Why should population attributable fractions be periodically recalculated?: An example from cardiovascular risk estimation in southern Europe

doi:10.1016/j.ypmed.2010.03.012

Preventive Medicine

Volume 51, Issue 1, July 2010, Pages 78-84

https://doi.org/10.1016/j.ypmed.2010.03.012 Get rights and content

Abstract

Objective

To determine the effect of age and study period on coronary heart disease (CHD) risk attributable to cardiovascular risk factors.

Methods

A cohort of cardiovascular disease (CVD)-free randomly participants from Girona (Spain) aged 35–74 years recruited in 1995 and 2000 and followed for an average of 6.9 years. A survey conducted in the same area in 2005 was also used for the analysis. Smoking, hypertension, diabetes, sedentary lifestyle, obesity, total cholesterol ≥ 240 mg/dl, low-density lipoprotein (LDL) cholesterol ≥ 160 mg/dl, and high-density lipoprotein cholesterol < 40 mg/dl were the risk factors considered. The composite end-point included myocardial infarction, angina pectoris, and CHD death.

Results

LDL cholesterol had the highest potential for CHD prevention between 35 and 74 years [42% (95% Confidence Interval: 23,58)]. The age-stratified analysis showed that the population attributable risk (PAF) for smoking was 64% (30,80) in subjects < 55 years; for those ≥ 55 years, the PAF for hypertension was 34% (1,61). The decrease observed between 1995 and 2005 in the population's mean LDL cholesterol level reduced that PAF in all age groups.

Conclusion

Overall, LDL cholesterol levels had the highest potential for CHD prevention. Periodic PAF recalculation in different age groups may be required to adequately monitor population trends.

Introduction

Public health decisions about resource allocation, prevention, and patient care are closely tied to the availability of information on prevalence, incidence, mortality, and case-fatality to address the illnesses with the greatest impact on the population and their determinant risk factors. A key example is our need to understand cardiovascular diseases (CVD), the main cause of death in the developed world, and the associated risk factors (World Health Statistics, 2009).

The population attributable fraction (PAF), the proportion of disease incidence in the population that can be attributed to a risk factor, combines the concepts of incidence (alternatively, relative risk or hazard ratio) and risk factor prevalence (Walter, 1976). The assumptions underlying valid PAF estimation include a causal relationship between risk factors and disease; immediate risk reduction among the exposed, equal to non-exposure, when the risk factor is eliminated from a population; and independence of the considered risk factors from other factors that influence disease occurrence (Northridge, 1995, Rockhill et al., 1998).

PAF magnitude varies with age and region, known to modify risk factor prevalence (Gabriel et al., 2008, Evans et al., 2001, Howard et al., 2009, Yusuf et al., 2004, Menotti et al., 2000). Moreover, secular changes in the prevalence of risk factors also affect PAF values (Evans et al., 2001, Grau et al., 2007). For some risk factors, relative risks or hazard ratios vary by length of follow-up (Menotti et al., 2005, Menotti and Lanti, 2003), which may slightly influence PAF estimates.

Estimates of relative risk and risk factor prevalence are typically obtained from published data on different populations, regions and time periods (Medrano et al., 2007). Accurate PAF estimation requires relative risk and risk factor prevalence data from the same population.

This analysis aimed to determine the effect of age and study period on 10-year coronary heart disease (CHD) risk attributable to cardiovascular risk factors in a population-based study (CHD-PAF).

Section snippets

Design

In northeast Spain, a cohort of participants ages 35–74 was randomly recruited from the Girona census in two surveys (1995 and 2000) used to determine 10-year hazard ratios and CHD-PAF for different risk factors by age groups.

Baseline examinations at recruitment and a 2005 survey in the same region permitted analysis of CHD-PAF modifications over 10 years. Participation in all three surveys was > 72%. Inclusion criteria and recruitment methodology are previously described (Grau et al., 2007).

Results

We followed 1802 men and 1932 women for 25,744 person-years, with 6.9 years mean follow-up. During follow-up, 220 participants (4.7%) were lost, 106 participants suffered an incident CHD event, and 136 died of other causes (Fig. 1). The acute myocardial infarction age-standardized incidence rate was 295 and 94/100,000 men and women, respectively.

