You are here: Controversies / Barry Blackwell: The lithium controversy: A historical autopsy / Comment by Paul Grof and Jules Angst
Friday, 24.03.2017

Comment by Paul Grof and Jules Angst


Reply to Barry Blackwell

Paul Grof and Jules Angst


Renaissance of interest

For quite a while, lithium treatment had fallen out of favor in the mainstream. Non- patentable and inexpensive, lithium could not compete with the skillful marketing of new profitable neuroleptics and antiepileptics and could not withstand other pressures exerted by the pharmaceutical industry. The finest example was the clever advertising of divalproex which, despite the absence of evidence for stabilizing patients, quickly became the best selling drug for bipolar disorder in the United States. But recently, a renaissance of interest in the use of lithium treatment has unexpectedly emerged.

Several motives may be converging here. Lithium’s rather unique antisuicidal properties, proven for some time (1) have recently been widely publicized. In neuroscience laboratories lithium has turned out neuroprotective (2, 3) and it might even become helpful in the management of several obstinate neurological and geriatric disorders (3). More important clinically, the re-evaluation of atypical neuroleptics in the treatment of bipolar disorders has lately curved sour. Furthermore, voices have now arisen suggesting that lithium may actually be the only true mood stabilizer, as it demonstrably acts against both polarities of manic-depressive disorder (4).

Perhaps it was this resurgence of interest that led colleagues to ask independently Barry Blackwell and myself to address again the history of the lithium controversy. And as Barry Blackwell completed his interesting reminiscences (5) and invites comments on his version of the autopsy, I happily oblige.

It is in this context that I think it is useful to dissect the lithium controversy. I concur with Blackwell that we can learn from the past. In psychiatry we now live in an era of conceptual turmoil and absorbing lessons from our history has become critical. Each story has at least two ways of interpreting. With the passage of time our differences have softened and I agree with most of what Barry Blackwell says in general but still part with him on the weighing of the usefulness of long-term lithium treatment.


Preceding events

What was the controversy actually about? Let me first briefly sum up, from my perspective, the events that preceded the disagreement. In the 1950’s the maintenance treatment offered to manic-depressive patients used to be psychoanalysis and maintenance ECT (6). The former was unfortunately not helpful and the latter effective but not favored by patients. Lithium was initially used only for the management of acute mania.

Since 1956 however anecdotal observations started emerging about other possible benefits of lithium. Schou (7) reported an observation of a manic patient who subsequently stopped having both manic and depressive recurrences when he maintained lithium during the free intervals.  Beneficial action against depressions was also mentioned by Vojtechovsky (8). Hartigan (9) and Baastrup (10) similarly noted that patients maintained on lithium had a marked reduction of both types of recurrence.

Baastrup and Schou (11) then carried out a longitudinal study of patients with many previous episodes of illness. Patients with both bipolar and unipolar disorder were involved. The analyses indicated that recurrences occurred in patients significantly less frequently during lithium treatment than before such treatment, or even disappeared completely.  Schou, Angst and I then decided to collaborate and to use a “mirror-image” design, utilizing the ample information we had about the previous course of illness of these manic-depressive patients. Against marked editorial resistance, our joined prospective observations on 250 lithium-treated patients were eventually published in the British Journal of Psychiatry (12). Together with clinical reports published earlier, an ample body of similar observations was emerging and demonstrating lithium as a useful drug in the treatment of manic-depressive illness.

Opposition against such interpretation emerged quickly however and, among experts, views about the issue became sharply divided. Some psychiatrists expressed strong support for lithium prophylaxis, based on their own clinical experience. Others disagreed. On methodological grounds, Blackwell and Shepherd (13) concluded that the claims for prophylactic efficacy were just a myth, supported by faulty evidence. They raised several critical points; their main objections were, first, a bias due to the open, non-blind evaluation of the recurrences and, second, a statistical approach which in their opinion weighted the facts in favor of the hypothesis.

But the objections could be effectively counteracted only by a tightly designed double-blind evaluation. 


