Double role of mannose-binding lectin in relation to carotid intima–media thickness in patients with rheumatoid arthritis

Results The median ccIMT was 0.67 mm. The investigated MBL2 genotypes were not significantly associated with ccIMT. Using a general linear model, ccIMT was not linearly associated with serum MBL but was highly associated with the quadratic term of serum MBL (MBL 2 ) ( P = 0.001) reflecting a U-shaped relation. MBL 2 was also significantly associated with ccIMT in a multivariable analysis adjusting for traditional and RA related cardiovascular risk modifiers ( P = 0.025). Conclusion In RA patients, a quadratic U-shaped relation between serum MBL and ccIMT was observed independently of the effects of traditional and RA related risk factors for CVD. These results provide further support to the notion that both high and low levels of MBL may be associated with CVD. Keywords Rheumatoid arthritis Atherosclerosis Cardiovascular disease Mannose-binding lectin Risk factors 1 Introduction Patients with rheumatoid arthritis (RA) have increased risk of atherosclerosis and cardiovascular disease (CVD) ( Kerekes et al., 2008; Kumeda et al., 2002; Solomon et al., 2003 ) which cannot be fully explained by an excess of traditional cardiovascular risk factors ( Del Rincon et al., 2001; Dessein et al., 2005 ). However, several studies have shown that markers of systemic inflammation are associated with atherosclerosis and CVD in addition to traditional risk factors of CVD, particularly in patients with RA ( Del Rincon et al., 2003; Dessein et al., 2005; Gonzalez-Gay et al., 2005 ). Common proinflammatory cytokines are involved in the development and progression of both RA and atherosclerosis and tumor necrosis factor α (TNF-α) expressed by affected joints in RA is thought to act at a distance promoting all stages of atherosclerosis development ( Full et al., 2009 ). Other immunologic responses may also link RA and atherosclerosis, e.g. activation of specific T-cell populations ( Full et al., 2009 ) and several studies indicate that genetic factors, factors concerning the innate immune system and glycosylation changes of various proteins could be important for the development of atherosclerosis and CVD in RA ( Gonzalez-Gay et al., 2007; Troelsen et al., 2007 ). Several studies have shown that potent anti-inflammatory medication provides surviving benefits by reducing cardiovascular morbidity and mortality ( Choi et al., 2002; Van Halm et al., 2006 ). Mannose-binding lectin (MBL) is a C-type lectin involved in innate immune defence. It binds microorganisms and cellular debris through the carbohydrate recognition domain. Serum MBL can directly opsonize pathogens and enhance the activity of phagocytic cells or it can activate complement via the lectin pathway. Differences in MBL serum concentrations can partly be explained by structural variant alleles in the human MBL gene ( MBL2 ) on chromosome 10. The normal genotype is associated with the highest serum concentrations of MBL. Heterozygosity for MBL2 structural variant alleles causes in average a 85–90% drop in the serum concentration of functional MBL and homozygosity for structural variant alleles is devoid of functional MBL. The normal MBL2 allele is named A, and the common designation for the variant alleles is O. In addition to the effect of the structural allelic variants, polymorphisms in the promoter region of the MBL2 gene affect serum levels of MBL. In particular, a base substitution at codon-221 (G to C; promoter allele X) is associated with lower MBL serum concentrations. However, the large inter-individual variation in MBL serum concentration cannot fully be explained by the commonly described MBL2 polymorphisms ( Garred et al., 2006 ). In a recent follow-up study we experienced that genetically determined high serum levels of MBL increased the risk of myocardial infarction and ischemic heart disease in patients with RA ( Troelsen et al., 2007 ). This comes in line with several clinical