Original Author: Paul Perry, Ph.D, BCPP
Latest Revisers: Paul Perry, Ph.D, BCPP, Brian C. Lund,
Pharm.D.
Creation Date: 1996
Peer Review Status: Internally Peer Reviewed
In an early animal study, Radomski et al (1950) found that lithium chloride in daily doses of 20 mg/kg/day administered in the presence of a low-sodium diet could produce renal impairment manifested by loss of total body sodium, azotemia, renal tubular damage, and reduced lithium excretion. In retrospect, these findings are not startling because it is now recognized that patients on lithium with a negative sodium balance because of diet, thiazide diuretics, illnesses associated with nausea and vomiting etc. will become intoxicated on lithium and once intoxicated are at risk of experiencing renal dysfunction secondary to lithium. The majority of cases of lithium's adverse renal effects are associated with lithium intoxication.
ANECDOTAL DATA
The earliest reported cases of lithium nephrotoxicity were associated with lithium poisonings. Isolated case reports of uremia or acute renal failure began to appear in the literature in 1965 in patients with lithium intoxication (Achong et al 1975, Amkisen and Skoldborg 1969, Dias and Hocken 1973, Hawkins and Dorken 1969, Lavender et al 1973, Verbov et al 1965). In one of the fatal cases (Achong et al 1975), a renal biopsy was performed because of acute renal failure. The biopsy demonstrated damage mainly in the proximal tubules with less marked changes in the glomeruli and the interlobular arteries. Yet in another case (Lavender et al 1973), renal functional changes such as proteinuria, elevated BUN and hypokalemia occurred, but no significant pathological changes of the kidney were observed at autopsy. Schou et al in 1968 described moderate elevations of the BUN and/or serum creatinine in seven of eight poisoned patients. The highest observed BUN and serum creatinine were 80 mg% and 3.5 mg%, respectively. Clinical signs of uremia were not observed. Transient proteinuria occurred in three patients and reverted to normal upon reversal of the lithium intoxication. The authors felt that in humans, the central nervous system rather than the kidneys were the primary target organ affected by lithium intoxication. Thus, both functional and morphological changes of the kidneys have been reported in lithium-poisoned patients. Whether the renal impairments are due to a direct effect of the lithium on the kidneys or secondary to systemic toxicicity is not resolved. Although a direct toxic effect of lithium on the kidneys has been shown in animals, the renal failure could also result from hypotension and renal ischemia resulting from patients going into shock during lithium intoxications (Dias and Hocken 1973, Hawkins and Dorken 1969, Verbov et al 1965).
The early reports suggest that lithium intoxication was a prerequisite for the occurrence of renal damage. Hestbech et al (1977) studied 14 chronic lithium users (duration of lithium therapy ranged between 1 2/3 to 14 years). These patients were referred to the nephrologist because of either acute lithium intoxication or lithium-induced diabetes insipidus. Renal biopsy specimens showed that 13 patients had a pronounced degree of focal nephron atrophy and/or interstitial fibrosis. Thirteen age-matched patients not on lithium were used as controls. The authors found significant differences between the two groups in the degree of focal cortical fibrosis, diffuse medullary fibrosis, distal tubular dilatation, and the presence of focal mononuclear cells. Although the majority (8) of the lithium-treated patients had lithium intoxication at the time of study, it was postulated that the chronic use of lithium also may have contributed to the chronic renal lesions in the five patients who had never been intoxicated. However, of the three patients who experienced a persisting reduction of creatinine clearance, all had been intoxicated. Despite these histological findings significant impairment of the GFR was not observed. The authors subsequently suggested that up to 26% of patients on maintenance lithium for greater than two years would develop pathological renal histolgical changes (chronic focal interstitial nephropathy) characterized functionally by progressive impairment of urinary concentrating ability. Hansen and Amdisen (1978) reported renal insufficiency (CrCl < 45 ml/min) occurring in 17 of the 23 patients intoxicated with lithium. Five of these cases did not regain normal renal function. In seven patients, renal biopsy studies suggested the presence of a chronic nephropathy. Although chronic lithium use had been implicated in the occurrence of lithium nephrotoxicity, the majority of the cases were still associated with intoxications.
Burrows et al in 1978 postulated that lithium could be nephrotoxic in chronic users, but the severity of the risk might be determined by the duration of lithium use. In their five cases of lithium-related renal lesions, two patients who had been treated for only a few months had the lesions mainly in the distal convoluted tubules and the collecting ducts with only minimal changes in the proximal tubules. In contrast, the three patients who had been taking lithium for over six years were found to have similar lesions in the distal tubules as well as more advanced changes including glomerular and interstitial cell involvements. None of these five patients had episodes of lithium intoxication.
