ORIGINAL ARTICLE

The association between anxiety and disease activity and qualityof life in rheumatoid arthritis: a systematic review and meta-analysis

Annabelle R. Machin1& Opeyemi Babatunde1

& Randula Haththotuwa1 & Ian Scott1,2 & Milica Blagojevic-Bucknall1 &

Nadia Corp1& Carolyn A. Chew-Graham1,3,4

& Samantha L. Hider1,2,4

Received: 3 October 2019 /Revised: 2 December 2019 /Accepted: 15 December 2019# International League of Associations for Rheumatology (ILAR) 2020

AbstractObjectives In people with rheumatoid arthritis (RA), mental health problems are common, but often not recognized or treated,contributing to increased morbidity and mortality. Most studies examining the impact of mental health problems in RA havefocused on depression. We aimed to determine the association between anxiety, and disease activity and quality of life (QoL) inpeople with RA.Methods A systematic review and meta-analysis were performed. A protocol was registered with PROSPERO (CRD2-17062580). Databases (Web of Science, PsycINFO, CINAHL, Embase, Medline) were searched for studies examining theassociation between anxiety and disease activity and QoL, in adults with RA, from inception to February 2019. Primary outcomemeasures were DAS28 and SF-36. Eligibility screening and data extraction were completed by two reviewers. Disagreementswere resolved by discussion or a third reviewer. Quality assessment was carried out using the Newcastle-Ottawa Scale.Results From 7712 unique citations, 60 articles were assessed for eligibility. The final review included 20 studies involving 7452people with RA (14 cross-sectional, 6 cohort). Eleven examined disease activity, 6 reported QoL outcome measures and 3included both. Anxiety was associated with increased disease activity and worse QoL. Meta-analysis showed anxiety to becorrelated with increased DAS28 scores (r = 0.23, CI 0.14, 0.31) and reduced physical (r = − 0.39, CI − 0.57, − 0.20) and mentalQoL (− 0.50, CI − 0.57, − 0.43).Conclusions Anxiety in people with RA is associated with increased disease activity and worse QoL. Improved recognition andmanagement of comorbid anxiety may help to improve outcomes for people with RA.

Key Points• This is the first systematic review and meta-analysis to examine the relationship between anxiety and disease activity and QoL in people with RA.• Anxiety was associated with higher disease activity both cross-sectionally and at up to 12-month follow-up.• Anxiety may have a more significant impact on disease activity in early RA, highlighting the importance of early recognition and management of

comorbid anxiety.• People with anxiety had poorer self-reported physical and mental QoL, although there was some heterogeneity in study findings, particularly for

physical QoL (I2 = 78.5%).

Keywords Anxiety . Disease activity . Quality of life .Meta-analysis . Rheumatoid arthritis . Systematic review

Electronic supplementary material The online version of this article(https://doi.org/10.1007/s10067-019-04900-y) contains supplementarymaterial, which is available to authorized users.

* Annabelle R. Machina.r.machin@keele.ac.uk

1 Primary Care Centre Versus Arthritis, School for Primary,Community and Social Care, Keele University, Staffordshire ST55BG, UK

2 Haywood Academic Rheumatology Centre, Midlands PartnershipFoundation Trust, Staffordshire, UK

3 West Midlands CLAHRC, West Midlands, UK

4 Midlands Partnership Foundation Trust, Staffordshire, UK

Clinical Rheumatologyhttps://doi.org/10.1007/s10067-019-04900-y

Introduction

Rheumatoid arthritis (RA) is a chronic systemic conditioncharacterized by joint pain and inflammation, affecting0.67% of the adult population [1]. Comorbid mental healthproblems are common; the estimated prevalence of anxietyis 14% [2], whilst prevalence estimates for depression varyfrom 16.8–38.8% depending on the method of case ascertain-ment [3], which are substantially higher than in the generalpopulation of England (prevalence of anxiety and depressionin adults is estimated at 5.9% and 3.3% respectively) [4].

In patients with RA, comorbid anxiety and depression areoften under-recognized and under-treated [5] and are associatedwith increased morbidity and mortality [6, 7]. Rathbun et al.have proposed a bidirectional relationship between depressionand physical health in RA [8]. Low mood may influence med-ication adherence [9], in addition to physical activity, leading todeconditioning and increased pain [10]. Psychological factorsmay also impact on disease activity by altering the immuneresponse, increasing pro-inflammatory cytokines and conse-quential inflammation [11]. Reciprocally, worse physical healthcan lead to depression [8].

Mental health problems can also impact on how diseaseactivity is assessed. People with RA and comorbid anxietyor depression can rate their disease activity higher than theirphysicians [12]. Patient-reported measures forming part of thedisease activity score in 28 joints (DAS28) are strongly influ-enced by psychological variables [13]. Consequently, separatereporting of the DAS28 components alongside an assessmentfor anxiety or depression could aid patient management [13].

To date, most studies examining the impact of mental healthproblems in RA have focused on depression. A systematicreview including 7 studies of low to moderate quality has foundthat depression may be associated with increased disease activ-ity in RA [8], whilst baseline depressive symptoms in peoplewith RA on biologic treatments have been linked to a reducedimprovement in DAS28 over time, compared with people with-out depression [14]. Further studies have shown depression inpeople with RA to be associatedwith reduced disease remission[15]. QoL has also been found to be significantly reduced inpeople with RA and depressive symptoms [6, 16].

When the impact of anxiety in RA has been examined, thishas often been in combination with depression. For example, asecondary analysis of data from a randomized controlled trialwas performed to determine whether symptoms of anxiety anddepression, as assessed using the EQ-5D, predicted treatmentresponse [17]. Both baseline and persistent mood problemswere associated with significantly increased DAS28 scores.

However, anxiety does not always co-exist withdepression—in the general population, at least 40% of individ-uals with anxiety do not have comorbid depression [18]. Arecent systematic review found the incidence of anxiety to behigher in people with RA (OR 1.2, 95% CI 1.03–1.39) [19].

This is important, as although some management options over-lap with depression [20], there are key management differencesfor people with anxiety, including applied relaxation and ahigher threshold for prescription of antidepressants or anxio-lytics [21]. As mentioned, people with RA and comorbid anxi-ety have also been found to rate their disease activity higher thantheir physician [12]. Consequently, it is important that anxiety isrecognized, both to ensure suitable psychological treatment andto facilitate appropriate management of RA. Despite this, fewstudies have examined the impact of anxiety alone in RA.

Understanding the impact of anxiety in RA on QoL anddisease activity will provide evidence to support the need torecognize and appropriately treat anxiety in people with RA.Consequently, a systematic review and meta-analysis wereperformed, to determine the relationship between anxietyand disease activity and QoL in patients with RA.

Materials and methods

The protocol for this systematic review was registered withPROSPERO number CRD2-17062580 [22].

