This article is more than 5 years old. Information may no longer be current.
First AGS Journal Club
This year, the American Glaucoma Society held its first “Journal Club.” The goal was to have an open discussion of some of the more important papers published in the field during the previous year.
Identifying predictive factors for visual field progression
A paper discussed during Journal Club reported predictive factors for glaucomatous visual field progression in the Advanced Glaucoma Intervention Study (AGIS).1 From the original pool of 789 eyes of 591 patients in AGIS, 509 eyes of 401 patients met the entry criteria of a reference visual field score of <16, at least 3 years of follow-up and a minimum of seven visual field examinations with a reliability score of <2. Patients were randomized to undergo either argon laser trabeculoplasty (ALT)-trabeculectomy-trabeculectomy (ATT) or trabeculectomy-ALT-trabeculectomy (TAT). Researchers in AGIS used a single method for longitudinal evaluation of visual fields. The Humphrey Visual Field Analyzer I was set for the central 24-2 threshold test, size III white stimulus, and full threshold strategy, with the foveal threshold test turned on. Data were recorded from 55 locations in the visual field and scoring was based on the number, pattern and depth of depression of threshold sensitivities. Scores ranged from 0, or no defect, to 20, indicating advanced glaucoma. Visual field tests were administered 3 months after the initial intervention and at 6-month intervals thereafter. To avoid missing important associations, a pointwise linear regression analysis was applied to a subgroup of patients from AGIS.
Progression occurred in 151 eyes (29.9%) according to pointwise linear regression criteria and in 138 eyes (27.1%) according to AGIS criteria (64% agreement). Multivariate logistic regression demonstrated that older age at the time of first intervention, greater IOP fluctuation (later described as intervisit IOP variability), increasing number of glaucoma interventions and longer follow-up were associated with a higher probability of visual field progression. Male gender and lower baseline IOP had potential associations with visual field progression (.05<P<.1). Exclusion of AGIS visual field score and cup:disc ratio at baseline from the analysis resulted in the association of the same four variables with visual field progression.
|
Similar analyses were performed using AGIS criteria for definitions of visual field outcomes. Multivariate logistic regression analysis revealed that positive predictive factors for visual field progression included lower baseline AGIS score (P<0001), greater IOP fluctuation (P<.0001), vertical cup:disc ratio >.60 (P=.001), greater number of glaucoma surgeries (P=.008), longer follow-up (P=.013) and older age at first intervention (P=.028). After exclusion of baseline factors related to glaucoma severity including cup:disc ratio and baseline AGIS score, the only significant risk factors associated with visual field progression were greater IOP fluctuation (P=.0017) and increasing number of glaucoma interventions (P=.0098). Eyes with an IOP fluctuation <3 mm Hg tended to remain stable over time; however, eyes with an IOP fluctuation >3 mm Hg tended to progress significantly (P=.0006).
Researchers concluded that IOP fluctuation and older age at first glaucoma intervention consistently predict visual field progression according to both pointwise linear regression analysis and AGIS criteria. Furthermore, longer follow-up and a higher number of glaucoma interventions may be additional risk factors for worsening visual fields.
Following a summarization of results, discussants led by Harry A. Quigley, MD, emphasized the need to identify the risk factors for visual field progression in glaucoma and focused on methods, analysis, results and the next steps required in this area of research.2 An important discussion point was a recommended change in terminology from “IOP fluctuation” to “interoffice IOP variability.” It was suggested that IOP fluctuation may imply short-term diurnal IOP changes but “interoffice IOP variability” refers to IOP changes over longer intervals of time (i.e., 6-month intervals as in the AGIS study). Another important point raised was that patients in AGIS do not represent typical patients with glaucoma; therefore, generalization is a problem. Approximately 75% of AGIS patients had IOP >22 mm Hg while on maximum medication regimens. In contrast, 50% of glaucoma patients in population-based studies, such as the Early Manifest Glaucoma Trial (EMGT),3 have IOP <22 mm Hg. The degree of damage among AGIS patients was advanced compared with glaucoma patients from the EMGT. On the other hand, AGIS patients were treated optimally in that they received free medications and were followed-up regularly at 6-month intervals.
|
Field progression criteria for AGIS were selected on the basis of short-term reproducibility of fields, whereas pointwise linear regression methodology was selected because it yields the lowest progression rate so as to prevent overcompensation. Univariate analyses were included in the final model to determine preoperative and postoperative factors that were associated with visual field progression as were variables that might potentially predict or confound detection of visual field progression from a clinical viewpoint. When correcting for baseline IOP, older age at time of first intervention, longer follow-up and increasing number of glaucoma interventions, both IOP over time and IOP variability were significant predictors of visual field progression. When both variables were put in the same model, one “stole power from the other” and mean IOP was not predictive of visual field progression. When multivariate logistic regression was repeated excluding IOP variability, mean IOP reached statistical significance as a predictive factor (P=.045). IOP variability was weakly correlated with mean IOP (r=0.22, P<.001). Collectively, these results suggest that IOP fluctuation is an independent and more significant predictor than mean IOP for visual field progression in glaucoma patients.
Determining what factors were not associated with progression was equally important. For example, the intervention sequence was not predictive of visual field progression. Patients with vascular disease were not at a greater risk for visual field progression. An interesting finding was that diabetic retinopathy might be protective (odds ratio <1), although the correlation was not significant (P=.29). It is speculated that early leakage of growth factors into the retina may have protective effects.
Results from the AGIS discussed here indicate that higher IOP and a relatively greater degree of fluctuation are risk factors for visual field progression. Age is a surrogate for duration or susceptibility. Progression should be viewed as a rate with linear pointwise linear regression and not as an all-or-none event. Finally, no specific standard deviation of IOP targets exists as glaucoma populations differ markedly.
