Clinical Characteristics, Visual Outcomes, and Prognostic Factors of Open Globe Injuries

Background and Objectives: Open globe injuries (OGI) remain an important cause of visual impairment and loss, impacting all ages. A better understanding of the factors influencing visual outcomes is important in an attempt to improve the results of the treatment of OGI patients. The author aimed to contribute to this knowledge with the analysis of clinical characteristics, prognostic factors, and visual outcomes of their cohort of OGI patients. Materials and Methods: A retrospective medical record review was performed for 160 patients (161 eyes) who sustained an open globe injury between January 2015 and December 2017 and presented to the Hospital of Lithuanian University of Health Sciences. Data analyzed included age, sex, type, cause, place of OGI, initial visual acuity (VA), final best-corrected visual acuity (BCVA), and tissue involvement. Open globe injuries were classified using the Birmingham Eye Trauma Terminology (BETT) and Ocular Trauma Classification System (OTCS). Univariate analysis was conducted to evaluate the prognostic factors. Results: The mean age of the patients was 41.9 years. The male-to-female ratio was found to be 8.4:1. The home was the leading place of eye injury (59.6%), followed by an outdoor environment (14.3%) and workplace (11.8%). Penetrating injury accounted for 43.5%, followed by intraocular foreign body injury (39.1%) and globe rupture (13%). Overall, 19.5% of patients regained a good final vision of ≥0.5, but for 48.1% of them, eye trauma resulted in severe visual impairment (BCVA ≤ 0.02). In the univariate analysis, a bad visual outcome of less than 0.02 was correlated with bad initial VA, iris dialysis, hypotony, vitreous hemorrhage, and vitreous prolapse at presentation. Phthisis bulbi was correlated with eyelid laceration, iris prolapse, iris dialysis, hyphema, vitreous prolapse, vitreous hemorrhage, and choroidal rupture at initial examination. Conclusions: Open globe injury remains an important preventable cause of ocular morbidity. This study provides data indicating that open globe injuries are a significant cause of visual impairment in our research group.


Introduction
Open globe injury remains a major cause of permanent visual impairment and blindness in the world [1]. Despite advances in ophthalmic surgery and equipment, loss of vision may be unfavorable in a significant number of cases [1]. Open globe injury, defined as a full-thickness injury of the eyewall, presents severe damage to the eye and often results in poor outcomes [2]. The accumulation of knowledge about the pathophysiology of eye injuries and their prognostic factors as well as advances in diagnostic and therapeutic methods have greatly improved the success rates for managing open globe injuries [3]. A better understanding of these prognostic factors may help to provide our patients with better and more realistic expectations of their final visual acuity [3]. Many studies have been conducted to evaluate the factors associated with poor prognosis in patients with open globe injuries [1,[3][4][5][6][7][8][9]. Several prognostic factors such as initial visual acuity, the involvement of ocular tissue, and both proper diagnosis and appropriate treatment may help to achieve a useful vision [1,[3][4][5][6][7][8][9].
A total number of 160 patients (161 eyes), diagnosed with OGI, were admitted to the hospital during the study period. The age of patients at the time of injury ranged from three to 82 years with a mean of 41.9 ± 1.5 (M ± SE) years (42.2 ± 1.5 years in males and 39.1 ± 6.1 years in females). Male patients constituted 89.4% of the cases (n = 144), making the male-to-female ratio 8.4:1. Bilateral OGI was noted in one patient. Among the unilateral injuries, no significant difference was observed between the affected eyes (right eye 50.9% vs. left eye 49.1%, p > 0.05). The study population was divided into three age groups: <18 years (12.4%, n = 20), 18-59 years (70.2%, n = 113), and ≥60 years (17.4%, n = 28). A significant predominance of males in the age group of 18 to 59 years and females in the age group of <18 years was found. Figure 1 presents the details for age and gender distribution. groups: <18 years (12.4%, n = 20), 18-59 years (70.2%, n = 113), and ≥60 years (17.4%, n = 28). A significant predominance of males in the age group of 18 to 59 years and females in the age group of <18 years was found. Figure 1 presents the details for age and gender distribution. Urban residents accounted for 47.2% of the total subjects, and residents from rural areas accounted for 52.8% (p > 0.05). Ten subjects (6.2%) admitted alcohol use before the injury.
Presenting VA was documented as NLP in all cases of perforating injuries. Three cases of these underwent primary wound closure and one eye was enucleated during the initial surgery. All patients with perforating injuries were lost for further follow-up.
