Serum lithium. A CAP survey perspective.

Lithium data from the 1975 College of American Pathologists Comprehensive Chemistry Survey program have been analyzed. The results have been organized into five method-dependent groups: four variations on emission flame spectrophotometry, and one group containing all manual atomic absorption methods. The bias observed between flame emission and atomic absorption methods was correlated with the sample mean. In addition, three paired specimens were analyzed by mean comparison. This study indicates that atomic absorption methods have a negative bias, but an apparent correlation with the sample mean indicates the probability of standardization error between the methods. This may be related to a protein depression effect and may be due to differences in viscosity between aqueous standards and serum-based samples. Both atomic absorption and flame emission methods demonstrate variability that is clinically satisfactory.