On the first Zassenhaus conjecture for integral group rings