Smoking determines the 10-year (2004–2014) prognosis in patients with Acute Coronary Syndrome: the GREECS observational study

Sampling procedure at baseline examination 2003–2004

GREECS is a prospective, observational study that was established in 2003. The main
goal of the study was to evaluate the annual incidence of ACS, as well as the role
of various CVD risk factors on the development and prognosis of ACS. From October
2003 to September 2004, n?=?2172 consecutive patients with discharge diagnosis of ACS (i.e., acute myocardial
infarction (AMI) or unstable angina (UA)) that were hospitalized in the cardiology
clinics or the emergency units of 6 major General Hospitals in Greece (i.e., Hippokration
hospital in Athens and the general prefectural hospitals in Lamia, Karditsa, Halkida,
Kalamata and Zakynthos island) were enrolled into the study (participation rate varied
from 80 to 95 %). The hospitals were selected in order to represent populations with
various socio-economic, cultural and regional characteristics. Of the enrolled patients,
n?=?1649 (76 %) were men (65?±?13 years) and n?=?523 (24 %) were women (62?±?11 years) (p for age and gender differences 0.001). With the exception of Athens, where there
are several other hospitals, all the other hospitals cover the whole population of
the aforementioned regions, including urban and rural areas. At entry, as well as
during hospitalization biomarkers suggesting cardiac injury and AMI were measured.
Moreover a 12-lead electrocardiogram (ECG) was performed and clinical symptoms were
evaluated in all patients, by a cardiologist. AMI and UA were defined following the
up-to-date definitions 18], 19]. Medical information was retrieved through hospital records.

Investigated measurements at baseline examination

The baseline examination included a variety of patients’ clinical, biochemical, socio-demographic
and lifestyle characteristics. Particularly, socio-demographic and lifestyle characteristics
included: age, sex, physical activity, diet and smoking, years at school, financial
and marital status and psychological evaluation. In particular, as regards smoking,
patients were asked whether they were current, former or never smokers. Current smokers
were defined as those who smoked at least 1 cigarette/day or have attempted to quit
smoking during the past 12 months, while the rest who smoked at some time were defined
as past or former smokers. The rest of the patients were defined as never or occasional
smokers 20]. Questions about years of smoking exposure, type of cigarettes smoked (i.e., light,
heavy), number of cigarettes/day, smoking at work or/and home place, were asked; for
the former smokers information about years of smoking cessation was also recorded.
Patients were divided into four quartiles (statistically) according to the distribution
of the packs smoked per year: (a) 1
^st
quartile (0 pack/years), (b) 2
^nd
quartile (30 pack/years), (c) 3
^rd
quartile (30–60 pack/years) and (d) 4
^th
quartile (60 pack/years). The certain classification provided a better distribution
of the sample resulting in balanced subgroups. Special attention was also given to
the baseline exposure of second-hand smoke (in years), for at least 30 min per day,
to cigarette smoke, at home, workplace, as well as in indoor recreational environments.
As regards other major characteristics, financial status was classified – according
to the Greek tax cut-offs – as: “low” (9000€), “moderate” (18,000€), “good” (48,000€)
and “very good” (48,000€). Dietary habits were evaluated using a validated food frequency
questionnaire and the level of adherence to the Mediterranean dietary pattern was
assessed using the MedDietScore (range 0–55) 21]. Higher values of this diet score indicate greater adherence to the Mediterranean
diet. Physical activity was evaluated through a self-reported questionnaire provided
by the American College of Sports Medicine 22] and it was defined as any engagement in activities of at least 3 times/week and for
at least 30 min. As regards medical history, it was retrieved during the physical
examination and through the patient’s medical records and included the detailed assessment
of hypertension, hypercholesterolemia, diabetes and any previous CVD event (i.e.,
prior to the baseline), as well as the pharmaceutical treatment and management of
these conditions. Body mass index (BMI) was calculated as weight (in Kg) divided by
height (in m) squared. Obesity was defined as BMI??29.9 kg/m
²
.

