"Liver function tests" is a misnomer for various lab tests that focus on the liver. A simple review with nice algorithms: Am. Fam. Phys. 39(3): 117, March 1989. The asymptomatic patient with abnormal "LFT's": Postgrad. Med. 81(6): 45, May 1, 1987. Essay Br. Med. J. 301: 250, 1990.
If we wanted to measure true "liver function", we'd be doing those obscure tests from the physiology books. Semi-meaningful measures of actual liver function include serum albumin (remember the three-week life span of an albumin molecule), serum bilirubin (see below), and (perhaps best, but seldom used) serum bile salts. Even better is the old BSP dye test, which unfortunately is hazardous (tissue necrosis, anaphylaxis).
*Certain other tests of absolute liver function (indocyanine green clearance, galactose elimination, others) are popular with surgeons who resect massive amounts of liver in heroic attempts to cure hepatocellular carcinoma: Ann. Surg. 211: 438, 1990; Dig. Dis. Sci. 36: 197, 1991; Am. J. Surg. 163: 515. 1992.
*Currently very popular with researchers: metabolism of radioactive caffeine (Clin. Pharm. Ther. 47: 516, 1990; Gut 32: 681, 1991).
The business of liver testing is fraught with pitfalls. Learning objectives for this unit are limited to knowing this handout at the recognition level, and supplying the appropriate tests when you suspect a particular disease.
* Not to memorize just now:
Terminology and assay:
"Direct bilirubin" ("conjugated bilirubin", fast van den Bergh reaction) consists of glucuronide, diglucuronide, and bilirubin bound covalently to albumin ("delta bilirubin").
The remaining bilirubin (i.e., the unconjugated component, "total" minus "direct")
The three mechanisms of jaundice:
(1) Hemolysis (all those causes of hemolytic anemia; also remember intramedullary hemolysis as in megaloblastic anemia and thalassemia, resorption of hematomas, red hepatization of lung, and lung infarcts): Increase only in unconjugated bilirubin
(2) Hepatocellular disease (injured hepatocytes cannot take up unconjugated bilirubin, and the small bile ducts between hepatocytes are disrupted): Increase in both unconjugated and conjugated bilirubin. Remember alcoholism, hepatitis A, B, C, D, autoimmune hepatitis, CMV, Epstein-Barr mononucleosis, drug effects (remember rifampin, isoniazid, methyldopa, and methotrexate; there are many others), ischemia (Ravel's "passive congestion"), ongoing necrosis (Ravel's "active cirrhosis"), and bacterial sepsis (often overlooked).
(3) Extrahepatic biliary obstruction (common duct stone, pancreatic head or common duct cancer): Increased conjugated bilirubin only. Remember that in renal failure, there is likely to be a mild increase in conjugated bilirubin (why?). (Note that occlusion of one of the two hepatic ducts will not produce jaundice, though it will raise alkaline phosphatase.) Certain drugs (notably the anabolic steroids) elevate conjugated bilirubin.
NOTE: Having cited these three elegant mechanisms, I am almost sorry to have to add that the most common reason (by far) for an elevated bilirubin on screening is "Gilbert's non-disease", a poorly-understood (Gastroent. 102: 577, 1992) liver condition affecting 5% of humankind. Usually the problem is a mild allele at the Crigler-Najjar UDP-glucuronosyltransferase locus (Lancet 345: 958, 1995 -- like p53, the enzyme is a tetramer which needs 4 good subunits, which is probably why one bad gene out of two causes more than 50% reduction in activity). Only unconjugated bilirubin is elevated, and values seldom exceed 4 mg/dL. If this fits, and the reticulocyte count is normal, you're probably safe to assume Gilbert's non-disease is the non-problem. If you must prove this, a fairly specific test is that a 72 hour fast (<400 calories/day) causes the bilirubin to double.
NOTE: Two other non-diseases, Dubin-Johnson and Rotor syndrome, feature problems with the transfer of conjugated bilirubin into the bile, resulting in conjugated hyperbilirubinemia. In Dubin-Johnson, the liver is darkly pigmented too. Other liver function test are normal in both non-diseases. The rare Crigler-Najjar syndromes feature partial or complete inability to conjugate bilirubin, with marked indirect hyperbilirubinemia.
