|Home | About | Journals | Submit | Contact Us | Français|
We sought to determine the magnitude of the association between mitral annular calcification (MAC) and vascular events in a multiethnic cohort.
MAC is common in the elderly, and is associated with atherosclerotic risk factors. Its impact on the risk of cardiovascular events is controversial.
The study cohort consisted of 1,955 subjects, aged ≥40 years, and free of prior myocardial infarction (MI) and ischemic stroke (IS). MAC was assessed by transthoracic 2D echocardiography. The association between MAC and MI, IS, and vascular death (VD) was examined by Cox proportional hazard models with adjustment for established cardiovascular risk factors. The effect of MAC thickness was also analyzed.
The mean age of the cohort was 68.0 ± 9.7 years and the majority of subjects were Hispanics (56.8%). 519 subjects (26.6%) had MAC. Of 498 patients with MAC thickness measurement available, 253 (13.1%) had mild to moderate MAC (1–4mm) and 245 (12.7%) severe MAC (≥4mm). During a mean follow-up of 7.4 ± 2.5 years, MI occurred in 100 (5.1%) subjects, IS in 104 (5.3%) subjects, and VD in 155 (8.0%) subjects. After adjustment for other cardiovascular risk factors, MAC was associated with an increased risk of MI (adjusted hazards ratio [HR] 1.75; 95% confidence interval [CI] 1.13–2.69: p=0.011) and VD (adjusted HR 1.53; 95%CI 1.09–2.15: p=0.015), but not IS (adjusted HR 1.34; 95%CI 0.87–2.05: p=0.18). Further analysis revealed that the impact of MAC was related to its thickness, with MAC >4mm being a strong and independent predictor of MI (adjusted HR 1.89: 95%CI 1.13–3.17: p=0.008) and VD (adjusted HR 1.81: 95%CI 1.21–2.72: p=0.002), and showing borderline association with IS (adjusted HR 1.59: 95%CI 0.95–2.67: p=0.084).
In this multiethnic cohort, MAC was a strong and independent predictor of cardiovascular events, especially MI and VD. The risk increase was directly related to MAC severity.
Mitral annular calcification (MAC) is a chronic process involving fibrosis and calcification of the mitral valve support ring. The prevalence of MAC has been reported to be as high as 15% in population-based studies,1 and up to 35% in patients with severe coronary artery disease.2 MAC has been associated with a high prevalence of risk factors for the development of atherosclerosis.3 Furthermore, association with clinical vascular events such as ischemic stroke (IS) has been reported in multiple community cohorts, although the incremental predictive value of MAC above other established risk markers has been questioned.4–6 The association of MAC with coronary heart disease (CHD) has been less well demonstrated In the Framingham database, 7 only modest association was observed with the combined outcome of myocardial infarction (MI), unstable angina, congestive heart failure, and non-hemorrhagic stroke
We analyzed the relationship between MAC and the risk of MI, IS, and vascular death (VD) in the multiethnic population of the Northern Manhattan Study (NOMAS), adjusting for the effect of traditional cardiovascular risk factors.
Subjects were participants in NOMAS, a population-based prospective cohort study in Northern Manhattan, NY. The methods of subject recruitment and enrollment into NOMAS have been described elsewhere.8,9 Briefly, random digit dialing was performed, and community participants were enrolled. NOMAS entry criteria included (1) age >39 years, (2) residence in Northern Manhattan for at least 3 months, and (3) no prior diagnosis of stroke. Subjects with prior myocardial infarction at enrollment were also excluded from this analysis. The study was approved by the Institutional Review Board at Columbia University Medical Center (CUMC).
