The p-values don't say anything about the size of the effect. With a large sample size, they are almost guaranteed to be small. In fact, the confidence interval

The p-values don't say anything about the size of the effect. With a large sample size, they are almost guaranteed to be small. In fact, the confidence interval after covariate adjustment comes very closing to containing the value 1, which suggests no substantive change between the groups.