- Email: [email protected]
Diffuse lamellar keratitis: evaluation of etiology, histopathologic findings, and clinical implications in an experimental animal model
laboratory science Diffuse lamellar keratitis: Evaluation of etiology, histopathologic findings, and clinical implications in an experimental animal model Mike P. Holzer, MD, Kerry D. Solomon, MD, David T. Vroman, MD, Luis G. Vargas, MD, Helga P. Sandoval, MD, Terrance J. Kasper, MD, David J. Apple, MD
Purpose: To induce diffuse lamellar keratitis (DLK) and investigate the potential causative agents in an animal model. Setting: Magill Research Center for Vision Correction, Storm Eye Institute, Medical University of South Carolina, Charleston, South Carolina, USA. Methods: In 70 eyes of 35 Dutch Belted rabbits, a corneal flap was cut and the interface randomly exposed to 1 of 7 substances: Pseudomonas aeruginosa endotoxin, 1 of 2 Staphylococcus aureus exotoxins, meibomian gland secretion, povidone⫺iodine 10%, Palmolive® Ultra soap, and Klenzyme姞 soap. Slitlamp examinations were performed 1, 3, 5, and 7 days postoperatively. The DLK was staged from 1 to 4. On day 7, the rabbits were killed and the eyes enucleated and processed for histopathologic examination. Results: At the end of the study, 54 eyes (46 exposed, 8 control) were available for evaluation. The 8 eyes studied concurrently in the control group remained clear and did not show interface inflammation. Thirty-one of 46 eyes (67%) treated with the various test substances developed DLK. The highest DLK rates were found with the cleaning soap Palmolive Ultra (100%; P ⫽ .022) and P aeruginosa lipopolysaccharide endotoxin (90%; P ⫽ .026). Conclusions: Interface inflammation was consistently induced in the animal model. All 7 agents caused DLK in at least some eyes. The histopathologic evaluation showed the morphologic profile of the marked inflammatory cellular reaction that occurred in almost all the specimens. J Cataract Refract Surg 2003; 29:542–549 © 2003 ASCRS and ESCRS
L
aser in situ keratomileusis (LASIK) is the most popular refractive surgery procedure for the correction of myopia and hyperopia with or without astigmatism.1–3 Although the surgical techniques and instruments for this procedure are extremely refined, postoperative complications may occur.4 Diffuse lamellar keratitis (DLK), a nonspecific inflammatory reaction Accepted for publication August 6, 2002. Reprint requests to Kerry D. Solomon, MD, Magill Research Center for Vision Correction, Medical University of South Carolina, Storm Eye Institute, 167 Ashley Avenue, Charleston, South Carolina 29425, USA. © 2003 ASCRS and ESCRS Published by Elsevier Science Inc.
at the interface level following LASIK, was first described in 1998 by Smith and Maloney.5 Clinically, it presents as a white granular infiltrate, usually appearing 1 to 5 days after surgery. It can, however, also occur in specific situations long after the procedure; for example, following epithelial defects or trauma.6 – 8 Maddox and Hatsis presented the same findings and based on the infiltrate’s characteristic wavy appearance, they termed it “Sands of the Sahara Syndrome” (A. Hatsis, MD, R. Maddox, MD, “Sands of the Sahara,” poster presented at the Symposium on Cataract, IOL and Refractive Surgery, San Diego, California, USA, April 1998). 0886-3350/03/$–see front matter doi:10.1016/S0886-3350(02)01691-7
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
The exact etiology of DLK is unknown, although multiple factors have been implicated in the development of the condition. Several factors have been reported to be associated with DLK, including epithelial defects, trauma, meibomian gland secretions, and bacterial endotoxins.6 –16 This study used an animal model to induce DLK and investigate possible causative factors related to its development.
Materials and Methods Seventy eyes of 35 female Dutch Belted rabbits weighing 1.5 to 2.0 kg were used. The rabbits were treated according to the guidelines of the Association for Research in Vision and Ophthalmology and the Division of Laboratory Animal Resources at the Medical University of South Carolina. They were randomly assigned to have their corneal interface exposed to 1 of 7 substances. A separate group of rabbits without substance exposure was used as the control group. Table 1 shows the groups and the substances. The rabbits were anesthetized with an intramuscular injection of a 1:1 mixture of ketamine hydrochloride (Ketaset威) and xylazine (Rompum威). A corneal flap was cut with a nasal or temporal hinge position using the Automated Corneal Shaper威 microkeratome (Bausch & Lomb Surgical) and an 8.5 mm suction ring with a 160 m plate. A new blade was used in each rabbit. The eyes were not proptosed, and posterior pressure was not necessary to cut the flaps. The flap was
Table 1. Groups and substances used in the study.
Group
Number of Eyes
1
11
2
9
Highly purified Staphylococcus aureus exfoliative toxin A (ETA) (Toxin Technology Inc.)
3
11
Highly purified Staphylococcus aureus enterotoxin B (SEB) (Toxin Technology Inc.)
4
9
Palmolive姞 Ultra soap (Colgate-Palmolive Co.)
5
6
Klenzyme姞 enzymatic cleaner (Steris Co.)
