Background/Purpose: The
presence of rheumatoid factor (RF) and antibodies against citrullinated
peptides (ACPA) is associated with progression of joint damage in patients with
rheumatoid arthritis (RA). RF and ACPA levels may change during the course of therapy,
but it not clear, if such changes in autoantibody levels also reflect a change
in prognosis of RA. Here, we aimed to investigate whether changes of RF and ACPA reflect a true change
in prognosis of RA.

Methods: A cohort of 450 RA outpatients was identified based on
classifiable RA by the 2010 ACR/EULAR classification
criteria, the presence of at least 2 available radiographs of hands and feet 3 years (30
to 42 months) apart, as well as availability of serological tests for RF and
ACPA taken within 3 months prior or after each radiograph. Radiographic images were
scored using the modified Sharp/van der Heijde (SvH) method.

Logistic regression analysis was used to examine the
effect of RF and ACPA at baseline and after 3 years on significant damage progression
(defined as increase of >=5 on the SvH score); results were adjusted for disease activity, expressed as
Simplified Disease Activity Index (SDAI).

Results: At baseline, RF and
ACPA were positive in 60.2% and 61.6%, respectively, of patients. Mean (SD) RF levels were 183.6 (517.2) IU/ml at baseline and 111.2 (247.2) IU/ml after 3 years; ACPA levels were 187.2 (232.6) IU/ml and 145.3 (169.5) IU/ml,
respectively. Over the 3-year course, RF levels changed
by a mean (SD) of 72.4 (409.7) IU/ml (p<0.001) and ACPA levels by 41,9
(120.8) IU/ml (p<0.001). SDAI levels at baseline and after 3 years were 13.8
(11.3) and 8.7 (8.8), respectively.

Univariate logistic regression showed a clear association of
radiographic progression with higher levels of RF at baseline (p=0.039), but
not after 3 years (p=0.515). When adjusting these effects for time-integrated
levels of disease activity, they essentially remained unchanged (p=0.052 and p=0.513,
respectively; table), despite the expected highly significant association of
disease activity and progression. Similarly to RF, the effect of ACPA on
radiographic progression was significant at baseline before and after adjusting
for disease activity, but in contrast to RF, remained so also after three years
(table).

Conclusion: RF levels seem to comprise
reversible and irreversible components. RF is associated with progression of RA,
although removal of the reversible component also eliminated this association.
This association was independent of disease activity during the observed
period. ACPA is poorly reversible and the association with
progression remained unchanged from baseline to year 3.