Subclinical Atherosclerotic Cardiovascular Disease and Early Age-Related Macular Degeneration in a Multiracial Cohort: Title and subTitle BreakThe Multiethnic Study of Atherosclerosis FREE

Despite new medical and surgical interventions, age-related macular degeneration (AMD) remains an important cause of vision loss in the United States.¹ Its pathogenesis remains poorly understood.^{2 - 3} Atherosclerosis has long been postulated to be associated with late stages of AMD through its effect on the choroidal circulation and possible deposition of lipids in the Bruch membrane.^{4 - 8} However, the associations of atherosclerotic cardiovascular disease (CVD) and its risk factors (eg, high blood pressure or high serum lipid levels) with the development of AMD have been inconsistent.⁹

Data from most population-based studies have not shown associations of clinical atherosclerotic CVD, as manifested by myocardial infarction or stroke, with AMD.^{10 - 19} Furthermore, high serum lipid levels and the use of lipid-lowering agents have not been consistently associated with the presence or absence of AMD.^{20 - 25} Few reports are available regarding the relation of signs of subclinical CVD with AMD.^{13 ,19 ,26} Data from the Rotterdam Eye Study showed that persons with subclinical CVD, as manifested by carotid artery plaques, were nearly 5 times as likely to have late AMD compared with those without such plaques.¹³ In addition, greater intima-media thickness (IMT) of the carotid artery and the presence of aortic calcification, both of which are measures of subclinical atherosclerosis, were also found to be associated with increased risk of incident AMD.¹³ Higher pulse pressure, a measure of decreased arterial distensibility that results from higher systolic and lower diastolic blood pressure, has been found to predict the incidence of exudative AMD in another population-based cohort.²⁷ This has not been found in other studies.¹¹

The purpose of this report is to evaluate the association of subclinical atherosclerotic CVD, serum lipid levels and lipid-lowering agents, and measures of arterial distensibility with AMD in the persons who participated in the large Multiethnic Study of Atherosclerosis (MESA). We hypothesized that these characteristics would be associated with early AMD in the cohort. We did not hypothesize any specific racial/ethnic differences for specific factors, although we thought that we might find some (eg, higher frequency of subclinical atherosclerosis and CVD risk factors in white individuals than in black individuals might result in higher frequency of early AMD in white individuals compared with black individuals).

Objectives of MESA, recruitment and participation of the cohort, and methods used have been described in detail elsewhere.²⁸ In brief, MESA is a longitudinal study supported by the National Heart, Lung, and Blood Institute with the goals of identifying risk factors for prevalence and progression of subclinical atherosclerosis in 4 different racial/ethnic groups.²⁸ The MESA cohort includes 6814 men and women aged 44 to 84 years at study entry with no history of clinical CVD (eg, no myocardial infarction, angina, stroke, or any medical condition that would prevent long-term participation) who were recruited from 6 field centers: Baltimore, Md; Chicago, Ill; Forsyth County, NC; Los Angeles, Calif; New York, NY; and St Paul, Minn. During the initial MESA visit (July 17, 2000, through August 29, 2002), data on an array of phenotypic measures of subclinical CVD were collected; fundus photography was performed (September 9, 2002, through February 7, 2004) for the first time at the second examination in 6176 participants. Tenets of the Declaration of Helsinki were followed, and institutional review board approval was granted. Written informed consent was obtained from each participant.

Fundus photography was performed using a standardized protocol that has been described in detail elsewhere.^{29 - 30} In brief, both eyes of each participant were photographed in a similar fashion using a 45° 6.3-megapixel digital nonmydriatic camera (Canon USA, Lake Success, NY). Two photographic fields were taken of each eye; the first centered on the optic disc and the second centered on the fovea.

Images were assessed for AMD at the University of Wisconsin according to standardized protocols described in detail elsewhere.³¹ In brief, among the AMD features evaluated within a circle with a radius of 3450 μm were drusen size, type, and area; increased retinal pigment; retinal pigment epithelial (RPE) depigmentation; pure geographic atrophy; and signs of exudative macular degeneration (eg, subretinal hemorrhage, subretinal fibrous scar, RPE detachment, and/or serous detachment of the sensory retina or laser or photodynamic treatment for neovascular AMD). For grading, a grid that consisted of 2 circles concentric with the center of the macula and 4 radial lines was superimposed over the image.

