| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Medicine and Therapeutics, Centre for Nutritional Studies, The Chinese University of Hong Kong, HONG KONG
Address reprint requests to: Professor Jean Woo, M.D., F.R.A.C.P., Department of Medicine and Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, N.T., HONG KONG. E-mail: jeanwoowong{at}cuhk.edu.hk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
Objective: To examine the association between food variety and body fatness in Hong Kong Chinese adults.
Design: One hundred and twenty Hong Kong Chinese adults (aged 18–50 y). Usual dietary intake over a one-week period of all subjects was assessed by using a food frequency questionnaire. Anthropometric parameters were measured using standardized methods.
Results: Varieties of grain and meat were negatively correlated with obesity indices (grain vs. BMI/body fat/waist/hip circumferences: partial r = –47/–0.43/–0.46/–0.42, p < 0.001; meat vs. BMI/body fat/waist/hip circumferences: partial r = –0.31/–0.24/–0.25/–0.29, p < 0.01). In contrast, there was a positive relationship between variety of snack and obesity indices (snack vs. BMI/body fat/waist/hip circumferences: partial r = 0.35/0.42/0.42/0.36, p < 0.001). A food variety ratio derived from varieties of snack, grain and meat, was a stronger predictor of body fat compared with dietary fat in a regression model.
Conclusion: Food variety may contribute to the local escalation in the prevalence of obesity. The variety of snack is the promoting factor for obesity while the variety of grains and meats may counteract its development. The findings of this study may have implications for treatment of obesity and the prevention of further weight gain.
Key words: food variety, snack, grain, meat, obesity, Chinese
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
25) and obesity (BMI 30) is 25%–30% and 3%–5%, respectively [6,7]. However, using the new definition of obesity for the Asia-Pacific region (BMI 23 for overweight; BMI 25 for obese), the actual prevalence would be much higher [8]. The etiology of obesity is multifactorial and involves complex interactions of many environmental, social and genetic factors. It is a consequence of energy imbalance where energy intake exceeds energy expenditure over a considerable period. Dietary intake is one of the important factors affecting energy balance and is considered to be a major modifiable factor, through which changes of body weight can be promoted. Among the many dietary factors, the percentage of energy from dietary fat is widely believed to be an important determinant of accumulation of body fat [9–12]. This causal relationship is controversial, as no correlation between dietary fat intake and body fat was found in epidemiological studies [13–15]. Furthermore, statistics show that in the past two decades in United States, the prevalence of obesity has increased whereas the fat consumption was reduced [16–17]. Therefore, body fatness could not be accounted for by dietary fat alone.
Recently, human and animal studies suggest that food variety may play a role in the accumulation of body fat [18–22]. An increase in food variety provides a surplus of energy intake, which results in an increase in body fat [18–22]. In addition, increased variety from energy dense foods has been shown to be positively associated with body fatness whereas there is an inverse correlation with increased variety from low energy dense foods [20]. These findings may help to explain the rising prevalence of obesity. It may also provide insight to weight management for overweight and obese individuals.
Most studies had been carried out in Caucasian populations, and the results may not be applicable in Chinese subjects who have different dietary habits. In this study, the validity of this concept is tested in Hong Kong Chinese. In addition, the differences in dietary patterns between obese and normal weight adults, and their associations with body fatness are also investigated. The information obtained could then be used to provide better public health nutritional recommendations.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
Dietary Assessment
Usual dietary intake over a one-week period of all subjects was assessed by using a local food frequency questionnaire, which has been validated [24]. On the basis of the previous local dietary surveys, the food items chosen in the questionnaire were those most frequently consumed [25,26], and they were categorized in the following food groups: bread/pasta/rice (19 items); vegetables (53); fruits (26 items); meat (35 items); fish (30 items); eggs (6 items); tofu (7 items); beverages (28 items); dimsum/snacks (45 items); soups (10 items); and oil/salt/sauces. Some food groups were combined as a single food group for food variety analysis and finally there were 6 food groups: grains (bread/pasta/rice); vegetables; fruits; meats (meat/fish/eggs/tofu); beverages (beverages/soups) and snacks (dimsum/snacks) (Appendix).
