Eur J Epidemiol (2013) 28:249–256
DOI 10.1007/s10654-013-9765-3
CARDIOVASCULAR DISEASE
Depressive symptoms, antidepressants and disability and future
coronary heart disease and stroke events in older adults:
the Three City Study
Renaud Péquignot • Christophe Tzourio • Karine Péres •
Marie-Laure Ancellin • Marie-Cécile Perier • Pierre Ducimetière
Jean-Philippe Empana
Received: 25 August 2012 / Accepted: 7 January 2013 / Published online: 22 January 2013
Ó Springer Science+Business Media Dordrecht 2013
Abstract To investigate the association between baseline
depressive symptoms and first fatal and non fatal coronary
heart disease (CHD) and stroke in older adults, taking anti-
depressants and disability into account. In the Three City
Study, a community-based prospective multicentric obser-
vational study cohort, 7,308 non-institutionalized men and
women aged C65 years with no reported history of CHD,
stroke or dementia, completed the 20-item Center for Epi-
demiologic Studies Depression Scale (CESD) question-
naire. First CHD and stroke events during follow-up were
adjudicated by an independent expert committee. Hazard
ratios (HRs) were estimated by Cox proportional hazard
model. After a median follow-up of 5.3 years, 338 subjects
had suffered a first non-fatal CHD or stroke event, and 82
R. Péquignot
M.-C. Perier
J.-P. Empana (&)
INSERM U970, Paris Cardiovascular Research Centre, Paris
Descartes University, Sorbonne Paris Cité, UMR-S970, 75015
Paris, France
e-mail: jean-philippe.empana@inserm.fr
R. Péquignot (&)
Service de Médecine et Réadaptation, Hôpitaux de Saint-
Maurice, 14 rue du Val d’Osne, 94415 Saint-Maurice CEDEX,
France
e-mail: r.pequignot@hopitaux-st-maurice.fr
C. Tzourio
INSERM U708, 75651 Bordeaux, France
K. Péres
INSERM U897, 33076 Bordeaux, France
M.-L. Ancellin
INSERM U888, Montpellier 1 University, 34093 Montpellier,
France
P. Ducimetière
Paris-Sud University, 94807 Villejuif, France
•
had died from a CHD or stroke. After adjustment for study
center, baseline socio-demographic characteristics, and
conventional risk factors, depressive symptoms (CESD C
16) were associated with fatal events only: fatal CHD plus
stroke (HR = 2.50; 95 % CI 1.57–3.97), fatal CHD alone
(n = 57; HR = 2.21 ; 95 %CI 1.27–3.87), and fatal stroke
alone (n = 25; HR = 3.27; 95 % CI 1.42–7.52). These
associations were even stronger in depressed subjects
receiving antidepressants (HR = 4.17; 95 % CI 1.84–9.46)
and in depressed subjects with impaired Instrumental
Activities of Daily Living (HR = 8.93; 95 % CI 4.60–
17.34). By contrast, there was no significant association with
non fatal events (HR for non-fatal CHD or stroke = 0.94;
95 % CI 0.66–1.33). In non-institutionalized elderly sub-
jects without overt CHD, stroke or dementia, depressive
symptoms were selectively and robustly associated with first
fatal CHD or stroke events.
Keywords Aging
Depression
Cardiovascular disease

