A representative example of each of grades at the late phase (15 minutes after intravenous injection of 10% fluorescein sodium) of fluorescein angiograms. Rating of fluorescein angiography is explained in the "Patients and Methods" section. Reproduced with permission from Miyake et al.11
Miyake K, Ota I, Ibaraki N, Akura J, Ichihashi S, Shibuya Y, Maekubo K, Miyake S. Enhanced Disruption of the Blood-Aqueous Barrier and the Incidence of Angiographic Cystoid Macular Edema by Topical Timolol and Its Preservative in Early Postoperative Pseudophakia. Arch Ophthalmol. 2001;119(3):387-394. doi:10.1001/archopht.119.3.387
To investigate the effects of timolol maleate with preservative and its preserved (PV) and nonpreserved vehicles (NPV) (benzalkonium chloride) on the blood-aqueous barrier and angiographic cystoid macular edema (CME) in early postoperative pseudophakia.
Patients and Methods
Patients with ocular hypertension, normal tension glaucoma, and primary open-angle glaucoma who underwent surgery for cataracts. The study included a double-masked trial for timolol, PV, and NPV and a single-masked trial on the effect of diclofenac sodium and fluorometholone acetate on all three. The patients were divided into 6 groups, each of which were simultaneously administered the following different combinations of compounds: timolol and diclofenac (group A), timolol and fluorometholone (group B), PV and diclofenac(group C), PV and fluorometholone (group D), NPV and diclofenac (group E), and NPV and fluorometholone (group F). The 6 groups were then compared using a laser flare cell meter to determine the degree of disruption of the blood-aqueous barrier and fluorescein angiography to investigate angiographic CME. The differences in mean daily fluctuations in intraocular pressure were compared on the preoperative baseline day and for 5 weeks postoperatively. Twice daily administration of 0.5% timolol maleate or the vehicles was started 2 days before surgery, and continued until 5 weeks after surgery. Diclofenac or fluorometholone drops were instilled in the eyes 4 times preoperatively, on the day of surgery, and 3 times daily for 5 weeks postoperatively.
The flare amount was higher on the third and seventh days in group B than in group D, but was the same after the seventh day. The incidence of angiographic CME was the same between both groups. These 2 factors were significantly lower in group F. These 2 factors were also significantly lower in the 3 groups that received diclofenac instead of fluorometholone, with no difference among these groups. The intraocular pressure decline was significant in groups that received timolol compared with groups that received PV or NPV.
Timolol and its preservative, benzalkonium chloride, cause disruption of the blood-aqueous barrier in early postoperative pseudophakia and increased incidence of angiographic CME. The concurrent administration of nonsteroidal anti-inflammatory drug such as diclofenac prevents these adverse effects without interfering with the drop in intraocular pressure caused by timolol. The addition of benzalkonium chloride to timolol contributes considerably to these adverse effects.
The present results suggest the cause of similar complications produced by other antiglaucoma eyedrops containing similar preservatives.
THE REPORTS have indicated that, especially with their long-term administration, topical antiglaucoma drugs cause inflammatory changes in the ocular surface, ie, conjunctiva or Tenon capsule,1- 5 and that their intraocular adverse effects, mainly in aphakias, include angiographic cystoid macular edema (CME) or inflammatory disruption of the blood-aqueous barrier.6- 12 The recently developed drug latanoprost also has been reported to have these intraocular adverse effects in pseudophakia.9- 12 It remains unclear whether the cause is the main constituents of the drugs or the preservative.
The most frequently used antiglaucoma drug thus far is timolol maleate, use of which has shown a vast improvement of the ocular and systemic adverse effects seen with the initial use of β-blockers.13- 15 Timolol is a superior drug that does not affect pupil diameter or accommodation, as did earlier antiglaucoma drugs,16 and it has almost no adverse effects, such as burning or conjunctival hyperemia, with short-term use. However, there have been a few reports of ocular toxic effects attributed to long-term continuous use of timolol. These include ocular cicatricial pemphigoid,17,18 conjunctival hyperemia and burning accompanying superficial punctate keratitis, decreased corneal sensation, and damage to the tear-film mucus layer.19- 24
All antiglaucoma drugs contain preservatives. Timolol contains benzalkonium chloride, the preservative most widely used in eyedrops. It is not known whether the adverse effects of timolol are due to the agent, the preservative, or both. Benzalkonium chloride has toxic effects on the cornea and conjunctiva surface,25,26 and the main focus of in vivo and in vitro research has been on the cytotoxic effects to the corneal tissues.27- 30 Little research has been performed to compare the effects of timolol with and without added preservatives. The results of the few experimental studies that have been performed suggest that the preservative contributes more to the toxic effects or inflammatory changes than does the timolol,5,31 but this theory has not been investigated in humans.