and experimental studies based on other study populations: high serum levels of MBL increase the risk of coronary artery disease in apparently healthy men ( Keller et al., 2006 ) and in patients with diabetes ( Hansen et al., 2004 ) and high serum levels of MBL and the lectin pathway of complement initiates the inflammatory reaction seen in relation to ischemia reperfusion ( Busche et al., 2008; Collard et al., 2000 ). However, also lack of functional alleles and low serum levels of MBL have been associated with increased risk of atherosclerosis and coronary artery occlusion in other populations ( Best et al., 2004; Ohlenschlaeger et al., 2004 ). These observations indicate that both high and low serum levels of MBL may be disadvantageous in atherosclerosis development and that a non-linear relationship between serum levels of MBL and cardiovascular risk may be expected. Intima–media thickness of the common carotid artery (ccIMT), determined by high-resolution B-mode ultrasonography is a useful non-invasive anatomic measure of subclinical CVD and constitutes an excellent surrogate marker of generalized atherosclerosis and coronary artery disease ( Simons et al., 1999 ). Increased IMT has been observed in patients with RA compared to controls ( Kerekes et al., 2008 ). Furthermore, IMT has been reported to increase with disease duration and to associate with markers of inflammation in patients with RA ( Del Rincon et al., 2003; Gonzalez-Gay et al., 2005 ). Based on the above-described findings we wanted to examine the hypothesis of a non-linear relationship between serum levels of MBL and ccIMT in patients with RA. 2 Patients and methods 2.1 Patients One hundred and fourteen consecutive unrelated white Danish RA patients attending the Department of Rheumatology, Rigshospitalet, Copenhagen University Hospital were included from October 2006 to March 2007. All patients fulfilled the American College of Rheumatology (ACR) 1987 classification criteria for RA. At inclusion, all patients were subjected to a detailed interview and underwent a clinical examination. Posteroanterior (PA) + PA oblique radiographs of hands and feet were taken and fasting serum samples were collected and stored at −80 °C. Within 4 weeks of recruitment ccIMT was measured in all patients. The study was approved by the local scientific ethical committee (reference no. KF 01286984). Written informed consent was obtained from each patient. 2.2 Clinical and laboratory assessment Traditional cardiovascular risk modifiers recorded were: (1) gender, (2) age, (3) systolic blood pressure, (4) diastolic blood pressure, (5) smoking calculated as “pack-years” (number of packs (=20 cigarettes) per day times number of years), (6) body mass index (BMI), (7) story of diabetes, (8) total cholesterol, (9) HDL cholesterol, (10) LDL cholesterol, (11) VLDL cholesterol, (12) triglyceride and (13) homeostatic model assessment of insulin resistance (HOMA1-IR) ((insulin (mU/L) × glucose (mmol/L))/22.5) ( Wallace et al., 2004 ). Potential RA related cardiovascular risk modifiers reflecting acute inflammation and chronic inflammatory damage were: (1) duration of RA, (2) joint destruction, evaluated according to the principles of the Sharp/von der Heide scoring method (Total Sharp Score, TSS, possible range: 0–448) ( Van der Heijde, 2000 ), (3) functional disability assessed by the Health Assessment Questionnaire (HAQ) score (possible range: 0–3), (4) serum anti-cyclic citrullinated peptide antibodies (anti-CCP), (5) C-reactive protein measured by means of a highly sensitive method (hsCRP), (6) serum IL-6, (7) serum TNF-α, (8) non-galactosylated IgG assessed by the FG0/FG1 ratio ( Royle et al., 2008 ). Present treatment with glucocorticoids, methotrexate (MTX) and anti-TNF-α drugs was also recorded. All patients were interviewed and examined by the same physician and the same radiologist evaluated all radiographs of hands and feet without knowledge of other study data. In 9 cases it was not possible to calculate TSS due to articular surgery and no available prior radiographs that could have been used for carrying forward the last observation. Analysis of lipids and plasma insulin were not performed in 3 patients, because of inadequate material. Clinical charts were used to verify information regarding treatment with anti-rheumatic drugs. 