Orbit, 2013; 32(6): 356–361 ! Informa Healthcare USA, Inc. ISSN: 0167-6830 print / 1744-5108 online DOI: 10.3109/01676830.2013.764452
ORIGINAL ARTICLE
Enucleation versus Evisceration in Ocular Trauma: A Retrospective Review and Study of Current Literature Chengjie Zheng and Albert Y. Wu
ABSTRACT Purpose: To compare variables and outcomes from ocular trauma leading to either enucleation or evisceration to better inform surgical decision making. Design: Retrospective chart review. Methods: We reviewed 441 patients between 2001 and 2012 presenting with ocular trauma to a Level 1 trauma center in Queens, New York; of these, there were 16 enucleations and 6 eviscerations. Retrospective chart review noted age, gender, mechanism of injury, initial and final visual acuity, time to surgery, length of follow-up, pain, degree of motility, and complications. A review of literature in the context of our study was performed. Results: 20 patients were male and 2 patients were female; average age was 44 (SD: 20.0, range 18–91). 9/16 patients were enucleated to prevent sympathetic ophthalmia, whereas only 1/5 patient was eviscerated for this indication (p = 0.1619). No cases of sympathetic ophthalmia were reported over an average follow-up of 316 days. Average length of follow-up varied significantly between the two groups, with an average of 370.4 days (SD: 566.9, range 0–1870) for enucleated eyes and 172.7 days (SD: 146.3, range 0–422) for eviscerated eyes (p = 0.42). Medpor implants were preferred in eviscerations (5/6 eviscerations), whereas hydroxyapatite implants were preferred in enucleations (10/16 enucleations, p = 0.04). Conclusions: Surgical decision-making in ocular trauma is largely based on surgeon preference and experience, with minimal evidence in the literature to support either enucleation or evisceration. We recommend evisceration over enucleation in cases of reliable patient follow-up due to the low incidence of sympathetic ophthalmia. Keywords: Enucleation, evisceration, ocular trauma, orbital trauma, orbital implant
INTRODUCTION
A handful of studies have sought to analyze and compare the two methods with regards to outcomes; however, most do not focus on trauma specifically. Specific advantages of evisceration have been reported, including implant motility, fornix outcomes, sulcus contours, and complication rates.1 However, most of these studies discuss eye removal for all indications, with trauma accounting for 5–51% of study populations.2,3 Studies that do focus on trauma, conversely, do not specifically analyze the differences between evisceration and enucleation.4 This study aims to compare outcomes with regard to enucleation or evisceration after ocular trauma to better inform a decision that is currently largely dictated by surgeon preference.
Ocular trauma is a significant source of morbidity in North America and around the world, with over 2 million cases of eye injury estimated by the American Society of Ocular Trauma. Although most cases of eye injury result in significant recovery, severe trauma can lead to complete loss of ocular architecture and visual potential, requiring evisceration or enucleation of the eye. Evisceration is increasingly preferred in trauma because of the ease of the procedure, potential superior mobility of the prosthetic implant, and minimization of psychological trauma for the patient. However, there still remains significant debate regarding the selection of approach.
20 13
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
Department of Ophthalmology, Mount Sinai School of Medicine, New York, USA
Received 19 September 2012; Revised 4 December 2012; Accepted 4 January 2013; Published online 2 August 2013 Correspondence: Albert Ya-Po Wu, Mount Sinai School of Medicine, Ophthalmology, 1 Gustave L. Levy Place, Box 1183, New York 10029, United States. E-mail: [email protected]
356
Enucleation versus evisceration in ocular trauma 357
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
METHODS Elmhurst Hospital Center is a 545-bed community hospital affiliated with Mount Sinai School of Medicine and serves as a Level 1 trauma center in Queens, NY (population 2 million). We retrospectively reviewed 441 electronic and paper charts between January 1, 2001 and March 30, 2012 of patients presenting with ocular trauma. Charts were identified by querying the Elmhurst Trauma Registry and the Medical Records Department for all orbital traumas leading to evisceration of the eye, enucleation of the eye, or orbital exenteration. First, 111 open globes were discovered, leading to 16 enucleations and 6 eviscerations; no orbital exenterations for trauma were identified. Surgeries were performed by 7 or more different attending physicians. Clinical outcomes were evaluated with respect to age, gender, mechanism of injury, initial and final visual acuity,
TABLE 1. Patient demographics and characteristics of presenting injuries.
Type of surgery Age in years (range, median) Male Gender (%) Right eye (OD) Left eye (OS) Penetrating injury Blunt injury Presence of fractures (N, %) Initial visual acuity LP (N, %)
Enucleation
Evisceration
Overall
16 49 (18–91, 49) 94% 9 7 8 8 7 (44%)
6 32 (18–57, 29) 83% 4 2 4 2 6 (100%)
22 44 (18–91, 42) 91% 13 9 12 10 13 (59%)
3
0
3 (14%)
time to surgery, length of follow-up, pain, degree of motility, and complications. Statistical analysis was performed using the student’s t-test for average values and Fisher’s exact test for comparison of proportions. A review of the literature was performed to compare results from existing studies to ours.
RESULTS We identified 16 enucleations and 6 eviscerations from 21 patients; one patient received bilateral evisceration for a total of 22 eyes (Table 1). The mean age was 44 (SD 20.0, range 18–91); 91% of the patients were male. Case numbers per year are shown in Figure 1. The Ocular Trauma Score uses specific diagnostic criteria to predict visual prognosis in patients with ocular trauma.5 The score is based on initial visual acuity with the presence of globe rupture, endophthalmitis, perforating injury, retinal detachment, and/or afferent pupillary defect (APD). All but 3 patients had an Ocular Trauma Score of 1, with a raw score of 37, by virtue of globe rupture with an initial visual acuity of no light perception (NLP). One of the remaining three also had an OTS score of 1 due to globe rupture with LP vision and an APD. The other two had an OTS score of 2 – one had an initial visual acuity of 20/200 but with globe rupture and retinal detachment, and the other had an initial visual acuity of light perception (LP) with globe rupture. The mechanism of injury was related to assault in 12 (57%) of patients; 5 involved penetrating gunshot wounds and 2 involved penetrating knife injuries. One patient was involved in a hit-and-run motor vehicle accident while crossing a pedestrian walk; the remainder involved accidental injuries.
