Pharmacokinetics and therapeutic efficacy of gentamicin in an experimental pleural empyema rabbit model

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, July 1987, P. 982-985

Vol. 31, No. 7

0066-4804/87/070982-04$02.00/0

Copyright © 1987, American Society for Microbiology

Pharmacokinetics and Therapeutic Efficacy of Gentamicin in Experimental Pleural Empyema Rabbit Model

an

SHOHET, A. YELLIN, J. MEYEROVITCH, AND E. RUBINSTEIN* Infectious Diseases Unit, Chaim Sheba Medical Center, Tel Aviv University Medical School, Tel Hashomer, Israel Received 6 August 1986/Accepted 10 April 1987 I.

The pharmacokinetics and therapeutic efficacy of gentamicin

were

investigated in

an

experimental pleural

empyema rabbit model. Pleural effusion was induced by the intrapleural administration of turpentine, and empyema was induced by direct inoculation of the effusion with Kkebsiella pneumoniae. Pleural empyema compared with effusion was characterized by lower pH, oxygen tension (PaO2), and glucose levels and higher leukocyte count, lactic acid concentration, and PaCO2. After a single administration, gentamicin was first detectable in the pleural fluid at 60 min, whereas peak levels in empyema were observed at 180 min. Gentamicin persisted in the empyema longer than in blood. Animals treated with gentamicin only had 60% bacterial cure on day 7; those treated with gentamicin in an oxygen chamber had 100% cure on day 5 (P = 0.004). Low oxygen tension diminished the antibacterial efficacy of gentamicin in this model. An increase in oxygen tension

improved the therapeutic results without alterationi of the pharmacokinetics of gentamicin.

Pleural empyema poses a serious medical problem with a high rate of complications. Although the optimal treatment of pleural empyema has not been well defined, it is usually prolonged and often requires drainage. Staphylococcus aureus, Streptococcus pneumoniae, and anaerobic bacteria are the organisms most frequently isolated from empyema after pneumonia, whereas aerobic gram-negative enteric bacilli are the most predominant pathogens in postoperative patients (4,20, 22, 23). Defective bacterial opsonization, absence or decrease in intact complement, and decreased leukocyte functions may all contribute to the low rate of spontaneous cure of pleural empyema (7, 21). Insufficient antibiotic penetration and other unfavorable local conditions may further hamper the success of antibiotic therapy, necessitating surgical intervention (10). In this investigation, we evaluated the pH, oxygen tension (PaO2), PaCO02, glucose, lactic acid concentrations, and leukocyte count of experimental empyema fluid in rabbits. The therapeutic effect of gentamicin alone and gentamicin together with oxygen was investigated as well.

ml of turpentine through the cannula, followed by 0.4 ml of normal saline to flush the dead space of the cannula, by the method of Sahn et al. (17). The quantity of turpentine was selected on the basis of amounts necessary to produce pleural effusion in dogs, rats, and rabbits (9, 17, 18). After the administration of turpentine, the animals were rotated for 1 min to assure turpentine contact with a wide area of the pleura. Turpentine produced a sterile, exudative pleural effusion that could be sampled through the rubber stopper within a few hours after instillation. At 24 and 96 h after turpentine instillation, samples from the pleural fluids of each rabbit for baseline measurements of leukocyte count, pH, PaO2, PaCO2, glucose, and lactate concentration and for culture were obtained. Immediately thereafter, 109 phosphate-buffered saline-washed Klebsiella pneumoniae organisms in 1 ml of normal saline were injected into the pleural space through the cannula (9). Samples from the pleural fluid which became purulent were then obtained from each rabbit at daily intervals. The K. pneumoniae strain had been isolated from the empyema fluid of a patient; it was gentamicin susceptible (MIC and MBC, 0.4 ,ug/ml). The MIC was determined by the inoculation of 105 CFU of the organism into tubes containing 1 ml of Mueller-Hinton broth and decreasing concentrations of gentamicin. The tubes were incubated overnight at 37°C, and the concentration in the first tube in which no visible growth occurred was considered the MIC. The MBC was determined by the subculture of 0.05 ml of broth from clear tubes onto MuellerHinton agar and incubation at 37°C for 24 h. The MBC was determined when no colonies were detectable on the agar (detection range, 0.1 to 200 ,ug/ml). The strain, maintained on a slant, was grown overnight before each experiment in tryptic soya broth at 370C. This strain, a known laboratory strain, did not change its degree of encapsulation and virulence to mice throughout the experiment. Gentamicin levels in blood, pleural effusion, and empyema fluids were measured after a single intramuscular injection of 5 mg/kg (body weight). Samples from pleural fluids (0.2 ml) were obtained at 15-min intervals, parallel to blood samples (0.4 ml). After the induction of empyema, rabbits were randomly assigned to three groups as follows.

