Abstract
Background
Dietary pattern analysis is emerging as a practical, effective tool for relating comprehensive dietary intake to risk of cardiovascular disease mortality. However, no studies have applied this technique to a population outside of the developed world.
Methods
We conducted prospective cohort analyses in 11,116 participants enrolled in the Health Effects of Arsenic Study in Araihazar, Bangladesh, measuring deaths attributable to disease of circulatory system, heart disease, and cerebrovascular disease. Participants were enrolled in 2000 and followed up for an average of 6.6 years. Dietary information was obtained through a previously validated food-frequency questionnaire at baseline.
Results
Principal component analysis based on our comprehensive, 39 item FFQ yielded 3 dietary patterns: (i) a “balanced” pattern, comprised of steamed rice, red meat, fish, fruit and vegetables; (ii) an “animal protein” diet, which was more heavily weighted towards eggs, milk, red meat, poultry, bread, and vegetables; and (iii) a “gourd and root vegetable” diet that heavily relied on a variety of gourds, radishes, pumpkin, sweet potato, and spinach. We observed a positive association between increasing adherence to the animal protein diet and risk of death from both disease of the circulatory system and heart disease; the hazard ratio was 1.13 (95% CI, 1.00-1.28, p = 0.05) and 1.17 (95% CI, 0.99-1.38, p = 0.07), respectively, in relation to one standard deviation increase in the factor scores for the animal protein diet pattern, after controlling for age, sex, body mass index, smoking status, and energy intake. The positive association was more significant among ever smokers; the hazard ratios (95% CI) for deaths from disease of the circulatory system and heart disease was 1.17 (1.02-1.34) and 1.20 (1.00-1.45), respectively, in relation to one standard deviation increase in the factor scores for the animal protein diet pattern.
Conclusions
An animal protein-rich diet in rural Bangladesh may increase risk of heart disease mortality, especially among smokers. This emphasizes the need to further explore and address the impact of dietary patterns on cardiovascular disease in populations undergoing epidemiologic transition into a future of higher chronic disease burden.
Keywords: Dietary patterns, cardiovascular disease, Bangladesh, coronary heart disease, epidemiologic transition
INTRODUCTION
Cardiovascular disease (CVD) is a vitally important, yet often overlooked, health burden in developing countries. Compared to the developed world, it contributes less, proportionally, to total mortality (~23% compared to ~49%). However, with the vast majority of the global population within its borders, over 60% of worldwide CVD mortality is attributable to the developing world (1). As these regions undergo further industrialization, a decrease in infectious disease mortality and concurrent increase in chronic disease mortality will only enhance the burden of CVD (2-4). Thus, it has become advantageous to better understand modifiable risk factors that contribute to development of this disease. One such factor is diet, which can influence development of obesity, hypertension, dyslipidemia, diabetes, and, ultimately, CVD.
Rather than analyzing intake of specific foods or nutrients, it has been suggested that evaluation of whole diet patterns may yield a more comprehensive understanding of how this complex factor affects the etiology of disease (5). Compared to the “single nutrient” approach, this analysis more fully accounts for biochemical interactions between nutrients, as well as interrelationships between dietary components that cause difficulty in distinguishing individual food or nutrient effects (6). Multiple studies have examined CVD risk and mortality based on food intake patterns (7-13). However, very few have studied the association in non-Western populations, and none have included populations in low-income or developing countries.
We have identified three major dietary patterns among participants of the Health Effects of Arsenic Longitudinal Study (HEALS) in rural Bangladesh (14). In the present study, we evaluate the association between diet patterns and the risk of deaths due to disease of the circulatory system, heart disease, and cerebrovascular disease.
SUBJECTS AND METHODS
Study participants
The HEALS is a prospective cohort study based in Araihazar, Bangladesh, designed to investigate adverse health outcomes associated with chronic exposure to arsenic in groundwater. An abbreviated explanation of study design is presented here, as full details are available elsewhere (Ahsan, 2006). Using population-based sampling, 11,746 men and women were recruited between 2000 and 2002, under the criteria that all were married (to reduce loss to follow-up), between 18-75 years old, and had resided in the study area for at least 5 years. Study participants underwent baseline clinical assessment, and detailed interviews were conducted with each to characterize lifestyle, demographics, and environmental exposures.
Measurement of food intakes
Trained interviewers administered a 39-item food-frequency questionnaire (FFQ) during baseline assessment of all participants in the HEALS. Details of this questionnaire and its subsequent validation are available elsewhere (15). Briefly, prior to baseline, local nutrition experts were consulted to identify all locally available food, and a sequence of focus groups further refined food items for inclusion. A FFQ pilot study among 120 local individuals not included in the cohort study was used to create the final form questionnaire, ensuring that only common foods were included and deeming any items consumed less than once per month in the previous year as insignificant. US Department of Agriculture Nutrient Database for Standard Reference (abbreviated version) (16) and an Indian food nutrient database (17) were used to convert food intakes to nutrient intake values.
To assess the validity of the FFQ, 200 HEALS participants were randomly selected to maintain 7 day food diaries during 2 different seasons. Analysis of the 189 respondents with complete diaries showed sufficient correlation (0.30 to 0.76) with data previously obtained in the FFQ, including common food items, macronutrients, and many major micronutrients (namely total fat, monounsaturated fat, polyunsaturated fat, saturated fat, protein, carbohydrate, dietary fiber, sodium, potassium, vitamin B-6, vitamin B-12, riboflavin, manganese, thiamine, and iron). The present study included 11,116 participants from the 11,746 cohort members who filled out a FFQ at baseline.
Assessment of cardiovascular mortality
The study endpoint of interest was cardiovascular disease deaths, defined as deaths due to disease of circulatory system (ICD-10 I00-I99), among cohort participants during the period from baseline to March 18, 2009 (end of third follow-up). Details of the methods for the assessment of causes of deaths are described elsewhere (18). Briefly, we adapted a validated verbal autopsy procedure, developed by International Centre for Diarrhea Disease Research, Bangladesh, in collaboration with the WHO, to ascertain the causes of deaths. During the follow-up, upon receipt of a death reported by family or neighbors, a study physician and a trained social worker administered the verbal autopsy form to the next of kin. Medical records from physicians who had treated the deceased were collected. For deaths that occurred in the hospital, information on death certificates and biopsies was ascertained. These data were reviewed monthly by an outcome assessment committee, consisting of physicians and consulting medical specialists blinded to the exposure status. Causes of deaths were coded according to the WHO classification (19) and the International Classification of Diseases, 10th Revision (ICD-10) (20). International Centre for Diarrhea Disease Research, Bangladesh has used this method to ascertain causes of deaths since 1971 (21-22) and documented an overall 95% specificity, with a 85% sensitivity for cancer deaths and up to 85% sensitivity for cardiovascular deaths (23).
