REGULAR ARTICLES

The Influence of Age on the Response of
Major Depression to Electroconvulsive
Therapy
A C.O.R.E. Report
M. Kevin O’Connor, M.D., Rebecca Knapp, Ph.D.
Mustafa Husain, M.D., Teresa A. Rummans, M.D.
Georgios Petrides, M.D., Glenn Smith, Ph.D.
Martina Mueller, M.S., Karen Snyder, B.S.
Hilary Bernstein, M.S., A. John Rush, M.D.
Max Fink, M.D., Charles Kellner, M.D.

As part of a C.O.R.E., multi-site longitudinal study comparing continuation electroconvulsive therapy (ECT) vs. continuation pharmacotherapy, the authors determined
the response of 253 patients with major depression to acute-phase, bilateral ECT by
use of the 24-item Hamilton Rating Scale for Depression. Remission rates for three
age-groups, Ն65 years; 46–64 years; and Յ45 years, were 90 percent, 89.8 percent,
and 70 percent, respectively. Age, as a continuous variable, positively influenced response to treatment. Bilateral, dose-titrated ECT is a highly effective acute treatment
for major depression, and older age confers a greater likelihood of achieving remission. (Am J Geriatr Psychiatry 2001; 9:382–390)

M

ajor depressive disorder (MDD) causes significant
morbidity and mortality across the age continuum,1 but may be most problematic in elderly patients,
where it often goes unrecognized and undertreated.2,3
In addition to its profoundly negative impact on quality
of life and the associated risk of suicide, major depression may contribute to premature mortality in elderly
patients because it diminishes adherence to necessary
medical treatment and compounds the greater burden
of medical illness borne by those over 65. Electroconvulsive therapy (ECT) is an effective treatment modality

that is commonly used in geriatric patients with major
depression.4
Controversy has existed regarding whether or not
ECT is as effective and well tolerated in elderly patients
as it is in younger ones. To summarize, the retrospective
studies5–9 contradict one another on this issue. Two recent prospective trials,10,11 however, have found at least
equivalent response rates between these age-groups.
The purpose of the present article is to examine
the influence of age on response to an acute course of
ECT. This work is based on results from 253 depressed

Received April 5, 2001; revised May 25, 2001; accepted June 4, 2001. From the Mayo Clinic, Rochester, Minnesota. Address correspondence to
Dr. O’Connor, Mayo Clinic, 200 First Street SW, Rochester, MN 55905.
Copyright ᭧ 2001 American Association for Geriatric Psychiatry

382

Am J Geriatr Psychiatry 9:4, Fall 2001

O’Connor et al.
patients participating in the CORE Continuation ECT vs.
Pharmacotherapy Trial, in which standardized assessments and uniform ECT parameters were utilized. Specifically, we sought to confirm our hypothesis that age
confers a greater responsivity to acute-phase ECT in major depression.

METHODS
Design Overview
This study protocol was reviewed and approved by
the Institutional Review Board of all four sites. Written
informed consent was obtained from each participant.
The Consortium for Research in ECT (CORE) Continuation ECT vs. Pharmacotherapy Trial, begun in February
1997, is an ongoing, 5-year, multicenter, NIH-funded
trial of the efficacy and safety of continuation ECT in
patients with MDD. In the first phase of the trial (Phase
I), patients receive acute ECT three times per week;
patients who meet strict Remitter criteria and who remain remitted for 1 week are randomly assigned to receive continuation ECT or continuation pharmacotherapy (nortriptyline plus lithium; Phase II) and followed
for 6 months to evaluate relapse. The 24-Item Hamilton
Rating Scale for Depression (24-item Ham-D)12 is administered 1–3 days after each ECT treatment. The primary
outcome for Phase I is remission status and for Phase II
is relapse status. The participating clinical sites are the
Medical University of South Carolina (MUSC) in Charleston, SC; Hillside Hospital North Shore/Long Island Jewish Health System (New York); Mayo Clinic (Rochester,
MN), and University of Texas Southwestern Medical
Center (Dallas).
Patients were included if they met criteria for major
depression as the principal diagnosis on the basis of the
structured interview for DSM-IV Axis I diagnoses (SCIDIV)13 and had a score of 20 or higher on the 24-item
Ham-D. Patients with the following diagnoses were excluded on the basis of history, clinical examination, and
performance on the SCID: schizoaffective disorder,
schizophrenia, bipolar disorder, substance dependence
within the past 12 months, neurologic disease or brain
injury, preexisting delirium, comorbid medical conditions contraindicating ECT or Nor-Li administration, and
previous treatment failure in the index episode on the
combination of a heterocyclic antidepressant and lithium. Patients thought likely to have dementia were ex-

