Is it possible to detect atrial fibrillation by simply using RR intervals?

Atrial fibrillation is the most common sustained arrhythmia in clinical practice worldwide. Several algorithms have been developed to detect atrial fibrillation, which either rely on atrial activity analysis or are based on the irregularity of RR intervals. This paper is addressed to study the latter type of algorithms. The main question is whether there is sufficient information in the sequence of RR intervals for reliable detection of atrial fibrillation and whether the atrial fibrillation can be diferentiatedfrom other significant ECG arrhythmias. We have tested various types of algorithms existing in the technical papers utilizing MIT-BIH databases. Except the atrial fibrillation all other arrhythmias have some regularity, self-similarity and some degree of predictability. Consequently, algorithms utilizing only the values of RR intervals without their order may misclassifY other irregular rhythms as atrial fibrillation. The best algorithms use the scatter plot of successive RR diferences or Sample Entropy. The error rate was about 5%. It is possible to create a robust atrial fibrillation detection algorithm relying only on RR intervals considering their places in the sequence.