Remission rates in early RA

The current aim of treatment in patients with early RA is remission and when it is not possible to achieve remission to aim for low disease activity. Several definitions of remission are reported in clinical trials and observational studies including the American College of Rheumatology (ACR) preliminary criteria for remission , the Routine Assessment of Patient Index Data (RAPID-3) or those based on existing composite scores such as the Disease Activity Score (DAS), the 28-joint DAS (DAS28) , the Clinical Disease Activity Index (CDAI) and the Simplified Disease Activity Index (SDAI) . In 2011, an international task force from the ACR and the European League Against Rheumatism (EULAR) published new remission criteria to be used in clinical trials . Table 1 shows variables included in each of these indices and the cut-off used to define remission. The inclusion of different variables and the variation in definitions and cut-off levels of remission may result in categorising different patients in remission. A few studies compared the classification of patients in remission according to these different remission indices and some of these findings are summarised below.

In a study by Klarenbeek et al. including patients recruited to the BeSt study, a multi-centre clinical trial comparing four treatment strategies, who were followed up for 5 years, the following composite indices were evaluated: DAS, DAS28, SDAI, CDAI and three variants of the DAS score with adjustments in the tender joint count (TJC) of the score. The percentage of patients in remission on at least three occasions during the fifth year ranged from 15.4% applying the new ACR/EULAR (66-/68-joint count) clinical trial definition to 47.9% when applying the less stringent DAS28 CRP definition of remission. In this study also a small variation was observed applying the same remission criteria, but including either a 28-joint count compared to a 66-joint count as observed in other studies. It should however be noted that in the BeSt study treatment decisions were guided by the DAS score, and that patients recruited to this clinical trial may not be representative of a general RA population. In one observational cohort of patients with early RA (mean standard deviation (SD) disease duration 0.3 (0.6) years), the percentage of patients in remission 1 year after inclusion using the 28-joint count vs the 38-joint count was also very similar, respectively, 9% vs 7% for the ACR/EULAR Boolean and 14% vs 10% for the SDAI clinical trial criteria for remission . Somewhat higher percentages of patients in remission were observed in patients with early RA 6 months after inclusion into the ESPOIR study (median [IQR] disease duration 4.8 [2.9–7.0] months) for ACR/EULAR Boolean (12.9%), SDAI (17.0%), DAS28 (32.5%), CDAI (17.9%) and RAPID3R (25.1%). Kappa agreement between the ACR/EULAR Boolean criteria and the other indices of remission ranged from good (from 0.70 for RAPID3R + SJ0 + D1 to 0.79 for SDAI) to low/moderate (0.46 for DAS28 to 0.67 for RAPID3R + SJ0) . In the CATCH study, measuring agreement between ACR/EULAR remission criteria (both 28- and 66-/68-joint scores), SDAI, CDAI and DAS28, the agreement between all ACR/EULAR criteria and DAS28-based criteria was fair to moderate ( ϰ range 0.39–0.55) and substantial for SDAI and CDAI ( ϰ range 0.73–0.81) . In this study, including 369 patients with early RA, the percentage of patients achieving remission by 6 months after inclusion ranged from 14% (ACR/EULAR trial 68/66 definition) to 30% (DAS28) and patients in DAS28 remission had overall higher swollen joint count (SJC), TJC, physician and patient-reported global assessment scores.

Early remission and long-term outcomes

Some conflicting results have been found when comparing these remission criteria in relation to long-term outcomes such as functional disability, radiographic damage and mortality. In the BeSt study, the estimated probability for a health assessment questionnaire (HAQ) score >0.5 and annual Sharp van der Heijde score (SHS) progression of ≥3 units showed high agreement between the different indices ranging from, respectively, 34% (ACR/EULAR 68/66 trial) to 41% (DAS28 CRP) for HAQ and from 9% (CDAI, ACR 66/68 trial, ACR 68/66 practice) to 12% (DAS CRP, DAS28 CRP, DAS TJC53, DAS TJC 44) for SHS . Similarly, comparing the ACR/EULAR remission criteria based on the 28- or 38-joint count during a 1-year follow-up study, despite a difference in remission rates and residual pain and inflammation of the joints in the feet between the indices, no statistically significant difference between the 28- and 38-joint definitions on the impact on functional disability and radiographic damage deterioration was observed . In contrast, the highest positive likelihood ratios (LR+) were found for SDAI in association with non-progression of radiographic damage (LR+ 4.0 (95% CI 1.7–9.4)) , for ACR/EULAR remission criteria with low functional disability (LR+ 5.67 (95% CI 1.34–24.00)) and for SDAI with absence power Doppler-positive synovitis (LR+ 6.46 (95% CI 1.95–21.34)) in two observational studies, while in both studies DAS28-CRP had significantly lower positive LRs for these outcomes. Thus, when a stricter definition of remission is used, less deterioration of functional disability and progression of radiographic damage are seen.