Before the interview, subjects were advised to make a brief record at home to facilitate the subjects’ memorizing what they had eaten without missing food items during the previous week. The record also helped shorten the time of interviewing. Each subject was then asked to complete the questionnaire including the food items, the size of each portion, the frequency of consumption on a daily and weekly basis. During the interview, portion size was explained to subjects by using food replica, food containers, and a catalogue of pictures of individual food portions. Data were scrutinized by the dietary pattern (for example, if meals were missed) to determine the number of times staple foods such as rice or noodles consumed over a one-week period. The amount of cooking oil was estimated according to the 1995 Hong Kong Adult Dietary Survey [23]. The type of oil used was also recorded to estimate the type of fat in the diet. Quantification of nutrients was achieved by multiplying each food item’s frequency of consumption by the portion size consumed, and the nutrient content of the food. The nutrient content was determined from food tables for Hong Kong, which were compiled from McCance and Widdowson [27], and two food tables used in China, as published by the Institute of Health of the Chinese Medical Science Institute [28] and Zhongshan University [29]. Besides, the macronutrient contents were also validated by chemical analysis (unpublished data). Nutrient density (nutrient per kcal) and nutrient per kg fat free mass were used to determine the differences between obese and normal weight subjects. Since underreporting is reported elsewhere [30–33], the accuracy of the reported energy intake was then compared with the predicted energy requirement by the equations of Liu [24,34,35].
Calculation of Food Variety
Food variety of each food group was calculated as the percentage of different food items consumed within the corresponding food group, regardless of the frequency and portion with which they were consumed [20]. Food variety ratio was defined as a composite variety variable. In the study by McCrory et al., it was calculated as the ratio of the variety of combined food groups which were inversely correlated to body fatness to the variety of combined food groups with positive relationships to body fatness (and hence it was expressed as the ratio of the variety of vegetables consumed to the variety of the sweets, snacks, condiments, entrees, and carbohydrates) [20]. In the present study, the variety ratio was calculated as a ratio of variety of combined food groups which were positively correlated to body fatness to the variety of combined food groups with inverse relationships to body fatness. Therefore, it was defined as the ratio of variety of snacks to the variety of grains and meats.
Anthropometric Measurement
Height and weight were measured to the nearest 0.5 cm and 0.1 kg, respectively, without shoes and in light clothing. Waist and hip circumferences were measured three times, and their means were used for analysis. Waist circumference was measured half way between umbilicus and xiphoid, and readings were taken at the end of expiration [36]. Hip circumference was measured at the level of the great trochanters. Waist circumference was suggested as an indicator of adiposity and was used to measure body fatness in many studies [37–39]. Body fat (expressed as % of weight) and fat free mass were calculated by using the equation of Lean (body fat for men = 0.567 waist (cm) + 0.101 age (y) – 31.8; body fat for women = 0.439 waist (cm) + 0.221 age (y) – 9.4) which was showed to be most robust prediction with the least bias when compare with the conventional underwater weighing method [30]. Fat free mass was then calculated by subtracting the fat mass from body weight (weight (kg) – weight (kg) x % of body fat).
