Edgar R Miller

Johns Hopkins University, Baltimore, Maryland, United States

Are you Edgar R Miller?

Claim your profile

Publications (112)872.92 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Left ventricular hypertrophy (LVH) is prevalent among hypertensive children; however, blood pressure (BP) does not predict its presence. The authors conducted a 1-year prospective cohort study to examine the hypothesis that obesity-related risk factors are associated with left ventricular mass index (LVMI) in hypertensive children, and the association between adiposity and LVMI is mediated by BP-dependent and -independent pathways. A total of 49 hypertensive children were enrolled: 51% were overweight/obese and 41% had LVH at baseline. Children overweight/obese at baseline and follow-up had a greater LVMI increase than those of healthy weight at each visit: mean change of 6.4 g/m(2.7) vs 0.95 g/m(2.7) . Baseline body mass index z score was independently associated with LVMI change (β=4.08, 1.54-6.61; P=.002). Only pulse pressure and serum aldosterone partially mediated this relationship. Hypertensive youth manifest multiple cardiovascular disease risk factors that worsen over time despite treatment. Of these, adiposity is most associated with LVH and increasing LVMI.
    Journal of Clinical Hypertension 11/2015; DOI:10.1111/jch.12717 · 2.85 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: For decades, dietary sodium intake in the United States has remained high, and few studies have examined strategies for maintaining recommended intakes. We examined the effects of a behavioral intervention, which emphasized spices and herbs, on the maintenance of sodium intake at the recommended intake of 1500 mg/d in individuals to whom the US Dietary Guidelines for Americans apply. We conducted a 2-phase study that included adults ≥18 y of age for whom Dietary Guidelines for Americans recommends 1500 mg Na/d. The study was conducted in Baltimore, Maryland, from 2012 to 2014. In phase 1, 55 individuals consumed a low-sodium diet for 4 wk. Participants were provided all foods, snacks, and calorie-containing drinks. In phase 2, 40 participants from phase 1 were randomly assigned to either a behavioral intervention to reduce sodium intake (n = 20) or a self-directed control group (n = 20) for 20 wk. The primary study outcome was the change in mean 24-h urinary sodium excretion during phase 2. Linear regression analyses were used to determine intervention effects on urinary sodium excretion. Participant characteristics were as follows: women: 65%; African American: 88%; hypertension: 63%; diabetes: 18%; mean age: 61 y; and mean body mass index (in kg/m(2)): 30. At the end of phase 2, mean 24-h sodium excretion was lower in the behavioral intervention than in the self-directed group (mean difference: -956.8 mg/d; 95% CI: -1538.7, -374.9 mg/d) after sodium intake at screening was controlled for (P = 0.002). These findings persisted in sensitivity analyses that excluded potentially incomplete urine collections [Mage's equation mean difference: -1090 mg/d (P = 0.001); Joosens' equation mean difference: -796 mg/d (P = 0.04)]. A multifactorial behavioral intervention emphasizing spices and herbs significantly reduced sodium intake. Because of the ubiquity of sodium in the US food supply, multilevel strategies addressing individual behaviors and the food supply are needed to improve adherence to recommendations. This trial was registered at clinicaltrials.gov as NCT01615159. © 2015 American Society for Nutrition.
    American Journal of Clinical Nutrition 08/2015; 102(3). DOI:10.3945/ajcn.114.100750 · 6.77 Impact Factor

