Atherosclerotic cardiovascular disease in rheumatoid arthritis:

aspects of pathogenesis

and risk

Bengt Wahlin

Department of Public Health and Clinical Medicine

Rheumatology

Umeå 2019

This work is protected by the Swedish Copyright Legislation (Act 1960:729)

Dissertation for PhD

ISBN: 978-91-7855-047-0 ISSN: 0346-6612

New Series No 2029 Tetrao urogallus

Electronic version available at: http://umu.diva-portal.org/ Printed by: UmU Print Service, Umeå University

Umeå, Sweden 2019

A man with a fork in a world of soup

i

Table of Contents

Abstract ......................................................................................... iii

Abbreviations ................................................................................. v

Enkel sammanfattning på svenska ................................................ vii

Background .................................................................................... 1 Rheumatoid arthritis ........................................................................................................ 1

Epidemiology of rheumatoid arthritis ...................................................................... 1 Aetiology of rheumatoid arthritis ............................................................................. 1 Antibodies ...................................................................................................................3 Pathogenesis ...............................................................................................................4 Diagnosis .................................................................................................................... 5 Pharmacological treatment of RA.............................................................................6 Measurement of disease activity and severity .........................................................6

Rheumatoid arthritis and atherosclerosis ....................................................................... 7 Atherosclerosis.................................................................................................................. 7

Cardiovascular risk factors .......................................................................................9 Cardiovascular risk in patients with RA ..........................................................................9

Traditional risk factors in patients with RA........................................................... 11 RA-associated cardiovascular risk factors ............................................................. 12 Impact of risk factors in patients with RA .............................................................. 14

Assessment of atherosclerosis ........................................................................................ 14 Flow-mediated dilation............................................................................................ 14 Pulse wave velocity .................................................................................................. 16 Intima-media thickness ........................................................................................... 16 Assessment of coronary artery calcium by computed tomography ..................... 16 Positron-emission tomography ............................................................................... 16

Management of cardiovascular risk ............................................................................... 17 Estimation of cardiovascular risk ........................................................................... 17

AIMS ............................................................................................ 19

Study populations ......................................................................... 20 Paper I ............................................................................................................................ 20 Paper II........................................................................................................................... 20 Paper III and IV ............................................................................................................. 20

Methods ....................................................................................... 22 Clinical measures of disease activity ............................................................................. 22 Cardiovascular risk factors ............................................................................................ 22 Risk estimations and outcomes..................................................................................... 22 Radiographs ................................................................................................................... 23 Bone mineral density ..................................................................................................... 23 Biomarkers ..................................................................................................................... 23

Markers of bone turnover ....................................................................................... 24 Antibodies .......................................................................................................................25

ii

CMV serologies ...............................................................................................................25 Flow cytometry ...............................................................................................................25 Carotid ultrasound ..........................................................................................................25 Coronary artery CT .........................................................................................................25 Genetic analyses............................................................................................................. 26 Statistical analyses ......................................................................................................... 26

Results ......................................................................................... 28 Paper I ............................................................................................................................ 28 Paper II........................................................................................................................... 30 Papers III and IV ........................................................................................................... 32

Discussion .................................................................................... 36

Conclusions .................................................................................. 41

Future perspectives ...................................................................... 42

Acknowledgements ....................................................................... 43

References .................................................................................... 44

iii

Abstract

Patients with rheumatoid arthritis (RA) have an increased prevalence and

severity of atherosclerosis, and a corresponding increased risk of cardiovascular

disease. The mechanisms causing this are not well elucidated, but both traditional

cardiovascular risk factors and RA-associated factors have been associated with

atherosclerosis and increased risk of cardiovascular events in patients with RA.

Cardiovascular risk estimation based on traditional cardiovascular risk factors,

often underestimates the risk in patients with RA. The aims of this thesis were to

examine factors and biomarkers associated with atherosclerosis in patients with

RA, and to evaluate an algorithm for cardiovascular risk estimation in patients

with RA.

Methods Patients with early RA in the four northernmost counties of Sweden

have since 1995 been included in a prospective study of both the progress of RA

and comorbidities. Besides clinical data, radiographs, genetic markers and

autoantibodies are registered. Paper I includes 665 patients aged 40-80 years

from that cohort, in whom the 10-year risk of a first cardiovascular event was

estimated with both Expanded Cardiovascular Risk Prediction Score in

Rheumatoid Arthritis (ERS-RA), and the general population based ACC/AHA

algorithm. The estimations were then compared to the actual outcomes. Paper II

examines factors associated with coronary artery calcification (CAC) in 22

patients with long-standing RA. Papers III and IV use data from a cohort of

patients <60 years of age at diagnosis of RA (n=79), in whom development of

atherosclerosis has been prospectively followed since diagnosis of RA. This is a

subset of patients from the larger cohort in paper I. Controls matched for age and

sex (n=44) are examined as well. In paper III, phenotypes of T-cells and IgG-

antibodies against cytomegalovirus (CMV) are analysed in relation to carotid

intima-media thickness (IMT). In paper IV, bone mineral density and markers

and regulators of bone metabolism are analysed in relation to IMT.

Results Cardiovascular risk estimation with the RA-specific algorithm ERS-RA

is not superior to estimation with the ACC/AHA algorithm. Both algorithms

underestimate the risk in patients with a high grade of inflammation and in

patients with an estimated moderate risk. In patients with long-standing RA,

presence of CAC is associated with inflammatory activity, both at time of

examination and in earlier stages of RA. Presence of anti-CMV IgG antibodies

and altered T-cells (both CD4+ and CD8+) lacking the co-stimulatory molecule

CD28 (CD28null) are associated with a higher IMT, and patients IgG-positive for

CMV have a rapid increase in IMT after onset of RA. Regulators of bone

metabolism (sclerostin, osteoprotegerin and osteocalcin) are associated with a

higher IMT in patients with RA.

iv

Conclusion Cardiovascular risk estimation in patients with RA still needs to be

improved. The fact that CMV-positivity, altered populations of T-cells and IMT

all are associated, and that also regulators of bone metabolism reflect IMT,

suggests that the pathogenesis of atherosclerosis in patients with RA is

multifactorial. This thesis provides knowledge of the accelerated development of

atherosclerosis in RA and could possibly be relevant also in other chronic

inflammatory diseases, where markers of accelerated atherosclerosis and

increased cardiovascular risk are lacking.

v

Abbreviations

ACC/AHA The American College of Cardiology/American Heart Association risk score

ACPA Antibodies against citrullinated peptides Anti-CCP Antibodies against cyclic citrullinated peptides AUC Area under the curve BMD Bone mineral density BMI Body mass index. BP Blood pressure CAC Coronary artery calcification CDAI Clinical diseae activity CIC Circulating immune complexes CMV Human cytomegalovirus CT Computed tomography CTX C-terminal cross-linked telopeptide CRP C-reactive protein DAS28 Disease activity score of 28 joints DMARD Disease modifying anti-rheumatic drug DXA Dual-energy x-ray absorptiometry ELISA Enzyme linked immune-assay ERS-RA Expanded Cardiovascular Risk Prediction Score for Rheumatoid

Arthritis ESR Erythrocyte sedimentation rate FMD Flow-mediated dilation HAQ Health assessment questionnaire HDL High-density lipoproteins Ig Immunoglobulin IL2sR Soluble receptor of interleukin-2. IMT Intima-media thickness LDL Low-density lipoproteins MDA-LDL Malondialdehyde-modified LDL NETS Neutrophil extracellular traps NO Nitric oxide OCN Osteocalcin OPG Osteoprotegerin OPLS Orthogonal projection of latent structures ox-LDL Oxidized LDL PAI Plasminogen activator inhibitor-1 mass PET Positron-emission tomography P1NP Procollagen type I N-terminal propeptide PTPN22 Protein tyrosine phosphatase, non-receptor type 22 PWV Pulse-wave velocity RA Rheumatoid arthritis RANKL Receptor activator of nuclear factor ĸB RF Rheumatoid factor ROC Receiver operating characteristic SE HLA-DRB1 shared epitope

vi

SCORE Systematic Coronary Risk Evaluation algorithm T0 Baseline examination T1.5 Examination 1.5 years from baseline T11 Examination eleven years from baseline TG Triglycerides TNF Tumor necrosis factor tPA Tissue plasminogen activator antigen VAS Visual analogue scale VIP Variable importance in projection

vii

List of papers

I. Wahlin B, Innala L, Magnusson S, Möller B, Smedby T, Rantapää-

Dahlqvist S, Wållberg-Jonsson S. Performance of the Expanded

Cardiovascular Risk Prediction Score for Rheumatoid Arthritis Is

Not Superior to the ACC/AHA Risk Calculator. J Rheumatol

2019;46(2):130-137

II. Wahlin B*, Meedt T*, Jonsson F, Henein MY, Wållberg-Jonsson S.

Coronary Artery Calcification Is Related to Inflammation in

Rheumatoid Arthritis: A Long-Term Follow-Up Study. BioMed Res

Int. 2016;2016:1261582.

III. Wahlin B, Fasth AER, Karp K, Lejon K, Malmström V, Rahbar A,

Wållberg-Jonsson S, Södergren A. Atherosclerosis in rheumatoid

arthritis: associations between anti-cytomegalovirus IgG antibodies,

CD4+CD28null T-cells, CD8+CD28null T-cells and intima-media

thickness. Submitted

IV. Wahlin B, Ramnemark A, Rantapää-Dahlqvist S, Wållberg-Jonsson

S, Södergren A. Osteoprotegerin and osteocalcin are associated with

atherosclerosis in patients with rheumatoid arthritis: a prospective

cohort study. Manuscript.

*Equal contributors

Paper I is reprinted with permission from The Journal of Rheumatology

Paper II is distributed under the Creative Commons Attribution License.

viii

Enkel sammanfattning på svenska

Reumatoid artrit (ledgångsreumatism; RA) är en kronisk inflammatorisk

ledsjukdom, som drabbar ungefär 0,7 % av Sveriges befolkning. Varför en individ

insjuknar i RA är inte klarlagt, men man har identifierat såväl bakomliggande

ärftliga faktorer som miljöfaktorer, i första hand rökning.

Det dominerande symptomet hos patienter med RA är inflammation i lederna,

vilket leder till värk och stelhet och ofta även permanenta skador i lederna.

Förutom att RA leder till lidande, har patienter med RA en ökad risk för hjärt-

kärlsjukdom, i första hand hjärtinfarkt och stroke, vilket leder till en förkortad

livslängd. Denna riskökning beror på ökad förekomst av ateroskleros, som är den

bakomliggande förklaringen till huvudparten av hjärt-kärlsjukdom. Orsaken till

den ökade förekomsten av ateroskleros hos patienter med RA är inte helt

klarlagd, och inte heller finns något etablerat verktyg för att bedöma risken för

framtida hjärt-kärlsjukdom hos patienter med RA.

Syftet med denna avhandling var att studera faktorer som skulle kunna vara

relaterade till den ökade förekomsten av ateroskleros hos patienter med RA, och

att utvärdera ett riskvärderingsverktyg för hjärt-kärlsjukdom hos patienter med

RA.

I delarbete I jämfördes träffsäkerheten med det för patienter med RA

utvecklade riskberäkningsverktyget ERS-RA med träffsäkerheten för ett

riskberäkningsverktyg för vanlig befolkning och med den faktiska förkomsten av

hjärt-kärlsjukdom. Inget av dessa verktyg var helt träffsäkert, utan båda hade

svagheter i att bedöma risken för framtida hjärt-kärlsjukdom, särskilt när

beräkningen baserades på patientens och den undersökande läkarens uppgifter

om höga blodfetter och högt blodtryck.

Delarbete II är en uppföljning av en tidigare studie av ateroskleros hos patienter

som haft RA i över 30 år, där det visades att det finns ett samband mellan

inflammation under hela sjukdomstiden och graden av ateroskleros, mätt som

förkalkning i hjärtats kranskärl.

I delarbete III och IV studerades RA-patienter som följts med avseende på

aterosklerosutveckling sedan de insjuknade i RA elva år tidigare. I delarbete III

konstaterades att infektion med cytomegalovirus (CMV), ett virus som de flesta

människor är bärare av, var associerat med en snabb utveckling av ateroskleros

efter insjuknande i RA. Patienter som var infekterade med CMV hade också ökad

förekomst av en typ av avvikande T-lymfocyter, CD28null-T-celler, som även de

var associerade med ateroskleros. I delarbete IV studerades markörer för

ix

benomsättning i relation till ateroskleros, eftersom patienter med benskörhet har

ökad risk för hjärt-kärlsjukdom. I denna studie visades att osteoprotegerin, en av

många molekyler som reglerar omsättningen av ben, var associerad till

ateroskleros. Även osteocalcin, ett hormonliknande protein som frisätts från ben,

hade samband med ateroskleros, medan det inte fanns något samband mellan

markörer för själva benomsättningen och ateroskleros.

Sammanfattningsvis visar denna avhandling att det behöver utvecklas nya

verktyg för beräkning av risken för hjärt-kärlsjukdom hos patienter med RA, och

att aktiv mätning av blodfetter och blodtryck är en förutsättning för att kunna

göra en korrekt riskbedömning. Vidare visar jag att det finns ett samband mellan

inflammation och förkalkning i hjärtats kranskärl. Man kan också dra slutsatsen

att CMV-infektion, CD28null-T-celler och ateroskleros är associerade, och att de

benrelaterade molekylerna osteoprotegerin och osteocalcin har ett samband med

ateroskleros, trots att benomsättningen i sig inte har något samband med

ateroskleros hos patienter med RA.

1

Background

Rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic autoimmune disease. At onset, inflam-

mation of the intra-articular synovium is the predominant symptom, causing

pain and stiffness. The inflamed synovium becomes hyperplastic and can erode

the cartilage and bone in the affected joints, causing permanent disability.

Furthermore, malaise and fatigue are common in RA, due to systemic inflam-

mation (Smolen et al. 2018).

Many patients with RA have inflammation not only in the synovium, but also in

other organs (extraarticular RA). The most recognized extra-articular manifest-

tation is rheumatic nodules, usually located in the subcutis, but sometimes in

other organs. More severe manifestations (e.g pleuritis, pericarditis, Felty´s

syndrome and vasculitis) are seen in about 10% of patients with RA (Turesson et

al. 2003, Myasoedova et al. 2011). Patients suffering from RA have a reduced life

expectancy, not only due to extra-articular RA, but mainly due to cardiovascular

disease (Mutru et al. 1985, Wallberg-Jonsson et al. 1997).

Epidemiology of rheumatoid arthritis

The prevalence of RA varies globally between regions and populations. The global

prevalence of RA has been estimated to 0.24%, whereas in Western Europe, the

prevalence is 0.6 % (Cross et al. 2014). This is in line with studies of the Swedish

population, where the prevalence of RA is estimated to 0.5% (Simonsson et al.

1999) or 0.7% (Neovius et al. 2011). Globally, the highest prevalences are seen in

indigenous peoples of North America, where prevalences ranging from 0.6 to

6.8% have been reported (McDougall et al. 2017).

In both the global population and the Swedish population, females have a higher

risk of developing RA, making the prevalence of RA about twice as high in females

as it is in males (Neovius et al. 2011, Cross et al. 2014). The incidence of RA varies

with age, and the highest incidence of RA in Sweden is seen in individuals 70-79

years of age (Eriksson et al. 2013). Mean age at onset of RA was in recent Swedish

studies 55-60 years (Innala et al. 2011, Eriksson et al. 2013).

Aetiology of rheumatoid arthritis

Rheumatoid arthritis is an autoimmune disease, involving both the innate and

the adaptive immune system (Smolen et al. 2018). Despite extensive research, it

is still unknown what initiates the autoimmune response causing RA, but several

risk factors have been identified. The risk factors identified are of several kinds -

2

environmental, genetic, as well as hormonal - and there are also indications of

interaction between them, making the aetiology of RA hard to elucidate and

hitherto insufficiently known.

Genetic risk factors

A hereditary component in the aetiology of RA was identified several decades ago,

when studies showed a higher frequency of RA in first degree relatives to patients

with RA, than in the common population (Lawrence 1970). The heritability of RA

has been estimated to 60% (MacGregor et al. 2000). Today more than 100 loci

have been associated with risk of RA (Okada et al. 2014). The most important

hereditary factor for development of RA, and the first one to be identified, was

the HLA DRB1 alleles called shared epitope (SE) (Gregersen et al. 1987). The

association between RA and SE was later specified to patients with the

rheumatoid factor (RF) antibodies (Padyukov et al. 2004) and antibodies against

citrullinated peptides (ACPA) (van der Woude et al. 2009).

The second most important risk loci in RA is the protein- tyrosine-phosphatase

non-receptor type 22 (PTPN22) (Begovich et al. 2004). As for SE, PTPN22 is

primarily associated with ACPA-positive RA (Kallberg et al. 2007).

Environmental risk factors

Smoking increases the risk of RA (Vessey et al. 1987), in a dose-dependent

manner (Sugiyama et al. 2010). The association between smoking and RA is

predominantly seen for RF-positive RA (Symmons et al. 1997, Sugiyama et al.

2010) and ACPA-positive RA (Klareskog et al. 2006). Inhalation of dust is

another risk factor for RA. This has been shown for exposure to silica (Khuder et

al. 2002), but also for exposure to textile dust (Too et al. 2016). The increased risk

in textile dust exposed individuals is not limited to developing ACPA-positive RA,

but also ACPA-negative RA.

