Article Text
PDF
Abstract
Purpose To determine the causes of culture negativity and compare clinical outcomes in patients with culture-negative (CN) keratitis versus culture-positive (CP) keratitis.
Methods Medical and microbiology records of 60 patients, who presented at the L V Prasad Eye Institute, Bhubaneswar, between January 2009 and December 2010, and who were clinically diagnosed to have microbial keratitis were retrospectively reviewed. Detailed ocular examination was performed before corneal scraping was sent for microbiological workup. The data collected from each record included age, gender, predisposing factors (ocular and systemic), clinical presentation, management and outcome of treatment.
Results The duration of symptoms and prior use of topical medication was significantly higher in the CN group (p=0.005, p=0.007, respectively). Infiltrate size (p=0.42) and history of ocular trauma (p=0.14) had no association. While patients in the CP group underwent 9 (30%) minor major surgical procedures and 14 (46.7%) major surgical procedures, patients in the CN group underwent 11 (36.7%) minor surgical procedures and 1 (3.3%) major surgical procedure. Success in treatment was achieved in 27 (90%) and 25 (83.3%) patients in CP and CN groups, respectively. The mean visual acuity (logMAR) at presentation in the CP group was 2.53 which improved to 1.83 at the last follow-up (p=0.0001). Similarly, the mean visual acuity (logMAR) at presentation in the CN group was 2.57 which marginally improved to 2.34 at the last follow-up (p=0.03).
Conclusions CN keratitis is associated with long duration of topical medication. The number of major surgical interventions in CN keratitis is significantly less compared with CP keratitis. However, the final outcome of treatment is similar in both CP and CN keratitis.
- Infection
- Cornea
- Microbiology
Statistics from Altmetric.com
Introduction
Microbial keratitis is a major ocular disease which can potentially threaten visual acuity. It is defined as loss of the corneal epithelium, with underlying stromal infiltration and suppuration associated with signs of inflammation with or without hypopyon.1 ,2 It is considered as one of the leading causes of visual loss in developing countries.3 Culture technique is considered to be a gold standard for identification of the causative agents.4 Culture isolation helps in deciding specific drug therapy, provides antibiotic sensitivity pattern, avoids emergence of drug resistance, decreases time for resolution and provides prognosis of the case depending on the microbe isolated. American Academy of Ophthalmology recommends the use of blood agar, chocolate agar, thioglycollate broth (THIO) and sabouraud dextrose agar (SDA) as standard media to be inoculated for all cases of microbial keratitis.5 Our institute protocol, apart from direct microscopic examination of corneal scrapings stained with potassium hydroxide with calcofluor white (KOH+CFW) and Gram stain, consists of inclusion of blood agar, chocolate agar, SDA, brain heart infusion (BHI) broth, THIO, Robertson's cooked meat broth (RCMB) and non-nutrient agar (NNA) for the investigation of non-viral corneal ulcers. The culture positivity rate of microbial keratitis at our institute is 60.4%.6 Similarly, studies published from other parts of India show a culture positivity rate of 67.8–70.4%.4 ,7
The most common cause of microbial keratitis in the above studies was fungus that was isolated in 34.5–62.7% of cases followed by bacteria that were isolated in 22.7–51.9% of cases.4 ,6 ,7 Clinical outcomes of culture-positive (CP) keratitis (bacterial and fungal keratitis) have been well studied in the literature.7–11 However, there exists a gap in knowledge regarding culture-negative (CN) keratitis, which constitutes around 30–40% of all infectious keratitis.
This study retrospectively analyses and compares the data of CN cases with CP cases regarding their demographics, cause for culture negativity, clinical course and final outcome.
