Citology Samples Methods

SAVE OFFLINE

DOI: 10.1111/vop.12574

ORIGINAL ARTICLE

Comparison of 3 corneal cytology collection methods for evaluating equine ulcerative keratitis: Cytobrush, kimura platinum spatula, and handle edge of scalpel blade Laura Proietto1 | Sarah S. Beatty2 | Caryn E. Plummer1 1 Departments of Small and Large Animal Clinical Sciences, College of Veterinary Medicine, University of Florida, Gainesville, FL, USA 2 Department of Comparative, Diagnostic, and Population Medicine, College of Veterinary Medicine, University of Florida, Gainesville, FL, USA

Correspondence C. E. Plummer Email: [email protected]

Abstract Purpose: To compare corneal cytology samples from three common sampling techniques: cytobrush (CB), Kimura platinum spatula (KS), and the handle edge of a scalpel blade (SB). Methods: Equine patients presenting to the University of Florida College of Veterinary Medicine with ulcerative keratitis were included. Following diagnosis of corneal ulcer and sampling for microbial culture, two cytology samples per technique were collected with sterile CB, KS, and SB in a randomized order. Cytologic evaluation was performed by two observers masked to collection method. Objective measures of sample cellularity, quality, distribution, and identification of infectious organisms were recorded per 10 monolayer cell populations using 509 magnification with oil immersion which were compared to culture results. Variables were compared using ANOVA with Student’s t test when appropriate and Cohen’s kappa (k) to evaluate inter- and intra-observer agreement (IOA) between observers and techniques. Results: Twenty equine patients (120 samples) were included. The IOA between observers was substantial (k = 0.75 ! 0.06) for cytological parameters. SB provided the most cellular samples (P < .01). There was a trend toward agreement (k = 0.12 ! 0.16) in technique for sample quality (P = .08). CB and SB had significantly poorer cell distribution than KS (P < .05). Infection was confirmed in 12 of 20 patients with SB and CB techniques having a significantly higher diagnostic yield than KS (P < .05) and was most consistent with infection confirmed on culture. Conclusions: The SB provided the most diagnostic samples but all three techniques are clinically useful in evaluating equine ulcerative keratitis. KEYWORDS cornea, cytology, diagnostics, exfoliative cytology, keratitis, ulcerative keratitis

1

| INTRODUCTION

Corneal ulceration is common in the horse and is a frequent cause for veterinary examination.1 Collecting samples for cytological evaluation and microbial testing is routine Veterinary Ophthalmology. 2019;22:153–160.

practice in investigating the underlying cause or extent of the lesion.2,3 Cytologic examination is a safe, practical, and relatively inexpensive diagnostic tool that provides useful information for efficient assessment and guidance of therapy.1,3-10 Based on the cellular population and