Food pattern derivation
Details of the methods were presented elsewhere (14). Briefly, we performed principal component analysis using the PROC FACTOR procedure in SAS to identify dietary patterns. An orthogonal rotation (the varimax option in SAS) was used to derive factors (dietary patterns) uncorrelated to one another for better interpretability. To determine the number of meaningful factors, conventional criteria for principle component analysis including eigenvalue, the scree test, proportion of variance accounted for, and the interpretation criterion were considered (24). For each factor, each food item received a factor loading, which represents the correlation coefficient between the food item and the factor. Three major dietary patterns, namely the “balanced” diet, “animal protein”, and “root vegetable” diet, were identified and named accordingly to the foods that loaded most heavily on the pattern (Appendix 1). The first of these was a “balanced” pattern, comprised of steamed rice, red meat, fish, fruit and vegetables. An “animal protein” diet was more heavily weighted towards eggs, milk, red meat, poultry, bread, and vegetables. The “gourd and root vegetable” diet was heavily reliant on a variety of gourds, radishes, pumpkin, sweet potato and spinach. To indicate a subject’s relative standing for each dietary pattern in the population, a factor score, which is a linear composite of the optimally-weighted food items by factor loadings, was constructed for each dietary pattern. We have previously evaluated association of these diet patterns in relation to blood pressure at baseline (14) and the risk of skin lesions (25).
Statistical Analyses
Descriptive analyses were first conducted to assess relationships of diet patterns with baseline lifestyle and CVD risk factors. We computed person-time from the baseline to the date of death from any cause or the date of third active in-person follow-up visit, whichever came first. Because deaths from other forms of heart disease were often a consequence of previous ischemic heart disease in our study population, we a priori estimated the hazard ratios for the combined category heart disease that included deaths from ischemic heart disease and other types of heart disease (26). Cox proportional hazards regression was used to compute the hazard ratios for deaths from disease of circulatory system, heart disease, and cerebrovascular disease, first in relation to baseline lifestyle and cardiovascular disease risk factors. For main analyses, we estimated hazard ratios in relation to quartiles of factor scores as well as one standard deviation (SD) increase of the factor scores for each diet pattern. The assumption of proportional hazards was examined by testing the cross-product terms between covariate variables and log function of survival time, and p-values for all the terms were > 0.10. We first adjusted for age at baseline, sex and energy intake. A second model was constructed to additionally adjust for a priori defined potential confounders including baseline body mass index (m/kg2) and smoking status (never, past, and current). In addition, we conducted stratified analyses by smoking status, one of the most important risk factors for cardiovascular disease mortality, to assess whether the association between diet patterns and cardiovascular disease mortality was similar within ever and never smokers. All analysis was conducted using SAS 9.2 (SAS Institute, Inc., Cary, North Carolina).
RESULTS
A total of 185 deaths of disease of circulatory system were observed. Among the 185 deaths, 82 were due to cerebrovascular disease (I60-I69); 96 were due to ischemic heart disease (I20-I25, n = 64) or other forms of heart disease (I30-I52, n = 32), which included mostly deaths due to heart failure (I50.0-I50.9, n = 31) and ventricular tachycardia (I47, n = 1); and 7 deaths were due to pulmonary heart disease, hypertensive heart disease, or multiple valve diseases (I08, I11, and I27). Each diet pattern was divided into quartiles by factor score and analyzed for differences in various sociodemographic, lifestyle, and CVD risk factors (Table 1). Adherence to the animal protein diet was associated with a higher proportion of males, television and land ownership, older age, longer education, and higher BMI. Conversely, those following the root vegetable diet were more likely to be women, not television or land owners, and less educated.
Table 1.
Sociodemographic characteristics by quartile (Q) of dietary factor score
| Quartile of factor scores for the Balanced diet pattern |
Quartile of factor scores for the Animal protein diet pattern |
Quartile of factor scores for the Gourd and root vegetable diet pattern |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Q1 Mean |
Q4 Mean |
p for trend1 or correlations 2, 3 |
Q1 Mean |
Q4 Mean |
p for trend1 or correlations 2, 3 |
Q1 Mean |
Q4 Mean |
p for trend1 or correlations 2, 3 |
|
| Male (%) | 47.7 | 42.6 | <0.01 | 31.8 | 63.7 | <0.01 | 50.4 | 37.5 | <0.01 |
| Nonsmokers (%) | 60.8 | 63.7 | 0.87 | 71.9 | 49.3 | 0.61 | 60.7 | 68.3 | 0.88 |
| Own a television (%) | 35.8 | 37.8 | 0.09 | 20.5 | 53.8 | <0.01 | 37.9 | 32.1 | <0.01 |
| Own land >1 acre (%) | 14.3 | 18.8 | <0.01 | 8.5 | 26.3 | <0.01 | 15.9 | 17.0 | 0.07 |
| Age (y) | 37.3 | 37.6 | 0.01 | 35.7 | 39.4 | 0.14** | 37.1 | 36.7 | −0.01 |
| Education (y) | 3.6 | 3.6 | 0.02 | 2.2 | 5.2 | 0.30** | 3.7 | 3.4 | −0.01 |
| Systolic blood pressure | 115.4 | 114.9 | −0.01 | 113.1 | 116.9 | 0.08** | 114.4 | 115.3 | 0.02* |
| Diastolic blood pressure | 75.0 | 73.7 | −0.03** | 72.7 | 75.5 | 0.09** | 73.8 | 74.2 | 0.01 |
| BMI (kg/m2) | 20.0 | 19.8 | 0.00 | 19.3 | 20.7 | 0.19** | 19.9 | 19.7 | −0.01 |
Trend tests were performed by using linear regression that included energy-adjusted factor scores as the dependent variable and age and sex as the independent variables.
Pearson correlations between the factor scores and natural logs of energy-adjusted nutrient values.
