- Email: [email protected]
A Control Experiment for Studies that Show Improved Visual Sensitivity with Intraocular Pressure Lowering in Glaucoma
A Control Experiment for Studies that Show Improved Visual Sensitivity with Intraocular Pressure Lowering in Glaucoma Andrew J. Anderson, MSc(Optom), PhD, Matthew J. Stainer, PhD Purpose: Contrast sensitivity sometimes increases in patients with open-angle glaucoma when intraocular pressure (IOP) is decreased. Although often interpreted as demonstrating reversible glaucoma-induced dysfunction, this result, if true, could simply reflect a general relationship between sensitivity and IOP in visual mechanisms unaffected by glaucoma. To investigate this relationship, we test the hypothesis that reducing IOP in eyes without glaucoma (ocular hypertension) does not increase perimetric contrast sensitivity. Design: Comparative case series. Participants: A total of 692 participants drawn from the Ocular Hypertension Treatment Study (OHTS) (22 clinical centers). Methods: Commercially available topical ocular hypotensive medications. Main Outcome Measures: Post hoc analysis of IOP and perimetric contrast sensitivity (mean deviation [MD] and pattern standard deviation [PSD]) both at baseline (0 months, immediately before ocular antihypertensive therapy) and at 6-month review. An additional 618 control eyes from OHTS that did not receive treatment were examined over the same period. Data from the second phase of OHTS also were examined, and control eyes then received treatment. Results: Treated eyes had a decrease in IOP at 6 months (5.1 mmHg, P<0.001) but no significant change in MD (0.04 decibels [dB], P ¼ 0.59) or PSD (0.03 dB, P ¼ 0.19), relative to controls. A similar decrease in IOP was found for eyes that began treatment in the second phase of OHTS, but no significant change in MD or PSD. Conclusions: Despite using a large sample size, we found no relationship between perimetric contrast sensitivity and IOP reduction in ocular hypertension, which suggests that previous sensitivity changes seen in patients with glaucoma, if true, are indicative of reversible glaucoma-induced dysfunction rather than a general relationship between sensitivity and IOP in visual mechanisms unaffected by glaucoma. Ophthalmology 2014;:1e5 ª 2014 by the American Academy of Ophthalmology.
Although it is well established that retinal ganglion cells die in glaucoma, what is less clear is whether glaucoma induces a potentially reversible ganglion cell dysfunction before this death. One piece of evidence in favor of reversible dysfunction comes from studies that report an increase in contrast sensitivity in patients with open-angle glaucoma when intraocular pressure (IOP) is decreased.1e5 Contrast sensitivity in such studies is commonly assessed by means of grating targets1,2,4,5 or standard automated perimetry.2,3 When contrast sensitivity improvements are found with standard automated perimetry, typically only average sensitivity changes (mean deviation [MD]) rather than the degree of patterned loss (pattern standard deviation [PSD]),2,3 suggesting that the relative depth of glaucomatous scotomata may not alter.3 Therefore, improved sensitivity could reflect a general relationship between sensitivity and IOP in normal visual mechanisms, rather than reversible glaucoma-induced dysfunction per se. That flicker sensitivity was found to correlate with IOP for normal eyes measured cross-sectionally suggests such a relationship could exist.6 However, whether altering IOP in a given eye would change visual sensitivity by the amount predicted by this population-derived correlation is unknown. If present, 2014 by the American Academy of Ophthalmology Published by Elsevier Inc.
such changes are likely to be relatively small, which may also in part explain why an effect of lowering IOP on contrast sensitivity is not universally found.7,8 Determining whether altering IOP modifies contrast sensitivity in eyes without glaucoma is therefore an important control experiment, without which previous studies of an increase in contrast sensitivity in treated glaucoma are difficult to interpret unambiguously. If lowering IOP in observers without glaucoma results in an increase in average contrast sensitivity like that reported in treated patients with glaucoma, this would argue against such shifts being the result of reversible glaucoma-induced ganglion cell dysfunction. Conversely, if such IOP lowering has no effect in observers without glaucoma, this would significantly strengthen the argument that previous studies1e4 have demonstrated glaucoma-induced dysfunction rather than simply a relationship between IOP and contrast sensitivity. A study lowering IOP in normal eyes would have 2 problems: First, IOP-lowering medications are not without side effects, and so it may be difficult to justify treating a large enough group of normal observers for a study of sufficient statistical power. Second, such observers would likely have ISSN 0161-6420/14/$ - see front matter http://dx.doi.org/10.1016/j.ophtha.2014.04.014
1
Ophthalmology Volume -, Number -, Month 2014 pre-treatment IOPs significantly lower than those of patients with glaucoma, and so a failure to find an effect of lowering IOP might be because IOP only causes a reduction in contrast sensitivity when it is above typical normal limits (e.g., >21 mmHg). A better approach might be to examine the influence of IOP-lowering in ocular hypertensive patients. Such patients have pre-treatment IOPs more in keeping with glaucoma, yet lack any observable glaucomainduced changes on standard diagnostic measures (e.g., visual field assessment, optic nerve head evaluation). It may be thought that because visual field results are statistically normal, there exists no opportunity for visual field sensitivity to further improve. The statistical limits defining normal sensitivity are large,9,10 relative to the changes in MD seen in glaucoma with lowered IOP (0.84 decibels [dB], medical lowering,2 and 1.7 dB, surgical lowering3), providing substantial opportunity for improved visual sensitivity even in statistically normal visual fields. There have been investigations of the influence of IOP lowering on contrast sensitivity in ocular hypertensive patients. Flammer and Drance11 found no change in mean sensitivity on automated perimetry in ocular hypertensive patients (n ¼ 16) treated with acetazolamide despite finding a þ2.4 dB improvement in eyes with glaucoma (n ¼ 9) and having a similar IOP reduction (w8e9 mmHg), although they questioned whether this may have been due to a pharmacologic effect of acetazolamide rather than via its effect on IOP. Tytla et al12 found that reducing IOP had variable effects, with approximately equal proportions of their ocular hypertensive groups showing unchanged, normal flicker sensitivity, others showing normal sensitivity post-IOP treatment, and some continuing to show abnormal sensitivity despite treatment. In a small (n ¼ 15) sample by Farinelli et al,13 acute