Eur J Epidemiol (2008) 23:731–737 DOI 10.1007/s10654-008-9291-x
CARDIOVASCULAR DISEASE
Cardiovascular risk factors and mortality in patients with coronary heart disease Christof Prugger Æ Ju¨rgen Wellmann Æ Jan Heidrich Æ Stefan-Martin Brand-Herrmann Æ Ulrich Keil
Received: 14 March 2008 / Accepted: 24 September 2008 / Published online: 15 October 2008 Ó Springer Science+Business Media B.V. 2008
Abstract A priority in preventive cardiology is to reduce the number of recurrent events and to prolong survival in patients with established coronary heart disease (CHD). Aim of the present study was to examine risk factors for long-term mortality in CHD patients who entered routine secondary prevention after a coronary event or intervention. Such patients, from the EUROASPIRE (EUROpean Action on Secondary Prevention through Intervention to Reduce Events) I and II studies in the region of Mu¨nster, Germany, were followed over a mean period of 8.0 years up to the end of 2005. Patients were up to 70 years of age at baseline when they were interviewed and examined using standardised methods. Baseline examination was carried out at least 6 months and at a mean of 19.5 months after the coronary event or procedure. In 367 patients from EUROASPIRE I and 380 patients from EUROASPIRE II, a total of 125 deaths (16.7%) occurred during follow-up. Multivariate analyses, using Cox proportional hazards models, established diabetes mellitus and smoking as predictors for all-cause mortality with estimated hazard rate ratios (HRRs) of 2.24 (95% confidence interval (CI): 1.43–3.49) and 1.95 (95% CI: 1.23–3.10), respectively. Significant associations were found between diabetes mellitus and cardiovascular (HRR 2.36; 95% CI: 1.31–4.24) as well as CHD mortality (HRR 2.40; 95% CI: 1.25–4.59). Systolic blood pressure was significantly associated with increased cerebrovascular disease mortality (HRR 1.04; 95% CI:
1.01–1.08 for 1 mmHg increase). In conclusion, long-term mortality in coronary patients from routine secondary prevention is substantial. Diabetes mellitus and smoking represent key issues in patients with established CHD. Keywords Coronary heart disease Germany Mortality Secondary prevention Abbreviations BMI CABG CHD CVD DALYs EUROASPIRE
HDL HRR OR PTCA RCT WHO
Body mass index Coronary artery bypass graft Coronary heart disease Cardiovascular disease Disability adjusted life years EUROpean Action on Secondary Prevention through Intervention to Reduce Events High density lipoprotein Hazard rate ratio Odds ratio Percutaneous transluminal coronary angioplasty Randomized controlled trial World Health Organisation
Introduction C. Prugger J. Wellmann J. Heidrich U. Keil (&) Institute of Epidemiology and Social Medicine, University of Mu¨nster, Domagkstr. 3, 48149 Mu¨nster, Germany e-mail:
[email protected] S.-M. Brand-Herrmann Leibniz-Institute for Arteriosclerosis Research, Department of Molecular Genetics of Cardiovascular Disease, University of Mu¨nster, Domagkstr. 3, 48149 Mu¨nster, Germany
Coronary heart disease (CHD) is the leading cause of morbidity and mortality worldwide and is expected to be the most common and disabling disease in the year 2020 [1]. In 2002, CHD accounted for 2.4 million deaths and was the condition causing the highest burden of disability adjusted life years (DALYs) in the WHO European Region [2]. Thus, prevention and control of CHD play a key role in health promotion.
