Clin Rheumatol

DOI 10.1007/s10067-017-3551-7

BRIEF REPORT

Assessment of six cardiovascular risk calculators in Mexican

mestizo patients with rheumatoid arthritis according
to the EULAR 2015/2016 recommendations for cardiovascular
risk management
Dionicio A. Galarza-Delgado 1 & Jose R. Azpiri-Lopez 2 & Iris J. Colunga-Pedraza 3 &
Jesus A. Cardenas-de la Garza 1 & Raymundo Vera-Pineda 2 & Griselda Serna-Peña 1 &
Rosa I. Arvizu-Rivera 1 & Adrian Martinez-Moreno 2 & Martin Wah-Suarez 1 &
Mario A. Garza Elizondo 3

Received: 26 October 2016 / Revised: 12 December 2016 / Accepted: 16 January 2017

# International League of Associations for Rheumatology (ILAR) 2017

Abstract Variability of the 10-year cardiovascular (CV) risk
predicted by the Framingham Risk Score (FRS) using lipids,
FRS using body mass index (BMI), Reynolds Risk Score
(RRS), QRISK2, Extended Risk Score—Rheumatoid
Arthritis (ERS-RA), and algorithm developed by the
American College of Cardiology and the American Heart
Association in 2013 (ACC/AHA 2013) according to the
European League Against Rheumatism (EULAR)
2015/2016 update of its evidence-based recommendations
for cardiovascular risk management in patients with rheuma-
toid arthritis (RA) has not been evaluated in Mexican mestizo
patients. CV risk was predicted using six different risk calcu-
lators in 116 patients, aged 40–75, who fulfilled the ACR/
EULAR 2010 classification criteria. Results were multiplied
by 1.5 according to the EULAR 2015/2016 update. Global
comparison of the risk predicted by all scales was done using
the Friedman test, considering a P value of ≤0.05 as statisti-
cally significant. Individual comparison between the algo-
rithms was made using the Wilcoxon signed-rank test, and a
* Dionicio A. Galarza-Delgado
dgalarza@medicinauanl.mx
1
Internal Medicine Department, University Hospital Dr. Jose Eleuterio
Gonzalez UANL, Francisco I. Madero y Gonzalitos S/N,
64460 Monterrey, Nuevo Leon, Mexico
2
Internal Medicine Department, Cardiology Division, University
Hospital Dr. Jose Eleuterio Gonzalez UANL, Francisco I. Madero y
Gonzalitos S/N, 64460 Monterrey, Nuevo Leon, Mexico
3
Internal Medicine Department, Rheumatology Division, University
Hospital Dr. Jose Eleuterio Gonzalez UANL, Francisco I. Madero y
Gonzalitos S/N, 64460 Monterrey, Nuevo Leon, Mexico
P value of ≤0.003 was considered statistically significant. All
calculators showed to be different in the Friedman test
(p ≤ 0.001). Median values of predicted 10-year CV risk were
11.02% (6.18–17.55) for FRS BMI; 8.47% (4.6–13.16) for
FRS lipids; 5.55% (2.5–11.85) for QRISK2; 5% (3.1–8.65)
for ERS-RA; 3.6% (1.5–9.3) for ACC/AHA 2013; and 1.5%
(1.5–4.5) for RRS. ERS-RA showed no difference when com-
pared against QRISK2 (p = 0.269). CV risk calculators
showed variability among them and cannot be used indistinct-
ly in RA-patients.
Keywords Cardiovascular diseases . Rheumatoid arthritis .
Risk assessment
Introduction
Rheumatoid arthritis (RA) is a chronic, systemic, inflammato-
ry, and multifactorial disease that mainly affects the synovial
joints [1]. RA-patients have an increased risk of cardiovascu-
lar (CV) morbidity and mortality [2, 3], with atherosclerotic
CV disease being the leading cause of death [4]. The reason
behind this increased CV risk is complex, and traditional CV
risk factors do not fully explain it [5, 6]. Chronic systemic
inflammation and RA-specific factors such as disease dura-
tion, functional status, disease activity, seropositivity, extra-
articular manifestations, and RA-medication have also been
related to atherosclerotic CV disease [7].
The European League Against Rheumatism (EULAR)
2009 evidence-based recommendations for the management
of CV risk in RA-patients recommends an annual evaluation
of these patients according to national guidelines, as well as a

