As my development project nears the first release, I've not been able to make progress on this issue, despite a fair bit of debugging. I'm as clueless as before, I'm afraid. I must point out, though, that this library works really well in all other respects and I was able to bring the hibernation current down to 72 uA which includes the Teensy 3.2 and my PCB that it sits on.
Although not directly related to this issue, I was chasing a strange problem where the hibernation current was 72 uA at first but after the SD card was accessed for the first time after coming out of sleep, the sleep current thereafter went up to 480uA. Removing the SD card prevented this rise in current, so was pretty sure it was related to the SD card. The SD card sits on my PCB and has a VCC high side MOSFET switch to turn it off during sleep. After a bunch of debugging, I was reading through the Samsung SD card specification and it mentioned the importance of making sure the SPI lines were low before turning the VCC on or off. Otherwise, the SD card would get into an "unknown" state. I then manually set the SS, MOSI, MISO, and CK lines to low using digitalWrite, plus a 50 msec delay, before turning on VCC via the MOSFET. This fixed the issue and hibernation current was back down to 72uA. Somehow, this "unknown" state prevent the MOSFET switch from fully turning off and kept some current being supplied to the SD card.
Maybe this will help others, hopefully.