|
BACKGROUND/RATIONALE:
RA is a systemic inflammatory disorder that can progress to joint destruction, severe disability, and premature mortality. RA afflicts approximately 1% of the US population and the VHA serves over 67,000 RA patients. DMARD therapy has been proven efficacious and cost effective for RA. Current guidelines by the American College of Rheumatology recommend that RA patients with active disease receive DMARD therapy unless contraindicated. Despite this recommendation, many RA patients are not receiving appropriate DMARD therapy while at the same time other patients are continuing on DMARD treatment with little clinical benefit while being exposed to significant potential toxicity.
There is a critical need for improved health services systems to assure the appropriate DMARD treatment of RA patients as personalized medical care. This goal can only be achieved by establishing solid predictive models to help optimize DMARD therapy to assure administration of these agents in the most efficacious and cost effective manner.
OBJECTIVE(S):
Specific Aim #1. Expand the current VARA database to consolidate information from multiple sources for VARA patients. This process will involve two steps. First, to expand of the capacity of the VARA database to extract and record CPRS data to include specific information on current and prior DMARD therapy and data on concurrent adverse events. Second, data abstraction and extraction from CPRS into the VARA database.
Specific Aim #2. Complete genotyping of all VARA patients for critical susceptibility alleles. This genotyping will involve seven SNPs encompassing each of the critical genes: PTPN22, PAD4, CTLA4, STAT4, and TRAF1-C5.
Specific Aim #3. Develop a predictive model for DMARD response through analysis of clinical efficacy response and adverse event history associated with clinical, serologic, and genetic features in VARA patients. After identifying associations with individual DMARDs, multi-variant analysis will be used to determine the weighted importance of the associations identified. A best fit model will be developed on the basis of these observations for future testing and validation in separate cohorts. Future work is anticipated for HSR&D funded investigations of this model in the prospective management of RA patients to optimize clinical outcomes.
METHODS:
Specific Aim #1. Expand the current VARA database to consolidate information from multiple sources for VARA patients. This process will involve two steps.
Step # 1. Expand the capacity of the VARA database to extract and record CPRS data. The first step will be to work with our computer consultants at Boulder Consultants to develop supplemental software to expand the capacity of the VARA database to include the specific information on current and prior DMARD therapy and data on adverse events described below.
Step #2 Data abstraction and extraction. Data collection from CPRS will involve both data abstraction by chart review and data extraction using electronic methodology.
Data abstraction. A data abstraction template has been developed for the collection of DMARD therapy, clinical response, adverse events, and laboratory data as part of a research project by Dr. Melissa Reily under the supervision of the principal investigator. This instrument has been tested and refined on 76 VARA patients in Salt Lake City. The time for abstraction of data from one CPRS chart is one hour per patient. Using this estimate of one hour for data abstraction and entry into the VARA web-based database, we will need 1000 hours of study coordinator time to abstract the information on the currently enrolled 1000 patients. This proposal will have two currently available staff work 100% of effort for four months to complete this work. This data abstraction will record DMARD response and adverse events in VARA patients.
Data extraction. Using extraction methodology similar to that described above, additional extraction templates will be developed for recording DMARD information. This extraction technology uses specifically designed templates and reminder systems within CPRS. These extraction templates will collect the information on DMARD use including start date, numbers of prescriptions dispensed, dose, and directions for administration. Key laboratory variables associated with DMARD toxicity will also be abstracted. As stated above, this current methodology is being used to extract clinical response data, co-morbidities, and DMARD history into the VARA website. This project will essentially expand these current capabilities.
Specific Aim #2. Complete genotyping of all VARA patients for critical susceptibility alleles. As described above, significant genotyping has already been completed in this population. Full HLA-DRB1 testing of the critical MHC region and ADME genes is currently ongoing and will be completed by the time this project is funded. However, there are several important deficiencies in the current genotyping database that must be completed before an analysis of the VARA population could be undertaken for analysis of DMARD efficacy and toxicity as proposed. Genotyping will involve seven SNPs encompassing each of the critical genes under investigation: PTPN22, PAD4, CTLA4, STAT4, and TRAF1-C5. The use of multiple SNPs will be required to appropriately evaluate potential genetic associations between these genes and clinical outcomes because a single SNP may not detect the clinically important variant.
