Lithium Nephropathy

Updated: Jan 18, 2022
  • Author: Eleanor Lederer, MD, FASN; Chief Editor: Vecihi Batuman, MD, FASN  more...
  • Print

Practice Essentials

Lithium is currently a drug of choice for treating persons with bipolar depression and is widely used in this population. Approximately 0.1% of the US population is undergoing lithium treatment for psychiatric problems. Approximately 30% of patients taking lithium experience at least one episode of lithium toxicity.

Lithium toxicity can be acute, acute-on-chronic, or chronic. [1, 2] The acute lithium nephrotoxicity picture is dominated by evidence of volume depletion, obtundation, and the potential for cardiovascular collapse. The chronic lithium nephrotoxicity picture is dominated by polyuria and evidence of chronic kidney disease. See Presentation.

Treatment of acute toxicity involves correction of electrolyte abnormalities, volume repletion followed by forced diuresis, and dialysis in severe cases. In patients with chronic toxicity, polyuria can be treated with medication and the chronic renal insufficiency can be treated with the measures routinely used for chronic kidney disease. See Treatment.



Historical background

The medicinal use of lithium has a long and illustrious history. Galen recommended bathing in alkaline mineral waters, which might have contained lithium, for the treatment of mania as early as 200 AD.

In the mid-1800s, lithium was proposed as a treatment of uric acid calculi and gout, as uric acid crystals are highly soluble in solutions containing lithium carbonate. This therapy proved ineffective, but lithium was noted to be a highly effective treatment of psychiatric disorders in the late nineteenth century. Unfortunately, the toxicity of lithium severely limited its widespread acceptance at that point. Lithium was used as a substitute and added to the soft drink 7 Up in the early twentieth century; toxicity again leading to its withdrawal.

However, in 1949, the Australian psychiatrist John Cade reported on the successful use of lithium for mania. Since then, multiple studies have been performed demonstrating the efficacy of lithium in patients with mood disorders, such as depression, manic depression, and melancholia. Simultaneously, renal effects associated with lithium administration, including polyuria and nocturia, were increasingly reported.

In the 1950s and for several decades following, intensive studies on lithium nephrotoxicity were spurred by the wide acceptance of lithium administration in psychiatric practice as an effective treatment of and prophylaxis for unipolar and bipolar affective disorders.Those studies documented slowly progressive nephrotoxicity, rarely progressing to end-stage renal disease, in patients receiving long-term lithium.13033 For the past 2 decades, alternative psychiatric agents have been adopted for the treatment of these disorders, in large part because of the growing recognition of lithium nephrotoxicity.



Lithium is a univalent cation of the white metal series, closely related to both sodium and potassium, but having no known role in human physiology. Lithium is completely absorbed by the GI tract. The drug is not protein bound and is completely filtered at the glomerulus. The majority of the filtered load is reabsorbed by the proximal tubule, but significant amounts are also absorbed in the loop of Henle and the early distal nephron. Up to 90% of the filtered load is reabsorbed by the nephron, 60% in the proximal tubule, and the remainder in the thick ascending limb of the loop of Henle, the connecting tubule, and the cortical collecting duct. Lithium can substitute for sodium in several sodium channels, particularly the sodium-hydrogen exchanger in the proximal tubule (NHE3), the sodium/potassium/2chloride exchanger in the thick ascending limb of the loop of Henle (NKCC2), and the epithelial channel of the cortical collecting tubule (ENaC).

Lithium can affect kidney function in several ways. Acutely and chronically, lithium salts produce a natriuresis that is associated with an impaired regulation of the expression of the epithelial sodium channel in the cortical collecting tubule. [3, 4] Specifically, lithium use partially inhibits the ability of aldosterone to increase apical membrane ENaC expression, resulting in inappropriate sodium losses. [5]

The most common complication of long-term lithium therapy is nephrogenic diabetes insipidus. [6, 7, 8] At the cellular level, antidiuretic hormone (ADH) is released from the posterior pituitary in response to increases in serum osmolarity or decreases in effective circulating volume, and this hormone acts on V2 receptors in the basolateral membrane of the principal cells in the cortical and medullary collecting tubules. The net result of the cascade involving a G protein (guanyl-nucleotide regulatory protein) and adenylate cyclase is an increase in the intracellular cyclic adenosine monophosphate (cAMP) level, which can play a dual role in antidiuresis regulation. cAMP acutely stimulates protein kinase A, which facilitates the insertion of aquaporin-2 (AQP2) water channels. These water channels are preformed and stored in cytoplasmic vesicles in the apical plasma membrane of the principal cells. This process leads to increased water permeability and, thus,antidiuresis.

