A shift of the postoperative graphs to the left for the slow component indicates a slowing down of the progression rate after surgery. No statistically significant difference was identified in the visual field decay rate for the fast component or the regression slopes for the mean deviation (MD) and the Visual Field Index (VFI).
Lee J, Morales E, Yu F, Afifi AA, Kim E, Abdollahi N, Nouri-Mahdavi K, Caprioli J. Effect of Cataract Extraction on the Visual Field Decay Rate in Patients With Glaucoma. JAMA Ophthalmol. 2014;132(11):1296-1302. doi:10.1001/jamaophthalmol.2014.2326
A visual field parameter that is resistant to cataract formation and extraction would help monitor glaucomatous visual field progression in patients with coexisting glaucoma and cataract.
To evaluate the effect of cataract surgery on the slow and fast components of visual field decay in a group of patients with glaucoma.
Design, Setting, and Participants
Retrospective, interventional, longitudinal study. Eighty-five eyes of 68 patients with open-angle glaucoma who had cataract extraction were included. All patients had 5 or more reliable visual field measurements before and after surgery.
A pointwise exponential regression was used to perform trend analysis on thresholds at visual field test locations before and after cataract surgery. The test locations were ranked according to the decay rate and were partitioned into slow and fast groups.
Main Outcomes and Measures
The slow and fast visual field rate components were measured before and after cataract surgery and were compared. Linear regressions of the mean deviation and the visual field parameter were performed against time and were compared before and after surgery.
The mean (SD) mean deviation was −5.5 (5.1) dB before cataract surgery and −5.0 (4.9) dB after cataract surgery (P = .002). The mean (SD) Visual Field Index was 86.4% (13.5%) before cataract surgery and 86.6% (13.3%) after cataract surgery (P = .30). The mean (SD) slow component rate decreased from 0.48% (0.73%) per year before surgery to 0.26% (0.42%) per year after surgery (P = .04). No statistically significant difference was identified in the fast component mean (SD) rate per year before surgery (3.37% [4.05%]) vs per year after surgery (3.46% [3.56%]) (P = .29).
Conclusions and Relevance
Cataract progression seems to be the main determinant for the slow visual field rate component and does not change the fast visual field rate component. We conclude that the method used can help reduce the confounding effects of cataract progression and cataract extraction on measured perimetric progression in glaucoma.
Cataract and glaucoma frequently coexist because the prevalences of cataract and open-angle glaucoma increase with aging.1,2 Trabeculectomy augments the risk of cataract formation by 78% in patients with glaucoma.3 Other treatments for glaucoma, including laser trabeculoplasty and medications, seem to increase this risk as well.4
Simulated cataract, minimally affecting visual acuity, can depress visual field sensitivities.5,6 The development of cataract causes diffuse visual field loss in patients with glaucoma, and differentiating perimetric deterioration caused by cataract from that caused by glaucoma worsening remains challenging.7 The visual field mean deviation (MD) and visual acuity improve significantly after cataract extraction,7- 13 although some studies14,15 reported no significant changes in the MD. Similarly, changes in the pattern SD (PSD) and the Visual Field Index (VFI) have not been found to be significant after cataract extraction in some studies,8,11,13,14,16 whereas other studies9,10,12,17,18 have shown that the PSD worsened and the VFI improved after cataract extraction in patients with coexisting glaucoma and cataract. For example, the development or the worsening of cataract confounded the results of the Collaborative Normal-Tension Glaucoma Study.19 The favorable effect of intraocular pressure (IOP) reduction on visual field progression in patients with normal-tension glaucoma was detected only when the effect of cataract on the visual field results was removed.19
The measurement of the visual field decay rate is an important step to appropriate decision making to preserve vision in patients with glaucoma.20 While numerous studies have focused on the short-term effects of cataract extraction on visual acuity or the visual field parameters, little is known about the association between cataract extraction and decay rates in localized areas of the visual field. Our group developed a technique based on trend analysis that uses a pointwise exponential model to fit the behavior of individual test locations in a visual field series.21 This method separates slow and fast rate components of visual field decay, while preserving spatial information across a wide range of disease severity.21,22 The objective of this study was to evaluate the effect of cataract extraction on the slow and fast components of visual field decay as determined with a pointwise exponential regression model.
