Purpose
To evaluate racial differences in the development of visual field (VF) damage in glaucoma suspects.
Design
Prospective, observational cohort study.
Methods
Six hundred thirty-six eyes from 357 glaucoma suspects with normal VF at baseline were included from the multicenter African Descent and Glaucoma Evaluation Study (ADAGES). Racial differences in the development of VF damage were examined using multivariable Cox proportional hazard models.
Results
Thirty one of 122 African-descent participants (25.4%) and 47 of 235 European-descent participants (20.0%) developed VF damage ( P = .078). In multivariable analysis, worse baseline VF mean deviation, higher mean arterial pressure during follow-up, and a race ∗ mean intraocular pressure (IOP) interaction term were significantly associated with the development of VF damage, suggesting that racial differences in the risk of VF damage varied by IOP. At higher mean IOP levels, race was predictive of the development of VF damage even after adjusting for potentially confounding factors. At mean IOPs during follow-up of 22, 24, and 26 mm Hg, multivariable hazard ratios (95% confidence intervals) for the development of VF damage in African-descent compared to European-descent subjects were 2.03 (1.15–3.57), 2.71 (1.39–5.29), and 3.61 (1.61–8.08), respectively. However, at lower mean IOP levels (below 22 mm Hg) during follow-up, African descent was not predictive of the development of VF damage.
Conclusion
In this cohort of glaucoma suspects with similar access to treatment, multivariate analysis revealed that at higher mean IOP during follow-up, individuals of African descent were more likely to develop VF damage than individuals of European descent.
Primary open-angle glaucoma (POAG) is the leading cause of irreversible blindness in the African-American population. Studies in the United States and Africa suggest that the prevalence of POAG is 4–5 times higher among individuals of African descent than in those of European descent, and that there is more visual impairment and faster progression of the disease in individuals of African descent than in other racial groups. There is also consistent evidence that healthy individuals of African descent have, on average, thinner corneas and greater optic disc and neuroretinal rim area, larger cups, larger cup-to-disc ratio measurements, and thicker retinal nerve fiber layer than do European-descent individuals. However, it remains unclear what clinical, ocular, genetic, demographic, and socioeconomic factors explain the racial differences in the development of glaucoma and subsequent severity of the disease. For example, in the Ocular Hypertension Treatment Study, African ancestry was not predictive of the development of glaucoma when corneal thickness, cup-to-disc disc ratio, and other clinical and demographic predictive factors were included in the final multivariable model.
The African Descent and Glaucoma Evaluation Study (ADAGES) is a prospective, multicenter, observational cohort study of glaucoma patients, glaucoma suspects, and healthy participants of African and European descent. ADAGES evaluates participants with a variety of measures of optic nerve head (ONH) structure and visual function, while documenting clinical, ocular, systemic, and sociodemographic predictive factors. The overall aim of ADAGES is to better identify, describe, and categorize these ocular, clinical, and sociodemographic factors that explain differences in glaucoma onset and rate of progression found between individuals of African descent and those of European descent after providing similar access to treatment.
The present ADAGES report compares the incidence and identifies predictive factors for the development of visual field (VF) damage among African-descent and European-descent glaucoma suspects.
Methods
Study Population
This was an observational cohort study of glaucoma suspects. Participants included in this study were selected from the ADAGES and Diagnostic Innovations in Glaucoma Study (DIGS). The multicenter ADAGES includes participants from the Hamilton Glaucoma Center at the Department of Ophthalmology, University of California, San Diego; the New York Eye and Ear Infirmary; and the Department of Ophthalmology, University of Alabama at Birmingham. DIGS includes participants recruited at the University of California, San Diego. By design, the procedures and testing relevant to this report are identical in ADAGES and DIGS. The protocols for DIGS and ADAGES were harmonized at study initiation so that the European-descent participants in DIGS can be used as a comparison group for the primarily African-descent participants enrolled in ADAGES. All the methods adhered to the tenets of the Declaration of Helsinki and to the Health Insurance Portability and Accountability Act. The institutional review boards at the University of California, San Diego; New York Eye and Ear Infirmary; and University of Alabama at Birmingham approved the methods. All participants of the study gave written informed consent. ADAGES and DIGS are registered as cohort clinical trials ( http://www.clinicaltrials.gov [identifiers NCT00221923 and NCT00221897 ; September 14, 2005]).