Materials and methods
Clinical and microbiology records of 60 patients, who presented at the L V Prasad Eye Institute, Bhubaneswar, between January 2009 and December 2010, were retrospectively reviewed. The retrospective evaluation of clinical and microbiology records was approved by the ethics committee of the institute (Hyderabad Eye Research Foundation; Ethics Ref No. LEC 04-13-38; LVPEI-B-35-2013). Patients presenting with loss of the corneal epithelium, with underlying stromal infiltration and suppuration associated with signs of inflammation with or without hypopyon, were clinically diagnosed as microbial keratitis and were included in the study. Patients with clinical diagnosis of epithelial dendrites, neurotrophic keratopathy, exposure keratopathy, immune-mediated peripheral ulcerative keratitis, immune stromal keratitis and disciform keratitis were excluded from the study. After a comprehensive ophthalmic work-up, corneal scrapings were subjected to corneal smear microscopy (stained with KOH+CFW and Gram stain) and culture examination. Type of media and the sequence of inoculation included blood agar, chocolate agar, SDA, BHI, THIO, RCMB and NNA. Corneal scraping was done using #15 surgical blade on a Bard-Parker handle under topical anaesthesia (0.5% proparacaine hydrochloride). Care was taken to collect the scraping directly from the corneal infiltrate without coming in contact with the conjunctiva and lids.
All media were incubated at 36°C±1°C except chocolate agar (5% CO2 at 36°C±1°C) and SDA (27°C). Positivity of growth in different media and results of smear examination were noted. The significance of the culture was based on either (criteria-1) confluent growth (10 or more colonies) on the inoculum of one solid medium, and/or (criteria-2) growth in more than one medium and/or (criteria-3) growth in one medium consistent with direct smear result. All media were observed daily for growth for 2 weeks. Time taken for growth was noted. Identification of bacterial colonies was done after consideration of colony characteristics, Gram reaction, morphology and results of biochemical tests. Fungal species were identified by observing the rate of growth, colour, consistency, texture and characteristic microscopic features. CP keratitis was defined as patient having clinical features of microbial keratitis and fulfilling at least one criterion of significant growth. CN keratitis was defined as patient having clinical features of microbial keratitis but did not reveal any micro-organism either in smear or culture. Patients diagnosed as CP microbial keratitis was treated based on smear results initially. Fungal keratitis was started on topical natamycin (5%) one hourly with atropine sulfate (1%) three times a day. Gram-positive microbial keratitis was started on either topical fortified cefazolin (5%) or gatifloxacin (0.3%) and gram-negative microbial keratitis was started on topical ciprofloxacin (0.3%) one-hourly along with topical atropine sulfate (1%) three times a day. The treatment was then modified based on culture and sensitivity report. CN keratitis was treated with topical gatifloxacin (0.3%) one hourly and topical atropine sulfate (1%) three times a day. The duration of therapy and surgical intervention was done as per the clinical judgement.
Data of 30 patients were collected in each group ( CP, CN keratitis). The data included demographic profile (age, gender, occupation), duration of symptoms, history of previous trauma, surgery or contact lens use, type and duration of previous medication used, systemic disease if any, visual acuity (first and last visits), infiltrate size (mm2), smear and culture report, medication received at our centre, time for resolution of infiltrate on medical treatment, minor surgical interventions (tissue adhesive with bandage contact lens (TA+BCL), superficial keratectomy), major surgical interventions (therapeutic penetrating keratoplasty, evisceration, pars plana vitrectomy (PPV), anterior chamber (AC) wash) and duration of follow-up and final outcome (stable corneal scar, clear graft, failed graft, phthisis bulbi). The outcome of treatment was deemed as ‘treatment success’ in case there was complete resolution of infiltrate with stable ocular surface and tectonically stable globe. Incomplete resolution of infiltrate, phthisis bulbi and evisceration was deemed ‘treatment failure’.
The data were analysed using Student's t test. Preoperative and postoperative visual acuities were compared using paired t-test. Qualitative data were analysed using Fischer's exact test. The values are given as mean±SD in parametric data or median+range in non-parametric data. A p value <0.05 is considered as statistically significant.
Results
A total of 60 patients were analysed which included 30 CP keratitis patients and 30 CN keratitis patients. The mean age was 47 and 45.5 years in CP and CN groups, respectively (p=0.31). Male to female ratio (2:1) was the same in both the groups. Sixteen patients in the CP group (16/30, 53.3%) and 13 in the CN group had agricultural occupation. The history of ocular trauma was present in 7/30 (23.3%) in the CP group and 6/30 (20%) in the CN group (3.3%) (p=0.14). None of the patients in the CP group and only one patient in the CN group had systemic risk factor (ie, diabetes mellitus, p=1.0). The mean duration of symptoms was 14.4±15.1 (range: 3–45) days in the CP group and 27.6±20.3 (range: 5–90) days in the CN group (p=0.005).