.01< p <.05,
.01< p <.05,
p <.01
Higher levels of body mass index and diabetes status were associated with an increased risk of death due to heart disease (Table 2). There was no association of educational level with mortality from either disease of the circulatory system overall or any of the subtypes of cardiovascular disease. Diastolic and systolic hypertension were related to increased risk of death from disease of the circulatory system overall and all subtypes of cardiovascular disease, and the associations were stronger for cerebrovascular disease. Participants who were current smokers were 1.7 times more likely to die from disease of circulatory system and 2.3 times more likely to die from heart disease.
Table 2.
Relationships between Baseline Risk Factors and Mortality from Disease of the Circulatory System
| Disease of the circulatory system |
Heart disease |
Cerebrovascular disease |
|||||
|---|---|---|---|---|---|---|---|
| Risk Factors | Follow-up Person-years |
N of Deaths |
Age and Sex Adjusted HR |
N of Deaths |
Age and Sex Adjusted HR |
N of Deaths |
Age and Sex Adjusted HR |
| Body Mass Index | |||||||
| <18.5 | 28768.09 | 83 | 1.02 (0.73-1.43) | 36 | 0.91 (0.56-1.47) | 44 | 0.72 (0.39-1.31) |
| 18.5-22 | 28926.07 | 58 | 1.00 | 30 | 1.00 | 25 | 1.00 |
| >22 | 15379.07 | 44 | 1.34 (0.90-1.98) | 30 | 1.79 (1.08-2.97) | 13 | 0.90 (0.46-1.75) |
| Education (years)1 | |||||||
| 0 years | 32214.23 | 84 | 1.00 | 43 | 1.00 | 38 | 1.00 |
| 1-5 years | 21758.49 | 44 | 0.85 (0.59-1.22) | 23 | 0.85 (0.51-1.41) | 19 | 0.83 (0.48-1.44) |
| 5+ years | 19067.65 | 57 | 1.20 (0.87-1.70) | 30 | 1.23 (0.76-1.98) | 25 | 1.23 (0.73-2.05) |
| Systolic Blood Pressure2 | |||||||
| <140 mmHg | 67510.56 | 123 | 1.00 | 75 | 1.00 | 41 | 1.00 |
| >140 mmHg | 5251.32 | 61 | 3.21 (2.33-4.42) | 21 | 2.04 (1.24-3.38) | 40 | 5.55 (3.01-8.77) |
| Diastolic Blood Pressure3 | |||||||
| <90 mmHg | 66910.57 | 144 | 1.00 | 85 | 1.00 | 53 | 1.00 |
| >90 mmHg | 5819.27 | 41 | 2.56 (1.81-3.63) | 11 | 1.23 (0.65-2.30) | 29 | 4,64 (2.95-7.31) |
| Smoking | |||||||
| Never | 47389.05 | 48 | 1.00 | 26 | 1.00 | 20 | 1.00 |
| Past | 4643.23 | 28 | 1.14 (0.65-1.98) | 11 | 1.24 (0.54-2.84) | 13 | 0.93 (0.41-2.12) |
| Current | 21040.97 | 109 | 1.73 (1.10-2.71) | 59 | 2.30 (1.22-4.35) | 49 | 1.59 (0.81-3.11) |
n=5 missing education
n=46 missing systolic blood pressure
n=51 missing diastolic blood pressure
Adjusted hazard ratios for disease of circulatory system, heart disease, and cerebrovascular disease in relation to each dietary pattern are shown in Table 3. An inverse association was observed between the balanced diet pattern and disease of the circulatory system; when diet pattern was used as a continuous variable, HR for disease of circulatory system was 0.86 (95% CI = 0.73-1.01, p = 0.07, model 2) for each SD increase in the balanced diet pattern. Individuals in the highest quartile of the balanced diet pattern had reduced risk for death from disease of the circulatory system (HR = 0.86, 95% CI = 0.57-1.30, model 2) and heart disease (HR = 0.79, 95% CI = 0.44-1.40, model 2). However, the point estimates were not statistically significance. Adherence to the animal protein diet was positively related to death from disease of the circulatory system, especially heart disease mortality. When the diet pattern was a continuous variable, the HRs for deaths from disease of the circulatory system and heart disease was 1.13 (95% CI = 1.00-1.28, p = 0.05, model 2) and 1.17 (95% CI = 0.99-1.38, p = 0.07, model 2), respectively, in relation to one standard deviation increase in factor scores of the animal protein diet pattern. There was a dose-response relationship between animal protein diet pattern and risk of heart disease mortality, with the highest quartile of the animal protein diet possessing a risk increase of more than 100% (HR (95% CI): 1.0 (reference), 1.43 (0.68-3.01), 1.65 (0.81-3.38), 2.11 (0.81-4.17); (p for trend = 0.02, model 1)). The significance of the association was slightly attenuated after controlling for body mass index, and smoking status, in addition to age, gender, and energy intake (p for trend=0.05, model 2). A non-significant inverse association between increasing adherence to the animal protein diet and stroke mortality was also observed (p for trend =0.22). Adherence to gourd and root vegetable diet pattern was not related to risk of overall circulatory system disease, heart disease, or cerebrovascular disease.
Table 3.
