Trend Study

D - Click here for explanation of classification scheme.

NEUTRAL: See Quality Criteria Checklist below.

• To analyze data from cross-sectional, screening surveys to compare the association of breakfast size on BMI
• To examine the effect of relative breakfast size on BMI in males and females
• To examine possible confounders of this relationship.

Residents of the general community around the Sydney Adventist Hospital from 1976 until 2000. After 2000, large corporations were invited to send their corporate executives for a comprehensive evaluation of the health and associated risk factors.

None noted.

Recruitment

Between 1976 and 2000, residents of the general community around the Sydney Adventist hospital were invited to take part in heart health screening surveys that assessed heart health, food consumption and other lifestyle practices. After 2000, these services were no longer offered, rather large corporations were invited to send their corporate executives for a comprehensive evaluation of their health and associated risk factors.

Design

Trend study.

Blinding Used

Respondents were randomly selected from these annual surveys in 1976, 1986 and 2005. 

Intervention

At the time of the screening, hospital staff asked respondents to complete a questionnaire that included a food frequency questionnaire, frequency of various moderated and vigorous physical activities, dietary habits, smoking and alcohol usage and lifestyle.

Statistical Analysis

• Independent T-tests were performed to assess differences in ages between men and women for each year selected (presented as means ± SEM)
• Univariate analysis of variance was conducted on each sex for:
  • Demographic variables: 
    • Age
    • Sex
    • Religion
    • Marital status
    • Education
  • Food variables:
    • Eating behaviors:
      • Eating between meals
      • Breakfast
      • Regular meals patterns
      • Dieting
    • Lifestyle variables:
      • Smoking
      • Physical activity
      • Sleep
      • Time urgency (where supplied)
• Each of the demographic, food and lifestyle variables found to be significantly associated with body mass index was tested by backward linear regression
• Spearman rho was used to calculate correlation coefficients between breakfast size and various foods, using two-tailed tests of significance
• Forward linear regression was used to test possible breakfast confounders of physical activity and vegetarianism
• All differences were considered significant if P<0.05. 

Timing of Measurements

All data was collected at the time of the screening.

Dependent Variables

Biomedical indices:

• Height
• Weight
• Body mass index.

Independent Variables

• Food Frequency Questionnaire: In 1976, 42 different foods and drinks were included in the questionnaire, as well as dietary practices (relative breakfast size, dieting to lose weight, frequency of cut fat from meat and eating between meals). In 1986, these foods and dietary practices were replicated with the addition of fish and cream. In 2005, the questionnaire differed slightly in that the number of meats, milks and dessert foods was reduced, and there was more information about takeaways, low-energy and low-fat foods and the relative size of dinner. 
• Breakfast Question: In 1976 and 1986, participants were asked to compare the amount they ate for breakfast with a standard breakfast that comprised a bowl of cereal, a serving of fruit or juice, a cup of milk, and a slice of toast with juice. In 2005, participants were asked whether their breakfast was a large meal, moderate, a small meal or a cup of coffee, piece of toast or less.

Control Variables

Possible Confounders: Vegetarianism, physical activity, and lifestyle factors. Vegetarianism was determined from the aggregate score when meat was consumed "never" or "less than once per week." Physical activity (in 1976 and 1986) was assessed by questions about frequency of participation in walking, running, cycling, swimming, tennis, vigorous gardening and other vigorous activity. In 2005, participants recorded number of days exercised, time (minutes) of each exercise session and intensity (low, moderate or high). Other lifestyle factors examined included hours of sleep per night, smoking status (current, past and never smoked), frequency of alcohol consumption and number of drinks consumed per sitting.

• Initial N:
  • 1976: 384 men and 338 women
  • 1986: 244 men and 229 women
  • 2005: 270 men and 62 women
• Attrition (final N):
  • 1976: 384 men and 338 women
  • 1986: 244 men and 229 women
  • 2005: 270 men and 62 women
• Age: Mean age of men and women:
  • 1976: Men=44.5±0.6; women=45.9±0.7
  • 1986: Men=44.9±0.9; women=43.7±1.1
  • 2005: Men=49.7±0.6; women=47.6±1.3
• Other relevant demographics:
  • Over-sampling in 1976 and 1986 accommodated the large number of vegetarians present in these years (29.3% in 1976 and 19.4% in 1986)
  • The proportions of men and women in the samples were similar to the general population in 1976 and 1986, but contained more men in 2005
  • The sample was older, was better educated and had more married individuals in both 1976 and 1986 and was older in 2005 than the general population
• Anthropometrics: BMI was determined but not reported as a mean for each sampling group
• Location: Sydney, Australia.

Key Findings

• Age comparison of men and women: No differences in the mean ages of male and female participants in any of the years examined; however, men and women in 2005 were older than in previous years
• Relationship of breakfast consumption to BMI:
  • Reported breakfast amount increased while BMI decreased in men (this trend was observed in all three years). However, this relationship was not observed among women.
  • Male vegetarian participants had lower BMI than did non-vegetarians in 1976, 1986 and 2005
  • The relationship between breakfast and BMI was much stronger among male vegetarians than male non-vegetarians
  • Cereals and fruit were strongly associated with breakfast in all three years; spreads were more important in 1986 than the other years
  • Increasing coffee consumption (particularly in 1986 and 2005) was associated with the trend toward smaller breakfasts or breakfast skipping
  • Other food or eating-related variables found to be positively associated with men's BMI included the frequency of consumption of chicken, wieners (sausages) and alcohol together with dieting to lose weight
  • Food variety and the frequency of consumption of spreads were negatively related to BMI for men in 1976
  • In 1986, eating between meals and the frequency of consumption of coffee and table salt were associated with increases in BMI, whereas physical activity was negatively associated with BMI. 

• Results of this study confirm that smaller breakfasts were associated with higher BMI
• Skipping breakfast or the consumption of small breakfasts together with the consumption of coffee, table salt, and energy-dense foods; eating between meals; and dieting to lose weight seemed to be associated with increased BMI among men but not women
• Increasing coffee consumption was associated with the trend toward smaller breakfasts or breakfast skipping.

Industry:	Sanitarium Health Food Company Food Company:
University/Hospital:	Deakin University in Melbourne, Australia
Other:	Australian Research Council grant

Different samples and different instruments used for the three cross-sectional surveys.