Search strategy and study eligibility

A search strategy was developed using comprehensive textword searching of the title, abstract or keywords and databaseSubject Headings, combining terms for anxiety and RA. Thesearch was kept broad in order to ensure that alternative de-scriptive terms for the outcomes of interest, QoL and diseaseactivity were captured. Systematic searches were conducted infive electronic databases (Web of Science, PsycINFO(EBSCO), CINAHL (EBSCO), Embase (Ovid), Medline(Ovid)) from inception to February 2019, using customizedsearch terms for each database (See supplementary data 1 forthe Medline search strategy). Reference checking and citationtracking of index papers were completed. In addition, a searchfor grey literature was conducted using “www.opengrey.eu”.

All articles including a population of adults (> 18 years)with RA, with an exposure of anxiety (reported separatelyfrom depression) and comparator or control group (if any) ofadults with RA, were included. Primary outcome measureswere DAS28 and short-form (SF)-36, with additional validat-ed outcome measures for disease activity and QoL included,as detailed in the results. Studies were excluded if they includ-ed people aged under 18 years, data not specific to anxiety andRA, if an interpreter could not be found for a paper not writtenin the English language or when efforts to retrieve a full textwere unsuccessful and the abstract contained insufficient data.

The SF-36 is a set of patient self-reported QoL measuresthat assesses eight health concepts, which can be aggregatedinto two summary measures, the Physical (PCS) and Mental(MCS) Component Summary scores [23]. The SF-36 was

Clin Rheumatol

chosen due to it being acceptable to patients and having highinternal validity [24]. The DAS28 score comprises of bio-chemical measures (erythrocyte sedimentation rate (ESR) orC-reactive protein (CRP)), a 28 swollen joint count (SJC) andtender joint count (TJC), and a patient global assessment(PtGA) on a Visual Analogue Scale (VAS) [25]. The DAS28score was selected as a primary outcome measure as it iscommonly used to monitor disease activity and treatment re-sponse [26].

The full inclusion and exclusion criteria are detailed insupplementary data 2. Titles, abstracts and full texts werescreened by paired independent reviewers (AM, RH or IS)using pre-specified selection criteria. Disagreements regard-ing inclusion were resolved by discussion or the opinion of athird reviewer.

Data extraction and study quality assessment

A customized and piloted data extraction tool was used toextract relevant data from the included articles. Extracted dataincluded country of origin, study design, methodology, sam-ple characteristics, main findings and relevant statistical mea-sures. Data extraction was completed independently by tworesearchers, AM and RH, and any disagreements resolvedthrough discussion. Where only abstracts were available, orinsufficient information was reported, authors were contactedvia email. Study quality was assessed using the Newcastle-Ottawa Scale [27].

Analysis

Extracted data were synthesized using a narrative synthesisframework as follows:

1. An idea of how anxiety relates to QoL and diseaseactivity was developed, to inform decisions about thereview question, studies to include and the interpreta-tion of findings.

2. A preliminary synthesis of the findings of includedstudies was developed. A summary of all the studieswas tabulated, then separate data analysis of primaryand additional outcome measures for QoL and diseaseactivity performed.

3. Relationships between the studies were explored. For ex-ample, potential sources of heterogeneity were consideredthat could explain differences in study findings.

4. The strength of evidence for drawing conclusions andgeneralizing findings to different populations wasassessed, considering methodological quality of includedstudies, risk of bias and overall evidence in relation toeach of the specified outcomes.

Furthermore, a meta-analysis of quantitative data on theprimary outcome measures was performed. The associa-tion between anxiety and the primary outcome measures(DAS28, SF-36) was most frequently reported cross-sec-tionally, using correlation coefficients. Therefore, correla-tion coefficients between anxiety and DAS28 in peoplewith RA were pooled. From the four studies reporting thecorrelation between anxiety and SF-36, two used the PCSand MCS [28, 29], and two reported correlation coeffi-cients for all of the SF-36 subscales [30, 31]. Therefore,correlation coefficients between anxiety and the PCS or“physical functioning” subscale scores were pooled to givean overall impression of the association between anxietyand physical QoL. Correlation coefficients between anxi-ety and the MCS or “mental health” subscale scores werepooled to give an overall estimate of the association be-tween anxiety and mental QoL in people with RA.

As there was evidence of significant heterogeneity be-tween some of the studies, and due to several studies hav-ing small sample sizes, a random effects model was used[32]. Correlation coefficients (r) were converted to Fisher’sz scores r(z′) and an associated standard error calculatedusing the standard formula SE[r(z′)] = √(1/(n − 3)). Ther(z′) values were pooled using the metan command inStata (Version 14.0), then transformed back to obtainpooled r values and plotted together with study-specificestimates. The Cochran Q statistic was used to assess het-erogeneity. In addition, the I2 statistic [32] was calculatedto examine the proportion of total variation in study esti-mates which could be explained by heterogeneity.

Results

Summary of included studies

Figure 1 shows the flow of studies within the review. Thecharacteristics of the 20 studies identified for inclusion in thisreview [28–31, 33–48] are summarized in Table 1 (studiesreporting disease activity outcomes) and Table 2 (studiesreporting QoL outcomes). Included studies involved a totalof 7452 patients, 80% of whom were female, with a meanage of 53.5 years. Sample sizes within primary studies rangedfrom 20 to 2919 with a mean of 372 patients. Sixteen of thestudies were full text articles [28–31, 36–43, 45–48], whilstfour were conference abstracts [33–35, 44]. Fourteen [28–31,35–37, 40, 42, 44–48] were cross-sectional in design, whilstsix were cohort studies [33, 34, 38, 39, 41, 43]. In terms ofoutcomes, eleven of the studies assessed disease activity only[28–44], six assessed QoL only [30, 31, 45–48] and threeassessed both disease activity and QoL [28, 29, 37]. TheHospital Anxiety and Depression Scale (HADS) was the mostfrequently used tool to identify anxiety.

Clin Rheumatol

The primary outcome measure for disease activity, DAS28,was reported in nine studies, with a further study reportingchange in DAS28 by ≤ 0.6 or treatment discontinuation dueto inefficacy as outcomes [46]. Additional outcome measuresfor disease activity across four studies included the LansburyArticular Index (LAI), Thompson Articular Index, Pain VASand a composite score comprising of CRP, SJC and TJC and aPhysician’s global assessment.

The primary outcome measures for QoL (SF-36 or thePCS, MCS or subscale scores of SF-36) were reported in fivestudies. Other outcome measures for QoL reported in fivestudies included Health-Related QoL (HRQoL) using EQ-5D, Life Satisfaction Index (LSI), Nottingham Health Profile(NHP), QoL-RA Scale and World Health Organisation QoL-BREF (WHOQoL-BREF).

Methodological quality assessment for includedstudies

Quality assessment using the Newcastle-Ottawa score is pre-sented in Tables 1 and 2 (for additional details see supplemen-tary data 3) suggesting most studies were of moderate meth-odological quality. Many lacked detail on the representative-ness of their RA sample [28, 29, 36, 37, 40, 44, 46], or on non-

responders [28–31, 33–38, 41, 43–47], whilst several hadsmall sample sizes, limiting generalizability [28, 30, 31, 46].Although most used validated tools to ascertain the exposure,several did not report if they controlled for potential confound-ing factors [34, 40, 44, 45]. Several studies relied on self-reported questionnaires to assess outcome measures, particu-larly for QoL, which were open to bias, whilst some did notdescribe their method of outcome assessment.