Transthoracic echocardiograms were obtained in the 1,955 subjects between 1993 and 2001. Sstudies were performed and measurements taken according to the recommendations of the American Society of Echocardiography.10 Interpretation of echocardiographic studies was performed blinded of clinical and demographic characteristics. Inter-observer reliability was periodically assessed by use of intraclass correlation coefficients for the variables measured, which ranged between 0.59 and 0.74. MAC was defined as an intense echocardiographic-producing structure with highly reflective characteristics that was located at the junction of the atrioventricular groove and the posterior or anterior mitral leaflet on the parasternal long-axis, apical 4-chamber or 2-chamber, or parasternal short-axis view. The severity of MAC, expressed as maximal thickness in millimeters, was measured from the leading anterior to the trailing posterior edge at its greatest width. Calcification thickness greater than 1mm and less than 4mm was considered mild to moderate, and greater than 4mm was considered severe.
Baseline evaluation was performed at enrollment as previously reported.8 Hypertensive status was defined as a systolic blood pressure (SBP) recording ≥40 mm Hg or a diastolic blood pressure (DBP) recording ≥90 mm Hg based on the mean of two measurements, a patient’s self-reported history of hypertension, or antihypertensive treatment. Diabetes mellitus was defined by a patient’s self-report, insulin use, oral hypoglycemic use, or a fasting glucose ≥126 mg/dL. Hypercholesterolemia was defined as total serum cholesterol >240 mg/dL, a patient’s self-report or the presence of lipid-lowering treatment. The presence of atrial fibrillation was based on a current or past ECG. Body mass index (BMI) was calculated as weight (kilograms) divided by height (meters) squared. Smoking was defined as current cigar or cigarette smoking. Alcohol consumption was defined as lifetime drinking of >1 drink per month.
All subjects were followed at 6 months and then annually. In-person follow-up visits were conducted at the medical center and included interview, vital signs, physical and neurological examination. Over 80% of patients with MI or IS in northern Manhattan are hospitalized at CUMC. Subjects hospitalized at other local hospitals were identified through active hospital surveillance of admission and discharge in accordance with the International Classification of Diseases, 9th Revision codes, and through local physicians. All outcome events were reviewed by a specially trained research assistant, and reported to a study physician for adjudication.
MI was defined by criteria adapted from the Cardiac Arrhythmia Suppression Trial 11 and the Lipid Research Clinics Coronary Primary Prevention Trial, 12 requiring at least 2 of 3 following: (1) cardiac pain determined to be typical angina; (2) cardiac enzyme abnormalities in creatine phosphokinase MB isoenzyme (CPK-MB) fraction or troponin I values; and/or (3) ischemic EKG abnormalities.
Stroke was defined as the first symptomatic occurrence of any type of stroke, including intracerebral hemorrhage, subarachnoid hemorrhage, and cerebral infarction, based on the World Health Organization criteria.13 Only IS, defined by TOAST criteria 14, was considered for this report. The vast majority (70%) of IS cases were hospitalized at the CUMC, allowing access to information on clinical syndrome, blood test and imaging studies. The presence of IS was determined by two neurologists independently, and the Principal Investigator of NOMAS (R.L.S.) adjudicated any disagreements.
For subjects who died, deaths were classified as vascular or nonvascular based on information obtained from the family, medical records, and death certificate. Causes of death were also validated by a study physician. Vascular causes of death included stroke, MI, heart failure, and cardiac arrhythmia (e.g., sudden or unwitnessed death).
