Non-neoplastic lesions of equine skin in the central United States and Canada: A retrospective study

Abstract

A retrospective study evaluated the clinical data and histologic features of non-neoplastic dermatologic lesions in skin biopsies from horses, donkeys, and mules submitted over a 10-year period to the Colorado State University Diagnostic Medicine Center and to the University of Saskatchewan Western College of Veterinary Medicine and Prairie Diagnostic Services. A total of 1793 non-neoplastic biopsies were available for analysis. Based on the histologic description and morphologic diagnosis, each submission was classified as inflammatory, or non-neoplastic/non-inflammatory. Results of this study demonstrate that non-neoplastic equine cutaneous biopsies are predominantly inflammatory, but histopathology alone may not be sufficient to definitively identify the underlying etiologies driving inflammatory lesions. Clinical correlation, including breed, age, and distribution of the cutaneous lesion, as well as ancillary tests such as bacterial or fungal culture are required to complete the clinical picture and arrive at the appropriate clinical diagnosis.

Introduction

The epidemiology of non-neoplastic skin disease can be affected by geography, climate, season, abundance of and life cycles of pathogens and vectors, as well as host-factors including genetics, age, and the presence or absence of underlying conditions including immune suppression, immune tolerance, and endocrinopathies. In veterinary dermatology, well-known geographic trends include the distribution of blastomycosis in the Ohio River valley or Cryptococcus gattii in the pacific northwest (1). Seasonal trends include those of many ectoparasites, such as the increased incidence of flea-allergy dermatitis in companion animals during the temperate seasons and of dermal acariasis in herd animal species during winter months, when animals are likely to be in close contact with each other (2,3). Seasonal variation in the severity of allergic conditions in atopic dogs has been related to the local abundance of the inciting allergen in the environment (4). Climate, including temperature and moisture patterns, are also associated with variations in certain skin conditions (5).

Relative to small animal dermatology, the factors influencing equine dermatologic diseases are less well-characterized. Some studies have identified trends over different geographic locations, across different seasons, and potential breed predispositions. For example, analysis of equine skin biopsies submitted to the Oregon State University Veterinary Teaching Hospital and Diagnostic Laboratory identified fungal granulomas as being more common in spring, and eosinophilic granulomas as being more common in the fall and winter (6).

While some non-neoplastic dermatologic conditions have been reviewed on the east and northwestern coasts (6–8), no data are available regarding the types and distributions of non-neoplastic skin lesions in horses of the central United States or central Canada. This geographic region represents a unique geographic environment distinct from both the northwest and eastern areas, thus significant variations in the distribution and incidence of equine dermatologic conditions may be expected. The objective of this study was to review a decade of skin biopsies from non-neoplastic equine skin conditions submitted to the Colorado State University Diagnostic Medicine Center (CSU), Fort Collins, Colorado, USA, and to the Prairie Diagnostic Services (PDS), Inc., University of Saskatchewan, Saskatoon, Saskatchewan, Canada, to identify trends for horses in the central North American environment, and to compare trends to those identified in other geographical locations.

Materials and methods

The computerized records of the CSU Diagnostic Medicine Center and PDS were queried for all equine skin biopsies submitted between July 1, 2000 and July 1, 2010 using the following “wild card” search terms: skin, dermal, cutaneous, subcutaneous, pyoderma, folliculitis, adenitis, epidermitis, epidermatitis, collagen, necrobiosis, and proud flesh. Biopsies that included neoplastic diseases were evaluated separately (9), except where they co-existed with a significant inflammatory or non-inflammatory/non-neoplastic (NINN) condition. For the remaining biopsies, the age, gender, and breed of each horse, if known, were entered into the database. Data regarding the date of histologic diagnosis and geographical location (town, state or province, and zip or postal code) were also collected. When the exact address for the horse was not available, the geographic location of the submitting veterinary clinic was used as a surrogate. Lesions were classified as “tack associated” if the clinical history made specific mention of tack, e.g., “alopecia in the girth area,” or “nodules under the saddle.” Date of biopsy submission was recorded. Seasons were broadly classified as spring (March–May), summer (June–August), fall (September–November), and winter (December–February).

