Who is at risk of post-MI depressive symptoms?

Abstract

The aim of this study was to identify cardiologic, psychologic, and demographic risk factors in two groups of patients with post-myocardial infarction (MI) depressive symptoms (in-hospital and during the postdischarge year).
Patients admitted for MI were assessed for depressive symptoms with the Beck Depression Inventory (BDI) during hospitalization and 3, 6, and 12 months post-MI. We contrasted both groups with nondepressed patients.
Pre-MI vital exhaustion, living alone, history of depressive disorder, history of MI, poor performance on exercise tolerance testing, and female gender were significantly and independently associated with in-hospital depressive symptoms. Pre-MI vital exhaustion, history of depressive disorder, female gender, poor ejection fraction, and longer hospital stay were independent predictors of the development of postdischarge depressive symptoms.
Post-MI depressive symptoms seem largely driven by the psychological and social consequences of the MI in patients vulnerable to depression, as indexed by a history of depression and vital exhaustion.

Full text

Who is at risk of post-MI depressive symptoms?
Titia A. Spijkerman
a,b,
T, Rob H.S. van den Brink
a,b
, Jaap H.C. Jansen
a,b
,
Harry J.G.M. Crijns
c
, Johan Ormel
a,b
a
Department of Psychiatry, University Hospital Groningen, Graduate School of Behavioral and Cognitive Neurosciences,
University of Groningen, The Netherlands
b
Graduate School for Experimental Psychopathology, University of Groningen, The Netherlands
c
Department of Cardiology, University Hospital Groningen, University Hospital Maastricht, The Netherlands
Abstract
Objective: The aim of this study was to identify cardiologic,
psychologic, and demographic risk factors in two groups of
patients with post-myocardial infarction (MI) depressive symp-
toms (in-hospital and during the postdischarge year). Methods:
Patients admitted for MI were assessed for depressive symptoms
with the Beck Depression Inventory (BDI) during hospital-
ization and 3, 6, and 12 months post-MI. We contrasted both
groups with nondepressed patients. Results: Pre-MI vital
exhaustion, living alone, history of depressive disorder, history
of MI, poor performance on exercise tolerance testing, and
female gender were significantly and independently associated
with in-hospital depressive symptoms. Pre-MI vital exhaustion,
history of depressive disorder, female gender, poor ejection fraction,
and longer hospital stay were independent predictors of the
development of postdischarge depressive symptoms. Conclusions:
Post-MI depressive symptoms seem largely driven by the psy-
chological and social consequences of the MI in patients vul-
nerable to depression, as indexed by a history of depression and
vital exhaustion.
D2005 Published by Elsevier Inc.
Keywords: Myocardial infarction; Depressive symptoms; Risk factors
Introduction
Depressive symptoms and depressive disorder are com-
mon among myocardial infarction (MI) patients. While in
hospital, 15– 45% of MI patients report depressive symptoms
[1–4] and 10 –27% a depressive disorder [3,5,6]. In addition,
approximately 10 –20% develop a depressive disorder after
discharge [1,6–8]. Depressive symptoms following MI are
consequential. Post-MI depressive symptoms have been
shown to be an independent risk factor for mortality,
comparable in strength to LV function [2,3,5,9–12] and
poor health status [13]. The negative impact of depressive
symptoms on cardiac prognosis and health status in general
is potentially preventable. Depressive disorder is often
treated with reasonable success in noncardiac patients
[14,15]. Whether treatment in depressed cardiac patients
has the same effectiveness and whether a successful treat-
ment of depressive disorder has a positive effect on the
cardiac prognosis are not clear yet [16,17]. In previous study
populations, less than 20% of the patients with post-MI
depressive disorder received treatment for their depressive
symptoms or disorder [1,5,6,18].
The present study aims to identify the cardiologic,
psychologic, and demographic risk factors, in as far these
are routinely available or can be easily obtained during MI
hospitalization, in two groups of patients with post-MI
depressive symptoms: (1) those who experience depressive
symptoms during hospitalization; and (2) those who develop
depressive symptoms in the year after discharge from the
hospital. The identification of MI patients with depressive
symptoms was chosen, instead of those who fulfil the more
stringent psychiatric criteria of depressive disorder, because
depressive symptoms have been shown to affect the cardiac
prognosis of MI patients in their own right [3,19]. The
identification of risk factors of post-MI depressive symp-
toms, which are routinely assessed or can easily be obtained,
0022-3999/05/$ – see front matter D2005 Published by Elsevier Inc.
doi:10.1016/j.jpsychores.2005.02.005
Abbreviations: DepreMI, depression after myocardial infarction; MI,
myocardial infarction; VE, vital exhaustion; LVEF, left ventricular ejection
fraction; BDI, Beck Depression Inventory.