All cross-sectional studies apart from one, a conferenceabstract [44], included a description of their approach tostatistical analysis, reporting confidence intervals and Pvalues as appropriate. Across the six included cohort stud-ies, the length of follow-up varied from 6 months to 5 years[33, 34, 38, 39, 41, 43].

Association between anxiety and disease activityin RA

The overall findings of the nine studies that reported the asso-ciation between anxiety and DAS28, the primary outcomemeasure for disease activity, are summarized in Table 1.These included five cross-sectional [28, 35, 37, 42, 44] andfour cohort [33, 34, 38, 39] studies, involving a total of 4718participants. From the five cross-sectional studies, three founda significant correlation [28, 35, 42] and two a non-significantcorrelation [37, 44] between anxiety and increased DAS28.All five of these studies reported correlation coefficients forthe cross-sectional association between anxiety and DAS28,and results were pooled via meta-analysis. Meanwhile, fromthe four cohort studies, two found significant associations be-tween baseline anxiety and increased DAS28 at 6 and12 months [34, 39], one showed a significant association at3 months but not at 9 and 15months [38], whilst another studyfound a non-significant association between anxiety andDAS28 at 6-month follow-up [33].

One further study reported disease activity in terms of non-response to treatment with methotrexate, being indicated as animprovement in DAS28 by ≤ 0.6 or treatment discontinuationdue to inefficacy [43]. On multivariable analysis, higher base-line anxiety was found to significantly predict treatment non-response after 6 months, OR 1.07 (95% CI 1.03–1.12).

Four studies also reported DAS28 subcomponents [33, 35,39, 42]. Results suggested anxiety was most significantly as-sociated with the more subjective assessments of disease ac-tivity, such as PtGA, Pain VAS and TJC. For example,Matcham et al.’s cohort study found anxiety to be positivelycorrelated at baseline and at 1-year follow-up with PtGA, TJC,SJC and ESR, although this correlation was only significantfor PtGA (r = 0.31, p < 0.05 at baseline, r = 0.43, p < 0.01 at 1-year follow-up) [39]. Onmultiple regression analysis, baselineanxiety was also significantly associated with PtGA and TJCat 1-year follow-up [39]. In another study, Pain VAS and TJCwere significantly higher in people with anxiety although

Fig. 1 Study flowchart

Clin Rheumatol

Table1

Sum

maryof

characteristicsof

studiesreportingdiseaseactiv

ityoutcom

emeasures

Author/year/country

Studydesign

Samplesize

(m:f)

Meanage

Exposuremeasure

Outcomemeasure

Results

Studyquality

a

Al-Fadl

(28)

2014.E

gypt

Cross-sectio

nal

26(8:18)

43.4

HAM-A

DAS2

8HAMA-A

andDAS2

8r=

0.47*

Medium

Dyball(33)2018.U

KCohort

2919

(701:2

218)

57.3

HADS-A

DAS2

8BaselineHADS-Aandchange

inDAS2

86months

Adjustedb=0.01

(n.s.),

95%

CI,−0.01,0.04.

Medium

Fragoulis

(34)

2018.U

KCohort

848(not

reported)

Not

reported

HADS-A

DAS2

8HADS-AandDAS28

at:

-baseliner=

novaluegiven***

-6monthsr=

0.230***,-12

monthsr=

0.217***

Medium

Grosso(35)

2015.Italy

Cross-sectio

nal

200(29:171)

62.4

HADS-A

DAS2

8Anxiety

vs.health

ycontrol.MeanDAS2

8score(SD):3.38

(1.18)

vs.2.48(0.78)**

Medium

Ichikawa(36)

1995.Japan

Cross-sectio

nal

92(16:76)

53.4

BAI

Pain

VAS

Anxiety

andPain

VASr=

0.432***

Medium

LAI

Anxiety

andLAIr=

0.237*

Karahan

(37)

2016.T

urkey

Cross-sectio

nal

148(32:116)

51.1

Zung’sSelf-RatingScale

DAS2

8BAIandDAS2

8r=

0.159,(n.s.)

Medium

Kojim

a(29)

2009.Japan

Cross-sectio

nal

120(22:98)

57.7

HADS-A

SJC,T

JC,C

RP,

Physicians’globalassessment

HADS-Aanddiseaseactiv

ityfactor

loading−0.10

(n.s.)

Medium

Kuijper

(38)

2018.

Netherlands

Cohort

281(91:190)

53.0

HADS-A

DAS2

8HADS-AandDAS2

8at:

-3months,b=0.043**,95%

CI,0.013,0.073

-9months,b=0.017(n.s.),95%

CI,0.010,0.044

-15months,b=0.012,(n.s.),95%

CI,−0.020,0.043

High

Matcham

(39)

2016.U

KCohort

56(12:44)

53.6

HADS-A

DAS2

8Anxiety

andDAS2

8-Baseline,r=

0.29*,1-year

follo

w-up,r=

0.33*

High

Unadjustedb0.04*.95%

CI,0.00,0.07.

Adjustedb=0.02

(n.s.),

95%

CI,0.00,0.05.

Miwa(40)

2002.Japan

Cross-sectio

nal

82(20:62)

62.0

HADS-A

VAS

Anxiety

andVAS

(a)Meanactiv

ity<0.5,r=

0.2935*

(b)Meanactiv

ity≥0.5,r=

−0.0269

(n.s.)

Medium

Overm

an(41)

2011.

Netherlands.

Cohort

545(168:377)

56.0

10-Item

Anxiety

Scale

Thompson

Articular

Index

Anxiety

andThompson

Articular

Index

b(SE

)=55.1736(21.0731)**

Medium

ESR

Anxiety

andESR

,b(SE)=

0.2448

(0.0759)**

Ruhalia(42)

2018.

Malaysia

Cross-sectio

nal

189

(22:167)

49.6

DASS

(21)

DAS2

8DASS

(21)

andDAS2

8r=

0.233**

Medium

Sergeant

(43)

2018.U

KCohort

1050

(343:707)

59.0

HADS-A

Non-responseto

treatm

ent

(improvem

entin

DAS2

8≤0.6/discontin

uation

dueto

inefficacy)

HADS-Aandnon-response

at6months

-Multiv

ariableanalysis

OR(95%

CI)=1.07

(1.03,1.12)**

High

Zulgerel(44)

2014.R

ussia

Cross-sectio

nal

51(2:49)

43.0

SpielbergChennai

DAS2

8Anxiety

andDAS2

8r=

0.126(n.s.)

