Effect of Timolol versus Pilocarpine on Visual Field Progression in Patients with Primary Open,angle Glaucoma Roger Vogel, MD, DO, MCOphth,1 Ronald Pitts Crick, FRCS, FCOphth, DOMS,2 K. Barry Mills, BSc, MBChB, FRCS Ed, DO, FCOphth,3 Patricia M. Reynolds, MB, BS, MRCS, LRCP, DO, MCOphth,2 William Sass, MD,4 Coleen M. Clineschmidt, BA,1 Robert Tipping, MS I Background: Relatively few studies have been conducted linking decreasing intraocular pressure (lOP) to preservation of visual field. This investigation was conducted to determine if this link could be made and to compare the long-term effect of two ocular hypotensive agents on preservation of visual field. Methods: In an observer-masked study, 189 patients with primary open-angle glaucoma received either timolol or pilocarpine by random allocation. The dose of anti glaucoma agent was increased from 0.25% to 0.5% twice daily for timolol or from 2% to 4% four times daily for pilocarpine if the initial lOP response was inadequate. After an on-treatment baseline, visual fields were followed every 4 months for 2 years using the Octopus program 32. Results: Compared with timolol, Significantly more patients receiving pilocarpine discontinued use because of inadequate lOP control (P :s;; 0.01). By comparing the mean visual field scores, it can be seen that the pilocarpine group had a significantly worse score at all timepoints from month 4 to month 24. The pilocarpine group also had a greater mean number of test loci with decreased sensitivity of 5 or more decibels (dB) at all timepoints. The mean within-patient regression slope for timolol was 0.01 dB/ month and for pilocarpine was -0.06 dB/month (P < 0.01). The study has shown that over a 2-year period, patients treated with pilocarpine 2% or 4% four times daily experienced a significantly greater visual field deterioration than that seen in patients receiving either 0.25% or 0.5% timolol twice daily. Conclusion: Although these data do not support a link between lowering of lOP and visual field preservation, treatment with timolol was associated with significantly less visual field loss than treatment with pilocarpine. Ophthalmology 1992;99: 1505-1511
Originally received: July 16, 1991. Revision accepted: May 21, 1992. I Clinical Research Ophthalmology, Merck Sharp and Dohme Research Laboratory, West Point, PA. 2 Kings College Hospital, Denmark Hill, London. 3 Manchester Eye Hospital, Manchester. • Niagara Falls Memorial Medical C enter, Niagara Falls. The authors, other than those who work for Merck Sharp and Dohme Research Laboratory, have no proprietary interest in the development or marketing of this drug. Reprint requests to Roger Vogel, MD, MCOphth, Clinical Research Ophthalmology, Merck Sharp and Dohme Research Laboratory, West Point, PA 19486.
In 1978, pilocarpine was the most widely prescribed therapeutic agent for the treatment of glaucoma. However, although effective in lowering intraocular pressure (lOP), it was a parasympathomimetic agent and consequently produced miosis and spasm of accommodation. Timolol reduced lOP without the side effects of miosis and accommodation spasm and was regarded as a major improvement in glaucoma therapy. A nonselective !3-blocker, timolol slightly reduced pulse rate and blood pressure.) Langham 2 concluded that these systemic side effects of timolol might cause a reduction of ocular perfusion pressure and ocular blood flow with possible deleterious effects on the visual field.
1505
Ophthalmology
Volume 99, Number 10, October 1992
When timolol was introduced in 1979, little evidence existed to support the traditional view that pilocarpine therapy prevented the progression of visual field loss in primary open-angle glaucoma. Some questioned whether lowering lOP as the prime objective of therapy was justified by the available evidence. 3 There was concern, however, whether timolol could provide visual field protection equal to that believed to be provided by pilocarpine in view of the possible vascular effects oftimolol. Conversely, it was concluded that there might be some advantage with timolol because its topical side effects were fewer than those associated with pilocarpine use, and it seemed highly likely that compliance would be better. To clarify the position, a prospective, masked, controlled study was conceived using the Octopus 201 Perimeter (Interzeag, Schlieren, Switzerland). The objective was to compare the lOP-lowering capability, the tolerability, and visual field conservation of topical timolol and pilocarpine at the concentrations usually prescribed.
