Article Text

Download PDFPDF

Ocular firework trauma: a systematic review on incidence, severity, outcome and prevention
  1. R P L Wisse,
  2. W R Bijlsma,
  3. J S Stilma
  1. Resident Ophthalmology, Department of Ophthalmology, University Medical Center Utrecht, The Netherlands
  1. Correspondence to Robert P L Wisse, Department of Ophthalmology, University Medical Center, Utrecht, HP E.03.136, PO Box 85500, 3508 GA, Utrecht, The Netherlands; r.p.l.wisse{at}


Aim To provide a systematic review on ocular firework trauma with emphasis on incidence and patient demographics, the extent of ocular trauma and visual function loss, and firework regulation effects on injury rates.

Methods A literature search was performed using predetermined inclusion and exclusion criteria. Demographic characteristics of ocular firework casualties were obtained and incidence rates of sustained trauma and vision loss calculated.

Results Twenty-six relevant articles were suitable for calculation of trauma incidence and patient demographics, of which 17 articles could be used for calculating trauma severity and vision loss. Victims were male (77%), young (82%) and often bystander (47%). Most of the trauma was mild and temporary. Penetrating eye trauma, globe contusions and burns accounted for 18.2%, with a 3.9% enucleation rate. Mean visual acuity was >10/20 in 56.8%, with severe vision loss (<10/200) in 16.4%. Countries using restrictive firework legislation show 87% less eye trauma (p<0.005).

Conclusions One in six ocular firework traumas show severe vision loss, mostly in young males. Bystanders are as frequently injured. Firework traumas are a preventable cause of severe ocular injury and blindness because countries using restrictive firework legislation have remarkable lower trauma incidence rates.

  • Fireworks
  • trauma
  • review
  • prevention
  • eye
  • public health
  • epidemiology

Statistics from

The use of fireworks for celebrative purposes is associated with firework casualties. Eye trauma contributes to an estimated 18% of the total number of firework injuries.1 Firework trauma predominantly occurs in young males, with potential devastating effects on vision, cosmetics and professional functioning. The burden of severe ocular and facial firework trauma carries along substantial costs for society.2 In this report we wish to provide background information on firework use and its healthcare implications.

The use of fireworks to commemorate national and religious events is a well-known phenomenon. New Year's Eve is globally celebrated with firework festivities and numerous other occasions are known, such as the Olympic games (Beijing 2008), Independence Day (USA), Guy Fawkes Night (Commonwealth), Fallas (Spain), Hari Raya (Malaysia), Mawlid and Eid al-Adha (Muslim countries), Charshanbe-Soori (Iran) or Purim (Jewish festivity).

The extent of the problem of ocular fire work trauma is not well known. Most Western countries use a trauma incidence registration system; some categorise by causative agent, some by injured body part. Examples are the US Consumer Product Safety Commission, Consumer Safety Unit of the Department of Trade and Industry (UK), National Trauma Registry Consortium (Australia/New Zealand) and the Consumer and Safety Foundation (The Netherlands). The National Electronic Injury Surveillance System (NEISS), maintained by the US Consumer Product Safety Commission, is arguably the best documented and best accessible source.1 Its estimates are stratified for ocular trauma. The Dutch trauma registration estimates the total number of firework casualties and does not stratify per affected body part.3 The Australian trauma registration includes severe bodily traumas only.4

Firework trauma registration systems do not report on trauma severity or outcome, and incidence figures are often estimates. Furthermore, the usefulness of most databases for determining ocular firework trauma is rather low because of poor classification. Numerous articles and reports have been published in the last decades on the nature of ocular firework trauma, the most being strictly national. International data on firework incidence rates, extent of ocular trauma and visual outcome are lacking.

The purpose of this study is to conduct a systemic review on ocular firework trauma with emphasis on (1) patient demographics and injury incidence, (2) the extent of ocular trauma and visual function loss and (3) the effect of firework regulation on injury rates.


Search strategy

We identified articles using a systematic search in MEDLINE (Pubmed), EMBASE and CENTRAL (Cochrane Library) databases up to 22 July 2009. We additionally searched online trial registers by index (,, and proceedings of the American Academy of Ophthalmology, Association for Research in Vision and Ophthalmology and European Society of Ophthalmology annual meeting abstracts, from 2006 onwards.

