Water, Sodium and Potassium Balance

Drink a litre of water quickly and the sodium concentration of your plasma dips for a while, although the dip does not require any sodium leaving your body. Lose a litre of sweat and it rises, although sweat carries sodium out. Plasma sodium concentration is a ratio whose numerator and denominator are controlled by two systems with different sensors, effectors and time courses. This lesson keeps them apart, puts numbers on the water side, then follows potassium, which runs on two clocks at once. The kidney traced the nephron and the hormones; body water and its compartments gave the volumes.

Water sets osmolality; sodium sets volume

Osmolality is dissolved particles per kilogram of water, in mOsm per kg; in an adult human at rest plasma typically runs about 275 to 295 (osmoregulation). Tonicity is the part of osmolality made by solutes that do not cross cell membranes freely, so it is the part that moves water into or out of cells: urea raises osmolality and adds almost nothing to tonicity.

The water loop controls osmolality. Osmoreceptor neurons in the lamina terminalis fire faster when plasma tonicity rises, since that is what draws water out of them; their output drives thirst and release of vasopressin, which raises the water permeability of the collecting duct. The effectors are water intake and renal water excretion, and the sensor reads a concentration, blind to how much sodium the body holds.

The sodium loop controls effective circulating volume, the part of extracellular volume that the cardiovascular sensors report: arterial and atrial baroreceptors, stretch of the afferent arteriole, and salt delivery to the macula densa (glomerular filtration). Its effectors, renin, angiotensin II, aldosterone, sympathetic outflow and the natriuretic peptides, set sodium excretion, and because extracellular water follows extracellular sodium osmotically, controlling sodium content is how volume is controlled.

Edelman’s isotope-dilution relation, treated in body water and its compartments, writes plasma sodium concentration, in simplified form, as exchangeable sodium plus exchangeable potassium over total body water, the fitted regression also carrying a slope and an intercept: a solute-to-water ratio in which cell potassium counts alongside extracellular sodium. A low value therefore means water in excess of solute. Total body sodium may be low, normal or high: a low concentration can go with an expanded extracellular volume, and also with real sodium loss replaced by hypotonic fluid.

Vasopressin opens the collecting duct, and thirst starts at almost the same osmolality

That OVLT neurons sense osmolality themselves was shown in slices. Ciura and Bourque recorded from neurons of the organum vasculosum of the lamina terminalis in mouse hypothalamic explants and acutely isolated cells and found that they depolarize and fire faster when bath osmolality rises, with no circulation present. Mice lacking the channel protein TRPV1 drank less than wild-type mice after a hypertonic challenge, although they still drank, so the response was attenuated and not abolished, and OVLT neurons from these mice lost their osmotic response. The central circuitry belongs to the hypothalamus and autonomic control, anticipatory thirst to tissues and homeostasis.

The two human thresholds sit close together. In ten healthy volunteers given hypertonic saline, thirst on a rating scale appeared near a plasma osmolality of 281 mosmol per kg and the regression threshold for vasopressin release was 285, with slopes of 0.3 cm of scale and 0.4 pmol per L per mosmol per kg above those thresholds. Those numbers come from one study of ten people, whose thirst threshold was lower than in earlier work; reviews still describe vasopressin as the first line of defence and thirst as a second line at higher osmolality. Release is not purely osmotic: vasopressin also rises when blood volume falls and with stress, so a volume-depleted animal can retain water while plasma is dilute (skin and human heat covers sweat losses).

At the collecting duct, vasopressin binds the V2 receptor on the basolateral membrane and raises cyclic AMP (G proteins, cAMP and PKA), and water channels are inserted into the apical membrane of principal cells. The channel is aquaporin-2, cloned from rat kidney in 1993 by Fushimi, Sasaki and colleagues, who showed that expressing it in Xenopus oocytes markedly raised osmotic water permeability (the kidney covers its trafficking, membrane transport aquaporin-1). With channels in place water leaves down the medullary gradient, and a water-deprived adult human reaches about 1,000 to 1,200 mOsm per kg, while with the hormone absent urine can fall near 50. Copeptin, cleaved from the same prohormone, is the stable plasma surrogate for vasopressin.

