Hapax in Job. -d is probably due to the ARM origin of the word. There is an old tradition of translating this word as 'testiculus' (a recent discussion supporting this opinion see in [Pope Job 272]).
Jastrow's explanation as 'blown up' does not look convincing; a similar shift of meaning see in AKK pah_allu 'thigh' vs. 'penis' in other SEM (see *paḥ(a)l- 'penis', No. ) or in AKK birku 'knee; lap; male and female sexual parts' [CAD b 255]
Syrian Aramaic:puḥdā (-u- < *a in the vicinity of p ?) 'femur, clunis' [Brock 562]
Mehri:ǝfh_ād (to read ǝfh_ād_?), pl. fǝh_yōd_ 'flesh and bone of upper leg from knee to hip' [JM 110].
ǝfh_ād is likely to be a misprint in [JM], since all the derivatives have -d_ as the last consonant [ibid.]; cf., however, QISHN fh_ād 'cuisse' [SS L 2 216]
Jibbali:fah_d_ 'flesh and bone of upper leg from knee to hip' [JJ 67]
Harsusi:ǝfḫād_
Notes:SOQ fáh_id [LS 335], fh_ed 'cuisse' [SSL 4 87] must be an Arabism because of -h_- in place of the expected *ḥ.
Note a derived meaning 'tribe': ARM: PLM pḥd/pḥz 'tribal union'; ARB fah_id_-/fah_d_-/fih_d_- 'la plus petite subdivision d'une tribu' [BK 2 552]; ESA: QAT fh_d_ 'clan, family, sub-tribe' [Ricks 129]; MSA: MHR fǝh_ǝ́d_ǝt 'tribe' [JM 110], JIB fh_ǝ́d_ǝt do. [JJ 67], SOQ fâḥid 'peuple' (also fh_edeh, which must be an Arabism because of -h_-) 'peuple' [LS 335] (an areal Arabian semantic evolution?).
On this semantic shift see [M. Cohen Genoux], [Kogan, Militarev].