Carbonic anhydrase inhibitors (CAIs) act primarily on proximal tubule cells to inhibit bicarbonate absorption ( Fig. 49.2 ) with additional effects in the loop of Henle and along the distal nephron. ^, Carbonic anhydrase (CA) is a metalloenzyme that regulates sodium-bicarbonate reabsorption and hydrogen ion secretion by renal epithelial cells. ^,

Mechanisms of action of carbonic anhydrase inhibitors in the proximal tubule.

The figure shows a functional model of proximal tubule cells; many transport proteins are omitted from the model for clarity. Inside the cell, carbonic anhydrase (CA) catalyzes the formation of HCO ₃ ⁻ from OH ⁻ and CO ₂ . Bicarbonate leaves the cell via the sodium-bicarbonate transporter 1 (NBC1). A second pool of carbonic anhydrase is located in the brush border. This participates in disposing of the carbonic acid formed from filtered bicarbonate and secreted H ⁺ . Both pools of carbonic anhydrase are inhibited by acetazolamide and other carbonic anhydrase inhibitors ( CAIs ; see text for details).

CA is expressed by many tissues including erythrocytes, kidney, gut, ciliary body, choroid plexus, and glial cells. Although at least 14 isoforms of CA have been identified, two play predominant roles in renal acid-base homeostasis, CA II and CA IV. CA II is widely expressed in renal tubular cells, where it accounts for 95% of renal CA. It is present in proximal tubule cells and intercalated cells of the aldosterone-sensitive distal nephron (ASDN). CA IV is expressed at the luminal border of the cells of the proximal thick ascending limb (TAL) of the loop of Henle and α-intercalated cells of the ASDN.

The prototypic CAI is acetazolamide. However, many diuretics have some CAI action. ^, This characteristic contributes to the weak inhibition of proximal reabsorption by furosemide and chlorothiazide and to the relaxation of vascular smooth muscle cells by high-dose furosemide.

CAIs block the catalytic dehydration of luminal carbonic acid at the brush border of the proximal tubule, decrease the intracellular generation of H ⁺ required for counter transport with Na ⁺ , and decrease the peritubular capillary fluid uptake. CAIs also are weak inhibitors of reabsorption in the TAL, but the natriuretic efficacy of CAIs and loop diuretics (see later) is additive, confirming their independent mechanisms of action. CAIs also inhibit bicarbonate reabsorption along the distal tubule, presumably by interfering with the action of α-intercalated cells. The first administration of a CAI causes a brisk alkaline diuresis. The excretion of Na ⁺ , K ⁺ , HCO ₃ ⁻ , and PO ₄ ²⁻ increases, whereas the excretion of titratable acid and NH ₄ ⁺ decreases sharply. Excretion of Ca ²⁺ remains essentially unchanged. There is substantial kaliuresis owing to the presence of nonreabsorbable HCO ₃ ⁻ and high flow rates in the distal nephron. However, hypokalemia is uncommon because acidosis partitions K ⁺ out of cells. Frank hypokalemia is only reported with very-high-dose CAIs (as used in intracranial hypertension) used in combination with thiazides. Bicarbonaturia is reduced over time as the serum bicarbonate concentration declines, thereby reducing the filtered load of bicarbonate. Moreover, the tubular fluid bicarbonate concentrations are increased downstream of the proximal tubule, providing a favorable gradient for passive paracellular reabsorption.

Long-term CAI administration causes only a modest natriuresis, despite the magnitude of CA-dependent proximal Na ⁺ reabsorption. Several factors account for this:

• 1.

  CA is required for the reabsorption of HCO ₃ ⁻ , whereas about two-thirds of the proximal Na ⁺ reabsorption is accompanied by Cl ⁻ .

• 2.

  Some proximal HCO ₃ ⁻ reabsorption persists, even after apparently full inhibition of CA.

• 3.

  Some of the HCO ₃ ⁻ that is delivered out of the proximal tubule can be reabsorbed in the loop of Henle and more distal sites.

• 4.

  The metabolic acidosis that develops limits the filtered load of HCO ₃ ⁻ and thereby curtails the natriuresis.

• 5.

  The increased delivery of filtered Na ⁺ to the macula densa elicits a tubuloglomerular feedback (TGF)–induced reduction in the glomerular filtration rate (GFR).

Micropuncture studies of mice with deletion of the proximal Na ⁺ -H ⁺ exchanger, NHE3, have shown that inhibition of proximal Na reabsorption is largely balanced by reduced GFR, supporting mechanism #5. Although CAIs only cause a modest natriuresis, the combined use of a thiazide and acetazolamide results in a brisk natriuresis, perhaps because acetazolamide inhibits the HCO ₃ ⁻ /Cl ⁻ exchanger pendrin that is upregulated secondary to upstream inhibition of the sodium chloride cotransporter (Na ⁺ -Cl ⁻ cotransporter [NCC]) by thiazide diuretics. ^,

Acetazolamide is readily absorbed. It is eliminated with a half-life (t _1/2 ) of 13 hours by tubular secretion, which is diminished during hypoalbuminemia. Methazolamide has less plasma protein binding, a longer t _1/2 , and greater lipid solubility, all of which favor penetration into aqueous humor and cerebrospinal fluid. This agent has less renal effect and therefore is used to treat glaucoma.

The use of CAIs as diuretics is limited by their transient action, the development of metabolic acidosis, and a spectrum of adverse effects. They can be used with NaHCO ₃ infusion to initiate an alkaline diuresis that increases the excretion of weakly acidic drugs (e.g., salicylates and phenobarbital) or acidic metabolites (e.g., urate). Chloride depletion metabolic alkalosis is best treated by the administration of Cl ⁻ with K ⁺ or Na ⁺ . However, if this produces unacceptable extracellular volume (ECV) expansion, acetazolamide (250–500 mg/day) and KCl can be used to increase HCO ₃ ⁻ excretion.

Metabolic alkalosis due to loop diuretics or thiazides can depress respiration in patients with chronic respiratory acidosis (e.g., due to chronic obstructive pulmonary disease). This provides the rationale for the administration of acetazolamide to such subjects that can reduce the arterial partial pressure of carbon dioxide (Pa co ₂ ) and improve the partial pressure of oxygen (Pa o ₂ ). Because both Pa co ₂ and the plasma bicarbonate concentration (P hco ₃ ) decrease, there is little change in blood pH, but a reduction in P hco ₃ limits the buffering capacity of blood. CAIs can increase the Pa co ₂ during metabolic acidosis or exercise, perhaps by depressing hypoxic ventilatory drive and hypoxic pulmonary vasoconstriction, and can cause ventilation-perfusion imbalance. Nevertheless, acetazolamide (250 mg bid) can improve blood gas parameters in patients with chronic obstructive pulmonary disease. Careful surveillance is required when CAIs are administered to such patients.

When used to treat glaucoma, CAIs diminish the transport of HCO ₃ ⁻ and Na ⁺ by the ciliary process, thereby reducing intraocular pressure. CAIs also limit the formation of cerebrospinal fluid and endolymph. Currently, their use to treat glaucoma is mainly as topical preparations. On the other hand, oral acetazolamide is still widely used in the management of idiopathic intracranial hypertension (IIH; also sometimes still referred to as pseudotumor cerebri ).

Acute mountain sickness is characterized by headache, nausea, drowsiness, insomnia, shortness of breath, dizziness, and malaise after an abrupt ascent. Acetazolamide is useful in dosages of 250 to 750 mg daily as prophylaxis against mountain sickness, probably through stimulating respiration and diminishing cerebral blood flow and cerebrospinal fluid formation. Acetazolamide, used in established mountain sickness, improves oxygenation and pulmonary gas exchange and stimulates ventilation in patients with central sleep apnea. In patients with IIH, acetazolamide administration with a low-sodium weight-reduction diet modestly improves visual field function.

The Acetazolamide in Decompensated Heart Failure with Volume Overload (ADVOR) trial randomized patients with acute decompensated heart failure (ADHF) to add-on therapy with intravenous acetazolamide (500 mg daily for 3 days) versus placebo during intravenous loop diuretic administration. It reported a significant ( P < 0.001) 46% improvement in decongestion over 3 days. While encouraging, the effects of longer-term administration were not studied.

CAIs are effective in the prophylaxis of hypokalemic periodic paralysis because they diminish the influx of K ⁺ into cells. Paradoxically, they are also useful in the treatment of hyperkalemic periodic paralysis.

Studies in mice showed that acetazolamide attenuates lithium-induced nephrogenic diabetes insipidus, with fewer adverse effects than thiazides and amiloride. This effect was ascribed to a reduction in the GFR and activation of TGF but may also indicate a specific effect in the collecting duct. Clinical data are needed to assess the corresponding effect in humans.

Other indications for CAIs are experimental. They include possible applications in diseases as diverse as obesity, cancer, and infection.

Patients taking CAIs may complain of weakness, lethargy, abnormal taste, paresthesia, gastrointestinal distress, malaise, and decreased libido. These symptoms can be diminished by NaHCO ₃ , but this agent increases the risk of nephrocalcinosis and nephrolithiasis. Overall, symptomatic metabolic acidosis develops in 50% of patients with glaucoma treated with oral CAIs.

Older patients or those with diabetes mellitus or chronic kidney disease (CKD) can experience serious metabolic acidosis if given a CAI. An alkaline urine favors partitioning of renal ammonia into blood rather than its elimination in urine. Therefore in patients with liver failure, an increase in blood ammonia may precipitate encephalopathy.

Acetazolamide increases the risk of nephrolithiasis by more than 10-fold. CAIs occasionally cause allergic reactions, hepatitis, and blood dyscrasias. They can cause osteomalacia when used with phenytoin or phenobarbital. While kaliuresis is common, frank hypokalemia is rare.

Osmotic diuretics are freely filtered but poorly reabsorbed. Mannitol is the prototypic osmotic diuretic, although sorbitol and glycerol have similar actions. In the water-permeable nephron segments of the proximal nephron and thin limbs of the loop of Henle, fluid reabsorption concentrates filtered mannitol sufficiently to diminish tubular fluid reabsorption. Ongoing Na ⁺ reabsorption lowers the tubular fluid (Na ⁺ ) and creates a gradient for back-flux of reabsorbed Na ⁺ into the tubule. Increased distal flow stimulates K ⁺ secretion.

Mannitol is a hypertonic solute that abstracts water from cells. The increase in total renal blood flow (RBF) relates in part to hemodilution and a decrease in hematocrit and viscosity. Mannitol increases the medullary blood flow and decreases the medullary solute gradient, thereby preventing urinary concentration. The rise in renal plasma flow and drop in plasma colloid osmotic pressure can increase the GFR.

