Evaluation of a Primary Open-Angle Glaucoma Prediction Model Using Long-term Intraocular Pressure Variability Data: A Secondary Analysis of 2 Randomized Clinical Trials | Glaucoma | JAMA Ophthalmology | JAMA Network
[Skip to Navigation]
Figure.  Forest Plots for Follow-up Intraocular Pressure (IOP) Measures From Landmark Cox Proportional Hazards Regression Models of the Ocular Hypertension Treatment Study (OHTS) Observation Group and the European Glaucoma Prevention Study (EGPS) Placebo Group
Forest Plots for Follow-up Intraocular Pressure (IOP) Measures From Landmark Cox Proportional Hazards Regression Models of the Ocular Hypertension Treatment Study (OHTS) Observation Group and the European Glaucoma Prevention Study (EGPS) Placebo Group

A, Forest plot of models that substitute a measure of follow-up IOP for baseline IOP in the model that includes baseline age, central cornea thickness (CCT), vertical cup-disc ratio (VCDR), and pattern SD (PSD). B, Forest plot of models that add a measure of long-term IOP variability to a model that includes mean follow-up IOP, baseline age, CCT, VCDR, and PSD. For the hazard ratios (HRs), approximately 1-SD change in the OHTS observation group data occurred as follows: mean IOP, 3 mm Hg; SD IOP, 1 mm Hg; maximum IOP, 4 mm Hg; range of IOP, 3.5 mm Hg; and coefficient of variation (CV) IOP, 0.05 mm Hg. The 95% CIs that include 1.0 are not statistically significant at P < .05.

Table 1.  Demographic and Clinical Characteristics at Baseline and Follow-up in the OHTS Observation Group and the EGPS Placebo Group by POAG Status
Demographic and Clinical Characteristics at Baseline and Follow-up in the OHTS Observation Group and the EGPS Placebo Group by POAG Status
Table 2.  Univariate HRs (95% CIs) and C Statistics From Landmark Cox Proportional Hazards Regression Models for the OHTS Observation Group and the EGPS Placebo Group
Univariate HRs (95% CIs) and C Statistics From Landmark Cox Proportional Hazards Regression Models for the OHTS Observation Group and the EGPS Placebo Group
1.
Leske  MC, Heijl  A, Hussein  M, Bengtsson  B, Hyman  L, Komaroff  E; Early Manifest Glaucoma Trial Group.  Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial.   Arch Ophthalmol. 2003;121(1):48-56. doi:10.1001/archopht.121.1.48 PubMedGoogle Scholar
2.
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(6):701-713. doi:10.1001/archopht.120.6.701 PubMedGoogle Scholar
3.
Lichter  PR, Musch  DC, Gillespie  BW,  et al; CIGTS Study Group.  Interim clinical outcomes in the Collaborative Initial Glaucoma Treatment Study comparing initial treatment randomized to medications or surgery.   Ophthalmology. 2001;108(11):1943-1953. doi:10.1016/S0161-6420(01)00873-9 PubMedGoogle Scholar
4.
Van Veldhuisen  PC, Ederer  F, Gaasterland  DE,  et al; The AGIS Investigators.  The Advanced Glaucoma Intervention Study (AGIS), 7: the relationship between control of intraocular pressure and visual field deterioration.   Am J Ophthalmol. 2000;130(4):429-440. doi:10.1016/S0002-9394(00)00538-9 PubMedGoogle Scholar
5.
Collaborative Normal-Tension Glaucoma Study Group.  The effectiveness of intraocular pressure reduction in the treatment of normal-tension glaucoma.   Am J Ophthalmol. 1998;126(4):498-505. doi:10.1016/S0002-9394(98)00272-4 PubMedGoogle Scholar
6.
Nouri-Mahdavi  K, Hoffman  D, Coleman  AL,  et al; Advanced Glaucoma Intervention Study.  Predictive factors for glaucomatous visual field progression in the Advanced Glaucoma Intervention Study.   Ophthalmology. 2004;111(9):1627-1635. doi:10.1016/j.ophtha.2004.02.017 PubMedGoogle Scholar
7.
Caprioli  J, Coleman  AL.  Intraocular pressure fluctuation a risk factor for visual field progression at low intraocular pressures in the advanced glaucoma intervention study.   Ophthalmology. 2008;115(7):1123-1129.e3. doi:10.1016/j.ophtha.2007.10.031PubMedGoogle Scholar
8.
Musch  DC, Gillespie  BW, Lichter  PR, Niziol  LM, Janz  NK; CIGTS Study Investigators.  Visual field progression in the Collaborative Initial Glaucoma Treatment Study the impact of treatment and other baseline factors.   Ophthalmology. 2009;116(2):200-207. doi:10.1016/j.ophtha.2008.08.051 PubMedGoogle Scholar
9.
Musch  DC, Gillespie  BW, Niziol  LM, Lichter  PR, Varma  R; CIGTS Study Group.  Intraocular pressure control and long-term visual field loss in the Collaborative Initial Glaucoma Treatment Study.   Ophthalmology. 2011;118(9):1766-1773. doi:10.1016/j.ophtha.2011.01.047 PubMedGoogle Scholar
10.
Fujino  Y, Asaoka  R, Murata  H,  et al; Japanese Archive of Multicentral Databases in Glaucoma (JAMDIG) Construction Group.  Evaluation of glaucoma progression in large-scale clinical data: the Japanese Archive of Multicentral Databases in Glaucoma (JAMDIG).   Invest Ophthalmol Vis Sci. 2016;57(4):2012-2020. doi:10.1167/iovs.15-19046 PubMedGoogle Scholar
11.
Stewart  WC, Kolker  AE, Sharpe  ED,  et al.  Factors associated with long-term progression or stability in primary open-angle glaucoma.   Am J Ophthalmol. 2000;130(3):274-279. doi:10.1016/S0002-9394(00)00487-6 PubMedGoogle Scholar
12.
Hong  S, Seong  GJ, Hong  YJ.  Long-term intraocular pressure fluctuation and progressive visual field deterioration in patients with glaucoma and low intraocular pressures after a triple procedure.   Arch Ophthalmol. 2007;125(8):1010-1013. doi:10.1001/archopht.125.8.1010 PubMedGoogle Scholar
13.
Lee  PP, Walt  JW, Rosenblatt  LC, Siegartel  LR, Stern  LS; Glaucoma Care Study Group.  Association between intraocular pressure variation and glaucoma progression: data from a United States chart review.   Am J Ophthalmol. 2007;144(6):901-907. doi:10.1016/j.ajo.2007.07.040 PubMedGoogle Scholar
14.
Bengtsson  B, Leske  MC, Hyman  L, Heijl  A; Early Manifest Glaucoma Trial Group.  Fluctuation of intraocular pressure and glaucoma progression in the early manifest glaucoma trial.   Ophthalmology. 2007;114(2):205-209. doi:10.1016/j.ophtha.2006.07.060 PubMedGoogle Scholar
15.
De Moraes  CG, Juthani  VJ, Liebmann  JM,  et al.  Risk factors for visual field progression in treated glaucoma.   Arch Ophthalmol. 2011;129(5):562-568. doi:10.1001/archophthalmol.2011.72 PubMedGoogle Scholar
16.
Medeiros  FA, Weinreb  RN, Zangwill  LM,  et al.  Long-term intraocular pressure fluctuations and risk of conversion from ocular hypertension to glaucoma.   Ophthalmology. 2008;115(6):934-940. doi:10.1016/j.ophtha.2007.08.012 PubMedGoogle Scholar
17.
Gordon  MO, Torri  V, Miglior  S,  et al; Ocular Hypertension Treatment Study Group; European Glaucoma Prevention Study Group.  Validated prediction model for the development of primary open-angle glaucoma in individuals with ocular hypertension.   Ophthalmology. 2007;114(1):10-19. doi:10.1016/j.ophtha.2006.08.031 PubMedGoogle Scholar
18.
Gordon  MO, Kass  MA.  The Ocular Hypertension Treatment Study: design and baseline description of the participants.   Arch Ophthalmol. 1999;117(5):573-583. doi:10.1001/archopht.117.5.573 PubMedGoogle Scholar
19.
Miglior  S, Zeyen  T, Pfeiffer  N, Cunha-Vaz  J, Torri  V, Adamsons  I; European Glaucoma Prevention Study Group.  The European glaucoma prevention study design and baseline description of the participants.   Ophthalmology. 2002;109(9):1612-1621. doi:10.1016/S0161-6420(02)01167-3 PubMedGoogle Scholar
20.
Ederer  F, Gaasterland  DE, Sullivan  EK; AGIS Investigators.  The Advanced Glaucoma Intervention Study (AGIS), 1: study design and methods and baseline characteristics of study patients.   Control Clin Trials. 1994;15(4):299-325. doi:10.1016/0197-2456(94)90046-9 PubMedGoogle Scholar
21.
Gordon  MO, Higginbotham  EJ, Heuer  DK,  et al; Ocular Hypertension Treatment Study.  Assessment of the impact of an endpoint committee in the Ocular Hypertension Treatment Study.   Am J Ophthalmol. 2019;199:193-199. doi:10.1016/j.ajo.2018.11.006 PubMedGoogle Scholar
22.
van Houwelingen  HC.  Dynamic prediction by landmarking in event history analysis.   Scand J Stat. 2007;34(1):70-85. doi:10.1111/j.1467-9469.2006.00529.x Google Scholar
23.
Fisher  LD, Lin  DY.  Time-dependent covariates in the Cox proportional-hazards regression model.   Annu Rev Public Health. 1999;20:145-157. doi:10.1146/annurev.publhealth.20.1.145 PubMedGoogle Scholar
24.
Matlach  J, Bender  S, König  J, Binder  H, Pfeiffer  N, Hoffmann  EM.  Investigation of intraocular pressure fluctuation as a risk factor of glaucoma progression.   Clin Ophthalmol. 2018;13:9-16. doi:10.2147/OPTH.S186526PubMedGoogle Scholar
25.
Hair  JF  Jr, Black  WC, Babin  BJ, Anderson  RE.  Multivariate Data Analysis. 7th ed. Pearson Education Ltd; 2010.
26.
Gordon  MO, Beiser  JA, Brandt  JD,  et al.  The Ocular Hypertension Treatment Study: baseline factors that predict the onset of primary open-angle glaucoma.   Arch Ophthalmol. 2002;120(6):714-720. doi:10.1001/archopht.120.6.714 PubMedGoogle Scholar
27.
Miglior  S, Pfeiffer  N, Torri  V, Zeyen  T, Cunha-Vaz  J, Adamsons  I; European Glaucoma Prevention Study (EGPS) Group.  Predictive factors for open-angle glaucoma among patients with ocular hypertension in the European Glaucoma Prevention Study.   Ophthalmology. 2007;114(1):3-9. doi:10.1016/j.ophtha.2006.05.075 PubMedGoogle Scholar
28.
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(10):1351-1360. doi:10.1001/archopht.123.10.1351 PubMedGoogle Scholar
29.
Miglior  S, Torri  V, Zeyen  T, Pfeiffer  N, Vaz  JC, Adamsons  I; EGPS Group.  Intercurrent factors associated with the development of open-angle glaucoma in the European glaucoma prevention study.   Am J Ophthalmol. 2007;144(2):266-275. doi:10.1016/j.ajo.2007.04.040PubMedGoogle Scholar
30.
Gao  F, Miller  JP, Xiong  C, Beiser  JA, Gordon  M; The Ocular Hypertension Treatment Study (OHTS) Group.  A joint-modeling approach to assess the impact of biomarker variability on the risk of developing clinical outcome.   Stat Methods Appt. 2011;20(1):83-100. doi:10.1007/s10260-010-0150-zPubMedGoogle Scholar
31.
van Houwelingen  HC, Putter  H.  Dynamic Prediction in Clinical Survival Analysis. CRC Press Inc; 2011. doi:10.1201/b11311
32.
Jiang  X, Torres  M, Varma  R; Los Angeles Latino Eye Study Group.  Variation in intraocular pressure and the risk of developing open-angle glaucoma: the Los Angeles Latino Eye Study.   Am J Ophthalmol. 2018;188:51-59. doi:10.1016/j.ajo.2018.01.013 PubMedGoogle Scholar
33.
Kim  JH, Caprioli  J.  Intraocular pressure fluctuation: is it important?   J Ophthalmic Vis Res. 2018;13(2):170-174. doi:10.4103/jovr.jovr_35_18 PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    June 4, 2020