Search terms used were derivatives of “eye” AND “trauma” AND “fireworks”. The following search string was used for the MEDLINE search: (ocular[tiab] OR orbit*[tiab] OR eye[tiab] OR globe[tiab] OR adnex*[tiab]) AND (trauma*[tiab] OR blast[tiab] OR explosi*[tiab] OR injury[tiab] OR damage*[tiab] OR thermal[tiab) AND (firework[tiab] OR fireworks[tiab] OR crackers[tiab] OR fire*[tiab]). Analogous search queries on title and abstract were used for EMBASE database. The Cochrane library was searched on index. References from relevant articles were hand searched.

Governmental and legislation reports were derived from relevant articles and were further hand searched. National trauma incidence registers were contacted for data on most recent ocular firework trauma incidence rates.

Eligibility criteria

Included articles had to supply information on firework trauma to the eye or its adnexa. For general firework trauma publications to be considered, the ophthalmic component should be adequately described. Paediatric studies were included. To be eligible for qualitative analysis, the extent of trauma and visual acuity had to be described. Excluded from analysis were non-firework injuries, occupational injuries, wound classification studies, operation techniques and case reports. No language restriction was applied.

Data collection

Data collection was conducted using a standardised extraction form (by RW). Poorly definable articles were reviewed by a second investigator (JS). We did not contact authors of the original studies in the case of missing data.

For each included article, we extracted information on author, publication year, region, study design, inclusion time, total number of patients, percentage ocular involvement, age, sex and percentage of bystanders/spectators. Patient characteristics (sex, age, percentage of bystander) were extracted from studies supplying adequately described data. Concerning age, studies used different notations. Most studies reported on percentage or mean number of youth involved, but cutoff ages were documented differently. We collated these in a broad patient group called “young”.

For articles eligible for qualitative analysis, we furthermore extracted information on sustained injury rates (enucleation/evisceration, globe rupture, intraocular foreign body, penetration/perforation, corneal/adnexal burn, contusion, corneal abrasion) and visual acuity (>10/20, 10/200–10/20, counting fingers/hand movements/light perception, no light perception (NLP)). Mean incidence rates were calculated using a weighted average by study size. Because all studies used dissimilar stratifications and cutoff points for visual acuity, categorisation was indispensable. A uniform trauma grading system, like the Birmingham Eye Trauma Terminology system, was not used in either study.5 Categories were created to minimise patient dropout due to categorisation impossibilities. Sustained injuries were reported to conform to study design; some articles only graded most severe trauma, others graded all traumas. For all geographical regions covered by included articles, legislation on firework use was summarised together with local incident rates when available.

Statistical analysis

Trauma incidence rates in different regions were compared to statistically assess legislation effects. Incidence rates were extracted from included literature and national trauma registration systems and were grouped according to the legislation status per region (restrictive vs permissive). Only incidence rates concerning ophthalmic traumas were used. The independent-sample t test was used for comparing means of ophthalmic trauma incidence.


Search results

Literature search revealed 105 original articles and incident registrations, of which 26 were eligible for data collection on patient characteristics and demographics.6–31 Seventeen of these articles reported adequately on sustained injury and visual acuity and were used for qualitative analysis.9 11 12 15–20 22–24 26–30 For details on article recruitment and selection, see figure 1. One relevant conference presentation on ocular firework trauma was found, concerning counselling of victims and family by Morris et al.32 No quantitative data could be extracted from this conference abstract.

Figure 1

Results of systematic literature search.

Quantitative data

Patient demographics

The number of described ophthalmic firework casualties ranged from 20 to 2156 per study, with 7742 cases in total. Victims were male (75%, range 66–95%), young (67%, range <14 to 30 years old) and in 47% of cases bystander (range 34–80%).

Study recruitment time ranged from 1 month to 22 years. Eleven studies reported on general firework trauma in which eye trauma was adequately described. Only six studies performed a follow-up, ranging from 1 month to 9 years. On average, eye trauma contributed 21.8% of general firework casualties (range 16–45%). Results are shown in table 1.