For principal analyses, MAC was both dichotomized (present/absent) and examined as a continuous variable. MAC was then categorized for further analysis into mild to moderate (1–4mm) or severe (more than 4mm), with 4mm being the median value in subjects with MAC. We also performed an additional quantitative analysis with MAC categorized by the 75 percentile. The distribution of demographics and vascular risk factors was evaluated in the total cohort and in subjects with and without MAC. Comparisons were made using t-tests for continuous variables and χ2 tests for categorical variables. Time to cardiovascular events was illustrated with Kaplan-Meier curves and log-rank test was used to compare different MAC subsets. Cox proportional hazard models were used to identify the risk factors of cardiovascular events, and hazard ratios (HRs) and 95% CIs of MAC were evaluated. Variables with association at the p < 0.1 level in univariable models were included in the multivariable models. An additional model was generated to adjust for additional variables such as higher education (higher than high school), work status (over 10hr/week), white blood cell count and echocardiographically derived left ventricular mass index. We tested for interactions between MAC and significant covariates. Statistical analyses were conducted using SAS 9.1 software (SAS Institute, Cary, NC)
Of 1,955 subjects, 519 (26.6%) had MAC. The majority of the subjects were Hispanic (56.8%), followed by black (22.7%) and white (20.5%). The cohort was predominantly elderly (mean age 68.0 ± 9.7 years), female (61.4%), and hypertensive (67.3%). The average LVM in our cohort was 174.9±57.4g. Other baseline cohort characteristics are summarized in Table 1. Antiplatelet and anticoagulant agents were more frequently used in subjects with MAC compared with subjects without it (27.9% vs. 20.9%, p=0.001, and 3.5% vs. 1.9% p=0.06, respectively). No difference in the use of cholesterol lowering agents were noted (11.8% vs. 12.7%, p=0.58).
Subjects were followed for 7.4 ± 2.5 years. At follow-up, there were 100 MI (50[3.5%] vs. 50 [9.6%], no MAC vs. MAC, p<0.001), 104 IS (65 [4.5%] vs. 39 [7.8%], p=0.004), and 155 VD (81 [5.6% vs. 74 [14.7%], p<0.001). The incidence of MI, IS and VD was 13.8, 10.7, and 20.3/1000 person-years respectively in subjects with MAC, and 4.7, 6.1, and 7.5/1000 person-years respectively in subjects without MAC (p<0.001 for MI, p=0.001 for IS, p<0.001 for VD). The event-free Kaplan-Meier curves for outcomes based on MAC presence and thickness are shown in Figure 1A–C
Variables associated with each clinical outcome in univariable analysis are listed in Table 2. MAC, age, diabetes and hypertension were associated with risk of all pre-defined clinical outcomes (MI, IS and VD). After adjustment for other risk factors, MAC was associated with an increased risk of MI (p=0.011) and VD (p=0.015), but not IS (p=0.18).
MAC was also associated with an increased risk of composite endpoint of death and MI (univariable HR=2.28, 95% CI 1.89–2.75, p<0.0001, multivariable HR=1.44, 95% CI (1.18–1.77), p=0.0004) as well as death and IS (univariable HR=2.11, 95% CI 1.75–2.54, p<0.0001, multivariable HR=1.38, 95% CI (1.13–1.69), p=0.002). The results were essentially unchanged when MAC was categorized by the 75 percentile (n=133; HR=1.55, 95% CI 1.17–2.06, p=0.002 for MI and death, and HR=1.50, 95% CI 1.13–2.00, p=0.005 for IS and death). Further, when an even more comprehensive set of confounding variables (eg. education higher than high school, work status over 10hr/week, white blood cell counts and echocardiographically derived left ventricular mass index) was added to the model, MAC was consistently associated with these endpoints (HR=1.32, 95% CI 1.06–1.65, p=0.012 for MI and death, HR=1.30, 95% CI 1.05–1.61, p=0.016 for IS and death, and HR=1.48, 95% CI 1.02–2.13, p=0.024 for VD).
When MAC was categorized into mild to moderate (1–4mm: n=253; 13.1%) or severe MAC (more than 4mm: n=245; 12.7%) mild to moderate MAC was a significant predictor for MI and VD, but the association was no longer significant after adjusting for significant clinical variables (Table 3). Severe MAC, however, was a strong predictor of all pre-defined clinical outcome (MI, IS and VD) and its association with MI and VD remained significant after adjustment. The association of severe MAC with IS was of borderline significance (p=0.08).