123
732
The Joint European Societies define patients with established CHD as a top priority group for prevention in clinical practice [3]. Compared to the general population, patients with CHD are at increased risk of total and cardiovascular disease (CVD) mortality [4]. There is epidemiological evidence that risk factors for recurrent non-fatal and fatal CHD resemble risk factors for primary events [3]. A number of cohort studies comprising CHD patients form randomized controlled trials (RCTs) identified significant independent predictors of recurrent CHD events and death [5, 6]. However, transferability of results from RCTs is often limited. This can be due to the experimental study design, a short follow up period, or selection of a non-representative study population. The EUROASPIRE (EUROpean Action on Secondary Prevention through Intervention to Reduce Events) study offers the opportunity to investigate the influence of cardiovascular risk factors on mortality in routine secondary prevention of CHD. EUROASPIRE is a series of cross-sectional studies to evaluate risk factor management, lifestyle interventions, and prophylactic drug use in patients surviving at least 6 months after coronary events or procedures [7, 8]. Few prospective observational studies have addressed the predictive value of cardiovascular risk factors on long-term mortality in CHD patients from clinical practice over a longer follow-up period [9–13]. Therefore, the present study investigates the relationships between cardiovascular risk factors and long-term mortality in CHD patients from routine secondary prevention using data from EUROASPIRE I and II in the region of Mu¨nster.
C. Prugger et al.
Baseline Patients were interviewed and examined using standardised methods at least 6 months and at a mean of 19.5 (standard deviation: 8.0) months after having experienced their coronary event or procedure. The interview assessed personal and demographic details, socioeconomic factors, CHD and diabetes history, history of cardiovascular risk factors, and current medication. The clinical examination included measurements of height, weight, breath carbon monoxide, and blood pressure. Height and weight were measured in patients wearing light indoor clothes and no shoes. Breath carbon monoxide was assessed by means of a smokerlyser (Bedfont Scientific, Model EC 50). Blood pressure was measured in a sitting position on the right upper arm using an automatic digital sphygmomanometer (Takeda UA 731 in EUROASPIRE I and Omron 711 in EUROASPIRE II). Venous blood samples were drawn for the analysis of total cholesterol, high density lipoprotein (HDL) cholesterol, triglycerides, and plasma glucose. All analyses were performed at the central study laboratory. Since devices for the measurement of blood pressure and total/HDL cholesterol differed between surveys, results from EUROASPIRE I had to be adjusted for comparability. Diastolic blood pressure was adjusted by adding 2.64 mmHg to original values, total and HDL cholesterol by multiplying original values by 1.13 [15, 16].
Follow-up Patients and methods EUROASPIRE I and II were conducted in 1995–1996 and in 1999–2000, respectively. The study population of EUROASPIRE I and II in the region of Mu¨nster has been described in detail elsewhere [14, 15]. Briefly, patients up to 70 years of age hospitalised for CHD were identified retrospectively from hospital records by one of the following discharge diagnoses: (a) first coronary artery bypass graft (CABG), (b) first percutaneous transluminal coronary angioplasty (PTCA) without previous CABG, (c) first or recurrent acute myocardial infarction, and (d) first or recurrent acute myocardial ischemia without signs of infarction. Patients’ events and procedures had to date back at least 6 months prior to inclusion into the study. Hereby, the health care system was given a period of time considered to be sufficient for the implementation of risk factor management, lifestyle interventions, and prophylactic drug use. Recruitment of patients started with the most recent hospital record and proceeded backwards in time until the required sample size was achieved. The objective was to obtain information from equal numbers of study participants in the diagnostic groups described above.
123
Vital status of patients was registered up to December 31, 2005 using municipal records. Death certificates were obtained from local health authorities. Certificates were reviewed independently by two physicians who classified specific cause of death according to the following predefined categories: (1) CHD, (2) cerebrovascular disease, (3) other vascular disease, (4) cancer, (5) other cause of death, and (6) unknown cause of death. If classification of cause of death differed, a consensus decision was made.
