We performed a cohort cross-sectional study. The evaluating period was between October 2013 and July 2014 from a university hospital in the north of Portugal. Consecutive RA patients were recruited by participating study physicians during routine clinic visits. We included adult RA patients, diagnosed according to the 2010 ACR/EULAR classification criteria, under biological therapy. Patients lacking ability to answer the questionnaires were excluded. Additional exclusion criteria were not applied. A total of 154 patients were included. The medical ethics committee approved the study protocol.

HRQoL was captured by the Short Form Health Survey (SF-36) and EQ-5D.

SF-36 measures physical and mental health on eight subscales (physical functioning – PF, physical role functioning – PRF, bodily pain – BP, general health perceptions – GHP, vitality – VT, social role functioning – SRF, emotional role functioning – ERF, mental health – MH), which are summed up in two scales, the physical and mental summary measures (PSM and MSM). Each scale comprises four dimensions scored from 0 to 100, with higher scores indicating better HRQoL. We calculated summary scores of the PCS and MCS, as well as single domain scores. These component scores are standardized on the basis of worldwide population norms to a mean of 50 and a standard deviation of 10.18,19

EQ-5D is applicable to a variety of different illnesses and treatments and provides a simple descriptive profile and a single index value for health status. The five dimensions included in the EQ-5D are: mobility; self-care; usual activities; pain/discomfort and anxiety/depression. Each dimension has three levels (none, some or major problems) and together defines 243 health states (3 to the power of 5 gives the 243 possible combination), to which has been added “unconscious” and “dead” for a total of 245 health states. EQ-5D scores range between −0.594 and 1 (full health).20

Limitations in activities of daily life were assessed by portuguese version of the Health Assessment Questionnaire (HAQ) which comprises eight categories. A score of 0 (no difficulty), 1 (some difficulty), 2 (much difficulty or need of assistance) or 3 (unable to perform) is given to each question; the highest score in each category represents the score for that category. The sum of scores is then divided by the number of categories, yielding a total score ranging from 0 (best) to 3 (worst).21

Visual Analog Scales (VASs) were used, on which patients had to indicate on a scale from 0 (none) to 100 (worst) mm their rating of disease activity (VAS-DA) and pain (VAS-P).

Data collection was performed in the rheumatology day-care hospital unit and through the longitudinal national database “reuma.pt”. The collected variables were C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), numbers of swollen and tender joints by using a 28-joint count (SJC28, TJC28); VAS-P and VAS-DA, and HAQ. Composite indices, such as the simplified disease activity indice (SDAI) and the Disease Activity Score using 28-joint counts (DAS28-4V) were calculated by informatic system. Age, gender, level of education, marital and employment status as well as additional disease characteristics – disease duration, rheumatoid factor (RF) and anti-cyclic citrullinated peptide antibody (anti-ACCP) status, extra-articular manifestations (rheumatoid nodules, ocular, rheumatoid vasculites and pleuropulmonary) and DMARDs use – were also recorded. The application of questionnaires, available in the computer system, was held through face-to-face interviews.

In the description of socio-demographic and clinical characteristics as well as in the description of PROs (eight domains of SF-36, PSM, MSM, EQ-5D, HAQ, VAS) were used means and standard deviations for continuous variables with symmetrical distribution and median and range (minimum and maximum) for the variables continuous with asymmetric distribution. Categorical variables were expressed as proportions. To compare HRQoL of our RA patients to that of the portuguese general population, we used SF-36 published data by Ferreira et al.22 Measures of central tendency were compared using the Mann–Whitney test. To investigate whether outcomes are different at different levels of disease activity we divided sample according to level of disease activity by SDAI: remission (REM) SDAI≤3.3, low disease activity (LDA) 3.3<SDAI≤11, moderate disease activity (MDA) 11<SDAI≤26 and high disease activity (HDA) SDAI>26. The comparison of outcomes was performed using Kruskal Wallis test and Fisher's exact test, for medians and categorical variables, respectively. To assess the relationship between clinical data and the PROs, Spearman correlations (ρ) were performed. To identify explanatory variables of HRQoL and functional capacity we conducted a multiple linear regression (stepwise selection) and included variables statistically significant in univariate analysis and those considered clinically relevant. We assessed correlations between the continuous variables and HRQoL with Spearman correlation to check for collinearity. Linear regression model assumptions were examined and satisfied. Residual plots were examined for goodness of fit. Analyses were performed using the program SPSS Version 21. The p-values below 0.05 were considered significant.

After analysing all domains of the SF-36 we observed that the domain with lowest score was the “Physical Performance” (median 25) followed by the domain “General Health” and “Physical Functioning” (median 40 for both). Mental domains showed higher values; among them, domain “Social Functioning” showed the lowest score (median 40; range 3–53). The domain “Role-Emotional” had a median of 100 (range 0–100) (Table 2).

Likewise, we found that in SF-36 summary measure, PSM was lower than MSM (median 32 vs. 52). With the application of EQ-5D also observed a low score (median 0.620, range −0.349 to 1.000). In our sample, 5% of patients (n=6) had scores on the EQ-5D less than zero (states worse than dead). Median score was 1.250 (range 0.000–2.875) for HAQ (Table 2).

When comparing primary outcomes between genders we found significant statistical differences only for the HAQ in which woman presented worse results (median 1.38 vs. 0.88; p=0.024). There were no significant differences in PROs when evaluated according to disease duration (≤10 or >10 years), RF and ACCP status, presence of extra-articular manifestations or current biological therapy.

