- Email: [email protected]
A double-masked, placebo-controlled evaluation of 0.5% loteprednol etabonate in the treatment of postoperative inflammation1,2
A Double-masked, Placebo-controlled Evaluation of 0.5% Loteprednol Etabonate in the Treatment of Postoperative Inflammation The Loteprednol Etabonate Postoperative Inflammation Study Group 2 Objective: This study aimed to compare the efficacy and safety of loteprednol etabonate (LE) 0.5% to placebo (vehicle) in controlling the anterior chamber cell and flare reaction in patients undergoing cataract surgery with intraocular lens (IOL) implantation. Design: Randomized, double-masked, placebo-controlled, parallel group multicenter study. Participants: A total of 203 patients undergoing elective cataract removal and posterior chamber intraocular lens implantation who, on the day after surgery, exhibited a minimum anterior chamber inflammation score (ACI, sum of cell and flare reaction) rating of 3 (0 –9 scale). Intervention: All patients received either LE 0.5% or placebo (vehicle) four times daily in the eye that was operated on for up to 14 days after surgery. Main Outcome Measures: Resolution of ACI by final, on-treatment visit was measured. Results: The proportion of patients with ACI resolved by the final visit was 56 (55%) of 102 in the LE group and 28 (28%) of 100 in the placebo group (P ⬍ 0.001). For all the individual components of ACI (cell and flare), as well as other signs and symptoms, the resolution rate and mean change from baseline favored LE. Expanding the efficacy criterion to include patients with mild inflammation at final visit, the efficacy of LE was 95 (93%) of 102 in contrast to 65 (65%) of 100 for placebo. Among the 39 patients who did not complete the study, the majority were discontinued for inadequate anti-inflammatory effect: 25 (25%) of 101 placebo patients and 5 (5%) of 102 of LE patients. The difference in the treatment failure rates, as well the difference in the time course of failures, was both clinically meaningful and statistically significant in favor of LE (P ⬍ 0.001). Both treatments were well-tolerated. No clinically significant elevations in intraocular pressure (ⱖ10 mmHg) were seen in the LE treatment group. One patient in the placebo treatment group met this criterion. Conclusions: Loteprednol etabonate showed a clinically meaningful reduction in the signs and symptoms of postoperative anterior chamber inflammation when compared with that of placebo and had an acceptable safety profile compared with placebo. Ophthalmology 1998;105:1780 –1786 The surgical removal of cataracts combined with implantation of intraocular lenses is a common operation among elderly persons in the developed world. Intraocular inflammation may be observed after the surgical procedure, consisting of a mild iritis with increased cells and protein (flare) Originally received: December 9, 1997. Revision accepted: February 17, 1998. Manuscript no. 97841. Supported by Pharmos Corp and Bausch and Lomb Pharmaceuticals. At the time the study was conducted, Dr. Howes, Ms. Gill, Ms. Lawson, and Mr. Smerick were employees of Pharmos Corp, the sponsor of this study. Dr. Zaccardelli and Ms. Fazio were employees of Bausch and Lomb Pharmaceuticals, the licensors of loteprednol etabonate. Dr. Novack, Ms. Hart, and Dr. Crockett are consultants to Pharmos Corp but have no proprietary interest in the firm or competing firms. None of the investigators had a proprietary interest in loteprednol etabonate. Measurement scales and the Loteprednol Etabonate Postoperative Inflammation Study Group 2 are listed in Appendixes 1 and 2, respectively, at the end of this article. Reprint requests to Gary D. Novack, PhD, PharmaLogic Development, Inc, 17 Bridgegate Drive, San Rafael, CA 94903.
1780
in the anterior chamber. Although this condition usually is self-limited, the use of anti-inflammatory agents during the postoperative period allows for more rapid resolution of inflammatory symptomatology and improved patient comfort. Although some patients undergo the procedure with little or no inflammation, it is not possible to predict in advance which patients will develop clinically significant postoperative inflammation. Therefore, anti-inflammatory agents are used routinely in nearly all patients as prophylaxis. Controlled studies evaluating the efficacy and safety of anti-inflammatory agents in postoperative inflammation in patients with cataract have been conducted with both corticosteroids1– 4 and nonsteroidal anti-inflammatory agents (cyclo-oxygenase inhibitors).2,3,5–13 Based on their mechanism of action, corticosteroids offer the widest range of anti-inflammatory activity by ameliorating the immediate effects of the preformed mediators of inflammation as well as attenuating the release of newly formed mediators. However, the use of topical corticosteroids may increase the risk
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation of an elevation in intraocular pressure (IOP) and, with longer term treatment, cataract formation. In addition, it is possible that topical corticosteroids carry a risk of systemic effects such as suppression of the hypothalamic–pituitary– adrenal axis.14 Cyclo-oxygenase inhibitors do not provide the full breadth of pharmacologic activity offered by corticosteroids, and although they do not generally cause an elevation of IOP, cyclo-oxygenase inhibitors may be less comfortable for postoperative use than other agents. Loteprednol etabonate (LE) is a novel, site-active corticosteroid that was designed to retain the anti-inflammatory efficacy of a corticosteroid while lowering the risk of typical corticosteroid untoward effects.15,16 Although structurally similar to other corticosteroids, the number 20 position ketone group is absent in LE, so that it undergoes a predictable transformation to an inactive metabolite.17 Therefore, LE is designed to exert immediate anti-inflammatory activity within the eye on instillation, followed by predictable hydrolysis to an inactive metabolite. Loteprednol etabonate has been studied in a variety of inflammatory conditions affecting the eye. In two largescale clinical studies, 0.5% LE was effective in relieving the signs and symptoms of giant papillary conjunctivitis.18,19 In addition, 0.5% LE also was effective in the prevention of signs and symptoms of seasonal allergic conjunctivitis.20 Finally, 0.5% LE has been shown to be effective in the treatment of patients with uveitis.21 The study described in this report was designed to evaluate the efficacy and safety of 0.5% LE in controlling the anterior chamber cell and flare reaction in patients undergoing cataract surgery with intraocular lens implantation.
Patients and Methods This was a prospective, randomized, double-masked, placebocontrolled, parallel group multicenter (17 sites) study. Eligible patients were scheduled for cataract removal and intraocular lens implantation. On the day after surgery for cataract removal and posterior chamber intraocular lens implantation, adult patients had to exhibit a minimum anterior chamber inflammation score (ACI, sum of cell and flare reaction) rating of 3 (0 –7 scale; refer to Appendix 1). Cataract extraction by both extracapsular and phacoemulsification techniques was permitted in the study, providing that cataract removal and lens implantation were performed in a single, uncomplicated procedure. Excluded from the study were subjects having monocular potential, a history of allergic hypersensitivity to corticosteroids or to any component of the study medication, expected concurrent ocular therapy with a nonsteroidal anti-inflammatory agent (NSAID), mast cell stabilizer, antihistamine, and/or decongestant for the succeeding 15 days, or use of any of the above during the cataract surgery or within 2 days before visit 1, therapy with systemic, inhaled or topical (ocular) corticosteroids within 2 weeks before the start of the study, IOP that was greater than 30 mmHg in the postoperative treated eye, or any type of glaucoma. Also not allowed to participate in the study were women who were pregnant or nursing. Women of childbearing potential were required to use adequate birth control and to have a negative pregnancy test result before surgery. This study was approved by each investigator’s governing institutional review board, and all patients gave written informed
consent before treatment with the investigational medication. Patients were allowed to participate in the study for one eye only. Before study enrollment, potential patients underwent a screening examination (nominal day, ⫺14 to day 0). This examination included patient history (including concomitant medications), measurement of IOP, and, in the operative eye, biomicroscopy and dilated ophthalmoscopy. The patients were informed that they would be eligible for the study, approximately 24 hours after surgery, if their operative eye had an ACI score of 3 or more. After surgery, patients received an anti-infective agent only (either tobramycin four times daily, supplied by the sponsor, or an agent of the investigator’s choice and source). The duration of dosing was 7 days for tobramycin and consistent with standard of care for other antibiotics. One day after surgery (22–34 hours), an ophthalmic examination was performed, which included biomicroscopy, measurement of visual acuity and IOP, and recording of patient symptoms. Patients who met the inclusion criteria with respect to ACI score were randomized to receive either 0.5% LE or placebo (vehicle; PROC PLAN, PC-SAS version 6.12; SAS Institute, Cary, NC). Medications were supplied in identical opaque containers, and both patient and investigator were masked to the treatment assignment. In case of emergencies requiring decoding of the treatment assignment to determine therapy, a tamper-evident sealed code was maintained by the sponsor. One drop of test medication was instilled into each eye in the investigator’s office, and the patients were instructed to begin dosing every 4 hours (four times daily at approximately 0800, 1200, 1600, and 2000 hours) while awake for the next 14 days. Patients were asked to return for all subsequent visits at approximately the same time as the first postoperative visit. Ophthalmic examinations were performed at days 3, 8, 15, and 17 (visits 2, 3, 4, and 5), which included biomicroscopy, measurement of visual acuity and IOP, and recording of patient’s symptoms. Dilated ophthalmoscopy was performed on day 15. Also recorded at this visit (or if discontinued from the study, the final on-treatment visit for the patient) was an investigator’s global assessment (IGA) based on an evaluation of the control of the patient’s ocular inflammatory signs as shown through the ophthalmic findings over the duration of the study (refer to Appendix 1). Investigational medication was discontinued at the visit, and patients were asked to return on day 17 for a follow-up safety examination. A 2-day variance was allowed for each study visit. The nature of phase III clinical trials requires, as much as possible, the study design to emulate best clinical practices. Thus, a priori, systemic nonsteroidal anti-inflammatory drugs (NSAIDs) were allowed. Also allowed were artificial tears and ocular antibiotics (single agents only). Allowed only in the perioperative period were ocular NSAIDs, ocular-hypotensive agents, and antimuscarinic agents. The primary efficacy outcome was the resolution of ACI (i.e., score of 0), and the primary efficacy endpoint was the final, on-treatment visit. By-visit observations were supportive of the final visit analysis. Secondary efficacy outcomes were cell and flare individually, treatment failures (patients who discontinued early for inadequate control or who had an increase of 3 or more in their ACI score), and the IGA. The study population was to include all randomized patients with at least one on-treatment evaluation (i.e., intent to treat). A priori, the placebo resolution rate for the ACI score 15 days after surgery was estimated to be 20%. A clinically significant improvement over the placebo resolution rate was considered to be 20%. With equal allocation to treatment groups, LE and placebo, a sample size of 91 evaluable patients per group would allow rejection of the null hypothesis (H0: LE ⫽ PL) with a type I error rate of 0.05 and a power of 80% under the alternative hypothesis.
