Management of Late Life Depression

1
Management of Late Life Depression

• J. Craig Nelson, M.D.
• Leon J. Epstein Professor of Psychiatry
• Director of Geriatric Psychiatry
• University of California San Francisco

2
Financial Disclosure J. Craig Nelson, MD

• During the past 12 months, I have had
  relationships with the following companies
• Lecture Honoraria none
• Consultation/Advisory Board Abbott, Biovail,
  Bristol- Myers Squibb, Corcept, Eli Lilly,
  Forest, Glaxo-SmithKline, Novartis, Orexigen,
  Organon, Pfizer
• Research Support none
• Pharmaceutical Stock Ownership none

3
Your meal will be out shortly. The salmon was a
little wilder than we anticipated.
4
Graying of the Population
Number in US Population Over Age 64 or 84
Millions of Persons
Year
Administration on Aging 2001
5
Prevalence of Depression
6
Give it to me straight, Doc, how many more golden
years do I have staring me in the face?
7
National Comorbidity Survey Replication 2003
N9,090
Overall prevalence of MDD 16.2 Lifetime
Prevalence of MDD by Age Group
Kessler RC et al. JAMA 20032893095-3105.
8
Cumulative Prevalence of Major Depression by
Birth Cohort
Kessler et al JAMA 2003
9
Prevalence of Depressionin 2552 Subjects with
and without Medical Disorders
14
12
P lt0.01
10
Percent
8
6
4

2
0
6 Months
Lifetime
Wells KB, et al. Am J Psychiatry.
1988145976-981.
10
Prevalence of Major Depression in US Seniors
Prevalence()
12-16
10-15
10-12
2-4
Mulsant BH, Ganguli M. J Clin Psychiatry.
199960(suppl)9-15.
11
Consequences of Depression
12
Adverse Effects of Depression

• Poor quality of life
• Diminished sense of well being
• Impairment in activities of daily living
• Increased risk of suicide
• Increased mortality
• Poor outcome of medical illness

13
Treatment of Depression
14
Treatment of Depression
Types of Treatment

• Evidence based Psychotherapy
• IPT, CBT, PST
• Psychosocial interventions
• Antidepressants
• ECT

15
Efficacy of Psychotherapy in Depressed Elders

• Extensive Review by Arean and Cook
• Relatively few studies with a supportive clinic
  visit control
• Strong support for
• CBT alone
• IPT with antidepressants
• Less support for
• IPT alone
• Brief dynamic psychotherapy
• Life review therapy

Arean P and Cook B. Biol Psychiatry
200252293-303
16
Supportive Clinic Visits Are Fairly Effective
The Largest Drug vs Placebo Study in 728 Patients
with Late Life Depression
CGI responders much or very much improved
100
Sertraline
Placebo
80
60
P.005
Responding ()
Medication clinic visits placebo account for
78 of response in the drug group
40
45
35
20
0
ITT Sample
Schneider L, Nelson JC et al. Am J Psychiatry
20031601277-1285.
17
The question for both psychotherapy and drug
treatment is whether any specific treatment is
better than supportive clinic visits.
18
Pharmacologic Treatments
19
What seems to be the trouble, Fred?
20
Algorithm for Pharmacologic Treatment of
Depression (Cont.)
Primary target for antidepressant Rx
21
Treatment of Older Patients
Other Factors to Consider

• Comorbid medical conditions
• Comorbid dementia
• Disabilities hearing, vision, other
• Gait Disturbance
• Nutrition
• Medical medications and the potential for drug
  interactions

22
Do Older Patients Respond to Treatment as Well as
Younger Patients?
23
The Largest Study of Older and Younger Patients
N5243

• Patients treated with escitalopram for 8 weeks
• Open label treatment with standard assessments
  (CGI, HADS)
• Mean dose and duration
• Older patients 11.3 mg/d for 54.4 days
• Younger patients 11.7 mg/d for 56.7 days