FIGURE 1. Frequency of enucleation and evisceration from 2001–2011 at Elmhurst Hospital Center in Elmhurst, NY. !
2013 Informa Healthcare USA, Inc.
358 C. Zheng & A. Y. Wu TABLE 2. Comparison of operative characteristics and post-operative follow-up.
Type of surgery Time to initial surgery (days) Time to enucleation/evisceration (days) Medpor implant Hydroxyapatite implant Length of follow-up (days)
Enucleation (standard deviation, range)
Evisceration (standard deviation, range)
Overall (p Value)
16 2.69 (6.84, 0–27) 18.4 (31.3, 0–127) 3 10 370.4 (566.9, 0–1870)
6 1.5 (1.38, 0–3) 2.5 (1.5, 1–5) 5 1 172.7 (146.2, 0–422)
22 2.4 (p = 0.68) 14.0 (p = 0.24) 8 (p = 0.04) 11 (p = 0.04) 316.5 (p = 0.42)
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
TABLE 3. Indications for evisceration or enucleation.
Inability to repair globe Prophylaxis against sympathetic ophthalmia after NLP vision in injured eye Blind painful eye Prior poor/absent vision
A total of 16 (73%) eyes underwent enucleation: 5 (31%) were primary enucleations, and the remainder (69%) were ruptured globe repairs resulting in secondary enucleation (Table 2). Of the five primary enucleations, two were in patients with a previously blind eye. Six (27%) eyes underwent evisceration: 5 were primary eviscerations, and one was a secondary evisceration for severe rupture and sympathetic ophthalmia prophylaxis 5 days after the initial repair. The most common indication for enucleation was prevention of sympathetic ophthalmia (SO) in patients with NLP vision after ruptured globe repair; the most common indication for evisceration was inability to repair the globe primarily (Table 3). Time to initial surgery was longer in the enucleation group, at an average of 2.7 days (SD: 6.8, range 0–27) versus 1.5 days (SD: 1.4, range 0–3) (p = 0.68). Time to final enucleation or evisceration was also longer in the enucleation group, at 18.4 days (SD 31.3, range 0–127) compared with 2.5 days (SD 1.5, range 1– 5) (p = 0.24). In the majority of eviscerations, an 18- or 20-mm Medpor implant was used; whereas a 20-mm hydroxyapatite implant wrapped in a donor scleral shell was most commonly used in enucleations (p = 0.04). Average length of follow-up varied significantly between the two groups, with an average of 370.4 days (SD 566.9, range 0–1870) for enucleated eyes and 172.7 days (SD 146.3, range 0–422) for eviscerated eyes (p = 0.42). Complications were rare within our study group, and were only documented for 3 (14%) patients. The first patient was a 55-year-old Hispanic male who was reportedly hit with a piece of glass in a previously blind eye. His vision was found to be NLP on presentation, and he subsequently underwent primary enucleation due to inability to repair the globe, with placement of a 20-mm hydroxyapatite implant in Vicryl mesh. On postoperative week 1 after enucleation, graft exposure with conjunctival dehiscence
Enucleation
Evisceration
Overall
4 9 1 2
5 1 0 0
9 10 (p = 0.16) 1 2
was noted. Repair was scheduled; however, the patient was lost to follow-up, and no documentation exists of subsequent surgery. The second patient was a 29-year-old African-American male with a gunshot wound to his right temple who underwent bilateral primary eviscerations; conjunctival dehiscence wasnoted 1 month postoperatively in the right eye, and reapposition was achieved with removal of the conformer. The last patient was a 47-year-old Asian male who presented with penetrating injury of his right upper eyelid and globe after working with a metallic knife. He underwent immediate ruptured globe repair with repair of lid margin lacerations upon presentation, followed by enucleation 4 days later for an NLP eye. Follow-up notes consistently documented poor levator function and lagophthalmos, but excellent motility. No cases of sympathetic ophthalmia were reported. Pain was only documented for one patient: a 54-year-old Hispanic woman with a history of blunt trauma to the right eye from moving furniture, who underwent secondary enucleation for an NLP eye 14 days after primary ruptured globe repair. She complained only of some slight discomfort upon movement of her prosthesis. Motility was not well-documented in the follow-up encounters, with no measurements recorded; however, 1 patient was documented to have poor motility. The case involved a 29-year-old Hispanic male with a gunshot wound that traveled from the left medial canthus through the right orbit; he was operated upon the next day with exploration of the right orbit, removal of 2 bullet fragments from the lateral orbital wall, evisceration of the right globe, and insertion of a 20 mm Medpor implant. In addition to the poor motility, he was also noted to have vitreous hemorrhage and commotio retinae in his left (uninjured) eye upon follow-up. Five other patients were specifically noted to have good motility; all underwent primary Orbit
Enucleation versus evisceration in ocular trauma 359 ruptured globe repair with subsequent secondary enucleation within 2 weeks of presentation.
DISCUSSION
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
Enucleation or Evisceration There little in the literature to guide the choice of evisceration versus enucleation, with limited evidence to support either approach in cases of ocular trauma. Current studies generally focus on either enucleation versus evisceration for all causes, or on the management of ocular trauma without specific comparison of either technique. Nakra et al. is one of the first and most extensive studies involving the systematic comparison of enucleation and evisceration for all causes, and examined 32 enucleated eyes versus 52 eviscerated eyes1. They demonstrated statistically significant superior implant motility, but not prosthetic motility, in eviscerated eyes; the latter often holds greater significance for patients. The study also found improved fornix outcomes and sulcus contours in eviscerated eyes, although this again did not necessarily translate to an overall improved aesthetic outcome. A lower complication rate with eviscerations was also noted as compared to enucleations. Based on these findings, the authors concluded that evisceration was a safe alternative to enucleation, with the potential for better aesthetic and motility outcomes. The applicability of these findings to trauma should be considered with regard to the study population; only 10% of the enucleated patients and 21% of the eviscerated patients were operated on due to trauma; the majority received surgery due to malignancy, a blind painful eye, or endophthalmitis.1 Tari et al. likewise compares 50 eviscerated eyes to 50 enucleated eyes and outcomes for each; of note, only 5 eyes in this study were operated on for trauma, all of which were enucleated.3 Consistent with the Nakra et al. study, the primary significant difference between the two groups resided in superior motility in eviscerated eyes, with horizontal movement of 10.25 1.99 mm in eviscerated eyes versus 6.90 1.74 mm in enucleated eyes (p50.001), and vertical movement of 8.45 1.89 in eviscerated eyes versus 5.69 1.63 in enucleated eyes (p50.001). Yousuf et al. compared enucleation versus evisceration more recently, and found no significant difference in the rate of complications between enucleations and eviscerations, contrary to the Nakra et al. study.2 In addition, they experienced an increased rate of implant exposure with hydroxyapatite implants compared with glass sphere implants, despite reports on hydroxyapatite implants being well-tolerated in the literature. Because of similar complication rates and !