Thus by 1978, anecdotal data was available that suggested that lithium intoxication was not a prerequisite for the occurrence of lithium nephrotoxicity. Chronic use of lithium also contributed to renal damage although the renal lesions seen in chronic lithium therapy were not as severe as those observed in lithium intoxications nor as frequent. In the next several years a number of case-control studies were conducted to assess the risk of lithium-induced nephrotoxicity.
CASE CONTROL STUDIES
Although the relationship between morphological changes observed on biopsy and functional impairment remains unclear, chronic renal damage is seen in a minority of patients on prophylactic lithium therapy. Rafaelson et al (1979) examined biopsies, creatinine clearance rates, and urine volumes in 37 patients treated with lithium longer than 5 years. Significant correlations were observed between urine volume or serum creatinine level to the presence of histological anomaly. Six of 37 (16%) renal biopsies were clearly pathological. Five of six pathological renal biopsies and seven of eleven borderline pathological biopsies came from patients who had a 24-hour urine volume >4000 ml or 24-hour creatinine clearance below 60 ml/min. Interestingly, nine patients had an increased serum creatinine. Of this group, six had been taking lithium on a divided daily dose schedule while three had been taking the drug once daily. Of the entire series, 80% of the patients were on a once-a-day dose, while most of the remainder were on a divided daily dose. Thus only 10% of patients on single daily dosing had abnormal biopsies while 86% of patients on divided daily dosing had abnormal biopsies. They hypothesized that an increase in 24-hour urine volume to greater than 3,000 ml during the first year of therapy may be the most easily observed warning signal of impending kidney damage.
Hullin (1979) examined urinary beta-microglobulins due to their sensitivity as an index of renal proximal tubular damage. No significant differences were found between lithium-treated patients and the control groups.
Kincaid-Smith et al (1979) compared the renal biopsies of 16 never intoxicated patients on lithium therapy for an average of 5.5 years to those of nine patients about to start lithium and a group of age- matched donor kidneys used for transplantation. There were no differences between the lithium and pre-lithium biopsies. However, there were significant histological differences between the affective disorder patients' biopsies and the donor kidneys. The authors raised the question as to whether the other psychotropic drugs used to treat affective illness could be nephrotoxic, such as TCAs.
In 1979, the World Psychiatric Association convened a meeting in Copenhagen to discuss lithium and renal effects. The consensus at this meeting was that the risk of lithium nephropathy in the absence of lithium toxicity is small. However, the patient with pre-existing renal damage or a history of renal disease placed on lithium should have renal function studies performed at regular intervals while on the drug (Anon 1979). Subsequent studies have reinforced the conclusion and recommendations of the WPA.
Gerner et al (1980) assessed the renal function in 43 manic-depressives who had taken lithium from 1-120 months (x = 27 mo.). All patients were euthymic at the time of the evaluation and had therapeutic lithium serum levels i.e., 0.6-1.2 mEq/L. None had diabetes, hypertension, autoimmune diseases, or renal disease before starting lithium. There was no significant correlation between decrease in creatinine clearance and duration of lithium treatment, and when creatinine clearance was corrected for age, the expected decrease with age was seen. BUN and uric acid were only slightly elevated in four and two patients, respectively. Minor elevations in urine methyl-histadine and glycine were seen in two and five patients, respectively, but were not sustained upon repeat urine amino acid chromatography and were postulated to be due to diet. Urinary lysozymes remained normal. The patients appeared to have a defect in maximal renal concentrating ability; their mean osmolality was 602 mOsm/kg (range = 100-965). There were no significant correlations of urine osmolality with duration of lithium treatment.
Donker et al (1979) studied 30 patients on long-term lithium therapy of 39 ± 3.6 months. The patients were subjected to an extensive battery of renal function testing that included urine concentration tests after water deprivation and intranasal administration of vasopressin; determination of glomerular filtration rate (GFR) and effective renal plasma flow (ERPF); and measurement of minimal pH after an oral dose of ammonium chloride, and of urinary beta-2-microglobulin excretion. Urine osmolality during treatment and after 22h water deprivation was 845 mOsm/kg H2O (range 524-1078) as compared to that in normals of 1002 mOsm/kg H2O (range 712-1349). The mean GFR was 101 ± 4ml/min, mean ERPF 360 ± 18 ml/min, and mean minimal urine pH 4.95 ± 0.06. Polyuria (>3000ml/24h) was present in 27% (8/30) of the patients, but only two of the patients' urine osmolality was abnormal. Serum levels of beta-2-microglobulins and lysozymes and urinary excretion of beta-2-microglobulins were normal in all patients. No correlations were established between urine osmolality and serum lithium levels. The GFR only correlated with age.
