Just play with the TOW (2) endpoint and the test statistic.
BUT there are several case, mainly based on knowing if the TOW population variances are equal or not.
Verify that the two sample t test at level x of H0= mu_x=mu_y versus Ha: mu_x not equal to mu_y rejects if and only if the confidence interval for mu_x-mu_y does not contain zero.
I know this statement is true...but I don't know how to verify mathmatically...Anyone could help? Thanks!