Over extended periods of time, increased cAMP levels also increase the production of AQP2 water channels at the genetic level by promoting a 5' untranslated region of the AQP2 gene. [9] Lithium impairs the ADH stimulatory effect on adenylate cyclase, thereby decreasing cAMP levels. [10] Li and colleagues have also performed studies suggesting that the ability of lithium to produce nephrogenic diabetes insipidus may be independent of its effect on cAMP generation and related to decreased AQP2 mRNA levels. [11] Thus, lithium most likely impairs water permeability in the principal cells by inhibiting water channel delivery and, over a prolonged period of time, by suppressing channel production. [3, 12, 13]

A minority of reports, however, propose that lithium-induced partial central diabetes insipidus may play a role in the polyuria that may develop in patients who show a modest response to exogenous ADH. Other studies show that ADH levels in patients treated with lithium are normal or elevated.

Over 30 case reports of lithium-induced nephrogenic diabetes insipidus appear in the medical literature. [6] Patients with urine-concentrating defects resulting from lithium treatment usually take weeks to months to recover following discontinuation of the drug; in rare situations, the problem can persist for years. Early reports in psychiatric patients suggested that this persistent concentrating impairment may be linked to underlying renal histological damage and may be worse with neuroleptic use and prolonged lithium therapy. In a 1987 review, Boton and colleagues showed a 54% correlation between impaired urine-concentrating ability and the duration and total dosage of lithium treatment. [14]

Lithium may also be responsible for a distal tubular acidification defect. [15] The defect is believed to be a variant of incomplete distal renal tubular acidosis, whereby the effect is exerted from the luminal side, requiring lithium cell entry. Patients taking lithium have normal phosphate and ammonia excretion. Lithium is not known to cause significant hyperkalemia.

The role of lithium in the production of acute renal failure is well accepted. The cause is generally due to severe dehydration and volume depletion due to the combination of natriuresis and water diuresis accompanied by elevated lithium levels, altered mental status, and subsequent poor oral intake. Acute renal failure has also been described as a result of lithium-induced neuroleptic malignant syndrome. [16] However, controversy still exists over its role in chronic renal failure. Boton and colleagues estimated (from an analysis of more than 1000 patients) that 85% of patients on long-term lithium therapy had normal glomerular filtration rates (GFRs); the remaining 15% had GFRs of more than 2 standard deviations below the age-corrected normal values, but very few patients had values less than 60 mL/min. [14]

Extensive reviews in 1988 and 1989 suggested that monitored long-term lithium treatment does not adversely affect the GFR, despite other reports of concurrent histological damage. Prospective studies of patients taking stable lithium also failed to show a decline in GFR in the absence of acute lithium intoxication. Although a minimal increase in the protein excretion rate has been reported in some patients who were taking lithium for at least 2 years, overt proteinuria is not a common complication. A rare association between minimal-change nephrotic syndrome and lithium administration has also been described.

Lithium does not appear to adversely affect proximal tubular function.



Lithium toxicity tends to occur in the context of suicide attempts or an intervening illness in an otherwise stable patient leading to poor intake, volume depletion, and subsequent increase in lithium levels. Alternatively, if the patient does not have an elevated lithium level, then the practitioner should look for other causes of diabetes insipidus.

For central diabetes insipidus, other etologic possibilities are as follows:

  • Familial disorder
  • Trauma-induced
  • Postsurgical
  • Neoplastic
  • Ischemic
  • Infectious
  • Autoimmune
  • Granulomatous
  • Idiopathic

For nephrogenic diabetes insipidus, etiologic possibilities include other renal causes, systemic disorders, drugs, dietary factors, and pregnancy.