CSLO ancillary study of the OHTS
A second paper discussed during Journal Club was the report of the confocal scanning laser ophthalmoscopy (CSLO) ancillary study of the Ocular Hypertension Treatment Study (OHTS).4 By following 438 participants in the OHTS, researchers hoped to determine whether baseline CSLO optic disc topographic measurements were associated with the development of glaucoma in patients with ocular hypertension and whether specific findings could predict the onset of glaucoma. At entry to the OHTS, all patients had to have two successive normal visual field tests and normal-appearing optic discs based on clinical examination and review of full-frame 35-mm pairs or split-frame simultaneous stereoscopic optic disc photographs. Univariate and multivariate proportional hazard models were used to determine the relationship between baseline CSLO optic disc topographic measurements and indices associated with the development of POAG. Baseline CSLO parameters and composite indices included disc area, cup area and volume, cup:disc ratio, rim:disc ratio, mean cup depth, retinal nerve fiber layer thickness and cross sectional area, rim area and volume, mean height contour, reference plane height, cup shape, Moorfields Regression Analysis (MRA) and Heidelberg retina tomograph (HRT) classification (outside normal limits vs. within normal limits), as well as the continuous variable HRT classification II.
|
Forty-one eyes of 36 patients developed POAG after the initial CSLO assessment; nine reached clinical endpoints based on visual field results alone, 31 on stereophotographs alone and one based on concurrent visual field results and stereophotographs. Although race, sex, family history of glaucoma, history of high or low blood pressure, myopia and baseline visual field mean deviation were not associated with the development of POAG, univariate analysis demonstrated that history of heart disease, thin central corneal thickness and larger stereophotograph-based horizontal and vertical cup:disc ratios were significantly predictive of POAG in this study population. Study parameters that tended to be associated with the development of POAG according to the same univariate analysis included baseline age, IOP and visual field pattern standard deviation; however, these parameters did not reach statistical significance when compared with patients who did not develop POAG. Application of both univariate and multivariate analyses demonstrated that several baseline topographic optic disc measurements were significantly associated with the development of POAG. These measurements included larger cup:disc ratio, mean cup depth, mean height contour, cup volume, reference plane height, as well as smaller rim area, rim:disc and rim volume. Furthermore, analysis of multivariate hazard ratio demonstrated that classification of “outside normal limits” by HRT classification, MRA overall and global classifications and MRA classifications for the temporal inferior, nasal inferior and temporal superior regions were significantly associated with the development of POAG. The positive predictive value of CSLO indexes was highest for MRA temporal superior classification (40%). The positive predictive value of HRT classification outside normal limits was approximately 14%. However, longer follow-up (>97 months) is required to evaluate the true predictive value of these CSLO measurements.
What are the implications of this study?5 First, it is the first prospective study utilizing a planned analysis (i.e., not a post hoc analysis) to assess the relationship between baseline CSLO topographic optic disc parameters and the development of POAG in patients with ocular hypertension who have normal visual field test results and non-glaucomatous optic discs at study entry. Application of the appropriate Cox hazard model allowed researchers to determine which baseline parameters independently predicted disease.
Weaknesses of the study included generization and transferability. Discussion during the AGS Journal Club focused on how well the patients in this study reflect patients seen in clinical practice on a daily basis. The quality control measures that were instituted in the study to assure good scanning are not typical of everyday practice. Therefore, in order to validate this study and apply the results to the typical clinical setting, the methods should be applied to a separate set of patients with ocular hypertension, which is currently being done.
Another criticism of the study is that evidence exists that larger optic cup and smaller neuroretinal rim measurements may represent early signs of glaucoma and not risk factors for developing POAG. One of the discussants suggested that some of the patients who had no changes in visual fields who went on to develop visual field defects most likely already had optic nerve damage when they entered OHTS.
The researchers concluded that baseline CSLO topographic optic disc measurements used alone or in combination with extraneous factors including central corneal thickness, IOP and history of vascular disease, are predictive of glaucomatous changes in patients with ocular hypertension. All agreed that studies with longer follow-up periods are needed to confirm these results.
References
- Nouri-Mahdavi K, Hoffman D, Coleman AL, et al. Predictive factors for glaucomatous visual field progression in the advanced glaucoma intervention study. Ophthalmology. 2004;111(9):1627-1635.
- Caprioli J, Quigley HA. In discussion of: Nouri-Mahdavi K, Hoffman D, Coleman AL, et al. Predictive factors for glaucomatous visual field progression in the advanced glaucoma intervention study. Ophthalmology. 2004;111(9):1627-1635.
- Heijl A, Leske MC, Bengtsson B, Hussein M; Early Manifest Glaucoma Trial Group. Measuring visual field progression in the Early Manifest Glaucoma Trial. Acta Ophthalmol Scand. 2003;18(3):286-293.
- Zangwill LM, Weinreb RN, Beiser JA, et al. Baseline topographic optic disc measurements are associated with the development of primary open-angle glaucoma. The confocal scanning laser ophthalmoscopy ancillary study to the Ocular Hypertension Treatment Study. Arch Ophthalmol. 2005;123:1188-1197.
- Girkin CA. In discussion of: Zangwill LM, Weinreb RN, Beiser JA, et al. Baseline topographic optic disc measurements are associated with the development of primary open-angle glaucoma. The confocal scanning laser ophthalmoscopy ancillary study to the Ocular Hypertension Treatment Study. Arch Ophthalmol. 2005;123:1188-1197.