Discussion
Bacterial septic arthritis is a life-threatening cause of severe lameness in the horse, often resulting from a penetrating injury (Baxter 2004). There is limited literature discussing the involvement of fungal organisms(Riley et al. 1992, Madison et al. 1995, Swerczeket al. 2001, Sherman et al. 2006, Cohen et al.2008, Doyle et al. 2013). The literature reports treatment using a combination of synovial debridement and lavage with systemic and IA antifungal chemotherapy. Cohen et al., (2008) report the first case using intravenous regional perfusion of antifungal chemotherapeutics.
Wickerhamomyces anomalus (formerly Candida pelliculosa ) is widespread within nature often found in soil and on plant material and is considered a level 1 biosafety organism, safe for healthy humans (Satora et al. 2014). It is an opportunistic pathogen of humans and animals(Kurtzman 2011) and has been isolated from infected human lungs (Kurtzman 2011), involved in neonatal fungemia (PAUL et al.2020) and ocular keratitis(Kamoshita et al. 2015) cases.Candida species are well known to colonise the skin and mucosal surfaces of humans and often form biofilms making their eradication challenging (PAUL et al. 2020).
Cohen et al., (2008) report a 3-year-old standard-bred filly that developed bacterial metacarpophalangeal joint septic arthritis following IA medication with hyaluronic acid. Subsequent fungal septic arthritis developed due to Candia utilis which was isolated on culture. Cytology revealed a TNCC of 15x109cells/L, TP 18g/L. The filly was managed with systemic and IA fluconazole in combination with IVRP amphotericin B and arthroscopic lavage. A lytic subchondral bone lesion was identified on radiographs. The filly developed a recurrence of the fungal infection 8 weeks following discharge that was managed similarly. Eleven months after discharge from the hospital the filly was pasture sound(Cohen et al. 2008). Similarly, Doyle et al., (2013) report a case of fungal septic arthritis (Candida glabrata) of a 2-year-old gelding that was suspected to have had IA medication of the right metacarpophalangeal joint with corticosteroids. Cytology revealed a TNCC of 17.2x109cells/L, TP 69g/L. The gelding was euthanised due to a poor prognosis and evidence of lytic radiographic changes on the proximal sesamoid bones. There were significant findings reported on post-mortem (Doyle et al. 2013). Swerczek et al., (2001) report the case of a 6-year-old Thoroughbred gelding with a Scedosporium prolificans infection of the left metacarpophalangeal joint with a recent history of intralesional hyaluronic acid and antimicrobials to treat an inflammatory interosseous tendon injury. The horse was treated with IA corticosteroids and antimicrobials before being euthanised 10 days later(Swerczek et al. 2001). In 1995, Riley et al., reported a case of fungal arthritis caused by Candida famata in a 4-year-old Thoroughbred that sustained a laceration to the left metacarpophalangeal joint. The laceration appeared to heal well with systemic anti-microbials. However, 21 days after the initial injury the horse was markedly lame at the walk with subchondral lucencies evident on radiography. Subsequent treatment consisted of systemic antimicrobials, systemic amphotericin-B and potassium iodine. The mare was retired from racing and was clinically normal after 1 year with a normal foal.
The case reported here and that reported by Cohen et al., (2008) and Doyle et al., (2013) both received IA corticosteroid therapy in the days leading up to the development of lameness. The use of IA corticosteroids has been associated with fungal and bacterial arthritis in human patients and this is likely a result of local immunosuppression(Habibet al. 2010, Cohen 2011, Bufalari et al. 2016, Leclere 2017, Fang et al. 2019). Further to the immunosuppression caused by corticosteroids, if combined with antimicrobials this may well increase the risk of fungal infection further. In one human study, Candida albicans was the most common fungal organism involved in prosthetic synovial sepsis with recent (within 90 days) antimicrobial administration as a significant risk factor(Riaz et al. 2020). This re-enforces the need for responsible antimicrobial use especially given the evidence suggests that concurrent antimicrobial administration with corticosteroids does not significantly reduce the risk of iatrogenic synovial sepsis (Braid 2019) and could increase the risk of fungal septic arthritis (Fang et al. 2019, Riaz et al.2020). The incidence of septic arthritis following intra-articular injection is reported to be <0.1% following injection of more than 300,000 joints (Pezzanite et al. 2022).
The literature is sparse concerning the safe use of anti-fungal chemotherapeutics within synovial structures with some extrapolating treatment regimens from ophthalmic cases on the basis that if it is safe to use in the eye it is likely safe to use in a synovial structure. Voriconazole and fluconazole were used IA without any noted adverse effects in this case. Amphotericin-B is a well-documented model for inducing arthritis in the horse and for this reason, was not a first-line treatment choice (McIlwraith et al. 1979, McIlwraith and Van Sickle 1981). The authors did consider intravenous regional perfusion as reported by Cohen et. al., 2008, but felt there was little advantage over IA use given that a therapeutic concentration was required within the joint and thus still presented the risk of chemically induced arthritis. Amphotericin -B was later used in combination with hyaluronic acid intra-articulary which resulted in a transient but marked increase in lameness post-administration.
