PURPOSE
To systematically review, synthesize and quantify studies reporting patterns of adherence and persistence with prostaglandin analogues (PGAs) in order to comprehensively understand real-world treatment behavior among patients with glaucoma who are prescribed PGAs. These data can inform the decision between a glaucoma therapy based on either topical PGA medications or procedural PGA-based intervention.
DESIGN
Systematic literature review (SLR) and meta-analysis (MA).
METHODS
We updated a 2011 SLR using electronic database searches (MEDLINE, Embase and the Cochrane Database of Systematic Reviews [CDSR]), supplemented by hand searches of SLR and MA bibliographies. We included observational studies conducted in adult glaucoma patients treated with PGAs that reported objective measures of persistence or adherence. In addition to estimated rates of adherence and persistence, adherence among patients on any/unspecified PGA was characterized by mean MPR/PDC (medication possession ratio/proportion of days covered, where values >80% indicate good adherence). Duration of therapy was defined as the time period between initial therapy prescription and time of therapy discontinuation or switch.
RESULTS
The SLR included 50 publications reporting on 47 unique studies and involving 961,000 patients. The subsequent MA included all but four studies which did not report the age distribution of patients. The mean proportion of patients on any/unspecified PGA who were adherent at Year 1 was 44% (95% CI: 31%-58%). Among patients on any/unspecified PGA, the mean MPR/PDC was 54% (95% CI: 38%-75%) at Year 1 and 60% at Year 2 (95% CI: 39%-94%). The mean proportion of patients on any/unspecified PGA who were persistent fell from 75% (95% CI: 66%-85%) at Month 6 to 31% (95% CI: 12%-55%) at Year 3, with a smaller decrease observed between Year 1 (56%; 95% CI: 45%-66%) and Year 2 (53%; 95% CI: 45%-62%), and a larger decrease between Years 2 and 3. The mean duration on therapy was 315.7 days (95% CI: 190.0%-441.5 days).
CONCLUSIONS
Suboptimal adherence and persistence with PGAs are common, with further decreases as duration lengthens. These findings may underscore the value of procedural glaucoma treatments that do not depend on daily patient engagement with topical medications.
G laucoma is a leading cause of permanent vision loss worldwide and is associated with a reduced quality of life. Topical hypotensive eyedrops are often utilized as first-line treatment for ocular hypertension (OHT) and glaucoma, with lifelong therapy required to prevent disease progression. Among topical medications, prostaglandin analogues (PGAs) are the most commonly used due to their safety and efficacy in lowering intraocular pressure (IOP) with convenient once-daily administration, while other options have more safety concerns, poorer tolerance, and/or require more frequent dosing. Although popular as first-line therapy, topical medications have several key limitations that make them suboptimal including ocular surface damage, ocular and periocular side effects, increased risk of future surgical failure, IOP fluctuations, visual field progression, diminished quality of life, systemic side effects, and poor adherence.
Poor adherence to hypotensive eyedrops is a major clinical challenge and has been reported in up to 80% of patients with glaucoma. , Early-stage glaucoma and OHT are asymptomatic, so patients with these conditions may be less inclined to maintain use of ocular hypotensive therapy since medication side effects can occur and no improvement in vision is experienced. In addition, poor adherence may arise due to various other factors, including complex dosing regimens that place high demands on patients’ daily routines, memory impairment or forgetfulness, medication costs, or an inability to adequately instill a drop into the eye. , , Age is also a key factor impacting adherence, with an adherence peak in middle to older ages and lower adherence in very young and very old patients.
Lower adherence in turn can lead to increased IOP fluctuations, which may contribute to visual field progression. In recognition of the limitations of medications, there is growing interest in a paradigm shift away from patient-administered topical medical therapy and toward a procedural approach—termed interventional glaucoma —in which therapies such as procedural pharmaceuticals, lasers such as selective laser trabeculoplasty (SLT), and procedures such as minimally invasive glaucoma surgeries (MIGS) are administered by physicians and providers, obviating adherence issues.
Indeed, in recent years, more interventional treatment options have become available, leading to a thoughtful reevaluation of the former treatment approach of topical medications first. The emerging interventional glaucoma paradigm advocates for earlier, more proactive procedural interventions to prevent vision loss, rather than having patients remain on topical medications until visual field loss has already occurred and more invasive treatments are needed. Such procedural interventions can include SLT, as supported by the findings of the Laser in Glaucoma and Ocular Hypertension (LiGHT) trial, MIGS, and an emerging array of procedural pharmaceuticals. Novel procedural pharmaceuticals in particular have been developed to address issues with poor topical medication adherence. Sustained-release intracameral implants are an attractive option to eliminate reliance on patient action for medication delivery, as they provide a targeted, automatic and continuous 24/7 release of drugs over the long term, and can be placed via minimally invasive procedures. To date, the available procedural pharmaceuticals use PGAs. Thus, it is important to comprehensively analyze current patterns of adherence and persistence to PGAs among patients with glaucoma in order to better understand the potential benefits of these newer PGA delivery systems.
