Assessment of Firework-Related Ocular Injury in the US

Key Points

Question  What are the factors associated with firework-related ocular injuries in the US?

Findings  In this cross-sectional study of a nationally representative hospital database, nearly 2000 firework-related ocular injuries occurred annually between 1999 and 2017, with most injuries occurring in July and January. Ocular burns were the most frequent type of injury, and bottle rockets were a common firework type that disproportionally caused serious ocular injury, including ruptured globe.

Meaning  These findings may inform preventive methods to decrease firework-related ocular morbidity, especially around national holidays.

Importance  Fireworks are popularly used for recreation but can lead to potentially blinding injuries. Studies quantifying the trend and national prevalence of firework-related ocular injuries are limited.

Objective  To characterize firework-related ocular injuries treated in emergency departments (EDs) in the US from 1999 to 2017.

Design, Setting, and Participants  This cross-sectional study used the National Electronic Injury Surveillance System, a stratified probability sample of more than 100 hospital-affiliated US EDs representing more than 5300 nationwide hospitals. Deidentified individuals in the National Electronic Injury Surveillance System database who experienced eye injury due to fireworks between January 1, 1999, and December 31, 2017, were included. Analysis began February 2019.

Main Outcomes and Measures  Annual prevalence of firework-related ocular injury, firework type stratified by demographic information, diagnosis, event location/date, and patient disposition.

Results  A total of 34 548 firework-related ocular injuries were seen in US EDs during the 19-year study period (estimated from 1007 individuals in the National Electronic Injury Surveillance System database). Overall, 664 individuals (65.9%) were 18 years or younger, 724 (71.9%) were male, and 512 (50.8%) were white. Twenty-eight patients (2.8%) sustained a ruptured globe, while 633 (62.9%) sustained ocular burn injuries, 118 (11.7%) had ocular foreign bodies, 97 (9.6%) had conjunctival irritation, and 46 (4.6%) experienced other severe eye trauma. Of 1007 individuals, 911 (90.5%) were treated and released without transfer, while 87 (8.7%) required admission or transfer to another hospital. The most common specified firework types included firecrackers (193 [19.2%]), bottle rockets (177 [17.6%]), sparklers (88 [8.7%]), roman candles (66 [6.6%]), and novelty devices (65 [6.5%]) (eg, poppers and snappers). Bottle rockets disproportionately caused the most severe injuries, including ruptured globe (odds ratio, 5.82; 95% CI, 2.72-12.46; P < .001). A total of 411 cases (74.9%) occurred at home. Injuries most commonly occurred near the time of Independence Day and New Year’s Day: 707 patients (70.2%) presented in July, 74 (7.4%) in June, 101 (10.0%) in January, and 47 (4.7%) in December.

Conclusions and Relevance  These findings support that firework-related ocular injuries range from mild irritation to ruptured globe, and most occur frequently in those who are young, male, and white. Focused preventive methods and regulations may be imperative in decreasing fireworks-related ocular morbidity, namely from bottle rockets and especially near national holidays.

Fireworks are commonly used for festive or recreational purposes in the US and are responsible for a significant number of hospital visits, disproportionately affecting children and men.^1-5 It is estimated that 97 562 firework-related injuries have been treated in emergency departments (EDs) within the US between 2000 and 2010, with 42% of injuries affecting the head and neck region.⁵ Fireworks also cause significant morbidity in Europe^6-9 and Asia.^10-14 The most recent report by the US Consumer Product Safety Commission estimated that 12 900 firework-related injuries occurred in 2017, of which ocular injury comprised an estimated 1538 (12%).¹ Ocular firework injuries can result in partial or complete permanent vision loss and are a serious public health concern.^15,16 Previous similar studies have either not focused on ocular injury,⁵ not used national data,^4,17 or did not relate firework type with ocular injury.¹⁸

The purpose of our study was to gather national epidemiologic data specific to ocular injuries from fireworks to discern specific, actionable associations that could effectively guide public policies and reduce the extent of firework-related ocular injuries.

The source of data for this study was the Consumer Product Safety Commission’s National Electronic Injury Surveillance System (NEISS), a database that records ED visits for consumer product-related injuries. NEISS data are derived from a stratified probability sample of more than 100 hospitals in the US, representing more than 5300 nationwide hospitals with at least 6 beds and a 24-hour ED. The recorded data include patient age, sex, injured body part, diagnosis, location of incident, case disposition, and a brief narrative of the event. Visual acuity, follow-up information, and comorbidities are not included in the NEISS. This study was exempt from institutional review board approval and no ethical approval was required because it used existing, publicly available, deidentified data. Patient permission was also waived for these reasons.

