Indian Journal of Ophthalmology

ORIGINAL ARTICLE
Year
: 2003  |  Volume : 51  |  Issue : 4  |  Page : 315--321

Epidemiological characteristics and laboratory diagnosis of fungal keratitis. A three-year study


M Jayahar Bharathi, R Ramakrishnan, S Vasu, R Meenakshi, R Palaniappan 
 Department of Microbiology, Aravind Eye Care System, Tirunelveli, Tamil Nadu, India

Correspondence Address:
M Jayahar Bharathi
Department of Microbiology, Aravind Eye Care System, Tirunelveli, Tamil Nadu
India

Abstract

Purpose: To study the epidemiological characteristics and laboratory diagnosis of fungal keratitis seen at a tertiary eye care referral centre in South India. Methods: A retrospective review of all culture-proven fungal keratitis seen over a 3-year period, September 1999 through August 2002. Results: Fungal aetiology were confirmed in1095(34.4%) of 3183 corneal ulcers. The predominant fungal species isolated was Fusarium spp (471; 42.82%) followed by Aspergillus spp (286; 26%). Males (712; 65.08%) were more often affected (P< 0.0001). A large proportion of the patients (732; 66.85%) were in the younger age group (21 to 50 years). A majority (879; 80.27%) came from rural areas (P<0.0001), and most patients (709; 64.75%) were farmers (P<0.0001). Ocular trauma (1009; 92.15%) was a highly significant risk factor (P<0.0001) and vegetative injuries (671; 61.28%) were identified as a significant cause for fungal keratitis (P<0.0001). 172 (15.71%) patients had concurrent diabetes mellitus. The sensitivity of 10% potassium hydroxide (KOH) wet mount preparation was higher (99.23%) than Gram-stained smear (88.73%) (P<0.0001). Incidence of fungal keratitis was higher between June and September. Conclusion: Agricultural activity and related ocular trauma were principal causes of mycotic keratitis. A potassium hydroxide (KOH) wet mount preparation is a simple, and sensitive, method for diagnosis.



How to cite this article:
Bharathi M J, Ramakrishnan R, Vasu S, Meenakshi R, Palaniappan R. Epidemiological characteristics and laboratory diagnosis of fungal keratitis. A three-year study.Indian J Ophthalmol 2003;51:315-321


How to cite this URL:
Bharathi M J, Ramakrishnan R, Vasu S, Meenakshi R, Palaniappan R. Epidemiological characteristics and laboratory diagnosis of fungal keratitis. A three-year study. Indian J Ophthalmol [serial online] 2003 [cited 2024 Mar 29 ];51:315-321
Available from: https://journals.lww.com/ijo/pages/default.aspx/text.asp?2003/51/4/315/14654


Full Text

Corneal blindness is a major public health problem worldwide and infectious keratitis is one of the predominant causes. [1],[2],[3],[4],[5],[6],[7] Corneal infection of fungal aetiology is very common[8],[9],[10],[11]and represents 30% to 40% of all cases of culture-positive infectious keratitis in South India. [12],[13],[14]Fungal keratitis is a major ophthalmic problem.[15]

Filamentous fungi are responsible for a larger proportion of these corneal infections in tropical climates than in temperate climates, particularly following trauma with vegetative matter. In tropical climatic conditions as in South Florida,[8] Bangladesh,[10],[11] South India[12],[14],[16] and Nepal[17] the incidence of fungal keratitis is reported to be from 17% to 40%. In temperate climates such as Britain[18] and Northern United States,[19] the proportion of fungi causing suppurative keratitis is very low. Similarly, at the high altitude of Johannesburg, South Africa, the incidence is not more than 2.3%.[20],[21] At least 70 genera of fungi have been associated with fungal keratitis.[22] Of these, Fusarium species and Aspergillus species are responsible for 70 % of cases. [23],[24],[25]

There are several studies of fungal corneal ulcer from South India.[12]-[14],[26],[27],[28]The aetiological and epidemiological pattern of corneal ulceration varies significantly with patient population, health of the cornea,[29] geographic region, climate and also tends to vary over time Hence, an understanding of the current status of the regional epidemiological features, risk factors, the presence of ocular and/ or systemic co-morbidities, occupational status and knowledge of region-wise aetiological agents is important in the prevention and appropriate management of fungal keratitis.

