Otitis externa in free-roaming cat population: Clinical examination, cytology, and mycology culture

Abstract

Background:

Free-roaming cats are a potential source of several infectious and parasitic diseases that can be easily transmitted to outdoor pet cats.

Aim:

The purpose of this study was to determine the prevalence and pattern of co-occurrence of causative agents of otitis externa and to identify associated risk factors in free-roaming cats living in the continental region of Croatia. In addition, we investigated the presence of Malassezia yeasts in the external ear canals of cats using mycology culture.

Methods:

Clinical assessment of otitis externa involved auricular and otoscopic examinations. Samples of cerumen and ear swabs were collected and analyzed via cytological examination and mycological culture.

Results:

The frequency of otitis externa in the population of free-roaming cats was 40.5% (30/74) based on the presence of Otodectes cynotis, Malassezia yeasts, cocci, rods, and/or neutrophils. Results revealed multiple diagnoses for the same ears and unevenly affected ears of the same cats. O. cynotis was detected in 25.7% (19/74) of the cats and was the primary cause of otitis in 63.3% (19/30) infected cats. Co-diagnosis of O. cynotis and Malassezia yeasts was common at both the ear (39.3%, 22/56 infected ears) and cat (40.0%, 12/30 infected cats) levels, while Malassezia was causing otitis in 30% (9/30) infected cats. Three or more diagnoses co-occurred in 15% of infected ears and 27% of infected cats. The clinical signs of otitis externa were absent in 8.9% (5/56) of infected ears, all affected by Malassezia alone or with O. cynotis. Malassezia yeasts were cultured in 83.3% (25/30) of the otitis cases, with heavy growth in 70% and Malassezia pachydermatis as the predominant species in all cases.

Conclusion:

The high prevalence of otitis externa caused by O. cynotis and Malassezia yeasts in free-roaming cats in Croatia provides important insights into the health status of this cat population. Although otitis externa and its causative agents are frequently found in free-roaming cats, the actual prevalence of otitis externa and all associated risk factors have not been thoroughly assessed.

Introduction

Feline otitis is a common health disorder associated with the presence of primary, secondary, and perpetuating factors (Brame and Cain, 2021). Reports indicate that otitis externa affects a significant proportion of stray, free-roaming, and feral cats, with prevalence rates ranging from 0.9% to 55.1% (Perego et al., 2014; Bollez et al., 2018; Tyler et al., 2019). The diagnosis of feline otitis is based on a history of disease and clinical examination followed by cytology, and sometimes cultural methods (Brame and Cain, 2021). In the populations of community cats, the clinician must rely only on clinical examination, otoscopy, and cytology to diagnose the disease since there is no available anamnesis. Moreover, unlike dogs in which the OTIS scoring system is used (Nuttall and Bensignor, 2014), no standardized grading system for otitis has been developed for cats. This makes it challenging to objectively include and interpret the signs of otitis, especially in cases of mild otitis or subclinical infection where most of the clinical signs may be absent. Cytology represents the second step in the diagnosis of otitis, but it is becoming increasingly complex because there is no unique interpretation of Malassezia and bacterial cell counts evaluated in ear swab samples from dogs, much less cats (Millward, 2018). Various cytologic interpretation criteria for Malassezia yeasts, coccoid, and rod-shaped bacteria in healthy cats and cats with otitis have been provided by previous studies using high power or immersion oil field magnification (HPF and OIF, respectively) (Ginel et al., 2002; Tater et al., 2003; Perego et al., 2014; Pressanti et al., 2014; Tyler et al., 2019). Additionally, Malassezia otitis is common in cats where lipophilic Malassezia pachydermatis is not the only species involved in its pathogenesis (Crespo et al., 2002; Shokri et al., 2010). Despite this fact, conventional cultural identification of all Malassezia yeasts is not routinely performed because the Sabouraud dextrose culture media (SDA) used in veterinary diagnostic laboratories supports the growth of only M. pachydermatis. Results from previous studies reported that a comparison between ear cytology and Malassezia culture methods for the diagnosis of otitis externa yielded low sensitivity and good specificity (Cafarchia et al., 2005). On the other hand, the diagnostic value of culture-based enumeration of Malassezia yeasts in canine dermatitis has been recognized for a long time (Bond et al., 1994), including a study proposing a quantitative polymerase chain reaction method to evaluate effective otitis treatment in dogs (Puig et al., 2019). The aim of this study was to determine the prevalence and pattern of co-occurrence of causative agents of otitis externa and to identify associated risk factors in free-roaming cats living in the continental region of Croatia based on observed clinical assessment and cytology. We also sought to determine the presence of Malassezia yeasts in the external ear canal using mycology culture. Part of this data was previously presented as an abstract (Kolenc et al., 2019).

Materials and Methods

Sample size and study population

No reports exist on the prevalence of otitis externa in free-roaming cats in Croatia. The reported prevalence in Italy, Belgium, and the United Kingdom varies from 0.9% to 55.1% (Perego et al., 2014; Bollez et al., 2018; Tyler et al., 2019). Based on this reported range and assuming the most conservative prior prevalence of 50%, the required sample size is 67 cats for the desired margin of error of at least 12% (at 95% confidence interval) (Dohoo et al., 2009). The study population comprised free-roaming cats that inhabited the city of Zagreb and the suburban area, captured as part of the trap-neuter-return project (TNP) in the spring, summer, and fall of 2019. We enrolled 75, of which one cat was not examined, resulting in a sample size of 74 for the study.

Collection of samples

The 74 enrolled cats were sampled by collecting cerumen samples and ear swabs. The location, breed, sex, and age of the animals, as well as the pregnancy data for intact females, were recorded. Based on the age estimation, the animals were divided into two groups: less than 1 year old (G1:<1 year) and older cats (G2: >1 year).