Low

rPearson’scorrelationcoefficient,bmultip

leregression

coefficient,CIconfidence

interval,SE

standard

error,ORodds

ratio

,SD

standard

deviation,

n.s.notsignificant,(p>0.05),*(p<0.05),**

(p<0.01),***(p<0.001)

BAIB

eck’sAnxietyInventory,CRPC-reactiveprotein,DAS28Disease

Activity

Scorein28

Joints,D

ASS21

Depression,AnxietyandStressScale,ESR

ErythrocyteSedim

entatio

nRate,HADSHospital

Anxiety

andDepressionScore,HAM-A

Ham

ilton

Anxiety

RatingScale,L

AILansbury’sActivity

Index,

RARheum

atoidArthritis,SJCSw

ollenJointCount,S

TAISpielberger

StateandTraitAnxiety

Inventory,TJCTender

JointC

ount,V

ASVisualA

nalogueScale

aQualityratedouto

f9usingNew

castle-O

ttawaScale:0

–2=lowquality,3–6

=medium

quality

and7–9=high

quality

Clin Rheumatol

Table2

Sum

maryof

characteristicsof

studiesreportingQoL

outcom

emeasures

Author/year/country

Studydesign

Sample

size

(m:f)

Meanage

Exposuremeasure

Outcomemeasure

Results

Studyquality

a

Al-Fadl

[28]

2014.E

gypt

Cross-sectio

nal

26(8:18)

43.4

HAM-A

PCS/MCS

Anxiety

andPC

Sof

SF-36,r=

−0.38*

Anxiety

andMCSof

SF-36,r=

−0.34*

Medium

Alpi(45)2016.C

olom

bia

Cross-sectio

nal

62(3:59)

59.7

HADS-A

QoL

-RAScale

Anxiety

andQoL

-RAScale,r=

−0.644**

Anxiety

andQoL

-RAsub-scores;

Physicalability,r

=−0.492**,Su

pport,

r=−0.454**

Pain,r

=−0.489***,T

ension,r

=−0.581**

Health

,r=−0.624**,Arthritis,r=

−0.510**

Support,r=

−0.593**,Mood,r=

−0.674**

Medium

Anxiety

andQOL-RA,regressioncoefficient

b=−0.453(t=−4.241,p=0.000)

Celiker

(46)

2001.T

urkey

Cross-sectional

20(0:20)

46.6

STAI

LSI

A-State(current

anxiety)

andLSI,r

=−0.5005*

A-Trait(long-standing

anxiety)

andLSI,

r=−0.5103

(n.s.)

Medium

Karahan

(37)2016.Turkey

Cross-sectio

nal

148(32:116)

51.1

BAI

WHOQoL

-BREF

Mild

anxietyandWHOQoL

-BREF–

n.s.

ModeratetosevereanxietyandWHOQoL

-BREF–

*Medium

Kojim

a(29)

2009.Japan

Cross-sectio

nal

120

(22:98)

57.7

HADS-A

PCS/MCS

HADS-AandPC

Sof

SF-36,r=

−0.25**

HADS-AandMCSof

SF-36,r=

−0.51***

Medium

Mok

(47)

2012.C

hina

Cross-sectional

200(42:158)

51.4

Chinese

Bilingual

StructuredInterview

SF-36

Anxiety

vsno

psychiatricdisorders,mean(SD)

SF36

=31.2(12.9)***vs.56.6(20)**

Medium

Nas

(30)

2011.T

urkey

Cross-sectio

nal

421(72:349)

50.2(riskof

anxietygroup)

HADS-A

SF-36

High-RiskAnxiety

andSF

-36subscales.All*

**Ph

ysicalfunctio

ning

(r=−0.28),physicalrole

(r=−0.24)

Bodily

pain

(r=−0.32),Generalhealth,r

=−0.29,

Vitality(r=−0.40),So

cialfunctio

ning

(r=−0.36)

Emotionalrole(r=−0.23)Mentalh

ealth

(r=−0.48)

Medium

Anxiety

(n=166)

vs.w

ithoutanxiety

(n=255).

SF-36subscales,adjusted

OR(95%

CI),signifi-

cant

initalics.

Physicalfunctioning,0.991

(0.983,0.999),Ph

ysical

role,0.994

(0.998,1.001),

Bodily

pain,0.990

(0.971,

1.001),G

eneral

health,0.987

(0.977,0.998),Vitality,

0.972(0.961,0.984),Social

functioning,0.984,

(0.975,0.995),Emotionalrole,0.992(0.986,0.998)

Mentalh

ealth,0.966

(0.954,0.977)

RA-Q

oLAnxiety

(n=166)

vs.w

ithoutanxiety

(n=255),R

A-Q

oL,

OR(95%

CI)=1.060(1.032,1.088)

NHP

Anxiety

(n=166)

vs.w

ithoutanxiety

(n=255),

NHPsubscales,OR(95%

CI).S

ignificantin

italics.

Pain,1.011(1.004,1.018),Physicalm

obility1.017

(1.009,1.025),Energy,1.015(1.009,1.021),Sleep

1.018(1.011,1.025),Social

isolation,1.017(1.011,

1.023),E

motionalreaction,1.026(1.019,1.033)

Clin Rheumatol

there was no significant difference in ESR or SJC [35]. Afurther study found anxiety to be significantly correlated withincreased Pain VAS (r = 0.341, p < 0.001) and TJC (r = 0.197,p = 0.007), but not SJC (r = 0.060, p = 0.412) (51). Dyballet al. [33] found baseline anxiety to be associated with smallerimprovements in PtGA at 6 months (b = 0.74, 95% CI 0.32,1.26, p = 0.001) [33].

Four studies reported additional disease activity outcomemeasures, detailed in Table 1. These included one cohort study[41] and three cross-sectional studies [29, 36, 40], involving839 patients. Overall, anxiety was associated with ESR [41],the Thompson articular index [41], LAI [33], pain VAS [36]and VAS as a measure of arthralgia, in individuals with lowermean activity [40]. However, one study found no associationbetween anxiety and disease activity, when assessed using acombination of the physician’s global assessment, TJC, SJCand CRP [29], potentially reflecting discrepancies betweenpatient and physician global assessments.

Relationship between anxiety and QoL in RA

The findings of the five studies reporting SF-36 [28–31, 47]are summarized in Table 2. These were all cross-sectionalstudies and involved 801 patients. In the majority of studies,anxiety had a significant negative association with SF-36 andits subscale scores indicating worse QoL.

In particular, a significant negative correlation was reportedbetween anxiety and the PCS and MCS of SF-36 [28, 29]. Afurther study found a significant negative correlation betweenanxiety and all SF-36 subscale scores apart from emotionalrole [31]. Another study found a significant negative correla-tion between high-risk anxiety and all SF-36 subscales [30],whilst a comparison of RA patients with and without anxietyfound that all SF-36 subscale scores apart from physical roleand bodily pain were significantly worse in patients with anx-iety. One study also reported mean SF-36 scores to be signif-icantly worse in people with anxiety compared with thosewithout a psychiatric disorder [47]. Four of these studies re-ported correlation coefficient between anxiety and SF-36 andits subscales [28–31], hence these were pooled via meta-analysis.

