In studying hemorrhage, it seemed necessary first to define precisely the conditions under which this important episode occurs, as we did for pain, itching, etc., since the course of hemorrhage, and particularly its response to treatment, seems to depend greatly upon certain peculiarities related to its pathogenesis. Hemorrhage results from a break in the continuity of a blood vessel, which can be induced by an external influence on a previously normal vessel, or can occur as a result of processes taking place in abnormal vessels. We have called the first type of hemorrhage "accidental" or "traumatic" and reserved the term "pathological" for the second, which appears to be a direct result of pathological changes in a vessel. The latter term is used in the same sense as it is used for fractures where "pathological" indicates only preexisting lesions in the bones.

In a traumatic hemorrhage, the therapeutic problem is limited to stopping the flow of blood. Mechanical means to close the bleeding vascular wound or agents able to increase the capacity of the blood to form clots can be used. They represent the only approach in accidental hemorrhage.

In pathological hemorrhage, other problems arise. The knowledge of the local pathological changes, which lead to the appearance of bleeding, is important both for preventing and controlling hemorrhage. The pathogenic factor involved will be considered along with the problem of hemo stasis.

Hemorrhagiparous Agents

Local and general factors are involved in pathological hemorrhage. Ulcerated lesions provide a favorable condition. So does local infection. Anticoagulants may act indirectly upon blood vessels.

Hemorrhagiparous properties of some substances became evident during research on their therapeutic use in cancer patients. For example:

P. T., a patient with carcinoma of the floor of the mouth widely ulcerated in the submaxillar region, bled sporadically from the lesion.

When a dose of 20 drops of Coramine (brand of nikethamide) was given for his general condition, a relatively severe hemorrhage appeared immediately afterward. There was another hemorrhage when a similar dose was administered 12 hours later. Suspecting a correlation between hemorrhage and medication, medication was discontinued and there was no hemorrhage during the following two weeks. When only 10 drops of Coramine solution was given again, a new hemorrhage appeared. After two more weeks without Coramine and without hemorrhage, a new dose was given to check the correlation between medication and bleeding, and was followed again by hemorrhage.

Reviewing cases lost through hemorrhage, we could find several in which lethal bleeding was preceded by administration of Coramine in the usual dose. This hemorrhagiparous effect was noted so consistently as to make us discontinue use of Coramine in cases with ulcerated lesions.

Alerted by this experience, we noted hemorrhagiparous activity in other agents. Glycerol, we found, could induce severe bleeding even when administered in doses as low as 5-10 drops, and two lethal hemorrhages were traced to such doses. Thiamine chloride in therapeutic doses—such as 100 mg. injections—produced hemorrhages. A similar hemorrhagiparous effect was noted for isamine blue. Glucose, administered intravenously in large doses to patients with ulcerated and infected cancerous lesions who had previously hemorrhaged, induced new bleeding. We must emphasize that all the hemorrhagiparous agents produced the bleeding effect, especially in subjects who had previously had hemorrhages from their lesions. This would indicate the importance of local changes. In such patients, other agents also induced bleeding. In some cases of gastro intestinal cancer, bleeding followed a normal dose of aspirin. But even in lesions which had never bled, the first group of hemorrhagiparous agents—Coramine, glycerol and thiamine—could induce bleeding.

In trying to correlate the pathogenesis of these hemorrhagiparous effects with existing offbalances, an interesting relationship was noted. A tendency to bleed was found to be promoted by sterols. Administration of cholesterol in doses as high as 5 cc. of a 2% solution in oil, two or three times a day, for several days, was followed by hemorrhage in patients with ulcerated lesions. This occurred only in a few cases. But a high sensitivity to other hemorrhagiparous agents often developed. The administration of minimal doses of glycerol produced bleeding in subjects treated with cholesterol and did so even more frequently when large amounts of insaponifiable fraction of organs were administered.

Hemorrhages which follow administration of these hemorrhagiparous agents are usually severe and of arterial character. Examination of such bleeding lesions has revealed transverse severance of small or even medium arteries. Only rarely was there oozing bleeding from capillaries or small veins. In several cases, administration of hemorrhagiparous substances induced petechiae or purpura, but this occurred only when thrombopenia was also present. The petechiae were seen at sites where local circulatory impairment already was present. We believe that this hemorrhagiparous effect must be emphasized for its clinical importance. Products such as Coramine (nikethamide) and thiamine are widely used as therapeutic agents and glycerol is a common vehicle for pharmaceutical preparations. Therefore, we must bear in mind their possible role in hemorrhages. In subjects with ulcerated lesions, their administration has to be banned or special precautions must be used. The same precautions have to be taken for the use of glucose in cases in which a previous hemorrhage has not been controlled by mechanical means.