Other renal causes include the following:

  • Chronic kidney failure
  • The diuretic phase of acute kidney injury
  • Obstructive uropathy

Systemic disorders include the following:

  • Familial X-linked syndrome
  • Electrolyte disturbances (hypokalemia, hypercalcemia)
  • Sickle cell anemia (trait or disease)
  • Sjögren syndrome
  • Amyloidosis
  • Sarcoidosis

Drugs include the following:

  • Demeclocycline
  • Loop diuretics
  • Angiographic dyes
  • Anticancer agents

Dietary factors include polydipsia, a low-protein diet, or a low-sodium diet.



Lithium is currently a drug of choice for treating persons with bipolar depression and is widely used in this population. Approximately 0.1% of the US population is undergoing lithium treatment for psychiatric problems. Approximately 20-54% of these patients have symptoms of urine-concentrating defects during and after lithium use, and up to 12% develop frank diabetes insipidus. [17] Some patients continue to have this problem for years after discontinuing lithium. One case report describes patients who still had diabetes insipidus 8 years later. In another report of a small subset of patients, up to 63% had persistent defects 1 year after stopping lithium. [7, 18]

Of note, approximately 30% of patients taking lithium experience at least one episode of lithium toxicity, correlating with a decrease in glomerular filtration rate. Researchers continue to debate the incidence and pathophysiology of long-term lithium nephropathy.

A study from the Netherlands, the incidence of chronic kidney disease (CKD) in patients treated with lithium was 0.012 cases per exposed patient-year. Longer duration of lithium exposure increased the odds of reaching stage 3 CKD. In patients who developed stage 3 CKD, the average decline in estimated glomerular filtration rate (eGFR) was 1.8 mL/min/year. The incidence of stage 4 CKD was only 0.0004 per patient-year. [19]  



Because of the frequent use and high incidence of associated urine-concentrating defects, lithium has been cited as the most common cause of nephrogenic diabetes insipidus. This complication is a major source of electrolyte disturbances and associated morbidity. The very narrow therapeutic window for this drug contributes substantially to the frequency of acute and chronic toxicity.

Race-,Sex-, and Age-related Demographics

The available literature does not suggest a racial predominance in lithium nephrotoxicity. A United Kingdom study of patients taking lithium reported greater risk of development of renal disorders in women than in men, with women younger than 60 years at higher risk than older women. The adverse effects occurred early in treatment. Patients with lithium concentrations higher than median were also at greater risk. [20]

In an Italian study, cross-sectional evaluation showed that estimated glomerular filtration rate (eGFR) was lower in women (by 3.47 mL/min/1.73 m²), in older patients (0.73 mL/min/1.73 m² per year of age), and in patients with longer duration of lithium treatment (0.73 mL/min/1.73 m² per year). These authors found that renal dysfunction tends to appear after decades of lithium treatment and to progress slowly and irrespective of whether lithium is continued at previous doses, continued at reduced doses, or discontinued. [21]

A population-based cohort study that assessed the effect of lithium maintenance therapy on eFGR in patients with affective disorders found no significant difference in the decline in eGFR in 305 patients taking lithium compared with 815 patients taking other first-line drugs (quetiapine, olanzapine, and semisodium valproate). The researchers concluded that their results contradict the idea that long-term lithium therapy is associated with nephrotoxicity in the absence of episodes of acute intoxication. [22]   

In a randomized, double-blind, placebo-controlled study in older adults who were receiving low doses of lithium for treatment of mild cognitive impairment (serum lithium range of 0.25 to 0.5 mEq/L), 4 years of treatment did not have any adverse effects on renal function. However, lithium treatment was associated with significant increases in the number of neutrophils, serum thyroid-stimulating hormone level, body weight, and more adverse events overall than placebo, and patients treated with lithium had higher rates of diabetes mellitus and arrhythmia. [23]

In a nested case-control study of Canadian mental health service users aged 66 years and older, lithium use ws independently associated with an almost 2-fold increase in risk of chronic kidney disease (CKD). The adjusted odds ratio (OR) in lithium users (n=529) was 1.76 (95% confidence index [CI], 1.41-2.1), In comparison, valproate users (n=498) were not at increased risk (adjusted OR = 1.03; 95% CI, 0.81-1.29). [24]