A COMPARATIVE ANALYSIS OF TRAUMATIC RETINAL DETACHMENT AFTER OPEN AND CLOSED GLOBE INJURIES IN CHILDREN

Comparing pediatric traumatic retinal detachment from open and closed globe injuries, this study finds better outcomes after closed globe injuries. Anatomical success rates have improved, yet functional results remained limited. Initial visual acuity strongly predicts final outcomes.

Abstract

Purpose:

Pediatric traumatic retinal detachment (RD) resulting from open globe injuries (OGIs) or closed globe injuries (CGIs) presents unique challenges due to complexity often resulting in lifelong sequelae. This study compares pediatric traumatic RD outcomes and prognostic factors following OGI and CGI.

Methods:

A retrospective analysis reviewed 47 cases of pediatric traumatic RD in children (age <18 years), who underwent RD surgery between 2002 and 2021. Among them, 25 cases were caused by CGI and 22 cases by OGI. Demographics, RD characteristics, surgical procedures, and anatomical and functional results were assessed. Predictive factors for visual outcomes were investigated.

Results:

In the CGI group, mean (±SD) age was 11 years ± 4 years, and 10 years ± 5 years in the OGI group. Closed globe injury traumatic RD had significantly better preoperative (CGI: logarithm of the minimum angle of resolution 1.39 ± 0.19 (mean ± standard error); OGI: logarithm of the minimum angle of resolution 2.12 ± 0.20) and follow-up (CGI: logarithm of the minimum angle of resolution 0.94 ± 0.19; OGI: logarithm of the minimum angle of resolution 1.85 ± 0.20) best-corrected visual acuity (BCVA) (P < 0.05). Initial BCVA improvement was observed in CGI only. In multivariable analysis, prognostic factors for favorable BCVA outcomes included higher preoperative BCVA, older age, and absence of proliferative vitreoretinopathy (P < 0.05).

Conclusion:

Visual prognosis for pediatric traumatic RD remains limited, favoring CGI cases compared with OGI. Baseline BCVA emerged as a major determinant of final visual acuity. Tailored management approaches can optimize treatment results.

Traumatic retinal detachment (RD) is a severe complication of ocular trauma, particularly in children, posing a substantial risk to visual function with potentially lifelong sequelae. Despite being rare, traumatic RD resulting from open (OGI) or closed globe injuries (CGIs) is a challenging condition for retinal surgeons to address, in many cases carrying an unfavorable visual prognosis.^1–12

Open globe injury involves a full-thickness injury of the eye, often caused by sharp objects or high-velocity projectiles, leading to retinal tears or breaks.^10,13–17 By contrast, CGI, often caused by a blunt impact such as sports-related injuries, falls, or physical assault, can also induce traumatic RD through indirect mechanical forces and vitreoretinal traction.^10,13,16 In some cases, blunt impact can also result in OGI through the rupture of the eyeball. Limited research exists on traumatic RD characteristics and outcomes in children, leaving a gap in optimal management and strategies. To our knowledge, the few published studies focusing on traumatic RD in children cover approximately 20 years (1980–2003) of medical records.^8,9,11 There is only one more recent study with data from the 2010s.¹⁰ Considering the remarkable advancements in vitreoretinal surgery over the past decades, treatment options for pediatric traumatic RD have improved ever since. However, it remains unclear whether these advancements had beneficial effects on visual prognosis.

Through a comparative analysis, the objective of this study was to expand the available data on clinical characteristics, surgical procedures, and functional and anatomical outcomes, as well as to find prognostic indicators for traumatic RD following CGI and OGI in children.