Clinical examination of the external ear canal and cytological evaluation of swab impressions

Pinna was assessed for the presence of alopecia, erythema, crust, ulceration, or swelling (aural hematoma) (Nuttall and Bensignor, 2014; Bollez et al., 2018; Englar, 2019). An otoscopy was performed to detect possible obstructions of the external ear canal with cerumen or exudate, foreign bodies, polyps, or neoplasia. The presence of cerumen was described according to its (i) color (light brown and dark brown), (ii) consistency (crumbly, waxy, greasy, and gelatinous; in the statistical analysis the presence or absence of crumbly appearance was captured in a variable ”earwax.discharge.crumbly’ with levels yes versus no), and (iii) amount as small, moderate, or large (which was for statistical analysis coded in a variable “earwax.discharge.crumbly” with levels 1–3, respectively) depending on the visualization of the external ear canal and the tympanic membrane. A mineral oil technique was used to examine cerumen samples for the presence of ear mites at low magnification. Swab impressions from the external ear canal were stained with Diff-Quik (BioGnost, Croatia), and counting of cells (Malassezia yeast cells, coccoid, rod-shaped bacteria, and neutrophils) was carried out in 10 different fields under OIF (Angus, 2004) and interpreted based on the counts (neutrophils) or the average of counts in the 10 fields (Malassezia, cocci, and rods).

The diagnosis of otitis externa at the ear level was made on the basis of cytological evaluation as follows: Otodectes positive if present ≥1 egg, juvenile or adult mite (Sotiraki et al., 2001; Perego et al., 2014) or neutrophil positive if ≥1 neutrophil is present. Additionally, cocci positive if present >5 coccoid bacteria, Malassezia positive if present >1 Malassezia yeast, and rods positive if present >1 rod-shaped bacterium per OIF (Coyner, 2020).

Culture of Malassezia yeasts

A swab was taken from each ear, plated onto modified Dixon agar, and incubated at 32°C for 2 weeks. The identification of Malassezia yeasts was based on the appearance of the colonies, which were yellowish to creamy in color and had a characteristic budding cell shape when viewed under a microscope. Lipid dependence was determined by observing SDA growth at 37°C for 72 hours (Gueho-Kellermann et al., 2010). The number of grown colonies on plates was expressed as colony-forming units (CFUs), and growth was classified as: scarce (<10 CFU), moderate (10–50 CFU), and heavy (>50 CFU) using a semi-quantitative method previously described (Crespo et al., 2002).

Statistical analyses

Otitis externa diagnoses at the ear level based on the five cytological evaluations described in the “Clinical examination of the external ear canal and cytological evaluation of swab impressions” section (i.e., the presence of Otodectes, Malassezia, cocci, rods, and neutrophils above a specified threshold) were the binary outcomes of interest. These outcomes were interpreted individually and combined. At the individual level, Otodectes, cocci, and Malassezia were statistically analyzed, although a small number of positive rod and neutrophil counts precluded their consideration as outcomes in statistical analysis. When combined, the five diagnoses were interpreted in parallel, meaning that an ear was considered positive for a “general otitis externa” outcome if it was positive for any one or more of the five diagnoses and was negative otherwise. The goals of statistical analyses were to (1) determine patterns of individual outcome co-occurrence and (2) identify explanatory variables associated with the occurrence of individual and combined outcomes. The considered explanatory variables (with corresponding levels) included: sex [female (F; baseline)], male (M), spayed female (SF), and neutered male (NM); pregnancy status [no (baseline) and yes]; age [G1:<1 year (baseline) and G2:>1 year]; sampling period [spring (baseline), summer, and autumn], and clinical signs in terms of the presence of a wound [“wound”: no (baseline) and yes]; stenosis [“stenosis”: no (baseline) and yes]; invisibility of the tympanic membrane [“TM.invisible” no (baseline) and yes]; the color of the earwax discharge [“earwax.discharge”:light/brown (baseline) and dark.brown]; the amount of earwax discharge [“earwax.discharge.amount”:1 (baseline), 2, and 3]; and the crumbly consistency of the earwax discharge [“earwax.discharge.crumbly”: no (baseline) and yes]. Additionally, binary variables for the presence/absence of Otodectes cynotis, Malassezia, cocci, rods, and neutrophils were considered explanatory variables for individual outcomes.

The binary individual and combined outcomes of interest and explanatory variables were subjected to descriptive analysis. We cross-tabulated the five individual diagnoses across both ears of a cat and across cats to determine the frequency of the co-occurrence of multiple diagnoses at the ear and cat levels. Additionally, for each individual outcome of interest, we interpreted whether the ear infection in a cat was unilateral (UNI) or bilateral (BI). These results revealed multiple diagnoses for the same ears and unevenly affected ears of the same cats. Therefore, subsequent univariable and multivariable statistical analyses of the associations of individual and combined outcomes with explanatory variables were conducted at the ear level. In the univariable analysis, categorical explanatory variables were tested using the chi-square test or Fisher’s exact test (as applicable), while a single continuous variable (weight) was tested using the Mann–Whitney U test (Wilcoxon test). Predictor variables displaying significant associations (p < 0.05) with outcomes at the univariable level were tested in multivariable logistic regression models. The final logistic regression model was selected through a stepwise backward elimination process using the likelihood ratio test and the Akaike information criterion. Two-way interactions between variables included in the final multivariable model were also assessed, but none were identified. The association between the outcome and a statistically significant explanatory variable was expressed as odds ratio (OR) and the associated 95% confidence interval (95% CI) based on the Wald test or Fisher’s exact test (as applicable). All statistical analyses were performed using R software (R Core Team, 2021) and JMP^® (Version 16.0. SAS Institute Inc., Cary, NC, 1989–2021). p-values were considered statistically significant at 0.05. Correction for multiple comparisons in the univariable analyses was not conducted due to the exploratory nature of the investigation. Because the analyses of associations were conducted at the ear level, the p-values for explanatory variables that vary at the cat level (i.e., season, sex, age, pregnancy, and weight) might be underestimated due to the unaccounted repeated observations of the same cat level observation twice (once for each ear). However, considering the smaller number of explanatory variables that vary at the cat versus ear level and that preliminary analysis indicated p-values for most cat-level variables were >0.05 in univariable analyses and there was only one instance of a cat-level variable (season) in final multivariable models, we considered this to be a reasonable approach that allowed streamlined statistical analyses and aided interpretation while having minimal effects on conclusions.

