Hourglassing is common in First order reduced Integration elements. But is it possible in second order and other higher order reduced integration elements also?
Hourglassing is a phenomenon due to underintegration, and not to the order of shape functions. So yes, it can appear in higher order reduced integration elements as well. However, commercial packages usually have internal algorithms that should amend the problem automatically on run-time. Check the documentation and run some simple tests, if you are in doubt.
in isoparametric, super parametric, sub parametric or other transformation if the displacement at a Gauss point reduce to zero, while at nodes are non zero, then the HG may happen. this kind of bad performance is the characteristics of numerical methods.
For standard isoparametric displacement based elements hourglass can occur only using a single gauss point and bilinear shape functions. In this case in fact you can apply bending to the element without modifying the distance between the single gauss points (placed on the element center) and each edge, i.e. at zero strains. This zero energy mode can not occur using quadratic elements with 4 gauss points, because in this case you can not apply bending to the element without changing the distance from each gauss point to each edge.