Hormonal risk factors

Rheumatoid arthritis is more common in females, with a female to male ratio of

4:1 in individuals less than 50 years of age and 2:1 in individuals over 60 years of

age (Alpízar-Rodríguez et al. 2017). This suggests that hormonal factors are

involved in the pathogenesis of RA, but the role and mechanisms of hormonal

factors remain controversial. The incidence of RA is lower during pregnancy, but

regarding for example breast feeding, hormonal replacement therapy and oral

contraceptives, the results are conflicting (Alpízar-Rodríguez et al. 2017).

Microbiota

Periodontal disease is associated with an increased incidence of RA (de Pablo et

al. 2008) and also to disease activity in patients with RA (Rodríguez-Lozano et al.

2019). It has been suggested that the increased risk at least in part is mediated by

3

the oral microbiota, in particular Porphyromonas gingivalis and Aggregatibacter

actinomycetemcomitans, bacteria that are able to citrullinate peptides. The gut

microbiota is different in patients with RA compared with the general population,

but the microbiota changes with time of disease and treatment of RA (Guerreiro

et al. 2018).

Interactions between risk factors

There is a clear interaction between genetic and environmental factors on

development of RA. Smoking interacts with shared epitope (SE), as well as

PTPN22, and the combination of all three of these risk factors increase the risk

20-fold for development of ACPA-positive RA (Kallberg et al. 2007). There is also

an association between textile dust and SE, where dust exposed individuals

carrying the SE had a 40-fold increased risk of anti ACPA-positive RA. Regarding

interaction between environmental risk factors, individuals exposed to both

smoking and silica dust have a higher risk of development of RA, than individuals

exposed to either smoking or silica dust (Stolt et al. 2010).

Antibodies

Rheumatoid arthritis is a disease, where extensive studies of antibodies have been

undertaken. Numerous antibodies have been detected, although there are

patients in whom no RA-specific antibodies can be detected. The first antibody to

be discovered was RF (Waaler 1940, Rose and Ragan 1948). Although RF is

present in about 75% of patients with RA, it is not restricted to patients with RA,

but common in many other inflammatory conditions and in healthy individuals,

where the prevalence of RF increases with age (van Boekel et al. 2002). The

presence of RF is associated with a more aggressive disease and a higher risk of

joint erosions (van der Heijde et al. 1992). The terms seropositive and sero-

negative RA are still used to denote RA with or without RF, although the possible

presence of other antibodies makes the terms somewhat inappropriate.

Of other antibodies in patients with RA, ACPAs are highly relevant. Although they

have been known for decades, their clinical importance has arisen in the last 15

years, when studies showed them to be highly specific of RA (Nishimura et al.

2007), prevalent before disease onset (Rantapää-Dahlqvist et al. 2003) and

associated with a more aggressive disease phenotype (Lindqvist et al. 2005,

Berglin et al. 2006). The ACPA test used in most clinical laboratories is the second

generation anti-cyclic citrullinated peptides (anti-CCP).

4

Pathogenesis

The pathogenesis of RA often begins years before onset of clinical symptoms. This

has been shown in studies where antibodies have been detected before disease

onset (Rantapää-Dahlqvist et al. 2003, Brink et al. 2013) and even up-regulation

of pro-inflammatory cytokines have been detected before onset of disease

(Kokkonen et al. 2010). What initiates the disease and what initiates the onset of

symptoms is insufficiently known, although some environmental and hereditary

risk factors are known, as described in previous pages.

At disease onset, inflammation of joints usually is the dominant symptom. This

inflammation is located in the synovium, and involves several mechanisms. As

described by others (McInnes and Schett 2011, Smolen et al. 2018), the synovitis

is caused by infiltration of adaptive immune cells into the synovial lining. A large

part of the inflammatory cells in the inflamed synovium consists of CD4+ T-cells

interacting with B-cells in what has been proposed as an ongoing maturation of

the immune response in the synovium. Cells representing the innate immune

system, in particular macrophages, are also abundant in the synovium in patients

Table 1. The 1987 American College of Rheumatology classification criteria for rheumatoid arthritis (Arnett et al. 1988). 1. Morning stiffness Stiffness in and around joints, lasting at least 1 hour

before maximal improvement. 2. Arthritis of 3 or

more joints At least 3 joint areas have had soft tissue swelling observed by a physician. The 14 possible areas are right or left proximal interphalangeal (PIP) joints, metacarpophalageal (MCP) joints, wrist, elbow, knee, ankle, and metatarsophalangeal (MTP) joints.

3. Arthritis of hand joints

Swelling (arthritis) of the PIP, MCP, or wrist joints.

4. Symmetric arthritis

Simultaneous involvement of the same joint areas (as defined in 2.) on both sides of the body.

5. Rheumatoid nodules

Subcutaneous nodules, over bone prominences, or extensor surfaces, or in juxtaarticular regions observed by a physician.

6. Serum rheumatoid factor

Demonstration of abnormal amounts of serum rheumatoid factor by any method for which the result has been positive in ≤5% of normal control subjects.

7. Radiographic changes

Radiographic changes typical for RA on posterioanterior hand and wrist radiographs, which must include erosions or unequivocal bony decalcification in or most marked adjacent to the involved joints (osteoarthritis changes alone do not qualify).

Critera 1-4 must have been present for at least 6 weeks. Rheumatoid arthritis is defined by the presence of 4 or more criteria.

5

with RA. Besides the infiltration of inflammatory cells, the synoviocytes are active

in the inflammation. The synovial lining becomes hypertrophic, and fibroblast-

like synoviocytes produce cytokines and chemokines, as well as matrix

metalloproteinases. The synoviocytes thereby cause destruction to the joint both

directly, and by promoting the inflammation. Production of receptor activator of

nuclear factor ĸB (RANKL) and pro-inflammatory cytokines activate osteoclasts,

that degrade bone matrix and causes bone erosions.

Diagnosis

Diagnosing RA is a clinical task where joint symptoms, clinical findings,

assessment of acute phase reactants, analysis of RF and anti-CCP and also x-ray

are considered. No diagnostic criteria exist, but classification criteria have been

used since 1956. In 1987, a set of criteria was developed by the American college

of rheumatology (Arnett et al. 1988), that was in use until 2010, when they were

Table 2. The 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for RA (Aletaha et al. 2010). Target population: patients who have at least 1 joint with definite clinical synovitis (swelling) and with the synovitis not better explained by another disease. Joint involvement 1 large joint 0 2-10 large joints 1 1-3 small joints 2 4-10 small joints 3 >10 joints (at least 1 small joint) 5 Serology Negative RF and negative ACPA 0 Low-positive RF or low-positive ACPA 2 High-positive RF or high-positive ACPA 3 Acute-phase reactants Normal CRP and normal ESR 0 Abnormal CRP or abnormal ESR 1 Duration of symptoms <6 weeks 0 ≥6 weeks 1 Joint involvement: tender or swollen joints on examination. Distal inter-phalangeal joints, first carpometacarpal joints, and first metatarsophalangeal joints are excluded from assessment. Large joints: shoulders, elbows, hips, knees, and ankles. Small joints: metacarpophalangeal joints, proximal interphalangeal joints, second through fifth metatarsophalangeal joints, thumb interphalangeal joints, and wrists. Low-positive: above upper limit of normal (ULN), but ≤3 times the ULN. High-positive: >3 times ULN Patients with scores ≥6 are classified as having definite RA.

6

replaced by new criteria from a collaboration between the American college of

rheumatology (ACR) and the European league against rheumatism (EULAR)

(Aletaha et al. 2010). The two sets of criteria are presented in Tables 1 and 2.

Pharmacological treatment of RA

Pharmacological treatment of RA aims to reduce or eliminate the inflammation,

in order to improve physical function and reduce joint erosion, as well as the risk

of extra-articular disease and comorbidities. In order to achieve this, several

drugs are used. Disease modifying anti-rheumatic drugs (DMARD) is a term to

designate drugs that not only relieve the symptoms, but also have an impact on

inflammation. Although corticosteroids have a disease modifying effect, they are

not included in the term DMARD. The cornerstone of DMARDs is methotrexate,

that has been extensively used since the effects on clinical measures of disease

activity (Weinblatt and Maier 1990) and radiological progression (Rau et al. 1991)

were identified. Other commonly used DMARDS are sulfasalazine, hydroxy-

chloroquine, leflunomide, cyclosporine and gold salts. These agents are all small

molecules, developed without exact knowledge of mode of action, and are called

synthetic DMARDs, since they are produced by chemical methods. The word

synthetic is to separate them from the biologic DMARDs, that are large molecules

produced in biological processes. The first biological DMARDs were the tumor

necrosis factor inhibitors, that were introduced in the 1990´s and since then have

been extensively used to treat RA. Other biological DMARDS used to treat RA are

the B-cell depletion antibody rituximab, the interleukin 6 inhibitor tocilizumab,

and the T-cell activation inhibitor abatacept. In the last years targeted synthetic

DMARDs, small moleules with a specific target for mode of action, have been

introduced. Inhibitors of janus kinases are the first agents in this class of

DMARDs. Beside the DMARDs described above, glucocorticoids are anti-

inflammatory and disease modifying in patients with RA, but have side effects

that limit their use. Both EULAR (Smolen et al. 2017) and the Swedish association

of rheumatology (Swedish society of rheumatology 2018) have guidelines for

treatment of RA.

Measurement of disease activity and severity

Measurement of disease activity in RA relies on several different assessments,

that are combined in scores that represent general disease activity. Clinical

examination, where swollen and tender joints are registered, is fundamental in

all scores, but also visual analogue scales (VAS) for patient´s assessment of global

health and pain are used, as well as physician´s assessment of global health.

Analysis of erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP)

are routinely used as laboratory measures of inflammation. Two composite scores

that have widespread use are the Disease Activity Score of 28 joints (DAS28) and

Clinical Disease Activity Index (CDAI). DAS28 is calculated in a somewhat

7

intricate way, using VAS for patient´s global health and pain, swollen and tender

joints, and ESR (Prevoo et al. 199528). CDAI is calculated as the sum of swollen

joints, tender joints, patient´s assessment of global health and physician´s

assessment of global health (Aletaha et al. 2005). In clinical trials where

improvement of disease activity is the primary outcome, the American College of

Rheumatology definition of improvement often is used. This measurement is

based on relative improvement in swollen and tender joints, and improvement in

patient´s and physician´s global assessments, pain, disability, and acute-phase

reactants (Felson et al. 1995).

Another measure of disease activity and severity is the Health Assessment

Questionnaire (HAQ), a patient reported scale that yields a numeric measure of

function in daily life (Fries et al. 1982). It is worth noting, that reduced function

possibly can have different causes in new-onset RA and in long-term RA: in the

new-onset patient, reduced function and elevated HAQ likely reflects ongoing

inflammation, whereas in long-term disease, joint erosions can contribute to

impairment of physical function, thus making HAQ a measure of not only present

inflammation.

Severity of RA can be assessed by conventional x-ray, where joint erosions and

joint space narrowing are detected. Scoring systems often used in research and

clinical trials are the modified Larsen score (Larsen 1995) and Sharp score

modified by van der Heijde (Van Der Heijde et al. 1989).

Rheumatoid arthritis and atherosclerosis

In the 1950s and 1960s, reports of increased mortality in patients with RA were

published (Cobb et al. 1953, Duthie et al. 1964). The most important casue of

increased mortality was later shown to be cardiovascular disease (Mutru et al.

1985), suggesting that patients with RA could have an increased prevalence of

atherosclerosis. The first study confirming this was published in 2001 (Wållberg-

Jonsson et al. 2001), an observation that has been corroborated in several studies

(Ambrosino et al. 2015).

Atherosclerosis

Cardiovascular disease is the leading cause of death in the world (Global Burden

of Disease 2018). It is a heterogenous class of diseases, that includes both

thrombo-embolic diseases and clinical manifestations of atherosclerosis, as well

8

as for example congenital heart disease, cardiomyopathy and diseases of the heart

valves. Of these different types of cardiovascular disease, atherosclerotic

cardiovascular diseases are by far the most common and cause 85% of deaths

from cardiovascular disease (Global Burden of Disease 2018).

Although the onset of symptoms of disease often is sudden, the underlying

atherosclerosis is a slowly progressive disease. The penomenon where plaques

are developed in the arteries is referred to as atherogenesis (Figure 1). As

reviewed by others (Hansson and Hermansson 2011, Libby et al. 2011), it starts

with endothelial activation, where the endothelial cells lining the inner arterial

surface interact with inflammatory cells via expression of adhesion molecules, as

a response to irritation from for example hypertension, dyslipidaemia or

inflammation. Parallel to the interaction with inflammatory cells, the endothelial

permeability is increased, promoting entry of low-density lipoproteins (LDL) into

the arterial wall. When LDL is modified to oxidized LDL, it is subject of endo-

cytosis by macrophages derived from monocytes, that have migrated through the

endothelium. As the intracellular cholesterol increases, the macrophages become

cholesterol filled foam cells. In this stage of the atherogenesis, fatty streaks in the

vessel wall can be detected. The fatty streaks can disappear, or progress into more

Figure 1. Progression of atherosclerosis.

Source: https://commons.wikimedia.org/wiki/File:Endo_dysfunction_Athero.PNG

9

advanced atheromas and subsequent plaques. Formation of plaques involves

synthesis of extracellular matrix and migration of smooth muscle cells from the

medial layer of the arterial wall, that forms a fibrous cap surrounding a lipid core

containing cholesterol crystals and often necrotic macrophages. In addition to

monocytes, T-cells migrate through the intima into the plaque, where they are

supposed to have immunoregulatory functions, and also B-cells, neutrophils,

dendritic cells and mast cells enter the plaque. Inflammation in the plaque makes

the fibrous cap weaker and more prone to rupture, exposing the pro-coagulant

material from the core to the coagulation system and thrombocytes, thus causing

a thrombus that obstructs the blood flow. A plaque with a thick fibrous cap,

containing extracellular matrixproteins and calcium, is less prone to rupture

(“stable”), whereas a plaque with a thin fibrous cap and a high grade of

inflammation is more prone to rupture (“unstable”) (Muller et al. 1989).

Cardiovascular risk factors

Several risk factors for development of cardiovacscular disease have been

identified. Age, male sex and a family history of premature cardiovascular disease

are fixed factors that are not modifyable for an individual, but in the large,

international INTERHEART study, the modifyable risk factors abnormal lipids,

smoking, hypertension, diabetes, abdominal obesity, psychosocial factors,

consumption of fruits, vegetables, and alcohol, and regular physical activity

accounted for over 90% of the risk of myocardial infarction (Yusuf et al. 2004).

A role of infections in the process of atherogenesis has been hypothesized.

Antibodies against human cytomegalovirus (CMV) have been associated with

atherosclerosis and cardiovascular disease in several studies (Sorlie et al. 2000,

Dahal et al. 2017). Another infectious agent suggestedly associated with athero-

sclerosis and cardiovascular disease is Chlamydia pneumoniae (Liu et al. 2005,

Dahal et al. 2017).

Cardiovascular risk in patients with RA

As stated above, patients with RA have an increased risk of cardiovascular

disease. Compared with the general population, the morbidity in cardiovascular

disease is increased by 50% and the mortality by 60% (Meune et al. 2009, Avina-

Zubieta et al. 2012). The exact mechanisms have been hard to identify, but both

traditional cariovascular risk factors and inflammation contribute to the

increased risk (Wållberg-Jonsson et al. 1999, Innala et al. 2011, Nurmohamed et

al. 2015). In a recent multi-national study using several cohorts of patients with

RA, 30% of the cardiovascular events were attributable to RA-characteristics,

whereas 49% were attributable to traditional risk factors (Crowson et al. 2018).

10

Fig

ure

2. M

eta-a

na

lysis o

f stud

ies on

card

iov

ascu

lar even

ts in p

atien

ts with

RA

. CV

D: ca

rdio

va

scular d

isease. A

MI: a

cute

my

oca

rdia

l infa

rction

. CV

A: cereb

rov

ascu

lar a

cciden

t. CH

F: co

ng

estive h

eart fa

ilure. F

rom

Av

ina

-Zu

bieta

, J.A

.et al. A

nn

als o

f

the R

heu

ma

tic Disea

ses 20

12. 7

1: 152

4–

152

9

11

When RA-characteristics and traditional risk factors were combined, 70 % of the

cardiovascular events were explained, leaving 30% of the events unaccounted for.

Figure 2 presents results from a meta-analysis of cardiovascular risk in patients

with RA.

Traditional risk factors in patients with RA

The risk of cardiovascular disease is in both patients with RA and the general

population to a great extent explained by the traditional risk factors. In the above

mentioned study of patients with RA, smoking and hypertension were the

traditional risk factors that had the largest impact on the cardiovascular risk

(Crowson et al. 2018). Some of the traditional risk factors are more prevalent in

patients with RA, compared with the general population.

Smoking is a shared risk factor for both cardiovascular disease and RA, and

smoking is more prevalent prior to onset of RA (Kokkonen et al. 2017) and

seropositive inflammatory arthritis (Goodson et al. 2004).

Dyslipidaemia is more common in patients with RA and ongoing inflammation.

