Electroconvulsive Therapy
in Old-Old Patients
David T. Manly, M.D.
Stanley P. Oakley Jr., M.D.
Richard M. Bloch, Ph.D.

The authors compared falls, cardiovascular factors, confusion, gastrointestinal, pulmonary, and metabolic side effects for “old-old” (Ͼ75 years) patient groups treated
with either electroconvulsive therapy (ECT) or pharmacotherapy. A subset of a pharmacotherapy patient group was selected to match for age, sex, and diagnosis in a
case-control design. Side effects were recorded from each selected patient’s medical
record and compared between groups. Patients receiving ECT showed fewer cardiovascular and gastrointestinal side effects. Patients receiving ECT had longer lengths
of stay and more favorable outcomes. Overall, there was a tendency for ECT to result
in fewer side effects and better treatment outcomes. ECT appears to be relatively safe
and more effective than pharmacotherapy for major depressive disorders in old-old
patients. (Am J Geriatr Psychiatry 2000; 8:232–236)

M

ajor depression is a common and treatable cause
of morbidity and mortality in elderly patients that
is often underdiagnosed and undertreated in primary
care settings and nursing homes. Studies in the past
have shown that the use of electroconvulsive therapy
(ECT) in major depression increases with age and that
patients over 65 receive a disproportionately high share
of ECT compared with younger patients.1 There are
controversies over the safety of ECT in elderly and medically ill patients, and questions have been raised as to
whether it is more effective than pharmacotherapy. As
the population in the United States continues to age,
the use and safety of ECT in elderly patients will be an
increasingly important clinical issue. The most rapidly
growing subgroup of our population is the “old-old”,
generally defined as 75 years of age and older. Population projections for the period between 1980 and 2040

estimate the general population will grow 41%, whereas
those over 65 will grow 160%, and those 75 and older
will grow 268%.2
Most studies of the treatment of depression have
included patients of all ages and have often actually excluded the elderly patient with medical complications.
Therefore, few studies have specifically addressed the
treatment of depression in elderly patients.3,4 This is
particularly true for the “old-old”, frail elderly patients,
and those with concomitant medical or neurological
disorders. A recent study addressing the use of ECT in
old-old patients retrospectively compared 39 patients
age 80 and older with 42 younger patients age 65
through 80. The study found ECT to be relatively well
tolerated, even in the older age-group. Confusion was
the most common complication in both age-groups.
The older age-group had more cardiovascular problems

Received April 6, 1999; revised June 1, 1999; accepted August 12, 1999. From the Department of Psychiatric Medicine, East Carolina University
School of Medicine, Greenville, North Carolina. Address correspondence to Dr. Manly, Park Ridge Hospital, P.O. Box 1569, Fletcher, NC 28732.
Copyright ᭧ 2000 American Association for Geriatric Psychiatry

232

Am J Geriatr Psychiatry 8:3, Summer 2000

Manly et al.
and falls, reflecting higher medical acuity, as measured
by American Society of Anesthesia (ASA) scores, number of medical problems, and number of cardiac medications. Outcome was better in the younger group, but
the older patients did well, with at least moderate improvement noted in 85%.5
The current study extends these findings by directly
comparing two groups of old-old patients suffering
from major depression. One of these age- and sexmatched groups was treated pharmacologically, and the
other was treated with ECT. We compared therapeutic
outcomes and complications.

METHODS
We conducted a retrospective chart review of inpatients
at a university medical center with both private and
academic psychiatric services. Charts were selected
from the years 1987 to 1993 for patients 75 years of age
and older whose discharge diagnoses included a major
affective disorder (unipolar or bipolar depression) and
who had received ECT. From these, data were obtained
regarding age, gender, medical diagnoses, number and
type of medications, number and laterality of ECT (if
available), complications, and outcome. On the basis of
matching age, gender, and discharge diagnosis of major
affective disorder, a comparison group of records was
sequentially identified for patients who had been
treated psychopharmacologically without ECT. The
same data elements were gathered from these charts.
Diagnoses for all patients had been determined by an
attending psychiatrist with DSM-III-R criteria as recorded in the discharge notes.6 Because all charts meeting ECT and age criteria were selected for the experimental group, and “computer-applied matching
criteria” selected the comparison group, group selection bias was minimized.
ECT was administered 2 or 3 times per week by use
of a brief pulse device (Mecta SR1). Lead placements
were bilateral in 19 patients, right-unilateral in 9, both
bi- and unilateral in 9, and not noted in 2 patients.
Medical status was assessed independently by one
of the investigators (DM or SO) using the American Society of Anesthesiology (ASA) 5-point rating scale.7 Each
selected chart was reviewed for notations of complications, which were defined as any unexpected event requiring intervention or a change in treatment plan. All