reduction of IOP by an average of 8.4 mmHg using glycerol produced no effect on visual field indices on automated perimetry, but decreased contrast sensitivity at mid-range spatial frequencies. Flammer and Drance14 have also argued that changes in perimetric sensitivity are dependent on medication, not on IOP reduction although they provide no statistical analysis for the relatively small differences they present. Overall, a clear consensus does not emerge from previous investigations, with studies often being of limited sample size and without appropriate control groups. In this study, we perform a post hoc analysis on data from the Ocular Hypertension Treatment Study (OHTS)15 to see whether reducing IOP improves contrast sensitivity in the absence of glaucoma, thereby not supporting the hypothesis of reversible glaucoma-induced ganglion cell dysfunction. The OHTS investigated the influence of IOPlowering treatment on the rate that ocular hypertensive patients developed glaucoma and found that the majority (>90%) of patients failed to develop the disease over a 6year follow-up period even when their ocular hypertension was not treated.16 Because the cohort investigated in OHTS is large (>1600 participants), our study can have a high statistical power, and so a failure to find an effect on contrast sensitivity from lowering IOP can be taken as strong evidence that an effect is genuinely absent. This large sample contrasts with the low numbers typically
2
investigated in previous studies examining the effect of IOP reduction on contrast sensitivity.1e4,7,8
Methods Subjects For our principal analysis, we evaluated measurements from the OHTS dataset, the characteristics of which have been reported.15 Briefly, OHTS enrolled more than 1600 patients with ocular hypertension (repeatedly normal optic nerve and visual fields, and IOP 24 and <32 mmHg), with half of them randomized to treatment with topical ocular hypotensive medications and the other half followed for observation only. Baseline data for each group were similar.17 For our treatment group, we compared 2 visits for those randomized to treatment: the visit immediately before medication commencing and the 6-month follow-up visit after this. Most treated patients received medication in both eyes except with the onset of treatment delayed for 6 months in the second eye. For our study, we selected only those treated patients who had a single eye treated for the first 6 months (n ¼ 692 [390 right eyes, 248 left eyes]) and used this eye for all analyses. For our control group for this portion of our study, a single eye from patients in the observation arm of OHTS was selected at random for analysis (n ¼ 618 [309 right eyes, 309 left eyes]) at similar times (initial visit and 6-month follow-up). For both our treatment and control groups, only those who did not convert to glaucoma within the first phase of OHTS (i.e., approximately a further 5 years16 beyond the data we analyze) were included to minimize the possibility that our cohort included those with nascent glaucoma. Participants in OHTS provided informed written consent approved by each participating center’s institutional review board.16 The IOP reduction in OHTS was approximately 6 mmHg,16 which is higher than the 2.3 mmHg reduction previously shown to significantly improve contrast sensitivity across a range of spatial frequencies in persons with glaucoma (n ¼ 16).5 In the second phase of OHTS (OHTS II), a large number of participants initially randomized to observation in OHTS were placed on medication, whereas the originally medicated group was continued on the same regimen.18 This gave us a further opportunity to assess the influence of lowering IOP in ocular hypertensive patients. When analyzing OHTS II data, for our treatment group we analyzed the last scheduled visit when medication had not been used and then 6 months later (typically the 6- and 12-month visits in OHTS II, respectively). We randomly selected an eye for those patients binocularly treated. For our control group in OHTS II, we analyzed similar visits (i.e., 6and 12-month follow-up visits) for those patients already on medication from OHTS. For both our treatment and control groups, we analyzed only those patients who never achieved a diagnosis of glaucoma over the duration of the study (approximately a further 5 years18 beyond the data we analyze), being 515 potential patients in the treatment group (i.e., medication given for the first time at the beginning of OHTS II) and 581 potential patients in the control group. Of the potential treatment group, we analyzed 269 patients (156 right and 113 left eyes) after the following sequential deletions: never actually medicated (114), already medicated at first visit (31), missing 1 measures (IOP, PSD, or MD) at the visits to be analyzed (67), and no 12-month followup visit (34). Of the potential control group, we analyzed 492 patients (246 right and 246 left eyes, randomly assigned) after the following sequential deletions: <3 follow-up visits (7), missing 1 measures (IOP, PSD, or MD) at the visits to be analyzed (25), and no 6-month follow-up visit (0).
Anderson and Stainer
Contrast sensitivity and IOP lowering
Each of our principal analyses outlined (1 for OHTS, and 1 for OHTS II) contains 2 groups: a treatment group in which medication was commenced and its effect determined at 6 months followup and a control group in which management remained constant over a 6-month period.
Analysis We used separate 1-way analyses of variance for each dependent variable of change (6-month minus baseline) in IOP, MD, and PSD, with an independent variable of group (treatment and control). Should lowering IOP result in an improvement in average visual field sensitivity, we would predict a significant alteration in MD but not in PSD.
Results Figure 1 shows the difference between baseline (0 month) and the 6-month visit for IOP, MD, and PSD for both our treatment and control (no medication) groups drawn from OHTS. Consistent with the aim of OHTS, patients in our treatment group had a significant reduction in IOP compared with our control group (5.1 mmHg; 95% confidence interval [CI], 5.42 to 4.68; F (1,1254) ¼ 709.8; P<0.001). Despite this IOP reduction, there was no significant alteration in MD (0.04 dB; 95% CI, 0.16 to 0.09; F (1,1254) ¼ 0.29; P ¼ 0.59) or PSD (0.03 dB; 95% CI, 0.08e0.02; F (1254) ¼ 1.69; P ¼ 0.19) compared with the control group. Figure 2 shows the difference between baseline (0 month) and the 6-month visit for IOP, MD, and PSD for both our treatment and control (on medication at both visits) groups drawn from OHTS II. A significant reduction in IOP was seen in our treatment group compared with our control group (5.1 mmHg; 95% CI, 4.52 to 5.67; F (1,797) ¼ 368; P<0.001). Despite this IOP reduction, there was no significant alteration in MD (0.05 dB; 95% CI, 0.24 to 0.15; F (1,797) ¼ 0.23; P ¼ 0.62) or PSD (0.06 dB; 95% CI, 0.07 to 0.19; F (1,797) ¼ 0.96; P ¼ 0.33) compared with our control group.
