High-Rate Internal Pressurization of Human Eyes to Predict Globe Rupture

LABORATORY SCIENCES High-Rate Internal Pressurization of Human Eyes to Predict Globe Rupture

Jill A. Bisplinghoff, BS; Craig McNally, BS; Stefan M. Duma, PhD

Objective: To determine the dynamic rupture pressure onal direction. There was no significant difference in the of the human eye by using an in vitro high-rate pressur- rupture pressure between the equatorial and meridional ization system to investigate blunt-impact eye injuries. directions (P=.16).

Methods: Internal pressure was dynamically induced in Conclusion: As the loading rate increases, the rupture the eye by means of a drop-tower pressurization system. pressure of the human eye increases. The internal eye pressure was measured with a small pressure sensor inserted into the eye through the optic nerve. Clinical Relevance: Eye injuries are expensive to treat, A total of 20 human eye tests were performed to deter- given that the estimated annual cost associated with adult mine rupture pressure and characterize rupture patterns. vision problems in the United States is $51.4 billion. De- termining globe rupture properties will establish injury Results: The high-rate pressurization resulted in a mean criteria for the human eye to prevent these common yet (SD) rupture pressure of 0.97 (0.29) MPa (7275.60 devastating injuries. [2175.18] mm Hg). A total of 16 eyes ruptured in the equatorial direction, whereas 4 ruptured in the meridi- Arch Ophthalmol. 2009;127(4):520-523

ORE THAN 1.9 MILLION thalmologist.16 In published research, the people experience eye rupture pressure of eyes has been exam- injuries in the United ined to determine the effect of ocular sur- States each year.1 Au- gery,5,16-20 and, to our knowledge, only 1 tomobile accidents,2-7 study has been published on the rupture sports-related impacts,8,9 consumer prod- pressure of healthy human eyes.21 Burn- M 10 17 ucts, and military combat are some of the stein et al studied the strength of post- causes of the severe eye injuries that oc- mortem human and porcine eyes after pho- cur.11 These injuries are expensive to treat, torefractive keratectomy by increasing given that the estimated annual cost as- intraocular pressure gradually with nitro- sociated with adult vision problems in the gen gas until globe rupture occurred. The United States is $51.4 billion.12,13 Of the eye was attached to the nitrogen gas sys- total number of eye injuries in the coun- tem by a 25-gauge butterfly needle that was try, more than 600 000 sports injuries oc- inserted into the anterior chamber at the cur each year and 40 000 of them require limbus. Pressure was increased by 0.03 MPa emergency care.14 A 2002 study found that (225.02 mm Hg) at 5-second intervals. more than 9000 globe ruptures occur in Burnstein et al found the mean (SD) rup- the United States each year.15 A blunt im- ture pressure of human eyes subjected to pact, like that of a baseball, will cause the photorefractive keratectomy to be 0.46 eye to compress, with an increase in in- (0.12) MPa (3450.28 [900.07] mm Hg), ternal pressure that can result in globe rup- whereas porcine eyes subjected to photore- ture (Figure 1). The rupture starts away fractive keratectomy were observed to rup- from the impact location, demonstrating ture at 0.53 (0.10) MPa (3975.33 [750.06] that the cause is the increase in internal mm Hg). In 2 eyes undergoing photothera- pressure. Determining globe rupture prop- peutic keratectomy, Burnstein et al found erties will provide the needed informa- rupture pressures of 0.55 and 0.68 MPa tion for establishing injury criteria for the (4125.34 and 5100.42 mm Hg, respec- Author Affiliations: 21 Department of Mechanical human eye to help prevent these com- tively). Voorhies performed both static Engineering, Virginia mon yet devastating injuries. and dynamic tests on healthy postmortem Polytechnic Institute and State Information about safe intraocular pres- human and porcine eyes by similar meth- University, Blacksburg. sures is of particular use to the clinical oph- ods. The static rupture pressure for hu-

(REPRINTED) ARCH OPHTHALMOL / VOL 127 (NO. 4), APR 2009 WWW.ARCHOPHTHALMOL.COM 520

Figure 1. Baseball impact with a human eye within a gelatin orbit resulting in globe rupture. A, Before impact. B, Rupture occurring on impact (arrow). C, Ruptured globe.

