Diagnosis and treatment of canine chronic hepatitis

Introduction 

Box 1. Key histopathologic features of chronic hepatitis.

Chronic hepatitis (CH) can occur in any breed of dog, and the onset can be insidious. It progresses to end stage cirrhosis when significant fibrosis and nodular regeneration develops, and is defined histopathologically by certain key features, as outlined in (Box 1) 1. It is critical to differentiate CH from a histologic diagnosis of non-specific reactive hepatitis, where mild-to-moderate inflammatory infiltrates in the portal, lobular and centrilobular regions are present without evidence of cell death or degeneration. These infiltrates are due to the escape of inflammatory cytokines and endotoxins from disease elsewhere in the splanchnic bed 2.

Etiology

In the majority of dogs with CH the etiology cannot be determined, so-called idiopathic CH 3 4, but various causes or potential causes are worthy of note.

Several studies have failed to identify hepatotropic viruses in dogs with CH, but histopathological and/or serologic evidence of Leptospira bacteria have been identified in laboratory colonies of dogs, and more recently molecular means have identified leptospiral organisms in dogs with granulomatous hepatitis 5. Whether it is the organism or an immune reaction to the organism which causes the CH is unknown. Leishmaniasis is associated with granulomatous CH, whilst other bacterial (Bartonella), rickettsial (Ehrlichia, Anaplasma) and protozoal (Neospora, Toxoplasma, Sarcocystis) infections may cause canine CH. These infections, however, are more often acute to subacute, and part of a more systemic disease process.

Several drugs and supplements have the potential to cause CH in the dog, and clinicians should be vigilant in obtaining a complete medication history 6. Most drugs can potentially cause acute liver injury, but a few, including anticonvulsants (phenobarbital, primidone and phenytoin), oxibendazole, lomustine (CCNU), amiodarone, mitotane and NSAIDs can lead to chronic hepatic inflammation.

Copper toxicity is also a potential etiology. Normally, many dogs consume excess copper (Cu) in their diet. Copper entering the liver must bind to Cu-binding proteins or be excreted in the bile, since free Cu causes oxidative stress leading to hepatocellular death. Normal hepatic Cu concentrations in the dog are 120-400 µg/g dry weight (DW) 7. Hepatic damage (as evidenced by increased serum alanine aminotransferase (ALT) activity and morphologic changes) begins when levels exceed 1000 µg/g DW, and damage is invariably present with values of 1500 µg/g DW or higher 7 8 9. There is, however, considerable phenotypic variability in an individual dog’s response to excess Cu. Some dogs have toxic hepatic Cu levels but no evidence of liver damage, while others have only mild elevations in Cu and severe damage 9 10 11. Although any breed of dog can accumulate copper, several breeds exhibit a predilection (Table 1) 7. In some dogs, e.g., the Bedlington Terrier, copper accumulation is due to genetic aberrations in copper-handling proteins. Accumulating evidence, however, suggests that dietary copper excess also contributes to the increasing incidence of copper-associated CH (Cu-CH) reported in the last two decades 10 11. Some 20 years ago many petfood companies switched dietary supplementation from copper oxide (which has a very poor bioavailability) to the more bioavailable copper chelates. This change, coupled with the fact that the US National Research Council has not established a maximum limit for dietary copper, has resulted in some commercial diets containing excess amounts of highly bioavailable copper 12 13. In Europe FEDIAF¹ has established a maximum value for copper concentrations in canine diets, although studies suggest that dogs, particularly those with breed predilections, may accumulate hepatic copper when fed diets with copper concentrations below this level 14 15. Several studies have now demonstrated that dogs (both with and without CH) over the last two decades have higher hepatic copper concentrations compared to similar populations of dogs pre-1998 10 11. A diagnosis of Cu-CH requires evaluation of a hepatic biopsy specimen which will show the presence of CH accompanied by rhodamine-positive copper accumulation, primarily in centrilobular hepatocytes, and elevated hepatic copper levels (> 400 µg/g DW, typically greater than 1000 µg/g DW). There are, however, several challenges in making the diagnosis of Cu-CH. These include lobe-to-lobe variability in copper concentration, the presence of significant fibrosis which can decrease copper levels, the challenge that regenerative nodules lack copper accumulation, and the fact that later-stage inflammatory/fibrotic changes complicate determination of lobular distribution. 