Levels of high-density lipoproteins, low-density lipoproteins and total cholesterol

are reduced in patients with active RA (Choy and Sattar 2009), but to a great

extent normalized when the disease activity is lowered (Steiner and Urowitz

2009). There are also qualitative changes in the lipoproteins in patients with RA

and inflammatory activity (Rantapää-Dahlqvist et al. 1991).

Glucose tolerance is impaired in patients with RA (Svenson et al. 1988, Hoes et

al. 2011) and the prevalence of diabetes has been reported to be higher (Solomon

et al. 2010).

Hypertension has been studied in patients with RA, with conflicting results

regarding the prevalence, as reviewed by Panoulas (Panoulas et al. 2008).

However, hypertension has been reported to be underdiagnosed and

undertreated in patients with RA (Stoep et al. 2016).

Obesity is a field of conflicting results in patients with RA, but an altered body

composition with sarcopenia and increased trunc fat is reported (Book et al.

2009, Dessein et al. 2016, Turk et al. 2018). This condition is often referred to as

rheumatoid cachexia.

12

RA-associated cardiovascular risk factors

Patients with RA have cardiovascular risk factors that are associated with the

disease. Some of the RA-associated risk factors are mentioned below.

Inflammation is associated with an increased risk of cardiovascular disease in

patients with RA (Wållberg-Jonsson et al. 1999, Innala et al. 2011). Plausible

mechanisms of this finding are inflammation-induced endothelial activation

(Libby et al. 2011) and incresed instability of plaques due to inflammation (Aubry

et al. 2007, Karpouzas et al. 2014).

Genetic susceptibility of RA and cardiovascular disease is common for some

genetic traits. Studies have mainly focused on the HLA DRB1 shared epitope, that

is a risk factor for anti-CCP-positive RA, but also a risk factor for myocardial

infarction in both patients with RA (López-Mejías et al. 2016) and the general

population (Björkbacka et al. 2010, Paakkanen et al. 2012). On the other hand,

PTPN22 is not associated with cardiovascular events nor atherosclerosis in

patients with RA (Palomino-Morales et al. 2010). A population-based Swedish

study identified an increased risk of acute coronary syndrome in siblings of

patients with RA (Westerlind et al. 2019).

Antibodies against citrullinated peptides have been suggested to be associated

with cardiovascular death in patients with RA (Ajeganova et al. 2016).

Table 3. Previous studies of T-cell markers and atherosclerosis in patients with RA. Reference T-cell markers Measures of

atherosclerosis Gerli et al. 2004 CD4

CD28 IMT

Pingiotti et al. 2007 CD4 CD28 CX3CR1

IMT, FMD

Winchester et al. 2016 CD4 CD8 CD28 CD56 CD57 HLA-DR

CAC

IMT: intima-media thickness. FMD: Flow-mediated dilation. CAC: coronary

artery calcification

13

Neutrophil extracellular traps (NETS) are chromatin and proteins released by

neutrophils to control infections (Brinkmann and Zychlinsky 2012). NETS are

dysregulated in patients with rheumatoid arthritis (Apel et al. 2018), but also in

atherogenesis (Qi et al. 2017).

CD28null T-cells are expanded in patients with RA (Schmidt et al. 1996).

Increased numbers of CD28null T-cells are associated with atherosclerosis and

an increased risk of cardiovascular disease in the general population (Dumitriu

et al. 2009). No studies have examined the relation between CD28null T-cells and

cardiovascular disease in patients with RA, but CD28null T-cells are associated

with atherosclerosis in patients with RA (Gerli et al. 2004). Table 3 summarizes

previous studies of T-cells and atherosclerosis in patients with RA.

Osteoporosis is associated with increased risk of cardiovascular disease and

atherosclerosis (den Uyl et al. 2011). Osteoporosis is more common in patients

with RA compared with the general population (Haugeberg et al. 2000). Patients

with RA and a fragility fracture have almost a doubled risk of a cardiovascular

event, compared with patients without fractures (Ni Mhuircheartaigh et al. 2017).

Table 4 summarizes previous studies of atherosclerosis and aspects of bone

turnover in patients with RA.

Table 4. Previous studies of bone markers and atherosclerosis in patients with RA. Reference Bone marker Measures of

atherosclerosis Tanaka et al. 2006 Bone mineral

density Pulse-wave velocity

Asanuma et al. 2007 Osteoprotegerin CAC Asanuma et al. 2013 Osteoprotegerin Carotid plaque Dessein et al. 2014 Osteoprotegerin Intima-media thickness

Markers of endothelial activation

Beyazal et al, 2015 Osteoprotegerin Pulse-wave velocity Intima-media thickness

Provan et al. 2017 Bone mineral density Osteoporosis

Cardiovascular death

Ni Mhuircheartaigh et al. 2017

Fragility fracture Cardiovascular event

CAC: coronary artery calcification

14

Impact of risk factors in patients with RA

The impact of cardiovascular risk factors in patients with RA is somewhat

different from what we know from the healthy population. According to a large

study of heterogenous international cohorts (Crowson et al. 2018) and a meta-

analysis (Baghdadi et al. 2015), hypertension, smoking, diabetes and hyper-

cholesterolaemia are associated with an increased cardiovascular risk, whereas

BMI and physical inactivity are not. When RA-associated factors are assessed, the

impact of high inflammatory activity and positivity for RF or anti-CCP were

comparable to the impact of smoking and hypertension, and higher than the

impact of blood lipids and diabetes, as presented in figure 3 (Crowson et al. 2018).

There are also several studies of patients with RA, where treatment with tumor

necrosis factor (TNF) inhibitors and methotrexate have been associated with a

lower cardiovascular risk (Choi et al. 2002, Barbhaiya and Solomon 2013), but

whether this is due to a cardioprotective effect of the drugs, reduction of

inflammation or selection of patients to treatment is not sufficiently known.

However, the risk of myocardial infarction in patients with RA is associated with

increased ESR and CRP (Meissner et al. 2016), and a Swedish study showed that

patients with a good response to treatment with TNF inhibitors have a

cardiovascular risk comparable to the risk in the general population (Ljung et al.

2016). On the other hand, endothelial function is improved independent of

inflammatory response, when treatment with methotrexate and TNF inhibitors

is initiated in patients with RA (Deyab et al. 2017), indicating that the effect

possibly is explained not only by reduced inflammation.

Assessment of atherosclerosis

Atherosclerosis is in most cases subclinical and causes no symptoms, until the

affected individual suffers from an atherosclerotic cardiovascular event.

Therefore, methods have been developed to study subclinical atherosclerosis.

Flow-mediated dilation

Flow-mediated dilation (FMD) uses ultrasound to measure the diameter of the

brachial artery before and after occlusion (Celermajer et al. 1992). The

endothelium produces vasodilating nitric oxide (NO) when the artery is occluded,

but dysfunctional endothelium, an early phenomenon in atherogenesis, has

reduced capacity of NO production, thus making FMD reduced. FMD is a method

that requires a standardized procedure, but performed adherent to guidelines,

15

Figure 3. Population attributable risk for cardiovascular risk factors and RA

characteristics overall (panel A), among women (panel B) and among men (panel

C). ACPA: antibodies against citrullinated peptides. CRP: C-reactive protein. ESR:

erythrocyte sedimentaion rate. RF: rheumatoid factor. From Crowson, C.S. et al.

Annals of the Rheumatic Diseases 2018. 77: 48–54.

16

FMD is reproducible and reliable (Greyling et al. 2016). Decreased FMD is a

predictor of future atherosclerotic cardiovascular disease (Perticone et al. 2001,

Matsuzawa et al. 2015).

Pulse wave velocity

Arterial stiffness is a sign of atherosclerosis, that can be measured as increased

pulse wave velocity (PWV). The method is based on the fact that the velocity of

the pulse wave increases with reduced elasticity in the arterial wall. Several

methods for assessing PWV and arterial stiffness have been developed (Laurent

et al. 2006). Increased PWV is a predictor of future atherosclerotic cardiovascular

disease (Blacher et al. 1999, Laurent et al. 2006, Simon et al. 2007).

Intima-media thickness

Once the atherogenesis is started, the thickness of the innermost part of the

arterial wall (intima and media layers) is increased. This can be measured using

B-mode ultrasound, where the intima-media thickness (IMT) is measured,

usually in the far wall of the common carotid artery proximal to the bulb (Pignoli

et al. 1986). Increased IMT is a predictor of future atherosclerotic cardiovascular

disease (Chambless et al. 1997, Simon et al. 2007).

Assessment of coronary artery calcium by computed tomography

X-rays are attenuated as they pass through calcified matter. This holds true also

for calcified plaques, that can be detected using x-ray techniques. The most

commonly used method for assessment of calcified plaques is computed

tomography (CT) of the coronary arteries and subsequent quantification of

coronary artery calcium (CAC) according to Agatson (Agatston et al. 1990). This

method detects both fully calcified plaques, that are considered stable, and

plaques with heterogenous calcification, that are more prone to rupture.

Nevertheless, a higher CAC is a predictor of future atherosclerotic cardiovascular

disease (Detrano et al. 1996, Simon et al. 2007).

Positron-emission tomography

Positron-emission tomography (PET) is a nuclear medicine functional imaging

method that uses radioactive tracers to observe metabolic processes. When PET

is combined with computed tomography (PET-CT), the anatomical structure

corresponding with the tracer uptake can be located. 18F-fluorodeoxyglucose (18F-

FDG), an analogue to glucose, is a tracer that is concentrated in tissues with

increased metabolism, e.g. sites of inflammation, and has been used to identify

inflammation in plaques, whereas 18F-sodium fluoride (18F-NaF) is a marker of

bone mineralization and vascular calcification (McKenney-Drake et al. 2018).

Long-term studies of the prognostic value of PET-CT are scarce, but arterial

17

uptake of 18F-FDG in aorta has been associated with future cardiovascular events

in one study (Iwatsuka et al. 2018). Traditional cardiovascular risk factors are

stronger associated with arterial wall uptake of 18F-NaF, than with uptake of 18F-

FDG (McKenney-Drake et al. 2018).

Management of cardiovascular risk

It has been estimated that 90% of cardiovascular disease in the general

population could be prevented (McGill Henry C. et al. 2008). Since athero-

sclerotic cardiovascular disease is the leading cause of increased mortality in

patients with RA, prevention of disease is an urgent task in this population. Very

recently an interventional study was published, where intense treatment of

traditional cardiovascular risk decreased the risk of atherosclerotic cardio-

vascular events and reduced the rate of progress of IMT in patients with RA

(Burggraaf et al. 2019). This implies that the cardiovascular risk in patients with

RA actually can be reduced by successful treatment of both RA and traditional

cardiovascular risk factors, but in order to maximize the impact of primary

prevention, patients with a high risk must be identified.

Estimation of cardiovascular risk

Several tools have been developed to calculate the risk of a cardiovascular event

or cardiovascular death for an individual, based on the traditional risk factors.

When used in patients with RA, these risk estimation tools do not perform well,

mostly due to underestimation of the risk (Crowson et al. 2012, 2017a, Arts et al.

2015a). A modification of the European Systematic Coronary Risk Evaluation

algorithm (SCORE) that included RA-characteristics was developed, but did not

perform much better than the original algorithm (Arts et al. 2015b). Neither did

the RA-specific calculator created by the Trans-Atlantic Cardiovascular Risk

Consortium for Rheumatoid Arthritis perform better than calculators for the

general population (Crowson et al. 2017b). Another RA-specific risk calculator,

the Expanded Risk Score for cardiovascular disease in RA (ERS-RA) was

developed using the North American patient cohort CORRONA (Solomon et al.

2015), but tested in a multi-national cohort of patients with RA, it did not

outperform risk calculators for the general population (Crowson et al. 2017a),

although ERS-RA performed well when externally validated in Swedish patients

with RA (Ljung et al. 2018). Table 5 presents previous studies of cardiovascular

risk scores in patients with RA. The EULAR recommendations for management

of cardiovascular risk, state that the estimated risk from general population risk

scores shall be multiplied by 1.5 for patients with RA (Agca et al. 2017).

18

Table 5. Previous studies of cardiovascular risk estimation scores in patients with RA. Reference Risk scores studied

Crowson et al, 2012 Framingham Reynolds

Arts et al, 2015a SCORE Framingham Reynolds QRiskII

Arts et al, 2015b SCORE Adapted SCORE

Crowson et al, 2017a ERS-RA Framingham Reynolds QRiskII ACC/AHA

Crowson et al, 2017b SCORE Framingham QRiskII ACC/AHA New developed score

Ljung et al, 2018 ERS-RA

ACC/AHA: American College of Cardiology/American Heart Association risk score. ERS-RA: Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis. SCORE: Systematic Coronary Risk Evaluation algorithm

19

AIMS

Although there is overwhelming evidence that patients with RA have an increased

prevalence of atherosclerosis and atherosclerotic cardiovascular disease, the

underlying pathogenic mechanisms are not sufficiently known. Neither is there

enough knowledge of how the risk of cardiovascular disease in patients with RA

should be estimated. The aims of this thesis were to address these tasks, with an

emphasis on the following:

- To validate the performance of a cardiovascular risk score specific for

patients with RA.

- To investigate factors associated with coronary artery calcification in

patients with long-term RA.

- To investigate the impact of cytomegalovirus and subsets of T-cells on

development of atherosclerosis after onset of RA.

- To investigate the associations between atherosclerosis, bone mineral

density, and regulators and markers of bone turnover after onset of RA.

20

Study populations

The studies included in this thesis are undertaken in cohorts of patients with RA

in Northern Sweden. Papers III and IV also include matched controls. Figure 4

presents the cohorts included in papers I, III, and IV.

Paper I

Paper I uses data from a cohort of patients with early RA in the four northernmost

counties of Sweden (Jämtland, Västernorrland, Västerbotten and Norrbotten),

that was initiated in 1995. All patients fulfilling the 1987 criteria for RA (Arnett et

al. 1988) and with symptoms of RA for less than one year, are invited to the

ongoing study. Paper I uses data from patients included until October 2009. The

number of patients was initially 950, but was reduced to 810 after exclusion of

patients with a previouscardiovascular disease, or without any data on inflam-

matory activity at inclusion. ERS-RA estimates 10-year risk of a cardiovascular

event for patients 20-80 years of age. However, the American College of

Cardiology/American Heart Association risk score (ACC/AHA), that was used as

comparator in the study, does not calculate 10-year risk for patients less than 40

years of age. Those patients were consequently excluded, leaving 665 patients

(458 female, 207 male) 40-80 years old for risk estimation with ERS-RA. The

upper limit of age for risk estimation with ACC/AHA is 79 years, making the

number of patients 662 (455 female, 207 male) in risk estimation with ACC/AHA.

Paper II

Paper II is in part a follow-up of an earlier study (Wållberg-Jonsson et al. 2001),

where all eligible patients diagnosed with RA at the Department of Rheumatology

at Umeå university hospital 1974-1978 and less than 66 years old at the time of

examination, were invited. All patients fulfilled the ACR 1987 criteria for RA

(Arnett et al. 1988). That previous study included 39 patients, but at the time of

follow-up, seven patients were deceased and one emigrated. Since seven patients

declined inclusion in the follow-up, most of them due to severe disability, and two

patients did not complete the CT examination, 22 patients (18 female, 4 male)

were included.

Paper III and IV

Papers III and IV report data from an ongoing longitudinal study of athero-

sclerosis in patients with RA. The cohort is a subpopulation of the cohort in paper

I. In the years 2000-2004, all patients with early RA, not more than 60 years old

and resident in Jämtland, Västerbotten or Norrbotten were consecutively invited

to participate in the study. Of the 87 patients invited, six declined participation

21

and two were excluded (one due to pregnancy, one due to advanced malignancy),

leaving 79 patients that were included in the baseline examinations (T0). Follow-

up was undertaken after eleven years (n=62), hereafter denominated T11. Of the

62 patients examined at T11, eight patients did not have data on IMT due to a

hardware crash and four were not assessed by flow-cytometry analysis, making

the number of patients 50 in paper III and 54 in paper IV. At baseline, 44 controls

matched for age and sex were included, of whom 31 were re-examined at T11. A

subset of 26 unselected patients and their 27 matched controls were subjects to

an additional examination 18 months after inclusion (T1.5).

Figure 4. Illustration of the cohorts in papers I, III and IV. All patients were intially

included in the regional cohort of patients with new-onset RA.

22

Methods

The studies included in this thesis include data from several different measures

of clinical, laboratory and radiological examinations, that are described in the

following sections.

Measures of disease activity

In papers I, III and IV, data on disease activity was read from the Swedish register

of early RA and completed by thorough reading of patients´ records, whereas in

paper II the data on disease activity was registered at the study examination. In

all papers, the data on disease activity includes number of swollen and tender

joints, VAS scales for pain and global assessment, HAQ, DAS28, and the

laboratory measures ESR and CRP. Calculation of CDAI at baseline was

undertaken in paper I. In paper II, haptoglobin, soluble receptor of interleukin 2

(IL2sR), and interleukin 6 were analysed as measures of inflammation, and

accumulated disease activity (Baecklund et al. 1998) was calculated at the

baseline examination 1997-1998.

Cardiovascular risk factors

In order to assess the presence of cardiovascular risk factors at the time of

diagnosis of RA, all patients´ records until the time of diagnosis of RA were

thoroughly read and cardiovascular risk factors registered according to a protocol

for every patient in paper I, III, and IV. In papers III and IV, patients´ records

were also read during the study. A survey including questions on cardiovascular

risk factors was sent to the patients in paper I when they were included in the

study. The patients in papers II, III and IV filled in surveys on cardiovascular risk

factors at the study examinations.

