Separation of anxiety and depressive disorders: blind alley
in psychopharmacology and classification of disease
Edward Shorter, Peter Tyrer
No new drugs for mood and anxiety disorders have reached the market for over a decade.
Why is there so little innovation in a sector that accounts for the largest proportion by far of sales
of psychiatric drugs?
The current division between anxiety and depression is
increasingly recognised as inadequate. In the commu-
nity, most mood disorders present as a combination of
depression and anxiety. Yet the Food and Drug Admin-
istration in the United States, which has become the
world bellwether of drug approval,indicates drugs either
for major depression or for the various forms of anxiety
recognised by the American Psychiatric Association’s
Diagnostic and Statistical Manual of Mental Disorders
(DSM). As a result, the pharmaceutical industry is com-
pelled to develop drugs for diagnoses that are of
questionable clinical relevance. This is one reason for
the big slowdown in drug discovery in psychiatric drugs.
A return to the former unitary classification of mood
and anxiety disorders as nervousness or cothymia might
represent a way out of this blind alley.
Origins of the new system
In 1980,the American Psychiatric Association revised its
standard system of diagnoses in the third edition of its
diagnostic manual (DSM-III).1This document erected a
firewall between depression and anxiety. Indeed, in
drafting this edition the association appointed separate
committees to study depression and anxiety and stated
that any overlap between the two disorders would
henceforth be considered mainly as comorbidity.
Although this division was controversial at the time,
DSM-III became the accepted psychiatric nosology
worldwide, and its successors dominate the picture
today.2Recent observers,however,suggest that
x The concept of “major depression”is far too hetero-
geneous to be useful3
x The subdivision of anxiety into separate micro-
diagnoses of panic, social anxiety disorder, etc, is
x The firewall between anxiety and depression
ignores the fact that the commonest form of affective
disorder is mixed anxiety-depression.5
Admittedly, the manual allows for diagnoses such
as dysthymic disorder, a chronic form of depression
that merges closely with major depression, and adjust-
ment disorder, a lost diagnosis originally offered as a
political sop to the large American psychotherapeutic
community.6However, neither diagnosis has proved
The crucial point is that the Food and Drug
Administration accepts psychiatric drugs almost exclu-
sively for DSM-style indications; European regulators
seem to be headed in the same direction. For example,
the European Agency for the Evaluation of Medicinal
Products has decided to “re-evaluate the existing
requirements” for the treatment of anxiety by focusing
on generalised anxiety disorder.7So the name of the
disease determines how it is treated: depression is
treated with antidepressants, anxiety with anxiolytics.
Who defines psychiatric diagnoses?
Officially, diagnostic decisions are made through
scientific consensus by the World Health Organization
and the American Psychiatric Association. Small com-
mittees of experts decide whether, for example, schizo-
phrenia is one disease or several, and the process of
decision making should be transparent and based on
good scientific evidence.But sometimes the evidence is
poor and influenced heavily by the pharmaceutical
industry. The industry exerts a major influence
through publication of sponsored supplements to
journals, which are often poorly peer reviewed and
promote unapproved treatments.8Such supplements
are particularly common for drugs for anxiety and
depression as these are the most common treated con-
ditions. Worldwide sales of antidepressants dwarf sales
of drugs for all other psychiatric disorders.
Industry is said to prefer the disease based
approach of DSM-III to any dimensional approach to
DSM-style diseases represent tidy diagnostic market
niches. Every new diagnosis represents a new licensing
opportunity. Companies only have to show the
effectiveness of an existing drug over placebo for the
new diagnosis in large clinical trials. This is expensive
but not difficult if the same methods can be used as for
previous studies with the old diagnoses.
