Chief Fellow, Stanford Cardiovascular Medicine: 2020-present
American Heart Association SFRN Postdoctoral Fellow, Heart Health Technology Center at the Stanford Center for Digital Health, 2020-present
General Cardiology Fellow, Stanford Cardiovascular Medicine, 2018-present
Chief Resident, Stanford Internal Medicine: 2017-2018
Resident/Intern, Stanford Internal Medicine: 2014-2017
Interests: Cardiovascular disease prevention; Health disparities; Digital health & health tech innovation to implementation; Clinical trials
Clinical Scholar, Medicine - Cardiovascular Medicine
- Diverse Racial/Ethnic Group Underreporting and Underrepresentation in High-Impact Cholesterol Treatment Trials. Circulation 2021; 143 (24): 2409-2411
IDENTIFYING REASONS FOR STATIN NONADHERENCE IN A DIVERSE, REAL-WORLD POPULATION USING ELECTRONIC HEALTH RECORDS AND NATURAL LANGUAGE PROCESSING
ELSEVIER SCIENCE INC. 2021: 1665
View details for Web of Science ID 000647487501671
COUNTY- LEVEL RISK FACTORS ASSOCIATED WITH RACE/ETHNICITY-SPECIFIC HEART FAILURE MORTALITY
ELSEVIER SCIENCE INC. 2021: 545
View details for Web of Science ID 000647487500545
Statin Use in Older Adults with Stable Atherosclerotic Cardiovascular Disease.
Journal of the American Geriatrics Society
BACKGROUND/OBJECTIVES: Older adults (>75years of age) represent two-thirds of atherosclerotic cardiovascular disease (ASCVD) deaths. The 2013 and 2018 American multi-society cholesterol guidelines recommend using at least moderate intensity statins for older adults with ASCVD. We examined annual trends and statin prescribing patterns in a multiethnic population of older adults with ASCVD.DESIGN: Retrospective longitudinal study using electronic health record (EHR) data from 2007 to 2018.SETTING: A large multi-specialty health system in Northern California.PARTICIPANTS: A total of 24,651 adults older than 75years with ASCVD.MEASUREMENTS: Statin prescriptions for older adults with known ASCVD were trended over time. Multivariable regression models were used to identify predictors of statin prescription (logistic) after controlling for relevant demographic and clinical factors.RESULTS: The study cohort included 24,651 patients older than 75years; 48% were women. Although prescriptions for moderate/high intensity statins increased over time for adults over 75, fewer than half of the patients (45%) received moderate/high intensity statins in 2018. Women (odds ratio (OR) = 0.77; 95% confidence interval (CI) = 0.74, 0.80), patients who had heart failure (OR = 0.69; 95% CI = 0.65, 0.74), those with dementia (OR = 0.88; 95% CI = 0.82, 0.95) and patients who were underweight (OR = 0.64; 95% CI = 0.57, 0.73) were less likely to receive moderate/high intensity statins.CONCLUSIONS: Despite increasing prescription rates between 2007 and 2018, guideline-recommended statins remained underused in older adults with ASCVD, with more pronounced disparities among women and those with certain comorbidities. Future studies are warranted to examine reasons for statin underuse in older adults with ASCVD.
View details for DOI 10.1111/jgs.16975
View details for PubMedID 33410499
Gender Disparities in Cardiology-Related COVID-19 Publications.
Cardiology and therapy
Female authors are underrepresented in cardiology journals, although prior work suggested improvement in reducing disparities over time. Early in the recent COVID-19 pandemic, female authorship continued to lag that of their male counterparts despite a surge in publications. The cumulative impact of the COVID-19 pandemic on authorship gender disparities remains unclear. We aimed to characterize gender disparities in COVID-19-related cardiology publications across the duration of the ongoing pandemic.We retrospectively analyzed COVID-19-related research articles published in the top 20 impact factor cardiology journals between March and June 2021. Gender representation data were extracted for any author, first authors, and senior authors.We found that 841 articles were related to COVID-19, with a total of 5586 authors and an average of 42 articles per journal. Less than a third (29.9%) of the total authors from publications were women. Women represented a smaller proportion of first authors (21.3%) and senior authors (16.4%).Female authorship has continued to lag male authorship for the duration of the ongoing COVID-19 pandemic. The pandemic may have impeded progress in reducing gender disparities in academic cardiology publications. The low proportions of first and senior female authors may reflect the impact of the pandemic on women in cardiology in leadership domains.
