A 3½-Year Study of Bacterial Keratitis From Odisha, India

ORIGINAL CLINICAL STUDY A 3½-Year Study of Bacterial Keratitis From Odisha, India Bibhudutta Rautaraya, MD,* Savitri Sharma, MD,* Md. Hasnat Ali, MBA,Þ Sarita Kar, MSc,* Sujata Das, MS, FRCS(Glasg),* and Srikant K. Sahu, MS* Purpose: To report the clinical and microbiological characteristics and treatment outcome of bacterial keratitis at a tertiary eye care center in eastern India. Design: Retrospective analysis of medical and microbiology records. Methods: All patients had undergone complete clinical and microbiological evaluation for microbial keratitis. Patients with laboratoryproven bacterial keratitis were included in the study. Results: Between July 2006 and December 2010, 1417 microbial keratitis was clinically diagnosed in the patients. Whereas no organisms were found in 27.8% (394/1417) of cases, 21.4% (303/1417) were bacterial. From 303 patients, 347 bacterial isolates were cultured, 260 (74.9%) of which were gram-positive, 67 were gram-negative, and 20 were acid-fast. Streptococcus pneumoniae was the predominant isolate (86/347 [24.7%]), followed by Staphylococcus species [64/347 (18.4%)]. Pseudomonas aeruginosa (29/347 [8.3%]) was the most common gram-negative bacterial isolate. High level of susceptibility to (96.2%) and vancomycin (96.5%) was found in gram-positive cocci, whereas susceptibility of Pseudomonas species to gatifloxacin was 95.1%. Fifty-three patients (17.5%) required tissue adhesive, and 47 (15.5%) needed penetrating keratoplasty. Healed corneal scar was achieved in 188 patients (62%), whereas 34 (11.2%) were lost to follow-up. Large stromal infiltrate size, older age group, and poor presenting visual acuity were significant factors that adversely affected final outcome (P G 0.05). Conclusions: Proportion of bacterial keratitis was low compared with other studies from India. Gram-positive bacteria were a common cause of bacterial keratitis with high susceptibility to and vancomycin. Gram-negative bacteria were sensitive to gatifloxacin with overall good treatment outcome. Key Words: bacterial keratitis, microscopy, culture, diagnosis, treatment (Asia-Pac J Ophthalmol 2014;3: 146Y150) The spectrum of microbial agents associated with corneal ulcer is wide and varies from one geographical location to another. Many differences in keratitis profile have been noted between populations living in rural or in urban areas, in western and in developing countries. 1 Proportion of bacterial keratitis also varies from one region to the other within the same country. 2 To have an effective strategy for diagnosis and treatment, it is important to have a perspective on etiologic agents in a given geographical region. To determine the impact of various epidemiologic patterns, diagnostic methods and treatment strategies From the *LV Prasad Eye Institute, Bhubaneswar, Odisha; and Centre for Clinical Epidemiology and Bio-statistics (CCEB), LV Prasad Eye Institute, Hyderabad, Andhra Pradesh, India. Received for publication February 15, 2013; accepted July 8, 2013. Supported by Hyderabad Eye Research Foundation, Hyderabad. The authors have no conflicts of interest to declare. Reprints: Savitri Sharma, MD, Fellow of Academy of Medical Sciences, Ocular Microbiology Service, LV Prasad Eye Institute, Patia, Bhubaneswar 751024, India. E-mail: savitri@lvpei.org. Copyright * 2014 by Asia Pacific Academy of Ophthalmology ISSN: 2162-0989 DOI: 10.1097/APO.0b013e3182a3f 301 on the outcome of infective keratitis, results from studies using standard procedures over a considerable period in a large number of patients would be most informative. At the LV Prasad Eye Institute (LVPEI), Bhubaneswar, India, every patient who reports to the cornea clinic with a stromal infiltrate in the cornea undergoes a standard protocol of clinical evaluation, diagnostic investigation, and therapeutic regimen, and all clinical and