ADH keeps flowing at Na 112 — no osmotic brake left
Brain adaptation — why correction, not hyponatraemia, is the killer
1–3 h: extrudes Na⁺, K⁺, Cl⁻. 24–48 h: exports organic osmolytes. Refilling takes 5–7 days — correct faster than that and you get demyelination. 48 h is the pivotal number.
Verbalis JG, Gullans SR. Brain Res. 1991;567:274 · Sterns RH, Silver SM. J Am Soc Nephrol. 2008;19:194 · Ronco Ch 56 · Koyner Ch 19 · Overgaard-Steensen C, Ring T. Crit Care. 2013;17:206
The urine-electrolyte toolkit: what the kidney is doing, in two numbers
EFWC = urine volume × [1 − ( U_Na + U_K )/serum Na] · Furst ratio = ( U_Na + U_K )/serum Na
Furst decides whether restriction can work.
SIADH — the kidney generates water
Na 118 · U_Na 110 · U_K 40
EFWC −0.27 L/d · Furst 1.27
The kidney adds pure water.
Heart failure — clearing, not enough
Na 128 · U_Na 12 · U_K 28 · UO 0.9 L/d
EFWC +0.62 L/d · Furst 0.31
Restriction will work here.
Furst thresholds at the bedside
<0.5 — restriction to ~1 L/day works 0.5–1.0 — needs ~500 mL/day, rarely tolerated >1.0 — restriction cannot work; go to solute or a vaptan
Koyner Ch 19 (Eq 19.2) · Ronco Ch 56 (Eq 3) · Rose BD. Am J Med. 1986;81:1033 · Furst H et al. Am J Med Sci. 2000;319:240 · Winzeler B et al. J Intern Med. 2016;280:609
02
Hyponatraemia: making the diagnosis
Three questions in order — is the sodium real, is the serum hypotonic, is ADH on — before any question about volume status.
Koyner Ch 19 · Ronco Ch 56, 62
Step 1 — is the sodium real, and is the serum actually hypotonic?
Pseudohyponatraemia — a lab artefact
Indirect ISE (main lab) is fooled by lipids and paraprotein. Direct ISE (blood gas) measures plasma water — immune.
Free bedside test: same blood on the gas analyser.
Translocational — real dilution, high tonicity
Glucose and mannitol draw water out of cells. The patient is hypertonic — never give hypertonic saline for the number.
The glucose correction: 1.6 or 2.4?
Katz: 1.6; Hillier: 2.4 overall, ~4.0 above glucose 400. It predicts where Na lands after insulin — not a target.
Katz MA. N Engl J Med. 1973;289:843 · Hillier TA et al. Am J Med. 1999;106:399 · Rohrscheib M et al. Am J Med Sci. 2015;349:537 · Koyner Ch 19 (Eq 19.4) · Ronco Ch 56 · DKA: Lecture 07
Step 2 — urine osmolality first, volume status second
Urine osmolality<100 = ADH suppressed. >100 = ADH is driving it.
→
03
Volume statusWeight trend, fluid chart, POCUS — not skin turgor.
→
04
U_Na + urate<20–30 = hypo- or hypervolaemic. >40 with urate <4 = SIADH.
Four documented failure modes
Diuretics raise U_Na in true hypovolaemia — the discriminator is abolished
CKD fixes urine osmolality near isotonic — neither extreme is interpretable
Recent saline rewrites both numbers; exam gets volume right ~50% of the time
Urate <4 is sensitive for SIADH but low in any ECF expansion
Koyner Ch 19 (Fig 19.3) · Chung HM et al. Am J Med. 1987;83:905 · Fenske W et al. Am J Med. 2010;123:652 · Spasovski G et al. Eur J Endocrinol. 2014;170:G1
SIADH, its mimics, and the neuro-ICU's favourite argument
Essential: osm <275, urine osm >100, U_Na >40, euvolaemia, no diuretic
Supporting: urate <4, BUN <10, no response to 0.9% saline
Exclude adrenal insufficiency: cortisol >500 nmol/L makes it unlikely
Water excretion = solute load / minimum U_osm. Both autocorrect explosively.
