To determine the impact of presence or absence of sutures in cases with post–penetrating keratoplasty (PKP) microbial keratitis.
A 10-year retrospective chart review of post-PKP patients admitted with microbial keratitis at the Royal Victorian Eye and Ear Hospital, Melbourne, between January 1998 and December 2008 was undertaken.
Patients were categorized in 2 groups, “sutures present” and “sutures absent.” Main parameters evaluated were clinical and microbiological profile and treatment outcome.
One hundred and twenty-two episodes of microbial keratitis were noted in 101 patients: 71 (58.2%) with sutures present and 51 (41.8%) with sutures absent. Overall, pseudophakic bullous keratopathy was the most common indication for keratoplasty ( P = .92). Ocular surface disorder was the commonest risk factor associated with the occurrence of infection in both groups ( P = .17). Infections caused by Moraxella sp . ( P = .001) were significantly more common in the “sutures absent” group. Surgical interventions were required for 47 episodes (39%), with corneal gluing performed in significantly higher number of cases in the “sutures absent” group (40% vs 15%; P = .05). Multivariate analyses did not reveal any significant associations. Final mean visual acuity outcome was poorer in the “sutures absent” group (logMAR 2.10 ± 0.92 vs 1.76 ± 0.96; P = .04).
Corneal graft infections, in the presence and absence of sutures, share similar indications and risk factors. However, infections caused by indolent microorganisms were more prevalent in grafts without sutures. This group of patients required a higher number of surgical interventions in the form of corneal gluing and the overall visual outcome was poor.
Microbial keratitis following corneal transplantation is a major complication associated with loss of graft clarity and poor visual outcome. The incidence of graft infection varies considerably between studies, notably higher in reports from the developing world. The risk factors associated with the occurrence of graft infection that have been reported to date include presence of sutures, persistent epithelial defects, ocular surface disorders, recurrence of herpes simplex keratitis, graft failure, graft hypoesthesia, and lid abnormalities. Suture-related problems have been correlated with the occurrence of microbial keratitis within 1 year following corneal transplantation. Although sutures themselves are recognized as a nidus of infection, there has been no study that has independently addressed the impact of presence or absence of sutures on the clinical and microbiological pattern of microbial keratitis in a corneal graft. In this retrospective study we comparatively evaluated the demographic, clinical, and microbiological profile; graft failure; and visual outcomes of incident microbial keratitis cases after penetrating keratoplasty in the presence and absence of sutures.
A retrospective chart review of all patients with microbial keratitis following penetrating keratoplasty, admitted to the Royal Victorian Eye and Ear Hospital, Melbourne, Australia, between January 1998 and December 2008, was undertaken. All cases with a corneal infiltrate and overlying epithelial defect were labeled as having microbial keratitis. Patients with herpetic keratitis infection based on clinical evidence and/or tested positive to polymerase chain reaction were excluded from the study. Incident cases were divided into 2 groups: “sutures present,” defined by microbial keratitis in a corneal graft with any number of sutures in situ, and “sutures absent,” defined by microbial keratitis in a corneal graft without any sutures. Main parameters reviewed were demographic data, predisposing risk factors, treatment modalities, microbiological profile, visual outcomes, and graft survival. Cases were included in the study if the attending doctor documented a diagnosis of microbial keratitis and advised admission to the hospital. Of the baseline cohort, 12 were lost to follow-up. Missing data were noted for 13 eyes.
At initial presentation, clinical history, visual acuity, and slit-lamp biomicroscopic findings were recorded. The corneal scrapings from all cases were subjected to microbiological evaluation and submitted for Gram stain, Blankophor preparation, chocolate agar, Sabourad’s dextrose agar, and thioglycollate broth. A swab was taken for the detection of herpesvirus by polymerase chain reaction. Intensive antimicrobial therapy was started after corneal scraping in the form of hourly ofloxacin 0.3% eye drops. Cases with presumed fungal keratitis were initially started on topical natamycin 5% or amphotericin 0.15% eye drops. Corticosteroid eye drops were stopped in all cases at the time of presentation. Treatment was modified as indicated by culture results, sensitivity pattern, and clinical response. In the event of failure of medical treatment, surgical intervention was performed in the form of tarsorrhaphy/botox, corneal gluing, or graft resuturing. An emergency corneal transplantation was performed in the event of nonresolving microbial keratitis. An elective corneal transplantation was performed whenever indicated, after complete resolution of infection.
