Therapeutic keratoplasty is recommended for cases where infectious corneal disease continues to progress despite maximal medical therapy, compromising the integrity of the globe and useful vision [8, 11]. This study reports on the outcomes of modified tectonic corneoscleral graft (TCG) for treating corneal infections involving the limbus and partial sclera. Patients in such severe cases are often advised to undergo evisceration due to the lower probability of a successful outcome and minimal expectations of eye salvation, making it difficult to design case-control studies. Therefore, the results of modified TCG in these particularly challenging cases are reported here.

The results of this study indicated that fungal infections were more prevalent than bacterial infections, with Fusarium being the most frequently identified pathogen. It was consistent with reports that fungal keratitis is a leading cause of blindness in Asia [13] and Fusarium was the most common causal agent of fungal keratitis in developing countries [14, 15]. Special attention should be given to Pythium insidiosum, which requires confirmation by DNA sequencing due to its high rates of postoperative recurrence and enucleation [16, 17]. In this study, two patients who experienced recurrence in graft and ultimately required enucleation were infected Pythium insidiosum.

For total corneal infections or sclerocorneal infections, large therapeutic penetrating keratoplasty and sclerokeratoplasty have been described. Jain et al. [18] performed large therapeutic penetrating keratoplasty (with an average graft size of 10.5 mm) on hopeless microbial keratitis cases otherwise advised for evisceration, with a follow-up period of 3 months and a failure rate of 21.4% (6/28). Thatte et al. [9] conducted sclerocorneal transplantation for patients with severe sclerocorneal infections. Despite creating an oblique incision during the procedure, edge-to-edge suturing was used. Postoperatively, 10% of patients developed secondary glaucoma and 21.6% experienced recurrence of infection. Hirst et al. [12] used an edge-to-edge suturing technique in sclerokeratoplasty, resulting in a 50% incidence of secondary glaucoma. Kumar et al. [19] used fibrin glue to adhere the partial-thickness scleral and full-thickness corneal bed to the recipient bed. However, this method cannot completely prevent glue from entering the anterior chamber and requires longer follow-up to assess its safety.

In this cohort, surgical modifications were made to enhance the success rate and minimize postoperative complications. First, the diseased cornea along the corneal limbus and adjacent infected sclera were excised to ensure the complete removal of all pathogens. Second, the fresh donor cornea, with a thinned scleral ring, was overlapped and securely sutured onto the implant bed, and the anterior chamber was reformed using a viscoelastic agent. The procedure not only reduces the complexity of side-by-side suturing at the corneal limbus but also diminishes the risk of secondary glaucoma by ensuring a 360° formation of the anterior chamber with viscoelastic agents [20]. We hypothesize that the postoperative aqueous humor drainage might occur from the posterior chamber into the anterior chamber, then through potential pathway into the sclera, and finally either through the scleral pathway or directly into the subconjunctival space. Remarkable improvements were observed, as the IOP of all patients was within the normal range after surgery. Intraoperative peripheral iridectomy may exacerbate inflammatory responses, increase surgical complexity, and, more importantly, compromise the iris-lens diaphragm, thereby elevating the risk of postoperative infection spread. Consequently, prioritizing infection control, we did not perform a peripheral iridectomy in these cases.

Therapeutic large-diameter keratoplasty may not always achieve complete elimination of the infection. The recurrence rate of infection for end-stage corneal disease ranged from 30.4 to 65.0% [8, 9, 12]. In this study, the procedure has been shown to effectively eradicate the infection, preserve the eyeball, and retain some useful vision. Thirty-six cases (94.7%) successfully eradicated the infection and better restored the anatomic integrity of the eye postoperatively. Poor visual outcomes are still possible due to the inherent risks of surgery and the consequences of the infectious disease itself [8, 10, 12]. Large-diameter penetrating keratoplasty may help reduce postoperative astigmatism and improve visual acuity outcomes [21]. In the present series, the rate of monocular blindness decreased from 100 to 58%, with 15 patients achieving a best corrected visual acuity greater than 0.05. In five cases, the eyeballs were preserved, but the visual acuity was limited to hand movement due to complications such as graft immune rejection, recurrent corneal epithelial defects, and choroidal detachment.

Previous studies on penetrating keratoplasty for treating infectious keratitis have shown graft survival rates of 78.4–95.0% [22,23,24]. However, there is limited data on graft survival rates for large-diameter penetrating keratoplasty. The study here depicted that the survival rate of ocular surface stability declined gradually with time, from 84.2% at one year to 57.7% after three years. Similarly, the survival of central graft clarity declined from 84.2 to 36.0% after three years due to postoperative complications. At the final follow-up (mean, 25.1 ± 8.6 months), 84.2% of the eyes had a stable ocular surface and 52.6% maintained a clear graft, indicating better mid-term results.

Complications may lead to transplantation failure, especially graft rejection [25, 26]. High-risk factors for graft rejection in these cases include (1) ultra-large diameter corneal graft [27, 28], (2) severe infection [29] and (3) fungal infection, in which topical glucocorticoids are avoided in the early post-surgical period [30, 31]. To prevent graft rejection, tacrolimus and glucocorticoids were applied locally in concentration gradients. In this series, the incidence of immune rejection was 36.8%, higher than that typically observed with conventional penetrating keratoplasty (5–18%) [26, 32]. However, further analyses revealed that the affected patients did not adhere to regular follow-ups and did not consistently use their topical medications. Six cases experienced immune rejection of corneal grafts within 3–4 weeks after surgery. Additionally, patients with fungal corneoscleral ulcers did not receive topical glucocorticoids in the early postoperative period, which may have contributed to the early postoperative immune rejection. Thus, tacrolimus, a potent immunosuppressant [33,34,35], has demonstrated anti-rejection properties after TCG [27]. Ensuring good compliance with postoperative care significantly reduces the risk of immune rejection following a successful operation. Another complication to be aware of is choroidal detachment (4/38). Sutures that are too loose in the overlapping technique may lead to postoperative wound leakage and a shallow anterior chamber. The two-step incision method of initial vertical and then oblique, as adopted by Hirst and colleagues [12], can better prevent wound leakage. Therefore, we recommend using at least 16 tightly sutures in TCG to minimize leakage while maintaining a potential space for aqueous humor drainage.

Although steroids remain the mainstay of treatment, the management of corneal graft rejection still relies on empirical approaches [36]. This is especially challenging in patients with severe fungal corneoscleral infections, where the use of topical and systemic corticosteroids after corneal transplantation poses a significant therapeutic dilemma [31]. It has been suggested that administering IV pulse methylprednisolone at a dose of 500 mg may result in transient lymphopenia [37]. In a randomized controlled trial by Hudde et al. [38], evaluating corticosteroid regimens for endothelial corneal allograft rejection, it was found that additional systemic treatment with 500 mg methylprednisolone offered no significant advantage over intensive local corticosteroid therapy alone. For oral corticosteroids in cases of acute graft rejection, higher doses than the standard 60–80 mg daily are recommended [37]. In this study, all patients who developed graft rejection or choroidal detachment 3–5 weeks postoperatively had pre-existing fungal corneoscleral infections. To prevent infection recurrence, we reduced the steroid dosage, opting for intravenous administration of 40 mg methylprednisolone combined with topical corticosteroid eye drops.

The limitations of the study include the relatively small sample size, the lack of parallel controls with other surgical methods, and the absence of survival analysis of graft transparency and comparison of fungal and bacterial outcomes. In terms of surgery, it is a technically demanding procedure, and attention should be paid to avoid complications such as wound leakage, shallow anterior chamber, and choroidal detachment during the overlapping suturing process.