• Users Online: 3038
  • Home
  • Print this page
  • Email this page

   Table of Contents      
Year : 1998  |  Volume : 46  |  Issue : 1  |  Page : 31-35

Early diagnosis of mycotic keratitis : Predictive value of potassium hydroxide preparation

L.V. Prasad Eye Institute, Hyderabad, India

Correspondence Address:
S Sharma
L.V. Prasad Eye Institute, Hyderabad
Login to access the Email id

Source of Support: None, Conflict of Interest: None

PMID: 9707845

Rights and PermissionsRights and Permissions

Potassium hydroxide (KOH) preparation is an underutilized modality in the diagnosis of mycotic keratitis. We have earlier shown its utility in the diagnosis of Nocardia and Acanthamoeba keratitis. The aim of this study was (i) to evaluate the sensitivity, specificity and predictive value of KOH preparation, and (ii) to compare its efficacy with other methods of corneal scraping examination, for the diagnosis of mycotic keratitis. The study was conducted in two phases. In phase I, randomized corneal scrapings were examined by KOH, Gram's stain, and lactophenol cotton blue (LPCB) in 91 infectious keratitis subjects. In phase II, 53 corneal scrapings were stained with KOH and calcofluor white (CFW), and viewed with bright field (KOH) and fluorescence (CFW) microscopy. The KOH and CFW readings were recorded by an observer masked to the clinical findings and culture results. Nineteen scrapings were examined by two masked observers. In 22 culture positive fungal keratitis patients in phase I, the sensitivity of KOH, Gram's stain, and LPCB methods was 100%, 86.4%, and 77.3%, respectively. In phase II, the specificities of KOH and CFW were identical (83.8%), while the sensitivities were 81.2% and 93.7%, respectively (p = 0.59), in 16 culture positive mycotic keratitis patients. There was no significant difference between the negative and positive predictive values of KOH and CFW. Furthermore, no significant interobserver variability was found in the specificity and sensitivity. The KOH method compares well with other microscopy methods in the diagnosis of keratomycosis and has a definite place in the armamentarium of diagnostic techniques.

Keywords: Pottasium hydroxide, calcofluor white, diagnosis, mycotic keratitis

How to cite this article:
Sharma S, Silverberg M, Mehta P, Gopinathan U, Agrawal V, Naduvilath T J. Early diagnosis of mycotic keratitis : Predictive value of potassium hydroxide preparation. Indian J Ophthalmol 1998;46:31-5

How to cite this URL:
Sharma S, Silverberg M, Mehta P, Gopinathan U, Agrawal V, Naduvilath T J. Early diagnosis of mycotic keratitis : Predictive value of potassium hydroxide preparation. Indian J Ophthalmol [serial online] 1998 [cited 2022 Dec 5];46:31-5. Available from: https://www.ijo.in/text.asp?1998/46/1/31/14980


Click here to view

Click here to view

Click here to view

Click here to view

Click here to view

Click here to view
Diagnosis and treatment of mycotic keratitis is one of the most difficult problems encountered by ophthalmologists. The potential to mimic other forms of infective keratitis, worsening with inadvertent use of steroids, and recalcitrant course of the disease are common dilemmas in the diagnosis and treatment of this entity.

Direct microscopic evaluation is the most valuable and rapid diagnostic tool for the detection of fungal elements in corneal scrapings.[1] Several techniques have been used for the direct microscopic identification of fungi from corneal scrapings. Giemsa and Gram's stain techniques are equally sensitive in detecting fungal elements.[2],[3] Gomori's methenamine silver and periodic acid-schiff stains are other methods used when conventional staining techniques yield no results. However, these methods are time-consuming, tedious, and difficult to interpret without considerable experience.[4] Though the use of calcofluor white (CFW)[5]and fluorescein conjugated lectins[6] yield rapid results, special infrastructural requirements make them inapplicable in most situations. Potassium hydroxide (KOH) has been used as a 10-20% suspension, either plain[7] or with ink[5] or with lactophenol cotton blue (LPCB).[8] However, the sensitivity of this method is highly variable, ranging between 33-94%.[9],[11] In order to evaluate its efficacy and comparability with other acclaimed methods, we undertook this study.

  Materials and Methods Top

This study was under taken in two phases. In the first phase, a prospective evaluation of the sensitivities of 10% KOH, LPCB, and Gram's stain was done. In the second phase, a double-masked evaluation of the sensitivity and specificity of KOH and CFW preparations was performed. Giemsa stain was not used in this study.