In the patients in papers I, II and IV, cholesterol, high-density lipoproteins

(HDL), low-density lipoproteins (LDL) and triglycerides were analysed at the

time of diagnosis of RA or soon thereafter, completed by repeated analysis at the

subsequent study examinations for the patients in papers III and IV. Analyses

were done by routine methods at the local laboratories of clinical chemistry at the

present hospital, where the patient was examined. In paper III, fasting blood

lipids were analysed at the study examination.

Risk estimations and outcomes

The risk of a cardiovascular event for the patients in paper I was estimated by

both the ERS-RA and the ACC/AHA algorithm. Risk estimations using both risk

scores were for every individual patient performed according to the algorithms,

23

based on collected data, completed by multiple imputation for missing values.

The risk was calculated from ERS-RA in two settings: the first estimation used

patients´ and physicians´ reported data on hypertension and hyperlipidaemia,

whereas the second estimation used measured values of blood pressure and blood

lipids. Hypertension was defined as blood pressure >140/90 and hyperlipidaemia

defined according to the Third Report of the National Cholesterol Education

Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High

Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report (Wilkins

2002). The estimations from ACC/AHA was analysed both as crude risk

estimation, and multiplied by 1.5 according to the EULAR recommendations

(Agca et al. 2017).

Data on the cardiovascular outcomes for the patients in paper I was retrieved

from the Cause-of-Death-register and National Inpatient register at the National

Board of Health and Welfare (Socialstyrelsen 2017). Events until December 31

2013 were included.

Radiographs

In patients included in paper IV, conventional radiographs of hands, wrists and

feet from baseline were analysed by trained rheumatologists and evaluated

according to Larsen score (Larsen 1995).

Bone mineral density

Bone mineral density (BMD) was measured in patients and controls resident in

Västerbotten, using dual-energy x-ray absorptiometry (DXA) (Lunar DPX-L,

software version 1:3, Lunar, Madison, WI, USA). This technique is based on an

X-ray generator using two energies (“dual energy X-ray”; DXA). Using DXA, bone

can be separated from surrounding tissues, due to the different attenuations of

the x-rays (Pietrobelli et al. 1996). BMD for the spine was derived from the whole

body composition scan using region of interest for the spine.

Biomarkers

In paper II, several previously analysed biomarkers were included in the analysis.

These biomarkers included presence of antibodies with affinity for oxidized LDL

of immunoglobulin class A (IgA) (ox-LDL IgA), immunoglobulin class G (IgG)

(ox-LDL IgG) and immunoglobulin class M (IgM) (ox-LDL IgM), as well as

antibodies directed against malondialdehyde-modified LDL of classes IgA (MDA-

LDL IgA), IgG (MDA-LDL IgG) and IgM (MDA-LDL IgM). Endothelial activation

and adhesion molecules were analysed, including soluble intercellular adhesion

molecule-1 (ICAM-1), E-selectin, and circulating immune complexes (CIC).

24

Hemostatic factors analysed were von Willebrand factor, plasminogen activator

inhibitor-1 mass (PAI), tissue plasminogen activator antigen (tPA), D-dimer,

fibrinogen and anti-cardiolipin antibodies of class IgA, IgG and IgM. Leptin,

lipoprotein(a) and homocystein levels were also analysed. The procedures of

analyses have previously been described in detail (Wållberg-Jonsson et al. 2001,

2002).

Markers of bone turnover

Markers of bone turnover were analysed in frozen samples collected at baseline

and T11, presented in table 6. The samples were drawn at the study examinations

at baseline and T11, and then stored at -80° Celsius.

Serum concentrations of dickkopf-1, osteopontin, osteoprotegerin (OPG),

sclerostin, osteocalcin (OCN), and parathyroid hormone (PTH) were determined

using a multiplex assay (HBNMAG-51K-07, Millipore Corporation, Billerica, MA)

according to manufacturer´s protocol. Serum concentration of receptor activator

of nuclear factor κB ligand (RANKL) was analysed using an enzyme linked

immune-assay (ELISA; Human RANKL ELISA (BioVendor, Karasek, Czech

Republic)) according to the manufacturer´s protocol. All analyses were done in

duplicates. Procollagen type I N-terminal propeptide (P1NP), C-terminal cross-

linked telopeptide (CTX), thyroid stimulating hormone (TSH), thyroxin, 25-OH

vitamin D, phosphate, and calcium were analysed at the laboratory of clinical

Table 6. Bone markers studied in paper IV. Bone marker Baseline T11 Dickkopf-1 X X Osteopontin X X Osteoprotegerin X X Sclerostin X X Osteocalcin X X Parathyroid hormone X X Fibroblast growth factor 23 X X RANKL X X P1NP X CTX X TSH X Thyroxin X 25-OH vitamin D X Phosphate X Calcium X RANKL: receptor activator of nuclear factor κB ligand. P1NP: procollagen type 1 N-terminal propeptide. CTX: C-terminal cross-linked telopeptide. TSH: thyroid stimulating hormone.

25

chemistry at Karolinska University Hospital, using routine methods. Table 6

presents the bone markers.

Antibodies

Analysis of RF was performed by Waaler-Rose hemagglutination test at the local

hospitals using local routine methods. ACPA was determined using an ELISA for

anti-CCP2 antibodies (Euro-Diagnostica, Malmö, Sweden) at the local hospitals.

CMV serologies

In paper III, analysis of anti-CMV IgG-antibodies was done from frozen plasma

samples collected at T0 and T11, stored at -80°C. An in-house ELISA was used at

Karolinska institutet (Sundqvist and Wahren 1982, Rahbar et al. 2004). Paired

samples from patients and matched controls were assessed on the same plate.

Flow cytometry

In paper III, flow-cytometry analysis of T-cells subsets was undertaken at T11,

using a LSRII flow cytometer (Becton-Dickinson, San Jose, CA). The cell surface

markers analysed were CD45 (common leukocyte antigen), CD3 (T-cell marker),

CD4, CD8, CD28, CD56, and CX3CR1. Obtained data was analysed using the

BDDiva software (BD Biosciences, San Jose, CA, USA).

Carotid ultrasound

Ultrasound examinations of the right common carotid artery were performed

with the patient in supine position and the head turned to the left, using a Sequoia

512 and an 8L5 linear transducer (both from Siemens, Acuson, Upplands Väsby,

Sweden) at 8MHz. Three longitudinal measurements of IMT in the far wall in an

approximately 10 mm long segment proximal to the bulb was undertaken. The

average of the three measurements was calculated and used in further analysis.

Coronary artery CT

In paper II, measurement of CAC was performed by multidetector computed

tomography (LightSpeed VCT, GE Healthcare, Milwaukee, WI, US). The images

were analysed on a workstation (Advantage 4.4, GE Medical systems, Milwaukee,

WI, USA) and CAC was detected by a dedicated software (SmartScore version 4.0,

GE Medical systems, Milwaukee, WI, USA). The CAC score was quantified

according to the Agatston method (Agatston et al. 1990). Patients with CAC 0-10

were classified as having low CAC, whereas patients with CAC >10 were classified

as having high CAC. This classification was based on a large study (Budoff et al.

2007), where CAC >10 was a strong independent predictor of mortality, but

26

individuals with CAC 1-10 had only a small increase in mortality compared to

individuals without CAC.

Genetic analyses

A polymerase chain reaction kit with sequence specific primers (Dynal, Oslo,

Norway) was performed for genotyping of the HLA-DRB1 shared epitope alleles

*0101, *0401, *0404, *0405, and *0408. The PTPN22 1858T/C polymorphism

was identified using a 5´nuclease assay. Detection of the genotypes was made

using ABI PRISM 7900HT Sequence Detector System and the data was processed

using the SDS 2.1 software (both Applied Biosystems, Foster City, CA, USA).

Statistical analyses

For comparison of mean or median values between groups, Mann-Whitney U test

or student´s t-test was used, depending on distribution. Kruskal-Wallis test was

used when more than two groups were compared, Bonferroni corrected in

pairwise comparisons. Linear and logistic regression models were used for

analysis of associations between variables. Variables with skewed distribution

were log-transformed before entering linear regression models. Linear regression

models initially included severeal independent variables chosen on clinical

assumptions and results from univariable regression, but variables that lowered

R2 were subsequently excluded. Goodness of fit of multivariable logistic

regression models was tested with Hosmer-Lemeshow test. Collinearity of

independent variables was tested in multiple linear regression models, and a

variable inflation factor of more than four was regarded intolerable. P-values

<0.05 were regarded statistically significant.

In paper I, data from onset of RA was used for calculation of every patient´s

individual 10-year risk for a cardiovascular event. When the duration of follow-

up was less than 10 years, the risk was adjusted proportionally to the time of

follow-up. Multiple imputation with five repetitions was performed when there

was a missing value of a risk factor. The accuracy of estimations was assessed in

terms of calibration (comparing observed and estimated events) and

discrimination (the ability to correctly rank individuals from low to high risk).

Calibration was assessed using deciles of estimated risk, comparing observed

events and estimated risk. Discrimination was assessed by calculation of the area

under the curve (AUC) of receiver operating characteristic (ROC) curves.

In paper II, multivariate discriminant analysis was performed using the

orthogonal projection to latent structures discriminant analysis (OPLS)

algorithm (Trygg and Wold 2002). The concept of OPLS is transition from a large

number of descriptive variables, to a small number of vectors, representing latent

27

variables. The first latent variable is the latent variable that best explains the

variation in the response variable, whereas the subsequent latent variables are

orthogonal to vector 1, meaning that they represent information that varies in a

non-random pattern, but is independent of the response variable. The models

were modified by repeated exclusion of variables without importance in the

model (variable importance in projection (VIP) <0.5). ROC-curves were made for

each final model, to evaluate sensitivity and specificity in predicting high or low

CAC.

Statistical analyses were performed using SPSS for Windows version 21-25 (IBM

SPSS INC, Chicago, Illinois, USA) or, for OPLS, Simca version 13.0 (Umetrics,

Umeå, Sweden).

28

Results

Paper I

In paper I, the risk estimations from the RA-specific risk score ERS-RA in 665

patients (458 female, 207 male) patients with new-onset RA were validated. The

estimations were compared with the actual outcomes, and with the risk

estimations from the general population ACC/AHA risk score. The risk score

ERS-RA was used in two settings: one based on patients´ and physicians´ reports

on hyperlipidaemia (“ERS-RA reported”), and one based on measeured values of

blood pressure and blood lipids (“ERS-RA measured”). The ACC/AHA risk score

was analysed in two settings as well: crude data (ACC/AHA), and multiplied by

1.5 as recommended by EULAR (ACC/AHAx1.5) (Agca et al. 2017).

The estimated median 10-year risk was 6.3% for ERS-RA (reported), 8.3% for

ERS-RA (measured), 6.3% for ACC/AHA and 9.2% for ACC/AHAx1.5. During the

follow-up, 73 patients (11.0%) developed a cardiovascular event (9 cases of death

due to cardiovascular disease, 25 patients suffered a myocardial infarction, 39

patients suffered a stroke). Of the female patients, 38 (8.3%) had an event,

compared with 35 (16.9%) of the male patients. One event occurred in a patient

80 years old at inclusion, making the number of observed events 72 when

ACC/AHA data was analysed. Figure 5 illustrates the percentage of observed and

estimated events in deciles of estimated risk. When the deciles of observed and

estimated events for each of the four risk estimation scores were compared with

the Hosmer-Lemeshow test, ERS-RA (reported) had the lowest agreement and

the two models of ACC/AHA the highest agreement between estimated and

predicted events. Discrimination (ranking the patients according to risk) was

good for all four risk estimation scores with AUC of 0.76-0.77.

All risk scores were less accurate in terms of discrimination in patients with 5-

15% estimated risk. ERS-RA (reported) underestimated remarkably, with lower

AUC than the other risk scores (0.52 vs 0.59-0.60) and also the highest grade of

underestimation in that subset of patients.

In high-inflammatory patients (defined as ESR ≥ 40), the calibration of all risk

scores was poor. The observed proportion of events in this subset of patients was

12.1%, whereas the estimated proportion was 7.2%, 9.8%, 7.6% and 11.4% for

ERS-RA (reported), ERS-RA (measured), ACC/AHA, and ACC/AHAx1.5

respectively. Thus, ACC/AHA adjusted according to EULAR recommendations

(ACC/AHAx1.5), produced higher risk estimations than both variants of ERS-RA,

although ERS-RA includes variables on disease activity.

29

The distribution of events in patients classified to a risk <7.5% or ≥7.5% was

investigated for the risk scores. The positive and negative predictive values of an

estmimated risk <7.5% or ≥7.5% did not differ much, but both the highest positive

and negative predictive values were seen for ERS-RA (measured). The risk scores

were compared to each other in terms of reclassification of patients with future

events from the low-risk (<7.5%) to the high-risk group (≥7.5%). Using ERS-RA

(reported) as comparator, ERS-RA (measured) and both variants of ACC/AHA

correctly reclassified patients with future events to the high-risk group (≥7.5%).

Figure 5. Observed and predicted proportion of events in deciles of estimated risk from ERS-RA (reported) (upper left), ERS-RA (measured) (upper right), ACC/AHA (lower left), and ACC/AHAx1.5 (lower right). Black bars are observed proportion of events, grey bars estimated proportion of events.

30

Paper II

Paper II reports the results from a long-term follow-up of the first study that

identified a higher prevalence of atherosclerosis in patients with RA. The first

study (“baseline”) was undertaken in 1997-1998 and included both IMT and

several biomarkers (Wållberg-Jonsson et al. 2001, 2002). In the present follow-

up, the presence of CAC was examined, and factors associated with presence of

CAC were investigated.

The study included 22 patients (4male/18female, mean age 65 years, RA-

duration 30-36 years) from the baseline (n=39) cohort of patients with sero-

positive RA. The amount of CAC was quantified according to Agatston (Agatston

et al. 1990). Patients with CAC 0-10 were classified as having low CAC (n=10),

whereas patients with CAC >10 (range 18-1700) were classified as having high

CAC (n=12), a classification based on previous studies (Budoff et al. 2007).

Patients with high CAC had significantly higher ESR (24.3 vs 9.9 mm/h), swollen

joint count (2 vs 0) and DAS 28 (3.61 vs 2.31) than patients with low CAC.

When patients with high or low CAC were analysed in multivariable OPLS

models, the information in the independent variables from baseline and follow-

up to a great extent separated the two groups of patients from each other. In the

first OPLS model (“model 1”), independent variables both from baseline and

follow-up were included. Figure 6 illustrates how patients with high CAC and low

CAC are separated by the score vectors in the model. Score vector 1 is the latent

variable that best separates patients with CAC 0-10 from patients with CAC >10,

whereas score vector 2 contains

information that does not

separate the two groups of

patients from each other, but still

varies in a non-random manner.

Figure 7 presents how the

variables in the model are related

to each other and how they

contribute to the discrimination

between patients with high CAC

and low CAC. The more to the

right a variable is plotted, the

stronger is the association with

high CAC. R2 for this model was

0.87. Subsequent ROC analysis

displayed a sensitivity of 89% and

a specificity of 85% in

discriminating between high or

-5

0

5

-5 0 5

Sco

re v

ecto

r 2

Score vector 1

Figure 6. Score scatter plot for OPLS model

1. Vector 1 separates CAC 0-10 from CAC>10.

Filled diamonds represent patients with CAC

>10, empty diamonds patients with CAC 0-10.

31

low CAC. When the baseline ultrasound variables (IMT and plaque) were omitted

from the model, R2 was 0.86, sensitivity was 80% and specificity 83%.

The next OPLS model to be investigated included baseline variables, but not the

follow-up variables, nor delta values. This model was tested to investigate if the

baseline variables could predict CAC status 13 years later. R2 for this model was

0.67, sensitivity 73% and specificity 82% in separating patients with high CAC

from patients with low CAC. Variable importance in projection is presented in

Figure 8. All variables except HDL and leptin were positively related to high CAC.

FU HT treatmentFU statin

Sex

Age

Smoking

Delta ESR

Delta Haptoglobin

Cholesterol LDL

HDL

LDL/HDL

IMT

FU HAQ

FU DAS28

FU SJC

IL2sR

E-selectin

PAI mass

tPA

CIC

OxLDL IgM

Plaque

Acc activity

BMI

CAC 0-10 CAC >10

-0,5

-0,4

-0,3

-0,2

-0,1

0

0,1

0,2

0,3

0,4

0,5

-0,6 -0,4 -0,2 0 0,2 0,4 0,6

Load

ing

vect

or

2

Loading vector 1

Figure 7. Loading plot for OPLS model 1. The dots represent the loadings of the single

variables in each of the two vectors (latent variables) in OPLS model 1. This plot displays

how the variables are related to each other and to CAC status. The distance from origo is

proportional to the importance in the OPLS model. Acc activity: accumulated disease

activity. BMI: body mass index. CAC: coronary artery calcification. CIC: circulating

immune complexes. DAS28: 28 joint disease activity score. ESR: erythrocyte

sedimentation rate. FU: follow-up. HAQ: health assessment questionnaire. HDL: high-

density lipoproteins. HT: hypertension. IL2sR: soluble receptor of interleukin-2. IMT:

carotid intima media thickness. LDL: low-density lipoproteins. OxLDL: oxidized LDL.

