当前位置: 首页 > 期刊 > 《美国呼吸和危急护理医学》 > 2006年第3期 > 正文
编号:11417431
Prospective Study of Dietary Patterns and Persistent Cough with Phlegm among Chinese Singaporeans
http://www.100md.com 《美国呼吸和危急护理医学》
     Division of Epidemiology, Department of Public Health Sciences, University of California, Davis, California

    Department of Community, Occupational, and Family Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore

    Division of Population Science, Fox Chase Cancer Center, Philadelphia, Pennsylvania

    University of Minnesota Cancer Center, Minneapolis, Minnesota

    Epidemiology Branch, National Institute of Environmental Health Sciences, Department of Health and Human Services, National Institutes of Health, Research Triangle Park, North Carolina

    Correspondence and request for reprints should be addressed to Stephanie London, M.D., Dr.P.H., National Institute of Environmental Health Sciences, P.O. Box 12233, Mail Drop A3-05, Research Triangle Park, NC 27709. E-mail: london2@niehs.nih.gov

    ABSTRACT

    Rationale: Using principal components analysis to examine dietary patterns complements the evaluation of individual food and nutrient intake in relation to health outcomes, but has not yet been applied to nonmalignant respiratory disease or symptoms. Objective: To examine the relation between patterns of dietary intake at baseline and new onset of persistent cough with phlegm in a population-based cohort of Singapore Chinese. Methods: A 165-item validated food frequency questionnaire was administered in-person at baseline in 1993. We identified 623 cases of incident cough with phlegm among 52,325 subjects by telephone interview from 1999 through 2004. We identified two distinct food patterns: a "meat-dim sum" pattern characterized by pork and chicken dim sum foods and noodle dishes, and a "vegetable-fruit-soy" pattern characterized by vegetables, fruit, and soyfood items. Main Results: The meat-dim sum pattern was positively associated with new-onset cough with phlegm (odds ratio, 1.43; 95% confidence interval, 1.08, 1.89; comparing fourth to first quartile, p for trend = 0.02), after adjustment for age, sex, total energy intake, smoking, education, and nonstarch polysaccharide intake, a protective factor for cough with phlegm in this cohort. Weaker associations were seen for more chronic symptoms and for incident asthma. A weak inverse association for the vegetable-fruit-soy pattern disappeared after adjustment for nonstarch polysaccharide intake. Conclusion: A diet rich in meats, sodium, and refined carbohydrates may increase risk of developing cough with phlegm, independently of the apparent beneficial effects of a diet high in fiber in this Singapore Chinese cohort.

    Key Words: asthma chronic bronchitis chronic obstructive pulmonary disease diet respiratory, signs and symptoms

    Although smoking is the major risk factor, diet contributes to the pathogenesis of nonmalignant lung disease and symptoms (1–3). Dietary nutrients may modulate oxidative stress–induced lung damage among both smokers and nonsmokers, although prospective data are few (4–7). We previously identified reduced risk of cough with phlegm, the symptoms of chronic bronchitis, with higher nonstarch polysaccharide intake in a cohort of Chinese Singaporeans (8).

    Epidemiologic analyses that focus on individual nutrients and food items in relation to chronic disease, while valuable, have limitations (9). When nutrients of interest correlate strongly, teasing out individual effects can be difficult. Furthermore, the single-nutrient approach cannot capture the complex interactions between individual nutrients, and their correlations with other dietary (10, 11), lifestyle (12, 13), and sociodemographic patterns (14, 15) that may confound associations with health outcomes. Principal components analysis is a dimension-reduction statistical technique that has been used to define dietary patterns in epidemiologic studies (16). The high level of intercorrelation of individual nutrients can impede traditional analyses, but aids in the identification of dietary patterns in principal components analysis. Looking beyond individual nutrients or foods to identify overall patterns of diet that reflect actual eating habits can better inform public health dietary recommendations.