Epidemiology
Risk factors
Introduction
In the middle-aged, depressive symptoms and major
depression have been associated with the onset of coronary
heart disease (CHD) and, particularly, sudden cardiac
arrest [1–3], while evidence for an association with risk of
stroke is emerging [4, 5]. As the prevalence of both
depression and cardiovascular disease increases with age
[6, 7], and given the world’s aging population, assessment
of the relationship between depression and risk of cardio-
vascular disease in the elderly could have major public-
health implications. Depressive symptoms have been
associated with all-cause mortality in the elderly [8, 9].
123
250
With respect to morbidity from cardiovascular disease, the
evidence is less clear, as associations between depression
and the incidence of CHD or stroke have been described in
some but not all studies [9–14]. In these previous studies,
the respective roles of antidepressants use and disability,
two frequent conditions in the elderly, have not been
carefully considered. Their contribution is important to be
evaluated as antidepressants use may represent a marker of
the severity of depressive symptoms, while depressive
symptoms may be a consequence of disability. Moreover,
while depression has been associated both with fatal and
non-fatal CHD events in the middle-aged [1], this question
has not been addressed in the elderly.
Therefore, using data from the Three City Study [15],
we aimed to clarify the association between depressive
symptoms and risk of fatal and non-fatal CHD and stroke
events in older adults, taking into account the respective
roles of antidepressant use and disability.
Methods
Population
The protocol of the Three City study has been previously
detailed [15]. Briefly, 9,294 subjects, aged C65 years and
non-institutionalized, were randomly selected from the
electoral rolls of three large cities in France between March
1999 and March 2001. The study’s protocol was approved
by the Ethics Committee of the University Hospital of
Kremlin-Bicêtre. All participants gave their informed
consent.
Baseline data collection
Trained interviewers conducted face-to-face interviews
using a standardized questionnaire, including demographic
characteristics, occupation, daily-life habits, functional
status, cognitive tests and past history of cardiovascular
disease. Brachial blood pressure was measured twice, after
at least 5 min of rest in a seated position using a validated
digital electronic tensiometer (OMRON M4, OMRON
Corp., Kyoto, Japan). Hypertension was defined as a blood
pressure [140/90 mmHg or the use of an antihypertensive
treatment. Global cognitive function was assessed with the
Mini Mental State Examination (MMSE).
Blood was collected following overnight fasting, and
centralized standard measurements were assessed for lip-
ids, glucose, and creatinine levels. Hypercholesterol-
emia was defined as a total cholesterol concentration of
[2.40 g/L or the use of lipid-lowering treatment. Diabetes
was defined as a fasting blood-glucose level of C7 mmol/L
and/or treatment for diabetes. Glomerular-filtration rate
123
R. Péquignot et al.
was estimated by the Modification of Diet in Renal Disease
(MDRD) equation. A bilateral B-mode ultrasound exami-
nation of the carotid arteries were offered to 6,631 subjects,
aged \86 years, to assess the presence of carotid plaques
and to quantify intima-media thickness of the common
carotid arteries 2 cm below the bifurcation in an area free
of carotid plaques (CCA-IMT) [16].
Assessment of depressive symptoms and use
of antidepressants
Depressive symptoms at baseline were evaluated using the
Center for Epidemiologic Studies Depression Scale
(CESD) questionnaire, a 20-item self-report rating scale
designed to evaluate the frequency and severity of
depressive symptoms [17]. Subjects with a score C16 out
of a total of 60 were identified as having depressive
symptoms.
The face-to-face questionnaire included an inventory of all
drugs used during the preceding month. To reduce underre-
porting, participants who were interviewed at home were
asked to show medical prescriptions, drug packages, and any
other relevant information; those interviewed at the study
center were asked to come with their prescription forms. Drug
names were coded using the World Health Organization
Anatomic Therapeutic Chemical classification. Prescribed
doses and treatment indications were not available.
Assessment of disability
Disability in Instrumental Activities of Daily Living
(IADL) was evaluated using the Lawton scale [18]. Par-
ticipants were considered disabled for IADL if they needed
help to perform at least one task from the scale.
Ascertainment of vascular events
At each follow-up visit that took place every 2 years,
subjects were asked to report any new severe medical event
or hospitalization since the last contact. All available
clinical information was collected for all subjects who
reported a possible CHD or stroke event, including emer-
gency medical service and hospitalization reports, plus
neuro-imaging reports for stroke, interviews with the
patient’s physician or family and, if there was a fatal event,
death certificates and autopsy reports [19, 20].
Coronary heart disease and stroke events were adjudi-
cated by an independent expert committee. CHD included
a definite hospitalized angina, hospitalized myocardial
infarction, CHD death (I210–I219, I251–259, I461 and
R960 ICD-10 codes), coronary balloon dilatation or arterial
bypass. Stroke was defined according to the criterion of the
World Health Organization, as a new focal neurological
Depression and vascular events in the elderly 251

deficit of sudden or rapid onset, of presumed vascular

origin that lasted 24 h or more, or that lead to death.
Confirmed stroke cases were further classified as ischemic,
hemorrhagic, or unspecified. Fatal CHD and fatal stroke
were defined as death caused by CHD or stroke that
occurred within 28 days after the occurrence of the event.