Time is not included as a criterion in most definitions of remission. In general practice, it is recognised that sustained remission is related with more favourable outcomes such as less functional disability and radiographic damage over time, less synovitis and a reduced likelihood of mortality but research is limited. To date, some of the studies investigating these associations are based on small sample sizes or include patients recruited before 2000. In a small study including patients with early RA ( n = 147) who had responded to methotrexate (MTX) after 3–4 months, radiographic progression was not statistically significant between those who were in remission at all time points over 2 years compared to other patients (SHS 3.79 (6.68) vs 4.75 (8.20), p = 0.6) . By contrast, patients recruited to the Early RA Study (ERAS), an inception cohort including patients since 1986, who were in sustained remission (DAS <1.6 at 3-, 4- and 5-year follow-up) vs those never or intermittently in remission had less erosions after 5 years (respectively, 54% and 78%, p < 0.001). In this UK study, patients who were in sustained remission were also less likely to become work disabled (1.3% vs 12%), had less orthopaedic interventions (5% vs 15%) and showed better functional outcomes (decrease in HAQ score 0.79–0.13 vs 0.92–1.1) . However, sustained remission did not result in a decreased mortality risk, probably due to the low number of deaths within 5 years. In another UK inception cohort, the Norfolk Arthritis Register (NOAR) cohort including patients recruited between 1990 and 1994 and between 2000 and 2004, a similar association between sustained remission during the first three years and functional disability was observed . Over a 5-year follow-up period, the number of times in remission (remission = no swollen joints and no tender joints out of a 51-joint count) correlated with the odds of functional disability, with a mean decrease in the probability of ∼64% for each additional time in remission (odds ratio (OR) 0.38, 95% CI 0.28–0.52). In the NOAR cohort and the Dutch Early Arthritis Clinic (EAC) cohort, patients who achieved remission (NOAR = no swollen and tender joints and EAC: disease-modifying antirheumatic drug (DMARD)-free remission for 1 year) within the first 3 years had improved survival compared to those without remission (NOAR: hazard ratio (HR) 0.72, 95% CI 0.55–0.94 and EAC: HR 0.60, 95% CI 0.39–0.93). Although the mortality rate in the Dutch cohort including patients recruited between 1993 and 1998 was increased compared to the national population , later analysis showed that only patients who did not achieve remission during the first 3 years of disease had a significantly increased mortality rate compared to the national population, whereas patients who achieved remission did not . Comparison between the older cohort with a more recent cohort of patients treated more aggressively also revealed that the mortality rate was similar in the more recent cohort, suggesting that aggressive treatment with the aim of remission leads to better survival.

Overall, we can say that current advances in the treatment of RA have made it possible to achieve remission in a larger RA population than a decade ago, but unfortunately still a considerable proportion of patients do not respond adequately to treatment resulting in decreased physical functioning, deterioration of joint damage and increased risk of mortality. In light of the ICF framework to reduce the burden of RA, the aim should therefore be sustained and drug-free remission in an early phase of the disease.

Remission rates in early RA

The current aim of treatment in patients with early RA is remission and when it is not possible to achieve remission to aim for low disease activity. Several definitions of remission are reported in clinical trials and observational studies including the American College of Rheumatology (ACR) preliminary criteria for remission , the Routine Assessment of Patient Index Data (RAPID-3) or those based on existing composite scores such as the Disease Activity Score (DAS), the 28-joint DAS (DAS28) , the Clinical Disease Activity Index (CDAI) and the Simplified Disease Activity Index (SDAI) . In 2011, an international task force from the ACR and the European League Against Rheumatism (EULAR) published new remission criteria to be used in clinical trials . Table 1 shows variables included in each of these indices and the cut-off used to define remission. The inclusion of different variables and the variation in definitions and cut-off levels of remission may result in categorising different patients in remission. A few studies compared the classification of patients in remission according to these different remission indices and some of these findings are summarised below.

