PURPOSE
To investigate whether baseline microvasculature dropout (MvD) is associated with subsequent structural and functional changes in preperimetric glaucoma.
DESIGN
Subgroup analysis of prospective cohort study data.
METHODS
This study included 93 eyes from 70 participants with a glaucomatous optic disk appearance but no repeatable visual field defects, who were followed with optical coherence tomography, optical coherence tomography angiography, and visual field (VF) over a mean of 4.9 years. Circumpapillary retinal nerve fiber layer thickness and circumpapillary capillary density (cpCD) were assessed. Eyes were grouped based on the presence or absence of MvD at baseline. Mixed-effects models were used to identify factors associated with structural and functional progression.
RESULTS
Mean age was 67.7 (95% CI, 65.4-70.0) years. Among the 93 eyes, those with baseline MvD (32 eyes) showed faster loss of cpCD (−0.88 [95% CI, −1.08 to −0.67] %/year) compared to eyes without MvD (−0.23 [95% CI, −0.42 to −0.04] %/year). In multivariable models, the presence of MvD was independently associated with faster cpCD loss (−0.63 [95% CI, −0.95 to −0.32] %/year, P <.001). In contrast, it was not associated with circumpapillary retinal nerve fiber layer loss (0.03 [95% CI, −0.33 to 0.38] µm/year, P =.886). Moreover, PPG eyes that had MvD at baseline developed VF loss more frequently than those without MvD. (62.5% vs 26.2%, P <.001).
CONCLUSIONS
Baseline MvD in PPG is associated with both faster vessel density loss and subsequent VF loss, and higher rates of developing glaucomatous visual field damage. Assessment of MvD can be used to estimate risk and guide the frequency of follow-up in patients with PPG.
INTRODUCTION
G laucoma is a chronic, progressive optic neuropathy and a leading cause of irreversible blindness worldwide, characterized by the loss of retinal ganglion cells and structural changes of the optic nerve head (ONH). Early glaucomatous damage such as neuroretinal rim loss or retinal nerve fiber layer thinning can arise before repeatable visual field (VF) damage that is detected on standard automated perimetry. Optical coherence tomography (OCT) and OCT angiography (OCTA), another diagnostic test that enables quantification of microvascular alterations in the peripapillary region and throughout the retina, can detect microvascular loss, sometimes prior to repeatable VF damage, as well. ,
Among OCTA findings, microvasculature dropout (MvD), a focal area of absent microvascular signal in the parapapillary choroid, has shown to be associated with glaucomatous damage. Although the mechanisms underlying the association between MvD and structural progression remain uncertain, superficial retinal vessel density, supplying the retinal nerve fiber layer (RNFL) and retinal ganglion cell layer, declines as glaucoma progresses. Further, loss of deep-layer vessel density, including the peripapillary choroidal vasculature around the optic disc can alter ONH perfusion and, perhaps, contribute to axonal damage. Reduced blood flow in this shared vascular network may lead to functional decline even without measurable axonal loss. ,
While previous studies have shown that faster superficial vessel density loss is associated with faster subsequent glaucomatous progression, , the temporal relationship between parapapillary microvascular abnormality and neural tissue loss at the preperimetric stage, when there is no detectable VF loss, remains uncertain. Prior studies in eyes with preperimetric glaucoma (PPG) were frequently cross-sectional or combined suspects with established glaucoma, making it difficult to determine the relationship between MvD and loss of neural tissue, and glaucoma progression.
In this study, we investigated whether the presence of MvD at baseline is associated with structural and functional changes in eyes with PPG.
METHODS
This subgroup analysis used prospective cohort study data from the Diagnostic Innovations in Glaucoma Study (DIGS) and included eyes with PPG. Written informed consent was obtained from all participants. The study protocol was approved by the Institutional Review Board of the University of California San Diego (UCSD), and all methods complied with the tenets of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act (HIPAA).