Barry Blackwells invaluable contribution

Before commenting on these methodological disagreements, I want to express my gratefulness to Barry Blackwell.  Even though his objections were incorrect, it was an invaluable service to moving ahead. Had it not been for his widely quoted procedural condemnation, I really wonder if lithium would now be in clinical practice, all over the world. The national regulatory bodies insist on double-blind tests. It was Blackwell's somewhat sarcastic, sharp, articulate arguments that made a strong impression and eventually forced the randomized double-blind trial. My feelings of gratefulness may not have been quite the same then but in hindsight they are strong, and I have expressed them repeatedly publicly.

When Mogens Schou, Jules Angst and I completed the replication study published in the British Journal of Psychiatry (12), the last thing on our minds was to switch any of these patients to placebo. Many of them had suffered from severe, frequently recurrent mood disorder, had been hospitalized numerous times and badly incapacitated by their illness.  On lithium they were stable for the first time and it seemed not only unethical but also unimaginable to ask them to stop it, in order to participate in a clinical trial with placebo. 

Before they went on lithium, I followed my patients for up to six years, failing to stop their depressions and manias. I could not imagine putting them and their families through the same misery again. In addition, in most of these patients the effect of lithium stabilization was so convincing, so different from the previous course of illness, that to use placebo just to prove that they again relapse appeared unethical and redundant. Furthermore, the Swiss and Czech findings were already an independent replication of the earlier Danish findings.

Finally, our analysis also indicated that the criticism aimed at our methodology was not correct. Barry Blackwell and Michael Shepherd raised two main methodological objections against the findings: that the marked recurrence reduction the patients experienced was to be expected naturally – that it was the result of a “regression to the mean” - and that the observations were not made blindly and thus biased by enthusiasm.

As for the duo’s first protestation, the patients experienced frequent episodes qualifying them to enter the trial. Blackwell and Shepherd felt that the less frequent episodes that followed were the result of the recurrence frequency regressing back to a mean of lower value. To demonstrate their point, they quoted Saran’s (14) data of 13 patients who entered the follow-up with frequent episodes but lost that frequency with the passage of time.

But the problem was that, because of the small number of patients, Saran’s example was neither representative nor applicable to the problem. Ottoson and Issakson (15) and Laurel and Ottosson (16) showed that the “mirror image” design is justified. In a sufficiently large sample of patients not receiving maintenance treatment the mean frequency of recurrences in the past becomes replicated in the future.

In essence, the individual clinical course of manic-depressive illness is capricious, overall seemingly random. Given this capriciousness, in a small group of patients the future frequency of recurrences will vary in any direction.  It may decrease - as it did for Saran’s 13 patients, it may increase, or it may remain about the same. But to obtain a predictable, anticipated mean frequency, one requires a sufficiently sizable cohort, as was the case in our open trials with a “mirror” design (17).

As for the Blackwell and Shepherd’s second objection – biased open assessment – blind evaluation is often very important but not a panacea.  As Schou later demonstrated (18), the results from long-term clinical trials of lithium were well comparable, regardless of whether the evaluations were carried out blindly or openly. Obviously, if one were evaluating the effects of an anxiolytic in neurotic patients, the placebo effect and bias would usually play a huge role. Double-blind arrangement would be indispensable. Blackwell illustrated clearly the relationship between observer enthusiasm and treatment outcome earlier with MAOI inhibitors (32).

But in a maintenance treatment of manic-depressive patients the task is markedly different: to assess if a patient who was previously symptom-free, develops in a free interval an acute manic or depressive episode.  If in this task there would be large, systematic discrepancies between different psychiatrists with a similar training, we could forget psychiatry altogether.

Parenthetically, the biases of the involved investigators were actually markedly different, and not all positive. Schou and Baastrup were openly enthusiastic, because of their previous promising observations. Jules Angst appeared curious but neutral as to the expected outcome.  And my previous attempts to prevent the recurrences of manic-depressive illness were so dismal (19, 20), that I did not believe anything could work preventatively. As I wrote earlier, I was hoping to prove Schou wrong.  Yet despite our different preconceptions, our results with long-term lithium treatment were comparable.

Lithium’s efficacy was subsequently proven in a trial (21) that employed the design with blind evaluation and randomization. As to the ethical concerns, using sequential analysis minimized the number of patients receiving placebo. Using sequential analysis can markedly reduce the number of patients needed to reach a statistically significant difference by utilizing, in addition, the probability hidden in the sequence in which the observations come in.