2.3 Genotyping and detection of MBL serum concentration Detection of MBL2 alleles was analysed as previously described ( Madsen et al., 1995 ). All patients were genotyped for three MBL2 structural polymorphisms and for a promoter polymorphism in position −221 (X/X, X/Y, Y/Y). The following inferred haplotypes were tested for YA/YA, YA/XA, XA/XA, YA/YO, XA/YO and YO/YO. Serum concentrations of MBL were measured in a double ELISA based on a monoclonal anti-MBL antibody ( Garred et al., 1992 ). 2.4 CcIMT measurement Measurement of ccIMT was performed by using an Acuson Cypress ultrasound apparatus with a 7 MHz linear ultrasound probe. A still picture of the common carotid artery proximal to the bulbus caroticus was produced, ensuring a distinct picture of intima and media on the far wall of the vessel (the intima–media). Thickness was measured 3 times using the electronic callipers of the ultrasound apparatus and the picture was digitally stored for future reference. Measurements were performed on both carotid arteries. A mean for both arteries was calculated and used in analysis. 2.5 Statistical analysis One-way analysis of variance was used to examine the relationship between ccIMT and inferred haplotypes (YA/YA, YA/XA, XA/XA, YA/YO, XA/YO and YO/YO). Associations between ccIMT, traditional CVD risk modifiers and potential RA related modifiers of CVD risk were assessed using the Spearman rank correlation. The relationship between ccIMT (dependent variable) and MBL serum concentrations as well as other risk factors (explanatory) variables was studied using a general linear model. For the MBL concentration, a quadratic term (MBL 2 ) was included in order to obtain the best non-linear fit to the data. The risk factors included in the analysis were those significantly correlated to ccIMT. Univariate and multivariable regression analyses were performed; the latter was furthermore reduced in a stepwise manner using a 5% significance level to retain variables in the regression model. Model validation was done using conventional methods: assessment of residuals, Cook's D and internal validation. Statistical processing of the study was done using SPSS 15 and SAS 9.1. Statistical significance was defined as P -value <0.05. 3 Results Clinical and laboratory variables at inclusion for all patients are shown in Table 1 . The median ccIMT was 0.67 mm and ranged from 0.37 to 1.1 mm. The median serum concentration of MBL was 1.2 mg/L and ranged from 0 to 5.7 mg/L. Serum MBL varied with the inferred haplotypes (YA/YA, YA/XA, XA/XA, YA/YO, XA/YO and YO/YO) as shown in Fig. 1 . Using one-way analysis of variance we found no significant difference in ccIMT and the examined haplotypes ( P -value = 0.625). 3.1 ccIMT and relation to traditional and RA related risk factors for CVD Results of Spearman rank correlation analysis between ccIMT, traditional and RA related cardiovascular risk modifiers are shown in Table 2 . ccIMT was significantly correlated to male gender, age, systolic blood pressure, diastolic blood pressure, smoking (pack-years), BMI, story of diabetes, duration of RA and inversely correlated with present treatment with MTX. 3.2 General linear model analysis Using the general linear model analysis we examined ccIMT in relation to (A) MBL (linear model) and (B) MBL + MBL 2 (quadratic model). Curve fits and equations for the two models are shown in Fig. 2 . We found a highly significant quadratic relation between ccIMT and MBL ( P -value = 0.002). There was no linear association between ccIMT and MBL ( P -value = 0.284). Results from univariate and multivariable analyses are shown in Table 3 . This table includes parameter estimates and