Patching for corneal abrasion

  • Review
  • Intervention

Authors


Abstract

Background

Published audits have demonstrated that corneal abrasions are a common presenting eye complaint. Eye patches are often recommended for treating corneal abrasions despite the lack of evidence for their use. This systematic review was conducted to determine the effects of the eye patch when used to treat corneal abrasions.

Objectives

The objective of this review was to assess the effects of patching for corneal abrasion on healing and pain relief.

Search methods

We searched CENTRAL (which contains the Cochrane Eyes and Vision Trials Register) (2016, Issue 4), Ovid MEDLINE, Ovid MEDLINE In-Process and Other Non-Indexed Citations, Ovid MEDLINE Daily, Ovid OLDMEDLINE (January 1946 to May 2016), EMBASE (January 1980 to May 2016), Latin American and Caribbean Health Sciences Literature Database (LILACS) (January 1982 to May 2016), System for Information on Grey Literature in Europe (OpenGrey) (January 1995 to May 2016), the ISRCTN registry (www.isrctn.com/editAdvancedSearch), ClinicalTrials.gov (www.clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/search/en). We did not use any date or language restrictions in the electronic searches for trials. We last searched the electronic databases on 9 May 2016. We also searched the reference lists of included studies, unpublished 'grey' literature and conference proceedings and contacted pharmaceutical companies for details of unpublished trials.

Selection criteria

We included randomised and quasi-randomised controlled trials that compared patching the eye with no patching to treat simple corneal abrasions.

Data collection and analysis

Two authors independently assessed the risk of bias and extracted data. Investigators were contacted for further information regarding the quality of trials. The primary outcome was healing at 24, 48 and 72 hours while secondary outcomes included measures of pain, quality of life and adverse effects. We graded the certainty of the evidence using GRADE.

Main results

We included 12 trials which randomised a total of 1080 participants in the review. Four trials were conducted in the United Kingdom, another four in the United States of America, two in Canada, one in Brazil and one in Switzerland. Seven trials were at high risk of bias in one or more domains and one trial was judged to be low risk of bias in all domains. The rest were a combination of low risk or unclear.

People receiving a patch may be less likely to have a healed corneal abrasion after 24 hours compared to those not receiving a patch (risk ratio (RR) 0.89, 95% confidence interval (CI) 0.79 to 1.00, 7 trials, 531 participants, low certainty evidence). Similar numbers of people in the patch and no-patch groups were healed by 48 hours (RR 0.97, 95% CI 0.91 to 1.02, 6 trials, 497 participants, moderate certainty evidence) and 72 hours (RR 1.01, 95% CI 0.97 to 1.05, 4 trials, 430 participants, moderate certainty evidence). Participants receiving a patch took slightly longer to heal but the difference was small and probably unimportant (mean difference (MD) 0.14 days longer, 95% CI 0 to 0.27 days longer, 6 trials, 642 participants, moderate certainty evidence).

Ten trials reported pain scores. Most studies reported pain on a visual analogue scale (VAS). It was not possible to pool the data because it was skewed. In general, similar pain ratings were seen between patch and no-patch groups. Data from two trials reporting presence or absence of pain at 24 hours was inconclusive. There was a higher risk of reported pain in the patch group but wide confidence intervals compatible with higher or lower risk of pain (RR 1.51, 95% CI 0.86 to 2.65, 2 trials, 193 participants, low certainty evidence). Five trials compared analgesic use between the patch and no-patch groups. Data from three of these trials could be combined and suggested similar analgesic use in the patch and no-patch groups but with some uncertainty (RR 0.95, 95% CI 0.69 to 1.32, 256 participants, low certainty evidence). Frequently reported symptoms included photophobia, lacrimation, foreign body sensation and blurred vision but there was little evidence to suggest any difference in these symptoms in people with or without a patch.

Activities of daily living (ADL) were assessed in one study involving children. There was little difference in ADL with the exception of walking which was reported to be more difficult with a patch on: VAS 1.7 cm (SD 2.1) versus 0.3 cm (SD 0.7).

Complication rates were low across studies and there is uncertainty about the relative effects of patching or not patching with respect to these (RR 3.24, 95% CI 0.87 to 12.05, 8 trials, 660 participants, low certainty evidence). Three trials reporting rates of compliance to treatment found that 22% of participants did not have their eye patches during follow-up. No-patch groups generally received more adjuvant treatment with antibiotics or cycloplegics, or both, than the patch group. There were limited data on the effect of patching on abrasions greater than 10mm2 in size.

Authors' conclusions

Trials included in this review suggest that treating simple corneal abrasions with a patch may not improve healing or reduce pain. It must be noted that, in these trials, participants who did not receive a patch were more likely to receive additional treatment, for example with antibiotics. Overall we judged the certainty of evidence to be moderate to low. Further research should focus on designing and implementing better quality trials and examining the effectiveness of patching for large abrasions.

Plain language summary

Eye patches for corneal abrasion

What is the aim of this review?
The aim of this Cochrane Review was to find out what effect using an eye patch for corneal abrasions has on healing and pain relief compared with not patching. Cochrane researchers collected and analysed all relevant studies to answer this question and found 12 studies.

Key messages
Patching probably does not speed up healing and may not have an important effect on pain relief. None of the studies provided information on the effect of patching on larger abrasions.

What was studied in the review?
The cornea is the transparent outer layer of the eye. Corneal abrasions can result from scratches or superficial damage to the cornea. These are common problems which can be very painful. A common treatment option is to place a patch over the eye. This may have an impact on how long it takes for the abrasion to heal. It may also provide pain relief.

What are the main results of the review?
The review authors found 12 relevant studies. 6 were from North America, 5 from Europe, and 1 from South America (Brazil). These studies compared the use of eye patches with no patching.

People receiving a patch may be less likely to have a healed corneal abrasion after 24 hours compared with people not receiving a patch (low certainty evidence). Using eye patches probably makes little or no difference to the number of people whose abrasion heals after 48 and 72 hours (moderate certainty evidence).

Corneal abrasions in people receiving patches probably take slightly longer to heal than in people not receiving patches but the difference is small and probably unimportant (moderate certainty evidence).

Using eye patches may lead to more pain at 24 hours (low certainty evidence). However, the range where the actual effect may be shows that eye patches may lead to more pain, but may also lead to less pain.

People with corneal abrasions frequently experience sensitivity to light, watery eyes, a foreign body sensation and blurred vision. There was little evidence to suggest any difference in these symptoms in people with or without a patch.

There were limited data available on quality of life, visual acuity and adverse effects.

How up-to-date is this review?
The review authors searched for studies that had been published up to 9 May 2016.

Laički sažetak

Zavoj oka za liječenje ogrebotine (abrazije) rožnice

Koji je cilj ovog Cochrane sustavnog pregleda?
Cilj ovog Cochraneovog pregleda je ustanoviti učinak zavoja oka kod ogrebotine (abrazije) rožnice na cijeljenje i ublažavanje boli u usporedbi bez primjene zavoja oka. Cochraneovi istraživači su sakupili i analizirali sve relevantne studija i našli 12 studija.

Ključne poruke
Zavoj oka vjerojatno ne ubrzava cijeljenje i nema važan učinak na ublažavanje boli. Niti jedna studija nije imala podatak o učinku zavoja oka na veće ogrebotine rožnice.

Što je istraženo u ovom pregledu literature?
Rožnica je prozirni vanjski sloj oka. Ogrebotine (abrazije) rožnice nastaju od površnih ozljeda rožnice. To je česta i vrlo bolna ozljeda. Standardno liječenje uključuje stavljanje zavoja na oko. Može imati učinak na trajanje cijeljenja ogrebotine. Također može ublažiti bol.

Glavni rezultati pregleda
Autori pregleda su našli 12 relevantnih studija. Šest je bilo iz Sjeverne Amerike, 5 iz Europe i 1 iz Južne Amerike (Brazil). Uspoređeni su rezultati sa i bez zavoja oka.

Kod zavoja na oku je manja šansa cijeljenja nakon 24 h u usporedbi kad nije stavljen zavoj (dokaz niske pouzdanosti). Primjena zavoja ima mali ili nikakav učinak na cijeljenje nakon 48 ili 72 h (dokaz srednje pouzdanosti).

Ogrebotina rožnice cijeli malo dulje ispod zavoja oka u usporedbi kad nije stavljen zavoj, ali je razlika u trajanju mala i vjerojatno nevažna (srednja pouzdanost dokaza).

Zavoj oka moguće je razlogom jače boli 24 h nakon ozljede (dokaz niske pouzdanosti). Primjena zavoja oka je neizvjesna u smislu ublažavanja ili pojačavanja boli.

Osjetljivost na svjetlo, suzenje i osjet stranog tijela i zamućenog vida su česta pojava kod ogrebotine rožnice. Malo je dokaza o učinku zavoja na te simptome.

Podaci o kvaliteti života, oštrini vida i nuspojavama su oskudni.

Koliko su ažurni ovi dokazi?
Autori su pretražili studije koje su objavljene do 9. svibnja 2016.

Bilješke prijevoda

Hrvatski Cochrane
Prevela: Vesna Kušec
Ovaj sažetak preveden je u okviru volonterskog projekta prevođenja Cochrane sažetaka. Uključite se u projekt i pomozite nam u prevođenju brojnih preostalih Cochrane sažetaka koji su još uvijek dostupni samo na engleskom jeziku. Kontakt: cochrane_croatia@mefst.hr

Streszczenie prostym językiem

Stosowanie obturatorów (zasłonek na oko) w otarciach rogówki

Jaki jest cel przeglądu?
Celem tego przeglądu Cochrane było sprawdzenie czy stosowanie zasłonki na oko w otarciach rogówki jest skuteczne w łagodzeniu bólu i pomocne w gojeniu w porównaniu do niestosowania zasłonki. Badacze Cochrane zebrali i przeanalizowali wszystkie istotne badania, aby odpowiedzieć na pytanie, które było celem przeglądu, i zidentyfikowali 12 badań.

Kluczowe informacje
Zasłanianie oka prawdopodobnie nie przyspiesza gojenia oraz może nie mieć żadnego istotnego wpływu na łagodzenie bólu. Żadne z badań nie opisywało wpływu zasłonek na większe otarcia rogówki.

Co badano w przeglądzie?
Rogówka to zewnętrzna, przezroczysta część gałki ocznej. Otarcie rogówki może powstać w wyniku zadrapań lub powierzchownego uszkodzenia rogówki. Są to powszechne problemy, które mogą być bardzo bolesne. Powszechnie stosowaną opcją leczenia jest zastosowanie zasłonki na oko, co może mieć wpływ na długość procesu gojenia. Ponadto może także zapewnić ulgę w bólu.

Jakie są głównie wyniki przeglądu?
Autorzy przeglądu odnaleźli 12 stosownych badań. 6 pochodziło z Ameryki Północnej, 5 z Europy oraz 1 z Ameryki Południowej (Brazylii). Badania te porównywały stosowanie zasłonek do ich niestosowania.

U osób otrzymujących zasłonki występuje mniejsze prawdopodobieństwo wygojenia się otarcia rogówki po 24 godzinach w porównaniu do osób, które nie stosują zasłonki (dane naukowe dostarczające mało pewności co do rzetelności tego oszacowania). Stosowanie zasłonki prawdopodobne wiąże się z niewielką różnicą lub nie wiąże się z jakąkolwiek różnicą co do liczby osób, u których rana wygoiła się po 48 i 72 godzinach (dane naukowe dostarczające umiarkowanej pewności co do rzetelności tego oszacowania).

Otarcie rogówki u osób stosujących zasłonki prawdopodobnie goi się nieco dłuższej niż u osób nie stosujących ich, jednak różnica ta była niewielka i prawdopodobnie nieistotna (dane naukowe dostarczające umiarkowanej pewności co do rzetelności tego oszacowania).

Stosowanie zasłonek może powodować więcej bólu w okresie 24 godzin (dane naukowe dostarczające mało pewności co do rzetelności tego oszacowania). Jednakże, zakres obserwowanych efektów zasłonek wskazuje, że mogą zarówno prowadzić do zwiększenia bólu jak i do jego zmniejszenia.

U osób z otarciem rogówki często występuje wrażliwość na światło, łzawienie oczu, uczucie ciała obcego oraz niewyraźne widzenie. Istnieje mało danych naukowych, które wskazywały na różnice w występowaniu powyższych objawów u osób stosujących zasłonki i niestosujących zasłonek.

Dane dotyczące jakości życia, ostrości widzenia oraz działań niepożądanych były ograniczone.

Jak aktualny jest przegląd?
Autorzy przeglądu poszukiwali badań opublikowanych do 9 maja 2016 r.

Uwagi do tłumaczenia

Tłumaczenie Magdalena Koperny, Redakcja Joanna Zając

Резюме на простом языке

Глазные защитные повязки при эрозиях роговицы

Какова цель этого обзора?
Целью этого Кокрейновского обзора было выяснить, влияет ли защитная повязка при эрозии роговицы на скорость заживления эрозии и облегчение боли. Исследователи Кокрейн нашли и проанализировали все соответствующие (имеющие отношение) исследования, чтобы ответить на этот вопрос, и нашли 12 исследований.

Ключевые выводы
Глазная защитная повязка, вероятно, не ускоряет заживление эрозии роговицы и не оказывает существенного влияния на боль. Ни в одном из исследований не представлена информация о влиянии глазной защитной повязки на обширные эрозии.

Что изучалось в данном обзоре?
Роговица - это наружная прозрачная оболочка глаза. Роговичные эрозии могут возникнуть в результате царапины или поверхностного повреждения роговицы. Это распространенная проблема и она может быть очень болезненной. Обычным вариантом лечения является наложение защитной повязки поверх глазного яблока. Предполагается, что это может оказать влияние на скорость заживления эрозии. Также это может обеспечить облегчение боли.

Каковы основные результаты этого обзора?
Авторы обзора нашли 12 соответствующих исследований. 6 исследований были из Северной Америки, 5 - из Европы и одно из Южной Америки (Бразилии). В этих исследованиях сравнивали эффекты использования глазной защитной повязки с отсутствием повязки.

Люди, которым накладывают глазную защитную повязку, имеют меньше шансов заживления эрозии роговицы через 24 часа по сравнению с теми, кому этого не делали (низкая определенность доказательств). Использование главной повязки, вероятно, не оказывает влияния на число люде, у которых эрозия роговицы заживает за 48 или 72 часа (умеренная определенность доказательств).

Эрозии роговицы у лиц, получающих глазные защитные повязки, вероятно, заживают немного дольше, чем у лиц без повязки, но эта разница невелика и, вероятно, несущественна (умеренная определенность доказательств).

Использование глазных повязок может привести к большей интенсивности боли в течение первых 24 часов (низкая определенность доказательств). Однако, авторы отметили диапазон времени, в котором использование глазной защитной повязки может приводить и к увеличению, и к уменьшению интенсивности боли.

Люди с эрозией роговицы часто сильно восприимчивы к свету, имеют слезотечение, чувство инородного тела и затуманенное зрение. Мало доказательств, чтобы предположить, что может быть какое-либо различие в этих симптомах у людей с повязкой или без неё.

Данные по качеству жизни, остроте зрения и побочным эффектам были ограничены.

Насколько актуален этот обзор?
Авторы этого обзора провели поиск исследований, которые были опубликованы по 9 мая 2016 года.

Заметки по переводу

Перевод: Арсланов Глеб Маратович. Редактирование: Зиганшина Лилия Евгеньевна. Координация проекта по переводу на русский язык: Cochrane Russia - Кокрейн Россия (филиал Северного Кокрейновского Центра на базе Казанского федерального университета). По вопросам, связанным с этим переводом, пожалуйста, обращайтесь к нам по адресу: cochrane.russia.kpfu@gmail.com; cochranerussia@kpfu.ru

Summary of findings(Explanation)

Summary of findings for the main comparison. Patching versus no patching for corneal abrasion
  1. 1Adverse effects examined included persistent symptoms (such as photophobia, lacrimation, foreign body sensation, blurred vision), corneal ulceration with subsequent development of a hypopyon and repeat presentations for recurrent erosions.

    2Downgraded one level for risk of bias as the studies were largely at high or unclear risk of bias and downgraded one level for imprecision as the confidence intervals were compatible with less healing in the patch group or no effect.

    3 Downgraded one level for risk of bias as the studies were largely at high or unclear risk of bias. Not downgraded for imprecision as confidence intervals reasonably narrow around 1 (no effect) and therefore we judged that important differences between the two groups are unlikely.

    4 Downgraded one level for risk of bias as the studies were largely at high or unclear risk of bias and downgraded one level for imprecision as the confidence intervals are wide and compatible with less pain or more pain in the patch group compared with the no-patch group.

