We use cookies to improve our service and to tailor our content and advertising to you.More infoYou can manage your cookie settings via your browser at any time. To learn more about how we use cookies, please see our cookies policy.
We appreciate the interest in our paper by Kumawat and Kumar, and the opportunity to address their comments. With regards to axial length, it is certainly known to be correlated with choroidal thickness. However, this information was not routinely obtained in our retina clinic and was not available for most patients in the study. We were able to account for spherical equivalent in our multivariate model, which may serve as a proxy for axial length. We agree that accounting for at least one of these variables is required for studies on choroidal thickness. We appreciate the Kumawat and Kumar’s suggestion to categorize patients based on stage of disease (proliferative vs nonproliferative). We had considered this approach, however the small number of patients with proliferative disease (only 4) made this less ideal from a statistical standpoint. Subretinal neovascularization was included in our univariate analysis but was not found to be statistically significant with regard to choroidal thickness, so it was not further considered in our multivariate assessments. Lastly, while inter-ocular asymmetry in choroidal thickness may exist, eye laterality was not found to be a significant variable affecting choroidal thickness in our univariate analysis (p = 0.87) and thus was also not included in multivariate models. We once again thank Dr. Kumawat and Dr. Kumar for their interest in our work.
We would like to thank Dr. Tarannum Mansoori for the interest in our study, “Intraocular pressure change after injection of intravitreal dexamethasone (Ozurdex) implant in Korean patients” and highlighting important issues about the intraocular pressure (IOP) measurement methods and the correlation of IOP with age.1
We used KT-800 Non-Contact Tonometer (Kowa, Tokyo, Japan) to measure IOP initially and rechecked with GAT if necessary. Unless the patient was diagnosed with glaucoma, NCT was initially used to measure both pre- and post-injection IOP.
As Dr. Tarannum Mansoori has pointed out, we also agree that GAT is the gold standard for IOP measurement. If the IOP measured with NCT was found to be high, it was always rechecked with GAT. GAT was used in 2 situations in our study. First, when the patients had a previous history of glaucoma, and second, when the patients’ IOP as measured with NCT was high (greater than 20). Thus, in cases of high IOP, the measurement involved NCT and was also always double checked with GAT. Moreover, multiple IOP measurements were obtained with GAT in cases of high IOP, and the average value of the measurements was regarded as the final IOP.
The range of age of the patients for injection of intravitreal dexamethasone was broad, from 16 to 88 years. We know that very young age (less than six years) or an older age are risk factors for steroid-induced glaucoma.2 However, regrettably, we have not performed any further analysis...
We would like to thank Dr. Tarannum Mansoori for the interest in our study, “Intraocular pressure change after injection of intravitreal dexamethasone (Ozurdex) implant in Korean patients” and highlighting important issues about the intraocular pressure (IOP) measurement methods and the correlation of IOP with age.1
We used KT-800 Non-Contact Tonometer (Kowa, Tokyo, Japan) to measure IOP initially and rechecked with GAT if necessary. Unless the patient was diagnosed with glaucoma, NCT was initially used to measure both pre- and post-injection IOP.
As Dr. Tarannum Mansoori has pointed out, we also agree that GAT is the gold standard for IOP measurement. If the IOP measured with NCT was found to be high, it was always rechecked with GAT. GAT was used in 2 situations in our study. First, when the patients had a previous history of glaucoma, and second, when the patients’ IOP as measured with NCT was high (greater than 20). Thus, in cases of high IOP, the measurement involved NCT and was also always double checked with GAT. Moreover, multiple IOP measurements were obtained with GAT in cases of high IOP, and the average value of the measurements was regarded as the final IOP.
The range of age of the patients for injection of intravitreal dexamethasone was broad, from 16 to 88 years. We know that very young age (less than six years) or an older age are risk factors for steroid-induced glaucoma.2 However, regrettably, we have not performed any further analysis regarding the correlation with age in our study. We plan to perform further analysis and look forward to sharing our results in the near future.
References
1 Choi W, Park SE, Kang HG, et al. Intraocular pressure change after injection of intravitreal dexamethasone (Ozurdex) implant in Korean patients. Br J Ophthalmol 2018.
2 Lam DS, Fan DS, Ng JS, et al. Ocular hypertensive and anti-inflammatory responses to different dosages of topical dexamethasone in children: a randomized trial. Clin Exp Ophthalmol 2005;33:252-8.
We thank the authors for their careful perusal of our study report and thoughtful observations. We agree that as demonstrated by the large population study1,2 referenced by us and by them, the rate of complications with cataract surgery is non-homogenous and increases dramatically with advanced stage cataracts – as much as 200%+ increase in rate of PCT in cases with high grade cataract, pseudoexfoliation and other comorbidities. In fact, with the co-existence of multiple factors, the compound rate can be even higher.
Our pilot study was in patients with advanced cataracts and multiple co-existing ocular pathologies and given the small sample size we are not surprised that the study point estimate for the PCT rate may be on the higher end of the overall range demonstrated by the larger population study. In addition, the randomized control design of the clinical trial further validates a PCT rate which was similar for both treatment and control groups. Certainly, an informed reader would appreciate that such a small trial is underpowered to be conclusive regarding the small difference between the two groups so no claims should be made about the slightly better rate of PCT and lower trend demonstrated in the miLOOP group.
What is important to appreciate from both the population study and our pilot data is that the rate of PCT is not the same for all cataract surgeries and there is a multiplier effect in certain subgroups and subpopulations. Our authorship team...
We thank the authors for their careful perusal of our study report and thoughtful observations. We agree that as demonstrated by the large population study1,2 referenced by us and by them, the rate of complications with cataract surgery is non-homogenous and increases dramatically with advanced stage cataracts – as much as 200%+ increase in rate of PCT in cases with high grade cataract, pseudoexfoliation and other comorbidities. In fact, with the co-existence of multiple factors, the compound rate can be even higher.
