Experimental intraoperative use of perfluorotributylamine, perfluorodecaline, and perfluoropolyether

The perfluorocarbon Iiquids (PFCLs) are a group of compounds with a specific gravity greater than water that are used as an adjunct in the surgical management of complicated vitreoretinal diseases. Thirty-two pigmented rabbits underwent vitrectomy with injection of one of three PFCLs or saline into the vitreous cavity: eight eyes received perfluorotributylamine (PFT A) ; eight eyes received perfluorodecaline (PFDC) ; eight eyes received perfluoropolyether (PFPE) ; the remaining eight eyes, which served as controls, received saline. One hour after placement in the eyes the PFCLs were removed from the vitreous cavity and the eyes were filled with saline. The rabbits were then euthanized and the eyes enucleated and processed for histologic examination. Light microscopy demonstrated no significant difference between eyes with the PFCLs compared with those of control animais. Electron microscopy revealed only minor changes, consisting of a few irregulary shaped defects in the outer segment discs of photoreceptors, in the PFDC group only. These findings suggest that short-term intraoperative use of PFT A, PFPE, and PFDC, at least in this experimental animal model, is not associated with any significant retinal pathology.


ABSTRACf
The perfluorocarbon Iiquids (PFCLs) are a group of compounds with a specific gravity greater than water that are used as an adjunct in the surgical management of complicated vitreoretinal diseases. Thirty-two pigmented rabbits underwent vitrectomy with injection of one of three PFCLs or saline into the vitreous cavity: eight eyes received perfluorotributylamine (PFT A) ; eight eyes received perfluorodecaline (PFDC) ; eight eyes received perfluoropolyether (PFPE) ; the remaining eight eyes, which served as controls, received saline. One hour after placement in the eyes the PFCLs were removed from the vitreous cavity and the eyes were filled with saline. The rabbits were then euthanized and the eyes enucleated and processed for histologic examination. Light microscopy demonstrated no significant difference between eyes with the PFCLs compared with those of control animais. Electron microscopy revealed only minor changes, consisting of a few irregulary shaped defects in the outer segment discs of photoreceptors, in the PFDC group only. These findings suggest that short-term intraoperative use of PFT A, PFPE, and PFDC, at least in this experimental animal model, is not associated with any significant retinal pathology.

INTRODUCfION
Vitreoretinal surgeons have long sought a vitreous substitute that is transparent, viscous, and heavier than water, that does not interfere with refraction and, most of all, is not toxic to the retina or other structures of the eye. Currently, silicone oil is most widely used as a vitreous substitute, although there are some problems with its use and controversy over its possible toxicity continues(l-S)_ ln addition, silicone oil has a low specific gravity, and thus is not of benefit in the repair of inferior breaks or for intraoperative management of complicated retinal detachments(9,IO) .
Liquid perfluorochemicals meet many of the cri teria that are desirable in a vitreous substitute, and several have been tested as potential vitreous substitutes(II-IS). Unfortunately, a number of these agents cause ocular toxicity or other problems(1 2 ,IS) .
Recent1y CHANG and coworkers<19-22) described the use of perfluorocarbon liquids (PFCLs) for unfolding retinal flaps and in management of prolifera tive vitreoretinopathy. PFCLs, when used experimentally, tend to cause retinal toxicity at about 30 days following injection,<12,14,16) and thus are used only in the intraoperative period, and are removed once the surgical goal of retinal reattachment has been achieved.
The purpose of this study was to evaluate any structural changes in the retina caused by three different PFCLs after one hour in rabbit eyes.

MATERIALS AND METHODS
All experiments adhered to the Association for Research in Vi sion and Op hthalmology Resolution on the Use of AnimaIs in Research. Thirty-two Dutch cross rabbits (2 to 3 kg) were used to evaluate the one hour ocular tolerance to three different PFCLs injected intraoperatively. Perfluorotributylamine (PFr A), perfluorodecaline (PFDC), and perfluoropolyether (PFPE) were the PFCLs studied. Each of these PFCLs was injected into eight eyes; as controls, eight eyes were similarly injected with saline.
Maximum mydriasis was achieved by topical application of phenylephrine hydrochloride 2.5% and tropicamide 1 %. AnimaIs were anesthetized with an intramuscular injection of 0.75 mL/kg ketamine hydrochloride (100 mg/mL) and xylazine hydrochloride (20 mg!mL) in a 1: 1 solution. Proparacaine hydrochloride 0.5% was used as a topical anesthetic. Proptosis was obtained by a retrobulbar injection of 1.0 cc of lidocaine hydrochloride and sterile water (1: 1 solution). Eyes were prepared for surgery in a ARQ. BRAS. OFfAL. 55, (3), 1992 Exp erimental intraoperative use 01 perfluorotributylamine, perfluorodecaline, and perfluoropolyether sterile fashion. Pars plana vitrectomies were performed using an Ocutome system under an operating microscope (Zeiss Co., Model OPMI-6, Oberkochen, Germany) using a planoconcave vitrectomy lens. Following a superior fomix based peritomy, two 20-gauge sclerotomies were made at a distance of 2-3 mm posterior to the comeoscleral 1imbus at the 10 and 2 o'clock positions. A 4-mm infusion cannula, attached to a bottle of saline, was fi xed in the temporal sclerotomy with a preplaced 6/0 vicry1 suture. The posterior vitreous was retnoved as completely as possible prior to the injections. The PFCLs were transferred from the sterile containers to a 5-cc syringe through a 20-gauge needle and permeable membrane filters (Millipore, pore size=0.22 �m). Fluid-fluid exchange was performed by injecting each studied PFCL intravitreal1y through the infusion cannula. Since PFCLs are heavier than saline, the former moved to the posterior pole, expelling the lighter saline through the nasal sclerotomy. Injection of the PFCL continued until it was seen to be exiting from the nasal sclerotomy site. The infusion cannula was then removed and the sclerotomy closed with a 6/0 vicry1 suture. If the eye were hypotonus, a small amount of PFCL was injected through the pars plana via a 25-gauge needle. The entire exchange procedure required approximately 2 cc to 3 cc of the heavy liquid as replacement; with this method, we were able to replace 50% to 60% of vitreous with one of the three above mentioned liquids. The PFCLs were removed from the eye exactly one hour after injection. To achieve this procedure, a 25-gauge needle, connected to an elevated (2 ft) bottle of infusion fluid, was introduced into the eye via the temporal pars plana. The nasal sclerotomy was reopened and a suction probe was used to aspirate the PFCL. White the infusion was flowing, the suction probe was directed to the posterior pole and PFCL aspirated with 100 mmHg suction pressure. Following this procedure, the needle and the probe were withdrawn; the sclerotomy was closed with a 6/0 vicry1 suture and the conjunctiva was stroked back to the limbus. AnimaIs were then sacrificed with an intracardiac overdose of pentobarbital (100 mg! kg) under deep anesthesia, and eyes were enucleated for histopathologic studies.
For light and electron microscopy, eyes were fixed in half strength Kamovsky's fixative. Specimens for light microscopy were embedded in Historesin (LKB) and 3 � sections were cut and stained with toluidin blue. For electron microscopy, tissues were processed in Polybed and 70 Á sections were cut, post-stained with urany1 acetate and lead citrate and viewed with a Zeiss EM 10 electron microscope.