    Dietary patterns have been associated with cancer (17) and heart disease (18, 19), but have not been studied in relation to nonmalignant respiratory disease or symptoms. Few data address empirically defined dietary patterns among Asians in the United States (13, 14, 20) or Chinese populations (21). To our knowledge, there are no published data on dietary patterns in a Singaporean Chinese population. The Singaporean Chinese diet differs from Western diets. It consists primarily of mixed dishes combining Chinese, Malay, and Indian influences; is generally high in refined carbohydrates (e.g., noodles and white rice), soyfoods, and green leafy vegetables; and uses relatively small quantities of meats, such as fish, chicken, and pork. It is common to consume meals outside the home from either food courts or dim sum restaurants, where typical foods include meat-filled dumplings, buns, and noodle dishes.

    We examined whether dietary patterns are related to risk of cough with phlegm in the Singapore Chinese Health Study. Some of the results of these analyses have been previously reported in the form of an oral presentation at the American Thoracic Society 2005 conference (22).

    METHODS

    Study Population

    The design of the Singapore Chinese Health Study has been previously described (23). Briefly, the cohort was drawn from men and women, aged 45 to 74 yr, who belonged to one of the major dialect groups (Hokkien or Cantonese) of Chinese in Singapore. Between April 1993 and December 1998, 63,257 individuals completed an in-person interview that included questions on diet and smoking. The institutional review boards at the University of Southern California, the National University of Singapore, and the National Institute of Environmental Health Sciences approved this study.

    Identification of Respiratory Symptoms

    Beginning in July 1999, a follow-up telephone interview was administered that included standardized questions on history of asthma and habitual cough and phlegm production (24). Of the baseline cohort, 52,325 completed the follow-up interview. As of December 31, 2004, 7,722 subjects had died. The average time from baseline to follow-up interview was 5.3 yr (range, 2–11 yr). Incident cases were subjects who reported that they usually had cough with phlegm on arising or during the rest of the day, which initiated at least 1 yr after the baseline interview. Subjects were divided into mutually exclusive categories on the basis of cough and/or phlegm, and their incidence or prevalence, with a joint reference category of noncases (n = 46,942) without cough or phlegm. Six hundred and twenty-three subjects had incident cough with phlegm, and of those, 380 had both symptoms for 3 or more months of the year. The remaining 5,383 individuals had either incident cough only (n = 549), incident phlegm only (n = 1,349), prevalent cough with phlegm (n = 727), or discordant incident/prevalent combinations of cough and/or phlegm (n = 2,132). Analyses of the 5,383 subjects with these secondary outcomes can be found in Table E1 in the online supplement.

    Subjects reporting asthma were considered to be incident (n = 406) if their reported age at onset of asthma on the follow-up questionnaire was older than the age at baseline interview. Among the 406 subjects, 35 also reported incident cough with phlegm. We completed validation home visits on 331 of the 406 (81.5%) subjects with incident asthma where we asked again about their history of asthma, age at onset, current symptoms, and medication use. Among the 331 respondents, 270 (81.6%) subjects confirmed asthma, 52 (15.7%) confirmed asthma but gave age at onset consistent with prevalent asthma, and nine (2.7%) did not confirm asthma. Further details on the asthma validation are in the online supplement.

    Identification of Dietary Patterns

    At baseline, a 165-item quantitative food frequency questionnaire, developed for and validated in this population, was administered to assess usual diet over the past year (23, 25). Using principal components analysis (16, 26) among the baseline cohort (n = 63,257), we identified patterns from the food frequency responses. Extraction of principal components was followed by orthogonal rotation. The number of components retained for rotation was based primarily on examination of scree plots and factor interpretability, but eigenvalues (> 1.0) and percentage of variance explained were also considered (16).

    For each dietary pattern, a component score was computed as a linear composite of the foods with meaningful loading scores (e.g., 0.30). Scores were calculated by taking the unweighted sum of standardized frequencies of intake for each food associated with the pattern, then dividing them into quartiles based on the distribution of the baseline cohort. Principal components analyses were conducted using the Factor Procedure in SAS version 9 (SAS Institute, Cary, NC). Examples of the SAS programming code used have been previously cited (17) and are available online (27).

    The sensitivity and reproducibility of the identified dietary patterns were also assessed. These methods can be found in the online supplement.

