Are Signs and Symptoms Associated With Persistent Corneal Abrasions in Children?

Background  Corneal abrasions are common eye injuries in children. Most are treated with antibiotic drops or ointment, patching of the affected eye, and follow-up within 24 hours to confirm resolution by fluorescein examination.

Objective  To determine if signs and symptoms at follow-up were associated with the presence of a persistent corneal abrasion or abnormal visual acuity.

Design  Retrospective case series.

Setting  A children's hospital.

Patients  Children who were aged 4 years or older with the diagnosis of corneal abrasion between May 1992 and December 1996 and who had a follow-up examination.

Results  Seventy-seven patients (57% male) were enrolled (median age, 7 years). The respective sensitivities, specificities, positive predictive values, and negative predictive values of selective signs and symptoms for persistent abrasions were as follows: for pain, 53%, 93%, 80%, and 80%; for photophobia, 57%, 100%, 100%, and 80%; for redness, 100%, 46%, 44%, and 100%; for pain and redness, 40%, 96%, 80%, and 80%; and for at least 1 sign or symptom, 95%, 48%, 47%, and 95%. Twenty-six patients had persistent corneal abrasions at follow-up. Six of these 26 patients were symptom free at follow-up, and 15 patients had only redness as a persistent sign. Five patients had abnormal visual acuity, one of whom was asymptomatic. All 3 patients with complications were symptomatic.

Conclusions  Signs and symptoms are inconsistently associated with persistent corneal abrasions. Asymptomatic patients may have persistent corneal abrasions, suggesting the need for selective follow-ups.

EYE INJURIES account for many visits annually to emergency departments and physicians' offices. Corneal abrasions, or epithelial defects, are the most common injuries.¹ In the pediatric age group, corneal abrasions are frequently caused by foreign bodies in the eye or direct trauma to the eye.

Nonverbal patients often present with fussiness alone^2,3 or in combination with a painful, watery, red eye. Older patients usually complain of severe pain and photophobia secondary to exposure of the cornea's sensitive nerve endings.^3-10 Tearing, redness, blurred vision, and foreign body sensation are signs and symptoms that can be present as well.¹¹ If the abrasion is near the visual axis of the eye, visual acuity may be affected.

The epithelial lining of the cornea is the eye's protective barrier to infection. Although rare, patients with an unhealed cornea may develop bacterial infections, ulcerations, iritis, or chronic erosion syndrome.^6,8-10 These complications may require long-term medical management or even surgery.^6,8-10 An unhealed corneal defect, especially if linear in shape, may also suggest a retained foreign body.

Treatment of corneal abrasions is directed at relieving pain, preventing infection, and promoting reepithelialization of the cornea. For patients who do not wear contact lenses, therapy often includes application of a broad-spectrum topical antibiotic, with or without a cycloplegic, followed by a unilateral pressure eye patch for at least 24 hours.^4-8

Reevaluations after 24 to 48 hours of treatment are considered routine to confirm healing of the epithelium, to rule out infection, and to check for retained foreign bodies.^3,6-9 These follow-ups are believed to be important as some patients may feel relief when their eye is patched despite development of infection, ulcerations, or retained foreign body. At the follow-up visit, patients are routinely questioned about persistence of symptoms, and the eye is reexamined with fluorescein. If the abrasion is still present, the patient is either repatched for an additional 24 hours or referred to an ophthalmologist.

The traditional 24- to 48-hour follow-up visit, however, is controversial in practice. Although little has been written on this subject, some physicians believe that the follow-up visit is unnecessary if patients are free of signs and symptoms suggesting persistence.^5,12 They typically instruct patients and parents to remove the eye patch at 24 hours and return only if symptoms and signs persist or recur,^5,12 therefore saving patient time and expense of a return visit.

The purpose of this study was to determine if signs and symptoms commonly associated with corneal abrasions are predictive of fluorescein examination findings after 24 hours of treatment. We hypothesized that if verbal patients (4 years or older) were symptom free, they would no longer have uptake of fluorescein indicating an unhealed cornea or have abnormal visual acuities.

