Bh. Dennis et al., DIET DESIGN FOR A MULTICENTER CONTROLLED FEEDING TRIAL - THE DELTA PROGRAM, Journal of the American Dietetic Association, 98(7), 1998, pp. 766-776
Objective To describe the process and results of diet standardization,
diet validation, and monitoring of diet composition, which were key c
omponents of protocol 1 of Dietary Effects on Lipoproteins and Thrombo
genic Activity (DELTA-1), the initial protocol in a program of multice
nter human feeding studies designed to evaluate the effects of mount a
nd type of fat on lipoproteins and hemostasis parameters in various de
mographic groups. Design DELTA-1 was based on a randomized, blinded, c
rossover experimental design. Three diets were fed for 8 weeks to 103
healthy men and women aged 22 to 67 years at 4 field centers. Diet A,
an average American diet, was designed to provide 37% of energy from f
at, 16% of energy from saturated fatty acids (SFAs); diet B (step 1 di
et) was designed to provide 30% of energy from fat, 9% of energy from
SFA; and diet C (low SFA diet) was designed to provide 26% of energy f
rom fat, 5% of energy from SFA. Key features of diet standardization i
ncluded central procurement of fat-containing foods, inclusion of stan
dard ingredients, precision weighing of foods-especially sources of fa
t and cholesterol-and use of standardized written procedures. Setting
For menu validation, a set of 12 menus for each diet was prepared in d
uplicate and chemically assayed. For monitoring of diet composition du
ring the study, an 8-day diet cycle (6 weekday and 2 weekend menus) wa
s sampled by every field center twice during each of 3 feeding periods
. Statistical analyses Means (+/-standard error) were calculated and c
ompared with target nutrient specifications. Results DELTA-1 was able
to provide a standardized diet that met nutrient specifications across
4 field centers over 24 weeks of participant feeding spanning a total
of 8 months. Applications Prestudy chemical validation of menus and c
ontinuous sampling and assay of diets throughout the study are essenti
al to standardize experimental diets and to ensure that nutrient targe
t goals are met and maintained throughout a controlled multicenter fee
ding study.