    5 Downgraded one level for risk of bias as the studies were largely at high or unclear risk of bias and downgraded one level for imprecision as the confidence intervals are wide and compatible with fewer adverse events or more adverse events in the patch group compared with the no-patch group.

Patching compared to no patching for corneal abrasion
Patient or population: participants with corneal abrasion
Settings: Hospitals
Intervention: Patching
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Certainty of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
People with no patches People with patches
Complete healing after 24 hours

620 per 1000

(273 to 1000)

552 per 1000
(490 to 620)
RR 0.89
(0.79 to 1)
531
(7 studies)
⊕⊕⊝⊝
low2
-
Complete healing after 48 hours

856 per 1000

(813 to 1000)

831 per 1000
(779 to 891)
RR 0.97
(0.91 to 1.02)
497
(6 studies)
⊕⊕⊕⊝
moderate3
-
Complete healing after 72 hours

914 per 1000

(809 to 1000)

923 per 1000

(797 to 1000)

RR 1.01 (0.97 to 1.05)430 (4 studies)⊕⊕⊕⊝
moderate3
-
Days to complete healingThe mean number of days in the no-patch group ranged from 1 to 2.3 daysThe mean number of days to complete healing in the patch group was 0.14 longer (0 to 0.27 days longer) MD 0.14 (0.00 to 0.27)

642

(6 studies)

⊕⊕⊕⊝
moderate3
-
Pain at 24 hours 157 per 1000

237 per 1000

(135 to 416)

RR 1.51

(0.86 to 2.65)

193

(2 studies)

⊕⊕⊝⊝
low4
Most studies reported pain on a visual analogue scale (VAS). It was not possible to pool the data because they were skewed. In general, similar pain ratings were seen between patch and no-patch groups.
Quality of lifeSee comments----Activities of daily living (ADL) was assessed in one study of children. There was little difference in ADL with the exception of walking which was reported to be more difficult with a patch on. VAS 1.7cm (SD 2.1) versus 0.3cm (SD 0.7)
Adverse effects1

9 per 1000

(0 to 50)

26 per 1000
(8 to 86)
RR 3.24
(0.87 to 12.05)
660
(8 studies)
⊕⊕⊝⊝
low5
-
Change in visual acuity one week from initial presentation---46
(1 study)
Not graded as no estimate of effectMean best-corrected Snellen acuity at baseline was 0.9 (SD 0.2) in the patch group and 1.1 (SD 0.3) in the no-patch group. At one week, mean acuity was 1.1 (SD 0.3) in the patch group and 1.1 (SD 0.2) in the no-patch group.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio; SD: standard deviation; MD: Mean deviation
GRADE Working Group grades of evidence
High certainty: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low certainty: We are very uncertain about the estimate.

Background

Description of the condition

The cornea is the transparent outer layer of the eye. Besides its crucial role in refraction, the cornea also functions as a protective barrier against infection and trauma. It comprises five layers, with the corneal epithelium located most anteriorly. Corneal abrasions are superficial defects involving the corneal epithelium that mostly arise from mechanical injuries. In fact, corneal abrasions are among the most frequently encountered ocular conditions in eye emergency departments (Lubeck 1988; Vaughan 1995). A published audit ranked corneal abrasions as the eighth most common condition diagnosed in a series of 274 consecutive cases in the accident and emergency department of a regional eye hospital in Hong Kong (Lai 2003). In the same audit, external eye foreign bodies were the second most commonly treated condition. Corneal foreign bodies are often associated with corneal abrasions as an epithelial defect remains on removal. Eye injuries lead to significant morbidity and lost productivity. A major United States automotive corporation found an annual incidence of 15 eye injuries per 1000 employees, with a third of workers unable to resume normal duties for at least one day (Wong 1998).

Corneal abrasions are also a common presenting condition in general accident and emergency departments. An audit of US emergency departments (McGwin 2005) found that contusions or abrasions (44.4%) were the most common cause of ophthalmic-based presentations in this setting, with foreign bodies (30.8%) identified as the next most common presentation. If one were to extrapolate the findings of a British audit of general emergency departments (Edwards 1987) and consider 3% of new admissions to be eye-related trauma, then the total number of new cases of eye injuries would be approximately 420,000 per year in England alone (DOH 2004). This number is based on the total number of presentations to accident and emergency departments in one year. Two American audits of US emergency departments have estimated the incidence of eye-related presentations as between 315 per 100,000 (McGwin 2005) and 447.1 per 100,000 population (Nash 1998). In England, there are a number of emergency departments specifically for ophthalmology presentations where the proportion of corneal abrasions and foreign body injuries is much higher (Lai 2003). A recent audit of an eye emergency service in Ireland reported that corneal abrasions and corneal foreign bodies were the most common causes for traumatic eye presentations, making up 28.4% and 26.9% of traumatic ophthalmic presentations (Vartsakis 2014). In addition general practitioners and optometrists would likely deal with a significant proportion of corneal abrasions and therefore one can infer that this estimate of the total yearly number of corneal abrasions is conservative at best and underestimates the true incidence of this condition.

Description of the intervention

The management of a corneal abrasion with a patch (some form of occlusion of the affected eye) and topical antibiotics was the recommended therapy for corneal abrasion in many references (Catalano 1992; Cullom 1994; Khaw 2004; Parrish 1988; Pavan-Langston 1991; Webster 1987).

Why it is important to do this review

The practice of patching corneal abrasions has been questioned, with a number of trials suggesting no benefit (Hulbert 1991; Kirkpatrick 1993). Many of these trials had small numbers of participants and therefore lacked the statistical power to demonstrate any differences. A systematic review on the use of patching was completed in 1998 (Flynn 1998). This was reviewed by the Centre for Reviews and Dissemination which identified several areas for improvement: only a single author was involved in assessing the validity and quality of studies; the methods for selecting studies were not stated; and eligible studies were restricted to those published in English (CRD 2006). A second review, published in a Japanese journal, was restricted to trials published in the English language with searches restricted to only one database (Yamada 2001). Furthermore, new randomised controlled trials on the topic have been published since these reviews were prepared.

Objectives

The objective of this review was to assess the effects of patching for corneal abrasion on healing and pain relief.

Methods

Criteria for considering studies for this review

Types of studies

This review included randomised and quasi-randomised controlled trials.

Types of participants

Participants in the trials were people of all ages with recent onset (less than 48 hours) of corneal abrasion due to mechanical injury, foreign body removal or contact lens use, as diagnosed by fluorescein or slit-lamp examination. We excluded trials of participants with corneal abrasions due to infection, peripheral corneal degeneration or chemical injury (these conditions can result in epithelial loss similar to primary abrasions due to mechanical injuries, but the pattern, progression, treatment and prognosis differs markedly).

Types of interventions

We examined the following comparisons:

  • eye patching versus no eye patching;

  • eye patching plus topical antibiotics versus topical antibiotics alone

Treatment may have included cycloplegics or analgesics, or both. Eye patching should have been for at least 24 hours of continuous intended use.

We considered the following methods of eye patching:

  • cotton wool covered with a net held with tape over a closed eye;

  • pressure patching with either double eye pad or bulk gauze (enough to exert pressure) on closed eye held with either bandage or plaster. Tape or plaster placed onto the skin of the eyelids to prevent eye opening;

  • any other form of occlusion of the affected eye adopted by the trialists.

Types of outcome measures

Primary outcomes

The primary outcomes for this review were:

  • proportion with complete healing after 24, 48 and 72 hours;

  • mean days to complete healing;

  • rate (proportion/length/area of epithelial defect recovered per unit of time).

Healing should have been ascertained using fluorescein staining or slit-lamp examination.

Secondary outcomes

The secondary outcomes for this review were:

  • pain assessment using 0 to 100 score, visual analogue scale (VAS) or any form of pain measurement adopted by the trialists;

  • use of analgesia;

  • quality-of-life measures;

  • assessment of activities of daily living (ADL);

  • insomnia assessments;

  • duration of medical leave;

  • other symptoms, for example photophobia, lacrimation, foreign body sensation and blurred vision;

  • measure of compliance to treatment;

  • use of topical cycloplegics;

  • visual acuity measured using a logMAR acuity chart.

Adverse effects (severe, minor)

We examined the following adverse effects:

  • infection or inflammation after commencement of trial as diagnosed by trialists;

  • recurrent corneal abrasions as diagnosed by repeated episodes of corneal abrasion after complete healing had occurred;

  • any other untoward events.

Follow-up

The minimum length of follow-up required was 24 hours after enrolment. Follow-up may have been repeated every 24 hours until complete healing of abrasion had been noted.

Search methods for identification of studies

Electronic searches

We searched CENTRAL (which contains the Cochrane Eyes and Vision Trials Register) (2016, Issue 4), Ovid MEDLINE, Ovid MEDLINE In-Process and Other Non-Indexed Citations, Ovid MEDLINE Daily, Ovid OLDMEDLINE (January 1946 to May 2016), EMBASE (January 1980 to May 2016), Latin American and Caribbean Health Sciences Literature Database (LILACS) (January 1982 to May 2016), System for Information on Grey Literature in Europe (OpenGrey) (January 1995 to May 2016), the ISRCTN registry (www.isrctn.com/editAdvancedSearch), ClinicalTrials.gov (www.clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/search/en). We did not use any date or language restrictions in the electronic searches for trials. We last searched the electronic databases on 9 May 2016.

See appendices for details of search strategies for CENTRAL (Appendix 1), MEDLINE (Appendix 2), EMBASE (Appendix 3), LILACS (Appendix 4), OpenGrey (Appendix 5), ISRCTN (Appendix 6), ClinicalTrials.gov (Appendix 7) and the ICTRP (Appendix 8).

Searching other resources

Additional handsearching focused on reference lists and abstracts/proceedings of scientific meetings held on the subject. In particular, we searched the proceedings of the Association for Research in Vision and Ophthalmology (ARVO). For the abstracts from 2002, we performed electronic searches using the key words: corneal abrasion; patch*; occlusion; abrasion; trauma; foreign bod*. For abstracts from 1993 to 2001 the ARVO proceedings index was used searching through sections on 'Cornea' and checking subsections of 'Wound healing', 'Abrasion', 'Epitheli*'.

We contacted the authors of relevant published studies to help identify unpublished data. In March 2004 we contacted companies and pharmaceutical firms that produce eye patches and topical antibiotics (including GlaxoSmithKline, Pfizer, Alcon, Troge, Sigma, Novartis, CibaVision) for unpublished data.

Data collection and analysis

Selection of studies

Two review authors (CHLL, BXL) independently screened the updated search results. When disagreements arose, the review authors assessed the studies separately once more and held a discussion to decide whether these studies should be included. Complete versions of all included studies were then independently reviewed. In instances where articles were not published in English, we obtained an accurate translation.

Data extraction and management

Two review authors (CHLL, BXL) independently extracted data from included studies onto a data collection template. We then compiled the individually extracted results and discussed any discrepancies in the data. We checked decisions we made against published study data. Data collected included study characteristics, interventions, follow-up and outcome data.

Assessment of risk of bias in included studies

Two review authors (CHLL, BXL) independently assessed the risk of bias of each included study in accordance with guidelines laid out in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

We evaluated the following domains.

  • Randomisation: how were participants allocated to either the intervention or control groups? How was the allocation sequence generation made unpredictable to study personnel?

  • Selection bias (allocation concealment): was the allocation of participants to either the treatment or control arm of the study concealed from study personnel involved in the initial assessment of participants? If so, how were allocations performed?

  • Performance bias: were study personnel masked to the interventions received by each participant to preclude any differences in the clinical management received?

  • Detection bias: were personnel who were involved in the follow-up assessments of study participants adequately masked to the allocations of participants to the respective treatment arms?

  • Attrition bias: how did the rates of follow-up compare between the treatment and control groups? Was the analysis based on an ‘intention-to-treat’ principle?

  • Reporting bias: have all results with both demonstrable and non-demonstrable differences been reported in the manuscript?

Each of these domains was graded by each author as either 'high risk', 'low risk', or 'unclear risk' of bias. Thereafter, we compared the risk of bias tables and resolved any differences in assessments through discussion.

Measures of treatment effect

We used the risk ratio (RR) for dichotomous outcomes (proportion healed, with pain, adverse effects and analgesic use) and the mean difference (MD) for continuous outcomes (days to complete healing). We used the standardised mean difference (SMD) for analysing mean reduction in pain scores as these were measured on different scales. For continuous variables we checked the summary figures for skewness using the method described by Altman 1996.

Unit of analysis issues

All included studies were parallel group studies (i.e. people were randomly allocated to treatment). It was not clearly described how eyes were dealt with. In general corneal abrasions may be expected to occur predominantly in one eye and we have assumed that one eye per person only was included. Only one trial reported two bilateral cases (Arbour 1997). In these cases one eye was patched and the other not patched. It was difficult to distinguish this information from the complete dataset and was therefore ignored this in the analysis.

Dealing with missing data

Ideally we would have conducted an intention-to-treat (ITT) analysis using imputed data if computed by the trial investigators using an appropriate method.

ITT data were not available so an available case analysis was performed. This assumes that data are missing at random. We collected data from each included trial on the number of participants excluded or lost to follow-up and by treatment group, if reported.

Assessment of heterogeneity

We examined the overall characteristics of the studies, in particular the type of participants and types of interventions, to assess the extent to which the studies were similar enough to make pooling study results sensible. We examined the forest plots of study results to see how consistent the results of the studies were, in particular looking at the size and direction of effects. We calculated I2 which is the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error (chance) (Higgins 2002).

Assessment of reporting biases

We used the risk of bias assessment tool to look for selective or incomplete reporting. (See Assessment of risk of bias in included studies).

There were insufficient studies included in any meta-analysis to formally assess publication bias. In future updates of this review, if there are 10 trials or more included in a meta-analysis, we will construct funnel plots and consider tests for asymmetry for assessment of publication bias, according to Chapter 8 of The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Data synthesis

Data analysis for this update was performed using Review Manager (RevMan) 5.3 (RevMan 2014). We pooled data using a random-effects model, unless there were three or fewer trials, in which case we used a fixed-effect model.

Subgroup analysis and investigation of heterogeneity

We considered two subgroup analyses.

Size of corneal abrasions

We planned to compare effects in people with large abrasions compared to people with small abrasions as we considered that the effects of patching may be different in these two groups. Large abrasions were defined as lesions measuring more than 10 mm2. This subgroup was defined at the protocol stage.

Abrasions caused by the removal of foreign bodies

We planned to compare effects in people with abrasions caused by the removal of foreign bodies compared to people with abrasions caused by other means as we considered that the effects of patching may be different in these two groups. This subgroup was not considered in our original protocol but was included in the first published version of this review (Turner 2006).

Sensitivity analysis

We repeated the analyses excluding studies at higher risk of bias, that is, quasi-randomised studies and studies that were not masked.

Summary of findings table

We prepared a 'Summary of findings' table presenting relative and absolute risks. We graded the overall certainty of the evidence for each outcome using the GRADE classification( www.gradeworkinggroup.org ). We included the following outcomes (see Summary of findings for the main comparison).

  • Complete healing after 24 hours

  • Complete healing after 48 hours

  • Complete healing after 72 hours

  • Days to complete healing

  • Pain at 24 hours

  • Quality of life

  • Adverse effects

  • Change in visual acuity one week from initial presentation

Results

Description of studies

Results of the search

The initial search strategy resulted in a total of 74 reports of trials. We screened these reports and retrieved 39 full-text articles for further assessment. 12 papers described randomised controlled trials (RCTs). One further paper was a letter that contained enough information about a new trial to include its results in the review (Rao 1994). One study was subsequently excluded when translated since the two groups in the trial had different types of patches applied and there was no control group without patching (Gregersen 1991). Another randomised controlled trial was excluded as the characteristics of the participants did not fit our selection criteria (Kurt 2003). 11 papers were included in the review. The remaining 25 papers were either letters or comments made about trials that had been conducted.

There were no current or prospective trials listed on the UK and US online databases. The pharmaceutical firms were not helpful in providing any unpublished trial information. A number of authors were able to provide further information regarding their trials (Arbour 1997; Kaiser 1995; Kirkpatrick 2003; Le Sage 2001).

An updated search in December 2007 yielded a further 75 reports of studies. The Trials Search Co-ordinator (now known as the CIS) reviewed these results and removed any references which were not relevant to the scope of the review. This search did not identify any references which met the inclusion criteria for the review.

