The most common electrolyte disturbance in hospitalized patients — and the one most often mismanaged. A physiology-first approach from osmoregulation to avoiding osmotic demyelination.
Hyponatremia is rarely a sodium deficit. It is almost always an excess of free water relative to solute — and the organ that fails is usually the kidney's ability to excrete dilute urine, not the body's ability to hold onto salt.
Serum sodium <135 mEq/L. The most common electrolyte abnormality in hospitalized patients — present in 15–30% of inpatients at some point during admission, and 1–4% on arrival. Independent predictor of mortality across heart failure, cirrhosis, pneumonia, and postoperative cohorts (adjusted OR ≈ 1.3–2.5 for in-hospital death even in mild disease).
The measured serum [Na+] is a ratio: Na mass ÷ plasma water volume. You can be "hyponatremic" with normal total body sodium (dilutional), high total body sodium (hypervolemic hyponatremia of HF/cirrhosis), or low total body sodium (hypovolemic hyponatremia). You cannot tell from the Na value alone.
Plasma tonicity is what cells actually care about. Normal plasma osmolality is 275–295 mOsm/kg, maintained within ≈1% of set-point by the osmoreceptor–ADH–thirst axis. True hyponatremia means hypotonic plasma — cells swell, brain water rises, and symptoms follow a Na-concentration gradient but driven by absolute osmolality.
Urea and ethanol raise measured osmolality but freely cross cell membranes — they add osmoles without generating a tonicity gradient. A uremic patient with Na 128 and BUN 140 has low Na but is not hypotonic and does not need saline.
The Edelman equation (refined in 1958 by Isaac Edelman, MD) is the single most important formula in sodium disorders. It says that serum [Na+] is set by the ratio of exchangeable sodium + potassium to total body water:
This means that losing potassium raises the denominator problem — every mEq of K lost to a thiazide is "as hyponatremic" as a mEq of Na lost. It also explains why potassium repletion alone can dangerously over-correct a chronic hyponatremic patient (the K goes intracellularly with Na pushed out, raising [Na+]).
Hyponatremia carries a dose-dependent mortality signal independent of underlying disease. In the Copenhagen inpatient cohort (n=14,290), even Na 130–134 was associated with adjusted HR 1.38 for 30-day mortality; Na <120 reached HR 3.47.
The association persists after adjustment for comorbidities, suggesting hyponatremia is more than a bystander biomarker — it is likely causal via cerebral edema, falls/fractures (2–4× risk), gait instability, osteoporosis, and attentional deficits even at "mild" Na 128–132.
Patients with chronic Na 125–132 show measurable deficits on attention tests and an ≈2× fall risk. Correction reverses the gait instability. There is no truly benign hyponatremia.
The inpatient distribution differs sharply from outpatient hyponatremia. In hospital, hypotonic IV fluids, SIADH from drugs and post-op stress, diuretics in older patients, and heart failure/cirrhosis decompensations dominate. Primary polydipsia and exercise-associated hyponatremia are rare inpatient causes unless the patient brings them in.
The elderly are disproportionately affected: baseline ADH set-point rises with age, diluting capacity falls, and polypharmacy (SSRIs, thiazides, PPIs) compounds the problem.
Severity (Na level), timing (acute vs. chronic), and volume status. Each axis points to a different threat and a different treatment. The three must be assessed together.
The cutoffs are not arbitrary — they mark inflection points in cerebral edema risk, symptom probability, and (most importantly) the risk of over-correction-induced osmotic demyelination.
| Band | Na (mEq/L) | Risk at this level | Clinical features |
|---|---|---|---|
| Mild 130–134 | 130–134 | Fall risk, subclinical gait instability, attentional deficits | Usually asymptomatic; confused with "incidental" |
| Moderate 125–129 | 125–129 | Nausea, headache, fatigue; ODS risk if corrected fast | H/a, confusion in elderly; can be truly asymptomatic chronic |
| Severe 120–124 | 120–124 | Encephalopathy, cerebral edema, seizures if acute | Altered mentation, vomiting, ataxia |
| Profound <120 | <120 | Seizure, coma, respiratory arrest, herniation (acute) | Medical emergency if symptomatic or acute |
A sodium of 108 that developed over weeks may be less immediately dangerous than a sodium of 126 that developed over hours — because the brain has had time to osmotically adapt. Always interpret severity with the time course.
