All other factors being equal (i.e. obstacle free path), the intersection's different slope and elevation at TORA start seem to play a role. I just managed to run some calculations and could replicate the situation for A321 with 33k engines at 72t obtaining higher FLEX from RWY 06 E here
http://lis.rlp.cz/ais_data/aip/data/valid/a2-pr-adc.pdf.
Upon closer scrutiny, you did caught me red handed (
why is that no surprise 
) and
my statement under L above
in weird cases, intersection departures may yield better payload / higher AST/FLEX is incorrect.
Max payload is reduced for intersection (2.5t).
For the low given weight AST/FLEX figure is 1 deg greater at the intersection. I suppose this is only valid as long as AST/FLEX is at max -25% and the thrust setting is affected by nothing else than density elevation. I am not qualified to speculate what impact the different input AST/FLEX have on FADEC commanded N1 or resulting EGT.
A321-211 0 kt 13C 1020 hPa CF:2 A/C: Off EAI: Off
LKPR 06 DRY ELEV 1202 SLP -0,4 deg RWY 3715 CWY 300 STPW 0 OBST 0 LineUp 90 deg
PLW 98324 kg Max Structural 89000
at 72000 kg Flex 55 V1 122-145 VR 145 V2 147 ACC ALT 2707
LKPR 06 E DRY ELEV 1191 SLP -0,0 deg RWY 3060 CWY 300 STPW 0 OBST 0 LineUp 90 deg
PLW 95877 kg Max Structural 89000
at 72000 kg Flex 56 V1 138-149 VR 150 V2 150 ACC ALT 2691
I sincerely wish the audience would save time spent over this negligible niche in favor of reading some of yours/OSs work on Vmca and wet considerations.
Yours,
FD (the un-real