An update search run in January 2014 identified a further 274 references. The Trials Search Co-ordinator removed 12 duplicates and screened the remaining 262 references, of which 240 were not relevant to the scope of the review. We reviewed the remaining 22 references and discarded 21 reports as not relevant. We obtained one full-text report (Menghini 2013) and have included this study in the review.

Update searches ran in May 2016 yielded a further 259 records (Figure 1). After 68 duplicates were removed the Cochrane Information Specialist (CIS) screened the remaining 191 records and removed 165 references which were not relevant to the scope of the review. We screened the remaining 26 references but none met the inclusion criteria for the review.

Figure 1.

Study flow diagram.

Included studies

We have summarised the characteristics of the 12 included studies below. Details can be found in the 'Characteristics of included studies' tables.

Setting and participants

The four earliest trials were conducted in the United Kingdom (UK) (Hulbert 1991; Jackson 1960; Kirkpatrick 1993; Rao 1994). Four further trials were conducted in the United States of America (USA) (Campanile 1997; Kaiser 1995; Michael 2002; Patterson 1996), two in Canada (Arbour 1997; Le Sage 2001), one in Brazil (Agostini 2004) and one in Switzerland (Menghini 2013). All of the participants had a recent, simple corneal abrasion. Three trials excluded participants with a corneal abrasion secondary to a corneal foreign body (Arbour 1997; Jackson 1960; Kirkpatrick 1993). Five trials included data on participants with corneal abrasions specifically related to removal of corneal foreign bodies (Agostini 2004; Hulbert 1991; Kaiser 1995; Le Sage 2001; Menghini 2013). There were a total number of 1080 participants with total post-randomisation exclusions of 183. One trial enrolled children (Michael 2002).

Interventions

Most trials had two treatment groups with participants randomised to receive a patch (a form of occlusion of the affected eye) or no patch. One trial had three intervention arms (Menghini 2013) and in this review we only considered participants in the patch and no-patch groups. All the trials included a form of concurrent medication used in both treatment groups, for example antibiotic or cycloplegic eye drops. There were often differences in the formulation and administration of these additional drops between the two groups. For instance, participants in the patched group may have received one dose of an ointment-based formulation in 24 hours, while the no-patch group may have been instructed to administer three or four doses of a topical solution.

Types of outcome measures
Main outcomes

The primary outcome measure in all included trials was a measure of corneal healing. Five trials measured the number of participants who had completely healed (no further fluorescein staining) on each day of follow-up (Hulbert 1991; Jackson 1960; Kirkpatrick 1993; Le Sage 2001; Patterson 1996). Two trials measured mean time to healing (Agostini 2004; Kaiser 1995). Two trials measured percentage of healing on each day of follow-up (Campanile 1997; Michael 2002). Five trials measured corneal abrasion dimension sizes at baseline and at each day of follow-up (Arbour 1997; Kirkpatrick 1993; Le Sage 2001; Menghini 2013; Rao 1994).

Other outcomes

Ten trials measured pain scores (Agostini 2004; Arbour 1997; Hulbert 1991; Kaiser 1995; Kirkpatrick 1993; Le Sage 2001; Menghini 2013; Michael 2002; Patterson 1996; Rao 1994). Some trials measured analgesia use, impact on quality of life, duration of medical leave, other associated symptoms and compliance to treatment. One trial examined change in visual acuity.

Adverse events

Four trials specifically mentioned short-term adverse events (Jackson 1960; Kaiser 1995; Menghini 2013; Michael 2002). Three trials reported long-term complications and follow-up two to seven months after the corneal abrasion (Arbour 1997; Kaiser 1995; Kirkpatrick 1993).

Economic measures

Economic measures were not evaluated in any of the included trials.

Excluded studies

See: Characteristics of excluded studies tables for details.

Risk of bias in included studies

Only one trial (Menghini 2013) received a 'low' risk of bias rating across all assessed components, while another (Michael 2002) received a similar rating in most domains (See Figure 2; Figure 3).

Figure 2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

All authors of included studies described elements of randomisation as part of their allocation process. However, we noted poor reporting of randomisation techniques across most studies. Only four studies were initially graded by both review authors who had judged risk of bias as having undergone ‘low-risk’ randomisation (Kirkpatrick 1993; Menghini 2013; Michael 2002; Patterson 1996). Kirkpatrick 1993, was assessed as having undergone an adequate randomisation process, following review of previous correspondence describing a coin-tossing randomisation technique. Menghini 2013 utilised a computer-based permuted-block randomisation technique to determine the allocation of participants across different treatment arms prior to the start of the study. This information was concealed from study personnel in numbered envelopes. Michael 2002 undertook a permuted-block randomisation process with allocations concealed within envelopes. These envelopes were pre-randomised within groups of four envelopes, which were numbered and subsequently handed out to participants in the designated sequence. Patterson 1996described a randomisation process involving the use of a computer-generated table.

Three studies described quasi-randomisation techniques (Agostini 2004; Jackson 1960; Le Sage 2001). In these, participants were allocated to treatment arms, either based on cases presenting to the department (Agostini 2004) or alternating days (Jackson 1960; Le Sage 2001), . We assessed these techniques as being at high risk of bias.  

Although the remaining studies alluded to randomisation techniques in their participant allocation processes, there was insufficient information to allow us to assess the adequacy and comprehensiveness of these randomisation processes with confidence (Arbour 1997; Campanile 1997; Hulbert 1991; Kaiser 1995; Rao 1994).

Blinding

It was inherently impossible to conceal the implemented treatment modality from study participants because of obvious differences between patching and not patching. This may have influenced behaviours undertaken outside the study setting (e.g. those who belonged to the no-patch group may have gone home and applied other treatments to their eye). However, in terms of detection bias, masking of assessors (although achievable in all trials) was only reported in five studies (Arbour 1997; Campanile 1997; Le Sage 2001; Menghini 2013, Michael 2002). In three of these studies, participants in the patched group were requested to remove their eye patch prior to their scheduled consultations (Le Sage 2001; Menghini 2013;Campanile 1997). Although clinical assessors in Arbour 1997 and Michael 2002 did not appear to be masked to the treatment allocation of participants, the assessment and reporting of digitised corneal images or grid template diagrams was performed by a reviewer who had no knowledge of the treatment allocation. The remaining studies were either graded as high risk (Agostini 2004; Hulbert 1991, Jackson 1960, Kaiser 1995 Kirkpatrick 1993) or unclear risk (Patterson 1996; Rao 1994) of detection bias. Most of these studies provided insufficient detail about the measures adopted to mask assessors to the respective treatment arms to allow us to assess their effectiveness with any confidence.

Incomplete outcome data

The reported percentage of participants lost to follow-up ranged from 0% to 38%. High dropout rates may impact upon the validity of results obtained and their analysis. Furthermore, many included studies involved small numbers of enrolled participants. In these instances, even a small number of participants lost to follow-up could alter baseline characteristics of the different treatment arms. Moreover, in many of these trials, comparisons were not reported between dropout rates amongst patched and non-patched groups, nor were reasons provided for the losses.

Equal distribution of attrited participants across different treatment arms may not necessarily be acceptable, as participants may have dropped out for different or opposing reasons. The lack of such information across many studies made it difficult to assess the potential impact of this aspect on the validity of the results obtained. Although the highest percentage of participants lost to follow-up was reported by Menghini 2013, they have argued that this would most likely not have impacted upon the significance of their findings, since all included participants demonstrated complete resolution of corneal abrasions earlier in the trial.

Trials also often reported exclusions of participants following randomisation due to certain exclusion criteria. For example, participants found to have residual foreign bodies or stains from a prior visit on their first follow-up visit were excluded from the study (Hulbert 1991). This may be a subjective assessment and could have resulted in the introduction of bias as assessors were not masked to the treatment group. The assessor may have noted very poor healing in the no-patch group for instance, and also happened to note some residual staining in the same participant, thereby finding grounds for exclusion. With the small numbers in this study, one or two exclusions of this nature could bias the results.

Although efforts to contact participants lost to follow-up was described in several trials, intention-to-treat (ITT) analysis was not performed in many of these studies. Even in Michael 2002, where ITT was mentioned, participants lost to follow-up were not included in their analyses as outcome data were not available and there was a post-randomisation exclusion for ineligibility. Therefore, in this instance, ITT referred to participants who were assigned to the 'patch' group but removed their eye-patch during the follow-up period. They were still included in the data analysis despite deviating from protocol. Intention-to-treat is important since it seeks to minimise biases arising from differences in known and unknown prognostic factors between groups. As none of the studies adequately addressed ITT, results obtained should be interpreted with caution, especially given the apparent drop-out rates.

Selective reporting

We identified a range of issues across several studies. Many of these relate to the amount of information and level of detail provided by the study authors. Such issues include insufficient information regarding how the trialists performed certain aspects of the study and processes. Additionally, four studies did not discuss other important information, such as a comparison of baseline demographic characteristics of participants between treatment arms (Campanile 1997; Hulbert 1991; Jackson 1960; Patterson 1996). Such information may potentially skew the final analysis of results obtained and impact upon the validity of the results.

Effects of interventions

See: Summary of findings for the main comparison Patching versus no patching for corneal abrasion

See also Summary of findings for the main comparison.

Primary outcomes

Complete healing after 24 hours

Seven trials reported data for the number of participants in each group who had completely healed on the first day of follow-up. Participants receiving a patch may be less likely to have complete healing after 24 hours than participants receiving no patch (RR 0.89, 95% CI 0.79 to 1.00; participants = 531; studies = 7; I2 = 0%) (see Analysis 1.1). However, upper limit of the 95% confidence intervals for this analysis included the point of no difference (RR = 1). We judged this to be low certainty evidence, downgrading for risk of bias and imprecision.

Complete healing after 48 hours

Six trials reported data on complete healing of corneal abrasions after 48 hours. A RR of 0.97 (95% CI 0.91 to 1.02; participants = 497; studies = 6; I2 = 0%) suggested a negligible difference between the patch and no-patch groups (see Analysis 1.2). We judged this to be moderate certainty evidence, downgrading for risk of bias.

Complete healing after 72 hours

Four trials reported healing after 72 hours. Most corneal abrasions had healed regardless of which treatment option was used (RR 1.01, 95% CI 0.97 to 1.05; participants = 430; studies = 4; I2 = 0%) (see Analysis 1.3). We judged this to be moderate certainty evidence, downgrading for risk of bias.

Days to complete healing

Six studies reported days to complete healing of corneal abrasions. Similar results were seen in the patch and no-patch groups (MD 0.14, 95% CI 0.00 to 0.27; participants = 642; studies = 6; I2 = 37%) (see Analysis 1.4). It is likely that this variable was skewed to a certain extent. However, using the rule of thumb proposed in Altman 1996, in general most results were not too skewed as the mean was more than twice the size of the standard deviation. We judged this to be moderate certainty evidence, downgrading for risk of bias.

Changes in corneal abrasion dimension size

Corneal abrasion size at initial presentation and follow-up was measured in six trials. A variety of techniques and comparisons were utilised in the trials to evaluate changes in corneal abrasion size at follow-up. Agostini 2004 reported mean residual corneal abrasion sizes at follow-up. The patch group (total number in patch group 27) was found to have a mean abrasion size of 0.07 mm2 (SD 0.19) on day one of follow-up compared to 0.10 mm2 (SD 0.20) in the no-patch group (total number in no-patch group = 27) (P value = 0.620). On day two of follow-up, the patch group had completely healed, whereas the mean abrasion size was 0.01 mm2 (SD 0.04) in the no-patch group (P value = 0.167). Arbour 1997 examined the linear and surface speed of re-epithelialisation. No marked differences were demonstrated in comparing the patch (total number in patch group = 17) and no-patch (total number in no-patch group = 16) groups. The authors also reported the percentage of erosions which closed at the first follow-up visit, which was 24% (6 participants) in the patch group compared to 27% (6 participants) in the no-patch group (P value = 0.80). Campanile 1997 calculated the percentage healed in 24 hours by comparing the measured defect on a grid sheet at presentation to that at follow-up. The authors reported the mean initial abrasion size of the patch group (total number in patch group = 31) as 6.1 mm2 and 5.6 mm2 for the no-patch group (total number in no-patch group = 35). This differed from a manual calculation of individual abrasion sizes provided. This revealed a mean of 6.0 mm2 (SD 7.3) for the patch group and 5.6 mm2 (SD 9.6) for the no-patch group, which contrasted with the reported mean abrasion size of the patch group of 6.1 mm2 (P value = 0.8117). In this study, a larger percentage healing score was reported in the no-patch group (97.091%) compared to the patch group (84.130%) (P value = 0.0283). Michael 2002 also reported the percentage of healing achieved at follow-up. In this study, 94.23% (SD 15.6) of healing was achieved by participants in the patch group (total number in patch group = 17) compared to 96.33% (SD 10.3) in the no-patch group (total number in no-patch group = 18) (95% CI -11.00 to 7.00). Menghini 2013 analysed the reduction in corneal abrasion area achieved at follow-up. A reduction of 3.4 mm2 (SD 3.3) was achieved in the patch group (total number in patch group = 17) and 3.5 mm2 (SD 3.1) in the no-patch group (total number in patch group = 26). All participants in this study displayed complete reduction of the corneal abrasion area by the second day of follow-up. Rao 1994 measured the mean and medial values of the minor and major axes at baseline and at follow-up. Statistical analysis did not demonstrate a difference in the dimensions of the corneal abrasion at presentation (P value = 0.58 for the minor axis and P value = 0.43 for the major axis) or on days one and two (P value > 0.4).

Secondary outcomes

Pain and discomfort

Pain and discomfort was recorded utilising a wide variety of methods (see Table 1). The means were not normally distributed for this outcome measure. Of the 10 trials that measured pain outcomes, only two reported results demonstrating evidence suggesting less pain in the no-patch group (Hulbert 1991; Kaiser 1995). Three other trials showed less pain in the no-patch groups (Kirkpatrick 1993; Le Sage 2001; Michael 2002) while five trials demonstrated less pain in the patch groups (Agostini 2004; Arbour 1997; Menghini 2013 Patterson 1996; Rao 1994). However, no demonstrable differences were identified. .

Table 1. Pain outcomes
  1. FB: corneal epithelial defects secondary to foreign body removal

    IQR: Interquartile range

    Mean baseline pain scores given in Michael 2002

    SD: standard deviation

    Traumatic: Traumatic corneal abrasions

    VAS: Visual analogue scale

    *Menghini 2013 provided P values for the comparison between 3 groups (patched, therapeutic contact lens, and ointment only)

`Study name Method of measuring pain or definition of pain (higher number on the scale indicating worse pain) Continuous variables  Dichotomous variables 
Patching No patching P valuePatchingNo patching P value
Number in group Median or mean score (SD or IQR) Number in group Median or mean score (SD or IQR) Number in group Number (%) with pain Number in group Number (%) with pain
Agostini 20040-10 pain scale275.59 (2.11)276 (1.86)0.455-----
Arbour 1997100 mm VAS (mean score)2515.4 (15.9)2223 (18.9)0.15-----
Arbour 1997100 mm VAS (mean of maximum score)2523.7 (22.8)2233.9 (27.3)0.18-----
Hulbert 1991Painful vs painless at 24h - - - - -1612 (75%)144 (29%)0.03
Kaiser 1995 (Traumatic)Pain score at presentation625.35 (2.07)584.91 (1.69)0.207-----

Kaiser 1995

(Traumatic)

Pain score Day 1622.84 (1.99)581.89 (1.03)0.003-----

Kaiser 1995

(Traumatic)

Pain score Day 2622.39 (1.44)581.79 (1.42)0.217-----

Kaiser 1995

(Traumatic)

Differences in pain score at 24 h622.51 (0.08)583.02 (0.66)< 0.05-----
Kaiser 1995 (FB)Pain score at presentation395.28 (1.19)425.07 (1.88)0.552-----
Kaiser 1995 (FB)Pain score Day 1392.53 (1.25)421.80 (0.99)0.009-----
Kaiser 1995 (FB)Pain score Day 2391.65 (0.98)421.58 (1.08)0.859-----
Kaiser 1995 (FB)Differences in pain score at 24 h622.75 (0.06)583.27 (0.89)< 0.05-----
Kirkpatrick 19930-100 VAS (differences in pain score 24 h)17-20.8 (20.3)20-27.6 (24.2)0.37-----

Le Sage 2001

 

Pain (% initial visit)-----8244 (54%)8138 (47%) -
Le Sage 2001Pain (% 24 h)-----8212 (15%)8111 (14%)-
Le Sage 2001Pain (% 48 h)-----820 (0%)812 (2%)-
Le Sage 200112 cmVAS826.0 (4.5-9.4)815.7 (3.9-7.7) -----
Michael 2002FACES pain scale at presentationUnclear 3-10 years old4.7Unclear 3-10 years old5.7------
Michael 2002VAS at presentationUnclear 11-17 years old5.6Unclear 11-17 years old6.6 ------
Menghini 2013Wong-Baker FACES Pain Rating Scale (presentation) -4.8 (1.7) -3.9 (1.5)0.243*-----
Menghini 2013Wong-Baker FACES Pain Rating Scale (3 h) -3.7 (2.4) -4.5 (3.3)0.694*-----
Menghini 2013Wong-Baker FACES Pain Rating Scale(4 h) -0.8 (1.6) -1.7 (2.7)0.227*-----
Menghini 2013Pain relief at 24 h184.1 (2.0)282.2 (3.0) 0.04*-----
Patterson 1996VAS pretreatment (mean)174.2165.2 ------
Patterson 1996VAS (24 h)171.11162.47 ------
Patterson 1996VAS (mean difference 24 h)173.09162.77> 0.50-----
Rao 1994VAS vertical (at presentation) -7.5 (6.35) -5.15 (5.15) ------
Rao 1994VAS vertical (D1) -2.5 (3.17) -1.4 (2.29) ------
Rao 1994VAS vertical (D2) -0.2 (0.74) -0.1 (0.64) ------

Six trials included information on the mean reduction in VAS between baseline and 24 hours' follow-up (Kaiser 1995; Kirkpatrick 1993, Le Sage 2001; Menghini 2013; Patterson 1996; Rao 1994).