    Statistical Analysis

    The following variables were included in all models because these factors were either related both to the outcome and to the dietary patterns or because there was compelling evidence from the literature for possible confounding (in the case of smoking and environmental tobacco smoke [ETS]): age at baseline (continuous), sex, dialect group (Hokkien, Cantonese), total energy intake (continuous logarithm), highest education level (no formal education, primary school, secondary school or beyond), energy-adjusted nonstarch polysaccharide intake (quartiles) (8), smoking status (never, past, current), daily amount smoked (never, past, current and 12 cigarettes, current and 13–22 cigarettes, current and 23 cigarettes), age at starting smoking (never, past, current and 20 yr, current and 15–19 yr, current and 14 yr), and adult ETS exposure at home or work (ever, never). All nutrient data were adjusted for total energy intake using the residual method (28). Individuals were categorized as never active smokers if they had never smoked at least one cigarette a day for 1 yr or longer. Further information on the evaluation of confounding, including variables evaluated for confounding but not retained, is available in the online supplement.

    RESULTS

    We identified two distinct dietary patterns from the baseline cohort using principal components analysis: "vegetable-fruit-soy" and "meat-dim sum" patterns. The individual food items and their loading factor values are listed in Table 1. The higher the loading of a given food item, the greater the contribution of that food to the pattern. The vegetable-fruit-soy pattern was characterized by vegetable, fruit, and soyfood intake; of the 32 foods included in the pattern, 23 were vegetables, five were soyfood items, and four were fruit items. The meat-dim sum pattern contained 31 food items, predominantly chicken, pork, fish, rice and noodle dishes, and preserved foods. Eleven of the 19 dim sum or snack items on the questionnaire were included in this pattern. No food or beverage items overlapped between the two patterns.

    Distributions for selected baseline sociodemographic and smoking characteristics are presented for the quartiles of each dietary pattern among subjects with follow-up in Table 2. Individuals in the highest quartile of vegetable-fruit-soy scores were less likely to be of the Hokkien dialect group, to smoke, and to lack formal education than those in the lowest quartile. Individuals with the highest meat-dim sum scores were more likely to be younger, male, current smokers, and have had formal education compared with the lowest quartile of the meat-dim sum scores. Mean body mass index was similar across all quartiles of both dietary patterns, and body mass index was not correlated with either the vegetable-fruit-soy pattern (r = 0.003) or the meat-dim sum pattern (r = 0.004).

    We present correlations between each dietary pattern and selected food items, food groups, and nutrients of interest among subjects with follow-up in Table 3. The strongest positive correlations with the vegetable-fruit-soy pattern were for nonstarch polysaccharides, vitamin E, total carotenoids, total fruit, vegetables, and soyfoods. The strongest positive correlations with the meat-dim sum pattern were for total energy, sodium, total red meat, preserved red meat, preserved fish/shellfish, and fresh eggs. Total dietary fat was similarly correlated with each pattern, but correlations among fat subtypes varied. For example, the omega-3 and omega-6 fatty acids were moderately correlated with the vegetable-fruit-soy pattern, but not with the meat-dim sum pattern, whereas saturated fat intake was moderately associated with the meat-dim sum pattern, but not with the vegetable-fruit-soy pattern.

    We assessed risk of cough with phlegm by quartiles of component scores of each dietary pattern (Table 4). Cough with phlegm was associated with the meat-dim sum pattern, with a statistically significant trend by quartile. There was a weak inverse association with the vegetable-fruit-soy pattern. The association with the meat-dim sum patterns was attenuated and lost statistical significance when restricted to the smaller number of subjects with more chronic symptoms. The positive association between cough and phlegm and the meat-dim sum pattern is independent of our previously reported inverse association with nonstarch polysaccharides (8). This was expected given the weak correlation between these two dietary factors (r = –0.09). In contrast, the weak inverse association with the vegetable-fruit-soy pattern was eliminated by adjusting for the more highly correlated (r = 0.49) nonstarch polysaccharide intake. The adjusted odds ratios (ORs) for the fourth versus first quartiles of nonstarch polysaccharide intake were not appreciably altered by inclusion of the meat-dim sum pattern in the model (OR, 0.61; 95% confidence interval [CI], 0.48, 0.79, with meat-dim sum pattern, vs. OR, 0.59; 95% CI, 0.47, 0.76, without meat-dim sum pattern).