Statistical analyses

Student and Chi squared tests were used to compare the

baseline characteristics between groups. In longitudinal
analysis, time was censored to the first event or to the last
follow-up visit if no event occurred. However, ten subjects
who suffered from a non-fatal event and then died from
CHD or stroke, at [28 days later, were analyzed for
morbidity and mortality. Univariate Kaplan–Meier curves
of fatal and non-fatal CHD and stroke events by baseline
depressive-symptom status were compared using the log-
rank test. The hazard ratios (HRs) and 95 % confidence
intervals (CI) of depressive symptoms for CHD and stroke
combined, and for each individual endpoint, were esti-
mated in separate Cox’s proportional hazard models.
Model 1 was adjusted for age, gender, and study center;
model 2 was further adjusted for educational level
(C12 years of schooling), living alone, hypertension,
impaired fasting glycemia or diabetes mellitus, alcohol
consumption (C3 glasses of alcohol daily), smoking status
(never or past vs. current), MMSE (continuous), and
hypercholesterolemia. Missing information on fasting
glycaemia (n = 352), hypercholesterolemia (n = 274),
alcohol consumption (n = 108), IADL (n = 49), and other
covariates (n = 2–27) were imputed using multiple impu-
tations by chained equations [21]. To assess the contribu-
tion of antidepressant-treatment use and IADL disability,
we cross-tabulated depressive-symptom status alternatively
with antidepressant-treatment use and IADL-disability
status. Non-depressed participants either free of antide-
pressants or of IADL impairment were considered as the
reference group to estimate hazard ratios. The proportional
hazard assumption of the Cox’s model was verified
graphically for all covariates. All analyses were two-sided
and p values \0.05 were considered statistically signifi-
cant. Statistical analyses were performed using SAS ver-
sion 9.2 (Cary, North Carolina, USA).
Results
The flow chart for the study population is shown in Fig. 1. Of
the 7,538 participants free of CHD, stroke, dementia, and
who had data available for the CESD score at baseline, 230
were lost to follow-up. These latter were older, had more
depressive symptoms, diabetes, and IADL impairment at
Fig. 1 Flow chart of the Three City Study. CHD coronary heart
disease, CESD Center for Epidemiologic Studies Depression Scale
baseline and had a higher death rate over 6 years (58 vs.
7 %) than the remaining 7,308 subjects (all p \ 0.001) who
constituted the final study sample.
Baseline characteristics (Table 1)
Median age was 73 years (interquartile range: 69–77) and
36.5 % were male. There were 1,657 subjects (22.7 %)
with depressive symptoms, 13.2 % in men and 28.1 % in
women (p \ 0.001). Among those with depressive symp-
toms, 16.0 % were receiving antidepressants (12.3 % in
men vs. 17.0 % in women, p = 0.03), of which 53.1 %
were receiving selective serotonin-reuptake inhibitors
(SSRI) and 27.4 % were receiving tricyclics.
Follow up
After a median follow-up of 5.3 years (inter quartile range
4.8–5.6), 338 subjects had suffered from a first non-fatal
CHD or stroke, including 222 non-fatal CHD events and
116 non-fatal strokes (102 ischemic, 10 hemorrhagic, 4 of
unknown cause). Moreover, 82 had died from CHD or
stroke including 57 CHD and 25 stroke events (Fig. 1).
Furthermore, 148 women and 190 men had suffered from a
first non fatal CHD or stroke events, and 40 and 42 had a
fatal CHD or stroke events respectively. Finally, 169 sub-
jects aged less than 75y (n = 4,346) and 169 aged 75 or
more (n = 2,962) experienced a first non fatal CHD or
stroke events, and 30 and 52 a fatal event respectively.

123
252 R. Péquignot et al.

