Equine liver disease in the field. Part 2: causes and management

The approach to equine liver disease was explored in part 1 of this series (Tallon and McGovern, 2020). This article discusses selected causes of liver disease in more detail, providing updates from recent literature. These focus on toxic and infectious aetiologies, which typically cause subclinical disease at a herd level. Identification of a specific cause is not possible in many clinical cases and, instead, treatment is discussed based on clinical findings and biopsy results. Dietary modifications and supplementation for horses with liver disease is also reviewed. A more complete summary of causes of liver disease is illustrated in Table 1 but discussion of each of these in detail is beyond the scope of this article.

Causes of liver disease: what's new?

Toxic hepatopathies

Toxic hepatopathy should be considered if multiple horses on a premises are affected and, if suspected, hay and soil analysis should be performed. Mycotoxins are common in hay and feed. No clear association between aflatoxin contamination of forage with liver disease has been found, but fumonisin B₁, a toxin produced by Fusarium spp., has been shown to cause liver disease (Ross et al, 1993). More recently, the toxin was consistently identified in the forage of horses with liver disease but not in controls (Durham, 2017). Mycotoxin testing kits are commercially available and if contamination is suspected or confirmed, forage should be changed. In-feed mycotoxin adsorbents can also be used. Supplementation with a polymeric glucomannan adsorbent was shown to attenuate increases in gamma glutamyl transferase associated with Fusarium spp. ingestion in one study (Raymond et al, 2003) but further research is currently lacking.

High dietary iron levels can result in iron accumulation in hepatocytes (haemochromatosis), causing chronic hepatitis. It is worth noting that this is rare in adult horses; levels need to be extremely high or occur over a prolonged period. Research identified chronic hepatopathy and haemochromatosis in multiple horses associated with high iron levels in drinking water (between 2 and 240 times the recommended threshold) over a period of several years (Theelen et al, 2019). Iron toxicity is difficult to diagnose; many horses with liver disease have high iron levels regardless of cause, and haemosiderin accumulation is seen in the liver in many disease processes. Iron levels in drinking water in the UK can be a maximum of 0.2 mg/litre (Drinking Water Inspectorate, 2017). Cyanobacteria (blue-green algae) in drinking water produce microcystins that cause massive hepatocyte necrosis (Mittelman et al, 2016) and should also be considered when analysing water sources.

Hay or pasture contaminated with Senecio spp. (the most important of these being ragwort, Senecio jacobaea (Figure 1a)) is the most common cause of toxic hepatopathy in the UK. Early stages of disease are characterised by biochemical and histological changes without clinical signs. The pyrrolizidine alkaloids found in the plant are metabolised to toxic derivatives in the liver. These compounds alkylate nucleic acids and stop mitosis. Instead, affected hepatocytes become enlarged (megalocytosis) and when they apoptose, irreversible fibrosis occurs, making early diagnosis of this disease important. Ragwort is common in the UK and although disease incidence is fortunately reduced as a result of increased awareness, cases are still seen, especially in horses kept on badly managed pastures. Alsike clover and red clover (Trifolium spp.) (Figure 1b) are also widespread in the UK and can cause biliary duct hyperplasia with periportal fibrosis. Megalocytosis is not a feature. These plants rarely cause poisoning unless they make up at least 20% of the diet (Divers and Barton, 2018). It is unclear if toxicity is caused by the clover plant itself or a mycotoxin produced by a fungus (Cymodothea trifolii) commonly found on the plant.

Parasitic diseases

Fasciola hepatica, or liver fluke, has an estimated seroprevalence of up to 10% in both UK and Irish horse populations, with 2% of horses found to have adult fluke within the liver at post-mortem examination (Quigley et al, 2017; Howell et al, 2019). Horses with liver disease had significantly higher odds of Fasciola hepatica seropositivity (Howell et al, 2019), although it is worth mentioning that the number of seropositive horses was small, and careful interpretation is therefore necessary. The fluke migrates through the liver and causes damage to the liver parenchyma and bile ducts, where adults come to reside. Necrosis of hepatocytes is rare. Infection may be associated with co-grazing with ruminants as the fluke is the same population as that found in ruminant species, although horses are comparatively resistant to infection. Wet pastures also pose a higher risk. There is currently no licensed treatment for Fasciola hepatica in horses but triclabendazole (15 mg/kg PO*) or closantel (10 mg/kg PO*) may be prescribed using the veterinary cascade. Faecal worm egg counts are insensitive as a result of low numbers of adults and eggs and a long pre-patent period. A commercial enzyme-linked immunosorbent assay is currently available and has been demonstrated to be superior to faecal analysis for diagnosis, although sensitivity is still relatively low (71%) (Howell et al, 2019).