6
6
Povidone–iodine 10% (Betadine姞, Professional Disposable International)
7
8
Meibomian gland secretion
8
10
Substance Pseudomonas aeruginosa lipopolysaccharide (LPS) endotoxin serotype 10 (Sigma-Aldrich)
Control
lifted with a Lindstrom spatula (American Surgical Instruments Corp.), and the agents were exposed to the stromal bed using a Merocel威 sponge (Medtronic Solan), which was pressed down slightly on the stroma once. This was done assuming that the amount of liquid absorbed by each sponge and later released to the interface was the same in all cases. Measurements in our laboratory show that if a Merocel sponge is pressed on a straight surface, approximately 40 L of fluid are released. The concentrations of the agents were as follows: LPS endotoxin 2.5 mg/mL; both Staphylococcus aureus toxins 0.075 mg/mL; povidone⫺iodine 10%, Klenzyme, and Palmolive Ultra were diluted 1:50. Concentrations of the specific bacterial toxins were determined following reports of rabbit studies.17,18 Meibomian gland secretions were obtained by pushing the upper lid of a rabbit eye toward a Lindstrom spatula. Secretions were exposed to the central part of the stromal bed, followed by repositioning of the flap and 1-minute drying time. In the control group, the flap was repositioned without exposing the stromal bed to a substance. All surgeries were completed without laser ablation. Postoperatively, the rabbits were examined by the same unmasked examiner on days 1, 3, 5, and 7 with a slitlamp microscope (Topcon SL-7E). The corneal findings were documented and slitlamp photographs taken. The DLK findings in each eye were staged 1 to 4 following the system of Linebarger and coauthors19: stage 1, white granular infiltrates in the periphery of the lamellar flap, outside the visual axis; stage 2, white infiltrates in the center of the flap, involving the visual axis; stage 3, more dense, white, clumped infiltrates in the central visual axis; and stage 4, stromal melting, permanent scarring, and visual morbidity. The rabbits received no topical or systemic medication at any time postoperatively. On day 7, the rabbits were anesthetized with a 1:1 mixture of ketamine hydrochloride and xylazine and then killed with an intravenous injection of 2 mL pentobarbital (Sleepaway威). The eyes were immediately enucleated, and the entire corneas were excised and fixed in neutral formalin 10% for 24 hours. The corneas were prepared for serial paraffin sections and evaluated with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), and Masson trichrome stains. Corneal tissues from 46 animal eyes treated to induce opacities resembling those seen clinically in DLK were examined histopathologically by light microscopy. In the histopathologic examination of these eyes, emphasis was placed on evaluation of the possible infiltration of inflammatory cells within anatomic structures of the cornea, especially within the stroma and near the interface. The degree of wound-healing activity was also evaluated, and the results were compared with the findings in the 8 control eyes. Statistical analysis was performed using the Fisher exact test. A P value less than 0.05 was considered statistically significant.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
543
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Results
Nine of 10 eyes (90%) exposed to Pseudomonas aeruginosa LPS endotoxin developed DLK. Compared to the control group, this was statistically significant (P ⫽ .026). The DLK rates ranged between 29% and 60% with the other agents (S aureus toxins ETA and SEB, meibomian gland secretion, and povidone⫺iodine); however, there were no statistically significant differences. The clinical grading (Table 3) showed that P aeruginosa LPS endotoxin caused the most severe grades of DLK; 7 of 9 DLK eyes were stage 2 or 3. In the Palmolive Ultra group, 1 of 7 eyes was stage 3. All other eyes were stage 1. The DLK grade was highest in the LPS group (1.8) followed by the Klenzyme (1.5) and Palmolive Ultra (1.28) groups. Even though the groups were too small to determine a statistically significant difference, the trends shown in Table 3 correlate well with the findings noted in Table 2.
At the end of the study, 54 eyes were included for analysis. Sixteen eyes were excluded because of a slipped (n ⫽ 12) or free (n ⫽ 3) flap; in 1 eye, it was not possible to cut a flap because of insufficient suction by the microkeratome apparatus. The group sizes after exclusion of these 16 eyes are shown in Table 2. Clinical Findings Throughout the 7-day follow-up, the 8 eyes in the control group remained clear and did not show interface inflammation. Thirty-one of the 46 eyes (67%) exposed to 1 of the substances developed DLK. This was statistically significant (P ⫽ .024) compared to the control group (0%) (Table 2). The appearance of DLK varied from slightly granular with a faint haze to clumpy white granules in the entire interface. In most cases, DLK first appeared on day 3. To evaluate the possibility of infection in the eyes, scraping samples for chocolate agar culture from 3 suspicious corneal interfaces were taken. None of these showed growth of microorganisms. The highest rate of DLK was in the soap groups, with all eyes developing DLK. The findings in the Palmolive Ultra group were statistically significant (P ⫽ .022) compared to those in the control group. The Klenzyme group was too small to determine a statistically significant difference.
Histopathologic Findings In the control corneas (Figure 1), the staining (H&E and Masson trichrome) revealed a typical and relatively homogeneous normal pattern. A hyperplastic epithelial plug (facette) at the edge of the incision with thickening of up to 10 epithelial cell layers marked the beginning of the incision, and the number of keratocytes was slightly increased in this area. Despite newly formed, irregularly arranged collagen fibers in the microkeratome wound site, the stroma generally main-
Table 2. Number of inoculated and control eyes and postoperative DLK rates. Eyes with DLK
Agent
Total Number of Eyes
Eyes Included
Number
Control
10
8
0
%
P Value (Versus Control)
Palmolive Ultra
9
7
7
100.0
.022*
Klenzyme
6
4
4
100.0
.077
11
10
9
90.0
.026*
8
5
3
60.0
.2
LPS Meibomian gland secretion Povidone–iodine
6
5
3
60.0
.2
SEB
11
8
3
37.5
.228
ETA
9
7
2
28.6
.471
Total
70
54
31
—
—
Total without controls
60
46
31
67.4
.024*
SEB ⫽ Staphylococcus aureus enterotoxin B, ETA ⫽ Staphylococcus aureus exfoliative toxin A, LPS ⫽ Pseudomonas aeruginosa LPS endotoxin *Statistically significant
544
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
tained its lamellar structure unaltered. The number of keratocytes increased slightly in the wound margin but not in the wound center. Although a distinct tendency was noted, the histopathologic morphology in each case of DLK did not always correlate with the clinical appearance. For example, the clinical reaction to Klenzyme was severe, but the histopathologic appearance was moderate. Compared to the reaction in the control eyes, it had an increased number of keratocytes and inflammatory cells in the anterior stroma. The histopathologic examination of the cornea exposed to P aeruginosa LPS endotoxin revealed a marked cellular reaction with inflammatory cells, keratocytes, and collagen formation. It was determined marked and correlated well with the clinical appearance. Figures 2 and 3 show an eye exposed to S aureus exotoxin ETA and a clinical DLK-like inflammatory reaction. The histopathological specimen revealed a marked inflammatory cellular reaction with plasma cells including Russel bodies, eosinophil granulocytes, and lymphocytes in the corneal stroma around the interface.
Discussion Refractive surgery is a rapidly growing field, and LASIK has become the predominant procedure for the correction of myopia and hyperopia with or without astigmatism.1 However, as the number of LASIK proceTable 3. Diffuse lamellar keratitis score based on slitlamp examination (highest score for each agent during the 7-day postoperative period).
Agent
Mean DLK Score ⴞ SD
Standard Error
LPS
1.80 ⫾ 0.91
0.29
Klenzyme
1.50 ⫾ 0.58
0.29
Palmolive Ultra
1.28 ⫾ 0.75
0.28
Povidone-iodine
1.00 ⫾ 1.00
0.44
SEB
0.87 ⫾ 1.35
0.48
ETA
0.71 ⫾ 1.25
0.47
Meibomian gland secretion
0.60 ⫾ 0.54
0.24
Differences in the mean values among the treatment groups were not significant (P ⫽ .301, 1-way repeated-measures analysis of variance). LPS ⫽ Pseudomonas aeruginosa lipopolysaccharide; SEB ⫽ Staphylococcus aureus enterotoxin B; ETA ⫽ Staphylococcus aureus exfoliative toxin A DLK scores: stage 1 ⫽ 1; stage 2 ⫽ 2, etc.
