Development and Use of Network Infrastructure for High-Throughput GWA Studies

The "Development and Use of Network Infrastructure for High-Throughput GWA Studies" project is collaboration between Group Health Cooperative (GH), the University of Washington (UW), and the Fred Hutchinson Cancer Research Center (FHCRC). It is one of 5 sites participating in The Electronic Medical Records and Genomics (eMERGE) Network funded by the National Human Genome Research Institute (NHGRI) with additional funding from the National Institute of General Medical Sciences (NIGMS). The overall eMERGE project is designed to assess whether linking biorepositories of patients in healthcare delivery systems with electronic medical records (EMRs) is an efficient strategy for high-throughput genome wide association (GWA) studies and whether information can be pooled across sites. The primary phenotype in the Group Health/UW Aging and Dementia eMERGE study project is dementia.

Phenotyped participants originated from four population-based sources. Seattle-area members of GH (a large integrated health care system in Washington State) consented and enrolled in 1) the University of Washington Alzheimer's Disease Patient Registry (ADPR) and 2) the Adult Changes in Thought (ACT) study, 3) Marshfield Clinic Personalized Medicine Research Project (PMRP), a population-based DNA, plasma and serum biobank of 20,000 adults, 4) Vanderbilt's BioVU, a de-identified DNA biobank.

The ADPR (PI: Eric B. Larson; NIH/NIA U01 AG 006781) is a population-based registry of incident dementia cases designed to identify all new dementia cases within GH from 1987 to 1996. Potential dementia cases were identified through referrals from primary care physicians, mental health services, and neurologists, as well as review of CT scans, neurology, emergency room, geriatrician, and mental health clinic logs, and hospital discharge diagnoses. Medical history, physical, laboratory testing, and neuropsychological testing were performed on all consenting potential cases. These data were reviewed by a consensus conference including a neuropsychologist, study physicians, and an epidemiologist. After discussion of the case, dementia status is assigned using DSM-III-R criteria. Alzheimer's disease status was assigned using the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria. The study base of the ADPR population was stable with an attrition rate of less than 1%/year.

The ACT study (PI: Eric B. Larson; NIH/NIA U01 AG 006781) is an ongoing community-based cohort study designed to determine the incidence of Alzheimer's Disease (AD), other types of dementia, and cognitive impairment and to determine risk factors for these conditions. The original cohort of 2,581 was enrolled in 1994-1996. An expansion cohort of 811 was enrolled in 2000-2002. Continuous enrollment to keep 2,000 persons enrolled and at-risk for dementia outcomes was begun in 2005. To date the study has accrued more than 4,000 participants. Dementia-free participants are enrolled in this cohort and evaluated every two years by study personnel with many evaluations including an assessment of cognitive functioning using the Cognitive Abilities Screening Instrument (CASI), which is an extended version of the Mini-Mental State Examination that has been widely used in epidemiological studies of the elderly. Persons with CASI scores lower than 86 or in whom interviewers suspect unusual cognitive decline are followed up with dementia diagnostic evaluations which include physical and neurological examinations and neuropsychological tests. Medical records are reviewed, and laboratory tests and neuroimaging are obtained if not available. The dementia diagnoses and dementia subtypes are determined during a consensus conference attended by the examining clinicians and other study physicians, a neuropsychologist, and research nurse. After discussion of the case, dementia status is assigned using DSM-IV criteria. Alzheimer's disease status was assigned using NINCDS-ADRDA criteria. Individuals diagnosed with dementia are seen again one year following diagnosis using the same procedures. Individuals not diagnosed with dementia are returned to the pool of at-risk individuals and are seen at their next routinely scheduled biennial visit. The ACT sub-sample is stable; for the original cohort, median enrollment in GH was 19 years prior to joining the ACT study, and 85% of the cohort has ≥ 10 years of GH enrollment.

DNA for the ADPR participants were obtained through a companion study, Genetic Differences in Cases and Controls (PI: Walter Kukull; NIH/NIA R01 AG007584). DNA obtained through both studies were extracted from blood using Gentra Systems Puregene methods. DNA concentration is determined by UV optical density. All samples are checked for quality by OD 260/280 ratio. For long-term storage, samples are aliquoted and stored at -70°C.