    The association between the meat-dim sum pattern and cough with phlegm was similar in nonsmokers (OR, 1.41; 95% CI, 0.96, 2.05, for fourth compared with first quartile; p for trend = 0.05) and ever smokers (OR, 1.56; 95% CI, 1.01, 2.40, for fourth compared with first quartile; p for trend = 0.2). The associations between each dietary pattern and cough with phlegm did not vary appreciably by sex, dialect group (Hokkien, Cantonese), age (at the median), body mass index (< 25, 25), or education (no formal education, primary school, secondary or greater) (data not shown).

    To better understand the association between cough with phlegm and the meat-dim sum pattern, we explored individual foods, food groups, and nutrients that were highly correlated with the pattern. Of the five individual foods with the highest loading factors (siew mai, other steamed snack, gravy noodle, chicken rice, otar otar [spiced fish paste]; Table 1), only siew mai had a statistically significant positive association with the symptoms (OR, 1.34; 95% CI, 1.01, 1.77, comparing 2.5 servings/mo vs. none), but this was attenuated (OR, 1.23; 95% CI, 0.92, 1.66) when included in the same model with the meat-dim sum pattern. The OR was unchanged for the meat-dim sum pattern (OR, 1.38; 95% CI, 1.05, 1.84). When we examined food groups and nutrients highly correlated with the meat-dim sum pattern, statistically significant positive association were seen in individual models for preserved red meat (OR, 1.40; 95% CI, 1.07, 1.84), preserved fish and shellfish (OR, 1.38; 95% CI, 1.06, 1.79), and sodium (OR, 1.36; 95% CI, 1.06, 1.75), comparing fourth to first quartile of intake and adjusting for age, sex, dialect, smoking, the meat-dim sum pattern, and nonstarch polysaccharide intake. When sodium intake, preserved fish and shellfish, and preserved red meat were included in the same model with the dietary pattern, all were attenuated and only preserved red meat retained statistical significance (OR, 1.33; 95% CI, 1.01, 1.75, for the top relative to the bottom quartile).

    Although our primary focus was on incident cough with phlegm, which was previously associated with nonstarch polysaccharide intake in this cohort, we also examined other outcomes. For both "all reported" and "confirmed" incident asthma, subjects in the upper three quartiles of the meat-dim pattern were at nonstatistically significantly increased risk without a trend (Table 5). There was no association with the vegetable-fruit-soy pattern, after adjustment for nonstarch polysaccharide intake. Although numbers become small on stratification among confirmed asthma, there were similar findings for the fourth relative to the first quartile among nonsmokers (OR, 1.23; 95% CI, 0.74, 2.05) and among ever smokers (OR, 1.43; 95% CI, 0.65, 3.14).

    We also examined incident cough without phlegm, incident phlegm without cough, as well as prevalent cough with phlegm (see Table E1). The meat-dim sum pattern was more weakly associated with prevalent than with incident cough and phlegm, and the vegetable-fruit-soy pattern was weakly inversely associated with incident phlegm alone.

    DISCUSSION

    To our knowledge, this is the first report of analyses of dietary patterns in relation to nonmalignant respiratory disease or symptoms. We identified two distinct dietary patterns in a large cohort of adult Chinese Singaporeans. The highest quartile of the meat-dim sum pattern was associated with 1.4-fold increase in risk of cough with phlegm. The association reported for meat-dim sum pattern was not attributed to individual foods that were the major contributors to the pattern. However, we cannot exclude the possibility that this pattern may be due in part to correlated individual nutrients, such as sodium, or food groups, such as preserved meats or fish. Our previously reported inverse association between nonstarch polysaccharide intake and cough with phlegm (8) was not changed by adjustment for the meat-dim sum pattern. The vegetable-fruit-soy pattern was not associated with risk of cough and phlegm when adjusted for nonstarch polysaccharide intake.

