Diabetes insipidusBMJ 2019; 364 doi: https://doi.org/10.1136/bmj.l321 (Published 28 February 2019) Cite this as: BMJ 2019;364:l321
- Miles Levy, consultant endocrinologist1,
- Malcolm Prentice, consultant endocrinologist2,
- John Wass, consultant endocrinologist3
What you need to know
In patients with polyuria, diabetes insipidus is very unlikely if urine osmolality is >700 mOsmol/kg
Patients with central diabetes insipidus who are admitted to hospital should have specialist input and safeguards in place to ensure that desmopressin is not omitted
Intercurrent illness with hypernatraemia in a patient with diabetes insipidus should be managed as a medical emergency
Diabetes insipidus is a rare but treatable condition that typically presents with extreme thirst (polydipsia) together with the passing of large amounts of dilute urine (polyuria). Distinguishing these symptoms from those of primary polydipsia, diabetes mellitus, and causes of urinary frequency without polyuria can be challenging. Diabetes insipidus is caused by a problem with vasopressin production in the pituitary gland (central diabetes insipidus), or action of vasopressin in the kidneys (nephrogenic diabetes insipidus). Desmopressin, an analogue of vasopressin, is an effective treatment for cranial diabetes insipidus. Between 2009 and 2016 there were four reported deaths in England resulting from omission of desmopressin, and a further 56 reported incidents where dosing errors resulted in harm.1 This Practice Pointer offers an approach to diagnosing suspected diabetes insipidus, and guidance on managing people with diabetes insipidus who have intercurrent illness or require hospital admission.
What is diabetes insipidus?
Diabetes insipidus is rare, with a prevalence of 1 in 25 000.2 Central diabetes insipidus usually results from pituitary pathology,3 either as a result of infiltrative or inflammatory pathology, or following surgery for a pituitary tumour, but may also be due to a congenital defect in the production of arginine vasopressin.3 Nephrogenic diabetes insipidus is usually caused by electrolyte disturbance, renal disease, or drug toxicity (commonly lithium2).
Arginine vasopressin causes water reabsorption at the collecting ducts of the kidney (fig 1). Deficiency of or resistance to the hormone, as seen in diabetes insipidus, leads to excessive water loss resulting in polyuria. Typically, the compensatory drive for thirst will provide adequate rehydration, but in severe cases where there is not ready access to water, someone with diabetes insipidus can become rapidly dehydrated, which may lead to hyperosmolality, hypernatraemia, and potentially death.1
How is diabetes insipidus diagnosed?
Extreme thirst and passing large quantities of pale urine are typical presenting symptoms of diabetes insipidus. It may be difficult to distinguish diabetes insipidus from differential diagnoses with these symptoms, but there are pointers in the history and investigation that can help (table 1). A particular challenge is primary polydipsia, which refers to a psychologically driven increase in fluid intake rather than impaired vasopressin regulation and is often seen in patients with severe mental illness and/or developmental disability,45 although it may simply be behavioural and occur in healthy individuals in the absence of psychiatric disease. Polyuria should be distinguished from urinary frequency in the history, the latter suggesting a urological problem.
In central diabetes insipidus, the history of polyuria and polydipsia is usually abrupt, presenting within weeks or months of onset.3 In nephrogenic diabetes insipidus, the onset is more insidious and patients have often had symptoms for months or years before the diagnosis is made.2
Symptoms suggestive of pituitary disease may include fatigue, dizziness, irregular periods, and galactorrhoea in women, or loss of libido and reduced secondary sexual characteristics in men.
Ask about a history of pituitary disease, major head injury, or neurosurgery, which are risk factors for central diabetes insipidus. Several genetic mutations have been identified for the condition, so a family history of central or nephrogenic diabetes insipidus may be highly relevant.23
Look carefully at medication history. Patients taking loop diuretics and nephrotoxic drugs are at risk of developing nephrogenic diabetes insipidus.7 Nephrogenic diabetes insipidus occurs in approximately 15% of patients taking lithium.7
The initial investigation of a patient presenting with polyuria and polydipsia is summarised in figure 2.
Diabetes mellitus—Exclude diabetes mellitus either by urinalysis or point of care testing, confirmed by formal measurement of fasting or random glucose.
Electrolyte disturbance—Take blood to exclude hypercalcaemia and hypokalaemia as these can cause nephrogenic diabetes inspidus.2
Urine volume—If 24 hour urine volume is less than 2.5 L, diabetes insipidus is highly unlikely and other causes of urinary symptoms should be considered. Urine volume can be measured by the patient themselves (eg, with a measuring jug), or a 24 hour collection bottle can be given and sent to the laboratory for measurement of volume only.