| Acute (<48 h) | Chronic (>48 h or unknown) | |
|---|---|---|
| Brain volume | ↑↑ (no adaptation) | Normal (organic osmolytes out) |
| Symptoms | Severe at Na 125 | Often mild at Na 115 |
| Main threat | Cerebral edema, herniation | Osmotic demyelination on over-correction |
| Correction speed | Aggressive (100 mL 3% bolus ×1–3) | <10 mEq/L per 24 h (safer: 6–8) |
| Examples | TURP, marathon, MDMA, post-op iatrogenic | SIADH, thiazide, HF, cirrhosis, adrenal |
If duration is unknown, treat as chronic and cap correction at 8–10 mEq/L per 24 h. The exception: severe symptoms (seizure, coma) regardless of chronicity → 3% saline bolus to reverse the acute neurologic threat, then revert to chronic rules.
The historical framework sorts hyponatremia into three volume buckets. In practice, volume assessment by physical exam alone is only ~50% accurate. Use it as a starting hypothesis that labs and response to a fluid challenge confirm or refute.
| Volume state | Clues | Classic causes |
|---|---|---|
| Hypovolemic | Orthostasis, dry mucosa, flat JVP, ↑BUN/Cr ratio, UNa <20 (extra-renal) or >30 (renal loss) | GI loss, diuretic, adrenal, CSW, salt-losing nephropathy |
| Euvolemic | No overt volume signs, normal BUN/Cr, UNa >30, Uosm >100 | SIADH, hypothyroid, glucocorticoid deficiency, polydipsia, beer potomania, reset osmostat |
| Hypervolemic | Edema, ascites, ↑JVP, S3, pulmonary congestion, UNa <20 | Heart failure, cirrhosis, nephrotic syndrome, advanced CKD/AKI |
Before treating any Na <135, confirm the patient is actually hypotonic. The Na probe (ion-selective electrode, direct method) normally reads the Na concentration in plasma water. Old indirect methods and severe protein/lipid excess can distort the result — this is the classical pseudo-hyponatremia.
For every 100 mg/dL glucose above 100, add 1.6–2.4 mEq/L to the measured Na to estimate the "dry" sodium. A Na of 128 with glucose 500 isn't hyponatremia — it's a corrected Na ≈ 135 and a DKA problem. Treat the DKA; the Na will rise on its own as glucose falls.
Severity stratifies immediate risk. <125 = probably needs hypertonic if symptomatic.
Acute = cerebral edema rules. Chronic = ODS rules dominate. When unknown, assume chronic.
Points to etiology and drives whether saline, fluid restriction, or diuresis is the right lever.
Plasma osmolality → urine osmolality → urine sodium → volume exam. In that order. Every hyponatremia work-up should be ordered before the first fluid bag is hung.
Uosm 100–200 doesn't cleanly distinguish anything — think partial ADH, polydipsia with some ADH tone, or recent fluid shift. UNa thresholds are disrupted by active diuretic use (always >30 regardless of volume) — repeat off diuretic when feasible. Fractional excretion of urea (FEUrea) <35% suggests hypovolemia even on diuretic, because urea handling isn't affected by thiazides/loops.
Bartter & Schwartz (1967) criteria, still the standard. All essential criteria must be met:
| # | Essential Criterion |
|---|---|
| 1 | Hyponatremia with hypotonic plasma (Posm < 275) |
| 2 | Inappropriately concentrated urine (Uosm > 100) |
| 3 | Clinical euvolemia |
| 4 | Elevated UNa (> 30) on normal salt & water intake |
| 5 | Normal thyroid & adrenal function |
| 6 | No recent diuretic use |
Low BUN, low uric acid (FEUA > 12%), correction with fluid restriction. FEUA < 12% on fluid-restricted patient should reclassify to hypovolemia or renal salt wasting.
| Test | Why | Normal / Threshold |
|---|---|---|
| Plasma osmolality | Confirm hypotonicity; exclude pseudo / translocational | 275–295 mOsm/kg |
| Urine osmolality | Is ADH suppressed? (<100 = yes) | <100 vs >100 |
| Urine sodium | Volume status in kidney's opinion | <20 vs >30 |
| Serum glucose | Correct for translocation (−1.6 per 100 >100) | Correct before acting |
| TSH / cortisol (AM or Cosyntropin) | Exclude hypothyroidism, adrenal insufficiency | TSH <5; cortisol >18 or stimulate to >18 |
| BUN / creatinine / uric acid | Volume axis & SIADH support (low BUN/UA) | BUN/Cr <10; FEUA >12% |
| Serum K, lipids, total protein | Edelman (K), pseudohyponatremia | — |
Bartter–Schwartz criteria (1967). Check each that applies. All essential criteria must be met to establish SIADH.
