Patients with rheumatoid arthritis (RA) have excess morbidity and mortality due to ischemic heart disease. It has been suggested that high serum levels of mannose-binding lectin (MBL) and agalactosyl IgG (IgG-G0) are associated with increased inflammation in RA. MBL also enhances inflammation-mediated tissue injury during postischemic reperfusion. This study was undertaken to examine whether these factors are associated with increased risk of ischemic heart disease in RA.
MBL alleles were genotyped in 229 Danish patients with RA. In addition, serum levels of MBL and IgG-G0 were measured. Incidences of ischemic heart disease, myocardial infarction, and death due to ischemic heart disease after the diagnosis of RA were assessed in a prospective study.
During a median followup of 9.5 years, ischemic heart disease was diagnosed in 8 of 27 patients with genetically determined high serum levels of MBL, as compared with 24 of the remaining 192 patients (data not available on 10 patients). After correction for other known risk factors, the hazard ratio (HR) was 3.6 (95% confidence interval [95% CI] 1.4–9.2). The corrected HR for myocardial infarction was 9.0 (95% CI 2.2–36.4). High serum levels of MBL also conferred an increased risk of death due to ischemic heart disease (age- and sex-adjusted HR 10.5, 95% CI 2.7–41.3). However, further analyses showed that these associations were present only in patients with high serum levels of IgG-G0.
Genetically determined high serum levels of MBL and high serum levels of IgG-G0 are associated with increased risk of ischemic heart disease, myocardial infarction, and premature death in patients with RA.
Excess morbidity and mortality due to cardiovascular disease, especially ischemic heart disease, has been demonstrated in patients with rheumatoid arthritis (RA) (1, 2). The increased incidence of cardiovascular disease in RA cannot be attributed to traditional cardiac risk factors alone (3). In this respect, inflammatory mechanisms seem to be most important; markers of systemic inflammation are associated with an additional risk of cardiovascular death among patients with RA (4, 5), and potent antiinflammatory medication reduces cardiovascular mortality (6).
Mannose-binding lectin (MBL) is a liver-derived serum protein involved in innate immune defense. Ligands for MBL are, among others, mannose and N-acetylglucosamine residues, which are expressed on a wide range of microorganisms. Serum MBL can directly opsonize pathogens and enhance the activity of phagocytic cells or activate complement via the lectin pathway. Serum levels of MBL vary widely from person to person due to 3 different variant alleles in exon 1 of the human MBL2 gene on chromosome 10 (7). The normal allele is named A, and the common designation for the variant alleles is O. Moreover, base substitutions in the promoter region of the MBL2 gene affect the expression of the protein (8); specifically, in an otherwise structurally normal MBL2 gene, the preservation of G at position −221 (termed promoter allele Y and resulting in haplotype YA) is associated with a high serum level of MBL.
MBL may play a dual role in cardiovascular disease. Lack of functional alleles in some populations is associated with increased risk of atherosclerosis and cardiovascular occlusion (9–13). However, experimental data and clinical observations suggest that MBL and the lectin pathway of complement may initiate the inflammatory reaction seen in relation to ischemia reperfusion injury (14–16). Moreover, subjects who are homozygous for the functional MBL2 genotype and have high serum levels of MBL have the highest rate of early restenosis after carotid eversion endarterectomy (17).
MBL may also have differential roles in the pathogenesis of RA. In some studies, MBL deficiency has been associated with erosive disease (18, 19). Other studies have yielded inconclusive results in this regard (20). Conversely, the normal genotype (A/A) and high serum levels of MBL have been associated with increased inflammation in late-onset and advanced RA (21).
Glycosylation changes in the IgG Fc region are characteristic of RA (22, 23). IgG from patients with RA has marked differences in the oligosaccharide moieties in the C2 region, whereby the content of oligosaccharide chains terminating in galactose is reduced, leading to exposure of N-acetylglucosamine. This glycosylation variant is called agalactosyl IgG, or IgG-G0. Increased serum levels of IgG-G0 have been associated with poor prognosis in patients with RA (23, 24). IgG-G0 has been shown in vitro to interact with MBL through the exposed N-acetylglucosamine, thereby activating complement (25).
Based on the findings described above, we wished to determine whether MBL and IgG-G0 confer any additional risk of ischemic heart disease in RA. The present study was undertaken to investigate this.