Statistics Readiness to participate among all patients identified in EUROASPIRE I and II was assessed by conditional logistic regression analysis considering age, sex, diagnostic group, and time-lag between coronary event and end of recruitment period. Patients’ demographic and clinical characteristics at the time of interview were compared between both EUROASPIRE surveys using the Kolmogorov–Smirnov-test for differences in continuous variables and the v2-test for those in proportions. Survival of patients
Cardiovascular risk factors and mortality in patients with coronary heart disease
was compared between surveys using the log-rank test. Subsequent analyses were based on pooled data from EUROASPIRE I and II. Mean length of follow-up was quantified by Kaplan–Meier estimates to accommodate for shorter follow up times in the deceased [17]. Cumulative incidence functions that account for competing risks were drawn for all-cause and cause specific mortality [18]. Multivariate analysis was performed using Cox proportional hazards models to estimate hazard rate ratios (HRRs) for the association between patients’ characteristics assessed at interview and all-cause as well as cause-specific mortality. Follow-up time for the Cox model started with hospitalization for the coronary event or intervention. Since patients entered the study at interview at least six months later, their event times are left truncated. To accommodate for this fact, a Cox model with ‘‘delayed entry’’ was used [19]. Cox models allowed for different baseline hazards in EUROASPIRE I and II and were adjusted for diagnostic group with myocardial ischemia as the reference category. Established main cardiovascular risk factors were included in the Cox models. Age, blood pressure, and total/HDL cholesterol ratio were used as continuous variables. The remaining risk factors were used as categorical variables. Statistical tests were two-tailed and a 5% level was used for statistical significance. All statistical analyses were performed using SAS software version 9.1 (SAS Institute Inc., Cary, NC, USA).
Results In EUROASPIRE I and II a total of 524 and 684 CHD patients were identified through hospital records, respectively. Overall, 44 patients from the first survey and 54 patients from the second survey could not be reached and 16 patients from EUROASPIRE I and 26 patients from EUROASPIRE II died before they could be contacted. Further 72 patients in EUROASPIRE I and 202 patients in EUROASPIRE II refused to participate in the study leaving 392 and 402 patients who attended the interview. Odds ratio (OR) for non-participation did not change significantly with increasing time-lag between coronary event and end of recruitment period (OR: 0.994; 95% confidence interval (CI): 0.976–1.013 per month increase). Only patients with complete information on age, sex, smoking status, history of previous CHD, diabetes, total/HDL cholesterol, blood pressure, and vital status were considered for the present analyses. Therefore, a total of 47 individuals were excluded from the original dataset: 2 patients for missing information on vital status, 24 patients for missing values on total/HDL cholesterol, 3 patients for missing values on blood pressure, and 18 patients for missing values on medication use. Further 6 deceased patients with
733
unknown cause of death were not considered in the analyses of cause-specific mortality. Table 1 presents demographic and clinical characteristics of the remaining 747 patients at the time of interview stratified by EUROASPIRE survey. Significant differences were found in body mass index (BMI), diastolic blood pressure, total cholesterol, HDL cholesterol, and use of lipid-lowering medication. cA total of 125 patients (16.7%) died until 31 December 2005, 84 from the first survey and 41 from the second survey. Women accounted for 25.0% and 22.0% of deaths, respectively. In EUROASPIRE I (EUROASPIRE II), 41 (16) patients died due to CHD, 4 (5) due to cerebrovascular disease, 1 (0) due to other vascular disease, 20 (10) due to cancer, 13 (9) due to other cause of death, and 5 (1) due to unknown cause of death. The proportion of patients from EUROASPIRE I and II surviving 5 years after the interview was 89.1% and 91.3%, respectively. All-cause mortality in patients from the first and second EUROASPIRE study did not differ significantly (log rank test, p-value 0.65). Therefore, patients from both EUROASPIRE surveys were considered together for further analyses. In pooled EUROASPIRE surveys, mean follow-up time from the index event and from the interview was 9.7 and 8.0 years, respectively. Overall, 53.6% of deaths were due to CVD of which 85.1% were attributable to CHD. Cumulative incidence functions for all-cause mortality as well as cause-specific mortality as function of time since the interview are presented in Fig. 1. The cause-specific cumulative incidence functions add up to the all-cause mortality figure. For example, it is estimated that about 12.5% of patients will die within the first 7 years after the interview, where CHD contributes 6.5%, cerebrovascular disease 0.9%, other vascular disease 0.1%, and non-CVD 4.9%. CVD mortality rate, defined as deaths due to CHD, cerebrovascular, and other vascular disease per 1,000 person-years (py) was 12.2 whereas all-cause mortality rate was 22.7/1,000 py. CVD and CHD mortality rates were highest in the acute myocardial infarction group with 17.1 and 14.4/1,000 py, respectively. Lowest CVD and CHD mortality rates were found in patients with PTCA (8.1 and 7.4/1,000 py). Tables 2 and 3 show results from Cox proportional hazards models. Diabetes (HRR: 2.24; 95% CI: 1.43–3.49) and smoking (HRR 1.95; 95% CI: 1.23–3.10) were independent predictors for all-cause mortality during the follow-up period. Multivariate analyses for cause-specific death revealed significant HRRs for diabetes as a predictor for CVD (HRR: 2.36, 95% CI: 1.31–4.24) and CHD mortality (HRR: 2.40, 95% CI: 1.25–4.59). Age was found to be an independent predictor in all Cox proportional hazards models with the strongest association in cerebrovascular mortality (HRR: 1.21, 95% CI: 1.01–1.44 per year increase). In the analysis of cerebrovascular