TCorresponding author. Tel.: +31 503612079; fax: +31 503619722.
E-mail address: t.a.spijkerman@acggn.umcg.nl (T.A. Spijkerman).
Journal of Psychosomatic Research 58 (2005) 425 – 432

is important for at least two reasons. First, it will help to
contribute to understanding the etiology of post-MI
depressive symptoms. Important questions in this are the
independent contribution of MI-related characteristics, pre-
MI depressive history and vital exhaustion, and demo-
graphic characteristics. Second, knowledge of the risk
factors may improve the recognition of depressive symp-
toms and disorder in medical patients. Depressive disorder
is underdiagnosed in medical patients, in the hospital, and in
general practice as well [20 –23]. If strong predictors of
post-MI depressive symptoms can be identified, this knowl-
edge may alert treatment providers of the possibility of
depressive symptoms or disorder.
Methods
Design
The present study was part of the depression after
myocardial infarction (DepreMI) study, which is a natural-
istic follow-up study of MI patients for the occurrence of
depressive symptoms in the post-MI year and its impact on
cardiac prognosis. Depressive symptoms were assessed at 0,
3, 6, and 12 months post-MI. Patients received usual
medical after-care for their MI. No information was
provided to the treating physicians with respect to patient’s
mental status. The study protocol was approved by the
Institutional Review Board of the four participating hospi-
tals. All participating patients signed an informed consent.
Patient sample
Patients were eligible if they met at least two of three
criteria: (1) chest pain for at least 20 min, (2) creatinine
phosphokinase (CPK) value twice or more than the normal
or creatinine phosphokinase MB (CPK-MB) value greater
than 10% of the CPK value, and (3) the presence of new
pathological Q wave on the electrocardiogram in at least
two leads. Patients were excluded if they had a life
expectancy less than 1 year because of comorbid noncardiac
disease (e.g., malignancies), had poor cognitive functions,
were unable to speak or read Dutch, had visual or auditory
problems that precluded participation, had an MI during
hospital admission for other reasons (except angina pecto-
ris), were not scheduled for follow-up visits in a participat-
ing hospital, or died before they could be approached or
decide about participation.
Assessment of depressive symptoms
Patients were screened for depressive symptoms with the
Beck Depression Inventory (BDI; [24]) four times post-MI:
in the hospital (time frame: since MI) and at 3, 6, and
12 months post-MI (time frame: the previous week). The
BDI is a widely used 21-item, self-report measure of the
presence and severity of symptoms of depression. Respond-
ents are instructed to rate each symptom on a 0 to 3 scale,
with 0 representing babsentQand 1–3 representing increasing
levels of severity. A total score of 10 or higher was consid-
ered indicative of the presence of depressive symptoms.
Demographic, psychiatric, and cardiologic data
We examined three types of potential predictors: demo-
graphic, psychiatric, and cardiologic. In addition, we made a
distinction between information that is routinely available in
cardiologic treatment for MI patients and information that
requires some additional inquiry by the physician.
Demographic characteristics included age at the time of
MI, gender, living alone, and achieved level of education.
Education was categorized in three levels: low, representing
primary school; medium, representing lower vocational
training or secondary school; high, representing higher
vocational training or university degree.
Psychiatric characteristics included the presence of vital
exhaustion in the month before the MI and history of
depressive disorder. All patients were interviewed with the
Composite International Diagnostic Interview (CIDI) at
approximately 3 and 12 months post-MI. The CIDI provides
information on lifetime depressive disorder, i.e., whether a
depressive episode meeting ICD-10 criteria had occurred in
the past, i.e., in the years before the index MI. Episodes with
an onset of more than 1 year before the MI (n=68, 85% of all
episodes with onset before the MI) were classified as having
a history of depressive disorder. Vital exhaustion was
assessed with the Maastricht Questionnaire (MQ; [25]).
For this study, the MQ was modified by changing the time
frame from the past 2 weeks into the months before the MI.
We defined high vital exhaustion as a score of 16 or more.