Discussion Our results show that a reduction in IOP in ocular hypertensive patients did not cause an accompanying change in the average sensitivity of the visual field measured using standard automated perimetry. This argues against a relationship between change in IOP and contrast sensitivity in eyes without glaucoma, and so strengthens previous interpretations that a change in contrast sensitivity in glaucomatous eyes when IOP is lowered is the result of glaucoma-induced, reversible visual dysfunction.1e4 However, note that our results do not provide any further support for whether sensitivity increases when IOP is reduced in glaucomatous eyes per se. The absence of adequate control groups in some previous studies2,4 substantially weakens their conclusions that sensitivity increases in glaucomatous eyes when IOP is lowered, although other studies have incorporated control groups5 or eyes1,3 to minimize such effects as perceptual learning19 or regression-to-the-mean. Because of the large size of our sample, coupled with the tightly controlled measuring of IOP and visual fields in OHTS,15 our negative finding can be taken with confidence. Although our reductions in IOP are smaller than those involving surgical interventions,1,3 they are larger5 or similar to4 those from
previous studies showing contrast sensitivity improvement on grating-based measures of contrast sensitivity with medical treatment of glaucoma. Therefore, our negative findings are unlikely to be explained by assuming that our IOP reduction is below a threshold value for which sensitivity change occurs. Our findings are consistent with earlier research by Flammer and Drance,11 but address several concerns within their study. That these authors showed no change in perimetric contrast sensitivity in ocular hypertension but did find changes in glaucomatous observers despite a smaller sample size argues against the idea that their study was insufficiently powered to find changes in their ocular hypertensive group. However, it is possible that the sensitivity change seen in their glaucoma group was not due to IOP reduction. First, Flammer and Drance11 questioned whether their findings of improved sensitivity in treated glaucoma may have been the result of a pharmacologic effect of acetazolamide, rather than a change in IOP. Second, these authors found most change in sensitivity in defective areas of the visual field, but suggested this effect was greater than that explicable through regressionto-the-mean effects noted in previous work.14 It is likely that this previous assessment of regression-to-the-mean was an underestimate, however, because it examined those with “very early stages” of glaucoma14 in contrast to the average Octopus perimeter sensitivity of 17 dB for the glaucoma group reported in their later study.11 It is now well established that perimetric variability increases as field sensitivity decreases,20,21 suggesting that regression-to-the mean effects assessed in early glaucoma should underestimate effects seen in more advanced disease. In addition, testeretest limits for standard achromatic perimetry are asymmetric in areas of substantial loss22 and would favor larger sensitivity improvements more than sensitivity declines on retest. The OHTS assessed participants at 6-month intervals, providing a limit to the temporal resolution with which the influence of IOP could be assessed in our study. It may be thought that this period is sufficient for our results to be potentially influenced by disease progression, and that this could have counterbalanced any improvement in sensitivity with IOP lowering. We do not believe this is likely. First, we only examined those participants who did not progress to a diagnosis of glaucoma over the 6-year period of each OHTS phase (OHTS I and OHTS II: see “Methods”), and so the likelihood of any significant visual field progression in the first 6 months of this period is remote. Furthermore, in the unlikely event that progression occurred, we would expect a significant change in PSD as subtle scotomas develop, even if MD remained constant because the average depth of this progression happened to be exactly matched by a sensitivity improvement caused by a lowering of IOP. We find no such change in PSD (Fig 1), giving assurance that our control and treatment groups are well matched in terms of any progression, however unlikely. Although we show that visual field sensitivity did not alter, this is not to say that other visual functions may not be perturbed by changing IOP in eyes without glaucoma. For example, IOP lowering can alter pattern electroretinogram responses in ocular hypertensive observers (Ventura et al23 have a summary of recent findings). There is evidence that the influence of a transient elevation in IOP is more profound in ocular hypertensive patients than controls, despite pressure being
3
Ophthalmology Volume -, Number -, Month 2014
Figure 1. Data from the Ocular Hypertension Treatment Study: difference in intraocular pressure (IOP), mean deviation (MD), and pattern standard deviation (PSD) between baseline (0 months) and 6-month follow-up for the control group (no medication) and treatment group placed on IOP-lowering medication immediately after baseline. Error bars give 1 standard error of the mean. Initial IOP, MD, and PSD values for the treatment and control groups were 25.7 mmHg, 0.24 decibels (dB), and 1.95 dB, and 24.8 mmHg, 0.15 dB, and 1.94 dB, respectively. Negative values and positive values mean worsening values at follow-up for MD and PSD, respectively. Tx ¼ treatment group.
elevated to identical levels in both groups,24 suggesting that ocular hypertensive patients may have an increased susceptibility to IOP change in some circumstances. However, the presence of such alterations in ocular hypertension cannot necessarily be interpreted as subclinical manifestations of glaucoma or that early glaucomatous changes are potentially reversible. For example, Tytla et al12 observed that the prevalence of flicker sensitivity abnormalities in previous studies of hypertensive eyes was high, and greater than expected given the natural history of the disease, a concern shared by other authors.25 Therefore,
subtle sensitivity defects in ocular hypertension may represent stable changes associated with ocular hypertension that are not necessarily related to the progressive changes that may occur subsequently in glaucoma. This is true irrespective of whether these subsequent glaucomatous changes are reversible or permanent. In conclusion, we show that reduction of elevated IOP in eyes without glaucoma does not produce a significant change in contrast sensitivity using automated perimetry. Our results provide a necessary control required to strengthen the argument that previous studies showing visual field improvement
Figure 2. Data from Ocular Hypertension Treatment Study Phase II: difference in intraocular pressure (IOP), mean deviation (MD), and pattern standard deviation (PSD) between baseline (0 months) and 6-month follow-up for the control group (always on medication) and treatment group placed on IOPlowering medication immediately after baseline. Error bars give 1 standard error of the mean. Initial IOP, MD, and PSD values for the treatment and control groups were 23.6 mmHg, 0.33 decibels (dB), and 2.11 dB, and 18.2 mmHg, 0.17 dB, and 2.06 dB, respectively. Negative values and positive values mean worsening values at follow-up for MD and PSD, respectively. Tx ¼ treatment group.