NOTE: Metastatic cancer in the liver produces jaundice only late, or with obstruction the common bile duct or both hepatic ducts.
Urine bilirubin and urobilinogen: Too insensitive for screening (Arch. Path. Lab. Med. 113: 73, 1989), and not particularly specific (but good enough to warrant an investigation if abnormal: South. Med. J. 81: 1229, 1988)
"Urine bilirubin" is of course conjugated bilirubin (glucuronide makes the molecule water-soluble). Why the urine of hepatitis patients is brown.
You are familiar with the metabolism of bilirubin into urobilinogen. (It's produced from bile in the gut, and the urobilinogen in the urine has been reabsorbed from the gut.)
If there's no urobilinogen in the urine, chances are it broke down on exposure to light or air, or the urine is dilute. If you suspect the patient has complete biliary obstruction (the other cause), ask about clay-colored stools instead!
Increased urobilinogen in the urine indicates excess bilirubin production (i.e., hemolysis) or failure of the liver to recycle urobilinogen (hepatocellular disease).
Alkaline phosphatase: a marker for poor bile flow
This enzyme, the real purpose of which (if any) is unknown, is induced in liver cells and appears in the bloodstream wherever there is backup of bile in any part of the biliary tree.
You are familiar with other sources of alkaline phosphatase. When liver (but not bone, placenta, or intestinal) alkaline phosphatase is elevated, leucine aminopeptidase, 5'-nucleotidase, and gamma-glutamyl transferase (transpeptidase; GGT) will be generally elevated. (Ask the lab which assays are available. ( GGT, even without other liver test abnormalities, is an empirical, pretty-good marker for alcohol abuse.) If these are normal and "alk-phos" is high on a screen, evaluate the patient for pregnancy or bone disease.
*Some labs are starting to pay attention to alkaline phosphatase isoenzymes, "slow" (AKA "fast", depends on the medium) fractions, etc., etc.
Elevated hepatic alkaline phosphatase suggests that at least some of the biliary tree are obstructed, i.e., that the patient has (1) extrinsic bile duct obstruction (tumor, common duct stone, sclerosing cholangitis) or damage (primary biliary cirrhosis, PBC), (2) hepatocellular damage sufficient to obstruct small bile ducts, or (3) a space-occupying lesion in the liver (usually metastatic tumor, but don't forget hepatocellular carcinoma, cholangiocarcinoma, abscesses, amyloidosis, fungal infection, echinococcus, sarcoid, or TB).
If you do primary care and order chemical profiles, you'll find plenty of people with isolated elevated GGT's. These people are likely to be drinkers but the liver is unlikely to have serious histopathology (Br. Med. J. 302: 388, 1991).
NOTE: Don't confuse an alkaline phosphatase on a chemical profile with a leukocyte alkaline phosphatase. (* The latter, of course, is the histochemical stain for rapid differentiation of leukemoid reaction from chronic granulocytic leukemia).
Transaminases: markers for hepatocyte integrity
Aspartate aminotransferase (AST, SGOT) leaks injured liver, heart, or skeletal muscle cells or erythrocytes (less often kidney).
Alanine aminotransferase (ALT, SGPT) is more specific for liver. (* Warning: It's also less stable in blood samples: Br. Med. J. 301: 557, 1990)
think of massive hepatic necrosis, hepatitis A, B, C, or D, or alcoholic hepatitis
think of infectious mononucleosis (Epstein-Barr, CMV), chronic active hepatitis (viruses B, C, or D; Wilson's (don't miss this one), autoimmune I, II, or III, many alpha1-protease inhibitor deficiency PiZZ), or chronic persistent hepatitis (transaminasitis), drug-related hepatitis, hypoxemia of the liver ("passive congestion"), recovering hepatitis, milder degrees of boozing, cocaine (? at least in mice, maybe people: Arch. Int. Med. 151: 1126, 1991), many cases of hemochromatosis (South. Med. J. 83: 1277, 1990; don't miss this one, either)
metastatic liver disease, cirrhosis without any current cell necrosis (i.e., the currently-sober cirrhotic)
NOTE: A few of these diseases are ones you can actually treat effectively, i.e., Wilson's disease (remove the copper using penicillamine), drug injury (stop the drug), hemochromatosis (remove the iron by phlebotomy), & autoimmune chronic active hepatitis (* many people use azathioprine). Treatment of chronic hepatitis B infection (interferon, etc.) and alpha1-protease inhibitor deficiency (recombinant protein) is less satisfactory but still worthwhile. And of course, "you are powerless over another person's drinking".
