The Safety of Intraocular Ketorolac in Rabbits

Abstract

purpose. To assess the safety of a possible substitute treatment for intraocular steroid injections, intraocular injections of ketorolac tromethamine, one of the nonsteroidal anti-inflammatory drugs, were performed in rabbits.

methods. Either 0.5% or 0.25% preservative-free ketorolac tromethamine ophthalmic solution (0.1 mL) was injected into the vitreous of the right eye of 15 rabbits. Physiologic saline solution (BSS; Alcon, Ft. Worth, TX) was injected into the left eye of each rabbit as a control. A standard electroretinogram and intraocular pressure measurements were obtained before injection, and repeated 1 day and 1, 2, 3, and 4 weeks after injection. After 4 weeks, the rabbits were euthanatized and the retinas examined by light and electron microscopy. Differences in the electroretinograms, intraocular pressure, and histopathology between the two eyes were recorded. Further, the elimination half-life of the drug in the vitreous was assessed.

results. There were no statistically significant differences in electroretinograms, or intraocular pressure measurements obtained between the ketorolac-injected eyes and the control eyes. The half life of the drug was measured to be 2.3 hours. No histopathologic changes were observed in study eyes compared with control eyes.

conclusions. Preservative-free ketorolac tromethamine is nontoxic to the retinas of rabbits when injected intravitreally and could be considered as an alternative to intraocular steroid injections.

Corticosteroids have been a standard treatment for various inflammatory ocular conditions for more than 50 years. ¹ They have traditionally been delivered topically or by periocular and orbital injections. Recently clinicians have been injecting triamcinolone directly into the vitreous with the hope of increasing local concentration and duration of therapeutic activity. Such injections are currently being performed to treat diseases such as cystoid macular edema, diabetic macular edema, uveitis, and even age-related macular degeneration. ² Results are promising; however, the technique has not been well studied. Reported side effects of intravitreal steroids include cataract development, increased intraocular pressure, endophthalmitis, retinal detachment, vitreous hemorrhage, and pseudohypopyon. ^{1 2 3 4 5} The incidence of increased intraocular pressure after intravitreal triamcinolone has been reported to be approximately 40%. ²  

Another alternative to treat such diseases may be intraocular nonsteroidal anti-inflammatory drugs (NSAIDs). These drugs have a history of successfully treating inflammatory systemic diseases, and were the only medical treatment available for uveitis before the availability of steroids. Topical ocular NSAIDs have been available commercially in the United States since the early 1990s. Ocular NSAIDs are currently indicated for use in prevention of intraoperative miosis, management of postoperative cystoid macular edema, treatment of allergic conjunctivitis and keratitis, treatment of postprocedure pain, and control of inflammation in patients known to have increased intraocular pressure when treated with steroids. ⁶ Adverse effects of topical NSAIDs include burning, stinging, conjunctival hyperemia, and delayed corneal epithelial healing. There are also reports of corneal thinning and ulceration in the immediate postoperative period. ⁶ Two NSAIDs have been reported to have a dose-related toxic effect on the retina, vitreous, and lens. ^{7 8} Generally, NSAIDs are known to be less cataractogenic, and they do not cause intraocular pressure elevation. Thus, we devised an experiment to determine whether preservative-free ketorolac tromethamine, an NSAID, is nontoxic to the retina when injected into the vitreous of rabbits. 

Methods

All animals (pigmented Dutch-belted rabbits) were used according to the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research and the University of Nebraska Medical Center guidelines for use of animals in experimental procedures. All surgical and examination procedures were performed with rabbits under anesthesia induced with an intramuscular injection of 35 mg/kg of body weight of ketamine hydrochloride and 5 mg/kg of body weight of xylazine hydrochloride (both from Phoenix Scientific, Inc., St. Joseph, MO). Before all intravitreal injections, the eyes were cleaned with few drops of 5% povidone iodine solution. 

