Jaundice is yellowing of the skin and whites of the eyes, caused by an accumulation of the yellow-brown pigment bilirubin in blood and tissues. Jaundice is the chief sign of many disorders of the liver and biliary system. Many otherwise healthy babies are affected briefly by jaundice soon after birth.
Types and causes
Bilirubin is formed from haemoglobin (the oxygen-carrying pigment in red blood cells) when old red cells are broken down, mainly by the spleen. It is absorbed by the liver, where it is made soluble in water and excreted in bile. There are three main types of jaundice: haemolytic, hepatocellular, and obstructive.
In haemolytic jaundice, too much bilirubin is produced for the liver to process. This condition results from excessive haemolysis (breakdown of red blood cells), which can have many causes.
In hepatocellular jaundice, bilirubin accumulates because it is prevented from passing from liver cells into the bile. This form of jaundice is usually due to acute hepatitis (inflammation of the liver) caused by taking certain drugs or by liver failure.
In obstructive jaundice, also called cholestatic jaundice, bile cannot leave the liver because of bile duct obstruction, which may be caused by gallstones or due to a tumour anywhere in the duct. Obstructive jaundice can also occur if the bile ducts are underdeveloped (as in biliary atresia) or have been destroyed by disease. Cholestasis (stagnation of bile in the liver) then occurs and bilirubin overflows into the blood.
Diagnosis and treatment
Blood tests, and possibly also a liver biopsy (removal of a sample of tissue for analysis), may be performed to identify the cause of the jaundice. Investigation of the bile duct may be carried out using such imaging techniques as ERCP and MRI. Treatment is for the underlying cause.
Yellowing of the skin and whites of the eyes in newborn babies, due to accumulation of the yellow-brown bile pigment bilirubin in the blood. Neonatal jaundice usually results from the liver being too immature to excrete bilirubin efficiently. The condition, which tends to be more common in breast-fed babies, is usually harmless and disappears within a week.
In rare cases, severe or persistent neonatal jaundice can be caused by the blood disorder haemolytic disease of the newborn; the genetic condition G6PD deficiency; hepatitis (inflammation of the liver); hypothyroidism (underactivity of the thyroid gland); biliary atresia (abnormal formation or absence of the bile ducts); or infection. Jaundiced babies usually require extra fluids and may be treated with photo-therapy (light therapy) or, in severe cases, exchange transfusion. If severe neonatal jaundice is not treated promptly, kernicterus (a form of brain damage) may occur.
Jaundice in detail - technical
Physiology of bilirubin
All haem molecules are degraded in macrophages by haem oxygenase to biliverdin, and then by biliverdin reductase to bilirubin, which is selectively removed by hepatocytes from sinusoidal blood and then conjugated, mainly by one of the two specific isoforms of the microsomal enzyme UDP-glucuronyl (glucuronate-glucuronosyl) transferase, chiefly with two glucuronic acid moieties. Conjugated bilirubin is excreted into the bile by the anionic conjugate transporter protein (MRP2), but in many liver diseases it readily refluxes back into blood and—since it is water soluble and less firmly bound to albumin than unconjugated bilirubin—about 1% is filtered across the glomerular membrane and darkens the urine (choluria). In the distal intestine conjugated bilirubin is deconjugated and reduced to a series of uro- and stercobilinogens that give the normal colour to faeces. Some colourless urobilinogen is normally absorbed from the colon and undergoes an enterohepatic circulation, with a small amount being excreted in urine. If this biliary excretion is impaired in liver disease, or increased in haemolysis, then excess urobilinogen is excreted in urine, where it is easily detected by routine clinical ‘stix’.
Jaundice is the clinical sign of hyperbilirubinaemia and usually indicates disease of the liver or biliary tree. Dark urine and, less commonly, pale stools indicate cholestasis. Stigmata of chronic liver disease are important, but do not define the cause of jaundice.
Unconjugated hyperbilirubinaemia—should be sought for by testing if serum bilirubin levels are raised, while other liver-related blood tests are normal. Causes include (1) haemolysis, which if severe enough to raise bilirubin levels is likely to cause an elevated reticulocyte count and reduction in plasma haptoglobin; (2) benign constitutional unconjugated hyperbilirubinaemia (Gilbert’s syndrome)—a common recessive condition (affecting at least 3% of the normal adult population) due to homozygous polymorphisms in the promoter region of the specific glucuronyl transferase gene; recognized by a fluctuating, elevated serum bilirubin concentration that rises excessively on fasting; patients require reassurance that the results do not indicate liver disease.