Hazard Ratios for Disease of the Circulatory System in relation to Diet Patterns
| HR (95% CI) by quartile of diet pattern |
HR (95% CI) in relation to one SD diet pattern score |
||||||
|---|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | p for trend2 |
HR (95% CI) | p-value | |
| Balanced diet pattern | |||||||
| Disease of the circulatory system | |||||||
| Deaths, n/person-years | 45/18335 | 55/18472 | 45/18779 | 46/19331 | 185/73073 | ||
| Model 1* | 1.0 | 1.30 (0.88-1.94) | 0.96 (0.64-1.46) | 0.87 (0.56-1.32) | 0.26 | 0.85 (0.72-1.00) | 0.05 |
| Model 21 | 1.0 | 1.28 (0.86-1.91) | 0.96 (0.63-1.46) | 0.86 (0.57-1.30) | 0.24 | 0.86 (0.73-1.01) | 0.07 |
| Heart disease | |||||||
| Deaths, n/person-years | 25/18335 | 25/18472 | 28/18779 | 23/19331 | 96/73073 | ||
| Model 1* | 1.0 | 1.02 (0.58-1.79) | 1.07 (0.62-1.85) | 0.80 (0.45-1.42) | 0.51 | 0.85 (0.67-1.06) | 0.15 |
| Model 21 | 1.0 | 1.01 (0.57-1.80) | 1.06 (0.61-1.85) | 0.79 (0.44-1.40) | 0.47 | 0.86 (0.69-1.08) | 0.19 |
| Cerebrovascular disease | |||||||
| Deaths, n/person-years | 17/18335 | 29/18472 | 15/18779 | 22/19331 | 82/73073 | ||
| Model 1* | 1.0 | 1.95 (1.07-3.56) | 0.87 (0.44-1.75) | 1.08 (0.57-2.04) | 0.51 | 0.89 (0.71-1.13) | 0.34 |
| Model 21 | 1.0 | 1.88 (1.03-3.46) | 0.87 (0.43-1.75) | 1.05 (0.56-1.99) | 0.49 | 0.89 (0.70-1.13) | 0.32 |
| Animal protein pattern | |||||||
| Disease of the circulatory system | |||||||
| Deaths, n/person-years | 28/18700 | 40/18625 | 48/18727 | 75/18865 | 185/73073 | ||
| Model 1* | 1.0 | 1.08 (0.66-1.77) | 1.10 (0.68-1.77) | 1.25 (0.80-1.97) | 0.29 | 1.16 (1.03-1.31) | 0.0146 |
| Model 21 | 1.0 | 1.07 (0.65-1.76) | 1.11 (0.69-1.81) | 1.20 (0.75-1.91) | 0.41 | 1.13 (1.00-1.28) | 0.05 |
| Heart disease | |||||||
| Deaths, n/person-years | 11/18700 | 20/18625 | 26/18727 | 44/18865 | 96/73073 | ||
| Model 1* | 1.0 | 1.43 (0.68-3.01) | 1.65 (0.81-3.38) | 2.11(0.81-4.17) | 0.02 | 1.24 (1.06-1.46) | 0.0068 |
| Model 21 | 1.0 | 1.38 (0.63-3.04) | 1.76 (0.84-3.71) | 1.94 (0.95-4.00) | 0.05 | 1.17 (0.99-1.38) | 0.07 |
| Cerebrovascular disease | |||||||
| Deaths, n/person-years | 17/18700 | 18/18625 | 20/18727 | 28/18865 | 82/73073 | ||
| Model 1* | 1.0 | 0.77 (0.40-1.51) | 0.70 (0.36-1.36) | 0.71 (0.38-1.33) | 0.32 | 1.06 (0.88-1.28) | 0.55 |
| Model 21 | 1.0 | 0.80 (0.41-1.56) | 0.69 (0.36-1.36) | 0.74 (0.39-1.41) | 0.38 | 1.08 (0.89-1.31) | 0.46 |
| Gourd and Root Vegetable Diet Pattern | |||||||
| Disease of the circulatory system | |||||||
| Deaths, n/person-years | 46/18732 | 52/18513 | 49/18765 | 44/18906 | 185/73073 | ||
| Model 1* | 1.0 | 1.06 (0.71-1.58) | 1.04 (0.70-1.56) | 1.02 (0.68-1.55) | 0.94 | 1.02 (0.89-1.18) | 0.77 |
| Model 21 | 1.0 | 1.05 (0.70-1.58) | 1.06 (0.70-1.59) | 1.04 (0.69-1.59) | 0.84 | 1.03 (0.89-1.18) | 0.70 |
| Heart disease | |||||||
| Deaths, n/person-years | 22/18732 | 33/18513 | 23/18765 | 23/18906 | 96/73073 | ||
| Model 1* | 1.0 | 1.42 (0.82-2.45) | 1.04 (0.58-1.86) | 1.13 (0.63-2.02) | 0.99 | 1.09 (0.91-0.32) | 0.35 |
| Model 21 | 1.0 | 1.40 (0.80-2.44) | 1.09 (0.50-1.98) | 1.17 (0.64-1.98) | 0.86 | 1.10 (0.92-1.33) | 0.30 |
| Cerebrovascular disease | |||||||
| Deaths, n/person-years | 22/18732 | 17/18513 | 24/18765 | 20/18906 | 82/73073 | ||
| Model 1* | 1.0 | 0.74 (0.39-1.39) | 1.06 (0.59-1.90) | 0.97 (0.53-1.79) | 0.78 | 0.94 (0.75-1.18) | 0.62 |
| Model 21 | 1.0 | 0.75 (0.39-1.41) | 1.03 (0.57-1.86) | 0.99 (0.54-1.86) | 0.79 | 0.95 (0.76-1.19) | 0.65 |
HRs were adjusted for age, gender, energy intake.
HRs were adjusted for age, gender, energy intake, body mass index, and smoking status.
P-value for trend was estimated using quartile value of dietary pattern as a continuous variable in the model.
Stratified analyses by smoking status are shown in Table 4. The association between the balanced diet pattern and risk of disease of circulatory system remained among never smokers, with an HR of 0.67 (95% CI, 0.45-1.01; p=0.05) for each SD increase in the score for the balanced diet pattern. Smokers adhering to the animal protein diet were shown to have significantly increased risk of disease of the circulatory system (HR=1.17, 95% CI, 1.02-1.34; p=0.03) and heart disease (HR=1.20, 95% CI, 1.00-1.45; p=0.05), although the test for interaction by smoking status was not significant.
Table 4.
