INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int J Geriatr Psychiatry 2003; 18: 829–838.

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/gps.940

Guideline for the management of late-life depression in

primary care
Robert C. Baldwin1*, David Anderson2, Sarah Black3, Sandra Evans4,5, Rob Jones6, Ken Wilson7
and Steve Iliffe8 On behalf of the Faculty of Old Age Psychiatry Working Group, Royal College
of Psychiatrists, London, UK
1
Manchester Mental Health and Social Care Trust, Manchester Royal Infirmary, Manchester, UK
2
Mossley Hill Hospital, Liverpool, UK
3
Wonford House Hospital, Exeter, UK
4
Homerton Hospital, London, UK
5
Queen Mary University of London, UK
6
Queen’s Medical Centre, Nottingham, UK
7
University of Liverpool, Liverpool, UK
8
Reader in General Practice, University College London, Department of Primary Care and Population Science,
Royal Free & UCL Medical School, London, UK

SUMMARY
Objective To develop a guideline for the primary care management of depression in later life based on best practice.
Method Source material included relevant guidelines, literature reviews and consensus documents coupled with an
updated literature review covering 1998–October, 2001. This material was summarised as a series of evidence-based state-
ments and recommendations agreed by consensus.
Results Good quality evidence exists for the pharmacological and psychological treatment of depressive episode
(major depression), although not specifically in primary care. There is some evidence of efficacy of antidepressants in
late-life dysthymia and minor depression associated with poor functional status. In depressive episode, current evidence
suggests acute treatment for at least six weeks and a continuation period of at least 12 months. Both tricyclic antidepressants
and Selective Serotonin Re-uptake Inhibitors are effective in longterm prevention. There is less data on how to manage
patients who do not respond in the acute treatment phase. More data is needed on sub-groups of patients with specific
co-morbid medical conditions and those who are frail. Collaborative care is effective in older depressed primary care
patients.
Conclusions There are effective treatments for depression in primary care. More research is needed to address the
optimum treatment of depression with medical co-morbidity and to elucidate the role of newer psychological inter-
ventions. Collaborative care between primary care and specialist services is a promising new avenue for management.
Copyright # 2003 John Wiley & Sons, Ltd.

key words — depression; geriatric; guideline; practice; management

INTRODUCTION older people (Beekman et al., 1999). Untreated, it

Depression is the most common mental health pro-
blem of later life, affecting approximately 15% of
* Correspondence to: Dr R. C. Baldwin, Manchester Mental Health
and Social Care Trust, York House, Manchester Royal Infirmary,
Oxford Road, Manchester M13 9WL, UK. Tel: +44161 276 5317.
Fax: +441612765303. E-mail: Robert.Baldwin@man.ac.uk
Copyright # 2003 John Wiley & Sons, Ltd.
shortens life, increases health care costs as well as
adding to disability from medical illnesses (Penninx
et al., 2000) and is the leading cause of suicide among
older people (Lebowitz et al., 1997). When treated,
quality of life improves (Shmuely et al., 2001).
Most depression is treated entirely in primary
care but there is evidence of poor detection (Marks
et al., 1979; Plummer et al., 1997) and sub-optimal
Received 3 March 2003
Accepted 17 June 2003
830 r. c. baldwin et al.
treatment (Iliffe et al., 1991). This guideline sum-
marises relevant evidence about the primary care
management of late-life depression. Its focus is on
medication management and validated psychological
interventions. It is not intended as a specialist guide
although it is of relevance to specialist services. In
the UK specialists are expected to provide guidance,
education and training in the management of common
disorders such as depression (Department of Health,
2001). Additionally, collaboration between primary
and specialist care providers is associated with better
treatment outcomes for depression (von Korff and
Goldberg, 2001; Unützer et al., 2002).
METHOD
Because there are a number of existing reviews of the
subject, the following strategy was used:
(1) Key reviews and consensus documents (indicated
by a star in the reference list) were used as the
starting point.
(2) A literature search covering the period between
1998 and October 2001 (inclusive) was con-
ducted—1999 being the published year of the
latest reviews to be examined. The search strategy
and its principal findings have been published
(Baldwin, 2002).
(3) Evidence-based statements were developed and
circulated among the authors for comment until a
consensus was reached. This is similar to the
approach used by Anderson and colleagues
(2000).
(4) Evidence was categorised using the system of
Shekelle et al. (1999, see Appendix). An alpha-
betical grading system, A–D is applied to arrive at
the overall strength of recommendations.
TYPES OF DEPRESSIVE DISORDER
IN OLDER PEOPLE
Table 1 lists the main symptoms that are found in late-
life depressive disorder. A classification based on cur-
rent evidence is outlined in Table 2; ICD-10 codes
(World Health Organisation, 1993) are used in Table
2 and the footnotes list alternative descriptions used in
DSM-IV (American Psychiatric Association, 1994).
Some authorities believe that it is better to consider
depressive disorder as a spectrum rather than as dis-
crete categories. For example, some so-called ‘sub-
threshold’ (minor) depressions (see later) are in reality
Copyright # 2003 John Wiley & Sons, Ltd.
Table 1. Symptoms of depressive disorder (World Health
Organisation, 1993; Alexopoulos et al., 2001)
Core symptoms:
 Depressed mood sustained for at least two weeks
(on most days, much of the time) AND/OR
 Loss of interest or pleasure in usual activities AND/OR
 Decreased energy, increased fatigue (in patients who are
physically ill this may amount to feelings of fatigue not
confined to exertion); diminished activity
Other symptoms:
 Suicidal thoughts or behaviour
 Loss of confidence or self-esteem
 Feeling of helplessness
 Inappropriate or excessive guilt
 Feelings of hopelessness or worthlessness
 Avoiding social interactions or going out
 Poor concentration and/or difficulty with memory
 Psychomotor retardation or agitation
 Sleep disturbance
 Reduced appetite with corresponding weight loss
cases of incompletely resolved depressive episode
(Judd et al., 1998).
In depressive episode at least five symptoms
(including two core ones) from those listed in
Table 1 should be present. In psychotic depression
delusions and/or hallucinations are present, often
with a content of guilt, hypochondriasis, poverty or
imminent disaster.
