Discussion
This study investigated the course of POAG progression in African ancestry individuals. We found that both structural and functional progression rates in these individuals were faster than rates reported in other ethnic groups in previous studies. We also identified several risk factors for faster disease progression in our cohort. A deeper understanding of the more rapid progression rates in African ancestry individuals and their risk factors can help to better identify high risk patients and to inform best decisions on treatment choice and escalation.
The median rates of structural and functional decline in our African ancestry POAG cohort were faster than rates reported in most non-African cohorts of POAG patients. In all below comparisons, we used the median rate of progression wherever available; if not reported, we used the mean. Eyes in our study experienced RNFL loss at a median rate of change of −1.60 µm/year, which is almost three times as much as the normal age-related rate of change (−0.52 µm/year).30 33 The Duke Glaucoma Registry and the Diagnostic Innovations in Glaucoma Study reported slower mean rates of decline (−0.76 μm/year and −0.98 μm/year, respectively) in these cohorts, which were primarily white.15 34 Similarly, the median rate of MD loss in the POAAGG cohort was −0.40 dB/year. Studies in predominantly white cohorts, such as the Duke Glaucoma Registry and Portsmouth Visual Field Database, reported slower rates of MD loss (median of −0.15 dB/year for both cohorts).30 35 The African Descent and Glaucoma Evaluation Study reported a mean rate of −0.24 dB/year in black patients.36 Two studies reported faster rates of functional loss in their cohorts (mean of −0.45 dB/year; median of −0.62 dB/year), but both studies included eyes with other forms of glaucoma, making it difficult to draw direct comparisons to our POAG cohort.37 38 These results suggest that African ancestry populations experience more rapid structural and functional progression of POAG than other ethnic groups. Further research is needed to confirm these conclusions as they were drawn from comparison with other cohorts.
There were a wide range of RNFL and MD progression rates among eyes in our cohort. There was a fairly even distribution of structural progression rates among the slow, moderate and fast categories. In contrast, functional progression rates were classified as slow for almost 90% of eyes. Additionally, while almost one-third of eyes were fast structural progressors, only 1% were fast functional progressors. The functional progression rate and structural progression rate were only weakly correlated (correlation coefficient=0.08). These findings emphasise the importance of closely monitoring patients with rapid RNFL decline, even if their visual fields do not show concurrent significant loss. Studies have shown that these individuals are at high risk for severe visual field loss during follow-up.15 It is also important to note that the clinical significance of these categorisations largely depends on a patient’s age and life expectancy. For example, a 50-year-old patient with ‘slow’ progression rates may be at higher risk for vision loss over their lifetime compared with a 90-year-old patient with ‘fast’ progression rates.17
Another important finding from this study was how progression rates relate to baseline RNFL and MD values. We found that thicker baseline RNFL was associated with more rapid RNFL decline and increased risk of being a fast structural progressor. This could be the result of the non-linear nature of glaucoma progression, where progression in eyes with early disease exhibits a steeper slope before plateauing in advanced stages. RNFL thinning is no longer observable through OCT imaging after reaching this ‘floor effect’ in advanced disease, but patients may still experience disease progression.39 Other studies have also reported associations between thicker baseline RNFL and faster rates of RNFL loss, both in healthy33 and glaucomatous eyes.40 On a similar note, our study found that higher (better) MD at baseline was associated with faster rates of MD loss. While some studies have shown that worse baseline MD values were risk factors for progression,10 16 19 studies such as AGIS also found that worse visual fields at baseline were protective against further visual field loss.20 These results may also be due to the steeper slope of glaucoma progression in early disease, when patients have better MD values, before rates flatten in later disease stages.17 Additionally, it is also possible that patients with better RNFL or MD values at baseline received less aggressive interventions, contributing to faster subsequent progression. Interestingly, our study also found that higher (better) MD at baseline was associated with slower rates of RNFL loss, which is the opposite of what would be expected based on the above results. However, it is possible that individuals with less visual field damage at baseline require a higher level of insult to the optic nerve to cause damage and thus have slower structural progression rates.
We found that several well-known risk factors for POAG were associated with faster progression, including elevated IOP and larger CDR. These variables were associated with increased risk of being a moderate/fast functional progressor and with faster rates of MD loss, respectively. Prior studies have shown that higher baseline or peak IOP is a major risk factor for faster visual field decay.16 19 Though IOP control is not always sufficient to slow disease progression, with approximately 30% of patients progressing despite treatment, it still remains an essential component of treatment.41 Increased cupping of the optic nerve is also an important clinical indicator of worsening POAG, as it corresponds to death of retinal ganglion cells (RGCs).42 However, higher baseline CDR was not a major risk factor for POAG progression in most studies in non-black populations. A previous study in our African ancestry cohort found that legally blind glaucoma cases were more likely to have a higher CDR at diagnosis than non-blind glaucoma cases.43 Other studies have shown that CDR is larger in individuals of African descent compared with Caucasians.44 These findings suggest that enlarged cupping at baseline may be a risk factor for disease progression, especially in African ancestry populations.