Both loops and thiazides raise UNa regardless of volume. Diuretic-induced hyponatremia often shows UNa >30 with clinically hypovolemic patient. Repeat UNa 24–48 h after hold if feasible, or use FEUrea.
Post-SAH, post-neurosurgical, meningitis. Same Uosm >100 and UNa >30 as SIADH, but patient is actually volume-depleted. Treatment diverges: SIADH → restrict; CSW → salt + volume. FEUA and response to saline discriminate.
Pregnancy, quadriplegia, psychiatric, chronic malnutrition. ADH turns on at a lower set-point (say 125 instead of 280). Na is low but stable; urine dilutes with water load. No treatment needed.
Hover or tap any etiology to see diagnostic clues, the mechanism, and treatment divergence. Click to pin — useful when cross-referencing multiple phenotypes.
All SIADH is not the same. The underlying mechanism — ectopic ADH vs. hypothalamic dysregulation vs. drug-induced vs. reset osmostat — changes prognosis and treatment duration.
| Type | Mechanism | Classic settings | Practical impact |
|---|---|---|---|
| Type A Erratic | Unregulated, high ADH release | Small-cell lung cancer (ectopic), other malignancy | Highest urine osm; may need vasopressin antagonist |
| Type B "Leaky" | Constantly elevated but submaximal ADH | CNS disease, infection, idiopathic | Responds to restriction, urea |
| Type C Reset osmostat | ADH regulated but at low [Na] set-point | Pregnancy, quadriplegia, malnutrition, psych | Na stable; does not need treatment |
| Type D Nephrogenic SIAD | Gain-of-function V2 receptor mutation | Rare genetic; ADH is low but kidney acts as if high | Does not respond to vaptans (V2 agonist-like) |
| Drug-induced pharmacologic | Various (see list below) | SSRIs (especially in elderly), carbamazepine, oxcarbazepine, cyclophosphamide, vincristine, MDMA | Withdraw drug; often rapid resolution |
Match the treatment lever to the mechanism. The answer is never "just give saline" — and never "just restrict water." Treatment choice follows directly from volume state and urine osm.
No framework replaces addressing the underlying problem. Hold the thiazide. Replace cortisol. Drain the malignant effusion. Treat the pneumonia. Every "refractory SIADH" is one cause search short.
Indications: severe symptoms (seizure, coma, obtundation, severe vomiting) regardless of duration, or acute (<48 h) Na <125.
0.5–2 mL/kg/h of 3% NaCl, recheck Na q2–4 h. Slower onset, requires more monitoring — the bolus approach has overtaken it in most modern protocols.
<10 mEq/L in 24 h (safer: <8), <18 in 48 h. In high-risk patients (alcoholism, malnutrition, hypokalemia, liver disease, Na <105) → <6–8/24 h, <12–16/48 h.
Works in most SIADH and hypervolemic hyponatremia. Fails when the kidney is so ADH-saturated that even small oral intake overwhelms excretion.
| Ratio | Expected response | Approach |
|---|---|---|
| < 0.5 | Restrict works | < 1 L/day |
| 0.5–1.0 | Partial | < 500–750 mL/day + salt tabs / urea |
| > 1.0 | Restrict alone will fail | Add urea 15–30 g or vaptan |
If UNa + UK > PNa, the patient is excreting electrolyte-free water negatively — even drinking pure water will lower serum Na further. You need a second lever (solute load or vaptan).
Adrogué–Madias (2000) formula: predicts the change in serum [Na] from 1 L of a given replacement fluid. Gives a realistic estimate even in complex cases, but results must be re-checked against measured Na every 2–4 h — not applied blindly.