Organic depressive episode is diagnosed when
there is a direct patho-aetiological link between the
onset of depression and either a systemic or neurolo-
gical condition or an ingested substance or drug. The
criteria for a depressive episode must be met. There
are a large number of possible systemic, neurological
Table 2. Types of depressive disorder
 Mild depressive episode (F32.0)1
 Moderate depressive episode (F32.1)1
 Severe depressive episode without psychotic symptoms (F32.2)1
 Severe depressive episode with psychotic symptoms (F32.3)1
 Recurrent depressive disorder (F33)
 Organic depressive episode (F06.32)2
 Bipolar affective disorder: current episode mild, moderate
or severe depression with or without psychotic symptoms
(FF31.3–F31.5)
 Dysthymia (F34.1)
 Mixed anxiety and depressive disorder (F41.2)
 Adjustment disorder with depressive reaction (F43.20, F43.21)
 Minor depressive disorder3
1
Equivalent to Major Depression in DSM-IV (Code 296).
2
Equivalent to Mood Disorder due to a General Medical Condition
in DSM-IV (Code 293.83).
3
This appears as a proposal for further consideration in DSM-IV;
see text.
Int J Geriatr Psychiatry 2003; 18: 829–838.
 guideline for late-life depression
conditions and prescription drugs which may act in
this way the most common being cancer, stroke,
Parkinson’s disease, hip fracture, myocardial infarc-
tion, heart failure, chronic obstructive pulmonary dis-
ease and asthma (Baldwin et al., 2002). Depression
accompanying dementia may be classified here but
in DSM-IV there are separate mood specifiers for
the dementias. The prevalence of organic depressive
disorder caused or aggravated by medication is not
known. It has been estimated that beta-blockers alone
could account for 1–10% of depressive disorders in
older people (Dhondt and Hooijer, 1995). Alcohol
can precipitate or prolong major depression or depres-
sive symptoms.
Bipolar disorder of first onset in older people is
uncommon and may be associated with organic dis-
ease. Older people with bipolar disorder beginning
at a younger age have a high level of health burden
(Bartels et al., 2000).
Dysthymia is a chronic depression with a duration
of at least two years and a number of symptoms
from Table 1. These range from a few to over five.
Dysthymia may be more common among older
adults than is generally realised (Devanand et al.,
1994).
In mixed anxiety and depressive disorder symp-
toms of depression and anxiety are both present but
below the threshold for either depressive episode or
generalised anxiety disorder.
Adjustment disorder with depressive reaction is
diagnosed when depressive symptoms below the
threshold for a diagnosis of depressive episode begin
within a month of a serious threat or loss. Symptoms
usually resolve within six months.
The term minor depressive disorder does not
appear in ICD-10 but has been proposed as a ‘poten-
tial category’ in DSM-IV. Depressive syndromes
below the criteria for depressive episode constitute
the most common type of depression among older
people in primary care (Beekman et al., 1999). Minor
depression is sometimes used as short-hand. It com-
prises similar symptoms to depressive episode but
with a lower symptom count and less functional
impairment. Minor depression and depressive episode
share similar risk factors such as poor health, isola-
tion, disability and poor satisfaction with life (Chiu
et al., 1999). Both are associated with reduced physi-
cal activity, less social contact and increased mortal-
ity. Minor depression in later life is associated with
functional impairment approaching that of depressive
episode (Lyness et al., 1999) and minor depression
may be a risk factor for depressive episode (Judd
et al., 1998).
Copyright # 2003 John Wiley & Sons, Ltd.
831
Table 3. Recommendations for when to refer for specialist advice
(World Psychiatric Association, 1999; Alexopoulos et al., 2001)
Category D
 When the diagnosis in doubt (for example, is this dementia?)
 When depression is severe, as evidenced by;
Psychotic depression
Severe risk to health because of failure to eat or drink
Suicide risk
 Complex therapy is indicated (for example in cases with medical
co-morbidity)
 When first-line therapy fails (although primary care physicians
may wish to pursue a second course of an antidepressant from a
different class)
MANAGEMENT OF DEPRESSIVE DISORDER
IN OLDER PEOPLE
General management principles include (World
Psychiatric Association, 1999):
* Monitoring the risk of self harm.
* Educating the patient (and care givers) about
depression and involving him or her in treatment
decisions.
* Treating the whole person—co-existing physical
disorder; attention to sensory deficits and other
handicaps; sign-posting the patient to appropriate
social care agencies; reviewing medication with a
view to withdrawing those unnecessary.
* Treating depressive symptoms with the aim of
complete remission (as residual symptoms are a
risk factor for chronic depression; Lebowitz et al.,
1997; Judd et al., 1998).
* Prompt referral of patients requiring specialist
mental health services (Table 3).
Acute treatment with antidepressants
Evidence of benefit
Antidepressant drugs are effective in older patients
with depressive episode (Ia). The response rate to
antidepressants is 50–60% as opposed about to 30%
for patients given placebo. Evidence of efficacy is
summarised in a Cochrane systematic review (Wilson
et al., 2001) and three meta-analyses (Mittmann et al.,
1997; Gerson et al., 1998; McCusker et al., 1998).
Based on the papers reviewed by Mittmann et al.
(1997), Katona and Livingston (2002) have developed
an alternative approach to efficacy using the concept
of Numbers-Needed-to-Treat (NNT). Data from the
three meta-analyses can be summarised: (1) there are
no significant differences in efficacy between differ-
ent classes of antidepressants; (2) adverse events
Int J Geriatr Psychiatry 2003; 18: 829–838.
832 r. c. baldwin et al.

appear to be similar across classes; (3) tricyclic
antidepressants may be more effective in more severe
depression, although one recent study did not confirm
this (Mulsant et al., 2001a). Although efficacy and
tolerability data from clinical trials show little
difference between SSRIs and tricyclic antidepres-
sants, there is evidence from naturalistic studies that
more older depressed patients with medical co-
morbidity have contraindications to tricyclics than
SSRIs (Cole et al., 2001) and under-treatment is more
common with tricyclics than SSRIs (Rojas-Fernandez
et al., 1999). Medical co-morbidity is discussed later.
