Hyponatraemia and hypernatraemia: pitfalls in testing

Abstract

Disorders of salt and water balance are extremely common in primary care. In many cases the cause is apparent and the result is not life threatening, but doctors should be aware of warning signs that may point to serious progressive disorders so that these can be diagnosed and managed early

Many situations involving the use and interpretation of laboratory tests are not supported by the high levels of evidence that can be achieved when interventions are assessed, but considerable consensus guidance is available on optimal use of laboratory tests. This article considers two scenarios involving salt and water balance that may be seen in primary care and discusses when further investigations may be helpful, and it gives a summary of evidence based and consensus guidance.

Summary points

• Hyponatraemia is common in primary care; hypernatraemia is rarer

• In both conditions, the common causes are usually clinically apparent

• When great or rapid changes occur, consider rarer causes

• Urine spot sodium concentration and osmolality help to differentiate the cause

• Unexpected results should raise suspicion of pseudohyponatraemia and pseudohypernatraemia

Although disorders of salt and water balance are extremely common in primary care, their causes are usually apparent, and the primary clinical question that arises is whether any change of dosage or drugs is required (typically diuretics in heart failure).

Serious and rapidly progressing sodium and water balance problems are rarer, and the practitioner often needs to decide when to initiate further investigation.

Case 1

A 72 year old man with chronic obstructive pulmonary disease due to longstanding smoking presented to his general practitioner with a two week history of lethargy and feeling nauseated. The patient was being treated with furosemide 20 mg and enalapril 5 mg a day for congestive cardiac failure (doses unchanged over recent months). This was clinically stable, and he had been noted previously to be hyponatraemic (sodium 131 mmol/l at last check two months previously) with a moderate reduction in renal function (creatinine 124 µmol/l, estimated glomerular filtration rate 54 ml/min/1.73 m²).

On examination the patient did not have fever and his blood pressure was 146/88 mm Hg. He had slight ankle oedema, which had been noted previously, and was in sinus rhythm, rate 86 beats/min. His chest contained a few diffuse rhonchi but no focal signs. Heart sounds were normal. Electrolytes were reported as serum sodium 124 mmol/l, potassium 4.3 mmol/l, urea 6.2 mmol/l, and creatinine 111 µmol/l. He reported no change in regularity of taking his tablets, and repeat testing was arranged for one week later. This returned a sodium of 123 mmol/l. The doctor telephoned the laboratory to discuss the results.

The measured osmolality of the serum sample submitted was 256 mmol/kg water (reference range 280-301 mmol/kg in people over 60 years old), and the doctor was asked to submit a urine specimen from the patient for osmolality, which was found to be 560 mmol/kg. Urinary sodium concentration was 36 mmol/l, and random plasma glucose concentration was 6.1 mmol/l. These results were interpreted as inappropriate concentration of urine and a urinary sodium concentration that was inappropriately high in a patient with hyponatraemia and hypo-osmolar serum. After discussion with the local endocrinologist, a synacthen test was done to exclude adrenocortical deficiency. This showed adequate adrenocortical reserve: baseline cortisol 465 nmol/l rising to 780 nmol/l after 250 µg of intramuscular tetracosactrin (normal response is a rise >200 nmol/l to >550 nmol/l after 30 minutes). The doctor was advised to start giving the patient 1 litre fluid restriction pending an urgent specialist opinion.

On subsequent investigation, a plain chest x ray showed no abnormality, although thoracic computed tomography showed hilar lymphadenopathy. Subsequently, bronchoscopy showed a small hilar bronchial tumour, which was diagnosed histologically as small cell lung cancer. Fluid restriction was continued, pending assessment for chemotherapy, and his sodium concentration rose to 128 mmol/l 10 days later.

Case 2

A 42 year old woman with type 2 diabetes controlled by metformin and gliclazide attended her general practitioner because of increased tiredness, thirst, and polyuria over the previous six weeks. She had lost 3 kg in weight over the same period. Glycated haemoglobin (Hb_A1c) four months before had been 7.2%, and capillary blood glucose concentrations measured at home had occasionally been in the region of 5-8 mmol/l.

