This paper presents and compares five analytical formulas for the approximation of stop-loss premiums. Two of them, based on the inverse Gaussian distribution, are not widely known. The authors also suggest a technique which improves the precision of these approximations for portfolios.