Previous research made by our group showed that after one i.m. dose of azithromycin at 10 mg/kg to lactating Holstein cows, the C_max in milk were 4.72 μg/mL in healthy cows and 3.51 μg/mL in mastitic cows and the C_max in plasma were 0.35 μg/mL and 0.44 μg/mL in healthy and mastitic cows, respectively (Turic et al., 2003a).

A study carried out in goats (Cárceles et al., 2005) reported that, after i.m. dose of 20 mg/kg azithromycin, the C_max in plasma was 0.65 ± 0.09 μg/mL which was higher than the one observed in the present work 0.18 ± 0.16 μg/mL. Furthermore the different species the higher dose could contribute to the explanation of this difference. The T_max in goats was 1.25 ± 0.25 h, whereas in this study it was 2 ± 1.56 h.

Another study performed in foals (Davis et al., 2002) suggests that azithromycin may be an alternative to erythromycin for treatment of Rhodococcus equi pneumonia. R. equi is a Gram-positive aerobe that behaves as a facultative intracellular microorganism, resembling S. aureus. The PK profile of this antibiotic in foals was determined in plasma, polymorphonuclear leukocytes (PMN), bronchoalveolar lavage (BAL) fluid, and alveolar cells after oral dose of 10 mg/kg. After 4 and 12 h oral administration, azithromycin concentrations in PMN were 89 and 202 times the corresponding plasma concentrations, respectively. In addition, azithromycin concentrations in PMN persisted for 120 h after oral administration, with a half-life of more than 49 h, although the drug was detectable in plasma for only 12–24 h (the plasma half-life was 16 ± 4.54 h). Azithromycin concentrations in PMN and alveolar cells at 120 h after oral administration were four times above the MIC reported for human isolated R. equi, ≤1 μg/mL (Mascellino et al., 1994).

In our study, milk azithromycin concentrations were more than six times the S. aureus MIC at 48 h after 2^nd i.m. administration. The PK/PD surrogate markers should be used with caution in the case of macrolides for which milk and tissue pharmacokinetics have an important role to predict clinical efficacy. Therefore, further studies are necessary for a better integration of PK and PD parameters to predict clinical efficacy for macrolides, especially for azithromycin (Benchaoui et al., 2004).

Azithromycin exhibited major penetration into milk and it was cleared rather slowly. Pharmacokinetic parameters indicated a high retention of the drug in peripheral compartments. The T½λ in milk after first administration was at least four times longer than that in plasma. Azithromycin T½λ suggested that milk concentrations exhibited a tendency to decrease more slowly than plasma ones. The same pattern was observed after a single 10 mg/kg i.m. dose of azithromycin to lactating Holstein cows (Turic et al., 2003a).