Graders were masked with respect to information about the participant. Each image was graded twice (a preliminary and a detail grade) using a modification of the Wisconsin Age-Related Maculopathy Grading System.^{29 ,31} Every digitized image was graded using the full complement of image enhancement tools (eg, magnification, contrast enhancement, lightening, and red free [Digital Healthcare Image Management Systems, Cambridge, England]) according to preset protocols. However, final scoring of an AMD lesion required its appearance on the nonenhanced image. For the purposes of this report, data from 5887 (98.9%) of the 6176 participants photographed, those in whom at least 1 eye could be evaluated for AMD, are included in the analyses.

Soft distinct drusen were defined by size (between 63 and 300 μm in diameter) and appearance (sharp margins and a round nodular appearance with a uniform density [color] from center to periphery). Soft indistinct drusen were the same size as the soft distinct drusen but had indistinct margins and a softer, less solid appearance. Increased retinal pigment appears as a deposition of granules or clumps of gray or black pigment in or beneath the retina. Retinal pigment epithelium depigmentation is characterized by faint grayish yellow or pinkish yellow areas of varying density and configuration without sharply defined borders. Early AMD was defined by the presence of either soft drusen (distinct or indistinct) and pigmentary abnormalities (increased retinal pigment or RPE depigmentation) or large soft drusen 125 μm or greater in diameter with a large drusen area (>500 μm in diameter circle) or large soft instinct drusen ≥125 μm in diameter in the absence of signs of late AMD. Late AMD was defined by the presence of any of the following: geographic atrophy or pigment epithelial detachment, subretinal hemorrhage or visible subretinal new vessels, subretinal fibrous scar, or laser treatment scar for AMD.

All systemic characteristics were determined at the baseline visit with an average of approximately 2 years between visits. Ankle-brachial index (ABI), a measure of peripheral atherosclerosis, was determined by measuring blood pressure with a Doppler probe in the brachial, dorsalis pedis, and posterior tibial arteries.³² The ABI, a ratio of ankle to brachial blood pressure, was computed separately for each leg, with the numerator the higher of the posterior tibial or dorsalis pedis systolic pressures and the denominator the average of the right and left brachial systolic pressures. It was defined as abnormal if the ABI was less than 0.9 on either side. Ultrasound imaging of the carotid arteries was performed using a scanner (General Electric Medical Systems, Milwaukee, Wis). Videotaped scans were interpreted centrally. Stenosis was estimated from Doppler recordings of peak systolic velocities and analyzed categorically. The IMT was measured between the lumen-intima and media-adventitia interfaces of the near and far walls of the common carotid artery (the 1-cm segment proximal to the bifurcation) and the internal carotid artery (including the bifurcation and 1 cm distal to the bifurcation). A maximum IMT for each of these 2 segments was standardized (by subtraction of the MESA population mean and division by its standard deviation), and the mean of the standardized IMT for the common and the internal carotid maxima was used in the analysis. A previous study³³ with the same protocol showed this standardized index to be more closely associated with incident coronary heart disease than alternative IMT indices. Minimal and moderate subclinical atherosclerotic CVD was defined by baseline values of the IMT index of less than or greater than the sex-specific 75th percentile values, respectively, with no abnormal ABI and no carotid stenosis exceeding 50%. The presence of abnormal ABI or carotid stenosis greater than 50% was classified as more severe disease. Plaques were classified by their sonographic density relative to the surrounding arterial wall as echolucent, isodense, or echogenic (including calcified when acoustic shadowing was observed).³⁴ Persons with more than 1 plaque were characterized by the plaque of greatest density. Computed tomographic scanning of the chest was performed either with an electrocardiograph-triggered (at 80% of the RR interval) electron-beam computed tomography scanner (Chicago, Los Angeles, and New York field centers; Imatron C-150, Imatron, San Francisco, Calif)³⁵ or with prospectively electrocardiograph-triggered scan acquisition at 50% of the RR interval with a multidetector computed tomography system³⁵ that acquired 4 simultaneous 2.5-mm slices for each cardiac cycle in a sequential or axial scan mode (Baltimore, Forsyth County, and St Paul field centers; Volume Zoom; Siemens, Forchheim, Germany, or Lightspeed, General Electric Medical Systems). Each participant underwent scanning twice. Scans were read centrally at the Harbor-University of California, Los Angeles, Research and Education Institute to identify and quantify coronary calcification. Agatston calcium scores among scanning centers and among participants were adjusted with a standard calcium phantom scanned simultaneously with the participant. The average Agatston calcium score was used in all analyses.³⁵ More than 50% of the population had a score equal to 0. In persons whose score was greater than 0, we arbitrarily chose to break these data into tertiles so that we had a total of 4 categories. We also investigated the Agatston calcium score as a continuous variable by looking at the log calcium score and including an indicator variable for whether the score was greater than 0 in the models.