A retrospective review of the medical records of children with the diagnosis of corneal abrasions between May 1992 and December 1996 at The Children's Hospital (TCH), Denver, Colo, was performed.

Inclusion criteria were the discharge diagnosis of a corneal abrasion and documentation of a follow-up visit at 24 to 48 hours after the initial visit. Patients were excluded if they were younger than 4 years, sustained other ocular trauma (eg, corneal laceration or hyphema), had preexisting conjunctivitis or other disease in the affected eye, were contact lens wearers (placing them at risk for Pseudomonas infection), had received prior medical treatment for the corneal abrasion, or had follow-up at a facility other than the TCH system.

Medical records were reviewed for demographic information, as well as signs and symptoms, during the initial visit. These included pain, photophobia, blurred vision, foreign body sensation, tearing, and redness. Visual acuity and treatment rendered at the initial visit were also reviewed and noted. Similar information, including evidence of any signs and symptoms, was then evaluated at the follow-up visit. Treatment rendered and disposition were also documented for this second visit.

Fluorescein evaluations are performed at our institution with a dry paper strip impregnated with fluorescein wetted with sterile isotonic sodium chloride solution. A drop of the solution or the strip itself is gently introduced into the patient's bulbar conjunctiva in the lower eyelid. The eyes are then evaluated with a handheld cobalt light with magnification or a Woods lamp, for evaluation of fluorescein uptake on the cornea.

Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for all signs and symptoms at the follow-up visit. χ² Analysis was performed on all independent variables. Only documented presence or absence of signs and symptoms extracted from medical records was used for statistical analysis.

Two hundred fifty-nine patients were identified who had initial visits at TCH and were discharged with the diagnosis of corneal abrasion. One hundred twenty-six patients had follow-up visits in the TCH system. Thirty-four patients were excluded because they were younger than 4 years, and an additional 15 were excluded secondary to having other ocular injuries at the time of presentation (hyphema, traumatic iritis, etc).

Seventy-seven patients met the inclusion criteria for this study. The age range of these patients was 4 to 21 years (median age, 7 years), and the majority of patients (57%) were male. Demographics are listed in Table 1.

The majority of patients (79%) sustained their corneal abrasion as a result of minor trauma from their own finger, a tree branch, or from a toy (Table 2). Nine patients (12%) had a foreign body removed at their initial visit.

Pain and redness in the affected eye were the most common presenting complaints in all ages (Table 3). Reports of blurred vision and foreign body sensation were more common in the older age groups. Overall, the youngest age group (4-6 years old) had fewer reported symptoms. Visual acuity was checked in 44 patients (53%) at the initial visit. Treatment rendered at the initial visit was cycloplegic agents in 12 patients (15%), antibiotic ointment or drops in 67 patients (87%), and an eye patch in 73 patients (95%).

The median time from the initial visit to the follow-up visit was 22 hours (range, 12-46 hours). Documentation of signs and symptoms at follow-up and percentage extracted from charts are listed below.

Fifty-one patients (66%) showed resolution of the corneal abrasions by fluorescein examination at the time of follow-up. The remaining 26 patients (34%) had persistent corneal abrasions. Six (23%) of the 26 patients with persistent corneal abrasions had no sign or symptom at follow-up. Fifteen patients (58%) had only redness as a sign of persistence, and 5 patients (19%) were symptomatic with other or a combination of signs and symptoms.

Sensitivities, specificities, positive predictive values, and negative predictive values of signs and symptoms individually and in combination are listed in Table 4. All combinations of signs and symptoms not listed (including combinations of 3 or more signs and symptoms) were found to not be significant.

Five of the 6 patients who were asymptomatic with a persistent corneal abrasion had eventual resolution of their abrasion, and 1 patient was lost to further follow-up.

The remaining 20 patients with persistent abrasions who were symptomatic at follow-up included the 15 who had redness only as a sign of persistence. At follow-up, 2 of these 20 patients had rust rings removed, 1 had a foreign body removed, and 3 failed to return for a third visit. All remaining patients had resolution of their abrasion at a third visit. No patients at the follow-up visit or subsequent visits had evidence of infection, ulceration of the cornea, or chronic erosion syndrome.