The prevalence of baseline hypertension, diabetes, and obesity (per BMI or waist circumference) increased with age, but older participants had lower

Discussion

PAF magnitude depends on the effect size of risk factor exposure and on this risk factor's prevalence. Effect size depends in turn on length of follow-up, age group, prevalence in that age group, period of time and region. Our study showed that age group and study period are important determinants of the magnitude of CHD-PAF for total, LDL, and HDL cholesterol, hypertension, diabetes and smoking. In our region, the risk factors with the highest population impact on CHD were smoking and LDL

Conclusions

Overall, LDL cholesterol levels had the highest potential for CHD prevention over 10 years in a Mediterranean population aged 35–74 years. In age-stratified analyses, PAF estimates were highest for smoking among participants ages 35–54 and for hypertension and low HDL cholesterol among those 55 and older. PAF may require periodic age-stratified recalculation of prevalence to adequately monitor the population trends in world regions.

Conflict of interest statement

The authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by Spain's Ministry of Science and Innovation through the Carlos III Health Institute [Red HERACLES RD06/0009], Health Research Fund [FIS-90/0672, FIS-93/0568, FIS 94/0539, FIS96/0026-01, FIS99/0655, FIS99/0013-01, FIS 99/9342, FIS 2003/HERMES PI20471, CP06/00100 to M.F., and CM08/00141 to M.G.]; Science and Technology Ministries Commission [CICYT-FEDER 1FD97-0626], and Ministry of Education and Science [SAF2003/1240]; the Government of Catalunya through the Catalan

References (45)

R. Gabriel et al.
Prevalence, geographic distribution and geographic variability of major cardiovascular risk factors in Spain. Pooled analysis of data from population-based epidemiological studies: the ERICE Study
Rev. Esp. Cardiol.
(2008)
M. Gil et al.
Analysis of trends in myocardial infarction case-fatality, incidence and mortality rates in Girona, Spain, 1990–1999
Rev. Esp. Cardiol.
(2007)
M. Grau et al.
Risk functions and the primary prevention of cardiovascular disease
Rev. Esp. Cardiol.
(2008)
G. Howard et al.
Advancing the hypothesis that geographic variations in risk factors contribute relatively little to observed geographic variations in heart disease and stroke mortality
Prev. Med.
(2009)
W.B. Kannel et al.
Concept and usefulness of cardiovascular risk profiles
Am. Heart J.
(2004)
J. Marrugat et al.
The estimated incidence and case fatality rate of ischemic and hemorrhagic cerebrovascular disease in 2002 in Catalonia
Rev. Esp. Cardiol.
(2007)
M.J. Medrano et al.
Coronary disease risk attributable to cardiovascular risk factors in the Spanish population
Rev. Esp. Cardiol.
(2007)
A. Reina et al.
Gender differences in management and outcome of patients with acute myocardial infarction
Int. J. Cardiol.
(2007)
A.K. Rowe et al.
Why population attributable fractions can sum to more than one
Am. J. Prev. Med.
(2004)
S. Yusuf et al.
Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study
Lancet
(2004)

S. Baba et al.

Cigarette smoking and risk of coronary heart disease incidence among middle-aged Japanese men and women: the JPHC Study Cohort I

Eur. J. Cardiovasc. Prev. Rehabil.

(2006)

R.P. Bogers et al.

Association of overweight with increased risk of coronary heart disease partly independent of blood pressure and cholesterol levels: a meta-analysis of 21 cohort studies including more than 300 000 persons

Arch. Intern. Med.

(2007)

A.V. Chobanian et al.

Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure

Hypertension

(2003)

M.L. Daviglus et al.

Preventing cardiovascular disease in the 21st century: therapeutic and preventive implications of current evidence

Am. J. Cardiovasc. Drugs

(2006)

R. Elosua et al.

Validation of the Minnesota Leisure Time Physical Activity Questionnaire in Spanish men. The MARATHOM Investigators

Am. J. Epidemiol.

(1994)

R. Elosua et al.

Validation of the Minnesota Leisure Time Physical Activity Questionnaire in Spanish Women. Investigators of the MARATDON Group

Med. Sci. Sports Exerc.

(2000)

J.R. Emberson et al.

Re-assessing the contribution of serum total cholesterol, blood pressure and cigarette smoking to the aetiology of coronary heart disease: impact of regression dilution bias

Eur. Heart J.

(2003)

A. Evans et al.

Trends in coronary risk factors in the WHO MONICA project

Int. J. Epidemiol.

(2001)

Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults

Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III)

JAMA

(2001)

Fourth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice European guidelines on cardiovascular disease prevention in clinical practice: executive summary

Eur. Heart J.

