RESPIRATORY PROPERTIES OF THE ARTERIAL BLOOD IN NOR - MAL MAN AND IN PATIENTS WITH DISEASE OF THE LIVER : POSITION OF THE OXYGEN DISSOCIATION CURVE By ANCEL KEYS AND ALBERT

It is now generally agreed that the oxygen saturation of the arterial blood in normal man in rest at low altitudes is remarkably constant. Until recently, subnormal arterial saturations could almost invariably be traced to obvious pulmonary or cardiac disorders. Recently, however, it was noted that relative arterial unsaturation is common among patients with cirrhosis of the liver and other conditions involving severe parenchymatous hepatic damage (Snell (1)). In such cases, circulatory disturbances, ascites and edema did not appear to account for the unsaturation, and it was pointed out that the anoxemia in these cases must have resulted either from some interference with gas exchange in the lungs or from an abnormality of the blood itself. In the cases which came to necropsy, however, pathological changes in the lungs sufficient to account for the phenomenon were not seen, though minor alterations, for example, edema of the alveolar walls, could not be absolutely excluded. By the process of elimination an alteration of the blood itself was suggested, and one of us (Snell (2)) produced evidence to indicate that the oxygen dissociation curve of the blood was abnormal in these cases of hepatic disease. Two major objections could be raised against the acceptance of this conclusion. In the first place, the dissociation curves had all been studied at constant partial pressure of carbon dioxide rather than at constant pH; even in normal man this procedure may erroneously indicate alterations in the position of the oxygen dissociation curve and, in pathological conditions, such a method of comparison may lead to gross errors. In the second place, normal standards for the position of the oxygen dissociation curve, based on modem technic, had not yet been definitely established. The present study is a reexamination of the problem, with special emphasis on the question of the position of the oxygen dissociation curve. This point is particularly important because the fundamental constancy of the position of the oxygen dissociation curve of human blood at constant pH has been demonstrated in a wide variety of pathological conditions: in nephritis (3), pernicious anemia (4), diabetic coma (5), and in men acclimatized to chronic anoxemia (6).