Lithium is one of the most widely used mood stabilizers and is considered a first-line treatment for bipolar disorder because of its proven antimanic and prophylactic effects.1 This medication also can reduce the risk of suicide in patients with bipolar disorder.2 However, it has a narrow therapeutic index. While lithium has many reversible adverse effects—such as tremors, gastrointestinal disturbance, and thyroid dysfunction—its perceived irreversible nephrotoxic effects makes some clinicians hesitant to prescribe it.3,4 In this article, we describe the relationship between lithium and nephrotoxicity, explain the apparent contradiction in published research regarding this topic, and offer suggestions for how to determine whether you should continue treatment with lithium for a patient who develops renal changes.
A lithium dilemma
Many psychiatrists have faced the dilemma of whether to discontinue lithium upon the appearance of glomerular renal changes and risk exposing patients to relapse or suicide, or to continue prescribing lithium and risk development of end stage renal disease (ESRD). Discontinuing lithium is not associated with the reversal of renal changes and kidney recovery,5 and exposes patients to psychiatric risks, such as mood recurrence and increased risk of suicide.6 Switching from lithium to another mood stabilizer is associated with a host of adverse effects, including diabetes mellitus and weight gain, and mood stabilizer use is not associated with reduced renal risk in patients with bipolar disorder.5 For example, Markowitz et al6 evaluated 24 patients with renal insufficiency after an average of 13.6 years of chronic lithium treatment. Despite stopping lithium, 8 patients out of the 19 available for follow-up (42%) developed ESRD.6 This study also found that serum creatinine levels >2.5 mg/dL are a predictor of progression to ESRD.6
Discontinuing lithium is associated with high rates of mood recurrence (60% to 70%), especially for patients who had been stable on lithium for years.7,8 If lithium is tapered slowly, the risk of mood recurrence may drop to approximately 42% over the subsequent 18 months, but this is nearly 3-fold greater than the risk of mood recurrence in patients with good response to valproate who are switched to another mood stabilizer (16.7%, c2 = 4.3, P = .048),9 which suggests that stopping lithium is particularly problematic. Considering the lifetime consequences of bipolar illness, for most patients who have been receiving lithium for a long time, the recommendation is to continue lithium.10,11
The reasons for conflicting evidence
Many studies indicate that there is either no statistically significant association or a very low association between lithium and developing ESRD,12-16 while others suggest that long-term lithium treatment increases the risk of chronic nephropathy to a clinically relevant degree (note that these arguments are not mutually exclusive).6,17-22 Much of this confusion has to do with not making a distinction between renal tubular dysfunction, which occurs early and in approximately one-half of patients treated with lithium,23 and glomerular dysfunction, which occurs late and is associated with reductions in glomerular filtration and ESRD.24 Adding to the confusion is that even without lithium, the rate of renal disease in patients with mood disorders is 2- to 3-fold higher than that of the general population.25 Lithium treatment is associated with a rate that is higher still,25-27 but this effect is erroneously exaggerated in studies that examined patients treated with lithium without comparison to a mood-disorder control group.
Renal tubular dysfunction presents as diabetes insipidus with polyuria and polydipsia, which is related to a reduced ability to concentrate the urine.28 Reduced glomerular filtration rate (GFR) as a consequence of lithium treatment occurs in 15% of patients23 and represents approximately 0.22% of patients on dialysis.18 Lithium-related reduction in GFR is a slowly progressive process that typically requires >20 years of lithium use to result in ESRD.18 While some decline in GFR may be seen within 1 year after starting lithium, the average age of patients who develop ESRD is 65 years.6 Interestingly, short-term animal studies have suggested that lithium may have antiproteinuric, protective, and pro-reparative effects in acute kidney injury.29
Anatomical anomalies in lithium-related glomerular dysfunction
In a study conducted before improved imaging technology was developed, Markowitz et al6 used renal biopsy to evaluate lithium-related nephropathy in 24 patients.6 Findings revealed chronic tubulointerstitial nephritis in all patients, along with a wide range of abnormalities, including tubular atrophy and interstitial fibrosis interspersed with microcyst formation arising from distal tubules or collecting ducts.6 Focal segmental glomerulosclerosis (FSGS) was found in 50% of patients. This might have been a result of nephron loss and compensatory hypertrophy of surviving nephrons, which suggests that FSGS is possibly a post-adaptive effect (rather than a direct damaging effect) of lithium on the glomerulus. The most noticeable finding was the appearance of microcysts in 62.5% of patients.6 It is important to note that these biopsy techniques sampled a relatively small fraction of the kidney volume, and that microcysts might have been more prevalent.
Recently, noninvasive imaging techniques have been used to detect microcysts in patients developing lithium-related nephropathy. While ultrasound and computed tomography (CT) can detect renal microcysts, magnetic resonance imaging (MRI), specifically the half-Fourier acquisition single-shot turbo spin-echo T2-weighted and gadolinium-enhanced (FISP three-dimensional MR angiographic) sequence, is the best noninvasive technology to demonstrate the presence of renal microcysts of a diameter of 1 to 2 mm.30 Ultrasound is sometimes difficult to utilize because while classic cysts appear as anechoic, lithium-induced microcysts may have the appearance of small echogenic foci.31,32 When evaluated by CT, renal microcysts may appear as hypodense lesions.
Continue to: Recent small studies...