¹ Fédération européenne de l'industrie des aliments pour animaux familiers

A diagnosis of immune-mediated CH is often considered when no other etiology has been identified. Although specific criteria for such a diagnosis have not been developed, an immune basis in dogs with idiopathic CH is suggested by the presence of a moderate-to-marked lymphocytic infiltrate on histopathology, positive serum auto-antibodies, a familial history of CH, an association with other autoimmune diseases (e.g., hypothyroidism, atopy, inflammatory bowel disease), gender (females are generally more likely to be affected) and a favorable response to immunosuppressive therapy 13. A presumptive clinical diagnosis of immune-mediated CH requires the meticulous elimination of other potential etiologies (infectious, environmental or foodborne toxins, drugs).

Signalment and clinical signs 

Chronic hepatitis can occur in any dog including crossbreeds, but several breed predilections exist (Table 1) 16. Chronic hepatitis is generally a disease of middle-aged dogs, but cases have been reported as young as 5 months and as old as 17 years. There is a female predisposition in Labradors, Dobermans, Dalmatians and English Springer Spaniels, and a male predisposition in Cocker Spaniels (Figure 1).

The most common clinical signs in affected dogs are non-specific and may include lethargy/depression and anorexia. Polyuria and polydipsia (PU/PD) are two of the earliest signs. More specific signs of liver disease, such as jaundice, hepatic encephalopathy and ascites, are less common and typically indicate the presence of advanced disease (Figure 2a) (Figure 2b) (Figure 2c).

Because of the reserve capacity of the liver, many dogs with CH are subclinical, with the disease picked up on routine blood screens that show increased serum liver enzyme activity. It is at this stage that diagnosis should be pursued, as therapeutic intervention in advanced disease is often less successful.

Clinical pathology

Serum alanine aminotransferase (ALT) is the best screening test for CH, although its sensitivity is only around 70-80%. Significant histological lesions can therefore exist in the absence of accompanying ALT elevation. The magnitude of the ALT elevation activity is typically greater than that seen for serum alkaline phosphatase (ALP) activity, and elevations in ALP occur later in the disease. In late-stage cirrhosis, serum levels of liver enzymes may fall markedly as hepatocytes are replaced by fibrous tissue. The frequency of other clinical pathologic signs is summarized in (Table 2). 

Total serum bile acids (TSBA) are not a screening test for CH. Using 20-25 mmol/L as a cut-off, the sensitivity for pre-and post-prandial bile acids to detect CH is only about 50%. Since bile acids are very sensitive to shunting of blood around the liver, the sensitivity for cirrhosis when portal hypertension and multiple acquired portosystemic shunts (MAPSS) exist increases to almost 100%. It is ill-advised to wait until TSBA are elevated before proceeding to hepatic biopsy, as by this point significant and perhaps irreversible hepatic changes have occurred.

As PU/PD is a common clinical sign in CH, accompanying isosthenuria is seen on urinalysis. A transient acquired Fanconi syndrome (glucosuria with normoglycemia) is associated with Cu-CH 7.

Imaging

Radiology of the liver in affected dogs is usually normal, so ultrasound is a required part of the routine work-up in all dogs with suspected CH. A summary of the ultrasound changes described in literature is shown in (Table 3). It is important to note that several studies have shown that there are no ultrasonographic criteria that can predict the presence of CH; in fact the liver may appear normal on a scan, even in the presence of significant disease 17 18 19. 

In advanced CH, the liver may be small and irregularly marginated (Figure 3) on ultrasound, with signs of portal hypertension. These include ascites, edema (especially visible in the gallbladder and pancreas), reduced portal flow velocity (mean velocity of < 10 cm/s, as compared to a normal range of 10.5-25.7 cm/s) or hepatofugal flow and the visualization of MAPSS, usually seen as a complex plexus of small tortuous vessels caudal to the left kidney 20.