Statistical Analysis
Data were analyzed by the SYSTAT software (version 10.0; SPSS, Inc, Chicago). Descriptive data were presented as mean ± SD unless otherwise indicated. An independent t test (2-tailed significant) was used to compare the mean nutrient intake and other dietary variables between obese and normal weight subjects. A chi-square test was also carried out to test the association between food variety and obesity status. Pearson’s partial correlation was calculated to test the associations between dietary factors and the body fatness, which was represented by BMI, percentage of body fat, waist and hip circumferences (the correlations were adjusted for age, gender, and other non-tested food groups and dietary fat). On the basis of the results obtained, food variety ratio was calculated and the associations of all the dietary variables with body fatness were recomputed by stepwise multiple linear regression analysis. For all statistical tests, a p value of <0.05 was accepted as significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
|
|
|
|
To examine the strength of the relationships between these variables, Pearson’s correlation test was carried out. Four indices of body fatness (BMI, percentage of body fat, waist and hip circumferences) were used for analysis. Age and gender, and other non-tested indices of body fatness were adjusted for in all analysis. The associations between each of the body fatness indices and each of food groups were similar. Only varieties of grains, meats and snacks but not varieties of beverages, fruits and vegetables showed significant associations with indices of body fatness. Varieties of grains and meats were inversely correlated with obesity indices (grains: partial r range from –0.42 to –0.47, p < 0.001; meats: partial r range from –0.24 to –0.31, p < 0.01). In contrast, there was a positive relationship between variety of snacks and obesity indices (partial r range from 0.35 to 0.42, p < 0.001) independent of other body fatness factors. Similar results were obtained when 6 food groups were tested separately and the non-tested food groups were taken into control as well (grains: partial r range from –0.29 to –0.33, p < 0.01–0.001; meats: partial r range from –0.32 to –0.38, p < 0.01–0.001; snacks: partial r range from 0.34 to 0.43, p < 0.001) (Table 4).
|
Stepwise multiple regression analysis was used to determine the predictors of percentage of body fat. Interestingly, only food variety ratio was included in the regression model of body fat, accounting for 25% of the variance (body fat = 25.47 + (21.89 x variety ratio), R2 = 0.25, p < 0.001).
|
DISCUSSION |
|---|
Variety of snacks was shown to have a strong positive association with body fatness. Snacking has been considered as one of the possible determinants in the aetiology of obesity [44–47]. A widely accepted hypothesis from Booth proposes that consumption of extra snacks increases total energy intake as they are eaten in addition to meals [45]. Our findings of a significant positive correlation between variety of snacks and total energy intake (Fig. 1), might be explained by this hypothesis. Almost all of the snacks available in market are energy dense with high fat and/or high sugar such as cookies, chocolates and potato crisps etc. The palatability and pleasure derived from fat and sugar of the snacks tend to promote consumption [48]. Moreover, due to the high fat content, snacks exert only a weak action on satiation and satiety and probably result in overconsumption [49,50]. Snacks are also easy and quick to eat, so that more could be eaten in a short time. Therefore, it is reasonable that those consuming more variety of snacks are likely to have higher body fat. However, other food groups also show a positive relationship, and the relationship between snack variety and total energy intake is also confounded by the same possible relationship between the variety from other food groups and total energy intake. Some studies also show that snacking has no significant effect on body fatness or total energy intake [51,52] as some snacks may be low fat and low energy content [52].
Although grains was the largest proportion of food that consumed, variety of grains was shown to have negative correlation to body fatness. One explanation may be that its taste and texture is lesser palatable and pleasurable, and also compared with rice which is the staple food of Hong Kong Chinese, noodles/spaghetti/cereals are more bulky and easily fill up the stomach, so that the quantity consumed will be less. According to the Exchange List from the American Dietetic Association and the American Diabetes Association, the calorific value of a bowl of rice (251 kcal/1050 kJ) is higher than a bowl of noodles/pasta (160–180 kcal/670–753 kJ) [53]. Furthermore, in general, rice gives a higher glycemic index (GI) than noodles/pasta (rice: 60–88, instant noodles: 40, pasta: 32–40; glucose: 100) [54,55]. Higher GI grains may increase hunger and promote overeating relative to grains with lower GI [56–59], and high GI diets have been reported to stimulate de novo lipogenesis and result in enlarged size of adipocytes while low GI diets inhibit these responses [60]. In addition, many fast food restaurants tend to serve rice (boiled/fried) with oily sauces, but not noodles/pasta (boiled/fried), which might increase the energy density to the diet. Therefore, increasing the variety of grains may effectively lower the amount of energy intake and the GI of grains consumed, in a population where the staple diet is rice. This could account for the observation that the variety of grains was inversely proportional to the total energy intake (Fig. 1). Therefore, substituting increased variety of grains, such as noodles or pasta, for variety of snacks may prevent further weight gain or may even result in a loss of body fatness.