NOTE: While AST and ALT generally follow the same upward and downward courses in liver disease, ALT rises less in alcoholism, ostensibly because of the pyridoxine deficiency (low intake, greater demand) which is so common in alcoholics. You will learn rules such as "ALT greater than 300 suggests that underlying problem is not alcoholism", "AST/ALT > 2 means alcoholism", etc.
NOTE: A common situation in clinical medicine is the isolated elevated AST (+ ALT) on screening. If your history and physical exam turn up no clue, one approach is to test the patient in 2-4 weeks off alcohol (the usual culprit) and medications, before proceeding to more expensive tests.
NOTE: Lactate dehydrogenase fraction V comes from liver and skeletal muscle (the latter will often be accompanied by elevated creatine kinase MM). Its rise is as sensitive an indicator of liver disease as is the AST.
*Someday we may order glutathione S-transferase instead of transaminases. Don't hold your breath. Clin. Chim. Acta. 172: 211, 1988.
Autoantibodies (Clin. Lab. Med. 12: 25, 1992)
Anti-mitochondrial antibodies remain a very exciting topic in basic biology.
Among the dozen-or-so known anti-mitochondrial antibodies of human disease, anti-M2 (the main one; Hepatology 10: 247, 1989), anti-M4, and anti-M8 are almost completely sensitive and specific for primary biliary cirrhosis (Proc. Nat. Acad. Sci. 85: 8654, 1988; elegant review Gastroent. 100: 822, 1991).
The related antibodies are responsible for "positive AMA" in other diseases and in relatives of biliary cirrhosis patients.
M2 is the dihydrolipoamide acetyl-transferase component of pyruvate dehydrogenase, and other antibodies in PBC attack the homologous components of alpha-ketoacid dehydrogenase and other enzymes complexes (NEJM 320: 1377, 1989; Proc. Nat. Acad. Sci. 85: 7317, 1988); * they seem to attack the same auto-epitope (E2, where lipoic acid binds).
As you'd expect, biliary epithelium expresses the epitope on its surface (J. Clin. Inv. 91: 2653, 1993).
Newer autoantibodies in PBC: Clin. Exp. Immunol. 74: 68 & 283, 1988; more: Gastroenterology 99: 187, 1990. The old fluorescent assay on HEp-2 cells has given way to ELISA assays.
The antibodies actually inhibit function of the enzymes (J. Immunol. 141: 2321, 1988). Their titers correlate well with disease activity and prognosis (Gastroenterology 99: 1786, 1990). A monoclonal antibody that does the same thing (Scand. J. Immunol. 33: 749, 1991).
Clinical aspects of primary biliary cirrhosis: Med. Clin. N.A. 73: 911, 1989 (no discussion of basic biology). The disease is not rare, and is probably mild or asymptomatic in most cases (Scand. J. Gastroent. 24: 57, 1989).
*Perhaps it is triggered by rough strains of enterobacteriaceae: Lancet 2: 1166, 1988 ("molecular mimicry" again, since the bacteria have a similar epitope; of course, this can't be a whole explanation).
The presence of anti-smooth muscle antibodies OR anti-nuclear antibodies (either must be >1:40) are used to diagnose autoimmune ("lupoid" -- no real relationship to lupus) chronic active hepatitis type I.
So far, the antigen recognized by "anti-SMA" has not yielded up its mysteries like the others. Low titers are seen occasionally in other liver diseases.
*Many of these patients, and others which have similar clinical pictures but lack the defining antibodies, are positive for anti-ASPG-R (asialoglycoprotein receptor). This may define a disease soon.