Different QoL outcome measures were reported in fivestudies [30, 37, 45, 46, 48], which are summarized inTable 2. These were all cross-sectional studies, involving759 patients. A significant negative correlation was foundbetween moderate or severe anxiety and the Turkish versionof WHOQoL-BREF score [37]. Current anxiety and the LSI[46], EQ-5D [48] and QoL-RA Scale [45] also had a signifi-cant negative correlation. In addition, multiple regressionanalysis found anxiety, when measured using the QoL-RAScale negatively impacted on QoL [45]. Patients with anxietyalso reported significantly worse NHP subscale scores, andT

able2

(contin

ued)

Author/year/country

Studydesign

Sample

size

(m:f)

Meanage

Exposuremeasure

Outcomemeasure

Results

Studyquality

a

Ozcetin

(31)

2007.T

urkey

Cross-sectional

34(8:26)

50.4

BAI

SF-36

Beckanxietyscores

andSF

-36subscalescores

Physicalfunctioning

(r=−0.672***),Ph

ysicalrole

(r=−0.451**),B

odily

pain

(r=−0.583***),

Generalhealth

(r=−0.706***),Vitality(r=−0.737***),So

cial

functio

ning

(r=−0.718***),Mental

health(r=−0.655***)

Emotionalrole(r=−0.326(n.s.))

Medium

Wan

(48)

2015.S

ingapore

Cross-sectio

nal

108

(22:86)

56.4

HADS-A

HRQoL

(EQ-5D)

Anxiety

andHRQoL

usingEQ-5D,r

=−0.58**

High

rPearson’scorrelationcoefficient,bmultip

leregression

coefficient,CIconfidence

interval,O

Rodds

ratio

,n.s.not

significant,(p>0.05),*(p<0.05),**

(p<0.01),***(p<0.001).

BAIB

eck’sAnxiety

Inventory,EQ-5DEuroQ

uol5

-Dim

ension

Scale,H

ADS-AHospitalA

nxiety

andDepressionScore(anxiety

scale),H

AM-A

Ham

ilton

Anxiety

RatingScale,HRQoL

Health

-Related

QualityofLife,LS

ILifeSatisfactionIndex,MCSMentalC

omponentSu

mmary,NHPNottin

gham

Health

Profile,P

CSPh

ysicalCom

ponentSu

mmary,RARheum

atoidArthritis,STAIS

pielbergerStateand

TraitAnxiety

Inventory,SF

-36Sh

ort-Fo

rm36,W

HOQoL

-BREFWorld

Health

OrganisationQualityof

Life-BREF,QoL

-RAscaleQualityof

LifeRheum

atoidArthritisScale

aQualityassessed

usingNew

castle-O

ttawaScale:0

–2=low,3–6

=medium

and7–9=high

Clin Rheumatol

RA-QoL subscale scores [30], compared with those withoutanxiety.

Meta-analysis

Figure 2 shows the variation in correlation coefficients acrossstudies for DAS28, physical and mental QoL outcomes. Therewas evidence of significant heterogeneity between studies forphysical QoL (Q statistic p = 0.003; I2 = 78.5%), with relative-ly lower heterogeneity between studies for mental QoL (p =0.341; I2 = 10.3%) and DAS28 (p = 0.427; I2 = 0%). The

strongest pooled association was found between anxiety andworse mental QoL (pooled r (CI) = − 0.50 (− 0.57, − 0.43)).

Strength of evidence

The overall strength of evidence for the relationship betweenanxiety in RA and disease activity and QoL has beenreviewed in Table 3, using the Modified Grading ofRecommendations Assessment, Development andEvaluation (GRADE) system [49]. Most studies reportingdisease activity outcomes had consistent effect sizes, moder-ate sample sizes and controlled for confounding factors,whilst heterogeneity on meta-analysis of the correlation be-tween anxiety and DAS28 was low. Therefore, the strengthof evidence for the association between anxiety and diseaseactivity was felt to be moderate. Meanwhile, although effectsizes were relatively consistent for QoL outcomes, almosthalf of the studies had small sample sizes and there wassignificant heterogeneity on meta-analysis of physical QoL,meaning the overall strength of evidence for the associationbetween anxiety and QoL was low.

Discussion

This is the first systematic review to examine the relationshipbetween anxiety and QoL and disease activity in patients withRA. Findings suggest that anxiety is associated with increaseddisease activity and reduced QoL.

This systematic review demonstrates that anxiety in pa-tients with RA is associated with increased disease activity,both cross-sectionally and at 3 months [38], 6 months [34] and12 months [39] follow-up. These findings complement previ-ous research, showing that depression in RA is associatedwithincreased disease activity [8].

Whilst one study (involving 52 participants) found the im-pact of baseline anxiety on disease activity to increase be-tween 6- and 12-month follow-up [39], two larger studies(involving 281 and 848 participants) found the impact of base-line anxiety to reduce over time [34, 38]. The reasons for thisobservation are unclear, with Kuijper et al. [38] hypothesisingthat in early RA, when treatment has not yet been optimized,anxiety could influence subjective components of DAS28,whilst once disease is better controlled, patients may adaptto living with RA, leading to a reduction in anxiety.

This review suggests the association between anxiety andincreased disease activity may, in part, be due to people withanxiety reporting higher PtGA scores and other subjectivemeasures of disease activity [33, 35, 39, 42], rather than anobjective increase in disease activity.

Considering QoL outcomes, this systematic review alsodemonstrates that anxiety in patients with RA is cross-sectionally associated with reduced QoL, complementing

Al-Fadl [28] 2014.

Karahan [37] 2016.

Matcham [39] 2016.

Zulgerel [44] 2014

Overall (I-squared = 0%, p= 0.427)

NOTE: Weights are from random effects analysis

First Author/ Yearn= number of par�cipants

r (95% CI)

0.47 (0.10, 0.73)

0.16 (-0.00, 0.31)

0.29 (0.03, 0.51)

0.13 (-0.16, 0.39)

0.23 (0.09, 0.36)Ruhalia [42] 2018.

0.23 (0.14, 0.31)

n= 26

n= 148

n= 56

n= 189

n= 51

a : Anxiety and DAS28

Kojima [29] 2009.

Overall (I-squared = 78.5%, p= 0.003)

NOTE: Weights are from random effects analysis

r (95% CI)

-0.38 (-0.67, 0.01)

-0.25 (-0.41, -0.07)

-0.28 (-0.37, -0.19)

-0.67 (-0.82, -0.43)

-0.39 (-0.57, -0.20)

: Anxiety and Physical QoLFirst Author/ Yearn= number of par�cipants

n= 26Al-Fadl [28] 2014.

n= 421

n= 34

Nas [30] 2011.

Ozce�n [31] 2007.

n= 120

Al-Fadl [28] 2014.

Kojima [29] 2009.

Nas [30] 2011.

Ozce�n [31] 2007.

