To investigate epidemiologic and clinical characteristics, prognostic factors, and visual outcomes of posttraumatic intraocular foreign body (IOFB).
Retrospective, consecutive case series.
Information from 1421 eyes (1299 inpatients) with IOFB selected from all open-globe injury cases (4968 eyes of 4865 inpatients) in 15 tertiary hospitals in China over a 5-year period were collected from a standardized preformulated database of eye injuries. Factors affecting retinal lesions or detachment, development of endophthalmitis, and visual outcome were analyzed statistically.
IOFBs represented 28.60% of all open-globe injuries. Of the 1421 eyes, endophthalmitis developed in 232 (16.76%), and independent protective factors included primary repair within 24 hours (odds ratio [OR], 0.59; P = .006) and self-sealing wounds (OR, 0.69; P = .013). At discharge or follow-up, the enucleation rate was 7.53%, best-corrected visual acuity (VA) improved in 54.33%, and 4.85% had no light perception. Wound length of at least 3 mm and IOFB located in the posterior segment increased the risk of retinal lesion or detachment, and both (OR, 1.66; P < .001; OR, 4.04; P < .001) were significant negative predictors for visual outcome, along with wound larger than IOFB in largest length (OR, 2.38; P = .002) and endophthalmitis (OR, 2.01; P = .0003). Better initial VA (OR, 0.76; P < .001) was a protective factor for final VA.
For IOFBs, primary wound closure by repair within 24 hours or self-sealing independently reduces the risk of endophthalmitis. Worse presenting VA, larger wounds, IOFB in posterior segment, wound larger than IOFB in largest length, and endophthalmitis predict worse visual outcome.
Intraocular foreign bodies (IOFBs) are a common and serious form of eye trauma that can result in visual loss. Apart from the prominent feature of IOFB retention, IOFBs also include common characteristics of penetration, rupture, or perforation, depending on the mechanism of injury. Because of the multitude of potential findings, prompt and full assessment of patients with possible IOFB retention based on the early signs is important to provide an effective therapeutic plan, intraoperative guidance, prognosis, and counseling. The treatment aim is effective repair of ocular abnormalities while avoiding complications, such as endophthalmitis, resulting from insufficient evaluation or unnecessary treatment, such as enucleation, resulting from erroneous judgment. The present study was based on a multicenter review of inpatients with IOFBs and open-globe injuries in China during a 5-year period. We analyzed the epidemiologic and clinical characteristics of posttraumatic IOFBs and identified factors that affect the development of retinal lesions, endophthalmitis, and visual outcome.
Consecutive medical records of all patients with open-globe injuries admitted to 15 tertiary referral hospitals in China between January 1, 2001, and December 31, 2005, were reviewed retrospectively. Each case was recorded using a standardized preformulated data sheet, and the records were maintained in a computerized eye injury database. Of the 4865 individuals (4968 eyes) with open-globe injuries, 1421 eyes of 1299 inpatients with a retained IOFB were eligible for analysis. Eyeballs removed within 24 hours after open globe injury (173 eyes) or after IOFB (37 eyes) were excluded from some analyses.
Detailed information of each injured eye with a retained IOFB was collected using a standardized data sheet, including patient information (age, sex, occupation, etc.); a thorough history (cause, circumstance, and timing of injury, etc.); previous vision, if known; clinical presentation; and treatments and outcomes at discharge or follow-up. In addition, the site and length of the wound in the globe wall; the number, material, location, and longest diameter of IOFBs; and the timing of IOFB removal all were recorded if known. Findings during surgery provided valid evidence. Slit-lamp examination, ophthalmoscopy, ultrasonography, computed tomography, and magnetic resonance imaging were performed, and the results were compared. Only when IOFBs were seen in ophthalmologic or imaging examination or surgery could the diagnosis of IOFBs be made. The data of routine imaging examination is not part of this report. Visual acuity (VA) reported herein were all best-corrected or pinhole VA.
Classification and definition of ocular trauma in this study was based on the Birmingham Eye Trauma Terminology. Open-globe injury indicates a full-thickness wound of the eyeball. Herein, IOFB indicates any open-globe injury with a retained IOFB, that is, penetrating, rupture, or mixed. A self-sealing wound was a wound of the globe wall that closed tightly by itself without primary repair. Diagnosis of endophthalmitis was made chiefly based on clinical characteristic symptoms and signs. Culturing of intraocular contents was performed only in some cases. The culture outcomes were not considered in the diagnosis of endophthalmitis. The size of the foreign body refers to the largest IOFB if there were multiple IOFBs in 1 eye.
