First described in 2008, rainbow glare is a rare optical side effect of femtosecond laser–assisted in situ keratomileusis (FS-LASIK)1 in which patients complain immediately postoperatively of a rainbow-like effect when exposed to a polychromatic light source such as artificial lights. This is typically described as worse in the dark around automobile or street lights. It has been exclusively reported in FS-LASIK with various incidence rates depending on models and laser parameters, ranging from 19.07% with the Intra-Lase 15-kHz laser1 to 5.8% with the IntraLase 60-kHz laser.2 This difference was attributed to lower energy level, with higher frequency. The Wavelight FS-200 femtosecond laser (Alcon Laboratories, Inc), a new generation high-frequency femtosecond laser (200 kHz), was described to also induce rainbow glare in 2013.3 A recent study on the Wavelight FS-200 laser reported increased incidence of rainbow glare after a software update and a significant decrease following its downgrade attributed to reduced pulse energy and spot/line separation.4 Krueger et al1 reported an even lower incidence of 2.32%, using a newer laser model with increasing numerical aperture of the focusing optics, resulting in a smaller working distance under the cone. Finally, most of the studies report unilateral symptoms despite bilateral similar flap settings. It is therefore still unknown which factors predispose to rainbow glare development in an eye.
In a case of persistent unilateral rainbow glare,5 the authors reported an intraoperative image suggestive of a raster pattern of femtosecond laser impacts on the affected eye that was not apparent on the fellow asymptomatic eye. Such a finding was insufficient to conclude on its own, but reinforces the hypothesis that femtosecond laser patterns play an important role in the development of rainbow glare. Indeed, it was demonstrated on theoretical models that a random laser pattern for flap creation was the only one capable of entirely preventing the occurrence of rainbow glare,6 which may explain why the condition is exclusively described in FS-LASIK, because current femtosecond lasers are not able to randomize their laser impacts, as opposed to excimer lasers.
The presumed hypothesis is that rainbow glare is caused by a grating pattern of geometrically aligned, regularly spaced femtosecond laser impacts, resulting in diffractive light scattering.1 This has been further established by finding the spacing between laser impacts, from the size of the spectral bands described by the patient, and confirming it by in vivo visualization on confocal microscopy.3 Transient hyperreflective dots arranged in a grating pattern are often seen after LASIK and are attributed to tissue response to the laser impacts,7 but they seem to persist when rainbow glare is present.3,5 Regardless of the factors associated with rainbow glare, the condition seems to be caused by a uniform array of periodically aligned photodisruption defects that act as a likely source of the grating pattern, resulting in diffractive light scatter.
Often transient, no consensus exists regarding management of persistent rainbow glare, which can sometimes be incapacitating. Gatinel et al5 hypothesized that a planar photoablation of the posterior surface of the flap would smooth its undersurface and reduce its diffractive properties. After the first successful therapeutic 16-µm planar photoablation of the posterior surface of the flap (undersurface photoablation), we now report the case of a unilateral rainbow glare after Wavelight FS-200 laser treatment that required repeated undersurface photoablation using the Wavelight EX-500 excimer laser.
A 31-year-old man with bilateral myopia of 6.50 diopters underwent FS-LASIK (Wavelight FS-200 and EX-500). Surgery was uncomplicated, but the patient immediately experienced an incapacitating rainbow glare in the right eye. FS-200 laser settings were similar for both eyes, but a subtle raster pattern was noticed intraoperatively on the right eye (Figure A, available in the online version of this article).
Energy and spot parameters used for the flap creation. Parameters of the Wavelight FS-200 femtosecond laser (Alcon Laboratories, Inc) were similar for both eyes, but the picture taken immediately after flap creation of the (A) right eye is evoking a raster shot pattern, not seen on the (B) left eye.
Postoperative uncorrected distance visual acuity (UDVA) in the right eye was 20/16, and corrected distance visual acuity (CDVA) was 20/16 (+0.75 −0.50 × 10°). Slit-lamp examination of the flap and investigations of the cornea were unremarkable except for hyperreflective dots arranged in a regular grating pattern on confocal microscopy (Figure 1).
Confocal microscopy of the right eye at (A) 21 months after first myopic treatment, (B) 1 day after first revision, and (C) 3 months after second revision. Regularly aligned hyperreflective femtosecond laser dots can be seen around the flap, which progressively disappear along with symptoms after each undersurface photoablation.
