Chronic Lithium Administration Research Articles

Chronic lithium chloride administration to rats elevates glucose metabolism in wide areas of brain, while potentiating negative effects on metabolism of dopamine D2-like receptor stimulation

The regional cerebral metabolic rate for glucose (rCMRglc) can be imaged in vivo as a marker of brain functional activity. The effects of chronic lithium administration on baseline values of rCMRglc and values in response to administration of dopamine D2-like receptor agonists have not been examined in humans or rats. Knowing these effects may elucidate and localize the therapeutic action of lithium in bipolar disorder. In unanesthetized rats, we used the 2-deoxy-D-glucose (2-DG) technique to image the effects of a 6-week control diet or LiCl diet sufficient to produce a plasma lithium concentration therapeutically relevant to bipolar disorder, on rCMRglc at baseline and in response to the dopaminergic D2-like receptor agonist, quinpirole (1 mg/kg i.v.), or to i.v. saline. Baseline rCMRglc was significantly elevated in 30 of 81 brain regions examined, in LiCl diet compared with control diet rats. Affected were visual and auditory structures, frontal cortex, amygdala, hippocampus, nucleus accumbens, caudate-putamen, interpeduncular nucleus, and substantia nigra. Acute quinpirole significantly decreased rCMRglc in four areas of the caudate-putamen in control diet rats, and in these and 19 additional brain areas in LiCl-fed rats. In unanesthetized rats, chronic lithium administration widely upregulates baseline rCMRglc and potentiates the negative effects on rCMRglc of D2-like receptor stimulation. The baseline elevation may relate to lithium's reported ability to increase auditory and visual evoked responses in humans, whereas lithium's potentiation of quinpirole's negative effects on rCMRglc may be related to its therapeutic efficacy in bipolar disorder.

Chronic fluoxetine upregulates activity, protein and mRNA levels of cytosolic phospholipase A2 in rat frontal cortex

Chronic lithium and carbamazepine, which are effective against mania in bipolar disorder, decrease the activity of cytosolic phospholipase A(2) (cPLA(2)) and the turnover rate of arachidonic acid in phospholipids in rat brain. Assuming that stages of bipolar disorder are related to brain arachidonic acid metabolism, we hypothesized that drugs effective in depression would increase cPLA(2) activity. To test this hypothesis, adult male CDF-344 rats were administered fluoxetine (10 mg/kg intraperitoneally (i.p.) or saline (control) (i.p.) chronically for 21 days. Frontal cortex cPLA(2) protein, phosphorylated cPLA(2), activity and mRNA levels were increased after chronic fluoxetine. Transcription factors (activator protein-1, activator protein-2, glucocorticoid response element, polyoma enhancer element-3 and nuclear factor-kappa B) that are known to regulate cPLA(2) gene expression were not significantly changed by chronic fluoxetine, but nuclear AU-rich element/poly(U)-binding/degradation factor-1 RNA-stabilizing protein was increased significantly. The results suggest that chronic fluoxetine increases brain cPLA(2) gene expression post-transcriptionally by increasing cPLA(2) mRNA stabilization. Chronic fluoxetine's effect on cPLA(2) expression was opposite to the effect reported with chronic lithium or carbamazepine administration, and may be part of fluoxetine's mode of action.

Lithium increases bcl-2 expression in chick cochlear nucleus and protects against deafferentation-induced cell death