Treatment during the initial visit did not differ between the patients who had resolution of their abrasion and those with persistence. One patient who was asymptomatic with persistence and 2 patients who were symptomatic had been treated with a cycloplegic. This compares with 9 patients who had resolution of their abrasions who had been treated with a cycloplegic (P=.71, χ²). The follow-up period for asymptomatic vs symptomatic patients with persistence also did not differ (median time of 22 hours for both).

Visual acuity was evaluated in only 44 (57%) of the initial patient visits, and 26 (34%) of the follow-up visits. In patients who had visual acuity tested at follow-up, 5 had abnormal acuities (>20/50 in the affected eye). Four of the patients were symptomatic with pain and blurred vision, and at a later follow-up showed resolution of their corneal abrasion and exhibited normal visual acuity. The other remaining patient with abnormal visual acuity had persistence of the abrasion at follow-up and was free of symptoms. This patient was lost to a third or subsequent follow-up.

Corneal abrasions are common pediatric eye injuries. A 1-year survey of all children admitted to the emergency department at Wills Eye Hospital, Philadelphia, Pa, in 1986 revealed that 446 (55%) of 810 patients presented with corneal abrasions.¹

Most corneal abrasions in children are caused by a foreign body in the eye, minor trauma, or self-infliction with a finger. Our study showed that 79% of the abrasions were caused by minor injuries. Nine (12%) of the corneal abrasions were caused by a foreign body irritating the eye, which was removed at the initial visit. These results are slightly different from those of adult studies of corneal abrasions that revealed that 48% were caused by minor trauma and 40% were caused by a known foreign body that was removed at the initial visit.¹¹ This difference possibly reflects differing environmental exposures of children vs adults. Children may be more susceptible to sustaining injuries during play activity on the playground, whereas adults are more likely to have occupational exposures to metal and wood particles at their jobs.

Signs and symptoms at presentation of corneal abrasions have traditionally been noted to include pain, photophobia, foreign body sensation, and redness of the conjunctiva.^3,4,6,10 We found that almost all children, even the younger age groups, had pain and redness. Few children, however, reported foreign body sensation or photophobia. This is similar to the findings of Poole² in patients older than 1 year. In that study, 95% of patients presented with pain, 76% presented with redness, and 76% presented with photophobia. Some symptoms not typically associated with corneal abrasions, such as swelling of the eyelids and discharge from the affected eye, were also seen in our study. They most likely are manifestations of local irritation and rubbing a painful eye.

In our study, the majority of patients were traditionally treated with an eye patch and antibiotics at their initial visit. Only 15% had a cycloplegic installed in the eye to help eliminate pain by decreasing ciliary spasm. This small number was also noted in an adult study of the management of corneal abrasions in which only 6% of patients were treated with a cycloplegic, even when it was thought to be common practice at the study facility.¹² This treatment, although considered in some practices to be standard care, clearly varies across the nation.

Recently in the adult literature, the traditional use of the unilateral eye patch for treatment of corneal abrasions has been challenged.^11-14 Studies in adults have shown a decreased healing time, decreased perception of pain, and better patient compliance with treatment not involving an eye patch.^11-14 No published study to date addresses this issue in pediatric populations. The majority of the patients in this study were treated with an eye patch.

In our study, one third of patients had persistence of their abrasion at follow-up. In one adult study of corneal abrasions, 18% of patched patients and 31% of unpatched patients had incomplete "24-hour healing."¹³ This high percentage of persistence, both in our study and in the adult study, may be related to the timing of follow-up as the mean time of follow-up was approximately 24 hours in both studies. In 2 other adult studies related to patching, healing times for the corneal abrasions were 2.60 ± 0.77 (mean ± SD) days for patched and 2.33 ± 0.66 for unpatched patients and 2.00 ± 0.71 days for patched and 1.55 ± 0.61 days for unpatched, respectively.^11,14 Since these studies showed a healing rate of about 48 hours, a higher healing rate may have been found if our patients' follow-ups had been at a later time.

