With the inception of excimer laser use in corneal refractive and therapeutic surgery, several issues concerning the safety of ultraviolet radiation have been addressed. Among these, photochemical injury (photokeratitis) is questioned as a phenomenon of UV radiation not previously considered in the far UV. Previous work by Pitts1"4 working with UV in the 210 to 320 nm range demonstrates an increase in threshold energy density for photokeratitis as wavelength is decreased. He also noted that with shorter wavelength exposures, a more superficial epithelial photokeratitis is observed. This would suggest that at 193 nm an even greater energy exposure might be required for photokeratitis, and that this keratitis might be of a more superficial nature. Since the threshold energy density for excimer laser corneal ablation is 50 mJ/cmp 2/p,5 and perhaps as low as 25 mJ/cmp 2/p,6 photokeratitis might be seen only following repeated subablative energy exposures. Hence, the potential safety hazard of photokeratitis from scattered 193-nanometer excimer laser light might be negligible.
MATERIALS AND METHODS
Three adult Dutch Belted rabbits were used one or more times in five series of experiments to determine the threshold of photokeratitis at 193 nm. In the first series, all three rabbits were used, but in subsequent series, only one or two of these rabbits were reused after total resolution of the keratitis. Each animal was inspected biomicroscopically before each series of exposures and subsequently anesthetized with an intramuscular injection of ketamine hydrochloride (25 mg/Kg) and xylazine hydrochloride (1 mg/Kg). Each eye was stabilized by proptosis or lid speculum and irradiated four times, once in each of the four corneal quadrants. Pulse energy densities of less than 15 mJ/cm2 were use at pulse frequencies of less than or equal to 5 Hz over a wide range of total energy exposures. Each series was separated by an interval of at least 15 days before a given rabbit's cornea was reused, and only after careful inspection. No additive effects were seen from repeated exposures separated by this span of time.
FIGURE I: Optical configuration for excimer laser energy measurement and projection to the cornea. Diverging rays allow low pulse energy density projection within the range of a stable laser energy output.
A Lambda Physik 102E excimer laser was used with an optical configuration shown in Figure 1. The emergent laser beam was focused with a 15-centimeter focal length fused silica quartz lens and then allowed to diverge through a 2.5-mIUimeter aperture to obtain a reduced beam of good quality. A second 3-millimeter aperture was placed directly in front of the animal eye to assure correct placement of each exposure. Energy measurements were made before and after each series of exposures and were averaged. Energy measurements were made between the first and second aperture and calculated with the coincident area of exposure to determine pulse energy density. Energy exposures were allowed in each quadrant of the cornea to increase the number of data points.
During the first two series of exposures, balanced salt solution (BSS) was used every 5 seconds in the left eye to assess if a tear film altered the threshold for photokeratitis. The right eye received tears only between exposures. Subsequent series used BSS every 2 minutes to best simulate a human experience.
The animals were observed grossly immediately following exposure and via slit-lamp biomicroscope at various intervals beginning V2 to 1 hour and ending 26 hours postexposure. Corneas were carefully examined for signs of epithelial haze and epithelial stippling, and rose bengal and fluorescein stain were used to help delineate early signs of photokeratitis. Fluorescein was used in all series, while rose bengal was used in only the latter two series. All signs of epitheliopathy as evidenced by rose bengal and fluorescein staining were recorded only when seen within the circular pattern of irradiation projected in each of the four corneal quadrants. This was done to differentiate clinical photokeratitis from staining noted because of rabbit desquamation or surface drying of cells.
First Series (Three Rabbits, Six Eyes)
Similar exposures were performed in each eye at between 2.9 to 3.4 mJ/cmp 2/p for total energy exposures ranging from 0.023 to 2.3 J/cmp 2. BSS was used Q5 min in the left eye and between exposures in the right to prevent drying. A slight epithelial haze and irregular appearance was noted by slit-lamp exam within the first 1 V2 hours at 1.2 and 2.3 J/cmp 2, in both the left and right eyes. The haze in the left eye was slightly less dense than that in the right eye due to frequent instillation of drops.
