Background
The extent to which moderate overweight (body mass index [BMI], 25.0-29.9 [calculated as weight in kilograms divided by height in meters squared]) and obesity (BMI, ≥ 30.0) are associated with increased risk of coronary heart disease (CHD) through adverse effects on blood pressure and cholesterol levels is unclear, as is the risk of CHD that remains after these mediating effects are considered.
Methods
Relative risks (RRs) of CHD associated with moderate overweight and obesity with and without adjustment for blood pressure and cholesterol concentrations were calculated by the members of a collaboration of prospective cohort studies of healthy, mainly white persons and pooled by means of random-effects models (RRs for categories of BMI in 14 cohorts and for continuous BMI in 21 cohorts; total N = 302 296).
Results
A total of 18 000 CHD events occurred during follow-up. The age-, sex-, physical activity–, and smoking-adjusted RRs (95% confidence intervals) for moderate overweight and obesity compared with normal weight were 1.32 (1.24-1.40) and 1.81 (1.56-2.10), respectively. Additional adjustment for blood pressure and cholesterol levels reduced the RR to 1.17 (1.11-1.23) for moderate overweight and to 1.49 (1.32-1.67) for obesity. The RR associated with a 5-unit BMI increment was 1.29 (1.22-1.35) before and 1.16 (1.11-1.21) after adjustment for blood pressure and cholesterol levels.
Conclusions
Adverse effects of overweight on blood pressure and cholesterol levels could account for about 45% of the increased risk of CHD. Even for moderate overweight, there is a significant increased risk of CHD independent of these traditional risk factors, although confounding (eg, by dietary factors) cannot be completely ruled out.
Moderate overweight (body mass index [BMI], 25.0-29.9 [calculated as weight in kilograms divided by height in meters squared]) and obesity (BMI, ≥30.0) (both henceforth called “overweight”) are highly prevalent in Western populations. Nearly two-thirds of US adults1 and 60% of Australians2 are overweight, and increasing trends are apparent throughout the world.3 Obesity is clearly associated with increased mortality4,5 and adverse health outcomes, including coronary heart disease (CHD).6
Because of the high prevalence of overweight and the expected future increases, it is essential to gain precise insight into the consequences of overweight for health and into the metabolic pathways that link the two. This study investigated the relationship between overweight and CHD and, specifically, the extent to which this relationship is mediated by adverse effects of overweight on blood pressure and cholesterol levels. This research is relevant to clinical practice because it provides an indication about the excess risk of CHD in overweight people that would persist after optimal treatment for hypertension and hypercholesterolemia was established. Furthermore, it addresses the question whether to incorporate overweight as an additional modifiable risk factor in commonly used risk stratification schemes such as Adult Treatment Panel III or Framingham.7,8 According to Adult Treatment Panel III, the independent component of risk not mediated through the major risk factors has not been quantified. Recently, cohort investigations have demonstrated that overweight is related to CHD apart from its association with traditional risk factors such as blood pressure and cholesterol levels.9,10 Another recent publication showed that the association between overweight and death from atherosclerotic cardiovascular causes was attenuated to statistically nonsignificant levels after adjustment for blood pressure, cholesterol level, and blood glucose level.11 An analysis based on many prospective cohort studies would add to the available evidence.
The present report describes a meta-analysis of the associations between overweight and risk of CHD for 302 296 healthy persons, mainly white. We report pooled estimates of relative risk (RR) adjusted in a standardized way from 21 prospective cohort studies that participated in a worldwide collaboration. The main outcome measure of the analysis was the age-, sex-, physical activity–, and smoking-adjusted RR of CHD, with and without adjustment for blood pressure and cholesterol levels.
Studies were identified by regularly checking the PubMed and MEDLINE databases, by examination of the reference lists of identified articles, and via suggestions by colleagues. An additional literature search was performed in MEDLINE (1996-2005) by using the following search strategy: obesity, body mass index, BMI, or overweight in either the title or in the Medical Subject Heading (MeSH) and either coronary heart disease in the title or coronary disease in MeSH, plus either prospective or cohort.