Coppen et al (1980) examined 101 patients maintained on slow-release lithium tablets for 1 to 12.5 years given as a single bedtime dose for the past seven years. A control group of patients matched for age and sex with affective illness who had never been treated with lithium were utilized as a control group. Creatinine clearance, serum creatinine, maximum urine osmolality after DDAVP (1-deamino-8-D-arginine vasopressin) administration, urine and plasma beta-2-microglobulins and the urinary excretion of N-acetyl-beta-glucosaminidase were not significantly different between patients on lithium and depressed patients never on lithium. The authors concluded that there was little evidence of serious renal functional impairment with lithium therapy.
Lokkegaard et al (1985) evaluated renal function of 153 manic-depressive patients treated with lithium for more than five years with a mean of ten years. The urine concentrating ability decreased and the 24-hour urine output of the patients increased as the dose increased. Interestingly, the glomerular filtration rate showed a slight decrease with was not projected as significant until after patients had been on lithium for a period of 17 years. This finding was independent of the daily dosing schedule.
Johnson et al (1984) showed that the impaired GFR in lithium patients correlated with episodes of acute toxicity rather than duration of stable maintenance lithium therapy. Overall, GFR fell within the established normal range based on sex and age, whereas measures of urinary concentrating ability were generally impaired. Additionally, he points out that in the few controlled studies of renal function that are available, measurements of GFR were similar between patients on maintenance lithium therapy and other patients with affective psychiatric disorders not on maintenance lithium therapy.
Despite reassurances that lithium does not cause severe renal dysfunction, reported reductions in GFRs ranged from 3-30%. DePaulo et al (1981) studied 99 lithium-treated patients, i.e. 32 ± 35 months of therapy at a dose of 1198 ± 407 mg/day, using several techniques to measure GFR. Eighty-six patients were suitable for the final analysis. Tests of the GFR were serum creatinine, creatinine clearance estimated by the Cockroft-Gault method, standard creatinine clearance determination, i.e. 24h urine collection, and insulin clearance. The Cockroft creatinine clearance (Ccr) was estimated by the equation (Cockroft and Gault 1976):
Ccr = (140-age) (weight in kg) / (72) (serum creatinine)
Ten of 86 (12%) patients who had no history of renal disease had mildly low creatinine clearance values. The mean serum creatinine was 1.25 mg% while the mean duration of therapy was 50 months in these ten patients. Modest but significant correlations of duration of lithium treatment with serum creatinine and creatinine clearance values suggested the possibility of an adverse effect of lithium on glomerular function. It should be noted, however, that 22 of the patients in this study had a history of lithium toxicity which may explain these findings. Also, in a subsequent longitudinal study DePaulo et al (1986) were unable to demonstrate a correlation between GFR and number of years on lithium. The authors recommended that clinicians initially perform duplicate classical and Cockroft creatinine clearance tests using the Cockroft value to confirm that standard creatinine clearance value.
The Cockroft-Gault equation is a practical means of monitoring lithium patients' renal function than creatinine clearances with 24-hour urine collections or urine concentrating tests. The complexity of these tests requires the physician to explain the potential problem to the patients. The patients may become unnecessarily worried with the result that they discontinue the lithium. On the other hand, the physician fearing noncompliance might treat the problem with benign neglect and not monitor patients' renal function. The Cockroft- Gault equation allows for both patient and physician concerns to be circumvented until a problem really exists.
In order to calculate the expected creatinine clearance by the Cockroft method so as to compare it to the apparent creatinine clearance for an individual, it is necessary to know the mean serum creatinine values for Cockroft's different age cohorts. The mean serum creatinines are as follows: 18-29 yo, 1.0 mg%; 30-39 yo, 1.1mg%; 40-49 yo, 1.2mg%, 50-59 yo, 1.5mg%; 60-69 yo, 1.4mg% and 70-79 yo, 1.8mg%. Thus as an example a 24 yo, 84 kg, male with serum creatinine of 1.9mg% is calculated to have an apparent creatinine clearance of 71 ml/min although according to Cockroft's norms he is expected to have a creatinine clearance of 135 ml/min. (Duplicate 24-hour creatinine clearances on this patient were 71 and 72 ml/min.)