Ethical approval

The sampling was approved by the Veterinary Ethics Committee of the Faculty of Veterinary Medicine, University of Zagreb, Croatia (602-04/18-01/5, 251-61-41/258-19-4-S.V.).

Results

Free-roaming cat population

A total of 74 free-roaming cats were examined. All cats were of the domestic short-hair breed. Twenty-six (26/74; 35.1%) cats were male (intact (24/26; 92.3%) or neutered (2/26; 7.7%), and 48/74 (64.9%) were female (intact (43/48; 89.6%) or spayed (5/48; 10.4%). Of the 43 intact cats, 42 were checked for pregnancy, and six (6/42; 14%) were pregnant. Most cats were under 1 year of age [55/73; 75.3%; the age of one animal (1/74; 1.4%) was not determined]. Thirty-five animals (35/74; 47.3%) were sampled in spring, 17/74 (23.0%) in summer, and 22/74 (29.7%) in autumn.

Otitis externa

Of the 74 cats subjected to statistical analysis in this study, 30 were positive for at least one of the five otitis diagnoses: O. cynotis, Malassezia, cocci, rods, and/or neutrophils (Supplementary Table S1). This analysis revealed that cats commonly had multiple etiologies detected at the individual ear (Tables 1 and Supplementary Table S1) and cat (Tables 2 and Supplementary Table S2) levels. Co-diagnosis of O. cynotis and Malassezia yeasts was particularly common at both the ear (39.3%, 22/56 infected ears) and cat (40.0%, 12/30 infected cats) levels. Interestingly, three or more diagnoses co-occurred in 15% of infected ears (Table 1) and 27% of infected cats (Table 2). Importantly, no clinical signs of inflammation of the external ear canal were observed in 8.9% (5/56) of infected ears, all of which were infected with O. cynotis and Malassezia or Malassezia alone (Supplementary Table S1).

Supplementary Table S1. Diagnosis of otitis externa by ear for the five otitis externa diagnoses individually.

ear	cat.ID	Otodectes otitis	Malassezia. otitis	Cocci.otitis	Rods otitis	Neutrophils otitis
R	C3	Yes	Yes	No	No	No
L	C3	Yes	Yes	No	No	No
R	C6	Yes	Yes	No	No	Yes
L	C6	No	Yes	No	No	Yes
R	C7	Yes	Yes	No	No	No
L	C7	Yes	Yes	No	No	No
R	C14	No	Yes	No	No	No
L	C14	No	Yes	No	No	No
R	C17	No	No	No	No	No
L	C17	No	Yes	No	No	No
R	C19	No	Yes	No	No	No
L	C19	No	Yes	No	No	No
R	C27	No	Yes	No	No	No
L	C27	Yes	Yes	No	No	No
R	C41	No	Yes	No	No	No
L	C41	No	Yes	No	No	No
R	C44	No	Yes	Yes	Yes	Yes
L	C44	No	Yes	Yes	Yes	Yes
R	C45	Yes	No	No	No	No
L	C45	Yes	No	No	No	No
R	C46	Yes	Yes	No	Yes	No
L	C46	Yes	Yes	No	No	No
R	C47	Yes	Yes	No	No	No
L	C47	Yes	Yes	No	No	No
R	C48	No	Yes	No	No	No
L	C48	No	No	No	No	No
R	C49	Yes	Yes	No	No	No
L	C49	Yes	Yes	Yes	No	Yes
R	C51	No	Yes	No	No	No
L	C51	No	Yes	No	No	No
R	C52	No	Yes	No	No	No
L	C52	No	Yes	Yes	No	No
R	C53	Yes	Yes	Yes	No	No
L	C53	Yes	Yes	Yes	Yes	No
R	C54	Yes	Yes	No	No	No
L	C54	Yes	Yes	No	No	No
R	C55	No	Yes	No	No	No
L	C55	No	Yes	No	No	No
R	C57	Yes	Yes	No	No	No
L	C57	Yes	Yes	No	No	No
R	C58	Yes	Yes	No	No	No
L	C58	No	Yes	No	No	No
R	C59	No	Yes	No	No	No
L	C59	Yes	Yes	No	No	No
R	C61	No	No	No	No	No
L	C61	No	No	Yes	No	Yes
R	C62	Yes	Yes	Yes	No	Yes
L	C62	Yes	Yes	No	No	No
R	C64	No	No	No	No	No
L	C64	Yes	Yes	No	No	No
R	C68	No	Yes	No	No	No
L	C68	No	Yes	No	No	No
R	C71	No	Yes	No	No	No
L	C71	Yes	Yes	No	No	No
R	C72	Yes	No	No	No	No
L	C72	Yes	Yes	No	No	No
R	C74	Yes	Yes	No	No	No
L	C74	Yes	Yes	No	No	No
R	C75	Yes	Yes	No	No	No
L	C75	Yes	Yes	Yes	No	No

Table 1. Frequency of occurrence of individual and combinations of otitis externa diagnoses by ear among the 30 cats diagnosed with otitis externa.

Diagnosis by ear	Frequency (No.)	Percentage (%)
Otodectes cynotis	3	5
Malassezia spp.	19	32
Cocci	0	0
Otodectes cynotis + Malassezia spp.	22	37
Otodectes cynotis + cocci	0	0
Malassezia spp. + cocci	1	2
Malassezia spp. + neutrophils	1	2
Cocci + neutrophils	1	2
Otodectes cynotis + Malassezia spp. + cocci	2	3
Otodectes cynotis + Malassezia spp. + rods	1	2
Otodectes cynotis + Malassezia spp. + neutrophils	1	2
Otodectes cynotis + Malassezia spp. + cocci + rods	1	2
Otodectes cynotis + Malassezia spp. + cocci + neutrophils	2	3
Malassezia spp. + cocci + rods + neutrophils	2	3
None	4	7

Table 2. Frequency of occurrence of individual and combined otitis externa diagnoses among the 30 cats diagnosed with otitis externa.

