All observers (1 to 5) agree that the concentrations of lipid constituents in the serum of normal individuals vary greatly. In fact, the degree of variability is so great that no significance can be attached to a single observation unless it is extremely abnormal. On the other hand, there is evidence that the degree of variability of certain lipid components-for example, cholesterol-in a given normal person, is far more restricted (4 to 7). Furthermore, the relative proportions of certain of the constituents-for example, the ratio of free to total cholesterol-are quite constant, not only in a given individual, but also in the population at large (4,8). A large series of measurements of serum lipids from normal persons and patients has, therefore, been analyzed, in an effort to discover criteria which may permit more accurate evaluation of such measurements in pathologic or disturbed physiologic states.
MATERIAL AND METHODSThe material consists of a large number of measurements of lipids in sera, collected during the last 10 years, from normal persons and patients with various diseases. The subjects were all in the postabsorptive state. The clinical material was derived chiefly from the medical and psychiatric services of the New Haven Hospital and Dispensary.The analytical methods of Man, Gildea, Peters, and Bogdanovitch (9 to 13) were employed. In a large proportion of the determinations, cholesterol, total fatty acid, and lipid phosphorus were measured; in a number, however, lipid phosphorus was omitted. In a few instances, both free and total cholesterol were determined; and in a still smaller fraction, iodine numbers were also included. For the last, the procedure of Yasuda (14) Sperry (8) found that the ratio of free to total cholesterol in a large series of normals varied only from 0.24 to 0.30. Brun (4) has reported variations from 0.24 to 0.32, which agrees with our own experience. The averages in all of the 3 series lay between 0.26 and 0.28. The maximum error in the calculation of cholesterol fatty acids, by the formula adopted, should not, therefore, exceed 6 per cent.The relative proportions of phospholipids are less certain. Only a few measurements of the concentration of sphingomyelin in serum have been published, and these differ greatly (15,16). Whether, however, this compound makes up 40 per cent of the phospholipid or is completely absent from serum, the apparent extreme possibilities, the error of estimating phospholipid fatty acids by the formula adopted will not exceed 10 per cent.Free fat is estimated, in terms of m.eq. of fatty acid per liter, by the formula, Fat fatty acid = total fatty acid-(Cholesterol ester fatty acid + phospholipid fatty acid). Being estimated by difference, it is subject to the summated errors of the other calculations. With the factors selected for the calculation of fatty acids of cholesterol esters and phospholipids, only one negative value,-4.3 707