Analysis and Interpretation of Primary and Derived Data Sets in Cardiology

Abstract

Investigators collect data and present them in a way that offers the best insight regarding the questions at hand. To facilitate understanding of certain aspects, it may occasionally be useful to rearrange primary data and formulate them as derived variables. For example, the travel distance divided by the invested time yields average velocity (as m/s). Problems may arise when interpreting ratios that fail to have a physical dimension. For example, current TV-sets have a fixed ratio for height and width, implying that we need an additional detail to define its size. Size then is determined by the diagonal, which can be calculated from the two sides using the Pythagorean theorem. Similarly, paired hemodynamic variables may be expressed as ratios. Again, a fixed ratio may refer to a variety of underlying primary data which require consideration if the ratio is unitless. In this survey, we evaluate several derived metrics commonly used in cardiovascular studies, and offer comprehensive analysis strategies.

Are left atrial strain metrics and a biochemical marker superior to traditional left atrial size measures when identifying occult atrial fibrillation?

“…Therefore, the question remains if recently launched metrics such as strain truly offer incremental value above traditional primary metrics (i.e., ESV and EDV) that also have the advantage of carrying a physical dimension. 14 Zhao et al 9 cite various publications that claim the superiority of strain-related metrics over traditional "anatomic" parameters (such as LA volume), when identifying subclinical myocardial impairment.…”

Statistics at high‐altitudes: Relevance for the interpretation of metrics that reveal cardiac morphology and performance

“…Only a few parameters reached statistical significance, which observation is not really surprising in view of the explorative design of the majority of these studies 5–8 . No hypotheses were formulated based on a transparent and realistic pathophysiological model, and no distinction was made between primary and derived variables 9 …”

“…[5][6][7][8] No hypotheses were formulated based on a transparent and realistic pathophysiological model, and no distinction was made between primary and derived variables. 9 Therefore, an investigation into the nature of various metrics currently in vogue may be instructive. Similar as in many other studies, the authors analyzed a variety of ratio-based metrics, including e'/a' , E/A, EF, FAC, GLS for both LV and RV, as well as MPI, all without phys-ical units.…”

Prediction of a super‐response to cardiac resynchronization therapy as guided by left ventricular end‐systolic volume size

“…The blue curve is based on nonlinear regression analysis by employing a robust mathematical expression which warrants that EF approaches 100% for the theoretical case when ESV becomes extremely small, as well as an asymptotic behavior for the higher ESV range. 6,7,11 (B) Typical trajectories for the gain in ejection fraction (EF), expressed in absolute percent points, in dependence of initial (baseline) endsystolic volume (ESV) for three levels of ESV reduction following cardiac resynchronization therapy. In the higher ESV range any fixed variation of ESV results in a relatively smaller change for EF, compared to a reverse remodeling starting point in the smaller ESV range (beyond ESV 50 ml).…”

“…The rectangular green area refers to baseline EF values ≤35% as considered in the study by Zand et al 3 The nonlinear behavior of the curve implies that equal changes for EF (as marked by the yellow boxes, here reflecting a 5% point step, as also marked on the curve by the colored circles and corresponding dots) translate into unequal variations of ESV (shown as the red and purple bars). The blue curve is based on nonlinear regression analysis by employing a robust mathematical expression which warrants that EF approaches 100% for the theoretical case when ESV becomes extremely small, as well as an asymptotic behavior for the higher ESV range 6,7,11 . (B) Typical trajectories for the gain in ejection fraction (EF), expressed in absolute percent points, in dependence of initial (baseline) end‐systolic volume (ESV) for three levels of ESV reduction following cardiac resynchronization therapy.…”