PPG was defined as the presence of glaucomatous optic neuropathy characterized by neuroretinal rim thinning or notching, disc excavation, or localized/diffuse RNFL defects confirmed by 2 experienced graders with the absence of repeatable glaucomatous VF loss. General inclusion criteria for DIGS were (1) age > 18 years; (2) open angles on gonioscopy; (3) best-corrected visual acuity ≥ 20/40 at enrollment. Exclusions comprised (1) prior ocular trauma or intraocular surgery (except uncomplicated cataract or glaucoma surgery); (2) coexisting retinal disease; (3) uveitis; (4) non-glaucomatous optic neuropathy; (5) diagnosis of systemic diseases such as Parkinson’s disease, Alzheimer’s disease, other types of dementia, or a history of stroke. Eyes with axial length ≥ 27-mm or poor-quality OCTA/OCT images were also excluded. Participants underwent biannual serial OCT and OCTA (AngioVue; Optovue, Fremont, CA) imaging. Eyes qualified for inclusion if they had at least 3 acceptable ONH OCT-A/OCT sessions over at least 1.5 years of follow-up.
ONH-centered OCT and OCTA images (4.5 × 4.5-mm 2; 304 B-scans with 304 A-scans per B-scan) were acquired simultaneously with the AngioVue system (software version 2018.1.0.43), which performed automatic segmentation with exact registration of the analyzed regions. Retinal layers were segmented automatically and circumpapillary RNFL (cpRNFL) and circumpapillary capillary density (cpCD) were calculated using the manufacturer’s software. OCTA imaging was integrated into the DIGS protocol beginning in 2015. Therefore, the baseline for all participants in the current study refers to each participant’s first acceptable OCTA scan acquired during ongoing longitudinal follow-up.
The cpCD was derived from the radial peripapillary capillary slab, the layer between the internal limiting membrane (ILM) and the posterior boundary of the RNFL. It was calculated as the proportion of the area occupied by capillaries within the annulus between an inner circle 2-mm in diameter and an outer circle 4-mm in diameter, both centered on the optic disc center automatically identified by the software. The software quantifies capillary density using a large-vessel mask designed to detect vessels measuring 3 pixels or more (approximately ≥ 33-µm). Image quality was reviewed according to the UCSD Imaging Data Evaluation and Analysis (IDEA Center) Reading-Center protocol for AngioVue. Scans were excluded if they showed: (1) scan quality < 4; (2) poor clarity; (3) residual significant motion artifacts (irregular vessel patterns or distorted disc margins on the enface angiogram); (4) cropping or localized signal loss; or (5) segmentation errors not amenable to correction.
The choroidal vessel density slab encompassed tissue below the retinal pigmented epithelium (RPE), including choroid and sclera. Choroidal MvD was defined as complete loss of the choriocapillaris without visible microvascular network within the β-zone parapapillary atrophy (PPA). The β-zone PPA was identified by the presence of Bruch’s membrane with absence of RPE, allowing visualization of choroidal vessels and sclera. Dropout was required to extend across ≥ 4 consecutive horizontal B-scans, measuring > 200-µm in diameter, and in contact with the OCT disc boundary. The disc boundary was automatically detected by AngioVue and manually corrected if needed by a trained observer masked to clinical data. MvD grading methods and inter-rater agreement are described in detail in our prior publications. , Briefly, 2 independent masked observers graded MvD with discrepancies resolved by adjudication, and inter-rater agreement was good (Kappa = 0.86-0.92). ,
Standard automated perimetry with Swedish Interactive Thresholding Algorithm (SITA) strategies (Humphrey Field Analyzer; Carl Zeiss Meditec, Dublin, CA) was performed during follow-up. Only reliable fields (fixation losses and false negatives ≤ 33%, false positives ≤ 15%) were included. VF was considered abnormal when the glaucoma hemifield test was outside normal limits or the pattern SD was outside the 95% normal confidence limits. Eyes were classified as developing VF defects or not, with VF defect defined as ≥ 2 consecutive abnormal VF examinations.
STATISTICAL ANALYSIS
Patient characteristics were summarized as mean with 95% CIs for continuous variables and as counts with percentages for categorical variables. Agreement between graders for MvD presence was quantified using Cohen’s kappa. Categorical characteristics were compared between baseline MvD groups using χ² tests. Baseline continuous characteristics and imaging metrics were compared with linear mixed-effects models to account for correlation between fellow eyes (random intercept for eye nested within subject). We were interested in whether the baseline MvD status is associated with longitudinal OCT and OCTA change. Change over time in OCT/OCTA results was analyzed with linear mixed-effects models including time as a continuous predictor and random slopes. Multivariable models included age, mean intraocular pressure (IOP) during follow-up, and other potential predictors with P <.10 in univariable model. Results are presented as coefficients (95% CI) with corresponding P values. All tests were two-sided with α = 0.05. Statistical analyses were performed using Stata, version 16.0 (StataCorp LLC).