    There are similarities between the newly identified dietary patterns in our Chinese Singaporean population-based cohort and those previously identified in U.S. populations. For example, in various U.S. populations, two primary patterns have been consistently described (29–31), the "Western" pattern, characterized by red and processed meats, sweets and desserts, French fries, and refined grains, and the "prudent" pattern, characterized by fruits, vegetables, legumes, fish, poultry, and whole grains (32). The meat-dim sum pattern, identified in our cohort, is also characterized by red meat intake, preserved foods, such as red meat and eggs, and rice and noodle-based dishes and desserts. Deep-fried foods are also common dim sum items. However, the U.S. prudent pattern and the Singaporean vegetable-fruit-soy pattern have fewer similarities, in that there is no contribution of fish, poultry, or whole grains to the Singapore pattern, and vegetables contribute more than fruits. Whole grain intake is exceedingly rare in Singapore, as indicated in a study of food availability and sources of fiber intake (33).

    Prior literature on empirically derived dietary patterns and lung disease or symptoms is sparse. In a prospective study of dietary patterns and lung cancer in men in the Netherlands, a modest positive association was seen for a dietary pattern characterized by pork, processed meat, and potatoes, and weak inverse associations were observed for two dietary patterns characterized by vegetable intake (34). However, dietary patterns identified in U.S. populations have been associated with biologic markers for systemic inflammation (35), increased insulin (29, 36), and carcinogen metabolism (20), which may be relevant for lung disease and symptoms.

    Our finding that a dietary pattern characterized by meat and starchy food intake is related to respiratory symptoms in Singapore Chinese is novel. This pattern is notable for the number of starch-laden dishes that had high loading for the pattern, particularly the noodle and rice dishes. These starchy items have a high glycemic index and thus may increase insulin resistance over time. Given that both hyperglycemia and chronic obstructive pulmonary disease, which is frequently associated with cough and phlegm, are related to impaired lung function (37, 38), oxidative stress (39, 40), and inflammation (41, 42), it is possible that they may share pathophysiologic mechanisms (43).

    The meat-dim sum pattern is also highly correlated with dietary sodium. Many popular Chinese dim sum dishes are high in sodium (44). The major sources of added sodium in the diet in our population are from sauces and condiments and soups, rather than table salt (45). Sodium from these sources was captured in our food frequency questionnaire (23). Although findings from principal component analyses do not allow for drawing conclusions about single-nutrient–disease associations, there is some evidence that dietary sodium has effects on asthma and airway hyperreactivity (reviewed in Reference 1). Relevant to our study, symptoms of bronchitis (46) have also been reported to be affected by sodium intake.

    In addition to sodium, preserved meat and preserved fish/shellfish were also highly correlated with the meat-dim sum pattern as well as with each other (r 0.33). Preserved meat and fish are high in sodium but also contain other compounds, such as nitrosamines, that may be relevant. It is difficult to tease out the individual effects of these correlated foods; modeling these foods and the pattern together attenuates all of them. The meat-dim sum pattern offers the advantage of summarizing the effects of these correlated foods.

    An alternative explanation for the positive association between cough with phlegm and the meat-dim sum pattern is confounding by lifestyle and socioeconomic factors (47–51). We addressed this, to the extent possible, by evaluating confounding by detailed smoking variables, ETS, occupation, education, body mass index, and dialect group. Positive associations were similar in men and women, in lean and overweight individuals, and in more and less educated individuals. Of note, the association with the meat-dim sum pattern was present both in nonsmokers and smokers, suggesting that it is not due to residual confounding by smoking.

    The other dietary pattern in this population, characterized by vegetable, fruit, and soyfood intake, was more weakly associated with cough and phlegm and this association disappeared after adjustment for nonstarch polysaccharide intake, with which it is highly correlated. The weak association between symptoms and this dietary pattern is not unexpected because vegetables, the major contributor to the pattern, are not related to cough and phlegm in these (7) or other data (1, 2). In contrast, the inverse association with nonstarch polysaccharides in our data appears to be driven by fruits and, to a lesser degree, soyfoods (7).

    The OR for the meat-dim sum pattern was attenuated and lost statistical significance when restricted to the smaller number of subjects with more chronic symptoms. This is similar to what we observed in some previous dietary analyses in this cohort (8). In addition to the reduced power with a smaller sample size, it is possible that diet, which may change over time, has a greater impact on recently occurring symptoms, rather than longer term symptoms. Last, long-term symptoms could lead to dietary changes, obscuring etiologic associations.