PAI: plasminogen activator inhibitor-1. SJC: swollen joint count. VAS: visual analogue

scale.

32

The last OPLS model (not shown) included initially all baseline variables but IMT

and plaque, yielding an R2 of 0.58. Sensitivity was 67% and specificity 80% in

predicting high or low CAC at follow-up 13 years later.

Papers III and IV

Papers III and IV comprise the same cohort of patients with early RA, in whom

atheosclerosis has been prospectively studied with ultrasound measurement of

carotid IMT. A cohort of matched controls is examined as well. Paper III reports

the associations between IMT measured at baseline (T0), after 1.5 years (T1.5)

and after 11 years (T11), infection with CMV at T0 and CD28null T-cells assessed

at T11. Paper IV reports the associations between IMT, bone mineral density at

T11, and markers of bone turnover measured both at T0 and T11. This paper

includes data from 50 patients with RA and 29 controls. In both patients and

Figure 8. OPLS variable importance in projection (VIP) model 2. The numerical value

for a variable represents an estimation of it´s importance in the model, i.e. to what grade

it discriminates between the two groups of patients. Whiskers illustrate the upper limit of

bootstrapped confidence intervals. Acc activity: accumulated disease activity. BMI: body

mass index. CIC: circulating immune complexes. ESR: erythrocyte sedimentation rate.

FU: follow-up. HAQ: health assessment questionnaire. HDL: high-density lipoproteins.

IL2sR: soluble receptor of interleukin-2. IMT: carotid intima media thickness. LDL: low-

density lipoproteins. ox-LDL: oxidized LDL. PAI: plasminogen activator inhibitor-1 mass.

tPA: tissue plasminogen activator.

33

controls, 66% of the individuals were IgG-positive for CMV at T0. Neither did the

concentration of antibodies differ between patients and controls.

Patients with RA who were CMV IgG-positive had a rapid increase in IMT after

onset of RA, compared with controls and CMV IgG-negative patients, as

presented in Figure 9. However, the number of individuals in each category was

small, contributing to p-values 0.01-0.08 in pairwise comparisons of IMT at T1.5.

Furthermore, CMV-positive patients had a higher percentage of CD28null T-cells

of both CD4+ (median 7% vs 0.8%) and CD8+ types (mean 52% vs 27%).

Regression models revealed significant associations between presence of CMV

IgG and CD28null T-cells of both CD4+ and CD8+ types (Table 7). Since the

median percentage of CD4+CD28null T-cells have been determined to 1.6-1.9%

in healthy individuals of the general population (Schmidt et al. 1996, Goronzy et

al. 2001), a level of CD4+CD28null ≥2% was regarded elevated in further analyses

of the relation between CD4+CD28null T-cells and other variables.

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Positive patientPositive controlNegative controlNegative patient

T0T1.5

T11

IMT,

mm

Figure 9. Median IMT in patients with early RA during 11 years of follow-up. Results

from carotid ultrasound in 20 patients with RA and 17 controls, grouped by CMV-status.

Pairwise comparisons of change in IMT from T0 to T1.5 (CMV-positive patients as

reference): CMV-positive controls p=0.06; CMV-negative controls p=0.01; CMV-negative

patients p=0.08. CMV-positive patients n=10; CMV-positive controls n=12; CMV-

negative controls n=5; CMV-negative patients n=10.

34

When relations between subsets of T-cells and IMT at T11 were analysed in linear

regression models, an elevated percentage of CD4+CD28null T-cells (≥2%) was

associated with a higher IMT at T11, adjusted for systolic blood pressure (Table

8). Furthermore, there was a positive association between the percentage of

CD8+CD28null T-cells and IMT, adjusted for blood pressure (Table 8). In

regression models including age, neither CD28null T-cells nor blood pressure

were significantly associated with IMT at T11. CX3CR1 was abundant in

CD28null T-cells of both CD4+ and CD8+ type, but CX3CR1 per se was not

associated with IMT (data not shown).

Paper IV presents data from examinations of 54 patients and 31 controls, in whom

several biomarkers associated with bone turnover were measured both at T0 and

T11. Moreover, BMD was measured at T11 in patients resident in Västerbotten.

When patients and controls were compared, there were no significant differences

in BMD or concentrations of bone biomarkers, except for OPG (higher in controls

at T0 and higher in patients at T11) and PTH (higher in controls at T0).

Table 7. Multivariable regression models with CD4+CD28null and CD8+CD28null at T11 as dependent variables. CD4+CD28null >2%, yes

(logistic regression) CD8+CD28null, % (linear regression)

Variable Exp. B (CI 95%) B (CI 95%) Age, year 1.0 (0,95; 1.15) 0.75 (0.1; 1.4)* Anti-CCP, yes 0.41 (0.07; 2.3) -2.1 (-13.3; 9.2) Positive CMV IgG at T0, yes

20.3 (3.2; 130)** 16.2 (3.1; 29.3)*

R2 0.55 R2 0.32 T0: baseline. T11: Follow-up eleven years from baseline. *: p<0.05. **: p<0.01.

Table 8. Multivariable linear regression models with IMT at T11 (1/10 mm) as dependent variable. Variable Model 1

B (95% CI) Model 2 B (95% CI)

CD8+CD28null, % 0.02 (0.006; 0.04)* CD4+CD28null≥2%, yes 0.93 (0.12; 1.7)* Systoloc blood pressure T11, mmHg

0.05 (0.02; 0.07)** 0.05 (0.02; 0.07)**

R2 0.23 R2 0.25 T11: Follow-up eleven years from baseline. *: p<0.05. **: p<0.01.

35

Table 9. Multivariable linear regression model with IMT at T11 (1/10 mm) as response variable. Variable B (CI 95%) B (CI 95%) Systolic BP at T11, mm Hg 0.02 (-0.005; 0.04) 0.04 (0.1; 0.06)** OCN at T11, ng/mL -0.05 (-0.1; -0.0001)* -0.1 (-0.1; -0.02)** OPG at T11, ng/mL 2.0 (0.2; 3.8)* 3.4 (1.4; 5.4)** Age, years 0.08 (0.05; 0.1)*** R2 0.48 R2 0.26 BP: blood pressure. OCN: osteocalcin. OPG: osteoprotegerin. T11: Follow-up eleven years from baseline. *: p<0.05. **: p<0.01 ***: p<0.001

In patients with RA, OPG and OCN at T11 were significantly associated with IMT

at T11, adjusted for age and systolic blood pressure (Table 9). Change in IMT from

T0 to T11 was associated with OPG at T0 (B 2.0, p=0.07) and OCN at T0 (B-0.07,

p=0.02). However, none of the collagen derived markers of pure bone turnover

P1NP and CTX, nor BMD measured at T11, was significantly associated with IMT

at T11, in linear regression models including traditional cardiovascular risk

factors.

The relation between IMT, OPG and OCN was more pronounced in patients with

erosions at baseline, where a linear regression model with IMT at T11 as

dependent variable and independent variables OPG (B 9.6, p=0.005) and OCN

(B -0.2, p=0.04) yielded an adjusted R2 of 0.45. Furthermore, RANKL at T11

showed a strong association with IMT at T11 in patients with erosions at onset of

RA (standardized β=0.98, p=3x10-5).

36

Discussion

As shown in paper I, cardiovascular risk estimation in patients with RA still needs

to be improved. The RA-specific risk score ERS-RA did not make more accurate

risk estimations than the ACC/AHA risk score designed for the general

population, although ERS-RA comprises not only traditional cardiovascular risk

factors, but also variables reflecting inflammatory activity in RA, corticosteroid

use and duration of RA. Inflammation is well known to be associated with a

higher risk of cardiovascular disease in patients with RA (Wållberg-Jonsson et al.

1999, Maradit-Kremers et al. 2005, Innala et al. 2011), so inclusion of this

parameter in a risk score is expected to make it perform better than risk scores

for the general population. ERS-RA has recently been externally validated in large

Swedish cohorts and found to perform well (Ljung et al. 2018), but it was not

compared to other risk scores. When compared to risk scores for the general

population, no RA-specific risk score has outperformed the general population

risk scores (Arts et al. 2015a, Crowson et al. 2017a). Neither did an RA-adapted

verision of SCORE (Arts et al. 2015b), nor the RA-specific risk score developed by

the ATACC-RA consortium perform (Crowson et al. 2017b) better than risk scores

for the general population. This raises questions: are assessments of the

traditional risk factors reliable in patients with RA? Are the variables reflecting

disease activity and severity in RA reliable in the context of risk estimation?

The first question could be answered no. We know that blood lipids in patients

with ongoing inflammation in general are lower (Choy and Sattar 2009) and that

there are qualitative changes in the blood lipids as well (Rantapää-Dahlqvist et

al. 1991). Furhermore, hypertension and hyperlipidaemia are underdiagnosed

and undertreated in patients with RA (Burggraaf et al. 2019). This points to the

need of activity in risk estimation – when risk factors are not assessed, they

cannot be included in risk estimation, and neither can they be treated. In ERS-

RA, not only smoking and diabetes, but also hypertension and hyperlipidaemia

are dichotomous variables. If blood lipids and blood pressure are not measured,

they will not be included in the risk estimation. This is illustrated in paper I,

where ERS-RA did not perform very well when risk estimation was done using

patient´s or physician´s diagnosis of hypertension and hyperlipidaemia, whereas

the estimations were more accurate when measured levels of blood pressure and

blood lipids were used.

When it comes to the question of variables reflecting RA in risk estimation, one

has to keep in mind what the commonly used variables represent. The HAQ is a

patient reported disability index, where a high score probably can represent

different pathologies in patients with early disase, and in patients with a long

duration of RA, respectively. In early RA, an elevated HAQ most probably is

37

caused by ongoing inflammation, whereas it in patients with long-standing

diasease can be caused by erosions and other long-term effects of RA. Thus, an

elevated HAQ is not a reliable marker of inflammation in patients with long-term

RA. Furthermore, disease activity measured by DAS28 is much affected by

patient´s reported components pain, tender joints and general health. According

to a recent meta-analysis (Duffield et al. 2018), about 20% of patients with RA

have concomitant fibromyalgia, elevating DAS28 with 1.24 points compared to

patients with RA alone. This difference is not casued by differences in the

objective inflammatory components ESR or swollen joints, but comes from

patient reported components tender joints and pain, making DAS28 and other

scores including these components unreliable in patients with RA and

fibromyalgia. Of course this causes problems in all aspects of management of RA,

not only risk estimation.

The process of development of cardiovascular risk scores is another potential

confounder in evaluation of RA-specific risk scores. Cardiovascular risk scores

are developed from cohorts of patients, in whom potential predictors of

cardiovascular events are registered. After a period of follow-up, the relation

between possible predictors of events and actual events during follow-up is

analysed, and an algorithm based on the predictors values at baseline is

contructed. Thus, a new risk score is a representation of risk factors from the past,

so if the population changes over time, the risk score can become obsolete. When

it comes to RA, the treatment strategies have obviously changed over time.

Recommendations today comprise intense treatment and ambitious targets of

treatment, in addition to the new drugs available (Smolen et al. 2017), making

modern populations of patients with RA different from earlier populations. It is

also plausible that genetic variation and geographical differences in treatment

strategies limit the generalizability of populations with RA. This points out the

need of developing and evalutating RA-specific risk scores in sub-populations of

patients with RA, where patients are classified according to duration of disease,

age at disease onset, time of disease onset, geography, genetic background,

disease activity, and treatment.

One aspect of risk estimation, is whether it is useful to identify patients with an

elevated risk. If the risk cannot be decreased, identification of high-risk

individuals is useless. A recent study examined the effect of cardiovascular

prevention in patients with RA, and found that intense treatment of cardio-

vascular risk factors decreased the progress of IMT and reduced the risk of

cardiovascular events, although the number of events was small (Burggraaf et al.

2019). This tells us that cardiovascular risk prevention by treatment of traditional

cardiovascular risk factors in patients with RA is possible. Considering the

observations of a relationship between inflammation and risk of myocardial

infarction and acute coronary syndrome (Ljung et al. 2016, Meissner et al. 2016),

38

it is reasonable that the risk of atherosclerotic cardiovascular disease to a great

extent is modifyable. The relation between inflammation and atherosclerosis has

been studied in the general population, where several trials of prevention of

cardiovascular events using anti-inflammatory agents have been performed, but

mostly with negative results (Vaidya et al. 2019). A recent study with negative

results, was the Cardiovascular Inflammation Reduction Trial (CIRT), where

methotrexate did not reduce the risk of cardiovascular disease in patients with

diabetes and a previous myocardial infarction (Ridker et al. 2018), an observation

that contradicts the observations in patients with RA. Unlike other substances

investigated in this area, canakinumab (Ridker et al. 2017) and colchicine (Nidorf

et al. 2013) both have been effective in cardiovascular prevention in the general

population. Interestingly, these two substances have a common effect:

canakinumab is a potent inhibitor of interleukin 1β, while colchicine by inhibiting

the NLRP3-inflammasome decreases the production of interleukin 1 β. This

implies that reduction of inflammation is not enough to have effect on

atherosclerosis: the mechanism is crucial. If this holds true in RA is yet to be

explored, but the reduced risk of cardiovascular events in patients with RA

treated with methotrexate (Choi et al. 2002), suggests that the mechanisms

linking inflammation and atherosclerosis are different in patients with and

without RA.

As shown by Crowson et al (Crowson et al. 2018), traditional cardiovascular risk

factors and RA-associated risk factors together explain 70% of the cardiovascular

risk in patients with RA. This suggests that more risk factors are yet to be

identified. My study presented in Paper IV, identifies OPG, and possibly OCN, as

potential markers of atherosclerosis in patients with RA. Previous studies of OPG

in patients with RA have been consistent with this: OPG has been associated with

CAC (Asanuma et al. 2007), carotid plaque (Asanuma et al. 2013), endothelial

activation (Dessein et al. 2014), IMT (Dessein et al. 2014, Beyazal et al. 2016),

and pulse wave velocity (Beyazal et al. 2016). These measures represent stages of

atherosclerosis from the early endothelial activation, to calcified plaques,

implying that the relation between OPG and atherosclerosis is consistent, further

corroborated by a study linking increased levels of OPG in patients with RA to

established cardiovascular disease (López-Mejias et al. 2015). The question is,

whether OPG is a risk factor or a marker of atherosclerosis. In prospective cohort

studies, higher concentrations of OPG have been independently associated with

future atherosclerotic cardiovascular disease, adjusted for traditional risk factors

(Vik et al. 2011, Mogelvang et al. 2013, Tschiderer et al. 2017), although OPG is

associated with several traditional risk factors (Mogelvang et al. 2012). In patients

with RA, no prospective studies have been published.

A potential confounder in studies of OPG, is the fact that it is produced not only

by osteoblasts, but also in several other tissues and cell types, including vascular

39

smooth muscle cells and endothelial cells (Rochette et al. 2018), from the latter

secreted as a response to inflammation (Zannettino et al. 2005). Despite this, the

relation between OPG and inflammation is inconsistent in patients with RA. A

meta-analysis did not show a significant relation with DAS28, the only measure

of inflammation reported (Wang et al. 2017). In Paper IV, OPG was associated

with ESR, but not with other measures of inflammation, as in previous studies of

atherosclerosis, where a relation with measures of inflammation was reported in

one in one study (Asanuma et al. 2007), but not in others (Dessein et al. 2014,

López-Mejias et al. 2015, Beyazal et al. 2016). Also in general population cohorts,

the relation between OPG and inflammation measured as high-sensitive CRP has

been contradictory (Vik et al. 2011, Mogelvang et al. 2012). This suggests that

OPG is not the final link between atherosclerosis and inflammation in patients

with RA, but still is a marker of atherosclerosis. However, more longitudinal

studies are required, since my study is the first in patients with RA, where a cohort

of patients is followed over time.

Paper II reports results from a long-term follow-up of the very first study that

reported increased atherosclerosis in patients with RA (Wållberg-Jonsson et al.

2001). In the baseline study IMT was measured, but in this follow-up CAC was

measured, and found to be associated with inflammation in all stages of RA. A

large amount of biomarkers was measured at baseline, but no biomarkers

reflecting or regulating bone turnover. Otherwise that study might have shed

some light on the relations between atherosclerosis, inflammation and bone

biomarkers in patients with RA in patients with long-standing disease.

The mechanisms linking OPG and other markers of bone turnover and athero-

sclerosis are not sufficiently elucidated, but bearing in mind that the process of

plaque calcification is an active process (Rochette et al. 2018), it is reasonable to

think of it as independent from the regulation of bone turnover. My study

supports this idea, since levels of the collagen degradation product CTX, released

from osteoclasts during bone resorption, was not associated with atherosclerosis.

Neither was P1NP, a product from bone formation, associated with athero-

sclerosis, indicating that neither resorption nor formation of bone was related to

IMT, albeit OPG was. On the other hand, OCN, a protein only secreted by

osteoblasts (Moser and van der Eerden 2018), was negatively associated with

atherosclerosis, providing a link between low osteoblast activity and

atherosclerosis. However, OCN has been found to be less a regulator of

calcification, than a hormone regulating such disparate things as brain

development, response to exercise and glucose metabolism (Moser and van der

Eerden 2018). In accordance with this finding, results from studies of

atherosclerosis and OCN in the general population have been disparate, with the

exception of histological studies of vascular calcification, where OCN or OCN-

40

positive cells consequently have been associated with calcification (Millar et al.