Mannitol is distributed exclusively to the extracellular fluid. It is filtered freely at the glomerulus. Consequently, the t _1/2 for plasma clearance of mannitol depends on the GFR and is prolonged from 1 to 36 hours in those with advanced CKD. It can be infused intravenously in daily doses of 50 to 200 g as a 15% or 20% solution or 1.5 to 2.0 g/kg of 20% mannitol over 30 to 60 minutes to treat elevated intraocular or intracranial pressure.

Mannitol has been evaluated for the prophylaxis of AKI, but controlled trials of its use in patients at risk for AKI have not had positive results. However, mannitol can protect against AKI in cadaveric kidney transplant recipients. The use of diuretics to convert oliguric to nonoliguric AKI is discussed later (see “Clinical Uses of Diuretics”).

A trial of mannitol therapy for cerebral edema complicating hepatic failure has demonstrated a markedly better survival of 47%, compared with only 6% in the control group. Mannitol is recommended for the management of severe head injury with cerebral edema. ^, It is more effective than loop diuretics or hypertonic saline in reducing brain water content. Mannitol can reverse the dialysis disequilibrium syndrome.

The effects of mannitol on plasma electrolyte concentrations are complex. The osmotic abstraction of cell water initially causes hypertonic hyponatremia and hypochloremia. Later, after the excess extracellular fluid (ECF) is excreted, the decrease in cell water concentrates K ⁺ and H ⁺ within cells, thereby increasing the gradient for their diffusion into the ECF and leading to hyperkalemic acidosis. Normally, these electrolyte changes are corrected rapidly by the kidney if renal function is adequate. Later, hypernatremic dehydration may develop if free water is not provided because urinary concentrating ability is inhibited.

Expansion of ECV, hemodilution, and hyperkalemic metabolic acidosis occur in patients with renal failure who cannot eliminate the drug. Circulatory overload, pulmonary edema, central nervous system depression, and severe hyponatremia require urgent hemodialysis. Doses of more than 200 g/day can cause renal vasoconstriction and AKI.

The primary action of loop diuretics occurs from the luminal aspect of the TAL ( Fig. 49.3 ). An electroneutral Na-K-2Cl cotransporter, termed “NKCC2,” is located at the luminal membrane. ^, This cotransporter, a member of the solute carrier family 12 (gene symbol SLC12A1 ), mediates Na ⁺ and Cl ⁻ movement across the cell. A high luminal K ⁺ conductance, via the renal outer medullary K ⁺ (ROMK) channel, allows most of the K ⁺ to recycle across the luminal membrane. This effectively converts NKCC2 into an electrogenic NCC transporter that is coupled with the electrogenic exit of Cl ⁻ across the basolateral membrane. Thus NKCC2 generates a transepithelial voltage, oriented positively with the lumen relative to interstitial fluid. The primary energy for transport across TAL cells is provided via the basolateral Na ⁺ -K ⁺ –adenosine triphosphatase (ATPase), which maintains a low intracellular (Na ⁺ ). Additional details concerning mechanisms of solute reabsorption by TAL cells can be found in Chapter 6 . Loop diuretics are organic anions that bind to the NKCC2 from the luminal surface.

Mechanisms of diuretic action along the loop of Henle.

A model of thick ascending limb cells. Na ⁺ and Cl ⁻ are reabsorbed across the apical membrane via the loop diuretic–sensitive Na-K-2Cl cotransporter 2 (NKCC2). Loop diuretics bind to and block this pathway directly. Note that the transepithelial voltage along the thick ascending limb is oriented with the lumen positive relative to blood ( circled value, given in millivolts [mV]). This transepithelial voltage drives a component of Na ⁺ (and calcium and magnesium; see Fig. 49.4) reabsorption via the paracellular pathway. This component of Na ⁺ absorption is also reduced by loop diuretics because they reduce the transepithelial voltage.

ClC-KB, Chloride channel protein; ROMK, renal outer medullary K ⁺ channel; TAL, thick ascending limb.

NKCC2 is expressed on the apical membranes of medullary and cortical TALs and macula densa segments. ^, Its abundance is increased by prolonged infusion of saline or furosemide. A closely related gene, NKCCl, encodes a protein that is widely expressed in transporting epithelia. In contrast to NKCC2, NKCCl is implicated in the uptake and secretion of Cl ⁻ and NH ₄ ⁺ at the basolateral membrane of the medullary CDs.

Hormones that stimulate cyclic adenosine monophosphate (cAMP), such as arginine vasopressin (AVP), enhance TAL reabsorption and should enhance the response to loop diuretics. In contrast, those that stimulate cyclic guanosine monophosphate (cGMP), such as nitric oxide and atrial natriuretic peptide (ANP), those that increase intracellular (Ca ²⁺ ), such as 20-hydroxyeicosatetraenoic acid (20-HETE), or those that activate the Ca ²⁺ (polyvalent cation)–sensing protein inhibit TAL reabsorption and reduce the response to loop diuretics.

The rat TAL also transports NH ₄ ⁺ , which can substitute for K ⁺ on NKCC2. There is a luminal Na ⁺ -H ⁺ countertransporter in the rat nephron that contributes to tubular fluid acidification. Loop diuretics block the luminal entry of Na ⁺ via NKCC2, but not the peritubular exit via the Na ⁺ -K ⁺ -ATPase, and thereby reduce the intracellular (Na ⁺ ) sufficiently to promote luminal Na ⁺ uptake via the Na ⁺ -H ⁺ countertransport process. This is one reason that furosemide stimulates acid excretion in the rat. ^, However, furosemide did not affect net acid excretion or urine pH in normal human subjects.

Loop diuretics reduce proximal fluid reabsorption modestly. This effect has been ascribed to a weak CAI action. However, furosemide depresses proximal reabsorption in tubules perfused with HCO ₃ ⁻ -free solutions. Moreover, bumetanide, which is a much less potent inhibitor of CA, also impairs proximal fluid reabsorption.

Furosemide exerts two contrasting effects on reabsorption in the superficial distal tubule. Increased delivery to the unsaturated distal tubule reabsorption process increases Na ⁺ reabsorption. Loop diuretics also inhibit NaCl transport in short descending limbs of the loop of Henle, and although the TAL is clearly the major site of action of loop diuretics, some studies suggest other nephron segments may make minor contributions. ^, Reabsorption of solute from the water-impermeable TAL segments dilutes the tubular fluid. Its inhibition by loop diuretics impairs free water excretion during water loading and free water reabsorption during dehydration.

Loop diuretics increase the fractional excretion of Ca ²⁺ by up to 30%. The predominant mechanism is a decrease in the magnitude of the lumen-positive transepithelial potential (see Figs. 49.3 and 49.4 ). A large fraction of transepithelial Ca ²⁺ transport along the TAL traverses a paracellular pathway involving claudins 16 and 19 and is driven by the lumen-positive transepithelial potential. By reducing the magnitude of this potential, loop diuretics lower passive calcium absorption along this segment.

Possible mechanisms of diuretic effects on calcium and magnesium excretion.

Typical cells from the proximal tubule (PT) , thick ascending limb (TAL) , and distal convoluted tubule (DCT) are shown. Calcium reabsorption occurs along the DCT largely via a transient receptor potential channel (TRPV5) and magnesium reabsorption via a transient receptor potential channel (TRPM6). Transepithelial voltages (representative but arbitrary values, given in millivolts [mV]) are shown. Net effects on electrolyte excretion are shown at the bottom . Normal conditions are at the left . Treatment with loop diuretics (LDs) is shown in the middle ; treatment with DCT diuretics is shown on the right . Loop diuretics reduce the magnitude of the lumen-positive transepithelial voltage, thereby retarding passive calcium and magnesium reabsorption via the paracellular pathway. Long-term treatment, especially with DCT diuretics, increases proximal Na ⁺ and Ca ²⁺ reabsorption; thus less calcium is delivered distally. Enhanced distal calcium absorption, driven by DCT diuretics, also occurs. Effects of DCT diuretics to increase magnesium excretion remain incompletely understood.

ClC-KB, Chloride channel protein; NCC, Na-Cl transporter; NKCC2, Na-K-2Cl cotransporter 2; ROMK, renal outer medullary K ⁺ channel; ↑︎, increase(d); ↓︎, decrease(d).

The loop of Henle is a major nephron segment for the reabsorption of Mg ²⁺ . Mg ²⁺ also traverses a paracellular pathway that involves claudins and is driven by the transepithelial potential difference. Loop diuretics can increase fractional Mg ²⁺ excretion by more than 60% (see Figs. 49.3 and 49.4 and see later discussion of adverse effects). Loop diuretics reduce urate clearance that is largely secondary to volume depletion. There is a reduction in RBF and GFR of rats infused acutely with furosemide that depends on a steep rise in intrarenal pressure that can collapse capillaries and also on a reflex effect from renal afferent nerve activation. Likewise, human subjects given loop diuretics experience a 20% fall in GFR in the first day. However, during more prolonged administration there is normally a well-maintained RBF and GFR but a marked redistribution of blood flow from the medulla and inner cortex to the outer cortex. The fall in papillary plasma flow depends on angiotensin II. Furosemide increases the renal generation of prostaglandins. Blockade of cyclooxygenase prevents acute furosemide-induced renal vasodilation.

Reabsorption via NKCC2 on the luminal surface of the macula densa cells activates the TGF but inhibits renin secretion. An increase in NaCl delivery to the macula densa as during a reduction in proximal tubular fluid reabsorption increases NaCl entry into macula densa cells. This increases the production of adenosine that interacts with adenosine 1 receptors on vascular smooth muscle and/or extraglomerular mesangial cells to activating phospholipase C. The ensuing cellular depolarization activates voltage-dependent Ca ²⁺ channels, which contract afferent arterioles and reduce the GFR (and thereby activate the TGF response). Loop diuretics block NaCl entry into macula densa cells and thereby block TGF completely. This can explain the initial fall in GFR with drugs such as CAIs or SGLT2 inhibitors that reduce proximal reabsorption and thereby enhance delivery to the macula densa or the initial rise in RBF observed in some acute studies of loop diuretics in human subjects. However, furosemide exacerbates impaired renal function and oxygenation in a rat model of AKI. An analysis of data from the Renal Optimization Strategies Evaluation Acute Heart Failure (ROSE) trial of subjects with ADHF receiving a mean daily intravenous dose of furosemide of 560 mg revealed that renal function deteriorated significantly in 21% yet there was no increase in markers of tubular damage. This was attributed to a “beneficial” or “physiologic” reduction in GFR in response to successful therapy with ACEIs or SGLT2 is rather than an adverse effect of furosemide. Moreover, furosemide may reduce RBF and thereby O ₂ delivery, yet this can be offset by reduced NaCl reabsorption and thereby O ₂ usage. Indeed, furosemide increases renal medullary oxygenation in normal human subjects. Clearly, the causes for an increase in serum creatine during diuresis in clinical practice can be complicated but do not necessarily portend bad outcomes, renal tubular damage, or adverse effects on renal hemodynamics.

