In the past, various methods for the puncture of retinal vessels have been proposed. Recent approaches include the cannulation of retinal veins with micropipettes or the catheterization of retinal vessels with a flexible microcatheter for the injection of rt-PA.
21,22,26 –32 However, although these methods have demonstrated the feasibility of retinal vessel puncture, there is still a lack of clinically applicable methods for the injection of drugs into the lumen of retinal vessels in a reliable manner. The major problem with available techniques is the limited visualization of the retinal vessel because of insufficient magnification with conventional surgical microscopes, making it difficult to puncture the vessel or to effectively control the targeted injection of drugs. In a recent study, Feltgen et al.
23 reported on a high risk of surgical complications after vitrectomy with central venous injection of rt-PA in patients with CRVO. In this study, successful punctures were made in 10 of 13 patients. Of these, neovascular glaucoma developed in six patients and enucleation was needed in two eyes.
In the present study, all maneuvers were performed through a pars plana approach under endoscopic guidance. We were able to puncture branch retinal veins at various distances from the optic disc, where the vein had a considerably smaller diameter and the underlying retina was less tightly adhered. Using bimanual manipulation of the endoscope, we injected rt-PA, normal saline solution, or hypertonic solution. Moreover, we were able to maintain the tip of the catheter in its intravascular position until removal after several minutes.
The endoscopic approach provides excellent magnification, thereby managing the challenges of visualization of retinal vessels and holding the microcatheter at an appropriate angle for venipuncture. In our study, a sleeve carrying the catheter was mounted on the fiberoptic tip of the endoscope. This has an advantage over settings with multiple instruments because the bimanual manipulation of a single device allows for precise maneuvers by the surgeon. In addition, the surgeon can obtain pseudo-depth perception from the dynamic movement of the endoscope relative to the retinal vessel that, when fixed, is absent. In previous attempts to facilitate the task of delivering drugs to retinal microvessels, various complex micromanipulation devices or expensive robots have been used.
21,22,27 Such micromanipulators require maximum stability, which is usually accomplished by a complex fixation of the patient's head or eye, thereby limiting accessibility and operability in a surgical setting.
Another major concern is the size and form of current devices for the puncture of retinal microvessels. Thus far, cannulas and microcatheters have been designed to be insertable into main branches close to the optic nerve head. Usually, these vessels measure at least 100 μm in diameter. However, the puncture of retinal vessels close to the optic disc carries the risk of damaging the optic nerve. Moreover, the management of BRVO would require the proximal injection of a fibrinolytic agent to the occluded portion of the vein (i.e., the puncture of a vessel as small as 50 μm or less in diameter). Several experimental studies have demonstrated that probing and catheterization of the vascular lumen of retinal vessels may be achieved with sutures, glass tubes, or polyimide tubes.
28 –32 Since the maximally achievable reduction in size of a cannula depends on stability, previous techniques using materials such as borosilicate, metal, or polyamide have been limited to an outer diameter ranging from 100 μm to a minimum of 50 μm.
29 In a recent experimental study, Tsilimbaris et al.
32 demonstrated that beveling of the catheter tip and diameter of the catheter are the most important factors affecting the ease of vascular catheterization. They achieved the best results with polyimide tubes with internal diameters of approximately 50 μm. Based on their results, they concluded that this size seems to represent the maximum catheter diameter that can be used for microvascular catheterization.
32 Other researchers have reported the successful use of a microcatheter instrument and a surgical technique for retinal vein cannulation with prolonged intravascular infusion of a thrombolytic agent in experimentally induced BRVO in dog eyes.
26 However, in this study, retinal microvessel puncture was performed at relatively short distances away from the optic disc (1–4 disc diameters), indicating that this approach does not overcome the problem of visualization. The cannulation of smaller branches of the retinal vasculature with conventional surgical microscopes remains a difficult task, regardless of the catheter system used for the targeted injection of drugs. In our study, we used a flexible catheter fabricated of a single-piece quartz glass tubing coated with polyamide and terminating in a sharp distal tip with an outer diameter ranging from 10 to 20 μm. By moving the endoscope-catheter system within the vitreous cavity, it was possible to puncture branch retinal veins at distances up to 15 disc diameters from the optic disc, where the vein had a considerably smaller diameter. Occasionally, we noted small traces of blood at the puncture site on removal of the catheter tip. However, we observed no visible damage to the vessel wall or mechanical damage to retinal tissue.
The finding that arteries tended to slide under the pressure of the sharp tip of the microcatheter reflects the fact that the puncture of retinal arteries is technically more challenging. To date, all clinical studies using endovascular surgery included patients with retinal vein occlusion. However, the difficulties we encountered during the puncture of retinal arteries were mainly related to the experimental setup with emptied retinal arteries in porcine cadaver eyes. Our findings indicate that an endovascular approach to the management of central retinal artery occlusion will require the puncture of main branches, where the retina is attached tightly to the margin of the optic disc.
To date, there is much evidence that thrombolysis is associated with a more favorable visual outcome in retinal vein or artery occlusion.
12 –17 However, it is a well-established observation that the systemic administration of thrombolytic agents carries the risk of life-threatening hemorrhagic complications.
18,19 Because this is a dose dependent problem, the targeted retinal delivery of fibrinolytic agents at low doses should be the ideal approach. In a previous experimental study
33 on the percutaneous cannulation of supraorbital arteries, we were able to demonstrate that intravascular volumes of the ophthalmic and the supratrochlear artery were 60 μL and 10 μL, respectively, indicating that the retinal intravascular administration of drugs would require relatively small injection volumes. This is consistent with our current observation that puncturing and injection of small volumes of fluid always resulted in a visible rapid blood flow toward the optic disc and a temporary dilation of branch retinal veins in porcine cadaveric eyes. Another critical factor that has to be taken into account when using endovascular surgery to restore retinal blood flow is time to treatment. Given that the process of adherence and organization of a venous thrombus does not begin until 5 to 10 days after thrombus formation, it seems plausible that thrombolytic therapy should be initiated in the early acute stage.
16,34
The findings of the present study support the concept of retinal endovascular thrombolysis as a potential therapeutic approach to RVO diseases. The endoscopically guided delivery of drugs to retinal vessels may contribute to improve safety and reproducibility of this treatment strategy. Here, we were able to demonstrate adequate retinal vessel visualization and puncture at various distances away from the optic disc using a novel catheter system mounted on the fiberoptic tip of a high-resolution gradient index microendoscope. Experiments to determine the efficacy of endovascular thrombolysis in retinal vein or artery occlusion were beyond the scope of the present study. However, a greater understanding of the technical feasibility and the identification of appropriate surgical techniques are important prerequisites for future clinical studies to establish endoscopic retinal vessel cannulation as an alternative to systemic thrombolysis.