Methods

Medical records of all patients under 18 years diagnosed with traumatic RD following CGI or OGI between January 2002 and December 2021 at the University Eye Hospital Tübingen, Germany were retrospectively analyzed. Demographic data, including age, sex, affected eye, and clinical characteristics such as injury type, time between trauma and traumatic RD diagnosis, RD characteristics (macular involvement, number of quadrants, retinal break/tear type, present proliferative vitreoretinopathy [PVR] grade B or higher), number of RD surgeries (excluding non-RD surgeries), number of non-RD surgeries, type of (primary) RD surgery (scleral buckle [SB], pars plana vitrectomy [PPV], combined surgery), and postoperative complications were recorded. We classified the injury type according to the Birmingham Eye Trauma Terminology.¹⁷ Birmingham Eye Trauma Terminology categorizes eye injuries into CGI and OGI. Closed globe injuries encompass contusions and lamellar lacerations, while OGIs can be further classified into globe ruptures and globe lacerations. Globe lacerations summarize penetrating injuries (single laceration, entry wound, no exit wound), perforating injuries, (full thickness entry and exit wound) and injuries involving intraocular foreign bodies (retained intraocular foreign object).¹⁷ Contusions do not involve wounds to the eyeball, whereas lamellar lacerations are characterized by partial-thickness injuries. Globe ruptures are caused by blunt objects resulting in a full-thickness injury. Full-thickness wounds caused by sharp objects are classified as lacerations. Injuries with a single full-thickness defect are classified as penetrating injuries, while perforating injuries have an entrance and exit wound caused by a sharp object. Ocular trauma score (OTS) was used to objectify the injury severity.¹⁸ Regarding the timing of surgery, at our institution, primary globe closure in OGI is performed immediately as an emergency procedure. In cases of traumatic RD in children, we perform surgery on the date of presentation or the following day.

Best-corrected visual acuity (BCVA) served for functional assessment at the time of traumatic RD diagnosis, 1 year (±3 months) and 2 years (±3 months) after RD surgery, and at the final available follow-up. Best-corrected visual acuity measured using (Early Treatment Diabetic Retinopathy Study) ETDRS charts was converted to logarithm of the minimum angle of resolution (logMAR) values. Anatomic outcomes were determined by single surgery anatomic success (SSAS), final anatomic success (FAS), and limited anatomic success (LAS) rates. Single surgery anatomic success was defined as retinal reattachment in the absence of any tamponade at least 3 months after primary RD surgery regardless of the surgical procedure. Final anatomic success was determined as anatomically reattached retina without any tamponade at the final available follow-up examination. Limited anatomic success was defined as reattached retina independent from persistent tamponades (with or without tamponade) at the final available follow-up. Potential prognostic factors of the final BCVA were evaluated.

The study was approved by the ethics committee of the University of Tübingen (project number: 488/2023BO2) and conducted in accordance with the Declaration of Helsinki.

Descriptive statistical analysis summarized demographic and clinical characteristics. A mixed regression model assessed BCVA changes between injury types (CGI, OGI) and time periods (preop, 1 year, 2 years, final follow-up), using random effects for eyes and fixed effects for injury type and time periods. Post hoc Tukey tests were used for pairwise comparisons between time periods. Comparison between CGI and OGI groups employed independent t-tests or Mann–Whitney test for numeric variables, depending on the distributional characteristics of the variables (normality, presence of significant outliers, homogeneity of variances). Normality was assessed using the Shapiro–Wilk test, and the presence of outliers was evaluated using the box-plot method. Homogeneity of variances was checked using the Levene test. Fisher's exact test was used for testing differences in factor variables. The aim of the additional analysis was to identify variables affecting the final BCVA. A quadratic model utilized preoperative BCVA as input and postoperative BCVA as output variable, assessing the significance of added risk factors. In a final multivariable model, we included preoperative characteristics that significantly showed influence on the final BCVA in the baseline model (baseline BCVA, final BCVA; Figure 1). The results included R² values and analysis of variance P-values. Statistical significance was defined as P < 0.05. All calculations were performed with “R,” version 4.1.2.¹⁹

Fig. 1.

Correlation of baseline and final visual acuity. A good baseline BCVA was correlated with a good final BCVA and vice versa. A quadratic fit was most appropriate for the correlation of baseline and final BCVA. The prediction is more uncertain for cases with bad preoperative vision. This baseline model explains 55.2% of the variance (R² = 0.55, P < 0.001).

