Purpose
To comparatively evaluate the clinical and microbiological profiles and treatment outcome of cases with post-penetrating keratoplasty (PK) infectious keratitis in failed and clear grafts.
Design
Retrospective, matched cohort study.
Methods
All cases of infectious keratitis following penetrating keratoplasty admitted to the Royal Victorian Eye and Ear Hospital, Melbourne, between January 1998 and December 2008 were identified through a retrospective medical chart review. Cases without any surface sutures were selected and divided into 2 groups, microbial keratitis with pre-existing graft failure and microbial keratitis in clear graft on presentation. Demographic, clinical, and microbiological profiles of cases in both groups were analyzed and compared.
Results
Patients in the failed-graft group were older ( P = .004) and had an early onset of graft infection ( P = .049), compared with patients in the clear-graft group. All patients in the failed-graft group were on long-term corticosteroid drops at the time of presentation (vs 76% in clear-graft group; P = .005). Moraxella sp was more frequently isolated in the failed-graft group (30.4%) compared with the clear-graft group (8%). A higher number of cases in the clear-graft group required surgical intervention in the form of corneal gluing and therapeutic corneal transplantation compared with the failed-graft group ( P = .03).
Conclusions
Prolonged use of corticosteroid eye drops is a major risk factor for the occurrence of postkeratoplasty infectious keratitis in failed and clear grafts. Infections in failed grafts occur earlier as compared to clear grafts, and indolent organisms like Moraxella are prevalent in patients with failed grafts in Australia.
Infectious keratitis following penetrating keratoplasty (PK) is a serious complication associated with poor graft survival and visual outcome. The reported incidence of graft infection after PK varies from 1.76% to 12.1%. Eyes with corneal grafts may be vulnerable to infection because of exposed sutures, epithelial defects, graft hypoesthesia, and prolonged use of topical corticosteroids and eye drops with preservatives. In addition, failed and decompensated grafts have been recognized as risk factors for the occurrence of infectious keratitis in grafted eyes. These eyes have unhealthy and loose epithelium, which suffers from frequent breakdown, consequently predisposing the eyes to infection. We analyzed and comparatively evaluated the clinical and microbiological profile of cases with post-PK microbial keratitis in failed and clear grafts.
Methods
This study was a retrospective analysis. All cases of patients with infectious keratitis following penetrating keratoplasty admitted to the Royal Victorian Eye and Ear Hospital, Melbourne, between January 1998 and December 2008 were identified through a medical chart review. Cases were identified from the database of clinical codes, based on the primary diagnoses recorded for each patient on discharge from the hospital. Medical records were then retrieved and data were recorded on a standardized form.
Infectious keratitis was defined by the presence of corneal infiltration along with an overlying epithelial defect. Diagnosis was based on the history and clinical examination. At initial presentation, all patients underwent a detailed clinical evaluation including measurement of visual acuity and slit-lamp biomicroscopic examination. Corneal scrapings were obtained for microbiological evaluation, specifically for Gram stain, Blankophor preparation, and culture in chocolate agar, Sabouraud’s dextrose agar, and thioglycolate broth. A swab was taken for detection of herpesvirus by polymerase chain reaction in suspected cases. Intensive antimicrobial therapy was started after corneal scrapings and treatment was modified as indicated by culture results, sensitivity pattern, and clinical response. Data reviewed included patient’s age, sex, predisposing factors for infection, clinical presentation, microbiological profile, treatment, and outcomes.
Suture-related graft infections are one of the leading causes of post-PK keratitis reported in the literature. In order to avoid bias from suture-related infections after PK, we included cases without any surface sutures in the present study. Incident cases of infectious keratitis were divided into 2 distinct groups: microbial keratitis with pre-existing graft failure and microbial keratitis in clear graft. For the purpose of our study, graft failure was defined by the presence of a cloudy cornea with loss of central graft clarity, with or without epithelial or stromal edema sufficient to compromise vision.
