Potential Channeling Bias in the Evaluation of Cardiovascular Risk: The Importance of Comparator Selection in Observational Research.

BACKGROUND: Comparator selection is an important consideration in the design of observational research studies that evaluate potential associations between drug therapies and adverse event risks. It can affect the validity of observational study results, and potentially impact data interpretation, regulatory decision making, and patient medication access. OBJECTIVE: The aim of this study was to assess the impact of comparator selection bias using two real-world case studies evaluating an increased rate of acute myocardial infarction (AMI). METHODS: Data from the Truven Health Analytics MarketScan® electronic medical claims database were used to conduct two retrospective observational cohort studies, utilizing a cohort new-user design, comparing AMI risk between testosterone replacement therapy (TRT) and phosphodiesterase-5 inhibitors (PDE5is) in men treated for hypogonadism, and triptans versus other prescribed acute treatments for migraine in adults. All patients were enrolled continuously in a health plan (no enrollment gap > 31 consecutive days) for ≥ 1 year before index. Baseline period was defined as 365 days prior to index. Exposure was defined by prescription and outcome of interest was defined as occurrence of AMI. Using Cox proportional hazard models, primary analysis for the TRT cohort compared AMI risk between propensity score (PS)-matched TRT-treated and untreated patients; secondary analysis evaluated risk between PS-matched TRT-treated and PDE5i-treated patients. For the triptan cohort, primary analysis compared AMI/ischemic stroke risk between PS-matched triptan-treated and opiate-treated patients; secondary analysis evaluated risk between PS-matched triptan-treated and nonsteroidal anti-inflammatory drug (NSAID)-treated patients and PS-matched non-prescription-treated migraine patients and general patients. RESULTS: No significant association between TRT and AMI was observed among TRT-treated (N = 198,528, mean age 52.4 ± 11.4 years) versus PDE5i-treated men (N = 198,528, mean age 52.3 ± 11.5 years) overall (adjusted hazard ratio [aHR] 1.01; 95% CI 0.95-1.07; p = 0.80). Among patients with prior cardiovascular disease (CVD), risk of AMI was significantly increased for TRT-treated versus PDE5i-treated patients (aHR 1.13; 95% CI 1.03-1.25). The triptan study included three comparisons (triptans [N = 436,642] vs prescription NSAIDs [N = 334,152], opiates [N = 55,234], and untreated migraine [N = 1,168,212]), and a positive control (untreated vs general non-migraine patients [N = 11,735,009]). Analyses of MI risk in migraine patients prescribed triptans versus NSAIDs/opiates had mixed results: the point estimate ranged from 0.33 to 0.84 depending on chosen study window. CONCLUSIONS: Cardiovascular outcomes were not worse in hypogonadism patients with TRT versus PDE5i; however, a potential association with AMI was found in patients with prior CVD receiving TRT versus PDE5i. Findings pointed to a pseudo-protective effect of triptans versus untreated migraine patients or those potentially older and less healthy patients exposed to prescription NSAIDs or opiates. Triptan users should not be compared with those using other anti-migraine prescriptions when evaluating cardiovascular outcomes in migraine patients. Presence of high cardiovascular risks may contribute to channeling bias-healthier subjects being selected to receive treatment-highlighting the importance of choosing comparators wisely in observational studies.

Treatment heterogeneity and overall survival in patients with advanced/metastatic gastric or gastroesophageal junction adenocarcinoma in the United States.