On examination she was not acutely unwell and not clinically dehydrated. Her blood pressure was 124/78 mm Hg lying and 112/70 mm Hg standing, and pulse 84 beats per minute in sinus rhythm. Urine glucose by test strip showed ++ glucose, and serumelectrolyte results were reported as sodium 151 mmol/l, potassium 3.8 mmol/l, urea 8.3 mmol/l, creatinine 105 µmol/l. A random plasma glucose concentration was 7.7 mmol/l.

Her serum electrolyte results three months earlier were sodium 142 mmol/l, potassium 4.1 mmol/l, urea 4.3 mmol/l, creatinine 96 µmol/l, and her serum calcium corrected for albumin had been 2.26 mmol/l (reference range 2.20-2.62 mmol/l).

She was advised to drink plenty of fluids and to return in three days for review, during which time her doctor contacted the local laboratory to discuss her results. When she was seen again, her symptoms were still present; her repeat blood results were similar; and a urine sample, recommended by the laboratory, returned an osmolality of 210 mmol/kg. Her measured serum osmolality was 308 mmol/kg (reference range 275-295 mmol/kg).

Urine output over 24 hours was 4.6 litres. She was referred urgently to an endocrinologist, who gave her a test dose of intranasal desmopressin, which greatly reduced urine output, and urinary osmolality rose to 570 mmol/kg water. She started taking regular desmopressin while having further investigations. Anterior pituitary function testing showed no impairment of corticotrophin, gonadotrophin, or growth hormone release, and serum prolactin was within the reference range. However, her thyroid function tests showed secondary hypothyroidism with thyroid stimulating hormone of 3.1 mIU/l (reference range 0.4-5 mIU/l) and free thyroxine (FT4) of 9 pmol/l (reference range 11-23 pmol/l), and she started thyroid replacement treatment.

Subsequent magnetic resonance imaging of the brain showed that she had a craniopharyngioma, which was successfully decompressed surgically. She later had external beam radiotherapy to the residual tumour.

Discussion

Hyponatraemia

Hyponatraemia is the commonest of the electrolyte disorders. One recent study reported a prevalence of 7.2% in a community setting, rising to almost 30% in hospital inpatients.¹ Both hyponatraemia and hypernatraemia are more prevalent in older patients.² Hyponatraemia has been defined as mild (serum sodium 125-135 mmol/l), moderate (115-124 mmol/l), and severe (<115 mmol/l),³ but rate of change may be as important as absolute values.⁴

In most cases the cause is apparent from the clinical setting of diuretic use or secondary hyperaldosteronism in cardiac, liver, or renal disease. The difficulties are knowing when to investigate further and how to distinguish the rarer, less obvious, causes.

Analytical imprecision and biological variation mean that differences of up to 5 mmol/l in measurement of serum sodium may be due to chance. Laboratory errors can occur, so a repeat sample is prudent to confirm an unexpected result. Changes of 5 mmol or more should alert practitioners to a potentially progressive process.

Chronic mild hyponatraemia in a patient with heart failure is common and is related to a combination (among others) of natriuresis, diuretics, and secondary hyperaldosteronism. Diagnosing renal salt and water handling disorders in patients who are taking drugs that influence salt and water handling, notably diuretics, can be difficult; where possible, a cautious reduction in these drugs can remove some of their renal effects that may hinder the interpretation of results. The confirmed fall of 7-8 mmol/l over a two month period in case 1 suggests that an additional process is involved. In their normal functioning state, the kidneys will pass dilute urine and retain sodium to correct the hyponatraemia. Comparing serum and urine osmolality and sodium excretion therefore offers a simple and rapid way of identifying patients who are concentrating urine inappropriately. The urine osmolality is typically but not necessarily less than that of the hypo-osmolar serum,⁵ and a urinary sodium concentration of >30 mmol/l has been cited as inappropriately high in the presence of hyponatraemia,⁶ although quoted thresholds vary and the urine osmolality may provide useful additional information to discriminate between defective renal concentration and states in which renal salt is lost.

