Intraoperative optical coherence tomography guided corneal sweeping for removal of remnant Interface fluid during ultra-thin Descemet stripping automated endothelial keratoplasty

Partial thickness corneal transplants currently account for a majority of keratoplasty procedures performed [1, 2]. Descemet stripping automated endothelial keratoplasty (DSAEK) and Descemet membrane endothelial keratoplasty (DMEK) are considered the procedures of choice for corneal endothelial decompensation [3]. Endothelial keratoplasty leads to faster visual recovery, fewer complications, superior visual outcomes and lower graft rejection rates compared to penetrating keratoplasty (PKP) [4].

Despite its relatively high safety profile, DSAEK is not without potential intraoperative and postoperative complications with early postoperative graft detachment being the most common one [5]. Graft detachment vary greatly between studies ranging between 0.9 to 36.4% in studies that include complex cases [6]. Several factors have been reported to be associated with graft detachment in DSAEK including previous failed PKP [6], prior glaucoma surgery [6, 7], compromised iris-lens diaphragm [8], combined phacoemulsification cataract surgery [9] and residual graft-host interface fluid [10].

Intraoperative optical coherence tomography (iOCT) has been shown to be feasible for numerous anterior and posterior segment ophthalmic procedures [11‐14]. Specifically, iOCT has been reported to be beneficial during DSAEK procedures and identification of remnant interface fluid [15, 16]. In 2017 Hallahan et al. reported that larger interface fluid volume, area, and thickness at the end of iOCT guided DSAEK were associated with early graft detachment [10].

The aim of this study was to describe and assess outcomes of intraoperative optical coherence tomography (iOCT) guided meticulous peripheral corneal sweeping for removal of interface fluid during ultra-thin DSAEK (UT-DSAEK).

Remnant interface fluid was identified in nearly 90% of the UT-DSAEK cases after instillation of the air bubble leading to subsequent meticulous peripheral corneal sweeping. The meticulous corneal sweeping led to resolution of interface fluid in an overwhelming majority of cases and reduced intraoperative CGT. In addition, there was an association between immediate postoperative rebubbling and remnant interface fluid despite corneal sweeping. This is, to the best of our knowledge, the first study to characterize iOCT guided meticulous peripheral corneal sweeping in UT-DSAEK.

The ability to identify subclinical interface fluid with the aid of iOCT during DSAEK was first described in 2010 by Ide et al. and Knecht et al. [17, 18] A reduction in interface fluid after air bubbling and a further subsequent reduction following venting incisions was described [18]. These were followed by several small case series describing reduction of remnant interface fluid as seen on iOCT following venting incisions [19, 20]. Price et al. described corneal sweeping followed by venting incisions in order to promote graft adherence [21]. Thereafter, Terry et al. reported repeatedly compressing the corneal surface from the center to the periphery to “milk out” any interface fluid during DSAEK in order to avoid venting incisions [22]. The efficiency of avoiding the need for venting incisions during iOCT guided DSAEK corneal sweeping was later demonstrated by the PIONEER study group [10, 15]. Indeed, in the current study, meticulous peripheral corneal sweeping was necessary in nearly 90% of UT-DSAEK cases and completely resolved interface fluid in 84% (n = 21/25) with venting incisions deemed necessary in a single case. Similar to DSAEK, it seems as though meticulous iOCT guided corneal peripheral sweeping may obviate the need for venting incisions in UT-DSAEK as well. We found it to be critical to sweep from the center all the way out until beyond the limbus to “milk” as much interface fluid as possible. This was done from the center into all 4 quadrants. Despite this maneuver, 16% (n = 4) of the cases had remnant interface fluid. A potential explanation is that when the anterior chamber is filled with air and the eye is rather firm, it may make meticulous sweeping more difficult. For cases with persistent interface fluid, surgeons may consider slightly reducing the pressure of the eye and attempting the sweep again, a technique that was not assessed in the current study.

In the current study the overall rate of rebubbling (17.9%) was on the higher side of what is reported in the literature. Studies reporting DSAEK results where complex cases were included have reported graft detachment rates ranging between 0.9 to 36.4% [6]. We speculate that this is because our threshold for rebubbling is low as many of the patients treated arrive from far away to receive treatment at our tertiary center. A majority (60%) of the grafts that required rebubbling had small amounts of remnant interface fluid despite peripheral corneal sweeping and/or required a venting incision. A small minority (4.4%) of the eyes that did not require rebubbling had remnant interface fluid. These findings are supported by the PIONEER study where larger residual interface fluid volume, area, and thickness at the end of surgery detected with iOCT were associated with early graft non-adherence in DSAEK [10]. They postulated that overall fluid burden may have played a role in non-adherence of grafts. It may also be that unknown factors which led to incomplete resolution of interface fluid following sweeping may have also contributed to subsequent graft detachment. In any event, similar to DSAEK, UT-DSAEK grafts with remnant interface fluid on iOCT are at a higher risk for immediate postoperative graft detachment requiring rebubbling. It is unclear whether or not venting incisions are indeed helpful in these situations as this was performed in only one eye in the current study. Perhaps non-expansile gas such as SF₆ or C₃F₈ may be considered with an emphasis on keeping the eye more pressurized than usual in these cases.

Several studies have evaluated the utility of iOCT in DMEK surgery. These reports demonstrated that iOCT aided in identifying remnants of host Descemet membrane, identifying and facilitating correct graft orientation [23‐25]. Recently, Patel et al. reported in a prospective study (n = 100) that for novice DMEK surgeons, complication rates and unscrolling times compared favorably with alternative tissue orientation methods while avoiding the need for external markings [25].

Most studies have focused on interface fluid at different stages with a paucity of data available regarding iOCT assessment of DSAEK graft thickness at different stages of surgery. Steverink et al. (n = 8) reported an increase in DSAEK graft thickness after insertion into the anterior chamber, however thickness following corneal sweeping was not reported [26]. Indeed, in the current study, CGT was thicker when measured after air bubbling than measured in the cornea bank 2–3 days before surgery. Interestingly, there was a small (7.7%) yet significant reduction following peripheral sweeping. Further prospective comparative studies with larger sample sizes are needed to assess whether the reduction in thickness is a result of the meticulous sweeping or not.

This study has several limitations including its retrospective nature, small sample size and unbalanced groups when comparing the no rebubbling group to the rebubbling group. However, this sample size was sufficient to demonstrate the efficacy of the technique in resolving remnant interface fluid, reduction in intraoperative graft size and association with rebubbling when fluid persists despite sweeping.

In summary, iOCT identified subclinical remnant fluid in nearly 90% of UT-DSAEK cases. An iOCT guided peripheral corneal sweeping was associated with a reduction in CGT as well as complete resolution of interface fluid in a majority of cases. This pilot study demonstrates that eyes with persistent remnant fluid despite peripheral corneal sweeping during UT-DSAEK are more likely to require subsequent rebubbling.