microbiological data are collected systematically. The aim of this retrospective bacterial keratitis study was to identify the predisposing factors, determine various bacterial species involved and their antibiotic susceptibility, describe treatment modalities, and analyze outcome. We have compared our results with 5 other studies from different parts of India. MATERIALS AND METHODS A retrospective review of microbiology and clinical records of patients at LVPEI, Bhubaneswar, from July 2006 to December 2010 identified 347 bacterial isolates from 303 patients of culture-proven bacterial keratitis. This study was approved by the institutional review board of LVPEI (ethics ref. no. LEC-11-260). Patients sociodemographic features, duration of symptoms, predisposing factors, other systemic diseases, therapy received before presentation, visual acuity at the time of presentation and last follow-up, clinical features, treatment given, response to treatment during follow-up and the clinical outcome were documented. The study included patients with laboratory-proven culture-positive bacterial keratitis. All patients were followed up to at least 3 weeks. Corneal scrapings were collected and processed from all patients as per the institutional protocol published earlier. 3 Our laboratory complies by the national guidelines for laboratory quality assurance and control. 4 For microscopy and culture, corneal scrapings were collected from each patient. For microscopic examination, the scrapings tended to be sequentially collected and respectively stained by 10% potassium hydroxide with 0.1% calcofluor white (fluorescence microscopy) and Gram and Giemsa stains. Modified Ziehl-Neelsen stain using 1% H 2 SO 4 was done when thin, beaded, branching filaments were observed in Gram stain or when there was clinical suspicion of Nocardia infection. Likewise, Ziehl-Neelsen stain using 20% H 2 SO 4 was done when unstained or partially stained thin long bacilli were seen in Gram stain or when atypical mycobacterial infection was suspected clinically. The scrapings were inoculated on 5% sheep blood/chocolate agar, brain heart infusion broth, thioglycolate broth, and Sabouraud dextrose agar with 0.005% chloramphenicol. The media were incubated, observed daily for any growth, which was processed as per standard protocol. The media were held for 2 weeks in case of no growth. The significance of any growth was determined as published earlier. 3 Antibiotic susceptibility testing was done by Kirby Bauer disc diffusion assay, and the results were interpreted as per the Clinical and Laboratory Standards Institute guidelines. Intensive antibiotic topical therapy was given depending on the direct microscopic results of the corneal scrapings. Subsequently, the antibiotic drop was changed, if required, depending on the culture and sensitivity report. The treatment was not changed if clinical 146 www.apjo.org Asia-Pacific Journal of Ophthalmology & Volume 3, Number 3, May/June 2014

Asia-Pacific Journal of Ophthalmology & Volume 3, Number 3, May/June 2014 Bacterial Keratitis TABLE 1. Occupation and Age-Wise Distribution of Patients With Bacterial Keratitis (n = 303) Serial number Occupation n (%) Age, y Male, n (%) Female, n (%) Total, n (%) 1. Farmer 88 (29) 0Y9 6 (2.8) 7 (8.1) 13 (4.3) 2. Housewife 48 (15.8) 10Y19 11 (5.1) 4 (4.6) 15 (4.9) 3. Student 24 (7.9) 20Y29 29 (13.4) 11 (12.6) 40 (13.2) 4. Business 12 (3.9) 30Y39 40 (18.5) 4 (4.6) 44 (14.5) 5. Mechanic 4 (1.3) 40Y49 29 (13.4) 19 (21.8) 48 (15.8) 6. Fisherman 4 (1.3) 50Y59 27 (12.5) 8 (9.2) 35 (11.6) 7. Gardener/carpenter/electrician 7 (2.3) 60Y69 43 (19.9) 26 (29.8) 69 (22.8) 8. Others 21 (6.9) 70Y79 29 (13.4) 8 (9.2) 37 (12.2) 9. Retired 95 (31.3) Q80 2 (0.9) 0 (0) 2 (0.6) Total 303 Total 216 (71.3) 87 (28.7) 303 (100) improvement was obvious with the initial antibiotic. One hourly topical drop was applied for the first 3 days round-the-clock