Overall (I-squared = 10.3%, p= 0.341)

NOTE: Weights are from random effects analysis

r (95% CI)

-0.34 (-0.64, 0.05)

-0.51 (-0.63, -0.36)

-0.48 (-0.55, -0.46)

-0.65 (-0.81, -0.41)

-0.50 (-0.57, -0.43)

a: Anxiety and Mental QoLFirst Author/ Yearn= number of par�cipants

n= 26

n= 120

n= 421

n= 34

b

c

Fig. 2 Forest plots of the meta-analysis of correlation coefficients be-tween A: Anxiety and DAS28, B: Anxiety and Physical QoL and C:Anxiety and Mental QoL

Clin Rheumatol

previous research showing that depression in RA is associatedwith poorer QoL [17]. The largest impact of anxiety on SF-36subscales was seen on the vitality, social functioning and men-tal health components [30, 31]. Effect sizes for reduced socialfunctioning and mental health scores could be interpreted assymptoms of anxiety as well as components of QoL, whilstthe correlation between anxiety and vitality scores could befurther confounded by associations of mood problems andfatigue in RA [50] likely to impact negatively on QoL.

A number of strengths and limitations need to be consid-ered when interpreting the results. The search strategy waskept broad; hence, we are confident that all available evidenceon this topic has been considered. However, nine potentiallyrelevant studies could not be included as the published con-ference abstracts lacked sufficient data, and the authors did notrespond to requests to provide additional data. Furthermore,twelve articles had to be excluded as anxiety was not reportedseparately from depression as an outcome. As anxiety fre-quently exists in isolation, future studies should considerreporting anxiety and depression separately, to enable theirindividual effects to be analyzed.

Meta-analysis was limited by the small number of studieswhich provided suitable data to enable statistical pooling ofresults. Where pooled estimates of correlation were obtained,these, as well as the associated measures of extent of hetero-geneity, should be interpreted with caution. Potential sourcesof heterogeneity were differences in the source populations,sampling methods and adjustments made for confoundingfactors between studies. In the future, availability of suitabledata may allow the impact of individual factors on the out-comes to be investigated using meta-regression analysis.

All available evidence regardless of the methodologicalstudy quality was incorporated within this review. Overall,the quality of the included studies was varied. Several poten-tial sources of bias were identified in terms of the samplingmethods, sample sizes, method of outcome measurement andloss of participants to follow-up.

Clinicians should consider reviewing DAS28 subcompo-nents in patients who fail to respond to treatment, especially ifsubjective measures (PtGA, TJC) are high relative to moreobjective scores (SJC, ESR), since this could indicate under-lying mood problems rather than high disease activity due to

Table 3 Strength of evidence for association between anxiety, disease activity and QoL outcomes

Outcome Evidence base Strength of association Strength of evidence(GRADE)a.b

Comments

Disease activity 14 studiesn = 66078 cross-sectional

(28, 29, 35–37,40, 42, 44)

6 cohort (33, 34, 38,39, 41, 43)

Meta-analysisAnxiety correlation with DAS28Pooled r (CI) = 0.23 (0.14, 0.31)Narrative synthesisCorrelation between anxiety and increased

disease activity- 7 studies significant- 3 studies non-significantAssociation between anxiety and increased

disease activity- 2 studies significant baseline to 12 M- 1 study significant 3 M,

non-significant 9 M/15 M- 1 study non-significant 6 M

Moderate⊕⊕⊕

Most studies had consistentlysmall/moderate effect sizes, areasonable sample size and con-trolled for confounding factors.There was low heterogeneity inthe meta-analysis.

Quality of life 9 Studiesn = 1139All cross-sectional

(28–31, 37,45–48)

Meta-analysisAnxiety correlation with Physical QoLPooled r (CI) = − 0.39 (− 0.57, − 0.20)Anxiety correlation with Mental QoLPooled r (CI) = −0.50 (− 0.57, − 0.43)Narrative synthesisCorrelation between anxiety and reduced QoL- 6 studies significant- 1 study non-significant with mild anxiety but

significant with moderate/severe anxiety- 1 study non-significant with long-standing

anxiety but significant with current anxiety

Low⊕⊕

Most studies had consistentlysmall/moderate effect sizes,though nearly half of the studieshad small sample sizes and therewas significant heterogeneity onmeta-analysis of mental QoL,limiting strength of evidence.

a GRADE assessment included risk of bias, inconsistency, indirectness, imprecision, large effect (strength of association) and dose-response gradientb Symbols for quality of evidence: ⊕⊕⊕⊕High, further research is unlikely to change our confidence in the estimate of effect; ⊕⊕⊕ Moderate,further research is likely to have an important impact on our confidence in the estimate of effect andmay change it;⊕⊕ Low, further research is likely tohave an important impact on our confidence in the estimate of effect and is likely to change the estimate; ⊕Very low, any estimate of effect is veryuncertain

Clin Rheumatol

inflammation, and as such managing mood alongside inflam-mation could lead to improved outcomes.

In conclusion, comorbid anxiety in patients with RA isassociated with increased disease activity, poorer treatmentoutcomes and reduced QoL. Better identification and manage-ment of comorbid anxiety by patients and clinicians has thepotential to improve outcomes for people with RA.

Funding information TThis research is funded by the National Institutefor Health Research School for Primary Care Research (NIHR SPCR)[AM]. This research is also partfunded by the Haywood Foundation[AM]. In addition, this is funded by a Practitioner Grant from theScientific Foundation Board of the Royal College of GeneralPractitioners (SFB RCGP) grant number 2018-02 [AM].

Compliance with ethical standards

Disclaimer The views expressed are those of the author and not neces-sarily those of the NIHR or the Department of Health and Social Care.

Disclosures None.

References

1. Abhishek A, Doherty M, Kuo C, Mallen C, Zhang W, Grainge MJ(2017) Rheumatoid arthritis is getting less frequent- results of anationwide population-based cohort study. Rheumatology 56(5):736–744. https://doi.org/10.1093/rheumatology/kew468

2. Covic T, Cumming SR, Pallant JF, Manolios N, Emery P,Conaghan PG, Tennant A (2012) Depression and anxiety in pa-tients with rheumatoid arthritis: prevalence rates based on a com-parison of the depression, anxiety and stress scale (DASS) and thehospital, anxiety and depression scale (HADS). BMC Psychiatry12:6. https://doi.org/10.1186/1471-244X-12-6

3. Matcham F, Rayner L, Steer S, Hotopf M (2013) The prevalence ofdepression in rheumatoid arthritis: a systematic review and meta-analysis. Rheumatology 52(12):2136–2148. https://doi.org/10.1093/rheumatology/ket169

4. NatCen Social Research (2016) Adult Psychiatric MorbiditySurvey: Survey of Mental Health and Wellbeing, England, 2014.NHS Digital. Available from: http://content.digital.nhs.uk/catalogue/PUB21748 [Accessed 1st August 2019]

5. CepoiuM,McCusker J, ColeMG, SewitchM, Belzile E, Ciampi A(2007) Recognition of depression by non-psychiatric physicians—asystematic literature review and meta-analysis. J Gen Intern Med23(1):25–36. https://doi.org/10.1007/s11606-007-0428-5

6. Bazzichi L, Maser J, Piccinni A, Rucci P, Del Debbio A, VivarelliL, Catena M, Bouanani S, Merlini G, Bombardieri S, Dell’Osso L(2005) Quality of life in rheumatoid arthritis: impact of disabilityand lifetime depressive spectrum symptomatology. Clin ExpRheumatol 23(6):783–738

