We have shown that the above-mentioned substances penetrate into the vitreous body of rabbits through an implanted microdialysis membrane. This is of importance for the development of new means of intraocular drug administration.

The method of microdialysis was used for collecting series of samples from the rabbit vitreous after systemic and intravitreous administration of ceftazidime. The purpose of the study was to compare the method with traditional pharmacokinetic sampling. Ceftazidime was injected intramuscularly (1 mg/kg) or intravitreally (1 mg) in rabbits, with a previously implanted microdialysis probe in the vitreous. The membrane was perfused with a buffer, and the dialysate was collected in samples where the concentration of the drug was analyzed by HPLC. After intramuscular administration, blood samples were taken to calculate systemic pharmacokinetics. The same procedures were repeated with rabbits with mild intraocular inflammation induced by the injection of 400 EU of endotoxin into the vitreous, 12-15 hr before drug administration. Pharmacokinetic parameters, such as half-life and AUC, were calculated. The penetration into the vitreous after intramuscular injection was higher (42%) in inflamed than in non-inflamed eyes (20%), suggesting an interference with the blood retinal barrier. Other kinetic parameters did not differ significantly between the groups. The advantage of the method is that fewer experimental animals can be used to obtain the necessary data compared to traditional pharmacokinetic methods. In conclusion, intraocular dialysis with chronically implanted probes is a technique well suited for pharmacokinetic studies of systemically administered ceftazidime or other drugs that will pass a dialysis membrane.

The technique of microdialysis in vivo, much used in brain experiments, has been adapted for intraocular use. A new probe was designed, made from a soft tube with the dialysis membrane mounted in a fenestrated protecting sleeve, facing one side. A special surgical procedure was developed for the long-term implantation of the probes. They stay functional in the eyes for several weeks or more. Previously published intraocular microdialysis studies were short-term experiments. The acute surgical trauma is likely to affect the concentration of many compounds in the vitreous, and this effect is minimized with long-term probes.

ABSTRACT.The treatment of cytomegalovirus retinitis has during the last years been done mainly with sustained release ganciclovir devices implanted in the vitreous. In the present study it is shown that ganciclovir can be administered into the rabbit vitreous by microdialysis. A concentration of about 10 ª6 M, which is considered within the therapeutic range, is achieved in the vitreous after a microdialysis perfusion. The method offers possibility of variation in the drug delivery that is not possible with the ganciclovir implants.

ABSTRACT. Intraocular microdialysis was used to administer the drugs 5-fluorouracil, benzyl penicillin, daunomycin and dexamethasone into the vitreal space of rabbits. The purpose of the study was to investigate if therapeutic concentrations could be obtained with this administration method. After administering the drugs in labelled form, the attained concentrations were assessed by counting the radioactivity in the entire vitreous. For benzyl penicillin, the concentration was 2 pM, for dexamethasone it was 1.2 x lop7 M, and for daunomycin it was 1.2 pM, which are considered to be within the therapeutic ranges. For 5-fluorouracil, the corresponding concentration was 5 X M which probably is below the therapeutic level, when comparing with single-dose injections.