123
734
C. Prugger et al.
Table 1 Patients’ demographic and clinical characteristics at interview Characteristic
EUROASPIRE I (N = 367)
EUROASPIRE II (N = 380)
Age (mean (SD), year)
60.6 (7.6)
61.1 (7.7)
0.274
Women (%)
22.6
20.3
0.476
Smoking (%)a
15.5
16.3
0.842
b
14.2
14.2
1.000
Diabetes (%)
P-value
BMI (kg/m2) BMI (mean (SD))
27.7 (3.4)
28.4 (3.9)
0.043
BMI C 30 (%)
23.2
30.0
0.039
SBP (mean (SD))
140.5 (23.1)
141.8 (22.0)
0.457
DBP (mean (SD))
84.9 (10.8)
86.5 (11.9)
0.013
57.2
61.1
0.298
Blood pressure (mmHg)
SBP C 140 and/or DBP C 90 (%) Lipids (mmol/l) Total cholesterol (mean (SD))
6.1 (1.1)
5.5 (1.1)
\0.001
HDL cholesterol (mean (SD))
1.4 (0.4)
1.3 (0.3)
\0.001
Total/HDL cholesterol ratio (mean (SD))
4.6 (1.5)
4.6 (1.4)
0.900
BP-lowering medication (%)c
84.5
88.9
0.084
Lipid-lowering medication (%)d
36.2
68.7
\0.001
Anti-platelet therapies (%)e
86.6
86.3
0.915
Use of medication
a
Self-reported smoking and/or [10 ppm carbon monoxide in breath
b
Self-reported diabetes
c
Use of beta-blockers, ACE inhibitors, calcium antagonists, diuretics, angiotensin II receptor antagonists, potassium channel openers, or other antihypertensive drugs
d
Use of statins or other lipid-lowering drugs
e
Use of aspirin or other anti-platelet drugs; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure
between all-cause or cause-specific mortality and BMI, diastolic blood pressure, and total/HDL cholesterol.
percent deceased
30
20
Discussion 10
0 0
1
2
3
4
5
6
7
8
9
10
Follow– up time since interview in years CHD
cerebrovascular
other CVD
non – CVD
Fig. 1 Cumulative incidence functions for patients from EUROASPIRE I and II by all-cause and cause-specific mortality
disease mortality, a significant association was found for systolic blood pressure (HRR: 1.04, 95% CI: 1.01–1.08 for 1 mmHg increase). The use of anti-platelet therapies was negatively associated with CHD mortality (HRR: 0.42, 95% CI: 0.22–0.83). No significant associations were found
123
In EUROASPIRE, patients are recruited at least 6 months after having experienced a coronary event or intervention in order to give the health care system ample time to implement risk factor management and to initiate lifestyle interventions and prophylactic drug use. Thus, the present study sums up to an analysis of mortality as a function of secondary prevention of CHD in clinical practice. Compared to other studies, EUROASPIRE offers the opportunity to follow patients over a relatively long period of time. Mortality due to CVD accounted for more than half of all deaths in our study population. Main predictor of all-cause and cause-specific mortality was a self-reported history of diabetes. Smoking was found to be a significant predictor for all-cause mortality. Each 1 mmHg rise in systolic blood pressure was associated with a significantly increased cerebrovascular disease mortality.
