Return to search

Correlation of habitual diet with plasma risk factors for coronary heart disease

The statistical correlations between habitual diet and plasma
risk factors for coronary heart disease CHD were analyzed using
multiple regression. Thirty-one male subjects between 30-56 years
kept complete dietary records for 7 days. Daily means of nutrient
consumption were calculated using a computerized data base.
Fourteen independent variables (total kilocalories, protein %,
carbohydrate %, fat %, ethanol %, caffeine, P/S ratio, cholesterol,
age, weight, height , weight**.75, Body Mass Index,
kilocalories/weight**.75) were created. Plasma samples were analyzed
and the following simple (total plasma cholesterol (TC), VLDL-C,
LDL-C, HDL2-C, HDL3-C, apo A-I, apo A-II, apo B) and derived
(VLDL-C+LDL-C, LDL-C/TC, LDL-C/HDL-C, HDL-C, HDL2-C/HDL3-C, HDL-C/TC,
apo B/apo A-I, apo B/apo A-II, apo A-Il/apo A-I) dependent variables
were created. Dependent variables were individually regressed against the entire set of independent variables. An F-value of 4.00
to enter an independent variable in the model and of 3.99 to remove
one were used to achieve significance of p<05.
Age appeared in 5 regression models (TC, apo B, apo B/apo A-I,
apo B/apo A-II, apo A-Il/apo A-I) and was positively correlated with
increased risk for CHD. Total kilocalories appeared in 4 models
(LDL-C, apo B, VLDL-C+LDL-C, LDL/TC) and was negatively correlated
with risk. Fat % appeared in 4 models (VLDL-C+LDL-C, LDL/TC,
LDL-C/HDL-C, HDL-C/TC) and was associated with increased risk. Body
Mass Index was entered in 2 models (HDL2-C, HDL-C) and was positively
correlated with risk. P/S ratio was negatively correlated with risk
in the three models (HDL2, apo A-I, HDL-C) in which it appeared. The
independent variable carbohydrate % was negatively associated with
risk in 2 models, LDL-C and apo B. Kcal/wt**.75 was also negatively
correlated with risk in the VLDL-C, LDL-C/TC and HDL-C/TC models.
One independent variable, EtOH %, was positively associated with risk
in the apo A-II and apo A-II/apo A-I models. Two dependent variables
did not have any independent predictors (HDL3-C, HDL2-C/HDL3-C)
entered in their regression models.
Six independent variables did not appear in any regression
model (protein %, caffeine, cholesterol, weight, height,
weight**.75). Independent variables positively correlated with
increased risk for CHD were therefore fat %, ethanol %, age, and Body
Mass Index. Independent variables correlated with decreased risk for
CHD were total kilocalories, carbohydrate %, P/S ratio, and
kilocalories/weight**.75. / Graduation date: 1985

Identiferoai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/27429
Date08 August 1984
CreatorsBills, Nathan D.
ContributorsOh, Suk Y.
Source SetsOregon State University
Languageen_US
Detected LanguageEnglish
TypeThesis/Dissertation

Page generated in 0.0019 seconds