Table 1

Patient demographics, study size and combined quantitative data

Incidence rates

Obtaining incidence rate estimates for ocular firework traumas was only possible from the US National Electronic Injury System and equalled 0.43/100 000 (2008).1 NEISS data were not classified per state. Australian and Dutch trauma incidence registries only supplied data on general firework trauma rates and are estimated 0.18/100 000 in New South Wales, Australia, and 6.7/100 000 in The Netherlands (2008).3 33

Several included studies yielded information on ocular firework trauma: Sundelin (2000) reported an annual incidence in Western Sweden of 1.0/100 000, Lee (1966) a 6-year steady average of 1.3/100 000 in Hong Kong, Wilson (1982) a 2.6/100 000 incidence on yearly basis in the USA, a 0.2/100 000 incidence for West Virginia, a 0.3/100.000 incidence for Georgia and a 2.2/100 000 incidence for Alabama. Faber (2009) reported a 4.8/100.000 incidence around Dutch New Year's Eve, Vernon (1988) a 0.25/100 000 incidence covering a 2-week period around Guy Fawkes Day in the UK and Kuhn (2000) compared US and Hungarian registries reporting incidences of 4.4/100 000 and 0.16/100 000, respectively.11 17 18 26 29 30

Qualitative data

Extent of ocular trauma

Trauma ranged from mild conjunctivitis to complete loss of the eye. Table 2 shows trauma rates. Most of the traumas were mild and temporary, with highest incidence rates for corneal abrasions (42.2%) and globe contusions (25.9%). Severe ocular trauma occurred in 18.2% of cases on average, with a 3.9% enucleation rate. Table 2 details on percentages and number of cases.

Table 2

Trauma severity

Vision loss

Visual acuity in included studies was graded inconsistently, necessitating the creation of six distinct visual acuity categories. Seven studies reported on vision loss only, without supplying data on the extent of this vision loss.11 16 18 20 24 26 29 The weighted average of >10/20 vision was 56.8% (range 23.0–93.0%). Severe vision loss (<10/200) was found in 342 of 2092 cases (16.4%), including NLP in 96 of 1602 cases (6.0%). If no incidence data were given on NLP, enucleation rate was used. See table 3 for detailed information.

Table 3

Vision loss

Firework regulation and its effect

Regulation per country

Where global effort is deployed to unify laws on firework production, storage and sales, international legislation on consumer handling of fireworks varies widely. Apparently, there is no agreement on federal (USA) or member state (EU) level. Firework device classification is extensive and varies per country. In general, legislation describes four main categories, ranging from small fireworks such as sprinklers to professional firework devices. The amount of explosive powder is the main criterion for categorisation.

Table 4 gives an overview of firework legislation in different countries. All legislation communications were found on the internet, using local language if no English alternative was available. US legislation differs widely per state, with central states in general being more permissive to the personal use of fireworks. Sale periods and maximum allowed weight of explosive powder per device are more harmonised. Complete firework bans are found in Hungary, Ireland, Australia and the Northeast USA (including New York, New Jersey and Massachusetts).

Table 4

Firework legislation per country

Causative device

Correlations between trauma rate and type of responsible firework device are unequivocal. Several studies report no apparent differences per causative agent,6 11 13 15 21 26 27 four studies reported more injuries due to firecrackers and ground explosives,20 23 28 31 whereas three studies showed a markedly higher trauma rate due to rockets and aerial devices.16 17 30

Effect of firework legislation

Countries or states using restrictive firework laws concerning the personal use of firework show a significant lower incident rate of ocular firework trauma. Mean injury rate is 0.35/100 000 (SD 0.28) in restrictive regions compared to 2.70/100 000 (SD 1.42) in permissive regions (p<0.005). Restrictive regions show 87% less ophthalmic firework trauma. For details on incidence distribution, see figure 2.

Figure 2

Legislation effects on firework trauma incidence per region.

