Objective
To assess quality of life (QOL) as measured by patient-reported outcomes (PRO) within the iStent inject ® pivotal trial.
Design
Randomized controlled trial analysis of secondary outcomes.
Methods
The Vision Function Questionnaire (VFQ-25) and Ocular Surface Disease Index (OSDI) questionnaire were administered at baseline and at months 1, 6, 12, and 24. PRO responders were defined as patients reaching improvement based on minimally important differences.
Results
A total of 505 patients were randomized (386 iStent inject ® [Glaukos], 119 surgery alone). The iStent inject ® group exhibited a greater percentage of PRO responders across all follow-up visits over 24 months, averaging 58.0% vs 45.8%; P < .05 for VFQ-25 composite scores and 56.7% vs 48.9%; P < .05 for OSDI composite scores. Odds of being a responder in the iStent inject ® group was 60% ( P < .05) higher for the VFQ-25 and 32% ( P > .05) higher for the ODSI. Driving (49.0% vs 28.8%; P < .05), ocular pain (59.3% vs 47.2%; P < .05), and general vision (71.8% vs 60.0%; P < .05) were the VFQ-25 subscales responsible for differences between treatment groups. At month 24, 76.5% of VFQ-25 responders and 62.5% of nonresponders were medication free regardless of treatment group (P < .05).
Conclusions
Exploratory analysis suggests that by reducing medication dependence, implantation with the micro-scale iStent inject ® device with cataract surgery may improve QOL vs cataract surgery alone over 24 months, with improvements influenced by ocular symptoms and vision-related activities.
P rimary open-angle glaucoma (POAG) is characterized by progressive, irreversible, and largely asymptomatic vision loss at early stages caused by optic nerve damage. Cataract, a clouding of the lens in the eye, progressively blocks light from entering the eye, causing vision loss and, if not removed, eventually reversible blindness. These conditions often simultaneously progress slowly over extended periods of time and often have a detrimental impact on patients’ vision-related quality of life (QOL). Cataract and glaucoma surgeries are commonly performed in combination to achieve 2 meaningful benefits: (1) clearance of crystalline lens opacity from the visual axis and (2) the reduction of intraocular pressure (IOP). In treating patients with POAG, however, maintaining and improving patients’ QOL remains an unmet need. This is because worsening of a patient’s QOL, particularly as it relates to eye comfort and fluctuating vision, is commonly associated with medical treatment using ocular hypotensive medications known to cause ocular surface disease (OSD) and with glaucoma surgical interventions. Several studies have shown the deleterious impact of OSD on QOL among patients with glaucoma. It has been also observed that patients with glaucoma filtering blebs may experience discomfort in the form of burning, foreign body sensation, tearing, pain, or vague discomfort , and that those with glaucoma filtering blebs experience significantly more dysesthesia than those without filtering blebs. Therefore, a range of outcomes including QOL should be considered to assess the quality or success of glaucoma treatment.
Owing to the gradual clinical progression of POAG, QOL assessment measures of mild-to-moderate-severity patients, who are typically asymptomatic, may not be highly sensitive to glaucomatous changes. Indeed, other randomized clinical trials specifically designed to detect QOL differences among patients with early-stage glaucoma have failed to show significant differences among treatment groups. Specifically, the LiGHT (Laser as Initial Treatment of Glaucoma and Ocular Hypertension) study comparing primary selective laser trabeculoplasty with eye drops for initial therapy showed no significant difference using a generic utility measure (EQ-5D) or glaucoma-specific measure (GUI). In the CIGTS (Collaborative Initial Glaucoma Treatment Study), which aimed to determine whether patients are managed better by initial treatment with IOP-lowering medications or by immediate trabeculectomy, both groups had very similar effects on QOL. , Another trial, the United Kingdom Glaucoma Treatment Study (UKGTS), showed that vision-related quality of life, assessed by 15-item Glaucoma Quality of Life and 9-item Glaucoma Activity Limitation, was similar for both groups of patients (ie, patients receiving a topical prostaglandin analogue or placebo eye drops). Interestingly, the EAGLE study, which compared clear-lens extraction with laser peripheral iridotomy and topical medical treatment as first-line therapy in those with newly diagnosed primary angle closure with raised IOP or primary angle-closure glaucoma, found EQ-5D, VFQ-25, and GUI scores at 36 months significantly favored the clear-lens extraction group. Although this study was conducted in angle closure glaucoma patients, usually a more severe and more symptomatic form of glaucoma than POAG, it demonstrates the impact of lens extraction on QOL.
The pivotal study of cataract surgery plus iStent inject ® (Glaukos Corporation, San Clemente, California, USA) for the reduction of IOP in patients with glaucoma and cataracts, which led to the FDA’s approval of iStent inject ® in 2018, included patient-reported outcome (PRO) measures in addition to standard efficacy and safety measures. These PRO data were gathered for exploratory analyses and were not intended for the trial endpoints. Although iStent inject ® has been studied extensively and is in widespread use worldwide because of its proven clinical efficacy and safety, changes in QOL within and between treatment groups and the assessment of outcomes from the patients’ perspective, such as QOL, are yet to be fully elucidated.
