Bilirubin, the principal tetrapyrrole, bile pigment and catabolite of haem, is an emerging biomarker of disease resistance, which may be related to several recently documented biological functions. Initially believed to be toxic in infants, the perception of bilirubin has undergone a transformation: it is now considered to be a molecule that may promote health in adults. Data from the last decade demonstrate that mildly elevated serum bilirubin levels are strongly associated with reduced prevalence of chronic diseases, particularly cardiovascular diseases (CVDs), as well as CVD-related mortality and risk factors. Recent data also link bilirubin to other chronic diseases, including cancer and Type 2 diabetes mellitus, and to all-cause mortality. Therefore, there is evidence to suggest that bilirubin is a biomarker for reduced chronic disease prevalence and a predictor of all-cause mortality, which is of important clinical significance. In the present review, detailed information on the association between bilirubin and all-cause mortality, as well as the pathological conditions of CVD, cancer, diabetes and neurodegenerative diseases, is provided. The mechanistic background concerning how bilirubin and its metabolism may influence disease prevention and its clinical relevance is also discussed. Given that the search for novel biomarkers of these diseases, as well as for novel therapeutic modalities, is a key research objective for the near future, bilirubin represents a promising candidate, meeting the criteria of a biomarker, and should be considered more carefully in clinical practice as a molecule that might provide insights into disease resistance. Clearly, however, greater molecular insight is warranted to support and strengthen the conclusion that bilirubin can prevent disease, with future research directions also proposed.

Recent epidemiological and clinical data show protection from CVD (cardiovascular disease), all-cause mortality and cancer in subjects with GS (Gilbert's syndrome), which is characterized by a mildly elevated blood bilirubin concentration. The established antioxidant effect of bilirubin, however, contributes only in part to this protection. Therefore we investigated whether mildly elevated circulating UCB (unconjugated bilirubin) is associated with altered lipid metabolism. The study was performed on GS and age- and gender-matched healthy subjects ( n =59 per group). Full lipoprotein profile, TAG (triacylglycerols), Apo (apolipoprotein)-A1, Apo-B, lipoprotein(a), the subfractions of LDL (low-density lipoprotein) and selected pro-inflammatory mediators were analysed. A hyperbilirubinaemic rodent model (Gunn rats, n =40) was investigated to further support the presented human data. GS subjects had significantly ( P <0.05) improved lipid profile with reduced total cholesterol, LDL-C (LDL-cholesterol), TAG, low- and pro-atherogenic LDL subfractions (LDL-1+LDL-2), Apo-B, Apo-B/Apo-A1 ratio and lower IL-6 (interleukin 6) and SAA (serum amyloid A) concentrations ( P =0.094). When the control and GS groups were subdivided into younger and older cohorts, older GS subjects demonstrated reduced lipid variables (total cholesterol and LDL-C, TAG and LDL-C subfractions, Apo-B/Apo-A1 ratio; P <0.05; Apo-B: P <0.1) compared with controls. These data were supported by lipid analyses in the rodent model showing that Gunn rat serum had lower total cholesterol (2.29±0.38 compared with 1.27±0.72 mM; P <0.001) and TAG (1.66±0.67 compared with 0.99±0.52 mM; P <0.001) concentrations compared with controls. These findings indicate that the altered lipid profile and the reduced pro-inflammatory status in hyperbilirubinaemic subjects, particularly in the older individuals, probably contribute additionally to the commonly accepted beneficial antioxidant effects of bilirubin in humans.

