AJCN EB Program 2010 Early Registration
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.
Agricola
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.
American Journal of Clinical Nutrition, Vol. 88, No. 2, 289-297, August 2008
© 2008 American Society for Nutrition

ORIGINAL RESEARCH COMMUNICATION

Food patterns and cardiovascular disease risk factors: The Swedish INTERGENE research program1,2,3

Christina M Berg1, Georgios Lappas1, Elisabeth Strandhagen1, Alicja Wolk1, Kjell Torén1, Annika Rosengren1, Nibia Aires1, Dag S Thelle1 and Lauren Lissner1

1 From the Department of Food Health and Environment, University of Gothenburg, Göteborg, Sweden (CMB); the Department of Public Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Göteborg, Sweden (GL, ES, KT, and DST); The Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden (AW); the Department of Medicine, Sahlgrenska University Hospital/Östra, Göteborg, Sweden (AR); Clinical Science, AstraZeneca R&D, Mölndal, Sweden (NA); and the Department of Biostatistitcs, University of Oslo, Oslo, Norway (DST).

2 Supported by grants from the Västra Götaland County Council, the Swedish Council for Working Life and Social Research, the Swedish Research Council, the Swedish Swedish Heart and Lung Foundation, and the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning.

3 Reprints not available. Address correspondence to CMB, Department of Food, Health and Environment, Göteborg University, Box 320, SE 405 30 Göteborg, Sweden. E-mail: christina.berg{at}ped.gu.se.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background:Analyzing the impact of the intake of many foods simultaneously provides additional knowledge about analyses of nutrients and might make it easier to implement recommendations for the public.

Objective:The objective was to examine food patterns in a Swedish population and determine how they are related to metabolic risk factors for cardiovascular disease.

Design:The study is based on data from the INTERGENE population study of women and men aged 25–74 y in western Sweden. Dietary patterns were identified with cluster analysis of 93 food frequencies reported by 3452 participants. Associations with features of the metabolic syndrome, including blood lipids, blood pressure, and anthropometric measures, were analyzed.

Results:Five distinct food patterns were identified, of which one was interpreted as a "healthy" reference pattern. This healthy cluster was distinguished by more frequent consumption of high-fiber and low-fat foods and lower consumption of products rich in fat and sugar. The 4 other clusters differed significantly from the reference cluster with respect to prevalence of cardiovascular disease risk factors and the metabolic syndrome. For example, body mass index and waist-to-hip ratio were significantly higher in a cluster characterized by high consumption of energy-dense drinks and white bread and low consumption of fruit and vegetables (P < 0.0001 and P = 0.004, respectively).

Conclusions:It is possible to distinguish food patterns that are related to obesity and obesity-related cardiovascular disease risk factors in contrast with a more healthy pattern conforming with current dietary guidelines. Thus, the results indicate no reason for questioning the current recommendations.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Dietary factors are important in the development of obesity and cardiovascular disease (CVD) (1). However, many issues remain unsettled, and we need more knowledge about optimal intake of nutrients and foods and how the intake varies between individuals. One reason for the divergent conclusions in epidemiologic studies is that different dietary factors covary and are related to other lifestyle factors. In experimental studies, matters can be simplified by focusing on one or a few components in diet, predominantly nutrients, but in reality dietary intake is more complex because the diet consists of many foods, which in turn consist of many nutrients. Analysis of combinations of foods makes it possible to consider the known and unknown interactions and potential cumulative effects of dietary exposure. Thus, a food pattern may be a stronger predictor of disease risk when many dietary associations exist for the disease, as for CVD (2), and can provide complementary information to analyses of associations with single food components. Furthermore, results from such analyses might be easier to implement in recommendations to the public. Several studies have suggested that certain dietary patterns predict CVD or mortality. The Lyon Diet Heart Study (3) demonstrated that a certain diet reduced risk recurrence after a first myocardial infarction over a period of 4 y. In another randomized intervention study (4), a diet rich in fruit, vegetables, and low-fat dairy foods reduced blood pressure considerably. Still, most evidence for an association between food pattern and CVD comes from observational studies. Common statistical methodologies for assessing dietary patterns are cluster or factor analyses (5). Cluster analysis, in contrast with factor analysis, generally defines nonoverlapping groups of subjects with specific dietary behaviors. Therefore, cluster data can be related to other variables in an interpretable way and might indicate the prevalence of more or less healthy dietary patterns in a specific population. Cluster analyses based on intake of energy and a few nutrients have demonstrated higher risk of future angina pectoris, myocardial infarction (6), and CVD mortality (7) with specific dietary patterns. However, in these studies, the dietary clusters that predicted a lower risk of CVD were not the most healthy according to current nutrition recommendations, eg, did not have the lowest intake of saturated fat. In another prospective study that clustered food intake variables, a heart-healthy food pattern was, however, associated with a lower risk of atherosclerosis (8). Furthermore, factor analyses have revealed that a "healthy" pattern is protective against coronary heart disease (9–11) and CVD mortality (12). Thus, CVD morbidity and mortality have been observed to be associated with food patterns in several studies, but a universal protective dietary pattern could not be distinguished.

