An explosive blast has sudden and devastating effects and usually causes great casualties[1]. Explosion-related eye injuries are a common cause of morbidities for survivors, whether the injury occurred during wartime, a disaster, or an explosion-related accident during peacetime[2-6]. Explosions rapidly generate impact waves with dramatic power, producing instantaneous peeling, implosion, and hemodynamic effects, and pressure differences (called primary or level Ⅰ explosion injuries); fragments with high kinetic energy (leading to secondary or level Ⅱ injuries); huge airflow capable of overturning the human body (leading to level Ⅲ injuries); and hyperthermic or high-pressure chemical reaction (leading to level Ⅳ injuries)[4,5,7]. The eyeball is extremely vulnerable to explosive injuries since it is an exposed and incompressible spherical organ full of liquid and rich in vascular networks, with fragile tissues and fine structures. Explosion-related eye injury can cause impaired visual acuity (VA) and even blindness[8]. Owing to the diversity and complexity of explosive injury mechanisms, the difficulty, and uncertainty of diagnosis and treatment are greatly increased in patients with ocular explosive injuries. Thus, identifying the mechanisms and classification of ocular blast injuries in detail is particularly important for timely and accurate treatment to minimize the visual disability rate. The present study was based on a multicenter review of inpatients who experienced explosive injuries. We thoroughly analyzed the mechanisms and classifications of explosion-related ocular injuries, compared the differences among various trauma categories, and assessed meaningful risk factors that affect the visual function outcomes.

Explosions can result in various degrees of injuries distributed across a wide range due to different mechanisms initiated by the blast[1,7]. Generally, injuries caused by explosion involve significantly more binocular trauma (29.96% in the present study; 3.33%–72.91% in previous reports[4-8,12,13]) compared with those related to ordinary causes (0%–2.13%)[14-17]. The number of eyes involved seems to be related to the type and power of the explosives, for example, 3.33%–16.32% for firework-related injuries[6,12] and 72.91% for injuries related to explosive military ammunition[13]. This explosive difference also leads to different types of MEI. In this study, OGI (52.57%, 716/1362) was more common than CGI (47.43%, 646/1362). In previous studies, the OGI ratio (56.60%–81.25%) was much higher than the CGI ratio (18.75%–43.4%), some of which might be related to the fact that most of the explosives involved were bombs, weapons, and mines[2,6-8,12,13]. Conversely, the proportion of CGI (64.50%–67.34%) was higher than that of OGI (26.53%–35.5%) in fireworks-related injuries[6,12]. In cases of ocular trauma derived from common causes, CGI was more frequent than OGI[14,18,19] with a few exceptions[20]. The causes and sites of explosion injuries, positional relationship between the injured and the explosives, differences in national or regional working conditions and living standards, and statistical range of eye injury (i.e., simple eye appendage injury or not) are possible reasons for the observed differences. Frimmel et al.[21] reported that chemical (45%) and thermal burns (26%) (level Ⅳ injuries) were far more common than MEI in firework-related eye injuries, probably because the majority of injuries in this study were eyelid and conjunctiva burns and did not involve the eyeball.

The ocular trauma classification can be analyzed and interpreted based on the mechanism of explosive injury. Primary (level I) explosion injuries develop when the explosive shockwaves produce pressure differences followed by an implosion effect in the eyeball. The maximum IOP during an explosion can reach a peak of 0.29 MPa (2175 mmHg) at 1.63 ms, which is two times larger than the physiological IOP in healthy eyes[22]. Karimi et al.[23] suggested that IOP could reach as high as 15,000–17,000 mmHg if ground blast reinforcement effects were considered in their three-dimensional (3D) finite element model of ocular blast injury. Blunt force crushing the eye can lead to contusion (a type of CGI) followed by rupture (a type of OGI) of the eyeball when its pressure limit is exceeded. In our study, contusion and rupture accounted for 41.06% of all injured eyes and contusion was more common than rupture, similar to some literatures[5,14]; however, bomb explosions cause more ruptures than contusion injuries[7]. Other factors that influence the blast effect include the peak overpressure achieved, its duration, distance from the explosion, and whether the eyes were pointed towards the explosion or not. Secondary (level Ⅱ) explosion injuries develop when fragments from the explosives or exogenous fragments pushed by the explosion tear the cornea or sclera (laceration of OGI, including IOFB, penetration and perforation), cut the eyeball surface and wall (lamellar laceration or superficial foreign body, types of CGI) or cause blunt contusion or rupture. Regarding action time, the speed of the primary blast wave was considerably higher and could reach the eyeball sooner than the debris driven by it. If the fragments reached the eyeball, the injury intensity was dramatically increased[24]. Therefore, level Ⅱ explosion injuries are secondary to level I injuries in mechanism, order, or both. Moreover, level Ⅱ explosion injuries cannot exist independently in theory and are considered compound injuries of different forms and degrees. In our study, IOFB (27.26%, 395/1449) was the most common type of eye injury, and other types accounted for only a small fraction. When specifically associated with war and weapons use, the proportion of IOFB among previous reports also differed (9.33%–54.20%)[2,7,13,20,25], and its proportion in OGI was as high as 46.32%–95.70%[13,26] (compared with 55.17% in our study). In ocular trauma from common causes, the proportion of IOFB (1.0%–21.3%) was lower than that of explosive injuries[14,16,27]. In a grade Ⅲ explosion injury, the human body is overturned by the explosion airflow or is hit by other objects. Our data were consistent with the classification for explosive injuries. Contusion, rupture, and IOFB were the injury types most closely associated with level Ⅰ and level Ⅱ explosion injuries.