Results

A total of 47 patients (34 boys, 13 girls) with traumatic RD were included in this study, of which 25 patients experienced CGI and 22 patients OGI. According to the Birmingham Eye Trauma Terminology classification,¹⁷ all the 25 CGI cases had suffered contusion, and none showed lamellar laceration. Of the 22 OGI cases, 5 patients had globe rupture caused by blunt trauma, whereas 17 patients had lacerations induced by sharp objects or projectiles (13 penetrating, 1 perforation, 3 intraocular foreign bodies). Types and mechanisms of injuries are summarized in Table 1. The mean age of the patients was 11 years ± 5 years (mean ± SD), with the majority (n = 26; 55.3%) aged 10 years or older. The mean (±SD) ocular trauma score in CGI was 68.57 (±14.61) and 45.85 (±10.56) in OGI (P < 0.001). The interval between trauma and diagnosis of traumatic RD (median [Q1, Q3]) was 62.00 (11.25, 302.75) days with no significant difference between CGI and OGI groups (CGI: 18.0 [9.0, 250.0] days; OGI: 100.0 [41.0, 341.0] days—P = 0.175). The time between primary globe closure in OGI and RD repair was 100 (28.5, 379.0) days. Retinal detachment surgery was always performed on the day of RD diagnosis or on the following day. Clinical characteristics are summarized in Table 2.

Table 1.

Injury Type (According to Birmingham Eye Trauma Terminology) and Mechanisms of Injury

CGI	25
Contusion	25
Lamellar laceration	0
OGI	22
Globe rupture	5
Globe laceration	17
Penetrating	13
Perforating	1
Intraocular foreign body	3
Mechanism of injury
External objects	24
Tree branch/wood	6
Nail/skewer/screwdriver	4
Knives	3
Firework	2
Glass	2
Others	5
Sports-related injuries	8
Falls	2
Physical assault	2
Road accident (car, bike)	2
Autoaggression, epilepsy	2
Undefined in medical records	4

The majority of injuries are caused by external objects by accident and sports-related injuries.

Table 2.

Clinical Characteristics of Retinal Detachments in Both Injury Groups, CGI and OGI

	Total (n = 47)	CGI (n = 25)	OGI (n = 22)	P
OTS, mean (SD)	58.00 (17.14)	68.57 (14.61)	45.85 (10.56)	P < 0.001
Time from trauma to RD (days), mean (SD); median	385.37 (832.99; 49)	360.00 (802.30; 11)	415.57 (887.14; 97)	P = 0.096*
Time primary globe closure to RD (days)	—	—	414.45 (871.41)
# of RD quadrants, mean (SD)	2.96 (1.18)	2.72 (1.28)	3.23 (1.02)	P = 0.154*
Total retinal detachments, n (%)	22 (46.8)	9 (36.0)	13 (59.1)	P = 0.217†
Macular involvement, n (%)	34 (72.3)	17 (68.0)	17 (77.3)	P = 0.764†
Giant tear				P = 0.831†
No, n (%)	36 (83.7)	20 (87.0)	16 (80.0)
Yes, n (%)	7 (16.3)	3 (13.0)	4 (20.0)
Retinal dialysis				P = 0.877†
No, n (%)	33 (76.7)	17 (73.9)	16 (80.0)
Yes, n (%)	10 (23.3)	6 (26.1)	4 (20.0)
Retinal redetachment, n (%)	18 (38.3)	10 (40.0)	8 (36.4)	P = 1.000†
PVR (≥ grade B) at 1st RD diagnosis, n (%)	21 (44.7)	10 (40.0)	11 (50.0)	P = 0.795†
PVR during follow-up, n (%)	32 (68.1)	15 (60.0)	17 (77.3)	P = 0.372†
IOFB, n (%)	3 (6.4)	0 (0.0)	3 (13.6)	P = 0.026 †
Vitreous hemorrhage after trauma, n (%)	7 (14.9)	3 (12.0)	4 (18.2)	P = 0.833†
Posterior vitreous detachment at the time of RD surgery, n (%)	21 (44.7)	5 (20.0)	16 (72.7)	P < 0.001 †
Hypotony (≤6 mmHg) at final follow-up, n (%)	6 (12.8%)	2 (8.0)	4 (18.2%)	P = 0.574†

^*Mann–Whitney test.

^†Fisher's exact test.

IOFB, intraocular foreign body; OTS, ocular trauma score.

