The user_updates column means the number of times a user action caused a modification of the index, not how many times the index was used to satisfy an update query. If there are more user updates by a significant amount than there are user seeks then the resources used to maintain the index structure may be greater than the benefit the index provides. Glenn Berry, Jason State, and Kendra Little all have scripts on their blogs that do a pretty good job of identifying indexes that may be candidates for dropping or modification.