Article Text
Abstract
Objective To investigate the predictive value of baseline depression/anxiety on the likelihood of achieving joint remission in rheumatoid arthritis (RA) and psoriatic arthritis (PsA) as well as the associations between baseline depression/anxiety and the components of the remission criteria at follow-up.
Methods We included 1326 patients with RA and 728 patients with PsA from the prospective observational NOR-DMARD study starting first-time tumour necrosis factor inhibitors or methotrexate. The predictive value of depression/anxiety on remission was explored in prespecified logistic regression models and the associations between baseline depression/anxiety and the components of the remission criteria in prespecified multiple linear regression models.
Results Baseline depression/anxiety according to EuroQoL-5D-3L, Short Form-36 (SF-36) Mental Health subscale ≤56 and SF-36 Mental Component Summary ≤38 negatively predicted 28-joint Disease Activity Score <2.6, Simplified Disease Activity Index ≤3.3, Clinical Disease Activity Index ≤2.8, ACR/EULAR Boolean and Disease Activity Index for Psoriatic Arthritis ≤4 remission after 3 and 6 months treatment in RA (p≤0.008) and partly in PsA (p from 0.001 to 0.73). Baseline depression/anxiety was associated with increased patient’s and evaluator’s global assessment, tender joint count and joint pain in RA at follow-up, but not with swollen joint count and acute phase reactants.
Conclusion Depression and anxiety may reduce likelihood of joint remission based on composite scores in RA and PsA and should be taken into account in individual patients when making a shared decision on a treatment target.
- rheumatoid arthritis
- psoriatic arthritis
- remission
- depression
- anxiety
Statistics from Altmetric.com
Introduction
In today’s treatment of rheumatoid arthritis (RA) and psoriatic arthritis (PsA), remission is the main target. However, far from all patients reach this target.1
Studies have suggested depression and anxiety to be associated with greater perception of pain in RA patients.2 3 Furthermore, depression has been associated with reduced pain threshold and tolerance.4 Depression is thought to promote sensitisation, to interfere with endogenous pain inhibition and to have a profound long-term influence on the shaping of pain responses and pain outcomes.4 Depression and anxiety are more common in RA and PsA compared with the general population, with reported prevalence between 10%–42% in RA and 9%–37% in PsA.5 6
Only few studies have assessed the predictive value of depression and anxiety in RA suggesting poorer treatment outcomes.3 7 Whether this can be confirmed in larger, prospective studies using various remission criteria, as well as in PsA, remains unexplored.
In this study, we aimed to investigate the predictive value of baseline depression and anxiety on the likelihood of achieving joint remission in patients with RA and PsA, in a large prospective multicentre observational register.8 Secondary objectives were to explore the associations between baseline depression and anxiety and the components of the remission criteria at follow-up.
Methods
Patients
We included RA and PsA patients from the Norwegian Disease-Modifying Anti-Rheumatic Drug (NOR-DMARD) register,8 who started first tumour necrosis factor inhibitor (TNFi) with or without comedication with methotrexate (MTX), or started MTX as first DMARD, all patients included only once (see online supplementary figure S1). NOR-DMARD is a prospective observational multicentre study initiated in December 2000 including patients >18 years with inflammatory arthropathies starting biological or synthetic DMARDs. Assessments were made at baseline, 3, 6 and 12 months and then yearly. The study comprises five centres in different geographical regions in Norway.
Supplementary file 2
The patients in the current analyses were included between 1 March 2006 (introduction of EuroQoL-5D-3L (EQ-5D-3L)9 and 6 November 2012 and followed until 15 April 2013. Written informed consent was obtained from each patient. The study was approved by the National Data Inspectorate and by the Regional Ethics Committee.