modification of the CV risk predicted by CV risk calculators,
in which a 1.5 factor should be added to the result if the patient
fulfilled two of the following three criteria: disease duration of
more than 10 years, rheumatoid factor (RF) or anti-cyclic
citrullinated peptide (anti-CCP) seropositivity, or presence of
certain extra-articular manifestations [2]. These recommenda-
tions were updated and recently published, establishing that
all CV calculators that do not include RA as a variable should
be multiplied by a 1.5 factor, regardless of patients’ disease
duration, serological profile, or extra-articular manifestations
[8].
Commonly used CV risk calculators such as Framingham
Risk Score (FRS) [9], Reynolds Risk Score (RRS) [10], and
QRISK2 [11] do not perform well in RA-patients [12, 13].
Therefore, it was deemed possible that an algorithm designed
specifically for RA could be more accurate [7, 12] Recently,
the Extended Risk Score—Rheumatoid Arthritis (ERS-RA)
cardiovascular risk calculator was developed and internally
validated using data from the Consortium of Rheumatology
Researchers of North America (CORRONA) registry in the
USA. This score incorporates several RA-specific factors such
as corticosteroid use, disease duration, clinical disease activity
index (CDAI), and the modified health assessment question-
naire (M-HAQ) [7].
There are no specific recommendations for the use of a
specific algorithm for the assessment of the CV risk for pa-
tients with RA in many countries, including Mexico, and
many of the calculators have not been externally validated
outside North American and European populations [14].
The objective of this study is to assess the variability of the
predicted 10-year CV risk in RA-patients according to the
2015/2016 update of the EULAR recommendations for car-
diovascular risk management in patients with RA [8] using six
different CV risk calculators: FRS using lipids, FRS using
body mass index (BMI), RRS, QRISK2, ERS-RA, and the
algorithm recommended by the 2013 ACC/AHA guideline
[15] on the assessment of CV risk.
Patients and methods
A cross-sectional study with prospective inclusion was de-
signed. Eligible criteria included patients aged 40–75 years
who fulfilled the 2010 ACR/EULAR criteria with at least
two visits to our referral center in the last year. Mexican mes-
tizo was defined as an individual born in Mexico, who has a
Spanish-derived last name and a family of Mexican ancestors
belonging to the third generation [16]. Exclusion criteria in-
cluded previous atherosclerotic cardiovascular disease
(ASCVD) defined as myocardial infarction, stroke and periph-
eral arterial disease, or pregnancy. Patients who withdrew con-
sent and those with missing data were eliminated. Those pa-
tients in which another connective tissue disease was
Clin Rheumatol
diagnosed at the time or after enrollment were excluded or
eliminated, respectively. The research protocol was approved
by our local ethics committee.
Patients’ data including age, gender, CV risk factors, med-
ical history associated with CV disease, current medication
and disease duration, BMI, blood pressure, and complete join
count were obtained at baseline. Hypertension, dyslipidemia,
and type 2 diabetes mellitus were defined as a previous diag-
nosis by a physician or the use of a related drug. Disease
activity was evaluated using the CDAI and the disease activity
score (DAS28), and disability using the M-HAQ. Presence of
severe extra-articular manifestations was defined according to
Mälmo criteria [17].
Lipid profile, erythrocyte sedimentation rate (ESR), and
high sensitivity C-reactive protein (hs-CRP) were performed
in all patients. Anti-CCP (IgG) antibodies were measured
using enzyme-linked immunosorbent assay (ELISA) kit
(Euroimmun, Lübeck, Germany), and values above 5 IU/ml
were considered positive. RF (IgM, IgG, IgA) was measured
using an ELISA kit (Euroimmun, Lübeck, Germany).
CV risk was calculated with FRS BMI, FRS lipids, RRS,
ACC/AHA 2013, QRISK2, and ERS-RA using the official
online sites. A multiplication factor of 1.5 was applied to the
predicted CV risk for all algorithms that do not include RA
diagnosis as a variable according to the EULAR 2015/2016
update on evidence-based recommendations for CV risk man-
agement in patients with RA.
Patients were classified into risk categories according to
their 10-year CV risk using each calculator threshold. Using
FRS (both lipids and BMI), patients were classified as low risk
(<10%), moderate risk (10–20%), and high risk (>20%). For
ACC/AHA 2013 algorithm, high risk was defined as a 10-year
CV risk of ≥7.5%, using QRISK2, ≥20%, and 20% for RRS.
A threshold has not been established for ERS-RA. Description
of quantitative variables was carried out according to normal-
ity, using mean ± standard deviation, or median, 25th and 75th
percentile in each case. A Kolmogorov-Smirnov test was used
for this purpose with a p value ≤0.05 determined as statistical-
ly significant. Analysis of variance was performed between all
scores using the Friedman test. For the individual comparison
of variance among each score, the Wilcoxon signed-rank test
was used, with a p value of ≤0.003 considered as statistically
significant because of a correction for repeated measurements.
Unless otherwise mentioned, a standard p value of ≤0.05 was
chosen for all other tests. The complete statistical analysis was
done using SPSS version 21 (IBM, NY, USA).
Results
Of a total of 128 eligible patients, 116 were included in the
final analysis. Twelve patients were eliminated: 9 because
missing laboratory data and 3 because a second connective
Clin Rheumatol