View details for DOI 10.1007/s40119-021-00234-6
View details for PubMedID 34268712
Effects of Canagliflozin on Cardiovascular, Renal, and Safety Outcomes in Participants With Type 2 Diabetes and Chronic Kidney Disease According to History of Heart Failure: Results From the CREDENCE Trial.
American heart journal
We aimed to assess the efficacy and safety of canagliflozin in patients with type 2 diabetes and nephropathy according to prior history of heart failure in the Canagliflozin and Renal Events in Diabetes With Established Nephropathy Clinical Evaluation (CREDENCE) trial. We found that participants with a prior history of heart failure at baseline (15%) were more likely to be older, female, white, have a history of atherosclerotic cardiovascular disease, and use diuretics and beta blockers (all P<0.001), and that, compared with placebo, canagliflozin safely reduced renal and cardiovascular events with consistent effects in patients with and without a prior history of heart failure (all efficacy P interaction >0.150). These results support the efficacy and safety of canagliflozin in patients with type 2 diabetes and nephropathy regardless of prior history of heart failure.
View details for DOI 10.1016/j.ahj.2020.12.008
View details for PubMedID 33358942
Multimethod, multidataset analysis reveals paradoxical relationships between sociodemographic factors, Hispanic ethnicity and diabetes.
BMJ open diabetes research & care
2020; 8 (2)
INTRODUCTION: Population-level and individual-level analyses have strengths and limitations as do 'blackbox' machine learning (ML) and traditional, interpretable models. Diabetes mellitus (DM) is a leading cause of morbidity and mortality with complex sociodemographic dynamics that have not been analyzed in a way that leverages population-level and individual-level data as well as traditional epidemiological and ML models. We analyzed complementary individual-level and county-level datasets with both regression and ML methods to study the association between sociodemographic factors and DM.RESEARCH DESIGN AND METHODS: County-level DM prevalence, demographics, and socioeconomic status (SES) factors were extracted from the 2018 Robert Wood Johnson Foundation County Health Rankings and merged with US Census data. Analogous individual-level data were extracted from 2007 to 2016 National Health and Nutrition Examination Survey studies and corrected for oversampling with survey weights. We used multivariate linear (logistic) regression and ML regression (classification) models for county (individual) data. Regression and ML models were compared using measures of explained variation (area under the receiver operating characteristic curve (AUC) and R2).RESULTS: Among the 3138 counties assessed, the mean DM prevalence was 11.4% (range: 3.0%-21.1%). Among the 12824 individuals assessed, 1688 met DM criteria (13.2% unweighted; 10.2% weighted). Age, gender, race/ethnicity, income, and education were associated with DM at the county and individual levels. Higher county Hispanic ethnic density was negatively associated with county DM prevalence, while Hispanic ethnicity was positively associated with individual DM. ML outperformed regression in both datasets (mean R2 of 0.679 vs 0.610, respectively (p<0.001) for county-level data; mean AUC of 0.737 vs 0.727 (p<0.0427) for individual-level data).CONCLUSIONS: Hispanic individuals are at higher risk of DM, while counties with larger Hispanic populations have lower DM prevalence. Analyses of population-level and individual-level data with multiple methods may afford more confidence in results and identify areas for further study.
View details for DOI 10.1136/bmjdrc-2020-001725
View details for PubMedID 33229378
Cardiovascular Effects of Dipeptidyl Peptidase-4 Inhibitors and Glucagon-Like Peptide-1 Receptor Agonists: a Review for the General Cardiologist.