In a study by Klarenbeek et al. including patients recruited to the BeSt study, a multi-centre clinical trial comparing four treatment strategies, who were followed up for 5 years, the following composite indices were evaluated: DAS, DAS28, SDAI, CDAI and three variants of the DAS score with adjustments in the tender joint count (TJC) of the score. The percentage of patients in remission on at least three occasions during the fifth year ranged from 15.4% applying the new ACR/EULAR (66-/68-joint count) clinical trial definition to 47.9% when applying the less stringent DAS28 CRP definition of remission. In this study also a small variation was observed applying the same remission criteria, but including either a 28-joint count compared to a 66-joint count as observed in other studies. It should however be noted that in the BeSt study treatment decisions were guided by the DAS score, and that patients recruited to this clinical trial may not be representative of a general RA population. In one observational cohort of patients with early RA (mean standard deviation (SD) disease duration 0.3 (0.6) years), the percentage of patients in remission 1 year after inclusion using the 28-joint count vs the 38-joint count was also very similar, respectively, 9% vs 7% for the ACR/EULAR Boolean and 14% vs 10% for the SDAI clinical trial criteria for remission . Somewhat higher percentages of patients in remission were observed in patients with early RA 6 months after inclusion into the ESPOIR study (median [IQR] disease duration 4.8 [2.9–7.0] months) for ACR/EULAR Boolean (12.9%), SDAI (17.0%), DAS28 (32.5%), CDAI (17.9%) and RAPID3R (25.1%). Kappa agreement between the ACR/EULAR Boolean criteria and the other indices of remission ranged from good (from 0.70 for RAPID3R + SJ0 + D1 to 0.79 for SDAI) to low/moderate (0.46 for DAS28 to 0.67 for RAPID3R + SJ0) . In the CATCH study, measuring agreement between ACR/EULAR remission criteria (both 28- and 66-/68-joint scores), SDAI, CDAI and DAS28, the agreement between all ACR/EULAR criteria and DAS28-based criteria was fair to moderate ( ϰ range 0.39–0.55) and substantial for SDAI and CDAI ( ϰ range 0.73–0.81) . In this study, including 369 patients with early RA, the percentage of patients achieving remission by 6 months after inclusion ranged from 14% (ACR/EULAR trial 68/66 definition) to 30% (DAS28) and patients in DAS28 remission had overall higher swollen joint count (SJC), TJC, physician and patient-reported global assessment scores.

Early remission and long-term outcomes

Some conflicting results have been found when comparing these remission criteria in relation to long-term outcomes such as functional disability, radiographic damage and mortality. In the BeSt study, the estimated probability for a health assessment questionnaire (HAQ) score >0.5 and annual Sharp van der Heijde score (SHS) progression of ≥3 units showed high agreement between the different indices ranging from, respectively, 34% (ACR/EULAR 68/66 trial) to 41% (DAS28 CRP) for HAQ and from 9% (CDAI, ACR 66/68 trial, ACR 68/66 practice) to 12% (DAS CRP, DAS28 CRP, DAS TJC53, DAS TJC 44) for SHS . Similarly, comparing the ACR/EULAR remission criteria based on the 28- or 38-joint count during a 1-year follow-up study, despite a difference in remission rates and residual pain and inflammation of the joints in the feet between the indices, no statistically significant difference between the 28- and 38-joint definitions on the impact on functional disability and radiographic damage deterioration was observed . In contrast, the highest positive likelihood ratios (LR+) were found for SDAI in association with non-progression of radiographic damage (LR+ 4.0 (95% CI 1.7–9.4)) , for ACR/EULAR remission criteria with low functional disability (LR+ 5.67 (95% CI 1.34–24.00)) and for SDAI with absence power Doppler-positive synovitis (LR+ 6.46 (95% CI 1.95–21.34)) in two observational studies, while in both studies DAS28-CRP had significantly lower positive LRs for these outcomes. Thus, when a stricter definition of remission is used, less deterioration of functional disability and progression of radiographic damage are seen.