tissue disease was diagnosed after enrollment. Consent was
not withdrawn by any patient. The population’s demographi-
cal and clinical characteristics are shown in Table 1. Women
accounted for the majority of our population (108 (93.1%)).
Mean age was 56.0 ± 8.7 years, with mean disease duration of
12.0 ± 7.9 years. Regarding cardiovascular risk factors, 10
(8.6%) were active smokers, 11 (9.5%) had family history of
early coronary heart disease, 38 (32.8%) had hypertension, 32
(27.6%) had dyslipidemia, and 14 (12.1%) had type 2 diabetes
mellitus (Table 1). Presence of atrial fibrillation and chronic
kidney disease was not reported in any patient.
Patients presented a mean DAS 28-CRP of 3.16 ± 1.38 and
a median CDAI of 10 (3–19). Severe extra-articular manifes-
tations were found in only 3 patients (2.6%). IgM-RF sero-
positivity was found in 101 (87.0%), IgA-RF in 77 (68.3%),
IgG-RF in 30 (25.8%), anti-CCP antibodies in 85 (73.2%),
and double seropositivity (anti-CCP antibodies and any RF
isotype) was found in 82 (70.6%).
Lipid profile values showed mean total cholesterol (TC) of
182.54 ± 30.59 mg/dl; low-density lipoprotein cholesterol
(LDL-C), 97.90 ± 26.03 mg/dl; high-density lipoprotein cho-
lesterol (HDL-C), 51.40 (44.12–63.15) mg/dl; and TC/HDL-
C ratio 3.58 ± 1.04.
Table 1 Demographical and clinical characteristics
Characteristic Value
Current medication is described in Table 2. A total of 96
(82.6%) patients were treated with methotrexate (MTX) with
a median dose of 20 mg per week. Use of biologic disease-
modifying antirheumatic drugs (bDMARD) was reported in
16 (13.8%) patients. Daily use of prednisone was documented
in 67 subjects (57.8%) with a median dose of 5 mg.
All CV risk scores are shown in Table 3. Median values of
predicted 10-year CV risk were 1.5% (1.5–4.5) for RRS; 3.6%
(1.5–9.3) for ACC/AHA 2013; 5% (3.1–8.65) for ERS-RA;
5.5% (2.35–10.47) for QRISK2; 8.47% (4.6–13.16) for FRS
lipids; and 11.02% (6.18–17.55) for FRS BMI.
According to ACC/AHA 2013, 27 (23.2%) patients were
classified as high risk. In our cohort, FRS lipids stratified 34
(29.3%) at moderate risk and 6 (5.1%) at high risk, while FRS
BMI rated 36 (31.0%) as moderate risk and 19 (16.3%) as
high risk. QRISK2 classified 11 (9.4%) of our patients as high
risk. RRS ranked 3 (2.5%) patients as high risk; moreover, 6
(5.1%) were stratified as moderate risk. ERS-RA does not
have a risk classification. Patients classified into the high risk
category are eligible to receive statin therapy.
Global analysis of variance using the Friedman test showed
a significant difference among them (p ≤ .001). A secondary
comparison between each algorithm with the Wilcoxon
signed-rank test delivered the FRS BMI as the CV risk score
with the highest values (p < 0.001). ERS-RA showed no dif-
ference when compared to QRISK2 (p = 0.269). RRS