A systematic literature review (SLR) was published in 2011 by Reardon et al to summarize findings from studies assessing adherence and persistence with ocular hypotensive agents in patients with glaucoma and OHT. The objective of the current study was to update the earlier review by identifying additional, newer literature reporting on persistence and adherence to PGAs. A meta-analysis (MA) of the studies identified by the SLR was then conducted to synthesize and quantify overall patterns of adherence and persistence with PGAs to illuminate treatment behavior among patients with glaucoma who were prescribed PGAs.
METHODS
This SLR update used a prespecified protocol based on the methodology of the Reardon et al 2011 SLR. This protocol was not prospectively registered. MEDLINE, Embase, and the Cochrane Database of Systematic Reviews (CDSR) databases were searched on 24 th August 2023 via the Ovid SP and Cochrane Library platforms for studies published since 2011 (under the assumption that all relevant studies published prior to 2011 had been captured by Reardon et al in the original SLR). Full database search terms are presented in Supplementary Tables 1 and 2.
Supplementary Table 1 displays search terms for MEDLINE, MEDLINE In-Process, and Embase (searched simultaneously via the Ovid SP platform).
Supplementary Table 2 shows the search terms for CDSR (searched via the Cochrane Library Wiley Online platform). In addition, the bibliographies of any SLRs and (network) MAs ([N]MAs) identified in the database searches and which met the SLR eligibility criteria at the abstract review stage were manually hand-searched. The bibliography of Reardon et al (2011) was also hand-searched to identify studies relevant to this update.
The eligibility criteria used for the SLR update are presented in Table 1 . In brief, publications with an English-language full-text were included if they reported objective measures of persistence or adherence (also known as compliance), with or without subjective measures, from observational studies of PGAs in adults with glaucoma. Abstracts and then full-texts were assessed for inclusion by two independent reviewers, with any disagreements resolved by discussion until a consensus was met. If necessary, a third independent reviewer made the final decision. Publications reporting on the same study were linked and considered as a single unit in subsequent stages.
| Domain | Inclusion | Exclusion |
|---|---|---|
| Population |
|
|
| Intervention |
|
|
| Comparator |
|
|
| Outcomes |
|
|
| Study design |
|
|
| Publication type |
|
|
| Other considerations |
|
|
a “Adherence” and “compliance” are synonymous. “Adherence” is the preferred term used throughout the manuscript text, but for the purposes of the systematic literature review, studies using both terms (“compliance” and “adherence”) were searched.
OUTCOMES
Adherence and persistence were characterized by the proportion of patients who were adherent or persistent with therapy, respectively. Among patients on any/unspecified PGA, adherence was further characterized by mean medication possession ratio/proportion of days covered (MPR/PDC). MPR was defined as the number of days’ supply of glaucoma prescription dispensed divided by the total number of days in the specified time periods (i.e. one or two years). PDC was defined as the proportion of days covered by the glaucoma prescription over the specified time periods (i.e. one or two years). Adherent patients were defined as those who had MPR/PDC>80%.
In the measurement of persistence, persistent patients were defined as those who did not switch or discontinue their therapies, with an allowable refill gap of up to 90 days. Duration of therapy was defined as the period between initial therapy prescription and therapy discontinuation or switch.
DATA ANALYSIS
The data extraction was performed in line with guidelines from the University of York Centre for Reviews and Dissemination (CRD). Key information from eligible studies was initially extracted by a single individual into a prespecified data extraction grid. When the initial extraction was complete, a second individual verified the extracted information and checked that no relevant information had been missed. The quality of each study was assessed using the Alberta Heritage Foundation for Medical Research (AHFMR) tool for quantitative studies by a single individual, with verification by a second reviewer.
Qualitative and quantitative assessments of study designs, eligibility criteria, and patient baseline characteristics were conducted to ensure that a MA was feasible. For studies reporting relevant data only in Kaplan-Meier plots, numerical estimates were obtained by digitizing the data and generating pseudo-individual patient data (IPD) using the Guyot algorithm. Standard parametric survival curves were then fitted to the pseudo-IPD to estimate the proportions of adherent/persistent patients at the specific timepoint(s) of interest.
