Systems and methods include accessing a plurality of cardiac indications. A heart rate variability metric is produced by analyzing the plurality of cardiac indications using a measurement from a class of nonlinear measurements. Nonlinear measurements include, but are not limited to, approximate entropy, X-Y scatter from a Poincaré plot, fractal dimension, and detrended fluctuation analysis, in various examples. Based on the heart rate variability metric, a cardiac ischemic state is detected.