Cardiologic data
Cardiologic data were obtained from medical records and
encompassed characteristics of the index MI, complications
during hospitalization, treatment of index MI, and history of
MI. Characteristics of the index MI included the following:
the site of the MI (anterior or otherwise); size of MI, as
assessed by the maximum creatinine kinase and creatinine
kinase MB scores; degree of heart failure, as assessed by
Killip class (a standardized four-point clinical assessment of
the degree of heart failure) and left ventricular ejection
fraction [LVEF (assessed either by nuclear (n=285), wall
motion score (n=72), angiography (n=41), or clinical
judgement (n=104); treating cardiologist indicated on the
basis of all available data the LVEF) and dichotomized as
z40% vs. b40%]; and maximum workload. The presence or
absence of any of the following complications during
hospitalization were recorded: atrial fibrillation (AF),
ventricular fibrillation (VF), ventricular tachycardia (VT),
and cardioversion for VF or VT. Aspects of MI treatment
recorded included the administration of thrombolysis at
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432426

admission, invasive procedures during admission (either
CAG, PTCA, CABG, or other cardiac operation), prescrip-
tion of h-blockers before discharge, and duration of hospital
stay. Finally, we recorded the presence of a previous MI.
Analysis
We first examined the course of BDI symptoms during the
post-MI year by presenting the mean BDI scores, proportion
of patients with BDI-10+, and across-time correlations. This
provides information on the stability of mean scores and
individual differences. The relationship between post-MI
depressive symptoms, the outcome variable, and demo-
graphic, psychiatric, and cardiologic characteristics, denoted
as putative predictors, was studied with logistic regression
analysis. Note that the term predictor does not imply causal
direction because it used here in a statistical sense. First, we
calculated the association between each predictor and the
presence of depressive symptoms. Second, we estimated the
combined predictive power of the predictors by hierarchical
multivariate analysis. The hierarchy of inclusion was based
on the availability of information on the predictor in routine
care for MI and the possibility that the predictor could be
influenced by a post-MI depressive symptom. Routinely
obtained data were entered first, then predictors potentially
influenced by post-MI depressive symptoms, and finally,
additionally obtainable data. These criteria lead to the
following stepwise inclusion using forward inclusion based
on the likelihood ratio (LR) test in the multivariate analysis.
Step 1: gender, age, living alone, education, maximum CK,
maximum CK MB, LVEF, anterior site, Killip
class, and history of MI
Step 2: cardiologic treatment post-MI, complications dur-
ing hospitalization, exercise tolerance, and dura-
tion of hospital stay in days
Step 3: history of depressive disorder and vital exhaustion
pre-MI
Separate analyses were performed for the prediction of
(1) in-hospital depressive symptoms and (2) postdischarge
symptoms in patients without depressive symptoms during
hospitalization.
Results
Sample characteristics
There were 1166 patients with a confirmed MI, of which
284 were excluded because of previously mentioned
reasons. We asked 882 patients to participate in the study.
From 528 patients (59.9%), informed consent was obtained
(81% men; mean age=60.7 years, S.D. age=11.7). Women
were significantly older than men (63.7 vs. 60.0 years;
t=2.71, Pb.01). Nonparticipants were significantly older
than the participants (66.7 vs. 60.7 years; t=7.01, Pb.01),
and more often female (40% vs. 19%; v
2
=43.7, Pb.01).
Course of depressive symptoms
In-hospital BDI data were available for 502 patients.
Twenty-one patients did not fill in the in-hospital BDI, and
for five patients, the first BDIs were received after more
than 75 days and were therefore disregarded. Depressive
symptoms during hospitalization were reported by 118
(23.5%) of the patients. Table 1 shows the mean BDI score
and the proportion patients with BDI scores of 10 or more
for each of the four BDI assessments during the post-MI
year. No significant differences in mean score and propor-
tion BDI-10+ between the four BDI assessments were
found. Hence, the mean BDI scores and proportion BDI-
10+ did neither increase nor decrease during the post-MI
year in this cohort. The across-time BDI correlations in
Table 1 show that also the differential stability is substantial,
suggesting that individual differences in BDI symptoms
during hospitalization tend to persist during the post-MI
year. This is also documented by the proportion of patients
with a positive in-hospital BDI score (10+) who still had a
positive BDI score at one (17.6%), two (31.4%), or all three
(40.2%) BDI follow-ups.
In-hospital depressive symptoms
Table 2 presents the associations between depressive
symptoms during hospitalization and the selected putative
predictors. The statistically significant associations suggest
that there is greater chance for in-hospital depressive
symptoms if MI patients are female, live alone, have a
relatively low level of education, a history of depressive
disorder, vital exhaustion in the month before MI, and/or a
maximum workload on exercise tolerance testing of less
than 110 W. A trend towards significance (.05NPb.10)
was found for age, history of MI, Killip class zII, and log
max CK. No association was found with site of MI, log
max CK (MB), arrhythmic events during admission,
LVEF, and treatment.