The initial management of this case reflects the standard approach to bacterial synovial sepsis with arthroscopic lavage, and systemic and local antimicrobials (Pille and Martens 2009). Within 6 days of the initial surgery, the causal organism had been identified as aCandida species which then underwent further typing. This rapid identification of a fungal organism allowed prompt anti-fungal treatment to be initiated. Unfortunately, it took 3 days to source an appropriate anti-fungal chemotherapeutic for IA use. The initial choice was made whilst awaiting fungal sensitivity results which did not arrive until day 23 and revealed only intermediate sensitivity to voriconazole. It should be noted that the initial 24-hour fungal culture obtained on blood agar was significant and somewhat surprising given the frequently unrewarding (Dumoulin et al. 2010) results of synovial fluid culture without enrichment. Culture and sensitivity in this case were pivotal as they guided treatment from the more commonly encountered bacterial synovial sepsis to one of fungal origin. It is also valuable in prognosticating given what we already know about fungal septic arthritis and the poor prognosis associated.
Antimicrobial therapy was continued despite a negative bacterial culture since bacterial isolation without enrichment is often not achieved with rates of isolation as low as 25% (Robinson et al. 2016) and there was potential that a bacterial component may have been missed particularly given the heavy fungal growth. It is considered that bacterial colonisation of this already severely compromised joint may have further complicated the management of this already challenging case. On reflection, the use of a blood culture enrichment technique may have been useful and allowed the cessation of antimicrobials at an earlier time in the interest of anti-microbial stewardship. There are complex relationships between many commensal organisms including bacteria and fungi. These relationships may be synergistic, saprophytic or antagonistic and due to their opportunistic nature co-infection is possible (Soll 2002, Peleg et al. 2010). However, there are certain gram-negative bacteria (Escherichia coli ,Pseudomonas aeruginosa, Acinetobacter baumannii and Burkholderia cepacia) that have shown an ability to kill or attenuate someCandida species. The exploitation of these attenuation mechanisms is likely to prove advantageous in the treatment of infections in the future particularly given the increasing problem of antimicrobial resistance(Peleg et al. 2010).
There were financial constraints and anti-fungal chemotherapeutic availability issues that dictated some of the clinical management, in this case, resulting in a delayed 2nd surgery. However, by day 23 the owners had reviewed their financial situation and an extension of the budget allowed for a 2ndarthroscopic lavage and injectable fluconazole was then available.
An interesting feature of this case was the persistently low TNCC values despite a worsening clinical picture. The authors cannot offer a definitive explanation as to why this may have occurred, although it does seem consistent across other reports that TNCC is generally lower than in bacterial cases with values pre-lavage frequently reported in the mid-teens(Sherman et al. 2006, Cohen et al. 2008, Doyle et al. 2013). In the human literature, it is reported that corticosteroid use can dampen the inflammatory response(Kohli and Hadley 2005) and whilst that is certainly possible pre-lavage it is hard to believe potent immunosuppression persists thereafter. Another feature exhibited by this case is the slow insidious onset of clinical signs similar to that reported in other cases within the literature with lameness not present until days and sometimes even weeks after the initial suspected causal event(Cohen et al. 2008, Doyle et al. 2013), this too could be linked to the anti-inflammatory properties of the corticosteroids. However, is more likely representative of the indolent nature of fungal arthritis(Kohli and Hadley 2005). This case highlights the importance of monitoring the response to treatment for synovial sepsis using not only the synovial fluid parameters but also the clinical presentation of the horse, the degree of lameness and the physical appearance of the joint (Murray 2012).
Regarding fungal sensitivity testing, it is important to note that there are no veterinary-specific clinical breakpoints for fungal species (Sellon and Long 2013). Instead, human-derived clinical breakpoints are used. It is important to consider this as several distinct aspects determine these clinical breakpoints: minimum inhibitory concentration (MIC) distribution of the fungal species against the drug of interest, the pharmacokinetics (PK) and pharmacodynamics (PD) of the drug of interest in the body and the clinical outcomes of treatment (Toutainet al. 2017).
The MIC distribution of a particular fungal species is unlikely to differ between human and animal origin but the PK/PD data can vary greatly between humans and animals and even between animal species. Consequently, human-derived breakpoints may not be entirely appropriate for use in animals.
A further issue is that, even in human medicine, fungal-species-specific breakpoints for W. anomalus have not yet been described(Kiddet al. 2021). Therefore, breakpoints for other Candida species are used for interpretation but there may be fungal-species-specific differences which might influence the appropriateness of these (Procop et al. 2020).
In this case, the isolate of W. anomalus was categorized as intermediate to voriconazole using human-derived breakpoints for otherCandida species . This is the best estimate of the status of that isolate but does not guarantee that it was truly sensitive or resistant.
However, as the voriconazole treatment was administered intra-articularly (i.e. at an increased concentration), the concentrations achieved may have been sufficient for the inhibition of fungal growth. Recently the European Committee on Antimicrobial Sensitivity Testing (EUCAST) reclassified the intermediate categories in some cases to “Susceptible, Increased Exposure” to allow for such situations (Arendrup et al. 2020). Although, we do not know how the intra-articular environment, with cellular infiltration and debris, might have impacted the activity of the drug.
The gross post-mortem and histological findings, in this case, were extensive and beyond the previous expectations of the authors. The fact that the joint space was almost occluded by synovial proliferation highlights the severity of the inflammatory process. Additionally, it begs the question: had the fungal infection been cleared, how would the authors resolve this marked inflammatory response within such a high-motion joint? Ultimately, the answer is likely that it wouldn’t have been possible to fully resolve and return this horse to athletic soundness irrespective of the significant osseous changes. This supports the outcome reported by Cohen et al., (2008) where 11 months after hospital discharge the filly only achieved pasture soundness.