2013 Informa Healthcare USA, Inc.
reduced operative time, the authors concluded that evisceration was preferable as a safe and efficient alternative to enucleation. Migliori addressed this debate with a review of the literature in 2002 with regard to evisceration versus enucleation for all causes, and noted varying reports on the duration and intensity of postoperative pain, with some suggesting that evisceration may lead to greater pain than enucleation.6 However, the review found consensus in the finding that while time course differed, ultimate and permanent relief of pain was able to be achieved for both enucleations and eviscerations performed on a blind, painful eye.6 Although these studies shed light on the controversy, there is a paucity of data that addresses trauma in particular. Trauma is often devastating and commonly involves accompanying injuries surrounding the globe, such as orbital fractures, eyelid lacerations, and general disruption of local anatomy and architecture. This can carry with it alterations in outcomes after surgery; for example, the risk of implant explosure has been shown to be higher in trauma by Tabatabaee et al., likely due to impaired integrity of the Tenon’s capsule.7 This highlights the need for the question of enucleation versus evisceration to be addressed specifically with regard to severe globe injuries in ocular trauma. Savar et al. represents one of the only large-scale studies to examine enucleation for globe trauma specifically, and reviews 660 open globes resulting in 55 enucleations or eviscerations. The study found that there was a low rate of eye removal overall, indicating that most ruptured globes could be repaired primarily. In addition, repairs that led to NLP vision in the affected eye could be observed for signs of sympathetic ophthalmia in most (57%) of cases without the need for secondary enucleation.4 This was based on the finding that only 2 patients (0.3%) developed sympathetic ophthalmia in the study group.4
Sympathetic Ophthalmia One of the primary concerns informing the decision to select enucleation over evisceration lies in the risk of sympathetic ophthalmia, from antigens causing panuveitis and visual risk in the non-injured eye. Cases were first reported in the late 19th century, leading to a decline in the rate of evisceration6; however, in recent years, extremely low rates of sympathetic ophthalmia (0.1–0.3% after ocular trauma) have led to a resurgence of preference for eviscerations over enucleations due to the aforementioned possible benefits in motility and aesthetic outcome.6 The etiology of sympathetic ophthalmia is thought to be due to several ocular antigens, including retinal S-antigen, retinoid binding protein, melanin
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
360 C. Zheng & A. Y. Wu associated antigens, and choroidal and retinal pigment epithelial antigens.6,8 HLA-DRB1-04 and DQA103 have also been implicated.6,8 Galor et al. completed one of the largest multicenter studies on sympathetic ophthalmia with 85 patients, and noted that trauma makes up the majority of cases of sympathetic ophthalmia. In addition, traumatically induced cases of SO generally occurs sooner than surgically induced cases, with a median of 6.5 months after the inciting event in trauma, compared with 14.3 months after surgery.9 Trauma also resulted in a 2.39 times higher risk of worse vision (20/50 or worse) compared with surgically-related cases of SO.9 The study also found, however, that systemic steroids were associated with disease inactivation, suggesting that they are effective in treating SO. This is consistent with Savar et al.’s finding that steroids and immunosuppressive therapy were able to preserve vision in the two cases of SO in their study.4 Interestingly, despite the wide acceptance of SO prophylaxis as an indication for enucleation, 22% of inciting eyes in the study had already undergone enucleation, suggesting that enucleation may not be as reliable of a safeguard against SO as assumed. This finding, combined with the already low rates of SO found in the literature, brings into question the acceptance of enucleation after trauma as prophylaxis.10,11,12 In our present study, the most common indication for enucleation was prevention of SO, highlighting the continued concern in our institution of the need for surgical prophylaxis. This is especially interesting when considering Bellan’s calculation that, based on an SO rate of 0.28%, 9999 enucleations would be required to prevent one case.4,10,13 This indicates that hospitals may benefit from some type of clinical risk stratification before considering routine prophylactic enucleation.
