Improving Cardiovascular Risk Prediction--Biomarkers and Beyond; Implications for Astronaut Selection and Monitoring During Prolonged Spaceflight

The most likely cause of a non-traumatic life- or mission-threatening medical event in astronauts would be from acute cardiovascular disease (CVD). Current risk prediction models utilize only traditional atherosclerosis risk factors and focus narrowly on coronary heart disease events rather than global cardiovascular risk, ignoring outcomes such as heart failure or atrial fibrillation that could also be potentially mission-threatening. Numerous studies have evaluated novel risk markers in an attempt to improve CVD risk prediction, with several promising imaging and blood-based biomarkers identified. Most of these studies have investigated the incremental predictive value of a single biomarker added to a traditional risk factor model, with a few reporting combinations of biomarkers. Moreover, few studies have evaluated strategies for risk prediction that cross testing modalities. Such a multi-modality approach has the potential to markedly improve CVD risk prediction among potential and existing astronauts, and would have direct relevance to the general population.

Our primary objective was to develop a consortium of biomarker and aerospace medicine leaders, with expertise in multiple different testing modalities, and with access to robust existing databases, to identify and validate novel strategies to enhance global CVD risk prediction over two time windows: 1) 10-20 years, representing the full career of the astronaut and 2) 2-5 years, representing the planning and operational phase of a manned mission to Mars. The team included many of the leading biomarker and imaging experts in the U.S. This team of collaborative investigators accessed data from multiple existing cohort studies to develop two distinct multi-modality risk prediction tools, one based on 10-year global CVD risk (Aim 2) and one based on 3-year CVD risk (Aim 3). These models evaluated novel testing modalities on top of standard risk factors, including coronary calcium (a measure of the extent of coronary atherosclerosis), multiple blood based protein biomarkers that reflect inflammation, cardiac injury, and cardiac stress, as well as ECG measurements of cardiac hypertrophy and imaging-based assessments of cardiac function. Finally, we proposed an exploratory aim to work with NASA researchers in the Human Research Program to explore the feasibility of transforming the Longitudinal Study of Astronaut Health (LSAH) into a prospective state-of-the-art cohort study of the astronaut corps.

The grant had 2 scientific aims (Aims 2 and 3). Aim 2 has been completed ahead of schedule with a major manuscript published, and Aim 3 analyses are ongoing with promising preliminary results. The primary scientific aims of the grant required pooling of data from large cohort studies. Each of these studies has a unique regulatory structure, scientific proposal system, and approval process. The goals of the first year of funding were to obtain the necessary approvals and data transfer agreements to being the data pooling process. This was accomplished during the first year, when we obtained approval for data transfer from the Dallas Heart Study (DHS), Multiethnic Study of Atherosclerosis (MESA), Atherosclerosis Risk in Communities (ARIC), and the Framingham Heart Study (FHS).

The goal for the second year of funding was to secure data transfer, and to perform harmonization of data elements. This goal was successful accomplished by the end of 2015, ahead of schedule. We then moved to analyses for Aim 2 of the grant. We derived in MESA and validation in DHS a multimodality risk prediction tool that led to marked improvement over traditional risk prediction algorithms both for predicting 10 year atherosclerotic risk as well as global and cause-specific CVD risk. Five screening tests (coronary calcium screening by CT, left ventricular hypertrophy by ECG, and elevated levels of NT-proBNP, hs-cTnT, and hs-CRP) markedly improved global CVD risk prediction compared with standard risk assessment strategies. We created a simple score, consisting of the number of abnormal CVD screening tests. In both MESA and DHS a > 25-fold gradient of risk for CVD was seen across the range of scores. Of particular relevance for NASA, participants with zero abnormal tests results have an extremely low risk for any CVD outcome over 10 years of follow-up. The findings replicate extremely well across the two distinct cohort studies. The final results were presented at the 2017 Human Research Program Investigators Workshop in Galveston, TX, and were published in Circulation in June 2017 (de Lemos et al. A Multimodality Strategy for Cardiovascular Risk Assessment: Performance in Two Population-Based Cohorts. Circulation. 2017; 135:2119-32. PMID: 28360032).

Analyses for Aim 3 are underway presently. These required data pooling due to the smaller number of events over short term follow-up. Preliminary results suggest that the multimodality strategy provides even more robust stratification of short term (3-year) risk compared with the 10-year risk models developed in Aim 2 and recently published.

Aim 4 was an exploratory aim, designed to explore the feasibility of transforming the Longitudinal Study of Astronaut Health into a prospective state-of-the-art cohort study of the astronaut corps. Although several productive discussions were held, we did not resolve whether such an ambitious study is feasible in the future.

More »