Conjugated hyperbilirubinaemia—routine liver-related blood tests cannot differentiate between intra- or extrahepatic causes of jaundice, unless the transferases are very high, in which case hepatitis (e.g. viral, alcoholic) is certain. Cholestasis should be sought by abdominal ultrasonography to detect a dilated intra- and/or extrahepatic biliary tree (and often also reveal its cause, e.g. gallstones, tumour). Further investigation depends on clinical context: (1) likely biliary disease—endoscopic retrograde cholangiopancreatography (ECRP), magnetic resonance cholangiography (MRC); (2) likely intrahepatic cholestasis—hepatitis A, B, and C serology, autoantibodies, serum caeruloplasmin/copper, plasma α1-antitrypsin concentration; liver biopsy.
Jaundice is the clinical sign of hyperbilirubinaemia, and hence usually indicates disease of the liver or biliary tree. The pigment in the tissues in best seen as yellowing of the sclera; eventually the skin and soft palate become tinted, but not saliva nor sputum. The urine usually becomes dark. Rarely, carotenaemia, from eating carrots or vitamin A in excess, can mimic jaundice, but then the colour is more prominent in the palms than the sclera.
Physiology of bilirubin
All haem molecules in haemoglobin or cytochrome enzymes are stoichiometrically (1:1) degraded in macrophages via biliverdin to bilirubin, especially in the spleen and liver, but also macrophages in other tissues, including skin, and renal tubular cells. Haem oxygenase breaks open the asymmetric tetrapyrrole haem molecule specifically at the α-methene bridge, releasing carbon monoxide and iron, and forming biliverdin. One principal isomer of biliverdin, namely IXα, is formed, although small amounts of the other three possible isomers (β, γ, and δ) can be detected in bile. The excretion of carbon monoxide in breath can be used quantitatively to determine the breakdown of haem to bilirubin, of which 200 to 350 mg (340–600 µmol) is produced daily. About 85% of biliverdin, and therefore bilirubin, is derived from the delayed breakdown of the haemoglobin in ageing red blood cells, while the remainder is either from the breakdown of haem proteins, chiefly in the liver, or from ineffective erythropoiesis in the bone marrow; these constitute the so-called ‘early labelled’ bilirubin, defined by isotopic studies in vivo.
Biliverdin is green and is directly excreted in bile by birds, amphibians, and reptiles, but not by mammals in whom biliverdin is reduced by the macrophage cytosolic enzyme biliverdin reductase chiefly to the yellow bilirubin IXα, which has then to be excreted. The reason for this species difference was obscure, for bilirubin is lipid soluble and potentially toxic, and has to be conjugated before it is excreted in bile, whereas biliverdin is water soluble and can be readily excreted in urine and bile by mammals. However, bilirubin is an antioxidant or free radical scavenger in plasma and bile, particularly when bound to copper, and this may be particularly important in the neonate, especially when levels of the antioxidant ascorbate are low. Hence bilirubin probably has a function and is not just a waste product. Bilirubin is surprisingly lipid soluble; this is due to internal hydrogen bonding in the molecule so that it forms a tight, nonpolar, nonlinear, three-dimensional structure. After its release from macrophages, it is firmly bound to plasma albumin, so that none enters the urine. At high concentrations in the blood it slowly diffuses into tissues, where it can be toxic, particularly in the neonatal brain (kernicterus), or the kidney. Jaundice is less obvious in unconjugated, than in conjugated, hyperbilirubinaemia since its diffusion into the tissues is more limited. Bilirubin is readily oxidized back to biliverdin; hence the green vomit of intestinal obstruction.
The circulating pool of bilirubin in the plasma (c.100 µmol) is almost all unconjugated. Routine measurements still rely on the Van den Bergh diazo reaction, which yields either an indirect (unconjugated bilirubin) or direct (conjugated) reaction and, although this overestimates the true level of conjugated bilirubin, the results indicate whether or not circulating bilirubin is chiefly unconjugated. The direct and indirect reactions depend on the slow reaction of the unconjugated bilirubin with the reagent; this is accelerated when solvents, such as methanol, which break the internal hydrogen bonding of bilirubin, are added. The normal range of plasma bilirubin is wide (c.5–19 µmol/litre), reflecting wide variation in the rate of conjugation in the liver, and is higher than in most other mammals in which hepatic clearance and excretion are more efficient. The distribution of values is Gaussian, so that the true upper limit of normal is arbitrary (see ‘Familial unconjugated hyperbilirubinaemia’, below). Hepatic enzyme-inducing drugs reduce the plasma level by increasing hepatic conjugation and hence the plasma clearance of bilirubin.