Interestingly, the only demographic, behavioural or systemic variable associated with progression in this study was BMI. We found that lower BMI was associated with faster rates of MD loss in multivariable analysis. While there have been some conflicting reports regarding the relationship between glaucoma and BMI, the majority of studies have reported either an inverse relationship between these variables45 or no significant relationship.46 47 A study in normal-tension glaucoma patients also found that a lower BMI was associated with faster visual field loss.48 Though the exact mechanisms of this association are not definitively known, some researchers refer to the linear relationship between BMI and cerebrospinal fluid pressure as an explanation. Lower cerebrospinal fluid pressure affects the trans-LC pressure difference gradient, leading to increased deformation of the LC, decreased neuroretinal rim area and increased visual field defects.49 50
Two qualitative variables were associated with structural progression in this study: presence of beta parapapillary atrophy (PPA) (risk factor) and LC baring (protective). Like other qualitative features, both features were observed on stereo disc images by trained graders and recorded on a standardised grading form. Beta PPA was distinguished by marked atrophy of the retinal pigment epithelium and the choriocapillaris, giving a white background with good visibility of choroidal vessels and/or sclera. Our group found that beta PPA was a risk factor for faster RNFL loss, which is consistent with prior studies51–53; some groups suggest that this feature can be a potential diagnostic marker for POAG.54 The LC is the primary site of axonal injury in glaucomatous eyes. Bundles of axons and retinal blood vessels pass from the eye to the neural canal through LC pores, which are specific openings between the laminar cribriform plates.55 In this study, baring of the LC (defined by graders as >3 visible LC pores) was protective against being classified as a fast structural progressor. Several prior studies reported that the number of LC pores does not differ between glaucomatous and normal eyes.56 57 However, most prior studies focused more on the shape and size of LC pores, rather than the quantity, finding that pores typically become larger and more elongated during glaucoma.56 57 More research is needed to understand the association reported in our study. We plan to investigate the shape of LC pores and examine longitudinal changes in future studies.
Finally, we found that females with the TC genotype in the TRIM66 gene were at higher risk of being a moderate or fast functional progressor. In a prior GWAS in our cohort, this gene was associated with case–control status and led to expression in relevant cell lines under conditions of oxidative stress. A subsequent genotype/phenotype study found that this variant was associated with thinner RNFL and larger CDR in males only.26 This protein plays an important role in protection of RGCs against oxidative stress.
This study has several limitations that could affect the results. First, we included glaucoma cases with ≥2 visits with RNFL measurements and ≥2 visits with MD measurements over ≥6-month follow-up. We based these inclusion criteria on the prior literature, with the goal of including as many African ancestry patients as possible; however, we acknowledge that slope estimates can be imprecise (especially for eyes with only two visits), and that this sample may not be representative of the entire POAG cohort. Additionally, patients excluded due to lack of follow-up visits may represent those with worse adherence to treatment and possibly faster progression. Second, we tracked visual field loss by using MD, which is a global measure that is not as sensitive to localised glaucomatous damage. MD values can also be affected by factors beyond worsening glaucoma, such as cataract.58 Third, the results in this study were dependent on the cut-offs used to classify progression rates as slow, moderate and fast. We based these cut-offs on the range of the instruments and prior reports in the literature,30 but fully acknowledge that the percentages in each group would shift depending on the chosen cut-offs. Because there was a low number of eyes in the fast functional progressor group, we needed to combine moderate and fast functional progressors in order to complete the subsequent risk factors analysis. Fourth, some risk factors examined in this study had a large amount of missing data and were not considered in subsequent multivariable analyses, which may limit the generalisability of study findings. Finally, another potential limitation to this paper is that, although the earliest enrolled patients in this cohort have been followed for 12 years, and earliest age of enrolment was 35 years of age to capture early disease, we cannot establish that all patients were at the same stage of disease when we evaluated risk factors for progression.
In conclusion, this study provides important information about the course of POAG progression and its risk factors in African ancestry individuals. We showed that African ancestry individuals experience faster structural and functional progression compared with other ethnic groups. Fast progression was more frequent for structural progression than for functional progression, emphasising the importance of monitoring RNFL OCT results to identify at-risk patients and providing timely treatment in early disease stages. We confirmed risk factors for progression, such as higher IOP, larger CDR and beta peripapillary atrophy, while also identifying lesser-known risk factors, such as the TRIM66 TC genotype and better RNFL and MD values at baseline. Better understanding of the course of glaucoma progression in African ancestry individuals can inform decisions on treatment choice and escalation. Combined with other important factors, such as age, lifetime expectancy and baseline of visual field loss, ophthalmologists can make an informed prognosis of the likelihood of lifetime visual disability for patients—and make every effort to slow progression and delay this outcome.