Study Findings Our study did not reflect the trend toward evisceration reflected in the literature; this may be due to the low number of cases involved. It does not show one particular surgeon’s preference, as the ruptured globes in this study were managed by seven or more attending physicians. The predominance of men in this study is consistent with trauma literature in general; penetrating injuries were roughly equal with blunt injuries. A higher fraction of the eviscerated eyes were associated with orbital fracture (6/6, 100%) compared with enucleated eyes (7/16, 44%), suggesting that severe injury does not necessarily favor enucleation. Interestingly, the time it took for patients to be operated on upon initial presentation was longer in
the enucleation group, leading to questions as to whether this is a cause of subsequent enucleation, or a reflection of the severity of injury. Its lack of statistical significance is likely due to the disproportionately smaller sample size in the evisceration group. It is plausible that treatment delay may have led to decompensation requiring eventual enucleation of the eye; it is also plausible that patients with more severe injuries may have required life-saving emergent care before their eyes were addressed. Furthermore, time to final removal of the eye was also longer in the enucleation group as opposed to the evisceration group, with a mean of 19.5 days that lies outside the recommended 2-week window for secondary enucleation.4 This may be due to operating room scheduling difficulties; however, the low rate of complications in this study suggests that outcomes may not have been affected, especially given a median of 6.5 months for the onset of SO cited above. In our study, implant type was also closely associated with the type of procedure chosen, with a Medpor implant used in almost all eviscerations, and a hydroxyapatite implant wrapped in a banked scleral shell used in almost all enucleations. This was not consistent with all institutions reported in the literature, and may represent institution preference; it does not represent a single surgeon’s preference in this study, as mentioned above. A recent study has suggested that Medpor implants are more prone to implant exposure than wrapped hydroxyapatite implants; our study was not powered to detect this difference.7 It should also be noted that although the use of donor sclera is common in the US, access to eye banks worldwide is hugely variable. Last, patients in our enucleation group had significantly longer follow-up times compared with the eviscerated group, which cannot be explained by any other variable. It is presumed that patients would consider loss of the eye through either surgery to be equally serious, and that motivation to maintain care would be similar. One possibility is that enucleations led to a higher rate of complications that was not documented, leading to greater need for return; however, this seems unlikely. The other possibility is that the smaller number of eviscerations skews the data, and that there is not sufficient sample size to draw conclusions on this difference. In either scenario, this length of follow-up has significant implications on the detection of sympathetic ophthalmia, as this condition can occur many years after the inciting event. Since our eviscerated group had shorter followup times, it is possible that cases of sympathetic ophthalmia that occurred in eviscerated eyes were not captured in this study’s data. Again, however, it might be expected that patients would return for follow-up if they did experience a complication related to vision loss in their contralateral eye. Orbit
Enucleation versus evisceration in ocular trauma 361
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
CONCLUSIONS Our retrospective review represents one of the first attempts to evaluate evisceration versus enucleation specifically in the context of trauma. It was conducted at a high-volume Level 1 trauma center, making it an ideal setting for this study. Results from our study showed a low rate of postoperative pain and complications, with no cases of sympathetic ophthalmia over an average follow-up of 316 days. Review of the literature with consideration of our cases shows that systematic enucleation of NLP eyes for SO prophylaxis should not be performed because of the extremely low incidence of SO, and the fact that it may be well treated with corticosteroids. This is especially true given the possible benefits in motility and complication rate demonstrated by previous studies in choosing evisceration. Thus, we recommend that evisceration to be performed in cases of ocular trauma when the globe cannot be salvaged, in patients that have a history of compliance and reliable follow-up. The greatest limitations to this study lie in its reliance on the completeness of previous documentation due to its retrospective nature. Initial identification of cases was dependent upon coding of diagnoses and procedures; it is plausible that not all cases of ocular trauma leading to anophthalmic surgery were coded correctly. In addition, the analysis of these cases hinges upon careful documentation by evaluating physicians at the time of trauma, during operative dictations, and during follow-up visits. Considerable variation was seen in the level of detail of documentation, limiting the ability to do a full comprehensive analysis. For example, motility measurements were not quantified, and comments regarding extraocular movements were only found for certain patients. Lastly, the smaller scale of this study is also a limiting factor, and likely reflects the rarity of globe removal after trauma, the size and community setting of the hospital, and the fact that it is not a major referral center. Future directions should focus on larger scale studies comparing the outcomes of enucleation versus evisceration in open globe injuries, with specific regard to motility, enophthalmos, aesthetic outcome, presence of concurrent fractures, type of implant used, and complication rate. A prospective, multi-center study would provide a sufficiently large population for statistics and standardization of data. In addition, we found that rates of prophylactic enucleations as seen in trauma centers and community centers like ours are not currently justified by SO incidence. This question should be further explored by examining the outcome of NLP eyes that are not secondarily enucleated after ruptured globe repair. Future studies on sympathetic ophthalmia should also focus on unearthing risk
!
2013 Informa Healthcare USA, Inc.
factors for SO that could better inform surgical decision-making in ocular trauma.
ACKNOWLEDGEMENTS This study was approved the by the Institutional Review Board at Mount Sinai School of Medicine, and by the Elmhurst Hospital Research Committee.
DECLARATION OF INTEREST The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
REFERENCES 1. Nakra T, Simon GJ, Douglas RS, et al. Comparing outcomes of enucleation and evisceration. Ophthalmology 2006;113(12):2270–2275. Epub 2006 Sep 25. 2. Yousuf SJ, Jones LS, Kidwell Jr ED. Enucleation and evisceration: 20 years of experience. Orbit 2012; 31(4):211–215. Epub 2012 May 29. 3. Tari AS, Malihi M, Kasaee A, et al. Enucleation with hydroxyapatite implantation versus evisceration plus scleral quadrisection and alloplastic implantation. Ophthal Plast Reconstr Surg 2009;25(2):130–133. 4. Savar A, Andreoli MT, Kloek CE, Andreoli CM. Enucleation for open globe injury. Am J Ophthalmol 2009; 147(4):595–600.e1. Epub 2009 Feb 1. 5. Kuhn F, Maisiak R, Mann L, et al. The Ocular Trauma Score (OTS). Ophthalmol Clin North Am 2002;15(2):163–165, vi. 6. Migliori ME. Enucleation versus evisceration. Curr Opin Ophthalmol 2002;13(5):298–302. 7. Tabatabaee Z, Mazloumi M, Rajabi MT, et al. Comparison of the exposure rate of wrapped hydroxyapatite (Bio-Eye) versus unwrapped porous polyethylene (Medpor) orbital implants in enucleated patients. Ophthal Plast Reconstr Surg 2011;27(2):114–118. 8. Kilmartin DJ, Dick AD, Forrester JV. Prospective surveillance of sympathetic ophthalmia in the UK and Republic of Ireland. Br J Ophthalmol 2000;84:259–263. 9. Galor A, Davis JL, Flynn Jr HW, et al. Sympathetic ophthalmia: incidence of ocular complications and vision loss in the sympathizing eye. Am J Ophthalmol 2009; 148(5):704–710.e2. Epub 2009 Aug 7. 10. Bilyk JR. Enucleation, evisceration, and sympathetic ophthalmia. Curr Opin Ophthalmol 2000;11(5):372–386. 11. Levine MR, Pou CR, Lash RH. The 1998 Wendell Hughes Lecture. Evisceration: Is sympathetic ophthalmia a concern in the new millennium? Ophthal Plast Reconstr Surg 1999; 15(1):4–8 12. du Toit N, Motala MI, Richards J, et al. The risk of sympathetic ophthalmia following evisceration for penetrating eye injuries at Groote Schuur Hospital. Br J Ophthalmol 2008;92(1):61–63. 13. Rasmussen ML, Prause JU, Johnson M, et al. Review of 345 eye amputations carried out in the period 1996–2003, at Rigshospitalet, Denmark. Acta Ophthalmol 2010; 88(2):218–221. Epub 2008 Dec 24.