Other antiglaucoma drugs also contain preservatives, and inflammatory changes, including the disruption of the blood-aqueous barrier and onset of CME, have been reported.6- 12 The cause of the complications with such drugs, however, remains unknown, in particular whether they are due to the drug or the preservative.
To investigate the effect of timolol and its preservative on the disruption of the blood-aqueous barrier as well as on the incidence of angiographic CME after removal of cataracts and implantation of intraocular lenses, timolol eyedrops, the vehicle containing benzalkonium chloride (PV), and vehicle not containing benzalkonium chloride (NPV) were administered as therapy for early postoperative pseudophakia after cataract surgery. The effects of concurrently administered steroidal and nonsteroidal eyedrops were also investigated. The blood-aqueous barrier function and angiographic CME were selected because they are the most sensitive clinical variables for detecting intraocular inflammatory changes.
Patients with ocular hypertension, normal tension glaucoma, and primary open-angle glaucoma participated in this study to investigate the effects of (preserved) timolol, PV, and NPV on the disruption of the blood-aqueous barrier and angiographic CME. The following 3 solutions were used in the study in a double-masked control trial: 0.5% timolol eyedrops (Timoptol; Banyu Pharmaceutical Co, Ltd, Osaka, Japan), PV containing benzalkonium chloride (0.1 mg) in 1 mL of sterile phosphate-buffered solution, and the same solution without benzalkonium chloride (NPV). The 2 vehicles are otherwise identical in regard to pH and osmolality. The eyedrops of the 2 vehicles used as controls were prepared specially in our hospital pharmacy. The effects of simultaneously administered 0.5% diclofenac sodium nonsteroidal anti-inflammatory eyedrops (Diclode; Wakamoto Pharmaceutical Co, Ltd, Tokyo, Japan) and 0.1% fluorometholone acetate steroidal anti-inflammatory eyedrops (Flumetholone; Santen Pharmaceutical Co, Ltd, Osaka) were also investigated. The preservative for diclofenac is chlorobutanol, and that for fluorometholone is benzalkonium chloride (both at dilutions of 0.1 mg/1 mL). Due to the nature of the present study, it would have been better to use the same preservative for both anti-inflammatory agents. Chlorobutanol, however, is also not used as a preservative for steroidal eyedrops such as betamethasone sodium phosphate and dexamethasone sodium phosphate. Moreover, diclofenac is the most effective nonsteroidal anti-inflammatory eyedrop in Japan for preventing inflammation after implant surgery. For these reasons, anti-inflammatory agents with different preservatives were used, and due consideration must be given to the above when interpreting the results of the present study. Although the relative strength of fluorometholone is controversial, fluorometholone has the same anti-inflammatory effect as betamethasone and dexamethasone,32- 35 and was chosen because it does not easily produce steroidal glaucoma. However, because fluorometholone is a milky white substance, a double-masked trial was impossble in this part of the study, and we therefore settled on a single-masked trial.
In all eyes, locally administered antibiotics were injection of gentamicin sulfate (Gentacin; Schering Plough Pharmaceutical Co, Ltd, Osaka) below the Tenon capsule and ofloxacin eyedrops (Tarivid, Santen Pharmaceutical Co, Ltd), which were not preserved. Systemically given antibiotics were also the same in all eyes, which included intramuscular injection of isepamicin sulfate(Isepacin; Schering Plough Pharmaceutical Co, Ltd) and oral capsule of cefdimir(Cefzon; Fujizawa Pharmaceutical Co, Ltd, Osaka).