PATIENTS AND METHODS
The patients included in the present study originated from 2 previously described study populations. Population 1 consisted of 189 consecutive unrelated white Danish RA patients (155 women, 34 men) who were enrolled between September 1995 and December 1995 (21). The 183 of these patients for whom followup data were available were included in the present analysis. Population 2 comprised 75 consecutive unrelated white Danish patients with early polyarthritis (61 women, 14 men) who were enrolled between June 1996 and March 1998 (19). Forty-six of these patients fulfilled the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) 1987 classification criteria for RA (26). Followup data were available for all 46 of these patients, and all were included in the present analysis. The study population in the present analysis thus comprised 229 patients who met the ACR criteria for RA. Both studies from which the patients originated were approved by local scientific ethics committees, and all patients had provided written informed consent. Independent of data from MBL and IgG-G0 analyses, information regarding traditional cardiac risk factors, medications, cardiovascular deaths, and hospital admissions was obtained from clinical charts and the National Board of Health.
Clinical and laboratory assessment
Clinical and laboratory variables assessed at the time of enrollment included present age, age at disease onset, C-reactive protein (CRP) level, serum IgM rheumatoid factor status, and functional ability assessed with the Health Assessment Questionnaire (HAQ) (27).
Traditional cardiac risk factors and medications
Clinical charts were used to collect information on traditional cardiac risk factors and treatment with methotrexate (MTX). Based on self-reports, cigarette smoking status was recorded as smoking (former and current) or not smoking (ever), and alcohol abuse was defined as >168 gm/week for women and >252 gm/week for men. Hypertension was defined as blood pressure readings ≥140/90 mm Hg alone or in combination with the need for antihypertensive agents during the period of followup. Patients were considered to have diabetes if the 1997 criteria of the American Diabetes Association were met (28) or clinical charts showed a history of diabetes and treatment with insulin or antidiabetic drugs. Based on the latest physician's assessment, nutrition status was recorded as poor (below normal), normal, or obese (beyond normal). Although we were able to retrieve information on treatment with MTX from clinical charts, use of various corticosteroids and nonsteroidal antiinflammatory drugs (NSAIDs) was not consistently recorded.
Adverse outcomes studied were ischemic heart disease including myocardial infarction, and death due to ischemic heart disease. Using the patient's Central Person Register (CPR) number, the National Patient Registry provided information on hospital admissions with primary discharge diagnoses of ischemic heart disease and myocardial infarction, and the Danish Registry of Causes of Death provided information on death due to ischemic heart disease from the time of enrollment to April 1, 2005. Adverse outcomes were coded according to International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes (I20–I25 for ischemic heart disease and I21–I22 for myocardial infarction).
The CPR number is a unique person identifier and has been assigned to all living Danish citizens since 1968. The National Patient Registry and the Danish Registry of Causes of Death are based on the CPR number. The National Patient Registry contains information on the CPR number, code for hospital and department, dates of admission and discharge, and codes for the primary and secondary diagnoses. The Danish Registry of Causes of Death contains information on the CPR number, date and place of death, and classification of underlying and contributory causes of death. The diagnosis of acute myocardial infarction is a validated event in these registries (29). To further validate the outcome variables in this study, all diagnoses were verified by review of the clinical charts.
Genotyping and detection of serum MBL
Assessments of MBL2 alleles and serum concentrations of MBL were carried out at the time of enrollment. MBL2 alleles were detected as previously described (8, 21). All patients were genotyped for MBL2 structural polymorphisms. Because of difficulties with polymerase chain reaction amplification, only 209 patients were genotyped for the promoter polymorphism at position –221 (X/X, X/Y, Y/Y) (21). Serum concentrations of MBL were measured in a double enzyme-linked immunosorbent assay using anti-MBL monoclonal antibody (30). Twenty-three sera were not included in this analysis because of technical difficulties (21).
Determination of aggregated IgG-G0
IgG-G0 analyses were available only in patients from population 1. Serum levels of IgG-G0 were measured indirectly by MBL binding to aggregated IgG-G0 (21).
We first performed a Cox proportional hazards regression analysis with adjustment for age and sex to examine the relationship between the presence of MBL genotype A/A, genotype YA/YA, high serum levels of MBL, and the risk of ischemic heart disease, myocardial infarction, and death due to ischemic heart disease. The relationship between high serum levels of MBL and risk of adverse cardiovascular outcomes was assessed using receiver operating curves. Second, using multivariate Cox proportional hazards regression analysis, we calculated the individual hazard ratios (HRs) for ischemic heart disease and myocardial infarction with respect to the following conditions at enrollment: serum MBL >3,000 μg/liter, age (per decade), plasma CRP >100 mg/dl, HAQ score above the median, presence of IgM rheumatoid factor, male sex, treatment with MTX, poor nutrition status, obesity, diabetes, hypertension, current or former smoking, and abuse of alcohol. This analysis was not performed for death due to ischemic heart disease since the relatively low number of events and large number of covariates resulted in extreme and uncertain estimates of the HRs. In the Cox regression analyses, we examined the time from enrollment to a particular event of interest. Data were censored at the time of death or the end of followup (whichever came first). Cumulative incidence estimates were plotted as a graphic representation of the risk of a particular event in the examined strata. All reported P values are 2-sided, and P values less than 0.05 were considered significant.