4
Anderson and Stainer
Contrast sensitivity and IOP lowering
with IOP reduction in glaucoma1e4 have demonstrated glaucoma-induced dysfunction rather than simply a general relationship between IOP and contrast sensitivity. Acknowledgments. The authors thank Mae Gordon and Michael Kass for feedback on the manuscript.
References 1. Gandolfi SA, Cimino L, Sangermani C, et al. Improvement of spatial contrast sensitivity threshold after surgical reduction of intraocular pressure in unilateral high-tension glaucoma. Invest Ophthalmol Vis Sci 2005;46:197–201. 2. Prata TS, Piassi MV, Melo LA Jr. Changes in visual function after intraocular pressure reduction using antiglaucoma medications. Eye (Lond) 2009;23:1081–5. 3. Gandolfi SA. Improvement of visual field indices after surgical reduction of intraocular pressure. Ophthalmic Surg 1995;26: 121–6. 4. Evans DW, Hosking SL, Gherghel D, Bartlett JD. Contrast sensitivity improves after brimonidine therapy in primary open angle glaucoma: a case for neuroprotection. Br J Ophthalmol 2003;87:1463–5. 5. Pomerance GN, Evans DW. Test-retest reliability of the CSV1000 contrast test and its relationship to glaucoma therapy. Invest Ophthalmol Vis Sci 1994;45:3357–61. 6. Tyler CW, Ryu S, Stamper R. The relation between visual sensitivity and intraocular pressure in normal eyes. Invest Ophthalmol Vis Sci 1984;25:103–5. 7. Wittström E, Schatz P, Lövestam-Adrian M, et al. Improved retinal function after trabeculectomy in glaucoma patients. Graefes Arch Clin Exp Ophthalmol 2010;248:485–95. 8. Aydin A, Wollstein G, Price LL, et al. Optical coherence tomography assessment of retinal nerve fiber layer thickness changes after glaucoma surgery. Ophthalmology 2003;110:1506–11. 9. Heijl A, Lindgren G, Olsson J. Normal variability of static perimetric threshold values across the central visual field. Arch Ophthalmol 1987;105:1544–9. 10. Heijl A, Lindgren G, Olsson J. A package for the statistical analysis of visual fields. Doc Ophthalmol Proc Ser 1987;49:153–68. 11. Flammer J, Drance SM. Effect of acetazolamide on the differential threshold. Arch Ophthalmol 1983;101:1378–80. 12. Tytla ME, Trope GE, Bunic JR. Flicker sensitivity in treated ocular hypertension. Ophthalmology 1989;97:36–43. 13. Farinelli AC, Mikelberg FS, Drance SM, et al. Effect of acute medical reduction of intraocular pressure on the visual field and optic disc in ocular hypertension. Can J Ophthalmol 1988;23:216–8.
14. Flammer J, Drance SM. The effect of a number of glaucoma medications on the differential light threshold. Doc Ophthalmol Proc Ser 1983;35:145–8. 15. Gordon MO, Kass MA; Ocular Hypertension Treatment Study Group. The Ocular Hypertension Treatment Study: design and baseline description of the participants. Arch Ophthalmol 1999;117:573–83. 16. Kass MA, Heuer DK, Higginbotham EJ, et al; Ocular Hypertension Treatment Study Group. 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: 701–13. 17. Gordon MO, Beiser JA, Brandt JD, et al; Ocular Hypertension Treatment Study Group. The Ocular Hypertension Treatment Study: baseline factors that predict the onset of primary openangle glaucoma. Arch Ophthalmol 2002;120:714–20. 18. Kass MA, Gordon MO, Gao F, et al; Ocular Hypertension Treatment Study Group. Delaying treatment of ocular hypertension: the Ocular Hypertension Treatment Study. Arch Ophthalmol 2010;128:276–87. 19. McKendrick AM, Battista J. Perceptual learning of contour integration is not compromised in the elderly. J Vis [serial online] 2013;13:5. Available at: http://www.journalofvision. org/content/13/1/5.long. Accessed April 16, 2014. 20. Chauhan BC, Tompkins JD, LeBlanc RP, McCormick TA. Characteristics of frequency-of-seeing curves in normal subjects, patients with suspected glaucoma, and patients with glaucoma. Invest Ophthalmol Vis Sci 1993;34:3534–40. 21. Spry PG, Johnson CA, McKendrick AM, Turpin A. Variability components of standard automated perimetry and frequency-doubling technology perimetry. Invest Ophthalmol Vis Sci 2001;42:1404–10. 22. Artes PH, Hutchison DM, Nicolela MT, et al. Threshold and variability properties of matrix frequency-doubling technology and standard automated perimetry in glaucoma. Invest Ophthalmol Vis Sci 2005;46:2451–7. 23. Ventura LM, Golubev I, Lee W, et al. Head-down posture induces PERG alterations in early glaucoma. J Glaucoma 2013;22:255–64. 24. Colotto A, Falsini B, Salgarello T, et al. Transiently raised intraocular pressure reveals pattern electroretinogram losses in ocular hypertension. Invest Ophthalmol Vis Sci 1996;37: 2663–70. 25. Ruben ST, Arden GB, O’Sullivan F, Hitchings RA. Pattern electroretinogram and peripheral colour contrast thresholds in ocular hypertension and glaucoma: comparison and correlation of results. Br J Ophthalmol 1995;79:326–31.
Footnotes and Financial Disclosures Originally received: February 12, 2014. Final revision: March 25, 2014. Accepted: April 17, 2014. Available online: ---.
Manuscript no. 2014-230.
