Late onset nephrogenic diabetes insipidus following cessation of lithium therapy.

SUMMARY

Lithium-induced diabetes insipidus occurs in up to one third of patients on long-term therapy. I t usually manifests during the treatment course, often with preceding signs or symptoms We wish to highlight a case in which diabetes inspidus developed many days after cessation of long-term lithium therapy. We seek to explain this delayed onset by a discussion of the molecular and genetic changes underlying the pathogenesis of lithium-induced nephrogenic diabetes insipidus 

Key Words: diabetes insipidus, lithium, aquaporin, thiazide 

**********

Diabetes insipidus (DI) is a disorder of inadequate urinary concentration leading to a polyuric state (arbitrarily defined as a urine output exceeding three litres in 24 hours) and is subdivided into either cranial or nephrogenic types. Cranial DI is due to decreased secretion of vasopressin from the pituitary gland. At least 80% of vasopressin-synthesising neurones must be destroyed before overt clinical features become manifest. In contrast, nephrogenic DI results from resistance of the distal renal tubules to the action of vasopressin (1). In some cases it is possible to distinguish between these two types (by the history, for example), but a simple therapeutic trial of desmopressin can also be used. After administration of desmopressin, patients with cranial DI will be identified by little or no weight gain, a reduction in urine flow, normal plasma osmolality and a high urine osmolality. In other words, the replacement of the endogenous vasopressin by the exogenous desmopressin causes normal urinary concentration and resolution of the polyuria. In contrast, nephrogenic DI is characterised by a lack of response to desmopressin (2).

There are many causes of nephrogenic DI, including congenital (a familial X-linked recessive disorder), metabolic (hypercalcaemia, hypokalaemia), vascular (sickle cell disease), chronic renal disease (pyelonephritis, sarcoidosis) and iatrogenic (lithium, demeclocycline) (2). Lithium-induced DI is well documented (found in up to one third of patients on long-term lithium therapy) but it usually presents whilst the patient is still receiving the medication, often with preceding signs or symptoms. We wish to highlight a case in which the DI did not manifest until many days after the cessation of the patient's lithium therapy.

CASE REPORT

A 72-year-old woman presented to the accident and emergency department after involvement in a road traffic accident. Her only medical history consisted of bipolar disorder, which was treated with 400 mg lithium and 1 mg trifluoperazine twice daily. The accident resulted in severe bilateral below-knee degloving injuries with fractures of the ankle and the fibular neck. She was cardiovascularly unstable due to severe blood loss from the degloving injuries and so was transferred to the operating theatre for wound debridement and haemostasis. Perioperatively she was transfused with 22 units of blood, four units of fresh frozen plasma, two units of platelets, three litres of colloid and 13 litres of crystalloid. Postoperatively she remained ventilated on the intensive care unit (ICU) where she developed disseminated intravascular coagulopathy, requiring further transfusions of blood, fresh frozen plasma, platelets and cryoprecipitate. The patient's lithium was not administered during her stay on the ICU and subtherapeutic levels were measured on the 10th day after admission.

On the first day of her stay on the ICU she developed mild renal impairment (urea 5.9 mmol/l, creatinine 108 [micro]mol/l) with oliguria. Although the creatine kinase was elevated at 1053 IU/l, massive rhabdomyolysis was excluded by a negative urine myoglobin test. On the fourth day the patient developed significant polyuria (150-450 ml/h) with a persistent hypernatraemia (maximum serum sodium 160 mmol/l). Administration of 2 [micro]g intravenous desmopressin (DDAVP) produced negligible reduction in the urine output and this therefore raised the possibility of a diagnosis of nephrogenic diabetes insipidus. At this point the urea was 6.6 mmol/l and creatinine was 118 [micro]mol/l; serum calcium and potassium levels were normal. Supportive therapy was continued whilst the polyuria persisted and on the 10th day in the ICU a further 1 [micro]g of DDAVP was given subcutaneously. Surprisingly, within an hour of the DDAVP injection the urine output fell significantly from greater than 200 ml/h to only 30-40 ml/h. However, there was no change in urine osmolality following the DDAVP and this confirmed the diagnosis of nephrogenic diabetes insipidus. The following morning the urine output once again increased to 400-450 ml/h and further doses of 2 [micro]g intravenous DDAVP had no effect. Oral bendrofluazide (2.5 mg daily) was commenced and over the next three days the urine output settled to 40-60 ml/h. The hypernatraemia also settled slowly, falling to 145 mmol/l after two weeks. Two days after the bendrofluazide was commenced the plasma and urine osmolalities were 311 and 447 mosmol/kg respectively.

After 35 days on the ICU the patient was discharged to the ward and underwent further plastic surgery to her legs. She was seen several months later out of hospital and recovering well.