Bilirubin is selectively removed by hepatocytes from sinusoidal blood, although its plasma clearance (c.50 ml/min) is low compared e.g. with that of bile acids, and so its extraction (1.5% of plasma pool/min) is dependent more upon hepatocyte distribution and function than on hepatic blood flow. It is initially surprising that bilirubin can be displaced from its plasma binding sites and enter hepatocytes. Specific hepatic cytoplasmic binding proteins have been described, but binding to the active site of the microsomal conjugating enzyme would be sufficient to maintain a low level of free bilirubin in the cytoplasm, which, without the need for specific transfer proteins, should alone produce a gradient sufficient to allow bilirubin slowly to enter the hepatocyte. This uptake of bilirubin is facilitated by the direct contact of plasma with the hepatocyte in the interstitial space of Disse through fenestrations in the endothelium of hepatic blood capillaries. Although uptake predominates, dynamic studies show that there is also considerable reflux of bilirubin out of the cell back into the plasma.
Within the hepatocyte bilirubin is principally conjugated by one of the two specific isoforms of the microsomal enzyme UDP-glucuronyl (glucuronate-glucuronosyl) transferase, chiefly with two glucuronic acid moieties. Minor quantities of bilirubin are conjugated with one glucuronic acid molecule (monoglucuronide) or with combinations of related sugars (xylose, glucose); a small amount of unconjugated bilirubin also appears in bile. The chemical properties of the conjugated molecules are quite different from those of unconjugated bilirubin, for there is no internal hydrogen bonding of bilirubin—they now become more linear, fully water-soluble molecules and are efficiently excreted in bile. In many liver diseases conjugated bilirubin readily refluxes back into blood and, since it is water soluble and less firmly bound to albumin than unconjugated bilirubin, about 1% is filtered across the glomerular membrane and darkens the urine (choluria). Hepatocytes have at least six specific active transporters for the canalicular excretion of the major components of bile, although not for cholesterol, and isolated autosomal recessive defects in them have now been identified. Conjugated bilirubin is excreted out of the endoplasmic reticulum and then across the microvillous intercellular canalicular membrane by the anionic conjugate transporter protein (MRP2). There is a specific canalicular bile acid export pump protein (BSEP) and one for phospholipid (MDR3). MRP2 also transports other multivalent anions, such as conjugated bromsulphthalein.
The urinary excretion of conjugated bilirubin is increased by the bile acids that also accumulate in liver disease. If renal function is normal, this renal excretion of bilirubin eventually matches its normal rate of production when conjugated bilirubin levels in the plasma reach about 600 µmol/litre. With renal failure, or haemolysis, plasma levels rise higher. Little bilirubin, even if conjugated, diffuses through renal dialysis membranes.
Recently it has been shown that deconjugated bilirubin can undergo a substantial enterohepatic circulation; it is absorbed from the colon, particularly when there is bile acid malabsorption and hence the concentration of bile acids in the colon is increased, for example as a result of disease or resection of the ileum. This reabsorption then increases the concentration of bilirubin re-excreted in bile, and may in part explain the increased incidence of pigment gallstones in patients with ileal disease. Oral ursodeoxycholic acid also increases the enterohepatic recycling of bilirubin perhaps by solubilizing bilirubin in the intestinal lumen, or by impairing the reabsorption of other bile acids in the ileum. This may explain the rim of calcification in an outer pigment layer of cholesterol gallstones during their treatment with ursodeoxycholic acid, and thus the frequent resistance to such dissolution therapy. Similarly, fasting increases unconjugated bilirubin levels in the plasma by increasing the reabsorption of bilirubin, because it reduces intestinal motility and improves absorption.
In the distal intestine conjugated bilirubin is deconjugated and reduced to a series of uro- and stercobilinogens that give the normal colour to faeces. Some colourless urobilinogen is normally absorbed from the colon and undergoes an enterohepatic circulation, with a small amount being excreted in urine. If this biliary excretion is impaired in liver disease, or increased in haemolysis, then excess urobilinogen is excreted in urine, where it can oxidize on standing to dark brown urobilins. Urobilinogen is easily detected by routine clinical ‘stix’. Ehrlich’s aldehyde reagent was at one time used; urine containing excess urobilinogen turns red with this reagent and the urobilinogen pigment can then be extracted into an organic solvent, such as chloroform. This is unlike the similar pigment formed from the more polar porphobilinogen adduct in acute porphyria, which remains in the upper aqueous phase.
Management of jaundice
Complex algorithms for the management of the patient with hyperbilirubinaemia or jaundice have been published, but a simple pragmatic approach is proposed here.
Raised plasma bilirubin levels, and eventually frank jaundice, are due to excessive unconjugated or conjugated bilirubin levels in blood, depending respectively on whether the abnormality in bilirubin metabolism is either in its production and/or conjugation, or in the subsequent hepatic excretion of conjugated bilirubin. Impaired excretion is almost always combined with impaired bile flow and is best termed cholestasis, when other liver-related blood tests are abnormal, especially the biliary enzymes alkaline phosphatase and γ-glutamyl transpeptidase, serum bile acids are also raised, there is often itching, and the microvilli lining the biliary canaliculi are injured. Examination of a liver biopsy specimen taken from a patient with cholestasis may show bile plugs under light microscopy. These findings, however, are often termed ‘obstructive jaundice’—an unfortunate term, since it implies extrahepatic obstruction of the biliary tree. Prolonged cholestasis, as from bile duct obstruction, down regulates the MRP2 exporter transporter that is also found in enterocytes, so that the oral bioavailability of many drugs is increased. The molecular events underlying some forms of intrahepatic cholestasis are now being unravelled (see below).