The distribution of IOFB by location and type was as follows: 39.7%-magnetic IOFB in the anterior segment, 7.9%-nonmagnetic IOFB in the anterior segment, 50.8%-magnetic IOFB in the posterior segment, and 1.6%-nonmagnetic IOFB in the posterior segment.  Urban residents accounted for 47.2% of the total subjects, and residents from rural areas accounted for 52.8% (p > 0.05). Ten subjects (6.2%) admitted alcohol use before the injury.
Presenting VA was documented as NLP in all cases of perforating injuries. Three cases of these underwent primary wound closure and one eye was enucleated during the initial surgery. All patients with perforating injuries were lost for further follow-up.
The distribution of IOFB by location and type was as follows: 39.7%-magnetic IOFB in the anterior segment, 7.9%-nonmagnetic IOFB in the anterior segment, 50.8%-magnetic IOFB in the posterior segment, and 1.6%-nonmagnetic IOFB in the posterior segment.
There was a significant association between the type of OGI and age. The frequency of globe rupture was found to be significantly higher in the age group of 60 years and older (Table 1).
Penetrating wound was found to be the most common type of OGI for both genders, but globe rupture significantly predominated among female patients ( Figure 2). There was a significant association between the type of OGI and age. The frequency of globe rupture was found to be significantly higher in the age group of 60 years and older (Table 1). Penetrating wound was found to be the most common type of OGI for both genders, but globe rupture significantly predominated among female patients ( Figure 2). Through the cause of OGI, hammering on metal and injuries caused by lawn equipment were significantly associated with IOFB (38.1% and 7.9%, respectively), sharp objects-with penetrating injury and IOFB (93.2% and 46.0%, respectively), blunt objects, and fall-with globe rupture (90.5% and 9.5%, respectively) (χ 2 = 216.131, df = 16, p < 0.001).
In terms of zone of injury, 75 (47.8%) eyes had zone 1 injury, 40 (25.5%) eyes had zone 2 injuries, and 42 (26.8%) eyes had zone 3 injuries. Zone 1 was significantly more frequently diagnosed with IOFB and penetrating injury versus rupture, and zone 3 with penetrating injury and rupture versus IOFB ( Table 2).
All patients with OGI underwent surgery. Ninety-two eyes (57.1%) were operated on within the first 24 h and forty-nine eyes (30.4%) during 25-48 h after presentation. A primary procedure such as wound/rupture repair accounted for 54.7% of cases (88 eyes). Initial wound repair in combination with pars plana vitrectomy (PPV) was performed in 55 eyes (34.2%), initial PPV without wound repair in 16 eyes (9.9%), and lensectomy/cataract extraction in 31 eyes (19.3%). The IOFBs were removed during primary open globe repair in all 63 eyes with IOFB. Twenty-four eyes (14.9%) underwent secondary eye surgery. Types of secondary procedures included PPV (14 eyes), lensectomy/cataract extraction (10 eyes), and implantation of the intraocular lens (18 eyes). PPV, as a tertiary procedure, was performed in only one eye. Two eyes (1.2%) underwent primary enucleation. One eye was enucleated during secondary surgery.  (Table 5).  Regarding the type of OGI, the final visual acuity of grade 5 (NLP) was significantly related to globe rupture versus IOFB injury. Grade 1 (≥0.5) was related to IOFB injury. In terms of zone of injury, zone 1 was significantly associated with the final BCVA of grade 2 (0.2-0.4), and zone 3 with the final BCVA of NLP. The distribution of final BCVA by zone and type of OGI is presented in Table 6.
Univariate logistic regression analysis showed that an initial VA of 0.02 and worse was the strongest predictive factor of the poor visual outcome, with an odds ratio (OR) of 7.143 (95% confidence interval (CI), 2.519-20.257). Initial diagnoses such as iris dialysis (OR 2.783; 95% CI, 1.079-7.176), hypotony (OR 2.546; 95% CI, 1.006-6.443), vitreous hemorrhage (OR 3.125; 95% CI, 1.227-7.959), and vitreous prolapse (OR 3.069; 95% CI, 1.022-9.216) were also found to be significant predictive factors for the poor visual outcome of ≤0.02. Other factors such as injury zone, OGI type, retinal detachment, endophthalmitis, time of initial surgery, time from injury to presentation, and number of surgeries were not statistically significant in the univariate analysis.
All factors found significant in the univariate logistic analysis were included in the multivariate logistic analysis to further evaluate their associations with the final VA and phthisis bulbi, but no statistically significant association was found because of the multicollinearity of those factors.

Discussion
This study aimed to evaluate the epidemiological, clinical characteristics, visual, and anatomical outcomes after severe open globe injuries in patients, presented to the Department of Ophthalmology of the Hospital of Lithuanian University of Health Sciences, the principal tertiary center for ocular injuries in the country, and to identify the possible prognostic factors, influencing the final visual and anatomical outcome.