Previous research from Sweden has revealed significant associations between food patterns, defined using cluster analysis, and components of the metabolic syndrome (13). In that study Wirfält et al defined clusters by percentage of total energy from specific food groups. Because certain foods with a relatively low contribution of energy (eg, vegetables) are likely to be important in healthy food patterns, in the present study we examined food patterns based on intake frequencies rather than on energy. This is consistent with the work of Bailey et al (14), who demonstrated that a cluster analysis using number of servings better identifies a more healthful eating pattern than does an analysis using the percentage of energy from foods groups as the clustering variable. Thus, the aim of the present study was to study dietary patterns associated with CVD risk by clustering the subjects into distinct groups representing different food intake patterns and to examine how these patterns relate to metabolic and anthropometric characteristics. A further aim was to describe the clusters with respect to socioeconomic, lifestyle, and health characteristics.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Population
INTERGENE is a population based research program that assessed the INTERplay between GENEtic susceptibility and environmental factors for the risk of chronic diseases in western Sweden. The survey started in April 2001 and continued until December 2004. The study population consisted of randomly selected women and men aged 25–74 y and living in the Västra Götaland Region at the time of sampling. For practical reasons, data were collected at screening locations placed in the most densely populated areas; hence, a few small communities were excluded from the sampling. Altogether, the sample consisted of 8625 eligible subjects. Of the invited cohort, 3610 responded: 1908 women (44%) and 1702 men (39%). The mean age of the women was 51.2 y and of the men was 51.6 y. For the purpose of this study, pregnant women (n = 16) were excluded. The participant characteristics and the study procedure are described in detail in an earlier report (15) and on the Internet at http://www.sahlgrenska.gu.se/intergene/. The study was approved by the local ethical committee.

Data collection
The invitees were asked to abstain from eating during the 4 h before the clinical examination. Body height and weight were measured to the nearest 1 cm and 0.1 kg, respectively, while the subjects were wearing light clothing and no shoes. Waist circumference was measured at a level midway between the lower rib margin and iliac crest, and hip circumference was measured as the maximum perimeter over the buttocks. Blood pressure was measured twice in each person, after a 2-min rest, with a validated (16) automatic device (Omron 711 Automatic IS; Omron Healthcare Inc, Vernon Hills, IL) while the subjects were in a supine position. Blood samples were collected after 4-hour fasting, for immediate serum lipid (total cholesterol, HDL cholesterol, and triglycerides) and glucose analysis. Serum total cholesterol and triglyceride concentrations were determined with enzymatic assays. Serum HDL-cholesterol concentrations were measured after dextran sulfate-magnesium precipitation of apolipoprotein B–containing lipoproteins. Serum glucose was analyzed with a hexokinase method (Roche Hitachi 917 and Roche ModularP, Roche, Basel, Switzerland). All samples were identified with bar codes. Nine subjects who had not fasted according to instructions were excluded from the glucose and triglyceride analyses.

The metabolic syndrome was defined by National Cholesterol Education Program, Adult Treatment Panel III (17), ie, the presence of ≥3 of the following 5 components: waist circumference ≥88 cm in women and ≥102 cm in men, serum triglycerides ≥1.7 mmol/L, HDL cholesterol <1.03 mmol/L in men and <1.29 mmol/L in women, blood pressure of 130 mm Hg systolic and/or 85 mm Hg diastolic, fasting plasma glucose ≥6.1 mmol/L.

Usual dietary intake was assessed with a modified version of a validated self-administered food-frequency questionnaire (FFQ) developed at the Karolinska Institute (18–20). It comprised a list of 92 frequency questions, of which 72 had a choice of 8 frequencies ranging from "0 times a month" to "3+ per day." For 20 foods commonly consumed (various milks, coffee, tea, soft drinks, sugar, cheese, and bread), open questions about servings (eg, slices of bread) per day or week were included. Validity of the FFQ-based food intake estimates assessed as the correlation between the FFQ and food records varied between 0.4 (white bread) and 0.9 (sugar); a mean correlation for 20 of the most commonly consumed food items was 0.6. Additional items covered supplement use; choice of fat for cooking, for dressing, and on sandwiches; meal patterns; and portion sizes of cooked food.