Eligible studies were prospective cohort studies conducted in healthy populations that consisted mainly of white persons for whom RRs of BMI or overweight for total incidence or mortality from CHD had been reported. We identified 70 studies that met our inclusion criteria; we were able to contact 62 investigators, and 31 of them agreed to collaborate (44% of the 70 eligible cohorts [an appendix listing the studies that were not included is available at the authors’ Web site: http://www.rivm.nl/bibliotheek/digitaaldepot/appendiximd06.htm]). For the present report, it was necessary that RRs from the cohort be available with multiple adjustments for age, sex, physical activity, and smoking, both with and without simultaneous adjustment for blood pressure and cholesterol levels. Of the complete collaboration, 21 cohorts fulfilled this criterion.
We requested that investigators from the participating cohort studies calculate RRs and 95% confidence intervals (CIs) with systematic univariate and multiple adjustments for age, sex, physical activity, smoking, blood pressure, and cholesterol levels. To minimize the amount of work and maximize participation, investigators could calculate the RRs of CHD in a way similar to that used in the original articles, eg, for the same BMI categories and follow-up time.
An appendix (available at the authors’ Web site) presents the methods that were used in the original studies to define smoking habits, physical activity, blood pressure, and blood cholesterol level, and whether BMI was analyzed as a continuous variable or in categories. One cohort (Nurses' Health Study) used BMI based on self-reported weight and height, instead of measurements.
Adjustment for smoking was generally conducted by inclusion of dummy variables to indicate never-smokers, ex-smokers, and current smokers in the multiple regression model. For blood pressure, the majority of studies (n = 15) used systolic pressure, and for blood cholesterol concentrations, total cholesterol (n = 19). Other measures were, for instance, diastolic blood pressure and elevated total cholesterol concentrations (yes or no). Physical activity was predominantly defined by means of various categories of intensity, but there was considerable diversity between studies. Descriptive statistics for each cohort (eg, mean age, follow-up time, mean BMI in each category, number of persons, and cases of CHD per category) and the definition of the variables were checked by the original investigators.
Dummy variables indicated whether adjustments were made for smoking, physical activity, blood pressure, and cholesterol levels. The RRs were plotted to visualize variation in results between studies. Relative risks with equivalent adjustments were pooled by means of a random effects model12 and were calculated for the categories moderate overweight (BMI, 25.0-29.9) and obesity (BMI, ≥ 30.0), as compared with the reference category (“normal” weight; BMI, 18.5-24.9). Studies were selected in which the foregoing definition of the overweight categories was used. In this selection, the lower limit of the normal weight category varied somewhat between cohorts (see the appendix at the author’s Web site). Therefore, the meta-analysis was also performed in subsets with equal lower limits. These analyses showed that the percentage decrease in RR after adjustment for blood pressure and cholesterol levels was similar (data not shown). The final number of cohorts for the analyses of the categories moderate overweight and obesity was 14.
We also calculated risk of CHD by using BMI as a continuous variable, eg, risk per 5-unit increase in BMI. In this case, if individual studies had provided only RRs for categories of BMI, we transformed the independent variable to its continuous form for each set of adjustments by applying the method of Greenland and Longnecker,13 but using number of cases as observed rather than their fitted values.14 Consequently, more cohorts were available for these analyses (n = 21) than for the analyses of categories of BMI.
Statistical significance of the change in RR after adjusting for blood pressure and cholesterol level was assessed by means of meta-regression analysis (in which results stemming from a single study shared the same random effect). The analyses were repeated for cohorts with measured BMI (instead of BMI based on self-reported weight and height of the participants) and cohorts in which measures for blood pressure and cholesterol concentrations were systolic blood pressure and total cholesterol concentrations (instead of, eg, diastolic blood pressure). Heterogeneity of RRs between studies was examined by χ2 tests. All analyses were performed with the MIXED procedure in SAS statistical software, version 9 (SAS Institute Inc, Cary, NC).15-36
Characteristics of cohorts
Table 1 presents characteristics of the study populations, which included a total of 302 296 persons. Most studies used either mortality from CHD or incidence of CHD (both fatal and nonfatal events) as their end point. A total of 18 000CHD events were observed during follow-up. We were able to extend follow-up for some studies beyond that reported in the original articles. Table 1 presents the data as used in the present analysis.