The following are caveats: (1) there are very few stu-
dies directly of primary care patients (Freudenstein et
al., 2001); (2) most treatment trials are of short dura-
tion; (3) residual symptoms are common; (4) there are
few studies of very old people; and (5) there are few
controlled trials of antidepressants in patients who are
frail, such as in nursing homes.
Wilson et al. (2001) concluded that there was insuf-
ficient evidence to recommend low dosage antide-
pressant treatment in depressive episode in primary
care. A controversial meta-analysis of mixed-aged
subjects concluded that low dosage triclyclics might
be effective but the range of conditions studied
appeared heterogeneous (Furukawa et al., 2002).
More research involving older depressed patients is
needed.
There are no important differences in speed of
onset of antidepressants (Ia–IIb), although in general
older people take longer to recover than younger
adults (Wilson et al., 2001; Mottram et al., 2002).
There is evidence for the efficacy of antidepressants
in dysthymia (Ia—mixed aged patients) and in some
patients with minor depression (Ib) (Williams et al.,
2000). A recent Cochrane review of antidepressants
in mixed-aged patients with dysthymia showed that
they are effective with no significant advantage of
one group over another (Lima and Moncrieff, 2002).
There is no evidence that antidepressants are effective
in the management of recent onset (less than 4 weeks)
minor depression or uncomplicated grief. In the study
by Williams et al. (2000) paroxetine was effective in
treating dysthymia and persistent minor depression in
patients with functional impairment.
Antidepressants are effective in depressed patients
with a range of physical co-morbid conditions,
although tolerability varies (Ia—mixed-aged patients;
Ib—elderly). A Cochrane systematic review, whilst
not specific to older people, has shown this (Gill
and Hatcher, 1999). The NNT, of 4, is similar to that
of the systematic review of antidepressants in older
people (Wilson et al., 2001). Some studies show small
advantages of newer over older antidepressants in
terms of tolerability. In one controlled trial fluoxetine
was well tolerated and effective in patients with a
range of systemic illnesses, although the study was
under-powered (Evans et al., 1997).
Recommendations based on current evidence are
summarised in Table 4.
No evidence of benefit
There is no evidence that antidepressants are effective
in sub-threshold (minor) depression complicating
dementia but there is some evidence that offering sup-
port to the patient and care-giver and ‘waiting and
seeing’ is effective (IIa, although the target condition
may be more akin to psychological distress than
strictly defined depression). Antidepressant drug trials
demonstrate that there is a high rate of spontaneous

Table 4. Recommendations for practice—acute treatment with antidepressants

 For psychotic depression specialist treatment is indicated; an antidepressant alone is unlikely to be effective (B)1
 For moderate to severe (non-psychotic) depressive episode antidepressant medication is effective (A)
 In primary care treatment there is no evidence that one class of antidepressant is any more effective than another (A)
 Although newer antidepressants are no more effective than older ones, they are marginally better tolerated in healthy older people and are
safer, especially in overdose (C)
 In patients with medical co-morbidity, SSRIs are better tolerated than older antidepressants and safer (C)
 Other newer antidepressants such as venlafaxine, mirtazepine or nefazodone2 may also be effective and well tolerated in frail patients (D)
 Patients with moderate to severe depressive episode complicating dementia should be considered for treatment with an SSRI or
moclobemide (A)
 First-line treatment for patients with mild depression complicating dementia is practical and emotional support, including for the care
giver (C)
 Patients with moderate to severe depressive episode complicating stroke should be considered for treatment with an antidepressant (C)
 ‘Watchful waiting’ is recommended for stroke patients with mild depression (C)
 Low dose antidepressant treatment is not recommended for older depressed patients (A)
1
Although one recent study showed that antipsychotic augmentation of an antidepressant was no better than the antidepressant alone
(Mulsant et al., 2001b).
2
Nefazodone is being withdrawn in some European countries.

Copyright # 2003 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2003; 18: 829–838.
 guideline for late-life depression
resolution of symptoms within four weeks in those
with mild depression associated with dementia
(World Psychiatric Association, 1999; Alexopoulos
et al., 2001). Indirect evidence of the positive role
of support on mood comes from a CBT-oriented trial
in which modest short-term improvements in beha-
vioural and psychological symptoms in the dementia
group followed intervention (Marriott et al., 2000).
There is evidence that antidepressants, particularly
newer ones, are tolerated by depressed patients living
in nursing homes but no good controlled trial data to
show that they are more effective than placebo or sup-
port. Uncontrolled studies suggest that SSRIs, moclo-
bemide or simply providing socialisation may be
effective (III). Tricyclic antidepressants although pos-
sibly effective may be less well tolerated (Chiu et al.,
1999). In an open study of SSRIs (Trappler and
Cohen, 1998) the response rate was poor if the patient
had associated dementia, but otherwise good. Like-
wise in an open pilot study with sertraline in patients
with minor (sub-threshold) depression outcome and
tolerability were good (Rosen et al., 2000). A sociali-
zation programme was associated with short-term
benefit in mood (Rosen et al., 1997). The lack of ran-
domised controlled trials means that this data must be
interpreted with caution.
Choosing an antidepressant
There is little evidence that one class of antidepres-
sant is more effective than another (Mittmann et al.,
1997). The choice is determined by patient character-
istics (such as severity of depression—perhaps
favouring tricyclics; safety in overdosage—favouring
newer antidepressants; prior response to a particular
agent; tolerability; anticipated side effects; drug
interactions; compliance; and frailty) and by local
protocols. In primary care, newer drugs such as
the SSRIs are gradually replacing tricyclics as the
drugs of first choice (Mamdani et al., 2000). Data sug-
gest that they are only marginally better tolerated
(Anderson et al., 2000) although they are safer in
overdose (Mittmann et al., 1997). Lofepramine, a tri-
cyclic derivative, is an effective antidepressant which
appears to be as safe as SSRIs. In frail patients it is
advisable to commence treatment with a low
dose and titrate gradually to the therapeutic dosage
range.
