Two syndromes produce identical laboratory findings and take opposite treatments, and the thing that separates them is not on the blood results. Getting it wrong causes cerebral infarction. Correcting it too fast causes something worse.
In both SIADH and cerebral salt wasting the serum osmolality is under 274 mmol/L, the sodium under 135, the urine osmolality over 100 and the urine sodium over 20. The laboratory cannot tell them apart. What separates them is volume status — the SIADH patient is normo- or hypervolaemic and the salt-waster is hypovolaemic — and the treatments are opposites. Restrict and diurese a salt-waster after subarachnoid haemorrhage and you produce hypovolaemia and cerebral infarction.
Hyponatraemia is common after neurosurgery, after subarachnoid haemorrhage, in hydrocephalus and in central nervous system infection. Two syndromes account for most of it, they produce the same numbers, and their treatments are opposites.
SIADH. The posterior pituitary continues releasing antidiuretic hormone despite a low serum sodium. Free water is retained, and the result is hyponatraemia with a normal or expanded volume. Symptoms of the hyponatraemia itself — nausea, vomiting, headache, confusion, seizures, coma — are the same whatever caused it.
Cerebral salt wasting. Raised brain natriuretic peptide drives renal excretion of sodium and water. The result is hyponatraemia with hypovolaemia. It is a common cause of low sodium after subarachnoid haemorrhage, and the volume depletion is not incidental — it contributes to delayed cerebral ischaemia in exactly the population least able to tolerate it.
Here is the part worth committing to memory, because it is the reason people get this wrong. Laboratory findings in both include:
That panel confirms real hypo-osmolar hyponatraemia with concentrated urine and renal sodium loss. It confirms a category. It does not identify a mechanism, and no additional blood test resolves it.
What does resolve it is clinical context, intravascular volume status and urine output. The SIADH patient is normo- or hypervolaemic. The salt-waster is hypovolaemic. That difference is available at the bedside in about ninety seconds, and it is the whole diagnosis.
Treatment of SIADH is free water restriction and a loop diuretic, to achieve a negative water balance. Treatment of cerebral salt wasting is fluid resuscitation to euvolaemia, sodium repletion with hypertonic saline or sodium tablets, and fludrocortisone.
Apply the first to a patient who needs the second and you are deliberately dehydrating someone who is already volume depleted. The stated consequence is worsening hypovolaemia and cerebral infarction in the subarachnoid haemorrhage patient.
There is no equivalent catastrophe in the other direction, which is worth knowing: over-resuscitating an SIADH patient is a mistake and a recoverable one. The asymmetry should shape which way you err when the volume status is genuinely unclear.
The indications given are a serum sodium less than 120 mEq/L accompanied by seizures, somnolence, or cerebral oedema. Below 120 and asymptomatic is not the same clinical situation.
Most commonly 3% saline is administered slowly, 25 to 100 mL/h, with frequent monitoring of serum sodium.
The rate limits are the point of the day’s title. Correction should not exceed 0.5 mmol/L per hour, or 8 mmol/L in 24 hours. Both, not either. That daily ceiling is the expert panel’s figure for patients at heightened risk of demyelination — sodium at or below 105, hypokalaemia, malnutrition, alcohol use disorder, advanced liver disease — where it allows 10 to 12 at average risk. This curriculum teaches 8 throughout, because at the moment the first correction is written you frequently do not yet know which group the patient is in, and the two errors are not symmetrical: correcting too slowly is an inconvenience, and correcting too fast is a demyelination syndrome. Rapid correction is associated with osmotic demyelination syndrome, and the risk is increased when the hyponatraemia has been present for longer than 48 hours. When the duration is genuinely unknown — which is most of the time in a patient who arrived unwell — the safe assumption is that it is chronic.
One randomized trial bears on the technique rather than the target: rapid intermittent boluses of hypertonic saline were compared with slow continuous infusion in symptomatic hyponatraemia, and the bolus approach was not inferior and carried a lower risk of overcorrection. That is a useful result and it does not change the ceiling.
While you are here, learn the mirror image, because it appears in the same patients and presents as a fluid balance rather than as a diagnosis. Diabetes insipidus: polyuria above 200 mL/h, urine specific gravity below 1.005, urine osmolality below 300 mOsm/kg, serum osmolality above 300 mOsm/kg, sodium above 145 mmol/L, and dehydration.
Critically ill patients present more severely than others because thirst and oral intake — the mechanisms that normally hide the problem — are not available to them. A patient who cannot ask for water will simply become hypernatraemic while the chart records a good urine output.
No randomized trial addresses this.
