This paper presents a novel Multiwavelet-based hybrid method for compressed sensing (CS) and reconstruction of ECG signals. The proposed method is patient-agnostic and does not require any prior patient-specific information. The proposed method removes low-frequency components by using multi-scaling wavelet functions, which sparsify the ECG signals in the multi-wavelet domain. A novel way of CS acquisition is used to acquire the high-frequency component of the ECG signal. The compressed sensed signal acquisition is performed based on sample difference thresholding. This approach ensures that a significant portion of the ECG signal will be acquired, without a need for a QRS complex or an R peak detector. The main compression is achieved by CS and linear and nonlinear quantization. Arithmetic encoding is used to improve compression further. The results show good reconstruction quality while maintaining a high compression ratio. The proposed method was evaluated for different multi-wavelet functions.