Efficacy of Eye Patching for Traumatic Corneal Abrasions: A Controlled Clinical Trial

ORIGINAL CONTRIBUTION Efficacy of Eye Patching for Traumatic Corneal Abrasions: A Controlled Clinical Trial From Research Center, Centre Hospitalier Affilié de l Université Laval (CHA), * and the Department of Social and Preventive Medicine, Faculty of Medicine, Laval University, Quebec City, Quebec, Canada. Author contributions are provided at the end of this article. Received for publication October, 999. Revisions received April 2, 2; December 5, 2; and January 6, 2. Accepted for publication February 27, 2. Presented at the North American Primary Care Research annual meeting, Montreal, Quebec, Canada, November 998. Supported by the Quebec Association of Emergency Medicine (AMUQ), the Foundation of the CHA (Enfant-Jesus Hospital), the CHA Research Center, the Quebec Federation of General Practitioners (FMOQ), and the Department of Family Medicine, Laval University. Address for reprints: Natalie Le Sage, MD, Département d urgence, Centre hospitalier Affilié de l Université Laval, Hopital de l Enfant-Jésus, 4, 8 ième rue, Québec, Québec, Canada GJ Z4; 48-649-5632, fax 48-649-5975; E-mail nlesage@videotron.ca. Copyright 2 by the American College of Emergency Physicians. 96-644/2/$35. + 47//5443 doi:.67/mem.2.5443 Natalie Le Sage, MD, MSc * René Verreault, MD, PhD * Louis Rochette, MSc * Study objective: We sought to evaluate the efficacy of eye patching in the treatment of traumatic corneal abrasions in terms of time to healing and reduction in pain and discomfort. Methods: One hundred sixty-three patients presenting at the emergency department of a large university-affiliated hospital with traumatic corneal abrasions were included in this singleblind prospective controlled trial. Eligible patients were assigned to of 2 treatment regimens: topical antibiotic ointment and occlusive patch over the affected eye (n=82) or topical antibiotic ointment 4 times a day without an occlusive patch (n=8). Patients were reexamined every 24 hours until corneal healing occurred. Healing evaluation was performed by the emergency physician, using a slit lamp with fluorescein staining, without knowledge of the patient s assignment to a treatment group. The degree of discomfort was assessed at each visit by using a visual analog scale. Results: Both treatment groups were similar regarding size of the corneal lesions, delay from trauma to first ED visit, presence of foreign body or siderosis, initial degree of discomfort, and presence of specific symptoms (irritation, foreign body sensation, photophobia, redness, and pain). In the patched group, cumulative incidences of healing were 5%, 78%, and 92% after, 2, and 3 days, respectively, compared with 6%, 83%, and 88% in the nonpatched group. In the patched group, symptoms of initial discomfort decreased by 4.8, 4., and 5.5 cm after, 2, and 3 days, respectively, compared with 3.3, 5., and 6.5 cm in the nonpatched group. Conclusion: Eye patching does not appear to be beneficial in the treatment of traumatic corneal abrasions compared with topical antibiotic ointment. [Le Sage N, Verreault R, Rochette L. Efficacy of eye patching for traumatic corneal abrasions: a controlled clinical trial. Ann Emerg Med. August 2;28:29-34.] AUGUST 2 38:2 ANNALS OF EMERGENCY MEDICINE 29

INTRODUCTION Corneal abrasions number among the most frequent ocular conditions encountered in the emergency department.,2 More than 2, cases are diagnosed each year in Quebec EDs. 3 Although complete recovery occurs in most cases within days without further difficulties, corneal abrasions generally cause significant local pain and discomfort. 4 It has been repeatedly argued that reducing friction of the eyelid over the cornea could improve comfort and accelerate healing of corneal abrasions. Accordingly, occlusive patching of the affected eye has long been considered as part of the standard treatment and is still commonly recommended in standard textbooks of ophthalmology and emergency medicine. 4-6 However, routine use of this patch has been increasingly questioned in recent years. 