Therapeutic penetrating keratoplasty for recalcitrant microbial keratitis: An update

Abstract

Therapeutic penetrating keratoplasty (TPK) is a surgical procedure for salvaging medically refractory infectious keratitis. By replacing diseased, infected host corneal tissue with a full-thickness donor cornea, TPK aims to eliminate infection, preserves globe integrity and maintain vision. However, recurrence of infection in the graft remains a significant concern, particularly in cases caused by resistant organisms such as Acanthamoeba, Pythium, microsporidia, and certain bacterial and fungal pathogens. Although often curative, TPK carries a higher risk of endothelial rejection, reduced graft survival, and increased complications due to its open-sky approach. Lamellar keratoplasty, particularly deep anterior lamellar keratoplasty, has emerged as a promising surgical alternative in cases where the infection has spared the deeper corneal layers. Compared with full-thickness therapeutic keratoplasty, it offers advantages including improved graft survival, lower rejection rates and fewer postoperative complications. Nevertheless, recurrence may still occur in deep or extensive infections, and advanced cases with descemetocele, perforation, or limbal or scleral involvement often still require TPK. This article on TPK summarizes peer-reviewed literature from the past decade, incorporating recent advances in corneal surgery while placing the modern-day application of classically described techniques into perspective to strengthen surgical decision-making. Emphasis is placed on resistant pathogens, surgical indications, preoperative evaluation, anesthesia considerations, operative techniques, prognostic factors, and postoperative management strategies. Through this review, ophthalmic surgeons are expected to gain practical guidance for optimizing outcomes in complex corneal infections while addressing the challenges posed by resistant and atypical pathogens across diverse clinical scenarios.

Key Words: Keratitis; Penetrating keratoplasty; Eye infections; Corneal ulcer; Surgical wound infection

Core Tip: Therapeutic penetrating keratoplasty is a surgical procedure wherein diseased infected host corneal tissue is replaced by a full-thickness donor cornea for eliminating infection and maintaining globe integrity. However, recurrence of infection in the graft remains a concern, especially in cases of resistant organisms such as Acanthamoeba, Pythium, or microsporidia or even in cases of some bacterial and fungal keratitis. The purpose of this review is to provide a comprehensive review and an update of peer-reviewed literature published over the last decade with emphasis on recalcitrant infections due to these pathogens.

INTRODUCTION

Microbial keratitis has an estimated annual incidence ranging from 113 cases to 799 cases per 100000 population in developing countries, resulting in an annual burden of 1.5-2 million new cases of unilateral blindness[1]. In cases of infection caused by virulent or atypical organisms, or in low-resource settings where access to timely diagnosis and appropriate medical care may be limited, corneal ulcers can progress rapidly to corneal melting or perforation[2,3]. In such severe or refractory cases, therapeutic penetrating keratoplasty (TPK) is often the last resort to save the eye. Depending on geography and clinical practice, approximately 3%-38% of severe microbial keratitis cases may ultimately require TPK[4].

TPK replaces a diseased cornea with donor tissue, with the dual goals of eradicating infection and preserving globe integrity in severe keratitis[5-7]. Although visual outcomes are secondary, they are also considered, as a transparent graft can provide useful vision if the infection is resolved and complications are effectively managed. Unlike elective optical keratoplasty (performed for visual rehabilitation in a quiet eye) or purely tectonic grafts (used to restore globe integrity in non-infectious corneal melting), TPK often serves as a sight-saving procedure by preventing globe rupture and endophthalmitis, both of which are devastating complications of uncontrolled corneal infection[8].

Numerous studies worldwide have reported the outcomes of TPK for infectious keratitis[9,10] demonstrating high rates of infection eradication (often > 85%-90%) and globe salvage[11]. However, these studies also highlight substantial challenges, including graft failure due to rejection or reinfection, complication rates such as secondary glaucoma (5.3%-60%) and often disappointing visual recovery[10]. Understanding these outcomes and the factors influencing them is critical for guiding clinical decisions, including optimization of the surgical procedure according to baseline clinical characteristics, timing of surgical intervention, and the use of adjunctive measures to improve success. Furthermore, in high-volume centers serving rural populations, implementing referral and telemedicine strategies in the context of recalcitrant microbial keratitis can help ensure timely surgical intervention, thereby minimizing the silent epidemic of corneal blindness[2].

This narrative review synthesizes current evidence on TPK for bacterial keratitis. It first outlines the various types of therapeutic keratoplasty and their respective roles. It then discusses indications for surgery and preoperative evaluation, with emphasis on timing and microbiological work-up. Intraoperative considerations, including choice of anesthesia and key surgical steps, are elaborated upon. Common complications, their management strategies, and prognosis in terms of both anatomical success and visual outcomes are also summarized. Postoperative management principles, particularly antimicrobial and immunosuppressive regimens, are highlighted. Emerging surgical modifications (e.g., lamellar keratoplasty and biologically preserved grafts) and adjunctive therapies (such as corneal cross-linking) are also discussed. Finally, the review examines community-based aspects, including rural eye care models, teleophthalmology, and public health initiatives aimed at improving access and outcomes in severe corneal ulcer cases.

METHODS

This narrative review was conducted in accordance with the PRISMA 2020 statement and followed Cochrane guidance for the qualitative synthesis of non-randomized evidence.

Search strategy

Four electronic databases - PubMed/MEDLINE, EMBASE, Scopus, and CENTRAL -were searched from 1 January 2015 to 30 April 2025. The search strategy combined controlled vocabulary and free-text terms related to therapeutic keratoplasty (e.g., “penetrating keratoplasty”, “therapeutic keratoplasty”, “tectonic graft”) with infection-specific qualifiers (e.g., “bacterial keratitis”, “infective corneal ulcer”, “fungal”, “viral”, “Acanthamoeba” etc.). No language or geographical restrictions were applied during the initial search stage. Key publications from the pre-2015 era were also reviewed to contextualize recent developments within classical knowledge and historical evidence, thereby enabling a comprehensive understanding of the subject matter.

Eligibility criteria

Studies were included if they evaluated TPK (with or without lamellar rescue) performed for microbial keratitis (pure or mixed series with extractable organism-specific data), enrolled at least five eyes, and reported at least one prespecified outcome, including anatomical success, therapeutic success, graft clarity, visual acuity, or complications. Only complete, peer-reviewed journal articles were included. Animal studies, duplicate datasets, and non-English language papers without available translations were excluded. When multiple publications analyzed the same cohort, the most comprehensive or latest study was retained.

Data extraction

A piloted, standardized data extraction worksheet was used to capture study design, demographic details, ulcer etiology, surgical parameters (including graft size and adjunctive procedures), duration of follow-up, and reported outcomes. Five reviewers independently screened titles, abstracts, and full texts of all eligible manuscripts. Each reviewer performed data extraction, which was subsequently verified by a second reviewer.

Primary outcomes included therapeutic success and anatomical success, defined as preservation of globe integrity at final follow-up. Secondary outcomes included graft clarity at the last visit, visual acuity (percentage achieving ≥ 6/60), and complications such as recurrence of infection, secondary glaucoma, graft failure, and related adverse events.

Statistical synthesis and heterogeneity assessment

Given the heterogeneity in study design, population characteristics, and outcome reporting, a formal meta-analysis was not feasible. However, pooled outcome estimates were calculated using weighted averages based on sample size for key endpoints, including therapeutic success, anatomical success, and visual outcomes. Between-study variability was assessed both descriptively and quantitatively using the I² statistic, where applicable, to estimate heterogeneity across studies. An I² value of < 25% was considered low, 25%-75% moderate, and > 75% high heterogeneity. Subgroup analyses were performed qualitatively based on pathogen type (bacterial, fungal, Acanthamoeba, Pythium, microsporidia), graft size, and timing of intervention to identify clinically relevant trends. Due to inconsistent reporting across studies, these subgroup analyses were interpretative rather than statistical. All pooled analyses and descriptive statistics were performed using IBM SPSS Statistics (Version 26; IBM Corp., Armonk, NY, United States).