followed by 2 hourly drops during waking hours until resolution of the ulcer. Patients also received 1% atropine sulfate cycloplegic drops. Additional procedures were undertaken at the discretion of the clinicians for patients not responding to medical therapy, and they included therapeutic penetrating keratoplasty (PK), cyanoacrylate glue application with bandage contact lens, or anterior chamber wash with antibiotics and evisceration. After treatment, an ulcer was considered healed when the epithelial defect was found to be less than 1 mm in maximum diameter with slit-lamp biomicroscopy and/or a visible scar. A healing time of less than 2 weeks from presentation was considered good result, and healing time of more than 3 weeks was considered a poor response. Healing in the interim period constituted moderate response. Factors affecting treatment outcome were evaluated by statistical software R, version 2.14.1 (GNU General Public License). Multivariate Cox proportional hazards regression model by Akaike information criterion in a stepwise algorithm was used to estimate the adjusted hazard ratio, and P e 0.05 was considered statistically significant. RESULTS The most common age of patients (Table 1) with bacterial keratitis was 60 to 69 years (22.8%), and the disease was more common in males [216/303 (71.34%)]. No predisposing factor was reported by the patients in 123 cases (40.5%) (Fig. 1). Among patients in whom predisposing factor could be elicited, trauma by plant was commonest [78/180 (43.3%)], and the most common people affected were farmers [88/208 (42.3%)]. During the study period, 1417 patients were clinically diagnosed as nonviral microbial keratitis and were investigated for bacteria, fungi, or parasites. Although no organisms were found in 394 clinical samples (27.8%), 303 (21.4%) were bacterial, 405 (28.6%) were fungal with or without bacteria, 21 (1.5%) were Acanthamoeba with or without bacteria, and 294 (20.7%) were microsporidia with or without bacteria. A total of 347 bacterial isolates were obtained from the corneal samples of 303 patients. More than 1 bacterial isolate was found in 44 patients (14.5%) (2 isolates in 33 [10.9%], 3 isolates in 11 [3.6%]). Of the 347 bacterial isolates, the percentage positivity of the corneal scraping in direct microscopy for staining procedures, that is, Gram [226/342 (66%), not done in 5 cases] and Giemsa stain, was equal [44/66 (66%), not done in 281 cases]. Of 347 bacterial isolates, 260 (260/347 [74.9%]) were gram-positive, 67 were gram-negative, and 20 were acid-fast (20% H 2 SO 4 ) or partially acid-fast (1% H 2 SO 4 ). Streptococcus pneumoniae was found in 86 (24.7%) of 347 followed by Staphylococcus species [64/347 (18.4%)], and the difference in proportions was statistically not quite significant by W 2 test (2-tailed, P = 0.0528). Dacryocystitis was detected in patients (22.3%) with S. pneumoniae infection. Among the gram-negative isolates, Pseudomonas aeruginosa [29/347 (8.3%)] was the most common (Table 2). Most gram-positive bacteria were susceptible to (96.2%) and vancomycin (96.5%), followed by chloramphenicol (88.9%) and gatifloxacin (86.8%). All S. pneumoniae isolates FIGURE 1. Predisposing factors found in patients with bacterial keratitis (n = 303). * 2014 Asia Pacific Academy of Ophthalmology www.apjo.org 147