MAC was related to our pre-defined vascular clinical events including MI and VD in our multiethnic population-based cohort, but the association was not statistically significant for IS after adjustment for atherosclerotic risk factors. The impact of MAC was related to its thickness, with MAC >4mm being a strong and independent predictor of MI and VD, and borderline for IS. This dose-response relationship between MAC severity and outcome further strengthens the significance of the finding.
Several explanations may account for the predictive power of MAC for vascular clinical events. MAC has been associated with hypertension, left atrial enlargement and atrial fibrillation.1–3 All of these do have predictive power for the specified outcomes. Thus, although there have been occasional reports of embolic calcification causing stroke,5 it is more likely that MAC represents a marker of risk rather than a causative factor. MAC could be an indicator of severity and duration of hypertension (similar to left atrial size) or hypercholesterolemia, or both.
Similarities exist between atherosclerosis in the vasculature and chronic degenerative changes in valvular structures although valve calcification is a more amorphous, disorganized process.15. Risk factors are common to both conditions. The attachment of the mitral valve to the annulus is a site of turbulence, and inflammation tends to initiate early in the subclinical atherosclerotic phase in this area.3 The calcifying process of a valve is initially characterized by macrophage and T-cell infiltration in response to endothelial injury.16 The burden of atherosclerotic risk factors is likely reflected on the thickness of calcification, and MAC may be an accurate marker that reflects cumulative exposure to vascular metabolic and mechanical stresses.
Our study demonstrated a strong and independent association of MAC with MI. This is in partial agreement with a previously published study from the Framingham cohort7. In that study, conducted in a predominantly white population, MAC was associated with the combined outcome of MI, unstable angina, congestive heart failure, and non-hemorrhagic stroke, but not with MI alone after adjustment for baseline covariates. This difference with our study might be explained in part by different participant characteristics, including the older age of our cohort and its tri-ethnic race-ethnicity composition. Moreover, different echocardiographic methodologies used (M-mode evaluation of MAC in the Framingham cohort compared with 2-dimensional technique in the present study) may also contribute to these differences. Two-dimensional evaluation provides better analysis of the spatial location and magnitude of MAC, and our definition of MAC (as any calcification greater than 1 mm in multiplane 2D assessment), along with the older population, have lead to its higher prevalence (25%) compared with the Framingham cohort(15%).1,7
Further, our categorized multivariable analysis clearly revealed that it is the severity of MAC, rather than its presence, that drives the association with vascular events, but the available data are insufficient to suggest preventative strategies, such as the use of antiplatelet or anticoagulation therapy, in patients with MAC. As mentioned earlier, further investigation is needed to define whether the use of medical treatment to affect MAC might affect clinical outcome.
In our study, MAC was marginally associated with IS risk, and only when MAC thickness was at least 4mm. Multiple epidemiological studies have analyzed stroke risk in patients with MAC, but few adjusted for possible clinical confounders.17 The incremental predictive value of MAC over other established risk factors remains poorly defined. In the Stroke Prevention in Atrial Fibrillation (SPAF) study, MAC was no longer associated with stroke when patients with non-rheumatic atrial fibrillation were excluded from the analysis.18 MAC was associated with a relative risk of stroke of 2.1 in the Framingham Study after adjustment for clinical risk factors, and after exclusion of atrial fibrillation.4 However, by limiting outcomes to cerebral infarcts instead of including all incident strokes, the relative risk was reduced to a non-significant 1.78. Therefore, our results are consistent with previous reports, but underscore the importance the quantitative measurement of MAC, which may be a better risk indicator than the mere presence of MAC.
Renal failure, even in mild degree may be a significant confounding factor for the studied outcome variables. Our cohort did include 80 subjects with mild renal insufficiency (defined as serum creatinine above 1.5 mg/dl), but the exclusion of these subjects did not change the main results of our study (results not shown).
Lastly, novel inflammatory markers (eg. C-reactive protein) were available only in a subset of our cohort, and could not be considered in the analysis. Further study is needed to examine the relationship between systemic biomarkers of inflammation, valvular or annular calcification and vascular outcomes.