Cardiovascular risk factors and mortality in patients with coronary heart disease
735
Table 2 HRRs for the association between patients’ characteristics assessed at interview and total and cardiovascular mortality in patients pooled from EUROASPIRE I and II studies Cardiovascular mortalitya (67 deaths)
Total mortality (125 deaths) HRR (95% CI)
P-value
HRR (95% CI)
P-value
Women
0.87 (0.56–1.36)
0.543
1.14 (0.65–2.00)
0.642
Ageb (year)
1.07 (1.04–1.11)
\0.001
1.07 (1.02–1.11)
0.002
Smokingc Diabetesd
1.95 (1.23–3.10) 2.24 (1.43–3.49)
0.005 \0.001
1.58 (0.79–3.15) 2.36 (1.31–4.24)
0.196 0.004
History of CHDe
1.23 (0.79–1.93)
0.363
1.75 (0.93–3.27)
0.082
Diagnostic group
0.072
–
0.904
–
Myocardial ischemia
1.00 (Ref.)
–
1.00 (Ref.)
–
PTCA
0.85 (0.49–1.46)
0.548
0.93 (0.39–2.18)
0.859
CABG
0.89 (0.52–1.51)
0.657
1.20 (0.55–2.63)
0.642
AMI
1.02 (0.59–1.77)
0.948
2.36 (1.07–5.19)
0.033
BMI C 30 kg/m2
0.96 (0.64–1.46)
0.858
0.88 (0.50–1.55)
0.660
SBP (mmHg)
1.00 (0.99–1.01)
0.734
1.00 (0.99–1.02)
0.755
DBP (mmHg)
0.99 (0.97–1.01)
0.518
0.98 (0.95–1.01)
0.120
Total/HDL cholesterol ratio
1.00 (0.88–1.14)
0.959
1.07 (0.91–1.26)
0.435
BP-lowering medicationf
1.89 (0.98–3.66)
0.058
2.12 (0.75–5.94)
0.155
Lipid-lowering medicationg
0.77 (0.51–1.14)
0.192
0.80 (0.46–1.38)
0.416
0.80 (0.48–1.32)
0.376
0.54 (0.28–1.04)
0.066
Anti-platelet therapies
h
a
Death due to CHD, cerebrovascular disease, or other vascular disease; HRRs, hazard rate ratios from Cox proportional hazards models adjusted for all co-variables listed in the table; analysis based on 741 patients found alive at 31.12.2005 or deceased with known cause of death
b
Age at event or procedure
c
Self-reported smoking and/or [10 ppm carbon monoxide in breath
d
Self-reported diabetes
e
CHD prior to the index event
f
Use of beta-blockers, ACE inhibitors, calcium antagonists, diuretics, angiotensin II receptor antagonists, potassium channel openers, or other antihypertensive drugs
g
Use of statins or other lipid-lowering drugs
h
Use of aspirin or other anti-platelet drugs; PTCA, percutaneous transluminal coronary angioplasty; CABG, coronary artery bypass graft; AMI, acute myocardial infarction; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure
The results of our study are in line with the findings from the EUROASPIRE I cohort study conducted in nine European countries including Germany where diabetes and smoking showed strong associations with total (HRR: 1.9 and 2.0), CVD (HRR: 2.5 and 2.2) and CHD (HRR: 2.5 and 2.4) mortality [9]. Obesity, hypertension, and raised plasma lipids, which are all well-established cardiovascular risk factors, did not show statistically significant associations with mortality. Our findings are partly consistent with results from other observational studies. The Stockholm Heart Epidemiology Program (SHEEP) found diabetes as the most important risk factor for recurrent fatal and nonfatal CHD in 1,635 patients after myocardial infarction followed for 7.2 years, whereas hypertension and high total cholesterol were of no predictive importance [10]. In a cohort of 1,038 CHD patients from Germany, smoking, diabetes, serum low density lipoprotein cholesterol levels, prior history of myocardial infarction, but not hypertension were independent predictors of mortality over 6.4 years
[11]. A study in 2,677 patients with myocardial infarction from a health maintenance organisation found diabetes, continued smoking, and low HDL cholesterol as independent predictors of mortality over 3.4 years [12]. A study among 459 patients who experienced a myocardial infarction in the Framingham Study followed for a maximum of 32 years (mean 9.7 years) up to the 1980s revealed diabetes, systolic blood pressure, and serum cholesterol as independent risk factors for coronary death [13]. Thus, with regard to blood pressure, lipids, and smoking status the results of our study differ to some extent from the findings of previous studies. There are several explanations for these discrepancies. In prospective observational studies, risk factor measurements at a single occasion (i.e. at baseline) typically lead to underestimation of the association between risk factors and disease (‘‘regression dilution bias’’) [20, 21]. Additionally, recent improvement of blood pressure and lipid levels as well as smoking cessation during followup in our study population might have further attenuated the