Despite excellent UDVA, the patient complained of intense rainbow glare, which was detrimental to his life and did not improve 21 months after the treatment. Interestingly, the patient provided drawings of his symptoms at the time and their evolution (Figure 2A). In light of the subjective impact on the patient's quality of life, a 10-µm undersurface photoablation was performed (Figure BA, available in the online version of this article) as previously described5 with immediate but incomplete improvement of both subjective symptoms and objective reduction of the grating pattern on confocal microscopy (Figure 1B). Briefly, the center of the pupil was marked on the cornea and lateral marks were added to improve flap repositioning. The flap was lifted using the previously created rhexis and turned backward, epithelium side facing downward, lying on a flat-domed spatula. The undersurface of the flap or the stromal side of the flap was prepared and smoothed using a wet sponge. The patient was asked to look downward and, after careful alignment of the Wavelight EX500 laser with the centrally marked flap, the treatment was delivered on the stromal side of the flap. Using a canula, the flap was then replaced and the interface was abundantly irrigated (Figure 3).
Illustrations provided by the patient describing his symptoms. (A) When exposed to a polychromatic light source directly shined into his eyes, the patient experienced and drew himself the optical abnormalities he saw. (B) After the first rainbow glare treatment by flap undersur-face photoablation, the patient had milder but persisting symptoms. The lateral spectral bands were attenuated, but the vertical lines were still described as “intense.” (C) This led to a second treatment, with complete resolution of the symptoms.
Undersurface flap photoablation treatment settings. (A) First revision was composed of a 10-µm undersurface flap photoablation and provided significant but limited improved, and (B) a second 10-µm undersurface flap photoablation was required.
Surgical steps of the undersurface flap photokeratectomy. (A) The center of the preexisting flap was marked and then (B) flipped over, epithelium side facing downward, lying and being held against a flat-domed spatula. (C) Undersurface flap photokeratectomy was then applied to the undersurface of the flap.
Twelve months after the revision, UDVA remained 20/16 and CDVA was 20/16 (+0.75 −0.25 × 65°) but the patient asked for additional treatment because of the persistence of a milder form of rainbow glare. Indeed, he provided a drawing of his vision when exposed to polychromatic light (flashlight from smartphone) for the past months after surgery (Figure 2B), which showed persisting vertical rainbow glare lines, or vertical spectral bands, and an attenuation of the lateral spectral bands. Although improvement was obtained after the first treatment, the persistence of incapacitating symptoms after 12 months led to another 10-µm undersurface photoablation (Figure BB), this time with complete resolution (Figure 2C). On confocal microscopy, although rare hyperreflective dots could still be seen, the grating pattern completely disappeared (Figure 1C). At last follow-up, 6 months after the second revision, UDVA was 20/12.5 and CDVA was 20/12.5 (+0.75 −0.25 × 5°) in the right eye with total absence of rainbow glare and no hypermetropic shift was observed despite a total undersurface ablation of 20 µm (Figure 4).
OPD scan of the right eye 6 months after the last treatment. Interestingly, no hypermetropic shift was seen after the undersurface photoablation.
Even with the latest technological advances in refractive surgery (especially FS-LASIK surgery), rainbow glare is a known postoperative complication, which luckily remains a benign and often spontaneously resolving optic condition. However, it sometimes causes dramatic life impairment and its impact on patients' lives should not be discarded. Although its theoretical cause has now been well described,1,3,5,6 it is still unclear as to how to best prevent its appearance and chronic persistence.
Although flap parameters and laser properties must play a definite role, the fact that it is not rare to see uni-lateral rainbow glare despite similar femtosecond laser flap parameters bilaterally is evocative of a subtle interaction between a multitude of factors. Fortunately, most rainbow glares are transient and self-resolving, but treatment must be offered to patients with persistent and debilitating rainbow glare, especially given the usual young, healthy patients to whom refractive surgery is offered.
To the best of the authors' knowledge, only under-surface flap photoablation has been described as a successful rainbow glare treatment. This technique is limited by the residual flap thickness to diminish the risk of ectasia, flap foldings, and hyperopia induced by overtreatment. In this case, flap parameters were known and the residual stromal component of the flap was measured with preoperative corneal optical coherence tomography, but in thinner flaps or with more extensive undersurface photoablation, the possibility to ablate through the residual stromal component of the flap is a risk one must carefully assess. On the other hand, insufficient photoablation of the undersur-face of the flap in case of rainbow glare (as it happened during the first revision of this case) would likely result in persisting symptoms and therefore incomplete patient satisfaction, possible repeated surgeries, and revision-induced complications. It is therefore key for the surgeon to find the adequate parameters for under-surface flap ablation. In another study, a 16-µm under-surface flap photoablation was sufficient for complete resolution of symptoms, and in this patient, a depth of 2 x 10 µm was required. We therefore do not recommend performing undersurface flap photoablation of less than 16 µm, and advise that at least 16 to 20 µm be ablated.