7-9 Occlusive dressings impair binocular vision, obscure half of the visual field, and may carry a significant risk for anaerobic infections, particularly in patients using contact lenses. Moreover, it has been suggested that patching may actually increase local discomfort, causing many patients to remove them prematurely during treatment. Some authors have even reported that eye patching may delay recovery.,2 A few recent studies have attempted to assess the efficacy of occlusive patching of the eye in the treatment of traumatic abrasions of the cornea. -7 Although these studies generally found little support for this treatment, many primary care physicians are still using eye patching because most of these trials were conducted in specialized ophthalmology clinics with small samples of patients and evaluated healing and discomfort without making use of single-blind procedures. The objective of this controlled single-blind study was to assess the efficacy of eye patching in traumatic corneal abrasions among patients presenting to the ED, where the vast majority of cases are routinely diagnosed and treated. 3 To our knowledge, our study is the first to have been conducted with a large number of patients attending an ED and with a singleblind assessment of effects. MATERIALS AND METHODS Each year, more than 66, patients are seen in our ED. The study included patients presenting with a traumatic corneal abrasion (with or without foreign bodies) at the ED of a large university-affiliated hospital in Quebec City between January, 995, and September 3, 996. To be eligible, subjects had to be between 8 and 6 years old, had to be able to report with sufficient details the circumstances and timing of the ocular trauma, and had to demonstrate corneal fluorescein uptake. Patients with corneal perforation, history of glaucoma, chemical burns, UV keratitis, bilateral abrasions, and suspected corneal ulcers were excluded. Eligible patients were assigned to of 2 treatment groups by alternating every other patient. Patients in the first group were treated at the initial visit and at each follow-up visit with topical erythromycin ointment and a double occlusive patch over the affected eye. Patients in the second group were given topical erythromycin ointment to be applied 4 times a day without a dressing. In both groups, removal of foreign bodies, siderosis, or both were performed whenever necessary. The use of a mydriatic agent or of an opioid analgesic prescription was left to the practitioner s discretion and recorded. All subjects provided informed written consent, and the study was approved by the Research Ethics Committee of C.H.A. Enfant-Jésus Hospital. When a patient met the inclusion criteria, a specially trained research nurse on call was responsible for the verification of inclusion and exclusion criteria for all potential patients at the initial visit to ensure randomization of eligible cases to treatment groups and preservation of the single-blind nature of the study. A standardized questionnaire was completed by all patients to record information on circumstances of the eye trauma, time elapsed since the trauma occurred, and the presence of specific symptoms, such as irritation, foreign body sensation, photophobia, redness, and pain. The patients degree of discomfort was assessed by using a visual analog scale. This scale is a well-established and validated 2-cm horizontal scale including qualitative markers to ensure adequate understanding by the patient. 8-2 All measurements of discomfort were determined with this visual scale by the patient without the help of the research nurse and recorded to the nearest millimeter. All patients were examined at initial visit by the emergency physician with a slit lamp, and the dimensions of the corneal lesions were recorded on a standardized form. All patients were reexamined every 24 hours until complete healing was observed. Delays from 8 to 3 hours between visits were judged acceptable, but the exact time of each visit was recorded. To preserve concealment, patients assigned to the occlusive patch treatment group were asked to remove the patch about half an hour before presenting to the ED for follow-up. At each followup visit, the following data were recorded for all patients: dimensions of the corneal lesions measured with the slit lamp, presence of symptoms, use of mydriatic agents, use of analgesics (type and dosage), and patient compliance to treatment. At each visit, patients were asked to use the 3 ANNALS OF EMERGENCY MEDICINE 38:2 AUGUST 2