In this manner it became possible to complete the double-blind trial within six months and with a minimum of patients having been given placebo. All of the patients who became ill again were those switched to placebo, none of the lithium patients experienced recurrences during the same time.


Methodology in diapers

I fully concur with Barry Blackwell that one of the main reasons for our disagreements was the fact that the methodology of maintenance trials in bipolar disorders was in diapers then. In fact we were developing the methodology while proceeding with the studies (22).

As I read Barry Blackwell's “autopsy”, I felt there were good reasons why we could not and cannot see lithium treatment in quite the same light. Our background, professional careers, experience and interests were different. As I understand it, his central interests were clinical trials, psychopharmacology and pharmacology. He was frustrated by many methodologically inadequate studies in the past and did not want to see another shabby study confusing psychiatrists. And, after his critique of lithium, he moved on to the industry and then academic and clinical practice. He seemed more interested in anxiety states than in following manic-depressive patients (5).

We, on the other hand, prior to lithium studied the natural course of manic-depressive illness in hundreds of patients (23, 24). Since the heated lithium debate I have treated more than thousand patients with lithium, some of them up to 40 years, and researched who does respond. With experience being so different, even now Barry Blackwell and mine evaluation of lithium cannot be the same.


The efficacy of lithium is neither a myth nor imipramine-like

Barry Blackwell feels that, after the initial trials, uncertainty about Lithium’s efficacy lingered until later studies published in 1984. He singles out Prien et al. (25), a double-blind trial carried out mainly in the US VA hospitals. The results are interpreted as indicating that in bipolar patients imipramine is better in cases of mild depressions and lithium in more severe cases. To claim that the efficacy of lithium is comparable to imipramine requires disregarding fully the rest of the published evidence. Such assertion seems to me idiosyncratic, neglecting the existing regulatory decisions, numerous clinical trials and expert consensus.

There is a body of double-blind clinical investigations together demonstrating prophylactic efficacy of lithium both against manias and depressions (18, 26); trials that have dealt well with Blackwell's methodological objections. On this basis, by the early 1970s, lithium was approved for long-term treatment in most Western countries, by regulatory agencies requiring solid double-blind evidence. Expert committees that have produced more than 25 guidelines for the treatment of bipolar disorder now quote lithium trials as the best (class I) evidence for efficacy. Despite Prien’s study, imipramine is nowhere recommended for long-term treatment of bipolar disorders.

Prien’s findings only can be interpreted if they are placed in the context of what had happened between 1968 and 1984 with diagnosing mood disorders. Manic-depressive illness was transforming into a much larger and more heterogeneous “bipolar spectrum disorders”. As lithium treatment had a striking success in patients with typical manic-depressive illness, the diagnostic fashion for mood disorders broadened (27, 28), and in “bipolar spectrum” disorders many patients with mood-incongruent symptoms and multiple comorbidities were included. In addition to manic-depressive illness the experimenting with lithium also expanded to other indications: schizoaffective conditions, cycloid psychoses, aggressive states, alcoholism, potentiation of antidepressants, and several other situations.

But lithium prophylaxis is the treatment of choice only for what used to be “manic-depressive illness”: in essence, remitting, episodically recurring bipolar and unipolar disorders. It may also be of partial help in other conditions but the effect is quantitatively and qualitatively different: for example one will see low efficacy and intense rebound after discontinuation. The diagnosis of manic-depressive illness used to require, among others, the exclusion of mood incongruent psychotic symptoms, the exclusion of other psychiatric diagnoses (i.e. exclusion of comorbidity) and the presence of episodic course.

This development created a very interesting situation. Recent studies have shown that bipolar disorder is now often underdiagnosed, particularly in recurrent depression. At the same time, as bipolar diagnosis is now given simply on the basis of a symptom set, without further analysis and exclusions, it is also grossly overused instead of other diagnoses. As a result the bipolar spectrum disorder has become very fashionable and highly prevalent, but the classical lithium responsive manic-depressive patients are only a minority subgroup.

For whatever it's worth, while the Prien et al. study was going on, two experienced American colleagues who knew my interest in lithium and participated in the study contacted me. They were very critical of the patient selection, partly due to the population of VA hospitals, and warned me not to believe the findings once the study is completed. In hindsight, the Prien et al. study can hardly be considered the main pillar for the evaluation of lithium's usefulness.