Age (years), mean ± SD 56.0 ± 8.7 Table 2 Drug treatment

Women, n (%) 108 (93.1) Drug Value
Family history of ASCVD, n (%) 11 (9.5)
Active smoker, n (%) 10 (8.6) MTX, n (%) 96 (82.6)
Dyslipidemia, n (%) 32 (27.6) Weekly dose of MTX (mg), median (p25–p75) 20 (15–25)
Type 2 diabetes mellitus, n (%) 14 (12.1) Leflunomide, n (%) 15 (12.9)
Antihypertensive treatment, n (%) 38 (32.8) Sulfasalazine, n (%) 17 (14.7)
Extra-articular manifestations, n (%) 3 (2.6) Prednisone, n (%) 67 (57.8)
Disease duration (years), mean ± SD 12.03 ± 7.96 Daily dose of prednisone (mg), median (p25–p75) 5 (5–10)
BMI, mean ± SD 28.08 ± 4.97 Chloroquine, n (%) 21 (18.1)
Systolic blood pressure (mmHg), median 119 (110–130) Hydroxychloroquine, n (%) 19 (16.4)
(p25–p75) bDMARD, n (%) 16 (13.8)
DAS 28 CRP, mean ± SD 3.16 ± 1.38
ARB II, n (%) 17 (14.7)
CDAI, median (p25–p75) 10 (3–19)
ACEI, n (%) 14 (12.1)
M-HAQ, median (p25–p75) 0.125 (0–0.59)
Beta-blockers, n (%) 6 (5.2)
Total cholesterol (mg/dl), mean ± SD 182.54 ± 30.59
CCB, n (%) 4 (3.4)
HDL-C (mg/dl), median (p25–p75) 51.40 (44.12–63.15)
Diuretics, n (%) 2 (1.7)
LDL-C (mg/dl), mean ± SD 97.90 ± 26.03
Statins, n (%) 17 (14.7)
Total cholesterol/HDL-C index, mean ± SD 3.58 ± 1.04
Metformin, n (%) 12 (10.3)
hs-CRP (mg/dl), mean ± SD 1.12 ± 0.83
Sulfonylureas, n (%) 3 (2.6)
SD standard deviation, ASCVD atherosclerotic cardiovascular disease, Insulin, n (%) 3 (2.6)
BMI body mass index, DAS 28 disease activity score calculator for 28 Aspirin, n (%) 2 (1.7)
joints, CDAI clinical disease activity index, M-HAQ modified health as-
sessment questionarie. HDL-C high-density lipoprotein—cholesterol, MTX methotrexate, bDMARD biological disease-modifying antirheumat-
LDL-C low-density lipoprotein—cholesterol, hs-CRP high sensitive C- ic drugs, ARB II angiotensin receptor blocker, ACEI angiotensin
reactive protein converting enzyme inhibitor, CCB calcium channel blockers
 Clin Rheumatol

