Hyponatraemia: A Water Problem, Not a Salt Problem
It is the commonest electrolyte abnormality on the ward, and one of the few where the intuitive treatment can kill. The name says "low sodium," so the reflex is to reach for salt. But a low serum sodium almost never means the body is short of salt — it means there is too much water diluting the salt that is there. Get the reasoning backwards and you either flood a patient already drowning in water, or you correct them so fast that you destroy the brain you were trying to protect. This chapter is about thinking in the right currency: water and volume, not salt.
An 78-year-old woman is brought in drowsy and confused after a fall. Her serum sodium is 118 mmol/L — dangerously low. The nurse, reading "low sodium," asks whether to start salty fluids. But look closer: she was started on a thiazide diuretic three weeks ago, she is not thirsty, her mouth is moist, and her blood pressure barely drops when she sits up. Her body is not short of salt at all — it is holding on to too much water, and the water has diluted her blood. Pour in more fluid too quickly and you court disaster from the opposite direction: correct a chronic low sodium too fast and, days later, she could develop slurred speech, paralysis and a locked-in state from a brain injury you inflicted. Everything now depends on two questions that have nothing to do with salt: how much water is in her, and how fast is it safe to move it.
The founding idea: sodium is a concentration, not an amount
Serum sodium tells you the ratio of salt to water, never the absolute quantity of either. This is the single insight the whole topic turns on. A concentration is a fraction: sodium over water. A number can fall because the numerator (salt) drops, or because the denominator (water) rises — and in real patients it is nearly always the water. The body defends its sodium concentration obsessively through the hormone ADH (antidiuretic hormone, vasopressin), which controls how much water the kidney holds on to. When ADH is high, the kidney reabsorbs free water, dilutes the blood, and sodium falls. So the clinical question is never "has this patient lost salt?" but "why is there relatively too much water, and where is that water sitting?" This is why the whole approach is built on volume status and osmolality — the framework developed in the Fluid status chapter — rather than on the sodium number in isolation.
Step one: is it even true hyponatraemia?
Before classifying anything, confirm you are looking at true, hypotonic hyponatraemia — a genuinely dilute plasma. Two impostors must be excluded first. The first is pseudohyponatraemia: a laboratory artefact in which very high blood lipids or proteins (severe hyperlipidaemia, paraproteinaemia) crowd the plasma sample so the reported sodium reads low while the real concentration in the water phase is normal — the plasma osmolality is normal, which gives it away. The second is translocational (hyperglycaemic) hyponatraemia: in uncontrolled diabetes, glucose is osmotically active and pulls water out of cells into the blood, diluting sodium — here the measured osmolality is high, not low. Correcting the glucose corrects the sodium; giving salt would be wrong. Only once osmolality confirms the plasma is truly hypotonic do you proceed to the volume classification. This translocational trap is developed further in the Fluid status chapter.
Step two: classify by volume status — the three buckets
Once the sodium is truly dilute, one clinical judgement — is the patient dry, normal, or overloaded? — splits the whole differential. HYPOVOLAEMIC hyponatraemia — the patient has lost BOTH salt and water (vomiting, diarrhoea, diuretics — especially thiazides — Addison's disease and other salt-wasting states), and the water has been partly replaced by drinking or by hypotonic fluids. The body is truly volume-deplete, so it senses low circulating volume, releases ADH to hold on to water, and the sodium falls. Clinically they look dry: postural drop, dry mucous membranes, low urine sodium (the kidney is avidly retaining sodium). Treatment is 0.9% ("normal") saline to refill the tank, plus fixing the cause. EUVOLAEMIC hyponatraemia — volume looks normal, and the classic cause is SIADH (syndrome of inappropriate ADH): ADH is secreted when it shouldn't be, so the kidney retains water for no physiological reason. Also in this bucket: hypothyroidism, glucocorticoid (cortisol) deficiency, and drug-induced SIADH. HYPERVOLAEMIC hyponatraemia — the patient is visibly overloaded (oedema, ascites): total-body sodium is actually increased, but total-body water is increased even more, so the ratio still falls. The causes are the big organ-failure states: heart failure, cirrhosis, and nephrotic/renal disease.
SIADH and the drugs that cause it
SIADH deserves its own paragraph because it is both the classic euvolaemic cause and, very often, iatrogenic. The physiology is simple: ADH is released inappropriately — from a tumour (classically small-cell lung cancer), from lung or brain pathology, or from a drug — the kidney retains free water, and the plasma dilutes while the urine stays inappropriately concentrated. The drug list is worth memorising because these are everyday prescriptions: SSRIs (antidepressants — a very common culprit in the elderly), carbamazepine, antipsychotics, cyclophosphamide, and again the thiazide diuretics. This is where hyponatraemia connects to the Central nervous system and Toxicology sections (SSRIs and carbamazepine), and to the Endocrine section (the whole ADH/vasopressin axis, plus hypothyroidism and cortisol deficiency as mimics). The management of true SIADH is fluid restriction first-line — you cannot pour water into a system that refuses to excrete it — together with treating the cause or, crucially, stopping the offending drug.