Our pilot study was in patients with advanced cataracts and multiple co-existing ocular pathologies and given the small sample size we are not surprised that the study point estimate for the PCT rate may be on the higher end of the overall range demonstrated by the larger population study. In addition, the randomized control design of the clinical trial further validates a PCT rate which was similar for both treatment and control groups. Certainly, an informed reader would appreciate that such a small trial is underpowered to be conclusive regarding the small difference between the two groups so no claims should be made about the slightly better rate of PCT and lower trend demonstrated in the miLOOP group.
What is important to appreciate from both the population study and our pilot data is that the rate of PCT is not the same for all cataract surgeries and there is a multiplier effect in certain subgroups and subpopulations. Our authorship team felt that for investigational transparency and authenticity it is important to report all results, including the ones we don’t like and the ones that give us pause in order to stimulate thoughtful discourse such as the one unfolding through the comments in your letter.
Sean Ianchulev, MD MPH
Professor of Ophthalmology
New York Eye and Ear Infirmary of Mount Sinai
References
1. Jaycock P, Johnston RL, Taylor H, et al. The Cataract National Dataset electronic multi-centre audit of 55,567 operations: updating benchmark standards of care in the United Kingdom and internationally. Eye 2009 23(1):38-49.
2. Narendran N, Jaycock P, Johnston RL, et al. The Cataract National Dataset electronic multicenter audit of 55,567 operations: risk stratification for posterior capsule rupture and vitreous loss. Eye 2009 23(1):31-7.
We are interested in the work of Ianchulev et al in their recent interventional randomized controlled trial.[1] What piqued our interest was the rate of posterior capsular tears (PCT). 4/53 (7.5%) patients in the miLOOP+phaco group experienced PCT, and 5/48 (10.4%) phaco-alone controls with PCT. These rates are much higher than standard phacoemulsification reports. The authors refer to a large study that identified advanced cataracts increased risk of PCT at comparable levels.[2] That same group published investigations expounding upon this.[3-4] Advanced cataracts were specifically identified as brunescent/white cataracts, contrasting Grade 3-4 in the miLOOP study (curiously described as LOCSIII classification in the manuscript).
Using the risk calculation,[3] the range of composite adjusted odds ratio (aOR) for the miLOOP study was 49.93 (25-28% risk) to an aOR of 0.87 (<1% surgical risk). The average patient from the miLOOP investigation had an aOR of 4.43, thus <5% PCT risk.
Our concerns: First, the authors state that “There was a trend towards a lower rate of capsular tear during the phaco portion with miLOOP-assisted phaco (7.5%) compared to standard phaco (10.4%).” Given the numbers representing these percentages this is an inappropriate description of this relationship.
Let us assume that a control group presented with a PCT rate similar to that reported in literature: <5%. Here, miLOOP-phaco PCT rate would be higher than the control...
We are interested in the work of Ianchulev et al in their recent interventional randomized controlled trial.[1] What piqued our interest was the rate of posterior capsular tears (PCT). 4/53 (7.5%) patients in the miLOOP+phaco group experienced PCT, and 5/48 (10.4%) phaco-alone controls with PCT. These rates are much higher than standard phacoemulsification reports. The authors refer to a large study that identified advanced cataracts increased risk of PCT at comparable levels.[2] That same group published investigations expounding upon this.[3-4] Advanced cataracts were specifically identified as brunescent/white cataracts, contrasting Grade 3-4 in the miLOOP study (curiously described as LOCSIII classification in the manuscript).
Using the risk calculation,[3] the range of composite adjusted odds ratio (aOR) for the miLOOP study was 49.93 (25-28% risk) to an aOR of 0.87 (<1% surgical risk). The average patient from the miLOOP investigation had an aOR of 4.43, thus <5% PCT risk.
Our concerns: First, the authors state that “There was a trend towards a lower rate of capsular tear during the phaco portion with miLOOP-assisted phaco (7.5%) compared to standard phaco (10.4%).” Given the numbers representing these percentages this is an inappropriate description of this relationship.
Let us assume that a control group presented with a PCT rate similar to that reported in literature: <5%. Here, miLOOP-phaco PCT rate would be higher than the control raising safety concerns. Yet the study showed a similarly high complication rate in the control group, statistically maintaining the miLOOP’s safety.
Finally, this occurs in the setting of considerable financial interest.
Regardless of the opportunities not otherwise afforded to these patients by their generous work, ethical standards and protection against bias is essential to protect patients. The complication rate was unacceptably high, and the trial should have been stopped and the study should not have been published.
References
1. Ianchulev T, Chang DF, Koo E, et al. Microinterventional endocapsular nucleus disassembly: novel technique and results of first-in-human randomized controlled study. Br J Ophthalmol 2019 103(2):176-180.
2. Jaycock P, Johnston RL, Taylor H, et al. The Cataract National Dataset electronic multi-centre audit of 55,567 operations: updating benchmark standards of care in the United Kingdom and internationally. Eye 2009 23(1):38-49.
3. Narendran N, Jaycock P, Johnston RL, et al. The Cataract National Dataset electronic multicenter audit of 55,567 operations: risk stratification for posterior capsule rupture and vitreous loss. Eye 2009 23(1):31-7.
4. Johnston RL, Taylor H, Smith R, et al. The Cataract National Dataset Electronic Multi-center audit of 55,567 operations: variation in posterior capsule rupture rates between surgeons. Eye 2010 24(5):888-93.
We appreciate Francisco-Javier Carrera-Hueso, Pedro Vazquez-Ferreiro, and Jaime Poquet-Jornet's careful reading of this paper. This commissioned review had a necessarily broad scope in order to summarize benefits and harms across three available therapies for the most common clinical indications. We agree there was quite a bit of information to present and that doing so in a succinct format is a challenge. However, we disagree with their contention that we did not follow current methodologic systematic review standards. We did indeed follow PRISMA reporting guidelines, as described in the Methods.
Regarding Table 1, the studies included in the summary table are the same as those described within the text and meta-analyses; we apologize for any confusion. In terms of format for the listing of studies in the meta-analyses, since studies are known primarily by their acronym, we used them in the figures whenever possible. The trials without specific names or acronyms were listed according to author and year.