Am J Geriatr Psychiatry 9:4, Fall 2001

cluded on the basis of a Mini-Mental State Exam
(MMSE)14 score below 20. Among the group of 326 subjects screened for participation (i.e., assigned a study
number) before January 2000, 73 were eliminated because they met exclusion criteria (16%, MDD not primary diagnosis; 12%, 24-item Ham-D Ͻ20; 14%, delirium, dementia, amnestic disorders, or substance
dependence within 12 months; 12%, contraindicated
comorbid general medical conditions; 8%, refused informed consent; 7%, neurologic disease or brain injury;
31%, other exclusion categories or category not specified).
Study Sample
As of January 1, 2000, 253 subjects had participated
in the acute ECT treatment phase (Phase I). Of these
253 patients, 217 completed Phase I as prescribed by
the protocol (Completer sample). The patients who
dropped out (premature exits) withdrew for a variety
of reasons including refusal of further treatments (8/
36), psychiatric or other intercurrent medical events (4/
36), treatment side effects (16/36; confusion and memory impairment, headaches, nausea, anxiety), and
protocol violations (8/36). The final mean 24-item HamD score for those who exited prematurely was
19.2‫( 9.6ע‬n‫ ,)63ס‬vs. 8.1‫( 3.6ע‬n‫ )712ס‬for the study
Completers.
ECT Procedures
The Thymatron DGX ECT device (Somatics, Inc.,
Lake Bluff, IL) was used for the administration of ECT.
A standardized dose titration method was used to determine seizure threshold. All treatments were administered using the bi-temporal electrode placement. After
seizure threshold was determined at the first treatment
session, all subsequent treatments were delivered at 1.5
times this energy setting.
Definition of Remitter Status
For Phase I, the primary outcome is remission
status, defined as a decrease from baseline in 24-item
Ham-D scores of at least 60% and a final 24-item Ham-D
of 10 or less on two consecutive occasions. Treatment
Responders were those who showed at least at 50%
decrease from baseline in 24-item Ham-D scores, regardless of final 24-item Ham-D score. Non-Remitters

383

Age and Response to ECT
were patients given at least 10 treatments without
achieving remission status.
Analysis Sets
We carried out all analyses with two analysis sets:
a) Completers (n‫ ,)712ס‬comprising all subjects who
either remitted or remained on study through the required 10 ECT treatment sessions without achieving full
remission status (Non-Remitters); b) full analysis set
(analogous to an intent-to-treat sample), comprising all
subjects who agreed to participate in the study and who
had at least one ECT and one post-baseline measurement (N‫.)352ס‬
Statistical Analyses
The primary dependent variables were remission
status and 24-item Ham-D scores; the primary independent variable was patient age. In separate analyses, age
was treated as a continuous variable and as a categorical
variable having three categories: 18–45, 46–64, and 65–
85. This categorization was selected to equalize approximately the sample size across the groups and because
it can be considered clinically as roughly corresponding
to a younger group, a middle-aged group, and an elderly
group.
Analyses were repeated with the oldest age category separated into Young-Old (65–74 years) and OldOld (Ͼ74 years). The results across all categories of
comparison for the oldest-old subjects were similar to
those for the young-old; hence, results will be reported
only for the combined older category (65–85 years). To
evaluate the comparability of the age-groups with respect to baseline and demographic characteristics, we
used one-way analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. For the dichotomous outcome (Remitter/Nonremitter), chi-square analyses were used to evaluate the
simple (unadjusted) relationship of age-group to the dichotomous outcome (Remitter/Nonremitter).
We compared end-of-treatment 24-item Ham-D (final 24-item Ham-D) and change from baseline in total
24-item Ham-D scores across age-groups by use of oneway ANOVA, followed by post-hoc pairwise comparisons (Tukey’s Honestly Significant Difference [HSD]
procedure), if warranted. We also carried out multivariate regression analyses, using change in 24-item HamD as the dependent variable, age as the primary inde-

