1. Urine osmolality and plasma and urine arginine vasopressin (AVP) were measured before, during and at the end of 23 h of water deprivation in four groups of subjects. There were eight non-polyuric manic-depressive patients taking lithium (the lithium group), seven manic-depressive patients taking other psychotropic drugs but not lithium (the other drug group), seven healthy subjects (the control group) and three lithium-treated patients with polyuria (the polyuric lithium group).
2. The lithium group had a resistance to the effect of AVP on the renal tubule with a fivefold increase in AVP excretion at a given urine osmolality. However, their water homoeostasis was intact as they lost no more weight during water deprivation than did the control group.
3. The relationship beween urine osmolality and AVP excretion has been defined from these results and the effect of lithium treatment on it has been characterized. The results suggest that lithium competitively inhibits the effect of AVP on the renal concentrating mechanism but does not decrease the ‘theoretical’ maximum urine osmolality.
4. The other drug group had a high basal urine flow which was due to a primary increase in water intake. The responsiveness of the renal concentrating mechanism to AVP was the same in this group as in the control group.
5. The polyuric lithium group had the highest basal urine flow and lost the greatest amount of weight during water deprivation. Neither urine osmolality nor AVP excretion was at a maximum in the basal state, as both increased during water deprivation. These relationships between urine osmolality and AVP excretion indicated a much greater resistance to AVP than in the other lithium-treated patients.
6. We suggest that these patients are polyuric as they become thirsty and drink before their AVP secretion increases to levels high enough to overcome the resistance to AVP. The results suggest, however, that in the basal state their water intake may be more than the minimum determined by the resistance to AVP.