Hemoglobin Standards in Normal Men.

Abstract

In an earlier report on normal blood standards, 1 a Newcomer hemoglobinometer, restandardized on the basis of 10 hemoglobin determinations made by the Van Slyke oxygen capacity method, was employed. Further standardization of this Newcomer instrument has led to the conclusion that the original correction curve was incorrect.

Additional hemoglobin determinations by the Newcomer and Van Slyke methods have been carried out on 22 different samples of blood. Of this number 19 determinations (Table I) which checked very closely on 4 different Van Slyke instruments have been selected and a correction curve plotted on the basis of the total of 29 determinations.

The method employed for the plotting of the correction curve is that recommended by Pearl. 2 The hemoglobin determinations on the 2 instruments showed a high degree of correlation, the correlation coefficient (r) being .8669 ± .0984. From the fact that the correlation ratio (n) for these 29 determinations is not significantly different from the correlation coefficient (n = .9349 ζ = n 2 — r 2 = .1163 ± .0835) it has been concluded that the correlation is linear. Correction curves have been calculated by the method of averages and the method of least squares. None of these has been found altogether satisfactory. A logarithmic curve of the formula y = 6.25 + 1.038 x + .0286 log x has finally been selected as the best fitting curve. (See Figure 1.)

As the result of this recalibration, the author has come to the conclusion that the hemoglobin figures already reported for normal young men in the South 1 are too low. By the new calibration the published average of 15.85 gm. becomes 17.0 gm., with 73% of the determinations ranging between 15.5 and 18.5 gm.