wileyonlinelibrary.com/journal/vop

© 2018 American College of Veterinary Ophthalmologists

|

153

154

|

identification of infectious organisms, an initial diagnosis can be made.5 Accurate and timely diagnoses are critical for instituting appropriate therapy to provide the best prognosis for the patient.1,10 As part of a complete evaluation, cytology should complement a fungal and aerobic microbial culture with susceptibility.5,10-12 Microbial culture is more sensitive than cytology for the confirmation and identification of infectious organisms but results are not available immediately such as with cytology.7 Results of aerobic culture and susceptibility may not be available for several days and up to two to 4 weeks for fungal isolation sometimes causing a devastating delay in appropriate therapy.7,12,13 To make an accurate diagnosis on cytologic examination, sufficient collection and preparation of samples are critical but also challenging.5,14 An ideal sample for cytologic evaluation is a well-distributed monolayer of adequate cellularity that is representative of the lesion and causes the patient minimal irritation when obtained.7,10 The collection of cells for corneal cytology has been described using numerous tools and techniques in various species and can be divided into the use of sterile swab, brush, or tool for scraping the ocular surface.1-3,5,7,15 Reported to be the least traumatic, swabs of various fiber types have been dismissed as a useful tool for harvesting corneal cytology.14-16 Cytologic collection with a swab has been significantly correlated with a low cell count as well as distortion of cellular morphology and detail leading to a low diagnostic yield.5,7,11,15,17 Impression cytology has also been described for assessing the ocular surface but is technically demanding and challenging to prepare reducing its usefulness in veterinary practice.2,10,18,19 A now commonly used method of harvesting cells is the use of a cytobrush or microbrush. Brush cytology has been consistently shown to obtain a sufficient number of cells arranged in a monolayer promoting accurate analysis.5,15,16 This is now considered the technique of choice for conjunctival, oral mucosal, and cervical cytology in human and veterinary medicine.6,10,14,20 The use of a scraping tool for exfoliative cytology also remains a common technique in both domestic species and human patients.18 The Kimura platinum spatula is still considered the gold standard for sample collection of the corneal surface in veterinary medicine.7 Iris repositioning spatulas, chemistry measuring spatulas, and the blunt handle edge of stainless steel scalpel blades are all described tools for this technique.5,7,15 Cytology collection via the scraping technique is considered the most precise but can cause distortion in the collected cells when applied to the slide.7 With all techniques, the operator must be careful of applying too much pressure to the compromised corneal

PROIETTO

ET AL.

surface as this can lead to iatrogenic trauma and corneal rupture.14 There are few reports of the use of corneal cytology in domestic species with most research focused on human reproductive and oral and ocular health.2,6,14,16,18,20-23 These diagnostic and screening strategies have been utilized in both specialized and nonspecialized veterinary practice for corneal lesions. Determining an optimal method for cytology collection and interpretation will provide guidelines for consistent evaluation of ulcerative lesions in the horse. The objective of this study was to compare the harvest and preparation of corneal cytology samples by the cytobrush (CB), Kimura platinum spatula (KS), and blunt handle edge of a stainless steel No. 10 surgical scalpel blade (SB). As part of this objective, it was sought to determine whether these samples were diagnostic when compared to the results of microbial culture. To the knowledge of the authors, this is the first study to compare these collection techniques simultaneously as well as determine interobserver agreement between evaluators of different levels of expertise in cytology evaluation.

2

| MATERIALS AND METHODS

This prospective single-center study was conducted at the University of Florida Large Animal Hospital (UFLAH) with approval from the Institutional Animal Care and Use Committee (IACUC).

2.1

| Animals

Animals were client-owned equine patients evaluated by the University of Florida College of Veterinary Medicine Ophthalmology Service between February and September of 2016. A clinical diagnosis of ulcerative keratitis, the indication for corneal cytology, and microbial culture as well as written client consent was required for inclusion. Patients that received therapy prior to examination were not excluded from the study.

2.2

| Examination

Ulcerative keratitis was diagnosed following a thorough sedated ophthalmic examination. Chemical sedation was achieved with 0.01-0.03 mg/Kg detomidine hydrochloride (Dormosedan!, 10 mg/mL, ZoetisTM, Parsipanny NJ) intravenously at the discretion of the clinician. Blockade of the auriculopalpebral nerve was accomplished with a threemL subcutaneous injection of 2% lidocaine hydrochloride 20 mg/mL solution (Lidocaine HCl, Vet One!, MWI Animal Health, Boise, ID). If superficial or stromal corneal

PROIETTO

|

ET AL.

155

tissue loss was evident on slit-lamp biomicroscopy, a sterile culturette swab (BDTM BBL CultureSwab!, Franklin Lakes NJ) was first carefully used to obtain a sample for microbial and fungal culture and susceptibility. Using a three-mL syringe, 0.25 mL of topical 0.5% tetracaine hydrochloride ophthalmic solution (Tetracaine HCl, Bausch + Lomb, Rochester NY) was instilled to the eye of interest and any ocular discharge was gently removed through irrigation with sterile eyewash. An additional instillation of topical tetracaine solution was performed as necessary for patient comfort prior to further examination. Samples for corneal cytology were collected as described below, and the remainder of the ophthalmic examination was performed to include fluorescein stain, applanation tonometry when indicated and direct ophthalmoscopy.