Scale & asymmetry
Under-treating acute kills by oedema in hours; over-treating chronic disables by demyelination in days.
The 0.9% saline challenge
Hypovolaemia: saline raises serum and urine Na. SIADH: urine Na rises, serum Na falls.
Spasovski G et al. Eur J Endocrinol. 2014;170:G1 · Renneboog B et al. Am J Med. 2006;119:71.e1 · Hoorn EJ et al. J Bone Miner Res. 2011;26:1822 · Koyner Ch 19 · Ronco Ch 56
03
Treating hyponatraemia without causing the second injury
Symptoms set the urgency, risk factors set the ceiling, and a sodium measured every two hours beats every formula ever published.
Koyner Ch 19 · Ronco Ch 56, 62
Symptoms decide the first hour — the number does not
3% saline bolus now — before the workup is complete
Moderately severe
Nausea, confusion, headache, somnolence
One 150 mL 3% bolus, then a hard 24-h ceiling
Absent
No attributable neurological signs
No hypertonic saline; remove the cause
"Asymptomatic" in a sedated patient is an assumption, not an observation
No threshold sodium triggers symptoms — the rate and the relative fall matter. An alert patient at 108 has adapted: treat that number as evidence of chronicity.
The acute (<48 h) hyponatraemias — the danger is under-treatment
Exercise-associated
Hypotonic overdrinking + AVP that will not switch off. Treat: 100 mL 3% every 10 min until encephalopathy lifts.
Post-operative
Hypotonic maintenance + surgical ADH. Arieff: 15 healthy women, mean Na 108 — all seized, 27% died.
Water loading & irrigant
MDMA, psychogenic polydipsia; TURP glycine absorption — a hypotonic irrigant, and glycine is an ineffective osmole.
Mirror image of everything else
These brains are swollen, not adapted — the hazard is herniation in hours. Correct promptly; overshooting the ceiling is the smaller risk.
Hew-Butler T et al. Clin J Sport Med. 2015;25:303 · Hew-Butler T et al. Front Med. 2017;4:21 · Arieff AI. N Engl J Med. 1986;314:1529 · Ronco Ch 56 · Koyner Ch 19
The 3% saline protocol — a bolus drug, not an infusion
01
Bolus100–150 mL 3% NaCl over 10–20 min; peripheral is fine.
→
02
ReassessSymptoms at 5–20 min; Na after the bolus.
→
03
Repeat ≤ 3×Stop at +4–6 mEq/L or symptom resolution.
→
04
Stop and thinkSymptoms persisting after +10 are not the sodium.
SALSA · bolus vs infusion
Overcorrection 17.2% vs 24.2% (NS), with less relowering — boluses are at least as safe, and faster.
Why 4-6 mEq/L is the whole target
A ~5 mmol/L rise cuts ICP by ~50% within an hour and aborts seizures. Beyond that, only demyelination risk.
Baek SH et al. JAMA Intern Med. 2021;181:81 (SALSA) · Spasovski G et al. Eur J Endocrinol. 2014;170:G1 · Koenig MA et al. Neurology. 2008;70:1023 · Ronco Ch 56 (Fig 56.3)
Worked example — Adrogué-Madias, and why formulae under-read
Adrogué-Madias: ΔNa per L = (infusate [Na⁺ + K⁺] − serum Na⁺)/(TBW + 1) · 3% = 513 · D5W = 0
It assumes a closed system — what a hyponatraemic patient stops being once ADH switches off. Use it for the first dose, then measure Na q2h.
58 kg woman, post-op, Na 110, generalised seizure
TBW = 29 L
dNa/L of 3% = (513 − 110)/30 = 13.4 mEq/L
150 mL bolus = +2.0 · three boluses = +6.0
The documented failure mode
Mohmand: below Na 120 the observed rise exceeded prediction in 74%; overcorrectors rose 2.4×. The formulae under-read in small patients.
