Osmotic diuretics such as mannitol are non-resorbable solutes which retain water in the tubular fluid (Site 1, Fig. 27.1). Their effect is to increase water rather than sodium loss, and this is reflected in their special use acutely to reduce intracranial or intraocular pressure and not states associated with sodium overload.

Loop of Henle

The tubular fluid now passes into the loop of Henle where 25% of the filtered sodium is reabsorbed. There are two populations of nephron: those with short loops that are confined to the cortex, and the juxtamedullary nephrons whose long loops penetrate deep into the medulla and are concerned principally with water conservation;¹ the following discussion refers to the latter.

The physiological changes are best understood by considering first the ascending limb. In the thick segment (Site 2, Fig. 27.1), sodium and chloride ions are transported from the tubular fluid into the interstitial fluid by the three-ion co-transporter system (i.e. Na⁺/K⁺/2Cl⁻ called NKCC2) driven by the sodium pump. The co-transport of these ions is dependent on potassium returning to the lumen through the rectifying outer medullary potassium (ROMK) channel; otherwise potassium would be rate limiting. As the tubule epithelium is ‘tight’ here, i.e. impermeable to water, the tubular fluid becomes dilute, the interstitium becomes hypertonic, and fluid in the adjacent descending limb, which is permeable to water, becomes more concentrated as it approaches the tip of the loop, because the hypertonic interstitial fluid sucks water out of this limb of the tubule. The ‘hairpin’ structure of the loop thus confers on it the property of a countercurrent multiplier, i.e. by active transport of ions a small change in osmolality laterally across the tubular epithelium is converted into a steep vertical osmotic gradient.

The high osmotic pressure in the medullary interstitium is sustained by the descending and ascending vasa recta, long blood vessels of capillary thickness that lie close to the loops of Henle and act as countercurrent exchangers, for the incoming blood receives sodium from the outgoing blood.²Furosemide, bumetanide, piretanide, torasemide and ethacrynic acid act principally at Site 2 by inhibiting the three-ion transporter, thus preventing sodium ion reabsorption and lowering the osmotic gradient between cortex and medulla; this results in the formation of large volumes of dilute urine. Hence, these drugs are called ‘loop’ diuretics.

Distal convoluted tubule

The ascending limb of the loop then re-enters the renal cortex where its morphology changes into the thin-walled distal convoluted tubule (Site 3, Fig. 27.1). Here uptake is still driven by the sodium pump but sodium and chloride are taken up through a different transporter, the Na–Cl co-transporter, called NCC (formerly NCCT). Both ions are rapidly removed from the interstitium because cortical blood flow is high and there are no vasa recta present; the epithelium is also tight at Site 3 and consequently the urine becomes more dilute. Thiazides act principally at this region of the cortical diluting segment by blocking the NCC transporter.

Collecting duct

In the collecting duct (Site 4), sodium ions are exchanged for potassium and hydrogen ions. The sodium ions enter through the epithelial Na channel (called ENaC), which is stimulated by aldosterone. The aldosterone (mineralocorticoid) receptor is inhibited by the competitive receptor antagonist spironolactone, whereas the sodium channel is inhibited by amiloride and triamterene. All three of these diuretics are potassium sparing because potassium is normally secreted through the potassium channel, ROMK (see Fig. 27.1), down the potential gradient created by sodium reabsorption.

All other diuretics, acting proximal to Site 4, cause potassium loss, because they dump sodium into the collecting duct. Removal of this sodium through ENaC increases the potential gradient for potassium secretion through ROMK. The potassium-sparing diuretics are normally considered weak diuretics because Site 4 is normally responsible for ‘only’ 2–3% of sodium reabsorption, and they usually cause less sodium loss than thiazides or loop diuretics. Nevertheless, patients with genetic abnormalities of ENaC show salt wasting or retention to a degree that significantly affects their blood pressure, depending on whether the mutation causes, respectively, loss or gain of channel activity. Although ENaC clearly does not have the capacity to compensate for large sodium losses, e.g. during loop diuretic usage, it is the main site of physiological control (via aldosterone) over sodium loss.

The collecting duct then travels back through the medulla to reach the papilla; in doing so it passes through a gradient of increasing osmotic pressure which draws water out of tubular fluid. This final concentration of urine is under the influence of antidiuretic hormone (ADH) whose action is to increase water permeability by increasing the expression of specific water channels (or aquaporins); in its absence water remains in the collecting duct. Ethanol causes diuresis by inhibiting the release of ADH from the posterior pituitary gland.

Diuresis may also be achieved by extrarenal mechanisms, by raising the cardiac output and increasing renal blood flow, e.g. with dobutamine and dopamine.

Classification

The maximum efficacy in removing salt and water that any drug can achieve is dependent on its site of action, and it is appropriate to rank diuretics according to their natriuretic capacity, as set out below. The percentages refer to the highest fractional excretion of filtered sodium under carefully controlled conditions and should not be taken to represent the average fractional sodium loss during clinical use.