2.3 | Cytology sample collection and preparation Clean glass microscope slides (StarfrostTM, Platinum Line, Germany) were labeled with the date, patient medical record number, site of cytology sample collection, and order of collection. This information was recorded separately in addition to the order of technique used for use in data analysis. All samples for cytology were collected by the same clinician (LRP). A total of seven samples were collected from each ulceration including two samples per technique for general analysis with Wright–Giemsa stain and lastly, one for gram stain to be submitted for analysis separately. The order of collection was determined before beginning data collection and randomized using a random number generator (Microsoft! Excel 2016). The collection instruments used included a cytology brush (CB; Microbrush, Grafton, WI), Kimura platinum spatula (KS; Bausch + Lomb©, Storz! ophthalmic instruments), and handle edge of a sterile surgical steel No. 10 disposable scalpel blade (SB; Aspen Surgical!, Caledonia, MI) as shown in Figure 1. Briefly, the bristled tip of the CB was turned three to four revolutions over the lesion in the same direction before gently rolling on the slide.15 The KS was scraped two to three times in the same direction over the surface and margin of the ulcer not to exceed a 45 degree angle between the edge and arc of cornea.5,7,15,23 The collected material was then carefully smeared on a slide with the blade of the spatula.5,7 The SB was used in a similar fashion with the flat handle edge of the blade contacting the corneal surface and gently scraped two to three times before transfer of material to a slide.5,18,23 The wrapper of the blade itself was used as a handle folded backwards to cover the cutting surface of the blade. All samples were submitted to the University of Florida Veterinary Diagnostic Laboratories for routine automated

F I G U R E 1 Cytobrush (CB), Kimura platinum spatula (KS), and blunt edge of # 10 stainless steel surgical blade (SB) used for collection of corneal cytology

processing and plating for microbial and fungal culture, gram stain, and real-time general cytologic evaluation by a board-certified clinical pathologist using methanolic Wright–Giemsa stain. The submitted slides were then collected and pooled together for evaluation in this study.

2.4

| Cytologic evaluation

Cytologic evaluation was performed by a board-certified veterinary clinical pathologist (SSB) and veterinary ophthalmology resident (LRP) with light microscopy and oil immersion (CargilleTM Type B immersion oil, CargilleSacher Laboratories Inc, Cedar Grove, NJ). Microscopy was performed with 109 magnification eyepiece lenses and 49, 109, 509 oil, and 1009 oil objective lenses. Evaluators performed their review of the slides separately and were masked to the sampling technique. Samples from multiple ulcers were evaluated at one time. All data were recorded on a standardized worksheet including sample cellularity (number of intact epithelial cells and their stages of maturation), quality (intact:fragmented cells) of sample, cellular distribution (number of multilayers), and the presence of microbial or fungal elements identified under 509 oil magnification. If infectious organisms were present, their classification was recorded as cocci, diplococci, rods, bacilli, and fungal hyphae as identified under 1009 oil immersion. These morphological characteristics were then compared with results of bacterial and/or fungal cultures.

156

2.5

|

PROIETTO

| Statistical analysis

Statistical analysis of variables and effect of the order of sample collection between the three different collection techniques was performed using a one-way analysis of variance (ANOVA) with subsequent t test where appropriate for pairwise comparisons to determine which groups were different. Cohen’s k coefficient was calculated to evaluate the interobserver agreement (IOA) between the microscopic evaluation conducted by SSB and LRP for the criteria adopted to evaluate collection techniques.2,4,24 Cohen’s k coefficient was also used to compare the agreement between collection methods within the same observer. Agreement was considered absent if k ≤ 0.2, slight if 0.21 ≤ k ≤ 0.4, moderate if 0.41 ≤ k ≤ 0.6, substantial if 0.61 ≤ k ≤ 0.8, and good if it exceeds k > 0.8.2,4,24 Sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values were calculated for each method. A Pvalue of ≤.05 was considered significant for all statistical analyses. Statistical analysis was performed using Microsoft Excel 2016 Data Analysis (Microsoft! Excel 2016).