Materials and Methods Seven investigators entered a total of 189 patients with primary open-angle glaucoma into this study. Drug supplies were not masked but were issued by a secondary investigator according to a random schedule. Patients were recruited as they presented and qualified for entry into the study. One center originally participating did not sequentially assign allocation numbers as patients qualified and data from these patients do not appear in the tables. Elimination of these data did not alter the results or conclusions of the study. After a pretreatment washout period of 7 days, measurements of visual field and lOP, a slit-lamp examination, symptoms, gonioscopy, ophthalmoscopy, visual acuity, and pulse and blood pressure were performed. With the exception of gonioscopy and ophthalmoscopy (which were repeated at the end of the study), all parameters mentioned were followed at each visit throughout the study. Admission criteria included: (1) lOP of 22 mmHg or greater in 1 or both eyes on at least 1 of 5 measurements taken (approximately 9:00, 10:30, 12:00, 14:30, and 16: 30 hours) on the same day after a washout period of at least 7 days taking no glaucoma therapy; (2) open anterior .chamber angles; and, (3) a visual field defect recorded by "the Octopus Program 32, which showed a depression of three or more contiguous test points greater than 5 dB below "normal" values for the patient's age as determined by the Octopus or greater than 5 decibels (dB) below adjacent contiguous points. All visual field tests were performed at approximately the same time during the day and at approximately the same time in relation to the instillation of assigned study drug. Patients were excluded if they had: (1) a history of severe ocular trauma or intraocular surgery; (2) a corneal ulcer, ocular infection, or herpetic keratitis within 3 months of the study start; (3) a history of angle-closure or secondary glaucoma; (4) bronchial asthma or chronic obstructive pulmonary disease; greater than first degree heart block, uncompensated heart failure, or bradycardia
1506
of significant degree; (5) any disease other than open-angle glaucoma producing visual field loss; (6) concomitant medications known to affect lOP; (7) pregnant or nursing women or women of childbearing potential not using adequate means of contraception. All visual fields were performed on an Octopus 201 perimeter using programs 32 (at baseline, dose adjustment, and every 4 months during the study) and 42 (at dose adjustment and at 24 months). Patients were started on either 0.25% timolol twice daily or 2% pilocarpine four times daily according to the random allocation schedule. Patients entered a dose adjustment period, lOP was measured after 2 weeks, and the concentration of the drug was increased (to either timolol 0.5% or pilocarpine 4%) iflOP lowering was insufficient (i.e., if the lOP was above 22 mmHg). If the drug concentration was increased, the patient had a further lOP examination 2 weeks later. The visual field obtained at the end of this dose adjustment period was used as the on-treatment baseline to take account of any learning effect and any decrease in sensitivity as a result of the miosis caused by pilocarpine. Patients were examined every 4 months throughout the 2-year study. If a patient was receiving the high concentration of test medication, he or she was withdrawn from the study if his visual field defect dramatically increased or his lOP became uncontrolled (i.e., rose above 25 mmHg or if the lOP was not at least 5 mmHg below the pretreatment level). Concomitant medication known to affect lOP was excluded. This included other fj-blockers, clonidine, and carbonic anhydrase inhibitors. Any medication to the fellow eye was excluded if only one eye was in the study. If only one eye qualified for the study but both required treatment, the study drug was administered to both eyes. Only the qualifying eye was included in the analysis. The analysis of efficacy data was performed using a "worse" eye approach. If both eyes were being treated in the study, the eye with the lower mean visual field score at study entry was used in the analysis. If both eyes were being treated and had equivalent mean scores, the right eye was used in the analysis. For each eye at each timepoint, a mean visual field score was calculated from the 76 points tested in the 32 program.
Results In accordance with the protocol, patients were excluded or withdrawn from the visual field analysis for a variety of reasons (Tables IA and IB), which affected them at different stages during the 24 months of the study. Patients' data were included up to the time at which they developed a reason for exclusion. To qualify for inclusion in the visual field analysis, patients were required to have on-treatment, baseline, visual field available within the first 3 months of starting study therapy. For the analysis of vital signs and visual acuity data, we used the evaluations on the day that study drug was first administered and again at the end of the study. For
Vogel et al . Timolol versus Pilocarpine Table lA. Patients Excluded from the Visual Field Analysis
Lack of baseline data Adverse event or death Ineffective therapy Other reasons (Table IB) NS
=
Group
Baseline
Timolol Pilocarpine Timolol Pilocarpine Timolol Pilocarpine Timolol Pilocarpine
6 2 0 0 7 16 9 7
Month 24
6 2
4 NS (P = 0.363) 7
14
33 16 NS 19
P = 0.007* (P = 0.575)
not significant.