Table 3 Results of the Wilcoxon signed-rank test

Score ACC/AHA 2013 FRS lipids FRS BMI QRISK2 RRS ERS-RA
median (p 25–p 75) 3.6 (1.5–9.3) 8.47 (4.65–13.16) 11.02 (6.18–17.55) 5.5 (2.35–10.47) 1.5 (1.5–4.5) 5 (2.35–10.47)

ACC/AHA 2013
3.6 (1.5–9.3) – p < .001 p < .001 p < .001 p < .001 p = .003
FRS lipids
8.47 (4.65–13.16) p < .001 – p < .001 p < .001 p < .001 p < .001
FRS BMI
11.02 (6.18–17.55) p < .001 p < .001 – p < .001 p < .001 p < .001
QRISK2
5.5 (2.35–10.47) p < .001 p < .001 p < .001 – p < .001 p = .269
RRS
1.5 (1.5–4.5) p < .001 p < .001 p < .001 p < .001 – p < .001
ERS-RA
5 (2.35–10.47) p = .003 p < .001 p < .001 p = .269 p < .001 –

FRS Framingham risk score, BMI body mass index, RRS Reynolds risk score, ERS-RA extended risk score —rheumatoid arthritis

predicted the lowest CV risk assessment when compared to
the rest of the scales (p < 0.001 for each comparison).
Discussion
In our cohort of Mexican mestizo patients with RA, global
comparison of the CV risk predicted by these 6 CV risk cal-
culators resulted in a significant difference among them.
Clinically, this could have an impact in the therapy and pre-
vention strategies chosen because a patient who is regarded as
low risk by one calculator could be classified as high risk by
another, and vice versa.
Individually, there was no statistical difference when ERS-
RAwas compared to QRISK2, the only CV risk calculator that
includes RA as a variable and which adds a 1.4 multiplication
factor [12]. This is an interesting finding for several reasons.
The ERS-RA was designed from the CORRONA registry,
including data of 23,605 RA-patients, in which information
regarding blood pressure, lipid profile, and glycosylated he-
moglobin levels was lacking, and variables such as hyperten-
sion, dyslipidemia, and diabetes mellitus were dichotomized
and determined only as present or absent on the basis of a
previous diagnosis or the use of medication for these diseases
related to these conditions. This could lead to an inaccurate
prediction of CV risk in patients with hyperglycemia, high
blood pressure, or altered lipid profile without a formal med-
ical diagnosis or treatment [7]. However, the CV risk calculat-
ed by ERS-RA did not differ from that predicted by QRISK2,
which includes the TC/HDL-C ratio and systolic blood pres-
sure as quantitative variables. Because ERS-RA has not been
externally validated, it is not known how well it performs in
RA-patients, making a comparison to previous results impos-
sible. Therefore, our results need to be proven in a larger
population with registered CV outcomes. These findings raise
a question—given that the ERS-RA does not require labora-
tory studies for the evaluation—could it be an easier and more
affordable way to evaluate CV risk in RA-patients in an out-
patient setting? Further studies on the economical and clinical
impact, which can be relevant in especially emerging econo-
mies, evaluating the effect that the use of the newly developed
ERS-RA would have in the management of patients with RA
are needed.
FRS using BMI does not include lipid values in its vari-
ables, and in our patients, it predicted the highest values of CV
risk. This could be explained by the lipid paradox in RA-
patients, in which high levels of systemic inflammation are
linked to low levels of LDL-C and an increased CV risk [6,
18]. However, there is no data on the performance of this
algorithm in RA-patients, and the accuracy of its predictions
is yet to be assessed. Characteristics of the six CV risks eval-
uated in this study are shown in Table 4.
Arts et al. reported an overestimation of the CV risk by
QRISK2 in European patients with RA [12], which differs
from our results. FRS was derived using data of a population
in the USA, and its performance has been validated in the
general population in Mexico [19]. However, QRISK2 was
validated from data of populations in England and Wales, and
data on its performance in our population is nonexistent.
Although the differences between populations could be an
explanation for our results, a prospective study is needed to
truly evaluate the 10-year accuracy of these calculators and to
corroborate or to refute our findings.
Our study includes a large number of women, who account
for the 93.1% of our population. In the general population, CV
risk factors are the same in women and men, though their
prevalence varies among sexes. While prevalence of hyper-
tension is higher in women older than 65 years and type 2
Clin Rheumatol