Table 1 Baseline characteristics by depressive-symptom status: the (HR = 2.50; 95 % CI 1.58–3.96), fatal CHD alone (HR =
Three City Study 2.22; 95 %CI 1.28–3.87), and fatal stroke alone (HR =
Variable CESD \ 16 CESD C 16 pa 3.27; 95 % CI 1.43–7.47) after adjustment for age, study
(n = 5,651) (n = 1,657) center and sex (model 1). These associations were unaf-
fected by further adjustment (model 2). The association
Age (years), mean (SD) 73.6 (5.3) 74.3 (5.5) \0.0001
between depressive symptoms and fatal CHD or stroke did
Male gender (%) 41.0 21.2 \0.0001
not differ with age (HR\75 years = 3.47, 95 % CI 1.56–7.71;
C12 years schooling (%) 38.6 32.9 \0.0001
HRC75 years = 2.11; 95 % CI = 1.19–3.76; p for interac-
Living alone (%) 32.1 47.2 \0.0001
tion = 0.68) or by gender (HRwomen = 2.43, 95 % CI
Currently smoking (%) 6.1 5.5 0.38
1.29–4.57; HRmen = 2.68; 95 % CI = 1.35–5.30; p for
[3 glasses of alcohol/day 9.2 5.7 \0.0001
(%)
interaction = 0.93) or by study centers (p for interac-
tion = 0.80). Moreover, depressive symptoms were asso-
BMI (kg/m2), mean (SD) 25.6 (3.9) 25.4 (4.4) 0.11
ciated with total stroke (HR = 1.54; 95 % CI 1.06–2.25)
Impaired fasting glycemia 12.6 12.7 0.86
or diabetes mellitus (%) but not with total CHD events (HR = 1.11; 95 % CI
Hypertensionb (%) 76.2 74.4 0.13 0.82–1.50) after multivariate adjustment (model 2).
c
Hypercholesterolemia (%) 56.7 58.4 0.25 As shown in Table 3, the risk of fatal CHD and stroke
Glomerular-filtration rate 21.2 30.0 0.11 increased in a graded manner across depressive symptoms
\60 mL/mind (%) and antidepressant treatments statuses, with a 4.17-fold
MMSE, mean (SD) 27.5 (1.9) 27.0 (2.2) \0.0001 increased risk in those with depressive symptoms receiving
Antidepressants (%) 3.9 16.0 \0.0001 antidepressants (p for trend \0.001). Of note, subjects
Tricyclic 1.1 4.3 receiving tricyclics at baseline had a 4.89-fold increased
SSRI 2.2 8.2 risk (95 % CI 2.20–10.89) of fatal CHD or stroke, and
Other 0.7 4.2 those receiving SSRIs had a 2.36-fold increased risk
Disabled for IADLe (%) 5.5 12.7 \0.0001 (95 %CI 1.06–5.23), compared to those not taking antide-
Intima-media thickness f
0.71 (0.12) 0.70 (0.11) 0.04 pressants, independently from baseline depressive symp-
(mm), mean (SD) toms and conventional risk factors. Similarly, the risk of
Carotid plaquesf (%) 44.3 40.7 0.03 fatal CHD or stroke gradually increased across the com-
bination of depression and incapacity groups; in particular,
BMI body-mass index, CESD Center for Epidemiologic Studies
Depression Scale, IADL Disability in Instrumental Activities of Daily there was a nine-fold increased risk in subjects who had
Living, MMSE Mini Mental State Examination both depressive symptoms and IADL disability (p for trend
a
Chi squared or student’s t test as appropriate \0.001). However, depressive symptoms remained unas-
b
Blood pressure [140/90 mmHg or antihypertensive treatment sociated with non-fatal CHD or stroke when combined with
c
Fasting total cholesterol [2.40 g/L or receiving lipid-lowering antidepressant treatment or IADL (not shown).
treatment
d
Estimated by the MDRD equation Sensitivity analysis
e
Need help to perform at least one task from the Instrumental
Activities of Daily Living The lack of association between depressive symptoms and
f
B-mode ultrasound examination of the carotid arteries performed in non-fatal events persisted after restricting CHD events to
5,487 participants aged \86 years
hard end points (myocardial infarction and CHD death):
HRs for non fatal myocardial infarction and for non-fatal
myocardial infarction plus stroke were respectively of 1.03
Univariate Kaplan–Meir analysis (95 % CI 0.63–1.70) and 1.18 (95 % CI 0.85–1.63).
Moreover, the significant association between depressive
Difference in the cumulative incidence of events between symptoms and fatal events remained significant after a
depressed and non-depressed subjects was statistically series of exclusions that were performed to address the
significant for fatal CHD and stroke events only, either issue of residual confounding : the HRs for fatal CHD or
combined or considered individually (Fig. 2a, b). stroke were 2.74 (95 % CI 1.58–4.74) and 2.60 (95 % CI
1.58–4.26) after exclusion of the 38 subjects who died
Hazard-ratio estimates within the first 2 years of follow-up and of the 373 subjects
with a self-reported heart failure at baseline respectively.
As shown in Table 2, depressive symptoms were not Likewise, the HR for fatal stroke was 2.74 (95 %CI
related to non-fatal CHD or stroke. Instead, they were 1.12–6.71) when the 117 subjects with self-reported atrial
associated with fatal CHD and stroke combined fibrillation at baseline were excluded; finally, exclusion of

123
Depression and vascular events in the elderly 253

Table 2 Hazard ratio (HR) for baseline depressive symptoms

(CESD C 16) for coronary heart disease and stroke events over
6 years
n Model 1 Model 2
HR 95 % CI HR 95 % CI