Common carotid arterial distensibility was assessed as a ratio of diameter change over the cardiac cycle to brachial artery pulse pressure:

Distensibility = [2(D_s − D_d)/D_s]/(P_s − P_d).

The Young modulus, an index of arterial stiffness, corrects for arterial dimensions by including a wall thickness term:

The Young modulus = [D_a² (P_s − P_d)]/2h(D_s − D_d).

For each measure, D_s is defined as systolic carotid artery diameter, D_das diastolic carotid diameter, D_aas mean diameter, P_s as systolic blood pressure, P_d as diastolic blood pressure, and h as carotid IMT.³⁶ Stiffer arteries have low distensibility and high Young modulus values. The natural log of each index was approximately normally distributed and was used in analysis. Resting blood pressure was measured 3 times with participants in the seated position with a Dinamap model Pro 100 automated oscillometric sphygmomanometer (Critikon, Inc, Tampa, Fla). The average of the last 2 measurements was used in the analysis. The diagnosis of hypertension was based on self-reported treatment for hypertension with 1 of 6 common classes of antihypertensive medications or a systolic blood pressure of 140 mm Hg or higher or a diastolic blood pressure of 90 mm Hg or higher. Pulse pressure was defined as systolic minus diastolic blood pressure.

Fasting blood samples were drawn, processed, and stored using standardized procedures.³⁷ Serum total cholesterol, high-density lipoprotein cholesterol (HDL-C), and triglyceride levels were measured at the Collaborative Studies Clinical Laboratory at Fairview-University Medical Center (Minneapolis, Minn). Serum total cholesterol and HDL-C levels were measured using a cholesterol oxidase method (Roche Diagnostics, Indianapolis, Ind), the HDL-C level was measured after precipitation of non–HDL-C with magnesium-dextran, and triglyceride levels were measured using Triglyceride GB reagent (Roche Diagnostics). All lipid measurements were performed on the Roche COBAS FARA centrifugal analyzer. The laboratory analytical coefficients of variation for total cholesterol, HDL-C, and triglyceride levels were 1.6%, 2.9%, and 4.0%, respectively. The low-density lipoprotein cholesterol (LDL-C) level was calculated in plasma specimens with a triglyceride value less than 400 mg/dL using the formula of Friedenwald et al.³⁸ Levels of oxidized LDL-C were measured in Leuven, Belgium, with a monoclonal antibody 4E6–based competition enzyme-linked immunosorbent assay.^{39 - 41}

Standardized questionnaires were used to obtain information about level of education, annual household income, smoking history, and medication use for high blood pressure, high cholesterol level, or diabetes. Participants who had smoked in the last 30 days were considered current smokers. Diabetes was classified on the basis of a fasting glucose level of 126 mg/dL (7 mmol/L) or higher, use of insulin or oral diabetes medications, or history of physician-diagnosed diabetes. Height and weight were measured with participants wearing light clothing and no shoes. Body mass index was calculated as weight in kilograms divided by the square of height in meters. Statin and antihypertensive medication use was based on self-report. Education was defined in 5 categories: less than high school, high school, some college (includes technical and associate's degrees), bachelor's degree, and graduate school.

Logistic regression was used to predict AMD outcomes. Each atherosclerotic characteristic was entered into a separate model while adjusting for age (in 4 groups: 44-54, 55-64, 65-74, and 75-84 years), sex, racial/ethnic group, and study site. Odds ratios (ORs) for continuous variables are presented per standard deviation increase. Several of the variables (eg, the Young modulus) were log transformed to minimize skewness in the data. Tests of trend were performed by including the categorical variable as a continuous ordered factor in the logistic regression models. Interactions between adjustment variables (eg, race) and risk factors were tested by including an interaction term in the logistic regression models. Models in which a significant relation was found were additionally adjusted for educational level, smoking status, alcohol use, diabetes status, and body mass index to examine if the risk factor was independent of these other potentially confounding variables. We used SAS statistical software, version 9.1 (SAS Institute Inc, Cary, NC) for all analyses.

