Detecting disease progression in mild, moderate and severe glaucoma

INTRODUCTION

Glaucoma is a chronic, progressive optic neuropathy associated with characteristic atrophy of the optic nerve and associated visual field abnormalities. Glaucoma is thus composed of both structural and functional elements. The management of glaucoma is inherently tied to structural and functional evaluation, with interventions triggered by clinical evidence of disease progression. As structure and function may progress differently in glaucoma, no single test is sufficient for the surveillance of glaucoma. Instead, integration of both structural and functional data, tailored to disease stage and individual patients’ circumstances, is needed. These structural and functional measurement techniques have evolved considerably overtime. In this review, we provide an overview of evidence-based strategies for detection of glaucomatous progression at different disease stages.

GENERAL PRINCIPLES OF DETECTING GLAUCOMA PROGRESSION

DETECTING DISEASE PROGRESSION IN MILD GLAUCOMA

Mild glaucoma is defined by glaucomatous optic nerve abnormalities without SAP visual field abnormalities (International Classification of Diseases/ICD-10), or by a visual field mean deviation less than −6 dB (Hodapp-Parrish-Anderson/HPA Criteria). The nature of the structure-function relationship in early glaucoma – and especially the question of which element can be detected first – has been the subject of intensive study and debate. Classic investigations comparing optic disc cupping and visual field change [16], and subsequent clinical-pathologic case studies of cadaveric eyes [17], demonstrated that RGC loss can precede detectible visual field changes, coining the concept of ‘preperimetric glaucoma’. However, other studies have additionally demonstrated perimetric progression in ocular hypertension prior to structural changes [18]. The current understanding is that either structural or functional loss may be initially detected in early glaucoma, highlighting the importance of carefully monitoring for both [5,19,20].

Changes in RNFL thickness can be a sensitive metric for evaluating early-stage glaucomatous changes [21]. It is important, however, to assess the numerical thickness values for abnormal asymmetry and progression. Overreliance on OCT colour coding may lead a clinician to overlook early thinning, known as ‘green disease’ or ‘progression in the green’ (Fig. 1) [22], or to falsely diagnose glaucoma based on yellow or red coding of anatomic disc variants [23]. Comparison of the two eyes of the same patient is important; an inter-eye difference in average RNFL thickness of greater than 9 μm has been shown to be a statistically significant indicator of early glaucoma [24].

As the macula has high density of RGCs, early glaucomatous damage can be detected in this region with OCT as well [5]. The rate of average GCIPL thinning has been shown to be greater in early glaucoma progressors, and can be assessed with OCT GPA [25]. Some studies have, however, suggested that GCIPL parameters are of lower predictive value in preperimetric glaucoma than RNFL metrics [26]. Other studies suggest that an indexed combination of OCT datapoints from the ONH and macula may be more sensitive for detecting early glaucomatous changes than any individual metric by itself [7]. Finally, studies using newer wide-field OCT technology – simultaneously capturing a large area of the posterior pole – have shown some promise in detecting early glaucoma [27^■,28].

Early visual field changes in mild glaucoma often affect the nasal aspect of the visual field as a function of early arcuate nerve fibre damage [29]. Early visual field progression patterns can vary, however, and their initial location and propagation have been associated with different ophthalmic and systemic factors [30]. A significant proportion of early visual field defects can also revert to normal on subsequent examinations [31]. Clinicians must therefore be mindful of any new, enlarging or persistently deepening defect of the visual field, regardless of location. Event-based progression analysis can be very useful in early glaucoma to detect areas of early/focal progression, even when global indexes and associated trend-based analyses appear to be relatively stable [13].

The optimal frequency of visual field testing in early/mild glaucoma is a subject of some debate. An average of three annual tests in the first 2 years after glaucoma diagnosis has demonstrated good sensitivity to promptly identify patients with rapidly progressing disease (e.g. mean deviation decline by greater than 1 dB/year) [32]. Other studies have suggested two consecutive baseline tests followed by two yearly tests is reasonably specific to rule out rapid progression [33]. Recent machine learning-based algorithms have demonstrated the potential to predict rapid progressors even with only one baseline visual field test, when combined with additional clinical data [34^■].

Finally, the importance of a complete ophthalmic history and physical examination should be emphasized in early glaucoma. Risk factors such as family history of glaucoma and measurement of central corneal thickness (CCT) are vital for risk stratification, as is the more recent inclusion of corneal hysteresis to inform a patient’s risk profile [35]. Intraocular pressure should be carefully measured at every encounter. Important qualitative and characteristic changes can be detected through direct and photographic examination of the optic nerve. In particular, disc haemorrhage is well documented as a marker of progressing and/or inadequately controlled glaucoma [36]. Other important qualitative physical examination findings include monitoring for progression of focal rim thinning/notching and peripapillary atrophy. Serial fundus photography is useful to catalogue and monitor these qualitative findings, especially for long-term monitoring of glaucoma suspects [37].