CHD or stroke
Non-fatal events 338 1.03 0.78–1.36 1.08 0.83–1.42
Fatal events 82 2.50 1.58–3.96 2.50 1.57–3.97
CHD
Non-fatal CHD 222 0.92 0.64–1.31 0.94 0.66–1.33
Fatal CHD 57 2.22 1.28–3.87 2.21 1.27–3.87
Stroke
Non-fatal stroke 116 1.34 0.88–2.05 1.36 0.89–2.09
Fatal stroke 25 3.27 1.43–7.47 3.27 1.42–7.52
The Three-City Study
Model 1 was adjusted for age, study center, and gender; model
2 = model 1 ? smoking status, alcohol consumption, high blood
pressure, impaired fasting glycemia or diabetes, hypercholesterol-
emia, living alone, education level, Mini Mental State Examination
(MMSE) score
CESD Center for Epidemiologic Studies Depression Scale, CHD
coronary heart disease

was categorized by 5 points increase, the HRs for fatal

events were 0.70 (95 % CI 0.32–1.51), 2.08 (95 % CI
1.06–4.07), 3.31 (95 % CI 1.63–6.72) and 2.85 (95 % CI
1.38–5.90) respectively for a score between 6–10, 11–15,
16–20, and 21 or more, as compared to a score between 0
and 5 (p for trend \0.001).

Discussion

In this large prospective community-based cohort of older

Fig. 2 a Non-adjusted cumulative incidence rates of fatal CHD and
stroke events by baseline depressive-symptom status. The Three City
Study. b Non-adjusted cumulative incidence rates of non-fatal CHD
and stroke events by baseline depressive-symptom status. The Three
City Study
the 1,612 subjects who self-reported a previous episode of
depression at baseline yielded a HR of 2.27 (95 % CI
1.41–3.65) for fatal CHD or stroke. Moreover, among the
subsample of subjects who had a B-mode ultrasound
examination of their carotid arteries at baseline, the HRs of
depressive symptoms for fatal CHD and stroke combined
were 3.48 (95 % CI 1.89–6.41) after adjustment for CCA-
IMT, and 3.11 (95 % CI 1.73–5.59) after adjustment for
carotid plaques, respectively. Lastly, when the CESD score
adults, depressive symptoms at baseline were consistently
associated with fatal CHD and stroke events over 6 years
of follow-up. Moreover, subjects who had depressive
symptoms and were receiving antidepressants, or those
who had depressive symptoms and were impaired for
IADL, had the strongest association with fatal vascular
events. In contrast, depressive symptoms were not related
to the onset of non-fatal CHD or stroke events, whether
combined or not with antidepressants or incapacity for
IADL.
Previous studies
A number of previous studies on the elderly have examined
the relationship between depressive symptoms and car-
diovascular disease, but have yielded mixed results [9–14].
They have focused either on total mortality, CHD, or
stroke, and only a few have assessed the association with

123
254 R. Péquignot et al.

Table 3 Hazard ratio (HR) for fatal coronary heart disease and stroke event over 6 years by combinations of depressive-symptom status and
antidepressant use or impaired IADL status at baseline
N n HR 95 % CI

No antidepressants and no depressive symptoms 5,430 43 1.00

Depressive symptoms only 1,392 26 2.54 1.53–4.21
Antidepressants only 221 6 4.03 1.69–9.61
Depressive symptoms and antidepressants 265 7 4.17 1.84–9.46
p for trend \0.0001
No impaired IADLa and no depressive symptoms 5,306 39 1.00
Depressive symptoms only 1,433 18 1.84 1.04–3.27
Impaired IADL only 311 8 2.55 1.14–5.70
Depressive symptoms and impaired IADL 209 15 8.93 4.60–17.34
p for trend \0.0001
The Three-City Study
HRs were adjusted for age, study center, gender, smoking status, alcohol consumption, high blood pressure, impaired fasting glycemia or
diabetes, hypercholesterolemia, living alone, education level, MMSE score, antidepressant medication, and impaired IADL (when appropriate)
CHD coronary heart disease, IADL Instrumental Activities of Daily Living
a
IADL was missing in 49 subjects, and was imputed. Two of them died of CHD or stroke