Agostini 2004 reported the mean scores on a pain scale on days one and two post follow-up. The patch group (total number in patch group = 27) was found to have a mean pain score of 0.44 (SD 1.05) on day one of follow-up compared to 0.96 (SD 2.24) in the no-patch group (n = 27) (P value = 0.283). On day two of follow-up, the patch group had a mean pain score of 0.11 (SD 0.57), whereas the mean pain score was 0.14 (SD 0.45) in the no-patch group (P value = 0.794).

Arbour 1997 presented the mean VAS scores with a mean of 15.4 (SD 15.9) in the patch group (total number in patch group = 25) compared with the no-patch group (total number in no-patch group = 22) where mean VAS was 23.0 (SD 18.9) (P value = 0.15). Interestingly, a mean maximal VAS score was also examined in this study. The patch group was found to have a score of 23.7 (SD 22.8) compared to 33.9 (SD 27.3) (P value = 0.18) in the no-patch group. Despite the results appearing to favour the patch group, 48% identified the patch as their principal source of discomfort.

Hulbert 1991 reported that 75% of patients in the patch group (total number in patch group = 16) compared to 29% of patients in the no-patch group (total number in no-patch group = 14) experienced discomfort (RR 7.5 95% CI 1.17 to 55.6).

Kaiser 1995 found a greater reduction in mean pain scores in the patch group (total number in patch group = 39) compared to the no-patch group (total number in no-patch group = 42) of participants with corneal abrasions secondary to the removal of foreign bodies (MD 0.80, 95% CI 0.35 to 1.26). Similarly, there was also a greater reduction in the patch group (total number in patch group = 62) compared to the no-patch group (total number in no-patch group =58) in participants with traumatic corneal abrasions (MD 1.10 95% CI 0.71 to 1.48).

Kirkpatrick 2003 reported the mean decrease in pain scores of 20.8 (SD 20.3) in the patch group and 27.6 (SD 24.2) in the no-patch group (P value = 0.37) on a 100 point VAS.

Le Sage 2001 measured the level of discomfort on a VAS and reported a mean reduction of 4.8 (interquartile range (IQR) 2.2 to 7.0) in the patch group (total number in patch group = 82) and 3.3 (IQR 1.5 to 5.8) in the no-patch group (total number in patch group = 81) on day one of follow-up. The authors also analysed the number of participants who reported experiencing pain at presentation and follow-up. 54% of participants reported experiencing pain at presentation, while 15% reported pain as a symptom at 24 hours' follow-up and none reported experiencing pain at 48 hours' follow-up. In the no-patch group, 47% of participants experienced pain at presentation, while 14% and 2% experienced pain at 24 hours' and 48 hours' follow-up respectively.

Menghini 2013 demonstrated a difference in mean pain relief achieved by the patched group, with a mean score reduction of 4.1 (SD 2.0 95% CI 3.0 to 5.1; 19 participants), and the non-patched groups, with a mean score reduction of 2.2 (SD 3.0, 95% CI 0.9 to 3.4; 28 participants), 24 hours following therapy (P value = 0.04). However, this analysis included another intervention group involving the use of a soft therapeutic contact lens. No statistical differences between the different treatment arms were identified on further post-hoc analysis.

Michael 2002 used pain scores with a picture scale of faces depicting different levels of pain for children aged between 3 to 10 years, while a VAS was used for those aged between 11 to 17 years. Children in the patch group had a mean VAS score of 1.7 (SD 2.1) and 0.3 (SD 0.7) in the no-patch group, with a mean difference of 1.4 (95% CI 0.3 to 2.5; unclear number of participants).

Patterson 1996 presented 24-hour mean pain scores without including individual data or standard deviations. The patch group (total number in patch group = 17) had a mean pain score of 1.11 versus 2.47 in the no-patch group (total number in no-patch group = 16). The mean change in pain score was 3.09 in the patch group and 2.77 in the no-patch group (P value = 0.50).

Rao 1994 reported mean and median VAS at presentation and at follow-up on days one and two. The reduction in mean VAS from baseline for pain in the patch group (total number in patch group = 20) were 3.18 on day one and 2.43 on day two, while in the no-patch group (total number in no-patch group = 20), the reduction in mean VAS from baseline for pain was 0.44 on day one and 1.65 for day two in the no-patch group. A comparison between pain scores on days one (P value = 0.44) or two (P value = 0.89) did not demonstrate any statistical differences.

Two trials (Hulbert 1991, Le Sage 2001) reported presence or absence of pain at 24 hours (RR 1.51, 95% CI 0.86 to 2.65, 2 trials, 193 participants). (Analysis 1.5) We judged this to be low certainty evidence, downgrading for risk of bias and imprecision.

Use of analgesia

The use of analgesia was reported in several included trials. However, comparison of analgesic use between the patch and no-patch groups was only performed in five trials (Arbour 1997; Le Sage 2001; Menghini 2013; Michael 2002; Rao 1994). Arbour 1997 examined analgesia use based on the categories 'mild', 'strong' and 'the use of any form of analgesia'. 36% of participants in the patch group (total number in patch group = 25) used analgesia compared to 41% in the non-patch group (total number in no-patch group= 22) (P value = 0.73). Comparable rates of strong analgesia use (52% and 54%; P value = 0.86) was reported. Similar proportions of participants were also found to have used any form of analgesia (84% in the patch group compared to 82% in the non-patch group, P value = 0.84). This was in contrast with reported rates of analgesia use by Le Sage 2001 which was approximately 5% in the patch and non-patch groups. Menghini 2013 reported rates of analgesia use in the patch (total number in patch group = 18) and non-patch groups (total number in no-patch group = 28) as 14% and 25% respectively (RR 0.95, 95% CI 0.69 to 1.32, 156 participants). Michael 2002 explored the doses of pain medications taken, with similar mean scores of 1.4 between the patch (SD 1.2) and non-patch groups (SD 1.5) (95% CI -1.0 to 0.95). Rao 1994 found no differences between the number of participants taking paracetamol (P value > 0.2) and the number of tablets taken in both groups (P value > 0.8) (total number in patch group = 20 , total number in no-patch group = 20).

Impact on quality of life

The impact of the intervention (eye patching) upon quality of life was examined and reported in two studies. Arbour 1997 found comparable rates of reported insomnia between the patch and no-patch groups (P value = 0.98, 47 participants). Michael 2002 (participants were children) measured 10 aspects of daily living on a VAS. These included: dressing, feeding, walking, running, going to the bathroom, play, impact on resting and sleeping duration, any trips or falls, bumping into things and excessive crying. The only statistical difference between the patch (total number in patch group = 17) and no-patch groups (total number in no-patch group =18) was found with walking (95% CI for difference in means 0.3 to 2.5).

Duration of medical leave

Menghini 2013 measured the duration of medical leave taken in the patch (total number in patch group = 18) and non-patch groups (total number in no-patch group = 28). The patch group took 2.0 days (SD 1.6) of medical leave compared to 1.9 days (SD 2.2) in the non-patch group. A comparison of these groups and another intervention arm involving the application of a therapeutic contact lens found no statistical differences between these three groups (P value = 0.553). Michael 2002 examined the number of participants missing at least a day of school. 47%of participants in the patch group (total number in patch group = 17) were found to have taken at least one day off school compared to 33% in the no-patch group (total number in no-patch group =18) (95% CI -0.18 to 0.46).

Reported symptoms

Frequently reported symptoms across trials include photophobia, lacrimation, foreign body sensation, blurred vision, irritation and eye redness (Table 2). The latest reported time point was included for each of the trials to reflect the persistence of symptoms despite therapy. Pain and discomfort were reported as a separate outcome measure.

Table 2. Adverse effects
  1. *Arbour 1997: 7 out of 25 patients (28%) in the patch group and 1 out of 22 patients (4.5%) in the non-patch group complained of persistent symptoms in the affected eye, including pain,foreign body sensation, photophobia, and tearing. However, no further breakdown of these symptoms were provided.

    **Both Kaiser 1995 and Le Sage 2001 have provided the number of participants experiencing symptoms at presentation, day 1 and day 2. Number of participants experiencing symptoms at day 2 have been included in this table

    ***Jackson 1960: 1 patient experienced a recurrent corneal abrasion at 4 weeks. No information was provided regarding the treatment arm which this patient belonged to.

Adverse effects Patched participants experiencing symptoms Total number of patched participants Non-patched participants experiencing symptom Total number of non-patched participants
Photophobia    
Agostini 20047271327
Arbour 1997*----
Kaiser 1995 (Traumatic corneal abrasions)**462258
Kaiser 1995 (Corneal foreign body)**139142
Le Sage 2001**582581
Total1721021208
     
Lacrimation    
Agostini 20041027827
Arbour 1997*----
Kaiser 1995 (Traumatic corneal abrasions)**8621058
Kaiser 1995 (Corneal foreign body)**839742
Total2612825127
     
Foreign body sensation    
Agostini 20047271327
Arbour 1997*----
Kaiser 1995 (Traumatic corneal abrasions)**1662858
Kaiser 1995 (Corneal foreign body)**539542
Le Sage 2001**282481
Total3021030208
     
Blurred Vision    
Agostini 2004727827
Arbour 1997*----
Kaiser 1995 (Traumatic corneal abrasions)**262358
Kaiser 1995 (Corneal foreign body)**239042
Total1112811127
     
Insomnia    
Arbour 1997925822
     
Dendritic Ulcer    
Kirkpatrick 1993020124
     
Hypopyon    
Jackson 19601910104
     
Recurrent Corneal Erosion    
Jackson 1960***----
Kirkpatrick 1993020124
     
Discomfort    
Hulbert 19911216414
Kirkpatrick 1993420024
Total1636438
     
Irritation    
Le Sage 2001e**1082781

Agostini 2004 found that 25% of participants in the patch group (total number in patch group = 27) reported experiencing photophobia compared to 45% in the no-patch group (total number in no-patch group = 27) (P value = 0.091). 37% of participants in the patch group experienced lacrimation compared to 29% in the no-patch group (P value = 0.563). 25% of participants in the patch group and 48% of participants in the no-patch group reported experiencing foreign body sensations (P value = 0.761). Blurring of vision was reported by 25% of participants in the patch group and 29% in the no-patch group (P value = 0.761).

Arbour 1997 did not examine the presence of these symptoms during the initial follow-up, but noted that 28% of participants (total number in patch group = 25) in the patch group compared to 4.5% (total number in no-patch group = 22) in the no-patch group reported experiencing residual symptoms at a six-month follow-up requiring further consultation.Kaiser 1995 examined these symptoms in two specific cohorts of participants, those that had experienced a "traumatic corneal abrasion" and others that had had corneal foreign bodies removed. Although the study authors provided values at various time points, we chose to include reported rates at day one of follow-up as this was the most commonly available time point across the different studies. In the 'traumatic corneal abrasions' cohort of participants, 24% of participants in the patch group (total number in patch group = 62) compared to 19% of participants in the no-patch group (total number in no-patch group = 58) reported experiencing photophobia on day one (P value = 0.515). 61% of participants in the patch group experienced lacrimation compared to 57% of participants in the no-patch group (P value = 0.711). 45% of participants in the patch group reported experiencing a foreign body sensation compared to 36% of participants in the no-patch group (P value = 0.356). 40% of participants in the patch group experienced blurred vision compared to 17% in the no-patch group (P value = 0.009). In the cohort of participants who had had corneal foreign bodies removed, 15% of participants in the patch group (total number in patch group = 39) compared to 12% of participants in the no-patch group (total number in no-patch group = 42) reported experiencing photophobia on day one (P value = 0.751). 43% of participants in the patch group experienced lacrimation compared to 59% of participants in the no-patch group (P value = 0.184). 36% of participants in the patch group reported experiencing a foreign body sensation compared to 31% of participants in the no-patch group (P value = 0.647). 23% of participants in the patch group experienced blurred vision compared to 14% in the no-patch group (P value = 0.395).

Le Sage 2001 also examined symptoms at various time points. In this study, 15% of participants in the patch group (total number in patch group = 82) compared to 17% in the non patch group (total number in no-patch group = 81) reported photophobia, while 21% of participants in the patch group experienced foreign body sensation compared to 19% in the non-patch group at 24 hours following initial evaluation. This was distinguished from "local irritation", which 39% of participants in the patch group reported experiencing at 24 hours compared to 49% of participants in the non-patch group. Interestingly, Le Sage 2001 also examined "eye redness" as a symptom. Following discussions between authors of this review, we concluded that this was more appropriately identified as a sign and was excluded from this compilation.

Pooling the results of these studies did not suggest any difference between patch and no-patch groups although there was uncertainty with wide confidence intervals.

Photophobia: RR 0.99, 95% CI 0.48 to 2.07; participants = 255; studies = 3; I2 = 0% (Analysis 1.7)

Lacrimation: RR 1.04, 95% CI 0.64 to 1.68; participants = 255; studies = 3; I2 = 0% (Analysis 1.8)

Foreign body sensation: RR 0.98, 95% CI 0.63 to 1.55; participants = 418; studies = 4; I2 = 49% (Analysis 1.9)

Blurred vision: RR 0.99, 95% CI 0.48 to 2.07; participants = 255; studies = 3; I2 = 0% (Analysis 1.10)

Patient compliance

Compliance to treatment was examined in three of the included studies (Kaiser 1995; Kirkpatrick 2003; Michael 2002). Thirty-three out of a total of 153 participants (22%) in the patch groups of these three trials were found not to have their eye patches during follow-up. Of these, 15 participants reported that the eye pads fell off, 17 participants described discomfort associated with the eye patches and one participant removed the patch "to sleep".

Use of mydriatic drops

Mydriatic agents were used routinely in eight trials (Agostini 2004; Arbour 1997; Campanile 1997;Kaiser 1995; Kirkpatrick 1993 Le Sage 2001; Michael 2002; Rao 1994), while their use by Jackson 1960 and Le Sage 2001 was left up to the discretion of the treating physician. Agents used included: cyclopentolate hydrochloride, homatropine, atropine and phenylephrine. No comparisons were performed between participants in any of the included trials.

Visual outcomes

Visual outcomes were measured in two trials (Menghini 2013; Michael 2002), but only reported in one included trial. Menghini 2013 measured the best-corrected Snellen visual acuity (BCVA) at baseline and at one-week follow-up. The authors reported that the BCVA Snellen at baseline was 0.9 (SD 0.2) in the patch group and 1.1 (SD 0.3) in the no-patch group. Follow-up assessment found an average of 1.1 (SD 0.3) in the patch group compared to 1.1 (SD 0.2) in the no-patch group.

Adverse effects

Several trials examined participants longitudinally for adverse effects, with follow-up periods ranging from one week to 12 months. Adverse effects reported include persistent symptoms, recurrent corneal erosions, corneal ulceration with the development of hypopyon and conjunctivitis.

Arbour 1997 attempted to perform a six-month follow-up on all included participants. They were able to contact 17 participants in the patch group (total number in patch group= 25) and 11 participants in the non-patch group (total number in no-patch group = 22). Out of these participants, 7 participants in the patch group and 1 in the no-patch group complained of persistent symptoms, which included pain, foreign body sensation, photophobia and lacrimation. 4 participants in the patch group and 1 in the no-patch group required consultation for these residual symptoms.