MATERIALS AND METHODS Fifteen male and female New Zealand White rabbits weighing 2 to 3 kg were anesthetized with ketamine (Ketalar; Parke, Davis & Co.) and xylazine (Rompun; Bayer AG), their chest fur was clipped, and the underlying skin was scrubbed with iodine. A 16-gauge Venflon cannula capped with a rubber stopper to prevent air entry was introduced into the, right pleural space through a longitudinal paravertebral skin incision performed between ribs 8 and 9 in a method described by Sahn and Potts (16). The cannula was placed by being pointed caudally. After insertion of the cannula, the cap was removed for a few seconds, allowing air to be sucked into the pleural space and thus creating a pneumothorax, which was evacuated immediately by suction. The cannula was then secured to the subcutaneous tissue, and the overlying skin was closed with Dexon sutures. Pleural effusion was induced by the instillation of 0.3 *

Corresponding author. 982

VOL. 31, 1987

GENTAMICIN PHARMACOKINETICS IN PLEURAL EMPYEMA MODEL

(i) Rabbits treated with gentamicin (group 1). Five rabbits were administered gentamicin at 5 mg/kg intramuscularly, divided into three daily doses beginning 24 h after the induction of empyema. Empyema samples were obtained at 24-h intervals for measurements of pH, oxygen and carbon dioxide content, glucose and lactate levels, and leukocyte count and for culture. Pleural samples were obtained by puncturing the rubber stopper with sterile syringes and after meticulously cleaning the skin with tincture of iodine. The barrel of the cannula was first flushed by 0.4 ml of sterile saline, and then 2-ml pleural fluid samples were drawn. None of the specimens obtained was blood tinged. Care was taken to avoid the'entry of air into the specimens. Gentamicin therapy was continued until no more fluid could be obtained from the pleural space. Thereafter, lavage of the pleural space with 3 ml of normal saline was performed, and cultures from the lavaged fluid were obtained. All the animals were sacrificed at the end of the experiment, and autopsies were performed with bacteriological cultures of the pleural cavity and lung. (ii) Rabbits treated with gentamicin and oxygen (group 2). Five rabbits were caged separately in closed chambers and received oxygen at a rate of 5 to 7 liters/min, inducing a chamber pressure of 760 to 800 mm Hg (101 x 103 to 107 x 103 Pa). In addition, these animals received a regimen identical to that of group 1. Pleural fluids and blood samples were obtained while the animals remained in the closed chamber. (iii) Rabbits without therapy (group 3). Five control rabbits received no therapy. Samples and autopsies were performed as for group 1. Determinations of pH, PaO2, and PaCO2 were performed by using an acid-base analyzer-meter after standard calibration procedures. Pleural fluid samples were placed on ice at the time of collection and were measured within 30 min. Pleural fluid samples for glucose determination were' frozen immediately to prevent glycolysis and were measured by the ferricyanide technique. Lactate concentrations were determined by the enzymatic method, and protein was measured with a temperature-compensated refractometer. Gentamicin in blood and pleural fluid was determined by solid-phase radioimmunoassay (Diagnostic Products Corp., Los Angeles, Calif.) with a detection limit of 0.1 ,ug/ml. The presence of pus did not interfere with radioimmunoassay results. The half-life of gentamicin in serum and pleural fluids was measured by the trapezoid method. For statistical analysis, the Fisher exact probability test was used. RESULTS Peak gentamicin levels in serum occurred at 15 min, and peak concentrations in pleural fluid occurred at 180 min after administration (Fig. 1). The mean peak concentration in pleural effusion was 30% of the peak gentamicin concentration in serum. Gentamicin persisted longer in the pleural fluid than in the blood. The half-life of gentamicin in the blood was 75 min, whereas in the pleural fluid it was 180 min. Gentamicin concentrations in the effusion and empyema fluids were similar. Group 1. Pleural fluid characteristics for group 1 during the therapy period are shown in Table 1. Marked increases in total leukocyte count and percent polymorphonuclear leukocytes, PaCO2, and lactate levels were found in the empyema fluid compared with the pleural effusion. PaO2 and glucose levels were markedly depressed in the empyema