The median followup time in the 229 patients was 9.5 years, encompassing 1,799 patient-years. The median duration of disease at enrollment was 6.3 years. Basic demographic, clinical, and serologic characteristics of the patients are shown in Table 1.
Table 1. Demographic, clinical, and serologic characteristics of the 229 patients with rheumatoid arthritis*
Except where indicated otherwise, values are the number (%). CRP = C-reactive protein; HAQ = Health Assessment Questionnaire.
Data not available on all patients.
Deaths due to ischemic heart disease and hospital admissions with the primary diagnosis of ischemic heart disease. In case of hospital admission with ischemic heart disease followed by death due to ischemic heart disease, the episode was only counted once.
Deaths due to myocardial infarction and hospital admissions with the primary diagnosis of myocardial infarction. In case of hospital admission with myocardial infarction followed by death due to myocardial infarction, the episode was only counted once.
One hundred thirty-three patients (58%) had genotype A/A, 82 (36%) had genotype A/O, and 14 (6%) had genotype O/O. Of the 209 patients tested for MBL2 promoter allele X/Y, 139 (67%) were homozygous for the functional allele Y, and 70 (33%) had the extended genotype YA/YA. The serum concentration of MBL varied with the MBL2 genotype and promoter genotype, as shown in Figure 1. During followup, a total of 37 patients developed ischemic heart disease, 14 developed myocardial infarction, and 12 died of ischemic heart disease.
After adjustment for age and sex, the risk of ischemic heart disease, myocardial infarction, and death due to ischemic heart disease was greater among patients with the A/A genotype than among those with the A/O and O/O genotypes, but the differences were not significant (Table 2). However, patients with the extended genotype YA/YA had a significantly greater risk of myocardial infarction (HR 5.0, 95% confidence interval [95% CI] 1.4–17.5) and death due to ischemic heart disease (HR 4.1, 95% CI 1.2–14.3) compared with those with the non-YA/YA genotype (Table 2).
Table 2. Incidence of ischemic heart disease, myocardial infarction, and death due to ischemic heart disease according to MBL genotype, extended MBL genotype, and serum level of MBL in patients with rheumatoid arthritis*
Ischemic heart disease
Death due to ischemic heart disease
Values are the number (%). HR = hazard ratio; 95% CI = 95% confidence interval.
Comparing A/A with A/O plus O/O, adjusted for age and sex.
Comparing YA/YA with YA/XA plus XA/XA plus YA/O plus XA/O plus O/O, adjusted for age and sex.
Comparing serum mannose-binding lectin (MBL) >3,000 μg/liter with serum MBL ≤3,000 μg/liter, adjusted for age and sex.
Patients with the YA/YA genotype had the highest serum levels of MBL (Figure 1). Using receiver operating curve analysis, we found that the critical serum level of MBL with regard to the risk of adverse cardiovascular outcome was ∼3,000 μg/liter. After adjustment for age and sex, patients with high serum levels of MBL (>3,000 μg/liter) had a significantly greater risk of ischemic heart disease (HR 3.1, 95% CI 1.4–7.1), myocardial infarction (HR 6.3, 95% CI 2.0–20.1), and death due to ischemic heart disease (HR 10.5, 95% CI 2.7–41.3) compared with patients whose serum levels of MBL were ≤3,000 μg/liter. The increased risk of ischemic heart disease and myocardial infarction in patients with high serum levels of MBL is shown graphically in Figure 2.
The Cox regression analyses shown in Table 3 demonstrate that only serum MBL levels >3,000 μg/liter (HR as compared with serum-MBL ≤3,000 μg/liter 3.6, 95% CI 1.4–9.2), age at enrollment, and poor nutrition status were independently associated with increased risk of ischemic heart disease, and that only serum MBL levels >3,000 μg/liter (HR as compared with serum MBL ≤3,000 μg/liter 9.0, 95% CI 2.2–36.4) and age at enrollment were independently associated with increased risk of myocardial infarction.