visual analog scale for rating pain and discomfort at the moment of the visit, as well as the average level of discomfort since the last visit. Patients not presenting for a visit were systematically contacted by telephone to encourage them to come to the ED for follow-up, and in case of refusal, information was collected on assessment of discomfort and presence of symptoms. Throughout the Figure. Profile trial figure. * Specific reasons for exclusion: patient had glaucoma, 2 patients had ocular alkali burns, 2 patients had abrasions in both eyes, and patient had already put a patch on his eye. 72 potentially eligible patients 3 patients refused to participate 6 patients had exclusion criteria * Assignation study, emergency physicians were kept unaware of group assignment and patients self-ratings of discomfort; they were simply asked to determine whether the patient was healed. Healing of corneal abrasions was defined as the absence of corneal lesions after topical administration of fluorescein or the presence of isolated minimal residual fluorescein uptake with mild or no symptoms. These criteria are generally accepted in the literature. Cumulative incidences of healing were compared by using Kaplan-Meier survival analysis. Reduction of discomfort over time was assessed from the scores on the visual analog scale by calculating the following timed differences: Score at initial visit Score at a specific follow-up visit. comparisons of the reduction in discomfort were performed by using median values with interquartile ranges. RESULTS Participant flow is shown in Figure. Baseline variables were similar in the patched and nonpatched groups (Table ), and between those who did and did not complete follow-up. Initial visit n=82 group Initial visit n=8 group Follow-up Visit 2 (24 h) n=68 tel. info. n=5 n=73 Visit 3 (48 h) n=66 Visit 4 (72 h) n=65 Lost during follow-up: n=7 but 5 of them gave reliable information by telephone (no symptoms at all) Complete protocol n=65 including tel. info. n=7 Follow-up Visit 2 (24 h) n=72 tel. info. n=3 n=75 Visit 3 (48 h) n=7 Visit 4 (72 h) n=7 Lost during follow-up: n= but 3 of them gave reliable information by telephone (no symptoms at all) Complete protocol n=7 including tel. info. n=73 Table. Personal and clinical characteristics at initial evaluation by treatment group. Characteristic (n=82) (n=8) Age (y) * 32 (28 38) 36 (3 46) Proportion of male patients (%) 82 9 Presence of foreign bodies (%) 6 62 Presence of siderosis (%) 23 25 Local irritation (%) 84 83 Foreign body sensation (%) 83 86 Photophobia (%) 4 35 Redness (%) 65 59 Use of mydriatics (%) 2 7 Score of discomfort (cm) * 6. (4.5 9.4) 5.7 (3.9 7.7) Surface of lesions (mm 2 ) * 3. (. 6.) 2. (. 4.) Time from trauma to initial visit (h) * 9.5 (3. 9.) 8. (3. 8.) Time from initial to first follow-up visit (h) * 23.3 (2 25.) 23.3 (2 25.) Time from first to second follow-up visit (h) * 23.8 (22.3 24.) 23.8 (23.5 24.7) * Values are given as median (interquartile range, 25th 75th percentile). n=68 for patched group, and n=72 for nonpatched group. n=66 for patched group, and n=7 for nonpatched group. AUGUST 2 38:2 ANNALS OF EMERGENCY MEDICINE 3

Table 2 shows the cumulative incidence of corneal healing over time. All analyses were also conducted within subgroups of patients with low ( 6 cm) and high (>6 cm) scores of discomfort on the visual scale at the initial visit, as well as those with and without foreign bodies, siderosis, or both. All results were very similar in all subgroups of patients examined (Figures 2-5). Reduction in local pain and discomfort over time was also very similar between the 2 treatment groups (Table 3). Rates of symptomatic improvement are contrasted in Table 4. Only 8 patients, 4 in each group, used oral analgesics (acetaminophen with or without codeine) during the study period. DISCUSSION We report the first large comparative trial of patching for corneal abrasions by using blinded outcome measures in an ED setting. This controlled single-blind clinical trial did not show any advantage of the occlusive patch over topical erythromycin (without dressing) in the treatment of traumatic corneal abrasions in the ED. Use of the occlu- Figure 3. for patients with a score of discomfort of 6 cm or greater (n=44 for the patched group and n=37 for the nonpatched group). Table 2. over time * by treatment group. Corneal Healing Probability (n=82) (n=8) After d After 2 d 8 3 After 3 d.92 8 * Kaplan-Meyer survival analysis..9.3.2. 2 3 Figure 2. for all patients (n=82 for the patched group and n=8 for the nonpatched group). Figure 4. for patients with corneal foreign bodies (n=49 for the patched group and n=5 for the nonpatched group)..9.3.2. 2 3.9.3.2. 2 3 32 ANNALS OF EMERGENCY MEDICINE 38:2 AUGUST 2