3

| RESULTS

Twenty horses with ulcerative keratitis were included in the study with a total of 120 samples collected for corneal cytology. No complications were encountered during sample collection, and less than one minute per technique was spent harvesting cells for cytology in all cases. On ophthalmic examination, all 20 ulcers were positive for retention of fluorescein stain. Applanation tonometry was performed in five of 20 patients with a mean of 12.8 mm Hg ! 3.65 mm Hg with no difference between ulcerated and nonulcerated globes (P = .07). Other findings on examination were consistent with changes related to corneal ulceration such as corneal edema, aqueous flare, and miosis without evidence of significant lenticular or fundic abnormalities.

3.1

| Cellularity of sample

There was significant agreement between SSB and LRP (k = 0.75, 95% CI 0.63-0.87) in assessing cellularity. The number of intact cells was significantly higher when using the SB method per evaluation of the 120 slides (P < .01), and IOA between methods for the same group of samples (ie, patient) was only slight. The order of collection was not significant (P = .34). There were 28 of 120 samples (23.33%) that were not of sufficient cellularity for analysis (P = .02). Of these, use of the KS resulted in the most nondiagnostic samples (15/40, P < .01).

ET AL.

There was no significant difference between techniques and the collection of epithelial cells from different cell layers and stages of cell maturation (P = .16). All techniques successfully resulted in collection of epithelial cells from the deepest, middle, and most superficial aspects of the ulcerative lesion as determined by the number of basal, intermediate, and mature epithelial cells present.

3.2

| Quality of sample

There was no significant difference in the ratio of intact to fragmented cells per collection technique (P = .63) or significance to the order of collection (P = .76) in the pooled sample population. When techniques were compared among the same patient, there was no agreement between techniques and moderate to substantial agreement between observers using the same parameter for evaluation of sample quality (Table 1).

3.3

| Distribution of sample

For all 120 slides, there was a significantly higher number of multilayered cells per sample for CB and SB compared to KS (P = .02 and P < .01, respectively). The mean number of multilayers is shown in Table 2, and overall cell distribution can be appreciated in representative samples from one patient in Figure 2. Within the same patient, the order of collection technique was not significant in determining the number of multilayers on the slide (P = .23), and there was only slight agreement between collection methods for the distribution of cells (k = 0.21).

3.4

| Infectious organism identification

Of the 20 patients with corneal ulcers, 12 (60%) were positively identified on culture as having either a bacterial (n = 6), fungal (n = 3), or mixed bacterial and fungal infection (n = 3). Morphologic identification of organisms

T A B L E 1 When techniques were compared within the same group of samples (ie, patient), there was no IOA between techniques for the same patient and moderate to substantial agreement between observers Cytobrush CB (Mean ! 95% CI)

Kimura Spatula KS (Mean ! 95% CI)

Scalpel Blade SB (Mean ! 95% CI)

IOA (k) between techniques

0.09 ! 0.33

0.28 ! 0.09

0.12 ! 0.16

IOA (k) between observers

0.56 ! 0.67

0.75 ! 0.06

0.58 ! 0.16

PROIETTO

|

ET AL.

on cytologic evaluation was consistent with culture results in 18 of 20 cases (90%). Collection method was significant for the identification of infectious organisms (P ≤ .01) but infectious organisms were not identified in six of 24 CB, nine of 24 KS, and four of 24 SB samples from the 12 infected ulcers. The SB was significantly more reliable in yield of infectious organisms than the CB which had a higher diagnostic yield than the KS. The SB has the highest sensitivity and NPV, and the KS has the highest specificity and PPV as shown in Table 3. Order of sample collection was not significant (P = .76), and 30 of 120 T A B L E 2 Cellularity and quality of samples for cytologic evaluation using each technique

Technique

Multilayers (Mean ! 95% CI)

Intact cells (Mean ! 95% CI)

Int:Frag Cells (Mean ! 95% CI)

Cytobrush CB

5.73 ! 1.07*

89.56 ! 19.35

3.02 ! 0.94

Kimura spatula KS

4.26 ! 0.96

82.56 ! 21.98

4.65 ! 3.91

Scalpel blade SB

5.88 ! 0.83*

144.64 ! 24.55*

3.95 ! 0.80

*P < .05.