Dark urine and, less commonly, pale stools indicate cholestasis. Many drugs, including alcohol, can cause unconjugated and conjugated hyperbilirubinaemia and should be rigorously sought. Fever (hepatitis, cholangitis, abscesses), travel (hepatitis, amoebiasis), sexual history (hepatitis A, B, or C), surgery and anaesthesia (postoperative jaundice, see below; biliary tract disease), herbal medicines (e.g. West Indian teas, Chinese herbs), and transfusions or blood products (hepatitis B or C) can be important clues.
Stigmata of chronic liver disease (e.g. spider naevi, facial telangiectases, parotid enlargement, Dupuytren’s contractures, muscle wasting, hepatosplenomegaly, dilated abdominal wall veins and ascites) are important, but do not define the cause of jaundice.
Testing for unconjugated hyperbilirubinaemia
If serum bilirubin levels are raised but other liver-related blood tests are normal, unconjugated hyperbilirubinaemia should be excluded by testing whether the bilirubin in blood is predominantly conjugated or unconjugated. An abnormal reticulocyte count will suggest haemolysis severe enough to raise bilirubin levels and blood film examination may be informative. Suspected haemolysis is investigated as described elsewhere in chapter.
If no cause of unconjugated hyperbilirubinaemia is identified, then benign constitutional unconjugated hyperbilirubinaemia (Gilbert’s syndrome) is diagnosed (see below).
The familial syndromes without cholestasis (Dubin–Johnson and Rotor) are rare (see below).
Routine liver-related blood tests cannot differentiate between intra- or extrahepatic causes of jaundice, unless the transferases are very high (e.g. >1000 IU/litre), in which case hepatitis (e.g. viral, alcoholic) is certain. A greatly raised alkaline phosphatase level does not necessarily imply an extrahepatic lesion; intrahepatic causes are common (Table 1). Research methods for assessing liver function (e.g. galactose tolerance test, aminopyrine breath test) are of no value in the management of the patient with jaundice.
Cholestasis should be investigated first with abdominal ultrasonography, which will accurately detect a dilated intra- and/or extrahepatic biliary tree and often also reveal its cause (e.g. gallstones, tumour). Oral cholecystography, or the now little used intravenous cholangiography, will fail in the presence of jaundice. If biliary disease is thus suspected, an endoscopic retrograde cholangiogram (ERCP) or, failing that, a fine-needle percutaneous transhepatic cholangiogram (PTC), will define the anatomy more accurately and often provide definitive therapy (removal of biliary stones, stenting), thus avoiding surgery. Magnetic resonance cholangiography (MRC) is increasing in sensitivity and availability and can now produce high-quality noninvasive images of the biliary tree and pancreas; it is useful even for intrahepatic biliary disease. It cannot, of course, be therapeutic. Endoscopic ultrasonography (EUS) can show accurately the presence of stones, biliary or pancreatic tumours and sclerosing cholangitis, while gamma-camera scans with technetium-labelled hydroxyiminodiacetic acid (HIDA) can be used to indicate biliary obstruction, particularly in the neonate, if ultrasonography is normal.
|Table 1 Intrahepatic cholestasis|
|Infection||Viral hepatitis A, B, C, or E|
|Bacterial: sepsis, miliary tuberculosis, leptospirosis|
|Neonatal hepatitis syndrome|
|Drugs: hepatotoxic, idiosyncratic|
|Oestrogens: pregnancy, contraceptive pill|
|Adenocarcinoma of kidney (non-metastatic)|
|Neonatal cholestatic syndromes|
|Reduced blood flow||Perioperative hypoperfusion/shock|
|Intrahepatic ducts||Biliary atresia|
|Malignant infiltration of ducts|
|Autoimmune||Chronic active hepatitis|
|Primary biliary cirrhosis|
|Rejection of liver graft|
If intrahepatic cholestasis is suspected because of a normal sized biliary tree on ultrasonography, the following tests should be considered: hepatitis A, B, and C serology, autoantibodies (antimitochondrial for primary biliary cirrhosis, antinuclear, smooth muscle and liver–kidney microsomal for autoimmune chronic hepatitis) and immunoglobulins, serum caeruloplasmin and copper for Wilson’s disease if less than 40 years of age, or plasma α1-antitrypsin concentrations for homozygous deficiency of this enzyme. Intrahepatic masses seen by ultrasonography will prompt measurement of α-fetoprotein for primary hepatoma, and other tumour markers. A percutaneous needle liver biopsy (or aspiration of an abscess) may then be indicated, provided that blood coagulation and the platelet count are normal. Guidelines for liver biopsy have been published. A transjugular venous approach for the biopsy is appropriate if the risks of bleeding are increased.