7. AngDC, Choi H, KroenkeK,Wolfe F (2005) Comorbid depressionis an independent risk factor for mortality in patients with rheuma-toid arthritis. J Rheumatol 32(6):1013–1019. https://doi.org/10.1007/s11920-008-0042-1

8. Rathbun AM, Reed GW, Harrold LR (2013) The temporal relation-ship between depression and rheumatoid arthritis disease activity,treatment persistence and response: a systematic review.Rheumatology 52(10):1785–1794. https://doi.org/10.1093/rheumatology/kes356

9. DiMatteo M, Lepper HS, Croghan TW (2000) Depression is a riskfactor for noncompliance with medical treatment: meta-analysis ofthe effects of anxiety and depression on patient adherence. ArchIntern Med 160(14):2101–2107. https://doi.org/10.1001/archinte.160.14.2101

10. Covic T, Adamson B, Spencer D, Howe G (2003) Abiopsychosocial model of pain and depression in rheumatoid ar-thritis: a 12-month longitudinal study. Rheumatology 42(11):1287–1294. https://doi.org/10.1093/rheumatology/keg369

11. Greenen R, Van Middendorp H, Bijldma JW (2006) The impact ofstressors on health status and hypothalamic-pituitary-adrenal axisand autonomic nervous system responsiveness in rheumatoid arthri-tis. Ann N YAcad Sci 1069:77–97. https://doi.org/10.1196/annals.1351.007

12. Duarte C, Ferreira R, Batista S, Medeiros C, Sousa J, Eugenio Get al (2015) Pain and anxiety are the major predictors of discrepan-cies between patients’ and physicians’ perception of disease activityin rheumatoid arthritis. Ann Rheum Dis 74:671–672. https://doi.org/10.1136/annrheumdis-2015-eular.4266

13. Cordingley L, Prajapati R, Plant D, Maskell D, Morgan C, Ali FR,Morgan AW, Wilson AG, Isaacs JD, Biologics in RheumatoidArthritis Genetics and Genomics Study Syndicate (BRAGGSS),Barton A (2014) Impact of psychological factors on subjective dis-ease activity assessments in patients with severe rheumatoid arthri-tis. Arthritis Care Res 66(6):861–868. https://doi.org/10.1002/acr.22249

14. Matcham F, Davies R, Hotopf M, Hyrich KL, Norton S, Steer S,Galloway J (2018) The relationship between depression and bio-logic treatment response in rheumatoid arthritis: an analysis of theBritish Society for Rheumatology Biologics Register.Rheumatology (Oxford) 57(5):835–843. https://doi.org/10.1093/rheumatology/kex528

15. Cook MJ, Diffin J, Scirè CA, Lunt M, MacGregor AJ, SymmonsDPM, Verstappen SM (2016) Predictors and outcomes of sustained,intermittent or never achieving remission in patients with recentonset inflammatory polyarthritis: results from the NorfolkArthritis Register. Rheumatology (Oxford) 55(9):1601–1609.https://doi.org/10.1093/rheumatology/kew210

16. Senra H, Rogers H, Leibach G, Altamar MLP, Plaza SLO, Perrin P,Durán MAS (2017) Health-related quality of life and depression ina sample of Latin American adults with rheumatoid arthritis. Int JRheumDis 20(11):1684–1693. https://doi.org/10.1111/1756-185X.12412

17. Matcham F, Norton S, Scott DL, Steer S, Hotopf M (2016)Symptoms of depression and anxiety predict treatment responseand long-term physical health outcomes in rheumatoid arthritis:secondary analysis of a randomized controlled trial .Rheumatology 55(2):268–278. https://doi.org/10.1093/rheumatology/kev306

18. Kaufman J, Charney D (2000) Comorbidity of mood and anxietydisorders. Depress Anxiety 12:69–76. https://doi.org/10.1002/1520-6394(2000)12:1+<69::AID-DA9>3.0.CO;2-K

19. Qui X, ZhangX, Cai L, Yan C, Yu L, Fan J et al (2019) Rheumatoidarthritis and risk of anxiety: a meta-analysis of cohort studies. ClinRheumatol 38(8):2053–2061. https://doi.org/10.1007/s10067-019-04502-8

20. National Institute of Health and Care Excellence. Depression inadults: recognition and management (CG90). NICE guideline;2009. Available from: https://www.nice.org.uk/guidance/cg90.Accessed 1 Aug 2019

21. National Institute of Health and Care Excellence. Generalised anx-iety disorder and panic disorder in adults: management (CG113).NICE guideline; 2011. Available from: https://www.nice.org.uk/guidance/cg113. Accessed 1 Aug 2019

22. Machin A (2017) A systematic review to examine the impact ofanxiety on quality of life and treatment response in patients with

Clin Rheumatol

rheumatoid arthritis. PROSPERO 2017 CRD42017062580Available from: http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42017062580

23. Ware JE Jr, Sherbourne CD (1992) The MOS 36-item short-formhealth survey (SF-36). Conceptual framework and item selection.Med Care 30(6):473–483. https://doi.org/10.1097/00005650-199206000-00002

24. Brazier JE, Harper R, Jones NM, O’Cathain A, Thomas KJ,Usherwood T (1992) Validating the SF-36 heath survey question-naire: new outcome measure for primary care. BMJ 305(6846):160–164. https://doi.org/10.1136/bmj.305.6846.160

25. Prevoo ML, van’t Hof MA, Kuper HH, van Leeuwen MA, van dePutte LB, van Riel PL (1995) Modified disease activity scores thatinclude twenty-eight–joint counts: development and validation in aprospective longitudinal study of patients with rheumatoid arthritis.Arthritis Rheum 38:44–48. https://doi.org/10.1002/art.1780380107

26. National Institute of Health and Care Excellence (2018)Rheumatoid arthritis: the management of rheumatoid arthritis inadults. London: National Institute for Health and Care Excellence.Available from: https://www.nice.org.uk/guidance/ng100[Accessed 1st August 2019]

27. Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos P,Tugwell P (2014) The Newcastle-Ottawa Scale (NOS) for assessingthe quality of nonrandomised studies in meta-analyses. The OttawaHospital Research Institute. Available from: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp

28. Al-Fadl EM, Ismail MA, Thabit M, ElSerogy Y (2014) Assessmentof health-related quality of life, anxiety and depression in patientswith early rheumatoid arthritis. Egypt Rheum 36(2):51–56. https://doi.org/10.1016/j.ejr.2013.12.004

29. Kojima M, Kojima T, Ishiguro N, Oguchi T, Oba M, Tsuchiya H,Sugiura F, Furukawa TA, Suzuki S, Tokudome S (2009)Psychosocial factors, disease status, and quality of life in patientswith rheumatoid arthritis. J Psychosom Res 67(5):425–431. https://doi.org/10.1016/j.jpsychores.2009.01.001