Definitions of outcomes

To ensure consistency across included studies, standardized definitions were applied. Therapeutic success was defined as complete eradication of infection without recurrence and without the need for repeat keratoplasty or globe removal during follow-up. Anatomical success was defined as maintenance of globe integrity without phthisis or the need for evisceration or enucleation. A favorable visual outcome was defined as best-corrected visual acuity ≥ 6/60 (Snellen equivalent), representing ambulatory vision, while visual acuity ≥ 6/18 was classified as a good visual outcome. Graft clarity was defined as the absence of stromal edema or opacity sufficient to permit visualization of iris details. Where studies used different criteria, outcomes were recalibrated to these standardized definitions whenever feasible.

TYPES OF THERAPEUTIC KERATOPLASTY

Several subtypes and techniques can be distinguished.

TPK

The classical full-thickness corneal graft encompasses all corneal layers, from epithelium to endothelium[12]. This remains the standard approach for advanced infectious keratitis, particularly when ulcers are deep, full-thickness, or perforated. In TPK, the entire diseased cornea is trephined and replaced with a donor corneal button. Donor graft size often slightly exceeds host trephination size (by approximately 0.5-1 mm) to ensure complete excision of infected margins and secure suturing[13]. TPK offers the advantage of removing all infected tissue in a single procedure but carries a higher risk of graft failure, as the donor endothelium may undergo rejection or decompensation in an inflamed eye[14].

Therapeutic lamellar keratoplasty

In cases where infection is confined to the anterior stroma, with an intact Descemet’s membrane and endothelium, deep anterior lamellar keratoplasty (DALK) may be attempted. In this procedure, the infected stroma is dissected away, and donor corneal tissue lacking the Descemet-endothelium complex is sutured in place. This approach removes infected tissue while avoiding the introduction of donor endothelium, thereby reducing the risk of graft rejection. However, therapeutic DALK (TDALK) is not feasible in cases of corneal perforation or when infection involves the endothelium (e.g., endothelial exudates or suppurative infiltrates adherent to Descemet’s membrane).

A comparative study from Singapore demonstrated that, in advanced keratitis cases suitable for DALK, TDALK achieved infection cure rates comparable to TPK (84.6% vs 88%). Notably, one-year graft survival was higher in TDALK (90% clear grafts) compared with TPK (78.4%), and a greater proportion of patients achieved good visual outcomes (≥ 6/9) (50% vs 20%)[14]. However, in clinical practice, only a minority of therapeutic cases - often fewer than 20% - are suitable for a lamellar approach. Bagga et al[15] concluded that DALK performed in advanced Acanthamoeba keratitis with infiltrates > 8 mm is associated with higher complication rates and poorer graft clarity. Therefore, careful case selection is essential for TDALK success, and evidence supporting its broader use continues to evolve.

Tectonic (urgent) keratoplasty

This term overlaps with TPK but generally emphasizes structural support (e.g., sealing perforation) as the primary objective rather than infection eradication. For example, in large perforated fungal ulcers with a high risk of endophthalmitis, emergency grafting may be performed primarily as an eye-saving procedure, with subsequent interventions addressing persistent infection if necessary. Tectonic grafts may be either full-thickness or lamellar patch grafts. In microbial keratitis, therapeutic and tectonic goals frequently coincide. “Urgent PK” refers to full-thickness grafting performed emergently for severe corneal ulceration[16]. Glycerol-preserved corneas may also be used for tectonic purposes when fresh donor tissue is unavailable[17].

Patch grafts and alternate techniques

In addition to formal keratoplasty, several alternative surgical interventions may be employed for severe corneal ulcers. Conjunctival flap surgery, in which vascularized conjunctiva is sutured over the cornea, can stabilize and promote healing in superficial infectious ulcers once microbial sterilization has been achieved[18]. Similarly, tissue adhesives such as cyanoacrylate glue, combined with a bandage contact lens, can temporarily seal small perforations (< 2 mm), thereby deferring or occasionally avoiding transplantation. These techniques may serve as bridges to keratoplasty or, in selected cases, eliminate its necessity if adequate healing occurs. Ultimately, however, deep or progressive bacterial ulcers unresponsive to medical therapy generally require definitive TPK or TDALK[19].

Table 1 summarizes the characteristics of major TPK studies, which predominantly consist of single-center cohort designs (mostly retrospective) involving patients with severe infectious keratitis undergoing therapeutic keratoplasty.

Table 1 Characteristics of included studies on therapeutic keratoplasty for recalcitrant infective keratitis.

Ref.	Country	Design	Eyes (TPK for keratitis)	Etiology (%)	Mean graft size (mm)	Mean follow-up (months)	Regrafts	Recurrence and complications	Graft clarity and survival (%)	Favourable visual outcomes
Robaei et al[57], 2015	United Kingdom	Retrospective case series	50 (26 TPK, 24 OPK)	Acanthamoeba	NA	62.2 months	26.9% TPK, 12.5% OPK	1 in TPK	-	54.2% TPK, 26.9% OPK
Kitzmann et al[62], 2009	United States	Retrospective comparative case series	31 (22 TPK, 8 OPK)	Acanthamoeba	8.25 TPK, 8.50 OPK	-	55% TPK, 11% OPK	41% TPK, 22% OPK	9 (41%) TPK, 8 (89%) OPK	11 (50%) TPK
								Glaucoma - 32% TPK, 11% OPK		8 (89%) OPK
Kashiwabuchi et al[58], 2008	Brazil	Retrospective	32	Acanthamoeba	NA	24 ± 16 months	12 (37.5%)	Recurrence: 6%	45%	44%
								Graft failure: 56.2%		Mean VA 20/50
								Glaucoma: 40%
Abu Dail et al[59], 2024	Germany	Retrospective case series	28	Acanthamoeba	8.1 ± 0.4 mm	53 ± 42 months	8 (29%)	1 (4%) recurrence, 21 % glaucoma	-	14 (50%)
Zhang et al[50], 2023	-	Retrospective	59 (36 TPK, 23 LK)	Acanthamoeba	NA	NA	NA	10, 6 in TPK, 4 in LK	-	> 20/60 in 14 (38.9%) TPK, 15 (65.2%) LK
Bagga et al[15], 2020	India	Retrospective	23 DALK; 10 large size > 8 mm group, 13 < 8 mm group	Acanthamoeba	9.5 mm, 8 mm	5 months	NR	2 (20%)	32% advanced group, 74.1% in moderate group	NR
								1 (7.7%)
Kwitko et al[60], 2024	Brazil	Retrospective cohort	23; 17 early TPK vs 6 late TPK after 5 months	Acanthamoeba	NA	117 ± 79.61, 105.50 ± 60.90	NR	Early TPK 11.8% (2/17) vs 50% (3/6) late TPK	NR	9 (52.9%) in early vs 2 (33.3%)
Acharya et al[44], 2024	India	Retrospective	16	Pythium insidiosum	10.3 mm, 11.7 mm on repeat	8.9 ± 9.8 months	7/16	9/16	1/16 (6 months)	Mean UCVA 2.04 ± 0.77
Kate et al[51], 2023	India	Retrospective	50	Pythium insidiosum	NA	18.4 months	29/50 (58%)	Secondary glaucoma 20%	10/50 (20%)	Median VA 20/125
Saksurakan et al[31], 2024	Thailand	Retrospective	18 TPK, 2 DALK, 1 femtosecond-ALK	Microsporidia	NA	46	1/18 in TPK	4 (22%) TPK	NR, 8 graft failures	Range 20/40 to perception of light only
							1/3 in ALK	1/3 in ALK

Missing data reflect non-reporting in original studies. Visual acuity outcomes have been standardized to Snellen equivalents where possible. Sample size stratification by pathogen type is included when explicitly reported. Range of visual outcome: 20/20 to no light perception; median corrected distance visual acuity 20/30. TPK: Therapeutic penetrating keratoplasty; NA: Not available; ALK: Anterior lamellar keratoplasty; DALK: Deep anterior lamellar keratoplasty; LK: Lamellar keratoplasty; NR: Not reported separately; OPK: Optical penetrating keratoplasty; UCVA: Uncorrected visual acuity; VA: Visual acuity.