Approximately 20–30% of neurons in the avian cochlear nucleus (nucleus magnocellularis) die following deafferentation (i.e. deafness produced by cochlea removal) and the remaining neurons show a decrease in soma size. Cell death is generally accepted to be a highly regulated process involving various pro-survival and pro-death molecules. One treatment that has been shown to modify the expression of these molecules is chronic administration of lithium. The present experiments examined whether lithium treatment can protect neurons from deafferentation-induced cell death. Post-hatch chicks were treated with LiCl or saline for 17 consecutive days, beginning on the day of hatching. On the 17th day, a unilateral cochlea ablation was performed. Five days following surgery, the nucleus magnocellularis neurons were counted stereologically on opposite sides of the same brains. Lithium reduced deafferentation-induced cell death by more than 50% (9.8% cell death as compared with 22.4% in saline-treated subjects). Lithium did not affect cell number on the intact side of the brain. Lithium also did not prevent the deafferentation-induced decrease in soma size, suggesting a dissociation between the mechanisms involved in the afferent control of soma size and those involved in the afferent control of cell viability. A possible mechanism for lithium’s neuroprotective influence was examined in a second set of subjects. Previous studies suggest that the pro-survival molecule, bcl-2, may play a role in regulating cell death following deafferentation. Tissues from lithium- and saline-treated subjects were examined using immunocytochemistry. Chronic administration of lithium dramatically increased the expression of bcl-2 protein in nucleus magnocellularis neurons. These data suggest that lithium may impart its neuroprotective effect by altering the expression of molecules that regulate cell death.

Chronic Carbamazepine Decreases the Incorporation Rate and Turnover of Arachidonic Acid but Not Docosahexaenoic Acid in Brain Phospholipids of the Unanesthetized Rat: Relevance to Bipolar Disorder

The basis for carbamazepine's efficacy in treating bipolar disorder is not agreed on. One hypothesis is that, similar to lithium and valproate (antibipolar drugs), carbamazepine might selectively decrease the kinetics of arachidonic acid (AA) in brain phospholipids. To assess whether it targets brain AA kinetics, we administered carbamazepine (25 mg/kg/day, IP) to rats for 30 days and then determined its effect compared with that of vehicle on incorporation and turnover rates of AA and docosahexaenoic acid (DHA) in brain phospholipids. In unanesthetized rats that had received carbamazepine or vehicle, [1-14C]AA or [1-14C]DHA was infused intravenously, and arterial blood plasma was sampled until the animal was killed at 5 min and its brain, after being microwaved, was used for acyl-coenzyme A (acyl-CoA) and phospholipid fatty acid analysis. Chronic carbamazepine, compared with vehicle, decreased the rate of incorporation of AA-CoA (27%-29%) and turnover of AA (25%-27%) but not of DHA-CoA or DHA in brain phospholipids. The results, which are comparable to published findings after chronic administration of lithium and valproic acid to rats, support the hypothesis that drugs effective against mania in bipolar disorder act by selectively downregulating the incorporation rate of AA-CoA and turnover of AA in brain phospholipids.

Chronic valproate does not alter the kinetics of docosahexaenoic acid within brain phospholipids of the unanesthetized rat

It has been reported that each of three drugs effective in treating bipolar disorder (lithium, carbamazepine, and valproate) decreases the turnover of arachidonic acid (AA, 20:4n-6) in brain phospholipids of the awake rat. It is also known that lithium and carbamazepine do so without decreasing the turnover of docosahexaenoic acid (DHA, 22:6n-3). The aim of this study was to see whether valproate also specifically targets the turnover of AA but not of DHA in brain phospholipids. Valproate was administered (200 mg kg(-1), i.p.) to rats for 30 days to produce a therapeutically relevant plasma concentration and then determine its effect compared with that of vehicle on incorporation and turnover rates of DHA in brain phospholipids. In unanesthetized rats that had received valproate or vehicle, [1-14C]DHA was infused intravenously, and arterial blood plasma was sampled until the animal was killed at 5 min; and its brain, after being microwaved, was subjected to chemical and radiotracer analysis. Using equations derived from our fatty acid model, it was found that chronic valproate compared with vehicle did not alter the rate of incorporation or turnover of DHA in brain phospholipids. Valproate-treated animals had higher concentrations of linoleic acid (18:2n-6) in several brain phospholipids, supporting the hypothesis that it alters brain n-6 fatty acid metabolism. The results, comparable to published findings following chronic administration of lithium and carbamazepine to rats, support the hypothesis that drugs are effective against mania in bipolar disorder act by downregulating incorporation and turnover of AA, but not of DHA, in brain phospholipids.