There is no evidence that different classes of anti-
depressants differentially affect co-morbid symptoms
such as anxiety or insomnia. In a recent consensus
guideline (level IV evidence) clinicians favoured
using a more sedative antidepressant for troublesome
Copyright # 2003 John Wiley & Sons, Ltd.
833
insomnia and would only consider hypnotics (with a
short half-life) for short-term use (Alexopoulos et al.,
2001). For residual anxiety, most clinicians favoured
increasing the antidepressant dose to the maximum
recommended level in patients without cognitive
impairment. They did not support the use of sedatives
in patients with dementia or cognitive impairment
(Alexopoulos et al., 2001).
Medical co-morbidity
Medical co-morbidity frequently complicates the
management of late-life depression. The following
is not exhaustive.
Depression in cardiovascular and cerebrovascular
disease
SSRIs are the preferred antidepressants for depres-
sion complicating cardiovascular and cerebrovascu-
lar disease, mainly because of safety and tolerability
(Level III, mainly by extrapolation from studies of
mixed-aged patients). Depression complicating
ischaemic heart disease worsens cardiovascular out-
comes (Lebowitz et al., 1997), but there are
insufficient data to show whether treatment with
antidepressants improves this. A large multi-centre
trial (SADHART) has demonstrated efficacy for
sertraline over placebo in mixed-aged patients with
recurrent depression (Glassman et al., 2002) and
tricyclics are associated with significant adverse
cardiac events compared to SSRIs (Roose et al.,
1998). An absence of elderly-specific data limits what
recommendations can be made, but older tricyclics
should be avoided in elderly patients with ischaemic
heart disease.
Antidepressants are effective in depressive episode
complicating stroke but spontaneous recovery is com-
mon in the first six weeks (Ib). Depression is common
after stroke. There is a high rate of spontaneous recov-
ery within six weeks of stroke but thereafter this
diminishes rapidly (Andersen et al., 1994). The few
controlled trials which exist point to the efficacy of
SSRIs (Andersen et al., 1994; Wiart et al., 2000)
and to adverse effects from older tricyclics (Gustafson
et al., 1995). Other treatments such as venlafaxine
and mirtazepine have been recommended (Alexopou-
los et al., 2001) but clinical trial data are lacking.
Depression in degenerative neurological disease
Dementia. For persistent depression in dementia
tricyclic antidepressants, SSRIs and moclobemide are
effective; newer drugs are better tolerated and
safer (Evidence level Ib). Depression occurs more
Int J Geriatr Psychiatry 2003; 18: 829–838.
834 r. c. baldwin et al.
commonly in vascular than Alzheimer dementia (Chiu
et al., 1999) although antidepressant trials have mainly
been conducted on patients with the latter. There are
placebo-controlled trial data for moclobemide (Roth
et al., 1996), citalopram (Nyth et al., 1992) and sertra-
line (Lyketsos et al., 2000) and a head-to-head compar-
ison of paroxetine and imipramine (Katona, 1998). All
were effective and in the latter study paroxetine was
better tolerated and safer than impramine.
Parkinson’s disease. There are no adequate trial data
upon which to make recommendations about antide-
pressants. SSRIs may aggravate parkinsonian symp-
toms but this is not proven. As with other diseases
affecting the sub-cortical brain regions, there is a very
high rate of depression in idiopathic Parkinson’s dis-
ease. However, there are too few intervention trials to
allow recommendations about individual antidepres-
sants. Case reports have suggested that parkinsonian
symptoms may be made worse by SSRIs. Caution
should therefore be exercised when treating this patient
group. In one open-label study (Dell’Agnello et al.,
2001) patients with Parkinson’s disease were allocated
to one of four SSRIs and treated prospectively without
adverse effects on their motor symptoms.
Duration of treatment
At least six weeks of treatment is recommended to
achieve optimal therapeutic effect (Ia) Older patients
take longer to recover from depression (Lebowitz
et al., 1997). A systematic review recommended at
least six weeks of treatment (Wilson et al., 2001)
and other data show ongoing recovery up to 12 weeks
provided there has been some earlier improvement
(Lebowtiz et al., 1997).
In patients showing virtually no response in the first
four weeks continuing the same treatment is unlikely
to lead to recovery (IIb). How long is a reasonable
treatment trial is weighed up by factors such as
whether the antidepressant dosage is optimal, the
patient is compliant and whether there are psychoso-
cial factors impeding recovery. Six weeks is a recom-
mended minimum treatment interval, but if there has
been hardly any response (less than 25% recovery) in
the first four weeks continuing the same treatment
is associated with a low probability of remission
(Mottram et al., 2002).
In patients who have not responded in the first four
weeks the choice lies between switching to an antide-
pressant from another class, augmenting with another
substance or referring to the specialist service (III–
IV). There is insufficient evidence to recommend
one strategy over another. Switching class is the
Copyright # 2003 John Wiley & Sons, Ltd.
most practicable option in the primary care setting.
Aggregated data from studies of mixed-aged patients
suggest a response rate of approximately 50% (WPA,
1999) but most trials are uncontrolled and therefore
this conclusion should be interpreted cautiously.
In patients who have shown at least a 25%
improvement it is recommended continuing the same
antidepressant and optimising the dose (IIb). If the
patient has clearly improved at four weeks but has
not recovered (roughly indicated by at least 25–50%
clinical improvement), then continuing the anti-
depressant at optimal dosage is associated with
further remission in a significant proportion of
patients (Mottram et al., 2002). Controlled trial data
are lacking so that it is recommended that the patient
be reviewed at intervals of two to four weeks and the
help of the specialist team sought if there is deteriora-
tion or recovery is uncertain.
Augmentation therapy for non-responsive depres-
sion is widely used by specialists but has a limited evi-
dence base in older people (III). This includes
compounds such as lithium salts which are not routi-
nely prescribed by primary care physicians and which
are associated with a relatively high incidence of
adverse effects (WPA, 1999). Likewise, Electrocon-
vulsive Therapy (ECT), is a treatment which may
be used by specialists if pharmacological treatment
fails or in cases where life is acutely threatened.