Enrollment in ADAGES took place between January 2003 and July 2006. Follow-up of this cohort is still ongoing. Detailed information about ADAGES is provided elsewhere. To be included in ADAGES/DIGS, at study entry participants were required to have best-corrected visual acuity of 20/40 or better, spherical refraction less than 5.0 diopters (D), cylinder correction less than 3.0 D, and open angles by gonioscopy. Those with coexisting ocular trauma, retinal disease, uveitis, nonglaucomatous optic disc neuropathy, or other diseases possibly affecting the VF were excluded from the study.
For this report, only ADAGES/DIGS participants of African descent or European descent classified as “suspect glaucoma” at the baseline visit with a minimum of 2 years of follow-up, and with at least 4 good-quality VF visits, were included in the analysis. Eyes with suspected glaucoma had a history of elevated intraocular pressure (ocular hypertension, OHT) and/or an optic disc appearance suspicious of glaucoma but normal visual fields at study entry. Elevated intraocular pressure (IOP) was defined as IOP >21 mm Hg or a history of ocular hypotensive treatment. Participants with evidence of consecutive repeatable VF damage at study entry in either eye were excluded.
Ocular Examination
The ocular testing completed for ADAGES and DIGS has been described elsewhere. In brief, participants underwent a comprehensive ophthalmic examination, including annual review of medical history, best-corrected visual acuity, slit-lamp biomicroscopy, IOP, dilated funduscopy examination, pachymetry, simultaneous stereoscopic optic disc photography, standard automated perimetry (SAP) with 24-2 Swedish interactive threshold algorithm (Carl Zeiss Meditec, Inc, Dublin, California, USA), and Heidelberg Retina Tomograph (HRT) (HRTII; Heidelberg Engineering, Inc, Heidelberg, Germany). At each bi-annual follow-up visit, VF testing, IOP measurements, and HRT imaging were completed.
Structural Damage
An optic disc appearance suspicious of glaucoma was defined as suspicion of neuroretinal rim thinning or notching, localized or diffuse retinal nerve fiber layer defect, or a between-eye asymmetry of the vertical cup-to-disc ratio more than 0.2. Two graders, masked to the participants’ race, age, and clinical diagnosis, evaluated the simultaneous stereophotographs (Kowa WX3D; Kowa Optimed, Inc, Torrance, California, USA, or Nidek 3Dx; Nidek Inc, Fremont, California, USA), according to a standard protocol using a stereoscopic viewer. In case of discrepancies between the 2 graders, adjudication was completed by a third experienced grader. Only photographs of adequate quality were used for evaluation. Participants with optic disc appearance suspicious of glaucoma in 1 or both eyes, but no VF damage, were included in the study.
Functional Damage
Normal SAP visual fields were defined as a mean deviation (MD) and pattern standard deviation (PSD) within the 95% confidence limits and a glaucoma hemifield test (GHT) result within the 99% normal limits. Reliable tests had 33% or less fixation losses and false negatives and 25% or less false positives. Trained reviewers from the University of California, San Diego–based Visual Field Assessment Center (VisFACT) ensured that each VF test included in ADAGES and DIGS was of good quality without lid or rim artifacts, evidence of learning effects, or other artifacts.
SAP visual fields were defined as abnormal if PSD was ≤5% and/or GHT was “outside normal limits.” Eyes that developed a repeatable VF defect, defined as 3 consecutive abnormal VF tests, were defined as “developed VF damage.” The development of VF damage was reviewed by an ophthalmologist (N.K.) to confirm that the damage was glaucomatous and the location of damage was consistent on all 3 visual fields.
Study Endpoint: Development of Visual Field Damage
A participant reached study endpoint when at least 1 eye developed repeatable VF damage (ie, abnormal VFs as defined above) at 3 consecutive visits. In order to analyze the data based on participant, rather than eye, eyes that did not develop repeatable glaucomatous VF damage were excluded from the analysis if their fellow eye developed repeatable visual field damage during follow-up.
Clinical History and Sociodemographic Information
Standardized interviews were conducted to obtain sociodemographic information including age, sex, and race. Clinical information including history of glaucoma treatment and history of systemic hypertension was obtained from medical charts.