384

pendent variable, with psychosis status (psychotic
depression vs. nonpsychotic depression), initial severity
(24-item Ham-D at baseline), age at onset of depression,
number of psychiatric hospitalizations, and number of
previous episodes as additional candidate covariables to
evaluate the relationship between change in 24-item
Ham-D and age-adjusted for the potential confounding
or moderator effects of the covariables.
All significance tests were two-tailed and used a significance level of 0.05.
Missing data. For analyses involving the full analysis
set (N‫ ,)352ס‬we assigned to premature exits (dropouts) the worst-case value of Nonremitter. For the endof-treatment 24-item Ham-D analyses, we considered
the final value observed for 24-item Ham-D total score
before exit as the observed end-of-treatment value for
premature exits. The average end-of-treatment 24-item
Ham-D for premature exits was not significantly different across the three age-groups (Young: 20.5‫ ;4.8ע‬Middle: 17.5‫ ;0.5ע‬Older: 17.7‫ ;1.4ע‬P‫ ;74.0ס‬one-way ANOVA), and the reasons for premature exits were similar
across these groups.

RESULTS
Baseline Demographic, Clinical,
and Treatment Characteristics
Table 1 presents demographic, clinical, and treatment characteristics for the total sample and for the
three age-groups. The average age of the total sample
(N‫ )352ס‬was 56‫ 2.61ע‬years, with 66% of the sample
being women and 90.1%, white, reflecting the demographics of the U.S. population who currently are referred for ECT. The proportion of patients with psychotic depression was 30% for the total group, with
24.1% in the youngest group vs. 35.5% in the oldest
group (P‫ ;72.0ס‬chi-square test). The age at onset of
depression was 40.8‫ 7.91ע‬years for the total group and
ranged from 25.6‫ 5.9ע‬years for the youngest group to
57.5‫ 8.71ע‬years for the oldest group. On average, the
youngest group had significantly more previous episodes of depression (3.6‫ )9.6ע‬compared with the middle (2.4‫ )5.3ע‬and the older groups (1.6‫;4.2ע‬
P‫ ;130.0ס‬one-way ANOVA) with 55.1% of the youngest
reporting more than two previous episodes, vs. 45% of
the middle and 38.8% of the oldest groups. Approxi-

Am J Geriatr Psychiatry 9:4, Fall 2001

O’Connor et al.
mately 29% of the middle and older groups reported no
previous episodes of depression, vs. 23% of the youngest group (P‫ ;546.0ס‬chi-square test). The average number of psychiatric hospitalizations for any psychiatric
illness (young: 2.6‫ ;0.2ע‬middle: 2.3‫ ;1.2ע‬older:
2.2‫ )5.1ע‬and the proportion with more than two psychiatric hospitalizations (young: 62.5%; middle: 57.7%;
older: 63.7%) were similar across the age-groups. The
average number of ECT treatments received by the total
sample was 7.8‫ ,3.3ע‬and was similar across the agegroups (young: 7.5‫ע‬SD3.3; middle: 8.0‫ ;4.3ע‬older:
7.7‫ ;3.3ע‬P‫ ;86.0ס‬one-way ANOVA). The severity of
illness at baseline was similar for the young, middle, and
older age-groups, with mean 24-item Ham-D baseline
scores of approximately 35 across all groups (P‫;49.0ס‬
one-way ANOVA).
TABLE 1.

Remission Status
For the full analysis set (treating premature exits as
Nonremitters), 75% of all patients reached full remission
status (Table 2), requiring an average of 7.8‫ 3.3ע‬ECT
treatments. The rate of remission was higher in both
the middle (86%) and older age-groups (80%) compared
with the youngest group (57%; P‫ ;4000.0ס‬chi-square
test). The proportion achieving full remission status was
even more pronounced among those who completed
the acute phase (n‫ %78 ;)712ס‬of all Completers
reached full remission criteria, requiring an average of
approximately eight ECT treatments (approximately 2.5
weeks of treatment). The differential remission rate
across age-groups was also evident among study Completers, with approximately 90% in the middle and old-

Demographic and baseline clinical and treatment characteristics for the total sample and for Young (18–45), Middle
(46–64), and Older (65–85) age categories

Variable

Total Sample
(N؄253)

18–45
(n؄79)

46–64
(n؄81)

65–85
(n؄93)

Age, years

P

56.2 (‫)2.61ע‬

37.3 (‫)1.7ע‬

54.7 (‫)4.5ע‬

73.8 (‫)6.5ע‬

Gender, % women

66.4%
(168/253)

66.6%
(51/79)

64.2%
(52/81)