Rautaraya et al Asia-Pacific Journal of Ophthalmology & Volume 3, Number 3, May/June 2014 TABLE 2. Type of Bacteria Isolated From Corneal Scrapings of Patients (n = 303) With Bacterial Keratitis Sl No. Organism Isolated n % Gram-positive cocci 1. S. pneumoniae 86 24.8 2. Staphylococcus species 65 18.7 3. S. aureus 47 13.5 4. Micrococcus species 4 1.2 5. Streptococcus species 13 3.7 6. Corynebacterium species 45 13 Acid-fast and partially acid-fast 7. Mycobacterium species 7 2 8. Nocardia asteroides 13 3.7 Gram-negative bacilli and cocci 9. P. aeruginosa 29 8.3 10. Pseudomonas species 8 2.3 11. Enterobacteriaceae 12 3.4 12. Acinetobacter species 6 1.7 13. Haemophilus species 5 1.4 14. Moraxella species 7 2 Total 347 Patients with more than 1 organism = 44. were susceptible to (100%). Staphylococcus aureus was least sensitive to all fluoroquinolones (29%Y68%). Vancomycin (100%) was the most effective antibiotic for Corynebacterium species. For gram-negative bacteria including Pseudomonas, gatifloxacin was most effective (95.1%), followed by ciprofloxacin (90%). Nocardia species were susceptible to amikacin (100%) and to all fluoroquinolones (85%Y100%), and all Mycobacterium species were susceptible to amikacin (Supplemental Digital Content 1, http://links.lww.com/apjo/a44). Among the topical antibiotics, gatifloxacin (25.4%) was the most common antibiotic prescribed empirically when direct microscopy result was negative, and it was combined with in 15.5% of patients when gram-positive organisms were seen in microscopy (Table 3). Based on culture report, the most common antibiotic that was prescribed was also gatifloxacin [35 (11.6%)]. Systemic ciprofloxacin (500 mg twice daily), amoxicillin/clavulanic (625 mg thrice daily), ampicillin/cloxacillin (500 mg 4 times a day), co-trimoxazole (160/800 mg twice daily), and doxycycline (100 mg twice daily) were prescribed to 53 patients with corneal stromal infiltrate extending beyond one third of the cornea. Surgical intervention was required for the treatment in several patients, as shown in Table 3. Tissue adhesive was required in 53 patients (17.5%) for impending perforation, and in 47 (15.5%), PK was done. All patients were followed up to at least 3 weeks. Thirty-four patients (11.2%) were lost to follow-up. In 188 patients (62.0%), the outcome of healed scar/resolving ulcer was achieved. This outcome was seen in 27 patients (14.3%) within 2 weeks that constituted good clinical outcome. In 150 patients (46.6%), there was improvement in visual acuity, where as in 37 patients (13.7%) there was decrease in visual acuity, and there was no change in acuity in 126 cases (39.6%). Details of treatment outcome achieved at the time of data collection are given in Table 4. To evaluate the factors affecting treatment outcome by multivariate Cox proportional hazards regression model, 6 parameters were analyzed, of which the final model retained 5 variables (Table 5). A healed or resolving ulcer (on medical therapy) at the last follow-up, irrespective of the time period, was defined as successful treatment outcome in this model. Complete data were available for 159 patients. The analysis showed that older age, stromal infiltrate size more than 25 mm 2, and poor visual acuity at the time of presentation were significantly associated with poor outcome. A comparison of our data with other publications from different parts of India is given in Supplemental Digital Content 2, http://links.lww.com/apjo/a45. Bacterial keratitis is reported from all over India with S. pneumoniae as either the most common or second most common organism involved. None of the other studies have reported treatment outcome. DISCUSSION The present study presents a detailed laboratory and clinical data of a large number of bacterial keratitis patients from eastern India. A comparison of the results of the current study with that of other parts of India is shown in Supplemental Digital Content 2, http://links.lww.com/apjo/a45. 2,5Y8 A fairly large number of patients have been analyzed in all these published reports. The proportion of culture-proven bacterial keratitis is within 15% to 27% in western, central, northeast India including the present study. The proportion is high in southern India 8,9 (32.8%Y35%) and in Ahmedabad as reported by Tewari et al 10 (38%). It is also high in western Maharashtra (42.08%) in a study by Deorukhkar et al. 11 In a study from western Nepal by Dhakhwa et al, 12 it is reported to be 23.19%. In the western world, the proportion of culture-proven bacterial keratitis is much higher (50%). 