A participant was considered to have systemic hypertension if any of the following criteria were met: (1) history of high blood pressure obtained from medical charts; (2) history of treatment for systemic hypertension obtained from medical charts; (3) measured average readings of systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg in a sitting position during a follow-up visit using DINAMAP@ PRO Monitor Model 100 (Critikon, Tampa, FL, USA). Blood pressure measurement cutoffs were selected based on the guidelines of the International Society of Hypertension/World Health Organization. Mean systolic, diastolic, and arterial pressure during follow-up were calculated using measurements at baseline and at each follow-up visit. Mean arterial pressure was estimated using the following formula: mean arterial pressure = (2/3) diastolic pressure + (1/3) systolic pressure. Only mean arterial pressure was included in the model.
The following ocular parameters were included in the univariate analysis as they have been reported to be significantly associated with the risk of developing glaucomatous damage: baseline axial length, central corneal thickness (CCT), spherical equivalent refraction, HRT measured disc area and rim area, photography-based horizontal and vertical cup-to-disc ratios, and VF MD. Overall mean IOP during follow-up was calculated using measurements at baseline and at each follow-up visit.
By design, ADAGES is an observational cohort that likely reflects current best practices of treatment decision making. Most IOP-lowering medications were provided at no cost to participants. A participant was considered to have a history of “glaucoma treatment” if his or her medical record indicated use of IOP-lowering medication and/or any glaucoma-related procedure during follow-up. Treatment decisions were not dictated by the study and were made at the discretion of the treating ophthalmologist at each site.
Statistical Analysis
We examined racial differences in sociodemographic characteristics and ocular and systemic factors using the Wilcoxon rank-sum test and Fisher exact test. Univariate Cox proportional hazards models were used to determine whether race and other risk factors were predictive of the development of VF damage among glaucoma suspects and to account for the correlation between fellow eyes in the presence of censoring. A censored eye is one in which, as of the last visit, VF damage has not developed. In the multivariate model, variables were chosen for analysis based on their importance in previous publications and their statistical significance ( P < .05) in the univariate models. In addition, we evaluated whether the effect of race varied by level of IOP by including a race ∗ IOP interaction term in the multivariable model. An IOP ∗ IOP interaction term was also included as a possible predictor, as the relationship between IOP and VF change can be nonlinear. We present a hazard ratio (and 95% confidence interval [CI]) for each predictor variable, which represents the increased or decreased risk of VF damage with respect to that predictor. Statistical analyses were performed using STATA software version 12 (Stata Co, College Station, Texas, USA).
Results
The study included 636 eyes of 357 participants with a mean age at entry of 58.1 ± 12.3 years (mean ± standard deviation [SD]). Overall, 231 participants (65%) were female. Two hundred thirty-five (67%) were of European descent and 122 (33%) were of African descent. The overall mean follow-up time was 7.1 years (±2.4 years); for African-descent participants it was 7.0 years (±2.4 years) and for European-descent participants it was 7.1 years (±2.4 years). Sixty-two of African-descent (62/122; 51%) and 137 of European-descent participants (137/235; 58%) received ocular hypotensive treatment during the follow-up period; this difference was not statistically significant ( P = .180).
There were significant differences between African-descent and European-descent participants for several baseline ocular and nonocular factors ( Table 1 ). African-descent participants had significantly larger HRT-measured disc area, thinner corneas, and higher blood pressure parameters (mean arterial pressure and diastolic pressure during follow-up) than European-descent participants (all P values ≤.002, significant after Bonferroni correction for multiple comparisons).
