Distribution of intramuscularly administered erythromycin into subcutaneous tissue chambers before and after inoculation with Pasteurella haemolytica

Research in Veterinary Science 1993, 54, 366-371

Distribution of intramuscularly administered erythromycin into subcutaneous tissue chambers before and after inoculation with Pasteurella haemolytica C. R. CLARKE, Department of Physiological Sciences, College of Veterinary Medicine, Oklahoma State University, Stillwater, Oklahoma 74078, USA, D. W. A. BOURNE, HSC

College of Pharmacy, University of Oklahoma, PO Box 26901, Oklahoma City, Oklahoma 73190, USA, A. K. LAUER, Department of Physiological Sciences, S. J. BARRON, Department of Medicine and Surgery, College of Veterinary Medicine, Oklahoma State University, Stillwater, Oklahoma 74078, USA

Distribution of erythromycin into subcutaneous tissue chambers was characterised pharmacokinetically and the effect of Pasteurella haemolytica infection on the extent of penetration was studied. Thermoplastic tissue chambers were implanted subcutaneously in the paralumbar fossae of six calves. Thirty-five days after implantation, the tissue chamber distribution of intramuscularly administered erythromycin (30 mg kg -l) was studied. Chambers were then inoculated with P haemolytica and the tissue chamber pharmacokinetics of erythromycin were again studied. Diffusion of erythromycin into tissue chambers was best described using a two-compartment model with tissue chambers representing a relatively inaccessible compartment. Despite changes in chamber fluid pH, the extent of erythromycin penetration into chambers was not affected by P haemolyticainoculation. Comparison of computer simulated concentration-time curves resulting from different routes of administration revealed that penetration of erythromycin into less accessible sites was more likely to be higher after intravenous administration than after intramuscular administration.

THE primary goal of antibacterial therapy is t ° achieve an effective concentration of.drug at the site of infection for a duration sufficient to cause elimination of offending bacteria. The magnitude and duration of concentrations attained at the site of infection are determined by the dose, do sage interval, route of administration and ability of the antibacterial agent to diffuse into the

interstitial fluids of peripheral tissues, where most infections are situated. Ideally, calculation of dosage regimens should be based on predicted tissue concentrations, but these are difficult to estimate because the inaccessibility of the interstitial compartment impedes direct measurement of interstitial drug concentrations. However, tissue concentrations of antibacterial agents may be estimated indirectly by sampling tissue fluid which accumulates within implanted tissue chambers. Chemical and cytological analyses have shown that tissue fluid collected from chambers implanted subcutaneously in cattle closely resembles interstitial fluid (Clarke et al 1989d). Inoculation of Pasteurella haemolytica into subcutaneous tissue chambers implanted in cattle results in an inflammatory response that is characterised by an accumulation of fibrin and a rapid influx of neutrophils (Clarke et al 1989b). These responses are considered to be important components of the pathogenesis of pulmonary lesions seen in bovine shipping fever (Yates 1982, S10combe et al 1985). Studies conducted using sterile and infected chambers have shown that diffusion of intravenously administered drugs between blood and tissue chamber fluid occurs slowly, suggesting that the tissue chamber model rePresents a fairly inaccessible peripheral tissue (Clarke et a11989a, c). Considering the similarities between P haemolytica infection in tissue chambers and pulmonary tissues and the kinetic relationship between tissue chamber fluid and blood, tissue chambers inoculated with P haemolytica

366

Tissue chamber pharmacokinetics can be used as models to study drug distribution into consolidated or abscessed pneumonic lesions. Erythromycin is a macrolide antibiotic which is considered to be effective in vitro against most isolates o f P haemolytica (Fales et al 1982). However, despite an apparent propensity oferythromycin to accumulate within tissues (Burrows et al 1989), the slow rate of absorption following intramuscular administration brings into question the adequacy of the concentration gradient necessary for distribution into inaccessible tissue sites such as pneumonic abscesses. Therefore, the objective of this study was to characterise kinetically the disposition of intramuscularly administered erythromycin in sterile and P haemolytica-infected tissue chambers and to compare intravenous and intramuscular routes of administration using computer simulation. Materials and methods

Animals Six beef calves of mixed breeding, ranging in weight from 109 kg to 227 kg, were determined to be in good health by clinical examination and anamnesis and were acclimatised for two weeks before initiation of the study. They were maintained in small pens for the duration of the experiment and were fed grass and alfalfa hay supplemented with a commercial grain mixture containing 14 per cent protein.