983

fluid. Protein concentration showed no significant change between the sterile effusion and empye,ma fluid. Empyema fluid disappeared in three rabbits on day 7 of therapy and in two rabbits on day 8 of therapy. Lavage fluids of the pleural cavities were sterile, and autopsy showed no significant findings in the pleural cavities and lungs. Group 2. Pleural fluid characteristics for group 2 during the therapy period (Table 1) were similar to those of group 1. Empyema fluid was undetectable in one rabbit on day 3 of therapy, in three rabbits on day 4, and in one animal on day 5. Cultures of the lavage fluids' after the disappearance of the empyema fluid were all sterile. Autopsy findings of the manipulated pleural spaces and lungs were unremarkable. Group 3. No pleural fluid could be aspirated on day 9 in four rabbits in group 3 and in one on day 10. Lavage fluid of the pleural cavities grew K. pneumoniae in two rabbits. Autopsies showed pleural abscess formation in all rabbits. DISCUSSION Our model mimics infected human pleural empyema fluid in the following characteristics: low pH of the infected fluid compared with the noninfected effusion (12), high leukocyte count and lactic acid levels (2), and low glucose levels (6). These features assist in the differentiation of infected and noninfected pleural effusions. The cure rate of rabbits treated with gentamicin and oxygen was found to be more rapid than those of the control group and the gentamicin-treated group. On day 4 of therapy, all rabbits in groups 1 and 3 were found to be infected in comparison with only 20% of group 2 (P = 0.02) (Fisher exact probability test). On day 5 of therapy, positive cultures of the empyema fluid were found in all rabbits in groups 1 and 3 and in none in group 2 (P = 0.004). The pharmacokinetics of gentamicin were not tested separately in groups 1 and 2. According to our previous experience (15), they do not change under such conditions. Continuous drainage was not used as part of the experimental protocol. All interpretations of antibiotic levels need

18_ 1 16

142

10_

/

4 2

-.

MINUTES

FIG. 1. Blood and pleural gentamicin levels after single injection of gentamicin (5 mg/kg).-, Level in blood; --- -, level in level in pleural effusion. pleural empyema; .

984

SHOHET F-,T AL.

ANTIMICROB. AGENTS CHEMOTHER.

TABLE 1. Biochemical values, leukocyte counts, and cultures of pleural effusion and empyemaa Empyema

Characteristic

Day

Effusion

Genttmicin treated

Gentamicin and oxygen treated

3.6 ± 4.6 1.8 ± 2.0 45 ± 3

7.5 ± 5 12 ± 4.3

6.82 + 0,06 6.78 t 0.03 6.77 ± 0.05

0.94 t 0.13 6.9 t 0.07 7.0 ± 0.15

6.96 ± 0.16 6.98 t 0.1 105.6 ± 34.7 63.5 ± 17

Cootrol Glucose (mg%O)