Current cardiology reports
2020; 22 (10): 105
PURPOSE OF REVIEW: Results from cardiovascular (CV) outcome trials have revealed important insights into the CV safety and efficacy of glucose-lowering agents, including dipeptidyl peptidase-4 inhibitors (DPP-4i) and glucagon-like peptide-1 receptor agonists (GLP-1RA).RECENT FINDINGS: Among patients with T2DM, DPP-4i have no significant effect on risk of major adverse CV events (MACE: CV death, myocardial infarction, or stroke) with mixed results regarding risk for heart failure (HF). While sitagliptin and linagliptin have neutral effects on HF risk, saxagliptin significantly increases the risk of HF. The CV safety of the GLP-1RA class of medications has been clearly demonstrated, and select agents, such as liraglutide, semaglutide, albiglutide, and dulaglutide, reduce the risk of MACE in patients with T2DM and established CV disease. CV outcome trials have demonstrated CV safety but not incremental efficacy for DPP-4i in most cases. Select GLP-1RA have proven efficacy for MACE and should be considered by cardiologists for CV risk mitigation in the care of patients with T2DM and established CV disease.
View details for DOI 10.1007/s11886-020-01355-5
View details for PubMedID 32770420
CANAGLIFLOZIN (CANA) REDUCES CARDIOVASCULAR (CV) AND RENAL EVENTS INDEPENDENT OF BASELINE HEART FAILURE (HF): A CREDENCE SECONDARY ANALYSIS
ELSEVIER SCIENCE INC. 2020: 1018
View details for Web of Science ID 000522979101006
Wearables for arrhythmia care: Challenges and future prospects
Cardiovascular Digital Health Journal
View details for DOI 10.1016/j.cvdhj.2020.09.001
Canagliflozin and cardiovascular outcomes in Type 2 diabetes.
SGLT2 inhibitors have risen to prominence in recent years as Type 2 diabetes mellitus medications with favorable effects on cardiovascular (CV) and renal outcomes. Canagliflozin is a US FDA-approved SGLT2 inhibitor that has demonstrated CV and renal outcome benefits in large scale placebo-controlled randomized trials of patients with Type 2 diabetes mellitus and elevated CV risk. Canagliflozin use may also be associated with serious and nonserious adverse effects requiring ongoing monitoring in patients initiated on this medication. This paper provides a detailed overview of canagliflozin including its pharmacologic profile, clinical efficacy and safety data, with discussion of both clinical trial results, as well as real-world evidence.
View details for DOI 10.2217/fca-2020-0029
View details for PubMedID 32748638
Spontaneous Coronary Artery Dissection and ST-Segment Elevation Myocardial Infarction in an Anomalous LAD Artery
JACC: Case Reports
View details for DOI 10.1016/j.jaccas.2019.11.061
- Under-Reporting and Under-Representation of Racial/Ethnic Minorities in Major Atrial Fibrillation Clinical Trials. JACC. Clinical electrophysiology 2020; 6 (6): 739–41
Machine learning and atherosclerotic cardiovascular disease risk prediction in a multi-ethnic population.
NPJ digital medicine
2020; 3: 125
The pooled cohort equations (PCE) predict atherosclerotic cardiovascular disease (ASCVD) risk in patients with characteristics within prespecified ranges and has uncertain performance among Asians or Hispanics. It is unknown if machine learning (ML) models can improve ASCVD risk prediction across broader diverse, real-world populations. We developed ML models for ASCVD risk prediction for multi-ethnic patients using an electronic health record (EHR) database from Northern California. Our cohort included patients aged 18 years or older with no prior CVD and not on statins at baseline (n = 262,923), stratified by PCE-eligible (n = 131,721) or PCE-ineligible patients based on missing or out-of-range variables. We trained ML models [logistic regression with L2 penalty and L1 lasso penalty, random forest, gradient boosting machine (GBM), extreme gradient boosting] and determined 5-year ASCVD risk prediction, including with and without incorporation of additional EHR variables, and in Asian and Hispanic subgroups. A total of 4309 patients had ASCVD events, with 2077 in PCE-ineligible patients. GBM performance in the full cohort, including PCE-ineligible patients (area under receiver-operating characteristic curve (AUC) 0.835, 95% confidence interval (CI): 0.825-0.846), was significantly better than that of the PCE in the PCE-eligible cohort (AUC 0.775, 95% CI: 0.755-0.794). Among patients aged 40-79, GBM performed similarly before (AUC 0.784, 95% CI: 0.759-0.808) and after (AUC 0.790, 95% CI: 0.765-0.814) incorporating additional EHR data. Overall, ML models achieved comparable or improved performance compared to the PCE while allowing risk discrimination in a larger group of patients including PCE-ineligible patients. EHR-trained ML models may help bridge important gaps in ASCVD risk prediction.