Those who participated in the follow-up examination and had fundus digital images gradable for AMD were younger than those who did not (Table 1). After we controlled for age and sex, participants at the second examination with gradable fundus photographs were more likely to be white. After we controlled for age and race/ethnicity, they were more likely to be male, and after we controlled for age, sex, and race/ethnicity, they were more likely to have more education, to not have carotid stenosis on ultrasound, and to have an Agatston score of 0 and were less likely to have a history of diabetes, to have ever smoked, to be using diuretics, to have an abnormal ABI, to have subclinical atherosclerotic CVD, to have a lower pulse pressure, and to have a smaller carotid IMT than those without gradable fundus digital images at the second examination (Table 1). Early AMD was present in 3.8%, late AMD in 0.5%, large drusen in 9%, and soft drusen in 15% of the cohort. The prevalence of early AMD was 4.8%, 2.1%, 4.0%, and 3.6% and the prevalence of late AMD was 0.6%, 0.3%, 0.2%, and 1.0% in white, black, Hispanic, and Chinese participants, respectively.

Table 2 provides the ORs, 95% confidence intervals (CIs), and P values for the relationship between early AMD and CVD risk factors and measures of subclinical CVD after controlling for age, sex, race/ethnicity, and study site. Only 2 statistically significant associations between factors studied and early AMD were found in the whole cohort: a positive association of serum HDL-C level and an inverse association of echolucent plaque density with early AMD (Table 2). Increasing levels of serum HDL-C were related to the presence of large retinal drusen, soft drusen, and large drusen area (Table 3). The presence of echolucent carotid artery plaques was related to lower odds of large drusen and soft drusen. Other relationships with specific early AMD lesions are presented in Table 3 and include an inverse association of statin use with large drusen and large drusen area, an inverse association of serum triglyceride levels with large drusen area, and more severe Agatston calcium score with soft drusen. None of the risk factors were associated with increased retinal pigment or RPE depigmentation.

In models for early AMD, statistically significant interactions were found between race/ethnicity and mean IMT, subclinical atherosclerotic CVD, increasing severity of maximum carotid artery stenosis, and Agatston calcium score (Figure). Greater mean IMT was directly associated with early AMD in white participants (OR per 0.83 [1 SD] μm, 1.23; 95% CI, 1.03-1.48; P = .02) and black participants (OR, 1.53; 95% CI, 1.13-2.06; P = .006) but inversely associated in Hispanic participants (OR, 0.44; 95% CI, 0.28-0.69; P<.001). These associations of early AMD and IMT were similar for the internal and common carotid arteries (data not shown). For increasing severity of maximum carotid stenosis graded, decreased odds were found for early AMD in Hispanic participants (OR per increasing category increase, 0.39; 95% CI, 0.23-0.66; P<.001), but no association was found in the other 3 ethnic groups (Figure). Similarly, more severe subclinical atherosclerotic CVD was associated with increasing odds of early AMD in black participants (OR, 1.95; 95% CI, 1.17-3.23; P = .01) and decreasing odds in Hispanic participants (OR, 0.39; 95% CI, 0.19-0.81; P = .01). Also, for increasing Agatston calcium score, higher odds for early AMD were found in white participants (OR per increasing category, 1.22; 95% CI, 1.01-1.47; P = .04) and nonsignificant decreasing ORs for the other racial/ethnic groups (P value for interaction = .02). A higher serum triglyceride level was not statistically significantly associated with early AMD in white participants (OR per 88 mg/dL, 1.07; 95% CI, 0.91-1.27; P = .43) but was protective in Chinese (OR, 0.43; 95% CI, 0.20-0.91; P = .03) and Hispanic (OR, 0.69; 95% CI, 0.47-1.01; P = .06) participants. These associations were mainly due to relationships with large soft drusen (data not shown). No other statistically significant interactions were found among racial/ethnic groups for echogenic plaque, HDL-C level, and other risk factors and early AMD (data not shown).