RRs FOR MODERATE OVERWEIGHT AND OBESITY
Figure 1 and Figure 2 present the RRs of CHD for the separate cohorts for categories of moderate overweight and obesity, adjusted for age, sex, physical activity, and smoking, with and without adjustment for blood pressure and cholesterol levels. In all individual studies, the RR decreased after adjustment for blood pressure and cholesterol concentrations.
Table 2 presents pooled RRs of CHD for categories of BMI. After adjustment for age, sex, physical activity, and smoking, moderate overweight was associated with an RR of 1.32 (95% CI, 1.24-1.40) and obesity with an RR of 1.81 (95% CI, 1.56-2.10). Additional adjustment for blood pressure and cholesterol levels statistically significantly reduced the RR to 1.17 (95% CI, 1.11-1.23) for moderate overweight and 1.49 (95% CI, 1.32-1.67) for obesity. This corresponds to a decrease in excess risk of 47% for moderate overweight and 40% for obesity, ie, [(1.81 − 1.49)/(1.81 − 1)] × 100 = 40%.
The RRs were similar for studies in which BMI was measured (n = 13) instead of based on self-reported weight and height, and for studies in which BMI was measured and adjustments were made for systolic blood pressure and total cholesterol levels (n = 11; instead of other indicators such as diastolic blood pressure [the appendix at the authors’ Web site] shows which measures of blood pressure and cholesterol levels were used]). In the latter subset of studies, the risk of CHD decreased statistically significantly by 50% for moderate overweight and 43% for obesity after additional adjustment for blood pressure and cholesterol levels.
Rr associated with a 5-unit increase in bmi
When BMI was analyzed as a continuous variable, the age-, sex-, physical activity–, and smoking-adjusted RR associated with a 5-unit increase was 1.29 (95% CI, 1.22-1.35; n = 21; Table 3). After excluding participants with a BMI less than 20, this RR was similar (n = 6 studies, not shown). The range between studies was 0.95 to 1.73. Additional adjustment for blood pressure and cholesterol level lowered the excess risk by 45% to 1.16 (95% CI, 1.11-1.21; range, 0.83-1.87). Significant heterogeneity existed between studies both with and without adjustment for blood pressure and cholesterol concentrations (P<.001).
In this large meta-analysis, involving 302 296 participants worldwide and 18 000 CHD events during follow-up, a 5-unit increment in BMI was associated with a 29% increase in risk of CHD and, after additional adjustment for blood pressure and cholesterol levels, with a 16% increased risk. Hence, the present study indicates that adverse effects of overweight on blood pressure and cholesterol levels could account for about 45% of the increased risk of CHD, and that there is still a significantly increased risk of CHD that is independent of these effects.
The strength of our analysis lies in the large number of cohorts and the systematic adjustments for relevant variables. It is clearly shown that adjusting for blood pressure and cholesterol decreases the estimated RR of CHD substantially. This was the case in all studies, despite the observed heterogeneity in RRs. Meta-regression analysis suggests that causes of heterogeneity are the age of the population and the follow-up time, but not, for example, the end point that was used (fatal and nonfatal incidence vs mortality of CHD) (data not shown). We did not exclude the first years of follow-up in our analysis to account for undiagnosed preexisting disease that may cause weight loss and death, leading to a J-shaped BMI-mortality curve. Previous research showed that this effect of excluding early deaths is only marginal.37 In general, publication bias, ie, less frequent publication of studies with absent or negative associations between BMI and CHD, could have increased the apparent RR. An indication for the absence of publication bias (as suggested by a funnel plot; not shown) is that studies with higher estimates of RR were not overrepresented among the studies with low precision (ie, the smaller studies). Furthermore, the decision to participate in the collaboration seemed to depend merely on practical issues, such as time needed to conduct the analyses, and not on the actual results of a study. The majority of the eligible studies not included in our analysis also reported positive associations between moderate overweight and obesity and the risk of CHD, although the results were not always statistically significant. (An overview of these results from the literature is available at the authors’ Web site).
