Introduction

Exfoliation syndrome (XFS) is an age-related generalized disorder of the extracellular matrix that is now considered the most common identifiable condition leading to the development of open-angle glaucoma,1 exfoliation glaucoma (XFG), and subsequent blindness1, 2, 3 XFG comprises over 50% of open-angle glaucomas in some countries2, 4, 5, 6, 7, 8 and its socioeconomic importance has increased considerably in recent years. 1, 2, 3 At presentation XFG varies from primary open-angle glaucoma (POAG) due to older age,5, 7, 9 worse Intraocular pressure (IOP) characteristics,10, 11, 12 lower visual acuity,2, 8 more visual field damage,13, 14 and heavier trabecular pigmentation.8, 14 XFG subjects often present with unilateral glaucoma,5, 6, 15 exhibit a high mean IOP,10, 12, 14 a wider fluctuation of 24-h IOP,10, 11 and the IOP level may be rising with progression of the disease.16 It is therefore more likely for adjunctive therapy for multiple drugs and maximum medical therapy to be employed in the medical treatment of XFG.17, 18, 19, 20, 21, 22, 23, 24

It is well documented that XFG has a worse prognosis than POAG,25, 26, 27, 28, 29 but it has never been elucidated to date whether this is due to the difference in age, the worse IOP characteristics of XFG, or to an often quoted ‘poorer response’ of XFG to medications. Although in the literature there is often a suggestion that XFG demonstrates a poorer response to medications,17, 20, 28 this belief stems from retrospective data on glaucoma cohorts with different baseline IOPs and relies on evidence from a time when only timolol and pilocarpine were available as treatment options for this glaucoma. Surprisingly, to date there is very limited information with regard to the response and success rate of XFG with the newer medications (brimonidine, latanoprost, etc.). Since glaucomatous damage occurs more rapidly in XFG2, 14 and the rate of subsequent blindness is twice that seen with POAG,16, 25, 28 determining the most effective initial therapy and the best medical stepwise therapy in XFG is especially important to help control these subjects.

Latanoprost is currently one of the most popular choices in glaucoma treatment worldwide.30, 31, 32, 33, 34, 35 It is thus important to determine the success of monotherapy with latanoprost in XFG and compare the diurnal control obtained with this medication vs that with a classic drug of choice, timolol maleate, to facilitate the future management of this important secondary glaucoma.

Materials and methods

Subjects

Subjects were enrolled in this multicentre study if they had a clinical diagnosis of XFG, or ocular hypertension and XFS in at least one eye (study eye). If both eyes were eligible one eye was chosen at random. At screening, the IOP had to be controllable on monotherapy in the study eye and could safely undergo washout from their current medication in such a way that clinical stability of vision and the optic nerve is assured throughout this trial. All subjects were 39 years of age or older and prior to enrollment signed an informed consent form. At baseline, the mean untreated IOP (two measurements within one hour at the 1000 hour baseline time point) had to be between 22–36 mmHg inclusive. All subjects had open angles consistent in appearance with XFG. In the contralateral eyes, not included in the study, the IOP had to be controllable on no drug therapy, on the study medicine, or with laser/surgery. In the study eye, visual acuity had to be 0.3 on the Snellen chart or better.

Subjects were excluded if they had any abnormality preventing reliable applanation tonometry in the study eye; concurrent infectious/noninfectious conjunctivitis; keratitis or uveitis in either eye; any history of allergic hypersensitivity or poor tolerance to any components of the medications used in this trial; intraocular conventional surgery or laser surgery in the study eye; and according to the investigator's best judgement, risk of visual field or visual acuity worsening as a consequence of participation in the trial. Subjects were also excluded if they exhibited progressive retinal or optic nerve disease apart from glaucoma; if they were females of childbearing potential not using reliable means of birth control; pregnant or lactating females; and had any clinically significant, serious, or severe medical or psychiatric condition. Individuals who did not wish to risk darkened irides or eyelash changes during the trial were also not enrolled as were not individuals with active or potential ocular inflammation (including cystoid macula oedema), history of ocular herpes simplex or macular oedema, any contraindication to β-adrenergic blockers, including reactive airway disease, decompensated heart failure, second or third-degree heart block, or sick sinus syndrome. Subjects using systemic β-blockers were not enrolled in this study.

Procedures

All subjects signed an Institutional Review Board-approved informed consent form before entering this study. At the screening exam (Visit 1, Day −30), and all subsequent exams, IOPs were measured with a recently calibrated Goldmann applanation tonometer, visual acuity was measured on a Snellen visual acuity chart, and slit-lamp biomicroscopy as well as stereoscopic evaluation of the fundi were performed. Visual field assessment was performed at Visit 1 on the Humphrey Field Analyzer (program 24-2) or Octopus (G1 program). If subjects were treated prior to enrollment, they discontinued their medication according to a standard timetable of washing out (4 weeks for prostaglandin analogues and β-blockers, etc.).

