Nephrogenic diabetes insipidus and carbonic anhydrase inhibitors

Nephrogenic diabetes insipidus (NDI) is incredibly difficult and frustrating to treat. If the cause of NDI is loop diuretics, hypokalemia or hypercalcemia, it is a trivial problem, however in patients with congenital NDI or lithium induced NDI the treatment options are limited and pretty ineffective. The standard game plan is administering thiazide-type diuretics paired with a low salt/low solute diet. Here is the page from The Companion:

I think the first question is one of my favorites in the entire book.
Recently there has been some noise that carbonic anhydrase inhibitors maybe the way to go in lithium induced NDI.
Lithum causes NDI in two stages. In the first stage, lithum induces a down-regulation of the aquaporin-2 channels. The second stage comes from loss of the aquaporin-2 expressing principal cells of the collecting duct. To compensate there is an increase of alpha-intercalated cells.
The theory that HCTZ improves NDI through volume deficiency first took a hit in 2014 when Anne Sinke, and company showed that mice with lithium induced NDI but without the thiazide sensitive sodium-chloride co-transporter (NCCT) could still reduce urine volume in response to thiazide diuretics.
Let that sink in for a moment…mice that do not have the thiazide sensitive NCCT respond to thiazides as expected in regards to lithium induced NDI. This finding required an explanation and the authors suggested that this could be due to the carbonic anhydrase activity of thiazide diuretics. 
Yes, I know that thiazides reliably cause metabolic alkalosis, but despite that they have some minimal carbonic anhydrase activity that is uncovered in the NCCT knockout mice. According to this site, thiazides were initially designed to be carbonic anhydrase inhibitors, and then found to have NCCT antagonism activity. (Fun fact if you ask google about the thiazide’s effect on carbonic anhydrase the top hit is from David Goldfarb (@Weddellite) back from his basic science days.)
Back to the carbonic anhydrase story. Carbonic anhydrase is critical for bicarbonate reabsorption in the proximal tubule because it allows hydrogen secreted by the Na-H exchanger to combine with filtered bicarbonate to form CO2 and water. The CO2 is able to freely able to enter the proximal tubule cell where it combines with a hydroxyl group to reform bicarbonate. 
The authors concluded that by blocking carbonic anhydrase, there was less sodium reabsorption in the proximal tubule leading to activation of tubular glomerular feedback lowering GFR.
This lead to a second series of experiments where the same team reproduced the same reduction in urine volume but this time with acetazolamide, rather than HCTZ. The authors also thought there was a second mechanism based on acetazolamide’s ability to decrease lithium entry into principal cells of the collecting duct (similar to amiloride) and/or a reduction in intra-renal prostaglandin synthesis.
Fast forward to November 2016 and we have an interesting letter in the NEJM regarding a human patient with NDI that was admitted for a craniotomy. Post-operatively he developed recalcitrant hypernatremia with severe polyuria rising to 12 liters a day. 

…we administered acetazolamide to a patient with severe lithium-induced nephrogenic diabetes insipidus that was resistant to hydrochlorothiazide treatment and other measures…

The team used 500 mg of acetazolamide bid to take his urine output from ridiculous to merely absurd. Compelling data.