The present study has 2 important implications. First, even moderate overweight is associated with increased risk of CHD (for obesity this was already no point of debate). Because high blood pressure and cholesterol levels are plausible intermediary factors in the causal pathways linking overweight and CHD,6 adjusting for them—in epidemiological analyses—certainly results in underestimating the total public health impact of overweight. Second, the fact that the RR of CHD remained statistically significant after adjustment for these intermediary factors adds to the evidence that overweight itself increases CHD risk independent of traditional risk factors. This implies that, even under the theoretical scenario that optimal treatment would be available against hypertension and hypercholesterolemia in overweight persons, they still would have an elevated risk of CHD. It also implies that overweight, which is easily measured, may be considered to be incorporated as an additional risk factor in commonly used risk stratification schemes such as Adult Treatment Panel III7 and the Framingham CHD prediction algorithm,8 even though the exact mechanism that underlies an “independent” effect remains to be resolved.
The present estimate of the RR adjusted for blood pressure and cholesterol level, ie, 1.16 per 5 BMI units, is similar to the recently reported RR of hospitalization for CHD of 1.16 per 4 BMI units after adjustment for smoking, systolic blood pressure, and total cholesterol level.9 Because hypertension is correlated with other features of the metabolic syndrome, such as fasting serum glucose level,11,38 part of the reduction in the RR after adjustment for blood pressure may be caused by adjustment for these correlated variables, resulting in an overestimate of the excess risk of CHD mediated by blood pressure and cholesterol concentrations. On the other hand, the use of a single measurement of blood pressure and cholesterol, as opposed to repeated measurements, may have underestimated the effect of adjustment for these variables.39
Several mechanisms could underlie an effect of overweight on CHD independent of traditional risk factors. These include a state of low-grade inflammation, endothelial dysfunction, hemostatic imbalance favoring coagulation, impaired endothelial vasodilatory responses, left ventricularhypertrophy due to an increased blood volume, and reduced heart rate variability due to withdrawal of vagal activity and sympathetic predominance.40 Obviously, overweight is associated with increased risk of type 2 diabetes mellitus.41 In Adult Treatment Panel III, obesity is not listed as a risk factor because it is said to operate through diabetes (and other risk factors).7 Therefore, inclusion of data on diabetes or glucose intolerance in our analysis (which were not available for the meta-analysis) would have further attenuated the RR of CHD associated with overweight. Indeed, a large Korean cohort study11 showed that the RR of death from atherosclerotic cardiovascular causes decreased considerably after adjustment for blood pressure, cholesterol level, and fasting blood glucose level. Interestingly, the authors stated that, in addition to these factors, other consequences of increased BMI are likely to contribute to the risk of cardiovascular disease.
Before invoking the plausible pathways mentioned previously, some alternative explanations for our findings must be mentioned. There may have been confounders for which we were unable to adjust, and that themselves, rather than overweight, determine the risk of CHD. For instance, we did not control for diet, which has been shown to be related to CHD,42 because detailed dietary data were usually not available. However, in the Nurses' Health Study—the largest study included—adjustment for diet had virtually no impact on the association between BMI and risk of CHD.29 Also, the possibility of residual confounding due to inaccurate assessment of smoking and physical activity cannot be excluded. Despite diversity in the assessment of physical activity, most studies used measures that discerned between low and at least moderate levels of activity and, in most studies, additional adjustment for physical activity did not substantially change the age-, sex-, and smoking-adjusted RRs substantially (data not shown), which might indicate that measurement error did not differ much between studies.
With respect to the measure of body fatness used, waist-to-hip ratio has been shown to be more strongly related to CHD.43,44 The INTERHEART study demonstrated a graded and highly significant association between myocardial infarction and waist-to-hip ratio, in contrast to BMI.43 However, that study had a cross-sectional case-control design, and the long-term influence of factors such as overweight in midlife therefore could not be assessed. Furthermore, BMI is by far the most common measure of body fatness, especially in cohort studies that started decades ago, and therefore is most suitable for a meta-analysis. The present findings apply to healthy persons, predominantly white. A meta-analysis of cohorts in the Asia-Pacific region showed that risk of cardiovascular disease and CHD increased linearly from BMI levels as low as 18.45 It is known that in Asians the risk of CHD is increased at lower levels of BMI than in whites46 and hence confirmation in primarily white populations was warranted. Our results are similar to those of a previous systematic review47 that included mainly white cohorts, which found a RR of 1.14 per 2 BMI units. However, no formal meta-analysis was attempted in that study, and there was more variation between cohorts in the covariates for which the RRs were adjusted.