*Splitters: ANA-negative anti-smooth-muscle antibody disease is now called "chronic active hepatitis type IV". Table:
Anti-liver kidney microsome 1 antibody (anti-LKM1) is the marker for autoimmune chronic active hepatitis type II.
This turned out to be a group of antibodies against P450db1 antigen and its kin (J. Clin. Invest. 83: 1066, 1989; Immunol. Res. 10: 503, 1991), and the new test is apparently completely sensitive and specific.
*More LKM1 stuff: anti-HCV-negative patients have type IIa; anti-HCV-positive patients have type IIb. Hepatitis D patients have type III, which is antibodies against uridine diphosphate glycuronosyl transferase, another drug-metabolizer (Lancet 344: 578, 1994).
Anti-soluble liver antigen antibody is the marker for autoimmune chronic active hepatitis type III (Lancet 2: 292, 1987).
Autoimmune Chronic Active Hepatitis
Type I Anti-smooth muscle antibodies / ANA
Type II Anti-LKM1
Type III Anti-soluble liver antigen
*Anti-hepatocyte plasma membrane antibody is found in chronic active hepatitis from several causes, and may be part of their shared pathogenesis, or an antibody against damaged liver cells. Not a useful test yet. Stay tuned (Clin. Immunol. Immunopath. 52: 291, 1989).
You will learn about prothrombin time (if long in the absence of a blood problem, reasonably specific for bad liver disease) and viruses (hepatitis A-D, CMV, EBV) in other units during the month.
For hepatitis C, order both ELISA2 and RIBA2. Hepatitis C testing right now is still fraught with ambiguity. As you'd expect, some people who meet the virus rid their bodies of it, some never become sick, and some develop chronic liver disease (NEJM 327: 916, 1992). Seroconversion (i.e., antibody production) may take up to two years (NEJM 327: 910, 1992).
BEWARE! When the first tests for hepatitis C antibodies became available, they were used extensively to screen blood donors ("low prior probability" for hepatitis C), and were found to be strong predictors for the blood being infective. Since then, it is becoming clear that many people with a host of liver diseases, but without risk factors for hepatitis C, test positive for antibodies against hepatitis C. This is true particularly on the older C-100 ELISA assay, but sometimes on the newer RIBA recombinant immunoblot assay too. Unlike in the clearly-infected, these antibodies tend to be suppressible with glucocorticoids (Scand. J. Gastroent. 25: 1175, 1990; Mayo. Clin. Proc. 66: 572, 1991). They may represent either reactivation of an old antibody response to a past hepatitis C infection as part of the hypergammaglobulinemia of liver disease (see for example Gut 35: 542, 1994). Or they may even be "false positives". Newer antibody tests: J. Clin. Path. 47: 357, 1994. Review of the whole mess: Arch. Path. Lab. Med. 118: 342, 1994.
Disturbingly, immunosuppressed people sick with hepatitis C virus disease (i.e., recipients of hepatitis C positive organ transplants with subsequent liver disease) typically fail to make the antibody (NEJM 327: 910, 1992).
Work continues on more sensitive and more specific tests for the real presence of hepatitis C virus. In the meantime, the literature has become greatly muddied, and people have been told wrong things that have really frightened them. Don't start your problem drinkers on alpha-interferon just because of a "positive hepatitis C test". See Lancet 3358 258, 1990, Lancet 338: 277, 1991, Am. J. Gastroent. 86: 1219, 1991; Gastroenterology 99: 1531, 1990; Dig. Dis. Sci. 37: 156, 1992; third-generation test (ELISA C-200/C22) Proc. Nat. Acad. Sci. 89: 4486, 1992. The polymerase chain reaction for viral RNA is now available and seems to be very sensitive and specific for real liver disease (i.e., on biopsy, i.e., there are few "healthy carriers", unlike hepatitis B) in patients with the antibody (no surprise; Lancet 340: 697, 1992). Around half of new mothers with the antibody have the virus, while the others seem to have rid themselves of it: NEJM 330: 744, 1994. Stay tuned.
