Purpose: To investigate the longitudinal variability of glaucoma risk calculation in ocular hypertensive (OHT) subjects.
Methods: We reviewed the charts of untreated OHT patients followed in a glaucoma referral practice for a minimum of 60 months. Clinical variables collected at baseline and during follow-up included age, central corneal thickness (CCT), intraocular pressure (IOP), vertical cup-to-disc ratio (VCDR), and visual field pattern standard deviation (VFPSD). These were used to calculate the 5-year risk of conversion to primary open-angle glaucoma (POAG) at each follow-up visit using the Ocular Hypertension Treatment Study and European Glaucoma Prevention Study calculator (http://ohts.wustl.edu/risk/calculator.html). We also calculated the risk of POAG conversion based on the fluctuation of measured variables over time assuming the worst case scenarios (final age, highest PSD, lowest CCT, highest IOP, and highest VCDR) and best case scenarios (baseline age, lowest PSD, highest CCT, lowest IOP, and lowest VCDR) for each patient. Risk probabilities (%) were plotted against follow-up time to generate slopes of risk change over time.
Results: We included 27 untreated OHT patients (54 eyes) followed for a mean of 98.3±18.5 months. Seven individuals (25.9%) converted to POAG during follow-up. The mean 5-year risk of conversion for all patients in the study group ranged from 2.9% to 52.3% during follow-up. The mean slope of risk change over time was 0.37±0.81% increase/y. The mean slope for patients who reached a POAG endpoint was significantly greater than for those who did not (1.3±0.78 vs. 0.042±0.52%/y, P<0.01). In each patient, the mean risk of POAG conversion increased almost 10-fold when comparing the best case scenario with the worst case scenario (5.0% vs. 45.7%, P<0.01).
Conclusions: The estimated 5-year risk of conversion to POAG among untreated OHT patients varies significantly during follow-up, with a trend toward increasing over time. Within the same individual, the estimated risk can vary almost 10-fold based on the variability of IOP, CCT, VCDR, and VFPSD. Therefore, a single risk calculation measurement may not be sufficient for accurate risk assessment, informed decision-making by patients, and physician treatment recommendations.
Primary open-angle glaucoma (POAG), a neurodegenerative condition of the optic nerve, is one of the leading causes of irreversible loss of vision and affects more than 3 million people in the United States alone.1 An extensive body of research has elucidated the most important risk factors associated with this disease.2–7 Within the glaucoma continuum, ocular hypertensive (OHT) subjects—with normal optic discs and visual fields but statistically high intraocular pressure (IOP)—are at particular increased risk of glaucoma development.8 However, not all OHT subjects develop glaucoma even if they remain untreated.2,3 To identify individuals most susceptible to developing this condition and assist in risk stratification of patients with ocular hypertension, predictive models have been proposed.8–11 These models are based upon the results and findings of the Ocular Hypertension Treatment Study (OHTS) and European Glaucoma Prevention Study (EGPS) that identified higher IOP, older age, lower central corneal thickness (CCT), greater vertical cup-to-disc ratio (VCDR), and worse pattern standard deviation (PSD) on visual field analysis as independent risk factors for the onset of POAG.2,8,9,11
The complexity of synthesizing the different variables for any individual OHT patient can result in large discrepancies in predicting glaucoma risk given its subjectivity and interobserver variability; therefore, this empirical categorization of risk can complicate the decision on whether to initiate treatment.12,13 Objective, computerized prediction models such as the 5-year glaucoma risk calculator developed by the OHTS and EGPS can offer greater facility and consistency in the management of OHT patients.13
The currently available risk model from the OHTS, however, relies on baseline variables determined at the initial evaluation.8–11,14–16 One of its key assumptions is that these initial measurements are the most pertinent to determining an individual’s risk and that these covariates remain relatively static over time. Karp et al,17 however, drawing from their reassessment of the Framingham Heart Study participants, demonstrated that baseline-only risk calculators can be less reliable than models that incorporate updated risk factors collected over time.
In the present study, we attempted to assess the variability of the OHTS glaucoma risk prediction model in OHT subjects by applying updated risk factor information obtained during follow-up and measuring the variability in the calculated risk over time.
This was a prospective, longitudinal study approved by the New York Eye and Ear Infirmary Institutional Review Board and followed the tenets of the Declaration of Helsinki. The glaucoma 5-year risk estimator used in this study (http://ohts.wustl.edu/risk/calculator.html) derives from the results and risk factors identified by the OHTS and EGPS.9 The pooled multivariate Cox proportional hazards model from these reports determined that baseline age, IOP, CCT, VCDR, and Humphrey VFPSD were significant in predicting the risk of conversion to POAG in 1 or both eyes from untreated ocular hypertension within 5 years.