    Evaluation of a Primary Open-Angle Glaucoma Prediction Model Using Long-term Intraocular Pressure Variability Data: A Secondary Analysis of 2 Randomized Clinical Trials

    Author Affiliations
    • 1Department of Ophthalmology and Visual Sciences, Washington University School of Medicine in St Louis, St Louis, Missouri
    • 2Division of Biostatistics, Washington University School of Medicine in St Louis, St Louis, Missouri
    • 3Department of Surgery, Washington University School of Medicine in St Louis, St Louis, Missouri
    • 4Policlinico di Monza University of Milano-Bicocca, Milan, Italy
    • 5Department of Oncology, Istituto di Ricerche Farmacologiche “Mario Negri,” Milan, Italy
    JAMA Ophthalmol. 2020;138(7):780-788. doi:10.1001/jamaophthalmol.2020.1902
    Key Points

    Question  Do long-term intraocular pressure variability data improve a prediction model for which individuals with untreated ocular hypertension will develop primary open-angle glaucoma?

    Findings  In this post hoc secondary analysis of 2 randomized clinical trials that included 709 individuals, the mean follow-up intraocular pressure improved a prediction model for developing primary open-angle glaucoma that included the following baseline factors: age, central corneal thickness, vertical cup-disc ratio, and pattern SD. Adding intraocular pressure SD, maximum, range, or coefficient of variation to a model that included mean follow-up intraocular pressure and baseline factors did not significantly increase predictive accuracy.

    Meaning  These findings suggest that the inclusion of data on long-term intraocular pressure variability are unlikely to improve prediction models for the development of primary open-angle glaucoma in individuals with untreated ocular hypertension.

    Abstract

    Importance  The contribution of long-term intraocular pressure (IOP) variability to the development of primary open-angle glaucoma is still controversial.

    Objective  To assess whether long-term IOP variability data improve a prediction model for the development of primary open-angle glaucoma (POAG) in individuals with untreated ocular hypertension.

    Design, Setting, and Participants  This post hoc secondary analysis of 2 randomized clinical trials included data from 709 of 819 participants in the observation group of the Ocular Hypertension Treatment Study (OHTS) followed up from February 28, 1994, to June 1, 2002, and 397 of 500 participants in the placebo group of the European Glaucoma Prevention Study (EGPS) followed up from January 1, 1997, to September 30, 2003. Data analyses were completed between January 1, 2019, and March 15, 2020.

    Exposures  The original prediction model for the development of POAG included the following baseline factors: age, IOP, central corneal thickness, vertical cup-disc ratio, and pattern SD. This analysis tested whether substitution of baseline IOP with mean follow-up IOP, SD of IOP, maximum IOP, range of IOP, or coefficient of variation IOP would improve predictive accuracy.

    Main Outcomes and Measures  The C statistic was used to compare the predictive accuracy of multivariable landmark Cox proportional hazards regression models for the development of POAG.

    Results  Data from the OHTS consisted of 97 POAG end points from 709 of 819 participants (416 [58.7%] women; 177 [25.0%] African American and 490 [69.1%] white; mean [SD] age, 55.7 [9.59] years; median [range] follow-up, 6.9 [0.96-8.15] years). Data from the EGPS consisted of 44 POAG end points from 397 of 500 participants in the placebo group (201 [50.1%] women; 397 [100%] white; mean [SD] age, 57.8 [9.76] years; median [range] follow-up, 4.9 [1.45-5.76] years). The C statistic for the original prediction model was 0.741. When a measure of follow-up IOP was substituted for baseline IOP in this prediction model, the C statistics were as follows: mean follow-up IOP, 0.784; maximum IOP, 0.781; SD of IOP, 0.745; range of IOP, 0.741; and coefficient of variation IOP, 0.729. The C statistics in the EGPS were similarly ordered. No measure of IOP variability, when added to the prediction model that included mean follow-up IOP, age, central corneal thickness, vertical cup-disc ratio, and pattern SD, increased the C statistic by more than 0.007 in either cohort.