Second Series (One Rabbit, Two Eyes)
Once again with pulse energy densities ranging from 2.3 to 3.4 mJ/cmp 2/p, a slight haze was noted immediately following an energy density exposure of 1.5 J/cmp 2. This effect was noted as a well defined very superficial epithelial surface disruption on slitlamp exam 2 hours later, but there was no evidence of an effect at 9 and 19 hours postexposure. Once again, the effect was seen in both eyes with slightly less intensity in the eye receiving more frequent instillation of balanced-salt-solution drops.
FIGURE 2: The photographic appearance of phofokeraffffs as seen by retroillumination at 21.9 J/cmp 2 and 14 hours after exposure. This lesion stained positively with fluorescein and corresponds to a time delayed keratitis caused by longer wavelength fluorescence of subablative 193-nanometer excimer laser light.
Third Series (Two Rabbits, Four Eyes)
In the third series, slightly higher pulse energy densities were used to achieve the latter energy density exposures recorded. Initially, no perceptible effect was seen with pulse energy densities from 2.5 to 3.6 mJ/cmp 2/p for total exposures up to 4.0 J/cmp 2. With a higher pulse energy density of 10.4 mJ/cmp 2/p, a notable haze was seen at 2 hours for exposures 6.2, 11.4, and 19.5 J/cmp 2, which at 14 hours postexposure was noted to stain with fluorescein at the 11.4 and 19.5 exposure level. At 19 hours, exposures in this range showed no perceptible effect.
Fourth Series (One Rabbit, TWo Eyes)
Pulse energy densities in this series ranged from 12.7 to 14.3 mJ/cmp 2/p with total energy exposures ranging from 1.4 to 71.0 J/cmp 2. Rose bengal stain was used in these latter series in addition to fluorescein stain. The observed effects are as follows. At 7.3 J/cmp 2, an immediate circle of haze was grossly noted which appeared within a large unbordered zone of stippling with rose bengal at 6 1M hours post exposure. At 21.9 J/cmp 2, haze is seen immediately with distinct bordered stippled staining of rose bengal at 3 ½ hours, becoming more dense at 6 ½ hours. Faint rose bengal with dense flourescein staining is seen at 14 hours (Fig 2) with complete dissipation between 19 to 26 hours. This same pattern is seen with even higher energy exposures of 71 J/cmp 2, suggesting a patterned deepening of effect with time for energies greater than 21.9 J/cmp 2.
Fifth Series (One Rabbit, Two Eyes)
Lower energy exposures at lower pulse energy densities were sought in this last series to better define the threshold for the deep time-delayed effect noted above. Pulse energy densities from 5.5 to 8.0 mJ/cmp 2/p were used to create exposures of 1.4 to 11.0 J/cmp 2. At 3.6 J/cmp 2, rose bengal staining was seen at 1 hour, but had disappeared by 2 Vi hours. This early disappearance did not occur for 5.1 J/cmp 2, but became deeper with time, eventually disappearing at 13 hours postexposure. Energy exposures of 6.6, 9.8, and 11.0 J/cmp 2 showed the same pattern of rose bengal staining, but also showed fluorescein staining at 13 hours for the 11.0 J/cmp 2 exposure and at 15 hours for both the 9.8 and 11.0 J/cmp 2 exposures. Fluorescein staining was still seen at 19 hours, but was gone at 26 hours post exposure for the 11.0 J/cmp 2 exposure.
The UV action spectrum of the cornea was described by Pitts et al in the late 1960s and early 70s.1"4 During these studies, they described the threshold for photokeratitis over the spectrum of UV wavelengths from 210 to 370 nm. During their experiments, initially in rabbits1 and later in primates and humans,2"4 they noted threshold photokeratitis as the presence of epithelial haze, epithelial debris, or epithelial granules in the superficial or deep layers of the corneal epithelium. Epithelial haze was seen as an irregular crackled appearance, epithelial debris as small glistening bodies on the surface, and epithelial granules as small white round spots located in the deeper epithelial layers. Clinically, they also noted tearing, stippling, hyperemia, haze, discharge, photophobia, pain, and blepharospasm as additional signs of photokeratitis.