(2007)

M. Grau et al.

Trends in cardiovascular risk factor prevalence (1995–2000–2005) in northeastern Spain

Eur. J. Cardiovasc. Prev. Rehabil.

(2007)

S. Greenland

Attributable fractions: bias from broad definition of exposure

Epidemiology

(2001)

Cited by (23)

Impact of cardiovascular risk factors in the Extremadura population: HERMEX cohort contributions for a preventive strategy
2020, Atencion Primaria
Determinar la fracción atribuible poblacional (FAP) de los factores de riesgo (FR) clásicos para la aparición de la enfermedad cardiovascular en una cohorte poblacional de Extremadura, a fin de recomendar intervenciones preventivas prioritarias.
Diseño: estudio de cohortes.
Muestra poblacional representativa de un área de salud de Extremadura (España).
Dos mil ochocientos treinta y tres individuos, de 25 a 79 años, seleccionados aleatoriamente e incluidos entre 2007 y 2009. Se registraron antecedentes y se midieron parámetros clínicos, siendo seguidos hasta el 31 de diciembre de 2015.
Variables explicativas: edad, sexo, obesidad, tabaquismo, hipertensión arterial (HTA), diabetes mellitus (DM) e hipercolesterolemia.
Primer evento de la variable combinada de infarto de miocardio, angina de pecho, ictus, isquemia de miembros inferiores y muerte cardiovascular. Se calcularon las hazard ratio mediante regresión de Cox, totalmente ajustadas y las FAP mediante la fórmula de Levin.
Se analizaron 2.669 sujetos de la cohorte inicial, al ser excluidos 103 por antecedentes de enfermedad cardiovascular y 61 por pérdidas. El seguimiento fue de 6,9 años (RI: 6,5-7,5). Se documentaron 134 eventos. Tasa de incidencia 7,42/1.000 personas-año. Las hazard ratio ajustadas (IC 95%) fueron: HTA 2,26 (1,40-3,67), hipercolesterolemia 2,23 (1,56-3,18), DM 1,79 (1,24-2,58), tabaquismo 1,72 (1,11-2,69). Las FAP (IC 95%) fueron HTA 31,1 (12,4-48,8), hipercolesterolemia 27,0% (14,8-40,6), tabaquismo 18,8% (3,3-35,0), DM 7,9% (2,6-15,2).
La HTA es el FR con mayor impacto en la salud cardiovascular de la población extremeña, seguido de hipercolesterolemia y tabaquismo, constituyendo objetivos prioritarios para una estrategia preventiva poblacional.
To determine the population attributable fraction (PAF) of the major risk factors (RF) for the occurrence of cardiovascular disease in an Extremadura population cohort and therefore recommend priority preventive measures in health.
Design, Cohort study.
Representative population sample of a health area of Extremadura (Spain)
2833 individuals, from 25 to 79 years old, randomly selected and recruited between 2007 and 2009. Antecedents and clinical parameters were recorded, a follow up until December 31, 2015 were done.
Explanatory variables: Age, sex, obesity, current smoking, arterial hypertension, diabetes mellitus (DM) and hypercholesterolemia.
First event of the combined variable of myocardial infarction, angina pectoris, stroke, peripheral arterial disease and cardiovascular death. Fully adjusted hazard ratios (HR) were calculated by Cox regression. The PAFs were calculated using Levin's formula.
2669 subjects were included, 103 had history of cardiovascular disease and 61 were lost. The follow-up was 6.9 years (IR 6.5-7.5). 134 events were recorded. Incidence rate 7.42/1,000 people-year. Adjusted HR (95% CI) were: hypertension 2.26 (1.40-3.67), hypercholesterolemia 2.23 (1.56-3.18), DM 1.79 (1.24-2.58) and current smoking 1.72 (1.11-2.69). The PAF (95% CI) were: hypertension: 31.1 (12.4-48.8), hypercholesterolemia 27.0% (14.8-40.6), smoking 18.8% (3.3–35.0) and DM 7.9% (2.6-15.2).
Hypertension confers the greatest burden of cardiovascular disease in the population of Extremadura, followed by hypercholesterolemia and smoking. These RF are priority objectives for a population-based preventive strategy.
Hospital Mortality in 415 798 AMI Patients: 4 Years Earlier in the Canary Islands Than in the Rest of Spain
2019, Revista Espanola de Cardiologia
Canarias tiene la mortalidad por diabetes más elevada de España. El objetivo es averiguar si existen diferencias con las restantes comunidades autónomas en la mortalidad hospitalaria por infarto agudo de miocardio (IAM), en los factores asociados con esta mortalidad y la fracción poblacional atribuible a la diabetes.
Estudio descriptivo de los ingresos hospitalarios por IAM en España desde 2007 hasta 2014, registrados en el Conjunto Mínimo Básico de Datos.
Se identificaron 415.798 IAM. Los pacientes canarios (16.317) eran más jóvenes que los del resto de España (63,93 ± 13,56 frente a 68,25 ± 13,94 años; p < 0,001); también el fallecimiento ocurrió 4 años antes en el archipiélago (a los 74,03 ± 11,85 frente a los 78,38 ± 11,10 años; p < 0,001). En esta comunidad alcanzó su prevalencia máxima el tabaquismo (el 44% de los varones y el 23% de las mujeres), que se asoció con un adelanto de 13 años en la edad al IAM. Las islas Canarias tuvieron la mayor mortalidad de pacientes tanto con diabetes (8,7%) como sin ella (7,6%), y también la mayor fracción poblacional de muerte por IAM atribuible a la diabetes (9,4; IC95%, 4,8-13,6). Tras ajustar por tipo de IAM, diabetes, dislipemia, hipertensión, tabaquismo, consumo de cocaína, insuficiencia renal, sexo y edad, Canarias presentó el mayor riesgo de mortalidad respecto a España (OR = 1,25; IC95%, 1,17-1,33; p < 0,001). Fue, además, una de las comunidades autónomas que no mejoró significativamente su riesgo de mortalidad por IAM durante el periodo estudiado.