    Conclusions and Relevance  Evidence from the OHTS and the EGPS suggests that long-term variability does not add substantial explanatory power to the prediction model as to which individuals with untreated ocular hypertension will develop POAG.

    Introduction

    Higher intraocular pressure (IOP) is an important risk factor for the development and progression of glaucoma and currently the only modifiable factor.1-5 However, mean follow-up IOP does not capture the dynamic changes in pressure from visit to visit, such as peaks, troughs, and ranges that may be independently associated with the development and course of the disease.

    Controversy still exists about the effects of long-term variability of IOP. Some studies6-13 have found that variability is an independent risk factor for glaucoma development or progression beyond the mean follow-up IOP, whereas other studies2,14-16 have not. This controversy may reflect objective differences in the studies, including diagnosis (ocular hypertension, normal-tension glaucoma, or high-tension glaucoma), the range of IOP in the patients, the duration of follow-up and frequency of pressure measurements, the method of tonometry, and whether the patients received medical or surgical therapy. Patients with worsening primary open-angle glaucoma (POAG) or IOP above a therapeutic goal may undergo multiple changes in treatment, which may confound the association between disease course and IOP variability. In addition, poor adherence and persistence with medication or poor tolerance of medication may be associated with IOP variability and disease progression. Data were previously analyzed on long-term IOP variability in the entire Ocular Hypertension Treatment Study (OHTS) and European Glaucoma Prevention Study (EGPS) cohorts.17 In view of the continuing controversy about the effects of long-term IOP variability and because practitioners are uncertain about how to integrate these measures into clinical care, we evaluated this topic again. To reduce the influence of potential confounding factors and to assess the association of long-term IOP variability with the incidence of POAG in patients with ocular hypertension, we restricted the analysis sample to include only participants enrolled in the observation group of the OHTS and the placebo group of the EGPS. To test reproducibility of results from OHTS, data from EGPS were analyzed separately and reported separately.

    Methods

    This post hoc secondary analysis of 2 randomized clinical trials used data from the OHTS and EGPS. The OHTS2,18 and the EGPS19 were randomized clinical trials that tested the safety and efficacy of topical ocular hypotensive medication in delaying or preventing the development of POAG in individuals with ocular hypertension. The OHTS and the EGPS protocols are described in their respective baseline design articles18,19 but are briefly described here. The protocols of the OHTS and EGPS were approved by the institutional review boards of each participating clinic and resource center. Each participant provided written informed consent. All data were deidentified.

    In the OHTS phase 1, participants were randomized in equal proportions to receive topical ocular hypotensive medication or close observation. The participants were randomized to observation on February 28, 1994, and followed up to the end of phase 1 on June 1, 2002, and had at least 2 postrandomization visits and complete data for baseline factors (age, IOP, central corneal thickness [CCT], pattern SD [PSD], and vertical cup-disc ratio [VCDR]). In the EGPS, participants were randomized to receive dorzolamide or placebo drops. The EGPS participants were randomized to placebo on January 1, 1997, and followed up to September 30, 2003; these individuals had at least 2 follow-up visits after the 6-month surrogate baseline and complete data for baseline predictors. Key similarities between the studies were (1) similar inclusion and exclusion criteria; (2) IOP measurements by Goldmann tonometry using the Advanced Glaucoma Intervention Study (AGIS) protocol20; (3) follow-up visits at 6-month intervals for 5 years or until a censoring event; (4) masked determination and reproducibility of optic nerve deterioration and/or visual field abnormality; and (5) masked end point committee to determine whether reproducible optic nerve deterioration and/or visual field abnormality were cause by POAG.

    In the OHTS and EGPS, the end point date for POAG was defined as the date of the first abnormal visual field or optic disc photograph that masked readers determined met the criteria for reproducible change and the end point committee attributed to POAG. The protocol for ascertainment of POAG end points is described in detail elsewhere.21 Data were censored after POAG diagnosis or after initiation of ocular hypotensive treatment.

    Measures of IOP

    In the OHTS and EGPS, IOP was assessed every 6 months by Goldmann applanation tonometry using the AGIS protocol.20 We calculated the following measures of follow-up IOP: (1) mean follow-up IOP, (2) SD IOP, (3) maximum IOP, (4) range of IOP, and (5) coefficient of variation (CV) IOP (SD IOP divided by mean follow-up IOP).

    Statistical Analysis

    We used an adaptation of conventional Cox proportional hazards regression modeling (ie, landmark analysis)22 because conventional Cox proportional hazards regression modeling does not permit calculation of predictive accuracy (C statistic) with time-dependent covariates.23 Previous studies16,24 have calculated measures of long-term IOP variability by aggregating all IOPs from baseline to a censoring event or end of follow-up. Aggregate measures do not capture time-dependent changes and differences in duration of follow-up. In a landmark analysis, conventional Cox proportional hazards regression models are recalculated at prespecified time points (landmark point) using participants still at risk of developing an end point. Time-dependent covariates (eg, measures of IOP variability) are refreshed with all IOPs from baseline to the landmark point. Van Houwelingen et al22 found that the time-dependent effects can be well approximated with this approach. A landmark analysis enables the application of a wealth of theory developed for conventional Cox proportional hazards regression models to calculate predictive accuracy (C statistic) and graphic tools, such as Kaplan-Meier estimates.

    A minimum of 5 to 10 future end points are needed per covariate for a reliable prediction model.25 Thus, with as many as 6 factors in a multivariable model (mean follow-up IOP and a measure of IOP variability plus 4 factors from the original baseline model: age, CCT, VCDR, and PSD), landmark models could be fit when at least 30 future POAG conversions were available. In the OHTS, 9 landmark analyses were possible at 12 months with 97 end points in the future and so forth until the last analysis at 60 months with 38 end points in the future. In the EGPS, 3 landmark analyses were possible at 18 months with 44 end points, 24 months with 34 end points, and 30 months with 30 end points in the future. At each landmark point, Cox proportional hazards regression models were run for each measure of follow-up IOP as follows. First, we ran univariate models to estimate the contribution of each measure of follow-up IOP with no adjustment for baseline factors. Second, we substituted a measure of follow-up IOP for baseline IOP in the original prediction model that included baseline age, CCT, VCDR, and PSD. Third, we added a measure of IOP variability (SD IOP, maximum IOP, range of IOP, or CV IOP) to a model that included mean follow-up IOP, baseline age, CCT, VCDR, and PSD. Time zero was reset at each landmark point.

    At each landmark point, predictive accuracy (C statistic) and hazard ratios (HRs) were calculated for each Cox proportional hazards regression model. The C statistics across landmark points of each model were averaged to estimate predictive accuracy over the entire follow-up period. The mean HRs across the landmark points were estimated using Cox proportional hazards regression models stratified by landmark points, and 95% CIs were obtained based on the robust sandwich variance. The homogeneity of the HRs across strata was assessed by testing the interaction term between landmark points and each IOP measure in the Cox proportional hazards regression model. To enable comparisons of HRs among different variables and models and across the OHTS and EGPS studies, IOP measures were standardized by scaling them to the SD of the variable in the OHTS data (ie, 3 mm Hg for mean IOP, 1 mm Hg for SD IOP, 4 mm Hg for maximum IOP, 3.5 mm Hg for range IOP, and 0.05 for CV IOP). All analyses were 2-sided, and statistical significance was set at P < .05. Statistical analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc) and the survival library in the statistical package R (R Foundation for Statistical Computing). The rationale for a landmark rather than conventional multivariable Cox proportional hazards regression models is provided in the eAppendix in the Supplement. Conventional Cox proportional hazards regression modeling was also performed, and these results compared with those form the landmark analyses are given in in eTable 1 in the Supplement. Data analyses were completed between January 1, 2019, and March 15, 2020.