To calculate the half-life of the drug in the vitreous, we injected 0.1 mL of 0.5% ketorolac tromethamine (500 μg dose) solution (Acular PF; Allergan, Inc., Irvine, CA) into both eyes of four rabbits. The rabbits were euthanatized at 1, 2, 6, and 24 hours after injection. Ketorolac was analyzed in the vitreous humor of both eyes with reversed-phase HPLC, used according to a method similar to that reported in plasma. ⁹ To each vitreous sample (100 μL), 0.8 mL of phosphate-buffered saline (pH 7.4) fortified with 2.8 μg of tolmetin (internal standard) was added. Ketorolac and tolmetin are acidic drugs with calculated pKa values of 4.47 and 4.22, respectively (SciFinder Scholar, American Chemical Society; http//:www.cas.org). Samples preparations were acidified by adding 100 μL of 0.5 M sodium acetate (pH = 4.0) solution and extracted with 2 mL diethyl ether by vortex agitation for 1 minute. The aqueous and organic phases were separated by centrifuging at 5000 rpm for 10 minutes (RT 6000B Refrigerated Centrifuge; Sorvall, Newtown, CT). The organic phase was transferred to a fresh test tube and evaporated to dryness (N-evap system; Organomation Associates, Inc., Berlin, MA). The dry residue was redissolved in 100 μL of deionized water, and 50 μL was injected onto an HPLC system (Waters, Milford, MA) that included a pump (model TM 616; Waters), a controller (model 600 S; Waters), an autoinjector (model 717 plus; Waters), and a PDA detector (model 996; Waters). The peak areas were integrated on a computer (Millennium software, ver. 2.15.01; Millennium, Torrance CA). The drugs were separated with a 12.5-cm long C-18 column (Nucleosil 100-5; Machery-Nagel, Düren, Germany) with a particle diameter of 5 μm and a pore size of 10.0 nm. The mobile phase for the assay consisted of acetonitrile and deionized water with 1% trifluoroacetic acid (50:50 vol/vol). Ketorolac and internal standard were monitored at 312 nm. The mobile phase was run for 10 minutes at 1 mL/min flow rate, and the retention times were 4 and 4.9 minutes, respectively, for ketorolac and tolmetin. 

To examine drug toxicity in vivo, preservative-free ketorolac tromethamine was injected into 15 rabbits’ right eyes. Two different concentrations of ketorolac, 0.5% and 0.25%, were injected. A total of 0.1 mL of preservative-free ketorolac was injected into the vitreous cavity of the experimental eyes and 0.1 mL of physiologic saline was injected into the vitreous cavity of the left eyes of the same animal. The vitreous cavity was entered through the superotemporal sclera (less than 1 mm posterior to the limbus) using a 27-gauge needle connected to a 1-mL syringe that contained 0.1 mL of different concentrations of ketorolac (experimental eyes) or 0.1 mL of saline solution (control eyes). A few seconds later, the eye was rinsed with povidone-iodine again, and 0.1 mL of aqueous humor was withdrawn from the anterior chamber via a paracentesis performed by a 12-o’clock limbal insertion of a 30-gauge needle connected to a 1-mL syringe. 

The eyes were evaluated by anterior and posterior biomicroscopy, indirect ophthalmoscopy, and electroretinography (ERG) before the injection, and at 1 day and 1, 2, 3, and 4 weeks after injection. 

To perform the ERG, the pupils were dilated with 1 drop of tropicamide 1%, and phenylephrine hydrochloride 2.5% (both from Bausch & Lomb Pharmaceuticals, Inc., Tampa, FL). Animals were dark adapted for 30 minutes, and standard ERGs were recorded in both eyes. The ERG setup consisted of a contact lens electrode for each eye, a reference needle electrode positioned at the lateral canthus, and a ground disc electrode that was placed in the mouth of the animals. ERGs were recorded (UTAS-E200 system; LKC Technologies, Gaithersburg, MD). An average of three separate ERGs was used at each time point for each eye. 

Statistical analysis of the data was then performed. To account for the repeated measurements within subjects, the method of generalized estimating equations (GEEs) was used. This approach characterizes the average response for eyes measured at the same time point as a function of treatment and provides robust estimates of the standard errors of the model parameters. This methodology also allows for the correlation induced by taking repeated measurements among both eyes in the same animal. 