Table 4 Characteristics of CV
risk calculators CV risk Characteristics
calculator

ACC/AHA This algorithm predicts the 10-year risk of a coronary death or nonfatal myocardial infarction,
2013 and fatal or nonfatal stroke, in patients aged 20 to 79.
Gender, age, race, diagnosis of diabetes, smoking, treated hypertension, systolic blood
pressure, and lipid profile values (TC and HDL-C) are included as variables [15].
In RA-patients, systemic inflammation leads to a decrease of TC, LDL-C, and HDL-C (the
lipid paradox), making the relationship between CV risk and lipid profile less clear [18]
ERS-RA Four RA-specific factors are included as variables: disease duration, disease activity,
disability, and use of corticoids.
It was developed and internally validated from data in the CORRONA registry, but it has not
been externally validated [7].
Laboratory data is not needed.
A threshold for high risk has not been established.
There is no need to multiply by 1.5 since it is a specific calculator for RA-patients.
FRS BMI This algorithm predicts the 10-year risk of coronary death, myocardial infarction, coronary
insufficiency, angina, ischemic and haemorrhagic stroke, transient ischemic attack,
peripheral artery disease, and heart failure.
Variables include age, gender, and traditional CV risk factors (presence of diabetes, smoking,
and treated hypertension), as well as values of systolic blood pressure and BMI.
CV risk can be calculated for individuals aged 30 to 74 [9].
Because this algorithm does not include laboratory data, it could be a quick and cost-effective
tool for the evaluation of CV risk.
FRS lipids Lipid values replace BMI in this algorithm. The relationship between CV risk and lipid profile
is less clear because of the lipid paradox [18].
QRISK2 This algorithm predicts the 10-year risk of angina, myocardial infarction or stroke in patients
aged 25 to 84.
RA, family history of angina or heart attack, atrial fibrillation, chronic kidney disease and type
1 diabetes mellitus are included among its variables [11].
There is no need to apply a 1.5 factor according to the EULAR recommendations because RA
is included as a variable.
RRS CV risk is calculated for patients aged 45 to 80.
Among its variables, this algorithm includes high sensitivity C-reactive protein and family
history of a heart attack in youngers than 60 years old.
Age, gender, systolic blood pressure, and values of TC and HDL-C are also included (which
can make the relationship between CV risk and lipid profile less clear in RA-patients).
Outcomes include stroke, myocardial infarction, revascularization, and death [11, 19].

diabetes is more prevalent in women older than 20, men have
a higher prevalence of smoking. For both women and men,
coronary heart disease (CHD) is the largest contributor to
CVD morbidity and mortality. More women die because of
stroke and heart failure than men, while mortality of CHD is
higher in men. However, the mortality rate from CVD is over-
all substantially higher in men [20].
Patients with carotid plaque detected by ultrasound are con-
sidered at high risk by the European Society of Cardiology [21],
and studies using its detection as surrogate marker for subclin-
ical atherosclerosis have shown that the cut off points that CV
risk scores use to define the high-risk strata are not accurate in
RA-patients (for example, 7.3–10.8% instead of 20% for FRS)
[22] and inappropriately high. The update of the EULAR rec-
ommendations could improve the predictive accuracy of CV
risk calculators, but it might not be enough to reach the high-
risk threshold in patients with carotid plaque.
Statin therapy is recommended when a patient is classified
as high risk according to each algorithm threshold. In our
patients, ACC/AHA 2013 classified the larger number of pa-
tients into the high-risk category (23.1%), a larger number of
patients in comparison to the 14.7% under statin therapy. As
stated before, surrogate markers of subclinical atherosclerosis
have been found in patients who would be classified as low
and moderate risk using FRS, and patients who would benefit
of therapy with statins could be left untreated with such
thresholds for high risk. This could mean that use of image
techniques such as carotid ultrasound could be an appropriate
and more precise tool to evaluate subclinical atherosclerosis in
RA-patients and to decide the initiation of therapy with statins.