  • [Show abstract] [Hide abstract]
    ABSTRACT: Uric acid (UA) is associated with high blood pressure in adolescents and with left ventricular hypertrophy (LVH) and cardiovascular disease (CVD) in adults. We sought to determine if UA is independently associated with CVD risk factors and left ventricular mass (LVM) over time in hypertensive youth. This was a 1-year prospective observational study of hypertensive children aged 3-19 years. Cross-sectional and longitudinal associations of serum UA with CVD risk factors and LVM were explored. Of the 49 children who completed both the baseline and 12-month assessments, at baseline the mean age was 13.8 years and mean UA was 5.5 mg/dL; 24 % had elevated UA, 51 % were overweight/obese and 39 % had LVH. Measures of adiposity, low high-density lipoprotein cholesterol, high-sensitivity C-reactive protein, LVM and LVH were all significantly associated with elevated UA at baseline, but not with change over time. Each 1 mg/dL increase in baseline UA was associated with a 2.5 g/m(2.7) increase in the LVM index at follow-up (95 % confidence interval 0.64, 4.39; p = 0.01); after adjustment for age, sex, race, body mass index z-score, change in UA, time, blood pressure and medication use, this association was no longer significant. Hypertensive children with elevated UA have a higher prevalence of obesity-related CVD risk factors. Among hypertensive children, UA may be a marker of adiposity and not an independent CVD risk factor.
    Pediatric Nephrology 07/2015; DOI:10.1007/s00467-015-3154-y · 2.86 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Fruit and vegetable consumption produces changes in several biomarkers in blood. The present study aimed to examine the dose-response curve between fruit and vegetable consumption and carotenoid (α-carotene, β-carotene, β-cryptoxanthin, lycopene, lutein and zeaxanthin), folate and vitamin C concentrations. Furthermore, a prediction model of fruit and vegetable intake based on these biomarkers and subject characteristics (i.e. age, sex, BMI and smoking status) was established. Data from twelve diet-controlled intervention studies were obtained to develop a prediction model for fruit and vegetable intake (including and excluding fruit and vegetable juices). The study population in the present individual participant data meta-analysis consisted of 526 men and women. Carotenoid, folate and vitamin C concentrations showed a positive relationship with fruit and vegetable intake. Measures of performance for the prediction model were calculated using cross-validation. For the prediction model of fruit, vegetable and juice intake, the root mean squared error (RMSE) was 258·0 g, the correlation between observed and predicted intake was 0·78 and the mean difference between observed and predicted intake was - 1·7 g (limits of agreement: - 466·3, 462·8 g). For the prediction of fruit and vegetable intake (excluding juices), the RMSE was 201·1 g, the correlation was 0·65 and the mean bias was 2·4 g (limits of agreement: - 368·2, 373·0 g). The prediction models which include the biomarkers and subject characteristics may be used to estimate average intake at the group level and to investigate the ranking of individuals with regard to their intake of fruit and vegetables when validating questionnaires that measure intake.
    The British journal of nutrition 04/2015; 113(09):1-14. DOI:10.1017/S0007114515000355 · 3.45 Impact Factor
  • Source