Picture a glass of squash — cordial diluted with water. If the drink tastes too weak, there are two ways to read it. The naive fix is "add more cordial" (give salt). But usually the real problem is that someone kept topping up the glass with plain water — the cordial is fine, there is simply too much water. The cure is to stop adding water and let some evaporate off (restrict fluid, or make the kidney pee out free water), not to keep dumping in syrup. And there is a catch the analogy captures perfectly: if you have let the glass sit weak for a long time and then suddenly concentrate it all at once, you shock the system. In the body that shock is the brain, and it is far more dangerous than the weak drink ever was.
The volume-status framework tells you the treatment before you name a single drug. Hypovolaemic → the tank is empty, so give 0.9% saline. Euvolaemic/SIADH → the kidney is hoarding water, so restrict fluid (and stop the drug). Hypervolaemic → there is already too much salt and water, so restrict both and offload with a loop diuretic while treating the failing organ. Three buckets, three opposite fluid strategies — which is exactly why "just give salt" is wrong two times out of three, and can be actively harmful in the overloaded patient.
The drugs and fluids that treat it
Beyond fluids in or fluids out, a small pharmacology targets the water-handling machinery itself. FLUID RESTRICTION is the unglamorous first-line for euvolaemic (SIADH) and a mainstay in hypervolaemic states — simply limiting water intake below the amount the kidney can excrete lets the sodium drift up on its own. When that is not enough, the VASOPRESSIN RECEPTOR ANTAGONISTS — the "vaptans," chiefly tolvaptan — block the V2 receptor in the collecting duct, which is precisely where ADH tells the kidney to reabsorb water. Blocking it produces an aquaresis: the kidney excretes electrolyte-free water, sodium rises, and volume is offloaded. Vaptans are used in SIADH and in hypervolaemic hyponatraemia, but they carry two serious cautions — hepatotoxicity (tolvaptan can injure the liver), and the risk of over-rapid correction, which is why they must be started in hospital with close monitoring. Older and now rarely used is DEMECLOCYCLINE, a tetracycline that treats SIADH by a side effect — it induces a nephrogenic diabetes insipidus, making the kidney partly resistant to ADH so it wastes water. Finally, for SEVERE, symptomatic hyponatraemia (seizures, coma), the drug is HYPERTONIC (3%) SALINE, given in small, controlled boluses to lift the sodium just enough to stop the brain swelling — never as an open-ended infusion.
Vomiting/diarrhoea or thiazide, patient looks dry → hypovolaemic → 0.9% saline + treat/stop the cause. Small-cell lung cancer or an SSRI, patient euvolaemic with concentrated urine → SIADH → fluid restriction, stop the drug, add tolvaptan if refractory. Heart failure or cirrhosis with oedema/ascites → hypervolaemic → fluid and salt restriction + a loop diuretic (furosemide) + treat the failing organ. A marathon runner who collapses seizing after drinking litres of water → acute water intoxication → cautious hypertonic (3%) saline. Same low number on the lab report; four completely different responses — decided entirely by volume status and how fast the sodium fell.
- Confirm true hypotonic hyponatraemia first — exclude pseudohyponatraemia (normal osmolality) and hyperglycaemic/translocational (high osmolality).
- Low sodium usually means relative water excess, driven by ADH — think water and volume, not salt.
- Classify by volume: hypovolaemic (dry), euvolaemic (SIADH), or hypervolaemic (overloaded).
- Hypovolaemic → 0.9% saline; euvolaemic/SIADH → fluid restriction + stop the drug; hypervolaemic → fluid/salt restriction + loop diuretic.
- Drug-induced SIADH: SSRIs, carbamazepine, antipsychotics, cyclophosphamide, thiazides.
- Vaptans (tolvaptan) block V2 for an aquaresis; hepatotoxic and hospital-started; demeclocycline induces nephrogenic DI.