Mice let each step be removed. Inducible deletion of aquaporin-2 in adult mice raised urine output from about 2 to 25 mL per day and dropped urine osmolality from about 2,000 to below 500 mosmol per kg, with no rise after water deprivation: without the channel, vasopressin cannot raise collecting-duct water permeability at all, although it still acts on medullary salt and urea transport. In cultured kidney cells, gain-of-function V2 receptor variants do the reverse, holding aquaporin-2 at the plasma membrane without vasopressin.

Free-water clearance splits urine into two parts

Treat urine as two volumes added: one carrying all the excreted solute at the osmolality of plasma, one of solute-free water. Osmolar clearance is the volume of plasma per minute whose solute appears in the urine; free-water clearance is the rest of the flow.

Cosm = Uosm x V / Posm, and CH2O = V - Cosm, with V the urine flow

A positive CH2O means urine more dilute than plasma, zero means isosmotic urine, and a negative value means concentrated urine, in which case minus CH2O is the solute-free water reabsorbed into blood (TcH2O). The division in Cosm is by plasma, not urine, osmolality.

Worked problem: splitting the urine four ways

Round illustrative numbers for healthy adults, not measurements. Assume a steady state over each collection and constant plasma osmolality.

1. Water diuresis, after a water load with vasopressin suppressed. V = 10 mL per min, Uosm = 60 mOsm per kg, Posm = 282 mOsm per kg. Solute excretion = 60 x 10 = 600 microOsm per min. Cosm = 600 / 282 = 2.13 mL per min. CH2O = 10 - 2.13 = +7.87 mL per min: of each 10 mL, about 2.1 mL carries the solute and about 7.9 mL is solute-free water.
2. Antidiuresis, after overnight water deprivation. V = 0.50 mL per min, Uosm = 1,000 mOsm per kg, Posm = 295 mOsm per kg. Solute excretion = 500 microOsm per min. Cosm = 500 / 295 = 1.69 mL per min. CH2O = 0.50 - 1.69 = -1.19 mL per min, so TcH2O = +1.19 mL per min returned to blood.
3. Solute sets the range. Hold solute excretion at 600 microOsm per min, or 600 x 1,440 = 864 mOsm per day. At a maximal Uosm of 1,200 mOsm per kg the smallest daily volume is 864 / 1,200 = 0.72 L; at a very dilute Uosm of 60 mOsm per kg the largest is 864 / 60 = 14.4 L. Water excretion is limited by solute delivery as much as by the hormone.
4. Hormone present while plasma is dilute. A volunteer given vasopressin with a water load: Posm = 265, Uosm = 500 mOsm per kg, V = 1.0 mL per min. Cosm = 500 / 265 = 1.89 mL per min, CH2O = -0.89 mL per min, Uosm/Posm = 1.9: concentrated urine while plasma is dilute, the signature of water retention that osmolality is not switching off.

Start from a worked case in the lesson (round illustrative numbers, not measurements) My own values Case 1: solute 600, Uosm 60, Posm 282 Case 2: solute 500, Uosm 1,000, Posm 295 Case 3: solute 600, Uosm 1,200 (smallest daily volume) Case 3: solute 600, Uosm 60 (largest daily volume) Case 4: solute 500, Uosm 500, Posm 265 Solute excretion 600 µOsm per min Urine osmolality Uosm 60 mOsm per kg Plasma osmolality Posm 282 mOsm per kg

Urine flow V = solute ÷ Uosm

Osmolar clearance Cosm

Free-water clearance CH₂O

Uosm ÷ Posm

Urine volume if this flow were held for 24 h, V × 1,440

Water returned to blood, −CH₂O

In this sketch: round illustrative numbers for healthy adults, not measurements; a steady state over each collection with constant plasma osmolality; the urine is treated as two volumes added, one carrying all the solute at plasma osmolality and one of solute-free water; you set solute excretion, Uosm and Posm independently, so some combinations may not occur in a healthy adult. The slider limits are made up for the sketch (solute excretion 300 to 900 µOsm per min, Posm 265 to 295 mOsm per kg), except that Uosm about 50 to about 1,200 is the lesson’s range and Posm 265 is the lesson’s fourth example. Values are shown rounded, so “≈” marks a rounded step. The bar’s scale changes with the numbers, so read the axis.

