But there is also this paragragh in the "analysis of clinical endpoints" section -
"The change of clinical scores from baseline to Week 48 was analyzed as the dependent variable, with treatment and visit week as fixed effects, treatment-by-visit as interaction effect, and baseline score, country, baseline concomitant AD medication, natural logarithm of the baseline plasma Nf-L concentration, baseline MMSE status, and SIGMAR1 receptor gene variant genotype status (single-nucleotide polymorphism SIGMAR1 rs1800866 presence or absence) as covariates in the model. For CGI-I, baseline CGI-S score was used as baseline. The primary comparison was the contrast (difference in the least squares mean) between blarcamesine and placebo at the last visit (Week 48), which was performed for the active treatment group as well as separately for the assigned (30 mg and 50 mg) treatment groups. For the primary analysis, the model assumed the missing data to be missing at random without imputation."