A total of 180 eyes (180 patients) initially were used in the study, with 30 each randomly assigned to the following 6 groups: timolol and diclofenac(group A), timolol and fluorometholone (group B), PV and diclofenac (group C), PV and fluorometholone (group D), NPV and diclofenac (group E), and NPV and fluorometholone (group F). The inclusion criteria for these 180 eyes were that the patient be 40 years or older and have cataract and ocular hypertension, normal tension glaucoma, or primary open-angle glaucoma. In the 5-week follow-up, only 1 eye per patient was considered in the study. In addition, only eyes in which disease had not advanced or eyes without progressive field loss during preoperative examination are included in the study. The following patients were excluded: those whose pupillary diameter was less than 4 mm when mydriasis occurred due to the cataract surgery; those hypersensitive to diclofenac, fluorometholone, timolol, or fluorescein sodium; those who had undergone previous eye surgeries; and those with ocular diseases other than glaucoma or cataracts or any systemic diseases and medications.
This study was approved by the institutional review board at Miyake Eye Hospital, Nagoya, Japan, and conducted in accordance with the Helsinki Declaration. Informed consent was obtained in writing from patients participating in the study after they were given full explanations of the nature of the study and the fluorescein fundus angiography used.
The surgery in all cases was performed through a 3-mm clear corneal incision, with no sutures or, when necessary, 1 stitch. Following continuous curvilinear capsulorrhexis and phacoemulsification, a foldable acrylic intraocular lens (Acrylsof; Alcon Laboratory, Fort Worth, Tex) was implanted into the lens capsule. All surgeries were performed by two of us (K.M. or I.O.). Preoperative glaucoma treatment was discontinued 3 weeks before the start of the study to wash out the effects of the antiglaucoma medications used. Timolol or vehicle was given 2 days before the surgery, and administration was continued twice daily for 5 weeks postoperatively. In addition, fluorometholone acetate or diclofenac sodium was given 4 times on the day of surgery (7, 3, and 2 hours and 30 minutes before surgery), and then 3 times per day for 5 weeks after surgery. Concurrent drugs included systemically or locally administered antibiotics.
Intramuscular injection of isepamicin was performed once on the day of surgery and continued for 5 days postoperatively. Three oral capsules of cefdimir were given on the day of surgery and for 4 days postoperatively. Injection of gentamicin below the Tenon capsule was performed just after the operation. Ofloxacin eyedrops were given twice a day for 2 days preoperatively and 3 times a day for 30 days postoperatively.
Clinical examinations were performed and observations were made concerning patients' background, surgical procedures, visual acuity, intraocular pressure, amount of anterior chamber flare measured by a laser flare cell meter, and presence of CME measured by means of fluorescein angiography.
Visual acuity was measured at 1 to 3 days before surgery and on postoperative days 1, 3, 7, 14, and 35. Baseline intraocular pressure was measured 3 days before surgery 4 times, at 8 AM, noon, and 4 and 8 PM. The mean fluctuations in intraocular pressure throughout the day were used for evaluation and compared with the mean values obtained in the same way at 5 weeks postoperatively. The degree of disruption of the blood-aqueous barrier was measured using a laser flare cell meter (FC1000; Kowa Co, Ltd, Tokyo) 4 and 6 days before surgery and 1 and 3 days and 1, 2, and 5 weeks after surgery.
After measurement of intraocular pressure, fluorescein angiography was performed to measure angiographic CME formation at 5 weeks postoperatively. The late phase (15 minutes after intravenous injection of 10% fluorescein sodium) of fluorescein angiograms was graded by one of us (S.M.) using the method previously mentioned36 in a double-masked manner. Briefly, 0° means there is no sign of fluorescein leakage; I°, slight fluorescein leakage into the cystic space but not sufficient enough to enclose the entire fovea centralis; II°, complete circular accumulation of the fluorescein in the cystic space but with a diameter of less than 2.0 mm; and III°, circular accumulation of the fluorescein larger than 2.0 mm in diameter. Figure 1 shows a representative example of each grade.
