You might be interested in the ECG circuit given in the datasheet for the AD8420. Apparently, the problem with the circuit in the AD620 datasheet is that the DC excursions that occur when the patient moves are amplified by the AD620 limiting the gain that the AD620 can be set at, because those DC excursions are amplified and can saturate the AD620. That's not ideal because the inamp is where you want most of the gain due to the high CMRR provided by the inamp. In the AD8420 datasheet circuit, the DC excursions are canceled out in the inamp, by using an integrator to drive the REF pin, and so the gain of the inamp can be set much higher without saturating the output with DC excursions. Then the inamp is followed only with a low pass filter with a 50 Hz corner to attenuate 60 Hz noise and other higher frequency noise.