70%
(65/93)

0.669a

Psychosis, % psychotic

30.4%
(77/253)

24.1%
(19/79)

30.9%
(25/81)

35.5%
(33/93)

0.266a

90.1%
(228/253)
6.7%
(17/253)
3.2%
(8/253)

86.1%
(68/79)
10.1%
(8/79)
3.8%
(3/79)

87.7%
(71/81)
6.2%
(5/81)
6.2%
(5/81)

95.7%
(89/93)
4.3%
(4/93)
0%
(0/93)

0.10b

Age at illness onset, years

40.8 (‫)7.91ע‬

25.6 (‫)5.9ע‬

37.1 (‫)4.41ע‬

57.5 (‫)8.71ע‬

Ͻ0.0001c

n previous episodes
% no previous episodes

2.5 (‫)6.4ע‬
27.4%
(61/223)
45.7%
(102/223)

3.6 (‫)9.6ע‬
23.2%
(16/69)
55.1%
(31/69)

2.4 (‫)5.3ע‬
29.0%
(20/69)
44.9%
(38/69)

1.6 (‫)4.2ע‬
29.4%
(25/85)
38.8%
(52/85)

0.031c
0.645

Length of current episode, years

0.95 (‫)9.1ע‬

0.96 (‫)0.2ע‬

0.88 (‫)5.1ע‬

1.00 (‫)1.2ע‬

0.905c

n psychiatric hospitalizations

2.4 (‫)9.1ע‬
(n‫)742ס‬
12.6%
(31/247)
61.5%
(152/247*)

2.6 (‫)0.2ע‬
(n‫)87ס‬
12.8%
(10/78)
62.5%
(48/78)

2.3 (‫)1.2ע‬
(n‫)87ס‬
14.1%
(11/78)
57.7%
(45/78)

2.2 (‫)5.1ע‬
(n‫)19ס‬
11.0%
(10/91)
63.7%
(58/91)

0.393c

7.8 (‫)3.3ע‬

7.5 (‫)3.3ע‬

8.0 (‫)4.3ע‬

7.7 (‫)3.3ע‬

0.681c

125.0 (‫)0.66ע‬

86.2 (‫)8.93ע‬

125.5 (‫)5.85ע‬

157.2 (‫)1.27ע‬

Ͻ0.0001c

Race, %
White
Black
Other

% Ն two previous episodes

No psychiatric hospitalizations, %
Ն two psychiatric hospitalizations, %
Total n ECT treatments
Seizure threshold, in milliCoulombs

0.130

0.828d
0.695d

Note: Values are mean (‫ע‬standard deviation) unless otherwise specified; ns do not always total 253 because of missing values for the
variable.
a
From a 3 ‫( 2 ן‬Psychosis: Age-Group ‫ ן‬Psychosis status; Gender: Age-Group ‫ ן‬Gender) chi-square test.
b
From a 3 ‫( 3 ן‬Age ‫ ן‬Race) chi-square test.
c
From a one-way analysis of variance comparing means across age categories.
d
From a 3 ‫( 2 ן‬Age ‫ ן‬Hospitalization) chi-square test.

Am J Geriatr Psychiatry 9:4, Fall 2001

385

Age and Response to ECT
est age-groups remitting, compared with 75% in the
youngest groups (Table 2; P‫ ;300.0ס‬chi-square test).
84% of all patients (N‫ )352ס‬experienced at least a 50%
reduction from baseline in 24-item Ham-D scores (treatment Responders), with the proportion of Responders
ranging from 68% in the youngest group to approximately 90% in the middle and older age-groups
(PϽ0.0001; one-way ANOVA; Table 2).
Change in 24-item Ham-D. As shown in Figure 1, patients in all age-groups showed, on average, greater than
a 20-point decrease in 24-item Ham-D scores over the
course of acute-phase treatment. However, patients in
the middle and older age-groups showed a greater decrease in 24-item Ham-D total scores over the acute
treatment course than patients in the youngest group.
Those in the youngest group had an average final 24item Ham-D of 12.7‫ ,3.9ע‬compared with an average
TABLE 2.