All data were collected in an electronic database and were cross-checked for errors. Statistical analysis was performed using SPSS software version 17.0 (SPSS, Inc, Chicago, Illinois, USA). Categorical variables were analyzed using the chi-square test. Continuous variables were evaluated for normality, and means were compared using a 2-tailed t test. Further multiple logistic regression analysis was conducted to predict the independent factors affecting the occurrence of endophthalmitis after the occurrence of an IOFB. A P value of less than .05 was considered statistically significant for all tests.
A total of 1421 eyes (28.60%) with a retained IOFB among 4968 eyes with open-globe injuries were included in the analysis. Bilateral IOFB occurred in 122 cases (9.39%). The age of inpatients with IOFB ranged from 1 to 78 years (mean, 31.39 ± 12.74 years), and males were 93.10% of the IOFB cases. Up to 76.35% of IOFBs occurred at work, and only 0.77% of the patients claimed to be wearing safety goggles at the time the injury occurred.
The common causes of IOFB included hammering (500 eyes; 35.19%), explosions (395 eyes; 27.80%; 243 eyes with explosions contributed to the 76.53% of eyes with multiple IOFBs), drilling and use of other machining tools (152 eyes; 10.7%), and others. The characteristics of the corneal or scleral wound and IOFBs are listed in Table 1 . The foreign bodies varied greatly in shape and size, with the diameter of the largest foreign body being 25 mm. The entry wound was larger than the largest diameter of the IOFB in 245 (17.24%) of the eyes.
|Zone||I: 67.00%; II: 23.29%; III: 9.71%|
|Length (mm)||<3 (44.48%); 3 to 5 (34.69%); > 5 to 10 (14.78%); > 10 (6.05%)|
|Intraocular foreign bodies|
|Number||Multiple (≥ 2) in 22.36% of eyes|
|Location||21.39% in anterior segment; 72.28% in posterior segment; 6.33% in anterior and posterior segments both; 8.3% in lens; 33.92% embedded in retina; 1.27% subretinally; 9.64% impacted in wound|
|Size (mm)||At least 15.20% > 5 mm|
|Material||53.48% magnetic; 69.07% metallic; 7.67% copper; 3.45% glass; 22.52% stone; 3.52% vegetable|
Ocular Findings and Predicting Factors for Retinal Lesions
The most common ocular findings at admission, during intraocular surgery, and at follow-up are listed in Table 2 . Macular lesions included macular detachment (61 eyes; 4.29%), macular hemorrhage (21 eyes: 1.48%), and macular hole (19 eyes; 1.34%). In 311 eyes with proliferative vitreoretinopathy (PVR), 74.28% (231 eyes) occurred within the initial wound tract or at the final resting site of the foreign body. Anterior PVR was observed in 12 eyes (3.86%), posterior PVR was observed in 276 eyes (88.75%), and both posterior and anterior PVR were observed in 23 eyes (7.39%). A chi-square test revealed that a wound size larger than 3 mm (especially those exceeding 10 mm) and foreign body located in the posterior segment increased the risk of retinal lesion. A wound larger than 3 mm (especially those exceeding 5 mm) and IOFB located in the posterior segment also increased the risk of retinal detachment ( Table 3 ).