Signs and symptoms at the follow-up visit, in our study, were poorly associated with the persistence of a corneal abrasion. In fact, 23% of the patients who had persistent abrasions were asymptomatic. The majority of those that were symptomatic (58%) had redness of the conjunctiva only as an indication that the epithelial defect persisted. The negative predictive values of all and even combinations of signs and symptoms were roughly 80%. Sensitivities, other than for redness at 100%, were poor overall with a range of 33% to 57%.

None of the patients in this study, including those who had multiple follow-up visits, showed signs of corneal infection or chronic corneal erosions. The risk of these complications, along with removal of a retained foreign body, has been used to justify early repeated examination and fluorescein staining. In one adult study of 99 patients with corneal abrasions, 4 developed corneal infections.¹² In another adult study of 201 patients, 1 developed chronic recurrent erosion syndrome and 2 subsequently developed conjunctivitis.¹¹ The 2 patients with conjunctivitis were deemed by the authors "not to be associated with the original corneal abrasion."¹¹ In one adult study of patching, half of the complications occurred in the group that was not patched.¹² In our study, 2 patients had rust rings removed and 1 had an unidentified foreign body removed at the time of follow-up. The true incidence of complications from persistent abrasions in the pediatric population, although commented on in most textbooks of pediatric ophthalmology,^8,9 is unknown.

Visual acuity, a critical part of any evaluation for eye trauma, was tested very infrequently in our patients' initial and follow-up visits. Some of the reasons surrounding this low number may be related to poor documentation of the actual testing in the medical record or the inability to test it acutely secondary to patient's age or pain at presentation. This low rate would be considered below what is acceptable in most ophthalmologic practices.

Visual acuity was abnormal in 5 patients at follow-up. One patient reported being asymptomatic with an abnormal acuity and he did not have a previously reported visual acuity recorded to determine whether this was his baseline value. Furthermore, he was eventually lost to follow-up. All other patients within a week had resolution of their symptoms and abnormal acuities. Because of the small number of patients who had visual acuity evaluations, no comments can be made on the effect of persistent abrasions and abnormal visual acuities.

Limitations to this study result largely from its retrospective design. The evaluation and documentation of important signs and symptoms were poor in some instances (ie, visual acuity).

Large numbers of patients eligible for the study were excluded due to follow-up outside the TCH system. Further patients who did have a follow-up and had persistent corneal abrasions were subsequently lost to follow-up. Also, the number of patients in this study was not large enough to comment on risk of infrequent corneal complications (infections, corneal ulcerations, or chronic erosion syndrome) from persistent abrasions in the pediatric population.

In conclusion, patients with persistent corneal abrasions, despite traditional antibiotic and eye patch therapy, can be asymptomatic at follow-up. These asymptomatic patients, however, had no complications or morbidity related to the persistent abrasion in our small study population. Patients who had complications, including rust rings and retained foreign body, continued to be symptomatic at follow-up.

Therefore, our results suggest that patients who continue to be symptomatic after 24 hours of treatment should be reevaluated to assess for complications, persistent foreign bodies, and persistence of abrasion. Selective follow-up for asymptomatic patients may then be entertained with the knowledge that some of these patients will have persistent but probably clinically uncomplicated abrasions. The incidence of corneal infection, corneal ulceration, and chronic erosion syndrome, resulting from persistent unhealed epithelial defects, is unknown and cannot be further commented on from this study.

Further studies are warranted to prospectively determine the association of signs and symptoms with persistence of corneal abrasions and to determine the clinical relevance of the persistent abrasions in the pediatric patient.

Accepted for publication October 14, 1999.

Presented at the Regional Ambulatory Pediatric Association Meeting, Carmel, Calif, February 8, 1997.

Corresponding author: Kristine K. Rittichier, MD, Emergency Department, Primary Children's Medical Center, 100 N Medical Dr, Salt Lake City, UT 84113.

Editor's Note: It looks like we can't depend on obvious signs or symptoms to determine the persistence of corneal abrasions. Eye, Eye, Eye!—Catherine D. DeAngelis, MD