Our study queried the NEISS to identify all cases with product code 1313 (firework-related injuries) from January 1, 1999, to December 31, 2017, subcategorized for code 77 (eyeball injuries). Orbital and periorbital soft tissue injuries were not included. Narratives were reviewed for confirmation of event and to obtain additional information regarding types of ocular injury and firework used. Eye injury visits were stratified by age, month, race/ethnicity, sex, diagnosis, case disposition, and location of incident. Pearson χ² was used to analyze associations between variables when appropriate. To determine the strongest contributors to the observed frequency distribution, adjusted residuals (ARs) were calculated, with AR less than −1.96 and greater than 1.96 considered statistically significant. For post hoc analyses, a Bonferroni-corrected 2-sided P value less than .025 was used for statistical significance (corrected for 2 outcomes). Analyses were performed using GraphPad Prism, version 7.0d (GraphPad Software). Analysis began February 2019.

A total of 1007 injuries in the NEISS database represented an estimated 34 548 firework-related ocular injuries in US EDs from 1999 to 2017, an average of 1840 injuries per year. Figure, A, demonstrates the annual trend in national estimate of firework-related ocular injuries. The spike in 2000 is possibly due to millennium celebrations.¹⁹ Patient demographics are depicted in Table 1. The mean (SD) age of patients was 17.3 (13.2) years, and 664 (65.9%) were younger than 18 years, compared with an overall US population rate of 22.4% younger than 18 years.²⁰ Most patients were male (724 [71.9%]) and white (512 [50.8%]), compared with US rates of 49.2% male and 76.5% white individuals.²⁰ Ocular burn was the most common ocular injury (633 [62.9%]), followed by ocular foreign body (118 [11.7%]) and conjunctival irritation (97 [9.6%]) (Table 2). Twenty-eight individuals (2.8%) sustained ruptured globes and 35 (3.5%) had hyphema. Most cases occurred at home (411 [74.9%]), and most patients were treated and released without transfer (911 [90.5%]).

Overall, 32.1% (n = 323) of fireworks causing injury were of an unspecified type. The next most common specified firework type included firecrackers (193 [19.2%]), bottle rockets (177 [17.6%]), and sparklers (88 [8.7%]). On χ² analysis, the distribution of firework type was associated with diagnosis (Table 3). Firecrackers were significantly associated with ocular burn (AR, 2.1), while public displays were significantly associated with foreign bodies (AR, 6.6) and sparklers with conjunctival irritation (AR, 2.5). Rocket-type fireworks were significantly associated with ruptured globe (AR, 5.1; odds ratio, 5.82; 95% CI, 2.72-12.46; P < .001) and other forms of severe eye trauma such as hyphema and corneal injury (AR, 9.3; odds ratio, 5.71; 95% CI, 3.85-8.47; P < .001) (all severe eye trauma: odds ratio, 6.78; 95% CI, 4.66-9.85; P < .001).

Type of firework was also associated with month of injury and age group (Table 3). Injuries most commonly occurred near Independence Day and New Year’s Day (Figure, B). January featured a significantly higher proportion of firecracker injuries (AR, 2.6), while injuries in July were disproportionately caused by roman candles (AR, 2.1), public displays (AR, 2.4), and large fireworks (eg, M-80 explosives and mortars) (AR, 2.5). Novelty (eg, poppers and snappers) and homemade devices were significantly associated with injury in other months (AR, 5.4). Sparklers were most likely to injure children younger than 5 years (AR, 8.1), bottle rockets primarily injured individuals aged 10 to 18 years, public displays and large fireworks were disproportionally recorded in those older than 31 years, and novelty/homemade devices were associated with injuries in children younger than 10 years.

Using a national database, we found that an estimated 1840 firework-related ocular injuries occur per year. No significant change has occurred in the annual rate of all firework-related overall injuries in the last 15 years.¹ Similar to previous studies, most injuries in this study occurred in July and December and affected younger individuals and male individuals.^4,16,21-23 The most common injuries were ocular burns, ocular foreign body, and conjunctival irritation. The majority of injuries occurred at home, which is not surprising given that ocular injuries from professional firework displays are rare,¹⁶ comprising 4.4% of injuries in our cohort.