INDICATION FOR SURGERY

The decision to proceed with therapeutic keratoplasty in microbial keratitis depends on failure of medical therapy or an imminent threat to ocular integrity. The included studies and clinical guidelines converge on several key indications.

Corneal perforation or impending perforation

This is an absolute indication for urgent keratoplasty[16]. Even in cases with a small leak, or when a large ulcer has thinned the cornea to less than approximately 100 μm with a high risk of imminent rupture, surgical intervention is necessary to restore globe integrity. In one study, over 56% of cases had already perforated by the time TPK was performed[2]. Christy et al[20] similarly reported perforated ulcers in 45.9% of cases as the primary indication for TPK in their series. Thus, the occurrence of iris prolapse or a collapsed anterior chamber may necessitate emergency TPK, often within hours, to preserve the eye.

Non-resolving/progressive infection despite maximal therapy

Failure to respond may result from resistant organisms, inaccessible infectious foci (e.g., deep stromal abscesses), or uncontrolled inflammation. Misdiagnosis, atypical presentation, and prior treatment may further complicate management[21]. Consistent with the standard of care for refractory infectious keratitis, failed medical therapy should prompt TPK as a salvage procedure[22].

In bacterial ulcers that fail to respond to intensive antimicrobial therapy - evidenced by enlarging infiltrates, worsening suppuration, or lack of improvement over several days - TPK should be considered before perforation occurs[16]. If, after 48-72 hours of treatment with fortified antibiotics, no improvement or deterioration is noted, surgical management is typically planned. Bacterial keratitis generally has better outcomes than fungal keratitis due to the availability of effective antibiotics; however, particularly aggressive pathogens such as multidrug-resistant Pseudomonas may necessitate urgent surgical excision[8].

The need for penetrating keratoplasty in fungal keratitis varies widely, ranging from 15% to 55%[23-27]. A secondary analysis of data from the Mycotic Ulcer Treatment Trial 2 identified several predictors of eventual keratoplasty, including increased infiltrate size, stromal depth, and the presence of hypopyon[28].

Atypical micro-organisms, confirmed or suspected antimicrobial resistance

In Acanthamoeba keratitis, TPK is indicated in cases of deep or ring infiltrates unresponsive to anti-amoebic therapy, particularly when associated with hypopyon, pan-corneal infection, or impending perforation[5,11]. Scleral extension or persistent inflammation despite dual therapy (biguanides and diamidines) also warrants surgical intervention.

In Pythium insidiosum keratitis, early and prompt TPK is often required because the organism is rapidly proliferative, highly virulent, and poorly responsive to conventional antifungal therapy[21,29]. Indications include worsening keratitis despite brief anti-Pythium therapy (e.g., topical linezolid and azithromycin), peripheral guttering, persistent endoexudates, scleral involvement, rapid corneal melt, or perforation. Because Pythium may extend beyond the limbus, larger grafts or sclerokeratoplasty may be necessary[30].

In microsporidial keratitis, TPK is indicated for deep stromal infections unresponsive to medical therapy (e.g., topical fumagillin and oral albendazole) or when perforation is threatened. Although uncommon, this infection often mimics herpes simplex keratitis and may ultimately require transplantation for definitive cure[11,31].

Similarly, if microbiological evaluation identifies multidrug-resistant Nocardia or atypical Mycobacteria, early surgical excision of the infectious focus is indicated[26]. Early intervention may prevent extension into adjacent limbal or scleral tissues[10].

Imminent scleral extension or endophthalmitis

Infections involving the peripheral cornea with scleral spread or early signs of endophthalmitis (e.g., vitritis) warrant therapeutic transplantation, often using a large-diameter graft, to remove all infected tissue. Some cases may require penetrating keratoplasty combined with scleral patch grafting or pars plana vitrectomy. Although bacterial ulcers more commonly perforate than spread circumferentially, Pseudomonas in particular may produce ring ulcers with rapid peripheral extension. In such cases, grafts measuring 10-12 mm may be required, although these larger grafts are associated with higher risks of glaucoma and poorer outcomes[10].

The timing of TPK in microbial keratitis remains a matter of clinical judgment, balancing the risks of surgery in an inflamed eye against the consequences of delayed intervention. Current guidelines emphasize early surgery, ideally before perforation develops[10]. Koçluk and Sukgen[32] compared early vs delayed TPK and reported a trend toward superior outcomes with earlier intervention, including 100% clear graft survival and no reinfection among patients treated within one week, compared with a 17% reinfection rate in delayed cases. Although limited by small sample size, these findings reinforce the principle that timely TPK may arrest infection and prevent devastating complications, whereas delayed surgery increases the risk of irreversible damage such as endophthalmitis, panophthalmitis, or eventual evisceration. Indeed, studies consistently report that eyes requiring evisceration often had significant preoperative delays or infections caused by particularly virulent pathogens[22,33].

In clinical practice, once the indication for TPK is established, the surgical team must act promptly, often performing keratoplasty on an emergency basis.

PRE-OPERATIVE WORK-UP

Prior to performing therapeutic keratoplasty, a thorough yet expedited preoperative work-up is essential to guide therapy and maximize surgical success. The preoperative evaluation is focused on identifying the causative pathogen, assessing the extent of infection, and preparing the patient for a procedure that is often performed under urgent conditions. Effective triage is critical; for example, in cases involving a very large central ulcer in the only functional eye, involvement of a senior surgeon or consideration of a Gundersen conjunctival flap as a temporary measure may be appropriate if donor tissue is not immediately available. Conversely, in a healthy phakic patient with a mid-sized refractory ulcer, simultaneous planning of keratoplasty with cataract extraction may be considered to avoid subsequent surgery, particularly given that post-TPK cataract formation may occur in up to 65% of cases[8].

Clinical assessment and documentation

Detailed documentation of the ulcer’s size, depth, and any extension to the sclera or adjacent tissues should be performed. Clinical drawings or high-quality photographs are valuable for surgical planning, particularly for determining graft size, which should fully encompass the lesion with at least a 0.5-1 mm margin of clear tissue on all sides[13]. The presence of hypopyon, endothelial exudates, or lens involvement should also be noted. Ocular B-scan ultrasonography is often performed when the view is obscured to evaluate for vitritis or retinal detachment, findings that may significantly alter surgical management, such as the need for concurrent vitrectomy in cases of endophthalmitis.