• Difference is statistically significant.
the analysis of safety data, all evaluations occurring after study drug was first administered were included.
Visual Fields Two approaches were used to analyze the visual field data. A timepoint analysis was done using the change from baseline for the six time periods (months 4, 8, 12, 16,20, and 24). Also, for patients with a baseline and at least two other visual field examinations, an analysis was done using a slope variable that represented each patient's change in visual field per unit time. Within these two approaches, two response variables were analyzed. The first was the arithmetic mean of the 76 points comprising the visual field (a measure of the overall sensitivity), and the second was the number of points that decreased by 5, 7, 10 or more decibels from baseline (a relative measure of the change in size of a scotoma).
mean decrease at each timepoint (Table 2). The change in mean visual field score is shown in Figure 1. Patients in both treatment groups had a significant number of points in the visual field that decreased in sensitivity by 5 dB or more at each timepoint. The number of points decreasing in this way was always greater in the pilocarpine group (Table 3).
Regression Slope Approach Least squares regression techniques were used to calculate a regression line for each patient that passed through the individual patient's baseline value. With respect to both response variables, the patients in the pilocarpine group were losing visual field score at a greater rate than patients in the timolol group. These differences achieved significance (P < 0.01). These results are summarized in Table 4.
Intraocular Pressure
Timepoint Approach Significant differences between the treatment groups were observed at months 4, 8, 12, 16,20, and 24 for both response variables with the results always favoring timolol. Patients in the timolol group had a mean increase (improvement) in mean visual field score at each timepoint, while patients in the pilocarpine treatment group had a
A partial diurnal curve consisting of 5 lOP measurements spanning 7.5 hours was collected at each visit. The maximum of these five measurements was used as the response variable. Intraocular pressure results are summarized in Table 5. Within both treatment groups, there was a significant decrease in both the maximum diurnal lOP and in the
Table lB. Patients Discontinued for "Other Reasons" Taking concomitant ,B-blocker Exclusion criteria developed Angle closure developed Lost to follow-up Patient uncooperative Protocol deviation Study closed before 24 months completed Visual field evaluation unsatisfactory Total
Timolol
Pilocarpine
1 0 0 5 1 2 6 1
1 1 1 7 0
16
19
4 4 1
1507
Ophthalmology Volume 99, Number 10, October 1992 Table 2. Mean Visual Field Thresholds (Decibels)
Timolol
Pilocarpine
Timolol
Pilocarpine
Timolol
Pilocarpine
Mean
75
63
Mean
65
48
Mean
62
41
Mean
65
37
Mean
58
37
Mean
58
35
Mean
46
26
18.5 6.2 18.5 6.1 19.2 5.8 18.7 6.1 18.3 6.2 18.3 6.0 18.9 5.9
16.9 5.7 16.9 5.7 16.4 5.9 17.2 5.4 17.4 5.6 17.1 5.7 17.8 4.1
18.6 6.3 19.6 6.2 18.9 6.6 18.4 6.6 18.5 6.6 19.3 6.4
16.3t 5.6 15.6t 6.0 16.6t 5.6 16.4t 5.7 16.0t 5.5 16.6t 4.4
Month
Ot 4 8 12 16 20 24
Receiving Treatment
Baseline
No:
SD
SD SD SD
SD SD SD
P Values
0.03
Originally received: July 16, 1991. Revision accepted: May 21, 1992. I Clinical Research Ophthalmology, Merck Sharp and Dohme Research Laboratory, West Point, PA. 2 Kings College Hospital, Denmark Hill, London. 3 Manchester Eye Hospital, Manchester. • Niagara Falls Memorial Medical C enter, Niagara Falls. The authors, other than those who work for Merck Sharp and Dohme Research Laboratory, have no proprietary interest in the development or marketing of this drug. Reprint requests to Roger Vogel, MD, MCOphth, Clinical Research Ophthalmology, Merck Sharp and Dohme Research Laboratory, West Point, PA 19486.
In 1978, pilocarpine was the most widely prescribed therapeutic agent for the treatment of glaucoma. However, although effective in lowering intraocular pressure (lOP), it was a parasympathomimetic agent and consequently produced miosis and spasm of accommodation. Timolol reduced lOP without the side effects of miosis and accommodation spasm and was regarded as a major improvement in glaucoma therapy. A nonselective !3-blocker, timolol slightly reduced pulse rate and blood pressure.) Langham 2 concluded that these systemic side effects of timolol might cause a reduction of ocular perfusion pressure and ocular blood flow with possible deleterious effects on the visual field.