Lithium-induced alterations in nucleoside triphosphate levels in human brain: a proton-decoupled 31P magnetic resonance spectroscopy study

We examined how lithium's demonstrated effects on various cellular processes in human brain would be reflected in the 31P magnetic resonance spectra of living human beings with respect to brain high-energy phosphate metabolites. Eight healthy volunteers received a baseline 31P magnetic resonance spectroscopy (MRS) scan, after which they received lithium carbonate, 900 mg/day, for 14 days. Follow-up MRS scans were obtained on day 7 and on day 14. We detected a lithium-induced decrease in alpha-, beta-, gamma- and total nucleoside triphosphate NTP levels with chronic administration of lithium. On day 7, significant decreases were noted in gamma-NTP (14%) and total NTP (11%) levels. There was a trend for a decrease in beta-NTP (11%) levels. On day 14, significant decreases were noted in alpha-NTP (7%) and total NTP (8%) levels. There was a trend for a decrease in beta-NTP (16%) levels. Lithium caused a 25% reduction in inorganic phosphate (P i) levels on day 14. The theoretical relevance of the lithium-induced alterations on brain high-energy phosphates to the lithium-induced modifications of neuroplasticity is discussed.

Effect of chronic lithium treatment on glucocorticoid and 5-HT1A receptor messenger RNA in hippocampal and dorsal raphe nucleus regions of the rat brain

The therapeutic mechanism of action of lithium in the treatment of bipolar disorder is not well understood. Dysfunction of both 5-HT1A receptor mediated neurotransmission and the glucocorticoid receptor is associated with mood disorders, and preclinical studies suggest that lithium treatment can modulate these receptor subtypes. In this study, we investigated the effect of chronic lithium treatment on 5-HT1A receptors and glucocorticoid receptors in the rat brain. Male Sprague-Dawley rats were treated with lithium (3 mmol/kg/day) or saline for 28 days via subcutaneous implanted mini-osmotic pumps. After 28 days of treatment, the expression of mRNA for 5-HT1A receptors and glucocorticoid receptors in the rat hippocampus and dorsal raphe nucleus was determined by in situ hybridization histochemistry. Chronic administration of lithium decreased mRNA coding for post-synaptic 5-HT1A receptors in hippocampal subregions but not for somatodentritic 5-HT1A receptors in the dorsal raphe nucleus. Chronic administration of lithium did not affect mRNA coding for glucocorticoid receptors in hippocampal subregions or in the dorsal raphe nucleus. Mean plasma lithium levels in the lithium-treated group were 0.50 ± 0.03 mmol/l; all animals appeared healthy and maintained a normal increase in body weight. Given recent reports implicating hypersensitive post-synaptic 5-HT1A receptors in bipolar manic patients, the present study suggests that down-regulation of this receptor population may be important in the therapeutic mechanism of action of lithium.

Effect of chronic lithium treatment on glucocorticoid and 5-HT1A receptor messengerRNA in hippocampal and dorsal raphe nucleus regions of the rat brain

The therapeutic mechanism of action of lithium in the treatment of bipolar disorder is not well understood. Dysfunction of both 5-HT(1A) receptor mediated neurotransmission and the glucocorticoid receptor is associated with mood disorders, and preclinical studies suggest that lithium treatment can modulate these receptor subtypes. In this study, we investigated the effect of chronic lithium treatment on 5-HT(1A) receptors and glucocorticoid receptors in the rat brain. Male Sprague-Dawley rats were treated with lithium (3 mmol/kg/day) or saline for 28 days via subcutaneous implanted mini-osmotic pumps. After 28 days of treatment, the expression of mRNA for 5-HT(1A) receptors and glucocorticoid receptors in the rat hippocampus and dorsal raphe nucleus was determined by in situ hybridization histochemistry. Chronic administration of lithium decreased mRNA coding for post-synaptic 5-HT(1A) receptors in hippocampal subregions but not for somatodentritic 5-HT(1A) receptors in the dorsal raphe nucleus. Chronic administration of lithium did not affect mRNA coding for glucocorticoid receptors in hippocampal subregions or in the dorsal raphe nucleus. Mean plasma lithium levels in the lithium-treated group were 0.50 +/- 0.03 mmol/l; all animals appeared healthy and maintained a normal increase in body weight. Given recent reports implicating hypersensitive post-synaptic 5-HT(1A) receptors in bipolar manic patients, the present study suggests that down-regulation of this receptor population may be important in the therapeutic mechanism of action of lithium.