ORIGINAL ARTICLE
Enucleation versus Evisceration in Ocular Trauma: A Retrospective Review and Study of Current Literature Chengjie Zheng and Albert Y. Wu
ABSTRACT Purpose: To compare variables and outcomes from ocular trauma leading to either enucleation or evisceration to better inform surgical decision making. Design: Retrospective chart review. Methods: We reviewed 441 patients between 2001 and 2012 presenting with ocular trauma to a Level 1 trauma center in Queens, New York; of these, there were 16 enucleations and 6 eviscerations. Retrospective chart review noted age, gender, mechanism of injury, initial and final visual acuity, time to surgery, length of follow-up, pain, degree of motility, and complications. A review of literature in the context of our study was performed. Results: 20 patients were male and 2 patients were female; average age was 44 (SD: 20.0, range 18–91). 9/16 patients were enucleated to prevent sympathetic ophthalmia, whereas only 1/5 patient was eviscerated for this indication (p = 0.1619). No cases of sympathetic ophthalmia were reported over an average follow-up of 316 days. Average length of follow-up varied significantly between the two groups, with an average of 370.4 days (SD: 566.9, range 0–1870) for enucleated eyes and 172.7 days (SD: 146.3, range 0–422) for eviscerated eyes (p = 0.42). Medpor implants were preferred in eviscerations (5/6 eviscerations), whereas hydroxyapatite implants were preferred in enucleations (10/16 enucleations, p = 0.04). Conclusions: Surgical decision-making in ocular trauma is largely based on surgeon preference and experience, with minimal evidence in the literature to support either enucleation or evisceration. We recommend evisceration over enucleation in cases of reliable patient follow-up due to the low incidence of sympathetic ophthalmia. Keywords: Enucleation, evisceration, ocular trauma, orbital trauma, orbital implant
INTRODUCTION
A handful of studies have sought to analyze and compare the two methods with regards to outcomes; however, most do not focus on trauma specifically. Specific advantages of evisceration have been reported, including implant motility, fornix outcomes, sulcus contours, and complication rates.1 However, most of these studies discuss eye removal for all indications, with trauma accounting for 5–51% of study populations.2,3 Studies that do focus on trauma, conversely, do not specifically analyze the differences between evisceration and enucleation.4 This study aims to compare outcomes with regard to enucleation or evisceration after ocular trauma to better inform a decision that is currently largely dictated by surgeon preference.
Ocular trauma is a significant source of morbidity in North America and around the world, with over 2 million cases of eye injury estimated by the American Society of Ocular Trauma. Although most cases of eye injury result in significant recovery, severe trauma can lead to complete loss of ocular architecture and visual potential, requiring evisceration or enucleation of the eye. Evisceration is increasingly preferred in trauma because of the ease of the procedure, potential superior mobility of the prosthetic implant, and minimization of psychological trauma for the patient. However, there still remains significant debate regarding the selection of approach.
20 13
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
Department of Ophthalmology, Mount Sinai School of Medicine, New York, USA
Received 19 September 2012; Revised 4 December 2012; Accepted 4 January 2013; Published online 2 August 2013 Correspondence: Albert Ya-Po Wu, Mount Sinai School of Medicine, Ophthalmology, 1 Gustave L. Levy Place, Box 1183, New York 10029, United States. E-mail: [email protected]
356
Enucleation versus evisceration in ocular trauma 357
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
METHODS Elmhurst Hospital Center is a 545-bed community hospital affiliated with Mount Sinai School of Medicine and serves as a Level 1 trauma center in Queens, NY (population 2 million). We retrospectively reviewed 441 electronic and paper charts between January 1, 2001 and March 30, 2012 of patients presenting with ocular trauma. Charts were identified by querying the Elmhurst Trauma Registry and the Medical Records Department for all orbital traumas leading to evisceration of the eye, enucleation of the eye, or orbital exenteration. First, 111 open globes were discovered, leading to 16 enucleations and 6 eviscerations; no orbital exenterations for trauma were identified. Surgeries were performed by 7 or more different attending physicians. Clinical outcomes were evaluated with respect to age, gender, mechanism of injury, initial and final visual acuity,
TABLE 1. Patient demographics and characteristics of presenting injuries.
Type of surgery Age in years (range, median) Male Gender (%) Right eye (OD) Left eye (OS) Penetrating injury Blunt injury Presence of fractures (N, %) Initial visual acuity LP (N, %)
Enucleation
Evisceration
Overall
16 49 (18–91, 49) 94% 9 7 8 8 7 (44%)
6 32 (18–57, 29) 83% 4 2 4 2 6 (100%)
22 44 (18–91, 42) 91% 13 9 12 10 13 (59%)
3
0
3 (14%)
time to surgery, length of follow-up, pain, degree of motility, and complications. Statistical analysis was performed using the student’s t-test for average values and Fisher’s exact test for comparison of proportions. A review of the literature was performed to compare results from existing studies to ours.
RESULTS We identified 16 enucleations and 6 eviscerations from 21 patients; one patient received bilateral evisceration for a total of 22 eyes (Table 1). The mean age was 44 (SD 20.0, range 18–91); 91% of the patients were male. Case numbers per year are shown in Figure 1. The Ocular Trauma Score uses specific diagnostic criteria to predict visual prognosis in patients with ocular trauma.5 The score is based on initial visual acuity with the presence of globe rupture, endophthalmitis, perforating injury, retinal detachment, and/or afferent pupillary defect (APD). All but 3 patients had an Ocular Trauma Score of 1, with a raw score of 37, by virtue of globe rupture with an initial visual acuity of no light perception (NLP). One of the remaining three also had an OTS score of 1 due to globe rupture with LP vision and an APD. The other two had an OTS score of 2 – one had an initial visual acuity of 20/200 but with globe rupture and retinal detachment, and the other had an initial visual acuity of light perception (LP) with globe rupture. The mechanism of injury was related to assault in 12 (57%) of patients; 5 involved penetrating gunshot wounds and 2 involved penetrating knife injuries. One patient was involved in a hit-and-run motor vehicle accident while crossing a pedestrian walk; the remainder involved accidental injuries.
FIGURE 1. Frequency of enucleation and evisceration from 2001–2011 at Elmhurst Hospital Center in Elmhurst, NY. !
