Hypothesis Paper

Mood-stabilizers: the archeology of theconcept

In recent years the treatment of bipolar mooddisorders has changed dramatically with sodiumvalproate, carbamazepine, lamotrigine, and otheranticonvulsants now used regularly in addition toor in lieu of lithium. There is a general acknow-ledgement that pharmaceutical company interest inthe area of bipolar disorders has played some partin sustaining a wider interest. But this widerinterest has also led to the emergence of conceptualmodels challenging traditional notions in thistherapeutic domain. For example antidepressantsare routinely used in the depressed phase of abipolar disorder, but there is in fact very littleevidence to support this practice (1) and somereason to believe that antidepressants paradoxic-ally may make the problem worse (2, 3). There areclear implications of such perspectives for thetheoretical models that underpin clinical practice.A great deal hinges on the concept of a mood-

stabilizer. For three decades, lithium stood as whatwould now be called a mood-stabilizer in contrastto the �antidepressants�. The answer to the questionwhat is a mood-stabilizer was simple – it waslithium. The emergence of other compounds forcesus to go further. What does it now take to show a

drug is a mood-stabilizer? Will defining a drug as amood-stabilizer then lead to people who respondto that drug being diagnosed as bipolar patients?Would this be appropriate?This paper attempts to shed light on these

questions by charting aspects of the developmentof anticonvulsants for mood disorders. An accom-panying paper will provide comparative data onthe incidence and prevalence of service utilizationfor patients diagnosed as having bipolar mooddisorders and other data such as inter-illnessintervals from the pre- and postlithium periods.

The archeology of mood-stabilization

The initial use of lithium for mania created animpression that the manic pole of manic-depressiveillness might almost involve a lithium deficiencystate. The possibility of what would now be termedmood-stabilizing effects arose in the late 1960s.Two studies by Schou and Baastrup laid the basisfor claims that lithium was prophylactic for manic-depressive episodes (4–6). But the response to theseclaims was vigorous with critics of the conceptarguing that the results of the naturalistic studies

Harris M, Chandran S, Chakraborty N, Healy D. Mood-stabilizers: thearcheology of the concept.Bipolar Disord 2003: 5: 446–452. ª Blackwell Munksgaard, 2003

Objective: To review the history of �mood-stabilizing� treatments.

Method: We have reviewed primary source data on the origin of theuse of current mood-stabilizers.

Results: This historical record on the origins of the mood-stabilizerspoints to a more ambiguous picture as regards pharmacotherapeuticspecificity to bipolar disorders than is commonly conceded.

Conclusions: This review suggests a need for alternative formulationsof the concept of a mood-stabilizer. An alternative to the currentlydominant paradigm is that these agents have treatment effects, whichneed to be matched more precisely with patients� constitutional types inorder to optimize outcomes.

Margaret Harris, Summit Chandran,Nabonita Chakraborty and DavidHealy

North Wales Department of Psychological

Medicine, Hergest Unit, Bangor, UK

Key words: bipolar disorder – carbamazepine –

lithium – mood-stabilizer – prophylaxis – valproate

Received 28 October 2002, revised and accepted

for publication 13 February 2003

Corresponding author: David Healy, North Wales

Department of Psychological Medicine, Hergest

Unit, Bangor, LL57 2PW, UK.

Fax: 44 1248 371397;

e-mail: healy_hergest@compuserve.com

Bipolar Disorders 2003: 5: 446–452Copyright ª Blackwell Munksgaard 2003

BIPOLAR DISORDERS

446

that formed the basis for claims for lithium’sprophylactic effects might simply reflect a regres-sion to the mean, or the effects of a withdrawalsyndrome. A �mirror-image� service utilizationstudy of patients before and after lithium byAngst, Weiss, Grof, Baastrup and Schou in 1970(7), and a randomized controlled trial (8) appearedto settle the issue – lithium was what would now betermed a mood-stabilizer.However, the data on service utilization from the

study by Angst et al., which did so much to laythe basis for the concept of a mood-stabilizer in the1970s, from the perspective of the present look lessclear-cut than standard interpretations of the studysuggest. In part, this is because by 1970, the �mood-stabilizing� properties of valpromide had alreadybeen discovered, and reports were just about toemerge of lithium’s benefits in conditions otherthan manic-depressive illness.

