Authors: Anitha Venugopal, Pooja Andhare, Anusha Vemula Rao, Rameshkumar Gunasekaran, Meenakshi Ravindran
Categories: Original Article, Culture-positive keratitis, culture-negative keratitis, microbial keratitis
Source: Indian Journal of Ophthalmology
To study the risk factors, clinical features, and treatment outcomes of patients with culture-negative keratitis (CNK).
A retrospective data review of 933 patients with CNK was performed from January 2018 to December 2020. The variables such as the history of injury, visual acuity, slit-lamp findings with measurements of size and depth of ulcer, microbiological evaluation, duct patency, blood glucose levels, and treatment were considered, and clinical outcome was analyzed.
Of the 933 patients with CNK, 763 (81.8%) were medically managed, with a mean treatment duration of 2.08 ± 1.7 weeks. Among them, 622 (66.7%) were both smear and culture-negative, and 311 (33.3%) showed only smear positivity. Smear-positive patients showed a positive correlation with the history of injury. A higher incidence of fungal growth on repeat culture was observed. Surgical interventions were done only in 18.2% of the patients; the rest were treated with topical medications alone.
High clinical suspicion, differentiation of causative organisms based on clinical findings, and initiating empirical therapy with broad-spectrum antibiotics and antifungals improve the ultimate prognosis in patients with CNK, even though a standard protocol for empirical medical treatment may differ among institutions and surgeons based on their clinical experience and geographical variations.
Keywords: Culture-positive keratitis, culture-negative keratitis, microbial keratitis
Microbial keratitis (MK) is one of the most common causes of preventable blindness globally. In developing countries like India, the annual incidence of MK may exceed 2 million cases per year.[1,2] It also affects the economic burden of the healthcare system in developed countries; 175 million dollars are spent annually on MK in the United States.[1,3] Timely diagnosis with microbiological examination and culture is the key to prompt and successful management of MK.[4] The diagnosis is intrinsically made by corneal scraping with smear examination, culture, and antibiotic sensitivity tests. Even though the culture is considered the gold standard approach, it has a positive rate of about 40–60%.[4,5,6] The predisposing factors for culture negativity (CN) include delayed presentations of patients to the clinic, average being 19–20 days, depth of the infiltrates, usage of multiple self-topical medications before first corneal scraping, and infection with rare or fastidious pathogens with indistinguishable clinical features such as Acanthamoeba, atypical mycobacteria, and fungi.[6,7,8] Hence, in such scenarios, corneal biopsy to reach the deeper infiltrates and performing other molecular diagnostic modalities such as polymerase chain reaction (PCR), in-vivo confocal microscopy (IVCM), immunohistochemistry (IHC), and histopathological examination (HPE) of the tissue after therapeutic keratoplasty would aid in appropriate diagnosis depending upon the availability of resources.[9] The history and clinical features often elucidate the possible underlying causative organisms.[10] This might need clinical expertise. The management in such cases is primarily an empirical antimicrobial therapy, close follow-up, and its clinical response, which aids in determining the prognosis of the disease.[9] Delay initiating appropriate medical treatment of infective keratitis might lead to the progression of the ulcer, perforation, and even visually blinding conditions like endophthalmitis.[8] The clinical course of CNK is often overlooked, and its management outcomes should be more frequently explored in the literature. This study aims to comprehensively review the patients with CNK, its clinical course, and visual and treatment outcomes.
A retrospective review of 933 patients with culture-negative microbial keratitis presenting at a tertiary eye care center in Tamil Nadu, South India, was assessed for over 3 years, from January 2018 to December 2020. Demographic data, mode of injury, duration of symptoms, visual acuity, infiltrate size (largest diameter in mm), smear and culture report, course of treatment, time of resolution of symptoms, surgical interventions (therapeutic penetrating keratoplasty), duration of follow-up, and outcome (corneal scar, clear or failed graft) of all the study patients were collected from electronic medical records and entered in the Microsoft Excel for analysis. The study was conducted as per the tenets of the Declaration of Helsinki and after obtaining institutional ethical committee approval (IEC/R/CS/2022/008).