123
736
C. Prugger et al.
Table 3 HRRs for the association between patients’ characteristics assessed at interview and CHD and cerebrovascular mortality in patients pooled from EUROASPIRE I and II studies CHD mortalitya (57 deaths) HRR (95% CI)
Cerebrovascular mortalitya (9 deaths) P-value
HRR (95% CI)
P-value
Women
1.30 (0.71–2.36)
0.397
0.56 (0.10–3.07)
0.500
Ageb (year)
1.07 (1.02–1.11)
0.005
1.21 (1.01–1.44)
0.038
Smokingc Diabetesd
1.99 (0.98–4.01) 2.40 (1.25–4.59)
0.056 0.008
0.00 (-) 2.24 (0.45–11.21)
0.995 0.325
History of CHDe
1.71 (0.86–3.40)
0.124
2.52 (0.42–15.26)
0.316
Diagnostic group
0.831
–
0.123
–
Myocardial ischemia
1.00 (Ref.)
–
1.00 (Ref.)
–
PTCA
1.02 (0.40–2.56)
0.973
0.42 (0.04–4.97)
0.489
CABG
1.31 (0.56–3.07)
0.538
0.72 (0.11–4.76)
0.732
AMI
2.46 (1.04–5.81)
0.040
1.32 (0.17–10.34)
0.793
0.79 (0.42–1.50)
0.478
1.71 (0.41–7.11)
0.461
BMI C 30 kg/m2 SBP (mmHg)
1.00 (0.98–1.01)
0.592
1.04 (1.01–1.08)
0.025
DBP (mmHg)
0.98 (0.95–1.01)
0.230
0.95 (0.89–1.03)
0.223
Total/HDL cholesterol ratio
1.09 (0.91–1.30)
0.354
1.08 (0.69–1.69)
0.751
BP-lowering medicationf
2.56 (0.78–8.36)
0.121
0.55 (0.06–5.19)
0.599
Lipid-lowering medicationg
0.80 (0.44–1.45)
0.462
0.88 (0.18–4.42)
0.879
Anti-platelet therapies
h
0.42 (0.22–0.83)
0.012
–
i
–i
HRRs, hazard rate ratios from Cox proportional hazards models adjusted for all co-variables listed in the table a
Analysis based on 741 patients found alive at 31.12.2005 or deceased with known cause of death
b
Age at event or procedure
c
Self-reported smoking and/or [10 ppm carbon monoxide in breath
d
Self-reported diabetes
e
CHD prior to the index event
f
Use of beta-blockers, ACE inhibitors, calcium antagonists, diuretics, angiotensin II receptor antagonists, potassium channel openers, or other antihypertensive drugs
g
Use of statins or other lipid-lowering drugs
h
Use of aspirin or other anti-platelet drugs All nine patients who died from cerebrovascular disease had used anti-platelet therapies by the time of interview. Therefore, use of anti-platelet therapies as an indicator variable had to be omitted from the model; PTCA, percutaneous transluminal coronary angioplasty; CABG, coronary artery bypass graft; AMI, acute myocardial infarction; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure
i
predictive value of baseline risk factor levels on mortality. Use of lipid lowering and antihypertensive agents in secondary prevention of CHD increased substantially over the last decade. Use of statins in our study population increased from 31.9% in 1995–1996 to 65.7% in 1999–2000 [15]. In EUROASPIRE I and II, 43.6% and 67.9% of study participants used beta-blockers whereas 31.4% and 44.8% used ACE inhibitors [15]. Furthermore, the quality of lipid lowering and antihypertensive treatment might have improved over the follow-up period. Epidemiological studies show that smoking prevalence decreases with increasing age [22, 23]. Our findings concerning baseline blood pressure, total/HDL cholesterol, and smoking status might therefore at least in part be explainable by improved control of these risk factors over the follow-up time. The observed predictive value of systolic blood pressure on cerebrovascular disease mortality might reflect its stronger
123