Adrogué HJ, Madias NE. N Engl J Med. 2000;342:1581 · Barsoum NR, Levine BS. Nephrol Dial Transplant. 2002;17:1176 · Mohmand HK et al. Clin J Am Soc Nephrol. 2007;2:1110 · Hanna RM et al. Clin Kidney J. 2016;9:530 · Koyner Ch 19 (Eq 19.3)
The 24-hour ceiling — and why it is usually the patient who breaches it
Patient
Target / 24 h
Absolute ceiling
Trigger to relower
Standard risk
4–8 mEq/L
8 mEq/L
>8 in 24 h
High ODS risk: Na <105, K <3, alcohol, malnutrition, liver disease
4–6 mEq/L
6 mEq/L
>6 in 24 h
Documented acute + severe symptoms
Abort symptoms
Still 8
>8 in 24 h
The aquaresis trap — overcorrection is delivered by the kidney, not the pump
The ADH stimulus disappears → urine jumps to 300–500 mL/h and Na climbs ~2 mEq/L/h with nothing running. Dilute urine >100–150 mL/h is the earliest alarm.
Spasovski G et al. Eur J Endocrinol. 2014;170:G1 · Verbalis JG et al. Am J Med. 2013;126(10 Suppl 1):S1 · Sterns RH et al. Semin Nephrol. 2009;29:282 · Ronco Ch 56
Desmopressin, used twice: rescue after the fact, or a clamp from the start
dNa per litre of D5W = (0 − serum Na)/(TBW + 1)
Give the water and stop the diuresis together — D5W alone simply feeds the urine bag.
Rescue — 70 kg man, Na 108 to 121 in 14 h
dNa/L D5W = (0 − 121)/43 = −2.8 mEq/L
D5W 10 mL/kg over 1 h = −2.0; repeat to Na ≤116
Desmopressin 2–4 µg IV, then 1–2 µg q6–8h.
Pre-emption — the DDAVP clamp
1. Stop maintenance fluid. 2. Desmopressin 2 µg IV q8h. 3. 3% NaCl 1–1.5 mL/kg over 6 h — the rise becomes linear. 4. Titrate to <8/day. 5. Stop both at Na 125–130.
Sood L et al. Am J Kidney Dis. 2013;61:571 · Perianayagam A et al. Clin J Am Soc Nephrol. 2008;3:331 · Gankam Kengne F et al. Kidney Int. 2009;76:614 · Soupart A et al. Clin Nephrol. 1999;51:383 · Koyner Ch 19 (Table 19.3)
Osmotic demyelination syndrome
Mechanism: raise tonicity faster than the osmolyte transporters refill the cell
Biphasic: improves as Na rises, deteriorates 2–6 days later — the delay hides the mistake
Imaging: MRI can be normal for 1–2 weeks — a negative scan never excludes it
Prognosis: not uniformly fatal — recovery over months is common
It is not only a hyponatraemia disease
Any large, fast rise in effective osmolality — including salt poisoning
Occurs without hyponatraemia in advanced liver disease and after transplant
Reported after correcting documented acute hyponatraemia too
King JD, Rosner MH. Am J Med Sci. 2010;339:561 · Shah MK et al. Am J Kidney Dis. 2018;71:436 · Gankam Kengne F, Decaux G. Nephron. 2018;140:39 · Ronco Ch 56 (Fig 56.2) · Koyner Ch 19
Equipoise: is the correction-rate limit still defensible in 2026?
MacMillan · NEJM Evid 2023
22,858 hospitalisations; rapid correction in 17.7%; ODS in only 12 — and 7 of those never corrected rapidly.
Suppadungsuk · Kidney Med 2025
26,710 patients: ODS 0.73% after rapid correction vs 0.10% without — OR 3.16. 89% of cases had a predisposing risk factor.
The synthesis — risk-stratify, do not deregulate
Keep the limit where risk factors exist; don't treat a benign overshoot in a low-risk patient as a catastrophe.
MacMillan TE et al. NEJM Evid. 2023;2:EVIDoa2200215 · Suppadungsuk S et al. Kidney Med. 2025;7:100953 · See XY et al. J Crit Care Med. 2024;10:209 · Reis LAC et al. J Bras Nefrol. 2026;48(1):e20250161 · Sterns RH, Rondon-Berrios H. Am J Kidney Dis. 2026;87:115
Chronic SIADH: raise the solute load, or block the receptor
1
Treat the cause, stop the drug — in self-limiting SIADH this is the whole treatment.