The age, surgical data, visual acuity, amount of aqueous flare, and intraocular pressure of the patients in the 6 groups were analyzed using a 1-way analysis of variance. When differences were found among any of these variables, we used the Tukey test for multiple comparison to determine the site of difference. The degree of the drop in intraocular pressure in each group was analyzed using a paired t test. Sex, type of glaucoma, antiglaucoma medications, and family history of glaucoma were analyzed using the χ2 method. The incidence of angiographic CME was analyzed using the Fisher exact test. In all tests, the level of significance was taken to be P<.05. The data are recorded as mean ± SD, unless otherwise indicated.
There was a posterior capsule rupture in 1 patient in group A, and a failure of the continuous curvilinear capsulorrhexis in 1 patient in group E. Several patients were not available for follow-up examinations of the fluorescein fundus angiography because of health or commitments. Analysis was conducted with the remaining total of 168 eyes, 28 in group A, 29 in group B, 29 in group C, 28 in group D, 27 in group E, and 27 in group F. Among these patients were some who could not be present for all laser flare cell meter and visual acuity tests because of health or commitments. There was no significant difference in the incidence of patients being lost to follow-up or dropped from the study among the 6 groups.
There were no significant differences between patients in the 6 groups in age, sex, type of glaucoma, medication history, or family history (Table 1). No differences were found among groups in operation time, nuclear hardness (evaluated by a method described previously37), ultrasound time, or amount of intraocular irrigating solution (Table 2). No difference in the change in postoperative visual acuity was seen (Table 3).
Before surgery, no significant difference was seen in intraocular pressure among the 6 groups. In a comparison between the baseline of preoperative diurnal intraocular pressure and intraocular pressure at 5 weeks, there was a significant drop at 5 weeks in patients of all 6 groups (P<.001 in groups A and B; P<.01 in groups C-F). Moreover, at 5 weeks there was a significantly larger decrease in intraocular pressure in groups A and B, which were administered timolol, compared with groups C, D, E, and F, in which both types of vehicle were used (P<.05). However, at 5 weeks, there were no significant differences in intraocular pressure between groups A and B or between groups C, D, E, and F (Table 4).
The incidence of angiographic CME 5 weeks after surgery in the groups that received fluorometholone eyedrops (groups B, D, and F), was significantly higher (P<.01) than in the groups that received diclofenac (groups A, C, and E) (Table 5). The incidence was also significantly higher (P<.01) in groups B and D than in group F. However, no significant differences were noted between groups A, C, and E, all receiving diclofenac. There was also no significant difference in incidence between groups B and D.
No differences in aqueous flare were found among any of the 6 groups before surgery or 1 day after surgery (Table 6). Group B had a significantly higher amount of aqueous flare than group A on postoperative day 3 and at postoperative weeks 1, 2, and 5. Similarly, groups D and F had significantly greater amounts of aqueous flare on postoperative day 3 and postoperative weeks 1 and 2 than groups C and E(P<.05 in all cases). Throughout the study period, there was no significant difference in amount of aqueous flare between group A and groups receiving vehicle with diclofenac (groups C and E). However, among groups receiving fluorometholone, the amount of flare was significantly higher in group B than in group D on postoperative day 3 and 1 week (P<.05). Similarly, group B had a significantly higher amount of flare than group F on postoperative day 3 and weeks 1 and 2 (P<.05).
Finally, in comparing groups receiving vehicle and fluorometholone, group D had a significantly higher flare value than group F at postoperative weeks 1 and 2 (P<.05).
In summary, the results showed that the incidence of CME was the same in the eyes treated with benzalkonium chloride–preserved timolol and fluorometholone and in the eyes treated with preserved fluorometholone and PV. Flare was higher on the third and seventh days in the former groups, but afterward was the same between both groups. These values for NPV and preserved fluorometholone were significantly lower. If preserved timolol, PV, NPV, and chlorobutanol-preserved diclofenac are used in place of preserved fluorometholone, there is no significant difference among the 3 groups in CME and flare. There was a greater incidence of CME and more flare in the 3 groups receiving fluorometholone than in any of the other 3 groups receiving diclofenac.
