Why Do Thiazides Decrease Polyuria in Diabetes Insipidus?

I was reviewing the treatment of diabetes insipidus the other day, and was reminded of the paradoxical effect of thiazide diuretics on urine output in diabetes insipidus. How does this work? The traditional thinking is that thiazide-induced blockade of the Na-Cl cotransporter in the distal tubule leads to a decrease in GFR. This decrease is compensated by an increase in proximal tubule sodium and water uptake. Because less water and solute are then delivered to the collecting duct, less water is lost as urine. However, some studies suggest that chronic use of thiazides does not result in a decrease in extracellular fluid volume: cardiac output returns to normal several weeks after initiating therapy, and infusion of salt-free dextran does not increase blood pressure. Studies in rats with central DI have also shown that replacement of renal sodium losses does not prevent the antidiuretic effect of thiazides. Experiments by Kim et al. suggest that thiazides may serve to upregulate aquaporin channels and ENaC subunits. In rates with lithium-induced nephrogenic DI, HCTZ reversed lithium-induced downregulation of AQP2. It also caused an increase in the abundance of ENaC channels. While these results are specific to Li-induced renal effects, they may at least partially explain how a thiazide can serve to decrease polyuria in patients with diabetes insipidus.