Three important points to remember about the use of the above equation to insure its accuracy are: 1) the calculated CrCl should be reduced by 15% in women, 2) a correction to lean or ideal body weight is necessary in excessively fat or edematous patients, and 3) the serum creatinine value should be drawn in the fasting state.
The continuing appearance of new long-term studies suggesting little or no effect of lithium on GFR (Waller and Edwards 1988, Christensen and Aggernaes 1990) have lead Schou (1988) to state that "on the basis of the total evidence available today, one must conclude that lithium treatment does not, even when given for many years, lead to any change of GFR, let alone to renal failure." As a result, routine measurements of creatinine clearance and serum creatinine were thought hardly "worth the expenditure of time, effort, and money." It was stated that it is far more important to educate patients about signs of toxicity and risk situations such as physical illness with fever, rigorous slimming, and concurrent administration of diruetics or NSAIDS.
Most investigators agree the risk of renal dysfunction is far less than the risk of psychiatric morbidity. However, the present data suggests that it is imperative to avoid lithium intoxications. Thus routine measurement of urine volume (semi-annual) and serum creatinine (annual) and the education of patients to the signs and symptoms of lithium well are recommended as the current standard of care.
LITHIUM INDUCED POLYURIA - MECHANISM
Polyuria (urine output > 3 L/d) is currently recognized as the most bothersome effect of lithium on the kidneys. It occurs in 20-30% of patients on lithium, but it doesn't progress with time. It is also not immediately reversible upon lithium discontinuation and often persists for more than a year (Boton et al 1987) Lithium produces polyuria and polydipsia by disturbing ADH function through a number of inter-related mechanisms. The primary effect of lithium on the kidney is to block ADH-mediated water reabsorption by inhibiting ADH-sensitive adenylate cyclase production of cAMP in the collecting duct cells. Cyclic AMP ordinarily increases collecting duct permeability to allow reabsorption of free water (Christiansen et al 1985, Forrest et al 1974). As a result of the hypo-osmolality the pituitary will increase ADH secretion. The increased ADH level is known to elevate urinary prostaglandins which, similar to lithium, also inhibit ADH-sensitive adenylate cyclase (Grindlinger and Boylan 1987). It has also been shown that lithium increases thirst response to an osmotic load in non-polyuric lithium treated patients compared to patients not receiving lithium (Penney and Hampton 1990). This suggests that while much of the polydipsia seen is probably due to a secondary effect of the direct effect of lithium on the kidney, a primary component of increased thirst may also contribute.
LITHIUM INDUCED POLYURIA - MANAGEMENT
Since 1977, several factors have been identified which probably affect the risk of adverse renal effects with lithium. These include the number of lithium doses ingested per day and the type of lithium product formulation being used.
Lithium is still being routinely administered in divided daily doses although the half-life is sufficiently long to permit once-a-day dosing. Perry et al (1981) assessed the feasibility of once-a-day in eight patients diagnosed as manic-depressives and under treatment with lithium by administering their entire 24 hour maintenance dose (1500 mg/d or less) at 8 PM for 12 consecutive days. Lithium half-life, steady-state serum lithium concentration, lithium clearance, serum creatinine and CrCl were unchanged. However, the mean 24 hour urine output was decreased from 2104 to 1566 ml/24h. Rafaelson et al (1979) as previously noted has found that urine outputs of > 3000 more likely to be associated with abnormal renal biopsies. It may be that significant polyuria is a harbinger of renal morphology. Alternatively, it may be that the polyuria results from changes in renal morphology.
The type of product formulation may influence renal effects as well. Because of its rapid absorption, a single oral dose of 600 mg of regular release lithium carbonate results in a peak serum level within usually 1 1/2 hours after drug administration. The US company, Rowell Laboratories have demonstrated, utilizing computer-simulated curves for a 600 mg po q 24h dose of lithium carbonate, that the peak serum lithium levels are decreased significantly when utilizing slow-release lithium carbonate (Greenwell 1980). Eleven patients received single 600 mg doses of the slow-release and regular-release lithium carbonation formulations in a cross-over design. The pharmacokinetic parameters were determined and then the computer-simulated curves were derived. The peak serum lithium level, occurring 1-2 h after ingestion of the regular release formulation, at steady state was 0.76 ± 0.14 meq/l. The peak serum lithium level, occuring usually 4-6 h after ingestion of the slow-release formulation was 0.42 ± 0.09 meq/l, a 45% decrease.