Some strengths of this study should be highlighted. This is
the first study to analyze the variability of 6 CV risk scores
using the 2015/2016 update of the EULAR recommendations
for cardiovascular risk management in RA, and it also incor-
porates the new ERS-RA and ACC/AHA 2013 algorithms,
which had not been compared in RA-patients. Furthermore,
this is the first study to assess CV risk scores in Mexican
mestizo RA-patients, a population widely underrepresented
in clinical trials, and our results could be extrapolated to other
Latin American populations.
However, some limitations are worth noting. First, because
our study is cross-sectional and lacking of CV outcomes, the
evaluation of the accuracy of the predicted risk and recom-
mendation for the use of one particular CV calculator is not
possible. Follow-up of these patients is needed. Second, gen-
eralizability of our results may be limited because all of our
patients were recruited from one single center. An additional
limitation of this analysis is the small number of patients
included.
In conclusion, in our cohort of Mexican mestizo RA-pa-
tients, the global comparison of CV risk calculators showed a
significant difference among them. In the individual compar-
ison, QRISK2 did not show a statistical difference when it was
compared to ERS-RA. In the current study, FRS BMI deliv-
ered the highest values of predicted CV risk. Prospective stud-
ies with CV outcomes using the 2015/2016 update of the
EULAR recommendations are needed to evaluate if there is
an improvement in the predictive accuracy of the standard CV
risk calculators.
Compliance with ethical standards
The research protocol was approved by our local ethics committee.
Consent was not withdrawn by any patient.
Disclosures None.
References
1. You S, Cho CS, Lee I, Hood L, Hwang D, Kim WU (2012) A
systems approach to rheumatoid arthritis. PLoS One 7:e51508
2. Peters M, Symmons D, McCarey D, Dijkmans B, Nicola P,
Kvien T et al (2010) EULAR evidence-based recommenda-
tions for cardiovascular risk management in patients with
rheumatoid arthritis and other forms of inflammatory arthritis.
Ann Rheum Dis 69:325–331
3. Pieringer H, Pichler M (2011) Cardiovascular morbidity and mor-
tality in patients with rheumatoid arthritis: vascular alterations and
possible clinical implications. QJM 104:13–26
4. Sarmiento-Monroy JC, Amaya-Amaya J, Espinosa-Serna JS,
Herrera-Díaz C, Anaya J-M, Rojas-Villarraga A (2012)
Cardiovascular disease in rheumatoid arthritis: a systematic litera-
ture review in Latin America. Arthritis 2012:371909
5. Barber CE, Smith A, Esdaile JM, Barnabe C, Martin LO, Faris P
et al (2015) Best practices for cardiovascular disease prevention in
Clin Rheumatol
rheumatoid arthritis: a systematic review of guideline recommen-
dations and quality indicators. Arthritis Care Res (Hoboken) 67:
169–179
6. Choy E, Ganeshalingam K, Semb AG, Szekanecz Z, Nurmohamed
M (2014) Cardiovascular risk in rheumatoid arthritis: recent ad-
vances in the understanding of the pivotal role of inflammation,
risk predictors and the impact of treatment. Rheumatology
(Oxford) 53:2143–2154
7. Solomon DH, Greenberg J, Curtis JR, Liu M, Farkouh M, Tsao P
et al (2015) Derivation and internal validation of an expanded car-
diovascular risk prediction score for rheumatoid arthritis: a consor-
tium of rheumatology researchers of North America registry study.
Arthritis Rheumatol 67:1995–2003