Horncrest Foundation, awards to the Department of Ophthalmology and Visual Sciences at Washington University, the National Institutes of Health Vision Core Grant P30 EY 02687, Merck Research Laboratories, Pfizer, Inc. (Whitehouse Station, NJ), and unrestricted grants from Research to Prevent Blindness, Inc (New York, NY).
Department of Optometry and Vision Sciences, The University of Melbourne, Parkville, Australia. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.
Abbreviations and Acronyms: CI ¼ confidence interval; dB ¼ decibels; IOP ¼ intraocular pressure; MD ¼ mean deviation; OHTS ¼ Ocular Hypertension Treatment Study; PSD ¼ pattern standard deviation.
Supported by Australian Research Council Future Fellowship FT120100407 (A.J.A.). The original OHTS was supported by awards from the National Eye Institute, National Center on Minority Health and Health Disparities, National Institutes of Health (Grants EY09341, EY09307),
Correspondence: Andrew J. Anderson, MSc(Optom), PhD, Department of Optometry and Vision Sciences, The University of Melbourne 3010, Australia. E-mail: [email protected].
5
Although it is well established that retinal ganglion cells die in glaucoma, what is less clear is whether glaucoma induces a potentially reversible ganglion cell dysfunction before this death. One piece of evidence in favor of reversible dysfunction comes from studies that report an increase in contrast sensitivity in patients with open-angle glaucoma when intraocular pressure (IOP) is decreased.1e5 Contrast sensitivity in such studies is commonly assessed by means of grating targets1,2,4,5 or standard automated perimetry.2,3 When contrast sensitivity improvements are found with standard automated perimetry, typically only average sensitivity changes (mean deviation [MD]) rather than the degree of patterned loss (pattern standard deviation [PSD]),2,3 suggesting that the relative depth of glaucomatous scotomata may not alter.3 Therefore, improved sensitivity could reflect a general relationship between sensitivity and IOP in normal visual mechanisms, rather than reversible glaucoma-induced dysfunction per se. That flicker sensitivity was found to correlate with IOP for normal eyes measured cross-sectionally suggests such a relationship could exist.6 However, whether altering IOP in a given eye would change visual sensitivity by the amount predicted by this population-derived correlation is unknown. If present, 2014 by the American Academy of Ophthalmology Published by Elsevier Inc.
such changes are likely to be relatively small, which may also in part explain why an effect of lowering IOP on contrast sensitivity is not universally found.7,8 Determining whether altering IOP modifies contrast sensitivity in eyes without glaucoma is therefore an important control experiment, without which previous studies of an increase in contrast sensitivity in treated glaucoma are difficult to interpret unambiguously. If lowering IOP in observers without glaucoma results in an increase in average contrast sensitivity like that reported in treated patients with glaucoma, this would argue against such shifts being the result of reversible glaucoma-induced ganglion cell dysfunction. Conversely, if such IOP lowering has no effect in observers without glaucoma, this would significantly strengthen the argument that previous studies1e4 have demonstrated glaucoma-induced dysfunction rather than simply a relationship between IOP and contrast sensitivity. A study lowering IOP in normal eyes would have 2 problems: First, IOP-lowering medications are not without side effects, and so it may be difficult to justify treating a large enough group of normal observers for a study of sufficient statistical power. Second, such observers would likely have ISSN 0161-6420/14/$ - see front matter http://dx.doi.org/10.1016/j.ophtha.2014.04.014
1
Ophthalmology Volume -, Number -, Month 2014 pre-treatment IOPs significantly lower than those of patients with glaucoma, and so a failure to find an effect of lowering IOP might be because IOP only causes a reduction in contrast sensitivity when it is above typical normal limits (e.g., >21 mmHg). A better approach might be to examine the influence of IOP-lowering in ocular hypertensive patients. Such patients have pre-treatment IOPs more in keeping with glaucoma, yet lack any observable glaucomainduced changes on standard diagnostic measures (e.g., visual field assessment, optic nerve head evaluation). It may be thought that because visual field results are statistically normal, there exists no opportunity for visual field sensitivity to further improve. The statistical limits defining normal sensitivity are large,9,10 relative to the changes in MD seen in glaucoma with lowered IOP (0.84 decibels [dB], medical lowering,2 and 1.7 dB, surgical lowering3), providing substantial opportunity for improved visual sensitivity even in statistically normal visual fields. There have been investigations of the influence of IOP lowering on contrast sensitivity in ocular hypertensive patients. Flammer and Drance11 found no change in mean sensitivity on automated perimetry in ocular hypertensive patients (n ¼ 16) treated with acetazolamide despite finding a þ2.4 dB improvement in eyes with glaucoma (n ¼ 9) and having a similar IOP reduction (w8e9 mmHg), although they questioned whether this may have been due to a pharmacologic effect of acetazolamide rather than via its effect on IOP. Tytla et al12 found that reducing IOP had variable effects, with approximately equal proportions of their ocular hypertensive groups showing unchanged, normal flicker sensitivity, others showing normal sensitivity post-IOP treatment, and some continuing to show abnormal sensitivity despite treatment. In a small (n ¼ 15) sample by Farinelli et al,13 acute reduction of IOP by an average of 8.4 mmHg using glycerol produced no effect on visual field indices on automated perimetry, but decreased contrast sensitivity at mid-range spatial frequencies. Flammer and Drance14 have also argued that changes in perimetric sensitivity are dependent on medication, not on IOP reduction although they provide no statistical analysis for the relatively small differences they present. Overall, a clear consensus does not emerge from previous investigations, with studies often being of limited sample size and without appropriate control groups. In this study, we perform a post hoc analysis on data from the Ocular Hypertension Treatment Study (OHTS)15 to see whether reducing IOP improves contrast sensitivity in the absence of glaucoma, thereby not supporting the hypothesis of reversible glaucoma-induced ganglion cell dysfunction. The OHTS investigated the influence of IOPlowering treatment on the rate that ocular hypertensive patients developed glaucoma and found that the majority (>90%) of patients failed to develop the disease over a 6year follow-up period even when their ocular hypertension was not treated.16 Because the cohort investigated in OHTS is large (>1600 participants), our study can have a high statistical power, and so a failure to find an effect on contrast sensitivity from lowering IOP can be taken as strong evidence that an effect is genuinely absent. This large sample contrasts with the low numbers typically