HRQoL of our RA patients is lower than the general Portuguese population. Our patients had significantly lower scores in all domains of the SF-36, except for “Emotional role functioning” (Fig. 1). Regarding summary measures there is a significant difference in MSF (median 32 vs. 50, p<0.001) but not in the MSM (median 50 vs. 52, p=0.503). The same was observed through EQ-5D in which our patients had lower median score than general Portuguese population (0.620 vs. 0.758, p<0.001). Comparing the HRQoL of general Portuguese population with our patients in clinical remission, we found similars scores for the several individual domains and the summary measures of SF-36 (Fig. 2).

Fig. 1.

Comparison of SF-36 values of RA patients with general Portuguese population. Significant differences of different domains of SF-36 was found (p<0.001 for all analyses, except emotional role functioning (p=0.079), p-value using Mann–Whitney test.

(0.14MB).

Fig. 2.

Comparison of SF-36 in patients with remission state with general Portuguese population.

(0.14MB).

Regarding EQ-5D values, we also found identical values between REM group and the general population (0.848 vs. 0.758).

The performed analysis by disease activity groups showed significant differences in all applied instruments (Table 2).

When we access the functional capacity using the domains of physical health of the SF-36 (PF, PRF, BP and GHP) we also found statistically significant difference, showing worst scores with highest disease activity (p<0.001 for all analyzes). For the SF-36 mental health domains we only observed progressive decrease in the score with increase of level of activity disease on the “mental health” domain (p=0.001) (Table 2 and Fig. 3). The same was true for the SF-36 summary measures, i.e., decreasing scores of PSM with increasing levels of disease activity (p<0.001) and overlapping MSM values (Table 2). When we assessed potential differences in HRQoL by other measures, such as EQ-5D, we found very similar results (p<0.001) (Table 2).

Fig. 3.

Comparison of SF-36 among levels of disease activity determined by SDAI (remission (REM SDAI≤3.3; low disease activity (LDA) 3.3<SDAI≤11; moderate disease activity (MDA) 11<SDAI≤26 and high disease activity (HDA) SDAI>26)). Significant differences of different domains of SF-36 was found (p≤0.001 for all analyses by Kruskal Wallis test).

(0.13MB).

Comparing functional disability by HAQ at the four levels of disease activity, we observed significant differences between groups, showing a HAQ increase with increasing disease activity (p<0.001) (Table 2).

In order to compare the benefit of remission with state of low disease activity with respect to the PRO we directly compared REM group with LDA (Table 2). In this analysis we observed a statistically significant difference in all PRO (SF-36, EQ-5D, HAQ and VAS) except in “Vitality” and “Social role functioning” domains and mental summary measure of the SF-36.

We evaluated the relationship between the various clinical and laboratory parameters with the PRO through the Spearman correlation (ρ).

The analyzed clinical and laboratory variables were: current age and age at disease onset, disease duration, education level, CRP, ESR, tender and swollen 28-joint count, DAS28 change, VAS-P and VAS-GH. We also studied the relationship between the PSM of SF-36 and initial and actual HAQ.

For PSM we concluded that variables with strongest relationship were those that correspond to other PRO, namely current and initial HAQ (ρ=−0.67 and −0.48, respectively), VAS-P (ρ=−0.62) and VAS-GH (ρ=−0.58) (p<0.001 for all analyses). The relationship with tender and swollen 28-joint count, although significant, was much lower (ρ=−0.35 and −0.20, respectively). There was no correlation with ESR (ρ=0.02, p=0.84) nor with CRP (ρ=0.07, p=0.365). For MSM no significant correlations were identified.

The HAQ also showed weak correlation with ESR, CRP and 28-joint count. The strongest correlations were observed with PSM of SF-36 (ρ=0.674), VAS-GH (ρ=0.609), VAS-P (ρ=0.601) and initial HAQ (ρ=0.595) (p<0.001 for all analyzes).

Again, we verified that variables with strongest correlation with EQ-5D were other PRO (VAS-GH ρ=−0.573 and VAS-P ρ=−0.560, p<0.001 for both) and not the clinical or laboratory classics data.

There was not a significant association between the SF-36, EQ-5D and HAQ with the variation of DAS28-4V.

The potential explanatory variables considered for summary measures of the SF-36, EQ-5D and HAQ were age, gender, level of education, employment status, disease duration, VAS-GH, VAS-P, tender and swollen 28-joint count, ESR, CRP and DAS28-4V change. For SF-36 and EQ-5D we also included HAQ questionnaire. The regression coefficients are documented in Table 3. HAQ, VAS-P and level of education were retained as explanatory variables for PSM. The explained variance (R2) was 55% (p<0.001) and HAQ had the stronger effect on PSM (β=−9.119, p<0.001). For MSM, the model has not retained any significant explanatory variable. Therefore, we found no demographic or clinical variable to explain mental health.

HAQ and VAS-GH were retained as explanatory variables for EQ-5D and the model's explanation of the variance was significant (46.4%). As observed for PSM, HAQ was the variable with stronger effect on the EQ-5D (β=−0.208, p<0.001).

Studying HAQ determinants, we found that higher age, female gender, current employment, VAS-P and VAS-GH explained 50% of the variation. Presence of current employment was the variable with the greatest impact on functional capacity (β=−0.313).

Clinical and laboratory classic data (joint count, acute-phase reactants and DAS28-4V change) showed no explanatory effect on HRQoL and functional capacity of these patients.

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that they have followed the protocols of their work center on the publication of patient data.

The authors declare that no patient data appear in this article.