The morphologic diagnosis and histologic description were used to characterize inflammatory lesions based upon predominant type of inflammation (suppurative/neutrophilic, lymphoplasmacytic, eosinophilic, pyogranulomatous, mixed, and granulomatous), and distribution of inflammation (superficial, perivascular, deep, diffuse, nodular). Report comments were evaluated for pathologists’ interpretative comments and suggested etiologies. For all cases, additional features entered in the database as dichotomous variables (present or not present in the histologic description) included: erosion/ulceration, solar elastosis, hyperkeratosis, collagen changes (degeneration, dysplasia), and the presence of plant material, mineral, multinucleated giant cells, acantholytic cells, and pigmentary incontinence. The presence of organisms was noted and further subdivided into bacteria, fungi, or parasites. Descriptive and comparative statistical analysis was performed using commercially available software (©IBM SPSS 2010, Somers, New York, USA). The frequency of lesions by season, cell type, histologic pattern, breed and laboratory were compared using chi-squared tests and odds ratios with 95% confidence intervals.

Results

General

A total of 5141 equine skin biopsies were submitted to CSU and PDS between July 1, 2000 and July 1, 2010. Samples from Colorado (n = 602), Alberta (n = 309), Saskatchewan (n = 237), Wyoming (n = 60), Montana (n = 57), South Dakota (n = 52), New Mexico (n = 44), Texas (n = 35), Manitoba (n = 28), Nebraska (n = 26), Minnesota (n = 14), and other mid-western states (including Kansas, Oklahoma, North Dakota, Iowa, Missouri, and Wisconsin) comprised over 75% of total submissions during the time period evaluated. Geographic location was not available for 28 horses.

There were 1512 inflammatory and 278 NINN equine skin biopsy submissions, comprising 29.4% and 5.4% of total submissions, respectively. The number of inflammatory and NINN submissions to CSU ranged from 40 to 144 cases per year, with an average of 114 submissions per year, while submissions to PDS ranged from 23 to 75 cases per year, with an average of 64 per year. Overall, inflammatory and NINN biopsy submissions were most frequent in summer (29.6%) followed by spring (26.3%), fall (24.0%), and winter (19.7%), a difference that was statistically significant (P < 0.0001).

Biopsies were collected from 870 geldings, 711 mares, 80 stallions, and 132 horses for which the gender was unspecified. Age of horses ranged from < 1 y to 36 years of age, with a median age of 9 y and mode of 8 y. The most common breeds included Quarter horses (45% of biopsies with known breed), Arabians (9.8%), Paints (9.4%), Warmbloods (9.4%), Thoroughbreds (7%), and Appaloosas (4.1%). The age for 238 horses was unspecified and breed was not available for 165 horses.

A total of 1512 inflammatory biopsies were submitted to CSU and PDS over the 10-year period, with an average of 153 submissions per year. Inflammatory lesions were most commonly biopsied from the trunk (20.8%). Seventeen percent of biopsies represented multifocal lesions, while 13.7% represented a generalized skin condition. Sixteen percent of biopsies were from the limb and 14.7% from the head.

Cellular infiltrate in the majority of inflammatory submissions was eosinophilic (40.4% of all inflammatory biopsies) followed by lymphoplasmacytic (20.4%), suppurative (18.1%), mixed cell type (14.3%), granulomatous (4.0%), and pyogranulomatous (1.8%), a frequency difference that was statistically significant (P < 0.0001). The predominant inflammatory cell type did not differ by breed (P = 0.93). Although eosinophilic inflammation was more often diagnosed in the summertime, there was no statistically significant difference in the type of inflammation by season (P = 0.052).

Predominant cell types were correlated with specific histologic patterns. In particular, suppurative inflammation was 6.0 (95% CI: 4.6 to 7.7) times more likely to have a superficial distribution pattern than inflammatory lesions with other cell types. Biopsies that were predominantly lymphoplasmacytic in nature were 51.1 (95% CI: 36.3 to 71.9) times more likely to have a superficial or perivascular distribution. Biopsies with eosinophilic and pyogranulomatous inflammation were 9.2 (95% CI: 7.0 to 12.1) times more likely to be distributed in nodular patterns. No specific pattern was detected for mixed inflammation.