1505
Ophthalmology
Volume 99, Number 10, October 1992
When timolol was introduced in 1979, little evidence existed to support the traditional view that pilocarpine therapy prevented the progression of visual field loss in primary open-angle glaucoma. Some questioned whether lowering lOP as the prime objective of therapy was justified by the available evidence. 3 There was concern, however, whether timolol could provide visual field protection equal to that believed to be provided by pilocarpine in view of the possible vascular effects oftimolol. Conversely, it was concluded that there might be some advantage with timolol because its topical side effects were fewer than those associated with pilocarpine use, and it seemed highly likely that compliance would be better. To clarify the position, a prospective, masked, controlled study was conceived using the Octopus 201 Perimeter (Interzeag, Schlieren, Switzerland). The objective was to compare the lOP-lowering capability, the tolerability, and visual field conservation of topical timolol and pilocarpine at the concentrations usually prescribed.
Materials and Methods Seven investigators entered a total of 189 patients with primary open-angle glaucoma into this study. Drug supplies were not masked but were issued by a secondary investigator according to a random schedule. Patients were recruited as they presented and qualified for entry into the study. One center originally participating did not sequentially assign allocation numbers as patients qualified and data from these patients do not appear in the tables. Elimination of these data did not alter the results or conclusions of the study. After a pretreatment washout period of 7 days, measurements of visual field and lOP, a slit-lamp examination, symptoms, gonioscopy, ophthalmoscopy, visual acuity, and pulse and blood pressure were performed. With the exception of gonioscopy and ophthalmoscopy (which were repeated at the end of the study), all parameters mentioned were followed at each visit throughout the study. Admission criteria included: (1) lOP of 22 mmHg or greater in 1 or both eyes on at least 1 of 5 measurements taken (approximately 9:00, 10:30, 12:00, 14:30, and 16: 30 hours) on the same day after a washout period of at least 7 days taking no glaucoma therapy; (2) open anterior .chamber angles; and, (3) a visual field defect recorded by "the Octopus Program 32, which showed a depression of three or more contiguous test points greater than 5 dB below "normal" values for the patient's age as determined by the Octopus or greater than 5 decibels (dB) below adjacent contiguous points. All visual field tests were performed at approximately the same time during the day and at approximately the same time in relation to the instillation of assigned study drug. Patients were excluded if they had: (1) a history of severe ocular trauma or intraocular surgery; (2) a corneal ulcer, ocular infection, or herpetic keratitis within 3 months of the study start; (3) a history of angle-closure or secondary glaucoma; (4) bronchial asthma or chronic obstructive pulmonary disease; greater than first degree heart block, uncompensated heart failure, or bradycardia
1506
of significant degree; (5) any disease other than open-angle glaucoma producing visual field loss; (6) concomitant medications known to affect lOP; (7) pregnant or nursing women or women of childbearing potential not using adequate means of contraception. All visual fields were performed on an Octopus 201 perimeter using programs 32 (at baseline, dose adjustment, and every 4 months during the study) and 42 (at dose adjustment and at 24 months). Patients were started on either 0.25% timolol twice daily or 2% pilocarpine four times daily according to the random allocation schedule. Patients entered a dose adjustment period, lOP was measured after 2 weeks, and the concentration of the drug was increased (to either timolol 0.5% or pilocarpine 4%) iflOP lowering was insufficient (i.e., if the lOP was above 22 mmHg). If the drug concentration was increased, the patient had a further lOP examination 2 weeks later. The visual field obtained at the end of this dose adjustment period was used as the on-treatment baseline to take account of any learning effect and any decrease in sensitivity as a result of the miosis caused by pilocarpine. Patients were examined every 4 months throughout the 2-year study. If a patient was receiving the high concentration of test medication, he or she was withdrawn from the study if his visual field defect dramatically increased or his lOP became uncontrolled (i.e., rose above 25 mmHg or if the lOP was not at least 5 mmHg below the pretreatment level). Concomitant medication known to affect lOP was excluded. This included other fj-blockers, clonidine, and carbonic anhydrase inhibitors. Any medication to the fellow eye was excluded if only one eye was in the study. If only one eye qualified for the study but both required treatment, the study drug was administered to both eyes. Only the qualifying eye was included in the analysis. The analysis of efficacy data was performed using a "worse" eye approach. If both eyes were being treated in the study, the eye with the lower mean visual field score at study entry was used in the analysis. If both eyes were being treated and had equivalent mean scores, the right eye was used in the analysis. For each eye at each timepoint, a mean visual field score was calculated from the 76 points tested in the 32 program.