man eyes was found to be 0.36 (0.20) MPa (2700.22 [1500.12] mm Hg), whereas the dynamic rupture pres- Potentiometer sure was 0.91 (0.29) MPa (6825.56 [2175.18] mm Hg). Weight The previous studies tested rupture pressure in human and porcine eyes; however, these studies were lim- Drop-tower ited in that no high-rate dynamic tests were performed that track

would realistically simulate a dynamic ocular injury, such Guide Upstream as sports-related injuries, automobile injuries, military com- pressure Internal eye bat, or injuries caused by consumer products.17,21,22 More- pressure over, the majority of these studies were focused on test- Relief Hydraulic ing specimens that had been subjected to corrective eye valve cylinder DAS surgery before testing and had segmented eye tissue. There- module fore, the purpose of this study was to analyze the high- rate rupture pressure of healthy human eyes by means of an internal pressurization system. Figure 2. Schematic diagram of the pressure system used to examine high-rate rupture pressures of human eyes. DAS indicates data acquisition system. METHODS

High-rate pressurization was accomplished with a custom pres- Precision Measurement Co, Ann Arbor, Michigan) was in- sure system that was built to examine rupture properties of the serted into the eye through the optic nerve. The pressure trans- human eye (Figure 2). The test setup consisted of a drop tower ducer was rated for a range of 0 to 3.45 MPa (0-25877.13 that was used to create a hydraulic system to pressurize the hu- mm Hg) and a frequency response of 10 000 Hz, which are ideal man eye in a dynamic event. To initiate the event, a weight was for this application. dropped onto a piston that was inserted into the hydraulic cyl- Twenty human eyes were procured from the Roanoke Eye inder. Preparation of the system included adding water through Bank (Roanoke, Virginia) and the North Carolina Eye Bank the cylinder to act as the medium for pressurization and to pro- (Winston-Salem). The eyes had not undergone corneal trans- duce an approximate initial intraocular pressure of 0.002 MPa plantation or other operations before testing. The eyes were never (15.00 mm Hg) before rupture. Connecting the eye to the sys- frozen and were kept in isotonic sodium chloride solution in tem was a 16-gauge intravenous needle inserted into the optic glass jars and refrigerated until testing. A previous study showed nerve (Figure 2). To secure the optic nerve to the needle, a medi- a lack of correlation between age and rupture pressure, as well cal suture was used while a cylindrical placement guide held as time from harvest to test and rupture pressure (R2=0.06),4 the eye in place below the needle. To ensure that the optic nerve so these factors were not taken into consideration. Statistical was sealed, it was covered with a flexible coupling and then analyses were performed by t test, with ␣=.05 used to deter- secured with a plastic fastener. mine statistical significance. All test procedures were re- The high-rate dynamic event occurred when the weight was viewed and approved by the Virginia Tech Institutional Re- suspended above the piston at a height of 17.78 cm and then view Board. released. The impact of the weight onto the piston caused the water to be displaced throughout the system, which created a high-rate increase in intraocular pressure resulting in the rup- RESULTS ture of the eye. To capture this event, high-speed video and data acquisition were used. A video camera (Ultima APX-RS; Pho- The high-rate pressurization of 20 human eyes resulted in tron Inc, San Diego, California) captured video at 10 000 frames a mean (SD) rupture pressure of 0.97 (0.29) MPa per second with a resolution of 512 ϫ512 pixels, while the data acquisition system collected data at 30 000 Hz. The pressure (7275.60[2175.18] mm Hg) (Table). The failure pres- transducer data and the high-speed video were correlated to sure ranged from 0.57 to 1.59 MPa (4275.35-11 925.98 determine the pressure at the time of rupture. mm Hg) (Figure 3). The dynamic loading rate was 36.50 To acquire the internal pressure of the human eye, an in situ (15.35) MPa/s (273 772.5 [115 134.5] mm Hg/s) for the pressure sensor was used. A small pressure sensor (model 060; 20 human eye tests. High-speed video for each test showed

