Sunday, 26.03.2017

Reply to Paul Grof and Jules Angst

Reply to Paul Grof and Jules Angst
By Barry Blackwell 

I thank Paul Grof for the kindness and generosity of his comments and must confirm what both he and Jules Angst suggest was my youthful inexperience at the time of the controversial “Prophylactic Lithium” article in the Lancet co-authored with Michael Shepherd. Much of my residency training (1962-1967) was preoccupied with human and pharmacology research on the interaction of MAO inhibitors and tyramine containing foods so I had virtually no practical experience with lithium.

It is also accurate that different career patterns have colored our opinions of the research and its practical implications. Both Drs. Grof and Angst have devoted significant portions of their careers to sophisticated research on the natural history and drug treatment of the bipolar affective disorders in large populations of patients. My own career trajectory has been entirely different , devoted to a wide spectrum of interests in pharmacology, psychosomatic medicine, medical education  and specific  topics such as patient compliance, homelessness, chronic pain, physician career development and administration of two academic departments . As a result my continuing interest and knowledge in the arena of bipolar disorder became that of a journeyman (albeit academic) psychiatrist with only a modest involvement in everyday clinical practice and a patchy knowledge of the evolving literature.

From my personal experience and those of colleagues I did learn how the depressive component of bipolar disorders often persists in subdued form despite lithium and is difficult to treat with imipramine (or anything else) without the risk of aggravating manic symptoms. So this debate in enlivened by the question of how much a body of academic research knowledge can be reliably and usefully transferred as relevant to everyday clinical practice.

How much of value I may have missed in this search is problematic. Paul Grof’s bibliography confirms his comment on the lengthy lapse in general interest and research on lithium’s effects in bipolar disorder. His list of references covers over seven decades (1940-2015) and of his 64 citations exactly half (32) appeared during the single decade (1961-1970) when this debate erupted. No other decade has more than three citations until 2010, since when five new publications appear of which Paul is a co-author on three.

Review of this and Jules Angst’s literature suggests that although our differences are largely reconciled lingering issues are worthy of debate.

It is a fact that one methodological concern raised by Shepherd and myself related to potential bias due to lack of a double blind and also true that we underestimated the understandable concerns about safety and suicide that later resulted in imaginative alternative research designs. A second was the possibility of statistical regression to the mean. A third, and perhaps major concern, was the heterogeneity of the patient sample. Both of these latter two concerns were elegantly displayed by the article’s graphic portrayal of episodes of illness and remission including recurrent manic, depressive and mixed forms. At a time when imipramine had established its efficacy as an antidepressant in single episodes of unipolar depression we questioned if it also might have a prophylactic effect. We tested this hypothesis using Baastrup and Schou’s statistical model in a sample of recurrent unipolar depressed patients treated with imipramine from the Maudsley Hospital data base and found it to be confirmed. This colored our conclusion that lithium was unlikely to be prophylactic for the entire spectrum of bipolar disorders.  In retrospect our overboard response to this finding might belong in the category known as “throwing out the baby with the bathwater.”

Much of Paul Grof’s and Jules Angst’s ongoing research has been devoted to a more specific clarification of what types of bipolar spectrum disorder respond in which manner to lithium, imipramine and other mood stabilizers. Grof notes (p.24) that, “recent studies have shown that bipolar disorder is now often undiagnosed, particularly in recurrent depression.” This conclusion may have been embedded in Baastrup and Schou’s original study and unkindly labelled by us as “bias”. My current assumption, based on Paul Grof’s information is that Schou’s brother failed both ECT and imipramine before responding dramatically and persistently to lithium despite being previously considered to suffer from recurrent unipolar depression. Perhaps this conviction was reflected in their diverse patient population and claim for ubiquitous benefit across the spectrum of disorders. Schou’s late life interest in this topic suggests he might have been seeking for subtle manifestations of hypomania between episodes of severe depression that would indicate a lithium responsive diathesis. Experienced clinician that Schou was, perhaps he was correct in this also.