All patients underwent Gram and KOH+CFW smear examination. Fungus was seen in 24/30 (80%) in KOH+CFW stain and in 19/30 (63.3%) in Gram stain in the CP group. Gram-positive cocci were seen in 3/30 (10%) cases in Gram stain in the CP group. The CN group showed no organism in either Gram or KOH+CFW stain (table 1). Based on culture reports and criteria of significance, 22/30 (73.33%) were diagnosed as pure fungal keratitis, 7/30 (23.33%) were diagnosed as pure bacterial keratitis and 1/30 (3.3%) was diagnosed as mixed (bacterial+fungal) keratitis (table 2). The most common fungus isolated was unidentified fungus (14/22; 63.63%). The most common bacteria isolated were Pseudomonas aeruginosa (4/7; 57.14%). No growth was observed in the CN group.
Smear results of cases having culture-positive keratitis (n=30)
Culture results of cases having culture-positive keratitis (n=30)
Twenty-one out of 30 patients (70%) in the CP group and 23/30 patients in the CN group were using topical medications prior to visit to our hospital (p=0.77). Eleven out of 21 patients (52.4%) in the CP group and 10/23 in the CN group were on combination therapy of topical antibiotics (gatifloxacin (0.3%), ofloxacin (0.3%) or moxifloxacin (0.5%)) and antifungal (natamycin (5%)) (p=0.57). Two patients in the CP group (6.7%) were on topical steroid treatment. The mean duration of prior topical medications were 9.5±8.6 (range: 3–35) days in the CP group and 16.6±11.1 (range: 5–60) days in the CN group (p=0.007). The frequency of instillation ranged between 3 and 24 times a day, most of them used it in suboptimal dosing frequency. In vitro susceptibility was done in eight cases of bacterial keratitis in the CP group (seven pure bacterial and one mixed keratitis), whose details are mentioned in table 3.
In vitro susceptibility, antibiotic used at presentation and antibiotic prescribed at L V Prasad Eye Institute
The dimension of the infiltrate (largest dimension×smallest dimension in mm2) was 26.7 (range: 1.5–64.8) mm2 in the CP group and 22.4 (range: 1–64) mm2 in the CN group (p=0.42). The CP group underwent 9 (30%) minor surgical procedures and 14 (46.7%) major surgical procedures. Nine patients (30%) underwent minor surgical procedures which included eight TA+BCL (26.7%) and one (0.33%) superficial keratotomy. Eleven (36.7%) patients required therapeutic keratoplasty, two (6.7%) required AC wash and one (3.3%) required PPV for endophthalmitis. Eight patients (26.7%) underwent more than one procedure. The CN group underwent 11 (36.7%) minor procedures and 1 (3.3%) major procedure. It included 11 TA+BCL as minor procedure and 1 (3.3%) patient underwent major surgical procedure, that is, therapeutic penetrating keratoplasty for large central perforation (table 4).
Comparison of surgical interventions in culture-positive (CP) keratitis and culture-negative (CN) keratitis (n=30)
The mean duration between the date of presentation and therapeutic keratoplasty was 10±5.8 (range: 4–20) days in the CP group. Ten out of 11 patients (90.9%) of therapeutic keratoplasty had fungus aetiology and one had bacterial aetiology (table 5). The indication of therapeutic keratoplasty included corneal perforation (>2 mm) and large infiltrate (>6 mm) not responding to medical therapy. Three out of 11 patients (27.3%) maintained a clear graft till the last follow-up; however, 7/11 patients had graft failure either due to graft rejection or late endothelial failure. Although these six patients (54.5%) had graft failure, all of them maintained the globe integrity till the last follow-up. One patient (9.1%) had phthisis bulbi in spite of undergoing therapeutic keratoplasty and PPV because of fulminant endophthalmitis. The number of patients undergoing major surgical intervention was statistically higher in the CP group compared with the CN group (p=0.0002).
