Clustered PK is a table because all data from the table stored in it. Other indexes point to data through clustered index. When you drop clusterd PK, whole table and all its indexes have to be rebuild.

>Primary keys are clustered. There are other indexes.
>
>Documentation says that the drop of a primary key will create an overhead as SQL Server has to reorg its way of finding records. I ended up doing this:
>
>

>ALTER INDEX [ClientNumero] ON [Client] REBUILD WITH (ONLINE=OFF)
>

>
>This allows to do it direclty without a drop, with an online clause to be off to provide more performance.