In a report using a British slow-release product, Wallin and Alling (1979) studied 56 patients receiving long-term lithium therapy (mean duration of treatment 63.5 months). Twenty-eight patients were treated with slow-release lithium while 28 received a regular-release formulation. Dosage administration schedules were not mentioned. Patients were matched for age, sex, and lithium intake. Urine osmolality was found to be significantly higher (846 mOsm/kg H2O versus 747 mOsm/kg H2O) in patients receiving the slow-release formulation. Unfortunately, this study failied to control for lithium concentration, duration of lithium treatment, or dosage schedule, three potentially important variables affecting urine concentrating ability, it is impossible to assess the efficacy of sustained release lithium in the prevention of polyuria. In addition, most published data suggests that once daily dosing and a low minimum lithium concentration resulting from once daily dosing are more important factors in the prevention of polyuria. Since sustained release dosage forms produce less peak/trough fluctuation than immediate-release dosage forms, it may be that sustained-release products are more likely to produce polyuria.
Of the two risk factors described above it appears that a number of daily doses may be more important than the product formulation based on recent study by Schou and coworkers (1982). He contrasted the effects on the kidney of lithium slow-release tablets being given twice daily to regular-release tablets being given once daily. For years, two Danish hospitals have used these two different treatment regimens. The regimens did not affect the glomerular filtration or proximal tubular reabsorption but distal tubular water reabsorption and polyuria were significantly less affected in patients given regular-release tablets once daily (median urine output = 2.38 1/2 4h) than in patients given slow-release tablets twice daily (median output = 2.83 1/2 4h). It is impossible to differentiate between the effect of slow-release tablets and dosage schedule in this study--either factor, or both, may have contributed to the results of the study.
An animal study by Plenge et al (1981) has explored the issue of dosage schedule more carefully. In this study, lithium was administered in equal doses to two groups of rats. One received lithium by daily intraperitoneal injections and thus had high peak and low trough serum lithium concentrations. The second group was given lithium with their food and had almost constant serum lithium concentrations. Noteworthy, is the fact that the first group of rats receiving bolus doses of lithium developed significantly less polyuria than the second group. Plenge et al (1982) studied renal structure and function in two groups of patients receiving long-term lithium. The drug was administered once daily between 8 and 10 p.m. or in a bid or tid dosage schedule. Kidney biopsy was performed and 24-h urine volume was measured in each patient. The function and structural changes were more pronounced in patients given their lithium in divided doses during the day. The mean daily urine output in the patients receiving the drug on the divided daily dosage schedule was 4.63 l/24 h versus 2.25 l/24 h in the patients on once daily dosing (p < 0.001). Additionally, the renal biopsies showed significant greater sclerotic glomeruli (14.5 vs 5.4%), fibrosis (9.5 vs 6.2%), and atropic tubules (14.4 vs 4.1%) in divided daily dosing group.
Mellerup et al (1988) have confirmed these findings. Serum lithium parameters were analyzed over a 24 hour period in patients receiving lithium once daily. Multiple linear regression analysis was performed using urine output as the independent variable and lithium dose, maximum serum lithium concentration, 12-hour serum lithium, minimum serum lithium concentration, length of treatment, and age of patients as the dependent variables. Only minimum serum lithium concentration and urine volume showed significant positive correlation. In a similar study, Hetmar et al (1986) found a highly significant positive correlation between urine volume and number of lithium doses per day. Although the lithium dose was 1.23 times higher in the multiple daily dose group, 12 hour lithium concentrations were similar between groups.
Two studies have been published that contradict the previously cited studies' suggestion that once daily lithium dosing prevents polyuria (Abraham et al, 1992; O'Donovan et al, 1993). In the Abraham et al study (1992), 20 chronic lithium-treated patients with an average duration of lithium treatment of 4.4 years were studied. Ten patients were on once daily dosing, the other 10 received lithium 2-4 times daily. After one month the subjects crossed over to the alternate schedule. While on once daily lithium, the mean urine volume and mean lithium concentrations were 2.4 L/dayand 0.69 mEq/L, respectively. While on multiple daily doses of lithium, the mean urine volume and mean lithium concentrations were 2.2 L/day and 0.64 mEq/L, respectively. Similarly, among 24 patients on multiple daily doses of lithium for at least one month, O'Donovan et al found no significant effect of dosing interval on 24 hour urine volume at 3 month's follow-up (1993). Lithium concentrations ranged from 0.62-0.73 mEq/L in this study. Although these studies have failed to confirm an effect of dosing interval on 24 hour urine volume, the conservative lithium concentrations they maintained may have prevented a difference from becoming apparent. This emphasizes the fact that at low concentrations, lithium is generally well tolerated.