2013 Informa Healthcare USA, Inc.
358 C. Zheng & A. Y. Wu TABLE 2. Comparison of operative characteristics and post-operative follow-up.
Type of surgery Time to initial surgery (days) Time to enucleation/evisceration (days) Medpor implant Hydroxyapatite implant Length of follow-up (days)
Enucleation (standard deviation, range)
Evisceration (standard deviation, range)
Overall (p Value)
16 2.69 (6.84, 0–27) 18.4 (31.3, 0–127) 3 10 370.4 (566.9, 0–1870)
6 1.5 (1.38, 0–3) 2.5 (1.5, 1–5) 5 1 172.7 (146.2, 0–422)
22 2.4 (p = 0.68) 14.0 (p = 0.24) 8 (p = 0.04) 11 (p = 0.04) 316.5 (p = 0.42)
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
TABLE 3. Indications for evisceration or enucleation.
Inability to repair globe Prophylaxis against sympathetic ophthalmia after NLP vision in injured eye Blind painful eye Prior poor/absent vision
A total of 16 (73%) eyes underwent enucleation: 5 (31%) were primary enucleations, and the remainder (69%) were ruptured globe repairs resulting in secondary enucleation (Table 2). Of the five primary enucleations, two were in patients with a previously blind eye. Six (27%) eyes underwent evisceration: 5 were primary eviscerations, and one was a secondary evisceration for severe rupture and sympathetic ophthalmia prophylaxis 5 days after the initial repair. The most common indication for enucleation was prevention of sympathetic ophthalmia (SO) in patients with NLP vision after ruptured globe repair; the most common indication for evisceration was inability to repair the globe primarily (Table 3). Time to initial surgery was longer in the enucleation group, at an average of 2.7 days (SD: 6.8, range 0–27) versus 1.5 days (SD: 1.4, range 0–3) (p = 0.68). Time to final enucleation or evisceration was also longer in the enucleation group, at 18.4 days (SD 31.3, range 0–127) compared with 2.5 days (SD 1.5, range 1– 5) (p = 0.24). In the majority of eviscerations, an 18- or 20-mm Medpor implant was used; whereas a 20-mm hydroxyapatite implant wrapped in a donor scleral shell was most commonly used in enucleations (p = 0.04). Average length of follow-up varied significantly between the two groups, with an average of 370.4 days (SD 566.9, range 0–1870) for enucleated eyes and 172.7 days (SD 146.3, range 0–422) for eviscerated eyes (p = 0.42). Complications were rare within our study group, and were only documented for 3 (14%) patients. The first patient was a 55-year-old Hispanic male who was reportedly hit with a piece of glass in a previously blind eye. His vision was found to be NLP on presentation, and he subsequently underwent primary enucleation due to inability to repair the globe, with placement of a 20-mm hydroxyapatite implant in Vicryl mesh. On postoperative week 1 after enucleation, graft exposure with conjunctival dehiscence
Enucleation
Evisceration
Overall
4 9 1 2
5 1 0 0
9 10 (p = 0.16) 1 2
was noted. Repair was scheduled; however, the patient was lost to follow-up, and no documentation exists of subsequent surgery. The second patient was a 29-year-old African-American male with a gunshot wound to his right temple who underwent bilateral primary eviscerations; conjunctival dehiscence wasnoted 1 month postoperatively in the right eye, and reapposition was achieved with removal of the conformer. The last patient was a 47-year-old Asian male who presented with penetrating injury of his right upper eyelid and globe after working with a metallic knife. He underwent immediate ruptured globe repair with repair of lid margin lacerations upon presentation, followed by enucleation 4 days later for an NLP eye. Follow-up notes consistently documented poor levator function and lagophthalmos, but excellent motility. No cases of sympathetic ophthalmia were reported. Pain was only documented for one patient: a 54-year-old Hispanic woman with a history of blunt trauma to the right eye from moving furniture, who underwent secondary enucleation for an NLP eye 14 days after primary ruptured globe repair. She complained only of some slight discomfort upon movement of her prosthesis. Motility was not well-documented in the follow-up encounters, with no measurements recorded; however, 1 patient was documented to have poor motility. The case involved a 29-year-old Hispanic male with a gunshot wound that traveled from the left medial canthus through the right orbit; he was operated upon the next day with exploration of the right orbit, removal of 2 bullet fragments from the lateral orbital wall, evisceration of the right globe, and insertion of a 20 mm Medpor implant. In addition to the poor motility, he was also noted to have vitreous hemorrhage and commotio retinae in his left (uninjured) eye upon follow-up. Five other patients were specifically noted to have good motility; all underwent primary Orbit
Enucleation versus evisceration in ocular trauma 359 ruptured globe repair with subsequent secondary enucleation within 2 weeks of presentation.
DISCUSSION
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
Enucleation or Evisceration There little in the literature to guide the choice of evisceration versus enucleation, with limited evidence to support either approach in cases of ocular trauma. Current studies generally focus on either enucleation versus evisceration for all causes, or on the management of ocular trauma without specific comparison of either technique. Nakra et al. is one of the first and most extensive studies involving the systematic comparison of enucleation and evisceration for all causes, and examined 32 enucleated eyes versus 52 eviscerated eyes1. They demonstrated statistically significant superior implant motility, but not prosthetic motility, in eviscerated eyes; the latter often holds greater significance for patients. The study also found improved fornix outcomes and sulcus contours in eviscerated eyes, although this again did not necessarily translate to an overall improved aesthetic outcome. A lower complication rate with eviscerations was also noted as compared to enucleations. Based on these findings, the authors concluded that evisceration was a safe alternative to enucleation, with the potential for better aesthetic and motility outcomes. The applicability of these findings to trauma should be considered with regard to the study population; only 10% of the enucleated patients and 21% of the eviscerated patients were operated on due to trauma; the majority received surgery due to malignancy, a blind painful eye, or endophthalmitis.1 Tari et al. likewise compares 50 eviscerated eyes to 50 enucleated eyes and outcomes for each; of note, only 5 eyes in this study were operated on for trauma, all of which were enucleated.3 Consistent with the Nakra et al. study, the primary significant difference between the two groups resided in superior motility in eviscerated eyes, with horizontal movement of 10.25 1.99 mm in eviscerated eyes versus 6.90 1.74 mm in enucleated eyes (p50.001), and vertical movement of 8.45 1.89 in eviscerated eyes versus 5.69 1.63 in enucleated eyes (p50.001). Yousuf et al. compared enucleation versus evisceration more recently, and found no significant difference in the rate of complications between enucleations and eviscerations, contrary to the Nakra et al. study.2 In addition, they experienced an increased rate of implant exposure with hydroxyapatite implants compared with glass sphere implants, despite reports on hydroxyapatite implants being well-tolerated in the literature. Because of similar complication rates and !