2
investigated in previous studies examining the effect of IOP reduction on contrast sensitivity.1e4,7,8
Methods Subjects For our principal analysis, we evaluated measurements from the OHTS dataset, the characteristics of which have been reported.15 Briefly, OHTS enrolled more than 1600 patients with ocular hypertension (repeatedly normal optic nerve and visual fields, and IOP 24 and <32 mmHg), with half of them randomized to treatment with topical ocular hypotensive medications and the other half followed for observation only. Baseline data for each group were similar.17 For our treatment group, we compared 2 visits for those randomized to treatment: the visit immediately before medication commencing and the 6-month follow-up visit after this. Most treated patients received medication in both eyes except with the onset of treatment delayed for 6 months in the second eye. For our study, we selected only those treated patients who had a single eye treated for the first 6 months (n ¼ 692 [390 right eyes, 248 left eyes]) and used this eye for all analyses. For our control group for this portion of our study, a single eye from patients in the observation arm of OHTS was selected at random for analysis (n ¼ 618 [309 right eyes, 309 left eyes]) at similar times (initial visit and 6-month follow-up). For both our treatment and control groups, only those who did not convert to glaucoma within the first phase of OHTS (i.e., approximately a further 5 years16 beyond the data we analyze) were included to minimize the possibility that our cohort included those with nascent glaucoma. Participants in OHTS provided informed written consent approved by each participating center’s institutional review board.16 The IOP reduction in OHTS was approximately 6 mmHg,16 which is higher than the 2.3 mmHg reduction previously shown to significantly improve contrast sensitivity across a range of spatial frequencies in persons with glaucoma (n ¼ 16).5 In the second phase of OHTS (OHTS II), a large number of participants initially randomized to observation in OHTS were placed on medication, whereas the originally medicated group was continued on the same regimen.18 This gave us a further opportunity to assess the influence of lowering IOP in ocular hypertensive patients. When analyzing OHTS II data, for our treatment group we analyzed the last scheduled visit when medication had not been used and then 6 months later (typically the 6- and 12-month visits in OHTS II, respectively). We randomly selected an eye for those patients binocularly treated. For our control group in OHTS II, we analyzed similar visits (i.e., 6and 12-month follow-up visits) for those patients already on medication from OHTS. For both our treatment and control groups, we analyzed only those patients who never achieved a diagnosis of glaucoma over the duration of the study (approximately a further 5 years18 beyond the data we analyze), being 515 potential patients in the treatment group (i.e., medication given for the first time at the beginning of OHTS II) and 581 potential patients in the control group. Of the potential treatment group, we analyzed 269 patients (156 right and 113 left eyes) after the following sequential deletions: never actually medicated (114), already medicated at first visit (31), missing 1 measures (IOP, PSD, or MD) at the visits to be analyzed (67), and no 12-month followup visit (34). Of the potential control group, we analyzed 492 patients (246 right and 246 left eyes, randomly assigned) after the following sequential deletions: <3 follow-up visits (7), missing 1 measures (IOP, PSD, or MD) at the visits to be analyzed (25), and no 6-month follow-up visit (0).
Anderson and Stainer
Contrast sensitivity and IOP lowering
Each of our principal analyses outlined (1 for OHTS, and 1 for OHTS II) contains 2 groups: a treatment group in which medication was commenced and its effect determined at 6 months followup and a control group in which management remained constant over a 6-month period.
Analysis We used separate 1-way analyses of variance for each dependent variable of change (6-month minus baseline) in IOP, MD, and PSD, with an independent variable of group (treatment and control). Should lowering IOP result in an improvement in average visual field sensitivity, we would predict a significant alteration in MD but not in PSD.
Results Figure 1 shows the difference between baseline (0 month) and the 6-month visit for IOP, MD, and PSD for both our treatment and control (no medication) groups drawn from OHTS. Consistent with the aim of OHTS, patients in our treatment group had a significant reduction in IOP compared with our control group (5.1 mmHg; 95% confidence interval [CI], 5.42 to 4.68; F (1,1254) ¼ 709.8; P<0.001). Despite this IOP reduction, there was no significant alteration in MD (0.04 dB; 95% CI, 0.16 to 0.09; F (1,1254) ¼ 0.29; P ¼ 0.59) or PSD (0.03 dB; 95% CI, 0.08e0.02; F (1254) ¼ 1.69; P ¼ 0.19) compared with the control group. Figure 2 shows the difference between baseline (0 month) and the 6-month visit for IOP, MD, and PSD for both our treatment and control (on medication at both visits) groups drawn from OHTS II. A significant reduction in IOP was seen in our treatment group compared with our control group (5.1 mmHg; 95% CI, 4.52 to 5.67; F (1,797) ¼ 368; P<0.001). Despite this IOP reduction, there was no significant alteration in MD (0.05 dB; 95% CI, 0.24 to 0.15; F (1,797) ¼ 0.23; P ¼ 0.62) or PSD (0.06 dB; 95% CI, 0.07 to 0.19; F (1,797) ¼ 0.96; P ¼ 0.33) compared with our control group.