Tab le 3 shows that, from the routinely available
predictors (Block 1), gender had the strongest relationship
with the presence of in-hospital depressive symptoms, and
that living alone and a history of MI improved the
Table 1
Mean score, proportion with BDI-10+, and across-time correlations of four
BDI assessments
BDI
assessment n
Mean
BDI
% BDI
10 z10
Pearson correlations
In-
hospital
3-
month
6-
month
12-
month
In-hospital 502 6.75 23.5%
3-month 489 6.90 23.9% .68
6-month 469 6.72 25.2% .58 .76
12-month 468 6.81 24.1% .59 .73 .79
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432 427

prediction of post-MI depressive symptoms. After the
inclusion (Block 2) of the predictor that might be
influenced by post-MI depressive symptoms, maximal
workload on exercise tolerance testing, gender was no
longer significantly associated (OR drops from 2.06 to
1.61). This means that maximum workload mediated the
association between gender and in-hospital depressive
symptoms. When also readily available additional pre-
dictors are considered (Block 3), vital exhaustion and
history of depressive disorder independently predicted in-
hospital depressive symptoms. The contribution of gender
further dropped from 1.61 to 1.35 after the addition of
history of depressive disorder, meaning that the influence
of gender is also mediated by the history of depressive
disorder. The model correctly classified 80.3% of the
patients. The prevalence of BDI-10+ in the hospital
increased from 23.5% in the total group to 64.3% in
patients with a history of depressive disorder and pre-MI
vital exhaustion.
Predictors of postdischarge onset of depressive symptoms
This analysis was limited to patients without in-hospital
depressive symptoms. In addition, 30 patients were
excluded because of missing data on one or more BDIs at
3, 6, or 12 months post-MI, leaving 354 patients for the
Table 2
Univariate association between putative predictors and in-hospital depressive symptoms
BDI b10 BDI z10
OR (95% CI) PIn-hospital (n=384) In-hospital (n=118)
Routinely available data not influenced by
the presence of depressive symptoms
Mean age in years 60.1 62.3 1.02 (0.99 –1.04) .08
Female gender 15.4% 29.7% 2.32 (1.43–3.76) b.01
Level of education
I. Low 17.0% 26.3% 1.00
II. Medium 65.4% 64.5% 0.64 (0.38 –1.07)
a
.09
III. High 17.6% 9.1% 0.33 (0.15 – 0.74)
a
b.01
Living alone 12.3% 27.8% 2.75 (1.65 –4.60) b.01
History of MI 12.2% 18.6% 1.64 (0.94 –2.86) .08
Anterior site of MI 29.9% 35.6% 1.29 (0.84 –2.00) .25
Log max CK 2.94 2.86 0.63 (0.39 –1.02) .06
Log max CK MB 1.89 1.83 0.67 (0.40 –1.12) .12
Killip class z2 12.8% 19.5% 1.65 (0.96 –2.86) .07
Arrhythmic events during admission 8.6% 11.0% 1.31 (0.67–2.59) .43
LVEF V40% score 22.1% 25.4% 1.20 (0.74 –1.94) .46
Routinely available data possibly influenced
by depressive symptoms
Mean max workload
(e.t.t. predischarge z110)
76.8% 55.4% 0.38 (0.21 – 0.66) b.01
Thrombolysis 45.1% 38.1% 0.75 (0.49 –1.15) .17
CAG or
PTCA b24 h
11.9% 9.3% 0.76 (0.38 –1.52) 44
CABG 3.9% 1.7% 0.43 (0.10 –1.88) .26
h-Blockers at discharge 77.9% 71.2% 0.70 (0.44 –1.12) .14
Duration of admission (in days) 10.8 11.7 1.01 (0.99 –1.04) .25
Obtainable data
History of depressive disorder 9.6% 27.3% 3.52 (2.03 – 6.09) b0.01
Pre-MI vital exhaustion (16+) 20.9% 60.9% 5.88 (3.75 –9.23) b0.01
a
Compared with education level Class I.
Table 3
Multivariate predictors of in-hospital depressive symptoms in order of entrance in hierarchical analysis
Predictors subsequently entered Association of Block 1 predictors Association of Blocks 1 and 2 predictors Association of Blocks 1–3 predictors
a
Variable OR (95% CI) POR (95% CI) POR (95% CI) P
Female vs. male 2.06 (1.24 –3.45) b.01 1.61 (0.94 –2.75) .08 1.35 (0.75–2.42) .32
Living alone 2.34 (1.38 –3.98) b.01 2.36 (1.38 – 4.05) b.01 2.75 (1.52– 4.95) b.01
History of MI 1.77 (1.00 –3.14) .05 1.64 (0.92 –2.93) .09 1.80 (0.96 –3.36) .07
Workload z110 vs. b110 W 0.45 (0.25 – 0.82) b.01 0.38 (0.19 – 0.73) b.01
History of depression 2.01 (1.07– 3.77) .03
Pre-MI vital exhaustion (16+) 5.40 (3.30 – 8.86) b.01
a
Proportion correctly classified by model: 80.3%.