We observed lower rates of hyphema and ruptured globe compared with other reports.^4,21,24 Sandvall et al¹⁷ reported that homemade fireworks and shells/mortars accounted for the majority of morbidity caused by fireworks. Although we did not capture nonocular causes of morbidity, bottle rockets were nearly 7 times more likely to cause severe eye trauma compared with other fireworks types, possibly owing to their unpredictable flight patterns. Bottle rockets also disproportionately caused injury in children aged 10 to 14 years.

Despite their dangers, consumer-grade fireworks are widely available for legal purchase in the US and are associated with significant morbidity and mortality, most commonly injuring the hands, head, and eyes.^1,4,5,16 Building on previous studies, our use of ARs helps to reveal more injurious (or mild) firework types in the interest of policy development, with more positive AR indicating stronger correlation. For instance, sparklers resulted in a large number of injuries to young children (AR = 8.1) but most frequently caused conjunctival irritation (AR = 2.5). While simple eye protection or other precautionary measures may be indicated for children, actual legislation to limit sale and use might be necessary to prevent severe eye trauma caused by bottle rockets (AR = 9.3).

The effects of increased regulation to prevent the use and purchase of consumer fireworks on incidence of related injury has been well documented in non-US countries.^16,21,23,25 Countries that have banned the sale of all fireworks have lower rates of firework-related injuries.^22,25 Eye protection, although shown to reduce the incidence of firework-related ocular injury, is not mandatory for US operation.^22,25 In Norway, free provision of protective eyewear decreased the incidence of firework injuries, albeit not to a statistically significant extent, whereas a ban on bottle rockets halved the incidence of injuries.²⁵

Despite their capacity to cause serious/permanent injury, restrictions on sale of fireworks vary by US state.²⁶ While the Consumer Product Safety Commission has previously regulated powder amount, fuse time, and warning labels, many states maintain lax regulations: Illinois, Ohio, and Vermont ban all but sparklers and/or novelty devices, and only Massachusetts bans all consumer fireworks.²⁷ However, not only have regulations loosened over time,^28,29 but consumers are often able to cross state lines to purchase fireworks given sparse enforcement.³⁰ Additional federal laws may be difficult to enact considering consumer culture. Therefore, future state-level policy may need to focus on restricting the use of specific firework types (eg, bottle rockets) that are more likely to cause ocular injury as well as to incentivize use of eye protection.

One limitation of this study is that we queried the NEISS for cases in which fireworks injured the eyeball only, excluding cases in which other parts of the head were primarily injured but may have featured ocular injury. Therefore, our results may underestimate the full extent of ocular injury. Further, this national database did not include visual acuity, clinical examination, or follow-up data, precluding us from commenting on patient morbidity or vision loss. Poor outcomes have been associated with open globe injury, poor initial visual acuity, intraocular foreign bodies, retinal detachment, and endophthalmitis.⁴ Such cases constituted at least 3% of our cohort, although this is likely an underestimate, given that 9% required inpatient services or hospital transfer. The study methodology also relies on limitations of the NEISS database itself, including categorization, summarization of data by nonophthalmologists, and whether or not victims were bystanders. We lacked state-specific NEISS data; future research evaluating the effect of regional legislation on annual injury incidence is warranted.

Annual rates of all firework injury have not significantly changed in the last 15 years. Eye injuries from fireworks disproportionately affect younger people and male individuals, most frequently occur around national holidays, and are most commonly ocular burns. Compared with other firework types, bottle rockets are nearly 7 times more likely to cause severe eye trauma but are variably prohibited, representing an opportunity for regulation.

Corresponding Author: Natasha Nayak Kolomeyer, MD, Glaucoma Service, Wills Eye Hospital, 840 Walnut St, Ste 1100, Philadelphia, PA 19107 (nkolomeyer@willseye.org).

Accepted for Publication: February 23, 2020.

Published Online: April 9, 2020. doi:10.1001/jamaophthalmol.2020.0832

Author Contributions: Mr Shiuey and Dr N. Kolomeyer had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: A. Kolomeyer, N. Kolomeyer.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: All authors.

Administrative, technical, or material support: A. Kolomeyer.

Supervision: A. Kolomeyer, N. Kolomeyer.

Conflict of Interest Disclosures: None reported.