dures has increased, the number of complications has also risen and more surgeons are encountering DLK. In their 1998 description of DLK, Smith and Maloney5 reported 13 eyes of 12 patients with noninfectious diffuse infiltrates in the lamellar interface between 2 and 6 days after LASIK. The clinical evolution of DLK is characterized by the appearance of white granular infiltrates in the stromal interface 1 to 5 days after LASIK. The infiltrate normally starts at the periphery of the corneal flap and has to be distinguished from epithelial cells, debris, and meibomian secretions.20 Clinically, DLK can be graded in 4 stages; this study followed the staging system described by Linebarger and coauthors.19 To date, the exact etiology as well as the specific prophylactic treatment regimens for DLK are unknown. There are reports linking epithelial defects, trauma, meibomian gland secretions, and bacterial endotoxins to DLK.6 –11,13–16 To our knowledge, this is the first extensive experimental study testing several possible causative agents and the first to look at and report the histopathology of corneas with DLK. Although DLK development in the rabbit model may not completely equate to that in the human eye, an experimental setting to investigate this disease in more detail is necessary. The results of the study showed that all tested agents could cause an inflammatory reaction at the interface level. This may confirm the theory that DLK has multiple etiologies. Some of the agents resulted in very high DLK rates: Palmolive Ultra (100%), Klenzyme (100%), and P aeruginosa LPS endotoxin (90%). In addition to the DLK rates, the intensity of the DLK response also varied with P aeruginosa LPS endotoxin and the soaps (Palmolive Ultra and Klenzyme), with the most intense inflammatory response following inoculation with LPS endotoxin. The finding that the soaps caused DLK in all eyes shows the importance of thoroughly rinsing surgical instruments, especially the microkeratomes. Soap residue left on instruments may gain access to the corneal stroma and cause DLK. Many microkeratome manufacturers recommend Palmolive Ultra to clean their instruments due to its lubricating effect, and the soap has been used by many surgeons. Recently, some manufacturers changed the cleaning recommendations for their microkeratomes from Palmolive Ultra to specific medical-device cleaners similar to the enzymatic cleaner Klenzyme.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
545
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Figure 1. (Holzer) Cornea of a control eye 7 days postoperatively. The incision is recognizable as a small epithelial plug (facette). The corneal flap (toward the right) covers the anterior one third of the stroma. The epithelium reveals a normal pattern except for the facette at the incision site, which consists of a thickening of up to 8 epithelial cell layers. A slightly increased number of keratocytes (arrows) are seen around the incision area. EP ⫽ epithelial plug, F ⫽ flap, I ⫽ interface, SB ⫽ stromal bed. a: H&E stain, original magnification ⫻200. b: Masson trichrome stain, original magnification ⫻200.
Figure 2. (Holzer) Cornea with a marked inflammatory reaction 7 days postoperatively. Clinically, the eye showed a DLK-like inflammatory reaction, which was induced by S aureus exotoxin ETA. The interface is located in the anterior one fourth of the cornea. Compared to the control eye (Figure 1), the photomicrographs show a marked cellular reaction with inflammatory cells and keratocytes. The inflammatory reaction consists of plasma cells including Russel bodies, scattered eosinophil granulocytes, and lymphocytes in the flap periphery. a: Masson trichrome stain, original magnification ⫻100. b: Masson trichrome stain, original magnification ⫻200.
Both cleaning agents were tested in this study, and both caused DLK in every eye. This was statistically significant when the Palmolive Ultra group was compared to the control group. However, because of the small number of eyes (n ⫽ 4), there was no statistically significant difference detectable for the Klenzyme group. Further investigation to identify the ideal cleaner and the optimal cleaning regimen for microkeratomes and other sur546
gical instruments is necessary to find a standardized cleaning protocol that is highly efficient and also practical for use during a routine surgical day. The results of the bacterial toxins confirm that these agents have the potential to cause DLK. Holland et al.15 were the first to report DLK related to endotoxins. They found a biofilm from the gram-negative bacteria Burkholderia pickettii in the reservoir of their sterilizer and
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Figure 3. (Holzer) Cornea with a marked inflammatory reaction 7 days postoperatively. Clinically, the eye showed a DLK-like inflammatory reaction, which was induced by S aureus exotoxin ETA. Compared to the control eye (Figure 1), the photomicrographs show a marked cellular reaction with inflammatory cells and keratocytes. The inflammatory reaction consists of plasma cells and scattered eosinophil granulocytes. a: PAS stain, original magnification ⫻400. b: H&E stain, original magnification ⫻400.
postulated that bacteria are released into the distilled water of the sterilizer reservoir. These bacteria would be killed during the sterilization cycle and endotoxins, which are a component of the outer membrane of gram-negative bacteria, would be released after the cells lysed. The endotoxins may persist through the sterilization process, adhere to the instruments, and be deposited in the interface during surgery. In high doses, endotoxins are toxic and can cause inflammation, fever, anaphylactic shock, blood coagulation, and sepsis.21 If present in the corneal interface, endotoxins may cause an inflammatory reaction, which was seen in the histopathology sections of the corneas with clinical DLK. In a small animal study by Peters et al.,16 clinical DLK developed in 4 rabbit eyes within 3 days of inoculation with Pseudomonas LPS endotoxin. Peters et al. also examined sources of endotoxin contamination in eye centers and detected them in tap water, filtered and distilled water, instrument washbasins, and sterilizer reservoirs. In our study, P aeruginosa LPS endotoxin caused DLK in 9 of 10 eyes, most of which had more severe stages than the other eyes that developed DLK. Therefore, surgeons may want to consider microbiologic examination of their sterilizers and the use of special cleaning procedures to lower endotoxin concentrations and minimize DLK occurrences.