The questionnaire also included information about health, socioeconomic, and lifestyle factors. Physical activity during leisure time was categorized into 4 levels with a validated (21) questionnaire that has been in use in Göteborg since the 1960s (22). In addition questions about activity level during working time as well as time spent at household work, biking and walking, sedentary activities, and exercise were included. On the basis of these items and on the intensity levels of different activities (ratio of work metabolic rate to resting metabolic rate) (23), individual physical activity levels (PALs) were estimated.

Statistical analysis
The hypothesis tested was that there is a healthy food pattern in the population and that this pattern is associated with a better risk factor profile. We used cluster analysis to derive food intake patterns. The K-means nonhierarchical method with Euclidean distance was used to categorize individuals in FASTCLUS (SAS package 9.1; SAS Institute Inc, Chicago, IL). Servings or frequencies per week were used as input variables. Subjects with >8 missing food frequency values or with no reported use of common foods reported were excluded. In individuals with fewer missing values, missing data were considered as "no consumption" of the actual food.

Unstandardized variables were used to give more frequently consumed foods more influence on cluster formation. To maximize variation, men and women were analyzed together. Because cluster analysis is sensitive to extreme values, the scale of serving per week (ie, the 20 foods with open answers for frequencies) was truncated to the same endpoint as the scale with 8 categories (ie, ≥21servings/wk). To identify the optimal number of clusters, several runs with varying numbers of clusters were conducted starting with 2 clusters. Each set of clusters was examined to find a solution that was reasonably sized and that included one cluster considered to be healthy. A cluster solution with 5 big clusters included a clear "healthy cluster," whereas the analysis with fewer or more potential clusters included ≥2 clusters that could be interpreted as healthier in some aspects. Therefore, the solution with 5 clusters was selected. Forty individuals were excluded because they belonged to small additional identified clusters. The dietary clusters were interpreted and named on the basis of their high or low intakes of some food groups relative to the population mean intake and to the relative ranking of mean intake of all food items included in the cluster analysis. The clusters were validated by performing a cluster analysis, with a cluster solution with 5 clusters, in a random sample of 50% of the participants. The clusters were similar to those emerging in the original analysis, except for the small cluster ("sweet"), which was principally characterized by high consumption of fruit and vegetables and not sugar.

The "healthy" cluster was considered as a reference cluster against which the other clusters were compared regarding metabolic risk factors. Differences from the reference cluster were assessed by using general linear models for risk factors and logistic regressions for the metabolic syndrome. All of the models were adjusted for age, sex, smoking, and PAL in a first step and these variables plus education in a second step. These analyses were performed with and without participants with self-reported diagnoses associated with the metabolic syndrome to avoid bias due to medication use or to changes in diet associated with diagnoses. Women and men were analyzed separately if sex interactions were present. The statistical analyses were performed by using Statistical Analysis Software 9.1 (SAS Institute Inc, Cary, NC).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Dietary patterns
Five distinct dietary patterns were identified by cluster analysis; "healthy", "sweet", "coffee", "traditional" and "fast energy". The characteristic food intake pattern of each cluster is illustrated in Figure 1Go in comparison with the average intake of selected food groups in the 5 clusters together, ie, the bold ring refers to the population mean and the gray figure refers to the relative consumption in the specific cluster in relation to the average intake of each food group. The most common food items in each cluster are shown in Table 1Go. The "healthy" cluster was characterized by more frequent consumption of high-fiber cereals, fruit and vegetables, low-fat milk products, tea, and fish and less frequent consumption of products rich in fat and sugar, white bread, coffee, beer and spirits. The "sweet" cluster differed from the other clusters by more frequent consumption of sugar, soft drinks, white bread, tea, and sweets and snacks. The "coffee" cluster resembled the whole population even though it was distinguished by frequent consumption of coffee and alcoholic drinks. Distinctive for the "traditional cluster" was consumption of medium-fat milk, offal, boiled coffee, and potatoes and low consumption of low-fat products and alcoholic drinks. Characteristics of the "fast energy" cluster were a frequent consumption of soft drinks, white bread, fast food, full-fat milk, cheese, beer, spirits, sweets and snacks and infrequent consumption of fruit, vegetables, high-fiber products, and low-fat dairy products.