Subjects returned to the clinic after the washout period for their untreated diurnal assessment (performed at 0800, 1000, 1400, and 2000) (Visit 2, Day 0). They were then randomized to the appropriate medication, either timolol maleate 0.5% (Vianex/MSD, Athens, Greece) one drop dosed at 0800 and 2000 hours, or latanoprost 0.005% (Pharmacia Hellas, Athens, Greece) dosed at 2000 hour and placebo (placebo artificial tears, supplied by the courtesy of Pharmacia Hellas) dosed at 0800 hours. Subjects returned for safety visits at 2 and 6 weeks (Visits 3 and 4). They returned at 3 months for a treated diurnal assessment of their IOP (Visit 5). During the diurnal IOP measurements at 3 months after initiation of therapy, the IOP was first measured at 0800 hours prior to instillation of the morning drop (ie, evaluation of the trough IOP), 2 h after instillation of the morning medication (peak IOP), at 1400 hours (afternoon IOP), and at 2000 hours (trough IOP for both drugs). At all time points the allowable range was ±1 h.

The medicine bottles were not masked. However, the physicians who performed the ocular examination, and those who obtained the patient history information, were not aware of the patient's treatment regimen. Further, the same physician recorded the patient's adverse event history for each visit. In addition, the same physician (but different from the physician who inquired about adverse events) measured the IOP at each visit.

At Visits 3–5, subjects were queried as to whether or not they had experienced symptoms suggesting an adverse event and about their general compliance to the masked medication dosing regimen since their last scheduled visit. Only compliant subjects were continued in the trial. Subjects were dismissed from the trial at the end of the treated diurnal assessment unless an unresolved adverse event existed.

Statistics

The statistics were analysed by Pharmaceutical Research Network, LLC (Charleston, SC, USA). All data analyses were two sided and a 0.05 α level was used. Both the reduction and absolute levels of IOPs between treatment groups were analyzed by an unpaired t-test for intergroup analysis. This study provided an 80% power that a difference of 2.0 mmHg could be excluded between the IOP of timolol maleate and latanoprost. A standard deviation of 3.5 mmHg was assumed. The mean deviation was also analysed by an unpaired t-test. A Bonferroni correction (α/5) was used to correct the P-values for multiple comparisons separately for the absolute IOP level and the reduction in pressure from baseline.

Safety parameters for ranked intergroup analysis were evaluated with a Mann–Whitney U and unranked were evaluated by a χ2 test as appropriate. Adverse events were evaluated with a χ2 or Fisher's Exact test as appropriate.

Results

Subjects

A total of 109 consecutive subjects were prospectively enrolled in six centres around Greece during 1 year. Out of these, 103 subjects (94.5%) completed the study. Of the six subjects who discontinued, two were lost to follow-up, one had poor IOP control on timolol, and three were discontinued due to adverse events. All subjects were Caucasian. There was no difference with regard to age, gender, baseline IOP, visual acuity, mean visual field defect in subjects with XFG, or mean cup-to-disc ratio between treatment groups (see Table 1). In total, 51 subjects were randomized to receive latanoprost 0.005% and 52 were randomized to receive timolol maleate 0.5%.

Table 1 Subject characteristics
Full size table

Intraocular pressure

Table 2 and Figure 1 demonstrate the IOP results with the two regimens. The baseline diurnal-untreated IOP for the latanoprost group was 24.9±3.2 mmHg, whereas the baseline diurnal-untreated IOP for the timolol group was 24.7±2.8 mmHg. Both treatments significantly reduced the baseline IOP at each time point and for the diurnal curve following 3 months of chronic dosing (P<0.0001)(Table 2). Latanoprost, however, had a significantly lower mean IOP at 0800 hours (P=0.0002). However, the IOPs were not statistically different at the 1000 hours (P=0.37), 1400 hours (P=0.03), and 2000 hours (P=0.71) time points and for the mean diurnal curve (P=0.07) after the Bonferroni correction.

Table 2 IOP reduction (mm Hg±standard deviation)
Full size table
Figure 1
figure 1

Diurnal mean IOPs for latanoprost at baseline (diamonds) and with latanoprost treatment (squares) and timolol at baseline (triangles) and with timolol treatment (circles).

Full size image

Likewise, latanoprost treatment reduced the IOP from baseline significantly more than timolol at 0800h (P<0.0001) (12 h following dosing) following three months of chronic dosing. Timolol reduced the IOP just as well as latanoprost at 1000, 1400, and 2000 hours (2, 8, and 12 h following dosing) and for the diurnal curve (P=0.03) after the Bonferroni correction.

Latanoprost treatment obtained a significantly better diurnal range (ie, smaller fluctuation of treated IOP) compared with timolol (2.4±1.2 vs 3.2±1.5 mmHg, P=0.0017) and a better mean maximum IOP (P=0.03).

Safety

Table 3 shows the ocular adverse effects for both medications and Table 4 shows the systemic adverse events. There was significantly greater conjunctival hyperaemia with latanoprost (P=0.01). One subject was withdrawn from the latanoprost group due to ocular intolerance, and one subject dropped out due to the occurrence of floaters in his study eye. In the timolol group, one subject was withdrawn due to bradycardia.