We conclude that moderate overweight and obesity are associated with a significant increase in risk of CHD, and thus that the worldwide increase in (moderate) overweight may drive the incidence of CHD upward. Although effects of confounders, such as specific dietary factors, cannot be completely ruled out, negative effects will be exerted both through adverse influences on blood pressure and cholesterol levels (accounting for approximately 45% of the increased risk) and through other pathways.
Correspondence: Rik P. Bogers, PhD, Centre for Prevention and Health Services Research, National Institute for Public Health and the Environment, PO Box 1, 3720 BA Bilthoven, the Netherlands (rik.bogers@rivm.nl).
Accepted for Publication: March 30, 2007.
Author Contributions: Dr Bogers had full access to all of the data in the study (that were provided to him by the collaborating investigators) and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Bogers and Bemelmans. Acquisition of data: Bogers, Woodward, and Menotti. Analysis and interpretation of data: Bogers, Bemelmans, and Boshuizen. Drafting of the manuscript: Bogers and Bemelmans. Critical revision of the manuscript for important intellectual content: Bogers, Bemelmans, Hoogenveen, Boshuizen, Woodward, Knekt, van Dam, Hu, Visscher, Menotti, Thorpe, Jamrozik, Calling, Strand, and Shipley. Study supervision: Bemelmans. All other investigators of the collaboration contributed by collecting data and calculating RRs.
Financial Disclosure: None reported.
Additional Contributions: Federica Barzi, PhD, helped with data extraction and analysis of some of the Australian and New Zealand studies.
Box Section Ref IDBMI-CHD Collaboration Investigators and Study Sites
Australian National Heart Foundation Risk Factor Prevalence Study: Tim Welborn, AO, PhD Department of Medicine, University of Western Australia, Crawley.
Caerphilly Cohort Study:
John W. G. Yarnell, MD, and Shicheng Yu, MD, Department of Epidemiology and Public Health, Queen's University of Belfast, Institute of Clinical Science, Royal Victoria Hospital, Belfast, Northern Ireland; Medical Research Council Collaborative Group, Department of Social Medicine, University of Bristol, Bristol, England.
Dubbo Study of Australian Elderly:
Leon A. Simons, MD, FRACP, University of New South Wales Lipid Research Department, St Vincent's Hospital, Sydney.
Finnish Mobile Clinic Health Examination Survey:
Paul Knekt, PhD, Department of Health and Functional Capacity, National Public Health Institute, Helsinki.
Fletcher Challenge:
Stephen MacMahon, PhD, Robyn Norton, PhD, and Mark Woodward, PhD, The George Institute, University of Sydney, Sydney, Australia; Rod Jackson, PhD, University of Auckland, Auckland, New Zealand.
Kuopio Ischaemic Heart Disease Risk Factor Study:
Haana-Maaria Lakka, MD, PhD, Department of Public Health and Clinical Nutrition, University of Kuopio, Kuopio, Finland.
Malmö Preventive Project:
Susanna Calling, MD, PhD, and Bo Hedblad, MD, PhD, Department of Clinical Sciences in Malmö, Epidemiological Research Group, Lund University, Malmö University Hospital, Malmö, Sweden.
Melbourne Collaborative Cohort Study:
Graham G. Giles, PhD, Cancer Epidemiology Centre, Cancer Council Victoria, Melbourne, Australia.
Multifactor Primary Prevention Study Göteborg:
Annika Rosengren, MD, Department of Medicine, Sahlgrenska University Hospital/Östra, Göteborg, Sweden.
National Health and Nutrition Examination Survey I Epidemiologic Follow-up Study:
Roland J. Thorpe Jr, PhD, The Johns Hopkins Medical Institutions, Baltimore, Maryland.
Nijmegen Cohort Study:
J. Carel Bakx, MD, PhD, Department of General Practice, University Medical Centre St Radboud Nijmegen, Nijmegen, the Netherlands.
Norwegian Counties Study:
Bjørn Heine Strand, PhD, Division of Epidemiology, Norwegian Institute of Public Health, Oslo.
Nurses' Health Study:
Frank B. Hu, MD, PhD, and Rob M. van Dam, PhD, Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts.
