Dhingra et al
Soft Drink Consumption and Metabolic Risk
fructose corn syrup content,12 less satiety and compensation, and a general effect of consuming refined carbohydrates (see review by Drewnowski and Bellisle13).
The aforementioned data raise the possibility that the consumption of soft drinks can fuel metabolic derangements, including insulin resistance, that can translate into a greater risk of developing abdominal obesity, high triglyceride lev- els, low levels of high-density lipoprotein cholesterol (HDL- C), elevated blood pressure, and impaired glucose tolerance; this constellation of metabolic traits has been collectively referred to as the metabolic syndrome.14 Higher prevalence of the metabolic syndrome poses greater risk for cardiovascular disease in the community,15 although the independent contri- bution of this entity to vascular risk beyond its components has been questioned.16
In the present prospective investigation, we tested the hypothesis that greater soft drink consumption increases the risk of developing metabolic risk factors (alone and in combination [metabolic syndrome]) in middle-aged adults in the community. Additionally, we evaluated whether metabol- ic risk varied on the basis of consumption of sugar-sweetened (“regular”) versus artificially sweetened (“diet”) soft drinks.
prevalent cardiovascular disease (n926). After exclusions, a total of 8997 person-observations (4871 in women) were eligible for the cross-sectional analyses. For prospective analyses, we excluded individuals with baseline metabolic syndrome (n2897 person- observations; metabolic syndrome as defined below) and those with any missing metabolic syndrome components on follow-up (n61 person-observations). The schema for selection of individuals eligi- ble for cross-sectional and longitudinal analyses is displayed in the Figure. All participants provided written informed consent, and the protocol for the study was approved by institutional review board of Boston Medical Center.
Measurement of Covariates
At each Framingham Heart Study examination, participants provided a medical history and underwent a complete standardized physical examination that included anthropometry, blood pressure measure- ments, and laboratory assessment of vascular risk factors. Fasting levels of blood glucose, triglycerides, and HDL-C were measured with standard assays. Blood pressure was measured by a physician using a mercury sphygmomanometer and with the participant resting in a seated position for 5 minutes; the average of 2 readings obtained on the participant’s left arm constituted the examination blood pressure. Physical activity was assessed by calculating a “physical activity index”; participants were asked specific questions regarding how many hours in a typical day they spent sitting, sleeping, or performing light-moderate or heavy physical activities.20 Alcohol intake was assessed by averaging the number of alcoholic beverages consumed per week. Participants who reported smoking 1 or more cigarettes per day in the year before the Framingham Heart Study examination were considered current smokers.
The Framingham Heart Study began in 1948 with the enrollment of 5209 participants into the original study cohort.17 In 1971, children of the original cohort participants and the spouses of the children were enrolled into the Framingham Offspring Study (n5124).18 Offspring study participants are evaluated approximately every 4 years. Information on daily consumption of soft drinks was collected via a physician-administered questionnaire at each study visit from the fourth (1987–1991) through the sixth (1995–1998) examination cycles. That examination questionnaire did not elicit information regarding consumption of regular versus diet soft drinks; however, such information was available from the self-administered food frequency questionnaires (FFQ; Willett questionnaire)19 completed by participants at the fifth (1992–1995) and sixth examination cycles (see below).
For the present investigation, we selected offspring cohort partic- ipants who attended any 2 consecutive examinations from the fourth through the seventh (1998–2001) examination cycles. We excluded participants with missing data on covariates (n207) and those with
Assessment of Soft Drink Consumption and Dietary Intake of Other Foods
At the index examinations, participants reported the average number of 12-oz servings of soft drinks (Coke, Pepsi, Sprite, or other carbonated soft drinks, separately categorized into caffeinated or decaffeinated drinks) consumed per day in the year preceding the examination. The responses to the questions were entered as integers (0 or more) separately for caffeinated and decaffeinated soft drinks. This questionnaire (referred to as the “examination cola question- naire”) did not separate nondrinkers from infrequent drinkers (1 drink per day). Accordingly, we compared individuals who reported consuming 1, 1, or 2 soft drinks per day with attendees who reported consuming 1 soft drink per day (infrequent drinkers and nondrinkers, who served as the referent).
Intake of regular and diet soft drinks was assessed from FFQs19 that were administered at the fifth and sixth examinations. We also
Examination 5 (From FFQ data)
Examination 6 (From FFQ data)
1501 = 8997 (1263)
Number attending the examination with 4 year follow-up
History of CVD
Missing covariates at baseline
Sample for cross-sectional 1522 analyses
Prevalent metabolic s yn d r o m e
Missing covariates on follow up
Selection of study sample from baseline exami- nations using the examination cola question- naire and from the sample with available FFQ data (within parentheses, for examinations 5 and 6). Eligible participants and exclusions are indicated in the Figure. CVD indicates cardio- vascular disease.
Sample for prospective
951 = 6039