Interactions were examined for smoking, hypertension status, and diabetes status for the risk factors and early AMD. Only 2 statistically significant factors were found. Statin use in those without hypertension was associated with a lower odds of early AMD (OR, 0.37; 95% CI, 0.16-0.86), whereas in those with hypertension, no association was found (OR, 1.02; 95% CI, 0.63-1.64; P for interaction = .047). Calcium channel blocker use in those with diabetes was associated with higher odds of early AMD (OR, 3.86; 95% CI, 1.53-9.70), whereas in those without diabetes, no association was found (OR, 0.70; 95% CI, 0.34-1.41).

After further multivariate analyses controlling for education level, smoking status, alcohol use, diabetes status, and body mass index, all associations described herein remained statistically significant at P<.05 except for the association between increasing levels of HDL-C and early AMD and between statin use and no large drusen or reduced larger drusen area; these associations were attenuated and of borderline statistical significance (.05<P<.10) (data not shown). The same factors were examined in relation to late AMD, but with only 27 affected persons, none of the associations was statistically significant (data not shown).

The MESA study provides data on the nature and extent of early AMD ascertained in 6 communities nationwide in 4 racial/ethnic groups in a cohort of persons 44 to 84 years of age without clinical CVD at study entry. Standardized procedures were used for measuring subclinical atherosclerotic CVD and its risk factors and for grading AMD lesions obtained from digital color fundus images of the macula. Data from this study showed few relationships of serum lipid levels and subclinical CVD with early AMD in the whole cohort. However, on the basis of our previous findings of racial/ethnic variation in specific AMD lesions in this cohort, we investigated racial/ethnic group interactions and found differences in associations of early AMD with carotid IMT and other measures of subclinical atherosclerotic CVD.³⁰

Carotid IMT, abnormal ABI, and severity of subclinical atherosclerotic CVD were not found to be associated with early AMD in the whole MESA cohort. However, positive associations were found of the severity of IMT and early AMD in black participants and inverse associations in Hispanic participants, controlling for other risk factors. The reasons for these differences between black and Hispanic participants are not obvious, but these findings suggest that atherogenesis may be more important in the pathogenesis of early AMD in 1 racial/ethnic group (eg, black individuals and not in Hispanic or Chinese individuals). It is also unclear whether these associations are due to ethnic/racial difference in genotype frequency of AMD susceptibility genes or are chance findings. This is the first study, to our knowledge, to show a relationship of subclinical atherosclerotic CVD with AMD in black people. No association was found between early AMD and atherosclerotic risk factors (eg, serum total cholesterol and LDL-C levels) in black participants in the National Health and Nutrition Examination Survey III and the Cardiovascular Health Study and with carotid IMT in the Atherosclerosis Risk in Communities study and the Cardiovascular Health Study.^{18 - 19 ,42} Similar to most population-based studies, few associations were found of measures of atherosclerosis or lipids with early or late AMD among white participants in this cohort.^{11 - 12 ,14 - 17 ,27}

There was an inverse association of the presence of echolucent plaques and no relation of isodense or echogenic plaques with early AMD in the MESA cohort. These findings are contradictory to the relationship of higher risk of late AMD in the presence of carotid artery plaque in the Rotterdam Eye Study.¹³ Echolucent plaques have lipid-filled, macrophage-rich cores with a thin fibrous caps, are vulnerable to rupture, and are predictive of cardiovascular events (eg, myocardial infarction).^{42 - 44} The finding is consistent with the protective effect found in some epidemiological studies of higher levels of total serum cholesterol for a lower incidence of AMD.¹²

In MESA, there was an association of higher serum HDL-C levels in the whole cohort with early AMD. In addition, a protective effect of higher serum triglyceride level was found in Chinese and Hispanic participants. These apparently unexpected findings are consistent with data from some, but not all, epidemiological studies.^{11 - 12 ,14 - 17 ,27 ,45} In the Beaver Dam Eye Study, a direct association was found between serum HDL-C level and the 10-year incidence of geographic atrophy.²⁷ Hyman et al⁴⁶ reported a significant positive association of serum HDL-C level (OR of highest-quintile vs lowest-quintile serum HDL-C level, 2.3; 95% CI, 1.1-4.7) with neovascular AMD. In the Rotterdam Eye Study, serum HDL-C level was directly associated with geographic atrophy (OR per micromoles per liter, 2.4; 95% CI, 1.0-5.6).¹⁴ However, other studies^{16 ,18 - 19 ,47} have not found associations of serum HDL-C level with either early or late AMD. Heavy alcohol consumption raises the HDL-C level and is reportedly associated with the incidence of AMD.⁴⁷ In MESA, the association of serum HDL-C level with early AMD was attenuated while further controlling for history of alcohol consumption (OR, 1.16 per standard deviation increase; 95% CI, 0.98-1.37; P = .08).