View details for DOI 10.1038/s41746-020-00331-1
View details for PubMedID 33043149
View details for PubMedCentralID PMC7511400
Finding missed cases of familial hypercholesterolemia in health systems using machine learning.
NPJ digital medicine
2019; 2: 23
Familial hypercholesterolemia (FH) is an underdiagnosed dominant genetic condition affecting approximately 0.4% of the population and has up to a 20-fold increased risk of coronary artery disease if untreated. Simple screening strategies have false positive rates greater than 95%. As part of the FH Foundation's FIND FH initiative, we developed a classifier to identify potential FH patients using electronic health record (EHR) data at Stanford Health Care. We trained a random forest classifier using data from known patients (n = 197) and matched non-cases (n = 6590). Our classifier obtained a positive predictive value (PPV) of 0.88 and sensitivity of 0.75 on a held-out test-set. We evaluated the accuracy of the classifier's predictions by chart review of 100 patients at risk of FH not included in the original dataset. The classifier correctly flagged 84% of patients at the highest probability threshold, with decreasing performance as the threshold lowers. In external validation on 466 FH patients (236 with genetically proven FH) and 5000 matched non-cases from the Geisinger Healthcare System our FH classifier achieved a PPV of 0.85. Our EHR-derived FH classifier is effective in finding candidate patients for further FH screening. Such machine learning guided strategies can lead to effective identification of the highest risk patients for enhanced management strategies.
View details for DOI 10.1038/s41746-019-0101-5
View details for PubMedID 31304370
View details for PubMedCentralID PMC6550268
Genetic Testing and Risk Scores: Impact on Familial Hypercholesterolemia.
Frontiers in cardiovascular medicine
2019; 6: 5
Familial Hypercholesterolemia (FH) is an inherited lipid disorder affecting 1 in 220 individuals resulting in highly elevated low-density lipoprotein levels and risk of premature coronary disease. Pathogenic variants causing FH typically involve the LDL receptor (LDLR), apolipoprotein B-100 (APOB), and proprotein convertase subtulisin/kexin type 9 genes (PCSK9) and if identified convey a risk of early onset coronary artery disease (ASCVD) of 3- to 10-fold vs. the general population depending on the severity of the mutation. Identification of monogenic FH within a family has implications for family-based testing (cascade screening), risk stratification, and potentially management, and it has now been recommended that such testing be offered to all potential FH patients. Recently, robust genome wide association studies (GWAS) have led to the recognition that the accumulation of common, small effect alleles affecting many LDL-c raising genes can result in a clinical phenotype largely indistinguishable from monogenic FH (i.e., a risk of early onset ASCVD of ~3-fold) in those at the extreme tail of the distribution for these alleles (i.e., the top 8% of the population for a polygenic risk score). The incorporation of these genetic risk scores into clinical practice for non-FH patients may improve risk stratification but is not yet widely performed due to a less robust evidence base for utility. Here, we review the current status of FH genetic testing, potential future applications as well as challenges and pitfalls.
View details for PubMedID 30761309
- Dietary Patterns and Long-Term Survival: A Retrospective Study of Healthy Primary Care Patients AMERICAN JOURNAL OF MEDICINE 2018; 131 (1): 48–55
Is ACS in Young Patients a "Canary in the Coal Mine" for Familial Hypercholesterolemia?
Journal of the American College of Cardiology
2017; 70 (14): 1741–43
View details for PubMedID 28958331
Metabolic Markers to Predict Incident Diabetes Mellitus in Statin-Treated Patients (from the Treating to New Targets and the Stroke Prevention by Aggressive Reduction in Cholesterol Levels Trials).