Diagnosis by a cat (either ear of a cat positive)	Frequency (No.)	Percentage (%)
Otodectes cynotis	0	0
Malassezia spp.	9	30
Cocci	0	0
Otodectes cynotis + Malassezia spp.	12	40
Otodectes cynotis + cocci	0	0
Malassezia spp. + cocci	0	0
Malassezia spp. + neutrophils	0	0
Cocci + neutrophils	1	3
Otodectes cynotis + Malassezia spp. + cocci	1	3
Otodectes cynotis + Malassezia spp. + rods	1	3
Otodectes cynotis + Malassezia spp. + neutrophils	1	3
Otodectes cynotis + Malassezia spp. + cocci + rods	1	3
Otodectes cynotis + Malassezia spp. + cocci+ neutrophils	2	7
Malassezia spp. + cocci + rods	1	3
Malassezia spp. + cocci + rods + neutrophils	1	3

Supplementary Table S2. UNI versus BI positive diagnoses in 30 cats with diagnosed otitis externa.

cat.ID	Otodectes otitis	Malassezia otitis	Cocci otitis	Rods otitis	Neutrophils otitis
C3	BI	BI	NEG	NEG	NEG
C6	UNI	BI	NEG	NEG	BI
C7	BI	BI	NEG	NEG	NEG
C14	NEG	BI	NEG	NEG	NEG
C17	NEG	UNI	NEG	NEG	NEG
C19	NEG	BI	NEG	NEG	NEG
C27	UNI	BI	NEG	NEG	NEG
C41	NEG	BI	NEG	NEG	NEG
C44	NEG	BI	BI	BI	BI
C45	BI	NEG	NEG	NEG	NEG
C46	BI	BI	NEG	UNI	NEG
C47	BI	BI	NEG	NEG	NEG
C48	NEG	UNI	NEG	NEG	NEG
C49	BI	BI	UNI	NEG	UNI
C51	NEG	BI	NEG	NEG	NEG
C52	NEG	BI	UNI	UNI	NEG
C53	BI	BI	BI	UNI	NEG
C54	BI	BI	NEG	NEG	NEG
C55	NEG	BI	NEG	NEG	NEG
C57	BI	BI	NEG	NEG	NEG
C58	UNI	BI	NEG	NEG	NEG
C59	UNI	BI	NEG	NEG	NEG
C61	NEG	NEG	UNI	NEG	UNI
C62	BI	BI	UNI	NEG	UNI
C64	UNI	UNI	NEG	NEG	NEG
C68	NEG	BI	NEG	NEG	NEG
C71	UNI	BI	NEG	NEG	NEG
C72	BI	UNI	NEG	NEG	NEG
C74	BI	BI	NEG	NEG	NEG
C75	BI	BI	UNI	NEG	NEG

BI = bilateral, UNI = unilateral, NEG = negative.

For each diagnosis, we interpreted whether the infection in a cat was absent (none), UNI, or BI (Supplementary Table S2). When diagnosed with otitis externa caused by O. cynotis and Malassezia spp., the infection was BI 68% (13/19) and 86% (24/28) of the time, respectively (Table 3). These findings support the need for treatment of both ears for O. cynotis. For cocci, when present, the infection was more likely (71%) UNI, supporting the UNI application of antibiotic treatment; however, because of the small sample size, more research is needed to support the generalizability of this finding.

Table 3. Frequency (No.) of UNI versus BI otitis externa by diagnosis.

	Otodectes cynotis	Malassezia spp.	Cocci	Rods	Neutrophils
BI	13	24	2	1	2
UNI	6	4	5	3	3
None	11	2	21	24	24
Total positive	19	28	7	4	5

Table 4 presents the results of univariable analyses of associations between explanatory variables and individual diagnoses of otitis based on the presence of O. cynotis, Malassezia yeasts, or cocci, as well as general otitis externa (i.e., positive for any or more of the five diagnoses). Supplementary Table S3 shows the ORs and associated 95% CIs for the tested associations in these univariable analyses.

Table 4. Univariable analysis of the associations between explanatory variables and individual and combined diagnoses.