This study was conducted in accord with the tenets set forth in the Declaration of Helsinki and was approved by the UCLA Institutional Review Board. The institutional review board approved that informed consent from patients was waived because this is a retrospective study. A retrospective review of medical records of patients with open-angle glaucoma who underwent cataract surgery with intraocular lens implantation at the Jules Stein Eye Institute, David Geffen School of Medicine at UCLA, between July 1, 2000, and June 30, 2010, was performed. Patients who underwent combined cataract and trabeculectomy surgery were excluded, as were any patients who underwent glaucoma surgery during the period of interest. All cataract surgery was performed by one of us (J.C.) with a temporal clear cornea, small-incision phacoemulsification technique using peribulbar anesthesia. The patients received a silicone foldable intraocular lens (STAAR Surgical) or an acrylic foldable intraocular lens (Alcon Laboratories Inc) implanted in the capsular bag.
All consecutive patients with open-angle glaucoma who had 5 or more reliable visual field measurements before and after surgery were included in the study. Reliability was defined as less than 30% fixation losses, less than 30% false-positive rates, and less than 30% false-negative rates. All visual field tests were performed using an automated visual field analyzer (Humphrey Field Analyzer; Carl Zeiss Ophthalmic Systems Inc) with a 24-2 test pattern, size III white stimulus with the Swedish Interactive Threshold Algorithm standard strategy. The last visual field before surgery and the first visual field after surgery were required to have been measured within 1 year of cataract surgery.
The technique of measuring rates of visual field decay has been published in detail21,22 and is summarized herein. Rates of visual field decay were calculated by a pointwise exponential regression analysis of threshold sensitivities at 52 test locations, excluding the 2 locations corresponding to the blind spot. The association between the response variable (threshold sensitivity) and the explanatory variable (follow-up duration) was characterized by the following exponential regression models: y = ea + bx or, equivalently, lny = a + bx, where a is the intercept, x is time, b is the mean annual rate of change in ln y, and eb represents the ratio of y in a given year to y in the year before. The decay rate is defined as 1 − eb. To facilitate an intuitive clinical understanding of the magnitude of decay rates, the coefficients of the exponential regressions were converted into percentage per year deterioration rates, where the percentage per year decay rate is (1 − eb) × 100.
The 52 visual field test locations were ranked according to the decay rate and were clustered into 2 subgroups (slow and fast components) based on the P value for the difference in the mean rates between 2 clusters. For each possible partition, starting with a minimum number of 5 locations in a cluster, we computed a t test statistic, and the corresponding P values were adjusted for multiple testing. Because multiple simultaneous t tests were performed, it is desirable to correct the P values to control for false-positive results. Accordingly, Benjamini-Hochberg correction was used to adjust the P values for multiple testing.23 For each eye, the mean slow and fast decay rates were calculated for the partitioned components during the preoperative and postoperative periods separately.
The MD and VFI values for the preoperative and postoperative periods were also analyzed; univariate linear regression analyses of these values were performed against time, and corresponding regression slopes (in decibels per year and in percentages per year) were compared before and after cataract extraction. The MD, PSD, and VFI values of the last visual field measurement before surgery and the first visual field measurement after surgery were also compared.
The preoperative best-corrected visual acuity was recorded from the visit just before the surgery. All eyes were refracted within 1 month after cataract extraction. Snellen visual acuities were converted to a scale of the logMAR for comparison. The mean IOP and the mean number of medications were compared during the preoperative and postoperative periods.
All statistical analyses were performed with available software (SPSS for Windows 21.0; SPSS Inc). The normality of numerical data distribution was checked with Kolmogorov-Smirnov test. Differences between the preoperative and postoperative data were compared with Wilcoxon signed rank test. Spearman rank correlation coefficient was used to estimate correlations between various outcomes (eg, the preoperative and postoperative visual field decay rates) and each of the following continuous variables: age, vertical cup-disc ratio (VCDR), MD, PSD, VFI, mean IOP, and mean number of glaucoma medications. Multivariable linear regression analysis with a stepwise variable selection method was used to identify significant parameters that may explain the visual field decay rates in the slow and fast components.
