Recently, we have reported from the BARFOT early (duration <1 year) rheumatoid arthritis study that over the first 5 years women had, compared with men, a similar degree of joint destruction despite a higher Disease Activity Score calculated on 28 joints (DAS28) and Health Assessment Questionnaire (HAQ).1 However, a recent study of another rheumatoid arthritis cohort demonstrated that female gender was an independent predictor of radiographic progression over 10 years.2 ,3 A possible explanation for this discrepancy may be the different length of follow-up. Therefore, we have extended our study to a follow-up time of 8 years.

In all, 416 of the initial 549 patients were assessed. Of these 64.7% were women, of whom 54.4% were anti-cyclic citrullinated peptide antibody positive and 58.2% were rheumatoid factor positive versus 59.2% and 60.8% of the men, p=0.38 and p=0.62, respectively. There were no significant differences between women and men in the treatment with glucocorticoids and disease-modifying antirheumatic drugs during these years.

The radiographs of hands and feet were scored by the van der Heijde modification of the Sharp score. As shown in table 1, even after 8 years there were no significant differences between the sexes in total score, erosion score or joint space narrowing. At 8 years, radiographic progression (2 units/year) had occurred in 43% of the women and 40% of the men, p=0.51 and 62% of both female and male patients had developed one or more erosion (p=0.97).

From table 1 it is also evident that DAS28 was still significantly higher in women than in men after 8 years. This was as before owing to higher scores for the DAS28 components general health and tender joints, whereas erythrocyte sedimentation rate and number of swollen joints were similar. Likewise, HAQ was significantly higher in women, whereas signals of functional impairment were similar.

In a multiple logistic regression model, total Sharp van der Heijde score at baseline and presence of anti-cyclic citrullinated peptide antibody were independent predictors of radiological progression, whereas female sex was not (table 2).

The fact that sex was not an independent predictor is at variance with the study by Syversen et al.2 In their study the range of disease duration at study start was quite large, up to 4 years, resulting in different total scores between progressors and non-progressors already at inclusion. This disparity in study design might explain our different findings, and this was further confirmed in a smaller study from the same group, in which female gender was not proved to be an independent predictor of radiological damage either at 5 or at 10 years.4

Women had worse DAS28 and HAQ than men also after 8 years, suggesting a more severe disease course in women. However, this is in conflict with the observation that joint destruction was similar in women and men. This contradiction seems to be explained by the possibility that women had scored worse in the DAS28 owing to higher pain perception. Higher DAS28 in women has also been described by others,5 and might clarify why remission is less common in women after 2 and 5 years of disease when the DAS and DAS28 remission criteria are used.6 ,7 In conclusion, the notion that women fare worse than men is not confirmed by this study. The issue might be resolved by applying DAS28 independent criteria for disease activity and remission to these data.