| Risk Factors | African-Descent Participants (Subjects, n = 122) | European-Descent Participants (Subjects, n = 235) | P Value c | ||||
|---|---|---|---|---|---|---|---|
| Developed Visual Field Damage (n = 31) | No Visual Field Damage (n = 91) | Total African Descent (n = 122) | Developed Visual Field Damage (n = 47) | No Visual Field Damage (n = 188) | Total European Descent (n = 235) | ||
| Demographic | |||||||
| Baseline age, y | 57.5 ± 14.0 | 53.7 ± 12.7 | 54.7 ± 13.1 | 64.1 ± 9.8 | 58.8 ± 11.7 | 59.8 ± 11.5 | <.001 |
| Follow-up time, y | 7.5 ± 2.0 | 6.8 ± 2.5 | 7.0 ± 2.4 | 8.3 ± 2.0 | 6.9 ± 2.4 | 7.1 ± 2.4 | .295 |
| Follow-up time from baseline to development of visual field damage, y | 2.4 ± 2.0 | N/A | 3.4 ± 2.5 | N/A | .088 d | ||
| Age at first abnormal visual field, y | 59.9 ± 13.7 | N/A | 67.4 ± 9.6 | N/A | .016 d | ||
| Sex, female/male | 22 (71%)/9 (29%) | 63 (69%)/28 (31%) | 85 (70%)/37 (30%) | 30 (64%)/17 (36%) | 116 (62%)/72 (38%) | 146 (62%)/89 (38%) | .163 |
| Ocular at baseline | |||||||
| Central corneal thickness, a μm | 537.4 ± 38.0 | 538.4 ± 38.0 | 538.1 ± 37.9 | 558.5 ± 41.7 | 562.3 ± 35.3 | 561.5 ± 36.6 | <.001 |
| Spherical equivalent refraction, a diopters | −0.11 ± 1.42 | −0.77 ± 1.86 | −0.61 ± 1.77 | −0.33 ± 2.19 | −0.59 ± 1.95 | −0.53 ± 2.00 | .715 |
| Axial length, a mm | 23.82 ± 0.81 | 24.13 ± 1.01 | 24.05 ± 0.97 | 23.98 ± 1.12 | 23.79 ± 1.07 | 23.83 ± 1.08 | .027 |
| Rim area, a mm 2 | 1.33 ± 0.42 | 1.35 ± 0.26 | 1.34 ± 0.30 | 1.24 ± 0.32 | 1.31 ± 0.25 | 1.30 ± 0.26 | .197 |
| Stereophotograph-based vertical cup-to-disc ratio a | 0.64 ± 0.15 | 0.64 ± 0.14 | 0.64 ± 0.14 | 0.67 ± 0.16 | 0.59 ± 0.15 | 0.61 ± 0.16 | .022 |
| Stereophotograph-based horizontal cup-to-disc ratio a | 0.63 ± 0.16 | 0.64 ± 0.15 | 0.64 ± 0.15 | 0.64 ± 0.15 | 0.58 ± 0.16 | 0.59 ± 0.16 | .006 |
| Disc area, a mm 2 | 2.30 ± 0.40 | 2.30 ± 0.58 | 2.30 ± 0.54 | 2.08 ± 0.51 | 1.97 ± 0.40 | 2.00 ± 0.42 | <.001 |
| Visual field mean deviation, a dB | −1.17 ± 1.68 | −0.29 ± 1.24 | −0.51 ± 1.4 | −0.84 ± 1.20 | −0.31 ± 1.21 | −0.42 ± 1.23 | .557 |
| Pattern standard deviation, a dB | 1.85 ± 0.25 | 1.58 ± 0.25 | 1.65 ± 0.27 | 1.72 ± 0.24 | 1.53 ± 0.26 | 1.57 ± 0.26 | .007 |
| Intraocular pressure, mm Hg | 17.7 ± 6.2 | 16.8 ± 5.7 | 17.0 ± 5.8 | 18.1 ± 7.7 | 18.5 ± 5.9 | 18.4 ± 6.3 | .054 |
| Diagnostic groups | .911 | ||||||
| Ocular hypertension | 12 (39%) | 42 (46%) | 54 (44%) | 16 (34%) | 86 (46%) | 102 (43%) | |
| Optic disc appearance suspicious of glaucoma | 19 (61%) | 49 (54%) | 68 (56%) | 31 (66%) | 102 (54%) | 133 (57%) | |
| Ocular during follow-up | |||||||
| Mean intraocular pressure, b mm Hg | 19.0 ± 4.9 | 16.5 ± 3.4 | 17.2 ± 4.0 | 17.9 ± 5.1 | 18.4 ± 3.9 | 18.3 ± 4.2 | .010 |
| Glaucoma treatment | 17 (55%) | 45 (50%) | 62 (51%) | 30 (64%) | 107 (57%) | 137 (58%) | .180 |
| Systemic (during follow-up) | |||||||
| Mean arterial pressure, b mm Hg | 100.9 ± 11.1 | 95.5 ± 9.2 | 96.9 ± 9.9 | 92.9 ± 9.6 | 91.8 ± 8.7 | 92.0 ± 8.7 | <.001 |
| Mean systolic pressure, b mm Hg | 141.2 ± 17.7 | 133.1 ± 14.2 | 135.1 ± 15.3 | 133.1 ± 15.5 | 129.6 ± 13.7 | 130.3 ± 14.1 | .005 |
| Mean diastolic pressure, b mm Hg | 80.8 ± 10.5 | 76.8 ± 8.3 | 77.7 ± 9.0 | 72.9 ± 8.3 | 72.9 ± 8.0 | 72.9 ± 8.0 | <.001 |
| Systemic hypertension | 28 (90%) | 70 (77%) | 98 (80%) | 32 (68%) | 143 (76%) | 175 (74%) | .238 |
b Measured at follow-up visits from baseline to the last visit or first abnormal visual field in 3 consecutive abnormal visual fields.