Viral hepatopathies

Three new flaviviruses thought to be associated with hepatopathy have been recently identified: non-primate hepacivirus (or hepacivirus A), Theiler's disease-associated virus (or pegivirus D) and equine pegivirus (pegivirus E). The first of these to be identified was non-primate hepacivirus, which replicates in the liver. There is little evidence of an association between non-primate hepacivirus infection and clinical disease, but experimental infection causes elevated levels of liver enzymes, with inflammation and necrosis seen on histopathology (Ramsay et al, 2015). This virus has been reported in the UK horse population.

Equine pegivirus has a documented prevalence of 18.2% (by polymerase chain reaction) in adult horses in north America, with a seroprevalence of up to 66.5% (Lyons et al, 2014; Ramsay et al, 2019). Infection was associated with a reduced risk of having increased liver enzyme activity in a study, suggesting it is unlikely to be a cause of hepatitis in horses (Ramsay et al, 2019).

The third virus, Theiler's disease-associated virus, is rare and has not been documented in the UK population to date. It was originally detected in an outbreak of serum hepatitis in the United States but all horses in the study were found to be co-infected with equine parvovirus hepatitis. In more recent cases of serum hepatitis, Theiler's disease-associated virus was not detected (Divers et al, 2018) and it was not associated with raised levels of liver enzymes in 802 racehorses (Ramsay et al, 2019), making it an unlikely cause of hepatitis in the horse.

Equine parvovirus hepatitis has been identified in the USA and appears to be the most clinically relevant hepatitis virus. The virus has been shown to be transmitted by biological products (tetanus antitoxin, stem cells, plasma) and causes serum hepatitis in some horses (Tomlinson et al, 2019a). In-contact horses have also been demonstrated to be affected (Tomlinson et al, 2019b) and more work is needed to characterise transmission and tissue tropism. Commercial screening for viral hepatopathies is available, although no specific anti-viral treatment is currently described. The importance of co-infection with these viruses is unknown.

Management of liver disease

Despite an appropriate diagnostic approach, a causative agent is not identified in many cases of liver disease, and treatment must be guided by clinical signs and biopsy findings. Even biopsy changes may also be non-specific in chronic disease and treatment is therefore supportive.

Inflammation

Inflammatory liver disease is typically characterised by a lymphocytic-plasmocytic infiltrate on histopathology and may be associated with several disease processes. Chronic active hepatitis is an idiopathic, progressive hepatopathy with both active inflammation and fibrosis seen on histopathology. This process may occur in conjunction with cholangiohepatitis, so biopsy is important to rule out an infectious component before starting treatment with corticosteroids. Inflammation is often a precursor for fibrosis; stellate cells proliferate and secrete collagen in response to tumour necrosis factor-α (TNFα) from macrophages (Theisse, 2015). Prednisolone or dexamethasone can be tapered and serum biochemistry repeated to assess response to therapy. In some cases, prolonged treatment of several months' duration may be required.

Bacterial infection

Antimicrobial therapy should be initiated if septic cholangiohepatitis is suspected; classical presenting signs include colic, pyrexia, icterus and possibly abnormal peritoneal fluid. Neutrophilic infiltrates are typically seen on histopathology. Infection is often associated with ascending enteric bacteria from the gastrointestinal tract. Ideally, antimicrobial therapy should be based on bacterial culture and sensitivity from the biopsy sample; however, this is often unrewarding (Box 1). In one study, 77% of cholangiohepatitis cases had negative culture results (Peek and Divers, 2000). Appropriate choices for empirical therapy include trimethoprim sulphonamides and ß-lactam/aminoglycoside combinations (Peek, 2004). Trimethoprim sulphonamides are often chosen because of their ease of administration. The use of protected antimicrobials such as fluoroquinolones and third generation cephalosporins should be based on culture and sensitivity results and response to treatment. If culture has been negative, adjustment should be made promptly if no clinical response is seen. Treatment should be continued until biochemical parameters have normalised and a clinical improvement is seen. A median treatment duration of 51 days (range 17–124 days) was reported in one study of horses with cholangiohepatitis (Peek and Divers, 2000).

Box 1.Case study: acute cholangiohepatitisPresentation: A 20-year-old Connemara pony presented with mild colic signs, inappetence and pyrexia (39.1°C) of approximately 48 hours' duration.
Serum biochemistry (abnormal results)

Haematology was unremarkable

Liver ultrasonography: heterogenous echogenicity with hypo- and hyperechoic areas diffusely distributed throughout the parenchyma (Figure 2). A liver biopsy was performed. Owing to the pyrexia and suspicion of cholestatic disease, empirical treatment with procaine penicillin, gentamicin sulphate and metronidazole was started. The pony also received flunixin meglumine.

Figure 2. Ultrasonographic image of the liver, showing heterogenous echogenicity with hypo- and hyperechoic areas diffusely distributed throughout the parenchyma (arrows).

