To provide accurate estimates of weight change attributable to alcohol intake using data from a national sample of US adults participating in first National Health and Nutrition Examination Survey (NHANES).

Adults aged 25 to 74 years who were examined from 1971 through 1975 and were eligible for inclusion in the NHANES I Epidemiologic follow-up study.

• Not adults aged 25 to 74 years who were examined from 1971 through 1975 for eligibility for inclusion in the NHANES I Epidemiologic follow-up study
• Adults who were screened for the NHANES I Epidemiologic follow-up study, but were found to:
  • Have missing vital status data at follow-up
  • Be alive, but were not re-interviewed
  • Have died
  • Have no information on their caloric intake at baseline
  • Be pregnant at baseline or follow-up
  • Have an unknown height or weight measurement at baseline or follow-up
  • Have missing drinking information on baseline or follow-up
  • Have missing values for one or more other covariates used in the analysis.

Recruitment

The current study is using secondary data from the NHANES I Epidemiologic follow-up study.

Design

• Cross-sectional and prospective analysis of data collected from NHANES I Epidemiologic follow-up study. Participants were re-weighed ten years after the initial participation in NHANES I
• In some analyses, weight change status was defined as major weight gain (weight change, 10kg or more gain) or major weight loss (weight change, 10kg or more loss) over the follow-up period.

Dietary Intake/Dietary Assessment Methodology

• Alcohol consumption, expressed as drinks per day, was assessed at baseline using quantity-frequency questions on the medical history interview
• Study population was categorized into six groups:
  • Non-drinkers: Those who reported that they did not drink during the past year
  • Infrequent drinkers: Those who had fewer than 12 drinks a year
  • Very light drinkers: Those who had 12 or more drinks a year but less than one drink a week
  • Light drinkers: Those who had one or more drinks a week but less than one drink per day
  • Moderate drinkers: Those who drank one to 1.9 drinks per day
  • Heavy drinkers: Those who had two or more drinks per day

Statistical Analysis

• Sex-specific cross-sectional analyses were conducted to estimate mean body differences between non-drinkers and drinkers at baseline, using multiple linear regression analysis to adjust for covariates
• That same technique and analysis was used to examine whether alcohol intake was related to the longitudinal change in body weight over the 10-year follow-up period
• Multiple polytomous logistic regression analysis was used to estimate relative odds of major weight gain or weight loss over the follow-up period by the amount of alcohol intake at baseline
• Used standard diagnostic tests and the residual plot for linear regression to assess goodness-of-fit
• Separate analyses were carried out after stratifying by health status.

Timing of Measurements

• Weight and height was measured at baseline and 10 years later at follow-up
• Alcohol consumption was measured at baseline.

Dependent Variables

• Body weight and height, mean BMI (body mass index), weight gain and weight loss
• Height was measured by using a level platform with a vertical bar (to which was attached a steel tape); weight was measure on a self-balancing scale and measured at follow-up with a portable spring scale (adjustments were made to include removed shoes and extra articles of clothing and the adjustment of -1.6kg for indoor clothing).

Independent Variables

• Alcohol consumption, calories from alcohol
• Alcohol was measured at baseline through the completion of a quantity-frequency questionnaire within the medical history interview; study population were divided into six categories depending on amount of alcohol consumed (i.e., non-drinkers; less than 12 drinks per year; less than one drink per week; one to 6.9 drinks per week; one to 1.9 drinks per day; two or more drinks per day).

Control Variables

• Age, race, height, education, health status (obtained at follow-up based on physician-diagnosed conditions reported), smoking status (questions asked in follow-up interview), physical activity (obtained from questions), total non-alcoholic calories at baseline (based on 24-hour recall), and dieting for weight loss (based on baseline interviews asking if participants were on special diets)
• In prospective analysis:
  • Years of follow-up
  • Reproductive status
  • Baseline BMI.