American journal of cardiology
2016; 118 (9): 1275-1281
The goal of this analysis was to evaluate the ability of insulin resistance, identified by the presence of prediabetes mellitus (PreDM) combined with either an elevated triglyceride (TG >1.7 mmol/l) or body mass index (BMI ≥27.0 kg/m(2)), to identify increased risk of statin-associated type 2 diabetes mellitus (T2DM). Consequently, a retrospective analysis of data from subjects without diabetes in the Treating to New Targets and the Stroke Prevention by Aggressive Reduction in Cholesterol Levels randomized controlled trials was performed, subdividing participants into 4 experimental groups: (1) normal fasting glucose (NFG) and TG ≤1.7 mmol/l (42%); (2) NFG and TG >1.7 mmol/l (22%); (3) PreDM and TG ≤1.7 mmol/l (20%); and (4) PreDM and TG >1.7 mmol/l (15%). Comparable groupings were created substituting BMI values (kg/m(2) <27.0 and ≥27.0) for TG concentrations. Patients received atorvastatin or placebo for a median duration of 4.9 years. Incident T2DM, defined by developing at least 2 fasting plasma glucose (FPG) concentrations ≥126 mg/dl, an increase in FPG ≥37 mg/dl, or a clinical diagnosis of T2DM, was observed in 8.2% of the total population. T2DM event rates (statin or placebo) varied from a low of 2.8%/3.2% (NFG and TG ≤1.7 mmol/l) to a high of 22.8%/7.6% (PreDM and TG >1.7 mmol/l) with intermediate values for only an elevated TG >1.7 mmol/l (5.2%/4.3%) or only PreDM (12.8%/7.6%). Comparable differences were observed when BMI values were substituted for TG concentrations. In conclusion, these data suggest that (1) the diabetogenic impact of statin treatment is relatively modest in general; (2) the diabetogenic impact is accentuated relatively dramatically as FPG and TG concentrations and BMI increase; and (3) PreDM, TG concentrations, and BMI identify people at highest risk of statin-associated T2DM.
View details for DOI 10.1016/j.amjcard.2016.07.054
View details for PubMedID 27614854
Novel Therapies for Familial Hypercholesterolemia.
Current treatment options in cardiovascular medicine
2016; 18 (11): 64-?
Both HeFH and HoFH require dietary and lifestyle modification. Pharmacotherapy of adult HeFH patients is largely driven by the American Heart Association (AHA) algorithm. A high-potency statin is started initially with a goal low-density lipoprotein cholesterol (LDL-C) reduction of >50 %. The LDL-C target is adjusted to <100 or <70 mg/dL in subjects with coronary artery disease (CAD) with ezetimibe being second line. If necessary, a third adjunctive therapy, such as a PSCK9 inhibitor (not yet approved in children) or bile acid-binding resin, can be added. Finally, LDL-C apheresis can be considered in patients with LDL-C >300 mg/dL (or >200 mg/dL with significant CAD, although now approved for LDL-C as low as 160 mg/dL with CAD). Due to the early, severe LDL-C elevation in HoFH patients, concerning natural history, rarity of the condition, and nuances of treatment, all HoFH patients should be treated at a pediatric or adult center with HoFH experience. LDL-C apheresis should be considered as early as 5 years of age. However, apheresis availability and tolerability is limited and pharmacotherapy is required. Generally, the AHA algorithm with reference to the European Atherosclerosis Society Consensus Panel recommendations is reasonable with all patients initiated on high-dose, high-potency statin, ezetimibe, and bile acid-binding resins. In most, additional LDL-C lowering is required with PCSK9 inhibitors and/or lomitapide or mipomersen. Liver transplantation can also be considered at experienced centers as a last resort.
View details for DOI 10.1007/s11936-016-0486-2
View details for PubMedID 27620638
Cardiometabolic Effects of Glucagon-Like Peptide-1 Agonists.
Current atherosclerosis reports
2016; 18 (2): 7-?
Cardiovascular disease is the leading cause of death among adults in the USA. Both type 1 (T1DM) and type 2 diabetes mellitus (T2DM) are known risk factors for cardiovascular disease. Despite the development of numerous effective anti-glycemic therapies, we have been unable to completely mitigate cardiovascular risk with glucose lowering alone, and prevention of cardiovascular disease in patients with diabetes is primarily achieved with the use of medications that address other risk factors such as anti-hypertensives or statins. Glucagon-like peptide-1 (GLP-1) is a key hormone in the pathophysiology of diabetes. GLP-1 agonists have been recently approved for the treatment of T2DM as well as for chronic weight management. In this review, we aim to explore the effects of GLP-1 agonists on cardiovascular health with a focus on cardiometabolic variables and cardiac function.