Slowdown in drug discovery
The increase in the number of diagnoses of anxiety
and depression has occurred at the same time as an
equivalent slowdown in production of new drugs
Anxious or depressed? In practice, many patients are both
Education and debate
History of Medicine
Program, Faculty of
Canada M5G 1VJ
professor of the history
College Faculty of
Medicine, St Mary’s
head of department
BMJ VOLUME 32719 JULY 2003 bmj.com
(figure). Although this association may not be causal, it
does raise the question: what are these new diagnoses
for? The basic drugs used most commonly for mood
and anxiety disorders were largely discovered in the
1950s and 1960s or even before—valproic acid was first
described in 1882. Only four new drugs apart from
benzodiazepines and selective serotonin reuptake
inhibitors were patented in the 1970s compared with
15 in the 1960s (table). The four drugs were
bupropion, buspirone, modafinil, and mirtazapine—
and mirtazapine is an analogue of mianserin, which
Organon patented in 1967.
Just four drugs for mood and anxiety were patented
in the 1980s—two hypnotics (zaleplon and zolpidem),
venlafaxine,and the serotonin reuptake inhibitor sertra-
line. Since 1990, no drugs have been patented in the
mood and anxiety area that have reached the US
market. (Escitalopram, the most recent marketed
selective serotonin uptake inhibitor and an isomer of
citalopram, was patented in July 1990.) Among the
popular drugs currently prescribed in the United King-
dom, trazodone was patented in 1968, buspirone in
1973,moclobemide in 1977,and reboxetine in 1979.
Reasons for slowdown
Why has development of drugs for mood largely come
to an end? We believe it is connected to the artificial
separation of anxiety and depression.9Before DSM-III,
treatments tended to be for symptoms,not diseases.The
drugs developed before this time, including ampheta-
mines, benzodiazepines, and tricyclic antidepressants,
were used to treat both depression and anxiety. Later,
drug indications were linked clearly to disease states, so
that antipsychotics, anxiolytics, and antidepressants
became mutually exclusive categories.10Thus, nosology
suddenly became the prime force in shaping the indica-
tions for a drug. The regulators encouraged industry to
produce antidepressants, so they were duly introduced,
but more accurate labels would have been antifatigue
drugs,tonics,or psychic energisers.
The growing search for blockbusters in the 1970s
resulted in a trend to rubbish earlier drugs in order to
put new patent protected drug classes on the market.
Despite clear evidence that benzodiazepines were effec-
tive, they were dismissed as drugs for neurotic women,
who then become addicted.11The tricyclic antidepres-
sants were castigated for their adverse effects, and in
1998, a consensus conference on tricyclics compared
with selective serotonin reuptake inhibitors agreed that
tricyclic antidepressants should not be considered as
first line antidepressant treatment except in patients
admitted to hospital with depression.12Eli Lilly, which
patented fluoxetine,paid for the conference.
This backlash against established drugs obscured
the notion that certain drug classes were effective for
both anxiety and depression, particularly the benzodi-
azepines. With the decline of the benzodiazepines in
the 1980s, anxiolytic therapy was placed on the back
burner and antidepressants rushed into the spotlight,
in the form of selective serotonin reuptake inhibitors.
In the world of pharmacotherapy, anxiety and depres-
sion had become as different as chalk and cheese.
Commerce and cothymia
But while drug discovery has faltered, the proliferation
of niche diagnoses continues apace and mixed
diagnoses are scorned.13For example, each new
anxiety diagnosis seems to create an opportunity for
the promotion of a phoney new drug indication, so we
have paroxetine for social anxiety disorder, fluvox-
amine for obsessive compulsive disorder, and sertra-
line for post-traumatic stress disorder.
