The Diet and Exercise for Elevated Risk Trial (DEER) was a one year-long, single-center randomized controlled clinical trial initiated in 1992 at Stanford Medical School’s Prevention Research Center25
. The primary and secondary outcomes were the effects on HDL-C and low density lipoprotein cholesterol (LDL-C), respectively, of a low-fat diet and/or increased physical activity intervention in individuals with elevated cardiovascular disease risk (defined as having low HDL-C and elevated LDL-C). Results of the original study have been published.25
This project utilizes the existing data set and stored baseline and one year plasma blood samples to examine, retrospectively, the effect of the diet and physical activity interventions on hs-CRP.
Specific eligibility criteria were: for men, age 30-64 years, and HDL-C < 45 mg/dL combined with LDL-C 126-189 mg/dL; for women, age 45-64 years, postmenopausal, HDL-C < 60 mg/dL combined with LDL-C 126-209 mg/dL. Exclusion criteria included body mass index (BMI) ≥ 34 kg/m2 for men and ≥ 32 kg/m2 for woman, and for both genders: blood pressure ≥ 160/85 mmHg; fasting triglycerides ≥ 500 mg/dL; fasting glucose ≥ 140 mg/dL; abnormal baseline maximal exercise treadmill test; history of heart disease, stroke, insulin-dependent diabetes mellitus, recent cancer, or diagnosis of a life-threatening disease; neuromuscular or orthopaedic disability which would preclude brisk walking; use of lipid lowering medication or antihypertensive medications; non-euthyroid; low hematocrit; smoking > 9 cigarettes per day; > 4 alcoholic drinks per day; inability to attend sessions; by judgment of a physician; or, unwillingness to accept random assignment to a diet or exercise intervention.
Participant eligibility and baseline information was assessed by telephone and clinic screening prior to randomization. All clinic staff performing measurement were blinded to the participants’ group assignment.
Venous blood was collected in the morning at two separate visits at baseline and at one-year, after participants had fasted, i.e. no food or drink (except water), for 12 hours, no alcohol consumption or vigorous physical activity for 24 hours, and abstained from smoking for one hour. Blood collected for plasma was mixed with 1.5 mg/mL of EDTA, placed on ice before and after refrigerated centrifugation, and alliquoted samples were stored and kept frozen at -80° C.
Plasma high-sensitivity CRP (hs-CRP) concentrations were determined on stored plasma samples randomly chosen from two available samples available at each time point. Immunoturbidimetric assay was performed on the Hitachi 917 analyzer (Roche Diagnostics - Indianapolis, IN), using reagents and calibrators from DiaSorin (Stillwater, MN). This assay has a sensitivity of 0.03 mg/L. The day-to-day variability of the assay at concentrations of 0.91, 3.07 and 13.38 mg/L are 2.81, 1.61 and 1.1%, respectively.
Total cholesterol and triglycerides were measured using enzymatic procedures. 26;27
HDL-C was measured using dextran sulfate—magnesium precipitation28
as well as enzymatic measurement of non-precipitated cholesterol.26
Very low density lipoprotein (VLDL) was calculated as triglycerides divided by five29
, unless triglyceride levels exceeded 400 mg/dL, in which case enzymatic methods were used 26
after ultracentrifugation for 18 hours30
. LDL-C was calculated as total cholesterol minus the sum of HDL-C + VLDL 29
. Lipoprotein values were averaged between the two baseline fasting values.
Blood pressure was measured from the brachial artery utilizing a mercury sphygmomanometer and stethoscope. Averages for two readings of the first and fifth Korotkoff phase were recorded as systolic and diastolic blood pressure.31
Body weight was measured with a standard medical beam balance scale. Height was measured using a Harpenden stadiometer. BMI was calculated as body weight, in kilograms, divided by height, in meters squared (kg/m2
). BMI categories were determined using National Institutes of Health guidelines (NIH): 1) normal weight (18.5-24.9 kg/m2
), 2) overweight (25.0 – 29.9 kg/m2
), and obese (>30 kg/m2
. Waist circumference was taken at the narrowest circumference of the torso when viewed from the front. Skinfold measures were made in triplicate on the right side of the body and averaged. For males, the locations of the skinfolds were chest, abdomen, thigh. For women, the locations were triceps, suprailiac and thigh. Body density was estimated using generalized equations 33;34
and percent body fat was calculated using the Siri equation 35
The presence of metabolic syndrome was determined using the joint American Heart Association/National Heart, Lung and Blood Institute’s guidelines.36
Clinical identification of the metabolic syndrome includes at least three of the following 1) waist circumference > 102 cm in men; > 88 cm in women, 2) triglycerides ≥150 mg/dL, or on drug treatment, 3) HDL-C < 40 mg/dL for men; < 50 mg/dL for women, or on drug treatment, 4) blood pressure ≥130/85 mmHg, or on drug treatment, and 5) fasting glucose ≥ 100 mg/dL, or on drug treatment.36
Eligible participants were recruited and then computer randomized within cohort into treatment groups, using a modified Efron procedure which weighted the probability of assignment in order to balance groups for sample size, HDL-C and LDL-C measures.37
The four treatment groups were: 1) control, 2) diet, 3) physical activity (PA) and 4) diet plus physical activity (diet+PA).
