This study suggests that optical keratoplasty is the best choice for AK patients whenever possible, which has better final BCVA and safety. Early keratoplasty is critical to AK cases with poor response to medical treatment and timely therapeutic keratoplasty within 5 months would be recommended. TDALK should be considered prior when taking therapeutic keratoplasty. Penetrating keratoplasty could be performed when corneal perforation and deep stroma involvement.
The meta-analysis showed that optical keratoplasty has better final BCVA, higher graft survival rate and less recurrence rate than therapeutic keratoplasty for AK patients. Besides, optical keratoplasty and therapeutic keratoplasty had similar incidence rate of postoperative complications. Thus, optical keratoplasty was the preferred type of keratoplasty for AK patients. These conclusions were in tune with numerous publications that the infection should be treated medically and the keratoplasty should be performed on uninflamed or controlled-inflamed AK if possible.10,16,22 The optical keratoplasty was defined as least 1–3 months after discontinue antiamoebic therapy.10,23,24 Of note, limbal stem cell deficiency (LSCD) was easily induced in AK patient who experienced long-term inflammation and medical therapy, which could result in graft failure. Thus, it is very important to diagnose and treat LSCD before optical keratoplasty. Moreover, optical lamellar keratoplasty was preferred to lower the risk of endothelial rejection due to corneal neovascularization and previous corneal inflammation.10
When AK did not response well to medication, an early therapeutic keratoplasty was proposed. The included studies considered the following symptoms or signs were indicators for therapeutic keratoplasty: unrelieved pain, decreased BCVA, expansion of stromal infiltrates, impending or actual corneal perforation and increased anterior chamber inflammation after at least 1 week intensive topical anti-amoeba therapy.10,11,21 An interesting thing to note was the relevance and inflection point at 5 months in Fig. 3, which meant the longer the disease duration, the worse the final BCVA, and the patients who performed therapeutic keratoplasty after 5 months of AK had worse final BCVA. These results further supported the advice that the better opportunity for therapeutic keratoplasty is the first 5 months of AK duration.25
The major factors that affecting the outcomes between early and late therapeutic keratoplasty included the graft size and inflammatory response. Previous studies concluded the graft size was an powerful indicator for the ultimate anatomical and functional outcomes in infectious keratitis.10,26,27 Meanwhile, late therapeutic keratoplasty usually meant long-term toxic medical therapy and persistent ulcer, which accompanied by severe inflammation and associated with postoperative complications, such as glaucoma, cataract, anterior synechiae and graft rejection.16,26,28 However, how to use corticosteroid to control the inflammation before and after therapeutic keratoplasty in AK patients remain questionable. Some studies started steroids after therapeutic keratoplasty immediately, and some researchers used steroids at least 2 weeks after therapeutic keratoplasty.10,16,26 A study reported FK506 might improve the graft success rate after therapeutic keratoplasty in AK patients.29 Meanwhile, anti-amoebic agents must be continued postoperative for at least 2 months.30 Clinicians must balance the relationship between anti-inflammation and anti-amoeba therapy, which dictate the success of keratoplasty.
TDALK has been proposed as an effective therapy for drug-resistant localized infectious keratitis. The currently study showed that the AK patients who performed TDALK had better final BCVA and who performed TPK were more likely to need multiple keratoplasties. Furthermore, a recent study concluded TDALK provided less high-order aberrations than TPK for AK patients.19 Sarnicola et al. performed TDALK within 1–2 months after AK onset.21 Early surgery might also play an important role in improving the prognosis. The potential risk of TDALK was the infection may not be eliminated completely. In the included studies, only 8.8% (3/34) patients had AK recurrence after TDALK, 20,21 which was like TPK. Besides, TDALK could also lower the risk of endothelial rejection. Hence, TDALK, instead of penetrating keratoplasty, should be considered first in medically unresponsive AK.
This study also had some limitations. First, there was no randomized clinical trial comparing different keratoplasty procedure in AK, which limited the evidence level of the study. Second, because of the limit reported information, the preoperative BCVA only compared between early and late therapeutic keratoplasty, but not between optical keratoplasty and therapeutic keratoplasty, TPK and TDALK. Third, considering only 1 study reported the complications of TDALK and none of the patients had complications, it is difficult to conduct the analysis of complications between TDALK and TPK. Forth, the better operation opportunity for TDALK could not be calculated based on current studies. Further studies with detail pre-TDALK and post-TDALK information should be needed.
In summary, this systemic review demonstrated the advantage of optical keratoplasty for AK patients. For those who poor response to anti-amoeba therapy, early therapeutic keratoplasty, especially deep anterior lamellar keratoplasty should be considered. Further adequately powered studies with preoperative and postoperative details, including preoperative disease duration, anti-amoeba therapy duration, clinical signs etc., would be helpful in making the best keratoplasty decision comprehensively.