P -values for MBL 2 and the traditional and RA related cardiovascular risk modifiers included in the analysis. In model 1 only MBL and MBL 2 was included. The analysis of model 1 shows that MBL 2 was significantly associated with ccIMT ( P < 0.001). Model 2 included MBL, MBL 2 , traditional and RA related modifiers of CVD risk that were significantly correlated to ccIMT in the correlation analysis. In this model, MBL 2 remained significantly associated with ccIMT ( P = 0.025). Finally we reduced the model stepwise by the removal of the least significant variable: (a) story of diabetes, (b) diastolic blood pressure, (c) systolic blood pressure, (d) male gender, (e) duration of RA, (f) present treatment with MTX ending up with the final model 3 where MBL 2 ( P = 0.008), age ( P < 0.001), smoking (pack-years) ( P = 0.023) and BMI ( P = 0.006) were all significantly associated with ccIMT. The quadratic model fitting the model MBL as the explanatory variable and ccIMT as the dependent variables was assessed as described in Section 2.5 . In particular, removing extreme MBL levels (four highest MBL levels) resulted in estimates almost equivalent to the analysis including all values. In general, the internal validation and assessment of residuals as well as Cook's D demonstrated a reasonable fit to the data. 4 Discussion The role of MBL in the development of atherosclerosis and CVD has been discussed for more than 10 years. Several clinical and experimental studies have been performed with conflicting results indicating that both high and low serum concentrations of MBL in various study populations may be associated with atherosclerosis. Since RA patients have increased incidence of atherosclerosis that cannot alone be explained by excess of traditional cardiovascular risk factors and since MBL in a previous study was associated with myocardial infarction and ischemic heart disease in patients with RA ( Troelsen et al., 2007 ) we wanted to examine if serum levels of MBL in RA patients were associated with increased subclinical atherosclerosis measured by ultrasonography. Using general linear model analysis including a quadratic term of MBL we found, for the first time in a clinical study, that serum MBL has a quadratic U-shaped relation with ccIMT indicating that MBL has a dual effect in the development of atherosclerosis in RA patients. Moreover, we showed that the quadratic relation between serum MBL and ccIMT was consistent even after adjustment for traditional and RA related risk factors for CVD. ccIMT ≥ 0.60 mm is a marker of generalized atherosclerosis ( Belcaro et al., 2001 ). In our study, the mean ccIMT was 0.68 mm, which is in accordance with previous findings in RA patients ( Kerekes et al., 2008; Kumeda et al., 2002 ). As expected ccIMT was correlated to several traditional CV risk factors, i.e. male gender, age, history of diabetes, increasing systolic blood pressure, increasing diastolic blood pressure, BMI and number of pack-years. However ccIMT was not correlated to lipid risk factors or HOMA1-IR. This may partly be explained by the fact that most of the RA patients participating in our study were treated with MTX and/or anti-TNF-α drugs which are shown to have beneficial effects on lipid profiles and insulin sensitivity ( Garcia-Gomez et al., 2009; Kiortsis et al., 2005 ). Among RA related modifiers of CVD risk we found a significant correlation between ccIMT and duration of RA. However, we did not find any significant correlations between ccIMT and markers of acute inflammation (hsCRP, IL-6, TNF-α, FG0/FG1 ratio) or cumulative inflammatory damage (TSS, HAQ score). In some studies, a significant association between ccIMT and CRP ( Del Rincon et al., 2003; Gonzalez-Gay et al., 2005 ) has been described. Other studies have not been able to find this association ( Dessein et al., 2005; Kerekes et al., 2008; Kumeda et al., 2002 ) or associations between ccIMT and IL-6 and TNF-α ( Kerekes et al., 2008 ). Several studies have also described positive