    American Journal of Kidney Diseases 04/2015; 65(6). DOI:10.1053/j.ajkd.2015.02.330 · 5.90 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Previous reports of the longitudinal association between achieved blood pressure (BP) and end-stage renal disease (ESRD) among patients with chronic kidney disease (CKD) have not incorporated time-updated BP with appropriate covariate adjustment. To assess the association between baseline and time-updated systolic blood pressure (SBP) with CKD progression. Observational, prospective cohort study. (ClinicalTrials.gov: NCT00304148). 7 U.S. clinical centers. Patients in the Chronic Renal Insufficiency Cohort Study (n = 3708) followed for a median of 5.7 years (25th to 75th percentile, 4.6 to 6.7 years). The mean of 3 seated SBP measurements made up the visit-specific SBP. Time-updated SBP was the mean of that and all previous visits. Outcomes were ESRD and the composite end point of ESRD or halving of the estimated glomerular filtration rate. Analyses investigating baseline and time-updated SBP used Cox proportional hazards models and marginal structural models, respectively. Systolic blood pressure was 130 mm Hg or greater at all visits in 19.2% of patients. The hazard ratio for ESRD among patients with SBP of 130 to 139 mm Hg, compared with SBP less than 120 mm Hg, was 1.46 (95% CI, 1.13 to 1.88) using only baseline data and 2.37 (CI, 1.48 to 3.80) using time-updated data. Among patients with SBP of 140 mm Hg or greater, corresponding hazard ratios were 1.46 (CI, 1.18 to 1.88) and 3.37 (CI, 2.26 to 5.03) for models using only baseline data and those using time-updated data, respectively. Blood pressure was measured once annually, and the cohort was not a random sample. Time-updated SBP greater than 130 mm Hg was more strongly associated with CKD progression than analyses based on baseline SBP. National Institute of Diabetes and Digestive and Kidney Diseases.
    Annals of internal medicine 02/2015; 162(4):258-265. DOI:10.7326/M14-0488 · 17.81 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Foods that have similar carbohydrate content can differ in the amount they raise blood glucose. The effects of this property, called the glycemic index, on risk factors for cardiovascular disease and diabetes are not well understood. To determine the effect of glycemic index and amount of total dietary carbohydrate on risk factors for cardiovascular disease and diabetes. Randomized crossover-controlled feeding trial conducted in research units in academic medical centers, in which 163 overweight adults (systolic blood pressure, 120-159 mm Hg) were given 4 complete diets that contained all of their meals, snacks, and calorie-containing beverages, each for 5 weeks, and completed at least 2 study diets. The first participant was enrolled April 1, 2008; the last participant finished December 22, 2010. For any pair of the 4 diets, there were 135 to 150 participants contributing at least 1 primary outcome measure. (1) A high-glycemic index (65% on the glucose scale), high-carbohydrate diet (58% energy); (2) a low-glycemic index (40%), high-carbohydrate diet; (3) a high-glycemic index, low-carbohydrate diet (40% energy); and (4) a low-glycemic index, low-carbohydrate diet. Each diet was based on a healthful DASH-type diet. The 5 primary outcomes were insulin sensitivity, determined from the areas under the curves of glucose and insulin levels during an oral glucose tolerance test; levels of low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides; and systolic blood pressure. At high dietary carbohydrate content, the low- compared with high-glycemic index level decreased insulin sensitivity from 8.9 to 7.1 units (-20%, P = .002); increased LDL cholesterol from 139 to 147 mg/dL (6%, P ≤ .001); and did not affect levels of HDL cholesterol, triglycerides, or blood pressure. At low carbohydrate content, the low- compared with high-glycemic index level did not affect the outcomes except for decreasing triglycerides from 91 to 86 mg/dL (-5%, P = .02). In the primary diet contrast, the low-glycemic index, low-carbohydrate diet, compared with the high-glycemic index, high-carbohydrate diet, did not affect insulin sensitivity, systolic blood pressure, LDL cholesterol, or HDL cholesterol but did lower triglycerides from 111 to 86 mg/dL (-23%, P ≤ .001). In this 5-week controlled feeding study, diets with low glycemic index of dietary carbohydrate, compared with high glycemic index of dietary carbohydrate, did not result in improvements in insulin sensitivity, lipid levels, or systolic blood pressure. In the context of an overall DASH-type diet, using glycemic index to select specific foods may not improve cardiovascular risk factors or insulin resistance. clinicaltrials.gov Identifier: NCT00608049.