Further details about the aims, measurements and baseline procedures of the GREECS
study may be found elsewhere 17]

10-year follow-up evaluation

During 2013–2014, the 10-year follow-up of the patients was performed by the study’s
investigators. Information from n?=?1918 of the initially enrolled patients was retrieved; the remaining n?=?254 patients were lost after the 1st year of follow-up and considered as censored
in the statistical analysis; no vital status information at 10-year was available
for these patients (i.e., loss to follow-up around 11 %). Vital status and development
of ACS was evaluated using WHO-ICD-9 coding (as it was also performed in the 30-day,
6-month and 1-year follow-up that has been reported in previous publications) 17]. All patients were interviewed by using a standard questionnaire. Smoking during
the 10-year follow-up period was also assessed; for current smokers, number of cigarettes/day
and years of smoking were asked, while for the former smokers information about the
year of smoking cessation was recorded. Moreover, exposure to secondhand smoke was
also obtained, following the same methodology described above. Regarding patients
who died within the decade and in order to have an accurate death diagnosis, relevant
information was retrieved from the medical records, or local mortality registries.

No differences were observed between those participated in the 10-year follow-up and
those lost in follow-up, in all baseline clinical and lifestyle factors (all p’s 0.50).

Endpoints at follow-up

The endpoints studied in the 10-year follow-up were recurrent fatal or non-fatal ACS
events. In particular, the development of a new AMI, angina pectoris, other identified
forms of ischemia (WHO-ICD coding 410–414.9, 427.2, 427.6), heart failure of different
types and chronic arrhythmias (WHO-ICD coding 400.0–404.9, 427.0–427.5, 427.9), were
recorded by the physicians of the study.

Bioethics

The study was approved by the Medical Research Ethics Committee of the participating
Institutions and was carried out in accordance with the Declaration of Helsinki (1989)
of the World Medical Association. All patients were informed about the aims and procedures
of the study and signed an informed consent.

Statistical analysis

Continuous variables are presented as mean values?±?standard deviation, while categorical
variables are presented as absolute and relative (%) frequencies. Associations between
normally distributed continuous variables (i.e., MedDietScore, body mass index and
age) and groups of the patients per smoking quartile were evaluated by the analysis
of variance (ANOVA), after controlling for equality of variances (homoscedacity).
Due to multiple comparisons the Bonferroni rule was applied to correct for the inflation
of Type – I error. Years of school variable that was abnormally distributed was tested
through Kruskal-Wallis. Associations between categorical variables (i.e., sex, physical
activity, financial status, hypertension, hypercholesterolemia, diabetes mellitus,
family history of CVD) were tested by the use of the chi-squared test. Survival curves
according to quartiles of pack-years of smoking were calculated and log-rank test
was implemented to evaluate median follow-up differences between quartiles. In order
to control residual confounding, which may exist between smoking and ACS incidence,
nested models were estimated. Thus, the association between patients’ smoking (i.e.,
smoking status estimated with pack-years and smoking cessation after the baseline
ACS or continue, within the decade) and the dependent variable (i.e., 10-year ACS
fatal/non fatal events), after controlling for the above mentioned potential confounders,
was evaluated by the use of nested Cox proportional hazard models. Proportionality
of the hazards was graphically tested by plotting the log (-log(survival)) vs. the
log- of survival time. First order interactions between age, sex, medical history
and smoking were also evaluated for potential stratifying analyses. Appropriate tests
for goodness-of-fit (i.e., deviance and Pearson’s residuals) were applied in order
to evaluate the robustness of the models’ estimates rather than create prediction
models. Results are presented as hazard ratios (HR) and their corresponding 95 % confidence
intervals (95 % CI). All statistical calculations were performed with the SPSS version
21 software (IBM Hellas Inc, Athens, Greece).