Therefore, the post hoc exploratory analyses presented in this study investigated the comparative QOL changes of cataract surgery plus implantation of the trabecular micro-bypass stent (iStent inject ®) vs cataract surgery alone from patients within the pivotal study of cataract surgery with iStent inject ®.
Methods
Data Source
Details of the pivotal trial of cataract surgery plus iStent inject ® for the reduction of IOP in patients with glaucoma and cataracts are described elsewhere. The study was designed to assess the safety and effectiveness at 2 years postoperative of the second-generation trabecular micro-bypass stent system in patients with mild-to-moderate POAG and cataract at 41 US clinical sites. A targeted total of 500 eyes were randomized in a 3:1 ratio to the treatment group or control group after completion of uncomplicated cataract surgery. At baseline, month 11, and month 23, patients using ocular hypotensive medication(s) were instructed to undergo medication washout to permit unmedicated diurnal IOP (DIOP) assessment at months 12 and 24, respectively. The primary effectiveness endpoint, a 20% reduction from baseline in DIOP without ocular hypotensive medication at month 24, was met: a greater percentage of patients in the iStent inject ® group achieved a 20% or more reduction in medication-free IOP at 24 months compared with the surgery-alone group (75.8% vs 61.9%, P = .005). The study was conducted with Institutional Review Board approval, and study procedures followed the tenets of the Declaration of Helsinki (2008), including written informed consent of all participating subjects. The study was registered with the National Library of Medicine (clinicaltrials.gov, NCT00323284).
In this trial, 3 PRO measures, the Visual Function Questionnaire (VFQ-25), Ocular Surface Disease Index (OSDI) questionnaire, and the Patient Health Questionnaire-9 (PHQ-9) were administered to patients at baseline and at months 1, 6, 12, and 24. The VFQ-25 and OSDI were identified as most relevant to the patient experience and so data from these PRO measures were analyzed. Data from the Safety population, defined as all randomized subjects analyzed according to the treatment actually received, were included in the analysis.
Patient-Reported Outcomes Measures: VFQ-25 and OSDI
The VFQ-25 is a widely-used vision-related QOL instrument. It consists of 25 questions that measure the impact of vision-specific symptoms on visual functioning, disability, and general health and well-being. The patients selected answers from among multiple possible responses to the individual questions, and their responses were then converted to a 0-to-100 score, wherein higher scores indicate better functioning. Scores for vision-related questions were averaged to 12 subscale scores, and an overall composite score was calculated by averaging the subscale scores (excluding the general health rating subscore) for ocular pain, general vision, near and distance activities, mental health, social functioning, role difficulties, dependency, driving, color vision, and peripheral vision.
The OSDI consists of 12 questions used to measure the severity and effects of ocular surface symptoms related to chronic dry eye diseases and their impact on ability to function. Responses to the individual questions were rated on a scale of 0 to 4 (0 = none of the time, 1 = some of the time, 2 = half of the time, 3 = most of the time, 4 = all the time). An overall composite score was calculated based on the total number of questions answered, and 3 subscale scores were computed using the questions that comprise the subscale. The composite and subscale scores ranged from 0 to 100, with higher scores indicating greater disability.
Definition of Patient-Reported Outcomes Responders
Following the publication of the FDA’s Guidance for Industry Patient-Reported Outcome Measures in 2009, there has been a shift in approach and methodology in studies of PRO. This trend has centered on PRO responders who meet a certain threshold (or a variety of thresholds) for PRO measures, which can be interpreted as a treatment benefit, and the proportions of such individuals in each trial arm. Ideally this within-person threshold of meaningful change (ie, the amount of within person difference observed in a PRO measure that can be interpreted as a treatment benefit) would be determined using existing literature or anchor-based methods.
Because no specific within-person threshold or minimum important difference (MID) has been established in the mild-to-moderate glaucoma population, a MID was calculated for the scores from each questionnaire using a distributional approach whereby the MID was defined as half the standard deviation (SD) of the baseline scores combined across treatment groups. This approach was based on a literature review by Norman and associates (2003), which showed that the threshold of discrimination for changes in health-related quality of life was approximately half of 1 SD. Because higher scores on the VFQ-25 reflect better functioning, PRO responders for the VFQ were defined as patients with change from baseline greater than or equal to the MID threshold. For the OSDI, because higher scores reflect greater disability, responders were defined as patients with a change from baseline less than or equal to the MID. , This PRO responder approach, which is the primary focus of our analysis, differs from a group-level MID used to evaluate treatment benefit.
Medication-Free Status at Month 24 by Patient-Reported Outcomes Responders
The pivotal trial data showed 63.2% of iStent inject ® treatment eyes vs 50.0% of control eyes were medication-free at month 24. To assess association between glaucoma-related medications and PRO responders, the treatment groups were pooled and the proportions of medication-free individuals at month 24 were calculated and compared by PRO response. This allowed comparison of the rate of PRO responders who were medication-free vs those who were medicated, regardless of treatment assignment.
Statistical Analysis
For baseline demographic and clinical characteristics, comparisons were made using t tests for mean age; χ 2 tests were used for age category, sex, race, number of hypotensive medications, and prostaglandin use.