The purpose of this study was to evaluate the specific pathogenic agents and study epidemiological characteristics of fungal keratitis presenting at a tertiary referral centre in the southern most part of India.

 Materials and Methods



Patients

A retrospective analysis was performed of all patients with culture-proven fungal keratitis seen over a period of 3 years, September 1999 to August 2002. A total of 3183 consecutive patients with corneal ulceration seen at a tertiary eye care referral centre in south India were analysed.Ulceration was defined as a loss of the corneal epithelium with underlying stromal infiltration and suppuration associated with signs of inflammation with or without hypopyon. Ulcers with typical features of viral infection and healing ulcers were excluded as were Mooren's ulcers, interstitial keratitis, sterile neurotropic ulcers, and any ulcer associated with autoimmune conditions. A standardised form was filled out for each patient, documenting sociodemographic features, duration of symptoms, predisposing factors, history of corneal trauma, traumatising agents, associated ocular conditions, other systemic diseases, therapy received prior to presentation, visual acuity at the time of presentation and all the clinical findings.

Clinical procedures

All patients received a slitlamp biomicroscopic examination by an ophthalmologist.The size of the epithelial defect after staining with 2% fluorescein was measured with the variable slit on the biomicroscope and recorded in millimeters.[12] In similar fashion the size and depth of the stromal infiltrate was recorded.A sketch of each ulcer was drawn on the form using standardised frontal and cross-sectional diagrams. The presence or absence of a hypopyon was recorded and the height measured in millimetres.Associated ocular conditions such as blepharitis, conjunctivitis, dacryocystitis, spheroidal corneal degeneration, dry eyes, bullous keratopathy, pre-existing viral keratitis, lid abnormalities, Bell's palsy, lagophthalmos, trichiasis, suture infiltrates and adherent leucoma were noted.[12], [30] The use of contact lenses and of topical corticosteroids and other systemic combinations were also recorded.[12], [30]

After a detailed ocular examination using standard techniques,an ophthalmologist took corneal scrapings under aseptic conditions from each ulcer using a sterile Bard-Parker blade (No 15).[12], [30],[31],[32]. The procedure was performed under magnification of slitlamp or operating microscope after instillation of 4% lignocaine (lidocaine) without preservative.[12] The scraping material obtained from the leading edge and the base of each ulcer was initially inoculated directly onto solid media such as sheep's blood agar, chocolate agar, or Sabouraud's dextrose agar in a row of C-shaped streaks. Deep inoculation in liquid media such as brain heart infusion broth without gentamicin sulphate and thioglycollate medium was also done.[12], [30],[31],[32]. The material obtained by scraping was spread onto labeled slides in a thin, even manner for 10% KOH wet mount, Gram's staining, and Giemsa staining.[12], [30],[31],[32]In cases of suspected actinomycete keratitis, Kinyoun's method of acid-fast staining was performed.[12], [30],[31],[32] When KOH smears were positive for amoebic cysts further corneal scrapings were performed and the materials were inoculated onto non-nutrient agar.[12], [30],[31],[32] Meticulous care was taken in collection of material and its aseptic transfer to the appropriate culture media.[12]

Laboratory procedures

All inoculated media were incubated aerobically.[12] The inoculated Sabouraud's dextrose agar were incubated at 27�C, examined daily, and discarded at 3 weeks if no growth was seen. The inoculated blood agar, chocolate agar, thioglycolate broth, and brain heart infusion broth were incubated at 37�C, examined daily, and discarded at 7 days if no growth was seen. Broth tubes were held upright in racks. The inoculated non-nutrient agar plates were incubated at 37�C after overlaying with Escherichia coli broth culture and were examined daily for the presence of Acanthamoeba species and discarded at 3 weeks, if there were no signs of growth. All laboratory methods followed standard protocols.[12], [30],[31],[32] Microbial cultures were considered significant if growth of the same organism was demonstrated on more than one solid phase medium, and/or if there was confluent growth at the site of inoculation on one solid medium, and/or if growth of one medium with consistent with direct microscopy findings (that is, appropriate staining and morphology with Gram-stain) and/or if the same organism was grown from repeated scraping.[12], [30],[31],[32]

Fungus grown on the primary isolation medium was subcultured onto an SDA medium and incubated for a period of 15 days to facilitate sporulation.[13] Following adequate growth of the fungal isolate on SDA, identification was done based on its macroscopic and microscopic features.[13] Pearson's chi-square test was used for the statistical analysis wherever required.