Bold entries signify statistical significance (P < 0.05)

Surgical Characteristics, Anatomical and Functional Outcomes, Complications

The mean number of RD surgeries (excluding primary globe closure, phacoemulsification, anterior segment surgery, silicone oil [SO] removal, etc.) was 1.68 (±0.95) in the CGI group and 2.23 (±1.34) in the OGI group, without a significant difference (P = 0.072), while the number of non-RD surgeries differed significantly (CGI: 1.40 [±2.10], OGI: 1.95 [±1.21]—P = 0.015). Surgical methods notably varied, with SB (encircling, segmental, radial) being the primary approach in 15 cases (31.9%) overall; this approach was utilized as the primary surgical procedure more frequently in the CGI group (CGI: 12 patients [48.0%] vs. OGI: 3 patients [13.6%]—P = 0.019), whereas primary PPVs were more common in the OGI group (CGI: 10 patients [40.0%] vs. OGI: 16 patients [72.7%]—P = 0.040). Combined primary surgery (SB + PPV) was performed in three cases of each group (P = 1.000). Among the patients primarily treated with SB, three cases in the CGI group required additional PPV with SO tamponade due to persistent or recurrent RD. Three cases in the OGI group needed secondary PPV (one gas, two SO). Throughout the follow-up period, SO was successfully removed in all three cases of the CGI group. In the OGI group, SO could be removed in one of the two cases, while the other case experienced PVR-related redetachment following SO removal and consequently required permanent SO tamponade. Silicone oil was used as primary tamponade in 29 of 47 eyes (CGI: 12 eyes, OGI: 17 eyes). During the complete follow-up, SO was used in 35 eyes (CGI: 15 eyes, OGI: 20 eyes). Open globe injury group had a significantly higher number of SO tamponades at the final follow-up (CGI: 20.0%, OGI: 63.6%—P = 0.002). Gas tamponade was used once in CGI, with no retinal redetachment during 13 years of follow-up, and twice in OGI, with one patient experiencing retinal redetachment and combined surgery (SB + PPV with SO).

Anatomical success rates were evaluated using three rates: SSAS, FAS, and LAS. The SSAS rate (reattached retina in the absence of tamponade at least 3 months after primary RD surgery) was nonsignificantly higher in the CGI group (CGI 11 eyes [44.0%] vs. OGI 4 eyes [18.2%]—P = 0.106). Final anatomic success (reattached retina in the absence of tamponade at final follow-up examination) at the final follow-up was achieved in 27 eyes (57.4%), with a significantly higher success rate in CGI compared with OGI (CGI: 19 eyes [76.0%] vs. 8 eyes [36.4%]—P = 0.007). Limited anatomic success at the final follow-up (reattached retina independent from persistent tamponades [with or without SO]) was achieved in 43 of 47 eyes (91.5%) in total, with both groups showing LAS rates of over 90% (CGI: 23 eyes [92.0%] vs, OGI: 20 eyes [90.9%]—P = 1.000).

Preoperative BCVA was logMAR 1.39 ± 0.19 (mean ± standard error) (≈20/500 Snellen) in CGI and 2.12 ± 0.20 (counting fingers/hand movements) (≈CF/HM) in the OGI group (Figure 2). Both groups showed visual improvement, which however was only statistically significant in the CGI group (preoperative BCVA vs. final BCVA—P < 0.001), while for the OGI group BVCA improvement was not significant (P = 0.138). The functional outcome was significantly better for traumatic RD after CGI (final BCVA, CGI: logMAR 0.94 ± 0.19 [≈20/160 Snellen]; OGI: logMAR 1.85 ± 0.20 [≈CF]—P < 0.05). No significant BVCA improvement could be determined in the OGI group. Best-corrected visual acuity differed significantly at baseline and all time points between CGI and OGI. Best-corrected visual acuity significantly (P < 0.001) improved between preoperative assessment and 1-year follow-up in the CGI group and showed no further significant improvement until the final follow-up examination (Figure 2). There was no statistically significant interaction between time and groups observed meaning that the changes between the groups are not different. In the CGI group, 8 patients had logMAR ≤1.0 (≥20/200 Snellen) preoperatively, and 15 showed maintained or improved BCVA at the final follow-up, while the OGI group had 3 such patients initially and only 2 at the final follow-up.

Fig. 2.

Visual acuity. Trajectories of BCVA for the individual eyes with traumatic RD after CGI and OGI and mean (±standard error) BCVA trajectory (red line) estimated with a mixed regression model. Open globe injury shows worse BCVA at all time points (P < 0.05). Only the differences from preop to 2 years and preop to final follow-up in the CGI group were statistically significant. No statistically significant changes were observed in the OGI group. There was no statistically significant interaction between time and groups observed, which means that the changes between the groups are not different.

