Bioelectric signals are often corrupted by noise. The most common form of noise is power-line interference and its harmonics. A convenient way for eliminating these unwanted components is to use a single or multiple notch filters. One of the problems about this approach is the effect of transient response of the filter at the beginning of its output in short time measurements. In this work, three initialization methods, which can be used to reduce/overcome this problem are reviewed and their performance and computational complexity are evaluated using ECG as an example signal. These methods are projection initialization, pole radius-varying filtering and vector projection. Additionally, some implementation variations and memory usage considerations are discussed. Our study shows that, pole radius-varying method is computationally cheap but introduces longer transient than the others. On the other hand, vector projection provides a more accurate reconstruction of the signal in the transient part of the output but with a more expensive computation. There are also two drawbacks about vector projection. One is its computational complexity dependency to the sampling frequency of the signal and the other is the fact that it cannot provide the results in real-time.