The outcomes of corneal sight rehabilitating surgery in Stevens-Johnson syndrome: case series.

PURPOSE: To summarize the outcomes of corneal sight rehabilitating surgery in Stevens-Johnson syndrome (SJS).

METHODS

This is a retrospective analysis of a consecutive case series. Twenty-four eyes of 18 SJS patients were included in this study. The ocular parameters, surgical procedures, postoperative complications, and additional treatments of the cases were reviewed.

RESULTS

A total of 29 corneal sight rehabilitating surgeries, which consists of 9 keratoplasties, 8 Keratolimbal allograft (KLAL) and 12 combined surgeries (keratoplasty and KLAL simultaneously) were performed on the 24 eyes. All patients were treated with glucocorticoid eyedrops and tacrolimus eyedrops for anti-rejection treatment without combining systemic immunosuppression, except two patients who were prescribed prednisone tablets for the management of systemic conditions. The mean follow-up period was 50.6 ± 28.1 months. The optimal visual acuity (VA) (0.74 ± 0.60 logarithm of the minimum angle of resolution [logMAR]) and endpoint VA (1.06 ± 0.82 logMAR) were both significantly better than the preoperative VA (1.96 ± 0.43 logMAR) (95% CI, p = 0.000). 57.1% patients (8/14) were no longer in the low vision spectrum, and 88.9% patients (8/9) were no longer blind. The mean epithelialization time was 7.1 ± 7.6 weeks. The success rate was 86.7%. Additional treatments for improving epithelialization included administration of serum eyedrops (n = 10), contact lens (n = 15), amniotic membrane transplantation (n = 6), and tarsorrhaphy (n = 8). Complications included delayed epithelialization (n = 4, over 12 weeks), glaucoma (n = 11), and severe allograft opacity (n = 4). Only one graft rejection was observed.

CONCLUSIONS

Keratoplasty and KLAL can remarkably enhance VA and improve low vision or even eliminate blindness for ocular complications of SJS. The outcome of the surgeries was correlated with the preoperative ocular situation and choice of operative methods.