Lithium has a narrow therapeutic range with serious adverse effects, predominantly in the central nervous system and thus requires routine monitoring. Acute ingestions are usually well tolerated, even when taken on a background of lithium therapy (acute on chronic poisoning), and do not usually require significant intervention unless renal failure or sodium depletion is present.
Chronic toxicity is more severe than acute toxicity and has a different clinical presentation. Repeated haemodialysis is often required in severe chronic poisoning. Long term neurological damage may occur and is probably multifactorial in origin.
Lithium is a small cation that distributes uniformly in body water replacing normal cations. Its therapeutic effect is not directly receptor mediated but appears to be due to downstream effects to alterations in intracellular enzymes including: a large variety of lithium-sensitive magnesium-dependent phosphoesterases, glycogen synthase kinase 3a and b and phosphoglucomutase.
In clinical practice central neurological toxicity is probably multifactorial and includes lithium effects on cellular ATP production, effects of osmolar and sodium shifts (related to nephrogenic diabetes insipidus), other risk factor in critical illness and possibly thiamine deficiency.
Nephrogenic diabetes insipidus is common in patients on chronic lithium treatment and is a major risk factor for development of chronic lithium toxicity. Lithium reduces the expression of water transporting aquaporins in the renal collecting ducts reducing the ability to concentrate urine.
The half-life in therapeutic use is 8–12 hours with normal renal function and hydration.
The half-life with toxic concentrations is often substantially longer (2–3 days even with active treatment to enhance elimination) as patients are usually significantly dehydrated and may have renal failure.
Therapeutic doses of lithium have high bioavailability and are rapidly absorbed from the small intestine. Peak concentrations occur in therapeutic use within 2–3 hours. However, sustained release preparations are available and may result in greatly prolonged absorption.
Following overdose, lithium concentrations are normally much lower than concentrations predicted by extrapolating therapeutic kinetics. This suggests either a saturable process of absorption and/or maintenance of high urinary lithium clearance.
Lithium is not protein bound and has a volume of distribution equal to body water (0.6 L/kg). Thus its clearance can be substantially increased by haemodialysis. It distributes into cells, where it exerts its major toxic effects, quite slowly. Complete equilibrium between serum and tissue concentrations may take several days to weeks. The same is true when serum concentrations are falling and thus CNS effects may persist for weeks. There is good transfer across the placenta; maternal use of lithium prior to delivery is associated with floppy babies.
Lithium is excreted unchanged in urine. After filtration, lithium is reabsorbed in the proximal tubule by the sodium transport mechanism. Reabsorption is proportional to the extent of sodium reabsorption; patients with sodium depletion have increased lithium reabsorption and reduced lithium clearance.
Alkaline urine increases lithium clearance as do diuretics that act on the proximal tubule. Hyponatraemia, dehydration and diuretics that act more distally (loop diuretics, thiazides and potassium sparing diuretics) all reduce lithium clearance.
Reduction of glomerular filtration for any reason (including NSAIDs) will also reduce lithium clearance
Lithium poisoning presentations exists on a clinical spectrum that is generally characterised into 3 clinical scenarios that have different prognosis and treatment responses:
All of these clinical presentations need to be considered in the context of the patient’s current renal function.
Chronic lithium treatment is associated with declining renal function due to a glomerular and tubulointerstitial nephropathy.
In toxicology, the major driver for lithium toxicity is lithium induced nephrogenic diabetes insipidus plus deterioration in renal function from any cause.
Virtually all patients will have some degree of nephrogenic diabetes insipidus and therefore, when oral intake is reduced, they may dehydrate relatively quickly as they have high renal free water loss.
Nephrogenic diabetes insipidus is common in patients on chronic lithium treatment. Up to 55% of patients report symptoms. Twenty per cent of patients produce more than 3 litres of urine a day, which is a major risk factor for development of chronic lithium toxicity.
Lithium reduces the expression of water transporting aquaporins in the renal collecting ducts probably through effects on cyclic AMP. This leads to a failure to concentrate urine appropriately.
While some patients or relatives will give a history of increased urine volumes and drinking for many patients, the symptoms may have become interpreted as normal.
Useful history includes how much water the patient takes to bed at night and whether they have nocturia and thirst. It is a mistake to assume that increased thirst in patients on lithium treatment is psychogenic polydipsia.
Diagnosis is normally confirmed by examining blood and urine electrolytes and osmolarity. In diabetes insipidus, urine osmolarity is inappropriately low in the presence of high plasma osmolarity.
The immediate short term management of NDI is difficult. Removal of lithium will not result in a rapid improvement in NDI as new aquaporin receptors need to be produced; in addition there is probably lithium trapping in the cortical cells. As the aquaporins are not available desmopressin is generally ineffective.