    JAMA The Journal of the American Medical Association 12/2014; 312(23):2531-41. DOI:10.1001/jama.2014.16658 · 35.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objective. Pediatric hypertension remains largely unrecognized. We hypothesized that an electronic medical record (EMR) alert would increase elevated blood pressure (BP) recognition in a pediatric primary care setting. Study Design. Pre-post evaluation of a real-time EMR alert and one-time provider educational session. A total of 1305 encounters of children 3 to 21 years with elevated intake BP and no prior hypertension diagnosis were included. Elevated BP recognition and relationship of recognition with cardiovascular disease (CVD) risk factors during the intervention was compared with an historical control. Results. Recognition increased from 12.5% to 42% (P < .001). Recognition increased soon after alert implementation and was sustained without evidence of "alert fatigue." During both periods, presence of CVD risk factors was associated with recognition. However, the magnitude was lesser in the intervention period. Conclusions. Real-time EMR alerts substantially increase elevated BP recognition in children. However, underrecognition of elevated BP persisted, highlighting the need for additional strategies to improve provider recognition. © The Author(s) 2014.
    Clinical Pediatrics 11/2014; 54(7). DOI:10.1177/0009922814559379 · 1.15 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Chronic kidney disease is common and is associated with increased cardiovascular disease risk. Currently, markers of renal tubular injury are not used routinely to describe kidney health and little is known about the risk of cardiovascular events and death associated with these biomarkers independent of glomerular filtration-based markers (such as serum creatinine or albuminuria).
    American Journal of Kidney Diseases 10/2014; 65(2). DOI:10.1053/j.ajkd.2014.07.025 · 5.90 Impact Factor
  • Edgar R Miller · Lawrence J Appel ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Orthostatic hypotension (OH) is common in adults, and when accompanied by symptoms (dizziness, light-headedness, or fainting) carries an increased risk of falls, fractures, and mortality.(1) Symptoms are attributed to transiently reduced cerebral perfusion. The underlying causes of OH are numerous and include dehydration, "autonomic dysfunction", and medications that affect vascular compliance or responsiveness to autonomic reflexes.(2) Risk factors for OH in population-based studies include age, hypertension, hypertension treatment, diabetes, and sedative/hypnotic medication use.(3.)
    Circulation 10/2014; 130(20). DOI:10.1161/CIRCULATIONAHA.114.012884 · 14.43 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Poverty is associated with chronic kidney disease (CKD) in the United States and worldwide. Poor dietary habits may contribute to this disparity.
    Journal of Renal Nutrition 09/2014; 25(2). DOI:10.1053/j.jrn.2014.07.008 · 1.87 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objective: Evidence strongly suggests the delivery of gout care is suboptimal. The 2012 American College of Rheumatology (ACR) guidelines emphasize a serum uric acid (SUA) target of <6 mg/dl when utilizing urate lowering therapy (ULT). However, the proportion and characteristics of Americans with gout on ULT, or with a ULT indication, achieving this target is unknown.Methods: We identified US adults with gout on ULT, and those with an indication for ULT, using the National Health and Nutrition Examination Surveys from 2007-2010. A ULT indication, by ACR guidelines, comprised chronic kidney disease stages 2-5 (CKD), a history of nephrolithiasis, or current ULT use. Demographic and clinical factors associated with a SUA ≥6 mg/dl were determined using Poisson regression.Results: In 2007-2010, an estimated 4.5 million US adults with gout had an indication for ULT; two-thirds had a SUA ≥6 mg/dl. In adjusted analyses among those with gout and CKD or nephrolithiasis, those 70 years and older were less likely (prevalence ratio [PR] 0.77; 95% CI 0.61- 0.97) to have a SUA ≥6 mg/dl. Regarding those taking ULT, hypertension was related to a reduced prevalence (PR=0.51; 95%CI 0.30-0.87) whereas diabetes mellitus (PR=1.42; 95%CI 1.06-1.90) and obesity (PR=1.74; 95%CI 1.19-2.56) were each associated with a higher prevalence of a SUA value ≥6 mg/dl.Conclusion: Half of all Americans with gout on ULT, and two-thirds with an indication for ULT, have a SUA above target. This study furnishes a meaningful baseline for assessing the effectiveness of the ACR guidelines in future years. © 2014 American College of Rheumatology.
    09/2014; 67(4). DOI:10.1002/acr.22469
  • Source

