In the majority of cases with purpuric eruptions, the symptoms do not implicate the alimentary tract, and there is no reason to connect it with the origin of the disease; but in a minority such symptoms may obtain, or a gastro-intestinal origin may be suspected.

In purpura haemorrhagica, for example, haemorrhage may occur from any part of the tract, epistaxis and intestinal haemorrhage being the most common, and the loss of blood in either case may be copious and prolonged; and both may occur without any special gastro-intestinal symptoms or with such as are evidently the results of a general and widespread affection.

Cases of the type which are known as Henoch's purpura may, however, be associated with severe gastro-intestinal symptoms, though their relationship is somewhat ill-defined. It occurs mainly among children, and attacks are apt to recur over a period of years. The principal feature is the paroxysmal occurrence of purpuric eruptions with arthritis and abdominal symptoms, the chief of which are colic, which is usually severe, vomiting and diarrhoea, blood being often present in the vomit and the stools. The initial symptoms vary, the abdominal disturbance sometimes preceding the eruption and sometimes succeeding it, while both may arise at the same time. Dr. Osier has emphasized the variety of the cutaneous lesions that may be met with, purpura, erythema, angioneurotic oedema and urticaria, sometimes appearing in successive attacks; and he apparently considers that the symptom-complex is most frequently the result of some general condition, the visceral manifestations being the result of gastro-intestinal lesions comparable to those that occur on the cutaneous surfaces. A haemorrhagic nephritis is an occasional complication, but one, however, which frequently leads to a fatal termination.

In certain cases purpura is of toxic origin and may follow the administration of such drugs as quinine, belladonna, iodide of potassium, etc., and in another group seems to be associated with hepatic or pancreatic lesions. It thus seems possible that in some cases of Henoch's purpura the abdominal disturbances may be the cause of the illness, and on general grounds the condition of the alimentary tract should be given particular attention.

It is always wise, at the onset of an attack, to ensure that the bowel is emptied of any irritant by the administration of a laxative, unless the diarrhoea has been severe and continued, or haemorrhage has occurred. Milk is in these cases the most suitable food, and may be diluted or predigested, if necessary.

In the sub-acute cases and in those where intestinal symptoms are absent, the dietary should be bland and unirritating, but the maximum of food that is possible should be given to replace, so far as may be, the loss of blood. The finer farinacea, Benger's and Mellin's foods, eggs, white fish, etc., can often be utilized.

During convalescence a more nourishing diet should be ordered, the items consisting of the "fresh" foods, and containing a suitable proportion of the red meats.

The diet between attacks should be carefully supervised, and the ingestion of irritating articles prevented; but it must be acknowledged that the most careful regulation often fails to prevent a recurrence of the attacks, and their origin must be sought elsewhere.