DETECTING DISEASE PROGRESSION IN MODERATE GLAUCOMA

Moderate glaucoma is defined as optic nerve damage consistent with glaucoma and glaucomatous visual field abnormalities in one hemifield greater than 5 degrees from fixation (ICD-10). HPA criteria define moderate glaucoma based on a visual field mean deviation between −6 and −12 dB and certain PSD criteria. Structural and functional deficits are typically present; a key challenge in this stage is the variability in how quickly, and in what order, these deficits progress. Both event and trend-based guided progression analyses can be of great use in detecting meaningful progression (Fig. 2). However, often structural and functional changes do not occur simultaneously.

In the moderate stage of glaucoma, the correlation between RGC count, average RNFL thickness and SAP mean deviation is nonlinear, with early RGC loss translating to relatively larger changes in RNFL thickness than mean deviation decline. This suggests that OCT is comparatively more sensitive to glaucoma progression than SAP in earlier stages of disease. Increasingly rapid progression of VF loss can subsequently be encountered as RNFL thinning plateaus. This asymmetric structure-then-function pattern of progression has been termed the ‘broken stick model’; Liu et al. [38] found that detectible visual field loss began after the neuroretinal rim reached one-third normal thickness. Thus, a finding of OCT RNFL and/or macular progression should be noted and treated even in the absence of VF change, as such changes often predict future VF progression [39^■]. The rate of OCT progression bears prognostic value with rapid thinning, defined in one recent study as mean OCT RNFL thinning greater than 2 μm/year, correlating with subsequent visual field defects [40^■]. Of note, age-related mean RNFL thinning proceeds at comparably modest rate of nearly 0.2–0.5 μm/year [3,6,41,42].

Detection of progression in moderate glaucoma often carries significant consequences, as management may require more aggressive medical or procedural interventions. In cases of uncertainty, the use of additional corroborating data points is useful. True progression is of higher likelihood, for example, in the setting of intraocular pressure persistently elevated above target [43^■■,44^■], or in the setting of recent disc haemorrhages. Inter-test variability increases proportionally with glaucoma severity; [45] as such, there should be a low threshold for repeating tests to confirm progression. In addition, clinicians should consider supplemental testing, such as a 10–2 HVF to assess possible new central field loss noted on a 24–2 testing scheme.

DETECTING DISEASE PROGRESSION IN SEVERE GLAUCOMA

Severe glaucoma is defined as glaucomatous visual field abnormalities affecting both hemifields and/or losses within 5 degrees of fixation (ICD-10) or mean deviation less than −12 dB (HPA). Test-retest variability (long-term fluctuation) increases in more advanced stages of glaucoma; variability can even occur with increasing frequency during a single examination (short-term fluctuation) [13,46]. Structural and functional losses ultimately approach the limit of our machines’ ability to reliably measure further losses. This lower bound threshold at which a metric’s further progression is no longer detectible is termed the ‘floor effect’. The floor effect does not represent a biological phenomenon, but a technological limitation. Measurements should be viewed with caution when approaching the floor range, as a sequence of similar readings can easily be misinterpreted as ‘stable’ when progression is still occurring (Fig. 3). Exact floor effect thresholds vary across devices; for OCT, average RNFL floors occur in the range of 40–50 μm [47]. SAP exhibits a relatively large range given the logarithmic nature of its measurement scale, and may continue to show progression beyond the RNFL floor.

Late-stage structural progression can still be measured via OCT measurements of the macula, as GCIPL tissue often remains above the measurement floor even after RNFL measurement reaches the floor (Fig. 3) [48,49]. Furthermore, the central macula has been shown to correlate relatively well with the central 10 degrees of the visual field, rendering it a clinically meaningful target to monitor progression even for patients with diffuse RNFL thinning and profoundly constricted peripheral visual fields [50].

For patients with visual field impairments that exceed the limitations of SAP, nonstandard perimetric tests can still be useful. When the paracentral 10 degrees are impacted, a centre-focused testing strategy such as HVF 10–2 may better assess progression. Increasing the perimeter stimulus size from the standard size III to size V has additionally been shown to reduce test variability in advanced glaucoma [51]. Switching to this new testing pattern does, however, impose its own specific challenges, including loss of common metrics such as mean deviation, PSD and VFI, and an inability to reliably compare results with prior SAP tests. With even further progression, kinetic (Goldmann) perimetry can also offer advantages over static perimetry for certain patients, providing sharper demarcation of visual field boundaries and more control in cases wherein fixation is a challenge [52]. However, comparison of kinetic perimetry is mostly qualitative.

In the most advanced cases, there may come a point where structural and functional tests are insufficient to yield useful information for progression monitoring. Instead, information on progression may be gleaned from patients’ subjective reports of their visual acuity and visual function. Grounding inquiries in concrete examples, such as a patient’s activities of daily life and safety, may yield meaningful information [53]. Subjective reporting is, nonetheless, quite variable, and self-reported progression may not reflect biological changes.

CONCLUSION

Structural and functional measurement techniques in glaucoma have evolved considerably over the recent years, providing clinicians with an expansive toolkit to detect glaucomatous progression. Despite the recent proliferation of complex computer algorithms and high-definition imaging modalities, no single metric has demonstrated universal superiority. The detection of glaucomatous progression at all stages of disease relies on clinical judgement, derived from the critical assessment of both structural and functional testing in conjunction with intraocular pressure measurements, clinical experience and each patient’s specific clinical context.