Disease activity (including remission)

There is evidence of a secular trend of improvement in disease activity in the routine clinic. Pincus et al. conducted two cross-sectional studies over the period of 1 year in their weekly clinic at the Vanderbilt University, Nashville, TN, USA – one in 1985 and one in 2000. All patients were assessed in a similar comprehensive fashion and were managed according to the treatment paradigms of the day . During that time the median (interquartile range, IQR) joint swollen count (out of 28) fell from 12 to 5; the median (IQR) erythrocyte sedimentation rate (ESR) from 33 (16–50) to 20 (9–33) and the median (IQR) DAS from 5.7 (4.9–6.5) to 4.4 (3.2–5.3).

There is also evidence of decreasing disease activity at baseline in inception cohorts of early RA patients. The Nijmegen group has reported falls in the baseline swollen joint count, ESR and DAS28 successive quinquennia since 1985 ( p < 0.0001) in patients recruited to their early arthritis clinic ( Table 1 ) . In Jyväskylä Central Hospital Finland, three cohorts of patients with early RA were recruited in 1983–85 ( n = 58), 1988–89 ( n = 77) and 1995–96 ( n = 62). Although the median ESR fell, the median swollen joint count did not ( Table 1 ).

At 5 years, the DAS28 had fallen within each of the Nijmegen cohorts and, with time, the 5 year outcome had improved significantly from 1990 to 2005 ( p = 0.014) ( Table 2 ).

Current treatment aims are early suppression of disease activity aiming for remission, hopefully leading to improved long-term effects such as decreased functional disability. Evidence shows that, with the passage of time, increasing proportions of patients are achieving remission on treatment with the biologic agents – perhaps reflecting their introduction progressively earlier in the disease course . Two studies report an association between the number of times a patient was in remission and Stanford Health Assessment Questionnaire (HAQ)-score measured up to 5 years after inclusion. The first study included patients with a mean disease duration of 3 years at baseline and examined HAQ-score after three years . The second came from the Norfolk Arthritis Register (NOAR), which is an inception cohort study of patients with recent onset inflammatory polyarthritis (IP). NOAR was established in 1990 and has been recruiting patients continuously since that time. At the time of the baseline assessment and at each subsequent annual assessment, the patient undergoes a joint examination and completes the British version of HAQ . This study reported an association between the number of times that a patient was in remission during the first 3 years of disease and the HAQ-score 5 years after inclusion into the register . In this study of early disease there was no association between the time until remission and subsequent functional disability in the patient populations recruited either from 1990 to 1994 or from 2000 to 2004. However, another study showed that, despite clinical remission, average HAQ-score increased significantly with increasing disease duration and the irreversibility of disability . Taken together, these studies suggest that, in early disease, clinical remission may be associated with long-lasting benefit in terms of function. In established disease, clinical remission may slow the rate of accumulation of physical disability but if there is already irreversible damage this is likely to progress despite disease suppression.

Physical function

There is evidence of substantial improvements in outcome with regards to physical disability up to around the year 2000. Krishnan and Fries examined functional disability data from 3035 patients with RA whose disease onset was from 1977 to 1998. Disability data were collected twice yearly using the HAQ disability index. Average disability declined by about 2–3% per calendar year of disease onset. This trend was consistent by age, sex, race, disease duration, clinical centre and baseline disability. After accounting for potential confounders, average disability levels in RA had declined by approximately 40% in the 20 or so years since 1977 . Pincus reported an improvement in the median (IQR) modified HAQ-scores measured cross-sectionally in his Vanderbilt University clinic from 1.0 (0.6, 1.4) in 1985 to 0.4 (0.1, 1.0) in 2000 . The median (IQR) disease duration was 7 (2, 15) and 9 (4, 18) years, respectively.

However, in patients with more recent onset of disease, improvements in physical disability are more difficult to discern. NOAR offers the opportunity to compare the HAQ at baseline in patients with symptom onset before and after the biologic era. The median baseline HAQ-score for patients recruited between 1990 and 1994 was (0.75 IQR 0.25, 1.38). It was somewhat higher in the cohort of patients recruited between 2000 and 2004 (median 0.88, IQR 0.38, 1.63) ( Table 3 ). A similar pattern was seen in the Nijmegen clinic in which, despite a significant fall in DAS28 at presentation, the median HAQ actually rose from 0.54 (IQR 0.2, 1.04) in 1985–1990 to 0.91 (IQR 0.5, 1.68) in 2000–2005 .