Time is not included as a criterion in most definitions of remission. In general practice, it is recognised that sustained remission is related with more favourable outcomes such as less functional disability and radiographic damage over time, less synovitis and a reduced likelihood of mortality but research is limited. To date, some of the studies investigating these associations are based on small sample sizes or include patients recruited before 2000. In a small study including patients with early RA ( n = 147) who had responded to methotrexate (MTX) after 3–4 months, radiographic progression was not statistically significant between those who were in remission at all time points over 2 years compared to other patients (SHS 3.79 (6.68) vs 4.75 (8.20), p = 0.6) . By contrast, patients recruited to the Early RA Study (ERAS), an inception cohort including patients since 1986, who were in sustained remission (DAS <1.6 at 3-, 4- and 5-year follow-up) vs those never or intermittently in remission had less erosions after 5 years (respectively, 54% and 78%, p < 0.001). In this UK study, patients who were in sustained remission were also less likely to become work disabled (1.3% vs 12%), had less orthopaedic interventions (5% vs 15%) and showed better functional outcomes (decrease in HAQ score 0.79–0.13 vs 0.92–1.1) . However, sustained remission did not result in a decreased mortality risk, probably due to the low number of deaths within 5 years. In another UK inception cohort, the Norfolk Arthritis Register (NOAR) cohort including patients recruited between 1990 and 1994 and between 2000 and 2004, a similar association between sustained remission during the first three years and functional disability was observed . Over a 5-year follow-up period, the number of times in remission (remission = no swollen joints and no tender joints out of a 51-joint count) correlated with the odds of functional disability, with a mean decrease in the probability of ∼64% for each additional time in remission (odds ratio (OR) 0.38, 95% CI 0.28–0.52). In the NOAR cohort and the Dutch Early Arthritis Clinic (EAC) cohort, patients who achieved remission (NOAR = no swollen and tender joints and EAC: disease-modifying antirheumatic drug (DMARD)-free remission for 1 year) within the first 3 years had improved survival compared to those without remission (NOAR: hazard ratio (HR) 0.72, 95% CI 0.55–0.94 and EAC: HR 0.60, 95% CI 0.39–0.93). Although the mortality rate in the Dutch cohort including patients recruited between 1993 and 1998 was increased compared to the national population , later analysis showed that only patients who did not achieve remission during the first 3 years of disease had a significantly increased mortality rate compared to the national population, whereas patients who achieved remission did not . Comparison between the older cohort with a more recent cohort of patients treated more aggressively also revealed that the mortality rate was similar in the more recent cohort, suggesting that aggressive treatment with the aim of remission leads to better survival.

Overall, we can say that current advances in the treatment of RA have made it possible to achieve remission in a larger RA population than a decade ago, but unfortunately still a considerable proportion of patients do not respond adequately to treatment resulting in decreased physical functioning, deterioration of joint damage and increased risk of mortality. In light of the ICF framework to reduce the burden of RA, the aim should therefore be sustained and drug-free remission in an early phase of the disease.

Rates of joint damage

Despite earlier diagnosis and more intensive treatment, joint damage still occurs in many patients resulting in loss of function, participation, problems with daily activities and increased morbidity and mortality. In two observational cohorts, the French ESPOIR cohort and the Canadian CATCH cohort , respectively, 22% and 28% of patients had hand or foot erosions at presentation (mean disease duration of 103 days and 6.1 months, respectively). Although the percentage of patients treated according to 21st-century standard care showing radiographic progression was high (79%), with a mean change of 8.8 in the total modified SHS score and 2.6 in erosion score over 3 years, the progression rates are lower than those observed more than a decade ago in a Dutch cohort, with a mean annual SvH score of 8.6 and mean annual erosion score of 5.4 . A meta-analysis of randomised controlled trials (RCTs) comparing biologic treatment with MTX treatment in early RA showed a favourable effect of biologic treatment on radiographic progression compared to MTX treatment .

Patients with erosions at first presentation are more likely to have worse radiographic progression and subsequent decrease in functional disability than patients without joint damage. It is thus important to detect possible joint destruction at an early stage and monitor progression routinely as recommended by the EULAR task force on the management of early RA . To date, X-rays are still considered to be the gold standard for imaging of bone loss and cartilage degradation in RA, but conventional radiographs may not show these findings in patients with early RA. Ultrasound (US) has been sensitive in detecting both erosions and synovitis in patients with early RA and is more effective for scoring erosions than conventional radiography . While validated semi-quantitative radiographic damage scoring methods exist to measure joint damage by conventional X-rays (e.g., modified SvH score or Larsen score), consensus on scoring abnormalities of the common pathological findings using US still needs to be established . Magnetic resonance imaging (MRI) has also shown to be highly associated with radiographic damage, but the use of MRI in research and clinical practice is less feasible due to high costs . Another approach to detect early joint destruction is by measuring biomarkers of bone and cartilage destruction and turnover in urine and blood . So far, the association between these biomarkers and radiographic progression has been moderate and more research is necessary before measuring biomarker levels for clinical purposes.