8. Agca R, Heslinga S, Rollefstad S, Heslinga M, McInnes I,
Peters M, et al. (2016) EULAR recommendations for cardio-
vascular disease risk management in patients with rheumatoid
arthritis and other forms of inflammatory joint disorders:
2015/2016 update. Ann Rheum Dis [journal on the internet].
2013 Oct 3[date cited 2013 Dec 7]. doi:10.1136/annrheumdis-
2016-209775
9. D'Agostino RB Sr, Vasan RS, Pencina MJ, Wolf PA, Cobain M,
Massaro JM et al (2008) General cardiovascular risk profile for use
in primary care: the Framingham heart study. Circulation 117:743–753
10. Ridker PM, Buring JE, Rifai N, Cook NR (2007) Development and
validation of improved algorithms for the assessment of global car-
diovascular risk in women: the Reynolds risk score. JAMA 297:
611–619
11. Hippisley-Cox J, Coupland C, Vinogradova Y, Robson J, Minhas
R, Sheikh A et al (2008) Predicting cardiovascular risk in England
and Wales: prospective derivation and validation of QRISK2. BMJ
336:1475–1482
12. Arts E, Popa C, Den Broeder A, Semb A, Toms T, Kitas G et al
(2014) Performance of four current risk algorithms in predicting
cardiovascular events in patients with early rheumatoid arthritis.
Ann Rheum Dis 74:668–674
13. Crowson CS, Matteson EL, Roger VL, Therneau TM, Gabriel SE
(2012) Usefulness of risk scores to estimate the risk of cardiovas-
cular disease in patients with rheumatoid arthritis. Am J Cardiol
110:420–424
14. Damen JA, Hooft L, Schuit E, Debray TP, Collins GS, Tzoulaki I
et al (2016) Prediction models for cardiovascular disease risk in the
general population: systematic review. BMJ 353:i2416
15. Goff DC Jr, Lloyd-Jones DM, Bennett G, Coady S, D'Agostino RB,
Gibbons R et al (2014) 2013 ACC/AHA guideline on the assess-
ment of cardiovascular risk: a report of the American College of
Cardiology/American Heart Association task force on practice
guidelines. Circulation 129:S49–S73
16. Guardado-Estrada M, Juarez-Torres E, Medina-Martinez I, Wegier
A, Macías A, Gomez G et al (2009) A great diversity of Amerindian
mitochondrial DNA ancestry is present in the Mexican mestizo
population. J Hum Genet 54:695–705
17. Turesson C, Jacobsson L, Bergström U (1999) Extra-articular rheu-
matoid arthritis: prevalence and mortality. Rheumatology (Oxford)
38:668–674
18. González-Gay MA, González-Juanatey C (2014) Inflammation and
lipid profile in rheumatoid arthritis: bridging an apparent paradox.
Ann Rheum Dis 73:1281–1283
19. Alcocer LA, Lozada O, Fanghänel G, Sánchez-Reyes L,
Campos-Franco E (2011) Estratificación del riesgo cardiovas-
cular global. Comparación de los métodos Framingham y
SCORE en población mexicana del estudio PRIT. Cir Cir 79:
168–174
20. Mosca L, Barrett-Connor E, Kass Wenger N (2011) Sex/gender
differences in cardiovascular disease prevention: what a difference
a decade makes. Circulation 124:2145–2154
Clin Rheumatol
21. Catapano AL, Reiner Z, De Backer G (2011) European Atherosclerosis
Society (EAS). ESC/EAS guidelines for the management of
dyslipidaemias the task force for the management of dyslipidaemias
of the European Society of Cardiology (ESC) and the European
Atherosclerosis Society (EAS). Atherosclerosis 217:3–46
22. Dessein PH, Corrales A, Lopez-Mejias R, Solomon A, Woodiwiss
AJ, Llorca J et al (2016) The Framingham score and the systematic
coronary risk evaluation at low cutoff values are useful surrogate
markers of high-risk subclinical atherosclerosis in patients with
rheumatoid arthritis. J Rheumatol 43:486–494