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432428

analysis. Of these patients, 67 (18.9%) reported onset of
depressive symptoms after discharge. The univariate asso-
ciations between late-onset depressive symptoms and
putative predictors are reported in Table 4.
The statistically significant associations suggest that
there is a greater chance for late-onset depressive symptoms
if MI-patients are female, have a maximum workload on
exercise testing at discharge of less than 110 W, stay longer
than average in the hospital, have a history of pre-MI
depressive disorder, and/or pre-MI vital exhaustion. A trend
towards significance was found for LVEF (P=.06). No
associations were found for age, living alone, all cardiologic
characteristics other than maximum workload and LVEF,
and treatment.
Tab le 5 shows that, from the routinely available
characteristics, gender and LVEF have an independent
contribution to the prediction of late-onset depressive
symptoms. If predictors that might be influenced by
depressive symptoms (Block 2) were added, the duration
of admission emerged as an additional predictor. The risk of
a late-onset depressive symptoms increased with 4% per
hospitalization day. Maximal workload on exercise toler-
ance testing did not contribute to the prediction of late-onset
depressive symptoms. Finally, when psychiatric background
data are added (Block 3), both the history of depressive
disorder and pre-MI vital exhaustion improved the predic-
tion model. The model correctly classified 82.2% of the
patients. The combination of a positive history of depressive
Table 4
Association between putative predictors and incident cases of depressive symptoms during follow-up
BDI b10 at 3, 6, or
12 months (n=287)
BDI z10 at 3, 6,
12 months (n=67) OR (95% CI) P
Routinely available data not influenced by the
presence of depressive symptoms
Mean average age in years 60.0 62.0 1.02 (1.00–1.04) 0.16
Female gender 12.9% 25.4% 2.23 (1.20 – 4.40) 0.01
Level of education
I. Low 14.8% 21.8% 1.00
II. Medium 66.1% 65.6% 0.67 (0.34–1.34)
a
0.26
III. High 19.2% 12.5% 0.44 (1.68–1.15)
a
0.09
Living alone 11.0% 17.9% 1.76 (0.85–3.64) 0.18
History of MI 12.2% 11.9% 0.98 (0.43–2.21) 0.95
Anterior site of MI 30.0% 29.9% 1.00 (0.56–1.78) 0.99
Log CK max 2.94 2.95 1.07 (0.57–2.00) 0.84
Log CK MB max 1.91 1.86 0.73 (0.37–1.41) 0.34
Killip class z2 11.5% 14.9% 1.35 (0.63–2.90) 0.44
LVEF V40% 20.6% 31.3% 1.76 (0.98–3.18) 0.06
Arrhythmic events during admission 7.7% 13.4% 1.87 (0.81–4.23) 0.14
Routinely available data possibly influenced
by depressive symptoms
Thrombolysis 43.6% 50.7% 1.34 (0.78–2.27) 0.29
CAG or PTCA b24 h 12.8% 10.6% 0.81 (0.34 –1.91) 0.63
CABG 3.8% 1.5% 0.38 (0.05–3.00) 0.36
Max workload e.t.t. at discharge z110 W 62.5% 80.5% 0.40 (0.19 – 0.84) 0.02
h-Blockers 77.7% 76.1% 0.92 (0.50 –1.71) 0.79
Duration of admission (in days) 10.3 12.8 1.03 (1.00 –1.07) 0.03
Obtainable data
History of depressive disorder 7.0% 20.3% 3.38 (1.57–7.28) b0.01
Pre-MI vital exhaustion (16+) 14.7% 43.3% 4.43 (2.47–7.95) b0.01
a
Class II and III compared with Class I.
Table 5
Multivariate predictors of in postdischarge depressive symptoms in order of entrance in hierarchical analysis
Predictors subsequently entered
Strength of association
Block 1 predictors Blocks 1 and 2 predictors Blocks 1–3 predictors
a
Variable OR (95% CI) POR (95% CI) POR (95% CI) P
Female vs. male 2.40 (1.24 – 4.63) b.01 2.47 (1.28 – 4.78) b.01 2.53 (1.25 – 5.12) .01
LVEF b40% 1.85 (1.02 –3.37) .04 1.60 (0.86– 3.00) .14 1.79 (0.92–3.50) .09
Duration of admission 1.03 (1.00 – 1.06) .07 1.04 (1.01–1.08) .02
History of depressive disorder 2.57 (1.10 –5.99) .03
Pre-MI vital exhaustion (16+) 4.76 (2.55–8.90) b.01
a
Proportion correctly classified by model: 82.2%.
