Our CSV handling automatically avoids adding an header rows when appending to an existing file. This enables runFrom like actions to gather lists into a single CSV file (with just 1 header). However, non-CSV output does not need or want this, instead it needs the header information for each appended list. For example, the pretty formatting of table output depends on the length of data in each column and that differs between each list. This means it is always expecting header information to process.
Do not combine append with outputType parameters in the same request