Drs. Gonzalez-Gay and Llorca share senior authorship.
HLA–DRB1 and persistent chronic inflammation contribute to cardiovascular events and cardiovascular mortality in patients with rheumatoid arthritis
Article first published online: 31 JAN 2007
Copyright © 2007 by the American College of Rheumatology
Arthritis Care & Research
Volume 57, Issue 1, pages 125–132, 15 February 2007
How to Cite
Gonzalez-Gay, M. A., Gonzalez-Juanatey, C., Lopez-Diaz, M. J., Piñeiro, A., Garcia-Porrua, C., Miranda-Filloy, J. A., Ollier, W. E. R., Martin, J. and Llorca, J. (2007), HLA–DRB1 and persistent chronic inflammation contribute to cardiovascular events and cardiovascular mortality in patients with rheumatoid arthritis. Arthritis & Rheumatism, 57: 125–132. doi: 10.1002/art.22482
- Issue published online: 31 JAN 2007
- Article first published online: 31 JAN 2007
- Fondo de Investigaciones Sanitarias. Grant Number: PI06-0024 (Spain)
- Rheumatoid arthritis;
- Cardiovascular disease;
- Cardiovascular event;
Cardiovascular (CV) disease is the most common cause of mortality in patients with rheumatoid arthritis (RA). We assessed the contribution of epidemiologic features, clinical features, routine laboratory markers of inflammation, and HLA–DRB1 alleles to CV mortality in patients with RA prospectively followed at a single referral center in Spain.
Patients fulfilling the 1987 American College of Rheumatology classification criteria for RA seen at the rheumatology outpatient clinic of Hospital Xeral-Calde, Lugo between March and September 1996 were included. HLA–DRB1 phenotype, epidemiologic data, and clinical data were assessed at that time. Patients were prospectively followed and clinical records were examined until patient's death or September 1, 2005.
A total of 182 consecutive patients were assessed. Compared with the general Spanish population, the age- and sex-standardized mortality ratio by CV cause was 1.78. CV mortality adjusted by age at disease onset and sex was associated with chronic inflammation determined by C-reactive protein level (CRP; hazard ratio [HR] 1.14, P < 0.001) and erythrocyte sedimentation rate (ESR; HR 1.05, P = 0.003). Patients with HLA–DRB1*04 shared epitope alleles (HR 4.15, P = 0.030), in particular those HLA–DRB1*0404 positive (HR 6.65, P = 0.002), had increased risk of CV mortality. Increased risk of CV events was also associated with CRP level (HR 1.09, P = 0.001), ESR (HR 1.03, P = 0.003), and HLA–DRB1*0404 (HR 4.47, P = 0.002).
Our results suggest that a chronically high inflammatory response in genetically predisposed individuals promotes an increased risk of CV events and CV mortality in RA.
Several studies have disclosed a high morbidity and mortality rate due to cardiovascular (CV) disease in patients with rheumatoid arthritis (RA) (1–7). Shared inflammatory mechanisms responsible for both synovial lesions in RA and atherosclerotic disease may also play a role in promoting accelerated atherosclerosis in patients with RA (8). Although in some cases traditional risk factors known to promote the progression of atherosclerotic lesions may be absent in patients with RA (6), accelerated atherosclerosis has been observed in these patients (9–12). High-resolution B-mode ultrasound of the common carotid artery, a surrogate marker of atherosclerotic disease, disclosed increased carotid artery intima-media thickness (IMT) and carotid plaques in patients with RA without classic atherosclerosis risk factors compared with matched controls (13). A significant linear trend for increased carotid IMT associated with increasing C-reactive protein (CRP) levels was also reported in these patients (14).
Genetic and environmental factors are believed to contribute to RA, with the genetic component accounting for up to 60% of disease susceptibility (15). The major genetic contribution is provided by genes of the HLA region, particularly the major histocompatibility complex class II HLA–DRB1 alleles. Several HLA–DRB1 alleles (HLA–DRB1*0401, *0404, *0405, *0408, *0101, *0102, 1001, and 1402) are associated with susceptibility to RA. These alleles encode a conserved amino acid sequence (QKRAA, QRRAA, or RRRAA), called the shared epitope, at position 70–74 in the third hypervariable region of the HLA–DRβ1 molecule (16). A number of studies have shown an association between HLA–DRB1*04 shared epitope alleles and extraarticular disease in patients with RA from different populations (17–20). Also, regarding medication, patients with RA from northwest Spain requiring treatment with cyclosporin A because of severe disease and incomplete response to methotrexate (MTX) were significantly more likely to have HLA–DRB1*04 shared epitope alleles than patients not requiring such treatment (21). In the same population, endothelial dysfunction, an early step in the development of atherosclerosis, was more commonly observed in patients with RA positive for HLA–DRB1*04 shared epitope who had long disease duration, particularly in those carrying the HLA–DRB1*0404 allele (22, 23). Taking all these considerations into account, in the present study we sought to determine the potential contribution of epidemiologic features, clinical features, routine laboratory markers of inflammation, and HLA–DRB1 alleles to CV mortality and CV events in a cohort of patients with RA who were prospectively followed at the referral center for an area of northwest Spain.