Chronic Lithium Administration to Rats Selectively Modifies 5-HT2A/2C Receptor-Mediated Brain Signaling Via Arachidonic Acid

The effects of chronic lithium administration on regional brain incorporation coefficients k* of arachidonic acid (AA), a marker of phospholipase A2 (PLA2) activation, were determined in unanesthetized rats administered i.p. saline or 1 mg/kg i.p. (+/-)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane hydrochloride (DOI), a 5-HT2A/2C receptor agonist. After injecting [1-(14)C]AA intravenously, k* (brain radioactivity/integrated plasma radioactivity) was measured in each of 94 brain regions by quantitative autoradiography. Studies were performed in rats fed a LiCl or a control diet for 6 weeks. In the control diet rats, DOI significantly increased k* in widespread brain areas containing 5-HT2A/2C receptors. In the LiCl-fed rats, the significant positive k* response to DOI did not differ from that in control diet rats in most brain regions, except in auditory and visual areas, where the response was absent. LiCl did not change the head turning response to DOI seen in control rats. In summary, LiCl feeding blocked PLA2-mediated signal involving AA in response to DOI in visual and auditory regions, but not generally elsewhere. These selective effects may be related to lithium's therapeutic efficacy in patients with bipolar disorder, particularly its ability to ameliorate hallucinations in that disease.

Effect of chronic lithium treatment on the turnover of α 2-adrenoceptors after chemical inactivation in rats

One of the most effective psychotherapeutic agents in the treatment of bipolar disease is lithium. Chronic lithium treatment affects some signal transduction mechanisms such as cAMP, cGMP, inositol 1,4,5 P 3, Gi protein, protein kinase C and can also modify gene expression in rat brain. In a previous study, we observed a greater inhibitory effect of lithium on cAMP production after blockade of α 2-adrenoceptors in rat cerebral cortex. Here we examine the influence of chronic lithium treatment on turnover of α 2-adrenoceptors after their inactivation by N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline (EEDQ) in rat cerebral cortex. After treatment with lithium for 10 days (120 mg/kg/day, i.p.), there was a significant increase in the appearance and disappearance rate constants of these adrenoceptors and a significant reduction of their half-life. These results suggest that chronic lithium administration alters the α 2-adrenoceptor turnover in rat brain.

Chronic lithium administration enhances noradrenergic responses to intravenous administration of the alpha2 antagonist idazoxan in healthy volunteers.

The acute and chronic effects of lithium carbonate administration at therapeutic blood levels on peripheral noradrenergic activity and sympathetic responses to alpha2 adrenoceptor blockade were examined in 10 medically and psychiatrically healthy volunteers. Supine resting levels of plasma norepinephrine and the increases in norepinephrine following intravenous infusion of 200 microg/kg of idazoxan, a selective alpha2 adrenoceptor antagonist, were determined before lithium (Li+) administration and after 5 days and after 4 weeks of daily Li+ treatment. Chronic Li+ treatment significantly increased mean resting plasma norepinephrine levels by 53.6%. The noradrenergic responses to infusions of idazoxan were slightly enhanced after 5 days of Li+ administration and significantly increased following 4 weeks of Li+ treatment. The possibility that Li+ produces functional alpha2 subsensitivity causing enhanced peripheral noradrenergic activity in humans is supported by the findings of increased mean resting plasma norepinephrine and increased response to idazoxan following chronic Li+ administration. Alteration of regulatory mechanisms in the noradrenergic system may be relevant to understanding the clinical effects of Li+ in manic-depressive illness.