Eighty-five eyes of 68 patients with open-angle glaucoma who had cataract extraction were included in the study. The study eyes were followed up for a mean (SD) of 11.6 (2.9) years, and a mean (SD) of 19.7 (5.3) visual field measurements were available during the entire follow-up period. The demographic, clinical, and ocular characteristics of the study sample are summarized in Table 1. The mean (SD) logMAR best-corrected visual acuity significantly improved from 0.33 (0.19) before surgery to 0.14 (0.11) (P < .001) after cataract extraction (Table 2). The mean (SD) IOP decreased from 13.4 (3.2) mm Hg to 12.4 (3.7) mm Hg after cataract extraction (P = .002). Similarly, the mean (SD) number of glaucoma medications was reduced from 1.5 (0.9) to 1.0 (0.9) after cataract extraction (P < .001).
The mean (SD) MD value improved from −5.5 (5.1) dB before surgery to −5.0 (4.9) dB after surgery (P = .002), while the mean (SD) PSD value worsened from 5.7 (4.1) dB before surgery to 6.0 (4.1) dB after surgery (P = .003) (Table 2 and Figure). No statistically significant difference was identified in the MD decay rates (slopes) after surgery (mean [SD], −0.29 [0.72] vs −0.29 [0.52] dB per year, P = .48). The improvement in the MD after surgery was inversely related to the last visual field MD before surgery (r = −0.32, P = .003). The mean (SD) VFI did not change after surgery (86.4% [13.5%] vs 86.6% [13.3%], P = .30), and no statistically significant difference was identified between the mean (SD) VFI rates of change before and after surgery (−0.89% [1.97%] vs −0.96% [1.57%] per year, P = .20).
The mean (SD) visual field decay rate for all locations was 2.05% (2.93%) per year before surgery and 2.19% (2.07%) per year after surgery (P = .22) (Table 3 and Figure). The mean (SD) visual field decay rates in the slow component were 0.48% (0.73%) per year before surgery and slowed to 0.26% (0.42%) per year after surgery (P = .04). No statistically significant difference was identified in the visual field decay rates in the fast component (mean [SD], 3.37% [4.05%] per year before surgery and 3.46% [3.56%] per year after surgery, P = .29). Fifty-seven percent (1020 of 1800) of test locations identified as the fast component before surgery remained fast after cataract surgery.
Correlation coefficients were estimated to explore the effects of age at the time of surgery, baseline VCDR, MD, PSD, VFI, mean IOP, and mean number of glaucoma medications on the slow and fast visual field decay rates before and after surgery. No statistically significant correlation was identified between the above variables and the slow visual field decay rates before and after surgery. However, the preoperative fast visual field decay rates were positively associated with the baseline VCDR and PSD (r = 0.39, P < .001 and r = 0.31, P = .004, respectively) and were inversely associated with the baseline VFI (r = −0.28, P = .009). A higher postoperative decay rate in the fast component was positively correlated with the baseline VCDR and PSD of the first visual field after surgery (r = 0.31 and r = 0.30, respectively, P = .005 for both) and demonstrated an inverse association with the MD and VFI of the first visual field after surgery (r = −0.23, P = .04 and r = −0.33, P = .002, respectively). Changes in the mean IOP correlated directly with changes in the slow and fast visual field decay rates (r = 0.32, P = .003 and r = 0.27, P = .01, respectively). A decrease in the mean postoperative IOP entailed a slowing of the slow and fast visual field decay rates after surgery. A worse baseline VFI predicted a higher fast visual field decay rate before and after surgery (β = −0.14 and β = −0.11, respectively, P < .001 for both, adjusted R2 = 0.11 for both).