The variety of meats was also negatively associated with body fatness although its variety was positively proportional to total energy intake (but it was only very slightly). The underlying reason was not clear. However, one of the possible reasons may be the low energy density of high protein foods which are high water content. Moreover, we can notice that the obese people had a higher protein intake (either percentage of total energy intake or energy density) than the normal weight people. Obese people may eat more from a limited variety of meats, or they prefer red meats (beef, pork) with a higher fat content rather than white meats (fish, chicken) with lower fat content.
The findings are slightly different compared with those of McCrory et al. [20]. In their study, variety of sweets, snacks, condiments, entrees, and carbohydrates were positively associated with body fatness while only variety of vegetables was negatively associated with body fatness. One explanation may be differences in the methods and criteria used to define body fatness (McCrory et al. used percent of body fat as the only parameter for the body fatness [20]). Moreover, the dietary assessments were based on different length of period (6 month period in McCrory vs. 1 week period in the present study), and the use of only plausible energy intake reports. Other possible reasons include different categorization of food groups and items provided in the questionnaires (Appendix). For example, beef and spaghetti were categorized into the food groups of meats and grains, respectively, in this study, while they were both grouped into entrees in the study of McCrory et al. Furthermore, in our study, there were 26 and 63 food items listed in the food groups of fruits and vegetables, respectively, compared with 10 items of fruits and 14 items of vegetables in McCrory’s study. Moreover, vegetables are usually eaten raw or boiled in the western cuisine, but in the Chinese style of cooking, vegetables are usually fried, and thus, the energy content is increased. Therefore, if people do have a variety of vegetables, they may have higher calorie intake as well. This is supported by the increase in total energy intake with the variety of vegetables in our study (Fig. 1). Therefore, increasing variety of vegetables may be expected to be associated with the increase of body fatness in Hong Kong Chinese. However, this was not observed in our study. It is possible that although a large variety of vegetables may be consumed, the quantities tend to be very small.
A study of the relationship between dietary variety and body fatness in mainland Chinese [43] showed similar results to McCrory’s. However, the definitions of dietary variety of these studies were different from ours. Nevertheless, the main conclusion is similar in that dietary variety or food variety did play a role in causing obesity. Therefore, the relationship between food variety and body fat is likely to be dependent on geographic region and culture.
On the other hand, there are limitations in this study. One of these is the equation used to estimate body fat has not been validated in Chinese and this may have introduced bias in the analysis. Furthermore, the current dietary assessments did not include condiments and sauce might underestimate the total energy intake. Moreover, the lack of validation of food variety may also make the association between food variety and body fatness underestimated. In addition, information on socioeconomic variables was not collected, and these variables may be expected to affect food choices [42]. Therefore, it is worthwhile doing further studies on these issues. There are also limitations in the reporting of dietary intake, in that women selectively under report fat intake [30,61,62], and that obese subjects selectively under report either fat or snack intake [31–33]. However, we attempted to detect this by comparing the reported total energy intake with that calculation from predicted equations. The mean reported value was not lower than the predicted value.
Notwithstanding the limitations, the findings of this study may have implications on the dietary advice in the treatment of obesity and the prevention of further weight gain. Several strategies may be made by focusing on the variety of snacks, meats and grains. If a large variety of snacks increase the total energy intake is true [47,51,63], body fat may be reduced by limiting the variety of snack. However, our study, merely demonstrated an association between snack variety and total energy intake, without demonstrating that snacks increase total energy intake. In addition, a recommendation to increase the variety of regular portions of lean meats and grains could be made. Education is very important in promoting changes in dietary habits. Therefore, importance of food variety in weight management should be discussed with the patients.
In conclusion, the results of this study suggest that there are differences in dietary pattern between obese and normal weight people. Variety of snacks promotes obesity while variety of grains and meats may counteract its development. The relative importance of food variety in relation to other factors associated with obesity such as psychological, social and genetic factors could be explored further in future studies.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
Vegetables:
Fruits:
Meats:
Beverages:
Snacks:
Received June 12, 2003. Accepted February 12, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSAPPENDIXREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
