Pupil block after DMEK

As many as 10 per cent of patients undergoing descemet membrane endothelial keratoplasty (DMEK) are at risk of pupillary block after surgery if the problem is not detected and managed, Matthew T Feng MD told the 2014 American Society of Cataract and Refractive Surgery symposium in Boston.

The 10 per cent rate suggested by a study Dr Feng and colleagues conducted at Price Vision Group, Indianapolis, US, is within the range reported for descemet stripping automated endothelial keratoplasty (DSAEK), he said. Dr Feng’s study found that patients undergoing DMEK combined with cataract surgery and intraocular lens (IOL) implantation are at higher risk.

Dr Feng recommended examining all DMEK patients one hour after surgery for signs of pupil block. These include the anterior chamber remaining full of air, high intraocular pressure (IOP) and obstructed peripheral iridotomy (PI).

Patients with signs of potential pupil block should be observed further. Those with full air fills that do not spontaneously resolve or develop symptoms of pupil block should be dilated and have some air removed, Dr Feng said. “In our practice, the pupil block rate is zero because we actively screen for and prevent it,” he added.

Watch the bubble

Dr Feng noted that air is injected into the anterior chamber at the end of DMEK to hold the graft in place until donor endothelial cell function suctions it up against the recipient stroma. To ensure the graft is held firmly even when the patient sits up, the target is an air fill that leaves a small 360-degree fluid meniscus when supine.

This corresponds to an 80-90 per cent air fill in most eyes, which leaves enough fluid over a patent PI to allow aqueous flow and circumvent pupillary block. (see Figure)

With DSEK and DSAEK, large bubbles have been shown to carry a risk of pupillary block (Lee JS et al, Cornea. 2009 Jul;28(6):652-6), with reported rates ranging from 0.3 (Terry MA et al, Ophthalmology. 2008 Jul;115(7):1179-86) to as high as 30 per cent for phakic patients (Tsui JYM et al, Cornea. 2011;30(3):291-295), Dr Feng noted.

“How is it then that in the English literature the published rate (for DMEK, which usually uses a larger bubble) is zero per cent?” he asked.

To better characterise pupillary block risk in DMEK, Dr Feng and colleagues conducted a retrospective interventional cohort study of 351 consecutive eyes over a one-year period that underwent DMEK procedures, with or without cataract surgery, for Fuchs', corneal edema or failed grafts. Only eyes with previous glaucoma surgery were excluded. Outcomes included bubble size after surgery, IOP and management required. Sub analyses were conducted by whether a single or triple DMEK was performed and the status of the PI, posterior capsule and vitreous.

One hour after surgery, 74 eyes (or 21 per cent) had a full air fill rather than the intended 80-90 per cent. Of these, 57 per cent (or 12 per cent overall) spontaneously resolved under observation. However, 43 per cent (or nine per cent overall) required intervention in the form of air removal, and these are the eyes Dr Feng believes are at risk of pupillary block. No eyes experienced pupillary block after discharge.

Statistically significant factors associated with full air fills at one hour were PI obstruction, high same-day postoperative IOP and DMEK triple procedure.

Dr Feng believes newly pseudophakic eyes may be at higher risk because a flexible IOL-iris diaphragm allows for more air to be pumped into the anterior chamber. Eyes requiring intervention had a higher incidence of IOP over 30mm Hg or PI obstruction than those that resolved spontaneously.

Have a plan

Dr Feng recommended adopting a management algorithm for patients with full air fills one hour postoperatively. For those with an obstructed PI, IOP above 30mm Hg, or rising IOP with symptoms of pupillary block, he advised shrinking the air bubble to 50-60 per cent and dilating. For those with high IOP and a potentially patent PI (eg obvious PI in an area of iris-corneal touch from air behind iris), he suggested shrinking air to 80 per cent. Air fill patients with a patent PI and lower IOP should be observed hourly but generally resolve spontaneously, he said.

Air removal

Eyes that underwent air removal achieved similar three-month visual results (median 20/25) and endothelial cell counts (median 2100-2200 cells/mm2) as those that resolved with observation, Dr Feng said. At three per cent, the air removal eyes actually had lower re-bubble rates than the 17 per cent seen in observation eyes, but the difference just missed significance (P = 0.06), he added.

Matthew Feng: mattfeng@pricevisiongroup.net