Relative hypovolaemia ADH xs water
Heart failure
Hypoalbuminaemia (liver/renal failure/protein losing enteropathy)
Loop diuretics
XS water
Dextrose IVI
Polydipsia (some solute has to be excreted with urine - as minimal urine concentrating ability is around 60mosmol/l no more water than this can be excreted).
Malnourishment (reduced protein intake means inadequate solute formation in urine so less water can be excreted - bear in mind with critically ill or post-op patients without nutrition).
Na loss
Hypoadrenal / ACEi

Mild >125
Moderate >120 – N+V
Severe <120 – CNS symptoms (

Cancel elective if <120 or symptomatic
Emergency – consider risk/benefit

  • Hyponatraemia
  • Normovolaemia (without peripheral oedema)
  • Urine sodium level >25mmol/l
  • Serum osmolality <280 mOsmol/l
  • Urine osmolality >serum osmolality
  • Fluid restriction (1000 ml isotonic fluid/24 h)
  • If fluid restriction alone fails or hyponatraemia <115 mmol/l
  • NaCl 3 % + frusemide
  • Aim for slow serum sodium correction (max 1 mmol/h, 12/24h) to avoid myelinolysis.
Cerebral Salt Wasting Syndrome (CSW)
  • Hyponatraemia
  • Hypovolaemia
  • Urine sodium levels >50 mmol/l
  • Fluid replacement with saline 0.9 %
  • NaCl 3 % if necessary for slow serum sodium correction

Be aware - high glucose, triglyceride or albumin levels can fake a hyponatraemia (pseudohyponatraemia)


Dehydration (osmotic diuresis, low ADH (DI))
Na overload

>155 - CNS symptoms due to hyperosmolar state and cellular dehydration
Correct slowly (over 48h) to prevent cerebral oedema

Cancel elective if >155
Emergency – beware too rapid correction

Diabetes Insipidus

Cranial DI is caused by post-traumatic failure of ADH secretion

  • Polyuria (>3000ml/24h or >125mls/h)
  • Hypernatraemia
  • Urine osmolarity <300 mOsmol/l, rises > 50 % after desmopressin (DDAVP)
  • Serum osmolarity>300 mOsmol/l
  • If diuresis <4000 ml /24 h then only fluid replacement
  • If diuresis >4000 ml /24 h then fluid replacement + DDAVP 1 mcg i.v.
  • Correct hypovolaemia with Hartmanns, then use Glucose 5 % for correction of hypernatraemia.
  • Plasma sodium lowering not faster than 0.7 mmol/h.(16.8 mmol/day)

