Efficacy and Tolerability of Prostaglandin-Timolol Fixed Combinations: a Meta-Analysis of Randomized Clinical Trials

Home , Latanoprost, Prostaglandin, Travoprost

Eur J Ophthalmol22 (2012; :1) 5-18 DOI: 10.5301/ejo.5000009

ORIGINAL ARTICLE

Efficacy and tolerability of prostaglandin-timolol fixed combinations: a meta-analysis of randomized clinical trials

Florent Aptel1, Michel Cucherat2, Philippe Denis3

1Department of Ophthalmology, Edouard Herriot Hospital, Lyon - France 2TherInfo, Lyon1 University, Lyon - France 3Department of Ophthalmology, Croix-Rousse Hospital, Lyon - France

Purpose. To evaluate the intraocular pressure (IOP)-lowering effects and tolerability of the 3 prostaglandin-timolol fixed combinations (PG-timolol FCs). Methods. Clinical trials comparing directly the PG-timolol FCs or comparing the PG-timolol FCs to their individual components were thoroughly searched. The main outcome measures were efficacy assessed by IOP (taken at 9 am, noon, 4 pm, and over the mean diurnal curve) change at 3 months (or after 1 to 6 months of treatment if no data were available at month 3) from baseline and tolerability assessed by the incidence of conjunctival hyperemia. Results. Twenty trials were identified (n=4684 patients). Intraocular pressure reduction was usually greater with the 3 PG-timolol FCs than the individual PG (mean difference [MD] 0.00 mmHg to 2.59 mmHg; p>0.1 to p<0.001). The incidence of hyperemia was significantly less with latanoprost- and bimatoprost-timolol FCs than with the individual PG (relative risk = 0.66 and 0.61; p=0.05 and p<0.001). From direct comparisons, IOP reduction was significantly greatest with bimatoprost-timolol FC, at 9 am, 4 pm, and over the mean diurnal curve compared to latanoprost-timolol FC (MD = 0.90 mmHg to 1.48 mmHg; p<0.001) and at all time points compared to travoprost-timolol FC (MD = 0.66 mmHg to 0.90 mmHg; p<0.001). The incidence of hyperemia was not significantly less with latanoprost-timolol FC than with bimatoprost-timolol FC (relative risk = 1.32; p>0.1). Conclusions. The 3 PG-timolol FCs provide a greater IOP reduction and lower incidence of hyperemia than the 3 PGs alone. The direct comparisons suggest a greater efficacy of the bimatoprost-timolol FC compared with latanoprost- and travoprost-timolol FCs.

Key Words. Effect size, Fixed combination, Glaucoma, Intraocular pressure, Meta-analysis, Prosta- glandin-timolol

Accepted: May, 16, 2011

INTRODUCTION been identified, among which elevated intraocular pressure (IOP) is the only currently treatable one. Over the last Glaucoma is a multifactorial disease of the optic nerve with decade, several large clinical studies have proven that IOP a characteristic progressive degeneration and death of the reduction can delay or prevent the progression of ocular retinal ganglion cells and their axons that form the optic hypertension to glaucoma, and can delay disease progres- nerve. The exact mechanisms of the retinal ganglion cells sion in established glaucoma (1). degeneration are not known, but many risk factors have The prostaglandin analogues (PGs) have become the first

