In my opinion Hans von Bülow's edition is still extremely valuable for the numerous comments he offers on interpretation below the music. This was a very influential edition. However, the text is unreliable.
Evidence of unreliability can be found almost immediately if you try. For example, compare the first page of Opus 101 in von Bülow's edition, available on IMSLP, and in a modern edition. For the latter I'll substitute the old but 'clean' Breitkopf und Härtel edition, because this is also available on IMSLP:
Breitkopf und Härtel:
and the manuscript:
von Bülow's suggestions are not sufficiently distinguished from what was originally present: a large number of extra hairpins, e.g. in measures 3-4, echoing 1-2; in 9-10 in the upper parts, and so on.
There is some hint of the practice of printing suggestions smaller than original markings, but it is too difficult to tell, too often.
Note that both editions amend a tempo present in the autograph to Tempo I. The Arrau edition published by Peters corrects this, and I assume other more modern editions do too. This is very insignificant.
It's also worth noting that these editions typically have an introduction where the editors explain their practice, including how to differentiate any additions from the original. This could be present in von Bülow's edition, and may account for things like pedal markings. However, it can't account for the hairpins.
Here is an especially bad mistake in the Appassionata that occurs in several other editions too:
But look at the manuscript:
The right hand chord is secco. This is a completely different effect.
Another famous mistake, this time from the first movement of the Waldstein:
That should be F flat:
A catalogue of von Bülow's inaccuracies would be quite a booklet, but again I believe the edition is valuable despite these flaws.