If the use of single daily dosing lowering of the serum lithium level does not adequately control the lithium-induced polyuria the the use of thiazide diuretics or the potassium-sparing diurtetic amiloride is indicated.
Levy et al (1973) studied a 54 year-old woman with lithium-induced polyuria of 5500 ml/24 h. Following the addition of chlorothiazide 500 mg/d to her therapy, her previously stable serum lithium level increased from 1.2 to 2.5 mEq/l at a lithium carbonate maintenance dose of 1800 mg/d. The patient experienced a lithium neurotoxicity reaction manifested by dysarthria, ataxia, lethargy and tremor. When her maintenance dose was reduced to 1200 mg/d and the diuretic continued, the serum lithium concentration declined to 1.2 mEq/l. It is important to note that her urine output decreased to 2 liters/24 h on the diuretic. However, when the diuretic therapy was discontinued for 11 days, a > 5 liter/24 h polyuria recurred. Juhl et al (1976) described a 29 year old male with > 9 liter/24 h polyuria. Following the addition of hydrochlorothiazide 50 mg/d to his therapy, the polyuria decreased to < 3 liters/24 h. However, the patient's daily lithium carbonate dose of 2700 mg/d which produced a steady-state concentration of 0.85 mEq/l had to be reduced to 2100 mg/d to achieve a steady-state concentration of 1 mEq/l. The mechanism of action of the thiazide diuretics in reversing the symptoms of polyuria is hypothesized to be the result of the thiazides decreasing sodium reabsorption in the distal tubules thereby resulting in a compensatory increase in sodium and water reabsorption in the proximal tubules. Therefore a less free water reaches the distal tubules. Bearing in mind the interaction of lithium and the thiazides, one should use caution when using the two drugs together. Serum lithium levels should be closely monitored and dosage adjustments probably will have to be made to avoid lithium intoxication. Serum potassium should be closely monitored for hypokalemia during the first 4 weeks of combined treatment.
The potassium-sparing diuretic, amiloride, may be a more preferable diuretic to treat lithium induced polyuria. Batlle et al (1985) studied the effect of amiloride on lithium-induced polyuria, i.e., 3.1-8.2 liters/24 h, in nine patients with affective illness being treated with prophylactic lithium therapy. Amiloride was selected because it is known to significantly improve the defect in free-water reabsorption in lithium-induced polyuria in animal models. All patients were receiving 20 mg/d within three weeks of the start of the study, except in the case of one subject who achieved this dose after three months. After an average course of 24 days of amiloride administration, the urine volume declined significantly from 4.7 to 3.1 liters/24 h. Additionally, the urine osmolality increased significantly from 228 to 331 mOsm/kg of H2O. The decrease in urine output was sustained for six months. After the addition of amiloride 10 mg/d, serum lithium levels measured 7-14 days later increased insignficantly from a mean of 1.1 mEq/l to 1.2 MEq/l. In one patient, however, an asymptomatic rise in the serum lithium level from 0.8 to 2.0 mEq/l was observed, at which time the daily lithium dose was decreased by 50%. The authors hypothesized that amiloride might have caused a volume contraction which decreased the glomerular filtration rate, thereby increasing lithium levels. Another patient required the lithium dose to be halved because the serum lithium level increased to 1.8 mEq/l by the end of the second month of amiloride. Serum potassium remained in the normal range throughout the study period. Kosten and Forrest (1986) replicated the finding that amiloride is effective in treating lithium-induced nephrogenic diabetes insipidus. Eight patients being treated with hydrochlorothiazide 50-150 mg/d for polyuria were switched to amiloride 10-20 mg/d. Within 4-6 days of amiloride the polyuria decreased significantly from 5906 ml/24 h to 5.1 l/24 h. Two of the eight patients who were continued on the drug for an additional 2-3 months had their polyuria decrease from 4.5 to 1.8 l/24 h. However, the data suggest that the chances are only 1 in 3 that a polyuria decrease of greater than 1 l/24 h will be observed. The mechanism whereby amiloride reduced urine volume in these patients is suggested to be prevention of lithium uptake into the collecting duct cells, thereby preventing lithium's inhibitory action on adenyl cyclase. Though amiloride is less likely to increase lithium levels than thiazide diuretics, the potential exists. Therefore, serum lithium and potassium levels (for hyperkalemia) should be monitored frequently during the first 4 weeks of combined treatment or after amiloride dose increases.