PATIENTS AND METHODS
Consecutive unselected patients who fulfilled the 1987 American College of Rheumatology (ACR; formerly the American Rheumatism Association) classification criteria for RA (24) who were seen at the rheumatology outpatient clinic of Hospital Xeral-Calde, Lugo (northwestern Spain) between March and September 1996 were included in this study. This hospital is the single referral center for CV and rheumatic diseases for a defined population of central Galicia in northwestern Spain (15). Epidemiologic and clinical data were assessed at that time. Also, blood samples were collected for HLA–DRB1 typing. All patients were prospectively followed and their clinical records were reviewed until the patient's death or September 1, 2005.
At the time of recruitment, patients' data regarding traditional CV risk factors (hypertension, diabetes mellitus, dyslipidemia, obesity, and smoking history); history of CV events; clinical manifestations, including data on tender and swollen joints and extraarticular manifestations of the disease (nodular disease, Felty's syndrome, pulmonary fibrosis, rheumatoid vasculitis, or secondary Sjögren's syndrome) (13); and radiologic and laboratory features (rheumatoid factor, glucose, lipid profile, CRP level, erythrocyte sedimentation rate [ESR]) were registered. Also, available information from the time of disease diagnosis on comorbidities, joint involvement, extraarticular manifestations, rheumatoid factor, and laboratory tests (glucose, lipid profile, CRP level, and ESR) was reviewed at that time. Then patients were followed and assessed every 3–6 months. All changes in RA medication were noted throughout followup, and patients' medical records were screened for comorbidities.
Information on CV events and cause of death over the extended followup was also registered. Plain radiographs of the hands and feet were obtained at least every 3 years. Larsen radiographic score (22) and HLA–DRB1 typing (22, 23, 25) were performed as previously reported.
Because in most cases information registered by a family physician on the onset of RA symptoms was insufficient or not available, the onset of the disease was considered as the time when a patient began to report pain and inflammation involving joints of the hands and/or the feet in a symmetrical fashion. Also, pain had to be worse in the morning, associated with morning stiffness, and improve with use of the affected joints. The time of disease diagnosis was established as the time in which a patient was diagnosed as having RA by an experienced physician, generally a rheumatologist, according to the 1987 ACR classification criteria (24).
Information about cholesterol, triglycerides and glucose levels, blood pressure, and smoking history was available in all cases from the time of disease diagnosis. Patients were considered to have dyslipidemia if they had hypercholesterolemia and/or hypertriglyceridemia (defined as diagnosis of hypercholesterolemia or hypertriglyceridemia by the patients' family physicians prior to the diagnosis of RA, or total cholesterol and/or triglyceride levels in fasting plasma >240 mg/dl and 160 mg/dl, respectively, at the time of disease diagnosis or over the extended followup). Patients were considered to have hypertension if before the diagnosis of RA they had been diagnosed as having hypertension by their family physicians, or if at the time of disease diagnosis or over the extended followup they had blood pressure >150/90 mm Hg in 2 different examinations performed on different days. Patients were considered to have diabetes mellitus if before disease diagnosis they had been diagnosed as having diabetes mellitus by their family physicians or if 2 fasting plasma glucose levels on different days at the time of disease diagnosis or over the extended followup were >125 mg/dl. Obesity was defined as body mass index (calculated as weight in kg divided by height in m2) >30 kg/m2 at enrollment or over the extended followup. As previously reported (26), smoking history was treated as a dichotomous variable (heavy versus nonheavy smoking) in this analysis. Heavy smokers comprised patients who smoked at the time of disease diagnosis or during the followup or who had smoked within the 10 years before the onset of RA symptoms or the disease diagnosis, and the remaining patients (nonheavy smokers) were those who had never smoked or had stopped smoking at least 10 years before the disease onset. For the purpose of this study, CV risk factors were measured at enrollment in 1996 and then monitored over time.