Wide Acceptance and Narrow Opposition

Lithium treatment for bipolar disorder has gradually been accepted in most countries of the world, including the Third World countries. Lithium is now available as an effective mood stabilizer worldwide but its use is geographically uneven. It should help in the Third World that lithium is inexpensive, particularly in comparison with new putative stabilizers.

But repeated questioning lithium’s efficacy does happen and comes particularly from those who had been using lithium outside of the established evidence and in naturalistic studies with looser diagnosing and monitoring. But careful analyses have shown that lithium remains effective for patients with clinical profile for which it was proven effective in the first place (29).

Despite overwhelming evidence of the efficacy in typical manic-depressive cases, continuing debates about lithium are likely to occur between opponents who incorrectly believe that they are discussing the same issue but have used lithium in other bipolar types. Unfortunately, the correct evaluation of the outcome of stabilizing treatment in recurrent mood disorders is much more challenging than one would assume. Capricious course, fluctuating compliance with medication, and a varying speed of stabilization all make it difficult to evaluate the relationship between the medication and a changed course of illness in any individual patient. Bipolar disorders have distinct subtypes responding preferentially to different mood stabilizers, and lithium offers a variety of markedly different benefits to patients outside the classical manic-depressive illness (30,31).



I do not know if Barry Blackwell really believes - as he seems to indicate in his writing – that the effect of lithium treatment on bipolar disorders is indeed comparable to the effect of imipramine, or whether this is just another expression of his mastery of hyperbole. But we certainly do approach this issue from different angles.

When I think of lithium stabilization, my thinking is unavoidably colored by my experience of treating many bipolar patients for more than five decades. Before lithium treatment we lost every year several patients to suicide and the life of those who continued living was marred by their illness: the impact of frequent episodes of manias, depressions and hospitalizations and the influence on their families and professional life. Since we have been using lithium, the situation changed dramatically and these problems have been minimized, and often completely eradicated.

For Mogens Schou the initial heated debates were stressful. He was a very compassionate physician and switching stabilized patients from lithium to placebo troubled him greatly. He was also an extremely meticulous researcher.  The possibility raised by Blackwell and Shepherd that he may have overlooked something important in methodology bothered him very much. He was extremely careful, as he kept moving between his laboratory and his clinical investigations.

The accusation of biased observation was not easy for him to swallow.  There was some irony in blaming him for not having carried out the observations double-blind as he, in fact, performed the first double-blind study in psychopharmacology fifteen years earlier.  Particularly unfortunate was, I thought, Michael Shepherd’s criticism ad hominem: he repeatedly stressed that Mogens Schou was a biased enthusiast because Mogens’ brother’s depressions responded well to lithium and, replying to questions, he never publicly conceded that lithium works.

From what Barry Blackwell has written about his professional life (5), his professional interests have been different than ours and he worked more along different lines. His critique of lithium was an important but a relatively short-lived involvement and reflected more his interest in methodology and history of clinical trials than in the treatment of bipolar patients. Thus I may be biased in favor of lithium but from his text I tend to conclude that he underestimates the helpfulness of lithium treatment and oversimplifies its use in bipolar disorders. Nevertheless, as I mentioned, in hindsight I see enormous value of his critique during the early days of lithium’s clinical trials.


Impact of Lithium Treatment on Psychiatry

Up until 1967 no medication had seemed capable of averting recurrences of affective disorders; therefore only acute episodes had been treated.  The introduction of long-term lithium treatment, lithium prophylaxis, changed things radically. From a practical point of view it was primarily lithium’s ability to prevent recurrences that made an impression. For research the introduction of lithium was a major stimulus for neurobiology, demonstrating that a simple element can produce major neurobiological changes. Lithium became the focus of attention of pharmacologists, biochemists, physiologists, psychiatrists, psychologists, and many others. It was probably the advent of lithium therapy that made psychiatric research truly interdisciplinary. Research on all aspects of the affective disorders has been greatly stimulated by the demonstration of the effectiveness of lithium in the treatment of these conditions.

