Use of Cardiac Magnetic Resonance and Echocardiography in Population-Based Studies

Abstract

Population research studies are often directed toward eliciting the association of physiological measurements (eg, left ventricular [LV] function) and clinical variables (eg, hypertension) with outcomes.1 Although cardiovascular magnetic resonance (CMR) is known to be accurate and versatile, until recently, the diffusion of CMR technology was too limited by technical and logistic challenges to consider its use in population studies on a large scale (>10 000). However, with recent technical developments, CMR has reached a level of maturity and ease of use, which makes its use in large-scale population studies a practical reality for the first time. The goal of this review is to facilitate the process of selecting imaging methods for population research studies based on design requirements and existing experience with the techniques. There are 4 fundamental aspects of imaging that are pertinent in population studies: validity, feasibility, accuracy, and reproducibility. ### Validity Selection bias2 is a critical issue in population studies. Inappropriate patient selection may lead to problems in extrapolating sample information even to the population from which the sample was derived, which is a core task of a population-based study. External validity may also be limited by selection. In contrast to this, greater variance in imaging measurements may require larger numbers, as discussed in the later section on accuracy and validity, but may have relatively less importance in population studies than in clinical trials.3 Patient selection is, therefore, critical to external validity. This has been considered carefully in the echocardiographic literature. For example, before the incorporation of 2-dimensional (2D) measurements of LV mass, in addition to M mode, the feasibility of LV mass measurement was somewhat limited, for example, in the Framingham Heart Study and Cardiovascular Health Study. In contrast, more recent studies, such as the Strong Heart Study and Losartan Intervention For Endpoint Reduction in …