PurposeTo investigate the clinical therapy for giant intraocular foreign bodies (IOFBs) and evaluate the best treatment method with minimum secondary injury.MethodsWe retrospectively analyzed the data of 73 eyes of 73 patients with ocular trauma caused by giant IOFBs between January 2016 and December 2018. The IOFB size, localization, shape, and magnetic properties were recorded. The best corrected visual acuity (BCVA), ocular tissue injuries, entrance wound, interval time from injury to second phase surgery, silicone oil removal, and globe recovery were also observed. The cases were divided into three groups based on the following IOFB extraction paths: limbus path, the pars plana path, and the entrance wound path. The BCVA, IOFB size and shape, the wound, endophthalmitis, and silicone oil removal were compared among the three groups.ResultsThe IOFBs were 46 cases of magnetic and 27 cases of nonmagnetic, with a shape of thin flat in 19 cases, thick flat in 12 cases, long in seven cases, and irregular in 35 cases. Multiple damages were caused by the giant IOFBs, mainly involving the severe cornea, lens, and retina injuries. The postoperative BCVA increased compared with the preoperative BCVA (z = −6.06, P < 0.01). The rate of recovery from blindness was 40.85% (29/71). The thin flat IOFB and long IOFB resulted in a better postoperative BCVA than the other two IOFB shapes (all P < 0.05). The irregular IOFB had a poorer silicone oil removal rate than the other three IOFB shapes (all P < 0.05). The IOFB extraction followed the limbus path in 18 cases, pars plana path in 27 cases, and entrance wound path in 28 cases. The IOFB length and width in the pars plana path group were significantly lower than that in the limbus path group (all P < 0.05), the preoperative BCVA of the pars plana path group was superior to that of the limbus path group (P < 0.05), and the IOFB length, width, and entrance wound length in the pars plana path group were significantly lower than in the entrance wound path group (all P < 0.05). But the postoperative BCVA in the pars plana path group was not better than that in the other two groups (all P > 0.05). The postoperative BCVA of the entrance wound path group was significantly superior to that of the limbus path group (z = −2.01, P = 0.04), while there was no difference between the two groups in IOFB length, width, entrance wound length, or preoperative BCVA (all P > 0.05).ConclusionThe entrance wound path would benefit to minimize secondary injury in giant IOFB extraction procedure, compared with the limbus and pars plana path.