The US Department of Veterans Affairs is pulling together one of the largest cohorts ever assembled with the aim of advancing genomic-based disease research and the development of personalized therapies.
Over the next five to seven years, the Million Veteran Program, or MVP, aims to enroll one million veterans who will provide blood samples, fill out surveys about health and health-related behaviors, and allow secure access to VA and VA-linked medical and health information. DNA will be extracted from the blood sample and correlated with disease phenotypes, though the VA has yet to determine whether it will use genome sequencing or microarray technology for the analysis.
The voluntary program, which has been years in the making, was launched at a single VA medical center in Boston in January and expanded nationally last week. The program will be rolled out at VA medical centers over the next year. So far, 20 sites across the country are open for enrollment.
Only authorized researchers within the VA system and their collaborators will have access to the database, which will be used to determine which genetic variations are associated with particular health issues.
Blood samples will be stored in a biobank in Boston and labeled with a barcode instead of personal information in order to protect the privacy of participating veterans. Researchers who are approved to access samples and data will not receive the name, address, social security number, or date of birth of participants. Furthermore, researchers will not be able to download the data but will access it through the VA's secure GenISIS (Genomic Information System for Integrative Science) computing environment. The VA has a comprehensive electronic health records system, but MVP data will not be added to veterans' EHRs because it is considered research data.
"One of the challenges that has affected genome association studies is the replication problem," Timothy O'Leary, director of clinical science research and development at the Veterans Health Administration, told PGx Reporter. "You do, for example, a genome-wide association experiment on one population that identifies a series of genes that are associated with your disease of interest. Then you do another GWAS and it comes up with a small number of overlapping genes that are associated and many non-overlapping genes, and then people start to get a little bit confused by this."
O'Leary said that he and his VA colleagues "felt we needed to be able to deal with very large datasets in order to achieve the replication." The target of a million participants "was chosen with that idea of replication in mind — that at least for common conditions we could achieve replication data sets."
O'Leary said that the primary goal of the database will be to help study health issues related to military service, such as post-traumatic stress disorder, which is thought to have a genetic component based on studies with twins. Another topic will be susceptibility to cancers that have been connected with military service, such as chronic lymphocytic leukemia, which has been associated with Agent Orange, an herbicide used in the Vietnam War.
The cohort will also be useful for studying other diseases and conditions, however, such as cardiovascular disease and stroke, adverse drug reactions, and response or non-response to common therapeutic agents.
"I think the list of things that can be done over time, and as the number of enrollees grows, is limited only by the imagination," O'Leary said.
The first project in which the MVP will play a role is an ongoing clinical study, dubbed CSP572, which is investigating the role of genetics in bipolar disease and schizophrenia. The MVP cohort will serve as the reference sample for that study, O'Leary said. The goal is to recruit 9,000 individuals with bipolar disorder, 9,000 with schizophrenia, and 20,000 participants for the reference population. He estimated that it will take a year or two to collect enough samples for the study.
O'Leary noted that previous GWAS studies have looked at bipolar disease and schizophrenia, but said that the VA project will be different because it will focus on the functional disability that goes along with these diseases.
"In both cases you can treat a lot of the symptoms, but it doesn't necessarily get people back to a normal life," he said. "We're trying to understand the genetics associated with that because we know that some folks with schizophrenia are able ultimately to have a relatively normal life, and others do not. And we need to home in on this really important set of questions if people are going to develop better therapies."
Five Years in the Making
O'Leary said the VA began planning the MVP about five years ago by assessing whether such a project could be conducted ethically and with the full support of veterans.
In 2006, the VA created a Genomic Medicine Program Advisory Committee, a 13-member panel of genetics and genomics experts intended to lay the groundwork for implementing a genomic medicine strategy within the VA medical system.
The agency followed that with a collaboration with the National Human Genome Research Institute and Johns Hopkins' Genetics and Public Policy Center. GPPC conducted a survey of nearly a thousand veterans to assess their attitudes toward a project like MVP. The results, published in Genetics in Medicine in 2009, indicated that a vast majority — 83 percent — were in favor of such a database, while 71 percent said they would serve as research subjects if asked.
Following that work, O'Leary said the VA gave some thought to a model for participation that would best protect veterans' interests and ultimately decided on an opt-in model with "explicit" informed consent.
"From our perspective, that's part of the nature of the partnership," he said. "We want the informed consent to participate to be clearly and freely given because a veteran that participates in the program is giving a tremendous gift of information to the research community."
In practice, the VA sends mailings to veterans who attend participating MVP sites. Those veterans who decide to participate receive a copy of the informed consent document ahead of time, "so we can make sure that any questions they have are fully answered," O'Leary said. When participants come into the collection site they receive a verbal presentation that touches on some of the key elements of the informed consent.
"I think one of the really critical elements is that we want to make sure that anybody that participates in this program knows exactly what they're doing," O'Leary said. "We've all heard of cases where somebody has participated in a clinical investigation [and] someone came up to them and said, 'If you want to participate, here's this 20-page document. Read it and sign.' I'm pretty convinced that it many cases, you're not getting real informed consent out of that."
Participants have the opportunity to withdraw from the study later, O'Leary said.
So far, about 20 percent of the veterans who are contacted have agreed to participate and the project has enrolled around 1,200 people, with another 3,500 people who have either returned the questionnaire or have scheduled appointments at collection centers.
The program hopes to enroll between 50,000 and 100,000 veterans in the first year.
Caveats and Unknowns
While the MVP promises to serve as a useful resource for large-scale gene association studies, it won't be a perfect cohort for all studies.
"One thing that you gain from a million veterans is the opportunity to have a good representation of demographic subgroups, but there may be limitations," O'Leary said. "For example, only around 10 percent of our veterans are women right now. That means you have 100,000 women potentially in this data set when it's completed. That's a large number, but for some gender-specific conditions it may still be limiting."
The program is also still weighing the issue of what sort of technology it will use to conduct the genomic analysis. While the initial thinking was to use GWAS for the schizophrenia study, the plummeting cost of sequencing has made the decision of what technology to use a year from now a guessing game.
"We're building the airplane as we're flying it and it's a little hard to know for sure exactly what technology we'll use and exactly what informatics support will be required," O'Leary said, "so we're keeping an open mind and looking to match the appropriate technology at the time with the questions that are being asked."
The timeline for the decision on the technology platform depends on the rate of recruitment. "I'm not sure when we'll make a call," he said, but "it will have to be made in a timeframe so that we can either contract for the services; or buy [them], get up and running, and get comfortable with the technology that we decide to use."
The final decision on how to approach the analysis won't be made until enrollment in the study is complete, which could take up to a year or more.
"I think one of the characteristics of this program is that we're going to have to be flexible to changes in the technology over time," he said. "These are very fast moving targets, and one of the characteristics of the program is that it's going to work very hard to be technologically relevant at all times."
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at btoner [at] genomeweb [.] com.