Canarias presenta una situación muy desfavorable respecto al resto de España en la mortalidad hospitalaria por IAM.
The Canary Islands has the highest mortality from diabetes in Spain. The aim of this study was to determine possible differences in mortality due to acute myocardial infarction (AMI) during hospital admission between this autonomous community and the rest of Spain, as well as the factors associated with this mortality and the population fraction attributable to diabetes.
Cross-sectional study of hospital admissions for AMI in Spain from 2007 to 2014, registered in the Minimum Basic Data Set.
A total of 415 798 AMI were identified. Canary Island patients (16 317) were younger than those living in the rest of Spain (63.93 ± 13.56 vs 68.25 ± 13.94; P < .001) and death occurred 4 years earlier in the archipelago (74.03 ± 11.85 vs 78.38 ± 11.10; P < .001). This autonomous community had the highest prevalence of smoking (44% in men and 23% in women); throughout Spain, AMI occurred 13 years earlier in smokers than in nonsmokers. Patients in the Canary Islands had the highest mortality rates whether they had diabetes (8.7%) or not (7.6%), and they also showed the highest fraction of AMI mortality attributable to diabetes (9.4; 95% CI, 4.8-13.6). After adjustment for type of AMI, diabetes, dyslipidemia, hypertension, smoking, cocaine use, renal failure, sex and age, the Canary Islands showed the highest risk of mortality vs the rest of Spain (OR = 1.25; 95%CI, 1.17-1.33; P < .001) and it was one of the autonomous communities showing no significant improvement in the risk of mortality due to AMI during the study period.
Mortality due to AMI during hospital admission is higher in the Canary Islands than in the rest of Spain.
Full English text available from: www.revespcardiol.org/en
Clinical Profile and Changes of Serum Lipid Levels in Epileptic Patients after Cerebral Infarction
2017, Journal of Stroke and Cerebrovascular Diseases
Citation Excerpt :
Dyslipidemia is a well-known risk factor for developing cardiovascular and cerebrovascular disease (CVD).1,2
Antiepileptic drugs (AEDs) may increase development of dyslipidemia and cerebrovascular disease (CVD). We examined the clinical profile and changes of serum lipid levels after AED monotherapy in patients with poststroke epilepsy (PSE) after cerebral infarction (CI).
Medical records were reviewed in consecutive 2144 CI patients. Monotherapy of valproate, carbamazepine (CBZ), phenytoin (PHT), zonisamide, levetiracetam, or lamotrigine was performed in PSE patients. Serum lipid levels were measured before and at 3 months after AED treatment.
The prevalence of PSE was 7.0% in CI patients. The TOAST etiology disclosed large-artery atherosclerosis in 68 patients (45%), cardioembolism in 63 patients (42%), and undetermined cause in 19 patients (13%). CVD risk profile showed obesity of 18 patients (12%), current smoker of 30 patients (20%), hypertension of 75 patients (50%), diabetes mellitus of 32 patients (21%), dyslipidemia of 15 patients (10%), and atrial fibrillation of 63 patients (42%). CBZ or PHT administration increased serum total cholesterol (TC) and low-density lipoprotein-cholesterol (LDL-C) levels significantly compared to baseline and AED-untreated controls. Those levels were not increased significantly in other AED and control groups. Serum high-density lipoprotein-cholesterol and triglyceride levels did not differ statistically in all groups.
The prevalence of post-CI epilepsy was 7.0%. The pathogenesis contributed to atherothrombosis and cardioembolism. CBZ or PHT administration increased serum TC and LDL-C significantly. Thus, we should pay more attention to serum lipid levels in patients receiving cytochrome P450 (CYP)-induced AEDs, and might considerer switching to non–CYP-induced AEDs in patients with unfavorable serum lipid changes.
Antiepileptic drug use for treatment of epilepsy and dyslipidemia: Systematic review
2015, Epilepsy Research
Citation Excerpt :
While a part of the increase risk may be explained by life-style factors, AED use may in fact be a major contributing factor (Chuang et al., 2012; Elliott et al., 2008; Nilsson et al., 1997). Dyslipidemia is a well-known risk factor for vascular disease (Grau et al., 2010; Karthikeyan et al., 2009). Furthermore, it appears that the use of certain, not all, AEDs can put patients at a higher risk of developing dyslipidemia (Mantel-Teeuwisse et al., 2001).
To characterize the association between commonly used anti-epileptic drugs (AEDs) and plasma lipid levels in patients with epilepsy.
We sought observational studies that reported association between commonly used AEDs and plasma lipid levels in patients. The primary outcome was low-density lipoprotein (LDL) cholesterol. High-density lipoprotein (HDL), total cholesterol and triglyceride were secondary outcomes. The control group included healthy controls, pre-treatment patients or patients treated with other AEDs. We conducted a systematic search of major bibliographic databases and review of reference lists of primary articles and reviews. Primary comparisons of interest were: AED monotherapy vs. no AED use, monotherapy with one AED vs. other AED, and AED polytherapy vs. no AED use.