CHD or stroke events in the same cohort [10, 11].
Importantly, none have distinguished between fatal and
non-fatal events.
Association with fatal events
The seemingly selective association between depressive
symptoms and fatal events in our study was observed for
CHD and stroke events, for men and women, after adjusting
for important confounding and mediating factors, which
underlines the robustness of our findings. Importantly, these
associations were strong and statistically significant despite
the relatively limited number of fatal events. Interestingly,
our findings on depressive symptoms and fatal CHD events
are consistent and expand those recently reported in two
population-based studies of middle-aged individuals. In the
EPIC-Norfolk study, men and women 40–79 years of age
with major depression had a 2.7-fold increased risk of CHD
mortality over 8 years of follow up [22]. In the Nurses’
Health Study including women 30–55 years of age,
depressive symptoms which were measured by the Mental
Health Index, were associated with fatal CHD and sudden
cardiac death in particular, but not with non-fatal myocar-
dial infarction [3]. In the present study, the significant
association between depressive symptoms and total stroke
was reflected in a specific association with fatal stroke only.
This is consistent with a recent meta analysis on published
data showing association of depression or depressive
symptoms with fatal stroke only [23].
The association between depressive symptoms and fatal
CHD and stroke may be due to several mechanisms. First,
it may be related to the known susceptibility of depressed
subjects to developing ventricular arrhythmias and sudden
cardiac death [2, 3], in particular due to the alteration of
their autonomic imbalance [24]. Second, those with
depressive symptoms may have suffered from more severe
vascular events and, therefore, may have been more likely
to die from their vascular event. Third, depression is
associated with poor health behaviours which may con-
tribute to the increased risk of fatal events [25]; our main
analyses were adjusted for main confounding factors
however. Fourth, residual confounding by chronic disease
(reverse causality phenomenon) and/or by an occult vas-
cular disease (the vascular depression hypothesis) cannot
be excluded [1, 26]. Our sensitivity analyses provided
consistent results however.
On the other hand, the apparent lack of significant
association between depressive symptoms and non-fatal
vascular events in our study might be related to survival
bias, i.e. those depressed subjects who were the most likely
to develop vascular morbidity had already died before
enrollment to the study. Similarly, exclusion of the 230
subjects who were lost to follow up could have also con-
tributed to the apparent lack of association with non fatal
events, as the latter were more depressed and had several
risk factors compared to the remaining subjects.
Depression, antidepressants, and disability
Although depressive symptoms alone were associated with
fatal CHD or stroke events, their combination with either
IADL disability or antidepressant-treatment use yielded the
greatest increased risk. We were not able to address in the
present study whether depressive symptoms were a cause