Paired urine and plasma osmolality—If 24 hour urine volume exceeds 2.5 L, paired serum and urine osmolalities can help to distinguish diabetes insipidus from polyuria caused by primary polydipsia. If baseline urine osmolality is >700 mOsmol/kg, diabetes insipidus is very unlikely as the ability to concentrate urine adequately has been demonstrated. Diabetes insipidus is likely if serum osmolality is high (>295 mOsmol/kg) and urine osmolality low (<300 mOsmol/kg). Because patients with diabetes insipidus compensate by drinking according to thirst, it can be difficult to distinguish diabetes insipidus from primary polydipsia on the basis of a one-off paired urine and plasma osmolality measurement.6 In this situation it may be necessary to perform more complex investigations in the specialist setting.
Referral and specialist management
Patients with suspected diabetes insipidus should be referred for specialist investigation and treatment. The urgency of referral depends upon the severity of symptoms. If thirst and polyuria are extreme and serum osmolality >295 mOsmol/kg, refer patients within days or a few weeks at most. Patients with known diabetes insipidus who have hypernatraemia should be seen as an emergency the same day.
Water deprivation test—Currently, in equivocal cases where the diagnosis of diabetes insipidus is not clear cut, the water deprivation test is the most common confirmatory test used in specialist care.6 In this test, the person is deprived of water for several hours while their urine output, urine osmolality, and serum osmolality are monitored over time. In patients with severe diabetes insipidus, water deprivation can be highly unpleasant and should be supervised by the endocrine team with continued measurement of serum and urine osmolality, urine volume, and weight. The test should be stopped if the patient is distressed or there is clear evidence of continued high output of dilute urine, an excessive rise in serum osmolality, and an excessive loss in weight, during fluid restriction. In people with diabetes insipidus there is continued polyuria and low urine osmolality despite water deprivation. If urine output falls and urine osmolality exceeds 750 mOsmol/kg, diabetes insipidus is excluded. In such cases primary polydipsia is more likely. In the second part of the water deprivation test, desmopressin is given to those with confirmed diabetes insipidus. Patients with central diabetes insipidus respond to desmopressin with a rise in urine osmolality and fall in urine volume. There is no response to desmopressin in patients with nephrogenic diabetes insipidus.
Other specialist investigations—There is an increasing move to measure copeptin, a marker of arginine vasopressin levels,8 in response to hypertonic saline infusion. In normal individuals, hypertonic fluid leads to an increase in vasopressin release, and therefore an increase in copeptin levels. In central diabetes insipidus, there is a blunted copeptin rise, and this is probably a more sensitive and specific diagnostic test than water deprivation. Detailed pituitary imaging such as magnetic resonance and positron emission tomography may help to differentiate between inflammatory and infiltrative pituitary disorders.
Central diabetes insipidus—adequate fluid replacement, treatment of the underlying condition, and desmopressin administration are the mainstays of management. Desmopressin can be taken orally or via an intranasal spray. Central diabetes insipidus usually responds to desmopressin immediately, and patients notice a substantial reduction in polyuria and thirst. Symptoms of under-replacement with desmopressin are thirst and polyuria, while symptoms of over-replacement are headache and mild confusion (due to hyponatraemia) and reduced urine output.
Empowering patients to manage their own condition is an important part of management, and the endocrine specialist nurse plays a key role in this regard. An important area of self management is preventing hyponatraemia, which is a common complication of desmopressin treatment because excessive fluid intake in the presence of continued desmopressin may lead to over-dilution of the blood due to excessive reabsorption of water It is good practice for patients to have a regular diuresis by omitting a desmopressin dose once or twice a week, or waiting until they have passed urine before taking their medication.9 The Pituitary Foundation produces a diabetes insipidus card and booklet for patients to carry, to alert the treating physician to the diagnosis in the event of an emergency.10 This ensures that if a patient is too unwell to give a history, the clinician first at the scene is aware of the diagnosis of diabetes insipidus and knows the importance of fluid and desmopressin administration. There is current ongoing work to produce patient information leaflets similar to sick-day hydrocortisone rules in patients with hypoadrenalism.