In our case, after the first revision, the main persisting symptoms were the vertical spectral lines. The lateral spectral bands constituting the diffractive pattern in rainbow glare are easily comprehensible when considering the horizontal regular spacing of the femtosecond laser spots. However, there are invariably bright vertical lines with the typical spectral dispersion associated with rainbow glare (Figure 5) that cannot be explained by the even spot distribution. Because each horizontal line is shifted due to the line-wrap of the circular perimeter, the spots are not vertically aligned. However, the lines created by the spots simulate regularly spaced thin diffractive slits, which could explain the origin of the vertical spectral lines. In our experience, these vertical spectral lines are more often described, and tend to be more intensely perceived by patients, which is unsurprising given their mechanism. Until femtosecond lasers are able to create a flap using a completely random pattern of impacts, which would supposedly eradicate the incidence of rainbow glare, an area of possible and more easily achievable improvement may be to randomly space the parallel horizontal lines in an attempt to reduce the risk of the vertical spectral lines.
Schematic representations of the cause of vertical rainbow glare (image provided courtesy of Damien Gatinel, MD, PhD). Although the regular spacing of the horizontal femtosecond laser impacts cause the horizontal diffraction and the constitution of the lateral spectral bands, it might be disconcerting to understand how the vertical lines are created because vertical impacts are slightly misaligned because the start of each horizontal line is offset because of the peripheral shift along the circular shape of the flap. This is supposedly due to the lines created by the parallel spots that simulate small regularly spaced thin diffractive slits, causing the vertical spectral lines.
This case reinforces the hypothesis that rainbow glare is caused by a uniform array of periodically aligned photodisruption defects that act as a likely source of the grating pattern, resulting in diffractive light scatter. Undersurface photoablation has been successful in treating rainbow glare and its efficacy can objectively be followed by confocal microscopy through progressively disappearing hyperreflective femtosecond laser impacts. Moreover, this case shows that undersurface photoablation can be repeated if needed, although a single treatment would be preferred. Undersurface photoablation is an efficient treatment of rainbow glare, and should include a thickness of at least 16 to 20 µm.
- Krueger RR, Thornton IL, Xu M, Bor Z, van den Berg TJ. Rainbow glare as an optical side effect of IntraLASIK. Ophthalmology. 2008;115(7):1187–1195.e1. doi:10.1016/j.ophtha.2007.10.007 [CrossRef]
- Bamba S, Rocha KM, Ramos-Esteban JC, Krueger RR. Incidence of rainbow glare after laser in situ keratomileusis flap creation with a 60 kHz femtosecond laser. J Cataract Refract Surg. 2009;35(6):1082–1086. doi:10.1016/j.jcrs.2009.01.026 [CrossRef]
- Gatinel D, Saad A, Guilbert E, Rouger H. Unilateral rainbow glare after uncomplicated femto-LASIK using the FS-200 femtosecond laser. J Refract Surg. 2013;29(7):498–501. doi:10.3928/1081597X-20130426-01 [CrossRef]
- Zhang Y, Chen YG. High incidence of rainbow glare after femtosecond laser assisted-LASIK using the upgraded FS200 femtosecond laser. BMC Ophthalmol. 2018;18(1):71. doi:10.1186/s12886-018-0734-1 [CrossRef]
- Gatinel D, Saad A, Guilbert E, Rouger H. Simultaneous correction of unilateral rainbow glare and residual astigmatism by undersurface flap photoablation after femtosecond laser-assisted LASIK. J Refract Surg. 2015;31(6):406–410. doi:10.3928/108 1597X-20150521-07 [CrossRef]
- Ackermann R, Kammel R, Merker M, Kamm A, Tünnermann A, Nolte S. Optical side-effects of FS-laser treatment in refractive surgery investigated by means of a model eye. Biomed Opt Express. 2013;4(2):220–229. doi:10.1364/BOE.4.000220 [CrossRef]
- Sonigo B, Iordanidou V, Chong-Sit D, et al. In vivo corneal confocal microscopy comparison of intralase femtosecond laser and mechanical microkeratome for laser in situ keratomileusis. Invest Ophthalmol Vis Sci. 2006;47(7):2803–2811. doi:10.1167/iovs.05-1207 [CrossRef]