2
Fluid restriction below urine output — fails when Furst >1.
3
Salt tablets + loop diuretic — solute excreted in more water.
4
Urea 15–30 g/day — 15 g = 250 mOsm of filtered solute. Safe in SIADH, HF, cirrhosis.
SALT-1/2 + EVEREST — why the vaptan role keeps shrinking
SALT: tolvaptan raised Na, but the benefit was lost after withdrawal. EVEREST: no effect on mortality. It corrects the number, and nothing else.
Schrier RW et al. N Engl J Med. 2006;355:2099 · Konstam MA et al. JAMA. 2007;297:1319 · Rondon-Berrios H et al. Clin J Am Soc Nephrol. 2018;13:1627 · Sterns RH et al. Kidney Int. 2015;87:268 · Ronco Ch 62
Quick poll
63 F, alcohol use disorder, admitted vomiting. Na 104, K 2.8, drowsy but rousable. U_osm 90, U_Na 12. She received 2 L of 0.9% saline and 40 mmol KCl. Four hours later: Na 113, urine 400 mL/h, U_osm 68.
A. Continue saline · B. Give 3%; 113 is still dangerously low · C. Stop saline; D5W plus desmopressin; bring Na back down · D. Stop saline and restrict; ODS is rare · E. Start tolvaptan
Hands up — then name her ODS risk factors out loud, and say what the KCl contributed
MacMillan TE et al. NEJM Evid. 2023;2:EVIDoa2200215 · Suppadungsuk S et al. Kidney Med. 2025;7:100953 · See XY et al. J Crit Care Med. 2024;10:209 · Sterns RH, Rondon-Berrios H. Am J Kidney Dis. 2026;87:115 · Koyner Ch 19
04
Hypernatraemia
Always a failure to drink — and in a sedated, intubated patient that failure is ours, which is why it reads as a marker of care quality.
Koyner Ch 20 · Ronco Ch 56
Less feared than hyponatraemia, more lethal — three mechanisms, one on the drug chart
Extrarenal water loss
Fever, burns, open abdomen, diarrhoea, lactulose, inadequately humidified ventilation.
8.4% bicarbonate is 1000 mmol/L; plus saline flushes, hypertonic saline for ICP, drug diluents.
The audit before the arithmetic
6–25% of ICU patients, mostly ICU-acquired. Weight loss = dehydration; weight gain = sodium excess — and the two treatments are opposite.
Palevsky PM et al. Ann Intern Med. 1996;124:197 · Lindner G et al. Am J Kidney Dis. 2007;50:952 · Darmon M et al. Nephrol Dial Transplant. 2010;25:2510 · Polderman KH et al. Crit Care Med. 1999;27:1105 · Koyner Ch 20 · Ronco Ch 56
Worked example — the deficit is only half the prescription
Water deficit (L) = weight × k × (serum Na/140 − 1); k = 0.6/0.5 men/women, 0.45 elderly women
Ongoing loss = urine volume × [1 − (U_Na + U_K)/serum Na]
The formula is a snapshot — no ongoing losses, no sodium input.
68 y woman, 60 kg, febrile; Na 162; urine 2.2 L/d
Deficit to 140 = 4.2 L — a two-to-three-day plan. Day-1 fall 10 = 1.8 L + EFWC 1.25 L/d + insensible 0.5 L = 3.6 L.
The rule of thumb, and when to distrust it
Shortcut: ignore the first 1 L of urine, replace half of 1–3 L, all above 3 L. Fails with dilute urine.