sive patch did not have any effect on rate of recovery, relief of pain, discomfort, or other symptoms during treatment. These findings remained the same regardless of the intensity of initial local discomfort or of other clinical characteristics of the corneal trauma (eg, the presence of foreign bodies or siderosis). Similar results have been reported in previous studies. In a randomized clinical trial published in 99, Hulbert 3 found no advantage of the patch in corneal abrasions with foreign bodies, but the study included only 3 subjects, and pain was assessed with a dichotomous question (ie, painless or painful) without any blinded procedures. Subsequently, Kirkpatrick et al, 4 in a trial including 44 patients, and Patterson et al, 5 in a study with 33 patients, found similar results in the treatment of corneal abrasions without corneal foreign bodies, but these studies did not include blinded assessment of pain and suffered from significant losses during follow-up. In a larger study including 2 patients and published in 995, Kaiser reported a better reduction in discomfort (6 mm on the visual analog scale), a slightly shorter recovery time (2 hours), and a lower frequency of troubled vision in favor of not using an occlusive patch. However, these differences, although statistically significant, were relatively small in magnitude and clinical significance. Again, in this study, assessments of recovery and discomfort were not performed with blinded procedures. A Canadian study with a small sample of 46 patients, using precise measurements of re-epithelialization, found no statisti- cally significant advantage of the occlusive patch in the treatment of corneal abrasions. 6 The study published in 997 by Campanile et al 2 included 64 patients with superficial corneal abrasions without foreign bodies. Recovery rate after 24 hours was significantly lower in patients with (84%) than in patients without (97%) an occlusive patch, but the effect on relief of pain or other symptoms was not examined. Although a meta-analysis 7 was published showing no clinical benefit of eye patching, and although some authors are already trying to find alternative treatments, 22,23 many primary care physicians are still searching for evidence against patching because most of these trials were conducted on small samples of patients or in specialized ophthalmology clinics. Unlike most previous reports, our study included a relatively large number of patients, Table 3. Reduction of discomfort by treatment group. Reduction of Discomfort * (n=82) (n=8) Discomfort at first visit 6. (4.5 9.4) 5.7 (3.9 7.7) Discomfort during first 24 h 4.7 (2. 5.8) 3.7 (2.2 5.8) Reduction after d 4.8 (2.2 7.) 3.3 (.5 5.8) Reduction after 2 d 4. (2.7 8.7) 5. (3. 7.3) Reduction after 3 d 5.5 (2.8 7.7) 6.5 (3.4 7.7) * Values are given as median (interquartile range, 25th 75th percentile). Figure 5. for patients without corneal foreign bodies (n=33 for the patched group and n=3 for the nonpatched group)..9.3.2. 2 3 Table 4. Number of patients presenting symptoms at initial visit and number (and proportions) of patients still presenting symptoms after 24 and 48 hours by type of symptom and by treatment group. (n=82) (n=8) Symptom Symptom Symptom Symptom Symptom Symptom Present Present Present Present Present Present at Initial at 24 h at 48 h at Initial at 24 h at 48 h Symptom Visit (%) (%) (%) Visit (%) (%) (%) Local irritation 84 39 2 83 49 9 Sensation of 83 2 2 86 9 5 foreign body Photophobia 4 5 6 35 7 6 Eye redness 65 32 8 59 36 9 Pain 54 5 47 4 2 AUGUST 2 38:2 ANNALS OF EMERGENCY MEDICINE 33

maintained a blind assessment of corneal healing by participating physicians during the follow-up, was conducted in the general setting of an ED where most cases of corneal abrasions are routinely diagnosed and treated, and suffered from minimal losses to follow-up. Unlike the other studies, the amount of analgesic medications is reported. Moreover, to our knowledge, no study assessing the patients degree of discomfort during treatment (ie, between visits) has yet been reported. There are some limitations to our study. First, it is possible that we did not detect a minimal difference that could exist between the 2 groups. However, this difference would probably be clinically insignificant. Second, the assessment of recovery used recognized clinical criteria on the basis of the opinion of the attending physician, but the definition of healing could allow some subjectivity. However, the use of a single-blind methodology enhances the validity of all healing measurements. Finally, 2% of the patients were lost to follow-up. This rate of noncompliance remains lower than those of most studies, which vary between % and 34%. -6 It should not have any significant effect on the results validity because it was low and similar in our 2 treatment groups. Dressing corneal abrasions with ocular patches has well-recognized negative consequences, such as impairment of binocular vision and reduction of the visual field. 