How thiazides work

The field was advanced significantly by a simple experiment in mice. What happens when thiazides are given to mice with NDI, when those mice lack the thiazide-sensitive NaCl co-transporter, NCC? Curiously, such mice still mount an anti-diuretic response to thiazides, suggesting that any effect is not mediated through the classic thiazide target. Thiazides are some of the oldest - and so presumably ‘dirtiest’ - drugs in clinical use; they are known to exert several off-target effects. One such - carbonic anhydrase inhibition - is likely responsible for the anti-diuretic effect in NDI. In the NCC-knockout experiment, thiazide treatment alkalinised the urine and reduced expression of NHE3, consistent with carbonic anhydrase inhibition and suppression of proximal tubular Na⁺ reabsorption.

The majority of filtered Li⁺ is reabsorbed in the proximal tubule, a fact well-known to physiologists who use fractional excretion of lithium as a proxy measure of proximal tubular sodium transport. So an effect here should have a number of consequences, all observed in mice: reduced blood levels of lithium, increased NaCl delivery to the macula densa (detected as reduced urinary PGE₂) and suppression of GFR by TGF. Without getting side-tracked too much, an important role for PGE₂ in regulating urinary water excretion emerged recently from a genome-wide assocation study in thiazide-associated hyponatraemia. In thiazide-associated hyponatraemia, enhanced luminal PGE₂ activity was thought to stimulate AQP2 expression and water reabsorption, exacerbating hyponatraemia. However, that is because susceptible patients carried a mutation in a prostaglandin transporter that prevented them from diverting PGE₂ to the basolateral side of the collecting duct, where it would ordinarily oppose AQP2 expression (i.e. favour a dilute urine). Therefore, the thiazide-induced reduction in PGE₂ in NDI might help to treat polyuria - unless of course Li²⁺ exposure changes prostaglandin transport in the collecting duct, which is entirely plausible.

The PGE₂ story notwithstanding, none of this should necessarily translate into reduced urine volumes, even if there are obvious clinical implications (e.g. for monitoring of Li⁺ levels). Instead, the effect on urine volume is likely to come from a direct effect of thiazides on principal cells, where they reduce Li⁺ uptake (through ENaC) and boost AQP2 expression. To my knowledge, the mechanism of this effect on principal cells has not been elucidated, and we can speculate that this might be due to CA inhibition or some PGE₂ effect.

Acetazolamide

If you need money, go to the bank. So why use a thiazide for its off-target effects when we could simply go for acetazolamide? Certainly this has been used effectively to limit polyuria in NDI. Like thiazides, in isolated cells and in mice, acetazolamide reduces Li⁺ uptake and boosts AQP expression. So, acetazolamide is a valid therapeutic option and perhaps the main reason to favour thiazides or amiloride is that we are generally more comforable with prescribing those agents.

Amiloride

There are also some clinical trial data for amiloride. In one tiny open-label trial (n = 9) and one RCT (n = 11), amiloride therapy resulted in a modest increase in urine osmolality. In the open-label trial, there was a reduction in 24hr urine volume, an outcome that was bizarrely not reported for the RCT.

One really attractive feature, is that amiloride blocks Li⁺ entry - through ENaC - into principal cells. Given that Li²⁺ uptake by principal cells causes collecting duct remodelling, loss of AQP2 expression and in fact the whole polyuria phenotype, amiloride strikes at the root cause of the problem. Of course we don’t know whether blocking Li²⁺ uptake by principal cells would also prevent progressive kidney fibrosis on Li²⁺ therapy - but we are highly unlikely to ever know that for certain, given the difficulties in performing clinical trials in this area.

Step 1) = get the basics right (in-patient)

The most important thing is to avoid any iatrogenic water deprivation. Classically, patients with NDI function very well when left to their own devices. They will drink copious amounts - often without even realising that they do so. (Ask the family and they will tell you that they keep a water bottle by their bed and carry one with them wherever they go.) Disaster only strikes when the patient is admitted to hospital and doctors and nurses try to take over the provision and prescription of water. Water is never provided in sufficient abundance and so the serum sodium is invariably 165 mM, day 2 post-appendicectomy. So in a hospitalised patient, the only really critical thing to do is to ensure adequate provision of drinking water - or NG water / IV glucose if the patient cannot drink.

Step 2) = get the basics right (out-patient)

In the well outpatient, things are a bit different; water supply should not be a problem. Instead the basics to get right are:

• in a patient still taking Li⁺, establish if the benefits of Li⁺ outweigh the risks (they usually do)
  
• check a serum calcium (Li⁺ can cause hyperPTH, which in turn could drive polyuria)
  
• ideally confirm and quantify any polyuria with a 24hr collection to measure urine volume (and possibly also solute excretion); this is perhaps most useful as a baseline measure before any intervention
  
• review diet to avoid excessive solute intake

Step 3) = the anti-diuretic

If polyuria is confirmed and it seems sensible to reduce urine volume (e.g. to improve quality of life), then it is reasonable to try an ‘anti-diuretic’. Factor in blood results (if tends to hypoK, favour amiloride; if tends to hyperK, favour thiazide), co-morbidities and likely effect on lithium levels. If still on lithium and unable to stop this, then there is an argument for amiloride, given its ability to block Li²⁺ uptake by principal cells. Review the benefits a few weeks after starting - ideally supported with an objective measure of urine volume. Remember that getting the basics right are much more important, so have a low threshold to bail out if the ‘anti-diuretic’ is causing problems.