Implantation of tissue chambers Two sterile tissue chambers were implanted subcutaneously in each calf, one in each paralumbar fossa. The perforated cup-shaped chambers were constructed of thermoplastic (Delrin; E~ Du Pont Nemours) and measured 4.6 cm internal diameter, 5.2 cm outer diameter and 1.5 cm in depth. Perforations in the walls and base of the chamber allowed unrestricted exchange of cells and solutes between interstitial and chamber fluids. The top of each chamber was covered with a silicone rubber membrane (Silastic; Dow Corning) through which chamber fluid samples could be withdrawn percutaneously. Tissue chamber assembly, sterilisation, surgical implantation and collection of chamber fluid samples have been described previously

367

(Clarke et al 1989d). The sterility of the surgical technique was monitored after implantation, by culturing an aspirate of chamber fluid aerobically and anaerobically on 5 per cent sheep blood agar.

Pharmacokinetic studies At 35 days after implantation, erythromycin (Gallimycin; CEVA Laboratories) was administered intramuscularly (30 mg kg-1) to each of the calves. Chamber fluid samples were collected by percutaneous aspiration at 0 (before administration), 0.5, two, four, six, eight, 12, 18 and 24 hours after administration. Blood samples were collected at 0 (before administration), 0-25, 0.5, one, two, four, six, eight, 12 and 24 hours after administration. Serum and chamber fluid samples were stored at -20°C until analysis. At 45 days after chamber implantation, chambers were inoculated with a two-hour culture of a field isolate of P haemolytica. A I ml inoculum containing 1 x 108 colony forming units (cFu) 1TI11 was injected into each chamber. Beginning 24 hours after inoculation, a second kinetic study was conducted. Erythromycin dose and sampling intervals were identical to those used before inoculation.

Preparation of inocula A P haemolytica biotype A serotype 1, originally isolated from the trachea of a feedlot calf was used (Corstvet et a11973). Inoculum cultures (stored frozen at 70°C in 60 per cent brain heart infusion IBm] broth and glycerol) were grown overnight on Bm agar (supplemented with 5 per cent citrated bovine blood) at 37°C in an atmosphere of 5 per cent carbon dioxide. Bacteria were scraped from the agar and inoculated into BHIbroth which was incubated at 37°C with rotatory shaking at 80 oscillations min-1 for 1.5 hours. Broth cultures were then centrifuged and bacteria were resuspended in phosphate-buffered saline solution (0.01 M, pH 7-18) at an approximate concentration of 108CFUmll. as determined photometrically (OD650 = 0.207). Measurement of albumin concentrations and pH Albumin concentrations, were measured using a commercially prepared bromcresol green

368

C R. Clarke, D. W. A. Bourne, A. K. Lauer, S. 3". Barron

method (Centrifichem; Baker Instruments) and analyzer (Encore; Baker Instruments). Chamber fluid pH was measured with a pH meter and electrode (Beckman Instruments). Erythromycin assay

Concentrations of erythromycin in serum and chamber fluid samples were measured by agargel diffusion bioassay, using Micrococcus luteus ATe 9341 as the test organism (Bennet et al 1966). Standard curves were constructed by dissolving appropriate amounts of erythromycin in blank matrix (serum, sterile chamber fluid or infected chamber fluid) and assaying them together with the samples. Pharmacokinetic analyses and simulations

Serum and tissue chamber fluid data were analysed simultaneously using a microcomputer program (MULT~-FORT~)for modelling and simulation of pharmacokinetic data (Bourne 1986). Choice of the appropriate pharmacokinetic model was based on the lowest weighted sum of squares. Areas under the computer-generated serum (AUCs) and chamber fluid (AUCi) curves were calculated using trapezoidal approximations between first (0 hour) and last sampling times (Edwards and Penney 1982). The extent of distribution of erythromycin into tissue chambers was described using the ratios of the areas under the chamber fluid and serum curves (AUCi/AUCs). After the appropriate pharmacokinetic model was identified and mean rate constants describing diffusion of erythromycin between inoculated chambers and serum were calculated, these model parameters were used to simulate predicted serum and tissue chamber fluid concentration-time profiles resulting from different routes of administration: intravenous and intramuscular routes were compared using a published estimate of bioavailability (60 per cent, Burrows et al 1989) and complete absorption (bioavailability 100 per cent). Computer simulations were conducted using the MULTI-FORTEprogram. Simulated concentrations resulting from each route of administration were compared with the minimum inhibitory concentration (MIC) for the P haemolytica isolate used (2 ~tg ml 1) (Clarke et al 1992).