In another experiment, intraocular pressure (IOP) measurements were performed for 0.5% ketorolac (n = 3), 0.25% ketorolac (n = 3), and noninjected eyes (n = 6) with a pneumotonometer (model 30 Classic; Mentor, Norwell, MA). The IOP was measured before the injection, at 1 day and 1, 2, 3, and 4 weeks after injection. 

After 4 weeks, the rabbits were euthanatized with an overdose injection of pentobarbital. The eyes were immediately enucleated and fixed in a solution of formalin 10% for light microscopy, or in a glutamate-formalin mixture (paraformaldehyde 2%, glutaraldehyde 2%, and phosphate buffer 0.1 M; pH 7.4) for electron microscopy examination. 

Thickness measurements of different retinal layers (outer segments, outer nuclear layer, inner nuclear layer, and the ganglion cell layer) were performed on all specimens available for the 0.5% ketorolac, 0.25% ketorolac, and saline-injected eyes by using light microscopy. 

Results

After intravitreous injection of a 500-μg dose of ketorolac, rabbit vitreous concentrations of the drug declined from 122 μg/mL at 1 hour to under 1 μg/mL at 24 hours (Fig. 1) . The vitreous concentration versus time plot could be explained by the monoexponential expression shown in the figure, where C _vit is the vitreous drug concentration in micrograms per milliliter and t is time in hours. Thus, the drug concentrations followed first-order elimination kinetics, with an elimination rate constant of 0.298 hour⁻¹ and a half-life of 2.3 hours. The “0” time concentration for the 500-μg dose was estimated to be 177.6 μg/mL, suggesting a volume of distribution of 2.8 mL, which is consistent with the vitreous volume in a rabbit eye. 

The GEE regression model described herein was used to assess the ERG results. There was no evidence that the b-wave amplitude produced after white light stimulation (0 dB) differed significantly for eyes treated with 0.25% or 0.5% ketorolac compared with saline-treated eyes (P = 0.55 for 0.25% ketorolac, and P = 0.59 for 0.5% ketorolac). Similarly, the response amplitudes produced after photopic white 30-Hz flicker stimuli did not differ significantly between eyes treated with 0.25% or 0.5% ketorolac compared with saline-treated eyes (P = 0.24 for 0.25% ketorolac, and P = 0.96 for 0.5% ketorolac). Descriptive statistics for each eye at each time point are provided in Tables 1 and 2 , and Figures 2 3 4 and 5 . Descriptive statistics for the difference in the response (right eye − left eye) are provided in Tables 3 and 4 . 

The mean IOPs for 0.5% and 0.25% ketorolac-treated and noninjected eyes are specified in Table 5 . There was a tendency for lower IOP for the ketorolac-injected eyes at 1 day and 1 week after injection. This tendency did not reach statistical significance. No significant IOP differences were recorded among the different study groups. 

Histopathology specimens were examined by light microscopy and electron microscopy. There were no apparent differences between the histology examinations of eyes injected with ketorolac and of eyes injected with saline. Figure 6shows the light microscopy image of an eye injected with 0.5% ketorolac, Figure 7shows a similar image of a control eye, and Figure 8shows an electron microscopy image of an eye injected with 0.5% ketorolac. Figures 9 and 10show similar images of eyes injected with 0.25% ketorolac and physiologic saline, respectively. 

Thickness measurements of the different retinal layers were performed. Mean thickness measurements for 0.5% and 0.25% ketorolac-treated eyes and saline-treated eyes are specified in Table 6 . No significant differences were found between the groups. 

Discussion

Corticosteroids have been an effective treatment for various inflammatory ocular conditions for more than 50 years. The recent trend of injecting triamcinolone directly into the vitreous holds promise for improving the treatment of diseases such as pseudophakic cystoid macular edema, diabetic macular edema, uveitis, and even age-related macular degeneration. However, the side effects of intraocular steroids are well known, and the risk-benefit ratio precludes their use in some patients. Therefore, alternatives are needed. 