RESULTS
Ninety-three eyes of 70 patients with PPG were included. Of these, 32 eyes (22 patients) demonstrated baseline MvD, while 61 eyes (48 patients) had no MvD ( Table 1 ). The mean age of all participants was 67.7 (95%CI, 65.4-70.0) years. Baseline cpCD was lower in the MvD group (45.6% vs 47.7%, P =.007), and baseline cpRNFL thickness was also reduced (82.7-µm vs 90.7-µm, P <.001) compared to no MvD group, while baseline IOP and VF mean deviation (MD) did not differ between groups ( P s >.05).
TABLE 1
Characteristics of Eyes Categorized by Presence of Baseline MvD.
| Variables | MvD (n =22, 32 Eyes) | No MvD (n =48, 61 Eyes) | Overall (n =70, 93 Eyes) | P Value |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age (years) | 66.2 (61.6 to 70.9) | 68.4 (65.7 to 71.1) | 67.7 (65.4 to 70.0) | .392 |
| Sex (% female) | 14 (63.6%) | 27 (56.3%) | 45 (55.6%) | .560 |
| Race (n, %) | — | — | — | .628 |
| African descent | 4 (18.2%) | 15 (31.3%) | 19 (27.1%) | — |
| Asian descent | 2 (9.1%) | 2 (4.2%) | 4 (5.7%) | — |
| European descent | 15 (68.2%) | 29 (60.4%) | 44 (62.9%) | — |
| Other/not reported | 1 (4.6%) | 2 (4.2%) | 3 (4.3%) | — |
| Self-reported hypertension, n (%) | 8 (36.4%) | 27 (56.3%) | 35 (50.0%) | .122 |
| Self-reported diabetes, n (%) | 1 (4.6%) | 8 (16.7%) | 9 (12.9%) | .160 |
| Eye characteristics | ||||
| Axial length (mm) | 24.7 (24.4 to 25.0) | 24.3 (24 to 24.7) | 24.5 (24.2 to 24.7) | <.001 |
| Spherical equivalent (D) | −1.6 (−2.4 to −0.8) | −1.3 (−1.9 to −0.7) | −1.4 (−1.9 to −0.9) | <.001 |
| CCT (µm) | 533.4 (519.1 to 547.7) | 555 (545.3 to 564.7) | 547.6 (539.4 to 555.8) | .032 |
| IOP Mean IOP during follow-up (mm Hg) | 16.6 (15.4 to 17.8) | 16.4 (15.6 to 17.3) | 16.5 (15.8 to 17.2) | .814 |
| Number of glaucoma medications at baseline, n | 2.2 (1.7 to 2.7) | 1.9 (1.5 to 2.3) | 2.0 (1.7 to 2.3) | .395 |
| Number of glaucoma medications at last visit, n | 2.2 (1.7 to 2.7) | 1.9 (1.5 to 2.3) | 2.0 (1.7 to 2.3) | .395 |
| Baseline VF MD (dB) | −0.5 (−1.1 to 0.0) | −0.3 (−0.7 to 0.1) | −0.4 (−0.7 to −0.1) | .604 |
| Baseline cpCD (%) | 45.6 (44.3 to 46.8) | 47.7 (46.9 to 48.6) | 47.0 (46.2 to 47.7) | .007 |
| Baseline cpRNFL (µm) | 82.7 (77.6 to 87.8) | 94.9 (91.4 to 98.3) | 90.7 (87.6 to 93.7) | <.001 |
| Follow-up period (years) | 5.3 (4.8 to 5.7) | 4.7 (4.3 to 5.1) | 4.9 (4.6 to 5.2) | .116 |
P Values <0.05 are shown in bold.
CCT = central corneal thickness; cpCD = circumpapillary capillary density; cpRNFL = circumpapillary retinal nerve fiber layer; IOP = intraocular pressure; MD = mean deviation; MvD = microvascular dropout; VF = visual field.
Values are shown in mean (95% confidence interval) to unless otherwise indicated.
During a mean follow-up period of 4.9 (95%CI, 4.6 to 5.2) years, eyes with MvD on baseline OCTA had a faster annual cpCD loss compared with eyes without baseline MvD (−0.87%/year vs −0.23%/year, respectively P <.001; Figure 1 ). In contrast, the rate of cpRNFL thinning did not differ significantly between eyes with and without MvD ( P >.05).
