Urinary bile acids in population screening for inapparent liver disease.

BACKGROUND/AIMS: We performed this study in order to evaluate the diagnostic potential of bile acids in random samples of urine for detection of latent liver disease and to compare a radioimmunoassay for urine bile acids with an enzymatic method that detects bile acid sulfates. This was a prospective cohort study carried out at the VA Medical Center involving 151 adults who attended a Community Health Fair at the hospital and wanted to know if they had liver disease.

METHODOLOGY

Urinary bile acids in random specimens of 5-10 ml urine were measured. Radioimmunoassay for primary bile acids and an enzymatic assay with or without sulfated bile acids, all corrected with creatinine for urine flow were performed. Serum primary bile acids were determined by radioimmunoassay. In addition, routine liver profile and clinical examination were carried out.

RESULTS

In 78 of 151 subjects there was at least one recent liver profile to match with the urine bile acids. Of these 78, 52 subjects with normal urine bile acids had a normal liver profile. In 11 subjects abnormal urine bile acids were associated with an abnormal liver profile. Nine of these 11 subjects were anti HCV positive, one was HIV positive. Urine bile acids correctly predicted the outcome of routine liver tests in 89% of 78 subjects. In nine cases there was a discordance between urinary bile acids and the liver profile. Failure to correctly predict the liver profile using urine, was reduced from nine subjects to three when urine bile acids were obtained twice at separate intervals. Urine bile acids predicted the outcome of anti HCV testing in 37 subjects with similar accuracy as serum ALT or AST. Urine bile acids correlated with serum bile acids at r=0.96, 0.88 and 0.76 for the radioimmunoassay, enzymatic assay that included sulfated bile acids and enzymatic assay without the sulfates, respectively.

CONCLUSION

Bile acids in a random sample of urine are useful for population screening for latent liver disease. Prediction of sub-clinical hepatitis C is comparable to that of serum ALT or AST. Inclusion of bile acid sulfates mildly increases the predictive value of urine. Urine bile acids highly correlate with serum bile acids, indicating their surrogate diagnostic value.