In this paper we prove that it is sufficient to use Pareto distribution to approximate the tail of a slowly varying heavy-tailed distribution. Several applications of the results are considered.