Patients and methods

A total of 549 Caucasian patients with RA (343 women and 206 men), with disease duration less than 1 year and included into the BARFOT (for “Better Anti-Rheumatic FarmacOTherapy”) study between October 1993 and February 1998, were recruited to the present study. At the 5-year visit the number of patients were 480 (306 women and 174 men).

All patients gave their informed consent and the local ethic committees approved the study, which was performed in accordance with the Helsinki Declaration.

Assessments

The following assessments were performed at inclusion in the BARFOT study (baseline) and at follow-up visits at 1, 2 and 5 years.

Disease activity

This was measured by DAS28.1 DAS28 between 2.6 and ⩽3.2 is considered as low disease activity, between >3.2 and ⩽5.1 as moderate activity and above 5.1 as high activity. Patient self-assessed pain was scored on a visual analogue scale (VAS) (0–100 mm).

Disability status

This was reported by the patients using the Swedish version of the HAQ.2 It consists of 20 items in 8 categories of activities of daily living (ADL). Total HAQ score ranges from 0–3.

Functional impairment

This was scored using the SOFI instrument,9 where assessors examine four items of hand function, three of arm function and four of leg function. The total SOFI score ranges from 0–44 and the domain scores for hand, arm and leg range from 0–16, 0–12 and 0–16, respectively.

Radiographic joint damage of hands, wrists and feet

This was scored according to the Sharp method as modified by van der Heijde,10 which allows for separate presentation of total score (TotSharp; range 0–448), erosion score (ES; range 0–280) and joint space narrowing score (JSN; range 0–168). Specially trained readers (KA and KF) assessed the radiographs in chronological order.

Statistics

The statistical analyses were performed using SPSS V.16.0 statistical software (SPSS, Chicago, Illinois, USA). To test differences between groups, the Mann–Whitney U test or the independent samples t test was used for continuous variables and the χ² test for proportions. For paired samples the Wilcoxon signed rank test was used. Correlations were examined by the Spearman rank correlation test. Area under the curve (AUC) analyses were calculated as summary measures integrating serial assessments of DAS28, HAQ and SOFI scores over the duration of the study. Analysis of covariance (ANCOVA) was used to correct for difference in mean age between groups. All significance tests were two-tailed and were conducted at the 5% significance level.

Results

The women were significantly younger than the men, mean (SD) age 54 (16) and 61 (13) years, respectively. Women had higher pain VAS, DAS28 and HAQ scores than men, whereas men had higher C-reactive protein (CRP) levels and SOFI scores; see tables 1 and 2.

Disease activity and pain

As shown in table 1, DAS28 had improved in both genders at 1 year and was then stable. However, the mean DAS28 was significantly higher in women than in men at all time points, also when adjusted for age.

As to the components of the DAS28, women had more tender joints at all time points compared with men and also worse scores of general health up to 1 year. The number of swollen joints was similar. Pain VAS scores were significantly higher in women at all time points but one.

Disability

HAQ scores decreased at follow-up in both genders, but were around 0.2 score units higher in women (p<0.001) at all time points, table 2. The higher HAQ scores for women depended mainly on the categories of eating, walking, reach, grip and additional common daily activities.

Functional impairment

As shown in table 2, men scored worse than women for total SOFI due to higher scores on SOFI hand and arm. The differences between genders were significant during the first 2 years. At 5 years, only SOFI arm was worse in men.

Radiographic joint destruction

The radiographic scores increased over time with no statistical differences between women and men at any time point (table 3).

Correlations between DAS28, HAQ and SOFI scores and radiographic joint destruction

In women and men, the DAS28 AUC correlated significantly with TotSharp at 5 years, r = 0.316 and r = 0.313, respectively, both p<0.001. Looking at the specific components, such correlations were found only for erythrocyte sedimentation rate (ESR) (r = 0.496 and r = 0.298) and swollen joints (r = 0.482 and r = 0.488), but not for tender joints (r = 0.022 and r = 0.027) or for general health (r = 0.06 and r = 0.176).

The HAQ AUC did not correlate significantly with TotSharp at 5 years for any gender. The SOFI AUC correlated significantly with TotSharp at 5 years in women, total SOFI, r = 0.384 and its components, r = 0.248 to 0.371, all p<0.001. In men neither total SOFI AUC nor its components correlated with TotSharp.

Medical treatment

At baseline, 73% of the women and 75% of the men were given disease-modifying antirheumatic drugs (DMARDs), mostly sulfasalazine or methotrexate. The treatment strategy changed to more methotrexate during the study and drug combinations and biologicals were introduced. This did not affect gender differences of outcome measures. Overall there were no differences in drug treatments between genders at any time point (p = 0.11 to 0.59).

Treatment with prednisolone was given at inclusion to 50% of the women and 55% of the men and at 5 years to 32% of the women and 31% of the men (not significant).

Discussion

During the first 5 years of RA women had higher DAS28 compared with men, mainly depending on worse scoring of tender joints and general health. Furthermore women scored around 0.2 units worse on the HAQ than men at all time points. Men had worse SOFI upper limb scores than women, implying greater functional impairment.

The higher DAS28 in women confirms earlier reports.3 4 6 This gender difference has been suggested to imply that women have a more aggressive disease. However, in the present study, women and men had similar total Sharp scores up to 5 years, which is in agreement with reported similar x ray scores after 2 years of RA by the Larsen assessment.4

Analysing the components of DAS28, both genders had similar swollen joint counts and, interestingly, swollen joints and the ESR AUC correlated significantly with TotSharp. However, tender joints and general health, which caused the higher DAS28 in women, did not correlate with TotSharp but with pain. Pain perception is more pronounced in women than in men with RA,6 which has also been found in the general population.11 The gender difference in DAS28 might therefore be caused by higher pain perception in women and need not solely be related to inflammation. Pain perception as well as the immune system are modulated by the gonadal hormones.12 13

In agreement with others,5 the women scored worse on total HAQ. A higher HAQ in women with RA, also found when adjusting for disease severity and physical function,14 has given rise to several theories. One is that men with RA underscore their disability,2 verified later as men with RA overestimated their functional ability by 0.21 HAQ units,7 which is the same gender difference in HAQ between genders found here. Of interest, healthy women and men have been reported to score HAQ similarly.15 16 Another explanation might be the impact of pain on HAQ scores,17 18 especially pronounced in women.18

A further explanation of the more favourable HAQ results in men could be greater male muscle strength, making it easier to perform daily tasks. Accordingly, muscle strength has been shown to have a major impact on the HAQ.8 Thus, when divided into subgroups with equal grip strength, women and men with RA had corresponding degrees of activity limitation as measured by the HAQ.19

Of interest, total SOFI score was worse in men than in women in agreement with an earlier report.20 SOFI includes assessments of range of movement, which have been found more limited in healthy men than in healthy women and thus suggested to be constitutional.16 It thus seems as if men with RA can balance their lower range of upper limb movements by superior muscle strength and less pain when reporting functional limitation as in the HAQ. This explanation is supported by the fact that only the SOFI AUC for women correlated with TotSharp in this study.

In summary, this study has shown that DAS28 and HAQ are worse in women with RA compared with men, in spite of similar radiographic destruction. This is suggested to depend on the fact that women experience more pain, which implies that they need to be treated in a more multimodal way than men with RA. This also calls for revision of current evaluation instruments with special attention to gender differences.