Dear Editor,

We thank Dr Arnolds and Ms Matta for their letter regarding our article Vision screening in children by Plusoptix Vision ScreenerTM compared with gold standard orthoptic assessment Dahlmann-Noor et al. (19 November 2008). We agree that a child vision screening tool that could be used effectively with minimal input from healthcare personnel would be desirable. This, together with the National Screening Committee ( UK) directive that all children should undergo vision screening between the age of 4-5 years stimulated our interest in the Plusoptix Vision Screener.

Contrary to their interpretation of Table 3 in our article, the Plusoptix is not able to detect most refractive errors. This study was a vision screening investigation, with a study population of mostly emmetropic children. When the prevalence of refractive errors is low, the Plusoptix may appear to perform well. However, it fails to identify children with significant hypermetropia even when it causes a reduction in visual acuity, probably because of device-induced fixation myopia.

Whilst the present study investigated the performance of the Plusoptix as a vision screening tool, its refractive performance is subject of a different study we conducted (Plusoptix Vision ScreenerTM: The accuracy and repeatability of refractive measurements using a new autorefractor. Dahlmann-Noor AH, Comyn O, Kostakis V, Misra A, Gupta N, Heath J, Brown J, Iron A, McGill S, Vrotsou K, Vivian AJ. Br J Ophthalmol. 2008 Nov 10. [Epub ahead of print]). This article is currently available online and will appear in print soon. In this second study, all subjects underwent cycloplegic refraction allowing for a meaningful assessment of the refractive performance of the Plusoptix. Many of Dr Arnold and Ms Matta’s points occurred to us as well, and are addressed by this second study. The idea that changing the cut-off criteria could improve sensitivity is sensible but when we did re-analyse data with different cutoff values (as explained in the discussion) this resulted in a lowering of specificity with very modest improvement of sensitivity as would be expected if the under-estimation of hypermetropia is caused by accommodation by the child on the machine at 1 metre.

Whereas the prevalence of amblyopia in most populations is between 2- 4%, the combined prevalence of amblyopia. strabismus and significant refractive errors reducing visual acuity is considerably greater. We report an orthoptic referral rate of 12.5% in our study with only two false positives in this group. The limitation of an orthoptist-only screening program is that refraction is not part of the screening, hence the desire to find a sensitive and specific refraction tool.

In our view, the Plusopix Vision Screener has many plus points, but until the problem with fixation-induced myopia is overcome, its lack of sensitivity limits its use as a stand-alone child vision screening device.

Again, we wish to thank Dr Arnolds and Ms Matta for their interest and hope that the answers we have provided clarify the concerns they have raised.

AJ Vivian and A. H. Dahlmann-Noor

Dear editor,

We would like to give a response to the letter of Dr. Anshuman Sinha concerning our article "23 Gauge versus 20 Gauge System for Pars Plana Vitrectomy: A Prospective Randomized Clinical Trial".

For analysis of postoperative pain we used a simplified verbal rating scale (ranging from 0, meaning no pain, to 3, meaning severe pain). The standardized verbal rating scale reaches from 0 (no pain) to 10 ("the worst pain possible"), which we think is too extensive for our patients and purpose. In literature verbal rating scales are described to be a valid, reliable and appropriate tool for determination of postoperative pain in clinical practice [1,2] Furthermore we believe that the assessment of postoperative pain is not likely to be biased if the assessor is not masked, because the perception of pain is a completely subjective parameter.

Regarding the sample size evaluation: The effect size was assumed to be 75% of the standard deviation and the power to be 80%. These assumptions were extrapolated from the observations of our previous study [3]. Unfortunately, this information had been dropped from a previous version of the manuscript for brevity.

It is well known that studies with single surgeons have better results than with multiple surgeons. On the other hand, it is recommended to choose a realistic situation to achieve more realistic results. Our surgeons were randomized and there was a general surgical protocol to be followed. This also is the reason why we used two different vitrectomy machines. Both of them work in a comparable way and the settings were preset and the same in both machines. We agree, that the lens status has an influence on the amount of vitreous removed and that an intraocular tamponade influences postoperative eye pressure. However, both factors were equally distributed in both surgical groups and therefore have no influence on the results. Concerning the way of suturing the 20 gauge wounds, this was done the same way in all patients using vicryl 7.0 sutures as described in the methods section.

A summary of nominal categorical data would have to be done by contingency tables and neither by mean nor by median and interquartile range. Since the data mentioned are not of nominal but ordinal nature, we assume that the author of this letter refers to ordinal data. We have to note that the summary presented in table 2 was done for the area under the curve (AUC) for the conjunctival injection and postoperative pain measured at days 0, 1, 2, 3, one week, one month and three months after surgery as indicated in figures 1 and 2. Although theses measurements are, strictly speaking, still of ordinal nature they have a large range of values (see table 2) and hence can be treated as metric variables. Moreover, it is common practice to summarize AUCs by mean and standard deviation. We also note that in a previous version of the paper we have given in addition to the mean and standard deviation also the median, first and third quartile. Unfortunately, these numbers as well as the information in the text that the AUCs are summarized have been omitted in the edited manuscript.

