Mechanism of action of diuretics

INTRODUCTION — Natriuretic diuretics are among the most commonly used drugs. They act by diminishing sodium reabsorption at different sites in the nephron, thereby increasing urinary sodium and water losses. A second class of diuretics, sometimes termed aquaretics, instead inhibit water reabsorption by blocking vasopressin receptors along the connecting tubule and collecting duct.

The ability to induce negative fluid balance has made diuretics useful in the treatment of a variety of conditions, particularly edematous states and hypertension.

A review of the mechanism and time course of action of diuretics will be presented here. Diuretic dosing and adverse effects are discussed separately:

●(See "Loop diuretics: Dosing and major side effects".)

●(See "Time course of loop and thiazide diuretic-induced electrolyte complications".)

MECHANISM OF ACTION — The natriuretic diuretics are generally divided into four major classes, which are distinguished by the site at which they impair sodium reabsorption [1,2]:

●Loop diuretics act in the thick ascending limb of the loop of Henle

●Thiazide-type diuretics in the distal convoluted tubule

●Potassium-sparing diuretics in the aldosterone-sensitive distal nephron (including the connecting tubule and collecting duct)

●Acetazolamide and mannitol act at least in part in the proximal tubule

To appreciate how diuretics act, it is first necessary to review the general mechanism by which sodium is reabsorbed. Each of the sodium-transporting cells contains Na-K-ATPase pumps in the basolateral membrane [3]. These pumps perform three major functions: They return reabsorbed sodium to the systemic circulation, they maintain the cell sodium concentration at relatively low levels, and they raise cell potassium concentration above its electrochemical equilibrium, thereby maintaining polarization of the inside of the cell with respect to the outside. The effect on intracellular sodium concentration is particularly important since it allows filtered sodium to enter the cells down a favorable concentration gradient via carrier-mediated transport.

This sodium entry process at the apical membrane must be mediated by a transmembrane carrier or a sodium channel since charged particles cannot freely cross the lipid bilayer of the cell membrane. Each of the major nephron segments has one or more unique sodium entry mechanisms and the ability to specifically inhibit this step explains the nephron segment at which the different classes of diuretics act.

The site of action within the nephron is one determinant of diuretic efficacy. Most of the filtered sodium is reabsorbed in the proximal tubule (approximately 60 to 65 percent) and the loop of Henle (20 percent). As a result, it might be expected that a proximally acting diuretic, such as the carbonic anhydrase inhibitor acetazolamide, could induce relatively large losses of sodium and water. However, this does not occur, since almost all of the excess fluid delivered out of the proximal tubule can be reabsorbed more distally, particularly in the loop of Henle and to a lesser degree the distal tubule. Transport in these segments is primarily flow dependent, varying directly with the delivery of chloride [4,5].

A similar process of distal compensation occurs with the administration of loop diuretics, as some of the extra sodium chloride leaving the loop of Henle is reabsorbed in the distal tubule. With chronic loop diuretic therapy, animal studies have demonstrated both distal tubular hypertrophy and a rise in Na-K-ATPase activity in distal tubular cells [6-8]. A similar increase in distal tubular sodium reabsorption appears to occur in humans as demonstrated by an enhanced diuresis following administration of a thiazide diuretic after chronic therapy with a loop diuretic compared to placebo (figure 1) [9]. There is now evidence from lithium clearance studies that this process contributes to diuretic resistance in patients [10]. However, the reabsorptive capacity of the distal and collecting tubules is relatively limited and, in most circumstances, the natriuretic response to a loop diuretic is not seriously impaired [2].