2013 Informa Healthcare USA, Inc.
reduced operative time, the authors concluded that evisceration was preferable as a safe and efficient alternative to enucleation. Migliori addressed this debate with a review of the literature in 2002 with regard to evisceration versus enucleation for all causes, and noted varying reports on the duration and intensity of postoperative pain, with some suggesting that evisceration may lead to greater pain than enucleation.6 However, the review found consensus in the finding that while time course differed, ultimate and permanent relief of pain was able to be achieved for both enucleations and eviscerations performed on a blind, painful eye.6 Although these studies shed light on the controversy, there is a paucity of data that addresses trauma in particular. Trauma is often devastating and commonly involves accompanying injuries surrounding the globe, such as orbital fractures, eyelid lacerations, and general disruption of local anatomy and architecture. This can carry with it alterations in outcomes after surgery; for example, the risk of implant explosure has been shown to be higher in trauma by Tabatabaee et al., likely due to impaired integrity of the Tenon’s capsule.7 This highlights the need for the question of enucleation versus evisceration to be addressed specifically with regard to severe globe injuries in ocular trauma. Savar et al. represents one of the only large-scale studies to examine enucleation for globe trauma specifically, and reviews 660 open globes resulting in 55 enucleations or eviscerations. The study found that there was a low rate of eye removal overall, indicating that most ruptured globes could be repaired primarily. In addition, repairs that led to NLP vision in the affected eye could be observed for signs of sympathetic ophthalmia in most (57%) of cases without the need for secondary enucleation.4 This was based on the finding that only 2 patients (0.3%) developed sympathetic ophthalmia in the study group.4
Sympathetic Ophthalmia One of the primary concerns informing the decision to select enucleation over evisceration lies in the risk of sympathetic ophthalmia, from antigens causing panuveitis and visual risk in the non-injured eye. Cases were first reported in the late 19th century, leading to a decline in the rate of evisceration6; however, in recent years, extremely low rates of sympathetic ophthalmia (0.1–0.3% after ocular trauma) have led to a resurgence of preference for eviscerations over enucleations due to the aforementioned possible benefits in motility and aesthetic outcome.6 The etiology of sympathetic ophthalmia is thought to be due to several ocular antigens, including retinal S-antigen, retinoid binding protein, melanin
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
360 C. Zheng & A. Y. Wu associated antigens, and choroidal and retinal pigment epithelial antigens.6,8 HLA-DRB1-04 and DQA103 have also been implicated.6,8 Galor et al. completed one of the largest multicenter studies on sympathetic ophthalmia with 85 patients, and noted that trauma makes up the majority of cases of sympathetic ophthalmia. In addition, traumatically induced cases of SO generally occurs sooner than surgically induced cases, with a median of 6.5 months after the inciting event in trauma, compared with 14.3 months after surgery.9 Trauma also resulted in a 2.39 times higher risk of worse vision (20/50 or worse) compared with surgically-related cases of SO.9 The study also found, however, that systemic steroids were associated with disease inactivation, suggesting that they are effective in treating SO. This is consistent with Savar et al.’s finding that steroids and immunosuppressive therapy were able to preserve vision in the two cases of SO in their study.4 Interestingly, despite the wide acceptance of SO prophylaxis as an indication for enucleation, 22% of inciting eyes in the study had already undergone enucleation, suggesting that enucleation may not be as reliable of a safeguard against SO as assumed. This finding, combined with the already low rates of SO found in the literature, brings into question the acceptance of enucleation after trauma as prophylaxis.10,11,12 In our present study, the most common indication for enucleation was prevention of SO, highlighting the continued concern in our institution of the need for surgical prophylaxis. This is especially interesting when considering Bellan’s calculation that, based on an SO rate of 0.28%, 9999 enucleations would be required to prevent one case.4,10,13 This indicates that hospitals may benefit from some type of clinical risk stratification before considering routine prophylactic enucleation.
Study Findings Our study did not reflect the trend toward evisceration reflected in the literature; this may be due to the low number of cases involved. It does not show one particular surgeon’s preference, as the ruptured globes in this study were managed by seven or more attending physicians. The predominance of men in this study is consistent with trauma literature in general; penetrating injuries were roughly equal with blunt injuries. A higher fraction of the eviscerated eyes were associated with orbital fracture (6/6, 100%) compared with enucleated eyes (7/16, 44%), suggesting that severe injury does not necessarily favor enucleation. Interestingly, the time it took for patients to be operated on upon initial presentation was longer in
the enucleation group, leading to questions as to whether this is a cause of subsequent enucleation, or a reflection of the severity of injury. Its lack of statistical significance is likely due to the disproportionately smaller sample size in the evisceration group. It is plausible that treatment delay may have led to decompensation requiring eventual enucleation of the eye; it is also plausible that patients with more severe injuries may have required life-saving emergent care before their eyes were addressed. Furthermore, time to final removal of the eye was also longer in the enucleation group as opposed to the evisceration group, with a mean of 19.5 days that lies outside the recommended 2-week window for secondary enucleation.4 This may be due to operating room scheduling difficulties; however, the low rate of complications in this study suggests that outcomes may not have been affected, especially given a median of 6.5 months for the onset of SO cited above. In our study, implant type was also closely associated with the type of procedure chosen, with a Medpor implant used in almost all eviscerations, and a hydroxyapatite implant wrapped in a banked scleral shell used in almost all enucleations. This was not consistent with all institutions reported in the literature, and may represent institution preference; it does not represent a single surgeon’s preference in this study, as mentioned above. A recent study has suggested that Medpor implants are more prone to implant exposure than wrapped hydroxyapatite implants; our study was not powered to detect this difference.7 It should also be noted that although the use of donor sclera is common in the US, access to eye banks worldwide is hugely variable. Last, patients in our enucleation group had significantly longer follow-up times compared with the eviscerated group, which cannot be explained by any other variable. It is presumed that patients would consider loss of the eye through either surgery to be equally serious, and that motivation to maintain care would be similar. One possibility is that enucleations led to a higher rate of complications that was not documented, leading to greater need for return; however, this seems unlikely. The other possibility is that the smaller number of eviscerations skews the data, and that there is not sufficient sample size to draw conclusions on this difference. In either scenario, this length of follow-up has significant implications on the detection of sympathetic ophthalmia, as this condition can occur many years after the inciting event. Since our eviscerated group had shorter followup times, it is possible that cases of sympathetic ophthalmia that occurred in eviscerated eyes were not captured in this study’s data. Again, however, it might be expected that patients would return for follow-up if they did experience a complication related to vision loss in their contralateral eye. Orbit
Enucleation versus evisceration in ocular trauma 361
Orbit Downloaded from informahealthcare.com by University of Waterloo on 02/12/15 For personal use only.