The S aureus exotoxins SEB and ETA also caused DLK but in a relatively smaller proportion of eyes (38% and 29%, respectively). This lower DLK rate might be related to the concentration of the toxins used. Exotoxins are generally an extremely potent stimulus for T-lymphocytes and inflammatory reactions.22 In the histopathological sections of the rabbit corneas, inflammatory cells, including lymphocytes and plasma cells, were found in the corneas that were clinically diagnosed with DLK. As S aureus can be found in the normal human eyelid flora,23 surgeons should pay careful attention to proper lid hygiene before and during surgery to avoid contamination of the stromal bed. In this study, meibomian gland secretions also caused DLK. However, because of the way the gland secretion was obtained in the rabbit eyes, bacteria and/or bacterial endotoxins/exotoxins may also have been exposed to the interface. There is 1 report of the possible correlation between meibomian secretion and DLK.14 The authors did not, however, find a correlation between their patients with DLK and the presence of significant meibomianitis or blepharitis. In a survey of refractive surgeons, lid margin disease was the factor most commonly thought to be related to DLK.24 Therefore, in our opinion, lid hygiene, including scrubbing and draping the lid margins, may be important to min-
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
547
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
imize DLK after LASIK, and patients with blepharitis, which is associated with Staphylococcus epidermidis and Propionibacterium acnes,23 should not be scheduled for LASIK until effectively treated. Povidone⫺iodine 10%, commonly used in a diluted solution prior to LASIK to scrub eyelids and lashes, also caused DLK in the rabbit eyes. In their initial report describing DLK, Smith and Maloney5 hypothesize that povidone⫺iodine may be a causative factor. In an animal study by Alp et al.,25 the toxicity of povidone⫺iodine 5% and 10% to the corneal endothelium was shown. Similar toxic effects on the corneal stroma may be a reason for the development of DLK. Although povidone⫺iodine is important for scrubbing the lids and for infection prophylaxis, surgeons should be careful to avoid contact of povidone⫺iodine with the corneal stroma, which may cause DLK. The animal model used in this study and the causative agents with the highest DLK rate (P aeruginosa LPS endotoxin and cleaning soaps) may help to establish an investigational setting for DLK. A critical point of this study, however, might be that it was not observer masked, which will be addressed in future studies. With the help of standardized studies, further research on DLK as well as improved prevention and treatment should be possible. Although the specific pathophysiologic reactions of DLK in humans are still unknown, the findings of this study may confirm that it is a multifactorial disease. There is some evidence that inflammatory reactions are correlated with DLK. The histopathologic examination of the eyes exposed to endotoxins and exotoxins showed severe responses with an inflammatory cell reaction, eyes that were previously clinically diagnosed with DLK. The histopathologic examination of the eyes did not always correlate with the clinical findings. This might be because the pathologic sections were done at the end of the study on day 7 and the peak of the DLK reaction had passed. Foreign material that gains access to the interface during surgery may be the stimulus for an inflammatory response that may be seen clinically as DLK. Therefore, it is important to avoid the inoculation of harmful agents into the stromal bed. Additionally, microbiological investigation of sterilizers and water sources used in LASIK facilities to detect possible high concentrations of endotoxins or exotoxins may be considered to minimize the incidence of DLK. 548
This study used an animal model for DLK and investigated possible causative agents. Additional studies are needed to learn more about the detailed pathophysiologic process of DLK and to better understand the treatment and prevention of this disease after LASIK.
References 1. Pallikaris IG, Siganos DS. Laser in situ keratomileusis to treat myopia: early experience. J Cataract Refract Surg 1997; 23:39 –49 2. Salchow DJ, Zirm ME, Stieldorf C, Parisi A. Laser-insitu keratomileusis (LASIK) zur Myopie- und Astigmatismuskorrektur. Ophthalmologe 1998; 95:142–147 3. Knorz MC, Jendritza B, Liermann A, et al. LASIK zur Myopiakorrektur; 2-Jahres-Ergebnisse. Ophthalmologe 1998; 95:494 –498 4. Ambro´ sio R Jr, Wilson SE. Complications of laser in situ keratomileusis: etiology, prevention, and treatment. J Refract Surg 2001; 17:350 –379 5. Smith RJ, Maloney RK. Diffuse lamellar keratitis; a new syndrome in lamellar refractive surgery. Ophthalmology 1998; 105:1721–1726 6. Haw WW, Manche EE. Late onset diffuse lamellar keratitis associated with an epithelial defect in six eyes. J Refract Surg 2000; 16:744 –748 7. Harrison DA, Periman LM. Diffuse lamellar keratitis associated with recurrent corneal erosions after laser in situ keratomileusis. J Refract Surg 2001; 17:463–465 8. Weisenthal RW. Diffuse lamellar keratitis induced by trauma 6 months after laser in situ keratomileusis. J Refract Surg 2000; 16:749 –751 9. Johnson JD, Harissi-Dagher M, Pineda R, et al. Diffuse lamellar keratitis: incidence, associations, outcomes, and a new classification system. J Cataract Refract Surg 2001; 27:1560 –1566 10. Shah MN, Misra M, Wilhelmus KR, Koch DD. Diffuse lamellar keratitis associated with epithelial defects after laser in situ keratomileusis. J Cataract Refract Surg 2000; 26:1312–1318 11. Steinert RF, McColgin AZ, White A, Horsburgh GM. Diffuse interface keratitis after laser in situ keratomileusis (LASIK): a nonspecific syndrome. Am J Ophthalmol 2000; 129:380 –381 12. Yavitz EQ. Diffuse lamellar keratitis caused by mechanical disruption of epithelium 60 days after LASIK [letter]. J Refract Surg 2001; 17:621 13. Schwartz GS, Park DH, Schloff S, Lane SS. Traumatic flap displacement and subsequent diffuse lamellar keratitis after laser in situ keratomileusis. J Cataract Refract Surg 2001; 27:781–783 14. Fogla R, Rao SK, Padmanabhan P. Diffuse lamellar ker-
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
15.
16.
17.
18.
19.
20.
atitis: are meibomian secretions responsible? [letter]. J Cataract Refract Surg 2001; 27:493–495 Holland SP, Mathias RG, Morck DW, et al. Diffuse lamellar keratitis related to endotoxins released from sterilizer reservoir biofilms. Ophthalmology 2000; 107: 1227–1233; discussion by EJ Holland, 1233–1234 Peters NT, Iskander NG, Penno EEA, et al. Diffuse lamellar keratitis: isolation of endotoxin and demonstration of the inflammatory potential in a rabbit laser in situ keratomileusis model. J Cataract Refract Surg 2001; 27: 917–923 Siqueira JA, Speeg-Schatz C, Freitas FI, et al. Channelforming leucotoxins from Staphylococcus aureus cause severe inflammatory reactions in a rabbit eye model. J Med Microbiol 1997; 46:486 –494 Schultz CL, Morck DW, McKay SG, et al. Lipopolysaccharide induced acute red eye and corneal ulcers. Exp Eye Res 1997; 64:3–9 Linebarger EJ, Hardten DR, Lindstrom RL. Diffuse lamellar keratitis: diagnosis and management. J Cataract Refract Surg 2000; 26:1072–1077 Bu¨ hren J, Baumeister M, Kohnen T. Diffuse lamellar keratitis after laser in situ keratomileusis imaged by confocal microscopy. Ophthalmology 2001; 108: 1075–1081
21. O’Reilly M, Newcomb DE, Remick D. Endotoxin, sepsis, and the primrose path. Shock 1999; 12:411–420 22. Marrack P, Kappler J. The staphylococcal enterotoxins and their relatives. Science 1990; 248:705–711; erratum, 1066 23. Groden LR, Murphy B, Rodnite J, Genvert GI. Lid flora in blepharitis. Cornea 1991; 10:50 –53 24. Solomon KD, Holzer MP, Sandoval HP, et al. Refractive surgery survey 2001. J Cataract Refract Surg 2002; 28: 346 –355 25. Alp BN, Elibol O, Sargon MF, et al. The effect of povidone iodine on the corneal endothelium. Cornea 2000; 19:546 –550 From the Magill Research Center for Vision Correction, Storm Eye Institute, Medical University of South Carolina, Charleston, South Carolina, USA. Presented in part at the Symposium on Cataract, IOL and Refractive Surgery, San Diego, California, USA, April 2001, and as a poster at the annual meeting of the American Academy of Ophthalmology, New Orleans, Louisiana, USA, November 2001. Supported in part by an unrestricted grant from Research to Prevent Blindness, New York, New York, USA. None of the authors has a financial interest in any product mentioned.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
549