30. Nas K, Sarac AJ, Gur A, Cevik R, Altay Z, Erdal A, Ersoy Y, KayaA, Tekeoglu I, Ugur M, Durmus B, Ardicoglu O, Senel K, BaysalO, Kamanli A, Karatay S, Yildirim K, Ozgocmen S (2011)Psychological status is associated with health related quality of lifein patients with rheumatoid arthritis. J Back Musculoskelet Rehabil24(2):95–100. https://doi.org/10.3233/BMR20110281

31. Ozcetin A, Ataoglu S, Kocer E, Yazici S, Yildiz O, Ataoglu A et al(2007) Effects of depression and anxiety on quality of life patientswith rheumatoid arthritis, knee osteoarthritis and fibromyalgia syn-drome. West Indian Med J 56(2):122–129. https://doi.org/10.1590/s0043-31442007000200004

32. DerSimonian R, Laird N (1989) Meta-analysis in clinical trials.Control Clin Trials 7(3):177–188. https://doi.org/10.1016/0197-2456(86)90046-2

33. Dyball S, Cordingley L, Isaacs JD, Wilson G, Morgan A, Hyrich Ket al (2018) Pre-treatment anxiety and depression levels are corre-lated with less improvement in self-reported disease activity inrheumatoid arthritis patients treated with biological therapy: resultsfrom the BRAGGS cohort. Rheumatology 57(3):key075.455.https://doi.org/10.1093/rheumatology/key075.455

34. Fragoulis GE, Cavanagh J, Derakhshan MH, Paterson C, Porter D,McInnes IB et al (2018) Depression and anxiety in an early rheu-matoid arthritis inception cohort. Associations with epidemiologi-cal, socioeconomic and disease features. Ann Rheum Dis 77:221.https://doi.org/10.1136/annrheumdis-2018-eular.7083

35. Grosso V, De Nard F, Breda S, Monti S, Todoerti M, Caporali Ret al (2015) Anxiety and depression in patients with rheumatoidarthritis treated with biological drugs: prevalence and relationshipwith disease activity. Ann Rheum Dis 74:1027. https://doi.org/10.1136/annrheumdis-2015-eular.3290

36. IchikawaY, Horiki T, Hosaka T, HoshinaY,Morita K, UchiyamaMet al (1995) Psychological stress, disease activity and functionaldisability in patients with rheumatoid arthritis: a comparison withother chronic rheumatic or autoimmune disease patients. Jpn JRheum 6(1):9–17 Available from: https://search.jamas.or.jp/link/ui/1996169788 [Accessed 1st August 2019]

37. Karahan AY, Kucuk A, Balkarli A, Kayhan F, Ozhan N, Nas Oet al (2016) Alexithymia, depression, anxiety levels and qual-ity of life in patients with rheumatoid arthritis. Acta med.Mediterr 32(5):1675–1682 Available from: http://www.actamedicamediterranea.com/archive/2016/medica-5/alexithymia-depression-anxiety-levels-and-quality-of-life-in-patients-with-rheumatoid-arthritis/pdf [Accessed 1st August2019]

38. Kuijper TM, Luime JJ, Xiong H, de Jong PHP, van der LubbePAHM, van Zeben D et al (2018) Effects of psychological factorson monitoring treatment effect in newly diagnosed rheumatoid ar-thritis patients over time: response data from the tREACH study.Scand J Rheumatol 17(30):178–184. https://doi.org/10.1080/03009742.2017.1349176

39. Matcham F, Ali S, Irving K, Hotopf M, Chalder T (2016) Aredepression and anxiety associated with disease activity in rheuma-toid arthritis? A prospective study. BMCMusculoskelet Disord 17:155. https://doi.org/10.1186/s12891-016-1011-1

40. Miwa Y, Yajima N, Shiozawa F, Yoda Y, Hanaoka R, Hanyuda Met al (2002) Relationship between psychological factors and arthral-gia in patients with rheumatoid arthritis. Mod Rheumatol 12(1):32–36. https://doi.org/10.3109/s101650200005

41. Overman C, Bossema ER, van Middendorp H, Meij LW,Verstappen SMM, Bulder M et al (2012) The prospective associa-tion between psychological distress and disease activity in rheuma-toid arthritis: a multilevel regression analysis. Ann Rheum Dis71(2):192–197. https://doi.org/10.1136/annrheumdis-2011-200338

42. Ruhalia AR, Cheng CH (2018) Self-reported symptoms of depres-sion, anxiety and stress among patients with Rheumatoid Arthritisin a Malaysian rheumatology centre - prevalence and correlates.Med J Malaysia 23(4):226–232 Available from: http://www.e-mjm.org/2018/v73n4/rheumatoid-arthritis.pdf [Accessed 1stAugust 2019]

43. Sergeant JC, Hyrich KL, Anderson J, Kopec-Harding K, Hope HF,Symmons DPM, RAMS Co-Investigators, Barton A, VerstappenSMM (2018) Prediction of primary non-response to methotrexatetherapy using demographic, clinical and psychosocial variables:results from the UK Rheumatoid Arthritis Medication Study(RAMS). Arthritis Res Ther 20:147. https://doi.org/10.1186/s13075-018-1645-5

44. Zulgerel D, NandinErdene D (2014) Quality of life, depression,anxiety in rheumatoid arthritis patients. Int J Rheum Dis 17(2):8Available from: https://onlinelibrary.wiley.com/doi/pdf/10.1111/1756-185X.12504 [Accessed 1st August 2019]

45. Alpi SV, Quiceno JM, Lozano F, Romero S (2017) Health-relatedquality of life, illness perception, happiness and negative emotionsin rheumatoid arthritis patients. Act.Colom.Psicol 20(1):49–59.https://doi.org/10.14718/ACP.2017.20.1.4

46. Celiker R, Borman P (2001) Fibromyalgia versus rheumatoid ar-thritis: a comparison of psychological disturbance and life satisfac-tion. J Musculoskelet Pain 9(1):35–45. https://doi.org/10.1136/ard.2009.107912

47. Mok CC, Lok EYC, Cheung EFC (2012) Concurrent psychiatricdisorders are associated with significantly poorer quality of life inpatients with rheumatoid arthritis. Scand J Rheumatol 41(4):253–259. https://doi.org/10.3109/03009742.2012.664648

48. Wan S, He HG, Mak A, Lahiri M, Luo N, Cheung PP et al (2015)Health-related quality of life and its predictors among patients with

Clin Rheumatol

rheumatoid arthritis. Appl Nurs Res 30:176–183. https://doi.org/10.1016/j.apnr.2015.07.004

49. Schünemann H, Brożek J, Guyatt G, Oxman A. (2013) TheGRADE handbook. Cochrane Training. Available from: https://training.cochrane.org/resource/grade-handbook

50. Matcham F, Ali S, Hotopf M, Chalder T (2015) Psychologicalcorrelates of fatigue in rheumatoid arthritis: a systematic review.

Clin Psychol Rev 39:16–29. https://doi.org/10.1016/j.cpr.2015.03.004

Publisher’s note Springer Nature remains neutral with regard to jurisdic-tional claims in published maps and institutional affiliations.

Clin Rheumatol