The biochemical parameters used to assess liver pathology may be divided into two classes: the enzymes that reflect liver damage and/or cholestasis (discussed below) and the indicators of liver function (bile acids, ammonia, albumin etc. covered in a separate article). Serum enzyme screening for hepatobiliary disease is common in veterinary practice. Clinical signs associated with liver disease are wide-ranging and often non-specific, and consequently laboratory profiles are often run in patients that have a constellation of clinical signs that includes one or more of those seen in liver disease. More frequently, patients that are just "not well", or those that are clinically normal (in for a "check up" often young patients coming in for vaccination/deworming or geriatric patients), or because surgery is being contemplated, have test "panels" conducted that almost invariably include liver enzymes. It is against this background that one should assess the usefulness of laboratory tests. It is just too simplistic to evaluate liver enzymes by comparing the serum activity in patients with liver disease with those that are healthy. For a test to be really useful, it must be abnormal in most, patients with liver disease most of the time and normal in most sick patients, with signs that could represent liver disease as well as the latter "healthy" groups, most of the time.

As patients may be presented some time after the damage first occurred, a one-off serum activity is difficult to interpret in terms of severity. Furthermore, pathophysiological processes in diseases that are not primarily hepatic may generate fairly substantial serum enzyme activity (secondary liver disease or induced enzyme activity). Even those primary diseases, such as parenchymal damage, cholangitis, cholangiohepatitis, chronic hepatitis and diffuse neoplasia, may be accompanied by negligible or no increases in serum enzyme activity.

This paper compares the diagnostic utility of liver enzymes, when assessed against normal animals with that derived from assessment in a "routine" clinical environment. Insights gained during a long-term toxicology trial are also presented. For the purpose of this paper, liver disease will be broadly categorised into Acute hepatopathy (AH), Chronic hepatopathy, including both chronic fibrotic as well as vascular diseases (CH), Biliary pathology (BP) and Endocrine hepatopathy (EH).

Alanine Aminotransferase (Transaminase) (ALT)

ALT is reported to be essentially liver specific in the dog and cat. As AlT is principally an enzyme that is used in gluconeogenesis using the Pyruvate-Glucose route and the liver is the principal site of this process, it is very liver specific in carnivores. In herbivores, however, where gluconeogenesis follows the Fatty-acid/Krebs-cycle route (eventually Oxaloacetate to Glucose), there is relatively little enzyme in the liver of these species and consequently it is neither sensitive nor specific for liver disease in large animals.

Peak serum activity is proportional to the number of hepatocytes affected, but gives no indication as to whether the disease is diffuse or focal, and does not reflect the severity of the disease or its reversibility (i.e., is NOT prognostic, no matter how high the number). In order to set a prognosis one utilises the reported plasma half life, which is of the order of two-and-a-half days, implying that, after serum peak levels, a halving every 2 to 3 days is a good prognostic indicator. Earliest sensitivity data (Center's 1985, 150-case report) range from 47% for Chronic hepatopathy, 77% in Biliary pathology, 40% in Endocrine hepatopathy and 100% in Acute hepatopathy. The author's own data (165 cases of suspected liver disease) were 45% for Chronic hepatopathy, 60% in Biliary pathology, 67% in Endocrine hepatopathy and 71% in Acute hepatopathy, respectively.

Data obtained from the OVAH database (1252 cases on which serum ALT was determined), the data were 47% (CH), 100% (BP), 72% (EH) and 75% (AH) respectively. The interesting, if somewhat disappointing, index derived from this survey, is that the Predictive value of a positive result (ALT > 40 U/l) for the presence of liver disease is only 18% (i.e., if one obtains a laboratory result of ALT > 40 U/l {top-normal for the laboratory the author used}, then the probability that the patient has hepatic disease is 18%). If one sets the cut-off at twice-top-normal, however, this improves to 29%. Furthermore, a "normal" AlT result predicts the absence of hepatic disease with almost a 90% certainty.

Alkaline Phosphatase (ALP)

There are a number of sources of AlP and there is a unique sensitivity to drug and cholestatic induction by de novo synthesis, solubilization from membranes by bile salts and/or elution from membranes. These factors are reported to make ALP a test of relatively low specificity. However, its clinical utility arises from its reported sensitivity in detecting hepatic disease involving the bile duct system. Interpretation of results in "sick" dogs, that do not have canine Cushing's disease, is often confounded by the difficult-to-predict effect of endogenously produced and/or iatrogenically administered corticosteroids. Plasma half life is reported to be about 72 hrs in dogs. Early published sensitivity data (Center, 1985) range from 44% for CH, 91% in BP, 90% in EH and 57% in AH. The author's data (165 cases) were 49%, 67%, 67% and 74% respectively.

Data obtained from the OVAH database (1277 cases on which serum ALP was determined), the data were 50%, 100%, 81% and 71% respectively. From this survey, the Predictive value of a positive result (ALP > 190 U/l {top-normal}) for the presence of liver disease is only 21%. If one sets the cut-off at twice-top-normal, however, this improves to 32%. So, surprisingly, these data suggest that ALP is more specific for liver disease, in general, than is generally believed (when compared with ALT).