Purpose: To induce diffuse lamellar keratitis (DLK) and investigate the potential causative agents in an animal model. Setting: Magill Research Center for Vision Correction, Storm Eye Institute, Medical University of South Carolina, Charleston, South Carolina, USA. Methods: In 70 eyes of 35 Dutch Belted rabbits, a corneal flap was cut and the interface randomly exposed to 1 of 7 substances: Pseudomonas aeruginosa endotoxin, 1 of 2 Staphylococcus aureus exotoxins, meibomian gland secretion, povidone⫺iodine 10%, Palmolive® Ultra soap, and Klenzyme姞 soap. Slitlamp examinations were performed 1, 3, 5, and 7 days postoperatively. The DLK was staged from 1 to 4. On day 7, the rabbits were killed and the eyes enucleated and processed for histopathologic examination. Results: At the end of the study, 54 eyes (46 exposed, 8 control) were available for evaluation. The 8 eyes studied concurrently in the control group remained clear and did not show interface inflammation. Thirty-one of 46 eyes (67%) treated with the various test substances developed DLK. The highest DLK rates were found with the cleaning soap Palmolive Ultra (100%; P ⫽ .022) and P aeruginosa lipopolysaccharide endotoxin (90%; P ⫽ .026). Conclusions: Interface inflammation was consistently induced in the animal model. All 7 agents caused DLK in at least some eyes. The histopathologic evaluation showed the morphologic profile of the marked inflammatory cellular reaction that occurred in almost all the specimens. J Cataract Refract Surg 2003; 29:542–549 © 2003 ASCRS and ESCRS
L
aser in situ keratomileusis (LASIK) is the most popular refractive surgery procedure for the correction of myopia and hyperopia with or without astigmatism.1–3 Although the surgical techniques and instruments for this procedure are extremely refined, postoperative complications may occur.4 Diffuse lamellar keratitis (DLK), a nonspecific inflammatory reaction Accepted for publication August 6, 2002. Reprint requests to Kerry D. Solomon, MD, Magill Research Center for Vision Correction, Medical University of South Carolina, Storm Eye Institute, 167 Ashley Avenue, Charleston, South Carolina 29425, USA. © 2003 ASCRS and ESCRS Published by Elsevier Science Inc.
at the interface level following LASIK, was first described in 1998 by Smith and Maloney.5 Clinically, it presents as a white granular infiltrate, usually appearing 1 to 5 days after surgery. It can, however, also occur in specific situations long after the procedure; for example, following epithelial defects or trauma.6 – 8 Maddox and Hatsis presented the same findings and based on the infiltrate’s characteristic wavy appearance, they termed it “Sands of the Sahara Syndrome” (A. Hatsis, MD, R. Maddox, MD, “Sands of the Sahara,” poster presented at the Symposium on Cataract, IOL and Refractive Surgery, San Diego, California, USA, April 1998). 0886-3350/03/$–see front matter doi:10.1016/S0886-3350(02)01691-7
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
The exact etiology of DLK is unknown, although multiple factors have been implicated in the development of the condition. Several factors have been reported to be associated with DLK, including epithelial defects, trauma, meibomian gland secretions, and bacterial endotoxins.6 –16 This study used an animal model to induce DLK and investigate possible causative factors related to its development.
Materials and Methods Seventy eyes of 35 female Dutch Belted rabbits weighing 1.5 to 2.0 kg were used. The rabbits were treated according to the guidelines of the Association for Research in Vision and Ophthalmology and the Division of Laboratory Animal Resources at the Medical University of South Carolina. They were randomly assigned to have their corneal interface exposed to 1 of 7 substances. A separate group of rabbits without substance exposure was used as the control group. Table 1 shows the groups and the substances. The rabbits were anesthetized with an intramuscular injection of a 1:1 mixture of ketamine hydrochloride (Ketaset威) and xylazine (Rompum威). A corneal flap was cut with a nasal or temporal hinge position using the Automated Corneal Shaper威 microkeratome (Bausch & Lomb Surgical) and an 8.5 mm suction ring with a 160 m plate. A new blade was used in each rabbit. The eyes were not proptosed, and posterior pressure was not necessary to cut the flaps. The flap was
Table 1. Groups and substances used in the study.
Group
Number of Eyes
1
11
2
9
Highly purified Staphylococcus aureus exfoliative toxin A (ETA) (Toxin Technology Inc.)
3
11
Highly purified Staphylococcus aureus enterotoxin B (SEB) (Toxin Technology Inc.)
4
9
Palmolive姞 Ultra soap (Colgate-Palmolive Co.)
5
6
Klenzyme姞 enzymatic cleaner (Steris Co.)
6
6
Povidone–iodine 10% (Betadine姞, Professional Disposable International)
7
8
Meibomian gland secretion
8
10
Substance Pseudomonas aeruginosa lipopolysaccharide (LPS) endotoxin serotype 10 (Sigma-Aldrich)
Control
lifted with a Lindstrom spatula (American Surgical Instruments Corp.), and the agents were exposed to the stromal bed using a Merocel威 sponge (Medtronic Solan), which was pressed down slightly on the stroma once. This was done assuming that the amount of liquid absorbed by each sponge and later released to the interface was the same in all cases. Measurements in our laboratory show that if a Merocel sponge is pressed on a straight surface, approximately 40 L of fluid are released. The concentrations of the agents were as follows: LPS endotoxin 2.5 mg/mL; both Staphylococcus aureus toxins 0.075 mg/mL; povidone⫺iodine 10%, Klenzyme, and Palmolive Ultra were diluted 1:50. Concentrations of the specific bacterial toxins were determined following reports of rabbit studies.17,18 Meibomian gland secretions were obtained by pushing the upper lid of a rabbit eye toward a Lindstrom spatula. Secretions were exposed to the central part of the stromal bed, followed by repositioning of the flap and 1-minute drying time. In the control group, the flap was repositioned without exposing the stromal bed to a substance. All surgeries were completed without laser ablation. Postoperatively, the rabbits were examined by the same unmasked examiner on days 1, 3, 5, and 7 with a slitlamp microscope (Topcon SL-7E). The corneal findings were documented and slitlamp photographs taken. The DLK findings in each eye were staged 1 to 4 following the system of Linebarger and coauthors19: stage 1, white granular infiltrates in the periphery of the lamellar flap, outside the visual axis; stage 2, white infiltrates in the center of the flap, involving the visual axis; stage 3, more dense, white, clumped infiltrates in the central visual axis; and stage 4, stromal melting, permanent scarring, and visual morbidity. The rabbits received no topical or systemic medication at any time postoperatively. On day 7, the rabbits were anesthetized with a 1:1 mixture of ketamine hydrochloride and xylazine and then killed with an intravenous injection of 2 mL pentobarbital (Sleepaway威). The eyes were immediately enucleated, and the entire corneas were excised and fixed in neutral formalin 10% for 24 hours. The corneas were prepared for serial paraffin sections and evaluated with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), and Masson trichrome stains. Corneal tissues from 46 animal eyes treated to induce opacities resembling those seen clinically in DLK were examined histopathologically by light microscopy. In the histopathologic examination of these eyes, emphasis was placed on evaluation of the possible infiltration of inflammatory cells within anatomic structures of the cornea, especially within the stroma and near the interface. The degree of wound-healing activity was also evaluated, and the results were compared with the findings in the 8 control eyes. Statistical analysis was performed using the Fisher exact test. A P value less than 0.05 was considered statistically significant.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
543
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Results
Nine of 10 eyes (90%) exposed to Pseudomonas aeruginosa LPS endotoxin developed DLK. Compared to the control group, this was statistically significant (P ⫽ .026). The DLK rates ranged between 29% and 60% with the other agents (S aureus toxins ETA and SEB, meibomian gland secretion, and povidone⫺iodine); however, there were no statistically significant differences. The clinical grading (Table 3) showed that P aeruginosa LPS endotoxin caused the most severe grades of DLK; 7 of 9 DLK eyes were stage 2 or 3. In the Palmolive Ultra group, 1 of 7 eyes was stage 3. All other eyes were stage 1. The DLK grade was highest in the LPS group (1.8) followed by the Klenzyme (1.5) and Palmolive Ultra (1.28) groups. Even though the groups were too small to determine a statistically significant difference, the trends shown in Table 3 correlate well with the findings noted in Table 2.