Variable	Level	Otodectes cynotis			Malassezia spp.			Cocci			General otitis externa
		Negative (No.)	Positive (No.)	p-value	Negative (No.)	Positive (No.)	p-value	Negative (No.)	Positive (No.)	p-value	Negative (No.)	Positive (No.)	p-value
Wound	No	114	25	NA	96	43	NA	132	7	NA	92	47	NA
	Yes	2	7	3E-04	0	9	5E-05	7	2	0.09	0	9	1E-04
Stenosis	No	110	32	NA	96	46	NA	135	7	NA	92	50	NA
	Yes	6	0	0.34	0	6	2E-03	4	2	0.04	0	6	2E-03
TM.invisible	No	104	18	NA	94	28	NA	118	4	NA	90	32	NA
	Yes	12	14	1E-05	2	24	3E-11	21	5	0.01	2	24	3E-10
Earwax. discharge	Light brown	85	8	NA	83	10	NA	93	0	NA	82	11	NA
	Dark brown	31	24	6E-07	13	42	6E-16	46	9	9E-05	10	45	2E-17
Earwax.discharge.amount	1	85	4	NA	82	7	NA	88	1	NA	81	8	NA
	2	17	9	2E-04	13	13	7E-06	25	1	0.40	11	15	6E-07
	3	14	19	5E-10	1	32	2E-21	26	7	4E-04	0	33	1E-22
Earwax.discharge.crumbly	No	103	7	NA	94	16	NA	108	2	NA	91	19	NA
	Yes	13	25	2E-14	2	36	2E-19	31	7	1E-03	1	37	1E-19
Rods.otitis	No	114	30	NA	96	48	NA	138	6	NA	NT	NT	NT
	Yes	2	2	0.20	0	4	0.01	1	3	6E-04	NT	NT	NT
Neutrophils.otitis	No	112	29	NA	95	46	NA	137	4	NA	NT	NT	NT
	Yes	4	3	0.17	1	6	0.01	2	5	5E-06	NT	NT	NT
Season	Spring	64	6	NA	57	13	NA	70	0	NA	57	13	NA
	Summer	24	10	0.01	17	17	9E-04	28	6	9E-04	15	19	1E-04
	Autumn	28	16	3E-04	22	22	4E-04	41	3	0.06	20	24	7E-05
Sex	F	66	20	NA	57	29	NA	82	4	NA	54	32	NA
	M	38	10	0.83	31	17	0.85	43	5	0.28	30	18	1.00
	SF	8	2	1.00	6	4	0.73	10	0	1.00	6	4	1.00
	NM	4	0	0.57	2	2	0.61	4	0	1.00	2	2	0.63
Discharge	No	109	31	NA	91	49	NA	131	9	NA	87	53	NA
	Yes	3	1	1.00	3	1	1.00	4	0	1.00	3	1	1.00
Stomatitis	No	72	24	NA	60	36	NA	90	6	NA	57	39	NA
	Yes	40	8	0.26	34	14	0.32	45	3	1.00	33	15	0.27
Pregnancy	No	55	17	NA	51	21	NA	71	1	NA	48	24	NA
	Yes	9	3	1.00	4	8	0.02	9	3	0.01	4	8	0.05
Age	G1	84	26	NA	71	39	NA	103	7	NA	67	43	NA
	G2	30	6	0.38	25	11	0.59	34	2	1.00	25	11	0.36
Otodectes otitis	No	NT	NT	NT	93	23	NA	112	4	NA	NT	NT	NT
	Yes	NT	NT	NT	3	29	2E-13	27	5	0.02	NT	NT	NT
Malassezia otitis	No	93	3	NA	NT	NT	NT	95	1	NA	NT	NT	NT
	Yes	23	29	2E-13	NT	NT	NT	44	8	1E-03	NT	NT	NT
Cocci otitis	No	112	27	NA	95	44	NA	NT	NT	NT	NT	NT	NT
	Yes	4	5	0.02	1	8	1E-03	NT	NT	NT	NT	NT	NT

p-values < 0.05 are shown in bold. Gray cells indicate that the analysis was not applicable and was therefore not conducted.

Supplementary Table S3. ORs and 95% confidence intervals for univariable associations of explanatory variables with individual and combined otitis externa diagnoses. TN indicate that the analysis was not applicable and thus not conducted.

		Otodectes			Malassezia			Cocci			General otitis externa
		OR			OR			OR			OR
Variable	Level comparison	Mean	Lower	Upper	Mean	Lower	Upper	Mean	Lower	Upper	Mean	Lower	Upper
Wound	Yes versus no	15.5	2.7	161.8	Inf	4.1	Inf	5.3	0.5	36.1	Inf	3.6	Inf
Stenosis	Yes versus no	0.0	0.0	3.1	Inf	2.3	Inf	9.3	0.7	79.5	Inf	2.1	Inf
TM.invisible	Yes versus no	6.7	2.7	16.9	39.1	8.8	361.7	6.9	1.4	37.7	32.9	7.5	300.2
Earwax.discharge	Dark versus light brown	8.2	3.3	20.2	26.8	10.9	66.2	Inf	3.7	Inf	33.5	13.2	85.1
Earwax.discharge.amount	2 versus 1	10.9	2.7	54.3	11.3	3.5	40.7	3.5	0.0	278.7	13.3	4.2	46.0
	3 versus 1	27.6	7.7	128.2	334.1	44.5	13,648.4	23.0	2.8	1,072.8	Inf	64.3	Inf
Earwax.discharge.crumbly	Yes versus no	28.3	10.2	78.3	100.6	22.4	947.2	11.9	2.1	123.5	169.1	25.5	6,912.5
Rods.otitis	Yes versus no	3.8	0.3	53.8	Inf	1.3	Inf	62.1	4.3	3,535.6	NT	NT	NT
Neutrophils.otitis	Yes versus no	2.9	0.4	18.0	12.2	1.4	574.0	75.3	9.3	1012.6	NT	NT	NT
Season	Summer versus Spring	4.4	1.5	13.6	4.4	1.8	10.8	Inf	2.7	Inf	5.6	2.2	13.7
	Autumn versus Spring	6.1	2.2	17.2	4.4	1.9	10.2	Inf	0.7	Inf	5.3	2.3	12.3
Sex	M versus F	0.9	0.3	2.2	1.1	0.5	2.4	2.4	0.5	12.6	1.0	0.5	2.2
	SF versus F	0.8	0.1	4.6	1.3	0.3	6.0	0.0	0.0	14.0	1.1	0.2	5.2
	NM versus F	0.0	0.0	5.4	1.9	0.1	28.2	0.0	0.0	41.6	1.7	0.1	24.2
Discharge	Yes versus no	1.2	0.0	15.2	0.6	0.0	8.0	0.0	0.0	24.9	0.5	0.0	7.0
Stomatitis	Yes versus no	0.6	0.2	1.5	0.7	0.3	1.4	1.0	0.2	4.9	0.7	0.3	1.4
Pregnancy	Yes versus no	1.1	0.2	5.0	4.8	1.1	24.0	22.0	1.6	1,247.2	3.9	0.9	19.7
Age	G2 versus G1	0.6	0.2	1.7	0.8	0.4	1.8	0.9	0.1	4.8	0.7	0.3	1.5
Otodectes otitis	Yes versus no	NT	NT	NT	37.8	10.4	210.5	5.1	1.0	27.6	NT	NT	NT
Malassezia otitis	Yes versus no	37.8	10.4	210.5	NT	NT	NT	17.0	2.2	771.1	NT	NT	NT
Cocci otitis	Yes versus no	5.1	1.0	27.6	17.0	2.2	771.1	NT	NT	NT	NT	NT	NT

