Thomas A.Ban, editor. Lithium in Psychiatry in Historical Perspective.


Gordon Johnson’s Comment


       The beginning of my experience with lithium was timely, with interest in lithium in psychiatry in the U.S.  escalating and my appointment to work with Sam Gershon, a leading researcher and advocate for approval for the clinical use of lithium. It could not have been more appropriate and engrossing.

       My initial research role was in the ongoing randomized clinical trial comparing lithium and chlorpromazine in the treatment of manic states (Johnson, Gershon and Hekimian 1968). A control group of schizoaffective patients was included (Johnson 1970). I was involved in most aspects of the trial such as patient selection and consent, but mainly clinical trial assessment and management of patients admitted to a hospital metabolic unit. Patients were assigned in a random manner to either drug. Serum lithium levels were known only to the psychiatrist assigning treatment. While this was a double-blind trial, the pharmacological differences between lithium and chlorpromazine were such that it was often evident which drug the patient was taking during the three-week trial period.   

       The clinical trial assessed the benefit and toxicity with an increasing dose until a therapeutic response or toxicity occurred. The blind was then broken and further management decided. The interim report showed 78% of lithium treated patients remitted compared to 36% of the patients assigned to chlorpromazine. The two drugs differed in their psychopharmacological action. Both drugs produced a decrease in motor activity, this occurring earlier with chlorpromazine. Normalization of affect and ideation occurred at optimum doses with lithium with no significant side effects. A more detailed analysis of the rating scales showed that lithium produced a controlled remission in acute mania in 10 to 14 days with significant changes in the core symptoms of motor hyperactivity, flight of ideas, euphoria and pressure of speech (Johnson, Gershon, Burdock et al. 1971).

       Fortuitously, I attended a meeting of The Veterans Administration and the National Institute of Mental health collaborative study group reporting on a similar comparative lithium and chlorpromazine study in mania with patients divided, however, into highly active and mildly active. This largely addressed a major issue with the delay in action of lithium in the treatment of acute mania. In the collaborative study there was a high dropout rate of highly active patients on lithium, which was not evident in the mildly active. Chlorpromazine was clearly superior to lithium in treating the highly active patients more quickly and had fewer dropouts. Lithium appeared to be the better treatment in the mildly active patient and left the patient less sluggish and fatigued. In keeping with these findings combined or sequential lithium and antipsychotic drugs became widely used in clinical practice. 

       While the safety of lithium depends on serum lithium monitoring a wider safety profile requires a better understanding of the neurotoxic effects. This is one of the most unpredictable, alarming and potentially serious complications of lithium therapy. Having worked in electroencephalography in London I negotiated a collaborative EEG study with the department of neurology with two major goals, firstly to gain insight in the relationship between brain function as shown in the EEG changes and serum lithium levels, toxic manifestations and clinical psychiatric changes. The second was to ascertain whether there were specific neurophysiological changes correlated with behavioral changes following acute and chronic administration of lithium in patients and controls (Johnson 1969). 

       Acute lithium administration produced minimal EEG changes from baseline without behavioral change. Following chronic administration EEG changes were highly correlated with the presence and severity of neurotoxicity and behavioral change. There was no association with clinical psychiatric changes. There was a relationship with serum lithium levels, but patient-specific variables were important such as organic brain damage, sodium depletion, drug interaction and concurrent physical illness. Clinical and EEG changes were evident for up to one month following cessation of lithium (Johnson, Maccario, Gershon and Korein 1970). A review of reported case studies dating back to the use of lithium as a salt substitute in the US was in keeping with these findings (Johnson 1976). Changes in the EEGs were nonspecific but in the absence of any other factor they were a sensitive index of neurotoxicity. 

       That continuation of lithium treatment following remission of mania may prevent relapse was observed in early studies. The long-term prophylactic effect of lithium against recurrences of mania and depression was first systematically studied by Baastrup and Schou (1967). A group of manic-depressive patients who had suffered two or more manic or depressive episodes during a single year or one or more episode per year for at least two years before being given lithium and had been on lithium continuously for at least 12 months. The patients were observed for 6.5 years: comparison of relapse frequency prior to and following lithium therapy showed a significant reduction. Recurrence occurred on average every eight months prior to lithium but every 60-85 months during lithium. Recurrences were prevented with equal efficacy in patients having both mania and depression or depression only. A more extensive multi-center trial using a similar, open before and after design showed similar results (Angst, Weis, Grof et al. 1970). 