In the present study, we found that benzalkonium chloride–preserved timolol eyedrops increase disruption of the blood-aqueous barrier in early postoperative pseudophakia and incidence of CME. The PV of timolol had the same actions. The difference in the effects of timolol and its PV was that the latter caused somewhat less disruption to the blood-aqueous barrier, although their effects on CME incidence were the same. These findings suggest that the adverse effects of timolol are due, at least in considerable part, to the preservative used. We also found that the concurrent administration of nonsteroidal anti-inflammatory eyedrops (diclofenac) can prevent these effects without adversely affecting the drop in intraocular pressure caused by timolol.
To our knowledge, our present findings are the first to suggest that benzalkonium chloride might affect the blood-aqueous barrier and onset of CME in early postoperative pseudophakic eyes. To the best of our knowledge, there has been no previous report describing preservatives other than benzalkonium chloride or antiglaucoma medications with such preservatives that produce these kinds of complications.
The following 2 points are new knowledge and of considerable significance. First, although numerous reports have described the adverse effects induced by timolol, including inflammatory changes on the ocular surface such as in the cornea, conjunctiva, and tear film,1- 5,17- 24 there are very few studies on intraocular lesions due to timolol; to our knowledge only a single report5 describes the chronic inflammatory changes of trabeculae induced by the long-term (mean, 5 years) administration of antiglaucoma eyedrops, including timolol. This is the first report on the acute effect. Second, although experimental evidence suggests that the preservative of timolol participates in these adverse effects,5,31 this is the first report of a clinical investigation in humans. We found that although the incidence of CME is the same with administration of preserved timolol and PV, there is a slight difference in disruption to the blood-aqueous barrier. It has been suggested from animal experiments as well that the adverse effects of timolol are rather small.5 Baudouin and associates5 reported no difference in the subconjunctiva infiltration of inflammatory cells in rats between preserved timolol and PV. Furthermore, using a tissue culture of human Tenon capsule fibroblasts, Williams and associates31 suggested that, rather than the direct stimulation of cell proliferation by β-blockers, other glaucoma medications and their preservatives cause cell proliferation due to chronic inflammation. If preservatives play a major role in the onset of early postoperative inflammatory complications in pseudophakic eyes, the mechanism for this effect needs to be established.
Recent in vivo and in vitro studies have shown that various chemical mediators are biosynthesized during the process of proliferation and pseudometaplasia of lens epithelial cells that have been damaged in surgery.38,39 This process is closely linked to the natural course of early postoperative inflammatory reactions such as pupillary fibrin formation.40,41 The synthesis of chemical mediators by lens epithelial cells also explains the significant elevation in the amount of aqueous flare 1 to 2 weeks after surgery compared with immediately afterward.38- 40 In the present study, it was found that the amount of aqueous flare was the same on postoperative day 1 in all 6 groups receiving different drugs. However, an increase was seen on postoperative day 3 and at 1 and 2 weeks in the groups given timolol, PV, or NPV with fluorometholone (groups B, D, and F). Such an elevation of aqueous flare is enhanced by timolol and its PV.