General health and medications

Although urgency is paramount, a focused systemic evaluation is important. Underlying immunosuppression should be identified, and inadvertent prior steroid use should be documented, as corticosteroids may worsen infection severity and impair healing. Diabetes mellitus has been associated with a significantly higher risk of anatomical failure (odds ratio approximately 2.8)[9]. Accordingly, preoperative optimization of glycemic control and discontinuation of potentially harmful treatments, including traditional or inappropriate topical medications, should be prioritized.

Consent and counseling

Patients and their families should be clearly informed that the primary goal of surgery is globe preservation, with visual rehabilitation considered a secondary objective. The potential need for future interventions, including repeat grafting, cataract surgery, or glaucoma management, should be discussed. In rural or underserved populations, socioeconomic barriers to follow-up must also be considered, as poor compliance may compromise long-term outcomes[18]. Counseling should additionally address the use of donor corneal tissue, particularly where cultural or religious concerns may influence patient acceptance.

Microbiological investigations

Identification of the responsible pathogen is critical for guiding both intraoperative decisions and postoperative therapy. Corneal scrapings should ideally be obtained prior to surgery for Gram staining, potassium hydroxide preparation, and microbial culture[18,34]. In emergency cases involving perforation, intraoperative sampling of the ulcer margin or excised corneal button may provide valuable microbiological data.

Determining the specific pathogen is especially important in differentiating bacterial from fungal infections, as postoperative steroid protocols differ substantially. For instance, steroid therapy is generally deferred in fungal keratitis. In one series, 56% of cases yielded positive cultures, evenly divided between fungal and bacterial pathogens[8]. When routine cultures are negative but suspicion for atypical organisms such as Acanthamoeba or Nocardia remains high, special staining techniques or polymerase chain reaction testing may be indicated. Culture of the excised corneal button may also identify pathogens when preoperative scrapings fail[22].

Donor tissue preparation

Close coordination of the surgeon and the eye bank is essential to secure suitable donor tissue. For TPK, optical-grade donor tissue is not mandatory; medium-quality corneal tissue is generally sufficient. During periods of donor shortage, such as the coronavirus disease 2019 pandemic, glycerol-preserved corneas have been used successfully as emergency alternatives. Corneas preserved in 100% glycerol at -80 °C, while lacking viable endothelium, can provide structural support and restore globe integrity[17,35].

Systemic and topical antimicrobials

In severe bacterial keratitis, particularly in cases involving perforation or high risk of dissemination, some surgeons administer systemic broad-spectrum antibiotics preoperatively (e.g., intravenous vancomycin and ceftazidime) as prophylaxis against intraoperative bacterial spread or endophthalmitis, although robust evidence supporting this practice remains limited[36]. Similarly, systemic antifungal agents such as oral voriconazole or ketoconazole may be initiated preoperatively if fungal co-infection has not been excluded[24].

In microsporidial keratitis, topical fumagillin (0.3%) and/or oral albendazole may be administered preoperatively[31,37]. In Acanthamoeba keratitis, standard therapy consists of biguanides (e.g., polyhexamethylene biguanide or chlorhexidine) alone or in combination with diamidines such as propamidine or hexamidine[38,39].

At a minimum, intensive topical antimicrobial therapy should be continued until the time of surgery. Once the patient is optimized and donor tissue is secured, attention can then be directed toward anesthesia selection and definitive surgical management.

CHOICE OF ANESTHESIA

The choice of anesthesia for TPK requires balancing patient comfort, surgical safety in an open-globe setting, and the need for complete ocular akinesia. General anesthesia (GA) is the preferred modality in most cases due to the precarious nature of the procedure and the need for maximal patient immobility, although retrobulbar or sub-Tenon’s anesthesia with or without sedation may be acceptable in carefully selected adult patients[13]. The primary objective is to maintain a completely immobile, pain-free eye throughout surgery, thereby allowing the surgeon to perform meticulous intraocular work on an inflamed and structurally compromised globe without undue risk.

Patient factors

Patients presenting with severe corneal ulcers are often in significant pain, highly photophobic, and anxious. Remaining supine under an operating microscope with intense illumination may be intolerable without profound sedation. GA ensures complete analgesia, eliminates movement, and provides optimal operating conditions, which are especially critical when the globe is open and intraocular contents are vulnerable to extrusion. Bajracharya and Gurung[8] strongly advocated for GA whenever feasible, noting that 71% of their patients had perforated ulcers and would likely have been unable to tolerate surgery under local anesthesia (LA)[8].

Open-globe precautions

Once trephination is performed, the eye becomes an open globe. Any coughing, straining, vomiting, or sudden movement may precipitate choroidal effusion or catastrophic expulsion of intraocular contents. Under GA, airway control and physiologic stabilization allow the anesthesia team to maintain low intraocular pressure (IOP) and prevent patient movement. By contrast, LA carries an inherent risk of patient restlessness or Valsalva maneuvers, which may compromise surgical safety. Prior suturing of a Flieringa ring to the sclera prevents collapse of the globe in cases of low scleral rigidity, like children and myopes. To mitigate elevated IOP, some surgeons perform pars plana decompression or anterior chamber paracentesis before trephination[12]. Additional techniques, such as pupil-scaffolding, have also been described to minimize vitreous upthrust and reduce the risk of lens extrusion[40]. Administration of intravenous mannitol (300 mL approximately 30 minutes preoperatively) may further reduce IOP and vitreous pressure during surgery.

Duration and complexity

Therapeutic keratoplasty cases are often surgically unpredictable, with potential intraoperative requirements including iris repair, cataract extraction, anterior vitrectomy, or vitreous biopsy. GA provides the necessary flexibility for prolonged or complex procedures. Under LA, discomfort may increase significantly if surgery extends posteriorly or involves substantial iris manipulation. Furthermore, inflamed eyes may respond suboptimally to local anesthetic agents, resulting in incomplete analgesia or akinesia. Gümüş et al[13] reported using GA in the majority of their therapeutic cases, reserving retrobulbar or sub-Tenon’s anesthesia for selected situations while supplementing LA with intravenous sedation to ensure complete akinesia.

Pediatric or uncooperative patients

In pediatric patients, GA is mandatory for TPK[41]. Similarly, adults with altered mental status, poor cooperation, severe systemic illness, or intolerable ocular pain should also undergo surgery under GA.

Role of LA

Despite the advantages of GA, LA may be considered in select cooperative adults with peripheral ulcers, no perforation, and limited anticipated intraocular manipulation. Retrobulbar or peribulbar blocks using lidocaine with hyaluronidase can provide effective anesthesia and akinesia, occasionally supplemented by facial nerve block for orbicularis control. However, the current literature lacks robust evidence defining the optimal anesthesia modality across varying clinical scenarios, and the choice remains largely dependent on surgeon experience, patient factors, and available institutional resources.

SURGICAL PROCEDURE

This procedure is a meticulous and often challenging undertaking that demands surgical precision, adaptability, and sound intraoperative judgment. The primary goals are complete excision of infected tissue, restoration of ocular integrity, and creation of a stable environment for subsequent visual rehabilitation.

Tissue handling and trephination

Preoperatively, the surgeon estimates trephine size based on slit-lamp assessment of ulcer dimensions plus an adequate safety margin. Intraoperatively, the final trephine diameter is selected using Castroviejo calipers to ensure complete excision of the lesion with approximately 0.5-1 mm of surrounding uninvolved cornea[12,20]. In most cases, it is preferable to err on the side of a larger trephination to avoid retention of peripheral infected tissue.

A disposable suction trephine is typically centered over the ulcer, often in a decentered fashion depending on lesion location. For instance, inferotemporal ulcers require appropriately offset trephination. The host cornea is initially trephined to approximately 80% depth to minimize sudden decompression. Entry into the anterior chamber is then completed using a sharp blade, such as a 15-degree knife[13]. In cases with a shallow or flat anterior chamber, viscoelastic should be injected promptly upon chamber entry to stabilize intraocular structures.