(REPRINTED) ARCH OPHTHALMOL / VOL 127 (NO. 4), APR 2009 WWW.ARCHOPHTHALMOL.COM 521

Test No. Rupture Pressure, MPa Time to Rupture, ms Loading Rate, MPa/s Rupture Direction Rupture Location 1 1.59 20.6 77.4 Equatorial Equator 2 0.70 21.2 32.8 Equatorial Equator 3 0.82 21.5 38.3 Equatorial Equator 4 0.80 23.8 33.7 Equatorial Equator 5 1.34 24.5 54.7 Equatorial Equator 6 1.30 25.2 51.6 Meridional Equator 7 1.32 25.2 52.4 Meridional Equator 8 1.50 25.6 58.6 Equatorial Equator 9 0.71 26.1 27.1 Equatorial Equator 10 0.57 26.5 21.6 Equatorial Cornea 11 0.95 26.8 35.3 Meridional Cornea 12 0.78 28.1 27.7 Equatorial Equator 13 0.91 28.7 31.7 Equatorial Equator 14 0.96 28.8 33.5 Meridional Equator 15 1.01 29.3 34.4 Equatorial Equator 16 0.98 29.8 32.9 Equatorial Equator 17 0.94 31.5 30.0 Equatorial Equator 18 0.88 37.3 23.7 Equatorial Equator 19 0.76 39.0 19.5 Equatorial Equator 20 0.60 45.9 13.2 Equatorial Equator Mean (SD) 0.97 (0.29) 28.3 (6.2) 36.5 (15.3)

Conversion factor: To convert the rupture pressure from megapascals to millimeters of mercury, multiply by 7500.617.

sults because of the capability of the manometer used. 1.6 The test series completed by Pinheiro et al22 used an ar- 1.4 tificial anterior chamber to investigate the integrity of the

1.2 cornea, thereby determining rupture pressure for the cor-

1.0 nea rather than the sclera. Although these studies are use- ful for their specific purposes, the current study is in- 0.8 tended as a more general approach that can be used for 0.6

Pressure, MPa scleral injuries that occur from increasing internal pres- 0.4 sure, such as a blunt impact. 0.2 The human eye is both anisotropic and viscoelas- tic.17,21,23 Previous dynamic tests performed by Voor- 0 5 10 15 20 25 30 35 40 hies21 were executed by means of a loading rate of ap- Time, ms proximately 2.77 MPa/s (20 776.71 mm Hg/s)21; however, for the current test series, a higher loading rate was de- Figure 3. Rupture pressure with respect to time for 20 human eyes. To convert the rupture pressure from megapascals to millimeters of mercury, sired to more accurately predict a dynamic ocular in- multiply by 7500.617. jury observed from impact with blunt objects. The high- rate rupture pressure device achieved an average loading rate of 36.5 MPa/s (273 772.5 mm Hg/s), which is an or- the location of each rupture. Of the 20 eyes tested, 18 rup- der of magnitude greater than those in the previous dy- tured at the equator, while the remaining 2 ruptured through namic tests. In this study, both the loading rate and the the cornea. The eyes were then examined to determine the duration of the event were changed by approximately an direction of the rupture. Four ruptured in the meridional order of magnitude from previous dynamic tests.21 The direction while 16 ruptured in the equatorial direction. A time to rupture was also measured and resulted in a mean 2-tailed unpaired, unequal variance t test showed that the of 28.27 (6.26) milliseconds, compared with 0.46 (0.23) difference in the rupture pressure between the equatorial seconds in previous dynamic tests performed by Voorhies. mean of 0.93 (0.30) MPa (6975.57 [2250.19] mm Hg) and In the study by Voorhies, a pressure transducer was the meridional mean of 1.13 (0.21) MPa (8475.70 [1575.13] placed at a location before the needle interface between mm Hg) was not significant (P=.16). the eye and the system and was assumed to reflect the internal pressure of the human eye.21 In comparing the COMMENT current results with the previous rupture pressure data, it is important that, because of pressure loss across the Previous studies have performed quasi-static and dy- needle during the event, the upstream pressure used for namic tests to determine the rupture pressure of human the previous testing is higher than that actually experi- eyes.17,18,21-23 Both Campos et al18 and Pinheiro et al22 tested enced by the eye (Figure 4). A measurement error is the effects of refractive surgery. Campos et al18 used por- introduced when the upstream pressure is used instead cine eyes and were unable to report rupture pressure re- of the direct internal pressure.