Table 3 Ocular side effects (number of events)
Full size table
Table 4 Systemic side effect (number of events)
Full size table

A number of systemic side effects were recorded in the study (Table 4). However, there were no differences between groups for any individual adverse event.

Discussion

The approach to the management of the subject with XFG may not be similar to primary POAG and the response to the new medications may differ when compared to POAG.3, 11, 19 Often the initial response to medications is greater with XFG, but in the long term these subjects can be more difficult to control with medical therapy than those with POAG.14, 17, 24 It is uncertain whether this is due to the declining efficacy of the medications themselves,2, 16, 20 the significantly higher mean IOP values,17 the greater fluctuation of 24-h IOP,23, 24 or the rising IOP during the course of XFG.16 In the past, XFG required surgical intervention more often and had a worse prognosis despite recourse to surgery earlier and more often.2, 3, 9, 12, 22, 25 However, these observations were made when medical therapy comprised mainly β-blockers, pilocarpine, and epinephrine. The impact of the newer medications on the success of medical therapy in XFG has yet to be taken into consideration.

To date, the specific target diurnal IOP for XFG has not yet been addressed. It is conceivable that the target IOP is different in XFG and it is possible that XFG, being a hypertensive glaucoma, may respond better to long-term IOP reduction compared with POAG. In a recent study, examining the role of IOP reduction in XFG progression, we documented a clear-cut benefit for IOP reduction in many subjects with XFG, with only 28% of XFG subjects deteriorating with a mean IOP lower than 17 mmHg over more than 5 years of follow-up.36 Thus, it is important to investigate how successful timely medical or surgical IOP reduction can be obtained.

Latanoprost is one of the most successful antiglaucoma medications available today and is fast becoming the gold standard of glaucoma therapy in POAG.30, 37 Unfortunately, to date there are limited data with respect to the efficacy of this medication in controlling the IOP in XFG.38

In this current study, we investigated the diurnal IOP reduction at four time points covering a period of 12 daytime hours to help establish the potency of these two medications in XFG. Both treatments significantly reduced the diurnal baseline IOP. Following a Bonferroni correction, the absolute IOP levels and the reduction in pressure from the nontreated baseline showed a nonstatistical trend in favour of latanoprost. The lack of significance may have resulted from, at least in part, the fact that the study was slightly underpowered (80% power to detect a 2.0 mmHg difference). Sponsored clinical trials typically are powered statistically to detect a 1.5 mmHg difference.

Latanoprost treatment, however, demonstrated a significantly better diurnal range and better mean peak IOP compared with timolol. The smaller fluctuation of treated IOP may confer a long-term benefit in the management of XFG since it has been shown in POAG that the less the fluctuation, the better the prognosis in the management of glaucoma.39, 40, 41

With regard to specific time points, latanoprost treatment reduced the IOP significantly more than timolol and provided a lower absolute IOP at 0800 hours (12 h following dosing). However, only a trend of a lower IOP existed (after the Bonferroni Correction) with latanoprost for the other time points and the diurnal curve following 3 months of chronic dosing. It is not surprising that latanoprost is more efficacious than timolol at 0800 hours since this is the trough point for timolol, but probably the peak efficacy for latanoprost. It has been shown that latanoprost works better 12–24 h after administration.35 Considering the proximity of our diurnal baseline IOP time points to that of the latanoprost regulatory studies, our results indicate a similar, but slightly less, IOP effect in comparison with that found for POAG in these previously published multicentre studies.31, 32, 34

With regard to diurnal efficacy, our study shows that both latanoprost and timolol appear to be at least as efficacious in XFG as previously reported to be in POAG.31, 32 The specific response of the various medications with XFG has been a topic of controversy.2, 3 Even with timolol maleate, the gold standard against which medications are measured, eyes with XFG have been reported to respond in a varied way in comparison with POAG. Aasved et al.42 have reported a worse response, whereas others suggest an equal,43 or greater11 ocular hypotensive effect. Konstas et al.11 evaluated the 24-h response with timolol 0.5% and showed that despite a greater initial IOP reduction in the XFG subjects, they still had a higher mean IOP and greater fluctuation in the 24-h IOP curve than subjects with POAG. Little previous information regarding the diurnal efficacy of latanoprost in EXG is available. Konstas et al.24 demonstrated a similar diurnal IOP response with adding latanoprost, as a third-line medication, to timolol and dorzolamide compared to pilocarpine 4%.

To date, it is not known if latanoprost, by diverting aqueous to the uveoscleral outflow, has more or less of a beneficial effect on XFG vs medications that reduce aqueous production like timolol or vis-à-vis POAG. Furthermore, it has not been elucidated if the pigment granules liberated in XFG are deposited in the uveoscleral outflow, and if so, whether this plays a role in the context of XFG.

This study suggests that latanoprost provides a statistically lower 0800 hours IOP with similar safety than timolol as monotherapy in the treatment of XFG. The favourable diurnal IOP response to latanoprost compared to timolol maleate in XFG would be useful in guiding ophthalmologists in the selection of first-line medication in the management of XFG. However, it should be noted that the present study is not a full 24-h study and thus nighttime IOPs have not been adequately assessed.