Urinary Cholesterol in Cancer: Chemical State of Urinary Cholesterol and Methods of Estimation

Abstract

DURING recent years two groups of workers have reported a notable increase in urinary cholesterol excretion in cancer. Bloch and Sobotka (1938) worked with pooled urine specimens from cancer wards and compared them with pooled specimens from cardiac and tuberculosis wards, and found approximately ten times as much cholesterol in the cancer urine as compared with the cardiac and tuberculosis urine. In this study the cholesterol was isolated in relatively large quantities (0.35 g. per 100 litres of urine), and its identity was proved by determination of melting-point, optical rotation and preparation of the acetate.This work was later extended by Sobotka, Bloch and Rosenbloom (1940) to individual cases of cancer using a less specific colorimetric method, depending on the Liebermann-Burchard reaction. They reported an increase of urinary cholesterol in 19 out of 92 cancer patients. Later, Bruger and Ehrlich (1943), using a similar method of estimation, found a higher incidence of hypercholesteroluria, 11 of their 32 cases giving results above normal limits. However, the actual concentration of cholesterol reported by Bruger and Ehrlich in normal urine was approximately twice as great as that found by Sobotka, Bloch and Rosenbloom. It appeared, therefore, that the methods used in these two studies must have been significantly different, and we felt that further work on the technique of estimation was needed. While admitting the possible non-specificity of the Liebermann-Burchard reaction, we felt that this type of method was the only one likely to find immediate clinical application, and it does appear to have given results consistent with the original work of Bloch and Sobotka. The present work was undertaken with the main object of finding out whether refinements of this colorimetric method of estimation would be of any value in the diagnosis of cancer.EXPERIMENTAL.Extraction method.The chloroform extraction method of Neustadt, Howard and Myerson (1946) was first tried, but the amount of colour obtained at the final stage from the normal urines was too small for accurate measurement. Adaptation of this method, using the whole of the extract from 200 ml. of urine (in place of 50 ml.), gave a brown rather than a green colour with the Liebermann-Burchard reagent, as shown in Fig. 1. An attempt was made to correct for this error by the introduction of a blank reading, which was obtained by adding to half the urinary extract in 5 ml. of chloroform 0-9 ml. of ethanol and 0-1 ml. of concentrated H2S04 in place of the Liebermann-Burchard reagent. This sometimes gave