Loop diuretics — When administered at maximum dose, the loop diuretics, furosemide, bumetanide, torsemide, and ethacrynic acid, can lead to the excretion of up to 20 to 25 percent of filtered sodium [1,11]. They act in the medullary and cortical aspects of the thick ascending limb, including the macula densa. At each of these sites, sodium entry is primarily mediated by a Na-K-2Cl carrier isoform 2 in the luminal membrane (NKCC2, encoded by SLC12A1) that is activated when all four sites are occupied (figure 2) [1,5,12,13]. It has been postulated that loop diuretics bind a chloride site on the carrier [13,14], and contemporary studies using cryoelectron microscopy have largely validated this view. According to the so-called "glide symmetry" model, the NKCC first binds a Na⁺, then a Cl⁻, followed by a K⁺ and a second Cl⁻. Binding of the second Cl⁻ seems to be interrupted by loop diuretics, owing to competitive inhibition [15]. Inhibition of NKCC1, encoded by SLC12A2, in the inner ear, is thought to be responsible for the ototoxicity that is seen occasionally with high-dose intravenous loop diuretic therapy. (See "Causes and treatment of refractory edema in adults".)

Loop diuretics also have important effects on renal calcium handling. The reabsorption of calcium in the loop of Henle is primarily passive, being driven by the electrochemical gradient created by NaCl transport and occurring through the paracellular pathway [16,17]. As a result, inhibiting the reabsorption of NaCl leads to a parallel reduction in that of calcium, thereby increasing calcium excretion. In the past, saline plus high doses of furosemide was a mainstay of the treatment of hypercalcemia. Volume expansion is still important, but other drugs, such as bisphosphonates, have largely replaced loop diuretics in this setting, except in hypervolemic patients. (See "Treatment of hypercalcemia".)

A potential concern is that the calciuric response to loop diuretics might lead to kidney stones and/or nephrocalcinosis. These complications have been primarily reported in premature infants in whom a loop diuretic can induce more than a 10-fold rise in calcium excretion [18,19]. Patients with Bartter syndrome frequently develop nephrocalcinosis, resulting from this phenomenon. (See "Inherited hypokalemic salt-losing tubulopathies: Pathophysiology and overview of clinical manifestations" and "Bartter and Gitelman syndromes in children: Clinical manifestations, diagnosis, and management".)

Thiazide diuretics — The thiazide diuretics primarily inhibit sodium transport in the distal convoluted tubule [1,2,20-22], although they may also have more modest effects along the proximal tubule and cortical collecting tubule [23-25]. The distal convoluted tubule reabsorbs a smaller proportion of the filtered load than the loop of Henle; as a result, the thiazide-type diuretics have a smaller natriuretic effect than loop diuretics and, when given in maximum dose, inhibit the reabsorption of at most 3 to 5 percent of filtered sodium [1,2]. Furthermore, the net diuresis may be partially limited by increased reabsorption in the connecting tubule and cortical collecting duct [7,26]. These characteristics make the thiazides less useful in the treatment of edematous states (unless given in combination with a loop diuretic for resistant edema), but are not a problem in uncomplicated hypertension where marked fluid loss is neither necessary nor desirable. (See "Use of thiazide diuretics in patients with primary (essential) hypertension" and "Causes and treatment of refractory edema in adults".)

In the distal convoluted tubule, an apical Na-Cl cotransporter (NCC, encoded by SLC12A3) is responsible for most sodium transport (figure 3) [1,20,27]. This transport protein is a member of the cation chloride cotransporter family and is structurally homologous to the NKCC proteins. Like loop diuretics, thiazides were believed to inhibit NaCl reabsorption by NCC by competing for a chloride site on the transporter [15,28,29]. Some thiazide-type drugs (such as chlorothiazide) also modestly impair sodium transport in the proximal tubule, due to partial inhibition of carbonic anhydrase [21,30]. This does not normally contribute to the net diuresis, however, since the excess fluid delivered out of the proximal tubule is reclaimed in the loop of Henle [21]. In addition, thiazides may impair sodium reabsorption along the collecting duct in the setting of hypovolemia by inhibiting a sodium-dependent chloride bicarbonate exchanger [25].