associations between ccIMT, radiographic scores and HAQ score ( Dessein et al., 2005; Kumeda et al., 2002 ), but in our study we found no significant correlations between ccIMT and TSS or HAQ score. These discrepancies could be due to still unidentified confounders, and it is likely that differences in study design, differences in medication, age and duration of disease, may influence the associations studied. Regarding ccIMT and treatment with anti-rheumatic drugs we observed that ccIMT was inversely correlated to present treatment with MTX. This comes in line with other studies where MTX has been shown to reduce cardiovascular morbidity and mortality ( Choi et al., 2002; Van Halm et al., 2006 ). We did not find any significant correlations between ccIMT and present treatment with glucocorticoids or anti-TNF-α drugs. In one-way analysis of variance we could not find any difference in ccIMT between the inferred haplotypes (YA/YA, YA/XA, XA/XA, YA/YO, XA/YO and YO/YO). This may be due to the high degree of variation in MBL serum concentrations within the various MBL haplotypes as visualized in Fig. 1 . However, the present study corroborates our previous findings of a higher degree of association between MBL serum concentrations, myocardial infarction and ischemic heart disease in patients with RA than between MBL genotypes and the outcome variables mentioned ( Troelsen et al., 2007 ). Moreover in a recent study of MBL and diabetes there were no associations between MBL2 genotypes (including 20 MBL2 single nucleotide polymorphisms) and type 1 diabetes or diabetic nephropathy whereas MBL serum concentrations were higher in patients with macroalbuminuria ( Kaunisto et al., 2009 ). Nevertheless, the individual MBL serum concentrations are very stable over time ( Graudal et al., 1998 ). Thus, still unknown polymorphic sites and other unknown regulatory mechanisms may determine MBL serum concentrations and thereby contribute to the lack of association between the presently measured MBL2 genotypes and ccIMT ( Bernig et al., 2005 ). Both proatherogenic and antiatherogenic roles of MBL have been demonstrated in several population-based studies and experimental studies ( Best et al., 2004; Busche et al., 2008; Collard et al., 2000; Hansen et al., 2004; Keller et al., 2006; Ohlenschlaeger et al., 2004 ). In a recent study a differential role of MBL in a mouse model of atherosclerosis development was indicated ( Matthijsen et al., 2009 ) but our study is the first to show that both high and low serum levels of MBL are associated with increased subclinical atherosclerosis within the same disease population. However other studies indicate that heterozygosity for the MBL2 allele and intermediary serum concentrations of MBL confer advantage in various disease settings. Heterozygosity for the MBL2 allele is associated with protection against clinical tuberculosis ( Soborg et al., 2003 ), confers a protective effect in critically ill patients attending an intensive care unit ( Hellemann et al., 2007 ) and has disease-retarding effects in HIV patients ( Catano et al., 2008 ). MBL can enhance the clearance of pathogens, IgG-G0 containing immune complexes as well as dying host cells and cellular debris ( Dommett et al., 2006 ). Based on these abilities several pathophysiological mechanisms may be suggested to explain the dual role of MBL in the development of atherosclerosis in RA patients. A theory of infection in atherosclerosis is not proven, but several studies have shown that chronic infections including periodontitis and infections with Helicobacter pylori and Chlamydia pneumoniae may be associated with the development of atherosclerosis and cardiovascular disease ( Palm et al., 2009 ). As infection with C. pneumoniae mainly leads to the development and progression of severe coronary artery disease in patients with variation in the MBL2 gene ( Rugonfalvi-Kiss et al., 2002 ) it could be hypothesized that deficiency of MBL results in defective clearance of infectious