The only statistically significantly associated cat-level explanatory variables of ear infection were pregnancy, season, and body weight. Specifically, ear infections were more likely in pregnant cats than in non-pregnant cats for Malassezia and cocci, but not for O. cynotis (Tables 4 and Supplementary Table S3). In addition, ear infection was more likely in summer and autumn compared with spring for all three individual diagnoses as well as the combined diagnosis, except that the association did not reach the statistical significance threshold for the autumn and cocci relationship (p-value = 0.06; Tables 4 and Supplementary Table S3). Cat weight was associated with Otodectes otitis diagnosis, with a significantly (p-value = 4.8E-03) lower weight in Otodectes-positive cats (mean = 2.94, median = 2.98, and range: 1.40–5.35 kg) compared with Otodectes-negative cats (mean = 3.40, median = 3.25, and range: 1.60–5.35 kg). The presence of the evaluated clinical predictors (wound, stenosis, tympanic membrane invisibility, dark brown color of discharge, and crumbly discharge appearance) statistically significantly increased the odds of the three individual diagnoses and the combined diagnosis, except for no detected association of stenosis with Otodectes and wound with cocci (Tables 4 and Supplementary Table S3). The presence of rods and neutrophils was not associated with the diagnosis of Otodectes, but rods and neutrophils increased the odds of diagnosis of Malassezia- and cocci-based otitis. Finally, Otodectes, Malassezia, and cocci were statistically significantly associated with each other, where otitis diagnosed based on the presence of one of the pathogens increased the odds of a positive diagnosis for the other two pathogens; the association between Otodectes and Malassezia was particularly strong (OR mean 37.8, 95%CI: 10.4–210.5, p-value 10^-13) (Tables 4 and Supplementary Table S3), which is unsurprising considering their common co-occurrence (Tables 1 and Supplementary Table S1).

No multivariable model was constructed for cocci. However, the final multivariable logistic regression models for O. cynotis, Malassezia yeasts, and general otitis externa differed (Tables 5 and Supplementary Table S3), which is unsurprising considering the differences in the patterns of diagnoses across cats and individual cats’ ears (Supplementary Tables S1 and S2). After controlling for other factors, only a few clinical explanatory variables were associated with the two causative agents, albite, with very wide 95% CIs. Specifically, the odds of otitis externa caused by the primary factor O. cynotis were higher in the presence of crumbly earwax discharge (OR: 5.9, 95% CI: 1.7–20.7, and p-value: 0.005) and co-occurrence with the secondary (opportunistic) factor, Malassezia yeast (OR: 12.4, 95% CI: 2.7–56.4, and p-value: 0.001) (Tables 5 and Supplementary Table S3). The odds of Malassezia yeast diagnosis increased in the presence of an invisible tympanic membrane (OR: 9.7, 95% CI: 1.1–86.8, and p-value: 0.04), large (but not medium) amount of earwax discharge (OR: 58.5, 95% CI: 5.4–630.3, and p-value: 0.001), crumbly discharge (OR: 14.5, 95% CI: 1.9–112.9, and p-value: 0.01), and co-occurrence with the primary agent, O. cynotis (OR: 8.8, 95% CI: 1.4–57.4, and p-value: 0.01) (Tables 5 and Supplementary Table S3). Finally, the odds of general otitis externa were higher in the presence of dark brown discharge (OR: 30.2, 95% CI: 7.2–126.1, and p-value: 3*10^-6) and crumbly discharge (OR: 65.9, 95% CI: 7.8–559.2, and p-value: 10^–4). The odds of general otitis were also higher in summer and autumn than in spring (OR: 12.8, 95% CI: 2.2–74.6, and p-value: 0.005; 11.7, 95% CI: 2.2–62.8, and p-value: 0.004, respectively) (Tables 5 and Supplementary Table S3).

Table 5. The final multivariable logistic regression models for the primary agent of otitis (O. cynotis), the secondary agent (Malassezia yeast), and general otitis externa (defined as an ear positive for O. cynotis, Malassezia yeasts, cocci, rods, and/or neutrophils). A multivariable model could not be built for the cocci-based diagnosis.

Outcome	Explanatory variable	OR
		Mean	Lower	Upper	p-value
Otodectes otitis	(Intercept)	0.0	0.0	0.1	3E-09
	Earwax.discharge.crumbly (yes versus no)	5.9	1.7	20.7	0.005
	Malassezia (yes versus no)	12.4	2.7	56.4	0.001
Malassezia otitis	(Intercept)	0.1	0.0	0.1	7E-10
	TM.invisible (yes versus no)	9.7	1.1	86.8	0.04
	Earwax.discharge.amount (2 versus 1)	1.8	0.3	10.1	0.50
	Earwax.discharge.amount (3 versus 1)	58.5	5.4	630.3	0.001
	Earwax.discharge.crumbly (yes versus no)	14.5	1.9	112.9	0.01
	Otodectes (yes versus no)	8.8	1.4	57.4	0.02
General otitis externa	(Intercept)	0.0	0.0	0.1	1E-07
	Earwax.discharge (dark versus light brown)	30.2	7.2	126.1	3E-06
	Earwax.discharge.crumbly (yes versus no)	65.9	7.8	559.2	1E-04
	Season (summer versus spring)	12.8	2.2	74.6	0.005
	Season (autumn versus spring)	11.7	2.2	62.8	0.004

Culture of Malassezia yeasts

During the examination of cats, Malassezia yeasts were cultured from the external ear canal of 31 of 74 (41.9%) cats, of which 74.2% (23/31) bilaterally and 25.8% (8/31) unilaterally. Interestingly, 22.6% (7/31) culture-positive cats were negative on cytological investigation for Mallasezia yeast, and 14.3% (4/28) were positive on cytology but had negative culture results. Malassezia yeasts were isolated from 83.3% (25/30; Table 6) of cats diagnosed with otitis based on the cytological investigation (i.e., presence of Otodectes, Malassezia, cocci, rods, and/or neutrophils). Out of the 30 cats with otitis, 4 (13.3%) exhibited moderate Malassezia growth, while 21 (70%) showed heavy Malassezia growth in at least one ear (Table 6). Five (16.7%) of the 30 cats diagnosed with otitis based on any etiology showed no Malassezia growth. Malassezia pachydermatis was the predominant species isolated from all positive cats.