The control group was instructed to maintain their usual lifestyle habits for the duration of the trial.
The dietary goals for the diet group were based on the National Cholesterol Education Program (NCEP) Step II Guidelines 38
: 1) reduce total fat to less than 30% of total calories, 2) reduce saturated fat to less 7% of total calories and 3) reduce dietary cholesterol to less than 200 mg/day. Each participant met with a dietician to individualize dietary recommendations and attended eight group sessions about the NCEP Step II Guidelines.
Participants in the PA group had an individualized physical activity prescription based on the results of a breath-by-breath maximal treadmill exercise test. All PA participants began with six weeks of aerobics sessions three days a week for one hour. After the adoption phase, PA participants were instructed to perform 20 minutes three times a week at 60-85% maximum heart rate, increasing duration over the course of the year to 45-60 total minutes. Progression of the PA program was negotiated between the exercise leader and the participant. PA individuals who were already active prior to randomization were asked to add at least 20 minutes three times a week to their existing programs, in order to elicit physiological changes from an increase in PA. Participants either opted to continue supervised training or adopt a home program for the remaining eight months. The typical PA program involved a minimum of ten miles per week of brisk walking, jogging, or running.
Participants in the diet+PA group received both the diet and PA treatments as individual treatments. To prevent contamination, diet+PA had separate diet and PA sessions from the other groups, and project staff leading the individual sessions made no reference to the other treatment groups.
The diet, PA and diet+PA groups did not emphasize weight loss as an intervention goal.
All intervention methods and laboratory analyses were approved by the Institutional Review Board of Stanford University. CRP secondary data analyses were approved by the Institutional Review Board of the University of Maryland.
All statistical analyses were performed by using SAS software version 9.1 (SAS Institute, Cary, NC). Participants with hs-CRP levels at baseline or follow-up over 10 mg/L were removed from the analysis to eliminate the acute effects of infection (n=9).39
Because the original trial was designed with sex-specific eligibility criteria to identify high risk individuals, it was powered for gender-stratified analyses.25
Because of this, and our a priori hypotheses, all analyses were stratified by gender
Wilcoxon rank sum tests were used to compare hs-CRP baseline level between metabolic syndrome status groups. Analysis of covariance (ANCOVA) was used to test the effects of diet and PA on hs-CRP change (ΔCRP) between treatment groups for: men overall, women overall, men with metabolic syndrome, women with metabolic syndrome, men without metabolic syndrome and women without metabolic syndrome. Older age is a known influence for both metabolic syndrome3
. Furthermore, hs-CRP also increases with metabolic syndrome status4
, thus it was necessary to account for these differences in the analysis. The analyses in men and women with metabolic syndrome are exploratory due to the limited sample size in each intervention group.
ΔCRP was calculated as a difference between the follow-up and baseline value and was a normally distributed variable, thus a transformation was not necessary. Differences between treatment groups for ΔCRP were compared for: 1) control versus diet, 2) control versus PA, 3) control versus diet+PA, 4) diet versus diet+PA and 5) PA versus diet+PA. An α level of 0.01 was adopted to control for type I error and to account for the multiple statistical comparisons. ANCOVA analyses controlled for baseline hs-CRP, cohort, baseline body fat (%), change in body fat (%), cigarettes per day, alcoholic drinks per day, age and for women, menopausal hormonal therapy (MHT) status. hs-CRP was significantly higher at baseline and follow-up for women on MHT, thus it was included as a covariate. However, change in CRP, was not significantly different between women with or without MHT, thus MHT-stratified analyses were not warranted. Despite weight loss not being a focus of the current intervention, preliminary analysis found significantly greater weight loss, BMI reductions, and percent body fat loss in the diet and diet+PA groups relative to control in both men and women. To control for group differences and to represent changes in body composition resulting from weight loss, percent body fat and baseline body fat were chosen as covariates for the models. Analyses were also run without the covariates for body fat.. In the case of significant differences in ΔCRP between intervention groups, a two-way ANCOVA was analyzed in which diet (yes/no), PA (yes/no), and its interaction were used to distinguish the most important lifestyle component(s).