Well. I will write not only for you but for everyone who think that ssd will die fast in case of full usage.
Sandfroce had a bug with trim. Apple's SSD react on this command. Wearlevelling is tech that help to use all blocks equally to avoid destruction only in part of blocks. It works better when free space is maximal (i.e. full free ssd) . And if you don't rewrite full ssd every time it will die faster (and you are right).
But will calculate. MLC chips (I think it is not SLC) Apple's SSD that have resource about 10000 cycles of rewriting. Let's get some good 250GB ssd and will write 5000IOPS 4KB blocks (~20MB/sec, about 1.5TB per day) around the clock. Than 1% of this ssd will die around in half of month (really fast!). How much less times indeed writes average user per day? I think at least 100 times less (about 15GB). Than 1% of ssd will die in 50 months, it is about 4 years. Let's think now that we rewrite only 1GB every day from 5GB free, so it will serve about 1.2 years. Let's remember about 30% blocks for remapping in ssd (250*0.3=75). It is 25 times of 5GB than more than 25 years of usage. But I think, person who have 5GB free does not rewrite 1GB per day.
But it is theoretically, in real life SSD die time to time even it's resource is OK as other complex techniques do. And even when smart shows everything is OK (ex. Google's investigation about hdds) .
My example of long time living SSD: It's my old Kingmax (controller based on cheap Sandforce) 60GB that was bought 5 years ago (it's smart doesn't work, bugs of first Sandforce controllers). It was always used with 1-3GB free and several times was fully rewritten. Now it is working but I don't use it because it is too small for me. I think that more than 5 years is more than enough for average user.
PS: i used my old ssd for long time and rewrite a lot of data to it (but it's smart broken and shows that I wrote only 32gb to it when it is full)