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432 429

disorder and vital exhaustion in the pre-MI months
improved the prevalence of BDI-10+ from 19.0% to 57.1%.
Discussion
Our findings should be interpreted in the context of four
limitations. First, because our objective was to evaluate the
influence on post-MI depressive symptoms of cardiologic,
psychologic, and demographic risk factors, which are
routinely available or can easily be obtained during MI
hospitalization, some potential predictors like social adver-
sity and personality were not assessed. It is likely that
detailed measures of social adversity and support would
have added to the explanation of post-MI in-hospital and
postdischarge depressive symptoms. The influence of
personality is probably reflected in the association of a
history of pre-MI depressive disorder with post-MI depres-
sive symptoms. Second, from the eligible patients, 41%
refused to participate. The nonparticipants were older and
more often female and, therefore, more at risk of depressive
symptoms. However, the prevalence of depressive symp-
toms in our sample was within the range reported in
comparable studies [4,5]. A third limitation is the retro-
spective assessment of history of depressive disorder and
pre-MI vital exhaustion, which both might have been biased
by current post-MI depressive symptoms. Finally, for 104
patients (21%), we had to rely on the clinical assessment of
LVEF by the treating cardiologist because nuclear, wall
motion, or angiographic data were lacking. We considered
the clinical assessment of the treating cardiologist superior
to a missing value.
Within these limitations, two findings stand out: (1) the
strength of the association of pre-MI vital exhaustion and,
to a lesser extent, history of depressive disorder in
predicting both in-hospital (OR=5.4) and late-onset depres-
sive symptoms (OR=4.8); and (2) the absent or weak
association between the characteristics of the MI (and its
treatment) and in-hospital and postdischarge depressive
symptoms. Although pre-MI vital exhaustion and history
of depressive disorder independently predict post-MI
depressive symptoms, they are strongly correlated (in
patients with history of depressive disorder, 70.1%
reported VE within the months pre-MI, and in patients
without a history of depressive disorder, 29.9%; v
2
=45.4,
Pb.01). Hence, the associations of pre-MI vital exhaustion
and depressive disorder with post-MI depressive symptoms
could simply reflect the persistence of symptoms of vital
exhaustion and depressive disorder from the pre- to the
post-MI period. However, this is unlikely, because both do
not only predict in-hospital depressive symptoms but
equally strongly late-onset depressive symptoms in those
without in-hospital depressive symptoms. Because patients
with in-hospital depressive symptoms were excluded from
the late-onset analysis, late-onset depressive symptoms
must represent a new episode.
The second finding that stands out is the absent or weak
association between MI severity and post-MI depressive
symptoms. From all MI characteristics, only maximum
workload was independently associated with in-hospital
depressive symptoms, and LVEF with late-onset depressive
symptoms. The exercise test score is not an entirely
objective indicator of cardiac function but may be con-
founded with in-hospital depressive symptoms because
performance on the test is also determined by motivation,
and motivation may be reduced in depressed patients. In line
with this interpretation is the fact that LVEF was not
associated with in-hospital depressive symptoms. The lack
of a robust association between MI severity and post-MI
depressive symptoms is consistent with other reports
[2,3,5,26].
Even when adjusted for the impact of the other, vital
exhaustion is a stronger predictor of post-MI depressive
symptoms than is a history of depressive disorder. We can
only speculate why this might be so. We see two
hypotheses. First, the difference in impact might simply
reflect the difference in time to event (MI) between the
history of depression and vital exhaustion. Albeit both were
assessed at T1, the time frame of the vital exhaustion
inventory refers to the few months preceding the MI,
whereas history of depressive disorder was defined as one or
more episodes with an onset of more than a year before the
MI. Second, vital exhaustion might contain nondepression
components that play a stronger role in the etiology of post-
MI depressive symptoms than the vulnerability reflected in a
history of depression.
Living alone was associated with in-hospital depressive
symptoms and a longer hospital stay with postdischarge
depressive symptoms. The influence of a longer hospital
stay remained present after adjustment for MI severity. This
effect of a longer hospital stay is probably not due to
confounding by MI treatment because MI treatment did not
predict postdischarge depressive symptoms. The effect of a
longer hospital stay might be due to a more complicated
course and social problems preventing discharge. Patients
who underwent CABG stayed, on average, 11.2 days longer
in the hospital than did the patients without CABG (ttest:
Pb.01). The independent influence of living alone on in-
hospital depressive symptoms supports a role for social
problems as a factor that influences both the duration of
hospital stay and postdischarge depressive symptoms.
Because we did not measure social problems, we cannot
test this assumption.