The cardinal safety rule: correct chronic hyponatraemia slowly
The brain adapts to a slow-falling sodium — and that adaptation is exactly what makes fast correction lethal. When sodium falls gradually over days, the brain protects itself: its cells extrude solutes so that water does not flood in, and the brain shrinks back to near-normal size despite the dilute blood. This clever adaptation sets a trap. If you now raise the serum sodium quickly, the blood becomes suddenly hypertonic relative to those adapted brain cells, water is dragged out of them, and they shrivel — producing osmotic demyelination syndrome (central pontine myelinolysis), a devastating and often irreversible iatrogenic injury: days later the patient develops dysarthria, quadriparesis, and in the worst cases a locked-in state. The rule that prevents it is a number worth memorising: correct chronic hyponatraemia no faster than about 8–10 mmol/L in 24 hours (lower still in high-risk patients — the malnourished, alcoholics, hypokalaemic). Monitor the sodium closely, and if it is rising too fast, actively re-lower it with electrolyte-free water (5% dextrose) and/or desmopressin (synthetic ADH) to reapply the brake. Slow is safe.
There is one important exception. ACUTE hyponatraemia — sodium that has crashed within hours, before the brain has had time to adapt — is the mirror image. Here the danger is not over-correction but the acute cerebral oedema itself: with no protective solute extrusion yet, water floods brain cells and they swell inside the rigid skull, causing seizures, brainstem herniation and death. This is the picture in the marathon runner or the person who drinks enormous volumes of water, and in MDMA (ecstasy) toxicity, where the drug both stimulates ADH and drives compulsive water drinking — a link back to the Toxicology section. Because the brain has not adapted, prompter correction with hypertonic saline is appropriate and even life-saving. The art is telling acute from chronic: when in doubt, or when the timeline is unknown, treat it as chronic and go slow — the cost of being too cautious is far smaller than the cost of osmotic demyelination.
- Chronic hyponatraemia: correct SLOWLY — aim ≤ 8–10 mmol/L per 24 h; over-rapid rise causes osmotic demyelination (central pontine myelinolysis).
- The brain adapts to chronic low sodium by extruding solutes — which is exactly why fast correction shrinks and injures it.
- If sodium overshoots, actively re-lower with 5% dextrose ± desmopressin.
- Acute hyponatraemia (marathon, water intoxication, MDMA) causes cerebral oedema — the exception where prompter correction is needed.
- Severe/symptomatic hyponatraemia (seizures, coma) → controlled boluses of hypertonic (3%) saline, not open infusion.
- When acute vs chronic is unclear, treat as chronic and correct slowly.
- Reaching for salt because the label says "low sodium." In SIADH and in the fluid-overloaded patient, giving saline can worsen the picture — the problem is water, not a salt deficit.
- Correcting a chronic low sodium too fast. Exceeding ~8–10 mmol/L in 24 h risks osmotic demyelination (central pontine myelinolysis) — a devastating, often irreversible injury inflicted by the treatment.
- Skipping the osmolality check and missing pseudohyponatraemia or hyperglycaemic hyponatraemia — treating a low number that isn't truly dilute plasma.
An 80-year-old woman started on a thiazide 3 weeks ago is admitted confused with sodium 116 mmol/L, present for at least several days. She is clinically euvolaemic. Her sodium is corrected with 0.9% saline and rises by 16 mmol/L over 24 hours. Four days later she develops dysarthria and quadriparesis. What happened?
- A low serum sodium is usually relative WATER excess (high ADH), not a salt deficit — think in volume and osmolality, not "give salt."
- Confirm true hypotonic hyponatraemia, then classify: hypovolaemic → 0.9% saline; euvolaemic/SIADH → fluid restriction + stop the drug; hypervolaemic → fluid/salt restriction + loop diuretic.
- Targeted drugs: vaptans (tolvaptan) block V2 for an aquaresis (hepatotoxic, hospital-started); demeclocycline induces nephrogenic DI; severe/symptomatic disease needs controlled 3% hypertonic saline.
- The cardinal rule: correct CHRONIC hyponatraemia slowly (≤ 8–10 mmol/L per 24 h) or risk osmotic demyelination; acute hyponatraemia (water intoxication, MDMA) causes cerebral oedema and is the exception needing prompter correction.
- Rang & Dale's Pharmacology — The kidney and body fluids; vasopressin and its antagonists.
- Katzung — Basic & Clinical Pharmacology: Agents affecting water excretion; ADH antagonists.
- BNF — Tolvaptan, demeclocycline, and fluid/electrolyte preparations.
- UK Kidney Association (Renal Association) — Clinical Practice Guideline on the Investigation and Management of Hyponatraemia in Adults.
- Spasovski G, et al. Clinical practice guideline on diagnosis and treatment of hyponatraemia (European Society of Endocrinology / ESICM / ERBP).
- Ganong's Review of Medical Physiology — Regulation of extracellular fluid osmolality and volume; vasopressin.