 Results



During the study period of three years, 3183 patients with the clinical diagnosis of corneal ulceration were evaluated at our institute. 1095 (34.4%) of 3183 patients grew fungus, 1043 (32.77%) had bacterial growth, 33 (1.04%) had Acanthamoeba species growth, 76 (2.39%) were mixed bacterial and fungal growth and the remaining 936 (29.41%) were culture negative. 1100 fungal isolates were recovered from 1095 fungal keratitis [Table 1] cases. 1090 (99.54%) cases had single species of isolates and remaining 5 (0.46%) cases had two species of fungal isolates. 822 (74.73%) of 1100 growth were hyaline (non-pigmented) group and 278 (25.27%) were dematiaceous (pigmented) group. No yeast or yeast-like organisms were isolated. The predominant fungal species isolated was Fusarium spp (471; 42.82%) followed by Aspergillus spp (286; 26%). Males (712; 65.02%) were affected more often ( P Fusarium spp was the predominant species in this study, similar to the reports from South Florida[8] and Ghana.[9] This is in contrast to most reports of Aspergillus spp from India.[22],[35],[38],[39] and Candida spp in other parts of the world. [40],[41],[42] The dematiaceous fungi are frequently reported as causes of keratitis in many tropical and subtropical regions.[43]

The incidence of fungal keratitis was significantly higher in males, in individuals from rural area and following corneal injury. The calculated odds are 1.47 times in males (95% confidence interval(CI) 1.26-1.71), 2.2 times in individuals from rural areas (95% CI, 1.87-2.65); 1.4 times in farmers (95%CI; 1.19 -1.61); 7.7 times following injury (95% CI, 6.12 - 9.85); 23.6 times following injury with vegetative matter (95% CI, 19.07 - 29.3). Male preponderance for corneal ulcers has been described in many studies.[9],[12],[13],[17],[22],[44] We observed that younger people, aged 21-50 years, are more often affected by fungal keratitis compared to those above 50 years, who are affected by bacterial corneal ulcers (unpublished data). Higher incidence of fungal keratitis has also been reported among farmers.[12],[13],[17],[22] We have also reported 70% incidence of Nocardia spp keratitis in farmers.[45]

Corneal trauma has always been identified as a cause of microbial keratitis,[12],[17],[46] and Schaefer et al have also identified co-existing ocular diseases as a major predisposing factor.[47] Predisposing ocular conditions and use of corticosteroids associated with development of fungal keratitis accounted for 86 (7.85%) patients. 172 (15.71%) patients had diabetes mellitus. Association of fungal keratitis with use of corticosteroids and diabetes mellitus has been reported earlier.[13],[22],[39],[44]

The clinical picture of fungal keratitis is pleomorphic. It varies from individual to individual, and largely depends upon the type of fungus, severity of the invading pathogen, liberation of toxin, resistance of host tissue and age of the patient.[48],[49] The unique biomicroscopic appearnce is a dry, raised, grayish white lesion and/ or as stromal infiltration with a feathery or hyphase border,[23],[48],[49] seen in 826 (75.43%) and 786 (71.78%) of our patients respectively. Satellite lesions seen in 110 (10%) patients also assisted in diagnosis.[23],[48],[49],[50] Other lesions that helped in diagnosis were heaped up or cheesy hypopyon in 609 (55.6%) patients, immune ring in 15 (1.37%) patients, corneal perforation in 15 (1.37%) and posterior corneal abscess in 11 (1%) patients. A dendritic pattern was observed in 46 (4.2%) patients in our study. It occurs in early stages of the fungal corneal infection. Sometimes these features may cause misdiagnosis and prompt a treatment with antiviral drugs or even corticosteroids.[23] The sensitivity and specificity of clinical diagnosis of fungal keratitis made by an ophthalmologist using slitlamp biomicroscope was 94.1% and 94.58% respectively. There was a high correlation between clinical and culture-based diagnosis of fungal keratitis and there was no significant difference between the two (P=1).

