Year : 2022 | Volume
: 34 | Issue : 1 | Page : 81--83
Department of Ophthalmology, Regional Institute of Ophthalmology, Thiruvananthapuram, Kerala, India
Department of Ophthalmology, Regional Institute of Ophthalmology, Thiruvananthapuram, Kerala
|How to cite this article:|
Mathew A. Journal Review.Kerala J Ophthalmol 2022;34:81-83
|How to cite this URL:|
Mathew A. Journal Review. Kerala J Ophthalmol [serial online] 2022 [cited 2022 Oct 7 ];34:81-83
Available from: http://www.kjophthal.com/text.asp?2022/34/1/81/343657
Muller Cell Cone-Associated Foveal Detachment as a Risk Factor for Visual Acuity Loss After Glaucoma Filtering Surgery
Citation: Chihara E, Chihara T, Matsuzaki S. Muller cell cone-associated foveal detachment as a risk factor for visual acuity loss after glaucoma filtering surgery. Retina. 2021;41:2571-2577 DOI: 10.1097/IAE.0000000000003216.
Short summary of the article: The authors assessed the association of hypotony-associated foveal lesions (FovLs) as a risk factor for long-standing visual loss after filtering surgery with mitomycin C for glaucoma. They concluded that Müller cell cone-associated FovLs lead to long-lasting visual acuity loss after filtering surgery.
An estimated 4.5 million people with moderate-to-severe vision impairment and 3.2 million people have blindness attributable to glaucoma. There are an estimated 12 million people with glaucoma in India and 1.5 million people with blindness attributable to glaucoma in India. Filtering surgery that aims at sustained management of intraocular pressures, thereby visual acuity is a recognized strategy in the management of glaucoma. However, visual acuity loss of four or more lines is found in 14.6% of patients after trabeculectomy with adjunctive mitomycin C, with permanent visual acuity loss reported in 28.3% of cases. Anterior segment complications may lead to short-term visual acuity loss after filtering surgery while posterior segment causes of visual acuity loss are usually long lasting. Hypotony maculopathy (HM) occurred in up to 20% of patients and is identified as a risk factor for visual acuity loss following trabeculectomy. However, visual acuity deterioration is also found in patients who developed hypotony following trabeculectomy without any manifest clinical features of HM. Eyes such as hypotony following trabeculectomy were noted to have 40 μm increase in central retinal thickness. Subclinical level macular abnormalities were noted in more than 50% of eyes with posttrabeculectomy hypotony and the presence of epiretinal membranes (ERM) was also documented in such cases. The authors hypothesized that hypotony-associated FovLs and postsurgical outer retinal changes after trabeculectomy with mitomycin C may lead to postsurgery visual acuity loss and designed a study to determine postsurgical changes in the outer retina assessed using optical coherence tomography (OCT) in those patients who underwent trabeculectomy with mitomycin C, and to assess the effect of ERM on posttrabeculectomy visual acuity deterioration.
There are several study designs that are possible to address these aims. The study involves an intervention, and a randomized clinical trial (RCT) can be considered. However, the study is not looking at a comparison of the effectiveness of interventions, and hence a RCT may not be necessary. Alternate designs that can be considered include a pre–post interventional design that compares presurgical baseline with postsurgical changes in the same subject and an interventional case series design. Both these alternate designs can be either prospective or retrospective. The retrospective design has limitations on selection bias, standardization of assessments, and completeness of data while offering benefits of faster completion of the study. The authors chose to do a retrospective interventional case series to answer the study questions. This may be a pragmatic choice given the relative rarity of the condition and the ability to recruit a reasonable number of cases in a reasonable period. However, the choice of this design limits the interpretation of results only to the generation of a hypothesis and does not allow for a formal statistical testing of hypothesis and generation of conclusive evidence of any association. The lack of association in a case series may be a true lack of association or maybe due to an inadequate sample size.
The authors recruited one eye each of 44 patients for the study. Six of these 44 patients dropped out after 6 months of follow-up and complete details could be analyzed in the remaining 38 patients. The data of the 6 patients who dropped out were treated as missing statistical data in the study. All recruited patients underwent trabeculectomy with mitomycin C and postsurgical assessments were done at 1 week, and 1, 3, 6, and 12 months after surgery. The authors studied the mean retinal thickness at the central area, inner paramacular and outer paramacular area were studied using the spectral domain-OCT. HM was defined as a low intraocular pressure (IOP), ophthalmoscopic radial retinal folds, choroidal folds, tortuous vessels, optic disk edema, and/or the presence of intraretinal fluid and choroidal folds detected using OCT. The presence of ERM, choroidal folds, and foveal detachment was assessed by two independent researchers using crossline and retinal map images from the device and lesions were judged as positive when both researchers agreed on the findings. The authors do not mention the agreement between observers and do not mention if postsurgery assessments are masked to presurgery assessment results. A selection bias is also possible depending on the patient population that accesses care at the study setting (the distribution of severity of cases) and the personnel recruiting cases.
