You need to specify the maximum operating frequency and/or fastest edge-rate. Impedance is a function of frequency.
You should also think carefully about the size of your design/length of traces. You may actually be worrying about something that isn't a problem.
Er is an intrinsic material property. It isn't effected by the dimensions (to the first order). FR-4 is a composite material consisting of glass fibers and resin Er about 2 and 6, with a typical weighted average of 4.6 or so.
The maximum operating frequency of my circuit is 1.5GHz. I need to
maintain impedance in 100ohm differential pairs and 75ohm lines about
30mm maximum lenght each
In your case, you are at the low end of the "gray" zone. In general, you can ignore reflections (the consequence of mismatched impedances) when the propagation distance is less than an order-of-magnitude of the wavelength. This works because any reflecting waves would have time to decay (die out) before the next clock edge. They'd have to bounce off the ends more than five times and still have enough energy to distort the signal beyond your trigger threshold.
In copper-over-FR-4, the speed of light is approximately half that of free-space (1.5e8 m/s). There the corresponding wavelength at 1.5GHz is 10cm. Your tracks are 3cm. If you can reduce the track length to 1cm (or less) you can easily ignore impedance control.
That said, don't take it as license to do something silly. You should still route cleanly, hold spacing and track width constant, neck smoothly, etc -- but you don't have to pay for "impedance control" services from the fabricator).
What I Would Do
There is no magic to "impedance control". The fabricator uses a model for the material (which they will happily share with you -- since it ties you in to their factory) to determine a track width/spacing and uses it. When you pay for impedance control services they simply add an extra step to the fab process where they will verify that they got it right.
Ask the fabricator for the target track widths and spacing.
Use it ;-)
I wouldn't pay for impedance control in this case at these frequencies given your route length. Remember the energy of the reflection is in proportion to the ratio of the impedance mismatch. If you are off only a little, and your distances are small, and your material is close to spec... you'll be fine.
If this is an expensive low-volume product where you can't afford respin or 5% performance degradation (et. al. r.) then by all means ignore this advice.