Progressive radiographic damage and deterioration in physical function occur in patients with RA who do not receive timely effective treatment. Several long-term prospective studies have looked at the relationship between radiographic damage and physical function in patients with RA.^8–10 A prospective 6-year follow-up study by Kuper et al¹⁰ showed that large joint damage occurs early in the disease, with at least 20% of patients having damage in at least one joint within 1 year of disease onset and 50% of patients having some damage within 6 years of disease onset. Radiographic changes were significantly related to damage in the hands and feet, the physical disability index and cumulative disease activity.¹⁰ In a review of the links between joint damage and disability in patients with RA, Scott et al reported that joint damage progresses continuously over the first 20 years of RA and that this damage accounts for 25% of the disability seen in patients.⁹ Radiographic damage is thus clearly a major determinant of long-term physical function in patients with established RA.

Although studies have shown that a relationship between physical function and radiographic damage exists, they have mostly examined this association looking at radiographic damage over the long term. It is not known whether progression of radiographic damage over short periods is associated with immediate impairment in physical function. This question is particularly relevant today because of the availability of powerful treatments that can prevent radiographic progression if started early enough. For example, in clinical trials, etanercept, either alone or in combination with methotrexate, has been shown to reduce radiographic progression in patients with recent and long-term RA.^11–13

Our study used data from the Trial of Etanercept and Methotrexate with Radiographic Patient Outcomes (TEMPO),^{11 14} to investigate the longitudinal relationship between physical function and both level of radiographic damage and radiographic progression over a 2-year period, in a cohort of patients with early or advanced active RA.

All patients were randomly assigned to one of three treatment groups: etanercept (25 mg twice-weekly subcutaneous doses), methotrexate (7.5–20 mg weekly oral doses) or combination therapy with etanercept and methotrexate.

Physical function was measured by the Health Assessment Questionnaire (HAQ) at baseline and at protocol-specified intervals throughout the study; scores at baseline and 1 and 2 years were used in the analysis. The HAQ is scored on a scale of 0–3, with higher scores indicating increased disability.

Radiographs of the hands, wrists and forefeet were taken at baseline, 6 months, 1 year and 2 years. Digitised images were scored by two readers, using the van der Heijde-modified Sharp method.¹⁷ Radiographic progression was determined for each interval by subtracting status scores of two time points. Radiographic progression was divided into four categories (negative, <0; zero, 0–1; minor, 1–5; and greater progression, >5) for comparisons with HAQ scores. The radiographic population consisted of patients who had an acceptable baseline and at least one acceptable post-baseline film.

In the linear mixed model, which was applied here in order to aggregate data of baseline, year 1 and year 2 data under simultaneous adjustment for within-patient correlation, the TSS was significantly positively associated with the HAQ score independently of the DAS (table 1). A number of additional variables were independently associated with the HAQ score, including age (positive correlation), sex (women had higher HAQ scores) and the DAS (positive correlation).

The HAQ score is primarily determined by disease activity. In the context of a clinical trial, it is expected that the disease activity would show important variation that would easily obscure a contributory and independent association with radiographic damage or progression. We first investigated the relationship between the DAS and the HAQ score. After adjustment for age and sex, the relationship between DAS and the HAQ score was almost linear (fig 1). This linear relationship justifies why in every subsequent analysis to determine the contribution of radiographic damage to explaining physical function we have adjusted for the known relationship between the DAS and the HAQ score.

To visualise the adjusted relationship between the HAQ score and the Sharp score, Sharp scores were divided into six categories of 10 Sharp units each and estimated marginal means were calculated. Figure 2 shows estimated marginal means of the fitted model (mean HAQ scores adjusted for age, sex, DAS, treatment and time) for each category of 10 Sharp units and clearly points to an increasing trend. A separate analysis, in which the Sharp score was replaced by the change in Sharp score, also indicated that the change in Sharp score (interval progression; Sharp units per year) was significantly and independently associated with the HAQ score (p=0.001; table 2). This implies that patients who show joint damage progression have worse physical function, independent of disease activity and the type of treatment they use, although the effects were small.