final 24-item Ham-D of 8.5‫ 2.6ע‬for the older groups
(PϽ0.0001; one-way ANOVA). Considering age as a continuous variable, the association between age and
change in 24-item Ham-D was positive and highly significant (P‫ .)8900.0ס‬This effect of age on 24-item HamD outcome remained significant in multiple-regression
analyses after adjustment for psychosis status, initial severity (24-item Ham-D baseline), and number of previous episodes (multiple-regression analyses beta-coefficient for age‫ ;90.0ס‬P‫ ;3100.0ס‬model r2‫.)25.0ס‬
Additional regression analyses that included clinical site
and clinical site ‫ ן‬age interaction terms did not indicate
a significant clinical site effect or a differential effect of
age across clinical sites.
Effect of presence of psychosis on age ‫ ן‬treatment response relationship. Overall, illness severity at baseline as measured by total 24-item Ham-D appears to be

Comparison of treatment response across Young (18–45), Middle (46–64), and Older (65–85) age-groups
Age-Groups

Response

Overall

18–45

46–64

65–85

P

Remitters, % (Completers; n‫)712ס‬

87.1%
(189/217)

75.0%
(45/60)

93.3%
(70/75)

90.2%
(74/82)

0.003e

Remitters, %a,b (full analysis set; N‫)352ס‬

74.7%
(189/253)

57.0%
(45/79)

86.4%
(70/81)

80.0%
(74/93)

0.0004e

Ն50% reduction in Ham-D, %c
(full analysis set; N‫)352ס‬

83.8%
(212/253)

68.4%
(54/79)

91.4%
(74/81)

90.3%
(84/93)

Ͻ0.0001e

35.0 (‫)0.7ע‬h

35.0 (‫)7.6ע‬

34.8 (‫)8.6ע‬

35.2 (‫)5.7ע‬

0.942g

Ham-D end (full analysis set; N‫)352ס‬

9.6 (‫)4.7ע‬h

12.7 (‫)3.9ע‬

8.0 (‫)6.5ע‬

8.5 (‫)2.6ע‬

Ͻ0.0001g

Ham-D changed (full analysis set; N‫)352ס‬

25.4 (‫)8.9ע‬h

22.3 (‫)9.9ע‬

26.8 (‫)9.8ע‬

26.6 (‫)8.9ע‬

0.003f

34.1 (‫)2.6ע‬
(n‫)06ס‬
36.8 (‫)8.7ע‬
(n‫)91ס‬

33.7 (‫)0.7ע‬
(n‫)65ס‬
37.3 (‫)7.5ע‬
(n‫)52ס‬

33.2 (‫)0.6ע‬
(n‫)06ס‬
38.8 (‫)8.8ע‬
(n‫)33ס‬

0.581g

Ham-D end
Nonpsychotic
Psychotic

12.3 (‫)1.9ע‬
13.7 (‫)1.01ע‬

8.6 (‫)0.6ע‬
6.5 (‫)4.4ע‬

9.2 (‫)5.6ע‬
7.3 (‫)6.5ע‬

0.014g
0.001g

Ham-D changed
Nonpsychotic
Psychotic

22.1 (‫)1.01ע‬
23.1 (‫)4.9ע‬

25.1 (‫)2.9ע‬
30.8 (‫)0.7ע‬

24.0 (‫)1.8ע‬
31.5 (‫)8.01ע‬

0.204g
0.006g

a

Ham-D baseline (full analysis set; N‫)352ס‬

Comparison by psychosis status
Ham-D baseline
Nonpsychotic
Psychotic

0.615g

Note: Values are mean (‫ע‬standard deviation) unless otherwise specified.
a
Remitter‫ס‬Ն50% reduction in Ham-D and final Ham-D Յ10.
b
For full analysis set, premature exits are considered Non-Remitters.
c
Percent reduction‫ס‬Ham-D change from baseline (Baseline-to-Final)/Ham-D baseline; last observed value used for dropouts.
d
Ham-D change from baseline (baseline to end of acute phase value).
e
From chi-square test.
f
From one-way analysis of variance comparing means across three Age categories. Pooled t-test Bonferroni-corrected P-values: Young vs.
Middle, P‫ ;900.0ס‬Young vs. Old, P‫ ;210.0ס‬Middle vs. Older, PϾ0.9.
g
From one-way analysis of variance comparing means across three Age categories.
h
Mean (‫ע‬standard deviation).