2017).

The hypothesis of infection causing atherosclerosis has been studied (Dahal et al.

2017), and several serological studies have linked infection with CMV to

atherosclerosis and cardiovascular disease (Sorlie et al. 2000, Dahal et al. 2017).

This hypothesis was tested in paper III, where IMT increased rapidly after onset

of RA in CMV IgG-positive patients (Figure 9). This suggests an additional effect

of RA and CMV, propagating the process of atherosclerosis when a CMV-positive

individual develops RA. Furthermore, the presence of CD28null T-cells was

higher in CMV IgG-positive patients, as in previous studies in patients with RA

(Hooper et al. 1999, Broadley et al. 2017), and the general population (Gratama

et al. 1987, Derhovanessian et al. 2011). CD28null T-cells are terminally

differentiated, with properties different from the normal CD28+ T-cells:

CD4+CD28null T-cells are no longer T-helper cells, but are cytotoxic, whereas

CD8+CD28null T-cells demonstrate both increase and decrease in cytotoxic

properties, pointing to the fact that the population of CD28null T-cells is

heterogenous (Mou et al. 2014). Nevertheless, CD28null T-cells were associated

with atherosclerosis in paper III, as well as in previous studies in patients with

RA (Gerli et al. 2004, Pingiotti et al. 2007, Winchester et al. 2016). It has been

suggested that CMV-infection through induction of CD4+CD28null T-cells is the

cause of increased cardiovascular mortality in patitents with RA (Broadley et al.

2017), and our results do not contradict this idea, although more eveidence

definitely is needed. Thinking of CMV and atherosclerosis, the hypothesis of an

immediate effect of CMV on atherosclerosis cannot be rejected: the endothelial

cells are in fact infected by CMV and become activated from infection with CMV,

and although the immediate response is downregulated, the endothelium is

chronically infected (Jeffery et al. 2013). Thus, it is reasonable to think of CMV in

itself as a potential cause of atherosclerosis.

41

Conclusions

From this study of atherosclerosis and risk of cardiovascular disease in patients

with RA, the following conclusions can be made:

Risk estimation scores for patients with RA need to be improved.

Blood pressure and blood lipids must me measured in patients with RA,

to enable accurate risk estimation.

The amount of coronary artery calcification is related to both present and

previous inflammation in patients with RA.

CD28null T-cells are associated with increased atherosclerosis in

patients with RA.

CMV IgG-positive patients with RA have increased atherosclerosis.

Osteoprotegerin, and possibly osteocalcin, could be used as indicators of

atherosclerosis in patients with RA.

Bone turnover per se is not associated with atherosclerosis in patients

with RA.

42

Future perspectives

As stated in previous parts of this thesis, there are still many areas where more

knowledge is needed. The first of them, is how to improve the cardiovascular risk

estimation in patients with RA. There is a need for an RA-specific risk score that

performs better than risk scores for the general population. This probably will not

be one uniform risk score applicable to all patients with RA, but rather a risk score

where the patients are categorized according to disease characteristics, and

perhaps genetic and geographic background. The ACC/AHA risk score includes

interaction between the included variables, since the impact of a variable depends

on other variables. This probably holds true even more for patients with RA,

where there are more potential interactions to include in the risk estimation, e.g.

the well-known interaction between inflammation and blood lipids. The first step

would be to recalibrate the ACC/AHA algorithm using patients with RA, and

investigate if this improves the performance. Secondly, characteristics of RA,

genetics, and geography could be included in the recalibrated and the original

ACC/AHA risk score. The characteristics of RA would preferably include not only

measures of present inflammation, RF- or ACPA-status and HAQ, but also

duration of disease, age at disease onset, time of disease onset, geography, genetic

background, and treatment. As a last step, inclusion of new variables could be

included, e.g. OPG, OCN and CMV IgG-positivity.

Another question to be answered, is if antiviral treatment for CMV could be

efficient in reducing atherosclerosis and atherosclerotic cardiovascular disease in

patients with RA. A randomized controlled trial using antiviral treatment could

provide not only an opportunity to treat cardiovascular risk, but also give useful

information whether CMV per se is involved in atherogenesis. If analysis of the

T-cell repertoire is performed along, some light could be shed on the interaction

between CMV, T-cells and atherosclerosis in patients with RA.

Numerous studies of biomarkers in terms of atherosclerosis have been performed

in patients with RA, but the histological studies are scarce. Patients with RA are

subject to vascular surgery, and studies of specimen from these patients would

give an opportunity to examine cell types and biomarkers in situ, in contrast to

soluble biomarkers and circulating cells. As a matter of fact, the disease is located

to the arterial wall, not what circulates in the lumen.

43

Acknowledgements

Many people have helped me along the way to make this possible. I want to thank

you all. Some require to be mentioned:

First of all, my supervisor and supporter Solveig Wållberg Jonsson, who

transformed me from a research consumer to a research producer, and with such

patience followed me to the bitter end. I am also very grateful to co-supervisor

Anna Södergren.

All people in the research groups at the Department of Rheumatology, and in

particular Solbritt Rantapää Dahlqvist, without whom nothing of this would

have been.

My colleagues at the clinic, who work together with solidarity and make every day

a nice day.

Kristina Juneblad, head of the clinic, Gerd-Marie Alenius, former head of the

clinic, and Helena Forssblad d’Elia, Professor and chief of research at the clinic,

for giving me the opportunity to spend time on research.

Statistician Fredrik Jonsson, for such patience and being a gifted teacher,

providing explanations that made me believe I even understood latent vectors in

multidimensional spaces.

Kristina Eriksson, Elisabet Lundström, Gun-Britt Johansson, Ann-Cathrin

Kallin, Sonja Odeblom, Viktoria von Zweigbergk, and Lena Uddståhl for

acquisition of data.

All my co-workers not mentioned elsewhere: Lena Innala, Staffan Magnusson,

Bozena Möller, Torgny Smedby, Thomas Meedt, Michael Henein, Andreas

Fasth, Kjell Karp, Kristina Lejon, Vivi Malmström, Afsar Rahbar, and Anna

Ramnemark.

Swedish society of Rheumatology, for financial support to presentation of my

results in a sauna-like Madrid.

My family: parents Anders and Ylva Britt for eternal love and support, not only

for being role models in making science in the evenings, brother Björn, and last

but not least my wife Johanna and our three little girls Tyra, Signe and Märta.

44

References

Agatston, A.S.; Janowitz, W.R.; Hildner, F.J.; Zusmer, N.R.; Viamonte Jr., M.; Detrano, R. 1990. Quantification of coronary artery calcium using ultrafast computed tomography. Journal of the American College of Cardiology 15: 827–832.

Agca, R.; Heslinga, S.C.; Rollefstad, S.; Heslinga, M.; McInnes, I.B.; Peters, M.J.L.; et al. 2017. EULAR recommendations for cardiovascular disease risk management in patients with rheumatoid arthritis and other forms of inflammatory joint disorders: 2015/2016 update. Annals of the Rheumatic Diseases 76: 17–28.

Ajeganova, S.; Humphreys, J.H.; Verheul, M.K.; Steenbergen, H.W. van; Nies, J.A.B. van; Hafström, I.; et al. 2016. Anticitrullinated protein antibodies and rheumatoid factor are associated with increased mortality but with different causes of death in patients with rheumatoid arthritis: a longitudinal study in three European cohorts. Annals of the Rheumatic Diseases 75: 1924–1932.

Aletaha, D.; Nell, V.P.; Stamm, T.; Uffmann, M.; Pflugbeil, S.; Machold, K.; et al. 2005. Acute phase reactants add little to composite disease activity indices for rheumatoid arthritis: validation of a clinical activity score. Arthritis Research & Therapy 7: R796–R806.

Aletaha, D.; Neogi, T.; Silman, A.J.; Funovits, J.; Felson, D.T.; Bingham, C.O.; et al. 2010. 2010 Rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis and Rheumatism 62: 2569–2581.

Alpízar-Rodríguez, D.; Pluchino, N.; Canny, G.; Gabay, C.; Finckh, A. 2017. The role of female hormonal factors in the development of rheumatoid arthritis. Rheumatology 56: 1254–1263.

Ambrosino, P.; Lupoli, R.; Di Minno, A.; Tasso, M.; Peluso, R.; Di Minno, M.N.D. 2015. Subclinical atherosclerosis in patients with rheumatoid arthritis. A meta-analysis of literature studies. Thrombosis and Haemostasis 113: 916–930.

Apel, F.; Zychlinsky, A.; Kenny, E.F. 2018. The role of neutrophil extracellular traps in rheumatic diseases. Nature Reviews Rheumatology 14: 467.

Arnett, F.C.; Edworthy, S.M.; Bloch, D.A.; Mcshane, D.J.; Fries, J.F.; Cooper, N.S.; et al. 1988. The american rheumatism association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis & Rheumatism 31: 315–324.

Arts, E.E.A.; Popa, C.; Broeder, A.A.D.; Semb, A.G.; Toms, T.; Kitas, G.D.; et al. 2015a. Performance of four current risk algorithms in predicting cardiovascular events in patients with early rheumatoid arthritis. Annals of the Rheumatic Diseases 74: 668–674.

Arts, E.E.A.; Popa, C.D.; Broeder, A.A.D.; Donders, R.; Sandoo, A.; Toms, T.; et al. 2015b. Prediction of cardiovascular risk in rheumatoid arthritis: performance of original and adapted SCORE algorithms. Annals of the Rheumatic Diseases 75: 674–680.

Asanuma, Y.; Chung, C.P.; Oeser, A.; Solus, J.F.; Avalos, I.; Gebretsadik, T.; et al. 2007. Serum osteoprotegerin is increased and independently associated

45

with coronary-artery atherosclerosis in patients with rheumatoid arthritis. Atherosclerosis 195: e135-141.

Asanuma, Y.F.; Shimada, Y.; Kouzu, N.; Yokota, K.; Nakajima, K.; Sato, K.; et al. 2013. Serum osteoprotegerin concentration is associated with carotid atherosclerotic plaque in patients with rheumatoid arthritis. Modern rheumatology / the Japan Rheumatism Association 23: 269–275.

Aubry, M.-C.; Maradit-Kremers, H.; Reinalda, M.S.; Crowson, C.S.; Edwards, W.D.; Gabriel, S.E. 2007. Differences in atherosclerotic coronary heart disease between subjects with and without rheumatoid arthritis. The Journal of Rheumatology 34: 937–942.

Avina-Zubieta, J.A.; Thomas, J.; Sadatsafavi, M.; Lehman, A.J.; Lacaille, D. 2012. Risk of incident cardiovascular events in patients with rheumatoid arthritis: a meta-analysis of observational studies. Annals of the Rheumatic Diseases 71: 1524–1529.

Baecklund, E.; Ekbom, A.; Sparen, P.; Feltelius, N.; Klareskog, L. 1998. Disease activity and risk of lymphoma in patients with rheumatoid arthritis: nested case-control study. BMJ : British Medical Journal 317: 180–181.

Baghdadi, L.R.; Woodman, R.J.; Shanahan, E.M.; Mangoni, A.A. 2015. The Impact of Traditional Cardiovascular Risk Factors on Cardiovascular Outcomes in Patients with Rheumatoid Arthritis: A Systematic Review and Meta-Analysis. PLoS ONE 10.

Barbhaiya, M.; Solomon, D.H. 2013. Rheumatoid Arthritis and Cardiovascular Disease: Update on Treatment Issues. Current opinion in rheumatology 25: 317–324.

Begovich, A.B.; Carlton, V.E.H.; Honigberg, L.A.; Schrodi, S.J.; Chokkalingam, A.P.; Alexander, H.C.; et al. 2004. A Missense Single-Nucleotide Polymorphism in a Gene Encoding a Protein Tyrosine Phosphatase (PTPN22) Is Associated with Rheumatoid Arthritis. American Journal of Human Genetics 75: 330–337.

Berglin, E.; Johansson, T.; Sundin, U.; Jidell, E.; Wadell, G.; Hallmans, G.; et al. 2006. Radiological outcome in rheumatoid arthritis is predicted by presence of antibodies against cyclic citrullinated peptide before and at disease onset, and by IgA-RF at disease onset. Annals of the Rheumatic Diseases 65: 453–458.

Beyazal, M.S.; Erdoğan, T.; Devrimsel, G.; Türkyılmaz, A.K.; Cüre, M.C.; Beyazal, M.; et al. 2016. Relationship of osteoprotegerin to pulse wave velocity andcarotid intima-media thickness in rheumatoid arthritis patients. Zeitschrift für Rheumatologie 75: 723–728.

Björkbacka, H.; Lavant, E.H.; Fredrikson, G.N.; Melander, O.; Berglund, G.; Carlson, J.A.; et al. 2010. Weak associations between human leucocyte antigen genotype and acute myocardial infarction. Journal of Internal Medicine 268: 50–58.

Blacher, J.; Guerin, A.P.; Pannier, B.; Marchais, S.J.; Safar, M.E.; London, G.M. 1999. Impact of aortic stiffness on survival in end-stage renal disease. Circulation 99: 2434–2439.

van Boekel, M.A.; Vossenaar, E.R.; van den Hoogen, F.H.; van Venrooij, W.J. 2002. Autoantibody systems in rheumatoid arthritis: specificity, sensitivity and diagnostic value. Arthritis Research 4: 87–93.

46

Book, C.; Karlsson, M.K.; Åkesson, K.; Jacobsson, L.T.H. 2009. Early rheumatoid arthritis and body composition. Rheumatology 48: 1128–1132.

Brink, M.; Hansson, M.; Mathsson, L.; Jakobsson, P.-J.; Holmdahl, R.; Hallmans, G.; et al. 2013. Multiplex analyses of antibodies against citrullinated peptides in individuals prior to development of rheumatoid arthritis. Arthritis and Rheumatism 65: 899–910.

Brinkmann, V.; Zychlinsky, A. 2012. Neutrophil extracellular traps: is immunity the second function of chromatin? The Journal of Cell Biology 198: 773–783.

Broadley, I.; Pera, A.; Morrow, G.; Davies, K.A.; Kern, F. 2017. Expansions of Cytotoxic CD4+CD28− T Cells Drive Excess Cardiovascular Mortality in Rheumatoid Arthritis and Other Chronic Inflammatory Conditions and Are Triggered by CMV Infection. Frontiers in Immunology 8: 195.

Budoff, M.J.; Shaw, L.J.; Liu, S.T.; Weinstein, S.R.; Mosler, T.P.; Tseng, P.H.; et al. 2007. Long-term prognosis associated with coronary calcification: observations from a registry of 25,253 patients. Journal of the American College of Cardiology 49: 1860–1870.

Burggraaf, B.; Stoep, D.F. van B. der; Vries, M.A. de; Klop, B.; Liem, A.H.; Geijn, G.-J.M. van de; et al. 2019. Effect of a treat-to-target intervention of cardiovascular risk factors on subclinical and clinical atherosclerosis in rheumatoid arthritis: a randomised clinical trial. Annals of the Rheumatic Diseases: annrheumdis-2018-214075.

Celermajer, D.S.; Sorensen, K.E.; Gooch, V.M.; Spiegelhalter, D.J.; Miller, O.I.; Sullivan, I.D.; et al. 1992. Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis. Lancet (London, England) 340: 1111–1115.

Chambless, L.E.; Heiss, G.; Folsom, A.R.; Rosamond, W.; Szklo, M.; Sharrett, A.R.; et al. 1997. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) Study, 1987-1993. American Journal of Epidemiology 146: 483–494.

Choi, H.K.; Hernán, M.A.; Seeger, J.D.; Robins, J.M.; Wolfe, F. 2002. Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study. Lancet (London, England) 359: 1173–1177.

Choy, E.; Sattar, N. 2009. Interpreting lipid levels in the context of high-grade inflammatory states with a focus on rheumatoid arthritis: a challenge to conventional cardiovascular risk actions. Annals of the Rheumatic Diseases 68: 460–469.

Cobb, S.; Anderson, F.; Bauer, W. 1953. Length of Life and Cause of Death in Rheumatoid Arthritis. New England Journal of Medicine 249: 553–556.

Cross, M.; Smith, E.; Hoy, D.; Carmona, L.; Wolfe, F.; Vos, T.; et al. 2014. The global burden of rheumatoid arthritis: estimates from the Global Burden of Disease 2010 study. Annals of the Rheumatic Diseases 73: 1316–1322.

Crowson, C.S.; Matteson, E.L.; Roger, V.L.; Therneau, T.M.; Gabriel, S.E. 2012. Usefulness of Risk Scores To Estimate the Risk of Cardiovascular Disease in Patients with Rheumatoid Arthritis. The American Journal of Cardiology 110: 420–424.

Crowson, C.S.; Gabriel, S.E.; Semb, A.G.; van Riel, P.L.C.M.; Karpouzas, G.; Dessein, P.H.; et al. 2017a. Rheumatoid arthritis-specific cardiovascular

47

risk scores are not superior to general risk scores: a validation analysis of patients from seven countries. Rheumatology (Oxford, England) 56: 1102–1110.

Crowson, C.S.; Rollefstad, S.; Kitas, G.D.; van Riel, P.L.C.M.; Gabriel, S.E.; Semb, A.G. 2017b. Challenges of developing a cardiovascular risk calculator for patients with rheumatoid arthritis. PLoS ONE 12: e0174656.