    Limitations of principal components analysis include the subjective nature of determining the number of patterns, labeling the patterns, and interpreting these patterns (52). Reduced rank regression has been suggested to be superior to principal components analysis (53). However, this was not an option for our analyses, because reduced rank regression requires a priori knowledge of dietary factors associated with the outcome of interest. Because there is no clear picture of the underlying biological mechanism or a set of dietary guidelines for preventing chronic lung disease, as there is for cardiovascular disease (54), we chose to use an a posteriori method.

    We entered individual foods and beverages, without grouping them, into the principal components analysis. One rationale for grouping foods before conducting principal components analysis has been to limit influences of within-person variability in individual food intake (55). However, we conducted sensitivity analyses and found a high degree of internal consistency and reproducibility with our patterns. Another reason we chose to assess individual foods and beverages is that a large proportion of items in our questionnaire are mixed dishes (23), which do not lend themselves to a priori grouping. This is in contrast to dietary questionnaires commonly used in the United States in which the vast majority of questionnaire items are single foods.

    In this first study of dietary patterns in relation to nonmalignant respiratory disease or symptoms, we identified two primary dietary patterns in the Chinese Singaporean population. One of the patterns, characterized by meat intake, meat-containing noodle, rice, and savory mixed dishes, was associated with an increased risk of developing cough with phlegm. The positive association with the meat-dim sum pattern was independent of nonstarch polysaccharide intake, which we previously identified as a protective factor for cough with phlegm in this cohort. The other pattern, characterized by vegetable, fruit, and soyfood intake was not independently associated with cough and phlegm after adjustment for nonstarch polysaccharide intake. Although reducing cigarette smoke exposure remains the most important preventative measure for respiratory disease, these data suggest that there may be deleterious dietary patterns independent of smoking. For the Singapore population, dietary recommendations based on our analysis of dietary patterns might include consuming fewer dim sum meals or choosing more healthful dim sum items, with a lower glycemic index and saturated fat and sodium content. Although studies in other populations would be required, consuming foods with a lower glycemic index and saturated fat and sodium content may also be a healthful modification to Western diets.

    Acknowledgments

    The authors thank Siew-Hong Low of the National University of Singapore for supervising the field work of the Singapore Chinese Health Study; Kazuko Arakawa of the University of Southern California for the development and management of the cohort study database; and Marsha Shepherd of Westat, Inc., for analysis and programming.

    FOOTNOTES

    Supported by the Division of Intramural Research, National Institute of Environmental Health Sciences (ZO1 ES43012), and the National Cancer Institute (RO1 CA80205).

    This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org

    Originally Published in Press as DOI: 10.1164/rccm.200506-901OC on November 4, 2005

    Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.

    REFERENCES

    Romieu I, Trenga C. Diet and obstructive lung diseases. Epidemiol Rev 2001;23:268–287.

    Hu G, Cassano PA. Antioxidant nutrients and pulmonary function: the Third National Health and Nutrition Examination Survey (NHANES III). Am J Epidemiol 2000;151:975–981.

    Schunemann H, Freudenheim J, Grant B. Epidemiologic evidence linking antioxidant vitamins to pulmonary function and airway obstruction. Epidemiol Rev 2001;23:248–267.

    Smit HA. Chronic obstructive pulmonary disease, asthma and protective effects of food intake: from hypothesis to evidence Respir Res 2001; 2:261–264.

    Burney P. The origins of obstructive airways disease: a role for diet Am J Respir Crit Care Med 1995;151:1292–1293.

    MacNee W. Oxidants/antioxidants and COPD. Chest 2000;117:303S–317S.

    Altose MD. Approaches to slowing the progression of COPD. Curr Opin Pulm Med 2003;9:125–130.

    Butler LM, Koh WP, Lee HP, Yu MC, London SJ. Dietary fiber and reduced cough with phlegm: a cohort study in Singapore. Am J Respir Crit Care Med 2004;170:279–287.

    Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol 2002;13:3–9.