Statistical analyses

The effects of infection on erythromycin disposition, albumin concentration, and pH were determined using paired t tests (Wilcoxon's sign rank test for those parameters which were not normally distributed) to compare pharmacokinetic parameters derived before and after P haemolytica inoculation. Differences were considered significant at the P< 0-05 level. Results

Disposition of intramuscular erythromycin was best explained using a two-compartment model with absorption into and elimination from a central compartment representing serum. Tissue chambers were represented by a relatively inaccessible peripheral compartment, which was linked to the central compartment via low rate constants (Fig 1). The model was most clearly identifiable when the distribution rate constants, k12 and k21were held equal. Parameters describing erythromycin disposition are summarised in Table 1. Concentrations of erythromycin in tissue chambers increased slowly, with peak concentrations being achieved approximately 10 to 12 hours after administration. These peak concentrations were considerably lower than corresponding peak serum concentrations, which occurred approximately four hours after administration (Fig 2). Inoculation with P haemolytica did not affect distribution (AUCi/AUCs)of erythromycin into tissue chambers, despite the significantly lower mean pH in infected chambers (Table 2). Other pharmacokinetic parameters were also not affected by infection (Table 1). Computer simulated tissue chamber fluid concentrations, following intravenous and intramuscular administration (Fig 3), revealed that the predicted peak concentration after intravenous dosing was considerably higher than that achieved after intramuscular dosing with 60 per cent bioavailability and that the intramuscular route was unlikely to result in tissue chamber concentrations which exceeded a MICof 2 gg ml-1. Indeed, even with complete absorption following intramuscular administration, the time during which simulated Concentrations exceeded 2 gg m1-1 (7.8 hours) was shorter than the corresponding value following intravenous administration (8.5 hours).

Tissue chamber pharmacokinetics TABLE 1: Mean (_+ SD) pharmacokinetie values derived from simultaneous analysis of serum and tissue chamber fluid collected before and after P haemolytica infection Determinant

Before inoculation

After inoculation

ka (h-1)

0.405 _+0.272

0.472 + 0.386

k12 (h-1)

0.149 +0.158

0.182 + 0.208

kl0 (h-~)

0.148 + 0-015

0-174 _+0.123

Vl/F (litres kg-~)

3.636 -+ 3.004

4.152 + 2.681

V2/F (litres kg-~)

4071 -+ 2.852

5.437 -+ 2-113

MRTs (h)

15.22 + 3.81

14-61 + 2.22

MRTi (h)

20'70 _ 2'65

40"73 -+ 30"88

AUCi/AUCs*

0"33 (0"15-0"40)

0"34 (0"26-0"43)

MRT Mean residence time in serum (s) and chamber fluid (i) AUC Area under the serum (s) and chamber fluid (i) curves * Median (range)

TAB LE 2: Mean (¢ SD)protein concentrations and pH of chamber fluid collected before and after P haemolytica inoculation Determinant

Before inoculation

After inoculation

Albumin (g dim)

1.07 _+0.39

139 +0-62

pH

7.58 _+0.08

7.44 _+0.10*

*Significantly different from value before inoculation

Discussion

Simultaneous analysis of serum and tissue chamber concentration-time data has shown that the relationship between vascular and tissue chamber fluid concentrations of intravenously administered drugs is best explained by a threecompartment model in which tissue chambers represent the third of three compartments arranged in series (Clarke et al 1989a). The twocompartment model identified in the present study is analogous to the three-compartment intravenous model, if the assumption is made that distribution between blood and well-perfused tissues (represented by an additional 'shallow' compartment in the intravenous model)is obscured by slow absorption after intramuscular administration. After both intravenous and intramuscular administration, tissue chambers are kinetically characterised as 'deep' Compartments and are probably representative of less accessible sites such as joints or infected tissues which are consolidated or abscessed. The relatively poor vascularity of these tissue sites together with long diffusional distances can be expected to result in slow drug diffusion and low diffusion rate constants. Indeed, studies of drug penetration into