We examined the safety profile of one such alternative, intraocular ketorolac. This drug is commercially available in a preservative-free formulation, and has been commonly used topically to treat ocular diseases such as pseudophakic cystoid macular edema for more than 10 years. ¹⁰ Topical ketorolac has fewer side effects than does topical corticosteroids and is often used as an alternative treatment in patients with contraindications to corticosteroids. To our knowledge, the safety of intraocular injection of ketorolac has never been reported. 

Other studies have been published on the intraocular use of other NSAIDs. ^{7 8} The authors found retinal and lenticular dose-related toxicity. Shen et al. ⁷ reported that concentrations of more than 400 μg diclofenac causes such toxicity. We used an injection of 500 μg ketorolac, which was not toxic to the retina or lens. It is possible that the toxic effects were caused by using a different drug, because no preservatives were included in either of the previous studies. 

Our study demonstrates the safety of intravitreal injections of ketorolac in rabbits, both functionally, and histologically. Major side effects of intravitreal steroid use such as high IOP and cataract formation were not observed in this study, though the duration of the follow-up was relatively short. 

The efficacy of NSAIDs is thought to be lower than that of steroids. Also, we have observed that ketorolac is eliminated from the vitreous humor of rabbits, with a half-life of approximately 2.3 hours. Despite the short half-life and a possible lower efficacy, prolonged drug effects would be feasible if the drug is retained in the target tissues. 

In conclusion, our data show that ketorolac tromethamine, an NSAID, is nontoxic to the retina when injected into the vitreous of rabbits. In the future, it might be studied as a possible treatment for retinal diseases, such as cystoid macular edema, diabetic macular edema, uveitis, and age-related macular degeneration. 

 Supported in part by an unrestricted grant from Research to Prevent Blindness.

 Submitted for publication June 16, 2005; revised October 24 and December 2, 2005; accepted March 15, 2006.

 Disclosure: E. Margalit, Allergan Inc. (F); L.J. Kugler, None; M.V. Brumm, None; J.L. Meza, None; U.B. Kompella, None; E.R. Escobar, None; G.R. Christensen, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Concentration of ketorolac tromethamine in rabbit vitreous humor after a single intravitreal injection of a 500-μg dose (0.5%). The data presented are a mean of the concentration in the two eyes.

Table 1.

 

Descriptive Statistics for Each Eye in Rabbits Treated with 0.5% Ketorolac in the Right Eye and Saline Solution in the Left Eye

	Preoperative			Day 1			Week 1			Week 2			Week 3			Week 4
		OD	OS	OD	OS	OD	OS	OD	OS	OD	OS	OD	OS
Scotopic
Median	131.10	135.98	123.17	102.44	93.60	115.24	127.44	101.83	96.34	108.54	110.37	99.39
Min	87.80	90.85	65.85	68.90	66.46	55.49	88.41	83.54	79.27	79.88	84.15	79.27
Max	256.10	256.71	149.39	131.10	157.93	151.83	143.29	138.41	128.05	139.02	137.80	153.05
Mean	138.42	144.51	114.94	101.83	101.30	107.77	117.96	106.01	103.22	105.58	110.54	103.83
SEM	18.23	17.95	10.22	7.36	12.48	11.37	8.06	6.26	7.42	7.11	6.78	9.62
Photopic
Median	35.00	35.50	24.50	17.90	28.05	26.10	30.50	29.00	22.20	24.60	35.90	35.90
Min	16.60	18.50	10.20	9.30	16.30	15.90	12.90	12.40	14.60	18.00	24.90	25.60
Max	73.40	55.40	35.90	31.70	50.50	45.10	42.90	46.30	39.00	40.70	53.70	56.60
Mean	39.63	37.20	24.24	18.69	29.05	27.93	30.60	29.04	24.70	25.77	37.26	37.16
SEM	5.91	4.59	3.23	2.77	3.65	3.37	3.76	4.80	3.63	2.75	3.76	4.39

 Scotopic: The stimulus that was used was a scotopic white flash of 0 dB. The b-wave amplitudes are presented for the scotopic stimulation. Photopic: White 30-Hz flicker was used as a stimulus. The response amplitudes are presented. All scotopic and photopic results are presented in microvolts.