Assessments
Assessments included 32 tender/swollen joint count (28 joint count with addition of ankles (0–2) and metatarso-phalangeal joints (0–2)), visual analogue scales (0–100 mm) for joint pain, patient’s and evaluator’s global assessments, Modified Health Assessment Questionnaire,10 erythrocyte sedimentation rate (mm/hour), C reactive protein (CRP; mg/L), disease duration, current smoking status (yes/no) and erosive disease at baseline (yes/no). Twenty-eight-joint Disease Activity Score (DAS28),11 Simplified Disease Activity Index (SDAI)11 and Clinical Disease Activity Index (CDAI)11 were computed, as well as a modified Disease Activity Index for Psoriatic Arthritis (DAPSA)12 using 32 instead of 66/68 joint count and joint pain instead of pain. The following criteria for depression/anxiety were applied: (1) EQ-5D-3L question 5: ‘I am not or moderately/extremely anxious or depressed’,9 (2) Medical Outcomes Survey Short Form-36 (SF-36) Mental Health subscale (SF-36MH)≤5613 and (3) SF-36 Mental Component Summary (SF-36MCS)≤38.13 The following remission criteria were explored: (1) DAS28 <2.6,11 (2) SDAI ≤3.3,11 14 (3) CDAI ≤2.8,11 (4) ACR/EULAR Boolean11 and (5) DAPSA ≤4.12
Statistics
For the patients’ demographic and baseline characteristics medians (25th and 75th percentiles) were calculated for non-normally and means (SD) for normally distributed data. Continuous measures were compared using Mann-Whitney U-test or independent t-test as appropriate. Proportions were analysed using X2 test.
In the main analyses, as well as in subgroup analyses of TNFi/MTX-treated patients, the predictive value of depression/anxiety for DAS28, SDAI, CDAI, ACR/EULAR Boolean and DAPSA remission after 3 and 6 months treatment was explored in prespecified logistic regression models adjusted for age, sex, disease duration and smoking.
The analyses were performed as completer analyses and without adjustment for multiple comparisons. Sensitivity analyses with additional adjustment for baseline erosive disease and CRP were performed.
In secondary analyses, the components of the remission criteria at 3 and 6 months follow-up were explored for patients with and without baseline depression/anxiety in prespecified multiple linear regression models adjusted for age, sex, disease duration and smoking, as well as with and without adjustment for baseline values of the explored outcome variables. Statistical tests were performed using SPSS for Windows V.21.0.
Results
Demographics and baseline disease activity measures according to EQ-5D-3L depression/anxiety status are shown in table 1. Number/total (%) of depressed/anxious patients according to SF-36MH≤56/ SF-36MCS≤38 was 238/1311 (18.2)/ 381/1277 (29.8) in RA and 115/724 (15.9)/ 182/708 (25.7) in PsA, respectively.
Associations between baseline depression/anxiety and achievement of joint remission
The proportion of patients achieving joint remission was overall lower in patient with versus without baseline depression/anxiety (figure 1).
Baseline depression/anxiety according to EQ-5D-3L, SF-36MH≤56 and SF-36MCS≤38 negatively predicted achievement of DAS28, SDAI, CDAI, ACR/EULAR Boolean and DAPSA remission after 3 and 6 months treatment in RA. Corresponding findings in PsA also showed consistently lower point estimates but did not reach significance for all the analyses (table 2).
The findings were consistent also in sensitivity analyses (see online supplementary table S1). In subgroup analyses of TNFi/MTX-treated patients, baseline depression/anxiety according to EQ-5D-3L, SF36-MH≤56 and SF36-MCS≤38 were found to be negative predictors of remission at 6 months in RA, but only partly in PsA (see online supplementary tables S2–4).
Supplementary file 1
Secondary outcomes analyses
Baseline depression/anxiety was associated with increased patient’s and evaluator’s global assessment, joint pain and tender joint count at 3 and 6 months in RA and with increased patient’s global assessment and joint pain in PsA, but not with swollen joint count or levels of acute phase reactants (see online supplementary tables S5a-c and S6a-c). Additional adjustment for baseline values of the explored outcome variables did not change the main results.
Discussion
In this prospective multicentre observational study, baseline depression and anxiety were found to be strong negative predictors of remission after 3 and 6 months treatment in RA, and partly in PsA, using joint-specific measures of remission. The findings were valid across different depression/anxiety and remission criteria, as well as in subgroup analyses of TNFi-treated and MTX-treated patients.
The thresholds SF-36MH≤56 and SF-36MCS≤38 were recently found to have good sensitivity and specificity to detect depression and anxiety in RA patients.13 The EULAR recommendation to use one single question to screen for depression in chronic inflammatory rheumatic diseases is in accordance with the EQ-5D-3L criterion,15 which may measure slightly different aspects compared with the more comprehensive SF-36 derived criteria.