386

Am J Geriatr Psychiatry 9:4, Fall 2001

O’Connor et al.
slightly higher for the psychotic-depressed than for the
nonpsychotic-depressed diagnostic groups, although
the disparity did not reach statistical significance (Table
2). Despite the higher baseline levels, the final 24-item
Ham-D values for the middle and older psychotic-depressed groups are lower than for the nonpsychoticdepressed group, giving a higher change from baseline
score for these groups. Although the reduction in 24item Ham-D score is large (Ͼ22 points for all ages)
among the nonpsychotic-depressed patients, the most
pronounced effect of an acute ECT treatment course is
seen among the older psychotic-depressed patients.
The middle and older groups experienced, on average
Ͼ30-point improvement in 24-item Ham-D over the
treatment course, compared with a 23-point improvement for the youngest group (Table 2; P‫ ;600.0ס‬oneway ANOVA; with middle and older groups significantly
higher than young; PϽ0.05; Tukey’s HSD post-hoc procedure). The trend toward increasing improvement
FIGURE 1.

with increasing age does not hold for the nonpsychoticdepressed groups (P‫ ;402.0ס‬one-way ANOVA; Table 2).

DISCUSSION
In 1984, a retrospective chart review authored by Karlinsky and Shulman5 examined their experience in administering ECT to a heterogeneous group of elderly
patients with various diagnoses. They concluded that
ECT in this population was not as effective and caused
greater morbidity than in the younger age-group. Burke
et al.,6 in another retrospective study, found that the
rate of remission was lower, and the complication rate,
particularly for cardiovascular events, was significantly
higher in patients over the age of 60 who were given
ECT. Cattan et al.7 performed a retrospective review of
the response to ECT in a group of 39 patients over the
age of 80 and compared it with that of 42 patients be-

Comparison of mean baseline, final, and change from baseline 24-item Ham-D across Young (18–45), Middle (46–64),
and Older (65–85) age-groups

Note: N‫ 352ס‬patients in the acute ECT treatment phase. Ham-D‫ס‬Hamilton Rating Scale for Depression. Baseline Ham-D: one-way ANOVA;
F[2, 252]‫ ;60.0ס‬P‫ .249.0ס‬Final Ham-D: one-way ANOVA; F[2, 252]‫ ;4.01ס‬P‫ .1000.0ס‬Post-hoc comparisons: Young vs. Middle, Bonferronicorrected P‫ ;3000.0ס‬Young vs. Old, Bonferroni-corrected P‫ ;200.0ס‬Middle vs. Old, Bonferroni-corrected P‫ .9.0ס‬Change from baseline: oneway ANOVA; F[2, 252]‫ ;8.5ס‬P‫ .300.0ס‬Post-hoc comparisons: Young vs. Middle, Bonferroni-corrected P‫ ;900.0ס‬Young vs. Old, Bonferronicorrected P‫ ;210.0ס‬Middle vs. Old, Bonferroni-corrected P‫.9.0ס‬

Am J Geriatr Psychiatry 9:4, Fall 2001

387

Age and Response to ECT
tween the ages of 65 and 80. They found that those over
the age of 80 responded less well to treatment and had
significantly more adverse outcomes, such as cardiovascular complications and falls. Response to treatment
was judged from clinical commentary in the patients’
records.
In contrast to these results, a number of retrospective studies have indicated more positive outcomes in
elderly patients. Tomac et al.8 examined the Mayo Clinic
experience in ECT treatment of 34 patients over the age
of 85. Response to treatment, based on chart reviews,
and non-blind 24-item Ham-D ratings performed during
the course of treatment revealed that ECT was safe and
effective with a variety of depressive disorders in this
age-group. The most common complication was transient post-ECT confusion. Likewise, Gormley et al.9
found ECT to be safe and effective in patients over the
age of 75, with 85 percent of their subjects obtaining
marked or moderate improvement. The most common
adverse events were confusion and hypomania, which,
in virtually all cases, resolved within 2 weeks of treatment completion.
In the first of two prospective comparison, Wilkinson et al.10 compared the response to ECT (as defined
by a 50% reduction in the Montgomery-Asberg Depression Rating Scale) of four age-groups with DSM-III-R
MDD and found that the response rate was 72 percent
in patients over the age of 75 and only 54 percent in
the younger age-group. Of particular note, they observed a concomitant increase in all age categories in
MMSE performance 72 hours after completion of treatment.
In a well-designed prospective study using standardized diagnostic criteria, near-uniform treatment
across age-groups, and carefully measured outcome criteria, Tew et al.11 reported the comparative response of
268 patients, separating them into categories of “oldold” (age Ͼ75 years), “young-old” (age 60–74), and
“adults” (age Յ59). They measured the 24-item Ham-D
response of these age-groups to dose-titrated unilateral
or bilateral ECT administered at 21⁄2 times the determined seizure threshold. The response of the old-old
subjects was intermediate (67%) between that of adults
(54%) and young-old (73%). It is noteworthy that the
younger patients had a longer length of illness and were
more resistant to heterocyclic antidepressant medication than either of the older age cohorts, two factors
associated with diminished responsivity to ECT in the
studies of Sackeim et al.15 and Prudic et al.16