Differences in lens status between groups at the final follow-up were significant, with a higher proportion of aphakic patients in the OGI group (77.3%) compared with the CGI group (16.0%) (P < 0.001), while 12 patients in the CGI group retained phakic lens status (25.5%) (P < 0.001); the number of pseudophakic patients showed no significant difference (CGI: nine eyes [36.0%], OGI: five eyes [22.7%]—P = 0.544).

Regarding other complications, seven patients had increased intraocular pressure or secondary glaucoma with no significant differences between groups (P = 0.299). In the CGI group, one patient developed posttraumatic glaucoma and required cyclokryocoagulation whereas another patient had elevated intraocular pressure levels due to SO emulsification, which was compensated by topical antiglaucoma medication. In the OGI group, five patients developed secondary glaucoma, of which one required cyclophotocoagulation and another had trabeculotomy. Persistent hypotony (intraocular pressure ≤ 6 mmHg) at the final available follow-up examination was found in two (8.0%) patients after CGI and four (18.2%) patients after OGI (P = 0.574), while phthisis was present at the final follow-up in one (4.0%) and five (22.7%) patients in CGI and OGI groups, respectively (P = 0.075).

Predictors of Final Visual Acuity

A strong association between the baseline BCVA and final BCVA was found, meaning that a good baseline BCVA was most indicative of a good visual outcome (Figure 1). A quadratic regression model (R² = 0.55, P < 0.001) provided the best fit for the data explaining 55.2% of the differences we observed in the final BCVA, leaving 44.8% potentially influenced by other factors not considered by baseline BCVA.

To explore other potential prognostic indicators, we used a second-degree polynomial model with BCVA as the input variable (Figure 3). We then separately analyzed the impact of additional factors (Figure 3) on the baseline model (baseline BCVA vs. final BCVA). By incorporating FAS, SSAS, age, or present PVR at diagnosis of traumatic RD into the baseline model, we discovered a significant correlation between each parameter and the final BCVA given the preoperative BCVA (SSAS: R² = 0.68, P < 0.001; FAS: R² = 0.76, P < 0.001; PVR at diagnosis of traumatic RD: R² = 0.63, P = 0.007). This confirms SSAS and FAS as indicators of success, associated with better visual prognosis. The presence of PVR at traumatic RD diagnosis leads to worse outcomes. In addition, older age at RD diagnosis correlates with better final BCVA when controlled for initial BCVA (R² = 0.61, P = 0.024). None of the other characteristics studied had a significant effect on the baseline model. For example, macular involvement (R² = 0.55, P = 0.796) and injury type (CGI, OGI; R² = 0.58, P = 0.144) did not contribute any additional explanatory power to the variation in postoperative BCVA beyond the strong correlation already captured by the baseline BCVA.

Fig. 3.

FAS, SSAS, PVR (at RD diagnosis), and age. Graphs including risk factors that showed significant influence on the baseline model (Fig. 1). A. FAS, (B) SSAS, (C/D) older age and no PVR at traumatic RD diagnosis are associated with better final BCVA. C. Includes cases without PVR, while (D) shows cases with PVR at the diagnosis of traumatic RD.

As we found that age and PVR at the time of diagnosis of traumatic RD were significantly related to the final BCVA alongside baseline BCVA, in a final multivariable model (Figure 3, C and D, Table 3), we included age, PVR, and baseline BCVA in a single prognostic model as independent variables. This multivariable model revealed that PVR worsened the prognosis of the final BCVA (P < 0.05) and age improved the final BVCA (P < 0.1). The coefficients given in Table 3 can be used to estimate postoperative BCVA for any given value of an independent variable using the following formula: $\text{postopBCVA}=1.665+\text{preopBCVA}*3.852+\left[\text{BCVA}\hspace{0.17em}{\left(\text{LogMar}\right)}^2\right]*1.643-\text{Age}*0.045+0.536*\text{PVR}\hspace{0.17em}\left(\ge \text{grade}\hspace{0.17em}\mathrm{B};\text{present}\hspace{0.17em}\text{PVR}=1,\text{absent}\hspace{0.17em}\text{PVR}=0\right).$

Table 3.