Augmentation with a psychological intervention
such as CBT or IPT may enhance recovery in slow
or incomplete responders (IIa). Augmentation is
usually taken to mean enhancing the effects of an
antidepressant with additional medication. However,
augmentation with a psychological intervention may
reduce time to remission (Reynolds et al., 1999a;
Williams et al., 2000; Thompson et al., 2001) and
has at least as good an evidence base as medication
augmentation.
Psychological therapy in the acute phase of treatment
Age should not be a barrier to receiving a psychologi-
cal therapy (Katona et al., 1995). With increased
knowledge about their efficacy, demand for psycholo-
gical treatments by older people is likely to increase,
with more opting for a psychological approach over
drug treatment (Un}utzer et al., 2002).
In mild to moderate depression episode CBT, IPT
and brief focal dynamic psychotherapy are effective
and there is emerging evidence for Problem-Solving
Therapy (IIa). Psychological treatments such as those
recommended by the English National Service
Framework for Older People include Cognitive
Int J Geriatr Psychiatry 2003; 18: 829–838.
 guideline for late-life depression
Behaviour Therapy (CBT), Inter-Personal Therapy
(IPT) and brief focal analytic psychotherapy (Depart-
ment of Health, 2001). These are effective treatments
for depressive disorder in older people but they are
under-used (Lebowitz et al., 1997). Other effective
interventions, although not specifically used for
depressive disorder, include life review and pro-
blem-solving therapy (Gatz et al., 1998; Pinquart
and Sorensen, 2001). The latter has been subject to
considerable research although less so among older
adults.
In moderate to severe (non-psychotic) depressive
episode treatment which combines antidepressant
medication with a psychological intervention such
as CBT or IPT is associated with better outcomes than
either intervention alone (Ib). This was demonstrated
by Reynolds et al. (1999a) using IPT with nortripty-
line and by Thompson et al. (2001) for desipramine
with CBT. In the latter study the combination had
greatest effect in those who were most depressed.
For sub-threshold depression, such as adjustment
disorder and minor depression, offering structured
support to the patient and care giver may be as effec-
tive as antidepressants or psychological interventions
(IIa). Offering support is not doing nothing. In antide-
pressant trials quite high rates of symptomatic recov-
ery have been observed in controls groups comprising
‘attention’ or ‘support’ (McCusker et al., 1998). In
patients with dementia a structured family interven-
tion is associated with improved mood in care givers
(Marriott et al., 2000).
A psychological intervention delivered in combina-
tion with an antidepressant is effective in preventing
recurrence in high risk patients (Ib). In a study of
middle-aged-to-older patients prophylactic IPT
helped reduce the risk of relapse and was most effec-
tive when combined with an antidepressant (Reynolds
et al., 1999b). In a one year study, CBT, in addition to
usual care, resulted in reduction in depression severity
at follow-up compared to those given a placebo
(Wilson et al., 1995).
Recommendations based on this evidence are sum-
marised in Table 5.
Table 5. Recommendations for psychological interventions
 It is recommended that primary care teams have access to staff
trained in the delivery of a psychological intervention such as
CBT, IPT or Problem-Solving Treatment (A)
 Psychological interventions should be available to patients as a
first choice option in mild to moderate depressive episode (A)
 Structured support to patients and care givers is recommended
for mild depressive symptoms (B)
Copyright # 2003 John Wiley & Sons, Ltd.
835
Table 6. Recommendations for ongoing treatment
 Patients with first onset depressive episode should be
continued for a minimum of 12 months on the same dosage
of antidepressant that got them well (B)
 Those with recurrent depression should be continued at the same
dosage that got them well for 12–36 months with a review held
with the patient to discuss risk factors at regular intervals (C)
 For high risk patients, treatment should be continued for a
minimum of 3 years (D)
 Patients at high risk of recurrence should be jointly managed
with the local specialist psychiatry service (D)
Duration of treatment
At the point of remission (when all symptoms have
resolved) it is recommended continuing the antide-
pressant at the same dose for another 12 months
for a first onset case (IIb) and expert opinion suggests
up to 36 months for recurrent cases (IV) (Alexopoulos
et al., 2001).
Both tricyclic antidepressants and some SSRIs are
efficacious in the prevention of relapse and recur-
rence over periods of 1–3 years (Ib). The Old Age
Depression Interest Group demonstrated this for
dothiepin (Old Age Depression Interest Group,
1993), a tricyclic, and a recent study has replicated
this using the SSRI, citalopram (Klysner et al.,
2002). As with younger adults, treatment with full
dosage (not reduced dosage) is recommended.
Those most at risk of a recurrence include patients
with multiple past episodes, poor health, chronic dis-
ability or severe social difficulties (III) Such patients
should be considered for long-term maintenance
treatment. Some authorities favour life-long treatment
for all older patients with late life depression but
others disagree. A pragmatic approach is to continue
treatment for 12 months and then review risk factors
with the patient.
Recommendations on duration of treatment and
ongoing care are summarised in Table 6.
Collaborative care
In mixed-aged depressed patients collaborative care is
effective (von Korff and Goldberg, 2001). Models
vary but the principles include the use of multi-
faceted interventions, identifying a case manager
(e.g. a practice nurse) and flexible collaboration
between primary and secondary (specialist) care to
improve access to the skills of the psychiatrist. Three
randomised controlled trials have shown that this
approach is effective in older patients (Waterreus
et al., 1994; Banerjee et al., 1996; Unützer et al.,
2002).
Int J Geriatr Psychiatry 2003; 18: 829–838.
836 r. c. baldwin et al.
In the largest of the three (Unützer et al., 2002),
1801 patients aged 60 years or above with depressive
episode, dysthymia or both, were randomised to either
an active intervention or usual care from the primary
care team. Interventions included antidepressant
treatment, the option of Problem-Solving Treatment,
care management, support with medication compli-
ance and education about depression. The NNT, 4,
is about the same as controlled trials of antidepres-
sants (Wilson et al., 2001; Katona and Livingston,
2002) with an average cost over 12 months of £343
($550; s395) per intervention subject. However, resi-
dual symptoms were quite common.