Table 2.

 

Descriptive Statistics for Each Eye in Rabbits Treated with 0.25% Ketorolac in the Right Eye and Saline Solutions in the Left Eye

	Preoperative			Day 1			Week 1			Week 2			Week 3			Week 4
		OD	OS	OD	OS	OD	OS	OD	OS	OD	OS	OD	OS
Scotopic
Median	114.02	129.88	90.24	87.20	92.68	103.05	89.02	124.39	118.90	90.85	117.99	110.98
Min	78.66	88.41	78.66	76.22	54.27	87.80	40.85	40.85	89.02	67.07	79.27	95.73
Max	198.78	167.68	128.05	139.02	122.56	124.39	195.73	167.07	183.54	154.88	160.37	126.22
Mean	123.00	126.22	98.43	100.96	90.59	106.62	103.92	114.98	121.86	101.13	119.21	111.99
SEM	15.26	11.17	7.53	9.25	8.76	6.11	18.41	17.40	12.80	13.01	11.24	4.88
Photopic
Median	31.00	33.20	18.00	16.60	26.30	24.60	29.50	41.20	29.30	28.30	33.90	33.45
Min	23.90	22.20	6.60	7.60	8.00	15.40	17.80	6.10	17.10	20.20	25.60	15.10
Max	56.80	41.50	26.30	26.30	45.10	51.20	65.90	53.40	53.70	51.50	42.00	39.50
Mean	33.76	32.99	15.83	16.49	25.13	29.03	32.54	32.77	29.84	30.47	33.98	29.18
SEM	4.48	3.01	2.52	2.61	4.32	4.56	5.89	6.51	4.59	4.24	2.80	4.39

 Scotopic and photopic stimuli are as described in Table 1 .

Figure 2.

 

The median amplitude of the electroretinogram b-wave amplitude (in microvolts), as a function of time, after a 0-dB white-flash light stimulus was given to all rabbits injected with 0.5% ketorolac in the right eye (solid line) and saline solution in the left eye (dashed line).

Figure 3.

 

The median amplitude of the flicker response (in microvolts), as a function of time, after a 30-Hz photopic light stimulus was given to all rabbits injected with 0.5% ketorolac in the right eye (solid line) and saline solution in the left eye (dashed line).

Figure 4.

 

The median amplitude of the electroretinogram b-wave (in microvolts), as a function of time, after a 0-dB white flash light stimulus was given to all rabbits injected with 0.25% ketorolac in the right eye (solid line) and saline solution in the left eye (dashed line).

Figure 5.

 

The median amplitude of the flicker response (in microvolts), as a function of time, after a 30 Hz photopic light stimulus was given to all rabbits injected with 0.25% ketorolac in the right eye (solid line) and saline solution in the left eye (dashed line).

Table 3.

 

Descriptive Statistics for the Difference between the Eyes in Rabbits Treated with 0.5% Ketorolac in the Right Eye and Saline Solution in the Left Eye

	Preoperative Difference	Day 1 Difference	Week 1 Difference	Week 2 Difference	Week 3 Difference	Week 4 Difference
Scotopic white φdB b-wave amplitude
Median	−4.57	14.03	−2.13	17.19	−7.32	12.20
Min	−107.93	−17.68	−85.37	−18.30	−24.39	−39.64
Max	98.78	42.68	43.30	42.68	37.81	27.44
Mean	−6.10	13.11	−6.48	11.95	−2.35	6.71
SEM	19.98	7.68	13.56	8.16	8.72	8.16
Photopic white 30-Hz amplitude
Median	−1.10	3.80	4.20	1.50	−3.40	2.00
Min	−19.80	−1.20	−24.60	−11.20	−21.40	−29.30
Max	37.50	26.10	14.40	13.50	13.90	17.80
Mean	2.43	5.55	1.13	1.56	−1.07	0.10
SEM	6.33	3.02	4.38	3.68	4.62	5.65

 Difference is OD − OS. All scotopic and photopic results are presented in microvolts.