Plasma bilirubin levels are exponentially and positively related to the half-life of circulating red blood cells, which determines bilirubin load, and negatively to the hepatic clearance rate of bilirubin. This relationship is analogous to that of muscle breakdown, plasma creatinine, and glomerular filtration rate. Hence, if the rate of haemolysis rises or clearance falls, bilirubin levels may rise rapidly in response to small changes of the load or removal rate from plasma, or both.
Haemolytic jaundice is most commonly encountered in the haemoglobinopathies of sickle-cell anaemia (homozygous SS or heterozygous SC disease) or homozygous thalassaemia major, although dark skin may render it difficult to detect. The ‘acholuric jaundice’ of hereditary spherocytosis is rare. Mildly elevated bilirubin levels are described in ineffective erythropoiesis of the bone marrow in vitamin B12 deficiency (pernicious anaemia), or in thalassaemia minor.
Drugs may cause haemolysis (e.g. methyldopa, sulphasalazine), or impair hepatic bilirubin clearance (e.g. rifampicin). Infections (e.g. malaria) or mismatched blood transfusions can produce massive haemolysis, but this overshadows the raised bilirubin levels. Autoimmune haemolytic anaemia, such as in lupus erythematosus, or haemolysis due to glucose-6-phosphate dehydrogenase deficiency, or to leaking prosthetic cardiac valves, can cause clinical jaundice.
Familial unconjugated hyperbilirubinaemia
A series of defects of the hepatic bilirubin conjugating enzyme UDP-glucuronyl transferase produce various degrees of unconjugated hyperbilirubinaemia due to impaired bilirubin clearance; they have long fascinated physiologists and more recently molecular biologists.
At least 3% of the normal adult population have mildly raised unconjugated bilirubin levels in blood that rise excessively on fasting. This ‘phenomenon’ is commonly termed ‘Gilbert’s syndrome’ (OMIM 143500), although it is unclear whether the eponym is justified. The raised concentrations of bilirubin develop in early adult life and are often associated with mild degrees of haemolysis. Any combination of an increased bilirubin load from the haemolysis and a mildly impaired clearance will increase plasma bilirubin concentrations more than would either alone, and hence together they bring the underlying condition to notice. Various associated defects of hepatic drug metabolism have also been described and these are probably linked genetic abnormalities. The syndrome is not a discrete entity, but rather different defects of conjugation and haemolysis that elevate bilirubin levels above an arbitrary upper limit of normal. Determination of the bilirubin-conjugating capacity of liver biopsy tissue has shown that the activity of glucuronyl transferase is reduced by 60 to 70%, and this impairs bilirubin clearance.
Gilbert’s syndrome is recognized by a fluctuating, raised serum bilirubin concentration with the other routine liver-related blood tests remaining normal, and a normal reticulocyte count to exclude overt haemolysis. It can be confirmed by measuring the unconjugated fraction of the bilirubin, which should be greater than 90%. Measuring the pronounced increase of plasma bilirubin that occurs after a 48-h fast on 400 kcal/day or provocation with intravenous nicotinic acid are research procedures. A liver biopsy is not needed. Reassurance that the results do not indicate liver disease and will not affect life insurance is important. Plasma bilirubin concentrations rise in patients with Gilbert’s syndrome during intercurrent illness and jaundice may then be observed.
It is said that Gilbert’s syndrome can follow an attack of viral hepatitis, although this may be due to ascertainment bias.
The genetic basis of Gilbert’s syndrome remains controversial but it appears that it is a recessive condition in which there are homozygous polymorphisms in the promoter region affecting expression of the specific glucuronyl transferase gene. Heterozygotes have normal bilirubin levels. There must be another factor responsible for the increased bilirubin concentrations as the heterozygote abnormality occurs in 40% of normal individuals. The variable bilirubin load from red cell breakdown is one factor that will influence the underlying prevalence of the anomaly in the population.
Two syndromes of more severe unconjugated hyperbilirubinaemia have been described, namely the rare type I Crigler–Najjar (OMIM 218800; 100 cases reported), which without treatment causes neonatal death, and the more common, and benign type II (OMIM 606785). Both are due to severe deficiency in the UDP-glucuronyl transferase enzymes.
In type I, first reported in 1952, with a recessive inheritance, neonates rapidly become progressively jaundiced in the first days of life (bilirubin levels reach 350–950 µmol/litre) and, if untreated, develop kernicterus or brain damage. Death usually occurs within a year but delayed kernicterus has been reported. There is no conjugated bilirubin in bile, but small quantities of unconjugated bilirubin can be found in bile and also cross the intestinal wall.