For academic psychiatry the acceptance of lithium treatment led to the important recognition that mood disorders are much more common than previously presumed, and that the existing classification systems must be reconsidered. As the history of the past four decades has shown, lithium therapy has made a significant contribution to modern psychiatry, both in relation to its specific uses in alleviating recurrent endogenous affective disorders, and in stimulating psychiatric research and conceptual thinking.


Paul Grof

1 November 2014




1.  Mueller-Oerlinghausen B, Ahrens B, Volk J, Grof P, Grof E, Schou M, Vestergaard P, Lenz G, Simhandl C, Thau K, et al. Reduced mortality of manic-depressive patients in long-term lithium treatment: an international collaborative study by IGSLI. Psychiatry Res. 1991;36:329-331.

2.  Hajek T, Cullis J, Novak T, Kopecek M, Hoschl C, Blagdon R, O'Donovan C, Bauer M, Young LT, Macqueen G, Alda M. Hippocampal volumes in bipolar disorders: opposing effects of illness burden and lithium treatment. Bipolar Disord. 2012;14:261-270.

3.  Quiroz J.A.C.; Drevets, W.C.; Henter, I.D.; Manji, H.K.: Mood Disorders (Chapter 3). in Translational Neuroscience. Edited by Barrett JEC, J.T.;Williams, M, Cambridge University Press; 2012. pp. 27-69.

4.  Grof P, Mueller-Oerlinghausen B. A critical appraisal of lithium's efficacy and effectiveness: the last 60 years. Bipolar Disorders. 2009;11:10-19.

5.  Blackwell B: The Lithium Controversy: An Historical Autopsy. in International Network for The History of Neuropsychopharmacology, Controversies. INHN June 2014.

6.  Geoghegan JS, GH;. rophylactic electroshock. Am J Psychiatry. 1949;105:494-496.

7.  Schou M. Litiumterapi ved mani: Praktiske retningslinier,. Nord Med. 1956;55:790-794.

8.  Vojtechovsky M: Zkusenosti s lecbou solemi lithia. Problemy Psychiatrie v Praxi a ve Vyzkumu. Prague, Czechoslovak Medical Press,; 1957. pp. 216-224.

9.  Hartigan GP. The use of lithium salts in affective disorders. Brit J Psychiat. 1963;109:810-814.

10.  Baastrup PC. The use of lithium in manic-depressive psychosis. Compr Psychiatry. 1964;5:396-408.

11.  Baastrup PC, Schou M. Lithium as a prophylactic agent: its effect against recurrent depressions and manic-depressive psychosis. Archives of General Psychiatry. 1967;16:162-172.

12.  Angst J, Weis P, Grof P, Baastrup PC, Schou M. Lithium prophylaxis in recurrent affective disorders. Br J Psychiatry. 1970;116:604-614.

13.  Blackwell B, Shepherd M. Prophylactic lithium: Another therapeutic myth? An examination of the evidence to date. The Lancet. 1968;I:968-971.

14.  Saran BM. Prophylactic lithium? Lancet1968 Aug 3;2(7562):284-5.:284-285.

15.  Isaksson A, Ottosson JO, Perris C: Methodologische Aspekte der forschung uber prophylaktische Behandlung bei affektiven Psychosen. in Das Depressive Syndrom. Edited by Hippius H, Selbach H. Munchen, Urban and Schwarzenberg; 1969. pp. 561-574.

16.  Laurell B, Ottosson JO. Prophylactic Lithium? The Lancet. 1968;2:1245.

17.  Grof P. Designing long-term clinical trials in affective disorders. J Affect Disord. 1994;30:243-255.

18.  Schou M: Phases in the development of lithium treatment in psychiatry. in The Neurosciences: Paths of Discovery II. Edited by Samson F AG, eds. Boston, Basel, Berlin., (Birkhäuser: Boston, Basel, Berlin. 1992) 148-166.; 1992. pp. 148-166.

19.  Grof P, Vinao O. Maintenance and prophylactic imipramine doses in recurrent depressions. Activitas Nervosa Superior. 1966;8:384-385.

20.  Grof PV, O.: Comparison of various prophylactic procedures in affective psychoses. in Das Depressive Syndrom. Edited by Hippius H SHe. Berlin1969. pp. 101-104.

21.  Baastrup PC, Poulsen JC, Schou M. Prophylactic lithium: Double blind discontinuation in manic-depressive and recurrent-depressive disorders. The Lancet. 1970;2:326.