association with stroke compared to CHD mortality [21]. Elevated HRRs for total, CVD, and CHD mortality in patients using blood pressure lowering medication might be owing to confounding by indication. Our study has a number of limitations. No repeated measurements of risk factors were available from study participants. As discussed above, this may have led to an underestimation of the associated effects. The time-lag between coronary event and entry in the study varies between study participants. Sampling from patients with longer time-lag may have lead to selection of patients who survived. Therefore, in the Cox models time-lag was considered appropriately. Participants with longer time-lag may also differ in some other respects from those with shorter time-lag. Since information on patients’ characteristics before the baseline interview is limited, we were unable to adjust for such potential differences in the
Cardiovascular risk factors and mortality in patients with coronary heart disease
analysis. However, as outlined in the result section, readiness to participate in the surveys did not change with increasing time-lag, making selection and unmeasured confounding less likely. Overall, the final analysis was based on 69.9% of patients contacted and found alive. Furthermore, an array of explanatory variables was introduced in the Cox models for cause-specific mortality, but the number of events was limited. No information was collected on recurrent non-fatal vascular disease during the follow-up period. However, considering the limitations in the availability and reliability of CHD morbidity data, use of death as an endpoint is a reasonable alternative. Accuracy of death certificates in providing cause of death is discussed controversially. However, as rather broad categories of cause of death were used, misclassification due to crossover between cause-specific death groups is unlikely. In summary, long-term mortality in coronary patients from routine secondary prevention surviving at least 6 months after hospitalisation is substantial. In this cohort of patients with established CHD, diabetes mellitus and smoking are strongest predictors of mortality and represent key issues of secondary prevention. Acknowledgments We thank all patients for participation in the study. We also thank the European Society of Cardiology (ESC), Sophia Antipolis, France, for financial support in EUROASPIRE I and II. Unrestricted educational grants were obtained from MSD Sharp & Dohme, Haar (EUROASPIRE I), and Pfizer, Karlsruhe and AstraZeneca, Wedel/Holstein (EUROASPIRE II).
8.
9.
10.
11.
12.
13.
14.
15.
16.
References 1. Lopez AD, Murray CC. The global burden of disease, 1990– 2020. Nat Med. 1998;4:1241–3. doi:10.1038/3218. 2. World Health Organisation. World Health Report 2004. Geneva: World Health Organisation; 2004. 3. European Guidelines on Cardiovascular Disease Prevention in Clinical Practice. Fourth Joint European Societies’ Task Force on Cardiovascular Disease Prevention in Clinical Practice. Eur J Cardiovasc Prev Rehabil. 2007;14:S1–113. 4. Chambless L, Keil U, Dobson A, et al. Population versus clinical view of case fatality from acute coronary heart disease: results from the WHO MONICA Project 1985–1990. Multinational MONItoring of Trends and Determinants in CArdiovascular Disease. Circulation. 1997;96:3849–59. 5. Marschner IC, Colquhoun D, Simes RJ, et al. Long-term risk stratification for survivors of acute coronary syndromes. Results from the Long-term Intervention with Pravastatin in Ischemic Disease (LIPID) Study. LIPID Study Investigators. J Am Coll Cardiol. 2001;38:56–63. doi:10.1016/S0735-1097(01)01360-2. 6. Wilhelmsen L, Pyorala K, Wedel H, et al. Risk factors for a major coronary event after myocardial infarction in the Scandinavian Simvastatin Survival Study (4S). Impact of predicted risk on the benefit of cholesterol-lowering treatment. Eur Heart J. 2001;22: 1119–27. doi:10.1053/euhj.2000.2481. 7. EUROASPIRE Study Group. EUROASPIRE: European action on secondary prevention through intervention to reduce events. A
17.