Five years after inclusion into the NOAR cohort, median (IQR) HAQ-scores were 0.75 (0.13, 1.63) for the 1990–1994 recruits and 1.0 (0.25, 1.75) for the 2000–2004 recruits. The Nijmegen group found that, although the HAQ improved from baseline in each cohort, the degree of improvement fell with time and the 5 year functional outcome was worse in recent time (although the differences were not significant) ( Table 2 ). Thus, physical disability does not seem to have improved either at presentation or at 5 years in recent times despite increasing use of DMARDs and MTX over time ( Table 4 ).

As the HAQ is calculated from a patient-completed questionnaire, it is possible that it is the perceptions of patients with IP that have not changed rather than their actual physical capabilities. To gain better insight into the secular changes in HAQ-score at the same time point in disease course, it is useful to compare changes in the severity of functional disability in patients with RA with those of the general population over time. In a Finnish study, data on HAQ-scores obtained in 2000 and 2005 from 863 patients with RA and 1176 community controls (age-and gender-matched) were evaluated . In both the patient and general population, the increase in mean HAQ-score was 0.01 units year ⁻¹ . The change in mean HAQ-score was mainly attributable to changes in persons older than 70 years. A similar result was observed in the NOAR cohort in which patients, both men and women, with a late onset of IP (age > 75 years) had worse functional ability compared with patients aged <75 years . In the NOAR, the Nijmegen and the Finnish studies, the mean age at IP presentation has been slowly increasing ( Tables 1 and 3 ). This may reflect the changing demography of society rather than any change in the age-specific incidence of IP. The changes in HAQ-score, especially in older people with or without RA, over time may be explained partly by increasing number of co-morbidities with advancing age. Radner et al. specifically looked at the impact of co-morbidity on functional disability and found a significant increase in HAQ-scores in RA patients with increasing levels of co-morbidity (adjusted for disease activity, gender and disease duration) .

Radiographic progression

In the past, the presence of radiological erosions was regarded as one of the hallmarks of RA. Now increasingly patients are diagnosed with RA before any erosions have appeared and the goal of current treatment strategies is to prevent the development of erosions. Patients may now be many years from symptom onset before erosions first appear. There is evidence of a secular trend in the prevalence and severity of erosive damage. In Finland, three cohorts of patients with early RA recruited at the Jyväskylä Central Hospital from 1983–85, 1988–89 and 1995–96 ( Table 1 ) were monitored prospectively for 5 years . X-rays of the hands and feet were taken at baseline and at 5 years. The Larsen score increased by a median of 12, 6 and 4 points by year 5 in the first, second and third cohorts respectively after adjustment for age, sex, rheumatoid factor (RF) and baseline Larsen score and ESR. As in all observational studies, it was difficult to distinguish whether these observations were the effect of improved natural history or of treatment. Very few patients in the first cohort received MTX, whereas 70% of those in the third cohort did. A more sophisticated study was published in 2006 that examined radiographic progression in 418 patients with an onset of RA in the 1970s, 1980s and 1990s, who were first seen at the Wichita Arthritis Center in Kansas, USA within 3 years of symptom onset . They found a significant trend towards less radiographic progression in more recent decades in the crude analysis. However, after adjustment for DMARD and steroid use and baseline predictors, the differences between the decades disappeared, suggesting that the observed improvement in radiographic outcome is attributable to effective treatment rather than to secular trends.

Extra-articular features

Extra-articular features of RA include nodules, vasculitis, neuropathy, pericarditis, interstitial lung disease (ILD) and eye involvement. There appears to have been a marked decline in the incidence of rheumatoid vasculitis in the last 10–20 years. Watts et al. concluded that there was no fall in the incidence of systemic rheumatoid vasculitis in the UK between 1980 and 1994 and Turesson et al. found no decline in vasculitis in the Mayo Clinic population in the first 10 years of disease in those first diagnosed from 1975 to 1994 . However, by 2004, Watts et al. had observed a sharp decline in the incidence of rheumatoid vasculitis in the Norfolk population from 15.2 per million (95% confidence interval (CI) 9.1, 23.8) in 1992–94 to 3.0 per million (95%CI 0.8–7.8) in 1998–2000 . In a study from the US Veteran’s Health Administration, which includes approximately 37,000 RA patients at any one time (92% male), the prevalence of rheumatoid vasculitis fell from 4.4% in 1985 to 3.3% in 2006 of RA outpatients and from 4.7% to 1.4% respectively of RA inpatients .