Analyses were performed by specific PGAs for the three most commonly studied PGAs (travoprost, latanoprost, bimatoprost) and an “any/unspecified PGA” group. All analysis steps were conducted in R (v4.3.0 onwards), using the meta package. , To account for both within-study sampling error and between-study variability, random-effects models were used to generate estimates. Between-study variance in the random effects models was estimated using the restricted maximum-likelihood estimator (REML). Summary estimates were reported as mean values with their corresponding 95% confidence intervals (CI).
Heterogeneity across the studies included in each analysis was reported using the I 2 statistic. Funnel plots were used to visualize any potential biases exhibited by studies, with a symmetrical distribution of studies falling within the funnel suggesting minimal publication bias.
RESULTS
RECORDS ANALYSIS
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram for the SLR is presented in Figure 1 . A total of 514 records were reviewed from the electronic databases, with 30 records ultimately included for data extraction. From the supplementary searches of reference lists of SLR and (N)MAs, 92 records were reviewed and one record fulfilled the eligibility criteria. Finally, 19 publications from the original SLR were relevant to the current scope and carried forward for data extraction. Ultimately, 50 publications reporting on 47 unique studies were included in the SLR, with a total of 961,000 patients enrolled ( Table 2 ). More than half ( n = 24, 51.1%) were national registry/database studies, 14 were retrospective cohort studies (29.8%), seven were prospective cohort studies (14.9%), one was a case series (2.1%) and one was a cross-sectional study (2.1%). The majority of studies ( n = 33, 70.2%) collected adherence and persistence data through prescription fill records, eight studies collected data through medical chart reviews (17.0%) and six studies collected data through electronic monitoring devices (12.8%). In the feasibility assessment, age was determined to be a critical variable to account for because older age is associated with a higher comorbidity burden which may impact an individual’s ability to adhere to and persist with treatment. Four studies did not report the age distribution of the patients and were therefore excluded from the subsequent MA of 588,935 patients, with a mean age of 67.5 years.
| Study | Study Design | Included Patient Number | Type of PGA Used | Type of Adherence Outcome Reported | Type of Persistence Outcome Reported |
|---|---|---|---|---|---|
| Electronic Monitoring Studies | |||||
| Cate et al. a , | Prospective cohort | 98 | Travoprost | Rate of adherence | NR |
| Dreer et al. 2012 | Prospective cohort | 116 | Travoprost | Rate of adherence | NR |
| Mansberger et al. a , | Case series | 58 | Travoprost | Rate of adherence; PDC | NR |
| Waterman et al. 2020 | Prospective cohort | 53 | Latanoprost | Rate of adherence | NR |
| Zarnowski et al. a , | Prospective cohort | 127 | Travoprost | Rate of adherence | NR |
| Gatwood et al. a , | Prospective cohort | 28 | Travoprost; bimatoprost | Rate of adherence | NR |
| Prescription Fill Record Studies | |||||
| Amiri et al. 2023 | National registry/ database | 10,014 | Latanoprost | PDC | Cumulative incidence of discontinuation; risk of treatment discontinuation |
| Castel et al. 2014 | Cross-sectional | 738 | PGAs (unspecified) | MPR | NR |
| Daniels et al. | National registry/ database | 67,785 | PGAs (unspecified) | NR | Treatment duration |
| Heo et al. 2019 | National registry/ database | 3,888 | Bimatoprost; latanoprost; travoprost | MPR | Rate of persistence; rate of patients switching treatment; risk of treatment switching; time to treatment discontinuation; risk of treatment discontinuation |
| Hwang et al. 2014 | National registry/ database | 3,134 | PGAs (unspecified) | NR | Rate of persistence |
| Imperato et al. 2022 | National registry/ database | 232,572 | Bimatoprost; latanoprost; travoprost | MPR | Rate of persistence |
| Iskedjian et al. 2011 | National registry/ database | 28,534 | Bimatoprost; latanoprost; travoprost | NR | Rate of persistence |
| Jang et al. 2021 | National registry/ database | 14,648 | PGAs (unspecified) | MPR | Rate of persistence; treatment duration |
| Kashiwagi et al. 2014 | National registry/ database | 2,483 | Bimatoprost; latanoprost; tafluprost; travoprost | NR | Rate of persistence |
| Na et al. 2018 | Retrospective cohort | 366 | Tafluprost | MPR | NR |