7 These may have significant effects on driving abilities, may interfere with activities of daily living, and may even delay returning to work. In addition, potential risks of severe anaerobic bacteria infections have been reported in patients wearing contact lenses. Our study confirms that the use of eye patching, although still widely used in primary care and in emergency medicine, should be abandoned for its lack of efficacy. Interestingly, our results indicate that although most patients present relatively rapid corneal healing and reduction of symptoms after 24 hours, pain and discomfort remain significant in the early period of follow-up, especially within the initial 24 hours of treatment, with or without patching. Clinicians should focus on more effective ways to relieve pain without complications during this critical early period. REFERENCES. Lubeck D, Greene JS. Corneal injuries. Emerg Med Clin North Am. 988;6:73-94. 2. Vaughan D, Asbury T. General Ophthalmology. 4th ed. Norwalk, CT: Lange Medical Publications; 995:26-29, 357-358. 3. Statistics of Health Ministry of the Province of Quebec (Canada). Quebec, Canada: Health Ministry of the Province of Quebec; 2. 4. Wright KW. Textbook of Ophthalmology. st ed. Baltimore, MD: William & Wilkins; 997. 5. Mitchell JD. Ocular emergencies. In: Tintinalli JE, Kelen GD, Stapczynski JS, eds. Emergency Medicine: A Comprehensive Study Guide. 5th ed. New York, NY: McGraw-Hill; 2:5-58. 6. Wu G. Ophthalmology for Primary Care. Philadelphia, PA: WB Saunders; 997:5-5. 7. Munro I. Traumatic corneal abrasion [editorial]. Lancet. 987;2:25. 8. Patel BCK. Efficacy of eye pad in corneal healing [letter]. Lancet. 99;337:7-7. 9. Leasty D. Is an eye pad needed in cases of corneal abrasion? [letter]. BMJ. 993;37:22.. Clemons CS, Cohen EJ, Arentsen JJ, et al. Pseudomonas ulcers following patching of corneal abrasions associated with contact lens wear. CLAO J. 987;3:6-64.. Kaiser PK. A comparison of pressure patching versus no patching for corneal abrasions due to trauma or foreign body removal. The Corneal Abrasion Patching Study. Ophthalmology. 995;2:936-942. 2. Campanile TM, St Clair DA, Benaim M. The evaluation of eye patching in the treatment of traumatic corneal epithelial defects. J Emerg Med. 997;5:769-774. 3. Hulbert MFG. Efficacy of eyepad in corneal healing after corneal foreign body removal. Lancet. 99;337:643. 4. Kirkpatrick JNP, Hoh HB, Cook SD. No eye pad for corneal abrasion. Eye. 993;7:468-47. 5. Patterson J, Fetzer D, Krall J, et al. Eye patch treatment for the pain of corneal abrasion. South Med J. 996;89:227-229. 6. Arbour JD, Brunette I, Boisjoly HM, et al. Should we patch corneal erosions? Arch Ophthalmol. 997;5:33-37. 7. Flynn CA, D Amico F, Smith G. Should we patch corneal abrasions? A meta-analysis. J Fam Pract. 998;47:264-27. 8. Huskisson EC. Measurement of pain. Lancet. 974;2:27-3. 9. Scoville B, Kreiglstein GK, Then E, et al. Measuring drug-induced eye irritation: a simple new clinical assay. J Clin Pharmacol. 985;25:2-28. 2. Ho K, Spence J, Murphy MF. Review of pain-measurement tools. Ann Emerg Med. 996;27:427-432. 2. Wewers ME, Lowe NK. A critical review of visual analogue scales in the measurement of clinical phenomena. Res Nurs Health. 99;3:227-236. 22. Kaiser PK, Pineda R. A study of topical nonsteroidal anti-inflammatory drops and no pressure patching in the treatment of corneal abrasions. Ophthalmology. 997;4:353-359. 23. Szucs PA, Nashed AH, Allegra JR, et al. Safety and efficacy of diclofenac ophthalmic solution in the treatment of corneal abrasions. Ann Emerg Med. 2;35:3-37. Author contributions: NL conceived the study and designed the trial. NL and RV obtained research funding. NL supervised the conduct of the trial and data collection. LR provided statistical advice on study design and analyzed the data. NL drafted the manuscript, and RV contributed substantially to its revision. NL takes the responsibility for the paper as a whole. We thank Mrs. Line Beauchesne for her assistance, the research nurses for their collaboration, and all emergency physicians of the C.H.A. Enfant-Jésus Hospital for their help in patient accrual. 34 ANNALS OF EMERGENCY MEDICINE 38:2 AUGUST 2