The purpose of this study is to establish consensus among experts on outcomes and methods to be used in clinical trials to assess adverse effects of anti-glaucomatous eye drops.
Modified Delphi method.
Clinical experts from Europe, North America, South America, the Middle East, and Asia were invited to participate in 2 sequential web-based surveys administered from June 27 to August 29, 2021. A total of 91 clinical experts were invited to participate. Of these, 71 (78%) experts from 23 different countries accepted the invitation and answered the first questionnaire. The importance of items was ranked using a 10-point scale (1 as not important, 10 as very important).
A total of 84 items were rated in round one by 71 participants. Of these, 68 (81%) reached consensus. In round 2, 19 items, including 3 additional items, were rated by 53 (75%) participants. Consensus was reached in 98% of investigated items. Eight outcomes were agreed as important to assess when conducting future trials: ocular surface, dryness, epithelial damage, local adverse effects related to eye drops as reported by patients, periocular surroundings and eyelids, quality of life questionnaires, hyperemia, visual acuity, tear film, and anterior chamber inflammation.
We propose a consensus-based series of outcomes and assessment methods to be used in clinical trials assessing adverse effects of antiglaucomatous eye drops. This, we hope, will improve the comparability of results from future trials and thus facilitate meta-analyses and progress in this field.
G laucoma is the leading cause of irreversible blindness worldwide, affecting approximately 80 million people. With a growing elderly population, this number is estimated to increase to 111 million by 2040. The only current treatment is lowering the intraocular pressure (IOP), which is most often done with eye drops. For most patients, treatment is lifelong and the long-term use of topical ocular medications can be associated with adverse events such as redness, irritation, feeling of dryness, burning and stinging, discomfort, tearing, itching, and foreign body sensation as seen in dry eye disease. The prevalence of dry eye disease in the general elderly population is approximately 15% compared to 60% in patients with glaucoma. The severity of local AEs has been shown to affect adherence and thus potentially the efficacy of antiglaucomatous eye drops. , It is therefore important to take AEs into account when treating patients with glaucoma.
Clinical trials studying topical treatment for glaucoma show diverse reporting of safety measures, which makes comparisons between treatments challenging and prevents indirect comparisons that, ultimately, leads to missed opportunities for potentially synthesizing evidence and informed decision-making. Examples include the study by Hedengran and associates evaluating the efficacy and safety of benzalkonium chloride (BAK)–preserved eye drops compared with alternatively preserved and preservative-free eye drops in the treatment of glaucoma and the study by Steensberg and associates evaluating the efficacy and safety of generic vs original prostaglandin analogues in the treatment of glaucoma. ,
Inconsistency in outcomes and methods prevented the studies from synthesizing evidence to address current concerns on the safety profiles of preserved eye drops and generic substitutions. Furthermore, inconsistencies hindered detailed AE comparisons that are essential for optimal and individualized patient care. In the current study, we wanted to establish consensus on outcomes and methods that could be recommended for future trials evaluating the safety of antiglaucomatous eye drops.
We used the Delphi method, an iterative process to establish consensus on a given subject. The Delphi method is characterized by a controlled feedback cycle where participants are presented with the median responses of the collected group compared to their own previous responses and are hereby encouraged to reevaluate their initial responses.
A core group of leading clinical experts in the fields of glaucoma and ocular surface was established. Core members included researchers experienced in dry eye disease, glaucoma management, and the pathology and biology of ocular surface damage associated with glaucoma medications. Core group members were identified through previous work in the mentioned fields. Members were required to have clinical trial experience, preferably as chief investigators, and to hold a faculty position equivalent to associate professor or higher. Each core member recommended 8 to 10 international glaucoma or ocular surface experts who were invited to participate in the Delphi process.
The study aimed for a majority of clinicians with clinical trial experience but also accepted scientists with relevant expertise in the participant group. Delphi participants were required to hold a medical degree with clinical experience in relevant fields and preferably had clinical trial experience. Participants without direct clinical trial experience were recommended by core group members based on relevant experience demonstrated through previous collaborations.
