Sodium-lithium countertransport: does it matter how it is calculated?

On the basis of a preliminary observation that sodium-lithium countertransport showed different intra-individual variations when calculated by two standard methods despite using the same sets of lithium efflux data, we decided to compare values by the two methods in a much larger number (50) of volunteers. Although there was significant correlation between the two sets of values (r = 0.936), mean value by M-1 was significantly higher than that by M-2 (P < 0.007). Using appropriate statistical methodology, the limits of agreement between M-1 and M-2 values (-0.066 to 0.096 mmol L cell-1 h-1) were considerable; and the 95% confidence limits of the bias (i.e. mean difference) did not include zero. Furthermore, neither the 95% confidence interval for the correlation coefficient nor that for the slope of regression line included 1.0. These suggest lack of agreement between the two sets of values. The probable cause of this difference is discussed.