For most of the UV wavelengths, the onset and termination of photokeratitis in humans followed a patterned temporal course.2"4 In general, dead cells could be seen biomicroscopically on the surface initially, followed by corneal clarity within the 1st hour. By 4 to 6 hours post exposure, mild epithelial haze and granules were seen which increased slightly with time and demonstrated a superficial collection of fluorescein stain. More pronounced fluorescein staining was not seen until symptoms first appeared at 11 to 12 hours postexposure, and these abated usually within 24 hours. In rabbits and primates, careful temporal analysis was not done, but observations were made at 12 to 18 hours when threshold signs were evident.1,2
FIGURE 3: The ultraviolet action spectrum for threshold photokeratitis in the rabbit.1,2 At 193 nm, the two points represent the threshold levels for photokeratitis due to direct excimer laser light (lower) and indirect fluorescence from the excimer laser (upper). The solid line represents the American Conference of Governmental Industrial Hygienists-Threshold Limit Values (ACGIH TLV) for safe energy exposure in both eye and skin."
With wavelengths shorter than 250 nm, finer more superficial granules and increased surface debris could be seen. These, along with haze, tended to occur much sooner, 1 to 2 hours, after exposure with clearing by 4 to 6 hours.2,4 Symptoms were also experienced earlier, 2 to 3 hours, and these tended to disappear by 6 hours. The shorter wavelights were felt to have absorbed more superficially with a more rapid response, while longer wavelengths produced deeper more serious epithelial changes in a delayed time course.
The action spectrum for threshold photokeratitis is demonstrated in Figure 3. Included in this graft are the two threshold levels for photokeratitis at the 193-nanometer excimer laser exposure. As can be seen from the graph, wavelengths longer than 320 nm (UVA) tended to require a great deal of total energy (greater than 10.0 J/cmp 2) to induce photokeratitis in rabbits. Below 310 nm, thresholds were more on the order of 0.01 to 0.10 J/cmp 2 with a maximum sensitivity at 270 nm (5 m J/cmp 2). The threshold once again sharply rose with wavelengths shorter than 210 nm (0.70 J/cmp 2). This trend suggests that the threshold for direct 193-nanometer light would be even greater than that of 210 nm with consequently an even more superficial absorption length.
As stated above, our results demonstrate two threshold levels for photokeratitis at 193 nm. The first represents a very superficial epithelial cell surface devitalization, while the second a deeper more delayed photokeratitis change. At 1.00 to 1.50 J/cmp 2, a very superficial epithelial haze could be seen immediately post exposure which remained within the 1st hour biomicroscopically. This disappeared within a matter of hours without subsequent return. At higher doses, on the order of 10 J/cmp 2, a delayed photokeratitis with deep rose bengal and even fluorescein staining was seen. The deeper rose bengal staining was seen after 5 to 7 hours and lessened during the onset of fluorescein staining at 13 to 15 hours postexposure. These stains would remain for several more hours and disappear between 19 to 26 hours postexposure.
Our first response is consistent with superficial absorption of 193-nanometer light, creating a subablative epithelial cell devitalization. This effect probably represents a photochemical process as thermal and mechanical effects are less likely at these low energy and power densities. This type of response was previously noted in early investigations with excimer laser when epithelial cells would show an epithelial whitening or grayish granular opacity at subablative energy densities.5·7 The observed threshold for epithelial whitening, 20 to 25 mJ/cmp 2,5 and grayish granular opacity, 5 to 10 m J/cmp 2,7 is within the realm of several of our subablative energy densities during this study. Our results, however, demonstrate that although this epithelial devitalization was present at pulse energy densities as high as 15 mJ/cmp 2/p, it was also present at densities as low as 2.7 mJ/cmp 2 and, therefore, suggest the absence of thermal and mechanical mechanisms.