Mood stabilizers: protecting the mood…protecting the brain

The mechanism underlying the therapeutic action of mood stabilizers in bipolar disorder is not completely understood. The discovery that anticonvulsant agents, such as valproate (VPA), were effective in the treatment of bipolar disorder suggested a common biochemical mechanism(s) with lithium. Recent research has focused on how VPA and lithium change the activities of cellular signal transduction systems, especially the cyclic AMP and phosphoinositide second messenger pathways. Despite being structurally dissimilar, VPA produces effects on the protein kinase C (PKC) signalling pathway that are similar to lithium, although the VPA effects appear to be largely independent of myo-inositol. Furthermore, the therapeutic benefit of either drug require a prolonged administration suggesting alterations at the genomic level. Studies have revealed that both VPA and lithium altered the expression of several early inducible genes belonging to the AP-1 family of transcription factors; this family is responsible for controlling the expression of a number of genes including cytoprotective proteins such as the anti-apoptotic protein, bcl-2. Evidence shows that chronic administration of VPA or lithium can stimulate bcl-2 expression as well as inhibit GSK-3β activity, which renders a cell less susceptible to apoptosis. Thus, the mood stabilizers may act to restore the balance among aberrant signalling pathways in specific areas of the brain and prevent degeneration.

Effects of chronic lithium and sodium valproate on concentrations of brain amino acids

The present study was designed to determine if the mood stabilizers, lithium and valproate, have common effects on concentrations of amino acid neurotransmitters which may be related to their mechanisms of action. Two separate groups of rats were administered therapeutic doses of lithium, sodium valproate, or saline for 2 weeks. Whole brain extracts were then examined using either high-field 1H NMR spectroscopy or HPLC. Both drugs decreased whole brain concentrations of aspartate, glutamate, and taurine while brain concentrations of γ-aminobutyric acid (GABA) and alanine decreased following chronic sodium valproate administration but not following chronic lithium administration. These findings indicate that lithium and sodium valproate share common effects on the concentrations of certain amino acid neurotransmitters in whole brain which may be related to their mechanisms of action in bipolar disorder.

Chronic lithium administration potentiates brain arachidonic acid signaling at rest and during cholinergic activation in awake rats.

Studies were performed to determine if the reported 'proconvulsant' action of lithium in rats given cholinergic drugs is related to receptor-initiated phospholipase A2 signaling via arachidonic acid. Regional brain incorporation coefficients k* of intravenously injected [1-14C]arachidonic acid, which represent this signaling, were measured by quantitative autoradiography in unanesthetized rats at baseline and following administration of subconvulsant doses of the cholinergic muscarinic agonist, arecoline. In rats fed LiCl for 6 weeks to produce a therapeutically relevant brain lithium concentration, the mean baseline values of k* in brain auditory and visual areas were significantly greater than in rats fed control diet. Arecoline at doses of 2 and 5 mg/kg intraperitoneally increased k* in widespread brain areas in rats fed the control diet as well as the LiCl diet. However, the arecoline-induced increments often were significantly greater in the LiCl-fed than in the control diet-fed rats. Lithium's elevation of baseline k* in auditory and visual regions may correspond to its ability in humans to increase auditory and visual evoked responses. Additionally, its augmentation of the k* responses to arecoline may underlie its reported 'proconvulsant' action with cholinergic drugs, as arachidonic acid and its eicosanoid metabolites can increase neuronal excitability and seizure propagation.

Dose–response effects of chronic lithium regimens on spatial memory in the black molly fish

Lithium is widely used in the management of bipolar disorder, yet memory impairment is a serious side effect. To assess the effects of lithium on spatial working and reference memories, we have employed a plus maze utilizing spontaneous alternation (SA) and place-learning paradigms in two experiments with the black molly fish. Four treatment groups were gavaged with 20 μl of a 10, 100, or 1000 mM lithium chloride (LiCl) solution or ddH 2O vehicle every 12 h for 22 to 24 days. On Day 15, subjects began an 8-day SA task or a 10-day place-learning task. Results indicate that there is a significant difference in SA performance among the treatment groups for Days 1, 2, and 3. Results of the place-learning task indicate that the 1 M dose group needed significantly more trials to reach criterion and made significantly fewer correct first choices than the other dose groups. Capillary ion analysis determinations of plasma and brain lithium levels illustrate linear dose–response relationships to doses administered. Regression analyses indicate that there is a relationship between SA performance and plasma/brain lithium levels during the initial part of testing. Collectively, the results indicate that chronic lithium administration impairs spatial working and reference memories.