The prespecified logistic regression model was adjusted for age, sex, disease duration and smoking, but not for recently identified predictors in NOR-DMARD for only one or some of the explored remission criteria, nor for patient-reported outcomes (PROs) or composite scores comprising PROs, as these may be symptoms of/ influenced by depression and anxiety.2 4 16 The validity of the findings reported in table 2 was confirmed in sensitivity analyses with adjustment for additional covariates (see online supplementary table S1).
This study is in line with a recent study in which self-reported lifetime history of depression was associated with decreased likelihood of 6 months CDAI remission in RA.7 However, associations between baseline depression and the components of CDAI at follow-up were not confirmed. Furthermore, in post hoc analyses of clinical trial data in RA. persisting depression/anxiety assessed by EQ-5D was associated with a decreased likelihood of DAS28 remission as well as higher patient’s global assessment and tender joint count after 2 years.3 By contrast, in a smaller study in RA, depression/anxiety did not reach significance as predictors of 1-year DAS28 remission.17 Still, baseline depression/anxiety was associated with higher tender joint count and patient’s global assessment at follow-up.17
We have recently reported reduced likelihood of remission among RA and PsA patients with discordance between baseline patient’s and evaluator’s global assessment as well as tender and swollen joint count.18 Discordance between more subjectively and more objectively weighted measures of disease activity may be reflected by psychosocial factors like depression.4
Major limitations of the study are lack of consensus on validated remission criteria in PsA, lack of established diagnoses of depression and anxiety, as well as lack of psoriasis severity information, as psoriasis may have major implications for mental health. Furthermore, the assessment of 32 and not 66/68 joint count may have led to overestimation of remission rates in PsA and missing data handled by completer analyses may have affected the generalisability of the results. Finally, the modification of DAPSA with 32 instead of 66/68 joint count and joint pain instead of pain is not validated for PsA and DAPSA is developed as remission criterion for PsA and not RA.12
The major strength of this study is the prospective observational multicentre design and the inclusion of large cohorts of RA and PsA patients over a long time span. Furthermore, this is the first study to assess the predictive value of depression and anxiety on achievement of remission in PsA and the first study using EQ-5D as well as SF-36-derived depression/anxiety criteria as predictors of remission in RA. The consistent findings for these different predictors support the validity and the robustness of the results.
Depression and anxiety are frequently occurring disorders but are not widely considered in routine care of arthritis patients, although of considerable economic impact to the society.6 19 Increased emphasis on the negative predictive value of depression and anxiety ought to be considered as part of a treat-to-target strategy in patients not reaching remission. In particular, alternative targets to composite scores may be considered in the shared decision making between the patient and healthcare provider. Thus, depression and anxiety should be taken into account according to recommendation number 5 (RA) and 6 (PsA) in the treat-to-target recommendations.20 21
In conclusion, depression and anxiety were found to be strong, negative predictors of joint remission at 3 and 6 months treatment in RA and partly in PsA, according to various remission as well as depression/anxiety criteria. Depression and anxiety were associated to more subjectively weighted measures, but not acute phase reactants and swollen joint count during follow-up. These observations support a focus on depression and anxiety as comorbidities in a treat-to-target strategy.
Acknowledgments
The authors would like to thank the patients for participating in this study and the local rheumatology staff for data collection.
References
Footnotes
Contributors BM, EKK, JS, HBH, KMF, EL, GH and TKK were responsible for study design. AW, SK, ER, FK and TKK were responsible for data acquisition. BM analysed the data and wrote the manuscript. All authors critically revised the manuscript and approved the final version.
Funding The study was funded through a clinical research fellowship from Diakonhjemmet Hospital, originating from South-Eastern Health Authority. Data collection in NOR-DMARD was partly funded through unrestricted grants from Abbvie, BMS, MSD, Pfizer (Wyeth), Roche and UCB.
Competing interests HBH has received fees for speaking and/or consulting from AbbVie, Pfizer, UCB, Roche, MSD, BMS and Novartis. TKK has received fees for speaking and/or consulting from AbbVie, Biogen, BMS, Boehringer Ingelheim, Celgene, Celltrion, Eli Lilly, Epirus, Hospira, Merck-Serono, MSD, Mundipharma, Novartis, Oktal, Orion Pharma, Hospira/Pfizer, Roche, Sandoz and UCB and received research funding to Diakonhjemmet Hospital from AbbVie, BMS, MSD, Pfizer, Roche and UCB. All other authors have declared that no competing interests exist.
Provenance and peer review Not commissioned; externally peer reviewed.