Globe stability during simulated vitrectomy with valved and non-valved trocar cannulas

To compare the effects of valved and non-valved cannulas on intraocular pressure (IOP), fluid leakage, and vitreous incarceration during simulated vitrectomy. Three-port pars plana incisions were generated in six rubber eyes using 23-, 25-, and 27-gauge valved and non-valved trocar cannulas. The models were filled with air and IOP was measured. Similar procedures were followed for 36 acrylic eyes filled with saline solution. Vitreous incarceration was analyzed in eleven rabbit and twelve porcine cadaver eyes. In the air-filled model, IOP loss was 89%-94% when two non-valved cannulas were unoccupied versus 1%-5% when two valved cannulas were unoccupied. In the fluid-filled model, with non-valved cannulas, IOP dropped while fluid leaked from the open ports. With two open ports, the IOP dropped to 20%-30% of set infusion pressure, regardless of infusion pressure and IOP compensation. The IOP was maintained in valved cannulas when one or two ports were left open, regardless of IOP compensation settings. There was no or minimal fluid leakage through open ports at any infusion pressure. Direct microscopic analysis of rabbit eyes showed that vitreous incarceration was significantly greater with 23-gauge non-valved than valved cannulas (P