157

samples were considered nondiagnostic (low cell number or not representative of culture results).

4

| DISCUSSION

To the knowledge of the authors, this is the first comparison of the three collection techniques (CB, KS, and SB) routinely used in clinical practice for cytologic sampling of ulcerative keratitis in veterinary patients. All techniques were easily and quickly performed without complication. The blunt edge of a No. 10 or No. 15 disposable SB has readily been used for the collection of corneal cells for cytology in both human and veterinary medicine.18 Use of the SB produced significantly more cellular samples compared to both the CB and the KS. Although there was more cell overlap of the SB and CB compared to the KS, more meaningful interpretation could be performed due to the cellularity of the sample. The SB was also superior in collecting a sample for accurate detection of infectious organisms. A SB is readily available, already in the inventory of most practices, packaged as a sterile instrument, and highly affordable.3,7,8 Although sharp compared to both the CB and the KS, careful and diligent use of the handle side should not result in greater damage or disruption to the wound or the cornea. Regardless of collection method,

F I G U R E 2 Cytology collected from a corneal ulcer of a thoroughbred mare with ulcerative keratitis and bacterial sepsis. These samples were from the same patient viewed with 10x (A-C), and 509 oil magnification (D-F) obtained via cytobrush (A,D), Kimura platinum spatula (B, E), and scalpel blade (C,F) methods. Scale bar represents 250 lm. Stages of epithelial cell maturation also indicated with thin arrow (parabasal/ basal), thick arrow (intermediate/wing), and arrowhead (superficial)

158

|

PROIETTO

ET AL.

T A B L E 3 SB has the highest sensitivity and negative predictive value (NPV), and KS has the highest specificity and positive predictive value (PPV) compared to results of bacterial and fungal culture after analysis of 10 serial monolayer cell populations under 509 oil immersion Technique Cytobrush CB Kimura spatula KS Scalpel blade SB

Sensitivity % (Mean ! 95% CI)

60.87 ! 19.42 50 ! 20.88

70.83 ! 16.55

Specificity % (Mean ! 95% CI) 87.5 ! 10.95 100 ! 20.59

93.75 ! 6.09

PPV % (Mean ! 95% CI) 87.5 ! 8.88

100

94.44 ! 4.70

NPV % (Mean ! 95% CI)

Nondiagnostic %

60.87 ! 11.83

25

68.18 ! 12.01

12.5

57.14 ! 9.41

37.5

KS produced the most nondiagnostic samples.

gentle pressure on the globe with the instrument, sedation, inclusion of an auriculopalpebral blockade, local analgesia, and adequate eyelid restraint with careful positioning should always be practiced.7,8,14,23 The CB provided reliable sample quality, which is consistent with prior reports comparing spatulas, swabs, impression cytology, and lavage for various tissue sample collections for cytologic evaluation.5,7,11,15,20 Bristles provide a greater surface area for sample collection and the ability to more safely reach deeper portions of the lesion.16,20 Multiple samples can be obtained with comparatively less risk of causing surface damage to the eye.10 Just as in this study with regard to identifying infectious organisms, a previous study comparing use of a metal spatula or CB for detection of oral neoplasia in human patients showed the superiority of the CB with fewer false-negative samples.20 The KS provided an equivalent cellularity to the CB though both were inferior to the SB. Significantly more samples collected with the KS were considered nondiagnostic as they were not of enough cellularity or were not representative of the lesion of interest. This is consistent with a pilot study where 40% of samples collected with the KS were nondiagnostic compared to the 37.5% in this study.25 The KS also produced the fewest number of multilayered samples, which is contradictory to a study where the KS technique resulted in more cell overlap than the CB.20 The KS is more expensive than either the CB or the SB and though reusable, must be flame sterilized or autoclaved between uses.8 Despite the consideration of the KS to be gold standard for cytology sampling, there are limited reports of its specific use and utility in scientific literature. More than one preparation did not affect the diagnostic utility of samples that were harvested after the first or second sample. Seven samples including one for gram stain were obtained per lesion in this study, and there was no bias due to order of collection in any parameter measured. Each collection method resulted in the presence of epithelial cells from the three stages of maturation indicating that they each reach an appropriate depth for the harvest of cells, and the cellular population is not altered with multiple sampling.