Crowson, C.S.; Rollefstad, S.; Ikdahl, E.; Kitas, G.D.; van Riel, P.L.C.M.; Gabriel, S.E.; et al. 2018. Impact of risk factors associated with cardiovascular outcomes in patients with rheumatoid arthritis. Annals of the Rheumatic Diseases 77: 48–54.

Dahal, U.; Sharma, D.; Dahal, K. 2017. An Unsettled Debate About the Potential Role of Infection in Pathogenesis of Atherosclerosis. Journal of Clinical Medicine Research 9: 547–554.

Derhovanessian, E.; Maier, A.B.; Hähnel, K.; Beck, R.; de Craen, A.J.M.; Slagboom, E.P.; et al. 2011. Infection with cytomegalovirus but not herpes simplex virus induces the accumulation of late-differentiated CD4+ and CD8+ T-cells in humans. Journal of General Virology 92: 2746–2756.

Dessein, P.H.; Solomon, A.; Hollan, I. 2016. Metabolic abnormalities in patients with inflammatory rheumatic diseases. Best Practice & Research. Clinical Rheumatology 30: 901–915.

Dessein, P.H.; López-Mejias, R.; González-Juanatey, C.; Genre, F.; Miranda-Filloy, J.A.; Llorca, J.; et al. 2014. Independent Relationship of Osteoprotegerin Concentrations with Endothelial Activation and Carotid Atherosclerosis in Patients with Severe Rheumatoid Arthritis. The Journal of Rheumatology 41: 429–436.

Detrano, R.; Hsiai, T.; Wang, S.; Puentes, G.; Fallavollita, J.; Shields, P.; et al. 1996. Prognostic value of coronary calcification and angiographic stenoses in patients undergoing coronary angiography. Journal of the American College of Cardiology 27: 285–290.

Deyab, G.; Hokstad, I.; Whist, J.E.; Smastuen, M.C.; Agewall, S.; Lyberg, T.; et al. 2017. Methotrexate and anti-tumor necrosis factor treatment improves endothelial function in patients with inflammatory arthritis. Arthritis Research & Therapy 19: 232.

Duffield, S.J.; Miller, N.; Zhao, S.; Goodson, N.J. 2018. Concomitant fibromyalgia complicating chronic inflammatory arthritis: a systematic review and meta-analysis. Rheumatology (Oxford, England) 57: 1453–1460.

Dumitriu, I.E.; Araguás, E.T.; Baboonian, C.; Kaski, J.C. 2009. CD4+ CD28 null T cells in coronary artery disease: when helpers become killers. Cardiovascular research 81: 11–19.

Duthie, J.J.R.; Brown, P.E.; Truelove, L.H.; Baragar, F.D.; Lawrie, A.J. 1964. Course and Prognosis in Rheumatoid Arthritis. Annals of the Rheumatic Diseases 23: 193–204.

Eriksson, J.K.; Neovius, M.; Ernestam, S.; Lindblad, S.; Simard, J.F.; Askling, J. 2013. Incidence of Rheumatoid Arthritis in Sweden: A Nationwide Population-Based Assessment of Incidence, Its Determinants, and Treatment Penetration. Arthritis Care & Research 65: 870–878.

Felson, D.T.; Anderson, J.J.; Boers, M.; Bombardier, C.; Furst, D.; Goldsmith, C.; et al. 1995. American College of Rheumatology. Preliminary definition of

48

improvement in rheumatoid arthritis. Arthritis and Rheumatism 38: 727–735.

Fries, J.F.; Spitz, P.W.; Young, D.Y. 1982. The dimensions of health outcomes: the health assessment questionnaire, disability and pain scales. The Journal of Rheumatology 9: 789–793.

Gerli, R.; Schillaci, G.; Giordano, A.; Bocci, E.B.; Bistoni, O.; Vaudo, G.; et al. 2004. CD4+CD28- T lymphocytes contribute to early atherosclerotic damage in rheumatoid arthritis patients. Circulation 109: 2744–2748.

Global Burden of Disease. 2018. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet (London, England) 392: 1736–1788.

Goodson, N.J.; Silman, A.J.; Pattison, D.J.; Lunt, M.; Bunn, D.; Luben, R.; et al. 2004. Traditional cardiovascular risk factors measured prior to the onset of inflammatory polyarthritis. Rheumatology (Oxford, England) 43: 731–736.

Goronzy, J.J.; Fulbright, J.W.; Crowson, C.S.; Poland, G.A.; O’Fallon, W.M.; Weyand, C.M. 2001. Value of Immunological Markers in Predicting Responsiveness to Influenza Vaccination in Elderly Individuals. Journal of Virology 75: 12182–12187.

Gratama, J.W.; Naipal, A.M.; Oosterveer, M.A.; Stijnen, T.; Kluin-Nelemans, H.C.; Ginsel, L.A.; et al. 1987. Effects of herpes virus carrier status on peripheral T lymphocyte subsets. Blood 70: 516–523.

Gregersen, P.K.; Silver, J.; Winchester, R.J. 1987. The shared epitope hypothesis. an approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis & Rheumatism 30: 1205–1213.

Greyling, A.; van Mil, A.C.C.M.; Zock, P.L.; Green, D.J.; Ghiadoni, L.; Thijssen, D.H.; et al. 2016. Adherence to guidelines strongly improves reproducibility of brachial artery flow-mediated dilation. Atherosclerosis 248: 196–202.

Guerreiro, C.S.; Calado, Â.; Sousa, J.; Fonseca, J.E. 2018. Diet, Microbiota, and Gut Permeability—The Unknown Triad in Rheumatoid Arthritis. Frontiers in Medicine 5: 349.

Hansson, G.K.; Hermansson, A. 2011. The immune system in atherosclerosis. Nature Immunology 12: 204–212.

Haugeberg, G.; Uhlig, T.; Falch, J.A.; Halse, J.I.; Kvien, T.K. 2000. Bone mineral density and frequency of osteoporosis in female patients with rheumatoid arthritis: results from 394 patients in the Oslo County Rheumatoid Arthritis register. Arthritis and Rheumatism 43: 522–530.

van der Heijde, D.M.; van Riel, P.L.; van Leeuwen, M.A.; van ’t Hof, M.A.; van Rijswijk, M.H.; van de Putte, L.B. 1992. Prognostic factors for radiographic damage and physical disability in early rheumatoid arthritis. A prospective follow-up study of 147 patients. British Journal of Rheumatology 31: 519–525.

Hoes, J.N.; Goes, M.C. van der; Raalte, D.H. van; Zijl, N.J. van der; Uyl, D. den; Lems, W.F.; et al. 2011. Glucose tolerance, insulin sensitivity and β-cell function in patients with rheumatoid arthritis treated with or without low-to-medium dose glucocorticoids. Annals of the Rheumatic Diseases 70: 1887–1894.

49

Hooper, M.; Kallas, E.G.; Coffin, D.; Campbell, D.; Evans, T.G.; Looney, R.J. 1999. Cytomegalovirus seropositivity is associated with the expansion of CD4+CD28- and CD8+CD28- T cells in rheumatoid arthritis. The Journal of Rheumatology 26: 1452–1457.

Innala, L.; Möller, B.; Ljung, L.; Magnusson, S.; Smedby, T.; Södergren, A.; et al. 2011. Cardiovascular events in early RA are a result of inflammatory burden and traditional risk factors: a five year prospective study. Arthritis research & therapy 13: R131.

Iwatsuka, R.; Matsue, Y.; Yonetsu, T.; O’uchi, T.; Matsumura, A.; Hashimoto, Y.; et al. 2018. Arterial inflammation measured by 18F-FDG-PET-CT to predict coronary events in older subjects. Atherosclerosis 268: 49–54.

Jeffery, H.C.; Söderberg-Naucler, C.; Butler, L.M. 2013. Human Cytomegalovirus Induces a Biphasic Inflammatory Response in Primary Endothelial Cells. Journal of Virology 87: 6530–6535.

Kallberg, H.; Padyukov, L.; Plenge, R.M.; Ronnelid, J.; Gregersen, P.K.; van der Helm-van Mil, A.H.M.; et al. 2007. Gene-gene and gene-environment interactions involving HLA-DRB1, PTPN22, and smoking in two subsets of rheumatoid arthritis. American Journal of Human Genetics 80: 867–875.

Karpouzas, G.A.; Malpeso, J.; Choi, T.-Y.; Li, D.; Munoz, S.; Budoff, M.J. 2014. Prevalence, extent and composition of coronary plaque in patients with rheumatoid arthritis without symptoms or prior diagnosis of coronary artery disease. Annals of the Rheumatic Diseases 73: 1797–1804.

Khuder, S.A.; Peshimam, A.Z.; Agraharam, S. 2002. Environmental risk factors for rheumatoid arthritis. Reviews on Environmental Health 17: 307–315.

Klareskog, L.; Stolt, P.; Lundberg, K.; Källberg, H.; Bengtsson, C.; Grunewald, J.; et al. 2006. A new model for an etiology of rheumatoid arthritis: smoking may trigger HLA-DR (shared epitope)-restricted immune reactions to autoantigens modified by citrullination. Arthritis and Rheumatism 54: 38–46.

Kokkonen, H.; Stenlund, H.; Rantapää-Dahlqvist, S. 2017. Cardiovascular risk factors predate the onset of symptoms of rheumatoid arthritis: a nested case-control study. Arthritis Research & Therapy 19: 148.

Kokkonen, H.; Söderström, I.; Rocklöv, J.; Hallmans, G.; Lejon, K.; Rantapää Dahlqvist, S. 2010. Up-regulation of cytokines and chemokines predates the onset of rheumatoid arthritis. Arthritis and rheumatism 62: 383–391.

Larsen, A. 1995. How to apply Larsen score in evaluating radiographs of rheumatoid arthritis in long-term studies. The Journal of rheumatology 22: 1974–1975.

Laurent, S.; Cockcroft, J.; Van Bortel, L.; Boutouyrie, P.; Giannattasio, C.; Hayoz, D.; et al. 2006. Expert consensus document on arterial stiffness: methodological issues and clinical applications. European Heart Journal 27: 2588–2605.

Lawrence, J.S. 1970. Heberden Oration, 1969. Rheumatoid arthritis--nature or nurture? Annals of the Rheumatic Diseases 29: 357–379.

Libby, P.; Ridker, P.M.; Hansson, G.K. 2011. Progress and challenges in translating the biology of atherosclerosis. Nature 473: 317–325.

50

Lindqvist, E.; Eberhardt, K.; Bendtzen, K.; Heinegård, D.; Saxne, T. 2005. Prognostic laboratory markers of joint damage in rheumatoid arthritis. Annals of the Rheumatic Diseases 64: 196–201.

Liu, R.; Yamamoto, M.; Moroi, M.; Kubota, T.; Ono, T.; Funatsu, A.; et al. 2005. Chlamydia pneumoniae immunoreactivity in coronary artery plaques of patients with acute coronary syndromes and its relation with serology. American Heart Journal 150: 681–688.

Ljung, L.; Rantapää-Dahlqvist, S.; Jacobsson, L.T.H.; Askling, J. 2016. Response to biological treatment and subsequent risk of coronary events in rheumatoid arthritis. Annals of the Rheumatic Diseases 75: 2087–2094.

Ljung, L.; Ueda, P.; Liao, K.P.; Greenberg, J.D.; Etzel, C.J.; Solomon, D.H.; et al. 2018. Performance of the Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis in a geographically distant National Register-based cohort: an external validation. RMD open 4: e000771.

López-Mejias, R.; Ubilla, B.; Genre, F.; Corrales, A.; Hernández, J.L.; Ferraz-Amaro, I.; et al. 2015. Osteoprotegerin Concentrations Relate Independently to Established Cardiovascular Disease in Rheumatoid Arthritis. The Journal of Rheumatology 42: 39–45.

López-Mejías, R.; Castañeda, S.; González-Juanatey, C.; Corrales, A.; Ferraz-Amaro, I.; Genre, F.; et al. 2016. Cardiovascular risk assessment in patients with rheumatoid arthritis: The relevance of clinical, genetic and serological markers. Autoimmunity Reviews 15: 1013–1030.

MacGregor, A.J.; Snieder, H.; Rigby, A.S.; Koskenvuo, M.; Kaprio, J.; Aho, K.; et al. 2000. Characterizing the quantitative genetic contribution to rheumatoid arthritis using data from twins. Arthritis and Rheumatism 43: 30–37.

Maradit-Kremers, H.; Nicola, P.J.; Crowson, C.S.; Ballman, K.V.; Gabriel, S.E. 2005. Cardiovascular death in rheumatoid arthritis: a population-based study. Arthritis and rheumatism 52: 722–732.

Matsuzawa, Y.; Kwon, T.; Lennon, R.J.; Lerman, L.O.; Lerman, A. 2015. Prognostic Value of Flow‐Mediated Vasodilation in Brachial Artery and Fingertip Artery for Cardiovascular Events: A Systematic Review and Meta‐Analysis. Journal of the American Heart Association: Cardiovascular and Cerebrovascular Disease 4.

McDougall, C.; Hurd, K.; Barnabe, C. 2017. Systematic review of rheumatic disease epidemiology in the indigenous populations of Canada, the United States, Australia, and New Zealand. Seminars in Arthritis and Rheumatism 46: 675–686.

McGill Henry C.; McMahan C. Alex; Gidding Samuel S. 2008. Preventing Heart Disease in the 21st Century. Circulation 117: 1216–1227.

McInnes, I.B.; Schett, G. 2011. The Pathogenesis of Rheumatoid Arthritis. New England Journal of Medicine 365: 2205–2219.

McKenney-Drake, M.L.; Moghbel, M.C.; Paydary, K.; Alloosh, M.; Houshmand, S.; Moe, S.; et al. 2018. 18F-NaF and 18F-FDG as molecular probes in the evaluation of atherosclerosis. European Journal of Nuclear Medicine and Molecular Imaging 45: 2190–2200.

Meissner, Y.; Zink, A.; Kekow, J.; Rockwitz, K.; Liebhaber, A.; Zinke, S.; et al. 2016. Impact of disease activity and treatment of comorbidities on the

51

risk of myocardial infarction in rheumatoid arthritis. Arthritis Research & Therapy 18: 183.

Meune, C.; Touzé, E.; Trinquart, L.; Allanore, Y. 2009. Trends in cardiovascular mortality in patients with rheumatoid arthritis over 50 years: a systematic review and meta-analysis of cohort studies. Rheumatology (Oxford, England) 48: 1309–1313.

Millar, S.A.; Patel, H.; Anderson, S.I.; England, T.J.; O’Sullivan, S.E. 2017. Osteocalcin, Vascular Calcification, and Atherosclerosis: A Systematic Review and Meta-analysis. Frontiers in Endocrinology 8: 183.

Mogelvang, R.; Pedersen, S.H.; Flyvbjerg, A.; Bjerre, M.; Iversen, A.Z.; Galatius, S.; et al. 2012. Comparison of osteoprotegerin to traditional atherosclerotic risk factors and high-sensitivity C-reactive protein for diagnosis of atherosclerosis. The American Journal of Cardiology 109: 515–520.

Mogelvang, R.; Haahr-Pedersen, S.; Bjerre, M.; Frystyk, J.; Iversen, A.; Galatius, S.; et al. 2013. Osteoprotegerin improves risk detection by traditional cardiovascular risk factors and hsCRP. Heart (British Cardiac Society) 99: 106–110.

Moser, S.C.; van der Eerden, B.C.J. 2018. Osteocalcin-A Versatile Bone-Derived Hormone. Frontiers in Endocrinology 9: 794.

Mou, D.; Espinosa, J.; Lo, D.J.; Kirk, A.D. 2014. CD28 Negative T cells: is their loss our gain? American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 14: 2460–2466.

Muller, J.E.; Tofler, G.H.; Stone, P.H. 1989. Circadian variation and triggers of onset of acute cardiovascular disease. Circulation 79: 733–743.

Mutru, O.; Laakso, M.; Isomäki, H.; Koota, K. 1985. Ten year mortality and causes of death in patients with rheumatoid arthritis. British medical journal (Clinical research ed.) 290: 1797–1799.

Myasoedova, E.; Crowson, C.S.; Turesson, C.; Gabriel, S.E.; Matteson, E.L. 2011. Incidence of extra-articular rheumatoid arthritis in Olmsted County, Minnesota in 1995–2007 vs 1985–1994: a population-based study. The Journal of rheumatology 38: 983–989.

Neovius, M.; Simard, J.F.; Askling, J.; ARTIS study group. 2011. Nationwide prevalence of rheumatoid arthritis and penetration of disease-modifying drugs in Sweden. Annals of the Rheumatic Diseases 70: 624–629.

Ni Mhuircheartaigh, O.; Crowson, C.S.; Gabriel, S.E.; Roger, V.L.; Melton, L.J.; Amin, S. 2017. Fragility Fractures Are Associated with an Increased Risk for Cardiovascular Events in Women and Men with Rheumatoid Arthritis: A Population-based Study. The Journal of Rheumatology 44: 558–564.

Nidorf, S.M.; Eikelboom, J.W.; Budgeon, C.A.; Thompson, P.L. 2013. Low-Dose Colchicine for Secondary Prevention of Cardiovascular Disease. Journal of the American College of Cardiology 61: 404–410.

Nishimura, K.; Sugiyama, D.; Kogata, Y.; Tsuji, G.; Nakazawa, T.; Kawano, S.; et al. 2007. Meta-analysis: diagnostic accuracy of anti-cyclic citrullinated peptide antibody and rheumatoid factor for rheumatoid arthritis. Annals of Internal Medicine 146: 797–808.