(REPRINTED) ARCH OPHTHALMOL / VOL 127 (NO. 4), APR 2009 WWW.ARCHOPHTHALMOL.COM 522

©2009 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Financial Disclosure: None reported. 7 Funding/Support: The Virginia Tech–Wake Forest Cen- Upstream pressure 6 ter for Injury Biomechanics funded the tissue testing. Additional Contributions: The US Army Aeromedical Re- 5 search Laboratory supported this program. 4 Upstream Internal eye pressure pressure 3 REFERENCES Pressure, MPa 2 Internal eye pressure 1. McGwin G Jr, Xie A, Owsley C. Rate of eye injury in the United States [published 1 correction appears in Arch Ophthalmol. 2005;123(9):1285]. Arch Ophthalmol. 2005;123(7):970-976. 0 10 20 30 40 50 2. Duma SM, Jernigan MV, Stitzel JD, et al. The effect of frontal air bags on eye injury Time, ms patterns in automobile crashes. Arch Ophthalmol. 2002;120(11):1517-1522. 3. Fukagawa K, Tsubota K, Kimura C, et al. Corneal endothelial cell loss induced by Figure 4. Pressure difference between upstream pressure and direct internal air bags. Ophthalmology. 1993;100(12):1819-1823. eye pressure. To convert the rupture pressure from megapascals to 4. Kennedy EA, Voorhies KD, Herring IP, Rath AL, Duma SM. Prediction of severe millimeters of mercury, multiply by 7500.617. eye injuries in automobile accidents: static and dynamic rupture pressure of the eye. Annu Proc Assoc Adv Automot Med. 2004;48:165-179. 5. Kisielewicz LT, Kodama N, Ohno S, Uchio E. Numerical prediction of airbag- The location of the pressure sensor inside the eye was caused injuries on eyeballs after radial keratotomy. Paper presented at: Society also a concern because of the change in pressure gradi- for Automotive Engineers International Congress and Exposition; February 24, ent as the water was injected into the eye. A test was per- 1998; Detroit, Michigan. formed with 2 sensors within 1 eye to compare the re- 6. Duma SM, Crandall JR. Eye injuries from airbags with seamless module covers. J Trauma. 2000;48(4):786-789. sults of 2 identical pressure sensors in different locations 7. Duma SM, Kress TA, Porta DJ, et al. Airbag-induced eye injuries: a report of 25 during the event. Regardless of the location of the pres- cases. J Trauma. 1996;41(1):114-119. sure sensors, the pressure output was the same. 8. Kennedy EA, Ng TP, McNally C, Stitzel JD, Duma SM. Risk functions for human All eyes were stored in isotonic sodium chloride so- and porcine eye rupture based on projectile characteristics of blunt objects. Stapp lution and were refrigerated before testing. Sclera is largely Car Crash J. 2006;50:651-671. 9. Vinger PF, Duma SM, Crandall J. Baseball hardness as a risk factor for eye injuries. acellular, and its strength comes from collagen fibers simi- Arch Ophthalmol. 1999;117(3):354-358. 4 lar to those found in ligaments. Because the tissue was 10. Heier JS, Enzenauer RW, Wintermeyer SF, Delaney M, LaPiana FP. Ocular inju- never exposed to a freeze-thaw cycle and was stored in ries and diseases at a combat support hospital in support of Operations Desert refrigerated isotonic sodium chloride solution, the globe’s Shield and Desert Storm. Arch Ophthalmol. 1993;111(6):795-798. 11. Kennedy EA, Inzana JA, McNally C, et al. Development and validation of a syn- integrity was preserved before the testing. thetic eye and orbit for estimating the potential for globe rupture due to specific The principal rupture location for the 20 tests was the impact conditions. Stapp Car Crash J. 2007;51:381-400. equator. Other research has also shown the equator as 12. Frick KD, Gower EW, Kempen JH, Wolff JL. Economic impact of visual impair- the prime location for a rupture occurring from a blunt ment and blindness in the United States. Arch Ophthalmol. 