In an unpublished study coordinated by Bruce Alexander, Pharm.D. and Brian Cook, D.O. at the Univerity of Iowa, placebo, hydrochlorothiazide (HCTZ), and amiloride were compared in the treatment of lithium-induced polyuria. In general, both HCTZ and amiloride porduced a one liter reduction in 24 hour urine volume compared to placebo. Response was variable, with some patients responding better to one treatment over another. Although HCTZ was more likely to produce significant changes in lithium concentration, this effect did not occur in all patients on HCTZ. For amiloride, significant elevations in lithium concentration also occurred, although less frequently. Both drugs affected potassium concentration. When patients were on HCTZ, they were more likely to experience hypokalemia and require potassium supplementation. When patients were on amiloride, hyperkalemia was more likely. Therefore, in terms of potential adverse effects and drug interactions, HCTZ and amiloride were comparable. Lithium and potassium concentrations must be monitored for either drug.
CONCLUSION
It has been consistently demonstrated that lithium decreases urinary concentrating ability to produce a nephrogenic diabetes insipidus as a result of the drug's inhibition of vasopressin to produce the symptoms of polyuria and polydipsia. These symptoms usually are most pronounced in pateints who have experienced lithium intoxications. This effect is usually reversible on discontinuation of the drug. Although histological changes have been described in the literature, no long-term studies have been able to associate a decrease in the glomerular filtration rate with the biopsy changes. The focal interstial nephropathy that has been reported is only associated with lithium intoxications. Thus the risk of renal damage and dysfunction appears to be extremely small in patients on stable maintenance lithium treatment who have no prior history of lithium intoxication.
REFERENCES
Abraham G, Delva N, Waldron J et al (1992). Lithium treatment: a comparison of once-and twice-daily dosing. Acta Psychiatr Scand 85:65-9.
Achong MR, Fernadez PG, McLeod PJ (1975). Fatal self-poisoning with lithium carbonate. Can Med Associ J 112:868-70.
Amdisen A, Skoldborg H (1969). Hemodialysis for lithium poisoning. Lancet 2:213.
Anon (1979). Lithium and the kidney. Grounds for cautious optimism. Lancet 1056-7.
Batlle DC, van Riotte AV, Gavairia M, Grupp M (1985). Amelioration of polyruia by amiloride in patients receiving long-term lithium therapy. New Engl J Med 312:408-14.
Boton R Moises M, and Batlle C (1987). Prevalence, pathogenesis, and treatment of renal dysfunction associated with chronic lithium therapy. Am J Kid Dis 10:329-45.
Burrows GD, Davies B, Kincaid-Smith P (1978). Unique tubular lesion after lithium. Lancet 2:1310.
Christensen EM and Aggernaes H (1990). Prospective study of EDTA clearance among patients in long-term lithium treatment. Acta Psychiatr Scand 81:302-3.
Christiansen D, Kusano E, Yusufi ANK (1985). et al. Pathogenesis of nephrogenic diabetes insipidus due to chronic administration of lithium in rats. J Clin Invest 75:1869-79.
Cockroft DW, Gault MH (1976). Prediction of creatinine clearance from serum creatinine. Nephron 16:31-47.
Coppen A, Bishop ME, Bailey JE, Cattell WR (1980). Renal function in lithium and non-lithium treated patients with affective disorders. Acta Psychiatry Scand 62:343-55.
DePaulo JR, Correa EI, Sapir D (1981). Renal glomerular function and long-term lithium therapy. Am J Psychiatry 324-27.
DePaulo JR, Correa EI, Sapir D (1986). Renal function and lithium: a longitudinal study. Am J Psychiatry 143:892-5.
Dias N, Hocken AG (1973). Oliguric renal failure complicating lithium carbonate therapy. Nephron 10:246-9.
Donker AJM, Prins E, Meiger S, Sluiter WJ, van Berkestijn JW, Dols LCW (1979). A renal function study in 30 patients on long-term lithium therapy. Clin Nephrol 12:254-62.
Forrest JN, Cohen AD, Torretti J, et al (1974). On the mechanism of lithium-induced diabetes insipidus in man and the rat. J Clin Invest 53:1115-23.
Gerner RH, Psarras J, Kirschenbaum MA (1980). Results of clincial renal functin tests in lithium patients. Am J Psychiatry 137:834-6.
Greenwell BE (1980). Pharmacokinetic parameters and computer dose simulations for two proprietary lithium preparations (Lithotabs and Lithobid). Personal Communication.
Grindlinger GA and Boylan MJ (1987). Amelioration by indomethacin of lithium-induced polyuria. Crit Care Med. 15:538-9.