123
Depression and vascular events in the elderly
or a consequence of disability in IADL. There may be a
complex interplay between these two factors. The current
study suggests that the combined effect of depressive
symptoms and disability for IADL on the risk of fatal
events was greater than the simple addition of their indi-
vidual effect. Moreover, the currently reported increased
risk of fatal events associated with tricyclics is consistent
with their described association with total mortality and
sudden cardiac arrest [27]. More striking was the observed
increased risk of fatal vascular events associated with
SSRIs, although these medications have been recently
associated with sudden cardiac death in middle-aged
women and with stroke risk (especially hemorrhagic) in
post-menopausal women [3, 28]. However the observa-
tional design of the current study precludes from drawing
any causal relationship between antidepressant-treatment
and fatal events; in particular, confusion by indication
cannot be excluded if antidepressant were prescribed in
more severe depression. Moreover, association between
antidepressant treatment and outcome may reflect a lack of
adherence to treatment [25].
Implications
Depressive symptoms were present in 23 % of the study
participants and were associated with a 2.5-fold increased
risk of CHD and stroke mortality, so that the population-
attributable risk due to depressive symptoms yielded 26 %
in the current study. This estimate gives an indication of
the potential benefit that may be expected by effective
interventions on depressive symptoms. But at present, the
potential impact of effective interventions for treating
depressive symptoms on vascular disease remains a chal-
lenging and unresolved issue [29–31]. Given the magnitude
of the association between depressed subjects with dis-
ability and fatal CHD and stroke, the present study also
suggests that this group of individuals may be targeted for
cardiovascular disease prevention in priority.
Limitations
The current study has several strengths including its pro-
spective design, the study sample size, the wide age range
among the elderly population, the participation of both men
and women, the possibility to adjust for several important
confounding factors, a strict definition of vascular events,
and the possibility to compare associations with individual
vascular end point including CHD and stroke. However,
this study is not without limitations. Depressive symptoms
were analyzed as a whole, whereas cognitive and somatic
dimensions may have specific associations with vascular
events [32]. Analysis was based on depressive symptoms
measured at baseline. Information on the duration of
255
depressive symptoms and of antidepressant treatment prior
to baseline were not available. Information on non-phar-
maceutical treatments for depression was not assessed in
the study. Finally, given the median age of 73, the current
participants were healthy survivors and the results may
differ for frailer elderly subjects.
In conclusion, in this large prospective study of non-
institutionalized elderly subjects, baseline depressive
symptoms were associated with CHD and stroke mortality,
but not with non-fatal CHD and stroke events. The current
study also suggests that depressed subjects with disability
are at particular increased risk of CHD and stroke mortality.
Acknowledgments The Three City Study was conducted under a
partnership agreement between the Institut National de la Santé et de la
Recherche Médicale (INSERM), the Victor Segalen–Bordeaux II
University, and Sanofi-Aventis. The Fondation pour la Recherche
Médicale financed the preparation and initiation of the study. The Three
City Study was also supported by the Caisse Nationale d’Assurance
Maladie des Travailleurs Salariés, the Direction Générale de la Santé,
the MGEN, the Institut de la Longévité, the Regional Councils of
Aquitaine and Bourgogne, the Fondation de France and the Ministry of
Research–INSERM Program ‘‘Cohortes et collections de données bi-
ologiques.’’ The funding sources had no role in the study design; the
collection, analysis and interpretation of data; in the writing of the
report; and in the decision to submit the article for publication.
Conflict of interest Dr. Tzourio serves on scientific advisory boards
for Merck Sharp & Dohme and the Fondation Plan Alzheimer, serves
on the editorial boards of Neuroepidemiology and the Journal of
Hypertension, and receives research support from the Agence Na-
tionale de la Recherche and Fondation Plan Alzheimer. Dr. Empana
has received consultancy honoraria from Lundbeck and speaker
honoraria from Pfizer.
References
1. Nicholson A, Kuper H, Hemingway H. Depression as an aetio-
logic and prognostic factor in coronary heart disease: a meta-
analysis of 6362 events among 146 538 participants in 54
observational studies. Eur Heart J. 2006;27:2763–74.
2. Empana JP, Jouven X, Lemaitre RN, et al. Clinical depression
and risk of out-of-hospital cardiac arrest. Arch Intern Med.
2006;166:195–200.
3. Whang W, Kubzansky LD, Kawachi I, et al. Depression and risk
of sudden cardiac death and coronary heart disease in women:
results from the Nurses’ Health Study. J Am Coll Cardiol. 2009;
53:950–8.
4. Larson SL, Owens PL, Ford D, et al. Depressive disorder, dys-
thymia, and risk of stroke: thirteen-year follow-up from the
Baltimore epidemiologic catchment area study. Stroke. 2001;32:
1979–83.
