Its an interesting question and there are a few things to consider.
First, Japanese have extensive records on battle casualties from Nanbokucho to late Muromachi, from various sources. In Nanbokucho for example after sometime in service you would write a letter detailing your wounds, lost equipment, kills, tropheys taken. Which today produces many Ph.Ds dealing with these letters.
By memory swords in their most effective periods go to like 35% of total casualties, but generally oscillate between 5 and 20%. The rest are bows, later yari etc.. In Nambokucho there is surprising percentage of stones, axes and other tools that are seldom imagined.
Also, most surviving swordsmen (i.e. capable of leaving a record) even in Nambokucho would have zero kills, and people with five+ claimed kills would be around 5% of the fighting force.
Second, swords overall are not made to survive battles. European sword some people estimate to survives 3 active engagements on average. Japanese are generally much more prone to hagire and catastrophic chipping so it is comparable at best. However, we are dealing with survivors - so either blades which never killed and thus were preserved, or the blades which somehow killed successfully.
So Edo period sword obviously probably never killed. Muromachi - we are talking about huge supply of swords, yet 5-10% sword casualty rates and most sword construction being cheap and iffy enough to be smelted if it kills and chips rather than being repolished. Its probably sub 1% chance it killed.
Nanbokucho - there is actually a solid chance it did.