Acetazolamide is the only carbonic anhydrase inhibitor with relevant diuretic effects. Acetazolamide binds tightly to carbonic anhydrase, and high concentrations are found in tissues containing this enzyme, particularly red blood cells and the renal cortex. Acetazolamide is mainly renally cleared as the unchanged molecule with approximately 50 % of its renal clearance as the result of active tubular secretion. Its administration is ordinarily accompanied by a brisk alkaline diuresis. Although carbonic anhydrase inhibitors are proximal tubular diuretics (where the bulk of Na⁺ reabsorption occurs), their net diuretic effect is modest, since Na⁺ reabsorption in more distal nephron segments offsets proximal Na⁺ losses. Acetazolamide use is constrained by both its transient action and the development of metabolic acidosis with prolonged use. Acetazolamide (250–500 mg daily) can correct the metabolic alkalosis that derives from chloride (Cl⁻) loss with thiazide and/or loop diuretic therapy and can be of some utility in patients with sleep apnea [5, 6].

Loop diuretics act predominately at the apical membrane in the thick ascending limb (TAL) of the loop of Henle, where they compete with Cl⁻ for binding to the Na⁺/K⁺/2Cl⁻ cotransporter, thereby inhibiting Na⁺ and Cl⁻ reabsorption [7]. Loop diuretics also have effects on Na⁺ reabsorption within other nephron segments, which are qualitatively minor compared with their action at the TAL. Other clinically important effects of loop diuretics include a decrease in both free water (H₂O) excretion and absorption during H₂O loading and dehydration, respectively, about a 30 % increase in fractional calcium (Ca⁺⁺) excretion, a significant increase in magnesium (Mg⁺⁺) excretion, and a transient increase followed by an ultimate decrease in uric acid excretion [8]. The increase in urine Ca⁺⁺ excretion with loop diuretic therapy increases parathyroid hormone release and therein a negative cardiac effect relating to the profibrotic effect of PTH on the myocardium [9].

The available loop diuretics include bumetanide, ethacrynic acid, furosemide, and torsemide. These compounds are heavily protein bound (to albumin), and therefore, in order to gain access to the tubular lumen (site of action), they must undergo secretion (the same applies to thiazide-type diuretics), which in their case is by way of probenecid-sensitive organic anion transporters localized to the proximal tubule [10, 11]. Tubular secretion of loop diuretics may be reduced in the company of elevated levels of endogenous organic acids, such as in chronic kidney disease (CKD), and by drugs that share the same transporter, such as salicylates and nonsteroidal anti-inflammatory drugs (NSAIDs). Loop diuretic protein binding can also be decreased by uremic toxins and/or fatty acids, which can in kind alter diuretic action.

Diuretic excretion rates approximate drug delivery to the medullary TAL and match up well with the observed natriuretic response [2, 4]. The relationship between urinary loop diuretic excretion rate and natriuretic effect is that of an S-shaped sigmoidal curve (Fig. 38.2) [11]. A normal dose-response relationship (as is typically seen in the untreated patient with hypertension) can be distorted (downward and rightward shifted) by a mixture of clinical conditions ranging from volume depletion (“braking phenomenon”) to HF or nephrotic syndrome (disease-state alterations) to various drug therapies. As an example of the latter, NSAIDs modify this dose-response relationship by inhibiting prostaglandin synthesis, blunting any expected diuretic effect. Finally, although the diuretic dose-response relationship can flatten in the setting of nephrotic-range proteinuria, the binding of loop diuretics to urinary protein and thereby limiting free drug access to receptor sites seems not to be the basis for this reduction in diuretic effect [12].

Diuretic-induced inhibition of Na⁺ reabsorption in one nephron segment elicits important adaptations in other nephron segments, which not only limits diuretics’ antihypertensive and fluid-depleting actions but also contributes to the magnitude of side effects. Although a portion of this resistance to diuretic effect is an expected consequence of their use, profound diuretic resistance from such adaptations can be encountered in patients with clinical disorders such as HF, cirrhosis, and/or proteinuric CKD. An understanding of how adaptation to diuretic therapy occurs is a prerequisite to minimizing the negative features of this process.