1781
Ophthalmology Volume 105, Number 9, September 1998 For the last observation carried-forward analysis, the last ontreatment visit was identified as visit 4 (completing patients) or the last on-treatment valid visit. Change in IOP from screening rather than from baseline was considered the appropriate prestudy parameter for measuring drug effect on this safety measure. This decision was based on the supposition that IOP at the postoperative day (baseline) would reflect changes caused by the surgical procedure or postsurgical inflammation or both. A clinically significant increase in IOP over screening was defined as 10 mmHg or more.22 Comparability of the treatment groups for prestudy patient characteristics was assessed for age by Student’s t test and for all other measures by Fisher’s exact test for independence. For race, the test was for white versus nonwhite. The test for equality of resolution rates was the Cochran–Mantel–Haensel test for independence (treatment ⫻ outcome) controlling for investigator (PCSAS). Planned subgroup analyses included age, gender, race, type of surgery, type of lens and baselines for ACI, cell, and flare. Logistic regressions for these prestudy characteristics were performed individually to determine whether their inclusion or the inclusion of their interaction with the treatment group in the model improved the prediction of success (resolution) compared to the model with treatment group alone. Because ACI scores are a categoric rather than continuous measure, nonparametric methods were used to analyze the treatment effect. All statistical tests for the efficacy measures were two-tailed with alpha ⫽ 0.05.
Table 1. Demographics
N Age (yrs) Mean (SD) Range ⬍ 70 ⱖ 70 Gender Male Female Race White Hispanic Black Other Iris color Light Dark Surgery type Phacoemulsification ECCE Lens type Foldable Nonfoldable
LE
Placebo
P*
102
100
69.4 (10.5) 25–95 43 (42%) 59 (58%)
71.7 (9.1) 47–99 45 (45%) 55 (55%)
0.684
43 (42%) 59 (58%)
41 (41%) 59 (59%)
0.868
75 (74%) 1 (1%) 25 (25%) 1 (1%)
70 (70%) 4 (4%) 26 (26%) 0 (0%)
0.578
50 (49%) 52 (51%)
49 (49%) 51 (51%)
0.998
81 (79%) 21 (21%)
79 (79%) 21 (21%)
0.943
67 (66%) 35 (34%)
68 (68%) 32 (32%)
0.728
0.096
* The P value for mean age was calculated by t test. All other P values were calculated by Fisher’s exact test for independence, not controlling for investigator. For race, the test was for white versus non-white.
Results There were 203 patients randomized to treatment with LE (n ⫽ 102) or placebo (n ⫽ 101). Excluded from all analyses was one patient in the placebo group who took investigational medication for less than 24 hours, started taking prednisolone acetate on day 2, and had no valid on-treatment visits. Among the 39 (19%) of 203 patients who did not complete the study, 32 (32%) of 101 received placebo and 7 (7%) of 102 received LE. The majority were discontinued for inadequate anti-inflammatory efficacy—25 placebo patients and 5 LE patients. As shown in Figure 1, the failures in the placebo group occurred throughout a 13-day postoperative period, as might be expected for patients not receiving topical anti-inflammatory medication after surgery. The relatively few failures in the LE group occurred early, indicating that patients not responding to the drug can be identified in a timely fashion. The difference in the treatment failure rates, as well as the difference in the time course of failures, were both clinically meaningful and statistically significant in favor of LE (P ⬍ 0.001).
Figure 1. Survival curve for treatment failure.
1782
The two treatment groups were similar in their demographics and prerandomization characteristics (Table 1), consistent with an efficient randomization. The mean severity of ACI at baseline (1 day postoperative just before randomization) was 3.5 ⫾ 0.1 (mean ⫾ standard error of the mean) for LE and 3.7 ⫾ 0.1 for placebo (P ⫽ 0.085). At the final visit, 56 (55%) of 102 LE patients had achieved resolution of cell and flare compared to 28 (28%) of 100 placebo patients. This 27% difference (treatment effect) in resolution rates was clinically meaningful and statistically significant (P ⬍ 0.001; Table 2). Also evaluated were the by-visit resolution rates. For both treatment groups, the resolution rates increased with time on treatment for the patients remaining on treatment at each of the scheduled visits. Ten percent of LE patients experienced resolution of ACI at visit 2, 34% at visit 3, and 58% at visit 4 compared to 9%, 17%, and 39%, respectively, for placebo. The difference in rates was statistically significant (P ⱕ 0.008) in favor of LE at visits 3 and 4. An additional tabulation of the ACI score at final visit is presented in Table 3. Listed in this table is the proportion of patients resolved (as in Table 2), as well as the proportion of patients with mild, moderate, and severe inflammation at the final visit. The proportion of patients treated with LE whose inflammation was either resolved or mild at final visit was 95 (93%) of 102 patients in contrast to 65 (65%) of 100 patients treated with placebo. At the final visit, the mean change from baseline in ACI was ⫺2.6 severity units for LE compared to ⫺1.5 for placebo. This difference was statistically significant with the median treatment effect (LE minus placebo) estimated to be a change of ⫺1 unit (P ⫽ 0.001). In terms of observed severity, the LE patients had a mean ACI score of 0.9 with 7% of patients with moderate or greater ACIs compared to a mean ACI of 2.2 for placebo with 35% of patients with moderate or greater ACIs. Over time, the mean severity in ACI decreased monotonically (i.e., a greater magnitude of effect at each subsequent visit) compared to baseline for both
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation Table 2. Anterior Chamber Inflammation Score
Visit Response at each visit* 2 (days 2–6) 3 (days 7–12) 4 (days 13–20) Final visit (LOCF)
Visit Change at each visit† 1 (baseline) 2 (days 2–6) 3 (days 7–12) 4 (days 13–20) Final visit (LOCF)
Resolved
Treatment Group
N at risk
N
%
LE Placebo LE Placebo LE Placebo LE Placebo
102 100 96 83 93 70 102 100
10 9 33 14 54 27 56 28
10 9 34 17 58 39 55 28
N/A
Treatment Group
N at risk
Mean Observed
Change
LE Placebo LE Placebo LE Placebo LE Placebo LE Placebo
102 100 102 100 96 83 93 70 102 100
3.5 3.7 2.2 2.8 1.4 2.2 0.7 1.4 0.9 2.2
N/A N/A ⫺1.3 ⫺0.9 ⫺2.2 ⫺1.4 ⫺2.8 ⫺2.3 ⫺2.6 ⫺1.5
Treatment Effect (%)
95% CI
1
⫺25%, 27%
0.834
18
⫺8%, 43%
0.005
19
⫺3%, 42%
0.008
27
6%, 48%
⬍0.001
Treatment Effect (median)
95% CI
P
N/A
N/A
P
0.0
⫺1.0, 0.0
0.060
⫺1.0
⫺1.0, 0.0
0.044
0.0
⫺1.0, 0.0
0.145
⫺1.0, ⫺1.0
0.001
⫺1.0
N/A ⫽ not applicable; CI ⫽ confidence interval; N at risk ⫽ number of patients with a score ⬎ 0 at baseline and a valid on-treatment evaluation for the visit; LOCF ⫽ last valid on-treatment observation, carried forward; Resolved ⫽ the proportion of patients at risk for whom the score was 0 at the endpoint. * The treatment effect is the median difference between treatments in resolution rates (LE ⫺ placebo) with investigators pooled. Positive treatment effects indicate that LE is favored over placebo for resolution rate. P value is from the CMH test for independence of treatment assignment and resolution, controlling for investigator. † Change is the mean change in severity compared with baseline for that measure. A negative number indicates improvement. Treatment effect is the median difference between treatments, and its 95% confidence interval is estimated by distribution-free methods with investigators pooled. A negative treatment effect indicates that LE is favored over placebo for change. P value is from the CMH test for equality of treatment group mean ranks for change in severity controlling for investigator.