Figure 1
View larger version (25K):
[in this window]
[in a new window]

  		FIGURE 1.. Mean intakes of foods in clusters in relation (%) to the mean intake (bold ring) in the 5 clusters (healthy, sweet, coffee, traditional, and fast energy) combined.

 

View this table:
[in this window]
[in a new window]

  		TABLE 1. Average consumption frequency per week of the most commonly consumed food items (>3 times/portion per week on average), by cluster, on a sliding scale1

 
Even though the clusters were distinguished by specific food patterns, some foods were commonly reported in all clusters, as shown in Table 1Go. Fat cheese, coffee, and bread were generally frequently consumed in the study population. On average, these foods were consumed daily in all clusters, even if the number of servings differed considerably and the choice of bread and coffee varied by group. Whole-meal bread was reported to be commonly consumed and selected rather than white bread in the "healthy," "traditional," and "coffee" clusters, whereas consumption of white bread was more common in the "sweet" and "fast energy" clusters. Milk was another frequently consumed food, with low-fat choices common only in the "healthy" cluster. Vegetables and fruit were more commonly consumed in the "healthy" cluster, but apples, potatoes, and tomatoes were consumed ≥3 times/wk on average in 4 of the clusters. However, the average individual in the "fast energy cluster" ate no fruit or vegetables half of the days in the week, but instead consumed an average of almost 2 glasses of soft drink/d compared with <2 glasses/wk in the healthy cluster. Thus, the "healthy" reference cluster differed considerably from the other clusters in the most commonly consumed foods, with a slightly more varied food choice.

Background characteristics of the clusters
The distribution of demographic, lifestyle, and health characteristics by dietary pattern is shown in Table 2Go. Individuals in the "healthy" cluster were older and the proportion of women was higher than in the other clusters. The reference pattern ("healthy" pattern) was characterized by subjects with a higher education and a more settled financial position, by a lower proportion of smokers, and a higher proportion of subjects who reported regular physical exercise, dietary supplement use, regular consumption of breakfast, and recent changes in diet. The proportion of self-reported CVD diagnoses and treatments was higher in the "healthy" cluster than in the other clusters. The associations between cluster and previous diagnoses were not independent of age.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Background characteristics by cluster

 
Dietary patterns in relation to cardiovascular disease risk factors
The association between dietary cluster and metabolic syndrome were analyzed in all participants and in the "healthy" participants, ie, those with no self-reported diagnoses associated with the metabolic syndrome. Results are presented separately for women and men because interactions between cluster and sex were significant. Compared with the reference cluster, the risk of metabolic syndrome was higher for women and men in the "fast energy" cluster, after adjustment for age, PAL, and smoking (Table 3Go). The "fast energy" cluster had a higher risk of all the components in the syndrome, although not significantly so. Having a "sweet" food pattern was associated with elevated plasma glucose and, in men, with low HDL cholesterol and elevated serum triglycerides.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Prevalences (age adjusted) and odds ratios (ORs) and 95% CIs for the metabolic syndrome and the components of the metabolic syndrome by cluster and stratified by sex1

 
The "healthy" reference cluster was compared with the other clusters with respect to selected metabolic risk factors. Mean values in the "healthy" cluster and significant differences from this cluster are shown in Table 4Go. The participants in the "healthy" cluster had a significantly lower body mass index (BMI) and waist-to-hip ratio (WHR) than did the other participants, ie, the other clusters combined. When each cluster was compared with the reference, the "traditional" and "fast energy" clusters had a significantly higher BMI and WHR than did the "healthy" cluster. Compared with the reference, blood pressure also tended to be higher in the "fast energy" cluster. Furthermore, serum triglycerides were higher in the "fast energy" cluster, and serum HDL concentrations were lower in the "traditional" and "fast energy" clusters than in the "healthy" cluster when all participants were included in the analyses. Some of these associations were not significant after adjustment for BMI and WHR. When subjects with a previous diagnosis were excluded, the same trend as for all participants was observed, even if the other associations became significant for the blood lipids. The analyses in Table 4Go showed similar results when additionally adjusted for education.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Metabolic risk factors by cluster1

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
We identified 5 unique dietary patterns, of which 1 was associated with a better CVD risk factor profile. Variations in healthy clusters were observed in previous studies, and some showed similarities with the present study (24–30), although based on different clustering methods of various dietary data in various populations. Thus, it seems that dietary recommendations that encompass several dimensions of a health-promoting diet (eg, consumption of vegetables, fruit, and fish and low-fat and high-fiber products) are adopted by some persons. Other characteristics of a healthy pattern, which also have been identified in previous research, are the consumption of dietary supplements (24, 28, 31) and the regular consumption of breakfast (32). In addition to identifying divergent food patterns, the results also indicate some general Swedish food patterns. Bread, cheese, coffee, and milk were frequently consumed food groups in all clusters, which indicated that it might be useful to include information on alternatives to these foods in dietary guidelines.