Sample collection and Patients presenting to the hospital with clinical features suggestive of MK with no growth detected on the first corneal scraping culture were included in the study. Patients with presumed viral keratitis, neurotropic ulcers, interstitial keratitis, atheromatous ulcers, and any ulcer associated with autoimmune conditions were excluded. After slit-lamp examination and documentation of clinical findings such as size, shape, depth, and extent of the ulcer, all the patients were subjected to corneal scraping. Corneal scrapings were collected under topical anesthesia using 0.5% proparacaine. Scraping specimens included two scrapings for smear examination (one each for Gram stain and 10% potassium hydroxide wet mount), followed by a subsequent scraping for culture on potato dextrose agar, 5% sheep blood agar, chocolate agar, and non-nutrient agar plates. Culture positivity (CP) was defined as the growth of at least three confluent colonies on one solid medium within 2 weeks of incubation. The absence of growth on any culture plates, even after 2 weeks, was considered culture negativity. A repeat microbiological examination was done in 339 (36.3%) patients with nonhealing keratitis.
Treatment for patients was started based on the clinical appearance of the ulcer and initial smear reports. Later, patients with culture positivity were treated according to the organism identified. All patients with culture-negative reports and who were not responding to treatment and showing signs of worsening, repeat cultures, and smears are performed by either corneal biopsy or scraping after 72 hours of treatment after discontinuing the medical therapy for 12 hours. Table 1 depicts the treatment approach adopted in our institution based on the clinical appearance of the ulcer and literature reference.[11,12,13,14,15,16,17,18]
The daily examination was done until the ulcer improved, gradually reducing the frequency of drops and follow-up over 2 weeks. Continue using eye drops four times daily for 2 more weeks after healing the ulcer. Cases unresponsive to medical therapy with worsening of symptoms for >3 weeks, an increase in the size of the ulcer/exudates/hypopyon, limbal involvement, and impending perforation [Fig. 1] were managed surgically with therapeutic penetrating keratoplasty (TPK). Treatment success was defined as complete healing of the ulcer with scar formation within 12 weeks of presentation with medical therapy. Progressive corneal ulceration, perforation, and the need for urgent therapeutic keratoplasty constituted treatment failure.
Figure 1 Slit-lamp image showing clinical worsening of infiltrate with signs of impending perforation
Statistical analysis was performed using STATA software version 14.2 (StataCorp, USA). Descriptive qualitative variables were reported using frequencies (n) and percentages (%). Quantitative variables were declared using the mean with standard deviation (SD) and median with interquartile range (IQR). The Chi-square test was used for categorical variables, and the Wilcoxon rank sum test was used for continuous variables to compare the variables between the trauma group and the no-trauma group. Snellen’s visual acuity values were converted into a logarithm of the minimal angle of resolution (log MAR) for statistical analysis. The Wilcoxon sign rank test compared postoperative visual acuity with baseline data. The treatment duration was compared with different infiltrate sizes, and the significance was analyzed using the Kruskal–Wallis test. A Spearman rank correlation was used to examine the correlation between infiltrate size and treatment duration. Two-tailed hypothesis testing was used, and the results with P < 0.05 were considered statistically significant.
A total of 2347 patients of microbial keratitis, excluding clinical viral keratitis, were registered during the study period from January 2018 to December 2020 [Fig. 2]; of these 2347 patients, 933 culture-negative keratitis (CNK) (39.7%) cases were included in the study. The duration of the mean follow-up was 15.63 (19.0) months. The demographic and clinical findings of the study patients are summarized in Table 2. The median (IQR) age at the presentation was 41 years (range 7–66 years). There was male preponderance in the study population as 678 (72.7%) were male patients, and only 255 (27.3%) were female. A history of ocular trauma with vegetative matter (wooden stick/thorn) was noted in 340 (36.4%) patients. Use of prior topical ocular medications was recorded in 667 eyes (71,4%) before presentation to the hospital.
Figure 2 Breakup of total number of study patients with CNK based on the culture and smear results
During the initial visit, 157 (16.8%) patients had an infiltrate size of >5 mm, 323 (34.6%) had an infiltrate size between 3 and 5 mm, and 453 (48.6%) had an infiltrate size less than 3 mm. Emphatically, the duration of treatment was prolonged with an increase in infiltrate size [rho = 0.55, P < 0.001, Fig. 3].