22.  Grof P, Schou M, Angst J, Baastrup PC, Weis P. Methodological problems of prophylactic trials in recurrent affective disorders. Br J Psychiatry. 1970;116:599-603.

23.  Angst J BP, ; Grof P, ; Hippius H, ; Poldinger W, ; Weiss P. Clinical course of affective disorders. Psychiatry. 1969;76:489-500.

24.  Angst J, Dittrich A, Grof P. Course of endogenous affective psychoses and its modification by prophylactic administration of imipramine and lithium. IntPharmacopsychiat. 1969;2:1-11.

25.  Prien R, Kupfer DJ, Mansky PA, Small JG, Tuason VB, Voss CB, Johnson WE. Drug therapy in the prevention of recurrences in unipolar and bipolar affective disorders: report of the NIMH collaborative study group comparing lithium carbonate, imipramine, and a lithium carbonate-imipramine combination. Archives of General Psychiatry. 1984;41:1096-1104.

26.  Coppen A, Peet M, Bailey J, Noguera R, Burns BH, Swani MS, Maggs R, Gardner R. Double-blind and open studies of lithium prophylaxis in affective disorders. Psychiatr Neurol Neurosurg. 1973;76:501-510.

27.  Baldessarini RJ. Frequency of diagnoses of schizophrenia versus affective disorders from 1944 to 1968. American Journal of Psychiatry. 1970;127:59-63.

28.  Grof P, Fox D. Admission rates and lithium therapy (letter to the editor). British Journal of Psychiatry. 1987;150:264-265.

29.  Berghöfer A, Alda M, Adli M, Baethge C, Bauer M, Bschor T, Grof P, Müller-

           Oerlinghausen B, Rybakowski JK, Suwalska A, Pfennig A.

           Stability of lithium treatment in bipolar disorder - long-term follow-up of 346 patients. Int J Bipolar Disord. 2013 Jul 31;1:11.

30.  Grof P. Has the effectiveness of lithium changed? Impact of the variety of lithium's effects. Neuropsychopharmacology. 1998;19:183-188.

31.  Grof P. Selecting effective long-term treatment for bipolar patients: monotherapy and combinations. J Clin Psychiatry. 2003;64 Suppl 5:53-61.

32.Blackwell B, Taylor D. An Operational Evaluation of MAOI. Proceedings Royal  soc.   

            Med. 1967.



Studies on the long-term natural history of mood disorders

Knowledge of the course of mood disorders is essential when deciding whether long-term prophylactic medication is justified. This is especially the case if the natural history of a disorder shows not spontaneous improvement but rather persistent recurrence or even an increase of episodes, reflected by shortening cycle lengths. A cycle is defined as an episode plus the subsequent interval, i.e. the time between the onset of two subsequent episodes.

In 1967 Angst and Weis [1], investigating a group of 375 subsequent hospital admissions of patients with mood disorders in Zurich (Switzerland), found a log-normal distribution of episode and cycle lengths in four subgroups (125 recurrent depression, 117 involutional melancholia, 45 bipolar and 85 schizo-affective psychoses). This signified that studies should no longer base on arithmetic means. More important, however, was that the longitudinal analysis of cycle lengths showed a clear acceleration of recurrences with an increasing number of episodes. This was most marked in patients with bipolar disorder, followed by schizo-affective disorder and was lowest in patients with depressive disorders.

These findings were reproduced in 386 patients from a further three centres (Basle, Berlin, Landeck) by Angst, Grof, Hippius and Weis in 1968 [2]. In 701 patients with bipolar disorder and 988 patients with recurrent major depression a progressive shortening of cycles correlated with age at onset, age, and number of episodes over 20 years (the subsequent cycle was about 10% shorter than the previous one. Despite the clear finding of an increasing recurrence risk by shortening of cycles, in our statistical testing of the long-term effect of prophylactic treatment we applied the very conservative mirror model, which assumes, as the zero hypothesis, merely an equal occurrence of the number of episodes before and under treatment during identical, individual observation periods.


Studies on prophylactic treatment of mood disorders with imipramine and lithium.

