Inserting multiple rows efficiently with ODBC?

I have a legacy application that used ESQL/C and utilized insert cursors that is being converted to ODBC. I am looking at using SQLBulkOperations, but the documentation indicates that a SQLFetch is the first step in using SQLBulkOperations for inserts. The application's logic retrieves all necessary data, modifies it, deletes all rows in the database that were previously retrieved and then inserts the modified data, so I don't have a valid SQLFetch anymore. Is there anyway to use SQLBulkOperations without first using SQLFetch or will I have to live with removing insert cursors and just doing prepared inserts? TIA Tim Gilbert