1 3 5

46

t

7.5

pH

1 3 5

7.16

t

0.05

PaO2 (mm Hg)b

1 3 5

75 ± 9.5

16 ± 4.0 11.5 ± 3.5 10 ± 8.3

59 ± 20.2 29.4 t 15.4 20.6 ± 3.0

PaCO2 (mm Hg)b

1 3 5

40 ± 5.4

80 ± 6.0 95 ± 3.0 91 ± 8.3

49.4 ± 28.9 86.6 ± 11.7 74.3 ± 13.6

Lactic acid (meq/liter)

1 3 5

72 ± 12

89.3 ± 10.4 106 ± 8.2 128 ± 8.6

112.6 ± 6.4 121.2 ± 8.9 109 ± 16

Leukocyte count

1 3

37,200 ± 14,100

24,800 ± 3,780

96

78

90

5/5 5/5 5/5 5/5 5/5 1/5 0/5

5/5 S/S 5/5 2/5 0/5

5/5 4/5 0/5

% Polymorphonuclear leukocytes Positive cultures (no. of rabbits/total no.)

a

b

Values are means

+

5 t 1.2 5 ± 2.2 0

65 ± 4.5 78 ± 5

80.2 ± 8.4 68.5 ± 14.5

3,900 ± 440

1 3

35

1 3 5 7 8 9 10

0/5

22,760

+

8,200

standard deviation.

Metric equivalent for 1 mm Hg is 133.3 Pa.

to be couched by the proviso that drainage was not used in this model, yet sampling could probably be regarded as intermittent though insufficient drainage. Our results demonstrate that gentamicin penetrates well into pleural effusion and pleural ipmpyema, as was previously reported in humans (13,'14). Oxygen therapy did not influence drug penetration into the pleural cavity, suggesting an oxygen-independent drug entry mechanism. The area under the curve of gentamicin concentration in serum was 1.47 times that of pleural effusion and 1.29 times that of empyema. Prolonged persistence of antibiotics in inflammatory exudates was previously described for several antibacterial agents (1, 8, 24). It is possible that in our model the leukocytes in the pleural cavity bound the' aminoglycoside and thus prolonged its half-life (3). Another possible explanation may be the geometry of the pleural site which limits the surface area available for diffusion relative to the volume of fluid collection. This may have produced a reservoirlike or trapping phenomenon that was previously described (5). Despite the therapeutic levels of gentamicin in the pleural fluid, the drug failed to exert its expected therapeutic effect (11, 15, 24). The explanation for the difference in the action of gentamicin in pleural empyema of animals treated with and without oxygen may be the increase in' MIC and MBC for the offending organism in anaerobic conditions. It has been reported that under anaerobic conditions the antibac-

terial activity of aminoglycosides decreases because of the inability of these compounds to bind to the important 30S ribosome. It is our impression that the low PaO2 in the empyema fluid was the main reason for the inefficacy of gentamicin in the present experimental model (15). The delivery of oxygen to areas of bacterial invasion may be, an essential prerequisite not only for certain antimicrobial agents to exert their activity but also for effective phagocytic killing of invading bacteria (19). The results of this study suggest that gentamicin reacjes therapeutic and prolonged concentrations in pleural effusion and empyema. Despite the adequate levels and susceptibility of the microorganism, the resolution of the empyema was slow, but it could be accelerated by combined treatment with gentamicin and oxygen. LITERATURE CITED 1. Bergan, T. 1981. Pharmacokinetics of tissue penetration of antibiotics. Rev. Infect. Dis. 3:45-66. 2. Brook, I. 1981. Lactic acid lvels in pleural empyema. Respiration 40:344-349. 3. Bryant, R. E., and D. Hammond. 1974. Interaction of purulent material with antibiotics used to treat Pseudomonas infections. Antimicrob. Agents Chemother. 6:702-707. 4. Emerspn, J. D., I. B. Bormchow, G. R. Daicoff, T, D. Bartley, and M. W. Wheat. 1971. Empyema. J. T4orac. Cardiovasc.