Chronic lithium downregulates cyclooxygenase-2 activity and prostaglandin E(2) concentration in rat brain.

Rats treated with lithium chloride for 6 weeks have been reported to demonstrate reduced turnover of arachidonic acid (AA) in brain phospholipids, and decreases in mRNA and protein levels, and enzyme activity, of AA-selective cytosolic phospholipase A(2)(cPLA(2)). We now report that chronic lithium administration to rats significantly reduced the brain protein level and enzyme activity of cyclooxygenase-2 (COX-2), without affecting COX-2 mRNA. Lithium also reduced the brain concentration of prostaglandin E(2) (PGE(2)), a bioactive product of AA formed via the COX reaction. COX-1 and the Ca(2+)-independent iPLA(2) (type VI) were unaffected by lithium. These and prior results indicate that lithium targets a part of the AA cascade that involves cPLA(2) and COX-2. This effect may contribute to lithium's therapeutic action in bipolar disorder.

Recent studies on interactions between n-3 and n-6 polyunsaturated fatty acids in brain and other tissues.

Recent literature provides a basis for understanding the behavioral, functional, and structural consequences of nutritional deprivation or disease-related abnormalities of n-3 polyunsaturated fatty acids. The literature suggests that these effects are mediated through competition between n-3 and n-6 polyunsaturated fatty acids at certain enzymatic steps, particularly those involving polyunsaturated fatty acid elongation and desaturation. One critical enzymatic site is a delta6-desaturase. On the other hand, an in-vivo method in rats, applied following chronic n-3 nutritional deprivation or chronic administration of lithium, indicates that the cycles of de-esterification/re-esterification of docosahexaenoic acid (22:6n-3) and arachidonic acid (20:4n-6) within brain phospholipids operate independently of each other, and thus that the enzymes regulating each of these cycles are not likely sites of n-3/n-6 competition.

Lithium inhibits the development of physical dependence to clonidine in mice.

Based on our previous finding that chronic lithium treatment reduced naloxone-precipitated withdrawal syndrome in morphine-treated mice, the effect of chronic lithium treatment was evaluated on the development of dependence to clonidine. Dependence was induced by injection of either morphine (50, 50 and 75 mg/kg, intraperitoneally with 3 hr interval for 3 consecutive days), or clonidine (2 mg/kg/day, intraperitoneally for 10 days). Naloxone (4 mg/kg, intraperitoneally) precipitated withdrawal signs in both morphine- and clonidine-treated mice. Yohimbine (5 mg/kg, intraperitoneally) precipitated withdrawal signs in the clonidine-treated mice, similar to morphine withdrawal signs; but failed to precipitate any significant sign in the morphine-treated mice. Coadministration of lithium was carried out by adding lithium chloride to drinking water (600 mg/l for 20 days; 10 days before the beginning of clonidine administration and 17 days before the administration of morphine to allow the lithium concentration to reach steady-state). The results indicated that chronic lithium administration significantly attenuated the withdrawal signs, precipitated either by yohimbine or naloxone, in clonidine-treated mice. As a conclusion, clonidine withdrawal signs are very similar to opioid withdrawal signs, and lithium is able to prevent the development of physical dependence to clonidine.