The ocular injury zone location according to OTCS (in which the eyeball tissues are divided according to repair ability and accessibility) is closely related to the prognosis of visual function. The ocular wall involved in the zone Ⅰ can be repaired, whether the injury is an OGI or CGI. Because the cornea is a refractive medium, corneal damage affects VA, which can be improved by certain treatments, such as corneal transplantation. Zone Ⅱ involves the iris, ciliary body tissue, and lens, which affect the pupil and IOP. Intraocular lenses can make up for the abnormal loss of function of the natural lens. The retina is the nerve layer responsible for eye visual perception and transmission, and is the key tissue involved in zone Ⅲ injuries, which are difficult to repair. Low IOP caused by ciliary body injury and visual function damage caused by retinal injury are the most stubborn and difficult injuries to cure and improve. A poor prognosis is generally associated with zone Ⅲ, followed by zones Ⅱ and I[28]. In our study, the ranking of ratios of eyes injured by explosives was zone Ⅲ > zone Ⅰ > zone Ⅱ among all cases of MEI; OGI zone Ⅰ and CGI zone Ⅲ were the most affected. Rupture and contusion were among the three most common types, and both generally involved zone Ⅲ. However, the location of general eye injuries or CGI from general causes has been most often reported to be in zone I[3,19] or Ⅱ[27,29], which is in contrast with our results. This comparison shows to some extent that explosion-related eye injuries predominantly involve the posterior segment (zone Ⅲ) of the eye and are therefore more serious than general eye injuries from general causes. Meanwhile, it also shows that ruptures and contusions, are mainly caused by level Ⅰ ocular explosion injury, and are the most important, non-negligible, and destructive types of blast trauma. A 3D fluid-structure interaction model of the human eye simulating the deformation of the eye components owing to primary blast injury showed that the amount of von Mises stresses (ability to resist deformation) on different ocular tissues was relevant to their mechanical and material properties. In terms of stress and strain, the eye components ranked from high to low were the sclera, ciliary body, cornea, lens, iris, retina, muscle, optic nerve, aqueous body, and vitreous body[30]. Although the optic nerve and retina (especially the macula) are located in zone Ⅲ (the rearward segment of the eyeball), they are most likely disrupted by even a small stress of strain leading to blindness. Conversely, the lowest stresses on the vitreous and aqueous humor, the root of the incompressible sphere, indirectly lead to global wall rupture, and the stress is the highest on the sclera and lowest on the extraocular rectus muscles. Another computational eye-specific model also indicated that the highest stresses and strains were near the rectus muscle insertions into the sclera[31]. This finding is likely the root cause for rectus insertions into the sclera, which is the most vulnerable site of globe rupture.

In our study, traumatic cataracts, VH, RD, and PVR all occurred more in OGI than in CGI, which indicates that, regardless of whether the sharp objects directly penetrate the eyeball or rupture it indirectly, the global structural integrity destructions indicate that the local or overall intraocular tissues undergo more severe, extensive, and diverse types of damage. No statistically significant difference was observed in the proportion of traumatic cataract in contusions and ruptures, which supports that zone Ⅱ (where the lens is located) is equally involved in both types of ocular trauma by blunt force. Tears of the eyeball wall along with the adjacent uveal vascular networks can inevitably lead to intraocular hemorrhage including VH which, then, was more common in rupture than contusion. Among contusion injuries, RD only developed when the blunt impact from the shockwave produced a strain in the retina high enough to exceed the specified threshold for RD[32]. Obviously, rupture can directly cause retinal/choroidal break, and the retina also can be pulled away by the lost or prolapsed vitreous body or vacuumed out because of the significant loss of intraocular contents and pressure. While no difference was observed in the proportion of PVR between contusions and ruptures, PVR was significantly more common in lacerations (especially perforating and IOFB injuries) than in ruptures. Therefore, we propose a bold hypothesis: the probability of PVR caused by blunt force is not necessarily related to whether the eyeball is broken or not, and the mechanism of the foreign object penetrating the eyeball significantly promotes the PVR occurrence and development.