Infection was confirmed in 12 of the 20 microbial cultures from patients with ulcerative keratitis with 90% of the adequate cytology samples supportive of this finding. It has previously been reported that corneal cytology had a sensitivity of 49% with a specificity of 80% when compared to culture in analysis of a limited number of monolayer evaluations such as in this study.26 The mean sensitivity and specificity in this prospective study were comparable at 60.7% and 93.75%, respectively. The final cytology reports from the pathology service had a sensitivity and specificity that were higher at 91.67% and 100%, respectively, after reviewing the entire slide in a clinical setting. Cytology has also previously been reported to correlate with culture results with a PPV and NPV of 73% and 52% in the horse which is similar to what was found in the present study.26 The pathogenic organisms identified here were consistent with previously reported pathogens and flora of the normal and diseased equine cornea.13,27-29 Of the infected corneal ulcers, 50% were bacterial, 25% were fungal, and 25% had a mixed population of bacterial and fungal pathogens present. Mixed bacterial keratitis and keratomycosis have been reported to lead to the most severe corneal disease.27 These types of infections reportedly occur in 30% of infectious keratitis cases which is consistent with the 25% of cases reported here.9,27,30 Accurate identification of offending pathogens can be challenging even utilizing both cytology and culture; therefore, guidelines for obtaining the most consistent samples are very important. In a previous study of 48 animals with ulcerative keratitis, only 26 of 35 infections were identified on cytology and 29 were identified via culture.31 In one case report, a diagnosis of Histoplasma was made with cytology though negative on microbial culture.17 The diagnosis provided with cytologic evaluation allowed for rapid implementation of appropriate therapy and salvage of the globe.17 In this study, 28 (23.33%) samples were not of adequate cellularity for evaluation and 30 (25%) were not representative of the disease as they were inconsistent with culture findings. There are many sources of confusion in cytologic evaluation that can lead to misdiagnosis and inefficacious therapy. Proper sampling technique is essential to accurate cytologic evaluation and best performed in a systematic

PROIETTO

ET AL.

fashion. Following sampling for cytology with any collection technique samples should be fixed immediately once applied to clean slides to reduce confusion from pre-existent debris or distortion of cells and loss of intracellular detail.7,8,21 Scrolled epithelial cells and vegetative material can be mistaken for fungal elements, and melanin granules or stain precipitation can give a false impression of a bacterial infection.7,8,21 Microscopic evaluation is most accurately performed in regions where the cellular population is not dense or sparse but distributed in a monolayer representative of the lesion.4 Results are also dependent on the operator, their proficiency in sample collection with the technique used, slide preparation, and evaluation.14 To assess this, the interobserver agreement (IOA) was calculated for each parameter between the clinicians (SSB and LRP) and between techniques per patient per clinician.24 There was moderate to substantial agreement between clinicians in cytologic evaluation and slight to no agreement between methods per patient samples. Because of this, it is important to use the technique that creates the most easily evaluated and interpreted sample. Agreement between a veterinary practitioner and clinical pathologist should be at least moderate (k = 0.6) for meaningful real-time evaluation of slides by the clinician that can then be submitted for assessment by a board-certified clinical pathologist. Poor agreement can lead to gross misdiagnoses and inappropriate therapy.4 The purpose of this investigation was to compare these three methods, and in doing so, identify the most accurate and consistent method of corneal cytology collection in the presence of an ulceration. Factors that should be considered when selecting an instrument for the collection of corneal cytology include ease of use, consistency, cost, and operator proficiency. Further evaluation would include verifying these results against clinicians from a more varied level of training, experience, and clinical background.