line of treatment because of their efficacy, convenience, Literature search and minimal systemic side effects. However, for many patients, a single medication is not sufficient to reach a All RCTs up to the end of August 2010 were searched us- pressure at which optic nerve damage will not progress. ing MEDLINE, EMBASE, the Cochrane Central Register of Results of the Ocular Hypertension Treatment Study have Controlled Trials, ClinicalTrials.gov, and Controlled-trials. shown that almost 40% of the patients required 2 or more com. Key words used for the search included glaucoma, medications to reduce IOP to a given target pressure (2). ocular hypertension, latanoprost, travoprost, bimatoprost, In recent years, the use of fixed combinations (FCs) con- timolol, prostaglandin, prostamide, fixed combination, and taining 2 medications in a single bottle has increased ocular lipids. All prospective RCTs were identified using a steadily. A beta-blocker (timolol) is the most common 2-level search strategy: l) databases in the public domain agent added to PG worldwide. The use of PG-timolol FCs were sought using Web-based search engines and 2) rele- has several advantages. FCs reduce the exposure to pre- vant RCTs were identified through a manual search of sec- servative such as benzalkonium chloride, which is known ondary sources, including the references of articles found to exert toxic effects for the ocular surface (3). Moreover, from the databases and information obtained directly from FCs simplify dosing regiments and may improve both the expert and pharmaceutical companies (29). All references compliance and the quality of life (4, 5). In many coun- were downloaded for consolidation, elimination of dupli- tries, the cost of FCs is less than the combined cost of the cates, and further scrutiny. Additional searches were con- component medications. Fixed combinations are there- ducted for known trial acronyms cited in review articles. fore now widely used in some countries, notably in Eu- rope, and are recommended in practice guidelines (6, 7). Selection Three PG-timolol FCs are marketed, namely the 0.005% latanoprost with 0.5% timolol (LTFC, Pfizer Inc, New York, All potentially relevant trials found in published articles and NY, USA), the 0.004% travoprost with 0.5% timolol (TTFC, abstracts were assessed for inclusion into the study. To be Alcon Inc., Fort Worth, TX, USA), and the 0.03% bimato- included, they had to be prospective, double- or investiga- prost with 0.5% timolol (BTFC, Allergan Inc., Irvine, CA, tor-masked, completely randomized, and controlled clinical USA). Only a few clinical trials have compared the effi- trials with an intent-to-treat or per-protocol analysis. Male or cacy and tolerability of the 3 PG-timolol FCs (8-12). By female patients with primary open-angle glaucoma, exfolia- contrast, many studies have compared the FCs to each tive glaucoma, or ocular hypertension had to be 18 years of the component medications used as monotherapy (13- of age or older, without systemic/ocular medications that 27). However, results were sometimes conflicting, with could affect IOP, and without intraocular laser/surgery within some studies not reporting a significant difference in ef- the past 3 months. The patients had to be treated with LTFC ficacy between prostaglandins and PG-timolol fixed com- once daily (0.005% latanoprost and 0.5% timolol maleate, binations (17-26). Prostaglandin-timolol FCs are therefore Pfizer Inc), latanoprost once daily (0.005% latanoprost, Pfizer not approved by regulatory authorities in some countries, Inc), TTFC once daily (0.004% travoprost and 0.5% timolol such as the United States of America. maleate, Alcon Inc.), travoprost (0.004%, Alcon Inc.), BTFC We undertook this systematic review and network meta- once daily (0.03% bimatoprost and 0.5% timolol maleate, analyses to evaluate and compare the IOP-lowering effects Allergan Inc.), bimatoprost once daily (0.03%, Allergan Inc.), and tolerability of the 3 PG-timolol fixed combinations. or timolol 0.5% monotherapy. Efficacy was determined as the mean IOP change between baseline and month 3 (or after 1 to 6 months of treatment METHODS if no data were available at month 3). Because IOP fluctua- tions have been suggested to be a risk factor for glaucoma This study was conceived, conducted, and reported ac- progression, 3 diurnal IOP measurements were used for de- cording to the Preferred Reporting Items for Systematic termining efficacy (9 am ± 2, 12 pm ± 2, 4 pm ± 2, and over the Reviews and Meta-analyses (PRISMA) statement for im- mean diurnal curve). The mean diurnal curve was defined as proving the quality of reports of randomized clinical trial the mean of all the diurnal measures recorded at month 3 (or (RCT) meta-analyses (28). after 1 to 6 months of treatment if no data were available at

month 3). Tolerability was evaluated based on conjunctival using a funnel plot of the logarithm of effect size versus the hyperemia reported by patients at month 3 (or after 1 to 6 standard error for each trial. months of treatment if no data were available at month 3). Likewise, where a SD of an IOP change was not reported, it was calculated according to the method of Follman et al Quality assessment (33), which involves making an assumption for the unre- ported within-individual correlation between baseline and Two investigators reviewed the primary studies (64 po- final values: tentially relevant RCTs) independently to determine if they Var (follow-up IOP – baseline IOP) = Var (follow-up IOP) + were appropriate for inclusion in the analysis. Quality of the Var (baseline IOP) – 2 × Rho SD (baseline IOP) × SD (follow- trials was assessed with the scoring system of Jadad et al up IOP) (30). Trial quality was evaluated on the basis of adequacy where Rho is the correlation between the measurement of allocation concealment using the criteria recommended at baseline and that at follow-up (33). A Rho value of 0.2 by Schulz et al (31) as well as Jüni and Egger (32), descrip- was used, as estimated from the results of all studies with tion of withdrawals, and masking of the patient and the complete baseline follow-up IOP and change SD values. In investigator. A PRISMA flow diagram of the study selection case neither an SD nor a standard error of the mean (SEM) process is illustrated in Figure l. of the follow-up measurement was available, baseline SD was used as an estimate of the SD of the follow-up mea- Data abstraction surement. Statistical significance was set at p<0.05 (34).

Two investigators used a standardized form to extract in du- plicate data. In some cases, a digital ruler was used to extract RESULTS data from figures (MicroFox-Software, Columbus, OH).