18.
19.
20.
21.
22.
23.
737
European Society of Cardiology survey of secondary prevention of coronary heart disease: principal results. Eur Heart J. 1997; 18:1569–82. EUROASPIRE Study Group. Lifestyle and risk factor management and use of drug therapies in coronary patients from 15 countries; principal results from EUROASPIRE II Euro Heart Survey Programme. Eur Heart J. 2001;22:554–72. doi:10.1053/euhj.2001. 2610. De Bacquer D, De Backer G, Ostor E, et al. Predictive value of classical risk factors and their control in coronary patients: a follow-up of the EUROASPIRE I cohort. Eur J Cardiovasc Prev Rehabil. 2003;10:289–95. doi:10.1097/00149831-20030800000012. Leander K, Wiman B, Hallqvist J, et al. Primary risk factors influence risk of recurrent myocardial infarction/death from coronary heart disease: results from the Stockholm Heart Epidemiology Program (SHEEP). Eur J Cardiovasc Prev Rehabil. 2007;14:532–7. doi:10.1097/HJR.0b013e328012e3cc. Vo¨lzke H, Henzler J, Menzel D, et al. Outcome after coronary artery bypass graft surgery, coronary angioplasty and stenting. Int J Cardiol. 2007;116:46–52. doi:10.1016/j.ijcard.2006.02.008. Kaplan RC, Heckbert SR, Furberg CD, et al. Predictors of subsequent coronary events, stroke, and death among survivors of first hospitalized myocardial infarction. J Clin Epidemiol. 2002;55:654–64. doi:10.1016/S0895-4356(02)00405-5. Wong ND, Cupples LA, Ostfeld AM, et al. Risk factors for longterm coronary prognosis after initial myocardial infarction: the Framingham Study. Am J Epidemiol. 1989;130:469–80. Enbergs A, Liese A, Heimbach M, et al. Evaluation of secondary prevention of coronary heart disease. Results of the EUROASPIRE study in the Mu¨nster region. Z Kardiol. 1997;86:284–91. doi:10.1007/s003920050060. Heidrich J, Liese AD, Kalic M, et al. Secondary prevention of coronary heart disease Results from EUROASPIRE I and II in the region of Mu¨nster, Germany. Dtsch Med Wochenschr. 2002;127:667–72. doi:10.1055/s-2002-23480. EUROASPIRE I and II Group. European Action on Secondary Prevention by Intervention to Reduce Events. Clinical reality of coronary prevention guidelines: a comparison of EUROASPIRE I and II in nine countries. Lancet. 2001;357:995–1001. doi: 10.1016/S0140-6736(00)04235-5. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17:343–6. doi: 10.1016/0197-2456(96)00075-X. Gooley TA, Leisenring W, Crowley J, et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695–706. doi:10.1002/(SICI) 1097-0258(19990330)18:6\695::AID-SIM60[3.0.CO;2-O. Hosmer DW, Lemeshow S. Applied survival analysis. Regression modeling of time to event data. New York: John Wiley & Sons; 1999. MacMahon S, Peto R, Cutler J, et al. Blood pressure, stroke, and coronary heart disease Part 1. Prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet. 1990;335:765–74. doi:10.1016/ 0140-6736(90)90878-9. Boshuizen HC, Lanti M, Menotti A, et al. Effects of past and recent blood pressure and cholesterol level on coronary heart disease and stroke mortality, accounting for measurement error. Am J Epidemiol. 2007;165:398–409. doi:10.1093/aje/kwk021. Husten CG, Shelton DM, Chrismon JH, et al. Cigarette smoking and smoking cessation among older adults: United States, 1964– 94. Tob Control. 1997;6:175–80. Junge B, Nagel M. Smoking behavior in Germany. Gesundheitswesen. 1999;61:S121–5.
123
