Hyperosmolar Coma and Lithium-induced Diabetes Insipidus
| Title: | Hyperosmolar Coma and Lithium-induced Diabetes Insipidus |
|---|---|
| Author: | Matz, Robert |
| Publisher: | Lancet |
| Date Published: | November 25, 1995 |
| Reference Number: | 46 |
SIR - MacGregor and colleagues (Aug 12, p 413) note that, "since no cause for the hyperglycaemia was evident, it was assumed to be due to dietary indiscretion". Hyperosmolality decreases pancreatic insulin secretion.1 Thus the inciting event in this case of non-insulin dependent diabetes mellitus may have been impairment of insulin secretion secondary to hyperosmolality from the lithium-induced nephrogenic diabetes insipidus.
During massive osmotic diureses induced by agents such as mannitol or glucose, by contrast with saline-induced osmotic diuresis, the urinary osmolality asymptotically approaches isotonicity with the serum. Indeed, during a solute diuresis, the urinary osmolality may reach isotonicity even though supraphysiological amounts of vasopressin (ADH) are present.2 In this sense an osmotic diuresis may cause partial nephrogenic diabetes insipidus (ie, resistance to the action of ADH).
Although the initial chest radiograph was normal there was clear indication of hypoxaemia and respiratory acidosis with arterial pH 7.32, PaCO2 40 mm Hg, and PaCO2 69 mm Hg, and a respiratory rate of 28. We described3 12 patients and found an additional nine reported with severely uncontrolled diabetes mellitus who presented with or rapidly developed adult respiratory distress syndrome (ARDS). Additionally we showed a widening alveolar arterial O2 gradient in several less critically ill uncontrolled diabetics as they were treated. This patient's arterial blood gases on admission and her progression to intubation suggest that she had a similar syndrome, perhaps related to her uncontrolled diabetes or treatment.
Finally, diabetes insipidus due to pituitary damage has been seen in association with the commonest fatal complication of diabetic ketoacidosis (DKA) in the under 20 age group--namely, cerebral oedema.4,5 Although this patient did not have this entity it is worthy of consideration, for completeness sake, in the differential diagnosis of the causes of persistent polyuria following treatment of severely uncontrolled diabetes mellitus.
Robert Matz
The Mount Sinai Medical Center, Mount Sinai Hospital,
Mount Sinai School of Medicine, New York, NY 10029, USA
- Gerich J, Penhos JC, Gutman RA, Recant L. Effect of dehydration and hyperosmolarity on glucose, free fatty acid and ketone body metabolism in the rat. Diabetes 1973; 22: 264-71.
- Howard RL, Bichet DG, Schrier RW. Hypernatremic and polyuric states. In: Seldin DW, Giebisch G, eds. The kidney: physiology and pathophysiology. 2nd edition. New York: Raven Press, 1992: 1753-78.
- Carroll P, Feinstein S, Nierenberg S, Matz R. The adult respiratory distress syndrome complicating diabetic ketoacidosis. Cardiovasc Rev Rep 1986; 9: 801-03.
- Taubin H, Matz R. Cerebral edema, diabetes insipidus and sudden death during the treatment of diabetic ketoacidosis. Diabetes 1968; 17: 108-09.
- Rosenbloom AL. Intracerebral crises during treatment of diabetic ketoacidosis. Diab Care 1990; 13: 22-33.
SIR--MacGregor and colleagues' report of a meeting of critical care physicians from Bowman Gray School of Medicine provided a thorough discussion of hyperosmolar coma and lithium-induced diabetes insipidus but failed to address the issue of lithium intoxication. A patient taking lithium who presents with impaired level of consciousness, pronounced volume depletion, and reduced renal function should be presumed to be lithium toxic until proven otherwise. No mention is made of serum lithium concentrations and no consideration is given to the role that toxic concentrations of lithium could be playing in the clinical presentation. Likewise, one wonders whether the patient would have benefited from haemodialysis, the treatment of choice for severe lithium intoxication. A further consideration with respect to the mechanism of action of lithium in causing nephrogenic diabetes insipidus is its effect on the water-channel protein aquaporin-2.1 In rats, chronic treatment with lithium greatly reduced the expression of aquaporin-2 in association with the development of polyuria and decreased urine osmolality.2
James W. Jefferson
Lithium Information Center, Dean Foundation for Health, Research and Education, Madison, WI 53717, USA
- Kanno K, Sasaki S, Hirata Y, et al. Urinary excretion of aquaporin-2 in patients with diabetes insipidus. N Engl J Med 1995; 332: 1540-45.
