What we usually do for big tables is to use the sampling option of "update statistics".
for example, we use the following command:
update index statistics <object_name> using 2000 values with sampling=10 percent
Try testing with sampling in a test/preprod environment.
Vincent