Loop diuretics also stimulate renin secretion. Although due in part to ECV depletion, a major component is from direct effects of loop diuretics to inhibit NaCl reabsorption in the macula densa. Macula densa reabsorption inhibits renin secretion acutely and inhibits renin synthesis chronically. Macula densa cells express cyclooxygenase-2 (COX-2) constitutively. Addition of loop diuretics to fluid bathing a macula densa cell line acutely increases the release of prostaglandin E2 (PGE2), followed by a delayed induction of COX-2 expression and renin secretion mediated by a rapid phosphorylation of p44/42 and p38 MAP kinases following activation of the prostaglandin E2 type 4 (EP4) receptor. Consequently, blockade of COX2 >1 with celecoxib or nonsteroid antiinflammatory drugs (NSAIDs) inhibits renin secretion strongly.

Loop diuretics are absorbed promptly after ingestion, but their bioavailabilities vary. Because bumetanide and torsemide are more completely absorbed than furosemide, changing from intravenous to oral dosing may require a doubling of the dose of furosemide dose but does not require changing the dose of bumetanide or torsemide. Moreover, there is considerable variation in furosemide absorption, both between patients and over time, that is accentuated by food intake.

The highly variable bioavailability of furosemide compared with torsemide and the absence of hypokalemia with usual doses, combined with the superior outcomes of torsemide in two clinical trials of patients with heart failure (HF), ^, have prompted consideration that torsemide become the loop diuretic of choice, at least for HF. However, the Torsemide vs. Furosemide After Discharge in all-Cause Mortality in Patients Hospitalized with Heart Failure (TRANSFORM-HF) pragmatic trial reported no difference in one-year mortality or all-cause hospitalizations. The reasons for these discrepant findings are unexplained. Nevertheless, a class problem shared with all generic loop diuretics is their short duration of action of 2 to 4 hours. The resulting torrential diuresis (termed the “Niagara effect”) can be troubling, especially for patients with prostatism or stress incontinence. Moreover, the limited time of action of a daily dose of diuretic on the tubules provides some 20 hours for the kidneys to regain salt and water losses (see Fig. 49.12 ). An extended-release formulation of torsemide that delivers torsemide to the circulation over 8 to 12 hours doubled salt and water losses in normal volunteers in the day after a single dose, without increasing potassium excretion or reducing the creatinine clearance. Remarkably, one dose of the extended-release torsemide given to subjects ingesting 300 mmol of Na ⁺ daily caused a negative daily Na ⁺ balance, whereas prior studies with furosemide at this high level of salt intake reported no net loss of Na ⁺ because of postdiuretic salt retention. Extended-release torsemide is likely to be especially beneficial for the >50% of patients with CHF who have symptoms of overactive bladder or urgency incontinence.

Effects of sodium-glucose-linked transport 2 inhibition on proximal tubule and loop of Henle reabsorption.

A simplified functional model of the early (S1 and S2) and late (S3) segments of the proximal tubule and the thick ascending limb of loop of Henle. SGLT2 is linked functionally to the sodium hydrogen exchanger 3 and to the urate transporter (URAT 1) by the microtubule-associated protein 3 (MAP3). Inhibition of SGLT2 blocks NHE3, thereby delivering more sodium from the proximal tubule while also reducing glucose reabsorption, thereby increasing tubular fluid osmolality to create an osmotic diuresis. More glucose is delivered to SGLT1 in the S3 segment that is activated to reabsorb sodium coupled in a 2:1 ratio with glucose that thereby offsets the natriuresis. The fall in tubular fluid sodium concentration promotes passive sodium back leak from the interstitium to the tubular fluid. Increased delivery of sodium chloride to the thick ascending limb (TAL) of the loop of Henle enhances responsiveness to loop diuretics while increased delivery to the macula densa activates the tubuloglomerular feedback mechanism (TGF).

Once absorbed, loop diuretics circulate largely bound to albumin (91%–99%). This limits their clearance by glomerular filtration. The diuretic volume of distribution varies inversely with the serum albumin concentration, but this is not usually a major determinant of diuretic responsiveness (see later). The metabolism of loop diuretics is composed of both hepatic and renal mechanisms; the relative fractions that are cleared by each mechanism differ among agents. Loop diuretics, thiazides, and CAIs are all secreted avidly by a probenecid-sensitive organic anion transporter (OAT) in proximal tubule cells ( Fig. 49.5 ). ^, Diuretics gain access to tubular fluid almost exclusively by proximal secretion. Studies have characterized this weak organic anion (OA ⁻ ) transport process. Four isoforms of an OAT have been cloned and are expressed in the kidney. ^, Peritubular uptake by an OAT is a tertiary active process (see Fig. 49.5 ). Energy derives from the basolateral Na ⁺ -K ⁺ -ATPase providing a low intracellular [Na ⁺ ] that drives an uptake of Na ⁺ coupled to α-ketoglutarate (α-KG) to maintain a high intracellular level of α-KG. This, in turn, drives a basolateral OA–α-KG countertransporter. OAT1 is expressed on the basolateral membrane of the S2 segment of the proximal tubule. OAT1 knockout dramatically impairs renal OA secretion and is accompanied by furosemide resistance. A similar effect was observed in OAT3-deficient mice, suggesting that both OAT1 and OAT3 mediate the secretion of loop diuretic by proximal cells.

Mechanisms of diuretic secretion by proximal tubule cells.

Cell diagram of the S2 segment of the proximal tubule shows secretion of anionic diuretics including loop diuretics and distal convoluted tubule (DCT) diuretics. Peritubular uptake by an organic anion transporter (primarily OAT1, although OAT3 may play a smaller role) occurs in exchange for α-ketoglutarate, which is brought into the cell by the Na ⁺ -dependent cation transporter NaDC-3. Luminal secretion can occur via a voltage-dependent pathway or in exchange for luminal hydroxyl or urate. A portion of the luminal transport traverses multidrug resistance–associated protein 4 (Mrp4) .

ATPase, Adenosine triphosphatase.

OATs translocate diuretics into the proximal tubule cell, where they can be sequestered in intracellular vesicles. They are secreted across the luminal membrane by a voltage-driven OA transporter and by a countertransporter in exchange for urate or OH ⁻ . The orphan transporter hNPT4 (human sodium phosphate transporter 4; SLC17A3) has been identified as an organic anion efflux transporter that likely also secretes furosemide and bumetanide. In addition, the multidrug resistance–associated protein 4 (MRP4) is a third transporter involved in the urinary excretion of diuretics. Approximately 50% of furosemide is eliminated by metabolism to the inactive glucuronide. Only the unmetabolized and secreted fraction is available to inhibit NaCl reabsorption. In contrast, bumetanide and torsemide are metabolized in the liver. ^, Unlike bumetanide or torsemide, the elimination of furosemide is reduced greatly in patients with CKD because its metabolism to the inactive glucuronide occurs predominantly in the kidney; in contrast, metabolic inactivation of bumetanide and torsemide in the liver are unaffected by uremia. ^, This difference prolongs the t _1/2 of furosemide in CKD, leading to drug accumulation. However, CKD increases the fraction of a dose of furosemide that is excreted unchanged, leading to an enhanced natriuretic response ( Fig. 49.6 ). Therefore there is a tradeoff in the selection of a loop diuretic in CKD: Furosemide can accumulate and cause ototoxicity at high doses, whereas bumetanide and torsemide retain their metabolic inactivation but are somewhat less potent.

(A−E) Comparison of the pharmacokinetics and dynamics of furosemide (F, 160 mg; metabolically inactivated in the kidney) and bumetanide (B, 4 mg; metabolically inactivated in the liver) in 10 subjects with chronic kidney disease (mean creatinine clearance, 12 ± 2 mL/min).

Significance of difference: ∗, P <.05; ∗∗∗, P <.005 . FE _Na , Fractional excretion of sodium; T _1/2 , half-life; U _Na , urinary sodium.

Redrawn from data in Voelker JR, Cartwright-Brown D, Anderson S, et al. Comparison of loop diuretics in patients with chronic renal insufficiency. Kidney Int. 1987;32:572−578.

Renal clearance of the active form of furosemide is reduced in CKD in proportion to the creatinine clearance. There is competition for both peritubular uptake and luminal secretion with other OAs including urate, hippurate, and other organic anions that accumulate in uremia. Metabolic acidosis depolarizes the membrane potential of proximal tubule cells, which decreases OA secretion, an effect that may explain why diuretic secretion is enhanced by alkalosis. Therefore the increased plasma levels of OAs and urate and the metabolic acidosis of CKD impair proximal tubule secretion of diuretics and hence impair their delivery to active sites in the nephron.

Despite its high plasma protein binding, the proximal secretion of active furosemide is potentiated by albumin. In the rabbit, an equal fraction of administered furosemide is taken up by probenecid-sensitive mechanisms in the S2 segment of the proximal tubule, where it is secreted into tubular fluid or by the S1 segment of the proximal tubule, where it is conjugated and excreted as the inactive glucuronide ( Fig. 49.7 ). Hypoalbuminemia inhibits the uptake and secretion of active furosemide by the S2 segment yet increases the uptake and metabolism by the S1 segment. Therefore it enhances furosemide metabolism and decreases the tubular secretion of the active diuretic. The consequences of this process are described later (see “Nephrotic Syndrome”).

Diagrammatic representation of the disposition of intravenous furosemide and the effects of hypoalbuminemia or probenecid in normal or hypoalbuminemic rabbits.

After intravenous administration of furosemide, 15% is metabolized by uridine diphosphate glucuronyl transferase (UDPGT) in the liver and gut to the inactive furosemide glucuronide (F-GC) . Of the remainder, 85% is transported by the kidney. Some 42% is taken up in the S1 segment of the proximal tubule (PT-S ₁ ) and metabolized to the inactive glucuronide, and the remainder is taken up by the S2 segment (PT-S ₂ ) and secreted in active form into the lumen. Both uptake processes are inhibited by probenecid. The plasma albumin concentration facilitates uptake and secretion by PT-S ₂ but inhibits uptake and metabolism by PT-S ₁ .