13 This has implications in response to empiric treatment. Geographic areas with high proportion of bacterial keratitis are more likely to benefit from empiric treatment with broad-spectrum antibiotics. A specific laboratory-based treatment is necessitated when other causes such as fungal are prevalent. Across the country, including the current study, bacterial keratitis seems to be more in males and in farmers except for the study from Ahmedabad, 10 where they have reported highest in housewives (21.3%). One of the studies from Nepal 12 has found highest bacterial keratitis in farmers (33.3%), but contradictory to our study, a study by Upadhyay et al 14 from Nepal has found males and females to be equally affected. Similar to fungal keratitis, the most common predisposing factor remains trauma to the cornea. Bacterial keratitis is common in the age group younger than 50 years in studies from western and northeastern India, in contrast to 60 to 69 years (29.8%, females; 19.9%, males) in the present study. The reason for this trend is not clear. Gram stain can diagnose up to 66.6% of the cases. In a study by Khanal et al, 15 direct microscopy was positive in 48% of cases. The rest were diagnosed only on culture as seen in the study done in central India by Satpathi and Satpathi 6 and the present study. Because clinical acumen would vary according to the level of training and experience, it seems appropriate for all practitioners to have a minimum laboratory facility available in their clinic for better management of microbial keratitis. The most common isolate in most recent studies from India and other countries is Staphylococcus species, gram-positive cocci that are normally considered nonpathogenic in the eye. 1 In Nepal, the scenario has changed from S. pneumoniae in a study by Upadhyay et al 14 in 1987 to Staphylococcus epidermidis in a study by Dhakhwa et al 12 in 2012. Although there was no statistical difference in the proportions of bacterial keratitis caused by staphylococci and S. pneumoniae in our study, a greater number of cases were caused by S. pneumoniae. The higher proportion may be owing to presence of dacryocystitis in 22.3% patients included in the study. The study from Florida 13 and Texas 16 in the United States has reported P. aeruginosa (25.7%) as the most common agent followed by S. aureus (19.4%). The 148 www.apjo.org * 2014 Asia Pacific Academy of Ophthalmology

Asia-Pacific Journal of Ophthalmology & Volume 3, Number 3, May/June 2014 Bacterial Keratitis TABLE 3. Treatment Received by Patients With Bacterial Keratitis (n = 303) Initial Treatment Topical Antibiotic Changes Made After Culture Report Systemic Antibiotic, n (%) Surgical Treatment Sl No. Antibiotic n (%) Antibiotic n (%) Antibiotic n (%) Surgical Procedure n (%) 1. Ciprofloxacin 74 (24.4) Gatifloxacin 35 (11.6) Ciprofloxacin 20 (6.6) TABCL 53 (17.5) 2. Gatifloxacin 77 (25.4) Cefazolin 22 (7.2) Ampiclox 20 (6.6) PK 47 (15.5) 3. Gatifloxacin + 47 (15.5) Ofloxacin 26 (8.5) Amoxiclav 8 (2.6) TABCL + PK 11(3.6) 4. Ofloxacin 40 (13.2) Chloramphenicol 2 (0.6) Co-trimoxazole 2 (0.6) Evisceration 8 (2.6) 5. Ciprofloxacin + 21 (6.9) Ciprofloxacin 13 (4.2) Cefazolin 1 (0.3) DCR 7 (2.3) 6. Amikacin 17 (5.6) Gatifloxacin + 5 (1.6) Doxycycline 1 (0.3) Orbitotomy 1 (0.3) 7. Ofloxacin + 2 (0.6) Chloramphenicol + 10.3) Amoxyclav + 1(0.3) Corneal tear repair 1 (0.3) vancomycin ciprofloxacin 8. Cefazolin 6 (1.9) Vancomycin 2 (0.6) No antibiotics 250 (82.5) No surgery 175 (57.7) 9. Tobramycin 3 (0.9) Ciprofloxacin + 2 (0.6) Total 303 Total 303 10. Moxifloxacin 1 (0.3) Moxifloxacin 2 (0.6) 11. Vancomycin 1 (0.3) Gentamicin 2 (0.6) 12. No antibiotics* 14 (4.6) No change done 191 (63.1) Total 303 Total 303 *Patients did not follow instructions. TABCL indicates tissue adhesive with bandage contact lens; DCR, dacryocystorhinostomy; PK, penetrating keratoplasty. common gram-negative bacilli isolated and reported in India is P. aeruginosa. 