All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) for Windows (version 12.0; SPSS Inc, Chicago, Illinois, USA). The t tests were used to test for significant differences between the proportion of responders and nonresponders, with a P value <.05 considered significant.
Results
We reviewed 650 case files of post-PK patients who were following up at our hospital over a period of 10 years. Of these, 122 episodes of post-PK microbial keratitis were identified and subsequently cases of 51 patients without any surface sutures were included for analysis. On further classification, 26 patients (51%) had pre-existing graft failure and 25 (49%) had clear graft at the time of presentation.
Preoperative Characteristics
The number of female patients (n = 17) was higher than the number of male patients (n = 9) in the failed-graft group as compared to the clear-graft group (12 male and 13 female patients). Mean age of patients was significantly higher in the failed-graft group (75.00 ± 14.76 vs 61.80 ± 16.31 years) ( P = .004). Mean donor age ( P = .54) and endothelial cell count ( P = .36) were slightly higher in the clear-graft group ( P = .54) ( Table 1 ).
| Keratitis in Failed Graft (n = 26) | Keratitis in Clear Graft (n = 25) | P Value | |
|---|---|---|---|
| Male/female | 9/17 | 12/13 | .49 |
| Mean age (y) ± SD (range) | 75 ± 14.76 (43.85-91.68) | 61.80 ± 16.31 (29.5-93.39) | .004 |
| Donor age (y) ± SD (range) | 42.09 ± 34.29 (11-82) | 47.64 ± 30.32 (11-81) | .54 |
| Donor endothelial cell count (cells/mm 2 ) ± SD (range) | 2527.17 ± 409.27 (1765-3174) | 2623.53 ± 323.21 (2105-3225) | .36 |
Overall, pseudophakic bullous keratopathy (PBK) was the most common indication for corneal transplantation. PK was performed in a significantly higher number of patients in the failed-graft group ( P = .02), whereas the clear-graft group had a higher number of patients who were operated for keratoconus ( P = .002) ( Table 2 ).
| Keratitis in Failed Graft | Keratitis in Clear Graft | Total | P Value | |
|---|---|---|---|---|
| Pseudophakic bullous keratopathy | 9 | 3 | 12 | .97 |
| Failed graft | 7 | 1 | 8 | .02 |
| Corneal dystrophy | 3 | 3 | 6 | >.99 |
| Keratoconus | 0 | 7 | 7 | .002 |
| Herpetic keratitis | 2 | 2 | 4 | >.99 |
| Chemical burns | 2 | 1 | 3 | .13 |
| Corneal scar (noninfectious) | 2 | 3 | 5 | .63 |
| Corneal perforation | 1 | 4 | 5 | .35 |
| Wound dehiscence | 0 | 1 | 1 | – |
Characteristics of Graft Infection
The mean time interval between corneal graft and graft failure was 69.42 ± 33.2 months (range: 29.6-146.56 months) in the failed-graft group. Mean interval between corneal grafting and onset of graft infection was significantly shorter in the failed-graft group (71.96 ± 32.38 months) compared with the clear-graft group (113.94 ± 97.20 months) ( P = .049). The mean interval between graft failure and infection in the failed-graft group (2.53 ± 1.7 months) was also significantly shorter than the mean interval between corneal graft and infection in the clear-graft group (113.94 ± 97.20 months; P < .0001).
Of the 26 patients included in the failed-graft group, 7 (27%) developed infection within 2 years of corneal transplantation, whereas 19 (73%) developed infection after 2 years from the time of corneal transplantation. In the clear-graft group, 9 of 25 patients (36%) developed infection within 2 years and 16 (64%) developed infection after 2 years from the time of cornea transplantation ( P = .49).