Details of therapeutic keratoplasty in 11 culture-positive keratitis patients and one culture-negative (CN) keratitis patient
The final outcome was assessed on the basis of the number of patients achieving the goal of complete elimination of infection and tectonically stable globe at the last follow-up visit. Patients achieving the goal were deemed as ‘treatment success’; however, patients having non-resolving keratitis and loss of globe integrity were deemed as ‘treatment failure’. Twenty seven out of 30 patients (90%) in the CP group achieved the goal of ‘treatment success’. Seventeen patients (56.7%) achieved this goal on medical treatment and/or minor procedure; however, 10 patients (33.3%) required therapeutic keratoplasty as major surgical intervention to achieve this goal. Three patients (10%) had treatment failure in the CP group. Two patients (6.7%) had phthisis bulbi due to advanced stage of disease at presentation and one had non-resolving keratitis till the last follow-up. Twenty-five out of 30 patients (83.3%) in the CN group achieved the goal of ‘treatment success’. Five patients (16.7%) had ‘treatment failure’ in the CN group. Four (13.3%) had non-resolving keratitis till the last follow-up and one had phthisis bulbi. The mean duration of medical treatment for corneal scar formation was 27±19.8 (range: 5–90) days in the CP group and 31±18.4 (range: 4–67) days in the CN group (p=0.44).
The mean visual acuity (logMAR) at presentation for CP keratitis was 2.53 which improved to 1.83 at the last follow-up (p=0.0001). The mean visual acuity (logMAR) at presentation for CN keratitis was 2.57 which marginally improved to 2.34 at the last follow-up (p=0.03). The final visual acuity at the last follow-up was statistically not significant between CP and CN groups (p=0.07).
Discussion
The identification of microbial agent causing infectious keratitis is completely based on laboratory evaluation of smears and cultures. However, the sensitivity of cultures for identification of the microbes is 60–70%, suggesting that the causative microbe remains undiagnosed in one-third of all infectious keratitis. The cause for culture negativity in microbial keratitis has been described in the past, among which the use of prior topical antibiotic therapy remains the most important one. A study done at Dutch Hospital has shown that the culture positivity rate is statistically lower (41.3%) in patients pretreated with antibiotics compared with those who did not receive it (68.7%).12 However, in the present study, both CP and CN keratitis patients were on prior topical medication. We found that the duration of symptoms and prior use of topical medication was significantly more in CN keratitis cases. McDonnell et al13 found that 48.1% of patients (38/79) referred with corneal ulcers were on antibiotic therapy without having obtained a prior culture. Ours being a tertiary care hospital, most of the patients were already on topical medication from the places they were referred. Forty-four out of 60 patients (73.3%) were on prior topical medication in the present study and none had undergone prior microbiological work-up. Prior topical medication did not differ significantly between the CP group and the CN group (p=0.77). A similar result was shown by Marangon et al14 where patients on prior therapy were only slightly more likely to be CN (p=0.317). We believe that only history of prior topical medication does not affect the culture positivity rate; however, long duration of symptoms and subsequent long duration of therapy may decrease the chance of isolation of the microbes in culture.15 Other factors such as age, gender, associated ocular and systemic risk factors were comparable in both the groups.
Medical management (ie, formation of stable corneal scar on topical medication and/or minor procedure) was successful in 24/30 (80%) in the CN group and 16/30 (56.66%) in the CP group (p=0.053), which was nearly statistically significant. A higher number of patients healed on medical management alone in CN keratitis. Our strategy for CN keratitis is to start a broad spectrum topical fourth-generation fluoroquinolones (gatifloxacin, 0.3%) assuming the CN keratitis to be bacterial unless proved otherwise. The mean duration of medical treatment for corneal scar formation was 27±19.8 (range: 5–90) days in the CP group and 31±18.4 (range: 4–67) days in the CN group (p=0.44). The time for corneal scar formation was slightly longer in CN keratitis. A similar finding of delayed healing time of smear-negative keratitis was observed in a previous study.15 It is most likely because of specific drug therapy which is received by the CP group leading to faster resolution.