The inheritance of type II is complex, and is reported both to be dominant with incomplete penetrance or recessive. Bilirubin levels are lower (<350 µmol/litre), and persistent mild jaundice is only noticed in childhood. Brain damage does not occur, and the only problem is cosmetic. One-third of the conjugated bilirubin in bile is present as the monogluronide (normally <10%). In the Gunn strain of laboratory rat severe unconjugated hyperbilirubinaemia occurs and glucuronyl transferase activity is absent in the liver, as it is in Crigler–Najjar type I. In type II, enzyme activity is less than 10% of normal, but measurable.
It has long been known that there is a spectrum of bilirubin levels in type II Crigler–Najjar and Gilbert’s syndromes and indeed both conditions have been observed within the same families, suggesting different degrees of enzyme activity. Phenobarbitone or other hepatic microsomal enzyme-inducing agents markedly reduce bilirubin levels in Gilbert’s and Crigler–Najjar type II syndromes, although unfortunately not in Crigler–Najjar type I, and increase the activity of glucuronyl transferase. Such treatment, however, is not needed.
Molecular analysis of the genes encoding human UDP-glucuronyl transferases has both clarified and complicated our understanding of the genetic basis of these disorders. The complementary DNAs for the two human isoforms of the enzyme have been sequenced; they differ from those that encode the other glucuronyl transferases, which conjugate e.g. steroids. The UDP-glucuronyl transferases map to human chromosome 2, where at least five exons encode the specific mRNAs of the isoenzymes. Analysis of DNA from patients with type I Crigler–Najjar syndrome has identified homozygous or heterozygous defects in the exons encoding particularly the most active of the two bilirubin transferase isoforms. Similar defects occur in the Gunn rat.
In Crigler–Najjar type II syndrome, mutations have been described in the gene encoding the more active bilirubin glucuronyl transferase isoform. Phenobarbitone induces the expression of the abnormal enzyme, explaining its efficacy in this condition. Probably a heterozygous combination of an abnormality of the promoter region (Gilbert’s defect) and a Crigler–Najjar I defect is responsible for the phenotype of type II Crigler–Najjar syndrome. This explains the presence of patients with Gilbert’s syndrome within families with type II Crigler–Najjar.
It seems likely that a series of rare abnormalities in the gene for bilirubin glucuronyl transferases will be found in each of the three arbitrary phenotypes, which are clinically defined by the degree of impairment of conjugation and hence plasma bilirubin levels.
Crigler–Najjar type I syndrome can now be successfully treated by whole-body blue-light phototherapy for 16 h daily or by plasmapheresis until liver transplantation can be carried out as a definitive treatment. Severe kernicterus is a contraindication to transplantation as it is not reversible. Some patients have received successful transplants. Hepatocyte transplantation, in which donor hepatocytes are infused into the portal vein, has been partially successful. Drugs that displace unconjugated bilirubin from albumin (sulphonamides, salicylates, penicillin) increase brain damage in type I Crigler–Najjar syndrome and must be avoided.
Unconjugated hyperbilirubinaemia, often with mild clinical jaundice, occurs in all full-term newborn infants, and is harmless and probably beneficial. Bilirubin concentrations are maximal at 2 to 5 days after birth, but the plasma bilirubin rarely exceeds 90 µmol/litre; neonatal jaundice is more severe in premature infants. It is attributed to a combination of immaturity of hepatic glucuronyl transferase and the added load of bilirubin from rapid haemolysis of surplus fetal red blood cells in the neonatal period. Before birth, fetal bilirubin is excreted by the mother, and meconium and stools are pale because of the limited excretion of bilirubin by the fetus.
If haemolysis is increased, as in rhesus or other fetomaternal incompatibility of red cell antigens when transplacental maternal antibodies cause intravascular haemolysis of fetal red blood cells, severe jaundice and kernicterus can occur. Acidosis and some drugs (sulphonamides, salicylates, penicillin) may increase kernicterus by displacing unconjugated bilirubin from albumin. Glucose 6-phosphate deficiency can also cause jaundice and anaemia in the neonatal period, usually in infants of Mediterranean, African, or Chinese ancestry.
Treatment with phenobarbitone induces hepatic glucuronyl transferase and lowers bilirubin levels, but its effect is slow unless it is given to the mother before birth. Exchange transfusion or plasmapheresis are more effective. Phototherapy, namely exposure of the near-naked infant to blue light in an incubator, is also effective. Being yellow, bilirubin absorbs light at approximately 450 nm, which oxidizes it to water-soluble, nontoxic products. Hence, exposure of the bilirubin in skin capillaries to light reduces its plasma concentration and the breakdown products are excreted safely in urine and bile. Reabsorption of bilirubin from the intestine can also be reduced by giving agar by mouth, thus interrupting its enterohepatic circulation. Nevertheless, it seems that moderate hyperbilirubinaemia does not adversely affect development.
