The Histology of Graft Adhesion in Descemet Stripping with Automated Endothelial Keratoplasty

The Histology of Graft Adhesion in Descemet Stripping with Automated Endothelial

Keratoplasty

Matthew Caldwell, Alan Proia, Francis DeCroos, and Natalie Afshari

Authors have no financial interest.

Supported by an award from Research to Prevent Blindness

Duke Eye Center

2

PURPOSE

• BACKGROUND–DSEK is the surgery of choice for endothelial

dysfunction.–The most common complication is graft

detachment.–The mechanism of graft adhesion is poorly

understood.

• OBJECTIVE–To describe the histologic findings of graft

adhesion in DSEK 2

METHODS

• IRB approval

• Search of DSEK surgeries followed by PK

• Clinical history reviewed

• Slides cut and stained with –H&E–PAS–Colloidal Iron

RESULTSEyes with PK following DSEK

Age Sex/Race History Duration

1 88 F/W Fuchs,PCIOL, DSEK 458

2 79 M/W PCIOL,PK, DSEK 103

3 81 M/W PCIOL, tube x2, PK x2, DSEK 232

4 90 F/W ACIOL, DSEK 573

5 87 F/W Tube x3, hypotony, ACIOL, DSEK 138

6 56 M/B Tube x2, CPC, PCIOL, DSEK 373

7 23 M/B Peters, glaucoma surg x3, PK, DSEK x2 75

8 53 F/W Fuchs, PK, PCIOL, DSEK 373

9 59 F/I PCIOL, PK, tube, synechiolysis, DSEK 490

10 64 F/W Fuchs, PK, PCIOL, DSEK 511

Pseudophakic=10 (ACIOL=2)

Glaucoma w/ tube = 5

Failed graft prior to DSEK = 6}All high risk patients with multiple ocular comorbidities

5

Interface Findings

• The interface was barely perceptible in most cases. Keratocyte density was unaltered at the interface.

• In 3 cases, increased cellularity was noted along the unopposed stromal edge in areas of detachment.

6

Interface Material

• Pigment granules and debris were seen in the interface to some extent in nearly all cases. The presence of glycosaminoglycans was not demonstrated with colloidal iron stain.

6

Interface Descemet Membrane

• Descemet membrane was present in the peripheral interface in 8 of 10 cases.

Interface Descemet Membrane

Completely

Separated

(3/10)

+DM in 2/3

• Adherent only to graft in both

Completely

Attached

(4/10)

+DM in 4/4

• More adherent to graft at split edges

Partially

Separated

(3/10)

+DM in 2/3

• More adherent to graft in 1, free floating in the other

Of the 10 specimens, 4 were completely attached grossly, 3 were partially separated, and 3 were completely separated. Interface Descemet membrane (DM) was noted in all three groups. In all groups the interface DM was preferentially adherent to the graft rather than the posterior stroma.

Peripheral Splitting Artifact

Peripheral splitting was present at at least one edge of all attached cases. This is presumed artifact due to shearing forces from specimen sectioning. This was present independent of the presence of Descemet membrane.

Peripheral splitting artifact. Note preferential adhesion of Descemet membrane to the graft.

Other Findings

• Fibrocollagenous pannus

• Anterior stromal scarring

• Paucity of endothelial cells

• Retrocorneal membrane

Other anterior and posterior findings unrelated to the interface were present in many cases.

Retrocorneal membrane

CONCLUSIONS

• DSEK grafts form adhesion without scarring or keratocyte proliferation.

• Pigment granules and Descemet membrane are common in the interface.

• DSEK grafts form a firm attachment to Descemet membrane (relative to the adhesion of DM to the native posterior stroma)

• Stripping of Descemet’s membrane may be unnecessary for graft adhesion