material thereby aggravating ongoing inflammatory processes leading to atherosclerosis development. Moreover lack of functional MBL may result in increased amounts of endogenous cellular debris, which in human atherosclerotic lesions contributes to gruel accumulation and plaque instability ( Bjorkerud and Bjorkerud, 1996 ). On the other hand, excessive complement activation through the MBL mediated lectin pathway may result in an inflammation-driven proatherogenic process stimulating atherosclerosis or development of myocardial infarction and ischemic heart disease ( Troelsen et al., 2007 ). In support of this theory terminal complement complexes have been associated with endothelial damage in rheumatoid nodules, indicating that complement activation may play a role in vascular damage in RA ( Kato et al., 2000 ). Moreover, oxidative stress seems to expose structures on endothelial cells that promotes binding of MBL ( Collard et al., 2000 ). MBL may interact with damaged endothelial cells, activate complement and increase the inflammatory reaction in vessels leading to atherosclerosis. However, the previously demonstrated interaction between MBL and IgG-G0 that increased the risk of CVD among patients with RA ( Troelsen et al., 2007 ) had in this study no effect on ccIMT. This may indicate that complement activation due to MBL binding to IgG-G0 only plays a major role in the late stages of atherosclerosis. In conclusion, the clinical U-shaped findings in this study supported by experimental studies where a differentiated role of MBL in atherosclerosis development was observed clearly indicate a dual pathophysiological contribution of MBL in atherosclerosis in RA. Future studies should focus on the mechanisms by which MBL influences atherosclerosis development and if our findings may be applicable beyond RA. Acknowledgements This work was supported by grants from the Danish Rheumatism Association, The Novo Nordisk Research Foundation, The Capital Region of Denmark and Rigshospitalet, Copenhagen University Hospital. References Belcaro et al., 2001 G. Belcaro A.N. Nicolaides G. Ramaswami M.R. Cesarone S.M. De L. Incandela P. Ferrari G. Geroulakos A. Barsotti M. Griffin S. Dhanjil M. Sabetai M. Bucci G. Martines Carotid and femoral ultrasound morphology screening and cardiovascular events in low risk subjects: a 10-year follow-up study (the CAFES-CAVE study(1)) Atherosclerosis 156 2001 379 387 Bernig et al., 2005 T. Bernig W. Breunis N. Brouwer A. Hutchinson R. Welch D. Roos T. Kuijpers S. Chanock An analysis of genetic variation across the MBL2 locus in Dutch Caucasians indicates that 3′ haplotypes could modify circulating levels of mannose-binding lectin Hum. Genet. 118 2005 404 415 Best et al., 2004 L.G. Best M. Davidson K.E. North J.W. MacCluer Y. Zhang E.T. Lee B.V. Howard S. DeCroo R.E. Ferrell Prospective analysis of mannose-binding lectin genotypes and coronary artery disease in American Indians—the strong heart study Circulation 109 2004 471 475 Bjorkerud and Bjorkerud, 1996 S. Bjorkerud B. Bjorkerud Apoptosis is abundant in human atherosclerotic lesions, especially in inflammatory cells (macrophages and T cells), and may contribute to the accumulation of gruel and plaque instability Am. J. Pathol. 149 1996 367 380 Busche et al., 2008 M.N. Busche M.C. Walsh M.E. McMullen B.J. Guikema G.L. Stahl Mannose-binding lectin plays a critical role in myocardial ischaemia and reperfusion injury in a mouse model of diabetes Diabetologia 51 2008 1544 1551 Catano et al., 2008 G. Catano B.K. Agan H. Kulkarni V. Telles V.C. Marconi M.J. Dolan S.K. Ahuja Independent effects of genetic variations in mannose-binding lectin influence the course of HIV disease: the advantage of heterozygosity for coding mutations J. Infect. Dis. 198 2008 72 80 Choi et al., 2002 H.K. Choi M.A. Hernan J.D. Seeger J.M. Robins F. Wolfe Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study Lancet 359 2002 1173 1177 Collard et al., 2000 C.D. Collard A. Vakeva M.A. Morrissey A. Agah S.A. Rollins W.R. Reenstra J.A. Buras S. Meri G.L. Stahl Complement activation after oxidative stress—role of the lectin complement pathway Am. J. Pathol. 