View details for DOI 10.1007/s11883-016-0558-5
View details for PubMedID 26782825
Lectin-Dependent Enhancement of Ebola Virus Infection via Soluble and Transmembrane C-type Lectin Receptors
2013; 8 (4)
Mannose-binding lectin (MBL) is a key soluble effector of the innate immune system that recognizes pathogen-specific surface glycans. Surprisingly, low-producing MBL genetic variants that may predispose children and immunocompromised individuals to infectious diseases are more common than would be expected in human populations. Since certain immune defense molecules, such as immunoglobulins, can be exploited by invasive pathogens, we hypothesized that MBL might also enhance infections in some circumstances. Consequently, the low and intermediate MBL levels commonly found in human populations might be the result of balancing selection. Using model infection systems with pseudotyped and authentic glycosylated viruses, we demonstrated that MBL indeed enhances infection of Ebola, Hendra, Nipah and West Nile viruses in low complement conditions. Mechanistic studies with Ebola virus (EBOV) glycoprotein pseudotyped lentiviruses confirmed that MBL binds to N-linked glycan epitopes on viral surfaces in a specific manner via the MBL carbohydrate recognition domain, which is necessary for enhanced infection. MBL mediates lipid-raft-dependent macropinocytosis of EBOV via a pathway that appears to require less actin or early endosomal processing compared with the filovirus canonical endocytic pathway. Using a validated RNA interference screen, we identified C1QBP (gC1qR) as a candidate surface receptor that mediates MBL-dependent enhancement of EBOV infection. We also identified dectin-2 (CLEC6A) as a potentially novel candidate attachment factor for EBOV. Our findings support the concept of an innate immune haplotype that represents critical interactions between MBL and complement component C4 genes and that may modify susceptibility or resistance to certain glycosylated pathogens. Therefore, higher levels of native or exogenous MBL could be deleterious in the setting of relative hypocomplementemia which can occur genetically or because of immunodepletion during active infections. Our findings confirm our hypothesis that the pressure of infectious diseases may have contributed in part to evolutionary selection of MBL mutant haplotypes.
View details for DOI 10.1371/journal.pone.0060838
View details for Web of Science ID 000317717300151
View details for PubMedID 23573288
Tailoring Encodable Lanthanide-Binding Tags as MRI Contrast Agents
2012; 13 (17): 2567-2574
Lanthanide-binding tags (LBTs), peptide-based coexpression tags with high affinity for lanthanide ions, have previously been applied as luminescent probes to provide phasing for structure determination in X-ray crystallography and to provide restraints for structural refinement and distance information in NMR. The native affinity of LBTs for Gd(3+) indicates their potential as the basis for engineering of peptide-based MRI agents. However, the lanthanide coordination state that enhances luminescence and affords tightest binding would not be ideal for applications of LBTs as contrast agents, due to the exclusion of water from the inner coordination sphere. Herein, we use structurally defined LBTs as the starting point for re-engineering the first coordination shell of the lanthanide ion to provide for high contrast through direct coordination of water to Gd(3+) (resulting in the single LBT peptide, m-sLBT). The effectiveness of LBTs as MRI contrast agents was examined in vitro through measurement of binding affinity and proton relaxivity. For imaging applications that require targeted observation, fusion to specific protein partners is desirable. However, a fusion protein comprising a concatenated double LBT (dLBT) as an N-terminal tag for the model protein ubiquitin had reduced relaxivity compared with the free dLBT peptide. This limitation was overcome by the use of a construct based on the m-sLBT sequence (q-dLBT-ubiquitin). The structural basis for the enhanced contrast was examined by comparison of the X-ray crystal structure of xq-dLBT-ubiquitin (wherein two tryptophan residues are replaced with serine), to that of dLBT-ubiquitin. The structure shows that the backbone conformational dynamics of the MRI variant may allow enhanced water exchange. This engineered LBT represents a first step in expanding the current base of specificity-targeted agents available.
View details for DOI 10.1002/cbic.201200448
View details for Web of Science ID 000311245800014
View details for PubMedID 23150430