High efficacy

Furosemide

and the other ‘loop’ diuretics can cause up to 25% of filtered sodium to be excreted. Their action impairs the powerful urine-concentrating mechanism of the loop of Henle and confers higher efficacy compared with drugs that act in the relatively hypotonic cortex (see below). Progressive increase in dose is matched by increasing diuresis, i.e. they have a high ‘ceiling’ of effect. In fact, they are so effective that over-treatment can readily dehydrate the patient. Loop diuretics remain effective at a glomerular filtration rate (GFR) below 10 mL/min (normal 120 mL/min).

Moderate efficacy

The thiazide family,

including bendroflumethiazide and the related chlortalidone, clopamide, indapamide, mefruside, metolazone and xipamide, cause 5–10% of filtered sodium load to be excreted. Increasing the dose produces relatively little added diuresis compared to loop diuretics, i.e. they have a low ‘ceiling’ of effect. Such drugs cease to be effective once the GFR has fallen below 20 mL/min (except metolazone).

Individual diuretics

High-efficacy (loop) diuretics

Moderate-efficacy diuretics

(See also Hypertension, Ch. 24.)

Thiazides

Thiazides depress salt reabsorption in the distal convoluted tubule (at Site 3), i.e. upstream of the region of sodium–potassium exchange at Site 4. Hence these drugs have the important effect of raising potassium excretion. Thiazides lower blood pressure, initially due to a reduction in intravascular volume but chronically by a reduction in peripheral vascular resistance. The latter is accompanied by diminished responsiveness of vascular smooth muscle to noradrenaline/norepinephrine; they may also have a direct action on vascular smooth muscle membranes, acting on an as yet unidentified ion channel.

Uses

Thiazides are given for mild cardiac failure and mild hypertension, or for more severe degrees of hypertension, in combination with other drugs.

Pharmacokinetics

Thiazides are generally well absorbed orally and most begin to act within an hour. Differences among the numerous derivatives lie principally in duration of action. The relatively water-soluble, e.g. cyclopenthiazide, chlorothiazide, hydrochlorothiazide, are most rapidly eliminated, their peak effect occurring within 4–6 h and passing off by 10–12 h. They are excreted unchanged in the urine and active secretion by the proximal renal tubule contributes to their high renal clearance and t_½ of less than 4 h.

The relatively lipid-soluble members of the group, e.g. polythiazide, hydroflumethiazide, distribute more widely into body tissues and act for > 24 h, which can be problematic if the drug is used for diuresis, but there is no evidence this property makes them more effective at controlling hypertension. With the exception of metolazone, thiazides are not effective when renal function is moderately impaired (GFR < 20 mL/min), because they are not filtered in sufficient concentration to inhibit the NCC.

Adverse effects

in general are discussed below. Rashes (sometimes photosensitive), thrombocytopenia and agranulocytosis occur. Thiazide-type drugs increase total plasma cholesterol concentration, but in long-term use this is less than 5%, even at high doses. The questions about the appropriateness of thiazides for mild hypertension, of which ischaemic heart disease is a common complication, are laid to rest by their proven success in randomised outcome comparisons (see Ch. 24).

Bendroflumethiazide

is a satisfactory member for routine use. For a diuretic effect the oral dose is 5–10 mg, which usually lasts less than 12 h, so that it should be given in the morning. As an antihypertensive, 2.5 mg is commonly prescribed. However, this dose does not achieve maximal blood pressure reduction in patients with salt-dependent hypertension (which includes most patients receiving a blocker of the renin–angiotensin system), and has not been shown to prevent strokes or heart attacks. New guidance is likely to recommend some return to higher doses – and/or greater use of the thiazide-like diuretics (below), with or without co-prescribed K⁺-sparing diuretic. Important potassium depletion is uncommon, but plasma potassium concentration should be checked in potentially vulnerable groups such as the elderly (see Ch. 25). If marked hypokalaemia occurs hyperaldosteronism should be excluded.

Hydrochlorothiazide is a satisfactory alternative. Other members of the group include benzthiazide, chlorothiazide, cyclopenthiazide, hydroflumethiazide and polythiazide.

Diuretics related to the thiazides

Several compounds, although not strictly thiazides, share structural similarities with them and probably act at the same site on the nephron; they therefore exhibit moderate therapeutic efficacy. Overall, these substances have a longer duration of action, are used for oedema and hypertension, and their profile of adverse effects is similar to that of the thiazides. They are listed below.

Chlortalidone acts for 48–72 h after a single oral dose.

Indapamide is structurally related to chlortalidone but lowers blood pressure at subdiuretic doses, perhaps by altering calcium flux in vascular smooth muscle. It has less apparent effect on potassium, glucose or uric acid excretion (see below).

Metolazone is effective when renal function is impaired. It potentiates the diuresis produced by furosemide and the combination can be effective in resistant oedema, although the risk of hypokalaemia is very high.

Xipamide is structurally related to chlortalidone and to furosemide. It induces a diuresis for about 12 h that is brisker than with thiazides; this may trouble the elderly.