2007;125(4):544- impact, such as a baseball.21,24 Because failures occur pri- 550. 13. Rein DB, Zhang P, Wirth KE, et al. The economic burden of major adult visual marily at the equator of the eye, this study, along with disorders in the United States [published correction appears in Arch Ophthal- 4 previous studies, suggests that the equator is the weak- mol. 2007;125(9):1304]. Arch Ophthalmol. 2006;124(12):1754-1760. est portion of the eye. 14. Hecker S. More than half a million Americans suffer eye injuries from sports- Accurate response characteristics of the eye must be related accidents: lack of proper eye protection can lead to painful injuries, vision loss and even blindness. Vision News. August 2007:1-2. known for a model to behave realistically and to predict 15. Smith D, Wrenn K, Stack LB. The epidemiology and diagnosis of penetrating eye when failure of local tissues will occur. Although the rup- injuries. Acad Emerg Med. 2002;9(3):209-213. ture location was primarily the equator, the ruptures oc- 16. Bryant MR, McDonnell PJ. Constitutive laws for biomechanical modeling of re- curred in 2 different directions: along the equator and fractive surgery. J Biomech Eng. 1996;118(4):473-481. along the meridian. However, the difference in the rup- 17. Burnstein Y, Klapper D, Hersh PS. Experimental globe rupture after excimer la- ser photorefractive keratectomy. Arch Ophthalmol. 1995;113(8):1056-1059. ture pressure between the equatorial and meridional di- 18. Campos M, Lee M, McDonnell PJ. Ocular integrity after refractive surgery: ef- rections was not significant. fects of photorefractive keratectomy, phototherapeutic keratectomy, and radial These data are essential for the continuing improve- keratotomy. Ophthalmic Surg. 1992;23(9):598-602. ment in prediction of eye injury. This research will pro- 19. Hanna KD, Jouve FE, Waring GO III, Ciarlet PG. Computer simulation of arcuate and radial incisions involving the corneoscleral limbus. Eye. 1989;3(pt 2):227-239. vide the necessary information to advance current math- 20. Wray WO, Best ED, Cheng LY. A mechanical model for radial keratotomy: to- ematical and computer models of eye injury risk ward a predictive capability. J Biomech Eng. 1994;116(1):56-61. assessment and potentially reduce the incidence of eye 21. Voorhies KD. Static and Dynamic Stress/Strain Properties for Human and Por- injuries from blunt impacts. cine Eyes [thesis]. Blacksburg: Virginia Polytechnic Institute and State Univer- sity; 2003. 22. Pinheiro MN Jr, Bryant MR, Tayyanipour R, Nassaralla BA, Wee WR, McDonnell Submitted for Publication: June 26, 2008; final revi- PJ. Corneal integrity after refractive surgery: effects of radial keratotomy and mini- sion received August 15, 2008; accepted August 18, 2008. radial keratotomy. Ophthalmology. 1995;102(2):297-301. Correspondence: Jill A. Bisplinghoff, BS, School of Bio- 23. Uchio E, Ohno S, Kudoh J, Aoki K, Kisielewicz LT. Simulation model of an eye- medical Engineering, Virginia Tech–Wake Forest Uni- ball based on finite element analysis on a supercomputer. Br J Ophthalmol. 1999; 83(10):1106-1111. versity, Center for Injury Biomechanics, 100F Ran- 24. Stitzel JD, Duma SM, Cormier JM, Herring IP. A nonlinear finite element model dolph Hall, Mail Code 0238, Blacksburg, VA 24061 of the eye with experimental validation for the prediction of globe rupture. Stapp ([email protected]). Car Crash J. 2002;46:81-102.

(REPRINTED) ARCH OPHTHALMOL / VOL 127 (NO. 4), APR 2009 WWW.ARCHOPHTHALMOL.COM 523