The diagnostic utility of AlP can be enhanced by exploiting the knowledge about enzyme induction.

As a rule, the increase in serum enzymes reflects a breakdown of cellular integrity (i.e., an increase in leakiness or frank membrane disruption) of the organ/s rich in that enzyme. This, of course, explains why the sensitivity of AlT for the chronic presentation hepatopathies is very poor because in many of those conditions there is very little loss of cellular integrity at any one moment in time.

In this context, AlP is an unusual enzyme, among those used diagnostically, in that the hepatocyte is, in fact, not a very rich source of the enzyme (explaining, to some extent, why it lags behind AlT as a diagnostic indicator of acute hepatopathies, more of that below). However, there are two pathological processes that INDUCE the liver to synthesize massive quantities of the enzyme, much of which is released and enters the blood. These two pathological processes are:

 Increased bile duct and bile canalicular pressure.

 Persistently high cortisol levels.

This bit of information allows for a very useful interpretative algorithm (rule of interpretation), namely: If the AlP activity is in excess of 2½-times top-normal (500 U/l in the database used for the statistics in this paper), then it is safe to assume that most of this enzyme is one (or both) of the inducible forms. That being the case, it is wise to spend time and effort on deciding which:

 Is there an increase in bile duct pressure?--(possibly answered by looking at urine bilirubin levels on a dipstick), or

 Is there a persistently high cortisol?--(possibly answered by urine SG, examination of skin/haircoat, typical "stress" leukogram, with or without highest Hct and/or platelet count).

Applying this algorithm to the above database, the following emerges:

 Cases with known gall-bladder or bile duct pathology78% of these were identified.

 Cases with high cortisol (Cushing's) 78% of these were identified.

Interestingly, 45% of the Diabetes mellitus cases and 44% of acute hepatopathies also "fit" the algorithm. It is tempting to speculate that a large proportion of the Diabetics could be Type II (insulin resistance) secondary to a primary hypercortisolaemia. Furthermore, 54% of hepatic neoplasia cases (principally secondary metastasis) including 63% of the Stage V lymphomas. This finding relating to neoplasia may be explained by the fact that metastatic neoplasms tend to start their growth within the liver in the portal triad areas, because that is where the hepatic vessels enter and consequently one of the earlier expected pathophysiological changes would be an increase in bile duct pressure.

An alternate algorithm is to add a rider to the first criterion (i.e., Is serum AlP > 500 U/l), namely: "....and is the increase in AlP disproportionately high when compared with the increase in AlT?". This tends to "fit" fewer cases (i.e., decreases the sensitivity, reducing the Cushing's cases identified from 78% down to 32%) but does make the algorithm more specific for the two types of pathology, reducing the acute hepatopathies (where the increase in AlP is probably simply the result of hepatocyte damage, as opposed to induction) from 44% down to 12%.

In horses AlP is also fairly sensitive for biliary-associated (cholestasis-induced) disease and may be useful in chronic liver disease (possibly due to bile duct pressure). Bone-derived AlP can cause quite significant elevations in young, growing foals (2-3 X top-normal). There are reports of GIT disease-derived elevations and placental pathology-derived AlP elevations. AlP is one of the most stable parameters in horse blood and in health will vary extremely little from day-to-day--indicating that one can set a reference range for each individual and detect early movement long before it gets outside the general horse reference range.

Gamma-Glutamyl Transpeptidase (GGT)

Although present in many tissues, the majority of serum GGT is derived from the liver. The enzyme is abundant in renal tubular epithelium, but it is accepted that serum levels are not elevated after kidney damage as the released enzyme is lost in the renal filtrate. Activity usually parallels that of ALP, but it is perhaps less influenced by hepatocyte necrosis and more by biliary epithelium disease than AlP.

It appears that serum GGT is more specific (87%) than serum ALP (51%) in the detection of hepatobiliary disease, but less sensitive (50%) than ALP (80%) (Center, 1992).

The measurement/interpretation of ALP and GGT in series (if GGT is increased, what is the AlP doing?) improves the specificity markedly from 46% to 91%. This is particularly useful in feline medicine where it has been shown that in Lipidosis the AlP is usually disproportionately increased compared with the GGT, which, itself is also 2 to 6-times top-normal compared with the cholangio-hepatopathies where the GGT is increase (about the same range as in Lipidosis) but the AlP is often only marginally increased (particularly when compared with the GGT).

In an aflatoxicosis toxicology trial conducted on a small number of dogs, GGT was shown to be very insensitive, only exceeding upper-normal on 34 of the 74 dog-days (11 of which by only one unit) in animals fed a toxic dose (representing a 46% sensitivity).

In ruminants, by contrast, GGT has been found to be a very reliable indicator of portal triad pathology (Aflatoxicosis, Lantana, Fasciola etc). As colostrum is very rich in GGT, its activity in calf serum can be used as an index of colostrum intake.