       While these studies provided evidence of the prophylactic value of lithium, they were criticized on the grounds that the evaluations were not blind and that improvement may have been due to spontaneous remissions (Blackwell and Shepherd 1968). Controlled evidence of lithium's effectiveness as a prophylactic agent in preventing both mania and depression in bipolar illness and depression in unipolar illness was provided by Baastrup, Poulsen, Schou et al. (1970) who discontinued lithium maintenance in matched pairs of 50 manic depressive patients and 34 patients with “recurrent depression” who had been maintained on lithium for a period of up to seven years.

       Patients were randomly assigned to placebo or lithium and the trial continued until a significant relapse frequency between pairs was obtained. The trial lasted five months, relapse occurring in 21 patients on placebo but in none of the patients on lithium. The volume of evidence establishing the prophylactic effect of lithium was substantial and lithium prophylaxis became the standard treatment for bipolar disorder.  This was arguably the most important advance since Cade reported the antimanic effects of lithium in 1949. However, the benefits in clinical outcomes were dependent on routine lithium monitoring.   

       Historically, patients receiving lithium were largely treated in hospital or university-related clinics by investigators or clinicians interested in its use. They provided consistent standards, regular monitoring, patient education and support. The current reality is that most patients are treated in the community and there is evidence that there are serious deficiencies in lithium surveillance that will impact on its effectiveness.

       A survey of lithium monitoring in hospital and general practice in the U.K found wide variation in monitoring practice, with less than 33% patients having lithium levels estimated at intervals of less than 13 weeks, while 76% had lithium levels at intervals less than 26 weeks; 25% of subjects had lithium levels greater than 1nmol/L and of these only 55% were followed by a repeat level within three weeks; 8% had consistently low levels less that 0.4. Only 15% met their ideal monitoring criteria, that is at least three, monthly lithium assessments, lithium levels within the therapeutic range 0.4-1.0, and action taken outside these levels and six monthly electrolyte and thyroid assessment (Ryman 1977). 

       Deficiencies in lithium monitoring practices were not confined to the U.K. My experience from patients referred to the University of Sydney department of psychiatry lithium clinic and from colleagues found monitoring a frequent problem.  Patient adherence to lithium is a problem all clinicians experience. Continuing treatment is the exception rather than the rule. Estimates of non-compliance in patients with bipolar disorder range from 12-60%. A survey of patients at a university Affective Disorders clinic found that 47% stopped taking lithium at least once and 34% had stopped taking lithium more than once (Jamison and Goodwin 1983).  Common side effects such as tremor were often given as the reason for stopping lithium. As these are sensitive to dose and plasma level, fine tuning can make a difference in compliance.

       Polyuria is a well-recognized complication of lithium treatment that has been attributed to a decreased responsiveness of the renal tubules to antidiuretic hormone (ADH) which in some patients may progress to a diabetes insipidus-like syndrome. Such increased urine volume has been considered a reversible pharmacological side-effect of lithium treatment, but with potential risk of lithium toxicity for patients during sodium depletion, excessive fluid loss or reduced intake. However, evidence of a non-specific interstitial nephropathy was reported in patients following lithium treatment and the degree of tubular damage found on renal biopsy correlated with the degree of impairment of renal concentrating ability which, in turn, was related to duration of lithium treatment, suggesting a progressive nephrotoxicity (Johnson, Hunt, Duggin et al. 1981). If lithium treatment produces a slowly progressive impairment of renal function in some patients, then cross-sectional studies are of limited value. Only prospective studies, using patients as their own controls, will enable identification of progressive impairment of renal function and significant risk factors.

       In collaboration with renal physicians an assessment of renal function was carried out in a sample of patients with bipolar disorder receiving lithium for an average period of 4.5 years. Overall, glomerular filtration rate (GFR) was within the established normal range based on sex and age, whereas measures of urinary concentrating ability were generally impaired. There was no relationship between duration of lithium treatment and either GFR or impairment of urinary concentrating ability. Moreover, there was no evidence of a progressive impairment of glomerular or tubular function in patients re-tested after two years (Johnson, Hunt, Duggin et al. 1984). The results of this study confirmed the safety of lithium administration in most patients and emphasized the importance of careful clinical monitoring to avoid lithium intoxication. In advancing age, there is a fall in GFR, and therefore, lithium elimination. Failure to adjust lithium dosage with rising plasma lithium levels exposes patients to lithium accumulation and intoxication. Lithium has become the most frequent cause of nephrogenic diabetes insipidus and management often involves shared care with a renal physician.  The benefit to risk with long-term maintenance requires that lithium levels should be kept at the lowest concentrations consistent with adequate therapeutic effect.