The origins of valproate/valpromide

The origins of valproate and valpromide lie in theSecond World War and efforts by German scien-tists to produce butter substitutes (9). These effortsled to the synthesis of valproic acid. After the warvalproic acid was used as a common diluent forother drugs. In 1963 George Carraz of the Labo-ratoire Berthier at Grenoble, when asked to testout a new product for possible anticonvulsantproperties dissolved the new compound in valproicacid. Testing failed to show any correlationbetween different doses of the experimental com-pound and anticonvulsant activity but yet themixture was anticonvulsant. Carraz realized thatthe anticonvulsant properties stemmed from valp-roic acid and titrating the dose of this demonstra-ted the issue conclusively.Carraz synthesized valproate (Depakine) and

valpromide (Depamide) derivatives of valproicacid. The conventional wisdom of the time had itthat an azote moiety would enhance the psycho-tropic properties of a compound, and it was thisthat led to the synthesis of valpromide. Valpromidein fact protects animals from epileptic convulsionstriggered by strychnine where valproate does not.Valpromide also crosses the blood brain barriermore readily leading to higher CNS concentrationsthan valproate.Carraz had a link with Sergio Borselli a psychi-

atric trainee with Pierre Lambert at Bassens Hospi-tal in Rhone-Alpe. This led to the primary tests ofthe anticonvulsant properties of both valproate andvalpromide in Bassens Hospital (10–12).At that point in time most large asylums in

Europe had significant populations of epileptic

patients – from 10 to 20%. This gave a readypopulation in which to try out a new anticonvul-sant. Borselli and Lambert initially found valpro-mide intensely sedative, particularly when added toother anticonvulsants such as phenobarbitone.When valpromide was finally administered on itsown, it became clear that it had psychotropic inaddition to neurotropic effects. This has beendescribed by Lambert as follows �patients felt morethemselves, the mental stickiness, viscosity that hadsometimes been there on older agents, was less. Wesaw the disappearance of tendencies to depression,sometimes even a mild euphoria� (13).Epilepsy was then thought to predispose to both

schizophreniform developments, and an epilepticpersonality disorder. Epileptic patients were seenas importunate, manipulative and viscous intheir personalities. These patients were frequentlydetained in hospital not because of their convul-sions but because of the social disturbances theycaused. They were thought to have impulse controldisorders, which underlay their inability to adaptto normal social life. The other feature of theirpersonalities was a certain obsessionality. Onvalpromide, these social disturbances and thecharacteristic importunate behavior of hospitalizedepileptics appeared to change. Female patients inparticular were less likely to end up in conflicts, lesslikely to provoke others in their surroundings, andless likely to self-harm. This led Lambert andBorselli to ask whether valpromide reduced self-harm tendencies; was it anti-masochistic?These issues return in the case of the discovery of

carbamazepine and pose a real question. Thedegree of control of convulsions is not significantlybetter now compared with before but it is clear thatepileptic patients do not end up in mental hospitalsin a way that they did before. Is this because ofa beneficial effect of these drugs on personalityand general integration that has been all butun-investigated? Is this beneficial effect whatunderpins mood-stabilization?At this time, lithium was unavailable in France

or was more generally thought of as being ineffec-tive. There was a premium therefore on findingeffective treatments for manic-depressive disorder.The standard maintenance regimes at the timeinvolved the use of antipsychotics such as chlor-promazine or levomepromazine. The sedativeproperties of valpromide led to its use in combi-nation with chlorpromazine for agitated and manicpatients, just as phenobarbitone had been used. Onrecovery, patients left on valpromide alone showedan enhanced compliance compared with patientson chlorpromazine. Altogether Lambert et al.studied the drug in approximately 250 patients