2,7,8 In Nepal and European countries, the picture is almost the same, but Pseudomonas is generally seen in contact lens users in the European world. 13 The most effective antibiotic varies from place to place within India. Studies from western India 2 have reported (94%) as the most effective antibiotic for gram-positive in contrast to southern India, 9 where moxifloxacin (99%) and vancomycin (99%) were found to be the best. In the present study from eastern part of India, vancomycin (96%Y100%) and (96.2%) were the better antibiotics for gram-positive bacteria. In the western world, resistance to fluoroquinolones in Staphylococcus is gradually increasing from 11% in 1990 to 28% in 1998. 13 Ceftazidime is the most useful antibiotic against gram-negative isolates from a study from western India, 2 whereas susceptibility to amikacin (100%) is highest in southern India. 9 Nevertheless, these 2 antibiotics do not normally find a place in topical therapy for bacterial keratitis. In the present study, gatifloxacin (87.5%Y95%) was most effective against gramnegative bacteria, which suggests that gatifloxacin could be used for monotherapy of bacterial keratitis due to gram-negative bacteria. In the study from Nepal, 12 ciprofloxacin and ofloxacin (79.31%) were found to be most effective against gram-negative bacteria, but the picture in the western world is reverse where susceptibility to fluoroquinolones and aminoglycosides is less among Pseudomonas isolates. 13 Our findings with respect to susceptibility of Nocardia and atypical mycobacteria to amikacin, vancomycin, and gatifloxacin were on expected lines, with all isolates being sensitive to these antibiotics. Topical amikacin (2.5%) made from injection formulation is the drug of choice for the treatment of Nocardia and atypical mycobacteria keratitis. 17 TABLE 4. Treatment Outcome in 303 Patients With Bacterial Keratitis Sl No. Treatment Outcome n (%) 1. Healed scar/resolving 188 (62) 2. Clear graft 14 (4.6) 3. Failed graft 30 (9.9) 4. Status quo 19 (6.3) 5. Worsening 10 (3.3) 6. Evisceration 8 (2.6) 7. LTFU 34 (11.2) Total 303 LTFU indicates lost to follow-up. TABLE 5. Multivariate Cox Proportional Hazards Regression Analysis of 5 Variables Affecting Visual Outcome in Patients With Bacterial Keratitis (n = 159) 95% CI HR LCL UCL P Age, y 0.97 0.95 0.99 0.01 VA at presentation 2.14 1.28 3.57 0.01 Size of infiltrate (925 mm 2 ) 2.96 1.07 8.16 0.04 Gram-negative bacteria 2.17 0.82 5.76 0.12 Hypopyon 0.38 0.14 1.03 0.06 P e 0.05 was considered statistically significant. CI indicates confidence Interval; HR, hazard ratio; LCL, lower confidence limit; UCL, upper confidence limit; VA, visual acuity. * 2014 Asia Pacific Academy of Ophthalmology www.apjo.org 149

Rautaraya et al Asia-Pacific Journal of Ophthalmology & Volume 3, Number 3, May/June 2014 The final outcome of healed scar was achieved in 62% of cases in the present study. Compared with fungal keratitis, response to medical treatment is usually good in bacterial keratitis with more than 75% resulting in healed scar. 3 Patients requiring PK were 47 (15.5%), and evisceration was done in 8 patients (2.6%). All these patients had presented late and were caused by both gram-positive and gram-negative organisms. As expected, early treatment results in favorable outcome. This is obvious in this study as a large number of patients with healed cornea had presented within 4 to 7 days of symptoms. Penetrating keratoplasty and tissue adhesive application for impending perforation were seen in patients who presented late after 10 days. Not surprisingly, the treatment outcome was poor in patients who had presented with poor visual acuity, had stromal infiltrate size larger than 25 mm 2, and who were older (Table 5). Presence of hypopyon and gram-negative