Aspergillus and Fusarium are reported as the most common causes of keratomycoses (Barnett et al. 2004; Brooks 1999; Stoppini et al. 2003; Lassaline et al. 2010) and were common in the conjunctival swabs from healthy horses in this study (Table 1). However, it was the black pigmented fungal genera, namely Exoserohilum, Alternaria, Curvularia and Cladosporium that accounted for the majority (57%) of fungi identified in cases of keratomycoses in Townsville from 2013 to 2017 (Table 2). As shown in Table 1, the prevalence of yeasts was low (15%) with Rhodotorula being the most common genus isolated (12%). Rhodotorula is often associated with water contamination or poor bedding quality (Khosravi et al. 2014) and despite its presence in the conjunctiva, it does not appear to be an important cause of equine keratomycosis. There was only one isolate each of the yeasts, Kodomaea ohmeri, Pseudocercospora kadsurae, Aureobasidium, and Cystobasidium sloofiae. Their clinical significance is uncertain as only Candida and Trichosporon species have been implicated in keratomycosis in Townsville.
Since most of the mould genera recovered from normal eyes in this study are similar to those from equine keratomycosis cases, it is likely that fungi in the conjunctival fornix are able to invade the injured cornea (Reed et al. 2013). Although not shown in the Tables, repeated sampling of the same eye yielded different fungal genera in most horses, reinforcing the opinion that fungi in the conjunctival fornix are temporary residents that originate from the environment (Reed et al. 2013).
As ninety-five percent of the population studied were females, no conclusions could be drawn concerning gender. No previous studies mention any differences in the conjunctival flora when comparing the gender of the horses (Hampson et al. 2019). Gender differences have been found in domestic rabbits and pigs (Cooper et al. 2001; Davidson et al. 1994). Also, no significant differences were found between breeds in most studies (Hampson et al., 2019) but one study from Iran found that Caspian miniature horses had high colony corming units (CFU) of fungi per sample compared to other breeds (Khosravi et al. 2014). They believe that this might be due to the horses’ smaller stature, which brings their head and conjunctiva into close contact with feed allowing the saprophytic fungi to be transferred to eyes.
Statistical analysis between two different age groups was done in this study using age groups (< 15 years,>15 years,) and no association was found between age and frequency of ocular fungal organisms. This statement agrees with many studies (Barsotti et al. 2006; Johns et al. 2011; Hampson et al. 2019; Sgorbini et al. 2008), in which the authors stated that age had no effect on the frequency or type of fungal organisms found. But one study by Andrew et al., (2003), found that there was an increased number of Gram-negative bacterial and fungal isolates in younger animals. This difference may be because of ocular surface defence mechanism differences between age groups (Cooper et al. 2001).
The average rainfall (mm) datasets were very different between the two periods studied, 50 mm, and 239 mm respectively. Even so, no association between the frequency or type of fungi and rainfall was shown in our study, which agrees with other studies (Samuelson et al. 1984; Hampson et al. 2019).
In other studies, seasonal factors played a large role in the prevalence of keratomycosis with the greatest prevalence of disease being in the late summer, early autumn (Hendrix et al. 1995; Gaarder et al. 1998; Grahn et al. 1993). Summer was chosen as the sampling period as this was the time when most of the cases of equine keratomycosis are diagnosed and when environmental conditions are optimal for fungal growth in the environment. It was also at this time when the fungal burden in the pasture or hay is expected to be high. The horses in this study were sampled in December and January, just prior to the expected peak in keratomycoses cases (Gaarder et al. 1998; Grahn et al. 1993). However, 38% (n = 5) of the cases of equine keratomycosis in Townsville (2013–2017) occurred in winter when it is dry, with average daytime temperatures of 26°C, indicating that cases could be non-seasonal. Laboratory cases however, represented only a small portion of the total cases diagnosed clinically as keratomycosis. They tended to be the more severe and recurrent cases where horse owners more readily agreed to laboratory testing. Thus, the laboratory results are biased and may not reflect the true seasonality of keratomycosis. The cases in dry winter months could be explained by more woody plants being present in the pasture as well as feeding with hay at that time. The woody plants and hay stems can traumatise the cornea allowing fungi to attach to and penetrate the cornea (Brooks 1999). Studies would have to be carried out in the winter months to determine whether there is a seasonal difference in fungal diversity and burden in the conjunctival fornix of horses.
The interpretation of susceptibility to antifungals is difficult since the clinical breakpoints for antifungals are lacking. So, the data from previous publications and CLSI documents were used to interpret the pattern in this study (Clinical-&-Laboratory-Standards-Institute 2017; Pearce et al. 2009; Marangon et al. 2004). In earlier studies, it was reported that fungi isolated from horses’ eyes were generally resistant to fluconazole and ketoconazole and susceptible to natamycin, nystatin, and miconazole (Moore et al. 1988). Other studies reported that most fungi were susceptible to itraconazole and voriconazole, and resistant to fluconazole (Brooks et al. 1998; Ledbetter et al. 2007; Pearce et al. 2009). The latter studies are similar to our findings, where 109 out of 121 fungal isolates were resistant to fluconazole. Antifungal testing of fungi isolated from the ulcerative corneas of keratomycosis cases in Townsville from 2013–2019 also indicated that fluconazole is the least effective antifungal drug (Table 4). Moreover, most studies have shown that equine-origin fungal isolates have poor susceptibility to fluconazole (Brooks et al. 1998; Grahn et al. 1993; Pearce et al. 2009;). We found that amphotericin B was not a very effective in vitro agent for the treatment of fungi associated with equine keratomycosis in this study (Table. 3). In this study, fungal isolates displayed high susceptibility to voriconazole and ketoconazole while voriconazole was the most effective drug against moulds in vitro. This agrees with what others have reported in Australia and internationally (Hampson et al. 2019; Pearce et al. 2009). The differences in susceptibility patterns between studies might be due to geographic, temporal, or climatic variation and evolving mechanisms of fungal resistance.
