A TALE OF TWO BUBBLES

Though some remain skeptical, there is increasing evidence to support the existence of the “Dua Layer”, a tough acellular layer of stroma-like tissue immediately adjacent to the Descemet’s membrane, said Harminder Dua MD, PhD, in his Rayner Medal Lecture delivered at the UKISCRS 38th Annual Congress in London, UK.

“I have chosen a topic that is still a little controversial, but I'm still sticking to the evidence my team and I have generated, and with some trepidation to the name,” said Dr Dua, Nottingham University, Nottingham, UK.

He noted that the origin of the hypothesis of the presence of this layer of tissue stemmed from several observations. One observation was that in deep anterior lamellar keratoplasty procedures the apparently bared Descemet’s membrane is generally much more sturdy and resilient than Descemet’s membrane when it is removed from the posterior stroma.

When a suture is passed through a button used in penetrating keratoplasty a distinct sharp edge is seen as the needle emerges from the deep stroma. This has been attributed to the Descemet’s membrane. When suturing a button prepared for deep anterior lamellar keratoplasty (DALK) by removing the Descemet’s membrane, such an edge is still visible, indicating that there is another layer of tissue that produces this edge in addition to the Descemet’s membrane.

Moreover, when Descemet’s membrane is peeled from the back surface of a donor button it comes off easily but when deep stroma is separated from the front surface of the Descemet’s membrane, which essentially is the same thing, more resistance and strands of collagen are encountered. This suggested that the two planes were different, the latter being between deep stroma and the new layer.

Furthermore, many corneal surgeons who perform lamellar keratoplasty find that there are two different types of bubbles that occur when using the big bubble technique. There are bubbles that start at the centre but do not extend all the way to the periphery, which Dr Dua terms as Type 1 bubbles, and there are bubbles that start on the periphery and spread from there across the centre, termed Type 2 bubbles.

The Type 1 bubbles generally have a diameter no greater than 8.5mm and the Type 2 bubbles have a diameter up to 12.0mm. On occasions the two types of bubbles occur at the same time with the Type 1 being usually complete and the Type 2 being partial. Such bubbles are termed “mixed bubbles”. Until Dr Dua reported his findings it was believed that mixed bubbles were due to a split in the banded and non-banded zones of the Descemet’s membrane. That misconception has now been dispelled.

In a series of experiments using donor eyes, Dr Dua and his associates were able to demonstrate that the Type 1 bubbles occur between the posterior stroma and what appears to be a previously undiscovered layer of tissue. Among their findings was that although Type 2 bubbles deflate when Descemet’s membrane is removed, the Type 1 bubbles do not (video).

In addition, they showed that it is still possible to create a Type 1 bubble in corneas where the Descemet’s membrane has been removed, whereas it is not possible to create a Type 2 bubble under those conditions.

Histological examination of this layer indicates that it is around 10 microns in thickness and is composed of acellular tissue consisting of predominantly Type 1 collagen bundles but with higher concentrations of collagen 6 than is present in other parts of the cornea.

They also showed that the periphery (approximately half a millimetre) of this layer along its circumference is populated by trabecular cells which lie between lamellae of the layer and lay down basement membrane. At this point the lamellae start to split and continue as the collagen core of the trabecular meshwork.

Dr Dua noted that to take advantage of the sturdy and resilient nature of the newly discovered layer, he and his associate Dr Amar Agarwal MD have developed a technique for deliberately including the pre-Descemet’s or Dua’s layer with the donor endothelium in a modification of Descemet’s membrane endothelial keratoplasty (DMEK) that has been christened pre-Descemet’s endothelial keratoplasty or PDEK.

Results in the first few patients have been published and appear promising, with good graft attachment and good postoperative visual recovery in all cases. Similarly with another colleague, Dr Ahmed Atef Zaki, he has published results to show that the toughness of the layer allows a complete cataract extraction with lens implant to be carried out during the DALK operation, a procedure termed “triple DALK”.

He added that the presence of Dua’s layer can help explain the otherwise unaccountable superior strength of DALK grafts compared to penetrating grafts; why some descemetoceles (covered with this layer) resist perforation and have some evidence to suggest that this layer also splits together with the Descemet’s membrane during acute hydrops in keratoconus.

Harminder Singh Dua:

harminder.dua@nottingham.ac.uk