Plasma digoxin levels and the interbeat interval signal in atrial fibrillation

Fifty-eight patients with atrial fibrillation treated with digoxin were studied to determine the correlation between serum digoxin levels and the ventricular rate. The study was a computer-based exercise, processing a signal consisting of R-R intervals, derived by point-digitising electrocardiograms. The means, variances and centers of gravity of power-spectra from the signal were correlated with serum digoxin levels, the peripheral pulse and with each other. A poor negative correlation of -0.31 was calculated between the means of interbeat-intervals and serum digoxin levels. Other Ecg-derived rate parameters did not correlate any better with serum digoxin. The mean interbeat-interval in a group of patients with higher serum digoxin levels (2 ng/ml) was significantly (p is less than or equal to 0.05) shorter than in a group with a low level (1 ng/ml). The negative correlation and this significant difference are best explained by the gradual increase in the dose administered to non-responders by the attending physicians who did not fear over-dosage because of frequent serum level determinations. It is concluded that the serum digoxin level is a poor predictor of the ventricular rate in patients with atrial fibrillation because of marked individual differences. These are due to the poor representation by serum levels of drug concentration at the point of interest (A-V node) and the non-linearity of the chronotropic effect of digitalis.