Berger et al6 found a 7.3 times higher incidence of firework-related injuries in US states that allow a wide variety of fireworks compared to states using restrictive laws. McFarland et al21 reported a threefold increase in firework trauma (9.7 vs 3.1/100 000) after allowing category III fireworks in the Washington State. The Dutch Consumer and Safety Foundation (2007) reported an increase in firework trauma from 4.0/100 000 to 6.7/100 000 after increasing the maximum allowed explosive powder from 250 to 500 g in the Netherlands.3


This systematic review shows that (1) firework traumas occur predominantly in young men, with bystanders accounting for half of injuries, (2) that most traumas are mild and temporary, but on average one in six casualties sustains eye trauma with severe vision loss (<10/200), and (3) that countries using restrictive firework laws show a 87% lower trauma incidence rate. Sporadically, bilateral blindness is the aftermath of an accident with an apparently safe firework device.

Demographic outcomes show comparable characteristics among different studies. In addition, the extent of sustained ocular trauma does not show much variation. This suggests that people sustaining firework injuries are much alike, regardless of nationality or local legislation. No correlation between bystander rate and trauma severity was seen.26 Eye trauma contributes 22% of general firework traumas, a consistent finding in several studies and comparable to US NEISS data.

Standardised reporting

It is recommended to use a unified registration like the Birmingham Eye Trauma Terminology system,5 so that international data can be compared effectively to increase data reliability. Especially, case series concerning patients recruited in a single ophthalmic unit are prone to referral bias. One could expect that heavier ocular trauma is referred to such a tertiary treatment centre. Web-based registration could enhance evaluation, international application, increase data quality and reduce bias.36

The effect of firework regulation on injury rates

Causative firework device

Internationally, thousands of different legal firework devices exist, not to mention the innumerable illegal homemade fireworks. Much can be said on the hazards of specific pyrotechnic devices; most countries use extensive legislation on firework classification, handling, storage and sales. Governmental reports on relationships between responsible fireworks (explosives vs aerial devices) and trauma are unequivocal. Contradictive legislation exists where in one country aerial devices are banned and explosives allowed (eg, Norway), and in the other country explosives are banned and aerials allowed (eg, South Africa, New Zealand). Considering a fireworks device regional popularity pertaining to its corresponding trauma rate, our data show no single device should be regarded safer over the other.

Existing firework legislation

The existing firework legislation has its effect as is demonstrated by US figures showing a decline in injury rate per million consumed pounds explosives, whereas firework consumption increased tenfold since 1976. This suggests that fireworks are safer constructed and safer used. Nevertheless, the estimated US firework injury rate remained at the same level as 30 years ago.1 37 The effect of a broad community awareness campaign is demonstrated by D'Argenio et al10 showing the number of firework casualties reduced by 50% in Napels, Italy. Chan et al7 reported a threefold increase after lifting restrictive firework laws in North Ireland. Allowing heavier-class fireworks is associated with increased numbers of severe trauma as demonstrated by Berger et al, McFarland et al and the Dutch Consumer and Safety Foundation (2007).3 6 21

Restricting the personal use of fireworks

The main contributing legislation factor for reducing firework traumas is the installation of restrictive laws on personal use of fireworks. Our data show that regions using restrictive firework legislation have an 87% lower ophthalmic injury rate (p<0.005).

Study limitations

This study has a few limitations. First, obtaining viable governmental data on firework injury rates is difficult. Dedicated internet sites often lack a good index or search option, and most legislation is written in local language. As a result we were unable to investigate incidence rates and legislation for all countries. Available trauma registry data are often estimates instead of actual reported cases of firework injury.

Furthermore, the methodological quality of included studies ranged from thoroughly designed prospective incidence studies to retrospective case series with small numbers. Trauma severity and visual acuity was graded inconsistently, necessitating the classification of data in distinct categories. Inevitably, valuable data were lost and the number of cases per category became fairly small compared with total number of included cases.

Conclusion and recommendation

Restricting the personal use of fireworks is a proven method to greatly reduce the number of firework casualties. A change in firework legislation demands political responsibility and certainly broad community support. Myriad aspects play a role in decision making; environmental issues, costs to society, incurred property loss, law enforcement, firework industry employment, etc. Our responsibility as a medical community is supplying data on the health risks of consumer fireworks and feeding the ongoing discussion on firework legislation with evidence-based arguments.


Preliminary results of our systematic review, titled “Plea for a Ban on Private Use of Fireworks”, were presented at the 203rd Annual Meeting of the Dutch Ophthalmic Society in Groningen on 25 March 2009.



  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.