The subjects included in this study are a subset of patients with untreated ocular hypertension from 1 clinical center who were enrolled in the OHTS. Only those patients who were followed for a minimum of 60 months were included in the present study. Untreated OHT was defined using the definition for recruitment into the OHTS.18 Subjects were between the ages of 40 and 80 years (inclusive), with an IOP≥24 and ≤32 mm Hg in one eye and an IOP≥21 and ≤32 mm Hg in the fellow eye, normal appearing cup-to-disc ratio with a difference not >0.2 between the 2 eyes, and normal and reliable perimetry based on full-threshold white-on-white Humphrey 30-2 visual field tests. Visual field tests were performed at 6-month intervals as part of the study protocol. If progression was detected or suspected, repeated testing was performed at closer intervals for confirmation. Those with best-corrected visual acuity worse than 20/40 in either eye, exfoliation syndrome or pigment dispersion, previous intraocular surgery (except for uncomplicated extracapsular cataract extraction with posterior chamber intraocular lens), secondary causes of elevated IOP (including angle-closure glaucoma or anatomically narrow angles), background diabetic retinopathy, or underlying systemic or ocular conditions that may contribute to visual field or optic disc abnormalities were excluded.
Clinical variables collected at baseline and during follow-up included the aforementioned predictive factors necessary for calculating the risk percentage (ie, age, CCT, IOP, VCDR, and VFPSD). Considerable variability in recorded VCDR as determined by biomicroscopy was felt to be attributable to the subjectivity of individual examiners.19,20 To maintain consistency, the mean VCDR over the course of follow-up was used for the risk calculation for each visit. Conversion to POAG was determined using the definitions provided by the OHTS protocol, that is, the presence of a reproducible Humphrey 30-2 visual field abnormality of the same type, location and index of abnormality, or significant optic disc deterioration represented by generalized or localized thinning of the neuroretinal rim on sequential stereoscopic optic disc photographs.18
The OHTS/EGPS 5-year glaucoma risk calculator was used to calculate each patient’s glaucoma risk on each follow-up visit using the requisite variables collected on that particular day. These values were used to determine the mean and range of calculated risk over the course of follow-up for each patient. Risk values were plotted against follow-up time to generate slopes of risk change over time (%/t) for each patient. The 2-sided independent-samples t test was used to compare data where appropriate. The nominal significance level for comparisons in this study was α=0.05.
To assess the theoretical degree of fluctuation possible in the calculated risk of POAG conversion, a worst case scenario (WCS) and best case scenario (BCS) risk was determined for each patient. For IOP fluctuation and PSD variability, we considered the highest and lowest measurements observed during the entire follow-up period for each enrolled untreated OHT subject. For CCT variability, we considered the effect of both its short-term variability (or diurnal fluctuation)21 and long-term variability (or measurement error),22 and assumed a value of ±10 µm. For VCDR, a value of ±0.2 was applied to model interobserver and intraobserver variability.20,23 Therefore, to determine the WCS risk for each patient, the assumed values applied were final age, highest PSD, mean CCT minus 10 µm, highest IOP, and mean VCDR plus 0.2. The converse extremes were assumed for the BCS risk calculations: baseline age, lowest PSD, mean CCT plus 10 µm, lowest IOP, and mean VCDR minus 0.2.
Twenty-seven patients (54 eyes) with untreated ocular hypertension were identified and included in the study. The mean age was 53.8±7.3 years, and 59% were women. Mean follow-up duration was 98.3±18.5 months.
Seven individuals (25.9%) converted to POAG (+POAG) during follow-up on the basis of Humphrey visual field abnormalities and/or progressive neuroretinal rim thinning of the optic disc. At baseline, a larger VCDR was observed in the +POAG group compared with the cohort that did not progress to POAG (−POAG) (P=0.043). The +POAG group also had higher IOP, thinner CCT, and greater PSD than the −POAG group, although these differences did not reach significance (Table 1).
The mean 5-year risk of conversion to POAG calculated at baseline was 14.0±12.0% and 22.5±11.5% in the −POAG and +POAG groups, respectively. Over the course of follow-up, the mean calculated risk was 13.4±10.7% (range, 2.9% to 49.2%) in the −POAG group and 26.2±15.7% (range, 14.0% to 52.3%) in the +POAG group (P<0.001). The mean risk percentage range (highest risk %−lowest risk %) in the −POAG and +POAG groups was 10.5±6.9% and 23.4±13.4%, respectively (P=0.43) (Table 2).