The percentages of VFQ-25 and OSDI PRO responders at the 1-, 6-, 12-, and 24-month follow-up visits were compared between treatment groups using a logistic generalized linear mixed model for repeated measures with stratification by baseline IOP (<25 mm Hg and ≥25 mm Hg) and an unstructured covariance matrix to estimate the correlation among within-subject repeated measurements. The percentages of responders within each treatment group, the percentage differences between treatment groups, and P values for the comparison between treatment groups were reported at each follow-up visit.
A χ 2 test was used to compare the percentage of patients free of medication at month 24 by PRO response.
To identify the factors associated with the likelihood of being a composite score responder for either the VFQ-25 or the OSDI, 2 mixed effects logistic models were explored: Model 1, which included fixed categorical effects for treatment group (iStent inject ® or surgery alone), baseline IOP (<25 mm Hg or ≥25 mm Hg), and baseline prostaglandin use (yes or no); and Model 2, which included the additional fixed categorical effects for age (≥60 or <60), sex (male or female), and race (black or non-black). The odds ratios (OR) and 95% confidence intervals (95% CI) for being a composite responder vs a nonresponder were summarized for each of the fixed effects for each model.
Because the analyses performed were post hoc and exploratory in nature, and there was no adjustment for multiplicity, all P values resulting from these analyses should be interpreted as descriptive statistics.
Results
Patient Baseline Demographic and Clinical Characteristics
At 41 US clinical sites, a total of 505 eyes were randomized 3:1 to cataract surgery with iStent inject ® implantation or cataract surgery only. Their baseline demographic and clinical characteristics were well balanced. The mean medicated IOP and unmedicated DIOP were 17.5 mm Hg and 24.8 mm Hg in the iStent inject ® group and 17.5 mm Hg and 24.5 mm Hg in the cataract-only group. Mean screening visual field mean deviation was -3.4 decibels in both groups. The treatment groups were also reasonably well balanced by baseline age (iStent inject ® group mean age = 69.0 years, cataract-only group mean age = 70.1 years), sex (iStent inject ® group 58.1% female, cataract-only group 54.2% female), number of ocular hypotensive medications at screening with 57.9% of the iStent inject ® group and 60.2% of the cataract-only group receiving on average 1 medications preoperatively. With regard to the preoperative ocular hypotensive medication class used at screening, 85.0%, 35.1%, 19.6%, and 19.1% of the iStent inject ® group and 79.7% 40.7%, 17.8%, and 16.1% of the cataract-only group were using prostaglandin analogue, beta-blockers, alpha agonists, and carbonic anhydrase inhibitors, respectively. In the iStent inject ® group, 41.3% of patients had a history of ocular surface disease while 43.2% in the cataract-only group did. The baseline characteristics of the Safety population (iStent inject ® group n = 386 and cataract surgery only n = 119) were comparable to those of the Intent-to-Treat population (iStent inject ® group n = 387 and cataract surgery only n = 118). Additional details of the baseline trial data are described elsewhere.
Comparison of Patient-Reported Outcomes Responders by iStent inject ® vs Cataract Only
Overall, compared with the cataract surgery alone group, the iStent inject ® group exhibited a greater percentage of PRO responders during all follow-up visits, averaging 58.0% vs 45.8%, P < .05) for the VFQ-25 composite scores and 56.7% vs 48.9% ( P < .05) for the OSDI composite scores over 24 months.
Treatment difference favored the iStent inject ® group of 14.1% at month 1, 13.4% at month 6, 11.6 % at month 12 ( P < .05), and 9.0% at month 24 ( P > .05). Over 24 months, the percentage of PRO responders averaged 58.0% in the iStent inject ® group compared to 45.8% in the cataract surgery–only group, a treatment difference of 12.1% ( P < .05).
When measured by the OSDI composite score, steady PRO improvements were observed within each group at all time intervals, with an average treatment difference favoring the iStent inject ® group of 7.8% over 24 months ( P < .05).
Although the trial measured the PRO at 5 time intervals over the course of 2 years, the majority of patients completed the 2 questionnaires consistently at each time interval, with completion rates to the questionnaire being similar in each group ( Figures 1 and 2 ). Completion rates were high and ranged from 91.6% to 99.2% across both instruments and all time points, and as such, there was no need to further investigate the impact of missing data, as they were accounted for in the analyses by the mixed models for repeated measures using the correlation among within-patient repeated measures.
Percentage of Responders Among the VFQ-25 and OSDI Subscales
Differences in PRO response rates at month 24 among the 12 VFQ-25 subscales ( Figure 3 and Table 2 ) were most pronounced in the general vision, ocular pain, and driving subscales ( P < .05). For both groups, the general vision subscale, followed by ocular pain, had the highest proportions of PRO responders. However, the iStent inject ® group had a greater proportion of PRO responders than the cataract-only group (71.8% vs 60.0%; P < .05 for general vision and 59.3% vs 47.3%; P < .05 for ocular pain). The driving subscale produced the largest difference between the 2 groups (49.0% vs 28.8%; P < .05), a 20.2% treatment difference favoring iStent inject ®.