Jackson 1960 reviewed participants over a two-month period following the conclusion of their study. During the initial healing process, one participant in the patch group developed corneal ulceration with hypopyon, requiring hospital admission, but was not included in the final cohort of participants analysed. Another 3 patients were reported to have developed complications after healing had occurred. One participant developed an acute conjunctivitis on the sixth day following the initial presentation while two other participants in the patch group (total number in patch group= 77) experienced recurrent corneal abrasions; one at four days, and another at five weeks from the first presentation.

Kaiser 1995 reported a follow-up period of between seven to 12 months. During this period, a participant from the patch group presented with a new corneal abrasion eight months after the initial treatment. This was included as part of our analysis. However, the authors were unable to confidently ascertain if this was a true recurrent erosion, or secondary to subsequent mechanical trauma. Two other participants returned with a diagnosis of viral conjunctivitis, however this was not thought to have been related to the previous corneal abrasion or its management.

Kirkpatrick 1993 described a follow-up period of 27 weeks, during which one participant who had initially presented with a corneal foreign body in the no-patch group (total number in no-patch group = 42) experienced two episodes of recurrent corneal erosions in the same eye. The presence of a dendritic ulcer was also identified in the affected eye of the patient in the no-patch group.

Participants in Menghini 2013 who were reviewed one week following their initial presentation did not experience any adverse effects (total number in patch group = 46). One participant experienced a second corneal foreign body in the same eye and was excluded from subsequent analysis.

Agostini 2004 (total number of evaluated participants = 54), Campanile 1997 (total number of evaluated participants = 64), Michael 2002 (total number of evaluated participants = 35) reported no complications in either group during the follow-up periods.

Neither the presence or absence of adverse effects were discussed by Hulbert 1991, Le Sage 2001 Patterson 1996 or Rao 1994,. Analysis of trials reporting adverse effects (see Analysis 1.11) suggested an increased risk of adverse events in the patch group but the confidence intervals were wide and also compatible with fewer adverse events in the patch group (RR 3.24, 95% CI 0.87 to 12.05, 660 participants)

Subgroup analysis
Size of corneal abrasions

The number of participants with large abrasions (more than 10 mm2) were limited in the trials. No trials specifically included or excluded participants with large abrasions so we were unable to do a subgroup analysis.

Two trials reported results for large abrasions separately (Campanile 1997; Kaiser 1995).

In Kaiser 1995 16 (13%) participants had large abrasions. The no-patch group showed longer mean time to healing with 4.20 days compared to the patched group with 3.45 days (P > 0.08). Results were not presented separately for the small abrasion group.

In Campanile 1997 13 people had large abrasions, 8 in the patch and 5 in the no-patch group. Six out of eight participants in the patch group were healed after 24 hours compared to two out of five participants in the no-patch group (RR 1.88, 95% CI 0.8, 5.90)). 51 people had small abrasions, 23 in the patch and 28 in the no-patch group. 15 out of 23 people in the patch group were healed after 24 hours compared to 25/28 in the no-patch group (RR 0.73, 95% CI 0.53, 1.01). These two risk ratios were very similar (test for interaction P value = 0.12) but the low numbers of participants in these analyses means that the power to detect a difference is low.

Abrasions caused by the removal of foreign bodies

We compared studies that included abrasions caused by removal of foreign bodies (Campanile 1997; Hulbert 1991; Le Sage 2001; Patterson 1996) to studies that excluded abrasions caused by removal of foreign bodies (Arbour 1997; Jackson 1960; Kirkpatrick 2003) for the primary outcome: complete healing after 24 hours. There was no evidence of any difference in effect between these two groups of trials (Analysis 1.12).

One trial reported foreign body abrasions separately (Kaiser 1995). Similar results were seen in people with foreign body abrasions compared to those with traumatic abrasions; participants healed faster and had less pain when they were not patched.

Sensitivity analysis

Table 3 shows the results of the sensitivity analysis excluding trials that were at high risk of bias, either because they were quasi-randomised (Agostini 2004; Le Sage 2001) or because they were not masked (Hulbert 1991; Kirkpatrick 1993) or both (Jackson 1960). In general effect estimates were similar after excluding these trials. Any differences favoured the no-patch group. This is perhaps unsurprising as it may be that participants and investigators were expecting patching to be beneficial.

Table 3. Sensitivity analysis excluding studies at high risk of bias
  1. RR: Risk ratio; MD: Mean difference SMD: standardised mean difference

    *These were studies that were quasi-randomised (Agostini 2004; Le Sage 2001) or not masked (Hulbert 1991; Kirkpatrick 1993) or quasi-randomised and non masked (Jackson 1960).

    **RR of less than 1 favours no-patch group

    § MD of greater than 0 favours no-patch group i.e. longer time to heal in patch group

    RR of less than 1 favours patch group

AnalysisIncluding all trialsExcluding studies at high risk of bias*
Complete healing after 24 hours**RR 0.89, 95% CI 0.79 to 1.00; studies = 7RR 0.94, 95% CI 0.74 to 1.21; studies = 3
Complete healing after 48 hours**RR 0.97, 95% CI 0.91 to 1.02; studies = 6RR 1.00, 95% CI 0.92 to 1.09; studies = 1
Complete healing after 72 hours**RR 1.01, 95% CI 0.97 to 1.05; studies = 4RR 1.00, 95% CI 0.92 to 1.09; studies = 1
Days to complete healing§MD 0.14, 95% CI 0.00 to 0.27; studies = 6MD 0.29, 95% CI 0.09 to 0.48; studies = 2
Analgesic useRR 0.95, 95% CI 0.69 to 1.32; studies = 3RR 0.95, 95% CI 0.69 to 1.29; studies = 2
Adverse eventsRR 3.24, 95% CI 0.87 to 12.05; studies = 8RR 5.01, 95% CI 0.91 to 27.44; studies = 4

Discussion

Summary of main results

There was little evidence of demonstrable differences across primary and secondary outcomes for both patched and non-patched groups. This may have been attributable in part to a high attrition rate in some studies, which may affect eventual analyses. It is therefore reasonable to conclude that patching of the eye is not useful for the treatment of simple, traumatic corneal abrasions.

Overall completeness and applicability of evidence

Wearing a patch renders an individual acutely monocular. This has important consequences for functional tasks requiring depth perception. These activities include walking or step climbing. Disorientation with sudden monocular vision may also limit other activities which require breadth of field, such as driving.

Acute corneal abrasions cause considerable pain which can impact upon visual acuity. This may be alleviated to an extent through patching, which theoretically obviates mechanical stimulation of the lesion by the eyelids during blinking. Interestingly, not patching was associated with a larger, albeit small, reduction in pain compared to patching. The authors here recognise the presence of other factors leading to ocular discomfort following a corneal abrasion, such as ciliary spasm, which may be alleviated via cycloplegics. We have also observed visual acuity to be conversely normal despite any pain. Although the study by Menghini 2013 showed an apparent improvement in BCVA in the patch group and none in the no-patch group, this may have been related to a state of co-operability during initial assessment correlating with the degree of pain (noted to be higher in the patch versus the non-patch group). Final visual acuities were noted to be similar in both groups as we would expect most corneal abrasions to be fully healed by the first week, barring any complications. Most patients also experience a rapid recovery of their vision, and the exclusion of a patch in their management may facilitate attendance at follow-up visits, if required, as they would not require accompaniment.

There were qualitative differences in the adjunctive treatment options applied to different treatment arms in each trial (see Table 4). These differences include the use of cycloplegics, analgesics and antibiotics. In the patch group, antibiotics were often administered prior to patching and at times re-administered a day later with removal of the patch for clinical assessment during follow-up. In contrast, the control group often received antibiotic drops or ointment more regularly during the day. Agostini 2004 instructed participants in the no-patch group to use an antibiotic ointment three times daily either for five days or until closure of the lesion, while the patch group only received application of the ointment prior to being patched daily. Similarly, in studies by Arbour 1997, Campanile 1997, Kirkpatrick 1993, Le Sage 2001, Menghini 2013, Michael 2002 and Patterson 1996, participants in the no-patch group applied topical antibiotics to the affected eye several times daily, while those in the patch group often had a single instillation of this treatment prior to application of the patch. It is possible that the use of the cycloplegics or ophthalmologic antibiotics may affect healing rates sufficiently to overshadow any benefit of patching. The use of cycloplegics may additionally induce comfort and reduce further mechanical trauma from rubbing.

Table 4. Additional treatments
  Mydriatics   Antibiotics   Analgesics  
Study Patch No Patch Differences Patch No Patch Differences Patch No Patch Differences
Agostini 2004G. cyclopentolate Hydrochloride 1% before dischargeG. cyclopentolate hydrochloride 1% before dischargeNo differencesEpitezan (amino acids 25 mg; chloramphenicol 5 mg; methionine 5 mg; retinol acetate 10.00 IU) at presentation and once daily with a new bandage every 24 h until healedEpitezan (amino acids 25 mg; chloramphenicol 5 mg; methionine 5 mg; retinol acetate 10.00 IU) TDS for 5 days or until closure of the epithelial defectFrequency of administrationAllowed to use oral analgesics and anti-inflammatoriesAllowed to use oral analgesics and anti-inflammatoriesNo differences
Arbour 1997G. homatropine hydrobromide 2% before dischargeG. Homatropine Hydrobromide 2% before dischargeNo differencesOc. sulfacetamide sodium 10% before application of eye patchOc. sulfacetamide sodium 10% BDSFrequency of administrationPO acetaminophen (325-650 mg) or codeine-acetaminophen (30/300 mg) 1-2 tablets QDSPO acetaminophen (325-650 mg) or codeine-acetaminophen (30/300 mg) 1-2 tablets QDSNo differences
Campanile 1997G. cyclopentolate hydrochloride 1% (one drop)G.cyclopentolate hydrochloride 1% (one drop)No differencesOc. erythromycin once before application of eye patchOc. erythromycin Q6H (24 h)Frequency of administrationNot specifiedNot specifiedInadequate information
Hulbert 1991Not specifiedNot specifiedInadequate informationG. chloramphenicol 0.5% 2 drops at each reviewG. chloramphenicol 0.5% 2 drops at each reviewNo differencesNot specifiedNot specifiedInadequate information
Jackson 1960G. atropine 1% PRNG. atropine 1% PRNInadequate informationOc sulfacetamide 10% TDS PRNOc sulfacetamide 10% TDS PRNInadequate informationNot specifiedNot specifiedInadequate information
Kaiser 1995G. phenylephrine 2.5%/tropicamide 1% before application of eye patchG. phenylephrine 2.5%/Tropicamide 1% TDSFrequency of administrationOc. erythromycin/polysporin once for 24 h, then remove patch and instil ointment TDS for 5 days or until abrasion is completely healedOc. erythromycin/polysporin TDS for 5 days or until complete healingFrequency of administrationMild oral analgesics including acetaminophen, ibuprofen, or aspirinMild oral analgesics including acetaminophen, ibuprofen, or aspirinNo differences
Kirkpatrick 1993G. homatropine hydrobromide 2% before application of eye patchG. homatropine hydrobromide 2% once dailyFrequency of administrationOc. chloramphenicol once before application of eye patchOc. chloramphenicol QDS for 3 days after complete healingFrequency of administrationSimple analgesics such as aspirin and paracetamolSimple analgesics such as aspirin and paracetamolNo differences
Le Sage 2001Mydriatics PRNMydriatics PRNInadequate informationOc. erythromycin at initial visit and at each reviewOc. erythromycin QDSFrequency of administrationOpioid analgesic PRNOpioid analgesic PRNNo differences
Menghini 2013NANANo differencesOc. ofloxacin before application of eye patchOc. ofloxacin QDSFrequency of administrationOral analgesics PRNOral analgesics PRNInadequate information
Michael 2002G. cyclopentolate hydrochloride 1% (one drop)G. cyclopentolate hydrochloride 1% (one drop)eNo differencesOc. erythromycin before application of eye patchOc. erythromycin TDS until follow-upFrequency of administrationPO ibuprofen (10 mg/kg per dose to a maximum dose of 400 mg) Q6-8gh PRN and Acetaminophen (15 mg/kh per dose to a maximum of 500 mg) q4-6 h for breakthrough painPO Ibuprofen (10 mg/kg per dose to a maximum dose of 400 mg) q6-8gh PRN and Acetaminophen (15 mg/kh per dose to a maximum of 500 mg) q4-6 h for breakthrough painInadequate information
Patterson 1996Not specifiedNot specifiedInadequate informationOc. tobramycin before application of eye patchG. tobramycin Q4H while awakeFrequency of administrationPO Ketoprofen 75 mg PRNPO Ketoprofen 75 mg PRNInadequate information
Rao 1994G. cyclopentolate hydrochloride 1%G. cyclopentolate hydrochloride 1%Inadequate informationOc. chloramphenicol 1%Oc. chloramphenicol 1%Inadequate informationPO Paracetamol PRNPO Paracetamol PRNInadequate information

Another theoretical problem associated with patching that has been proposed is that the warm, moist environment may support bacterial proliferation (Parrish 1988). Other problems include decreased oxygenation of the cornea, reduced epithelial turnover and decreased elimination of cellular metabolism waste products, which may interfere with the washout of bacteria. However, the complication of infection was reported in such low numbers in the trials that it would require much larger numbers of participants enrolled in the studies to demonstrate a difference, if any, between patching and no patching.

Furthermore, this review has revealed a lack of evidence to offer recommendations in the management of large corneal abrasions (that is, more than 10 mm2). In these circumstances, some practitioners may advocate the use of a patch with the rationale that this may promote epithelial proliferation and migration. This treatment option has however not been investigated satisfactorily in randomised controlled trials. Only one trial performed a subgroup analysis (Kaiser 1995). Even though this was one of the largest trials included in this review, there were insufficient large abrasions to make the subgroup analysis informative. Kaiser 1995 recommends that people with large abrasions as well as deeper stromal and partial-thickness defects should be managed with the currently accepted standard of care of an antibiotic ointment, mydriatic drops and a pressure patch. However, as mentioned, this treatment regimen is not evidence-based.

Quality of the evidence

The overall quality of the evidence was moderate to low. The certainty of the effect estimates was downgraded on the basis of study limitations and imprecision. Two of the studies were quasi-randomised, two were not masked and one study was quasi-randomised and not masked. Lack of masking may be problematic, especially as assessment of the outcome consisted of a subjective component. The other trials were generally poorly reported. Only one trial was judged to be at low risk of bias in all domains. In general the results of the studies were consistent. Excluding studies at high risk of bias, either because they were quasi-randomised or not masked, did not affect the conclusions of the study.

Potential biases in the review process

Many of the included trials in this review are not recent studies. This has complicated the process of both retrieving unpublished data and seeking clarification from the authors of these studies. Previous attempts at contacting authors of the relevant studies were unsuccessful. Although we have tried to identify all relevant trials available, the search strategies adopted for this review have mostly revolved around electronic searches. Such an approach may result in the omission of unpublished data, which could result in the introduction of publication bias.

Agreements and disagreements with other studies or reviews

This review contains a more thorough search of the grey literature than either of the previous reviews on this topic (Flynn 1998; Yamada 2001) and has included studies published in non-English journals. This is important because exclusion of even a small number of trials could alter the results of this meta-analysis. This review also corrects errors in both aforementioned reviews, with respect to the trial by Jackson 1960, where data were incorrectly assigned to day two rather than day one in the meta-analysis. However, the findings of this review are in agreement with the conclusions of these reviews (Flynn 1998; Yamada 2001), which have neither demonstrated an improvement in healing rates nor a reduction in pain with patching compared to non-patching.

Authors' conclusions

Implications for practice

The results of this current review do not support patching of corneal abrasions. However, the meta-analysis should be interpreted cautiously, as we identified differences between treatment regimes, such as the frequency of antibiotic dosing. Furthermore, limited data exists for the treatment of large (> 10 mm2) corneal abrasions.

Implications for research

It would be helpful for future studies in this area to focus on minimising possible confounding variables in the management of participants with corneal abrasions without patching. Only by standardising the management of participants in either treatment arm and minimising the number of differences can we confidently determine the efficacy of eye patching.

It would also be helpful to develop a sufficiently powered trial to investigate either large corneal abrasions or corneal abrasions with partial-thickness defects involving the deeper stromal tissue, which would answer some useful and practical questions. A trial designed around this topic would need to take into account the flaws that we found in the trials included in this review.

Acknowledgements

  • Mansur Rabiu for his contributions to the first published version of this review (Turner 2006).

  • Michael Clarke (Supervisor of an MSc thesis at the University of Oxford).