Table 6. Semiquantitative evaluation of the growth of Malassezia yeasts cultured in 30 cats with otitis externa at the ear and cat levels. The cat level category was assigned based on the ear with the highest observed Malassezia growth.

Malassezia yeasts	Ears of cats with otitis externa (%)	Cats with otitis externa (%)
No growth	16/60 (26.7)	5/30 (16.7)
Scarce growth (<10 CFU)	1/60 (1.7)	0/30 (0.0)
Moderate growth (10–50 CFU)	4/60 (6.7)	4/30 (13.3)
Heavy growth (>50 CFU)	39/60 (65.0)	21/30 (70.0)

Of the 30 cats diagnosed with otitis, 28 (93.3%) had these yeasts present based on the cytological evaluation. Supplementary Table S4 shows the strong univariable associations between the result of Malassezia yeast culture and Malassezia otitis diagnoses at the ear level: OR: 23.8, 95%CI: 9.8–57.7, and p-value: 3*10^–15. Similarly, analysis at the cat level indicated a strong association between culture and the diagnosis of Malassezia otitis: OR: 33.4, 95%CI: 8.8–126.6, and p-value: 2*10^–9 (not shown here).

Supplementary Table S4. Association between Malassezia yeasts culture and evaluated individual and combined otitis diagnoses outcomes.

		Negative	Positive	OR
Outcome	Level	No.	%	No.	%	Mean	Lower	Upper	p-value
Otodectes otitis	No	87	75%	7	22%	1.0	NA	NA	NA
	Yes	29	25%	25	78%	10.7	4.2	27.4	3E-08
Malassezia otitis	No	83	86%	11	21%	1.0	NA	NA	NA
	Yes	13	14%	41	79%	23.8	9.8	57.7	3E-15
Cocci otitis	No	93	67%	1	11%	1.0	NA	NA	NA
	Yes	46	33%	8	89%	15.9	2.0	722.4	1E-03
General otitis externa	No	81	88%	13	23%	1.0	NA	NA	NA
	Yes	11	12%	43	77%	24.4	10.1	59.0	2E-15

Discussion

Free-roaming cat populations are a potential source of several infectious and parasitic diseases that can be easily transmitted to outdoor pet cats (Gerhold and Jessup, 2013). In this study, general otitis externa (diagnosed based on the presence of O. cynotis, Malassezia yeasts, cocci, rods, and/or neutrophils) was found in 40.5% (30/74) of the cats. This prevalence is lower than the previously reported rate of 55.1% in stray cats in Italy (Perego et al., 2014) but significantly higher than the 2% reported in Belgium (Bollez et al., 2018). Cross-tabulation of the five individual diagnoses across cats and ears of individual cats revealed that the five evaluated diagnoses often but not always co-occur in an individual ear (Tables 1 and Supplementary Table S1) or cat (Tables 2 and Supplementary Table S2). This suggests the multicausal etiology of otitis externa in cats and underscores the importance of comprehensive cytological evaluations, such as those performed here, to reduce the risk of false-negative diagnoses. The higher prevalence of general otitis externa at the ear level was observed in summer and autumn compared with spring, thereby contributing to the fact that otitis is more common in warmer seasons (Topală et al., 2007). In this study, neither sex nor age was a significant predisposing factor for the development of otitis, consistent with previous reports in stray cats (Perego et al., 2014). Although not statistically significant, it has been observed that male cats and cats younger than 1 year are more likely to develop otitis. A previous study noted a higher incidence of otitis in 35.3% of cats up to 1 year old (Topală et al., 2007). One of the reasons is that younger cats are more susceptible to developing otitis from ear mites due to maternal transmission and social interactions (Fanelli et al., 2020). Univariable statistical analyses indicated that pregnancy increased the likelihood of developing Malassezia and cocci otitis; however, after controlling for other factors in multivariable analysis, pregnancy was no longer a significant predictor (possibly due to the small sample size of the subset of intact cats checked for pregnancy to which this analysis was applicable). Nevertheless, hormonal fluctuations can disrupt the balance of the skin microbiome during pregnancy, creating favorable conditions for the proliferation of Malassezia yeasts and dysbiosis.

