Riboflavin, vitamin B2, serves in its phosphorylated form as the prosthetic group for a number of tissue enzymes important in biological oxidations (1). The involved substrates include glucose, lactic acid, the d-amino acids, and other compounds of biological interest (1, 2, 3). Riboflavin deficiency has been described in man (4), and it seems possible that further clinical study of this vitamin may reveal alterations in its metabolism which will increase our knowledge of the r6le played by these enzymatic oxidation systems in health and in disease. To aid in interpreting such a study, a convenient method has been sought for following the urinary excretion of riboflavin.It is known that riboflavin occurs in urine in the form of uroflavin, a pigment almost identical with it in composition,' properties, and vitamin activity (5), but the methods usually recommended for its determination (5,6,7,8,9) were not found by the writer to be sufficiently convenient or accurate for routine use. More specific and accurate determinations are possible, and smaller and more convenient volumes of urine (I to 10 cc. instead of 100 to 500 cc.) can be used when the pigment is estimated not by its optical density but by its intense greenish-yellow fluorescence.2 In order to do this, the effect of interfering substances must be avoided. With urines of normal individuals containing 0.5 gamma or more flavin per cc., this is readily accomplished by diluting down the salts and other materials to the point where they no longer affect fluorescence (10) and by destroying the greater part of other pig-1 The carbon analysis of the purest preparation isolated did not quite agree with that for riboflavin (Koschara (5)).2 Pfaltz & Bauer, New York City. Model A is a suitable fluorometer provided with an additional cell for the measurement of optical density. ments and fluorescent materials in a brief permanganate toxidation (5). With urines of certain patients containing less than 0.5 gamma per cc., some preliminary concentration of the uroflavin is necessary, and can be much more conveniently and effectively carried out by using an adsorption column than by adding large quantities of adsorbing agent directly to urine (11,12). In measuring the fluorescence of the column eluates or the diluted urine samples, the usual ultra violet radiation is not satisfactory since it excites the fluorescence of too many substances. Measurements of flavin fluorescence are much more specific and much more accurate when carried out with an exciting beam of visual wavelength, Mercury line at 4358A°, and when the fluorometric photoelectric cell is protected by a yellow filter.3 With such an arrangement, the relationship between fluorescence and riboflavin concentration is linear for column eluates and diluted urine samples (Figures 1 and 2), particularly when the effect of any occasional alteration in optical density is corrected by means of the empirical curve illustrated in Figure 3.
METHODS