  • Catey Bunce, Gabriela Czanner, Roberta Scherer, Stephanie Watson, Suzanne Brodney-Folse, Dawn Lim for their comments on the review and/or protocol.

  • Anupa Shah and Richard Wormald at Cochrane Eyes and Vision, London.

  • Jennifer Evans assisted with the methodological aspects of this review.

  • Iris Gordon from Cochrane Eyes and Vision prepared and executed the updated electronic searches.

  • Matthew Page and Miranda Cumpston from the Australasian Cochrane Centre for their assistance with this review.

Data and analyses

Download statistical data

Comparison 1. Patching versus no patching
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Complete healing after 24 hours7531Risk Ratio (M-H, Random, 95% CI)0.89 [0.79, 1.00]
2 Complete healing after 48 hours6497Risk Ratio (M-H, Random, 95% CI)0.97 [0.91, 1.02]
3 Complete healing after 72 hours4430Risk Ratio (M-H, Random, 95% CI)1.01 [0.97, 1.05]
4 Number of days to complete healing6642Mean Difference (IV, Random, 95% CI)0.14 [0.00, 0.27]
5 Pain at 24 hours2193Risk Ratio (M-H, Fixed, 95% CI)1.51 [0.86, 2.65]
6 Analgesic use3256Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.69, 1.32]
7 Photophobia3418Risk Ratio (M-H, Fixed, 95% CI)0.80 [0.45, 1.42]
8 Lacrimation2255Risk Ratio (M-H, Fixed, 95% CI)1.04 [0.64, 1.68]
9 Foreign body sensation3418Risk Ratio (M-H, Fixed, 95% CI)0.98 [0.63, 1.55]
10 Blurred vision2255Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.48, 2.07]
11 Adverse events8660Risk Ratio (M-H, Random, 95% CI)3.24 [0.87, 12.05]
12 Complete healing after 24 hours: subgroup analysis7724Risk Ratio (M-H, Random, 95% CI)0.88 [0.80, 0.97]
12.1 Studies including abrasions caused by removal of foreign bodies4290Risk Ratio (M-H, Random, 95% CI)0.89 [0.78, 1.03]
12.2 Studies excluding abrasions caused by removal of foreign bodies3241Risk Ratio (M-H, Random, 95% CI)0.87 [0.67, 1.12]
12.3 Studies only including abrasions caused by removal of foreign bodies2193Risk Ratio (M-H, Random, 95% CI)0.87 [0.73, 1.03]
Analysis 1.1.

Comparison 1 Patching versus no patching, Outcome 1 Complete healing after 24 hours.

Analysis 1.2.

Comparison 1 Patching versus no patching, Outcome 2 Complete healing after 48 hours.

Analysis 1.3.

Comparison 1 Patching versus no patching, Outcome 3 Complete healing after 72 hours.

Analysis 1.4.

Comparison 1 Patching versus no patching, Outcome 4 Number of days to complete healing.

Analysis 1.5.

Comparison 1 Patching versus no patching, Outcome 5 Pain at 24 hours.

Analysis 1.6.

Comparison 1 Patching versus no patching, Outcome 6 Analgesic use.

Analysis 1.7.

Comparison 1 Patching versus no patching, Outcome 7 Photophobia.

Analysis 1.8.

Comparison 1 Patching versus no patching, Outcome 8 Lacrimation.

Analysis 1.9.

Comparison 1 Patching versus no patching, Outcome 9 Foreign body sensation.

Analysis 1.10.

Comparison 1 Patching versus no patching, Outcome 10 Blurred vision.

Analysis 1.11.

Comparison 1 Patching versus no patching, Outcome 11 Adverse events.

Analysis 1.12.

Comparison 1 Patching versus no patching, Outcome 12 Complete healing after 24 hours: subgroup analysis.

Appendices

Appendix 1. CENTRAL search strategy

#1 MeSH descriptor Cornea
#2 MeSH descriptor Corneal Diseases
#3 MeSH descriptor Epithelium, Corneal
#4 MeSH descriptor Eye Injuries
#5 (#1 OR #2 OR #3 OR #4)
#6 MeSH descriptor Wounds and Injuries
#7 injur* or abrasion* or erosion* or trauma* or wound* or foreign bod*
#8 (#6 OR #7)
#9 eye* or cornea*
#10 (#8 AND #9)
#11 (#5 OR #10)
#12 MeSH descriptor Occlusive Dressings
#13 patch* or bandage* or plaster* or wool* or dress* or pad* or gauze or occlusi*
#14 (#12 OR #13)
#15 (#11 AND #14)

Appendix 2. MEDLINE (OvidSP) search strategy

1 randomized controlled trial.pt.
2 (randomized or randomised).ab,ti.
3 placebo.ab,ti.
4 dt.fs.
5 randomly.ab,ti.
6 trial.ab,ti.
7 groups.ab,ti.
8 or/1-7
9 exp animals/
10 exp humans/
11 9 not (9 and 10)
12 8 not 11
13 exp cornea/
14 exp corneal diseases/
15 exp epithelium corneal/
16 exp eye injuries/
17 or/13-16
18 exp "wounds and injuries"/
19 (injur$ or abrasion$ or erosion$ or trauma$ or wound$ or foreign bod$).tw.
20 or/18-19
21 (eye$ or cornea$).tw.
22 20 and 21
23 17 or 22
24 exp occlusive dressings/
25 (patch$ or bandage$ or plaster$ or wool$ or dress$ or pad$ or gauze or occlusi$).tw.
26 or/24-25
27 23 and 26
28 12 and 27

The search filter for trials at the beginning of the MEDLINE strategy is from the published paper by Glanville (Glanville 2006).

Appendix 3. EMBASE (OvidSP) search strategy

1 exp randomized controlled trial/
2 exp randomization/
3 exp double blind procedure/
4 exp single blind procedure/
5 random$.tw.
6 or/1-5
7 (animal or animal experiment).sh.
8 human.sh.
9 7 and 8
10 7 not 9
11 6 not 10
12 exp clinical trial/
13 (clin$ adj3 trial$).tw.
14 ((singl$ or doubl$ or trebl$ or tripl$) adj3 (blind$ or mask$)).tw.
15 exp placebo/
16 placebo$.tw.
17 random$.tw.
18 exp experimental design/
19 exp crossover procedure/
20 exp control group/
21 exp latin square design/
22 or/12-21
23 22 not 10
24 23 not 11
25 exp comparative study/
26 exp evaluation/
27 exp prospective study/
28 (control$ or prospectiv$ or volunteer$).tw.
29 or/25-28
30 29 not 10
31 30 not (11 or 23)
32 11 or 24 or 31
33 exp cornea/
34 exp cornea disease/
35 exp cornea epithelium/
36 exp eye injury/
37 or/33-36
38 exp injury/
39 (injur$ or abrasion$ or erosion$ or trauma$ or wound$ or foreign bod$).tw.
40 or/38-39
41 (eye$ or cornea$).tw.
42 40 and 41
43 37 or 42
44 (patch$ or bandage$ or plaster$ or wool$ or dress$ or pad$ or gauze or occlusi$).tw.
45 43 and 44
46 32 and 45

Appendix 4. LILACS search strategy

injur$ or abrasion or erosion or trauma or foreign bod$ and eye$ or cornea$ and patch$ or bandage$ or plaster$ or wool$ or dressing$ or pad$ or gauze or occlus$

Appendix 5. OpenGrey search strategy

(cornea OR eye) AND (patch OR bandage OR plaster OR wool OR dressing OR pad OR gauze OR occlusion)

Appendix 6. ISRCTN search strategy

(cornea OR eye) AND (injury OR abrasion OR erosion OR trauma) AND (patch OR bandage OR plaster OR wool OR dressing OR pad OR gauze OR occlusion)

Appendix 7. ClinicalTrials.gov search strategy

(cornea OR eye) AND (injury OR abrasion OR erosion OR trauma) AND (patch OR bandage OR plaster OR wool OR dressing OR pad OR gauze OR occlusion)

Appendix 8. ICTRP search strategy

(cornea OR eye) AND (patch OR bandage OR plaster OR wool OR dressing OR pad OR gauze OR occlusion)

Appendix 9. Glossary

This section contains a list of abbreviations that are used in this systematic review.

ADLs - Activities of Daily Living, routine daily self-care activities which include eating, dressing, toileting, the ability to transfer or walk as well as maintain continence

BCVA - Best Corrected Visual Acuity

BDS (Latin: bis die sumendum)-Two times daily

G. (Latin: guttae or gutta)- Drops or drop

ITT - Intention To Treat

I.U - International Units

NA- Not Applicable

Oc. (Latin: oculentum)- Ointment

PRN (Latin: pro re nata)- As required

Q4H (Latin: quaque quarta hora)- Every 4 hours

Q6H (Latin: quaque sexta hora)- Every 6 hours

TDS (Latin: ter die sumendum)- Three times daily

What's new

DateEventDescription
9 May 2016New citation required but conclusions have not changedThe risk of bias tables and text of the review have been updated.
9 May 2016New search has been performedAn updated search of the literature was performed in May 2016. One new study (Menghini 2013) was identified and incorporated into this review.

History

DateEventDescription
26 October 2008AmendedConverted to new review format.

Contributions of authors

Original review

Conceiving the review: AT
Designing the review: AT, MR
Co-ordinating the review: AT
Undertaking manual searches: AT, MR
Screening search results: AT, MR
Organising retrieval of papers: AT
Screening retrieved papers against inclusion criteria: AT, MR
Appraising quality of papers: AT, MR
Abstracting data from papers: AT, MR
Writing to authors of papers for additional information: AT
Obtaining and screening data on unpublished studies: AT, MR
Data management for the review: AT, MR
Entering data into RevMan: AT, MR
Analysis of data: AT, MR
Interpretation of data: AT, MR
Writing the review: AT

2016 update

Screening search results: CL, BL
Organising retrieval of papers: CL, BL
Screening retrieved papers against inclusion criteria: CL, BL
Appraising quality of papers: CL, BL
Abstracting data from papers: CL, BL
Data management for the review: CL, BL
Entering data into RevMan: CL, BL
Analysis of data: CL, BL
Interpretation of data: CL, BL
Writing the review: CL, BL
Reviewing draft of the review: AT

Declarations of interest

None of the authors has any conflict of interest or proprietary interests related to this submission.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Rhodes Trust, UK.

  • National Institute for Health Research (NIHR), UK.

    • Richard Wormald, Co-ordinating Editor for Cochrane Eyes and Vision (CEV) acknowledges financial support for his CEV research sessions from the Department of Health through the award made by the National Institute for Health Research to Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology for a Specialist Biomedical Research Centre for Ophthalmology.

    • The NIHR also funds the CEV Editorial Base in London.

    The views expressed in this publication are those of the authors and not necessarily those of the NIHR, NHS, or the Department of Health.

Differences between protocol and review

The protocol for this review was originally published in 2004 (Turner 2004). Since that time there have been improvements in Cochrane methods, in particular the assessment of risk of bias and production of 'Summary of findings' tables and GRADE assessment. These new methods have been incorporated.

We have included the following additional outcomes: insomnia assessments and duration of medical leave and have added 72 hours follow-up for complete healing.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Agostini 2004

Methods

Parallel study design

Duration: 4 months
Post-randomisation exclusions: 34% (28 patients excluded for not properly following treatment - 14 in patch group and 14 in non-patch group)
Outcomes assessor and participants not masked
Potential confounder: ointment containing vitamin A used in both groups

Participants

Inclusion criteria

  • Foreign body only

  • > 18 years old

Exclusion criteria

  • Contact lens users

  • Corneal foreign body with infiltrate or organic matter

  • Other trauma to eye in addition to corneal foreign body

  • History of corneal eye disease

Setting: Brazil
Number of participants: 82
Exclusions after randomisation: 28
Total available for analysis: 54
Mean age: 34 years old
Sex: Male:Female 53:1

Interventions
  • Gauze pads firmly taped to prevent blinking. Used until complete healing

  • Cyclopentolate 1% administered on initial presentation to both groups. Epitezan (retinol acetate, methionine, amino acids, chloramphenicol ointment) used in both groups 3 times/day for 5 days or until closure of epithelial defect in non-patch group and once/day in patched group with new bandage

  • Participants in both groups were permitted to use oral analgesics and anti-inflammatories

Outcomes
  • Area of erosion, difference in lesion area

  • Symptoms

  • Mean time to healing

  • Pain scores VAS

Notes

Date study conducted: July 2001 to October 2001

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskAllocation into either treatment group was performed by alternating patients. This represented a quasi-randomisation technique
Allocation concealment (selection bias)High riskLikely difficult to conceal method of allocation from investigators involved in treatment. The enrolment of participants and how they were counselled on the treatment process was also not discussed in the paper.
Masking of participants and personnel (performance bias)
All outcomes
High riskNo information on the masking process was given. In particular, with regards to the allocation process, post consultation and subsequent follow-up visits
Masking of outcome assessment (detection bias)
All outcomes
High riskThe use of mydriatics or analgesia may confound results which rely on participants’ subjective input, such as the reporting of pain via the pain scale. Was there any information collection on the presence of anterior chamber cells/flare as well as predominant location of corneal abrasion i.e. centrally in line with visual axis or peripherally which may further contribute to the presence of discomfort? Also, no mention was made as to whether patches were taken off prior to consultation (i.e. masking of the assessor) and if the assessors were aware of the defined endpoints of the study
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskDrop-out rates (34%) were accounted for and discussed adequately by the authors. However high drop out rates may have influenced the final outcome obtained. There was no mention of an intention to treat approach, which should have been performed
Selective reporting (reporting bias)Unclear riskThere was no mention about how missing data was dealt with. Authors reported results where both demonstrable and non-demonstrable results were identified.

Arbour 1997

Methods

Randomised controlled trial with parallel design

Duration: 22 months

Method of randomisation: unclear
Post-randomisation exclusions: 2% (1 patient excluded because of non-compliance with treatment (unclear distribution))
Outcomes assessors masked. Fluorescein stain (slit lamp) and photo/template drawing used to assess epithelial erosion
Potential confounder: different use of antibiotic ointment in two groups

Participants

Inclusion criteria

  • Presence of epithelial erosion greater than 1 mm in smallest diameter

  • Secondary to trauma or recurrent erosion syndrome

  • Sparing of Bowman membrane

Exclusion criteria

  • Signs of stromal, endothelial or neural corneal involvement (on slit lamp examination)

  • Previous patch

  • Corneal abrasion secondary to removal of foreign body

  • Signs of infection

  • Contact lens wear

  • Use of topical or systemic steroids

  • Inability to comply with daily follow-up

  • Patient refusal

Setting: Canada
Number of participants: 46 (48 eyes)
Exclusions after randomisation: 1
Total available for analysis: 45 (2 participants had bilateral involvement and 1 eye was subsequently patched while the other eye was not)
Mean age: in patched group 41.6 years and 39.8 years in non-patched group.
Sex: in patched group 32% female and in non-patched group 36% female

Interventions
  • Double eye pad tightly taped with sufficient pressure to prevent lid opening

  • Homatropine 2% and sulfacetamide 10% were applied topically before patch applied. In non-patch group, homatropine 2% also used, but antibiotic ointment applied twice daily

  • Oral analgesia (acetaminophen (paracetamol) and codeine, or acetaminophen alone) was prescribed as needed for both groups

Outcomes
  • Slit lamp examination for stromal oedema and anterior chamber inflammation

  • Corneal photograph: analysis of digitised images for linear and surface rates of re-epithelialisation (mm/hour), perimeter, area of erosion, diameter of largest circle included in erosion, shape index (ratio between perimeter and area of erosion)

  • Discomfort associated with patch

  • Pain VAS score

  • Insomnia

  • Symptoms including: pain, foreign body sensation, photophobia,and tearing

  • Analgesic requirements

Notes

It is noted that participants with recurrent erosion syndrome were included in the study. The proportion of the test group and criteria that was used to classify these participants was not elaborated upon. "Participants with recurrent erosion syndrome" refers to patients with corneas with defective adhesions at the level of the epithelial layer. This may affect healing rates and perception of pain

Date study conducted: January 1992 to October 1993

Funding: Quebec Eye Bank Foundation Inc and Fonds de la Recherche en Santé du Québec to the Guy-Bemier Research Centre, Maisonneuve- Rosemont Hospital, Montreal

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk

Randomisation methods utilised in allocating participants was not described. In participants with bilateral corneal abrasions, it was also unclear as to how treatment allocations were made for each eye

The authors have described prescribing either acetaminophen or codeine-acetaminophen. However, no information was provided on how analgesic requirements were determined and how “any” differed from mild and strong. No further information was also provided on the type of analgesia given to participants with bilateral involvement

Allocation concealment (selection bias)Unclear riskNo information was provided about how allocations were performed even though trial has been described as 'randomised' and how concealment of the allocator, if any, was undertaken
Masking of participants and personnel (performance bias)
All outcomes
Unclear riskIt is not clear if the two observers, JDA and IB were masked from the 2 intervention arms and the final outcomes of the study
Masking of outcome assessment (detection bias)
All outcomes
Unclear riskThe digitised corneal images were assessed by a masked observer. However, the treatment end-point was not clearly defined and it was unclear as to whether re-epithelialisation was determined clinically, or through the assessment of the digitised corneal images. Asessors do not appear to be masked to other study outcomes.
Incomplete outcome data (attrition bias)
All outcomes
Low riskOne participant was excluded because of non-compliance with treatment. Fair rationale provided for excluding epithelial defects that had already closed on the first follow-up visit, in calculating the linear and surface speeds of re-epithelialisation
Selective reporting (reporting bias)Unclear riskThe element of discussion with a resultant applied score may potentially alter the conclusion. No further information was provided on scoring components and discrepancies noted between both observers and reconciliation, if any.