Final Model With Baseline BCVA, Age, and PVR as Independent Variables

	Dependent Variable	P
	Final BCVA (logMAR)
BCVA (logMAR) preop	3.852 (0.689)	<0.01
BCVA (logMAR) preop2	1.643 (0.641)	0.015
Age (years) at RD	−0.045 (0.024)	0.070
PVR (≥ grade B) at 1st RD diagnosis	0.536 (0.218)	0.018
Constant	1.665 (0.325)
Observations	43
Adjusted R2	0.626

The results are given as coefficient (standard error).

Discussion

Based on 20 years of available data at a tertiary care hospital, the aim of this study was to compare clinical, surgical, and functional aspects of traumatic RD following CGI and OGI in children. Despite poor visual outcomes for both injury types, traumatic RD after CGI showed more favorable results, aligning with prior studies.^8–11 As most of these studies analyzed data from the 1980s to the early 2000s, limited recent data on pediatric traumatic RD left uncertainty about surgical advancement's impact on functional and anatomical outcomes. Our study population had a mean age (11 ± 5 years) similar to previous studies.^8–11,13,20 Approximately half (55.3%) of our patients were 10 years or older, which holds significance due to the heightened risk of severe amblyopia in young patients potentially impeding functional rehabilitation efforts.

While BCVA significantly differed between CGI and OGI at traumatic RD diagnosis, other clinical features showed no significant differences before surgery. Late traumatic RD diagnosis was observed in several cases, possibly due to children's difficulty in expressing symptoms adequately and appraising symptoms as not self-limited, which contrasts with what they usually have been experiencing as healthy individuals. In addition, it is important to note that the vitreous in young individuals is well formed and firmly attached to the retina.²¹ Following ocular trauma, vitreous liquefaction may occur, while the retina is still attached. Initially, when the vitreous is still well attached, the leakage of fluid through traumatic retinal breaks into the subretinal space is likely to be minimal.²¹ However, as vitreous liquefaction progresses, leakage into the subretinal space may increase, inducing a slowly progressive RD, which potentially explains the delayed onset of RD observed in many patients. This highlights the need for close monitoring and frequent follow-up examinations after ocular trauma.

Delayed diagnosis of RD results in higher PVR rates.¹ Accordingly, PVR negatively impacted the final BCVA in our study. The high prevalence of PVR (44.7%) at traumatic RD diagnosis, consistent with previous studies, implies a frequent tractive component complicating surgical treatment.^1,9,11 Proliferative vitreoretinopathy rates vary in previous studies, possibly due to the heterogeneity in patients and trauma. By contrast, potentially suggesting more severely injured eyes and/or delayed presentation, our study's findings revealed high PVR rates during the complete follow-up after CGI (60.0%) and OGI (77.3%), other studies on pediatric traumatic RD reported lower PVR rates ranging from 0% to 10.0% following CGI^8,9 and 5.4% to 64% for traumatic RD after OGI.^1,8,9,11 These findings emphasize that PVR poses a significant challenge not only in the initial management of pediatric traumatic RD but also often necessitates additional surgical interventions due to tractive redetachments.

The total RD rate in children is high and associated with worse visual prognosis and a lower rate of successful anatomical repair.³ In our study, substantial total RD rates were observed in both CGI (36%) and OGI (59%), aligning with earlier research (CGI: 22%; OGI: 30%–64%).^8,11

Surgical management of pediatric traumatic RD is complex, particularly due to strong vitreoretinal adherence and high PVR prevalence. In this study, 48% of cases of CGI-related traumatic RD were treated with SB, while nearly all OGI cases of PPV with SO tamponade was the therapy of choice, which frequently resulted in a permanent tamponade (63.2% with SO tamponade at the final follow-up). In line with this, Sheard et al⁹ found 94.1% of OGI-related traumatic RD cases underwent PPV, compared with 46.7% in the CGI group. We observed several cases in the OGI group that required encircling SB in a secondary procedure. Potentially, a combined surgery in a primary procedure could improve outcomes, especially in cases with present PVR. A previous study recommended combined surgery for traumatic RD with PVR in both CGI and OGI cases.²² Investigations on the efficacy of standalone or combined PPV with SO tamponade and encircling SB for pediatric traumatic RD are required to potentially improve outcomes, especially in traumatic RD after OGI.