The initial dose of a diuretic ordinarily produces a brisk diuresis and in most instances results in a net negative Na⁺ balance. The new equilibrium state established with diuretic therapy is one where body weight decreases and with repetitive dosing stabilizes, since adaptive processes intervene and preclude a continuous volume loss. In nonedematous patients given either a thiazide or a loop diuretic, this adaptation or braking phenomenon occurs within a matter of days and limits weight loss to 1–2 kg; this has been demonstrated in normal subjects given the loop diuretics furosemide or bumetanide. Furosemide administered to subjects ingesting a high-Na⁺ diet (270 mmol/24 h) produces a brisk natriuresis, which results in a negative Na⁺ balance for the following 6 h. This is followed by a prolonged period of time up until the next dose, where Na⁺ excretion is reduced to a level considerably below that of intake [19]. This post-diuresis Na⁺ retention compensates for initial Na⁺ losses, with the result being no net weight loss. In fact, this same pattern of Na⁺ loss and compensatory retention can persist for as long as a month after furosemide administration [20]. Sodium intake, prior to and after diuretic dosing, will influence the end results of the braking phenomenon. For example, if Na⁺ intake is kept low, Na⁺ balance will remain negative in the hours after the initial natriuresis with a net fall in body weight.

Diuretic resistance is encountered in a number of disease states, such as CKD, nephrotic syndrome, HF, and cirrhosis with ascites, and most times relates to prominent pharmacodynamic alterations that are derived from disease-state organ dysfunction [11]. Combination diuretic use, which produces sequential nephron segment blockade and thus an additive diuretic response, is frequently necessary and is regularly employed in these conditions. Pharmacokinetic determinants of diuretic response, including the dose administered, absolute bioavailability, and tubular transport rate/capacity are pertinent considerations factors in how a diuretic response is generated [11]. Pharmacodynamic factors, such as improvement in the underlying disease-state condition promoting diuretic resistance, are perhaps more important to overall response and often result in modification of the dose-response relationship. Stratagems used to maximize the diuretic response to loop diuretics include correcting abnormal hemodynamic parameters, utilizing larger doses of well-absorbed orally administered loop diuretics, or using constant intravenous infusions or bolus diuretic therapy [21]. Bolus loop diuretic therapy does not appear to be much different than that of loop diuretic infusions in initiating and maintaining a diuretic response [22]. If these measures fail, then diuretic combinations are useful. Perhaps the most effective is the combination of metolazone and a loop diuretic. The rationale for and use of various diuretic combinations, with particular emphasis on the metolazone-loop diuretic combination, is well established but requires a deft touch in selection of medication doses in that over-diuresis is not uncommon [23].

The main site of action for thiazide-type diuretics is the early distal convoluted tubule where the coupled reabsorption of Na⁺ and Cl⁻ is inhibited (Fig. 38.1). Besides effects on Na⁺ excretion, thiazide diuretics also impair urinary diluting capacity (while preserving urinary concentrating mechanisms), reduce Ca ⁺⁺ and uric acid excretion, and are modestly magnesuric. The latter can be particularly prominent with long-acting thiazide-type diuretics, such as chlorthalidone [24].

There are two classes of K⁺-sparing diuretics: competitive antagonists of aldosterone, such as spironolactone and eplerenone, and compounds – such as amiloride and triamterene – which work in an aldosterone-independent manner. Drugs in this class inhibit active Na⁺ absorption in the late distal tubule and the collecting duct. In so doing, basolateral Na⁺– and K⁺-ATPase activity drops and intracellular K⁺ concentration is reduced. The resultant decrease in the electrochemical gradient for both K⁺ and H⁺ reduces secretion of these cations. Potassium-sparing diuretics also reduce Ca⁺⁺ and Mg⁺⁺ excretion and can be of some utility in patients prone to the development of calcium-containing stones [33]. Since K⁺-sparing diuretics are only modestly natriuretic, their clinical utility resides more in their K⁺-sparing properties, especially when more proximally acting diuretics increase distal Na⁺ delivery, or in states of primary aldosteronism. Potassium-sparing diuretics do, however, have a significant natriuretic effect when they are one of several components of a diuretic regimen, so-called sequential nephron blockade, and/or when used in suitably titrated doses in patients with cirrhosis/ascites [34].