  Phase I Top


Between February and July 1993, 91 patients were clinically diagnosed to have microbial keratitis, based on history and slitlamp biomicroscopic examination. In all patients, scrapings from corneal ulcer were subjected to complete microbiological evaluation which included direct microscopic examination and inoculation into various culture media [Table - 1].


After instillation of local anesthetic (lignocaine hydrochloride 4%) drops in the affected eye of the patient, corneal scrapings were taken from the edge as well as base of the ulcer using a No. 15 Bard-Parker blade. Tissue scrapings were taken for a minimum of three direct microscopic preparations, the sequence for which was randomized. The first smear was that of KOH preparation in which 10% KOH containing 10% glycerol was used as the reagent. LPCB preparation was made using commercially available reagent. For both these preparations corneal scrapings were placed over a glass slide and a drop of the reagent was added over the tissue specimen and a coverslip was used, taking care to avoid trapping of air bubbles. Whenever tissue specimens appeared too thick for microscopy, the coverslip was gently pressed with a metal or glass rod to spread the tissue. For Gram's stain preparation, the specimen was smeared evenly over the slide, which was then fixed with 95% methyl alcohol for 5 minutes and stained by the standard Gram's staining method.

All three direct preparations were examined immediately under the microscope for the presence of fungal elements. The wet preparations of KOH and LPCB were examined under low magnification (xl25) initially and then under higher magnification (x500). The Gram's stained preparation was examined under oil immersion objective (x1,250).

Employing the gold standard of culture positivity, 22 culture positive cases of mycotic keratitis constituted the study group of phase I. The results of the three methods of direct examination were compared in these cases and the sensitivity of each was calculated. The statistical significance of the observations was assessed by the chi-square test applying the Yates correction factor.

  Phase II Top


Fifty consecutive patients with a clinical diagnosis of infectious keratitis were included in the study. All patients were subjected to microbiological investigation for etiological diagnosis, prior to treatment.


Corneal scrapings were obtained from all patients by the method described in phase I. The initial two scrapings were used for direct microscopic examination, and the subsequent scrapings were inoculated into various media [Table - 1].

Of the two scrapings used for direct evaluation, one was stained by Gram's technique and the other was prepared by adding one drop of 0.1% CFW (Polysciences, USA) with 0.1% Evans blue and one drop of 10% KOH.[12] A coverslip was placed over the specimen.

Fifty-three scrapings from 50 patients (more than one scraping from 2 patients), stained with KOH and CFW, were examined by an observer masked to the results of Gram's smear and culture findings. Initial observation was made under bright field microscopy, and the same smear was subsequently examined under fluorescence microscopy (Olympus BH2-RFC, Cube-BH2-DMU, DM400, EF20UG1, BF17L420). Nineteen out of 53 scrapings were also examined by a masked second observer under bright field and fluorescence microscopy. A scale of 0-3 was employed to record the quantity of fungal elements seen by both the observers.

The sensitivity and specificity of bright field observation (representing KOH) and fluorescence microscopy (representing CFW) were calculated from the results of 53 scrapings, while only 19 samples could be considered (because of non-availability of one observer) for the assessment of interobserver variability of KOH and CFW. Fisher's exact probability test was used to determine the significance of the differences in sensitivity and specificity of KOH and CFW.

  Results Top

Phase I

All 22 cases of keratomycosis were caused by filamentous fungi; Aspergillus species in 9, dematiaceous fungi in 9, and Fusarium species in 4 cases. Fungal filaments were seen in the corneal scrapings of all patients by the KOH method (100%) compared to 19 (86.4%) by LPCB and 17 (77.3%) by Gram's method. Though the percentage positivity was higher for KOH, the difference in sensitivity with LPCB and Gram's stain did not reach statistical significance (p=0.23, 0.057, respectively).

Phase II

Among 50 cases of infectious keratitis, 16 were culture-proven mycotic ulcers, with Fusarium solani in 5, Aspergillus species in 4, dematiaceous fungi in 5, and unidentified hyaline fungi in 2 cases.

The observations on smears and cultures of all 16 patients of mycotic keratitis are shown in [Table - 2]. While the sensitivity of KOH and CFW was 81.2% and 93.7%, respectively (p=0.59), the overall specificity of KOH and CFW was identical (83.8%) in the 53 scrapings. The positive predictive values of KOH and CFW were 68.4% and 71.4%, while the negative predictive values were 91.2% and 96.9%, respectively. There was no statistically significant difference between the predictive values of KOH and CFW (p=0.89 and 0.65, respectively).