Biswas 2011 only reported the percentage of patients gaining ≥15 letters at the 18-month endpoint, not at 12 months, so the study could not be included in the 12-month analysis for this outcome. The study did report mean change in BCVA at both endpoints, so it is included in both the 12 month and 18-24 month analyses in Figure 2. In terms of analyzing cost-effectiveness, only two trials meeting inclusion criteria (CATT and DRCR) di...
We appreciate Francisco-Javier Carrera-Hueso, Pedro Vazquez-Ferreiro, and Jaime Poquet-Jornet's careful reading of this paper. This commissioned review had a necessarily broad scope in order to summarize benefits and harms across three available therapies for the most common clinical indications. We agree there was quite a bit of information to present and that doing so in a succinct format is a challenge. However, we disagree with their contention that we did not follow current methodologic systematic review standards. We did indeed follow PRISMA reporting guidelines, as described in the Methods.
Regarding Table 1, the studies included in the summary table are the same as those described within the text and meta-analyses; we apologize for any confusion. In terms of format for the listing of studies in the meta-analyses, since studies are known primarily by their acronym, we used them in the figures whenever possible. The trials without specific names or acronyms were listed according to author and year.
Biswas 2011 only reported the percentage of patients gaining ≥15 letters at the 18-month endpoint, not at 12 months, so the study could not be included in the 12-month analysis for this outcome. The study did report mean change in BCVA at both endpoints, so it is included in both the 12 month and 18-24 month analyses in Figure 2. In terms of analyzing cost-effectiveness, only two trials meeting inclusion criteria (CATT and DRCR) directly compared cost outcomes between any of the 3 agents; both were included in our review.
Regarding the differing figures for mean difference calculated by the letter authors, we state in the Data Synthesis and Analysis portion of the text that “we used the mean difference (MD) between arms reported in the study whenever available; otherwise, mean between-group differences were calculated based on reported data.” For example, Kodjikian 2013 reported a mean difference of 2.36 (95% CI, -0.72 to 5.44) analysis for the GEFAL study, so that data was used in our analyses.
The use of quantitative estimation of publication bias is controversial and not without its own drawbacks (1). Publication bias is a type of reporting bias in which positive (often smaller) studies are preferentially published over negative studies. We examined the possibility of including unpublished research by searching, as mentioned in the manuscript, trial registries, and regulatory agency sites to identify all relevant registered and in-progress trials. We also requested unpublished data from drug manufacturers, but did not identify any additional trials.
We describe qualitatively the results contributed by each study according to its risk of bias. We found no indication of consistent differences in results according to ROB. Statistical measures also suggested minimal heterogeneity in the meta-analytic estimates presented.
We thank the authors of the letter for their interest in our article, and appreciate the opportunity to respond to their concerns.
Sincerely,
Allison Low
Devan Kansagara
References:
1. Guyatt GH, Oxman AD, Montori V, et al. GRADE guidelines: 5. Rating the quality of evidence—publication bias. J Clin Epidemiol. 2011;64(12):1277-1282.
We thank Dr. Bunce et al for their interest in our paper.1 We would like to apologize for not mentioning the Statistics Notes Series2-12 from the UK Ophthalmology Research Section of the NIHR Statistics group. Given that our paper’s purpose is to evaluate whether the correlated eye data were analyzed properly in published ophthalmic clinical science papers, we did not cite these papers because we think most of them serve as introductions of general statistical methods instead of specific statistical methods for correlated eye data.
We agree these Statistics Notes Series are very helpful to the vision research community to improve the statistical analysis and interpretation of ophthalmic data. We applaud the UK Ophthalmology Research Section of the NIHR Statistics group for their collaborative efforts in improving the quality of statistics for ophthalmic research through these series of publications and workshops. Similarly in the USA, we have been promoting the proper analysis of correlated eye data through tutorial papers13-14 and the ARVO short course. We believe all these efforts will lead to improvement in the statistical practice for ophthalmic data.
We also agree that there are varying degrees of misuse of statistical methods in analyzing correlated eye data. Ignoring the inter-eye correlation when data from both eyes are analyzed is very bad practice as it can lead to the invalid conclusion, while analyzing correlated ocular data at person-level does...
We thank Dr. Bunce et al for their interest in our paper.1 We would like to apologize for not mentioning the Statistics Notes Series2-12 from the UK Ophthalmology Research Section of the NIHR Statistics group. Given that our paper’s purpose is to evaluate whether the correlated eye data were analyzed properly in published ophthalmic clinical science papers, we did not cite these papers because we think most of them serve as introductions of general statistical methods instead of specific statistical methods for correlated eye data.
We agree these Statistics Notes Series are very helpful to the vision research community to improve the statistical analysis and interpretation of ophthalmic data. We applaud the UK Ophthalmology Research Section of the NIHR Statistics group for their collaborative efforts in improving the quality of statistics for ophthalmic research through these series of publications and workshops. Similarly in the USA, we have been promoting the proper analysis of correlated eye data through tutorial papers13-14 and the ARVO short course. We believe all these efforts will lead to improvement in the statistical practice for ophthalmic data.
We also agree that there are varying degrees of misuse of statistical methods in analyzing correlated eye data. Ignoring the inter-eye correlation when data from both eyes are analyzed is very bad practice as it can lead to the invalid conclusion, while analyzing correlated ocular data at person-level does not take full advantage of data and may lead to the inefficient conclusion.
We are saddened by the death of Professor Altman. We have learnt and benefited so much from the papers published by him. Professor Altman’s death is an immense loss to the applied statistics community.
References
1. Zhang HG, Ying GS. Statistical approaches in published ophthalmic clinical science papers: a comparison to statistical practice two decades ago. Br J Ophthalmol. 2018 Sep;102(9):1188-1191.
2. Stephenson J, Bunce C, Doré CJ, Freemantle N; Ophthalmic Statistics Group. Ophthalmic statistics note 11: logistic regression. Br J Ophthalmol. 2016 Dec;100(12):1594-1595.