369

sterile and infected abscesses have revealed disposition profiles which are very similar to those measured in tissue chambers: delayed peak concentrations and elimination half-lives that are longer than corresponding serum elimination half-lives (Joiner et al 1981). Disposition data derived from tissue chamber studies should not be used to predict drug concentrations in well perfused tissues, as this is likely to result in underestimation of peak drug concentrations. The tissue chamber model is relevant only to highly consolidated areas of infected lung, which are usually present in subclinically or chronically infected patients. Therefore, therapeutic strategies based on tissue chamber data should recognise that disposition into tissue chambers represents a worst-case scenario of concentration profiles in the pneumonic lung. Besides the bloodflow, factors affecting distribution of antibacterials into tissues include: the pH gradient between blood and interstitial fluids; the proportion of drug bound by proteins; and the nature of the cellular barriers between the circulation and the site of activity. Bacterial infection of the interstitial space can be expected to alter all of these factors, thus affecting the rate and extent of drug distribution. Despite the lower pH in infected chambers, which should favour ion trapping of basic drugs, inoculation of'P haemolytica did not affect tissue chamber penetration oferythromycin. This was in contrast to an earlier study which demonstrated higher concentrations of trimethoprim and sulphadiazine in P haemolytica-infected chambers than in sterile chambers (Clarke et al 1989c). However, tissue chamber distribution of sulphadiazine/ trimethoprim was measured 36 hours after inoculation, when protein concentrations in chamber fluid were significantly higher than preinoculation levels. In the present study, although protein concentrations appeared to be increasing at 24 hours after inoculation, they were not significantly dif: ferent from preinoculation levels. Examinations of computer-simulated concentration-time curves in Fig 3 Suggests that penetration of drugs into deep tissue sites is more extensive after intravenous administration than after intramuscular administration, particularly when intramuscular bioavailabi!ity and absorption rate constant are low. The relatively raPid elimination of drug from serum after intravenous dosing is compensated for by slower elimination

370

C. R. Clarke, D. W. A. Bourne, A. K. Lauer, S. J. Barton

[Drug at injection 1

site ]k.

l

~?diiii::i~iiii?:i~i~ii~iiiiiii:,iiii?diiiiiii |ili Compartment (1)iil ~ i i

iiiiiil

Peripheral

t~::i::ii::~::~ii::ii~::~::::::i::::i~::~i~i~::~iL::~i~i~::::!~i~i~l

[iiii::iU:i::::i::i;::i::ii::!::i~ii::i::::::ii~::!i::iii~iii~l

-,@® kao

i

:

FIG1: Two-compartment pharmacokinelic model which best described disposition of erythromycin in serum and chamber f l u i d after intramuscular administration. The first-order rate constants kl~ and k~, describe diffusion from compartment 1 into compartment 2 and diffusion from compartment 2 into compartment 1, respectively. • A n a l y s i s o f concentration-time data necessitated estimation of compartment volumes/bieavailability(V~/F for compartment 1 and VJF for cgmpartment 2)

E

4!

I"

r--.... Eo

;~[

"'~'~-----.. ~ ' -

2ii/

0:'-

..........

,0

4

,

8

- .-

........

,

12 Time (h)

16

20

References

24

FiG 2: Mean (+st) erythromycin concentrations in serum (s) and cham-

ber fluid (CF), before (N) and after (I) inoculation with P haemolytica 3-

-E

2

Representative MIC for

L U

0

•

0

-

,

4

-

-

-

,

8

-

-

•

i

•

•

12 Time (h)

-

i

16

-

•

•

,

20

-

.

•

,

for a considerably longer period than do serum concentrations. The therapeutic implication of this dispositional profile is that animals with consolidated or pneumonic lesions may be more likely to respond to intravenous administration of antibacterial agents than to intramuscular administration of equivalent doses. Pneumonic pasteurellosis that has progressed to the stage where consolidated lung fields can be auscultated is notoriously difficult to treat. However, a con; clusion that the intravenous route is preferable to routes requiring absorption is subject to the limitation that the two-compartment model used to simulate intravenous concentration-time data in t h e present study may overestimate tissue chamber concentrations. Direct comparison between the two routes of administration in vivo was not possible because preliminary studies indicated that intravenous administration of gallimycin frequently caused severe systemic reactions. Obviously, in vivo experiments comparing distribution of other antibacterial agents into relatively inaccessible tissues after intravenous and intramuscular administration will have to be conducted before more definitive conclusions can be drawn concerning the most appropriate route and antibacterial agent to be used in cases of chronic pneumonic pasteurellosis.