Domains and methods were initially extracted from clinical trials included in the systematic review and meta-analysis performed by Hedengran and associates. All mentioned domains and methods in the studies were included. , , Quality of life questionnaires were extracted from a literature review by Okumura and associates. Items were then augmented and altered by the core group. Items were translated into survey-format questions and organized into outcome domains, for example, visual acuity, with subgroups concerning measurement method, for example, Snellen chart.
Participants were asked to rank outcomes and methods on a scale from 1 to 10 based on importance (1 as not important and 10 as very important). Consensus on an outcome being important was achieved when the mean value of an item was 7 or higher and the interquartile range (IQR) of scores was 3 or lower. Consensus in 90% of the investigated domains was considered a valid status for conclusion. Items marked “no preference or unfamiliar with method” or not answered by more than 50% of the experts were concluded as “no preference.”
Invitations were sent to potential participants via email with an individual link to the REDCap questionnaire. Each survey round was conducted over 2 to 4 weeks. The study aimed for 2-4 Delphi rounds. In the first round, participants were asked to rank items on a scale from 1 to 10 and were given the opportunity to comment and suggest additional outcomes or methods. The mean, median, and IQR of each item were calculated. Items with interquartile range (IQR) of scores greater than 3 (ie, lack of consensus) were included in the following round. Items were then presented with the median response from the collected group and the individual’s own response from the previous round. Experts were also asked to rate new items suggested by participants in the previous round.
The Delphi study was conducted from June 27 to August 29, 2021, using Research Electronic Data Capture (REDCap) hosted by the University of Copenhagen. Only 2 rounds were required to reach consensus.
A total of 91 clinical experts were invited to participate. Of these, 71 (78%) experts from 23 different countries ( Figure 2 ) accepted the invitation and answered the first questionnaire. Only participants who completed round 1 were invited to participate in round 2. Of the invited participants, 53 (75%) answered the second questionnaire. Participants were evenly distributed in the fields of glaucoma (55%) and ocular surface (45%). The vast majority of participants were clinical experts, of whom 56 (79%) had been directly involved in clinical trials. For geographical distribution, see Figure 2 .
In the first round, participants were asked to rate 84 items, of which 81% reached consensus. In the second round, participants were asked to rate 19 items, 3 of which were new items suggested by the participants in the first round, and the remaining 16 items that did not reach consensus in the first round. Of the investigated domains, 98% had reached consensus after round 2, and the study was concluded ( Figure 1 ).
Experts agreed that the following outcomes were important when assessing AEs in clinical trials studying topical treatment for glaucoma (from highest to lowest mean score): ocular surface, dryness, epithelial damage, local AEs related to eye drops as reported by patients, periocular surroundings and eyelids, quality of life (QoL), hyperemia, visual acuity, tear film, and anterior chamber inflammation. One or more methods were recommended for each outcome except hyperemia, where there was no preference for any of the presented methods. The full results of all examined domains and methods are available in Table 1 .