3. Bunce C, Stephenson J, Doré CJ, Freemantle N; Ophthalmic Statistics Group.Ophthalmic statistics note 10: data transformations.Br J Ophthalmol. 2016 Dec;100(12):1591-1593.
4. Skene SS, Bunce C, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 9: parametric versus non-parametric methods for data analysis. Br J Ophthalmol. 2016 Jul;100(7):877-878.
5. Bunce C, Quartilho A, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 8: missing data--exploring the unknown. Br J Ophthalmol. 2016 Mar;100(3):291-4.
6. Cipriani V, Quartilho A, Bunce C, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 7: multiple hypothesis testing—to adjust or not to adjust. Br J Ophthalmol. 2015 Sep;99(9):1155-7.
7. Cook JA, Bunce C, Doré CJ, Freemantle N; Ophthalmic Statistics Group. Ophthalmic statistics note 6: effect sizes matter. Br J Ophthalmol. 2015 May;99(5):580-1.
8. Saunders LJ, Zhu H, Bunce C, Doré CJ, Freemantle N, Crabb DP; Ophthalmic Statistics Group. Ophthalmic statistics note 5: diagnostic tests—sensitivity and specificity. Br J Ophthalmol. 2015 Sep;99(9):1168-70.
9. Nash R, Bunce C, Freemantle N, Doré CJ, Rogers CA; Ophthalmic Statistics Group.
Ophthalmic Statistics Note 4: analysing data from randomised controlled trials with baseline and follow-up measurements. Br J Ophthalmol. 2014 Nov;98(11):1467-9.
10. Ophthalmic statistics note: the perils of dichotomising continuous variables. Cumberland PM, Czanner G, Bunce C, Doré CJ, Freemantle N, García-Fiñana M; Ophthalmic Statistics Group.Br J Ophthalmol. 2014 Jun;98(6):841-3.
11. Bunce C, Patel KV, Xing W, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 2: absence of evidence is not evidence of absence. Br J Ophthalmol. 2014 May;98(5):703-5.
12. Bunce C, Patel KV, Xing W, Freemantle N, Doré CJ; Ophthalmic Statistics Group Ophthalmic statistics note 1: unit of analysis. Br J Ophthalmol. 2014 Mar;98(3):408-12.
13. Ying GS, Maguire MG, Glynn R, Rosner B. Tutorial on Biostatistics: Linear regression analysis of continuous correlated eye data. Ophthalmic Epidemiology 2017;24:130-140.
Ying GS, Maguire MG, Glynn R, Rosner B. Turtorial on biostatistics: Statistical analysis for correlated binary eye data. Ophthalmic Epidemiology 2018;25:1-12.
We thank Drs. Robert and Vidal for their comments. After carefully reading their original series of 5 patients and observing their videos [1], our impression was that both series could definitely refer to the same unique phenomenon.
As stated before [2], a drawback of our series was the inability to acquire eye movement recordings for any of our patients due to technical obstacles and parental refusal. We found Robert and Vidal’s ability to do so in one of their patients very important to the understanding and definition of the phenomenon [1]. Clearly their recordings demonstrate an upbeating nystagmus that would be expected in patients with tonic downgaze, assuming the eyes drift down while saccadic correcting movements are upward towards primary gaze. Hopefully, additional supporting recordings will be added to the literature in the future, allowing us to conclude that this is a representing finding for all of these patients.
This condition was apparently described under different titles over the years owing to scarce descriptions in the literature and difficulty providing convincing support for one definition over the other. This is an important step in that direction. We agree that with their addition of data, the term should include “upbeat nystagmus” and therefore suggest the term “benign infantile positional tonic downgaze with upbeat nystagmus”.
1. Robert MP, Michel S, Adjadj E, Boddaert N, Desguerre I, Vidal PP. Benign intermittent upbeat nystag...
We thank Drs. Robert and Vidal for their comments. After carefully reading their original series of 5 patients and observing their videos [1], our impression was that both series could definitely refer to the same unique phenomenon.
As stated before [2], a drawback of our series was the inability to acquire eye movement recordings for any of our patients due to technical obstacles and parental refusal. We found Robert and Vidal’s ability to do so in one of their patients very important to the understanding and definition of the phenomenon [1]. Clearly their recordings demonstrate an upbeating nystagmus that would be expected in patients with tonic downgaze, assuming the eyes drift down while saccadic correcting movements are upward towards primary gaze. Hopefully, additional supporting recordings will be added to the literature in the future, allowing us to conclude that this is a representing finding for all of these patients.
This condition was apparently described under different titles over the years owing to scarce descriptions in the literature and difficulty providing convincing support for one definition over the other. This is an important step in that direction. We agree that with their addition of data, the term should include “upbeat nystagmus” and therefore suggest the term “benign infantile positional tonic downgaze with upbeat nystagmus”.
1. Robert MP, Michel S, Adjadj E, Boddaert N, Desguerre I, Vidal PP. Benign intermittent upbeat nystagmus in infancy: a new clinical entity. Eur J Paediatr Neurol. 2015; 19: 262-5.
2. Sternfeld A, Lobel D, Leiba H, Luckman J, Michowiz S, Goldenberg-Cohen N. Long-term follow-up of benign positional vertical opsoclonus in infants: retrospective cohort. Br J Ophthalmol. 2018; 102: 757-60.
We read with great interest the nice series from Sternfeld et al. about so-called “benign positional vertical opsoclonus in infants”. [1] As stated by the authors, the very specific condition they describe is not uncommon in the population, yet still poorly described in the scientific literature. Additionally, it is called differently by different authors, one reason for it being the difficulty to assess through the naked eye the very nature of the high frequency eye movements, as shown in video n°1.