In horses GGT is considered to be the single most sensitive and specific test for liver disease although its role in allowing differentiating of parenchymal vs portal triad disease appears to be less convincing. Foals tend to have high GGT although it does not appear to be an index of colostrum intake as it is in other species. There is some evidence that it may be stress-induced and this may explain why it has been found to be a surprisingly good predictor of racing success (high activity = poor placing).

Glutamate Dehydrogenase (GLD)

In all species, GLD is a mitochondrial enzyme, and therefore requires fairly substantial cell damage before it is released.

Increased GLD is reported to be a fairly specific indicator of liver damage. Although abundant in renal tubular epithelium, it is accepted that serum levels are not elevated after kidney damage as the released enzyme is lost in the renal filtrate. Sensitivity of this enzyme for hepatic disease is reported to be high at 92%, and it is believed to reflect necrosis of hepatocytes.

In the same aflatoxicosis toxicology trial (see above), GLD was shown to be more sensitive than GGT, exceeding upper-normal on 45 of the 74 dog-days (most elevations being substantial) in animals fed a toxic dose (representing a 61% sensitivity).

Unfortunately the reagent kit for this test has become increasingly more difficult to access since Roche-Boehringer stopped production.

Aspartate Aminotransferase (Transaminase) (AST)

Unfortunately, erythrocytes are fairly rich in AsT and haemolysed blood (and even serum that has been in contact with the cells for a long time) will give false-high results.

In herbivores, gluconeogenesis follows the Fatty-acid/Krebs-cycle route (eventually Oxaloacetate to Glucose), using AsT as the principal "transfer" enzyme. Consequently, AsT is fairly liver sensitive and specific in these species. As carnivores utilise the Pyruvate-Glucose route there is relatively little AsT in the liver of these species and consequently it is neither sensitive nor specific for liver disease in small animals.

Two AST isoenzymes are found, one cytosolic and one mitochondrial. The half life is reported to be about 12 hours in dogs. This enzyme is released into the blood with increased cell membrane permeability and cellular necrosis (when the mitochondrial isoenzyme is released). Like ALT, the magnitude of increase is reported to be proportional to the number of hepatocytes injured but does not indicate the functional status of the organ. It is reported to be a good indicator of degree of necrosis as well as a good screening test with a sensitivity of 88%. Since AST is found in many organs, including muscle, heart and liver it is not a specific indicator of hepatic damage and if elevations occur in the absence of elevations of ALT, measurement of serum Creatine kinase, a muscle specific enzyme, may be required to rule out muscle damage. It is not routinely used to assess hepatic injury in dogs as it does not appear to have an advantage over ALT as a marker of hepatocyte damage. However, AST was increased in 2/3 of 14 dogs with hepatic abscesses. In the late stages of hepatic lipidosis when the lipid accumulation becomes excessive, AST is reported to elevate. Increased AST is reported to be a sensitive indicator of metastatic (secondary) liver disease in dogs. An examination of the paper does reveal an interesting statistic, namely that most of the neoplasia cases with elevated AsT were lymphoma cases and it is just possible that the enzyme was derived from the lymphocytes rather than the liver. The Faculty of Veterinary Science, Onderstepoort does not, generally, make use of this enzyme in dogs, preferring to rely on ALT and ALP.

In the horse, the fact that myopathy (even from minor incidents and certainly in horses in training) is common, makes this enzyme much less specific and AsT data can only really be interpreted in the light of CK (or clinical indications of muscle damage). In horses, exercise can increase serum activity as much as 30% and especially in early training, resting levels are 50-100% greater than resting levels of horses not in training. It is reported that AsT is correlated with the degree of liver insult in early/acute liver disease and then tails off as the disease becomes chronic.

In herbivores, especially, but also in small animals, muscle trauma (including "down" animals), rhabdomyolysis, white muscle disease (vitamin E-selenium deficiency), and infectious myositis (black leg or Clostridial myositis), and muscular dystrophy may result in marked increases. Serum CK activity will also increased. Note that as AST has a longer half life than CK, increases in AST persist for longer than increases in CK. Therefore, in chronic muscle disease, AST may be elevated, whilst CK levels may be normal. When there is active muscle disease, both CK and AST are elevated (and CK will decline more rapidly as the injury resolves). Equine rhabdomyolysis and bovine Quarter Evil can produce massive serum enzyme elevations.

Anything New?

Until time-and-while a novel serum, liver-specific enzyme is proposed, the use of ALT, ALP and GGT in cats and herbivores (with the possible occasional use of GDH), probably supplies the average veterinary clinician with sufficient information about hepatocyte integrity, cholestasis and steroid induction provided that the limited sensitivity and poor predictive value are borne in mind. An exciting development is that serum ARGINASE, always known to be diagnostically useful but plagued by methodological hurdles, will soon be available in the form of an ELISA kit. Its advantage is extreme specificity for liver and the fact that it indicates necrosis. Furthermore, ARG has a very short half life and can thus be used to monitor response to treatment.