Antibiotic levels in serum are.commonly used to guide ahtibiotic therapy. The (7) and Raebtirn (5) studied interstitial fluid antibiotic levels in the abraded forearms of human volunteers. Animal models in which subcutaneous tissue cages were used to measure the antibiotic levels in interstitial fluid have been described (2, 4). There is no model, however, which allows measurement of the in vivo killing of bacteria along with a determination of antibiotic levels in tissues.We developed a canine model that allows measuremnent of the antibiotic levels in interstitial fluid and the bacterial killing rate by using a modified Sykes-Moore chamber. This model can be used to study the penetration of antibiotics into body tissues and to compare the in vivo efficacy of antibiotics.MATERIALS AND METHODS Adult mongrel dogs of either sex ranging in weight from 14 to 34 kg were used. They were placed under general anesthesia by induction with sodium pentothal, orotracheally intubated, and maintained with intravenous sodium pentothal, pentobarbital, and lactated Ringer solution with 5% glucose. The dogs were routinely maintained under general anesthesia for 9 h without difficulty.Each animal was placed in the supine position, and its abdomen was shaved and prepped with povidone-iodine solution. Oblique abdominal skin incisions were made bilaterally. The external oblique muscle belly was freed by blunt and sharp dissection from the surrounding tissues with the origin and insertion of the muscle left intact. Bleeding from the skin and subcutaneous tissue was meticulously con-* Corresponding author. trolled with electrocautery. This dissection produced four potential spaces in the tissue planes: between the skin and anterior surface of the external oblique muscle and between the internal oblique muscle and the posterior surface of the external oblique muscle, bilaterally. These pockets were used to place the modified Sykes-Moore chamber in intimate contact with the muscle belly.Our Sykes-Moore chambers consist of two stainless steel rings with a glass cover slip, a self-sealing rubber gasket, a membrane filter (pore size, 0.22 ,um; Millipore Corp., Bedford, Mass.), a surface area of 3.14 cm2, and a volume of 0.8 ml (Fig. 1). The rubber gasket allows sampling of the chamber fluid. When the gasket is placed with the Millipore filter in contact with body tissues, the filter allows the free diffusion of interstitial fluid into the chamber but is impermeable to bacteria. The chambers were assembled and then steam autoclaved at 121°C under 15 lb (ca. 6.8 kg) of pressure for 15 min.The chambers were inoculated with clinical isolates of either Staphylococcus aureus or Escherichia coli immediately before being placed into the dog. The bacteria had been grown for 18 h in Trypticase soy broth (BBL Microbiology Systems, Cockeysville, Md.). A 1:100 dilution of bacteria in phosphate-buffered saline (PBS) at a pH of 7.4 was prepared, and 0.8 ml of this suspension was injected into the chamber through the rubber gasket. The concentration of S...