had resolution of cell compared to 31 (31%) of 100 placebo patients. The 29% difference (treatment effect) in resolution rates was statistically significant (P ⬍ 0.001). One LE patient had a flare rating of 0 at baseline and was excluded from the flare analysis. At final visit, 68 (67%) of 101 LE patients had resolution of flare compared to 36 (36%) of 100 placebo patients. The 31% difference (treatment effect) in resolution rates was statistically significant (P ⬍ 0.001). Patients included in the intent-to-treat population were eligible for inclusion in analysis of the IGA of treatment. Two placebo patients had no global assessments recorded by the investigator at their end-of-treatment evaluation. Thus, there were 102 LE patients and 98 placebo patients in the analysis of IGA. The mean assessment was 0.8 (full-to-reasonable control) for LE and 2.0 (slight improvement) for placebo. This difference was statistically
treatment groups. The treatment effect favored loteprednol etabonate and was statistically significant at visit 3. When subgroup analyses were performed, regardless of the treatment group, patients with light irides were more likely to have resolution of ACI than those with dark irides (P ⫽ 0.004), and the ACIs for patients with phacoemulsification were more likely to resolve than those with extracapsular cataract extraction (P ⫽ 0.048). The interaction of treatment with baseline cell score was a significant addition to the model (P ⫽ 0.048), indicating that the treatment effect (LE–PL) was greater in patients with more severe ACI ratings than in those with milder ratings. There was no evidence for a difference in treatment effect across categories (i.e., interaction) for any other variable. The efficacy of LE was similar when cell and flare were analyzed individually. At final visit, 61 (60%) of 102 LE patients
Table 3. Anterior Chamber Inflammation Score: Distribution at Final Visit 0 (resolved)
1–2 (mild)
0–2 (resolved or mild)
3–6 (moderate)
7–9 (severe)
Treatment Group
N at risk
N
%
N
%
N
%
N
%
N
%
Loteprednol etabonate Placebo
102 100
56 28
56 28
39 37
38 37
95 65
93 65
6 34
6 34
1 1
1 1
1783
Ophthalmology Volume 105, Number 9, September 1998 significant with the median treatment effect estimated to be a difference of ⫺1 scale unit (P ⬍ 0.001). Ninety-one (89%) of 102 LE patients were reported to have at least reasonable control compared to 48 (49%) of 98 placebo patients. The resolution rates for supportive signs and symptoms all numerically favored LE, with the exception of dryness. The difference between the treatment groups was statistically significant (P ⱕ 0.05) for chemosis, erythema, bulbar injection, ciliary flush, pain, photophobia, tearing, and discomfort. For dryness and hyphema, 10 or fewer patients (i.e., ⬍5% of the population) had a finding at baseline. When the LE-treated patients were evaluated several days after therapy stopped, 91% of patients remained unchanged or improved from their visit 4 values. The average IOP in the affected eye at screening was 16.4 mmHg for LE (n ⫽ 102) and 16.3 mmHg for placebo (n ⫽ 100; P ⫽ 0.847). One patient (placebo treatment group) did not have an IOP at screening and was excluded from the analysis of change in IOP. No patients in the LE treatment group had an IOP elevation of 10 mmHg; one patient in the placebo treatment group had an IOP elevation of 10 mmHg or more during the on-treatment phase of the study (from 19 mmHg at screening to 48 mmHg at day 11, study medication discontinued, treated with ocular-hypotensive agents and fluorometholone, and resolved by day 22). Overall, there was a mean decrease in IOP for both treatment groups for all visit-day intervals (1–2 mmHg) when compared to the baseline for IOP obtained at the screening visit. No difference between treatment groups in mean change from screening was detected for any interval (P ⬎ 0.206). At least one treatment-emergent medical event (mostly consistent with typical postoperative signs and symptoms) was observed in 55 (54%) of 102 LE patients and 76 (75%) of 101 placebo patients. This difference in favor of LE was statistically significant (P ⫽ 0.002). A total of four patients (all in the placebo group) were terminated from further study participation due to an adverse event. At day 2, one patient had tearing, blurred vision, pain, and photophobia, and a second patient had toxic keratitis. At day 8, one patient had corneal edema. At day 11, one patient had elevated IOP. An additional patient in the LE group was hospitalized for observation on study day 11 due to itching of the arms, legs, and trunk; fatigue; and leg cramps. The itching, which was treated with oral hydroxyzine, resolved within 2 days and was judged to be unrelated to the study medication, and treatment was not stopped. Other than these five patients, there were no serious, life-threatening, or sight-threatening events in this elderly, postoperative population during the study period.
Discussion The objective of this study was to evaluate the efficacy and safety of 0.5% LE in controlling the anterior chamber cell and flare reaction in patients undergoing cataract surgery with intraocular lens implantation. The LE was more effective than its vehicle in primary, secondary, and supportive measures of efficacy. In the primary measure of efficacy, the resolution of ACI in the eye that was operated on at final visit, 56 (55%) of 1020 LE patients had achieved resolution compared to 28 (28%) of 100 placebo patients. This 27% difference (treatment effect) in resolution rates was clinically meaningful and statistically significant (P ⬍ 0.001). Expanding the efficacy criterion to include patients with mild inflammation at final visit, the efficacy of LE was 95 (93%) of 102 in contrast to 65 (65%) of 100 for placebo.
1784
Secondary and supportive analyses of the primary efficacy measure were complementary of the primary analysis of ACI. Resolution rate at each visit, cell and flare individually at final visit and each examination, treatment failures, and investigator global assessment were all consistent in showing a clinically and statistically significant treatment effect of LE. Regardless of the treatment group, patients with light irides were more likely to have resolution of ACI than those with dark irides, and the ACIs for patients with phacoemulsification were more likely to resolve than those with extracapsular cataract extraction. The treatment effect of LE relative to its vehicle was greater in patients with more severe baseline ACI ratings than in those with milder ratings (i.e., treatment by baseline interactions). With this one exception (which is consistent with a baseline effect), there was no evidence for a difference in treatment effect across categories (i.e., treatment by factor interaction) for any variable. Thus, the efficacy of LE over its vehicle generally was consistent within all subgroups. As noted previously, both corticosteroids and nonsteroidal anti-inflammatory agents (cyclo-oxygenase inhibitors) have been evaluated for the treatment of postsurgical (cataract) inflammation.2,3,9,10,13 We attempted to evaluate the efficacy of LE observed in this current study in light of some of these previous reports for other agents. Approved in the United States for treatment of postoperative inflammation in recent years are rimexolone, diclofenac, and ketorolac, all on the basis of similarly designed, placebo-controlled, treatment trials (in contrast to prophylactic studies). In a double-masked study of 197 patients, rimexolone was more effective than placebo in suppressing cell and flare. The magnitude of this effect was in the same general range as in the current study.23 The diclofenac study appears to be available only in abstract form with limited data (Vickers FF, McGuigan LJB, Ford C, et al. The effect of diclofenac sodium ophthalmic on the treatment of postoperative inflammation. Invest Ophthalmol Vis Sci Suppl 1991;32:793), and we were unable to find a publication regarding ketorolac for this indication. A host of other studies compared either corticosteroids or NSAIDs to other agents or, in some cases, to a placebo treatment. These studies were predominantly prophylactic in contrast to treatment studies. Efficacy measures included clinical signs of inflammation (cells and flare) as observed by an ophthalmologist and, in some cases, an instrument that measures anterior chamber light properties and translates them to cells and flare. Additional studies used a measure of blood–aqueous barrier function (aqueous fluorophotometry). For the most part, active agents were more effective than their placebo and similar in efficacy to other active agents. The latter comparisons typically were of low power because of sample size and the smaller definition of a clinically significant difference.2,3,8 –11,13,24 Although it is difficult to compare across studies, given the changes in surgical technique over the years, the treatment effect of LE seen in the current vehiclecontrolled study is at least as great as the prophylactic effect in any of the previous reports. We considered using a Kowa Laser Cell Flare meter to evaluate postoperative inflammation.25 However, the logis-
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation tics of placing such a device at all study centers was substantial. As well, we thought it important to show the anti-inflammatory efficacy of LE in a method used by all surgeons: observations at the slit lamp. With respect to safety, both LE and placebo had favorable safety profiles, including ocular tolerance. There was no evidence for any safety concern with LE regarding treatment-emergent medical events. Furthermore, there was no issue with elevation in IOP in LE-treated patients in this study. Granted that the duration of treatment in this study, 14 days, is short of the traditional period for corticosteroidinduced elevation in IOP.26,27 However, in corticosteroidresponders treated for 6 weeks, LE had a very low incidence IOP elevation relative to prednisolone acetate.28 Furthermore, in other studies in which patients received LE for 4 to 6 weeks, the incidence of elevations in IOP of 10 mmHg or greater was similar to vehicle and lower than prednisolone acetate.29 Thus, LE was both clinically and statistically significantly favored over its vehicle in the treatment of postoperative inflammation. There was no evidence of rebound of signs and symptoms after cessation of therapy. Further, the safety profile of LE was excellent, with little evidence of clinically significant elevations in IOP or deleterious effects on postoperative recovery. Loteprednol etabonate showed a clinically meaningful reduction in the signs and symptoms of postoperative anterior chamber inflammation compared with that of placebo. It also had an acceptable safety profile compared with that of placebo. Acknowledgment. The authors thank Ellen R. Strahlman, MD, MHSc, for her contribution.