c P value comparing total African-descent participants to total European-descent participants, using Fisher exact test for categorical variables, Wilcoxon rank-sum test for continuous variables. If adjusting for multiple comparisons using Bonferroni correction, significant P value is defined as ≤.002.
d P value comparing African-descent participants who developed visual field damage to European-descent participants who developed visual field damage.
Overall, 101 eyes (101/636; 16%) of 78 participants (78/357; 22%) developed repeatable VF damage during the follow-up period. Twenty-six participants developed VF damage in both eyes; 52 developed VF damage in 1 eye. Thirty-one African-descent participants developed VF damage (31/122; 25.4%) compared to 47 (47/235; 20.0%) European-descent participants. The univariate Cox proportional hazards risk of developing VF damage in African-descent compared to European-descent participants was hazard ratio (HR), 1.43; 95% confidence interval (CI), 0.87–2.35 ( P = .163) ( Table 2 ). Given that participants of African descent compared to European descent were younger at baseline (54.7 ± 13.1 vs 59.8 ± 11.5 years, respectively, P < .001) and developed VF damage at an earlier mean age (59.9 ± 13.7 and 67.4 ± 9.6 years, respectively, P = .016), it became important to adjust for these age differences in the analysis. The association between race and the development of VF damage was statistically significant after adjusting for baseline age (HR African descent, 1.66; 95% CI, 1.00–2.74, P = .049), and of borderline significance after adjusting for baseline age and VF MD in multivariable models (HR African descent, 1.57; 95% CI, 0.96–2.58, P = .073).
| Predictive Factors | Hazard Ratio (95% CI) | ||
|---|---|---|---|
| Univariate | Adjusted for Baseline Age | Adjusted for Baseline Age and Visual Field Mean Deviation | |
| Demographic | |||
| African descent | 1.43 (0.87–2.35) | 1.66 (1.00–2.74) | 1.57 (0.96–2.58) |
| Older age per year | 1.03 (1.01–1.06) | 1.03 (1.01–1.06) | 1.02 (1.00–1.05) |
| Male sex | 0.87 (0.52–1.45) | ||
| Ocular (at baseline) | |||
| Diagnostic group: optic disc appearance suspicious of glaucoma vs ocular hypertension | 1.55 (0.94–2.56) | 1.44 (0.87–2.38) | 1.39 (0.84–2.30) |
| Axial length (per 1.0 mm greater) | 0.96 (0.79–1.16) | ||
| Central corneal thickness (per 40 μm thinner) | 1.22 (0.95–1.57) | 1.26 (0.97–1.64) | 1.29 (0.99–1.68) |
| Spherical equivalent refraction (per 1.0 D greater) | 1.14 (1.00–1.29) | 1.06 (0.92–1.22) | 1.11 (0.96–1.28) |
| Disc area (per 0.4 mm 2 greater) | 1.17 (0.98–1.39) | 1.24 (1.03–1.48) | 1.23 (1.02–1.48) |
| Stereophotograph-based vertical cup-to-disc ratio (per 0.1 greater) | 1.21 (1.01–1.46) | 1.22 (1.02–1.46) | 1.18 (1.00–1.41) |
| Rim area (per 0.2 mm 2 greater) | 0.86 (0.10–7.54) | ||
| Baseline visual field mean deviation (per 0.1 dB lower) | 1.04 (1.02–1.05) | 1.03 (1.02–1.05) | 1.03 (1.02–1.05) |
| Ocular (during follow-up) a | |||
| Mean intraocular pressure a (per 1.0 mm Hg greater) | 1.01 (0.93–1.09) | 1.01 (0.93–1.09) | 1.01 (0.93–1.09) |
| Mean intraocular pressure ∗ mean intraocular pressure | 1.02 (1.01–1.02) | 1.02 (1.01–1.02) | 1.01 (1.01–1.02) |