Stabilizing the open globe

Once the infected corneal button is excised, the globe is rendered open and must be stabilized immediately. The anterior chamber is irrigated with balanced salt solution to remove inflammatory debris and infectious material. Intracameral antimicrobials are commonly administered at this stage - for example, vancomycin plus ceftazidime for bacterial keratitis, or amphotericin B or voriconazole for fungal infections[23,42,43]. Gümüş et al[13] specifically described anterior chamber irrigation with vancomycin and ceftazidime prior to graft placement in bacterial cases.

Viscoelastic agents, such as sodium hyaluronate, are then injected to maintain anterior chamber depth and protect intraocular tissues.

Debridement and inspection

Residual infected tissue at the wound margins must be carefully trimmed, particularly in fungal or Pythium keratitis, where microscopic extension is common[12,20]. Necrotic Descemet’s membrane should be excised where necessary. Fibrin, hypopyon, endothelial exudates, pupillary membranes, and other inflammatory debris should be meticulously removed to minimize postoperative inflammation.

In cases of scleral extension, adjunctive procedures such as scleral deroofing or double freeze-thaw cryotherapy may be required[44]. Posterior synechiae should be gently broken when present.

If advanced infection has resulted in cataract formation or lens contamination, simultaneous cataract extraction may be indicated. Preservation of the posterior capsule is preferred whenever possible, as it serves as a barrier against posterior extension of infection and endophthalmitis. Similarly, localized vitreous exudates may necessitate limited anterior vitrectomy through the graft opening[45].

Donor graft preparation

The donor cornea is generally trephined 0.5-1.0 mm larger than the host bed. For example, an 8.0 mm host trephination is typically matched with an 8.5 mm donor button[12,13]. Oversizing improves wound apposition, reduces postoperative wound gape, and compensates for tissue shrinkage. Surgeons should also ensure availability of larger donor tissue in case intraoperative extension of diseased tissue necessitates graft upsizing.

When glycerol-preserved or otherwise opaque donor tissue is used, careful orientation is essential[17]. In cases of active infection, some surgeons dip the donor tissue in antibiotic solution briefly before placing (though evidence for this is limited)[10]. In cases involving limbal or scleral extension, donor tissue with a scleral rim may be necessary.

Graft suturing

The donor button is positioned epithelial side up onto the recipient bed. Typically, 16 interrupted 10-0 non-absorbable monofilament nylon sutures are placed, beginning with cardinal sutures at 12 o’clock, 6 o’clock, 3 o’clock, and 9 o’clock, followed by evenly spaced sutures to ensure centration and secure wound closure[32]. Although some surgeons prefer a combined interrupted-running technique, interrupted sutures are often favored in TPK because they allow selective removal if recurrent infection develops at the graft-host junction.

Sutures in therapeutic cases are generally placed deeper and tighter than in elective penetrating keratoplasty to ensure a watertight seal, given the inflamed and compromised host tissue. Refractive considerations are secondary to tectonic integrity and slight wound override is tolerated if it ensures stability[12].

Intraoperative antibiotics and adjuncts

At the conclusion of surgery, additional intracameral antimicrobial agents may be administered prophylactically. Subconjunctival antibiotics are also commonly used, with corticosteroids added selectively in non-fungal cases[7]. In recalcitrant fungal keratitis, intrastromal antifungal injections around the graft-host junction may be considered, although evidence remains anecdotal[13].

In Pythium keratitis, intracameral linezolid (200 μg/0.1 mL) has been described as an adjunctive therapy[44].

Peripheral iridectomy

For grafts measuring ≥ 9 mm, particularly those involving the limbus, or in eyes with shallow anterior chambers, a peripheral iridectomy is often performed to reduce the risk of postoperative pupillary block glaucoma[46]. This is generally completed before closure of the final graft quadrant.

Throughout the procedure, strict infection control is paramount. Instruments contacting infected host tissue should either be re-sterilized or replaced before donor tissue handling. Many surgeons employ separate instrument sets for host and donor phases. Final ocular surface irrigation with povidone-iodine may also be performed as an additional precaution.

As reported by Zhang et al[9], meticulous surgical technique and rigorous aseptic precautions contributed significantly to high anatomical success rates of 87.7% over a 16-year period. Across major studies, anatomical success rates following TPK range from approximately 85% to 100% (Table 2). However, successful surgery is only the initial step, and vigilant postoperative management remains essential to preserve globe integrity and optimize long-term outcomes.

Table 2 Pooled outcomes of therapeutic keratoplasty in severe keratitis.

Outcome measure	Pooled result (95%CI)	Interpretation
Anatomical success (globe intact)	89% (88%-91%)	Eye preserved without requiring evisceration or enucleation in 9 out of 10 cases
Therapeutic success (infection cured)	91% (89%-92%)	Infection eradicated post-TPK in majority of eyes
		Recurrence of primary infection occurs in 9% overall (higher if fungal)
Clear graft at final follow-up	45% (43%-48%)	Proportion of grafts remaining clear (no significant opacity or failure) at last follow-up
		Many grafts turn opaque or fail over time despite initial success
Functional visual recovery	20% (range 15%-25% across studies)	Eyes achieving ambulatory vision (≥ 6/60) after TPK were a minority
		Poor initial ocular condition and graft complications limit visual acuity in most cases

Pooled estimates (random-effects) are derived from the 10 included studies using weighted averages based on sample size (95% confidence intervals in parentheses). “Functional vision” generally corresponds to ambulatory vision (approx ≥ 6/60 or ≥ 20/200 Snellen) at final follow-up; because definitions varied, a range is given instead of a pooled value. Statistical analyses were performed using IBM SPSS Statistics (Version 26). Heterogeneity across studies was assessed using I² statistics where applicable. TPK: Therapeutic penetrating keratoplasty; CI: Confidence interval.

COMPLICATIONS AND THEIR MANAGEMENT

TPK for infectious keratitis, while often sight-saving, is associated with a substantially higher incidence of complications than elective optical keratoplasty due to the underlying infectious and inflammatory milieu. Meticulous surgical technique and intensive postoperative management are therefore essential to minimize morbidity and optimize both anatomical and functional outcomes[47].

Recurrence of infection

Recurrence refers to reappearance of the original infection within the graft or at the graft-host interface. Reported recurrence rates range from 0% to 30%, with significantly higher rates observed in fungal and atypical infections[33,48,49]. Gram-negative bacterial ulcers also demonstrate comparatively poorer therapeutic outcomes. For example, although Anshu et al[14] achieved 100% anatomical success even in large perforations, 12% of cases subsequently developed reinfection. Similarly, one series reported unresolved infection in 44% of Gram-negative cases compared with approximately 10% of Gram-positive infections[22]. In Acanthamoeba keratitis, postoperative recurrence has been reported in nearly one-sixth of eyes[50].

Risk factors for recurrence include large ulcer diameter (> 6 mm), limbal involvement, dense endothelial plaques, extensive hypopyon, and significant stromal infiltration[33].

Management is done by intensive antimicrobial therapy guided by cultures from the trephined button or edge if possible. Management typically involves aggressive antimicrobial therapy guided by culture results from the excised corneal button or graft-host interface. While localized recurrences may occasionally respond to intensified medical therapy, repeat TPK is often required. Koçluk and Sukgen[32] reported repeat grafting in delayed surgical cases, while Gümüş et al[13] noted that 22% of cases required re-keratoplasty, with reinfection accounting for 77% of these procedures. Moon et al[45] highlighted that late recurrence can occur even a year later[46]. Aggressive surgical re-intervention at first sign of recurrence is often needed, sometimes combined with systemic therapy. In cases progressing to endophthalmitis, pars plana vitrectomy with intravitreal antimicrobials or even evisceration may be necessary[36,49].