The condition combines a positional tonic downgaze and abnormal vertical eye movements. Oculomotor recordings of infants presenting with this clinical picture do actually confirm that these movements comprise downbeating slow phases and upbeating saccades 2, as clinically seen in video n°2, and are therefore a vertical kind of nystagmus. We therefore proposed to refer to them as benign intermittent upbeat nystagmus in infancy. [2] As stated by the authors, the association of a tonic downgaze and an upbeat nystagmus is logical and has been related to posterior semicircular canal predominance. [3]
In addition to the evidence of slow phases in this condition and to the fact that, to the best of our knowledge, no pulse of purely vertical saccades has ever been recorded, the very possibility for the oculomotor system to produce such movements is questionable. Opsoclonus, by definition, designates a succession of multidirectional saccades. We therefo...
We read with great interest the nice series from Sternfeld et al. about so-called “benign positional vertical opsoclonus in infants”. [1] As stated by the authors, the very specific condition they describe is not uncommon in the population, yet still poorly described in the scientific literature. Additionally, it is called differently by different authors, one reason for it being the difficulty to assess through the naked eye the very nature of the high frequency eye movements, as shown in video n°1.
The condition combines a positional tonic downgaze and abnormal vertical eye movements. Oculomotor recordings of infants presenting with this clinical picture do actually confirm that these movements comprise downbeating slow phases and upbeating saccades 2, as clinically seen in video n°2, and are therefore a vertical kind of nystagmus. We therefore proposed to refer to them as benign intermittent upbeat nystagmus in infancy. [2] As stated by the authors, the association of a tonic downgaze and an upbeat nystagmus is logical and has been related to posterior semicircular canal predominance. [3]
In addition to the evidence of slow phases in this condition and to the fact that, to the best of our knowledge, no pulse of purely vertical saccades has ever been recorded, the very possibility for the oculomotor system to produce such movements is questionable. Opsoclonus, by definition, designates a succession of multidirectional saccades. We therefore believe that the term “vertical opsoclonus” should be avoided.
1. Sternfeld A, Lobel D, Leiba H, Luckman J, Michowiz S, Goldenberg-Cohen N. Long-term follow-up of benign positional vertical opsoclonus in infants: retrospective cohort. Br J Ophthalmol. 2018; 102: 757-60.
2. Robert MP, Michel S, Adjadj E, Boddaert N, Desguerre I, Vidal PP. Benign intermittent upbeat nystagmus in infancy: a new clinical entity. Eur J Paediatr Neurol. 2015; 19: 262-5.
3. Brodsky MC, Donahue SP. Primary oblique muscle overaction: the brain throws a wild pitch. Arch Ophthalmol. 2001; 119: 1307-14.
We read with great interest the recent paper by Zhang and Ying exploring statistical approaches in published ophthalmic clinical science papers.1 We very much agree with the main conclusion drawn by the authors that collaborative efforts should be made in the vision research community to improve statistical practise for ocular data. In this vein, however, we were disappointed not to see reference to the Statistics Notes Series that has been published in this very journal. These have been written with a view to tackling some of the more prevalent statistical issues within ophthalmology and we would encourage readers to make use of these.2- 12. Within the UK this view that there needs to be greater collaboration in the vision research community has led to the formation of the Ophthalmology Research Section of the NIHR Statistics group which is championing cross- professional collaboration and active discussion in relation to statistical issues. It is always important when reviewing misuse of statistics in biomedical research to distinguish between misuse that leads to distorted or incorrect results and those methods which do not fully use data to maximum potential given that this loss of information might be viewed as unethical. In this regard we find the results from Zhang et al pleasing in that the proportion of papers which analysed at the level of the individual because of the nature of the observation rose from 15.2 % in 1995 to 50 % in 2017. A finding which is...
We read with great interest the recent paper by Zhang and Ying exploring statistical approaches in published ophthalmic clinical science papers.1 We very much agree with the main conclusion drawn by the authors that collaborative efforts should be made in the vision research community to improve statistical practise for ocular data. In this vein, however, we were disappointed not to see reference to the Statistics Notes Series that has been published in this very journal. These have been written with a view to tackling some of the more prevalent statistical issues within ophthalmology and we would encourage readers to make use of these.2- 12. Within the UK this view that there needs to be greater collaboration in the vision research community has led to the formation of the Ophthalmology Research Section of the NIHR Statistics group which is championing cross- professional collaboration and active discussion in relation to statistical issues. It is always important when reviewing misuse of statistics in biomedical research to distinguish between misuse that leads to distorted or incorrect results and those methods which do not fully use data to maximum potential given that this loss of information might be viewed as unethical. In this regard we find the results from Zhang et al pleasing in that the proportion of papers which analysed at the level of the individual because of the nature of the observation rose from 15.2 % in 1995 to 50 % in 2017. A finding which is statistically significant (P< 0.001, two-tailed Chi-square test). This is pertinent in that historically single person-level responses were often created by ophthalmologists (taking the worse eye only for example or combining information from each eye) prior to the development of the more advanced techniques that make better use of the data (by using data from both eyes instead of one) such as generalised estimating equations. A review by Gange et al has shown that point estimates and tests of significance that used this simple person-level approach are similar to those obtained from the more complex methods and concluded that there would therefore be unlikely benefit from a reanalysis with more sophisticated methods 13 . Misuse of statistics in medical research is unethical and this is a message that has been strongly advocated by Professor Doug Altman since 1980. 14 Professor Altman died in June of 2018 and his loss is hugely felt by the applied statistics community. Let us make 2019 a year of improved statistics throughout all research in vision and eyes.
1 Zhang HG, Ying GS. Statistical approaches in published ophthalmic clinical science papers: a comparison to statistical practice two decades ago. Br J Ophthalmol. 2018 Sep;102(9):1188-1191.
2 Stephenson J, Bunce C, Doré CJ, Freemantle N; Ophthalmic Statistics Group. Ophthalmic statistics note 11: logistic regression. Br J Ophthalmol. 2016 Dec;100(12):1594-1595.
3 Bunce C, Stephenson J, Doré CJ, Freemantle N; Ophthalmic Statistics Group.Ophthalmic statistics note 10: data transformations.Br J Ophthalmol. 2016 Dec;100(12):1591-1593.
4 Skene SS, Bunce C, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 9: parametric versus non-parametric methods for data analysis. Br J Ophthalmol. 2016 Jul;100(7):877-878.
