Potential pitfalls in the evaluation of the usefulness of hemodialysis for the removal of lithium.

A 50-year old female who was comatose from an overdose of lithium was treated with hemodialysis. Serum lithium concentrations declined 47% during a 3-h hemodialysis but increased afterwards, peaking 8 h after hemodialysis was stopped. Hemodialysis clearances were estimated by equations using extraction ratios of lithium from whole blood, serum, and red cells, and flows of whole blood, serum or red cells. The amount of lithium removed was calculated from these clearances as well as measured directly in the dialysate. Errors were introduced into the calculation of the amount of lithium removed by hemodialysis unless whole blood concentrations of lithium and whole blood flows were used. These arose because extraction of lithium from serum (0.7 +/- 0.3, mean +/- SD) was greater than that from whole blood (0.49 +/- 0.06) or from red blood cells (0.18 +/- 0.12). Despite the rapid decrease in serum concentrations of lithium during hemodialysis and rebound afterwards, the patient's neurologic status did not change concurrently. The patient did not regain consciousness until lithium concentrations fell to less than 0.4 meq/L in serum and 0.1 meq/L in cerebrospinal fluid. The lack of parallel change in serum concentrations and coma probably reflects the lag time in equilibration between lithium concentrations in serum and brain.