The mainstay of management is meticulous fluid balance. Once lithium concentrations are low, adjunct treatment that can be trialled includes thiazide diuretics, NSAIDs, sodium restriction and careful fluid restriction. Amiloride blocks lithium uptake into the distal cells and may prevent NDI; some patients with established NDI will have a significant reduction in urine volume with doses of 10–20 mg day, however this may take weeks (Bedford et al, 2008).
|Concentration (mmol/L)||Observed effects|
|2.5||Myoclonic and other involuntary movements, ataxia and confusion|
|3.0||Marked delirium, coma, seizures|
Clinical Grading (from Hanse and Amdisen, 1978)
|1||Mild||Nausea, vomiting, tremor, hyperreflexia, agitation, muscle weakness, ataxia|
|2||Moderate||Stupor, rigidity, hypertonia, hypotension|
|3||Severe||Myoclonus, collapse, seizures, coma|
The non CNS clinical features and the difference between acute and chronic poisoning are shown below. Some patients with an overdose who are on chronic therapy do not fit neatly into this acute/chronic dichotomy.
|Correlation of Li concentrations with clinical features||Poor||Better|
|Gastrointestinal effects (nausea, vomiting, diarrhoea)||42%||20%|
|Seizures, CNS effects||Delayed, less common, avoidable||Common with concentrations > 2.0 mmol/L|
|Renal effects||usually insignificant||Universal|
|ECG changes||Generally normal||QT prolongation. Ventricular arrhythmias may occur|
|Thyroid Disease||No||Hypothyroidism 20%, goitre very common|
|Recovery||Rapid and usual||Delayed up to 2–3 weeks, potential for chronic neurological injury|
Recovery after chronic poisoning may be delayed up to 2–3 weeks (until CNS lithium is cleared). In a series of 20 patients with chronic lithium toxicity, there was 10% mortality and 10% long term disability (Hansen & Amdisen), the rate of death in this series was much greater than subsequent series. The major residual neurological signs are cerebellar ataxia and dysarthria and little improvement occurs after 6 months.
Acute lithium poisoning in patients not previously on lithium is generally well tolerated when renal function is normal and the patient is sodium replete. Gastrointestinal symptoms such as nausea and diarrhoea are common. High concentrations of lithium can cause QT prolongation. In this situation equilibration into the CNS is much slower than renal clearance.
Most patients who present with an acute ingestion of lithium on background of chronic treatment have a clinical course similar to an acute ingestion in a lithium patient (Oakley et al, 2001 Waring et al 2007). However, as these patients already have significant CNS concentrations and may have nephrogenic diabetes insipidus they are at higher risk of developing CNS toxicity. They will require more careful management of fluid status and monitoring of renal function
Lithium has dose related toxicity in therapeutic use. The major adverse effects are predominantly in the central nervous system and develop over several days. Chronic poisoning is thus much more severe than acute poisoning for a given serum concentration.
The initial symptoms which can occur within the therapeutic range are tremor and polyuria (due to nephrogenic diabetes insipidus (NDI)). NDI, if pre-existing, is a major risk factor for chronic toxicity. Later symptoms are impaired level of consciousness, myoclonus, dysarthria and ataxia. With more severe poisoning the patient becomes comatose, has convulsions, acute renal failure and death.
Patients with signs of lithium toxicity or a lithium concentration greater than 1.5 mmol/L should be admitted. A concentration greater than 2.5–3.0 mmol/L, with corresponding clinical signs, would be an indication for haemodialysis in chronic lithium toxicity. The central nervous system findings are protean and dose dependent. A rough correlation between serum concentrations and clinical findings at presentation in chronic poisoning is shown above.
Note that adverse effects are common at the upper end of the therapeutic range. The correlation at high concentrations (> 3.0 mmol/L) is less clear as serum concentrations are usually rising rapidly due to dehydration even in chronic poisoning.
A classic clinical scenario is a downward spiral where a patient, often with some NDI, develops an intercurrent illness which affects their fluid/sodium balance status, causing reduced lithium clearance, rising concentrations and increasing toxicity often associated with reduced oral intake and increased renal and gastrointestinal losses. In this setting, patients can quite rapidly deteriorate into severe toxicity
Maintenance levels are typically 0.4–0.8 mmol/L. Many labs still report the upper end of the therapeutic range as 1.2 mmol/l; this range was derived from early naïve interpretations of kinetics.
Lithium concentrations are helpful in making the diagnosis of lithium exposure and guiding treatment in chronic poisoning. (See table)
The lithium concentration combined with the patients calculated GFR (Glomerular Filtration Rate) can be used to predict the fall in lithium concentration over 36 hours. In chronic lithium toxicity this may help inform the decision to use hemodialysis to enhance elimination.
The lithium concentration alone should not be used to guide treatment in acute poisoning. Patients with lithium concentrations within the therapeutic range may still be toxic; in particular if their concentrations had been high in the last week. This reflects the slow equilibration of lithium out of the CNS as well as slow restoration of intracellular concentrations.
Patients should have electrolytes and renal function measured at regular intervals Along with serum and urine osmolarity and electrolytes.
Renal tubular abnormalities can lead to acidosis, hypokalaemia, hyponatraemia and hypernatraemia.