Change in functional disability over time

For patients with RA, functioning, i.e., being able to perform daily activities and not being restricted in participation, is an important aspect of their life. Compared to the general population patients with RA have significantly higher HAQ scores (0.71 vs 0.17) . In the WHO ICF model ( Fig. 1 ), functioning (and its component body functions and structures, activities and participation) is seen in relation with a health condition (e.g., RA) and personal and environmental factors, which is a bidirectional relationship. It is therefore for this reason that functional ability is frequently included as a patient-reported outcome measure in clinical studies. In observational studies applying the HAQ at multiple time points in patients with early RA or inflammatory polyarthritis (IP) (symptom/disease duration <2 years), the HAQ score at entry in these studies ranged from a median of 0.6 to a mean of 1.3 . The majority of these studies suggested that the mean/median HAQ score over time is J-shaped ( Fig. 2 ). High HAQ scores at presentation are mainly thought to be associated with high levels of disease activity . Thereafter, an initial improvement is seen after treatment commencement followed by an insidious decline in function with radiographic joint damage having a larger effect on functional disability than disease activity . Overall, the annual rate of change in functional disability is estimated to be ∼0.02–0.03 units per year .

Change in mean/median HAQ scores over time in patients with early rheumatoid arthritis.

Contextual factors and functional disability

Both disease activity and radiographic damage are directly related to functional disability, but other personal and environmental factors may also contribute to disability. Identification of factors associated with worsening of functional disability, both modifiable and non-modifiable factors, will help in deciding on the best intervention for the individual patient. Women and older patients have higher HAQ scores at diagnosis and over time. When grouping patients according to age of onset, distinctive trajectories of functional disability were noticeable in NOAR in which patients with increasing age at onset (early-onset (<55 years), late-onset (55–75 years) and very-late-onset RA (75+ years)) had a faster rate of disability deterioration which was independent of treatment and co-morbidity measures . The recognition of co-morbidities such as CVD and respiratory manifestations is important as it has an impact on functional disability . Other factors associated with worse functional disability include low socio-economic status , low educational levels and psychological scores .

Pain and fatigue as outcome measures

In the context of the ICF framework, the experience of pain and fatigue is influenced by biologic, psychological and social factors and vice versa. For many patients, pain and fatigue are significant problems, but are not always included in common disease activity indices (e.g., DAS28 and CDAI). Results from an Outcome Measures in Rheumatology Clinical Trials (OMERACT) RA Flare group initiative, using a Delphi approach amongst 114 patients and 100 health-care professionals, showed that both patients (93%) and health-care professionals (93%) rated pain as the most important domain to be included in defining flare . Although fatigue was found to be important, especially for patients with established disease, it did not quite reach consensus by the total group (68% with 70% being the cut-off). Fatigue has been endorsed by ACR and EULAR as an important domain for inclusion in clinical trials and is also included in the Rheumatoid Arthritis Impact of Disease (RAID) score, as is pain .

These recommendations have led to an increasing number of studies reporting on pain and fatigue as primary or secondary outcomes. A recent review showed that the use of non-biologic DMARDs, either monotherapy or combination therapy, resulted in significant reductions of pain in patients with early RA . In the COMET trial , combining the data of patients treated with etanercept plus methotrexate or methotrexate monotherapy, patients who achieved remission showed the greatest improvement in patient-reported outcomes including Visual Analogue Scale (VAS) fatigue and VAS pain compared to patients with low disease activity.

Causes and consequences of pain and fatigue

High pain levels in early RA are associated with female gender , younger age , high DAS , worse mental health , high depression score and smoking . In established RA, pain is highly associated with joint destruction. Little is known about the causes and consequences of fatigue and a theoretical framework explaining the experience of fatigue in RA is lacking. The possible causes of fatigue include being of the female gender, illness-related aspects (e.g., pain), physical functioning (functional disability, sleep quality and disturbances, VAS global health), cognitive and emotional functioning (e.g., anxiety, depression and perceptions of self-efficacy) and social and environmental aspects (interpersonal events and social support) . Whereas measures of inflammation are associated with pain, less consistent results are found for fatigue. Consequences of fatigue include illness-related aspects (ocular and oral dryness and morning stiffness), physical functioning (physical health-related quality of life), cognitive and emotional functioning (depression, global psychological distress, mental health-related quality of life and satisfaction with health) and social and environmental aspects (work ability, negative and daily events, and parenting). These findings were based on results of both cross-sectional and longitudinal studies not fully explaining possible causal relationships .