    Randall E, Marshall JR, Graham S, Brasure J. Patterns in food use and their associations with nutrient intakes. Am J Clin Nutr 1990;52:739–745.

    Kant AK, Schatzkin A, Block G, Ziegler RG, Nestle M. Food group intake patterns and associated nutrient profiles of the US population. J Am Diet Assoc 1991;91:1532–1537.

    Slattery ML, Edwards SL, Boucher KM, Anderson K, Caan BJ. Lifestyle and colon cancer: an assessment of factors associated with risk. Am J Epidemiol 1999;150:869–877.

    Maskarinec G, Novotny R, Tasaki K. Dietary patterns are associated with body mass index in multiethnic women. J Nutr 2000;130:3068–3072.

    Lv N, Cason KL. Dietary pattern change and acculturation of Chinese Americans in Pennsylvania. J Am Diet Assoc 2004;104:771–778.

    Gex-Fabry M, Raymond L, Jeanneret O. Multivariate analysis of dietary patterns in 939 Swiss adults: sociodemographic parameters and alcohol consumption profiles. Int J Epidemiol 1988;17:548–555.

    Hatcher L. A step-by-step approach to using SAS for factor analysis and structural equation modeling. Cary, NC: SAS Institute, Inc.; 1994.

    Tseng M, Breslow RA, DeVellis RF, Ziegler RG. Dietary patterns and prostate cancer risk in the National Health and Nutrition Examination Survey Epidemiological Follow-up Study cohort. Cancer Epidemiol Biomarkers Prev 2004;13:71–77.

    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;161:1857–1862.

    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;72:912–921.

    Tsai YY, McGlynn KA, Hu Y, Cassidy AB, Arnold J, Engstrom PF, Buetow KH. Genetic susceptibility and dietary patterns in lung cancer. Lung Cancer 2003;41:269–281.

    Gao X, Yao M, McCrory MA, Ma G, Li Y, Roberts SB, Tucker KL. Dietary pattern is associated with homocysteine and B vitamin status in an urban Chinese population. J Nutr 2003;133:3636–3642.

    Butler L, Koh WP, Lee HP, Tseng M, Yu MC, London SJ. Dietary patterns and risk of cough with phlegm . Proc Am Thorac Soc 2005;2:A848.

    Hankin JH, Stram DO, Arakawa K, Park S, Low SH, Lee HP, Yu MC. Singapore Chinese Health Study: development, validation, and calibration of the quantitative food frequency questionnaire. Nutr Cancer 2001;39:187–195.

    Ferris BG. Epidemiology Standardization Project (American Thoracic Society). Am Rev Respir Dis 1978;118:1–120.

    Sun CL, Yuan JM, Arakawa K, Low SH, Lee HP, Yu MC. Dietary soy and increased risk of bladder cancer: the Singapore Chinese Health Study. Cancer Epidemiol Biomarkers Prev 2002;11:1674–1677.

    Kline P. An easy guide to factor analysis. New York: Routledge; 1994.

    Tseng M. Basic program to use principal components analysis to identify dietary patterns [Internet; accessed April 28, 2005]. Available from: http://www.fccc.edu/research/labs/tseng/TsengDOD01.html

    Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986;124:17–27.

    Hu FB, Rimm E, Smith-Warner SA, Feskanich D, Stampfer MJ, Ascherio A, Sampson L, Willett WC. Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am J Clin Nutr 1999;69:243–249.

    Slattery ML, Boucher KM, Caan BJ, Potter JD, Ma KN. Eating patterns and risk of colon cancer. Am J Epidemiol 1998;148:4–16.

    Tseng M, DeVellis RF. Fundamental dietary patterns and their correlates among US whites. J Am Diet Assoc 2001;101:929–932.

    Fung T, Hu FB, Fuchs C, Giovannucci E, Hunter DJ, Stampfer MJ, Colditz GA, Willett WC. Major dietary patterns and the risk of colorectal cancer in women. Arch Intern Med 2003;163:309–314.

    Lee HP, Gourley L. Fat and fibre in the Singapore diet. Ann Acad Med Singapore 1987;16:408–411.

    Balder HF, Goldbohm RA, van den Brandt PA. Dietary patterns associated with male lung cancer risk in the Netherlands Cohort Study. Cancer Epidemiol Biomarkers Prev 2005;14:483–490.