|Findings||No. of Patients (%)|
|Breach of lens capsule||456 (32.09)|
|Lens subluxation||38 (2.67)|
|Lens luxation||6 (0.42)|
|Vitreous hemorrhage||807 (56.79)|
|Retinal detachment||329 (23.15)|
|Retinal breaks||79 (5.56)|
|Occlusion of retinal blood vessel||51 (3.59)|
|Macular lesions||101 (7.11)|
|Choroid detachment||38 (2.67%)|
|Sympathetic ophthalmia||3 (0.21%)|
|Traumatic cataract||1039 (73.12)|
|Proliferative vitreous retinopathy||311 (21.89)|
|Variable||Eyes with Retinal Lesion, n (%)||P Value||Eyes with Retinal Detachment, n (%)||P Value|
|Length of corneal/scleral wound (mm)|
|< 3 (n = 632)||337 (53.32)||95 (15.03)|
|3 to 5 (n = 493)||322 (65.31)||108 (21.91)|
|> 5 to 10 (n = 210)||138 (65.71)||<.001||75 (35.71)||<.001|
|> 10 (n = 86)||80 (93.02)||33 (38.37)|
|Intraocular foreign body location|
|Anterior segment (n = 304)||52 (17.11)||17 (5.59)|
|Posterior segment (n = 1027)||770 (74.98)||<.001||288 (28.04)||<.001|
|Both (n = 90)||55 (61.11)||24 (26.67)|
Endophthalmitis and Its Affecting Factors
Endophthalmitis was diagnosed in 232 (16.76%) eyes with an IOFB, excluding eyes that were removed within 24 hours after injury. Among these, intraocular specimens from 14 (63.64%) of 22 eyes produced a positive micro-organism culture. Based on univariate analysis, the IOFB material and diameter, regardless of whether the IOFB was removed, timing of IOFB removal, and self-sealing of wounds were not significant factors for the development of endophthalmitis. The location of the IOFB and the timing of the primary repair were statistically significant factors in the development of endophthalmitis ( Table 4 ). Logistic regression analysis provided somewhat different results: repair within 24 hours after injury was a significant protective factor. In addition, cases with self-sealing of the wound had a statistically significantly lower rate of endophthalmitis, whereas IOFB location in the posterior segment was not significant ( Table 5 ).
|Factor||Endophthalmitis, No. (%)||P Value|
|Material of IOFB (n = 1336) b|
|Metal (n = 923)||155 (16.79)|
|Nonmetal (n = 407)||60 (14.74)||.36|
|Coexisting (n = 6)||0 (0)|
|Vegetable IOFB (n = 1336) b|
|Yes (n = 47)||12 (25.53)|
|No (n = 1289)||203 (15.75)||.11|
|IOFB location (n = 1296) c|
|Anterior segment (n = 299)||35 (11.71)|
|Posterior segment (n = 997)||174 (17.45)||.02|
|Diameter of IOFB (n = 1157), mm d|
|<1 (n = 37)||4 (10.81)|
|1∼5 (n = 941)||143 (15.20)|
|> 5 to 10 (n = 117)||20 (17.09)||.36|
|> 10 (n = 62)||14 (22.58)|
|Timing of IOFB removal (n = 1384), h|
|< 24 (n = 209)||26 (12.33)|
|≥ 24 (n = 1175)||206 (17.53)||.07|
|Timing of primary repair (n = 695), h e|
|< 24 (n = 415)||50 (12.05)|
|≥ 24 (n = 280)||49 (17.50)||.044|
|Self-sealing of wounds (n = 1384)|
|Yes (n = 712)||129 (18.12)|
|No (n = 672)||103 (15.33)||.19|
a Excluding 37 eyes removed within 24 h after injury.
b Nature of 48 IOFBs were unknown.
c Excluding 88 eyes with foreign bodies in both anterior and posterior segments.
d Excluding 176 eyes with retained IOFB at discharge and 51 eyes unknown of foreign body size.
|Factor||P Value||Odds Ratio||95% Confidence Interval|
|Within 24 h of primary repair||.006||0.59||0.409 to 0.861|
|IOFB in posterior segment||.19||1.26||0.885 to 1.807|
|Nonmetal||.53||0.95||0.820 to 1.109|
|IOFB removal ≥ 24 h||.63||1.06||0.674 to 1.274|
|Minor diameter||.07||0.82||0.664 to 1.016|
|Self-sealing of wounds||.013||0.69||0.511 to 0.925|
Foreign bodies were removed during primary repair in 59.68% of all eyes. Because of multiple or occult foreign bodies, 63 eyes (4.43%) underwent 2 surgeries to remove the IOFB, and 1 eye underwent 3 surgeries. The delay of foreign body removal after the injury ranged from 15 days to 22 years (the latter a case in which a light bulb explosion resulted in the retention of glass fragments in the ciliary body and the patient was admitted 22 years later for recurrent iridocyclitis) in 18.50% of the cases and exceeded 1 year in 2.53%. Vitrectomy was performed in 807 (56.79%) eyes, among which 186 (23.05%) eyes underwent the procedure at least twice, 72 (8.92%) underwent retinotomy, 300 (37.17%) underwent retinal cryocoagulation, and 42 (5.20%) underwent a subretinal procedure. For eyes complicated by endophthalmitis, 22 eyes received an injection of medicine or irrigation of the vitreous cavity. Some inpatients (5.63%; 80 eyes) opted not to undergo surgical treatment for economic or other reasons and were not observed with symptoms of endophthalmitis in the duration of hospital stay. Some injuries, such as those resulting from explosion, resulted in many tiny foreign bodies located in the corneal stroma, iris, or ciliary body that were difficult to remove completely. In others, for example a foreign body in a lens without cataract, surgery to remove the IOFB was delayed. At the latest follow-up, IOFBs were retained in 176 (12.39%) eyes.