Nelson JC, Rush AJ, Bose A, et al. Submitted
24
CGI Response Rates in Depressed Patients gt or lt
60 years N4887
ITT/LOCF CGI Response Rates
Nelson JC, Rush AJ, Bose A, et al. Submitted
25
Comparison of Response in Older and Younger
Adults with Major Depression
Walsh and Sysko. J Clin Psychopharmacol
200525(suppl 1)S29-33.
26
How Effective Are Antidepressants in Older Adults?
27
I tell you, mock jury duty beats cancer testing.
28
Meta-analysis of Antidepressant Studies in
Late-Life Depression

• Second-generation antidepressants (nontricyclics)
• Random-assignment, placebo-controlled clinical
  trials
• Patients with major depression
• Age 60 years and over
• Living in the community
• Not limited to a specific medical illness

Nelson JC, et al. Unpublished data.
29
Literature Search Flow
30
Meta-analysis of Late-Life Depression Studies
Nelson JC, et al. Presented at the American
Association for Geriatric Psychiatry Annual
Meeting March 1-4, 2007 New Orleans, La.
31
The odds ratio is the probability of an event
(e.g. response) occurring in the treatment group
divided by the same probability in the control
group. The overall odds ratio for the
meta-analysis is the mean of the odds ratios
computed for each contrast weighted for sample
size and event rate. The odds ratio for this
meta-analysis was 1.40 (95 CI 1.24-1.57,
Plt0.001). The risk difference was 0.08.The
number needed to treat was 13.
Meta-analysis of Late-Life Depression Studies
(Cont.)
32
Mean Pooled Response Rates In 10 Studies With 13
Contrasts
NNT7
NNT20
Response ()

Nelson JC, Delucchi K, Schneider LS. Am J
Geriatric Psychiatry in press
33
Discontinuation Rates In 10 Studies With 13
Contrasts
Discontinuation for any reason OR1.22, 95 CI
1.06-1.40, P0.005 Discontinuation for adverse
events OR1.84, 95 CI 1.51-2.24, Plt0.001

Nelson JC, Delucchi K, Schneider LS. Am J
Geriatric Psychiatry in press
34
Response Rates in Placebo-Controlled Trials and
Comparison Trials
Expectation of receiving placebo lowers response
rates
Sneed JR, et al. Unpublished data.
35
Efficacy of Antidepressants

• Best established for relapse/recurrence
  prevention
• In mixed age adults, a meta-analysis of 31
  studies with 4410 patients found1
• Antidepressant treatment reduced relapse rates
  70
• Relapse on drug 18, on placebo 41

1. Geddes et al. Lancet 2003361653-61.
36
Preventing Recurrence
Efficacy of Combined Therapy

• 96 Patients, gt 59 years
• Diagnosis recurrent major depression without
  psychosis or dementia
• Fully remitted after NT IPT
• Randomized to NTIPT, NT, IPT Placebo, Placebo
  (NT levels 80 to 120 ng/ml)
• Followed for 3 years

Reynolds et al. JAMA. 1999(Jan)281(1)39-45
37
Recurrence Rates in Late Life MDD
3 Year Outcomes
100
80
60
Recurrence
40
20
0
NTIPT
NTMC
MCPlacebo
IPTPlacebo
All Active Rxs gt Placebo, p0.001
NTnortriptyline IPTinterpersonal
psychotherapy MCmedication clinic Reynolds C,
et al. JAMA 1999
38
2 Year Recurrence Rates Paroxetine and IPT
116 Patients with MDD, 70 years and older
Overall comparison, p0.02 Any paroxetine group
vs any placebo group, 36 vs 63, p0.02
NNT4
Reynolds et al. N Engl J Med 20063541130-38.
39
Efficacy of Antidepressants in Older Adults
Comparison Studies
40
Mirtazapine vs. ParoxetinePatients ?65 Years
with MDD, N246
2
5
Mirtazapine (n 126) 25.6 mg
2
0
Paroxetine (n 120) 26.5 mg
1
5
HAM-D17 Total Score