Collectively, our findings suggest that post-MI depres-
sive symptoms largely result from the psychological and
social consequences of the MI for patients who are
vulnerable to depressive symptoms, a vulnerability indexed
by premorbid depression and vital exhaustion, who lack
social support, and whose MI has a more complicated
course. Given the findings on the etiology of depressive
disorder in non-MI populations [27–30], feelings of loss,
humiliation, and defeat may be more pronounced in
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432430

this subgroup of MI patients due to their personality and
social situation.
The clinical significance of our study lies in the
identification of predictors of post-MI depressive symptoms.
We found, in multivariate analysis, that in-hospital depres-
sive symptoms were associated with pre-MI vital exhaus-
tion, a history of depressive disorder, a poor result on the
exercise test, and living alone. Depressive symptoms that
develop after discharge were associated, in multivariate
analysis, with pre-MI vital exhaustion, history of depressive
disorder, female gender, duration of admission, and,
marginally, with LVEF. Information on these predictors is
routinely available or can easily be collected during
admission and on consultation after discharge. This infor-
mation can be used to decide whether to explore the
presence of depression. The prevalence of BDI-10+ during
hospitalization increased from 23.5% in the total group to
64.3% in patients with a history of depressive disorder and
pre-MI vital exhaustion. For postdischarge depressive
symptoms, the prevalence increased from 19.9% to
57.1%. The regression model correctly classified 80.3%
(in-hospital) and 82.2% (postdischarge) of the patients.
Work is in progress to establish the screening properties of
an easy-to-administer brief version of the BDI in post-MI
patients with regard to not only DSM-IV major depression
but also the persistence of depressive symptoms.
Conclusions
Pre-MI vital exhaustion and a history of depressive
disorder are strong predictors of both in-hospital and
postdischarge depressive symptoms during the postdi-
scharge year. In-hospital depressive symptoms are, besides
these indicators, also associated with a poor performance on
exercise testing and living alone. Additional predictors of
late-onset depressive symptoms are female gender, duration
of admission, and, marginally, LVEF. Post-MI depressive
symptoms seem largely driven by the psychological and
social consequences of the MI in patients vulnerable to
depression, as indexed by a history of depression, vital
exhaustion, and living alone.
Acknowledgments
We thank the Dutch Organization for Scientific Research
(Zon MW) for their financial support of this study (grant
904-57-100).
References
[1] Schleifer SJ, Macari-Hinson MM, Coyle DA, Slater WR, Kahn M,
Gorlin R, et al. The nature and course of depression following
myocardial infarction. Arch Intern Med 1989;149(8);1785 – 9.
[2] Ladwig KH, Kieser M, Konig J, Breithardt G, Borggrefe M.
Affective disorders and survival after acute myocardial infarction.
Results from the post-infarction late potential study. Eur Heart J
1991;12(9);959 – 64.
[3] Bush DE, Ziegelstein RC, Tayback M, Richter D, Stevens S, Zahalsky
H, et al. Even minimal symptoms of depression increase mortality risk
after acute myocardial infarction. Am J Cardiol 2001;88(4);337 – 41.
[4] Lane D, Carroll D, Ring C, Beevers DG, Lip GY. Mortality and
quality of life 12 months after myocardial infarction: effects of
depression and anxiety. Psychosom Med 2001;63(2);221 – 30.
[5] Frasure-Smith N, Lesperance F, Talajic M. Depression following
myocardial infarction. Impact on 6-month survival. JAMA 1993;
270(15);1819 – 25.
[6] Travella JI, Forrester AW, Schultz SK, Robinson RG. Depression
following myocardial infarction: a one year longitudinal study. Int J
Psychiatry Med 1994;24(4);357 – 69.
[7] Lesperance F, Frasure-Smith N, Talajic M. Major depression before
and after myocardial infarction: its nature and consequences.
Psychosom Med 1996;58(2);99 – 110.
[8] Honig A, Lousberg R, Wojciechowski FL, Cheriex EC, Wellens HJ,
van Praag HM. Depression following a first heart infarct; similarities
with and differences from bordinaryQdepression. Ned Tijdschr
Geneeskd 1997;141(4);196 – 9.
[9] Ahern DK, Gorkin L, Anderson JL, Tierney C, Hallstrom A, Ewart C, et
al. Biobehavioral variables and mortality or cardiac arrest in the Cardiac
Arrhythmia Pilot Study (CAPS). Am J Cardiol 1990;66(1);59 – 62.
[10] Denollet J, Sys SU, Brutsaert DL. Personality and mortality after
myocardial infarction. Psychosom Med 1995;57(6);582 – 91.
[11] Irvine J, Basinski A, Baker B, Jandciu S, Paquette M, Cairns J, et al.
