To amplify on that, if the investigators did repeated statistical analyses after dividing the patients up according to different eGFR values (eg, looking at the data when eGFR<60 is used, then again using eGFR<50, etc....) the P values would have to be adjusted for the fact you are doing repeated tests. It gets complicated, and the numbers of subjects with low GFR would not be enough for this type of procedure.