Rule to keep: Cosm = Uosm × V ÷ Posm (divide by plasma osmolality, not urine osmolality) and CH₂O = V − Cosm. Positive means urine more dilute than plasma, zero isosmotic urine, negative concentrated urine, and then minus CH₂O is the solute-free water reabsorbed into blood. This sketch is a rule only and does not predict plasma sodium concentration: free-water clearance is a renal, solute-based index of one collection and cannot say which way plasma sodium will move next, since that depends also on what is drunk and on non-renal losses.

One limit matters more than the arithmetic. Free-water clearance is a renal, solute-based index of one collection, and it cannot say which way plasma sodium concentration will move next. Osmolar clearance counts urea and other ineffective osmoles, which is why electrolyte-free clearance describes the effect on sodium concentration better, and whole-body balance must also count what is drunk and non-renal losses.

Sodium balance runs on volume sensors, and aldosterone escapes

Renin release from granular cells of the afferent arteriole has three inputs: a fall in renal perfusion pressure sensed in the arteriolar wall; a fall in sodium chloride at the macula densa, inversely related to renin secretion in perfused preparations (glomerular filtration); and sympathetic stimulation of beta-adrenergic receptors on those cells. Angiotensin II raises proximal sodium reabsorption and stimulates aldosterone release, and aldosterone is also raised directly by a rise in plasma potassium (the adrenal glands, nuclear receptors). Pushing the other way, atrial natriuretic peptide is released from atrial granules mainly when atrial-wall stretch rises with intravascular volume, and lowers blood pressure through natriuresis and diuresis, vasorelaxation and antagonism of the renin-angiotensin system (blood pressure regulation, tubular transport and diuretics).

If aldosterone set sodium excretion, giving it continuously should retain sodium continuously. August, Nelson and Thorn tested that in 1958 in two normal men on constant metabolic diets who received large amounts of aldosterone daily while sodium intake, urinary sodium and body weight were measured each day. The prediction was sustained retention, steady weight gain and oedema; what they saw was retention and weight gain at first, then urinary sodium returning to about the control level while the hormone continued, with oedema slight. They called it escape.

The study rules in that once extracellular volume has expanded by some amount, other mechanisms offset the hormone and restore sodium excretion in normal humans, and rules out aldosterone as the sole setter of that output. It does not show the mechanism: pressure natriuresis, natriuretic peptides, reduced proximal reabsorption and WNK-kinase and paracrine effects are all proposed, and a 2026 review still calls the question open. It rests on two men, and this lesson tells only its sodium result.

Sodium content and extracellular water may also not move together: in rats on a high-salt diet, sodium accumulates in skin, reported as hypertonic in some studies and as isotonic in others, and in healthy men on fixed salt intake in long confinement studies total body sodium swung by roughly 200 to 400 mmol on roughly monthly and longer rhythms, with no parallel change in body weight or extracellular water.

Potassium is shifted in minutes and excreted over hours

About 98 percent of body potassium sits inside cells, mostly in skeletal muscle; extracellular potassium is 3.5 to 5 mEq per L, and daily intake is about 70 mEq. A single meal can carry a sizeable fraction of the whole extracellular content, so the first defence is the muscle cell rather than the kidney, and why small outside changes matter for excitable tissue is in cardiac electrophysiology and the ECG.