- Marples D, Christensen S, Christensen EI, et al. Lithium-induced downregulation of aquaporin-2 water channel expression in rat kidney medulla. J Clin Invest 1995; 95: 1838-45.
SIR--MacGregor and colleagues describe a woman with hyperosmolar coma due to a combination of non-ketotic diabetes mellitus and lithium-induced nephrogenic diabetes insipidus. An alternative approach to the management of severe hypernatraemia, as often seen in hyperosmolar non-ketotic diabetic coma, is to use haemofiltration to provide a sodium clamp, allowing a controlled reduction in serum osmolality and appropriate water balance. This procedure was first described in 1990 by Moss et al1 who applied continuous arteriovenous haemodiafiltration (CAVHDF) to a 4-year-old with pneumonia, hypernatraemia (serum sodium 189 mmol/L), and acute renal failure. Peritoneal dialysis was technically impossible and CAVHDF was initiated via radial artery and jugular vein. Ultrafiltrate volumes were small (maximum 104 mL/h) and blood pressure dependent, and dialysis/replacement fluids had a sodium concentration of 132 mmol/L. Because of the low blood flows, serum sodium slowly returned to 147 mmol/L over 4 days despite the large sodium gradient. The child eventually died from irreversible hypotension.
More recently in this unit, continuous venovenous haemofiltration (CVVHF) has been used to correct severe hypernatraemia associated with diabetic hyperosmolar, non-ketotic episodes. Gambro haemofilters have been used, producing 20-40 mL/min ultrafiltrate, replacing with prepared solutions of known sodium concentration, reducing serum sodium slowly to a target of 150 mmol/L.
Replacement with fluid of a sodium concentration 5-10 mmol/L below that of serum is done for several hours and then replacement sodium concentration is adjusted downwards. This cycle is repeated until serum sodium achieves the target concentration. Glycaemic control is obtained with intravenous insulin. Overall fluid balance is, of course, determined by the volume of replacement fluid and can accommodate polyuria. This approach allows a gradual, controlled return to normal serum sodium concentration and optimum fluid balance.
*Andrew Allen, Lui Forni, David Wright, Philip Hilton
*Department of Renal Medicine and Department of Intensive Care, St. Thomas' Hospital, Guy's and St. Thomas' Hospital Trust, Lambeth Palace Road, London SE1 7EH, UK
- Moss GD, Primavesi RJ, McGraw ME, Chambers TL. Correction of hypernatraemia with continuous arteriovenous haemodiafiltration. Arch Dis Child 1990; 65: 628-30.
SIR--In MacGregor and colleagues' patient the calculated osmolality on admission and day 4 were higher than the measured osmolality: the calculated osmolality on admission was 381 mOsmol/kg (22 mOsmol/kg higher) and on day 4 it was 392 (32 mOsmol/kg higher). Later, the difference between calculated and measured decreased to 11 mOsmol/kg. This is an unusual finding. Calculated and measured osmolality are usually closely related except in the presence of unmeasured osmoles such as mannitol or alcohol, or in the presence of pseudohypernatraemia. In all these instances measured osmolity is higher than the calculated osmolality. I wonder whether there is any explanation for the higher calculated osmolality in this patient?
R. Swaminathan
Division of Chemical Pathology, Guy's Hospital, London SE1 9RT, UK
Author's reply
SIR--Matz and Jefferson draw attention to a difficulty often encountered at a referral hospital--an incomplete database relative to the clinical scenario on presentation to the referring hospital. In our patient, the admitting physician made the diagnosis of non-ketotic hyperosmolar coma (NKHOC) due to diabetes mellitus. A search for the cause of the acute deterioration in glucose control was undertaken (blood counts, cultures, &c), and in the absence of an immediately identifiable cause, the NKHOC was attributed to dietary indiscretion, even though a dramatic alteration in diet should not lead to such severe hyperglycaemia. As a result, the hypernatraemia and the hyperosmolar state associated with lithium-induced nephrogenic diabetes insipidus (LINDI) were not diagnosed, and the initial fluid administration was with predominantly sodium-rich crystalloids. Fortunately, the initial management of this patient focused on volume resuscitation, which undoubtedly saved the patient's life, despite the ultimate problems of hypernatraemia, hyperosmolality, and LINDI.