Figure 3 Correlation of infiltrate size and duration of treatment using a scatter diagram
All patients were started on empirical therapy based on the clinical findings and positive smear report. Treatment was later modified depending on the treatment response. Out of the 933 patients with CNK, 763 (81.8%) were medically managed with a mean treatment duration of 2.08 (±1.7) weeks. Among them, 622 (66.7%) were both smear and culture-negative, and 311 (33.3%) showed only smear positivity. The smear-positive patients were studied further regarding their history of trauma [Table 3]. The smear-positive patients following trauma showed a higher incidence of fungal hyphae in 81 (23.8%) and Gram-negative bacilli in 32 (9.4%) patients. Those patients without a history of trauma had a higher incidence of Gram-positive cocci on smear, amounting to 99 (16.6%) cases. In contrast, fungal hyphae and Gram-negative bacilli were noted in 49 (8.3%) patients each.
A repeat microbiological examination was done in 339 (36.3%) patients with nonhealing keratitis. Of these, 169 (50%) patients showed positive culture reports on rescraping. In 104 (61.5%) patients, the culture showed fungus growth; 6 (3.5%) patients were positive for Acanthamoeba, 58 (34.3%) had mixed growth (bacterial and fungal), and one (0.05%) patient was positive for Pythium sp. The treatment was modified according to the repeat culture and sensitivity report in 35 (10.3%) patients, and the same treatment was continued in the rest of the cases as the clinical judgement was correlating with the culture positivity.
About 170 (18.2%) patients underwent therapeutic keratoplasty (TPK) for various indications, such as a perforated ulcer or impending perforation, progressive ulcer with involvement of limbus, and nonhealing ulcer for more than 3 weeks of medical therapy. After TPK, the infected corneal button was sent for culture and sensitivity testing. Among them, 114 (67.1%) eyes were positive for fungus (unidentified hyaline fungus), 3 (1.8%) were positive for Pythium sp., and 53 (31.2%) were once again culture-negative. Post TPK, the graft remained clear for at least 6 months in 61 (35.9%) cases, failed, and remained opaque with quiet eyes in 53 (31.2%). The remaining 37 (21.8%) had failed opaque grafts with secondary glaucoma and were on long-term antiglaucoma medications [Fig. 4]. A total of 19 (11.1%) cases underwent regraft for graft melt or graft reinfection, out of which 3 (15.7%) eyes with pythium keratitis had graft melted even after resurgery and went for phthisis bulbi, and the remaining 16 (84.2%) eyes were lost to follow-up. Collectively, 286 (30.6%) cases out of a total of 933 CNK cases were found to be culture-positive on re-evaluation (169 rescraping cases and 117 button cultures post TPK) [Fig. 5]. Post-treatment visual acuity (VA) was compared with baseline VA at presentation using the Wilcoxon sign rank test [Fig. 6]. There was a statistically significant difference in the post-treatment visual acuity (median VA = 0.18, i.e., 6/9) compared to baseline VA (median VA = 0.30, i.e., 6/12, P < 0.001).
Figure 4 Final clinical outcomes following TPK
Figure 5 Results of repeat corneal scraping results of our study patients with nonresolving keratitis
Figure 6 Post-treatment median BCVA (log MAR) was plotted using bar graph against pretreatment
Infective keratitis (IK) is the major culprit behind corneal blindness in our country. The incidence of IK varies from 5.3 per one lakh population in USA to 113 per one lakh population in South India.[8,16] From these statistics, it is clear that developing countries like India are under significantly more burden of IK than developed countries.