388

The most intriguing finding of our study is the superior antidepressant response of the elderly patients
to ECT as compared with the younger population—90%
vs. 64%. One possible explanation may be that the influence of age on response to ECT is confounded by the
presence of psychosis. The observed superior response
of the elderly patients to ECT may be a function of the
higher percentage of patients with psychotic depression seen in the older group. Among the nonpsychotically-depressed patients, the response to ECT, although
high for all age-groups, was not significantly affected by
age. Among the psychotically-depressed patients, however, there was a strong and statistically significant positive association between age and response to treatment.
This differential effect of age on ECT response for psychotic and nonpsychotic patients may partially explain
previously reported conflicting results. That is, failure
to find an age effect in some studies may be attributable
to the particular mix of psychotic- and nonpsychoticdepressed patients in a given sample of study participants. However, the presence of psychosis alone may
not account completely for the higher response to ECT
among the older age-groups. The earlier age at onset
and the greater number of lifetime episodes observed
in our younger age-groups, two factors generally
thought to reflect the severity of depression, may be
additional factors contributing to the different responsivity reported here.
There are several limitations to this study. We did
not assess Axis II psychopathology in the three agegroups. Personality disorder has been associated with
diminished responsivity to antidepressant treatment
and theoretically could have influenced treatment outcome. Although efforts were made to assess the relative
resistance to previous antidepressant drug therapy in
these populations, these data are incomplete and could
not be analyzed at this time. They may appear in future
C.O.R.E. reports. The most important limitation pertains to the measuring of treatment side effects. Because
the focus of this NIMH-sponsored investigation was on
continuation treatment, we conducted limited assessment of adverse events during Phase I, aside from the
monitoring of the patients by the clinicians caring for
them. However, the dropout rates across all three
groups due to treatment side effects were statistically
equivalent.
Taken together with the study of Tew et al.,11 two
well-designed, prospective trials involving 521 patients
with MDD have now shown that there is a statistically

Am J Geriatr Psychiatry 9:4, Fall 2001

O’Connor et al.
significant superior response of older patients with major depression to ECT. For those over the age of 65 completing this study, the rate of remission was 90%, with
those in the old-old group (over 75 years of age) having
rates similar to those in the 65–74 group. This remission
rate compares favorably with most treatments given in
medicine.
There are important differences between our study
and that of Tew et al.;11 68% of their patients received
unilateral treatment delivered at 21⁄2 times the determined seizure threshold. We treated all our patients
with bilateral electrode placement at 11⁄2 times seizure
threshold; 39% of the patients (87/223) who, they reported, were initially given unilateral treatments failed
to respond to a minimum of five such treatments and
required conversion to bilateral lead placement. In the
Tew study, parameters of treatment were selected by
the clinicians caring for these patients, and the end result of choosing such parameters was that 39% required
an average of approximately 12.8 treatments and did
not achieve as high a rate of remission as those who
participated in our study, who were given an average of
7.4 treatments. Thus, technical treatment issues substantially influenced the length and expense of treatment. Solely from the standpoint of efficacy, our study
would argue for the use of bilateral treatment at 11⁄2
times seizure threshold. Overall, these parameters resulted in a significantly higher remission rate and fewer
treatments to achieve remission status.
A number of studies have suggested that the neurobiology of depression may be different in elderly patients. Musetti et al.17 and Brodaty et al.18 have reported

that elderly patients with depression are less likely to
have a positive family history of mood disorder, suggesting that genetic factors are less important in geriatric depression. Alexopoulos et al.19 have postulated
that there is a subset of geriatric patients who experience major depression as a result of cerebrovascular disease, particularly if it disrupts prefrontal brain circuitry.
Coffey et al.20 have found more diffuse or confluent subcortical hyperintensities on brain MRI scans in geriatric
MDD patients, as compared with younger patients, with
MDD. Finally, Fink21 has hypothesized that ECT is effective through its stimulatory effect on the hypothalamus
and by causing repetitive secretion of regulatory neuropeptide such as corticotropin-releasing factor; Fink
(in an unpublished letter, 2001) also posited that neuroendocrine dysfunction is possibly more common in
elderly patients. The findings in this study of the greater
frequency of psychotic features and a substantially
higher responsivity to ECT in older patients as compared with younger patients, would also support a different pathophysiology of depression in elderly patients. Further studies designed to examine the basic
biology of a positive therapeutic response to ECT in
older patients as compared with that of younger patients would add to our understanding of the biology of
depression in elderly patients.