Am J Geriatr Psychiatry 8:3, Summer 2000

complications were classified into one of six categories:
cardiovascular, pulmonary, neurologic (including confusion), falls, gastrointestinal, and metabolic. Complications did not include the expected physiological responses to ECT, such as transient hypertension,
tachycardia, or postictal confusion immediately after
the stimulus. The results of treatment were evaluated
by use of a global rating scale used by previous investigators.5,8,9 A good outcome was defined as a complete
resolution of symptoms with a return to premorbid
baseline and with no or only minor residual sequelae
that do not interfere with social functioning. A moderate outcome was defined as some improvement, but
with residual symptoms. A poor outcome was defined
as no improvement or worsening of condition and level
of functioning.
The frequencies of various treatment complications
and cardiac histories for each group were compared by
use of the Fisher’s exact test statistic. Other comparisons between groups, such as ASA scores or length of
hospital stay, utilized standard t-tests for independent
groups.10

RESULTS
Table 1 shows subject group characteristics. The ECT
and pharmacological control populations are well
matched for age, gender, and race. The two groups were
also medically comparable. ASA scores did not differ
between groups (t[76]‫ ;65.1ס‬P‫ .)21.0ס‬The majority of
patients in both groups had scores of 2 or 3. No patients
in either group had a score greater than 3. The average
number of medications (4.6 vs. 4.2) and average number of medical diagnoses (3.1 vs. 2.9) were also similar
between groups. However, the groups did differ in
terms of the overall length of hospital stay, with the ECT
group requiring an average of 19 more days of hospitalization than the medication-only control group
(t[76]‫ ;48.4ס‬P‫ס‬Ͻ0.001).
The complications in both groups are noted in Table 2. The only neurological complication in the ECT
group was confusion, which turned out to be common
in the control group as well. Ten patients with ECT experienced confusion, whereas six experienced confusion in the control group (Fisher’s P‫[ 2.0ס‬NS]).
Gastrointestinal side effects were noted only in the
psychopharmacologically treated group (Fisher’s

233

ECT in Old-Old Patients
P‫ .)720.0ס‬These included nausea/vomiting, constipation, and dry mouth. All except dry mouth were associated with the use of tricyclic antidepressants.
The most common complication in the control
group was cardiovascular in nature. Six pharmacologically treated patients experienced cardiovascular complications. Interestingly, no patients in the ECT group
were noted to have cardiac complications (Fisher’s
P‫ ,)310.0ס‬although, as shown in Table 3, there was no
difference in the frequency of cardiac diagnoses. Overall, there were more complications in the control group,
with 16 patients out of 39 having complications as comTABLE 1.

pared with only 10 out of 39 in the ECT group, although
this difference did not reach statistical significance.
Table 4 shows the outcome data for the two treatment groups. Significantly more ECT patients had a
good outcome (76.9%) than the pharmacologically
treated group (33.3%; Fisher’s PϽ0.001). In contrast,
more patients in the pharmacologically treated group
achieved only a moderate outcome (56.4% vs. 23.1%;
Fisher’s P‫ )300.0ס‬whereas four of the pharmacologically treated group had a poor outcome (10.3%; Fisher’s
P‫ .)60.0ס‬It should be noted that no ECT-treated patients had a poor outcome in this study.

Subject demographics
Electroconvulsive Therapy
39

Age, years

Significance (P)

39

NS

83.69‫38.3ע‬

n

Pharmacotherapy

83.43‫62.3ע‬

NS

Gender (female/male)

36/3

34/5

NS

Race (white/black)

38/1

37/2

NS

Medications

4.64‫30.2ע‬

4.18‫75.2ע‬

NS

Medical diagnoses

2.89‫6.1ע‬

3.08‫2.1ע‬

NS

Cardiac diagnoses
ASA score

24

24

NS

42.51‫29.02ע‬

23.36‫94.11ע‬

0.001

2.28‫05.0ע‬

Length of stay

2.10‫15.0ע‬

NS

Note: ASA‫ס‬American Society of Anesthesia “medical acuity” score.

TABLE 2.

Complications

Category

Electroconvulsive Therapy

Cardiovascular

Pharmacotherapy

Significance (P)

2

Falls

4

NS
0.013

0

6

10

5

NS

Gastrointestinal

0

4

0.027

Pulmonary

1

0

NS

Metabolic

0

0

NS

Total complications

13

19

NS

Patients with complications, n

10

16

NS

Confusion/Neurologic

Note: Several patients in each group had more than one complication.

TABLE 3.

