Statistics from Altmetric.com
Prospective rather than retrospective studies are more likely to be published
During the past two decades the essentially anecdotal nature of medical practice has been largely replaced by “evidence based medicine.”1 Evidence based medicine incorporates the most reliable reproducible data from clinical studies, particularly clinical trials. Indeed, the multicentre, prospective, randomised clinical trial has become the gold standard by which other clinical data are now judged. The impact of this change on medical publications has been profound and cannot be overemphasised. Prospective rather than retrospective studies are more likely to be published. Studies with inappropriate or no controls are often rejected outright. Appropriate use of statistics is now essential for publication of even the most straightforward clinical study. All of these changes are appropriate and make it more likely that data published today will still be useful in the future. On the other hand, do we really believe that no useful data can be obtained from a thoughtful small case series or even from the lowly case report?
In this issue of the BJO we introduce a new feature entitled “Hypothesis.” This feature will not be published in every issue of the journal, and we will not solicit papers for it. However, when one of the editors identifies a paper that raises what appear to be important clinical issues (we will not include any laboratory science studies) but does not contain all the necessary data to address the issues we will consider publishing the paper in this series. We will do so only after obtaining specific permission from the authors to publish it under this banner. We recognise that some authors may not want their work published in this format. Lambert and coworkers have agreed to have their paper “Weaning children with accommodative esotropia out of spectacles” (p 4) initiate the series. What are the important clinical questions raised by the authors and why might we not publish it under one of our usual headings?
Accommodative esotropia is usually divided into two distinct subtypes. One type results from an anomalous relation between the central innervational controls of accommodation and convergence in the presence of a modest need for accommodation (high AC/A ratio type).2 The second type occurs when there is a normal linkage of these functions that is overstressed by excessive demand (high hyperopia type).3 Both types of accommodative esotropia are treated by discouraging the accommodative innervational effort by providing the patient, at least initially, with their full cycloplegic hyperopic correction. The goal of the treatment is to reduce the esotropia to within 8 prism dioptres or less. This angle allows the development of at least peripheral fusion and probably an increase in fusional amplitudes.4 The usual practice is to only reduce the hyperopic correction as the patient’s refractive error becomes less with age. Raab and Spierer have reported, therefore, that the majority of their adolescent patients still require a spectacle correction to control their accommodative esotropia.5 Are there any risks in continuing this usual practice pattern?
The multicentre, prospective, randomised clinical trial has become the gold standard by which other clinical data are now judged
There is now convincing evidence that infant rhesus monkeys who wear plus lenses become more hyperopic as the result of doing so.6,7 In other words, providing the full hyperopic refractive error in a spectacle correction to an infant rhesus monkey interferes with the normal emmetropisation process whereby young animals normally become less hyperopic as they become older. Even before these experimental studies had been published Repka et al had warned that prescribing the full cycloplegic hyperopic correction in children with accommodative esotropia might result in a similar interference with normal emmetropisation.8 Therefore, it seems there is good reason to study whether some or all accommodative esotropes can be weaned aggressively from their hyperopic spectacles without compromising the management of their strabismus and related amblyopia.
Lambert and coworkers have studied a very small group of accommodative esotropes where they attempted to do just that. The study is retrospective and with historical controls only. The study population may not be representative of a large unselected accommodative esotropic population. The authors describe their study group 1 (six patients) as not being high hyperopes (range +1.75 to +2.50 dioptres) and yet only one of the patients had a high AC/A ratio. Moreover, their study groups included only patients who were orthotropic while wearing their hyperopic spectacles. Yet, we know that a significant number of accommodative esotropes will obtain peripheral fusion only with their spectacle correction and not foveal fusion.8 Despite these shortcomings the findings of this study are thought provoking.
All of the patients in the authors’ study group 1 were weaned out of their spectacles successfully without compromising ocular motor alignment. Weaning was commenced at a median age of 6.3 years and completed by median age of 9.0 years. In their not precisely comparable historical control group all three patients remained in hyperopic corrections (final correction +2.75 to +4.50 dioptres). In contrast, in the group (group 1) that was weaned from spectacles the peak refractive error was +3.25 dioptres and the final refractive error at the time of spectacle discontinuation was +1.56 dioptres (range 0.0 to +3.06 dioptres). All four children in group 2 who could not be weaned from their spectacles had a normal AC/A ratio but their baseline hyperopia was +4.50 dioptres (range +3.00 to +5.00 dioptres). This would seem to confirm a generally held clinical opinion that high hyperopic accommodative esotropes are not likely ever to be weaned from their spectacle correction. This may be due to the fact that they do not seem to “lose” their hyperopia as they get older. In this group the hyperopia peaked at +5.55 dioptres and at the completion of the study had only decreased to +5.05 dioptres.
The major difference between groups 1 and 2 in the study by Lambert and coworkers is that the hyperopic refractive error in group 1 fell within the normal range for non-strabismic children of a similar age. In contrast, group 2 consisted entirely of “high” hyperopes. I would again emphasise that conspicuous by its relative absence is a study group of patients with a high AC/A ratio but with “normal” levels of hyperopia for their age. Most large studies of accommodative esotropia document that patients with these clinical features constitute at least 50% or more of all patients with accommodative esotropia.8,9 Only one of the 10 patients studied by Lambert and coworkers exhibited a high AC/A ratio.
The authors recognise and discuss their study’s limitations with its “selection bias,” “small number of children,” and lack of randomisation. They, therefore, advise caution in interpreting the results of their study. We hope that the limitations of this study will not discourage the authors and others from further more detailed study of the interesting questions raised by it. Which accommodative esotropes can be weaned from spectacles and are they more likely to undergo normal emmetropisation than those who cannot be weaned from their spectacle correction? These are important clinical questions. We look forward to publishing the study that adequately details the answers.
Note in Proof
Prospective rather than retrospective studies are more likely to be published
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.