1
0

5
0
7
14
21
28
42
56
Study Days
Plt0.05
Schatzberg, ACNP, 2000
41
Effects of Treatment on Suicidal Thinking and
Behavior
42
2003 U.S. Suicide Rates by Age and Sex
CDC, National Center for Injury Prevention and
Control. http//webapp.cdc.gov/cgi-bin/broker.exe
43
Suicidal Thinking during Sertraline Treatment in
Late Life Depression

• Large placebo controlled trial performed
• 8 week randomized, double-blind trial
• Sertraline 50-100 mg/d vs placebo
• 752 patients enrolled, 728 with at least one post
  treatment assessment

Nelson et al. Am J Geriatric Psychiatry
200715573-580
44
Suicidal Thinking during Sertraline Treatment in
Late Life Depression
MMRM analysis, SERT vs PLBO, p0.02
Nelson et al. Am J Geriatric Psychiatry
200715573-580
45
Emergent suicidal thinking during treatment in
248 patients without suicidal thinking at
baseline.
Nelson et al. Am J Geriatric Psychiatry
200715573-580
46
Odds Ratios for Spontaneously Reported Suicidal
Thinking And Behavior by Age
An analysis of 372 placebo-controlled RCTs of
12 second-generation antidepressants in 99,839
patients
FDA Report Nov 2006
47
Moderator Variables
48
If the drug-placebo difference in response is
modest, might determination of moderators of
response help to identify those unlikely to
respond or those who might have a robust response
to treatment?
Why Look for Moderator Variables?
49
Treatment for Depression with Heart Disease
The SADHART Study
pns
p0.003
p0.001
More severe 2 or more prior episodes and HAMD gt
18 Glassman AH et al. JAMA Aug 14 2002,
288(6)701-709.
50
Critical Distinction

• Most commonly in the literature, predictors are
  cited that predict overall treatment response.
• Less commonly are predictors of the specific
  effects of drug treatment identified.

51
Age of Onset and Response Late Onset 60 Years
Old
Roose et al Am J Psychiatry 2004
52
Are there moderating variablesfor response?

• The literature suggests that the following are
  associated with reduced response
• Older Age (gt 75 years)
• Lesser Severity ( median baseline HAMD)
• Single episode (versus recurrent)
• Late onset ( 60 years)
• Anxious depression (HAMD anxious factor)
• Cognitive impairment (MMSE lt 24)
• We also examined sex

53
Odds Ratios of ITT Drug and Placebo Response
Rates in Patients lt 75 and 75 Years
Nelson JC, Delucchi K, Schneider LS. Unpublished
data
54
)
55
Response Rates for Drug and Placebo in Patients lt
75 and gt 75 years
Response gt 50 improvement on HAMD or MADRS

Odds ratios in young-old and old-old were 1.39
(CI 1.21, 1.60) and 1.13 (1.16, 1.48) and did not
differ significantly in the two groups, ?2 2.45,
p0.06. Risk difference 60 0.03 (CI -0.02,
0.08) lt 60 0.08 (CI 0.05, 0.11)
56
Summary of Odds Ratios and Confidence Intervals
for Age, Sex, and Severity
Difference in odds ratios between subgroups,
one-tailed tests except for sex
57
Odds Ratios of ITT Drug and Placebo Response
Rates in Men and Women
Nelson JC, Delucchi K, Schneider LS. Unpublished
data
58
(No Transcript)
59
Response Rates for Drug and Placebo in Men and
Women
Response gt 50 improvement on HAMD or MADRS