Hazard Ratios for Disease of the Circulatory System in relation to Diet Patterns by smoking status
| HR (95% CI) in relation to one SD diet pattern score |
|||||
|---|---|---|---|---|---|
| Non-smokers |
Smokers |
||||
| HR (95% CI) | p-value | HR (95% CI) | p-value | P for interaction | |
| Balanced diet pattern | |||||
| Disease of the circulatory system | |||||
| Deaths, n/person-years | 48/47,389 | 137/25,684 | |||
| Model 1* | 0.67 (0.45-1.01) | 0.05 | 0.90 (0.74-1.10) | 0.29 | 0.16 |
| Heart disease | |||||
| Deaths, n/person-years | 26/47,389 | 70/25,684 | |||
| Model 1* | 0.83 (0.51-1.35) | 0.45 | 0.85 (0.64-1.13) | 0.25 | 0.76 |
| Cerebrovascular disease | |||||
| Deaths, n/person-years | 20/47,389 | 62/25,684 | |||
| Model 1* | 0.52 (0.25-1.04) | 0.06 | 0.99 (0.76-1.30) | 0.96 | 0.10 |
| Animal protein pattern | |||||
| Disease of the circulatory system | |||||
| Deaths, n/person-years | 48/47389 | 137/25,684 | |||
| Model 1* | 1.02 (0.74-1.39) | 0.92 | 1.17 (1.02-1.34) | 0.03 | 0.44 |
| Heart disease | |||||
| Deaths, n/person-years | 26/47,389 | 70/25,684 | |||
| Model 1* | 1.21 (0.84-1.73) | 0.30 | 1.20 (1.00-1.45) | 0.05 | 0.87 |
| Cerebrovascular disease | |||||
| Deaths, n/person-years | 20/47,389 | 62/25,684 | |||
| Model 1* | 0.78 (0.42-1.46) | 0.43 | 1.11 (0.90-1.36) | 0.34 | 0.19 |
| Gourd and Root Vegetable Diet Pattern | |||||
| Disease of the circulatory system | |||||
| Deaths, n/person-years | 48/47,389 | 137/25,684 | |||
| Model 1* | 1.16 (0.90-1.49) | 0.25 | 0.98 (0.83-1.16) | 0.79 | 0.36 |
| Heart disease | |||||
| Deaths, n/person-years | 26/47,389 | 70/25,684 | |||
| Model 1* | 1.22 (0.90-1.66) | 0.20 | 1.04 (0.83-1.31) | 0.72 | 0.88 |
| Cerebrovascular disease | |||||
| Deaths, n/person-years | 20/47,389 | 62/25,684 | |||
| Model 1* | 1.01 (0.64-1.59) | 0.96 | 0.94 (0.72-1.21) | 0.62 | 0.86 |
HRs were adjusted for age, gender, energy intake, body mass index.
DISCUSSION
We evaluated the impact of dietary patterns on cardiovascular mortality in a cohort of 11,116 participants in rural Bangladesh. The most noteworthy finding was increased risk of mortality due to overall cardiovascular disease (disease of the circulatory system) and heart disease in those adhering to the animal protein diet, especially among smokers.
Our study population presented a novel context within which to explore these topics, as FFQ-based pattern analyses of diet and cardiovascular mortality have thus far been presented mainly for westernized societies (7-11, 27) and, to a much lesser extent, some populations in East Asia (12-13). Bangladesh provides an especially interesting backdrop for such a study, with the country in the midst of an “epidemiologic transition” from the major causes of mortality being based in nutritional deficiencies and infectious diseases to primarily degenerative diseases, such as cancer, cardiovascular disease, and diabetes (2-4). Though Bangladesh is still in the early stages of this transition, they are rapidly improving nutritional status while adopting a diet that is higher in fat, iron, and calcium (28, 29). In the 15 years preceding 1996, this contributed to the prevalence of chronic energy deficiency falling from 78% to 64% (28).
Several studies have explored the relationship between dietary patterns and CHD, yielding somewhat inconsistent results. A prospective cohort study of US male health professionals showed significant CHD risk associated with a “Western diet,” characterized by more frequent intake of red meat, refined starch, sweets, and high-fat dairy (7). A Western diet was similarly associated with increased CHD risk among women in the Nurses’ Health Study (9). Two recent studies of Asian populations compared more “Westernized,” meat-heavy diets with traditional Chinese or Japanese dietary patterns, finding higher risk of CHD in the former (P(trend)=0.05 [12]; P(trend)=0.176 [13]). Additionally, two studies found a positive association between Western diet and plasma biomarkers of cardiovascular risk (c-peptide, insulin, glycated hemoglobin [30], leptin and homocysteine [31]), while negatively correlating with concentration of plasma folate, a cardioprotective agent (30, 31). It should be noted that two published reports have not supported the correlation between a Western or animal protein diet and CHD (10, 11).
A recent literature review summarized factors that are “protective” against CHD to include vegetables, nuts, and monounsaturated fats, as well as Mediterranean, prudent, and high-quality diets. “Harmful” factors included trans-fatty acids, foods with a high glycemic index, and a Western dietary pattern (32). It has been suggested that failure to detect increased CHD risk among Western diet adherents in some countries could be due to a given culture’s propensity to include healthy or cardioprotective food items (such as plants [11], light breads, or juices [10]) in their diet, even if they are consumed in lesser proportion because of displacement by more unhealthy, Western food items. This may explain the marginal significance observed for the meat-based diet of our study as well, considering that the animal protein diet still displayed high factor loadings for multiple fruits. This tendency for cardioprotective foods to appear across dietary patterns in Bangladesh could also explain our marginal significance for decreased risk of cardiovascular disease associated with consumption of a balanced diet.
Our finding that a diet higher in saturated fat may be protective against stroke, though not statistically significant, has been suggested in previous studies (33-36). Among a cohort of Japanese and Japanese-American men, Reed and colleagues observed that low-fat and low-protein diets correlated with higher cerebral infarction mortality (33). Similarly, an analysis of ischemic stroke in the Framingham Heart Study yielded a decline in risk associated with increased intake of total fat, saturated fat, and monounsaturated fat (35). The paradoxical findings in the associations of animal protein diet with heart disease and cerebrovascular disease may be explained through a pathological difference in occlusive processes of small and large arteries, in particular a potential protective effect of cholesterol on small vessel endothelium (35, 37, 38). Risk of large vessel stroke correlates with risk of CHD (also a large vessel process), and is the predominant cause of stroke in Western countries. Small vessel strokes are the predominant cause of stroke in non-Western countries, with this subtype having been shown to account for 76% of cerebral infarctions in a Japanese autopsy study (39) (compared to 20% in Western countries [38]). Future studies are needed to further investigate diet patterns and subtypes of stroke.
Cigarette smoking has been shown in multiple studies to have a cumulative effect on development of cardiovascular disease, a finding that was supported in our population (among smokers, HR in relation to one SD diet pattern score for animal protein pattern: p = 0.03, disease of the circulatory system; p = 0.05, heart disease) (40-42). It is most strongly associated with increased risk of aortic aneurysm and peripheral vascular disease, with cardiovascular dysfunction mediated by oxidative stress that yields inflammation, thrombosis, and oxidation of low-density lipoprotein cholesterol (43, 44). Future studies within our population would be needed to analyze the degree of increase in inflammatory markers or carotid artery intima-medial thickness relative to smoking status and dietary intake.