Various success rates are used to assess the outcome of RD surgery, often categorized as anatomical or functional. In our study, we used the SSAS rate as the best possible anatomical outcome, reaching 31.9%, and notably a nonsignificantly higher rate of 44.0% after CGI compared with OGI (18.2%). Comparing success rates with other studies is difficult due to nonstandardized definitions (present tamponade, follow-up time, etc.) and heterogeneous trauma severity. Sheard et al⁹ reported a “primary success” (attached retina after single surgery) in a total of 70.6% of eyes treated with SB and 46.7% in the group treated with primary PPV, but specifics were unclear (present tamponade, rates for CGI and OGI not given separately). Wang et al¹¹ defined anatomic success as retinal reattachment at the final follow-up (≥6 months after surgery) in the absence of SO tamponade, with a rate of 57.1% for OGI, yet surgical details were unspecified (e.g., number of surgeries needed). Similarly, other studies reported overall anatomic success rates of 65% for CGI and 46% for OGI, regardless of persistent SO tamponade (CGI 30%, OGI 54%) and number of surgeries.⁸ A recent Turkish study achieved a 72% anatomical success rate with PPV for traumatic RD, but no details on persistent tamponades were given.¹⁰ Despite limited comparability, the vague definition of these rates might be congruent to our study's LAS rate definition, which was above 90% for both injury types. Recent data suggest improved success rates compared with earlier studies.^8–11

Functional success definitions also vary across studies. Yaşa et al¹⁰ defined it as BCVA ≥ 5/200 Snellen (logMAR 1.6), with 37% achieving this definition after PPV for traumatic RD. In our study, 51% reached this level, but we suggest defining success as BCVA of logMAR ≤1.0 (≥20/200 Snellen), achieved by 40% (CGI: 60%; OGI 9.1%) of the patients in our cohort. Previous studies reported logMAR ≤1.0 (≥20/200 Snellen) rates of approximately 10% (combined CGI and OGI), 45% (CGI), and 23% (OGI).^8,10 Wang et al¹¹ found 36% reaching logMAR ≤0.7 (≥20/100) after OGI. Sheard et al's⁹ outcomes were favorable in CGI (71%) and OGI (60%), yet not directly comparable, as this included trauma cases without traumatic RD. However, favorable outcomes of logMAR ≤1.0 (≥20/100) were observed in 71.0% of the CGI cases (total n = 33, 30 cases with traumatic RD) and in 60.0% of the OGI cases (total n = 28, 17 cases with traumatic RD). Comparability of our data from 2002 to 2021 with older studies with data from the 1980s to 2000s is limited due to heterogeneity in trauma severity, inclusion criteria, and outcome definitions.^8,9,11 Advances in vitreoretinal surgery since the 1980s have surely improved the management of complex RD management.^23–27 Smaller instruments benefit pediatric vitreoretinal surgery, but assessing the impact of technical advancement on pediatric traumatic RD remains challenging due to limited data. While anatomic success rates seem to have improved over the past decades, our data do not clearly indicate enhanced functional outcomes. Functional outcomes in both CGI and OGI were within the range of prior studies, suggesting injury type's greater influence than current surgical advancements. Single surgery anatomic success and FAS correlated with superior final BCVA but are end-of-treatment assessments. Preoperative indicators would be more useful for forecasting visual outcomes. In this study, baseline BCVA was the best predictor of the final BCVA, which is consistent with previous studies.^9,11,12 Higher final BCVA was also seen, as in previous studies,^9–11 in the absence of PVR and older age. Prospective studies with defined success criteria are needed to assess potential outcome improvements.

This study has limitations due to its retrospective design and some missing data at a single follow-up, yet highlights that traumatic RD in children remains a complex and challenging condition. With limited available data, this research provides recent insights into managing traumatic RD in children. Our findings could help in the decision making of treatment strategies, improve visual outcomes, and offer patients valuable prognostication. Given the complexity of RD associated with severe ocular injury, a personalized approach remains indispensable.

Conclusion

Traumatic RD is a complex pathology often resulting in severe visual impairment. Our retrospective analysis of pediatric traumatic RD following CGI and OGI unveils the intricate challenges in managing these cases. Open globe injury–related traumatic RD often requires more invasive and extensive surgical interventions. While surgical advancements seem to have improved anatomical outcomes over time, achieving optimal functional outcomes, especially in cases with PVR, remains a significant hurdle. Personalized approaches are a prerequisite for improved outcomes. Further research and developments to refine treatment strategies and forecast visual prognoses accurately in these young patients are needed.