5. Ohira T, Iso H, Satoh S, et al. Prospective study of depressive
symptoms and risk of stroke among Japanese. Stroke. 2001;32:
903–8.
6. Jorm AF. The epidemiology of depressive status in the elderly:
implications for recognition, intervention, and prevention. Soc
Psychiatry Psychiatr Epidemiol. 1995;30:53–9.
7. Goldberg RJ, Yarzebski J, Lessard D, et al. A two-decades (1975
to 1995) long experience in the incidence, in-hospital and long-
123
256
term case-fatality rates of acute myocardial infarction: a com-
munity-wide perspective. J Am Coll Cardiol. 1999;33:1533–9.
8. Ryan J, Carriere I, Ritchie K, et al. Late-life depression and
mortality: influence of gender and antidepressant use. Br J Psy-
chiatry. 2008;192:12–8.
9. Ariyo AA, Haan M, Tangen CM, et al. Depressive symptoms and
risks of coronary heart disease and mortality in elderly Ameri-
cans. Cardiovascular Health Study Collaborative Research
Group. Circulation. 2000;102:1773–9.
10. Penninx BW, Guralnik JM, Mendes de Leon CF, et al. Cardio-
vascular events and mortality in newly and chronically depressed
persons [70 years of age. Am J Cardiol. 1998;81:988–94.
11. Wassertheil-Smoller S, Applegate WB, Berge K, et al. Change in
depression as a precursor of cardiovascular events. SHEP
Cooperative Research Group (systolic hypertension in the
elderly). Arch Intern Med. 1996;156:553–61.
12. Mendes de Leon CF, Krumholz HM, Seeman TS, et al. Depres-
sion and risk of coronary heart disease in elderly men and
women: New Haven EPESE, 1982–1991. Established populations
for the epidemiologic studies of the elderly. Arch Intern Med.
1998;158:2341–8.
13. Salaycik KJ, Kelly-Hayes M, Beiser A, et al. Depressive symp-
toms and risk of stroke: the Framingham Study. Stroke.
2007;38:16–21.
14. Bos MJ, Lindén T, Koudstaal PJ, et al. Depressive symptoms and
risk of stroke: the Rotterdam Study. J Neurol Neurosurg Psy-
chiatry. 2008;79:997–1001.
15. 3C Study Group. Vascular factors and risk of dementia: design of
the Three-City Study and baseline characteristics of the study
population. Neuroepidemiology. 2003;22:316–25.
16. Empana JP, Zureik M, Gariepy J, et al. The metabolic syndrome
and the carotid artery structure in noninstitutionalized elderly
subjects: the three-city study. Stroke. 2007;38:893–9.
17. Radloff LS. The CES-D Scale: a self-report depression scale for
research in the general population. Appl Psychol Meas.
1977;1:385–401.
18. Lawton MP, Brody EM. Assessment of older people: self-main-
taining and instrumental activities of daily living. Gerontologist.
1969;9:179–86.
19. Straczek C, Ducimetiere P, Barberger-Gateau P, et al. Higher
level of systemic C-reactive protein is independently predictive
of coronary heart disease in older community-dwelling adults: the
three-city study. J Am Geriatr Soc. 2010;58:129–35.
20. Bineau S, Dufouil C, Helmer C, et al. Framingham stroke risk
function in a large population-based cohort of elderly people: the
3C study. Stroke. 2009;40:1564–70.
123
R. Péquignot et al.
21. Horton NJ, Kleinman KP. Much ado about nothing: a comparison
of missing data methods and software to fit incomplete data
regression models. Am Stat. 2007;61:79–90.
22. Surtees PG, Wainwright NWJ, Luben RN, et al. Depression and
ischemic heart disease mortality: evidence from the EPIC-Nor-
folk United Kingdom Prospective Cohort Study. Am J Psychia-
try. 2008;165:515–23.
23. Pan A, Sun Q, Okereke OI, et al. Depression and risk of stroke
morbidity and mortality: a meta-analysis and systematic review.
JAMA. 2011;306:1241–9.
24. Carney RM, Blumenthal JA, Freedland KE, et al. Low heart rate
variability and the effect of depression on post-myocardial
infarction mortality. Arch Intern Med. 2005;165:1486–91.
25. Carney RM, Freedland KE, Miller GE, et al. Depression as a risk
factor for cardiac mortality and morbidity: a review of potential
mechanisms. J Psychosom Res. 2002;53:897–902.
26. Alexopoulos GS, Meyers BS, Young RC, et al. ‘Vascular
depression’ hypothesis. Arch Gen Psychiatry. 1997;54:915–22.
27. Hippisley-Cox J, Pringle M, Hammersley V, et al. Antidepres-
sants as risk factor for ischaemic heart disease: case-control study
in primary care. BMJ. 2001;323:666–9.
28. Smoller JW, Allison M, Cochrane BB, et al. Antidepressant use
and risk of incident cardiovascular morbidity and mortality
among postmenopausal women in the Women’s Health Initiative
study. Arch Intern Med. 2009;169:2128–39.
29. Gallo JJ, Bogner HR, Morales KH, et al. The effect of a primary
care practice-based depression intervention on mortality in older
adults: a randomized trial. Ann Intern Med. 2007;146:689–98.
30. Glassman AH, O’Connor CM, Califf RM, et al. Sertraline Anti-
depressant Heart Attack Randomized Trial (SADHEART) Group.
Sertraline treatment of major depression in patients with acute MI
or unstable angina. JAMA. 2002;288:701–9.
31. Berkman LF, Blumenthal J, Burg M, et al. Enhancing recovery in
coronary heart disease patients investigators (ENRICHD). Effects
of treating depression and low perceived social support on clin-
ical events after myocardial infarction: the enhancing recovery in
coronary heart disease patients (ENRICHD) randomized trial.
JAMA. 2003;289:3106–16.
32. Linke SE, Rutledge T, Johnson BD, et al. Depressive symptom
dimensions and cardiovascular prognosis among women with
suspected myocardial ischemia: a report from the National Heart,
Lung, and Blood Institute-sponsored Women’s Ischemia Syn-
drome Evaluation. Arch Gen Psychiatry. 2009;66:499–507.