As with the loop diuretics, the thiazides also affect calcium handling [31]. The distal tubule is a major site of active calcium reabsorption in the nephron, an effect that is independent of sodium transport [16]. Although the thiazides inhibit the reabsorption of sodium in this segment, they increase the reabsorption of calcium [32]. Thiazides also increase calcium reabsorption along the proximal tubule as a consequence of reducing extracellular fluid volume, which enhances proximal reabsorption. Both proximal and distal effects appear to be necessary for calcium excretion to decline [33,34]. The fall in calcium excretion can be useful in the treatment of recurrent kidney stones due to hypercalciuria. (See "Kidney stones in adults: Prevention of recurrent kidney stones", section on 'High urine calcium'.)

Thiazide diuretics also affect water handling and can cause hyponatremia. The distal convoluted tubule is a part of the diluting segment, and therefore blocking salt reabsorption there will impair urinary dilution. More importantly, these drugs cause mild extracellular fluid volume depletion that, by reducing tubule flow to the loop of Henle, will also limit free water excretion. Thiazides also stimulate the tubuloglomerular feedback mechanism, thereby reducing glomerular filtration rate [35], and also appear to have direct effects on water channel abundance along the connecting tubule and collecting duct [36]. (See "Diuretic-induced hyponatremia", section on 'Mechanisms of thiazide-induced hyponatremia'.)

Potassium-sparing diuretics — The five potassium-sparing diuretics, amiloride, triamterene, spironolactone, eplerenone, and finerenone, act in the principal cells in the connecting and collecting tubules [1]. Sodium entry in these segments occurs through aldosterone-sensitive sodium channels (figure 4) [37,38]. The reabsorption of cationic sodium without an anion creates a lumen-negative electrical gradient that then favors the secretion of potassium (through selective potassium channels) and hydrogen ions. Thus, inhibition of sodium reabsorption at this site can lead to hyperkalemia and metabolic acidosis due to the concurrent reductions in potassium and hydrogen ion excretion [1,2].

The potassium-sparing diuretics decrease sodium current by two different mechanisms [39,40]:

●Amiloride and triamterene are cations that directly block sodium channels but do not affect the mineralocorticoid receptor. Another cation, the antibiotic trimethoprim, can exert similar effects, especially when given in high doses, leading to hyperkalemia [41]. (See "Electrolyte disturbances with HIV infection".)

●Spironolactone and eplerenone competitively inhibit the mineralocorticoid receptor [39,40]. Eplerenone is better tolerated than spironolactone since it has greater specificity for the mineralocorticoid receptor, resulting in a lower incidence of endocrine side effects (eg, gynecomastia, menstrual abnormalities, impotence, and decreased libido) that are mediated by nonselective binding to estrogen and progesterone receptors. In patients treated with diuretics, mineralocorticoid receptor inhibitors may also act in the distal tubule, diminishing the number of Na-Cl cotransporters [42] indirectly, by raising the serum potassium concentration [43].

Spironolactone or eplerenone is specifically indicated in patients with primary aldosteronism or heart failure because blocking the mineralocorticoid receptor may reduce the adverse effects of excess aldosterone on the heart. These drugs are also preferred in patients with cirrhosis.

•(See "Secondary pharmacologic therapy for heart failure with reduced ejection fraction".)

•(See "Ascites in adults with cirrhosis: Initial therapy", section on 'Diuretic regimen'.)

•(See "Treatment of primary aldosteronism", section on 'Pharmacologic therapy'.)

Finerenone, a nonsteroidal mineralocorticoid receptor antagonist [44], can slow kidney function loss in patients with diabetic kidney disease. The incidence of hyperkalemia may be lower with this drug than with other mineralocorticoid receptor antagonists. (See "Treatment of diabetic kidney disease", section on 'Type 2 diabetes: Treat with additional kidney-protective therapy'.)

The potassium-sparing diuretics have relatively weak natriuretic activity, leading to the maximum excretion of only 1 to 2 percent of filtered sodium [1]. Thus, they are primarily used in combination with a loop or thiazide diuretic, occasionally to cause additional sodium excretion but more commonly to diminish the degree of potassium loss [1,2]. (See "Causes and treatment of refractory edema in adults".)  

