Statistical Significance of Erythrocyte Counts During Responses to Liver Extract in Normal Individual.

Abstract

Marked increases in the number of circulating erythrocytes have recently been reported following the administration of the Cohn-Minot liver extract to normal individuals. 1 The outstanding features of the curves obtained were 2: rapid and considerable increases as well as marked irregularities in the number of erythrocytes leading to deep and abrupt depressions alternating with peaks. The curves were constructed from daily blood counts.

It seems desirable to supplement these observations by more frequent determinations of the number of erythrocytes, as well as by an inquiry into the statistical significance of the erythrocyte counts. One of the persons (H. B-d) who had previously in 2 separate experiments responded to the liver extract both with peaks up to 7,870,000 and 7,700,000 erythrocytes respectively and with deep depressions was chosen for the experiment. The erythrocyte counts of this individual have been consistent over a number of years. As early as 1906 the counts gave values of 5,500,000 (mean of 5 or 6 determinations) and in the 2 recent experiments the curves started at a level of 5,950,000 and 5,400,000 respectively.

In order to determine the significance of the variations in the erythrocyte counts in the normal individual, observations were made at 3 hour intervals during the day before liver extract was started. Following this, the Cohn-Minot liver extract (Lilly 343; lots 29157-815244-815245) was given for 6 days in the dosages shown in Graph 1. On the 1st day counts were made at 3 hour intervals from 9:00 a. m. to 9:00 p. m. On the 2nd and 3rd days counts were made at 3 hour intervals from 6:00 a. m. to 12:00 midnight, and on the remaining 5 days counts at 2:00 a. m. were substituted for those at midnight. Each time erythrocyte counts were made, blood was taken from 3 fingers, using a different pipette for each finger. Three counts were then made from each pipette, making only one count for each filling of the counting chamber, giving a total of 9 counts for each period of the day. In making each count the cells covering 80 small squares were counted. The mean, standard deviation and probable error of the mean were calculated for each series of 9 counts. Comparisons were made of the means for different times of the normal day with the means for all other times of that day, and the significance of the difference tested by the usual formula for the probable error of a difference. The differences of the means for the pre-liver day were found to be statistically insignificant. Therefore all of the data for the pre-liver day were grouped in one series to establish a mean for use as a normal base.