Background: Gastric or gastroesophageal junction (GEJ) adenocarcinoma is the most common form of gastric cancer diagnosed in the United States (US) each year. Diagnosis typically is in later stages of disease when it has advanced. Patients have been treated with a variety of regimens. Methods: The goal of this retrospective study was to understand if treatment patterns were becoming more homogeneous or remaining heterogeneous using the Herfindahl-Hirschman index (HHI) and if treatments were becoming more concordant to treatment guidelines published by the National Comprehensive Cancer Network (NCCN). HHI scores were calculated for each site by 2-year increments. Trend analyses were conducted for HHI scores over time using a linear regression model. Concordance to Category 1 and any category NCCN guidelines was determined based on the date treatment was initiated with the version of the NCCN guidelines at that time. Time trend analyses were conducted using linear regression models. This study utilized data from the Flatiron Advanced Gastric/Esophageal cohort. This study also examined overall survival (OS) rates estimated by the Kaplan-Meier method by line of therapy. Results: There were no statistically significant differences in HHI scores in the first-line setting over time, suggesting heterogeneity has not improved. Concordance to NCCN treatment guidelines for any category significantly increased over time, however Category 1 regimen concordance remained low in the first-line setting. Concordance over time improved in second-line treatment. Median OS from the start of first-line therapy was 13.57 months. There was no relationship between OS time from initiation of first-line therapy and HHI score, concordance with NCCN guidelines, or concordance with NCCN Category 1 guidelines in the first-line setting. Conclusions: Treatment heterogeneity persists in gastric cancer care, though there is a significant association between heterogeneity and concordance with both Category 1 and any category in the NCCN treatment guidelines, and that concordance has increased over time.

Risk of Venous Thromboembolism Among Patients with Atopic Dermatitis: A Cohort Study in a US Administrative Claims Database.

INTRODUCTION: Atopic dermatitis (AD) is associated with risk factors for venous thromboembolism (VTE). However, the risk of VTE among this population is unknown. The aim of this study was to assess the risk of VTE among adults with AD and compare the risk vs. matched non-AD controls. METHODS: This retrospective study used claims data from the IBM Watson MarketScan® Commercial Claims and Encounters, Medicare Supplemental, and Medicaid databases to identify adults aged 18 years or older with AD. Incidence rates (IR) per 100 person-years (PY) of VTE were reported for three cohorts: overall AD, moderate-to-severe AD, and non-AD controls matched by age, sex, and calendar time to the overall cohort. Cox proportional hazards regression was used to estimate hazard ratios (HR) for VTE risk. RESULTS: Overall, 198,685 patients with AD were identified. Crude VTE IRs were 0.24 for AD overall, 0.31 for moderate-to-severe AD, and 0.25 for non-AD controls. VTE risk was similar in patients with AD vs. non-AD controls (partially adjusted HR 1.00, 95% confidence interval [CI] 0.92, 1.09). VTE risk was greater in patients with moderate-to-severe AD vs. non-AD controls in partially adjusted models (HR 1.24, 95% CI 1.13, 1.36), but not after adjustment for healthcare use and VTE risk factors (HR 0.95, 95% CI 0.85, 1.07). CONCLUSIONS: AD was not an independent risk factor for VTE, and the risk of VTE among patients with AD was low. These findings provide valuable context for understanding VTE risk among patients with AD, which is particularly relevant as advanced therapies for the treatment of moderate to severe AD, such as janus kinase inhibitors, become available.

Defining treatment regimens and lines of therapy using real-world data in oncology.

Retrospective observational research relies on databases that do not routinely record lines of therapy or reasons for treatment change. Standardized approaches to estimate lines of therapy were developed and evaluated in this study. A number of rules were developed, assumptions varied and macros developed to apply to large datasets. Results were investigated in an iterative process to refine line of therapy algorithms in three different cancers (lung, colorectal and gastric). Three primary factors were evaluated and included in the estimation of lines of therapy in oncology: defining a treatment regimen, addition/removal of drugs and gap periods. Algorithms and associated Statistical Analysis Software (SAS®) macros for line of therapy identification are provided to facilitate and standardize the use of real-world databases for oncology research.

Lay abstract Most, if not all, real-world healthcare databases do not contain data explaining treatment changes, requiring that rules be applied to estimate when treatment changes may reflect advancement of underlying disease. This study investigated three tumor types (lung, colorectal and gastric cancer) to develop and provide rules that researchers can apply to real-world databases. The resulting algorithms and associated SAS^® macros from this work are provided for use in the Supplementary data.

Acute Migraine Prescription Patterns Vary by Baseline Cardiovascular Risk and Clinical Characteristics: A Real-World Evidence Study.