Table 3. Risk of ischemic heart disease (n = 29) and myocardial infarction (n = 12) among 178 patients with rheumatoid arthritis*
Patients with incomplete information on risk factors (n = 51) were not included in the analysis. CRP = C-reactive protein; HAQ = Health Assessment Questionnaire; MTX = methotrexate (see Table 2 for other definitions).
HRs calculated by multivariate Cox regression analysis.
MBL >3,000 μg/liter
Age at enrollment (per decade)
CRP >100 mg/dl
IgM rheumatoid factor
HAQ above the median score
Nutrition status below normal
Nutrition status above normal
Current or former smoker
The median value of IgG-G0 was 436 AU (range 127–2,173). High levels of IgG-G0 (i.e., above the median) in combination with high serum levels of MBL conferred increased risk of ischemic heart disease (HR 9.2, 95% CI 2.7–31.9, P < 0.0001), myocardial infarction (HR 12.2, 95% CI 2.2–68.7, P = 0.005), and death due to ischemic heart disease (HR 46.7, 95% CI 3.4–646, P = 0.004) (Table 4). However, in patients with low levels of IgG-G0 there were no statistically significant associations between high serum levels of MBL and adverse cardiovascular outcomes (P > 0.7, P > 0.3, and P > 0.9 for ischemic heart disease, myocardial infarction, and death due to ischemic heart disease, respectively).
Table 4. Risk of ischemic heart disease, myocardial infarction, and death due to ischemic heart disease according to serum levels of MBL among patients with high and patients with low serum levels of IgG-G0*
Ischemic heart disease
Death due to ischemic heart disease
Values are the number (%). IgG-G0 = agalactosyl IgG (see Table 2 for other definitions).
Median value was 436 AU (range 127–2,173). High levels and low levels were defined as values above and below the median, respectively.
Comparing serum MBL >3,000 μg/liter with serum MBL ≤3,000 μg/liter, adjusted for age and sex.
We found in this study that high serum levels of MBL are a major risk factor for ischemic heart disease, including myocardial infarction, as well as death due to ischemic heart disease. The fact that similar results were obtained in patients with MBL2 polymorphisms resulting in high levels of MBL indicates that these findings are attributable to a mechanism linked to the MBL2 gene and do not represent an epiphenomenon such as an acute-phase reaction. Moreover, high serum levels of MBL were still found to be a major risk factor for ischemic heart disease overall and myocardial infarction in patients with RA after adjustment for CRP level, other disease markers, and traditional cardiac risk factors.
Besides ischemic heart disease, we also investigated development of myocardial infarction since this is a validated event in the registries of the Danish Board of Health (29). The annual rate of myocardial infarction was 0.78% overall (95% CI 0.43–1.3%), which is slightly lower than the 1.29% reported in a recent English study (1). As expected, ischemic heart disease and myocardial infarction were associated with increasing age. However, hypertension, smoking, diabetes, and obesity did not confer significant additional risk of ischemic heart disease or myocardial infarction. This parallels the findings of previous studies of RA, in which increased incidence of cardiovascular disease could not be attributed to traditional cardiac risk factors alone (3, 4). Unfortunately, we were unable to systematically account for lipid status and body mass index since this trial was not primarily designed as a study of cardiovascular outcomes.
Another limitation of this study was the lack of systematic recording of data on treatment with antirheumatic drugs other than MTX at enrollment. Several antirheumatic drugs may have effects on atherogenesis in RA. In a study by Choi et al, treatment with MTX provided substantial benefit in terms of reduction of cardiovascular mortality (6) and, although several pro-atherosclerotic effects of corticosteroids have been described, it has been suggested that their antiinflammatory effects may balance out their adverse cardiovascular effects, thus reducing the risk of accelerated atherosclerosis in corticosteroid-treated patients with RA (31). In a recent study, relative risk estimates for myocardial infarction were shown to be elevated among current and new users of all classes of non-aspirin NSAIDs (32). However, this finding has not been evaluated specifically in patients with RA. Treatment with anti–tumor necrosis factor α (anti-TNFα) drugs may have a significant effect on endothelial function in patients with RA (33); however, TNFα inhibitors may also worsen congestive heart failure (34).