Statistical analysis Trial flow and study characteristics

Analyses were carried out using public domain R language A total of 192 of 228 abstracts and 16 of 36 articles were and rmeta package (R Foundation for Statistical Comput- excluded (Fig. l). Twenty RCTs enrolling a total of 4684 pa- ing, Vienna, Austria). For categorical outcomes, treatment tients were retained for the meta-analysis (Tab. I). Figure effect was analyzed by the relative risk (RR) and its associat- 2 shows the network of interventions. Numbers of trials ed 95% confidence interval (CI). For continuous outcomes, reported in Figure 2 do not add up to numbers reported in treatment effect was analyzed by mean differences (MD) and Figure 1 or Table I because of multi-arm trials. their 95% CIs. Pooled estimates were calculated through the use of random-effect meta-analysis using inverse-vari- Efficacy ance weighting. For individual trials, MD was determined by the difference between the mean change in the treated Figures 3-6 show pooled estimates of IOP reduction com- group and that in the control group. Linear comparison was parisons obtained from trials comparing the FCs head to conducted because the data required for logarithmic com- head and comparing the FCs to their individual compo- parison were not available. None of the studies provided the nents. Intraocular pressure reduction was significantly logarithms of IOP differences. The standard error of MD was greatest with BTFC, at 9 am, 4 pm, and over the mean di- obtained by calculating the square root of the sum of the urnal curve compared to LTFC (9 am: MD = 1.37 mmHg; variances in outcomes in the treated and control groups. p<0.001; noon: MD = 1.48 mmHg; p<0.1; 4 pm: MD = 0.90 Heterogeneity between studies was analyzed with the mmHg; p<0.001; over the diurnal curve: MD = 1.24 mmHg; standard χ2 test (Cochran). We also used the I2 statistic p<0.001) and at all time points compared to TTFC (9 am: MD that is independent of the number of studies and quantifies = 0.70 mmHg; p<0.001; noon: MD = 0.90 mmHg; p<0.001; heterogeneity on a scale of 0% to 100%. Very large hetero- 4 pm: MD = 0.80 mmHg; p<0.001; over the diurnal curve: geneity between studies is usually denoted by I2 values of MD = 0.66 mmHg; p<0.001). The findings for LTFC 75% or more. Publication bias was assessed graphically and TTFC indicate that IOP reduction was significantly

Fig. 1 - Flow diagram of the study selection process. IOP = intraocular pressure.

Fig. 2 - Network formed by interventions. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

TABLE I - CHARACTERISTICS OF IDENTIFIED CLINICAL TRIALS

Trials Treatment Blinding Efficacy No. effective Mean % Women POAG/OHT % With- Jadad Funding assessment randomiza- age, y drawal score period, mo tions

Martinez and BTFC vs LTFC Investigator- 3 54 71.4 48.1 54/0 0 3 Galician Sanchez (8) masked (24 OAG, government 30 PEX) Centofanti BTFC vs LTFC Investigator- 3 82 64.7 51.2 82/0 0 3 Allergan et al (9) masked

Martinez and BTFC vs LTFC Investigator- 1 36 72.2 47.2 36/0 0 3 Galician Sanchez (10) masked (17 OAG, government 19 PEX) Topouzis TTFC vs LTFC Double-masked 3 to 12 407 64.8 59 268/64 18.7 5 Alcon et al (11) (233 OAG, 8 PDG, 27 PEX) Centofanti TTFC vs BTFC Double-masked 3 91 65.9 NS 89/0 2.2 4 Allergan et al (12) Brandt BTFC vs Double-masked 3 1061 61 52.6 585/454 6.9 5 Allergan et al (13) Bimatoprost vs Timolol Hommer BTFC vs Double-masked 1 268 59.9 51.1 86/182 2.9 5 Allergan et al (14) Bimatoprost Palmberg LTFC vs Double-masked 3 335 65.2 52.2 250/85 11.9 5 Pfizer et al (15) latanoprost (237 OAG, vs timolol 8 PEX, 5 PDG) Higginbotham LTFC vs Double-masked 3 394 63.9 52 278/116 14.2 5 Pfizer et al (16) latanoprost (270 OAG, vs timolol 7 PDG, 1 PEX) Magacho LTFC vs Double-masked 1 28 56.5 NS 25/3 0 5 NS et al (17) latanoprost Konstas LTFC vs Investigator- 3 37 65.8 62.2 37/0 5.4 3 NS et al (18) latanoprost masked Olander LTFC vs Double-masked 1 350 63.8 59.2 291/57 2.6 5 Pfizer et al (19) latanoprost (263 OAG, 13 PEX, 6 PDG, 5 mixed glaucoma, 4 other) Higginbotham LTFC vs Double-masked 6 418 62.3 48.6 309/109 17.5 5 Pharmacia et al (20) latanoprost (278 OAG, -Upjohn vs timolol 9 PEX, 13 PDG, 8 mixed) Pfeiffer LTFC vs Double-masked 6 436 63.6 55 372/64 9.2 4 Pharmacia et al (21) latanoprost (22 PEX, 8 PDG, vs timolol 6 mixed) Diestelhorst and LTFC vs Double-masked 1 92 62 59.8 92/0 5.6 4 Pharmacia Almegård (22) latanoprost (87 OAG, 5 PEX) vs timolol Konstas LTFC vs Investigator- 2 29 63.7 55.2 29/0 0 3 Pfizer et al (23) timolol masked Konstas LTFC vs Investigator- 2 33 62.4 61.8 33/0 2.9 4 Pfizer et al (24) timolol masked Konstas TTFC vs Double-masked 2 34 63.9 56 34/0 5.9 5 Alcon et al (25) travoprost Barnebey TTFC vs Double-masked 3 263 63 51.1 178/80 1.9 4 Alcon et al (26) travoprost (174 OAG, vs timolol 2 PEX, 2 PDG) Schuman TTFC vs Double-masked 3 236 62.2 58.1 132/104 1.3 5 Alcon et al (27) timolol (125 OAG, 6 PDG, 1 PEX)

BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; OHT = ocular hypertension; PDG = pigment dispersion glaucoma; PEX = pseudoexfoliation glaucoma; POAG = primary open-angle glaucoma; TTFC = travoprost-timolol fixed combination.

Fig. 3 - Comparison of treatment efficacy at 9 am (MD = mean difference with results in mmHg; 95% CI = 95% confidence interval). First column: number of patients, mean intraocular pressure (IOP) reduction, and standard deviation with the studied treatment. Second column: number of patients, mean IOP reduction, and standard deviation with the control treatment. In the graph, the vertical lines indicate means, the horizontal lines indicate 95% CI, the center of the diamond indicates mean, and the 2 extremities of the diamond indicate 95% CI. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

Fig. 4 - Comparison of treatment efficacy at noon (MD = mean difference with results in mmHg; 95% CI = 95% confidence interval). First column: number of patients, mean intraocular pressure (IOP) reduction, and standard deviation with the studied treatment. Second column: number of patients, mean IOP reduction, and standard deviation with the control treatment. In the graph, the vertical lines indicate means, the horizontal lines indicate 95% CI, the center of the diamond indicates mean, and the 2 extremities of the diamond indicate 95% CI. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

Fig. 5 - Comparison of treatment efficacy at 4 pm (MD = mean difference with results in mmHg; 95% CI = 95% confidence interval). First column: number of patients, mean intraocular pressure (IOP) reduction, and standard deviation with the studied treatment. Second column: number of patients, mean IOP reduction, and standard deviation with the control treatment. In the graph, the vertical lines indicate means, the horizontal lines indicate 95% CI, the center of the diamond indicates mean, and the 2 extremities of the diamond indicate 95% CI. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

Fig. 6 - Comparison of treatment efficacy over the diurnal curve (MD = mean difference with results in mmHg; 95% CI = 95% confidence interval). First column: number of patients, mean intraocular pressure (IOP) reduction, and standard deviation with the studied treatment. Second column: number of patients, mean IOP reduction, and standard deviation with the control treatment. In the graph, the vertical lines indicate means, the horizontal lines indicate 95% CI, the center of the diamond indicates mean, and the 2 extremities of the diamond indicate 95% CI. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

Fig. 7 - Comparison of conjunctival hyperemia (RR = relative risk; 95% CI = 95% confidence interval). First column: number of patients reporting conjunctival hyperemia and total number of patients with the studied treatment. Second column: number of patients reporting conjunctival hyperemia and total number of patients with the control treatment. In the graph, the vertical lines indicate RRs, the horizontal lines indicate 95% CI, the center of the diamond indicates RRs, and the 2 extremities of the diamond indicate 95% CI. BTFC = bimatoprost-timolol fixed combination; LTFC = latanoprost-timolol fixed combination; TTFC = travoprost-timolol fixed combination.