Effects of imipramine and lithium on wet-dog shakes mediated by the 5-HT 2A receptor in ACTH-treated rats

We examined the influence of imipramine and lithium on wet-dog shakes induced by the (±)-DOI, 5-HT 2A receptor agonist in adrenocorticotropic hormone (ACTH)-treated rats. The administration of imipramine for 14 days decreased the (±)-DOI-induced wet-dog shakes response; chronic administration of lithium for 14 days, however, had no effect. Chronic ACTH (100 μg/rat sc) treatment increased the wet-dog shake response induced by (±)-DOI. This effect of ACTH for 14 days, increasing the (±)-DOI-induced wet-dog shakes, was not inhibited by a 14-day administration of imipramine. Chronic coadministration of imipramine and lithium, lasting 14 days, decreased the wet-dog shakes response induced by (±)-DOI in rats treated with ACTH for 14 days. These findings indicate that lithium inhibits the hyperfunction of the 5-HT 2A receptor in rats treated with ACTH when coadministered with imipramine.

Lithium toxicity: an iatrogenic problem in susceptible individuals.

Lithium toxicity, manifesting primarily as neurotoxicity, is a significant health problem and is primarily iatrogenic in nature. Despite 50 years of medical experience with lithium, factors contributing to the development of severe neurotoxicity remain poorly documented. We hypothesized that severe neurotoxicity represents the most clinically significant manifestation of lithium toxicity. We proposed that this occurs primarily in the context of chronic therapeutic administration ('chronic poisoning'), rather than in the context of an overdose. Furthermore we hypothesized that patients who developed chronic poisoning did so in the presence of identifiable factors which predictably impair lithium clearance. A retrospective analysis of 97 cases of lithium poisoning, treated at a regional centre over a 13-year period was performed. Demographic data and factors considered likely to relate to the risk of developing lithium toxicity were recorded. Patients were classified according to mode of poisoning (acute, acute on chronic, or chronic) and according to severity of neurotoxicity (nil, mild, moderate, severe). The risk of developing severe neurotoxicity as a result of each mode of poisoning was assessed. The association between various risk factors and the development of chronic poisoning was assessed using a logistic regression model. Twenty-eight cases were rated as suffering severe neurotoxicity; in 26 this developed in the context of chronic poisoning and in two in the context of acute on chronic poisoning. All patients who developed severe neurotoxicity had at least one putative risk factor present, regardless of mode of poisoning. Length of stay was significantly longer for cases with severe neurotoxicity compared to those without severe neurotoxicity (12 vs. 2 days, P < 0.001). Peak serum lithium concentrations were significantly higher in cases with severe neurotoxicity compared to those without (2.3 vs. 1.6 mmol/L, P = 0.02). Patients presenting with chronic poisoning had a substantially higher risk of severe neurotoxicity than those presenting after an overdose of lithium (Odds Ratio [OR] 136, 95% CI 23-1300). A logistic regression model showed three factors contributed independently to the risk of chronic poisoning. These were: nephrogenic diabetes insipidus (adjusted OR 26.96, 95% CI 2.89-251.94), age over 50 years (adjusted OR 6.20, 95% CI 1.36-28.32) and thyroid dysfunction (adjusted OR 9.30, 95% CI 1.36-63.66). A fourth factor, baseline endogenous creatinine clearance below normal limits, was significant at the P = 0.05 level (adjusted OR 6.49, 95% CI 0.98-43.01). Hyperparathyroidism was noted in three cases of chronic poisoning suffering severe neurotoxicity. Severe lithium neurotoxicity occurs almost exclusively in the context of chronic therapeutic administration of lithium, and rarely results from acute ingestion of lithium, even in patients currently taking lithium. As such it is an iatrogenic illness, occurring in patients who have identifiable clinical risk factors: nephrogenic diabetes insipidus, older age, abnormal thyroid function and impaired renal function. Although administration of drugs which impair lithium clearance appeared to contribute minimally to chronic lithium poisoning in the absence of other factors, these drugs may well 'uncover' the predisposing risk factors and certainly should not be considered safe to use as a consequence of this study. The serious morbidity suffered by lithium toxic patients, and the cost to society due to long hospital stays, might be reduced by careful prescribing, vigilant monitoring and awareness of these factors, as they develop in otherwise stable patients. Review of existing therapeutic guidelines may be warranted.