Modified from Pichette V, Geadah D, du Souich P. The influence of moderate hypoalbuminemia on the renal metabolism and dynamics of furosemide in the rabbit. Br J Pharmacol. 1996;119:885−890.

There is a sigmoidal relation between fractional sodium excretion and the log of the urinary diuretic concentration that quantitates the amount of diuretic delivered to its active site on the tubular lumen ( Fig. 49.8 ). Inhibition of proximal secretion with probenecid shifts the curve of the plasma dose-response to the right. Thus natriuresis is related to the urinary, but not the plasma concentration, of diuretic. The administration of indomethacin or other nonsteroidal antiinflammatory drugs (NSAIDs) reduces the responsiveness of the tubule to furosemide. This reduction is due predominantly to reduced generation of PGE2 because a natriuretic response to furosemide can be restored in indomethacin-treated rats by the infusion of PGE2. A reduced dietary salt intake diminishes the renal tubular response to furosemide (see Fig. 49.8 ).

Relationship between the excretion of Na ⁺ and furosemide (log scale) following a bolus intravenous injection of 40 mg furosemide in normal subjects with a normal NaCl intake (1), with a normal NaCl intake after indomethacin (2), with a low Na ⁺ intake (20 mmol/24 hours) (3), and for the third day of furosemide administration with a low Na ⁺ intake (4).

Modified from Wilcox CS, Mitch WE, Kelly RA, et al. Response of the kidney to furosemide. J Lab Clin Med. 1983;102:450−458; and Chennavasin P, Seiwell R, Brater DC. Pharmacokinetic-dynamic analysis of the indomethacin-furosemide interaction in man. J Pharmacol Exp Ther. 1980;215:77−81.

These are discussed later (see “Clinical Uses of Diuretics” and “Adverse Effects of Diuretics”).

Thiazides and thiazide-like diuretics are moderately active natriuretic drugs that increase the excretion of sodium, chloride, and potassium while reducing the excretion of calcium. Their major site of action is the distal convoluted tubule (DCT), where they block coupled reabsorption of Na ⁺ and Cl ⁻ ( Fig. 49.9 ). ^{, ,} The true thiazides (benzothiadiazines) include chlorothiazide, hydrochlorothiazide, and bendroflumethiazide, whereas metolazone and chlorthalidone have similar actions but distinct molecular structures. Chlorthalidone is also an inhibitor of CA activity and has a prolonged half-life. Due to these properties, it is used widely as an antihypertensive.

Mechanisms of distal convoluted tubule (DCT) and collecting duct (CD) diuretics.

(A) Mechanism of action of DCT diuretics. In rat, mouse, and humans, two types of DCT cells have been identified, referred to here as DCT1 and DCT2. Na ⁺ and Cl ⁻ are reabsorbed across the apical membrane of DCT1 cells only via the thiazide-sensitive Na ⁺ -Cl ⁻ cotransporter (NCC) . This transport protein is also expressed by DCT2 cells where Na ⁺ can also cross through the epithelial Na ⁺ channel (ENaC; see text for details). Thus the transepithelial voltage along the DCT1 is near to 0 mV, whereas it is finite and lumen negative along the DCT2. (B) Mechanism of action of CD diuretics. The late distal convoluted tubule cells (DCT2 cells) and connecting tubule (CNT) or cortical collecting duct (CCD) cells are shown. Na ⁺ is reabsorbed via ENaC, which lies in parallel with a renal outer medullary K ⁺ channel (ROMK). The transepithelial voltage is oriented with the lumen negative, relative to the interstitium (shown in the circled value ), generating a favorable gradient for transepithelial K ⁺ secretion. Drugs that block the epithelial Na ⁺ channel reduce the voltage toward 0 mV (effect indicated by dashed line ), thereby inhibiting K ⁺ and H ⁺ secretion.

ClC-KB, Chloride channel protein; KCC4, Potassium-chloride cotransporter 4.

The predominant effect of the thiazide diuretics is to inhibit NCC (gene symbol SLC12A3 ) in the DCT ^{, ,} with additional natriuretic effects from a variable degree of inhibition of CA. Patients with Gitelman syndrome, who have a loss-of-function mutation in the NCC, and mice lacking NCC demonstrate a dramatically impaired natriuretic response to thiazides, confirming that the principal effect of these drugs is to inhibit NCC. Na ⁺ reabsorption by the proximal tubule is enhanced during long-term treatment with thiazides.

Thiazides are organic anions that bind to the NCC transport protein from the luminal surface. Inhibition has been studied using two approaches. The avid binding of [ ³ H]-metolazone and [ ³ H]-bumetanide to kidney membrane proteins is inhibited competitively by Cl ⁻ , suggesting that Cl ⁻ and diuretics compete for the same binding site. Findings have delineated the crystal structure of NCC, allowing for assessment of the 3D binding sites for ions and thiazides. These studies demonstrate that the thiazide binding site overlaps the proposed chloride binding site.

Thiazides increase potassium excretion indirectly ^, from hyperaldosteronism, increased distal flow, and increased CNT/CCD sodium delivery. They reduce the clearance of urate secondary to enhanced proximal reabsorption from ECV depletion and competition for uptake of urate from blood to PT cells. Thiazides reduce Ca ²⁺ excretion by three nonredundant mechanisms (see Fig. 49.4 ). ^, First, blockade of luminal NaCl entry reduces the tubule cell intracellular [Na ⁺ ] sufficiently to enhance basolateral Na ⁺ -Ca ²⁺ exchange. Second, thiazide-induced blockade of luminal NaCl entry reduces cell Cl ⁻ concentration, thereby hyperpolarizing the membrane voltage that increases calcium entry via the transient receptor potential channel subfamily V, member 5 (TRPV5) channel. ^, Third, thiazides stimulate proximal reabsorption of Ca ²⁺ secondary to ECV depletion, which accounts for a continuing effect of thiazides to reduce Ca ²⁺ excretion in mice with TRPV5 deleted. However, mice deficient in the DCT-specific protein parvalbumin fail to increase sodium excretion with hydrochlorothiazide but still develop hypocalciuria. Similarly, hypocalciuria occurs independent of hypovolemia in patients with Gitelman syndrome, who have a genetic inactivation of NCC. Thiazides cause a sustained reduction in renal Ca ²⁺ excretion that is accompanied by a small rise in serum Ca ²⁺ concentration.

Mg ²⁺ excretion is enhanced by prolonged thiazide therapy (see Fig. 49.4 ). Long-term thiazide treatment of mice diminishes expression of the transient receptor potential channel melastatin 6 (TRPM6) substantially. This would be expected to reduce Mg ²⁺ reabsorption. ^, The Mg ²⁺ depletion that can occur during chronic thiazide administration may augment K ⁺ depletion.

The aquaporin 2 (AQP2) water channel is at the terminal diluting segment of the kidney in the connecting tubule. Thiazides impair solute excretion and thereby maximal urinary dilution but not maximal urinary concentration. However, they also enhance water absorption from inner medullary collecting ducts in an AVP-independent manner that is accompanied by an increase in AQP2 expression. These effects summate with central effects on thirst and enhanced production of PGE2 to induce hyponatremia (see later discussion of adverse effects).

Thiazides are readily absorbed and extensively bound to plasma proteins. They are eliminated largely through secretion by the S2 segment of the proximal tubule, mostly via OAT1 and OAT3. ^, The t _1/2 is prolonged in renal failure and in older adults, accompanied by a reduced natriuretic efficacy. However, chlorthalidone remains effective in CKD IV. The more lipid-soluble drugs (e.g., bendroflumethiazide and polythiazide) are more potent, have a more prolonged action, and are more extensively metabolized. Chlorthalidone has a particularly prolonged action. Indapamide is sufficiently metabolized to limit accumulation in renal failure. Extrarenal effects of thiazide diuretics including effects on platelet aggregation and vascular permeability vary among the types of thiazide diuretics. While thiazide-like diuretics (e.g., chlorthalidone and indapamide) were believed to have superior cardioprotective effects compared with thiazide-type diuretics (e.g., chlorothiazide and hydrochlorothiazide), the large VA diuretic comparison project that contrasted outcomes for cardiovascular disease (CVD) in hypertensives randomized to hydrochlorothiazide versus chlorthalidone reported no differences and a higher rate of hypokalemia with chlorthalidone.

While there is little known about the pharmacogenetics of diuretics, available studies have suggested that this may account for some differences in individual responses. Diuretic therapy in carriers of a variant of α-adducin, a cytoskeleton protein important for the function of renal Na ⁺ -K ⁺ -ATPase, was associated with a lower risk of combined myocardial infarction and stroke than other antihypertensive therapies. Polymorphisms in with-no-lysine kinase 1 (WNK1), which is involved in the regulation of NCC, also affected the response to a thiazide diuretic. Finally, polymorphisms in Nedd4-2, a ubiquitin ligase that regulates NCC, predict the blood pressure (BP) response to hydrochlorothiazide in White subjects and cardiovascular outcome with hydrochlorothiazide in Black subjects.

The clinical indications for using these drugs are discussed later (see “Clinical Uses of Diuretics”), as are their adverse effects (see “Adverse Effects of Diuretics”).

Distal K ⁺ -sparing diuretics act on the-ASDN comprising the late DCT, connecting tubule, and cortical CD, where they inhibit luminal Na ⁺ entry via ENaC (see Fig. 49.9 ). ^, They depolarize the lumen-negative transepithelial voltage, thereby diminishing the electrochemical gradient for K ⁺ and H ⁺ secretion and increasing the serum potassium concentration (S _K ) and reducing the serum bicarbonate concentration (S _HCO3 ^– ).

Both amiloride and triamterene are organic cations that block ENaC directly from the luminal surface. Amiloride also inhibits NHE3, but its affinity for NHE3 is sufficiently low that the distal effects predominate. Amiloride appears to bind ENaC, where it competes with Na ⁺ predominantly at a short amino acid stretch within the extracellular loop. Amiloride and triamterene also reduce the excretion of Ca ²⁺ and Mg ²⁺ . ^,

Spironolactone and eplerenone are antagonists of the mineralocorticoid receptor (MRAs) that were developed on the basis of the steroidal structure of aldosterone and shown to block the MR competitively. Eplerenone was developed as an MRA with fewer side effects from blockade of androgen receptors. Finerenone was developed as a nonsteroidal, dihydropyridine-based MR antagonist.