PRIME (Prospective Epidemiological Study of Myocardial Infarction) Study:
Pierre Ducimetière, PhD, Institut National de la Sant et de la Recherche Médicale (INSERM), Unit 258: Cardiovascular and Metabolic Epidemiology, Villejuif, France; Phillipe Amouyel, MD, PhD, INSERM, Unit 508, Institut Pasteur de Lille, Lille, France; Dominique Arveiler, MD, Laboratoire d’Epidémiologie et de Santé Publique, Strasbourg, France; Alun Evans, MD, FRCP, Department of Epidemiology and Public Health, Queen’s University Belfast, Belfast; and Jean Ferrières, MD, MSc, INSERM, Unit 558, Faculté de Médicine Purpan, Toulouse, France.
Scottish Heart Health Study:
Hugh Tunstall-Pedoe, MD, FRCP, FFPH, and Mark Woodward, PhD, Cardiovascular Epidemiology Unit, Ninewells Hospital and Medical School, Dundee.
US Railroad Cohort, Rome Railroad Cohort, and Italian Rural Areas of the Seven Countries Study:
Alessandro Menotti, MD, PhD, Associazione per la Ricerca Cardiologica, Rome, Italy.
Ventimiglia di Sicilia Heart Study:
Carlo M. Barbagallo, MD, PhD, Dipartimento di Medicina Clinica e Patologie Emergenti, Università degli Studi di Palermo, Palermo, Italy.
Whitehall Study:
Martin J. Shipley, MSc, Department of Epidemiology and Public Health, University College London, London, England.
Zutphen Elderly Study:
Daan Kromhout, PhD, Division of Human Nutrition, Wageningen University, Wageningen, the Netherlands; Ivon E. J. Milder, PhD, Centre for Prevention and Health Services Research, National Institute for Public Health and the Environment, Bilthoven, the Netherlands.
1.Flegal
KMCarroll
MDOgden
CLJohnson
CL Prevalence and trends in obesity among US adults, 1999-2000.
JAMA 2002;288
(14)
1723- 1727
PubMedGoogle ScholarCrossref 2.Cameron
AJWelborn
TAZimmet
PZ
et al. Overweight and obesity in Australia: the 1999-2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab) [published correction appears in
Med J Aust. 2004;180(8):418].
Med J Aust 2003;178
(9)
427- 432
PubMedGoogle Scholar 3.York
DARossner
SCaterson
I
et al. American Heart Association, Prevention Conference VII: obesity, a worldwide epidemic related to heart disease and stroke, group I: worldwide demographics of obesity.
Circulation 2004;110
(18)
e463- e470
PubMedGoogle ScholarCrossref 4.Mokdad
AHMarks
JSStroup
DFGerberding
JL Actual causes of death in the United States, 2000 [published correction appears in
JAMA. 2005;293(3):293-234, 298].
JAMA 2004;291
(10)
1238- 1245
PubMedGoogle ScholarCrossref 5.Flegal
KMGraubard
BIWilliamson
DFGail
MH Excess deaths associated with underweight, overweight, and obesity.
JAMA 2005;293
(15)
1861- 1867
PubMedGoogle ScholarCrossref 6.National Task Force on the Prevention and Treatment of Obesity, Overweight, obesity, and health risk.
Arch Intern Med 2000;160
(7)
898- 904
PubMedGoogle ScholarCrossref 7.National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III), 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) final report.
Circulation 2002;106
(25)
3143- 3421
PubMedGoogle Scholar 8.Wilson
PWD’Agostino
RBLevy
DBelanger
AMSilbershatz
HKannel
WB Prediction of coronary heart disease using risk factor categories.
Circulation 1998;97
(18)
1837- 1847
PubMedGoogle ScholarCrossref 9.Yan
LLDaviglus
MLLiu
K
et al. Midlife body mass index and hospitalization and mortality in older age.
JAMA 2006;295
(2)
190- 198
PubMedGoogle ScholarCrossref 10.Mann
DMLee
JLiao
YNatarajan
S Independent effect and population impact of obesity on fatal coronary heart disease in adults.
Prev Med 2006;42
(1)
66- 72
PubMedGoogle ScholarCrossref 12.van Houwelingen
HCArends
LRStijnen
T Advanced methods in meta-analysis: multivariate approach and meta-regression.
