Purchase this article with an account.
Blake V. Acohido, Jeffrey D. Holiman, Joshua D. Galloway, Christopher G. Stoeger, Winston Chamberlain; Viscoelastic Bubble Dissection: A New Method for Tissue Preparation in Descemet Membrane Automated Endothelial Keratoplasty. Invest. Ophthalmol. Vis. Sci. 2012;53(14):30.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To study the use of viscoelastic in preparation of eye bank corneas for Descemet membrane (DM) automated endothelial keratoplasty (DMAEK). To compare endothelial cell death (ECD) between viscoelastic bubble dissection DMAEK (vDMAEK) tissues to air bubble dissection DMAEK (aDMAEK) to manual DM peeling (DMEK).
Corneal scleral rims, without specific pathology, from adult donors aged 18 to 75 years were obtained. 13 tissues used in DMAEK preparation were precut with microkeratome to generate a deep lamellar incision. DM was bared in the central 6.5 mm zone of 6 vDMAEK tissues with an injected mixture of 1% sodium hyaluronate and in 7 aDMAEK tissues with injection of air. Controlled central bubbles less than 6.5 mm in size were considered successful in barring central DM. 8 additional tissues were prepared using conventional DMEK technique with manual DM peeling. Centrally trephined tissue from the three techniques were examined for endothelial damage with calcein AM viability dye. Images were captured with an inverted light microscope, and digitally stitched together. ECD was calculated from the number of non-stained pixels over total number of pixels.
Controlled central bubble formation was successful in 100% (6/6) of tissues prepared by vDMAEK, compared to 73% (8/11) for aDMAEK and 100% (8/8) for DMEK. OCT imaging of DMAEK tissues dissected with viscoelastic and air showed centrally barred DM with residual, noncompact, stromal remnants measuring between 30 and 100 microns in thickness. Viscoelastic was easily removed from tissue during preparation. By inspection, areas of grafts with negative staining fell into 2 patterns 1) small zones of discrete cell loss and 2) larger zones of possible DM stretching. Tissues prepared by standard DMEK resulted in a mean ECD of 22% (95% Confidence Interval (CI): 16-29%). By comparison, vDMAEK grafts had significantly more mean cell loss of 40% (CI: 31-48%, p = 0.03). Grafts prepared by aDMAEK yielded a mean ECD of 28% (CI: 24-31%). This value was not statistically different from DMEK or vDMAEK, with p-values of 0.28 and 0.14, respectively.
Viscoelastic more predictably dissects central DM than air and can easily be washed away from grafts after preparation. Endothelial attenuation by calcein AM staining is significantly higher in vDMAEK tissue than DMEK but not aDMAEK. The reticular pattern of cell loss in the DMAEK grafts may reflect DM stretching with breaks in intercellular adhesion vs. absolute cell loss.
This PDF is available to Subscribers Only