Primary graft failure, delayed epithelialization or non-healing

Primary graft failure may occur due to endothelial decompensation, severe inflammation, or technical factors. Delayed epithelial healing is particularly common in glycerol-preserved grafts, with prolonged epithelial defects significantly increasing the risk of stromal melt and secondary infection[35]. In severe infections such as Pythium keratitis, median graft survival may be as short as 2.4 months, with clear graft survival rates as low as 20%[51].

Persistent epithelial defects should be aggressively managed with lubricants, bandage contact lenses, tarsorrhaphy, or amniotic membrane transplantation. Gümüş et al[13] reported amniotic membrane use in approximately 5% of cases to promote healing. Secondary optical keratoplasty may be considered once infection is fully eradicated and ocular inflammation has subsided.

Secondary glaucoma

Secondary glaucoma is among the most frequent postoperative complications, with reported incidence ranging from 20% to 43%[8]. Mechanisms include inflammatory angle closure, steroid-induced ocular hypertension, peripheral anterior synechiae, pupillary block, and trabecular meshwork dysfunction. Sharma et al[10] reported significantly higher glaucoma rates in perforated ulcers compared with non-perforated cases (29% vs 12%).

Initial management includes topical beta-blockers, alpha agonists, and carbonic anhydrase inhibitors, while prostaglandin analogues are generally avoided due to inflammatory concerns. Persistent or severe glaucoma may require surgical intervention, often in the form of glaucoma drainage devices, as trabeculectomy tends to perform poorly in inflamed grafted eyes. Peripheral iridectomy during surgery can reduce the risk of pupillary block[10].

Graft rejection

Corneal graft rejection may occur early due to heightened ocular inflammation, although differentiating rejection from recurrent infection or primary graft failure may be challenging. Reported rejection rates vary, with some studies documenting rates around 22% within the first postoperative year[2]. Suspected rejection in a previously clear graft warrants aggressive topical corticosteroid therapy, often supplemented with systemic immunosuppression when infection is no longer active. High-risk repeat grafts may benefit from adjunctive systemic agents such as cyclosporine or tacrolimus.

Structural graft complications

Structural complications include wound dehiscence, graft-host instability, and graft melt[42]. Graft melt is particularly concerning in fungal or persistent infections and may occur in approximately 6% of glycerol-preserved graft cases[35]. Management includes aggressive lubrication, anti-collagenase therapy (e.g., doxycycline and vitamin C), resuturing for minor wound leaks, or repeat grafting when necessary. In severe cases, conjunctival flap procedures may serve as globe-preserving alternatives.

Cataract

Cataract formation is highly prevalent following TPK, occurring in approximately two-thirds of phakic eyes[8]. Cataractogenesis may result from surgical trauma, chronic inflammation, corticosteroid use, or direct lens involvement. Once infection is resolved and the ocular surface is stable, cataract extraction may be undertaken, often with intraocular lens implantation when appropriate[52].

Special care is required during cataract surgery to preserve endothelial function and avoid graft compromise. The incisions should be peripheral or sclerocorneal to the extent possible, avoiding the graft-host junction and suture lines. The soft-shell technique is recommended using intracameral dispersive and cohesive viscoelastics. Phacoemulsification must be performed at a deeper plane to avoid fluid turbulence and energy dissipation near the endothelium, choosing the method of nucleotomy carefully to avoid unnecessary phaco power. Extracapsular cataract extraction or manual small-incision cataract surgery may sometimes offer superior endothelial preservation compared with phacoemulsification[53].

Retinal detachment and endophthalmitis

Although relatively rare, posterior segment complications such as retinal detachment and postoperative endophthalmitis can be devastating. Endophthalmitis rates of approximately 1.7% have been reported[13]. Prompt vitrectomy and intravitreal antimicrobial therapy are essential, although visual prognosis remains guarded[45].

Importance of postoperative surveillance

Close postoperative follow-up is critical for early detection and management of complications. Patients are typically monitored daily during the first postoperative week and at least weekly thereafter during the early postoperative period. Vigilant follow-up has been repeatedly associated with improved outcomes, as early identification of recurrence, elevated IOP, or graft compromise allows timely intervention. For example, Chatterjee and Agrawal[33] kept patients under intensive monitoring and could catch recurrences at a median of 16 days post-op - early enough to attempt re-grafting.

Timing of surgery

Studies consistently demonstrate that delayed surgery is associated with higher recurrence rates, poorer graft survival, and increased risk of severe complications. For instance, although 27% of Chatterjee’s cases had recurrence, they limited evisceration to 8.3% with timely re-intervention[33]. Likewise, Singh et al[22] noted that even with some failures, a portion of those eyes ended up with a failed graft but a quiet eye (infection cured) - which can later undergo repeat PK for vision.

Finally, complication rates are higher when TPK is done late or in suboptimal conditions. Early surgery means less intraocular damage and likely fewer sequelae. This is evidenced by Koçluk’s data where earlier TPK had zero recurrences and better clarity[32], and by Moon’s finding that delayed (> 30 days) cases had significantly more recurrences and worse outcomes[46]. Therefore, prevention of postoperative morbidity begins not only with surgical precision but also with optimal timing and comprehensive perioperative care.

In summary, although TPK is frequently globe-saving, long-term success requires proactive anticipation and management of complications. Early surgical intervention, careful patient selection, aggressive postoperative antimicrobial therapy, glaucoma surveillance, and patient education significantly improve outcomes.

PROGNOSIS FOR THERAPEUTIC SUCCESS

The primary measure of success in TPK is not visual acuity, as in elective keratoplasty, but rather therapeutic success - defined as eradication of infection and preservation of globe integrity. Visual rehabilitation, while important, is considered a secondary or functional outcome. Both aspects are critical in evaluating long-term prognosis.

Anatomical and therapeutic success

Overall, the prognosis for globe salvage and infection control following TPK is generally favorable, with more than 85% of cases achieving these primary objectives (Table 2). Across major studies, anatomical integrity and infection eradication rates typically range from 89% to 100%, particularly in cohorts dominated by bacterial keratitis[9,46]. These rates decline modestly when fungal or atypical infections are included; for example, Gümüş et al[13] reported a therapeutic cure rate of 79.7% in a cohort that included approximately 20% fungal ulcers. Even earlier studies from the 1990s demonstrated comparable success rates ranging from 70% to 100%[6].

Bacterial keratitis consistently demonstrates superior therapeutic outcomes compared with fungal infections, with significantly higher cure rates, better graft clarity, and lower recurrence rates[8]. Fungal keratitis is associated with greater postoperative recurrence and may require repeat TPK in up to 15% of cases[22,33]. This disparity is likely attributable to the ability of fungal organisms to invade adjacent scleral tissue or persist within microscopic residual foci despite apparently adequate excision. Similarly, virulent Gram-negative pathogens such as Pseudomonas aeruginosa are associated with markedly poorer outcomes compared with Gram-positive infections[22].

Atypical pathogens such as Acanthamoeba, herpes simplex virus, Microsporidia, Nocardia, and Pythium insidiosum also demonstrate elevated rates of recurrence and graft failure. For example, Saksurakan et al[31] reported an initial cure rate of 76.2% for microsporidial keratitis after first TPK, increasing to 95.2% following repeat procedures. Nocardia also have higher failure rates.