The syndromes described as resulting from inappropriate secretion of antidiuretic hormone and chronic dilutional hyponatraemia have causes that include ectopic production, increased pituitary secretion, increased sensitivity to antidiuretic hormone, and resetting of the body osmostat in chronic disease states. The last of these usually leads to stable mild hyponatraemia.

For clinical purposes the principal differentiation is between states that are chronic and relatively stable and those producing progressively severe and rapidly worsening hyponatraemia. Severe hyponatraemia can occur with any of a large number of drugs (hence the importance of a drug history) or as a result of secretion from tumours, classically small cell lung carcinoma but also several others. It may also occur with head injury and non-malignant lung disease.⁵ Other chronic illness states can produce hyponatraemia through a similar mechanism,⁷ although this is usually mild and relatively slow to change.⁵

Failure of the kidney to preserve sodium (urinary sodium concentrations are typically less than 10 mmol/l in hyponatraemia), combined with appropriately dilute urine, should prompt consideration of renal salt losing states, including intrinsic renal disease and endocrine causes (hypoadrenalism). Because secretion of antidiuretic hormone over-rides the body osmostat to maintain circulating volume,⁸ hypoadrenalism must be excluded before any fluid restriction is considered.

Pseudohyponatraemia (redistributional hyponatraemia) may be seen in a variety of situations unrelated to salt and water homoeostasis: hyperglycaemia, hyperproteinaemia, and chylomicronaemia.³ In these conditions serum osmolality is normal (hyperproteinaemia with chylomicronaemia) or raised (chylomicronaemia).

Hypernatraemia

Hypernatraemia is rarer in primary care, and in most cases the cause will be apparent from the clinical setting, typically gastrointestinal losses from vomiting or diarrhoea, or poor fluid intake. In case 2 the magnitude of the rise over three months acted as the indicator of a potentially progressive process.

Other than urine glucose concentrations, which may have been positive because of a low renal glucose threshold, the results in case 2 excluded poor diabetes control as a cause for her polyuria, and the finding of inappropriately dilute urine in a patient whose serum is hyperosmolar (and would under normal circumstances have been concentrating her urine in order to retain water) indicate diabetes insipidus, requiring urgent referral.

Hypernatraemia with hyperosmolar serum and inappropriately dilute urine is diagnostic of diabetes insipidus. This may be caused by defective production of antidiuretic hormone from the hypothalamus, or, more rarely, release from the posterior pituitary gland, or from loss of renal response to circulating ADH (nephrogenic diabetes insipidus). Additional investigations can exclude several of the secondary causes of nephrogenic diabetes insipidus: hypercalcaemia, hypokalaemia, and drugs (notably lithium). Further characterisation of the cause of the diabetes insipidus may include dynamic testing (water deprivation test under carefully supervised conditions); a challenge with desmopressin; and imaging to identify space-occupying hypothalamic lesions, classically craniopharyngioma, or the rarer granulomatous diseases such as Wegner's granulomatosis or neurosarcoidosis. In most cases the thirst response will attempt to correct the hyperosmolar state, and with the exception of patients such as elderly or bedbound people, who may have limited access to fluids through immobility or institutionalisation (reviewed by Milionis et al⁵), large confirmed rises in serum sodium concentration should prompt suspicion of diabetes insipidus.

Urine osmolality must be interpreted in light of serum osmolality. Quoted “reference ranges” for urine osmolality (typically 50-1200 mmol/kg) simply reflect the range of the concentrating capacity of the kidney, which depends on body hydration status.