Statins have been suggested to limit oxidative damage and endothelial dysfunction, factors thought to be involved in AMD.²² In MESA, statin use was associated with lower frequency of large drusen and large drusen area. Although these findings are consistent with data from some case-control studies,^{20 - 22 ,24 - 25 ,48} they have not been found in most other epidemiological studies.

Although our study has many strengths (eg, standardized measurements of subclinical atherosclerosis, objective grading of AMD, and a multiethnic cohort), caution must be exercised when interpreting the findings for several reasons. We examined many variables, some of which had been identified in previous studies as associated with early AMD. The significance of any associations found may have been due to the large number of analyses and may have led to chance findings or type I error. For example, in this report, we assessed 20 study factors and constructed 20 logistic regression models to produce the findings presented in Table 2 and 120 logistic regression models (20 study factors multiplied by 6 different study outcomes) to produce the findings in Table 3. With a type I error of 5%, we would expect to see 7 significant findings by chance. For this reason, caution must be observed in interpreting the reported significant associations. In addition, the power to analyze some relations was limited by the infrequency of early AMD (4%) and the risk factor under study (eg, abnormal ABI; prevalence, 4%). The cross-sectional nature of the study design limits the ability to determine what is antecedent and what is consequent. Another possible cause of concern is the possibility of misclassification of AMD status because of the use of minified 45° nonstereoscopic fundus images in the analyses compared with 30° stereoscopic images. However, we have no reason to believe that this will result in systemic biases in evaluating associations. The MESA cohort did not include people with clinical CVD, which might account for the differences found with earlier population-based studies in which more severe CVD was manifested (eg, plaque-AMD relationship reported in Rotterdam). Uncontrolled confounding may have resulted in the race/ethnicity interactions found. Finally, we cannot determine whether there were biases caused by nonparticipation or selective mortality that affected the relationships we investigated.

In summary, these data from MESA show few associations between subclinical atherosclerotic CVD and CVD risk factors (eg, hypertension or lipids) with early AMD at a given point within the whole cohort. However, with our earlier findings of differences in the prevalence of early AMD lesions among the 4 racial/ethnic groups, the finding of an association of early AMD with some signs of subclinical atherosclerotic CVD (eg, carotid IMT) that is different among the 4 racial/ethnic groups suggests that care must be taken in generalizing from one racial/ethnic group to others regarding such relations. The observation of differences in associations between subclinical CVD and early AMD among the different racial/ethnic groups warrants further investigation.

Correspondence: Ronald Klein, MD, MPH, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 N Walnut St, 450 WARF, Madison, WI 53726-2397 (kleinr@epi.ophth.wisc.edu). Reprints are not available from the author.

Submitted for Publication: June 19, 2006; final revision received August 9, 2006; accepted August 10, 2006.

Financial Disclosure: None reported.

Funding/Support: This research was supported by National Institutes of Health, National Heart, Lung, and Blood Institute grants HL69979-03 (Drs R. Klein and Wong), N01-HC-95159 through N01-HC-95165, and N01-HC-95169; General Clinical Research Center Grant M01-RR00645 from the National Center for Research Resources; and Research to Prevent Blindness (Drs R. Klein and B. E. K. Klein, Senior Scientific Investigator Awards). In addition, the National Institutes of Health, National Heart, Lung, and Blood Institute provided funding for entire eye examinations, including collection and analyses of data, as an ancillary study to MESA (contracts N01-HC-95159 through N01-HC95166). Research to Prevent Blindness provided further additional support for data analyses.

Additional Information: A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org (see MESA P&P Policy/Acknowledgments on Web site).

Acknowledgment: We thank the other investigators, the staff, and the participants of MESA for their valuable contributions.