The observation on smears and cultures of all 3183 cases highlight the value of the traditional method of KOH wet mount preparation in the diagnosis of microbial keratitis. The sensitivity of KOH wet mount preparation (99.23%) was higher than that of Gram-stained smear (88.73%) in the detection of fungal filaments. In the Hyderabad study the sensitivity and specificity of KOH wet mount preparation in the detection of fungal filaments was 81.2% and 83.8%[51] respectively. The value of 10% KOH wet mount preparation in the diagnosis of fungal keratitis lies in its ability to clear the scraping of cellular debris, thereby rendering hyphal fragments more refractile on microscopic examination.[52] The staining quality of Gram-stain is often variable, hyphal elements frequently appear as linearly stained precipitates and it is usually not possible to determine whether they are coenocytic or septate. If the stained smear of scraping from an ulcer is thick, the hyphae will be interspersed through necrotic tissue and their identification may be difficult or impossible.[52] KOH has been used as 10 to 20% suspension, alone,[51] with ink[53] or with lactophenol cotton blue[54] with variable sensitivity.[8], [51], [55] Additional benefit of KOH preparation is its ability for rapid detection and early diagnosis of Nocardia and Acanthamoeba .[56], [57]

The incidence of fungal keratitis was higher during paddy harvesting and also during the time of year when agriculture activity was greater. In our study it was more prevalent during June through September. The peak incidence correlates with windy and dry weather during the month of June through September. In coastal Karnataka, higher incidence is reported in October, June and January[38] and in Hyderabad, higher incidence of fungal keratitis is reported during the winter (October to January) and monsoon (June to September) seasons.[13] A hot, humid, windy climate and an agriculture-based occupation of a large population make fungal keratitis more frequent in tropical zones.[13]

In summary, this study presents the epidemiological and laboratory findings of the largest series of fungal keratitis among the population in our geographic region. Fungal keratitis continues to be a cause of concern to ophthalmologists. Predominance of agricultural activity is the principal causative factor. KOH wet mount preparation is as sensitive as other conventional methods in the diagnosis. Diagnosis can be delayed in the absence of proper mycological laboratory techniques.