These drugs were used for many years primarily in combination with other diuretics to reduce the excretion of K ⁺ and net acid. Later, it was found that long-term administration of spironolactone was more effective than furosemide in reducing cirrhotic ascites. The PATHWAY-2 trial randomly crossed over patients with uncontrolled hypertension despite taking three antihypertensives including a diuretic through four 12-week periods of add-on therapy with spironolactone (25–50 mg daily), the β-adrenoreceptor antagonist bisoprolol (5–10 mg), the α-adrenoreceptor antagonist doxazosin in a modified release formulation (4–8 mg) or placebo. The placebo-corrected fall in systolic blood pressure (SBP) with spironolactone that averaged 8.7 mm Hg was twice as great as for the other drugs. Spironolactone is now recommended as the first choice of add-on therapy for resistant hypertension. The Randomized Aldactone Evaluation Study (RALES) reported that add-on therapy with spironolactone for 2 years for patients with HF with reduced ejection fraction (HFrEF) lessened the risk of death or hospitalization for HF significantly by >30%. Generally, similar conclusions were drawn from trials of spironolactone in patients with HF with preserved ejection fraction (HFpEF) or of eplerenone in HFrEF. Thus MRAs have become drugs of first choice for the treatment of resistant hypertension and HF, but the extent to which these effects are due to their diuretic action remains unclear.

Triamterene is well absorbed. It is hydroxylated rapidly to active metabolites. The drug and its metabolites are secreted by the organic cation pathway in the proximal tubule and eliminated with half-lives of 3 to 5 hours. Triamterene and its active metabolites accumulate in patients with cirrhosis because of decreased biliary secretion and in older adults ^, and patients with CKD because of decreased renal excretion.

Amiloride is incompletely absorbed. Its duration of action is approximately 18 hours. It is secreted into the tubular fluid by the organic cation transport pathway. Other organic cations, such as cimetidine, inhibit its secretion and prolong its half-life. It accumulates in renal failure and may worsen renal function. Spironolactone is readily absorbed and circulates bound to plasma proteins. Its intrinsic half-life is short, but it is metabolized to active compounds with considerably prolonged actions. Spironolactone is metabolized to canrenones (t _1/2 = 16 hours) and to sulfur-containing metabolites, predominantly 7-α-thiomethylspirolactone (t _1/2 = 13 hours). Canrenones are metabolized by the cytochrome P-4503A system. This accounts for an effective t _1/2 for spironolactone of approximately 20 hours and the need for only once-daily dosing. It takes 10 to 48 hours to become maximally effective. It is lipid soluble and enters distal renal tubules from the plasma. Eplerenone has fewer antiandrogenic and proestrogenic effects. ^, However, it is metabolized with a half-life of 3 hours and therefore should be given twice daily. Finerenone is highly selective for MR and has a similarly short half-life as eplerenone.

Distal K ⁺ -sparing agents are used to prevent or treat the hypokalemic alkalosis that can arise from thiazide or loop diuretic therapy. Spironolactone is indicated as a first-line agent for ECV expansion in the setting of cirrhotic ascites, ^{, ,} where it is used in combination with furosemide at a dose ratio of spironolactone 100 mg/furosemide 40 mg. It is also indicated for HF (in doses of 25 and 50 mg/day), to treat hypertension associated with hyperaldosteronism and for resistant hypertension. It reduces proteinuria and progressive loss of kidney function in CKD. ^, Amiloride reduces proteinuria to an equivalent degree (however, see the following discussion of adverse effects). Eplerenone is indicated to prevent cardiac remodeling and systolic dysfunction in the setting of recent myocardial infarction. ^, Finerenone is indicated for treatment of diabetic kidney disease. It slows the progression of CKD and improves cardiovascular outcomes in this patient group. ^,

Hyperkalemia is the most common complication of the distal K ⁺ -sparing diuretics and MRAs. The risk is dose dependent and increases considerably in patients with CKD or in those receiving K ⁺ supplements, angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), NSAIDs, β-blockers, heparin, or ketoconazole. The incidence of hyperkalemia-associated morbidity and mortality in Canada rose sharply after the publication of the RALES trial. While finerenone does increase the risk of hyperkalemia, it may do so less frequently than spironolactone.

Gynecomastia and reduced libido may occur in men treated with spironolactone, especially as the dose is increased. ^, Women may experience menstrual irregularities, hirsutism, or swelling and tenderness of the breast. Impaired net acid excretion can cause metabolic acidosis, which worsens hyperkalemia.

Amiloride and triamterene accumulate in renal failure, ^, and triamterene accumulates in cirrhosis. Therefore these drugs should be used with caution in patients with these conditions. Triamterene occasionally precipitates in the urinary collecting system and causes obstruction. It can cause AKI when given with indomethacin.

Dopamine infused into normal subjects in low doses (1−3 μg/kg per minute) causes a modest increase in the GFR, reduces proximal reabsorption via a cAMP-induced inhibition of the Na ⁺ -H ⁺ antiporter, and increases Na ⁺ excretion. Fenoldopam is a selective dopamine type 1 receptor agonist, with little cardiac stimulation. Unfortunately, these beneficial effects are lost in patients who are critically ill and/or receiving vasopressors. A comprehensive review of controlled trials concluded that low-dose dopamine universally fails to improve renal outcomes in patients at high risk for AKI and that it had no effect on renal function, need for dialysis, or mortality in critically ill patients with early renal dysfunction. Thus there is currently no justification for the use of low-dose dopamine for renal protection. Dopamine infusion at higher rates has a role as a pressor agent in septic shock or refractory HF, but the benefits can be offset by arrhythmias.

Vasopressin receptor (VR) antagonists are nonpeptide molecules that competitively inhibit one or more of the human VRs, V1aR, V1bR, and V2R. Conivaptan is a combined V1aR/V2R antagonist for intravenous use, whereas tolvaptan, mozavaptan, and lixivaptan are orally active V2R-selective antagonists. All these agents cause a free water diuresis without appreciable natriuresis or kaliuresis (“aquaretics.)” This effect is mainly attributed to the inhibition of V2R in the collecting duct, which prevents vasopressin from recruiting AQP2 water channels to increase water reabsorption. Therefore vasopressin receptor antagonists can be used to treat hypervolemic or euvolemic hyponatremia. Coinhibition of V1aR, which is located in vascular smooth muscle, could be beneficial to reduce coronary vasoconstriction, myocyte hypertrophy, and vascular resistance in patients with HF, but definitive studies are lacking. The results of many clinical trials in patients with liver cirrhosis, HF, or hyponatremia secondary to SIADH report that VR antagonists effectively raised serum sodium levels (S _Na ) and help to correct hyponatremia with reduction in body weight, dyspnea, and ascites in patients with HF. However, the Efficacy of Vasopressin Antagonist in Heart Failure Outcome Study with Tolvaptan (EVEREST) trial failed to show a beneficial effect of tolvaptan on the primary outcome of death and rehospitalization. Thus V2R antagonists can effectively correct hyponatremia but have not yet shown a benefit in primary outcomes for HF.

Neprilysin inhibitors reduce the metabolism of natriuretic peptides and therefore contribute to natriuresis. The combined use of an ARB and neprilysin inhibitor reduced cardiovascular death and hospitalization for HF in patients with class II, III, or IV HF.

Several novel diuretics are being developed targeting urea channels, ROMK, or pendrin. A urea channel inhibitor UT-B (SLC14A1) induced osmotic diuresis in rats by inhibiting urea reabsorption in the renal medulla without commensurate urinary sodium and potassium loss. Dimethylthiourea inhibits the urea channels UT-B and UT-A1. It produced a greater diuresis in rats than furosemide while maintaining S _Na . It prevented hyponatremia in a model of SIADH. ROMK inhibitors induce a robust natriuresis in rats that exceeds that of a maximal dose of furosemide but does not cause kaliuresis. A ROMK inhibitor caused greater diuresis and natriuresis when combined with hydrochlorothiazide or amiloride but not bumetanide and blocks the TGF response, confirming that the diuretic target is predominantly the TAL and MD. Pendrin is a Cl ⁻ /HCO ₃ ⁻ exchanger expressed in the luminal membrane of intercalated cells involved in regulation of Na ⁺ reabsorption and acid-base balance. Small-molecule inhibitors of pendrin are not effective diuretics when given alone but enhance furosemide diuresis by 30% to 60%. Thus these drugs could become valuable to address loop diuretic resistance.

Diuretic therapy increases plasma renin activity (PRA) and serum aldosterone concentration (SAC). The initial rise in PRA with loop diuretics is independent of volume depletion or the SNS and is related to the inhibition of NaCl reabsorption at the macula densa, perhaps following prostacyclin and prostaglandin E2 (PGE2) release. During longer-term use of loop diuretics, the increases in PRA are augmented by ECV depletion, a fall in blood pressure and activation of the SNS.

Activation of the RAAS in patients treated with diuretics and salt restriction for edema can limit the natriuresis. ACE inhibition potentiates the diuretic and natriuretic responses to furosemide in HF, despite a drop in BP. However, severe volume depletion and azotemia can complicate overzealous combined therapy with ACE inhibitors and high doses of diuretics or in those with stenosis of both renal arteries or the artery to a single or dominant kidney. Thus the combination of diuretics and ACE inhibitors can be effective but requires careful surveillance.

Diuretic-induced increases in SAC promote renal K ⁺ losses that can be countered by ACE inhibitors and MRAs.

PGE2 acting on EP ₄ Rs inhibits NaCl reabsorption via NKCC2 and free water and Na ⁺ reabsorption in the CDs via changes in cAMP. ^, Loop diuretics, thiazides, triamterene, and spironolactone increase renal prostaglandin production substantially. Inhibition of PG synthesis by NSAIDs can diminish the natriuresis and diuresis induced by furosemide, hydrochlorothiazide, spironolactone, or triamterene (see Fig. 49.12 ). Microperfusion of the loop segment with PGE2 restores the response to furosemide in indomethacin-treated rats. Indomethacin also blunts furosemide-induced renal and capacitance vessel vasodilation and renin release. NSAIDs blunt furosemide-induced natriuresis while a COX-2 inhibitor blocks furosemide-induced renin secretion. Natriuresis is facilitated by COX1 in the distal nephron. Thus COX-1 products mediate a part of furosemide-induced natriuresis, whereas COX-2 products mediate much of the renin secretion.

