We improve the proof and results of Theorem 1 in [3] by means of the modified classical backward algorithm which is a generalization of the scalar case. We give also a more refined truncation error bound for this class of vector continued fractions under a slightly stronger condition.