156 2000 1549 1556 Del Rincon et al., 2001 I. Del Rincon K. Williams M.P. Stern G.L. Freeman A. Escalante High incidence of cardiovascular events in a rheumatoid arthritis cohort not explained by traditional cardiac risk factors Arthritis Rheum. 44 2001 2737 2745 Del Rincon et al., 2003 I. Del Rincon K. Williams M.P. Stern G.L. Freeman D.H. O’Leary A. Escalante Association between carotid atherosclerosis and markers of inflammation in rheumatoid arthritis patients and healthy subjects Arthritis Rheum. 48 2003 1833 1840 Dessein et al., 2005 P.H. Dessein B.I. Joffe M.G. Veller B.A. Stevens M. Tobias K. Reddi A.E. Stanwix Traditional and nontraditional cardiovascular risk factors are associated with atherosclerosis in rheumatoid arthritis J. Rheumatol. 32 2005 435 442 Dommett et al., 2006 R.M. Dommett N. Klein M.W. Turner Mannose-binding lectin in innate immunity: past, present and future Tissue Antigens 68 2006 193 209 Full et al., 2009 L.E. Full C. Ruisanchez C. Monaco The inextricable link between atherosclerosis and prototypical inflammatory diseases rheumatoid arthritis and systemic lupus erythematosus Arthritis Res. Ther. 11 2009 217 Garcia-Gomez et al., 2009 C. Garcia-Gomez J.M. Nolla J. Valverde J.A. Gomez-Gerique M.J. Castro X. Pinto Conventional lipid profile and lipoprotein(a) concentrations in treated patients with rheumatoid arthritis J. Rheumatol. 36 2009 1365 1370 Garred et al., 2006 P. Garred F. Larsen J. Seyfarth R. Fujita H.O. Madsen Mannose-binding lectin and its genetic variants Genes Immun. 7 2006 85 94 Garred et al., 1992 P. Garred H.O. Madsen J.A.L. Kurtzhals L.U. Lamm S. Thiel A.S. Hey A. Svejgaard Diallelic polymorphism may explain variations of the blood-concentration of mannan-binding protein in eskimos, but not in black-Africans Eur. J. Immunogenet. 19 1992 403 412 Gonzalez-Gay et al., 2007 M.A. Gonzalez-Gay C. Gonzalez-Juanatey M.J. Lopez-Diaz A. Pineiro C. Garcia-Porrua J.A. Miranda-Filloy W.E. Ollier J. Martin J. Llorca HLA-DRB1 and persistent chronic inflammation contribute to cardiovascular events and cardiovascular mortality in patients with rheumatoid arthritis Arthritis Rheum. 57 2007 125 132 Gonzalez-Gay et al., 2005 M.A. Gonzalez-Gay C. Gonzalez-Juanatey A. Pineiro C. Garcia-Porrua A. Testa J. Llorca High-grade C-reactive protein elevation correlates with accelerated atherogenesis in patients with rheumatoid arthritis J. Rheumatol. 32 2005 1219 1223 Graudal et al., 1998 N.A. Graudal C. Homann H.O. Madsen A. Svejgaard A.G. Jurik H.K. Graudal P. Garred Mannan binding lectin in rheumatoid arthritis. A longitudinal study J. Rheumatol. 25 1998 629 635 Hansen et al., 2004 T.K. Hansen L. Tarnow S. Thiel R. Steffensen C.D. Stehouwer C.G. Schalkwijk H.H. Parving A. Flyvbjerg Association between mannose-binding lectin and vascular complications in type 1 diabetes Diabetes 53 2004 1570 1576 Hellemann et al., 2007 D. Hellemann A. Larsson H.O. Madsen J. Bonde J.O. Jarlov J. Wiis T. Faber J. Wetterslev P. Garred Heterozygosity of mannose-binding lectin (MBL2) genotypes predicts advantage (heterosis) in relation to fatal outcome in intensive care patients Hum. Mol. Genet. 16 2007 3071 3080 Kato et al., 2000 H. Kato M. Yamakawa T. Ogino Complement mediated vascular endothelial injury in rheumatoid nodules: a histopathological and immunohistochemical study J. Rheumatol. 27 2000 1839 1847 Kaunisto et al., 2009 M. Kaunisto L. Sjolind R. Sallinen K. Pettersson-Fernholm M. Saraheimo S. Frojdo C. Forsblom J. Fagerudd T. Hansen A. Flyvbjerg M. Wessman P.H. Groop Elevated MBL concentrations are not an indication of association between the MBL2 gene and type 1 diabetes or diabetic nephropathy Diabetes 58 2009 1710 1714 Keller et al., 2006 T.T. Keller S.I. van Leuven M.C. Meuwese N.J. Wareham R. Luben E.S. Stroes C.E. Hack M. Levi K.T. Khaw S.M. Boekholdt Serum levels of mannose-binding lectin and the risk of future coronary artery disease in apparently healthy men and women Arterioscler. Thromb. Vasc. Biol. 26 2006 2345 2350 Kerekes et al., 2008 G. Kerekes Z. Szekanecz H. Der Z. Sandor G. Lakos L. Muszbek I. Csipo S. Sipka I. Seres G. Paragh J. Kappelmayer E. Szomjak K. Veres G. Szegedi Y. Shoenfeld P. Soltesz Endothelial dysfunction and atherosclerosis in rheumatoid arthritis: a multiparametric analysis using imaging techniques and laboratory markers of inflammation and autoimmunity J. Rheumatol. 