The mean age of our population was 41.9 ± 1.5. This value is similar to those that have been reported in some other studies [5,6,12]. Many authors have reported the younger mean age of the patients [1,[13][14][15][16], and this fact can be explained by different study designs or the specificity of the country.
We identified the predominance of the male gender for OGI (89.4%). This tendency has also been found in other epidemiological studies, with a male proportion varying between 66.0% and 96.7% [1][2][3]6,7,12,13,15,[17][18][19][20][21][22]. This variation could be explained that men are at more risk of being exposed to dangerous situations in the workplace or during outdoor activities as well as during gender-based behavior [3]. Our study demonstrated that age was strongly associated with the incidence of OGI. The rate of OGI was found to be significantly higher in male patients in the age range between 18 and 59 years. These findings are consistent with those of other studies in which the peak age of male OGI ranged from 20 to 49 years (52.6%) [8], 21-50 years (55.03%) [6], and 41-60 years (40.9%) [16]. We also found that the risk of OGI was significantly higher for females younger than 18 years old. This finding could be explained by the fact that at a younger age, males and females are engaged in similar daily activities.
In our study, the home was the leading place of eye injury (59.6%), followed by an outdoor environment/street and workplace, which only accounted for 11.8% of all OGI. In line with our results, the previous literature has also reported that the home was the most frequently associated place of eye trauma [1,7,16,23]. Other studies have confirmed that the majority of open globe injuries are occupational, ranging from 22.0% to 50.0% of cases [1][2][3][4]6,13,19,20].
Injury by sharp objects is among the most common mechanisms of injury. Glass accounted for the majority of such injuries [4]. In our series, sharp objects such as metal fragments, sharp instruments, and broken glass, accounted for 61.5% of cases. These results are consistent with the data published by Rahman et al. [4]. In our series, hammering on metal was the second leading cause of injury and accounted for 14.9% of all causes of OGI, and this rate was higher than that reported by Makhrash et al. (8.3%) [7] and Rahman et al. (4%) [4]. Blunt objects accounted for 11.8% of all causes of OGI and were responsible for ninety percent of globe ruptures in our study. Other studies reported a blunt mechanism to be responsible for 28% [4] and 20% [7] of all OGI cases. Fall was found to be one of the main causes of the blunt mechanism of injury, especially in older female patients [3], significantly predominated in older age groups [24]. Our results could not prove this fact, but we found a significant risk for globe rupture in older women.
An analysis of the zones of injury revealed that almost half of our patients had zone 1 injury (47.8%) and more than one-fourth of them had zone 3 injuries (26.8%). Similar data were presented by Wang et al. [16]. Numerous studies found similar rates for zone 1, which was the most common location of the injury, but in contrast, zone 2 was indicated as the second leading location of injury in these studies [1,5,7,8,13,14,19]. Furthermore, in our series, zone 3 injury was significantly related to initial diagnoses such as eyelid laceration, eyelid contusion, uveal prolapse, iris dialysis, iris laceration, lens dislocation, hyphema, hypotony, vitreous prolapse, choroidal hemorrhage, and initial VA of grade 5 at initial examination. The current investigation demonstrated that the zone of injury was also found to be associated with visual outcomes. Wounds involving zone 3 had significantly poorer presenting and final VA versus those involving zones 1 or 2. These results are supported by previous studies that have reported a significant association between the posterior extension of injury and a worse final VA [2,3,[7][8][9]13]. Regarding the mechanism of injury, our study confirmed that globe rupture was associated with a lower rate of visual survival and functional success than a laceration. These results are consistent with the data published by Fujikawa et al. [3] and Feng et al. [21].
In our study, 14.0% of all OGIs presented with an initial VA of grade 1 (≥0.5). Initial VA of ≤0.02 (grades 4 and 5) accounted for 55.3% of injuries at presentation. These findings were consistent with the studies performed by Meng et al. [6], Fu et al. [12], and Bauza et al. [19]. In contrast to these results, Pimolrat et al. reported that presenting VA less than 6/60 was determined in 92% of cases [26].
In our series, initial grade 1 was significantly related to IOFB injury. Grade 2 was significantly associated with penetrating injury, IOFB, and zone 1. Grade 4 (0.02-LP) showed a significant association with globe rupture. Grade 5 (NLP) was significantly related to rupture and penetrating injury versus IOFB and to zone 3 of injury.
Analysis of initial and final visual acuity was evaluated and compared in seventyseven cases of follow-up patients. No statistically significant difference was found between the grades of injury at the initial and final examination. In our study, 19.5% of injured eyes regained a good vision of ≥0.5. Comparisons with other studies are complicated due to the differences in study design and the great variability in the nature and severity of eye injuries themselves. Other studies reported achieving ≥0.4 in 29% [18], ≥0.5 in 22.29% [6], 26.8% [13], and 38.5% [19] [19].