Two hundred and seventy-two (17.9%) cases with inflammatory diagnoses had histologically visible bacteria, fungi, and/or parasites. Organisms were 5.0 (95% CI: 3.8 to 6.5) times more likely to be seen in lesions with suppurative or pyogranulomatous inflammation relative to all other categories. Bacteria or fungi were identified histologically in 27% of all suppurative biopsies and 65.7% of all pyogranulomatous biopsies. Of the suppurative biopsies with intralesional organisms, the majority (70%) contained bacteria, and the remainder fungal elements (26.2%) or both bacteria and fungi (3.8%). Of pyogranulomatous samples with intralesional organisms, the majority contained fungal organisms (76%) and the remainder bacterial organisms. Dermatophytes were 2.1 (95% CI: 2.0 to 8.5) times more common in PDS cases compared with CSU cases. Parasites were rarely seen within histologic specimens, and were only identified in 8 biopsies. Six of these biopsies were cases of external parasitism (1 contained Damalina sp. lice, 1 Chorioptes sp. mites, 1 a partial section of a tick, and 3 an unidentified ectoparasite). The other 2 contained the profiles of nematode larvae consistent with Habronema sp. All types of organisms were slightly less likely to be identified in inflammatory biopsies submitted in winter than those submitted in spring, summer, or fall; however, this difference in frequency was not statistically significant (OR 1.3, 95% CI: 0.95 to 1.90).

Plant material was identified in < 2% of inflammatory biopsies (26 biopsies) and tended to elicit a mixed inflammatory response. Plant material was 5.4 (95% CI: 2.7 to 10.9) times more likely to be seen within mixed or pyogranulomatous inflammation relative to all other inflammatory types. Only 2.5% of inflammatory biopsies (38 biopsies) were reported to be associated with tack including the girth, saddle, or bridle. Tack associated lesions were 2.6 (95% CI: 1.4 to 4.9) times more likely to be associated with eosinophilic inflammation relative to all other categories combined. Tack associated lesions were more commonly diagnosed in the summer, but this was not statistically significant.

There were 278 non-neoplastic, non-inflammatory biopsies (NINN). Eighty-seven of these cases were diagnosed with fibrosis, acanthosis, hyperkeratosis, epithelial hyperplasia, sebaceous hyperplasia, or some combination of these features. Fifty-two cases were diagnosed as granulation tissue with minimal or no significant inflammation (2.9% of total submissions). Forty-four biopsies (15.8% of NINN samples, 2.5% of total biopsies submitted) were cystic lesions (33 cases of epidermal inclusion cysts, 10 follicular cysts, and 1 apocrine cyst). Only 4 NINN biopsies (1.4%) featured erosion or ulceration, and these comprised 1 skin tag, 1 submission with severe solar elastosis, 1 submission with osseous metaplasia, and 1 epidermal cyst.

Discussion

Our study had several drawbacks inherent to retrospective studies and reviews of biopsy report records. Although numerous pathologists were involved in biopsy review and biopsies were pooled from only 2 diagnostic labs, high agreement has been shown in previous reviews of histopathology (10). In addition, it is important to consider that variances in the diagnosis of non-neoplastic and neoplastic skin conditions across the different geographic locales may reflect variances in economics, the ability of practitioners to diagnose causes in the field, and the association of veterinarians with a particular veterinary school or diagnostic laboratory.

Non-neoplastic lesions comprised nearly 35% of total cutaneous biopsies submitted to CSU and PDS in the 10-year period evaluated. In contrast, non-neoplastic cutaneous lesions comprised 88% of total submissions received by Cornell University (7). While this may suggest that non-neoplastic lesions are more common in the Northeast than in central North America, it may also be due to higher numbers of equine practitioners selectively submitting biopsies in that region. Non-neoplastic nodular and proliferative lesions comprised 8.7% of biopsy submissions to Oregon State University (6); however, it is difficult to compare our data to this data set for the Pacific northwest since that study specifically evaluated nodular and proliferative non-neoplastic lesions, and thus is not inclusive of all inflammatory/non-inflammatory conditions.

Overall, skin biopsy submissions were more common in spring, summer, and fall than in winter. All types of inflammatory lesions were more common in spring, summer, and fall relative to winter, but eosinophilic biopsies appeared to have the most striking seasonal trend. Several factors may influence seasonal variability including cyclical presence of biting insects such as mosquitos and Cullicoides, and obscuration of lesions by thicker winter hair coats and winter blankets. In addition, riding activity may be lessened in the winter and therefore lesions may be observed less frequently.

Inflammatory lesions were much more common than NINN lesions in this database. The predominant cell types present generally corresponded with distribution. For example, in biopsies with predominantly suppurative inflammation, the distribution of the inflammation was superficial. This suggests a response to acute injury such as erosion, ulceration, or necrosis of the skin surface, or a reaction to superficial bacteria. Eosinophilic inflammation and pyogranulomatous inflammation both tended to be nodular. This pattern fits with an inflammatory response centered on damaged tissue or localized pathogens. Lymphoplasmacytic inflammation was more likely to be perivascular, most compatible with a hypersensitivity response.