Extraarticular manifestations were considered to be present if they were found at any point from the time of disease diagnosis until the patient's death or September 1, 2005. The definition of ischemic heart disease (IHD) included acute coronary syndromes with or without persistent ST-segment elevation and chronic coronary heart disease. IHD was diagnosed if any of the following criteria were satisfied: a recorded diagnosis of ischemic cardiopathy on account of some acute coronary syndrome (acute myocardial infarction or unstable angina), the presence of pathologic Q waves in the electrocardiogram, and coronary images showing >50% stenosis of at least 1 coronary vessel. Ischemic dilated cardiomyopathy was also included in this category if the patient had deteriorated systolic function and a dilated left ventricle with evidence of ischemic cardiopathy by echocardiographic and/or catheterization studies (27).
A patient was considered to have a cerebrovascular accident when he or she had a stroke and/or transient ischemic attacks (TIAs). Strokes were classified according to their clinical features and were confirmed by computed tomography and/or magnetic resonance imaging. TIAs were diagnosed if the symptoms were self limited in <24 hours, without residual neurologic damage (28, 29).
Because Hospital Xeral-Calde is the referral center for the population of the Lugo area, the first CV event was defined as an event (case) of CV complication diagnosed at the hospital in a patient without a previous history of CV disease. Specific information on CV events and CV death was collected based on patients' medical records. For attributing death to a CV cause, 2 clinicians independently reviewed clinical records of patients with CV events; discrepancies were resolved by consensus. Informed consent was obtained from all patients. The local institutional committee approved the study.
A separate Cox regression model was used to estimate the influence of each potential predictor on CV mortality; patients who were alive at the end of followup and patients who died by causes other than CV events were considered as censored. Similar models were obtained for predicting the first CV event. To perform this analysis, patients who died by any other cause different from CV events and patients who were alive who had not had any CV event at the end of the followup were considered as censored. Proportional hazard assumption was tested using Schoenfeld residuals. Results were expressed as hazard ratios (HRs) with 95% confidence intervals (95% CIs) and were computed as both crude and age- and sex-adjusted for each potential predictor. Two-sided P values less than 0.05 were considered to indicate statistical significance.
CV mortality in this cohort was age- and sex-adjusted (indirect method) using the Spanish CV mortality data as a reference. Spanish data were provided by the National Institute for Statistics (Instituto Nacional de Estadística; available at http://www.ine.es). All statistical tests were performed with Stata, version 8/SE (StataCorp, College Station, TX).
A total of 182 consecutive patients were assessed in this study. Of these, HLA–DRB1 phenotype information was available for 178 (98%). At the time of the onset of this study (period of enrollment) in 1996, 174 (96%) were receiving disease-modifying antirheumatic drugs (DMARDs). At that time, MTX (median dosage 15 mg/week) with or without chloroquine was the most common standard DMARD therapy used for RA in our division. Other medications prescribed in 1996 or during the followup are summarized in Table 1.
|Age at disease onset, years||49.3; 50 (39–59)|
|Age at disease diagnosis, years||54.1; 57 (46–64)|
|Age at the onset of study, years||59.7; 61 (51–70)|
|Delay to disease diagnosis, years||5.2; 2.0 (1–7)|
|Disease duration before 1996, years||10.5; 8 (4–14)|
|Followup from diagnosis, years||13.4; 13 (10–16)|
|Women, no. (%)||131 (72)|
|Rheumatoid factor positive, no. (%)||141 (78)|
|Extraarticular manifestations, no. (%)||56 (31)|
|Cardiovascular risk factors, no. (%)|
|In 1996 (onset of study)||58 (32)|
|During followup||70 (39)|
|Patients receiving DMARD treatment in 1996, no. (%)||174 (96)|
|DMARD therapies used in 1996, no. (%)|
|Cyclosporin A†||54 (30)|
|Patients treated with gold prior to the onset of methotrexate therapy, no. (%)||80 (44)|
|Other medications administered during followup, no. (%)|
|Patients receiving prednisone for at least 1 year, no. (%)||173 (95)|
|CRP, mg/liter||10.3; 9 (5–14)|
|ESR, mm/hour||28.2; 25 (16–35)|
|Shared epitope positive, no./total no. (%)||119/178 (67)|
|HLA–DRB1*04 shared epitope positive, no./total no. (%)||87/178 (49)|
|HLA–DRB1*0401, no./total no. (%)||53/178 (30)|
|HLA–DRB1*0404, no./total no. (%)||16/178 (9)|
All patients were also treated with nonselective cyclooxygenase 2 inhibitor nonsteroidal antiinflammatory drugs. Prior to enrollment and/or throughout the study, 173 (95%) patients received prednisone (median dosage 5 mg/day) for at least 1 year. The mean ± SD prednisone cumulative dose at the end of the study was 13.5 ± 9.0 g. Twenty-six patients were receiving statins. The main epidemiologic and clinical features of this series of patients are also summarized in Table 1. Of note, there was a long delay to the diagnosis of RA from the initial symptoms of the disease (median 2.0 years) (Table 1).