A first study by Angst, Dittrich and Grof 1969 [3] dealt with patients treated with imipramine (N=63) or lithium (N=91) in Prague and Zurich. Statistically the mirror model was applied with Wilcoxon signed rank tests. Under imipramine there was a significant deterioration in the course of depression during the second intra-individual period, whereas lithium showed a positive effect in bipolar disorders (p<.025) and a trend to an effect in recurrent depression or involutional melancholia (p<.07). The negative and positive effects of the two drugs were comparable and significant in both samples (Prague and Zurich). This was the first statistically-based demonstration of the efficacy of lithium in treating mood disorders.

In a larger analysis of lithium data undertaken in Glostrup (DK), Prague and Zurich (Angst, Weis, Grof, Baastrup and Schou 1970 [4]) equal observation periods (before and under treatment) were again compared with regard to hospital admissions and number of episodes before and during lithium prophylaxis. In all three centres there was a reproducible significant decrease in the number of episodes (p<.001) during the lithium period. Taking all 244 patients together, there were significantly fewer hospital admissions for patients with manic-depressive disorders (p<.001), recurrent depression (p<.002) and schizo-affective disorders (p<.01). The average observation periods of the three groups compared were 2x38.5 months for bipolar disorder, 2x26.7 months for recurrent depression and 2x28.1 months for schizo-affective psychoses.

Thus in contrast to the repeatedly confirmed deterioration of the spontaneous course of mood disorders, a significant improvement was achieved with lithium but not so with imipramine.


Personal reminiscences

Ervin Varga from Budapest and I worked under Michael Shepherd for a few months, researching 981 records of patients treated for depression in the Maudsley Hospital. At that time my monograph showing the differences between unipolar depression, bipolar disorder and schizo-affective disorder (published in 1966) had already been accepted for print. Bleuler, Stroemgren and Aubrey Lewis agreed that the results could not be true, but fortunately Eliot Slater believed in the correctness of the findings. Stroemgren and Bleuler changed their minds after Perris published similar results. Michael Shepherd had by self-admission never treated a patient with lithium and I would be interested to know whether Blackwell himself had done so at the time of their joint article in the Lancet.

During these years we held annual meetings of the IGSAD (International group of studies of affective disorders, founded by Ottosson, Perris, Winokur and Angst). I invited Michael Shepherd to join the group. Some members (Christian Baastrup, Mogens Schou, Max Hamilton, Martin Roth, Paul Grof, Jules Angst etc.) discussed the ethical and feasibility problems of a placebo controlled study on lithium and decided finally on the design of a cessation study carried out in Glostrup, which was published by Baastrup et al. 1970 [5].

The positive results of this study were first presented to an IGSAD meeting and were the subject of intense debate. Michael Shepherd remained silent during the discussion and when asked for his opinion, replied "no comment". At the end of that year, I asked him to retire from the group, which he did; we remained on very friendly terms for the rest of his life.


Jules Angst

15 November 2014



1.         Angst J, Weis P (1967) Periodicity of depressive psychoses. In: Brill H, Cole JO, Deniker P, Hippius H, Bradley PB (eds) Neuropsychopharmacology. Proceedings of the Fifth International Congress of the Collegium Internationale Neuropsychopharmacologicum. Washington D.C. 1966, 28-31 March 1966 1966. International Congress Series No.129. Excerpta Medica Foundation, pp 703-710

2.         Angst J, Grof P, Hippius H, Poeldinger W, Weis P La psychose maniaco-dépressive est-elle périodique ou intermittente? In: De Ajuriaguerra J (ed) Cycles biologiques et psychiatrie. Paris, 1968. Masson, pp 339-351

3.         Angst J, Dittrich A, Grof P (1969) Course of endogenous affective psychoses and its modification by prophylactic administration of Imipramine and Lithium. Int Pharmacopsychiat 2:1-11

4.         Angst J, Weis P, Grof P, Baastrup PC, Schou M (1970) Lithium prophylaxis in recurrent affective disorders. Br J Psychiatry 116:604-614

5.         Baastrup PC, Poulsen JC, Schou M, Thomsen K, Amdisen A (1970) Prophylactic lithium: Double-blind discontinuation in manic- depressive and recurrent-depressive disorders. Lancet 2:326-330


Paul Grof and Jules Angst

January 22, 2015