Considering oxLDL as the main variable, the sample size
was estimated taking into account a reduction of 14.1 U/L
and an interquartile range of 16.3 U/L, according to the
study carried out by Khan et al. [11].With a bilateral confidence
index of 95% (a Z 0.05) and a statistical power of
80% (b Z 0.80) the sample size was estimated in 44 subjects.
Considering a possible drop-out rate of 15%, the final
sample size was established in 50 subjects. Normality of
the variables was assessed using ShapiroeWilk test. Data
are expressed as mean (SD) for normally distributed variables
or as median and interquartile range for non-normally
distributed. Comparisons between baseline and end
point, and between studies groups were accordingly analysed
by Student paired t-test, Wilcoxon test, independent
t-test or U ManneWhitney, depending on the normality of
the variables. Spearman correlation tests were applied to
evaluate the relationship between changes (D) on oxLDL,
total cholesterol, and glucose levels. Multiple linear
regression analysis was used to assess the effect of cocoa
supplementation over %DoxLDL adjusting for different
models, in which the independent variables were weight,
total cholesterol and LDL-c, differently combined. p < 0.05
was considered significant. The software used was SPSS
15.0 for Windows (SPSS Inc, Chicago, USA).