5 Bunce C, Quartilho A, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 8: missing data--exploring the unknown. Br J Ophthalmol. 2016 Mar;100(3):291-4.
6 Cipriani V, Quartilho A, Bunce C, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 7: multiple hypothesis testing—to adjust or not to adjust. Br J Ophthalmol. 2015 Sep;99(9):1155-7.
7 Cook JA, Bunce C, Doré CJ, Freemantle N; Ophthalmic Statistics Group. Ophthalmic statistics note 6: effect sizes matter. Br J Ophthalmol. 2015 May;99(5):580-1.
8 Saunders LJ, Zhu H, Bunce C, Doré CJ, Freemantle N, Crabb DP; Ophthalmic Statistics Group. Ophthalmic statistics note 5: diagnostic tests—sensitivity and specificity. Br J Ophthalmol. 2015 Sep;99(9):1168-70.
9 Nash R, Bunce C, Freemantle N, Doré CJ, Rogers CA; Ophthalmic Statistics Group.
Ophthalmic Statistics Note 4: analysing data from randomised controlled trials with baseline and follow-up measurements. Br J Ophthalmol. 2014 Nov;98(11):1467-9.
10 Ophthalmic statistics note: the perils of dichotomising continuous variables. Cumberland PM, Czanner G, Bunce C, Doré CJ, Freemantle N, García-Fiñana M; Ophthalmic Statistics Group.Br J Ophthalmol. 2014 Jun;98(6):841-3.
11 Bunce C, Patel KV, Xing W, Freemantle N, Doré CJ; Ophthalmic Statistics Group. Ophthalmic statistics note 2: absence of evidence is not evidence of absence. Br J Ophthalmol. 2014 May;98(5):703-5.
12 Bunce C, Patel KV, Xing W, Freemantle N, Doré CJ; Ophthalmic Statistics Group Ophthalmic statistics note 1: unit of analysis. Br J Ophthalmol. 2014 Mar;98(3):408-12.
13 Gange SJ, Linton KL, Scott AJ, DeMets DL, Klein R. A comparison of methods for correlated ordinal measures with ophthalmic applications. Stat Med. 1995 Sep 30;14(18):1961-74
14 Altman DG. Statistics and ethics in medical research. Misuse of statistics is unethical. Br Med J. 1980 Nov 1;281(6249):1182-4.
We read the recent article in the journal by Monfermé and coauthors (1) on phenotypic associations of TYR R402Q compound heterozygosity with keen interest. Given our own and others previous findings (2-4) that the R402Q-S192Y haplotype seemed to have a stronger biological effect, we were surprised that a triallelic effect was not supported in the clinical discussion by Monfermé et al. When we examined these results more closely, we noted that their sample included 52 S192Y variant allele carriers out of the entire collection of 69 patients, of whom 31 (44.9%) carried the R402Q-S192Y double variant haplotype. In our own collections from the Australian general population (BNMS and BLTS, Duffy et al., in submission), only 6% of R402Q carriers also carried S192Y in cis. When we examined the European 1000 Genomes subsamples, there too only 7% of R402Q haplotypes had the 192Y rather than 192S wild type allele. So, even though Monfermé et al could not detect a statistically significant additive effect on OCA trait severity due to the S192Y polymorphism within their sample, it is clear that the S192 variant allele is markedly overrepresented compared to the general population, suggesting that it must have some relationship to albinism. This genetic association of the TYR R402Q-S192Y double variant haplotype with OCA1 is now clearly causative (2, 3), and explains some of the missing heritability that has previously been seen in OCA patients (4).
We read the recent article in the journal by Monfermé and coauthors (1) on phenotypic associations of TYR R402Q compound heterozygosity with keen interest. Given our own and others previous findings (2-4) that the R402Q-S192Y haplotype seemed to have a stronger biological effect, we were surprised that a triallelic effect was not supported in the clinical discussion by Monfermé et al. When we examined these results more closely, we noted that their sample included 52 S192Y variant allele carriers out of the entire collection of 69 patients, of whom 31 (44.9%) carried the R402Q-S192Y double variant haplotype. In our own collections from the Australian general population (BNMS and BLTS, Duffy et al., in submission), only 6% of R402Q carriers also carried S192Y in cis. When we examined the European 1000 Genomes subsamples, there too only 7% of R402Q haplotypes had the 192Y rather than 192S wild type allele. So, even though Monfermé et al could not detect a statistically significant additive effect on OCA trait severity due to the S192Y polymorphism within their sample, it is clear that the S192 variant allele is markedly overrepresented compared to the general population, suggesting that it must have some relationship to albinism. This genetic association of the TYR R402Q-S192Y double variant haplotype with OCA1 is now clearly causative (2, 3), and explains some of the missing heritability that has previously been seen in OCA patients (4).
1 Dermatology Research Centre, The University of Queensland, The University of Queensland Diamantina Institute, Brisbane, Qld, Australia; 2 QIMR Berghofer Medical Research Institute, Brisbane, Australia
References
1. Monferme S, Lasseaux E, Duncombe-Poulet C, et al. Mild form of oculocutaneous albinism type 1: phenotypic analysis of compound heterozygous patients with the R402Q variant of the TYR gene. Br J Ophthalmol 2018 doi: 10.1136/bjophthalmol-2018-312729
2. Jagirdar K, Smit DJ, Ainger SA, et al. Molecular analysis of common polymorphisms within the human Tyrosinase locus and genetic association with pigmentation traits. Pigment cell & melanoma research 2014;27(4):552-64. doi: 10.1111/pcmr.12253
3. Norman CS, O'Gorman L, Gibson J, et al. Identification of a functionally significant tri-allelic genotype in the Tyrosinase gene (TYR) causing hypomorphic oculocutaneous albinism (OCA1B). Sci Rep 2017;7(1):4415. doi: 10.1038/s41598-017-04401-5
4. Gronskov K, Jespersgaard C, Bruun GH, et al. A pathogenic haplotype, common in Europeans, causes autosomal recessive albinism and uncovers missing heritability in OCA1. Sci Rep 2019;9(1):645. doi: 10.1038/s41598-018-37272-5
Dear Editor,
We appreciate the interest in our paper by Kumawat and Kumar, and the opportunity to address their comments. With regards to axial length, it is certainly known to be correlated with choroidal thickness. However, this information was not routinely obtained in our retina clinic and was not available for most patients in the study. We were able to account for spherical equivalent in our multivariate model, which may serve as a proxy for axial length. We agree that accounting for at least one of these variables is required for studies on choroidal thickness. We appreciate the Kumawat and Kumar’s suggestion to categorize patients based on stage of disease (proliferative vs nonproliferative). We had considered this approach, however the small number of patients with proliferative disease (only 4) made this less ideal from a statistical standpoint. Subretinal neovascularization was included in our univariate analysis but was not found to be statistically significant with regard to choroidal thickness, so it was not further considered in our multivariate assessments. Lastly, while inter-ocular asymmetry in choroidal thickness may exist, eye laterality was not found to be a significant variable affecting choroidal thickness in our univariate analysis (p = 0.87) and thus was also not included in multivariate models. We once again thank Dr. Kumawat and Dr. Kumar for their interest in our work.