In Pythium keratitis, outcomes are particularly guarded[29], with globe-threatening complications and evisceration rates reported as high as 57.5% in some series[54]. Delayed presentation, dense stromal infiltration, posterior stromal involvement, and extensive hyphal burden are major predictors of poor prognosis[41].

Ulcer-specific factors associated with poorer therapeutic outcomes include lesion size > 6 mm, involvement of more than 50% of corneal surface area, limbal or scleral extension, perforation, and delayed intervention. If endophthalmitis is present, the prognosis for the eye is guarded, hence several studies exclude these cases from their cohorts or count them as therapeutic failures[33]. Delayed presentation with infections ongoing more than a month pre-TPK have significantly higher recurrence, suggesting that these chronic cases might harbor microbes in the sclera[46].

Conversely, timely surgery with smaller graft sizes, particularly before perforation develops, is associated with improved anatomical outcomes[10]. Also, adherence to post-op therapy influences prognosis; patients who can comply with rigorous medication schedules and follow-ups likely do better[32].

Visual prognosis

Despite generally favorable rates of globe preservation, visual prognosis following TPK remains guarded. Most studies report relatively low rates of meaningful visual recovery. Sharma et al[10] found that only approximately 15% of eyes achieved final visual acuity better than 6/60, while Chatterjee and Agrawal[33] reported that only 17% reached this threshold at three months. Bajracharya and Gurung[8] similarly observed that only 25.4% of patients achieved better than 6/60 vision, with merely 4% reaching ≥ 6/18.

Pooled evidence suggests that approximately 20% of eyes regain ambulatory vision (≥ 6/60), while fewer than 10% achieve good visual outcomes (≥ 6/18). In Acanthamoeba keratitis, approximately 39% of eyes have been reported to recover ≥ 20/60 vision following TPK[13], with TPK performed within 5 months of Acanthamoeba keratitis leading to better final visual acuity than deferred cases[38]. Outcomes are improved when early intervention, smaller grafts, and aggressive postoperative rehabilitation strategies are employed. Christy et al[20] noted that with further interventions (like secondary optical grafts or endothelial keratoplasty for endothelial graft failure), some patients eventually improved - but at 3-year follow-up, only about 8% had a clear graft and presumably good vision[20].

Importantly, even limited visual recovery may represent a significant functional benefit compared with painful blindness or an eviscerated eye. Many patients retain hand motions to counting-fingers vision, which may still substantially improve independence and quality of life[10]. For a rural laborer or an elderly patient, this may mean the difference between independent ambulation vs total dependence.

The best visual outcomes are typically observed in carefully selected cases undergoing lamellar therapeutic approaches or staged visual rehabilitation. For instance, Anshu et al[14] reported that TDALK achieved visual acuity ≥ 6/9 in 50% of selected cases where deeper ocular structures remained uninvolved. In bacterial TPK cases, up to 42% of patients may achieve ≥ 20/200 vision, particularly when secondary interventions such as optical grafting, cataract extraction, glaucoma control, or contact lens rehabilitation are pursued[22].

Role of secondary optical rehabilitation

In many cases, the initial therapeutic graft is intended primarily as a globe-saving procedure rather than a definitive visual intervention. Once infection is eradicated and the eye is clinically quiet, secondary optical keratoplasty may be considered to improve visual outcomes[55]. This staged approach is common and often necessary. Christy et al[20] reported that 17% of patients underwent subsequent elective optical keratoplasty by three years following initial TPK.

Secondary grafts performed in quiet eyes generally have better survival and visual outcomes than primary therapeutic grafts, although success may still be limited by pre-existing vascularization, glaucoma, cataract, or retinal damage. Therefore, preoperative counseling should emphasize that TPK often represents the first stage in a prolonged rehabilitative process, with future surgeries potentially required to maximize visual function.

Overall perspective

In summary, TPK offers high rates of therapeutic and anatomical success, particularly when performed early and in bacterial keratitis, but functional visual outcomes remain limited in many patients. Early diagnosis, timely surgical intervention, aggressive postoperative care, and staged rehabilitative strategies are essential to optimize both globe salvage and eventual visual potential. While TPK frequently succeeds in preserving the eye, achieving meaningful long-term visual rehabilitation often requires additional surgical and medical interventions.

POST-OPERATIVE MANAGEMENT

Post-operative management following TPK for bacterial keratitis is as critical as the surgical procedure itself. It is a complex, intensive, and often prolonged process aimed at preventing recurrence of infection, controlling inflammation, promoting graft survival, and managing complications. A multidisciplinary approach involving ophthalmologists, microbiologists, and nursing staff is essential for optimal outcomes[56]. Christy et al[20] reported that at 6 months, 86% of eyes maintained anatomical success compared to 98.5% initial anatomical success at 1 month, reflecting the importance of effective post-operative management.

Antimicrobial therapy

In all cases, aggressive topical antimicrobial therapy is continued postoperatively, typically consisting of 3-4 weeks of intensive topical (broad-spectrum or targeted) antibiotics, with prolonged antifungal or amoebicidal therapy for fungal or Acanthamoeba cases. Antibiotic drops (fortified vancomycin 5% and ceftazidime 5%, or a fluoroquinolone) are generally continued for 4-6 weeks[13]. If the etiology is unclear or mixed, a combination of antibacterial and antifungal drops (fortified antibiotics and natamycin hourly) is maintained. After several weeks, therapy may be adjusted if post-transplant cultures show no growth[18].

Some surgeons administer subconjunctival depot antibiotic injections at the conclusion of surgery to provide high local drug concentrations during the immediate postoperative period. Typically, antimicrobial drops are administered hourly or every 2 hours initially, then gradually tapered over subsequent weeks if no signs of recurrent infection develop[18]. If cultures identify a specific bacterium, therapy is narrowed accordingly[3].

Adjunctive antimicrobial routes are used in high-risk cases and may include intracameral injections, intravitreal injections when posterior segment involvement is suspected, and systemic antibiotics for severe bacterial infections, especially in cases involving Pseudomonas or scleral extension. Rigorous antimicrobial therapy is generally maintained for at least 3 months to ensure complete eradication[22].

In Acanthamoeba keratitis, combination therapy with topical biguanides (polyhexamethylene biguanide or chlorhexidine) and diamidines (propamidine or hexamidine) is continued for several weeks to months after grafting[57-60]. Propamidine isethionate with neomycin has also been recommended for at least 6 months, as inadequate postoperative therapy has been identified as a major cause of recurrence[1,61]. Oral anti-amoebic agents may also be used.

In Pythium keratitis, recurrence rates may be as high as 56%, with endoexudates being the most common manifestation. Postoperatively, topical 0.2% linezolid and 1% azithromycin are continued 6-8 times daily for a minimum of one month[30]. Systemic linezolid 600 mg twice daily plus azithromycin may also be administered for several weeks[51].

Topical corticosteroids

The use of corticosteroids following TPK is complex, as steroids reduce inflammation and rejection risk but may increase the likelihood of recurrent infection if initiated prematurely.

In bacterial keratitis, topical corticosteroids are generally introduced relatively early, often within the first 3-5 days postoperatively, once infection control is reasonably assured. Some surgeons may initiate low-dose steroids on postoperative day 1 in carefully selected cases. Topical prednisolone acetate 1% or dexamethasone 0.1% is commonly used to suppress inflammation and reduce immunologic graft reactions. In Gümüş et al’s series[13], dexamethasone 0.1% was initiated six times daily almost immediately after surgery and tapered gradually over 12 months. Since these grafts are at a higher risk of rejection due to a vascularized host bed, the steroid was continued long-term, at least weekly or biweekly to mitigate the risk.