The odds ratio for drug vs placebo in men and
women were OR1.64 (CI 1.34, 2.00) vs 1.28
(1.10, 1.49) respectively, X23.73, df1,
p0.053. Risk difference Female 0.06 (CI
0.02, 0.10) Male 0.12 (CL 0.07, 0.17)
60
Summary of Odds Ratios and CIsfor Course, Age of
Onset, and Anxiety
Difference in odds ratios between subgroups,
one-tailed tests
61
ITT Response Rate in Early Onset (lt 60 years) and
Late Onset (gt 60 years) Patients
Nelson JC, Delucchi K, Schneider LS. Unpublished
data
62
(No Transcript)
63
Response Rates for Drug and Placebo by Age of
Onset
Response gt 50 improvement on HAMD or MADRS

The odds ratio for drug vs placebo in early and
late onset patients were OR2.03 (CI 1.49, 2.76)
vs 1.40 (1.11, 1.76) respectively, X22.98,,
df1, p0.08. Risk difference Onset lt 60 0.16
(0.09, 0.23) Onset 60 0.08 (0.03, 0.14)
64
In the 4 studies that used a low MMSE entry
criterion, the number of individuals per study
with scores less than 24 were too low, i.e. lt 10,
to provide meaningful data.
Cognitive Impairment
65
Other Moderators

• Medical illness
• Prior work suggests response declines
• Medical burden not usually assessed
• Comorbid Pain
• Severe pain associated with less good response
• Cognitive Impairment
• Executive dysfunction
• Several reports suggest ED patients do less well
  than non-ED patients
• Drug-placebo differences not reported
• Memory Impairment
• May reduce response to psychotherapy

66
Severity of Medical Comorbidity And
Antidepressant Response and Remission
80
Remission
Response
70
Prediction of Response plt0.03
60
50
Prediction of Remission plt0.02
Percent of Patients
40
30
20
10
0
0
1
2
3
4
5
6
CIRS Scores
Greater Medical Burden
Iosifescu DV et al. Am J Psychiatry
20031602122-2127
67
Baseline Pain Symptoms Predict Treatment Failure
in Depressed Patients
Odds of failing to respond to SSRI treatment by
level of baseline pain in 573 depressed patients
in primary care at 3 months. Compared with 109
patients having no pain. Response rated on SCL
20.
Odds Ratio1 If no pain present
Bair MJ, et al. Psychosom Med 20046617-21.
68
Moderators Executive Dysfunction

• Patients with late life depression and ED are
  less responsive than non-ED depressed patients
• Kalayam B, et al. Arch Gen Psychiatry
  199956713-718.
• Baldwin R, et al. Psychological Medicine. 2004
  34(1)125-36
• Alexopoulos G, et al. Biological Psychiatry 2005
  58204-210
• Sneed et al. Am J Geriatric Psychiatry
  200715553-563
• But,
• Drug-placebo differences not reported
• Not established if antidepressants are less
  effective or if major
  depression with ED is just less responsive
• Not established if all antidepressants are less
  effective

69
Problem Solving Therapy
In Older Depressed Patients with Executive
Dysfunction

• 25 Patients with MDD over 65 years of age
• HAMD gt 18
• Executive dysfunction
• Randomly assigned to treatment with problem
  solving therapy (PST) or supportive therapy (ST)
  for 12 weeks

Plt.01
Alexopoulos G, Raue P, Arean P. Am J Geriatric
Psychiatry 2003 1146-52
70
Moderators Memory Impairment Effect on
Psychotherapy Outcomes
Weekly Ham D scores n24
Scott Mackin, unpublished data
71
Summary

• Depression is common in late life
• Depression adversely affects quality of life,
  functioning, and medical illness
• Depression in late life is often not treated or
  not adequately treated
• Depression responds to treatment

72
Summary

• Psychotherapy effective but few studies are well
  controlled for effects of supportive clinic
  visits
• Antidepressants effective effects are greater
  with 10-12 weeks of treatment
• Nonspecific factors are important for response
• Attention, Education, Activation, and Reassurance
  

73
Questions