The effect of a thiazide diuretic on BP may be divided into three successive phases: acute, subacute, and chronic, which correspond to periods of about 1–2 weeks, several weeks, and several months, respectively [51]. In the “acute” response phase, the BP-lowering effect of a diuretic is coupled to a reduction in ECF volume and a corresponding decrease in cardiac output. The early response (first 2–4 days of treatment) to a thiazide-type diuretic, in the setting of a “no-salt-added” diet (100–150 mmol/day), results in a net Na⁺ loss ranging from 100 to 300 mmol, which translates into a 1–2 L decrease in ECF volume. This decrease in plasma volume reduces venous return and diminishes cardiac output, the basis for the initial BP decrease with a thiazide diuretic. In due course the thiazide diuretic effects on volume and cardiac output lessen in importance, although BP remains lowered. During the subacute phase of a treatment response (first few weeks), plasma volume returns to slightly less than pretreatment levels, despite the continued administration of a diuretic. The subacute response phase with thiazide-type diuretics is a transitional period during which both volume and resistance factors both contribute to the observed BP reduction (Fig. 38.3).

In the chronic response phase of therapy, the BP reduction with a thiazide diuretic reverts to that of a reduction in total peripheral resistance (TPR). The decrease in TPR during prolonged therapy has been attributed to several factors including changes in the ionic content of vascular smooth muscle cells, altered ion gradients across smooth muscle cells and/or potassium channel activation, and changes in membrane-bound ATPase activity [28, 52]. The ability of thiazide-type diuretics to reduce BP seems to be linked to the presence of some level of renal function; thus, these drugs do not reduce BP in patients undergoing maintenance hemodialysis [25].

A mechanistic understanding of both diuretic action and the counterregulatory forces triggered by diuresis provide for a well-reasoned approach to the treatment of hypertension. The early action of diuretics to reduce ECF volume is optimized if dietary Na⁺ is restricted at the onset of therapy. This limits the repercussions of the braking phenomenon, which is an inevitable occurrence with uninterrupted diuretic use [11]. Some limitation in dietary Na⁺ intake may also be relevant to how diuretics might chronically reduce TPR. It is thought that intracellular Na⁺/Ca⁺⁺ content is favorably adjusted in vascular smooth muscle cells with the acute volume reduction observed during the first several days of thiazide diuretic therapy. How the development of volume contraction specifically translates into a reduction in TPR remains unclear [28, 52]. Whatever the mechanism, it can be quite long lived, because a residual BP reduction can be seen several weeks after the withdrawal of thiazide diuretics (even without interposing non-pharmacologic treatments for maintenance of BP control) [53]. This residual BP-reducing effect upon cessation of thiazide-type diuretics has not been specifically compared to that observed with non-diuretic antihypertensive drug classes in the same patient populations.

The concept of class effect has been applied to both loop diuretics and thiazide-type diuretics in respect to the management of hypertension. The loop diuretic effect on BP is a function of at least two processes: first, the manner in which volume removal is effected and, second, the capacity of these compounds to independently reduce total peripheral resistance. Small doses of the long-acting loop diuretic, torsemide, may cause significant BP reduction in essential hypertensives, a process independent of the level of diuresis, and one not seen with sub-diuretic doses of furosemide [54]. Intra-arterially infused furosemide does not directly dilate human forearm arterial vessels even at supratherapeutic concentrations [55]; in bioequivalent doses however, furosemide is just as effective as torsemide in reducing 24-h ambulatory BP in stage II–III CKD patients [56]. Until comparison studies amongst loop diuretics are conducted in diverse populations, it is premature to presume that these compounds are distinguishable, independent of volume removal, in their ability to reduce BP.

The idea of class effect for thiazide-type diuretics is still advanced by some, but it has minimal experimental support. Much of the recent debate on thiazide-type diuretic class effect has centered on the similarities and differences between chlorthalidone and HCTZ [57]. The concept of class effect with thiazide-type diuretics should be considered in two ways: first the effect on BP and second the event rate reduction. These two compounds are fundamentally different diuretics, in that chlorthalidone has a considerably longer duration of diuretic action than HCTZ [58].

One study evaluated chlorthalidone 25 mg/day compared to HCTZ 50 mg/day. In spite of the lower dose of chlorthalidone, clinic BP was reduced to a similar degree as the higher dose of HCTZ. Notably, nighttime BP was significantly lower with chlorthalidone, which is likely related to its longer duration of action [58]. A recent meta-analysis further compared the effects of chlorthalidone and HCTZ on systolic BP and serum K⁺ and found that chlorthalidone had a greater effect on BP with similar effects on serum K⁺ at equivalent doses [59]. As to the issues of outcomes, chlorthalidone has been used in several of the major clinical trials in the United States and has had a more consistent pattern of favorable outcomes than is the case with HCTZ [60–62]. These more favorable outcomes with chlorthalidone could be due to a greater reduction in nighttime BP and/or a reduction in the early morning surge in BP [29].