We would like to thank Dr. Tarannum Mansoori for the interest in our study, “Intraocular pressure change after injection of intravitreal dexamethasone (Ozurdex) implant in Korean patients” and highlighting important issues about the intraocular pressure (IOP) measurement methods and the correlation of IOP with age.1
Show MoreWe used KT-800 Non-Contact Tonometer (Kowa, Tokyo, Japan) to measure IOP initially and rechecked with GAT if necessary. Unless the patient was diagnosed with glaucoma, NCT was initially used to measure both pre- and post-injection IOP.
As Dr. Tarannum Mansoori has pointed out, we also agree that GAT is the gold standard for IOP measurement. If the IOP measured with NCT was found to be high, it was always rechecked with GAT. GAT was used in 2 situations in our study. First, when the patients had a previous history of glaucoma, and second, when the patients’ IOP as measured with NCT was high (greater than 20). Thus, in cases of high IOP, the measurement involved NCT and was also always double checked with GAT. Moreover, multiple IOP measurements were obtained with GAT in cases of high IOP, and the average value of the measurements was regarded as the final IOP.
The range of age of the patients for injection of intravitreal dexamethasone was broad, from 16 to 88 years. We know that very young age (less than six years) or an older age are risk factors for steroid-induced glaucoma.2 However, regrettably, we have not performed any further analysis...
We thank the authors for their careful perusal of our study report and thoughtful observations. We agree that as demonstrated by the large population study1,2 referenced by us and by them, the rate of complications with cataract surgery is non-homogenous and increases dramatically with advanced stage cataracts – as much as 200%+ increase in rate of PCT in cases with high grade cataract, pseudoexfoliation and other comorbidities. In fact, with the co-existence of multiple factors, the compound rate can be even higher.
Our pilot study was in patients with advanced cataracts and multiple co-existing ocular pathologies and given the small sample size we are not surprised that the study point estimate for the PCT rate may be on the higher end of the overall range demonstrated by the larger population study. In addition, the randomized control design of the clinical trial further validates a PCT rate which was similar for both treatment and control groups. Certainly, an informed reader would appreciate that such a small trial is underpowered to be conclusive regarding the small difference between the two groups so no claims should be made about the slightly better rate of PCT and lower trend demonstrated in the miLOOP group.
What is important to appreciate from both the population study and our pilot data is that the rate of PCT is not the same for all cataract surgeries and there is a multiplier effect in certain subgroups and subpopulations. Our authorship team...
Show MoreWe are interested in the work of Ianchulev et al in their recent interventional randomized controlled trial.[1] What piqued our interest was the rate of posterior capsular tears (PCT). 4/53 (7.5%) patients in the miLOOP+phaco group experienced PCT, and 5/48 (10.4%) phaco-alone controls with PCT. These rates are much higher than standard phacoemulsification reports. The authors refer to a large study that identified advanced cataracts increased risk of PCT at comparable levels.[2] That same group published investigations expounding upon this.[3-4] Advanced cataracts were specifically identified as brunescent/white cataracts, contrasting Grade 3-4 in the miLOOP study (curiously described as LOCSIII classification in the manuscript).
Using the risk calculation,[3] the range of composite adjusted odds ratio (aOR) for the miLOOP study was 49.93 (25-28% risk) to an aOR of 0.87 (<1% surgical risk). The average patient from the miLOOP investigation had an aOR of 4.43, thus <5% PCT risk.
Our concerns: First, the authors state that “There was a trend towards a lower rate of capsular tear during the phaco portion with miLOOP-assisted phaco (7.5%) compared to standard phaco (10.4%).” Given the numbers representing these percentages this is an inappropriate description of this relationship.
Let us assume that a control group presented with a PCT rate similar to that reported in literature: <5%. Here, miLOOP-phaco PCT rate would be higher than the control...
Show MoreTo the Editor,
We appreciate Francisco-Javier Carrera-Hueso, Pedro Vazquez-Ferreiro, and Jaime Poquet-Jornet's careful reading of this paper. This commissioned review had a necessarily broad scope in order to summarize benefits and harms across three available therapies for the most common clinical indications. We agree there was quite a bit of information to present and that doing so in a succinct format is a challenge. However, we disagree with their contention that we did not follow current methodologic systematic review standards. We did indeed follow PRISMA reporting guidelines, as described in the Methods.
Regarding Table 1, the studies included in the summary table are the same as those described within the text and meta-analyses; we apologize for any confusion. In terms of format for the listing of studies in the meta-analyses, since studies are known primarily by their acronym, we used them in the figures whenever possible. The trials without specific names or acronyms were listed according to author and year.
Biswas 2011 only reported the percentage of patients gaining ≥15 letters at the 18-month endpoint, not at 12 months, so the study could not be included in the 12-month analysis for this outcome. The study did report mean change in BCVA at both endpoints, so it is included in both the 12 month and 18-24 month analyses in Figure 2. In terms of analyzing cost-effectiveness, only two trials meeting inclusion criteria (CATT and DRCR) di...