Dr.Grof is dismissive of the Prien et al (1984) study and the weight it accords in support of imipramine’s potential benefit in recurrent depressive disorder. He cites an impressive body of contradictory evidence which includes personal phone communications from two experienced American colleagues working with Prien who were, “very critical of the patient selection, partly due to the population of V.A. hospitals, and warned me not to believe the findings once the study was completed.” The precise scientific basis for this conclusion is not revealed but the allegation is sadly reminiscent of the “ad hominem” feelings evoked by us in Schou’s sensitive response to our better articulated concerns. I regret this.

Overall I strongly agree with Grof and Angst that this kind of historical dissection can be beneficial to contemporary understanding of the evolution of neuropsychopharmacology. We all make mistakes and owning them may benefit posterity. Perhaps the lithium controversy belongs in the larger context pervading our entire field. The first two decades of clinical psychopharmacology were filled with expectations that we would “discover the right drug for the right patient.” The story of lithium, the first of our truly psychotropic drugs, so well portrayed by Paul and Jules, shows how far we have come but have yet to go in achieving that end. As Dr.Grof notes, expert committees around the world have produced 25 guidelines for the treatment of bipolar disorder despite which he notes the lax diagnostic practices, overuse and unrealistic expectations for lithium.

Despite the best efforts of dedicated researchers to define therapeutic specificity the general practitioners in our field (of which I was one) continue to operate on a “trial and error basis” when selecting drug treatment for an individual patient. This is contributed to by our still incomplete knowledge of the natural history, genetic origins and phenotypic presentations of the disorders, an unhelpful DSM system of diagnosis,  complicated by side effect sensitivity, drug interactions, differing drug profiles, variable compliance and misleading commercial mythologies regarding drug specificity.

If this sounds like, “masterful hyperbole” (of which Paul Grof accuses me, p.25) please read our essay, “Sir Aubrey Lewis and Psychopharmacology” (Blackwell and Goldberg, 2015) to better understand the differences between hyperbole (OED: deliberate exaggeration, not to be taken literally) and empiricism (OED: knowledge based on observation and experiment). The latter philosophy of science was the model in which I was trained, a style for which the Maudsley was both renowned and denigrated and of which our article, “Prophylactic Lithium: a Therapeutic Myth” remains, for all its faults, a paradigm.

This deconstruction of the lithium controversy brings to mind a final concern. At their inception over half a century ago both the ACNP and the CINP established policies and memberships dedicated to translational research and dialog. This had dual implications; that basic science might illuminate clinical research while clinical research of the caliber conducted by Grof and Angst would translate to improved everyday diagnosis and treatment by practitioners in the general fields of psychiatry and medicine.

During my own residency training “Descriptive Psychiatry” was the prevailing idiom in European psychiatry – a dedicated interest in the nosology and natural history of mental disorders, illuminated by biological, psychological and social influences and insights although treatment options were sparse. As the new drugs appeared this interest survived initially but began to wane as the connection between clinical features and outcomes was influenced by discoveries, speculations and false hopes involving neurotransmitters, receptors, neural pathways, hormonal and genetic influences. The membership and interests of the ACNP began to tilt unevenly in the direction of neuroscientists (often with dual doctorates), the number of talented clinicians dwindled by attrition while clinical research and data analyses were increasingly usurped by industry. At the same time the NIMH withdrew from new drug research. The well intentioned DSM nosology is capable, if scrupulously used, of sustaining interest in descriptive psychiatry and sophisticated biopsychosocial formulations but its multi-axial potential has been degraded to become primarily an Axis 1 diagnosis for insurance purposes and ubiquitous use of the Not Otherwise Specified (NOS) categories.

I hope that this renaissance of interest in natural history, nosology and treatment outcomes in bipolar spectrum disorders, sparked by Grof and Angst’s research will have a wider influence on the future direction of our field.


Blackwell BB, Goldberg PD. Sir Aubrey Lewis and psychopharmacology. INHN Biographies 02.05.2015.


Barry Blackwell

February 5, 2015