Insulin and catecholamines drive potassium into muscle by stimulating the Na+-K+-ATPase (insulin, glucagon and glucose), insulin directly and epinephrine through beta2-adrenergic receptors. In six healthy volunteers an epinephrine infusion lowered plasma potassium by 0.82 meq per L, and in a second experiment at half that infusion rate a selective beta2-adrenergic antagonist given as a probe prevented the fall, which locates the receptor but not the tissue. Potassium-deprived rats show the dependence from the other side: their muscle becomes resistant to insulin’s stimulation of potassium uptake, so less potassium is taken into cells when insulin rises.

Acid-base changes shift potassium too, but mineral-acid loading moves far more out of cells than organic-acid loading, and the shift is not a direct potassium-for-hydrogen exchange (acid-base balance).

Renal excretion is the slower arm, and it is secretory rather than filtration-driven. Less than 10 percent of filtered potassium reaches the distal nephron in mammals; most urinary potassium is secreted by principal cells, which take it in across the basolateral Na+-K+-ATPase and release it to the lumen through ROMK; flow-activated BK (maxi-K) channels add secretion at high flow, and much of that BK pathway sits in intercalated cells. Two major influences set the rate: mineralocorticoid activity, and delivery of sodium and water to those cells, with higher flow raising secretion. Plasma potassium itself and other cellular signals change it too. So high aldosterone does not always mean potassium loss. It rises when volume is low, a state where distal delivery is low and potassium is retained, and it also rises when plasma potassium is high, a state where potassium is excreted. How one hormone does both is the aldosterone paradox, called poorly understood in a 2011 review; much of the account below has filled in since.

Part of the answer came from the segment upstream. Terker, Zhang, McCormick and colleagues, with Ellison as senior author, put mice on diets varying sodium and potassium and found that potassium deficiency activated the NaCl cotransporter NCC in the distal convoluted tubule even on a high-salt diet, with sodium retention and raised blood pressure. In cultured cells they found the step between: low plasma potassium hyperpolarizes the distal cell, intracellular chloride falls, and chloride-sensitive WNK kinases activate NCC, conventionally through the kinase SPAK. A later inducible deletion of the channel subunit Kir4.1 in adult mouse kidney caused salt wasting, low blood pressure and marked potassium wasting, and the distal cells no longer responded to plasma potassium. The distal convoluted tubule thus reads plasma potassium as a voltage: on the authors’ account, when potassium is low, more sodium is taken back upstream, less sodium and water reach the secretory cells and potassium secretion falls. In humans the evidence includes inherited Kir4.1 disorders, controlled potassium interventions and urinary vesicle measurements, while the voltage and chloride steps rest on experimental models.

Design problem

Brown, Brown and Murphy infused epinephrine into six healthy volunteers and plasma potassium fell by 0.82 meq per L; in a second experiment at half that infusion rate a selective beta2-adrenergic antagonist prevented the fall. Extend the study. First, predict plasma potassium when epinephrine is given after placebo pretreatment, after a beta1-selective antagonist and after a beta2-selective antagonist. Second, design measurements that would show where the potassium went. Third, state what the design cannot show.

One way to do it

Predictions. After placebo pretreatment the fall appears, since that is the epinephrine response itself, and under beta2 block it is prevented, as reported. Under beta1 block it should largely persist, because the receptor carrying the effect here was beta2; that arm is a prediction, not a result.

Design. Use a crossover in the same volunteers, with a saline infusion as the no-epinephrine control. Collect urine over the same minutes as each infusion: urinary potassium unchanged while plasma potassium falls supports a shift into cells rather than a loss from the body, once intake, extracellular volume and non-renal losses are controlled. To locate the uptake, sample arterial and deep forearm venous blood across resting muscle for venous potassium below arterial potassium, and biopsy muscle for Na+-K+-ATPase activity, which tests mechanism but not net uptake. Measure insulin and glucose, since insulin also moves potassium, and record heart rate.

Limits. Six volunteers is a small sample, and plasma measurements alone cannot say which tissue took the potassium up, so the muscle claim stays an inference. Blocking a receptor shows it is required in this preparation, not that beta2 stimulation sets plasma potassium minute by minute in ordinary life.