i am sure people's practical experience is more valuable, but...

if we assume that all the resistance to flow for the filter comes from the tubing (and none from the filter itself, a simplification i appreciate). flow is proportional to length divided by radius to the power 4.

so assume there is 2000mm of tube, and 8mm radius, and then compare this to 1900mm tube with 8mm radius and 100mm with 6mm radius in the device. i won't bore you with the maths, but this increases overall resistance to flow by about 11%.

obviously there is some resistance to flow from the filter itself, and from the inline device, e.g. bioballs in the reactor cyclinder, therefore the actual increase in total resistance is likely to be less.

does a roughly 10% reduction in flow by adding a 12mm reactor as opposed to a 16mm reactor seem reasonable based on actual experience? or is the resistance from the filter and inline device such that the reduction is acutally much less?