The Marshfield study, Personalized Medicine Research Project (PMRP), a population-based DNA, plasma and serum biobank of 20,000 adults. PMRP was started in 2002 as a 3 Phase project. Completed in April 2004, the objectives of Phase I were to build and develop a large population-based biobank with DNA, plasma and serum samples to facilitate genomics research. Phase 1 was also to educate, inform and consult with the Marshfield Epidemiologic Study Area population and communities concerning potential studies, create the DNA foundation of the personalized medicine database and build the bioinformatics tools to store securely and analyze genotypic and phenotypic data. Phase I tasks included operation of the ethics and security advisory board, scientific advisory board and community advisory group. More than 18,000 residents aged 18 and above from 19 different zip codes surrounding Marshfield, WI were invited to participate in Phase I. After providing written informed consent, participants completed brief questionnaires that included questions about demographics, some environmental exposures, family history of disease, and adverse drug reactions, as well as family members living in the area. Participants provided 50ml of blood from which DNA was extracted and plasma and serum samples were stored. The informed consent process allowed access to electronic medical records and included language about non-disclosure of personal research results. A tick-off box was included so that participants could either allow or decline subsequent recontact for future research studied. The objectives of Phase II are to create the phenotypic database, establish the scientific and administrative infrastructure to support genetic mapping of the DNA and the initial discovery projects and genotype a sufficient portion of the genetic material to support these discovery projects. The objectives of Phase III are to expand the discovery projects, complete the genotyping of the genetic database and expand physician/health care provider education and community consultation.

The Vanderbilt BioVU model uses discarded blood leftover from clinical care and de-identified clinical information, both of which are tracked with a Research Unique Identifier that is permanently disconnected from the medical record number used to generate it. The Vanderbilt IRB determined that the project does not qualify as "human subject" research under §46.102(f)(1)(2). The program has been reviewed by OHRP twice (the latest immediately prior to launch), and the determination of the Vanderbilt IRB that the project is consistent with the 8/10/2004 guidance has been validated. The model includes a simple opt out mechanism developed for patients who do not wish to have their samples included. Planning (including community involvement, ethics and IRB discussions, focus groups, development of operating procedures and pilot studies to assess them) began in May 2004, and the resource commenced sample collection in February 2007 and as of May 2009 contains approximately 56,000 samples. The current resource will be available to all Vanderbilt investigators who sign a Data Use Agreement. The resource represents Vanderbilt's broad patient population and thus provides a potential resource for research in a range of common and rare diseases as well as drug response or medication safety.

BioVU patient protection model: There has been extensive involvement by the ethics community and the IRB to put in place procedures to provide privacy protection. The overall protection plan is multifaceted and includes a combination of technology and policy: (1) creation of the SD described above; (2) return of only the specific clinical data items requested by investigators (rather than the complete de-identified chart); (3) submission of any specific proposed project to the IRB and to a separate protocol review committee, both of which must approve; (4) implementing a Data Use Agreement that permits only approved queries and data analyses, specifically prohibits attempts at re-identification (at the risk of institutional sanctions), and mandates redeposit of all genetic information generated in a research study back into a genomic database; and (5) tracking and auditing all use.

• Study Weblinks:
  • eMERGE
  • Group Health Center for Health Studies
  • Marshfield Clinic Research Foundation
  • Personalized Medicine Research Project
• Study Design:
  • Case-Control
• Study Type:
  • Case-Control
• dbGaP estimated ancestry using GRAF-pop
• Total number of consented subjects: 3756
• Subject Sample Telemetry Report (SSTR)

• BioProject
• PubMed
• BioSample
• MeSH
• PMC

Our study includes 257 ADPR participants (all of whom are dementia cases) and 2613 ACT participants (470 dementia cases, 2125 dementia controls) with viable specimens for genotyping.

Alzheimer Disease Patient Registry (ADPR):

Included Seattle-area members of GH enrollees who were 60 years or older, with symptoms of dementia, and determined to be dementia cases.

Patients were excluded if their physicians felt it was inappropriate for them to be involved the study (for example they were too ill), if dementia was diagnosed before January 1, 1987, or they had no memory problems.

Adult Changes in Thought (ACT):

Included a simple random sample of Seattle-area members of Group Health Cooperative enrollees who were 65 years or older and had either 1) a baseline Cognitive Abilities Screening Instrument (CASI) of value of at least 86 of 100 or 2) a baseline CASI of 80-86 who scored 86 or higher on retesting 2 months later and whose Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) score was normal, or 3) a baseline CASI less than 80 who on complete diagnostic evaluation did not meet clinical criteria for dementia.

Exclusion criteria included residents of a nursing home, participation in other studies, diagnosis of dementia, or 1) scoring 80-86 on the baseline CASI whose score remained low on retesting 2 months later, 2) scoring 80-86 on the baseline CASI with acquired cognitive impairment corroborated by the IQCODE and Blessed Dementia Rating Scale or no corroborating information available after a complete diagnostic evaluation and a consensus review ruled them ineligible. Living participants were also excluded if they did not provide consent to have their information submitted to dbGaP.

