The case against seronegative rheumatoid arthritis