Background: Saliva lithium (Li) determinations have been explored as a potential alternative to plasma levels. Unfortunately, this method has been shown to be unreliable. Saliva has two major components: the aqueous and the mucopolysaccharide portions.
Methods: Since Li is likely to distribute only in the aqueous fraction, we dialyzed the saliva through a 3000 Da filter to isolate the aqueous component and determine the Li level in it.
Results: Lithium levels in the dialyzed saliva agreed more closely with plasma levels (Spearman's rho = 0.901, p < 0.001) than did whole saliva (Spearman's rho = 0.775, p = 0.012).
Conclusion: Dialysis of saliva may contribute to more accurate saliva Li levels.