Turesson et al. reported no decrease in other extra-articular features of RA in the Mayo Clinic population or in the clinic population from Malmo, Sweden in those with an onset prior to 1995. RA-associated ILD is an under-recognised extra-articular feature of RA. In the early rheumatoid arthritis study (ERAS) inception cohort, following patients with early RA (<2 years at study entry), the annualised incidence rate of RA-ILD was 4.1 per 1000 (95%CI 3.0, 5.4) and the 15-year cumulative incidence was 62.9 per 1000 (95%CI 43.0, 91.7) . About one-third of the patients had already developed RA-ILD at baseline. In another study, including a small study population of non-smoking patients with RA without a history of respiratory disorders, the prevalence of pulmonary lung function test abnormalities was higher than expected but did not change over a 10 year period . In a retrospective cohort study using annual cross-sectional data collected from 1985 to 2006 from the US Veteran’s Health Administration system, an increase in RA lung disease was observed, increasing from 19 to 22 cases per 1000 over 10 years . Possible explanations include improved diagnostic sensitivity, but it may also be due to increased use of MTX and biological therapies.

MTX pulmonary toxicity has been reported in a number of studies . Almost all these studies were conducted in the previous millennium when the prescribed dosage of MTX was relatively low while less has been published in the last decade since MTX became the DMARD of first choice in many countries and the mean dose of MTX used has increased. In the CAMERA study, comparing an intensive treatment approach with a conventional treatment approach with MTX performed in the 21st century, 2.7% of patients in the intensive treatment group and 2.1% of patients in the conventional treatment group experienced pulmonary disorders (i.e., pleuritis, restrictive lung disease and exacerbation of chronic obstructive lung disease) .

A number of studies have investigated the association between biologic agents and lung disorders, such as RA-ILD in patients with RA, and results are controversial. In case reports and case series, a rapid progression of pre-existing RA-ILD or development of new-onset RA-ILD was observed. However, most of these cases occurred within 3 months of starting biological treatment and a causal relation is therefore less likely. In a study from the British Society for Rheumatology Biologics Register comparing RA-ILD, no increase in RA-ILD mortality was observed in patients receiving biological therapy compared with those receiving conventional DMARD therapy.

Co-morbidity

There has been an increasing awareness of the importance of co-morbidities in patients with RA, in recent years. This is in part because of better comparisons with age- and gender-matched counterparts in the general population and also because, as a consequence of the ageing population, the mean age at RA onset and of prevalent cases is increasing. The two co-morbidities of particular concern are cardiovascular disease (CVD) and respiratory disease. In addition, as smoking and diabetes are both risk factors for the development of RA, they add to the burden of co-morbidity .

Mortality

It is now well recognised that RA is associated with reduced life-expectancy and, as yet, there is no clear evidence that this is improving. A meta-analysis of 24 studies of mortality in RA published between 1970 and 2005 reported a weighted combined all-cause standardised mortality ratio (meta-SMR) of 1.50 (95%CI 1.39, 1.61). Over 50% of the excess deaths in RA were due to CVD. The meta-SMR for CHD was 1.59 (95% CI 1.46, 1.73) and for stroke it was 1.52 (95% CI 1.46, 1.73) . Despite changes in RA course over the past decades, the SMR for CV death has not changed over a 50-year period . Of course, the mortality from CVD in the general population has fallen in this time – nevertheless, it is disappointing that the relative mortality between an RA population and the general population should not have changed despite improvements in disease control.

Gonzalez et al. reported on the overall mortality experience of 822 RA patients who were residents of Olmsted County, Minnesota and diagnosed between 1955 and 2000. . The mortality rates for men and women (2.5 and 2.4 per 100 person years) were fairly constant throughout the decades. During the same time period, the mortality rate in the general population fell substantially, resulting in a widening mortality gap.

Mortality studies of patients treated with biologic agents are only just starting to appear. As yet, there have been no mortality studies confined to patients with a disease onset in the biologic era. A number of studies have suggested that MTX treatment, and in particular good response to MTX, results in improved survival . A study from Sweden reported a hazard ratio of 0.65 for patients treated with anti-tumour necrosis factor (TNF) agents compared with those treated with non-biologic DMARDs after adjustment for age, gender, physical function and baseline co-morbidity .