Associated Risk Factors
At the time of presentation, all patients (26/26; 100%) in the failed-graft group and 19 patients (19/25; 76%) in the clear-graft group were using topical corticosteroids ( P = .005). Overall, 50.9% of patients (n = 26) were using antiglaucoma eye drops on presentation (n = 16, failed-graft; n = 10, clear-graft; P = .11). Seven patients in the clear-graft group had ocular surface disease (n = 7) compared with 3 patients in the failed-graft group (n = 3; P = .15). In addition, 3 patients in the failed-graft group and 2 patients in the clear-graft group had nonhealing epithelial defects. One patient in the clear-graft group had contact lens–related infection in the graft.
There were no significant identifiable systemic risk factors in either group.
Clinical Presentation and Microbiological Profile
There was no difference in the location of the corneal infiltrate in either group (central: failed-graft = 12, clear-graft = 11, P = .89; midperipheral, failed-graft = 10, clear-graft = 10, P = .89; peripheral, failed-graft = 4, clear-graft = 4, P = .93). Six patients in the failed-graft group (6/26; 23%) and 2 patients in the clear-graft group (2/25; 8%) had an associated hypopyon.
Microbiological cultures were positive for bacterial organisms in 23 of 26 patients (88%) in the failed-graft group compared with 19 of 25 (76%) in the clear-graft group ( Table 3 ). Overall, coagulase-negative Staphylococcus was the most commonly isolated organism. Moraxella sp was more common in the failed-graft group, present in 7 of 23 patients (30.4%), as compared with the clear-graft group (2/25, 8%). One patient in the failed-graft group and 5 patients in the clear-graft group had coexistent herpetic infection. Candida infection was seen in 1 patient in the clear-graft group. Mixed bacterial isolates were seen in 10 of 23 patients (43.5%) and 7 of 19 patients (36.8%) in the failed-graft and clear-graft groups, respectively.
| Keratitis in Failed Graft | Keratitis in Clear Graft | Total | P Value | |
|---|---|---|---|---|
| Coagulase-negative Staphylococcus | 4 | 7 | 11 | .25 |
| Moraxella sp | 7 | 2 | 9 | .06 |
| Staphylococcus aureus | 4 | 4 | 8 | .92 |
| Corynebacterium sp | 4 | 1 | 5 | .17 |
| Pseudomonas aeruginosa | 2 | 0 | 2 | .13 |
| Streptococcus pneumoniae | 1 | 1 | 2 | >.99 |
| Enterococcus faecalis | 1 | 1 | 2 | >.99 |
| Mycobacterium sp | 0 | 2 | 2 | .14 |
| Propionibacterium acnes | 0 | 1 | 1 | .31 |
| Herpes simplex virus | 1 | 5 | 6 | .07 |
| Candida albicans | 0 | 1 | 1 | .31 |
The initial treatment of infectious keratitis consisted of intensive topical antibiotics in the form of hourly ofloxacin hydrochloride 0.3% eye drops in all patients. Nine patients in the failed-graft group and 8 patients in the clear-graft group were treated with corticosteroids. Overall, 23 of 26 patients (88%) in the failed-graft group and 17 of 25 patients (68%) in the clear-graft group healed with corneal scarring ( P = .08). In addition to the medical treatment, 7 of 26 patients (27%) required an adjunctive surgical procedure in the failed-graft group compared with 19 of 25 patients (76%) in the clear-graft group ( Table 4 ). Botulinum toxin injection was used for nonhealing epithelial defect in 5 cases. Corneal transplantation surgery was performed to manage corneal perforation in 2 of 26 patients (8%) in the failed-graft group compared with 8 of 25 patients (32%) in the clear-graft group ( P = .03). Of the 8 patients with corneal perforation in the clear-graft group, microbiological results showed Staphylococcus spp (n = 5), Streptococcus pneumoniae (n = 1), Propiniobacterium (n = 1), and Mycobacterium (n = 1). Two patients with corneal perforation in the clear-graft group grew coagulase-negative Staphylococcus and Corynebacterium . Corneal gluing was performed in 24% of patients (6/25) in the clear-graft group and 12% (3/26) in the failed-graft group ( P = .260). One patient in the failed-graft group (1/26; 4%) underwent enucleation because of extensive, nonresolving bacterial keratitis.