In this study, ear mites were found in 25.7% (19/74) of the examined cats. This is similar to previous findings from studies on stray and feral cats (Perego et al., 2014; Fanelli et al., 2020). They were the primary cause of otitis in 63.3% (19/30) of the cats in our study, which is slightly higher than the 53.3% prevalence reported in Italy (Perego et al., 2014) and considerably higher than the 2% reported in Belgium (Bollez et al., 2018). Studies have found that the prevalence of O. cynotis in cats can vary greatly, with reported rates ranging from 0.9% to 47.1% (Sotiraki et al., 2001; Akucewich et al., 2002; Tyler et al., 2019; Fanelli et al., 2020; Hiblu et al., 2021). All observed variations may be attributed to differences in environmental conditions, geographic locations, and acaricide treatments (Perego et al., 2014; Bollez et al., 2018; Fanelli et al., 2020). The primary factor of otitis in cats is O. cynotis, which often leads to an imbalanced microenvironment in the external ear canal that can promote excessive growth of resident bacteria or yeasts in the ear and otitis manifestation (Tielemans et al., 2021). The majority of otitis externa in this study was caused by O. cynotis together with Malassezia spp. (40%, 12/30), followed by Malassezia spp. alone (30%, 9/30) at the cat level, and the findings were similar at the ear level [37% (22/60) and 32% (19/60), respectively] of cats with otitis (Tables 1 and 2). These prevalences are significantly higher than the observed prevalences of 8.7% and 5.8% in Italian stray cats (Perego et al., 2014). In another study, 70.1% of pet cats had otitis caused by O. cynotis and M. pachydermatis, which is a significantly higher percentage. However, our results are consistent with the reported 28.3% prevalence of Malassezia infections (Nardoni, 2014). The proliferation of coccoid bacteria was detected in 23.3% (7/30) and rod-shaped bacteria in 13.3% (4/30) of cats with otitis. These findings are consistent with previously reported rates of 21.4% and 16%, respectively (Perego et al., 2014). Although bacterial cultures were not performed, these findings suggest that coccoid bacteria are more prevalent in causing otitis in cats, as previously reported (Roy et al., 2011; Nardoni, 2014). Finally, neutrophils were detected in 16.7% (5/30) of the cats with otitis, but never alone (Table 2).

Using the cultural method, Malassezia yeasts were isolated in 83.3% (25/30) of otitis cases. This is a lower percentage compared to 95% positive Persian and domestic short-haired cats (Shokri et al., 2010) but higher than that previously reported for pet cats (Crespo et al., 2002; Cafarchia et al., 2005). The heavy growth of Malassezia yeasts was found in 70% of cats with otitis, indicating that they play a significant role as secondary infectious agents. Among the positive cats, M. pachydermatis was the most frequently detected species, which is consistent with the findings of previous culture-based studies (Cafarchia et al., 2005; Shokri et al., 2010; Nardoni, 2014). This result is in contrast with the lower recovery rates observed in other studies (Crespo et al., 2002; Dizotti and Coutinho, 2007).

In clinical practice, the presence of a small amount of cerumen in the external ear canal of healthy cats is considered normal (Brame and Cain, 2021). However, increased cerumen content could indicate otitis externa, as confirmed by the results of this study. The presence of moderate or large amounts of cerumen increased the odds of otitis diagnosis caused by the primary agent O. cynotis and the secondary agent Malassezia yeasts, as well as general otitis, by at least 10-fold (ORs > 10), while a similar impact on cocci was seen only in the presence of a large amount of cerumen.

Diagnosing otitis externa requires a combination of patient history, clinical signs, and cytology. This study demonstrated associations between individual clinical signs and cytology (Table 4 and Supplementary Table S3). However, after controlling for the effect of other factors in multivariable analysis, the crumbly appearance of discharge was the only clinical sign that remained statistically significantly associated with otitis (individual and combined diagnoses; ORs > 10). The discharge amount was predictive of Malassezia otitis (OR > 1.8), and the dark color of the discharge was predictive of general otitis externa (OR = 30.2). Different clinical signs associated with individual diagnoses are expected, considering different patterns of diagnosis co-occurrence (Tables 1, 2, and Supplementary Tables S1 and S2) and different interactions of pathogens with the ear tissues. Importantly, no clinical signs of inflammation of the external ear canal were observed in 8.9% (5/56) of infected ears detected in this study, and in all instances, the ears were infected with O. cynotis accompanied by the secondary agent, Malassezia yeasts, or solely with Malassezia yeasts alone. In a previous study, O. cynotis was found in 10.8% of cats without concurrent clinical signs of otitis externa (Akucewich et al., 2002). As expected, a strong but not perfect association between cytology evaluation and mycology culture (Supplementary Table S4), indicating that culturing is not particularly suitable for screening otitis externa. However, a culture-based approach is recommended because of increasing reports of treatment failures in Malassezia infections and their resistance to antifungal agents (Angileri et al., 2019; Bond et al., 2020). The authors acknowledge the lack of unique defined thresholds for the diagnosis of otitis when semiquantitative or quantitative cytology is used. In addition, the evaluation of cell counts under HPF or OIF could also lead to differences that could affect the comparison of the results discussed and lead to biased critical opinions among the scientific reader. These facts present known challenges in interpreting the final results. While our research is the first of its kind to investigate the prevalence, risk factors, and pattern of occurrence of otitis externa in free-roaming cats in Croatia, its design has several limitations. We enrolled a sample of free-roaming cats captured as part of the TNP, which may have introduced selection bias, as captured cats may be different from those not captured. Although the confounding bias was controlled through multivariable analyses, residual confounders for which data were not collected may have been present. The relatively small sample size may have resulted in insufficient statistical power to detect additional risk factors for otitis externa.

Conclusion

A comprehensive understanding of the etiology of feline otitis and its risk factors, along with further data on this condition, is crucial for advancing our knowledge, particularly concerning free-roaming cat populations. Although the study design has certain limitations and the diagnostic criteria present known weaknesses, the findings offer valuable insights into the prevalence of otitis externa among free-roaming cats in a specific geographic region.

Conflict of interest

The authors declare no conflict of interest.

Funding

This research received no specific grant.

Authors contribution

SH designed the study and wrote the first draft of the manuscript and edited and revised the manuscript. MK performed sampling, laboratory analyses, and culturing, and assisted in writing the first draft. ZŠ was involved in designing the study, and performed double-blinded cell counting, and sample preservation. RI and MF performed statistical analysis, interpreted results, and wrote data analysis. ZŠ, JH, MP, RI, and VS helped in writing the manuscript, critically revised it, and edited it. All authors have read and approved the final manuscript.

Data availability

The data supporting the findings of this study are presented in the manuscript and supplementary tables. Further details can be obtained from the corresponding author upon reasonable request.