Mood-stabilizers: the archeology of the concept

447

and concluded that valpromide had distinct psych-otropic effects that were of benefit in the treatmentof both acute manic states and in the maintenancetreatment of manic depressive illness (14). This ledto a study looking more closely at 32 patients andat the impact of valpromide on rates of hospita-lization before and after exposure to this agent. Inline with the earlier findings of Angst et al. forlithium, there appeared on valpromide to be a fallin the number of manic episodes by 50% and adecrease of 60% in the duration of hospitalization(15).Valpromide at this stage, however, was not

promoted by Laboratoire Berthier as valproatewas selling well as an anticonvulsant both inFrance and abroad. Valproate also began to beused for mood disorders and in 1980 Emrich et al.(16) reported on its usefulness for the managementof mania without knowing about the prior use ofvalpromide.

The origins of carbamazepine

In the early 1970s lithium was not available inJapan. This led to the use of a wider variety ofagents to manage manic-depressive disorders thanwere being used in countries where lithium wasavailable. Japanese hospitals were also in theprocess of institutionalizing with a large increasein the hospital population following the discoveryof chlorpromazine in contrast to the reductionselsewhere (17). Japanese psychiatry was neuropsy-chiatrically oriented and the treatment of epilepsylay within the domain of psychiatrists. This meantthat a considerable number of patients were treatedin asylums for epileptic or related conditions. As aresult, carbamazepine, a tricyclic agent, came intouse within the asylum following its release as ananticonvulsant during the 1960s.The availability of and sedative properties of

carbamazepine almost inevitably led to its usein lieu of other sedative agents such as thebarbiturates in manic patients. In an echo ofthe valpromide story, it was noted that the use ofcarbamazepine contributed something distinctiveto the management of both epileptic and manicpatients (18, 19).These factors laid the basis for a clinical trial of

carbamazepine in the treatment of manic-depres-sive illness (20–22). When written up in English,this trial in which carbamazepine was comparedwith chlorpromazine demonstrated comparableresults to chlorpromazine (23). However the articlehad a poor reception in the Western literature withthe criticism that almost homeopathic doses ofchlorpromazine had been used (250 mg/day) and

that therefore there was no evidence of efficacy forcarbamazepine (19). This was during a periodwhen megadose regimes of neuroleptic agents wereused in the West against which a 250 mg dose ofchlorpromazine may well have looked indistin-guishable to placebo as a comparator. The proto-col used however was exactly the same protocolused to investigate lithium and these results werenot contested. The results of carbamazepine andlithium indeed appeared to be comparable (24, 25).Carbamazepine however did not emerge into

wider use in the West until its psychotropic effectswere documented by Ballenger and Post in 1980.By the time it emerged into wider use, it was clearthat carbamazepine had interesting psychotropicproperties. It had been used in Japan for a widerange of conditions and it was noted to be useful instabilizing aggressive outbursts in young men.Young men with impulse control disorders repor-ted that a break was interposed between them andtheir impulses so that they were allowed a pause forreflection that they did not have before. This use ofcarbamazepine entered into the Western literatureas a use in the management of episodic dyscontrolsyndrome (19, 26).

Parallel developments

Lithium was undergoing a parallel evolution.When first introduced in the 1950s, it had appearedto be a specific treatment for manic-depressiveillness. From there, it migrated during the 1960s tobecome a prophylactic treatment. In the early1970s, a study by Sheard in prisoners demonstratedan anti-irritability, or anti-impulsive action that ledto a reduction in violent behavior among prisoninmates (27). This study, which was immediatelyreplicated (28), questioned the basis for the sup-posed specificity of lithium’s effects to manic-depressive illness. Paradoxically at the same timethe concept of bipolar disorder was broadening outto encompass anyone who might respond to amood stabilizer. The licensing of lithium in theUnited States in conjunction with other historicalprocesses was leading to a re-diagnosis to manic-depressive illness of many patients formerly diag-nosed as having schizophrenia (29, 30).But there is another neglected history here.