Arginine vasopressin (AVP) increases after the administration of furosemide in response to volume depletion. Plasma AVP levels are increased in many edematous states, such as HF and liver cirrhosis, but are decreased in patients who develop hyponatremia during thiazide treatment. AVP stimulates K ⁺ secretion in the rat distal tubule. Diuretic-induced AVP release contributes to hypokalemia because the kaliuretic response to furosemide is reduced by 40% in subjects whose AVP release is suppressed by a water load. AVP can increase the activity of both NCC and ENaC, thereby possibly aggravating sodium retention in edematous states. ^,

While a dose of furosemide raises the heart rate and plasma catecholamine concentrations, ^, blockade of α ₁ -adrenergic receptors with prazosin does not modify the ensuing renal salt retention. However, blockade of B1 adrenoreceptors does blunt renin release. Short-term, furosemide-induced ECV depletion in the conscious rat activates the SNS that stabilizes the BP.

A loop diuretic and a thiazide or thiazide-like drug (e.g., hydrochlorothiazide and metolazone) are synergistic in normal subjects and subjects with edema or renal insufficiency. Metolazone is equivalent to bendrofluazide in enhancing NaCl and fluid losses in furosemide-resistant subjects with HF or nephrotic syndrome but is available for intravenous use. Chlorothiazide is another intravenous option in decompensated HF. Prolonged furosemide therapy augments the response to a thiazide. ^, Even patients with advanced CKD (eGFR <30 mL/min/1.73 m ² ) who are unresponsive to a loop diuretic can have a marked natriuresis when a thiazide is added to loop diuretic therapy, but close surveillance is required because of a high associated incidence of hypokalemia, excessive ECV depletion, and azotemia.

Amiloride or triamterene increase furosemide natriuresis only modestly but curtail the excretion of K ⁺ and net acid and preserve total body K ⁺ . Distal K ⁺ -sparing agents are generally contraindicated in advanced renal failure because they may cause severe hyperkalemia and acidosis.

HF can occur with reduced or preserved ejection fraction (HFrEF or HFpEF), yet both forms require diuretics as cornerstones of therapy. The approach to cardiac failure depends on the cause and whether there is acute decompensation or a compensated chronic state. ^, This section first reviews the role of diuretics in acute decompensated HF (ADHF) and HF secondary to acute coronary syndrome and thereafter discusses the role of diuretics as maintenance therapy in stable chronic HF. The reader is also referred to a review by Ellison and Felker and the joint guidelines published by the American College of Cardiology and American Heart Association for more detailed recommendations.

Most patients with cirrhotic ascites and peripheral edema have expansion of the ECV owing to arteriolar underfilling and diminished effective arterial blood volume from peripheral vasodilation. This can increase proximal reabsorption, which limits the response to loop diuretics and thiazides. However, the RAAS is activated in cirrhosis and contributes to salt retention, portal hypertension, and likely to progressive liver fibrosis. Thus the use of MRAs, often in high concentration, are rational for cirrhosis and are usually well tolerated. The American Association for the Study of Liver Disease practice guidelines have suggested that first-line treatment of patients with cirrhosis and ascites should consist of sodium restriction (2 g or 88 mmol daily) and diuretics (oral spironolactone, with or without oral furosemide). This can be initiated with 100 mg spironolactone and 40 mg furosemide, with titration upward maintaining the same diuretic ratio ^, to maximally recommended doses of 400 mg of spironolactone and 160 mg of furosemide.

Patients with cirrhosis and ascites usually cannot tolerate ACE inhibitors or ARBs because of a fall in BP. While sodium restriction is indicated in cirrhotic ascites, fluid restriction is only required for developing hyponatremia. ^, Ascitic fluid is largely cleared by the lymphatics. Diuretics reduce thoracic duct lymph flow, perhaps because of reduced lymph production. Furosemide, but not ethacrynic acid, inhibits lymphatic contractions directly by blockade of NKCC1. Thus diuretics decrease ascites formation by decreasing venous and portal hydraulic pressures and concentrating the plasma proteins. ^{, ,}

The maximal daily ascites drainage into the systemic circulation is limited to 300 to 900 mL. Therefore the maximum daily weight loss in nonedematous patients should not exceed 0.3 to 0.5 kg. However, in patients with ascites and edema, daily diuretic-induced weight losses of 1 to 3 kg do not perturb the plasma volume or renal function. In contrast, the same diuretic regimen maintained after the peripheral edema has cleared or been given to nonedematous patients reduces plasma volume by as much as 24% and raises the risks of hyponatremia, alkalosis, and azotemia. Furthermore, the reduced serum albumin levels and RH increased portal venous pressure, coupled with preexisting diuretic use, can lead to true “underfill edema.” Thus a diuretic prescription that is initially safe and effective must be reviewed continuously. Patients with ascites but without peripheral edema are more prone to the development of adverse effects of diuretics.

The most common problems with furosemide in cirrhosis are electrolyte disturbances and volume depletion. Hypokalemia, which is related to preexisting K ⁺ depletion and hyperaldosteronism, can be countered with spironolactone, eplerenone, or a distal K ⁺ -sparing agent. More severe diuretic resistance requires paracentesis. Controlled trials in patients with refractory ascites have shown that large-volume paracentesis is more effective than diuretic therapy in reducing hospital stay and electrolyte complications but does not influence mortality. Even repeated, large-volume paracenteses (4–6 L/day) are safe if intravenous albumin (40 g with each procedure) is administered. Most investigators, however, recommend paracentesis only for cases that are relatively resistant to diuretics and dietary Na ⁺ restriction ( Fig. 49.15 ).

Treatment algorithm for the management of fluid retention in patients with hepatic cirrhosis and ascites.

CV, Cardiovascular; IV, intravenous; S _Na , serum Na ⁺ concentration; TIPS, transjugular intrahepatic portosystemic shunt.

From Ellison DH, Wilcox CS. Diuretics: use in edema and the problem of resistance. In Brady HR, Wilcox CS, eds. Therapy in Nephrology and Hypertension. ed 2. London: Elsevier Science; 2003.

Patients with mild cirrhosis of the liver have a somewhat reduced natriuretic response to furosemide, with little change in diuretic kinetics. However, in those with advanced disease, furosemide absorption is slowed and its volume of distribution is increased because of hypoalbuminemia and an expanded ECV. Additionally, its elimination is delayed because of hypoalbuminemia, which limits proximal tubule diuretic secretion, and a low RBF, which limits renal clearance. However, resistance to loop diuretics in early cirrhosis is largely due to decreased responsiveness to the drug, which correlates with elevated serum aldosterone levels. With the development of ascites, a further decrease in natriuretic response correlates with decreased delivery of furosemide to the urine and with further stimulation of the RAAS.

Diuretic resistance is common in advanced cirrhosis. In addition to the usual causes (see Box 49.2 ), it may herald the development of infection, bleeding, or a critical drop in cardiac output. Patients whose disease is refractory and who are disabled by recurrent paracentesis may show response to body compression or a transjugular intrahepatic portosystemic shunt. Intravenous loop diuretics are generally discouraged because they may precipitate hepatorenal syndrome.

Urinary albumin losses and reduced hepatic albumin synthesis in the nephrotic syndrome eventually lead to hypoalbuminemia. The ensuing fall in plasma oncotic pressure increases the flux of fluid into the interstitial spaces, leading to underfill edema. ^, Additionally, primary renal salt retention can lead to overfill edema ( Fig. 49.16 ). Patients with minimal change disease often have contracted plasma volume and stimulated RAAS, whereas those with other causes of nephrotic syndrome usually have an expanded plasma volume and suppressed RAAS. Micropuncture studies of sodium-retaining animal models of the nephrotic syndrome have demonstrated pronounced NaCl reabsorption in the distal nephron and TAL. ^, The proteinuric kidney of a rat model of unilateral nephrotic syndrome has an enhanced Na ⁺ reabsorption in the CDs and a diminished response to ANP reinforced by hyperaldosteronism. Renin and aldosterone levels are highly variable in patients with nephrotic syndrome.

Comparison of clinical and biochemical characteristics and responses in patients with nephrotic syndrome and underfill edema versus overfill edema.

ANP, Atrial natriuretic peptide; AVP, arginine vasopressin; ↓︎ECV, decrease in extracellular fluid volume; GFR, glomerular filtration rate; PA, plasma aldosterone; PRA, plasma renin activity; ↑︎, increase; ↓︎, decrease; ↔︎, no effect; +, potential side effect.

Modified from Schrier RW, Fassett RG. A critique of the overfill hypothesis of sodium and water retention in the nephrotic syndrome. Kidney Int. 1998;53:1111−1117.

One study has found that a fractional sodium excretion of 0.2% differentiates volume-contracted from volume-expanded nephrotic syndrome in children; treatment with diuretics alone has proven effective and safe for the volume-expanded group. Hypoalbuminemia reduces the binding of furosemide to plasma proteins and thereby enlarges its volume of distribution. Whereas one study has reported that premixing furosemide with albumin in the syringe before intravenous injection enhances the diuresis of patients with the nephrotic syndrome, this finding has not been confirmed. ^{, ,} Indeed, two studies have shown that patients with a serum albumin level of 2 g/dL can deliver normal quantities of furosemide into the urine. ^, Iso-oncotic plasma volume expansion with albumin in patients with nephrotic syndrome fails to induce negative NaCl balance or enhance the response to furosemide and is not generally recommended. ^,

A more logical approach to diuretic resistance is to limit albuminuria with an ACE inhibitor, ARB, or SGLT2 inhibitor that also may combat the associated edema and progressive loss of renal function. The addition of a loop diuretic to an ACE inhibitor or an ARB reduces proteinuria further but increases the serum creatinine concentration.

The tubular secretion of CAIs and loop diuretics by the S2 segment of the proximal tubule depends on albumin. However, in the rabbit, the uptake of loop diuretics into the S1 portion of the proximal tubule, where furosemide is inactivated by glucuronidation, is inhibited by albumin (see Fig. 49.7 ). ^, Albumin infusion into nephrotic patients does indeed increase renal furosemide excretion, whereas hypoalbuminemia enhances its metabolic clearance.

The interaction of furosemide with its target in the lumen of the TAL is restricted by binding to filtered albumin. Addition of albumin to the tubular perfusate of the loop of Henle attenuates the response to perfused furosemide and is reversed by coperfusion with warfarin, which displaces it from its albumin-binding site. However, Agarwal and colleagues reported that displacing furosemide from albumin via the coadministration of sulfisoxazole did not affect natriuresis in patients with nephrotic syndrome. However, this study’s results are not definitive because subjects did not have diuretic resistance.

Animal studies have demonstrated five mechanisms that could impair the responsiveness to loop diuretics in patients with nephrotic syndrome: 1 and 2. decreased delivery and/or decreased tubular secretion of the diuretic; 3. increased renal furosemide metabolism; 4. decreased blockade of tubular transport by the diuretic; and 5. increased NaCl reabsorption by other nephron segments.