Thirty-nine (21%) patients experienced CV events. The type of CV event (n = 43) in these 39 patients is described in Table 2. Only 4 patients experienced CV events prior to disease diagnosis. Two of these patients experienced CV events a long time after the onset of RA manifestations, and they had a long delay from the onset of symptoms of RA to diagnosis (one had a TIA 17 years after the onset of disease symptoms and 3 years before the disease diagnosis, and another had a stroke 16 years after disease onset and 2 years before disease diagnosis). The other 2 patients experienced CV events prior to the onset of RA manifestations (one had IHD [angina] 2 years before disease onset and 3 years before disease diagnosis, and another had acute myocardial infarction 7.5 years before the onset of RA symptoms and 8 years before disease diagnosis). At the end of this study, 17 (9%) patients had died because of CV complications. Acute myocardial infarction was the leading cause of CV death in this series (Table 2).
|Number of patients who experienced CV events||39 (21)|
|Before disease diagnosis||4 (2)|
|After disease diagnosis||35 (19)|
|Type of CV event†|
|Ischemic heart disease||19 (10)|
|Acute myocardial infarction||14 (8)|
|Heart failure||7 (4)|
|Cerebrovascular accidents||15 (8)|
|Peripheral arteriopathy||2 (1)|
|Number of patients who died because of CV complications||17 (9)|
|Causes of CV death|
|Acute myocardial infarction||8 (4)|
|Heart failure||5 (3)|
|Peripheral arteriopathy and mesenteric thrombosis||1 (1)|
Compared with the general Spanish population, the age- and sex-standardized mortality ratio by CV cause in this series of patients with RA from Lugo was 1.78. Age at the onset of the disease (by each year) was significantly associated with increased CV death (P = 0.001). When CV mortality was adjusted by age at disease onset and sex, some differences according to clinical features and HLA–DRB1 phenotype were observed (Table 3). Although the presence of a positive rheumatoid factor was associated with an increased HR (3.24), this association was not statistically significant. Likewise, the increased risk of CV mortality associated with traditional CV risk factors did not achieve statistical significance. However, a significantly increased risk of CV death was associated with chronic inflammation determined by CRP level (HR by each mg/liter 1.14, 95% CI 1.06–1.23, P < 0.001) and ESR (HR by each mm/hour 1.05, 95% CI 1.01–1.08, P = 0.003). Also, patients with HLA–DRB1*04 shared epitope alleles (HR 4.15, 95% CI 1.15–15.0, P = 0.030), particularly those who were HLA–DRB1*0404 positive (HR 6.65, 95% CI 1.98–22.33, P = 0.002), had an increased risk of CV mortality (Table 3). Of particular interest, none of the 59 patients who were negative for the rheumatoid shared epitope died because of CV complications. These results did not change when CV mortality was also adjusted by MTX treatment (data not shown).
|Rheumatoid factor positive||3.24||0.42–24.94||0.260|
|Mean CRP, mg/liter||1.14||1.06–1.23||< 0.001|
|Mean ESR, mm/hour||1.05||1.01–1.08||0.003|
|Radiographic index (Larsen)|
|Hands and feet||1.13||0.86–1.47||0.387|
|Cardiovascular risk factors|
|HLA–DRB1*04 shared epitope||4.15||1.15–15.0||0.030|
Because most patients were treated with steroids, we could not establish the potential impact of steroids on the development of CV events or CV mortality. In this regard, 38 of 173 patients treated with prednisone for at least 1 year developed CV events compared with 1 of 9 patients not treated with prednisone (odds ratio 2.25, 95% CI 0.29–102.5). Due to the small number of patients not treated with prednisone, it was difficult to statistically confirm any prednisone–CV relationship in this series.