Through the 1960s a variety of other anticonvul-sants were also used for non-epileptic indications.These included diphenylhydantoin (31, 32), becla-mide (33) and sulthiame (34, 35). These drugs wereused a variety of psychotic and behavioral condi-tions including what were later called conductdisorders in children and are now liable to bediagnosed as juvenile onset bipolar disorders. This

Harris et al.

448

use of drugs like sulthiame and diphenylhydantoin,however, unlike the use of carbamazepine andvalproate did not get linked to bipolar disorders atthe time and did not lead to the concept of mood-stabilization.

The emergence of mood-stabilization

Based on reports of the effectiveness of carbamaze-pine for mood disorders, Post et al. (36–38)suggested that the efficacy of this tricyclic anticon-vulsant might be explained if episodes of mooddisorder were conceptualized as convulsive equiv-alents. Mood stabilization might then involve areduction of the kindling effects that primedsubsequent episodes. While Schou almost 20 yearsearlier had talked about mood-normalizers (39),Post’s formulations, which linked a proposedmechanism of action with prophylactic effects,inaugurated a new era of mood-stabilization,although the concept was nevertheless slow to takeshape – the term mood-stabilizer in fact onlyappears sporadically in the literature until the earlyto mid-1990s.As this new concept took shape, the proposed

effect of mood-stabilizers was relatively disease-specific and furthermore was one that should occurregardless of any beneficial non-specific functionaleffects such agents might also have. In addition, itfollowed from Post’s proposals that the longerthe period the person was left untreated and thegreater the number of episodes they had the greaterthe propensity to future episodes would be. Thisconceptualization coincided with contemporarythinking about lithium and it mandated earlyintervention. Evidence that valproate had similarmood stabilizing properties to carbamazepineappeared to endorse the kindling hypothesis.The kindling model put a premium on investi-

gating other anticonvulsants. Beneficial psycho-tropic effects in patients being treated for epilepsy,echoing those previously seen with valproate andcarbamazepine, were also described for lamotri-gine, gabapentin, vigabatrin and other anticonvul-sants. However, not all anticonvulsants appear tobe of benefit in manic-depressive disorders. Thecurrent status of gabapentin is uncertain (40), andit would seem that vigabatrin is unhelpful, tiaga-bine may be of limited utility and topiramate is notroutinely helpful, although it may have some utilityin refractory cases.The findings that some anticonvulsants have

minimal effects for mood disorders suggest that thenotion that agents that reduce kindling will neces-sarily be beneficial in manic-depressive ordersneeds to be reviewed. One possibility is that agents

with selective effects on limbic systems will befound to be useful, whereas others will not. Analternative, however, is that these differences inefficacy may be parsimoniously explained in termsof differential functional effects of lithium, carb-amazepine, valproate, valpromide and lamotrigine.While slow to emerge, the notion of mood-

stabilization has all but replaced the earlier notionof prophylaxis. If we ask whether any of the newermood-stabilizing agents can be demonstrated to betruly prophylactic, we reach the paradox at theheart of the mood-stabilization debate. Mood-stabilizers are agents, which ideally would showprophylactic effects without evidence of benefits inthe acute state. However, current �mood-stabilizers�are only on the market because of demonstrablebenefits in acute states.This sets up a number of paradoxes. Antipsych-