Cardiac history
Electroconvulsive Therapy

Hypertension

Pharmacotherapy

Significance (P)

19

18

NS

Arteriosclerotic cardiovascular disease

6

7

NS

Myocardial infarction

2

2

NS

Congestive heart failure

3

6

NS

Total cardiac diagnoses

30
24

33
24

NS
NS

Patients with cardiac history, n

Note: Several patients in each group had more than one cardiac diagnosis.

234

Am J Geriatr Psychiatry 8:3, Summer 2000

Manly et al.
DISCUSSION
Treatment of depression in elderly patients is frequently
complicated by comorbid medical conditions and polypharmacy. These issues are particularly prominent in
“old-old” patients, a group that has been inadequately
studied in the past. In fact, old-old patients have frequently been excluded from treatment studies because
of their age and frequent comorbidity. Our study was
undertaken to evaluate the safety and efficacy of ECT in
this group. Our hypothesis was that ECT would be
found to be more effective, with fewer side effects than
pharmacologic treatment in hospitalized old-old patients with major affective disorders. This study confirms our hypothesis and extends the result of Cattan et
al.5 Our study is particularly useful in that ECT-treated
patients were compared with an age- and gendermatched, pharmacologically treated group of comparable medical status.
Length of stay was considerably longer in the ECTtreated group, 41.5 days compared with 23.3 days in
the medically-treated control subjects. This finding is
consistent with several previous studies showing significantly longer length of stays in ECT-treated inpatients.11–14 In one study of 101 elderly (more than 65
years old) patients with depression, 46 patients who
received ECT were hospitalized for 43.7 days as compared with 55 patients not treated with ECT, who were
hospitalized for only 24.4 days.11 A recent retrospective
naturalistic study12 of geriatric patients hospitalized for
depression found that the length of stay for ECT patients
was twice that of patients treated without ECT (48 days
vs. 24 days). Furthermore, they found that when ECT
was the treatment of choice initially, the length of stay
was 40 days. However, when ECT was used secondarily,
following a medication trial, the length of stay increased
to 56 days. Yet another study13 found the mean length
of time between admission and the first ECT was 21.4
days. This suggests that significant factors in the longer
lengths of stay for ECT patients are medication trials
prior to ECT, weaning off medications, and medical evalTABLE 4.

uation and stabilization of elderly patients before ECT.
Furthermore, it has become common practice to initiate
maintenance antidepressant medication after the completion of ECT, which can further extend the length of
stay.
The incidence of complications noted in Table 2 is
in agreement with similar studies. Confusion was the
most common complication, with 10 of 39 ECT patients
(25.6%) experiencing confusion and 7.6% of patients
experiencing other medical complications. Our overall
complication rate was 33%. Mulsant et al.13 reviewed
the literature, citing 14 studies with a total of over 1,000
geriatric patients (age 65 and older) treated with ECT
and found the incidence of confusion to be about 10%
and also found 6% of patients having some other medical complications during ECT. The same investigators
performed a prospective naturalistic study of elderly inpatients and found a complication rate very similar to
ours. They found that 31% of patients suffered confusion, and 7% had some other medical complication. Cattan et al. found confusion to be the most common complication, occurring in 45% of patients age 65–80 and
in 59% of patients over the age of 80. It should be noted
that one possible factor in the discrepancy in the rate
of confusion between our study and Cattan et al.5 is that
patients in Cattan’s study were treated with a sine-wave
ECT apparatus, which is known to cause significantly
greater confusion.15,16 Individual studies show a wide
range of incidence of complications, ranging from none
to over 50%. The most common complications seem to
be cardiovascular, confusion, and falls. Our study differs
in that we had no cardiovascular complications noted
within our ECT population. As shown in Table 3, both
groups had the same frequency of cardiac diagnoses
(24/39) and so were equally at risk of cardiac side effects. The lack of cardiac complications, although surprising, may be, in part, due to the fact that we did not
include transient phenomena, which should be considered “normal” physiological concomitants of ECT, such
as transient hypertension, sinus tachycardia, or occasional PVCs. A recent retrospective study by Gormley

Outcome, n (percent)
Electroconvulsive Therapy

Good
Moderate
Poor

Pharmacotherapy

Significance (P)

30 (76.9)

13 (33.3)

0.001

9 (23.1)

22 (56.4)

0.003

0

4 (10.3)