Low-efficacy diuretics

Spironolactone

(Aldactone) is structurally similar to aldosterone and competitively inhibits its action in the distal tubule (Site 4; exchange of potassium for sodium); excessive secretion of aldosterone contributes to fluid retention in hepatic cirrhosis, nephrotic syndrome, congestive heart failure (see specific use in Ch. 25) and primary hypersecretion (Conn’s syndrome). Spironolactone is also useful in the treatment of resistant hypertension, where increased aldosterone sensitivity is increasingly recognised as a contributory factor.

Spironolactone itself has a short t_½ (1.6 h), being extensively metabolised, and its prolonged diuretic effect is due to the most significant active product, canrenone (t_½ 17 h). Spironolactone is relatively ineffective when used alone but is more efficient when combined with a drug that reduces sodium reabsorption proximally in the tubule, i.e. a loop diuretic. Spironolactone (and amiloride and triamterene; see below) usefully reduces the potassium loss caused by loop diuretics, but its combination with another potassium-sparing diuretic must be avoided as hyperkalaemia will result. Dangerous potassium retention is particularly likely if spironolactone is given to patients with impaired renal function. It is given orally in one or more doses totalling 100–200 mg/day. Maximum diuresis may not occur for up to 4 days. If after 5 days the response is inadequate, the dose may be increased to 300–400 mg/day. Lower doses (0.5–1 mg/kg) are required to treat hypertension.

Adverse effects. Oestrogenic effects are the major limitation to its long-term use. They are dose dependent, but in the Randomized Aldactone Evaluation Study (RALES)³ (see Ch. 25) even 25 mg/day caused breast tenderness or enlargement in 10% of men. Women may also report breast discomfort or menstrual irregularities, including amenorrhoea. Minor gastrointestinal upset also occurs and there is increased risk of gastroduodenal ulcer and bleeding. These are reversible on stopping the drug. Spironolactone is reported to be carcinogenic in rodents, but many years of clinical experience suggest that it is safe in humans. Nevertheless, the UK licence for its use in essential hypertension was withdrawn (i.e. possible use long term in a patient group that includes the relatively young), but is retained for other indications.

Eplerenone

is a spironolactone analogue licensed for use in heart failure that appears to be free of the oestrogenic effects; probably because of its lower affinity for the oestrogen receptor. It is useful in patients who need an aldosterone-receptor blocking agent, but are intolerant of the endocrine effects of spironolactone.

Amiloride

blocks the ENaC sodium channels in the distal tubule. This action complements that of the thiazides with which it is frequently combined to increase sodium loss and limit potassium loss. One such combination, co-amilozide (Moduretic; amiloride 2.5–5 mg plus hydrochlorothiazide 25–50 mg) is used for hypertension or oedema. The maximum effect of amiloride occurs about 6 h after an oral dose, with a duration of action greater than 24 h (t_½ 21 h). The oral dose is 5–20 mg daily.

Triamterene

(Dytac) is a potassium-sparing diuretic with an action and use similar to that of amiloride. The diuretic effect extends over 10 h. Gastrointestinal upsets occur. Reversible, non-oliguric renal failure may occur when triamterene is used with indometacin (and presumably other NSAIDs). It may also give the urine a blue coloration.

Indications for diuretics

• Oedema states associated with sodium overload, e.g. cardiac, renal or hepatic disease, and also without sodium overload, e.g. acute pulmonary oedema following myocardial infarction. Note that oedema may also be localised, e.g. angioedema over the face and neck or around the ankles with some calcium channel blockers, or due to low plasma albumin, or immobility in the elderly; in none of these circumstances is a diuretic indicated.

• Hypertension, by reducing intravascular volume and probably by other mechanisms too, e.g. reduction of sensitivity to noradrenergic vasoconstriction.

• Hypercalcaemia. Furosemide reduces calcium reabsorption in the ascending limb of the loop of Henle, which action may be utilised in the emergency reduction of raised plasma calcium levels in addition to rehydration and other measures.

• Idiopathic hypercalciuria, a common cause of renal stone disease, may be reduced by thiazide diuretics.

• The syndrome of inappropriate secretion of antidiuretic hormone secretion (SIADH) may be treated with furosemide if there is a dangerous degree of volume overload (see also p. 423).

• Nephrogenic diabetes insipidus, paradoxically, may respond to diuretics which, by contracting vascular volume, increase salt and water reabsorption in the proximal tubule, and thus reduce urine volume.

Therapy

Congestive cardiac failure

The main account appears in Chapter 25, where the emphasis is now on early use of angiotensin-converting enzyme (ACE) inhibitors and β-adrenoceptor antagonists that are specifically diuretic sparing. But oral diuretics are easily given repeatedly, and lack of supervision can result in insidious over-treatment. Relief at disappearance of the congestive features can mask exacerbation of the low-output symptoms of heart failure, such as tiredness and postural dizziness due to reduced blood volume. A rising blood urea level is usually evidence of reduced glomerular blood flow consequent on a fall in cardiac output, but does not distinguish whether the cause of the reduced output is over-diuresis or worsening of the heart failure itself. The simplest guide to the success or failure of diuretic regimens is to monitor body-weight, which the patient can do equipped with just bathroom scales. Fluid intake and output charts are more demanding of nursing time, and often less accurate.