Appendix 1 Measurement Scales Cell: Determined using a slit beam at 1-mm height by 1-mm width with maximum luminance of the Haag–Streit (or equivalent) slit lamp. Pigment cells and erythrocytes were to be ignored. 0 ⫽ Less than or equal to 5 cells 1 ⫽ 6 to 10 cells 2 ⫽ 11 to 20 cells 3 ⫽ 21 to 40 cells 4 ⫽ More than 40 cells 5 ⫽ Hypopyon Flare: Determined using the widest slit beam at 1-mm height with maximum luminance of the Haag–Streit (or equivalent) slit lamp. 0 ⫽ None to trace 1 ⫽ Mild (clearly noticeable, visible) 2 ⫽ Moderate (without plastic aqueous) 3 ⫽ Marked (with plastic aqueous) 4 ⫽ Severe (with fibrin deposits or clots or both) Anterior chamber inflammation (ACI): Sum of cell and flare score: 0 ⫽ None to trace
1–2 ⫽ Mild 3– 4 ⫽ Moderate (required at baseline) 5– 6 ⫽ Marked 7– 8 ⫽ Severe 9 ⫽ Hypopyon and severe flare IGA (for entire study) 0 ⫽ Fully controlled: inflammation is suppressed or cured, virtually no signs or symptoms 1 ⫽ Reasonably controlled: inflammation is diminished, moderate decrease in signs or symptoms 2 ⫽ Slightly controlled: Small decrease in signs and symptoms 3 ⫽ Unchanged: No response or overall change in signs and symptoms 4 ⫽ Worse: Overall increase in signs and symptoms Other signs and symptoms (chemosis, erythema, palpebral injection, bulbar injection, corneal edema, pain, photophobia, itching, tearing, dryness, discharge, and discomfort): 0 ⫽ Absent 1 ⫽ Mild 2 ⫽ Moderate 3 ⫽ Severe Hyphema and ciliary flush were scored as present (1) or absent (0).
Appendix 2 The Loteprednol Etabonate Postoperative Inflammation Study Group 2 is as follows: Investigators Cecil Beehler, MD, Ft. Myers, FL; Bruce Bodner, MD, Norfolk, VA; Bradley Bowman, MD, Dallas, TX; David Cooke, MD, St. Joseph, MI; J. Luther Crabb, MD, Memphis, TN; L. Raymond DeBarge, MD, Fort Oglethorpe, GA; Eric Donnenfeld, MD, Rockville Centre, NY; Marcel Estopinal, MD, Nashville, TN; Kenneth Fox, MD, Fredericksburg, VA; Peter Kastl, MD, PhD, New Orleans, LA; Kenneth Olander, MD, PhD, Milwaukee, WI; Elizabeth Sharpe, MD, Mt. Pleasant, SC; Robert Shofner, MD, Nashville, TN; Janis Stahl, MD, Golden, CO; Dara Stevenson, MD, New Orleans, LA; William Stewart, MD, Charleston, SC; Tom Walters, MD, Austin, TX. Study Coordination Sue Gill, John Howes, PhD, Cathy Lawson, Michael Smerick, Jr., BS (Pharmos Corp, Alachau, FL), David Zaccardelli, PharmD, Roselyn Fazio, BS (Bausch and Lomb Pharmaceuticals, Tampa, FL), and Gary D. Novack, PhD (Pharmacologic Development, Inc, San Rafael, CA). Biostatistics and Data Management R. Stephens Crockett, PhD, and Kathryn Hart, MA (D.A.T.A., Inc, Murray, KY).
References 1. Corboy JM. Corticosteroid therapy for the reduction of postoperative inflammation after cataract extraction. Am J Ophthalmol 1976;82:923–7.
1785
Ophthalmology Volume 105, Number 9, September 1998 2. Flach AJ, Kraff MC, Sanders DR, Tanenbaum L. The quantitative effect of 0.5% ketorolac tromethamine solution and 0.1% dexamethasone sodium phosphate solution on postsurgical blood–aqueous barrier. Arch Ophthalmol 1988;106: 480 –3. 3. Othenin–Girard P, Tritten JJ, Pittet N, Herbort CP. Dexamethasone versus diclofenac sodium eyedrops to treat inflammation after cataract surgery. J Cataract Refract Surg 1994;20: 9 –12. 4. Ramsell TG, Bartholomew RS, Walker SR. Clinical evaluation of clobetasone butyrate: a comparative study of its effects in postoperative inflammation and on intraocular pressure. Br J Ophthalmol 1980;64:43–5. 5. Araie M, Sawa M, Takase M. Topical flurbiprofen and diclofenac suppress blood–aqueous barrier breakdown in cataract surgery: a flurophotometric study. Jpn J Cln Ophthalmol 1983;27:535– 42. 6. Bizzotto MF, De Franco I, Carillio F. Efficacy of flurbiprofen versus placebo on postoperative course following cataract extraction. Drugs Exp Clin Res 1984;10:421–5. 7. Blaydes JE Jr, Kelley EP, Walt JG, et al. Flurbiprofen 0.03% for the control of inflammation following cataract extraction by phacoemulsification. J Cataract Refract Surg 1993;19: 481–7. 8. Flach AJ, Jaffe NS, Akers WA. The effect of ketorolac tromethamine in reducing postoperative inflammation: double-mask parallel comparison with dexamethasone. Ann Ophthalmol 1989;21:407–11. 9. Flach AJ, Lavelle CJ, Olander KW, et al. The effect of ketorolac tromethamine solution 0.5% in reducing postoperative inflammation after cataract extraction and intraocular lens implantation. Ophthalmology 1988;95:1279 – 84. 10. Roberts CW. Pretreatment with topical diclofenac sodium to decrease postoperative inflammation. Ophthalmology 1996; 103:636 –9. 11. Sabiston D, Tessler H, Sumers K, et al. Reduction of inflammation following cataract surgery by the non-steroidal antiinflammatory drug, flurbiprofen. Ophthalmic Surg 1987;18: 873–7. 12. Sabiston DW, Robinson IG. An evaluation of the anti-inflammatory effect of flurbiprofen after cataract extraction. Br J Ophthalmol 1987;71:418 –21. 13. Ostrov CS, Sirkin SR, Deutsch WE, et al. Ketorolac, prednisolone, and dexamethasone for postoperative inflammation. Clin Ther 1997;19:259 –72. 14. Yablonski ME, Burde RM, Kolker AE, Becker B. Cataracts induced by topical dexamethasone in diabetics. Arch Ophthalmol 1978;96:474 – 6. 15. Bodor N. Novel approaches to the design of safer drugs: soft drugs and site-specific chemical delivery systems. Advances in Drug Research 1984;13:255–331.