greater with TTFC at 9 am (MD = 0.30 mmHg; p<0.001) clinical trials (Clinicaltrials.gov, Controlled-trials.com, and and noon (MD = 0.30 mmHg; p<0.001), no significantly Cochrane Central Register of Controlled Trials). As most greater with TTFC over the mean diurnal curve (MD = governmental or professional organizations, ethics com- 0.10 mmHg; p>0.1), and significantly greater with LTFC mittees, and scientific or medical journals have policies on at 4 pm (MD = 0.30 mmHg; p<0.001). clinical trials for many years, it is unlikely that we missed Intraocular pressure reduction was greater with the 3 PG- any relevant trial. The inspection of the funnel plots did not timolol FCs than the individual PG (LTFC vs latanoprost: MD show clear asymmetry; therefore we did not perform sta- = 1.31 mmHg to 2.01 mmHg; p<0.01; TTFC vs travoprost: tistical tests for funnel plot asymmetry. MD = 2.02 mmHg to 2.59 mmHg; p<0.001; BTFC vs bi- The heterogeneity between trials in our meta-analysis was matoprost: MD = 0.00 to 1.14 mmHg; p>0.1 to p<0.001). important for some comparison and endpoint. A meta- analysis is more reliable if the trials which are combined Tolerability are highly similar. We therefore used a random effect model which deals with heterogeneity and gives conservative es- Figure 7 shows pooled estimates of incidences of self-re- timates. The robustness of the statistical method used to ported conjunctival hyperemia obtained from trials com- evaluate treatment effect (based on mean differences with paring the FCs head to head or to their individual com- their 95% CIs for continuous outcomes and RRs for cat- ponents. The incidence of hyperemia was significantly egorical outcomes) can be confirmed as, if anything, its less with latanoprost- and bimatoprost-timolol FCs than use could have only led to an underestimation of the differ- with the individual PG (LTFC vs latanoprost relative risk ences between the drugs. = 0.66; p=0.05; BTFC vs bimatoprost relative risk = 0.61; We used the Jadad scoring system to assess the method- p<0.001). From direct comparisons, the incidence of hy- ologic quality of the RCTs included in the meta-analysis. peremia was not significantly less with latanoprost-timolol Studies are scored according to the quality of randomiza- FC than with bimatoprost-timolol FC (relative risk = 1.32; tion, blinding, and accountability of all patients, including p>0.1). No direct comparison between TTFC and BTFC withdrawals and dropouts, allocating trials a score of be- was available. tween 0 (very poor) and 5 (rigorous). Usually, the method- ologic quality of a RCT is considered to be high and appropriate for inclusion in a meta-analysis when the score DISCUSSION is equal to or more than 3. As 5 of the 20 studies included in the present work have a score of 3, 5 of the 20 have a score of 4, and 10 of the 20 have a score of 5, we can Principal findings consider that the studies included were generally good in quality. The 3 PG-timolol FCs provide a greater IOP reduction at Only one RCT provided valid data for the direct comparison all time points and a lower incidence of conjunctival hype- of LTFC to TTFC and BTFC to TTFC. The analysis of direct remia than the 3 PGs alone. The direct comparisons sug- comparisons therefore did not provide more information gest a greater efficacy of the BTFC compared to LTFC and than that provided by the initial RCT itself. Other valid clini- TTFC. The tolerability of the 3 FCs does not seem to be cal trials directly comparing the FCs could be performed to significantly different, however, limited data are available provide further evidence. A recently published study com- about this point. paring TTFC to LTFC in patients with exfoliative glaucoma supports the findings of the present meta-analysis (35). Strengths and potential biases Relation to other studies The risk of publication bias was minimized by an exhaus- tive search of the available literature. Publication bias may Only a few reviews on PG-timolol FCs have been published arise from the tendency to underpublish research with a to date (36, 37). Such reviews present some selected stud- negative or inconclusive outcome. To prevent this bias, ies and summarize the results of the individual studies. we searched for all the relevant trials from the registers of However, the exact search methodologies, including the