Show MoreWe thank Dr. Bunce et al for their interest in our paper.1 We would like to apologize for not mentioning the Statistics Notes Series2-12 from the UK Ophthalmology Research Section of the NIHR Statistics group. Given that our paper’s purpose is to evaluate whether the correlated eye data were analyzed properly in published ophthalmic clinical science papers, we did not cite these papers because we think most of them serve as introductions of general statistical methods instead of specific statistical methods for correlated eye data.
We agree these Statistics Notes Series are very helpful to the vision research community to improve the statistical analysis and interpretation of ophthalmic data. We applaud the UK Ophthalmology Research Section of the NIHR Statistics group for their collaborative efforts in improving the quality of statistics for ophthalmic research through these series of publications and workshops. Similarly in the USA, we have been promoting the proper analysis of correlated eye data through tutorial papers13-14 and the ARVO short course. We believe all these efforts will lead to improvement in the statistical practice for ophthalmic data.
We also agree that there are varying degrees of misuse of statistical methods in analyzing correlated eye data. Ignoring the inter-eye correlation when data from both eyes are analyzed is very bad practice as it can lead to the invalid conclusion, while analyzing correlated ocular data at person-level does...
Show MoreWe thank Drs. Robert and Vidal for their comments. After carefully reading their original series of 5 patients and observing their videos [1], our impression was that both series could definitely refer to the same unique phenomenon.
As stated before [2], a drawback of our series was the inability to acquire eye movement recordings for any of our patients due to technical obstacles and parental refusal. We found Robert and Vidal’s ability to do so in one of their patients very important to the understanding and definition of the phenomenon [1]. Clearly their recordings demonstrate an upbeating nystagmus that would be expected in patients with tonic downgaze, assuming the eyes drift down while saccadic correcting movements are upward towards primary gaze. Hopefully, additional supporting recordings will be added to the literature in the future, allowing us to conclude that this is a representing finding for all of these patients.
This condition was apparently described under different titles over the years owing to scarce descriptions in the literature and difficulty providing convincing support for one definition over the other. This is an important step in that direction. We agree that with their addition of data, the term should include “upbeat nystagmus” and therefore suggest the term “benign infantile positional tonic downgaze with upbeat nystagmus”.
1. Robert MP, Michel S, Adjadj E, Boddaert N, Desguerre I, Vidal PP. Benign intermittent upbeat nystag...
Show MoreDear Editor,
We read with great interest the nice series from Sternfeld et al. about so-called “benign positional vertical opsoclonus in infants”. [1] As stated by the authors, the very specific condition they describe is not uncommon in the population, yet still poorly described in the scientific literature. Additionally, it is called differently by different authors, one reason for it being the difficulty to assess through the naked eye the very nature of the high frequency eye movements, as shown in video n°1.
Show MoreThe condition combines a positional tonic downgaze and abnormal vertical eye movements. Oculomotor recordings of infants presenting with this clinical picture do actually confirm that these movements comprise downbeating slow phases and upbeating saccades 2, as clinically seen in video n°2, and are therefore a vertical kind of nystagmus. We therefore proposed to refer to them as benign intermittent upbeat nystagmus in infancy. [2] As stated by the authors, the association of a tonic downgaze and an upbeat nystagmus is logical and has been related to posterior semicircular canal predominance. [3]
In addition to the evidence of slow phases in this condition and to the fact that, to the best of our knowledge, no pulse of purely vertical saccades has ever been recorded, the very possibility for the oculomotor system to produce such movements is questionable. Opsoclonus, by definition, designates a succession of multidirectional saccades. We therefo...
We read with great interest the recent paper by Zhang and Ying exploring statistical approaches in published ophthalmic clinical science papers.1 We very much agree with the main conclusion drawn by the authors that collaborative efforts should be made in the vision research community to improve statistical practise for ocular data. In this vein, however, we were disappointed not to see reference to the Statistics Notes Series that has been published in this very journal. These have been written with a view to tackling some of the more prevalent statistical issues within ophthalmology and we would encourage readers to make use of these.2- 12. Within the UK this view that there needs to be greater collaboration in the vision research community has led to the formation of the Ophthalmology Research Section of the NIHR Statistics group which is championing cross- professional collaboration and active discussion in relation to statistical issues. It is always important when reviewing misuse of statistics in biomedical research to distinguish between misuse that leads to distorted or incorrect results and those methods which do not fully use data to maximum potential given that this loss of information might be viewed as unethical. In this regard we find the results from Zhang et al pleasing in that the proportion of papers which analysed at the level of the individual because of the nature of the observation rose from 15.2 % in 1995 to 50 % in 2017. A finding which is...
Show MoreTo the Editor:
We read the recent article in the journal by Monfermé and coauthors (1) on phenotypic associations of TYR R402Q compound heterozygosity with keen interest. Given our own and others previous findings (2-4) that the R402Q-S192Y haplotype seemed to have a stronger biological effect, we were surprised that a triallelic effect was not supported in the clinical discussion by Monfermé et al. When we examined these results more closely, we noted that their sample included 52 S192Y variant allele carriers out of the entire collection of 69 patients, of whom 31 (44.9%) carried the R402Q-S192Y double variant haplotype. In our own collections from the Australian general population (BNMS and BLTS, Duffy et al., in submission), only 6% of R402Q carriers also carried S192Y in cis. When we examined the European 1000 Genomes subsamples, there too only 7% of R402Q haplotypes had the 192Y rather than 192S wild type allele. So, even though Monfermé et al could not detect a statistically significant additive effect on OCA trait severity due to the S192Y polymorphism within their sample, it is clear that the S192 variant allele is markedly overrepresented compared to the general population, suggesting that it must have some relationship to albinism. This genetic association of the TYR R402Q-S192Y double variant haplotype with OCA1 is now clearly causative (2, 3), and explains some of the missing heritability that has previously been seen in OCA patients (4).
Yours,...
Show MorePages