At the end of the study, 54 eyes were included for analysis. Sixteen eyes were excluded because of a slipped (n ⫽ 12) or free (n ⫽ 3) flap; in 1 eye, it was not possible to cut a flap because of insufficient suction by the microkeratome apparatus. The group sizes after exclusion of these 16 eyes are shown in Table 2. Clinical Findings Throughout the 7-day follow-up, the 8 eyes in the control group remained clear and did not show interface inflammation. Thirty-one of the 46 eyes (67%) exposed to 1 of the substances developed DLK. This was statistically significant (P ⫽ .024) compared to the control group (0%) (Table 2). The appearance of DLK varied from slightly granular with a faint haze to clumpy white granules in the entire interface. In most cases, DLK first appeared on day 3. To evaluate the possibility of infection in the eyes, scraping samples for chocolate agar culture from 3 suspicious corneal interfaces were taken. None of these showed growth of microorganisms. The highest rate of DLK was in the soap groups, with all eyes developing DLK. The findings in the Palmolive Ultra group were statistically significant (P ⫽ .022) compared to those in the control group. The Klenzyme group was too small to determine a statistically significant difference.
Histopathologic Findings In the control corneas (Figure 1), the staining (H&E and Masson trichrome) revealed a typical and relatively homogeneous normal pattern. A hyperplastic epithelial plug (facette) at the edge of the incision with thickening of up to 10 epithelial cell layers marked the beginning of the incision, and the number of keratocytes was slightly increased in this area. Despite newly formed, irregularly arranged collagen fibers in the microkeratome wound site, the stroma generally main-
Table 2. Number of inoculated and control eyes and postoperative DLK rates. Eyes with DLK
Agent
Total Number of Eyes
Eyes Included
Number
Control
10
8
0
%
P Value (Versus Control)
Palmolive Ultra
9
7
7
100.0
.022*
Klenzyme
6
4
4
100.0
.077
11
10
9
90.0
.026*
8
5
3
60.0
.2
LPS Meibomian gland secretion Povidone–iodine
6
5
3
60.0
.2
SEB
11
8
3
37.5
.228
ETA
9
7
2
28.6
.471
Total
70
54
31
—
—
Total without controls
60
46
31
67.4
.024*
SEB ⫽ Staphylococcus aureus enterotoxin B, ETA ⫽ Staphylococcus aureus exfoliative toxin A, LPS ⫽ Pseudomonas aeruginosa LPS endotoxin *Statistically significant
544
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
tained its lamellar structure unaltered. The number of keratocytes increased slightly in the wound margin but not in the wound center. Although a distinct tendency was noted, the histopathologic morphology in each case of DLK did not always correlate with the clinical appearance. For example, the clinical reaction to Klenzyme was severe, but the histopathologic appearance was moderate. Compared to the reaction in the control eyes, it had an increased number of keratocytes and inflammatory cells in the anterior stroma. The histopathologic examination of the cornea exposed to P aeruginosa LPS endotoxin revealed a marked cellular reaction with inflammatory cells, keratocytes, and collagen formation. It was determined marked and correlated well with the clinical appearance. Figures 2 and 3 show an eye exposed to S aureus exotoxin ETA and a clinical DLK-like inflammatory reaction. The histopathological specimen revealed a marked inflammatory cellular reaction with plasma cells including Russel bodies, eosinophil granulocytes, and lymphocytes in the corneal stroma around the interface.
Discussion Refractive surgery is a rapidly growing field, and LASIK has become the predominant procedure for the correction of myopia and hyperopia with or without astigmatism.1 However, as the number of LASIK proceTable 3. Diffuse lamellar keratitis score based on slitlamp examination (highest score for each agent during the 7-day postoperative period).
Agent
Mean DLK Score ⴞ SD
Standard Error
LPS
1.80 ⫾ 0.91
0.29
Klenzyme
1.50 ⫾ 0.58
0.29
Palmolive Ultra
1.28 ⫾ 0.75
0.28
Povidone-iodine
1.00 ⫾ 1.00
0.44
SEB
0.87 ⫾ 1.35
0.48
ETA
0.71 ⫾ 1.25
0.47
Meibomian gland secretion
0.60 ⫾ 0.54
0.24
Differences in the mean values among the treatment groups were not significant (P ⫽ .301, 1-way repeated-measures analysis of variance). LPS ⫽ Pseudomonas aeruginosa lipopolysaccharide; SEB ⫽ Staphylococcus aureus enterotoxin B; ETA ⫽ Staphylococcus aureus exfoliative toxin A DLK scores: stage 1 ⫽ 1; stage 2 ⫽ 2, etc.