Discussion Our results show that a reduction in IOP in ocular hypertensive patients did not cause an accompanying change in the average sensitivity of the visual field measured using standard automated perimetry. This argues against a relationship between change in IOP and contrast sensitivity in eyes without glaucoma, and so strengthens previous interpretations that a change in contrast sensitivity in glaucomatous eyes when IOP is lowered is the result of glaucoma-induced, reversible visual dysfunction.1e4 However, note that our results do not provide any further support for whether sensitivity increases when IOP is reduced in glaucomatous eyes per se. The absence of adequate control groups in some previous studies2,4 substantially weakens their conclusions that sensitivity increases in glaucomatous eyes when IOP is lowered, although other studies have incorporated control groups5 or eyes1,3 to minimize such effects as perceptual learning19 or regression-to-the-mean. Because of the large size of our sample, coupled with the tightly controlled measuring of IOP and visual fields in OHTS,15 our negative finding can be taken with confidence. Although our reductions in IOP are smaller than those involving surgical interventions,1,3 they are larger5 or similar to4 those from
previous studies showing contrast sensitivity improvement on grating-based measures of contrast sensitivity with medical treatment of glaucoma. Therefore, our negative findings are unlikely to be explained by assuming that our IOP reduction is below a threshold value for which sensitivity change occurs. Our findings are consistent with earlier research by Flammer and Drance,11 but address several concerns within their study. That these authors showed no change in perimetric contrast sensitivity in ocular hypertension but did find changes in glaucomatous observers despite a smaller sample size argues against the idea that their study was insufficiently powered to find changes in their ocular hypertensive group. However, it is possible that the sensitivity change seen in their glaucoma group was not due to IOP reduction. First, Flammer and Drance11 questioned whether their findings of improved sensitivity in treated glaucoma may have been the result of a pharmacologic effect of acetazolamide, rather than a change in IOP. Second, these authors found most change in sensitivity in defective areas of the visual field, but suggested this effect was greater than that explicable through regressionto-the-mean effects noted in previous work.14 It is likely that this previous assessment of regression-to-the-mean was an underestimate, however, because it examined those with “very early stages” of glaucoma14 in contrast to the average Octopus perimeter sensitivity of 17 dB for the glaucoma group reported in their later study.11 It is now well established that perimetric variability increases as field sensitivity decreases,20,21 suggesting that regression-to-the mean effects assessed in early glaucoma should underestimate effects seen in more advanced disease. In addition, testeretest limits for standard achromatic perimetry are asymmetric in areas of substantial loss22 and would favor larger sensitivity improvements more than sensitivity declines on retest. The OHTS assessed participants at 6-month intervals, providing a limit to the temporal resolution with which the influence of IOP could be assessed in our study. It may be thought that this period is sufficient for our results to be potentially influenced by disease progression, and that this could have counterbalanced any improvement in sensitivity with IOP lowering. We do not believe this is likely. First, we only examined those participants who did not progress to a diagnosis of glaucoma over the 6-year period of each OHTS phase (OHTS I and OHTS II: see “Methods”), and so the likelihood of any significant visual field progression in the first 6 months of this period is remote. Furthermore, in the unlikely event that progression occurred, we would expect a significant change in PSD as subtle scotomas develop, even if MD remained constant because the average depth of this progression happened to be exactly matched by a sensitivity improvement caused by a lowering of IOP. We find no such change in PSD (Fig 1), giving assurance that our control and treatment groups are well matched in terms of any progression, however unlikely. Although we show that visual field sensitivity did not alter, this is not to say that other visual functions may not be perturbed by changing IOP in eyes without glaucoma. For example, IOP lowering can alter pattern electroretinogram responses in ocular hypertensive observers (Ventura et al23 have a summary of recent findings). There is evidence that the influence of a transient elevation in IOP is more profound in ocular hypertensive patients than controls, despite pressure being
3
Ophthalmology Volume -, Number -, Month 2014
Figure 1. Data from the Ocular Hypertension Treatment Study: difference in intraocular pressure (IOP), mean deviation (MD), and pattern standard deviation (PSD) between baseline (0 months) and 6-month follow-up for the control group (no medication) and treatment group placed on IOP-lowering medication immediately after baseline. Error bars give 1 standard error of the mean. Initial IOP, MD, and PSD values for the treatment and control groups were 25.7 mmHg, 0.24 decibels (dB), and 1.95 dB, and 24.8 mmHg, 0.15 dB, and 1.94 dB, respectively. Negative values and positive values mean worsening values at follow-up for MD and PSD, respectively. Tx ¼ treatment group.
elevated to identical levels in both groups,24 suggesting that ocular hypertensive patients may have an increased susceptibility to IOP change in some circumstances. However, the presence of such alterations in ocular hypertension cannot necessarily be interpreted as subclinical manifestations of glaucoma or that early glaucomatous changes are potentially reversible. For example, Tytla et al12 observed that the prevalence of flicker sensitivity abnormalities in previous studies of hypertensive eyes was high, and greater than expected given the natural history of the disease, a concern shared by other authors.25 Therefore,
subtle sensitivity defects in ocular hypertension may represent stable changes associated with ocular hypertension that are not necessarily related to the progressive changes that may occur subsequently in glaucoma. This is true irrespective of whether these subsequent glaucomatous changes are reversible or permanent. In conclusion, we show that reduction of elevated IOP in eyes without glaucoma does not produce a significant change in contrast sensitivity using automated perimetry. Our results provide a necessary control required to strengthen the argument that previous studies showing visual field improvement
Figure 2. Data from Ocular Hypertension Treatment Study Phase II: difference in intraocular pressure (IOP), mean deviation (MD), and pattern standard deviation (PSD) between baseline (0 months) and 6-month follow-up for the control group (always on medication) and treatment group placed on IOPlowering medication immediately after baseline. Error bars give 1 standard error of the mean. Initial IOP, MD, and PSD values for the treatment and control groups were 23.6 mmHg, 0.33 decibels (dB), and 2.11 dB, and 18.2 mmHg, 0.17 dB, and 2.06 dB, respectively. Negative values and positive values mean worsening values at follow-up for MD and PSD, respectively. Tx ¼ treatment group.
4
Anderson and Stainer
Contrast sensitivity and IOP lowering
with IOP reduction in glaucoma1e4 have demonstrated glaucoma-induced dysfunction rather than simply a general relationship between IOP and contrast sensitivity. Acknowledgments. The authors thank Mae Gordon and Michael Kass for feedback on the manuscript.
