An analysis of these cases indicated two striking characteristics. One was the time sequence of the three diphasic changes which was uniform for all attacks in all subjects. The second was that although they differed in intensity, all the attacks were qualitatively alike.

In further studies we tried to understand the meaning of these changes. Similar changes, but without hemoglobin in serum and urine, and with only slight chill as a clinical manifestation, could be observed when the hands of normal individuals were placed in icy water for 10 minutes. Two similar diphasic phenomena were seen to appear. While their intensity was greatly reduced, the time of appearance of the two diphasic phenomena observed was basically the same as in patients with cold hemoglobinuria. This would indicate that the leucolysis in hemoglobinuria a frigore corresponds to a physiological response which induces hemolysis because the red cells have been sensitized by cold. This sensitization is recognized in the Donath Landsteiner reaction. It appears probable that leucolysis liberates the complement necessary to induce the hemolysis of the sensitized red cells.

We have tried to correlate these changes with other processes encountered in normal and abnormal physiology. In differential studies of blood smears obtained during an induced attack of hemoglobinuria, nuclear shadows were observed in the smears at the time of the leucopenia of the first phase. This was exceptionally manifest for the second diphasic complex, when a marked leucopenia occurs. We have seen above that this leucopenia has been correlated to the lysis of the leucocytes.

In patients with cold hemoglobinuria, we were able to show that, if with any physical exercise after the hands were out of the icy water, such as even walking in the room, the severity of the attack induced was much greater than when they were allowed to rest quietly. Not only was the severity of the chill and the degree of hemoglobinuria corresponding to the second diphasic phenomenon greatly increased, but all other changes were similarly intensified. The number of leucocytes decreased to less than 200 per cubic mm. Blood coagulation time went to values as high as over1/2 hour with almost no retraction of the clots. The antitryptic power of the serum, obtained only after centrifugation, reached the lowest values observed. Figs. 266 and 267 show the changes for the same subject with and without physical exercise. These findings could explain observations indicating the importance of rest, after blood transfusions. With the direct transfusion method, where 500 cc. of blood was administered in less than 10 minutes, chill was seldom seen in patients resting quietly, while it was constantly seen to appear in subjects taking any exercise immediately following the transfusion. (Fig. 265)

In the patient with hemoglobinuria a frigore

Fig. 267. In the patient with hemoglobinuria a frigore, exercise after immersion of the hand in icy water induces a violent chill with abundant hemoglobinuria.

The administration of 0.03 gm

Fig. 268. The administration of 0.03 gm. Pantopon prevents the appearance of manifestations after the immersion of the hand in icy water, in the subject with hemoglobinuria a frigore.

This relationship of leucolysis to the pathogenesis of the diphasic complex phenomenon in hemoglobinuria a frigore was confirmed when the attack could be prevented by a pretreatment which influenced the leucocytes. The study of the influence exercised by various agents in vitro upon the leucocytes of a pleural effusion preparation from rabbits, when collargol solutions were added, has shown that only morphine and other opium alkaloids were able to prevent lysis. Adrenaline and quinine—among others —appeared inactive. Administered to patients with paroxysmal hemoglobinuria, these last substances did not prevent attacks. However, morphine in minimal doses, such as 5 mgr. (1/12 g.) or less, by intravenous injection, entirely prevented the development of any clinical manifestations as well as analytic changes. Minimal or no leucopenia was seen, and there were no changes in coagulation time, retraction of the clot, antitryptic power, refraction index of the serum, esterase, etc. (Fig. 268)

In addition to confirming the hypothesis that leucolysis intervenes in the pathogenesis of the diphasic complex, the influence of morphine has furnished a means of preventing the attacks.

The presence of two or three successive diphasic phenomena in the case of paroxystic hemoglobinuria attack has revealed another important relationship. As mentioned above, hemolysis has been seen to occur only in the first phase of the diphasic phenomenon. Hemoglobin disappears from the serum between the complexes. It is classically accepted that in cold hemoglobinuria two factors must intervene to produce hemolysis. One is sensitization of the red cells which results from the influence of cold; the other is the presence of the complement. It appears that while sensitization persists for a while after exposure to cold, hemolysis will occur when the second condition is also fulfilled. Complement is released in the first phase of the diphasic phenomena since it is during the first phase that hemolysis occurs through the changes induced by the lysis of the leucocytes. The correlation between these two processes is clear from the fact that, under the influence of morphine, leucolysis does not take place and consequently hemolysis is prevented.