    Results

    Data from the OHTS consisted of 97 POAG end points from 709 of 819 participants (416 [58.7%] women; 177 [25.0%] African American and 490 [69.1%] white; mean [SD] age, 55.7 [9.59] years; median [range] follow-up, 6.9 [0.96-8.15] years). Data from the EGPS consisted of 44 POAG end points from 397 of 500 placebo participants (201 [50.1%] women; 397 [100%] white; mean [SD] age, 57.8 [9.76] years; median [range] follow-up, 4.9 [1.45-5.76] years). Descriptive statistics for demographic, baseline, and follow-up measures in the OHTS and EGPS are reported in Table 1. The OHTS participants had a median of 14 IOP measurements (range, 3-17) and a median follow-up of 6.9 years (range, 1.0-8.1 years). The EGPS participants had a median of 10 IOP measurements (range, 3-10) with a median follow-up of 4.9 years (range, 1.5-5.8 years).

    In the OHTS, the highest correlations (Pearson correlation coefficient) were among measure of follow-up IOP variability (SD IOP, range of IOP, and CV IOP), which ranged from 0.868 to 0.917. Maximum IOP was moderately correlated with other measures of IOP variability (Pearson correlation coefficient range, 0.325-0.640) and highly correlated with mean follow-up IOP (Pearson correlation coefficient, 0.852). Mean follow-up IOP was modestly correlated with other measures of IOP variability with correlations from −0.119 to 0.256 (eTable 2 in the Supplement). The correlational structure among measures of follow-up IOP in the EGPS closely resembled those of the OHTS (eTable 3 in the Supplement).

    Results from the univariate landmark Cox proportional hazards regression models are reported in Table 2. In the OHTS, all measures of follow-up IOP (mean follow-up IOP: HR, 2.03; 95% CI, 1.68-2.46; SD IOP: HR, 1.29; 95% CI, 1.14-1.46; maximum IOP: HR, 1.93; 95% CI, 1.64-2.26; and range of IOP: HR, 1.40; 95% CI, 1.18-1.66), except CV IOP (HR, 1.14; 95% CI, 0.96-1.36), were statistically significant in univariate models, which did not include baseline factors (Table 2). The largest differences in measures of follow-up IOP between participants who did not develop POAG and those who developed POAG were in mean follow-up IOP (HR, 2.03; 95% CI, 1.68-2.46) and maximum IOP (HR, 1.93; 95% CI, 1.64-2.26) (Table 2). Univariate models with the highest C statistics were for mean follow-up IOP (C statistic, 0.705) and maximum IOP (C statistic, 0.707) (Table 2). Similarly, in the EGPS, the greatest difference between participants who did not develop POAG and those who did was in mean follow-up IOP (HR, 1.58; 95% CI, 1.18-2.12) and maximum IOP (HR, 1.66; 95% CI, 1.21-2.21). C statistics were 0.622 for both measures in univariate models. SD IOP (HR, 1.17; 95% CI, 0.91-1.51), range of IOP (HR, 1.30; 95% CI, 0.86-1.97), and CV IOP (HR, 1.04; 95% CI, 0.79-1.37) were not statistically significant and had C statistics of 0.552 (SD IOP), 0.556 (range of IOP), and 0.523 (CV IOP).

    We examined whether the predictive accuracy of the original prediction model that included baseline age, IOP, CCT, VCDR, and PSD could be improved by substituting baseline IOP with a measure of follow-up IOP. The predictive accuracy of the original model as measured by the C statistic was 0.741, in which a C statistic of 0.50 indicates random accuracy and 1.00 indicates perfect accuracy. When baseline IOP was substituted by mean follow-up IOP, the C statistic for the model increased to 0.784 and 0.781 when baseline IOP was substituted by maximum IOP. The C statistic for the model with SD IOP (C statistic, 0.745) showed an improvement, but substituting baseline IOP with either range of IOP (C statistic, 0.741) or CV IOP (C statistic, 0.729) did not improve the C statistic. In the EGPS, the C statistic for the original prediction model with baseline IOP, age, CCT, VCDR, and PSD was 0.723, which did not increase when baseline IOP was substituted by mean follow-up IOP (C statistic, 0.710), SD IOP (C statistic, 0.694), maximum IOP (C statistic, 0.706), range of IOP (C statistic, 0.694), or CV IOP (C statistic, 0.694) (Figure, A).

    To assess whether any measure of IOP variability could improve the predictive accuracy of a model that included mean follow-up IOP and baseline factors (age, CCT, VCDR, and PSD), we added measures of IOP variability to the model 1 at time (SD IOP, maximum IOP, range of IOP, or CV IOP). In the OHTS, no measure of IOP variability when added to this model was statistically significant (Figure, B): SD IOP (HR, 1.10; 95% CI, 0.96-1.26), maximum IOP (HR, 1.07; 95% CI, 0.75-1.53), range of IOP (HR, 1.11; 95% CI 0.93-1.33), and CV IOP (HR, 1.15; 95% CI, 0.97-1.36). Furthermore, adding a measure of IOP variability to the baseline model that included mean follow-up IOP did not meaningfully increase the C statistic (0.784). C statistics were 0.791 for SD IOP, 0.785 for maximum IOP, 0.789 for range of IOP, and 0.791 for CV IOP. The largest increase in the C statistic from adding IOP variability to the model was only 0.007 (Figure, B). Results in the EGPS were comparable to those of the OHTS. In the EGPS, no measure of IOP variability was statistically significant when added to the model that included mean follow-up IOP and baseline factors (SD IOP: HR, 1.11; 95% CI, 0.88-1.41; maximum IOP: HR, 1.45; 95% CI, 0.70-3.02; range of IOP: HR, 1.20; 95% CI, 0.82-1.77; and CV IOP: HR, 1.11; 95% CI, 0.84-1.45). Nor did the C statistic of 0.710 for the model with mean follow-up IOP and baseline factors increase with the addition of SD IOP (C statistic, 0.708), maximum IOP (C statistic, 0.709), range of IOP (C statistic, 0.708), or CV IOP (C statistic, 0.710) (Figure, B). None of the interactions between mean follow-up IOP and measures of IOP variability were statistically significant in the OHTS or EGPS.

    Discussion

    Because IOP is currently the only modifiable risk factor for glaucoma, understanding how dynamic variation in IOP is associated with onset and progression of the disease may play a crucial role in management. The OHTS26 previously reported that a baseline model that included age, IOP, CCT, VCDR, and PSD was useful in identifying which participants with ocular hypertension were at higher risk for developing POAG. This baseline model was confirmed in the EGPS27 and by Medeiros et al.28 In this study, we examined whether predictive accuracy could be improved by adding mean follow-up IOP and measures of long-term IOP variability to the baseline prediction model. This question was tested using data from patients with ocular hypertension not receiving active treatment to avoid possible confounding owing to treatment changes, medication intolerance, and/or medication adherence and persistence that could be associated with increased IOP variability and the risk of POAG. Previous reports29,30 from the OHTS and EGPS on the association between IOP variability and risk of developing POAG pooled treated and untreated participants, possibly obscuring the true association between IOP variability and risk of developing glaucoma. Because of the continuing controversy about IOP variability and disease progression, we reanalyzed data using only participants in the OHTS observation group and the EGPS placebo group.

    In the OHTS, all measures of follow-up IOP (mean follow-up IOP, SD IOP, maximum IOP, and range of IOP) except CV IOP were moderately predictive of POAG in univariate landmark Cox proportional hazards regression models, with C statistics ranging from 0.604 to 0.707. The highest C statistics were observed for mean follow-up IOP (C statistic, 0.705) and maximum IOP (C statistic, 0.707) owing largely to their high intercorrelation of 0.852. The C statistic for the original prediction model (0.741), which included baseline age, IOP, CCT, VCDR, and PSD, increased to 0.784 when baseline IOP was substituted with mean follow-up IOP and 0.781 when substituted with maximum IOP. None of the other measures of follow-up IOP performed as well. We then assessed whether predictive accuracy could be improved by adding a measure of long-term IOP variability to a model that included mean follow-up IOP and baseline factors. In these models, none of the measures of long-term IOP variability were statistically significant, and their addition to the model did not meaningfully improve the C statistic. The largest improvement in the C statistic was 0.007. These results were replicated in the EGPS cohort.