Univariate regression analysis to determine how the calculated risk changes over time was performed for each patient. In the +POAG group, the mean change in risk increased by 1.3±0.78%/y in comparison with 0.04±0.51%/y in the −POAG group (P=0.0041). Notably, 10 of the 20 patients in the −POAG group had a negative risk %/t slope, whereas all patients in +POAG group had an uptrend in their risk percentage over time.
Analysis of the BCS and WCS revealed a mean 5-year risk of 4.2±3.8% and 40.5±21.6%, respectively, in the −POAG group (P<0.001). Similarly, in the +POAG group, the mean 5-year risk was 7.2±4.3% in the BCS and 60.5±24.1% in the WCS (P<0.001). No statistically significant difference was observed when comparing the BCS between the 2 groups (P=0.15) or the WCS (P=0.08) (Table 3).
The task of weighing the significance of each predictive factor to arrive at a global assessment of an individual’s risk for developing a clinical endpoint often presents itself as an impractical challenge. Mansberger and Cioffi12 highlighted the variability in risk stratification of OHTs by ophthalmologists, hindering agreement on whether to initiate treatment. Risk calculators deduced from population-based multivariate analyses provide an evidence-based attempt to circumvent this problem.
To determine whether risk calculators have any tangible effect on clinical decision-making, Boland et al13 recently reported that in managing OHTs, a significant proportion of ophthalmologists alter their treatment recommendations when provided risk calculations, doing so with greater consistency and confidence. Thus, prediction models potentially offer improvements in quality of care by reducing uncertainty and enabling providers to make repeatable treatment decisions.
However, although the current available prediction models are instrumental in guiding clinical judgment, they are by no means perfect. Most prediction models, for instance, including the OHTS risk calculator, derive from variables measured only at baseline. This assumes that these variables remain unchanged over time or that the baseline measurements are most predictive.
Substantial fluctuation, however, can exist from visit to visit in several of the factors applied to the OHTS risk calculator. Inherent fluctuations in IOP measurements (due to underlying biological mechanisms), CCT measurements (due to device limitations and technique), and VCDR estimates (due to examiner subjectivity) can potentially cause large variations in calculated risk values. In our analysis, this led to considerable variability in the calculated risk percentages with a mean range of 10.5±6.9% and 23.4±13.4% in the −POAG and +POAG groups, respectively. Because we applied the mean VCDR during follow-up to our calculations, thereby eliminating the inherent variability that exists between observers’ estimates of VCDR by ophthalmoscopy, these values likely underestimate the true range of risk percentages a clinician might calculate in real-time in the office setting.19,20
In theory, even greater deviations in calculated risk may potentially manifest at any given follow-up visit if a patient happens to present with a combination of real-time risk factor measurements that maximally distort the predicted risk. For instance, in the BCS analysis, the mean calculated 5-year risk percentages in the −POAG and +POAG groups were 4.2±3.8% and 7.2±4.3%, respectively. Conversely, under the WCS, these values inflate dramatically to 40.5±21.6% and 60.5±24.1%. Consequently, caution must be exercised when applying the glaucoma risk calculators and interpreting their results in daily practice.
The reliance on baseline characteristics of the OHTS glaucoma risk calculator, as with most multivariate prediction models derived from prospective studies, assumes that the risk of progression to glaucoma is linear. However, because of within-subject changes in risk factors during follow-up, correlation between the baseline and updated values may diminish with time.24 This may as a result underestimate the true association between disease rate and the updated risk factor measurements. We determined that the calculated risk for progression to glaucoma increases by 1.3±0.78%/y in OHTs that progress to POAG. This represents a rate that is 30-fold greater than that found in the −POAG group in which the calculated risk increased by only 0.04±0.51%/y. Risk assessment is likely to improve with models that integrate intercurrent risk factors, paralleling what has been accomplished with prediction models for coronary heart disease.17
To our knowledge, our study is the first to demonstrate that calculated risk based on the available glaucoma risk models can fluctuate significantly and should be interpreted with caution. It also indicates that further refinement of the glaucoma prediction models in OHT patients, by accounting for variations in risk factor measurements over time, is needed. Additional work is necessary to define the frequency with which updated risk factors should be incorporated into prediction models. Although multivariate risk calculators provide a rapid and simplified method for global risk assessment, clinical judgment of its applicability must still be exercised, particularly when a patient does not conform to the model population from which the analyses originate.