Cataract extraction had a differential effect on the slow and fast components of visual field decay. No statistically significant difference was identified in the mean visual field decay rate for all locations after cataract removal. However, when individual visual field test locations were partitioned into slow and fast components, the mean slow visual field decay rate slowed significantly after cataract surgery, whereas no statistically significant difference was identified in the fast visual field decay rate. As expected, visual acuity improved after cataract extraction, and the mean IOP and the mean number of glaucoma medications decreased after cataract extraction.7 Worsening of cataract may lead not only to loss of visual acuity but also to a diffuse reduction in visual field sensitivities, with depression of the MD in healthy individuals and in patients with open-angle glaucoma.8,16,24 Earlier studies7- 13,17,18 reported that the MD improved after cataract extraction and that a greater change in the MD occurred in eyes with more improvement in visual acuity.
No statistically significant difference was identified in the MD rates of change after cataract extraction in the present study. Although the MD is a weighted mean of the total deviation values, Smith et al25 found that linear regression of MD lacked sensitivity for detecting glaucoma progression. Because MD represents the mean damage across the entire visual field, the rate of the MD change would be expected to be insensitive to localized worsening of visual field damage.26 In this study, the PSD values increased significantly after surgery. Actual glaucomatous visual field defects may be masked by the relative depression caused by cataract. These findings are in agreement with the results of an earlier study10 reporting that the postoperative visual acuity improvement was significantly correlated with deterioration of the PSD. Some studies8,11,12,14,16 found that the PSD was not affected by lens opacity or cataract extraction. However, other studies7,9,10,13,17,18 found that the PSD worsened after cataract extraction in glaucomatous eyes. Hayashi et al7 reported an increase in the PSD among patients with dense scotomas but not among patients without dense field defects.
We found that the visual field decay rate for the slow component alone slowed significantly after cataract extraction, while no statistically significant difference was identified in the visual field decay rate for the fast component. Because the slow component represents more diffuse and nonspecific deterioration related to aging and media opacity, these results suggest that cataract formation is a main contributing factor for deterioration of the slow component.21,22 In contrast, cataract extraction minimally affects the fast component because that component reflects more focal, rapidly deteriorating test locations from glaucoma.21,22
Nouri-Mahdavi et al27 reported that the rate of visual field progression during the early follow-up period was a strong predictor of future progression in glaucomatous eyes. Therefore, the development of a visual field parameter that is resistant to cataract formation and extraction would offer a distinct advantage in monitoring glaucomatous visual field progression in patients with coexisting glaucoma and cataract. Given the lack of sensitivity of an MD regression analysis for detecting change in glaucoma, the MD regression slope may overestimate glaucomatous visual field damage in eyes with worsening cataract.25,26,28 The PSD, considered an index for localized visual field defect, may underestimate visual field damage because the PSD change can be masked by cataract as the diffuse component of visual field loss manifests as glaucoma deteriorates.7,25 In addition, it is inappropriate to interpret rates of the PSD with linear models because its value reaches a maximum with moderate glaucoma damage and decreases as the visual field continues to advance.25
We found that the VFI and the VFI rates of change after cataract extraction were similar to those before surgery. Our findings agree with the results of previous studies.13,28 Bengtsson and Heijl28 developed a VFI with significant weighting of the central areas of the visual field to express functional loss as a percentage compared with an age-matched reference and to calculate the rate of visual field deterioration. They concluded that the VFI was less affected by cataract and cataract surgery than the MD. However, data regarding the differences in visual field rates of progression with the VFI before and after cataract extraction were not presented. Another study12 reported that cataract extraction resulted in a statistically significant change in the VFI, together with the MD. The authors proposed that overall increase in sensitivity after cataract extraction may increase the VFI because of the effect of the central weighting of the VFI.12
Some controversy exists concerning the effect of cataract extraction on the PSD and the VFI. The different findings of previous studies7- 14,16- 18,28 may be the result of the inclusion of patients with various stages of glaucoma damage and various severities of cataract at the time of surgery in different populations.