At the final follow-up, 9.59% of eyes had been removed following explosive injuries in our study. In previous reports, this percentage was only 0.61%–10.00% for general or firework-related eye injuries[4-6,12,21,28,33-36], although it was as high as 11.90%–34.46% for injuries caused in war or by bomb or mine blasts[2,8,13,26,37,38]. The NLP in our study was 13.32% for the presenting VA and 14.56% for the final VA, which was significantly lower than the values provided in previous reports (initial NLP range 18.7%–53.1%)[3,4,7,8,13,28,37] or the final NLP in 14.04%–39.00%[3,4,8,12,28,36,37] of injuries related to firecrackers, bomb or mine explosions, war or weapon use, or OGI. This difference is probably related to the fact that our study included data for various causes of explosive injuries and types of eye injuries from mild to heavy, whereas Feng et al.[36] reported exclusively on eyes with severe explosive injuries that required vitrectomy. Additionally, the high proportion of low VA observed in our study [LP-4/200: 59.22% (initial); 31.26% (final)] as well as in previous reports [LP-hand moving (HM)/counting fingers (CF): 34.00%–56.47% (initial)[3,4,8,28,37]; 18.0%–55.6% (final)[3,4,8,13,28,37]] should not be ignored. The challenges of eye trauma treatment require increased clinical focus. In previous studies, after treatment, final VA level ≥20/40 was observed in 13.00%–42.11% of patients[3,4,8, 12] (24.71% in ours); the VA improved in 31.0%–67.3% of patients[3,5,6,26,37] (60.39% in ours). OGIs were often associated with worse visual prognosis in relation to highly lethal explosives, which is consistent with our comparison of presenting and final VA levels among injury types. And when compared with CGI, the proportion of patients with a final VA <0.1 was higher[3] and the mean VA was lower[13]. We further compared the classification of different ocular trauma mechanisms based on the proportions of patients with VA levels ≤4/200, VA reduction and enucleation rates, and inferred the following. (1) In OGI, the percentages of low VA levels and enucleation rates for ruptures caused by blunt force trauma are higher than those for lacerations caused by sharp objects, although the percentage of VA change or improvement after treatment is smaller among ruptures. Briefly, the harm caused by ruptures is greater than that caused by laceration and yet VA differentiation in some laceration cases is poor and not optimistic. (2) Among the three types of laceration by sharp objects, IOFB and perforating injuries can cause in a higher proportion of patients with low vision and a lower proportion of patients with VA improvement compared with penetrating injuries; however, the three types do not differ in final VA reduction and enucleation rates. (3) Rupture as a type of OGI and contusion as a type of CGI are both caused by blunt force. The proportion of patients with low vision and the enucleation rate after a rupture injury is higher, and the final VA changes (whether improvement or reduction) is lower. All these prove once again that a contusion occurs before the eyeball is subjected to instantaneous shock and compression of the blunt force beyond the threshold that the eyeball wall can withstand, and once this threshold is exceeded, a rupture injury occurs. Based on the results of this study, recovery from a high-severity of globe rupture with extensive ocular tissue damage is difficult. Few studies have compared visual function and prognosis of ocular trauma according to the injury mechanism. However, based on our analysis, determining the mechanism of injury and accurately classifying the ocular trauma contributes to the prioritization of ocular trauma treatment, injury evaluation, and outcome prediction.