Nephrogenic diabetes insipidus—is managed with fluid replacement and cause specific treatment, under the care of a renal specialist.2 Other treatments include diets low in salt and protein, diuretics, and non-steroidal anti-inflammatory drugs.2
Management of diabetes insipidus for the non-specialist
The Society for Endocrinology has recently produced guidance for inpatient management of acutely unwell patients with diabetes insipidus.11 These guidelines suggest that all patients admitted to hospital with central diabetes insipidus are identified on admission and that the endocrinology or alternative appropriate clinical team is alerted. Any patient with central diabetes insipidus who is admitted to hospital needs close monitoring of fluid replacement as well as appropriate administration of desmopressin.
All patients undergoing elective surgery should be highlighted in the pre-assessment process with a clear perioperative plan. They recommend hospitals develop an alert system to highlight all patients requiring ongoing desmopressin therapy to ensure doses are not missed.
A Society for Endocrinology survey of UK based endocrinologists suggests that the problem of delayed administration of desmopressin and fluids when patients are admitted to hospital is widespread.12 Fifty five per cent of respondents had concerns about management of patients with diabetes insipidus in their hospital, and 47% reported at least one patient coming to harm because of delayed administration of desmopressin or insufficient fluid replacement. An NHS patient safety alert has reported a series of critical incidents occurring in patients with central diabetes insipidus.1 Between 2009 and 2016, four inpatient deaths caused by desmopressin omission were reported in England.1 One was a 22 year old man with a benign pituitary tumour who died after a routine orthopaedic procedure. This safety alert identified several themes from these incidents, including a lack of awareness of the critical nature of desmopressin among medical, pharmacy, and nursing staff; poor availability of desmopressin within inpatient clinical areas; and omission due to nil-by-mouth status or acute illness. A small survey of non-specialist nursing staff found that some were not aware that diabetes insipidus was a different condition from diabetes mellitus.1
Central diabetes insipidus—In those who are unwell with intercurrent illness, it is important to accurately assess fluid status and measure serum electrolytes. Patients with hypernatraemia should be managed as a medical emergency in a high dependency setting. Monitor serum sodium every four hours during fluid resuscitation. In patients who have impaired consciousness, it may be necessary to administer desmopressin by the intravenous, subcutaneous, or intramuscular route.
Nephrogenic diabetes insipidus—Patients are similarly at risk of hypernatraemia and severe dehydration. Seek specialist input ideally from the renal team. Treat the cause of the intercurrent illness, consider withdrawal of drugs which may be causing diabetes insipidus, as well as fluid resuscitation.
Education into practice
Do your patients with diabetes insipidus have alerts on their electronic patient record that highlight the risk of desmopressin omission?
How might you offer training to staff to highlight the difference between diabetes insipidus and diabetes mellitus?
Do you offer safety cards to patients with diabetes insipidus?
How patients were involved in the creation of this article
No patients were directly involved in the writing of this article. A person with diabetes insipidus wrote the Patient perspective.
Patient perspective—The water deprivation test
The test begins with the words “YOU WILL NOT DRINK ANY FLUIDS FOR THE NEXT EIGHT HOURS.” If you do have suspected diabetes insipidus, you’ll now be in a total panic! I would typically have at least five litres in eight hours—often much more. You are then told that all trips to the toilet will be escorted—just in case you find a dripping tap, or (bliss) a can of icy coke on the way. Half hourly blood tests will break the monotony and you’ll carry a measuring jug at all times. After several hours of the test, any hope of a trickle of saliva has long gone—your tongue is firmly welded to the roof of your mouth. Can these doctors and nurses possibly imagine what you are going through? Why do they remark on your grey pallor and shuddering body—this is dehydration! Even with no fluids going in, your bladder will still twinge, like an annoying buzzing wasp urging you to empty it yet again. Where on earth is all this pee coming from? Finally, the doctor approaches with a cheery “the test is now over, you are free to drink.” You’ll gulp any fluid in sight—the entire ward’s water jugs, the domestic’s dirty water bucket! On completion of the test and if diagnosed, you’ll receive an injection of desmopressin—the most wonderful medicine ever produced for a person with diabetes insipidus. That injection gives you back a normal bladder output, and the raging thirst is quelled. Temporarily, of course, but heaven while it lasts.
Competing interestsThe BMJ has judged that there are no disqualifying financial ties to commercial companies. The authors declare the following other interests: None.
Further details of BMJ policy on financial interests is here: https://www.bmj.com/about-bmj/resources-authors/forms-policies-and-checklists/declaration-competing-interests
Provenance and peer review: commissioned, based on an idea from the author; externally peer reviewed.
Patient consent obtained for Patient perspective.