INTRODUCTION: Migraine is a debilitating neurological disease and one of the most common disorders in the world. Although the triptans, potent 5-HT1B/1D receptor agonists, are an effective and widely used acute treatment of migraine, few studies have assessed how their cardiovascular risk warnings could impact prescription patterns. This study characterized cardiovascular risk factors and other aspects of people with migraine in real-world settings and confirmed patterns of acute migraine care. METHODS: This retrospective study included five cohorts: people with migraine prescribed acute treatments [triptans, opiates, prescription nonsteroidal anti-inflammatory drugs (NSAIDs)], untreated people with migraine, and individuals without migraine diagnosis. Baseline demographic and clinical characteristics were used to develop and validate a 1-year myocardial infarction (MI) risk prediction model among untreated people with migraine. This validated prediction model generated disease risk scores (DRSs) for MI among untreated cohorts. RESULTS: Patients in the study included 436,642 prescribed a triptan, 55,234 prescribed opiates, and 334,152 prescribed NSAIDs; as well as 1,168,212 untreated persons with migraine and 11,735,009 nonmigraine participants. Those prescribed triptans were younger, had fewer cardiovascular risk factors and hospitalizations, and lower concomitant medication use than those in the NSAID and opiate cohorts. The distribution of the DRS showed that compared to patients prescribed NSAIDs (4.2%) or opiates (3.5%), a smaller proportion of patients prescribed triptans (1.3%) were at high risk for MI at 1 year (> 10%). CONCLUSION: People with migraine who had more cardiovascular risk factors and greater 1-year MI risk score were disproportionately prescribed opiates and NSAIDs compared to triptans. Future research should explore unmet needs for patients with disorders for which triptan therapy is contraindicated.

Treatment patterns and survival outcomes for patients receiving second-line treatment for metastatic colorectal cancer in the USA.

BACKGROUND: Colorectal cancer is the third most common cause of cancer death in the USA. It is important to identify patients who may experience poor outcomes from available treatments. METHODS: In this retrospective observational study, treatment patterns and survival outcomes were described among adult patients from the Flatiron Health electronic medical records database who were treated with at least two lines of therapy for metastatic colorectal cancer in the USA between January 2013 and May 2018. Patients with rapid progression were defined as those whose time from start of first- to second-line therapy was ≤ 183 days. RESULTS: A total of 14,315 patients formed the study cohort. The most common first-line treatments were FOLFOX (5-fluorouracil, leucovorin, and oxaliplatin) plus bevacizumab, received by 34.7% (n = 4962) of patients, followed by FOLFOX alone (17.1%, n = 2445). Of all patients, 6991 (48.9%) also received second-line anti-cancer therapy and of those, 3338 (47.7%) had rapid progression and 3653 (52.3%) did not. Median overall survival from the start of first- and second-line therapy was 20.8 months (95% CI 20.2-21.3) and 14.5 months (95% CI 13.9-15.0) for the entire study population, respectively. Median overall survival from the start of second-line therapy was 14.1 (95% CI 13.2-14.8) for patients with rapid progression and 14.6 months (95% CI 13.8-15.4) for patients without rapid progression. CONCLUSIONS: Patients diagnosed with metastatic colorectal cancer lived less than 2 years in this real-world database. While the time to initiation of second-line therapy was by definition longer among patients without rapidly progressing disease, survival outcomes were comparable from initiation of second-line therapy.

Hi-MC: a novel method for high-throughput mitochondrial haplogroup classification.

Effective approaches for assessing mitochondrial DNA (mtDNA) variation are important to multiple scientific disciplines. Mitochondrial haplogroups characterize branch points in the phylogeny of mtDNA. Several tools exist for mitochondrial haplogroup classification. However, most require full or partial mtDNA sequence which is often cost prohibitive for studies with large sample sizes. The purpose of this study was to develop Hi-MC, a high-throughput method for mitochondrial haplogroup classification that is cost effective and applicable to large sample sizes making mitochondrial analysis more accessible in genetic studies. Using rigorous selection criteria, we defined and validated a custom panel of mtDNA single nucleotide polymorphisms that allows for accurate classification of European, African, and Native American mitochondrial haplogroups at broad resolution with minimal genotyping and cost. We demonstrate that Hi-MC performs well in samples of European, African, and Native American ancestries, and that Hi-MC performs comparably to a commonly used classifier. Implementation as a software package in R enables users to download and run the program locally, grants greater flexibility in the number of samples that can be run, and allows for easy expansion in future revisions. Hi-MC is available in the CRAN repository and the source code is freely available at https://github.com/vserch/himc.

Application and comparison of generalized propensity score matching versus pairwise propensity score matching.

AIM: A comparison of conventional pairwise propensity score matching (PSM) and generalized PSM method was applied to the comparative effectiveness of multiple treatment options for lung cancer. MATERIALS & METHODS: Deidentified data were analyzed. Covariate balances between compared treatments were assessed before and after PSM. Cox proportional hazards regression compared overall survival after PSM. RESULTS & CONCLUSION: The generalized PSM analyses were able to retain 61.2% of patients, while the conventional PSM analyses were able to match from 24.1 to 77.1% of patients from each treatment comparison. The generalized PSM achieved statistical significance (p < 0.05) in 8/10 comparisons, whereas conventional pairwise PSM achieved 1/10. The noted differences arose from different matched patient samples and the size of the samples.

Evaluation of dystonia in children and adolescents treated with atomoxetine within the Truven MarketScan database: a retrospective cohort study.

OBJECTIVE: Atomoxetine is a non-stimulant drug indicated for the treatment of attention-deficit/hyperactivity disorder in children aged ≥6 years, adolescents, and adults. In this retrospective cohort study, the incidence and risk of dystonia in children and adolescents treated with atomoxetine was compared to a propensity score-matched cohort of stimulant users. METHODS: Data between 1 January 2006 and 31 December 2014 from patients aged 6-17 years in the Truven Health Analytics MarketScan database were used to generate two cohorts of patients: (1) atomoxetine users and (2) stimulant (methylphenidates or amphetamines) users. A Cox proportional hazards regression model was used to compare incidence of dystonia across propensity score-matched cohorts. RESULTS: Of the 70,657 atomoxetine users, 70,655 users were propensity score-matched to a stimulant user. In the atomoxetine- and stimulant-treated cohorts, the crude incidence rates of dystonia were 54.9 (95% CI: 27.1-82.7) and 77.9 (95% CI: 49.1-106.8) per 100,000 person-years, respectively. The hazard ratio for occurrence of dystonia with atomoxetine use relative to stimulant use was 0.68 (95% CI: 0.36 - 1.28; P = 0.23). CONCLUSION: In this large retrospective cohort study, there was no significant difference in incidence or risk of dystonia among patients treated with atomoxetine compared to stimulants.

Extracting Country-of-Origin from Electronic Health Records for Gene- Environment Studies as Part of the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) Study.

We describe here the extraction of country-of-origin, an acculturation variable relevant for gene-environment studies, in a biorepository linked to de-identified electronic health records (EHRs) assessed by the Epidemiologic Architecture for Genes Linked to Environment (EAGLE), a study site of the Population Architecture using Genomics and Epidemiology (PAGE) I study. We extracted country-of-origin from the unstructured clinical free text using regular expressions within the MySQL relational database system in a cohort of 15,863 subjects of mostly non-European descent (including 11,519 African Americans, 1,702 Hispanics, and 1,118 Asians). We performed searches for 231 world countries (including independent sovereign states, dependent areas, and disputed territories) and common misspellings in >14 gigabytes of data including >13 billion characters of clinical text. Manual review of a fraction of the initial country-of-origin assignments established rules for data cleaning and quality control to achieve final country-of-origin status for each subject. After data cleaning, a total of 1,911/15,893 (12.02%) subjects were assigned to a country-of-origin outside of the United States. Mexico was the most commonly assigned country outside of the United States (264 subjects; 13.8% of subjects with a foreign country-of-origin assignment). The distribution of the countries assigned followed expectations based on known migration patterns to the United States with an emphasis on the southeastern region. These data suggest country-of-origin can be successfully extracted from unstructured clinical text for downstream genetic association studies.

Reducing Clinical Noise for Body Mass Index Measures Due to Unit and Transcription Errors in the Electronic Health Record.

Body mass index (BMI) is an important outcome and covariate adjustment for many clinical association studies. Accurate assessment of BMI, therefore, is a critical part of many study designs. Electronic health records (EHRs) are a growing source of clinical data for research purposes, and have proven useful for identifying and replicating genetic associations. EHR-based data collected for clinical and billing purposes have several unique properties, including a high degree of heterogeneity or "clinical noise." In this work, we propose a new method for reducing the problems of transcription and recording error for height and weight and apply these methods to a subset of the Vanderbilt University Medical Center biorepository known as EAGLE BioVU (n=15,863). After processing, we show that the distribution of BMI from EAGLE BioVU closely matches population-based estimates from the National Health and Nutrition Examination Surveys (NHANES), and that our approach retains far more data points than traditional outlier detection methods.

Trans-ethnic fine-mapping of genetic loci for body mass index in the diverse ancestral populations of the Population Architecture using Genomics and Epidemiology (PAGE) Study reveals evidence for multiple signals at established loci.

Most body mass index (BMI) genetic loci have been identified in studies of primarily European ancestries. The effect of these loci in other racial/ethnic groups is less clear. Thus, we aimed to characterize the generalizability of 170 established BMI variants, or their proxies, to diverse US populations and trans-ethnically fine-map 36 BMI loci using a sample of >102,000 adults of African, Hispanic/Latino, Asian, European and American Indian/Alaskan Native descent from the Population Architecture using Genomics and Epidemiology Study. We performed linear regression of the natural log of BMI (18.5-70 kg/m2) on the additive single nucleotide polymorphisms (SNPs) at BMI loci on the MetaboChip (Illumina, Inc.), adjusting for age, sex, population stratification, study site, or relatedness. We then performed fixed-effect meta-analyses and a Bayesian trans-ethnic meta-analysis to empirically cluster by allele frequency differences. Finally, we approximated conditional and joint associations to test for the presence of secondary signals. We noted directional consistency with the previously reported risk alleles beyond what would have been expected by chance (binomial p < 0.05). Nearly, a quarter of the previously described BMI index SNPs and 29 of 36 densely-genotyped BMI loci on the MetaboChip replicated/generalized in trans-ethnic analyses. We observed multiple signals at nine loci, including the description of seven loci with novel multiple signals. This study supports the generalization of most common genetic loci to diverse ancestral populations and emphasizes the importance of dense multiethnic genomic data in refining the functional variation at genetic loci of interest and describing several loci with multiple underlying genetic variants.

Safety profiles of non-small cell lung cancer patients treated with pemetrexed plus carboplatin: a real-world retrospective, observational, cohort study.

OBJECTIVE: Pemetrexed plus carboplatin (PCb) is a frequently used first-line treatment in advanced non-small cell lung cancer (NSCLC). This study examined the characteristics and safety profile of a NSCLC population treated with PCb area under the concentration-time curve 5 (PCb5) or 6 mg/mL•min (PCb6) under real-world conditions. RESEARCH DESIGN AND METHODS: A retrospective, observational, cohort study was conducted, utilizing data from the IMS Oncology US clinic-based, longitudinal, patient-level electronic medical records (EMR), including patients with NSCLC on PCb5 or PCb6 regimens initiated concomitantly on or after the diagnosis of lung cancer during 2004-2014. Patient characteristics and incidence of adverse events (AEs) were described for each cohort. Propensity scores were calculated based on baseline demographic and clinical factors. Propensity score stratification was used to further adjust for cohort differences. RESULTS: In total, 636 NSCLC patients receiving PCb5 (37% aged ≥70 years) and 184 patients receiving PCb6 (34% aged ≥70 years) who met the inclusion criteria were identified in the EMR. Patients with more comorbidities were more likely to have received PCb5. Overall incidence rates (IRs) per 100 person-years were similar for neutropenia in both cohorts, were numerically higher for anemia (IR = 43.6 vs 101.0) and thrombocytopenia (IR = 1.5 vs 17.9), and were numerically lower for nausea (IR = 14.4 vs 9.9) in the PCb6 vs PCb5 cohort. Within the PCb6 cohort, the IR per 100 person-years was higher for neutropenia for ≥70 year-old patients (IR = 41.1) compared to <70 year-old patients (IR = 14.5). After propensity score stratification, adjusted IRs showed similar patterns. LIMITATIONS: Limitations included lack of power for AEs other than anemia, given the nature of EMR. CONCLUSIONS: Results from this real-world analysis add to existing evidence from randomized clinical trials about PCb safety profiles in the overall NSCLC population and in elderly patients. These results may guide physicians when making treatment decisions.

Population Stratification in the Context of Diverse Epidemiologic Surveys Sans Genome-Wide Data.

Population stratification or confounding by genetic ancestry is a potential cause of false associations in genetic association studies. Estimation of and adjustment for genetic ancestry has become common practice thanks in part to the availability of ancestry informative markers on genome-wide association study (GWAS) arrays. While array data is now widespread, these data are not ubiquitous as several large epidemiologic and clinic-based studies lack genome-wide data. One such large epidemiologic-based study lacking genome-wide data accessible to investigators is the National Health and Nutrition Examination Surveys (NHANES), population-based cross-sectional surveys of Americans linked to demographic, health, and lifestyle data conducted by the Centers for Disease Control and Prevention. DNA samples (n = 14,998) were extracted from biospecimens from consented NHANES participants between 1991-1994 (NHANES III, phase 2) and 1999-2002 and represent three major self-identified racial/ethnic groups: non-Hispanic whites (n = 6,634), non-Hispanic blacks (n = 3,458), and Mexican Americans (n = 3,950). We as the Epidemiologic Architecture for Genes Linked to Environment study genotyped candidate gene and GWAS-identified index variants in NHANES as part of the larger Population Architecture using Genomics and Epidemiology I study for collaborative genetic association studies. To enable basic quality control such as estimation of genetic ancestry to control for population stratification in NHANES san genome-wide data, we outline here strategies that use limited genetic data to identify the markers optimal for characterizing genetic ancestry. From among 411 and 295 autosomal SNPs available in NHANES III and NHANES 1999-2002, we demonstrate that markers with ancestry information can be identified to estimate global ancestry. Despite limited resolution, global genetic ancestry is highly correlated with self-identified race for the majority of participants, although less so for ethnicity. Overall, the strategies outlined here for a large epidemiologic study can be applied to other datasets accessible for genotype-phenotype studies but are sans genome-wide data.

KIDNEY DISEASE GENETICS AND THE IMPORTANCE OF DIVERSITY IN PRECISION MEDICINE.

Kidney disease is a well-known health disparity in the United States where African Americans are affected at higher rates compared with other groups such as European Americans and Mexican Americans. Common genetic variants in the myosin, heavy chain 9, non-muscle (MYH9) gene were initially identified as associated with non-diabetic end-stage renal disease in African Americans, and it is now understood that these variants are in strong linkage disequilibrium with likely causal variants in neighboring APOL1. Subsequent genome-wide and candidate gene studies have suggested that MYH9 common variants among others are also associated with chronic kidney disease and quantitative measures of kidney function in various populations. In a precision medicine setting, it is important to consider genetic effects or genetic associations that differ across racial/ethnic groups in delivering data relevant to disease risk or individual-level patient assessment. Kidney disease and quantitative trait-associated genetic variants have yet to be systematically characterized in multiple racial/ethnic groups. Therefore, to further characterize the prevalence of these genetic variants and their association with kidney related traits, we have genotyped 10 kidney disease or quantitative trait-associated single nucleotide polymorphisms (SNPs) (rs2900976, rs10505955, rs10502868, rs1243400, rs9305354, rs12917707, rs17319721, rs2467853, rs2032487, and rs4821480) in 14,998 participants from the population-based cross-sectional National Health and Nutrition Examination Surveys (NHANES) III and 1999-2002 as part of the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study. In this general adult population ascertained regardless of health status (6,293 non-Hispanic whites, 3,013 non-Hispanic blacks, and 3,542 Mexican Americans), we observed higher rates of chronic kidney disease among non-Hispanic blacks compared with the other groups as expected. We performed single SNP tests of association using linear regressions assuming an additive genetic model adjusted for age, sex, diastolic blood pressure, systolic blood pressure, and type 2 diabetes status for several outcomes including creatinine (urinary), creatinine (serum), albumin (urinary), eGFR, and albumin-to-urinary creatinine ratio (ACR). We also tested for associations between each SNP and chronic kidney disease and albuminuria using logistic regression. Surprisingly, none of the MYH9 variants tested was associated with kidney diseases or traits in non-Hispanic blacks (p>0.05), perhaps attributable to the clinical heterogeneity of kidney disease in this population. Several associations were observed in each racial/ethnic group at p<0.05, but none were consistently associated in the same direction in all three groups. The lack of significant and consistent associations is most likely due to power highlighting the importance of the availability of large, diverse populations for genetic association studies of complex diseases and traits to inform precision medicine efforts in diverse patient populations.

TESTING POPULATION-SPECIFIC QUANTITATIVE TRAIT ASSOCIATIONS FOR CLINICAL OUTCOME RELEVANCE IN A BIOREPOSITORY LINKED TO ELECTRONIC HEALTH RECORDS: LPA AND MYOCARDIAL INFARCTION IN AFRICAN AMERICANS.

Previous candidate gene and genome-wide association studies have identified common genetic variants in LPA associated with the quantitative trait Lp(a), an emerging risk factor for cardiovascular disease. These associations are population-specific and many have not yet been tested for association with the clinical outcome of interest. To fill this gap in knowledge, we accessed the epidemiologic Third National Health and Nutrition Examination Surveys (NHANES III) and BioVU, the Vanderbilt University Medical Center biorepository linked to de-identified electronic health records (EHRs), including billing codes (ICD-9-CM) and clinical notes, to test population-specific Lp(a)-associated variants for an association with myocardial infarction (MI) among African Americans. We performed electronic phenotyping among African Americans in BioVU≥40 years of age using billing codes. At total of 93 cases and 522 controls were identified in NHANES III and 265 cases and 363 controls were identified in BioVU. We tested five known Lp(a)-associated genetic variants (rs1367211, rs41271028, rs6907156, rs10945682, and rs1652507) in both NHANES III and BioVU for association with myocardial infarction. We also tested LPA rs3798220 (I4399M), previously associated with increased levels of Lp(a), MI, and coronary artery disease in European Americans, in BioVU. After meta-analysis, tests of association using logistic regression assuming an additive genetic model revealed no significant associations (p<0.05) for any of the five LPA variants previously associated with Lp(a) levels in African Americans. Also, I4399M rs3798220 was not associated with MI in African Americans (odds ratio = 0.51; 95% confidence interval: 0.16 - 1.65; p=0.26) despite strong, replicated associations with MI and coronary artery disease in European American genome-wide association studies. These data highlight the challenges in translating quantitative trait associations to clinical outcomes in diverse populations using large epidemiologic and clinic-based collections as envisioned for the Precision Medicine Initiative.

Cryptic relatedness in epidemiologic collections accessed for genetic association studies: experiences from the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study and the National Health and Nutrition Examination Surveys (NHANES).

Epidemiologic collections have been a major resource for genotype-phenotype studies of complex disease given their large sample size, racial/ethnic diversity, and breadth and depth of phenotypes, traits, and exposures. A major disadvantage of these collections is they often survey households and communities without collecting extensive pedigree data. Failure to account for substantial relatedness can lead to inflated estimates and spurious associations. To examine the extent of cryptic relatedness in an epidemiologic collection, we as the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study accessed the National Health and Nutrition Examination Surveys (NHANES) linked to DNA samples ("Genetic NHANES") from NHANES III and NHANES 1999-2002. NHANES are population-based cross-sectional surveys conducted by the National Center for Health Statistics at the Centers for Disease Control and Prevention. Genome-wide genetic data is not yet available in NHANES, and current data use agreements prohibit the generation of GWAS-level data in NHANES samples due issues in maintaining confidentiality among other ethical concerns. To date, only hundreds of single nucleotide polymorphisms (SNPs) genotyped in a variety of candidate genes are available for analysis in NHANES. We performed identity-by-descent (IBD) estimates in three self-identified subpopulations of Genetic NHANES (non-Hispanic white, non- Hispanic black, and Mexican American) using PLINK software to identify potential familial relationships from presumed unrelated subjects. We then compared the PLINKidentified relationships to those identified by an alternative method implemented in Kinship-based INference for Genome-wide association studies (KING). Overall, both methods identified familial relationships in NHANES III and NHANES 1999-2002 for all three subpopulations, but little concordance was observed between the two methods due in major part to the limited SNP data available in Genetic NHANES. Despite the lack of genome-wide data, our results suggest the presence of cryptic relatedness in this epidemiologic collection and highlight the limitations of restricted datasets such as NHANES in the context of modern day genetic epidemiology studies.

Towards a phenome-wide catalog of human clinical traits impacted by genetic ancestry.