bacterial etiology did not significantly influence outcome. In our study, the most prescribed antibiotic before and after culture was gatifloxacin, as gatifloxacin was effective in 87% of gram-positive and 95% in gram-negative organisms. Ofloxacin (12.2%) was the most common antibiotic used in patients who received prior treatment before coming to the institute. The second common antibiotic used in patients before coming to our hospital was gatifloxacin (9.2%). Our study highlights that the proportion of bacterial keratitis in patients coming to our hospital from eastern India is less compared with other parts of India, and the predominant bacterium is S. pneumoniae, which remains sensitive to a large number of antibiotics. The medical treatment outcome is generally good, although patients presenting late may require surgical intervention. REFERENCES 1. Bourcier T, Thomas F, Borderie V, et al. Bacterial keratitis: predisposing factors, clinical and microbiological review of 300 cases. Br J Ophthalmol. 2003;87:834Y838. 2. Kumar A, Pandya S, Kavathia G, et al. Microbial keratitis in Gujarat, western India: findings from 200 cases. Pan Afr Med J. 2011;10:48. 3. Gopinathan U, Garg P, Sharma S, et al. Review of epidemiological features microbiological diagnosis and treatment outcome of microbial keratitis: experience over a decade. Indian J Ophthalmol. 2009;57:273Y279. 4. Arora DR. Quality assurance in microbiology. Indian J Med Microbiol. 2004;22:81Y86. 5. Chawla B, Agarwal P, Tandon R, et al. In vitro susceptibility of bacterial keratitis isolates to fourth generation fluroquinolones. Eur J Ophthalmol. 2010;20:300Y305. 6. Satpathi P, Satpathi S. Study of microbial keratitis in central India. J Infect Dev Ctries. 2012;6:295Y298. 7. Basak SK, Basak S, Mohanta A, et al. Epidemiological and microbiological diagnosis of suppurative keratitis in Gangetic West Bengal, Eastern India. Indian J Ophthalmol. 2005;53:17Y22. 8. Bharathi MJ, Ramakrishnan R, Shivakumar C, et al. Etiology and antibacterial susceptibility pattern of community-acquired bacterial ocular infections in a tertiary eye care hospital in south India. Indian J Ophthalmol. 2010;58:497Y507. 9. Bharathi MJ, Ramakrishnan R, Vasu S, et al. Aetiological diagnosis of microbial keratitis in south IndiaVa study of 1618 cases. Indian J Med Microbiol. 2002;20:19Y24. 10. Tewari A, Sood N, Vegad MM, et al. Epidemiological and microbiological profile of infective keratitis in Ahmedabad. Indian J Ophthalmol. 2012;60:267Y272. 11. Deorukhkar S, Katiyar R, Saini S. Epidemiological features and laboratory results of bacterial and fungal keratitis: a five year study at a rural tertiary-care hospital in western Maharastra, India. Singapore Med J. 2012;53:264Y267. 12. Dhakhwa K, Sharma MK, Bajimaya S, et al. Causative organisms in microbial keratitis, their sensitivity pattern and treatment outcome in western Nepal. Nepal J Ophthalmol. 2012;4:119Y127. 13. Alexandrakis G, Alfonso EC, Miller D. Shifting trends in bacterial keratitis in South Florida and emerging resistance to fluoroquinolones. Ophthalmology. 2000;107:1497Y1502. 14. Upadhyay MP, Karmacharys PC, Koirla S, et al. Epidemiologic characteristics, predisposing factors and etiologic diagnosis of corneal ulceration in Nepal. Am J Ophthalmol. 1991;111:92Y99. 15. Khanal B, Deb M, Panda A, et al. Laboratory diagnosis in ulcerative keratitis. Ophthalmic Res. 2005;37:123Y127. 16. Pachigolla G, Blomquist P, Cavanagh HD. Microbial keratitis pathogens and antibiotic susceptibilities: a 5-year review of cases at an urban county hospital in north Texas. Eye Contact Lens. 2007;33:45Y49. 17. Garg P. Fungal, Mycobacterial and Nocardia infections and the eye: an update. Eye. 2012;26:245Y251. "Wine and cheese are ageless companions, like aspirin and aches, or June and moon, or good people and noble ventures." V M. K. F. Fisher 150 www.apjo.org * 2014 Asia Pacific Academy of Ophthalmology