Voriconazole and itraconazole appear to be a better choice for Aspergillus than ketoconazole and fluconazole based on the in vitro susceptibility data generated by this study. This agrees with a study from Florida, where Aspergillus spp. showed poor susceptibility to fluconazole compared to itraconazole and miconazole (Brooks et al. 1998; Ledbetter et al. 2007). In a recent study done on a human invasive fungal infection, voriconazole appeared to be highly effective as a treatment for human aspergillosis (Chopin et al. 1997). Moreover, Meletiadis et al. (2007) showed that voriconazole has a broader spectrum of activity than other triazole antifungals and evidence of a concentration-dependent sigmoid pattern of fungicidal effect on Aspergillus species. Voriconazole can effectively penetrate the cornea in clinically normal horse eyes (Clode et al. 2006). Given orally it will reach 50% of its original concentration in the aqueous humour.
For Fusarium spp., natamycin is highly recommended by many studies (Brooks et al. 1998; Ledbetter et al. 2007). But natamycin was not tested in this study due to the lack of availability of the drug. Natamycin is a polyene compound that would be expected to have a similar susceptibility behaviour to amphotericin B and nystatin. According to these in vitro results, nystatin worked well on Fusarium, but amphotericin B did not. The recommended antifungals for Fusarium appeared to be voriconazole, ketoconazole, and nystatin.
For Penicillium spp., ketoconazole and nystatin seemed to be better choices than fluconazole and amphotericin B based on the in vitro susceptibility data generated by this study. A similar pattern was established in Exserohilum, Cladosporium, and Curvularia species. The current study demonstrated poorer susceptibility of Nigrospora and Trichophyton spp. to every antifungal tested. But this might be because only very few isolates of these species were tested. The Phoma species of fungi exhibited susceptibility to nystatin only. However, Phoma was not isolated in Townsville from the keratomycosis cases.
Only one isolate each of Epicoccum, Alternaria, Verticillium, and Rhizopus species were tested. All of them, except Epicoccum, were susceptible to ketoconazole while itraconazole appeared to work well on Epicoccum and Verticillium. Since the number tested was very small, the susceptibility pattern of these species cannot be determined.
The yeasts were highly susceptible to ketoconazole in this study but less susceptible to fluconazole and voriconazole (Table. 3). This study provided evidence that certain costly antifungal medications such as fluconazole and amphotericin B may no longer be effective against most fungi and should be considered as the less ideal drug of choice in treatment plans for equine keratomycosis.
Although ketoconazole proved to be effective in this study, some studies disagree (Chopin et al. 1997; Ledbetter et al. 2007). Ketoconazole is poorly absorbed orally in horses, unless it is administered daily via stomach tube with hydrochloric acid. Furthermore, it inhibits hepatic P450 enzymes reducing enzymic breakdown of other drugs.
Most fungal isolates in this study were identified to species level based on their morphological criteria. Sequencing of the ITS region of the 23sRNA gene was done on representative isolates or those that could not be identified to species level. Both methods only identify to species level. Morphological identification is cheaper but requires a high level of mycological skills as the fungi must be encouraged in a variety of ways to produce fruiting bodies for identification purposes. Sequencing is technically simpler, but more expensive and cannot be carried out in house.
A weakness of our study was the number of fungi isolates tested for MIC was small, so, there might be a chance that these fungi were clonally related. The geographical range used in this study is small, only representing the Townsville region, and additional studies are needed to be able to represent the entire North Queensland. In vitro testing of fungi isolates provides relative but not an absolute indication of clinical response. The concentration of drugs achievable in the cornea might differ according to the pharmacokinetic and pharmacodynamic properties of the medications used. Additionally, host factors (immune responses and genetics) play a role in the in vivo effectiveness of the treatment (Hampson et al. 2019; Pearce et al. 2009). Moreover, the standardized in vitro criteria are not well established in horses for MIC testing, impeding the interpretation of the relationship between the MIC result and clinical outcomes. This method can only forecast the in vivo efficacy of antifungal drugs and the correlation between in vitro testing and successful in vivo treatment of fungi is not well established (Espinel-Ingroff 2003; Odds et al. 1998). Nevertheless, these MIC results can help clinicians when susceptibility tests are not available and as a preliminary guide in choosing the therapeutic protocol for equine keratomycosis and add to the information on fungal flora found in the eyes of horses living in tropical regions of the world.