The technology for these analyses is highly standardized and can be completed rapidly within the time frame of this grant. DNA samples for this analysis are currently in storage and ready for testing. All genotyping will be done at Salt Lake City under the direction of Dr. Roger Wolff. This information in concert with demographic and other pertinent clinical information, clinical disease activity, serologic finding, and foundational genotyping for HLA-DRB1 and ADME genes will allow construction of the models described below.
Specific Aim #3. Develop a predictive model for DMARD response through analysis of clinical efficacy response and adverse event history associated with clinical, serologic, and genetic features in VARA patients. Computer software and co-investigators in our TREP are available to assist with the rapid analysis of these data. Data analysis is anticipated to be completed within two months after the conclusion of data abstraction and extraction. The first step will be to determine the frequency of polymorphisms for each SNP in the VARA patients. The gene associations will also be tested for linkage disequilibrium.
The population genetics will be evaluated to identify if an association exists between the polymorphism of the selected genes with DMARD response characteristics and adverse event histories. As an initial analysis, we will focus on methotrexate, the most commonly employed DMARD, in the VARA cohort. As noted above, we estimate a 60% response rate and 40% adverse event rate in patients receiving methotrexate in the 76 VARA patients at Salt Lake City reviewed to date. In our preliminary work, 68 subjects (89%) of the initially reviewed VARA cohort in Salt Lake City had received methotrexate as a DMARD at some point during their therapy. Thus, estimate of power analysis for this DMARD is based on use of 90% of the entire VARA population. Assuming that the outcomes in Salt Lake City are representative of the entire VARA cohort, the ability to confidently identify associations is outlined in the following power analysis for VARA subjects using projected data on methotrexate assuming a 60% clinical response rate and 40% adverse event rate. All adverse events will be consolidated for this initial analysis with subsequent work focusing on individual adverse event types. With anticipated sample sizes of 1,000 VARA patients (900 exposed to methotrexate) and the data listed above for our single SNP analysis of PTPN22, STAT4 and TRAF-1, we expect to be well-powered to detect small differences in the rates of these alleles in association with clinical response and adverse events. Thus with a two-sided = 0.05, we will have 80% power to detect differences in the prevalence of PTPTN22 homozygosity of just 5-7% in comparison of the subjects with and without a clinical response.
Similar evaluations would be completed for other DMARDs and biologic agents. We recognize that the power of the evaluations may change with the lower number of subjects receiving these drugs.
After identifying associations with individual DMARDs, multi-variant analysis will be used to determine the relative importance of the associations identified. Recognizing the potential role of complex interactions between genetic, serologic, and other patient characteristics, we will also use state-of-the art Classification and Regression Tree (CART) analysis. CART software is available to investigators in the SLC TREP and our co-investigators in Omaha. A best fit model will be developed on the basis of these observations for future testing and validation in separate cohorts and hopefully eventually tested prospectively in the management of RA patients.
FINDINGS/RESULTS:
We have been able to identify 1100 RA subjects who will qualify for these evaluations. We are currently moving forward with the data abstraction and preparation for genotyping as noted in the methods.
IMPACT:
This work with be the foundation for future HSR&D projects to prospectively test this model and prove that directed therapy will optimize outcomes in RA patients. The overall impact of this type of intervention could result in improved efficacy, reduced adverse events, and a more cost effective application of resources to the treatment of RA patients.
PUBLICATIONS:
None at this time.
DRA:
Aging and Age-Related Changes, Chronic Diseases, Health Services and Systems
DRE:
Pathophysiology, Treatment, Technology Development and Assessment
Keywords:
Pharmaceuticals, Genomics, Arthritis
MeSH Terms:
none
|