Stat Med 2002;21
(4)
589- 624
PubMedGoogle ScholarCrossref 13.Greenland
SLongnecker
MP Methods for trend estimation from summarized dose-response data, with applications to meta-analysis.
Am J Epidemiol 1992;135
(11)
1301- 1309
PubMedGoogle Scholar 14.Hartemink
NBoshuizen
HCNagelkerke
NJJacobs
MAvan Houwelingen
HC Combining risk estimates from observational studies with different exposure cutpoints: a meta-analysis on body mass index and diabetes type 2.
Am J Epidemiol 2006;163
(11)
1042- 1052
PubMedGoogle ScholarCrossref 15.Welborn
TADhaliwal
SSBennett
SA Waist-hip ratio is the dominant risk factor predicting cardiovascular death in Australia.
Med J Aust 2003;179
(11-12)
580- 585
PubMedGoogle Scholar 16.Yarnell
JWPatterson
CCSweetnam
PMLowe
GD Haemostatic/inflammatory markers predict 10-year risk of IHD at least as well as lipids: the Caerphilly collaborative studies.
Eur Heart J 2004;25
(12)
1049- 1056
PubMedGoogle ScholarCrossref 17.Yu
SYarnell
JWSweetnam
PMMurray
L What level of physical activity protects against premature cardiovascular death? the Caerphilly study.
Heart 2003;89
(5)
502- 506
PubMedGoogle ScholarCrossref 18.Simons
LAFriedlander
YMcCallum
JSimons
J Risk factors for coronary heart disease in the prospective Dubbo Study of Australian elderly.
Atherosclerosis 1995;117
(1)
107- 118
PubMedGoogle ScholarCrossref 19.Knekt
PReunanen
AJärvinen
RSeppänen
RHeliövaara
MAromaa
A Antioxidant vitamin intake and coronary mortality in a longitudinal population study.
Am J Epidemiol 1994;139
(12)
1180- 1189
PubMedGoogle Scholar 20.MacMahon
SNorton
RJackson
R
et al. Fletcher Challenge–University of Auckland Heart & Health Study: design and baseline findings.
N Z Med J 1995;108
(1013)
499- 502
PubMedGoogle Scholar 22.Lakka
HMLakka
TATuomilehto
JSalonen
JT Abdominal obesity is associated with increased risk of acute coronary events in men.
Eur Heart J 2002;23
(9)
706- 713
PubMedGoogle ScholarCrossref 23.Jonsson
SHedblad
BEngstrom
GNilsson
PBerglund
GJanzon
L Influence of obesity on cardiovascular risk: twenty-three-year follow-up of 22,025 men from an urban Swedish population.
Int J Obes Relat Metab Disord 2002;26
(8)
1046- 1053
PubMedGoogle ScholarCrossref 25.Rosengren
AWedel
HWilhelmsen
L Body weight and weight gain during adult life in men in relation to coronary heart disease and mortality: a prospective population study.
Eur Heart J 1999;20
(4)
269- 277
PubMedGoogle ScholarCrossref 26.Thorpe
RJ
JrFerraro
KF Aging, obesity, and mortality: misplaced concern about obese older people?
Res Aging 2004;26
(1)
108- 129
Google ScholarCrossref 27.Bakx
JCVeldstra
MIvan den Hoogen
HM
et al. Blood pressure and cardiovascular morbidity and mortality in a Dutch population: the Nijmegen cohort study.
Prev Med 2001;32
(2)
142- 147
PubMedGoogle ScholarCrossref 28.Strand
BHTverdal
A Can cardiovascular risk factors and lifestyle explain the educational inequalities in mortality from ischaemic heart disease and from other heart diseases? 26 year follow up of 50,000 Norwegian men and women.
J Epidemiol Community Health 2004;58
(8)
705- 709
PubMedGoogle ScholarCrossref 29.Li
TYRana
JSManson
JE
et al. Obesity as compared with physical activity in predicting risk of coronary heart disease in women.
Circulation 2006;113
(4)
499- 506
PubMedGoogle ScholarCrossref 30.Mahamat
ARichard
FArveiler
D
et al. Body mass index, hypertension and 5-year coronary heart disease incidence in middle aged men: the PRIME study.