Nephrotic edema is best managed with dietary salt and fluid restriction. Most patients show an initial response to a loop diuretic when it is required. Spironolactone or eplerenone is effective in some patients. Decreasing renal function or the administration of indomethacin causes marked resistance to loop diuretics in these patients. The combination of a thiazide diuretic with furosemide dissipates edema but at the expense of marked kaliuresis. Studies have suggested that primary sodium retention in nephrotic syndrome is related to the presence of plasma proteinases in nephrotic urine, such as plasmin, which can cleave the extracellular loop of ENaC that projects into the tubular lumen, leading to ENaC activation. ^, Some studies have shown that this mechanism is also present in patients with proteinuric CKD. ^, Importantly, the ENaC blockers amiloride and triamterene have now been reported to reduce albuminuria in patients with proteinuric kidney disease.

Idiopathic edema predominantly affects women. It causes fluctuating salt retention and edema, exacerbated by orthostasis. The effect of diuretic withdrawal during controlled salt intake were studied in 10 such patients. Although their body weight increased by 0.5 to 5.0 kg within 2 to 8 days, 7 returned to their original weight by 3 weeks, without the reinstitution of diuretic therapy. The investigators concluded that diuretic abuse could cause idiopathic edema. However, this conclusion has been challenged. Nevertheless, normal subjects given a single dose of bumetanide with simultaneous saline infusion to prevent volume depletion are unable to excrete a salt load for at least 2 days. Remarkably, 83% of habitual furosemide abusers who consume high doses over prolonged periods demonstrate medullary nephrocalcinosis and tubulointerstitial fibrosis. Patients with idiopathic edema are best treated with salt restriction.

Hypertension is discussed in Chapter 45 , Chapter 46 , Chapter 47 .

A review of 11 randomized trials of loop diuretics or mannitol for prophylaxis or treatment of established AKI found no benefit. However, diuretics can be used to convert oliguric to nonoliguric AKI that may facilitate management. One study of 58 patients reported that sustained diuresis can be provoked in most patients given 1 g furosemide orally three times daily, but this large dose produced deafness in two patients, which was permanent in one and therefore cannot be recommended. Older observational and randomized studies have indicated that furosemide does not improve the prognosis of AKI, and some have suggested that it may even be worsened. Evidence of adverse effects of ECV overload in critically ill patients has been growing. The Fluid and Catheter Treatment Trial randomized patients with respiratory distress syndrome to different treatment strategies. Those with AKI exhibited reduced mortality with treatments that reduced central venous pressure (CVP) with higher diuretic doses. Thus loop diuretics can be used safely in this population. Furosemide can reduce the need for dialysis by diminishing hyperkalemia, acidosis, or fluid overload. One protocol is to give 40 mg of furosemide, 1 mg of bumetanide, or 25 mg of torsemide intravenously and to double the dose every 60 minutes if there is no response, up to a total daily dose of 1 g of furosemide or the equivalent. Bumetanide and torsemide are metabolized by the liver and so may need to be dosed at a lower level in patients with cirrhosis, whereas furosemide is metabolized by the kidney and therefore accumulates to a greater degree in patients with renal insufficiency (see Fig. 49.6 ). The management of AKI is discussed further in Chapter 28 .

The fractional reabsorption of NaCl and fluid by the renal tubules of subjects who are in balance is reduced in proportion to the fall in the GFR. The renal clearance of loop diuretics falls in parallel with the GFR because of a decreased renal mass and the accumulation of organic acids that compete for proximal secretion. Thus although the maximal increase in fractional excretion of Na ⁺ produced by furosemide is maintained in CKD, ^{, ,} the absolute response to diuretics is limited by a reduction in the absolute rate of NaCl reabsorption and a reduction in the delivery of the diuretic to its target (see Fig. 49.14 ). Although CKD decreases proximal reabsorption, there is enhanced fractional reabsorption in the loop segment, distal tubule, and CDs, with a relative increase of threefold to fourfold per residual nephron in the expression of the NKCC2 in the TAL and the NCC in the DCT.

Torsemide has a greater oral bioavailability than furosemide in CKD. For refractory cases, a loop diuretic infusion (e.g., bumetanide, 1 mg/hour for 12 hours) produces a greater natriuresis and less myalgia than two bolus injections. While thiazides have been used effectively in mild CKD for many years, one randomized, placebo-controlled trial has demonstrated that thiazides are also effective in CKD stage 4. ^{, ,} Moreover, when used in combination with a loop diuretic that increases NaCl delivery and reabsorption at the distal tubule, larger doses of thiazides are effective in patients with moderate azotemia, although at the cost of a sharp further rise in the serum creatinine and blood urea concentrations and a high incidence of hypokalemia and electrolyte disorders. Moreover, high plasma levels of furosemide can cause ototoxicity. Therefore care should be taken not to exceed the ceiling dose ( Table 49.1 ). ^, Epidemiologic studies have correlated diuretic use with kidney failure, but this is likely a noncausal association. In fact, continuing diuretics in patients undergoing dialysis who have residual renal function was associated with lower interdialytic weight gain, less hyperkalemia, and lower cardiac-specific mortality. The presence of hyponatremia in CKD patients using diuretics indicates volume overload, not depletion, and is associated with an earlier initiation of renal replacement therapy. (See Wilcox and Sica and Gehr for further reading on diuretics in CKD.)

Furosemide increases the distal delivery of NaCl and fluid and stimulates aldosterone secretion and phosphate elimination, all of which enhance acid elimination. In addition, there is a direct effect of both furosemide and thiazide on distal acidification via an increased abundance of the H ⁺ -ATPase B1 subunit. Hence furosemide can be used in patients with hyperkalemic (type IV) renal tubular acidosis to increase renal acid excretion. ^, Because this is usually due to hypoaldosteronism, mineralocorticoid therapy may also be indicated.

Ca ²⁺ excretion is increased by osmotic or loop diuretics but is decreased by thiazides and distal agents. Hypercalcemia activates the Ca ²⁺ -sensing receptor ^, that inhibits fluid and NaCl reabsorption in the TAL and impairs renal concentrating ability. The ensuing diuresis and frequent ECV depletion further limit Ca ²⁺ excretion by reducing the GFR and enhancing proximal fluid and Ca ²⁺ reabsorption. Therefore the initial therapy for hypercalcemia is volume expansion with saline, with or without bisphosphonates or steroids, depending on the cause. Loop diuretics may help prevent or treat fluid overload, but there is little evidence to support a role in the treatment of hypercalcemia.

The thiazide-like diuretic indapamide reduced stone formation in hypercalciuric and even normocalciuric patients by reducing the excretion of Ca ²⁺ and oxalate. However, a randomized controlled trial did not identify significant differences in the symptomatic or radiologic recurrence of calcium-containing kidney stones with placebo, 12.5, 25, or 50 mg of hydrochlorothiazide, although there was a reduction in radiologic recurrence alone with the two highest doses. Of note, once-daily dosing, high sodium intake, and the fact that not all patients were stone free at baseline may have offset some of the positive effects of hydrochlorothiazide. Combined therapy with thiazides and citrate may be effective in patients who don’t respond to thiazides alone. Ca ²⁺ reabsorption can be enhanced further by the addition of amiloride or a low-salt diet. KHCO ₃ produces a greater reduction in Ca ²⁺ excretion than KCl when given with hydrochlorothiazide.

Bone cells express an Na ⁺ -Cl ⁻ cotransporter that, when blocked by a thiazide, enhances Ca ²⁺ uptake into bone. Thiazides inhibit osteocalcin, an osteoblast-specific protein that retards bone formation, and directly stimulate the production of the osteoblast differentiation markers. ^, They inhibit bone reabsorption and augment bone mineralization, independently of parathyroid hormone. Thus thiazides may promote bone mineralization both by reducing renal Ca ²⁺ excretion and by direct effects on bone. These biologic effects are supported by data from epidemiologic studies and clinical trials. Thiazide therapy in older persons is associated with an increase in bone mineral density and a reduction in hip and pelvic fractures. In a placebo-controlled trial in postmenopausal women, hydrochlorothiazide (50 mg/day) significantly slowed cortical bone loss. Surprisingly, despite having opposite effects on Ca ²⁺ excretion, a thiazide and a loop diuretic both enhance bone formation in postmenopausal women, at least in the short term. However, loop diuretics alone have been associated with hip bone loss in older men, increased risk of fractures in postmenopausal women, and increased risk of revision following primary total hip arthroplasty in men and women and are best avoided for this indication.

Potassium-sparing diuretics, often given with potassium supplements, may be used in Gitelman syndrome to treat hypokalemia. Spironolactone (200−300 mg/day) was shown to be more effective than amiloride (10−30 mg/day). A later crossover trial compared the efficacies of indomethacin (75 mg/day), amiloride (30 mg/day), and eplerenone (150 mg/day) in correcting hypokalemia. Although all three drugs increased plasma potassium levels significantly, indomethacin was most effective but also associated with the most adverse effects.

Thiazides can reduce urine flow by up to 50% in patients with central or nephrogenic diabetes insipidus. ^, This paradoxic effect is related to decreased GFR, enhanced water reabsorption in the proximal and distal nephron, ^, and an increase in papillary osmolarity leading to distal water reabsorption. A small placebo-controlled crossover trial and an animal study have shown that amiloride prevents lithium-induced polyuria. ^, This effect is attributed to amiloride blocking the cellular entry of lithium into the principal cell via ENaC, where it can downregulate the water channel AQP2 via glycogen synthase kinase 3.