Breastfeeding slightly increases serum bilirubin levels and about 1 in 40 breast-fed infants develop jaundice, which remits on transfer to cow’s milk within 24 h; this jaundice does not always recur when breast milk is reintroduced. Breastfeeding increases the enterohepatic cycling of bilirubin from the intestine, since stool weights and frequency are less than when taking formula feeds, and hence intestinal motility is decreased. Steroid molecules in breast milk may also inhibit glucuronyl transferase activity in the neonatal liver.
Hypothyroidism increases jaundice and should be sought in neonates with unexplained hyperbilirubinaemia since it may not be associated with obvious cretinism. The rare Crigler–Najjar type I syndrome (see above) presents with florid jaundice in the first few days of life.
Sickle-cell anaemia and β-thalassaemia
Jaundice is common in homozygous sickle-cell anaemia due to the unconjugated hyperbilirubinaemia from persistent haemolysis. During crises jaundice often deepens in association with increasing anaemia, suggesting accelerated haemolysis, although transient bone marrow failure may also occur. Occasionally, conjugated hyperbilirubinaemia with dark urine occurs during these episodes, and hepatic histology may show areas of necrosis due to thrombosis and bile thrombi. Patients with sickle-cell anaemia are also prone to pigment gallstones, due to the excessive bilirubin constantly being excreted, and these can cause extrahepatic biliary obstruction; conjugated hyperbilirubinaemia and dark urine may then be clues. Unconjugated hyperbilirubinaemia occurs in homozygous thalassaemia as a result of increased red cell destruction and the intramedullary haemolysis associated with ineffective erythropoiesis; there may also be unexplained episodes of intrahepatic cholestasis.
There are many causes of intrahepatic cholestasis (Table 1 above).
Conjugated hyperbilirubinaemia and cholestasis in the neonate, with dark urine and pale stools, is always pathological and if it continues beyond 2 weeks of age requires urgent investigation. There are many causes.
In many instances the cause is never established and then, although it was once called neonatal hepatitis, it is better termed the hepatitis syndrome; hepatic histology shows hepatitis, sometimes with giant cells. Some babies recover, while perhaps half progress to hypoplasia of the intrahepatic bile ducts, which then overlaps with intrahepatic biliary atresia.
Infections, particularly urinary, can cause transient cholestasis. Syphilis is now rare, as is toxoplasmosis. Various viral infections (rubella, cytomegalovirus) can cause neonatal jaundice. The hepatotropic hepatitis B virus contracted from an HBe antigen-positive mother rarely causes jaundice. Metabolic diseases that may cause neonatal jaundice include galactosaemia, hereditary fructose intolerance (fructosaemia), and tyrosinosis—all of which need to be diagnosed quickly so as to start dietary treatment early—as well as homozygous α1-antitrypsin deficiency, and intravenous feeding per se. Other genetic diseases include trisomy 13 and trisomy 18 (one-quarter of babies developing the hepatitis syndrome) and cystic fibrosis.
Several familial syndromes presenting with neonatal cholestasis have been described, some with other congenital abnormalities, such as arteriohepatic dysplasia (Alagille’s syndrome), and others solely with cholestasis featuring persistent jaundice, raised serum bile acids, hepatosplenomegaly, steatorrhoea and failure to thrive, such as Byler’s syndrome in Amish families, which is genetically related to benign recurrent cholestasis. Bile duct hypoplasia, cirrhosis, and liver failure often follow unless liver transplantation is carried out. There are several different mutations of the genes described in the various syndromes of progressive familial intrahepatic cholestasis (PFIC), namely those encoding the canalicular transport proteins for bile acids (BSEP), phospholipid (MDR3), or bilirubin and other anions (MRP2), or for bile acid synthesis. Canalicular excretion of bile acids and conjugated bilirubin are severely impaired, and cholestasis develops early in life. Some are more common in infants born to mothers with obstetric cholestasis. Surprisingly, external biliary drainage may be beneficial.
Extrahepatic cholestasis in the neonate is most commonly due to biliary atresia, but a choledochal cyst or bile duct perforation can also cause jaundice at this age. Biliary atresia appears to represent a form of sclerosing cholangitis with progressive loss of intra- and extrahepatic ducts. HIDA scans, percutaneous liver biopsy, and retrograde cholangiography can establish the diagnosis without laparotomy.