Our second response is consistent with photokeratitis of longer wavelength and represents an indirect response due to fluorescent emission excited by the 193-nanometer radiation. This response due to fluorescence demonstrates a delayed and deeper biomicroseopic and staining pattern consistent with wavelengths greater than 250 nm.2'4 The presence of deep rose bengal staining and especially fluorescein staining indicate the greater depth of photokeratitis and are similar to the longer wavelength findings shown by Pitts et al.2"4
FIGURE 4: Comparative energy magnitude diagram for 193-nanometer excimer laser induced photokeratitis. Multiple subablative pulses have an additive effect creating photokeratitis at energy exposures which exceed the threshold for single pulse photoablation and optical plasma formation.
The temporal course of onset of epithelial changes, rose bengal staining, and fluorescein staining also confirm the above conclusions. In our results, superficial epithelial haze and stippling could be seen immediately postexposure for most energy densities greater than 1.0 to 1.5 J/cmp 2. This was confirmed biomicroscopically within the 1st hour, and was no longer noticeable several hours later. This temporal response would appear to mimic the tendency for more superficial and sooner photokeratitis as seen by Pitts with wavelengths less than 250 nm.2,4 The fact that our changes were noted almost immediately certainly suggests a sooner photokeratitis response, but this may also suggest a more immediate process other than photochemical keratitis may be taking place. Pitts, however, does report that immediately following threshold photokeratitis irradiation that dead cells could be seen on the epithelial surface with clearing within the 1st hour.2 It may be possible that a similar process is occurring here with overlap of a very superficial short-lived photokeratitis if indeed this photokeratitis does exist. Nevertheless, the presence of faint haze several hours postexposure does suggest photokeratitis is present and that it follows the tendency of shorter penetration depth and shorter duration for wavelengths less than 250 nm as observed by Pitts.2
The temporal response of photokeratitis for exposures on the order of 10 J/cmp 2 was of much longer duration and, in many ways, very similar to the longer UV wavelength exposures of Pitts.2,4 Our studies showed an initial faint rose bengal staining which increased in density between 5 to 13 hours postexposure. At 13 to 15 hours postexposure, fluorescein staining could be seen which would last up to 19 to 26 hours. Pitts' results for wavelengths greater than 250 nm revealed epithelial haze and granules between 4 to 6 hours and fluorescein staining at 11 to 12 hours with dissipation at 24 hours postexposure.2-4 In both cases, a temporal response of increasing density and depth of photokeratitis with time could be seen, both roughly following a similar time course.
There have been attempts by several authors to determine the actual wavelengths of fluorescence produced at 193 nm. Tuft et al8 have shown two broad peaks of fluorescence at 310 and 460 nm. The former peak is within the range of sensitive wavelengths producing photokeratitis and extends from 260 to 320 nm within this range. Tuft et al in a later abstract9 and Muller-Stolzenburg et al10 also report fluorescence at 300 and 280 nm, respectively. At these wavelengths of fluorescence, the threshold for photokeratitis is relatively small (5 to 50 mJ/cmp 2). For photokeratitis to occur at a 193-nanometer exposure of 10 J/cmp 2 would suggest an efficiency of fluorescence of less than 1% at these wavelengths. This is a relatively small amount of fluorescence and other data suggest it may even be less.8
Figure 4 illustrates the comparative magnitude of energy required for photokeratitis at 193 nm. A single low energy density pulse has no effect until it exceeds the ablation threshold. At much higher pulse energy densities, optical plasma formation occurs, and it is only at these energy levels that sufficient energy exposure for photokeratitis is found. Therefore, any scattered or reflected radiation during excimer laser keratectomy would be insufficient at subablative magnitudes to cause a significant photochemical injury to the cornea. Even if one considered using 150 to 200 mJ/cmp 2/p during a myopic photorefractive keratectomy, it would take 50 to 67 pulses before the photokeratitis threshold would be achieved. This total energy would correspond only to the directlyirradiated tissue which at these high-pulse energy densities would result in corneal ablation rather that photokeratitis. Any scattered energy to adjacent cornea of observer cornea would be at least one order of magnitude less, would be subablative, and consequently it would take at least one order of magnitude greater number of pulses, 500 to 670, striking the same portion of the cornea as scattered radiation to achieve the photokeratitis threshold. Hence, even chronic scatter of excimer radiation is safe and the use of protective goggles during excimer laser keratectomy may not be necessary.