References

1Chirambo MC, Tielsch JM, West KP, Katz J. Blindness and visual impairment in southern Malawi. Bull WHO 1986;64:567-72.
2Chirambo MC. Causes of blindness among students in blind school institutions in a developing country. Br J Ophthalmol 1976;60:665-68.
3Rapoza PA, West SK, Katala SJ, Taylor HR. Prevalence and causes of vision loss in Central Tanzania. Int Ophthalmol 1991;15:123-29.
4Brilliant LB, Pokhrel RP, Grasset NC, Lepkowski JM, Kolstad A, Hawks W, et al . Epidemiology of blindness in Nepal. Bull WHO 1985;63:375-86.
5Khan MU, Hague MR, Khan MR. Prevalence and causes of blindness in rural Bangladesh. Ind J Med Res 1985;82:257-62.
6Gilbert CE, Wood M, Waddel K, Foster A. Causes of chilhood blindness in East Africa: results in 491 pupils attending 17 school for the blind in Malawi, Kenya and Uganda. Ophthalmic Epidemiol 1995;2:77-84.
7Thylefors B, Negrel AD, Segaram PR, Dadzie KY. Available data on blindness (update 1994). Ophthalmic Epidemiology 1995;2:5-39.
8Liesegang TJ, Forster RK. Spectrum of microbial keratitis in South Florida. Am J Ophthalmol 1980;90:38-47.
9Hagan M, Wright E, Newman M, Dolin P, Johnson G. Causes of suppurative keratitis in Ghana. Br J Ophthalmol 1995;79:1024-28.
10Williams G, McClellan K, Billson F. Suppurative keratitis in Bangladesh: the value of Gram stain in planning management. Int Ophthalmol 1991;15:131-35.
11Dunlop AA, Wright ED, Howlader SA, Nazrul I, Husain R, McCellan, et al. Suppurative corneal ulceration in Bangladesh. A study of 142 cases examining the microbiological diagnosis, clinical and epidemiological features of bacterial and fungal keratitis. Aust NZ J Ophthalmol 1994;22:105-10.
12Srinivasan M, Gonzales CA, George C, Cevallos V, Mascarenhas JM, Asokan B, et al . Epidemiology and aetiological diagnosis of corneal ulceration in Madurai, South India. Br J Ophthalmol 1997;81:965-71.
13Gopinathan U, Garg P, Fernandes M, Sharma S, Athmanathan S, Rao GN. The epidemiological features and laboratory results of fungal keratitis: A 10-year review at a referral eye care center in south India. Cornea 2002;21:555-59.
14Bharathi MJ, Ramakrishnan R, Vasu S, Meenakshi R, Palaniappan R. Aetiological diagnosis of microbial keratitis in South India. Indian J Med Microbiol 2002;20:19-24.
15Leber T. Keratomycosis aspergillina als ursache von Hypopy on Keratitis. Arch Ophthalmol 1879;25:285.
16Thomas PA. Keratomycosis (mycotic keratitis). In: Hay RJ, editor . Bailliere's Clinical Tropical Medicine and C ommunicable Diseases . London: Bailliere, 1989;Vol 4, pp.269-86.
17Upadhyay MP, Karmachanya PCD, Koirala S, Tuladhar NR, Bryan LE, Smolin G, et al. Epidemiologic characteristics, predisposing factors, and etiologic diagnosis of corneal ulceration in Nepal. Am J Ophthalmol 1991;111:92-99.
18Coster DJ, Wilhelmus K, Peacock J, Jones BR. Suppurative keratitis in London. IV th Congress of the European Society of Ophthalmology. Royal Society of Medicine Internal Congress and Symposium Series No 40 . London, 1981:395-98.
19Asbell P, Stenson S. Ulcerative keratitis. Survey of 30 years laboratory experience. Arch Ophthalmol 1982;100:77-80.
20Carmichael TR, Wolpert M, Koornhof HJ. Corneal ulceration at an urban African hospital. Br J Ophthalmol 1985;69:920-26.
21Ormerod LD. Causation and management of microbial keratitis in subtropical Africa. Ophthalmology 1987;94:1662-68.
22Agarwal PK, Roy P, Das A, Banerjee A, Maity PK, Banerjee AR. Efficacy of topical and systemic itraconazole as a broad-spectrum antifungal agents in mycotic corneal ulcer: A preliminary study. Indian J Ophthalmol 2001;49:173-76.
23Agarwal V, Biswas J, Madhavan HN, Mangat G, Reddy MK, Saini JS, et al . Current perspectives in infectious keratitis. Indian J Ophthalmol 1994;42:171-91.
24Bennett JE. Diagnosis and treatment of fungal infections. In: Fauci AS, Braunwald E Isselbacher KJ, Wilson JD, Martin JB, Kasper DL, et al , editors. Harrison's Principles of Internal Medicine . 14 th ed. New York: McGraw-Hill;1998. Vol 1, pp 1148-54.
25O'Day DM. Fungal keratitis. In: Pepose JS, Holland GN, Wilhemus KR, editors. Ocular infections and immunity . St.Louis: Mosby;1997. p 263-64.
26Thomas PA, Ravi CMK, Rajasekaran J. Microbial keratitis: a study cases and review of the literature. Journal of Madras State Ophthalmic Association 1986;23:13-21.