Strengths of our study include a large, population-based sample size and a comprehensive, validated FFQ (15). Additionally, religious motivations lead to restricted alcohol consumption in Bangladesh, thus removing a potential cofounder. Furthermore, because health screening is not as robust as in more developed countries, there is a smaller chance of participants changing their diet because of risk factors for chronic disease. For example, similar studies of CHD in the United States have had to consider whether dietary groupings were skewed by individuals following a defined “healthy” or “fruits and vegetables” diet only because screening tests or medical intervention caused them to change their diet after decades of more unhealthy food consumption. Limitations include the delineation of food patterns itself. Though pattern analysis can be argued to be a more effective, realistic method of assessing food consumption and risk of cardiovascular disease, there is still a measure of subjectivity in selection of food groups for factor analysis, post-analysis identification of patterns, and choosing the method of vector rotation (5, 45). Furthermore, identification of diet patterns in Bangladesh is more nuanced than in countries where dramatically different patterns may exist; even among higher economic classes in Bangladesh, there is still a prioritized consumption of staple foods (46). This dietary homogeneity also limits the generalizability of our results to populations of more varied culture or socioeconomic background.
Our results underscore the importance of better understanding epidemiologic transition in developing countries such as Bangladesh, and responding with appropriate public health measures – not only in consideration of communicable diseases, environmental contamination, and nutrition deficiency, but also recognizing the potential for additional, widespread health issues on the horizon. As economic development continues in this country, changes in dietary patterns will usher in a future of more overweight individuals and higher chronic disease mortality. Recent studies have shown that from 1996 to 2004 alone, the proportion of overweight females has increased drastically in both urban (from 10 to 20%) and rural areas (from 2 to 6%) (47). This could be explained through the consumption of higher fat foods correlating closely with socioeconomic advances. A recent study in Bangladesh identified meat and dairy consumption as increasing to a greater degree than any other food items when comparing the richest economic class to the poorest (46). Though this trend will produce a double burden of overnutrition and undernutrion in Bangladesh, certain geographical, economic, or social characteristics will likely dictate which epidemic is present in a given subset of the population. Herein lies the need for focused interventions that will not only address those currently afflicted by malnourishment and communicable illness, but also stave off the burden of chronic disease that will otherwise coincide with increased prosperity and development in Bangladesh as well as other developing countries.
Acknowledgement
The authors of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology [48].
Funding: The research is supported by US National Institute of Health funding. The National Institute of Environmental Health Sciences and the National Cancer Institute provide support under the following grants: P42ES010349, R01ES017541, R01CA102484, R01CA107431, R01ES017541, R01ESO11601, P30ES09089, CA016087, ES000260, and CA014599.
Statement of the independence of researchers from funders: The views expressed in this paper are those of the authors and not necessarily those of any funding body or others whose support is acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Appendix 1
Factor Loading Matrix1 for Dietary Patterns Derived from Principle component Analysis
| “Balanced” diet | “Animal protein” | “Root vegetable” | |
|---|---|---|---|
| Tea | - | 0.48 | - |
| Big fish (fresh water) | - | 0.44 | - |
| Eggs (hen eggs) | - | 0.49 | - |
| Milk | - | 0.49 | - |
| Poultry (fowl) | - | 0.47 | - |
| Puffed rice | - | 0.22 | - |
| Lentil | - | 0.37 | - |
| Wheat bread (brown) | −0.16 | 0.49 | - |
| Beef or mutton | 0.21 | 0.45 | - |
| Bitter gourd (a kind of squash) | 0.43 | 0.18 | 0.28 |
| Banana | 0.16 | 0.47 | - |
| Jack Fruit | 0.30 | 0.22 | - |
| Watermelon | 0.20 | 0.29 | - |
| Mango | 0.39 | 0.22 | - |
| Steamed rice | 0.24 | −0.29 | - |
| Guava | 0.27 | - | - |
| Beans (scarlet runner) | 0.62 | - | - |
| Potato | 0.22 | - | - |
| Bottle gourd (a kind of squash) | 0.46 | - | - |
| Small fish (fresh water) | 0.39 | - | - |
| Eggplant | 0.33 | - | 0.17 |
| Cauliflower | 0.64 | - | 0.16 |
| Parwar/patol (a kind of squash) | 0.22 | 0.17 | 0.37 |
| Green papaya | 0.22 | - | 0.35 |
| Cabbage | 0.51 | - | 0.24 |
| Okra | 0.29 | - | 0.37 |
| Yam | 0.31 | - | 0.26 |
| Spinach | 0.24 | - | 0.25 |
| Dried fish | - | - | 0.22 |
| Ridge gourd/jhinga (a kind of squash) | - | - | 0.53 |
| Snake gourd/chachinga (a kind of squash) |
- | - | 0.57 |
| Ghosala/dhundal (a kind of squash) | - | - | 0.51 |
| Radish | - | - | 0.48 |
| Spinach stalks | - | - | 0.48 |
| Sweet potato | - | - | 0.29 |
| Water Rice | - | - | 0.17 |
| Pumpkin | - | - | 0.39 |
| Tomato | - | - | - |
| Salted fish | - | - | - |
Factor loadings represent the correlation between factor scores and intakes of food items. Positive Factor loadings ≤ 0.15 and negative factor loadings ≥ −0.15 were omitted in the table for simplicity
Footnotes
Contributors: YC and HA designed the study. HA and YC obtained funding. GS maintained and supervised the HEALS cohort database. FP, TI, AA, MRZ, and RH supervised the field work. SS and YC analyzed the data. TRM, YC, HA, and FP helped to interpret the results. TRM, YC, and HA wrote the paper. HA is guarantor.
