We have not consistently recorded this information, but it could be interesting doing so (also on a retrospective basis). Hence, what I am saying is purely empirical. Transfusion increments are widely around 1 g/dl, sometimes the benefit is minimal, sometimes around 2 g/dl. Of course values may be subject by a host of different variables: test variables (timing of second hb determination - during transfusion - sampling error - or just after transfusion? infusion of packed RBC requires some time - maybe in older patients more - to redistribute fluids from the extravascular compartment to "dilute" the concentrated RBC transfused; type of determination - with CBC or with POCT/ABG testing?), besides patient's variables that I assume may be of some importance (body size, state of hydration and of concomitant i.v. fluid/diuretic therapy, ongoing overt/occult bleeding...)
Moreover sometimes physicians administer a pro re nata dose of i.v. diuretic (eg furosemide) in patients with congestive heart failure after the transfusion to counteract the possible haemodynamic consequence of an acute expansion of circulating volume, so it is possible that further haemoconcentration may be due to this practice.