This paper critically examines the usefulness of serum albumin measurement in the light of current laboratory practice and knowledge of the pathophysiology of albumin metabolism. The main conclusions and recommendations are as follows: Albumin measurement forms a limited, but useful part of the investigation of liver disease; a normal serum albumin concentration makes the diagnosis of cirrhosis unlikely, while a low level in viral hepatitis suggests either severe hepatocellular damage or other complications. Albumin measurement is essential in selecting patients for, and in determining the amount and frequency of, albumin replacement. Serum albumin concentration provides a useful indication of prognosis in myeloma. In the long-term management of patients undergoing enteral or parenteral nutrition, serum albumin concentration is one of several parameters which, together, are useful in predicting the outcome of treatment. (v) The serum albumin concentration may provide a clue to the aetiology of unexplained oedema. Serum albumin measurement is useful in indicating the level of ionised calcium and of unbound unconjugated bilirubin.

The predictive value of plasma protein changes in disease is very largely unknown, as relevant clinical and laboratory studies are lacking. The usefulness of zone electrophoresis as a method of detecting a constellation of plasma protein changes is even less clear. With the advent of precise automated techniques for specific plasma protein measurement, zone electrophoresis has little to offer except in the identification and quantitation of paraproteins, where it is essential.