otics and antidepressants demonstrably producetreatment effects in the depressive and manic polesof bipolar disorder. Chlorpromazine was first usedin the management of mania and neuroleptics havebeen the standard agents for the management ofmanic states ever since. If by an antidepressant ismeant an agent that demonstrates a treatmenteffect in a trial with depressed patients, then mostneuroleptics are antidepressants (41), although itshould be noted that despite this evidence of short-term effects, few clinicians would regard theseagents as antidepressants in the longer run. Thesefindings in fact may do more to demonstrate thepitfalls of short-term trials than anything else. Thefunctional effects that these agents produce haveface validity as therapeutic principles in the man-agement of both depressive and manic states.Indeed ironically, while antidepressants may

cause manic reactions, one of the only controlledtrials done of imipramine in mania demonstratedthat it had beneficial effects in some patients (39). Itwas a consideration of results such as these in factthat led Schou to the concept of a mood normalizerin 1963.

Mood or psyche stabilizers?

The dominant conception of a mood-stabilizer atpresent appears to be that such a drug attacks aspecific underlying physiological abnormality with-out necessarily producing any obvious functionaleffect. The implication is that all mood-stabilizersare in some way modifying the same mechanism.Secondary messengers appear to be the favoritetarget at present, but there are no common specificeffects reported to date.While this conception can draw on histor-

ical notions about the specificity of lithium, the

Mood-stabilizers: the archeology of the concept

449

subsequent history of lithium as well as thediscoveries of the first psychotropic properties ofvalpromide and carbamazepine point to a need fora new term such as a psycho-stabilizer. Thestandard model is now a nosolytic one, in whichbenefits are specific and nosolytic. A psycho-stabilizer, in contrast, would produce a serenic,sedative or anti-irritability effect that would havedemonstrable benefits across a range of syndromes.The literature on lithium culminating in Sheard’s

1971 trial now suggests that far from being specificfor manic-depressive illness lithium may be anagent that among other things reduces the sensi-tivity to events in the environment so that thedisruptive impact of these events on internal moodstates is minimized. Such an effect has a clearfunctional utility that conceivably could producebenefits across a range of psychosyndromes, otherthan manic-depressive illness. Carbamazepine andvalproate appear to produce somewhat differentbut broadly serenic effects. For this alternativemodel to attract support it would be necessary tospecify the differences between these agents andlithium in sufficient detail to account for theconventional clinical wisdom and trial evidencethat carbamazepine and valproate may be moreuseful than lithium for mixed mood disorders andless beneficial in classic manic-depressive illness.This latter formulation of course stems from aperiod when lithium was viewed of as being all butspecific to manic-depressive illness in a way thatwas never the case for carbamazepine or valproate.Some specification of differential functional

effects is possible. Valpromide and valproate werediscovered initially through their use in mania andbecause of their particular �sedative� properties.Sedation is a therapeutic principle that makes sensein the management of manic states. Lamotrigine incontrast appears to be more effective in the depres-sive poles ofmanic-depressive disorder and this is anagent that far from being sedating is more likely tobe described in terms of its euphoriant properties(42). There is some basis therefore for arguing thatlithium, carbamazepine, valproate and lamotrigineall have functional effects that have face validity interms of managing various phases of manic-depres-sive disorders. The implication of this formulationhowever is that while these agents are now thoughtof as being a homogenous group, theymay in fact bequite diverse agents all of which have a certain utilitywhen used judiciously in manic depressive states.

Psychotropic utility and psychotropic efficacy

Recently two further conceptual issues have beenraised in the domain of mood disorder therapeutics.

First, in addition to the possible deleterious effectsof antidepressants on bipolar mood disorders, thepossibility has also been raised that antidepressantsmay have equally problematic effects in the unipo-lar domain. Fava in particular has argued thatantidepressants while efficacious in resolving acutedisorders may in fact lead to further episodes by asensitization process (43).A related area of interest in the treatment of

unipolar disorders has lain in the notion thattreatment should aim at restoring well-being ratherthan simply ameliorating the main features ofacute episodes. This domain links to the issueof relapse to antidepressants in that the existence ofsubclinical or residual symptoms is the biggestsingle predictor of future relapse (44).In a study that bears on both these points

Tranter et al. (45) have recently provided evidencethat subjects may be constitutionally predisposedto respond optimally to agents selective to partic-ular systems and that these agents have distinctivefunctional effects. The implication of these data isthat individuals may respond less well if at all toagents acting primarily on the wrong system forthem. Such sub-optimal responses can be expectedto be more likely to lead to further illness episodesthan would optimal responses. There is no reasonto believe that similar considerations will not alsoapply to the mood-stabilizers.