|Domain and Method||Conclusion a||Mean||% of Participants With No Preference b|
|Landolt C||No preference||7.4||55|
|Periocular surroundings and eyelids||Important||7.9||0|
|Meibomian gland dysfunction|
|Meibum quantity and quality||Important||7.8||17|
|Eyelid margin foam||Not important||6.6||26|
|Abnormal eyelid margin findings of vascularity||Important||7.5||25|
|Plugging of gland orifices||Important||7.6||14|
|Eyelid margin irregularity||Important||7.1||25|
|Eyelid margin thickening||Important||7.4||17|
|Partial glands||Not important||6.4||28|
|Morphology of meibomian glands by noninvasive meibography, “Meiboscore”||Important||7.0||32|
|2% fluorescein staining||Important||8.7||14|
|Noninvasive TBUT with keratometry||Important||7.0||32|
|Noninvasive TBUT with Tearscope||Not important||6.7||39|
|Infrared thermography||Not important||5.2||48|
|Blink rate (blink/minute)||Important||7.0||23|
|Tear film production|
|Schirmer 1 (without anesthetics)||Important||7.0||23|
|Schirmer 2 (Schirmer 1 but with nasal provocation)||Inconclusive (IQR higher than 3)||3.5||32|
|Basic secretion (with anesthetics)||Not important||5.2||26|
|Tear film clearance test||Not important||5.8||23|
|Height of tear film meniscus, µm||Not important||6.3||28|
|Tear film quality|
|Tear film osmolality||Not important||6.5||27|
|Tear film component: Inflammatory markers, eg, cytokines||Important||7.2||27|
|Tear film component: Mucins||Not important||6.9||31|
|Tear film component: Lipids||Important||7.0||28|
|Tear film component: Microbiome||Not important||6.3||34|
|Ocular surface, dryness, epithelial damage||Important||8.3||0|
|Efron system||No preference||6.5||53|
|Annunziato system||No preference||5.4||78|
|Cornea and Contact Lens Research Unit (CCLRU) grading scales||No preference||6.4||65|
|Vistakon system||No preference||5.7||79|
|Oxford grading scale||Important||7.7||34|
|National Eye Institute grading system (NEI)||Important||7.5||44|
|Lissamine green, van Bijsterveld score||Not important||6.3||38|
|Lissamine green, SICCA Ocular Staining Score (OSS)||Not important||6.6||28|
|Fluorophotometry||Inconclusive (IQR >3)||4.2||49|
|McMonnie and Chapman-Davies (MC-D) scale||No preference||6.6||69|
|The Institute for Eye Research (IER) scale for bulbar redness||No preference||7.1||65|
|The Validated Bulbar Redness (VBR 5) scale||No preference||6.9||68|
|Japan Ocular Allergy Society (JOAS) conjunctival hyperemia severity grading||No preference||6.3||78|
|CCLRU grading scales||No preference||6.0||68|
|Efron scale||No preference||7.2||65|
|Annunziato pictorial||No preference||6.4||79|
|Vistakon-Synoptik photographic grades||No preference||6.0||79|
|Jenvis grading scale, bulbar redness||No preference||6.4||78|
|Jenvis grading scale, limbal redness||No preference||6.4||78|
|Cell density||Not important||6.0||7|
|Modified Nelson’s grading system (goblet cells)||Not important (as domain is not important)||7.2||24|
|Staging of conjunctival squamous metaplasia||Not important (as domain is not important)||6.1||23|
|In vivo confocal microscopy, conjunctival||Not important (as domain is not important)||6.6||18|
|In vivo confocal microscopy, corneal||Not important (as domain is not important)||7.1||16|
|Anterior chamber inflammation||Important||7.2||26|
|Laser-flare meter||Not important||6.2||42|
|Slit lamp, Standardization of Uveitis Nomenclature (SUN) grading||Important||7.8||32|
|Patient-reported local adverse events related to eye drops||Important||8.0||3|
|Foreign body sensation||Important||8.6||16|
|Irritation, burning, stinging||Important||9.0||14|
|Symptoms associated with orbital fat atrophy||Not important||6.8||23|
|Quality of life questionnaires||Important||7.8||3|
|Ocular Surface Disease Index (OSDI)||Important||8.6||24|
|Impact of Dry Eye in Everyday Life (IDEEL)||No preference||6.8||62|
|Dry Eye-Related Quality-of-life Score (DEQS)||Important||7.8||45|
|University of North Carolina Dry Eye Management Scale (UNC DEMS)||No preference||6.1||80|
|Chinese version of DEQS||No preference||4.8||85|
|25-Item National Eye Institute Visual Function Questionnaire (NEI VFQ-25)||Important||7.3||42|
|The Glaucoma Symptom Scale (GSS)||Important||7.8||48|
|Standard Patient Evaluation of Eye Dryness Questionnaire (SPEED)||Important||7.4||34|