Six additional patients, though culture negative, had a heavy load of fungal filaments (>10 filaments in each microscopic field, x250) under bright field and fluorescence microscopy. The KOH appearance of corneal scraping from one such patient is shown in the Figure. Fungal filaments were seen along with Acanthamoeba cysts and Chlamydospores. The corneal scrapings did not grow fungus but Acanthamoeba was grown in culture. In only one scraping of another patient, fungus was not seen in either KOH or CFW, but the cultures yielded dematiaceous fungus [Table - 2]. In two corneal scrapings fungal elements were detected in CFW but not in KOH.

Among 19 patients whose scrapings were examined by two masked observers, 7 were culture-positive fungal ulcers. In these 7 cases, the specificities of both KOH and CFW were the same (83.3%) for both observers. In addition, there was no significant interobserver variability in the sensitivities of KOH and CFW preparations [Table - 3].

  Discussion Top

Direct microscopic detection of fungal elements in corneal scrapings is the quickest modality of diagnosis of mycotic keratitis. Of the various methods available for demonstration of fungi in corneal scrapings, use of KOH is one of the oldest methods. It has been traditionally used by dermatologists to detect fungal elements in direct wet skin preparations. Proteinaceous components, such as host cells are partially digested by the alkali, leaving intact the polysaccharide-containing fungal cell walls. Glycerol (10%) or thimerosal (0.1%) is generally added to the KOH preparation in order to preserve the sample for a longer period. The KOH preparation has been utilized as a principal modality to examine corneal scrapings in suspected fungal ulcers by several investigators.[13][14][15] However, recently published reviews related to ocular infections have not shown enthusiasm about the utility of KOH in the diagnosis of keratomycosis.[16],[17] Several investigators in their anecdotal references in the literature have cited the sensitivity of KOH method to be lower than newer methods.[6],[9],[10],[17]

Our present study highlights the value of the traditional method of KOH wet film preparation in the diagnosis of mycotic keratitis. Our experience of over nine years has shown the role of KOH not only in the diagnosis of mycotic keratitis but also in the diagnosis of Nocardia[18] and Acanthamoeba.[19] In phase I of this study, fungus was identified in 100 percent of the cases with KOH compared to 86.4% with Gram's and 77.3% with LPCB preparations. Our results with LPCB preparation are comparable with those of an earlier study,[4] where fungal hyphae were seen in 78% of culture-proven cases of mycotic keratitis.

Results of the phase I study could have been affected by the difference in quantity of material collected and sampling from different sites of the ulcer of varying severity. In order to avoid this sampling variation the same sample was observed for KOH and CFW in the second phase. CFW method was selected for comparison for the following reasons: (i) it is one of the most acclaimed methods of fungus detection,[20] and (ii) it was technically possible to directly compare KOH and CFW using the same sample.

Of the 53 corneal scrapings studied by KOH and CFW, the sensitivity of fungus detection was 81.2% and 93.7%, respectively. However, the difference was not statistically significant. The specificity for both the methods was identical (83.8%). In addition, the difference between either the positive predictive value or the negative predictive value of both tests was not significant.

Surprisingly, of the 15 smears positive for fungal filaments in CFW, only two were deemed negative in KOH [Table - 2]. Though the sensitivity of the CFW was not significantly higher than KOH, the morphology of smaller fungal filaments was better appreciated in CFW than KOH. This observation underscores the fact that the efficacy of KOH was only marginally lower than CFW in detection of fungal elements.

In both phases of this study, a quantitation of fungal filaments was carried out. Less than 5 filaments per smear were seen in 4 of 22 cases (18.1%) in phase I, and 1 of 16 cases (6.2%) in phase II of the study. The remaining cases demonstrated heavy fungal material, indicating that the majority of the patients were in advanced stages of the disease. This could possibly account for the presence of detectable levels of fungal filaments in an unstained preparation such as KOH. While not denying that KOH is likely to be less sensitive in early fungal ulcers, especially caused by yeast-like fungi, we recommend the continued usage of this method as one of the modalities of direct corneal scraping examination in the diagnosis of mycotic keratitis.

The large number of fungal filaments observed in the corneal scrapings of the six culture-negative cases in our study questions the reliability of the gold standard (that is, culture positive) of laboratory diagnosis in mycotic keratitis. Similar observation has been made by others.[11] It seems likely that the filaments were non-viable at the time of culture in at least three of our cases who had received antifungal treatment prior to reporting at our institute. The cause for non-viable filaments in other cases is not known. While three of these cases resulted in healed corneas following antifungal therapy, one case was lost to follow up and one resulted in pthisis bulbi. In one patient the infection spread inside causing endophthalmitis and Aspergillus flavus was cultured from the vitreous.