    Lopez-Garcia E, Schulze MB, Fung TT, Meigs JB, Rifai N, Manson JE, Hu FB. Major dietary patterns are related to plasma concentrations of markers of inflammation and endothelial dysfunction. Am J Clin Nutr 2004;80:1029–1035.

    Fung TT, Rimm EB, Spiegelman D, Rifai N, Tofler GH, Willett WC, Hu FB. Association between dietary patterns and plasma biomarkers of obesity and cardiovascular disease risk. Am J Clin Nutr 2001;73:61–67.

    Davis WA, Knuiman M, Kendall P, Grange V, Davis TM. Glycemic exposure is associated with reduced pulmonary function in type 2 diabetes: the Fremantle Diabetes Study. Diabetes Care 2004;27:752–757.

    Walter RE, Beiser A, Givelber RJ, O'Connor GT, Gottlieb DJ. Association between glycemic state and lung function: the Framingham Heart Study. Am J Respir Crit Care Med 2003;167:911–916.

    Barnes PJ. Mediators of chronic obstructive pulmonary disease. Pharmacol Rev 2004;56:515–548.

    Oberley LW. Free radicals and diabetes. Free Radic Biol Med 1988;5:113–124.

    Moller DE. Potential role of TNF-alpha in the pathogenesis of insulin resistance and type 2 diabetes. Trends Endocrinol Metab 2000;11:212–217.

    Pinto-Plata VM, Mullerova H, Toso JF, Feudjo-Tepie M, Soriano JB, Vessey RS, Celli BR. C-reactive protein in patients with COPD, control smokers, and nonsmokers. Thorax 2005.

    Funk GC, Doberer D, Petkov V, Block LH. Hyperglycemia, bronchial artery sclerosis, and lung function. Am J Respir Crit Care Med 2004; 169:427. [Author reply].

    Food and Public Health Branch of the Food and Environment Hygiene Department. Nutrient values of Chinese dim sum: risk assessment studies. Hong Kong: The Government of the Hong Kong Special Administrative Region; April 2005. Report No. 17. Available from: http://www.fehd.gov.hk/safefood/report/dim_sum/dim_sum_index.html

    Lee HP, Ong CN, Chia CB. An estimation of sodium chloride intake by adults in Singapore. Singapore Med J 1983;24:346–349.

    Schwartz J, Weiss ST. Dietary factors and their relation to respiratory symptoms. The Second National Health and Nutrition Examination Survey. Am J Epidemiol 1990;132:67–76.

    Lawlor DA, Davey Smith G, Bruckdorfer KR, Kundu D, Ebrahim S. Observational versus randomised trial evidence. Lancet 2004;364:755.

    Lawlor DA, Davey Smith G, Kundu D, Bruckdorfer KR, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence Lancet 2004;363:1724–1727.

    Concato J, Horwitz RI. Beyond randomised versus observational studies. Lancet 2004;363:1660–1661.

    Khaw KT, Day N, Bingham S, Wareham N. Observational versus randomised trial evidence. Lancet 2004;364:753–754; author reply 754–755.

    Brotman DJ. Observational versus randomised trial evidence. Lancet 2004;364:757.

    Martinez ME, Marshall JR, Sechrest L. Invited commentary: factor analysis and the search for objectivity. Am J Epidemiol 1998;148:17–19.

    Hoffmann K, Schulze MB, Schienkiewitz A, Nothlings U, Boeing H. Application of a new statistical method to derive dietary patterns in nutritional epidemiology. Am J Epidemiol 2004;159:935–944.

    Kroke A. Re: application of a new statistical method to derive dietary patterns in nutritional epidemiology. Am J Epidemiol 2004;160:1132. [Author reply, 1132–1133.]

    Balder HF, Virtanen M, Brants HA, Krogh V, Dixon LB, Tan F, Mannisto S, Bellocco R, Pietinen P, Wolk A, et al. Common and country-specific dietary patterns in four European cohort studies. J Nutr 2003;133:4246–4251.(Lesley M. Butler, Woon-Puay Koh, Hin-Pen)