Industry has been busy behind the scenes in this
handy convergence of eccentric new diagnoses and the
market nicheing of compounds. For example, in May
1984, Robert Spitzer, the chief disease designer of
DSM-III and DSM-III-R, convoked a meeting of the
anxiety working group, cosponsored by “the Psychop-
harmacology Unit of the Division of Medical Affairs of
the Upjohn Company.” At the end of the discussion of
the relation between panic and agoraphobia, Spitzer
announced, “Consensus favors the Upjohn model.”14It
is now routine for psychopharmacologists, such as
Brown University’s Martin Keller,to receive as much as
$500 000 (£320 000) in consulting fees from industry
in a given year.15USA Today has calculated that at 55%
of the meetings of the various advisory committees of
the FDA, “half or more of the FDA advisers had a con-
flict of interest.”16
Sometimes the relation between academic psychia-
trists and industry veers over the line of acceptability in
the form of ghost writing—academics lending their
names to articles drafted by industry hacks. This has
been a problem in psychopharmacology since the
1950s.17But only last year, Vienna psychiatry professor
Siegfried Kasper was identified in the Austrian press as
signing an industry ghostwritten article about an anti-
No of drugs and diagnoses
1950s 1960s1970s1980s 1990s
Inverse relation between number of new drugs patented for mood
and anxiety disorders and number of anxiety and depressive
diagnoses in DSM revisions
Drugs for mood and anxiety disorders advertised in major
American psychiatry journals, 1960-2000 by decade that they
were patented or chemical formula first made public*
or made public
*Main sources for drug advertisements: American Journal of Psychiatry,
Diseases of the Nervous System, Mental Hospitals, and Hospital and
Community Psychiatry. Sources for patent data: Pharmaceutical Manufacturing
Encyclopedia (2nd ed), Merck Index (13th ed).
Education and debate
BMJ VOLUME 32719 JULY 2003bmj.com
depressant.18Under normal circumstances, the inter- Download full-text
penetration of industry and academe can be fruitful, as
talent and ideas wash back and forth. Yet when drugs
start earning the kind of money usually associated with
the oil industry, there is potential for trouble.
We believe that the failure to advance the treatment of
anxiety and depression is related to wrong classification.
If you don’t have natural disease categories, you can’t
develop drugs for them. If the Food and Drug Adminis-
tration will accept only drugs that are effective for DSM
diagnoses, and if the diagnoses are artefacts, the drugs
are bound to be less valuable, even if in the short term
they increase their market share. Companies must start
developing drugs for mixed anxiety and depression and
forget about dividing this giant illness segment into
salami slices. Doctors could encourage this change by
being more cynical about pitches from drug representa-
tives claiming to have “the latest” in anxiolytic
medication.Ask instead for the latest in nervousness.19
Contributors and sources: ES has carried out research on
historical aspects of psychiatry over many years, wherever possi-
ble studying documents that are key to the making of decisions
by policy makers and psychiatric bodies. Most recently, he has
Administration and the American Psychiatric Association. PT
has carried out clinical trials into the drug and psychological
treatment of anxiety and depression since 1969 and has been
involved in field trials for the 10th revision of the International
Classification of Diseases (ICD-10).
Competing interests: PT has received support from the Mental
Health Foundation to evaluate the outcome of anxiety and
1 American Psychiatric Association. Diagnostic and Statistical Manual of
Mental Disorders, Third Edition. Washington, DC: American Psychiatric
Blashfield RK, Fuller AK. Predicting the DSM-V. J Nerv Ment Dis
Kendell R, Jablensky A. Distinguishing between the validity and utility of
psychiatric diagnoses. Am J Psychiatry 2003;160:4-12.
Tyrer P. Comorbidity or consanguinity. Br J Psychiatry 1996;168:669-71.
Jenkins R, Lewis G, Bebbington P, Brugha T, Farrell M, Gill B, et al. The
national psychiatric morbidity surveys of Great Britain—initial findings
from the household survey. Psychol Med 1997;27:775-89.
Casey P, Dowrick C, Wilkinson G. Adjustment disorders: fault line in the
psychiatric glossary. Br J Psychiatry 2001;179:479-81.
European Agency for the Evaluation of Medicinal Products. Concept
paper on the development of a committee for proprietary medicinal
products note to guidance on clinical investigation of medicinal products
for the treatment of generalised anxiety disorder. www.emea.eu.int/pdfs/
human/ewp/428402en.pdf (accessed 23 May 2003).
Bero LA. The publication of sponsored symposiums in medical journals.
N Engl J Med 1992;327:1135-40.
Tyrer P. The case for cothymia: mixed anxiety and depression as a single
diagnosis. Br J Psychiatry 2001;179:191-3.