Pseudohypernatraemia is unusual, but it may be seen in hypoproteinaemic states when sodium has been measured by routine laboratory methods.⁹

Questions and answers: learning points

The related series of questions and answers can be found in the eighth review of best practice in primary care pathology published in the Journal of Clinical Pathology.¹⁰ The key recommendations from these reviews are in boxes 1 and 2.

Box 1: How should I investigate a patient with low serum sodium concentration?

• Establish history of fluid intake and current treatments

• Assess fluid status, to identify whether hypovolaemia or hypervolaemia is present

• Repeat to confirm and establish whether acute and changing or chronic and stable. Changes of up to 5 mmol/l can reflect non-significant variation.

• Persistent and stable serum sodium 132-135 mmol/l in a clinically well patient may reflect a statistical population outlier and may not require investigation

• Serum sodium 125-131 mmol/l:

• • Check serum potassium, urea, creatinine, triglycerides, and protein and plasma glucose

  • If cause is not clinically apparent, check

  • Urine sodium and osmolality if inappropriate secretion of antidiuretic hormone (SIADH) suspected:

  • Urine sodium >30 mmol/l suggests renal sodium loss or SIADH

  • Urine osmolality >100 mmol/kg water in the context of hyponatraemia is consistent with SIADH

  • Exclude Addison's disease and hypothyroidism

  • Consider reset osmostat syndrome in patients with debilitating illness and stable hyponatraemia.

• Serum sodium 115-124 mmol/l:

• • Check, as above

  • Seek specialist advice unless long term stable and cause has been established

  • Consider immediate admission if sodium is falling rapidly or neurological signs or symptoms are present

• Serum sodium <115 mmol/l:

• • Immediate admission usually indicated

Box 2: How should I investigate a patient with high serum sodium concentration?

• Repeat to confirm and establish whether acute and changing or chronic and stable. Changes of up to 5 mmol/l can reflect non-significant variation

• Establish history of thirst, fluid intake and losses, and current treatments

• Assess fluid status to look for hypovolaemia

• Exclude diabetes mellitus

• Persistent serum sodium 146-148 mmol/l without clinical features of hypovolaemia may mean simply that the patient's result is a statistical outlier.

• Serum sodium 149-154 mmol/l:

• • Review serum potassium, urea, creatinine, calcium, and plasma glucose results

  • Consider lithium-induced nephrogenic diabetes insipidus and lithium toxicity when appropriate

  • Request urine and serum osmolality if diabetes insipidus is suspected

  • Consider specialist advice if clinical cause is not apparent and oral rehydration is not possible in a dehydrated patient

• Serum sodium ≥155 mmol/l:

• • Seek specialist advice or admission in addition to above measures

Evidence note

Definitions of mild moderate and severe hyponatraemia are based on a consensus opinion document³; they serve only as a guide, as the rate of decline and the clinical context will influence individual patients' responses. We found no definitions of the severity of hypernatraemia, and the guideline values offered are approximate equivalents to total body water loss, representing an approximate threshold at which oral hydration is practicable in a patient who is otherwise not at risk. Rate of change and clinical context are equally important.

The clinical situations in which hyponatraemia and hypernatraemia can occur are too varied for any action level to represent any more that a general guide. Interpretation of changes is based on the statistical likelihood of two or more consecutive results representing true and not random changes. A difference of 2.8 standard deviations, or approximately 5 mmol/l between two consecutive readings, is 95% likely to be significant.¹¹ A difference of approximately 0.7 SD offers a 50% chance of a result not being obtained by chance. For sodium this means that differences of up to 2 mmol/l are as likely to be random findings as true change, whereas those of 5 mmol/l or more are unlikely to have been obtained by chance.

The evidence on investigating for rarer causes and for pseudohypernatraemia or pseudohyponatraemia comes from observational studies. The guidance would seem to be justified on the basis of the potential clinical severity of several of the causes in pseudohypernatraemia and avoiding unnecessary investigation, missed diagnoses, and inappropriate treatment in pseudohyponatraemia.