Biopsy acquisition

The diagnosis of CH requires hepatic tissue sampling. Fine-needle aspirates are not adequate to make a diagnosis, and often result in misclassification of the disease process. Percutaneous ultrasound-guided biopsy with large (14 or 16G) needles can provide adequate samples for diagnosis if multiple cores are taken 21. The diagnostic accuracy of 18G needle biopsies, however, is questionable, as these result in a relatively small sample size, are subject to fragmentation when fibrosis is present, and may not enable sampling of abnormalities located in lobes other than the readily accessible left medial or lateral lobes. This is a problem, since there is often heterogeneity between different liver lobes in terms of histological severity and copper deposition. In general, accurate diagnosis of CH requires the pathologist to evaluate 10-12 portal regions, which is hard to attain unless multiple percutaneous biopsies are procured. Multiple biopsies can, however, increase the risk of bleeding.

Laparoscopy is the preferred technique to obtain hepatic biopsies. This enables gross evaluation of the entire liver, the extrahepatic biliary system and surrounding structures while permitting acquisition of multiple large specimens that consistently result in samples with a mean of 16-18 portal triads per biopsy specimen. For the diagnosis of CH, five samples from at least two different lobes are obtained; three for histopathologic evaluation and one each for aerobic/anaerobic culture and heavy metal quantification. 

Risks of biopsy include anesthetic complications (especially in patients with advanced liver disease), hemorrhage, air embolism (on laparoscopy), infection, pneumothorax and vagotonic shock. The primary concern is hemorrhage 22. Assessment of the bleeding risk in dogs with liver disease that have deficiencies in both pro- and anti-coagulants as well as in regulators of fibrinolysis is difficult. Prolongations of PT and aPTT occur in about 40% of affected dogs. Decreased fibrinogen, anti-thrombin and protein C activity also occur in many dogs, and mild anemia and thrombocytopenia are occasionally present. Based on the human literature and the limited work to date in dogs, guidelines for bleeding risk assessment are outlined in (Box 2) 13 21 22. 

Percutaneous ultrasound-guided liver biopsy (Figure 4) carries a greater bleeding risk than techniques in which hemostasis can be controlled locally (e.g., laparoscopy), although the risk of complications (defined as the need for transfusion or fluid resuscitation) appears to be low for both methods, at around 1-5% 13 22.

When high-risk patients are identified, it is unknown if therapeutic interventions – such as administration of blood products or vitamin K – will decrease the risk of hemorrhage following hepatic sampling. The exception is the dog with low von Willebrand factor activity, which should be given cryoprecipitate and desmopressin. Thus the recommendation in high-risk dogs is to pay strict attention to technique and monitor them carefully in hospital after the biopsy procedure for 12 hours, and to be prepared to rescue with blood product support if necessary 13.

Biopsy interpretation

Evaluation of liver biopsy specimens requires H&E, Sirius red or Massons’ trichrome (for fibrosis) and Rhodanine (for copper) staining 21. The pathologist should comment on the type, location and quantity of inflammation, fibrosis, degenerative change (lipidosis, vacuolar change, lipogranulomas), the presence, location and extent of cell death and ductular reaction, and the lobular distribution and amount of copper staining (Figure 5). Special stains for infectious organisms are indicated in some cases, particularly in pyogranulomatous hepatitis. An exchange of information between the clinician and the pathologist may be necessary to maximize the value obtained from biopsies. In some cases, evaluation of samples by a pathologist (and internist) with expertise in hepatic histopathology and medicine should be considered.

Treatment

Treatment is targeted to etiology. Suspected infectious agents should be treated with appropriate antimicrobial therapy, and toxic agents or drugs eliminated from the dog’s environment. Any increase in hepatic copper in a dog with CH should be treated. Treatment of Cu-CH is summarized in Table 4 and involves dietary Cu restriction and methods to chelate Cu or prevent intestinal absorption (penicillamine and zinc) 7. Concurrent administration of hepatoprotectant and anti-oxidant medications is indicated (S-adenosylmethionine, vitamin E +/- ursodeoxycholate). Some dogs with Cu-CH have intense inflammatory infiltrates and may benefit from a short course of anti-inflammatory corticosteroids.