RESULTS
No complications were encountered during injection or after removal of the PFCLs. All of the eyes, both experimental and control, demonstrated a similar appearance on light microscopy (Fig. 1). The size, shape and number of nuclei in both the outer and inner nuclear layers were within normal limits. The thickness of the retina corresponded to that of normal retina, and the ganglion cell layer showed no abnormalities.
Exp erimental intraoperative use of perfluorotributylamine, perfluorodecaline, and perfluoropolyether  However, two of the eight eyes in which PFDC had been injected showed irregular1y shaped defects in the outer segment discs of the photoreceptors.

DISCUSSION
Th e search for a new and better vitreous substitute has intensified recent1y because of problems related to the use of intra ocular silicone oil that have recently been reported (23)(24)(25). Perfluorocarbon derivatives are generally inert, and many of them are used as blood substitutes because they are good carriers of oxygen and carbon dioxide (26)(27)(28)(29).
PFPE,PFr A, and PFDC are fluorinated, synthetic, transparent compounds that provide excellent thermal and chemical stability. Their high specific gravity (1.88 to 1.94 g/mL) provi de excellent mechanical tamponade, especially for inferior retinal breaks. The low surface tension (16-20 dynes/cm2) results in a high interface tension with saline, which limits their passage through retinal breaks. Th eir lower viscosity makes them very easy to handle, and they can be removed easily from the eye at any time during the surgical procedure. The higher viscosity (68 centistokes) of PFPE prevents formation of the fishegg phenomenon, which is so common with PFrAC14) and PFDCC15).
The differences in physical characteristics between these three PFCLs were not clinicalIy significant. Intraoperatively, the PFCL/saline interface was always visible, alIowing easy fluid-fluid exchange. The refractive index is slightly dissimilar to saline, but this did not alter the refractive error of the eye, and conventional contact lenses could be used throughout surgery. PFDC had a slightly greater ARQ. BRAS. OFTAL. 55, (3), 1992 Exp erimental intraoperative use of perfluorotributylamine, perfluorodecaline, and perfluoropolyether tendency to disperse during injection and the interface was not as easily seen because its refractive index was more similar to that of saline. Removal of alI of these PFCLs had to be done slowly and progressively, especialIy in the PFPE-injected eyes. We did not observe the formation of smalI bubbles (fish egging phenomenon) when using these liquids in conjunction with the infusion fluid utilized in our experiments. The complications resulting from the intraoperative use of PFCLs appear to be minimal. ln some cases, as the liquid was injected, some dispersion did occur, with smalI droplets of PFCL being visible around the edge of the large bubble. Upon contact, however, these smalI droplets tended to coalesce with the large bubble. Furtherrnore, the low viscosity of the PFCLs used in this study permitted use of conventional 20-gauge microsurgical instruments. ln a previous study using PFr A, the outer segments of rabbit retina demonstrated irregular1y shaped defects, referred to as "moth-eaten" by the authors, that reverted to normal after removal of the PFCU14). ln another study, TERAUCHI and co-workers (16) observed toxic effects in eyes in which PFr A was used as a vitreous substitute for one month, but no structural changes were found if PFr A were removed from the eye one hour after injection. The irregularly shaped defects in the outer segment discs noted in our study were similar to those described by CHANG et al (14). Since these changes are reported to be reversible, however, this effect may not preclude the use of PFDC intraoperatively.
We believe our results support the intraoperative use of these PFCLs. Our findings suggest that one hour exposure of the retina to these PFCLs bubbles is not related to any significant toxic effect in the eyes of experimental rabbits. Additional studies are underway to further confirm the nontoxic nature of these PFCLs when used as short term substitutes for vitreous.