52

Nurmohamed, M.T.; Heslinga, M.; Kitas, G.D. 2015. Cardiovascular comorbidity in rheumatic diseases. Nature Reviews. Rheumatology 11: 693–704.

Okada, Y.; Wu, D.; Trynka, G.; Raj, T.; Terao, C.; Ikari, K.; et al. 2014. Genetics of rheumatoid arthritis contributes to biology and drug discovery. Nature 506: 376–381.

Paakkanen, R.; Lokki, M.-L.; Seppänen, M.; Tierala, I.; Nieminen, M.S.; Sinisalo, J. 2012. Proinflammatory HLA-DRB1*01-haplotype predisposes to ST-elevation myocardial infarction. Atherosclerosis 221: 461–466.

de Pablo, P.; Dietrich, T.; McAlindon, T.E. 2008. Association of periodontal disease and tooth loss with rheumatoid arthritis in the US population. The Journal of Rheumatology 35: 70–76.

Padyukov, L.; Silva, C.; Stolt, P.; Alfredsson, L.; Klareskog, L. 2004. A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis. Arthritis and Rheumatism 50: 3085–3092.

Palomino-Morales, R.; Gonzalez-Juanatey, C.; Vazquez-Rodriguez, T.R.; Rodriguez, L.; Miranda-Filloy, J.A.; Pascual-Salcedo, D.; et al. 2010. Lack of association of PTPN22, STAT4 and TRAF1/C5 gene polymorphisms with cardiovascular risk in rheumatoid arthritis. Clinical and Experimental Rheumatology 28: 695–701.

Panoulas, V.F.; Metsios, G.S.; Pace, A.V.; John, H.; Treharne, G.J.; Banks, M.J.; et al. 2008. Hypertension in rheumatoid arthritis. Rheumatology 47: 1286–1298.

Perticone, F.; Ceravolo, R.; Pujia, A.; Ventura, G.; Iacopino, S.; Scozzafava, A.; et al. 2001. Prognostic significance of endothelial dysfunction in hypertensive patients. Circulation 104: 191–196.

Pietrobelli, A.; Formica, C.; Wang, Z.; Heymsfield, S.B. 1996. Dual-energy X-ray absorptiometry body composition model: review of physical concepts. American Journal of Physiology-Endocrinology and Metabolism 271: E941–E951.

Pignoli, P.; Tremoli, E.; Poli, A.; Oreste, P.; Paoletti, R. 1986. Intimal plus medial thickness of the arterial wall: a direct measurement with ultrasound imaging. Circulation 74: 1399–1406.

Pingiotti, E.; Cipriani, P.; Marrelli, A.; Liakouli, V.; Fratini, S.; Penco, M.; et al. 2007. Surface Expression of Fractalkine Receptor (CX3CR1) on CD4+/CD28− T Cells in RA Patients and Correlation with Atherosclerotic Damage. Annals of the New York Academy of Sciences 1107: 32–41.

Prevoo, M.L.; van ’t Hof, M.A.; Kuper, H.H.; van Leeuwen, M.A.; van de Putte, L.B.; van Riel, P.L. 1995. Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis. Arthritis and rheumatism 38: 44–48.

Provan, S.A.; Olsen, I.C.; Austad, C.; Haugeberg, G.; Kvien, T.K.; Uhlig, T. 2017. Calcium supplementation and inflammation increase mortality in rheumatoid arthritis: A 15-year cohort study in 609 patients from the Oslo Rheumatoid Arthritis Register. Seminars in Arthritis and Rheumatism 46: 411–417.

53

Qi, H.; Yang, S.; Zhang, L. 2017. Neutrophil Extracellular Traps and Endothelial Dysfunction in Atherosclerosis and Thrombosis. Frontiers in Immunology 8: 928.

Rahbar, A.R.; Sundqvist, V. -a. S.; Wirgart, B.Z.; Grillner, L.; Söderberg‐Naucler, C. 2004. Recognition of cytomegalovirus clinical isolate antigens by sera from cytomegalovirus-negative blood donors. Transfusion 44: 1059–1066.

Rantapää-Dahlqvist, S.; Wållberg-Jonsson, S.; Dahlén, G. 1991. Lipoprotein (a), lipids, and lipoproteins in patients with rheumatoid arthritis. Annals of the Rheumatic Diseases 50: 366–368.

Rantapää-Dahlqvist, S.; de Jong, B.A.W.; Berglin, E.; Hallmans, G.; Wadell, G.; Stenlund, H.; et al. 2003. Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis. Arthritis and Rheumatism 48: 2741–2749.

Rau, R.; Herborn, G.; Karger, T.; Werdier, D. 1991. Retardation of radiologic progression in rheumatoid arthritis with methotrexate therapy. A controlled study. Arthritis and Rheumatism 34: 1236–1244.

Ridker, P.M.; Everett, B.M.; Thuren, T.; MacFadyen, J.G.; Chang, W.H.; Ballantyne, C.; et al. 2017. Antiinflammatory Therapy with Canakinumab for Atherosclerotic Disease. New England Journal of Medicine 377: 1119–1131.

Ridker, P.M.; Everett, B.M.; Pradhan, A.; MacFadyen, J.G.; Solomon, D.H.; Zaharris, E.; et al. 2018. Low-Dose Methotrexate for the Prevention of Atherosclerotic Events. New England Journal of Medicine.

Rochette, L.; Meloux, A.; Rigal, E.; Zeller, M.; Cottin, Y.; Vergely, C. 2018. The role of osteoprotegerin in the crosstalk between vessels and bone: Its potential utility as a marker of cardiometabolic diseases. Pharmacology & Therapeutics 182: 115–132.

Rodríguez-Lozano, B.; González-Febles, J.; Garnier-Rodríguez, J.L.; Dadlani, S.; Bustabad-Reyes, S.; Sanz, M.; et al. 2019. Association between severity of periodontitis and clinical activity in rheumatoid arthritis patients: a case–control study. Arthritis Research & Therapy 21: 27.

Rose, H.M.; Ragan, C. 1948. Differential agglutination of normal and sensitized sheep erythrocytes by sera of patients with rheumatoid arthritis. Proceedings of the Society for Experimental Biology and Medicine. Society for Experimental Biology and Medicine (New York, N.Y.) 68: 1–6.

Schmidt, D.; Goronzy, J.J.; Weyand, C.M. 1996. CD4+ CD7- CD28- T cells are expanded in rheumatoid arthritis and are characterized by autoreactivity. Journal of Clinical Investigation 97: 2027–2037.

Simon, A.; Chironi, G.; Levenson, J. 2007. Comparative performance of subclinical atherosclerosis tests in predicting coronary heart disease in asymptomatic individuals. European Heart Journal 28: 2967–2971.

Simonsson, M.; Bergman, S.; Jacobsson, L.T.; Petersson, I.F.; Svensson, B. 1999. The prevalence of rheumatoid arthritis in Sweden. Scandinavian Journal of Rheumatology 28: 340–343.

Smolen, J.S.; Landewé, R.; Bijlsma, J.; Burmester, G.; Chatzidionysiou, K.; Dougados, M.; et al. 2017. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological

54

disease-modifying antirheumatic drugs: 2016 update. Annals of the Rheumatic Diseases 76: 960–977.

Smolen, J.S.; Aletaha, D.; Barton, A.; Burmester, G.R.; Emery, P.; Firestein, G.S.; et al. 2018. Rheumatoid arthritis. Nature Reviews. Disease Primers 4: 18001.

Socialstyrelsen. 2017. Statistics and Data. Socialstyrelsen. (http://www.socialstyrelsen.se/statistics). Accessed on 23 Jan. 2017.

Solomon, D.H.; Love, T.J.; Canning, C.; Schneeweiss, S. 2010. The Risk of Diabetes Among Patients with Rheumatoid Arthritis, Psoriatic Arthritis, and Psoriasis. Annals of the rheumatic diseases 69: 2114–2117.

Solomon, D.H.; Greenberg, J.; Curtis, J.R.; Liu, M.; Farkouh, M.E.; Tsao, P.; et al. 2015. Derivation and Internal Validation of an Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis: A Consortium of Rheumatology Researchers of North America Registry Study. Arthritis & Rheumatology 67: 1995–2003.

Sorlie, P.D.; Nieto, F.J.; Adam, E.; Folsom, A.R.; Shahar, E.; Massing, M. 2000. A Prospective Study of Cytomegalovirus, Herpes Simplex Virus 1, and Coronary Heart Disease: The Atherosclerosis Risk in Communities (ARIC) Study. Archives of Internal Medicine 160: 2027–2032.

Steiner, G.; Urowitz, M.B. 2009. Lipid Profiles in Patients with Rheumatoid Arthritis: Mechanisms and the Impact of Treatment. Seminars in Arthritis and Rheumatism 38: 372–381.

Stoep, D.F. van B. der; Zeben, D. van; Klop, B.; Geijn, G.-J.M. van de; Janssen, H.J.W.; Meulen, N. van der; et al. 2016. Marked underdiagnosis and undertreatment of hypertension and hypercholesterolaemia in rheumatoid arthritis. Rheumatology 55: 1210–1216.

Stolt, P.; Yahya, A.; Bengtsson, C.; Källberg, H.; Rönnelid, J.; Lundberg, I.; et al. 2010. Silica exposure among male current smokers is associated with a high risk of developing ACPA-positive rheumatoid arthritis. Annals of the Rheumatic Diseases 69: 1072–1076.

Sugiyama, D.; Nishimura, K.; Tamaki, K.; Tsuji, G.; Nakazawa, T.; Morinobu, A.; et al. 2010. Impact of smoking as a risk factor for developing rheumatoid arthritis: a meta-analysis of observational studies. Annals of the Rheumatic Diseases 69: 70–81.

Sundqvist, V.A.; Wahren, B. 1982. Development of sensitive and enzyme-linked immunoassays for herpesvirus antigens and some applications. Developments in Biological Standardization 52: 245–253.

Svenson, K.L.; Pollare, T.; Lithell, H.; Hällgren, R. 1988. Impaired glucose handling in active rheumatoid arthritis: relationship to peripheral insulin resistance. Metabolism: Clinical and Experimental 37: 125–130.

Swedish society of rheumatology. 2018. Riktlinjer för läkemedelsbehandling vid reumatoid artrit. Svensk Reumatologisk Förening. (http://svenskreumatologi.se/grupper/arbetsgrupper/riktlinjer-for-lakemedelsbehandling-vid-ra/). Accessed on 20 Feb. 2019.

Symmons, D.P.M.; Bankhead, C.R.; Harrison, B.J.; Brennan, P.; Silman, A.J.; Barrett, E.M.; et al. 1997. Blood transfusion, smoking, and obesity as risk factors for the development of rheumatoid arthritis. Results from a primary care-based incident case-control study in Norfolk, England. Arthritis & Rheumatism 40: 1955–1961.

55

Tanaka, K.; Inaba, M.; Goto, H.; Nagata-Sakurai, M.; Sakai, S.; Yamada, S.; et al. 2006. Paraarticular trabecular bone loss at the ultradistal radius and increased arterial stiffening in postmenopausal patients with rheumatoid arthritis. The Journal of Rheumatology 33: 652–658.

Too, C.L.; Muhamad, N.A.; Ilar, A.; Padyukov, L.; Alfredsson, L.; Klareskog, L.; et al. 2016. Occupational exposure to textile dust increases the risk of rheumatoid arthritis: results from a Malaysian population-based case-control study. Annals of the Rheumatic Diseases 75: 997–1002.

Trygg, J.; Wold, S. 2002. Orthogonal projections to latent structures (O-PLS). Journal of Chemometrics 16: 119–128.

Tschiderer, L.; Willeit, J.; Schett, G.; Kiechl, S.; Willeit, P. 2017. Osteoprotegerin concentration and risk of cardiovascular outcomes in nine general population studies: Literature-based meta-analysis involving 26,442 participants. PloS One 12: e0183910.

Turesson, C.; O’Fallon, W.M.; Crowson, C.S.; Gabriel, S.E.; Matteson, E.L. 2003. Extra-articular disease manifestations in rheumatoid arthritis: incidence trends and risk factors over 46 years. Annals of the Rheumatic Diseases 62: 722–727.

Turk, S.A.; van Schaardenburg, D.; Boers, M.; de Boer, S.; Fokker, C.; Lems, W.F.; et al. 2018. An unfavorable body composition is common in early arthritis patients: A case control study. PLoS ONE 13.

den Uyl, D.; Nurmohamed, M.T.; van Tuyl, L.H.; Raterman, H.G.; Lems, W.F. 2011. (Sub)clinical cardiovascular disease is associated with increased bone loss and fracture risk; a systematic review of the association between cardiovascular disease and osteoporosis. Arthritis research & therapy 13: R5.

Vaidya, K.; Martínez, G.; Patel, S. 2019. The Role of Colchicine in Acute Coronary Syndromes. Clinical Therapeutics 41: 11–20.

Van Der Heijde, D.; Van Riel, P.; Gribnau, F.; Nuver-Zwart, I.; Van De Putte, L. 1989. EFFECTS OF HYDROXYCHLOROQUINE AND SULPHASALAZINE ON PROGRESSION OF JOINT DAMAGE IN RHEUMATOID ARTHRITIS. The Lancet 333: 1036–1038.

Vessey, M.P.; Villard-Mackintosh, L.; Yeates, D. 1987. Oral contraceptives, cigarette smoking and other factors in relation to arthritis. Contraception 35: 457–464.

Vik, A.; Mathiesen, E.B.; Brox, J.; Wilsgaard, T.; Njølstad, I.; Jørgensen, L.; et al. 2011. Serum osteoprotegerin is a predictor for incident cardiovascular disease and mortality in a general population: the Tromsø Study. Journal of thrombosis and haemostasis: JTH 9: 638–644.

Waaler, E. 1940. On the Occurrence of a Factor in Human Serum Activating the Specific Agglutination of Sheep Blood Corpuscles. Acta Pathologica Microbiologica Scandinavica 17: 172–188.

Wallberg-Jonsson, S.; Ohman, M.L.; Dahlqvist, S.R. 1997. Cardiovascular morbidity and mortality in patients with seropositive rheumatoid arthritis in Northern Sweden. The Journal of rheumatology 24: 445–451.

Wållberg-Jonsson, S.; Johansson, H.; Ohman, M.L.; Rantapää-Dahlqvist, S. 1999. Extent of inflammation predicts cardiovascular disease and overall

56

mortality in seropositive rheumatoid arthritis. A retrospective cohort study from disease onset. The Journal of rheumatology 26: 2562–2571.

Wållberg-Jonsson, S.; Cvetkovic, J.T.; Sundqvist, K.-G.; Lefvert, A.K.; Rantapää-Dahlqvist, S. 2002. Activation of the immune system and inflammatory activity in relation to markers of atherothrombotic disease and atherosclerosis in rheumatoid arthritis. The Journal of Rheumatology 29: 875–882.

Wållberg-Jonsson, S.; Backman, C.; Johnson, O.; Karp, K.; Lundström, E.; Sundqvist, K.G.; et al. 2001. Increased prevalence of atherosclerosis in patients with medium term rheumatoid arthritis. The Journal of rheumatology 28: 2597–2602.

Wang, P.; Li, S.; Liu, L.-N.; Lv, T.-T.; Li, X.-M.; Li, X.-P.; et al. 2017. Circulating osteoprotegerin levels are elevated in rheumatoid arthritis: a systematic review and meta-analysis. Clinical Rheumatology 36: 2193–2200.

Weinblatt, M.E.; Maier, A.L. 1990. Longterm experience with low dose weekly methotrexate in rheumatoid arthritis. The Journal of Rheumatology. Supplement 22: 33–38.

Westerlind, H.; Holmqvist, M.; Ljung, L.; Frisell, T.; Askling, J. 2019. Siblings of patients with rheumatoid arthritis are at increased risk of acute coronary syndrome. Annals of the Rheumatic Diseases.

Wilkins, L.W.&. 2002. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report. Circulation 106: 3143–3143.

Winchester, R.; Giles, J.T.; Nativ, S.; Downer, K.; Zhang, H.-Z.; Bag-Ozbek, A.; et al. 2016. Association of Elevations of Specific T Cell And Monocyte Subpopulations in Rheumatoid Arthritis with Subclinical Coronary Artery Atherosclerosis. Arthritis & rheumatology (Hoboken, N.J.) 68: 92–102.

van der Woude, D.; Houwing-Duistermaat, J.J.; Toes, R.E.M.; Huizinga, T.W.J.; Thomson, W.; Worthington, J.; et al. 2009. Quantitative heritability of anti-citrullinated protein antibody-positive and anti-citrullinated protein antibody-negative rheumatoid arthritis. Arthritis and Rheumatism 60: 916–923.

Yusuf, S.; Hawken, S.; Ôunpuu, S.; Dans, T.; Avezum, A.; Lanas, F.; et al. 2004. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. The Lancet 364: 937–952.

Zannettino, A.C.W.; Holding, C.A.; Diamond, P.; Atkins, G.J.; Kostakis, P.; Farrugia, A.; et al. 2005. Osteoprotegerin (OPG) is localized to the Weibel-Palade bodies of human vascular endothelial cells and is physically associated with von Willebrand factor. Journal of Cellular Physiology 204: 714–723.