Benign recurrent intrahepatic cholestasis (BRIC)
In this rare syndrome, recurrent reversible episodes of cholestasis start in childhood or adult life. Each attack is characterized by jaundice, anorexia, and itching for several months, which then subsides with no residual effects. Hepatic histology only shows cholestasis. Phenobarbitone or ursodeoxycholic acid may shorten and attenuate attacks. So far two mutations have been identified in patients with BRIC, one of which is similar to that in Byler’s disease.
Jaundice due to halothane hepatitis, post-transfusion viral hepatitis, incompatible blood transfusion, drugs, and bile duct damage is described elsewhere.
Prolonged intrahepatic cholestasis used to be common after cardiac surgery. It is related to the length of surgery and intraoperative cardiac function, and may be due to reduced hepatic blood flow during surgery. Improvements in intra- and postoperative care seem to have improved hepatic function and rendered the syndrome uncommon. Transfused red blood cells are prone to rapid haemolysis and this increases the bilirubin load, while impaired renal function reduces the urinary excretion of conjugated bilirubin. Drug-induced liver injury should be considered.
Prolonged parenteral nutrition is sometimes associated with cholestasis, fatty liver, and eventually fibrosis, especially in the neonate. The mechanism is unclear.
Cholestasis after liver transplantation has multiple causes, including rejection.
It is increasingly likely that the sensitivity of many examples of idiosyncratic drug-induced cholestasis is due to genetic abnormalities of the bile acid, phospholipid, or bilirubin excretion systems.
Cholestasis of pregnancy (obstetric cholestasis)
Slight impairment of the hepatic excretion of bilirubin can be demonstrated during normal pregnancy or after the administration of oestrogens, but in less than 1% of pregnancies bilirubin and alkaline phosphatase levels rise during the third trimester and intolerable itching and frank jaundice develop, all of which rapidly remit after delivery. The severity of obstetric cholestasis increases in successive pregnancies. There is an increased incidence of premature births, fetal distress, and intrauterine death, and so premature induction of labour may be needed. The incidence is higher in South America than Europe, and it is commoner in mothers of babies with familial intrahepatic cholestasis. This is not surprising because the syndrome is caused by mutational dysfunction of the biliary canalicular transporter proteins MRP3 and BSEP, which is exacerbated by the high levels of oestrogens and progesterone in pregnancy. Thus the contraceptive pill frequently causes a milder syndrome in the same susceptible women. Ursodeoxycholic acid is reported to ameliorate the condition and is safe, at least during late pregnancy. Phenobarbitone may help the itching, although there is a small risk of impairing neonatal respiration. Cholestyramine has also been used.
Other causes of jaundice in late pregnancy should be remembered, including acute fatty liver, extrahepatic biliary obstruction, such as from gallstones, and toxaemia.
Pregnancy, by affecting bilirubin excretion, may bring to notice the jaundice of primary biliary cirrhosis or the Dubin–Johnson/Rotor syndromes.
Abnormal liver-related blood tests, and occasionally cholestatic jaundice, often develop during bacterial/viral infections, unrelated to the administration of drugs. In animals this has been shown to be due to endotoxins and cytokines that rapidly down-regulate and translocate the canalicular transport protein MRP2, which excretes conjugated bilirubin into the canaliculus. At the same time other pump proteins are up-regulated, a complex rearrangement that may protect the hepatocyte against oxidative damage. Jaundice is especially common in patients with glucose-6-phosphatase deficiency when they develop sepsis, such as pneumonia, since the haemolysis exacerbates the jaundice. This combination of high conjugated bilirubin levels and sepsis is particularly damaging to the kidney.
Dubin–Johnson and Rotor syndromes
These are two rare, familial forms of nonhaemolytic, conjugated hyperbilirubinaemia without cholestasis.
The Dubin–Johnson syndrome (OMIM 237500), first described in 1954, is a chronic, relapsing jaundice, without itching or raised serum bile acids. Other liver-related blood tests are normal, but there are associated defects in the excretion of other anions, such as bromsulphthalein, radiographic dyes, and urobilinogen. Hence cholecystography fails, there is excess urobilinogen in the urine, and a delayed rise of the plasma levels of bromsulphthalein after an injection of the dye due to reflux of the conjugated anion from hepatocytes. Jaundice increases during pregnancy or when taking the contraceptive pill because oestrogens further impair bilirubin excretion. A black pigment accumulates in the liver so that at laparoscopy the liver appears strikingly black, as do needle biopsy specimens. Urinary coproporphyrin excretion is abnormal. Some at least seem to be due to abnormality of the MRP2 canalicular bilirubin transporter. The inheritance seems to differ between families, and a similar condition occurs in a mutant strain of Corriedale sheep, although then photosensitivity also occurs, and in a laboratory rat model. Other families have been described in which there are similar findings but no hepatic pigment, the so-called Rotor syndrome (OMIM 237450). No treatment of either syndrome is required apart from reassurance, and support when seeking life insurance.