31 studies in 4126 people were identified. Carbamazepine, phenytoin and valproic acid were the most commonly studied drugs and were also implicated in causing considerable changes in plasma lipid levels in treated patients. There was an increase in LDL and total cholesterol levels with use of these three drugs; however, carbamazepine and phenytoin were also associated with higher levels of HDL. We could not identify one particular AED which was worse than the other in head-to-head comparison. We were unable to identify a particular polytherapy regimen that was worse than others.
We found evidence to suggest that some AEDs may negatively alter lipids levels in patients with epilepsy. Both treating physicians and people with epilepsy need to be vigilant in managing their vascular risk factors to avoid vascular disease.
Differential patterns of effects of age and sex on metabolic syndrome in Taiwan: Implication for the inadequate internal consistency of the current criteria
2014, Diabetes Research and Clinical Practice
Citation Excerpt :
These criteria include central obesity, high blood pressure, high blood sugar, low high-density lipoprotein cholesterol (HDL-C), and high triglycerides. MetS and its components also contribute to CVD-related and all-cause mortality and post a global threat and burden to public health [8,10–15]. Plenty of human studies showed that the prevalence of MetS components differ in age, sex, and ethnicity [15–19].
Current criteria of metabolic syndrome (MetS) give equal weight to each component and apply mostly the same cut-off values to all ages. The contribution of each component to MetS and the effects of age and sex on each component and MetS were explored.
We carried out a survey on residents aged 40–74 years of the northern coastal area of Taiwan. The prevalent rates of MetS in 646 males and 961 females were 32.4% and 27.8%, respectively. Logistic regression analyses were used to assess the main and interactive effects of age and sex. The Cronbach's α coefficient was calculated as the indicator of internal consistency of MetS components.
There were significant age trends for MetS components, except for low HDL-C in both sexes and high fasting triglyceride in males. Logistic regression analyses showed that the effects of age and sex on MetS and its component were all different. The age-sex-specific Cronbach's α coefficients for MetS ranged from 0.43 to 0.61. The age trends of the coefficients in males and females were opposite. The exclusion of some components from the MetS resulted in an increase of the coefficients.
Our results indicate that the internal consistency of MetS was questionable. It seems that the currently defined MetS components of MetS did not formulate a single pathophysiological entity. Given equal weight to each component and used the same cut-off values for the subjects of all age groups in both sexes need to be reconsidered.
Cardiovascular risk factors attributable to obesity and overweight in Switzerland
2012, Nutrition, Metabolism and Cardiovascular Diseases
Citation Excerpt :
Overall, our results indicate that, even when underestimated, PAFs derived from self-reported data can actually be in agreement with the literature; however, care should be taken when using such PAFs. Thus, and as reported previously [14], population-specific, up-to-date PAFs based on measured data are preferred. In some instances, PAFs for overweight alone were higher than for obesity alone.
Obesity increases the risk for cardiovascular risk factors (CVRFs), including hypertension, dyslipidaemia and type 2 diabetes. In this study, we assessed the burden of overweight and obesity on CVRFs in Switzerland, using Swiss-specific population attributable fractions (PAFs).
The number of cases of CVRFs that could have been prevented if the increase in overweight and obesity in Switzerland had been contained was estimated using gender-specific, age- and smoking-adjusted PAFs for overweight and obesity. PAFs were estimated from the Swiss Health Survey 2007 (self-reported) and the CoLaus study (measured) data.
PAFs from self-reported were lower than from measured data. Using measured data, overweight and obesity contributed to 38% of hypertension cases in men (32% in women). In men, overweight had a larger impact than obesity (22.2% and 15.6%, respectively), while the opposite was observed for women (13.6% and 18.1%, respectively). In men, 37% of dyslipidaemia (30% in women) could be attributed to overweight and obesity; overweight had a higher contribution than obesity in both sexes. In men, 57% of type 2 diabetes (62% in women) was attributable to overweight and obesity; obesity had a larger impact than overweight in both sexes. Overall, approximately 27,000 cases of type 2 diabetes, 63,000 cases of high blood pressure and 37,000 cases of dyslipidaemia could have been avoided if overweight and obesity levels were maintained at 1992 levels.
A large proportion of CVRFs is attributable to overweight and/or obesity and could have been prevented by containing the overweight/obesity epidemic.

