Lithium-induced nephrotoxicity

I recently saw a patient in my outpatient clinic who was referred for management of chronic kidney disease presumed to be caused by chronic lithium-induced nephrotoxicity. The patient presented with a serum creatinine ranging from 1.4-1.5 mg/dL over the past year (baseline was 1.0 mg/dL ten years ago), minimal proteinuria, a history of polyuria and diabetes insipidus, and a non-anion gap metabolic acidosis (serum bicarbonate 15-18 mEq/L). She had been taking lithium as prescribed by her psychiatrist for 20+ years and was now questioning whether she should continue the medication given her renal insufficiency. I thought I'd use this opportunity to briefly discuss the principal forms of nephrotoxicity that are observed with chronic lithium administration. For those fellows who are beginning to think about the boards, this is a commonly covered topic.

Lithium is freely filtered by the glomerulus and up to 80% of the filtered load is reabsorbed. While the proximal tubule reabsorbs most of the filtered lithium, a small portion is reabsorbed in the distal nephron by the epithelial sodium channel (ENaC).

1. Nephrogenic diabetes insipidus (NDI). This is the most common adverse effect of lithium, occurring in up to 40% of patients. Chronic lithium ingestion can lead to ADH resistance and cause polyuria and polydipsia. Lithium enters the principal cells of the collecting duct via ENaC and accumulates intracellularly, interfering with the ability of ADH to increase water permeability. While the mechanisms are incompletely understood, they may involve downregulation of and decreased trafficking of aquaporin 2 channels from the cytosol to the luminal membrane. Cessation of lithium therapy can reverse this defect; however, NDI can persist permanently even after withdrawal of the medication. The use of amiloride has been shown to help improve the NDI (Bedford et al. Clin J Am Soc Nephrol 2008), as it competes with lithium for entry into the collecting duct cells via ENaC and prevents intracellular lithium accumulation.

2. Metabolic acidosis. Patients may also develop a hyperchloremic non-anion gap metabolic acidosis, thought to be secondary to impaired proton secretion in the collecting duct or excessive back-diffusion of acid equivalents.

3. Chronic tubulointerstitial nephropathy. This is the most common form of chronic kidney disease associated with chronic lithium administration. Renal biopsies in patients with lithium-induced CKD classically demonstrate interstitial fibrosis and thickened tubular basement membranes, often associated with focal segmental glomerulosclerosis, and renal cysts originating primarily from the distal tubule. Duration of lithium treatment and cumulative dose seem to be the main risk factors for the development of nephrotoxicity. CKD induced by lithium typically occurs in patients who have been treated with lithium for >10-20 years. In one study, 21% of patients treated with lithium for a mean of 16.8 years had a serum creatinine of 133 umol/L or higher on two consecutive occasions (Lepkifker et al. J Clin Psychiatry 2004). Renal impairment is usually mild, and the development and progression of renal insufficiency is generally insidious, although there have been a reported cases of patients progressing to ESRD (15-30% progress to ESRD within an average of 20 years in one series). Patients are normotensive, proteinuria is absent or minimal, and urine sediments are typically bland.

Interestingly, in spite of the established association between lithium and renal impairment, monitoring of renal function in patients taking chronic lithium has been surprisingly insufficient in some countries. Currently, the American Psychiatric Association recommends that serum creatinine be measured every 2-3 months for the first 6 months of lithium treatment and then every year afterwards.

4. Hypercalcemia. While lithium-induced hypercalcemia was first documented in 1973, its association with hyperparathyroidism was not noted until 1978. Up to 25-30% of patients treated with lithium can develop hypercalcemia, and the prevalence of hyperparathyroidism may be as high as 10-15% of patients treated for >10 years. Many of these patients have parathyroid hyperplasia, and discontinuation of lithium does not always lead to rapid correction of serum calcium levels.

Since lithium is such an effective medication at controlling symptoms of bipolar disorder, many patients are reluctant to switch to another medication (e.g. valproate, lamotrigine, and olanzapine) which may not have as predictable of a therapeutic effect. The decision to withdraw lithium in a patient with nephrotoxicity is often a difficult one and should be jointly made by the patient, the patient's psychiatrist, and the patient's nephrologist.