Persons in the "healthy" diet cluster had a higher socioeconomic status, smoked less, and exercised more than did the population as a whole, which indicates a generally healthy lifestyle. On the other hand, they reported more CVD diagnoses and treatments. These diagnoses and treatments were significantly associated with recent dietary changes (data not shown). Medical treatment might improve the risk factor profile, and dietary advice may result in reporting healthier food choices as a matter of what is desired and as a reflection of actual changes. Indeed, some previous studies have observed that a healthier eating pattern is associated with consuming a medically prescribed diet (6, 33), reporting changed food habits (34, 35), or having a diagnosis of cancer, arthritis, heart disease (6), or diabetes (36). Thus, the results indicate that a healthy food pattern is associated with a better risk factor profile, although recognized risk is probably an important reason for adopting these dietary behaviors, despite the fact that half of those in this cluster had made dietary changes during the past 5 y indicating that they recently consumed a less healthy diet.

According to a review by Giugliano et al (37), many diets high in refined starch, sugar, and saturated and trans fat and low in n–3 fatty acids and natural antioxidants and fiber from fruit, vegetables, and whole grains cause the inflammation associated with the metabolic syndrome by affecting the immune system. In the present study, the "fast energy" food pattern was associated with the metabolic syndrome and characterized by a frequent consumption of full-fat milk products and foods rich in refined carbohydrates, and a low consumption of fruit, vegetables, nuts, and whole grains. This cluster was also associated with high consumption of soft drinks which support the association between metabolic syndrome and soft drink consumption observed in a previous study (38).

Although conclusions are inconsistent (39) in studies examining the association between food patterns and anthropometric measures, several studies have found food cluster differences in central or general obesity or anthropometric measures. In studies in which a more healthy cluster can be identified, BMI or WHR tended to be low (although not always the lowest) compared with clusters with less healthy food patterns (6, 14, 25, 26, 28, 29, 31, 33, 36, 40). Even so, many reported associations were not significant, and inverse associations were found in some studies (27, 30, 41). Prospective studies have found a healthy food pattern to be protective against overweight (34) and to prevent increases in BMI and waist circumference (24). In the present study, both BMI and WHR were significantly lower in the "healthy" food cluster than in 2 of the other clusters, which might support the current dietary recommendations for a healthy weight.

In the Framingham Study (25), a "heart healthy" cluster had a lower average systolic blood pressure than did the "empty calorie" and "wine and moderate eating" clusters, but in most studies no significant associations were observed for blood pressure or hypertension (13, 30, 36). The present study found lower blood pressure levels in the healthy cluster than in the "fast energy" cluster. A lower body weight, a lower alcohol consumption, and a higher consumption of fruit and vegetables might contribute to this.

Some studies using cluster analysis have observed associations between food pattern and serum lipids (13, 25, 30, 33, 36, 42, 43). In general, a pattern characterized by high alcohol consumption has been shown to be associated with high serum HDL cholesterol (25, 33, 43, 44). In the present study, mean serum HDL cholesterol as well as alcohol consumption was highest in the "coffee" cluster. Mean HDL cholesterol was also significantly higher in the "healthy" cluster than in the "traditional" and "fast energy" clusters, probably as a result of the higher weight in the 2 latter clusters, as indicated by the results. This also seems to be one explanation for the higher average serum triglyceride concentration in "fast energy" cluster than in the "healthy" reference cluster.

A limitation of the present study was that the examinations were not standardized to the morning or fasting overnight. That the blood samples were drawn at different times of the day probably affected the plasma glucose and serum triglycerides values. That the fasting duration was only ≥4 h might also have influenced the results, but probably only to a limited extent (45, 46). Furthermore, overreporting of foods considered to be healthy and underreporting of other foods may selectively misclassify individual to clusters. Such biases might be associated with nutrition knowledge, which in turn might be a result of diagnosed risk factors. The low response rate might also have biased the picture of the food patterns in the population. It is less likely that the associations observed are a result of a selection bias. Nonparticipation may affect the results if it is associated with CVD risk and eating patterns, eg, if obese individuals with a healthy food pattern are overrepresented among the dropouts.

