Long-Term Increase of Radiographic Damage and Disability in Patients with RA in Relation to Disease Duration in the Era of Biologics. Results from the SCQM Cohort

Abstract

Objectives: There is little information on the relation between disease duration, disability and radiographic outcome since the introduction of biologics into the therapy of rheumatoid arthritis (RA). No long -term cohort studies have been conducted on this subject so far. To analyse radiographic damage, disability, and disease activity in RA-patients dependent on disease duration in the Swiss national RA cohort (SCQM).

Methods: The primary outcome was the association between the radiographic destruction, assessed by Ratingen scores, and disease duration. All patients with at least one clinical visit were analysed with polynomial and multiple negative binomial models.

Results: The disease duration in the 8678 patients with available radiographs analysed ranged between less than 1 and more than 65 years (median 8.3). Disease duration and radiographic destruction were significantly associated with an average increase of Ratingen scores by 8.3% per year. Apart from disease duration, positive rheumatoid factor was the strongest predictor for radiographic destruction. While DAS28-scores remained stable in patients with a disease duration of more than 5 years (median DAS28 2.8), HAQ-DI scores increased continuously by 0.018 for each additional year.

Conclusion: In this RA cohort, patients show a continuous increase of articular destruction and physical disability in parallel with disease duration. Even when nowadays a satisfactory control of disease activity can be achieved in most patients, RA remains a destructive disease leading to joint destruction and physical disability in many patients.

Keywords: ACPA; DAS28; HAQ-DI; Ratingen score; disease duration; radiographic destruction; rheumatoid arthritis; rheumatoid factor.

Figure 1

Patient distribution according to disease duration illustrated as individual data (black line) and as a histogram representing the subgroups described in Table 1.

Figure 2

Smoothed time trends for radiographic damage, disease activity, and patient centred outcome: (A) Ratingen scores, (B) DAS28 scores, and (C) Health assessment questionnaire-disability index (HAQ-DI) scores in dependence of disease duration as modelled by local polynomial regression using a smoothing parameter of 0.5. Results are shown with 95% confidence intervals (CI, grey shade).

Figure 3

(Column A) Results for the negative binomial regression model with the Ratingen score as outcome. The estimate for each of the parameters indicates the percentage change in Ratingen score associated with a change of said parameter. While most variables included in the model are binary, for the continuous variables, such as time since diagnosis and age, the estimates refer to an increase of the parameter by 1 unit. Values greater than 1 indicate a positive association, smaller than 1 indicate a negative association. Results for the linear regression models for (Column B) the DAS28 score and (Column C) the HAQ-DI. The estimate for each of the parameters indicates how much the score changes in absolute, if only this parameter changes. While most variables included in the model are binary, for the continuous variables, such as time since diagnosis and age, the estimates refer to an increase of the parameter by 1 unit. Values greater than 0 indicate a positive association, smaller than 0 indicate a negative association.

Figure 4

Correlation of HAQ-DI with radiographic destruction and disease activity stratified according to disease duration: The dependency of the HAQ-DI score from (A–C) Ratingen scores and (D–F) DAS28 scores for patients with a disease duration of <5 years (A,D) 5-15 years (B,E), and >15 years (C,F) are modelled by local polynomial regression fittings using a smoothness parameter of 0.1. Results are shown with 95%-CI (grey shade).