dures has increased, the number of complications has also risen and more surgeons are encountering DLK. In their 1998 description of DLK, Smith and Maloney5 reported 13 eyes of 12 patients with noninfectious diffuse infiltrates in the lamellar interface between 2 and 6 days after LASIK. The clinical evolution of DLK is characterized by the appearance of white granular infiltrates in the stromal interface 1 to 5 days after LASIK. The infiltrate normally starts at the periphery of the corneal flap and has to be distinguished from epithelial cells, debris, and meibomian secretions.20 Clinically, DLK can be graded in 4 stages; this study followed the staging system described by Linebarger and coauthors.19 To date, the exact etiology as well as the specific prophylactic treatment regimens for DLK are unknown. There are reports linking epithelial defects, trauma, meibomian gland secretions, and bacterial endotoxins to DLK.6 –11,13–16 To our knowledge, this is the first extensive experimental study testing several possible causative agents and the first to look at and report the histopathology of corneas with DLK. Although DLK development in the rabbit model may not completely equate to that in the human eye, an experimental setting to investigate this disease in more detail is necessary. The results of the study showed that all tested agents could cause an inflammatory reaction at the interface level. This may confirm the theory that DLK has multiple etiologies. Some of the agents resulted in very high DLK rates: Palmolive Ultra (100%), Klenzyme (100%), and P aeruginosa LPS endotoxin (90%). In addition to the DLK rates, the intensity of the DLK response also varied with P aeruginosa LPS endotoxin and the soaps (Palmolive Ultra and Klenzyme), with the most intense inflammatory response following inoculation with LPS endotoxin. The finding that the soaps caused DLK in all eyes shows the importance of thoroughly rinsing surgical instruments, especially the microkeratomes. Soap residue left on instruments may gain access to the corneal stroma and cause DLK. Many microkeratome manufacturers recommend Palmolive Ultra to clean their instruments due to its lubricating effect, and the soap has been used by many surgeons. Recently, some manufacturers changed the cleaning recommendations for their microkeratomes from Palmolive Ultra to specific medical-device cleaners similar to the enzymatic cleaner Klenzyme.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
545
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Figure 1. (Holzer) Cornea of a control eye 7 days postoperatively. The incision is recognizable as a small epithelial plug (facette). The corneal flap (toward the right) covers the anterior one third of the stroma. The epithelium reveals a normal pattern except for the facette at the incision site, which consists of a thickening of up to 8 epithelial cell layers. A slightly increased number of keratocytes (arrows) are seen around the incision area. EP ⫽ epithelial plug, F ⫽ flap, I ⫽ interface, SB ⫽ stromal bed. a: H&E stain, original magnification ⫻200. b: Masson trichrome stain, original magnification ⫻200.
Figure 2. (Holzer) Cornea with a marked inflammatory reaction 7 days postoperatively. Clinically, the eye showed a DLK-like inflammatory reaction, which was induced by S aureus exotoxin ETA. The interface is located in the anterior one fourth of the cornea. Compared to the control eye (Figure 1), the photomicrographs show a marked cellular reaction with inflammatory cells and keratocytes. The inflammatory reaction consists of plasma cells including Russel bodies, scattered eosinophil granulocytes, and lymphocytes in the flap periphery. a: Masson trichrome stain, original magnification ⫻100. b: Masson trichrome stain, original magnification ⫻200.
Both cleaning agents were tested in this study, and both caused DLK in every eye. This was statistically significant when the Palmolive Ultra group was compared to the control group. However, because of the small number of eyes (n ⫽ 4), there was no statistically significant difference detectable for the Klenzyme group. Further investigation to identify the ideal cleaner and the optimal cleaning regimen for microkeratomes and other sur546
gical instruments is necessary to find a standardized cleaning protocol that is highly efficient and also practical for use during a routine surgical day. The results of the bacterial toxins confirm that these agents have the potential to cause DLK. Holland et al.15 were the first to report DLK related to endotoxins. They found a biofilm from the gram-negative bacteria Burkholderia pickettii in the reservoir of their sterilizer and
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
Figure 3. (Holzer) Cornea with a marked inflammatory reaction 7 days postoperatively. Clinically, the eye showed a DLK-like inflammatory reaction, which was induced by S aureus exotoxin ETA. Compared to the control eye (Figure 1), the photomicrographs show a marked cellular reaction with inflammatory cells and keratocytes. The inflammatory reaction consists of plasma cells and scattered eosinophil granulocytes. a: PAS stain, original magnification ⫻400. b: H&E stain, original magnification ⫻400.
postulated that bacteria are released into the distilled water of the sterilizer reservoir. These bacteria would be killed during the sterilization cycle and endotoxins, which are a component of the outer membrane of gram-negative bacteria, would be released after the cells lysed. The endotoxins may persist through the sterilization process, adhere to the instruments, and be deposited in the interface during surgery. In high doses, endotoxins are toxic and can cause inflammation, fever, anaphylactic shock, blood coagulation, and sepsis.21 If present in the corneal interface, endotoxins may cause an inflammatory reaction, which was seen in the histopathology sections of the corneas with clinical DLK. In a small animal study by Peters et al.,16 clinical DLK developed in 4 rabbit eyes within 3 days of inoculation with Pseudomonas LPS endotoxin. Peters et al. also examined sources of endotoxin contamination in eye centers and detected them in tap water, filtered and distilled water, instrument washbasins, and sterilizer reservoirs. In our study, P aeruginosa LPS endotoxin caused DLK in 9 of 10 eyes, most of which had more severe stages than the other eyes that developed DLK. Therefore, surgeons may want to consider microbiologic examination of their sterilizers and the use of special cleaning procedures to lower endotoxin concentrations and minimize DLK occurrences.
The S aureus exotoxins SEB and ETA also caused DLK but in a relatively smaller proportion of eyes (38% and 29%, respectively). This lower DLK rate might be related to the concentration of the toxins used. Exotoxins are generally an extremely potent stimulus for T-lymphocytes and inflammatory reactions.22 In the histopathological sections of the rabbit corneas, inflammatory cells, including lymphocytes and plasma cells, were found in the corneas that were clinically diagnosed with DLK. As S aureus can be found in the normal human eyelid flora,23 surgeons should pay careful attention to proper lid hygiene before and during surgery to avoid contamination of the stromal bed. In this study, meibomian gland secretions also caused DLK. However, because of the way the gland secretion was obtained in the rabbit eyes, bacteria and/or bacterial endotoxins/exotoxins may also have been exposed to the interface. There is 1 report of the possible correlation between meibomian secretion and DLK.14 The authors did not, however, find a correlation between their patients with DLK and the presence of significant meibomianitis or blepharitis. In a survey of refractive surgeons, lid margin disease was the factor most commonly thought to be related to DLK.24 Therefore, in our opinion, lid hygiene, including scrubbing and draping the lid margins, may be important to min-
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
547
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
imize DLK after LASIK, and patients with blepharitis, which is associated with Staphylococcus epidermidis and Propionibacterium acnes,23 should not be scheduled for LASIK until effectively treated. Povidone⫺iodine 10%, commonly used in a diluted solution prior to LASIK to scrub eyelids and lashes, also caused DLK in the rabbit eyes. In their initial report describing DLK, Smith and Maloney5 hypothesize that povidone⫺iodine may be a causative factor. In an animal study by Alp et al.,25 the toxicity of povidone⫺iodine 5% and 10% to the corneal endothelium was shown. Similar toxic effects on the corneal stroma may be a reason for the development of DLK. Although povidone⫺iodine is important for scrubbing the lids and for infection prophylaxis, surgeons should be careful to avoid contact of povidone⫺iodine with the corneal stroma, which may cause DLK. The animal model used in this study and the causative agents with the highest DLK rate (P aeruginosa LPS endotoxin and cleaning soaps) may help to establish an investigational setting for DLK. A critical point of this study, however, might be that it was not observer masked, which will be addressed in future studies. With the help of standardized studies, further research on DLK as well as improved prevention and treatment should be possible. Although the specific pathophysiologic reactions of DLK in humans are still unknown, the findings of this study may confirm that it is a multifactorial disease. There is some evidence that inflammatory reactions are correlated with DLK. The histopathologic examination of the eyes exposed to endotoxins and exotoxins showed severe responses with an inflammatory cell reaction, eyes that were previously clinically diagnosed with DLK. The histopathologic examination of the eyes did not always correlate with the clinical findings. This might be because the pathologic sections were done at the end of the study on day 7 and the peak of the DLK reaction had passed. Foreign material that gains access to the interface during surgery may be the stimulus for an inflammatory response that may be seen clinically as DLK. Therefore, it is important to avoid the inoculation of harmful agents into the stromal bed. Additionally, microbiological investigation of sterilizers and water sources used in LASIK facilities to detect possible high concentrations of endotoxins or exotoxins may be considered to minimize the incidence of DLK. 548
This study used an animal model for DLK and investigated possible causative agents. Additional studies are needed to learn more about the detailed pathophysiologic process of DLK and to better understand the treatment and prevention of this disease after LASIK.