Table 4.

 

Descriptive Statistics for the Difference between the Eyes in Rabbits Treated with 0.25% Ketorolac in the Right Eye and Saline Solutions in the Left Eye

	Preoperative Difference	Day 1 Difference	Week 1 Difference	Week 2 Difference	Week 3 Difference	Week 4 Difference
Scotopic white φdB b-wave amplitude
Median	3.66	1.22	−19.51	−13.11	21.95	8.54
Min	−78.05	−34.75	−37.80	−51.83	18.29	−26.83
Max	40.85	11.58	28.66	28.66	28.66	44.52
Mean	−3.22	−2.53	−16.03	−12.71	23.48	7.22
SEM	13.95	5.65	9.30	13.38	1.65	9.34
Photopic white 30-Hz amplitude
Median	1.20	−2.50	−6.10	−6.80	−1.05	6.20
Min	−10.00	−7.80	−14.40	−13.90	−7.30	−9.00
Max	25.30	10.40	13.10	12.50	15.20	12.90
Mean	0.77	−0.66	−3.90	−4.28	2.32	4.80
SEM	4.53	2.39	3.70	4.15	3.68	3.08

 Difference is OD − OS. All scotopic and photopic results are presented in microvolts.

Table 5.

 

Intraocular Pressure Measurements

Dose	Before Injection	1 Day	1 Week	2 Weeks	3 Weeks	4 Weeks
0.5%	25.3, 1	23.3, 4.6	18.2, 4.2	28, 1.3	24, 3.9	26.5, 1.5
0.25%	27.8, 3.3	16.6, 2.8	21, 4.1	27.2, 2	21.2, 4.4	27, 1
No injection	28.3, 2.3	27.8, 1.8	23, 3.8	27.3, 1.1	22.7, 1.2	27, 0.9

 Right eyes were injected with 0.5% (n = 3) or 0.25% (n = 3) ketorolac. Left eyes were not injected (n = 6). Results are presented as average, SD in mm Hg.

Figure 6.

 

Light microscopy examination of the retina, performed 1 month after injection of 0.5% ketorolac. The ganglion cell layer is facing the bottom of the micrograph. H&E stain; magnification, ×700; bar, 50 μm.

Figure 7.

 

Light microscopy examination of the retina, performed 1 month after injection of saline solution. The ganglion cell layer is facing the upper side of the micrograph. H&E stain; magnification, ×300; bar, 50 μm.

Figure 8.

 

Electron microscopy of an eye injected with 0.5% ketorolac. The photoreceptor outer and inner segments, as well as the retinal pigment epithelium, are visible. Magnification, ×2400; bar, 10 μm.

Figure 9.

 

Electron microscopy of an eye injected with 0.25% ketorolac. The photoreceptors’ outer and inner segments as well as the retinal pigment epithelium are visible. Magnification, ×2400; bar, 10 μm.

Figure 10.

 

Electron microscopy of a control eye injected with saline solution. The photoreceptor outer segments and the retinal pigment epithelium are visible. A small portion of the photoreceptor inner segments appears at the top. The choriocapillaris, filled with erythrocytes, is visible at the bottom. Magnification, ×2400; bar 10 μm.

Table 6.

 

Retinal Layer Thickness

Dose	Outer Segment	ONL	INL	GCL
0.5%	36.1, 6.9	42.8, 3.4	22.2, 3.8	14.2, 2.0
0.25%	34.7, 4.2	43.0, 5.1	20.8, 4.4	11.9, 1.6
Saline solution	33.8, 5.7	41.6, 4.0	21.3, 3.5	13.3, 1.8

 Right eyes that were injected with 0.5% or 0.25% ketorolac. Left eyes were injected with balanced salt solution ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Results are presented as average, SD in micrometers.