Amiloride is also effective in the treatment of polyuria and polydipsia due to lithium-induced nephrogenic diabetes insipidus. The resistance to antidiuretic hormone in this disorder appears to result from lithium accumulation in the collecting tubule cells by movement through the sodium channels in the luminal membrane. Blocking these channels with amiloride partially reverses and may even prevent the concentrating defect, presumably by diminishing lithium entry into the tubular cells [45,46]. (See "Renal toxicity of lithium", section on 'Arginine vasopressin resistance (nephrogenic diabetes insipidus)'.)

Amiloride is better tolerated than triamterene. It can be given once a day and is associated with few side effects other than hyperkalemia. Triamterene, in comparison, is a potential nephrotoxin [47], possibly leading to crystalluria and cast formation (in up to one-half of patients) [48], and rarely to triamterene stones [49] or to acute kidney injury due either to intratubular crystal deposition or the concurrent use of a nonsteroidal antiinflammatory drug [50,51]. These stones, which are more likely to occur in patients with a prior history of stone disease, are faintly radiopaque; their formation is pH-independent, and they usually contain some calcium oxalate (although pure triamterene stones can occur) [49,52]. (See "Triamterene nephrotoxicity".)

Carbonic anhydrase inhibitors (acetazolamide) — Acetazolamide inhibits the activity of carbonic anhydrase, which plays an important role in proximal bicarbonate, sodium, and chloride reabsorption. As a result, this agent produces both NaCl and NaHCO3 loss [53,54]. The net diuresis, however, is relatively modest for two reasons:

●Most of the excess fluid delivered out of the proximal tubule is reclaimed in the more distal segments, particularly the loop of Henle.

●The diuretic action is progressively attenuated by the metabolic acidosis that results from the loss of bicarbonate in the urine.

The main indication for the use of acetazolamide as a diuretic is in edematous patients with metabolic alkalosis in whom loss of the excess bicarbonate in the urine will tend to restore acid-base balance [54]. This effect may be particularly important in patients with hypercapnic chronic lung disease in whom conventional diuretic therapy can produce metabolic alkalosis; the compensatory hypoventilation induced by the rise in arterial pH can exacerbate the hypoxemia and retard weaning from mechanical ventilation. (See "Treatment of metabolic alkalosis", section on 'Edematous states'.)

Osmotic diuretics (mannitol) — Mannitol is a nonreabsorbable sugar alcohol that acts as an osmotic diuretic, inhibiting sodium and water reabsorption in the proximal tubule and more importantly the loop of Henle [55,56]. As with loop diuretics, mannitol produces a relative water diuresis in which water is lost in excess of sodium and potassium [55].

Mannitol is not generally used in edematous states, since initial retention of the hypertonic mannitol can induce further volume expansion, which, in heart failure, might precipitate pulmonary edema. Mannitol can also produce a clinically important increase in the plasma osmolality by two different mechanisms (see "Complications of mannitol therapy"):

●The preferential water diuresis induced by the repeated administration of mannitol can, if the losses are not replaced, lead to a water deficit and hypernatremia [57].

●Hypertonic mannitol may be retained in patients with kidney failure, directly increasing the plasma osmolality. In this setting, water movement out of the cells down an osmotic gradient will lower the plasma sodium concentration by dilution [58]. This is an important condition to recognize since treatment must be aimed at the hyperosmolality, not the hyponatremia.

Vasopressin receptor antagonists — The diuretics described in the preceding sections increase sodium and water excretion. In contrast, the vasopressin receptor antagonists (also called aquaretics) inhibit the action of antidiuretic hormone (vasopressin), resulting in a selective water diuresis. These drugs are used for the treatment of hyponatremia, since water loss will raise the serum sodium concentration, and for the treatment of autosomal dominant polycystic kidney disease. (See "Overview of the treatment of hyponatremia in adults", section on 'Vasopressin receptor antagonists' and "Autosomal dominant polycystic kidney disease (ADPKD): Treatment".)