Regarding the type of OGI, the final visual acuity of grade 5 (NLP) was significantly related to globe rupture. A good visual outcome of grade 1 (≥0.5) was significantly related to IOFB injury. In contrast, the analysis presented by Atic et al. showed a strong association between IOFB and poor visual outcomes [1]. Our results also demonstrated that the zone of injury was associated with visual outcomes. Wounds involving zone 3 had significantly bad visual outcomes of NLP versus those involving zones 1 or 2. These findings were supported by numerous investigations [1,3,8,13,19].
In univariate analysis, an initial visual acuity of ≤0.02 (grades 4 and 5) was significantly associated with the final poor visual outcome of ≤0.02. The univariate analysis also demonstrated that the presence of iris dialysis, hypotony, vitreous hemorrhage, and vitreous prolapse was significantly associated with final BCVA ≤0.02 (grades 4 and 5) in this study. However, multivariate analysis did not reveal differences in these predictive factors, conceivably because of their multicollinearity.
In agreement with our results, initial visual acuity ≤0.02 [3,4,[6][7][8][9]19] and vitreous hemorrhage [3,7] were reported to be an important predictive factor of the poor visual outcome by other researchers. The findings of numerous studies indicated that globe rupture [3], presence of retinal detachment [3,7,8], dislocation of the crystalline lens [3], presence of RAVD [3,8], larger wound (>10 mm) [8], zone 3 injury [3,7,19], and aphakia [7] were the most significant predictors of final visual outcome, determined in univariate or multivariate analysis. Our study did not find a statistically significant association between these factors and visual prognosis. This could be explained by different study designs, the great variability in the nature and severity of the eye injuries themselves, or by the particularities of countries where the investigation occurred.
According to our results, diagnoses defined at the last follow-up visit such as corneal scars, glaucoma, traumatic cataract, vitreous opacities, PVR, and retinal detachment were mostly related to globe rupture and penetrating injury. A traumatic cataract is the most common vision limiting complication, and it can occur any time from day 1 to several years after OGI [1]. We found that there was no significant association between traumatic cataracts and poor visual outcomes. These results are supported by data published by Atic et al. [1], but, in contrast, Fujikawa et al. found a significant correlation between lenticular involvement and bad visual outcome [3]. Atic et al. reported that retinal detachment was found to be a predictor of poor outcome [3], but our results could not prove this finding.
In terms of the type of OGI, IOFB injury showed the best anatomic success in comparison with penetration and rupture in our study.
In the current investigation, seven eyes ended up with phthisis bulbi at the end of the follow-up period, two eyes underwent primary enucleation, and one eye was enucleated at a subsequent surgical procedure. Feng et al. reported 15 cases of NLP, enucleation, or phthisis [21], Souylu et al. found that in 17.7% of eyes, phthisis bulbi occurred during the follow-up period [18].
Phthisis bulbi and enucleation were defined to present an unfavorable anatomic and visual outcome, significantly associated with predictive factors such as rupture, zone 3, large scleral wound, ciliary body damage, severe intraocular hemorrhage, closed funnel retinal detachment and retinal prolapse, and choroidal hemorrhage [21]. Univariate logistic regression analysis showed that eyelid laceration, iris prolapse, iris dialysis, hyphema, vitreous prolapse, vitreous hemorrhage, and choroidal rupture at initial presentation were statistically significant predictive factors for final phthisis bulb, in our study. Other studies also concluded that vitreous hemorrhage was also a predictor of poor outcome.
Rahman et al. reported a higher rate of secondary enucleations (12%) after OGI. They found RAVD, the presence of lid laceration, a blunt mechanism of injury, and initial VA worse than 6/60 on presentation to be significant risk factors associated with eventual enucleation [4]. In our series, two eyes underwent primary and one eye secondary enucleation, and no significant associations between clinical factors and enucleation were found.
Limitations of our study need to be noted. First, it was conducted as a retrospective study design and contained a considerable number of unrecorded data. Second, the study was related to variable follow-up times, and in some of the cases, the follow-up duration was relatively short.

Conclusions
In conclusion, a patient with OGI should be carefully examined both at the time of admission and during the follow-up period. This study confirms that some clinical characteristics such as initial visual acuity, iris dialysis, hypotony, vitreous hemorrhage, and vitreous prolapse may have the potential to correctly predict final visual outcomes. In the future, a prospective study of OGI, aimed at a more detailed evaluation of prognostic factors and prediction of functional outcomes, could provide solid evidence for building better management strategies in cases of OGI.