Outcomes and Affecting Factors for Final Vision
According to the outcome at discharge or follow-up, 146 (10.27%) eyes were filled with silicone oil, 219 (15.41%) eyes had an intraocular lens, 565 (39.76%) eyes had no lens, 238 (16.75%) eyes had a secondary cataract, 39 (2.74%) eyes had a permanent macular lesion, 21 (1.48%) had bulbus phthisis, and 15 (1.06%) eyes had optic nerve atrophy. A total of 107 (7.53%) eyes were enucleated after admission for no possibility of reconstruction of anatomic structure. A summary of presenting and final VA is shown in Table 6 . Compared with the initial VA, the final VA improved in 54.33% of cases, remained unchanged in 32.65%, and deteriorated in 13.02%.
|Visual Acuity or Enucleation||Presenting||Final|
|20/40 or better||126 (8.87%)||243 (17.10%)|
|20/100 to 20/50||103 (7.25%)||180 (12.67%)|
|5/200 to 19/100||137 (9.64%)||178 (12.53%)|
|LP to 4/200||926 (65.16%)||644 (45.32%)|
|NLP||129 (9.08%)||69 (4.85%)|
|Enucleation or evisceration||0||107 (7.53%)|
Table 7 presents predictors of poor visual outcome (VA < 20/400) identified by univariate analysis. The final VA had a significantly higher probability of being worse than 20/400 if the injured eyes were characterized according to an initial presenting VA of less than 20/200 on admission, the IOFB had a diameter of at least 2 mm, the IOFB located in the posterior segment or in both the anterior and posterior segments, the entry wound was in zone II or III, the length of the entry wound was at least 3 mm, the length of the entry wound was larger than the largest diameter of the IOFB, there were multiple IOFBs, or endophthalmitis developed. Final VA, however, was not affected by the timing of IOFB removal or the timing of primary repair within 24 hours after injury. Logistic regression showed that better initial VA was the protective factor, and IOFB located in the posterior segment, longer entry wound, the length of the entry wound being larger than the largest diameter of the IOFB, and endophthalmitis were the independent risk factors for final VA worse than 20/400 ( Table 8 ).
|Factor||Final VA < 20/400, No. (%)||P Value|
|Initial presenting VA (n = 1421)|
|< 20/200 (n = 1104)||789 (71.47)|
|≥ 20/200 (n = 317)||68 (21.45)||<.001|
|Diameter of IOFB (n = 1194), mm b|
|< 2 (n = 329)||191 (58.05)|
|≥ 2 (n = 865)||572 (66.13)||.0115|
|IOFB location (n = 1421)|
|Anterior segment (n = 304)||87 (28.62)|
|Posterior segment (n = 1027)||688 (66.99)||<.001|
|Both (n = 90)||82 (91.11)|
|Entry wound (n = 1421)|
|I zone (n = 952)||540 (56.72)|
|II/II zone (n = 469)||317 (67.59)||<.001|
|Length of entry wound (n = 1421), mm|
|< 3 (n = 632)||288 (45.57)|
|3 to 5 (n = 493)||322 (65.31)||<.001|
|> 5 (n = 296)||247 (83.45)|
|Length of entry wound larger than largest diameter of IOFB (n = 1421)|
|Yes (n = 245)||236 (96.33)|
|No (n = 1176)||621 (52.81)||<.001|
|Number of IOFBs (n = 1421)|
|Multiple (n = 318)||244 (76.73)|
|One (n = 1103)||613 (56.58)||<.001|
|Endophthalmitis (n = 1421)|
|Yes (n = 232)||174 (75.00)|
|No (n = 1189)||683 (57.44)||<.001|
|Timing of IOFB removal (n = 1384), h c|
|< 24 (n = 209)||111 (53.11)|
|≥ 24 (n = 1175)||709 (60.34)||.059|
|Timing of primary repair (n = 695), h d|
|< 24 (n = 415)||289 (69.64)|
|≥ 24 (n = 280)||182 (65.00)||.22|