Eosinophilic inflammation was the dominant cell type. This is consistent with previous descriptions of the frequency of eosinophil dominated inflammation (7). Interestingly, the incidence of eosinophilic granulomas in our population (12.1%) was roughly 3 times higher than those reported by studies from Oregon State University (3.5%) and Cornell University (3.9%) (6,7). The reason for this is unknown, but this difference is consistent with the fact that equine dermatologic conditions can vary significantly over large geographical areas. Since eosinophilic inflammation is associated with allergic or hypersensitivity reactions, it is interesting to hypothesize that the horses in our data set may live in an environment with different or increased exposure to insects, external parasites, and botanical antigens than horses in other locations.

Cutaneous fungal infections are reported to be rare in horses, comprising < 0.5% of skin biopsy submissions (7). In a retrospective study of non-neoplastic nodular equine skin biopsies received by Oregon State University, fungal granulomas accounted for roughly 10% of all non-neoplastic lesions (8). In the current study, fungal elements were identified in 5.9% of total submissions and were evenly divided between cases of follicular dermatophytosis and fungal granulomas. Thus the incidence of fungal disease in our study population appears to be less than reported in the Pacific Northwest. This may be related to the more temperate and rainy environment of the Pacific Northwest, which may better promote the development of fungal infections relative to the drier environment of the central US and Canada (6). Contrary to the fungal granulomas evaluated at OSU, samples submitted to CSU were mainly characterized by pyogranulomatous instead of lymphoplasmacytic inflammation. Seasonal occurrence of fungal granulomas was consistent with that seen in Oregon horses with cases occurring year-round but more commonly biopsied in spring, summer, and fall; peak months were May and August.

Habronemiasis reportedly accounts for 8.6% and 1.0% of non-neoplastic skin lesions in horses of the Pacific Northwest and Northeast, respectively (6,7). Intralesional nematodes and nematode larvae were rarely identified in our database (< 1% of all lesions). Despite this, habronemiasis was included as a differential diagnosis for nearly all cases of nodular eosinophilic dermatitis in our database. Lesions clinically consistent with habronemiasis may not always contain histologic evidence of nematodes; in 1 case series, only 44% of biopsies clinically consistent with Habronema based on history, location, and the presence of sulfur granules had histologically evident nematodes (11). While biopsy site selection, depth of biopsy, and number of sections evaluated may partly explain the low incidence of Habronema identified in our study population, it is also possible that habronemiasis can be readily diagnosed in the field, or that horses with suspicious skin lesions may be treated empirically with anthelmintics, and thus lesions consistent with habronemiasis may resolve and not require biopsy. It is also possible that some lesions of habronemiasis were biopsied late in the course of disease, when the inciting organisms may have been degraded but the inflammatory reaction persisted. Though the Habronema spp. are distributed worldwide, the prevalence of infection has not been specifically reported across North America and may be rare or present at considerably lower levels in the central areas of the US and Canada, compared with the Pacific Northwest (12–16). Consistent with prior reports (7,11), cases of habronemiasis in the present study occurred in summer and fall, and were characterized by nodular aggregates of eosinophilic inflammation.

Roughly half of all submissions had no distinct cause identified in the morphologic diagnosis of the diagnostic report (Table 1). This finding indicates that biopsy alone cannot be relied upon to determine a definitive diagnosis of non-neoplastic dermatologic conditions. It is possible that the biopsied tissue represented a lesion in which the primary cause had resolved, such as in cases where an infectious agent or foreign body has been eliminated over time by the inflammatory response. It is also possible that some biopsy sections did not include representative sections of the etiology, for instance in cases with a small foreign body, rare larval nematodes, embedded fungal hyphae, or dermatophytes. Thus the number of sections examined and the biopsy site selection may have interfered with the ability to definitively diagnose the dermatologic condition. In many cases, the histologic inflammatory lesion is non-specific or non-diagnostic in the absence of additional information. Thus interpretation of the predominant cell type and histologic distribution lend important clues as to the etiology, especially when evaluated in light of clinical information.

Table 1.