The key variables to answer the research question includes Best corrected Visual Acuity (BCVA) compared to presurgery VA, postsurgery changes in IOP (hypotony) and postsurgery changes in the retina compared to presurgery assessment based on OCT findings. The changes in means can be assessed using parametric tests such as a paired t-test or analysis of variance and nonparametric tests such as the Wilcoxon rank-sum test or Mann–Whitney U-test. A nonparametric test will be preferable as the distribution will not be normal given the low sample size. Changes in proportions can be assessed using a Fisher's exact test rather than a Chi-square test as we anticipate cell counts <5 in a 2 × 2 table as the sample size is low. The authors performed nonparametric tests for means and Fisher's exact test for proportions. A regression analysis was performed to explore associations. The use of a regression model is appropriate for associations, although formal testing of hypothesis in a case series must be interpreted with caution due to inadequate sample size. Regression is indicated when one of the variables is an outcome and the other one is a potential predictor of that outcome, in a cause-and-effect relationship. If the outcome is a continuous variable, a linear regression model is indicated, and, if it is binary, a logistic regression is used. The authors also did a nonparametric Spearman correlation for visual acuity and IOP, although this is not essential within the context of their aims. A key aspect to remember while interpreting results is that the study is a case series design, the statistical testing in a case series design is to provide evidence to support the hypothesis or to generate a hypothesis, to determine sample sizes for further studies, and that the results do not provide conclusive evidence of a lack of association.
The authors reported a significant reduction in VA at 3 months (P < 0.001) and 12 months (P = 0.0046, Wilcoxon signed-rank test). After a large drop in the VA at 3 months, there was a mild recovery of the Early Treatment Diabetic Retinopathy Study letter score from 77.1 ± 14.7 letters to 80.7 ± 11.4 letters at 12 months; however, this difference was not statistically significant (P = 0.155, Wilcoxon signed-rank test). The difference in deterioration of the VA between the phakic and pseudophakic eyes was not statistically significant. The association between the deterioration of VA at 3 months or 12 months and presurgical IOP, the number of presurgical medications, lowest postsurgical IOP, and postsurgical IOP at 1 week was not statistically significant in a forward–backward stepwise multiple regression analysis. Five of the 44 patients had foveal detachment and HM was noted in 13 of the 44 eyes. ERM was found in five of the six eyes (83%) with FovLs and in seven of the 38 eyes (18%) without FovLs (P = 0.0037, Fisher exact test). Five of the 12 eyes (42%) with ERM developed FovLs; however, only one of the 32 eyes (3%) without ERM developed FovLs (P = 0.013). Both HM and FovLs were significant independent risk factors for visual acuity deterioration at the 3rd month, but only FovLs were a significant risk factor for visual acuity deterioration at the 12th month. The authors noted paramacular thickness was highest in eyes with foveal detachment possibly mediated by fluid accumulation in the retina by means of increased choroidal effusions that lead to increased retinal thickness following hypotony. The authors suggest that this increase in retinal thickness could be responsible for lifting the Internal limiting membrane (ILM) leading to the foveal detachment. The damage brought about to the ellipsoid zone by the foveal detachment could persist even if the foveal detachment resolves and this could explain why postsurgical FovLs continue to act as a risk factor for visual acuity deterioration at the 12th month. Eighty-three percent of eyes that had FovLs had an ERM as well and imply that in addition to the traction on the retina brought about by the accumulation of intraretinal fluid in the context of hypotony, the presence of ERM could further exacerbate the traction on the retina and lead to foveal detachment. Choroidal folds may lead to changes in the orientation of photoreceptors (Stiles–Crawford effect) and affect VA. In diabetic maculopathy, the first sign of macular lesions is fluid accumulation in the inner retina and detachment of the photoreceptor occurs in an advanced stage of the disease. By contrast, detachment of the photoreceptor occurs in the early stage of hypotony eyes and tends to diminish over time. This suggests that the main cause of FovLs in hypotonia is not the accumulation of exudates but choroidal effusion and mechanical traction on the photoreceptor.
We must consider the limitations of the study design as we explore the translation of these findings to clinical practice. There were no formal sample size estimates reported and the sample size of the study is small (44 eyes). We cannot be sure the lack of significant associations is a true lack of association. This is a single-center study from a tertiary center and based on a single ethnic population. This limits the representativeness of the study, and we cannot be sure the findings may apply to Asian Indian Eyes. The retrospective study has limitations in terms of bias, standardization of measures, and missing data. The authors do not report on coexisting anterior segment factors that may contribute to visual acuity deterioration following trabeculectomy and if they were adjusted for in the regression analysis. The authors do not report the agreement on OCT reading.
My conclusions from this article are that the case series provides insights for a plausible cause of VA loss following trabeculectomy with mitomycin C, and the structural, functional, and anatomic basis for this cause of loss of visual acuity is biologically plausible. However, we need to study this further as a larger study in Asian Indian eyes before we translate this to routine clinical practice in India. This may need a multicentric study involving glaucoma practices to obtain a reasonable sample size and to optimize representativeness. An additional consideration in the translation to clinical practice is that translation to individual patient care may be optimized if we assess the diagnostic effectiveness of this sign (FovLs) for visual acuity loss and report likelihood ratios and posttest probabilities besides the measures of association as an odds ratio.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.