What to do when there is unit root, you use differences and everything is insignificant?

[Matrassenverkoper]

When testing for unit root I found unit root in all my independent variables so i took differences to correct for unit root.
In my new regression I used the first lag of my dependent ( I tested for this) and the differences of all the other variables and regressed it on my dependent variable. However, all in sudden all my variables are insignificant, how is this possible and more important how can I correct this?

[Jorge Eduardo Perez Perez]

There's nothing to correct, this probably means that the signficant results you had before differencing were spurious because of the nonstationarity of the data.

[Eric de Souza]

Use the error correction model. Googling will find you lots of reference on this.