As far as predictors of poor final VA, initial VA is generally acknowledged as a key factor that affects final vision status[3,14,28,39-42], which is consistent with our results. To some extent, initial VA after ocular trauma reflects the severity of intraocular tissue damage in most eye injuries. In terms of eye injury type, IOFB (mainly associated to level Ⅱ explosion injury), rupture, contusion (resulting from level Ⅰ injuries) and NMEI (level Ⅳ injuries) were independent risk factors for poor vision. Indeed, these findings make it evident that explosion injuries cause great harm to eyeballs. The rupture severity discussed above is an important risk factor of poor vision, which is supported by the literature[5,26,38,41]. Contusion, although not a scoring factor for low vision in OTS (a simplified predictive tool for ocular trauma)[42], showed significance in this study probably owing to the inclusion of explosion injuries, and zone Ⅲ (posterior segment) accounting for the vast majority of injuries (75.32%). Indeed, injuries located in zone Ⅲ were recognized as an independent risk factor for low vision in this study and many others[4,38,41]. Consistent with our results, full-thickness laceration of the eyeball ≥5 mm (or larger wounds)[26,40,42], VH[19,40,41], RD[3,5,40,41], TON[5], endophthalmitis[39,42] and PVR[36] were identified as risks factors for poor VA in some previous analyses. Regarding eye structure and function, damage that involves the posterior segment is not easily repaired. Involvement of the retina or choroid was revealed as predictive factor of poor VA in previous reports[5,18,27,36]. Few studies have been conducted on ciliary body detachment (CBD), as most case analyses do not include this observation. In addition to initial VA, rupture, endophthalmitis and RD, the OTS includes perforation and relative afferent pupillary defect (RAPD)[11]. Because of the small number of perforating cases, few effective statistical analyses can be conducted, which may be the reason it has not been identified as an independent risk factor in many studies, including ours. In certain instances, especially of OGI, RAPD cannot be examined, evaluated and recorded after injury, and therefore was not included in this study. The OTS has general significance as a reference for the evaluation of visual function after ocular trauma, although the particularities of explosive ocular injury should be considered due to its special characteristics compared with conventional eye injuries. The predictive value of these risk factors is useful for patients counseling as well as managing expectations and guiding clinical decisions.

This multi-center review revealed that explosions cause high rates of bilateral ocular trauma, complex injury conditions and are associated with increased incidence of zone Ⅲ (posterior segment) involvement and poor prognosis. Eyeball contusions and ruptures associated with primary (level I) explosion injuries, lacerations (especially those caused by IOFB) secondary (level Ⅱ) to explosion-related injuries, and NMEI related to grade Ⅳ explosive injury (burns) were the main mechanisms identified in our study. Rupture injuries tend to predict a worse prognosis and lower probability of recovery. Compared with pure penetrating injuries, the harm caused by IOFB and perforation is more severe. Moreover, we suggest that PVR caused by blunt force is not necessarily related to eyeball rupture, and the mechanism of penetration by a foreign object into the eyeball, especially by IOFBs or direct penetrating of the retinal/choroid, promotes the PVR occurrence and development. Worse presenting VA, larger-sized eyeball wall wound, clinical signs (VH, TON, RD, CBD, endophthalmitis and zone Ⅲ OGI), injury type (IOFB, rupture and contusion) and PVR development were associated with worse visual outcomes among injured eyes following an explosion. Future studies that include large and homogeneous case and control populations are required to evaluate the prognostic scoring system of eye blast injuries among different types of eye injuries caused by general causes.

3D: Three-dimensional; BETT: Birmingham eye trauma terminology; CBD: Ciliary body detachment; CF: Counting fingers; CGI: Close globe injury; HM: Hand moving; IOFB: Intraocular foreign body; IOP: Intraocular pressure; LP: Light perception; MEI: Mechanical eye injury; NLP: No light perception; NMEI: Non-mechanical eye injury; OGI: Open globe injuries; OR: Odds ratio; OTCS: Ocular trauma classification system; OTS: Ocular trauma score; PVR: Proliferative vitreous retinopathy; RAPD: Relative afferent pupillary defect; RD: Retinal detachment; TON: Traumatic optic neuropathy; VA: Visual acuity; VH: Vitreous hemorrhage.

The online version contains supplementary material available at https://doi.org/10.1186/s40779-022-00438-4.

Additional file 1: Table S1 General features of the explosive eye injuries (total of 1115 patients). Fig. S1 Distribution of the number of patients and the proportions of male patients and bilateral eye injuries in different age groups.

Thank all participants in this study.

YZ and MNZ conceived and designed this study. YZ, XK and QW undertook the data analysis, results interpretation and manuscript preparation. YZ and MNZ were responsible for results interpretation and revision of the manuscript. QW, ZZ and JY performed the data collection and management. All authors read and approved the final manuscript.

Not applicable.

Not applicable.

The authors declare that they have no competing interests.

¹Senior Department of Ophthalmology, the Third Medical Center of Chinese PLA General Hospital, Beijing 100039, China. ²Department of Ophthalmology, the First Medical Center of Chinese PLA General Hospital, Beijing 100853, China. ³Medical Security Center, Chinese PLA General Hospital, Beijing 100853, China. ⁴Information Center, Logistics Support Department, Central Military Commission, Beijing 100842, China. ⁵Information Management Department, Chinese PLA General Hospital, Beijing 100853, China.