In RA, systemic inflammation and disease severity are associated with increased risk of cardiovascular disease (4, 5). In our study, CRP levels >100 mg/dl and high HAQ scores did not confer additional risk for ischemic heart disease. However, these measures were obtained only at enrollment and thereby represent a one-time picture of disease activity from 10 years prior to the present study. In this respect, it would have been of great use to have serial radiographic data on the patients in order to evaluate the cumulative inflammationdriven damage. However, this was not feasible due to varying study designs in the original trials. Moreover, in this study IgM rheumatoid factor was not associated with ischemic heart disease. As previously reported (35), we found poor nutrition status to be a major risk factor for ischemic heart disease in RA.
MBL may enhance inflammation-mediated tissue injury during postischemic reperfusion by complement activation (14–16). MBL has also been suggested to be a mediator in the pathogenesis of microvascular complications in type 1 diabetes (36) as well as graft survival after kidney transplantation (37). Recently, it was shown that homozygous carriers of the normal MBL2 genotype had the highest rate of early restenosis after carotid eversion endarterectomy, compared with patients with genotypes encoding low serum MBL levels (17). Taken together, these observations indicate an important adverse effect of MBL and the lectin pathway of complement activation on cardiovascular disease pathology. Terminal complement complexes have been associated with endothelial cell damage in rheumatoid nodules, indicating that complement activity may play a role in vascular damage in RA (38). It is likely that such endothelial cell damage would lead to the endothelial dysfunction that has been described in RA (39, 40), which again is predictive of the development of coronary artery disease (41).
It is of particular interest that MBL, among other factors, binds N-acetylglucosamine and damaged host cells, exposing nucleic acids (42–44). Oxidative stress seems to expose structures on endothelial cells that promote binding of MBL (14). Thus, one explanation for our observations could be that MBL amplifies the inflammatory reaction in vessels through complement, by its interaction with damaged endothelial cells. Another likely explanation derives from the observation that oligosaccharides in IgG from patients with RA contain increased levels of IgG-G0 glycoforms (22;23): MBL binds to aggregated IgG-G0 (21), and the MBL–aggregated IgG-G0 complexes may activate complement and initiate an inflammatory reaction (25). The latter two possible explanations are not mutually exclusive.
Our observation that high serum levels of MBL conferred increased risk of ischemic heart disease, including myocardial infarction, and death from ischemic heart disease only in patients with high levels of IgGG0 is in accordance with the hypothesis that complexes of MBL and IgG-G0 are harmful in RA (25). Thus, it may be hypothesized that complexes of MBL and IgG-G0 deposited in the vessel wall could participate in an inflammation-driven pro-atherogenic process. This hypothesis also implies that pharmacologic intervention targeted to MBL itself or to downstream processes could be a potential treatment option in selected patient groups.
In contrast to these observations, MBL variant alleles resulting in dysfunctional MBL or lack of MBL have been reported to be associated with accelerated atherosclerosis and cardiovascular occlusion in other patient groups (9–13). This paradox is probably a key issue in the understanding of the role of MBL in human disease. Thus, it is likely that a fine balance in the vessel wall determines whether MBL may be advantageous or disadvantageous. If MBL mediates excessive complement activation, the process may become detrimental. However, under conditions that allow the occurrence of the physiologic opsonic function of MBL, which is advantageous, lack of MBL may increase inflammation due to dysfunctional sequestration of harmful material from the vessel wall. Observations in our RA and systemic lupus erythematosus cohorts (12) may be good examples of such differential effects. The relatively high incidence of MBL haplotypes that specify low-level and dysfunctional MBL in many ethnic groups suggests that, under certain circumstances, a relative lack of MBL may offer advantages for the host, which could include protection against vascular complications, and that the MBL polymorphisms represent a balanced genetic system (30, 45, 46).
In summary, we have shown that genetically determined high levels of MBL confer a substantially increased risk of ischemic heart disease in RA patients and that this association leads to increased risk of myocardial infarction and premature death. Moreover, our results provide, for the first time, in vivo evidence in support of the notion that the presence of complexes of MBL and IgG-G0 represents a significant threat in vivo for RA patients. Since IgG-G0 may also be increased in several inflammatory (47, 48) and infectious (49) disease settings, future studies will hopefully illuminate whether the present findings may be applicable beyond RA.
Dr. Troelsen 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. Troelsen, Garred, and Jacobsen.
Acquisition of data. Drs. Troelsen, Garred, and Madsen.
Analysis and interpretation of data. Drs. Troelsen, Garred, Madsen, and Jacobsen.
Manuscript preparation. Drs. Troelsen, Garred, Madsen, and Jacobsen.
Statistical analysis. Drs. Troelsen, Garred, and Jacobsen.
ROLE OF THE STUDY SPONSORS
The funding organizations had no role in study design, data collection, data analysis, data interpretation, or writing of the report.