Route: enteral water ideal; D5W if the gut is unavailable
Koyner Ch 20 (Eq 20.1-20.2) · Adrogué HJ, Madias NE. N Engl J Med. 2000;342:1493 · Ronco Ch 56 (Eq 3) · Overgaard-Steensen C, Ring T. Crit Care. 2013;17:206
How fast? The evidence here is thinner than the dogma
What consensus says
No faster than 0.5 mmol/L/h, or 12 mmol/L/day, if chronic or of unknown duration
Documented acute (<48 h) may be corrected faster
Raised ICP: ~4 mmol/L/24 h
Fix the cause in parallel: hyperglycaemia, mannitol, SGLT2i, lithium
What the data say
Chauhan: no harm above 0.5 mmol/L/h in adult ICU cohorts
The glucose loop — and the fluid that is not as isotonic as it looks
Large-volume D5W → hyperglycaemia → osmotic diuresis → more hypernatraemia. With intracranial pathology use 0.9% saline; Ringer's is mildly hypotonic.
Chauhan K et al. Clin J Am Soc Nephrol. 2019;14:656 · Sterns RH. Clin J Am Soc Nephrol. 2019;14:645 · Kitisin N et al. J Crit Care. 2025;87:155012 · Pattamin N, Chuasuwan A. J Crit Care. 2026;92:155354 · Koyner Ch 20 · Ronco Ch 56
Diabetes insipidus: classify it, dose it, and do not test it the textbook way
Condition
Baseline U_osm
Response to desmopressin
ICU causes
Complete central DI
<290 mOsm/kg
Rises >200
Neurosurgery, TBI, SAH
Partial central DI
400–500 mOsm/kg
Rises ~200
Incomplete pituitary injury
Nephrogenic DI
Inappropriately low
No change
Lithium, hypercalcaemia, amphotericin
Diagnosis in a patient who cannot drink
No water deprivation test — the dilute-urine hypernatraemic patient has already performed one. Polyuria >4 mL/kg/h with U_osm <300; confirm with desmopressin.
Dosing, and the trap after pituitary surgery
Desmopressin IV 1–2 µg bid. Triphasic: DI → SIADH at days 5–10 → permanent DI. A standing order causes hyponatraemia in phase 2.
Koyner Ch 20 (Tables 20.1-20.2) · Ronco Ch 56 (Table 56.3, Box 56.2) · Fenske W et al. N Engl J Med. 2018;379:428 · Boton R et al. Am J Kidney Dis. 1987;10:329 · Garofeanu CG et al. Am J Kidney Dis. 2005;45:626 · Garrahy A et al. Clin Endocrinol. 2019;90:23
The other reason we make people hypernatraemic: tonicity as a drug
Raising plasma tonicity ~5 mOsm/kg draws water out of brain — ICP falls within 20–30 min
Targeting Na 150–155 for days is common with weak support — prescribed hypernatraemia
In a neuro-ICU patient, Na 155 may be deliberate — ask before writing free water
COBI
370 TBI patients, continuous 20% hypertonic saline vs standard care: no difference in 6-month GOS-E. Prophylactic HTS does not improve outcome; rescue remains standard.
Roquilly A et al. JAMA. 2021;325:2056 (COBI) · Schwimmbeck F et al. Neurocrit Care. 2023;38:456 · Koyner Ch 20, 48 · Ronco Ch 56 · ICP: Lecture 08
05
Dysnatraemia meets the machine
Dialysis is a tonicity intervention whether or not you prescribed it as one — and at the extremes the machine, not the physician, sets the correction rate.
Koyner Ch 20, 48 · Ronco Ch 56, 133
Dialysis disequilibrium: the mechanism, and the deliberately bad prescription that prevents it
Reverse urea hypothesis: urea leaves plasma in minutes but brain cells over 12–24 h
Late in or after the session: headache, nausea → encephalopathy, seizures, coma
At risk: first-ever treatment, BUN >175 mg/dL, children and elderly, severe acidosis
Mid-session encephalopathy: stop dialysis, secure the airway, hypertonic saline
Prevention — deliberately under-dialyse the first session
1. Blood flow 50–200 mL/min, small dialyser. 2. Short first session (~2 h), repeated daily. 3. Dialysate Na 143–146. 4. Mannitol in hour 2. 5. Or choose CRRT.
