Incremental advances in psychiatric molecular genetics and nosology

I begin my commentary on Stein et al’s impressive, in‐depth and balanced paper about the current status of psychiatric research and practice ¹ with a few general thoughts, and then dig more deeply into two main points to further explicate what for me is a central theme in the paper.

In my view, the authors are exactly right to seek to discredit the frequently repeated claim that psychiatry is in crisis. At the same time, they provide a healthy skepticism of the claims, long echoed in our field, that major breakthroughs are “just around the corner”. They instead advocate for a far more realistic projection of modest incremental advances.

At a different level, I also applaud the authors’ commitment to explanatory pluralism. Nearly all psychiatric disorders are highly multifactorial. Despite the appeal of monocausal models, which have over history been repeatedly proposed for psychiatric disorders (and were correct for one – general paresis of the insane), they have, with this one exception (plus perhaps a small number of severe, rare forms of autism) represented false hopes. At a deeper philosophical and historical level, I also believe that the history of psychiatry has been defined by a joint commitment to brain and mind which has led to endless controversies but also a rich tradition of attempted integrations ² . Abandoning either of these approaches would lead to an impoverishment of our field.

In the spirit of Stein et al’s paper, I want to comment in more depth about two areas with which I am familiar: psychiatric genetics and psychiatric nosology. They illustrate in different ways the failure of overly‐enthusiastic paradigm shifts in the field of psychiatry and the success of slow incremental advances.

With the advent of molecular genotyping methods in the 1980s, and the early successful mapping of the Mendelian locus for Huntington’s disease, the psychiatric field, with more exuberance than was justified by the available data, yearning for a dramatic paradigm shift, sought, with poorly powered samples, single major genes for schizophrenia and bipolar illness. The result was painful and predictable – dramatic false‐positive findings followed by the inability to replicate.

Then came the ill‐conceived candidate gene era, where the field flaunted well‐understood rules of multiple testing. Furthermore, it was imagined that genes involved in the structure of neurotransmitter receptors and/or the uptake or degradation of these neurotransmitters were true candidate genes. However, these genes were not involved in the etiology of the disorders but in the action of pharmacological treatments – a classic category mistake. In a triumph of exuberance over common sense, these studies also yielded almost entirely false‐positive findings.

Then came the more mundane and much more effortful brute force method of genome‐wide association studies (GWAS). This humbler method was based on the fact that, for psychiatric disorders, we knew next to nothing about specific etiologic mechanisms of illness. The field properly did its multiple testing homework. What was unknown was the expected effect sizes of the risk alleles. Not surprisingly, initial estimates here were far too optimistic. The first studies with sample sizes thought to be adequate were entirely negative. Then something unexpected happened. The field abandoned important parts of the more typical academic model of inter‐group competition for a model of inter‐group cooperation, forming the Psychiatric Genomic Consortium ³ . Positive results, that have replicated well, finally began to flow in, first as a trickle and then as a cascade ⁴ .

As genotyping costs fell, samples of exome sequence became large enough to also begin to yield positive results – this time identifying specific biological processes, and not the statistical signal of a genomic region that is obtained from positive GWAS results ⁵ . This is only a start. The pathway from genetic variants to pathophysiology, let alone to druggable targets, will certainly be long and complex, and success is by no means certain. Like many of the chronic diseases of humanity, we now know that psychiatric disorders suffer from the “curse of polygenicity”.

I next want to turn to psychiatric nosology with some comments complementary to those provided by Stein et al, and with an apology for a parochial emphasis on the US perspective. The Feighner criteria represented a major break with the “Great German Professor Principle” whereby the influence of a psychiatric nosology was based largely on the reputation of the proposer. By contrast, the Feighner criteria grew out of a journal club run by S. Guze and E. Robins at Washington University St. Louis that tried to develop criteria from the then rather skimpy empirical literature, inevitably complemented, when data were lacking, by clinical experience ⁶ . We might call this an “empirically aided expert consensus” model.

This model was applied with only modest changes for the Research Diagnostic Criteria, the DSM‐III and the DSM‐III‐R, the last of which I was able to observe personally. Literature was discussed, but largely to support clinical opinions, with some growth during the process of the importance of having at least some research basis for proposals for change. Systematic literature reviews became much more common in the DSM‐IV, although they varied widely in quality, and the idea of trying to systematically evaluate a set of validators was not widely adopted.

The DSM‐5 was initially conceptualized by its leaders as a paradigm shift in nosology, in particular with the intention of moving from descriptive to etiologically based diagnoses. However, when the work groups for DSM‐5 began meeting, none of them felt that the available data were adequate to support such a change. Although several had approached the DSM‐5 leadership to develop a clear set of guidelines for changes in DSM‐5, these requests were not acted upon ⁷ . However, in the middle of the DSM process, the DSM leadership requested such a document, that was developed quickly by a small group. While widely disseminated, the recommendations were not systematically adopted by all work groups.

With rising concerns about the heterogeneity of the approach across the work groups, the leadership of the American Psy­chiatric Association requested the formation of a Scientific Review Committee which, building on the previously proposed criteria, further developed them in a conscious attempt to move the process from an “empirically aided expert consensus” model to a more empirically driven process in which the focus would shift from personal expert opinion to systematic review of research evidence for validity and reliability ⁸ .

This is a very challenging process, and will never be as simple as the evaluation of efficacy of a drug treatment, which can focus largely on results from randomized controlled trials and reports of side effects. What we see in the DSM‐US based psychiatric nosologic process is a gradual shift from an expert consensus to a more data‐driven decision making, in line with the developments of the broader medical field ⁹ .

I am convinced that a move toward eti­ological diagnoses in psychiatry will result from incremental advances, not one dra­matic change. The DSM‐5 already contains an etiologic diagnostic criterion for narcole­psy – evidence for a hypocretin deficiency. In the coming years, if genetic risk factors (e.g., polygenic risk scores) or imaging findings can add to the diagnostic validity or reliability of specific diagnostic categories, then they can be added with the usual diagnostic review process. Eventually, psychiatric diagnostic criteria may come to resemble those seen in other areas of medicine, for example, rheumatology, where the operationalized criteria are a mix of symptoms, signs, course of illness, and specific biolo­gical find­ings.