Depression and risk of sudden cardiac death after acute myocardial
infarction: testing for the confounding effects of fatigue. Psychosom
Med 1999;61(6);729 – 37.
[12] Welin C, Lappas G, Wilhelmsen L. Independent importance of
psychosocial factors for prognosis after myocardial infarction. J Intern
Med 2000;247(6);629 – 39.
[13] Ruo B, Rumsfeld JS, Hlatky MA, Liu H, Browner WS, Whooley MA.
Depressive symptoms and health-related quality of life: The Heart and
Soul Study. JAMA 2003;290(2);215 – 21.
[14] Williams JW, Mulrow CD, Chiquette E, Noel PH, Aguilar C,
Cornell J. A systematic review of newer pharmacotherapies for
depression in adults: evidence report summary. Ann Intern Med
2000;132(9);743 – 56.
[15] Geddes JR, Freemantle N, Mason J, Eccles MP, Boynton J. Selective
serotonin reuptake inhibitors (SSRIs) for depression (Cochrane
Review). Cochane Libr 2002;(1). http://www.mrw.interscience.wiley.
com/cochrane/clsysrev/articles/CD001851/frame.html.
[16] van den Brink RH, van Melle JP, Honig A, Schene AH, Crijns HJ,
Lambert FP, et al. Treatment of depression after myocardial infarction
and the effects on cardiac prognosis and quality of life: rationale and
outline of the Myocardial INfarction and Depression-Intervention
Trial (MIND-IT). Am Heart J 2002;144(2);219 – 25.
[17] Berkman LF, Blumenthal J, Burg M, Carney RM, Catellier D, Cowan
MJ, et al. Effects of treating depression and low perceived social
support on clinical events after myocardial infarction: the Enhancing
Recovery in Coronary Heart Disease Patients (ENRICHD) Random-
ized Trial. JAMA 2003;289(23);3106 – 16.
[18] Frasure-Smith N, Lesperance F, Juneau M, Talajic M, Bourassa MG.
Gender, depression, and one-year prognosis after myocardial infarc-
tion. Psychosom Med 1999;61(1);26 – 37.
[19] Frasure-Smith N, Lesperance F, Talajic M. Depression and 18-month
prognosis after myocardial infarction. Circulation 1995;91(4);
999 – 1005.
[20] Mayou R, Hawton K, Feldman E. What happens to medical patients
with psychiatric disorder? J Psychosom Res 1988;32(4–5);541– 9.
[21] Freedland KE, Lustman PJ, Carney RM, Hong BA. Underdiagnosis of
depression in patients with coronary artery disease: the role of
nonspecific symptoms. Int J Psychiatry Med 1992;22(3);221– 9.
[22] Tiemens BG, Ormel J, van den Brink RH, Jenner JA, van der Meer K,
van Os TW, et al. Recognition of depression and generalized anxiety
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432 431

disorder in primary care. The development of a screening instrument.
Tijdschr Soc Gezondheidsz 1995;73:520 – 7.
[23] Simon GE, Goldberg D, Tiemens BG, Ustun TB. Outcomes of
recognized and unrecognized depression in an international primary
care study. Gen Hosp Psychiatry 1999;21(2);97 – 105.
[24] Beck ATWCH, Mendelson M, Erbaugh J. An inventory for measuring
depression. Arch Gen Psychiatry 1961;4:561 – 71.
[25] Appels A, Hoppener P, Mulder P. A questionnaire to assess
premonitory symptoms of myocardial infarction. Int J Cardiol
1987;17(1);15 – 24.
[26] Lane D, Carroll D, Ring C, Beevers DG, Lip GY. Effects of
depression and anxiety on mortality and quality-of-life 4 months
after myocardial infarction. J Psychosom Res 2000;49(4);229 – 38.
[27] Brown GW, Harris T. Social origins of depression. London: Tavistock
Publications Limited; 1978.
[28] Monroe SM, Rohde P, Seeley JR, Lewinsohn PM. Life events and
depression in adolescence: relationship loss as a prospective risk
factor for first onset of major depressive disorder. J Abnorm Psychol
1999;108(4);606 – 14.
[29] Ormel J, Wohlfarth T. How neuroticism, long-term difficulties, and
life situation change influence psychological distress—a longitudinal
model. J Pers Soc Psychol 1991;60(5);744 – 55.
[30] Ormel J, Oldehinkel AJ, Brilman EI. The interplay and etiological
continuity of neuroticism, difficulties, and life events in the etiology of
major and subsyndromal, first and recurrent depressive episodes in
later life. Am J Psychiatry 2001;158(6);885 – 91.
T.A. Spijkerman et al. / Journal of Psychosomatic Research 58 (2005) 425–432432