We decided to base the cluster analysis on servings and frequencies of several foods. An important concept in cluster analyses is to explore relations between dietary guidelines to the public and health outcomes, and these guidelines are usually expressed in terms of servings. Cluster and factor analyses are often referred to as empirically derived and, therefore, are not limited by current knowledge, in contrast with scores that are based on dietary guidelines and assumptions about what is "healthy." However, in general, cluster and factor analyses are based on highly aggregated food groups that, in turn, are formed by the presence or absence of certain nutrients and components known to be related to the outcome variables. One of the reasons for aggregating foods into a limited number of food groups is to avoid exclusion of subjects due to missing data; however, this might be done at the sacrifice of identification of key food behaviors with implications for health. If these analyses should be data driven and combine information of many aspects of the diet, they must be based on intake of food categories reflecting both cultural and nutrient differences. Probably, much information about cultural and/or nutrient-specific patterns would be lost if, for example, milk products with varying fat content are grouped together or if noncaloric beverages are excluded or grouped into one food category. In the present study, tea consumption was associated with a "healthy" food pattern, soft drinks with a "fast energy" pattern, and boiled coffee with a "traditional" pattern. Potentially, beverages can be used as markers for more or less healthy behaviors, as suggested by Duffey and Popkin (47).

In conclusion, in the present population, the cluster of subjects consuming a "healthy" diet had better anthropometric, blood pressure, and blood lipid values than did the other dietary patterns. The group of people with this healthy pattern were distinguished from the population in general by more frequent consumption of high-fiber cereals, low-fat milk products, fruit, vegetables, tea, and fish and less frequent consumption of products rich in fat and sugar, coffee, beer, and spirits, or in other words a diet corresponding with the current recommendations. However, the fact that the diagnosis of disease or risk seems to be an important motive for conforming to dietary recommendations (or reporting to do so) has to be taken into consideration in future research, because medical treatment and biased reporting might lead to erroneous conclusions regarding the associations between healthy food patterns and health outcomes.