Campanile 1997

Methods

Randomised controlled trial with parallel design

Duration: 7.5 months

Method of allocation: unclear
Post-randomisation exclusions: 13.5% (10 participants: 3 removed the patch secondary to discomfort and pain; 3 who were not patched were lost to follow-up; 4 excluded because of deviation from treatment protocol by the provider (unclear distribution))
Outcomes assessors masked (participants asked to remove patch at home and not mention management at follow-up assessment). Fluorescein stain (slit lamp) and grid template/hand drawing used to assess outcomes
Potential confounder: different dosing intervals of erythromycin ophthalmic ointment in 2 groups

Participants

Inclusion criteria

  • 5 years old and above

  • Traumatic corneal epithelial defects from corneal abrasion or removal of foreign body

Exclusion criteria:

  • < 5 years old

  • Corneal disease

  • Allergy to erythromycin

  • Ocular infection

  • Additional trauma-related injuries

Setting: USA
Number of participants: 74
Exclusions after randomisation: 10
Total available for analysis: 64
Mean age: 31 (range 5-74)
Sex: 54.7% male

Interventions
  • Double gauze pads with first pad folded in half to prevent eyelid opening

  • One drop of cyclopentolate 1% on initial presentation

  • Patched group received 1 dose of erythromycin ointment

  • Unpatched group received erythromycin ointment 4 times per day

Outcomes
  • Percent healing (24 hours) collected from grid template hand drawing

  • Mean initial abrasion size

Notes

Date study conducted: May 1995 to January 1996

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk

It is unclear how randomisation was achieved in the study

The baseline characteristics, exclusion and or inclusion criteria for participants was not delineated: mechanism of injury or concomitant injury, or whether any prior treatment was attempted, whether eye received any prior surgery or trauma or any other prior ocular injury

How was consent taken and how would patients aged 5-16 (minors) have been counselled?

Allocation concealment (selection bias)Unclear riskInformation on the randomisation and allocation methods are limited, together with a comparison on baseline characteristics and numbers per age group in each study group
Masking of participants and personnel (performance bias)
All outcomes
Low riskAttempts to mask allocated interventions at initial visit and follow-up were identified
Masking of outcome assessment (detection bias)
All outcomes
Low riskParticipants were instructed to remove any traces of being in the patched group prior to presenting for their follow-up assessment. All participants were tasked not to reveal their treatment regime. Furthermore, it is unclear if the assessors were aware of the defined endpoints of the study
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Drop-out rates between both treatment arms appear to be similar. However it is unclear which group the 4 individuals belonged to. The reasons protocol was not followed for the 4 excluded individuals could also have been revealed. No ITT analysis was performed

Further information on possible reasons for loss to follow-up would have been helpful – as to whether healing was thought to have occurred or worsening of condition with possible re-attendance elsewhere

Selective reporting (reporting bias)Unclear risk

Limited data reported in the article. How were the individual characteristics (i.e. gender, age) distributed between cases and controls? Was there any demonstrable difference in the distribution between both groups? If so, have these factors been adequately controlled for?

"There were no complications associated with either group"

The above statement was made without further elaboration on the specifics which may have constituted “complications” – again, may have been lost in those who were excluded from the study. Compliance was also not addressed in the study

Hulbert 1991

Methods

Randomised controlled trial

Duration: 3 months

Method of allocation: unclear
Post-randomisation exclusions: 9.1% ( 3 patients with persistent residual foreign body or stain (unclear distribution))
Masking of outcomes assessor unclear

Fluorescein staining (x 4 magnification) used to assess abrasion size

Participants

Inclusion criteria:

  • corneal epithelial defects resulting from removal of foreign body

Exclusion criteria:

  • participants in whom residual foreign body or stain remained after first attendance for removal

Setting: UK
Number of participants: 33
Exclusions after randomisation: 3
Total available for analysis: 30
Age: NR

Sex: NR

Interventions
  • 2 drops of chloramphenicol 0.5% into affected eye of all patients at each review

  • Gauze with enough bulk to exert slight pressure on the closed eye for patients in the patch group, secured with bandage

Outcomes
  • Number of days for complete healing, marked by absence of corneal staining

  • Level of discomfort indicated by descriptors "painful" or "painless" on day 1 and 2

Notes

Date study conducted: October 1997 and January 1998

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details on how randomisation was achieved
Allocation concealment (selection bias)Unclear riskNo information regarding allocation concealment, and how it was performed
Masking of participants and personnel (performance bias)
All outcomes
High riskNo masking of participants and personnel to the allocated interventions appears to have been undertaken in this study
Masking of outcome assessment (detection bias)
All outcomes
High riskWas there only one assessor and was the assessor masked to the 2 intervention arms? No masking of the allocated interventions appears to have been undertaken as part of the ongoing assessment and management of study participants. It is also unclear if the assessors were aware of the defined endpoints of the study
Incomplete outcome data (attrition bias)
All outcomes
Low riskAll enrolled participants were followed up until study completion. No attrition or loss to follow-up was noted. Fair exclusion criteria applied
Selective reporting (reporting bias)Unclear riskNo discussion of baseline demographic information of the study population. All outcomes were analysed in the final evaluation

Jackson 1960

Methods

Quasi-randomised controlled trial

Duration: 2 months

Method of allocation: alternate days for first attendance
Post-randomisation exclusions: 29% (14 patched and 24 non-patched participants lost to follow-up; exclusion of 16 for incomplete records, 6 for non-compliance with treatment, 5 with more serious lesions discovered subsequently)
Outcomes assessor not masked. Fluorescein stain (no slit lamp mentioned) was used to assess corneal abrasions
Potential confounder: different usage of antibiotic ointment in two groups

Participants

Inclusion criteria

  • Superficial abrasion of the cornea

Exclusion criteria:

  • Any complicated lesion such as embedded foreign body or any other part of the eye

Setting: UK
Number of participants: 222
Exclusions after randomisation: 65
Total available for analysis: 157
Age: NR

Sex: NR

Interventions
  • A cotton-wool pad covered by net, taped on

  • All participants received sulfacetamide 10% drops and if considered necessary, 1% atropine drops

  • Participants without the eye pad continued to use sulfacetamide 10% antibiotic drops 3 times per day

Outcomes
  • Number of days until complete healing, marked by absence of fluorescein staining

Notes

Date study conducted: NR

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High risk

Quasi-randomisation. Allocation based on alternate days of presentation

How were participants counselled on the treatment process above and on participation consent in the study?

Allocation concealment (selection bias)High riskParticipants were allocated to either treatment group based on alternate days of presentation. No evidence of concealment of allocation sequence
Masking of participants and personnel (performance bias)
All outcomes
High riskMasking of participants to allocation not possible. However, no masking of assessors to treatment allocations described.
Masking of outcome assessment (detection bias)
All outcomes
High riskNo masking of assessors to treatment allocations described. Additionally, it is unclear if the assessors were aware of the defined endpoints of the study
Incomplete outcome data (attrition bias)
All outcomes
High risk

The authors did not mention what constituted a "more serious lesion" that resulted in exclusion. This may potentially confound results, if for instance, participants with larger abrasions were deliberately removed from the study

Marked proportion of participants lost to follow-up, particularly in the non-patched group. ITT analysis should have been performed

Selective reporting (reporting bias)Unclear risk

No discussion of baseline demographic information of the study population

Participants who dropped out from the study were reported. However, no statistical analyses of the results was performed

Kaiser 1995

Methods

Randomised controlled trial with parallel design

Duration of trial: unknown

Method of randomisation: unknown
Post-randomisation exclusions: 9.9% (n = 22 excluded for ineligibility, lost to follow-up or dropped out from study (unclear distribution))

Outcomes assessors not masked.
Fluorescein stain (with slit lamp) was used to assess corneal erosions. Vertical and horizontal size of corneal abrasion measured
Potential confounder: different mydriatic drops, antibiotic ointments and oral analgesic regimens used across all study participants in both groups

Participants

Inclusion criteria

  • Traumatic corneal abrasion or removal of superficial foreign body of less than 36 hours' duration

  • > 18 years old

  • No history of eye trauma or disease in the affected eye

  • No other signs of ocular trauma

  • Simple epithelial defect without stromal oedema, loss, or infiltrate

  • No treatment before entering the study

Exclusion criteria:

  • Contact lens wearers

  • Corneal dystrophies

Setting: USA
Number of participants: 223
Exclusions after randomisation: 22
Total available for analysis: 201
Mean age: 36.17 (SD 11.93, range 19-78 years old)
Sex: 71% male

Interventions
  • Double eye patch with first pad folded in half, placed over closed eyelids and bandaged. Applied tightly to prevent eye movement.

  • Patched group received mydriatic drops (2.5% phenylephrine/1% tropicamide) and erythromycin or polysporin antibiotic ointment once only before application of the patch. Told to remove the patch after 24 hours and administer the antibiotic ointment 3 times per day for 5 days or until abrasion healed.

  • The non-patch group was treated with the same topical agents 3 times per day for 5 days or until abrasion completely healed

  • Both groups permitted to use mild oral analgesics (including acetaminophen, ibuprofen or aspirin)

Outcomes
  • Number of days until complete healing. Healing considered to have occurred if pain score of 2 or below reported, or when only minor non-confluent superficial punctate staining of the corneal epithelium remained following fluorescein administration

  • Pain scores (0-10)

  • Daily questions about symptoms such as photophobia, lacrimation, foreign body sensation and blurred vision

  • Long-term complications, including recurrent erosions (for a period of 7-12 months)

Notes

Date study conducted: NR

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo information on the randomisation process undertaken was provided in the article
Allocation concealment (selection bias)Unclear riskNo information was provided regarding allocation concealment
Masking of participants and personnel (performance bias)
All outcomes
High riskNo information on whether the two observers assessing the study were masked from the interventions received by participants
Masking of outcome assessment (detection bias)
All outcomes
High riskNo information on masking. We assume that in absence of reporting on this outcome assessors were not masked
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskApproximately 10% of participants were excluded from the study. Further elaboration on this aspect of the study, reporting on how the demographical characteristics of the participants lost to follow-up compared with the rest of the study population would have been helpful
Selective reporting (reporting bias)Low riskApart from the 22 excluded participants, it appears that all other participants were included in analysis. Appropriate discussion pertaining to the limitations of the study

Kirkpatrick 1993

Methods

Randomised controlled trial with parallel design

Duration: 5 months

Method of randomisation: coin-tossing
Post-randomisation exclusions: 15.9% (7 participants in total; 3 from patch group and 2 from non-patched group lost to follow-up. 1 patient in the non-patch group developed a dendritic ulcer and another patient in the non-patch group was patched after reattending the Emergency Department)
Outcomes assessors not masked

Fluorescein stain (with slit lamp) used to assess abrasion. Approximate area calculated and recorded
Potential confounder: different dosing regimen of chloramphenicol antibiotic ointment in 2 groups

Participants

Inclusion criteria

  • > 18 year old

  • Simple corneal abrasion within the last 24 hours

  • No treatment prior to trial inclusion

Exclusion criteria

  • Previous history of eye trauma or disease in affected eye

  • Signs of significant ocular trauma

  • Presence of foreign body.

Setting: UK
Number of participants: 44
Exclusions after randomisation: 7
Total available for analysis: 37 (4 women and 13 men in patch group, 13 women and 7 men in no-patch group)
Mean age: 36.1 years old for all participants, 36.3 (SD 11.0) years old in patch group and 35 (SD 11.5) years old in non-patch group

Interventions
  • Double eye pad with bandage until healed, then chloramphenicol 4 times per day for 3 days; 1 dose of chloramphenicol ointment and homatropine 2% topically before eye pad applied.

  • Non-patch group received chloramphenicol ointment 4 times per day and homatropine 2% once per day. Participants told to continue using chloramphenicol ointment 4 times a day for 3 days after complete healing

  • Permitted to take simple analgesics such as paracetamol and aspirin for pain relief

Outcomes
  • Time to complete healing (epithelial edges apposed with only minor fluorescein staining) - number of days

  • Size of abrasion (maximum and minimum dimensions and approximate area)

  • Pain score VAS 0-100 to assess level of discomfort

Notes

Date study conducted: August 1991 to December 1991

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskAdequate randomisation technique adopted utilising a coin-toss method. However, recently published articles have suggested that many variables may in fact influence the results obtained from coin-tossing. Therefore this technique may not truly result in random allocation of participants
Allocation concealment (selection bias)Unclear riskA coin-toss method has been reported by the authors. However, no information on how interventions were allocated to participants. Furthermore, it was unclear whilst reading the study, who was involved in assessing and allocating these participants, and whether this could have potentially influenced the allocation of participants to either treatment arm
Masking of participants and personnel (performance bias)
All outcomes
High riskAssessors do not appear to have been masked to the allocated interventions during the study
Masking of outcome assessment (detection bias)
All outcomes
High riskAssessors do not appear to have been masked to the allocated interventions during the study. This may in turn have affected how findings were documented
Incomplete outcome data (attrition bias)
All outcomes
High risk

16% of patients were excluded from the study. The authors have been thorough in documenting the reasons and explaining the rationale behind why participants were excluded from the study. However, attempts to contact participants lost to follow-up should have been made, outcomes established, and an ITT analysis adopted

 

Selective reporting (reporting bias)Low riskComprehensive coverage of all information available

Le Sage 2001

Methods

Quasi-randomised controlled trial

Duration: 21 months

Method of randomisation: alternate allocation to groups

Pre-randomisation: n = 3 refused to participate, n = 6 did not meet the inclusion criteria
Post-randomisation exclusions: 17.2% (n = 17 from patch group (lost to follow-up, but 5 gave information over the telephone) and n = 11 from the non-patched group (lost to follow-up, but 3 gave information over the telephone))
Outcomes assessors masked (patched participants asked to remove patch 30 minutes before presenting for follow-up). Slit lamp used with dimensions of abrasion recorded on standardised form
Potential confounder: different usage of antibiotic ointment in two groups, and variability in mydriatics and opioid analgesics prescribed to study participants

Participants

Inclusion criteria

  • 18-60 years old

  • Able to report sufficient details of circumstances and timing of trauma

  • Able to demonstrate corneal fluorescein uptake

Exclusion criteria

  • Corneal perforation

  • History of glaucoma

  • Chemical burns

  • UV keratitis

  • Bilateral abrasions

  • Suspected corneal ulcers

Setting: Canada
Number of participants: 172 eligible participants; 3 refused to participate and 6 met exclusion criteria (final number of participants: 163)
Exclusions after randomisation: 28
Total available for analysis: 135
Mean age in patch group 32 (range 28-38) years old and in non-patch group 36 (range 31-46) years old
Sex: 82% male in patch group and 90% male in non-patch group

Interventions
  • Double occlusive patch (worn until follow-up ceased)

  • Topical erythromycin ointment was used 4 times per day in non-patch group and once per day in the patch group

  • Use of mydriatics or opioid analgesics subject to practitioner's discretion

  • Removal of foreign bodies, siderosis, or both performed when necessary

Outcomes
  • Time to complete healing

  • Dimensions of corneal abrasions

  • Presence of symptoms

  • VAS for pain and discomfort

  • Compliance

  • Analgesic use

  • Symptoms

  • Mydriatics use

Notes

With regards to removal of foreign bodies, siderosis, or both, was a bigger epithelial defect created? What proportion of participants required such intervention and at which stage of follow-up? This may confound the assessment of healing process and participant’s assessment of discomfort

Date study conducted: January 1995 to September 1996

Funding: Quebec Association of Emergency Medicine; Foundation of the CHA (Enfant-Jesus Hospital), the CHA Research Center, the Quebec Federation of General Practitioners, and the Department of Family Medicine, Laval University

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskA systematic non random approach was used for allocation
Allocation concealment (selection bias)High risk

Quasi-randomisation technique adopted in this study may have resulted in the introduction of selection bias. It was also unclear if the research nurse was masked to the randomisation sequence. Otherwise, she could have potentially confounded the selection/allocation process

Is there a particular reason why subjects had to be between 18-60 years of age? This may not correspond with the criteria set out in the other studies. insufficient information was provided on the reasons for inclusion and exclusion criteria. Definition of “traumatic corneal abrasion” was also not addressed – an abrasion in response to a mere scratch or corresponding blunt trauma with subsequent reports of discomfort may differ. In addition, the type of corneal abrasion in a population type of particular corneal thickness may not be representative

Masking of participants and personnel (performance bias)
All outcomes
Low risk

The authors have described this as a 'single-blind' study. Participants were asked to remove the patch half an hour before presenting to the ED for follow-up

It is however not clear if participants were briefed and told not to disclose how they were managed prior to presenting for follow-up. Additionally, ideally the assessor should not have access to previous clinical documentation

In participants who were lost to follow-up but were able to provide reliable information over the telephone, how were their VAS scores taken? Numerous studies have demonstrated poor correlation between the VRS and VAS pain scores

We have however chosen to give this aspect of the study the benefit of the doubt and have assessed it as possessing a low risk of bias

Masking of outcome assessment (detection bias)
All outcomes
Low riskAs above
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

17 (20.7%) of participants in the patched group were lost during follow-up, which represents a significant proportion of participants; 5 (7%) gave information by telephone (no symptoms). 13.6% of participants in the non-patched group were lost during follow-up; 3 (3.7%) gave information by telephone (no symptoms)

 

Table 1 – missing information noted in dropouts for follow-up visits: 16 (out of 82) over 2 follow-up visits in the patched group versus 10  (out of 81) over 2 follow-up visits in the non patched group. How were they accounted for? Was any information collected on the reasons they had declined further up and would persistence in a response predispose to responder bias?