The physiology of each syndrome is settled. In SIADH, antidiuretic hormone continues to be released despite a low serum sodium, free water is retained, and the patient becomes hyponatraemic and normo- or hypervolaemic. In cerebral salt wasting, raised brain natriuretic peptide drives renal excretion of sodium and water, so the patient becomes hyponatraemic and hypovolaemic. Neither mechanism is in dispute.
The overlap in laboratory findings is a fact, not an artefact. Both produce a serum osmolality under 274 mmol/L, a serum sodium under 135 mEq/L, a urine osmolality over 100 mmol/L and a urine sodium over 20 mmol/L. The source text is explicit that the two must be distinguished on clinical context, intravascular volume status and urine output — because their treatments differ greatly.
The rate limits are consensus. Correction of chronic hyponatraemia should not exceed 0.5 mmol/L per hour or 8 mmol/L in 24 hours, because rapid correction is associated with osmotic demyelination syndrome, and the risk rises when the hyponatraemia has been present longer than 48 hours. These figures come from expert consensus and observational series rather than from randomized comparison — the harm they exist to prevent is too rare and too catastrophic to randomize.
The daily ceiling is stratified in the source, and this curriculum teaches the conservative figure. The expert panel sets 8 mmol/L in 24 hours for patients at heightened risk of osmotic demyelination — a sodium at or below 105, hypokalaemia, malnutrition, alcohol use disorder or advanced liver disease — and allows 10 to 12 for those at average risk. This day teaches 8 throughout, because the risk factors are frequently unknown at the moment the first correction is written, and because the cost of being wrong in the permissive direction is a demyelination syndrome rather than a slow correction.
One randomized trial does bear on how you deliver the correction. A comparison of rapid intermittent bolus against slow continuous infusion of hypertonic saline in symptomatic hyponatraemia found the bolus approach was not inferior and was associated with a lower risk of overcorrection. That addresses technique, not target.
Strength, in the guideline's own words The 2013 expert panel recommendations on hyponatraemia are consensus recommendations from a multidisciplinary panel, not a graded guideline.
A serum sodium of 126, a serum osmolality of 262, a urine osmolality of 480 and a urine sodium of 60 is a complete set of results that identifies the syndrome in nobody. Those numbers are compatible with SIADH and with cerebral salt wasting equally. They establish that the hyponatraemia is real and hypo-osmolar with inappropriately concentrated urine, and that is all they do.
What they look like, especially at three in the morning, is a diagnosis of SIADH — because SIADH is the commoner condition, because the pattern is the one taught with those numbers attached, and because the treatment is easy to start. Restrict fluid, add a loop diuretic, recheck in the morning.
Do that in a patient with subarachnoid haemorrhage who is actually salt wasting and you have deliberately dehydrated someone whose hypovolaemia is already contributing to delayed cerebral ischaemia. The stated consequence is cerebral infarction.
The discrimination is clinical and it is not subtle once you look for it: intravascular volume status, urine output, and the context. A hypovolaemic patient with a high urine output and a falling sodium after SAH is wasting salt, whatever the urine sodium says — because the urine sodium is high in both.
The second half of the pitfall is the correction itself, and it catches people who got the diagnosis right. Overcorrection is rarely a single bad decision. It is an hourly rate that looked fine at each check and breached the daily ceiling by evening, in a patient whose hyponatraemia turned out to have been present for a week rather than a day.
Full text · Preston goes through LibKey to whatever Preston holds — the PDF where there is one — and works on a phone. The doi link always reaches the publisher's record, whether or not we hold a subscription. If you are reading on a computer, LibKey Nomad adds the same one-click PDF button to PubMed and publisher pages everywhere else; it is a desktop browser extension, so it does nothing on a phone or tablet.
On day 4 after aneurysmal subarachnoid haemorrhage, a patient's serum sodium is 126 mEq/L with a serum osmolality of 264 mmol/L, urine osmolality 420 mmol/L and urine sodium 65 mmol/L. Urine output has been 250 mL/h, the heart rate is 105 and the patient looks dry. What is the best action?
What rate limits apply to the correction of chronic hyponatraemia, and why?
Why do the standard laboratory findings fail to distinguish SIADH from cerebral salt wasting?
The concern is osmotic. A solution below plasma osmolality drives water into tissue that is already injured and swelling. Balanced isotonic crystalloid, alone or with colloid, is the appropriate choice; glucose-containing solutions should also be avoided, and hydroxyethyl starch avoided in general as the one colloid linked with increased mortality.
Restrictive transfusion reduced 30-day mortality among critical care patients, while the same comparison among surgical patients trended in the opposite direction, and current evidence comparing restrictive with liberal targets in surgical patients does not point either way. Against that, transfusing patients with neurologic injury who are not in haemorrhagic shock has been associated with worse outcomes, particularly above 10 g/dL. A range with frequent measurement is what that evidence actually supports.