Research Based Case Determination:

In both ADRP and ACT research cohorts, the diagnosis of dementia was based on the Diagnostic and Statistical Manual (DSM) of mental disorders (ACT: IV; ADPR: III R) and the diagnosis of Alzheimer's Disease was based on the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) working group criteria. The dementia and Alzheimer's Disease research definitions are not applicable to the Marshfield and Vanderbilt participants.

EMR Based Case Determination:

For comparability with other eMERGE sites (Vanderbilt (n=272 cases) and Marshfield (n=224 cases)), study investigators also determined dementia status using a criterion based entirely on electronic medical records. The EMR derived definition can not be applied to the older GH ADPR cohort. Applying the automated algorithm to EMR data, dementia was defined as 5 occurrences of a dementia related ICD-9 code OR a prescription for a medication used to treat dementia. ICD-9 codes included Alzheimer's disease (331.0), senile dementia - uncomplicated (290.0), presenile dementia (290-10-290.13), senile dementia delusional [paranoid] features ( 290.20), senile dementia with depressive features ( 290.21), senile dementia with delirium or confusion (290.3), arteriosclerotic dementia (290.40-290.43), dementia due to alcohol (291.0-291.2), dementia due to drugs (292.82), dementia, unspecified (294.8), dementia not classified as senile, presenile, or arteriosclerotic (294.10-294.11), Pick's disease of the brain (331.11), other frontotemporal dementia (331.19), dementia with Lewy bodies (331.82). Prescription medications included donepezil, Aricept, galantamine, Razadyne, Reminyl, Nivalin, rivastigmine, Exelon, tacrine, Cognex, memantine, Namenda, Axura, Ebixa, and Akatinol.

• Primary Phenotype: Dementia
• Alzheimer Disease
• Aging
• Aged
• Age Factors

• Principal Investigator
  • Eric Larson, MD, MPH. Group Health Research Institute, Seattle, WA, USA.
• Co-investigator, Principle Investigator at University of Washington
  • Gail Jarvik, MD, PhD. University of Washington, Seattle, WA, USA.
• Co-investigator, Principle Investigator at Fred Hutchinson Cancer Research Center
  • Christopher Carlson, PhD. Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
• Co-Investigators
  • Wylie Burke, MD, PhD. University of Washington, Seattle, WA, USA.
  • Paul K. Crane, MD, MPH. University of Washington, Seattle, WA, USA.
  • Stephanie Malia Fullerton, PhD. University of Washington, Seattle, WA, USA.
  • Walter Kukull, PhD, MS. University of Washington, Seattle, WA, USA.
  • Ge Li, MD, PhD. University of Washington, Seattle, WA, USA.
  • Katherine M. Newton, PhD. Group Health Research Institute, Seattle, WA, USA.
• Research Associate
  • Leslie Spangler, VMD, PhD. Group Health Research Institute, Seattle, WA, USA.
• Project Managers
  • Kelly Ehrlich. Group Health Research Institute, Seattle, WA, USA.
  • Monica Fujii. Group Health Research Institute, Seattle, WA, USA.
• Programmers
  • David Carell, PhD. Group Health Research Institute, Seattle, WA, USA.
  • Robert Harrison. Group Health Research Institute, Seattle, WA, USA.
  • Gene Hart, MS. Group Health Research Institute, Seattle, WA, USA.
  • Noah Weston. Group Health Research Institute, Seattle, WA, USA.
• Research Scientist-UW
  • Susan Brown Trinidad, MA. University of Washington, Seattle, WA, USA.
  • David Crosslin, PhD. University of Washington, Seattle, WA, USA.
• Research Scientist-Fred Hutchinson Cancer Research Center
  • Andrew McDavid. Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
• Principal Investigator at Marshfield
  • Catherine A. McCarty, PhD. Marshfield Clinic Research Foundation, Marshfield, WI, USA.
• Co-Investigator at Marshfield
  • Peggy Peissig, MBA. Marshfield Clinic Research Foundation, Marshfield, WI, USA.
• Principal Investigator at Vanderbilt
  • Dan Roden, MD. Vanderbilt University, Nashville, TN, USA.
• Co-Investigator at Vanderbilt
  • Josh Denny, MD. Vanderbilt University, Nashville, TN, USA.
• Funding Source
  • U01 HG004610-01 or RFA-HG-07-005 ). National Institutes of Health, Bethesda, MD, USA.
• Institute
  • National Human Genome Research Institute, Bethesda, MD, USA.
• Genotyping Center
  • Johns Hopkins University Center for Inherited Disease Research (CIDR), Baltimore, MD, USA.
• Funding Source for Genotyping
  • U01HG004438-01. National Human Genome Research Institute, National Institutes of Health, Bethesda, MD, USA.