Histopathology: multifocal moderate neutrophilic cholangiohepatitis and lymphangitis with mild biliary hyperplasia, multifocal cholestasis and moderate portal fibrosis, with an overall score of 6/14 (Durham et al, 2003). No severe irreversible changes or neoplasia were seen in the examined sections. Culture of biopsy samples did not yield any bacterial isolates.Progression and treatment: bile acid concentrations continued to rise despite treatment (239 umol/litre), and the pony became increasingly dull. Based on the lack of clinical response and lack of culture results, antimicrobials were empirically adjusted to enrofloxacin, metronidazole and procaine penicillin. The pony's demeanour improved and repeat ultrasonography at 7 days post admission showed a marked improvement in the appearance of the liver parenchyma. Serum biochemistry identified a reduction in levels of serum bile acids (134 umol/litre), although gamma glutamyl transferase levels had increased (360 umol/litre). Based on the continued decrease in levels of serum bile acids, antimicrobial treatment was discontinued after 14 days. Bile acid levels continued to decrease, with the last measurement (25.8 umol/litre) at 18 days. The pony was discharged at this point, and no further treatment was required. Serial blood samples following discharge identified a progressive reduction in gamma glutamyl transferase levels.Comment: failure to obtain a positive culture from biopsy is commonly encountered. In this case, a protected antimicrobial was used empirically based on lack of clinical response to initial antimicrobial therapy. It can be difficult to know how long to continue antimicrobial treatment for. Here, serum bile acid concentrations and clinical response were used to guide treatment duration. This case also highlights that gamma glutamyl transferase activity may continue to increase despite clinical improvement (Divers and Barton, 2018).

Fibrosis

Early fibrotic changes may be reversible with prompt recognition and treatment. Fibrosis is often preceded by inflammation or occurs in conjunction with it, so use of glucocorticoids may help to prevent fibrosis. The use of other anti-fibrotic therapies is extrapolated from human studies, but evidence is lacking in horses. Pentoxifylline (10 mg/kg PO BID*), a phosphodiesterase inhibitor, is used to treat chronic hepatitis in humans and its use has also been described in horses (Divers, 2015). Side effects are limited. Long-term colchicine treatment in humans with hepatic fibrosis appears to exert an anti-inflammatory, anti-fibrotic and immunomodulatory effect (Nikolaidis et al, 2006). Its use has been reported in horses at a dose of 0.03 mg/kg PO SID*. Colchicine should not be used if pyrrolizidine alkaloid toxicity is suspected or if megalohepatocytes are visualised on histopathology, as it is known to inhibit mitosis (Divers, 2015). Fibrosis in these situations is also irreversible so anti-fibrotic therapy will have little effect in any case. Reported side effects of colchicine in the horse include bone marrow suppression (Peek et al, 2007).

Diet

Dietary modifications are only indicated if liver function is compromised. Minimal evidence exists for any benefit from a low-protein diet in the vast majority of cases of liver disease and it is generally considered important that protein requirements are met. This should only be considered for cases with hepatic encephalopathy (Box 2), with a view to reducing circulating ammonia levels, although this approach is no longer recommended in humans (Yao et al, 2018). The equine diet is generally low in protein in any case, making adjustments challenging. A palatable diet should be provided, removing any possible sources of toxins. The liver is primarily responsible for glucose metabolism, and this should be supported with small, frequent, high carbohydrate meals. Branched chain amino acids include valine, leucine and isoleucine. In human patients, branched chain amino acids may increase the oxidation rate of carbohydrates and decrease the oxidation rate of fat and a meta-analysis found that branched chain amino acidenriched snack supplementation in the late evening to improve abnormal energy substrate metabolism in human patients with liver cirrhosis (Yao et al, 2019). An accurate dose is not known in horses but doses of 10 g/adult horse PO BID* have been described (Divers, 2015).

Box 2.Management of hepatic encephalopathy

• Hepatic encephalopathy is a result of severely reduced hepatocellular function and is not associated with any specific cause
• Severity of signs is correlated to reduction in function but is not related to the underlying disease process
• The condition is potentially reversible, and management should be instigated while awaiting results of blood tests or biopsy
• Clinical signs associated with hepatic encephalopathy include excitement, depression, cortical blindness, foot stamping, head pressing, yawning, vocalization, bilateral laryngeal paralysis, dysphagia and gastric impaction. Seizures are not typically seen
• Hepatic encephalopathy is usually severe enough to require referral for intensive care and this should be considered if the animal is severely dehydrated, dysphagic, ataxic, or severe acid–base abnormalities are identified
• Treatment aims to reduce ammonia concentration and production and to control neurological signs
• In a field situation:
• Small amounts of sedation (alpha-2 agonists, such as xylazine or detomidine) may be administered as needed to control neurological signs. Over-sedation may affect organ perfusion and worsen cerebral signs if the head is lowered
• Hypertonic saline (4 ml/kg IV) and/or mannitol (0.5–2 mg/kg IV over 10 minutes) to control increases in intracranial pressure
• Lactulose (0.3 ml/kg PO QID) is commonly used in an attempt to reduce ammonia formation in the gastrointestinal tract, although evidence for use in horses is lacking
• Placement of a tracheostomy tube should be considered before referral if bilateral laryngeal paralysis is suspected
• Avoid nasogastric intubation where possible; this may result in haemorrhage, leading to ingestion of serum protein and further generation of ammonia
• Antibiotics such as metronidazole and neomycin may be administered to reduce the number of ammonia-forming bacteria within the gastrointestinal tract.