From our experience, we conclude that KOH method is as sensitive as other conventional methods in the diagnosis of mycotic keratitis. In addition, we recommend that all smear-positive cases of mycotic keratitis be treated accordingly despite negative culture reports.

  References Top

O'Day DM, Akrabawi PL, Head WS, Ratnar HB. Laboratory isolation techniques in human and experimental fungal infections. Am J Ophthalmol 1979;87;688-93.  Back to cited text no. 1
Wilhelmus KR, Liesegang TJ, Osato MS, Jones DB. Cumitech 13A: Laboratory Diagnosis of Ocular Infections. Specter SC, coordinating editor. Washington, DC:American Society for Microbiology;1994.  Back to cited text no. 2
Jones DB. Initial therapy of suspected microbial cornealulcers. II. Specific antibiotic therapy based on corneal smears. Surv Ophthalmol 1979;24:97-116.  Back to cited text no. 3
Thomas PA, Kuriakose T, Kirupashanker MP, Maharajan VS. Use of lactophenol cotton blue mounts of corneal scrapings as an aid to the diagnosis of mycotic keratitis. Diagn Microbiol Infect Dis 1991;14:219-24.  Back to cited text no. 4
Arffa 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.  Back to cited text no. 5
Robin JB, Nielson S, Tronsdale MD. Fluorescein-conjugated lectin identification of a case of human keratomycosis. Am J Ophthalmol 1986;102:797-98.  Back to cited text no. 6
Sharma 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.  Back to cited text no. 7
Reddy PS, Satyendran OM, Satapathy M, Vijaya HK, Reddy PR. Mycotic keratitis. Indian J Ophthalmol 1972;20:101-8.  Back to cited text no. 8
Liesegang TJ, Forster RK. Spectrum of microbial keratitis in south Florida. Am J Ophthalmol 1980;90:38-47.  Back to cited text no. 9
Asbell P, Stenson S. Ulcerative keratitis: survey of 30 years' laboratory experience. Arch Ophthalmol 1982;100:77-80.  Back to cited text no. 10
Vajpayee RB, Angra SK, Sandramouli S, Honavar SG, Chhabra VK. Laboratory diagnosis of keratomycosis:comparitive evaluation of direct microscopy and culture results. Ann Opthalmol 1993;25:68-71.  Back to cited text no. 11
Baron EJ, Finegold SM. Optical methods for laboratory diagnosis of infectious diseases. In: Bailey and Scott's Diagnostic Microbiology. 8th edition. Philadelphia:CV Mosby;1990. P77.  Back to cited text no. 12
Forster RK, Rebell G, Wilson LA. Dematiaceous fungal keratitis: clinical isolates and management. Br J Ophthalmol 1975;59:372-76.  Back to cited text no. 13
Sandhu DK, Randhawa IS, Singh D. The correlation between environmental and ocular fungi. Indian j Ophthalmol 1981;29:177-82.  Back to cited text no. 14
Laverde S, Moncada LH, Restrepo A, Vera CL. Mycotic keratitis:5 cases caused by unusual fungi. Saboraudia 1973;11:119-23.  Back to cited text no. 15
Rosa RH, Miller D, Alfonso EC. The changing spectrum of fungal keratitis in south Florida. Ophthalmology 1994;101:1005-13.  Back to cited text no. 16
Rao NA. A laboratory approach to rapid diagnosis of ocular infections and prospects for the future. Am J Ophthalmol 1989;107:283-91.  Back to cited text no. 17
Srinivasan M, Sharma S. Nocardia asteroides as a cause of corneal ulcer. Arch Ophthalmol 1987;105:464.  Back to cited text no. 18
Sharma S, Srinivasan M, George C. Acanthamoeba keratitis in non-contact lens wearers. Arch ophthalmol 1990;108:676-78.  Back to cited text no. 19
Marines HM, Osato MS, Font RL. The value of calcofluor white in the diagnosis of mycotic and Acanthamoeba infections of the eye and ocular adnexa. Ophthalmology 1987;94:23-26.  Back to cited text no. 20


  [Figure - 1]

  [Table - 1], [Table - 2], [Table - 3]


    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  Materials and Me...
  In this article
Phase I
Phase II
Article Figures
Article Tables

 Article Access Statistics
    PDF Downloaded0    
    Comments [Add]    

Recommend this journal