Hansen HE, Amdisen A (1978). Lithium intoxication. Quart J Med (NS) 47:123-44.
Hawkins JB, Dorken PR (1969). Lithium. Lancet 1:839-40.
Hestbeck J, Hansen HE, Amdisen A, Olsen S (1977). Chronic renal lesions following long-term treatment with lithium. Kidney Int 12:305-13.
Hetmar O, LBolwig TG, Brun C, et al (1986). Lithium: long-term effects on the kidney. I. renal function in retrospect. Acta Psychiatr Scand. 73:574-81.
Hullin RP (1979). Effects on renal and thyroid function and bond metabolism in long-term maintenance treatment with lithium salts. In Lithium: Controversies and Unresolved Issues, Cooper TB, Gershon S, Kline NS, Schou, M (eds), Amsterdam: Excerpta Medica pp 341-345.
Johnson GFS, Hunt GE, Duggin GG, et al (1984). Renal Function and Lithium Treatment: Initial and Follow-up Test in Manic Depressive Patients. J Affective Disord. Vol 6, pg 249-63.
Juhl RP, Kronfol Z, Tsuang MT (1976). Lithium-induced nephrogenic diabetes insipidus. Am J Hosp Pharm 33:843-5.
Kincaid-Smith P, Burrows GD, Davies BM, Holwill B, Walter M, Walker RG (1979). Renal biopsy findings in lithium and pre-lithium patients. Lancet 2:700-1.
Kosten TR, Forrest JN (1986). Treatment of severe lithium-induced polyuria with amiloride. Am J Psychiatry 143:1563-68.
Lavender S, Brown JN, Berrill WT (1973). Acute renal failure and lithium intoxication. Postgrad Med J 49:277-9.
Levy ST, Forrest JN, Henninger GR (1973). Lithium-induced diabetes insipidus: manic symptoms, brain and electrolyte correaltes, and chlorothiazide treatment. Am J Psychiatry 130:1014-8.
Lokkegaard H, Andersen NF, Henriksen E, et al (1985). Renal function in 153 manic-depressive patient treated with lithium for more than five years. Acta Psychiatr Scand 71:347-55.
Mellerup ET, Dam H, Plenge P, et al (1988). Serum lithium minimum and diuresis. Psychiatry Res 14:309-13.
O'Donovan C, Hawkes J, Bowen R (1993). Effect of lithium dosing schedule on urinary output. Acta Psychaitr Scand 87:92-5.
Penney MD and Hampton D (1990). The effect of lithium therapy on arginine vasopressin secretion and thirst in man. Clin Biochem. 23:233-6.
Perry PJ, Dunner FJ, Hahn RL, Tsuang MT, Berg MJ (1981). Lithium kinetics in single daily dosing. Acta Psychiatry Scand 64:281-94.
Plenge P, Mellerup ET, Norgaard T (1981). Functional and structural rat kidney changes by peroral or parenteral lithium treatment. Acta Psychiat. Scand 63:303-13.
Plenge P, Mellerup ET, Bolwig TG, et. al (1982). Lithium treatment: does the kidney prefer one daily dose instead of two? Acta Psychiatr Scand 66:121-8.
Radomski JL, Fuyat HN, Nelson AA, Smith PK (1950). The toxic effects, excretion and distribution of lithium chloride. J Pharmacol Exp Ther 100:429-44.
Rafaelson OJ, Bolwig TG, Ladefoged J, Brun C, Hetmar O, Larsen S (1979). Kidney function and morphology in long-term treatment lithium treatment. In Lithium: Controversies and Unresolved Issues, Cooper TB, Gershon S, Kline N, et al (eds), Amsterdam: Excerpta Medica pp 578-583, 666.
Schou M, Amdisen A, Trap-Jensen J (1968). Lithium poisoning. Am J Psychiatry 125:520-7.
Schou M, Amdisen A, Thomsen K, et al (1982). Lithium treatment regimen and renal water handling: The significance of dosage pattern and tablet type examined through comparison of results from two clinics with different treatment regimens. Psychopharmacology 77:387-90.
Schou M (1988). Effects of long-term lithium treatment on kidney function: an overview. J Psychiat Res 22:287-96.
Verbov JL, Phillips JD, Fife DG (1965). Case of lithium intoxication. Postgrad Med J 41:190-2.
Waller DG, Edwards JG, and Papasthatis-Papayanni S (1988). A longitudinal assessment of renal function during treatment with lithium. Q J Med 68(255):553-8.
Wallin L, Alling C (1979). Effect of sustained-release lithium tablets on renal function Br Med J 2:1332.