inclusion and exclusion criteria, were not detailed. More- junctival hyperemia induced by a PG has been described. over, reviews do not use statistical techniques to combine Unfortunately, the RCTs included in the present study did the results of the individual studies. As the previously pub- not allow us to compare the tolerability of the TTFC and lished reviews did not combine the individual trials to obtain BTFC. pooled estimates of the IOP reduction or tolerability, their nonquantitative results cannot be compared to those of Clinical implications the present study. A recently published meta-analysis had compared the PG-timolol FCs to the unfixed combinations, The 3 PG-timolol FCs provide a greater IOP reduction and and the morning to the evening dosing of the PG-timolol lower incidence of hyperemia than the 3 PGs alone. This FCs (38). This work showed that the concomitant use of demonstrates that timolol is an efficient adjunctive drug to latanoprost and timolol has a greater ability to reduce IOP PGs. Prostaglandin-timolol FCs may thus be considered when compared with the fixed combination. Comparisons as second-line therapy in patients who have failed to at- of morning and evening dosing of a PG-timolol FC did not tain target IOP with timolol monotherapy. The PG-timolol show a significant difference in mean IOP-lowering effect. FCs have not been approved for use in the United States This work did not directly compare the efficacy and toler- because regulatory studies demonstrated insufficient ability of the 3 PG-timolol FCs, and did not compare the 3 pressure reduction between LTFC dosed in the morning FCs to their individual components. and latanoprost dosed in the evening. Our findings on 12 By contrast, previous meta-analyses have evaluated and studies comparing the 3 PG-timolol FCs to the PGs alone compared the efficacy and tolerability of the 3 PGs. In one showed that the FCs have an overall greater ability to re- of these studies performing an intercomparison of the 3 duce IOP when compared to timolol or the original PG. The PGs, we found that bimatoprost has a greater ability to differences in IOP reduction were almost always statisti- reduce IOP compared to latanoprost and travoprost (39). cally significant, up to 2.59 mmHg when timolol is added The effect size measurements (weighted mean) ranged to travoprost. In some cases, particularly when BTFC is from 0.52 mmHg (vs travoprost at 4 pm) to 1.17 mmHg (vs compared to bimatoprost, the differences in IOP reduction latanoprost at noon). The same drug (0.5% timolol given may seem modest. It should be mentioned that BTFC was once daily in the fixed combination bottle) is added to the given in the morning in the included studies, and thus the 3 PGs to obtain the 3 FCs evaluated in the present study. diurnal time points of IOP measurements did not include However, as the additive efficacy of one given drug may the peak IOP reduction of the PG (about 12 hours after in- vary depending on the other drug, it was not trivial to sug- stillation). As the hypotensive effect of the PG is likely pre- gest that the combination of bimatoprost and timolol may dominant compared to the effect of the beta-blocker (given keep a greater efficacy than the combinations of latano- once daily in the fixed combination formulation), diurnal prost or travoprost to timolol. It should be noted that we evaluation of a PG-timolol FC may lead to underestimate did not compare the timolol-PG fixed combinations to the the differences in IOP reduction ability compared to a PG unfixed combinations in the present work. Other previous given in the evening. A 24-hour evaluation would probably or forthcoming work have compared the fixed to the un- have led to an even greater difference between the IOP fixed timolol-PG combinations, and have found statistically reduction ability of the FCs and the PG alone. A recently significant but often limited differences in terms of IOP re- published study has compared the morning and evening duction ability (37, 40). dosing of BTFC, and has shown that evening dosing gives In terms of tolerability, we found in the above mentioned rise to statistically better 24-hour IOP control than morn- meta-analysis a higher incidence of conjunctival hyperemia ing dosing (41). Moreover, we demonstrated that the FCs with bimatoprost than with latanoprost or travoprost. In allow a lower incidence of hyperemia than the PGs alone. the present study, the incidence of conjunctival hyperemia Ocular hyperemia is one of the primary side effects lead- is significantly less with BTFC than with the bimatoprost ing to poor compliance and early discontinuation. Stud- alone (RR=0.61; p<0.001), and not significantly higher ies have demonstrated that adherence is a crucial issue in than with the LTFC (RR=1.32; p>0.1). Also, the exact phar- glaucoma treatment, and that poor adherence increases macologic mechanisms are not yet known; the ability of the risk of progression. beta-adrenergic receptor antagonists to decrease the con- The direct comparison suggests a greater efficacy of the