Similar results were obtained when searching for predictors of the first CV event (Table 4). An increased risk of CV events was associated with CRP level (HR by each mg/liter 1.09, 95% CI 1.04–1.14, P = 0.001) and ESR (HR by each mm/hour 1.03, 95% CI 1.01–1.06, P = 0.003). Also, patients positive for HLA–DRB1*0404 were 4 times more likely to develop CV events (HR 4.47, 95% CI 1.75–11.39, P = 0.002). However, HLA–DRB1*04 shared epitope was not significantly associated with a higher risk of CV events (HR 1.76, 95% CI 0.85–3.63, P = 0.125) (Table 4). In addition, patients with more severe radiographic joint damage in the feet also had an increased risk of CV events (HR 1.37, 95% CI 1.05–1.77, P = 0.019) (Table 4). This observation was in accordance with a former study by our group that found an increased frequency of erosive disease in patients from Lugo with RA and HLA–DRB1*04 shared epitope positive alleles, largely due to HLA–DRB1*0404 and *0405 but not *0401 (25).
|Rheumatoid factor positive||2.39||0.72–7.90||0.155|
|Mean CRP, mg/liter||1.09||1.04–1.14||0.001|
|Mean ESR, mm/hour||1.03||1.01–1.06||0.003|
|Radiographic index (Larsen)|
|Hands and feet||1.06||0.88–1.27||0.555|
|Cardiovascular risk factors|
|HLA–DRB1*04 shared epitope||1.76||0.85–3.63||0.125|
Atherosclerosis is the leading cause of mortality in patients with RA. Recognition of the inflammatory basis for CV disease in the general population has increased interest in understanding the CV morbidity associated with RA. In the present study, our goal was to establish whether CV mortality was increased in a cohort of patients with RA from northwest Spain. We also reviewed the frequency of CV events and the causes of CV death in this series of patients. Finally, we attempted to determine the potential implication of clinical features, laboratory markers of inflammation, and HLA–DRB1 phenotype in the incidence of CV complications in patients with RA. Our results demonstrated an increased CV mortality in patients with RA from northwest Spain. CV mortality and CV events were associated with chronic inflammation determined by CRP level and ESR and HLA–DRB1*0404 shared epitope allele.
In the last few years, some studies have emphasized the importance of DMARD treatment, in particular MTX therapy, to decrease the high CV mortality observed in patients with RA (30, 31). However, the present study demonstrates that CV death is still increased in patients with RA treated with MTX.
Epidemiologic observations suggest that mechanisms other than the classic atherosclerotic risk factors may play a pivotal role in the increased prevalence of CV complications found in patients with RA (6). Solomon et al showed that the biomarkers of inflammation associated with CV disease are generally elevated in women with RA compared with those without this chronic disease (32). In this regard, the markers of systemic inflammation confer a statistically significant additional risk for CV death among patients with RA, even after controlling for traditional CV risk factors and comorbidities (33).
We previously reported the presence of subclinical atherosclerosis, manifested by increased carotid artery IMT and increased number of carotid plaques, in patients with RA without classic CV risk factors or CV events (13). As previously suggested by Sattar et al (8), further evidence confirmed that in patients with RA from Lugo the magnitude and chronicity of inflammatory response strongly correlated with the development of subclinical atherosclerosis (14). More recently, the study by Goodson et al was the first to demonstrate that baseline CRP level is a predictor of all-cause mortality, and specifically of CV mortality, in patients with inflammatory polyarthritis in a 10-year period following the onset of the rheumatic disease (34). The results shown in the present study reinforce the role of routine laboratory markers of inflammation as predictors of CV events and CV mortality in patients with RA.