Matz's reference to the fact that hyperosmolality decreases pancreatic secretion of insulin is especially important in this case.1 In instances of hyperglycaemia, as long as the patient is capable of increasing intake of water and renal perfusion is maintained, it is difficult to increase serum glucose concentrations above the renal threshold of about 16 mmol/L. Similarly, in patients with LINDI, osmolality remains fairly normal as long as the patient can respond to thirst and maintain oral intake of fluids. When fluid intake is limited (eg, in nausea, emesis, alteration in mental status), osmolality increases, renal perfusion decreases, glucose concentrations rise, and a vicious cycle ensues, progressing to a severe hyperosmolar state, such as was demonstrated by our patient.
The potential of acute lithium intoxication was considered in this case, but lithium concentrations were 1.1 mmol/L on admission and <0.1 mmol/L on transfer 5 days later. The idea of haemodialysis or CVVHF to correct serum sodium concentrations is very appealing, and certainly would have been used if our patient had shown a deterioration in renal function (oliguria). Unfortunately, we were faced with trying to meet the demands of massive polyuria (up to about 2.5 L/h), and adding additional CVVHF fluid replacement would have further stressed our ability to provide such huge volumes of fluids. In addition, although a more rapid reduction in sodium concentrations may have corrected the hyperosmolar state, accelerating the correction of hypernatraemia may have potentiated cerebral oedema, although there continues to be debate over the cause and treatment of cerebral oedema in patients with diabetic ketoacidosis and NKHOC.2
The physiological mechanisms by which lithium interferes with fluid and sodium homoeostasis continue to be elucidated. In addition to the reduction of cAMP in the distal tubule and collecting duct and the impairment of the cAMP effect on the tubule, there are two reports of alterations of the water-channel protein aquaporin-2 (AQP2).3,4 Lithium down-regulates AQP2 expression, which causes a vasopressin resistance that parallels the development of LINDI.4 Discontinuation of lithium is followed by a very slow return of normal AQP2 expression, which may lead to prolonged symptoms of diabetes insipidus, similar to what happened in our patient.
Swaminathan is concerned that calculated osmolality was considerably higher than measured on the admission and day 4 laboratory values. We have been able to identify several possible reasons why there is such a discrepancy. First, the osmolality measurements were done at different times than the electrolyte concentrations, and may not accurately reflect the declining glucose concentrations and fluid shifts occurring during the first few hours of admission. Second, the glucose concentration on admission and the sodium concentration on day 4 were very high, and fell near the extremes of measurement error for the analysers used. It is possible that the associated error range for these concentrations encompasses the difference between calculated and measured osmolality. Finally, erroneous results are a possibility, although the osmolality, glucose, and sodium measurements were repeated on numerous occasions with similar results.
Drew A. MacGregor
Department of Anaesthesia, Bowman Gray School of Medicine, Winston-Salem, NC 27157, USA
- Gerich J, Penhos JC, Gutan RA, Recant L. Effect of dehydration and hyperosmolarity on glucose, free fatty acid and ketone body metabolism in the rate. Diabetes 1973; 22: 264-71.
- Rosenbloom AL. Intracerebral crises during treatment of diabetic ketoacidosis. Diab Care 1990; 13: 22-33.
- Kanno K, Sasaki S, Hirata Y, et al. Urinary excretion of aquaporin-2 in patients with diabetes insipidus. N Engl J Med 1995; 332: 1540-45.
- Marples D, Christensen S, Christensen EI, Ottosen PD, Nielsen S. Lithium-induced downregulation of aquaporin-2 water channel expression in rat kidney medulla. J Clin Invest 1995; 95: 1838-45.
| Bowman Gray School of Medicine, Department of Anaesthesia, Winston-Salem, NC 27157, USA (D.A. MacGregor) November 1995 The Lancet, 42, Bedford Square, London, WC1B 3SL |
This translation by the NDI Foundation is to assist the lay reader. To provide a clear, accessible interpretation of the original article, we eliminated or simplified some technical detail and complicated scientific language. We concentrated our translation on those aspects of the article dealing directly with NDI. The NDI Foundation thanks the researchers for their work toward understanding and more effectively treating this disorder.
© Copyright NDI Foundation 2007 (JC)