CONCLUSIONS
As with younger adults, there is good evidence for the
efficacy of pharmacological and psychological treat-
ments in late-life depressive episode (major depres-
sion), although specific data for primary care are
very limited. There is some evidence of efficacy of
antidepressants in late-life dysthymia and persistent
minor depression with impaired function. A minimum
of six weeks treatment with antidepressant monother-
apy is recommended although in slow responders
(showing some improvement in the first 4–6 weeks)
treatment can be continued for up to 12 weeks. Low
dose antidepressants are not currently recommended.
Continuation treatment should last at least 12 months.
Both tricyclic antidepressants and SSRIs are effective
in preventing recurrence of depressive episode.
There is less data on how to manage patients who
do not respond in the acute treatment phase. If a
patient does not respond to a trial of an antidepressant
given in adequate dosage for a sufficient time the
options are to switch to an antidepressant from
another class or to augment with another medication
or a psychological intervention. The latter may not be
a realistic option in primary care when there is a lack
of trained personnel. However, consideration should
be given to referral to practice counsellors who are
more widely available and may have the requisite
skills to help with mild depression and adjustment
disorders. Restricting access to practice counsellors
based on age is not acceptable. Referral to a specialist
should be considered for patients who have not
responded to acute treatment.
There are weaknesses in the literature concerning
the treatment of very frail patients with depression
and those who have depression co-morbid with other
disorders such as dementia and a range of medical
conditions. Collaborative care is effective in older
depressed patients.
Copyright # 2003 John Wiley & Sons, Ltd.
KEY POINTS
* Both antidepressants and psychological inter-
ventions are effective treatments for later life
depressive disorder.
* There is little data concerning the efficacy of
treatment specifically in primary care.
* Acute treatment lasts at least 6 weeks and
continuation treatment at least 12 months.
* Newer antidepressants are better tolerated and
safer in patients with co-morbid medical
disorder but they are not more effective.
* For patients who do not respond to first-line
antidepressant treatment, the optimum strategy
is unclear because of a lack of evidence. A
common strategy, augmentation, is usually
taken as meaning medication although aug-
mentation with a psychological intervention is
also effective.
* Secondary prevention is effective using differ-
ent classes of antidepressants or with psycho-
logical treatments.
ACKNOWLEDGEMENTS
This report has been endorsed by the Executive Com-
mittee of the Faculty of Old Age Psychiatry of the
Royal College of Psychiatrists, London, UK. We
would like to thank professor Cornelius Katona for
helpful comments on the first draft.
DECLARATION OF INTEREST
Professor Baldwin has received travel grants and hos-
pitality from Janssen Pharmaceuticals, Novartis and
Lilly within the past three years in order to attend
Board meetings of the International Psychogeriatric
Association but has not been involved in antidepres-
sant drug trials in the past ten years; Professor Wilson
has received a grant from Pfizer for conducting a
controlled trial of sertraline in older depressed people;
Dr Evans was a Research Fellow in 1989–1991
funded by Janssen Research Foundation; Dr Jones
has received travel grants and hospitality from Lilly
within the past three years to support attendance at
the International Psychogeriatric Association meet-
ings. He has received support from the Royal College
of Psychiatrists to research guidelines for dementia
but has not been involved in any drug trials in the past
ten years; Drs Anderson and Iliffe have no conflict of
interest.
Int J Geriatr Psychiatry 2003; 18: 829–838.
 guideline for late-life depression
REFERENCES
(*)Alexopoulos GS, Katz IR, Reynolds CF, Carpenter D, Docherty
JP. 2001. The expert consensus guideline series: pharmacother-
apy of depressive disorders in older patients. Postgrad Med Spe-
cial Report. (October): 1–86. Expert Knowledge Systems,
L.L.C; McGraw-Hill Healthcare Information Programs: Minnea-
polis, MN.
American Psychiatric Association (APA). 1994. Diagnostic and
Statistical Manual Version IV. APA: Washington, DC.
Andersen G, Vestergaard K, Lauritzen L. 1994. Effective treatment
of post-stroke depression with the selective serotonin reuptake
inhibitor citalopram. Stroke 25: 1099–1104.
(*) Anderson IM, Nutt DJ, Deakin JFW. 2000. Evidence-based
guidelines for treating depressive disorders with antidepressants:
a revision of the 1993 British Association for Psychopharmacol-
ogy guidelines. J Psychopharmacol 14: 3–20.
(*) Baldwin RC, Chiu E, Katona CLE, Graham N. 2002. Guidelines
for Depressive Disorder in Later Life: Practising the Evidence.
Martin Dunitz: London.
Baldwin RC. 2002. Research into depressive disorder in later life:
who is doing what? A literature search from 1998–2001. Int
Psychogeriatr 14: 335–346.
Banerjee S, Shamash K, Macdonald AJ, Mann AH. 1996. Rando-
mized controlled trial of effect of intervention by psychogeriatric
teams on depression in frail elderly people at home BMJ 313:
1058–1061.
Bartels SJ, Forester B, Miles KM, Joyce T. 2000. Mental health ser-
vice use by elderly patients with bipolar disorder and unipolar
depression. Am J Geriatr Psychiatry 8: 160–166.
Beekman ATF, Copeland JRM, Prince MJ. 1999. Review of com-
munity prevalence of depression in later life. Br J Psychiatry
174: 307–311.
(*) Chiu E, Ames D, Draper B, Snowdon J. 1999. Depressive dis-
orders in the elderly: a review. In Depressive Disorders, Maj M,
Sartorius N (eds). John Wiley & Sons Ltd: Chichester, UK; 313–
363.
Cole MG, Elie LM, McCusker J, Bellavance F, Mansour A. 2001.
Feasibility and effectiveness of treatments for post-stroke
depression elderly inpatients: a systematic review. J Geriatr
Psychiatry Neurol 14: 37–41.
Dell’Agnello G, Ceravolo R, Nuti A, Bellini G, Piccinni A,
D’Avino C. 2001. SSRIs do not worsen Parkinson’s disease:
evidence from an ope-label, prospective study. Clin Neurophar-
macol 24: 221–227.