CONCLUSIONS Our retrospective review represents one of the first attempts to evaluate evisceration versus enucleation specifically in the context of trauma. It was conducted at a high-volume Level 1 trauma center, making it an ideal setting for this study. Results from our study showed a low rate of postoperative pain and complications, with no cases of sympathetic ophthalmia over an average follow-up of 316 days. Review of the literature with consideration of our cases shows that systematic enucleation of NLP eyes for SO prophylaxis should not be performed because of the extremely low incidence of SO, and the fact that it may be well treated with corticosteroids. This is especially true given the possible benefits in motility and complication rate demonstrated by previous studies in choosing evisceration. Thus, we recommend that evisceration to be performed in cases of ocular trauma when the globe cannot be salvaged, in patients that have a history of compliance and reliable follow-up. The greatest limitations to this study lie in its reliance on the completeness of previous documentation due to its retrospective nature. Initial identification of cases was dependent upon coding of diagnoses and procedures; it is plausible that not all cases of ocular trauma leading to anophthalmic surgery were coded correctly. In addition, the analysis of these cases hinges upon careful documentation by evaluating physicians at the time of trauma, during operative dictations, and during follow-up visits. Considerable variation was seen in the level of detail of documentation, limiting the ability to do a full comprehensive analysis. For example, motility measurements were not quantified, and comments regarding extraocular movements were only found for certain patients. Lastly, the smaller scale of this study is also a limiting factor, and likely reflects the rarity of globe removal after trauma, the size and community setting of the hospital, and the fact that it is not a major referral center. Future directions should focus on larger scale studies comparing the outcomes of enucleation versus evisceration in open globe injuries, with specific regard to motility, enophthalmos, aesthetic outcome, presence of concurrent fractures, type of implant used, and complication rate. A prospective, multi-center study would provide a sufficiently large population for statistics and standardization of data. In addition, we found that rates of prophylactic enucleations as seen in trauma centers and community centers like ours are not currently justified by SO incidence. This question should be further explored by examining the outcome of NLP eyes that are not secondarily enucleated after ruptured globe repair. Future studies on sympathetic ophthalmia should also focus on unearthing risk
!
2013 Informa Healthcare USA, Inc.
factors for SO that could better inform surgical decision-making in ocular trauma.
ACKNOWLEDGEMENTS This study was approved the by the Institutional Review Board at Mount Sinai School of Medicine, and by the Elmhurst Hospital Research Committee.
DECLARATION OF INTEREST The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
REFERENCES 1. Nakra T, Simon GJ, Douglas RS, et al. Comparing outcomes of enucleation and evisceration. Ophthalmology 2006;113(12):2270–2275. Epub 2006 Sep 25. 2. Yousuf SJ, Jones LS, Kidwell Jr ED. Enucleation and evisceration: 20 years of experience. Orbit 2012; 31(4):211–215. Epub 2012 May 29. 3. Tari AS, Malihi M, Kasaee A, et al. Enucleation with hydroxyapatite implantation versus evisceration plus scleral quadrisection and alloplastic implantation. Ophthal Plast Reconstr Surg 2009;25(2):130–133. 4. Savar A, Andreoli MT, Kloek CE, Andreoli CM. Enucleation for open globe injury. Am J Ophthalmol 2009; 147(4):595–600.e1. Epub 2009 Feb 1. 5. Kuhn F, Maisiak R, Mann L, et al. The Ocular Trauma Score (OTS). Ophthalmol Clin North Am 2002;15(2):163–165, vi. 6. Migliori ME. Enucleation versus evisceration. Curr Opin Ophthalmol 2002;13(5):298–302. 7. Tabatabaee Z, Mazloumi M, Rajabi MT, et al. Comparison of the exposure rate of wrapped hydroxyapatite (Bio-Eye) versus unwrapped porous polyethylene (Medpor) orbital implants in enucleated patients. Ophthal Plast Reconstr Surg 2011;27(2):114–118. 8. Kilmartin DJ, Dick AD, Forrester JV. Prospective surveillance of sympathetic ophthalmia in the UK and Republic of Ireland. Br J Ophthalmol 2000;84:259–263. 9. Galor A, Davis JL, Flynn Jr HW, et al. Sympathetic ophthalmia: incidence of ocular complications and vision loss in the sympathizing eye. Am J Ophthalmol 2009; 148(5):704–710.e2. Epub 2009 Aug 7. 10. Bilyk JR. Enucleation, evisceration, and sympathetic ophthalmia. Curr Opin Ophthalmol 2000;11(5):372–386. 11. Levine MR, Pou CR, Lash RH. The 1998 Wendell Hughes Lecture. Evisceration: Is sympathetic ophthalmia a concern in the new millennium? Ophthal Plast Reconstr Surg 1999; 15(1):4–8 12. du Toit N, Motala MI, Richards J, et al. The risk of sympathetic ophthalmia following evisceration for penetrating eye injuries at Groote Schuur Hospital. Br J Ophthalmol 2008;92(1):61–63. 13. Rasmussen ML, Prause JU, Johnson M, et al. Review of 345 eye amputations carried out in the period 1996–2003, at Rigshospitalet, Denmark. Acta Ophthalmol 2010; 88(2):218–221. Epub 2008 Dec 24.