    We analyzed data using a landmark analysis adaptation of the Cox proportional hazards regression model, which has been used widely in many other fields of medicine to incorporate time-varying covariates.31 In landmark analysis, conventional Cox proportional hazards regression models are run at each landmark point as time 0, thus adjusting for time-dependent covariates and variable duration of follow-up and enabling the calculation of predictive accuracy (C statistic). In contrast, conventional Cox proportional hazards regression models use all data accrued up to a censoring event, thereby failing to take into account that affected individuals have a shorter follow-up than unaffected individuals because of censoring. Because factors such as maximum IOP and IOP range are sensitive to the number of IOP measurements, conventional Cox proportional hazards regression models can lead to a paradoxical result (ie, higher maximum IOP and greater IOP range can appear to be protective of POAG). This statistical paradox is shown using the OHTS data in eTable 1 in the Supplement.

    Our analyses suggest that IOP variability does not substantially add to the prediction model as to which patients with ocular hypertension will develop POAG. Mean follow-up IOP was a stronger predictive factor than any measure of long-term IOP variability in the prediction model. In addition, mean follow-up IOP is easy to calculate, more consistent over time, and less dependent on duration of follow-up. The results of this report agree with those reported by the Diagnostic Innovations in Glaucoma Study,16 which also found no association between long-term IOP variability and the development of POAG in 126 individuals with untreated ocular hypertension. In contrast, the population-based Los Angeles Latino Eye Study32 reported an association of higher IOP variability (maximum IOP, range IOP, and SD IOP) with the risk of POAG in untreated individuals for whom IOP variability was calculated from 3 IOP measurements at baseline and 3 IOP measurements 5 years later. This association between greater IOP variability and development of POAG was statistically significant only in the patients in the lowest tertile of IOP. A recent review33 concluded that the association of IOP variability with development of POAG was strongest among treated patients with low mean levels of IOP.

    Limitations

    This report has several limitations, and the results must be interpreted cautiously. The OHTS and EGPS were limited to participants with ocular hypertension with baseline IOPS in the range of 21 to 32 mm Hg. Furthermore, the OHTS and EGPS participants were receiving no active ocular hypotensive therapy. Thus, these results should be not be extrapolated to patients with lower levels of IOP, treated patients, or patients with glaucoma. Both the OHTS and EGPS used the AGIS protocol for measuring IOP, which is more extensive and standardized than is typical in many clinical settings.20 The IOP measurements were taken during typical office hours and did not include diurnal or nocturnal measurements.

    Conclusions

    In this study, mean follow-up IOP improved a prediction model for developing POAG that included the following baseline factors: age, CCT, VCDR, and PSD. Adding SD IOP, maximum, range, or coefficient of variation to a model that included mean follow-up IOP and baseline factors did not significantly increase predictive accuracy. These findings suggest that the inclusion of data on long-term IOP variability are unlikely to improve prediction models for the development of POAG in individuals with untreated ocular hypertension.

    Back to top
    Article Information

    Accepted for Publication: April 14, 2020.

    Published Online: June 4, 2020. doi:10.1001/jamaophthalmol.2020.1902

    Correction: The article was corrected on July 9, 2020, to fix the spelling of a surname in the byline.

    Corresponding Author: Mae O. Gordon, PhD, Department of Ophthalmology, Washington University School of Medicine in St Louis, 660 S Euclid Ave, Campus Box 8096, St Louis, MO 63110 (mae@wustl.edu).

    Author Contributions: Drs Gordon and Gao had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Gordon, Gao, Miller.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Gordon, Gao, Huecker, Margolis, Kass.

    Critical revision of the manuscript for important intellectual content: Gordon, Gao, Miller, Margolis, Miglior, Torri.

    Statistical analysis: Gordon, Gao, Huecker, Miller.

    Obtained funding: Gordon, Gao, Kass.

    Administrative, technical, or material support: Gordon, Huecker, Margolis.

    Supervision: Gordon, Torri.

    Conflict of Interest Disclosures: Drs Gordon, Gao, and Kass reported receiving grants from the National Eye Institute during the conduct of the study. Dr Miller reported receiving grants from the National Institutes of Health during the conduct of the study. No other disclosures were reported.

    Funding/Support: This research was funded by grants 5UG1EY025180 (Dr Kass), 5UG1EY02518 (Dr Gordon)1, 5UG1EY025182 (Dr Gordon), 5UG1EY025183 (Dr Kass), and 5R03EY015498 (Dr Gordon) from the National Eye Institute.

    Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    The Ocular Hypertension Treatment Study members: Clinical centers: Bascom Palmer Eye Institute, University of Miami, Miami, Florida. Investigators: Donald L. Budenz, MD; Francisco E. Fantes, MD; Steven J. Gedde, MD; Richard K. Parrish II, MD. Coordinators and Staff: Madeline L. Del Calvo, BS. Eye Consultants of Atlanta (formerly M. Angela Vela, MD, PC), Atlanta, Georgia: Investigators: Thomas S. Harbin, Jr, MD; Paul McManus, MD; Charles J. Patorgis, OD; Coordinators and Staff: Laura Brannon, COMT; Montana L. Hooper, COT; Stacey S. Goldstein, COMT; June M. LaSalle Gartlir, COA; Debbie L. Lee, COT; Michelle D. Mondshein; Marianne L. Perry, COT; Julie M. Wright, COT. Cullen Eye Institute, Baylor College of Medicine, Houston, Texas. Investigators: Ronald L. Gross, MD; Silvia Orengo-Nania, MD. Coordinators and staff: Pamela M. Frady, COMT, CCRC; Benita D. Slight, COT, EMT-P, CCRP. Devers Eye Institute, Portland, Oregon. Investigators: George A. (Jack) Cioffi, MD; Elizabeth Donohue, MD; Steven Mansberger, MD; E. Michael VanBuskirk, MD. Coordinators and staff: Kathryn Sherman; JoAnne M. Fraser, COT. Emory University Eye Center, Atlanta, Georgia. Investigators: Allen D. Beck, MD; Anastasias Costarides, MD, PhD. Coordinators and staff: Donna Leef, MMSc, COMT. Henry Ford Medical Center, Detroit, Michigan. Investigators: G. Robert Lesser, MD; Deborah Darnley-Fisch, MD; Nauman R. Imami, MD. Coordinators and staff: Melanie Gutkowski, COMT, CO; Jim Bryant, COT; Ingrid Crystal Fugmann, COMT; Monica R. Gibson, COT; Wendy Gilroy, COMT; Monica Lacoursiere; Sue Loomis, COT; Lauren S. Turner. Johns Hopkins University School of Medicine, Baltimore, Maryland. Investigators: Donald J. Zack, MD, PhD; Donald A. Abrams, MD; Nathan G. Congdon, MD, MPH; Robert A. Copeland, MD; David S. Friedman, MD, MPH; Ramzi Hemady, MD; Eve J. Higginbotham, MD; Henry D. Jampel, MD, MHS; Irvin P. Pollack, MD; Harry A. Quigley, MD; Alan L. Robin, MD; Nancy E. Williams, MD. Coordinators and staff: Rachel Scott, BS, COA; Rani Kalsi; Felicia Keel, COT; Lisa Levin; Robyn Priest-Reed, MMSc. Charles R. Drew University, Jules Stein Eye Institute, University of California, Los Angeles. Investigators: Anne L. Coleman, MD, PhD; Richard S. Baker, MD; Michelle C. Banks, MD; Y. P. Dang, MD; Simon K. Law, MD; Dana P. Tannenbaum, MD. Coordinators and staff: Jackie R. Sanguinet, BS, COT; Bobbie Ballenberg, COMT; Salvador Murillo. W. K. Kellogg Eye Center, Ann Arbor, Michigan. Investigators: Terry J. Bergstrom, MD; Sayoko E. Moroi, MD, PhD. Coordinators and staff: Carol J. Pollack-Rundle, BS, COMT; Michelle A. Tehranisa, COT. Kresge Eye Institute, Wayne State University, Detroit, Michigan. Investigators: Bret A. Hughes, MD; Monica Y. Allen, MD; Mark S. Juzych, MD; Mark L. McDermott, MD; John M. O’Grady, MD; John M. Ramocki, MD; Dian Shi, MD; Dong H. Shin, MD, PhD. Coordinators and staff: Juan Allen; Laura L. Schulz, CNA; Linda A. VanConett, COT; Mary B. Hall. University of Louisville, Louisville, Kentucky. Investigators: Joern B. Soltau, MD; Gustavo E. Gamero, MD; Judit Mohay, MD; Gökhan Özdemir, MD; Robb R. Shrader, MD. Coordinators and staff: Sandy Lear, RN; Kathleen Coons, COT. Mayo Clinic/Foundation, Rochester, Minnesota: Investigators: David C. Herman, MD; Douglas H. Johnson, MD; Paul H. Kalina, MD. Coordinators and staff: Becky A. Nielsen, LPN; Nancy J. Tvedt. New York Eye & Ear Infirmary, New York. Investigators: Jeffrey M. Liebmann, MD; Robert Ritch, MD; Celso A. Tello, MD. Coordinators and staff: Jean L. Walker, BS, COA; Deborah L. Simon, COA. Ohio State University, Columbus. Investigators: Paul Weber, MD; N. Douglas Baker, MD; Robert J. Derick, MD; David Lehmann, MD; Omar Mobin-Uddin, MD. Coordinators and staff: Kathryne McKinney, COMT; Cynthia Hutchinson; Diane Moore, COA; Cynthia A. Williams. Pennsylvania College of Optometry/MCP Hahnemann University School of Medicine, Philadelphia. Investigators: G. Richard Bennett, MS, OD; Elliot Werner, MD; Myron Yanoff, MD. Coordinators and staff: Lindsay C. Bennett, BA; Mary Jameson, Opt, TR. Scheie Eye Institute, University of Pennsylvania, Philadelphia. Investigators: Jody R. Piltz-Seymour, MD; Teresa L. Brevetti, MD. Coordinators and staff: Jane L. Anderson, MS, CCRC; Cheryl McGill, COA; Janice T. Petner, COA. University of California–Davis, Sacramento. Investigators: James D. Brandt, MD; Jeffrey J. Casper, MD; John T. Dragicevich, MD; Thomas R. Johansen, MD; Esther S. Kim, MD; Michele C. Lim, MD; Michael B. Mizoguchi, MD; Alan M. Roth, MD; Ivan R. Schwab, MD. Coordinators and staff: Ingrid J. Clark, COA; Denise M. Owensby, BS, COA; Marilyn A. Sponzo, COA. University of California–San Diego, La Jolla. Investigators: Robert N. Weinreb, MD; J. Rigby Slight, MD. Coordinators and staff: Eva Kroneker, COA. University of California–San Francisco. Investigators: Michael V. Drake, MD; Allan J. Flach, MD; Robert Stamper, MD. Coordinators and staff: Fermin P. Ballesteros, Jr; Marjan Karimabadi; Valerie Margol; Peggy Yamada, COT. University Suburban Health Center, South Euclid, Ohio. Investigators: Kathleen A. Lamping, MD; Gregory A. Eippert, MD. Coordinators and staff: Angela K. McKean; Bettina J. Modica; Tonya Sims. Washington OHTS Center, Washington, DC. Investigators: Douglas E. Gaasterland, MD; Frank S. Ashburn, MD; Arthur L. Schwartz, MD; Howard S. Weiss, MD. Coordinators and staff: Robin L. Montgomery, COA; Donna M. Claggert; Karen D. Schacht, COT; Anne M. Bocckl, MS. Washington University School of Medicine, St Louis, Missouri. Investigators: Martin B. Wax, MD; Edward M. Barnett, MD; Bernard Becker, MD; Michael A. Kass, MD; Allan E. Kolker, MD; Carla J. Siegfried, MD; Regina M. Smolyak, MD. Coordinators and staff: Arnold D. Jones, COA; Diana L. Moellering, COT. Committees Executive/Steering Committee: Ingrid Adamsons, MD (nonvoting); Douglas R. Anderson, MD; James D. Brandt, MD; Anne L. Coleman, MD, PhD; Michael Drake, MD; Donald F. Everett, MA; Douglas Gaasterland, MD; Mae O. Gordon, PhD; Dale K. Heuer, MD; Eve J. Higginbotham, MD; Chris A. Johnson, PhD; Michael A. Kass, MD (chair); John L. Keltner, MD; Richard K. Parrish II, MD; Jody R. Piltz-Seymour, MD; M. Roy Wilson, MD; Jane L. Anderson, MS, CCRC; Patricia A. Morris (nonvoting); Carol J. Pollack-Rundle, COMT; Ann K. Wilder, RN, BSN (nonvoting). Data and Safety Monitoring Committee: Ingrid Adamsons, MD (nonvoting; open session); Roy Beck, MD, PhD; John Connett, PhD; Claude Cowan, MD; Barry Davis, MD, PhD (chair); Donald F. Everett, MA (nonvoting); Mae O. Gordon, PhD (nonvoting); Michael A. Kass, MD (nonvoting; open session, November 4, 1993, to April 30, 1998; closed session, May 1, 1998 to present); Ronald Munson, PhD; Mark Sherwood, MD; Gregory L. Skuta, MD. Endpoint Committee: Dale K. Heuer, MD; Eve J. Higginbotham, MD; Richard K. Parrish II, MD; Mae O. Gordon, PhD. Resource Centers: Coordinating Center: Washington University School of Medicine, St Louis, Missouri. Investigators: Mae O. Gordon, PhD; J. Philip Miller, AB. Coordinators and Staff: Joel Achtenberg, MSW; Mary Bednarski, MAS; Julia Beiser, MS; Karen Clark, BS; Christopher Ewing, BA; Elizabeth Hornbeck, BS; Ellen Long, CCRA; Carolyn R. Miles, MA; Patricia Morris; Denise Morrison; Ann K. Wilder, RN, BSN, CCRP. Chairman’s Office: Washington University School of Medicine, St Louis, Missouri. Investigator: Michael A. Kass, MD. Coordinators and staff: Deborah Dunn. Project Office: National Eye Institute, Rockville, Maryland. Donald F. Everett, MA. Optic Disc Reading Center: Bascom Palmer Eye Institute, University of Miami, Miami, Florida. Investigators: Richard K. Parrish II, MD; Douglas R. Anderson, MD; Donald L. Budenz, MD. Coordinators and staff: Maria Cristina Wells-Albornoz, MPH; William Feuer, MS; Ditte Hess, CRA; Heather Johnson; Joyce Schiffman, MS; Ruth Vandenbroucke. Visual Field Reading Centers: University of California–Davis, Sacramento; Discoveries in Sight, Devers Eye Institute, Portland, Oregon. Investigators: John L. Keltner, MD (Sacramento); Chris A. Johnson, PhD (Portland). Coordinators and Staff: Kimberly E. Cello, BS; Shannan E. Bandermann, MA; Bhupinder S. Dhillon, BS; Daniel Redline, BS. Ancillary Study Reading Centers: Confocal Scanning Laser Ophthalmoscopy Reading Center: University of California–San Diego, La Jolla. Investigators: Robert N. Weinreb, MD; Linda Zangwill, PhD. Coordinators and staff: Keri Dirkes, MPH; Amanda R. Smith, BS. Short Wavelength Automated Perimetry Reading Center: Devers Eye Institute, Legacy Portland Hospitals, Portland, Oregon. Investigator: Chris A. Johnson, PhD. Coordinator: Erna Hibbitts. Corneal Endothelial Cell Density Reading Center: Mayo Clinic/Foundation, Rochester, Minnesota. Investigator: William M. Bourne, MD. Coordinators and staff: Becky A. Nielsen, LPN; Thomas P. Link, CRA, BA; Jay A. Rostvold.