Regardless of the etiology, a chronic elevation of proinflammatory cytokines is deleterious and promotes accelerated atherogenesis (8, 35). Interestingly enough, atherosclerosis and RA share similar inflammatory mechanisms that include involvement of proinflammatory cytokines such as tumor necrosis factor α (TNFα) and interleukin-6 (8, 35). Inflammatory mechanisms implicated in the development of synovial lesions in patients with RA might also involve the vessel walls and promote the development of atherosclerotic lesions in these patients. Hürlimann et al demonstrated that TNFα blockade improved endothelial function in patients with RA after a short-term treatment with anti-TNFα monoclonal antibody infliximab (36). We observed a rapid but transient positive effect on endothelial function following anti-TNFα infliximab therapy in patients with RA treated with this drug for at least 1 year (37). This effect on endothelial function highlighted the importance of proinflammatory cytokines in the mechanisms of atherosclerosis mediated by endothelial dysfunction in RA. Another recent study from our group supported the evidence of progression of subclinical atherosclerosis in patients with RA with severe disease despite periodic treatment with anti-TNFα monoclonal antibody infliximab (38). These observations suggest that the potential effect of anti-TNFα therapy is transient and progression of atherosclerosis in patients with RA is in line with changes in systemic inflammatory levels. As emphasized by Sattar and McInnes (35), systemic levels of proinflammatory cytokines generally remain dysregulated in patients with RA compared with individuals without this chronic disease and may promote the development of vascular disease and atherosclerosis.
There is still controversy as to whether the influence of HLA–DRB1 molecules on RA is due to effects on disease susceptibility or on disease severity and progression. However, a number of studies have demonstrated an association between severe disease outcome and the presence of the RA shared epitope (39–43). We also provided further evidence demonstrating that the presence of the shared epitope in the RA population of northwest Spain was associated with the development of severe disease, which was likely to require aggressive drug therapy (21). More recently, our group demonstrated a correlation between endothelial dysfunction, an early step in the development of atherosclerosis, and HLA–DRB1*0404 allele in patients with RA (22, 23). The present study is the first to support a potential implication of HLA–DRB1 alleles, in particular HLA–DRB1* 0404, in the development of CV events and CV death in RA. This study also strengthens evidence of the role of chronic inflammation and suggests that genetic factors and inflammation may play an additive role in the CV mortality found in patients with RA. However, there are several limitations in this study. First, increased levels of homocysteine have been found in patients with RA (11). Homocysteine is directly toxic to endothelial cells, increases low-density lipoprotein oxidation, and has prothrombotic effects (11). However, homocysteine serum levels were not assessed in this series of patients with RA. Another point of potential concern in this study was that most patients from our series were exposed to a significant mean dose of corticosteroids. This is particularly worrisome because corticosteroid level may be a risk factor for atherosclerosis, although no relationship between corticotherapy and atherosclerosis in RA as measured by carotid IMT has been found in previous studies (13, 44, 45), and the association between cortisol and the carotid artery IMT progression in the general population remains unclear (46, 47). Finally, RA is a polygenic disease and the potential effect of other genes different from those of the HLA–DRB1 region in the accelerated atherosclerosis observed in patients with RA needs to be addressed.
Additional therapies should be considered to delay the progression of atherosclerosis in patients with RA. Statins improve endothelium-dependent coronary vasodilatation in patients with atherosclerosis. This vascular improvement, which is independent of the cholesterol-lowering actions of statins, has been found to be associated with the antiinflammatory and immunomodulatory properties of statins (35, 48). Interestingly, Hermann et al have recently confirmed a dramatic improvement of endothelial function in patients with RA following statin therapy (49). Therefore, drugs not specifically intended to treat rheumatic manifestations might be considered in the management of RA to reduce the increased incidence of CV events in patients with this disease.
In conclusion, our results suggest that a chronically high inflammatory response in genetically predisposed individuals promotes an increased risk of CV events and CV mortality in patients with RA. The clinical significance of these results may allow rheumatologists to identify a subgroup of patients with RA who have an increased CV risk. Future studies should confirm these data, in particular the influence of HLA–DRB1 alleles in the increased risk of CV events and CV mortality in RA.
Dr. Gonzalez-Gay had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Drs. Gonzalez-Gay, Gonzalez-Juanatey, Ollier, Martin, and Llorca.
Acquisition of data. Drs. Gonzalez-Gay, Lopez-Diaz, Piñeiro, Garcia-Porrua, and Miranda-Filloy.
Analysis and interpretation of data. Drs. Gonzalez-Gay, Gonzalez-Juanatey, Ollier, Martin, and Llorca.
Manuscript preparation. Dr. Gonzalez-Gay.
Statistical analysis. Drs. Gonzalez-Gay and Llorca.
- 27Ischemic heart disease in patients from Northwest Spain with biopsy proven giant cell arteritis: a population based study. J Rheumatol 2005; 32: 502–6., , , , , , et al.
- 34Baseline levels of C-reactive protein and prediction of death from cardiovascular disease in patients with inflammatory polyarthritis: a ten-year followup study of a primary care-based inception cohort. Arthritis Rheum 2005; 52: 2293–9., , , , , .