Long-term administration of timolol eyedrops in humans reportedly induces the invasion of inflammatory cells below the conjunctiva and Tenon capsule, as well as inflammatory lesions that can be detected immunohistochemically.5 Any of these effects may be present in healthy experimental animals or humans, and they take a considerable time.5 However, because the subjects in the present study already had postoperative inflammation, one might argue that the adverse effects and inflammatory response would appear sooner. It therefore appears that biosynthesis of inflammatory mediators is accelerated as a result of the encouragement of lens epithelial cell proliferation and pseudometaplasia by timolol and its PV, and thus the disruption of the blood-aqueous barrier is intensified 1 or 2 weeks after surgery. In other words, timolol and its PV are not direct inflammatory mediators. If these substances contributed directly to the disruption of the blood-aqueous barrier, the breakdown should have been greater in the eyes administered timolol or PV 1 day after surgery. Moreover, concurrent administration of a nonsteroidal anti-inflammatory agent with timolol or its PV suppressed disruption of the blood-aqueous barrier to the same extent as administration of a nonsteroidal anti-inflammatory agent with the NPV. This suggests that prostaglandin E or other substances synthesized during the process of proliferation and pseudometaplasia of lens epithelial cells, which is accelerated by timolol or its PV, act as a direct mediator. This corresponds well to the hypothesis that the appearance of postoperative inflammation and CME is related to prostaglandin E2 and cytokines biosynthesized due to surgical stimulation.36,38- 44
Administration of glaucoma drugs such as epinephrine bitartrate or latanoprost into postoperative aphakic or pseudophakic eyes is known to cause disruption to the blood-aqueous barrier and CME.6- 9,11,12 The present findings have suggested a mechanism for these phenomena, ie, the same preservative, benzalkonium chloride, is used in epinephrine and latanoprost. It therefore seems plausible that, similar to the present findings, the preservative rather than the agent in these medications is the main factor triggering the complications. Miyake and associates11 have found recently, in a study comparing latanoprost and its NPV, that the use of latanoprost increases the incidence of CME and the disruption of the blood-aqueous barrier in early postoperative pseudophakic eyes. Benzalkonium chloride is also used as a preservative in latanoprost, indicating the possibility that the contribution of the preservative is larger than that of a substituted analogue of prostaglandin F2α in the onset of complications. It has been reported recently that latanoprost increases the recurrence and severity of uveitis and herpetic keratitis.10,45 If the present findings indeed show that the preservatives in antiglaucoma drugs are responsible in large part for the onset of inflammation, then the suitability of timolol and other antiglaucoma drugs for patients with uveitis and herpetic keratitis will have to be reconsidered seriously.
It was found that the inflammatory complications that appear with administration of timolol in patients with early postoperative pseudophakia can be suppressed with the concurrent administration of a nonsteroidal anti-inflammatory agent. This has also been shown by Miyake and associates11 to be true with latanoprost. Although this is clinically significant, the level of postoperative concurrent administration and the critical period for administration require further investigation. In the 6 groups compared in this study, no significant difference was found in the changes in postoperative visual acuity, even when there was angiographic CME. This would suggest that, clinically, the CME that appeared was not severe enough to affect visual acuity. The present results, however, indicate that the long-term administration of timolol in pseudophakic eyes should be undertaken cautiously with patients who seem to be predisposed to blood-retinal barrier breakdown. In cases in which timolol must be given, the concurrent administration of nonsteroidal anti-inflammatory agent is recommended.
Since preservatives added to the agents and the vehicle in the present study were different, care must be exercised in interpretation. Benzalkonium chloride at 0.1 mg/1 mL was used for timolol, fluorometholone, and PV. A comparison of the incidence of CME and the amount of anterior chamber flare in the 3 groups receiving preserved fluorometholone (groups B, D, and F), revealed a higher value in groups B and D than in group F, but little difference between groups B and D. These findings suggest that incidence of CME and the amount of anterior chamber flare increases with the amount of benzalkonium chloride used, and that the effect of timolol is relatively mild.
In the present study, we have also indicated that the nonsteroidal anti-inflammatory agent diclofenac suppressed CME and flare in all groups, but the results are difficult to interpret because chlorobutanol is used as the preservative for diclofenac, whereas benzalkonium chloride is used for fluorometholone. The toxic effects of benzalkonium chloride and chlorobutanol against corneal epithelial cells have been studied in animal experiments, and it has been reported that the toxic effects of benzalkonium chloride are stronger.46- 49 Studies have been performed on the inflammatory effects of benzalkonium chloride,5 but no comparisons have been made between it and chlorobutanol. The present results can thus be interpreted in 2 ways. One is that diclofenac has a stronger anti-inflammatory effect than fluorometholone; the other is that, of the 2 preservatives used, the inflammatory effect of chlorobutanol was smaller than that of benzalkonium chloride. This defect in the design of the present study means that, in the future, it will be necessary to compare anti-inflammatory agents to which the same preservative has been added.
Accepted for publication August 22, 2000.
Corresponding author and reprints: Kensaku Miyake, MD, Shohzankai Medical Foundation, Miyake Eye Hospital, 1070-Kami 5, Higashiozone-cho, Kita-ku, Nagoya, 462-0823, Japan (e-mail: email@example.com).