5

| CONCLUSION

All sampling techniques provided diagnostic samples but the Kimura platinum spatula yielded the lowest quality samples and is therefore of less diagnostic utility than the handle edge of a steel surgical blade or brush cytology. In our hands, the scalpel blade consistently provided the highest cellularity sample but is of equal clinical utility as the cytobrush. REFERENCES 1. Maggs DJ, Miller PE, Ofri R, eds. Basic diagnostic techniques. In: David JM, Paul EM, Ron O, eds. Slatter’s Fundamentals of Veterinary Ophthalmology. 4th edn. St Louis, MO: Elsevier Saunders; 2007:81-106.

|

159

2. Perazzi A, Bonsembiante F, Gelain ME, et al. Cytology of the healthy canine and feline ocular surface: comparison between cytobrush and impression technique. Vet Clin Pathol. 2017;46(1):164-171. 3. Gilger BC. Equine ocular examination: routine & advanced diagnostic techniques. In: Gilger BC, ed. Equine Ophthalmology, 2nd edn. St Louis, MO: Elsevier Saunders; 2011:1-51. 4. Lanaux TM, Rozanski EA, Simoni RS, et al. Interpretation of canine and feline blood smears by emergency room personnel. Vet Clin Pathol. 2011;40(1):18-23. 5. Elizabeth A, Giuliano CPM. Eyes and ocular adnexa. In: Rick LCRDT, ed. Diagnostic Cytology and Hematology of the Horse. 2nd edn. St Louis, MO: Mosby Inc; 2002:43-64. 6. Cakir LGK, Kutsal O, Tunc AS. Evaluation of brush cytology (cytospin technique) and cultural results in the diagnosis of keratoconjunctivitis in a goat herd. Ankara Univ Vet Fak Derg. 2014;61:35-41. 7. Featherstone HJ, Heinich CL. Part 1: ophthalmic examination and diagnostics. In: Gelatt KN, Gilger BC, Kern TJ, eds. Veterinary Ophthalmology. 5th edn. Ames, IA: Wiley Blackwell; 2013:533-613. 8. Dwyer A. Practical general field ophthalmology. In: Gilger BC, ed. Equine Ophthalmology, 2nd edn. St Louis, MO: Elsevier Saunders; 2011:52-92. 9. Cutler TJ. Corneal epithelial disease. Vet Clin North Am Equine Pract. 2004;20:319-343, vi. 10. Yagmur M, Ersoz C, Ersoz TR, Varinli S. Brush technique in ocular surface cytology. Diagn Cytopathol. 1997;17(2):88-91. 11. Crispin SM. Examination of the feline eye and adnexa. EJCAP. 2007;17:1-15. 12. Lynne S, Sandmeyer BSB. Kate Robinson, Bruce H. Grahn. Diagnostic Ophthalmology. The. Can Vet J. 2014;55(3):281-283. 13. Brooks DE, Andrew SE, Biros DJ, et al. Ulcerative keratitis caused by beta-hemolytic Streptococcus equi in 11 horses. Vet Ophthalmol. 2000;3:121-125. 14. Bohn AA, Ferris RA, McCue PM. Comparison of equine endometrial cytology samples collected with uterine swab, uterine brush, and low-volume lavage from healthy mares. Vet Clin Pathol. 2014;43(4):594-600. 15. Bauer GA, Spiess BM, Lutz H. Exfoliative cytology of conjunctiva and cornea in domestic animals: a comparison of four collecting techniques. Vet Comp Ophthalmol. 1996;6:181-186. 16. Cocchia N, Paciello O, Auletta L, et al. Comparison of the cytobrush, cottonswab, and low-volume uterine flush techniques to evaluate endometrial cytology for diagnosing endometritis in chronically infertile mares. Theriogenology. 2012;77:89-98. 17. Richter M, Hauser B, Kaps S, Spiess BM. Keratitis due to Histoplasma spp. in a horse. Vet Ophthalmol. 2003;6:99-103. 18. Athmanathan S, Bandlapally SR, Rao GN. Collection of corneal impression cytology directly on a sterile glass slide for the detection of viral antigen: an inexpensive and simple technique for the diagnosis of HSV epithelial keratitis - a pilot study. BMC Ophthalmol. 2001;1:3. 19. Zhu YF, Zheng LB, Yao YF. Impression cytological study for ocular surface disorders of late stage eye burns. Eur Rev Med Pharmacol Sci. 2016;20:605-612. 20. Nanayakkara PG, Dissanayaka WL, Nanayakkara BG, Amaratunga EA, Tilakaratne WM. Comparison of spatula and cytobrush cytological techniques in early detection of oral malignant and premalignant lesions: a prospective and blinded study. J Oral Pathol Med. 2016;45:268-274.