bimatoprost-timolol FC compared with latanoprost- and REFERENCES travoprost-timolol FCs. The differences in IOP reduction were relatively small, but maintained at all time points. The 1. Weinreb RN, Khaw PT. Primary open-angle glaucoma. Lancet clinical significance of such IOP differences is not known. 2004; 363: 1711-20. 2. Kass MA, Heuer DK, Higginbotham EJ, et al. The Ocular Hy- However, some large clinical trials have found that small pertension Treatment Study: a randomized trial determines that additional IOP reduction can reduce the risk of further vi- topical ocular hypotensive medication delays or prevents the sual field defect progression. These findings are likely clini- onset of primary open-angle glaucoma. Arch Ophthalmol 2002; cally relevant, particularly in patients requiring a greater 120: 701-13. IOP reduction or at greater risk of glaucoma progression. 3. Baudouin C, Labbé A, Liang H, Pauly A, Brignole-Baudouin F. We found a much higher incidence of hyperemia with the Preservatives in eyedrops: the good, the bad and the ugly. Prog TTFC than the LTFC. This result should probably be taken Retin Eye Res 2010; 29: 312-34. 4. Robin AL, Novack GD, Covert DW, Crockett RS, Marcic TS. with caution. First, only one RCT provided data for the Adherence in glaucoma: objective measurements of once-daily comparison of self-reported conjunctival hyperemia with and adjunctive medication use. Am J Ophthalmol 2007; 144: TTFC and LTFC (11). Secondly, an objective assessment of 533-40. conjunctival hyperemia was also performed by the investi- 5. Covert D, Robin AL. Adjunctive glaucoma therapy use associ- gator in this double-masked study, by grading conjunctival ated with travoprost, bimatoprost, and latanoprost. Curr Med hyperemia based on a scale ranging from 0 to 3. Differenc- Res Opin 2006; 22: 971-6. 6. Terminology and Guidelines for Glaucoma (European Guide- es in average hyperemia scores were always small and not lines), 3rd ed. Savona: DOGMA, 2008; 86-104. statistically significant at 8 of the 10 study visits. Therefore, 7. Weinreb RN, Liebmann J. Medical Treatment of Glaucoma. we believe that the above-mentioned higher incidence of Amsterdam: Kugler Publications, 2010; 117-169. hyperemia with the TTFC compared to the LTFC should be 8. Martinez A, Sanchez M. Bimatoprost/timolol fixed combination interpreted with caution. vs latanoprost/timolol fixed combination in open-angle glau- In summary, the present meta-analysis indicates that the 3 coma patients. Eye 2009; 23: 810-8. 9. Centofanti M, Oddone F, Vetrugno M, et al. Efficacy of the fixed PG-timolol FCs provide a greater IOP reduction and lower combinations of bimatoprost or latanoprost plus timolol in pa- incidence of hyperemia than the 3 PGs alone, and suggests tients uncontrolled with prostaglandin monotherapy: a multi- a greater efficacy of the bimatoprost-timolol FC compared center, randomized, investigator-masked, clinical study. Eur J with latanoprost- and travoprost-timolol FCs. Ophthalmol 2009; 19: 66-71. 10. Martinez A, Sanchez M. A comparison of the safety and intraocular pressure lowering of bimatoprost/timolol fixed com- The authors report no proprietary interest or financial support. bination versus latanoprost/timolol fixed combination in patients with open-angle glaucoma. Curr Med Res Opin 2007; 23: 1025-32. 11. Topouzis F, Melamed S, Danesh-Meyer H, et al. A 1-year study to compare the efficacy and safety of once-daily travoprost 0.004%/timolol 0.5% to once-daily latanoprost 0.005%/timolol 0.5% in patients with open-angle glaucoma or ocular hypertension. Eur J Ophthalmol 2007; 17: 183-90. 12. Centofanti M, Oddone F, Gandolfi S, et al. Comparison of travoprost and bimatoprost plus timolol fixed combinations in open-angle glaucoma patients previously treated with latanoprost plus timolol fixed combination. Am J Ophthalmol 2010; 150: 575-80. 13. Brandt JD, Cantor LB, Katz LJ, Batoosingh AL, Chou C, Address for correspondence: Bossowska I, Ganfort Investigators Group II. Bimatoprost/ Florent Aptel, MD timolol fixed combination: a 3-month double-masked, ran- Department of Ophthalmology domized parallel comparison to its individual components in Edouard Herriot Hospital patients with glaucoma or ocular hypertension. J Glaucoma 5, place d’Arsonval 69437 Lyon cedex 03 2008; 17: 211-6. France 14. Hommer A, Ganfort Investigators Group I. A double-masked, [email protected] randomized, parallel comparison of a fixed combination of bi-

