Statistical Analysis
Estimation of parameters and statistical analyses on PK and PD data were
conducted using non-compartmental analysis in Phoenix® WinNonlin®
version 8.2 (Certara USA Inc, Princeton, NJ) and SAS® version 9.4 (SAS
Institute, Cary, NC). For the additional statistical analyses, SAS® 9.4
was also used.
Descriptive statistics were calculated according to the variable nature:
i) for continuous variables, number of subjects (n), arithmetic mean,
standard deviation (SD), standard error of the mean and range (minimum
and maximum); median, quartiles, coefficient of variation (CV%),
geometric mean, geometric standard deviation, and geometric CV% are
given, if pertinent; ii) for categorical variables, number of subjects,
relative and absolute frequencies.
Actual p-values were reported, no multiplicity or adjustment of type I
error was implemented; therefore, an alpha of 0.05 (p≤0.05) was set for
statistical significance. Hypotheses testing was defined to
statistically compare the main primary PD endpoints of the study with
zamicastat and placebo (SBP, DBP, and HR), with clinical margins of 5.5
mmHg, 3.7 mmHg and 4.1 bpm for mean SBP, DBP, and HR, respectively.
Hence the null hypotheses were as follows: