A Genome-Wide Association Study of Peripheral Arterial Disease

The Electronic Medical Record Phenotypes and Community-engaged Genomic Association Study aims to identify genetic variants mediating susceptibility to peripheral arterial disease (PAD). The study leverages a biorepository of blood samples of 1688 PAD cases and 1649 controls, and the electronic medical record (EMR) to annotate the biorepository. PAD cases were identified from the vascular laboratory database as having an ankle brachial index (ABI) <0.9 at rest or after exercise or having non-compressible vessels. Controls were without prior history of atherosclerotic vascular disease and when tested, no evidence of ischemia on a stress test. Phenotypes and environmental exposures including age, ethnicity, demographic and anthropometric data are derived from the Mayo Electronic Medical Record (EMR). Comorbidities were determined using algorithms for diabetes and hypertension based on ICD-9 codes and medication use. Relevant laboratory data, including lipid levels, fasting blood sugar and serum creatinine at index date or within a 1-year window of the index date were extracted. Medication classes at index date were identified using Mayo's Natural Language Processing-based system with RxNorm codification and NDF-RT terminologies mapping. Smoking status was confirmed by natural language processing of clinical notes. Genotyping of ~600,000 SNPs across the genome is being conducted at the Broad Institute using the Illumina 660W platform. Statistical analyses will be conducted to identify genetic variants associated with susceptibility to PAD.]]>
Pseudocode to Select T2DM (Type 2 Diabetes Mellitus) Cases & ControlsHypertension Phenotype CriteriaMayo Algorithm for Identifying PAD patients from the EMRPAD patients. The PAD patients were identified from the non-invasive Vascular Laboratory at the Mayo Clinic Rochester. Patients seen at Mayo Clinic who are suspected of having PAD/claudication are referred for lower extremity non-invasive arterial evaluation at this laboratory. We used the following criteria to define presence of PAD: 1) an ankle brachial index of <0.9 at rest or 1 min after exercise on the treadmill, along with an abnormal continuous wave Doppler signal in one of the lower extremity arteries, or 2) presence of non-compressible vessels. We excluded patients with PAD secondary to vasculitis, radiation to the abdomen or lower extremities, trauma to lower extremity artery, thrombophilia and in situ thrombosis formation, and end-stage renal disease. Control subjects. Controls subjects were identified from patients referred to the Cardiovascular Health Clinic for screening for cardiovascular disease. A large proportion of these patients underwent stress exercise ECG to rule out coronary artery disease. We excluded patients who had a positive stress ECG, those who were younger than age 50, or those who had history of prior coronary heart disease. A proportion of the subjects who underwent exercise ECG also underwent measurement of ankle brachial index. The prevalence of an abnormal ABI in patients who had a negative stress ECG was <1%.]]>
Motivation for the study: Sequencing of the human genome, subsequent validation of single nucleotide polymorphisms (SNPs), and advances in genotyping technology with decreasing costs have made possible studies that involve genotyping of 500,000 or more SNPs across the genome (genome-wide association studies). Combining genome-wide association data with clinical information from the electronic medical record (EMR) provides new opportunities to identify genetic variants that influence susceptibility to the common 'complex' diseases that afflict humans; this knowledge can be used to improve diagnosis, prognosis, prevention, and treatment of these diseases. Existing health care systems can be leveraged for genomic research given the potential for speed, efficiency, cost-effectiveness, and rapid translation to improved care of patients. However, optimal approaches for conducting genome-wide research in biorepositories to identify genes that influence susceptibility to complex diseases have yet to be established. An important goal of the project is to delineate optimal strategies for extracting comprehensive and unbiased phenotype/exposure data from the Mayo EMR, obtaining suitable consent for genome-wide association studies and broad data sharing, and interacting with participants, communities, and IRBs. We will attempt to define potential sources of bias and error in phenotypic and exposure data derived from EMR, and identify approaches for improving data collection and standardization for genomic research. Additionally, we will identify sources of concern among biorepository participants and other relevant groups regarding current consent and consultation processes for application of genome-wide technologies, and for sharing of data; define needs for additional consent and/or consultation, and recommend approaches for improving consent and consultation processes in biorepository-based research. Using state of the art statistical genetic methods, we will identify genetic variants associated with two distinct phenotypes of PAD using phenotype and covariate information derived from EMR data. We will make the resulting data rapidly and widely available for research use and work collaboratively to share experience and expertise across participating biorepositories.]]>