| Patel et al. 2016 | National registry/ database | 11,234 | Bimatoprost | PDC; odds of high adherence; odds of low adherence | Rate of persistence; time to treatment discontinuation; rate of patients remaining on index study treatment; odds of receiving treatment |
| Quek et al. 2011 | National registry/ database | 2,781 | Bimatoprost; latanoprost; travoprost | NR | Rate of persistence |
| Rahman et al. 2011 | National registry/ database | 1,006 | Bimatoprost; latanoprost; travoprost | NR | Time to initial treatment change; risk of changing initial treatment; time to initial treatment discontinuation; risk of initial treatment discontinuation |
| Schultz et al. 2016 | National registry/ database | 21,227 | PGAs (unspecified) | PDC | NR |
| Sheer et al. 2016 | National registry/ database | 73,256 | PGAs (unspecified) | PDC | NR |
| Shirai et al. 2021 | National registry/ database | 1,725 | Latanoprost; tafluprost; travoprost | PDC | Rate of persistence |
| Stein et al. 2015 | National registry/ database | 8,427 | Bimatoprost; latanoprost; travoprost; | PDC | NR |
| Menino et al. 2023 | Retrospective cohort | 3,548 | PGAs (unspecified) | Rate of adherence | Rate of persistence |
| 45 and Up Study | Prospective cohort | 12,899 | PGAs (unspecified) | MPR; PDC | Rate of persistence |
| Campbell et al. 2014 | National registry/ database | 12,985 | Bimatoprost; travoprost | PDC | Rate of patients remaining on study treatment |
| Mammo et al. 2020 | Retrospective cohort | 184 | PGAs (unspecified) | Rate of adherence | NR |
| Bhosle et al. 2007 | Retrospective cohort | 268 | Latanoprost | MPR | Rate of persistence |
| Reardon et al. | Retrospective cohort | 7,873 | Bimatoprost; latanoprost; travoprost | Days covered | Medication possession |
| Wilensky et al. b , | Retrospective cohort | 2,424 | Bimatoprost; latanoprost; travoprost | Proportion of days adherent | Rate of persistence |
| Yeaw et al. 2009 | Retrospective cohort | 3,310 | Bimatoprost; latanoprost; travoprost | PDC | Rate of persistence |
| Yousuf et al. 2011 | Retrospective cohort | 184 | PGAs (unspecified) | Rate of adherence | NR |
| Zimmerman et al. b , | Retrospective cohort | 6,271 | Latanoprost; bimatoprost; travoprost | NR | Rate of persistence; rate of patients switching treatment; rate of treatment discontinuation by physician; rate of patients switching or discontinuing treatment; rate of patients restarting treatment |
| De Natale et al. 2009 | Retrospective cohort | 815 | Travoprost | NR | Rate of treatment failure; time to treatment failure |
| Djafari et al. 2009 | National registry/ database | 181 | PGAs (unspecified) | Rate of adherence | NR |
| Owen et al. 2009 | National registry/ database | 5,670 | PGAs (unspecified) | NR | Rate of persistence; rate of treatment failure |
| Glaucoma adherence and persistency study a , | National registry/ database | 13,977 | Latanoprost; bimatoprost; travoprost | MPR | Rate of patients remaining on initially prescribed treatment |
| Rait et al. b , | National registry/ database | 357,099 | Latanoprost; bimatoprost; travoprost | NR | Rate of persistence; rate of treatment discontinuation |
| Reardon et al. 2004 | National registry/ database | 28,741 | Latanoprost; bimatoprost; travoprost | NR | Rate of persistence; risk of discontinuation or switching; risk of discontinuation |
| Medical Chart Review | |||||
| Arias et al. 2010 | Retrospective cohort | 148 | Bimatoprost; latanoprost; travoprost | NR | Rate of persistence |
| Day et al. 2004 | Retrospective cohort | 1,182 | Bimatoprost; latanoprost | NR | Risk of treatment discontinuation |
| Hahn et al. b , | National registry/ database | 6,271 | Bimatoprost; latanoprost; travoprost | NR | Rate of persistence; rate of patients switching therapies; rate of treatment discontinuation by physician; rate of patients restarting treatment; duration of treatment |
| Kashiwagi et al. 2010 | National registry/ database | 1,955 | Latanoprost | NR | Time on treatment |
| Kuwayama et al. 2017 | Prospective cohort | 4,265 | Tafluprost | NR | Rate of persistence |
| Lanza et al. 2022 | Retrospective cohort | 190 | PGAs (unspecified) | NR | Duration of treatment |
| Maccabi Glaucoma Study | National registry/ database | 5,934 | Prostaglandin analogues (unspecified) | NR | Time to discontinuation |
| Nakakura et al. 2021 | Retrospective cohort | 328 | Latanoprost | NR | Rate of persistence; time to treatment discontinuation; duration of persistence |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