    Annals of internal medicine 06/2014; 160(11):809-810. DOI:10.7326/L14-5011-6 · 17.81 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Diabetes and hypertension, common conditions in antiretroviral-treated HIV-infected individuals, are associated with glomerular hyperfiltration, which precedes the onset of proteinuria and accelerated kidney function decline. In the Multicenter AIDS Cohort Study, we examined the extent to which hyperfiltration is present and associated with metabolic, cardiovascular, HIV and treatment risk factors among HIV-infected men. Cross-sectional cohort using direct measurement of glomerular filtration rate by iohexol plasma clearance for 367 HIV-infected men and 241 HIV-uninfected men who were free of chronic kidney disease. Hyperfiltration was defined as glomerular filtration rate above 140-1 ml/min per 1.73 m per year over age 40. Multivariate logistic regression was used to estimate the odds ratios (ORs) of prevalent hyperfiltration for metabolic, cardiovascular, HIV and cumulative antiretroviral exposure factors. Among individuals without chronic kidney disease, the prevalence of hyperfiltration was higher for HIV-infected participants (25%) compared to uninfected participants (17%; P = 0.01). After adjustment, HIV infection remained associated with hyperfiltration [OR 1.70, 95% confidence interval (CI) 1.11-2.61] and modified the association between diabetes and hyperfiltration, such that the association among HIV-uninfected men (OR 2.56, 95% CI 1.33-5.54) was not observed among HIV-infected men (OR 1.19, 95% CI 0.69-2.05). These associations were independent of known risk factors for hyperfiltration. Indicators of hyperglycemia and hypertension were also associated with hyperfiltration as was cumulative zidovudine exposure. Hyperfiltration, a potential modifiable predictor of kidney disease progression, is significantly higher among antiretroviral-treated HIV-infected men. Furthermore, HIV-infection nullifies the association of diabetes and hyperfiltration present in HIV-uninfected men.
    AIDS (London, England) 01/2014; 28(3):377-86. DOI:10.1097/QAD.0000000000000094 · 5.55 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Some observational studies have identified elevated uric acid concentration as a risk factor for diabetes, while others have found an inverse relationship. We examined both the association of uric acid level with incident diabetes and the change in uric acid concentration after a diabetes diagnosis. We analyzed data from the Atherosclerosis Risk in Communities (ARIC) Study and quantified the independent association between uric acid level and incident diabetes via Cox proportional hazards models. The association between duration of diabetes and change in uric acid level was examined via linear regression. Among 11,134 participants without diagnosed diabetes at baseline (1987-1989), there were 1,294 incident cases of diabetes during a median of 9 years of follow-up (1987-1998). Uric acid level was associated with diabetes even after adjustment for risk factors (per 1 mg/dL, hazard ratio = 1.18, 95% confidence interval: 1.13, 1.23), and the association remained significant after adjustment for fasting glucose and insulin levels. Among participants with diabetes (n = 1,510), every additional 5 years' duration of diabetes was associated with a 0.10-mg/dL (95% confidence interval: 0.04, 0.15) lower uric acid level after adjustment. We conclude that uric acid concentration rises prior to diagnosis of diabetes and then declines with diabetes duration. Future studies investigating uric acid as a risk factor for cardiovascular disease should adequately account for the impact and timing of diabetes development.
    American journal of epidemiology 01/2014; 179(6). DOI:10.1093/aje/kwt320 · 5.23 Impact Factor