TIME COURSE OF DIURESIS — The therapeutic effectiveness of a diuretic is related to a number of factors, including its site of action, its bioavailability, its duration of action, and dietary salt intake (see "Loop diuretics: Dosing and major side effects"). As an example, a short-acting loop diuretic, such as furosemide, produces a significant natriuresis during the six-hour period that the diuretic is acting [59,60]. Sodium excretion then falls to very low levels during the remaining 18 hours of the day, because the associated volume depletion leads to the activation of sodium-retaining mechanisms. This phenomenon is typically called "postdiuretic NaCl retention."

The net result in patients on a high sodium intake is that there is no net sodium loss. In this setting, one or more of the following changes must be present to induce a negative sodium balance:

●The patient can be placed on a low-sodium diet, thereby minimizing the degree of sodium retention once the diuretic effect has dissipated [60]. This is the preferred method since it can also limit concurrent potassium losses [61].

●The diuretic can be given two or more times per day.

●The dose of the diuretic can be increased, although the larger initial diuresis may induce symptomatic hypovolemia.

Several factors contribute to the compensatory antinatriuresis following the institution of diuretic therapy. The initial fluid loss leads to activation of the renin-angiotensin-aldosterone and sympathetic nervous systems; angiotensin II, aldosterone, and norepinephrine can all promote tubular sodium reabsorption [62-64]. However, in a study of normal volunteers who were treated with furosemide, blocking both of these pathways with prazosin (an alpha-1-adrenergic blocker) and an angiotensin-converting enzyme (ACE) inhibitor did not prevent the secondary renal sodium retention [59]. In this setting in which both vasoconstrictor hormones were inhibited, there was a mean 13 mmHg fall in the systemic blood pressure. Hypotension, in the absence of neurohumoral activation, directly promotes sodium retention [59,65,66].

These observations permit a more complete understanding of the volume regulatory actions of angiotensin II and norepinephrine. In the presence of volume depletion, the combined vasoconstrictor and sodium-retaining effects of these hormones result in both maintenance of the systemic blood pressure and an appropriate fall in sodium excretion.

Reestablishment of the steady state — Even if a net diuresis is induced, the diuretic response at a given dose is short-lived as a new steady state is rapidly established in which sodium intake and output are again equal but the extracellular volume has fallen due to the initial period of negative sodium balance. In this setting, the diuretic-induced sodium losses are counterbalanced by several factors (this is frequently called "the braking phenomenon") [67]:

●Neurohumorally mediated increases in tubular reabsorption at nondiuretic sensitive sites, such as the proximal tubule (due to angiotensin II and to a lesser degree norepinephrine) and the collecting tubules (due to aldosterone) [67,68].

●Flow-mediated increases in tubular reabsorption distal to the site of action of the diuretic as distal sodium delivery is enhanced [2,9]. Administration of a loop diuretic, for example, leads to hypertrophy and increased Na-K-ATPase activity in both the distal and collecting tubules [6-8]. A thiazide diuretic, on the other hand, acts in the distal tubule and the distal adaptations are limited to the Na-reabsorbing cells in the collecting tubules [7,8,26].

●Diminished diuretic entry to the intratubular sites of action also may contribute if kidney perfusion becomes impaired [69].

The attainment and maintenance of the new steady state requires that both diuretic dose and sodium intake are relatively constant. This limitation on the net diuresis is physiologically appropriate since progressive volume depletion and shock would eventually ensue if urinary sodium excretion were persistently greater than intake.

What is generally underappreciated is how rapidly the steady state is reestablished. This issue was addressed in a study in which three normal subjects on a constant sodium and potassium intake were treated with hydrochlorothiazide at a dose of 100 mg/day [70]. Sodium was lost for only three days and potassium for six to nine days; after this period, intake and output were again equal, but the early losses persisted (figure 5). Other studies in stable patients have shown that the maximum diuresis occurs after the first dose of the diuretic. As soon as fluid loss occurs, activation of sodium-retaining mechanisms will limit the response to the second dose (figure 6) [71]. Similarly, the response to a continuous intravenous diuretic infusion attenuates over time on the first day (figure 6).