10 Shorter E. Looking backwards: a possible new path for drug discovery in
psychopharmacology. Nature Rev Drug Discov 2002;1:1003-6.
11 Romach M, Busto U, Somer G, Kaplan Hl, Sellers E. Clinical aspects of
12 Mendlewicz J, Lecrubier Y. Antidepressant selection: proceedings from a
TCA/SSRI consensus conference. Acta Psychiat Scand Suppl 2000;403:5-8.
13 Hudson JI, Pope HGJ. Affective spectrum disorder: does antidepressant
response identify a family of disorders with a common pathophysiology?
Am J Psychiatry 1990;147:552-64.
14 Spitzer R.Letter to “Dear Colleague,”8 May1984.Anxiety meeting folder.
J Williams papers, DSM-III-R files, box 3. Washington, DC: American
Psychiatric Association Archives, 1994.
15 Bass A. Drug companies enrich Brown professor. Boston Globe 1999 Apr
10: http://www.prozactruth.com/moneytrail.htm (accessed 9 Jun 2003).
16 Cauchon D. FDA advisers tied to industry. USA Today 2000 Sep 25:1.A
http://pqasb.pqarchiver.com/USAToday (accessed 9 Jun 2003).
17 Denber HCB. Notes from another time and place. In: Ban T, Healty D,
Shorter E.The rise of psychopharmacology and the story of CINP. Budapest:
18 Schoenbauer R. Uni-Psychiater unter “Ghostwriting” Verdacht. Der
Standard 2002 May 11-12.
19 Healy D. Some continuities and discontinuities in the pharmacotherapy
of nervous conditions before and after chlorpromazine and imipramine.
Hist Psychiatry 2000;11:393-412.
(Accepted 14 May 2003)
Corrections and clarifications
Severe acute respiratory syndrome:patients were
During the editing process the affiliations of two
coauthors listed at the end of this letter by Moira
Chan-Yeung and colleagues were wrongly inserted
(21 June, p 1393). Their job titles were correct, but
the hospitals they work at were wrong. Thomas S T
Lai is a consultant physician at Princess Margaret
Hospital, Hong Kong, and Wilson K S Yee is a
consultant physician at Kwong Wah Hospital,
Cost effectiveness analysis of different approaches of
screening for familial hypercholesterolaemia
Occasionally, an error only comes to light a year or
more after publication, as is the case with this
paper by Dalya Marks and colleagues, which was
published in June last year (BMJ 2002;324:1303-6).
The error occurred only in the full length version
of the paper (on bmj.com); the abridged version
was correct. Somehow, in the production process,
we managed to publish one table twice (as table 3
and table 4). The table is correct as table 4, but
table 3 should have shown the “comparison of
number needed to be invited to screening and
screening costs (ages 16-54 years) for different
strategies using clinical and genetic confirmation of
the diagnosis” (in other words, the table that was
called table 2 in the abridged version). The correct
table 3 now appears with the full version (http://
Managing pulmonary embolism
The BMJ long ago gave up checking every
reference that is supplied. We do, however, expect
full details and are alert to any “suspicious looking”
element in a reference. Some errors, of course, are
difficult to spot. In this editorial by Karin Janata
(21 June, pp 1341-2), the technical editor and
proofreader could not have known that the page
numbers in reference 2 were wrong because they
did not look particularly odd. The British Thoracic
Society’s guidelines for management of suspected
acute pulmonary embolism are indeed published
in volume 58 of Thorax (2003), but they appear on
pages 470-83 [not 1-14].
Since 1980 anxiety and depression have been
considered as discrete diseases
Drug development has subsequently been
tailored to new diagnoses
The increase in the number of mood and anxiety
diagnoses has coincided with a fall in the rate of
development of new drug types
The proliferation of niche diagnoses is liked by
industry because it creates new licensing
opportunities for drugs
The focus of drug development needs to return
to mixed anxiety-depression disorders
Education and debate
BMJ VOLUME 32719 JULY 2003 bmj.com