INTRODUCTION. The aim of the study is to investigate the ability of an amniotic membrane implant combined with penetrating keratoplasty to reduce early and mid-term complications in complex cases such as penetrating burns, infective ulcers, keratitis, or corneal graft failure.

MATERIALS AND METHODS. Fifty patients: 12 with keratitis, 24 with infective ulcers, 10 with penetrating trauma, and four with a history of corneal graft failure, were divided into two groups. The first group of 25 patients (Group A) underwent penetrating keratoplasty alone, while the second group of 25 patients (Group B) received penetrating keratoplasty associated with an amniotic membrane implant. Amniotic membrane implantation was performed in a ‘patch modality’, and the membrane was sewn to the graft with the epithelium/basement membrane side facing inwards. All patients were evaluated, respectively, 3, 15, 30, 90, and 180 days after surgery. At each visit a slit-lamp examination was performed together with corneal thickness and endothelial cell count assessment. All the data were subjected to statistical analysis with Student’s t-test.

RESULTS. At the slit-lamp examination in Group A 48% of patients showed Descemet folds at one and three months, respectively, while 80% of patients of Group B did not show any Descemet folds. In Group A we registered two cases of early graft failure and two cases of shallow anterior chamber, while none of these complications appeared in Group B. In Group A the average corneal thickness at 15, 30, 90, and 180 days post-operatively was, respectively, 695 ± 43 μm, 658 ± 31 μm, 588 ± 12 μm, and 518 ± 20 μm, while in Group B it was found to be, respectively, 667 ± 12 μm, 632 ± 17 μm, 562 ± 16 μm, and 516 ± 10 μm. Differences in corneal thickness between Group A and Group B were found to be statistically significant (p = 0.05). Endothelial cell count was in Group A 2582 c/mm2 at 15 days, 2500 c/mm2 at one month, 2335 c/mm2 at three months, and 2111 c/mm2 at six months, while the average count for Group B patients was 2607 c/mm2, 2503 c/mm2, 2299 c/mm2, and 2086 c/mm2, respectively. Differences in endothelial cell count between the two groups did not show any statistical significance (p > 0.05).

CONCLUSIONS. Amniotic membrane implantation associated with penetrating keratoplasty reduces early and mid-term complications of corneal grafts in patients with high risk of failure. This positive effect may be due to the anti-inflammatory, neurotrophic, and anti-angiogenetic properties of the amniotic membrane. Longer observation and larger case series can be useful in evaluation of the functional outcome of grafts in these patients.