0.06

Overall

Am J Geriatr Psychiatry 8:3, Summer 2000

0.003

235

ECT in Old-Old Patients
et al.17 also found no cardiac complications in a review
of 93 courses of ECT in 67 patients over the age of 75.
In this study we used a retrospective naturalistic
design. Several notable weaknesses are inherently possible in this design. The nonblinded nature of the datagathering could have resulted in bias in the assessment
of the outcome. Furthermore, the difference in length
of stay could have led to an overestimation of the response rate to ECT due to the fact that patients were
evaluated at different points in time. Also, when data
are collected retrospectively, the evaluation of complications depends on the documentation of the attending
physician at the time of occurrence. Subtle biases in the
recognition or documentation of side effects of ECT
could, therefore, account for some of the findings in the
current study. Patient selection biases must also be considered. Patients viewed as likely to have side effects
may be less likely to be given ECT in the first place. The
fact that this study replicates and extends similar ECT
side-effect advantages in elderly patients5 suggests that
at least local biases in terms of selection, recognition,
or documentation are less likely to have accounted for
these findings. A final methodological consideration involves the possibility that multiple comparisons may
have taken advantage of chance differences between

groups. The low level of various types of side effects
was a relatively consistent pattern for ECT and makes it
highly unlikely that the findings reflect a consistent multiple-comparison effect.
Our study supports a dramatic positive effect of
ECT in the treatment of affective disorders, even in the
old-old age-group. Our results are in agreement with the
study by Rubin and colleagues11 that found a 98% response rate, with 45 of 46 patients showing a moderateto-good response to ECT, as measured by the Beck Depression Inventory, Global Depression Scale, and
nonblinded global assessment. Mulsant et al.’s study13
using the Ham-D scale in elderly patients demonstrated
that 100% had at least a slight degree of improvement,
with two-thirds of patients showing at least a 50% reduction in Ham-D scores.
Our study confirms that ECT is both safe and extremely effective in geriatric patients with severe depression and extends earlier results to the old-old population over 75 years of age. If carefully monitored, even
“old-old” patients can be safely treated with ECT. Further prospective studies using objective outcome measures such as the Ham-D18 and documentation of complications are indicated to elaborate upon these
findings.

References
1. Thompson JW, Weiner RD, Meyers CP: Use of ECT in the United
States in 1975, 1980, and 1986. Am J Psychiatry 1994; 151:1657–
1661
2. Rice DP: Demographic realities and projections of an aging population, in Health Care for Aging Society. Edited by Andreopoulos
S, Jogness J. New York, Church Livingstone, 1989, pp 15–45
3. Finkel SI: Efficacy and tolerability of antidepressant therapy in the
old-old. J Clin Psychiatry 1996; 57(suppl5):23–28
4. Salzman C, Schneider L, Lebowitz B: Antidepressant treatment of
very old patients. Am J Geriatr Psychiatry 1993; 1:21–29
5. Cattan RA, Barry PP, Mead G, et al: Electroconvulsive therapy in
octogenarians. J Am Geriatr Soc 1990; 38:753–758
6. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC,
American Psychiatric Association, 1987
7. American Society of Anesthesiologists: New classification of physical status. Anesthesiology 1963; 24:111
8. Burke WJ, Rutheford JL, Zorumski CF, et al: Electroconvulsive therapy and the elderly. Compr Psychiatry 1985; 26:480–486
9. Burke WJ, Rubin EJ, Zorumski CF, et al: The Safety of ECT in geriatric psychiatry. J Am Geriatr Soc 1987; 35:516–521

236

10. StatMost 32 for Windows. DataMost Corporation, Salt Lake City,
UT, 1996
11. Rubin EH, Kinsherf DA, Wehrman SA: Response to treatment of
depression in the old and very old. J Geriatr Psychiatry Neurol
1991; 4:65–70
12. Philibert RA, Richards L, Lynch CF, et al: Effect of ECT on mortality
and clinical outcome in geriatric unipolar depression. J Clin Psychiatry 1995; 56:390–394
13. Mulsant BH, Rosen J, Thornton JE, et al: A prospective naturalistic
study of electroconvulsive therapy in late-life depression. J Geriatr
Psychiatry Neurol 1991; 4:3–13
14. Koessler D, Fogel BS: Electroconvulsive therapy for major depression in the oldest old. Am J Geriatr Psychiatry 1993; 1:30–37
15. Abrams R: Electroconvulsive Therapy, 2nd Edition. New York, Oxford University Press, 1992
16. Sackeim HA: Electroconvulsive therapy in late-life depression, in
Clinical Geriatric Psychopharmacology, 3rd Edition. Edited by
Salzman C. Baltimore, MD, Williams & Wilkins, 1998, pp 262–309
17. 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
18. Hamilton M: Development of a rating scale for primary depressive
illness. Br J Soc Clin Psychol 1967; 6:278–296

Am J Geriatr Psychiatry 8:3, Summer 2000