-

24

FIG 3: Computer generated curves simulating erythromycm concentrations in tissue chamber fluid after intravenous administration, intramuscular administration with 60 per cent bioavailability (intramuscular F=0.6)), and intramuscular administration with completeabsorption (intramuscular F=I)

from the deep site, resulting in effective tissue concentrations which may exceed a given MIC

BENNET, J. V . , B R O D I E , J. L . . BENNER, E. J. & KIRBY, W. M. M. (1966) Simplified, accurate method for antibiotic assay of clinical specimens. Applied Microbiology 14, 170-177 BOURNE; D. W. A. (1986) Multi-fore, a microcomputer program for modelling and simulation of pharmacokinetic data. Computer Methods and Programs in Biomedicine 23, 277-281 BURROWS, G. E., GENTRY, M. & EWING, B. S. (1989) Serum and tissue concentrations of erythromycin in calves with induced pneumoulc pasteurellosis. American Journal of Veterinary Research 50, 1166-1169 BURROWS, G. E., GRIFFIN, D. D., PIPPIN, A. & HARRIS, K. (1989) A comparison of the various routes of administration of erythromycin in cattle. Journal of Veterinary Pharmacology and Therapeutics 12, 289-295 CLARKE, C. R., BARRON, S. J., AYALEW, S . & BURROWS, G. E. (1992) Response of Pasteurella haemolytica to erythr0mycin and dexamethasone in calves with established infection. American Journal of Veterinary Research 53, 684-688 CLARKE, C. R., SHORT, C. R., BOURNE, D. W. A. & USENIK E. A. (1989a) Subcutaneously implanted tissue chambers - a pharmacokinetic study. Journal of Veterinary Pharmacology and Therapeutics 12, 312-321 CLARKE. C. R.. SHORT. C. R., CORSTVET, R. E. & NOBLES, D. (1989b) Interaction between Pasteurella haemolytica, sulfadiazine/trimethoprim, and bovine viral diarrhea virus. American Journal of Veterinary Research 50, 1557-1565 CLARKE, C. R., SHORT, C. R., CORSTVET, R. E. & NOBLES, D. (1989c) Effect of Pasteurella haemolytica infection on the distribution of sulfadiazine and trimethoprim into tissue chambei:s

Tissue c h a m b e r p h a r m a c o k i n e t i c s implanted subcutaneously in cattle. American Journal of Veterinary. Research 50, 1551-1556 CLARKE, C. R., SHORT, C. R., USENIK, E. A. & RAWLS, R. (1989d) Subcutaneously implanted tissue chambers: a pathophysiological study. Research in Veterinary Science 47, 195-202 CORSTVET, R. E., PANCIERA, R. J., RINKER, H. B., STARKS, B. & HOWARD, C. (1973) Survey of tracheas of feedlot cattle for Haemophilus somnus and other selected bacteria. Journal of the American Veterinary Medical Association 163, 870-873 EDWARDS, C. H. & PENNY, D. E. (1982) Calculus and Analytical Geometry. New Jersey, Prentice-Hall FALES, W. H., SELBY, L. A., WEBBER, J. J., HOFFMAN, L. J., KINTNER, L. D., NELSON, S. L., MILLER, R. B., THORNE, J. G., MCGINTY, J. T. & SMITH, D. K. (1982) Antimicrobial resistance among Pasteurella spp recovered from Missouri and Iowa cattle with bovine respiratory disease complex.

371

Journal of the American Veterinary Medical Association 181, 477479 JOINER, K. A., LOWE, J. L., DZINK, J. L. & BARTLETT, J. G. (1981) Antibiotic levelsin infected and sterile subcutaneous abscesses in mice. Journal of lnfectious Diseases 143, 487-494 SLOCOMBE, R. F., MALARK, J., INGERSOLL, R,, DERKSEN, F. J. & ROBINSON, N. E. (1985) Importance of neutrophils in the pathogenesis of acute pneumonic pasteurellosis in calves. American Journal of Veterinary Research 46, 2253-2258 YATES, W. D. G. (1982) A reviewof infectious bovine rhinotracheitis, shipping fever pneumonia and viral-bacterial synergism in respiratory disease of cattle. Canadian Journal of Comparative Medicine 46, 225-263 Received February 11, 1992 Accepted November 10, 1992