Linear regression of the MD and the VFI has the advantage of providing a rapid and intuitive guide for the physician when analyzing serial visual fields of an individual patient. However, the MD change is not sensitive to regional changes or focal components of damage. Likewise, the VFI provides no spatial or pattern information about the visual field. Artes et al29 reported that the VFI may be less able to detect visual field damage in early stages of glaucoma because of the ceiling effect caused by pattern deviation maps.
As opposed to these global visual field parameters, the method used herein enables the measurement of the rate of visual field worsening for the slow and fast rate components separately, without loss of spatial or pattern information.21,22 This method does not rely on proprietary, machine-stored normative data for measuring the rate of progression.21 While we believe that most of the glaucoma signal resides in the fast component, we cannot exclude the possibility that some slow, diffuse glaucomatous visual field loss is present in the slow component.
In the present study, no statistically significant difference was identified in the MD regression slope after surgery. However, when the visual field locations were clustered into slow and fast decay components, we found that the slow component decay rate was significantly slowed, while the fast component decay rate remained unchanged. This suggests that our approach can unmask significant visual field changes occurring at the same time when cataract is progressing and help with earlier detection of glaucoma deterioration.
We found positive correlations between the baseline VCDR and the visual field decay rates in the fast component before and after cataract extraction, as well as between the baseline PSD and the visual field decay rates in the fast component before and after cataract extraction. These findings are in agreement with earlier studies30- 33 reporting that the baseline cup-disc ratio was predictive of future visual field damage and that the baseline visual field loss was associated with glaucomatous visual field progression.
This study has several limitations. One limitation is its retrospective nature; however, the follow-up period and the number of visual fields measured did not differ before and after surgery, and a single surgeon performed all surgery with a standardized method. Patients enrolled in the study had mild to moderate glaucoma based on the mean baseline and final MDs. These results may not apply to patients with more advanced glaucomatous damage. Patients enrolled in this study had mild glaucoma (mean baseline MD, −4.6), and only 0.3% (11 of 3801) of test locations had baseline threshold sensitivities between 0 to 10 dB. Therefore, we believe that any floor effect is minimal in this study. We see measurable decay rates even for those few locations with a baseline sensitivity of less than 10 dB. The effect of lens opacities on visual field test results may be influenced by lens opacity location and severity,13 which was not addressed in this study. Seventy-one percent (48 of 68) of patients in the study were of white race/ethnicity; therefore, the effects of cataract extraction on the visual field decay rate may be different in other populations. We did not consider the presence of posterior capsular opacity in the interpretation of visual fields.
In summary, this study demonstrates that the slow component rate of visual field decay is significantly slowed after cataract extraction, while no statistically significant difference was identified in the fast component rate of visual field decay in a longitudinal visual field series. The results suggest that cataract formation minimally affects the fast component decay rate. We conclude that the method used herein can help reduce the confounding effects of cataract progression and cataract extraction on measured perimetric progression in glaucoma.
Submitted for Publication: March 9, 2014; final revision received May 4, 2014; accepted May 6, 2014.
Corresponding Author: Joseph Caprioli, MD, Jules Stein Eye Institute, David Geffen School of Medicine at UCLA, 100 Stein Plaza, Los Angeles, CA 90095 (firstname.lastname@example.org).
Published Online: July 31, 2014. doi:10.1001/jamaophthalmol.2014.2326.
Author Contributions: Dr Caprioli had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Lee, Afifi, Kim, Caprioli.
Acquisition, analysis, or interpretation of data: Lee, Morales, Yu, Abdollahi, Nouri-Mahdavi, Caprioli.
Drafting of the manuscript: Lee, Afifi, Kim, Caprioli.
Critical revision of the manuscript for important intellectual content: Lee, Morales, Yu, Afifi, Abdollahi, Nouri-Mahdavi, Caprioli.
Statistical analysis: Lee, Morales, Yu, Kim, Nouri-Mahdavi, Caprioli.
Administrative, technical, or material support: Abdollahi.
Study supervision: Caprioli.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported.
Funding/Support: This study was supported by unrestricted grants from Research to Prevent Blindness, Inc.
Role of the Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.