Thyroid function should be checked.
Hypercalcaemia is a potential problem, hyperparathyroidism is associated with lithium treatment
Rehydration and correction of any hyponatraemia are the cornerstone of initial management.
Patients with nephrogenic diabetes insipidus require careful fluid balance that normally requires matching of the previous hour’s urine output along with correcting any pre-existing deficit. Electrolytes need to be monitored frequently to avoid large changes in sodium which may contribute to neurological injury.
Once the sodium depletion is corrected, patients with NDI typically require a normal sodium intake plus an increase in free water. This is often easier to acheive if the patient is able to drink.
Patients with CNS involvement should receive thiamine, initially 500 mg tds.
Interacting drugs should be ceased.
Patients with an abnormal ECG should be monitored.
Indications for admission
Admission to ICU should be done for all patients requiring haemodialysis and for those with ECG changes.
Activated charcoal does not bind to lithium. Since little clinical effect occurs after acute poisoning in the presence of normal renal function aggressive measures to decontaminate the bowel are not generally indicated. Whole bowel irrigation with polyethylene glycol could be considered in patients presenting within 1-2 hours of ingestion of a very large dose (> 50 g). Sustained release preparations may have a longer course because of delayed absorption but if renal function remains normal and the patient is sodium replete the lithium is rapidly cleared without much clinical effect and decontamination is not particularly indicated.
Haemodialysis increases lithium clearance. Repeated haemodialysis is frequently required due to rebound in lithium concentrations within 6-12 hours of ceasing haemodialysis. This occurs due to redistribution of lithium from peripheral tissues. Lithium concentrations should be measured 6-12 hours post dialysis to determine if further dialysis is indicated.
For symptomatic lithium toxicity, continuous veno-venous haemodialysis (CVVHD) can be used to continue enhanced lithium clearance after initial haemodialysis has reduced the initial peak in lithium concentration. CVVHD for 24 hours is about as effective as a 4 hour haemodialysis and may be used without prior haemodialysis in asymptomatic patients with elevated lithium concentrations and renal failure.
Indications for haemodialysis
The major indication for haemodialysis is lithium toxicity in the setting of impaired renal lithium excretion. If renal lithium clearance is normal most patients can be managed conservatively.
Dialysis should be continued until serum lithium concentration 6–12 hours following dialysis is less than 0.6 mEq/L. Rebound in lithium concentrations is not usually a problem after CVVHD.
Recovery after chronic poisoning may be delayed up to 2–3 weeks (until CNS lithium is cleared). In early case series (1970s) death due to medical complications and long term disability each occur in about 10% of chronic poisoning (Hansen and Amdisen, 1978). Death is now rare in the context of good supportive care.
The major residual neurological signs are cerebellar ataxia and dysarthria and little improvement occurs after 6 months.
Hansen HE, Amdisen A . Lithium intoxication. Report of 23 cases and review of 100 cases from the literature. Q J Med 1978;47:123–44
Waring WS, Laing WJ, Good AM, Bateman DN. Pattern of lithium exposure predicts poisoning severity: evaluation of referrals to a regional poisons unit. QJM 2007 May 100(5):271–276.
Markowitz GS, Radhakrishnan J, Kambham N, Valeri AM, Hines WH, D'Agati VD. Lithium nephrotoxicity: a progressive combined glomerular and tubulointerstitial nephropathy. J. Am. Soc. Nephrol 2000 Aug;11(8):1439–1448
Bocchetta A, Bernardi F, Pedditzi M, Loviselli A, Velluzzi F, Martino E, Del Zompo M. Thyroid abnormalities during lithium treatment. Acta Psychiatr Scand 1991;83:193-198.
Ellenhorn MJ: Medical Toxicology: Diagnosis and Treatment of Human Poisoning. Williams & Wilkens, Baltimore, 1997.
Harvey NS & Merriman S. Review of clinically important drug interactions with lithium. Drug Safety 1994;10(6):455-63.
Schou M. Long lasting neurological sequelae after lithium intoxication. Acta Psychiatr Scand 1984;70:594-602
Oakley PW, Whyte IM, Carter GL. Lithium toxicity: an iatrogenic problem in susceptible individuals. Aust N Z J Psychiatry 2001 Dec;35(6):833-40
Beckmann U, Oakley PW, Dawson AH, Byth PL. Efficacy of continuous venovenous haemodialysis in the treatment of severe lithium toxicity. J Toxicol Clin Toxicol 2001;39(4):393-7
Oakley PW, Dawson AH, Whyte IM. Lithium: thyroid effects and altered renal handling. J Toxicol Clin Toxicol 2000;38(3):333-7
Bedford JJ, Weggery S, Ellis G, McDonald FJ, Joyce PR, Leader JP, Walker RJ. Lithium-induced nephrogenic diabetes insipidus: renal effects of amiloride. Clin J Am Soc Nephrol. 2008 Sep;3(5):1324–31.