References 1. Pallikaris IG, Siganos DS. Laser in situ keratomileusis to treat myopia: early experience. J Cataract Refract Surg 1997; 23:39 –49 2. Salchow DJ, Zirm ME, Stieldorf C, Parisi A. Laser-insitu keratomileusis (LASIK) zur Myopie- und Astigmatismuskorrektur. Ophthalmologe 1998; 95:142–147 3. Knorz MC, Jendritza B, Liermann A, et al. LASIK zur Myopiakorrektur; 2-Jahres-Ergebnisse. Ophthalmologe 1998; 95:494 –498 4. Ambro´ sio R Jr, Wilson SE. Complications of laser in situ keratomileusis: etiology, prevention, and treatment. J Refract Surg 2001; 17:350 –379 5. Smith RJ, Maloney RK. Diffuse lamellar keratitis; a new syndrome in lamellar refractive surgery. Ophthalmology 1998; 105:1721–1726 6. Haw WW, Manche EE. Late onset diffuse lamellar keratitis associated with an epithelial defect in six eyes. J Refract Surg 2000; 16:744 –748 7. Harrison DA, Periman LM. Diffuse lamellar keratitis associated with recurrent corneal erosions after laser in situ keratomileusis. J Refract Surg 2001; 17:463–465 8. Weisenthal RW. Diffuse lamellar keratitis induced by trauma 6 months after laser in situ keratomileusis. J Refract Surg 2000; 16:749 –751 9. Johnson JD, Harissi-Dagher M, Pineda R, et al. Diffuse lamellar keratitis: incidence, associations, outcomes, and a new classification system. J Cataract Refract Surg 2001; 27:1560 –1566 10. Shah MN, Misra M, Wilhelmus KR, Koch DD. Diffuse lamellar keratitis associated with epithelial defects after laser in situ keratomileusis. J Cataract Refract Surg 2000; 26:1312–1318 11. Steinert RF, McColgin AZ, White A, Horsburgh GM. Diffuse interface keratitis after laser in situ keratomileusis (LASIK): a nonspecific syndrome. Am J Ophthalmol 2000; 129:380 –381 12. Yavitz EQ. Diffuse lamellar keratitis caused by mechanical disruption of epithelium 60 days after LASIK [letter]. J Refract Surg 2001; 17:621 13. Schwartz GS, Park DH, Schloff S, Lane SS. Traumatic flap displacement and subsequent diffuse lamellar keratitis after laser in situ keratomileusis. J Cataract Refract Surg 2001; 27:781–783 14. Fogla R, Rao SK, Padmanabhan P. Diffuse lamellar ker-
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
LABORATORY SCIENCE: DLK IN AN ANIMAL MODEL
15.
16.
17.
18.
19.
20.
atitis: are meibomian secretions responsible? [letter]. J Cataract Refract Surg 2001; 27:493–495 Holland SP, Mathias RG, Morck DW, et al. Diffuse lamellar keratitis related to endotoxins released from sterilizer reservoir biofilms. Ophthalmology 2000; 107: 1227–1233; discussion by EJ Holland, 1233–1234 Peters NT, Iskander NG, Penno EEA, et al. Diffuse lamellar keratitis: isolation of endotoxin and demonstration of the inflammatory potential in a rabbit laser in situ keratomileusis model. J Cataract Refract Surg 2001; 27: 917–923 Siqueira JA, Speeg-Schatz C, Freitas FI, et al. Channelforming leucotoxins from Staphylococcus aureus cause severe inflammatory reactions in a rabbit eye model. J Med Microbiol 1997; 46:486 –494 Schultz CL, Morck DW, McKay SG, et al. Lipopolysaccharide induced acute red eye and corneal ulcers. Exp Eye Res 1997; 64:3–9 Linebarger EJ, Hardten DR, Lindstrom RL. Diffuse lamellar keratitis: diagnosis and management. J Cataract Refract Surg 2000; 26:1072–1077 Bu¨ hren J, Baumeister M, Kohnen T. Diffuse lamellar keratitis after laser in situ keratomileusis imaged by confocal microscopy. Ophthalmology 2001; 108: 1075–1081
21. O’Reilly M, Newcomb DE, Remick D. Endotoxin, sepsis, and the primrose path. Shock 1999; 12:411–420 22. Marrack P, Kappler J. The staphylococcal enterotoxins and their relatives. Science 1990; 248:705–711; erratum, 1066 23. Groden LR, Murphy B, Rodnite J, Genvert GI. Lid flora in blepharitis. Cornea 1991; 10:50 –53 24. Solomon KD, Holzer MP, Sandoval HP, et al. Refractive surgery survey 2001. J Cataract Refract Surg 2002; 28: 346 –355 25. Alp BN, Elibol O, Sargon MF, et al. The effect of povidone iodine on the corneal endothelium. Cornea 2000; 19:546 –550 From the Magill Research Center for Vision Correction, Storm Eye Institute, Medical University of South Carolina, Charleston, South Carolina, USA. Presented in part at the Symposium on Cataract, IOL and Refractive Surgery, San Diego, California, USA, April 2001, and as a poster at the annual meeting of the American Academy of Ophthalmology, New Orleans, Louisiana, USA, November 2001. Supported in part by an unrestricted grant from Research to Prevent Blindness, New York, New York, USA. None of the authors has a financial interest in any product mentioned.
J CATARACT REFRACT SURG—VOL 29, MARCH 2003
549