It has a argument batch to define categorical batches to remove. It has an argument group for biological variables of interest to preserve. Alternatively, you can use covar_mod to provide a design matrix (model.matrix()) containing your experimental design WITHOUT the batch variable. The latter two arguments are used to preserve the biological effects. That obviously only works if batch is not nested with any of it.

As always when I read "meta-analysis", I say that it's not a meta-analysis to combine data into a single quantitative analysis, as they're not independent anymore. Meta would be to analyze separately and aggregate results, like Fisher, Stouffer, RRA, intersect, whatever.