       At the 1997 Collegium Internationale Neuro-Psychopharmacologicum meeting a Festschrift was presented to Sam Gershon. The review concluded that lithium had emerged as an exceptional drug. It was estimated that 0.77 to 1.5 of every 1,000 persons in the U.S. population were on lithium treatment. This period could be considered the high point of lithium's impact, its proven clinical indications encompassing not only the treatment of mania and recurrent bipolar disorder but also playing a role as augmentation therapy in refractory depression and in the management of recurrent depression.  However, doubts had arisen concerning the clinical utility of lithium as measured by its effectiveness in the maintenance treatment of bipolar disorder in ordinary clinical practice, with several reports finding the benefits modest compared to the results of clinical trials. Additionally, the use of alternative medications had challenged lithium's role in acute mania and long-term management of bipolar disorder (Keck and McElroy 1996). Surveys of drug use had found a high proportion of bipolar patients receiving multiple medications, including neuroleptics, antidepressants, benzodiazepines and anticonvulsants as well as lithium (Solomon, Keitner, Ryan and Miller 1996). One investigator concluded that lithium monotherapy is not sufficient for most manic patients who were referred for tertiary care (Small, Klapper and Milstein 1991).

       The National Centre for Health Statistics in the U.S. found outpatients with mania four times more likely to be treated with multiple concurrent psychotropic medications than were outpatients with non-psychiatric diagnoses. Lithium was part of a polypharmacy regime in more than 70% of the ambulatory visits in which it was prescribed and out of all psychotropic drugs lithium was most likely to be used in combination with other psychotropic drugs. They concluded the two best predictors of psychotropic polypharmacy were the prescription of lithium and a diagnosis of mania (Nichol, Stimmel and Lange 1995).

       These follow up studies suggest that for many patients with bipolar disorder lithium is ineffective or partially effective leading to combination medication strategies. In others initial benefit may not be maintained or escape from effective prophylaxis occurs. It is unclear whether this represents a lessening of the pharmacological action over time or that the driving force of the illness overwhelms the prophylactic effects of lithium (Coryell, Winokur, Solomon et al. 1997; Maj, Pirozzi and Magliano 1996). Response to these findings occasioned mixed comments. Schou (1993) trenchantly observed that the naturalistic studies reveal what happens to carelessly selected manic-depressive patients when they are given long-term lithium treatment under conditions of insufficient information, support and supervision.

       In contrast, Guscott and Taylor (1994) reviewing outcome studies with lithium in medical practice and comparison with clinical trials, stated "that effectiveness in naturalistic studies show poorer results than efficacy studies in all areas of medicine." Furthermore, several reports had found evidence suggesting a mortality lowering effect of lithium, with a reduction in expected deaths from suicide and cardiovascular complications in patients on lithium maintenance in long-term follow up studies (Schou 1995).

       In 1999 meetings were held to mark the 50th anniversary of Cades discovery of the therapeutic effects of lithium (Cade 1949). That lithium's future was now in doubt came as a surprise. However, questions concerning the efficacy and effectiveness of lithium had been widely canvassed, creating uncertainty in the minds of clinicians (Moncrieff 1997). Also, the emergence of newer therapeutic drugs such as the anticonvulsants and the novel antipsychotics (Dardennes, Even, Bange and Heim 1995) compelled re-assessment of lithium's place in the treatment of bipolar disorder (Goodwin and Jamison 1990). This was accompanied by a vigorous marketing campaign by pharmaceutical manufacturers to establish a pre-eminent place for their products. There is little argument that the availability of new treatments represents a significant advance to patients and clinicians struggling to improve outcomes in the management of mania and the long-term prevention of recurrent episodes.

       However, in this context lithium use is in danger of being overlooked, an outcome not only to lithium's disadvantage but also to patient outcomes. Lithium lacks an advocate to press its case. The future of lithium should be secure. It is the best established agent for long term maintenance treatment of recurrent bipolar disorder and its use is associated with a potential impact on suicidal behavior. It is a first line agent for the treatment of mania. Lithium augmentation of antidepressant agents is an effective and proven strategy in treatment refractory depression. The safety profile of lithium is well established, with no likely surprises. Evidence suggests, however, that lithium monitoring to optimize effectiveness and minimize risks is often inadequate. Declining interest in lithium use will further "deskill" clinicians and affect their choice of lithium even where the evidence shows it to be the preferred treatment. In the worst-case scenario static or declining use may lead the manufacturer to withdraw lithium formulations from the market. Recent examples are readily available such as has occurred with the tricyclics and monoamine oxidase inhibitors. We like to think otherwise but market forces and government subsidies determine drug availability. Lithium remains a unique and important treatment. Its continued use, availability and advocacy remains the responsibility of the profession.