References 1. Gandolfi SA, Cimino L, Sangermani C, et al. Improvement of spatial contrast sensitivity threshold after surgical reduction of intraocular pressure in unilateral high-tension glaucoma. Invest Ophthalmol Vis Sci 2005;46:197–201. 2. Prata TS, Piassi MV, Melo LA Jr. Changes in visual function after intraocular pressure reduction using antiglaucoma medications. Eye (Lond) 2009;23:1081–5. 3. Gandolfi SA. Improvement of visual field indices after surgical reduction of intraocular pressure. Ophthalmic Surg 1995;26: 121–6. 4. Evans DW, Hosking SL, Gherghel D, Bartlett JD. Contrast sensitivity improves after brimonidine therapy in primary open angle glaucoma: a case for neuroprotection. Br J Ophthalmol 2003;87:1463–5. 5. Pomerance GN, Evans DW. Test-retest reliability of the CSV1000 contrast test and its relationship to glaucoma therapy. Invest Ophthalmol Vis Sci 1994;45:3357–61. 6. Tyler CW, Ryu S, Stamper R. The relation between visual sensitivity and intraocular pressure in normal eyes. Invest Ophthalmol Vis Sci 1984;25:103–5. 7. Wittström E, Schatz P, Lövestam-Adrian M, et al. Improved retinal function after trabeculectomy in glaucoma patients. Graefes Arch Clin Exp Ophthalmol 2010;248:485–95. 8. Aydin A, Wollstein G, Price LL, et al. Optical coherence tomography assessment of retinal nerve fiber layer thickness changes after glaucoma surgery. Ophthalmology 2003;110:1506–11. 9. Heijl A, Lindgren G, Olsson J. Normal variability of static perimetric threshold values across the central visual field. Arch Ophthalmol 1987;105:1544–9. 10. Heijl A, Lindgren G, Olsson J. A package for the statistical analysis of visual fields. Doc Ophthalmol Proc Ser 1987;49:153–68. 11. Flammer J, Drance SM. Effect of acetazolamide on the differential threshold. Arch Ophthalmol 1983;101:1378–80. 12. Tytla ME, Trope GE, Bunic JR. Flicker sensitivity in treated ocular hypertension. Ophthalmology 1989;97:36–43. 13. Farinelli AC, Mikelberg FS, Drance SM, et al. Effect of acute medical reduction of intraocular pressure on the visual field and optic disc in ocular hypertension. Can J Ophthalmol 1988;23:216–8.
14. Flammer J, Drance SM. The effect of a number of glaucoma medications on the differential light threshold. Doc Ophthalmol Proc Ser 1983;35:145–8. 15. Gordon MO, Kass MA; Ocular Hypertension Treatment Study Group. The Ocular Hypertension Treatment Study: design and baseline description of the participants. Arch Ophthalmol 1999;117:573–83. 16. Kass MA, Heuer DK, Higginbotham EJ, et al; Ocular Hypertension Treatment Study Group. 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: 701–13. 17. Gordon MO, Beiser JA, Brandt JD, et al; Ocular Hypertension Treatment Study Group. The Ocular Hypertension Treatment Study: baseline factors that predict the onset of primary openangle glaucoma. Arch Ophthalmol 2002;120:714–20. 18. Kass MA, Gordon MO, Gao F, et al; Ocular Hypertension Treatment Study Group. Delaying treatment of ocular hypertension: the Ocular Hypertension Treatment Study. Arch Ophthalmol 2010;128:276–87. 19. McKendrick AM, Battista J. Perceptual learning of contour integration is not compromised in the elderly. J Vis [serial online] 2013;13:5. Available at: http://www.journalofvision. org/content/13/1/5.long. Accessed April 16, 2014. 20. Chauhan BC, Tompkins JD, LeBlanc RP, McCormick TA. Characteristics of frequency-of-seeing curves in normal subjects, patients with suspected glaucoma, and patients with glaucoma. Invest Ophthalmol Vis Sci 1993;34:3534–40. 21. Spry PG, Johnson CA, McKendrick AM, Turpin A. Variability components of standard automated perimetry and frequency-doubling technology perimetry. Invest Ophthalmol Vis Sci 2001;42:1404–10. 22. Artes PH, Hutchison DM, Nicolela MT, et al. Threshold and variability properties of matrix frequency-doubling technology and standard automated perimetry in glaucoma. Invest Ophthalmol Vis Sci 2005;46:2451–7. 23. Ventura LM, Golubev I, Lee W, et al. Head-down posture induces PERG alterations in early glaucoma. J Glaucoma 2013;22:255–64. 24. Colotto A, Falsini B, Salgarello T, et al. Transiently raised intraocular pressure reveals pattern electroretinogram losses in ocular hypertension. Invest Ophthalmol Vis Sci 1996;37: 2663–70. 25. Ruben ST, Arden GB, O’Sullivan F, Hitchings RA. Pattern electroretinogram and peripheral colour contrast thresholds in ocular hypertension and glaucoma: comparison and correlation of results. Br J Ophthalmol 1995;79:326–31.
Footnotes and Financial Disclosures Originally received: February 12, 2014. Final revision: March 25, 2014. Accepted: April 17, 2014. Available online: ---.
Manuscript no. 2014-230.
Horncrest Foundation, awards to the Department of Ophthalmology and Visual Sciences at Washington University, the National Institutes of Health Vision Core Grant P30 EY 02687, Merck Research Laboratories, Pfizer, Inc. (Whitehouse Station, NJ), and unrestricted grants from Research to Prevent Blindness, Inc (New York, NY).
Department of Optometry and Vision Sciences, The University of Melbourne, Parkville, Australia. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.
Abbreviations and Acronyms: CI ¼ confidence interval; dB ¼ decibels; IOP ¼ intraocular pressure; MD ¼ mean deviation; OHTS ¼ Ocular Hypertension Treatment Study; PSD ¼ pattern standard deviation.
Supported by Australian Research Council Future Fellowship FT120100407 (A.J.A.). The original OHTS was supported by awards from the National Eye Institute, National Center on Minority Health and Health Disparities, National Institutes of Health (Grants EY09341, EY09307),
Correspondence: Andrew J. Anderson, MSc(Optom), PhD, Department of Optometry and Vision Sciences, The University of Melbourne 3010, Australia. E-mail: [email protected].
5