| Systemic (during follow-up) a | |||
| Mean arterial pressure (per 1.0 mm Hg greater) | 1.04 (1.01–1.06) | 1.03 (1.01–1.06) | 1.04 (1.02–1.07) |
| Mean systolic pressure (per 1.0 mm Hg greater) | 1.02 (1.01–1.04) | ||
| Mean diastolic pressure (per 1.0 mm Hg greater) | 1.03 (1.00–1.06) | ||
| Systemic hypertension | 1.00 (0.57–1.77) | ||
a Measured at follow-up visits from baseline to the last visit or first abnormal visual field in 3 consecutive abnormal visual fields.
Other factors that were positively associated with the development of VF damage based on univariate analysis included older age (HR per year, 1.03; 95% CI, 1.01–1.06), higher mean systolic pressure during follow-up (HR per 1.0 mm Hg, 1.02; 95% CI, 1.01–1.04), higher mean diastolic pressure during follow-up (HR per 1.0 mm Hg, 1.03; 95% CI, 1.00–1.06), and higher mean arterial pressure during follow-up (HR per 1.0 mm Hg, 1.04; 95% CI, 1.01–1.06) ( Table 2 ). Lower baseline VF MD (HR per 0.1 dB, 1.04; 95% CI, 1.02–1.05), larger photograph-based vertical cup-to-disc ratio (HR per 0.1, 1.21; 95% CI, 1.01–1.46), and spherical equivalent refraction (HR per 1.0 D greater, 1.14; 95% CI, 1.00–1.29) were among the ocular parameters predictive of VF damage.
The multivariable model included self-described race, CCT, disc area, mean IOP during follow-up, the interaction between self-described race and mean IOP, and the variables listed above with P ≤ .05 in univariate Cox proportional hazards analysis.
This multivariable model ( Table 3 ) suggests that in this ADAGES cohort of glaucoma suspects, worse baseline VF MD (HR per 0.1 dB, 1.04; 95% CI, 1.02–1.06) and higher mean arterial pressure during follow-up (HR per 1.0 mm Hg, 1.03; 95% CI, 1.00–1.06) were associated with the development of VF damage. The significance of the race ∗ IOP interaction term ( P = .003) suggests that racial differences in the probability of developing VF damage varied depending on mean IOP during follow-up, while the IOP ∗ IOP interaction term ( P < .0001) suggests a nonlinear relationship between IOP and the development of VF damage. For this reason we calculated the multivariable HR for race at different IOP levels ( Table 3 ). We found that at the highest quartile of mean IOP level during follow-up (IOP >21 mm Hg), African descent was predictive of the development of VF damage; at the lowest 3 quartiles of mean IOP levels during follow-up (≤21 mm Hg), African descent was not predictive of the development of VF damage ( Figure 1 ). Specifically, at IOP levels of 22 mm Hg, 24 mm Hg, and 26 mm Hg, the multivariable hazard ratios (HR African descent compared to European descent [95%CI]) for developing VF damage were 2.03 (1.15–3.57), 2.71 (1.39–5.29), and 3.61 (1.61–8.08), respectively, even after controlling for CCT, baseline age and VF MD, disc area, and other systemic and ocular factors ( Figure 2 ). However, when we did not specify the IOP in the model, the mean IOP of the entire cohort (approximately 17.8 mm Hg) is used as the default IOP. Using the default mean IOP during follow-up value in the multivariable model, race was not associated with the development of VF damage (HR African descent [95%CI] 1.12 [0.66–1.90]).