    The European Glaucoma Prevention Study members: Principal Investigators: Stefano Miglior, MD (project coordinator); Thierry Zeyen, MD, PhD (now with the University of Leuven, Leuven, Belgium); Norbert Pfeiffer, MD; Jose Cunha-Vaz, MD, PhD. Clinical Centers: Belgium. University of Antwerpen: Thierry Zeyen, MD, PhD (now with the University of Leuven, Leuven, Belgium) Marie-Claire Linsen, MD, Liliana Pellicci, MD, Andre’ Janssens, Anja van der Veken, MD, Fanny Nerinckx, MD, Vicky Boeyden, MD. University of Bruxelles: Michele Detry-Morel, MD. University of Gent: Philippe Kestelijn, MD, Kathleen Van den Abeele, MD, Kristin Jacobs, MD, Christian Decock, MD. University of Leuven: Marc Goethals, MD, Sayeh Pourjavan, MD, Karolien Maris, MD. Germany. University of Mainz: Norbert Pfeiffer, MD, Inga Kersten, MD, Annette Vogel, MD, Ulrike Herkel, MD, Oliver Schwenn, MD, Marlene Maser-Wahle. University of Freiburg: Jens Funk, MD, Beate Schmidt, MD, Nuschin Akbariyeh. University of Heidelberg: Reinhard Burk, MD, Andrea Thomsen, MD. University of Wuerzburg: Franz Grehn, MD, Dirk Marquardt, MD. Italy. University of Milan, S. Paolo Hospital: Nicola Orzalesi, MD, Stefano Miglior, MD (now with University of Milano Bicocca, Milan, Italy), Luca Rossetti, MD, Maurizio Ferrante, Luciana Mandelli, Sara Marini, Sara Bagno. University of Milan, L. Sacco Hospital: Giancarlo Bertoni, MD, Mirella Blini, MD, Massimo De Molfetta, MD. University of Verona: Luciano Bonomi, MD, Roberta Morbio, MD, Michele Marraffa, MD. University of Parma: Giovanni Maraini, MD, Stefano Gandolfi, MD, Sally Williams, Luca Cimino, MD, Patrizia Dondi, MD. Oftalmico Hospital, Rome: Bruno Lumbroso, MD, Gianluca Manni, MD, Francesco Cocco, MD, Giacomo Glorialanza, MD. S. Giovanni Hospital, Rome: Carlo Maria Villani, MD, Augusto Pocobelli, MD, Massimo Cesareo, MD, Gaetano Cupo, MD. Fatebenefratelli Hospital, Rome: Riccardo Neuschuler, MD, Claudio Pernini, MD, Roberta Catalani, M. Portugal. Coimbra, AIBILI: Jose Cunha-Vaz, MD, PhD, Luisa Ribeiro, MD, Ilidio Faria, MD, Jose Moura Pereira, MD, Salvador Chingui, MD, Lilianne Duarte, MD, Fausto Carvalheira, MD, Alda Baltar, Adozinda Simao. Viseu, S. Teotonio Hospital: Jose Arede, MD, Ana Magalhaes. Lisbon, S. Jose’ Hospital: Pedro Abrantes, MD, Maria Reina, MD, Jose Pedro Silva, MD, Graca Romano, Sofia Silva. Data Management and Biostatistical Center: Istituto Mario Negri, Milan: Valter Torri, MD, Davide Poli, Cinzia Nerviani, Angelo Tinazzi. Steering Committee: Stefano Miglior, MD, Thierry Zeyen, MD, PhD, Norbert Pfeiffer, MD, Jose Cunha-Vaz, MD, PhD, Ingrid Adamsons, MD, Joseph Caprioli, MD, Valter Torri, MD. Data and Safety Monitoring Committee: Richard Wormald, MD, Moorsfields Eye Hospital, London, Anders Hejil, MD, PhD, University of Malmo, Malmo, Sweden, Juani Airaksinen, MD, University of Oulu, Oulu, Finland, J. Michaelis, MD, PhD, University of Mainz, Germany, Irene Floriani, PhD (nonvoting), Institute of Mario Negri, Milan, Italy. Coordinating Center: University of Milan, S. Paolo Hospital: Stefano Miglior, MD, Luca Rossetti, MD, Maurizio Ferrante, Luciana Mandelli, Sara Bagno. Optic Disc Storage Center: University of Antwerpen: Thierry Zeyen, MD, PhD, Andre’ Janssens. Optic Disc Readers: University of Milan, S. Paolo Hospital: Stefano Miglior, MD. University of Antwerpen: Thierry Zeyen, MD, PhD, Marie-Claire Linsen, MD. University of Mainz: Annette Vogel, MD, Inga Kersten, MD, Oliver Schwenn, MD. Merck and Co Inc: Ingrid Adamsons, MD, MPH, Arthur Shedden, MD, Elyssa Gottfried, Jill Hutzelmann, Cynthia Rusk, Scott Reines, MD, PhD, Reynold Spector, MD, Janet Hombrey, MD, Herbert Snyder, John Gacos, Steve Snapinn, PhD, Al Getson, PhD, John Amos, Kathleen Serruys, William Malbecq, Eddie John, Justin Chapman, Jerry Beck, Carlos Tessi, MD, Naveen Rao, MD. MSD, Italy: Francesco Paolo Bottari, MD, Raffaella D’Achille, Ann Wilkins. MSD, Germany: Karin Magerl, Claudia Bauer. MSD, Belgium: Christiane Derouwaux. MSD, Portugal: Margarida Cunha, Raquel Santana, Luis Gouveia Andrade, MD, Susana Bule, Rui Melo, Helena Baumont, Isabel Bonaventura, MD.