Data were stratified by the presence and absence of sutures in infected corneal grafts. Odds ratios (OR) and 95% confidence intervals (CI) were reported for the association between indication (failed graft, corneal scar, keratoconus, pseudophakic bullous keratopathy), risk factors (steroid, ocular surface disease, regraft, glaucoma, herpes simplex keratitis), and surgical intervention (botox, therapeutic keratoplasty, repeat keratoplasty, glue) with infected corneal grafts in the presence and absence of sutures. Indication, risk factors, and surgical intervention exposures were analyzed for the entire cohort. Two models were created; Model 1 adjusted for age and gender and Model 2 further adjusted for ocular surface disease, regraft, glaucoma, and herpes simplex keratitis. All P values were 2-sided and based on χ 2 (for categorical variables) and t test (for continuous variables). Stata IC version 10.0 (Stata Corp, College Station, Texas, USA) was used for all analyses.
Of the 122 episodes of microbial keratitis following penetrating keratoplasty in 101 patients, 71 (58.2%) had sutures present and 51 (41.8%) had sutures absent ( Table 1 ). Mean recipient age was higher in the “sutures absent” group, 69 ± 16.7 years (range: 29-94 years), as compared to the “sutures present” group, 66.4 ± 19.0 years (range: 22-94 years). There were more women than men in both groups, with no significant intergroup differences. Mean donor age was significantly higher in the “sutures present” group ( P = .02), whereas the endothelial cell count was comparable between both groups ( Table 1 ). Pseudophakic bullous keratopathy was the most common surgical indication for corneal grafting in both groups, followed by comparative rates of failed previous grafts, keratoconus, and corneal scars. After adjusting for ocular surface disease, regraft, glaucoma, and herpes simplex keratitis, keratoconus conferred increased odds (OR 1.46, 95% CI 0.13-16.6, P = .76) of developing microbial keratitis in corneal grafts with sutures present compared to sutures absent. However, the strength of this association was weak ( Table 2 ). Nine out of 71 eyes (13%) were grafted for rheumatoid melt in the “sutures present” group only ( P = .01) ( Table 1 ). At the time of presentation, mean visual acuity was poorer in the “sutures absent” group (logMAR 2.32 ± 0.84) compared to the “sutures present” group (logMAR 1.98 ± 0.92) ( P = 0.62).
|Suture Present N = 71 (%)||Suture Absent N = 51 (%)||P-Value|
|Age (recipient)||66.4 ± 19.0||69.0 ± 16.7||0.43|
|Age (donor)||57.4 ± 27.8||44.1 ± 32.2||0.02|
|Donor endothelial cell density (cells/mm 2 )||2521 ± 351||2571 ± 363||0.13|
|Indications for Corneal Transplantation|
|Pseudophakic bullous keratopathy||15 (21)||12 (24)||0.92|
|Failed previous graft||11 (15)||9 (18)||0.75|
|Keratoconus||7 (10)||4 (8)||0.69|
|Corneal scars||4 (6)||3 (6)||0.94|
|Rheumatoid melt||9 (13)||0||0.01|
|Indication for Surgery||Age-Gender-Adjusted OR (95% CI) a||P||Multivariate-Adjusted OR (95% CI) b||P|
|Failed graft||0.87 (0.14-5.40)||0.89||1.04 (0.16-6.89)||0.96|
|Keratoconus||1.31 (0.14-12.2)||0.82||1.46 (0.13-16.6)||0.76|
|Pseudophakic bullous keratopathy||0.94 (0.16-5.41)||0.94||0.75 (0.12-4.71)||0.76|
|Present||0.53 (0.23-1.24)||0.15||0.46 (0.19-1.12)||0.09|
|Ocular surface disease c|
|Present||1.02 (0.43-2.47)||0.97||1.36 (0.54-3.42)||0.52|
|Present||0.51 (0.19-1.35)||0.18||0.30 (0.09-0.99)||0.05|
|Present||2.01 (0.59-6.91)||0.27||4.74 (1.00-22.3)||0.05|
|Herpes simplex keratitis f|
|Present||0.61 (0.22-1.71)||0.35||0.54 (0.18-1.62)||0.27|
At the time of presentation, 76% of patients in the “sutures present” group and 64% of patients in the “sutures absent” group were using topical corticosteroid eye drops. Within the “sutures present” group, ocular surface disorders (44%) and corneal regrafting (28%) were the other associated factors ( Table 3 ). Ocular surface disorder was inclusive of dry eyes/poor tear film and persistent epithelial defects that resulted in compromised ocular surface integrity. In the “sutures absent” group, presence of glaucoma (48%) and corneal regrafting (48%) were more prevalent than ocular surface disease (37%). After adjusting for confounders, ocular surface disease (OR 1.36, 95% CI 0.54-3.42, P = .52) and glaucoma (OR 4.74, 95% CI 1.00-22.3, P = .05) conferred increased odds of developing microbial keratitis in corneal grafts with sutures present compared to sutures absent. However, the strength of these associations was weak ( Table 2 ).