DISCUSSION

One third of patients on long-term lithium therapy develop DI. The mechanism behind the effect of lithium on renal tubules has been well researched and there is now a better understanding of the effects of lithium at genetic and molecular levels. Water reabsorption in the kidney takes place by a number of mechanisms, but one important structure that is involved is a Vsopressin-regulated water channel known as aquaporin-2 (AQP2), which is found in the collecting ducts'. Vasopressin causes an increase in cytosolic 3:5-cyclic phosphatase which then acts as the second messenger to deliver AQP2 to the plasma membrane (3). Water is then reabsorbed from the urine via these channels.

Studies in rats have demonstrated that treatment with high-dose lithium results in dramatically reduced whole kidney abundance of AQP2 (4). This is due to markedly reduced levels of AQP2 mRNA expression in the kidneys of rats on lithium therapy (5). Furthermore, it is known that in some forms of DI there is also impaired delivery of the AQP2 channels to the apical plasma membrane (3). Thus, the combination of decreased production of AQP2 and reduced delivery to its site of action results in impaired reabsorption of water from the urine, i.e. nephrogenic diabetes insipidus.

Lithium-induced nephrogenic DI is well documented. What was unusual about this case was the delay of onset of the DI following the cessation of the lithium therapy (on admission). Polyuria did not develop until four days after the lithium had been discontinued and it persisted until the 13th day of admission, well after a subtherapeutic plasma level of lithium had been demonstrated. The molecular changes underlying lithium-induced DI go some way to explain this. Alterations in mRNA expression and AQP2 production and transport will take time to develop and thus their reversal will not be immediate on cessation of lithium. Studies have indeed shown that recovery of AQP2 expression can be slow, even after correction of the underlying condition (3).

The use of thiazide diuretics such as bendrofluazide in the treatment of nephrogenic DI may initially seem counter-intuitive. However, the AQP2 water channel has one more role to play. Further studies have shown that rats which were treated with lithium and developed DI and reduced AQP2 expression subsequently recovered their urinary concentrating abilities when treated with hydrochlorothiazide. Furthermore, it was demonstrated that the thiazide treatment caused a significant partial recovery in AQP2 abundance (6). Thus, thiazide diuretics are a useful tool in the management of diabetes insipidus.

In summary, this case highlights a delay in the development of diabetes insipidus in a patient who had been on prolonged lithium therapy. Whilst diabetes insipidus is a not uncommon complication of lithium, this case was unusual as the polyuria developed many days after the cessation of the patient's usual medication, without any preceding signs or symptoms. We have sought to explain this delay by a more detailed discussion of the molecular and genetic changes underlying lithium-induced diabetes insipidus. An understanding of these changes also sheds light on the well-documented use of thiazide diuretics in the treatment of diabetes insipidus. Whilst molecular genetics sometimes seems far removed from the hands-on management of a patient in clinical medicine, this is a good example of how knowledge of the former can have a direct impact on the latter.

REFERENCES

(1.) Waise A, Fisken RA. Unsuspected nephrogenic diabetes insipidus. Clinical Review. BMJ 2001; 323:96-97.

(2.) Ball SG, Baylis PH. The Neurohypophysis. In: Wass JAH, Shalet SM, ed. Oxford Textbook of Endocrinology and Diabetes, 1st ed. Oxford University Press 2002; 87-99.

(3.) Marples D, Frokiaer J, Knepper MA, Nielsen S. Disordered water channel expression and distribution in acquired nephrogenic diabetes insipidus. Proc Assoc Am Physicians 1998; 110:401-406.

(4.) Kwon TH, Laursen UH, Marples D, Maunsbach AB, Knepper MA, Frokiaer J et al. Altered expression of renal AQPs and Na(+) transporters in rats with lithium-induced NDI. Am J Physiol Renal Physiol 2000; 279:17552-564.

(5.) Laursen UH, Pihakaski-Maunsbach K, Kwon TH, Ostergaard Jensen E, Nielsen S, Maunsbach AB. Changes of rat kidney AQP2 and Na,K ATPase mRNA expression in lithium-induced nephrogenic diabetes insipidus. Nephron Exp Nephrol 2004; 97: e1-16.

(6.) Kim GH, Lee JW, Oh YK, Chang HR, Joo KW, Na KY et al. Antidiuretic effect of hydrochlorothiazide in lithium-induced nephrogenic diabetes insipidus is associated with upregulation of aquaporin-2, Na-Cl co-transporter, and epithelial sodium channel. J Am Soc Nephrol 2004; 15:2836-2843.

H. Paw *, M. E. Slingo ([dagger]), M. Tinker ([double dagger]) Intensive Care Department, York Hogoital, York, United Kingdom

* B.Pharm., M.R.Pharm.S., M.B., B.S., ER.C.A., Consultant in Anaesthesia and Intensive Care.

([dagger]) M.A. (Hons Cantab), M.B., B.Chir., Pre-registration House Officer.

([double dagger]) B.Sc.(Hons), M.B.Chb., Senior House Officer.

Address for reprints: Dr H. Paw, Anaesthesia and Intensive Care, York Hospital, Wigginton Road, York, Y031 8HE, U.K.