VOL . 31, 1987

GENTAMICIN PHARMACOKINETICS IN PLEURAL EMPYEMA MODEL

Surg. 62:967-972. 5. Etta, L. L. V., L. R. Peterson, C. E. Fasching, and D. N. Gerding. 1982. Effect of the ratio of surface area to volume on the penetration of antibiotics into extracellular spaces in an in-vitro model. J. Infect. Dis. 146:423-428. 6. Hirsch, A., P. Ruffle, M. Nebut, J. Bignon, and J. Chretien. 1979. Pleural effusion: laboratory tests in 300 cases. Thorax 34:106-112. 7. Lew, P. D., R. Zubler, P. Vaudaux, J. J. Farquet, F. A. Waldvogel, and P. H. Lambert. 1979. Decreased heat labile opsonic activity and complement levels associated with evidence of C3 breakdown products in infected pleural effusions. J. Clin. Invest. 63:326-334. 8. McCune, R. 1960. Delivery of antimicrobial drugs across inflammatory membrane in rabbits. J. Clin. Invest. 39:846-853. 9. Menkin, V. 1934. Studies of inflammation. The cytological picture of an inflammatory exudate in relation to its hydrogen ion concentration. Am. J. Pathol. 10:193-210. 10. Morin, J. E., D. D. Munro, and L. D. McLean. 1972. Early thoracotomy for empyema. J. Thorac. Cardiovasc. Surg. 64:530-536. 11. Nichols, R. L., J. W. Smith, E. N. Fossedal, and R. E. Condon. 1979. Efficacy of parenteral antibiotics in the treatment of experimentally induced intra-abdominal sepsis. Rev. Infect. Dis. 1:302-309. 12. Potts, D. E., D. C. Levin, and S. A. Sahn. 1976. Pleural fluid pH in parapneumonic effusions. Chest 70:328-331. 13. Riff, J. J., and G. G. Jackson. 1971. Pharmacology of gentamicin in man. J. Infect. Dis. 124:598-606. 14. Ristuccia, A. M., and B. A. Cunha. 1982. The aminoglycosides.

985

Med. Clin. North Am. 66:303-312. 15. Rubinstein, E., Z. Dreznick, and Z. Mark. 1982. Gentamicin and cefsulodin efficacy in a rat abscess model. Surg. Gynecol. Obstet. 155:363-368. 16. Sahn, S. A., and D. E. Potts. 1978. Turpentine pleurisy in rabbits: a model of pleural fluid acidosis and low pleural fluid glucose. Am. Rev. Respir. Dis. 118:893-901. 17. Sahn, S. A., D. A. Taryle, and J. T. Good. 1979. Experimental empyema. Am. Rev. Respir. Dis. 120:355-361. 18. Spector, W. G., and D. A. Willoughby. 1959. The demonstration of the role of mediators in turpentine pleurisy in rats by experimental suppression of the inflammatory changes. J. Pathol. Bacteriol. 77:1-17. 19. Spitznagel, J. K. 1983. Microbial interactions with neutrophils. Rev. Infect. Dis. 5(Suppl. 4):806-812. 20. Sullivan, K. M., R. D. O'Toole, R. H. Fisher, and K. N. Sullivan. 1973. Anaerobic empyema thoracis. Arch. Intern. Med. 131:521-527. 21. Suter, S., V. E. Nydegger, L. Roux, and F. A. Waldvogel. 1981. Cleavage of C3 by neutral proteases from granulocytes in pleural empyema. J. Infect. Dis. 144:499-508. 22. Varkey, B., H. D. Rose, K. Kutty, and J. Politis. 1981. Empyema thoracis during a ten year period. Arch. Intern. Med. 141: 1771-1776. 23. Weese, W. C., E. R. Shindler, I. M. Smith, and S. Rabinovitch. 1973. Empyema of the thorax then and now. Arch. Intern. Med. 131:516-520. 24. Weinstein, W. M., A. B. Onderdonk, J. B. Bartlett, T. J. Louie, and S. L. Gorbach. 1979. Antimicrobial therapy of experimental intra-abdominal sepsis. J. Infect. Dis. 132:282-286.