|Associated Factors||Suture Present N = 71 (%)||Suture Absent N = 51 (%)||P-Value|
|Topical corticosteroid use||54 (76)||33 (64)||0.17|
|Ocular surface disorders||17 (44)||19 (37)||0.17|
|Regraft||11 (28)||13 (48)||0.18|
|Glaucoma||8 (21)||13 (48)||0.19|
|Previous herpes simplex keratitis||8 (21)||3 (11)||0.31|
History of previous herpetic keratitis was higher, but not significant, in grafts with sutures present (21%) compared to sutures absent (11%) ( Table 3 ). None of the patients had received fortified antibiotics prior to admission.
Examination of corneal smears revealed gram-positive organisms in 50 of 122 (41%) episodes. Gram-positive cocci were the commonest isolate overall, 48% in the “sutures present” and 14% in the “sutures absent” group. No fungal isolates were noted. Microbiological culture results were positive in 99 of 122 episodes (81%), 77% of which were gram-positive and 23% gram-negative. Culture yields of gram-positive organisms were higher in the “sutures present” than the “sutures absent” group ( Table 4 ). Of the gram-positive organisms, the commonest organism isolated was Staphylococcus aureus , higher with sutures present (51%) than with sutures absent (40%). Staphylococcus epidermidis , Streptococcus pneumoniae , and other Streptococcus spp. occurred with relatively similar frequency in both groups ( Table 4 ). Of the gram-negative organisms, Moraxella sp . ( P = .001) was the most significant organism isolated within the “sutures absent” group ( Table 4 ). Polymicrobial infections were present in 18 eyes, 61% in the “sutures present” and 39% in the “sutures absent” group, with no intergroup differences.
|Organism (Gram Positive)||Suture Present N = 51 (%)||Suture Absent N = 25 (%)||P||Organism (Gram Negative)||Suture Present N = 12 (%)||Suture Absent N = 11 (%)||P|
|Staphylococcus aureus||26 (51)||10 (40)||0.120||Moraxella sp.||1 (8)||9 (82)||0.001|
|Staphylococcus epidermidis||7 (14)||1 (4)||0.110||Pseudomonas aeruginosa||1 (8)||2 (18)||0.480|
|Other Streptococcus sp.||7 (14)||3 (12)||0.810||Stenotrophomonas maltophilia||3 (25)||0||0.0456|
|Coagulase negative Staphylococci||3 (6)||4 (16)||0.220||Klebsiella sp.||2 (17)||0||0.120|
|Streptococcus pneumoniae||2 (4)||1 (4)||0.490||Serratia sp.||2 (17)||0||0.120|
|Corynebacterium sp.||4 (8)||4 (16)||0.340||Proteus mirabilis||2 (17)||0||0.120|
|Bacillus sp.||1 (2)||0||0.310||Pseudomonas fluorescens||1 (8)||0||0.310|
|Enterococcus faecalis||1 (2)||1 (4)||0.650|
|Propionibacterium sp.||0||1 (4)||0.320|