In fungal or uncertain infections, corticosteroids are delayed significantly, often for at least 2 weeks after confirmed infection control. In Acanthamoeba keratitis, steroids are used with extreme caution (e.g. in severe inflammation), and are generally postponed until at least 2 weeks postoperatively, while anti-amoebic therapy is continued for at least 2 months and withheld only after ensuring absence of amoebae by repeated scrapings[38,62].

In Pythium keratitis, steroids are typically withheld until microbiological and clinical evidence and culture confirms infection eradication. If both preoperative and postoperative cultures remain positive, steroid initiation may be delayed for at least 3 weeks. If the clinical scraping is culture positive and postoperative button culture is negative, it is recommended to initiate steroids after 2 weeks of TPK[30,44]. Similarly, in microsporidial keratitis, steroids are introduced only after infection resolution. This entity is often mistaken for herpes keratitis – and one should avoid antiviral medication as they are ineffective[11].

Glaucoma management

IOP is monitored closely, often daily during the early postoperative period, due to the high risk of secondary glaucoma from inflammation, retained viscoelastic, or steroid response. Elevated IOP is treated promptly with topical beta-blockers and alpha-agonists. Steroid-induced glaucoma may require tapering or switching to lower-potency agents such as loteprednol. If IOP spikes above 25 mmHg, oral acetazolamide is added. If severe inflammation is the cause, steroids or cycloplegics often help in IOP control[12]. In a study on tele-ophthalmology data from India, 2.3% of missed referrals were for raised IOP - highlighting that even in community follow-up, attention to IOP is vital[63].

Supportive care

Supportive management focuses on epithelial healing and structural integrity. Bandage contact lenses may be used to promote epithelialization. Preservative-free lubricants are administered frequently, and adjunctive oral vitamin C and doxycycline may help reduce collagenolysis and promote healing[6].

Suture management

Sutures are generally retained longer than in elective keratoplasty due to delayed wound healing. Unless loose or problematic, sutures may remain for 6-12 months because the graft-host wound healing may be compromised by infection. If vascularization or rejection is creeping along a suture track, that particular suture might be removed earlier. Astigmatic suture adjustment is low on priority but selective suture removal can be done at 3-6 months to improve refraction.

Monitoring

Frequent postoperative follow-up is mandatory: Daily for the first week, every few days during the second week, weekly for the first month, and monthly thereafter. Clinical evaluation includes visual acuity, graft clarity, wound integrity, anterior chamber status, and IOP monitoring. Graft infiltrates, endothelial rejection line, or recurrence or scleral extension, if detected, require immediate action[6].

In resource-limited or rural settings, inpatient observation, temporary lodging, or teleophthalmology may improve compliance and outcomes. Telemedicine networks can facilitate remote monitoring and prompt referral for complications[63].

Adjunctive immmunomodulation

In high-risk large or sclerolimbal grafts, topical cyclosporine A (0.05%-0.1%) or tacrolimus (0.01%) may be introduced after the early postoperative period to reduce rejection risk. Long-term low-dose corticosteroids (e.g. prednisolone 1-2 times a week) may be maintained for 1-2 years in vascularized graft beds[13]. If the patient is at risk of non-compliance, some centers give a subconjunctival steroid depot (e.g., triamcinolone acetate) around 1 month post-operatively to provide immunosuppression.

Visual rehabilitation

Once infection is eradicated and the eye is stable after 3-6 months, visual rehabilitation may include rigid gas-permeable lenses, suture manipulation, or secondary optical keratoplasty. Typically, re-grafting is delayed for at least 6-12 months to ensure ocular quiescence. Meanwhile, amblyopia in younger patients is treated by refraction and patching the better eye if needed to maximize final outcomes.

Overall, successful postoperative management requires close collaboration among corneal surgeons, glaucoma specialists, microbiologists, and rehabilitation teams. In underserved populations, coordinated referral systems and teleophthalmology can substantially improve long-term graft survival and functional outcomes. Many rural patients will need reminders and perhaps involvement of local health workers to continue their drops appropriately after they leave the tertiary center. A rural center can also follow up the patient’s graft weekly and communicate with the surgeon remotely. Das et al[63] demonstrated in a large network of such centers, that implementing protocols to effectively triage red eyes can ensure those needing further care are promptly referred in case of early rejection, infection recurrence or other complications.

EMERGING THERAPIES AND UNMET CLINICAL NEEDS

Recent advances in the management of refractory microbial keratitis have introduced adjunctive and alternative strategies to conventional TPK. Corneal collagen cross-linking has demonstrated antimicrobial and stroma-stabilizing effects, particularly in early bacterial keratitis; however, its role in advanced infections remains limited due to poor penetration and risk of endothelial toxicity. Biologically preserved grafts, including glycerol-preserved corneas, have emerged as valuable alternatives in resource-limited settings and during tissue shortages. While they provide excellent tectonic support, their lack of viable endothelium limits long-term optical outcomes, often necessitating secondary procedures. Lamellar techniques such as TDALK offer theoretical advantages of reduced rejection and better graft survival; however, their applicability is restricted to cases without deep stromal or endothelial involvement. A critical unmet need remains in the management of Pythium keratitis, which is associated with aggressive progression, high recurrence rates, and poor graft survival. Early surgical intervention, larger graft margins, and adjunctive antimicrobial strategies are being explored but lack standardized protocols. Another key challenge is determining the optimal timing for secondary optical keratoplasty, balancing graft survival against visual rehabilitation. Prospective studies are required to define evidence-based timing strategies. Overall, future research should focus on pathogen-specific surgical algorithms, integration of molecular diagnostics, and development of targeted antimicrobial therapies.

LIMITATIONS

This article has several limitations. First, as a narrative review, there is an inherent risk of selection bias in study inclusion despite adherence to PRISMA-based search strategies, based on the publication timeline of studies pertaining to particular pathogens. Second, the majority of included studies were retrospective cohort designs, limiting the strength of evidence and introducing potential confounding factors. Third, heterogeneity in outcome definitions, follow-up durations, and reporting metrics limited direct comparability across studies and precluded formal meta-analysis. Although pooled estimates were calculated, these should be interpreted cautiously. Fourth, pathogen-specific subgroup analyses were largely qualitative due to inconsistent data reporting. Finally, publication bias and underreporting of negative outcomes may influence the overall interpretation of therapeutic success rates. These limitations highlight further need for standardized reporting and prospective multicentric studies in therapeutic keratoplasty.

CONCLUSION

TPK remains an indispensable tool in the armamentarium of ophthalmic surgeons for the management of refractory infectious keratitis. Penetrating keratoplasty continues to serve as the principal therapeutic intervention in severe microbial keratitis, while lamellar techniques are reserved for carefully selected cases. Prognosis for anatomical and therapeutic success is generally favorable, strongly supporting the use of TPK as a globe-preserving procedure in eyes that would otherwise be at high risk of irreversible loss. Visual prognosis, however, remains considerably more variable and is often limited, reinforcing the primary objective of TPK as eradication of infection and preservation of ocular integrity rather than immediate visual rehabilitation.

Post-operative management following TPK is intensive and prolonged, requiring aggressive and long-term antimicrobial therapy, judicious corticosteroid use, vigilant monitoring for complications such as glaucoma or recurrence, and long-term follow-up. With meticulous multidisciplinary care, most eyes can achieve infection control and anatomical preservation. Long-term graft maintenance and eventual visual rehabilitation through secondary interventions remain the ultimate goals of comprehensive management.