Hypernatraemia in the critically ill

Hypernatraemia is a relatively common problem in critically ill patients. A recent study 1 in one centre showed 7% of critically ill patients in their 3 units (surgical, medical and neurological) developed hypernatraemia during their ICU stay. Hypernatraemia was shown to be an independent predictor of mortality. Sepsis, hypokalaemia, hypoalbuminaemia, renal dysfunction, mannitol, sodium bicarbonate were independently associated with hypernatraemia. Interestingly, during the development of hypernatraemia, fluid balance was negative in 60% of cases but positive in 40%. This tells us that hypernatraemia develops not just when there is negative water balance but when there is positive sodium balance.
Critically ill patients by necessity receive large quantities of sodium containing fluids. It is however becoming increasingly recognized that this causes a number of problems
2. These fluids equilibrate with the ECF volume thus exacerbating the organ oedema and dysfunction caused by a systemic inflammatory response. Whilst the human body is very good at retaining sodium the mechanisms for its excretion are much less efficient. Activation of the sympathetic nervous system and renin-angiotenesin system, hyperchloraemia (renal vasoconstriction) and increased catabolism (urea competes with sodium for excretion) all cause sodium retention. The physiological reduction in urine volumes that this causes can be misinterpreted as hypovolaemia and more sodium containing fluids are administered exacerbating sodium overload. In critical illness the kidneys also lose their concentrating ability meaning a large amount of water is required to excrete solutes. Renal concentrating ability often falls to 500mosmol/l. A positive osmolar balance of 2500 (will often be much more) would therefore require 5L of water for excretion (free water is the best diuretic). Many drugs also either present a large sodium load (many antibiotics) or cause sodium retention (such as steroids). So, even if elevated plasma levels of sodium do not occur, total body sodium is usually high. Our patients will only get ‘better’ when this excess sodium is excreted which takes a significant amount of time. Disturbances of plasma sodium can also cause cerebral oedema and myelinolysis if they develop, or are treated, too quickly.
Patients in critical care units have plasma sodium levels monitored at least daily and often much more often as they are measured routinely by ABG machines. Therefore plasma hypernatraemia is a completely preventable problem requiring only the administration of adequate amounts of free water. A patient with a plasma sodium of 155 and an ECF volume of 15L has 2325 mmol of total body sodium which means an additional 1L of ECF water would be required to normalise the sodium to 140mmol/l. As only 1/3 of any water given will remain in the ECF the total water required would be 3L. Of course the ECF is not simply a bucket and physiological processes will have a significant influence on this. It does however serve as a rough reminder of the amount of water required. Free water (either added to enteral feed or as 5% dextrose IV) should be given in normal physiological amounts with additional water titrated against plasma sodium concentration. Sodium levels should of course be corrected at a speed which will not cause osmotic fluid shifts within the brain.
Total body hypernatraemia is more difficult to avoid as fluids containing sodium in excess of daily requirements (1mmol/kg) are necessary for initial fluid resuscitation coupled with the fact that physiological mechanisms of sodium retention are activated. We can however limit sodium doses by a number of ways.
During rescitation, some would advocate using colloid in addition to crystalloid to limit fluid volumes. But colloids should be contraindicated in critically ill patients. Level 1 evidence shows that only around 50% more crystalloid is required than colloid for comparable resuscitation goals (in contrast to the widely held belief that 3-4 times as much crystalloid is required). There is also emerging evidence that gelatin and HES solutions (including 130/0.4 HES) cause renal failure
3, coagulation failure 4, widespread tissue uptake, intractable pruritis and lysosomal storage disease. There are no adequate safety studies for the use of colloids and none in critical illness. The use of balanced crystalloid rather than 0.9% saline and having clearly defined endpoints for identifying intravascular filling will limit total sodium dose.
Subsequently we can use dextrose where possible for diluting drugs and particularly avoid using sodium containing fluids for maintenance which leads to gross fluid and sodium overload.
The ARDS network study referenced has demonstrated the safety and advantages of restrictive fluid regimes in critically ill patients. It is worth noting that just because a patient is fluid responsive does not necessarily mean they need fluid.


Use balanced crystalloid (CSL) for resuscitation.
Stop fluid administration once no longer fluid responsive.
CSL is a resuscitation fluid and
not a maintenance fluid. After fluid resuscitation 5% dextrose or dextrose saline 4% 0.18% should be used for maintenance (with monitoring of sodium in case of hyponatraemia).
Dilute drugs in dextrose when possible.
As soon as enteral feed established add enteral water to enhance diuresis and excretion of sodium.
In summary I quote part of the title of my first reference - ‘Hypernatraemia….too little water and too much salt’. Do your best to avoid these two things.


Hypernatraemia in critically ill patients: too little water and too much salt. Hoorn EJ et al. Nephrol Dialysis Transplant 2008; 23: 1562-1568.

British Consensus Guidelines on Intravenous Fluid Therapy for Adult Surgical Patients. GIFTASUP. Powell-Tuck et al. On behalf of BAPEN Medical - a core group of BAPEN, the Association for Clinical Biochemistry, the Association of Surgeons of Great Britain and Ireland, the Society of Academic and Research Surgery, the Renal Association and the Intensive Care Society.

Intensive Insulin Therapy and Pentastarch Resuscitation in Severe Sepsis. Brunkhorst et al. N Engl J Med 2008;358:125-3

Wiedermann et al. The National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Comparison of Two Fluid-Management Strategies in Acute Lung Injury. N Engl J Med 2006;354:2564-75.