Figure 3 shows the effect of changes in radiographic damage on physical function, with progression of radiographic damage stratified into four categories (negative, zero, minor and greater progression). HAQ scores tended to increase with increasing radiographic progression, although the differences were small and probably not clinically meaningful. Figure 3 also suggests that patients with negative progression scores have lower HAQ scores than patients with positive progression scores, but this effect was not statistically significant.

In a 10-year longitudinal study,¹⁸ we have previously observed that radiographic damage and disease activity independently contribute to changes in physical function in RA regardless of disease duration. Others have reported the longitudinal relationship between joint damage and physical function after adjustment for disease activity.¹⁹ In that particular study, performed on an inception cohort of patients with RA followed up for many years while receiving standard conventional antirheumatic treatment, the main driver of physical function was determined to be disease activity. But Welsing et al also found a significant contribution of joint damage (measured by van der Heijde’s modification of the Sharp score) to explain the variation in physical function.¹⁹

There are important differences between these previous studies^{18 19} and our current study: First, we analysed a trial population using highly effective treatments. In our study, radiographic progression was minimal, although the level of baseline damage was, on average, substantial. Apart from that, the active trial treatment resulted in very low levels of disease activity in the majority of the trial population. Even with such a low level of variation in the determinants of function (disease activity and radiographic damage/progression) a relationship between radiographic damage progression and the HAQ score was suggested by this analysis. Second, in our study, we investigated the contribution of radiographic progression itself, rather than radiographic damage at any particular time point. Our analysis suggests that both the level of damage, and the progression rate, although very low, contributed to variation in physical function. Third, our observations were made in a 2-year follow-up study, whereas the previous studies^{18 19} investigated function over a period of 9–12 years, in which deterioration of physical function is likely to be far more substantial. Taken together, however, the longitudinal studies provide evidence that physical function in patients with RA is determined not only by disease activity but also by radiographic damage and the rate at which radiographic damage increases over time. Fluctuations in the progression rate and in the disease activity cause demonstrable effects on physical function within very short time intervals. These data add to the ever-growing evidence that it is important not only to suppress disease activity but also to halt radiographic progression to maintain physical function.

It is not a coincidence that studies using longitudinal data analyses are consistent with the existence of a relationship between radiographic damage and physical function. It has been hypothesised that this relationship exists, but it has turned out to be rather difficult to confirm it statistically in studies using cross-sectional analytical approaches.^{20 21} A longitudinal data analysis provides far more statistical power because it uses all available data of a prospectively followed cohort in an aggregated manner, while adjusting for spurious intra-patient correlation. Apart from that, longitudinal data analysis provides insight into how changes in disease activity and radiographic progression—for example, as a result of effective treatments, influence physical function.

Physical function is often considered to be a rather static outcome. However, recent experience from randomised clinical trials evaluating highly effective treatments has shown that impairment of physical function in patients with RA is reversible to some extent. In a recent analysis, Aletaha et al reported that the irreversible part of the HAQ score was between 0% and 33%, increasing with the duration of RA.²² We hypothesise that this irreversible impairment is caused by structural damage, while the reversible impairment is primarily inflammation. If true, early effective intervention to limit structural damage may be helpful in decreasing irreversible physical impairment.

Many studies support the use of early treatment of RA. Landewé et al showed that the best way to prevent radiographic damage is to start treatment before such damage has occurred.²³ If this is not possible, aggressive treatment should be started within 12–24 months after the diagnosis of RA to limit radiographic progression over time.²³ Anderson et al showed that the rate of radiological progression is set during the early stages of RA and suggested resetting this progression rate through pharmacological treatment as soon as possible after the disease is recognised.²⁴ Other studies support the finding that radiographic damage occurs early in the onset of RA and progresses throughout the disease.^{2 3 25–27} A recent meta-analysis showed that starting treatment early has a long-term benefit on inhibition of structural damage.²⁸ This analysis demonstrated that the association between radiographic progression and physical function was similarly important in relatively early RA (disease duration 3 years) as compared with more advanced RA (disease duration >3 years), which emphasises the importance of minimising radiographic progression at all stages of the disease.