To evaluate the extent and cause(s) of dialysis-related hypoxemia, we studied 10 patients, 7 days apart using acetate (AC) and bicarbonate dialysate (HCO₃). We measured arterial blood gases, WBC, minute ventilation (V_E) and inspired and expired gas concentrations and calculated the respiratory quotient (R) and the alveolar-arterial oxygen difference (A-a)DC·2 before and during hemodialysis. 8 patients developed hypoxemia. Arterial PO₂ (PaO₂) dropped similarly at 30 min from 93 ± 5 to 78 ± 6 (p < 0.05) and 89 ± 4 to 79 ± 5 mm Hg (p < 0.05) with AC and HCO₃, respectively. R and V_E decreased during AC (p < 0.05). (A-a)DC·2 increased at 30 min and correlated with the drop in PaO₂ during both AC (r = 0.68, p < 0.025) and HCO₃ (r = 0.76, p < 0.025). The fall in PaO₂ also correlated with the fall in WBC count for both AC and HCO₃ (r = 0.63, p < 0.005). The increase in arterial pH during HCO₃ (up to 7.45 ± 0.01) was significantly greater than that during AC (up to 7.42 ± 0.01) (p < 0.025), and coincided with a relative decrease in V_E. We conclude that (1) HCO₃ does not prevent hypoxemia, and (2) hypo ventilation V/Q abnormalities and increase in arterial pH, contribute variably to dialysis related hypoxemia depending on the type of dialysate and the time during dialysis.

SUMMARYExpanding the frequency response of the electrocardiogram and its derivative to 1,000 cps revealed notching in certain parts of the QRS complex which correlates with the presence of primary myocardial disease. Chi-square analysis of data from 169 patients with myocardial involvement indicated that notching on the downstroke of leads X, V4, and V6 separated abnormal from normal patients at the 1% level of significance, whereas fine and coarse slurring showed reverse correlation at the 1% level of significance. This suggests that notching and not slurring is the important evidence of disease. Neither notching nor slurring was significant at the peak of the R wave in any lead. Study of individual cases revealed that complete right and left bundlebranch blocks do not mask high-frequency components caused by myocardial disease nor do they produce high-frequency components. Conclusions regarding specific diagnostic criteria, however, should serve only as guidelines. ADD,ITIONAL INDEXING WORDS:Bundle-branch block Right ventricular hypertrophy M UCH of the clinically useful information derived from the electrocardiogram comes from gross alterations in the QRS complexes and T waves. These changes involve frequencies mostly below 100 cps' or in the low-frequency range, but patients with diffuse myocardial disease often will show reduced voltage and many high-frequency components in the QRS complexes. Oppenheimer and Rothschild2 first correlated these changes with a disseminated patchy sclerosis involving the subendocardial layer of the heart, and