The research reported in this article was supported by a grant from the National Institutes of
Health (MH 55484) awarded to Consortium on Research in ECT.

References
1. Wells KB, Stewart A, Hays RD, et al: The functioning and wellbeing of depressed patients. JAMA 1989; 262:914–919
2. NIH Consensus Development Panel on Depression in Late Life:
Diagnosis and Treatment of Depression in Late Life. JAMA 1992;
268:1018–1024
3. Lebowitz BD, Pearson JL, Schneider LS, et al: Diagnosis and Treatment of Depression in Late Life: Consensus Statement Update.
JAMA 1997; 278:1186–1190
4. American Psychiatric Association: The Practice of Electroconvulsive Therapy: Recommendations for Treatment, Training, and
Privileging. Washington, DC, American Psychiatric Association,
1990, pp 71–72
5. Karlinsky H, Shulman KI: The clinical use of electroconvulsive
therapy in old age. J Am Geriatr Soc 1984; 32:183–186
6. Burke WJ, Rubin EH, Zorumski CF, et al: The safety of ECT in
geriatric psychiatry. J Am Geriatr Soc 1987; 35:516–521
7. Cattan RA, Berry PP, Mead G, et al: Electroconvulsive therapy in
octogenarians. J Am Geriatr Soc 1990; 38:753–758
8. Tomac TA, Rummans TA, Pileggi TS, et al: Safety and efficacy of

Am J Geriatr Psychiatry 9:4, Fall 2001

electroconvulsive therapy in patients over age 85. Am J Geriatr
Psychiatry 1997; 5:126–130
9. Gormley N, Cullen C, Walters L, et al: The safety and efficacy of
electroconvulsive therapy in patients over age 75. Int J Geriatr
Psychiatry 1998; 13:871–874
10. Wilkinson AM, Anderson DN, Peters S: Age and the effects of
ECT. Int J Geriatr Psychiatry 1993; 8:401–406
11. Tew JD, Mulsant BH, Haskett RF, et al: Acute efficacy of ECT in
the treatment of major depression in the old-old. Am J Psychiatry
1999; 156:1865–1870
12. Hamilton M: Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol 1967; 6:278–296
13. First MB, Spitzer RL, Gibbon M, et al: Structured Clinical Interview for DSM-IV Axis I Disorders–Patient Edition (SCID–I/P, Version 2.0). Biometrics Research Department, February 1997 (ECT
Study Version)
14. Folstein MF, Folstein SE, McHugh PR: Mini-Mental State: a practical method for grading the cognitive state of patients for the
clinician. J Psychiatr Res 1975; 12:189–198

389

Age and Response to ECT
15. Sackeim HA, Prudic J, Devanand DP, et al: The impact of medication resistance and continuation pharmacotherapy on relapse
following response to electroconvulsive therapy in major depression. J Clin Psychopharmacol 1990; 10:96–104
16. Prudic J, Haskett RF, Mulsant B, et al: Resistance to antidepressant
medications and short-term clinical response to ECT. Am J Psychiatry 1996; 153:985–992
17. Musetti L, Perugi G, Soriani A, et al: Depression before and after
age 65: a re-examination. Br J Psychiatry 1989; 155:330–336

390

18. Brodaty H, Peters K, Boyce P, et al: Age and depression. J Affect
Disord 1991; 23:137–149
19. Alexopoulos GS, Meyers BS, Young RC, et al: Vascular depression
hypothesis. Arch Gen Psychiatry 1997; 54:915–922
20. Coffey CE, Figiel GS, Djang WT, et al: Subcortical hyperintensity
in MRI: a comparison of normal and depressed elderly subjects.
Am J Psychiatry 1990; 147:187–189
21. Fink M: Electroshock revisited. American Scientist March–April
2000; 88:162–167

Am J Geriatr Psychiatry 9:4, Fall 2001