Ethical approval: The study procedures were approved by the Ethical Committee of the Bangladesh Medical Research Council and the Institutional Review Boards of Columbia University and the University of Chicago. Verbal consent was obtained from study participants.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- [1].Reddy KS, Yusuf S. Emerging epidemic of cardiovascular disease in developing countries. Circulation. 1998 Feb 17;97(6):596–601. doi: 10.1161/01.cir.97.6.596. [DOI] [PubMed] [Google Scholar]
- [2].Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases - Part I: General considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation. 2001 Nov 27;104(22):2746–2753. doi: 10.1161/hc4601.099487. [DOI] [PubMed] [Google Scholar]
- [3].Omran AR. Epidemiologic Transition - Theory of Epidemiology of Population Change. Milbank Meml Fund Q. 1971;49(4):509. &. [PubMed] [Google Scholar]
- [4].Islam M. Mortality and Epidemiological Transition in Bangladesh: Lessons and Experiences to the Developing Countries in the New Millennium. J Epidemiol Commun H. 2011 Aug;65:A116–A116. [Google Scholar]
- [5].Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol. 2002 Feb;13(1):3–9. doi: 10.1097/00041433-200202000-00002. [DOI] [PubMed] [Google Scholar]
- [6].Willett WC. Diet and Health - What Should We Eat. Science. 1994 Apr 22;264(5158):532–537. doi: 10.1126/science.8160011. [DOI] [PubMed] [Google Scholar]
- [7].Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC. Prospective study of major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr. 2000 Oct;72(4):912–921. doi: 10.1093/ajcn/72.4.912. [DOI] [PubMed] [Google Scholar]
- [8].Kant AK, Schatzkin A, Graubard BI, Schairer C. A prospective study of diet quality and mortality in women. Jama-J Am Med Assoc. 2000 Apr 26;283(16):2109–2115. doi: 10.1001/jama.283.16.2109. [DOI] [PubMed] [Google Scholar]
- [9].Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu FB. Dietary patterns and the risk of coronary heart disease in women. Arch Intern Med. 2001 Aug 13;161(15):1857–1862. doi: 10.1001/archinte.161.15.1857. [DOI] [PubMed] [Google Scholar]
- [10].Osler M, Heitmann BL, Gerdes LU, Jorgensen LM, Schroll M. Dietary patterns and mortality in Danish men and women: a prospective observational study. Brit J Nutr. 2001 Feb;85(2):219–225. doi: 10.1079/bjn2000240. [DOI] [PubMed] [Google Scholar]
- [11].Harriss LR, English DR, Powles J, et al. Dietary patterns and cardiovascular mortality in the melbourne collaborative cohort study. Am J Clin Nutr. 2007 Jul;86(1):221–229. doi: 10.1093/ajcn/86.1.221. [DOI] [PubMed] [Google Scholar]
- [12].Shimazu T, Kuriyama S, Hozawa A, et al. Dietary patterns and cardiovascular disease mortality in Japan: a prospective cohort study. International Journal of Epidemiology. 2007 Jun;36(3):600–609. doi: 10.1093/ije/dym005. [DOI] [PubMed] [Google Scholar]
- [13].Cai H, Shu XO, Gao YT, Li HL, Yang G, Zheng W. A prospective study of dietary patterns and mortality in Chinese women. Epidemiology. 2007 May;18(3):393–401. doi: 10.1097/01.ede.0000259967.21114.45. [DOI] [PubMed] [Google Scholar]
- [14].Chen Y, Factor-Litvak P, Howe GR, Parvez F, Ahsan H. Nutritional influence on risk of high blood pressure in Bangladesh: a population-based cross-sectional study. Am J Clin Nutr. 2006 Nov;84(5):1224–1232. doi: 10.1093/ajcn/84.5.1224. [DOI] [PubMed] [Google Scholar]
- [15].Chen Y, Ahsan H, Parvez F, Howe GR. Validity of a food-frequency questionnaire for a large prospective cohort study in Bangladesh. Brit J Nutr. 2004 Nov;92(5):851–859. doi: 10.1079/bjn20041277. [DOI] [PubMed] [Google Scholar]
- [16].USDA National Nutrient Database for Standard Reference. 2002 Release 15. http://www.ars.usda.gov/ba/bhnrc/ndl.
- [17].Gopalan C, Rama Sastri BV, Balasubramanian SC. In: Nutritive value of Indian foods. Indian Council of Medical Research NIoN, editor. Hyderabad, India: 1989. [Google Scholar]
- [18].Pierce BL, Kalra T, Argos M, et al. A prospective study of body mass index and mortality in Bangladesh. International Journal of Epidemiology. 2010 Aug;39(4):1037–1045. doi: 10.1093/ije/dyp364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Aho K, Harmsen P, Hatano S, Marquardsen J, Smirnov VE, Strasser T. Cerebrovascular-Disease in the Community - Results of a Who Collaborative Study. B World Health Organ. 1980;58(1):113–130. [PMC free article] [PubMed] [Google Scholar]
- [20].Word Health Organization . Manual of the international classification of diseases, injuries and causes of death. WHO; Geneva: 2007. [Google Scholar]
- [21].Ronsmans C, Vanneste AM, Chakraborty J, Van Ginneken J. A comparison of three verbal autopsy methods to ascertain levels and causes of maternal deaths in Matlab, Bangladesh. International Journal of Epidemiology. 1998 Aug;27(4):660–666. doi: 10.1093/ije/27.4.660. [DOI] [PubMed] [Google Scholar]
- [22].Baqui AH, Black RE, Arifeen SE, Hill K, Mitra SN, Al Sabir A. Causes of childhood deaths in Bangladesh: results of a nationwide verbal autopsy study. B World Health Organ. 1998;76(2):161–171. [PMC free article] [PubMed] [Google Scholar]
- [23].Sohel N, Persson LA, Rahman M, et al. Arsenic in Drinking Water and Adult Mortality A Population-based Cohort Study in Rural Bangladesh. Epidemiology. 2009 Nov;20(6):824–830. doi: 10.1097/EDE.0b013e3181bb56ec. [DOI] [PubMed] [Google Scholar]
- [24].Hatcher L. A step-by-step approach to using the SAS system for factor analysis and structural equation modeling. SAS Institute; Cary, NC: 1994. [Google Scholar]