35 2008 398 406 Kiortsis et al., 2005 D.N. Kiortsis A.K. Mavridis S. Vasakos S.N. Nikas A.A. Drosos Effects of infliximab treatment on insulin resistance in patients with rheumatoid arthritis and ankylosing spondylitis Ann. Rheum. Dis. 64 2005 765 766 Kumeda et al., 2002 Y. Kumeda M. Inaba H. Goto M. Nagata Y. Henmi Y. Furumitsu E. Ishimura K. Inui Y. Yutani T. Miki T. Shoji Y. Nishizawa Increased thickness of the arterial intima–media detected by ultrasonography in patients with rheumatoid arthritis Arthritis Rheum. 46 2002 1489 1497 Madsen et al., 1995 H.O. Madsen P. Garred S. Thiel J.A.L. Kurtzhals L.U. Lamm L.P. Ryder A. Svejgaard Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein J. Immunol. 155 1995 3013 3020 Matthijsen et al., 2009 R.A. Matthijsen M.P. de Winther D. Kuipers I. van der Made C. Weber M.V. Herias M.J. Gijbels W.A. Buurman Macrophage-specific expression of mannose-binding lectin controls atherosclerosis in low-density lipoprotein receptor-deficient mice Circulation 119 2009 2188 2195 Ohlenschlaeger et al., 2004 T. Ohlenschlaeger P. Garred H.O. Madsen S. Jacobsen Mannose-binding lectin variant alleles and the risk of arterial thrombosis in systemic lupus erythematosus N. Engl. J. Med. 351 2004 260 267 Palm et al., 2009 F. Palm C. Urbanek A. Grau Infection, its treatment and the risk for stroke Curr. Vasc. Pharmacol. 7 2009 146 152 Royle et al., 2008 L. Royle M.P. Campbell C.M. Radcliffe D.M. White D.J. Harvey J.L. Abrahams Y.G. Kim G.W. Henry N.A. Shadick M.E. Weinblatt D.M. Lee P.M. Rudd R.A. Dwek HPLC-based analysis of serum N-glycans on a 96-well plate platform with dedicated database software Anal. Biochem. 376 2008 1 12 Rugonfalvi-Kiss et al., 2002 S. Rugonfalvi-Kiss V. Endresz H.O. Madsen K. Burian J. Duba Z. Prohaszka I. Karadi L. Romics E. Gonczol G. Fust P. Garred Association of Chlamydia pneumoniae with coronary artery disease and its progression is dependent on the modifying effect of mannose-binding lectin Circulation 106 2002 1071 1076 Simons et al., 1999 P.C. Simons A. Algra M.L. Bots D.E. Grobbee G.Y. van der Common carotid intima–media thickness and arterial stiffness: indicators of cardiovascular risk in high-risk patients. The SMART study (second manifestations of arterial disease) Circulation 100 1999 951 957 Soborg et al., 2003 C. Soborg H.O. Madsen A.B. Andersen T. Lillebaek A. Kok-Jensen P. Garred Mannose-binding lectin polymorphisms in clinical tuberculosis J. Infect. Dis. 188 2003 777 782 Solomon et al., 2003 D.H. Solomon E.W. Karlson E.B. Rimm C.C. Cannuscio L.A. Mandl J.E. Manson M.J. Stampfer G.C. Curhan Cardiovascular morbidity and mortality in women diagnosed with rheumatoid arthritis Circulation 107 2003 1303 1307 Troelsen et al., 2007 L.N. Troelsen P. Garred H.O. Madsen S. Jacobsen Genetically determined high serum levels of mannose-binding lectin and agalactosyl IgG are associated with ischemic heart disease in rheumatoid arthritis Arthritis Rheum. 56 2007 21 29 Van der Heijde, 2000 D. Van der Heijde How to read radiographs according to the Sharp/van der Heijde method J. Rheumatol. 27 2000 261 263 Van Halm et al., 2006 V. Van Halm M.T. Nurmohamed J.W. Twisk B.A. Dijkmans A.E. Voskuyl Disease-modifying antirheumatic drugs are associated with a reduced risk for cardiovascular disease in patients with rheumatoid arthritis: a case control study Arthritis Res. Ther. 8 2006 R151 Wallace et al., 2004 T.M. Wallace J.C. Levy D.R. Matthews Use and abuse of HOMA modeling Diabetes Care 27 2004 1487 1495