1786
16. Bodor N. Soft drugs: principles and methods for the design of safe drugs. Med Res Rev 1984;4:449 – 69. 17. Druzgala P, Wu WM, Bodor N. Ocular absorption and distribution of loteprednol etabonate, a soft steroid, in rabbit eyes. Curr Eye Res 1991;10:933–7. 18. Asbell P, Howes J. A double-masked, placebo-controlled evaluation of the efficacy and safety of loteprednol etabonate in the treatment of giant papillary conjunctivitis. CLAO J 1997;23:31– 6. 19. Friedlaender MH, Howes J, for The Loteprednol Etabonate Giant Papillary Conjunctivitis Study Group I. A doublemasked, placebo-controlled evaluation of the efficacy and safety of loteprednol etabonate in the treatment of giant papillary conjunctivitis. Am J Ophthalmol 1997;123:455– 64. 20. Dell SJ, Shulman DG, Lowry GM, et al. A controlled evaluation of the efficacy and safety of loteprednol etabonate in the prophylactic treatment of seasonal allergic conjuctivitis. Am J Ophthalmol 1997;123:791–7. 21. Neumann R, Howes J. Loteprednol etabonate. A novel ocular steroid with improved safety profile. In: Nussenblatt RB, Whitcup SM, Caspi RM, Gery I, eds. Advances in Ocular Immunology: Proceedings of the 6th International Symposium on the Immunology and Immunopathology of the Eye. Bethesda, MD; Amsterdam; New York: Elsevier, 1994 (Int’l Congress Series no. 1068). 22. Stewart RH, Smith JP, Rosenthal AL. Ocular pressure response to fluorometholone acetate and dexamethasone sodium phosphate. Curr Eye Res 1984;3:835–9. 23. Assil KK, Massry G, Lehmann R, et al. Control of ocular inflammation after cataract extraction with rimexolone 1% ophthalmic suspension. J Cataract Refract Surg 1997;23: 750 –7. 24. Kraff MC, Sanders DR, McGuigan L, Raanan MG. Inhibition of blood-aqueous humor barrier breakdown with diclofenac. A fluorophotometric study. Arch Ophthalmol 1990;108:380 –3. 25. Drews RC. Management of postoperative inflammation: dexamethasone versus flurbiprofen, a quantitative study using the new Flare Cell Meter. Ophthalmic Surg 1990;21:560 –2. 26. Armaly MF. Statistical attributes of the steroid hypertensive response in the clinically normal eye. I. The demonstration of three levels of response. Invest Ophthalmol 1965;4:187–97. 27. Becker B. Intraocular pressure response to topical corticosteroids. Invest Ophthalmol 1965;4:198 –205. 28. Bartlett JD, Horwitz B, Laibovitz R, Howes JF. Intraocular pressure response to loteprednol etabonate in known steroid responders. J Ocul Pharmacol 1993;9:157– 65. 29. Novack GD, Howes J, Crockett RS, Sherwood MB. Change in intraocular pressure during chronic use of loteprednol etabonate. J Glaucoma, in press.
1780
in the anterior chamber. Although this condition usually is self-limited, the use of anti-inflammatory agents during the postoperative period allows for more rapid resolution of inflammatory symptomatology and improved patient comfort. Although some patients undergo the procedure with little or no inflammation, it is not possible to predict in advance which patients will develop clinically significant postoperative inflammation. Therefore, anti-inflammatory agents are used routinely in nearly all patients as prophylaxis. Controlled studies evaluating the efficacy and safety of anti-inflammatory agents in postoperative inflammation in patients with cataract have been conducted with both corticosteroids1– 4 and nonsteroidal anti-inflammatory agents (cyclo-oxygenase inhibitors).2,3,5–13 Based on their mechanism of action, corticosteroids offer the widest range of anti-inflammatory activity by ameliorating the immediate effects of the preformed mediators of inflammation as well as attenuating the release of newly formed mediators. However, the use of topical corticosteroids may increase the risk
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation of an elevation in intraocular pressure (IOP) and, with longer term treatment, cataract formation. In addition, it is possible that topical corticosteroids carry a risk of systemic effects such as suppression of the hypothalamic–pituitary– adrenal axis.14 Cyclo-oxygenase inhibitors do not provide the full breadth of pharmacologic activity offered by corticosteroids, and although they do not generally cause an elevation of IOP, cyclo-oxygenase inhibitors may be less comfortable for postoperative use than other agents. Loteprednol etabonate (LE) is a novel, site-active corticosteroid that was designed to retain the anti-inflammatory efficacy of a corticosteroid while lowering the risk of typical corticosteroid untoward effects.15,16 Although structurally similar to other corticosteroids, the number 20 position ketone group is absent in LE, so that it undergoes a predictable transformation to an inactive metabolite.17 Therefore, LE is designed to exert immediate anti-inflammatory activity within the eye on instillation, followed by predictable hydrolysis to an inactive metabolite. Loteprednol etabonate has been studied in a variety of inflammatory conditions affecting the eye. In two largescale clinical studies, 0.5% LE was effective in relieving the signs and symptoms of giant papillary conjunctivitis.18,19 In addition, 0.5% LE also was effective in the prevention of signs and symptoms of seasonal allergic conjunctivitis.20 Finally, 0.5% LE has been shown to be effective in the treatment of patients with uveitis.21 The study described in this report was designed to evaluate the efficacy and safety of 0.5% LE in controlling the anterior chamber cell and flare reaction in patients undergoing cataract surgery with intraocular lens implantation.
Patients and Methods This was a prospective, randomized, double-masked, placebocontrolled, parallel group multicenter (17 sites) study. Eligible patients were scheduled for cataract removal and intraocular lens implantation. On the day after surgery for cataract removal and posterior chamber intraocular lens implantation, adult patients had to exhibit a minimum anterior chamber inflammation score (ACI, sum of cell and flare reaction) rating of 3 (0 –7 scale; refer to Appendix 1). Cataract extraction by both extracapsular and phacoemulsification techniques was permitted in the study, providing that cataract removal and lens implantation were performed in a single, uncomplicated procedure. Excluded from the study were subjects having monocular potential, a history of allergic hypersensitivity to corticosteroids or to any component of the study medication, expected concurrent ocular therapy with a nonsteroidal anti-inflammatory agent (NSAID), mast cell stabilizer, antihistamine, and/or decongestant for the succeeding 15 days, or use of any of the above during the cataract surgery or within 2 days before visit 1, therapy with systemic, inhaled or topical (ocular) corticosteroids within 2 weeks before the start of the study, IOP that was greater than 30 mmHg in the postoperative treated eye, or any type of glaucoma. Also not allowed to participate in the study were women who were pregnant or nursing. Women of childbearing potential were required to use adequate birth control and to have a negative pregnancy test result before surgery. This study was approved by each investigator’s governing institutional review board, and all patients gave written informed
consent before treatment with the investigational medication. Patients were allowed to participate in the study for one eye only. Before study enrollment, potential patients underwent a screening examination (nominal day, ⫺14 to day 0). This examination included patient history (including concomitant medications), measurement of IOP, and, in the operative eye, biomicroscopy and dilated ophthalmoscopy. The patients were informed that they would be eligible for the study, approximately 24 hours after surgery, if their operative eye had an ACI score of 3 or more. After surgery, patients received an anti-infective agent only (either tobramycin four times daily, supplied by the sponsor, or an agent of the investigator’s choice and source). The duration of dosing was 7 days for tobramycin and consistent with standard of care for other antibiotics. One day after surgery (22–34 hours), an ophthalmic examination was performed, which included biomicroscopy, measurement of visual acuity and IOP, and recording of patient symptoms. Patients who met the inclusion criteria with respect to ACI score were randomized to receive either 0.5% LE or placebo (vehicle; PROC PLAN, PC-SAS version 6.12; SAS Institute, Cary, NC). Medications were supplied in identical opaque containers, and both patient and investigator were masked to the treatment assignment. In case of emergencies requiring decoding of the treatment assignment to determine therapy, a tamper-evident sealed code was maintained by the sponsor. One drop of test medication was instilled into each eye in the investigator’s office, and the patients were instructed to begin dosing every 4 hours (four times daily at approximately 0800, 1200, 1600, and 2000 hours) while awake for the next 14 days. Patients were asked to return for all subsequent visits at approximately the same time as the first postoperative visit. Ophthalmic examinations were performed at days 3, 8, 15, and 17 (visits 2, 3, 4, and 5), which included biomicroscopy, measurement of visual acuity and IOP, and recording of patient’s symptoms. Dilated ophthalmoscopy was performed on day 15. Also recorded at this visit (or if discontinued from the study, the final on-treatment visit for the patient) was an investigator’s global assessment (IGA) based on an evaluation of the control of the patient’s ocular inflammatory signs as shown through the ophthalmic findings over the duration of the study (refer to Appendix 1). Investigational medication was discontinued at the visit, and patients were asked to return on day 17 for a follow-up safety examination. A 2-day variance was allowed for each study visit. The nature of phase III clinical trials requires, as much as possible, the study design to emulate best clinical practices. Thus, a priori, systemic nonsteroidal anti-inflammatory drugs (NSAIDs) were allowed. Also allowed were artificial tears and ocular antibiotics (single agents only). Allowed only in the perioperative period were ocular NSAIDs, ocular-hypotensive agents, and antimuscarinic agents. The primary efficacy outcome was the resolution of ACI (i.e., score of 0), and the primary efficacy endpoint was the final, on-treatment visit. By-visit observations were supportive of the final visit analysis. Secondary efficacy outcomes were cell and flare individually, treatment failures (patients who discontinued early for inadequate control or who had an increase of 3 or more in their ACI score), and the IGA. The study population was to include all randomized patients with at least one on-treatment evaluation (i.e., intent to treat). A priori, the placebo resolution rate for the ACI score 15 days after surgery was estimated to be 20%. A clinically significant improvement over the placebo resolution rate was considered to be 20%. With equal allocation to treatment groups, LE and placebo, a sample size of 91 evaluable patients per group would allow rejection of the null hypothesis (H0: LE ⫽ PL) with a type I error rate of 0.05 and a power of 80% under the alternative hypothesis.