Bedlington Terriers, Dalmatians and young dogs with markedly elevated (> 3000 µg/g DW) hepatic Cu will likely need lifetime dietary therapy combined with Cu chelation. In other dogs, the time necessary to achieve normal Cu balance with penicillamine and low copper diet is poorly defined. Some work in Labrador Retrievers suggests that duration of chelation is related to initial hepatic Cu concentration, with about 6, 9 and > 12 months required for 1000, 1500 and 2000 µg/g DW respectively (Figure 6). Whether this is true for other breeds is unknown. Expert opinion is that some dogs “de-copper” more easily than others and that this is often independent of hepatic Cu concentration 13.

Ideally, re-biopsy with qualitative and quantitative Cu determination should define when chelation can be stopped. If this is not possible then serum ALT is used as a surrogate marker, whilst recognizing that levels can be normal despite ongoing histologic inflammation; chelation should therefore be continued 2-3 months beyond normalization of serum ALT. Although some limited work suggests that fine-needle aspirates stained with rhodamine may be useful in monitoring hepatic copper, this is not recommended until further studies are done.

In some dogs, normal Cu balance is restored yet serum ALT and histologic evidence of inflammatory disease persists. These dogs either did not have Cu-CH, or the damage has exposed neoepitopes and incited a self-perpetuating immune disease.

Typically, all affected dogs stay on a Cu-restricted diet, but this is often insufficient to maintain normal hepatic Cu, although it is hard to predict which dogs will need additional therapy. In general dogs with high initial hepatic Cu (> 2000 µg/g), dogs with a family history of Cu-CH, and dogs in which serum ALT does not normalize within a 6-8 month period of chelation will often receive an appropriate diet in combination with maintenance penicillamine or zinc therapy.

Limited studies suggest that some dogs with idiopathic Cu-CH truly have immune disease and will go into remission on appropriate therapy, although prospective clinical trials of immunosuppression in suspected immune CH are lacking. Corticosteroids, azathioprine, mycophenolate and cyclosporine have been advocated to treat presumptive immune hepatitis in dogs (Table 5), although again none have been evaluated in prospective clinical trials. Often dogs with presumptive immune disease are also placed on concurrent hepatoprotectant medications. 

Again, the optimum therapeutic endpoint for judging treatment response is normalization of hepatic histology, but this is often impossible, so ALT activity may be used as a surrogate marker. The timeframe for remission in dogs with immune CH is unknown. Control of enzymes may take 2-3 years in humans, although better long-term response is seen when enzyme activity is controlled within 3 months. As histologic improvement lags behind clinical and laboratory improvement (in humans by 3-8 months), treatment recommendation adjustments must go beyond laboratory remission by several months before attempting treatment withdrawal. Stable laboratory data within reference intervals over 12-18 months may be adequate to consider tapering medication. The relapse rate in dogs is unknown but in humans is up to 50%. Recapture of disease control with re-induction of primary therapy is often promptly achieved.

Prognosis and complications

Once diagnosed, dogs with CH typically experience progressive disease. Survival times have been reported in several studies, all of which studies were retrospective 13, and dogs were treated with a variety of medications and diets. In 10 studies with survival data (n=364 dogs) the mean survival was 561 days +/- 268 days. In dogs with biopsy-proven cirrhosis, survival was considerably less, 23 +/- 23 days (n=39). The clinicopathologic factors associated with a poor prognosis are hyperbilirubinemia, increased PT and aPTT, and hypoalbuminemia. The presence of ascites and the degree of fibrosis on biopsy are also negative prognostic signs; the one exception may be in Cocker Spaniels with CH, where dogs with ascites can have prolonged survival.

Complications of CH in the dog include portal hypertension, ascites, hepatic encephalopathy, gastrointestinal ulceration and coagulopathies (both bleeding and thrombosis) 20 23 24. Bleeding is more common with end-stage disease, and thrombosis occurs more often where other pro-thrombotic factors are involved, such as systemic inflammation, surgery or corticosteroid therapy 20. The incidence of secondary bacterial infection is poorly documented in dogs with CH, but appears to be low at around 5% 24.

Chronic hepatitis can occur in any breed of dog and the onset can be insidious; significant histological lesions can exist in the absence of accompanying serum liver enzyme elevation. Multiple biopsy samples are required for a definitive diagnosis, although biopsy can carry some risk for the patient. Targeted treatment is preferred wherever possible, although the causative agent may not be identified in many cases, and therapy should be continued for some months after clinical signs resolve.