   ACKNOWLEDGMENTS
 
The authors' responsibilities were as follows—CMB: designed the study, conducted and synthesized the analyses, and wrote the manuscript; NA and GL: provided statistical support and advice; KT, DST, LL, and AR: designed the research program and organized and founded the data collection; DST, LL, and AR: contributed to the study design and manuscript; ES: coordinated the data bank; and AW: developed the FFQ. All authors read and approved the final version of the manuscript. No conflicts of interest were reported.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Hu FB, Willett WC. Optimal diets for prevention of coronary heart disease. JAMA 2002;288:2569–78.[Abstract/Free Full Text]
  2. Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol 2002;13:3–9.[CrossRef][Medline]
  3. de Lorgeril M, Salen P, Martin JL, Monjaud I, Delaye J, Mamelle N. Mediterranean diet, traditional risk factors, and the rate of cardiovascular complications after myocardial infarction: final report of the Lyon Diet Heart Study. Circulation 1999;99:779–85.[Abstract/Free Full Text]
  4. Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med 1997;336:1117–24.[Abstract/Free Full Text]
  5. Schulze MB, Hoffmann K. Methodological approaches to study dietary patterns in relation to risk of coronary heart disease and stroke. Br J Nutr 2006;95:860–9.[CrossRef][Medline]
  6. Diehr P, Beresford SA. The relation of dietary patterns to future survival, health, and cardiovascular events in older adults. J Clin Epidemiol 2003;56:1224–35.[CrossRef][Medline]
  7. Farchi G, Mariotti S, Menotti A, Seccareccia F, Torsello S, Fidanza F. Diet and 20-y mortality in two rural population groups of middle-aged men in Italy. Am J Clin Nutr 1989;50:1095–103.[Abstract/Free Full Text]
  8. 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;104:208–14.[CrossRef][Medline]
  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;161:1857–62.[Abstract/Free Full Text]
  10. 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;72:912–21.[Abstract/Free Full Text]
  11. Akesson A, Weismayer C, Newby PK, Wolk A. Combined effect of low-risk dietary and lifestyle behaviors in primary prevention of myocardial infarction in women. Arch Intern Med 2007;167:2122–7.[Abstract/Free Full Text]
  12. Osler M, Heitmann BL, Gerdes LU, Jorgensen LM, Schroll M. Dietary patterns and mortality in Danish men and women: a prospective observational study. Br J Nutr 2001;85:219–25.[Medline]
  13. Wirfalt E, Hedblad B, Gullberg B, et al. Food patterns and components of the metabolic syndrome in men and women: a cross-sectional study within the Malmo Diet and Cancer cohort. Am J Epidemiol 2001;154:1150–9.[Abstract/Free Full Text]
  14. Bailey RL, Gutschall MD, Mitchell DC, Miller CK, Lawrence FR, Smiciklas-Wright H. Comparative strategies for using cluster analysis to assess dietary patterns. J Am Diet Assoc 2006;106:1194–200.[CrossRef][Medline]
  15. Berg CM, Lissner L, Aires N, et al. Trends in blood lipid levels, blood pressure, alcohol and smoking habits from 1985 to 2002: results from INTERGENE and GOT-MONICA. Eur J Cardiovasc Prev Rehabil 2005;12:115–25.[CrossRef][Medline]
  16. Golara M, Jones C, Randhawa M, Shennan AH. Inflationary oscillometric blood pressure monitoring: validation of the OMRON-MIT. Blood Press Monit 2002;7:325–8.[CrossRef][Medline]
  17. National Institutes of Health. Third report of National Cholesterol Education Programme Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Bethesda, MD: National Institutes of Health, 2001. (NIH publication 01-3670.)
  18. Messerer M, Johansson SE, Wolk A. The validity of questionnaire-based micronutrient intake estimates is increased by including dietary supplement use in Swedish men. J Nutr 2004;134:1800–5.[Abstract/Free Full Text]
  19. Wolk A, Ljung H, Vessby B, Hunter D, Willett WC. Effect of additional questions about fat on the validity of fat estimates from a food frequency questionnaire. Study Group of MRS SWEA. Eur J Clin Nutr 1998;52:186–92.[Medline]
  20. Wolk A, Vessby B, Ljung H, Barrefors P. Evaluation of a biological marker of dairy fat intake. Am J Clin Nutr 1998;68:291–5.[Abstract]
  21. Aires N, Selmer R, Thelle D. The validity of self-reported leisure time physical activity, and its relationship to serum cholesterol, blood pressure and body mass index. A population based study of 332 182 men and women aged 40–42 years. Eur J Epidemiol 2003;18:479–85.[CrossRef][Medline]
  22. Saltin B, Grimby G. Physiological analysis of middle-aged and old former athletes: comparison with still active athletes of the same ages. Circulation 1968;38:1004–15.
  23. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 2000;32(suppl):S498–504.[Medline]
  24. Newby PK, Muller D, Hallfrisch J, Qiao N, Andres R, Tucker KL. Dietary patterns and changes in body mass index and waist circumference in adults. Am J Clin Nutr 2003;77:1417–25.[Abstract/Free Full Text]
  25. Millen BE, Quatromoni PA, Nam BH, O'Horo CE, Polak JF, D'Agostino RB. Dietary patterns and the odds of carotid atherosclerosis in women: the Framingham Nutrition Studies. Prev Med 2002;35:540–7.[CrossRef][Medline]