Devanand DP, Nobler MS, Singer T, et al. 1994. Is dysthymia a dif-
ferent disorder in the elderly? Am J Psychiatry 151: 1592–1599.
Department of Health. 2001. National Service Framework for Older
People. Department of Health: London.
Dhondt ADF, Hooijer C. 1995. Iatrogenic origins of depression in
the elderly. Int J Geriatr Psychiatry 10: 1–8.
Evans M, Hammond M, Wilson K, Lye M, Copeland J. 1997.
Placebo-controlled treatment trial of depression in elderly physi-
cally ill patients. Int J Geriatr Psychiatry 12: 817–824.
Freudenstein U, Jagger C, Arthur A, Donner-Banzhoff N. 2001.
Treatments fro late-life depression in primary care—a systema-
tic review. Fam Practice 18: 321–327.
Furukawa TA, McGuire H, Barbui C. 2002. Meta-analysis of effects
and side effects of lose dosage tricyclic antidepressants in
depression: systematic review. BMJ 325: 991–995.
Gatz M, Fiske A, Fox LS, et al. 1998. Empirically validated psycho-
logical treatments for older adults. J Ment Health Aging 4: 9–46.
Gerson S, Belin TR, Kaufman A, Mintz J, Jarvik L. 1999. Pharma-
cological and psychological treatments for depressed older
Copyright # 2003 John Wiley & Sons, Ltd.
837
patients: a meta-analysis and overview of recent findings. Harv
Rev Psychiatry 7: 1–28.
Gill D, Hatcher S. 1999. Antidepressant drugs in depressed patients
who also have a physical illness (Cochrane Review). In The
Cochrane Library, Issue 3. Update Software: Oxford.
Glassman AH, O’Connor CM, Califf RM, et al. 2002. Sertraline
treatment of major depression in patients with acute MI or
unstable angina. JAMA 288: 701–709.
Gustafson Y, Nilsson I, Mattsson M, Astrom M, Bucht G. 1995.
Epidemiology and treatment of post-stroke depression. Drugs
Aging 7: 298–309.
Iliffe S, Haines A, Gallivan S, Booroff A, Goldenberg E, Morgan P.
1991. Assessment of elderly people in general practice. 1. Social
circumstances and mental state. Br J Gen Pract 41(342): 9–12.
Judd LL, Akiskal HS, Maser JD, et al. 1998. A prospective 12-year
study of subsyndromal and syndromal depressive symptoms in
unipolar major depressive disorders. Arch Gen Psychiatry 55:
694–700.
Katona C, Freeling P, Hinchcliffe K, Blanchard M, Wright A. 1995.
Recognition and management of depression in late life in general
practice: consensus statement. Prim Care Psychiatry 1: 107–
113.
Katona CLE. 1998. A double-blind comparison of the efficacy and
safety of paroxetine and imipramine in the treatment of depres-
sion with dementia. Int J Geriatr Psychiatry 13: 100–108.
Katona C, Livingston G. 2002. How well do antidepressants work
in older people? J Affect Disord 69: 47–52.
Klysner R, Bent-Hansen J, Hansen HL, et al. 2002. Efficacy of cita-
lopram in the prevention of recurrent depression in elderly
patients: placebo-controlled study of maintenance therapy. Br J
Psychiatry 181: 29–35.
(*) Lebowitz BD, Pearson JL, Schneider LS, et al. 1997.
Diagnosis and treatment of depression in late life. JAMA 278:
1186–1190.
Lima MS, Moncrieff J. 2002. Drugs versus placebo for dysthymia.
In The Cochrane Library. Update Software Ltd: Oxford.
Lyketsos CG, Sheppard J-ME, Steele CD, et al. 2000. Randomised,
placebo-controlled, double-blind clinical trial of sertraline in the
treatment of depression complicating Alzheimer’s disease: initial
results from the depression in Alzheimer’s disease study. Am J
Psychiatry 157: 1686–1689.
Lyness JM, King DA, Cox C, Yoediono Z, Caine ED. 1999. The
importance of subsyndromal depression in older primary care
patients: prevalence and associated functional disability. JAGS
47: 647–652.
Mamdani M, Parikh SV, Austin PC, Upshur REG. 2000. Use of
antidepressants among elderly subjects: trends and contributing
factors. Am J Psychiatry 157: 360–367.
McCusker J, Cole M, Keller E, Bellavance F, Berard A. 1998.
Effectiveness of treatments of depression in older ambulatory
patients. Arch Int Medicine 158: 705–712.
Marks J, Goldberg D, Hillier V. 1979. Determinants of the ability of
general practitioners to detect psychiatric illness. Psycholog Med
9: 337–353.
Marriott A, Donaldson C, Tarrier N, Burns A. 2000. Effectiveness
of cognitive-behavioural family intervention in reducing the bur-
den of care in carers of patients with Alzheimer’s disease. Br J
Psychiatry 176: 557–562.
(*) Mittmann N, Herrmann N, Einarson TR, et al. 1997. The effi-
cacy, safety and tolerability of antidepressants in late life depres-
sion: a meta-analysis. J Affect Disord 46: 191–217.
Mottram PG, Wilson KCM, Ashworth L, Abou-Saleh M. 2002. The
clinical profile of older patients’ response to antidepressants—
an open trial of sertraline. Int J Geriatr Psychiatry 17: 574–578.
Int J Geriatr Psychiatry 2003; 18: 829–838.
838 r. c. baldwin et al.
Mulsant BH, Pollock BG, Nebes R, et al. 2001a. A twelve week,
double-blind, randomized comparison of nortriptyline and par-
oxetine in older depressed inpatients and outpatients. Am J Ger-
iatr Psychiatry 9: 406–414.
Mulsant BH, Sweet RA, Rosen J, et al. 2001b. A double-blind ran-
domized comparison of nortriptyline plus perphenazine versus
nortriptyline plus placebo in the treatment of psychotic depres-
sion in late life. J Clin Psychiatry 62(8): 597–604.
Nyth AL, Gottries CG, Lyby K, et al. 1992. A multicenter clinical
study of citalopram and placebo in elderly depressed patients
with and without concomitant dementia. Acta Psychiatrica
Scand 86: 138–145.