    References
    1.
    Leske  MC, Heijl  A, Hussein  M, Bengtsson  B, Hyman  L, Komaroff  E; Early Manifest Glaucoma Trial Group.  Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial.   Arch Ophthalmol. 2003;121(1):48-56. doi:10.1001/archopht.121.1.48 PubMedGoogle Scholar
    2.
    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(6):701-713. doi:10.1001/archopht.120.6.701 PubMedGoogle Scholar
    3.
    Lichter  PR, Musch  DC, Gillespie  BW,  et al; CIGTS Study Group.  Interim clinical outcomes in the Collaborative Initial Glaucoma Treatment Study comparing initial treatment randomized to medications or surgery.   Ophthalmology. 2001;108(11):1943-1953. doi:10.1016/S0161-6420(01)00873-9 PubMedGoogle Scholar
    4.
    Van Veldhuisen  PC, Ederer  F, Gaasterland  DE,  et al; The AGIS Investigators.  The Advanced Glaucoma Intervention Study (AGIS), 7: the relationship between control of intraocular pressure and visual field deterioration.   Am J Ophthalmol. 2000;130(4):429-440. doi:10.1016/S0002-9394(00)00538-9 PubMedGoogle Scholar
    5.
    Collaborative Normal-Tension Glaucoma Study Group.  The effectiveness of intraocular pressure reduction in the treatment of normal-tension glaucoma.   Am J Ophthalmol. 1998;126(4):498-505. doi:10.1016/S0002-9394(98)00272-4 PubMedGoogle Scholar
    6.
    Nouri-Mahdavi  K, Hoffman  D, Coleman  AL,  et al; Advanced Glaucoma Intervention Study.  Predictive factors for glaucomatous visual field progression in the Advanced Glaucoma Intervention Study.   Ophthalmology. 2004;111(9):1627-1635. doi:10.1016/j.ophtha.2004.02.017 PubMedGoogle Scholar
    7.
    Caprioli  J, Coleman  AL.  Intraocular pressure fluctuation a risk factor for visual field progression at low intraocular pressures in the advanced glaucoma intervention study.   Ophthalmology. 2008;115(7):1123-1129.e3. doi:10.1016/j.ophtha.2007.10.031PubMedGoogle Scholar
    8.
    Musch  DC, Gillespie  BW, Lichter  PR, Niziol  LM, Janz  NK; CIGTS Study Investigators.  Visual field progression in the Collaborative Initial Glaucoma Treatment Study the impact of treatment and other baseline factors.   Ophthalmology. 2009;116(2):200-207. doi:10.1016/j.ophtha.2008.08.051 PubMedGoogle Scholar
    9.
    Musch  DC, Gillespie  BW, Niziol  LM, Lichter  PR, Varma  R; CIGTS Study Group.  Intraocular pressure control and long-term visual field loss in the Collaborative Initial Glaucoma Treatment Study.   Ophthalmology. 2011;118(9):1766-1773. doi:10.1016/j.ophtha.2011.01.047 PubMedGoogle Scholar
    10.
    Fujino  Y, Asaoka  R, Murata  H,  et al; Japanese Archive of Multicentral Databases in Glaucoma (JAMDIG) Construction Group.  Evaluation of glaucoma progression in large-scale clinical data: the Japanese Archive of Multicentral Databases in Glaucoma (JAMDIG).   Invest Ophthalmol Vis Sci. 2016;57(4):2012-2020. doi:10.1167/iovs.15-19046 PubMedGoogle Scholar
    11.
    Stewart  WC, Kolker  AE, Sharpe  ED,  et al.  Factors associated with long-term progression or stability in primary open-angle glaucoma.   Am J Ophthalmol. 2000;130(3):274-279. doi:10.1016/S0002-9394(00)00487-6 PubMedGoogle Scholar
    12.
    Hong  S, Seong  GJ, Hong  YJ.  Long-term intraocular pressure fluctuation and progressive visual field deterioration in patients with glaucoma and low intraocular pressures after a triple procedure.   Arch Ophthalmol. 2007;125(8):1010-1013. doi:10.1001/archopht.125.8.1010 PubMedGoogle Scholar
    13.
    Lee  PP, Walt  JW, Rosenblatt  LC, Siegartel  LR, Stern  LS; Glaucoma Care Study Group.  Association between intraocular pressure variation and glaucoma progression: data from a United States chart review.   Am J Ophthalmol. 2007;144(6):901-907. doi:10.1016/j.ajo.2007.07.040 PubMedGoogle Scholar
    14.
    Bengtsson  B, Leske  MC, Hyman  L, Heijl  A; Early Manifest Glaucoma Trial Group.  Fluctuation of intraocular pressure and glaucoma progression in the early manifest glaucoma trial.   Ophthalmology. 2007;114(2):205-209. doi:10.1016/j.ophtha.2006.07.060 PubMedGoogle Scholar
    15.
    De Moraes  CG, Juthani  VJ, Liebmann  JM,  et al.  Risk factors for visual field progression in treated glaucoma.   Arch Ophthalmol. 2011;129(5):562-568. doi:10.1001/archophthalmol.2011.72 PubMedGoogle Scholar
    16.
    Medeiros  FA, Weinreb  RN, Zangwill  LM,  et al.  Long-term intraocular pressure fluctuations and risk of conversion from ocular hypertension to glaucoma.   Ophthalmology. 2008;115(6):934-940. doi:10.1016/j.ophtha.2007.08.012 PubMedGoogle Scholar
    17.
    Gordon  MO, Torri  V, Miglior  S,  et al; Ocular Hypertension Treatment Study Group; European Glaucoma Prevention Study Group.  Validated prediction model for the development of primary open-angle glaucoma in individuals with ocular hypertension.   Ophthalmology. 2007;114(1):10-19. doi:10.1016/j.ophtha.2006.08.031 PubMedGoogle Scholar
    18.
    Gordon  MO, Kass  MA.  The Ocular Hypertension Treatment Study: design and baseline description of the participants.   Arch Ophthalmol. 1999;117(5):573-583. doi:10.1001/archopht.117.5.573 PubMedGoogle Scholar
    19.
    Miglior  S, Zeyen  T, Pfeiffer  N, Cunha-Vaz  J, Torri  V, Adamsons  I; European Glaucoma Prevention Study Group.  The European glaucoma prevention study design and baseline description of the participants.   Ophthalmology. 2002;109(9):1612-1621. doi:10.1016/S0161-6420(02)01167-3 PubMedGoogle Scholar
    20.
    Ederer  F, Gaasterland  DE, Sullivan  EK; AGIS Investigators.  The Advanced Glaucoma Intervention Study (AGIS), 1: study design and methods and baseline characteristics of study patients.   Control Clin Trials. 1994;15(4):299-325. doi:10.1016/0197-2456(94)90046-9 PubMedGoogle Scholar
    21.
    Gordon  MO, Higginbotham  EJ, Heuer  DK,  et al; Ocular Hypertension Treatment Study.  Assessment of the impact of an endpoint committee in the Ocular Hypertension Treatment Study.   Am J Ophthalmol. 2019;199:193-199. doi:10.1016/j.ajo.2018.11.006 PubMedGoogle Scholar
    22.
    van Houwelingen  HC.  Dynamic prediction by landmarking in event history analysis.   Scand J Stat. 2007;34(1):70-85. doi:10.1111/j.1467-9469.2006.00529.x Google Scholar
    23.
    Fisher  LD, Lin  DY.  Time-dependent covariates in the Cox proportional-hazards regression model.   Annu Rev Public Health. 1999;20:145-157. doi:10.1146/annurev.publhealth.20.1.145 PubMedGoogle Scholar
    24.
    Matlach  J, Bender  S, König  J, Binder  H, Pfeiffer  N, Hoffmann  EM.  Investigation of intraocular pressure fluctuation as a risk factor of glaucoma progression.   Clin Ophthalmol. 2018;13:9-16. doi:10.2147/OPTH.S186526PubMedGoogle Scholar
    25.
    Hair  JF  Jr, Black  WC, Babin  BJ, Anderson  RE.  Multivariate Data Analysis. 7th ed. Pearson Education Ltd; 2010.
    26.
    Gordon  MO, Beiser  JA, Brandt  JD,  et al.  The Ocular Hypertension Treatment Study: baseline factors that predict the onset of primary open-angle glaucoma.   Arch Ophthalmol. 2002;120(6):714-720. doi:10.1001/archopht.120.6.714 PubMedGoogle Scholar
    27.
    Miglior  S, Pfeiffer  N, Torri  V, Zeyen  T, Cunha-Vaz  J, Adamsons  I; European Glaucoma Prevention Study (EGPS) Group.  Predictive factors for open-angle glaucoma among patients with ocular hypertension in the European Glaucoma Prevention Study.   Ophthalmology. 2007;114(1):3-9. doi:10.1016/j.ophtha.2006.05.075 PubMedGoogle Scholar
    28.
    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(10):1351-1360. doi:10.1001/archopht.123.10.1351 PubMedGoogle Scholar
    29.
    Miglior  S, Torri  V, Zeyen  T, Pfeiffer  N, Vaz  JC, Adamsons  I; EGPS Group.  Intercurrent factors associated with the development of open-angle glaucoma in the European glaucoma prevention study.   Am J Ophthalmol. 2007;144(2):266-275. doi:10.1016/j.ajo.2007.04.040PubMedGoogle Scholar
    30.
    Gao  F, Miller  JP, Xiong  C, Beiser  JA, Gordon  M; The Ocular Hypertension Treatment Study (OHTS) Group.  A joint-modeling approach to assess the impact of biomarker variability on the risk of developing clinical outcome.   Stat Methods Appt. 2011;20(1):83-100. doi:10.1007/s10260-010-0150-zPubMedGoogle Scholar
    31.
    van Houwelingen  HC, Putter  H.  Dynamic Prediction in Clinical Survival Analysis. CRC Press Inc; 2011. doi:10.1201/b11311
    32.
    Jiang  X, Torres  M, Varma  R; Los Angeles Latino Eye Study Group.  Variation in intraocular pressure and the risk of developing open-angle glaucoma: the Los Angeles Latino Eye Study.   Am J Ophthalmol. 2018;188:51-59. doi:10.1016/j.ajo.2018.01.013 PubMedGoogle Scholar
    33.
    Kim  JH, Caprioli  J.  Intraocular pressure fluctuation: is it important?   J Ophthalmic Vis Res. 2018;13(2):170-174. doi:10.4103/jovr.jovr_35_18 PubMedGoogle Scholar
    ×