  26. Greenwood DC, Cade JE, Draper A, Barrett JH, Calvert C, Greenhalgh A. Seven unique food consumption patterns identified among women in the UK Women's Cohort Study. Eur J Clin Nutr 2000;54:314–20.[CrossRef][Medline]
  27. Pryer JA, Cook A, Shetty P. Identification of groups who report similar patterns of diet among a representative national sample of British adults aged 65 years of age or more. Public Health Nutr 2001;4:787–95.[Medline]
  28. Pryer JA, Nichols R, Elliott P, Thakrar B, Brunner E, Marmot M. Dietary patterns among a national random sample of British adults. J Epidemiol Community Health 2001;55:29–37.[Abstract/Free Full Text]
  29. Liese AD, Schulz M, Moore CG, Mayer-Davis EJ. Dietary patterns, insulin sensitivity and adiposity in the multi-ethnic Insulin Resistance Atherosclerosis Study population. Br J Nutr 2004;92:973–84.[CrossRef][Medline]
  30. Villegas R, Salim A, Collins MM, Flynn A, Perry IJ. Dietary patterns in middle-aged Irish men and women defined by cluster analysis. Public Health Nutr 2004;7:1017–24.[Medline]
  31. Engeset D, Alsaker E, Ciampi A, Lund E. Dietary patterns and lifestyle factors in the Norwegian EPIC cohort: the Norwegian Women and Cancer (NOWAC) study. Eur J Clin Nutr 2005;59:675–84.[CrossRef][Medline]
  32. Ruxton CH, Kirk TR. Breakfast: a review of associations with measures of dietary intake, physiology and biochemistry. Br J Nutr 1997;78:199–213.[CrossRef][Medline]
  33. Huijbregts PP, Feskens EJ, Kromhout D. Dietary patterns and cardiovascular risk factors in elderly men: the Zutphen Elderly Study. Int J Epidemiol 1995;24:313–20.[Abstract/Free Full Text]
  34. Quatromoni PA, Copenhafer DL, D'Agostino RB, Millen BE. Dietary patterns predict the development of overweight in women: The Framingham Nutrition Studies. J Am Diet Assoc 2002;102:1239–46.[Medline]
  35. Wirfalt E, Mattisson I, Gullberg B, Berglund G. Food patterns defined by cluster analysis and their utility as dietary exposure variables: a report from the Malmo Diet and Cancer Study. Public Health Nutr 2000;3:159–73.[Medline]
  36. Millen BE, Quatromoni PA, Pencina M, et al. Unique dietary patterns and chronic disease risk profiles of adult men: the Framingham nutrition studies. J Am Diet Assoc 2005;105:1723–34.[CrossRef][Medline]
  37. Giugliano D, Ceriello A, Esposito K. The effects of diet on inflammation: emphasis on the metabolic syndrome. J Am Coll Cardiol 2006;48:677–85.[Abstract/Free Full Text]
  38. Dhingra R, Sullivan L, Jacques PF, et al. Soft drink consumption and risk of developing cardiometabolic risk factors and the metabolic syndrome in middle-aged adults in the community. Circulation 2007;116:480–8.[Abstract/Free Full Text]
  39. Togo P, Osler M, Sorensen TI, Heitmann BL. Food intake patterns and body mass index in observational studies. Int J Obes Relat Metab Disord 2001;25:1741–51.[CrossRef][Medline]
  40. Ledikwe JH, Smiciklas-Wright H, Mitchell DC, Miller CK, Jensen GL. Dietary patterns of rural older adults are associated with weight and nutritional status. J Am Geriatr Soc 2004;52:589–95.[Medline]
  41. Fraser GE, Welch A, Luben R, Bingham SA, Day NE. The effect of age, sex, and education on food consumption of a middle-aged English cohort-EPIC in East Anglia. Prev Med 2000;30:26–34.[CrossRef][Medline]
  42. Newby PK, Muller D, Tucker KL. Associations of empirically derived eating patterns with plasma lipid biomarkers: a comparison of factor and cluster analysis methods. Am J Clin Nutr 2004;80:759–67.[Abstract/Free Full Text]
  43. Tucker KL, Dallal GE, Rush D. Dietary patterns of elderly Boston-area residents defined by cluster analysis. J Am Diet Assoc 1992;92:1487–91.[Medline]
  44. Martikainen P, Brunner E, Marmot M. Socioeconomic differences in dietary patterns among middle-aged men and women. Soc Sci Med 2003;56:1397–410.[CrossRef][Medline]
  45. Emberson JR, Whincup PH, Walker M, Thomas M, Alberti KG. Biochemical measures in a population-based study: effect of fasting duration and time of day. Ann Clin Biochem 2002;39:493–501.[Abstract/Free Full Text]
  46. Troisi RJ, Cowie CC, Harris MI. Diurnal variation in fasting plasma glucose: implications for diagnosis of diabetes in patients examined in the afternoon. JAMA 2000;284:3157–9.[Abstract/Free Full Text]
  47. Duffey KJ, Popkin BM. Adults with healthier dietary patterns have healthier beverage patterns. J Nutr 2006;136:2901–7.[Abstract/Free Full Text]
Received for publication December 14, 2007. Accepted for publication April 16, 2008.




This article has been cited by other articles:


Home page
 J. Nutr.Home page
A. L. Brantsaeter, M. Haugen, S. O. Samuelsen, H. Torjusen, L. Trogstad, J. Alexander, P. Magnus, and H. M. Meltzer
A Dietary Pattern Characterized by High Intake of Vegetables, Fruits, and Vegetable Oils Is Associated with Reduced Risk of Preeclampsia in Nulliparous Pregnant Norwegian Women
J. Nutr., June 1, 2009; 139(6): 1162 - 1168.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.
Agricola
Right arrow Articles by Berg, C. M
Right arrow Articles by Lissner, L.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS