There's a hacky workaround, but it depends on how many drives you have and what your tolerance for downtime is. I had 2x8TB drives in RAIDZ1 configuration (sda, sdb), and I added an extra 8TB drive (sdc).

My solution was:

offlining one drive (sdb), degrading vpool1

creating a vpool2 in RAIDZ1, with sdb, sdc, and an 8TB sparse file on a flash drive

offlining the sparse file, degrading vpool2

sending a ZFS snapshot from vpool1 to vpool2

destroying vpool1, and adding sda to vpool2 to replace the sparse file