1781
Ophthalmology Volume 105, Number 9, September 1998 For the last observation carried-forward analysis, the last ontreatment visit was identified as visit 4 (completing patients) or the last on-treatment valid visit. Change in IOP from screening rather than from baseline was considered the appropriate prestudy parameter for measuring drug effect on this safety measure. This decision was based on the supposition that IOP at the postoperative day (baseline) would reflect changes caused by the surgical procedure or postsurgical inflammation or both. A clinically significant increase in IOP over screening was defined as 10 mmHg or more.22 Comparability of the treatment groups for prestudy patient characteristics was assessed for age by Student’s t test and for all other measures by Fisher’s exact test for independence. For race, the test was for white versus nonwhite. The test for equality of resolution rates was the Cochran–Mantel–Haensel test for independence (treatment ⫻ outcome) controlling for investigator (PCSAS). Planned subgroup analyses included age, gender, race, type of surgery, type of lens and baselines for ACI, cell, and flare. Logistic regressions for these prestudy characteristics were performed individually to determine whether their inclusion or the inclusion of their interaction with the treatment group in the model improved the prediction of success (resolution) compared to the model with treatment group alone. Because ACI scores are a categoric rather than continuous measure, nonparametric methods were used to analyze the treatment effect. All statistical tests for the efficacy measures were two-tailed with alpha ⫽ 0.05.
Table 1. Demographics
N Age (yrs) Mean (SD) Range ⬍ 70 ⱖ 70 Gender Male Female Race White Hispanic Black Other Iris color Light Dark Surgery type Phacoemulsification ECCE Lens type Foldable Nonfoldable
LE
Placebo
P*
102
100
69.4 (10.5) 25–95 43 (42%) 59 (58%)
71.7 (9.1) 47–99 45 (45%) 55 (55%)
0.684
43 (42%) 59 (58%)
41 (41%) 59 (59%)
0.868
75 (74%) 1 (1%) 25 (25%) 1 (1%)
70 (70%) 4 (4%) 26 (26%) 0 (0%)
0.578
50 (49%) 52 (51%)
49 (49%) 51 (51%)
0.998
81 (79%) 21 (21%)
79 (79%) 21 (21%)
0.943
67 (66%) 35 (34%)
68 (68%) 32 (32%)
0.728
0.096
* The P value for mean age was calculated by t test. All other P values were calculated by Fisher’s exact test for independence, not controlling for investigator. For race, the test was for white versus non-white.
Results There were 203 patients randomized to treatment with LE (n ⫽ 102) or placebo (n ⫽ 101). Excluded from all analyses was one patient in the placebo group who took investigational medication for less than 24 hours, started taking prednisolone acetate on day 2, and had no valid on-treatment visits. Among the 39 (19%) of 203 patients who did not complete the study, 32 (32%) of 101 received placebo and 7 (7%) of 102 received LE. The majority were discontinued for inadequate anti-inflammatory efficacy—25 placebo patients and 5 LE patients. As shown in Figure 1, the failures in the placebo group occurred throughout a 13-day postoperative period, as might be expected for patients not receiving topical anti-inflammatory medication after surgery. The relatively few failures in the LE group occurred early, indicating that patients not responding to the drug can be identified in a timely fashion. The difference in the treatment failure rates, as well as the difference in the time course of failures, were both clinically meaningful and statistically significant in favor of LE (P ⬍ 0.001).
Figure 1. Survival curve for treatment failure.
1782
The two treatment groups were similar in their demographics and prerandomization characteristics (Table 1), consistent with an efficient randomization. The mean severity of ACI at baseline (1 day postoperative just before randomization) was 3.5 ⫾ 0.1 (mean ⫾ standard error of the mean) for LE and 3.7 ⫾ 0.1 for placebo (P ⫽ 0.085). At the final visit, 56 (55%) of 102 LE patients had achieved resolution of cell and flare compared to 28 (28%) of 100 placebo patients. This 27% difference (treatment effect) in resolution rates was clinically meaningful and statistically significant (P ⬍ 0.001; Table 2). Also evaluated were the by-visit resolution rates. For both treatment groups, the resolution rates increased with time on treatment for the patients remaining on treatment at each of the scheduled visits. Ten percent of LE patients experienced resolution of ACI at visit 2, 34% at visit 3, and 58% at visit 4 compared to 9%, 17%, and 39%, respectively, for placebo. The difference in rates was statistically significant (P ⱕ 0.008) in favor of LE at visits 3 and 4. An additional tabulation of the ACI score at final visit is presented in Table 3. Listed in this table is the proportion of patients resolved (as in Table 2), as well as the proportion of patients with mild, moderate, and severe inflammation at the final visit. The proportion of patients treated with LE whose inflammation was either resolved or mild at final visit was 95 (93%) of 102 patients in contrast to 65 (65%) of 100 patients treated with placebo. At the final visit, the mean change from baseline in ACI was ⫺2.6 severity units for LE compared to ⫺1.5 for placebo. This difference was statistically significant with the median treatment effect (LE minus placebo) estimated to be a change of ⫺1 unit (P ⫽ 0.001). In terms of observed severity, the LE patients had a mean ACI score of 0.9 with 7% of patients with moderate or greater ACIs compared to a mean ACI of 2.2 for placebo with 35% of patients with moderate or greater ACIs. Over time, the mean severity in ACI decreased monotonically (i.e., a greater magnitude of effect at each subsequent visit) compared to baseline for both
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation Table 2. Anterior Chamber Inflammation Score
Visit Response at each visit* 2 (days 2–6) 3 (days 7–12) 4 (days 13–20) Final visit (LOCF)
Visit Change at each visit† 1 (baseline) 2 (days 2–6) 3 (days 7–12) 4 (days 13–20) Final visit (LOCF)
Resolved
Treatment Group
N at risk
N
%
LE Placebo LE Placebo LE Placebo LE Placebo
102 100 96 83 93 70 102 100
10 9 33 14 54 27 56 28
10 9 34 17 58 39 55 28
N/A
Treatment Group
N at risk
Mean Observed
Change
LE Placebo LE Placebo LE Placebo LE Placebo LE Placebo
102 100 102 100 96 83 93 70 102 100
3.5 3.7 2.2 2.8 1.4 2.2 0.7 1.4 0.9 2.2
N/A N/A ⫺1.3 ⫺0.9 ⫺2.2 ⫺1.4 ⫺2.8 ⫺2.3 ⫺2.6 ⫺1.5
Treatment Effect (%)
95% CI
1
⫺25%, 27%
0.834
18
⫺8%, 43%
0.005
19
⫺3%, 42%
0.008
27
6%, 48%
⬍0.001
Treatment Effect (median)
95% CI
P
N/A
N/A
P
0.0
⫺1.0, 0.0
0.060
⫺1.0
⫺1.0, 0.0
0.044
0.0
⫺1.0, 0.0
0.145
⫺1.0, ⫺1.0
0.001
⫺1.0
N/A ⫽ not applicable; CI ⫽ confidence interval; N at risk ⫽ number of patients with a score ⬎ 0 at baseline and a valid on-treatment evaluation for the visit; LOCF ⫽ last valid on-treatment observation, carried forward; Resolved ⫽ the proportion of patients at risk for whom the score was 0 at the endpoint. * The treatment effect is the median difference between treatments in resolution rates (LE ⫺ placebo) with investigators pooled. Positive treatment effects indicate that LE is favored over placebo for resolution rate. P value is from the CMH test for independence of treatment assignment and resolution, controlling for investigator. † Change is the mean change in severity compared with baseline for that measure. A negative number indicates improvement. Treatment effect is the median difference between treatments, and its 95% confidence interval is estimated by distribution-free methods with investigators pooled. A negative treatment effect indicates that LE is favored over placebo for change. P value is from the CMH test for equality of treatment group mean ranks for change in severity controlling for investigator.