Etiologies and syndromes of inflammatory and non-neoplastic, non-inflammatory equine skin biopsies

Etiology	CSU	PDS	Total/%
Unknown	563	347	910/50.8%
Eosinophilic syndromes	359	158	517/28.8%
Eosinophilic	209	91	300/16.7%
Eosinophilic granuloma	150	67	217/12.1%
Bacterial	49	12	61/3.4%
Benign cyst	51	8	59/3.3%
Autoimmune suspect	22	34	56/3.1%
Fungal (non-dermatophyte)	34	10	44/2.5%
Granulation tissue	21	21	42/2.3%
Dermatophytosis	11	25	36/2.0%
Suspect hypersensitivity	22	5	27/1.5%
Foreign material	4	6	10/0.6%
Parasite	5	3	8/< 0.5%
Solar elastosis	5	0	5/< 0.5%
Photosensitivity suspect	0	4	4/< 0.5%
MEED suspect	0	2	2/< 0.5%
Calcinosis circumscripta	0	2	2/< 0.5%
Amyloid	0	1	1/< 0.5%

CSU — Colorado State University, PDS — Prairie Diagnostic Service, MEED — Multisystemic eosinophilic epitheliotropic disease.

Diagnoses that fell into the category of undetermined etiologies included lymphoplasmacytic dermatitis, pyogranulomatous or granulomatous dermatitis, hyperplastic and hyperkeratotic dermatitis, chronic suppurative dermatitis, ulcerative dermatitis, non-nodular eosinophilic dermatitis, and dermal fibrosis. The predominant cell type present can provide clues about the particular etiology. Lymphoplasmacytic inflammation typically was found in either superficial or perivascular patterns, consistent with chronic allergic dermatitis. Nodular eosinophilic inflammation may be consistent with eosinophilic granuloma, while when identified in a perivascular pattern fits with hypersensitivity reaction. Pyogranulomatous inflammation tended to occur in a nodular pattern, which is consistent with a dermal pathogen or foreign body to which these cells may react. Suppurative inflammation was predominantly present in a superficial pattern, suggesting a reaction to ulceration or superficial pathogens such as bacteria. Though the number of submissions was small, tack-associated lesions tended to be associated with eosinophilic inflammation, suggesting that hypersensitivity to tack antigens may play a role in the development of these lesions. These lesions were more common in summer, which may be potentially related to increased use of horses and possibly the introduction of new tack materials during the summer. Location of lesions on the body can also provide the practitioner with clues as to etiology. Distributional differences for non-neoplastic dermatoses have been previously described for horses in the Northeast (7).

In the present study, < 4% of the biopsies were associated with concurrent bacterial or fungal culture. Culture of skin lesions is an extremely important tool when used in conjunction with histopathology. Alone, histopathology may recognize the presence of a bacterial or fungal organism, but may be unable to determine whether the suspicious organism is truly pathogenic, an opportunistic pathogen taking advantage of the primary skin disease, or simply a contaminant. Without the histologic picture to help identify associated tissue injury and inflammation, culture alone may not be enough to determine whether an isolated organism is the true cause of the lesion in question. Similarly, without a coincident culture, determining if organisms identified histologically are significant to the disease process may pose a diagnostic challenge for the pathologist and a therapeutic dilemma to the clinician. Though it is likely that more cultures were submitted than were recorded in our database (such as if a culture sample was submitted prior to or after biopsy without being linked to the previous submission), the low count in our database is still surprising.

The evaluation of equine dermatologic conditions requires an approach that combines clinical and histological features with ancillary diagnostics. Equine practitioners should consider submitting tissue for ancillary diagnostics concurrently with biopsy specimens, and should clearly indicate this on the submission form so that the histologic changes can be interpreted in light of the culture results. Consultation with a veterinary microbiologist or board-certified veterinary dermatologist may be warranted for certain cases. In combining all available information and collaborating with other veterinary professionals and specialists, significant improvements can likely be made in the diagnosis, treatment, and understanding of equine cutaneous lesions.

In summary, results of this study demonstrate that non-neoplastic equine cutaneous biopsies are predominantly inflammatory, but the etiology driving inflammatory lesions is largely undetermined by histopathology alone. Histopathology is only 1 of many important diagnostic tools available to the equine practitioner. The histologic description and morphologic diagnosis must be considered in light of host factors such as breed and age of the horse, the distribution of the lesion on the animal, timeline of lesion development, response to therapy, and environmental factors including husbandry and season. In light of our findings, combining the insights gained from histopathology with the clinical aspects of the case and with culture results and other diagnostics is extremely important when it comes to reaching a definitive diagnosis for inflammatory lesions of equine skin.