Perspectives for the future

A number of consequences stem from the aboveformulation. It has proven all but impossible todemonstrate prophylactic efficacy for agents, otherthan perhaps lithium, in the case of manic-depres-sive disorders. A proper trial demonstrating sucheffects would run for many years and woulddemonstrate a reduced frequency of episodes, anincrease in the inter-illness interval compared withplacebo and would also demonstrate that theseeffects outweigh any disruption produced by with-drawal syndromes on discontinuation. In practiceit has proved impossible to sustain a seriously illpatient group in such a trial.However another method of evaluating treat-

ments opens up if the focus switches to their moreimmediate functional effects. If patients on any ofthese agents identify a specifically useful effectproduced by that agent, this would de factoproduce a rationale for continuing treatment withthat agent in that particular person. Trials couldconceivably compare outcomes in patient groupswho could identify beneficial functional effectscompared with those who could not do so. Theultimate benefits of such an approach however will

Harris et al.

450

depend critically on practitioners identifying useful�psyche stabilizing� effects in an individual andproceeding with treatment on this basis rather thantreating on the basis that the agent has supposedlybeen demonstrated to be a mood stabilizer andtherefore will be beneficial regardless of whetherthe taker describes beneficial functional effects.In summary, there would seem to be few good

experimental or theoretical grounds to think that�mood-stabilizers� target the underlying deficit inmood disorders any better than the previousgeneration of antidepressants or neuroleptics. Inother words, the response of patients with mood-disorders to anticonvulsants is simply not the sameas the response of epileptic patients to anticonvul-sants. But there are abundant clinical and histor-ical grounds to think that �mood-stabilizers� havecomparatively unexplored psychotropic effects thatmight well account for their apparent benefits inboth acute and non-acute phases of bipolar disor-ders. Work is needed to specify these effects and tomatch therapeutic effects to patient types.

References

1. Ghaemi SN, Sachs GS, Chiou A, Pandurangi AK,Goodwin FK. Is bipolar disorder still underdiagnosed?Are antidepressants overutilized? J Affect Disord 1999; 52:135–144.

2. Ghaemi SN. On defining �mood stabilizer�. Bipolar Disord2001; 3: 154–158.

3. Ghaemi SN, Lenox MS, Baldessarini RJ. Effectiveness andsafety of long-term antidepressant treatment in bipolardisorder. J Clin Psychiatry 2001; 62: 565–569.

4. Baastrup PC, Schou M. Lithium as a prophylactic agent:its effect against recurrent depressions and manic-depres-sive psychosis. Arch Gen Psychiatry 1967; 16: 162–172.

5. Baastrup PC, Schou M. Prophylactic lithium. Lancet 1968;1: 1419–1422 & II, 349–350.

6. Baastrup PC, Poulsen JC, Schou M, Thomsen K, AmdisenA. Prophylactic lithium: double-blind discontinuation inmanic-depressive and recurrent depressive disorders. Lan-cet 1970; II: 326–330.

7. Angst J, Weiss P, Grof P, Baastrup PC, Schou M. Lithiumprophylaxis in recurrent affective disorder. Br J Psychiatry1970; 116: 604–614.

8. Coppen A, Noguera R, Bailey J et al . Prophylactic lithiumin affective disorders. Controlled trial. Lancet 1971; ii:275–279.