had resolution of cell compared to 31 (31%) of 100 placebo patients. The 29% difference (treatment effect) in resolution rates was statistically significant (P ⬍ 0.001). One LE patient had a flare rating of 0 at baseline and was excluded from the flare analysis. At final visit, 68 (67%) of 101 LE patients had resolution of flare compared to 36 (36%) of 100 placebo patients. The 31% difference (treatment effect) in resolution rates was statistically significant (P ⬍ 0.001). Patients included in the intent-to-treat population were eligible for inclusion in analysis of the IGA of treatment. Two placebo patients had no global assessments recorded by the investigator at their end-of-treatment evaluation. Thus, there were 102 LE patients and 98 placebo patients in the analysis of IGA. The mean assessment was 0.8 (full-to-reasonable control) for LE and 2.0 (slight improvement) for placebo. This difference was statistically
treatment groups. The treatment effect favored loteprednol etabonate and was statistically significant at visit 3. When subgroup analyses were performed, regardless of the treatment group, patients with light irides were more likely to have resolution of ACI than those with dark irides (P ⫽ 0.004), and the ACIs for patients with phacoemulsification were more likely to resolve than those with extracapsular cataract extraction (P ⫽ 0.048). The interaction of treatment with baseline cell score was a significant addition to the model (P ⫽ 0.048), indicating that the treatment effect (LE–PL) was greater in patients with more severe ACI ratings than in those with milder ratings. There was no evidence for a difference in treatment effect across categories (i.e., interaction) for any other variable. The efficacy of LE was similar when cell and flare were analyzed individually. At final visit, 61 (60%) of 102 LE patients
Table 3. Anterior Chamber Inflammation Score: Distribution at Final Visit 0 (resolved)
1–2 (mild)
0–2 (resolved or mild)
3–6 (moderate)
7–9 (severe)
Treatment Group
N at risk
N
%
N
%
N
%
N
%
N
%
Loteprednol etabonate Placebo
102 100
56 28
56 28
39 37
38 37
95 65
93 65
6 34
6 34
1 1
1 1
1783
Ophthalmology Volume 105, Number 9, September 1998 significant with the median treatment effect estimated to be a difference of ⫺1 scale unit (P ⬍ 0.001). Ninety-one (89%) of 102 LE patients were reported to have at least reasonable control compared to 48 (49%) of 98 placebo patients. The resolution rates for supportive signs and symptoms all numerically favored LE, with the exception of dryness. The difference between the treatment groups was statistically significant (P ⱕ 0.05) for chemosis, erythema, bulbar injection, ciliary flush, pain, photophobia, tearing, and discomfort. For dryness and hyphema, 10 or fewer patients (i.e., ⬍5% of the population) had a finding at baseline. When the LE-treated patients were evaluated several days after therapy stopped, 91% of patients remained unchanged or improved from their visit 4 values. The average IOP in the affected eye at screening was 16.4 mmHg for LE (n ⫽ 102) and 16.3 mmHg for placebo (n ⫽ 100; P ⫽ 0.847). One patient (placebo treatment group) did not have an IOP at screening and was excluded from the analysis of change in IOP. No patients in the LE treatment group had an IOP elevation of 10 mmHg; one patient in the placebo treatment group had an IOP elevation of 10 mmHg or more during the on-treatment phase of the study (from 19 mmHg at screening to 48 mmHg at day 11, study medication discontinued, treated with ocular-hypotensive agents and fluorometholone, and resolved by day 22). Overall, there was a mean decrease in IOP for both treatment groups for all visit-day intervals (1–2 mmHg) when compared to the baseline for IOP obtained at the screening visit. No difference between treatment groups in mean change from screening was detected for any interval (P ⬎ 0.206). At least one treatment-emergent medical event (mostly consistent with typical postoperative signs and symptoms) was observed in 55 (54%) of 102 LE patients and 76 (75%) of 101 placebo patients. This difference in favor of LE was statistically significant (P ⫽ 0.002). A total of four patients (all in the placebo group) were terminated from further study participation due to an adverse event. At day 2, one patient had tearing, blurred vision, pain, and photophobia, and a second patient had toxic keratitis. At day 8, one patient had corneal edema. At day 11, one patient had elevated IOP. An additional patient in the LE group was hospitalized for observation on study day 11 due to itching of the arms, legs, and trunk; fatigue; and leg cramps. The itching, which was treated with oral hydroxyzine, resolved within 2 days and was judged to be unrelated to the study medication, and treatment was not stopped. Other than these five patients, there were no serious, life-threatening, or sight-threatening events in this elderly, postoperative population during the study period.
Discussion The objective of this study was to evaluate the efficacy and safety of 0.5% LE in controlling the anterior chamber cell and flare reaction in patients undergoing cataract surgery with intraocular lens implantation. The LE was more effective than its vehicle in primary, secondary, and supportive measures of efficacy. In the primary measure of efficacy, the resolution of ACI in the eye that was operated on at final visit, 56 (55%) of 1020 LE patients had achieved resolution compared to 28 (28%) of 100 placebo patients. This 27% difference (treatment effect) in resolution rates was clinically meaningful and statistically significant (P ⬍ 0.001). Expanding the efficacy criterion to include patients with mild inflammation at final visit, the efficacy of LE was 95 (93%) of 102 in contrast to 65 (65%) of 100 for placebo.
1784
Secondary and supportive analyses of the primary efficacy measure were complementary of the primary analysis of ACI. Resolution rate at each visit, cell and flare individually at final visit and each examination, treatment failures, and investigator global assessment were all consistent in showing a clinically and statistically significant treatment effect of LE. Regardless of the treatment group, patients with light irides were more likely to have resolution of ACI than those with dark irides, and the ACIs for patients with phacoemulsification were more likely to resolve than those with extracapsular cataract extraction. The treatment effect of LE relative to its vehicle was greater in patients with more severe baseline ACI ratings than in those with milder ratings (i.e., treatment by baseline interactions). With this one exception (which is consistent with a baseline effect), there was no evidence for a difference in treatment effect across categories (i.e., treatment by factor interaction) for any variable. Thus, the efficacy of LE over its vehicle generally was consistent within all subgroups. As noted previously, both corticosteroids and nonsteroidal anti-inflammatory agents (cyclo-oxygenase inhibitors) have been evaluated for the treatment of postsurgical (cataract) inflammation.2,3,9,10,13 We attempted to evaluate the efficacy of LE observed in this current study in light of some of these previous reports for other agents. Approved in the United States for treatment of postoperative inflammation in recent years are rimexolone, diclofenac, and ketorolac, all on the basis of similarly designed, placebo-controlled, treatment trials (in contrast to prophylactic studies). In a double-masked study of 197 patients, rimexolone was more effective than placebo in suppressing cell and flare. The magnitude of this effect was in the same general range as in the current study.23 The diclofenac study appears to be available only in abstract form with limited data (Vickers FF, McGuigan LJB, Ford C, et al. The effect of diclofenac sodium ophthalmic on the treatment of postoperative inflammation. Invest Ophthalmol Vis Sci Suppl 1991;32:793), and we were unable to find a publication regarding ketorolac for this indication. A host of other studies compared either corticosteroids or NSAIDs to other agents or, in some cases, to a placebo treatment. These studies were predominantly prophylactic in contrast to treatment studies. Efficacy measures included clinical signs of inflammation (cells and flare) as observed by an ophthalmologist and, in some cases, an instrument that measures anterior chamber light properties and translates them to cells and flare. Additional studies used a measure of blood–aqueous barrier function (aqueous fluorophotometry). For the most part, active agents were more effective than their placebo and similar in efficacy to other active agents. The latter comparisons typically were of low power because of sample size and the smaller definition of a clinically significant difference.2,3,8 –11,13,24 Although it is difficult to compare across studies, given the changes in surgical technique over the years, the treatment effect of LE seen in the current vehiclecontrolled study is at least as great as the prophylactic effect in any of the previous reports. We considered using a Kowa Laser Cell Flare meter to evaluate postoperative inflammation.25 However, the logis-
Loteprednol Etabonate Postoperative Inflammation Study Group 2 䡠 LE in Postoperative Inflammation tics of placing such a device at all study centers was substantial. As well, we thought it important to show the anti-inflammatory efficacy of LE in a method used by all surgeons: observations at the slit lamp. With respect to safety, both LE and placebo had favorable safety profiles, including ocular tolerance. There was no evidence for any safety concern with LE regarding treatment-emergent medical events. Furthermore, there was no issue with elevation in IOP in LE-treated patients in this study. Granted that the duration of treatment in this study, 14 days, is short of the traditional period for corticosteroidinduced elevation in IOP.26,27 However, in corticosteroidresponders treated for 6 weeks, LE had a very low incidence IOP elevation relative to prednisolone acetate.28 Furthermore, in other studies in which patients received LE for 4 to 6 weeks, the incidence of elevations in IOP of 10 mmHg or greater was similar to vehicle and lower than prednisolone acetate.29 Thus, LE was both clinically and statistically significantly favored over its vehicle in the treatment of postoperative inflammation. There was no evidence of rebound of signs and symptoms after cessation of therapy. Further, the safety profile of LE was excellent, with little evidence of clinically significant elevations in IOP or deleterious effects on postoperative recovery. Loteprednol etabonate showed a clinically meaningful reduction in the signs and symptoms of postoperative anterior chamber inflammation compared with that of placebo. It also had an acceptable safety profile compared with that of placebo. Acknowledgment. The authors thank Ellen R. Strahlman, MD, MHSc, for her contribution.