J Hypertens 2003;21
(3)
519- 524
PubMedGoogle ScholarCrossref 31.Menotti
ASeccareccia
FBlackburn
HKeys
A Coronary mortality and its prediction in samples of US and Italian railroad employees in 25 years within the Seven Countries Study of cardiovascular diseases.
Int J Epidemiol 1995;24
(3)
515- 521
PubMedGoogle ScholarCrossref 32.Tunstall-Pedoe
HWoodward
MTavendale
RA’Brook
RMcCluskey
MK Comparison of the prediction by 27 different factors of coronary heart disease and death in men and women of the Scottish Heart Health Study: cohort study [published correction appears in
BMJ. 1998;316(7148):1881].
BMJ 1997;315
(7110)
722- 729
PubMedGoogle ScholarCrossref 33.Barbagallo
CMCavera
GSapienza
M
et al. Prevalence of overweight and obesity in a rural southern Italy population and relationships with total and cardiovascular mortality: the Ventimiglia di Sicilia project.
Int J Obes Relat Metab Disord 2001;25
(2)
185- 190
PubMedGoogle ScholarCrossref 35.Batty
GDShipley
MJJarrett
RJBreeze
EMarmot
MGDavey Smith
G Obesity and overweight in relation to disease-specific mortality in men with and without existing coronary heart disease in London: the original Whitehall study.
Heart 2006;92
(7)
886- 892
PubMedGoogle ScholarCrossref 36.Buijsse
BFeskens
EJKok
FJKromhout
D Cocoa intake, blood pressure, and cardiovascular mortality: the Zutphen Elderly Study.
Arch Intern Med 2006;166
(4)
411- 417
PubMedGoogle Scholar 37.Allison
DBFaith
MSHeo
MTownsend-Butterworth
DWilliamson
DF Meta-analysis of the effect of excluding early deaths on the estimated relationship between body mass index and mortality.
Obes Res 1999;7
(4)
342- 354
PubMedGoogle ScholarCrossref 38.Caballero
AE Endothelial dysfunction in obesity and insulin resistance: a road to diabetes and heart disease.
Obes Res 2003;11
(11)
1278- 1289
PubMedGoogle ScholarCrossref 39.Rosner
BSpiegelman
DWillett
WC Correction of logistic regression relative risk estimates and confidence intervals for random within-person measurement error.
Am J Epidemiol 1992;136
(11)
1400- 1413
PubMedGoogle Scholar 40.Saltzman
EBenotti
PN Effects of obesity on the cardiovascular system. Bray
GABouchard
C
Handbook of Obesity: Etiology and Pathophysiology. 2nd ed. New York, NY Marcel Dekker Ltd2003;825- 843
Google Scholar 41.Field
AECoakley
EHMust
A
et al. Impact of overweight on the risk of developing common chronic diseases during a 10-year period.
Arch Intern Med 2001;161
(13)
1581- 1586
PubMedGoogle ScholarCrossref 43.Yusuf
SHawken
SOunpuu
S
et al. INTERHEART Study Investigators, Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study.
Lancet 2005;366
(9497)
1640- 1649
PubMedGoogle ScholarCrossref 44.Silventoinen
KJousilahti
PVartiainen
ETuomilehto
J Appropriateness of anthropometric obesity indicators in assessment of coronary heart disease risk among Finnish men and women.
Scand J Public Health 2003;31
(4)
283- 290
PubMedGoogle ScholarCrossref 45.Ni Mhurchu
CRodgers
APan
WHGu
DFWoodward
MAsia Pacific Cohort Studies Collaboration, Body mass index and cardiovascular disease in the Asia-Pacific region: an overview of 33 cohorts involving 310 000 participants.
Int J Epidemiol 2004;33
(4)
751- 758
PubMedGoogle ScholarCrossref 46.Jee
SHPastor-Barriuso
RAppel
LJSuh
IMiller
ER
IIIGuallar
E Body mass index and incident ischemic heart disease in South Korean men and women.
Am J Epidemiol 2005;162
(1)
42- 48
PubMedGoogle ScholarCrossref 47.Whitlock
GLewington
SMhurchu
CN Coronary heart disease and body mass index: a systematic review of the evidence from larger prospective cohort studies.
Semin Vasc Med 2002;2
(4)
369- 381
PubMedGoogle ScholarCrossref