The threshold for a very mild form of photokeratitis characterized by a very superficial corneal haze in the rabbit cornea has been found to lie between 1.0 and 1.5 J/cmp 2 at 193 nm. The extremely superficial absorption of 193-nanometer laser radiation reported previously,6 as well as extrapolations of thresholds from longer wavelengths suggests that photokeratitis at 193 nm, if possible, would be of relatively high energy threshold, short duration, and superficial depth. Our experiments confirm this data, and suggest that photokeratitis at this wavelength is present almost immediately and/or may be coincident with and initial cell-surface devitalization. These results also imply that potential side effects and safety hazards from scattered 193nanometer laser radiation during excimer laser keratectomy will be remote because of the extremely short penetration depth.
A second, more classical photokeratitis is seen at a 193-nanometer threshold exposure level of 10 J/cmp 2. This more-deeply-penetrating, longerduration photokeratitis is felt to be secondary to fluorescent emission of longer UV wavelengths. Its presence at only the highest exposures suggest a very small efficiency of fluorescence and a relatively low risk clinical safety hazard.
1. Pitts DG, Kay KR. Photo-ophthalmic threshold for the rabbit. American Journal of Optometry and Archives of the American Academy of Optometry. 1969;46:561-572.
2. Pitts DG. A comparative study of the effects of ultraviolet radiation on the eye. American Journal of Optometry and Archives of the American Academy of Optometry. 1970;47: 535546.
3. Pitts DG, Tredici TJ. The effects of ultraviolet on the eye. Am Ind Hyg Assoc J 1971;32:235-246
4. Pitts DG. The human ultraviolet action spectrum. American Journal of Optometry and Physiologic Optics. 1974;51:946960.
5. Krueger RR, Trokel SL, Schubert HD. Interaction of ultraviolet laser light with the cornea. Invest Ophthalmol Vis Sci. 1985;26:1455-1464.
6. Puliafito CA, Wong K, Steinert RF. Quantitative and ultrastructural studies of excimer laser ablation of the cornea at 193 and 248 nanometers. Lasers Surg Med. 1987;7:155-159.
7. Taboada J, Archibald CJ. An extreme sensitivity in the corneal epithelium to far UV ArF excimer laser pulses. Proceedings of the scientific program, Aerospace Medical Association. San Antonio, Tex; 1981: 98-99.
8. Tuft SJ, Al-Dhahir R, Dyer P, Zehao Z. Characterization of the fluorescence spectra produced by excimer laser irradiation of the cornea. Proceedings at the International Corneal Laser Society. Sarasota, FIa; May, 1989.
9. Tuft SJ, Martin CA, Loree TR, Johnson TM. Quantification of the fluorescence spectra produced by ArF laser ablation of the cornea and sclera. Invest Ophthalmol Vis Sci. 1990(suppl); 31:447.
10. Muller-Stolzenburg NW, Buchwald HJ, Schruender S, et al. Corneal fluorescence under 193-nanometer excimer laser irradiation. Second International Congress Laser Technology in Ophthalmology (abstracts). Lasers and Light in Ophthalmology. 1989;2A.ll.
11. Sliney DH. The merits of an envelope action spectrum for ultraviolet radiation exposure criteria. Am Ind Hyg Assoc J. 1972;33:644-653.