Old Age Depression Interest Group. 1993. How long should the
elderly take antidepressants? A double blind placebo-controlled
study of continuation/prophylaxis therapy with dothiepin. Br J
Psychiatry 162: 175–182.
Penninx BW, Deeg DJ, van Eijk JT, Beekman AT, Gurainik JM.
2000. Changes in depression and physical decline in older adults:
a longitudinal perspective J Affect Disord 61: 1–12.
Pinquart M, Sorensen S. 2001. How effective are psychotherapeutic
and other psychosocial interventions with older adults? A meta-
analysis. J Ment Health Aging 7: 207–243.
Plummer S, Ritter SAH, Leach RE, Mann AH, Gournay KJ. 1997.
A controlled comparison of the ability of practice nurses to
detect psychological distress in patients who attend their clinics.
J Psychiatric Ment Health and Nursing 4: 221–223.
Reynolds CF 3rd, Miller MD, Pasternak RE, et al. 1999a. Treatment
of bereavement-related major depressive episodes in later life: a
controlled study of acute and continuation treatment with nor-
triptyline and interpersonal psychotherapy. Am J Psychiatry
156: 202–208.
Reynolds III CF, Frank E, Perel JM, et al. 1999b. Nortriptyline and
interpersonal psychotherapy as maintenance therapies for recur-
rent major depression: a randomized controlled trial in patients
older then 59 years. J Am Med Assoc 281: 39–45.
Rojas-Fernandez C, Thomas VS, Carver D, Tonks R. 1999. Subop-
timal use of antidepressants on the elderly: a population-based
study in Nova Scotia. Clin Therapeut 21: 1937–1950.
Roose SP, Laghrissi-Thode F, Kennedy JS, et al. 1998. Comparison
of paroxetine and nortripyline in depressed patients with
ischemic heart disease. JAMA 279: 287–291.
Rosen J, Rogers JC, Marin RS, Mulsant BH, Shahar A, Reynolds
CF, III. 1997. Control-relevant intervention in the treatment of
minor and major depression in a long-term care facility. Am J
Geriatr Psychiatry 5: 247–257.
Rosen J, Mulsant BH, Pollock BG. 2000. Sertraline in the treatment
of minor depression in nursing home residents: a pilot study. Int J
Geriatr Psychiatry 15: 177–180.
Roth M, Mountjoy CQ, Amrein R. 1996. Moclobemide in elderly
patients with cognitive decline and depression: an international
double-blind placebo-controlled study. Br J Psychiatry 168:
149–157.
Shekelle PG, Woolf SH, Eccles M, Grimshaw J. 1999. Developing
guidelines. BMJ 318: 593–596.
Shmuely Y, Baumgarten M, Rovner B, Berlin J. 2001. Predictors of
improvement in health-related quality of life among elderly
patients with depression. Int Psychogeriatr 13: 63–73.
Thompson LW, Coon DW, Gallgher-Thompson D, Sommer BR,
Koin D. 2001. Comparison of desipramine and cognitive/beha-
vioral therapy in the treatment of elderly outpatients with
mild-to-moderate depression. Am J Geriatr Psychiatry 9: 225–
240.
Trappler B, Cohen CI. 1998. Use of SSRIs in ‘very old’ depressed
nursing home residents. Am J Geriatr Psychiatry 6: 83–89.
Copyright # 2003 John Wiley & Sons, Ltd.
Un}utzer J, Katon W, Callahan C, et al. 2002. Collaborative care
management of late-life depression in the primary care setting.
JAMA 288: 2836–2845.
von Korff M, Goldberg D. 2001. Improving outcomes in depres-
sion. BMJ 323: 948–949.
Waterreus A, Blanchard M, Mann A. 1994. Community psychiatric
nurses for the elderly: few side-effects and effective in the treat-
ment of depression. J Clin Nurs 3: 299–306.
Wiart L, Petit H, Joseph PA, Mazaux JM, Barat M. 2000. Fluoxetine
in early poststroke depression: a double-blind placebo-controlled
study. Stroke 31: 1829–1832.
Williams JW, Barrett J, Oxman T, et al. 2000. Treatment of dysthy-
mia and minor depression in primary care: a randomised con-
trolled trial in older adults. JAMA 284: 1519–1526.
Wilson KCM, Scott M, Abou-Saleh M, Burns R, Copeland JRM.
1995. Long-term effects of cognitive-behavioural therapy and
lithium therapy on depression in the elderly. Br J Psychiatry 167:
653–658.
Wilson K, Mottram P, Sivanranthan A, Nightingale A. 2001. Antide-
pressant versus placebo for depressed elderly (Cochrane Review).
In The Cochrane Library, Issue 2. Update Software: Oxford.
World Health Organisation (WHO). 1993. The ICD-10 classifica-
tion of mental and behavioural disorders: research criteria.
WHO: Geneva.
(*) World Psychiatric Association (WPA). 1999. WPA International
Committee for Prevention and Treatment of Depression. Depres-
sive Disorders in Older Persons. NCM Publishers Inc: New York,
NY. Available on-line: http://www.wpanet.org/sectorial/edu4.html
APPENDIX: CATEGORIES OF EVIDENCE
AND STRENGTH OF RECOMMENDATIONS
Categories of evidence for causal relationships
and treatment (Shekelle et al., 1999)
Ia Evidence from meta-analysis of randomized con-
trolled trials
Ib Evidence from at least one randomized controlled
trial
IIa Evidence from at least one controlled study without
randomization
lIIb Evidence from at least one other type of quasi-
experimental study
III Evidence from non-experimental descriptive stu-
dies, such comparative studies, correlation studies
and case-control studies
IV Evidence from expert committees reports or
opinions and/or clinical experience of respected
authorities.
Strength of recommendation
A Directly based on category I evidence
B Directly based on category II evidence or extrapo-
lated recommendation from category I evidence
C Directly based on category III evidence or extra-
polated recommendation from category I or II
evidence
D Directly based on category IV evidence or extra-
polated recommendation from category I, II or III
evidence.
Int J Geriatr Psychiatry 2003; 18: 829–838.