9. Meijer JW, Meinardi H, Binnie CD. The developmentof antiepileptic drugs. In: Parnham MJ, Bruinvels J eds.Discoveries in Pharmacology Vol 1. Amsterdam: Elsevier,1983.

10. Carraz G, Lebreton S, Boitard M, Borselli S, Bonnin J.A propos de deux nouveaux anti-epileptiques de la serien-dipropylacetique. Encephale 1965; 458–465.

11. Lambert P-A, Borselli S, Marcou G, Bouchardy M, CarrazG. Proprietes neuro-psychotropes du Depamide: actionpsychique chez les epileptiques et les malades presentantdes troubles caracteriels. CR Congr de Psychiatrie et deNeurologie de langue Francaise, Masson, Paris 1966;1034–1039.

12. Lambert P-A, Borselli S, Midenet J, Baudrand C, MarcouG, Bouchardy M. L’a�ction favorable du Depamide surl’evolution a long terme des psychoses maniaco-depres-sives. CR Congr de Psychiatrie et de Neurologie de langueFrancaise, Masson, Paris 1968: 489–495.

13. Comite Lyonnais de Recherches Therapeutiques en Psy-chiatrie. The birth of psychopharmacotherapy: explorationsin a new world – 1952–1968. In: D. Healy ed. The Psycho-pharmacologists Vol 3. London: Arnold, 2000: 1–54.

14. Lambert P-A, Carraz G, Borselli S, Bouchardy M. Ledipropylacetamide dans le traitement de la psychosemaniaco-depressive. L’Encephale 1975: 25–31.

15. Lambert P-A. Acute and prophylactic therapies of patientswith affective disorders using valpromide. In: Emrich HM,Okuma T, Muller AA, eds. Anticonvulsants in AffectiveDisorders. Amsterdam: Elsevier, 1984: 33–44.

16. Emrich HM, von Zerssen D, Kissling W, Moller H-J,Windorfer A. Effect of sodium valproate on mania:GABA-hypothesis of affective disorders. Arch PsychiatrNervenkr 1980; 229: 1–16.

17. Healy D. The Creation of Psychopharmacology. Cam-bridge Mass: Harvard University Press, 2002.

18. Takezaki H, Hanaoka M. The use of carbamazepine(Tegretol) in the control of manic-depressive psychosis andother manic-depressive states. Seishin Igaku 1971; 13: 173–183.

19. Okuma T. The discovery of the psychotropic effects ofcarbamazepine. In: Healy D ed. The Psychopharmacolo-gists Vol 3. London: Arnold, 2000: 259–280.

20. Okuma T, Kishimoto A, Inoue K et al. Anti-manic andprophylactic effects of carbamazepine (Tegretol) on manic-depressive psychosis – A preliminary report. Folia Psy-chiatr Neurol Japonica 1973; 27: 283–297.

21. Okuma T, Kishimoto A, Inoue K et al. Anti-manic andprophylactic effects of carbamazepine (Tegretol) on manicdepressive psychosis. Seishin Igaku 1975; 17: 617–630.

22. Okuma T, Kishimoto A. Anti-manic and prophylactic ef-fects of Tegretol. In: Abstracts of the 6th World Congressof Psychiatry. Honolulu, Ciba-Geigy, Summit, NJ, 1977.

23. Okuma T, Inanaga K, Otsuki S. Comparison of the anti-manic efficacy of carbamazepine and chlorpromazine: adouble blind controlled study. Psychopharmacology (Berl)1979; 66: 211–217.

24. Takahashi R, Sakuma A, Itoh H. Comparison of theantimanic efficacy of lithium carbonate and chlorproma-zine in mania. Report of collaborative study group ontreatment of mania in Japan. Arch Gen Psychiatry 1975;32: 1310–1318.

25. Okuma T, Yamashita I, Takahashi R. Comparison of theantimanic efficacy of carbamazepine and lithium carbonateby double-blind controlled study. Pharmacopsychiatry1990; 23: 143–150.