Microbiological smear and culture evaluation are necessary for the appropriate management of IK since the prognosis depends on the accuracy of treatment based on the identification of the infective organism. But, the microbiological culture sensitivity is only 60–70%.[6,9] Due to various reasons such as inadequate sample collection, growing a minuscule sample on a culture plate, difficulty in collecting samples from uncooperative patients, pretreatment with antimicrobial agents, and contamination of the culture plates might lead to the technical difficulty in identifying the causative organisms and a few organisms like fungi, Acanthamoeba, and Microsporidia have the propensity for penetrating deep into the cornea, which requires deep corneal scrapings or corneal biopsy for an adequate sample.[19,20,21,22] Gram-negative organisms are more difficult to identify on Gram staining due to their lighter color.[10,11,23] The sensitivity of Gram staining is only 36–40% for bacteria, 42.1% for fungus, and 91% for Acanthamoeba with KOH wet mount.[19,20,21,22] The overall specificity of staining methods in diagnosing MK is only between 29 and 62%.[19,22,23] Prior use of antimicrobial agents, mechanical damage to the cell wall, an insufficient sample, poor staining techniques, failure to examine the whole slide, excessive heat fixation, and contamination of the stain or culture plates used are some of the laboratory-associated factors leading to adverse sensitivity reports. Hence, in such scenarios, experienced clinicians prefer treating MK based on the clinical picture alone.[7,8] Grant Carlisle et al.[22] noted very high CNR in 69% of his study subjects and attributed it to prior topical antibiotics. Similarly, among our patients, 71.4% used prior topical drops. In such instances, repeat culture helped us to modify the empirical antimicrobial therapy according to the infective agent identified.[22] We had repeated cultures in 36.3% of the patients; among them, CPR was 50%, which was quite significant. The appropriate therapy was changed depending on the repeat culture reports, which ranged from 5 to 15%, comparable to our observations (10.3%).[24] Thirty-five patients (10.3%) is the subset of the 169 who required treatment modification compared to the 134 patients who did not require treatment modification as clinical suspicion matched with the positive culture report on repeat scraping.
The culture positivity rate also depends on the size of the infiltrate and the amount of scraping material available. Morlet et al.[24] found that ulcers larger than 5 mm had 2.68 times the chance of a positive culture result, but we observed only 16.8% of our patients with infiltrate size >5 mm, which was probably the reason for the culture-negative results. Higher incidences of fungus were observed with repeat cultures (61.5%). Hence, it is a requisite to perform repeat cultures in patients with nonhealing keratitis promptly.
Many studies have documented the association of a higher incidence of fungal keratitis with trauma.[18,25,26] Likewise, we also identified fungus in 71.6% of smears of the patients (113) with a positive history of trauma. Henceforth, detailed history taking relating to trauma is mandatory in patients with CNK.
TPK remains the ultimate option for clinically uncontrolled infectious keratitis, with the primary goal being eradicating the infection process and preserving globe integrity.[27] Song et al.[28] reported a high incidence of surgical intervention (74%) in patients living in rural areas exposed to agricultural activities with a high fungal infection rate (48.7%). Even though we had a higher incidence of fungal keratitis in our study patients, the surgical intervention was less (18.2%), likely due to the early addition of antifungals as the empirical therapy in our patients. Hence, it becomes pivotal to add antifungals to the empirical treatment at the time of presentation in the case of CNK to achieve a good prognosis. The post-treatment improvement in visual acuity was statistically significant in our patients, which might be due to the small-sized infiltrates, around 48.5% (n = 453). This might be one of the reasons behind the high culture negativity rate in our study. PCR and confocal microscopy are an attractive diagnostic modality for microbial keratitis; it has higher specificity and sensitivity. Moreover, the procedure is complicated, needs technical support, and is expensive, impeding its use in developing countries. Performing molecular methods, if feasible to improve the diagnosis in cases of CNK, would aid in improving the culture positivity rate and thereby be helpful in initiating an appropriate therapy for the microbe identified[29] since the accurate diagnosis of the organisms and focused treatment on the target microbes carries good prognosis.
Ours was a retrospective observational study; the authors have used the microbial smear and culture methods as a diagnostic tool; the advanced diagnostic modalities like PCR/confocal microscopy were not considered, and we have not included the description of species for studying the results in our study. These are some of the limitations.
CNK is a challenging scenario when further molecular diagnostic methods are not available. As a clinician, it is mandatory to achieve high culture yield by doing repeat corneal scrapings and corneal biopsies in deep-seated keratitis with the aim to identify the inciting microbes in these eyes. Furthermore, clinical acumen of identifying the type of microbial keratitis along with microbiological reports benefits management of MK. Also, aiming for newer diagnostic modalities whenever it is feasible would help in accurate therapy and help in achieving good prognosis in these cases.
Nil.
There are no conflicts of interest.