1. The plasma disappearance rate and the increment in plasma unconjugated bilirubin after intravenous administration of 5.9 μmol of rifamycin SV (RSV)/kg body wt. were investigated in 51 subjects with Gilbert's syndrome and 35 control subjects of both sexes. 2. Both the plasma disappearance rate and the unconjugated hyperbilirubinaemia after RSV administration were higher ( P <0.001) in Gilbert's syndrome. Females, both normal and with the syndrome, showed a significantly shorter t 1/2 and a lower hyperbilirubinaemic response as compared with males. A linear correlation ( P <0.001) was present between RSV plasma half-life and the hyperbilirubinaemic response. 3. In vitro , RSV was shown to inhibit sulphobromophthalein (BSP) uptake in rat liver plasma-membrane vesicles with a K i of 20 μmol/l. Evidence that this effect was due to competition for bilitranslocase was sought on preparations of the purified protein. Under these experimental conditions, RSV inhibited BSP binding with a K i of 17 μmol/l. 4. Since RSV competes with BSP for binding to bilitranslocase in vitro , the data are interpreted as suggesting that reduced bilitranslocase function might underlie the delayed RSV plasma clearance and the exacerbated unconjugated hyperbilirubinaemia present in Gilbert's syndrome.

1. To test the hypothesis that the presence of food in the intestine plays a role in the control of serum bilirubin levels, the effect of a 1674 kJ (400 kcal) diet, including non-absorbable material able to produce an intestinal bulk, was compared with the effect of caloric restriction alone. 2. In normal individuals ( n = 10) the rise of plasma bilirubin was comparable after the two trials. In patients with Gilbert's syndrome ( n = 18), bilirubin concentration rose to a much greater extent during the caloric restriction as compared with controls, and to a lesser extent ( P <0.001) after the bulky diet, as compared with caloric restriction alone in the same patient. 3. It is proposed that a factor originating from the intestine, perhaps a hormone, the release of which is dependent on the presence of food-bulk, plays a role in the control of hyperbilirubinaemia associated with caloric restriction.

1. Clofibrate (20 mg day −1 kg −1 body weight) given orally for 2 weeks to 18 subjects with Gilbert's syndrome reduced the total serum bilirubin concentration from 44.4 (22.1–71.7) μmol/l (median, range) to 15.0 (7.9–28.9) μmol/l, mainly by decreasing the ‘indirect’ fraction to 34.5% of pretreatment values. 2. In contrast to treatment with phenobarbitone, clofibrate did not change the ratio of bilirubin mono- to di-conjugates in bile. 3. The initial plasma disappearance rate of indocyanine green and of bromosulphophthalein did not show consistent changes during clofibrate treatment but the ‘apparent’ maximal biliary secretion capacity of bromosulphophthalein ( T m ) decreased in most of the patients studied, whereas the relative storage capacity ( S ) tended to increase. 4. As expected, the cholesterol saturation of bile increased in all subjects. 5. These results suggest that clofibrate increases mainly the glucuronidation of bilirubin, promoting as such the overall hepatic transport in Gilbert's syndrome. However, it decreases the maximal biliary secretion of bromosulphophthalein, possibly by an increased hepatic retention. These phenomena might be linked to the augmented content of Z protein in the liver.

1. The kinetics of non-radioactive bilirubin were investigated in 100 patients with chronic non-haemolytic unconjugated hyperbilirubinaemia and in nine normal volunteers. 2. The patients with Gilbert's syndrome were divided on the basis of a two-compartment model into two distinct groups: group 1 (74 patients) with decreased hepatic uptake and conjugation of bilirubin and a bilirubin turnover rate within the normal range, and group 2 (26 patients) having normal uptake, impaired conjugation and increased bilirubin turnover rate. 3. On the basis of these findings, studies were made of early pigment production in four patients of group 1, in six patients of group 2 and in four of the normal volunteers, with [ 14 C]glycine and ş-[ 3 H]aminolaevulinic acid. 4. After receiving injections of labelled haem precursors, the four patients in group 1 had a normal rate of incorporation of both tracers, whereas abnormalities in incorporation were found in group 2. Four patients showed an increase of the first component of the early peak of [ 14 C]bilirubin as well as of the early peak of [ 3 H]bilirubin, suggesting an increased turnover of hepatic haem. Two patients showed an increased incorporation of [ 14 C]glycine characterized by the fusion of the two early peaks, whereas the incorporation of the δ-[ 3 H]aminolaevulinic acid was normal, indicating the presence of a primary shunt hyperbilirubinaemia. 5. The results confirm that Gilbert's syndrome is a heterogeneous condition with respect to bilirubin turnover rate. The population of Gilbert's syndrome with increased bilirubin turnover rate comprises not only patients with an increased haem erythroid turnover, but also patients with increased metabolism of nonerythroid haem protein in the liver.