Koyner Ch 48 (Tables 48.2-48.3) · Mistry K. Int J Nephrol Renovasc Dis. 2019;12:69 · Patel N et al. Semin Dial. 2008;21:493 · Ronco Ch 133 · Modality: Lecture 08
Extreme dysnatraemia on CRRT: the sodium arithmetic
Hypernatraemia — raise the circuit sodium
Na 175 against a standard 140 bath drops far faster than 10 mEq/L/day. Add 3% NaCl: a 5 L bag needs 360 mL to reach 165 mEq/L. Step the target down 8–10 mEq/L/day.
Hyponatraemia — lower the circuit sodium
Na 104 against 140 corrects 20–30 mEq/L in hours. Dilute with sterile water: a 5 L bag needs 930 mL to reach 118 mEq/L.
Why CRRT, not intermittent haemodialysis
Machine dialysate sodium adjusts only over ~130–150 mEq/L. CRRT compounds the bath and spreads correction over 24 h. Write a target per 24-h block, not a rate.
Koyner Ch 19–20, 48 · Ronco Ch 56, 133 · Yessayan L et al. Am J Kidney Dis. 2014;64:305 · CRRT prescription: Lectures 08–09
Quick poll
48 M, undiagnosed CKD, obtunded. Na 108, K 6.9, urea 61 mmol/L (BUN 171), HCO3 9, pulmonary oedema, anuric. He needs kidney replacement therapy tonight.
A. Standard 4 h IHD, dialysate Na 140 · B. 2 h IHD, blood flow 150, dialysate Na 145 · C. CVVHDF with fluid compounded to Na ~118, stepped up 6–8/day · D. Hypertonic saline to Na 120, then IHD · E. Medical management only
Two hazards are running at once — name both before you vote
Koyner Ch 48 (Tables 48.2-48.3) · Mistry K. Int J Nephrol Renovasc Dis. 2019;12:69 · Yessayan L et al. Am J Kidney Dis. 2014;64:305 · Ronco Ch 133 · CRRT prescription: Lectures 08–09
Case 1 · Day 2 after hip arthroplasty — treat first, then read the trap
Treat3% NaCl 150 mL over 20 min, before the workup; repeat ×2, stop at +5.
→
02
Stop the causeStop dextrose-saline, treat nausea, review the opioid
→
03
AnticipateAquaresis when pain settles; dilute urine >150 mL/h
→
04
CapNa q2h; ceiling 6 in 24 h
The part everyone forgets
K 3.2 → 4.0 is +1.5 mEq/L Na before any saline — give the potassium, then subtract its effect.
Spasovski G et al. Eur J Endocrinol. 2014;170:G1 · Adrogué HJ, Madias NE. N Engl J Med. 2000;342:1581 · Koyner Ch 19
Case 2 · Fluid restriction that is going nowhere
74 M, small-cell lung carcinoma, ventilated. Day 5 of a 1 L/day fluid restriction. Euvolaemic; no diuretic; TSH normal; cortisol 690 nmol/L.
Na 122S osm 258U osm 610U Na 115U K 35Urine 900 mL/dUrate 1.4FE urate 18%
Do the arithmetic that tells you why this restriction was never going to work — then prescribe, and say what a cortisol of 320 would change.
Think 60 seconds — one calculation, then a plan
Spasovski G et al. Eur J Endocrinol. 2014;170:G1 · Verbalis JG et al. Am J Med. 2013;126(10 Suppl 1):S1 · Maesaka JK et al. Front Med. 2018;5:319 · Koyner Ch 19 (Table 19.4)
Case 3 · Day 3 after decompressive craniectomy
55 M, 70 kg, TBI. Sedated, ventilated, ICP 12. Urine 300 mL/h × 6 h; balance −3.5 L.
Na 172U_osm 95U_Na 12U_K 8Urine 7.2 L
Diagnosis and plan
Deficit to 140 = 9.6 L; renal loss 6.4 L/day — treat the DI.
Desmopressin 1–2 µg q6–8h; Na q4h.
The four traps in this patient
ICP: cap the fall at 4–6/24 h; 0.9% saline, not Ringer's
Was the 172 prescribed? Check for hyperosmolar therapy
Triphasic: a standing order becomes SIADH at days 5–10