Surrogate end-points are used when the true end-points are costly or time consuming.In a typical set up we observe a fixed proportion of true-and-surrogateresponses, and the remaining proportion are only-surrogate responses. It is obviousthat the inclusion of such only-surrogate end-points increase the efficiency ofassociated estimation. In this present paper we want to quantify the gain in efficiencyas a function of the proportion of available true responses. Also we obtain theexpression of the gain in true sample size at the expense of surrogates to achieve afixed power, as a function of the proportion of true responses. We present ourdiscussion in the two-treatment set up in the context of odds ratio.