Appendix 1 Measurement Scales Cell: Determined using a slit beam at 1-mm height by 1-mm width with maximum luminance of the Haag–Streit (or equivalent) slit lamp. Pigment cells and erythrocytes were to be ignored. 0 ⫽ Less than or equal to 5 cells 1 ⫽ 6 to 10 cells 2 ⫽ 11 to 20 cells 3 ⫽ 21 to 40 cells 4 ⫽ More than 40 cells 5 ⫽ Hypopyon Flare: Determined using the widest slit beam at 1-mm height with maximum luminance of the Haag–Streit (or equivalent) slit lamp. 0 ⫽ None to trace 1 ⫽ Mild (clearly noticeable, visible) 2 ⫽ Moderate (without plastic aqueous) 3 ⫽ Marked (with plastic aqueous) 4 ⫽ Severe (with fibrin deposits or clots or both) Anterior chamber inflammation (ACI): Sum of cell and flare score: 0 ⫽ None to trace
1–2 ⫽ Mild 3– 4 ⫽ Moderate (required at baseline) 5– 6 ⫽ Marked 7– 8 ⫽ Severe 9 ⫽ Hypopyon and severe flare IGA (for entire study) 0 ⫽ Fully controlled: inflammation is suppressed or cured, virtually no signs or symptoms 1 ⫽ Reasonably controlled: inflammation is diminished, moderate decrease in signs or symptoms 2 ⫽ Slightly controlled: Small decrease in signs and symptoms 3 ⫽ Unchanged: No response or overall change in signs and symptoms 4 ⫽ Worse: Overall increase in signs and symptoms Other signs and symptoms (chemosis, erythema, palpebral injection, bulbar injection, corneal edema, pain, photophobia, itching, tearing, dryness, discharge, and discomfort): 0 ⫽ Absent 1 ⫽ Mild 2 ⫽ Moderate 3 ⫽ Severe Hyphema and ciliary flush were scored as present (1) or absent (0).
Appendix 2 The Loteprednol Etabonate Postoperative Inflammation Study Group 2 is as follows: Investigators Cecil Beehler, MD, Ft. Myers, FL; Bruce Bodner, MD, Norfolk, VA; Bradley Bowman, MD, Dallas, TX; David Cooke, MD, St. Joseph, MI; J. Luther Crabb, MD, Memphis, TN; L. Raymond DeBarge, MD, Fort Oglethorpe, GA; Eric Donnenfeld, MD, Rockville Centre, NY; Marcel Estopinal, MD, Nashville, TN; Kenneth Fox, MD, Fredericksburg, VA; Peter Kastl, MD, PhD, New Orleans, LA; Kenneth Olander, MD, PhD, Milwaukee, WI; Elizabeth Sharpe, MD, Mt. Pleasant, SC; Robert Shofner, MD, Nashville, TN; Janis Stahl, MD, Golden, CO; Dara Stevenson, MD, New Orleans, LA; William Stewart, MD, Charleston, SC; Tom Walters, MD, Austin, TX. Study Coordination Sue Gill, John Howes, PhD, Cathy Lawson, Michael Smerick, Jr., BS (Pharmos Corp, Alachau, FL), David Zaccardelli, PharmD, Roselyn Fazio, BS (Bausch and Lomb Pharmaceuticals, Tampa, FL), and Gary D. Novack, PhD (Pharmacologic Development, Inc, San Rafael, CA). Biostatistics and Data Management R. Stephens Crockett, PhD, and Kathryn Hart, MA (D.A.T.A., Inc, Murray, KY).
References 1. Corboy JM. Corticosteroid therapy for the reduction of postoperative inflammation after cataract extraction. Am J Ophthalmol 1976;82:923–7.
1785
Ophthalmology Volume 105, Number 9, September 1998 2. Flach AJ, Kraff MC, Sanders DR, Tanenbaum L. The quantitative effect of 0.5% ketorolac tromethamine solution and 0.1% dexamethasone sodium phosphate solution on postsurgical blood–aqueous barrier. Arch Ophthalmol 1988;106: 480 –3. 3. Othenin–Girard P, Tritten JJ, Pittet N, Herbort CP. Dexamethasone versus diclofenac sodium eyedrops to treat inflammation after cataract surgery. J Cataract Refract Surg 1994;20: 9 –12. 4. Ramsell TG, Bartholomew RS, Walker SR. Clinical evaluation of clobetasone butyrate: a comparative study of its effects in postoperative inflammation and on intraocular pressure. Br J Ophthalmol 1980;64:43–5. 5. Araie M, Sawa M, Takase M. Topical flurbiprofen and diclofenac suppress blood–aqueous barrier breakdown in cataract surgery: a flurophotometric study. Jpn J Cln Ophthalmol 1983;27:535– 42. 6. Bizzotto MF, De Franco I, Carillio F. Efficacy of flurbiprofen versus placebo on postoperative course following cataract extraction. Drugs Exp Clin Res 1984;10:421–5. 7. Blaydes JE Jr, Kelley EP, Walt JG, et al. Flurbiprofen 0.03% for the control of inflammation following cataract extraction by phacoemulsification. J Cataract Refract Surg 1993;19: 481–7. 8. Flach AJ, Jaffe NS, Akers WA. The effect of ketorolac tromethamine in reducing postoperative inflammation: double-mask parallel comparison with dexamethasone. Ann Ophthalmol 1989;21:407–11. 9. Flach AJ, Lavelle CJ, Olander KW, et al. The effect of ketorolac tromethamine solution 0.5% in reducing postoperative inflammation after cataract extraction and intraocular lens implantation. Ophthalmology 1988;95:1279 – 84. 10. Roberts CW. Pretreatment with topical diclofenac sodium to decrease postoperative inflammation. Ophthalmology 1996; 103:636 –9. 11. Sabiston D, Tessler H, Sumers K, et al. Reduction of inflammation following cataract surgery by the non-steroidal antiinflammatory drug, flurbiprofen. Ophthalmic Surg 1987;18: 873–7. 12. Sabiston DW, Robinson IG. An evaluation of the anti-inflammatory effect of flurbiprofen after cataract extraction. Br J Ophthalmol 1987;71:418 –21. 13. Ostrov CS, Sirkin SR, Deutsch WE, et al. Ketorolac, prednisolone, and dexamethasone for postoperative inflammation. Clin Ther 1997;19:259 –72. 14. Yablonski ME, Burde RM, Kolker AE, Becker B. Cataracts induced by topical dexamethasone in diabetics. Arch Ophthalmol 1978;96:474 – 6. 15. Bodor N. Novel approaches to the design of safer drugs: soft drugs and site-specific chemical delivery systems. Advances in Drug Research 1984;13:255–331.
1786
16. Bodor N. Soft drugs: principles and methods for the design of safe drugs. Med Res Rev 1984;4:449 – 69. 17. Druzgala P, Wu WM, Bodor N. Ocular absorption and distribution of loteprednol etabonate, a soft steroid, in rabbit eyes. Curr Eye Res 1991;10:933–7. 18. Asbell P, Howes J. A double-masked, placebo-controlled evaluation of the efficacy and safety of loteprednol etabonate in the treatment of giant papillary conjunctivitis. CLAO J 1997;23:31– 6. 19. Friedlaender MH, Howes J, for The Loteprednol Etabonate Giant Papillary Conjunctivitis Study Group I. A doublemasked, placebo-controlled evaluation of the efficacy and safety of loteprednol etabonate in the treatment of giant papillary conjunctivitis. Am J Ophthalmol 1997;123:455– 64. 20. Dell SJ, Shulman DG, Lowry GM, et al. A controlled evaluation of the efficacy and safety of loteprednol etabonate in the prophylactic treatment of seasonal allergic conjuctivitis. Am J Ophthalmol 1997;123:791–7. 21. Neumann R, Howes J. Loteprednol etabonate. A novel ocular steroid with improved safety profile. In: Nussenblatt RB, Whitcup SM, Caspi RM, Gery I, eds. Advances in Ocular Immunology: Proceedings of the 6th International Symposium on the Immunology and Immunopathology of the Eye. Bethesda, MD; Amsterdam; New York: Elsevier, 1994 (Int’l Congress Series no. 1068). 22. Stewart RH, Smith JP, Rosenthal AL. Ocular pressure response to fluorometholone acetate and dexamethasone sodium phosphate. Curr Eye Res 1984;3:835–9. 23. Assil KK, Massry G, Lehmann R, et al. Control of ocular inflammation after cataract extraction with rimexolone 1% ophthalmic suspension. J Cataract Refract Surg 1997;23: 750 –7. 24. Kraff MC, Sanders DR, McGuigan L, Raanan MG. Inhibition of blood-aqueous humor barrier breakdown with diclofenac. A fluorophotometric study. Arch Ophthalmol 1990;108:380 –3. 25. Drews RC. Management of postoperative inflammation: dexamethasone versus flurbiprofen, a quantitative study using the new Flare Cell Meter. Ophthalmic Surg 1990;21:560 –2. 26. Armaly MF. Statistical attributes of the steroid hypertensive response in the clinically normal eye. I. The demonstration of three levels of response. Invest Ophthalmol 1965;4:187–97. 27. Becker B. Intraocular pressure response to topical corticosteroids. Invest Ophthalmol 1965;4:198 –205. 28. Bartlett JD, Horwitz B, Laibovitz R, Howes JF. Intraocular pressure response to loteprednol etabonate in known steroid responders. J Ocul Pharmacol 1993;9:157– 65. 29. Novack GD, Howes J, Crockett RS, Sherwood MB. Change in intraocular pressure during chronic use of loteprednol etabonate. J Glaucoma, in press.