  • Annals of internal medicine 12/2013; 159(12):850-1. DOI:10.7326/0003-4819-159-12-201312170-00011 · 17.81 Impact Factor
  • Priya Palta · Laura J Samuel · Edgar R Miller · Sarah L Szanton ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Objective To perform a systematic review and meta-analysis that quantitatively tests and summarizes the hypothesis that depression results in elevated oxidative stress and lower antioxidant levels.Methods We performed a meta-analysis of studies that reported an association between depression and oxidative stress and/or antioxidant status markers. PubMed and EMBASE databases were searched for articles published from January 1980 through December 2012. A random-effects model, weighted by inverse variance, was performed to pool standard deviation (Cohen's d) effect size estimates across studies for oxidative stress and antioxidant status measures, separately.ResultsTwenty-three studies with 4980 participants were included in the meta-analysis. Depression was most commonly measured using the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. A Cohen's d effect size of 0.55 (95% confidence interval = 0.47-0.63) was found for the association between depression and oxidative stress, indicating a roughly 0.55 of 1-standard-deviation increase in oxidative stress among individuals with depression compared with those without depression. The results of the studies displayed significant heterogeneity (I(2) = 80.0%, p < .001). A statistically significant effect was also observed for the association between depression and antioxidant status markers (Cohen's d = -0.24, 95% confidence interval = -0.33 to -0.15).Conclusions This meta-analysis observed an association between depression and oxidative stress and antioxidant status across many different studies. Differences in measures of depression and markers of oxidative stress and antioxidant status markers could account for the observed heterogeneity. These findings suggest that well-established associations between depression and poor heath outcomes may be mediated by high oxidative stress.
    Psychosomatic Medicine 12/2013; 76(1). DOI:10.1097/PSY.0000000000000009 · 3.47 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To examine the association between 25-hydroxyvitamin D [25(OH)D] deficiency and anemia in a cohort of otherwise-healthy children and to determine whether race modifies the association between 25(OH)D status and hemoglobin (Hgb). Cross-sectional study of 10 410 children and adolescents ages 1-21 years from the 2001-2006 National Health and Nutrition Examination Survey. Anemia was defined as Hgb less than the 5th percentile for age and sex based on National Health and Nutrition Examination Survey III (1988-1994) data. Lower 25(OH)D levels were associated with increased risk for anemia; <30 ng/mL, adjusted OR 1.93, 95% CI 1.21-3.08, P = .006, and <20 ng/mL, OR 1.47, 95% CI 1.14-1.89, P = .004. In linear regression, small but significant increases in Hgb were noted in the upper quartiles of 25(OH)D compared with the lowest quartile (<20 ng/mL) in the full cohort. Results of race-stratified linear regression by 25(OH)D quartile in white children were similar to those observed in the full cohort, but in black children, an increase in Hgb in the upper 25(OH)D quartiles was only apparent compared with the lowest black race-specific quartile (<12 ng/mL). 25(OH)D deficiency is associated with increased risk of anemia in healthy US children, but the 25(OH)D threshold levels for lower Hgb are lower in black children in comparison with white children.
    The Journal of pediatrics 10/2013; 164(1). DOI:10.1016/j.jpeds.2013.08.060 · 3.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Determinants of oxidative capacity, such as fitness and level of adiposity, are strongly associated with type 2 diabetes. Whether decreased oxidative capacity itself is a cause or consequence of insulin resistance and diabetes is unknown. We examined the association of plasma lactate, a marker of oxidative capacity, with incident diabetes in 8045 participants from the Atherosclerosis Risk in Communities (ARIC) Study with no history of subclinical or diagnosed diabetes at baseline (1996-1998). Incident diabetes was self-reported during annual telephone calls. During a median follow-up of 12 years, there were 1513 new cases of diabetes. In Cox proportional hazards models, baseline plasma lactate (per 10 mg/dL) was significantly associated with diabetes (hazard ratio, 1.20; 95% confidence interval, 1.01-1.43), even after adjustment for diabetes risk factors, fasting glucose, and insulin. The upper quartile of baseline lactate (≥8.1 mg/dL) was also significantly associated with diabetes risk (hazard ratio, 1.20; 95% confidence interval, 1.02-1.41) compared with the lowest quartile (≤5.1 mg/dL). Significant associations persisted among persons without insulin resistance (homeostatic model assessment insulin resistance index < 2.6 U) (P-trend < .01). These findings suggest that low oxidative capacity may precede diabetes. Future studies should evaluate the physiological origins of elevated lactate to better understand its possible role in the pathogenesis of diabetes.
    Annals of epidemiology 10/2013; 23(12). DOI:10.1016/j.annepidem.2013.09.005 · 2.00 Impact Factor

Publication Stats

7k Citations
872.92 Total Impact Points


  • 2003-2015
    • Johns Hopkins University
      • • Department of Epidemiology
      • • Department of Medicine
      • • Welch Center for Prevention, Epidemiology, and Clinical Research
      Baltimore, Maryland, United States
  • 2002-2014
    • Johns Hopkins Bloomberg School of Public Health
      • Department of Epidemiology
      Baltimore, Maryland, United States
  • 1997-2014
    • Johns Hopkins Medicine
      • • Department of Medicine
      • • Welch Center for Prevention, Epidemiology and Clinical Research
      • • Department of Epidemiology
      Baltimore, Maryland, United States
  • 2010
    • University of California, San Francisco
      • Division of Hospital Medicine
      San Francisco, CA, United States
  • 2006
    • National Institute on Aging
      Baltimore, Maryland, United States
    • National Institutes of Health
      베서스다, Maryland, United States
  • 2005
    • Yonsei University
      • Graduate School of Public Health
      Seoul, Seoul, South Korea
  • 2004
    • Duke University
      Durham, North Carolina, United States