The confusion is compounded by the likelihood that at least some acute viral infections can trigger the chronic autoimmune hepatitis syndrome. Hepatitis A is implicated (Lancet 337: 1183, 1991); one group wants to exonerate hepatitis C (Arch. Int. Med. 151: 1548, 1991), another implicates it (Gastroenterology 106: 1672, 1994) etc. We can hope this gets unscrambled in the next few years.
Transferrin saturation of 60% or greater suggests a problem with iron overload, and 80% or greater is very suggestive of hemochromatosis. Follow up your suspicions with a serum ferritin, which is generally >800 ng/mL (usually much greater). Please don't miss incipient iron overload -- it happens too often.
The new rule seems to be that transferrin saturation >62% and/or serum ferritin >400 g/L is a positive screen (Gastroenterology 107: 453, 1994).
The new awareness of hemochromatosis has prompted many requests for deferoxamine tests, quantitative iron studies on liver tissue, etc., etc. A biopsy can help you tell primary hemochromatosis (lots of iron throughout the lobule and usually in bile ductules too) from cirrhosis with secondary iron overload (extremely common, iron is less and is unevenly distributed) from hemosiderosis due to hemolysis (iron is mostly in Kupffer cells). Right now, the literature considers tissue iron assay on liver biopsy to be more sensitive and specific (* the "iron index", which distriminates between alcoholism and hemochromatosis, is micromoles of iron per gram of tissue, with 2 or more being an indication of primary hemochromatosis; some new formulas divide iron content by age: Gastroent. 100: 1397, 1991; Gut 32: 207, 1991). A clinical history helps, too. There's plenty being written now about hemochromatosis screening (West. J. Med. 15x: 323, 1991 is a good one); a get-real approach is outlined in NEJM 328: 1616, 1993 (transferrin saturation >60% in a man or 50% in a woman is suspicious, check serum ferritin; if high, bleed and maybe biopsy liver; if normal, follow yearly.)
Ceruloplasmin is lower than 20 mg/dL in most cases of Wilson's disease. Also check urine copper (more than 100 micrograms in 24 hours). Remember that Wilson's is likely to present as a "psychiatric problem": Postgrad. Med. 95: 135, 1994, and it is notoriously impossible fully to rule it out using any single lab: Gastro. 113: 350, 1997.
Serum bile acids are a very sensitive indicator of hepatocellular disease.
Alpha-fetoprotein: Tumor marker for hepatocellular carcinoma; levels over 1000 pg/mL are a reasonably specific indicator.
*We can now distiguish the isoform for hepatocellular carcinoma from the kind seen at low concentrations in the serum of patients with benign liver diseases: Cancer 74: 25, 1994.
*Des-gamma-carboxy prothrombin (a PIVKA) is an importnat marker for hepatocelular caricnoma (why?). Mechanism: Cancer 74: 1533, 1994.
Alpha-1 protease inhibitor ("alpha1-antitrypsin"): to detect patients in whom deficiency of this factor causes hepatitis and cirrhosis. WARNING: Patients who are deficient, but not totally lacking, can have normal levels during the acute phase reaction.
*Apolipoprotein A1 is advocated by a French group as a way of telling drinkers with cirrhosis from those without. The marker is reported to be high in non-fibrotic liver disease, low when cirrhosis supervenes (Gastroent. 100: 1397, 1991). Other tests to see which drinker has cirrhosis may eventually include serum laminin, serum procollagen, and serum type VI collagen (Hepatology 15: 637, 1992).
* Future directions: In the next few years, watch for (1) better algorithms for the diagnosis of hemochromatosis, and a greater awareness of this entity by physicians and patients; (2) better tests for hepatitis C infection, and a clarification of the role of all the hepatitis viruses in "idiopathic" liver diseases; (3) identification of the target molecule and mechanism of injury in autoimmune hepatitis type I; (4) identification of the genes, molecular mechanisms and real pathogenesis in Wilson's, Gilbert's, and familial hemochromatosis. For your interest: Hepatitis C antigen and antibody in Science 244: 359 & 362, 1989. How it was found: Ann. Int. Med. 115: 644, 1991. Chronic hepatitis A: Liver 9: 223, 1989.