matoprost 0.03%/timolol 0.5% with non-fixed combination use tion ophthalmic solution. Am J Ophthalmol 2005; 140: 1-7. in patients with glaucoma or ocular hypertension. Eur J Oph- 27. Schuman JS, Katz GJ, Lewis RA, et al. Efficacy and safety thalmol 2007; 17: 53-62. of a fixed combination of travoprost 0.004%/timolol 0.5% 15. Palmberg P, Kim EE, Kwok KK, Tressler CS, Canada and United ophthalmic solution once daily for open-angle glaucoma or States Fixed Combination Latanoprost/Timolol Study Group. A ocular hypertension. Am J Ophthalmol 2005; 140: 242-50. 12-week, randomized, double-masked study of fixed combi- 28. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. nation latanoprost/timolol versus latanoprost or timolol mono- Preferred reporting items for systematic reviews and meta- therapy. Eur J Ophthalmol 2010; 20: 708-18. analyses: the PRISMA statement. BMJ 2009; 339: b2535. 16. Higginbotham EJ, Olander KW, Kim EE, Grunden JW, Kwok 29. Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre KK, Tressler CS, United States Fixed-Combination Study SR, Hedges Team. Optimal search strategies for retrieving Group. Fixed combination of latanoprost and timolol vs indi- scientifically strong studies of treatment from Medline: ana- vidual components for primary open-angle glaucoma or ocular lytical survey. BMJ 2005; 330: 1179. hypertension: a randomized, double-masked study. Arch Oph- 30. Jadad AR, Moore RA, Carroll D, et al. Assessing the qual- thalmol 2010; 128: 165-72. ity of reports of randomized clinical trials: is blinding neces- 17. Magacho L, Reis R, Shetty RK, Santos LC, Avila MP. Efficacy of sary? Control Clin Trials 1996; 17: 1-12. latanoprost or fixed-combination latanoprost-timolol in patients 31. Schulz KF, Altman DG, Moher D. Allocation concealment in switched from a combination of timolol and a nonprostaglandin clinical trials. JAMA 2002; 288: 2406-9. medication. Ophthalmology 2006; 113: 442-4. 32. Jüni P, Egger M. Allocation concealment in clinical trials. 18. Konstas AG, Boboridis K, Tzetzi D, Kallinderis K, Jenkins JN, JAMA 2002; 288: 2407-9. Stewart WC. Twenty-four-hour control with latanoprost-timolol- 33. Follmann D, Elliott P, Suh I, Cutler J. Variance imputation for fixed combination therapy vs latanoprost therapy. Arch Oph- overviews of clinical trials with continuous response. J Clin thalmol 2005; 123: 898-902. Epidemiol 1992; 45: 769-73. 19. Olander K, Zimmerman TJ, Downes N, Schoenfelder J, Xala- 34. Olkin I. Statistical and theoretical considerations in meta- com/Latanoprost Study Group. Switching from latanoprost analysis. J Clin Epidemiol 1995; 48: 133-47. to fixed-combination latanoprost-timolol: a 21-day, random- 35. Konstas AG, Mikropoulos DG, Embeslidis TA, et al. 24-h In- ized, double-masked, active-control study in patients with traocular pressure control with evening-dosed travoprost/ glaucoma and ocular hypertension. Clin Ther 2004; 26: timolol, compared with latanoprost/timolol, fixed combina- 1619-29. tions in exfoliative glaucoma. Eye 2010; 24: 1606-13. 20. Higginbotham EJ, Feldman R, Stiles M, Dubiner H, Fixed 36. Tabet R, Stewart WC, Feldman R, Konstas AG. A review of Combination Investigative Group. Latanoprost and timolol additivity to prostaglandin analogs: fixed and unfixed com- combination therapy vs monotherapy: one-year randomized binations. Surv Ophthalmol 2008; 53(suppl 1): S85-92. trial. Arch Ophthalmol 2002; 120: 915-22. 37. Higginbotham EJ. Considerations in glaucoma therapy: 21. Pfeiffer N, European Latanoprost Fixed Combination Study fixed combinations versus their component medications. Group. A comparison of the fixed combination of latanoprost Clin Ophthalmol 2010; 4: 1-9. and timolol with its individual components. Graefes Arch Clin 38. Webers CA, Beckers HJ, Zeegers MP, Nuijts RM, Hendrikse Exp Ophthalmol 2002; 240: 893-9. F, Schouten JS. The intraocular pressure-lowering effect of 22. Diestelhorst M, Almegård B. Comparison of two fixed com- prostaglandin analogs combined with topical β-blocker ther- binations of latanoprost and timolol in open-angle glaucoma. apy: a systematic review and meta-analysis. Ophthalmology Graefes Arch Clin Exp Ophthalmol 1998; 236: 577-81. 2010; 117: 2067-74. 23. Konstas AG, Pikilidou MI, Tsironi S, et al. 24-hour intraocular 39. Aptel F, Cucherat M, Denis P. Efficacy and tolerability of pressure and blood pressure levels with latanoprost/timolol prostaglandin analogs: a meta-analysis of randomized con- fixed combination versus timolol. Curr Eye Res 2009; 34: trolled clinical trials. J Glaucoma 2008; 17: 667-73. 369-77. 40. Quaranta L, Riva I, Russo A, et al. Prostaglandin analogues 24. Konstas AG, Lake S, Economou AI, Kaltsos K, Jenkins JN, (PGAs) and timolol fixed combination (Fc) vs. extemporane- Stewart WC. 24-Hour control with a latanoprost-timolol fixed ous combination (Ec) or monotherapy (Mt) in the treatment combination vs timolol alone. Arch Ophthalmol 2006; 124: of primary open-angle glaucoma (POAG) and ocular hyper- 1553-7. tension (OHT): a systematic review and meta-analysis. In- 25. Konstas AG, Mikropoulos D, Haidich AB, Ntampos KS, Stew- vest Ophthalmol Vis Sci 2011; 52: E-Abstract 215. art WC. Twenty-four-hour intraocular pressure control with the 41. Konstas AG, Holló G, Mikropoulos D, et al. Twenty-four- travoprost/timolol maleate fixed combination compared with hour intraocular pressure control with bimatoprost and the travoprost when both are dosed in the evening in primary open- bimatoprost/timolol fixed combination administered in the angle glaucoma. Br J Ophthalmol 2009; 93: 481-5. morning, or evening in exfoliative glaucoma. Br J Ophthal- 26. Barnebey HS, Orengo-Nania S, Flowers BE, et al. The safety mol 2010; 94: 209-13. and efficacy of travoprost 0.004%/timolol 0.5% fixed combina-

18 © 2011 Wichtig Editore - ISSN 1120-6721 Copyright of European Journal of Ophthalmology is the property of Wichtig Editore and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.