The CP group had significantly more major surgical intervention compared with the CN group (46.7% in the CP group: 3.3% in the CN group) (p=0.0002). The most common surgical intervention was TA+BCL (11/30; 36.7%) in CN keratitis and therapeutic keratoplasty (11/30; 36.7%) in CP keratitis. Our results are similar to previous study where they have compared surgical interventions in smear-negative keratitis and smear-positive keratitis. Both the groups had nearly the same number of surgical interventions (47% in the smear-negative group; 51.9% in the smear-positive group); however, the number of patients undergoing major surgical intervention (ie, therapeutic keratoplasty, AC wash, evisceration) was significantly higher in the smear-positive group compared with the smear-negative group (7.3% in the smear-negative group; 21.5% in the smear-positive group) (p=0.01). CP keratitis represents a more severe form of disease requiring major surgical intervention more often than CN keratitis. The majority of patients in the CP keratitis group had fungal aetiology requiring surgical intervention more frequently, which may have altered the results in this comparative study. However, it is worth noting that in a tropical country like India, fungus is much more common cause of keratitis compared with bacteria and thus the CP group represents closely to the geographic region of the study.
Ninety per cent of patients in the CP group (27/30) and 83.3% of patients in the CN group had successful outcomes. Treatment failure was seen in three patients (10%) of the CP group. All three patients had fungal aetiology. Two patients(6.7%) had severe disease at presentation with infiltrate size >6 mm. Both patients (6.7%) had phthisis bulbi at the end of follow-up. The third patient had 4×3 mm infiltrate but lost to follow-up after 8 days.
Five patients (16.7%) in the CN group had treatment failure. Four patients (13.3%) had worsening keratitis till the last follow-up. One patient (3.3%) had lens expulsion while doing TA+BCL and later went into phthisis bulbi.
Our primary goal of treatment was elimination of infection and a tectonically stable globe as visual acuity is a poor indicator of treatment outcome in microbial keratitis. Both CP and CN groups showed significant difference in visual acuity pretreatment and post-treatment. However, the final visual acuity was not significantly different between the two groups (p=0.07).
We assume that a subset of CN keratitis patients represent the inflammatory component of resolving corneal ulcers which eventually form corneal scar. Histopathology report of corneal button of a single therapeutic keratoplasty case in the CN group showed destruction with loss of lamellar arrangement of collagen fibres with patchy infiltrate of neutrophils, cell debris and few lymphocytes along with myofibroblastic transformation of fibrokeratocytic nuclei. Routine and special stain (periodic acid fast stain, Gomori methanamine stain) for fungus and Gram stain for bacteria showed no organism, suggesting that it is only the inflammatory component which is active in the corneal lesion.
The other subset of CN keratitis patients who do not respond to the CN keratitis treatment protocol (‘treatment failure’) may represent rare but not uncommon diagnosis of viral aetiology such as necrotising stromal keratitis which closely resembles infectious keratitis.16 Recurrent attacks of pain and redness, old corneal scars (herpetic footprints in the same or fellow eye), white stromal infiltrate (compared with dirty yellow infiltrate of bacterial or fungal keratitis), angry-looking extensive stromal vascularisation and epithelial defect eccentric to stromal infiltrate are some of the clinical signs to differentiate herpetic necrotising stromal keratitis from infectious keratitis. However, these signs may not always be present and strong clinical suspicion is the only key to manage this condition.17
Conclusion
CN keratitis is associated with long duration of topical medication. The number of major surgical interventions in CN keratitis is significantly less compared with CP keratitis. However, the final outcome of treatment is similar in both CP and CN keratitis.
References
Footnotes
Contributors All the authors included in this paper fulfil the criteria of authorship. SD and SS: concept and design; analysis and interpretation. YB and SD: writing the article. SD and SS: critical revision of the article. YB, SD, MKK, SKS and SS: final approval of the article. YB and MKKsav: data collection. SD, SKS and SS: Provision of materials, patients or resources. YB, SD, MKK and SKS: Literature search. SD, SKS and SS: administrative, technical or logistic support.
Funding Hyderabad Eye Research Foundation, Hyderabad.
Competing interests None declared.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- At a glance