• Initial N: 14,407 adults aged 25 to 74 years eligible for inclusion in the NHANES 1 study
• Attrition (final N):
  • The final number of participants was 7,230 adults, after exclusions
  • Adults who were screened for the NHANES I Epidemiologic follow-up study were excluded from analysis if (number and percent):
    • Vital status data was missing at follow-up: 1,024 (7.1%)
    • They were alive but were not re-interviewed: 838 (5.8%)
    • They had died: 2,022 (14%)
    • Caloric intake information was missing at baseline: 2,270 (15.7%)
    • They were pregnant at baseline or follow-up: 101 (0.7%)
    • Height or weight measurements were unknown at baseline or follow-up: 560 (3.9%)
    • Drinking information was missing on baseline or follow-up: 81 (0.6%)
    • Values for one or more other covariates used in the analysis was missing: 281 (2%).
• Age:
  • Range: 25 to 74 years
  • Mean age (years):
    • Non-drinkers: Men, 55; women, 51
    • Those that drank one to 6.9 drinks per week: Men, 46; women, 40
    • Those that drank one to 1.9 drinks per day: Men, 46; women, 40
    • Those that drank two or more drinks per day: Men, 46; women, 40
• Ethnicity: Greater than 80% of participants were white
• Other relevant demographics:
  • Majority of the participants in the study had less than 12 years of education
  • Summary demographics at baseline:
    • Fewer women drank than did men (67% vs. 81%)
    • Men drinkers tended to drink more (15% of the men drank two or more drinks per day vs. 3% of the women)
    • Among men and women, those who were white, younger and better educated were more likely to drink
    • Men and women drinkers had significantly higher intakes of total daily calories than non-drinkers (this relation persisted even after excluding calories contributed by alcohol)
  • Sample size with respect to drinking status at baseline:
    • Non-drinkers: Men, 500; women, 1,505
    • Those that drank less than 12 drinks per year: Men, 344; women, 1,075
    • Those that drank less than one drink per week: Men, 296; women, 625
    • Those that drank one to 6.9 drinks per week: Men, 801; women, 1,043
    • Those that drank one to 1.9 drinks per day: Men, 290; women, 222
    • Those that drank two or more drinks per day: Men, 383; women, 146
  • Mean daily calories (kcal) with respect to drinking status at baseline:
    • Non-drinkers: Men: 2,056; women, 1,406
    • Those that drank less than 12 drinks per year: Men, 2,143; women, 1,452
    • Those that drank less than one drink per week: Men, 2,290; women, 1,557
    • Those that drank one to 6.9 drinks per week: Men, 2,344; women, 1,565
    • Those that drank one to 1.9 drinks per day: Men, 2,465; women, 1,596
    • Those that drank two or more drinks per day: Men, 2,562; women, 1,614
  • Calories from alcohol (kcal) with respect to drinking status at baseline:
    • Non-drinkers: Men, zero; women, zero
    • Those that drank less than 12 drinks per year: Men, one; women, one
    • Those that drank less than one drink per week: Men, six; women, six
    • Those that drank one to 6.9 drinks per week: Men, 35; women, 29
    • Those that drank one to 1.9 drinks per day: Men, 103; women, 98
    • Those that drank two or more drinks per day: Men, 336; women, 270
• Anthropometrics:
  • Although BMI of men was similar for drinkers and non-drinkers, the BMI of women drinkers was lower than that of women non-drinkers
  • Mean BMI (kg/m²) with respect to drinking status at baseline:
    • Non-drinkers: Men, 25.6; women, 26.6
    • Those that drank less than 12 drinks per year: Men, 26.2; women, 25.7
    • Those that drank less than drink per week: Men, 26.0; women, 24.7
    • Those that drank one to 6.9 drinks per week: Men, 25.8; women, 24.4
    • Those that drank one to 1.9 drinks per day: Men, 25.7; women, 24.1
    • Those that drank two or more drinks per day: Men, 25.8; women, 24.5
• Location: United States.

Key Findings

• Prospectively, both men and women drinkers tended to gain less weight than did non-drinkers (P=0.006 for trend in women, P=0.11 for trend in men)
• Over the 10 year period 180 men (6.9%) and 482 women (10.4%) had a major weight gain (10kg or more gain) and 173 men (6.6%) and 379 women (8.2%) had a major weight loss (10kg or more loss)
• Drinkers had more stable weight over the 10-year follow-up period
• Drinkers were less likely to have major weight gain or loss (gaining or losing 10kg or more) than were non-drinkers
• Compared with non-drinkers, for those who consumed one to 6.9 drinks per week:
  • Women had an odds ratio (OR)=0.7 (95% CI: 0.5 to 0.9) for major weight gain and an OR=0.7 (95% CI: 0.5 to 1.1) for major weight loss
  • Men had an OR=1.0 (95% CI: 0.6 to 1.6) for major weight gain and an OR=0.7 (95% CI: 0.5 to 1.2) for major weight loss
• For those who consumed two or more drinks per day:
  • Women had an OR=0.5 (95% CI: 0.3 to 1.0) for major weight gain and an OR=0.8 (95% CI: 0.4 to 1.6) for major weight loss
  • Men had an OR=0.9 (95% CI: 0.5 to 1.6) for major weight gain and an OR=1.0 (95% CI: 0.6 to 1.7) for major weight loss.

Mean Difference in Body Weight (kg) Between Non-drinkers and Different Groups of Drinkers at Baseline: First National Health and Examination Survey, 1971 to 1975

**Number in parentheses, 95% confidence interval

Other Findings

Cross-sectional association between alcohol intake and body weight:

• At baseline, male drinkers tended to be heavier than non-drinkers, while women drinkers weighed less than non-drinkers
• Among men, mean body weight among different groups differed little from that of non-drinkers after controlling for potential confounding factors
• Among women, the cross-sectional association at baseline was much stronger than observed for men; that is, body weight decreased substantially with increasing alcohol intake. This inverse relationship decreased but remained significant even after controlling for potential confounding factors.

• For the US adult population, the study found that alcohol was inversely associated with body weight among women, and both men and women drinkers tended to gain less weight than did non-drinkers over a ten-year period
• Drinkers were less likely to have either a major weight gain or loss than non-drinkers
• The study results suggest than alcohol intake is not a risk factor for obesity.

Authors identified these limitations in the study:

• Alcohol intake was self-reported and thus likely to be underestimated (if this reporting error does not vary with body weight status, the resulting misclassification would bias the association between alcohol intake and body weight toward null and the weak inverse relation between alcohol and weight gain observed may be a conservative estimate)
• While 24-hour dietary recalls can provide reasonable accurate estimates of mean caloric intake for groups, they may not adequately represent and individual's intake
• Alcohol intake and body weight may have fluctuated over the follow-up periods in ways not captured by the baseline and follow-up measures and the findings may not provide much insight on the relation between changes in alcohol intake and body weight changes over time.