| Scenario | First-line | Second-line / adjunct | Watch for |
|---|---|---|---|
| Severe symptomatic (any cause) | 3% NaCl 100 mL bolus ×1–3 | DDAVP clamp if concern for over-correction | Don't exceed 6–8 mEq/24 h cap in chronic |
| SIADH — chronic, asymptomatic | Fluid restriction <800 mL/day | Urea 15–30 g, salt tabs + loop, tolvaptan (30-day) | UNa+UK/PNa > 1 predicts restriction failure |
| Hypovolemic (GI loss, 3rd space) | 0.9% NaCl 1–2 L | Treat underlying cause | Beware rapid rise once volume repleted — ADH drops, free water dumps |
| Thiazide-induced | Hold thiazide; low-rate 0.9% NaCl if hypovolemic | K replacement if low (can raise Na via Edelman) | Over-correction common in elderly women |
| Hypervolemic (HF, cirrhosis) | Fluid restriction + loop diuretic | Tolvaptan (SALT-1/2), treat HF/cirrhosis | Do not give isotonic saline — will worsen edema |
| Adrenal insufficiency | IV hydrocortisone 100 mg | 0.9% NaCl for volume | Cortisol replacement alone often enough; may over-correct |
| Hypothyroidism (severe) | Levothyroxine + supportive | Restriction if mild | Only severe/myxedematous hypothyroidism causes clinically relevant hyponatremia |
| Primary polydipsia / beer potomania | Restrict intake | Low-dose saline if severe; watch for rapid auto-correction | Highest over-correction risk — kidney was never the problem |
A 2013 strategy popularized by Sterns et al.: give DDAVP 1–2 μg IV/SC q6–8 h prophylactically alongside hypertonic saline. This suppresses any endogenous free-water diuresis and lets you titrate Na by dialing the 3% NaCl up or down instead of chasing a brisk aquaresis.
Sood et al. (Am J Kidney Dis 2013) prospectively used DDAVP clamping in 25 severe hyponatremia patients. 0 ODS events; correction stayed within target in 92%. Has become the default in modern nephrology/ICU protocols.
Hyponatremia has a smaller but sharp evidence base. Select any trial to see design, primary endpoint, and the practical teaching.
The core evidence base clusters in three eras: (1) observational Sterns work on correction rates in the 1980s–90s that defined ODS risk, (2) the vaptan RCT wave (SALT-1/2, EVEREST, TEMPO) 2006–2012, and (3) modern DDAVP clamp / protocol-driven work 2013–present.
Osmotic demyelination syndrome (ODS) — historically "central pontine myelinolysis" — is the dreaded iatrogenic complication of over-rapid correction of chronic hyponatremia. Understanding the mechanism makes the correction rules intuitive.
During chronic hyponatremia (>48 h), the brain extrudes organic osmolytes (myo-inositol, glutamate, taurine, betaine) to shrink back to normal volume at a low plasma Na. This adaptation takes days. If Na is corrected rapidly, water is osmotically pulled out of brain cells faster than the osmolytes can be re-accumulated. Oligodendrocytes are uniquely vulnerable — they lose water, demyelinate, and the classical "locked-in" pontine picture develops 2–6 days later.
ODS is biphasic. The patient first improves as hyponatremia is corrected — mentation clears, they wake up, they seem well. Then 2–6 days later, a new deficit emerges: dysarthria, dysphagia, quadriparesis, altered mental status, classic "locked-in" picture in severe cases. MRI may be negative for the first 1–2 weeks, lagging the clinical picture.
Negative MRI in the first week does not rule out ODS. If you over-corrected and the patient deteriorates 3–7 days later, treat presumptively and repeat imaging.
| Site | Frequency |
|---|---|
| Central pons (CPM) | ~50% |
| Extrapontine (EPM) — basal ganglia, thalamus, cerebellum | ~30% |
| Combined CPM + EPM | ~20% |
"Trident" sign on axial T2: sparing of corticospinal tracts and ventrolateral pons gives a characteristic three-pronged bright pontine signal.
The "18 in 48, 12 in 24, 6 in 24 high-risk" rule traces to Richard Sterns' seminal observational work in the 1980s. Modern data (MacMillan 2023, Gankam Kengne 2019) support even more conservative caps and show that relowering after over-correction reduces ODS incidence to near-zero in animal models and case series. The fear of relowering is unjustified — over-correction without rescue is what causes ODS.
Classical teaching held that ODS was uniformly catastrophic. Modern data are more nuanced. In a 2011 series (Louis et al., 33 patients), 25% recovered completely, 37% had minor residual, 25% had significant disability, and 13% died. Recovery takes months — don't withdraw care prematurely. Supportive care (rehab, PT/OT, swallow), aggressive treatment of secondary infection/aspiration, and patience are the mainstays. No proven specific therapy; case reports on steroids, IVIG, and plasma exchange remain anecdotal.