Supplementation of antioxidants and vitamins should also be considered. A meta-analysis of randomised human trials did not find any benefit associated with the use of antioxidants for liver disease (Bjelakovic et al, 2011). However, there is a lack of evidence either way in horses and minimal adverse effects are reported. Vitamin C and E would be commonly used orally.

Milk thistle extract (silymarin) contains a compound called silibinin which was shown to have antifibrotic and antioxidative properties in experimental studies (Kim et al, 2013; Federico et al, 2017). However, a systematic review of human literature did not demonstrate any significant benefit in patients with chronic liver disease (Rambaldi et al, 2007). In healthy horses, minor improvements in antioxidant capacity were demonstrated (Hackett et al, 2013a). Although bioavailability has been shown to be <1% in healthy horses (Hackett et al, 2013b) and silibinin concentrations vary between commercial products, no adverse effects have been reported in horses and the product remains in widespread use.

S-adenosyl methionine supplementation is widely used in liver disease in small animal medicine. The compound is made in the liver and undergoes conversion to glutathione, an antioxidant. Sadenosyl methionine synthesis may be reduced in chronic liver disease, exacerbating hepatic injury. In people, a meta-analysis suggested that S-adenosyl methionine could improve liver function in chronic liver diseases (Guo et al, 2015) but evidence in the horse is currently lacking.

Vitamin C (ascorbic acid) is synthesised by the equine liver and has many antioxidant functions. Under normal circumstances this synthesis meets demand, and dietary supplementation is not required. Consideration should be given to supplementation in liver disease, particularly if the diet is lacking in fresh grass. It is worth noting that many commercial liver supplements do not contain ascorbic acid. A recommended dose is 20 mg/kg of ascorbyl monophosphate, as ascorbic acid is poorly absorbed in horses (Deaton et al, 2003).

Vitamin E should also be supplemented if the diet is lacking in fresh green forage. A soluble, liquid formulation containing natural vitamin E (α-tocopherol) is recommended; this has superior bioavailability and antioxidant properties (Brown et al, 2017). Vitamin A is also stored in the liver and similarly supplementation is not required if the diet contains green forage. Over-supplementation can lead to hepatotoxicity as a result of the effects of retinol on stellate cells in the liver (García-Cortés et al, 2016). Selenium and zinc are stored in the liver and supplementation in rats with fatty liver disease resulted in improved biochemical parameters (Shidfar et al, 2018). Iron-containing supplements should not be given.

Horses with photosensitisation (Figure 3) should be stabled during the day or have non-pigmented skin appropriately covered if turned out. Topical emollients may provide some relief. Care should be taken in horses that are diagnosed with liver disease — non-pigmented skin should be closely monitored for the development of photosensitisation.

Monitoring

Serum biochemistry is a useful tool for monitoring progression and response to treatment. An interval of 2–4 weeks should be left to allow for the longer half-life of some liver enzymes but should be repeated sooner if deterioration occurs. Repeat biopsies are not always necessary if there is a clinicopathological improvement but should be considered in cases that are non-responsive to treatment or if serum biochemistry was not initially diagnostic. An interval of 12–16 weeks should allow adequate time to assess response to treatment.

Conclusions

Identification of an underlying cause of liver disease can be challenging and is not achieved in many clinical cases. Treatment is often guided by clinical signs and histopathological changes seen on biopsy samples. Recent research has provided more information on toxic and infectious causes of liver disease. Dietary modification is not required for many cases of subclinical disease.

Key Points

• Consideration should be given to toxic and infectious causes of liver disease, particularly if more than one horse is showing evidence of liver disease on serum biochemistry.
• Further research is needed to understand the relationship between viral infection and clinical liver disease.
• Broad-spectrum antimicrobials should be initiated for suspected cases of cholangiohepatitis, even if culture is negative.
• Reduction of dietary protein should only be considered for cases of hepatic encephalopathy or severe liver failure where increased ammonia levels are a concern.
• Much of the recommended dietary modifications are extrapolated from human medicine and there is a distinct lack of evidence base for their use in horses.