Right, I see your problem. Unfortunately there is no easy way to avoid this scenario. Not on our side anyways.
The best you can do is to download the population source yourself, request the isochrones, neglect the total_pop property and do a raster statistical analysis with isochrone geometry on the population raster you have now on your local machine. To get rid of overlapping areas, you just dissolve all of the isochrone geometries (you might have to a geometric difference of the ranges in case you’re looking at multiple time/distance ranges at once and then dissolve each range separately).
Find the source here:
http://cidportal.jrc.ec.europa.eu/ftp/jrc-opendata/GHSL/GHS_POP_GPW4_GLOBE_R2015A/GHS_POP_GPW42015_GLOBE_R2015A_54009_250/V1-0/
From GHSL:
https://ghsl.jrc.ec.europa.eu/ghs_pop.php
Careful, it’s 1 GB zipped.