Statistics from Altmetric.com
Editor,—Fielder and Quinn1 have provided an interesting hypothesis and speculation on the origin of myopia in premature infants’ eyes following treatment for ROP. However, they have overstated the strength of the evidence in their opening statement that ‘. . . clinical research has shown that the degree of myopia is significantly less following laser therapy when compared with cryotherapy for severe ROP.’ The three references quoted contain evidence only from historical comparison groups, raising questions about the strength of the evidence for this conclusion.2-4
The reports describe the incidence4 or severity2 of myopia in two studies which compared laser with cryo treated ROP. Patient allocation of these 66 infants to the two treatments was not random, with the earlier historical comparison groups receiving cryotherapy. The third report3 describes 13 laser treated infants with no comparison group. Infants who lost vision were largely unreported. (In our own literature search a fourth report5 describes myopia outcomes in 17 infants treated with cryotherapy and 26 infants treated with laser therapy, but these are probably included among the infants reported by Algawi et al.4 An additional full report of reduced myopia following laser compared with cryotherapy was identified and the English abstract suggests a historically controlled comparison (Italian not yet translated),6 while a final abstract reports less myopia following laser in 14 infants compared with 20 non-randomly assigned cryotherapy infants7.)
Two methodological issues therefore raise a significant potential for biased conclusions: historical controls and non-reporting of treatment failures.
The criteria for choosing infants for treatment in clinical practice are constantly under evolution as experience and judgment leads to potentially better choices of the time for treatment. Thus, the earlier treated (cryo) infants are more likely to have had more severe disease and a poorer overall prognosis and more sequelae than the later treated infants (laser). The effect of this type of bias would be that infants receiving laser treatment (compared with cryo treatment) would have less severe ROP at the time of treatment, a better prognosis because more of these would have regressed even without cryotherapy, and therefore they would be expected to have less myopia, even if there were no true myopia effect of cryo v laser treatment.
The number of infants who progressed to retinal detachments following either treatment is critical to evaluate because only infants with attached retinas can be assessed for myopia. If cryotherapy ‘rescued’ eyes that would have detached but leaves them myopic, that would still be a better outcome than if laser therapy failed to salvage those eyes but no myopia was found in the fewer eyes that did remain attached. It might be clearer to think of an analysis of the proportion of the originally treated eyes that were not myopic or detached at follow up. Of course, this type of bias is unlikely to be a problem in this case, since randomised controlled trials of cryo v laser therapy have shown laser therapy to be at least as effective as cryotherapy in preventing unfavourable retinal outcomes.8-10
We agree with Fielder and Quinn that these are important observations, but they are not yet strong conclusions. We can hope that laser will reduce the sequelae of severe ROP even more than does cryotherapy, but a thorough search of the existing literature reveals only 90 infants entered in randomised, concurrent comparison trials of cryotherapyv laser ablation, and none of these has reported myopia outcomes to date.8-10 Given the consistency of observations from the non-randomised studies which used either historical controls247 or no controls,3it may well be that there is a true reduction in the rate of myopia following laser. However, the validity of this conclusion and the accurate estimation of the magnitude of such an effect depend on data from randomised controlled trials. The investigators who conducted those controlled trials should seek out and publish myopia outcomes from those children to permit a best estimate of potential true differences. If the hypothesised differences in myopia rates are true, there are important questions for basic science laboratories to answer, and clear clinical practice implications.
Editor,—Upon review of our recently published editorial ‘Myopia of prematurity: nature, nurture, or disease?’ based on an article by Laws et al, comparing refractive outcomes in eyes that received laser vcryotherapy for severe retinopathy of prematurity (ROP), we are concerned that we may have conveyed two false impressions:
(1) The data presented by Laws et al do not necessarily ‘confirm’ in a scientific sense the finding that lasered eyes had less myopia than eyes that underwent cryotherapy for severe ROP. Rather the study by Laws et al and the other two studies they cited were unrandomised, uncontrolled studies and they all contributed to the suggestion that lasered eyes may be less likely to have severe degrees of myopia than those which are treated by cryotherapy.
(2) Laws et al found that the ‘cryoed’ eyes were highly myopic without citing data on the natural course of untreated eyes which regress after developing severe retinopathy. As we indicated in our editorial, the distribution in cryo treated and control eyes in the CRYO-ROP study was virtually the same for eyes with myopia up to 6 dioptres and hyperopic eyes. This suggests that the abnormal refractions are probably due to the disease of ROP and not the surgical intervention of cryotherapy. Finally, we would reiterate the point that there is ‘vital work to do not only with regard to refractive development’, but also to improve our care of the eyes of these preterm infants.
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.