View all citing articles on Scopus

^☆: See the roster of REGICOR Investigators at: www.regicor.org/regicor_inv

View full text

Why should population attributable fractions be periodically recalculated?: An example from cardiovascular risk estimation in southern Europe☆

Abstract

Objective

Methods

Results

Conclusion

Introduction

Section snippets

Design

Results

Discussion

Conclusions

Conflict of interest statement

Acknowledgments

Rev. Esp. Cardiol.

Rev. Esp. Cardiol.

Rev. Esp. Cardiol.

Prev. Med.

Am. Heart J.

Rev. Esp. Cardiol.

Rev. Esp. Cardiol.

Int. J. Cardiol.

Am. J. Prev. Med.

Lancet

Cigarette smoking and risk of coronary heart disease incidence among middle-aged Japanese men and women: the JPHC Study Cohort I

Eur. J. Cardiovasc. Prev. Rehabil.

Association of overweight with increased risk of coronary heart disease partly independent of blood pressure and cholesterol levels: a meta-analysis of 21 cohort studies including more than 300 000 persons

Arch. Intern. Med.

Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure

Hypertension

Preventing cardiovascular disease in the 21st century: therapeutic and preventive implications of current evidence

Am. J. Cardiovasc. Drugs

Validation of the Minnesota Leisure Time Physical Activity Questionnaire in Spanish men. The MARATHOM Investigators

Am. J. Epidemiol.

Validation of the Minnesota Leisure Time Physical Activity Questionnaire in Spanish Women. Investigators of the MARATDON Group

Med. Sci. Sports Exerc.

Re-assessing the contribution of serum total cholesterol, blood pressure and cigarette smoking to the aetiology of coronary heart disease: impact of regression dilution bias

Eur. Heart J.

Trends in coronary risk factors in the WHO MONICA project

Int. J. Epidemiol.

Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III)

JAMA

Eur. Heart J.

Trends in cardiovascular risk factor prevalence (1995–2000–2005) in northeastern Spain

Eur. J. Cardiovasc. Prev. Rehabil.

Attributable fractions: bias from broad definition of exposure

Epidemiology