Two curves were assessed by the
two independent sensors in each test cell. The parameters calculated by the software
were the mean values of the parameters determined with each curve. Repeated
measurements were advised when the coefficient of correlation was 20% of the mean were observed. The parameter
that reflected the overall platelet activity was the area under the aggregation curve,
which was displayed by the device in units (U).
Statistical analyseswere performed with the BMDP program package (Dixon, 1992).
For the majority of variables, data descriptions represent the arithmetic mean and standard
deviation (SD). After testing for deviations from the normal distribution using
the normal probability plots of the residuals (Q-Q-plot), one-way repeated measures
ANOVAswere performed after logarithmic respectively square root transformation
of the data if required. When the ANOVAs showed significant changes over time, post
hoc Dunnett’s tests were used to compare each time point to T0. For each variable,
the level of statistical significance was set at P ≤ 0.05. An a priori power analysis using
the program BiAS for Windows (Version 9.08) revealed a sample size of 10 horses
needed to detect an expected mean reduction of 70% in platelet aggregation 24 h after
the loading dose of clopidogrel with a probability of 0.95 and an accuracy of ±10%.