27Kalavathy CM, Thomas PA, Rajasekaran J. Spectrum of microbial keratitis in south India. Journal of Madras State Ophthalmic Association 1985;22:37-48.
28Sharma S, Srinivasan M, George C. The current status of Fusarium species in mycotic keratitis in south India. Indian J Med Microbiol 1993;11:140-47.
29Ansons AM. Corneal ulceration caused by penicillin-resistant Neisseria gonorrhoeae . Arch Ophthalmol 1987;105:1325.
30Reddy M, Sharma S, Rao GN. Corneal ulcer. In: Dutta LC, editor. Modern Ophthalmology . 2 nd ed, New Delhi: Jaypee Brothers Medical Publishers; 2000. pp 200-216.
31Jones DB, Liesegang TJ, Robinson NM. Laboratory Diagnosis of Ocular Infections . Washington DC: Cumitech 13, American Society for Microbiology; 1981.
32Sharma S, Athmanathan. Diagnostic procedures in infectious keratitis. In: Nema HV, Nema N, editors. Diagnostic Procedures in Ophthalmology . New Delhi: Jaypee Brothers Medical Publishers; 2002. pp 232-253.
33Ananthanarayanan R, Paniker CKJ. Textbook of Microbiology . 3rd ed. Madras: Orient Longman Ltd; 1986. pp 570-587.
34Cooper BH. Taxonomy, classcification, and nomenclature of fungi. In: Lennette EH, Balows A, Hausler WJ, Shadomy HJ, editors. Manual of clinical microbiology . 4th ed. Washington, D.C: American Society for Microbiology;1985. pp 495-499.
35Chander J, Sharma A. Prevalence of fungal corneal ulcers in Northern India. Infection 1994;22:207-9.
36Dutta LC, Dutta D, Mohanty P, Sharma J. Study of fungal Keratitis. Indian J Ophthalmol 1981;29:407-40.
37Varenkar MP, Shubhangi B, Pinto MJM, Naik PA. Study of mycotic keratitis in Goa. Indian J Med Microbiol 1998;16:58-60.
38Kotigadde S, Ballal M, Jyothirlatha, Kumar A, Rao SPN, Shivananda PG. Mycotic keratitis: a study in coastal Karnataka. Indian J Ophthalmol 1992;40:31-33.
39Mohanty PK, Ambekar VA, Deodhar LP. Prevalence of mycotic infections of external eye. Indian J Ophthalmol 1984;32:153-55.
40Jones DB. Decision-making in the management of microbial keratitis. Ophthalmology 1981;88:814-20.
41Musch DC, Sugar A, Meyer RF. Demographic and predisposing factors in corneal ulceration. Arch Ophthalmol 1983;101:1545-48.
42Ormerod LD, Hertzmark E, Gomez DS, Stabiner RG, Schanzlin DJ, Smith RE. Epidemiology of microbial keratitis in Southern California: A multivariate analysis. Ophthalmology 1987;94:1322-33.
43Forster RK, Rebell G, Wilson LP. Dematiaceous fungal keratitis: clinical isolates and management. Br J Ophthalmol 1975;59:372-76.
44Forster RK, Rebell G. The diagnosis and management of keratomycoses. Arch Ophthalmol 1975;93:975-78.
45Bharathi MJ, Ramakrishnan R, Vasu S, Meenakshi R, Chirayath A, Palaniappan R. Nocardia asteroides keratitis in south India. Indian J Med Microbiol 2003;21:31-36.
46Thylefors B. Epidemiological patterns of ocular trauma. Aust NZ J Ophthalmol 1992;20:95-98.
47Schaefer F, Bruttin O, Zografos L, Guex-Crosier Y. Bacterial keratitis: A prospective clinical and microbiological study. Br J Ophthalmol 2001;85:842-47.
48Forster RK. Fungal keratitis and conjunctivitis: Clinical disease. In: Somalin G, Thoft RA, editors. The Cornea. New York: Little, Brown and Company;1994. pp 239-252.
49Jones DB. Diagnosis and management of fungal keratitis. In: Tasman W, Jaeger EA, editors. Duane's Clinical Ophthalmology . Philadelphia: J. B. Lippincott company; 1994. Vol 4, Chapter 21.
50Kaufman HE, Wood RM. Mycotic keratitis. Am J Ophthalmol 1965;59:993-1000.
51Sharma S, Silverberg M, Mehta P, Gopinathan U, Agarwal V, Naduvilath TJ. Early diagnosis of mycotic keratitis: predictive value of potassium hydroxide preparation . Indian J Ophthalmol 1998;46:31-35.
52Wilson LA, Sexton RR. Laboratory diagnosis in fungal keratitis. Am J Ophthalmol 1968;66:646-53.
53Arffa RC, Avni I, Ishibashi Y, Robin J, Kaufman HE. Calcofluor and ink-potassium hydroxide preparations for identifying fungi. Am J Ophthalmol 1985;100:719-23.
54Reddy PS, Satyendran OM, Satapathy M, Vijaya HK, Reddy PR. Mycotic keratitis. Indian J Ophthalmol 1972;20:101-8.
55Vajpayee RB, Angra SK, Sandramouli S, Honavar SG, Chhabra VK. Laboratory diagnosis of keratomycosis: Comparative evaluation of direct microscopy and culture results. Ann Ophthalmol 1993;25:68-71.
56Srinivasan M, Sharma S. Nocardia asteroides as a cause of corneal ulcer. Arch Ophthalmol 1987;105: 464.
57Sharma S, Srinivasan M, George C. Acanthamoeba keratitis in non-contact lens wearers. Arch Ophthalmol 1990;108:676-78.