160

|

21. Moosa NY, Khattak N, Alam MI, et al. Comparison of cervical cell morphology using two different cytology techniques for early detection of pre-cancerous lesions. Asian Pac J Cancer Prev. 2014;15:975-981. 22. Fujihara TTT, Saito K, Kitajima Y, Kobayashi TK, Tsubota K. Evaluation of human conjunctival epithelium by a combination of brush cytology and flow cytometry: an approach to the quantitative technique. Diagn Cytopathol. 1997;17:456-460. 23. Clode AB. Corneal cytology & culture collection. NAVC Clinician’s Brief. 2012;2012:75-77. 24. Landis JR, Kock GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174. 25. LLR Proietto, CE Plummer, SB Beatty, editor. Comparison of the cytobrush, kimura platinum spatula, and blunt scalpel edge corneal cytology collection techniques for evaluation of equine ulcerative keratitis: a pilot study. ACVO Annual Conference; Monterey, CA; 2016. 26. U Jeffery, K Gervais, F Mowat, S Hostetter, Whitley D, editor. Ability of corneal cytology to predict bacterial culture results. American Society for Veterinary Clinical Pathology (ASVCP) 47th Annual Meeting; 2012; Seattle, WA, 2012. 27. Moore CP, Fales WH, Whittington P, Bauer L. Bacterial and fungal isolates from Equidae with ulcerative keratitis. J Am Vet Med Assoc. 1983;182:600-603.

PROIETTO

ET AL.

28. Whitley RD, Elizabeth CB, Moore CP. Microbial isolates of the normal equine eye. Equine Vet J. 1983;15(S2):138-140. 29. Whitley RDMC. Microbiology of the equine eye in health and disease. Vet Clin North Am Large Anim Pract. 1984;6:451466. 30. Andrew SE, Brooks DE, Smith PJ, Gelatt KN, Chmielewski NT, Whittaker CJ. Equine ulcerative keratomycosis: visual outcome and ocular survival in 39 cases (1987-1996). Equine Vet J. 1998;30:109-116. 31. Massa KLMC, Hartmann FA, Miller PE, Korsower CS, Young KM. Usefulness of aerobic microbial culture and cytologic evaluation of corneal specimens in the diagnosis of infectious ulcerative keratitis in animals. J Am Vet Med Assoc. 1999;215:1671-1674.

How to cite this article: Proietto L, Beatty SS, Plummer CE. Comparison of 3 corneal cytology collection methods for evaluating equine ulcerative keratitis: Cytobrush, kimura platinum spatula, and handle edge of scalpel blade. Vet Ophthalmol. 2019;22:153–160. https://doi.org/10.1111/vop.12574