 

An ITT analysis approach appears to have been adopted by the authors. However, it is unclear as to how this was performed

Selective reporting (reporting bias)Unclear riskIt is unclear how participants who were lost to follow-up (especially those that were uncontactable) were accounted for whilst calculating the percentage of those with remaining symptoms at 24 h and 48 h, the cumulative incidences, and reduction in discomfort? (since n = 82 and n = 81 was used)

Menghini 2013

Methods

Randomised controlled trial with 3 parallel study groups

Duration: 19 months (October 2008 – April 2010)

Pre-randomisation exclusion: exclusion criteria stated but no indication of the proportion of participants excluded based on criteria

Post-randomisation exclusion: 4 patients

Outcomes assessors masked from allocated intervention. Area of corneal abrasion recorded on a digital anterior segment camera and processed with a measuring tool (Synedra View ®) by two different ophthalmologists

Participants

Inclusion criteria

  • All adults 18 years or older diagnosed and treated for a superficial corneal foreign body

Exclusion criteria

  • Infectious keratitis

  • Advanced trauma with stromal loss

  • Corneal abnormalities including epithelial, stromal or endothelial dystrophies

  • Chemical trauma

  • Limbal stem cell deficiency

  • Use of chronic topical eye medication

  • Collagen vascular disease

  • Children (under the age of 16)

Setting: Department of Ophthalmology, University Hospital of Zurich, Switzerland

Number of participants: 66

Uncertain number of participants excluded pre-randomisation; 3 participants (4.5% 1 patch group (PG) , 2 ointment group (OG) were excluded post-randomisation as they failed to attend their first follow-up visit. 1 patient presented with second corneal foreign body in same eye during 1 week follow-up (unclear which treatment arm this patient belonged to )

Total available for analysis: 66

Mean age: 28.7 years in patched group and 34.3 years in non-patch (ointment only) group

Sex: all male

No demonstrable differences between groups in terms of age, gender, time to presentation, initial corneal abrasion area and initial pain score

Interventions
  • PG* (patch group): double-firm pressure patch taped over injured eye after application of ofloxacin ointment

  • CLG (contact lens group): bandage contact lens inserted and application of ofloxacin eye drops applied 4 times daily

  • OG* (ointment (non-patch) group): ofloxacin ointment applied 4 times daily

  • A number of participants in each group (PG 3; CLG 2; OG 7) were noted to have taken oral analgesics. No demonstrable difference in the number of participants in each group

*Interventions of interest in this study

Outcomes

Primary outcome measure

  • reduction in corneal abrasion area from the time of removal of foreign body (documented by photography)

Secondary outcome measures

  • Pain assessed with a modified (10-point) Wong-Baker FACES Pain Rating Scale

  • Duration of medical leave

  • Presence of residual corneal opacities (yes/no)

  • Use of oral analgesics

  • Visual acuity (Snellen, converted to logMAR)

  • Conjunctival injection score

  • Any serious adverse events (e.g. microbial keratitis)

Notes

With regards to removal of foreign bodies, siderosis, or both, was a bigger epithelial defect created? What proportion of participants required such intervention and at which stage of follow-up? This may confound the assessment of healing process and participant’s assessment of discomfort.

Differences in the size of abrasions was noted. This statistical distribution (both intra and intergroup) was however noted to be homogenous

Variations in the delivery vehicle of ofloxacin (drops vs ointment) noted. However, the two groups that we are interested in (PG and OG) both used ointments, although the frequency of administration differed due to differences in intervention

Populations between treatment arms were not very different. However, it was interesting that only men were recruited into this study. This may impact upon the external validity of the study

Date study conducted: October 2008 to April 2010

Funding: no funding received for this study

Declaration of interest: reported no conflict of interest

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk

Computer based randomisation technique. This randomisation generator randomises subjects to either treatment using a permuted-block randomisation technique.  Proper randomisation technique utilised

Interestingly, only male participants were captured over the duration of this study; a two-year period

Allocation concealment (selection bias)Low riskAllocation to treatment modality was conducted by a study nurse using numbered closed envelopes that were randomised before the start of the study
Masking of participants and personnel (performance bias)
All outcomes
Low riskThe patch was removed by the study nurse 30 min prior to ophthalmic examination at the follow-up visit. In the CLG, a therapeutic contact lens was inserted and the participants instructed to use ofloxacin eye drops 4 times a day. The bandage contact lens was removed by the study nurse 30 min prior to ophthalmic examination
Masking of outcome assessment (detection bias)
All outcomes
Low riskDocumentation of the corneal abrasion area was performed by photography (magnification × 10) using a digital anterior segment camera. The assessment of corneal abrasion area was done by processing the digital photographs by two different ophthalmologists
Incomplete outcome data (attrition bias)
All outcomes
Low riskLow initial drop out rates for initial follow-up. Although the high subsequent dropout rates at 7 days across all 3 treatment arms were initially concerning, with no mention of ITT analysis, subsequent mention of complete corneal abrasion area reduction achieved by day 1  (n = 60) and day 2 (n = 3) demonstrates that participants lost to follow-up on day 7 were not likely to influence the results achieved
Selective reporting (reporting bias)Low risk

All enrolled participants were reported and accounted for. Authors reported both demonstrable and non-demonstrable differences in results

No demonstrable differences were noted in demographics and baseline characteristics between treatment groups. No demonstrable differences between drop-out rates on the second visit, or in both primary and secondary outcome measures were noted

Michael 2002

Methods

Randomised controlled trial with parallel study design.

Duration: 9 months

Method of allocation: block randomisation (groups of 4) placed in envelopes and numbered
Post-randomisation exclusions: 5.4% (n = 2 from non-patch group (1 lost to follow-up, 1 had a retained foreign body causing new abrasions)
Outcomes assessors masked to treatment allocation when measuring per cent healing from photographs. Photographs or eye template diagrams drawn based on slit lamp biomicroscopy findings used to record abrasion size
Potential confounder: different antibiotic ointment dosing regimens in both groups. Differences in analgesic regimens, with paracetamol (acetaminophen) taken for breakthrough pain if ibuprofen was insufficient

Participants

Inclusion criteria

  • Children 3-17 years old

  • Isolated corneal abrasion

  • Ability to return for a follow-up examination

Exclusion criteria

  • Any other ocular trauma

  • Any prior treatment for the corneal abrasion

  • Erythromycin allergy

  • Did not speak English

  • Eye infection, eye injuries or conditions requiring inpatient care

  • Monocular vision

Setting: USA
Number of participants: 37
Exclusions after randomisation: 2
Total available for analysis: 35
Mean age of all participants was 10 years
Sex: 62.9% (n = 22) were men

Interventions
  • Double eye patch with first pad folded in half and placed over closed eyelids, both secured with tape to provide pressure and prevent eyelid opening.

  • All participants received 1% cyclopentolate (one drop).

  • In the patched group, 1 application of erythromycin ointment was applied topically.

  • The non-patched group received erythromycin ointment 3 times per day.

  • Ibuprofen (10 mg/kg, max 400 mg) was administered to all participants every 6-8 hours as required. Paracetamol (acetaminophen) (15 mg/kg max 500 mg) every 4-6 hours for breakthrough pain

Outcomes
  • Per cent healing at 24 hours measured from either digital photograph printouts or eye template diagrams

  • Visual outcomes

  • Analgesia use documented in a pain medication diary

  • Assessment of interference with ADLs either from parent or child on a VAS

  • School attendance

  • Any complications (such as infection, inflammation and increased abrasion size)

Notes

Date study conducted: July 1999 to March 2001

Funding: Katharine B. Richardson Grant Funding

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskAdequate randomisation achieved via the use of permuted block randomisation
Allocation concealment (selection bias)Low riskThe use of concealment in the selection process with allocation via sealed envelopes appears to be adequate
Masking of participants and personnel (performance bias)
All outcomes
Unclear risk

Study personnel were not masked to the allocated interventions. This may be a potential source of bias. However, if most of the participants had their lesions documented photographically rather than from the template, and percentage healing assessed by a masked reviewer, this may mitigate the risk of bias stemming from this issue. However, no information was provided regarding the number of eye template diagrams that had to be used

It is notable that attempts for assessment without introducing interviewer bias were made, including the use of the pain medication diary, which required participants to self-report their analgesic use

Masking of outcome assessment (detection bias)
All outcomes
Low riskPersonnel assessing study participants at follow-up were not blinded to treatment instituted. However, percentage healing from either digital photograph printouts or eye template diagrams was assessed by a masked reviewer. In light of this, we judged this study as having undergone sufficient measures to minimise detection bias
Incomplete outcome data (attrition bias)
All outcomes
Low riskMinimal attrition was noted in this study. In addition, an ITT analysis was adopted with available data
Selective reporting (reporting bias)Low riskIt is notable that attempts for assessment without introducing interviewer bias were made, including the use of the pain medication diary, which required participants to self-report their analgesic use

Patterson 1996

Methods

Randomised controlled trial.

Duration: unknown

Method of randomisation: allocation according to computer-generated table
Post-randomisation exclusions: 34% ("almost equally divided" between groups)
No masking of outcomes assessor. Fluorescein examination under 5-power magnification (no slit lamp) used to document lesion
Potential confounder: different antibiotic preparations and dosing regimens in two groups

Participants

Inclusion criteria

  • Eye pain and documented corneal abrasion (on fluorescein staining)

  • Cause may be foreign body or mechanical disruption

Exclusion criteria

  • Under 12 years of age

  • Significant coexisting eye disease

  • Chemical/thermal injuries

Setting: USA
Number of participants: 50
Exclusions after randomisation: 17
Total available for analysis: 33
Age: NR

Sex: NR

Interventions
  • Double pressure patch, first patch folded in half, placed over closed eyelids, both patches entirely covered with tape

  • Patched group received topical tobramycin ointment before application of patch

  • Non-patched group received tobramycin drops to be used every 4 hours while awake

  • Ketoprofen 75 mg prescribed as required

Outcomes
  • Pain scores VAS

  • Analgesia use - ketoprofen at 24 hours

  • Complete healing at 24 hours

Notes

Date study conducted: NR

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskNo information about the randomisation process undertaken as part of this study was provided beyond describing the use of a computer-generated table. It is assumed that a random generator was used to generate the allocation sequence
Allocation concealment (selection bias)Unclear riskNo further information given on the allocation process
Masking of participants and personnel (performance bias)
All outcomes
Unclear riskNo description of attempts made to mask personnel from treatment allocations made in this study
Masking of outcome assessment (detection bias)
All outcomes
Unclear riskNo information was provided
Incomplete outcome data (attrition bias)
All outcomes
High riskA high rate of attrition at 34% (17 participants) was noted. No breakdown of these participants into either treatment arm was provided, except for the statement that participants lost to follow-up were "almost equally divided" between groups. ITT analysis not performed
Selective reporting (reporting bias)Unclear riskNo discussion of baseline demographic information of the study population

Rao 1994

  1. a

    ADLs - activities of daily living

    ITT - intention-to-treat

    NR - not reported

    VAS - visual analogue scale

Methods

Randomised controlled trial

Duration of trial: unknown
Outcome measurement technique not described
Post-randomisation exclusions: 0%.

No mention of masking of assessors to treatment allocation

Unclear regimen of antibiotic ointment and mydriatic drops

Slit lamp biomicroscopy to assess location and size of the corneal abrasion

Participants

Inclusion criteria

  • Patients with corneal abrasions

  • No exclusion criteria described

Setting: UK
Number of participants: 40

Exclusion post-randomisation: 0

Total available for analysis: 40

Interventions
  • Firm padding of the eye.

  • Both groups received topical cyclopentolate 1% and chloramphenicol ointment 1%

  • Paracetamol 500 mg with dosing depending on participants' requirements for analgesia

Outcomes
  • Dimensions of corneal abrasion

  • Maximum or minimum length on days 1 and 2

  • Daily pain scores on vertical VAS

  • Analgesic use (quantity)

Notes

Date study conducted: NR

Funding: NR

Declaration of interest: NR

Trial registration number: NR

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo description of the randomisation process undertaken in sequence generation for this study
Allocation concealment (selection bias)Unclear riskNo further information was provided as to whether any allocation concealment was performed
Masking of participants and personnel (performance bias)
All outcomes
Unclear riskNo information about participant and personnel masking was given by the authors
Masking of outcome assessment (detection bias)
All outcomes
Unclear riskAlthough attempts were made to reduce the effect of responder bias with respect to discomfort level, we are not told if the observer involved in assessing healing of the corneal abrasion was masked to the two treatment arms. Steps to mask assessors to the interventions performed in either group were not described in this article
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo patients were lost to follow-up
Selective reporting (reporting bias)Unclear risk

No discussion of baseline demographic information of the study population was provided.

The authors appear to have reported the various parameters that were examined as part of this study. However, attrition rates and how this data (if any) was subsequently handled were not addressed in the study. This could potentially skew the results obtained and result in reporting bias

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Alper 1997Letter - no new data presented
Anonymous 2001Letter - no new data presented
Daugherty 2002Mini-review - no new data presented
Douglas 1999Letter - no new data presented
Easty 1993Letter - no new data presented
Gregersen 1991There is no control 'no-patch' group. Two different types of patching are compared to each other in the randomised trial
Hart 1997Retrospective chart review/audit - no new data presented
Health 1996Letter - no new data presented
Hirst 1997Letter - no new data presented
Jampel 1995Letter - no new data presented
Kurt 2003Randomised controlled trial that included participants who had received an abrasion up to 7 days prior to being included in the study. The mean time to seek ophthalmological attention was 2.1 days. This means the study falls outside the study criteria defined as being abrasion of recent onset (< 48 hours)
Le Claire 1996Letter - no new data presented
Mackway-Jones 1999Mini-review - no new data presented
Melton 1991Letter - no new data presented
Mindlin 1996Letter - no new data presented
Olson 1996Letter - no new data presented
Perlman 2000Letter - no new data presented
Rogers 1995Letter - no new data presented
Sabiston 1972Letter - no new data presented
Sabri 1998National survey of corneal abrasion treatment - no new data presented
Schechter 1997Letter - no new data presented
Seiff 1996Letter - no new data presented
Slawson 1996Letter - no new data presented
Soli 2001Letter - no new data presented
Solomon 2000Patching the eye is compared with no patching, however the groups differ significantly. The no-patch group is treated with an alternative therapy (topical indomethacin)
Spitz 1997Letter - no new data presented
Yamada 2001A meta-analysis of previous studies. No new data presented

Ancillary