26. Lewin J, Sumners D. Successful treatment of episodicdyscontrol with carbamazepine. Br J Psychiatry 1992; 161:261–262.

27. Sheard MH. Effect of lithium on human aggression. Nat-ure 1971; 230: 113–114.

28. Tupin JP, Smith DB, Clanon TL, Kim LI, Nugent A,Groupe A. The long-term use of lithium in aggressiveprisoners. Comp Psychiatry 1973; 14: 311–317.

29. Zarate CA, Tohen M, Baraibar G, Zarate SB, BaldessariniRJ. Shifts in hospital diagnostic frequencies: bipolar disor-der subtypes, 1981-1993. J Affect Disord 1997; 43: 79–84.

30. Stoll AL, Tohen M, Baldessarini RJ et al. Shifts in diag-nostic frequencies of schizophrenia and major affectivedisorders at six North American psychiatric hospitals,1972-1988. Am J Psychiatry 1993; 150: 1668–1673.

Mood-stabilizers: the archeology of the concept

451

31. Loomer A, Conners CK. Diphenylhydantoin in childrenwith severe temper tantrums. Arch Gen Psychiatry 1970;23: 80–89.

32. Conners CK, Kramer R, Rothschild GH, Schwartz L,Stone A. Treatment of young delinquent boys with diph-enylhydantoin sodium and methylphenidate. A controlledcomparison. Arch Gen Psychiatry 1971; 24: 156–160.

33. Raptis C, Garcia Boruguero D, Weber MM, Dose M,Bremer D, Emrich HM. Anticonvulsants as adjuncts for theneuroleptic treatment of schizophrenic psychoses; a clinicalstudy of beclamide. Acta Psychiatr Scand 1990; 81: 162–167.

34. Liu MC. Clinical experience with sulthiame. Br J Psychi-atry 1966; 112: 621–628.

35. Moffat WR, Siddiqui AR, MacKay DN. The use ofsulthiame with disturbed mentally subnormal patients. BrJ Psychiatry 1970; 117: 673–678.

36. Ballenger JC, Post RM. Carbamazepine in manic-depres-sive illness: A new treatment. Am J Psychiatry 1980; 137:782–790.

37. Silberman EK, Post RM, Nurnberger J, Theodore W,Boulenger AP. Transient sensory, cognitive and affectivephenomena in affective illness. Br J Psychiatry 1985; 146:81–89.

38. Post RM, Rubinow D, Ballenger JC. Conditioning andsensitization in the longitudinal course of affective illness.Br J Psychiatry 1986; 149: 191–201.

39. Schou M. Normothymotics, �mood-normalizers�: are lith-ium and imipramine drugs specific for affective disorders?Br J Psychiatry 1963; 109: 803–809.

40. Pande AC, Crockatt JG, Janney CA, Werth JL, TsarouchaG. Gabapentin in bipolar disorder: a placebo-controlledtrial of adjunctive therapy. Bipolar Disord 2000; 2: 249–55.

41. Robertson M, Trimble M. Major Tranquillisers used asAntidepressants. J Affect Disord 1982; 4: 173–193.

42. Calabrese JR, Bowden CL, Sachs GS, Ascher JA, Mona-ghan E, Rudd GD. A double-blind placebo-controlledstudy of lamotrigine monotherapy in outpatients withbipolar 1 depression. J Clin Psychiatry 1999; 60: 79–88.

43. Fava GA. Potential sensitizing effects of antidepressantsdrugs on depression. CNS Drugs 1999; 12: 247–256.

44. Fava GA. Subclinical symptoms in mood disorders:pathophysiological and therapeutic implications. PsycholMed 1999; 29: 47–61.

45. Tranter R, Healy H, Cattell D, Healy D. Functional var-iations in agents differentially selective to monoaminergicsystems. Psychol Med 2002; 32: 517–524.

Harris et al.

452