Furosemide or torsemide given to normal subjects or infused into rats reduces the GFR and RBF over the first hours or first day independent of volume losses or angiotensin II. However, when given over weeks or months, diuretics normally do not decrease the GFR, ^, although a decrease in GFR can be precipitated by vigorous diuresis in patients with CKD, severe edema, or cirrhosis and ascites. A rise in the ratio of blood urea nitrogen to creatinine suggests ECV depletion. This change can be ascribed to a combination of a decreased renal urea clearance because of greater urea reabsorption in the distal nephron and increased urea generation due to greater arginine uptake by the liver, which is metabolized by arginase. Combining diuretics with ACE inhibitors or NSAIDs raises the risk of AKI. ^,

This effect is relatively specific for thiazides that inhibit urinary dilution, whereas loop diuretics inhibit urinary concentration and dilution. Indeed, thiazides are 12-fold more likely than loop diuretics to cause hyponatremia. The mechanism of thiazide-induced hyponatremia has long remained elusive. Although one could postulate that the natriuretic effect causes plasma volume depletion with AVP release, patients with thiazide-induced hyponatremia actually gain weight, suggesting that they enhance renal water reabsorption. Clark and colleagues have shown that older age and thiazide diuretics are additive in impairing maximal free water excretion following a water load. Some 80% of thiazide-induced hyponatremia occurs in females, most of whom are older and with low body mass. ^, Hyponatremia can develop during rechallenge with a thiazide, ^{, ,} often within the first 2 weeks of therapy. ^, One study has shown that patients with thiazide-induced hyponatremia displayed increased free water reabsorption independent of AVP and that this is accompanied by increased renal PGE2 excretion ^, ( Fig. 49.17 ). Pharmacogenetic analysis has identified a loss of function variant in SLCO2A1, which encodes a prostaglandin transporter in the distal nephron. This prevents access of PGE2 secreted into tubular fluid by thiazide diuretics from being reabsorbed across the nephron to activate EP1/3 receptors on the basolateral membrane of the collecting duct, where they act as antagonists to AVP. Moreover, the enhanced tubular fluid delivery of PGE2 enhances its activation of EP4 receptors on the luminal membrane of the collecting ducts. This has the opposite effect of enhancing the response to AVP by enhancing the insertion of aquaporin (AQP) channels into the collecting duct cells, thereby enhancing free water reabsorption at any level of AVP. The outcome is enhanced water reabsorption leading to hyponatremia ( Fig. 49.18 ).

Urinary prostaglandin E2 (PGE2) concentration in thiazide-induced hyponatremia (TIH) cases, compared with normonatremic thiazide- and non−thiazide-treated controls, stratified according to their genotype at position 396 of the prostaglandin transporter gene, SLCO2A1. TIH cases who carry at least one copy of the risk allele (396T) have significantly elevated urinary PGE2 levels relative to those homozygous for 396A, whereas no such effect was observed in normonatremic controls. Data are represented as mean ± SEM. ∗ P <.05; ∗∗ P <.01; ∗∗∗ P <.001. UCr, Urinary creatinine.

From Ware JS, Wain LV, Channavajjhala SK, et al. Phenotypic and pharmacogenetic evaluation of patients with thiazide-induced hyponatremia. J Clin Invest. 2017;127(9):3367–3374.

Hypothesis for the role of SLCO2A1 in contributing to thiazide-induced hyponatremia in individuals carrying the SLCO2A1 A396T variant.

Thiazide diuretics increase renal synthesis and secretion of prostaglandin E2 into the tubular fluid of the distal nephron. Tubular fluid PGE2 is normally reabsorbed by the prostaglandin transporter in the distal nephron. PGE2 activates the PG type 1 and 3 receptors (EP1R and EP3R) on the basolateral membrane of the principal cells of the collecting duct (CD) that inhibit the insertion and activation of the aquaporin type 2 channel (AQP2) into the luminal membrane. Inactivation of the PGT limits PGE2 from accessing the basolateral EP1 and EP3 receptors and increases its luminal concentration and activation of the luminal EP4 receptor that prompts the insertion and activity of AQP2. This results in increased reabsorption of free water and the development of hyponatremia coupled with increased excretion of PGE2.

ADH, Antidiuretic hormone; AQP2, aquaporin-2; PGE2, prostaglandin E2; PGT, phosphate glucuronyl transferase; TIH, thiazide-induced hyponatremia. (Data from Palmer BF, Clegg DJ. Altered prostaglandin signaling as a cause of thiazide-induced hyponatremia. Am J Kidney Dis . 2018;71(6):769–771.)

Mild hyponatremia can be treated by withdrawal of diuretics, restriction of the daily intake of free water to 1.0 to 1.5 L, increasing solute intake, restoring of any K ⁺ and Mg ²⁺ losses, and replenishing of NaCl if the patient is clearly volume depleted. ^,

It is becoming increasingly clear that even mild chronic hyponatremia can produce symptoms. Therefore the development of diuretic-induced hyponatremia should probably be regarded as an indication to discontinue thiazides.

Four mechanisms have been identified that increase renal K ⁺ elimination during therapy with thiazides or loop diuretics ( Fig. 49.19 ): increased tubular flow, increased sodium delivery to CNT/CCD, secretion of AVP, aldosterone, and alkalosis. Flow-dependent K ⁺ secretion by the distal nephron provides a universal mechanism for increased K ⁺ secretion in response to diuretics that act more proximally. The basal and flow-dependent components of K ⁺ secretion have been shown to be mediated largely by distinct channels. Whereas basal secretion traverses ROMK channels, flow-dependent K ⁺ secretion is mediated by calcium-activated maxi or big K ⁺ (BK) channels. Normally, a flow-dependent rise in K ⁺ secretion during increased water intake is offset by a drop in AVP concentration that diminishes distal K ⁺ secretion. Diuretic therapy, however, is unusual because it combines increased distal tubule flow with maintained or increased AVP release due to nonosmotic stimulation, which is the second mechanism contributing to kaliuresis. Indeed, inhibition of AVP release in normal subjects undergoing a furosemide diuresis inhibits the kaliuresis. Nonosmotic AVP release is common in edematous subjects. Therefore enhanced release of AVP and increased distal tubule fluid delivery combine to promote ongoing K ⁺ and H ⁺ losses during diuretic therapy for edema. Diuretic-induced aldosterone secretion also promotes distal K ⁺ secretion. ^{, ,} The effects of flow and aldosterone on distal K ⁺ secretion are normally counterbalanced during changes in salt intake, just as are the effects of flow and AVP induced by changes in water intake (see Fig. 49.20 ). Diuretic treatment, however, uncouples the two because it enhances the secretion of aldosterone and AVP but increases distal flow, thereby accounting for the particular importance of aldosterone and AVP in promoting K ⁺ loss with diuretics. Finally, diuretic-induced alkalosis enhances the distal secretion of K ⁺ .

Hypothesis for the hyperglycemic actions of thiazide diuretics.

ACEI, Angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ECV, extracellular fluid volume; SNS, sympathetic nervous system; CO, cardiac output; ↑︎, increase(d); ↓︎, decrease(d).

From Wilcox CS. Metabolic and adverse effects of diuretics. Semin Nephrol. 1999;19:557−568.

The serum potassium concentration (S _K ) of hypertensive patients not receiving KCl supplements falls by an average of 0.3 mmol/L with furosemide and 0.6 mmol/L with thiazides. In normal subjects receiving a single daily dose of furosemide, there is no detectable change in K ⁺ balance on a high-salt diet. However, on a low-salt diet, this same medication produced a negative K balance. These findings imply that the primary cause of hypokalemia during loop diuretic administration is dependent on the clinical scenario. With a low-sodium diet or more frequent dosing, renal potassium losses may play a greater role. With a high-sodium diet and less frequent dosing, redistribution of K ⁺ into cells related to metabolic alkalosis may play a larger role. ^, For thiazide diuretics, which have significantly greater decreases in serum potassium with standard dosing, findings concerning the role of NCC in the physiology of potassium handling may provide some insight. Some studies have demonstrated that acute and chronic deactivation of NCC plays a key role in excreting a potassium load. This deactivation is critical for delivery of sodium to the CNT and CCD so that ENaC can reabsorb sodium, providing the driving force for potassium secretion via ROMK and Maxi-K channels. Similarly, thiazides can mimic this effect, resulting in significant potassium excretion in the absence of a potassium load, primarily when coupled with aldosterone-induced ENaC stimulation. This observation helps to explain why higher doses of thiazide diuretics, as used previously, are no longer recommended for patients with normal kidney function. High doses are more likely to lead to hypokalemia, which by, activating NCC, counteracts diuretic effectiveness.

Mild, diuretic-induced hypokalemia (serum potassium, 3.0−3.5 mmol/L) increases the frequency of ventricular ectopy. Some investigators have proposed that this does not pose a risk of clinically significant cardiac dysrhythmia. ^, In contrast, others have reported a dose-dependent risk of cardiac arrest in patients receiving thiazides that is prevented by therapy with a K ⁺ -sparing diuretic such as amiloride. Studies in patients with HF report that mortality increases as S _K falls below 4.0 mmol/L.

Adverse effects of hypokalemia are clearly important in certain cases:

• 1.

  Severe hypokalemia (S _K < 3.0 mmol/L) requires treatment because it is associated with a doubling of serious ventricular dysrhythmias, muscular weakness, and rhabdomyolysis.

• 2.

  Mild hypokalemia can precipitate dangerous dysrhythmias in patients with cardiac dysfunction due to left ventricular hypertrophy, coronary ischemia, HF, prolonged QT interval, anoxia, or ischemia and in patients with known dysrhythmias.

• 3.

  Hypokalemia enhances the toxicity of cardiac glycosides by diminishing the renal tubule secretion of digoxin and by enhancing its binding to cardiac Na ⁺ -K ⁺ -ATPase, thereby exaggerating its actions on the heart.

• 4.

  Hypokalemia stimulates renal ammoniagenesis. This effect is dangerous for patients with cirrhosis and ascites who are prone to hepatic encephalopathy due to hyperammonemia. Moreover, the accompanying diuretic-induced alkalosis partitions ammonia into the brain.

• 5.

  Catecholamines partition K ⁺ into cells and lower S _K . Myocardial infarction provokes sufficient catecholamine release to lower S _K by approximately 0.5 mmol/L, which is potentiated in patients who have received prior thiazide therapy.

• 6.

  Hypokalemia impairs insulin release and predisposes to hyperglycemia.

• 7.

  Hypokalemia limits the antihypertensive action of thiazides. In a placebo-controlled study of hypokalemic subjects receiving thiazide diuretics, coadministration of KCl that restored S _K also reduced BP significantly.

Therefore it is prudent to prevent even mild degrees of hypokalemia by increasing the intake of K ⁺ with Cl ⁻ (see Fig. 49.19 ). However, this often requires 40 to 80 mmol daily and, in the presence of alkalosis, hyperaldosteronism, or Mg ²⁺ depletion, hypokalemia is quite unresponsive to dietary KCl. A more effective, convenient, and predictable strategy is to prescribe a combined therapy with a distal K ⁺ -sparing agent such as amiloride or triamterene, which maintains S _K (see Fig. 49.19 ), ^, prevents diuretic-induced alkalosis, and provides further natriuresis and antihypertensive efficacy. Other strategies include the administration of an ACE inhibitor, ARB, MRA, or SGLTi to counter hyperaldosteronism, which would promote distal K ⁺ secretion.