Angst J, Weis P, Grof P, Baastrup P, Schou M. Lithium prophylaxis in recurrent affective disorders. British Journal of Psychiatry 1970;116:604-14 

Baastrup PC, Schou M. Lithium as a prophylactic agent-its effects against recurrent depression and manic depressive psychosis. Arch Gen Psychiatry 1967;16(2):162-72. 

Baastrup PC, Poulsen JC, Schou M, Thomsen K, Amdisen A. Prophylactic lithium: double blind discontinuation in manic depressive and recurrent depressive disorders. Lancet 1970;2(7668):326-30. 

Blackwell B, Shepherd M. Prophylactic lithium: another therapeutic myth? An examination of the evidence to date Lancet 1968;1(7549):968-71. 

Cade JF. Lithium salts in the treatment of psychotic excitement. Med J Aust 1949; 2(10):349-52.

Coryell W, Winokur G, Solomon D, Shea T, Leon A, Keller M. Lithium and recurrence in a long term follow up of bipolar disorder. Psychol Med ;27(2):281-9.

Dardennes R, Even C, Bange F, Heim A. Comparison of carbamazepine and lithium in the prophylaxis of bipolar disorder. A meta-analysis. Br J Psychiatry 1995;166(3):378-81.

Goodwin F, Jamison K. Manic-depressive illness. Oxford University Press; 1990, p. 688. 

Guscott R, Taylor L. Lithium prophylaxis in recurrent illness efficacy, effectiveness and efficiency. Br J Psychiatry 1994;164(6):741-6. 

Jamison K, Goodwin, F. Psychotherapeutic issues in bipolar illness in psychiatry update. The APA Annual Review Washington DC 1983;2:319-43. 

Johnson G. Lithium and the EEG: an analysis of behavioral, biochemical and electrographic changes. Electroencephalogr Clin Neurophysiol;27(7):656-7. 

Johnson G. Differential responses of manic depressive and schizoaffective patients to lithium carbonate. Diseases of the Nervous System 1970;31:9, 613-15. 

Johnson GF. Lithium Neurotoxicity. Aust N Z J Psychiatry 1976;10(1):33-8. 

Johnson G, Gershon S, Burdock EI, Floyd A, Hekimian L.  Comparative effects of lithium and chlorpromazine in the treatment of acute manic states. British Journal of Psychiatry 1971;119(550):267-76. 

Johnson G, Gerson S, Hekimian L. Controlled evaluation of lithium and chlorpromazine in the treatment of manic states: an interim report. Compr Psychiatry 1968;9(6):563-73. 

Johnson G, Hunt G, Duggin GG, Tiller DJ, Horvath JS. Lithium nephrotoxicity. Med J Aust 1981;1(7):376.

Johnson GF, Hunt GE, Duggin GG, Horvath JS, Tiller DJ. Renal function and lithium treatment: initial and follow-up tests in manic-depressive patients. J Affect Disord 1984;6(3-4):249-63. 

Johnson G, Maccario M, Gershon S, Korein J. The effects of lithium on electroencephalogram, behavior and serum electrolytes. The Journal of Nervous and Mental Disease 1970;151(4):273-89. 

Keck PE, McElroy SL. Outcome in the Pharmacologic treatment of bipolar disorder. J Clin Psychopharmacol 1996;16(2 Suppl 1):15S-23S. 

Maj M, Pirozzi R, Magliano L. Late non-response to lithium prophylaxis in bipolar patients: prevalence and predictors. J Affect Dis 1996;39(1):39-42. 

Moncrieff J. Lithium: Evidence Reconsidered. Br J Psychiatry 1997;171:113-9. 

Nichol MB, Stimmel GL, Lange SC. Factors predicting the use of multiple psychotropic medications. J Clin Psychiatry 1995;56(2):60-6.

Ryman A. Lithium monitoring in hospital and general practice. Psychiatry Bull 1977:21(9):570-2. 

Schou, M. Lithium prophylaxis about 'naturalistic' or clinical practice studies. Lithium 1993;4,77-81. 

Schou M. Mortality lowering effect of prophylactic lithium treatment: a look at the evidence. Editorial Pharmacopsychiatry, Pharmacopsychiatry 1995;28(1):1.

Small JG, Klapper MH and Milstein V.  Carbamazepine compared with lithium in the treatment of mania. Arch Gen Psychiatry 1991;48:915-21.  

Solomon DA, Keitner GI, Ryan CE, Miller IW. Polypharmacy in bipolar I disorder. Psychopharmacology Bulletin 1996;32(4):579-87.



October 14, 2021