1. Leske MC.Open-angle glaucoma—an epidemiologic overview.Ophthalmic Epidemiol.2007;14:166–172.
2. Gordon M, Beiser J, Brandt J, et al..The Ocular Hypertension Treatment Study: baseline factors that predict the onset of primary open-angle glaucoma.Arch Ophthalmol.2002;120:714–720.
3. .Predictive factors for open-angle glaucoma among patients with ocular hypertension in the European Glaucoma Prevention Study.Ophthalmology.2007;114:3–9.
4. Leske MC, Connell AMS, Wu SY, et al..Risk factors for open-angle glaucoma: the Barbados Eye Study.Arch Ophthalmol.1995;113:918–924.
5. Tielsch JM, Katz J, Sommer A, et al..Family history and risk of primary open angle glaucoma: the Baltimore Eye Survey.Arch Ophthalmol.1994;112:69–73.
6. Sommer A, Tielsch JM, Katz J, et al..Relationship between intraocular pressure and primary open angle glaucoma among white and black Americans: the Baltimore Eye Study.Arch Ophthalmol.1991;109:1090–1095.
7. Tielsch JM, Katz J, Sommer A, et al..Hypertension, perfusion pressure, and primary open-angle glaucoma: a population-based assessment.Arch Ophthalmol.1995;13:216–221.
8. Weinreb RN, Friedman DS, Fechtner RD, et al..Risk assessment in the management of patients with ocular hypertension.Am J Ophthalmol.2004;138:458–467.
9. Gordon MO, Torri V, Miglior S, et al..Validated prediction model for the development of primary open-angle glaucoma in individuals with ocular hypertension.Ophthalmology.2007;114:10–19.
10. Mansberger SL.A risk calculator to determine the probability of glaucoma.J Glaucoma.2004;13:345–347.
11. Medeiros FA, Weinreb RN, Sample PA, et al..Validation of a predictive model to estimate the risk of conversion from ocular hypertension to glaucoma.Arch Ophthalmol.2005;123:1351–1360.
12. Mansberger SL, Cioffi GA.The probability of glaucoma from ocular hypertension determined by ophthalmologists in comparison to a risk calculator.J Glaucoma.2006;15:426–431.
13. Boland MV, Quigley HA, Lehman HP.The impact of risk calculation on treatment recommendations made by glaucoma specialists in cases of ocular hypertension.J Glaucoma.2008;17:631–638.
14. Anderson KM, Odell PM, Wilson PW, et al..Cardiovascular disease risk profiles.Am Heart J.1991;121:293–298.
15. Assmann G, Cullen P, Schulte H.Simple scoring scheme for calculating the risk of acute coronary events based on the 10-year follow-up of the Prospective Cardiovascular Münster (PROCAM) Study.Circulation.2002;105:310–315.
16. Wilson PW, D’Agostino RB, Levy D, et al..Prediction of coronary heart disease using risk factor categories.Circulation.1998;97:1837–1847.
17. Karp I, Abrahamowicz M, Bartlett G, et al..Updated risk factor values and the ability of the multivariable risk score to predict coronary heart disease.Am J Epidemiol.2004;160:707–716.
18. Kass MA, Heuer DK, Higginbotham EJ, et al..The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma.Arch Ophthalmol.2002;120:701–713.
19. Lichter PR.Variability of expert observers in evaluating the optic disc.Trans Am Ophthalmol Soc.1976;74:532–572.
20. Tielsch JM, Katz J, Quigley HA, et al..Intraobserver and interobserver agreement in measurement of optic disc characteristics.Ophthalmology.1988;95:350–356.
21. Kotecha A, Crabb DP, Spratt A, et al..The relationship between diurnal variations in intraocular pressure measurements and central corneal thickness and corneal hysteresis.Invest Ophthalmol Vis Sci.2009;50:4229–4236.
22. Miglior S, Albe E, Guareschi M, et al..Intraobserver and interobserver reproducibility in the evaluation of ultrasonic pachymetry measurements of central corneal thickness.Br J Ophthalmol.2004;88:174–177.
23. Schwartz JT.Methodologic differences and measurement of cup-disc ratio: an epidemiologic assessment.Arch Ophthalmol.1976;94:1101–1105.
24. Clarke R, Shipley M, Lewington S, et al..Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies.Am J Epidemiol.1999;150:341–353.
© 2014 by Lippincott Williams & Wilkins.