We furthermore agree that each complication represents a serious situation. However, from the data of our study we can not conclude that a clear indication for higher risk is given using 23 gauge vitrectomy. This question remains open and a larger prospective randomized study is needed to get a clear result of the safety of 23 gauge surgery. Since the data do not indicate a difference in the development of serious complications between the two treatment groups we made the cited careful assertion "seems to be a safe system".

Regarding the Consolidated Standards for Reporting Trials (CONSORT) statement [4], we certainly know and appreciate this publication and followed these recommendations. Moreover, our article represents the only prospective randomized study concerning this topic. We do appreciate your considerations, but we strongly believe that we have performed a well-designed study. Still, there are always things that can be improved.

References

1.Williamson A, Hoggart B. Pain: a review of three commonly used pain rating scales. J Clin Nurs. 2005 Aug;14(7):798-804.

2.Randall C. Cork, Ihab Isaac, Ahmad Elsharydah, Sarosh Saleemi, Frank Zavisca and Lori Alexander: A comparison of the verbal rating scale and the visual analog scale for pain assessment. The Internet Journal of Anesthesiology. 2004: Volume 8, Number1.

3.Kellner L, Wimpissinger B, Stolba U, Brannath W, Binder S. 25-gauge vs 20-gauge system for pars plana vitrectomy: a prospective randomised clinical trial. Br J Ophthalmol. 2007 Jul;91(7):945-8.

4.Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 2001;357:1191-4.

Dear editor,

We read with interest the paper by Wimpissinger et al comparing sutureless 23-gauge system to a standard 20-gauge system for pars plana vitrectomy, in a randomized clinical trial.[1] We aim to highlight a few issues in the design, methodology and analysis.

The authors have analysed postoperative pain and conjunctival injection as primary outcome measures. The effort by the assessor to elicit a response on the degree of pain is likely to be biased if he is not masked, as in this study. In assessment, the authors have used descriptive terms like no pain, mild, moderate and severe pain. The authors could have used other standardized methods described to measure pain such as numerical scores, faces scale, visual analogue scales and multidimensional instruments like various questionnaires. Literature in Ophthalmology has precedence of using a visual analogue scale for assessment of pain.[2]

The authors determined the sample size by statistical pre-study evaluation (by two group t-test with a 0.05 two-sided significance level). Since the authors have not provided data on effect size (delta), and power (1-beta), it is not clear how a sample size of 60 with 30 patients in each arm was arrived at.

The other confounding variable likely to be of significance are the issues like randomization of 4 surgeons (instead of preferably single surgeon), use of different machines for 20Ga and 23Ga surgeries, unspecified cut rates for 20Ga system (likely to influence time of vitrectomy), the lens status (likely to influence the extent of vitrectomy possible and aimed at), type of tamponade (likely to influence postoperative intraocular pressures) and suturing technique of conjunctiva (likely to influence subjective pain) after 20Ga surgery.

In analysis, nominal type of categorical data (as used for assessing postoperative pain and conjunctival injection) would have been better analysed in terms of median and interquartile range rather than taking their minimum, maximum and average values.

The authors have concluded that "The 23 gauge system seems to be a safe system suitable for a broad spectrum of vitreoretinal diseases". We disagree with this conclusion as the data provided shows that 3 (10%, 95% CI -0.74 to 20.74%) patients developed serious complications, 2 of whom required re-surgery.[1]

An international group developed the Consolidated Standards for Reporting Trials (CONSORT) statement in an attempt to improve documented suboptimal reporting of RCTs.[3] In an attempt to further improve the reporting of RCTs, the same international group published the revised CONSORT statement in 2001 which outlined a checklist of 22 items.[4]

References

1. B Wimpissinger, L Kellner, W Brannath, K Krepler, U Stolba, C Mihalics and S Binder. 23-gauge versus 20-gauge system for pars plana vitrectomy: a prospective randomised clinical trial. Br J Ophthalmol 2008; 92; 1483-1487.

2. Mathew MR, Williams A, Esakowitz L, Webb LA, Murray SB, Bennett HG. Patient comfort during clear corneal phacoemulsification with sub- Tenon's local anesthesia. J Cataract Refract Surg 2003 ; 29(6): 1132-6.

3. Begg C, Cho M, Eastwood S, et al. Improving the quality of reporting of randomized controlled trials. The CONSORT statement. JAMA 1996;276:637-9.

4. Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 2001;357:1191-4.

Dear Editor:

I read with interest the article by Dr. Cohn, et al regarding the natural history of autosomal dominant optic atrophy (DOA). The authors describe an average of 10-year follow up for 69 patients with genetically confirmed DOA. In their study, 6 (9%) patients enjoyed improvement in visual acuity by 2 or more lines.

I found this surprising, and I wonder if the authors could provide further information regarding this group. Is it the opinion of the authors that these patients actually improved or that this may represent testing artifact or bias? Were they significantly younger than the rest of the cohort? Was their follow up significantly shorter? By how much did the acuities improve among this group? Was their baseline acuity more likely to come from outside records?

Sincerely,

Michael S Lee, MD
Minneapolis, MN