1. Intravenous doses of bilirubin (3.4 μmol/kg) were given to normal subjects and patients with Gilbert's syndrome. Both groups displayed an identical initial disappearance of a substantial proportion of the bilirubin but, late in time, the Gilbert's patients exhibited reduced clearance with a sustained elevation of the plasma bilirubin and no reflux into the plasma space of conjugated bilirubin. Increasing the dose in normal subjects (by factors of 3 and 6) failed to reproduce the response found in the Gilbert's patients. 2. In the bile-containing duodenal aspirates of Gilbert's patients the average proportion of bilirubin found as bilirubin diglucuronide was 68% (normal 88%) and of bilirubin monoglucuronide, 23% (normal 7%). Both differences were significant at the P < 0.001 level. In the Gilbert's patients restriction of caloric intake to 1569 kJ/day for 2 days characteristically raised the serum bilirubin with no modification of the biliary pigment pattern; phenobarbital (180 mg/day for 2 weeks) decreased the plasma bilirubin to the normal range with, concomitantly, a reversion of the biliary pigment pattern towards normal. 3. We conclude that there is no hepatic uptake defect in Gilbert's syndrome but that there is decreased activity in the conjugation process underlying the addition of the second glucuronic acid moiety to bilirubin, to form bilirubin diglucuronide.

1. The kinetics of the plasma disappearance of bilirubin (2 mg/kg intravenously) were studied in 106 patients with Gilbert's syndrome and in 13 normal subjects. 2. All patients had significant decreases in hepatic bilirubin clearance and transfer rates from plasma to liver, resulting in increased values for plasma retention at 4 h. The calculated value for unconjugated bilirubin production was normal in 40% of patients and increased in the remainder. 3. In 29 of the Gilbert's patients their bromosulphthalein kinetics were studied 1 week before the bilirubin test. These results were essentially normal and it was concluded that the hepatic clearance mechanisms for bilirubin and bromosulphthalein are different. 4. In 10 patients the bilirubin transport maximum ( T m ) was found to be low whereas the relative storage capacity ( S ) was normal. Phenobarbitone treatment in four patients resulted in an increase in T m , and S decreased in two patients and remained unchanged in the other two. 5. These results support the hypothesis that there are several variants of Gilbert's syndrome and that the bilirubin tolerance test is a useful diagnostic test.

1. The clearance of a tracer dose of [ 14 C]bilirubin from the plasma was studied in patients with Gilbert's syndrome, congenital non-haemolytic jaundice, haemolytic jaundice and in normal subjects. Clearance was significantly impaired in the patients with Gilbert's syndrome and in those with congenital non-haemolytic jaundice when compared with the normal subjects, and was normal in the patient with haemolytic jaundice. 2. Treatment for 2 weeks with phenobarbitone (180 mg/day) lowered the plasma bilirubin concentration and improved all indices of plasma clearance of the isotope in patients with Gilbert's syndrome, so that they became indistinguishable from those of normal subjects. The improvement in plasma [ 14 C]bilirubin clearance in these patients was associated with modest increases in hepatic bilirubin glucuronyl transferase in some subjects. 3. Phenobarbitone treatment improved plasma bilirubin concentrations and plasma [ 14 C]bilirubin clearance in patients with congenital non-haemolytic jaundice, so that they resembled those seen in patients with untreated Gilbert's syndrome. Despite this improvement hepatic bilirubin glucuronyl transferase activity remained undetectable. 4. These results are compatible with the hypothesis that Gilbert's syndrome is a manifestation of a relative deficiency of hepatic bilirubin glucuronyl transferase, and differs from congenital non-haemolytic jaundice only in severity.