Mary Kaye Wojczynski, Ph.D.
Project Overview:
Coronary heart disease (CHD) remains the number-one killer in the United States. Many health conditions associated with heart disease are known. Additionally, both heart disease and its associated health conditions are known to “run in families”. Thus, a family history of heart disease and its related health conditions is a common part of a standard physical exam. But what exactly does it mean for a disease to “‘run in families”? Formally, researchers interpret this phrase in three ways: 1) shared family lifestyle habits (such as smoking, poor diet, lack of exercise, etc.) are responsible for the clustering among family members; 2) disease risk genes are being transmitted through the family generations; or 3) a combination of both family and genetics. However one frames the “‘run in families”’ question, it is necessary to have family datasets that have thorough disease classification, vital status, lifestyle, genotype, and clinical data in order to formally research the effect of family on disease.
The National Heart Lung and Blood Institute (NHLBI) Family Heart Study (FamHS) is such a family dataset. It was begun in 1992 when NHLBI wanted to develop a national resource in which to search for cardiovascular genes and familial factors predisposing to heart disease. Investigators at Washington University in St. Louis successfully competed to become the Coordinating Center for this endeavor (Province, PI), giving us inside access to this unique resource. Approximately 25,000 subjects from three-generation families were studied in four different areas of the country for all major disease conditions. About half of those families were chosen at random for study, but the other half were chosen because they had high rates of heart disease. About 6,000 of those people were also given an intense, four-hour physical exam in which all of the major known risk factors for heart disease were carefully measured (along with a number of promising new novel risk factors). Furthermore, FamHS successfully obtained funding for a second clinical exam between 2002 and 2004, where many participants were reassessed for the same conditions and risk factors. Additionally, African ancestry participants were newly recruited during this exam, allowing the FamHS to assess a minority population as well. In recent years, this resource also obtained funding for genotyping of genome-wide single nucleotide polymorphisms and exomic variants.
All of the clinical risk factor data on the FamHS families and genotypic data are already collected, cleaned, and ready for analysis at Washington University. All that is needed (and all that is being asked in this application) is to follow up on these subjects, to find out which ones have survived and which have died in the intervening years since the last National Death Index (NDI) search. Fortunately, the NDI is a single, efficient resource for getting nearly complete information on all deaths that have occurred in the U.S. since the last clinical exam. Additionally, since a proportion of the FamHS participants are Medicare-eligible, we would also like to search the Centers for Medicare & Medicaid Services databases to obtain information on incident events since the last clinic visit.
We will then use state-of-the-art statistical models to assess the importance and interplay of the measured clinical factors on survival and co-occurring diseases, with special emphasis on enhancement of CHD risk prediction models. The results of these analyses should more clearly define if family clustering increases current risk prediction, if newer CHD risks increase current prediction, and if validated genetic variants associated with CHD and its co-occurring diseases add to risk prediction models. Furthermore, the results could have the potential to be incorporated into new risk prediction models which could lead to earlier CHD intervention, more aggressive intervention or treatment strategies, and potential advancement of personalized medicine for CHD.
Progress Report:
The overall goal of this research was to obtain health outcomes data for participants of the NHLBI Family Heart Study (FamHS). The FamHS includes 4,491 extensively phenotyped and genotyped participants, of whom 622 were of African ancestry. With data on mortality and morbidity outcomes for the participants, we can expand the usefulness of the FamHS by performing epidemiologic outcomes research in light of genotypic profile, thus providing a new vantage point from which to examine the complex associations between genotype and phenotype in this population.
Aim 1: Obtaining mortality outcomes. We have obtained from the National Death Index (NDI) the vital status and cause of death (if deceased) through 2005 of all 4,491 FamHS participants.
Status: Conversion into SAS datasets useable for analysis is ongoing.
Aim 2: Obtaining morbidity outcomes. Our sample is restricted to those who are Medicare-eligible, determined using age 65. For these participants, we propose to query the Centers for Medicare & Medicaid Services (CMS) databases for incident events and event dates described by International Classification of Disease (ICD) codes, aimed at cardiovascular disease outcomes as well as chronic conditions related to cardiovascular disease (e.g. stroke, diabetes, angina, hypertension, hyperlipidemia, asthma, chronic obstructive pulmonary disease, atrial fibrillation, myocardial infarction, coronary artery bypass graft). The information would come from medical claims files from CMS, including the Medicare Enrollment segment, Chronic Conditions segment, Medicare Provider and Analysis Review (MedPAR) file, outpatient claims and carrier claims files, with the assistance of the Research Data Assistance Center (ResDAC).
We originally sought data for year 2004, the first year after the final recruitment of African ancestry (AA) participants (in 2004, there are 1,812 European ancestry [EA] and 127 AA Medicare-eligible participants). However, the decision was made to change the years of follow-up, so requested a search in 2003 for 564 participants (507 EA and 57 AA) who are Medicare-eligible (≥ 65 years old) and a search in 2004 for 1,212 participants (1,085 EA and 127 AA).
Status: Data has been obtained and is awaiting conversion into SAS datasets for analyses.
Aim 3: Morbidity and mortality outcome prediction analyses. We will perform epidemiologic analyses using the mortality and morbidity data as outcomes, incorporating traditional and novel risk factors, family histories of disease and validated genetic risk variants to assess health outcomes prediction.
Status: Waiting for analysts to recode Aim 1 and Aim 2 data into useable SAS datasets. I will then perform the analyses.
NOTE:
In the beginning of the funding, I was asked to cut my budget in half without changing the scope of work. This has been the source of numerous delays, especially as funding is not sufficient to pay for timely conversion the NDI and CMS data into useable SAS datasets. Once the datasets have been converted and can be analyzed, I anticipate producing at least one, if not two, publications in the ensuing 18 months.
Despite the delays, the experience I have gained in dealing with red tape, budget management, and administration and personnel time issues will be invaluable to my future as a scientist.