A similar course in which there is a limited duration of diuresis at a constant diuretic dose occurs in edematous states, such as heart failure and cirrhosis. In heart failure, for example, the diuretic-induced reduction in cardiac filling pressures leads to a decline in cardiac output and activation of the renin-angiotensin system, both of which promote sodium retention [72]. (See "Use of diuretics in patients with heart failure".)

These findings are important clinically for patients treated with loop diuretics. As long as dose and dietary intake are stable, fluid and electrolyte complications associated with loop diuretics tend to occur within the first four weeks of therapy. Unlike thiazides, hyponatremia is not an early complication of loop diuretics. (See "Time course of loop and thiazide diuretic-induced electrolyte complications".)

Sequential evaluation of patients with hypertension has revealed that all of the fall in the plasma potassium concentration following therapy with a thiazide diuretic occurs within the first two to four weeks, with subsequent stabilization at the new level [73]. Similar considerations apply to the use of a potassium-sparing diuretic to correct thiazide-induced hypokalemia; the plasma potassium concentration rises during the first two to three weeks and then remains relatively constant [74].

The sequence is different in patients who are markedly volume expanded due to renal sodium retention (eg, in acute glomerulonephritis). In this setting, the renin-angiotensin system is suppressed and will not be activated by initial sodium loss, since hypervolemia persists. Thus, the second and subsequent doses may produce as large a natriuresis as the original dose until most of the excess fluid has been removed. Even in this setting, however, the first dose still represents the maximum response that will be seen.

SUMMARY

●Diuretics diminish sodium reabsorption at different sites in the nephron, thereby increasing urinary sodium and water losses, at least over the short term. The four major classes of diuretics, which are distinguished by the site at which they act, include loop, thiazide-type, and potassium-sparing diuretics, and diuretics that act in the proximal tubule such as acetazolamide and mannitol. (See 'Introduction' above and 'Mechanism of action' above.)

●Loop diuretics (furosemide, bumetanide, torsemide, and ethacrynic acid) decrease sodium reabsorption in the medullary and cortical thick ascending limb by inhibiting the Na-K-2Cl carrier in the luminal membrane. (See 'Loop diuretics' above.)

●Thiazide diuretics decrease sodium reabsorption in the distal tubule, by inhibiting the Na-Cl cotransporter. Thiazide-type diuretics have a smaller natriuretic effect than loop diuretics and the diuresis may be limited by increased sodium reabsorption in the cortical collecting tubule. (See 'Thiazide diuretics' above.)

●The five potassium-sparing diuretics (amiloride, triamterene, spironolactone, eplerenone, and finerenone) decrease the number of open sodium channels in the connecting tubule and collecting duct. Amiloride and triamterene directly inhibit sodium channel activity, while spironolactone, eplerenone, and finerenone block the mineralocorticoid receptor. (See 'Potassium-sparing diuretics' above.)

●Acetazolamide inhibits the activity of carbonic anhydrase, which plays an important role in proximal bicarbonate, sodium, and chloride reabsorption causing both NaCl and NaHCO3 loss. Mannitol is an osmotic diuretic and inhibits sodium and water reabsorption in the proximal tubule and loop of Henle. (See 'Carbonic anhydrase inhibitors (acetazolamide)' above and 'Osmotic diuretics (mannitol)' above.)

●The therapeutic effectiveness of a diuretic is related to its site of action, duration of action, and dietary salt intake. The diuretic response at a given dose is short-lived as a new steady state is rapidly established in which sodium intake and output are again equal but the extracellular volume has fallen due to the initial period of negative sodium balance. This antinatriuretic response is mediated by activation of the renin-angiotensin-aldosterone and sympathetic nervous systems, which promote tubular sodium reabsorption. (See 'Time course of diuresis' above.)

●Because a new steady state is rapidly established in stable patients, the fluid and electrolyte complications associated with a loop diuretic dose typically occur within the first four weeks of therapy. (See 'Reestablishment of the steady state' above.)