When comparing PK parameters by grouping quarters according to health status, it was observed that azithromycin was eliminated more slowly from mastitic quarters. Although this was an unexpected finding (the pK_a partition hypothesis suggests the opposite), it is coincident with previously reported data (Turic et al., 2003a). Milk pH in the experimental animals ranged between 6.5 and 7.5 with the majority of values around 7.0. Average pH from all mastitic quarters was 7.13 ± 0.23 and from healthy quarters was 6.90 ± 0.21 (see Table 6). This is a normal finding for animals carrying subclinical mastitis. Azithromycin is a weak base with a pKa value of 8.74, as a consequence, by application of the Henderson–Hasselbach equation, there would be approximately double azithromycin molecule dissociation in mastitic milk and more than three times in milk of healthy animals in comparison with plasma (see Table 6). Alkaline drugs (like azithromycin) are trapped in acidic compartments. This theoretical considerations could not, however, be confirmed by the experimental findings reported here. Azithromycin (i.m.) gave rise to very low plasma AUCs, which could be explained by its very high liposolubility and penetration into tissues. Although higher AUC was expected in milk of healthy animals (more acidic), which is a common finding with the classic macrolide antibacterials, we found exactly the opposite end. In our experiment, the highest concentrations were determined in the milk of mastitic animals, with an AUC_milk/AUC_plasma ratio of 743.60 (see Table 6). Our explanation for this unexpected finding is the amount of somatic cells (SCC) present in mastitic milk in comparison with the normal milk. In the former case, the number of SCC was several times above those in normal milk. Mastitic milk normally exhibits very high cell counts as consequence of the inflammatory reaction. As it is known, azithromycin is able to reach high concentrations at infected sites, as a result of increased delivery from phagocytes (Wildfeuer et al., 1994, 1996; Labro, 1998; Shryock et al., 1998; Davis et al., 2002). On this basis, we consider that the inflammatory reaction (and the high amount of cells) in the infected quarters is the main reason for the differences found between mastitic and healthy quarters. A significant azithromycin fraction could be trapped in the milk-cell compartment without participating of the plasma:milk equilibrium, largely dependent on the pK_a– pH relationship. The AUC_0-∞ (P ≤ 0.05) and the MRT were higher in whole milk from mastitic quarters, which may indicate that the drug is present in higher amounts and persist during longer time in mastitic quarters than in healthy ones. At the same time, the F_milk of azithromycin was higher in the mastitic quarters indicating a different PK profile of azithromycin depending on the quarter status. The previous data, reported after a single 10 mg/kg i.m. dose of azithromycin to lactating Holstein cows, support our observations (Turic et al., 2003a). The Cl_mam/F (P ≤ 0.05) showed that azithromycin elimination was faster in healthy quarters than in mastitic quarters.

After an intramammary azithromycin syringe containing 125 mg in each mammary quarter of lactating Holstein cows, difference between the AUC in milk of mastitic and healthy cows resulted statistically significant. The AUC of azithromycin in mastitic cows was 1336.82 μg·h/mL and in healthy cows was 458.02 μg·h/mL (Errecalde et al., 2003). Showing a similar pattern, after an intramammary azithromycin syringe containing 500 mg in each mammary quarter of Holstein cows at drying off, the F_milk in mastitic cows was higher than the F_milk in healthy cows (Turic et al., 2003b).

As expected, the variability of the concentrations in mastitic quarters was higher than that in healthy quarters. Variability of the concentrations in healthy quarters is inherent to the animal physiology and the quarters themselves. In mastitic quarters, there is an extra factor related to the pathological status. Therefore, in addition to the physiological variability there are effects related to infection and inflammation levels, the somatic cell count, the presence of abscesses and/or fibrosis in glandular tissue, milk compositional changes and other circumstances related to mastitis.

As a consequence of the animal selection criterion, the productive levels of both groups were not extremely different. However, separating the groups according to median production allowed us to identify observations that may be extrapolated to situations of major differences in productive levels. The AUC_0-∞ was higher in quarters of low-producing cows than in quarters of high-producing cows. It is possible that a lower milk production causes a slower antibiotic elimination, with a lower Cl_mam/F value and more prolonged T½λ₂. On the basis of these results, we could suggest that low-producing cows have a high tendency to exceed LMR in milk while high-producing cows can eliminate (and dilute) the drug fast enough so as to diminish ‘contact time’ and clinical efficacy possibilities.

According to the excellent availability obtained in milk and a relatively low MIC determined in vitro, we consider azithromycin as a potential antimastitic drug and especially an alternative for the treatment of S. aureus mastitis. Nevertheless, before its recommendation, the withdrawal time must be carefully calculated, especially in consideration of its long persistence. This long persistence is a disadvantage of AZT, although its use in dry cow therapy might be an interesting alternative. Nonetheless, clinical studies are required to establish the optimal PK/PD ratio of azithromycin in milk and mammary tissue. Additionally, we consider that therapeutic protocols should be adjusted according to the level of production of the animals to be treated, because optimal concentrations must be reached avoiding cow under/over dose. That would result in public health benefits, as it is a way to avoid the presence of chemical contaminating agents in milk. Moreover, unwanted exposure of bacteria to sub-optimal antimicrobial concentrations, with the risk of resistant strains emergence, would be minimized.