- [25].Pierce BL, Argos M, Chen Y, et al. Arsenic Exposure, Dietary Patterns, and Skin Lesion Risk in Bangladesh: A Prospective Study. American Journal of Epidemiology. 2010 Jun 1;171:S3–S3. doi: 10.1093/aje/kwq366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Chen Y, Graziano JH, Parvez F, et al. Arsenic exposure from drinking water and mortality from cardiovascular disease in Bangladesh: prospective cohort study. Brit Med J. 2011 May 5;342 doi: 10.1136/bmj.d2431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Heidemann C, Schulze MB, Franco OH, van Dam RM, Mantzoros CS, Hu FB. Dietary patterns and risk of mortality from cardiovascular disease, cancer, and all causes in a prospective cohort of women. Circulation. 2008 Jul 15;118(3):230–237. doi: 10.1161/CIRCULATIONAHA.108.771881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Hels O, Hassan N, Tetens I, Thilsted SH. Food consumption, energy and nutrient intake and nutritional status in rural Bangladesh: changes from 1981-1982 to 1995-96. Eur J Clin Nutr. 2003 Apr;57(4):586–594. doi: 10.1038/sj.ejcn.1601567. [DOI] [PubMed] [Google Scholar]
- [29].Jahan K, Hossain M. Bangladesh National Nutrition Survey, 1995 – 96. Institute of Nutrition and Food Science, University of Dhaka; Dhaka: 1998. Nature and Extent of Malnutrition in Bangladesh. [Google Scholar]
- [30].Kerver JM, Yang EJ, Bianchi L, Song WO. Dietary patterns associated with risk factors for cardiovascular disease in healthy US adults. Am J Clin Nutr. 2003 Dec;78(6):1103–1110. doi: 10.1093/ajcn/78.6.1103. [DOI] [PubMed] [Google Scholar]
- [31].Fung TT, Rimm EB, Spiegelman D, et al. Association between dietary patterns and plasma biomarkers of obesity and cardiovascular disease risk. Am J Clin Nutr. 2001 Jan;73(1):61–67. doi: 10.1093/ajcn/73.1.61. [DOI] [PubMed] [Google Scholar]
- [32].Mente A, de Koning L, Shannon HS, Anand SS. A Systematic Review of the Evidence Supporting a Causal Link Between Dietary Factors and Coronary Heart Disease. Arch Intern Med. 2009 Apr 13;169(7):659–669. doi: 10.1001/archinternmed.2009.38. [DOI] [PubMed] [Google Scholar]
- [33].Reed DM, Resch JA, Hayashi T, Maclean C, Yano K. A Prospective-Study of Cerebral-Artery Atherosclerosis. Stroke. 1988 Jul;19(7):820–825. doi: 10.1161/01.str.19.7.820. [DOI] [PubMed] [Google Scholar]
- [34].Reed DM. The Paradox of High-Risk of Stroke in Populations with Low-Risk of Coronary Heart-Disease. American Journal of Epidemiology. 1990 Apr;131(4):579–588. doi: 10.1093/oxfordjournals.aje.a115542. [DOI] [PubMed] [Google Scholar]
- [35].Gillman MW, Cupples LA, Millen BE, Ellison RC, Wolf PA. Inverse association of dietary fat with development of ischemic stroke in men. Jama-J Am Med Assoc. 1997 Dec 24;278(24):2145–2150. [PubMed] [Google Scholar]
- [36].Sauvaget C, Nagano J, Hayashi M, Yamada M. Animal protein, animal fat, and cholesterol intakes and risk of cerebral infarction mortality in the adult health study. Stroke. 2004 Jul;35(7):1531–1537. doi: 10.1161/01.STR.0000130426.52064.09. [DOI] [PubMed] [Google Scholar]
- [37].Kuller L, Reisler DM. Explanation for Variations in Distribution of Stroke and Arteriosclerotic Heart Disease among Populations and Racial Groups. American Journal of Epidemiology. 1971;93(1):1–9. doi: 10.1093/oxfordjournals.aje.a121223. [DOI] [PubMed] [Google Scholar]
- [38].Warlow CP. Epidemiology of stroke. Lancet. 1998 Oct;352:Siii1–Siii4. doi: 10.1016/s0140-6736(98)90086-1. [DOI] [PubMed] [Google Scholar]
- [39].Masuda J, Tanaka K, Omae T, Ueda K, Sadoshima S. Cerebrovascular diseases and their underlying vascular lesions in Hisayama, Japan—a pathological study of autopsy cases. Stroke. 1983;14:934–940. doi: 10.1161/01.str.14.6.934. [DOI] [PubMed] [Google Scholar]
- [40].Wilson PW, Hoeg JM, D’Agostino RB, et al. Cumulative effects of high cholesterol levels, high blood pressure, and cigarette smoking on carotid stenosis. N Engl J Med. 1997 Aug 21;337(8):516–522. doi: 10.1056/NEJM199708213370802. [DOI] [PubMed] [Google Scholar]
- [41].Millen BE, Quatromoni PA, Nam BH, et al. Dietary patterns, smoking, and subclinical heart disease in women: opportunities for primary prevention from the Framingham Nutrition Studies. J Am Diet Assoc. 2004 Feb;104(2):208–214. doi: 10.1016/j.jada.2003.11.007. [DOI] [PubMed] [Google Scholar]
- [42].Meyer J, Doring A, Herder C, Roden M, Koenig W, Thorand B. Dietary patterns, subclinical inflammation, incident coronary heart disease and mortality in middle-aged men from the MONICA/KORA Augsburg cohort study. Eur J Clin Nutr. 2011 Jul;65(7):800–807. doi: 10.1038/ejcn.2011.37. [DOI] [PubMed] [Google Scholar]
- [43].Benowitz NL. Cigarette smoking and cardiovascular disease: pathophysiology and implications for treatment. Prog Cardiovasc Dis. 2003 Jul-Aug;46(1):91–111. doi: 10.1016/s0033-0620(03)00087-2. [DOI] [PubMed] [Google Scholar]
- [44].Ambrose JA, Barua RS. The pathophysiology of cigarette smoking and cardiovascular disease: an update. J Am Coll Cardiol. 2004 May 19;43(10):1731–1737. doi: 10.1016/j.jacc.2003.12.047. [DOI] [PubMed] [Google Scholar]
- [45].Martinez ME, Marshall JR, Sechrest L. Invited commentary: Factor analysis and the search for objectivity. American Journal of Epidemiology. 1998 Jul 1;148(1):17–19. doi: 10.1093/oxfordjournals.aje.a009552. [DOI] [PubMed] [Google Scholar]
- [46].Thorne-Lyman AL, Valpiani N, Sun K, et al. Household Dietary Diversity and Food Expenditures Are Closely Linked in Rural Bangladesh, Increasing the Risk of Malnutrition Due to the Financial Crisis. J Nutr. 2010 Jan;140(1):182S–188S. doi: 10.3945/jn.109.110809. [DOI] [PubMed] [Google Scholar]
- [47].Corsi DJ, Kyu HH, Subramanian SV. Socioeconomic and Geographic Patterning of Under- and Overnutrition among Women in Bangladesh. J Nutr. 2011 Apr;141(4):631–638. doi: 10.3945/jn.110.131599. [DOI] [PubMed] [Google Scholar]
- [48].Coats AJS, Shewan LG. Statement on Authorship and Publishing Ethics in the International Journal of Cardiology. Int J Cardiol. 2011;153:239–40. doi: 10.1016/j.ijcard.2011.10.119. [DOI] [PubMed] [Google Scholar]