BACKGROUND: Racial/ethnic differences for commonly measured clinical variables are well documented, and it has been postulated that population-specific genetic factors may play a role. The genetic heterogeneity of admixed populations, such as African Americans, provides a unique opportunity to identify genomic regions and variants associated with the clinical variability observed for diseases and traits across populations. METHOD: To begin a systematic search for these population-specific genomic regions at the phenome-wide scale, we determined the relationship between global genetic ancestry, specifically European and African ancestry, and clinical variables measured in a population of African Americans from BioVU, Vanderbilt University's biorepository linked to de-identified electronic medical records (EMRs) as part of the Epidemiologic Architecture using Genomics and Epidemiology (EAGLE) study. Through billing (ICD-9) codes, procedure codes, labs, and clinical notes, 36 common clinical and laboratory variables were mined from the EMR, including body mass index (BMI), kidney traits, lipid levels, blood pressure, and electrocardiographic measurements. A total of 15,863 DNA samples from non-European Americans were genotyped on the Illumina Metabochip containing ~200,000 variants, of which 11,166 were from African Americans. Tests of association were performed to examine associations between global ancestry and the phenotype of interest. RESULTS: Increased European ancestry, and conversely decreased African ancestry, was most strongly correlated with an increase in QRS duration, consistent with previous observations that African Americans tend to have shorter a QRS duration compared with European Americans. Despite known racial/ethnic disparities in blood pressure, European and African ancestry was neither associated with diastolic nor systolic blood pressure measurements. CONCLUSION: Collectively, these results suggest that this clinical population can be used to identify traits in which population differences may be due, in part, to population-specific genetics.

Leveraging Epidemiologic and Clinical Collections for Genomic Studies of Complex Traits.

BACKGROUND/AIMS: Present-day limited resources demand DNA and phenotyping alternatives to the traditional prospective population-based epidemiologic collections. METHODS: To accelerate genomic discovery with an emphasis on diverse populations, we--as part of the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study--accessed all non-European American samples (n = 15,863) available in BioVU, the Vanderbilt University biorepository linked to de-identified electronic medical records, for genomic studies as part of the larger Population Architecture using Genomics and Epidemiology (PAGE) I study. Given previous studies have cautioned against the secondary use of clinically collected data compared with epidemiologically collected data, we present here a characterization of EAGLE BioVU, including the billing and diagnostic (ICD-9) code distributions for adult and pediatric patients as well as comparisons made for select health metrics (body mass index, glucose, HbA1c, HDL-C, LDL-C, and triglycerides) with the population-based National Health and Nutrition Examination Surveys (NHANES) linked to DNA samples (NHANES III, n = 7,159; NHANES 1999-2002, n = 7,839). RESULTS: Overall, the distributions of billing and diagnostic codes suggest this clinical sample is a mixture of healthy and sick patients like that expected for a contemporary American population. CONCLUSION: Little bias is observed among health metrics, suggesting this clinical collection is suitable for genomic studies along with traditional epidemiologic cohorts.

Extracting Primary Open-Angle Glaucoma from Electronic Medical Records for Genetic Association Studies.

Electronic medical records (EMRs) are being widely implemented for use in genetic and genomic studies. As a phenotypic rich resource, EMRs provide researchers with the opportunity to identify disease cohorts and perform genotype-phenotype association studies. The Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study, as part of the Population Architecture using Genomics and Epidemiology (PAGE) I study, has genotyped more than 15,000 individuals of diverse genetic ancestry in BioVU, the Vanderbilt University Medical Center's biorepository linked to a de-identified version of the EMR (EAGLE BioVU). Here we develop and deploy an algorithm utilizing data mining techniques to identify primary open-angle glaucoma (POAG) in African Americans from EAGLE BioVU for genetic association studies. The algorithm described here was designed using a combination of diagnostic codes, current procedural terminology billing codes, and free text searches to identify POAG status in situations where gold-standard digital photography cannot be accessed. The case algorithm identified 267 potential POAG subjects but underperformed after manual review with a positive predictive value of 51.6% and an accuracy of 76.3%. The control algorithm identified controls with a negative predictive value of 98.3%. Although the case algorithm requires more downstream manual review for use in large-scale studies, it provides a basis by which to extract a specific clinical subtype of glaucoma from EMRs in the absence of digital photographs.