Results In accordance with the protocol, patients were excluded or withdrawn from the visual field analysis for a variety of reasons (Tables IA and IB), which affected them at different stages during the 24 months of the study. Patients' data were included up to the time at which they developed a reason for exclusion. To qualify for inclusion in the visual field analysis, patients were required to have on-treatment, baseline, visual field available within the first 3 months of starting study therapy. For the analysis of vital signs and visual acuity data, we used the evaluations on the day that study drug was first administered and again at the end of the study. For
Vogel et al . Timolol versus Pilocarpine Table lA. Patients Excluded from the Visual Field Analysis
Lack of baseline data Adverse event or death Ineffective therapy Other reasons (Table IB) NS
=
Group
Baseline
Timolol Pilocarpine Timolol Pilocarpine Timolol Pilocarpine Timolol Pilocarpine
6 2 0 0 7 16 9 7
Month 24
6 2
4 NS (P = 0.363) 7
14
33 16 NS 19
P = 0.007* (P = 0.575)
not significant.
• Difference is statistically significant.
the analysis of safety data, all evaluations occurring after study drug was first administered were included.
Visual Fields Two approaches were used to analyze the visual field data. A timepoint analysis was done using the change from baseline for the six time periods (months 4, 8, 12, 16,20, and 24). Also, for patients with a baseline and at least two other visual field examinations, an analysis was done using a slope variable that represented each patient's change in visual field per unit time. Within these two approaches, two response variables were analyzed. The first was the arithmetic mean of the 76 points comprising the visual field (a measure of the overall sensitivity), and the second was the number of points that decreased by 5, 7, 10 or more decibels from baseline (a relative measure of the change in size of a scotoma).
mean decrease at each timepoint (Table 2). The change in mean visual field score is shown in Figure 1. Patients in both treatment groups had a significant number of points in the visual field that decreased in sensitivity by 5 dB or more at each timepoint. The number of points decreasing in this way was always greater in the pilocarpine group (Table 3).
Regression Slope Approach Least squares regression techniques were used to calculate a regression line for each patient that passed through the individual patient's baseline value. With respect to both response variables, the patients in the pilocarpine group were losing visual field score at a greater rate than patients in the timolol group. These differences achieved significance (P < 0.01). These results are summarized in Table 4.
Intraocular Pressure
Timepoint Approach Significant differences between the treatment groups were observed at months 4, 8, 12, 16,20, and 24 for both response variables with the results always favoring timolol. Patients in the timolol group had a mean increase (improvement) in mean visual field score at each timepoint, while patients in the pilocarpine treatment group had a
A partial diurnal curve consisting of 5 lOP measurements spanning 7.5 hours was collected at each visit. The maximum of these five measurements was used as the response variable. Intraocular pressure results are summarized in Table 5. Within both treatment groups, there was a significant decrease in both the maximum diurnal lOP and in the
Table lB. Patients Discontinued for "Other Reasons" Taking concomitant ,B-blocker Exclusion criteria developed Angle closure developed Lost to follow-up Patient uncooperative Protocol deviation Study closed before 24 months completed Visual field evaluation unsatisfactory Total
Timolol
Pilocarpine
1 0 0 5 1 2 6 1
1 1 1 7 0
16
19
4 4 1
1507
Ophthalmology Volume 99, Number 10, October 1992 Table 2. Mean Visual Field Thresholds (Decibels)
Timolol
Pilocarpine
Timolol
Pilocarpine
Timolol
Pilocarpine
Mean
75
63
Mean
65
48
Mean
62
41
Mean
65
37
Mean
58
37
Mean
58
35
Mean
46
26
18.5 6.2 18.5 6.1 19.2 5.8 18.7 6.1 18.3 6.2 18.3 6.0 18.9 5.9
16.9 5.7 16.9 5.7 16.4 5.9 17.2 5.4 17.4 5.6 17.1 5.7 17.8 4.1
18.6 6.3 19.6 6.2 18.9 6.6 18.4 6.6 18.5 6.6 19.3 6.4
16.3t 5.6 15.6t 6.0 16.6t 5.6 16.4t 5.7 16.0t 5.5 16.6t 4.4
Month
Ot 4 8 12 16 20 24
Receiving Treatment
Baseline
No:
SD
SD SD SD
SD SD SD
P Values
0.03
