vacuumdb: ERROR: Invalid UNICODE character sequence found

I get this error when I try to analyze a column containing md5 hashes.

The data type of the column is: bytea. The database encoding is UNICODE.

Does anybody know a workaround for this?

Thanks.

$ vacuumdb --analyze --table 'files(md5hash)'
ERROR: Invalid UNICODE character sequence found (0xdb51)
vacuumdb: vacuum files(md5hash) gutenberg failed

$ psql --version
psql (PostgreSQL) 7.3.3
contains support for command-line editing

gutenberg=> SELECT version();
version
-----------------------------------------------------------------
PostgreSQL 7.3.3 on i686-redhat-linux-gnu, compiled by GCC 2.96
(1 row)

--
Marcello Perathoner